Split (1)

train · 5.09k rows

pmcid stringlengths 6 6	title stringlengths 9 374	abstract stringlengths 2 4.62k ⌀	fulltext stringlengths 167 106k	file_path stringlengths 64 64
514614	The new reconstruction technique in the treatment of the skin cancers located on the eyelid: Posterior temporalis fascia composite graft	Background Difficulty of reconstruction of the eyelids arises from the need to reconstruct different supporting and covering structures in a single operation. Defects in the anterior lamella of the eyelids can be readily repaired with skin grafts or flaps but posterior lamellar reconstruction needs more complex applications. Methods We performed posterior lamellar eyelid reconstruction with posterior parts of the temporalis fascia, since their anatomical and histological features are very similar to the defects. Nine patients with skin tumors located on the periorbital region were treated with local skin flaps and deep layer of the temporalis fascia. Results Grafts were harvested very easily. There was no complication related with graft or donor site. Biopsy was performed in three cases and normal conjunctival elements were seen. Functional and acceptable aesthetically results were achieved in all patients. Conclusion Ideal reconstructive material for replacement of the posterior lamina is still lacking. Tarsal reconstruction can be made with deep temporalis fascia with success since the thickness of the both tissues are very similar and also since the loose areolar layer of the temporalis fascia is very thin and highly vascularized, this layer can be used in reconstruction of the conjunctiva. According to our knowledge this is the first report of using of the posterior part of temporalis fascia as a composite graft in the literature.	Introduction Reconstruction of eyelid defects after tumor excision should aim at obtaining full globe protection without visual disruption and restoring the area to an appearance as close to normal as possible [ 1 ]. Reconstruction of the eyelids requires special considerations and complete understanding of the specialized anatomy of the region. The eyelids consist of an anterior lamella of skin, orbicularis muscle; posterior lamella of tarsus and conjunctiva. Full thickness defects of eyelid after tumor resection require reconstruction of these layers. However reconstruction of the skin and subcutaneous tissue can be easily reconstructed with skin grafts and local flaps, most important subject is reconstruction of the posterior lamellar segment of the eyelid. Tarsal plate is dense, fibrous tissue (not a cartilage!) that gives the eyelid its contour and provides its skeleton [ 2 ]. Tarsal substitutes including banked sclera, nasal cartilage, ear cartilage, and periosteum can be beneficial for posterior lamellar repair [ 3 ]. Reconstruction of the conjunctiva is more complex issue. Tarso-conjunctival grafts, buccal mucosa, hard palate mucosa and amniotic membrane have been used in the reconstruction of the conjunctival defects and all techniques were reported with individual advantages and disadvantages. Temporalis fascia has a specific and complex anatomy [ 4 - 6 ]. Mainly temporalis fascia consists of two layers: superficial temporalis fascia and deep temporalis fascia. The superficial temporalis fascia is separated from the deep temporalis fascia by a distinct plane of loose areolar tissue: loose areolar layer [ 5 ]. In our study we performed posterior lamellar reconstruction of the lower eyelid defects with loose areolar layer for conjunctiva and deep temporalis fascia for tarsus in the patients who required total eyelid reconstruction secondary to the skin tumors. Loose areolar layer (subaponeurotic layer or subgaleal fascia) has highly vascularized histological structures and this feature may allow growing of the bulbar conjunctiva very easily [ 4 ]. Since the deep layer of temporalis fascia is very similar to tarsus, reconstruction with deep layer temporalis fascia provides good structural stability as original skeleton. The cosmetic and functional outcomes of our technique are encouraging. Methods Nine patients with lower eyelid defects after removal of malignant tumors were treated with local flaps combined with temporalis fascia grafting in the Department of the Plastic and Reconstructive Surgery, Adnan Menderes University Hospital. Data of the patients were summarized in Table 1 . Table 1 Clinical cases No: Age Sex Lesion Size of defect (mm) Flap type Pathology Complication 1 52 Female Left lower eyelid 20 × 5 Mustardé Basal cell Ca None 2 66 Male Left lower eyelid 38 × 12 Mustardé Basal cell Ca None 3 60 Male Right lower eyelid 28 × 8 Mustardé Basal cell Ca None 4 48 Male Right lower eyelid 26 × 10 Mustardé Basal cell Ca None 5 36 Female Inner right canthus 40 × 10 Tri-lobed flap Basal cell Ca None 6 56 Female Left lower eyelid 22 × 8 Bi-lobed flap Basal cell Ca None 7 71 Female Right lower eyelid 45 × 10 Tri-lobed flap Basal cell Ca None 8 52 Male Left inner canthus 42 × 15 Forehead flap Basal cell Ca None 9 21 Female Right inner canthus 50 × 15 Forehead flap Basal cell Ca None Ages of patients were ranged from 21 to 71. Full thickness defects measured from 15 × 5 mm to 50 × 15 mm. All patients were operated under the general anesthesia. The follow up periods ranged from 6 months to 24 months. After ethical approval from our department and patient consent, biopsy performed in three cases to evaluate the technique. Surgical Technique Tumor excision was performed with safety surgical margins in all patients. Mustardé cheek flap was prepared in four patients, forehead flap and tri-lobed flap in two patients and one bi-lobed flap were designed. "Y" skin incision was performed for harvesting of the temporalis fascia (Figure 1a,1b and 1c ). Figure 1 Schematic view of the graft harvesting. 1a skin incision, 1b: fascia incision 1c: harvesting of the fascia. Against possible hair damage we did not use solution containing of epinephrine for haemostasis. The scalp is incised at angle that follows the direction of the hair follicles, which is facilitated by not shaving the hair. The incision is made through the scalp to the superficial fascia. Superficial fascia was incised to the temporalis muscle in dimensions 40 × 20 mm. This layer was elevated above the loose areolar tissue as distal based flap and loose areolar layer and deep temporalis fascia was dissected and harvested (Figure 2 ). Figure 2 Three layers of the fascia of temporalis region. From above to below: superficial temporalis fascia, loose areolar fascia and deep layer. Loose areolar layer and deep layer can be harvested easily. Superficial fascia was re-located in its original location. Fascia was sutured with 4-0 vicryl. Haemostasis was performed with bipolar cautery. Scalp was closed with 3-0 nylon sutures. Harvested fascia consisted of the layers: loose areolar tissue and deep temporalis fascia. Continuity of these two layers were not disrupted and used as a composite graft. Thin and membranous loose areolar tissue was used for conjunctiva reconstruction and dense and firm deep temporalis fascia was used for tarsus reconstruction. These two layers were adapted to the flap with 6-0 catgut sutures. Flaps were transposed to the defects. Loose areolar tissue was sutured to the edge of palpebral conjunctiva with continuous 7-0 vicryl. One edge of the deep temporalis muscle was sutured to the periosteum, which is located inside the lateral orbital rim and the other edge sutured to the medial canthal tendon with 4-0 nylon. Prophylactic antibiotics were used topically in post-operative first week. Results All patients tolerated the operation well (figure 3 , 4 , 5 ) Figure 3 View of the patient. 3a: Patient with skin tumor located on lower eyelid. Tumor excision and Mustardé cheek flap with posterior temporal fascia graft were applied to this patient. 3b: Post-operative six months view of the reconstructed conjunctiva with loose areolar layer. Figure 4 Post-operative six months of the patient. Conjunctiva was reconstructed with "loose areolar layer" of the temporalis fascia; tarsal plate was reconstructed with deep temporalis fascia. (Patient pulled down his eyelid with his finger) Figure 5 Figure View of the patient. 5a: Patient with skin tumor located on inner canthus and lower eyelid. Tumor excision and bi-lobed forehead skin flap with posterior temporal fascia graft were applied to this patient. 5b: View after excision of the tumor. 5c: Harvesting of the fascia for posterior lamella reconstruction. 5d: Post-operative six months view of the patient. There was no complication. Patient did not accept the further revision. There was no early or late complication in donor site and flap (Figure 6 ). Figure 6 Post-operative six months view of the donor site. There is no scar. We observed neither infection nor irritation signs nor symptoms. There was no graft lysis. Patients were followed minimum six months. There were no complaints in this period in the patients. In all patients, the functional and aesthetic results were achieved. There were no signs and symptoms related with shrinkage of the grafts. Microscopically normal conjunctival elements were seen in the biopsy of the reconstructed conjunctiva with Hematoxylene-eosin staining (Figure 7 ). Figure 7 Biopsy form grafted area of the lower eyelid. Hematoxylene-eosin staining, × 100 magnification. Discussion Difficulty of the reconstruction of eyelids arises from the need to reconstruct different supporting and covering structures, i.e. the conjunctiva, tarsus, orbicularis muscle, canthal ligaments and skin. Many flaps and reconstruction techniques were described in the literature for anterior lamellar part of the eyelids. In our study, we used the posterior part of the temporalis fascia for the posterior lamellar part of the eyelids. Fox and Edgerton first used the fascia of the temporalis region in reconstructive surgery [ 7 ]. Brent and Byrd used the temporoparietal fascia for ear reconstruction and they also stated that these anatomical structures could be used as a free microvascular autograft [ 8 ]. Temporalis fascia widely used ophthalmic reconstructive material in plastic surgery especially in socket reconstruction as a flap [ 9 ]. This tissue were also used in open rhinoplasty, facial paralysis, Peyronie disease, reconstruction of temporomandibular joint, repair of the perforation of nasal septum, lip augmentation and finally malar augmentation were reported in the literature [ 10 - 16 ]. Superficial layer of the fascia was mainly used in all these studies. But temporalis region has a specific fascial anatomy. In the temporoparietal region, there are four and in some places five different layers, excluding the skin, subcutaneous tissue and the temporalis muscle [ 6 ]. Although many reports were presented in the literature about the anatomy of the fascia of temporoparietal region, there is no consensus in terminology. There are several names in current use for each layer. In our study we used posterior part of the tissue: subaponeurotic plane and deep temporal fascia. The superficial temporal fascia is separated from the deep temporal fascia by distinct plane of loose areolar tissue [ 5 ]. This layer has been termed the "loose areolar layer" or "subaponeurotic layer" or "subgalael fascia". This tissue is well developed and highly rich supplied by the branches from the superficial temporal artery. It is easy to dissect as a discrete layer 6 . However we performed our operation under the general anesthesia, Miller pointed that the temporalis area can be easily anesthetized by infiltration local anesthesia [ 17 ]. The graft can be obtained quickly under the direct vision. There is minimal amount of postoperative pain and no visible scar. We speculate that highly vascularized histological structure of the loose areolar tissue allows growing of the bulbar conjunctiva very easily and conjunctival elements were seen in histological examinations in the biopsy from the graft in post-operative six months. Tolhurst et al. presented detailed anatomical research about the subgaleal fascia [ 6 ]. They summarized the advantages of this layer as: (a) this layer is thin and will conform to the shape of underlying soft tissue or cartilage with aesthetically pleasing fidelity, (b) well vascularized, and if handled with care, will readily support split and full thickness skin graft, (c) can be harvested easily, with minimal donor-site morbidity. We used this anatomical structure for conjunctiva reconstruction. Conjunctiva is a mucous membrane that covers the posterior aspect of the eyelids (palpebral conjunctiva) and the anterior surface of the globe (bulbar conjunctiva). Defects larger than 25% of the eyelid usually require a free graft to make for the loss of specialized tissue. Conjunctiva harvested from another lid is the ideal match and physiologically but it is thin, difficult the handle, has tendency to contract, and can only be harvested sparingly to avoid interfering to donor fornices. In addition this technique requires two operations when it is performed as a lid sharing procedure. Oral mucosa is abundant and simple to remove but tends to contract and infection. Nasal mucosa is easier to handle, and minimal contraction but this procedure has some problem of poor access and thickness. Today, most performed techniques are hard palate mucosal graft, amnion membrane, chondormucosal Septum [ 18 - 20 ]. Tarsal plate was reconstructed with deep layer of the temporalis fascia in our study. The deep temporalis fascia is a white, dense, though, uniform fascial layer similar in strength, appearance, and thickness to the sheet of the rectus abdominis muscle [ 5 ]. The tarsal plates are composed of dense fibrous connective tissue that provides structural support to the eyelid [ 2 ]. Matsumoto et al used the fascia lata for the reconstruction of the tarsal plate [ 21 ]. They performed this operation with cheek flap for skin coverage, and buccal mucosa graft for the conjunctiva defect. They applied two different graft materials (mucosa and fascia) to the flap. One of the most popular grafts in posterior lamella reconstruction used to come from the chondromucosal nasal septum. Major disadvantages of this technique are the difficulty of the shape and adapt to the eyelid. Palatal mucosal grafts were widely used. Such grafts meet both the mucosal and supporting requirements. The use of hard palate mucosal grafting in lower eyelid reconstruction was first described by Siegel [ 22 ]. Hard palate is ideal material for posterior lamellar reconstruction but this technique has serious complications such as: infection, oronasal fistula, and post-operative discomfort. Also increased operating time for graft harvest and occasional keratinization of the surface with potential ocular surface irritation could be listed as disadvantages of the technique [ 23 ]. In our series we did not observe irritation signs and symptoms. It is agreed among most authors that the ideal reconstruction should use identical or similar tissue to that of the original structure. Undoubtedly, reconstructions of the posterior lamella with original tissue i.e. free tarso-conjunctival grafts or flap is the best treatment modality. If the donor sites of the tarso-conjunctival grafts or flap were not suitable or enough, deep temporalis fascia grafting would be an alternative in the reconstruction of the posterior lamella. We believe that application of the deeper part of the temporalis fascia for the reconstruction of the posterior lamella has many advantages: easy to perform, single composite graft for two layers, without or minimal donor site morbidity, no infection risk, no shrinkage of the graft, easy to adapt, wide donor site, no irritation and excellent cosmetically results. There is no data about the using of the posterior part of the temporalis fascia in reconstruction of the tarsus and conjunctiva in the literature. In conclusion, loose layer of the temporalis fascia is ideal tissue for the growing of the conjunctiva and the thickness of the deep layer of the temporalis fascia is very similar to the tarsus. Both tissues can be used in reconstruction of the posterior lamellar reconstruction of the eyelids with success. Competing interests None declared Authors' contributions EC conceived of the study, and participated in its design and coordination. NS participated in design of study. All authors read and approved the final manuscript.	/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC514614.xml
423133	Plasticity of DNA Replication Initiation in Epstein-Barr Virus Episomes	In mammalian cells, the activity of the sites of initiation of DNA replication appears to be influenced epigenetically, but this regulation is not fully understood. Most studies of DNA replication have focused on the activity of individual initiation sites, making it difficult to evaluate the impact of changes in initiation activity on the replication of entire genomic loci. Here, we used single molecule analysis of replicated DNA (SMARD) to study the latent duplication of Epstein-Barr virus (EBV) episomes in human cell lines. We found that initiation sites are present throughout the EBV genome and that their utilization is not conserved in different EBV strains. In addition, SMARD shows that modifications in the utilization of multiple initiation sites occur across large genomic regions (tens of kilobases in size). These observations indicate that individual initiation sites play a limited role in determining the replication dynamics of the EBV genome. Long-range mechanisms and the genomic context appear to play much more important roles, affecting the frequency of utilization and the order of activation of multiple initiation sites. Finally, these results confirm that initiation sites are extremely redundant elements of the EBV genome. We propose that these conclusions also apply to mammalian chromosomes.	Introduction Biochemical studies performed in higher eukaryotes have shown that DNA replication initiates at specific sites, or within initiation zones, suggesting the involvement of particular DNA sequences called replicators (reviewed by DePamphilis 1999 ). In contrast, functional studies, as well as studies of DNA replication performed in early embryos of various vertebrates and invertebrates, have suggested that initiation of DNA replication can take place with limited sequence specificity (reviewed in Gilbert 2001 ). The presence of specific initiation sites and of initiation zones has also been proposed to explain the latent replication of the Epstein-Barr virus (EBV) genome in human cell lines. During latent replication, the EBV genome is maintained as a circular episome (∼175 kb in size), and the host cell provides both the replication machinery and the licensing apparatus that limit the genome's duplication to once per cell cycle (reviewed in Kieff 1996 ; Yates 1996 ). Initiation site oriP was the first initiation site identified in the EBV genome. In the presence of the viral protein EBNA1, this DNA sequence confers autonomous replication to plasmids transfected into human cell lines ( Yates et al. 1984 ). In addition, initiation of DNA replication at oriP was recently shown to be regulated by geminin, and to correlate with the binding of various cellular components of the replication complex (Orc1, Orc2, Orc3, Orc4, Orc6, Mcm2, Mcm3, and Mcm7) ( Chaudhuri et al. 2001 ; Dhar et al. 2001 ; Schepers et al. 2001 ; Ritzi et al. 2003 ). These and other reports have been interpreted as evidence that oriP contains a replicator (e.g., Koons et al. 2001 ). However, other initiation sites have also been described ( Kirchmaier and Sugden 1998 ), and a study performed by two-dimensional (2D) gel electrophoresis at neutral pH has suggested the presence of a large initiation zone ( Little and Schildkraut 1995 ). In addition, reports from different laboratories have shown that various portions of the EBV genome, including oriP, can be deleted without affecting the maintenance of the episomes in replicating cells (see Discussion and references therein). Therefore, the presence of specific replicator sequences and their relationship with the sites of initiation of DNA replication also remain to be demonstrated in this system. We recently began to study the replication of individual EBV episomes using fluorescence microscopy ( Norio and Schildkraut 2001 ). In a previous study, we collected various images of the Raji EBV genome ( Norio and Schildkraut 2001 ). The analysis of those molecules demonstrated that the duplication of different EBV episomes begins at different initiation sites located within the initiation zone identified by 2D gel electrophoresis. However, the number of molecules analyzed was not sufficient to infer the precise dynamics of activation of the initiation sites (i.e., to detect events having a short life or occurring infrequently during the duplication of the episomes). In the present study, we performed an extensive analysis of the replication dynamics of the EBV genome in two human Burkitt's lymphoma cell lines (Raji and Mutu I). By utilizing a different procedure to stretch DNA molecules we were able to collect a large number of images of the EBV genome representative of different stages of duplication. This allowed us to determine how DNA replication initiates, progresses, and terminates throughout the EBV genome and to precisely measure the duplication time of specific portions of the EBV genome. These improvements allowed us to obtain important new results as well as to extend previous observations. Here we show that initiation events are not limited to a specific portion of the EBV genome (namely the initiation zone detected by 2D gel electrophoresis), but, unexpectedly, take place throughout the EBV genome. Multiple initiation events were also detected in individual EBV episomes. Hence, if the initiation sites do correspond to replicators, the latter must necessarily be highly redundant (present at a frequency of one or more every 20 kb). Our new results also indicate that, in these two EBV strains, both the frequency and the order of activation of the initiation sites vary considerably throughout the viral genome. This variation involves initiation sites such as oriP, the sequence of which is highly conserved in the two EBV strains ( Salamon et al. 2000 ). Hence, the utilization of the initiation sites is largely independent of their DNA sequence, and it is affected by the genomic context (i.e., the presence/absence of initiation sites activated earlier or the presence of transcription). Finally, we noticed that the initiation sites that tend to be activated earlier during the duplication of each episome are located in clusters, each of which spans several kilobases. The locations of these clusters are different in the Raji and Mutu I strains. Therefore, the utilization of the initiation sites (particularly their order of activation) appears to be regulated at the level of genomic regions rather than at the level of individual initiation sites. Results Fluorescent Hybridization Immunostaining of Individual EBV Episomes Stretched on Microscope Slides In order to study DNA replication, we used a procedure that we call single molecule analysis of replicated DNA (SMARD). This procedure labels the replicating DNA in a way that allows us to determine the position, the direction, and the density of the replication forks in a steady-state population of replicating molecules (in this case, EBV episomes). This in turn allows us to determine how DNA replication initiates, progresses, and terminates throughout the genomic region analyzed. SMARD presents several advantages over procedures that utilize different labeling schemes and allows us to overcome most of the limiting factors that have traditionally affected studies of replication performed on DNA fibers (see Materials and Methods ). In our procedure, an asynchronous population of cells is sequentially labeled with 5′-iodo-2′-deoxyuridine (IdU) and 5′-chloro-2′-deoxyuridine (CldU) ( Norio and Schildkraut 2001 ). The length of each labeling period is longer than the time required to completely replicate the EBV genome (3.5–4 h; see Materials and Methods ). This allows some of the replicating EBV episomes to become substituted with the halogenated nucleotides along their entire length ( Norio and Schildkraut 2001 ). The incorporation of these nucleotide analogs is later detected by immunofluorescence of individual DNA molecules stretched on microscope slides. In these molecules, the transitions from IdU to CldU mark the positions of the replication forks at the time of the switch from the first to the second labeling period (see below). Hence, the results of this analysis are presented as a series of images of EBV episomes representative of the different stages of duplication that were present at the end of the first labeling period. In addition, the use of long labeling periods makes the data collected by SMARD suitable for quantitative analysis, allowing us to calculate the duplication time of different genomic regions. In the experiments described in this study, agarose-embedded total DNA was prepared from cells labeled with halogenated nucleotides. The circular EBV episomes were converted to linear molecules by digestion with a restriction endonuclease (PacI or SwaI). After pulsed field gel electrophoresis, the EBV DNA was recovered by agarase treatment ( Norio and Schildkraut 2001 ) and stretched on microscope slides by capillary action (see Materials and Methods ). Using this procedure we obtained relatively high numbers of stretched molecules even when the starting amount of purified DNA was very small. The hybridization of specific biotinylated probes (visualized with Alexa Fluor 350-conjugated avidin; shown in blue in the figures throughout the manuscript) was used to identify the EBV molecules and their orientation ( Norio and Schildkraut 2001 ). In addition, the halogenated nucleotides were visualized using specific monoclonal antibodies and secondary antibodies conjugated with Alexa Fluor 568 (shown in red in the figures throughout the manuscript; IdU) and Alexa Fluor 488 (shown in green in the figures throughout the manuscript; CldU). The detection procedure and the analysis of the images are described in Materials and Methods and in Figures 1 and 2 . The use of long labeling periods, and the analysis of molecules substituted with the halogenated nucleotides along their entire length, present several advantages. In particular this procedure provides multiple internal controls that could not have been performed if short labeling times had been used (see Materials and Methods ). Figure 1 Fluorescent Hybridization Immunostaining of Individual EBV Episomes Image of two stretched DNA molecules in the same optical field. The hybridization signals (p107.5 and pSalF) and the immunostaining to detect the halogenated nucleotides are shown in different pseudocolors (red = IdU, green = CldU, and blue = hybridization probes). The top panel shows the merged image. The different color channels are shown separately in the lower panels. One of the two stretched molecules is a PacI-linearized EBV episome (molecule above) and can be recognized by the presence of the hybridization signals. The molecule below is a piece of cellular genomic DNA of similar size (no hybridization signals). The presence of the hybridization signals decreases the intensity of the immunostaining along the same portion of the EBV episome. This confirms that both halogenated nucleotides and hybridization probes are located on the same DNA molecule. The blue dots visible in the bottom panel represent hybridization background (this background was digitally removed from Figures 3B, 4B, and 6B). The EBV episome is substituted along its entire length with both IdU (red regions) and CldU (green regions). Yellow arrowheads indicate the approximate position of the replication forks at the time of the switch from the first to the second labeling period. The background visible in the red and green channels is mainly other DNA molecules containing halogenated nucleotides (white horizontal arrowheads). Some of these molecules attached to the glass before becoming fully extended and appear thick, displaying a brighter immunostaining. Small dots are also visible (magenta vertical arrowheads), sometimes overlapping with the DNA molecules (white asterisks); however, they were not considered in our analysis because they are too short to be unequivocally ascribed to DNA replication. Figure 2 Stretching DNA Molecules by Capillary Action (A) Lengths of 219 unbroken Raji EBV episomes with a recognizable hybridization pattern. These molecules were stretched by the movement of a DNA solution between a silanized microscope slide and a nonsilanized coverslip. About 94% of these molecules have a size of 70 μm (±15 μm), corresponding to about 2.4 kb/μm. (B) Schematic of the PacI-linearized Raji EBV genome with the positions of various genetic elements shown to scale. The initiation site oriP is shown in green with the FR element and the DS element indicated by green boxes. The positions of EBER genes (black box), terminal repeats (smaller red box), internal repeats 1 (larger red box), and the restriction sites utilized in this study (PacI and SwaI) are also indicated. (C) Images of 6 PacI-linearized Raji EBV episomes aligned with the EBV map after hybridization and immunostaining of the DNA molecules and digital adjustment of length. The hybridization signals are shown in blue. Immunostaining to detect the halogenated nucleotides is shown by red and green. Vertical light blue lines indicate the positions of the ends of the hybridization probes and yellow lines, the position of the PacI site used to linearize the EBV episomes. All the molecules shown in this figure represent EBV episomes duplicated during either the first labeling period (red) or the second labeling period (green). The quality of the alignment of the images with the EBV map is comparable to the alignment previously obtained with combed EBV episomes ( Norio and Schildkraut 2001 ). The resolution of analysis is limited to about 5 kb primarily because of discontinuities in the fluorescent signals, as previously reported for similar assays ( Parra and Windle 1993 ; Jackson and Pombo 1998 ; van de Rijke et al. 2000 ). The Raji EBV Genome Contains a Region That Usually Replicates First and a Region That Usually Replicates Last In order to define precisely how the Raji EBV genome replicates, we recovered the images of 245 PacI-linearized EBV episomes that incorporated halogenated nucleotides along their entire length (112 fully stained in red, 84 fully stained in green, and 49 stained in both red and green). The results of this experiment are shown in Figure 3 . In the episomes that incorporated both kinds of halogenated nucleotides, the red to green transitions (arrows in Figure 3 B) define the position of the replication forks at the time of the switch from the first to the second labeling period. The red portions of these molecules are nested around the ends of the PacI-linearized episomes ( Figure 3 B). This indicates that DNA replication proceeded in a similar manner in most of episomes. Figure 3 SMARD Performed on PacI-Linearized EBV Episomes Replicated in Raji Cells (A) Map of the PacI-linearized EBV genome with the positions of various genetic elements shown to scale. Below the EBV map, light blue bars indicate the positions of the hybridization probes (p107.5 and pSalF) utilized during SMARD to identify the molecules of interest and their orientation. Gray bars (a–h), and black bars (1–10), indicate the positions of the restriction fragments analyzed by 2D gel electrophoresis. (B) PacI-linearized Raji EBV episomes after hybridization and immunostaining of the DNA molecules (aligned with the map). These molecules incorporated both halogenated nucleotides, and the images are ordered (from 1 to 48) by increasing content of DNA labeled during the first labeling period (red). One additional molecule was unsuitable for precise measurements and is not shown. Vertical light blue lines indicate the positions of the ends of the hybridization probes and yellow lines, the position of the PacI site. Arrowheads mark the approximate position of the red-to-green transitions. Asterisks indicate the position of short colored patches not necessarily related to DNA replication. (C) Replication profile of the Raji EBV episomes. This profile was obtained using both the images shown in (B) and the images collected in a previous SMARD experiment ( Norio and Schildkraut 2001 ), for a total of 69 episomes. Starting from the PacI site, genomic intervals of 5 kb are indicated on the horizontal axis by numbers from 1 to 35. The vertical axis indicates the percentage of molecules stained red within each 5-kb interval. (D) Profile of replication fork abundance and direction throughout the EBV genome. Genomic intervals of 5 kb are indicated on the horizontal axis as for (C). The vertical axis indicates the percentage of molecules (out of a population of 69 episomes) containing replication forks (red-to-green transitions) within each 5-kb interval. The forks moving from left to right are depicted in orange. The forks moving from right to left are depicted in yellow. (E) Map of the EBV genome aligned with the horizontal axes of histograms (C) and (D), and with the restriction fragments analyzed by 2D gel electrophoresis (black and gray bars below the map). Green I s indicate the presence of replication bubbles. Red T s indicate the presence of replication intermediates produced by random termination events. Replication bubbles were detected by 2D gel electrophoresis across the region marked by the red dashed line (approximately corresponding to the RRF). (F) Transcription of the Raji EBV genome. Red arrows mark the positions of regions that can be transcribed during latency. The level of transcription derived by nuclear run-on according to Kirchner et al. (1991) is shown as gray scale (black = highest level; white = lowest level or not transcribed). The EBER genes represent the most intensively transcribed portion of the EBV genome. Intermediate levels of transcription were detected across and downstream from the long transcription unit of the EBNA genes. According to Sample and Kieff (1990) , the level of transcription along the EBNA genes region decreases from left to right (I–III). Intermediate levels of transcription were also reported for the two hatched regions. However, these regions contain either repeated sequences (the terminal repeats) or cross-hybridize with other transcribed regions (oriLytR and oriLytL); therefore, their actual level of transcription could be lower. However, the progression of DNA replication throughout the EBV genome is better described by the replication profile of the molecules analyzed ( Figure 3 C). This profile was obtained by dividing the map of the episomes into intervals of 5 kb (horizontal axis) and then indicating the percentage of molecules stained in red within each interval (vertical axis). From this profile we can easily identify a genomic region that usually replicates first (RRF; more frequently stained in red), a genomic region that usually replicates last (RRL; less frequently stained in red), and two transition regions. The RRF contains the initiation sites most frequently utilized to begin the duplication of the Raji EBV episomes. More than 80% of the molecules analyzed were stained in red throughout intervals 1–7 and 31–35 (horizontal axis; Figure 3 C). In the molecules representing early stages of episomal duplication (upper portion of Figure 3 B), initiation events took place either within the RRF (molecules 2–21) or in adjacent portions of the EBV genome (i.e., molecule 1). Interestingly, low levels of replication bubbles had been previously detected by 2D gel electrophoresis within various restriction fragments located in the same portions of the EBV genome (black bars 1–3 and 5–7 in Figure 3 A). Therefore, the initiation sites activated earlier during the duplication of each episome are located within what appears to be an initiation zone that spans several tens of kilobases (encompassing intervals 1–7 and 27–35 in Figure 3 C). The RRL appears in the replication profile of the Raji EBV episomes as a large valley ( Figure 3 C). The bottom of the valley spans about 40 kb (intervals 11–18), and its flat appearance indicates that throughout this region the episomes terminate their duplication with similar probability. Note, however, that termination events can also occur within the transition regions (green in molecules 43 and 48; Figure 3 B). Interestingly, the level of transcription across the RRL is higher than in the rest of the Raji EBV genome, while across the RRF it is either very low or absent ( Figure 3 F) ( Sample and Kieff 1990 ; Kirchner et al. 1991 ). The presence of RRF and RRL was confirmed by a second SMARD experiment in which we digested the EBV episomes with SwaI. This enzyme cleaves twice in the viral genome, producing fragments of 105 and 70 kb. The larger fragment was expected to contain most of the RRF (now located near the center of DNA molecules), and a small portion of the RRL. We recovered 209 fully substituted 105-kb fragments (94 red, 79 green, and 36 red and green). These molecules were analyzed as described for the PacI-linearized EBV episomes ( Figure 4 ). We found that both the RRF (intervals 1–13), and the RRL (intervals 17–21) encompass the same genomic sequences occupied in the previous SMARD experiment (compare Figures 3 C and 4 C). Initiation events located within the RRF are visible in molecules 1–4 ( Figure 4 B). We conclude that the results obtained by SMARD are reproducible and do not depend on the particular restriction enzyme used for digesting the DNA molecules. Figure 4 SMARD Performed on SwaI-Digested EBV Episomes Replicated in Raji Cells (A) Map of the approximately 105-kb fragment obtained by digesting EBV episomes with the restriction endonuclease SwaI. (B) Images of 36 EBV molecules ordered and marked as in Figure 3 B. Molecules 18 and 19 are distorted, but the positions of the red-to-green transitions are clear. (C) Replication profile of the SwaI-digested EBV episomes shown in (B). Starting from the SwaI site, intervals of 5 kb are indicated on the horizontal axis by numbers from 1 to 21. The vertical axis indicates the percentage of molecules stained red within each 5-kb interval. (D) Profile of replication fork abundance and direction. Intervals of 5 kb are indicated on the horizontal axis as for (C). The vertical axis indicates the percentage of molecules containing replication forks in each 5-kb interval. The partitioning of the EBV genome is different from Figure 3 D. Hence, the four pausing sites that in Figure 3 D were located within interval 8 are here located between interval 13 and interval 14. As a consequence, the peak visible in Figure 3 D is here split into two smaller adjacent peaks. (E) Map of the approximately 105-kb SwaI fragment aligned with the horizontal axes of histograms (C) and (D). Replication Forks Move Without Significant Pausing throughout the Raji EBV Genome with the Exception of the Genomic Region near oriP The movement of the replication forks throughout the EBV genome is described by the profiles of replication fork abundance (see Figures 3 D and 4 D). These profiles were obtained by dividing the map of the EBV genome into intervals of 5 kb (horizontal axis) and then indicating the percentage of molecules containing red-to-green transitions within each interval (vertical axis). As seen earlier, these transitions indicate the position, and the direction, of the replication forks at the time of the switch from the first to the second labeling period. The significant accumulation of red-to-green transitions visible within interval 8 ( Figure 3 D) indicates that replication forks were not moving freely across this portion of the EBV genome. A similar accumulation is visible for the SwaI-digested molecules in the same portion of the EBV genome ( Figure 4 D). This result was expected since four different pausing sites had been previously described within and near oriP ( Little and Schildkraut 1995 ; Norio and Schildkraut 2001 ). However, no other major accumulation of forks is visible. Therefore, replication forks move mostly unimpeded across the Raji EBV genome. From the profiles of replication fork abundance we can also determine the prevalent direction of the replication forks throughout specific portions of the EBV genome. For example, throughout most of the RRL, replication forks move in both directions at similar frequencies (yellow and orange bars within intervals 11, 13–15, and 17 in Figure 3 D and within intervals 18 and 19 in Figure 4 D). The bidirectional movement of the replication forks also characterizes the central portion of the RRF (intervals 4–7; Figure 4 D). However, this was not evident from the profile of replication fork abundance of the PacI-linearized EBV episomes ( Figure 3 D) because the extremities of the molecules can be distorted or not fully stretched (such as in molecule 5 in Figure 3 ). As a consequence, in the PacI experiment, the position of the replication forks could not always be observed properly within the RRF. From this we conclude that within the central portions of RRF and RRL, replication forks move in both directions with a similar probability. Within the rest of the EBV genome, however, the movement of the replication forks is mostly unidirectional (Figures 3 D and 4 D). For example, replication forks move mainly rightward from interval 11 throughout oriP and beyond ( Figure 4 D; forks infrequently moving in the opposite direction may not appear in this kind of histogram unless extremely large numbers of molecules are imaged). This direction bias is compatible with a previous 2D gel analysis of the oriP region in Raji cells ( Little and Schildkraut 1995 ) and is not affected by the pausing of the replication fork. We conclude that the direction of movement of the replication forks is mainly a consequence of the dynamics of initiation of DNA replication throughout the viral genome. Active Initiation Sites Are Not Limited to the RRF Early studies performed by electron microscopy identified Raji EBV episomes with multiple replication bubbles but could not identify the position of these initiation events ( Gussander and Adams 1984 ). In order to detect the presence of these events and to determine their location we analyzed the immunostaining patterns of the DNA molecules described above. Multiple initiation events should produce molecules containing multiple red patches, each surrounded by green. The qualitative examination of the replication patterns shown in Figure 3 B revealed some of these immunostaining patterns. In molecules 17 and 43, for example, an early initiation event apparently took place within the RRF (large region stained in red). However, shorter red regions are also present, indicating the occurrence of initiation events at later times. Throughout this manuscript, when we refer to multiple initiation events, we will mean that they occur on the same DNA molecule. In addition, if the activation of the initiation sites is not synchronous (as in the molecules described above), we will refer to the initiation events used to begin the duplication of the EBV genome as primary and any subsequent initiation event as secondary. The secondary events visible in molecules 17 and 43 are both located within the long transcription unit of the Epstein Barr nuclear antigen [EBNA] genes (see Figure 3 F). In particular, molecule 43 shows a secondary initiation event that occurred when the duplication of the EBV episome was almost complete (red patch near the pSalF hybridization signal). Therefore, initiation events are not limited to the RRF of the Raji episomes. Initiation events located throughout the EBV genome, as well as multiple initiation events, were also identified in a much larger fraction of Mutu I EBV episomes (see below). We conclude that the entire EBV genome constitutes a large initiation zone, although the frequency of the initiation events is reduced throughout RRL (see below). DNA Replication Proceeds at Different Speeds throughout Different Portions of the Raji EBV Genome In the previous sections we showed that different portions of the EBV genome are not equivalent with respect to when and where DNA replication begins and how DNA replication progresses. Here we wanted to determine the quantitative aspects of DNA replication in different portions of the EBV genome. The data obtained by SMARD can be analyzed quantitatively and used to determine the average time required to duplicate any portion of the EBV genome ( Td; Figure 5 A). By knowing Td and the length of the segment analyzed, the corresponding duplication speed (Sd) can also be calculated ( Figure 5 A). Importantly, these measurements are based on all the images collected during each SMARD experiment, including the molecules entirely stained in red or in green (several hundred). Therefore, the conclusions reached by this analysis are not solely dependent on the appearance of the immunostaining patterns in a small fraction of the DNA molecules. In addition, the quantitative analysis is performed on relatively large genomic segments; therefore, it is not significantly affected by the resolution at which the positions of the replication forks are determined. Figure 5 Duplication Speed of Various Segments of the Raji EBV Genome (A) A detailed procedure to calculate Td using SMARD was published elsewhere ( Norio and Schildkraut 2001 ). Here we describe how to calculate Td ab for a generic genomic region (a)–(b) (i.e., any portion of the EBV genome) using information derived from the replication patterns of a longer region (A)–(B) (i.e., the whole EBV genome). Depicted are the hypothetical staining patterns for 32 DNA molecules representing the genomic region (A)–(B) after double-labeling with two halogenated nucleotides (red and green). The molecules that started and completed their duplication during the first labeling period are fully red (R). The molecules that started their duplication during the first labeling period, completing it during the second labeling period are stained in both red and green (RG). In the total population of molecules, the fraction of R molecules increases when the length of the first labeling period (Tp1) increases. The fraction of RG molecules is proportional to the time required to duplicate the genomic region (A)–(B). Some of these molecules (marked rg ) are stained in red and green also within the region (a)–(b). The fraction of rg molecules is proportional to the time required to duplicate the genomic region (a)–(b). This relationship is expressed by the equation reported at the bottom of the figure and allows us to calculate Td ab using parameters that can be easily measured on individual DNA molecules ( N R , the number of R molecules; N RG , the number of RG molecules; N rg , the number of rg molecules). Finally, the ratio between the size of the genomic segments analyzed (L ab ) and Td ab represents the duplication speed of the segment (Sd ab ). The results obtained from the PacI and the SwaI experiments are reported in (B) and (C). Double-headed arrows indicate the genomic segments analyzed quantitatively. Segments marked with the same letter in the two maps correspond to identical portions of the Raji EBV genome. Above each arrow is indicated the corresponding Sd value in kilobases per minute. The sizes of these segments are as follows: A–G, 25 kb; H, 20 kb; I, 35 kb; K, 10 kb; L, 75 kb; and J, 10 kb. A comparison of the values obtained from the two experiments shows remarkable similarities. The largest variation was found for segment G . However, in both experiments this segment was located at the end of the DNA molecules. Therefore, the variability in stretching in this portion of the molecules may have affected the collection of the data. The red dashed line below the map indicates the position of the RRF. We calculated the value of Sd for each portion of the Raji EBV genome, depicted with double-headed arrows in Figure 5 (segments A–K; Td is reported in Table 1 ). The results obtained from the PacI and the SwaI experiments were analyzed independently but show remarkable similarities (compare the values reported above segments marked with the same letter in Figures 5 B and 5 C). Therefore, the quantitative analysis is highly reproducible. Table 1 Quantitative Analysis of Different SMARD Experiments The segments of the EBV genome (left column) are shown as double-headed arrows in Figures 5 and 6F. For each segment the table reports the number of EBV genomes completely stained in red (N R ), the number stained in both red and green (N RG ), the number stained in both red and green within the genomic segment analyzed (N rg ), and Td (calculated using the equation shown in Figure 5A). Note that the values of N R , N RG , and N rg reported for the PacI-linearized Raji EBV episomes include data from a previously published SMARD experiment ( Norio and Schildkraut 2001 ). Tp1 was 240 min for the Raji experiments and 210 min for the Mutu I experiment We found that different portions of the EBV genome replicate at different speeds, with values that range from a minimum of 0.3 kb/min (segment K; Figures 5 B and 5 C) to a maximum of 3.5–4.7 kb/min (segment I ). More details are provided later in the text. However, it is important to note that the highest Sd values were detected within the central portion of the Raji RRF (segments A, I, and G ). This result can be explained in two ways. First, replication forks may move faster throughout the RRF. Alternatively, the RRF could contain a significant level of multiple initiation events. For the reasons mentioned below we favor the second possibility. Multiple Initiation Events Can Take Place on the Same Raji EBV Episome within the RRF Three lines of evidence indicate that multiple initiation events take place within the Raji RRF. Two lines of evidence are discussed in this section (the presence of multiple red patches in the immunostaining patterns of some EBV episomes and the detection of termination events by 2D gel analysis); the third is discussed in the last section of Results (differences in duplication speed across segments of the RRF of different sizes). The first line of evidence is provided by the immunostaining pattern of the EBV molecules. Although discontinuities in the immunostaining make it difficult to detect multiple initiation events when the distance between converging forks is 5 kb or less, the replication patterns of some of the molecules are compatible with the presence of multiple initiation events within RRF (molecule 8 in Figure 3 B and molecules 5 and 9 in Figure 4 B). For example, in molecule 9 three regions stained in red (divergent arrows) are separated by two regions stained in green about 4 kb in size (convergent arrows); shorter patches are also visible (asterisks) and might indicate the presence of additional initiation events. In these molecules, the genomic regions stained in red are very close to each other and approach the resolution limits of SMARD. If these signals were produced by multiple initiation events, we should conclude that they took place at about the same time and with a short interorigin distance. Alternatively, secondary initiation events might have taken place in proximity to an incoming replication fork (generated by a primary initiation event). In both cases, the different replication bubbles would rapidly fuse into a larger bubble, making the detection of these events extremely difficult. Since these patterns are too short to be unequivocally ascribed to DNA replication, the presence of multiple initiation events within the RRF was confirmed using a replication mapping approach independent of SMARD. A second line of evidence is provided by the structure of the replication intermediates examined by 2D gel electrophoresis in exponentially growing Raji cells ( see Materials and Methods ). We analyzed nine restriction fragments, indicated in Figure 3 A as gray bars (a–i). We also reexamined the hybridization patterns of ten different fragments analyzed in a previous study (black bars 1–10 in Figure 3 A) ( Little and Schildkraut 1995 ). In total, we considered 19 restriction fragments spanning approximately 65% of the Raji EBV genome. The summary of these 2D gel analyses is shown in Figure 3 E. Replication intermediates indicative of initiation events were found in several restriction fragments (marked by a green I above the corresponding fragment in Figure 3 E). The restriction fragments containing replication bubbles are contiguous and span the genomic regions underlined by the red dashed line at the bottom of Figure 3 E (approximately corresponding to the RRF). As expected, termination events (marked by a red T in Figure 3 E) were detected in many of the fragments located within the RRL. Importantly, random termination events were also detected in most of the fragments in which we detected bubble arcs. However, the source of these events could not be identified in our previous studies. As discussed earlier, SMARD shows unequivocally that the RRF completes its duplication before forks originating elsewhere have the time to reach its central portion. We conclude that the termination events detected by 2D gel electrophoresis derive from the collision of replication forks generated by multiple initiation events taking place within the RRF. An estimate of the frequency of these multiple initiation events is reported later in the text. In Mutu I Cells, the Order of Activation of the Initiation Sites Varies Throughout the Viral Genome Previous observations have suggested that some initiation sites (such as oriP) are used at a different frequency in different EBV strains ( Little and Schildkraut 1995 ). However, it was not known whether these changes were the result of modifications in the activity of individual initiation sites or involved multiple initiation sites throughout the EBV genome. In order to determine the extent of these differences, we performed SMARD in another Burkitt's lymphoma cell line called Mutu I ( Gregory et al. 1990 ). A brief description of this cell line and of the conditions used for SMARD is reported in Materials and Methods . From this experiment we recovered the images of 271 PacI-linearized EBV episomes substituted along their entire length with halogenated nucleotides (122 red, 107 green, 42 red and green). The results of this analysis are shown in Figure 6 . Figure 6 SMARD Performed on PacI-Linearized EBV Episomes Replicated in Mutu I Cells (A) Map of the PacI-linearized EBV genome with the positions of various genetic elements shown to scale. Below the EBV map, light blue bars show the positions of the hybridization probes (pWW and pSalF) utilized to identify the EBV molecules and their orientation. Black bars indicate the positions of two short deletions present in the Raji EBV genome ( Polack et al. 1984 ). (B) Images of 40 PacI-linearized EBV episomes ordered and marked as in Figure 3 B (from a population of 42 molecules collected in this experiment). Some molecules are distorted but the positions of the red-to-green transitions are clear. Two additional molecules were unsuitable for precise measurements and are not shown. (C) Replication profile of the PacI-linearized EBV episomes shown in (B). Starting from the PacI site, intervals of 5 kb are indicated on the horizontal axis by numbers from 1 to 35. The vertical axis indicates the percentage of molecules stained red within each 5-kb interval. (D) Profile of replication fork abundance and direction. Intervals of 5 kb are indicated on the horizontal axis as for (C). The vertical axis indicates the percentage of molecules containing replication forks in each 5-kb interval. Three different pausing sites contribute to the significant accumulation of replication forks within interval 8 (the two EBER genes and the FR element). A fourth pausing site (the DS element) is located within interval 9, producing a minor accumulation of replication forks. (E) Map of the EBV genome (to scale) aligned with the horizontal axes of histograms (C) and (D). Red arrows mark the positions of regions transcribed during the type I latency that characterize the Mutu I EBV episomes. The red dashed line above the map indicates the position of the RRF. (F) Duplication speed for various segments of the Mutu I EBV genome. Double-headed arrows indicate the genomic segments analyzed quantitatively. Above each arrow is indicated the corresponding Sd value in kilobases per minute. Segments A′–G′ divide the whole EBV genome into seven parts of identical size, corresponding, respectively, to the intervals 1–5, 6–10, 11–15, 16–20, 21–25, 26–30, and 31–35 on the horizontal axes of Figures 3C and 4C. The sizes of these segments are as follows: A′–G′, 25 kb; and K′ , L′, and M′, 10 kb. Due to the presence of small differences in the DNA sequence (see text), segments A′–G′ are similar but not identical to segments A–G in Figure 5 . The most striking differences in the replication of the two EBV strains were found in the order of activation of the initiation sites. In Raji episomes, primary initiation events are restricted to an 80-kb region approximately corresponding to the RRF (this study and also Norio and Schildkraut 2001 ). In contrast, in Mutu I episomes primary initiation events occur at many locations throughout the viral genome (molecules 1, 4, 5, and 8; Figure 6 B). Multiple initiation events (mostly largely spaced), firing either synchronously (molecules 2 and 5) or asynchronously (molecules 11, 15, 31 and 34), are also frequent in Mutu I. This explains the heterogeneity detected in the positions of the termination events (green patches in molecules 32–40), as well as in the replication patterns of the episomes at intermediate stages of duplication. Hence, DNA replication initiates, progresses, and terminates differently in different Mutu I episomes. Even if the order of activation of the initiation sites varies from molecule to molecule, more than 80% of the Mutu I episomes are stained in red between intervals 6 and 9 (a 20-kb region that includes oriP; Figure 6 C). This indicates that this EBV strain also contains an RRF, although its genomic location differs from that found in Raji (compare Figure 3 C with Figure 6 C). In summary, these results confirm that initiation sites are not confined to a specific portion of the EBV genome and allow us to conclude that their utilization in different viral strains can change throughout the EBV genome. The RRFs Are Produced by Clusters of Initiation Sites Frequently Activated at the Beginning of the Duplication of the Episomes Modifications in the activity of individual initiation sites (such as oriP) could potentially explain the different location of the RRF in Raji and Mutu I episomes. However, the following considerations indicate that this is not the case. Initiation events occurring at oriP take place in the vicinity of the dyad symmetry (DS) element and produce replication forks that stall in the leftward orientation at the family of repeats (FR) element ( Gahn and Schildkraut 1989 ). Initiation events taking place to the left of oriP (such as in Raji episomes) instead produce replication forks that pause in the rightward orientation (interval 8 in Figure 3 D). In the Mutu I genome, oriP is at the center of the RRF, but replication forks accumulate in both orientations within interval 8 (see Figure 6 D). This indicates that primary initiation events occur on both sides of oriP. Examples of initiation events that took place near the DS element, or to the left of oriP, are visible, respectively, in molecule 5 and in molecules 1 and 3 ( Figure 6 B). The presence of initiation events on both sides of oriP is also supported by the replication pattern of molecule 2, in which two initiation events spaced approximately 10 kb apart are visible on the same episome. Therefore, in the Mutu I EBV genome, the RRF (∼20 kb in size) contains multiple active initiation sites that have a shared tendency to be activated at the beginning of the duplication of each episome. Similar conclusions apply to the RRF of the Raji EBV genome (∼80 kb in size), in which primary initiation events were detected at various locations (this study and also Norio and Schildkraut 2001 ). This could explain why only weak bubble arcs were detected by 2D gel electrophoresis throughout the RRF, even though the duplication of the episomes usually begins within this genomic region (this study and also Little and Schildkraut 1995 ). We conclude that the RRFs in the Raji and Mutu I EBV genomes are similar in that they contain various initiation sites that have a shared tendency to be activated at the beginning of the duplication of each episome. The Duplication Speed of Various Portions of the EBV Genome Is Different in Raji and Mutu I EBV Episomes In the previous sections we have shown that the order of activation of the initiation sites in Raji and Mutu I EBV episomes is not conserved. Here, we wanted to determine whether the quantitative aspects of DNA replication were also different. SMARD was used to calculate Sd for each portion of the Mutu I EBV genome, depicted as double-headed arrows in Figure 6 F (segments A′ – M′; see also Table 1 ). We found that DNA replication proceeds at different speeds throughout different portions of the Mutu I EBV genome (from a minimum of 0.5 kb/min in segment K′, to a maximum of 3.8 kb/min in segment G′ ). This is very similar to the range of speeds found in the Raji episomes (0.3–4.7 kb/min). Therefore, DNA replication appears to progress with similar kinetics in the two EBV strains. We also noticed that similar portions of the EBV genome have different Sd values in the two viral strains. Segments A′–G′ divide the Mutu I EBV genome in seven parts of identical size (∼25 kb; Figure 6 F). These segments encompass portions of the EBV genome similar to segments A–G in the Raji genome (see Figure 5 ). However, the values of Sd differ significantly in almost every segment. Interestingly, in Mutu I episomes, the highest Sd values were not detected within the RRF (segment G′ ). Instead, the RRF contained some of the lowest Sd values (segment B′ ). This is probably due to the presence of strong pausing sites within this portion of the Mutu I EBV genome. Nevertheless, segment B′ replicates faster than the corresponding portion of the Raji EBV genome (segment B in Figure 5 ), a phenomenon that could be explained by the presence of multiple initiation events on both sides of oriP in a fraction of the EBV episomes. In any case, these results indicate that there is no simple correlation between the Sd of a genomic segment and its location within the RRF or the RRL. Replication Forks Progress at Similar Rates Across Different Portions of the EBV Genome and in Different EBV Strains Previous observations have suggested that in mammalian cells the speed of the replication forks can vary ( Housman and Huberman 1975 ). Here, we wanted to determine if some of the differences detected in the replication of Raji and Mutu I episomes could be ascribed to modifications in the rate of progression of the replication forks as proposed for other systems ( Anglana et al. 2003 ). For a genomic segment replicated by forks moving in a single direction, Sd corresponds to the average speed of the replication forks (provided pausing sites are absent). This allowed us to measure the average speed of the replication forks in various portions of the EBV genomes in which these conditions are satisfied. In Raji episomes, we found that the average speed of the replication forks was about 1.0 kb/min throughout both segment H and segment J (see Figure 5 ); these segments are replicated by forks moving predominantly in one direction (corresponding, respectively, to intervals 22–25 and 9–10 in Figure 3 D). Interestingly, a similar value (1.1 kb/min) was found for two different portions of the Mutu I EBV genome in which replication forks also move predominantly in one direction (segments M′ and L′ in Figure 6 F, corresponding, respectively, to intervals 3–6 and 9–11 in Figure 6 D). Therefore, in both EBV strains, we found that the average speed of the replication forks is approximately 1.0 kb/min within every segment that could be analyzed. Studies performed in different systems have suggested that transcription could interfere with the progression of the replication forks ( Brewer 1988 ; Liu and Alberts 1995 ). In the Raji EBV genome, segments J and H (see Figure 5 ) are located within the long transcription unit of the EBNA genes (see Figure 3 F). Throughout segment J, replication forks progress in the same direction of transcription, whereas in segment H their orientation is reversed (see Figure 3 D). Nevertheless, as demonstrated above, replication forks move at the same speed in both segments. Replication forks also move at a similar speed across two nontranscribed regions in the Mutu I EBV genome (segments M′ and L′; see Figure 6 ). We conclude that in our systems the progression of the replication forks is not significantly influenced by transcription. This could be so either because the level of transcription is not sufficiently high or because, as suggested by others, transcription and DNA replication do not occur at the same time in mammalian cells ( Wei et al. 1998 ). The Duplication Speed of a Genomic Segment Reflects the Average Number of Replication Forks Involved in Its Replication Variation in the utilization of the initiation sites and similarity in the speed of the replication forks suggest that the former should have a stronger influence on the duplication speed of a genomic segment. If we assume that the speed of the replication forks is constant throughout the EBV genome (except for the regions containing pausing sites), Sd becomes a function of the number of replication forks actively synthesizing DNA. High Sd values would indicate that a large number of replication forks participate in the replication of a genomic segment. If we apply this assumption to the central portion of the Raji RRF, we can see that the values of Sd for segments A and G (see Figure 5 ) are compatible with the presence 2–3 active forks/segment (corresponding to about one initiation event per duplication cycle within each of these 25-kb segments). Importantly, a larger segment spanning the same portion of the EBV genome (segment I; 35 kb in size) replicates even faster (3.5–4.7 kb/min; see Figure 5 ). This increase could not be explained if the changes in duplication speed were caused by modifications in the speed of the replication forks. However, it is precisely what would be expected if an average of two initiation events take place within the 35-kb segment of the RRF (as suggested by the immunostaining pattern of the episomes and supported by the 2D gel analysis). Therefore, the observed duplication speeds support a model in which, within the RRF, initiation sites spaced 25 kb apart or less can become licensed on the same EBV episome. We also noticed that the EBV genome duplicates faster in Mutu I than in Raji cells ( Table 1 ). This is in agreement with the higher level of widely spaced multiple initiation events detected in Mutu I (compare Figures 3 B and 6 B). We conclude that the differences in Sd across the EBV genome and between the two EBV strains reflect different frequencies of initiation and termination events. Discussion Conserved and Nonconserved Features in the Latent Replication of Different EBV Strains In this study, we determined how DNA replication initiates and progresses in EBV episomes latently replicating in two human Burkitt's lymphoma cell lines (Raji and Mutu I). Previous experiments had suggested that some variability in the utilization of oriP might exist among different EBV strains ( Little and Schildkraut 1995 ). Here, however, we found that the replication dynamics vary across the entire EBV genome to an extent that could have not been predicted from previous studies. As exemplified by the replication profiles, the immunostaining patterns of the episomes is strikingly different in the Raji and Mutu I strains (compare Figures 3 and 6 ). This indicates that the order of activation of the initiation sites is not conserved. Differences were also found in the direction of movement of the replication forks (see Figures 3 D and 6 D) and in the duplication speed of different portions of the EBV genome (see Figures 5 and 6 F). The last, in particular, indicates that the frequency of initiation and termination events varies across the EBV genome and between the two viral strains. We did not find a simple correlation between the Sd of a genomic segment and its location within the RRFs. For example, the high Sd value for segment G′ indicates the presence of active initiation sites outside the Mutu I RRF. Therefore the factors that influence the order of activation of the initiation sites are at least partially distinct from the factors that control their frequency of activation. The EBV episomes replicating in these two cell lines have a similar size and genomic organization. However, the number of internal repeats 1 (also called Bam HI W) is reduced by one unit in the Mutu I strain (not shown), while the Raji EBV genome contains two short deletions (see Figure 6 A) ( Polack et al. 1984 ). In principle, these differences could affect some initiation sites. On the other hand, initiation events were detected throughout the EBV genomes. It is unlikely that localized modifications of the DNA sequence (affecting one or few initiation sites) could account for all the differences in the replication of Raji and Mutu I episomes. Primary or secondary events were detected within almost every 25-kb segment of the Mutu I EBV genome (see Figure 6 F), such as segment A′ (molecule 11; see Figure 6 B), segment B′ (molecules 1, 2, 3, and 5), segment C′ (molecule 34), segment D′ (molecule 8), segment F′ (molecules 5 and 15), and segment G′ (molecules 4 and 31). Similarly, in Raji episomes, replication bubbles were detected within every restriction fragment analyzed by 2D gel electrophoresis throughout a region of about 80 kb (the sizes of these fragments ranged from 3 to 12 kb; Little and Schildkraut [1995] and this study). Using SMARD, low frequencies of secondary initiation events were also detected in the remaining portion of the Raji EBV genome. Therefore, even if SMARD and the 2D gel analysis do not have the resolution to pinpoint the locations of the initiation events at the nucleotide level, our results clearly indicate that the average distance between the initiation sites is below 25 kb. This study also revealed modifications in the pausing of the replication forks in the oriP region. Accumulation of replication forks is clearly present in both EBV strains within this genomic region. However, only 25% of replicating Mutu I episomes contain replication forks at this location (see Figure 6 D), compared with 43% of Raji episomes (see Figure 3 D). Quantitative estimates of the average pausing of the replication forks suggested values of about 30 min in Raji episomes and 10 min in Mutu I episomes (data not shown). The decreased pausing could reflect the presence of active initiation sites on both sides of oriP in the Mutu I strain (as suggested by the immunostaining pattern of the episomes). Alternatively, a decreased efficiency of the pausing sites could also contribute to the significant reduction in pausing detected in Mutu I. We also found that some features of the episomal duplication do not vary. In both Raji and Mutu I cells, replication forks move freely throughout the EBV genome (except near oriP; see Figures 3 D and 6 D), and their progression rate appears to be constant. This indicates that modifications in the speed of the replication forks do not contribute significantly to the differences in DNA replication described above. This contrasts with results obtained by another laboratory for an amplified genomic locus ( Anglana et al. 2003 ), in which the slower progression of the replication forks—caused by a reduction in nucleotide pools—was presented as the cause for a more frequent activation of initiation sites. Instead, the changes in DNA replication detected in our experiments appear to be caused by differences in the order and frequency of activation of groups of initiation sites encompassing large genomic regions (see next section). Another common feature between Raji and Mutu I cells is the presence of a genomic region that usually replicates first during the duplication of each episome. The position of this RRF differs in the two EBV strains. However, the direction of movement of the replication forks throughout the RRFs is similar. For example, within the central portion of each RRF, replication forks move in both directions, while along its distal portions, replication forks move predominantly outward (see Figures 3 D, 4 D, and 6 D). We conclude that the direction of movement of the replication forks throughout the EBV genome is mainly a consequence of the dynamics of the initiation of DNA replication. Utilization of Initiation Sites is Regulated at the Level of Genomic Regions Rather Than at the Level of Individual Initiation Sites Even if the EBV episomes utilize the same replication machinery (provided by the host cell), several aspects of their duplication are not conserved between Raji and Mutu I. In mammalian cells, prereplication complexes are believed to form at the end of mitosis, when general transcription is shut off ( Okuno et al. 2001 ; Dimitrova et al. 2002 ; Mendez and Stillman 2002 ). However, the selection of specific initiation sites occurs only later in G1, at the origin decision point ( Wu and Gilbert 1996 ). This suggests that there must be a mechanism that, during G1, restricts the utilization of initiation sites to specific regions of the mammalian genomes. In this study, we have shown that initiation sites are present throughout the EBV genome and that their utilization differs dramatically in different EBV strains. It is therefore reasonable to assume that the utilization of the initiation sites in the EBV episomes is restricted by the same mechanisms responsible for the selection of the initiation sites in mammalian chromosomes. One of the questions we tried to answer is whether initiation of DNA replication is regulated at the level of individual initiation sites. Clues to a possible regulatory mechanism can be found in the replication profiles of the EBV episomes. The RRFs are localized in specific portions of the EBV genome that differ in the two EBV strains. These regions are tens of kilobases in size (about 80 kb in Raji episomes and 20 kb in Mutu I episomes) and encompass multiple initiation sites. The early activation of an individual initiation site could be sufficient to generate a RRF. However, our results have demonstrated that within each RRF various initiation sites have a similar tendency to be activated at the beginning of the duplication of each episome. Therefore, the order of activation of the initiation sites varies at the level of genomic regions rather than at the level of individual initiation sites and might reflect the presence of a particular chromatin organization. Recent findings have shown that histone acetylation can influence the timing of replication origin firing in yeast ( Pasero et al. 2002 ; Vogelauer et al. 2002 ). In this study we found that even if initiation events were detected at many locations within the MutuI episomes, primary initiation events occurred predominantly within the RRFs. Modifications in chromatin structure could be used in mammalian cells to regulate the order of activation of the initiation sites across genomic regions that encompass multiple initiation sites. The early activation of the initiation sites located in these regions would increase the chance of passively replicating the neighboring initiation sites contributing, at least in part, to the process of selection of the initiation sites. In addition to changes in the order of activation of the initiation sites, other mechanisms could influence their utilization by affecting their frequency of activation. We noticed that in Raji episomes the frequency of initiation events across the RRL appears to be reduced compared to that of the RRF. This difference is reflected in the higher levels of Sd detected within the RRF compared to the RRL ( Figure 5 ) and in the absence of bubble arcs outside the RRF ( Figure 3 E). The replication profile of the EBV episomes also indicates that the RRL (the genomic region stained in red in less than 40% of the EBV episomes) is larger in Raji (intervals 9–22 in Figure 3 C) than in Mutu I (intervals 17–25 in Figure 6 C) and that it mirrors the positions of the longest transcription units active in each strain (see Figures 3 E and 6 E). This suggests that the presence of a long transcription unit could delay the duplication of the corresponding genomic region. This delay is unlikely to be caused by an impaired progression of the replications forks, since we have shown that the average speed of the replication forks is not significantly influenced by transcription (see Figures 5 B, 5 C, and 6 F). An alternative possibility could be that transcription decreases the frequency of initiation events across the genomic regions traversed by RNA polymerases, as previously suggested by others ( Kalejta et al. 1998 ; Snyder et al. 1988 ; Haase et al. 1994 ; Tanaka et al. 1994 ). Perhaps the passage of RNA polymerase removes, or inactivates, prereplication complexes deposited on the DNA at the end of mitosis (see next section). The observation that initiation events appear more diffusely across the Mutu I EBV genome than in Raji is consistent with this hypothesis and might reflect the presence of larger nontranscribed regions in Mutu I ( Gregory et al. 1990 ). Further experiments will be required to shed light on this phenomenon. However, the detection of some initiation events within the transcribed regions (i.e., molecules 17 and 43 in Figure 3 B and molecule 8 in Figure 6 B) suggests that the relationship between transcription and replication could be more complex. Initiation Sites Are Redundant Elements of the EBV Genome In this study, we have shown that initiation events are not confined to a specific portion of the episomes, suggesting that DNA sequences capable of functioning as initiation sites must be rather common. This can explain why, under various experimental conditions, individual initiation sites do not appear to play an essential role in the replication of EBV episomes. For example, a hundred-nucleotide deletion encompassing the DS element of oriP is sufficient to abrogate both initiation of DNA replication ( Norio et al. 2000 ) and the binding of ORC and MCM proteins at this genomic location ( Chaudhuri et al. 2001 ; Schepers et al. 2001 ). This deletion, however, has no apparent effect on the stable replication of the episomes in established cell lines ( Norio et al. 2000 ; Kanda et al. 2001 ). Therefore, other efficiently licensed initiation sites are present in different portions of the EBV genome. Large deletions are also well tolerated (deletions I, II, III, and IV in Figure 7 ), even when they encompass portions of the EBV genome known to contain multiple initiation sites (such as the Raji RRF). One of the deleted EBV genomes shown in Figure 7 (genome IV; Kempkes et al. 1995a ; Kempkes et al. 1995b ) was recently analyzed to detect binding sites for ORC and MCM proteins. Significant binding of these proteins was detected only at oriP ( Schepers et al. 2001 ). However, we found that in this mini-EBV genome, oriP is used at a frequency that approaches 100% (B. Chaudhuri and C. L. Schildkraut, unpublished data). Therefore, the absence/reduction of replication complexes at other locations correlates with an increased usage of the licensed initiation site oriP. Interestingly, these short versions of the EBV genome were specifically engineered to preserve the latency genes by removing most of the untranscribed regions. Therefore, a possible effect of transcription could be to reduce the number of replication complexes present throughout the EBV genome. Reductions in initiation events throughout transcribed regions could be relevant in the maintenance of genomic stability. In fact, it has been reported that at least three extremely large transcription units ( FHIT, WWOX, and Parkin; each ∼1 Mb in size) encompass known common fragile sites in mammalian genomes ( Ohta et al. 1996 ; Ried et al. 2000 ; Krummel et al. 2002 ; Denison et al. 2003 ). We conclude that initiation sites are redundant elements of the EBV genome and that the deletion of some of them can be compensated for by an increased usage of the remaining sites. Figure 7 EBV Strains Carrying Large Deletions Map of a generic EBV genome linearized with the restriction endonuclease PacI. The number of terminal and internal repeats 1 can vary in different EBV strains; therefore, broken lines were inserted in the map at these positions. The deletions present in four EBV strains are shown as black bars below the map. Deletion I is 12 kb in size and is present in episomes of cell lines obtained by transformation with B95–8 EBV isolates ( Raab-Traub et al. 1980 ; Parker et al. 1990 ). This deletion encompasses a portion of the EBV genome corresponding to the central portion of the Raji RRF. The remaining deletions were artificially engineered in the EBV genome. Deletion II is described in Robertson et al. (1994) . Deletion III is described in Robertson and Kieff (1995) . Deletion IV is described in both Kempkes et al. (1995a) and Kempkes et al. (1995b) . Role of oriP in the Replication of EBV Episomes Initiation site oriP is the best characterized initiation site of the EBV genome. Initiation of DNA replication has been detected at this site in every EBV strain analyzed to date by 2D gel electrophoresis. However, the frequency of the initiation events at oriP varies in different EBV strains, and it is particularly low in Raji ( Little and Schildkraut 1995 ). The infrequent use of this initiation site in Raji does not appear to be caused by an inability to assemble a prereplication complex. In fact, in this cell line, both ORC and MCM proteins efficiently bind oriP ( Chaudhuri et al. 2001 ; Ritzi et al. 2003 ). Changes in the DNA sequence of oriP are also an unlikely cause for this difference since only single nucleotide polymorphisms between the Raji and Mutu I strains have been described at this location ( Salamon et al. 2000 ). In our current study we show that primary initiation events are frequently detected by SMARD near oriP in Mutu I but not in Raji episomes. This could be explained, in part, by the decreased frequency of utilization of this site in the Raji strains. However, this is unlikely to be the only reason. Even infrequent primary initiation events occurring at oriP would produce replication forks that pause in the leftward orientation within interval 8 (as seen in Mutu I), but none of the Raji EBV episomes showed forks paused in this orientation (compare Figure 6 D with Figure 3 D). Therefore, an alternative explanation is that in Raji episomes the activation of oriP is delayed compared to initiation sites located in the RRF. This would cause an increase in the passive replication of this genomic region, explaining the reduced frequency of activation detected by 2D gel electrophoresis. In this context, the residual activity of oriP in Raji could represent secondary initiation events that take place in proximity to replication forks that originated in the RRF and paused near oriP. These events would produce small red patches, such as the one marked by an asterisk in molecule 35 of Figure 3 , that are too close to the paused forks to be unequivocally identified by SMARD as separate initiation events. We conclude that oriP is one of the initiation sites preferentially utilized to begin the duplication of the Mutu I episomes, while in Raji only secondary initiation events usually occur at this site. Various groups have suggested that different cellular proteins could participate in regulating the activity of oriP ( Dhar et al. 2001 ; Shirakata et al. 2001 ; Deng et al. 2002 ). However, it is currently not clear why, in Raji episomes, oriP is not among the preferred initiation sites. Interestingly, it has been reported that oriP is more extensively methylated in Raji than in other EBV strains ( Salamon et al. 2000 ). It is therefore tempting to speculate that epigenetic modifications of the DNA template, or modifications of the chromatin structure, could be responsible for the differences in the order of activation detected in these EBV strains. The epigenetic regulation of oriP activity could be particularly important during the establishment of latent replication, since it has been demonstrated that an epigenetic event is required for the establishment of oriP-dependent replication ( Leight and Sugden 2001 ). Conclusions—Flexible Utilization of Initiation Sites in Higher Eukaryotes In this study we have shown that, while the basic features of DNA replication are conserved (i.e., the progression of the replication forks), the activity of the initiation sites (order and frequency of activation) varies significantly in different EBV strains and across different portions of the EBV genome. Importantly, using SMARD we are now beginning to detect similar modifications in the utilization of initiation sites across transcriptionally active chromosomal loci of the mouse genome (data not shown). Additional mechanisms could regulate DNA replication at transcriptionally silent loci, as suggested by the complete absence of initiation events throughout an approximately 450-kb portion of the mouse IgH locus in non–B cell lines ( Zhou et al. 2002 ; data not shown). These results are compatible with the flexible utilization of initiation sites also suggested by other laboratories ( Kalejta et al. 1998 ; Lunyak et al. 2002 ). It is therefore likely that the large redundancy in initiation site usage and the regulation of initiation site activity at the level of genomic regions represent common features of DNA replication in mammalian cells. In particular, our results suggest that long-range changes in chromatin structure or chromosomal organization could be far more important than local modifications at individual initiation sites in regulating DNA replication. This could represent an efficient way for eukaryotic cells to control the replication of their very large genomes, and could have broad implications for the maintenance of genomic stability. By using SMARD on primary cells, we will soon be able to determine if similar dynamics are also present in nontransformed mammalian cells. Methods Cell cultures, EBV strains, and double-labeling of replicating DNA Raji cells were grown in exponential phase (as described in Little and Schildkraut 1995 ), keeping the cell density between 3 × 10 5 and 8 × 10 5 cells/ml. The experiments presented in this manuscript were performed at approximately 5 × 10 5 Raji cells/ml, using two labeling periods (240 min each) with 25 μM IdU (first label) and 25 μM CldU (second label). IdU was added directly to the growing culture, followed by low-speed centrifugation of the cells at the end of the first labeling period and resuspension in warm medium containing CldU. Early passages of the Mutu I cells (clone c179 p44; Gregory et al. 1990 ) were provided by Alan B. Rickinson and grown for only seven additional passages (keeping the cell density between 4 × 10 5 and 8 × 10 5 cells/ml) before the replicating DNA was labeled. The conditions used for growth and labeling were the same as those used for Raji cells, with the exception that the labeling periods were only 210 min each. This Mutu I cell line is characterized by the presence of a small fraction of cells in which EBV replicates lytically, producing molecules linearized at the terminal repeats. However, this did not affect our analysis of the latently replicating episomes because only the DNA molecules that were circular before the digestion with PacI were recovered from the agarose gels and analyzed by SMARD. Only ten EBV genes (out of about 100) can be expressed during latency (six EBNAs; two latent membrane proteins [LMPs]; two nontranslated Epstein Barr encoded RNAs [EBERs]; reviewed in Kieff 1996 ). In EBV-associated diseases, where the viral genome is maintained as a circular episome, the phenotype of the infected cell influences the viral patterns of expression ( Babcock et al. 2000 ). Three different latent transcription patterns have been described ( Kerr et al. 1992 ): type I (only EBNA1 and EBERs expressed), type II (only EBNA1, the LMPs, and the EBERs expressed), and type III (all the EBNAs, the LMPs, and the EBERs expressed). Although both Raji and Mutu I are human Burkitt's lymphoma cell lines, their transcription profiles are different. The Mutu I cell line used in this study was an early passage of a type I clone isolated in the Alan B. Rickinson laboratory. Raji cells, instead, have a type III -like transcription pattern and they also carry a deletion of the EBNA3C gene ( Polack et al. 1984 ). Improved method to stretch a large number of EBV molecules on individual slides In order to collect a sufficient number of images of the EBV genome, the population of replicated episomes needed to be enriched by a partial purification using pulsed field gel electrophoresis. However, starting from the limited amount of DNA that can be purified from a pulsed field gel, we could not stretch a sufficient number of DNA molecules on microscope slides by molecular combing ( Bensimon et al. 1994 ). As a consequence, the collection of several hundred images of EBV episomes would have required the analysis of a very large number of microscope slides and the use of large amounts of pulsed field–purified DNA. In order to solve this problem, we decided to stretch the DNA molecules using a modification of the method originally introduced for the optical mapping of restriction sites on individual DNA molecules ( Aston et al. 1999 and references therein) as well as for other applications ( Henegariu et al. 2001 ). The stretching was achieved by the movement of a DNA solution (a few microliters) gently deposited at the interface between a silanized microscope slide and a nonsilanized coverslip. In this way it was possible to complete our analysis using just few microscope slides and a fraction of the pulsed field–purified DNA derived from the digestion of 10 6 cells. The molecules stretched by capillary action vary in their orientation (see Figure 1 ) and in their size (see Figure 2 A). Nevertheless, the EBV molecules were clearly identified by the two hybridization signals. These images were aligned with the map of the EBV genome by computer adjustment of the image size for the entire DNA molecule (see Figures 2 B and 2 C), as we did previously when EBV episomes were stretched by molecular combing ( Norio and Schildkraut 2001 ). Hybridization, probe detection, and immunostaining of the individual DNA molecules stretched on microscope slides Hybridization was performed as previously described ( Parra and Windle 1993 ) using probes prepared by nick translation in the presence of biotin-16-dUTP (Roche, Basel, Switzerland). The probes used in this study, pSalF, p107.5, and pWW (provided by John L. Yates), were detected using a modification of the DIRVISH procedure ( Heng et al. 1995 ). Briefly, five layers of Alexa Fluor 350 conjugated NeutrAvidin (Molecular Probes, Eugene, Oregon, United States) and biotinylated anti-avidin antibodies (Vector Laboratories, Burlingame, California, United States) were deposited on the microscope slide, washing with PBS, 0.03% Igepal CA-630 (Sigma, St. Louis, Missouri, United States) after each step. The purpose of the hybridization signals is to identify and orient the EBV episomes. Since the DNA molecules studied by SMARD are substituted along their entire length with halogenated nucleotides, they are very easy to detect even in presence of substantial hybridization background (i.e., blue dots in the lowest panel of Figure 1 ). This hybridization background does not affect SMARD, therefore it was digitally removed from the images of the molecules shown in Figures 3 B, 4 B, and 6 B (as described in other studies performed on stretched DNA molecules; Pasero et al. 2002 ). Immunostaining to detect IdU and CldU was performed simultaneously with the detection of the biotinylated DNA probes. Mouse anti-IdU (Becton-Dickinson, Palo Alto, California, United States) and rat anti-CldU (Accurate Chemical, Westbury, New York, United States) were used as primary antibodies (monoclonal), while Alexa Fluor 568-conjugated goat anti-mouse (Molecular Probes) and Alexa Fluor 488-conjugated goat anti-rat (Molecular Probes) were used as secondary antibodies. The immunostaining has almost no background. As described previously ( Norio and Schildkraut 2001 ), the specificity of the immunostaining was tested on DNA fully substituted with IdU or with CldU. No cross-reaction of the antibodies was detected using the detection procedure utilized in this study, and both antibodies were unable to recognize the unlabeled DNA. In practice, the background visible in the red and green channels is mainly represented by other DNA molecules containing halogenated nucleotides (white horizontal arrowheads in the central panels of Figure 1 ). These molecules can be fully or partially stretched (sometimes collapsed or broken in pieces), but are usually clearly distinguishable from the unbroken, fully substituted EBV molecules. By using appropriate dilutions of the DNA sample during stretching, we minimized the overlap of different molecules. Advantages in the labeling scheme utilized for SMARD and internal controls Studies performed by fiber autoradiography have previously shown that the results obtained using DNA fibers (such as the average size of the replicons) are significantly affected by the length of the labeling period utilized to label the replicating DNA (reviewed in Berezney et al. 2000 ). In these studies bias could also be introduced during the collection of the data as a result of the criteria utilized by the experimenter in the choice of the images analyzed. In addition, if synchronized cells are considered, the length of the labeling period also defines the potential resolution at which initiation sites can be mapped, and the estimate of the replication fork speed. Replacing the radioactive detection of the labels with fluorescence microscopy does not solve any of these problems, nor does the statistical analysis of the data. These problems are completely eliminated by the labeling scheme that characterizes SMARD ( Norio and Schildkraut 2001 ). For our experiments we utilized exponentially growing cells and labeling periods that are longer than the time required to fully replicate the genomic region of interest. In practice, since the replication of a specific genomic region can proceed differently in various DNA molecules, we utilize labeling periods that are sufficiently long to insure the duplication of even the slowest replicating molecules. In addition, only the molecules completely replicated during these labeling periods are examined. By studying this particular population of molecules, we introduce an objective criterion in the collection of the data, eliminating possible biases. Therefore, the molecules replicated during these labeling periods will faithfully represent the distribution of the replication forks in the steady-state population of replicating molecules ( Norio and Schildkraut 2001 ). Using long labeling periods, and limiting our analysis to the molecules entirely substituted with the halogenated nucleotides, also provides multiple internal controls. These controls cannot be performed when short labeling periods are used. Since the molecules are immunostained throughout their length, their images can be easily aligned to the map of the genomic region analyzed. This allows us to detect the presence of unevenly stretched molecules that can therefore be discarded. In addition, since the immunostaining is visible along the entire length of the molecules, the loss of signal caused by the breakage of some molecules is immediately revealed. The complete substitution of the DNA molecule with halogenated nucleotides also allowed us to easily detect overlaps between different DNA molecules. These overlaps can occur relatively frequently during the stretching of the molecules and their frequency increases as the density and the size of the DNA molecules increases. It is also worth noting that the presence of hybridization probes decreases the intensity of the immunostaining along the corresponding portion of the DNA molecules (see Figure 1 ). This causes a significant loss of information along the hybridized regions, but it also represents an additional control indicating that the immunostaining is indeed present on the DNA molecules that we intend to study (rather than on adventitiously overlapping molecules). Finally, our labeling scheme allows us to insure that the replication proceeded normally during the labeling of the replicated DNA and that no bias was introduced during the collection of the images. In fact, when these conditions are satisfied, the number of molecules fully substituted with IdU is expected to be very similar to the number of molecules fully substituted with CldU. These controls represent a strong proof that the images of the molecules are representative of a steady-state population of replicating molecules. Analysis of the replication intermediates by 2D gel electrophoresis at neutral pH The procedures for the enrichment of replication intermediates, 2D gel electrophoresis, and Southern analysis were essentially as described previously ( Little and Schildkraut 1994 ; Norio et al. 2000 ). Preparations of replication intermediates from Raji cells were digested with different restriction enzymes depending on the fragment analyzed: EcoRI/DraI for fragments a–e, EcoRI/EcoRV for fragments f and I, EcoRI/HindIII for fragment g, and EcoRI/XbaI for fragment h . The positions (EBV strain B95–8 coordinates) of the restriction fragments analyzed by 2D gel electrophoresis were as follows: fragment a, DraI (79202)–EcoRI (82920); fragment b, EcoRI (82920)–DraI (88865); fragment c, DraI (88865)–EcoRI (91421); fragment d, EcoRI (93162)–EcoRI (95239); fragment e, EcoRI (95239)–DraI (103226); fragment f, EcoRV (100583)–EcoRV (116863); fragment g, HindIII (110942)–EcoRI (125316); fragment h, XbaI (161383)–EcoRI (1); and fragment I, EcoRV (126415)–EcoRI (137221). The probes used to detect the restriction fragments were as follows: pHindLHI for fragments a–c, pHindE for fragments d–e, pSalF for fragment f, the pHindC fragment XbaI (121146)–BglII (120341) for fragment g, the p107.5 fragment XhoI (169423)–XhoI (167487) for fragment h, and the pHindC fragment HpaI (131959)–XbaI (133151) for fragment i . The plasmids pHindLHI, pHindE, pSalF, pHindC, and p107.5 were kindly provided by John L. Yates. Two different preparations of replication intermediates were used to study the replication patterns of fragments a–c.	/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC423133.xml
503397	A quantitative analysis of qualitative studies in clinical journals for the 2000 publishing year	Background Quantitative studies are becoming more recognized as important to understanding health care with all of its richness and complexities. The purpose of this descriptive survey was to provide a quantitative evaluation of the qualitative studies published in 170 core clinical journals for 2000. Methods All identified studies that used qualitative methods were reviewed to ascertain which clinical journals publish qualitative studies and to extract research methods, content (persons and health care issues studied), and whether mixed methods (quantitative and qualitative methods) were used. Results 60 330 articles were reviewed. 355 reports of original qualitative studies and 12 systematic review articles were identified in 48 journals. Most of the journals were in the discipline of nursing. Only 4 of the most highly cited health care journals, based on ISI Science Citation Index (SCI) Impact Factors, published qualitative studies. 37 of the 355 original reports used both qualitative and quantitative (mixed) methods. Patients and non-health care settings were the most common groups of people studied. Diseases and conditions were cancer, mental health, pregnancy and childbirth, and cerebrovascular disease with many other diseases and conditions represented. Phenomenology and grounded theory were commonly used; substantial ethnography was also present. No substantial differences were noted for content or methods when articles published in all disciplines were compared with articles published in nursing titles or when studies with mixed methods were compared with studies that included only qualitative methods. Conclusions The clinical literature includes many qualitative studies although they are often published in nursing journals or journals with low SCI Impact Factor journals. Many qualitative studies incorporate both qualitative and quantitative methods.	Background Quantitative studies provide answers or insights for many important questions or issues in health care and clinical research. Other important questions dealing with why, how, contexts, and experiences of individuals or groups, can be best addressed using qualitative methods. Other issues benefit from interleaving or integration of both research traditions. Miller and Crabtree [ 1 ], describe their experiences working in family medicine, a clinical domain where balancing qualitative and quantitative research styles benefits both patients and their families and health care professionals. They embrace holding "quantitative objectivism in one hand and qualitative revelations in another" and encourage others to use findings from both paradigms in understanding and practicing effective health care. Creswell and colleagues expand on this theme by stating that "When used in combination, both quantitative and qualitative data yield a more complete analysis, and they complement each other" [ 2 ]. Most studies in the major clinical journals have been quantitative studies. Very few qualitative studies and even fewer that combine both qualitative and quantitative approaches are published. An example of the breadth of qualitative studies and how findings and results can be combined across paradigms is a study by Jolly and Wiles [ 3 , 4 ] who used mixed methods to study a nurse-led intervention for 422 adults after myocardial infarction and 175 adults with new-onset angina in 67 general practices in the United Kingdom. Their study showed statistically insignificant results at 1 month for eating healthy food, participating in exercise programs, and successful smoking cessation. Although patients in the nurse-led group were more likely to attend a rehabilitation program (37% vs. 22%, P = 0.001) attendance was disappointingly low. The researchers interviewed a group of patients using qualitative methods and found that people felt survival after a myocardial infarction indicated that the event had not been all that serious. Health care professionals often communicated simplified data about recurrence and being "back to normal" in 6 weeks. Because of these two issues, patients felt that their cardiac problems had probably been mild and therefore were not sufficiently motivated to implement major lifestyle changes. Another example of the use of mixed methods was research done by Willms and Wilson and their colleagues [ 5 - 7 ] on smoking cessation. They found the meanings that patients who smoked attributed to their cigarettes (peer acceptance, coping during a time of stress and feeling out of control, feeling more like an adult, and smoking as more glamorous, tough, and rebellious) had more influence on cessation than did such external conditions as nicotine gum or counseling. Until the complex issues of why individuals smoke were dealt with, few were motivated to change their attitudes towards smoking and thus stop smoking. Another effective example of integrated qualitative (ethnography) and quantitative (epidemiology) methods was a study done by Borkan and colleagues [ 8 ] to determine predictors of recovery after hip fracture in elderly patients. Traditional predictors such as age, type of break, and comorbidity, were collected by using standard questionnaires. In-depth interviews were used to collect injury narratives focusing on internal explanations of the fracture, sense of disability, and view of the future after hip fracture. None of the epidemiology factors predicted successful outcomes but those who perceived their fracture as more external or mechanical as opposed to an internal or organic problem (e.g., related to chronic disease) were more likely to have good recovery. Persons who perceived their disability in the context of autonomy, independence, and connection with the outside world also showed better ambulation at 3 and 6 months than persons with a more narrow and confined view of the fracture and its resulting disability. Donovan and colleagues [ 9 ] used mixed methods to study prostate cancer screening and treatment choices to determine why study recruitment was lower than expected. Rousseau and Eccles and their colleagues [ 10 , 11 ] used qualitative methods (case interviews) to explain the limited use of computerized guidelines for asthma and angina in a primary care study done in the United Kingdom. Many other examples exist; Creswell and colleagues describe 5 additional mixed methods studies in primary care as well as provide criteria for evaluating mixed methods studies [ 2 ]. We postulate that qualitative studies, either stand-alone reports or studies with mixed methods, are occurring more frequently in health care. This paper was done to describe the publishing of qualitative studies in 1 year of clinical literature, document and present the range of content and techniques in these studies, and establish a baseline for subsequent studies. We defined our sample to include all articles published in a set of major general medical, mental health, or nursing journals during 2000. We determined how many qualitative studies were published and in which journals, and extracted design methods and healthcare content, and how often studies used mixed methods and analyses. Because the nursing literature published a higher proportion of qualitative studies in our sample we also compared studies published in nursing journals with other journals to ascertain if quantitative differences exist across disciplines in the use of qualitative methods. Our analysis is a quantitative review of qualitative studies in health care in 2000. Methods The Health Information Research Unit of the Department of Clinical Epidemiology and Biostatistics, Faculty of Health Sciences at McMaster University in Hamilton, Ontario, Canada was the editorial office for four evidence-based summary journals in 2000: ACP Journal Club (internal medicine content), Evidence-Based Medicine (family/general practice content), Evidence-Based Nursing (general care nursing content), and Evidence-Based Mental Health (mental health care content). Their purpose is to provide enhanced abstracts and commentaries on important high-quality original studies and review articles for their respective clinical audiences. To identify these studies and review articles, 6 research staff read major clinical journals to ascertain if articles were in 1 or more categories of therapy, diagnosis, prognosis, etiology, economics, clinical prediction guides, differential diagnosis, and qualitative studies and if so, did each meet predefined methodology criteria for study quality[ 12 ]. For 2000 we intensified our data collection to provide data to update and develop new clinical retrieval searching hedges for MEDLINE, PsycINFO, CINAHL, and EMBASE using methods described by Haynes and colleagues[ 13 ]. One hundred and seventy journals provided data for this article. The staff of the Health Information Research Unit has established quality criteria for the 8 categories of clinical literature that must be met before articles are judged appropriate for clinical application and publication in an abstract journal. Qualitative studies have 3 criteria: • content relates to how people feel or experience certain situations, specifically those that relate to health care • data collection methods and analyses are appropriate (primary analytical mode is inductive rather than deductive) • units of collection and analysis are ideas, thoughts, concepts, phrases, incidents, or stories that become categories or themes. The reading methods have been developed during the past 13 years and inter-rater reliability kappa (chance adjusted agreement) for identifying categories and applying criteria is consistently > 80%. For this paper, KAM, one of the readers, analyzed the qualitative studies. Qualitative systematic reviews were excluded leaving only reports of original studies. These were assessed to extract journal title, qualitative study type, data collection methods, research question, persons studied, setting, and disease or health condition considered. In addition, studies with mixed methods were further analyzed although we did not use stringent criteria for assessing the quality [ 2 ] of the combination of methods. We identified mixed methods articles using a loose criterion of "some numerical or statistical analysis of quantitative data or qualitative data that had been turned into quantitative data". (An example of quantifying qualitative data is the study done by Borkan and colleagues [ 8 ] on hip fracture.) The analysis had to be fairly substantial–for example, a simple descriptive analysis of baseline demographics of the participants was not sufficient to be included as a mixed methods article. In addition, Giacomini and Cook [ 14 , 15 ], as part of the Evidence-Based Working Group in the Users' Guides to the Medical Literature, describe attributes that they have identified as belonging to high-quality qualitative studies: participant selection, data collection, and analysis methods. These aspects were also extracted for analysis in this report. Data were taken from article abstracts and if needed, the full text was reviewed. Methodologies assessed were phenomenology, grounded theory, ethnography, case studies, narrative analysis, participant action, critical incident techniques, and discourse analysis. Author descriptions were used and if an additional methodology was found it was added to the list of types using definitions and descriptions from the Handbook of Qualitative Analysis, 2 nd edition by Denzin and Lincoln [ 16 ]. Data collection and sampling procedures were also extracted. Multiple designations were allowed. To assess the reproducibility a random 10% (n = 35) sample of citations was reviewed using predefined decision rules by another researcher trained in research methods. Results The 170 journals included 60 330 articles of which 31 496 (52%) contained original data or were review articles. 3830 of these (6%) passed criteria for being high-quality and clinically relevant in 1 of the 8 categories. 367 articles met quality criteria for original studies or reviews of qualitative studies. Table 1 lists the journals that published at least 1 qualitative study. Twelve systematic reviews were excluded leaving 355 qualitative studies for assessment. Approximately 0.6% of all articles in the 170 journals and 9% of all high-quality, clinically relevant studies were qualitative studies. The reproducibility of the categorization was measured by kappa (chance adjusted agreement): 0.92 for disease/condition, 0.83 for groups studied, 0.81 for setting, 0.73 for data collection, and 0.63 for data analysis type. The agreement for data analysis type was disappointing but not surprising in that 20% of the studies did not label their analyses necessitating assignment of analysis type by the data extractors. Agreement was low for participant selection methods (kappa 0.5) and therefore data on participant selection methods are not reported. The 355 qualitative studies appeared in 48 journals (mean 7.4 articles per journal, range 1 to 86). These 48 journals were only 28% of the 170 clinical journals being read. Most of the qualitative studies were published in nursing journals: The 17 nursing titles included 214 qualitative studies (61% of all of the qualitative studies). Few qualitative studies were published in the high-circulation, general healthcare journals. Using SCI Impact Factor ranking for 2000, only 4 of the top 20 journals (Table 2 ) published qualitative studies. These 4 journals published 15 qualitative studies with BMJ publishing 12. The highest-ranking journal with qualitative studies was Annals of Internal Medicine , ranked number 6. JAMA , ranked number 2, published articles about qualitative studies in 2000 [ 14 , 15 ] but did not publish any qualitative studies. Mixed qualitative and quantitative studies 37 qualitative studies (11%) included qualitative and quantitative methods and analyses. These were published in 17 journals with only 1 article in BMJ from the top 20 journal titles in Table 1 . Social Science and Medicine published 10 of these mixed methodology studies–the most of any title studied. Content Content of the studies is shown in Table 3 . Many studies dealt with a range of participants and settings. Patients (56%), family (22%), and other non-health care professionals (14%) were studied more often than health professionals (nurses (21%), physicians (11%), and others (5%)). Non health care settings occurred more often with home or similar settings being studied in 44% of studies and other community settings in 16%. Health care settings were the hospital (25%), clinic (17%), nursing home (5%), and the emergency department (2%). Disease/condition breakdowns represented common health care situations: cancer (11%), mental health (10%), pregnancy and childbirth (9%), cerebrovascular disease (10%), general issues such as vaccinations or Internet use, and nonspecific spectrum of diseases (e.g., all patients in a clinic) (12%). Many uncommon issues were also assessed. For example, Tongprateep [ 17 ] reports a phenomenology study designed to help nurses better understand essential elements of spirituality and health among rural Thai elders. Analysis of the 37 articles with mixed methods showed similar patterns for settings, persons studied, and disease/condition evaluated except that more physicians were studied (P < 0.025) and more situations dealing with injury (P < 0.001) were evaluated. For the 211 articles in Nursing journals, very little difference was also seen except that fewer physicians were studied (P < 0.001) and more studies were done outside clinical settings (P < 0.001). Phenomenology (37%), grounded theory (35%), and ethnography (18%) were used most often with some case studies (7%), narrative analysis (6%), participant action (3%) research, critical incident techniques (1%), and discourse analysis (1%) (Table 4 ). More than one qualitative method was used in 8% of studies. This pattern of methodology choice was similar for the 37 mixed methods studies and the 211 Nursing articles except that mixed studies methods used relatively more case studies and the Nursing studies used fewer of them (P < 0.025). The mixed methods studies did not included participatory action research, critical incident technique, or discourse analysis studies, methods that could be difficult to combine with quantitative studies. Semi-structured interviews were used (77%) with some focus groups (18%) and observation (14%). These methods are major data gathering techniques in qualitative studies. Questionnaires (7%), document analysis (6%), and structured (4%) and unstructured interviews (1%) were used less often. For mixed methods studies, patterns were similar although questionnaires were used more frequently (24% vs. 7%, P < 0.01). Nursing studies did not differ for data gathering techniques. Sampling is important in all studies–often no single right way exists for a study question. Purposive, snowball, and theoretical sampling are often used in qualitative studies and random and consecutive sampling for quantitative studies. All methods were represented in this analysis but the breakdowns are not reported because of low inter-rater agreements for categorization and missing author information. Discussion In 2000 the major clinical journals published many qualitative studies–approximately 9% of all high-quality, clinically relevant articles. Most of the qualitative studies were reports of original research although 12 (3%) were systematic reviews. Most of the qualitative studies were in nursing journals although some medical journals such as BMJ and Annals of Internal Medicine also published several. Three of the high circulation medical journals ( New England Journal of Medicine, Lancet, and JAMA ) and 16 of the top 20 clinical journals, based on SCI Impact Factors, did not publish any qualitative studies. This is likely a reflection on the emphasis on a positivist, numerical approach that many of these journals embrace. The difference in proportion of qualitative studiers published in nursing journals is probably because of two historical, but linked factors. Qualitative studies have roots in women's studies and the nursing profession has always dealt with the patient as much more of a whole person rather than basic sciences facts and numbers. Both of these factors lead to more emphasis on understanding and embracing qualitative methods for research and practice. This view is substantiated by the fact that MEDLINE indexes most of the qualitative studies under the term Nursing Methodology Research until 2003. A substantial proportion of the qualitative studies (11%) included both qualitative and quantitative (mixed) data. In general, these mixed methods studies were similar to the single methodology studies except they did more assessments of physicians and relied more on questionnaires to gather data for analysis. The presence of these mixed methods or multipardigmatic studies as described by Miller and Crabtree [ 1 ] and Creswell [ 18 ] is encouraging for those who espouse harnessing methodologies appropriate for exploring, explaining, and interpreting the complexities and ranges of issues in health care practice and research. It is also interesting comparing qualitative studies in Nursing and non-Nursing journals. Regardless of the differences in proportion of qualitative studies published, from a content point of view few differences exist between the Nursing and non-Nursing journals except that more physicians were studied in the non-Nursing journals and fewer studies were done in clinical settings–not unsurprising findings. This indicates that the content and methods of qualitative studies seem to be similar across disciplines or if the methods are combined with quantitative methods. This review of the publication of qualitative studies is limited in several ways. The proportion of journals studied was very low in relation to the total number of journals published. MEDLINE indexes over 4000 journals and this number is still a relatively small proportion of all journals that deal with health care. In addition, all of the journals searched were published in English so we do not know about qualitative studies in other languages. Although our criteria were relatively strict for including qualitative studies, our criteria for mixed methods studies could certainly have been stronger. We did not count the number of high-quality quantitative studies that could have included some qualitative analyses. We studied only 1 year of publishing; much could have changed since 2002. Qualitative studies provide insight into social, emotional, and experiential aspects of health and health care and as such, they have an important place in understanding health and health care. Hopefully more studies will be published and more will be published in the high impact (high circulation) journals. This paper provides a basis for measuring increases. Conclusion Qualitative studies are being done and are published in a wide range of healthcare journals. These journals however are not the highest impact journals. It is encouraging to see that the number of qualitative studies that were published in 2000 and also the number of studies that combined qualitative and quantitative methods. More can be done however to complete and publish qualitative studies, and where appropriate, integrate the best of both methodologies. Both qualitative and quantitative researchers and clinicians need to work together to make this happen. Journal editors can also encourage submission of qualitative and mixed methods studies and facilitate publication of those they do receive. List of Abbreviations SCI ISI Science Citation Index Competing interests None declared. Author Contributions This work was done in partial fulfillment of PhD requirements for KAM. Both authors have supplied intellectual input in designing and implanting the survey. KAM has collected and analyzed the data and both authors have contributed to writing the paper and agree on its content. Pre-publication history The pre-publication history for this paper can be accessed here:	/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC503397.xml
549513	Quantification of the magnification and distortion effects of a pediatric flexible video-bronchoscope	Background Flexible video bronchoscopes, in particular the Olympus BF Type 3C160, are commonly used in pediatric respiratory medicine. There is no data on the magnification and distortion effects of these bronchoscopes yet important clinical decisions are made from the images. The aim of this study was to systematically describe the magnification and distortion of flexible bronchoscope images taken at various distances from the object. Methods Using images of known objects and processing these by digital video and computer programs both magnification and distortion scales were derived. Results Magnification changes as a linear function between 100 mm (×1) and 10 mm (×9.55) and then as an exponential function between 10 mm and 3 mm (×40) from the object. Magnification depends on the axis of orientation of the object to the optic axis or geometrical axis of the bronchoscope. Magnification also varies across the field of view with the central magnification being 39% greater than at the periphery of the field of view at 15 mm from the object. However, in the paediatric situation the diameter of the orifices is usually less than 10 mm and thus this limits the exposure to these peripheral limits of magnification reduction. Intraclass correlations for measurements and repeatability studies between instruments are very high, r = 0.96. Distortion occurs as both barrel and geometric types but both types are heterogeneous across the field of view. Distortion of geometric type ranges up to 30% at 3 mm from the object but may be as low as 5% depending on the position of the object in relation to the optic axis. Conclusion We conclude that the optimal working distance range is between 40 and 10 mm from the object. However the clinician should be cognisant of both variations in magnification and distortion in clinical judgements.	Introduction The flexible bronchoscope has been used in pediatrics for more than 20 years [ 1 ] yet there are only a limited number of publications on the systematic examination of the physical properties of magnification and distortion found in endoscopes of any size, let alone bronchoscopes specific for the pediatric sized airways. [ 1 - 6 ]. Some understanding of specific bronchoscopic magnification and distortion is an important issue to the clinician as these instruments are being used more regularly to define the nature and severity of airway lesions such as tracheal stenosis and malacia disorders from which important medical and surgical decisions ultimately follow[ 7 - 14 ]. The pediatric bronchoscope comes in a variety of sizes but the resultant magnification from these instruments depends not only on the size but also on the image processing that occurs. Magnification of an object refers to the virtual or real visualized and measured increase in size of that object after rays of light from that object have passed through a lens system. Magnification varies across the field of view and with the distance of the lens from the object [ 15 ]. Distortion refers to the variations of the faithful representation in scale and perspective of the various elements of the object and its plane of context[ 15 ]. There are various forms of distortion including geometrical, curvilinear, anamorphic and perspective distortion recognised in photometrics [ 15 ]. Geometrical distortion refers to the changes in the peripheral details because of elongation of the elements of an object. Simply, it refers to the distortion that is created when there is increasing obliquity of the angle of viewing of the object. Curvilinear distortion is described where straight lines are rendered as curved either inward (pincushion) or outwards (barrel) curves. This form of distortion is a result of asymmetry of lens configuration and is a feature of bronchoscopes and indeed most endoscopes in general[ 2 , 5 , 16 ]. The ultimate image quality of any bronchoscope is dependent on the lens characteristics in general and the subsequent transmission of the image to the image processor and display module. The more recently developed bronchoscopes such as the flexible videobronchoscope (FVB's) have moved away from fibreoptics to transmit images. They use charge coupled devices (CCD's) which are mounted adjacent to the lens thus essentially converting the image to electrical energy at the lens. This early conversion of the image has the effect of producing "less noise" thus allowing greater image clarity. Despite the latter, all of these issues mentioned are important to the clinical use of FVB's and subsequent patient management and as the Olympus company was unable to provide the necessary details on magnification and distortion, we undertook to quantify these effects in order to enhance our decision making quality during FVB procedures. The aim of this paper is to define the magnification and distortion in a commonly used pediatric video bronchoscope: the Olympus BF Type 3C160. Methods Equipment A flexible bronchoscope (Olympus BF Type 3C160, Olympus, Tokyo, Japan) and light source (Olympus Evis Exera CLV-160, Olympus, Tokyo, Japan) and processor (Olympus Evis Exera CV-160, Olympus, Tokyo, Japan) were connected to a TV monitor (Sony HR Trinitron, Sony Corporation, Shinagawa-ku, Tokyo, Japan) and a colour video printer (Sony Mavigraph, Sony Corporation, Shinagawa-ku, Tokyo, Japan) as supplied by the Olympus company. The image signals from the Evis Exera CV-160 were co-recorded by a digital camera (Sony Mini DV Digital Handycam, Sony Corporation, Shinagawa-ku, Tokyo, Japan). These digital video images were viewed and stored as uncompressed 640 × 480 pixel, millions of colours as TIFF files. These images were then analysed by an image processing program (Image J, Wayne Rasband, National Institute of Health, USA) and a computer (MAC Power Book G4. Apple Inc,, Cupertino, Ca., USA). Measurement of Magnification and Distortion A precisely measured object consisting of 4 concentric circles with emanating radii on 1 mm × 1 mm graph paper was created and drawn to precision using design software (Auto CAD, Autodesk, San Rafael, Ca, USA) (Fig 1 ). These circles were 2, 6, 10 and 20 mm diameter respectively. A precision made device that fixed the hand piece and supported the length of the bronchoscope so that the central or geometric axis of the bronchoscope (Fig. 2 ) would align with the centre point of a moveable screen was engineered by Queensland Radium Institute (QRI, Brisbane, Australia) engineering. The device's moveable stage allowed the object to be positioned over a range of distances from 100 mm to 3 mm without moving the scope itself. The measured distances were verified with electronic callipers to a degree of accuracy of 0.02 mm. The calibration process for calculation of magnification and distortion was performed at the distance of 100 mm, the point at which the object to image ratio was 1. All subsequent ratios were therefore larger than the calibration value and represented the magnification of the object. Figure 1 Photographic Object: A square of graph paper 3 cms × 3 cms with 1 mm × 1 mm units and concentric circles of varying diameter (2 mm, 6 mm, 10 mm and 20 mm) and diameter markings (AE, BF, CG, DH) all generated by Auto CAD. Figure 2 The end view of the tip of the Olympus BF Type 3160 bronchoscope displaying the "off set" lens and lights and the geometric centre. Magnification was calculated by repeatedly measuring the image diameters AE, BF, CG, DH across the 4 concentric circles (Fig. 1 ) at distances of 100, 80, 60, 40, 30, 25, 20, 15, 10, 5 and 3 mm and then dividing these measurements by the corresponding value of the object. These magnification measurements were repeated in 3 separate experiments and the mean value for all of the diameters of that particular circle at that particular distance was taken as the final magnification for that circle at that distance. Between the distances of 100 mm and 40 mm the calculations of the inner circle were incomplete or impossible because of the effects of the light spot obscuring the object (Fig. 3a ). Similarly as the lens approached very close to the object (5 mm to 3 mm) the outer circles became incomplete or outside the field of view (Fig. 3b ). Figure 3 Effects of distance from object on the image appearance within the Field of View (FOV) and Light intensity obscuring parts of the image. Note the barrel appearance of the graph paper at the periphery in image "A" while in image "C" these effects are clearly offset or asymmetrical. Geometrical distortion was defined as the radial length ratios from each interlocking circle with radii A taken as the denominator for all subsequent radial measurements within the outer circle and then other elements of radius A for the other circles. The uniformity of the ratios was then tested in quadrant or sectors: a sector or quadrant being an area defined in accordance with mathematical convention[ 17 ] whereby sector 1 is the NNE sector and sector 2 is NNW continuing in a anticlockwise direction for sectors 3 and 4. The comparisons were as follows: sector 1 with sector 3; sector 2 with sector 4 or the diagonal sectors with their respective opposites. This provided an understanding of the homogeneity of the distortion profile or geometrical distortion and also gives reference to curvilinear distortion. The changes in magnification across the field of view was further defined as the differences in calculated measurements for a known distance of 2 mm across the field of view at a fixed distance (15 mm) from the object. These measurements were made in the horizontal plane. As with the magnification calculations, the mean distortion value from the entire set of 3 experiments was taken as the defined distortion at the particular distance from which the measurements were made. All of the aforementioned experiments were then repeated on 3 separate Olympus BF Type 3C160 bronchoscopes. The experiments were performed separately with the central geometrical axis of the bronchoscope and then the optic axis of the bronchoscope (Fig. 2 ) aligned with the centre point of the stage and object. Statistics Magnification and distortion were expressed as mean values and comparisons between geometric and optical axis measurements were assessed by the Student's t test. A repeat measures ANOVA was used to assess the effects of distance, circle and sectorial effects. The reliability of the repeated measurements of magnification between tests and between bronchoscopes was assessed by intra-class correlation. Data storage and statistical calculations were performed on SPSS for MAC version11 (SPSS Inc., Chicago, Il, USA). Results The mean linear magnification for the bronchoscopes aligned along the central geometrical axis of the bronchoscope over the range of the depth of field from 100 mm to 3 mm is shown in Additional file 1 . The magnification progressively increased in a linear fashion between 100 and 10 mm from the object and then exponentially increased to 40 times between the distances of 10 mm and 3 mm from the object. The magnification factor was approximately ×10 at 10 mm and ×40 at 3 mm from the object. The graphic representation of these values as tightly fitting linear becoming exponential curves is displayed in Figure 4 . The goodness of fit (R 2 value) for each of the curves for circles A,B,C,D was 0.999, 0.999, 0.999 and 0.999 respectively. When the bronchoscope was aligned to the object through the optical axis, the magnification was reduced generally but by as much as 22% at 5 mm and 14.5% at 10 mm from the object ( Additional file 2 ). Figure 4 Effects of position within the Field of View: Magnification changes from 40 mm to 3 mm from the object displaying the variable exponential changes in magnification close to the object. The mean intra-class correlation alpha level for the repeated measurement of magnification ranged from 0.9369 (95%CI: 0.6167 to 0.9878) to 0.9811 (95%CI: 0.8827 to 1.000). These intraclass correlation data indicate and support the goodness of fit of the regression curves. The variability in magnification measurement between the bronchoscopes was extremely small and there are virtually no differences between the bronchoscopes (Table 1 ). The reliability coefficient average alpha value for the measurements from 3 separate bronchoscopes assessed at 10 mm was 0.9996 (95%CI: 0.9991 to 0.9999). Table 1 Inter-scope comparison of magnification with optical axis aligned. Bronchoscope No. Distance mm Circle A Circle B Circle C Circle D 324 30 3.0938 3.1601 3.2913 292 30 3.0984 3.1963 3.3258 039 30 3.0820 3.2083 3.3191 324 15 5.5356 6.2137 6.4443 6.4918 292 15 5.5016 6.1527 6.3062 6.3832 039 15 5.4997 6.1107 6.6364 6.4358 324 10 8.9428 9.5024 9.6755 292 10 8.5663 9.0650 9.3453 039 10 8.5663 9.0649 9.3449 324 5 17.6819 19.8996 292 5 15.9314 17.0974 039 5 15.6890 16.7272 324 3 32.4274 292 3 26.1051 039 3 24.7775 The across field magnification was greatest in the centre of the field and least at the periphery with some 38.5% reduction in magnification across a 20 mm diameter object but only 15.4% across an object of 6 mm diameter (Figure 5 ). The overall distortion including geometric distortion for the bronchoscopes aligned along the central axis ranged from near zero at 40 mm, 5% at 5 mm but at 3 mm it had risen to 30% (Figure 6 ). When the object very closely approximated the optical axis in alignment (object centre is within 2 mm of the lens's optical axis), the overall magnification factors changed (see additional file 2 ) but importantly this value of distortion at 3 mm from the object was markedly reduced to 5%. Distortion was significantly different in different quadrants when the scopes were aligned along their central geometrical axes to the object. Figure 5 Across lens magnification: the near linear changes in magnification across the horizontal plane from the centre to periphery of the lens. Figure 6 Mean central distortion ± 95% CI for the central or geometrical axis aligned bronchoscope. The appearance of the curvilinear distortion is shown in Figure 3 and clearly displays barrel distortion. The measurements show that curvilinear distortion was different for the different parts of the field of view. In particular, the magnification at the centre of the field of view at any distance was different to the periphery (Figure 5 ). These differences were not evenly distributed across the field of view, however the differences between quadrants were extremely small. A univariate analysis found significant effects for ring, distance and sector analysis. However the three factor repeat measures ANOVA analysis revealed no significant differences between measures by sector (adjusted p-value using Box's conservative test = 0.13). When the data was stratified by distance, all 8 p-values were not significant (0.13 to 0.31), indicating that a mild effect by sector may exist. Discussion This is the first reported study detailing systematically the magnification and distortion factors that surround the optic properties of a commonly used pediatric video-bronchoscope. The magnification progressively increases in a linear fashion over the distances of 100 mm to 10 mm with changes from 1 to near 10 fold magnification. However below 10 mm, the magnification changes exponentially thus making appreciation of the actual or real size of the object very difficult even when the distance from the object is known [ 3 , 5 , 6 ]. This study also shows that the magnification of the instrument changes is accordance with the axis of orientation of the bronchoscope to the object. These data indicate that the optimal operating distance is between 40 mm and 10 mm where the mean magnification is linear and is 3 to 9.5 fold and the distortion of any form can be under 5%. However it also means that it is important for the bronchoscopist to maintain a "perspective of operating distance" or distance from the object when judgements of size are to be made. This study has also shown that the Image J and BTV programs and MAC power book G4 can be readily interfaced with Olympus BF Type 3C160 to produce an objective measurement package. In addition this system could be readily adapted to most image acquisition systems and varieties of flexible bronchoscopes types (FVB and fibre optic) that must be in clinical use around the world. With respect to distortion, this study has shown that geometric distortion can result with very significant errors in size perception with up to 30 percent distortion occurring if the lens is placed very close to the object (<5 mm) and the object is not aligned through the optical axis. This could be corrected by movement of the object closer to the line of the optical axis of the lens. However there are no markers on the bronchoscope to indicate that this position has been gained. At close proximity to this point there is separation of the light beams, however while the use of this is possible in laboratory settings, it is not currently possible in the clinical context where there are differences in reflectance and absorption of light. The second issue is that barrel distortion is generally regarded as homogeneous and independent of working distance however when the lens is offset as it is in this type of instrument it cannot be, given the differences between the geometrical and optical axes of the bronchoscope. In the clinical context these issues indicate that clinicians must be aware that their judgements always contain "distortion effects" but particularly when working very close to the object eventhough the image appears clear and is centred on the screen. The limitations of this study are the fact that the object positioning could only be moved with micrometer precision in the longitudinal plane. However an instrument that allowed for lateral movement of the object with micrometer precision may have produced less distortion but in the clinical context such precision could only ever be transient and thus these methods have allowed for a greater appreciation of the extent and complexity of distortion. Even though computer based systems have been used to correct for some of the elements of distortion[ 3 , 5 , 6 ] none exist in routine use within the current hardware of the bronchoscopes themselves. In those that have been described [ 5 ], it is unclear as to what type of distortion has been corrected. These authors have aligned their image centre or Field of View (FOV) as the centre for their measurements. However as shown in this study, correction for the offset position of the lens is important to any real measurement when the object is closer than 5 mm. If this is not done, then correction formulations will overestimate the levels of distortion in some instances, as there would be a high likelihood that the objects of view would pass through all axes of orientation during an in vivo procedure. Despite these arguments, the overall distortion of any type is extremely low and in the clinical context these values are highly acceptable as far as measurements are concerned. Also it is unlikely that an object will be viewed at this distance for any period of time given the "blurring effects" of the distortion. However during passage of the bronchoscope across lesions, clearly such transients in magnification and distortion could be misleading. The clinical relevance of this work may not be obvious. However, we argue that it has a number of important clinical applications and that all bronchoscopists should be cognisant of these factors, which may influence their judgement of the size of lesions. Firstly, the distance of the tip of the scope to the lesions being assessed must be measured and not approximated by judgement as distances under 5 mm may result in very considerable differences in magnification and thus the estimates of size or even presence or absence of a lesion. The most obvious example of this is in assessments of a curvaceous left main stem bronchus that may initially appear as malacia but appears to become "normal" as it is viewed at a closer distance. Indeed in defining malacia a viewing or operating distance must be stated otherwise these perspective difficulties will prevail and comparative statements and judgements become meaningless. Another example occurs when assessing malacia or airway cross sectional area changes across the respiratory cycle. In that scenario the movement of the object is away from the lens during inspiration making the image relatively smaller than is real for the increase in lung volume. These optic effects combined with the venturi effects from the bronchoscope itself on reducing intraluminal pressure again compound the issue and reduce the image size or its relative changes. Bronchoscopists need to be aware of these apparent paradoxical effects, that the lumen is greatest at the end of expiration. Indeed at this point in time and until the distance of airway movement can be measured in vivo, measurements can only be made precisely at one point in the respiratory cycle, and that generally is the end expiratory point. If end inspiration is used there is likely to be intrusion of the mucosa into the field of view during the subsequent expiration. The most recent published example of most of these effects can be seen in the images provided in that paper by Okasaki et al [ 18 ] where the instrumentation magnification is not known, the measurements are not calibrated at a defined time in the respiratory cycle, the viewing distance is not described and there is clearly image change across the induced respiratory cycle. Secondly, paradoxical concepts also apply to the use of the flexible bronchoscope in assessment and or removal of foreign bodies situated peripherally in the airways. When dealing with peripheral or markedly angulated bronchial branches it is not always possible to position the image in the centre of the screen. Here estimates of size may be misconstrued by the interactions of variable magnification across the lens, curvilinear distortion and potentially geometric distortion. These effects and the "iceberg effects" of the presented foreign body size usually mean that the foreign body appears smaller than it really is and the inexperienced bronchoscopists might tend to dismiss the object as trivial or inconsequential. To the contrary attempts to remove it should be maximized. Finally, in terms of research, defining the position of the bronchoscope in terms of the lesion being assessed and photographed is vital if realistic comparisons are to be made across time or between groups. In this regard techniques that allow for distance and lesions measurement and assessment need to be developed and integrated into routine clinical use. Using our technique described in this paper, we are quantifying airway size in a variety of airway lesions (eg tracheobronchomalacia) associated with significant respiratory morbidity. The fact that performance of each bronchoscope was remarkably similar in terms of magnification and distortion measurements obviously reflects company's production quality. In a clinical session where a number of bronchoscopes might be used and in research where inter-group comparisons might be desired, this information suggests that a significant level of confidence can be maintained in perception terms for the bronchoscopists. This does not negate the need for individual bronchoscope calibration and for objective and accurate measurements in bronchoscopic work. Detailed data from the Olympus Company was not available for comparison; however our own validation experiments show these data to be correct and accurate. Despite the latter, as there are now many instruments available to the clinician, we suggest that companies and manufacturing regulators make readily available the magnification and distortion characteristics of each instrument type or size so that more effective clinical appraisals can occur. Although this study has shown that the Olympus BF Type 3C160 video-bronchoscopes produce remarkably consistent magnification across the working ranges of 100 mm to 3 mm from the object, an understanding of the influence of distance on magnification and distortion of the image obtained by a flexible bronchoscope is an essential step in the development of an invivo technique of measuring airway sizes. We have provided graphic appreciation of these effects and the importance of optical versus geometric axis orientation. The optimal working distances for this bronchoscope are between 40 mm and 10 mm from the object. The study also provides reasonable working magnification factors for this type of bronchoscope and as such could allow for a better appreciation of real or actual airway sizes. Supplementary Material Additional File 1 Mean magnification and the mean whole of field magnification at defined distances. Click here for file Additional File 2 Comparison of bronchoscope axis and optical axis aligned magnification measurements. Click here for file	/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC549513.xml
535802	Family structure and risk factors for schizophrenia: case-sibling study	Background Several family structure-related factors, such as birth order, family size, parental age, and age differences to siblings, have been suggested as risk factors for schizophrenia. We examined how family-structure-related variables modified the risk of schizophrenia in Finnish families with at least one child with schizophrenia born from 1950 to 1976. Methods We used case-sibling design, a variant of the matched case-control design in the analysis. Patients hospitalized for schizophrenia between 1969 and 1996 were identified from the Finnish Hospital Discharge Register, and their families from the Population Register Center. Only families with at least two children (7914 sibships and 21059 individuals) were included in the analysis. Conditional logistic regression with sex, birth cohort, maternal schizophrenia status, and several family-related variables as explanatory variables was used in the case-sibling design. The effect of variables with the same value in each sibship was analyzed using ordinary logistic regression. Results Having a sibling who was less than five years older (OR 1.46, 95% CI 1.29–1.66), or being the firstborn (first born vs. second born 1.62, 1.87–1.4) predicted an elevated risk, but having siblings who were more than ten years older predicted a lower risk (0.66, 0.56–0.79). Conclusions Several family-structure-related variables were identified as risk factors for schizophrenia. The underlying causative mechanisms are likely to be variable.	Background Several factors relating to family structure have been suggested as risk factors for schizophrenia. They include sibship size, number of older siblings, birth order in the sibship, age difference to older siblings, and age-at-onset, and sex of the affected siblings. These family-related risk factors may provide useful hints about the nature of environmental and genetic risk factors for schizophrenia [ 1 ]. In several studies, siblings of female probands have had a higher risk of developing schizophrenia than siblings of male probands in several studies, with the number of cases varying from 123 to 500 [ 2 - 7 ]. The effect of the age-at-onset of proband is less clear. Some studies have detected no effect [ 5 , 8 ], but in one, early age-at-onset of proband predicted a higher risk for siblings [ 4 ]. There are mixed results about the effect of birth order on risk. Two studies reported lowest risk for the firstborns [ 9 , 10 ], and in a Swedish population based study with 167 cases, male subjects fourth or higher in birth order had over a three-fold risk compared to males with lower birth order [ 11 ]. On the other hand, a very large scale Danish cohort study involving 2669 cases showed no effect for birth order [ 12 ]. A significantly higher risk for first-born males was reported from a Finnish cohort study with 100 cases [ 13 ], and in accordance with this, a recent study from Pakistan with 453 cases reported highly significant excess of firstborns among patients with schizophrenia [ 14 ]. The number and age of older siblings at the time of birth is hypothesized to be connected to the risk of schizophrenia through infection pressure. In a Swedish register-based study (270 cases), having siblings who were 3–4 years older was associated with increased risk (relative was 2.02) [ 10 ]. The Danish cohort showed no effect of birth order on risk, but found that having less than two years age difference to the nearest older or younger sibling increased the risk of schizophrenia [ 12 ]. However, in an Irish case-control study with 2969 cases and 5904 controls, the number of older siblings did not predict a diagnosis of schizophrenia over neurosis [ 15 ]. Several recent studies have found that advanced paternal age is a risk factor for schizophrenia [ 16 - 18 ]. All risk factors concerning birth order, family size, number of older or younger siblings, and paternal and maternal age are correlated. For example, the number of older siblings is dependent on both the individual's birth order and family size, which are connected to maternal age. If all these correlated variables are not in the model simultaneously, spurious associations may be found. Therefore, a large, population-based data set that contains a sufficient number of sibships with variable structures is needed to analyze family-related variables. Our aim was to investigate which of these variables were risk factors for schizophrenia in Finland among those who were born between 1950 and 1976. Methods Register data, study design and statistical analyses Using the Finnish Hospital Discharge Register of Finland, we obtained information on all hospitalizations with a diagnosis of schizophrenia (ICD-8 295, ICD-9 295, ICD-10 F20*). The information was sought for all Finns born between 1950 and 1976 for the period between 1969 and 1996. Sibship was defined as a group of individuals having the same mother. Since many of the variables we examined are related to maternal characteristics, we considered it essential that each sibship has the same mother. It should be noticed that generally the mother has a stronger influence on the child than the father especially in early childhood: e.g. mother's womb provides the fetal environment, and after birth the mother has more of a close physical relationship to the child through breast feeding. Siblings and parents of the probands were identified using the data from the Population Register Centre, and their information was then linked back to the Hospital Discharge Register to obtain data on relatives' hospitalizations. Only those who were born between 1950 and 1976 were qualified as probands, but their younger siblings were included in analyses as family members. We assumed our ascertainment was complete, because the two registers used, the Population Register and the Finnish Hospital Discharge Register, are nationwide and accurate [ 19 ]. The reliability of register diagnoses of schizophrenia is good [ 20 - 22 ] and according to previous studies, more than 95 percent of individuals with schizophrenia receive hospital treatment at some point in their lives [ 22 ]. Altogether, we identified 17549 sibships with 35481 siblings, of whom 31004 were born between 1950 and 1976. There were 12 siblings in the largest family, and the highest number of cases in one family was six (Table 3 ). All families that contained only one child (n = 7947) were excluded from the study because almost all variables we investigated required there to be more than one child in the family. Table 3 A family study of schizophrenia in Finland in birth cohorts born from 1950 to 1976. Family structure of families (N = 2605) with at least one sibling with schizophrenia; also siblings born before 1950 and after 1976 are included. N. of cases in the family Family size 2 3 4 5 6 7 8 9 10 11 12 1 3289 2012 984 369 171 62 22 9 5 1 0 2 282 271 190 82 35 16 12 5 1 0 0 3 0 22 26 11 10 8 4 0 1 0 1 4 0 0 4 2 1 1 1 0 0 0 0 5 0 0 0 1 0 1 0 0 0 0 0 6 0 0 0 0 0 0 1 1 0 0 0 We used two types of explanatory variables: variables that were unique for each person, and variables that had the same value for all members of the sibship. The unique variables were sex, birth cohort, birth order, number and ages of older siblings at the time of birth, and the age of mother and father at the time of birth, and the shared variables were the lowest age-at-onset in sibship, sex of the proband with lowest age-at-onset in the sibship, and the sibship size, and whether the mother had schizophrenia or not. Study design and statistical analyses We investigated the effect of unique variables using the case-sibling design [ 23 ], in which each case was matched with one or more unaffected siblings. We modeled the data with conditional logistic regression using each case in the family as the conditioning unit, which is the standard method for case-sibling design [ 23 ]. The sibling with the lowest age-at-onset of schizophrenia was the first case or proband and all unaffected siblings born between 1950 and 1976 and alive at age when first sibling had his/her first illness episode, were chosen as controls. Thus, the matching was carried out with respect to age-at-onset. If there were later cases in the family, controls were chosen in the same way but excluding those siblings who already had developed schizophrenia. Thus, a sibling could serve as a control before developing schizophrenia. The number of cases with at least one control was 9013 and the number of controls was 12046. As shown in Table 4 , our data included a substantial number of families with more than 5 siblings. Table 4 Basic characteristics of sibships. no schizophrenia (%) schizophrenia (%) All Sex male 4636 47 5207 52 9843 female 7410 66 3806 34 11216 Siblings under 5 years no 5496 52 4876 48 10372 yes 6550 61 4137 39 10687 Siblings 5–10 years no 7727 52 7172 48 14899 yes 4319 70 1841 30 6160 Siblings over 10 years no 10094 55 8377 45 18471 yes 1952 75 636 25 2588 Birth order 1 2934 44 3781 56 6715 2 4297 57 3260 43 7557 3 2629 68 1262 32 3891 4+ 2186 75 710 25 2896 Birth cohort 1950–1955 3756 51 3648 49 7404 1956–1960 3566 57 2689 43 6255 1961–1965 2825 62 1748 38 4573 1966–1976 1899 67 928 33 2827 Maternal age under 20 441 45 546 55 987 20–25 2509 50 2462 50 4971 26–30 3419 56 2692 44 6111 31–35 2896 62 1778 38 4674 over 35 2781 64 1535 36 4316 Paternal age under 20 101 49 106 51 207 20–25 1281 49 1355 51 2636 26–30 2727 53 2413 47 5140 31–35 2765 60 1879 40 4644 over 35 4159 63 2407 37 6566 not available 1013 54 853 46 1866 Sex of youngest proband male 6880 57 5187 43 12067 female 5166 57 3826 43 8992 Lowest age-at-onset in sibship under 20 2797 54 2351 46 5148 20–25 4410 57 3350 43 7760 26–30 2930 59 2014 41 4944 over 30 1909 60 1298 40 3207 Family size 2 3623 46 4233 54 7856 3 3623 59 2551 41 6174 4 2373 65 1284 35 3657 5+ 2427 72 945 28 3372 Mother schizophrenic no 11537 57 8561 43 20098 yes 509 53 452 47 961 Shared variables, which were the lowest age-at-onset in sibship, sex of the proband with the lowest age-at-onset in the sibship, and the sibship size, were modeled using ordinary logistic regression without conditioning. The probands, who were the affected siblings with the lowest age-at-onset in the families, were not included in this model. We tested the significance of the interactions between sex and other variables, and interactions between sex of the proband and other variables using likelihood ratio tests. Results In the conditional logistic regression, females had lower risk of developing schizophrenia (Table 1 ). Having siblings who were less than five years older was associated with increased risk of schizophrenia, while having siblings who were more than ten years older was associated with a lower risk of schizophrenia. The firstborn had a higher risk of schizophrenia than later-borns. Young maternal age was associated with increased risk of schizophrenia, while young paternal age decreased risk. Table 1 The effect of unique family-related variables on the odds of developing schizophrenia: results from the multivariate conditional logistic regression model which used the matched case-sibling design, odds ratios with 95% confidence intervals. Sex (female vs. male) 0.507(0.478, 0.537) Siblings who were less than 5 years older (yes vs. no) 1.463(1.288, 1.661) Siblings who were 5–10 years older (yes vs. no) 0.905(0.800, 1.025) Over 10 years older siblings (yes vs. no) 0.661(0.556, 0.785) Birth order 1 (reference) 2 0.618(0.535, 0.715) 3 0.542(0.441, 0.666) 4+ 0.504(0.382, 0.667) Birth cohort 1950–1955 (reference) 1956–1960 1.143(1.032, 1.267) 1961–1965 1.199(1.015, 1.415) 1966–1976 0.956(0.746, 1.224) Maternal age under 20 (reference) 20–25 0.794(0.671, 0.939) 26–30 0.688(0.555, 0.852) 31–35 0.614(0.470, 0.803) over 35 0.605(0.433, 0.845) Paternal age under 20 (reference) 20–25 1.382(0.990, 1.927) 26–30 1.432(1.004, 2.043) 31–35 1.340(0.915, 1.963) over 35 1.332(0.880, 2.015) not available 0.213(0.018, 2.491) In the ordinary logistic regression, siblings had higher risk if the proband had early, under 20 years, age at onset (Table 2 ). A sibship size of four or more was associated with increased risk of schizophrenia, as was having a mother with schizophrenia. The effect of parental age and having siblings who were less than 5 years older in the sibship disappeared in the ordinary logistic regression, but the effect of being the first born and having siblings who were more than 10 years older remained the same. Table 2 The effect of shared family-related variables on siblings' odds of developing schizophrenia: results from the multivariate logistic regression where the proband was removed from the analysis, odds ratios with 95% confidence intervals. Unique family-related variables (sex, ages of other siblings, birth order, birth cohort, maternal age, and paternal age) were also included in models as background variables. Sex of the youngest proband (female vs. male) 0.983(0.864, 1.117) Lowest age-at-onset in the sibship under 20 (reference) 20–25 0.574(0.497, 0.664) 26–30 0.323(0.267, 0.391) over 30 0.156(0.116, 0.209) Family size 2 (reference) 3 1.128(0.943, 1.349) 4 1.282(1.043, 1.575) 5+ 1.286(1.019, 1.623) Mother schizophrenic(no vs. yes) 1.822(1.431, 2.321) None of the tested interactions were significant. We investigated the effect of proband's sex and age-at-onset, birth order, age difference to siblings, birth cohort, and parental ages on the risk of developing schizophrenia. Proband's early age-at-onset and female sex increased the risk of schizophrenia among siblings. Male sex, having siblings who were less than 5 years older, and being the firstborn also increased the risk of schizophrenia, while having over 10 years older siblings and small family size decreased it. The higher risk of schizophrenia among the firstborns confirms the findings from some other recent studies. The most likely explanation is the increased risk of pregnancy and delivery complications among primiparous (first delivery) women. In the 1966 and 1986 Northern Finland birth cohorts, primiparity was a risk factor for having a low birth weight infant and for perinatal mortality [ 24 ]. In an earlier Finnish cohort study, infants of multiparous women were significantly heavier than infants of primiparous women [ 25 ]. As low birth weight and some obstetric complications increase the risk of later development for schizophrenia [ 26 ], they may explain the increased risk of schizophrenia among the firstborns. Having siblings who were less than five years older was a risk factor for schizophrenia in our study. Sham et al. found that having siblings who were 3–4 years older was a risk factor for schizophrenia [ 10 ], while Westergaard et al. found that a short interval to the nearest older sibling was a risk factor for schizophrenia [ 12 ]. Both suggested that common infections brought to the family by young siblings, causing fetal or early childhood infections, could lie behind the association [ 10 , 27 ]. Consistently with this, Brown et al found in a cohort study that second trimester exposure to respiratory infections increased the risk of later development of schizophrenia spectrum disorders two-fold [ 28 ]. Discussion Our results concerning the effect of paternal and maternal age are not comparable to previous studies [ 16 - 18 ], because all patients who did not have siblings were excluded from the analysis. Our findings suggest that in families with at least two children, young maternal age at birth is a risk factor for schizophrenia. This could be linked to obstetric complications. However, although some studies have found that obstetric complications are more common among adolescent mothers [ 29 ], this is not the case in Finland [ 24 , 30 ]. Thus, obstetric complications probably do not explain why the offspring of young mothers have an increased risk of schizophrenia. Consistently with a recent meta-analysis [ 31 ], the risk of schizophrenia was 50 percent lower for females than males. If risk factors for schizophrenia are equally distributed among females and males but females are for some reason more resilient to them, it could be assumed that females who develop schizophrenia have more schizophrenia-predisposing genes than males. This would explain why siblings' risk of schizophrenia was higher when the proband was female. Our results showed that the higher was the age-at-onset of the first proband in the family, the lower was the siblings' risk (Tables 1 and 2 ), which is in line with previous family studies [ 5 , 32 , 33 ] and also with studies that found an earlier age-at-onset among patients with familiar rather than sporadic schizophrenia [ 34 ]. Also, a study of parents of patients with childhood- and adult-onset schizophrenia found that parents of patients with childhood-onset schizophrenia had a significantly higher morbid risk of schizophrenia spectrum disorders than parents with adult-onset schizophrenia [ 35 ]. All these findings are consistent with the hypothesis that in schizophrenia, as in many other multifactorial neuropsychiatric diseases [ 36 ], early age at onset is a marker of greater genetic vulnerability to the disorder. A case-sibling design has some limitations. It can only provide approximate estimates of relative risk. In our analyses only families with two or more siblings were informative, and therefore our results are not applicable to one child families. However, whether matching was taken into account or not did not largely affect the results. Our study was based on a nationwide, representative sample of patients with schizophrenia and their siblings, and our sample size was considerably larger than in any of the previous studies. Register-based information is not as accurate as information based on face-to-face interview, but Finnish registers are quite reliable. The two registers we used, the National Population Register and the National Hospital Discharge Register, are nationwide. The National Hospital Discharge Register covers all public and private hospitals in Finland, and the National Population Register registers all Finnish citizens. However, as information on parents enabling the identification of sibships first appeared in the Population Register only in 1974, sibship data may be more accurate in the younger cohorts. Information in both registers is accurate. Some proportion of siblings of the study population are really half-siblings, but even in this case, the case-sibling design is applicable, because it does not assume that familial aggregation is due to a genetic mechanism. When entries in the National Hospital Discharge Register were compared with data on hospital case notes, the primary diagnosis in the register and hospital case notes was identical in 99 percent of cases for schizophrenia and in 98 percent of cases for all mental disorders [ 19 ]. The reliability of register diagnoses of schizophrenia is also good, with a greater propensity for false-negatives than false-positives [ 20 - 22 , 37 ]. Conclusions We investigated concurrently several family-related factors which independently have been suggested to be risk factors for schizophrenia. We were able to verify some of the previous associations, while others were refuted. The risk factors we identified probably have different underlying causative mechanisms. For example, the risk associated with being the firstborn might be explained by increased risk to obstetric complications, while the combination of decreased risk among females but increased risk among siblings of female probands might be explained by genetic factors. The risk factors we identified should be studied further in samples that include more detailed information on possible causative factors, such as infections and obstetric complications. Competing interests The author(s) declare that they have no competing interests. Authors contributions JH participated in the design of the study, performed the data-analysis and drafted the manuscript. JS participated in the design of the study, and the drafting of the manuscript. JL participated in the design of the study and its coordination. All authors read and approved the final manuscript. Pre-publication history The pre-publication history for this paper can be accessed here:	/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC535802.xml
555847	Factor structure of the Hospital Anxiety and Depression Scale (HADS) in German coronary heart disease patients	Background Depression and anxiety in patients with coronary heart disease (CHD) are associated with a poorer prognosis. Therefore the screening for psychological distress is strongly recommended in cardiac rehabilitation. The Hospital Anxiety and Depression Scale (HADS) is a widely used screening tool that has demonstrated good sensitivity and specificity for mental disorders. Methods We assessed mental distress in in-patient cardiac rehabilitation in Germany. The factor structure of the German language version of the HADS was investigated in 1320 patients with CHD. Exploratory factor analysis and confirmatory factor analysis were used to determine the underlying factor structure of the instrument. Results Three-factor models were found to offer a superior fit to the data compared to two-factor (anxiety and depression) models. The German language HADS performs similarly to the English language version of the instrument in CHD patients. The German language HADS fundamentally comprises a tri-dimensional underlying factor structure (labelled by Friedman et al. as psychomotor agitation, psychic anxiety and depression). Conclusion Despite of clinical usefulness in screening for mental disturbances the construct validity of the HADS is not clear. The resulting scores of the tri-dimensional model can be interpreted as psychomotor agitation, psychic anxiety, and depression in individual patient data or clinical investigations.	Background Coronary heart disease (CHD) is of profound interest to health and clinical psychology due to the high levels of anxiety and depression observed in patients following the occurrence of a coronary event [ 1 - 5 ]. CHD was the most leading diagnosis for treatment in hospitals in Germany for men (320,000 patients per year), and, after childbirth and breast cancer, the third reason for in-patient treatment for women (150,000 patients per year) [ 6 ]. Most in-patient rehabilitation hospitalizations in Germany for men (about 60,000 per year) were caused by CHD [ 7 , 8 ]. Recent research showed, that at least one in five patients in cardiac rehabilitation suffer from a psychological disorder [ 9 ]. Accurate identification of significant anxiety and depression as soon as possible following a cardiac event is essential in order to facilitate delivery of an effective and comprehensive treatment package which takes into account psychological as well as coronary disease symptoms [ 10 ]. This is particularly relevant since anxiety and especially depression have been demonstrated to be predictors of mortality in this clinical group [ 11 ]. The availability of easy to administer, reliable and valid screening tools would logically be a critical component of a clinical protocol seeking to identify CHD patients with psychological disturbance. A suitable measure would readily identify those patients for whom additional referral to a clinical psychologist or to a liaison psychiatry service would be more appropriate. A candidate screening tool that has been widely and increasingly used with CHD patients is the Hospital Anxiety and Depression Scale (HADS: [ 12 ]), an easily administered 14-item self-report measure comprising 7 anxiety items and 7 depression items from which separate anxiety and depression sub-scale scores are calculated [ 13 ]. The HADS was designed to exclude symptoms that might arise from the somatic aspects of illness such as insomnia, anergia, and fatigue, therefore the instrument has been designed for use within the clinical context of general medicine. The HADS has been used for screening purposes in a diverse and broad range of clinical groups [ 14 - 24 ]. A number of investigations have suggested that the HADS is a suitable instrument to accurately assess anxiety and depression in CHD patients [ 10 , 17 , 24 - 27 ]. A fundamental assumption underpinning the clinical usefulness of the HADS across a broad range of clinical groups, including CHD, is that the instrument reliably assesses anxiety and depression as two distinct and separable dimensions [ 28 ]. On the other hand, recent psychometric evaluations of the HADS in a range of clinical populations have suggested that the proposed factor structure of the instrument may indeed be compromised by the physiological aspects of the disease or by changes in health status [ 23 , 29 , 30 ]. Conversely, there is accumulating evidence that the fundamental factor structure of the HADS comprises three factors instead of two [ 24 , 26 , 31 - 33 ]. The finding that the three-factor structure offers a superior fit to clinical data than the two-factor (anxiety and depression) model formulated as part of the original instrument development by Zigmond and Snaith [ 12 ] has implications in terms of the use, scoring and future development of this assessment tool. Dunbar et al. [ 32 ] found a three-factor structure of the HADS in a non-clinical population (for an overview see table 1 ) and interpreted their findings in relation to the conceptually rich 'tripartite' model proposed by Clark and Watson [ 34 ]. Extending these observations to a clinical population, Friedman et al. [ 33 ] found a three-factor structure to the HADS in a patient group being treated for major depression, which incidentally, was similar to that observed by Dunbar et al. [ 32 ]. Martin and Newell [ 35 ] found that Friedman et al.'s [ 33 ] three-factor model offered the best-fit to their data examining individuals with significant facial disfigurement compared to competing two-factor models. Martin and Thompson [ 22 ] observed a three-factor structure to the HADS in myocardial infarction patients and, in a later study, Martin et al. [ 26 ] extended further the findings of both Dunbar et al. [ 32 ] and Friedman et al. [ 33 ] to myocardial infarction patients finding additional support for the three-factor structure suggested by these researchers to underlie the HADS. A recent study [ 24 ] of the psychometric properties of the HADS in Chinese acute coronary syndrome (ACS) patients has established further support for the three-factor structure of the HADS furnishing evidence that the three-dimensional structure of the instrument appears to be consistent across diverse cultures. Caci et al. [ 31 ] suggested a three-factor underlying structure to the HADS that represents a modification to the three-factor model identified by Friedman et al. [ 33 ] and replicated by Martin et al. [ 26 ]. However, Caci et al.'s [ 31 ] model was based on a student population and it should be remembered that the presence of significant pathology or physiological change states does impact on the underlying factor structure of this instrument [ 23 , 30 , 36 ]. The most consistent contemporary observation in terms of the underlying factor structure of this instrument in cardiac patients is strongly indicative that the HADS comprises three underlying dimensions. However, these cardiac studies have used either clinical populations from the United Kingdom [ 22 , 26 ] or from the Far East [ 24 ]. Table 1 Characteristics of each factor model tested in earlier studies. Model, author, year Number of factors Clinical population n Factor extraction method # Allocation of items to scales Zigmond & Snaith (1983) 2 Medical 100 No Factor analysis conducted anxiety: 1, 3, 5, 7, 9, 11, 13 depression: 2, 4, 6, 8, 10, 12, 14 Moorey et al. (1991) 2 Cancer 568 PCA anxiety: 1, 3, 5, 9, 11, 13 depression: 2, 4, 6, 7, 8, 10, 12, 14 Dunbar et al. (2000) 3 Non-clinical 2,547 CFA autonomic anxiety: 3, 9, 13 negative affectivity: 1, 5, 7, 11 anhedonic depression: 2, 4, 6, 7, Friedman et al. (2001)* 3 Depressed 2,669 PCA psychomotor agitation: 1, 7, 11 psychic anxiety: 3, 5, 9, 13 depression: 2, 4, 6, 8, 10, 12 Razavi et al. (1990) 1 Cancer 210 PCA All items included Caci et al. (2003) 3 Non-clinical 195 CFA anxiety: 1, 3, 5, 9, 13 restlessness: 7, 11, 14 depression: 2, 4, 6, 8, 10, 12 The three-factors are correlated in this model. # PCA: Principal Components Analysis; CFA: Confirmatory Factor Analysis. There has been little systematic investigation of the factor structure of the translated HADS in German cardiac patients, though interestingly a German language version of the instrument has been developed in Germany using cardiac patients within the context of the original assumed two-dimensional (anxiety and depression) structure [ 37 ]. To date, the HADS has been found to comprise a two-factor structure consistent with the anxiety and depression sub-scales proposed by Zigmond and Snaith [ 12 ] in cardiac [ 38 ] and non-clinical [ 39 ] populations in Germany. However, these large studies did not investigate the possibility that alternative factor models may provide a better explanation of the data. Identification of a coherent three-factor underlying structure of the HADS has a number of significant implications in terms of the validity of the tool as a screening instrument. Firstly, referral to mental health services could be undermined based on a two-dimensional (anxiety and depression) interpretation of HADS scores in cardiac patients. Secondly, further replication of a three-factor structure of the HADS in a German cardiac population would be valuable in determining if the HADS should be more effectively used as a screening instrument when comprised of three sub-scales in this group. Thirdly, replication of a consistent three-factor structure in the German-translated version of the HADS would provide strong evidence that the three-dimensional structure is implicit to the instrument and not a language-based artifact. Finally, the widespread international use of the HADS provides a compelling rationale to establish the psychometric properties of the instrument not only in broad diagnostic categories, but also across culturally-diverse groups. The purpose of the present study was to determine whether the three-factor structure of the HADS identified by Martin and colleagues [ 26 ] in myocardial infarction patients in the UK and Martin et al. [ 24 ] in Chinese ACS patients has the same psychometric properties as that of the German-translated version of the HADS in a cohort of German patients presenting with CHD. The present study addresses two research questions: 1) Do exploratory factor analysis (EFA) and confirmatory factor analysis (CFA) techniques concord to the proposed bi-dimensional structure of the HADS in German CHD patients? 2) Does a three-dimensional factor structure provide a superior fit compared to competing bi-dimensional factor structures? Methods Design The study used a cross-sectional design with all measures taken at one observation. The dependent variables were sum scores obtained on the HADS (all items), and the anxiety (HADS-A) and depression (HADS-D) sub-scales. Exploratory factor analysis (EFA) and confirmatory factor analysis (CFA) methods were used to address the research questions using a pooled HADS data set from all patients. Ethical approval for the study was given by the local ethical committee of the University Hospital of Freiburg. Written informed consent was obtained from all participants prior to the commencement of the study. Procedure The study was conducted in three German cardiac rehabilitation hospitals. The patients stay in these hospitals for three to four weeks for a comprehensive cardiac rehabilitation program consisting of medical advice, exercise, patient education, relaxation and psychosocial interventions. All cardiac patients who agreed to take part in the PROTeCD-study (Psychotherapeutic Resource-Orientated Treatment for Cardiac Patients with Depression) were screened for mental distress with the German version of the HADS [ 37 ] at admission to hospital. Sociodemographic data were collected by self report and somatic data were reported by physicians at study entry. Statistical analysis Exploratory factor analysis Exploratory factor analysis was performed on the full 14-item HADS using SPSS 12.0 statistical software. The criterion chosen to determine that an extracted factor accounted for a reasonably large proportion of the total variance was based on an eigenvalue greater than 1. A maximum likelihood factor extraction procedure was chosen since this method of factor condensation is consistent with our previous research [ 22 ] and is particularly useful for extracting psychologically meaningful factors [ 40 ]. An oblimin non-orthogonal factor rotation procedure was chosen [ 40 ] due to the possibility that extracted factors are likely to be correlated. Determination of a significant item-factor loading was set at a coefficient level of 0.30 or greater, this level based on a rationale of generating a more complete psychological interpretation of the data set, this being a level consistent with investigators who have used EFA [ 22 , 30 , 36 , 41 ]. Confirmatory factor analysis Confirmatory factor analysis was conducted using the Analysis of Moment Structures (AMOS) version 4 [ 42 ] statistical software package. Eight models were tested. These were Zigmond and Snaith's [ 12 ] original two-factor model, Moorey et al.'s [ 43 ] two-factor model, Razavi et al.'s [ 44 ] single-factor model, two versions of Clark and Watson's [ 34 ] three-factor model as evaluated by Dunbar and colleagues [ 32 ], Friedman et al.'s [ 33 ] three-factor model and two versions of Caci et al.'s [ 31 ] three-factor model. The characteristics and the allocation of the items to the factors in each tested modelare shown in Table 1 . For all models, independence of error terms was specified. Factors were allowed to be correlated where this was consistent with the particular factor model being tested. Multiple goodness of fit tests [ 45 ] were used to evaluate the eight models, these being the comparative fit index (CFI; [ 46 ]), the Akaike information criterion (AIC; [ 47 ]), the consistent Akaike information criterion (CAIC; [ 48 ]), the normed fit index (NFI; [ 45 ]), the goodness of fit index (GFI; [ 49 ]) and the root mean squared error of approximation (RMSEA). A CFI greater than 0.90 indicates a good fit to the data [ 50 ]. A NFI and GFI greater than 0.90 indicates a good fit to the data [ 51 ]. A RMSEA with values of less than 0.08 indicates a good fit to the data [ 52 ], while values greater than 0.10 suggest strongly that the model fit is unsatisfactory. The AIC and CAIC are useful fit indices for allowing comparison between models [ 32 ]. The Chi-square goodness of fit test was also used to allow models to be compared and to determine the acceptability of model fit. A statistically significant χ 2 indicates a significant proportion of variance remains unexplained by the model [ 45 ]. Results Participants 1320 patients (1035 male) enrolled in an in-patient cardiac rehabilitation programme in three hospitals in Germany provided complete HADS data sets for analysis. Inclusion criteria for participation in the study was a confirmed diagnosis of CHD; for details see [ 53 ]. The patient group comprised patients with a diagnosis of myocardial infarction (N = 666), coronary artery bypass graft (N = 382), percutaneous transluminal coronary angioplasty (N = 303) and unstable angina pectoris (N = 40). It is noted that diagnostic N exceeds total cohort N because many patients will have multiple CHD diagnoses. Patients were required to have had a diagnosis of CHD and had a recent cardiac event (MI, CABG, PTCA) in the past weeks. Female patients (mean age = 62.88; SD = 12.14) were significantly older, ( t (401.14) = 4.14, p < 0.001) than male patients (mean age = 59.58; SD = 10.59). Descriptive findings The mean HADS-A sub-scale score was 6.14 (SD = 4.15, range 0–20) and the mean HADS-D sub-scale score was 5.41 (SD = 4.00, range 0–20). Using Snaith and Zigmond's [ 28 ] cut-off criteria of HADS-A and HADS-D scores of eight or over, 467 participants (35%) demonstrated possible clinically relevant levels of anxiety and 373 participants (28%) possible clinically relevant levels of depression. Adopting Snaith and Zigmond's [ 28 ] higher threshold for sensitivity of HADS-A and HADS-D scores of eleven or over, 204 participants (15%) demonstrated probable clinically relevant levels of anxiety and 161 participants (12%) probable clinically relevant levels of depression. Exploratory factor analysis The Kaiser-Meyer-Olkin (KMO) measure of sampling adequacy and the Bartlett Test of Sphericity (BTS) were conducted on the data prior to factor extraction to ensure that the characteristics of the data set were suitable for the EFA to be conducted. KMO analysis yielded an index of 0.94, and BTS (χ 2 = 7758.34, df = 91, p < 0.001) was highly significant indicating the data satisfied the psychometric criteria for the factor analysis to be performed based on data distribution characteristics. Examination of individual item skew and kurtosis characteristics confirmed the suitability of the maximum likelihood factor extraction procedure [ 54 ]. Following extraction and oblimin rotation, two factors with eigenvalues greater than 1 emerged from analysis of the complete HADS and accumulatively accounted for 53% of the total variance. Factor loadings of individual HADS items in relation to the two-factor solution are shown in Table 2 . Factor scores were calculated for each participant using regression and revealed the two extracted factors to be highly statistically and positively correlated, r = 0.82, p < 0.001, explaining 67% of the common variance between factors. Table 2 Factor loadings of HADS items following maximum likelihood factor extraction with oblimin rotation HAD Scale item Factor 1 Factor 2 Anxiety sub-scale (1) I feel tense or wound up (AGI) 0.23 0.45 (3) I get a sort of frightened feeling as if something awful is about to happen (ANX) -0.02 0.75 (5) Worrying thoughts go through my mind (ANX) 0.09 0.69 (7) I can sit at ease and feel relaxed (AGI) 0.35 0.33 (9) I get a sort of frightened feeling like 'butterflies' in the stomach (ANX) -0.02 0.72 (11) I feel restless as if I have to be on the move (AGI) -0.01 0.43 (13) I get sudden feelings of panic (ANX) -0.02 0.75 Depression sub-scale (2) I still enjoy the things I used to enjoy 0.95 -0.14 (4) I can laugh and see the funny side of things 0.87 -0.06 (6) I feel cheerful 0.72 0.01 (8) I feel as if I am slowed down 0.38 0.21 (10) I have lost interest in my appearance 0.40 0.08 (12) I look forward with enjoyment to things 0.65 0.09 (14) I can enjoy a good book or TV programme 0.42 0.19 Note: Bold indicates that item loading on a factor is 0.30 or above. Factors in final model in figure 1; AGI: psychomotor agitation; ANX: psychic anxiety; DEP: remains the same Confirmatory factor analysis The factor models tested and accompanying fit indices are shown in Table 3 . χ 2 goodness of fit analyses for all models were highly statistically significant ( p < 0.001) indicating that a proportion of the total variance was unexplained by each model. Examination of the fit indices revealed that the best fit to the data was offered by Friedman et al.'s [ 33 ] three-factor model (see figure 1 ). This model provided consistently the best fit across all but one model fit assessment criteria. It was also found that both models of Dunbar et al.'s [ 32 ] three-factor model evaluation of Clark and Watson's [ 34 ] 'tripartite' model provided a 'best fit' to the data on a number of the fit indices tested (CFI, NFI and GFI) as did model 1 (CFI, NFI, and GFI) and model 2 (CFI, AIC, NFI and GFI) of Caci et al.'s [ 31 ] three-factor model. The two-factor models of Zigmond and Snaith [ 12 ] and Moorey et al. [ 43 ] offered poorer fits to the data compared to all three-factor models evaluated, however against accepted model fit convention, these two-factor models still offered an acceptable fit to the data. The single-factor model of Razavi et al. [ 44 ] was observed to offer the poorest fit to the data across all model fit estimates. Table 3 Factor structure of the HADS determined by testing the fit of models derived from factor analysis. All χ 2 analyses were statistically significant at p < 0.01 (χ 2 degrees of freedom in parentheses). Model χ 2 RMSEA CFI CAIC AIC NFI GFI Zigmond and Snaith (1983) 481.47(76) 0.06 0.95 718.84 539.47 0.94 0.95 Moorey et al. (1991) 480.77(76) 0.06 0.95 718.15 538.77 0.94 0.95 Caci et al. (2003) model 1 391.55(74) 0.06 0.96 645.30 453.55 0.95 0.96 Caci et al. (2003) model 2 352.02(62) 0.06 0.96 589.40 410.02 0.95 0.96 Dunbar et al. (2000) model 1 396.56(73) 0.06 0.96 658.49 460.56 0.95 0.96 Dunbar et al. (2000) model 2 # 399.52(73) 0.06 0.96 661.45 463.52 0.95 0.96 Friedman et al. (2001) 361.41(74) 0.05 0.96 584.16 423.41 0.95 0.96 Razavi et al. (1990) 986.48(77) 0.09 0.88 1215.67 1042.48 0.87 0.88 Note: Bold indicates best model fit as a function of model fit index criterion. * Three-factor model excludes item 10. # Hierachical arrangement of factors. Figure 1 Standardised factor loadings and between-factor correlations of the Friedman et al. [33] model. Boxes represents HADS items labelled as shown as in table 2. Circles represents factors. One-way and two-way arrows indicate factor loadings and between-factor correlations, respectively. Discussion The findings of the current study offer a further important contribution to the evidence base regarding the underlying factor structure of the widely used HADS. It is worthy of note that high levels of HADS assessed anxiety and depression were observed in the study. This finding is consistent with investigators using this instrument in cardiac populations in other parts of the world [ 22 , 24 , 26 ] and verification of the need to screen for psychological disturbance in patients presenting with CHD. The findings from the factor analyses conducted on the HADS data are of pertinent methodological as well as clinical interest. EFA of the HADS revealed two factors, the loadings of individual items being consistent with the anxiety and depression sub-scale domains. However, it was also observed that the HADS-A item-7 'I can sit at ease and feel relaxed' was jointly loading on both anxiety and depression factors, this split-loading slightly in favour of the depression factor. A recent EFA of the HADS conducted with patients with significant facial disfigurement [ 35 ] also revealed item-7 to be split-loaded between anxiety and depression latent domains. Martin and Newell [ 35 ] suggest that in circumstances of split-loading such as those observed in the current study, a two-factor solution may offer the most parsimonious solution in EFA, but may not provide the best identification of factors in terms of model fit. Martin and Newell's [ 35 ] rationale for this is that EFA is not a model evaluation technique, therefore identification of factors based on arbitrary cut-points such as eigenvalues and scree plots is likely to produce a psychometrically reductionist account of sophisticated relationships between observed and latent variables. Martin and Newell [ 35 ] proposed that the lack of apriori model specification in EFA provides a convincing psychometrically plausible explanation of inconsistencies between EFA and CFA in extracted factors and interpretation of data. Indeed, Martin and Newell [ 35 ] found a similar finding to that of the current investigation, EFA support for a two-factor model and CFA support for the superiority of three-factor compared to two-factor models. Interestingly, Dagnan et al. [ 55 ] and Mykletun et al. [ 56 ] identified three-factor initial solutions within the HADS but chose to dismise the third-factor, without a sound psychometric rationale. It is likely that an expectation of a presumed two-factor model makes it difficult to reconcile an unexpected three-factor model emerging from the data, therefore it is explainable why these researchers might choose to dismiss a third factor. The findings from the CFA revealed the best model fit to be provided by Friedman et al.'s [ 33 ] three-factor mode (see figure 1 ). The 'next best' fit to the data is offered by Caci et al.'s [ 31 ] three-factor model 2. It was also observed that the remaining three-factor models tested [ 31 , 32 ] not only offered a good fit to the data but also provided a superior fit to the data compared to the two-factor models evaluated on a number of estimates of model fit. The two-factor models of Zigmond and Snaith [ 12 ] and Moorey et al. [ 43 ] did however offer an acceptable fit to the data. The uni-dimensional model of Razavi et al. [ 44 ] was found to offer a poor fit to the data, a finding consistent with previous research on the HADS across a variety of clinical groups [ 19 , 24 , 26 , 30 , 35 ]. There remains little doubt from the CFA analysis that the best fit to the data is offered by three-factor models irrespective of the clinical population from which the three-factor model was derived. The findings from the CFA have furnished compelling support of the HADS as a tri-dimensional instrument, consistent with contemporary research with this instrument across diverse clinical presentations [ 19 , 26 , 30 , 31 , 33 , 35 ]. Conclusion In conclusion, the current study found the German language version of the HADS to have an underlying three-dimensional factor structure following CFA in CHD patients, an observation consistent with UK [ 26 ] and Chinese [ 24 ] CHD populations. The traditional interpretation of the HADS as a two-factor (anxiety and depression) structure was also found to offer an acceptable fit to data, though inferior to that of the three-factor models. It can be concluded that the HADS may serve as useful screening purpose by being scored as two sub-scales of anxiety and depression. The clinical utilisation of the HADS continues to be invaluable in screening for mental disorders. Our results suggest that the assessment of the efficacy of interventions in evaluation studies by the HADS may be biased by problems in construct validity. Two decades have passed since the HADS was introduced to the clinical screening battery. The findings of this study and those of others, suggests that despite the clinical usefulness in screening the individual results of the HADS could be interpreted more precise in clinical routine. The differentiation of the anxiety scale in "psychomotor agitation" and "psychic anxiety" in the best fitting model may be helpful in the interpretation of individual results of patients. These results may improve our understanding of the process of adaptation in patients with somatic illness. A separate analysis of subscales in clinical trials may reduce bias caused by somatic medical conditions of the patients. Agitation might be more likely biased by the medical status of the patients. Authors' contributions JB designed the study, carried out the data collection and clinical assessment. JB drafted part of the manuscript and was involved in the interpretation of the findings. CRM developed the statistical framework, carried out the statistical analysis and drafted part of the manuscript. Both authors have no competing financial or other interest in relation to this manuscript. Funding The study was funded by the Federal Ministry of Education and Research, Germany; Regional Pension Insurance Institute, Baden-Wuerttemberg, Germany, LVA 02 804 and is part of the Rehabilitation Research Network South-West (Germany).	/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC555847.xml
514576	Role of Tax protein in human T-cell leukemia virus type-I leukemogenicity	HTLV-1 is the etiological agent of adult T-cell leukemia (ATL), the neurological syndrome TSP/HAM and certain other clinical disorders. The viral Tax protein is considered to play a central role in the process leading to ATL. Tax modulates the expression of many viral and cellular genes through the CREB/ATF-, SRF- and NF-κB-associated pathways. In addition, Tax employs the CBP/p300 and p/CAF co-activators for implementing the full transcriptional activation competence of each of these pathways. Tax also affects the function of various other regulatory proteins by direct protein-protein interaction. Through these activities Tax sets the infected T-cells into continuous uncontrolled replication and destabilizes their genome by interfering with the function of telomerase and topoisomerase-I and by inhibiting DNA repair. Furthermore, Tax prevents cell cycle arrest and apoptosis that would otherwise be induced by the unrepaired DNA damage and enables, thereby, accumulation of mutations that can contribute to the leukemogenic process. Together, these capacities render Tax highly oncogenic as reflected by its ability to transform rodent fibroblasts and primary human T-cells and to induce tumors in transgenic mice. In this article we discuss these effects of Tax and their apparent contribution to the HTLV-1 associated leukemogenic process. Notably, however, shortly after infection the virus enters into a latent state, in which viral gene expression is low in most of the HTLV-1 carriers' infected T-cells and so is the level of Tax protein, although rare infected cells may still display high viral RNA. This low Tax level is evidently insufficient for exerting its multiple oncogenic effects. Therefore, we propose that the latent virus must be activated, at least temporarily, in order to elevate Tax to its effective level and that during this transient activation state the infected cells may acquire some oncogenic mutations which can enable them to further progress towards ATL even if the activated virus is re-suppressed after a while. We conclude this review by outlining an hypothetical flow of events from the initial virus infection up to the ultimate ATL development and comment on the risk factors leading to ATL development in some people and to TSP/HAM in others.	Introduction Human T-cell leukemia virus type-I (HTLV-1) is the first discovered human retroviral pathogen [ 1 ]. It has been firmly implicated with the etiology of an aggressive malignancy known as adult T-cell leukemia (ATL) and of a neurological progressive inflammatory syndrome called tropical spastic paraparesis or HTLV-1 associated myelopathy (TSP/HAM). In addition, there are indications that it might be also associated with certain other clinical disorders [ 2 , 3 ]. In culture HTLV-1 can infect a wide variety of cell types from different species. However, in natural human infections this virus targets mainly mature CD4 + helper T-cells [ 4 - 6 ], resulting in benign expansion the infected cells [ 7 ]. Clonal or oligoclonal expansion of the infected CD4 + cells is mostly associated with development of ATL and 90–96% of the HTLV-I DNA is, indeed, found to segregate with CD4 cells in the peripheral blood of ATL patients [ 4 ], whereas CD4/CD8 double-positive leukemic cells are detected in rare cases [ 8 ]. CD8 + T-cells might also be infected [ 9 , 10 ], but their expansion is rather polyclonal and frequently occurs in asymptomatic carriers. Therefore, their disease association is unclear yet [ 11 ]. Shortly after infection the virus enters into a latent state, rendering the infected individuals asymptomatic seropositive carriers. About 5% of these individuals develop one of the viral associated diseases 10 to 40 years after infection. During latency the viral gene expression in the peripheral blood lymphocytes (PBLs) of such carriers is very low. Viral RNA is undetectable by Northern blot analysis in most of the infected cells (i.e. viral DNA harboring cells) freshly isolated from their peripheral blood [ 5 ], although it can be detected in some carriers by the highly sensitive RT/PCR analysis [ 12 ]. Furthermore, very little or no viral proteins are detectable in the carriers' PBLs [ 12 , 13 ]. Notably, despite this low virus expression, healthy carriers contain antibodies against viral antigens. They also display anti HTLV-1 specific cytotoxic T-lymphocytes (CTL) activity at variable levels that seem to be determined by hosts' genetic determinants, particularly by those associated with their HLA antigens [ 3 , 14 , 15 ]. Experimental evidence has been reported, pointing to the critical role of these two anti HTLV-1 immune response arms in keeping this low viral expression. It has been repeatedly shown that PBLs isolated from such carries start eliciting high viral gene expression within few hours of growing in culture [ 10 , 13 , 16 ]. However, Tochikura et al. have noted that addition of sera from HTLV-1 carriers or patients to the culture medium reduces this viral expression at an efficiency which correlates to their titer of anti HTLV-1 antibodies and that removal of these antibodies by protein A abolishes this inhibition. No such inhibition has been observed with sera of uninfected control donors [ 13 ]. Other workers have analyzed the level of HTLV-1 expression in PBLs grown in whole blood samples of various infected individuals and found that depletion of CTLs from these samples remarkably increases in the number of virus-expressing CD4+ cells compared to that found in the same samples without CTL-depletion [ 10 , 16 ]. Furthermore, these authors have demonstrated a similar increase by blocking the CTL-mediated cytolytic activity with concanamycin A. These data strongly suggest that anti HTLV-1 CTL activity, mounting in infected individuals, eliminate cells with high level of viral antigens and keep, thereby, the overall virus expression in the carriers' PBLs at low level. In view of this low virus expression, the viral load in HTLV-1 infected individuals has been noted to expand primarily through proliferation of the proviral DNA-harboring cells rather than through repeated cycles of cell-to-cell infection of new uninfected cells [ 17 ]. As discussed in our recent review article [ 18 ], this expansion pattern is widely considered to account for the maintenance of high sequence stability of the viral genome throughout the hundreds of thousands years of evolution since its emergence from its simian T-lymphotropic retrovirus origin. This stability is in striking contrast to the high genetic diversity of HIV-1 which is known to spread within the infected individuals through repeated infections of new cells by cell-free virions [ 19 ]. Although the mechanism of HTLV-1 pathogenicity is not fully understood yet, it is widely believed that a virally encoded transactivator protein, called Tax, plays a central role in this mechanism. It should, therefore, be noted that while the low level of the virus gene expression detected in latently infected carriers might be sufficient for maintaining their anti HTLV-1 seropositivity and CTL activity, the low Tax level, resulting from this reduced viral expression is, most likely, below its pathogenic threshold. This implies that generating an HTLV-1 related disease requires an activation of the dormant virus in order to elevate Tax to its pathogenic level. In this article we present a comprehensive review of the wide range of Tax molecular interactions and biological effects that might be closely relevant to the mechanism of ATL genesis and summarize this information by proposing hypothetical flow of a stepwise pathway leading to this malignancy or to TSP/HAM. HTLV-1 genomic structure and gene expression HTLV-1 is a complex retrovirus that, in addition to the two long terminal repeats (LTRs) and the gag, protease, pol and env genes, which are typical to most other retroviruses, its genome contains an additional region called pX, which resides between the env gene and the 3'-LTR,. This region includes four partially overlapping reading frames (ORFs), of which the most investigated ones are ORFs III and IV that encode for the viral regulatory Rex and Tax proteins respectively (see illustration in Fig. 1A ). The gag, protease and pol precursor polypeptide is translated from the full genomic length viral RNA, whereas the env precursor polypeptide is translated from a singly spliced viral RNA. These precursor polypeptides are cleaved into the mature functional proteins by the viral protease. Tax and Rex are translated from a doubly spliced viral RNA, using two alternative translational initiation codons as illustrated in Fig. 1B . Figure 1 Schematic illustration of the HTLV-I genome organization (A) and its various mRNA species with their specific splicing and encoded protein products (B) (See the text for detailed explanation). Tax is present predominantly in the nucleus due to its nuclear localization signal (NLS) residing at its amino terminus [ 20 , 21 ]. However a substantial portion of Tax is present also in the cytoplasm due to its newly identified nuclear export signal (NES) [ 22 ]. Tax, which acts as a dimer [ 23 ], was originally discovered as a transactivator of viral RNA transcription from a promoter located at the 5'-LTR [ 24 ], but later proved to modulate the synthesis or function of a wide range of cellular regulatory proteins [ 25 - 27 ]. Rex, on the other hand, acts to promote the export of the unspliced and singly spliced viral RNAs species from the nucleus to the cytoplasm [ 28 ] by binding to a Rex responsive element (RxRE) residing in the 3' R region of the viral RNA [ 29 ]. In addition, there are some indications that Rex may also inhibit splicing and degradation of the viral RNAs [ 30 ]. Thus at high level of Rex there is a preferential export of the gag-protease-pol- and of the env-encoding RNA species and low export of the Tax/Rex-encoding RNA. This leads to a decline in the level of Rex and Tax proteins and consequently to a reduced viral RNA transcription. As a result, the Tax/Rex-encoding RNA is preferentially exported from the nucleus. In this manner Rex maintains these different RNA species at an optimal balance required for the virus production. Consistent with this notion Ye et al. [ 31 ] have shown that cells harboring proviral DNA with defective Rex reading frame produce high level of the doubly spliced tax/rex encoding mRNA and high level of functional Tax protein, but low level of p19 Gag protein and undetectable Rex protein. An alternatively spliced RNA encodes for another protein from ORF III, termed p21 Rex , but its biological function is unclear [ 32 ]. More recently interest has been focused also on ORF I that encodes for p12 and p27 and ORF II that encodes for p13 and p30 proteins [ 33 ]. In contrast to Tax and Rex, which are encoded by a bicistronic pX mRNA formed by double splicing of the viral RNA [ 21 , 34 , 35 ], the other four accessory proteins are encoded by different pX mRNAs formed by alternative splicing events [ 33 , 36 ]. Pique et al. [ 37 ] have detected CTL activity in HTLV-I infected individuals against specific peptides from each of these ORF I and ORF II proteins, indicating that each of them is produced natural human infections. The functional role of these accessory proteins is not completely clear yet. Certain studies have demonstrated that deletions within frame I and II do not affect the replication and infectivity of HTLV-1 [ 36 ] nor its capacity to immortalize primary T-cells [ 36 , 38 ]. In contrast, by using molecular HTLV-1 clone, the group of Albrecht and Lairmore has provided evidence for the critical role of these accessory proteins in the viral replication and pathogenesis [ 33 ]. It has been shown that ablation of frame I markedly reduced the virus ability to infect quiescent peripheral blood lymphocyte (PBLs) [ 39 ] and to replicate in a rabbit model [ 40 ]. The explanation suggested by these investigators for the discrepancy between theirs and the others' results regarding p12 is that the other groups examined the role of this protein in IL-2/mitogen-activated PBLs, whereas their own data indicate that p12 is required for HTLV-1 infection in quiescent PBLs, since when they added a mitogen and IL-2 to their cultures the p12-defective HTLV-1 clone became highly infective [ 33 , 41 ]. Notably, p12 localizes to the endoplasmic reticulum (ER) and is associated with two ER-resident proteins; calerticulin and calnexin. Calerticulin is a calcium-binding protein that participates in calcium signaling and linked to activation of the transcription factor nuclear factor of activated T-cells (NFAT) [ 42 ]. In this manner, p12 can activate the HTLV-1 DNA clone-harboring quiescent PBLs and provide the physiological requirements for its infectivity, or vice versa, mitogen/IL-2 activation of the PBLs can override the deficiency imposed by the p12-defective clone. Since HTLV-1 targets quiescent T-cells in natural infection, these findings suggest an important role of p12 protein for the virus in vivo infectivity. The frame II encoded p30 protein has been shown to localize to the nucleus and to function as a transcription factor. Transient transfection experiments have demonstrated that this protein can modulate the expression of various promoters and to activate HTLV-I LTR expression independently of Tax [ 43 ]. It was also shown to interact with the transcriptional co-activators CREB-binding protein (CBP) and p300 [ 44 ]. Together, these and other data indicate that p30 may account for the activation of several genes in HTLV-1 infected cells [ 44 ] and play an important role in the virus replication [ 45 ] and maintaining high viral load in in-vivo infection [ 33 , 39 , 46 ]. In contrast, a recent study by Nicot et al. [ 47 ] have shown that p30 rather inhibits HTLV-I expression by binding to the tax/rex-encoding doubly spliced viral RNA and retaining it in the nucleus. In this manner p30 prevents the synthesis of Tax and Rex proteins and interferes, thereby, with the production of viral particles. Furthermore, high level of p30 has been found to interfere with Tax-induced activation of HTLV-I LTR [ 44 ]. In view of these data it has been suggested that by reducing HTLV-I expression high level of p30 protects the infected cells from the anti HTLV-I immune response and contribute, in this manner to the virus persistence [ 33 ]. The other frame II-encoded protein, p13 localizes in the mitochondria and alters its morphology and function [ 48 ]. This protein has been shown to be also essential for maintaining high viral load in rabbit [ 45 , 46 ]. It has been also demonstrated that p13 interferes with the phosphorylation of the guanine nucleotide exchanger Vav protein in T-cells [ 49 ]. Fig. 1 describes schematically the viral genome organization, its various mRNA species and the encoded proteins. Since Tax protein is widely regarded as a key element in the HTLV-1 related leukemogenic process. We will discuss in the following sections the molecular activities and biological effects of Tax that seem to contribute to its oncogenic potential. Modulation of viral and cellular gene expression by Tax Tax-mediated activation of CREB/ATF-dependent gene expression As noted before, Tax was initially discovered as a transactivator of the HTLV-1 gene expression [ 24 ]. It activates the viral LTR through three imperfectly conserved 21 bp repeats called Tax responsive elements (TxRE) [ 50 ], which contain a centered sequence TGACG(T/A)(C/G)(T/A) that is imperfectly homologous to the consensus cAMP responsive element (CRE; TGACGTCA) [ 51 ]. This element, which is also referred to as domain B of the TxRE, is flanked by a short G-rich stretch (AGGC) at its 5' side, termed domain A and a C-rich stretch (CCCC) at its 3' side, termed C domain C [ 27 , 51 ] (Fig. 2A ). Although several basic leucine zipper (bZIP)-containing proteins, belonging to the CRE-binding/activating transcription factor (CREB/ATF) family, can bind to this viral CRE [ 52 ] only few of them can efficiently mediate the Tax-induced transactivation of HTLV-1 LTR [ 53 - 56 ]. A recent investigation of the effect of negative transdominant constructs against various bZIP proteins of this family has provided evidence that CREB is the most prominent factor that cooperates with Tax in activating HTLV-1 LTR expression [ 53 ]. Numerous earlier studies have demonstrated that in the absence of Tax, CREB forms unstable complex with the viral CRE, whereas Tax acts to stabilize this complex. By interacting with the bZIP region of CREB Tax enhances CREB dimerization and increases, thereby, its affinity to CRE [ 54 , 57 - 59 ]. This Tax-CREB-TxRE complex is further stabilized by direct binding of Tax to domains A and C of the TxRE through its N-terminus [ 60 , 61 ] (Fig. 2A ). This stabilized binding enables Tax to recruit to the ternary Tax-CREB-TxRE complex the co-activators CREB binding protein (CBP) and its homologous protein p300 by binding to their KIX domain through its kinase-inducible domain (KID) [ 62 ] and the p300/CBP-associated factor (P/CAF), which binds through it carboxy terminus to a distinct site located around amino acid 318 to 320 of the Tax protein [ 63 ]. These three co-activators exert their effect by histone acethylation, which induces chromatin conformational modification at the site of the target promoter and facilitates, thereby, the interaction of the enhancer-bound transcriptional activators with the TATAA box-associated basal transcriptional factors [ 27 ] (Fig. 2A ). Interestingly, however, Jiang et al. have shown that P/CAF can bind Tax without CBP or p300 and enhances its stimulatory effect on HTLV-1 LTR transcriptional expression independently of histone acetylation [ 63 ]. In contrast, several other studies have indicated that CREB2 (called also ATF-4), a member of another bZIP protein family, plays a more central role in Tax activation of HTLV-1 gene expression. These studies show that while in the absence of Tax, CREB can activate HTLV-1 LTR expression only if phosphorylated by protein kinase A (PKA), CREB2 can markedly activate the viral LTR without phosphorylation and that this protein mediates a much stronger activation of the viral LTR by Tax than CREB does [ 64 - 66 ]. Figure 2 Schematic illustration of the DNA elements and the activator and co-activator proteins involved in Tax-induced transcriptional activation of (A) HTLV-I LTR and (B) SRF-dependent promoters (See the text for detailed explanation). Of particular note are also the recent observations that when two copies of the TxRE are placed upstream to TATAA boxes from HTLV-1 LTR or from other promoters, the strongest activation by Tax is detected with the TATAA box of the HTLV-1 LTR, indicating that this TATAA box contains a specific Tax responsive element. Furthermore, these studies have also revealed that beside of the enhancing effect Tax on the association of the TATAA-box binding protein (TBP) to the TATAA site, Tax has an additional stimulatory effect that is directed towards a step occurring after the assembly of the basal transcriptional factors onto the TATAA box [ 53 ]. Many cellular genes contain in their promoters a consensus CRE element and are activated by signals that elevate the cellular cAMP level. The elevated cAMP activates PKA to phosphorylate CREB which, in turn, binds to CRE and to CBP/p300. However, there is a substantial controversy on whether Tax can activate only the viral CRE in its context with the CG-rich flanking domains in the viral LTR [ 25 , 61 ], or also CRE located in cellular promoters [ 67 , 68 ]. In addition, there are data demonstrating that Tax uses the CREB/ATF factors to repress the expression of certain genes, like the cyclin A [ 69 ], p53 [ 70 ] and c-myc [ 71 ]. This CRE-dependent effect of Tax on such cellular genes may contribute to the initiation of an oncogenic process by impairing the cell cycle and growth control. Tax mediated activation of SRF-dependent gene expression HTLV-1 infected and Tax-expressing T-cell lines display increased expression of immediate early genes such as c-Fos, c-Jun, JunB, JunD and Fra-1, which are components of the dimeric transcription factors AP1, Egr-1 and Egr-2 [ 72 ], fra-1 [ 73 ], Krox-20 and Krox-24 [ 74 ]. Formation of these transcription factors is normally activated by the serum responsive factor (SRF) in response to various mitogenic signaling agents like serum, lysophosphatidic acid (LPA), lipopolysaccharide (LPS), 12-O-tetradecanoylphorbol-13-acetate (TPA), cytokines and tumor necrosis factor-α (TNFα). SRF acts through an SRF responsive element (SRE) residing in the promoters of these genes [ 75 ]. The SRE region actually contains two binding sites; a CArG box [CC(A/T) 6 GG], and an upstream Ets box [GGA(A/T)]. After binding to the CArG box, SRF protein interacts with the ternary complex factors (TCFs), which consequently bind to the upstream Ets box. In addition, SRF requires for its transcriptional activity the CBP/p300 and p/CAF co-activators [ 76 ]. Tax activates these immediate early genes by interacting with SRF [ 77 , 78 ] and with TFCs, CBP/p300 and P/CAF [ 76 ] (Fig. 2B ). Moreover, AP-1, which is highly expressed in HTLV-1 infected T-cells [ 79 ], regulates the expression of multiple genes essential for cell proliferation, differentiation and prevention of apoptosis [ 80 ], so that by activating SRF, Tax can also indirectly induce a wide variety of such cellular genes. Thus, constitutive activation of such genes in HTLV-1 infected T-cells independently of specific external signals might be a trigger for initial steps in the oncogenic transformation of HTLV-1 infected T-cells in culture as well as in human infection. Tax-mediated activation of NF-κB-dependent gene expression A substantial part of Tax oncogenic potential is attributed to its ability to activate transcription factors of the NF-κB family, since these factors regulate the expression of numerous cellular genes [ 81 ] associated with diverse biological processes, such as embryonic development, immune and inflammatory responses, cell growth, apoptosis, stress responses and oncogenesis [ 25 , 82 - 84 ]. The NF-κB factors are functionally related to the c-Rel proto-oncogene and include the p50(NF-κB1), p52(NF-κB2), p65(RelA), RelB and c-Rel proteins, which act in various combinations of homo- and heterodimers displaying distinct specificities. They share a common domain of 300 amino acids, termed Rel homology domain (RHD), which is involved in their dimerization, DNA binding and nuclear localization. The p65:p65 and p65:p50 κB are the most prominent dimers involved in NF-κB-dependent transcriptional activation, whereas the p50:p50 dimer is rather inhibitory [ 85 ]. In non-activated state NF-κB factors are trapped in the cytoplasm, tightly associated with inhibitory proteins called IκBs, primarily with IκBα and IκBβ. These inhibitors contain ankyrin repeats through which they bind to the RHD of the NF-κB factors and mask their nuclear localization signal (NLS) [ 86 ]. In addition, these complexes contain the catalytic subunit of protein kinase A (PKAc) which binds in the cytoplasm to both IκBα and IκBβ and is held there in an inactive state [ 87 ] (see illustration in Fig. 3 No. 1). NF-κB factors are activated in response to a wide variety of inflammatory cytokines and mitogens, such as TNF-α, IL-1, IL-6, IL-8, GM-CSF, bacterial lipopolysaccharide (LPS) and stress-inducing factors [ 81 , 83 , 84 ] (see Fig. 3 , No. 2a and 3a). This activation proceeds in two phases, one taking place in the cytoplasm and the other in the nucleus. Figure 3 Schematic illustration of the factors and the molecular interactions associated with the release the NF-κB factors from their IκB inhibitors in the cytoplasm by external signaling stimuli and by HTLV-I Tax (See the text for detailed explanation). The cytoplasmic phase includes phosphorylation of IκBα on serine32 and serine36 and of IκB on serine19 and serine23 (Fig. 3 , No. 6), which is followed by their ubiquitination and subsequent proteosomal degradation [ 88 ] (Fig. 3 , No. 7). The release from IκBs, activates the associated PKAc, which phosphorylates the free p65(RelA) factor at its serine276 (Fig. 3 , No. 8). As will be discussed later in more details, this phosphorylation is essential for the transcriptional activity the p65(RelA)-containing dimers [ 87 ]. In addition, degradation of the IκBs releases the sequestered NF-κB dimers to translocate to the nucleus [ 88 ] (Fig. 3 , No. 9). The phosphorylation of IκBs is carried out by an IκB kinase (IKK) complex comprised of two catalytic subunits, IKK and IKK and a regulator subunit, IKK which is called also NF-κB essential modulator (NEMO) [ 89 , 90 ] (Fig. 3 , No. 2a and 3a). IKKα and IKK share a 52% amino acid identity and a similar domain structure that includes amino-terminal kinase domain, a dimerization leucine zipper domain, and helix-loop-helix motifs, which are involved in regulating their kinase activity [ 89 , 90 ]. The phosphorylating function of the IKK complex is activated by upstream kinases such as the NF-κB inducing kinase (NIK) (Fig. 3 , No. 2b), the mitogens-activated protein kinase/ERK kinase kinase-1 (MEKK1) (Fig. 3 , No. 3b) and certain other signal-activated kinases [ 91 ]. NIK phosphorylates mainly the IKKα subunit (Fig. 3 , No.2b), whereas MEKK1 activates both IKKα and IKKβ [ 92 ] (Fig. 3 , No. 3b). Activation of IKKα results from its phosphorylation at serine176 and serine180, whereas IKK is activated by its phosphorylation at serine177 and serine181 [ 93 , 94 ]. Despite their high homology, IKKβ is much more active than IKKα in phosphorylating the IκBs [ 93 , 95 , 96 ]. This predominant activity of IKKβ over IKKα may be partially explained by the observation that in addition to the phosphorylation of IKKβ by MEKK1, IKKβ is directly phosphorylated also by IKKα, [ 97 , 98 ] (Fig. 3 , No. 2b). A recent study has suggested an additional function for IKKα by showing that p65(RelA) needs to be phosphorylated by this kinase at serine536 in order to be transcriptionally active [ 99 ]. The third subunit, IKKγ/NEMO is devoid of kinase activity. Its role is to serve as a universal scaffold which connects between the two catalytic IKK subunits and their upstream activating factors into a large IKK complex [ 100 , 101 ] (Fig. 4 , No. 2b and 3b). Iha et al., [ 102 ] have shown that these various factors assemble to the IKK complex through different domains of the IKKγ/NEMO protein, which could be selectively inactivated, thus attenuating certain NF-κB activating signals without affecting others. Figure 4 Schematic illustration of the factors and molecular interactions occurring in the nucleus which are involved in regulating the transcriptional competence of the NF-κB factors after reaching the nucleus and the function of HTLV-I Tax in this regulation (See the text for detailed explanation). Recently, much interest has been attracted to the nuclear regulation of the NF-κB transcriptional competence. It has been shown that after reaching the nucleus p65(RelA) can bind the CBP/p300 and P/CAF coactivators which are essential for the transcriptional competence of p65(RelA):p65(RelA) and p65(RelA):p50 dimers [ 103 ]. This binding depends on p65(RelA) phosphorylation at serine276 by PKA and certain other signal activated serine kinases [ 85 , 87 , 104 - 108 ] (see illustration in Fig. 5 , No. 1a and 1b). This phosphorylation is blocked by an NF-κB-inducible protein termed SINK, which binds to p65(RelA). This binding does not affect the nuclear localization of p65(RelA), nor its binding to the target DNA sites. Instead, by inhibiting p65(RelA) phosphorylation SINK prevents its association with the CBP/p300 and P/CAF co-activators, thus creating a negative feedback control of p65(RelA) transcriptional activity [ 109 ]. Another inhibitor protein, called RelA-associated inhibitor (RAI), has been identified in the nucleus of certain cell types where it can interact with p65(RelA) and inhibit its transcriptional activity by blocking its DNA binding. It has been proposed that this protein provides an alternative cell-type specific control of NF-κB-dependent gene expression [ 110 ]. Figure 5 Schematic presentation of Tax biological effects which contribute to its oncogenic potential. In addition to its cytoplasmic inhibitory function IκBα plays an important regulatory role in the nucleus too. IκBα has an NLS signal which enables its translocation to the nucleus where it is protected from the signal-induced degradation described above [ 111 ]. Within the nucleus IκBα binds to the nuclear p65(RelA) and abrogates its transcriptional activity by inhibiting its DNA-binding [ 112 ]. IκBα has also a nuclear export signal (NES) which mediates the export of the p65(RelA):IκBα complex back to the cytoplasm via its interaction with the nuclear exporting protein CRM1 [ 113 ] (see Fig. 4 , No. 2a, 2b, 2c and 2d). It has been proposed that as long as the signal-induced cytoplasmic degradation of the NF-κB-associated IκBα is active, induction of corresponding NF-κB-dependent gene expression can keep going on, whereas upon termination of this signal the export of the p65(RelA):IκBα complex from the nucleus may serve as an immediate terminator of this gene expression. However, the nuclear association of IκBα with p65(RelA) has been noted to depend on p65(RelA) acetylation status. The nuclear p65(RelA) can be acetylated by p300 and this acetylation avoids the binding of p65(RelA) to IκBα, thus preserving its transcriptional activity [ 114 ]. On the other hand, the nuclear p65(RelA) can bind to specific isoforms of histone deacetylase (HDAC) which deacetylate it and inhibit, thereby, its transcriptional activity by facilitating its association to IκBα [ 115 ]. (see Fig. 4 . No. 2e). In contrast to this nuclear IκBα function, it has been noted that signals imposing persistent NF-κB activation, do so by enhancing the level of unphosphorylated IκBβ, which binds to p65(RelA) in the cytoplasm without masking its NLS or interfering with its DNA binding [ 116 ] (Fig. 4 , No 3a, 3b and 3c). It has been proposed that under such conditions IκBβ escorts p65(RelA) to the nucleus, where it protects it from the inhibitory effect of the nuclear IκB and maintains, in this manner, a persistent NF-κB transcriptional activation [ 116 ]. IKK has also been found to have an important role in the nucleus (Fig. 4 , No 4a) where it seems to affect the NF-κB transcriptional activity in several different ways. In one study the nuclear IKKα has been shown to bind CBP and p65(RelA) and to recruit, in this manner, the CBP co-activator to NF-κB-responsive promoters, where it acetylates histone H3 and facilitates, thereby, the expression of these promoters [ 117 ] (Fig. 4 , No. 4b). Another study has shown that the nuclear p65(RelA)-associated IKKα stimulates the NF-κB-responsive promoters by directly phosphorylating histone H3 with its kinase activity [ 118 ] (Fig. 4 , No 4c), and a third study has demonstrated that the nuclear IKKα phosphorylates the nuclear p65(RelA) and facilitates, thereby, its association with CBP/p300 [ 99 ] (Fig. 4 , No 4d). IKKγ/NEMO too has been noted to translocate to the nucleus where it regulates the NF-κB transcriptional activity by competing with the nuclear p65(RelA) and IKKα for CBP/p300 [ 119 ] (Fig. 4 , No. 5a and 5b correspondingly). In contrast to the transient NF-κB activation by external signals, NF-κB factors are constitutively activated by HTLV-1 Tax protein in Tax-expressing and HTLV-1-infected cells. Reported studies suggest that Tax may exert this activation in three ways: a) The most widely accepted concept is that Tax associates with the IKK complex through the adaptor IKKγ/NEMO subunit. Tax also binds to the upstream kinases, MEKK1 and NIK and enhances their kinase activity. In this manner Tax connects these activated kinases to IKKγ/NEMO and recruits their kinase activity to phosphorylate IKK and IKKβ [ 25 , 102 , 120 - 122 ] which, in turn, phoshphorylate IκBα and IκBβ (see Fig. 3 , No. 4a, 4b and 6). A recent study have proposed that IKKγ/NEMO assembles into the large IKK complex as a homodimer or homotrimer and that its binding to Tax enhances its oligomerization [ 123 ]. b) Tax can bind directly to IKK and IKK and activates their kinase activity independently of their phosphorylation by the upstream signal-induced kinases [ 124 ] (Fig. 3 , No. 5 and 6), c) Tax can bind directly to the IκBs and induce their proteosomal degradation independently of their phosphorylation by IKK [ 90 , 125 ] (Fig. 3 , No. 10a, 10b and 10c). Thus, Tax induces phosphorylation-dependent or independent degradation of both of the IκBs and enables, thereby, the nuclear translocation of the released NF-κB factors independently of exogenous signals (Fig. 3 , No. 9). A number of studies indicate that the nuclear Tax plays an important role in establishing the transcriptional activity of the NF-κB factors reaching to the nucleus. Bex et al. [ 126 ] have demonstrated that the nuclear Tax localizes in transcriptionally active structures containing the NF-κB factors p50 and p65(RelA), RNA polymerase II, nascent RNA and splicing factors. Other studies have shown that Tax physically binds to the NF-κB factors p65(RelA) [ 127 ], c-Rel [ 127 ], p50 [ 128 ] and p52 [ 129 ] and enhances, thereby, their dimerization [ 130 ], which is essential for their binding to the NF-κB responsive element in the target promoters [ 127 , 128 ]. Tax has noted also to associate with these factors when they are already bound to their DNA targets and facilitates their transcriptional activity [ 127 , 128 ]. In contrast, our own experiments, to be published elsewhere (manuscript in preparation), indicate that the binding of Tax to the free p65(RelA) factor occurs already in the cytoplasm and then the two proteins translocate to the nucleus together (see Fig. 3 , No. 11a and 11b). In addition, it has been demonstrated that by its ability to bind the NF-κB factors [ 127 - 129 ] on one hand and the CBP/p300 [ 62 , 131 ] and P/CAF [ 63 ] co-activators on the other hand, Tax recruits these co-activators to the NF-κB factors independently of the above mentioned serine276 phosphorylation on p65(RelA) (see illustration in Fig. 4 , No 6). However, a recent study has indicated that in order to be transcriptionally active, p65(RelA) needs to be phosphorylated by IKKα at serine536 even when activated by Tax [ 99 ]. This phosphorylation is mediated by Tax [ 99 ] through its capacity to physically bind IKK and IKKβ and induce their kinase activity [ 124 ]. Tax biological effects contributing to its oncogenic potential Enhancing T-cell proliferation There is ample of literature, reviewed in ref. [ 2 , 3 , 21 , 132 , 133 ], which demonstrate a modulation of expression of a wide range of cellular genes by Tax. cDNA profile analyses have detected several hundreds of Tax modulated cellular genes [ 82 , 134 ]. Some of them are directly involved in activation of T-cells proliferation, such as interleukin 2 (IL-2) [ 135 ] and the α subunit of its receptor (IL-2Rα) [ 136 ], which together establish an autocrine loop [ 137 ], IL-15 [ 138 ] and its receptor (IL-15R) [ 139 ], granulocyte-macrophage colony stimulating factor (GM-CSF) [ 140 ], tumor necrosis factor-α (TNF-α) [ 141 ], the MAD1 [ 142 ] and others [ 21 ]. Tax also activates cyclin D2 [ 143 ], cyclin D3 [ 144 ] and cdk6 [ 145 ], which are involved in the cell cycle progression, and inactivates p16 INKA4 [ 146 ] which acts to restrain excessive cycle progression. In addition, Tax affects the functions of many regulatory proteins by physical binding to them. A recent protein profile analysis has revealed that Tax can form complexes with 32 different proteins. Many of them belong to the signal transduction and cytoskeleton pathways and transcription/chromatin remodeling [ 147 ]. Constitutive deregulation of such regulatory factors in HTLV-1 infected T-cells can set the cells into uncontrolled continuous proliferation. Induction of such a continuous proliferation of mature T-cells is likely one of the first steps in the initiation of the ATL leukemogenic process since it renders the cells more accessible to spontaneous and exogenously induced mutagenesis. Induction of genetic instability Enhancing mutagenesis via telomerase inhibition Telomeres are specialized nucleoprotein structures located at the ends of each chromosome. In human they consist of up to 15 kb long double stranded tracts of tandem TTAGG repeats, ending with a 3' single-stranded overhangs and are associated with a number of functional proteins [ 148 ]. These structures prevent chromosomes from fusing end-to-end with each other on one hand and protect them from degradation by exonucleases on the other hand. They also enable the cells to distinguish between ends of broken DNA and natural chromosomal ends and prevent these natural ends from initiating DNA damage-specific checkpoint or repair cascades [ 148 ]. Telomeres are formed by telomerase, which are present in germ and embryonal cells and in many cancers but not in normal adult somatic cells [ 149 ]. Hence, in the absence of telomerase activity the telomeres of normal somatic cells are progressively shortened in each cell division until they reach a critical length, at which point the cells enter a quiescent viable state and are subsequently eliminated by apoptosis. However, the same shortening process may abrogate the telomere's protective effects and allow, thereby, end-to-end chromosomal fusion that forms dicentric and multimeric chromosomal structures. Such structures can break during mitosis at variable points, resulting in aneuploidy and extensive non-reciprocal chromosomal translocations and rearrangements. This chromosomal instability can lead to accumulation of various mutations, including such that inactivate important checkpoints or induce a telomere-restoring mechanism, which may result in immortalization and carcinogenesis of the cells [ 150 ]. This implies that telomerase inactivation may, actually, play an important role in tumor initiation. Consistent with this notion many established human cancers maintain stabilized telomere length either due to mutational telomerase reactivation [ 149 ], or activation of other alternative mechanisms [ 151 ]. Of note in this context is, that unlike other types of human leukemia and lymphoma, ATL cells display numerous unique chromosomal aberrations [ 152 , 153 ] resembling those resulting from telomere dysfunction, frequently seen in solid tumors [ 149 ]. Moreover, HTLV-1 Tax has been recently found as capable of inactivating telomerase in a variety of cells [ 154 ], suggesting that telomerase inhibition in infected cells of HTLV-1 carriers might be one of the mechanisms by which Tax initiates the ATL-related leukemic process. This possibility is supported by data demonstrating that infection of primary peripheral blood T-lymphocytes with HTLV-1 results in an initial decline of the cell viability which parallels with reduction in telomerase activity and that this decline is subsequently followed by an outgrowth of selected immortal survivors displaying increased telomerase activity [ 155 ]. Additional support comes from the close correlation observed between telomerase activity in ATL cell and the clinical stage of the disease. Leukemic cells of acute ATL patients display the highest telomerase activity, whereas patients with less severe clinical stage, whose leukemic cells elicit high telomerase activity, were noted to rapidly progress to the acute form, suggesting that the increased telomerase activity is not a side result of the acute ATL conditions, but rather one of the causes leading to this stage [ 156 ]. Interference with DNA repair As noted above, HTLV-1 infected T-cells show high frequency of chromosomal abnormalities [ 152 , 153 ]. The first indication that Tax is associated with cellular genetic aberration came from the observation that Tax represses the expression of polymerase-β which is involved in DNA repair [ 157 ]. This notion was later substantiated by demonstrating the capacity of Tax to enhance mutation rate [ 158 ] and other types of genetic instability [ 159 ] via impairing the chromosomal segregation fidelity and interfering with several modes of DNA repair such as the, mismatch repair (MMR), base excision repair (BER) and nucleotide excision repair (NER) [ 160 - 165 ]. Particular interest has been focused in the last few years on Tax interference with NER, since this mode of repair is regarded as a major mechanism of maintaining the genome stability and its abrogation has been linked to increased cancer incidence [ 166 ]. However, there are several unresolved questions regarding the role of some factors in these DNA repair pathways. For example, Tax has been shown to enhance the expression of PCNA [ 167 ], an essential cofactor of DNA polymerase-δ and , which are involved with DNA replication and repair [ 168 ]. Mutagenesis analysis have suggested that the ability of Tax to stimulate PCNA correlates with its ability to inhibit NER [ 162 ]. Based on this observation it has been proposed that the increase of PCNA molar ratio over polymerase-δ interferes with the DNA repair activity of this polymerase without affecting its function in DNA replication [ 160 , 162 ]. However, this explanation seems to over-simplify the complex role of PCNA in coordinating polymerase-δ activities between DNA replication and DNA repair. Of note in this context is that moderate levels of p53 also stimulates PCNA [ 169 ], but yet NER is rather enhanced in these conditions [ 161 ]. Another explanation has been based on p53 ability to elevate the level of p21 WAF-1 which binds to PCNA [ 170 ]. It has been proposed that this binding of p21 WAF-1 directs PCNA towards inhibition of DNA replication without affecting NER [ 168 ] and that Tax can prevent this pathway by its capacity to inhibit p53 transcriptional activities [ 70 , 171 , 172 ]. However, we [ 173 ] and others [ 174 ] have shown that Tax also elevates p21 WAF-1 and therefore, should accordingly, be anticipated to enhance NER rather than to inhibit it. Furthermore, other studies have demonstrated that p21 WAF-1 is not needed for NER [ 175 ] or may even inhibit it [ 176 ]. Therefore, more intensive studies are needed to resolve these conflicts. Of note however, p53 has been found to act also at the early damage-recognition step of NER [ 177 ]. It would be interesting to find out whether Tax directs its inhibitory effect towards this early step of NER. In addition, p53 has been proved to be also directly involved in BER [ 178 ], suggesting that Tax interference with BER [ 164 ] might also be exerted through its inhibitory effect on p53 function. At any rate, this genetic destabilization by Tax is certainly an important element of Tax oncogenic potential. Inhibition of topoisomerase I Topoisomerase I (Topo-I) is involved in DNA synthesis and maintenance of the genome stability by participating in DNA repair and chromosome condensation. It alters DNA topology by transiently breaking one strand of the DNA, passing the other strand through the break and finally resealing the break [ 179 ]. Tax has been found to bind to topo-I and inhibit its activity [ 180 ], whereas Topo-I activity has been shown to be stimulated by p53 [ 181 ]. Thus Tax may interfere with Topo-I activity also through its effect on p53. This might be another way for Tax to destabilize the cellular genome, but more intensive investigation is required to further substantiate this possibility. Tax-mediated protection of HTLV-1 infected T-cells from stress-induced cell cycle arrest and apoptosis All the above effects of Tax which enhance mutations and other chromosomal aberrations and interfere with DNA repair should be expected to induce cell cycle arrest or apoptosis. This, in turn, should prevent further progression of the leukemogenic process in the infected cells, unless such cells can, somehow, escape the cell cycle arrest and apoptosis. There is a substantial controversy over the influence of Tax on the cell response to stress insults. While many studies have demonstrated that Tax protects cells from stress-induced cell cycle arrest or apoptosis [ 182 - 186 ], others have shown that it enhances the cell sensitivity to these stress-induced effects [ 186 - 190 ]. Indeed, cDNA microarray analysis of HTLV-1 Tax expressing cells exposed to DNA damage stress signal revealed elevated expression of pro- as well as of anti-apoptosis genes [ 191 ]. However, results from our [ 182 ] and other laboratories reviewed in ref. [ 186 , 191 ], suggest that in HTLV-1 producing T-cells the anti-apoptotic effects of Tax override its potential pro-apoptotic effects. Besides, Tax has been shown to suppress a wide range of factors participating in the apoptosis cascade on one hand and to stimulate factors acting as apoptosis inhibitors on the other hand [ 144 , 185 , 192 , 193 ]. A possible explanation for the above noted controversy is that in most of the studies presenting Tax pro-apoptotic effect, Tax was over-expressed through highly potent promoters. Excessive levels of Tax may, reasonably, sensitize the cells to apoptosis. However, these experimental conditions do not reflect the situation in HTLV-1 expressing T-cells, in which Tax cannot exceed the optimal level required for its replication due to the Rex-mediated fine regulation of the balance between the viral RNA species encoding for the gag, pol, prot and env proteins and those which encode for the Tax and Rex proteins [ 28 , 29 ] described earlier in this review. In addition, Tax has been shown to enhance the cell cycle progression and to release cells from stress-induced cell cycle arrest [ 26 , 145 , 194 ]. Experimental models for Tax oncogenicity Numerous studies have been focused on investigating Tax oncogenic potential in cultured cells and animal models. Most of them used plasmids expressing w.t. Tax or various Tax mutants under the control of HTLV-1 LTR or other different promoters. It was only few years ago that an infectious clone of the entire HTLV-1 genome was constructed and appropriate cell culture techniques were developed to introduce this clone into human primary PBLs. This clone was found as capable of propagating in cultured mammalian cells and to transform primary human T-lymphocytes [ 195 - 197 ]. This clone opened an opportunity to study the oncogenic potential of Tax and of various Tax mutants in the context of the entire viral genome. Some of the studies to be discussed in this section have been performed with such clones. Transformation of rodent cells Tax has been shown to induce neoplastic transformation of the rat fibroblast Rat-1 [ 198 - 203 ] and the mouse fibroblast NIH/3T3 [ 201 ] cell lines. Tax has been also shown to cooperate with the ras oncogene in transforming primary embryo fibroblasts [ 203 ]. Transformation was determined by formation of foci of morphologically transformed cells, colony formation in soft agar and tumor formation in nude mice. Several mechanisms have been proposed to mediate the transformation of Rat-1 cells: 1) Involvement of NF-κB [ 199 ], 2) involvement of CREB/ATF [ 68 ], 3) involvement of phosphoinositide-3 kinase-PKB/Akt [ 202 ] and 4) Stimulation of p21 WAF-1 which prevents apoptosis and enhances the replication of the transformed cells [ 204 ]. The cooperation of Tax with ras in transforming primary embryo fibroblasts is postulated to be mediated by SRF through the CArG elements of SRF responding genes [ 198 ]. It should be emphasized, however, that maintenance of this transformation phenotype requires the continuous presence of active Tax and that no genetic mutation can be identified in these transformed cells [ 200 ]. Therefore, the process leading to this transformation is unlikely to reflect the entire pathway leading to ATL because the leukemic ATL barely express Tax [ 3 ] and they are characterized by intensive chromosomal aberration [ 152 , 153 , 165 ]. For the most, this transformation may reflect only the very early steps of the initiation of the ATL process. Immortalization and transformation of primary human T-lymphocytes A closer insight into the ATL leukemogenesis has been gained through studies using primary human T-lymphocytes. Such experiments have revealed that after infection in culture with HTLV-1 or permanent transfection with Tax, primary human T-cells undergo two stages of cellular changes. In the first stage the cell become immortalized but still remain dependent on IL-2 for their growth [ 205 ]. This immortalization has been shown to result from Tax-induced stimulation of the G1 phase-specific cyclin-dependent kinases CDK4 and CDK6, increased expression of signal transduction genes like cyclin G1, c-fgr, hPGT [ 206 ] and p21 WAF-1 [ 204 ] and to be associated with mutations conferring increased telomerase activity [ 155 ]. Studies with different Tax mutants, deficient of CREB/ATF- or NF-κB-activation, have yielded conflicting results as to which of these two major regulatory pathways is involved in this Tax-mediated immortalization. While certain studies have shown that this immortalization depends on Tax ability to activate NF-κB [ 207 , 208 ], others have demonstrated that Tax mutants deficient of NF-κB activation still retain their capacity to induce this immortalization [ 209 ]. On the other hand, experiments with an infectious molecular clone of the entire HTLV-1 genome have shown that disruption of Tax ability to activate CREB results in preferential immortalization of CD8+ lymphocytes, rather than preferential immortalization of CD4+ lymphocytes seen with the wild-type infectious clone [ 210 ]. In addition, it has been found that disruption of Tax capacity to interact with CBP/p300 does not affect its immortalizing potential [ 195 , 210 ]. In the second stage few IL-2-independent clones of transformed cell emerge. Such transformed cells display an IL-2-independent constitutive activation of the IL-2 receptor (IL-2R) signaling pathway that includes the Janus kinases JAK1 and JAK3, and the signal transducers and activators of transcription STAT3 and STAT5, which are constitutively active in such cells [ 206 , 211 - 213 ]. Other studies have shown in such transformed cells a constitutive high expression of the growth factor independence-1 (Gfi-1) which is also involved in coffering their IL-2-independent growth [ 214 ]. In addition, intensive studies have been recently focused on factors participating in a negative regulation of the IL-2R associated pathway in HTLV-1 transformed T-cells. One of these factors is the SH2-containing tyrosine phosphatase SHP1. A gradual loss of this phosphatase has been noted to correlate with the progression of HTLV-1 infected primary T-cells from the immortalization (IL-2 dependence) to the transformation (loss of IL-2 dependence) stage [ 213 , 215 ]. Changes in other negative regulators of Jak/STAT/IL-2R pathway have been noted to vary between different transformed clones and, therefore, their role in acquiring the IL-2-indepence is unclear yet [ 213 ]. Also notable is that in contrast to the rodent cells, the HTLV-1 transformed primary human T-cells display high mutation rate [ 158 ] and other genetic aberrations [ 159 , 165 , 216 ]. Of particular interest in this context is the observation that exposure of HTLV-1 infected primary T-cells to carcinogens enhances a stepwise progression from their IL-2-dependent immortalized state to the autonomous transformed state [ 217 ]. Also interesting is the association noted between chromosome changes in such cells and their growth potential [ 216 ]. Many of the above changes observed in the HTLV-1/Tax immortalized and transformed primary human T-cells are quite analogous to those found in ATL cells. However, while fresh leukemic cells from ATL patients, as well as cell lines derived from these leukemic cells, are successfully engrafted in SCID mice and their leukemic infiltration to various organs is similar to that seen in ATL patients [ 218 ], primary human T-cells immortalized or transformed in culture by HTLV-1 infection, Tax transfection or intruding a molecular clone of the entire HTLV-1 genome, do not show such tumorigenicity in these mice [ 196 , 219 ]. This observation can be explained by postulating that during their progression through multiple selection steps in the infected patient the ATL leukemic cells accumulate selected genetic changes conferring their tumorigenic phenotype, whereas under culture conditions there is no selective pressure for preferential accumulation of such particular mutations. Tumor induction in transgenic mice Tax transgenic mice have been widely used as models for investigating the oncogenic effects of Tax in-vivo, hoping to get closer insight to the ATL leukemogenic process in human [ 220 ]. A wide range of different tumors have been described in such animals and it appears that the promoter used to express Tax determines at least partially the type of the developing tumors. Transgenic mice expressing Tax through HTLV-1 LTR were found to develop neurofibrosarcomas [ 221 ], mesenchymal tumors [ 222 ] or skeletal bone abnormalities [ 223 ], but not leukemias or lymphomas. Mice expressing Tax through the promoter of CD3ε were found to develop mesenchymal tumors at wound sites and salivary and mammary adenoma [ 224 ]. Only mice expressing Tax through the granzyme B promoter showed Tax expression in mature T-lymphocytes and developed large granular lymphocytic leukemia [ 225 ]. These studies suggest that Tax alone is capable of inducing tumors in various tissues, including lymphoid cells. A possible explanation for the failure of HTLV-1 LTR-Tax to induce leukemia in such animals may be provided by the observation that expression of this construct can be detected in various non-lymphoid organs like the brain, saliva glands, spleen, thymus, skin, muscle, bones and mammary glands [ 223 , 226 ] but not in the bone marrow [ 223 ]. It has been proposed that activation of HTLV-1 LTR expression in lymphoid cells requires the cooperation of the accessory proteins encoded by ORF1 and/or ORF II of the pX region with Tax protein. Therefore, Tax alone cannot activate the expression of the HTLV-1 LTR-Tax construct in these cells, whereas this cooperation is not needed in other organs. [ 3 , 33 , 226 ] Various modes of tumor induction by the transgenic Tax have been noted so far. Hall et al. [ 224 ] have shown that the mesenchymal and the mammary adenomal tumor induced by the CD3ε-Tax transgene displayed high levels of apoptosis which is associated with high levels of Myc, Jun and p53. In contrast Portis et al. [ 227 ] have demonstrated a Tax mediated functional inactivation of p53 in the early stage of the large granular lymphocytic tumor formation by the granzyme B-Tax transgene and p53-inactivating mutations in a later stage of the tumor progression. Other studies have demonstrated the importance of Tax-mediated activation of NF-κB in the induction of both the lymphoid [ 228 ] and non-lymphoid [ 229 ] tumors. As noted before, continuous Tax expression is required for maintaining the neoplastic phenotype of Rat-1 cells transformed by Tax in culture [ 200 ]. In contrast, suppression of Tax expression in transformed fibroblasts derived from tumors of Tax transgenic mice did not affect their growth rate and ability to form tumors in animals [ 230 ], indicating that Tax was involved only in the initiation of the in-vivo tumorigenic process, after which the cells continued to progress through several genetic changes rendering their neoplastic phenotype independent of Tax. Conclusive comments about the pathways leading to ATL and TSP/HAM In view of the above described pleiotropic effects of Tax, which are summarized in Fig. 5 , it is widely accepted that the viral Tax protein is a key element in ATL genesis [ 2 ]. This implies that generating this malignancy requires active viral gene expression in the infected T-cells of HTLV-1 carriers in order to keep Tax protein at an effective level. However, as noted before, shortly after establishing the host immune response against the viral antigens, HTLV-1 virus expression is kept very low and is the level of Tax. This low Tax level accounts, likely, the "carrier" state of the infected individuals by supporting a limited but continuous expansion of the infected CD4+ cells [ 17 ] and for their anti HTLV-1 seropositive states. However, this low Tax level, plausibly, is insufficient for exerting all the above described oncogenic effects leading to ATL. Therefore, over 95% of the infected individuals do not develop this malignancy, or other HTLV-1 related clinical disorders, during their entire life [ 2 , 3 ]. Thus, generating this malignancy would, plausibly, require activation of the dormant virus in order to elevate Tax to its oncogenic threshold. Our previous studies [ 182 , 231 - 233 ] indicate that this activation can be induced by a variety of stress agents which are widely present in the daily human surrounding. However, such agents normally induce also cell cycle arrest or apoptosis, which could be expected to prevent the subsequent progression towards ATL. This paradoxical conflict was resolved by our observation that Tax protects HTLV-1 producing human T-cells from stress-induced apoptosis [ 182 ], implying that the Tax protein, emerging after activation of the latent virus, can rescue the host cells from the stressed-induced apoptosis. After this activation step there may, actually, be a progression to either ATL or TSP/HAM. Intensive studies, reviewed in ref. [ 3 , 234 ], indicate that genetic factors of the host, mainly those associated with the HLA histocompatibily complex class I, are the major factors determining whether the progression will proceed towards ATL or TSP/HAM [ 235 ]. Of note is that TSP/HAM is characterized by high virus expression [ 236 - 238 ]. Such high virus expression is widely considered to be a predisposing factor for TSP/HAM development [ 239 ]. We [ 240 ] and others [ 235 ] discussed in details in earlier review articles, how this high virus expression accounts for most of the pathological and immunological manifestations of this syndrome and correlates with its severity [ 238 ]. On the other hand, no or very little Tax can be detected in the leukemic cells of ATL patients [ 3 , 12 , 13 , 16 , 234 ]. It has been proved that anti Tax CTLs mounted in these patients eliminate the rare cells with high Tax expression and keep, thereby, Tax at very low level [ 10 , 16 ]. This difference seems to be determined by the HLA type of the host [ 14 ]. It can be postulated that the immune response of people with HLA types of high risk for TSP/HAM, permits permanent high expression of the activated virus, whereas the immune response of people with other HLA types probably act to re-suppress the activated virus. Therefore, progression towards ATL can, presumably, proceed only if a mutation, that abrogates one of the important cellular checkpoints, occurs before the activated virus is re-suppressed. This speculation is supported by reports showing that the leukemic cells of most ATL patients carry one or more mutations which deregulate the formation or function of cellular factors associated with T-cell replication, cell cycle arrest or apoptosis, such as the IL-2 receptor [ 241 ], the JAK/STAT proteins [ 242 ], the growth factor independence 1 (Gfi) [ 214 ], p53 [ 243 - 245 ], p15 INK4B and p16 INK4A [ 246 ], p27 KIP1 [ 247 ], p16 (CDKN2) [ 248 ], pRb [ 249 ], surviving [ 193 ], Fas (Apo1/CD95) [ 250 ] and caspases [ 251 ]. Interestingly, the leukemic cells of most ATL patient are defective in the mitotic spindle checkpoint [ 252 ], which likely accounts for the frequent clastogenic and aneugenic chromosomal abnormalities detected in these cells [ 153 ]. However, no mutation in mitotic checkpoint genes has been identified in such cells [ 252 ]. Instead, the viral Tax protein has been shown, in one study, to inactivate the function of the mitotic spindle checkpoint protein MAD1 [ 142 ]. In another study the MAD1 and MAD2 checkpoint proteins, which normally reside in the nucleus, have been found in HTLV-1 infected cells to localize predominantly in the cytoplasm [ 252 ]. Since Tax is hardly detected in circulating ATL cells, it is rather unlikely to ascribe this checkpoint loss in the ATL cells to Tax activity. It seems more reasonable to speculate that this loss results from mutations in other genes which might indirectly affect the proper subcellular localization of these proteins. It is also reasonable to assume that this and the other mutations in the above mentioned regulatory genes are, most likely, acquired in the pre-leukemic stage during a certain time-gap when the cells are highly susceptible to mutagenesis. We like to propose that this time-gap is the time when Tax is still highly active in a large number of circular T-cells due to the putative activation of the latent virus. It is plausible to assume that the level of the anti Tax antibodies existing in latent HTLV-1 carriers is sufficient to repress Tax expression in the few high virus-expressing CD4+ cells [ 13 ] existing before activation of the latent virus and that existing level of anti Tax CTLs in such carriers is sufficient to eliminate these cells [ 10 , 16 ], but neither of these Immune response arms is sufficient to handle the overwhelming number of such high virus-expressing cells resulting from the virus activation. However, this situation is likely temporary and may plausibly last until the anti Tax antibodies and CTLs boosted to mount to a sufficiently higher level that can overcome this large number of high virus-expressing cells. This is the time-gap during which we thing that the pre-leukemic cells are most susceptible to mutagenesis and can acquire one or more of the above mentioned checkpoint-abrogating mutations. If, after the occurrence of such mutations, Tax expression is re-suppressed by the mounting level of the anti Tax antibodies and cells that still remain with high virus expression are eliminated by the mounting CTLs, this will not stop the remaining mutant cells to further accumulate additional mutations and progress towards ATL. This re-suppressed Tax expression will avoid exposure of the progressing cells to anti Tax CTLs. However, since the probability for such particular mutations to occur during this limited time-gap might be very low, it is quite possible that multiple episodes of such virus activation and re-suppression may occur in HTLV-1 carriers before progression to ATL can be turned on. Such flow of events, which is illustrated in Fig. 6 , may explain why ATL usually develops after much longer clinical latency than TSP/HAM. Figure 6 Schematic hypothetical flow of the events occurring between the initial infection with HTLV-I and ATL or TSP/HAM development (See the text for detailed explanation). Authors' contributions Authour 1; (A.I) prepared the Fig.ures and together with author 2 (S-K.Y) covered the cited publications and prepared the draft of this review. Author 3 (A.M) designed the outlines of the review and together with the other two authors prepared the final version for submission. All authors read and approved the final manuscript.	/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC514576.xml
516771	Dam inactivation in Neisseria meningitidis: prevalence among diverse hyperinvasive lineages	Background DNA adenine methyltransferase (Dam) activity is absent in many, but not all, disease isolates of Neisseria meningitidis , as a consequence of the insertion of a restriction endonuclease-encoding gene, the ' dam replacing gene' ( drg ) at the dam locus. Here, we report the results of a survey to assess the prevalence of drg in a globally representative panel of disease-associated meningococci. Results Of the known meningococcal hyper-invasive lineages investigated, drg was absent in all representatives of the ST-8 and ST-11 clonal complexes tested, but uniformly present in the representatives of the other hyper-invasive lineages present in the isolate collection (the ST-1, ST-4, ST-5, ST-32 and ST-41/44 clonal complexes). The patterns of sequence diversity observed in drg were consistent with acquisition of this gene from a source organism with a different G+C content, at some time prior to the emergence of present-day meningococcal clonal complexes, followed by spread through the meningococcal population by horizontal genetic exchange. During this spread a number of alleles have arisen by mutation and intragenic recombination. Conclusion These findings are consistent with the idea that possession of the drg gene may contribute to the divergence observed among meningococcal clonal complexes, but does not have a direct mechanistic involvement in virulence.	Background Neisseria meningitidis , the causative agent of meningococcal meningitis and septicaemia, is a common inhabitant of the human nasopharynx, being asymptomatically carried by approximately 10% of the population [ 1 , 2 ]. There is evidence for extensive horizontal genetic exchange in populations of this antigenically and genetically variable Gram-negative bacterium [ 3 - 7 ] but, despite the diversity of carried meningococci, only a limited number of genotypes – the hyper-invasive lineages – are responsible for most reported disease [ 8 ]. These lineages have been identified by the techniques of multilocus enzyme electrophoresis [ 9 ] and multilocus sequence typing (MLST) [ 10 ] as clonal complexes. In recent years, the ST-1 complex (formerly subgroup I), ST-4 complex (subgroup IV), and ST-5 complex (subgroup III) have dominated meningococcal disease in Africa and Asia, while members of the ST-11 (ET-37) complex, ST-8 complex (cluster A4), ST-41/44 complex (lineage 3), and ST-32 (ET-5) complex have caused most disease in other parts of the world. Meningococci occasionally cause epidemic outbreaks of varying scale up to global pandemics, although at a given point in time disease in a given geographical locale is often dominated by a limited number of clonal complexes [ 8 ]. DNA adenine methyltransferase (Dam) is an enzyme involved in the mismatch repair system of bacteria. During DNA replication, only the parental strand is fully methylated by the enzyme, since methylation is not immediate, allowing a mismatch on the newly synthesised strand to be excised and replaced [ 11 ]. Disruption of Dam activity suggests that mismatch repair will be less effective, potentially allowing the creation of frameshift mutations in the homopolymeric or simple repeat motifs within the promoter or coding regions of surface antigen genes leading to the reversible switching of their expression state. This is an important mechanism found at bacterial loci that encode gene products that are advantageous under certain conditions but not others; such genes have been termed contingency loci [ 12 ]. There is extensive evidence for such loci in the meningococcal genome [ 13 ]. In particular, the ability to vary the expression state of surface-exposed cellular components is important for within host adaptation, for example by allowing evasion of the immune response in the nasopharynx. These mechanisms are also important in the occasional transition from harmless colonisation to invasive disease [ 14 - 16 ]. Examples of such loci described in the meningococcus to date include: the capsular polysaccharide [ 17 ]; the Opc outer membrane protein [ 18 - 20 ]; pili [ 21 ]; the PorA protein [ 22 ]; and opacity proteins [ 19 , 23 ]. Dam methylation has been implicated in modifying the virulence of a number of bacterial pathogens [ 24 - 26 ], but its role in N. meningitidis has been contentious. Bucci et al ., 1999 [ 27 ] suggested that the absence of Dam activity was responsible for high rates of phase variation in the siaD capsule gene, resulting in an increased virulence of some strains. In addition, the latter work suggested that all pathogenic isolates lacked Dam activity, the dam gene inactivated by the insertion of a putative restriction endonuclease, named ' dam replacing gene' ( drg ), with the genotypes dam + / drg - and dam - / drg + being mutually exclusive. However, two subsequent studies [ 28 , 29 ] have found no effect. The drg gene encodes a restriction enzyme, Nme BII that is similar to the Streptococcus pneumoniae Dpn I restriction endonuclease which cleaves at GmeATC, but not at GATC, sequences [ 30 , 31 ]. In the present study the prevalence of drg and its association with disease-associated isolates of N. meningitidis was examined by a survey of its distribution in a collection of isolates chosen to represent the known diversity of pathogenic meningococci [ 10 ]. The results suggest that, while drg is present in some lineages but not others, this gene has spread in the meningococcal population by horizontal genetic exchange, possibly after the introduction of the drg gene from an exogenous source. Results Distribution of dam and drg among invasive meningococci In all cases, there was an exact correlation with the presence of dam and the absence of drg , and vice-versa. Of the 84 isolates tested, 23 were indicated to be dam + / drg - and 61 to be dam - / drg + by restriction analysis and PCR amplification (see Additional file 1 ). The presence of dam or drg was associated with particular clonal complexes. Of the hyper-invasive lineages included in the analysis, all ST-8 complex and ST-11 complex isolates were dam + / drg - , while all of the representatives of the ST-1, ST-4, ST-5, ST-41/44, and ST-32 complexes were dam - / drg + . Diversity of drg gene fragment sequences The level of nucleotide diversity, as represented by number of alleles and polymorphic sites, along with the p-distances seen within drg gene fragments was similar to that seen with the housekeeping genes used for MLST (Table 1 ), with a total of 19 different fragment sequences ( drg-1 – drg-19 ) present in this collection. The ratio of non-synonymous to synonymous substitutions ( d N / d S , 0.11) was also similar to that seen in the MLST gene fragments. Three drg gene fragments ( drg-13 , drg-14 and drg-15 ) exhibited polymorphisms resulting in single base frameshifts, two resulting in stop codons (Figure 1 ) and one fragment ( drg-16 ) had an insertion corresponding to a 24-base direct repeat. The G+C content of drg at 42.4% was lower than that of the housekeeping genes, which were in the range 51.1–57.4%. Split decomposition analysis of the drg fragment sequences generated a star phylogeny with some net-like phylogenetic structure involving the drg-2 allele (Figure 2 ). The distribution of the different drg sequences amongst the clonal complexes is also shown in Figure 2 . Discussion The meningococcus is an example of an accidental pathogen, a commensal organism that rarely causes disease and which gains no evolutionary benefit from this process [ 32 ]. A complete understanding of the diseases caused by such organisms is dependent upon an appreciation of the mechanisms by which harmless carriage develops into invasive disease [ 33 ]. The observation that some meningococcal genotypes, the hyper-invasive lineages, are more likely to cause disease than others [ 8 , 10 ] suggests that comparative studies of disease-associated and carried meningococci may identify genetic factors responsible for disease [ 34 ]. In this context, the suggestion that the possession of a single genetic change, the insertion of the drg gene at the dam locus, was essential for virulence in meningococci [ 27 ] was attractive as it provided a single characteristic associated with the disease phenotype. Further, this proposal provided a plausible mechanistic explanation, namely the promotion of rapid switching of expression of contingency genes, for example those encoding the capsule which represents the best defined meningococcal virulence determinant [ 35 ]. This is consistent with the evidence that inactivation of Dam has been shown to affect rates of recombination and phase change in other species [ 25 , 36 ]. The data presented here, however, do not support the contention that the drg insertion is preferentially associated with disease-associated meningococcal isolates, even those expressing serogroup B capsular polysaccharide. There was no evidence of a difference between disease-associated and carried meningococcal isolates with 76% (45/59) of disease-associated isolates and 75% (11/15) of carrier isolates possessing the drg insertion. The isolate collection employed by the present investigation was chosen to be globally representative of meningococcal disease in the latter part of the 20 th century, with multiple examples of each of the major hyper-invasive meningococci identified over this period. When analysed by clonal complex, which identifies hyper-invasive lineages, it was apparent that the drg insertion was absent from all of the isolates representing the ST-8 and ST-11 clonal complexes, both major hyper-invasive lineages but, conversely, drg was present in all representatives of the hyper invasive lineages represented by the ST-1, ST-4, ST-5, ST-41/44, and ST-32 complexes. This association was independent of the serogroup expressed by the isolates and provided a likely explanation of the earlier observations [ 27 ]. As meningococcal disease in a given locale at a given time tends to be dominated by a limited number of clonal complexes [ 8 ], it is likely from the results provided here that a given sample of disease-associated isolates collected from a given locale will be uniform at the dam locus. At a given point in time they may be dominated by drg containing meningococci, while at other times isolates without this insertion might dominate. For example, as both the ST-1 and ST 32 complexes contain the drg insertion, this insertion would be prevalent among the disease isolates recovered by the pandemic outbreaks caused by these clonal complexes [ 37 , 38 ]; conversely, the spread of ST-11 meningococci [ 39 , 40 ] would result in an increased prevalence of disease-associated meningococci without the drg insertion. This observation highlights the importance of assembling genetically defined isolate collections with a prescribed sampling frame if comparisons of meningococci with potential for virulence are to be undertaken. It has been suggested that the drg gene has been introduced into Neisseria from an exogenous source [ 30 ]. The difference in the G+C% content of the drg gene compared to the meningococcal housekeeping genes was consistent with this idea; however, the levels of diversity among the drg alleles, which was similar to that observed in the meningococcal housekeeping genes, along with the reported occurrence of drg in other Neisseria species [ 30 ], suggested that this was a relatively old event predating the emergence of present-day clonal complexes. The ratio of non-synonymous to synonymous substitutions ( d N / d S ), which was also similar to that observed in meningococcal housekeeping genes, suggested that the drg gene had been subject to stabilising selection for conservation of function since its putative introduction into the meningococcal population, although the possession of frame shift mutations in some of the drg alleles demonstrated that, as with N. lactamica , not all meningococci possessing the drg insertion expressed the endonuclease [ 30 ]. In other respects the distribution of the gene among clonal complexes and the patterns of nucleotide sequence variation observed were similar to those seen in meningococcal housekeeping genes, suggesting that it is subject to similar selection pressures [ 41 ]. The presence of multiple methylation systems is a characteristic of the Neisseria and may be related to the transformable nature of these organisms [ 42 ]. Whole genome comparisons of meningococci belonging to different clonal complexes frequently yield different restriction modification systems as the principal detectable genetic differences [ 43 , 44 ]. It is attractive to speculate that the genetic isolation of clonal complexes may be promoted by such differences [ 45 ] and in this respect the insertion of drg in the dam locus, replacing a methylase with an endonuclease, may provide a potent barrier to genetic exchange from dam + to drg + meningococci; indeed, while there was evidence for the occasional horizontal genetic exchange of drg alleles among the genetically diverse meningococci that possess the insertion, there was no evidence for the transformational loss of drg from a given clonal complex among the isolates examined here. This concept of genetic isolation among lineages receives some support from the possession of porB2 alleles by members of the dam + ST-8 and ST-11 complexes and porB3 alleles by the clonal complexes with the drg insertion [ 9 , 39 ], but the evidence for limitations in gene flow between these subgroups of meningococci is not at present conclusive. Conclusions While it is possible that the possession of the drg insertion may influence meningococcal population structure, the data presented here do not support a direct association of this genotype with meningococcal virulence. Methods Bacterial isolates The 84 isolates used in this study (see Additional file 1 ) were a subset of a collection of 107 diverse meningococci assembled to develop and validate the MLST scheme for Neisseria meningitidis [ 10 ]. The collection included representatives of the major hyper-invasive lineages and has been characterised at many genetic loci (full details of this isolate collection are available at ). Restriction digestion of chromosomal DNA The methylation status of the chromosomal DNA of each isolate was determined by restriction digestion with the restriction endonucleases Dpn I, Dpn II, and Sau 3AI, followed by separation of the digestion products by agarose gel electrophoresis. Each enzyme specifically recognises a cognate target site of GATC and cleaves this site depending on whether it is methylated or not: Dpn I cleaves this target sequence only when N 6 -methyladenine is present within the recognition sequence; Dpn II cleaves only unmethylated sites, and Sau 3AI cleaves both methylated and unmethylated target sequences. PCR amplification and nucleotide sequence determination The presence of the dam gene was determined by polymerase chain reaction (PCR) performed using dam specific primers, DamF1 (5' – TAAAATGGGCAGGCGGCA – 3') and DamB2 (5' – CGTAAGGGGGATCGCAAT – 3'). These amplify a 534 bp fragment from the 5' end of the dam gene in dam + strains but not from dam -strains. The presence of drg was determined by PCR using primers DrgF1 (5' – CATGAATTTATTTTTCGATA – 3') and DrgB2 (5' – AATTTGCAACTGTTGGCG – 3') that bind to drg internal sites and produce a 705 bp fragment in isolates containing the drg gene. PCR amplification was also performed using primer pairs Drg5F (5' – TGTCTAAAGAACTCAAAG – 3') / DrgB3 (5' – CGGTATCGAAAAATAAAT – 3') and Drg3F (5' – ATCCATCCAATTTCCCCA – 3') / DamB5 (5' – AAATGCCGTCTGAA – 3') based on dam and drg coding regions in order to confirm that dam inactivation was due to drg insertion. Amplicons corresponding to the drg gene were purified by precipitation with poylethylene glycol and sodium chloride as described previously [ 46 ] and their nucleotide sequences determined on both strands by cycle sequencing. BigDye™ terminators (ABI, Foster city, California) and the same primers as used for amplification were employed in the extension reactions and the labelled reaction products separated on an ABI 3700 automated DNA sequencer. Each fragment was sequenced at least once on each strand and the sequences assembled with the STADEN suite of computer programs. Each unique drg sequence was assigned an arbitrary allele number in order of discovery. Data analysis Percentage G+C was calculated using the program START [ 47 ]. Split decomposition analysis of the drg gene fragments was performed using SPLITSTREE, version 3.1 [ 48 ]. The proportion of non-synonymous to synonymous substitutions was calculated by the method of Nei and Gojobori [ 49 ] using the program MEGA2 [ 50 ]. Authors' contributions KAJ carried out the nucleotide sequencing and analysis of the data. LS performed the PCR screening of the isolates. ERM and MCJM conceived of the study and participated in its design and coordination. Supplementary Material Additional File 1 Bacterial isolates used in the study Click here for file	/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC516771.xml
516017	Aberrant CBFA2T3B gene promoter methylation in breast tumors	Background The CBFA2T3 locus located on the human chromosome region 16q24.3 is frequently deleted in breast tumors. CBFA2T3 gene expression levels are aberrant in breast tumor cell lines and the CBFA2T3B isoform is a potential tumor suppressor gene. In the absence of identified mutations to further support a role for this gene in tumorigenesis, we explored whether the CBFA2T3B promoter region is aberrantly methylated and whether this correlates with expression. Results Aberrant hypo and hypermethylation of the CBFA2T3B promoter was detected in breast tumor cell lines and primary breast tumor samples relative to methylation index interquartile ranges in normal breast counterpart and normal whole blood samples. A statistically significant inverse correlation between aberrant CBFA2T3B promoter methylation and gene expression was established. Conclusion CBFA2T3B is a potential breast tumor suppressor gene affected by aberrant promoter methylation and gene expression. The methylation levels were quantitated using a second-round real-time methylation-specific PCR assay. The detection of both hypo and hypermethylation is a technicality regarding the methylation methodology.	Background Allelic loss of heterozygosity (LOH) of the human chromosome 16q in several sporadic cancer types, including breast, prostate and ovary cancers, suggests this chromosome arm harbors tumor suppressive loci [ 1 - 3 ]. The most frequent region of allelic loss occurs within 3-megabases (Mb) at 16q24.3 between the marker D16S498 and the telomere [ 1 , 4 ]. Sequencing of the 3-Mb region has identified approximately 100 genes [ 4 ]. Eight of these have been excluded as potential tumor suppressors for breast cancer based on mutation analysis in tumor DNA [ 5 ]. Recently, a messenger RNA (mRNA) expression survey was completed within 2.4-Mb of this region examining the expression profiles of over 75 genes in a panel of breast tumor cell lines [ 6 ]. It was found that only three genes exhibited significantly aberrant expression profiles. These genes were highly expressed in some cell lines and lowly expressed in others. This led to the hypothesis that this aberrant expression may reflect a role for these genes as tumor suppressors in determining cancer phenotype and behavior. One of these genes was the core-binding factor, alpha subunit 2, translocation to 3; termed CBFA2T3. CBFA2T3 encodes for a protein that belongs to the eight-twenty-one (ETO) family, which also includes the genes CBFA2T1, CBFA2T2 in mammalian cells and nervy in Drosophila [ 7 ]. The mammalian members of this protein family are involved in therapy-related chromosomal translocations causing acute myeloid leukemia [ 8 ]. The CBFA2T3 gene encodes two alternative transcripts, CBFA2T3A (NM_005187) and CBFA2T3B (NM_175931) (Figure 1A ). CBFA2T3A reportedly functions as a nuclear transcriptional co-repressor via its interaction with histone deacetylase (HDAC) complexes [ 9 ]. CBFA2T3B functions as a kinase anchorage protein in T lymphocytes and may play a role in inflammatory response [ 10 ]. Recently, it was demonstrated that CBFA2T3B functions as a transcriptional repressor and exhibits in vitro characteristics consistent with tumor suppressor activity [ 11 ]. This gene was found to be lowly expressed in a number of breast tumor cell lines and upon re-introduction it reduced their growth on plastic and in soft agar. Figure 1 CBFA2T3 gene structure and promoter methylation patterns. (A) CBFA2T3 encodes two alternative transcripts, CBFA2T3A and CBFA2T3B. CBFA2T3A is encoded by exons 1A and 2–12. CBFA2T3B is encoded by exons 1B-12 splicing out exon 3. Relative exon sizes are shown. The exon 1A start site contains no CpG island. The black box marks the location of a high-density CpG island located five prime to the exon 1B start site. The black arrowheads mark the primers used for real-time RT-PCR. (B) CBFA2T3B contains a CpG island of approximately 160 CpG sites spanning 1-kb of sequence. The single black bars represent CpG sites scaled relative to each other. CAT ELISA promoter constructs and primers used for MSP, real-time MSP and bisulfite sequencing are shown. The asterisk marks the location of the amplicon and internal primers used for second-round real-time MSP. (C) CBFA2T3B promoter methylation patterns were examined in hypermethylated cell lines using sodium bisulfite sequencing. A characteristic sinusoidal pattern of approximately six high to low frequency methylation levels every 40–150 bp was detected. The high-frequency cytosine methylation levels residing within a consensus Sp1 binding site located approximately minus 450 bp from the transcriptional start of exon 1B are shown. Mutation analysis in breast tumor cell lines and primary breast tumor samples failed to identify any CBFA2T3 sequence aberrations [ 11 ]. It was recognized that aberrant promoter methylation might be the mechanism responsible for the altered expression of CBFA2T3 in breast tumors. The expression of several tumor suppressor genes has been found inactivated or reduced in tumors in association with promoter hypermethylation [ 12 ]. Promoter hypermethylation can occur in conjunction with allelic loss and or mutation and is regarded as an alternative form of 'knockout' in biallelic inactivation. Accumulating evidence now suggests that promoter hypermethylation may affect genes that reside within regions of frequent allelic loss more often than mutation [ 13 ]. Alternatively, several oncogenes have been found to be over-expressed in tumors in association with promoter hypomethylation [ 14 ]. As such, aberrant promoter methylation is considered a fundamental process in developing cancers and has recently received considerable interest as a rapid non-invasive molecular screening tool for the early detection of tumor cells in a range of bodily fluids and biopsy specimens [ 15 ]. In this study, the methylation status of a high-density CpG island promoter region located five prime to the exon 1B sequence of the CBFA2T3B transcript is described (Figure 1A ). We explored whether this region is aberrantly methylated in breast tumor cell lines and primary breast tumor samples and whether a correlation exists between methylation and gene expression. Both aberrant hypo and hypermethylation levels were detected in breast tumors in correlation with elevated and reduced expression. The phenomenon of hypo and hypermethylation relates to the amount of DNA used in the methylation methodology as discussed. Results Isoform-specific analysis of CBFA2T3 gene expression levels CBFA2T3 encodes two alternative transcripts, CBFA2T3A and CBFA2T3B (Figure 1A ). It was recently demonstrated that CBFA2T3 expression levels are aberrant in breast tumor cell lines. In this study, the total expression levels were assayed using real-time RT-PCR and primers that span exons 4–5 of the CBFA2T3 gene. The tumor suppressor activity previously shown for CBFA2T3B, led us to examine the expression profile of this transcript in breast tumor cell lines using a TaqMan probe specific for exon 1B. CBFA2T3B displayed aberrant expression similar to the total (Figure 2A ). This aberrant expression was found to be endogenously low in all samples examined. The expression of this gene was so low that it could not be reliably detected using Northern Blots or RNase protection (data not shown). In addition, expression levels of the CBFA2T3A transcript were assayed in breast tumor cell lines using a TaqMan probe specific for exon 1A. CBFA2T3A expressed at lower levels than CBFA2T3B, but due to the identification of several complex splice variants between exon 1A and exons 1B, 2, 3 and 4, further analysis is required (data not shown). An example of the raw data real-time RT-PCR expression analysis of the CBFA2T3 transcripts is shown ( Additional file 1 ). Figure 2 CBFA2T3 gene expression levels, 5-Aza-dC re-expression and promoter activity. (A) CBFA2T3 gene expression levels were assayed using real-time RT-PCR. Several breast tumor cell line and normal tissue sample expression levels are shown. The y-axis represents mRNA mlcls expressed per 10 4 cells shown on a log scale (mean ± SD, n = 6). Fold changes in expression relative to normal breast expression are shown above each sample. The white diamonds and white squares represent expression levels of the housekeeping genes cyclophilin A (CYPA) and ATPase coupling factor 6 subunit (ATP5A), respectively. CBFA2T3B and CBFA2T3 expressed at endogenously low yet aberrant levels in breast tumor cell lines. Using normal breast as a reference, CBFA2T3B (600 mRNA mlcls per 10 4 cells) and CBFA2T3 (1,800) were low compared to ATP5A (600,000) and CYPA (1,600,000). CBFA2T3 expression ranged 30,000-fold from 4 to 120,000 mRNA mlcls per 10 4 cells in MDA-MB-231 and BT-483, respectively. In contrast, CYPA and ATP5A expression ranged 2-fold and 20-fold, respectively (100–200 and 15–300 mRNA mlcls per cell). Expression levels were also examined in several primary breast tumor samples for which total RNA was available (see Figure 6). (B) CBFA2T3 re-expression levels were examined in MDA-MB-231 cells using 5-Aza-dC. Fold changes in CBFA2T3B, CBFA2T3 and SYK expression levels are shown upon exposure to 5-Aza-dC relative to control untreated cells. > 100-fold re-expression was detected in CBFA2T3 and CBFA2T3B. 5-Aza-dC had no affect on CBFA2T3A expression (data not shown). > 1,000-fold re-expression was also detected in SYK, a control gene known to be hypermethylated and down-regulated in MDA-MB-231 cells. (C) CBFA2T3B promoter activity was assayed using CAT ELISA. Promoter constructs labeled A to D (see Figure 1B) were inserted upstream of the CAT reporter in pBLCAT3 (Boehringer Mannheim). pBLCAT2 driven by the tyrosine kinase promoter was used as a positive control. 2.0 × 10 6 293T cells were transfected in triplicate using Lipofectamine 2000 (Invitrogen) with 1.5 μg of construct and control vector and 0.3 μg of the internal pSVβ-galactosidase control vector (Stratagene). Cells were lysed after 24 h and CAT concentrations determined using ELISA. Of the four constructs labeled A to D, the 1-kb B construct promoted a 30-fold increase in CAT expression (mean ± SD are triplicates, n = 3). Qualitative analysis of CBFA2T3B promoter methylation levels Based on the idea that this altered expression might develop via aberrant methylation, the promoter activity and methylation status of a high-density CpG island located five prime to the exon 1B sequence of CBFA2T3B was examined (Figures 1A and 1B ). The promoter activity was assayed using chloramphenicol acetyl-transferase (CAT) ELISA and confirmed that a 1-kb region spanning the island was capable of promoting a 30-fold increase in CAT expression (Figure 2C ). To determine if this promoter region is aberrantly methylated in breast tumors, 24 breast tumor cell lines, 20 primary breast tumors, 20 normal breast counterparts and 24 normal whole blood samples were screened using methylation-specific PCR (MSP). Four separate 100–200 bp regions spanning the promoter were amplified using primers specific for either unmethylated or methylated cytosines at the CpG sites shown (Figure 1B ). The full results of this analysis are summarized in Additional file 2 . In general, a low 'basal' level of methylation was detected in all samples at the various regions examined. An example of this basal methylation is shown for the normal blood samples at region two (Figure 3A ). Unlike the bloods, several breast tumor cell lines, primary breast tumors and normal breast counterpart samples displayed complex high to low methylation levels. An example of this complex methylation is shown for the primary breast tumors and their normal counterparts at region four (Figure 3B ). This analysis revealed that only few cell lines displayed clear hypermethylation (e.g. MDA-MB-231) or hypomethylation (e.g. BT-483) in association with reduced and elevated expression. To support this association, it was found that treatment of MDA-MB-231 cells with the demethylating agent 5-aza-2'-deoxycytidine (5-Aza-dC) was capable of increasing both CBFA2T3 and CBFA2T3B expression levels by > 100-fold relative to controls (Figure 2B ). Overall, however, it was difficult to comprehend this complex promoter methylation and an association with expression thus suggesting that further analysis was required. Figure 3 CBFA2T3B promoter methylation levels examined using MSP. The full results of this analysis in breast tumor cell lines, primary breast tumors, normal breast counterparts and normal whole blood samples are summarized in Additional file 2. (A) An example of the characteristic basal methylation levels detected in all samples is shown for normal blood samples examined at region 2. (B) An example of the complex high to low methylation levels is shown for 20 primary breast tumor samples with adjacent normal breast counterparts. The pUC19 DNA/MspI marker concentrations reflect an approximate concentration of 100 to 1 unmethylated to methylated mlcls. The asterisks indicate the samples examined by bisulfite sequencing. Bisulfite sequence analysis of CBFA2T3B promoter methylation patterns To understand this complex promoter methylation, sodium bisulfite sequencing of the hypermethylated breast tumor samples was used to examine the pattern and frequency of methylation with this region. This analysis revealed that specific cytosines appeared more susceptible to methylation compared to others (Figure 4 ). A characteristic sinusoidal pattern of approximately six high to low frequency methylation levels every 40–150 bp was detected in cell lines and tumor samples (Figures 1C and 4 ). Based on this pattern, primers were designed specific for the real-time MSP quantitation of high-frequency cytosine methylation levels residing within a consensus Specificity protein (Sp1) binding site located approximately minus 450 bp from the transcriptional start of exon 1B (Figures 1B and 1C ). As Sp1 proteins are commonly known to regulate gene transcription, it was considered that variable methylation at this site may be reflective of elevated or reduced expression and thus suitable for resolving and correlating the complex promoter methylation levels. Figure 4 CBFA2T3B promoter methylation patterns examined using bisulfite sequencing. The methylation patterns in breast tumor cell lines, primary breast tumors, normal breast counterparts and normal whole blood samples are shown. The y-axis represents protected 5-methylcytosines scored as percent cytosines methylated per 5–10 clones. The complete methylation maps displaying 160 CpG sites spanning 1-kb of sequence are shown for breast tumor cell lines only. The white bars indicate the Sp1 sites targeted by second-round real-time MSP. Bisulfite sequencing of the CBFA2T3B promoter region in hypermethylated breast tumor samples MDA-MB-231, MDA-MB-468 and 14T revealed a characteristic sinusoidal methylation pattern. This sinusoidal pattern was also detected in samples SK-BR-3, 14N and 17T at levels approximately 1 tenth of the hypermethylated samples. No methylation was detected in the normal blood samples 4B and 5B or the breast tumor samples 3T and MCF-7. Quantitative analysis of CBFA2T3B promoter methylation levels To assay CBFA2T3B promoter methylation levels at this Sp1 site, a bisulfite sequencing amplicon spanning this region was initially PCR amplified and column purified from 24 breast tumor cell lines, 55 primary breast tumors, 22 normal breast counterparts and 46 normal whole blood samples. Second-round real-time MSP was performed on the amplicons using internal forward primers to detect for either unmethylated or methylated cytosines at the Sp1 site. Standard curve dilutions of previously prepared internally primed clones representative of either unmethylated or methylated sequence were used to extrapolate methylation levels and normalize for differences in amplification efficiencies. The methylated cytosines were expressed as a fractional ratio of unmethylated cytosines to determine the methylation indices [mi = m/(m + u)]. On plotting the indices, a clear difference between the tumor, normal groups and complex promoter methylation levels was revealed (Figure 5 ). The normal blood samples maintained a specific basal methylation level and were similar to normal breast counterparts with methylation indices ranging from 0.006–0.09 and 0.002–0.08, respectively. Median methylation index levels were 0.02 in normal bloods and 0.01 in normal breast counterparts. In contrast, several breast tumor cell lines and primary breast tumors were highly variable relative to the normal samples with methylation indices ranging from 0.0002–0.8 and 0.0001–0.9, respectively. A Levene's test revealed a statistically significance difference in variances of methylation indices with the tumor groups being more varied than the normals (P = .001). It was predicted that 83–75% of breast tumor cell lines and 78–69% of primary breast tumors displayed aberrant methylation levels outside the methylation index interquartile ranges of the normal blood (0.02–0.04) and normal breast counterpart samples (0.008–0.03), respectively. Half of the aberrations detected in breast tumor cell lines were either hypo or hypermethylated relative to the normal breast counterpart interquartile ranges. Up to 22% of the primary breast tumors were hypomethylated and 47% hypermethylated relative to the normal breast counterpart interquartile ranges. An example of the raw data real-time MSP and melt curve analysis showing the aberrant methylation levels in breast tumor cell lines compared to normal whole blood samples is shown ( Additional files 3 and 4 ). Figure 5 CBFA2T3B promoter methylation levels assayed using second-round real-time MSP. The methylation levels in normal whole blood samples (nwb), normal breast counterparts (nbc), primary breast tumors (pbt) and breast tumor cell lines (btcl) were assayed at the Sp1 site shown in Figure 1C using real-time MSP. The y-axis represents methylation levels plotted as methylation indices [mi = m/(m + u)] on a log scale. Each white circle represents a different sample. The breast tumor cell lines examined are shown in descending order from high to low methylation. The horizontal bars mark the median methylation indices calculated for each group. The asterisks mark the interquartile ranges for normal groups. The median methylation indices and (interquartile ranges) were 0.02 (0.02–0.04) for nwb, 0.01 (0.008–0.03) for nbc, 0.03 (0.009–0.08) for pbt and 0.02 (0.002–0.3) for btcl. The median methylation index variance in each tumor group was statistically significantly different than the normal groups (P = .001); nwb/btcl (P < .0001), nwb/pbt (P = .009), nbc/btcl (P = .01), nbc/pbt (P = .05). The normal group median methylation index variances were not significantly different; nwb/nbc (P = 0.6). Correlation of CBFA2T3B promoter methylation and gene expression levels To correlate CBFA2T3B promoter methylation levels with gene expression, the methylation indices from 24 breast tumor cell lines and 20 primary breast tumor samples were plotted against their expression. A statistically significant inverse correlation between CBFA2T3B promoter methylation and exon 1B specific expression was established (r 2 = 0.63; r = -0.8, P = .0002). Based on the possibility that five prime RNA degradation and secondary structures may have affected the exon 1B complementary DNA (cDNA) synthesis, a correlation between methylation and the total expression is alternatively shown (Figure 6 ). CBFA2T3B promoter hypermethylation and reduced expression inversely correlated with hypomethylation and elevated expression (r 2 = 0.72; r = -0.9, P < .0001). At a hypermethylated index of around 0.9, approximately 4 mRNA molecules (mlcls) per 10 4 cells were detected compared to a hypomethylated index of 0.0001 and 120,000 mRNA mlcls per 10 4 cells. The number of CBFA2T3B promoter mlcls methylated per cell for each breast tumor cell line was also calculated by multiplying the methylation index values by the number of 16q24.3 DNA mlcls per cell as previously determined by FISH [ 16 ]. These values were plotted against expression in aim to improve the original correlation (r 2 = 0.77; r = -0.9, P < .0001) ( Additional file 5 ). By plotting these data sets a power regression was derived which could be used to solve unknown x or y values. Figure 6 CBFA2T3B promoter methylation levels versus gene expression. The data for 24 breast tumor cell lines and 20 primary breast tumor samples are shown. The y-axis represents methylation levels assayed using real-time MSP and plotted as methylation indices [mi = m/(m + u)] on a log scale. The x-axis represents total expression levels assayed using real-time RT-PCR and plotted as mRNA mlcls per 10 4 cells on a log scale. Each white circle represents a different primary breast tumor sample and the black circles represent the breast tumor cell lines. The breast tumor cell lines examined are shown in descending order from high to low methylation. A statistically significant inverse correlation was established between promoter hypermethylation (mi = 0.9) and reduced expression (4 mRNA mlcls per 10 4 cells) versus hypomethylation (0.0001) and elevated expression (120,000) (r 2 = 0.72; r = -0.9, P < .0001). A power regression (y = cx b ) describes the relationship between methylation and expression. Discussion In this study, it has been demonstrated that expression of the CBFA2T3B isoform is altered in breast tumors and that this correlates strongly with aberrant CpG island promoter methylation. Moreover, a comprehensive method for the detection, quantitation and correlation of promoter methylation and gene expression levels has been developed. MSP was used in combination with sodium bisulfite sequencing to identify sites within the CBFA2T3B promoter region displaying high-frequency methylation in breast tumors. Second-round real-time MSP was used to quantitate methylation levels at these sites in breast tumor cell lines, primary breast tumors, normal breast counterparts and normal whole blood samples. The CBFA2T3B promoter methylation levels were calculated as methylation indices and the indices from breast tumors were plotted against their expression. Validated methylation detection using the MSP and real-time MSP methods Throughout the development of this method, it was recognized that a pre-requisite for valid MSP amplification required that sufficient amounts of bisulfite modified DNA were used in order to detect methylation and avoid stochastic effects when quantitating low target mlcl numbers and parameters relating to particle distribution statistics. The amount of DNA modified and subsequently amplified is an important parameter in terms of actually detecting potential CpG methylation in a given specimen as discussed. Most MSP studies use 50 ng of bisulfite modified DNA for amplification. As a normal diploid cell contains an average 6.6 ρg of DNA, this equates that approximately 15,000 alleles or mlcls are made available as starting templates for MSP. If for example, 1 mlcl in 1,000 of these is methylated (mi = 0.001) for a given gene at a specific CpG site(s), then approximately 15 mlcls will act as potential starting templates. Routinely, in unmodified DNA specimens a 'purified' 100–300 bp amplicon diluted to this mlcl number should be detected at around 28–30 cycles of real-time PCR under standard primer efficiencies. Initially, this is a late cycle threshold (C T ) for the detection of amplification and is prone to stochastic effects. Under MSP conditions, the DNA has been bisulfite modified which introduces numerous variables that further reduce the overall probability of detection. Upon modification multiple sequence permutations at the CpG site(s) of interest may arise (e.g. from 3 CpG sites a total 8 possible C to TpG permutations exist as 2 3 = 8). If primers are designed to detect a permutation present at say 4 mlcls in the above example of 15 mlcls (i.e. at attogram starting amounts), then under MSP, a 100–300 bp amplicon at this mlcl number will not amplify until after 35 cycles or may not amplify at all. Moreover, because the initial cycles of MSP are asymmetric the starting template mlcl numbers are further reduced. In addition, it is well known that the DNA is substantially degraded following bisulfite modification with studies demonstrating up to 84–96% degradation [ 17 ]. If 1 μg of DNA is modified, as routinely reported, and say 80% is degraded, then based on an unlikely recovery rate of 100%, only 20% of the original pool of templates used in the 50 ng will be available for amplification. Thus, in the above example if only 20% of the 15 mlcls and its possible permutations are available then they may not amplify at all. The pool of primer specific methylated templates could in fact be instantly diminished. When considering the levels of bisulfite-mediated degradation in 1 μg of DNA, even alleles methylated at levels as high as 1 in 10 might be difficult to detect or become spurious. For instance, if assuming that a 100 times more template is available (i.e. 1,500 mlcls) yet permutations exist (i.e. 400 mlcls) and the recover rate is only 4%, then as little as 16 mlcls will be available as starting templates for amplification. Thus, several technical thresholds exist for the detection of methylation using MSP. The use of 1 μg of DNA for modification, 50 ng for amplification, sequence permutations, PCR efficiency and bisulfite-mediated DNA degradation all reduce the overall probability and validity of detection. To warrant potential detection of methylation in a given specimen at specific CpG sites above a threshold of say 1 in 1,000, it was found necessary to use at least 10X coverage of the 1 μg amount or 10 μg of DNA for modification and 400–500 ng for MSP amplification. Taking into account the above example, this should provide that at least 30–40 mlcls are made available as starting templates, which should amplify within 28 cycles. Specimens with methylation levels lower than this, such as 1 in 50,000, will not be reliably detected and require the use of second-round nested MSP amplification. In this case, the first-round will still require sufficient amounts as any initial stochastic effects may result in poor reproducibility for the second-round. Moreover, as demonstrated in this study, bisulfite sequence analysis of the CBFA2T3B promoter methylation patterns in the hypermethylated MSP samples indicated that the potential for variable methylation frequencies do exist. In this case, it was found that even with sufficient amounts of DNA the qualitative MSP was unreliable and presented complex methylation data. The use of MSP primers for sites that are not methylated or methylated at low to high frequencies, in combination with these other technical thresholds, created complex stochastic methylation data merely decipherable by examining several regions. As a result, this concealed the true methylation status in a majority of the samples under investigation. Only those samples with clear hypermethylation (e.g. MDA-MB-231) or hypomethylation (e.g. BT-483) were greatly reproducible and associated with reduced and elevated expression (Figure 2A and Additional file 2 ). To overcome these technical thresholds, a second-round real-time MSP assay was developed. A sequencing amplicon displaying high-frequency methylation sites was initially PCR amplified under 10X coverage conditions and column purified from all samples. Second-round real-time MSP was performed on approximately 100 ρg of the amplicon using internal forward primers to detect for either unmethylated or methylated cytosines at the high-frequency sites. Using this approach, methylation was detectable in every sample examined. All samples amplified with 28 cycles and a methylation index was calculated for these. In comparison, when the standard MSP conditions were assayed by real-time using SYBR Green I, it was found that the C T of methylation detection was below 28 cycles for most samples (data not shown). Extremely low methylation levels, such as those in BT-483, were undetectable using this method but were detected using the second-round. In the latter case, up to 1 methylated mlcl in 10,000 (mi = 0.0001) could be reliably detected within 25 cycles of amplification (Figure 5 and Additional file 3 ). Notably, no wild-type, unmethylated or methylated cross-amplification was detected when using more DNA. If this occurs then there may be a problem with primer design or modification as the conversion should be complete. In fact, with more DNA a bias in amplification for the target sequence should be created. In addition, several other technicalities were overcome. The use of only a single second-round primer to quantitate the methylation levels of the high-frequency sites limited the number of possible 5-methyl-CpG sequence permutations that could perturb the accuracy of quantitation. This coupled with the quantitation of methylation levels from absolute standard curves enabled the normalization of reaction efficiencies and calculation of absolute methylation indices [mi = m/(m + u)] and absolute methylation ratios (u/m). The methylation ratios are simply an alternative way of expressing the methylation data ( Additional file 6 ). The calculation of these single methylation values based on the quantitative normalization discriminates against biased amplifications and comparisons of unmethylated to methylated 'band' intensities when using MSP. Moreover, the absolute quantitation of indices and ratios offers improvement over the relative methods, such as comparative cycle thresholds (ΔΔC T ), as they have biological significance, are less consuming, more accurate, and do not require the dynamic range in amplification efficiencies of target and references to be similar to enable valid quantitation. The concept of gene promoter hypo and hypermethylation Thus, by using this method it was possible to detect CBFA2T3B promoter methylation in all samples. This phenomenon may actually be widespread for genes under the control of methylation as in accordance with the replication model of maintenance methylation [ 18 - 20 ]. This model might suggest that the CpG island is not just randomly methylated; the CpG island is 'always' methylated by memory yet propagated at variable levels for cell type-specific expression or at aberrant levels in association with cancers. Remarkably, in this study it was found that the methylated CBFA2T3B CpG island is propagated as a sinusoidal pattern at aberrant levels in both permanent cell lines and recently resected tumor specimens suggesting that a methylation memory does exist. In fact, it was recognized that the bisulfite sequencing detection levels of this sinusoidal pattern complemented the second real-time MSP quantitations. For example, a methylation ratio in SK-BR-3 of 10:1 (mi = 0.1) was concordant with a 1 in 10 or 10% 5-methylcytosine per 10 clone frequency (Figure 4 and Additional file 6 ). If by comparing the other bisulfite sequencing levels with their methylation ratios this might suggest that up to 30, 60, 90 or 5,000 clones of 4B, 5B, 3T and MCF-7, respectively, would require sequencing in order to detect 1 methylated sinusoidal mlcl. Based on the idea that promoter methylation may occur ubiquitously, the concept arises; what is hypo and hypermethylation. In this study, it was demonstrated that hypo and hypermethylation are merely a prediction of levels outside a majority or 'interquartile range' of methylation levels found in the normal samples. For instance, of all the methylation levels detected in the cell lines, only 37.5% of these were predicted as either hypo or hypermethylated relative to the interquartile ranges of normal breast methylation. Moreover, the number of predicted hypo and hypermethylated samples are much lower when compared to the 'full range' of normal methylation ratios as shown in Additional file 6 . This is considerably lower for the primary breast tumor samples (i.e. only 24% hypermethylated relative to normal breast) and is likely due to the heterogeneity of breast tumors in concealing the true tumor-related methylation levels. Regardless, by using a more sensitive detection method, the phenomenon of hypo and hypermethylation has appeared. The phenomenon hypomethylation is a reflection of the low methylation levels that can be detected in the 'hypomethylated' samples relative to the methylated and hypermethylated samples. Alternatively, the phenomenon of hypermethylation is readily observed yet indicates that the detection of methylation alone does not simply represent hypermethylation. The CBFA2T3B gene is methylated in all samples and according to MSP in a high-percentage of samples. In effect, several studies which have examined the methylation status of potential tumor suppressors and cancer-related genes often demonstrate methylation in a high-percentage of samples and state hypermethylation based on the detection of methylation. Examples include, ras-effector nore1A (RASSF1A) [ 21 , 22 ], stratifin (14-3-3σ) [ 23 ], p15 and p16 [ 24 ], O6-methylguanine-DNA methyltransferase (MGMT) [ 25 ], mismatch repair gene (hMLH1) [ 26 ] and hyperplastic colon polyps gene (HPP1) [ 27 ]. In these cases, as with CBFA2T3B, it is probable that only a small percentage of this methylation has cancerous significance in terms of 'methylation-induced silencing', which is routinely confirmed by correlation with 'absence' of expression. Notably, several studies use insufficient amounts of RNA and cDNA for expression analysis and state the absence of amplification to represent inactivation rather than a level of reduction, particularly in cases of low endogenous expression. The need for quantitation to classify methylation levels has been recognized for genes such as glutathione S-transferase P1 (GSTP1) [ 28 ] and adenomatous polyposis coli (APC) [ 29 ], although in these cases the use of low DNA may nevertheless affect intra and interassay reproducibility. In biological terms, several scenarios exist as to how these aberrant methylation levels might develop. In the case of hypomethylation, a correlation could be made with severely duplicated chromosome copies of 16q24.3 (e.g. BT-483, MCF-7). A possible scenario here is that the hypomethylation is apparent because of a duplicated copy number and thus, for example, not a direct cause of reduced DNA methyltransferase activity or over-expression per se. In the case of hypermethylation, a correlation could be made with 16q24.3 LOH (e.g. MDA-MB-231, MDA-MB-468). An emerging scenario here is that the hypomethylation induces 16q LOH to promote aberrant DNA methyltransferase activity and hypermethylation [ 13 ]. Accumulating evidence suggests that the hypermethylation itself, or 'aberrant' methylation, may be targeted to constantly methylated CpG islands (i.e. methylation induces methylation), and in addition targeted to transcriptionally inert CpG islands [ 19 ]. In this study, it was found that the CBFA2T3B promoter region is in fact constantly methylated at a median methylation index of 0.02 (i.e. 2 mlcls in 100 are methylated). When comparing this median methylation index with a median gene expression index from the breast tumor cell lines, it was calculated that only approximately 20 mRNA mlcls per 98 'active' alleles are transcribed. These calculations are shown in Additional file 5 . This suggests that the CBFA2T3B gene is not only constantly methylated but also largely transcriptionally inert with the remaining active alleles possibly trans-factor dependent for expression. Although the nature of such targeted aberrations are unknown, several studies demonstrating that altered maintenance and or de novo DNA methyltransferase activities can induce tumorigenesis, clearly demonstrates that the controlled methylation of CpG islands is crucial for normal cell development [ 30 - 34 ]. Accordingly, an in silico prediction of several Sp1, homeotic, epidermal and insulin growth factor recognition sites within the CBFA2T3B promoter region may implicate a role for this element in epithelial development. How the CBFA2T3B CpG island is maintained and dispersed at specified levels within a population of cells is unknown but likely relates to the methylation machinery in controlling distribution within a cell type-specific population (i.e. methylation memory and or phenotype). In this study, it was shown that aberrant deviations outside these specified levels occur profoundly in a majority of breast tumors, particularly the pure tumorigenic cell lines, and as such might comprise an element in tumor formation. Conclusions Overall, this study lends further support to the idea that CBFA2T3B is aberrantly regulated in breast cancer. Additional clues supporting a role for this gene in tumor suppression may reside within its protein structure. CBFA2T3B contains a characteristic zinc finger myeloid-nervy-DEAF-1 (zf-MYND) domain and nervy homology regions [ 7 ]. Several studies demonstrate that these regions function in transcriptional co-repression via interaction with HDAC and or nuclear co-repressor complexes [ 9 , 35 - 38 ]. Accumulating evidence suggests the zf-MYND domain, which is also common to developmental proteins RP-8, DEAF-1, suppressin, Blu, BS69, PDCD2 and Bop, may interact with co-repression complexes to regulate cell-cycle transcription during cell type-specific differentiation [ 39 - 46 ]. Abnormal regulation may be central to tumor formation as supported by reports that several zf-MYND-like proteins display tumor suppressive activity. Recently, much emphasis has been placed on the development of methylation-based tumor biomarkers for early breast cancer detection to predict disease outcome and strategies for therapy [ 15 ]. Interesting data indicates that the use of nipple aspirate fluids may provide for a rapid non-invasive source of screening material for methylation biomarker analysis [ 47 ]. Similar studies are underway to evaluate if the methylation status of CBFA2T3B presents biomarker utility. Although in this case, because the normal breast fluids and breast tumors themselves are histologically complex tissues containing a variety of cell types, it is recognized that further studies are required to refine the methylation index interquartile ranges. This may involve large-scale comparisons between normal and tumor cells captured using laser microdissection. Ideally, such normal controls would be resected at autopsy or reduction mammoplasty from non-risk category individuals. Methods Sample collection and nucleic acid isolation 24 breast tumor cell lines were obtained from the American Type Culture Collection and cultured under recommended conditions. 46 normal whole blood samples, 55 primary breast tumor samples with pathologically classified grade III lesions and 22 adjacent normal breast counterparts samples were obtained with clinical research approval from the Flinders Medical Centre, Department of Surgery. Breast tumor cell line, primary breast tumor and normal breast counterpart genomic DNA was isolated using GenElute for Mammalian Tissues (Sigma). Whole blood genomic DNA was isolated using the QIAamp DNA Blood Kit (Qiagen). Breast tumor cell line and primary breast tumor total RNA was isolated using RNAqueous-4PCR (Ambion). Nucleic acid concentrations were determined using RiboGreen (Molecular Probes). Breast tumor cell line, primary breast tumor and normal tissue (Clontech) total RNA extracts were DNase I treated (Ambion). Real-time reverse transcription-PCR (RT-PCR) 10–20 μg of total RNA was oligo(dT) 16 reverse transcribed at 55°C for 2 h using MMLV (Promega) with addition of RNAguard (Promega) and 5% DMSO. CBFA2T3, CYPA, ATP5A and SYK expression levels were assayed by real-time RT-PCR using SYBR Green I (BMA). CBFA2T3 isoform expression levels were assayed using TaqMan probes specific for exons 1A and 1B (GeneWorks). Real-time RT-PCR was performed on a Rotor-Gene 2000 (Corbett Research) using standard 25 μl HotStar Taq conditions (Qiagen) on cDNA equivalent to 100–1,500 ng total RNA. 0.35X final SYBR Green I or 200 nM probe was used for detection. Amplifications were at 95°C for 10 min, 45 cycles at 94°C for 20 s, annealing temperature for 30 s and 72°C for 30 s. Primer sequences and annealing temperatures are shown ( Additional file 7 ). Unknown expression levels were extrapolated from standard curve dilutions of column purified cDNA amplicons. Replicate standard curve assays (n ≥ 2) were used with C T coefficient variations averaging < 15% over six orders within replicates and between dilutions. mRNA mlcl numbers were quantitated from samples at the parameter C T from standard curves with known mlcls/μl calculated from the dilution mass in μg/μl ÷ M.W. of ssRNA transcript (× 6.02 × 10 17 mlcls/μmole). mRNA mlcls per cell were calculated at the C T concentration ÷ amount of total RNA (ρg) per reaction multiplied by 4 based on a 4 ρg total RNA per cell estimation. mRNA was shown as raw expression data (n > 4) and or normalized against CYPA or ATP5A. Several other housekeeping and cancer-related gene expression levels were quantitated to ensure the mRNA mlcl per cell estimations were compatible with other methods (data not shown). Sodium bisulfite modification 10–20 μg of genomic DNA was digested overnight at 37°C with restriction enzymes XbaI, XhoI, HindIII and EcoRI and cleaned using nucleotide purification columns (QIAvac 24, Qiagen). The digested DNA was pooled and bisulfite modified [ 17 ]. 10 μg of DNA was diluted in 500 μl of water and denatured with 55 μl of 2 M NaOH for 20 min at 37°C. DNA was mixed with 300 μl of 10 mM hydroquinone (Sigma), 5.2 ml of 3.6 M NaHSO 3 (pH 5.0) (Sigma), overlaid with paraffin oil and deaminated in the dark for 16 h at 55°C. DNA was desalted using Qiagen purification columns, eluted in 500 μl water and desulfonated with 55 μl of 3 M NaOH for 20 min at 37°C. DNA was neutralized and precipitated with 800 μl of 10 M ammonium acetate (pH 7.0), 20 μl linear acrylamide and 5 ml cold 100% EtOH. Modified DNA was pelleted, resuspended in 40 μl 1 mM Tris-Cl (pH 8.0) and the concentrate stored at -80°C. Methylation-specific PCR (MSP) and sodium bisulfite sequencing Four 100–200 bp regions spanning approximately 1-kb of the CBFA2T3B promoter region were amplified from ≥ 300 ng of bisulfite modified DNA under standard HotStar Taq conditions using primers specific for either unmethylated or methylated cytosines. Hypermethylated breast tumor samples were used for bisulfite sequencing to generate a CBFA2T3B promoter region methylation map. Four 300–400 bp regions spanning the promoter were amplified under standard conditions as described above using primers simultaneous for both unmethylated and methylated cytosines. The amplicons were sub-cloned into pGEM (Promega) and 5 to 10 clones sequenced using BigDye (Applied Biosystems). Primer sequences and annealing temperatures are shown ( Additional file 7 ). Demethylation assay MDA-MB-231 was treated with the demethylating agent 5-Aza-dC (Sigma). Approximately 2.0 × 10 5 cells per T75 flask were seeded in 10 ml RPMI-1640 supplemented with 10% FCS, 15 mM HEPES, 10 mg/liter PGS and cultured for 48 h at 37°C with 5% CO 2 . Cells were treated with 50 μm 5-Aza-dC for 120 h and replenished with fresh medium and 5-Aza-dC every 24 h. These concentrations are not inhibitory to cell growth [ 48 ]. Concentrations ranging from 1–5 μM 5-Aza-dC had no affect on expression levels (data not shown). Control untreated cells were cultured in parallel and supplemented with DMSO. Replicate T75 flasks for both treatments and controls were performed (n = 4). Total RNA was isolated and analyzed for CBFA2T3 isoform and SYK expression levels using real-time RT-PCR. MDA-MB-231 cells were also treated with the HDAC inhibiting agents trichostatin A (Sigma), sodium butyrate (Sigma) and apicidin (Calbiochem). Only trichostatin A elicited re-expression levels similar to 5-Aza-dC (data not shown). Real-time methylation-specific PCR (MSP) A sequencing amplicon was initially tested for bisulfite PCR amplification and cloning bias by scoring percent methylation frequencies of overlapping amplicons. Bisulfite PCR bias was also tested by amplification on proportional mixtures of hypo and hypermethylated bisulfite treated DNA. This amplicon was PCR amplified and column purified from all samples. Second-round real-time MSP was performed on the amplicon using internal forward primers to detect for either unmethylated or methylated cytosines at the Sp1 site displaying high-frequency cytosine methylation. The methylated and unmethylated primer specificities were tested by real-time amplification on serial dilutions of unmethylated and methylated clones, respectively. Second-round amplicons were also sequenced to ensure specificity. Second real-time MSP was performed on a Rotor-Gene 2000 using internal forward primers under standard 25 μl HotStar Taq conditions with approximately 100 ρg of first-round cleaned amplicon and 0.35X SYBR Green I. Amplifications were at 95°C for 10 min, 45 cycles at 94°C for 20 s, annealing temperature for 30 s and 72°C for 30 s. Unknown unmethylated and methylated cytosine levels were extrapolated at the Sp1 site from standard curve dilutions of internally PCR amplified and column purified cloned amplicons originally sequenced and found to be representative of either unmethylated (e.g. BT-483) or methylated (e.g. MDA-MB-231) sequence. Unmethylated and methylated mlcl numbers were quantitated from all samples at the parameter C T from standard curves with known mlcls/μl calculated from the dilution mass in μg/μl ÷ M.W. of dsDNA clone sequence (× 6.02 × 10 17 mlcls/μmole). CBFA2T3B promoter methylation levels were expressed as methylation indices [mi = m/(m + u)] and ratios (u/m) [ 49 , 50 ]. Statistical analysis CBFA2T3B promoter methylation index medians and interquartile ranges were determined for each group of tissue samples. Statistical comparisons between groups were performed using the two-way Levene's test for Equality of Variances, i.e. H0: median methylation index variances are the same in each group; HA: median methylation index variances are not the same in each group. Variances in the median methylation indices between each group were considered statistically significant when P ≤ .05. Correlations between CBFA2T3B promoter methylation and gene expression levels were determined by calculating a Spearman's rank coefficient. All statistical analyses were performed using GraphPad Prism Version 4.0. Authors' contributions AJB completed the work and manuscript. AEG aided in bisulfite sequencing. OLDM performed promoter analysis. DFC, GRS and GK supervised the work. All authors read and approved the final manuscript. Supplementary Material Additional File 1 CBFA2T3 gene expression levels assayed using real-time RT-PCR The raw data CBFA2T3B expression levels and preliminary CBFA2T3A expression levels in breast tumor cell lines are shown. The y-axis represents the fluorescence detection scale and the x-axis represents the C T of amplification. The CBFA2T3 gene expresses at endogenously low levels and requires at least 500 ng of reverse transcribed total RNA to cDNA template for reproducible detection. In contrast, the housekeeping genes such as CYPA require only 100 ng of template. When 500 ng is used, the CYPA expression levels are off the fluorescence scale. Note the C T values are below 35 cycles for the down-regulated cell lines such as MDA-MB-231. This low-level of expression is undetectable by conventional RT-PCR. Moreover, because of this low endogenous expression the CBFA2T3 mRNA could not be reliably detected using Northern Blots or RNase protection (pdf file). Click here for file Additional File 2 CBFA2T3B promoter methylation levels examined using MSP The methylation levels in 24 breast tumor cell lines (labeled), 20 primary breast tumors (1T-20T), 20 normal breast counterparts (1N-20N) and 24 normal whole blood samples (1–24) are shown. Four separate regions (1–4) spanning 1-kb of CBFA2T3B promoter sequence were amplified using primers to detect for either unmethylated or methylated cytosines (see Figure 1B for primer locations and Additional file 7 for primer sequences and annealing temperatures). The asterisks indicate the samples examined by bisulfite sequencing. Unmethylated (U) and methylated (M) band intensities were scored as high (black dot), low (gray dot) or negative (white dot). Low-level basal methylation defined by high unmethylated to low methylated band intensities was detected in all normal bloods in ≥ 2/4 regions. High methylation was also detected in 21% of bloods in region four. Samples were considered hypermethylated when high intensity bands amplified in all four regions or hypomethylated when no methylation was detected. Based on this, 21% of breast tumor cell lines were hypermethylated (e.g. MDA-MB-231 and MDA-MB-468), 16% hypomethylated (e.g. BT-483 and MDA-MB-361) and 63% displayed a combination of basal to high methylation. Primary breast tumor samples were complex. 57% displayed basal methylation in ≥ 2/4 regions, 14% displayed high methylation in ≥ 2/4 regions and 29% tested negative in ≥ 3/4 regions. Normal breast counterpart samples were similar with 62% displaying basal methylation in ≥ 2/4 regions, 14% displaying high methylation in ≥ 2/4 regions and 24% testing negative in ≥ 3/4 regions. The detection of high-methylation in MCF-7 is due to spurious amplification events. Several overlapping primers spanning this region in this cell line were methylation negative (data not shown). This cell line is hypomethylated according to real-time MSP (pdf file). Click here for file Additional File 3 CBFA2T3B promoter methylation levels assayed using second-round real-time MSP CBFA2T3B promoter methylation levels were assayed using second-round real-time MSP. The raw data methylation levels in normal whole blood samples and breast tumor cell lines are shown. The y-axis represents the fluorescence detection scale and the x-axis represents the C T of amplification. Second round real-time MSP was performed on a bisulfite sequencing amplicon using internal forward primers to detect for either unmethylated (uF) or methylated (mF) cytosines at the Sp1 CpG sites shown in Figure 1C. The C T for methylated amplification in breast tumor cell lines was highly aberrant compared to normal blood samples. Note the late unmethylated C T obtained in MDA-MB-231 compared the early hypermethylated C T . In contrast, BT-483 shows an early unmethylated and late hypomethylated C T (pdf file). Click here for file Additional File 4 CBFA2T3B promoter methylation melt curve analysis CBFA2T3B promoter methylation melt curves were examined following second-round real-time MSP. Raw data melt curves of second round amplicons in normal blood samples and breast tumor cell lines are shown. Curves were calculated from the negative derivative in fluorescence over temperature versus temperature (-dF/dT m versus T m ). Normal blood samples displayed consistent peak levels for both the unmethylated and methylated mlcls. In contrast, breast tumor cell lines displayed highly aberrant peak levels depicted by broad melt transitions and heterogeneous melt curves reflective of the aberrant concentration and composition of 5-methylcytosines (pdf file). Click here for file Additional File 5 CBFA2T3B promoter methylation levels versus gene expression Methylation indices were calculated as methylated promoter mlcls per 10 4 cells for breast tumor cell lines with pre-determined 16q24.3 DNA mlcls per cell and plotted against CBFA2T3 mRNA mlcls per 10 4 cells. The y-axis represents methylation levels assayed using real-time MSP and the x-axis represents expression levels assayed using real-time RT-PCR. Both data sets are shown on a log scale. Each black circle represents a different breast tumor cell line. The asterisk marks the median methylation and median gene expression levels. The median methylation index of 0.02 (i.e. 2 mlcls or alleles methylated in 100) is calculated from the median methylation level of 450 methylated alleles per 10 4 cells divided by the number of unmethylated 'active' alleles in the 10 4 cells or 20,000 alleles, i.e. [450 ÷ (20,000 - 450) = 0.02]. The median gene expression index of 0.2 (i.e. 20 mRNA mlcls expressed in 98 'active' alleles) is calculated from the median expression level of 4,500 mRNA mlcls per 10 4 cells divided by the number of unmethylated 'active' mlcls in the 10 4 cells, i.e. (4,500 ÷ 19,550 = 0.2). This calculation equates to approximately 4–5 mRNA mlcls expressed per 10 cells and suggests that the CBFA2T3B gene is largely transcriptionally inert. The remaining active alleles may be trans-factor dependent for expression. An inverse correlation between promoter methylation and expression levels per population of cells was established (r 2 = 0.77; r = -0.9, P < .0001). In hypermethylated MDA-MB-231, approximately 17,000 promoter mlcls were methylated (i.e. mi = 0.85 as 17,000 in 20,000 are methylated) and 4 mRNA mlcls expressed per 10 4 cells. In hypomethylated BT-483, approximately 5 promoter mlcls were methylated (mi = 0.0002) and 120,000 (± 40,000) mRNA mlcls expressed per 10 4 cells. This elevated expression equates that 12 (± 4) mRNA mlcls per cell are expressed from an estimated four-promoter mlcls per cell (i.e. four-16q24.3 DNA mlcls per cell) or that one 'active' unmethylated CBFA2T3B promoter mlcl per cell transcribes 2–4 mRNA mlcls. From a methylation index of approximately 2% and greater, a large increase in y may lead to a small decrease in x. Under this condition, the relationship may be asymptotic. The regression equation y = 25,801 x -0.58 (r 2 = 0.7669) describes the methylation and expression relationship. When plotted as methylation index values the equation was y = 1.05 x -0.61 (r 2 = 0.7665) (pdf file). Click here for file Additional File 6 CBFA2T3B promoter methylation levels assayed using second-round real-time MSP The methylation levels in normal whole blood samples, normal breast counterparts, primary breast tumors and breast tumor cell lines were assayed at the Sp1 site shown in Figure 1C and plotted as absolute methylation ratios (u/m). All samples are labeled corresponding in part to Additional file 2. The asterisks indicate the samples examined by bisulfite sequencing. The gray highlights indicate the normal blood and normal breast counterpart 'full' methylation ranges. Basal methylation ratios in normal bloods averaged 60:1 (cumulative mean) unmethylated to methylated CBFA2T3B promoter mlcls. This average coincided with conventional MSP band intensities of 100:1 based on pUC19 DNA/MspI marker concentrations (see Figure 3) and the median methylation index of 0.02 (i.e. 2 mlcls in 100 are methylated). Normal blood ratios ranged 20:1 to 160:1 unmethylated to methylated mlcls. Normal breast counterparts were similar to normal bloods averaging 100:1 but with a larger range of 10:1 to 350:1. Relative to the normal samples, breast tumors displayed highly aberrant methylation ratios clearly resolved by second-round real-time MSP. 75% of breast tumor cell lines were aberrantly methylated outside the full range of normal blood basal methylation. 58% were outside the range of normal breast. Half of the aberrations were either hypo or hypermethylated relative to both normal blood and normal breast. Similar to cell lines, 51% of primary breast tumors were aberrantly methylated relative to normal blood. 35% were aberrant relative to normal breast (i.e. 24% were hypermethylated and 11% hypomethylated). Aberrant methylation ratios ranged from 1:10 in hypermethylated MDA-MB-231 to 7,000:1 in hypomethylated BT-483 (pdf file). Click here for file Additional File 7 Primers, probes and annealing temperatures used (pdf file). Click here for file	/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC516017.xml
535816	Elevated responses to constant facial emotions in different faces in the human amygdala: an fMRI study of facial identity and expression	Background Human faces provide important signals in social interactions by inferring two main types of information, individual identity and emotional expression. The ability to readily assess both, the variability and consistency among emotional expressions in different individuals, is central to one's own interpretation of the imminent environment. A factorial design was used to systematically test the interaction of either constant or variable emotional expressions with constant or variable facial identities in areas involved in face processing using functional magnetic resonance imaging. Results Previous studies suggest a predominant role of the amygdala in the assessment of emotional variability. Here we extend this view by showing that this structure activated to faces with changing identities that display constant emotional expressions. Within this condition, amygdala activation was dependent on the type and intensity of displayed emotion, with significant responses to fearful expressions and, to a lesser extent so to neutral and happy expressions. In contrast, the lateral fusiform gyrus showed a binary pattern of increased activation to changing stimulus features while it was also differentially responsive to the intensity of displayed emotion when processing different facial identities. Conclusions These results suggest that the amygdala might serve to detect constant facial emotions in different individuals, complementing its established role for detecting emotional variability.	Background Facial expressions and facial identities are important cues for the evaluation of social contexts [ 1 , 2 ]. Two main types of information have to be processed while seeing other persons' faces: a face has to be identified as belonging to a unique individual, establishing facial identity, while facial expressions have to be interpreted for emotional context, which is crucial for the social interaction [ 3 ]. Facial expression itself conveys two levels of information: first, facial emotions of others signal information about the emotional state and the benevolence or hostility of that person towards oneself; second, they convey information about the person's evaluation of the environment. The same emotional expression on several other persons' faces often signals a high degree of consistency in their evaluation of the current environment; they are therefore a particularly valid cue for one's own situational appraisal of this environment. In particular, the existence of specialized brain systems for the perception of fear expressions as a form of threat related to physical attack [ 4 ], points to the importance the brain attaches to social signals of potential environmental threats [ 5 , 6 ]. Visual analysis of human faces has been suggested to be achieved by a core system comprising the fusiform gyrus together with the inferior occipital gyrus, the superior temporal sulcus, and the amygdala [ 7 , 8 ]. The amygdala in particular has been shown to play a pivotal role in the processing and recognition of emotional facial expression, especially fear [ 9 , 10 ]. Since early functional imaging studies [ 11 , 12 ] showed amygdala activation to fearful face stimuli, a variety of imaging studies have led to a more detailed understanding of amygdala function in facial expression. Temporal properties of amygdala responses to stimulus repetition (i.e. habituation, e.g. [ 11 , 13 - 15 ]), the influence of attentional state [ 16 ] and awareness (e.g. [ 17 - 19 ]), as well as the amygdala's response to different types of emotional expression (most recently e.g. [ 20 ]) have been explored. Despite many studies investigating facial affect, only few have addressed the interaction of facial affect and identity. Preceding experiments on face processing were not explicitly designed to investigate differences in processing of facial identity and facial expressions, rather they examined the effects of changes in appearance, viewing angle, selective features and physical stimulus properties on face processing (e.g. [ 14 , 21 , 22 ]). Recent advances to assess the effects of facial identity and emotional expression include one report on increased left amygdala activation to blocks of multiple novel vs. single identical faces displaying neutral or emotionless expressions [ 23 ]. Animated emotional expressions or identities, morphing either from neutral to emotional expression or from one neutral face to another, have been shown to elicit stronger activations in bilateral amygdala than static displays of these same stimuli [ 24 ]. Notably, these studies were not designed to investigate the interaction of identity and emotional expression as identity was only varied within the neutral expression condition. Most recently, Winston and colleagues have employed a factorial design to explore fMRI-adaptation to repeated presentations of two facial expressions in the context of an event-related fMRI experiment [ 25 ]. In contrast to earlier findings (e.g. [ 11 , 12 ]), the authors did not observe any significant signal changes in the amygdala in the context of their factorial design, suggesting that sparse stimulus presentations interleaved with a checkerboard might not be potent enough to evoke a task-related fMRI activation in this region. Variability and constancy of identity and emotional expression are concepts that require stimulus integration over time, thus a sufficient number of stimulus presentations might be critical for a brain region such as the amygdala in order to detect stability or changes in the stimulus sequence. Therefore, we tested this interaction in a 2 × 2 factorial design (see Figure 1 ), with either constant or variable facial identity (factor 1), and either constant or variable emotional expression (factor 2), in a block design fMRI study. The conditions used were: (a) constant identity, constant expression (C I C E ), (b) variable identity, constant expression (V I C E ), (c) constant identity, variable expression (C I V E ), (d) variable identity, variable expression (V I V E ). To control for attentional effects during the procedure, an oddball task was included to avoid confounds by an emotional judgment or a gender differentiation task (the latter being a property of facial identity, [ 26 ]). Given that numerous reports demonstrate amygdala activation in the processing of sequential facial expressions [ 11 , 12 , 27 - 29 ] and given the importance of different facial identities as a valid cue for making social inferences (see above), we hypothesized that constant emotional expressions displayed in different faces would elicit strong responses in the amygdala despite its well-described habituation to repeated stimuli [ 13 , 15 , 23 ]. Furthermore, we expected that constant fearful expressions (shown in different identities) would elicit the strongest amygdala activation given the evidence from lesion and imaging studies that this expression is a particularly potent activator of the amygdala [ 11 , 12 , 17 , 30 , 31 ]. Because we expected to find a stimulus repetition effect in the fusiform gyrus [ 14 , 32 ] we systematically examined responses to changing versus constant stimulus features, i.e., facial expression and identity in this area. Results Analysis of the behavioral data showed that subjects maintained a high degree of task attention (98.03 % hits to oddball stimuli (facial stimuli at reduced luminance)). There was a trend of a main effect of identity in the reaction times (RTs) to the oddball stimuli (F 1,10 = 4.859, p < 0.06), with longer RTs in the conditions of changing identities, but no significant effects of emotion or an interaction (emotion: F 1,10 = 0.755, p > 0.40; interaction: F 1,10 = 1.656, p > 0.20). For the analysis of the imaging data we chose a statistical threshold of p < 0.05, corrected for a reduced search volume of interest for regions with a priori hypotheses (amygdala, fusiform gyrus). For other brain areas the threshold was set to a threshold of p < 0.05, corrected for the entire brain. Post-hoc contrasts of specific experimental conditions for detailed characterization of the experimental effects were carried out at an uncorrected significance threshold. In the amygdala we expected to find higher responses to constant emotional expressions in different facial identities. Thus in our 2 × 2 factorial design, this can be formally tested by an interaction. Data for all experimental conditions in the peak activation voxels found in our masked interaction contrast (thresholded at p < 0.01, for the purpose of visualization) are shown in the top left and right panels of Figure 2 with the corresponding significant voxels in the SPMs (arrows in middle panels). Applying a reduced search volume corresponding to an anatomical mask of the amygdala (see Methods) to these activations we found a highly significant peak in the right and another peak approaching significance in the left amygdala. We report the latter peak in light of bilateral amygdala activation in similar tasks shown previously [ 11 , 14 ]. Locations and statistics for this analysis for each peak are displayed in Table 1 . Post-hoc contrasts of the experimental conditions revealed that amygdala responses to V I C E were only significantly more active than responses to C I C E (left amygdala (-15 0 -15): T = 3.75, p < 10 -4 ; right amygdala (18 0 -18): T = 4.76, p < 10 -4 ), and also more active than the other two experimental conditions (C I V E and V I V E ): left amygdala (-15 0 -15): V I C E vs. C I V E : T = 1.77, p < 0.05; V I C E vs. V I V E : T = 2.40, p < 0.01; right amygdala (18 0 -18): V I C E vs. C I V E : T = 2.19, p < 0.01; V I C E vs. V I V E : T = 3.30, p < 0.001. In a further post-hoc analysis of the response to V I C E we decomposed this condition with respect to different emotional expressions. This analysis revealed that the V I C E activation is primarily caused by the response to maximally fearful expressions and (to a lesser degree) to other emotional expressions, but only marginal to neutral expressions (Figure 2 , bottom graphs). Post-hoc pair-wise comparisons between maximally fearful and the other emotional expressions exceeded statistical thresholding (left amygdala (-15 0 -15): Fear Max vs. Fear Min: T = 1.95, p < 0.05, Fear Max vs. Neutral: T = 2.45, p < 0.01, Fear Max vs. Happy Min: T = 2.53, p < 0.01, Fear Max vs. Happy Max: T = 1.91, p < 0.05; right amygdala (18 0 -18): Fear Max vs. Neutral: T = 1.85, p < 0.05, Fear Max vs. Happy Max: T = 2.17, p < 0.05). In order to characterize the time-course of the amygdala responses to the emotion-specific responses in condition V I C E we calculated the fitted responses for the most intense and the neutral expressions by multiplying the parameter estimates (regression coefficients) for the respective boxcar and exponential decay regressor with the canonical response function created by SPM during design specification. Figure 3 shows the time course for each emotion-specific response for each peak voxel in both amygdalae. While the overall height of the fitted responses parallel the parameter estimates of the boxcar regressors in Figure 3 , we found a trend toward within-block habituation to fearful expressions, and an increase in the fitted responses for happy expressions (especially in the right amygdala) while neutral expression maintained the level of activation across the entire block. Based on previous work [ 22 , 32 ] we expected activation in the fusiform gyrus to be dependent on stimulus changes irrespective of whether emotional expression, facial identity, or both were varied. We also expected to detect an influence of emotion type and intensity on fusiform activity [ 12 , 22 ]. Accordingly, we compared conditions with at least one changing stimulus feature (identity, emotion) to the condition in which the same picture was repeated for the entire block. Here we found evidence for an effect of changing stimulus features. Figure 4 depicts significantly activated voxels in the lateral fusiform gyrus for this contrast (thresholded at p < 0.001 for the purpose of visualization), with arrows pointing to the voxels of peak activation. Location and statistics corrected for a reduced search volume for this region reported by Vuilleumier and colleagues [ 33 ]; see Methods) are shown in Table 1 . These locations correspond well to previously reported regions linked to human face processing [ 34 ]. Pair-wise post-hoc contrasts of the four experimental conditions revealed that the lateral fusiform gyrus was significantly more activated by all experimental conditions with at least one changing stimulus feature (V I C E , C I V E , V I V E ) than by the condition with repeated presentation of the same picture (C I C E ): left lateral fusiform gyrus (-39 -54 -24): V I C E vs. C I C E : T = 2.75 p < 0.01, C I V E vs. C I C E : T = 3.09, p < 0.01, V I V E vs. C I C E : T = 3.77, p < 0.001; right lateral fusiform gyrus (36 -54 -24): V I C E vs. C I C E : T = 2.98, p < 0.01, C I V E vs. C I C E : T = 2.79, p < 0.01, V I V E vs. C I C E : T = 3.01, p < 0.01. A further post-hoc analysis of the condition V I C E revealed an effect of emotion intensity with the maximal intensity expressions eliciting stronger activations than those with reduced intensities and the neutral expression. Pair-wise contrasts of the emotion-specific V I C E condition revealed that maximally fearful expressions elicited significantly larger activation than neutral and emotional expressions of minimal intensity (left lateral fusiform gyrus (-39 -54 -24): Fear Max vs. Fear Min: T = 2.6, p < 0.01; Fear Max vs. Neutral: T = 2.79, p < 0.001; Fear Max vs. Happy Min: T = 2.38, p < 0.01; right lateral fusiform gyrus (36 -54 -24): Fear Max vs. Fear Min: T = 1.72, p < 0.05; Fear Max vs. Neutral: T = 1.94, p < 0.05; Fear Max vs. Happy Min: T = 2.27, p < 0.01; indicated by asterisks () in the bottom panels of Figure 4 ). However, only in the left lateral fusiform gyrus the same post-hoc tests were significant when contrasting maximally happy expressions with all others (Happy Max vs. Happy Min: T = 1.84, p < 0.05; Happy Max vs. Neutral: T = 2.54, p < 0.01; Happy Max vs. Fear Min: T = 1.99, p < 0.05, indicated by pluses (+) in the bottom panels of Figure 4 ). Discussion We investigated the interaction of facial identity and expression in 2 × 2 factorial blocked fMRI study which uniquely enabled us to compare all combinations of facial identity and expression within the same experiment, therefore the results of the present study complement and extend the results of previous studies on facial identity and expression processing in the human brain. In support for our hypotheses we found distinct response patterns in bilateral amygdala and bilateral lateral fusiform gyrus. While the lateral fusiform gyri can be characterized by a binary response pattern corresponding to an effect of changing stimulus feature, the amygdala responded maximally to constant emotional facial expressions in combination with changing facial identity. In addition, in the V I C E experimental condition we found an effect of emotion intensity in the fusiform gyrus, with stronger activations to maximally intense expression irrespective of valence when compared with neutral or modestly intense expressions. While we found a (non-significant) trend for the same modulation of the V I C E activation by emotion intensity in the left amygdala the general pattern when including the right amygdala rather conforms to a specific sensitivity for maximally fearful expression which elicited the strongest activation. Fusiform gyrus activations The binary response pattern conforming to sensitivity for changes in stimulus features in the lateral fusiform gyrus might be explained by reference to the proposed network of face processing in humans [ 7 ]. Haxby and colleagues hypothesize that this part of the distributed network is especially sensitive to facial stimulus configurations (and the variability of these configurations). Evidence supporting this conclusion has been reported in a study by Vuilleumier and colleagues who demonstrated a significant repetition suppression effect in the fusiform gyrus in response to the second compared to the first presentation of a given facial identity [ 22 ]. Similarly, Rotshtein and colleagues demonstrated stronger activation in face related voxels in the lateral occipital complex (LOC) for blocks with different identities (resembling the condition V I C E of the present study) compared to blocks of the repetition of the same identity (our condition C I C E ) [ 14 ]. Another study investigating the influence of facial expression and identity on fusiform activation found reduced activation to facial identity but not to expression and no interaction of these two factors [ 25 ]. Our data support and extend these findings with respect to identity in the fusiform gyrus, as we found evidence for an effect of changing stimulus features in the lateral fusiform gyrus irrespective of the dimension of variation (facial identity, emotional expression collapsed over all emotions, or both). Additionally, we found an effect of emotion intensity in the fusiform gyrus activity in condition V I C E with higher activation to maximally fearful and happy expressions relative to with neutral expressions, paralleling earlier findings which report stronger activation to fearful than to neutral expressions [ 22 , 24 , 27 , 33 , 35 ]. Rotshtein and colleagues show significantly larger activation for aversive versus happy expressions in repeated presentations (C I C E ), suggesting an identity repetition × emotion interaction [ 14 ], while others fail to find an effect of negative (or positive) expression in the fusiform gyrus [ 25 ]. This apparent negativity bias might be a confound of stimulus selection as researchers use negative material more frequently than positive stimuli. In fact, the underlying mechanism might be an enhanced attentional processing of arousing facial expressions [ 36 ], which are usually also the most negative faces. Our finding of an effect of emotion intensity supports this notion more convincingly as we also show increased activation in the fusiform gyrus to happy expressions. An underlying arousal dimension that exerts a modulatory effect on the activation in this region might also explain the lack of an effect of expression in the findings of Winston and colleagues [ 25 ], as they did not include a neutral (low arousal) expression in their stimulus set. Two recent studies suggest that the low frequency information of facial stimuli might drive the modulatory effect of arousal, which is thought to be a feedback influence of the amygdala, which also displays a preference for low frequency information [ 22 , 35 ]. Amygdala activations Our 2 × 2 factorial blocked fMRI design allowed us to compare between the possible conditions of constant and variable facial identity and expression. We show for the first time a maximum of activation of the amygdala to variable facial identities displaying the same emotion (V I C E ) compared to all three other main conditions (C I C E , C I V E and V I V E ). Furthermore, within this condition (V I C E ) maximally fearful expressions elicited the strongest amygdala activity. The result of the present study replicates and extends, in part, the earlier findings that also utilized blocked presentations of facial expression with stimulus configurations resembling one or more of our experimental conditions (V I C E blocks: [ 11 , 12 , 28 , 37 , 38 ]; C I C E and V I C E blocks: [ 14 ]; V I C E and V I V E blocks: [ 39 ]). In light of our findings, the significant amygdala activation to fearful expression in the aforementioned studies might have been obtained because they employed a stimulus presentation conforming to our condition V I C E . Interestingly, in a study comparing dynamically morphed and static presentations of facial stimuli, LaBar and colleagues [ 24 ] also report amygdala activation when neutral faces are morphed into each other (a dynamic version of our condition V I C E ) as compared to static and repeated presentation of the same facial stimuli (the condition C I C E in the present study). A closer inspection of our findings in Figure 2 reveals a similar trend, as the neutral expression in the V I C E condition still elicited a larger signal change than the condition C I C E . Thus, our comprehensive factorial design can relate these earlier findings in the broader context of constancy and variability of identity and expression. Many imaging and lesion studies have documented activations or behavioral impairments following the presentation of fear-related stimuli (for review see [ 9 ]). Our decomposition of the amygdala activation in condition V I C E supports this claim. However, recent studies, which used different sensory modalities and carefully controlled for the often confounded dimensions of valence and intensity, argued for an effect of emotion intensity (arousal) in the amygdala [ 40 - 42 ]. Further support for this interpretation comes from a comprehensive study investigating the effects of different expressions of basic emotions during direct and incidental stimulus processing [ 20 ]. The authors found no specific effect for a particular emotional expression; they rather report amygdala activations when comparing high vs. low intensity exemplars of the facial stimuli. Although we also find trends for an effect of emotion intensity (Figure 2 , bottom left panel), the question of specific amygdalar fear sensitivity or a more general arousal sensitivity remains equivocal based on our findings. Interestingly, Winston and colleagues [ 25 ], employing a similar design as in the present study, failed to observe significant signal changes in the amygdala. The divergent findings might be explained by the difference between our blocked and their event-related design, in which they sequentially presented pairs of facial stimuli. Sparse stimulus presentations in that study might not be sufficient to trigger activation in the amygdala, as this structure might decode a more sustained stimulus train for assessing the biological relevance of the current situation (see below). Furthermore, in their study, the pairs of faces presented were separated by an face-outlined checkerboard [ 25 ] which might have prevented the induction of an emotional state that would be encoded by the amygdala. The temporal properties of amygdala responses to facial expressions are crucial. Many studies show habituation of the amygdala response to blocks of directly adjacent fearful vs. neutral or happy expressions with constant identities over the course of 30 to 80 sec [ 13 , 15 , 38 ]. This is comparable to the C I C E condition in the present study in terms of presented stimuli but not with respect to block length and block order. Interestingly, when the C I C E condition in the present study is analyzed for the comparison of maximum fear vs. neutral blocks a non-significant trend for left ventro-lateral amygdala activity (data not shown) is seen, indirectly supporting those findings. Other studies used fearful and neutral V I C E conditions showing within-block and across-block habituation with fixed alternating block order [ 11 ] or in the comparison of V I C E vs. C I C E blocks [ 14 , 23 ] in designs with pseudo-randomized block order. Likewise, we also show a highly significant difference between the V I C E and C I C E conditions in the amygdala. The response profile of the amygdala for maximal fearful expressions in the V I C E condition trends toward a within-block habituation effect (Figure 3 ). In contrast, the response profile for neutral expressions exhibits a sustained response during the entire block albeit at a much lower overall level. This aspect parallels findings by Breiter and colleagues [ 11 ], although direct comparison is limited because that study showed habituation on a larger timescale (between blocks). The within-block increase of amygdala activation to happy expressions, however, suggests that the temporal nature of signal changes in the amygdala might be more complex. This delayed onset of activation to happy faces is in accordance with an evolutionary interpretation of our findings. The amygdala is located in a critical position on the efferent pathway that is involved in the preparation of autonomic responses to threatening situations [ 43 ]. Because happy expressions are usually a valid signal for non-threatening situations that do not require fight or flight responses, amygdala activation is not needed at an early stage. The explanation also holds for the early but overall attenuated response to neutral expressions that are inherently ambiguous, which thus prompt for sustained perceptual processing and, potentially, for preparation of defensive behavior [ 44 ]. Additionally, Wright and colleagues [ 23 ] showed greater amygdala activation to neutral V I C E blocks compared with neutral C I C E blocks. Clearly, further research is needed to characterize the temporal evolution of amygdala activation during this type of blocked stimulus presentation in more detail. One might argue that the responses in the conditions with variable emotional expression (C I V E , V I V E ) were reduced because within those blocks variable emotional and neutral expressions were intermixed, possibly leading to smaller signal changes within those blocks. However, subjects were presented with the same number of stimuli of each emotional expression in every experimental condition. Therefore, differences cannot be attributed to a varying number of emotional expressions seen in different conditions. Thus, if the conditions with variable emotions elicit a smaller signal change because of the neutral faces within them, this effect should also apply to the blocks of constant neutral expressions in different faces (V I C E condition). This effect is in fact shown in the lower bar graphs in Figure 2 , especially in comparison to the other emotional V I C E conditions. Because we found an overall elevated response in the amygdala to the condition with constant emotions in changing identities (V I C E ) compared to conditions with variable emotions (C I V E , V I V E ; see Figure 2 upper bar graphs), we argue that the observed differences between our experimental conditions represent a true effect of the sequential stimulus configuration within this experimental design. Although we can not ultimately exclude the possibility of confounding order effects within the categories of variable emotion conditions (C I V E , V I V E ), this does not diminish the main point of this study, namely that the human amygdala is most responsive to the sequential presentation of faces with constant emotion (fearful) and varying identity. To further confirm and generalize this finding future studies are needed using additional emotions (e.g. anger, sadness, etc.). Novelty effects of changing facial identity and stimulus order of the different conditions might also be claimed as an explanation for the strong responses to the condition V I C E [ 23 ]. But in contrast to the latter study, subjects in our study were familiarized with the stimuli before the experiment and the same facial identities were presented repeatedly throughout the experiment, thus minimizing stimulus novelty. Furthermore, if novelty detection were the main factor driving the amygdala response, the condition V I V E should have also elicited a strong amygdala response which it did not (see Figure 2 ). There was also no systematic sequence of blocks of constant facial identities followed by variable facial identities in the present study which has been shown to relatively increase the activation to the multiple identity condition presented secondly [ 23 ]. It is interesting to note that the peak activation within the amygdala is located in the medio-dorsal part of this structure. Previous imaging studies of the processing of facial affect have predominantly reported their activations in the dorsal amygdala (for review see [ 10 ], Figure 3 ). The dorsal amygdala has also been associated with the representation of ambiguous stimuli, such as fearful expressions that do not signal a potential threat directly [ 10 , 29 , 44 ]. However, given the resolution and post scan smoothing of the functional images in this and other functional imaging studies, the localization of amygdala activity should be discussed with caution. The potential biological relevance of constant facial emotion Sensitivity for constant facial emotions in different identities can be seen as a process that integrates stimuli with respect to the conveyed emotion over a certain amount of time. Thus, environmental stimuli (such as facial expressions) are compared to each other to detect changes to stability in the emotion, a concept that can be described as emotion constancy . In this context, constant facial emotions in different individuals signal a high degree of consistency in others' appraisals of the environment, and constitute a more valid cue for one's own appraisal. Furthermore, facial expressions differ in their degree of saliency which often reflects the biological relevance of the stimulus that evoked the expression [ 18 , 44 ]. For example, a fearful face as a reaction to a threatening stimulus is more salient and calls more immediately for appropriate action than a neutral facial expression. We found elevated response in the amygdala especially to these salient (fearful) expressions because these constant salient facial emotions call for an immediate situational appraisal and subsequent action. Our findings gain support and extend the findings of earlier imaging studies on the processing of human facial emotions which employed an experimental design similar to our condition V I C E and found significant activation in the left dorsal amygdala with constant fearful expressions [ 11 , 12 ]. Taken together, these results, as well as our own findings in this experiment, suggest that the same emotional expressions displayed in many different faces are potent stimuli that activate the amygdala and might serve the detection of emotion constancy. Conceptually, conditions with variable emotionality (C I V E , V I V E ) can be seen as noise in the context of the detection of constant emotions among others and thus, it is not surprising that the amygdala shows less signal change in these conditions. The perception of emotional facial expressions in others yields insights into their evaluations of the environment and guides one's own emotional and behavioral reactions. Encountering the same emotional expression in many different people (emotion constancy) is an especially valid and readily available cue for making subsequent inferences about the potential harmfulness of an environmental situation. These inferences would have direct implications for evolutionary survival and must have been a central feature of human ancestral cognitive abilities [ 45 ]. They also remain essential for safe locomotion in our current complex environment as they reduce the time spent in a potentially threatening situation, because the perceiving subject can avoid energy-intensive and time-consuming search for potential threats in the environment [ 46 ]. For example, encountering the same fearful expression in several different people (emotion constancy) strongly implies an activation of the fear system [ 45 , 47 ] and that precaution and avoidance behavior are adequate reactions in this situation. Our data, in particular the large response to constant maximally fearful expressions in different individuals, suggest that the amygdala plays a central role in the neurobiological realization of this environmental evaluation. Conclusions Emotional facial expressions are an important cue for the appraisal of an environmental situation. Our study has demonstrated a new perspective on the functional characterization of the amygdala involved in the perceptual processing of human faces by incorporating the dimensions of constancy and variability detection in these stimuli. These are essential for the assessment of the temporal dynamics of social situations. Methods Experimental design A 2 × 2 factorial design (identity × emotional expression) across 4 different block types was used: (a) constant identity, constant expression (C I C E ), (b) variable identity, constant expression (V I C E ), (c) constant identity, variable expression (C I V E ), (d) variable identity, variable expression (V I V E ). Figure 1 schematically displays our experimental design showing three consecutive stimuli from one block within each cell of the factor table. Subjects 13 Subjects (8 females, 10 right-handed) participated in this fMRI study. The mean age was 25.6 (SD 7.8). The data sets of two subjects were excluded from further image analysis because of radio frequency artifacts caused by the scanner leaving a total of 11. All subjects were fully informed about the experimental procedure and signed a consent statement which was approved by the local ethics committee. Experimental procedure The procedure was completed in one imaging session with 4 runs each containing 20 blocks of stimulus presentation. Within one run the blocks were interleaved with 15 s of central fixation (rest period). Each run started with a rest period of 20 s. 20 stimuli were presented per block at a rate of 1 Hz. Stimuli were shown for 900 ms interposed with a 100 ms gray blank of mean luminance to make transitions between stimuli less abrupt. Numbers and types of both facial expressions and facial identities were counterbalanced across the entire experiment. Conditions varied only in terms of the sequential configuration of the stimuli within the blocks. Stimulus order within each block was pseudo-randomized and fixed. The sequence of blocks within each run was counterbalanced, pseudo-randomized and fixed to minimize stimulus order effects. Additionally, the sequence of the four runs was pseudo-randomized across subjects assuring different orders of runs in each subject. Stimuli consisted of 4 different faces (facial identities) drawn from the Ekman series of facial affect [ 48 ] with 5 different emotional expressions ranging from fearful to happy expressions, including neutral. Two intermediate expressions displayed fearful and happy emotions at reduced (50 %) intensities. Those face pictures were interpolations using computer morphing procedures [ 49 ] similar to those in other studies. Subjects were instructed to fixate on a small red fixation dot presented in the middle of the viewing monitor while simultaneously attending to the entire stimulus presentation. In order to maintain and control for attentional effects during the procedure we included an oddball task as we were seeking to avoid confounds by an emotional judgment or a gender differentiation task (latter being a property of facial identity, [ 26 ]). As oddball stimulus we occasionally presented the facial stimuli with reduced luminance of the entire face while leaving the stimulus visible which was not expected to affect activation in high-order visual areas [ 50 ]. We chose to manipulate the entire facial stimulus for the oddball task in order to keep the subjects attention on the entire stimulus rather than on some small feature. Subjects were instructed to respond with a button press whenever an oddball target appeared. The number of oddball targets per block ranged from 2 to 4 to avoid subjects' expectancy effects. In order to prevent a systematic effect of the number of oddballs in the data analysis, the number of oddball stimuli was counterbalanced across blocks and conditions. Subjects were familiarized with the oddball task in a practice session prior to the first scanning run. Image acquisition Imaging was performed on a 1.5 T Magnetom Vision (Siemens, Erlangen, Germany) scanner. 43 transversal slices of echo-planar (EPI) T2 weighted images in each volume with a slice thickness of 2 mm and 1 mm gap (TR = 3.5 s, TE = 40 ms, flip angle 90°, FoV 192 × 192 mm 2 , matrix 64 × 64) were acquired. A total of 204 volumes were collected per run. Image processing Image processing and statistical analysis were carried out using SPM99 for the single subject analysis and SPM2 for the group analysis [ ] All volumes were realigned to the first volume, spatially normalized to a standard EPI template [ 51 ] using sinc interpolation and finally smoothed with a 11 mm isotropic full width at half maximum (FWHM) Gaussian filter to account for anatomical differences between subjects and to allow statistical inference using Gaussian Random Field theory. Statistical analysis The data of 11 subjects were included in the statistical analysis. Data analysis was performed using the mass univariate general linear model as implemented in SPM99 and commenced by specifying the design matrix for each subject using a boxcar and an exponential decay regressor for modeling the hemodynamic response to each experimental condition. The boxcar regressor models the mean activation within the block while the exponential decay regressor (time constant 4 s) models decreases and increases (through negative contrast weights) within the block. The conditions with constant emotional expressions (C I C E , V I C E ) allowed a decomposition into emotion-specific components. Thus each of these two conditions involved 10 regressors (boxcar and exponential decay for 5 different expression), while each of the other conditions (C I V E , V I V E ) were specified with two regressors. Data were high-pass filtered at 1/120 Hz. Serial autocorrelation was controlled by superimposing a known autocorrelation in form of temporal smoothing using a low-pass filter at 4 sec filter width. Successively, contrasts for each experimental condition were computed by averaging the same block type across runs and multiplying the design matrix with the contrast vectors. These single-subject contrast images were then taken to the second level oneway ANOVA [ 52 , 53 ] in SPM2 allowing for an appropriate non-sphericity correction [ 54 ]. This correction is equivalent to the Greenhouse-Geisser procedure in multivariate ANOVA analyses and allows for correct assessment of the error covariance matrix, hence securing valid inference in the group comparisons. In order to detect voxels that show elevated responses to the same expression displayed in different individuals we constructed the interaction contrast of our 2 × 2 design. Hence, we created the contrast [(V I C E > C I C E ) > (V I V E > C I V E ), p < .01 for the purpose of visualization] for the amygdala and masked it with the contrast [(V I C E > C I V E ), p < .05] to exclude regions showing higher activations to C I V E than to V I C E . For our hypothesis in the fusiform gyrus we used a contrast that compared conditions with at least one changing stimulus feature (facial identity, emotional expression, or both) with the condition in which the same stimulus was shown for the entire block [(V I C E + C I V E + V I V E ) > 3 × C I C E ]. T-statistics for the assessment of significant regional activation were assembled into Statistical Parametric Maps (SPMs) which refer to the probabilistic behavior of Gaussian random fields [ 55 ]. Our threshold was set at p < .05 (corrected). Because we had region-specific hypotheses for the amygdala and the lateral fusiform gyrus, we applied a reduced search volume to our amygdala activation which was derived by an anatomical mask created with MRIcro [ 56 ] on the template brain of the Montreal Neurological Institute (MNI, [ 57 ]). With additional visual reference to a high-resolution anatomical atlas [ 58 ] we outlined the amygdala on each slice of the MNI template brain. Thus, the amygdala search volume comprised 77 voxels, or 2071 mm 3 on the right side and 74 voxels, or 2005 mm 3 on the left side. Similarly, we applied a 10 mm radius sphere to our activation peaks in the lateral fusiform gyrus centered on coordinates reported by Vuilleumier and colleagues when contrasting faces vs. houses in a functional localizer task [ 33 ]; see Table 1 ). For additional brain areas not included in our volumes of interest we corrected for the entire brain volume. For the emotion-specific analyses of condition V I C E (Figure 2 and 4 , bottom panels) we referred to specific contrasts for each emotional expression created at the single-subject level. These contrast images were raised to another second level one-way ANOVA to test, for example, for significant activations to maximally fearful faces within this condition. These statistical comparisons were carried out at an uncorrected significance threshold. List of abbreviations EPI: echo-planar imaging fMRI: functional magnetic resonance imaging LOC: lateral occipital complex RT: reaction time SPM: statistical parametric map V I V E : variable identity, variable emotion V I C E : variable identity, constant emotion C I C E : constant identity, constant emotion C I V E : constant identity. variable emotion Fear/Happy Max: maximally fearful/happy expression Fear/Happy Min: minimal fearful/happy expression. Authors' contributions J.G. and O.T. designed, coordinated, and conducted data collection, analysis, and interpretation. C.B. and C.W. conceived of the study and participated in its design, analysis and interpretation. All authors read and approved the final manuscript.	/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC535816.xml
548143	Estrogen, mitochondria, and growth of cancer and non-cancer cells	In this review, we discuss estrogen actions on mitochondrial function and the possible implications on cell growth. Mitochondria are important targets of estrogen action. Therefore, an in-depth analysis of interaction between estrogen and mitochondria; and mitochondrial signaling to nucleus are pertinent to the development of new therapy strategies for the treatment of estrogen-dependent diseases related to mitochondrial disorders, including cancer.	Introduction Estrogen is considered to elicit different growth responses in various tissues through binding to the estrogen receptor (ER) α and ERβ [ 1 - 3 ]. The modulation of estrogen responsive gene transcription by the ER is termed a " genomic " action of estrogen. In contrast, " non-genomic " effects of estrogen are characterized by a rapid onset of action within seconds to minutes after hormone exposure through post-translation modification of signaling proteins. ER-mediated signaling pathways are considered to support the growth of normal, preneoplastic and neoplastic cells [ 1 - 3 ]. In contrast to the classical genomic pathways of estrogen action that occur over the course of several hours or days, recent studies have shown evidence of a rapid signaling pathway mediated by cell surface ERs and non-genomic estrogen-induced signal transduction pathways which contribute to cell proliferation [ 4 ]. Recently, we have identified that estrogen stimulated the growth of HEK 293 cells in an ER-independent manner given that this cell line does not contain ER (unpublished Singh KP, Venkat S, Roy D). These findings suggest that in addition to ER mediated actions, other factor(s) must be involved in the stimulation of cell growth by estrogen. More recently, mitochondria have been implicated in the control of cell proliferation [ 5 ]. For instance, the mitochondrial peripheral benzodiazepine receptor (PBR) has been implicated in the regulation of human breast cancer cell proliferation [ 6 ]. Similarly, we have demonstrated that mitochondria can modulate the expression of nuclear cell cycle genes and human breast tumor growth [ 7 , 8 ]. For instance, the growth of estrogen-dependent and estrogen-independent cells, is inhibited by controlling mitochondrial biogenesis [ 8 ]. In this paper, we critically review the role of mitochondria in the growth of estrogen-dependent cancer and non-cancer cells. Mitochondria are important targets of estrogen action. The cross-talk between the cell nucleus and the mitochondria appears to control estrogen-induced signaling involved in the apoptosis, proliferation, and differentiation of both normal and malignant cells. Mitochondria through its interaction with the cytoskeleton, export of cleaved signaling peptides, or generation of ROS appears to transduce signals to the nucleus for the activation of transcription factors involved in the cell cycle progression of estrogen-dependent cells. The understanding of the regulation of mitochondrial biogenesis by estrogenic compounds would open a new way to better understand steroidal and non-steroidal estrogen action at the cellular level. Estrogen actions at the mitochondria Besides apoptosis, respiration, and oxidative phosphorylation; mitochondria also control ion homeostasis and the synthesis of heme, lipids, amino acids, and nucleotides. Steroidogenesis is also controlled by mitochondria. Estrogen biosynthesis-related enzymes, 3 β-hydroxysteroid dehydrogenase and aromatase, have been demonstrated in the mitochondria of ovarian tumor epithelial cells [ 9 ]. Estrogen transport to mitochondria Although estrogen synthesis occurs in the mitochondria, exogenously added estrogen is also transported to this organelle. For instance, in vivo exposure of ovariectomized rats to tritiated 17-β-estradiol (E2) showed with increasing time, the translocation of this hormone from the plasmalemma mainly to the mitochondria (75%) rather than the nuclei in liver, adrenal gland, and spleen tissues [ 10 ]. The lipophilic property of E2 allows this molecule to easily diffuse into lipid bilayers. Since mitochondria are enriched with lipids, the organelle has the ability to act as an estrogen-sink within the cell. Although passive diffusion of estrogen into the mitochondria exists, a rapid delivery of E2 via receptor-mediated endocytosis from the plasma membrane to the mitochondria has been reported as a potential new pathway in HepG2 cells [ 11 ]. The uptake of E2-BSA from the medium by HepG2 cells occurred as early as 30 min. post-exposure and the ligand could be viewed in organelles that resembled vesiculated mitochondria found in steroid producing cells of the adrenal cortex and testes [ 11 - 13 ]. Estrogen effects on mitochondria morphology The impact of estrogen on mitochondrial morphology has previously been reported in the human breast cancer cell line MCF7. Transmission electron microscopy (TEM) revealed that an 8 day treatment of MCF7 cells with E2 (10 nM) resulted in large, clear mitochondria [ 14 ]. These alterations in mitochondria structure were observed as early as 2 days after treatment with a physiologically relevant dose of estrogen. The delamellated cristae of E2 treated mitochondria resemble an early anaerobic state of mitochondria development seen in embryonic rats and primates in which the cell depends on glycolysis [ 15 ]. However, it is not known whether the reported estrogen induced morphological changes had an affect on mitochondrial function. Estrogen receptor localization in mitochondria The function of estrogen at the mitochondria is not clear, however, recent studies have identified ERα and ERβ within the mitochondria implicating its role in the regulation of mitochondrial genome transcription. Subcellular fractionation of rabbit ovarian and uterine tissue revealed isoforms of ERα and ERβ in the mitochondrial enriched fraction as detected by western Blot analysis using ER specific antibodies [ 16 ]. More recently, ERβ was shown to be localized in the mitochondria of human lens epithelial cells (HLE-B3), human heart, rat primary neuron and primary cardiomyocyte, and in a murine hippocampal cell line HT-22 [ 17 , 18 ]. Western blot analysis using polyclonal antibodies against human ERα and -β showed both ERs present in purified mitochondria isolated from the human breast cancer cell line MCF7 [ 19 ]. Using immunohistochemistry with confocal microscopy and immunogold electron microscopy, ERα and ERβ were identified in the MCF7 mitochondrial matrix. Mitochondrial ERα and ERβ were reported to account for 10% and 18%, respectively, of total cellular ER-α and -β when MCF7 cells were treated with E2. Treatment of MCF7 with E2 (10 -8 M and 10 -9 M) significantly increased the mitochondrial level of ERα and ERβ by 2.5-fold. Estrogen influence on mitochondrial gene expression An increasing body of evidence has shown that mitochondrial transcription is enhanced by estrogen treatment. For instance, a 16-fold increase in cytochrome oxidase II (CO II) mRNA is reported in the GH4C1 rat pituitary tumor cell line when treated for 6 days with E2 (0.5 nM) [ 20 ]. The mitochondrial gene for subunit III of cytochrome oxidase (CO III) is induced as early as 3 h following a single dose of E2 in the hippocampus of ovariectomized female rats [ 21 ]. Other mitochondrial transcripts have also been reported to increase in the human hepatoma cell line, HepG2, and rat hepatocytes when exposed to ethinyl estradiol (EE). A 40 h exposure to an EE concentration ranging from 0.5 to 10 μM resulted in a 2- to 3-fold induction of CO I, CO II, and NADPH dehydrogenase subunit 1 (NADPH-DH1) mRNA [ 22 ]. E2 (20 μM), although less potent than EE, showed a similar effect of induction from 1.5- to 1.8-fold in mitochondrial transcripts CO I, CO II, and NADPH-DH1 when treated for 12 h. The E2 catechol metabolite 4-OH-E2 caused a greater response in CO I and CO II transcript levels as compared to E2 after 24 h of treatement with a dose of 10 μM. The mitochondrial gene for ATP synthase subunit 6 (ATPase 6) was also elevated in female rat liver tissue exposed to EE (5 μg/day) for 42 days. An increase in the transcript level of COX7RP (cytochrome c oxidase subunit IV-related protein) was reported after a 6 h E2 (100 nM) treatment in MCF7 cells [ 23 ]. It is not known whether the COX7RP transcript is translated to a functional protein in the mitochondria, but the study proposed that COX7RP may represent a regulatory subunit of cytochrome c oxidase that modulates a high state of energy production in estrogen sensitive target tissues. More recently, a 12 h E2 (0.3 μM) treatment of MCF7 cells was demonstrated to enhance the mitochondrial transcript levels of CO I approximately fourfold and CO II~2.5-fold [ 19 ]. The mechanism of estrogen-induced mitochondrial gene transcription is not clearly understood. The involvement of estrogen responsive elements (EREs) and/or the ER may be a possible mechanism in the increase of these mitochondrial transcripts. Sequences with partial similarity to the ERE consensus sequence, (AGGTCANNNTGACCT), have been reported in the mouse mitochondrial genome [ 24 ]. These partial EREs were detected in genes CO I and CO II which may account for the observed increases in these two transcripts in rat GH4C1 pituitary cells and rat hepatocytes [ 20 , 22 ]. Other genes in which the various EREs were detected include 12S rRNA, 16S rRNA, tRNA-gln, cytochrome oxidase b, unidentified reading frame (URF) 4, URF5, and the D-loop region [ 24 ]. In the human mitochondrial genome, we identified partial or ERE 1/2 sites in the D-loop region, CO II, tRNA-met, 12S rRNA, 7S rRNA, URF1, and URF5 (unpublished Felty Q and Roy D). The presence of these partial EREs in the mitochondrial genome may lend support to a novel ER signal transduction pathway. A mechanism of ER translocation into the mitochondria and ER binding to mitochondrial EREs remains unclear. Using electrophoresis mobility shift assay (EMSA) and plasmon resonance analysis, the recombinant human ERα- and ERβ-containing mitochondrial proteins were demonstrated to specifically bind putative EREs in the mtDNA D-Loop, and this ER binding was enhanced by E2 treatment and inhibited by ICI 182780 [ 19 ]. Based on this evidence, it is biologically plausible that ER mediates mitochondrial transcription in the same manner as the glucocorticoid receptor (GR) which is translocated into the mitochondria and binds glucocorticoid response elements (GRE) after treatment with glucocorticoid [ 25 , 26 ]. Whether estrogen-induced mitochondrial transcription participates in the development and growth of estrogen dependent breast cancer is not known. Long-term stilbene estrogen (diethylstilbestrol = DES) treatment of Syrian hamsters produced tumors in the kidney with a 5- to 10-fold higher transcript level of CO III than age-matched control kidneys [ 27 ]. Estrogen and the electron transport chain Besides transcription of mitochondrial genes, estrogen has also been demonstrated to effect mitochondria at the protein level. Estrogen has been demonstrated in several studies to inhibit mitochondrial respiratory complex I, II, III, IV, and mitochondrial ATP synthase (F 0 F 1 -ATPase) [ 28 - 31 ]. Several studies have reported estrogen specific inhibition of mitochondrial respiratory proteins, but it is not clear whether estrogen can modify mitochondrial proteins at the post-translational level. However, there is a report that E2 increased the phosphorylation of a 76 kDa protein in the mitochondrial fraction of the rat corpus luteum [ 32 ]. The presence of protein kinases within the mitochondria together with evidence for estrogen-induced phosphorylation of mitochondrial proteins suggest that estrogen may regulate mitochondrial respiratory physiology at the post-translational level [ 33 ]. Besides a direct interaction with mitochondrial proteins, estrogen may indirectly effect the electron transport chain (ETC) through an increase of membrane fluidity. Given that thyroid hormone increases mitochondrial membrane fluidity [ 34 ] and that physiologic concentrations of estrogen can alter the fluidity of human red blood cell membranes [ 35 ], it is likely that a low dose of estrogen may facilitate electron transfer by increasing respiratory protein interactions through more membrane fluidity. Thus, the rate of electron transfer may increase as a consequence of more frequent collisions/interactions between respiratory chain complexes and electron carriers. The three proton pumps of the ETC depend on electron flow to generate the mitochondrial membrane potential (ΔΨ m ). In human neuroblastoma cells, the following ER ligands: tamoxifen (30.2 μM), clomiphene (10.6 μM), and nafoxidine (2.8 μM); were reported to modulate ΔΨ m while E2 did not change ΔΨ m at concentrations up to 100 μM [ 36 ]. Assuming that estrogen can effect the fluidity of the inner mitochondrial membrane, a rise in ΔΨ m could result from an increased rate of electron transfer. The formation of ROS in mitochondria is reported to occur at high ΔΨ m [ 37 , 38 ] and therefore suggests that estrogen modulation of ΔΨ m may be a possible mechanism for the generation of ROS. Whether physiologic concentrations of estrogen can increase ΔΨ m in target tissues is not clear, but these lines of evidence suggest that estrogen may modulate ΔΨ m in a dose- and tissue-specific manner. This effect is significant because estrogen-induced mitochondrial reactive oxygen species (ROS) may participate in cell signaling. In the following section, we provide a detailed review of the effects of E2 on mitochondrial respiratory complexes. NADH dehydrogenase (complex I) The effect of natural estrogens and synthetic estrogens on the mitochondrial ETC has been demonstrated in several studies. Human NADH dehydrogenase (complex I) is the largest respiratory chain complex consisting of 7 mitochondrial genome encoded subunits and more than 41 subunits encoded from the nuclear genome. This integral membrane protein is located within the mitochondrial inner membrane and the matrix. Two electrons enter the ETC from the oxidation of NADH by ubiquinone (CoQ) at complex I which is coupled with proton movement across the inner membrane from the matrix to the intermembrane space. Natural estrogens, 17-α-estradiol, E2, and estrone, at concentrations of approximately 10 μM were demonstrated to inhibit mitochondrial electron transport in homogenates of rat uterus, liver, and skeletal muscle [ 39 ]. The synthetic estrogen DES was also reported to inhibit electron transfer from complex I to CoQ at a half-maximal inhibitory concentration range of 0.2–2.6 μM [ 28 ]. Although DES at a dose of 20–30 μM could inhibit electron transfer by 90% it was much less effective than rotenone and piericidin A which could inhibit electron tranfer by 90–95% at a dose of 30–50 nM due to a tighter binding affinity to complex I. The researchers postulated that at relatively low doses, DES reversibly inhibits electron transfer at complex I. Additionally, DES displayed specific binding to a site in which rotenone and piercidin A bind to complex I. Since photoreactive analogues of rotenone have been reported to label complex I [ 40 ], these results implicate estrogen specific binding to respiratory complex I. Another class of compounds, phytoestrogens, which are found in our diet can also inhibit the activity of complex I. Phytoestrogens, genistein (found in soy beans) and resveratrol (found in red wine), can inhibit the activity of complex I, but are considered less active than rotenoid compounds [ 41 ]. There are very few reports that have investigated co-treatment of estrogen with mitochondrial inhibitors of oxidative phosphorylation (OXPHOS). Complex I inhibitors rotenone, piericidin A, and amytal have been used in co-treatment with DES and/or estrogens to elucidate the site of estrogen action on electron transfer [ 28 , 39 ]. In MCF7 cells, it was demonstrated that co-treatment with rotenone (10 nM) and E2 (10 nM) strongly inhibited ornithine decarboxylase activity by 86% [ 42 ]. More recently, a study reported that treatment with 10 μM of 2-methoxyestradiol (2-Me) induced apoptosis in Ewing sarcoma cells through hydrogen peroxide (H 2 O 2 ) production [ 43 ]. Since a 2 h prior treatment of rotenone (6 μM) inhibited 2-Me induced apoptosis and H 2 O 2 production, it was suggested that the H 2 O 2 source was the mitochondria. Although the biological significance of estrogen interaction with complex I is unknown, evidence of estrogen inhibition of electron transfer support a novel role of estrogen in the formation of mitochondrial ROS. Succinate dehydrogenase (complex II) Human succinate dehydrogenase (complex II) is a membrane bound protein located on the matrix side of the inner mitochondrial membrane. Complex II consists of 4 subunits all encoded from the nuclear genome. Two electrons enter the mitochondrial ETC from the oxidation of FADH 2 by CoQ at complex II. Although previous studies have reported complex I and cytochrome bc 1 reductase (complex III) as the major sites of ROS generation, ubiquinone radicals have been reported to contribute to basal levels of ROS at complex II [ 44 ]. Physiologically relevant ROS generation supported by the complex II substrate succinate occurs at complex I through reversed electron transfer [ 45 ]. Research studies of the protective effect of estrogen on the brain use the complex II inhibitor, 3-nitroproprionic acid (3-NPA) to model the condition of ischemia. Although the mechanism of estrogen neuroprotection is not clearly understood, estrogen has been proposed to modulate cerebral energy/glucose metabolism. Both 17-β-E2 and 17-α-E2 have been demonstrated in vivo to reduce ischemic brain damage induced by middle cerebral artery occlusion in ovariectomized rats [ 46 , 47 ]. To model an in vitro state of interrupted energy metabolism as seen in cerebral ischemia and chronic neurodegenerative disease, the human neuroblastoma cell line SK-N-SH was treated with 3-NPA (10 mM) [ 48 ]. This study demonstrated that pretreatment of SK-N-SH with E2 (2 μM) restored ATP levels to 80% at 12 h as compared to the control cells treated with 3-NPA alone. Whether mitochondrial function can be preserved with a physiological concentration of estrogen is not clear as this study used a high dose (2 μM) which can be cytotoxic in certain tissues. The maintenance of ΔΨ m and ATP levels by estrogen when faced with 3-NPA toxicity was proposed to be due to the antioxidant effect and/or ATP increasing effects. At the 2 μM E2 dose an antioxidant effect is likely because estrogen is reported to be an effective neuroprotective antioxidant in the micromolar dose range [ 49 ]. However, another possibility may be due to an estrogen-induced increase in complex II activity which would overcome inhibition by 3-NPA as succinate dehydrogenase activity has been reported to increase in the brain of E2 treated rats [ 29 ]. Estrogen has been shown to cause multiple effects at the level of complex II that include the inhibition of electron transfer, maintenance of ΔΨ m and ATP levels, and the enhancement of succinate dehydrogenase activity. Cytochrome bc 1 reductase (complex III) The human cytochrome bc 1 reductase (complex III) is an integral membrane protein located in the inner mitochondrial membrane consisting of 10 subunits encoded from the nuclear genome and 1 subunit encoded from the mitochondria. Complex III transfers electrons from ubiquinol (CoQH 2 ) to cytochrome c which is coupled to proton movement across the inner membrane. The synthetic stilbene estrogen, DES (10 μM–50 μM), is reported to inhibit complex III at the site of electron flow from ubiquinone to cytochrome c 1 [ 50 ]. The ER ligand tamoxifen is also reported to inhibit electron transfer at the site of complex III in isolated rat liver mitochondria [ 30 ]. The phytoestrogen resveratrol has been reported to bind both ER α/β, increase expression of estrogen responsive genes, and stimulate cell proliferation of MCF7 and T47D breast cancer cells [ 51 , 52 ]. Interestingly, resveratrol has been shown to inhibit complex III activity (20%) by competition with ubiquinol (CoQH 2 ) and preserve mitochondrial function by its action on complex III in isolated rat brain mitochondria [ 53 , 54 ]. Besides the inhibition of electron transfer at complex III, the phytoestrogen genistein (50 μM) induced mitochondrial permeability transition (MPT) in isolated rat liver mitochondria. Since estrogen can inhibit electron transfer at complex III, the potential of estrogen to modulate the formation of mitochondrial ROS exists at multiple respiratory complexes. Cytochrome c oxidase (complex IV) In humans, the mitochondrial cytochrome c oxidase (Complex IV) is located in the inner membrane and catalyzes the transfer of electrons to oxygen with water being the final product of the reduction reaction coupled with proton movement across the inner membrane. The mitochondria genome encodes 3 subunits of complex IV while the other 10 subunits are encoded from the nuclear genome. The effect of estrogen and ER ligands on complex IV has been reported to increase and decrease enzymatic activity. During the oestrus cycle in rats, E2 decreased complex IV activity in brown adipose tissue, which suggests a role for E2 in the modulation of oxidative capacity [ 55 ]. The ER ligand tamoxifen was reported to restore complex IV activity to normal levels in disrupted rat liver mitochondria [ 56 ]. In rat liver mitochondria, tamoxifen was demonstrated to have a biphasic effect on complex IV in which a low concentration of 10–15 μM increased enzyme activity while a higher concentration of 50 μM inhibited activity by 50% [ 30 ]. Although the biological effect of estrogen actions on complex IV remains to be elucidated, it has been suggested that estrogen modulation of complex IV activity may increase energy production in estrogen sensitive tissues [ 23 ]. Furthermore, it has been proposed that ΔΨ m and ROS formation may be controlled by a hormone mediated reversible phosphorylation of complex IV [ 57 ]. Thus, the possibility for estrogen control of ROS formation by modulating complex IV activity may provide a mechanism of estrogen-induced redox signaling by the mitochondria. Mitochondrial ATP synthase (F 0 F 1 -ATPase/Complex V) Mitochondrial ATP synthase (F 0 F 1 -ATPase/Complex V) is composed of two distinct parts: 1) the F 1 -ATPase portion which protrudes into the matrix and synthesizes ATP when protons pass through it down their electrochemical gradient. 2) the F 0 -ATPase which forms a transmembrane proton channel through the inner membrane. F 0 F 1 -ATPase is encoded by 2 mitochondrial genome encoded subunits and 14 nuclear encoded subunits. F 0 F 1 -ATPase like respiratory chain complexes I-IV represents another enzyme in the mitochondrial inner membrane that is sensitive to estrogen. Inhibition of rat liver F 0 F 1 -ATPase by DES (10 μM) has been demonstrated and the F 0 portion was reported to contain a distinct binding site for DES [ 58 ]. This specific binding site for the F 0 portion has made DES a unique probe for the rapid isolation of functional F 0 from rat liver mitochondria [ 59 ]. Using E2-BSA conjugates, a 23 kDa estrogen binding protein was identified in rat brain mitochondria [ 60 ]. The 23 kDa protein was identified as the oligomycin-sensitivity conferring protein (OSCP) which forms the stalk region between F 0 and F 1 subunits of F 0 F 1 -ATPase in mitochondria. The OSCP was proposed to be the specific site on F 0 F 1 -ATPase where estrogen modulates ATPase activity. Differential effects of estrogen on F 0 F 1 -ATPase activity in isolated rat heart, liver, and brain mitochondria were observed when treated with E2, DES, and resveratrol [ 31 ]. E2 (13 nM) stimulated F 0 F 1 -ATPase activity in the heart by 10%, but not in the liver and brain. The phytoestrogen resveratrol (13–15 μM) inhibited F 0 F 1 -ATPase activity in the heart and liver while lower doses (133 pM–1.3 nM) stimulated F 0 F 1 -ATPase activity in the liver by 10%. Both phytoestrogens resveratrol (19 μM) and genistein (55 μM) can inhibit F 0 F 1 -ATPase activity in rat brain mitochondria [ 61 ]. In rat heart, liver, and brain the mitochondrial F 0 F 1 -ATPase activity was inhibited by DES (6.7 μM) 61%–67% [ 31 ]. In rat brain mitochondria, 17α-estradiol and E2 partly inhibited F 0 F 1 -ATPase activity at low concentrations of 15 nM and 3.4 nM, respectively, while 17α-E2 preserved mitochondrial function altered by the stress of anoxia-reoxygenation [ 62 ]. Although these studies of estrogen effects on F 0 F 1 -ATPase have been shown in isolated mitochondrial preparations, F 0 F 1 -ATPase inhibition by E2 was demonstrated to occur with intact human osteolclastic FLG 29.1 cells [ 63 ]. These studies of F 0 F 1 -ATPase activity demonstrate that estrogens and estrogen-like compounds possess cell-type and dose-specific effects on mitochondrial function. Modulation of mitochondrial ROS by estrogens Within the cell, mitochondria are considered to be a major source of ROS, which include superoxide anion (O 2 •- ), H 2 O 2 , and the hydroxyl free radical ( • OH) [ 64 - 66 ]. Since mitochondria consume 85% of the oxygen used by the cell, the mitochondrial ETC generates a substantial amount of intracellular ROS [ 67 ]. As electrons pass through the mitochondrial ETC, some electrons leak out to molecular oxygen (O 2 ) to form O 2 •- which is dismutated by manganese superoxide dismutase (MnSOD) to form H 2 O 2 [ 64 , 68 ]. During mitochondrial respiration, 2% of the electron flow is reported to result in the formation of H 2 O 2 [ 66 ]. However, lower values of free radical leak were reported in the range of 0.4%–0.8% for heart mitochondria respiring on physiological concentrations of succinate (<0.5 mM) [ 37 ]. In support of these findings, intact rat skeletal muscle, heart, and liver mitochondria were reported not to produce measurable amounts of ROS when respiring on complex I and complex II substrates [ 69 ]. In addition, this study reported a lower estimate of electron flow (0.15%) that contributed to H 2 O 2 production under resting conditions. These results suggest that mitochondria produce low levels of ROS that can be effectively scavenged by the cell's antioxidant defenses at resting conditions. It is this point, the low basal level of ROS produced by the mitochondria at rest, which makes mitochondrial ROS ideal signaling molecules since its contribution to the intracellular level of ROS is not at so high a level to induce oxidative stress; instead, a low oxidant level provides a physiologically safe window for redox signaling which allows the cell to regulate mild to moderate oxidative changes and critically respond to them by activating cellular processes such as proliferation and differentiation rather than triggering cell death. Characteristics of mitochondrial ROS Mitochondria are a predominant source of ROS in most cell types with unique characteristics that may allow it to participate in growth signal transduction. First, mitochondria are unique because they are a regulatable source of ROS in response to external stimuli. For example, cortical neurons exposed to N-methyl-D-aspartate (NMDA) were reported to couple a rise in intracellular calcium with mitochondrial O 2 •- production [ 70 ]. Tumor necrosis factor alpha (TNF-α) is another example of stimulated mitochondrial generation of O 2 •- in L929 cells and this ROS generation is coupled to the cytokine by the TNF-α receptor [ 71 , 72 ]. Few other examples exist of mitochondria producing ROS in response to external stimuli, but more recently integrins (cell surface receptors that interact with the extracellular matrix) were reported to modulate mitochondrial ROS production for signal transduction [ 73 ]. Although signal pathways involved in triggering mitochondrial ROS remain largely unknown, it has been proposed that mitochondria participate in integrin signaling in a nonapoptotic manner, which leads to gene expression and cell differentiation. Mitochondrial ROS can enter the cytosol as either H 2 O 2 or O 2 •- where it can participate in redox signaling. Within the mitochondria, MnSOD can dismutate O 2 •- to H 2 O 2 which is a highly diffusible signaling molecule that can exit the mitochondria. In addition to H 2 O 2 , O 2 •- was demonstrated to be released by mitochondria to the cytosol via the voltage-dependent anion channels (VDACs) [ 74 ]. In regard to turning the mitochondria ROS signal off, cellular antioxidant defenses such as SOD, catalase, and glutathione peroxidase easily degrade ROS, which terminates the signal. Therefore, mitochondrial ROS fulfill the prerequisites of a 2 nd messenger since they are short-lived (rapidly generated and degraded), produced in response to a stimulus, highly diffusible, and ubiquitously present in most cell types. Mitochondria are highly dynamic structures capable of changing their shape (by elongation, branching, swelling) and their location inside a living cell [ 75 ]. It is becoming clear that the morphological, functional, and genetic differences (heteroplasmy) that exist within the mitochondria population may reflect a division of labor within the cell. Mitochondria have been reported to be morphologically heterogeneous and unconnected within individual cells [ 76 ]. Pancreatic acinar cells were reported to contain distinct groups of mitochondria classified by their cellular location that included perinuclear, subplasmalemmal, and perigranular mitochondria [ 77 ]. In light of this finding, the highly diffusible H 2 O 2 generated by mitochondria may become a specific signaling molecule as a function of location. For example, perinuclear mitochondria may generate H 2 O 2 that only transduces signals to the nucleus or the subplasamlemmal mitochondria may only activate signal cascades of plasma membrane origin. Additionally, perinuclear, subplasmalemmal, and perigranular mitochondria were independently activated by intracellular calcium signals in their immediate environment, which supports distinct calcium functions for each type of mitochondria. It is significant that mitochondria can create subcompartments or 'microzones' within the cytoplasm because signal transduction depends on the close proximity of substrates and effector molecules to be an efficient process. In addition, given the presence of other endogenous ROS sources besides the mitochondria such as NADPH oxidase, peroxisomes, cytochrome p450, xanthine oxidase, cyclooxygenase, lipooxygenase, and γ-glutamyl transpeptidase [ 78 ]; and because ROS is involved in a variety of signal cascades, understanding how mitochondrial ROS is activated at the right place and at the right time is vital in understanding the organelle's role in signal transduction. Compartmentalization has already been reported to play a key role in redox signaling and we consider this attribute when describing the mitochondria as a signal transducer [ 79 ]. In adult cells, mitochondrial clustering functions to create steep gradients of low molecular weight species such as O 2 , ATP, and pH resulting in specialized microzones that may facilitate signal specificity [ 80 ]. In the cytosol, the volume occupied by mitochondria in cells is highly variable and ranges from 15% to 50%. Based on volume, mitochondria compose a significant compartment within the cytosol that harbors signaling molecules. H 2 O 2 produced within the mitochondria is highly diffusible in contrast to O 2 •- , which cannot diffuse through membranes making it easily compartmentalized. Thus, mitochondrial generated O 2 •- may be kept separated from the cytosol until an appropriate stimulus releases it through VDACs. Another route for O 2 •- release may be through the mitochondrial permeability transition pore (MPTP) as low molecular weight compounds up to molecular weight 1500, can be exchanged between the mitochondrial matrix and the cytosol via this pore [ 81 ]. Since the MPTP is reported to reversibly open/close naturally in intact cells without resulting in apoptosis, mitochondrial signaling molecules could be exchanged with the cytosol by the transient 'flickering' (open/closing) of the MPTP in response to certain stimuli [ 82 ]. In addition to location and compartmentalization, protein scaffolds mediate the selective activation of the mitogen activated protein kinase (MAPK) signaling pathway which raises the question of whether mitochondria may also act as a protein scaffold for signaling complexes [ 83 ]. A-kinase anchoring protein (AKAP) is reported to tether protein kinase A (PKA) to the mouse mitochondrial outer membrane [ 84 , 85 ]. What makes AKAPs unique is their ability to simultaneously bind multiple signaling enzymes such as other kinases and phosphatases [ 86 ]. This multivalent scaffold has been described as a 'transduceosome' capable of integrating signals from multiple pathways [ 87 ]. Whether these multivalent scaffolds exist on mitochondria is not clear at this time, but these signaling complexes could be a mediator of signals between the mitochondria and the nucleus during cell division. For example, AKAP84/121 has been demonstrated to concentrate on mitochondria in interphase and on mitotic spindles during metaphase transition alluding to its role in the cell cycle [ 88 ]. In addition to AKAP, a mitochondrial signaling complex has been reported to activate MAPKs. The PKCε can form signaling complexes with ERKs, JNKs, and p38 MAPKs in the murine heart [ 89 ]. Activated PKCε was shown to increase phosphorylation of mitochondrial ERK and p38 MAPKs. Whether the anchoring of PKC to mitochondria depends on AKAP is not known at this time. Mechanisms of estrogen-induced mitochondrial ROS Studies of the mitochondrial ETC have reported only two ROS forming sites, the FMN group of complex I and the ubiquinone site in complex III [ 45 , 90 ]. The topology of ROS production, on which side of the mitochondrial inner membrane O 2 •- is produced, was reported to occur on the cytosolic side by complex III and on the martix side of the inner membrane by complex I [ 69 ]. Estrogen is known to act as either an antioxidant or pro-oxidant depending on the concentration [ 91 ]. Whether physiological concentrations of estrogen can stimulate mitochondrial ROS at complex I and complex III is not clear since most studies have been performed with cytotoxic doses. In the following section, we provide evidence for potential mechanisms of estrogen induced mitochondrial ROS. Electrons feed into the mitochondrial ETC at complex I and complex II. At complex I and complex II, ubiquinone or co-enzyme Q 10 (CoQ) oxidizes NADH and FADH 2 , respectively. CoQ functions as a mobile electron carrier within the mitochondrial inner membrane and transfers 2 electrons from both NADH and FADH 2 to complex III [ 92 ]. CoQ is known to offer protection from heart disease by increased ATP production and antioxidant actions [ 93 ]. It is a highly lipophillic compound due to its structure which includes an isoprenoid tail of usually 10 isoprene units in length, hence the designation Q 10 . Other than its role as an electron carrier and antioxidant, CoQ is also reported to act as a pro-oxidant. Although the pro-oxidative action of CoQ within the mitochondria is a matter of debate, O 2 •- formation occurs from a single electron transfer from ubisemiquinone to molecular oxygen. Exogenously added CoQ has been demonstrated to enhance O 2 •- generation in isolated respiratory complex I and III [ 94 ]. Further evidence in support of CoQ redox-cycling come from a study that demonstrated H 2 O 2 derived from decomposing O 2 •- was inhibited after the removal of CoQ [ 90 ]. Upon the re-addition of CoQ, H 2 O 2 was detected again which indirectly demonstrated that the O 2 •- may originate from CoQ. Although some reports make a case for O 2 •- formation by CoQ redox cycling in mitochondria, arguments against its role as a source of O 2 •- come from a study which demonstrated that O 2 •- formation did not occur in a water-free nonpolar reaction system that mimics the lipophilic nature of the inner mitochondrial membrane [ 95 ]. However, pretreatment of the membrane with toluene which increased its permeability to protons provided conditions in favor of O 2 •- formation by CoQ. Thus, it was proposed that under certain pathological conditions in which the inner mitochondrial membrane is protonated, CoQ becomes a significant source of O 2 •- [ 96 ]. In line with this result, a study demonstrated that CoQ (100 μM) enhanced the release of H 2 O 2 from mitochondria in the presence of antimycin A (2 μM) and to a lesser extent with Ca 2+ (10 μM) [ 97 ]. The antimycin A and Ca 2+ pretreatment was thought to induce a leaky inner mitochondrial membrane thereby allowing protons to interact with CoQ and enhance ROS production by redox cycling. Interestingly, CoQ shares similar characteristics to catechol metabolites of estrogen. The catechol metabolites 2- and 4-OH-E2 contain hydroxyl groups that can be oxidized to semiquinones, which in the presence of molecular oxygen can be further oxidized to quinones with the formation of O 2 •- [ 98 ]. Since CoQ undergoes reduction/oxidation (redox) reactions which result in the radical semiquinone intermediate (semiubiquinone) and quinone, it is biologically possible for catechol estrogens to participate in shuttling electrons and to act as a pro-oxidant like CoQ. Given that MCF7 cells treated with the o-quinone form of estrogen and NADPH produce significant amounts of H 2 O 2 in the mitochondrial subfraction [ 99 ]; and that mitochondrial enzymes catalyze redox reactons of stilbene estrogen in the mitochondria [ 100 ], the capacity of catechol estrogens to redox cycle within mitochondria suggests that these metabolites could facilitate mitochondrial ROS formation. Assuming that estrogen is a weak electron carrier compared to CoQ, it may have a tendency to leak electrons to molecular oxygen instead of transferring its electrons to complex III. The phytoestrogen and dietary flavinoid quercetin is reported to act as a pro-oxidant. Foods of plant origin contain flavinoids known to act as antioxidants or pro-oxidants depending on the concentration and metal chelates which catalyze the oxidative process [ 101 ]. Estrogen is similar to some flavinoids with respect to inhibition of the mitochondrial respiratory chain at complex I and complex II [ 102 , 103 ]. Based on these reports, the formation of O 2 •- by estrogen at the level of electron transport could be one mechanism of increased mitochondrial ROS. Mitochondrial Ca 2+ , [Ca 2+ ] m , accumulation is reported to promote the generation of ROS [ 104 ]. For example, an increase in [Ca 2+ ] m was reported to stimulate H 2 O 2 production by rat brain mitochondria in the presence of rotenone [ 105 ]. Using confocal microscopy, we have shown a time-dependent increase in [Ca 2+ ] m with E2 (100 pg/ml) treatment of MCF7 cells [ 106 ]. Another study reported an increase in [Ca 2+ ] m with E2 (10 ng/ml) treatment of hippocampal neurons [ 107 ]. The mechanism for these increases in [Ca 2+ ] m is not clear, however, the inhibition of Na-dependent Ca 2+ efflux from mitochondria was reported to increase calcium retention in E2 (1 nM) treated synaptosomal mitochondria [ 108 ]. There is some evidence which suggests allosteric inhibition of a respiratory complex may be a mechanism for hormone induced ROS formation. Allosteric inhibition of the respiratory complex IV (COIV) or cytochrome c oxidase is reported to occur by cAMP-dependent phosphorylation; and this inhibition is turned-off by Ca 2+ -activated dephosphorylation [ 109 ]. A study proposed that a hormone stimulated increase of cellular Ca 2+ may activate a mitochondrial protein phosphatase which dephosphorylates cytochrome c oxidase. In turn, cytochrome c oxidase is activated which results in a rise of ΔΨ m and ROS [ 57 ]. Interestingly, the ER ligand tamoxifen (15 μM) showed a slight stimulatory effect on cytochrome c oxidase [ 30 ]. Since estrogen is capable of increasing [Ca 2+ ] m , it is possible for estrogen to signal the formation of mitochondrial ROS through a similar mechanism. The inhibition of respiratory complex I is known to favor ROS generation. Rat brain mitochondria that respired on complex I substrates produced a substantial level of ROS when inhibited with rotenone concentrations as low as 20 nM [ 110 ]. Since estrogen is known to inhibit respiratory complex I, we speculate that complex I interactions with the hormone could favor ROS production in a manner similar to rotenone. The phytoestrogen genistein is another flavinoid besides quercetin that acts like a pro-oxidant at the level of mitochondria. Genistein (50 μM) treatment of rat liver mitochondria was shown to increase ROS formation through interaction with respiratory complex III which resulted in the opening of the membrane transition pore [ 111 ]. Besides hormone interactions with respiratory enzymes, post-translational modifications such as phosphorylation-/-dephosphorylation that affect the activity of mitochondrial proteins should also be considered in ROS generation. The cAMP-dependent protein kinase is reported to phosphorylate 6-, 18-, 29-, 42-kDa mitochondrial proteins in bovine heart and phosphorylate the human 18-kDa subunit which promotes the activity of complex I [ 112 - 115 ]. Since estrogen is reported to stimulate cAMP-dependent protein kinase activity in hippocampal neurons, it raises the possibility for estrogen to induce cAMP accumulation in mitochondria [ 116 , 117 ]. If estrogen increased cAMP levels within mitochondria, then cAMP-dependent phosphorylation of mitochondrial respiratory complexes may modulate ΔΨ m and/or [Ca 2+ ] m in favor of ROS generation. Several isoforms of the enzyme nitric oxide synthase (NOS) are reported to exist which include inducible-(i), endothelial-(e), neuronal-(n), and mitochondrial-(mt) NOS. Estrogen has been reported to induce various isoforms of NOS. The activity of eNOS is modulated by estrogen in human aortic endothelial cells, uterine artery, heart, and skeletal muscle [ 118 , 119 ]. Estrogen has also been shown to stimulate protein expression of nNOS in human neutrophils and the transcription of iNOS in rat macrophages [ 120 , 121 ]. These effects are not limited to E2 because the estrogen-like chemical bisphenol A and the phytoestrogen resveratrol are also reported to stimulate NO synthesis [ 122 , 123 ]. These lines of evidence demonstrate a significant role of estrogen compounds in the modulation of NOS and NO. In regards to redox signaling, a NO-dependent inhibition of cytochrome c oxidase has been proposed to generate O 2 •- which is dismutated into the membrane permeable second messenger H 2 O 2 [ 124 ]. Since estrogen is capable of inducing NOS activity and expression, we postulate that an estrogen induced rise in NO could participate in a similar manner whereby generating mitochondrial H 2 O 2 . Interestingly, NO induced mitochondrial biogenesis has been demonstrated in several cell lines, which include brown adipocytes, 3T3-L1, U937, and HeLa cells [ 125 ]. We have shown that estrogen can influence mitochondrial biogenesis (data unpublished) and postulate that estrogen-induced NO could be one possible mechanism. More specifically, since mtNOS activity is dependent on Ca 2+ , we propose that an estrogen-induced rise in [Ca 2+ ] m could stimulate mtNOS activity ultimately leading to the generation of ROS via NO-dependent inhibition of cytochrome c oxidase [ 126 ]. Estrogen-induced growth of cells in relation to mitochondria The rapid stimulation of intracellular ROS by platelet-derived growth factor (PDGF), epidermal growth factor (EGF), and nerve growth factor (NGF) suggests that this underlying mechanism of cell growth may be shared with other growth factors including estrogen [ 127 ]. Exogenous addition of low concentrations of H 2 O 2 and/or O 2 •- has been demonstrated to stimulate cell growth in a variety of cell types including muscle cells, fibroblasts, amnion cells, prostate cancer cells, and aortic endothelial cells [ 78 ]. The molecular signaling mechanism that initiates ROS production by mitochondria is not clear, however, other cell processes besides apoptosis may be coupled to this signaling event. Tumor necrosis factor alpha (TNF-α) induces gene expression via mitochondrial respiratory chain dependent activation of NF-κB, AP-1, JNK, and MAPKK [ 73 ]. The proliferative response of endothelial cells to hypoxia was demonstrated to be initiated upstream by mitochondrial ROS which activated the MEK/ERK pathway [ 128 ]. Although other endogenous ROS sources besides mitochondria such as NAD(P)H oxidase exist, mitochondria will be the focus of this paper for the following reasons: ( i ) mitochondria are the principal source of intracellular ROS in epithelial cells. ( ii ) the growth of adenocarcinomas occur in tissue of epithelial cell origin. A characteristic of rapidly dividing cancer cells is their capacity to produce significant amounts of intracellular ROS, which has been implicated in the promotion of accelerated cell cycle activity in neoplastic cells. Mitochondria have long been suspected to play a role in the development and progression of cancers. The ROS molecules H 2 O 2 and NO have been demonstrated to stimulate mitochondrial biogenesis, a process that depends on the flow of molecules into and out of the organelle [ 125 , 129 ]. Since mitochondrial proteins are encoded in two separate genomes (mitochondria and nuclear genome), biogenesis is a coordinated effort in which mitochondria transmit signals to the nucleus and vice versa. The question of how mitochondria transmit these signals in the process of cell proliferation has risen from reports of its involvement in cell growth. Cerebral granular cells isolated from newborn rats with high mtNOS activity were reported to exhibit maximal proliferation rates which depended on NO and H 2 O 2 levels. In addition, MnSOD displayed an increased pattern of activity similar to mtNOS [ 130 ]. NO has been proposed to inhibit cytochrome c oxidase in favor of O 2 •- production and therefore MnSOD may dismutate O 2 •- generated by NO-dependent inhibition into the signaling molecule H 2 O 2 . Ethinyl estradiol, E2, and estrogen catechol metabolites at a dose of 0.25 to 5 μM are reported to increase mitochondrial O 2 •- in cultured rat hepatocytes and HepG2 cells [ 131 ]. Although the biological significance of the estrogen-induced mitochondrial O 2 •- is not known at this time, ROS has been demonstrated to modulate ER protein expression in various cell lines. Treatment of human breast cancer cells MCF7 and T-47D with H 2 O 2 (2.5 μM) increased the protein level of ERβ [ 132 ]. In addition, PMA (100 ng/ml) treatment increased the expression of ERβ in the macrophage cell line J774A.1. Evidence for the involvement of redox signaling with estrogen-induced cell proliferation has been demonstrated in several studies. Liposomes containing SOD or catalase inhibited in vitro estrogen-induced proliferation of Syrian hamster renal proximal tubular cells [ 133 ]. The cytokines IL-1β and TNF-α are known to cause the release of O 2 •- from human fibroblast cells. Co-treatment with an inhibitor of IL-1β and TNF-α synthesis, pentoxifylline, inhibited stilbene estrogen-induced increase in myeloperoxidase activities, 8-hydroxydeoxyguanosine (8-OHdG) formation, mutations in the testicular genome, and prevented estrogen-induced testicular preneoplastic lesions [ 3 ]. Recently, we have shown that estrogen-induced stimulation of macrophage cells and MCF7 cells in part occurs through ROS [ 134 , 135 ]. We have also observed inhibition of estrogen-induced MCF7 cell growth by ROS scavengers such as N-acetylcysteine, ebselen, and catalase (unpublished Singh M, Felty Q and Roy D). ROS can modulate effector molecules such as PKC, p53, extracellular regulated kinase (ERK), nuclear factor-κB (NF-κB), and the c-fos/c-jun heterodimer (AP-1); and these effector molecules are known to participate in growth signal transduction [ 136 ]. Therefore, estrogen-induced production of mitochondrial ROS may activate cell growth in estrogen-sensitive tissues. Oxidative stress has been shown to affect mitochondrial proteins of chronically estrogenized Syrian hamster kidney. A decrease in thiol/sulfhydryl groups was reported in the mitochondrial fraction at a preneoplastic stage of carcinogenesis [ 137 ]. Estrogen-induced oxidative stress may be responsible for these post-translational modifications in mitochondrial proteins. This finding is significant in the context of cell signaling because redox reactions involving cysteine thiol groups transduce signals by breaking or forming protein dithiol/disulfide bridges [ 138 ]. Since estrogen can induce mitochondrial ROS, we infer that the oxidation of thiols in response to estrogen converts the oxidative stress to a change in protein function involved in cell growth. Oxidative stress modifies mitochondrial matrix protein thiols [ 139 ]. Similarly, thiols on protein subunits 51-kDa and 75-kDa of NADH dehydrogenase (complex I) have been reported to form mixed disulfides with glutathione (glutathionylation) in response to mitochondrial oxidative stress. This post-translational modification was reversible and correlated with an increase in mitochondrial O 2 •- production [ 140 ]. Evidence in support of a ROS signal transduction pathway originating from complex I comes from a study which reported that the mitochondrial complex I inhibitor, rotenone, blocked ROS mediated signaling. Interestingly, this study demonstrated that a co-treatment of rotenone (10 nM) and E2 (10 nM) inhibited ornithine decarboxylase activity by 86% in MCF7 cells [ 42 ]. Since ornithine decarboxylase activity is a marker for cell growth, it appears that a signal transduction pathway for estrogen-induced cell growth may originate from the mitochondria assuming that rotenone inhibition is specific to complex I. The antitumor arotinoid, mofarotene (Ro 40-8757), has been demonstrated to down-regulate mitochondrial encoded NADH dehydrogenase subunit 1 (MtND1) expression in breast cancer cell lines MDA-MB-231, ZR-75-I, and MCF7 [ 141 ]. Since MtND1 has been reported to form part of the rotenone-binding site in complex I [ 40 ], the absence of MtND1 may remove an important site of estrogen action in mofarotene treated cells and may account for the anti-proliferative effects of this compound. Whether these protein interactions and/or modifications can occur as a result of estrogen exposure remains to be investigated. From these investigations, we infer that estrogen mediated cell growth via mitochondrial generated ROS signaling molecules may exist and merits future exploration to address this novel pathway. The mitochondrial thioredoxin system has been demonstrated to play a role in cell cycle progression. In general, the two antioxidant oxidoreductase enzymes thioredoxin (Trx) and thioredoxin reductase (TrxR) that compose the system modulate signal transduction properties of ROS by the reduction of intracellular disulfides. Trx acts as a protein disulfide reductant for ribonucleotide reductase and several transcription factors including p53, NF-κB, and AP-1 [ 142 ]. Once oxidized the active disulfide site is reduced by TrxR re-generating the reductant form of Trx. Enzyme isoforms Trx-2 and TrxR2 are reported to exist in the mitochondria. A biological role for TrxR2 in cell growth was demonstrated in HeLa cells using a dominant negative form of TrxR2 (TrxR2DN) [ 143 ]. An increase of G 1 to S phase transition, cell growth, and transcription of cell cycle genes was induced by TrxR2N expression. TrxR2DN expression was suggested to increase intracellular H 2 O 2 , which in turn signaled cell proliferation. Although it is not clear whether estrogen can modulate the H 2 O 2 levels of mitochondria, estrogen (10 nM–100 nM) treatment of primary human endometrial stromal cells in vitro show an increase in Trx protein and mRNA which implicate Trx involvement in cell growth and differentiation of estrogen responsive tissue [ 144 ]. Alterations in cellular redox status by increased expression of TrxR2 have been suggested to play a role in the growth of hepatocellular carcinomas [ 145 ]. Whether estrogen can signal cell growth through Trx2 and/or TrxR2 is not known, but these findings suggest that estrogen may modulate signal transduction of mitochondrial derived ROS via the thioredoxin system. Transduction of estrogn-induced mitochondrial signals to nucleus Mitochondrial ROS fulfill the characteristics of a 2 nd messenger since they are short-lived (rapidly generated and degraded), produced in response to a stimulus, highly diffusible (H 2 O 2 ), and ubiquitously present in most cell types. It is not known whether mitochondrial ROS like H 2 O 2 are involved in signaling pathways that control estrogen-induced cell proliferation. In this section, we provide evidence in support of redox signaling pathways of mitochondrial origin, which may be involved in cell cycle progression of estrogen-dependent cells. Redox sensor proteins and transcription Protein kinases are known to participate in phosphorylation signal cascades, however, zinc finger domains contained in some proteins may allow them to participate in redox signaling networks. Zinc finger structures within a protein consist of at least two zinc-coordinated thiolates. Upon oxidation zinc is released from the protein, which converts cysteine thiol groups to disulfide. A conformational change in the protein may result in either its activation or inhibition. There are several protein kinases such as a-raf and PKC that contain zinc finger domains. In addition, zinc finger domains are also found in hormone receptors such as the GR and ER [ 146 ]. Both PKC and c-raf have been demonstrated to be redox activated at the zinc finger domain [ 147 , 148 ]. For instance, ROS can trigger the release of zinc ions from PKC which results in its activation. Another protein kinase, c-raf, known to participate in the MAPK signal cascade was also demonstrated to be redox activated at the zinc finger domain. The mitochondrial localization of protein kinases src, Akt, a-raf, and PKC is evidence that this subcellular compartment harbors oxidant sensitive proteins that may facilitate cross-communication between redox and phosphorylation networks [ 33 , 89 ]. Although the role of the protein kinase a-raf in the mitochondria is not clear, a-raf mRNA is highly expressed in normal murine tissues such as the epididymis, ovary, kidney, and urinary bladder [ 149 ]. In Hela cells, epidermal growth factor rapidly (2 min.) and transiently activated a-raf, which in turn phosphorylated the MAP kinase activator MEK1 [ 150 ]. Therefore, mitochondrial ROS may activate MAPK signaling via a-raf. It is interesting to note that E2 can stimulate the phosphorylation of a-raf and cell cycle progression in MCF7 cells [ 151 ]. Whether the estrogen induced phosphorylation of a-raf depends on ROS is not known. Mitochondrial PKCδ and PKCε could also activate the raf/MEK/ERK pathway or directly activate MAPKs, respectively [ 152 , 153 ]. Rapid effects of estrogen have been demonstrated to mediate the DNA binding activity and phosphorylation of transcription factors. E2 treatment of rat adipocytes doubled AP-1 DNA binding and phosphorylated CREB protein within 15 min [ 154 ]. The redox sensitive protein Akt is known to phosphorylate an upstream kinase, IKKα, which stimulates the degradation of Iκ-B [ 155 ]. Estrogen-induced mitochondrial ROS may stimulate Akt leading to the degradation of Iκ-B and activation of the transcription factor NF-κB. Whether estrogen treatment can activate Akt via mitochondrial derived ROS is not clear, however, phosphorylation and translocation of Akt to the mitochondria was demonstrated when cells are treated with estrogen [ 156 ]. Given that E2 can stimulate mitochondrial ROS generation; ER, src, a-raf, Akt, and PKC are targets of oxidative stimuli localized at the mitochondria; and the transcription factors AP-1, NF-κB, and CREB are stimulated by oxidants [ 127 , 131 , 157 ]; it is possible that estrogen specific effects at the level of mitochondria can activate these transcription factors. Based on these studies we postulate that estrogen-induced mitochondrial ROS stimulates oxidant sensors a-raf, Akt, or PKC, which in turn activate transcription factors such as NF-κB, CREB, or AP-1 via the MEK/ERK pathway resulting in the transcription of cell cycle genes containing DNA responsive elements for NF-κB, CREB, or AP-1 and ultimately estrogen-induced cell proliferation (Fig. 1 ). Figure 1 Hypothetical scheme outlining three E2-induced signaling pathways from mitochondria. ( i ) E2 binding to a plasma membrane receptor and/or mitochondrial respiratory complexes generates ROS which leads to kinase activation. ( ii ) E2-induced rise in mitochondrial calcium leads to the activation of calcium-dependent proteases which process signal peptides, in turn responsible for kinase activation. ( iii ) E2-induced cytoskeleton modifications by mitochondria leads to kinase activation. Increased kinase activity results in the activation of transcription factors responsible for cell cycle progression. Estrogen and mitochondrial-cytoskeleton interactions Mechanical signals associated with cytoskeletal tension generation and cytoskeleton restructuring are a requirement for anchorage dependent cells to pass through the late G 1 restriction point [ 158 ]. Since these cytoskeleton dependent effects on the G 1 checkpoint are independent from the MAPK signaling pathway, a new question rises of whether mitochondria can modulate cell growth by interacting with the cytoskeleton. Mitochondria are reported to be associated with three major cytoskeletal structures, which include microtubules, microfilaments of actin, and intermediate filaments [ 159 ]. Mitochondrial tubulin and microtubule associated proteins (MAPS) are reported to bind to porin or VDAC a component of the permeability transition pore [ 160 ]. The association of the cytoskeleton with VDAC could be biologically significant because the actin filament severing and capping protein gelsolin has been reported to modulate ΔΨ m by its interactions with VDAC [ 161 ]. Epithelial cell spreading results from the binding of integrins to the extracellular matrix which depends on the actin cytoskeleton [ 162 , 163 ]. The survival of epithelial cells depends on this interaction with the extracellular matrix, which if disrupted leads to a specific form of apoptosis called anoikis [ 164 ]. Actin filaments are necessary to cluster integrin receptors and proteins linked to their cytoplasmic domain into focal adhesion complexes [ 165 ]. These focal adhesion complexes provide a direct link between the extracellular matrix and the actin cytoskeleton. Anchorage-independent growth is a property of cancer cells, which may depend on the mitochondria based on evidence from the following studies. Long-term exposure of cells to ethidium bromide, an intercalation agent which inhibits mtDNA replication, results in the depletion of mitochondria. Mitochondria-depleted (ρ°) brain and breast tumor cells have been shown to lose their ability for anchorage-independent growth [ 166 ]. In addition, human ρ° cell lines derived from ovarian carcinoma, cervical carcinoma, and osteogenic sarcoma were demonstrated to be non-tumorigenic or poorly tumorigenic when administered subcutaneously to nude mice [ 167 ]. Taken from these reports is the interesting possibility that cancer cells maintain tension on the cytoskeleton via the contraction and expansion of mitochondria instead of binding to the extracellular matrix. Actin assembly and disassembly is regulated by the protein gelsolin. Since gelsolin is reported to prevent apoptotic mitochondrial changes by binding and closing VDAC, perhaps other diverse functions are modulated by interactions between the mitochondria and cytoskeleton. More recently, it was demonstrated that mitochondrial ROS production is stimulated by integrin induced changes [ 73 ]. Although integrin receptors are linked to the actin cytoskeleton, it is not clear whether the signal that is transduced to the mitochondria occurs via the cytoskeleton. Furthermore microtubules have also been implicated in the biogenesis of mitochondria based on the inhibition of mitochondrial mass increase and mtDNA replication caused by the microtubule-destabilizing drug colchicine and the estrogen metabolite 2-methoxyestradiol in mammalian cells [ 168 ]. Mitochondria biogenesis is reported to occur in the G 1 phase of the cell cycle but also starts in the late S phase [ 169 ]. Although mitochondria were not reported to be an upstream signaling event for generating cytoskeleton tension, physical effects of mitochondria such as shape changes and stretching or contraction could generate tension based on its association with the cytoskeleton. Changes in cytoskeleton tension may be mediated by mitochondria in response to estrogen. For instance, transmission electron microscopy (TEM) showed an increase in mitochondria size of MCF7 cells treated with E2 [ 14 ]. This change in mitochondrial size may generate mechanical forces that, in turn, may transduce a signal to the nucleus. Another possibility is that estrogen stimulated mitochondrial ROS may affect the elasticity of the actin network. Actin filaments are a prevalent feature of the cytoskeleton, which partially determine the overall mechanical strength of a cell. Thiol oxidation forms disulfide-bonded actin dimers resulting in interfilament cross-links [ 170 ]. Estrogen-induced oxidative stress has been recently reported to oxidatively modify cysteine residues of proteins [ 137 ]. Thus, estrogen-induced mitochondrial ROS could stimulate the formation of actin dimers that modulate cytoskeleton tension which in turn transduces a signal to the nucleus (Fig. 1 ). Mitochondrial proteolysis and peptides as signals The mitochondrial protein cytochrome c is known to be released to the cytosol where it initiates a signal for apoptosis. Given the role of cytochrome c the existence of other mitochondrial protein signaling molecules is a likely possibility that could mediate a diverse number of cellular processes including cell growth. It has been shown that mitochondria have the capability to export mitochondrial-matrix proteins to other cellular compartments such as the nucleus, peroxisome, endoplasmic reticulum, and secretory vesicles [ 171 ]. For example, mitochondrially transmitted factors (MTFs) are peptides derived from mitochondrial encoded proteins that are presented on the cell surface as minor histocompatability antigens. MTFs are derived from the mitochondrial encoded NADH dehydrogenase subunit 1 gene in murine and humans while rat MTFs are derived from the mitochondrial encoded ATPase 6 gene [ 172 , 173 ]. Although the synthesis and cell surface expression of MTFs was inhibited by the mitochondria specific protein synthesis inhibitor chloramphenicol, it is not clear whether post-translational modifications of mitochondria proteins are also responsible for MTFs; given that chloramphenicol is also an inhibitor of proteolysis in rat liver mitochondria [ 174 ]. Thus, mitochondria may serve as a subcellular compartment of proteolysis that generates signaling peptides that are exported to the cytosol. It is possible that proteolysis plays a significant role in mitochondrial protein processing because the chemical rhodamine 6G (R6G) inhibits matrix catalyzed processing of rat-liver mitochondrial precursors which include iron-sulfur protein, cytochrome c 1 , and core protein I of the cytochrome bc 1 complex; the α and β subunits of F1 ATPase and subunit IV of cytorochrome oxidase [ 175 ]. The molecular mechanism of R6G inhibition of protein processing was not identified, but it was proposed to be due to an interaction between R6G and a matrix protease. The import of proteins into mitochondria has been investigated in great detail while the process of export is minimally explored. A novel ATP-binding cassette (ABC) transporter, Mdl1, located in the inner mitochondrial membrane of yeast is required for the export of mitochondrial peptides with a molecular mass of 2100 to 600 daltons generated by proteolysis [ 176 ]. It was suggested that the export of peptides from the mitochondria may allow the mitochondria to communicate with its environment. This novel mode of communication may exist based on studies that demonstrate mitochondrial specific cleavage and export of cytokines. The cytokine IL-1β is localized in the mitochondria of LPS stimulated human peripheral blood monocytes and the 31 kDa form of IL-1β is reported to be cleaved into the 17 kDa mature active form within the mitochondria upon exposure to the HIV coat glycoprotein 120 [ 177 , 178 ]. Since macrophages secrete IL-1β by the ABC transporter, it is possible that proteins up to 17 kDa may be exported from the mitochondria [ 179 ]. In addition to IL-1β, mitochondrial Trx protein may also participate in a similar protein export mechanism. A truncated form of Trx, Trx80, is reported to be a potent mitogenic cytokine, however, it is not known whether Trx80 is derived from mitochondrial Trx [ 180 ]. Based on these reports it appears that cleaved proteins may be released as growth factors in response to an estrogen-induced rise in mitochondrial calcium that activates proteolysis. Thus, cleaved signaling peptides from the mitochondria may stimulate cell growth in an autocrine manner (Fig. 1 ). Several proteases are known to exist in the mitochondrial matrix such as ATP-dependent human Lon protease, ATP-dependent human Clp proteinase chain P (hClpP), and Ca 2+ dependent neutral protease [ 181 - 183 ]. Since endogenous proteolysis is a mechanism that regulates cell cycle progression, we postulate that E2-induced rise of mitochondrial calcium can activate calcium dependent proteases such as calpeptin and possibly hClpP. Once activated mitochondrial matrix proteins could be cleaved into bioactive forms that are exported to the cytosol. The heterogeneous association between ERα cleavage products and regulatory proteins has been suggested to play a role in physiological or pathological processes [ 184 ]. Low molecular weight ERα isoforms (~35–28 kDa) have been identified in mitochondria [ 16 ]. The 44-kDa protein related to the nuclear RXRα receptor is reported to be enzymatically cleaved and imported into the mitochondrial matrix [ 185 ]. It may interact with mitochondrial proteins or bind the organelle genome. Once in the cytosol mitochondrial clevage products may also regulate the function of various enzymes. For example, a truncated ERα 46-kDa protein in human endothelial cells mediated an acute activation of eNOS in response to a 15 min E2 (30 nM) treatment [ 186 ]. A significant finding from this report is that the truncated ERα 46-kDa mediates acute responses of estrogen rather than transcriptional responses in endothelial cells. Recently, it was reported that physiologic concentrations of E2 (<10 nM) induced NOS1 and activates the cGMP signal transduction pathway leading to sustained expression of Trx and MnSOD in human SH-SY5Y cells [ 187 ]. Thus, the acute activation of mtNOS by ERα cleavage products is yet another interesting possibility for redox signaling. Conclusion In summary, mitochondria are a major target of estrogen. It appears that mitochondria through its interaction with the cytoskeleton, export of cleaved signaling peptides, and/or generation of ROS may transduce signals to the nucleus for the activation of transcription factors, such as, AP-1, NF-κB, and CREB involved in the cell cycle progression of estrogen-dependent cells. These interactions between estrogen and mitochondria merit furture investigations, which may shed new light on the role of mitochondria in cell growth.	/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC548143.xml
314301	Gene Expression Signature of a Fibroblast Serum Response Predicts Cancer Progression	null	The idea that cancer cells go through a fateful transition that turns them into fast-growing, invasive, metastasizing tumors first surfaced in the early 1970s. During this conversion, blood vessels form around the tumor, providing a dedicated supply of blood to fuel the tumor's aggressive behavior. By the mid-1980s histological analysis revealed a similarity between the tumor “microenvironment” and that of a healing wound, prompting Harvard pathologist Harold Dvorak to describe cancer as a wound that does not heal. When the body sustains a wound, it coordinates an emergency response defined by rapid cell proliferation, invasion and “remodeling” of connective tissues and extracellular matrix (the network of proteins and molecules around cells), cell migration, and blood vessel formation (angiogenesis). These processes, which are restorative in normal wound healing, may promote cancer by supporting tumor formation, invasion, and metastasis. With no systematic method to measure the “wound-like” features in cancer, however, scientists have no way to evaluate the risk that a wound-healing genetic program may pose in cancer progression. A molecular understanding of the wound-healing process and its connection to cancer would provide insight into the nature of these similarities and perhaps provide molecular indicators of tumor progression. In an effort to create a framework for evaluating this relationship, Howard Chang and his colleagues at Stanford University developed a model to predict cancer progression based on the gene expression profile of a cellular response to serum in cell culture. Part of the problem with evaluating the physiological status of a tumor based on its genetic profile is that current techniques indicate only the expression, not the effect, of genes. To develop a strategy for interpreting biological outcomes from a gene expression profile, Brown's team modeled a physiological process by exposing cultured fibroblasts to serum—the soluble fraction of coagulated blood—and tracking gene expression. Serum is encountered in the body where blood leaks out of blood vessels (in essence, all the sites of injury) and is thought to be a major initiator of the wound response. Fibroblasts exist in the connective tissue of epithelial organs (which include the digestive tract, lungs, and mammary glands) and contribute to organ development, wound healing, inflammation, and a condition called fibrosis. (Fibrosis involves the same type of extracellular matrix remodeling seen in wound healing and cancer.) And fibroblasts can promote tumor formation and metastasis when mixed with epithelial cancer cells. Though fibroblasts from different sites in the body differ in their properties and gene expression profiles, Chang et al. found that they share a common expression pattern in response to serum. From this expression profile, the researchers identified a core group of genes—a genetic signature—associated with a serum response. Because many of the genes in the signature were known to be involved in various wound-healing processes—such as matrix remodeling, cell motility, and angiogenesis—Chang et al. used this signature as a surrogate marker to measure how much tumors may be like wounds. When they compared the wound-like genetic signature with the expression profiles of various clinical tumor samples, they found the signature was always present in certain cancers—prostate and liver-cell carcinomas—and occurred variably in others—breast, lung, and gastric carcinomas. In each of these three latter types of tumors, patients with tumors carrying the serum-activated wound-like genetic signature had a significantly increased risk of metastasis and death compared to patients with tumors that lacked the signature. Therefore, Chang et al. conclude that a wound-like phenotype is a general risk factor for metastasis and the aggressive behavior in many of the most common cancers. These results reveal a robust and useful similarity between the molecular programs in normal wound healing and tumor progression and metastasis. Although Chang et al. point out that their results do not indicate whether this fibroblast “fingerprint” is merely a marker for cancer progression or plays a role in orchestrating this pathway, they conclude that the genetic program activated in response to serum also contributes to tumor invasion and metastasis. This serum-response expression profile, the authors propose, provides a valuable new tool for predicting tumor behavior and determining a patient's prognosis. Genomics predicts tumor behavior	/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC314301.xml
314467	Looking from the Past to the Future	The history of one of the world's first electronic archives for scientific journals	JSTOR is successful for reasons its founders did not intend. Bill Bowen's inspired vision was of a solution to libraries' ever-voracious demands for space to house paper volumes. The idea was that libraries could save space by removing volumes available in electronic format. Few libraries have discarded the volumes digitised in JSTOR, but many libraries without the paper volumes have been able to offer their users access to the important journal runs JSTOR has digitised. Paper holdings have not decreased dramatically, but electronic holdings have increased. So a space-saving service became an access service. As an access service, JSTOR is a creation of its time. Understandable though the decision to use page images may have been eight years ago, future user-friendly access requires searching capabilities across full-text, which page images cannot supply. Likewise, the decision to digitise the back-runs of around 100—now 218—paper journals was a bold decision at the time, but the future for access to journal literature lies in electronic versions of thousands rather than hundreds of titles, both current and retrospective. When we reach that point, JSTOR will still have a valued place in the content on offer, but it is difficult to see JSTOR providing thousands rather than a few hundred titles. Its technical solutions and financial models look dated as both subscription-based and open-access publishers improve their services to authors and to readers. As the number of journal articles accessible over the networks increases, JSTOR will be seen as a small-scale pioneer from which we learned valuable lessons. Roger Schonfeld ends his very detailed description of JSTOR with a chapter on ‘Lessons Learned’, many of which are relevant to current access initiatives. The need for grant funding to launch any such initiative has to be accompanied by a sound business plan to ensure long-term economic viability. JSTOR has achieved that transition, and its success provides a model for others. Much of the credit must go to JSTOR's enterprising president, Kevin Guthrie, who found the quickest way through the maze of conflicting advice—much of which could have resulted in JSTOR's reaching a deadend—and convinced the library and publishing communities to buy into a product that was only a promise. Meeting user needs for easy access to high-quality content was the key to the fulfilment of that promise. JSTOR's public image is of quality in depth—long runs of core journals—and that image has to become the hallmark of the new open-access initiatives as they develop. It is understandable that some mistakes were made on the way. The difficulty that JSTOR financial planning had in coming to terms with consortial purchases delayed its growth as an access service. Although selling to consortia of academic libraries may not have improved JSTOR's financial position in the short-term, consortia are a route to spreading access and therefore securing longer-term financial stability (as the major publishers have realised through their ‘Big Deals’ in selling hundreds of journals to hundreds of libraries in a consortium). Some opportunities were also delayed—not lost—through too slow an adaptation of the JSTOR purchasing model for selling outside the United States, the United Kingdom being the exception. The UK deal was with JISC, the Joint Information Systems Committee of the UK Higher Education Funding Councils, acting more as a negotiating agent than a consortium, and this model could have been applied in other countries. More countries would have valued access to JSTOR earlier, but the approach to non-US deals had to be imaginative. For all vendors, there has to be an understanding of the political, social, economic, and educational structure of the country into which the product is being sold, an understanding that takes time to acquire but that pays dividends. Open-access publishers do not have to sell their product to users of their journals, but local knowledge is essential in ‘selling’ their services to authors. The globalisation of publishing has combined with the globalisation of the networks and with the globalisation of research to provide opportunities for high-quality research conducted outside North America and Western Europe to be published in peer-reviewed open-access journals more readily than in the traditional subscription-based journals. Roger Schonfeld's book draws out many of the significant points about JSTOR's place in the history of electronic publication through a minute examination of the process leading to JSTOR as it is today. There is so much detail in the book that the reader may feel that its comprehensiveness cannot be questioned, but one small omission of which I have personal knowledge makes me question the value of so much detail. The omission concerns the interest by my institution, University College London, in joining JSTOR before the JISC deal was considered. Not a detail of world-shattering significance, but it does illustrate the fact that outside the United States, as well as within, the early interest in JSTOR came from individual institutions rather than from consortia. I sympathise with Roger Schonfeld in attempting to write such a comprehensive history, but what is the point of appearing to be comprehensive when comprehensiveness is an impossible goal? Would a briefer history have been just as valuable? Leaving aside quibbles and caveats about the book and about JSTOR, this remains a fascinating and instructive history of an important and ground-breaking initiative. Bill Bowen's vision may not have developed in quite the way he expected, but the ‘bottom-line’ is that the vision did become a successful reality. The problem of ever-expanding libraries has not gone away in the ten years since JSTOR was conceived, but the ultimate solution—the availability of electronic content—has become closer, and JSTOR's success has encouraged others to develop services that are more in accord with 2003 than 1993. One lesson Roger Schonfeld does not draw out is the pace of change in electronic publishing, and if so much has been achieved since 1993, what promise is held out by the next ten years'!	/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC314467.xml
314473	Mechanism Suggests How HIV Protein Disrupts Immune Cell Migration	null	One of the cornerstones of immune system function is movement. When word spreads that a virus has entered the body, chemical signals tell lymphocytes to proliferate and travel to the site of infection. Efforts to combat HIV have focused on understanding how the virus disrupts this immune response in the hopes of developing drugs to block its replication as well as vaccines to control the virus itself. Toward this end, scientists are investigating how each of the virus's nine genes—which all appear to have multiple functions—contribute to HIV infection. When HIV infects a cell, viral enzymes copy its RNA genes into DNA, which can then invade the infected cell's chromosomes. The viral DNA might lay dormant or it might use the cell to reproduce more viruses, which go on to infect other cells. The course of infection is determined by interactions between circulating T cells and antigen-presenting cells (cells that present evidence of infection), like macrophages, which may unwittingly aid the virus by transferring it to the T cells. Macrophages, for example, produce proteins that tell T cells to come check out an infection. A viral protein called Nef sparked intensive research after observations that patients with a rare strain of HIV lacking Nef took a very long time to develop AIDS symptoms. Nef has been linked to molecules involved in macrophage- and other antigen-signaling pathways and may use the molecules to appropriate these pathways for its own ends—enhancing virulence by facilitating viral replication. How Nef does this is not entirely clear. Now Jacek Skowronski and his colleagues at Cold Spring Harbor Laboratory in New York have identified the key molecules that Nef enlists to coopt the signaling machinery of immune cells. To understand how this might happen, biochemically speaking, Skowronski's lab first needed to determine which molecules Nef associates with. An adaptor protein, Nef does not directly catalyze reactions, but binds to enzymes that do. The researchers identified two proteins, DOCK2 and ELMO1, that form a complex with Nef. DOCK2 regulates enzymes, called Rac1 and Rac2, that are required for normal lymphocyte migration and antigen-specific responses. ELMO1 has also been shown to help DOCK2 activate Rac. Because DOCK2 activates Rac as part of two different signaling pathways—one activated by the T cell receptor, which mediates T cell activation, and one by a chemokine receptor, which controls T cell migration—the researchers investigated whether Nef could affect these important pathways by modulating Rac activity. They found that Nef in fact activates Rac by binding to the DOCK2–ELMO1 complex. And they went on to show that HIV uses these components of the chemokine receptor pathway to disrupt T cell migration. To generate an effective immune response, it is crucial that T cells travel to sites within lymphatic tissues where they interact with other lymphocytes. By inhibiting T cell migration, the researchers propose, Nef prevents these critical interactions, thereby providing a mechanism for stifling the immune response. These results, the authors argue, provide the biochemical evidence that Nef targets a protein “switch” that can interfere with important aspects of T cell function. In this way, Nef subverts the immune response pathways controlled by receptors on the surface of T cells to effectively disarm the immune system and turn T cells into viral replication factories. Understanding how Nef interacts with these proteins to spread infection could lay the foundation for valuable new therapies aimed at inhibiting and arresting HIV infection by blocking Nef-mediated effects. Pathway by which Nef disrupts T cell migration	/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC314473.xml
544590	HIV-1 encoded candidate micro-RNAs and their cellular targets	MicroRNAs (miRNAs) are small RNAs of 21–25 nucleotides that specifically regulate cellular gene expression at the post-transcriptional level. miRNAs are derived from the maturation by cellular RNases III of imperfect stem loop structures of ~ 70 nucleotides. Evidence for hundreds of miRNAs and their corresponding targets has been reported in the literature for plants, insects, invertebrate animals, and mammals. While not all of these miRNA/target pairs have been functionally verified, some clearly serve roles in regulating normal development and physiology. Recently, it has been queried whether the genome of human viruses like their cellular counterpart also encode miRNA. To date, there has been only one report pertaining to this question. The Epstein-Barr virus (EBV) has been shown to encode five miRNAs. Here, we extend the analysis of miRNA-encoding potential to the human immunodeficiency virus (HIV). Using computer-directed analyses, we found that HIV putatively encodes five candidate pre-miRNAs. We then matched deduced mature miRNA sequences from these 5 pre-miRNAs against a database of 3' untranslated sequences (UTR) from the human genome. These searches revealed a large number of cellular transcripts that could potentially be targeted by these viral miRNA (vmiRNA) sequences. We propose that HIV has evolved to use vmiRNAs as a means to regulate cellular milieu for its benefit.	Findings Initially discovered in Caenorhabditis elegans as regulators of temporal control of post-embryonic development [ 1 , 2 ], miRNAs are small RNAs involved in the specific regulation at the post-transcriptional level of cellular genes in various organisms such as flies, plants and mammals [ 3 , 4 ]. To date, more than two hundred human miRNAs have been described [ 5 ]. Structurally, miRNAs are 21 to 25 nucleotide RNAs derived from the maturation of a hairpin precursor transcript which can be encoded by the 3' untranslated region of genes, introns of genes, or by specific chromosomal regions composed of tandem clusters of miRNA sequences. Precursor RNAs for miRNAs are structured as imperfect RNA hairpins containing mismatches and bulges. In mammalian cells, the maturation of miRNA occurs in two steps consecutively involving two cellular RNase III proteins, the nuclear Drosha and the cytoplasmic Dicer [ 6 ]. Accordingly, a miRNA precursor is specifically recognized in the nucleus by Drosha which cleaves the RNA to release an imperfect stem-loop structure of ~ 70 nucleotides, the pre-miRNA. This structure is then exported by exportin-5 into the cytoplasm and further cleaved there by Dicer into corresponding imperfect RNA duplexes of 21 to 25 nucleotides, the miRNA [ 7 ]. Mechanistically, either one of the two strands of the mature miRNA can be incorporated into the RNA-induced silencing complex (RISC). miRNA-armed RISCs can then specifically recognize and interact via imperfect complementarity with RNA targets to induce repression of translation and (less frequently) mRNA cleavage. The precise molecular mechanism of translational silencing remains unclear; however, in such instances, it has been observed that protein synthesis is inhibited while the stability of the mRNA is not altered [ 8 - 10 ]. Recently, in addition to plants, insects, invertebrate animals, and mammals, Pfeffe r et al . identified virus-encoded miRNA sequences in Epstein-Barr virus (EBV) infected cells [ 11 ]. They reported that EBV encodes five miRNAs each capable of regulating viral genes involved in latency as well as modulating the expression of host cell genes. Thus, it would appear that EBV has evolved to use the miRNA pathway for its replicative benefit. To query whether this stratagem might also be employed by other viruses, we have analyzed putative miRNA-encoding capacity of HIV-1. We wondered if HIV-1 maintains RNA structures that resemble pre-miRNAs. As a proof-of-principle, we examined pre-miRNA structures in one specific example of HIV-1, the genome of the pNL4-3 molecular clone. Because HIV has well-described stem-loops such as TAR ( t rans- a ctivation responsive RNA) and RRE ( R ev- r esponsive e lement) [ 12 ], one might think that pre-miRNA structures would be prevalently found in this virus. However, when we set search parameters to include RNA structure of ~ 70 nucleotides in total size with an imperfect stem of 21 to 25 base-pairs, only a few thermodynamically reasonable candidates were revealed. Using a new scanning method StemEd [ 13 ], we uncovered 5 pre-miRNA candidates. As shown in figure 1a , these sequences (#1 to #5) are discretely separated in different regions of the HIV genome: near TAR, in capsid gag , near the gag-pol frameshift, in the nef gene, and in the 3'LTR [ 14 , 15 ]. The corresponding predicted folding for each candidate and their deduced mature virally-encode miRNA (vmiRNA) sequences are presented in figure 1b . Figure 1 Sequences and localization of HIV-encoded miRNA candidates. a) Locations for 5 predicted pre-miRNAs candidates in the pNL4-3 genome are shown. b ) The folded structures of the 5 viral pre-miRNAs from pNL4-3 (Accession Number AF 324493) [17] are illustrated. Folded pre-miRNAs and their corresponding predicted mature viral miRNA (red) are listed. Nucleotide positions (where 1 is the initiation of transcription) in the pNL4-3 genome are presented in the right column. The 5 HIV-1 encoded pre-miRNA candidates can in principle yield 10 mature vmiRNAs. To ask, whether these putative vmiRNAs, if expressed in infected cells, could be used by HIV-1 to modulate host cell gene expression profiles (i.e. suppress the translatability of cellular mRNAs), we checked each vmiRNA sequence against a 3'UTR database for human genes. Because the exact rules governing suppression of translation based on miRNA complementarity to 3' UTR remain unclear, we collected all "hits" that had 6 or fewer mismatches with an additional constraint that the 5'-most nucleotide of the vmiRNA cannot be mis-matched with the target sequence. Surprisingly, based on the above criteria, a very large number of cellular targets were found (Figure 2 ). On average, it is suggested that each vmiRNA could target 50 to 100 cellular RNAs. If all 10 vmiRNAs were functionally competent, this would argue that HIV-1 could potentially modulate the expression of 500 to 1,000 cellular transcripts using this mechanism. Intriguingly, in the setting where a limited number of mis-matches are permitted, the same vmiRNA apparently can target functionally distant cellular proteins such as IkB-kinase-β and proteasome 26S subunit or macrophage colony-stimulating factor M-CSF and CDC42 effector protein 1 (Figure 2 ). This suggests that vmiRNAs could pleiotropically affect the expression pattern of cellular proteins. Figure 2 Potential cellular targets for each of the vmiRNAs. The two deduced mature vmiRNAs predicted from each precursor miRNA are shown. The mature vmiRNA sequences were individually searched against a database of human 3'UTRs using imperfect complementarity criteria as described in the text. The number of potential candidate cellular RNA targets is enumerated. Most of the cellular targets are incompletely characterized expressed sequence tag (EST) clones, with a subset of targets being known genes. For each predicted vmiRNA, we list two examples of known cellular gene targets at the right. A full list of targets is available upon request. Here, we introduce the concept that the HIV genome could reasonably encode 5 candidate pre-miRNAs. We further suggest that a large number of cellular transcripts could potentially be targeted if these 5 pre-miRNAs were processed into 10 predicted mature vmiRNAs (Figure 2 ). Studies are in progress to verify experimentally the expression of our candidate vmiRNAs in HIV-1 infected cells. If HIV-1 encoded vmiRNA candidates can be shown to be functional, their action could, in part, explain the frequently observed landscape changes in host cell gene expression profiles during HIV-1 infection as revealed by micro-array studies [16]. We are also currently examining how vmiRNAs might additionally affect HIV-1 gene expression. Competing interests The authors declare that they have no competing interests. Authors' contributions YB and KTJ conceived of the ideas and wrote the paper. SYL did the computation work for the study. MLY participated in the discussion of the data.	/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC544590.xml
524516	Effect of smoking on vitamin A, vitamin E, and other trace elements in patients with cardiovascular disease in Bangladesh: a cross-sectional study	Background Data regarding the impact of cigarette smoking on trace elements are scarce and inconsistent. In this study, we evaluated the effect of smoking on serum concentrations of trace elements among adult males with heart disease. Methods This cross-sectional study included 100 adults hospitalized with heart disease in Bangladesh. The major variables of interest included mean serum concentrations of trace elements and proportion of subjects with bacterial growth on throat swab culture. Results Smokers had significantly lower serum concentrations of retinol, alpha-tocopherol, selenium, and zinc and increased concentrations of copper. Throat swab cultures were more often positive for Streptococcus β- hemolyticus in smokers than controls. Conclusions Smoking decreases serum concentrations of trace elements. Smoking control programs are needed in Bangladesh to improve health and nutrition of the people who are already nutritionally deficient.	Introduction Smoking is a widely accepted practice in Bangladeshi men and is associated with socialising, sharing, and male identity [ 1 ]. According to an earlier cross-sectional study, approximately 50% of males and 3% of females are tobacco smokers in Bangladesh [ 2 ]. Although smoking is a recognized risk factor for several diseases including emphysema, chronic bronchitis, cardiovascular diseases, and cancer [ 3 - 5 ], very little is known about the nutritional consequences of smoking. In animal models, administration of benzo(a)pyrene, a constituent present in cigarette smoke induced vitamin A depletion [ 6 ]. Vitamin A deficiency per se causes emphysema. Some other trace elements, such as iron, zinc, and vitamin E were found to be deficient among healthy smokers compared to non-smokers. However, the available data are inconsistent regarding the effect of smoking on trace elements. In this study, we documented the effect of different doses of smoking on trace elements among hospitalized patients with heart disease in Bangladesh. Methods and Materials A cross-sectional study was conducted among 100 male patients admitted to the National Institute of Cardiovascular Disease (NICVD), Dhaka, Bangladesh from January through December 1998, after obtaining informed consent from the participants. The study protocol was reviewed and approved by the Human Subjects Ethical Committee of the NICVD. The patients who had a history of smoking 10 or more cigarettes per day were considered smokers, and those who never smoked were controls. All the patients, including controls, were admitted with heart disease. The study did not include females in this study because smoking is not a norm among females in that society. All the smokers and the first 20 of the non-smokers who met the selection criteria were eligible for the study. This study included only heavy smokers who smoked at least 10 sticks per day, and excluded mild or casual smokers to leave a buffer zone of comparison between smokers and non-smokers. Patients were stratified according to their smoking status as follows: Control ( n = 20), non-smokers; Grade I ( n = 20), 10 – 15 sticks/day; Grade II ( n = 30), 16 – 20 sticks/day; Grade III ( n = 30), 21 and more sticks/day. Body weight and height were measured at admission. Weight (kg) was measured to the nearest 0.1 kg, with participants wearing light clothing and no shoes and using a beam balance with non-detachable weights. Height (cm) was measured with a stadiometer to the nearest 0.5 cm. Body mass index (BMI) was calculated using the following formula: weight (kg) / height (m) 2 . Five ml of venous blood was obtained from each patient at admission in a test-tube, which was wrapped with aluminum foil to avoid degradation of vitamin A in light. Serum samples were separated by centrifugation, and kept stored at -20°C until further analysis. For the analyses of trace elements, serum samples were stored at separate ion-free vials. Throat swab cultures were detected for the growth of bacteria. Laboratory Methods Serum retinol (vitamin A) and serum α-tocopherol (vitamin E) concentrations were determined by high performance liquid chromatography (HPLC) according to Bieri et al. [ 7 ]. In brief, serum retinol and α-tocopherol were extracted with hexane after deproteinization with absolute ethanol containing retinyl acetate and α-tocopherol acetate (Sigma Chemical Co., St Louis, MO, USA) as internal standards for retinol and α-tocopherol respectively. Retinol and α-tocopherol were separated by HPLC (model PU 4010; Pye-Unicam) on a reverse-phase C 18 column using methanol-water (97.5:2.5, v/v) as the mobile phase. Coefficient of variation (CV) values of ten replicates from a pooled serum sample for retinol and α-tocopherol were 2.3 and 3.3% respectively. Serum zinc was measured by flame atomic absorption spectrophotometry (AAS, Analyst 800, Perkin-Elmer, Norwalk, CT, USA) using a modification of the method described by Kirgbright [ 8 ]. Serial replication of aliquots from a pooled serum sample and quality control sera were used to check the precision and accuracy of the analytical methods. The within-run CV for zinc in a pooled serum sample was between 2.2 and 4%, based on six to seven samples in each of the five runs. Serum concentrations of copper and selenium were measured by the AAS method mentioned above. Statistical Methods Data were analyzed using SPSS for windows, version 11.0 (SPSS Inc., Chicago, IL). Descriptive statistics of the major variables of interest were calculated to determine the distribution of the data. All the variables except serum iron concentrations were normally distributed. Variables between smokers and non-smokers were compared by Student t-test for continuous data and by Chi-square test for categorical data. A probability level of 5% was considered statistically significant. Results Of the 110 patients enrolled, 10 dropped out; eight had incomplete data, one withdrew early, and one left the hospital without notice. The mean ± SD age of the study patients was 43.67 ± 6.79 y (range, 28 – 57 y). The groups did not differ significantly in terms of age and anthropometric measurements (Table 1 ). Table 1 Baseline characteristics of the study subject Control (Non-smoker) n = 20 Smoker All n = 100 Grade I 10–15 sticks/d n = 20 Grade II 16–20 sticks/d n = 30 Grade III 21+ sticks/d n = 30 Age (year) 41.60 ± 8.30 44.40 ± 7.52 43.57 ± 5.85 44.67 ± 6.06 43.67 ± 6.79 Weight (kg) 60.95 ± 3.32 63.25 ± 3.40 62.50 ± 3.64 62.90 ± 2.81 62.46 ± 3.34 Height (m) 1.64 ± 0.03 1.65 ± 0.03 1.65 ± 0.03 1.63 ± 0.04 1.65 ± 0.03 BMI 22.32 ± 1.35 23.11 ± 1.31 22.90 ± 1.63 23.17 ± 1.10 22.91 ± 1.38 Values are mean ± SD . Serum retinol concentrations were below normal (0.70 μmol/L) among all smokers and the majority (60%) of the controls. Table 2 shows that the smokers who smoke 16 sticks or more cigarettes per day had significantly lower concentrations of serum retinol compared with controls. Percentage decrease of alpha-tocopherol was most striking of all the trace elements. Zinc concentrations did not change among grade I and grade II smokers but decreased among grade III smokers compared with controls. Concentrations of copper increased but selenium decreased among smokers than controls. Table 2 Effect of smoking on serum concentrations of retinol, alpha-tocopherol and other trace elements Trace element (μmol/L) Control (Non-smoker) n = 20 Smoker Grade I 10–15 sticks/d n = 20 Grade II 16–20 sticks/d n = 30 Grade III 21+ sticks/d n = 30 Retinol .72 ± .06 .66 ± .12 .45 ± .09* .28 ± .07* Alpha-tocopherol 13.5 ± 1.6 10.0 ± .7 8.6 ± 1.9* 7.9 ± 1.2*** Copper .44 ± .25 .46 ± .16 .61 ± .19* .62 ± .27* Selenium .013 ± .001 .004 ± .001* .003 ± .001* .003 ± .001* Zinc .46 ± .23 .46 ± .15 .46 ± .20 .42 ± .15* Values are mean ± SD . * P < .05; P < .01; * P < .001 compared with the control. A significantly higher proportion of smokers compared with controls had bacterial growth on their throat cultures, mostly due to Streptococcus β- hemolyticus (Table 3 ). Table 3 Effect of smoking on bacterial growth on throat swab culture Organism Control (Non-smoker) n = 20 Smoker Grade I 10–15 sticks/d n = 20 Grade II 16–20 sticks/d n = 30 Grade III 21+ sticks/d n = 30 Streptococcus β hemolytica 2.8 70.5 71.4 72.5 Aerobacter aerogenes 12.2 25.4 26.7 27.0 No growth 85.0 4.1 1.9 1.0 Values are percentages. P < .001 for all values of smokers compared with the control. Discussion In this study, adult male smokers with heart disease had significantly decreased serum concentrations of retinol, alpha-tocopherol, and selenium, and increased concentrations of copper, compared to non-smokers. Depression of trace elements in blood was more with increasing doses of smoking. In a study in Turkey, plasma selenium, copper, zinc and iron concentrations, and the activities of related erythrocyte antioxidative enzymes copper-zinc superoxide dismutase (Cu-Zn SOD), catalase, and glutathione peroxidase (GSH-Px) were measured in tobacco smokers and compared with those of nonsmokers [ 9 ]. Plasma thiocyanate levels were measured as an index of smoking status. While plasma copper concentration and erythrocyte Cu-Zn SOD activity were significantly higher, plasma selenium concentration and erythrocyte GSH-Px activities were significantly lower in tobacco smokers than in nonsmokers. We did not measure antioxidative enzyme levels in blood; however, our study had consistency with earlier findings of decreased serum concentrations of selenium and increased concentrations of copper among smokers. Kocyigit et al. [ 9 ] did not observe any significant effect of smoking on zinc or iron status. Our observation of a significantly depressed zinc status only among heavy smokers (those who smoked 21 or more sticks per day) compared to non-smokers was consistent with findings of Uz et al. in Turkey [ 10 ]. Several studies documented that smoking may increase oxidative stress and impair oxidant defense system [ 11 ]. Serum selenium glutathione peroxidase, glutathione reductase, and extracellular superoxide dismutase activities were found lower in smokers than in non-smokers. Serum ascorbic acid and folate concentrations were lower in smokers than in non-smokers, whereas serum thiobarbituric acid-reactive substances (TBARS) were higher. However, Kim et al. (2003) did not observe any effect of smoking on serum copper, iron, and magnesium concentrations [ 11 ]. In a later study, Kim et al. (2004) further evaluated the influence of short– and long-term cigarette smoking on blood antioxidant status among Korean teenage girls (aged 14 to 18 y) and adult males (aged 36 to 51 y) [ 12 ]. Extracellular superoxide dismutase activities and concentrations of serum vitamin C and folate were lower in both short-term and long-term smokers. Serum copper concentrations were higher only among long-term smokers compared to non-smokers. In our study, we observed increased serum concentrations of copper among grade II and grade III smokers (those who smoked 16 or more sticks per day) but not among grade I smokers (those who smoked 10–15 sticks per day), as compared to non-smokers. Both the studies suggest that probably an increasing dose of smoking modify serum copper status more compared to those who smoke less or do not smoke at all. However, cigarette smoking, irrespective of dose or duration, had negative effects on antioxidant status in the Korean study [ 12 ]. Increasing evidence suggests that smoking is a causal factor for coronary heart disease and stroke. In a prospective study in Japan [ 4 ], 19,782 men and 21,500 women aged 40 to 59 years who were free of prior diagnosis of stroke, coronary heart disease, or cancer and reported their smoking status were followed. During a 461,761 person-year follow-up, relative risks (95% CIs) for current smokers compared with never-smokers were 1.27 (1.05 to 1.54) for total stroke, 0.72 (0.49 to 1.07) for intraparenchymal hemorrhage, 3.60 (1.62 to 8.01) for subarachnoid hemorrhage, and 1.66 (1.25 to 2.20) for ischemic stroke. One of the limitations of our study is that it is difficult to establish any causal association of heart disease and deficiency of trace elements or increased isolation of Streptococcus β- hemolyticus among our study subjects, because it is a cross-sectional study. However, epidemiologic evidence has suggested a modifying role for antioxidant micronutrients, including tocopherols and carotenoids, in atherosclerosis and heart disease. In an experimental study, Handelman et al. (1996) exposed freshly obtained human plasma to the gas phase of cigarette smoke to assess its effects on tocopherols, carotenoids, and retinol. Exposure to cigarette smoke led to the depletion of most of the lipophilic antioxidants in human plasma [ 13 ]. In addition to the impact on health, tobacco smoking represents a major economic burden for impoverished Bangladeshis. Average male cigarette smokers spend more than twice as much on cigarettes as per capita expenditure on clothing, housing, health and education combined. The typical poor smoker could easily add over 500 calories to the diet of one or two children with the daily tobacco expenditure [ 14 ]. It may be noted that most of the study subjects were undernourished, as indicated by an average BMI of 23. Strong tobacco control measures are needed in the context of Bangladesh to decrease tobacco expenditures and thus significantly increase resources and improve health and nutrition of the people. Conclusion This study demonstrated that increasing amount of cigarette smoking negatively impact serum concentrations of retinol, alpha-tocopherol, selenium, and zinc. Cigarette smoking may act as an important adjunct to the deficiency of those trace elements in a population who are already nutritionally compromised. Competing interests The authors declare that they have no competing interests. Authors' contributions SKB participated in the design of the study and collected the samples. AKM performed the statistical analysis and drafted the manuscript.	/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC524516.xml
524502	Serum S100B in primary progressive multiple sclerosis patients treated with interferon-beta-1a	S100B belongs to a family of calcium-binding proteins implicated in intracellular and extracellular regulatory activities. This study of serum S100B in primary progressive multiple sclerosis (PPMS) is based on data obtained from a randomized, controlled trial of Interferon β-1a in subjects with PPMS. The key questions were whether S100B levels were associated with either disability or MRI findings in primary progressive MS and whether Interferon β-1a has an effect on their S100B levels. Serial serum S100B levels were measured using an ELISA method. The results demonstrated that serum S100B is not related to either disease progression or MRI findings in subjects with primary progressive MS given Interferon β-1a. Furthermore there is no correlation between S100B levels and the primary and secondary outcome measures.	Introduction S100B belongs to a family of calcium-binding proteins implicated in intracellular and extracellular regulatory activities [ 2 ]. Intracellularly, it exhibits regulatory effects on cell growth, differentiation, cell shape and energy metabolism. Extracellularly, S100B stimulates neuronal survival, differentiation, astrocytic proliferation, neuronal death via apoptosis, and stimulates (in some cases) or inhibits (in others) activity of inflammatory cells. Several studies suggest that S100B has a role in the pathogenesis of multiple sclerosis (MS). Phenotypically and functionally similar T cells specific against S100B can be detected in the peripheral blood of MS patients making S100B a putative candidate auto-antigen in MS [ 15 ]. Furthermore, S100B may act as a cytokine [ 2 , 10 , 11 ] and in vitro studies show that, at high levels, S100 can induce the neuronal expression and secretion of pro-inflammatory interleukin-6. In addition, elevated levels of S100B have been detected in the cerebrospinal fluid (CSF) of MS patients during acute phases or exacerbations of the disease [ 10 ] and it has therefore been proposed that elevated S100B may be indicative of active cell injury [ 11 ]. Interferon-β (IFN-β) is effective in reducing relapse rate in relapsing-remitting [ 6 , 14 , 17 ] and secondary progressive MS [ 3 ] but the mechanisms behind the beneficial action of IFNβ are not fully understood. Two potential sites of action are on cytokine production [ 1 , 4 , 12 ] and on the entry of leukocytes into the CNS [ 8 , 9 , 16 , 18 ]. In this clinically negative phase II study [ 7 ], we assessed the effect of IFNβ-1a on serum levels of S100B at 3-month intervals in subjects with primary progressive MS (PPMS). The key questions were whether serum S100B levels correlated with disability or MRI findings in patients with PPMS, and whether IFN-β has an effect on levels of serum S100B. Methods Patients and examination Fifty patients with PPMS were recruited in a phase II trial of IFNβ-1a (Avonex ® , Biogen) and were assessed three monthly over a study period of 2 years. Fifteen of these patients were treated with IFNβ-1a 30μg intramuscularly (im) weekly (IFN30), 15 received IFNβ-1a 60μg im weekly (IFN60) and 20 with placebo. IFNβ-1a was reduced to half dose in 5 subjects receiving 60μg im weekly, and in 2 subjects receiving IFNβ-1a 30μg im weekly. Seven subjects withdrew from treatment [ 7 ] (see Figure 1 ). Figure 1 Fifty subjects with PPMS were randomised in a phase II trial of Interferon β-1a and were assessed 3 monthly over a 2-year study period. n = number of subjects with PPMS Neurological examination was performed at each visit and disability was measured using Kurtzke's expanded disability status scale (EDSS). Progression was defined as a sustained (3 months apart) increase of at least 1.0 on the EDSS scale between 0 to 5 and 0.5 for subjects with EDSS score of 5.5 and above. Fourteen healthy subjects served as controls. All subjects provided informed consent prior to their inclusion in the study. This study was approved by the ethics committee and has therefore been performed with the ethical standards laid down in the 1964 Declaration of Helsinki. MR imaging and analyses MRI was performed at baseline and 6 monthly for 2 years. Only baseline and year 2 data were included in this study. Brain and spinal cord atrophy, ventricular volume, T1 and T2 lesion load were measured as described elsewhere [ 7 ]. Serum S100B levels Serum samples were centrifuged and stored at -20°C. Serum S100B levels were quantified using a modified ELISA method as previously described by Green et al. [ 5 ]. Ninety-six-well plates were coated with 100μl 0.05 M carbonate buffer containing 10μl monoclonal anti-S100B (Affiniti Research Products, Exeter, UK). The plates were washed with 0.67 M barbitone buffer containing 5 mM calcium lactate, 0.1% BSA and 0.05% Tween and then blocked with 2% BSA and washed again. Diluted serum (1:1) in 0.67 M barbitone buffer containing 5 mM calcium lactate was added in duplicate. After incubation and wash 0.1% HRP conjugated polyclonal anti-S100B (Dako, Copenhagen, Denmark) was used as detecting antibody. The OPD colour reaction was stopped with 1 M hydrochloric acid and the absorbance read at 492 and 405 nm. The antigen concentration was calculated against a standard curve ranging from 0.01 to 2.5 ng/ml. Statistical analyses Median, interquartile range and significance of group differences (Mann-Whitney U tests) were evaluated. Changes of serum level over time were examined using variance components regression models of serum response variable on time as predictor, with random subject-specific intercepts and fixed common slopes. Curvature was assessed using a quadratic term in time; modification of curve over time by treatment was assessed using additional terms for treatment and treatment by time interaction in the model. Two sets of treatment terms were used: i) indicators of assigned weekly dose ii) average weekly dose over follow-up (including changes to dose regime) as continuous variable. Modification of the curve over time by MRI variable values were similarly examined using terms for MRI variable and MRI variable by time interaction. Direct associations between serum level and MRI/clinical variables were examined by regression models of 24 month serum on 24 month MRI variable, adjusting for baseline serum and MRI values (this type of model takes into account change from baseline), with additional terms for treatment and treatment by MRI variable interaction, the latter to assess possible modifications of the relationship by treatment. Software used were the SPSS software package (version 11.0 for Windows) and Stata 7.0 (Stata Corporation. Stata Statistical Software: Release 7.0. College Station, Texas, USA). Results Serum S100B between subjects with PPMS and controls The median and interquartile ranges for all subjects are described in Table 1 . There were no significant differences between any of the groups in relation to age. When comparing S100B levels at baseline of subjects with PPMS and controls, the difference was not statistically significant (p = 0.3). Table 1 Age and serial serum S100B levels expressed as median (interquartile range). n = number of subjects; mo, months; N/A, non-applicable. Control (n = 14) Placebo (n = 20) IFN30 (n = 15) IFN60 (n = 15) Male:Female 6:8 n = 14 15:5 n = 20 10:5 n = 15 7:8 n = 15 Age (years) 32 (29–44) n = 14 43 (36–51) n = 20 51 (39–53) n = 15 52 (43–54) n = 15 S100B-0mo 0.08 (0.08–0.10) n = 14 0.09 (0.02–0.10) n = 20 0.06 (0.04–0.10) n = 12 0.07 (0.04–0.10) n = 14 S100B-3mo N/A 0.08 (0.05–0.10) n = 20 0.06 (0.05–0.10) n = 15 0.07 (0.04–0.10) n = 14 S100B-6mo N/A 0.10 (0.03–0.20) n = 18 0.07 (0.04–0.10) n = 15 0.07 (0.04–0.10) n = 14 S100B-9mo N/A 0.08 (0.03–0.10) n = 18 0.05 (0.02–0.10) n = 15 0.06 (0.04–0.10) n = 14 S100B-12mo N/A 0.07 (0.03–0.10) n = 17 0.08 (0.05–0.10) n = 15 0.08 (0.04–0.10) n = 14 S100B-15mo N/A 0.07 (0.02–0.10) n = 19 0.07 (0.04–0.10) n = 14 0.09 (0.04–0.10) n = 14 S100B-18mo N/A 0.07 (0.04–0.10) n = 18 0.06 (0.02–0.09) n = 13 0.09 (0.03–0.20) n = 14 S100B-21mo N/A 0.06 (0.05–0.10) n = 18 0.08 (0.04–0.10) n = 14 0.08 (0.04–0.20) n = 13 S100B-24mo N/A 0.07 (0.02–0.10) n = 18 0.07 (0.05–0.10) n = 15 0.06 (0.04–0.10) n = 13 Serum S100B change over time There was no change over time in the serum S100B levels. The shape of the serum trajectory did not vary between the treatment regimes, i.e. placebo vs. IFN30 vs. IFN60. Serum S100B versus Clinical and MRI parameters (Table 2 ) There was no evidence that the 24-month serum S100B values were associated with either changes in the T1 or T2 loads, or ventricular or cord volumes at 24 months, after adjusting for the baseline values of each subject. There was no correlation with disease progression on the EDSS. There was also no evidence that these relationships were modified by treatment assignment (intention-to-treat analysis) (Table 2 ) or the overall average dose, which included the changes to treatment regime (non-intention-to-treat analysis) (Table 2 ). Table 2 Serum S100B versus Clinical and MRI variables. Estimated mean change in 24-month serum S100B associated with unit increase in mean value of T1 and T2 lesion load, ventricular and spinal cord volume, adjusted for baseline values of both serum S100B and of MRI parameters. Baseline adjustment ensures that the coefficient assesses the 'effect' of the 24-month MRI parameters value relative to its baseline. * Test of treatment interactions with row variable. Variable Coefficient P-value 95% Confidence Interval (CI) P-value for treatment modification*: Assignment average dose 24 month T1 load -4 × 10 -6 0.35 -1 × 10 -5 , 4 × 10 -6 0.76 0.59 24 month T2 load -3 × 10 -6 0.16 -7 × 10 -6 , 1 × 10 -6 0.57 0.89 24 month ventricular volume 7 × 10 -7 0.75 -3 × 10 -6 , 5 × 10 -6 0.46 0.24 24 month cord volume -2 × 10 -3 0.54 -9 × 10 -3 , 5 × 10 -3 0.58 0.88 Discussion These results suggest that serum S100B levels in patients with PPMS were not affected by intramuscular IFNβ-1a and that there was no observable change in S100B over time. Furthermore, we did not observe any correlation between S100B levels and clinical disability or between S100B and quantitative MRI measures. This study therefore suggests Although there is evidence that S100B elevation in MS is related to inflammatory activity [ 10 , 11 , 13 ], this study has shown that S100B was not sensitive to disease progression in PPMS. This supports the view that PPMS is less inflammatory than other forms of MS and that serum S100B would be ineffective as a surrogate marker of disease progression in this subgroup. It would be valuable to identify surrogate markers of clinical progression in PPMS to aid the development of effective therapeutic intervention, since clinical trials with a disability endpoint are very large and resource consuming. It is possible that such markers would need to be less related to acute inflammation and more dependant on other neuropathology such as axonal loss and regeneneration.	/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC524502.xml
554783	Genetic linkage analysis supports the presence of two susceptibility loci for alcoholism and heavy drinking on chromosome 1p22.1-11.2 and 1q21.3-24.2	Background In order to confirm a previous finding of linkage to alcoholism on chromosome 1 we have carried out a genetic linkage study. Methods DNA from eighteen families, densely affected by alcoholism, was used to genotype a set of polymorphic microsatellite markers at loci approximately 10 centimorgans apart spanning the short arm and part of the long arm of chromosome 1. Linkage analyses were performed using the classical lod score and a model-free method. Three different definitions of affection status were defined, these were 1. Heavy Drinking (HD) where affected subjects drank more than the Royal College of Psychiatrists recommended weekly amount. 2. The Research Diagnostic Criteria for alcoholism (RDCA) 3. Alcohol Dependence Syndrome (ADS) as defined by Edwards and Gross (1976) and now incorporated into ICD10 and DSMIV. Results Linkage analyses with the markers D1S1588, D1S2134, D1S1675 covering the cytogenetic region 1p22.1-11.2 all gave positive two point and multipoint lods with a maximum lod of 1.8 at D1S1588 (1p22.1) for the RDCA definition of alcoholism. Another lod of 1.8 was found with D1S1653 in the region 1q21.3-24.2 using the HD affection model. Conclusion These results both support the presence of linkage in the 1p22.1-11.2 region which was previously implicated by the USA Collaborative Study of the Genetics of Alcoholism (COGA) study and also suggest the presence of another susceptibility locus at 1q21.3-24.2.	Background Epidemiological studies report that alcoholism as defined by the Research Diagnostic Criteria (RDC) affects almost 10–15% of the general population in the USA and 1–5% in Europe [ 1 ]. Drinking behaviour is clearly partly determined by cultural and psychological factors, but genetic factors also play an important part as shown by family, twin and adoption studies [ 2 ]. The overall heritability for alcoholism has been estimated to be around 50% to 60% [ 3 ]. However there are almost certainly multiple aetiological subtypes of alcoholism, which will eventually be shown to have heterogeneous genetic and cultural components [ 2 , 4 ]. Many genetic association studies of alcoholism have sought to identify candidate susceptibility genes, but few linkage studies have been undertaken so far. Two genome-wide linkage studies have been performed on US populations. One was carried out on a well defined population of south western USA American Indians [ 5 ] and the other on a large sample of families by the Collaborative Study of the Genetics of Alcoholism (COGA) [ 6 , 7 ]. In the native American Indian study two chromosomal regions provided suggestive evidence for linkage. One was on chromosome 11p close to the DRD4 dopamine receptor and the tyrosine hydroxylase genes and the other on chromosome 4p near the β1 GABA receptor gene. Three loci in the ADH cluster on chromosome 4 also gave evidence of linkage on two point but not multi-point linkage analyses [ 5 ]. On the other hand the COGA study, the multi-point linkage analysis provided suggestive evidence of linkage on chromosome 1 and 7 with more modest evidence for a locus on chromosome 2. In addition, there was suggestive evidence for a protective locus on chromosome 4 near the ADH gene cluster [ 6 ]. This study [ 6 ] implicated chromosome 1 in two distinct regions on 1p21-35. Two-point analysis of affected sib-pairs showed a significant increase of allele sharing for two adjacent markers D1S532 and D1S1588. The multipoint linkage analysis reported a lod of 2.93 for D1S1588. A second region near D1S224, 60 cM apart from the first locus, had a multipoint lod score of 1.65. In the COGA replication set of families [ 7 ], linkage near the marker D1S224 with a maximum multipoint lod score of 1.6 was reported. In the combined COGA sample a LOD score of 2.6 was reported near markers D1S2614 and D1S1588. Re-analysis of COGA data [ 8 ] for linkage to an alcoholism-related phenotype consisting of alcoholism and depression in the combined COGA samples found a maximum lod of 5.12 near the markers D1S1648 and D1S1588. Furthermore, the region on chromosome 1p near D1S1588 and D1S1631 was also identified as demonstrating possible linkage to the "low level of response to alcohol" phenotype with a maximum lod score of 2.0 [ 9 ]. Linkage was also supported in the COGA dataset of an endophenotype characterized by a later age of onset of regular drinking and higher harm avoidance to a region near D1S518 on 1q [ 10 ]. Mouse linkage analyses have also shed light on the genetic determinants of alcohol consumption. The extensive syntenic homology between the mouse and human genomes enables predictions about which human loci are syntenic with mouse alcohol related loci. Buck and co-workers, who studied several mouse alcohol related phenotypes, predicted that genes related to physical dependence on ethanol may localize to human chromosome regions 1q21-43, 2q11-32, 5p15, 5q14-21, and 9p24-22, 10q23-26 [ 11 ]. The purpose of the current study was to test the hypothesis that the positive linkage reported on human chromosome 1p21-23 by the COGA study and elsewhere on chromosome 1 could be replicated in a United Kingdom sample of families multiply affected by alcoholism. Results Selection criteria for the families eliminated all but 9% of the families contacted from all sources. Press contacts were found to be the most productive source of suitable families willing to participate in the study, accounting for over 50% of subjects recruited. There appeared to be no bias in the affection status of subjects recruited from the different sources [ 12 ]. The affection status of all the subjects is summarized in table 1 . The individuals interviewed were 297 and 176 subjects were genotyped for the purpose of this study. Alcohol consumption and severity of dependence indices measured using the severity of alcohol dependence questionnaire (SADQ) confirmed that ADS subjects were the most severely affected, the next level of lesser severity of alcoholism was defined by the Research Diagnostic Criteria of Alcoholism (RDCA) and then by the Heavy Drinking category (HD). The mean maximum regular lifetime consumption of the HD subjects (including the RDCA and ADS subjects, who are all HD) is 138 units/week. 80% of HD subjects reported a period of at least one month during which they were drinking in excess of 50 units/week for males or 35 units/week for females, which are the amounts above which the Royal College of Psychiatrists advises that drinking is likely to be harmful (Royal College of Psychiatrists, 1986). Almost half the HD subjects (49%) gave a history of a one month or longer period during which they were drinking in excess of 150 units/week for males, and 100 units/week for females. The mean SADQ score for ADS subjects was just below 30 (SD 14.2). The range for this rating scale was from 0 to 60. Stockwell and co-workers [ 13 ] suggested that scores above 30 indicate severe dependence. In our sample, 40% of the ADS subjects were severely dependent according to this definition. Table 1 Proportion of the individuals for each affection status category and their mean age ( TT Lifelong abstainers; SD Social Drinkers; HD Heavy drinkers; RDCA Research Diagnostic Criteria Alcoholism; ADS Alcohol Dependence Syndrome) Affection Status Family Sample Number Genotyped % Mean age ± SD (yrs) in all cases and relatives TT 7 4 58.2 ± 13 SD 43 24 50.4 ± 17.1 HD 22 13 41.5 ± 14.1 RDCA 50 28 41.5 ± 13.5 ADS 54 31 42.1 ± 11.9 HD affection model Assuming dominant transmission, two-point heterogeneity LODS (HLOD) for linkage to the HD affections status gave a score of 1.8 near the marker D1S1653. The three-point multipoint analysis produced an HLOD of 1.8 with the markers D1S1653 and D1S1677. Analyses with the same markers assuming recessive transmission produced an HLOD of 1.1 with a two-point analysis and an HLOD of 1.5 with a three-point analysis. The model-free (MFLINK) analysis produced an MLOD of 1.5 for both the two-point and three-point analyses. Three-point analyses of D1S1677 and D1S1679 produced an HLOD of 1.7 assuming recessive transmission and a MALOD of 1.3 for the model-free three point analysis. The marker D1S1675, on the short arm of chromosome 1, produced an HLOD of 1.2 with both a two-point analysis and a three-point analysis in combination with D1S3723 assuming dominant transmission, Model-free analyses of D1S1588 produced a MALOD of 1.5 and 1.2 using two-point and three-point analyses, respectively. Figure 1 shows the two-point linkage HLOD results for the HD category under both the dominant and the recessive model and also the model-free MFLINK analyses. The HD category is a cumulative definition which includes all the more severely affected categories for linkage analysis as well as those at the borderline of becoming alcoholic. The HD category is justified because drinking at this level is well recognised to be a predictor of future alcoholism, an observation that has also been incorporated in another genome linkage study of alcoholism which also employs a similar definition [ 14 ]. We have consistently used this affection status category in all our previous linkage analyses and it has not been applied in an arbitrary manner that could have inflated any lod scores. Figure 1 Two-point linkage analysis for the Heavy Drinking (HD) category under the dominant and the recessive model (HLOD) and model-free analysis (MALOD) RDCA affection model Assuming recessive transmission the marker D1S1679 was linked to RDC alcoholism with a two-point HLOD of 1.0 and a three-point HLOD of 1.6 with the markers D1S1677 and D1S1679. The MLOD two point MFLINK score with D1S1677 was 0.9 and this increased to 1.4 with a three-point model-free analysis. For the RDCA level of affection status the highest MALOD score of 1.8 was found near the telomere of chromosome 1p near the marker D1S1588 under the model free analysis. Figure 2 shows two-point linkage analysis for the RDCA category under the dominant and recessive models and also the model-free analyses. Figure 2 Two-point linkage analysis for the RDC alcoholism (RDCA) category under the dominant and the recessive model (HLOD) and model-free analysis (MALOD) ADS affection model The marker D1S1591 gave the highest maximum admixture lod (MALOD) of 1.00 with the model-free method of analysis for the ADS category. The results of MFLINK multipoint analysis for this and the other affection status models are shown in table 2 . Table 2 MFLINK two-point MALODs for the HD, RDCA and ADS diagnostic categories. In bold MALOD score value ≥1. Markers HD RDCA ADS D1S1591 0 0 1 D1S2134 0.1 01 0.2 D1S1665 0.2 0.5 0 D1S532 0.5 0.4 0.4 D1S1728 0.1 0.4 0.4 D1S551 0.4 0.8 0.5 D1S1588 1.5 1.8 0 D1S1631 0 0.3 0.01 D1S3723 0.2 0.3 0.04 D1S1675 0.8 0.4 0.4 D1S1595 0.7 0.3 0.3 D1S1653 1.5 0.3 0.3 D1S1677 0.4 0.5 0.06 D1S1679 0.8 1 0.5 Conclusion The COGA study first reported evidence of a locus linked to alcoholism on chromosome 1p with the markers D1S532 and D1S1588 linked to Alcohol Dependence [ 6 ]. Curtis and coworkers, in their reanalysis of the initial COGA dataset computed a MALOD lod score of 1.75 with the marker D1S1588 [ 15 ]. The COGA group repeated the analysis in a new set of families and the lod score at D1S1588 was 1.6 in this replication data set. The combined analysis of the two COGA samples gave a lod of 2.6 [ 7 ]. The COGA study also conducted a regression method of linkage so as to include unaffected and discordant pairs. D1S1588 then gave a peak lod score of 2.9. Our own results of a lod of 1.8 at the region near D1S1588 offers support but not clear confirmation for the COGA findings. A second region on chromosome 1p, near D1S224, approximately 60 cM proximal to D1S1588, gave a multipoint lod score of 1.7 in the COGA study [ 6 ]. The current British study gave a MALOD score of 1.2 over this region with the marker D1S1675 under the HD model, and a MALOD of 1.2 with the marker D1S3723 for the RDCA category. Dick et al [ 10 ] who analysed the COGA data have suggested that there is a locus on the long arm of chromosome 1 near the marker D1S518 linked to the endophenotype of a late age of onset of regular drinking and high harm avoidance. In a mouse recombinant inbred strain linkage study, a quantitative trait locus for physical dependence on alcohol was mapped to the murine chromosome 1 which is a region syntenic with human chromosome 1q21-43. In the present UK linkage analyses a three point HLOD score of 1.8 was found with the combined markers D1S1653 and D1S1677 for the HD phenotype. The positioning of the lod reported by Dick [ 10 ], the mouse genetic study and the UK linkage study are all compatible with a second locus for alcoholism which is on the proximal part of the long arm of chromosome 1 at 1q21.3-24.2. In the UK linkage study the two regions on chromosome 1 are implicated with two different affection status categories. We could speculate that the two different loci on chromosome 1 could be mediating an effect on alcoholism through two loci that have different types of effect on susceptibility. The most interesting and probably important finding from the COGA linkage study is that the stratification of families that have members with both alcoholism and depression maximizes the lod to 5.12 on chromosome 1p near the markers D1S1648 and D1S1588 [ 14 ]. Without stratification based on clinical phenotype in the combined COGA sample a LOD score of 2.6 was obtained. These findings provide a strong indication that the 1p locus near D1S1588 is mediating its effect on alcoholism through a genetic susceptibility for depression and anxiety. A recent genome wide scan for quantitative-trait loci carried out by Fullerton and colleagues [ 16 ] identified a locus on chromosome 1p, D1S2868, linked to neuroticism. The authors speculated that this locus influences traits genetically related to neuroticism and maps near the locus, D1S1588, identified by Nurnberger et al. [ 14 ] for the endophenotype of alcoholism and depression. It is a generally accepted clinical fact that it is very difficult to diagnose whether depression or anxiety are predisposing to alcoholism or resulting from it. It is likely that both effects occur simultaneously. The twin and family study data to date, strongly supports the hypothesis that alcoholism both causes and augments pre-existing depression and anxiety [ 17 - 19 ]. The resolution of this issue at a biological and psychosocial level is a high priority for psychiatry as a whole because of the very high population prevalence of these comorbid disorders in most populations. The full understanding of these linkage findings in alcoholism may help us unravel basic aetiological mechanisms and help create new treatment and preventive strategies based on a fuller understanding of genetic and environmental factors involved in comorbidity. It would seem justified to start allelic association studies on chromosome 1 in order to fine map susceptibility genes for patients comorbid for depression, anxiety and alcoholism. A positive finding would gain support from the prior linkage studies of both alcoholism and affective disorders and vice versa. The presence of locus heterogeneity for genetic effects on depression, anxiety and alcoholism will require relatively large case-control studies of alcoholics with adequate clinical assessment to detect alcoholics with and without comorbid disorders. Methods a) Family Sample Prior to commencing the study, ethical permission for this research project was obtained from the University College London Medical School Clinical Investigations Panel which has been updated in 2003 with multicentre research ethics committee (MREC) approval. Information stored on computer was registered under the 1984 Data Protection Act. Caucasian families multiply affected by alcoholism, and suitable for linkage studies, were ascertained with the following selection criteria. (a) Presence of two or preferably more cases of alcoholism as defined by the RDC. (b) Large families, preferably with two or more generations willing and able to participate. (c) Evidence of unilineal inheritance of alcoholism in the parental generation, which meant that families were excluded where both parents were affected. (d) Willingness of as many family members as possible, especially affected individuals, to participate in the research project. The families were identified from hospital records, clinicians, advertisements in the media and other contacts [ 12 ]. The families were all white Caucasian in origin. The subjects were diagnosed using the following interview schedules: the Schedule for Affective Disorders and Schizophrenia – Lifetime version (SADS-L) which provides an RDC diagnosis of alcoholism [ 20 ], a "lifetime modification" of sections 1 and 3 of the Clinical Alcoholism Interview schedule [ 21 ], the Lifetime Drinking History [ 22 ] and a lifetime modification of the Severity of Alcohol Dependence Questionnaire (SADQ) [ 13 ]. Heavy Drinking (HD) was defined as drinking in excess of the Royal College of Psychiatrists recommendations, i.e. more than 14 units per week for females or 21 units per week for males, for over one month (Royal College of Psychiatrists, 1986). Diagnoses were made at three levels and these were employed in the linkage analyses in a hierarchical manner. All subjects fulfilling criteria for the Alcohol Dependence Syndrome (ADS) also fulfilled the criteria for RDC alcoholism (RDCA) and HD. All subjects meeting criteria for RDCA by definition also met the criteria for HD. Individuals who were drinking regularly below these limits, or who drank infrequently, were classified as social drinkers (SD), these individuals were considered as unaffected in the linkage analyses. Subjects who had never drunk alcohol, or had only one or two drinks in a lifetime, were considered to be lifelong abstainers (TT) and were also considered as unaffected. In order to check the RDCA, HD, SD and TT assessments, the interview schedules were examined by two independent psychiatrists. Where discrepancies arose between two raters, a consensus diagnosis was reached by process of joint discussion and consideration of the data. b) Laboratory procedures Fifty nanograms of total genomic DNA was extracted from venous blood samples and amplified by Polymerase Chain Reaction (PCR) with oligonucleotide primers using standard methodology [ 23 ]. Short Tandem Repeat markers from the Research Genetics Set 9 genome screening panel and key markers from the COGA study were typed. A M13 tail was added to one of the oligonucleotide primers which was used to hybridise with a complementary oligonucleotide pre-labelled with infra red dyes that fluoresce at either 700 or 800 nm. The resulting images from argon laser scanning allow a highly detailed visualisation of genetic polymorphisms [ 24 ]. Gel electrophoresis and pattern visualization was performed using LI-COR Model 4200 automated fluorescent DNA sequencers [ 25 ]. Ten markers on 1p21-35 were genotyped: D1S1591, D1S2134, D1S1665, D1S1728, D1S532, D1S1588, D1S551, D1S1631, D1S3723 and D1S1675. Four markers on 1q were genotyped: D1S1595, D1S1653, D1S1677 and D1S1679. Marker order and intermarker distances were as compiled in the Ensembl database . Positions of the markers are shown in table 3 . Table 3 Marker location and type of polymorphism Markers Mb Type D1S1591 39 TETNUC D1S2134 48 TETNUC D1S1665 73 TETNUC D1S532 77 TETNUC D1S1728 81 TETNUC D1S551 82 TETNUC D1S1588 92 TRINUC D1S1631 105 TRINUC D1S3723 107 TETNUC D1S1675 114 TETNUC D1S1595 153 TETNUC D1S1653 155 TETNUC D1S1679 159 TETNUC D1S1677 161 TETNUC Genotypes were read blind to diagnostic information. Tests for Mendelian inheritance of marker data were performed and inconsistent genotypes were repeated or omitted. c) Linkage analysis The linkage analysis was performed using the classical lod score method and using likelihood-based model-free analysis carried out with the MFLINK program [ 26 ]. For lod score analyses, the programs MLINK and LINKMAP were used from the FASTLINK package [ 27 ]. Three affection models were used: ADS, RDCA and HD. Each affection model was analysed assuming dominant transmission, recessive transmission and using the model-free method. For dominant models the frequency of the abnormal allele was set to 0.02 and for recessive models to 0.2. The penetrance for normal and abnormal genotypes respectively were set to 0.02 and 0.5 for ADS, to 0.04 and 0.7 for RDCA and to 0.2 and 0.9 for HD. These penetrance values were chosen to produce models approximately consistent with prevalence data from previous epidemiological research of 0.04, 0.06 and 0.2 for ADS, RDCA and HD respectively [ 28 , 29 ]. Two-point and three-point analysis was carried out for all affection definitions and transmission models. Classical linkage analysis was carried out under the assumption that locus heterogeneity might be present, yielding an HLOD statistic. Authors' contributions IG carried out the genotyping and produced the manuscript CCHC collected the multiplex family sample and participated in the design of the study WK participated in genotyping AD participated in genotyping AM participated in genotyping DC performed the linkage analysis HMDG conceived the study and participated in its design and coordination All authors read and approved the final manuscript	/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC554783.xml
521687	Factors influencing preoperative stress response in coronary artery bypass graft patients	Background In many studies investigating measures to attenuate the hemodynamic and humoral stress response during induction of anaesthesia, primary attention was paid to the period of endotracheal intubation since it has been shown that even short-lasting sympathetic cardiovascular stimulation may have detrimental effects on patients with coronary artery disease. The aim of this analysis was, however, to identify the influencing factors on high catecholamine levels before induction of anaesthesia. Methods Various potential risk factors that could impact the humoral stress response before induction of anaesthesia were recorded in 84 males undergoing coronary aortic bypass surgery, and were entered into a stepwise linear regression analysis. The plasma level of norepinephrine measured immediately after radial artery canulation was chosen as a surrogate marker for the humoral stress response, and it was used as the dependent variable in the regression model. Accordingly, the mean arterial blood pressure, heart rate and the calculated pressure-rate product were taken as parameters of the hemodynamic situation. Results Stepwise regression analysis revealed that the oral administration of low-dose clonidine (mean dose 1.75 μg·kg -1 ) on the morning of surgery was the only significant predictor (p = 0.004) of the high variation in preoperative norepinephrine plasma levels. This intervention decreased norepinephrine levels by more than 40% compared to no clonidine administration, from 1.26 to 0.75 nmol·l -1 . There was no evidence for dose-responsiveness of clonidine. All other potential predictors were removed from the model as insignificant (p > 0.05). The use of beta-blocker, ace-inhibitors, ejection fraction, and body mass index were significant determinants for the hemodynamic situation (heart rate, mean arterial pressure, pressure rate product) of the patient during the pre-induction period. Conclusion The oral administration of clonidine is the only significant predictor for the observed variation of norepinephrine levels during the preoperative period. Lack of significant dose responsiveness suggests that even a low dose of the drug can attenuate the preoperative stress response and thus is recommended in cardiovascular high risk patients.	Background There is increasing evidence that sympathetic nervous system mediated cardiovascular stimulation with increased catecholamine blood levels is the principal mechanism responsible for perioperative tachycardia and hypertension, myocardial ischemia and infarction [ 1 - 3 ]. Even short-lived changes may have detrimental effects on the coronary circulation of high-risk patients, with higher rates of morbidity and lethality [ 4 , 5 ]. Thus, many studies have concentrated on the stressful stimulus of endotracheal intubation, and a number of pharmacological attempts have been used to attenuate the hemodynamic response, including the use of high doses of opioids, α 2 -adrenergic receptor agonists, β-adrenergic blocking drugs or other antihypertensive drugs [ 3 , 4 , 6 - 19 ]. However, little attention has been paid to the stress response of cardiac high risk patients when entering the operating area, during initiation of routine monitoring, and finally during awake venous and arterial canulation. Especially the latter procedure can cause significant discomfort for patients even when performed under local anaesthesia [ 20 ]. In several trials a high inter-individual variation of pre-induction norepinephrine levels, heart rate and blood pressure could be noticed [ 21 ]. Thus, this observational analysis was performed in patients undergoing coronary aortic bypass surgery (CABG-surgery) in order to identify patients at high risk for increased sympathoadrenergic stress response during the immediate preoperative period using norepinephrine levels and the hemodynamic status as surrogate measures. Methods After Ethics Committee approval was obtained, patients gave their written and informed consent. Eighty-four consecutive male patients undergoing CABG-surgery were enrolled into this observational study. The only exclusion criterion was emergency operation. Due to the observational character of the study, drugs administered preoperatively (including clonidine) were given at the discretion of the anaesthetist performing the preoperative examination. Each patient received an oral premedication with clorazepate 20 mg in the evening before surgery and in the morning of surgery. In about half of the patients, benzodiazepine premedication was combined with clonidine 75–300 μg. Patients were maintained on their regular cardiac and antihypertensive medication up to the day of surgery but all inhibitors of platelet aggregation were discontinued 3–7 day preoperatively. After arrival at the operating theatre an i.v. line was initiated and 500 ml hydroxyethyl-starch (10%, 200000 Dalton) was infused. A 12-lead channel ECG with an automatic ST-segment analysis, oxygen saturation and invasive blood pressure monitoring were connected to the patients. A radial artery catheter was then inserted after local anaesthesia with 1 ml mepivacaine 1%. Heart rate (HR) and mean arterial blood pressure (MAP) were recorded every 15 seconds online using a Laptop computer connected to a Solar 9500 monitor (General Electrics, USA). HR and MAP were multiplied to receive the pressure-rate product (PRP). These variables were used as a measure of the hemodynamic stress response. To determine humoral stress response, an arterial blood sample for the measurement of norepinephrine plasma level was taken immediately after placing the radial artery catheter (this measurement was the main outcome of the study) and 5, 15, and 60 minutes after endotracheal intubation was performed (these measurements were used in an additional explorative analysis, see figure 1 ). 10 ml plastic lithium-heparin tubes were used for this purpose. Specimens were placed on ice immediately after sampling, spun in a centrifuge for 20 minutes and plasma was separated and stored at -70°C pending analysis. Plasma norepinephrine levels were determined by high-performance liquid chromatography (HPLC) with electrochemical detection (Millipore, Billerica, Mass. USA). The lower limit of detection for norepinephrine was 0.018–0.024 nmol·l -1 and the same-day coefficient of variation for norepinephrine measurements determined by repeated measures of a standardized probe was 3%. Figure 1 Norepinephrine levels in patients with and without clonidine premedication. General anaesthesia was standardized. After administration of midazolam 0.05 mg·kg -1 and a three minute period of preoxygenation, anaesthesia was induced using a continuous infusion of propofol (10 mg·kg -1. h -1 ) and sufentanil (10 μg.kg -1. h -1 ). After loss of consciousness propofol was reduced to 3 mg·kg -1. h -1 and sufentanil to 1.5 μg·kg -1. h -1 . Endotracheal intubation was performed after administration of pancuronium bromide 0.1 mg·kg -1 . To allow comprehensive analysis of potential factors associated with a reduced stress response, the following data were recorded prospectively: • Age • Bodyweight • Height • Body mass index (BMI) • Clorazepate dose per kilogram bodyweight • Clonidine (yes – no) • Clonidine per kilogram bodyweight • Time from morning premedication until observational period • Inhibitors of the angiotensine converting enzyme system (ace-inhibitors) • Beta-blocking drugs • Calcium antagonists • Angiotensin-2 receptor inhibitors • Left ventricular ejection fraction (EF) • Number of affected vessels Statistical analysis A power analysis had revealed that 80 patients provide a power of more than 95% to detect an R 2 of 0.3 and higher, attributed to 14 independent variables using an F-test with a significance level of 0.05. All potential relevant factors were subjected to a stepwise linear regression analysis using a backward technique. In each step the least significant factor was eliminated when F was lower than 3.96. The quality of the regression model was judged using the Durbin-Watson statistic (a value between 0 and 4 indicating the amount of autocorrelation within the model with an optimum of 2.0), and by checking if the standardized residuals are normally distributed. All calculations were performed using SPSS 11.0 for Windows. All continuous data are presented as mean and standard deviation when normally distributed and as median (25 th –75 th percentile) when normal distribution had to be rejected using the Kolmogorov-Smirnov-test. Results Stepwise regression analysis revealed that the single administration of low-dose clonidine (mean dose 1.75 μg·kg -1 ) on the morning of surgery was the only significant predictor (p = 0.004) of the high variation in preoperative norepinephrine plasma levels. This intervention decreased norepinephrine levels by 40% compared to no clonidine administration (from 1.26 to 0.75 nmol·l -1 ). In this analysis, the dichotomous variable (clonidine administration: yes-no) was a better predictor for the norepinephrine levels than variables including the clonidine dose (absolute dose or dose per body weight), indicating that our data provide no evidence for a strong dose-responsiveness of clonidine in this context. All other of the investigated factors (see methods) were removed from the regression model as not significant. The two factors that were eliminated with a p < 0.1 during the last but one and during the final step were body mass index (removed in step 11 with a p = 0.064) and age (removed in step 12 with a p = 0.076). Both factors were associated with increased norepinephrine levels. The overall quality of the regression model was excellent. The Durbin-Watson coefficient was 2.04 (very near to the optimum of 2.0) and the standardized residuals were normally distributed. For mean arterial blood pressure, heart rate and the calculated pressure rate product, however, preoperative clonidine administration was not an influencing factor. For the mean arterial pressure (MAP), a higher ejection fraction (EF) was a statistically significant predictor (p = 0.024). Each 10% increase of EF was associated with a 2.7 mmHg higher MAP. Administration of an ace-inhibitor was the second predictor in the final model of MAP (p = 0.03). These patients had a 7.5 mmHg lower MAP than patients without ace-inhibitors. For heart rate (HR) there were three significant predictors that remained in the model. Administration of beta-blockers and ace-inhibitors were both associated with a decreased HR (p = 0.004). Each of them decreased HR between 6–7 beats per minute (bpm). Additionally, a higher BMI was associated with a 1.3 bpm higher HR per kg·m -2 (p = 0.001). Since the PRP is the product of HR and MAP, it is not surprising that similar variables contributed to its prediction. These were the administration of beta-blockers (p = 0.017) and ace-inhibitors (p = 0.004), each of them reducing the PRP, whereas the EF was associated with an increase in PRP (p = 0.014). No patient had signs of cardiac ischemia on arrival at the operating theatre until induction of anaesthesia (defined as ST-T change > 0.1 mV in any ECG lead). There were no major adverse events during the entire induction period and surgery. Discussion The α 2 -adrenergic receptor agonist clonidine acts by decreasing central sympathetic nervous system activity in all hyperadrenergic situations. In addition to its sedative, anxiolytic, analgesic and antihypertensive properties [ 6 , 22 ] it has shown to improve congestive heart failure, to optimize the myocardial oxygen supply / demand ratio in ischemic heart disease [ 23 , 24 ] and to reduce attacks of angina pectoris [ 25 , 26 ]. In many investigations attention has been drawn to the stressful stimulus of endotracheal intubation [ 3 , 4 , 6 - 19 ], as it has been shown that even short-lasting sympathetic cardiovascular stimulation may have detrimental effects on the coronary circulation of patients with coronary artery disease (CAD), with higher rates of morbidity and lethality [ 4 , 5 ], However, little emphasis has been paid to the preoperative period where patients may be stressed by or because of the upcoming procedure. Furthermore, transfer to the operating theatre, initiation of routine monitoring, and venous and arterial canulations are stressors for the patients. In this context it is certainly a drawback of our study that we did not record the level of preoperative sedation or anxiolysis using clinical measurements on appropriate scales. Instead of this, only a rough judgment was made (awake versus asleep but rousable) that did not allow a valid analysis of the data. Previous data, however, have shown potent anxiolytic and sedative properties of the drug [ 27 ]. Thus, it was the major aim of this observational study to identify factors that might contribute either to increased humoral stress or that might help to attenuate this response. Our results show, that a single application of low dose oral clonidine was the only factor that was associated with significantly decreased norepinephrine levels on arrival at the pre-induction area. The question that arises from this observation is, if this is simply an association (or statistically spoken a co-linearity between other protective factors) or if clonidine premedication is the cause for lower norepinephrine levels. In our opinion the latter is the case. Firstly, there were no differences considering any other variables between those patients who had received clonidine and those who had not (see table 1 ). Thus, it is unlikely that other factors were responsible for the reduced norepinephrine levels. Secondly, there is good evidence from the literature that clonidine is a powerful drug that attenuates stress response of various causes [ 3 , 4 , 6 - 19 ]. Table 1 Patients' demographic data and preoperative condition. Data are presented for all 84 patients that were included in the study, and separately for those patients receiving clonidine and those without oral clonidine premedication. Values are expressed as mean ± standard deviation, median (25 th –75 th percentile), or n = (%). All patients Patients with clonidine morning premedication Patients without clonidine morning premedication n = 84 n = 42 n = 42 Age (years) 66 ± 9 65 ± 9 67 ± 8 Bodyweight (kg) 82 ± 10 82 ± 11 82 ± 10 Height (cm) 173 ± 6 173 ± 6 173 ± 5 BMI (kg · m -2 ) 27.3 ± 3.1 27.3 ± 3.2 27.4 ± 3.0 EF (%) 62 ± 14 61 ± 13 63 ± 16 Affected vessels (n = / %) n = 1 3 (4%) 0 (0%) 3 (7%) n = 2 16 (19%) 9 (21%) 7 (17%) n = 3 64 (76%) 33 (79%) 31 (74%) n = 4 1 (1%) 0 (0%) 1 (2%) Pre-treated with (n= / %) ace-inhibitors 43 (51%) 17 (40%) 26 (62%) Beta-blockers 55 (65%) 26 (62%) 29 (69%) Calcium-antagonists 12 (14%) 5 (12%) 7 (17%) Clonidine dose (n= / %) 75 μg 8 (19%) 150 μg n/a 33 (79%) n/a 300 μg 1 (2%) Time from morning premedication until observational period (hours) 1.0 (0.5–4.5) 2.5 (0.5–5.0) 1.0 (0.5–4.5) Heart rate [bpm] 66 ± 11 66 ± 10 66 ± 12 Mean arterial blood pressure [mmHg] 102 ± 16 100 ± 15 104 ± 17 Pressure rate product [mmHg·bpm] 6750 ± 1640 6660 ± 1600 6850 ± 1690 Plasma norepinephrine level (nmol·l -1 ) 1.00 ± 0.82 0.75 ± 0.48 1.26 ± 1.00 However, it is interesting to notice that the mean dose administered to our patients (1.75 μg·kg -1 ) was low compared to all other trials. Data concerning the appropriate dose of clonidine to attenuate the stress response to intubation vary considerably between 0.625 and 10 μg·kg -1 . For example, one trial demonstrated that clonidine 0.625 and 1.25 μg·kg -1 i.v. were sufficient to attenuate pressure response to laryngoscopy and intubation [ 28 ], whereas in another one [ 19 ] evaluating the dose-response effects to laryngoscopy and intubation, 2 μg·kg -1 clonidine i.v. was equally effective as placebo, and only 4 and 6 μg·kg -1 significantly attenuated hemodynamic and adrenergic reactions in an equal manner. It could also be shown that 4 or 6 μg·kg -1 were necessary to reduce norepinephrine levels before induction of anaesthesia, however 2 μg·kg -1 where not sufficient in this setting [ 19 ]. In our trial as well as in all other studies with even much higher doses, clonidine was well tolerated and did not produce any adverse hemodynamic effects. In our analysis there was no strong evidence for a dose responsiveness of orally administered clonidine. First, in the regression model catecholamine levels could better be predicted by the dichotomous variable and second, there was only a weak correlation between the weight adjusted clonidine dose on the one hand and norepinephrine levels on the other hand (Pearson's correlation coefficient was -0.31, Spearman's rank correlation coefficient was -0.30). Furthermore, a post-hoc comparison between the patients receiving either 75 or 150μg clonidine did not show relevant differences (p = 0.91 using the Mann-Whitney U-test). Higher age and higher body mass index showed a non-significant tendency to increase the catecholamine concentration. No other of the investigated factors (body weight, height, time from morning premedication until observational period, benzodiazepine dose per kilogram bodyweight, ace-inhibitors, beta-blocking drugs, calcium antagonists, EF, number of affected vessels) had statistically significant impact on norepinephrine levels. An explorative post-hoc analysis of the impact of clonidine premedication (none versus any dose) and clonidine dose on norepinephrine levels during the entire induction period proves the results of the main analysis. There was a pronounced reduction of norepinephrine plasma levels after induction of general anaesthesia with lower values in the clonidine-group. However, a statistically significant interaction term (p = 0.012) suggests that the fall of norepinephrine levels are more marked in the untreated group and thus mainly caused by induction of general anaesthesia rather than effects of clonidine (figure 1 ). Conclusion This observational trial demonstrates that patients undergoing coronary artery bypass graft surgery have a great variation of norepinephrine levels when entering the operating theatre. We could identify oral clonidine premedication as the only predictor for increased humoral stress response. There was no strong evidence for a dose dependency, indicating that even small doses, like 75–150 μg attenuate the humoral stress response before coronary artery bypass graft surgery. Clonidine did not have a negative impact on hemodynamic parameters. Competing interests None declared. Authors' contributions AMM processed the data and wrote the manuscript. GG conceived the study, collected the clinical data and participated in its design. US collected the clinical data. MK designed the study and collected the clinical data. HAA performed the laboratory investigations. HW participated in the conception of the study. LHJE designed the study, performed the statistical analysis and extensively revised the manuscript. All authors read and approved the final manuscript. Pre-publication history The pre-publication history for this paper can be accessed here:	/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC521687.xml
524258	A Clear View of Mycobacterial Infection	null	Fighting an infection might seem to be a battle between David and Goliath, given the relative sizes of bacterial infectious agents and the animals they infect. But on closer examination it is more often a chess match between two skilled opponents who have the uncanny ability to anticipate each other's moves. Mycobacterium tuberculosis causes tuberculosis (TB) in people, and related species that infect other animals are used as model systems for the study of TB. Much progress has been made in identifying the armaments (or virulence factors) of the bacteria. But the interplay, or chess match, between the bacterium and the animal it infects is much less clear. One of the host's first moves against the mycobacterium is the formation of a granuloma. Granulomas are tightly aggregated structures that consist of macrophages—one of the first lines of defense of the immune system—within which the infecting bacterium grows. Besides these and related cells that are present at the site of infection, additional macrophages and other immune cells are recruited in the formation of the granuloma. Although granulomas are required for the elimination of the infection, Lalita Ramakrishnan and colleagues have now shown that the bacteria have a game plan of their own. Mycobacteria co-opt granulomas for their growth and spread One problem in understanding the interaction between the mycobacterium and the host has been that it occurs deep in the lung of the infected animal, which makes it difficult to analyze how each of the animal or bacterial factors affect the strategic interplay between the host and pathogen. To overcome this limitation, Ramakrishnan and colleagues used zebrafish embryos, which are transparent and can be infected by a relative of the TB pathogen, M. marinum . This enables the researchers to watch cells as they are recruited into the granuloma. Some of the virulence factors of mycobacteria are encoded in an area of the genome called the RD1 locus. In a mouse model, a strain of the bacteria missing RD1 causes far less pathology than a strain with the full complement of genes. The RD1 locus is also absent in the bacterial strain M. bovis that is used as an attenuated TB vaccine. But the precise role of RD1 in infection remains obscure. By visualizing in zebrafish infections of a virulent strain of M. marinum and a strain with an RD1 deletion, Ramakrishnan and colleagues have observed that RD1 is actually required for granuloma formation but isn't needed for the bacteria to infect macrophages. What's more, macrophages that are infected with mycobacteria that contain RD1 produce a signal that further recruits macrophages to granulomas. This might seem an odd virulence strategy, as macrophages are required for mycobacterial elimination. But in this ongoing chess match, the virulent mycobacterium exploits the host's defense—granuloma formation—by providing additional macrophages for the bacteria to infect. The end game of the chess match remains unclear. While granulomas are required for protection against mycobacteria, they are not completely effective. Thus, these bacteria have developed a strategy to recruit the normally defensive cells of the host to their advantage, but it remains to be shown what tips the balance between the macrophages' ability to clear the infection and their unwitting participation in the development of TB.	/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC524258.xml
509321	Evolution's “Molecular Clock”: Not So Dependable After All?	null	DNA mutates, and it's a good thing it does. If it didn't, there could only be one kind of life, not the millions there are today, and species could not adapt to new challenges. This is because mutations in genes—the coding portion of DNA—are the raw material for evolution. However, genes make up a surprisingly small fraction of our DNA. If the genome were a cookbook, its 30,000-odd genetic recipes would be scattered among millions of pages of apparently meaningless nonsense. Mutations affect all DNA, not just the genes, and this provides population geneticists with a veritable toolbox of methods useful, for example, in DNA profiling. Importantly, all these methods rely on the idea of a “molecular clock,” the notion that mutations rain down on noncoding DNA like a fine drizzle, so constantly that genetic similarity is a good measure of evolutionary time. Thus, if orangutans diverged from humans twice as long ago as did chimpanzees, on any given piece of DNA we would find twice as many differences between the orangutan sequence and the human sequence as between humans and chimps. The mutations are marking time. If the molecular clock works, scientists can do wonderful things like estimating how long ago it was that the common ancestor of all humans lived, or when birds evolved from dinosaurs. The clock assumes that mutations occur independently of each other and at a constant rate. By analyzing thousands of noncoding DNA sequences scattered throughout the human genome, Edward Vowles and William Amos have found that the clock is anything but constant. Instead, a mutation in one spot in the genome affects the chance of getting another mutation nearby. Non-random base frequencies around microsatellites Not all noncoding DNA is made up of benign tracts of random letters. Some sequences appear to be more difficult to copy than others, and these trouble spots can give rise to alphabetic stuttering. DNA is made up of four component chemical units, called nucleotides, which are often referred to by their initial letters: A, C, G, and T. Stuttering occurs when the same pairs or triplets of letters occur together, for example ACACAC. Such regions are called microsatellites, and instead of mutating by swapping one letter for another, as most nucleotides do, these sequences evolve mainly by gaining and losing triplets or pairs like “AC.” In this study, Vowles and Amos used the published sequence of the human genome to track down and compare thousands upon thousands of microsatellites. If the molecular clock ran smoothly, they would expect to find no similarity at all between the DNA sequences surrounding any pair of unrelated microsatellites. To their surprise, they found the complete reverse, with entirely unrelated microsatellites showing widespread and obvious similarities in their flanking DNA. This meant that mutations near microsatellites were not random, but favored certain letters in certain positions. Just as a new shipwreck will attract its own special community of marine life, so microsatellites appear gradually to change the surrounding DNA towards a common pattern. The result is convergent evolution, an unusual state of affairs where, as time goes by, DNA sequences become more similar, not less. As yet, the exact mechanisms remain unclear, though it probably has something to do with how comfortably different combinations of letters sit next to each other. In English, “U” always follows “Q” and “B” never follows “V.” Similar rules may apply to DNA, albeit on a much subtler level. For example, if a microsatellite contains alternating As and Cs, the flanking regions also tend to have As at alternate positions, in phase with the As in the microsatellite. It is as if the DNA prefers the pattern in the microsatellite to extend into the flanking DNA, rather than abruptly stopping at the end of the microsatellite. These findings suggest that it may be wise to take the notion of a molecular clock at face value. With a perfect clock, two or three identical mutations would be highly unlikely, but we now know that this may be possible near microsatellites. Vowles and Amos estimate that as much as 30% of the genome may show evidence of convergent evolution, simply because microsatellites are so common. These mutation biases probably exist to a lesser extent in most sequences. Once scientists understand more fully how and where these biases operate, they may be able to estimate more accurately the risk of any given mutation occurring, be it one that causes human disease or makes a virus more virulent. These findings represent yet another windfall from the Human Genome Project, and act as a powerful reminder that unexpected results always lurk around the corner as we delve deeper into the secret world of the genome.	/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC509321.xml
521693	Assessing response bias from missing quality of life data: The Heckman method	Background The objective of this study was to demonstrate the use of the Heckman two-step method to assess and correct for bias due to missing health related quality of life (HRQL) surveys in a clinical study of acute coronary syndrome (ACS) patients. Methods We analyzed data from 2,733 veterans with a confirmed diagnosis of acute coronary syndromes (ACS), including either acute myocardial infarction or unstable angina. HRQL outcomes were assessed by the Short-Form 36 (SF-36) health status survey which was mailed to all patients who were alive 7 months following ACS discharge. We created multivariable models of 7-month post-ACS physical and mental health status using data only from the 1,660 survey respondents. Then, using the Heckman method, we modeled survey non-response and incorporated this into our initial models to assess and correct for potential bias. We used logistic and ordinary least squares regression to estimate the multivariable selection models. Results We found that our model of 7-month mental health status was biased due to survey non-response, while the model for physical health status was not. A history of alcohol or substance abuse was no longer significantly associated with mental health status after controlling for bias due to non-response. Furthermore, the magnitude of the parameter estimates for several of the other predictor variables in the MCS model changed after accounting for bias due to survey non-response. Conclusion Recognition and correction of bias due to survey non-response changed the factors that we concluded were associated with HRQL seven months following hospital admission for ACS as well as the magnitude of some associations. We conclude that the Heckman two-step method may be a valuable tool in the assessment and correction of selection bias in clinical studies of HRQL.	Background The potential impact of missing survey responses is often ignored in health-related quality of life (HRQL) studies [ 1 , 2 ]. Missing data from study participants can cause bias in parameter estimates of models predicting HRQL outcomes [ 2 - 4 ]. Unfortunately, regression models are frequently interpreted with the assumption that available data are representative of the entire study population. Researchers may compare clinical characteristics of respondents with and without missing surveys, but rarely attempt to assess the impact of these differences on the regression model parameter estimates and ultimately on the results of the study. The assumption that there are minimal or no effects on parameter estimates is only reasonable if one can demonstrate that data missing from a study are truly missing at random, making them ignorable, which is rarely the case [ 2 , 4 ]. Newer statistical techniques have been developed to assess and correct for bias resulting from missing HRQL surveys [ 2 , 3 ]. One technique which has received little attention in the medical literature to date is the Heckman two-step method [ 5 - 8 ]. The Heckman method was developed by an economist, James Heckman, to address problems of self-selection among women participating in the labor force. This method makes it possible to assess whether selection bias is present, identify factors contributing to the selection bias, and to control for this bias in estimating the outcomes of interest. The Heckman method attempts to control for the effect of non-random selection by incorporating both the observed and unobserved factors that affect non-response. The objective of this study was to demonstrate the use of the Heckman two-step method to assess and correct for bias due to missing HRQL surveys. To accomplish this goal, we evaluated HRQL outcomes in a cohort of patients with acute coronary syndromes (ACS). Methods Study population We analyzed data from the VA Access To Cardiology study, which was a multi-center prospective cohort study of 2,733 veterans with a confirmed diagnosis of acute coronary syndromes (ACS), including either acute myocardial infarction or unstable angina [ 9 ]. Baseline patient characteristics (demographic, cardiac history, non-cardiac history and hospitalization variables) were collected at the time of ACS hospitalization. HRQL outcomes were then assessed by the Short-Form 36 (SF-36) health status survey which was mailed to all patients who were alive 7 months following ACS discharge. A second mailed survey was sent to non-respondents. If no response was obtained from the mailed surveys, attempts were made to contact the patients by phone. Of the 2,733 patients in the study, 1,660 (61%) completed the survey, 306 (11%) died, and 767 (28%) were alive and did not complete the survey. Of those completing the survey, most responded to the first mailing with much smaller numbers responding to the second mailing or to phone calls. Variables The outcome variables were the Physical Component Summary (PCS) and Mental Component Summary (MCS) scores from the 7-month SF-36 health status survey. The PCS and MCS scores reflect a patient's overall physical and mental health status, respectively [ 10 , 11 ]. The PCS and MCS scores are continuous variables with a range of 0–100, where higher scores indicate better health status. We constructed a dichotomous variable to indicate whether the patient responded to the SF-36 or not. It is important to note that models of HRQL outcomes may be biased both because subjects may have died before survey administration (survivor bias) or because of survey non-response in subjects that were alive at the time of survey administration [ 2 ]. However, the best way to handle patients who die prior to administration of the HRQL survey remains controversial [ 12 ]. Since the focus of this paper was to demonstrate the use of the Heckman model rather than methods of dealing with death in HRQL studies, we included only those patients who survived 7-months and were therefore eligible to complete the survey. Candidate predictor variables included a wide array of demographic, cardiac, and non-cardiac variables from the index hospitalization, and selected variables from the interim period between discharge and the 7-month SF-36 health status survey (Table 1 ). These variables were derived from the established literature on risk factors for adverse post-MI outcomes (mortality, functional status, and HRQL) [ 13 - 20 ]. Patient demographic and clinical data from the index hospitalization and 7-month follow-up period were abstracted from the electronic medical record and/or from national VA patient care databases. Table 1 Comparison of SF-36 Survey Responders to Non-Responders Variables SF 36 responders SF 36 non-responders P-value Demographics Mean patient age 65.5 62.8 <.01 Male gender 97.7% 97.5% 0.69 Caucasian race 78.4% 80.0% 0.21 Prior Cardiac History Hx of MI 38.0% 36.6% 0.33 Hx of Chronic heart failure 18.2% 18.3% 0.90 Hx of Coronary artery bypass graft surgery 26.3% 23.0% 0.01 Hx of Percutaneous coronary intervention 17.2% 12.6% <.01 Prior Non-Cardiac History History of trauma 0.4% 0.7% 0.31 Substance or alcohol abuse 15.6% 21.0% <.01 Current smoker 35.7% 43.0% <.01 Dementia 2.0% 3.3% <.01 COPD 29.7% 26.1% 0.01 Stroke 10.8% 13.2% 0.01 Depression 28.7% 40.5% <.01 Arthritis 54.8% 54.4% 0.92 Diabetes mellitus 31.6% 32.0% 0.73 Peptic ulcer disease 11.1% 9.6% 0.26 Events during Index Admission and Interim Period* Coronary revascularization during index admission 26.6% 23.9% 0.05 Cardiogenic shock 1.1% 1.1% 0.94 Hypotensive episode 8.7% 9.2% 0.59 Do not resuscitate order 1.2% 2.9% <.01 Positive stress test 12.1% 11.1% 0.32 Admitted to Tertiary VA 63.9% 69.5% 0.01 ST-segment elevation MI on ECG 19.0% 21.4% 0.05 Mean serum creatinine 1.17 1.19 0.58 Left bundle branch block on ECG 3.6% 3.2% 0.52 Cardiac catheterization during index admission 33.4% 31.4% 0.16 Discharge diagnosis Unstable Angina (vs. MI) 49.4% 46.5% 0.06 * Interim period is defined as the time period between discharge from index ACS hospitalization and the 7-month HRQL survey assessment. MI: myocardial infarction; Hx: History; COPD: chronic obstructive pulmonary disease; ECG: electrocardiogram Analyses Baseline characteristics of the patients who did and did not complete an SF-36 were compared using analysis of variance for continuous variables and chi-square for categorical variables (Table 1 ). Then, a series of regression models were developed, including 1) initial PCS and MCS models which did not account for potential bias due to missing surveys, 2) Heckman selection models (modeling response to the SF-36), and finally 3) final PCS and MCS models (accounting for potential bias due to missing surveys). We used robust regression for all equations (Stata version 8.0 SE), controlling for cluster sampling by VA medical center. In prior analyses, we established that the intra-class correlation, the measure of the effect of clustering by medical center in this case, was significant. As a result, it was necessary to control for bias due to autocorrelation, or similarity among patients within a medical center, compared to patients at a different medical center. Stata uses the Huber-White estimator to control for the bias due to clustering. This technique deflates the standard errors of the parameter estimates, in this case the coefficients, correcting the inference statistics. Overview of the Heckman method There are two steps in the Heckman method. The first step is the development of a selection equation (i.e. a model of factors associated with survey non-response). This step includes derivation of a variable from the selection equation called the Inverse Mills Ratio (IMR). The second step of the Heckman method is the insertion of the IMR variable into the initial regression models (e.g. those not accounting for potential bias due to missing surveys) from a given study in order to assess for, and attempt to control for, selection bias. Heckman method: Step one The first step in the Heckman method is to create the selection model, which estimates whether or not the quality of life survey was completed. The Heckman selection equation is usually estimated using a probit estimator [ 5 , 21 ]. The probit estimator requires a binary outcome variable, in this case whether the patient responded to the SF-36 or not (coded 1 for responder, 0 for non-responder). The candidate predictor variables for the selection model were those listed in Table 1 . Although the Heckman selection equation will usually have multiple variables, some of which will be the same variables that enter the multivariable models of HRQL outcomes, it is important that the Heckman selection equation contain at least one variable that can legitimately be excluded from the initial models to safeguard against colinearity between the Heckman selection equation and the initial regression models. This means that this variable (or set of variables) is, in theory, a factor influencing whether someone responded to the questionnaire, but not a factor in predicting their component scores on the SF-36. This variable or set of variables is called an instrument in econometrics, and should be a strong predictor of response in the selection equation. We therefore stress that it is essential that the candidate variables considered for the Heckman selection equation be as comprehensive as possible, not omitting any variables that may contribute to whether a person responds to the survey. Once the Heckman selection equation is estimated, the residuals (error term) from this equation are used to form a new variable called the Inverse Mills Ratio (IMR). The formula to create the IMR variable depends on the distributional assumption in the outcome equation. In most HRQL applications, the quality of life score is the outcome of interest and is usually estimated using multivariable linear regression. In this case, the distributional assumption of the error term is the standard normal distribution, so that the ratio of the standard normal probability density function (pdf) and cumulative density function (cdf) applied to the residuals for each individual in the data set is created. The ratio of pdf/cdf is the IMR. Each individual in the study sample receives an individual value of the IMR based on the residual observed for that individual in the selection equation. In this study, the value of the IMR for each individual represented the predicted probability that they completed the 7-month SF-36 survey. It is important to note that the IMR is a function not only of observed or measured variables that are included in the selection equation, but also of unobserved or unmeasured variables. These are captured through the error term or residual in the selection equation, and included through the non-linear function used to estimate the IMR. As a result, adding the IMR into the outcome equation introduces a term that attempts to capture both observed and unobserved variables that affect selection, or non-response. We estimated the Heckman model using the maximum likelihood estimation method in Stata version 8.0. In this approach, the outcome and selection models are estimated jointly, which can result in slightly different selection models for different outcomes, in this case the PCS and MCS scores from the SF-36. However, for clarity of presentation of the Heckman process, we present only one table of selection equation results (the probit estimation of the probability of returning the SF36), assuming that the patient survived to the 7-month survey point. Heckman method: Step two The second step of the Heckman method is to include the IMR as a separate predictor variable in the initial regression models. In this study, the IMR variable derived from our Heckman selection model was inserted into the initial PCS and MCS models. Once this variable is inserted, two factors can be evaluated to help determine whether there is significant bias from missing responses in the initial models. First, one can examine the significance of the IMR variable itself. If significant, it suggests there was significant bias in the initial model. However, one potential limitation of the Heckman method is that if the Heckman selection model is not well-specified, and the variables in the selection model do not predict response/non-response well, the IMR may be weaker than expected and the Heckman method may have limited power to detect bias. Therefore, a second factor to examine following the addition of the IMR variable into the initial outcome models is whether or not there have been significant changes in any of the parameter estimates of the other predictor variables in the model. While somewhat arbitrary, changes in parameter estimates of >10% may indicate that these estimates were biased due to missing surveys. Where possible, one should apply clinical judgment about whether changes in parameter estimates are 'biologically important' [ 22 ]. With these factors taken together, the insertion of the IMR variable into the initial risk models allows the assessment of whether or not there was bias in the initial models, and suggests which initial predictors may have been most associated with this bias. Furthermore, by including the effect of unmeasured as well as measured variables from the selection equation, bias due to selection is controlled. Results Baseline characteristics Compared to patients that completed the 7-month SF-36 survey, patients who did not respond to the survey were older, more likely to be current smokers and more likely to have a history of alcohol or substance abuse, dementia, stroke, or depression (Table 1 ). Furthermore, the non-responders were more likely to have had ST-segment elevation on their ECG, more likely to have been admitted to a tertiary care VA hospital, and were more likely to have had a do not resuscitate order during their index hospitalization. Survey non-responders were less likely to have a history of prior coronary artery bypass graft (CABG) surgery, prior percutaneous coronary intervention (PCI), or chronic obstructive pulmonary disease (COPD), and were less likely to receive coronary revascularization during index hospitalization. Initial PCS and MCS models The initial multivariable PCS and MCS models not accounting for potential selection bias from missing HRQL surveys are presented in Tables 3 and 5 . Table 2 Heckman Selection Model (N = 1605) 95% Confidence Interval Variables Coefficient Lower Limit Upper Limit Mean patient age 0.01 0.00 0.02 Male gender 0.44 -0.07 0.95 Caucasian race -0.06 -0.25 0.13 Hx of MI 0.02 -0.16 0.20 Hx of Chronic heart failure -0.01 -0.20 0.18 Hx of Coronary artery bypass graft surgery 0.09 -0.13 0.31 Hx of Percutaneous coronary intervention 0.28 0.01 0.54 History of trauma -0.29 -1.37 0.80 Substance or alcohol abuse -0.21 -0.40 -0.01 Current smoker -0.10 -0.29 0.09 Dementia 0.02 -0.76 0.79 COPD 0.23 0.05 0.41 Stroke -0.12 -0.39 0.15 Depression -0.26 -0.48 -0.05 Arthritis 0.08 -0.07 0.23 Diabetes mellitus -0.02 -0.25 0.21 Peptic ulcer disease 0.04 -0.20 0.28 Coronary revascularization during index admission 0.12 -0.06 0.30 Cardiogenic shock 0.04 -0.81 0.89 Hypotensive episode 0.05 -0.29 0.38 Do not resuscitate order -0.68 -1.09 -0.27 Positive stress test -0.01 -0.25 0.24 Admitted to Tertiary VA -0.09 -0.28 0.10 ST segment elevation MI on ECG -0.04 -0.29 0.22 Mean serum creatinine 0.01 -0.06 0.07 Left bundle branch block on ECG -0.02 -0.41 0.38 Cardiac catheterization during index admission 0.04 -0.10 0.19 Discharge diagnosis Unstable Angina (vs. MI) 0.06 -0.20 0.32 MI: myocardial infarction; Hx: History; COPD: chronic obstructive pulmonary disease; ECG: electrocardiogram; Instruments not included in outcomes equations Table 3 Initial PCS Model 95% Confidence Interval Variables Coefficient Lower Limit Upper Limit Mean patient age (per 1 year increment) -0.06 -0.11 -0.01 Male gender -1.04 -4.31 2.23 Caucasian race -0.69 -1.72 0.35 Hx of MI -0.85 -2.46 0.75 Hx of Chronic heart failure -2.96 -4.36 -1.55 Hx of Coronary artery bypass graft surgery -3.12 -4.29 -1.95 Hx of Percutaneous coronary intervention -1.09 -2.68 0.51 Substance or alcohol abuse -0.30 -1.89 1.30 Current smoker -0.57 -1.66 0.52 Dementia -0.84 -3.17 1.50 COPD -3.76 -4.77 -2.76 Stroke -3.01 -4.34 -1.67 Depression -4.61 -6.07 -3.15 Arthritis -1.89 -2.94 -0.83 Diabetes mellitus -2.33 -3.56 -1.09 Coronary revascularization during index admission 1.91 0.38 3.44 Cardiogenic shock 1.62 -3.88 7.12 Hypotensive episode 0.13 -1.93 2.19 Positive stress test 1.15 -0.47 2.76 Admitted to Tertiary VA -0.88 -2.00 0.24 ST-segment elevation MI on ECG 2.34 0.81 3.86 Mean serum creatinine (per 1 mg/dl increment) -0.64 -1.13 -0.16 Left bundle branch block on ECG 0.09 -2.43 2.61 Cardiac catheterization during index admission 0.63 -0.45 1.71 Discharge diagnosis Unstable Angina (vs. MI) -0.66 -1.54 0.21 Inverse Mills Ratio * * Inverse Mills Ratio: Variable derived from the Heckman Selection Equation; MI: myocardial infarction; Hx: History; COPD: chronic obstructive pulmonary disease; ECG: electrocardiogram Table 4 PCS Model Corrected For Response Bias (N = 1605) 95% Confidence Interval Variables Coefficient Lower Limit Upper Limit Mean patient age (per 1 year increment) -0.07 -0.11 -0.03 Male gender -1.00 -4.05 2.04 Caucasian race -0.67 -1.62 0.28 Hx of MI -0.89 -2.39 0.62 Hx of Chronic heart failure -2.95 -4.22 -1.69 Hx of Coronary artery bypass graft surgery -3.16 -4.23 -2.09 Hx of Percutaneous coronary intervention -1.24 -2.75 0.28 Substance or alcohol abuse -0.17 -1.85 1.52 Current smoker -0.48 -1.48 0.51 Dementia -0.77 -2.89 1.34 COPD -3.91 -4.91 -2.91 Stroke -2.92 -4.14 -1.70 Depression -4.68 -6.06 -3.31 Arthritis -1.71 -2.75 -0.67 Diabetes mellitus -2.33 -3.45 -1.20 Coronary revascularization during index admission 1.76 0.32 3.19 Cardiogenic shock 1.44 -3.72 6.60 Hypotensive episode 0.48 -1.50 2.46 Positive stress test 1.06 -0.38 2.50 Admitted to Tertiary VA -0.78 -1.81 0.25 ST-segment elevation MI on ECG 2.41 0.97 3.85 Mean serum creatinine (per 1 mg/dl increment) -0.66 -1.11 -0.22 Left bundle branch block on ECG 0.05 -2.30 2.39 Cardiac catheterization during index admission 0.60 -0.41 1.60 Discharge diagnosis Unstable Angina (vs. MI) -0.62 -1.42 0.19 Inverse Mills Ratio * -2.15 -5.66 1.37 * Inverse Mills Ratio: Variable derived from the Heckman Selection Equation; MI: myocardial infarction; Hx: History; COPD: chronic obstructive pulmonary disease; ECG: electrocardiogram Table 5 Initial MCS Model 95% Confidence Interval Variables Coefficient Lower Limit Upper Limit Mean patient age (per 1 year increment) 0.08 0.04 0.13 Male gender 0.14 -3.07 3.36 Caucasian race -0.69 -1.51 0.12 Hx of MI -1.19 -2.36 -0.02 Hx of Chronic heart failure 0.01 -1.05 1.07 Hx of Coronary artery bypass graft surgery 0.27 -0.94 1.48 Hx of Percutaneous coronary intervention 0.49 -0.76 1.75 Substance or alcohol abuse -2.27 -3.92 -0.62 Current smoker -0.51 -1.70 0.67 Dementia -2.80 -6.71 1.10 COPD -1.14 -2.15 -0.12 Stroke -2.10 -4.78 0.58 Depression -1.13 -2.56 0.30 Arthritis -11.68 -13.60 -9.76 Diabetes mellitus -0.49 -1.55 0.57 Coronary revascularization during index admission 1.09 -0.17 2.35 Cardiogenic shock -2.68 -8.31 2.95 Hypotensive episode -0.09 -2.52 2.33 Positive stress test -0.01 -2.44 2.43 Admitted to Tertiary VA 0.02 -1.37 1.40 ST-segment elevation MI on ECG 1.88 0.64 3.12 Mean serum creatinine (per 1 mg/dL increment) -0.34 -1.00 0.32 Left bundle branch block on ECG -1.87 -4.72 0.97 Cardiac catheterization during index admission -0.40 -1.62 0.81 Discharge diagnosis Unstable Angina (vs. MI) 0.24 -0.97 1.46 Inverse Mills Ratio * * Inverse Mills Ratio: Variable derived from the Heckman Selection Equation; MI: myocardial infarction; Hx: History; COPD: chronic obstructive pulmonary disease; ECG: electrocardiogram Variables significantly associated with better 7-month physical health status included ST-segment elevation MI on electrocardiogram and coronary revascularization during the index ACS hospital admission. Variables significantly associated with worse 7-month physical health status included older age, history of prior CABG surgery, chronic heart failure, arthritis, COPD, stroke, depression, diabetes, and elevated serum creatinine during index ACS admission. Variables significantly associated with better 7-month mental health status included older age and ST-segment elevation MI on electrocardiogram. Variables significantly associated with worse 7-month mental health status included a history of prior MI, alcohol and/or substance abuse, COPD, and arthritis. Heckman selection model The Heckman selection model (modeling response to the SF-36) is presented in Table 2 . Older age, prior PCI, and history of COPD were associated with an increased likelihood of survey response, whereas a history of alcohol or substance abuse, depression, and have had a do not resuscitate order during their index hospitalization were associated with a decreased likelihood of survey response. Final PCS and MCS models The final multivariable models for PCS and MCS (after addition of the IMR variables from the Heckman selection model) are presented in Tables 4 and 6 . There was little evidence of selection bias for the PCS model. None of the results of inference testing for significance changed between the initial model and the model with the IMR variable added. Furthermore, the changes in magnitude of parameter estimates were not large, and the parameter estimate on the IMR variable itself was not significant. Table 6 MCS Model Corrected For Response Bias (N = 1605) 95% Confidence Interval Variables Coefficient Lower Limit Upper Limit Mean patient age (per 1 year increment) 0.06 0.00 0.12 Male gender 0.25 -2.65 3.15 Caucasian race -0.51 -1.24 0.22 Hx of MI -1.24 -2.07 -0.40 Hx of Chronic heart failure 0.17 -0.94 1.27 Hx of Coronary artery bypass graft surgery -0.01 -1.21 1.18 Hx of Percutaneous coronary intervention -0.19 -1.36 0.99 Substance or alcohol abuse -1.55 -3.39 0.29 Current smoker -0.17 -1.46 1.13 Dementia -2.81 -6.90 1.27 COPD -1.76 -2.80 -0.72 Stroke -1.75 -4.25 0.75 Depression -1.33 -2.71 0.04 Arthritis -10.80 -12.83 -8.78 Diabetes mellitus -0.35 -1.43 0.72 Coronary revascularization during index admission 0.70 -0.48 1.89 Cardiogenic shock -2.58 -7.28 2.12 Hypotensive episode -0.16 -2.26 1.94 Positive stress test 0.01 -2.24 2.25 Admitted to Tertiary VA 0.24 -1.10 1.58 ST-segment elevation MI on ECG 1.92 0.91 2.92 Mean serum creatinine (per 1 mg/dL increment) -0.37 -0.97 0.24 Left bundle branch block on ECG -2.00 -4.63 0.63 Cardiac catheterization during index admission -0.46 -1.63 0.72 Discharge diagnosis Unstable Angina (vs. MI) 0.04 -1.23 1.32 Inverse Mills Ratio * -8.93 -10.73 -7.13 * Inverse Mills Ratio: Variable derived from the Heckman Selection Equation; MI: myocardial infarction; Hx: History; COPD: chronic obstructive pulmonary disease; ECG: electrocardiogram By contrast, when the IMR variable was inserted into the initial MCS model, we found evidence of selection bias. In this case, the parameter estimate for history of alcohol or substance abuse changed from significant to insignificant with the introduction of the IMR variable. Therefore, it appears that a history of alcohol or substance abuse was associated with lower likelihood of responding to the survey, but not directly associated with mental health status. In addition, a number of parameter estimates changed quantitatively, with larger changes than those observed in the PCS findings. Finally, the coefficient on the IMR variable itself was significant, and was negatively associated with MCS, implying that unobserved variables in the selection equation appear to be associated with worse MCS scores. Discussion The goal of this study was to demonstrate the use of the Heckman two-step method to assess and correct for bias due to missing HRQL surveys in a clinical study of acute coronary syndrome patients. We created initial multivariable models of 7-month post-ACS physical and mental health status using data only from survey respondents. Then, using the Heckman method, we modeled survey non-response, derived an Inverse Mills Ratio variable for each patient that captured the likelihood of survey response, and incorporated this variable into our initial models to assess and correct for potential bias from survey non-response. We found that our initial model of 7-month physical health status was not biased due to survey non-response. In contrast, our initial model of 7-month mental health status was biased. After controlling for bias due to non-response, a history of alcohol or substance abuse was no longer associated with mental health status. Furthermore, the magnitude of the parameter estimates for several of the other predictor variables in the MCS model changed after accounting for bias due to survey non-response. Given these results, biased parameter estimates of the association between these variables and mental health status would have been reported if we had used the standard approach to evaluating the predictors of HRQL outcomes in this population. Furthermore, we might have concluded that alcohol/substance abuse was significantly associated with mental health status outcomes following ACS, and may therefore be an important target for interventions to improve post-ACS HRQL (e.g. improving alcohol screening and treatment). While alcohol/substance abuse may be important for other reasons, it would have been incorrect to conclude that it was associated with HRQL in our study population. Rather, it was a marker for survey non-response. This analysis demonstrates the utility of the Heckman method in its application for assessing and correcting survey response bias in clinical studies of HRQL. Health-related quality of life data are usually not missing at random, and failure to account for missing HRQL assessments can bias estimates of associations and may lead to inappropriate conclusions about the determinants of HRQL outcomes [ 1 - 3 ]. Often, HRQL data are missing in systematic ways that can be estimated and controlled for. This study demonstrates the use of one technique to accomplish this, the Heckman two-step method [ 5 - 8 ]. To date, the Heckman method has rarely been utilized in studies reported in the medical literature, although it was previously used in one study assessing the impact of selection on medication use among older patients [ 7 ]. There are other statistical techniques, or approaches, to assess and correct for bias resulting from missing HRQL surveys, including index function models, propensity scores, instrumental variables, and multiple imputation methods [ 2 - 4 , 23 ]. The Heckman method is one example of an index function model. Generally, the index function approach to missing HRQL data is to model whether or not HRQL surveys were completed (i.e. the dependent variable is survey completion). This allows an estimation of the 'likelihood' that a given patient would complete a survey based on their clinical characteristics and/or other process or structure of care variables. This information, in turn, is used to assess and correct for bias in the primary model of interest (i.e. the model of quality of life outcome). Therefore, a primary strength of the Heckman method is that it not only permits the assessment of selection bias, it corrects for the bias, and does so in an informative way that may yield new insights into the association between patient characteristics or processes of care and outcomes of interest such as HRQL. In the Heckman method, the assumption is made that the error term in the outcome equation is standard normal, the distribution assumed in classical linear regression. Other index functions allow other distributional assumptions to be made for the error term in the outcome equation, such as logistic. Propensity score approaches can be analogous to the Heckman method in that a multivariable model of survey non-response is developed and the probability of survey non-response is used to stratify the study population and/or the propensity score is used as an independent variable in the primary HRQL models. In other words, propensity scores are similar to the Heckman method in that the predicted probability of non-response is used as the basis for assessing the impact of missing data and controlling for it [ 23 ]. Unlike a propensity score, however, which is often entered directly into the outcome equation as a predictor, the non-linear transformation from the prediction into the IMR variable in the Heckman method is one of the safeguards against colinearity in the outcome equation. Instrumental variable approaches are also used to address similar questions to those addressed by the Heckman method. In the full instrumental variable approach, a single exogenous variable (called the instrument) is used to stratify the full sample, removing the effect of the correlated error terms that lead to biased estimates [ 24 ]. An instrumental variable approach can be a very powerful approach to controlling selection bias. However, it can be very difficult to find an appropriate and adequate instrumental variable. The Heckman approach offers a more flexible, if less powerful, approach, and adds information about the underlying processes by which selection arose. It should be noted that propensity scores can be used as instrumental variables, when a suitable instrument is found. Finally, multiple imputation methods can be employed to address missing HRQL data [ 2 ]. In contrast to the Heckman and other approaches described thus far, multiple imputation methods derive missing values from existing data in the dataset, thereby creating a 'complete' dataset and eliminating the need to drop patients from analysis. Imputation thereby eliminates bias from missing data per se (i.e. there is no longer missing data), but is highly dependent on the validity of deriving the missing HRQL survey data from the existing dataset. It is important to note that in this paper, we are focused on missing surveys rather than incomplete surveys (i.e. missing data elements within a survey). In this regard, multiple imputation will most often be employed in studies with serial measurements of HRQL over time, such that HRQL data before and/or after the time point of interest can inform the missing data imputation. The Heckman method can be used even for a single point in time, cross-sectional assessment of HRQL, as in our analysis in which we measured HRQL at only one time point. The Heckman method has several limitations. First, the selection equation must have at least one variable that is associated with survey response but not the outcome of the study (i.e. HRQL). In some clinical applications, the inability to identify such a variable may make it difficult to use the full Heckman method to control bias. However, it is still possible to use the first step of estimating a selection equation to assess the degree to which selection bias may affect the parameter estimates in an outcome equation. If there are variables that are significant in both the selection equation and the outcomes equation, it is likely that there is bias due to selection effects in the outcome equation. Acknowledging this and commenting on the likely magnitude of effect may provide helpful guidance to clinicians and other researchers. Another limitation of the Heckman method is that this technique depends heavily on the quality of the data available for the selection equation. If the amount of variance explained is relatively low, then there is a possibility that selection bias in the outcomes equation may not be detected. In other words, the Heckman method can be under-powered for the detection of bias in some cases. Finally, the Heckman method is very sensitive to how the model is specified; in other words, omitting variables that are associated with either non-response or with the outcome of interest (in this case, health related quality of life measures) can lead to inaccurate findings and biased estimates of the parameters in the final models. Careful attention to specifying the models, and ensuring that model specification follows what is known in the literature to be associated with the outcomes of interest is essential [ 4 ]. Conclusions This study demonstrated the use of the Heckman two-step method to assess and control for bias from missing HRQL surveys in a clinical study. We found that our mental health status model was significantly biased due to missing HRQL assessments. Recognition and correction of this bias changed the parameter estimates of association and the factors that we concluded were associated with HRQL seven months following hospital admission for an acute coronary syndrome. We conclude that the Heckman two-step method may be a valuable tool in the assessment and correction of selection bias in clinical studies of HRQL. Authors' contributions AES conceived of the study, participated in design and coordination, conducted analyses, and drafted the manuscript; MEP conducted analyses and contributed to the manuscript; DJM and JAS reviewed and contributed to the manuscript; JSR participated in the design and coordination of the study and participated in the drafting and revision of the manuscript. All authors read and approved the final manuscript.	/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC521693.xml
524270	Serial Block-Face Scanning Electron Microscopy to Reconstruct Three-Dimensional Tissue Nanostructure	Three-dimensional (3D) structural information on many length scales is of central importance in biological research. Excellent methods exist to obtain structures of molecules at atomic, organelles at electron microscopic, and tissue at light-microscopic resolution. A gap exists, however, when 3D tissue structure needs to be reconstructed over hundreds of micrometers with a resolution sufficient to follow the thinnest cellular processes and to identify small organelles such as synaptic vesicles. Such 3D data are, however, essential to understand cellular networks that, particularly in the nervous system, need to be completely reconstructed throughout a substantial spatial volume. Here we demonstrate that datasets meeting these requirements can be obtained by automated block-face imaging combined with serial sectioning inside the chamber of a scanning electron microscope. Backscattering contrast is used to visualize the heavy-metal staining of tissue prepared using techniques that are routine for transmission electron microscopy. Low-vacuum (20–60 Pa H 2 O) conditions prevent charging of the uncoated block face. The resolution is sufficient to trace even the thinnest axons and to identify synapses. Stacks of several hundred sections, 50–70 nm thick, have been obtained at a lateral position jitter of typically under 10 nm. This opens the possibility of automatically obtaining the electron-microscope-level 3D datasets needed to completely reconstruct the connectivity of neuronal circuits.	Introduction Many cellular structures are so small that they can only be resolved in the electron microscope. Furthermore, it is often crucial to visualize and reconstruct the three-dimensional (3D) structure of biological tissue. One prime example of where 3D information is indispensable is in exploring the connectivity of local networks of neurons. While axonal and dendritic processes have been traced using the light microscope from the very beginning of cellular neuroscience (Cajal 1911), light microscopic tracing is only possible if staining is restricted to a small subset of cells, as results, for example, from the Golgi method ( Golgi 1873 ) or from the mosaic expression of fluorescent proteins ( Feng et al. 2000 ). However, in many cases, in order to understand computational algorithms, the reconstruction of a complete neural circuit may be necessary. For this, the resolution of the light microscope is insufficient because dendritic and axonal processes can have diameters that are substantially below the wavelength of light. This lack of resolution (1) results in the inability to resolve densely packed neighboring processes, which is absolutely necessary to reconstruct network topology, and (2) does not allow a sufficiently precise estimation of the neuronal geometry, which may be necessary for biophysical modeling of cellular behavior. So far, only the electron microscope (EM) can provide the spatial resolution needed to track neural processes or to identify synapses unambiguously. Most commonly used to image biological tissue is the transmission electron microscope (TEM) ( Ruska and Knoll 1932 ), in which a broad beam of electrons is directed at a sample that is thin enough to allow a substantial fraction of the electrons to pass through and then be focused onto film or another electron-sensitive spatially resolving detector. Specimens are typically thin slices that are cut from blocks of plastic-embedded tissue, with the resulting electron micrographs providing a two-dimensional cross section through the tissue. Scanning electron microscopy (SEM) ( Ardenne 1938a , 1938b ), in which a tightly focused beam of electrons is raster-scanned over the specimen while secondary or backscattered electrons are detected, is used in biological imaging mostly as a surface visualization tool, creating a 3D appearance but no actual 3D datasets. Truly 3D information in the TEM can be obtained using either tilt-series-based tomography ( Hoppe 1981 ; Frank 1992 ; Baumeister 2002 ) or serial ultrathin sections ( Sjostrand 1958 ; Ware and Lopresti 1975 ; Stevens et al. 1980 ). Tomography is a very promising technique for obtaining high-resolution structural data of macromolecules, organelles, and small cells but may not be applicable when larger volumes need to be reconstructed, because section thickness is limited to around 1 μm. In the high-voltage EM, somewhat thicker sections can be viewed, which facilitates the correlation with light microcopy ( Martone et al. 2000 ). Serial-sectioning TEM, which can be used to obtain 3D data of much thicker volumes than is possible with tomography, is a mostly mature technology and has contributed enormously to our understanding of the local 3D ultrastructure, for example of dendritic spines ( Harris 1999 ). A whole animal, the nematode Caenorhabditis elegans has been reconstructed in this way ( White et al. 1986 ). This is still considered a seminal effort, partly because serial sectioning is such a labor-intensive and time-consuming method, which requires intensive operator involvement in cutting sections, gathering data, and reconstructing volumes. The number of sections that need to be handled can be reduced by combining tomography with serial sectioning ( Soto et al. 1994 ), but the need to manually handle the sections remains. For these reasons the number of larger-scale 3D serial reconstructions has been rather limited, even though many biological problems, not only the tracing of neural circuits, require truly 3D information. Volume information equivalent to that from reconstructing serial sections can, of course, be obtained if the sections are imaged before being cut, that is, by repeatedly imaging the block face. Block-face imaging is used even in light microscopy ( Odgaard et al. 1990 ; Ewald et al. 2002 ), where optical-sectioning techniques, such as confocal ( Minski 1961 ) or multi-photon microcopy ( Denk et al. 1990 ), are readily available. It is, of course, impossible to image the block face in the TEM. The SEM, as a surface-imaging technique, is, on the other hand, well suited for this task. This was recognized several decades ago by Leighton, who also constructed a microtome for cutting sections inside the microscope chamber ( Leighton 1981 ). Imaging the block face removes not only the need to deal manually with ribbons of fragile sections but also the difficulty of aligning the images of sections, which are often distorted. The prevalent contrast mode in the SEM is the detection of so-called secondary electrons, which are low-energy electrons that are emitted when the primary electron beam strikes the sample surface. The secondary-electron signal depends strongly on the orientation of the surface, leading to topographic images that characteristically resemble obliquely illuminated solid objects. Since the microtome-cut block surface is devoid of topographic features, very little contrast is generated unless an additional preparatory step, such a plasma etching ( Kuzirian and Leighton 1983 ; Hukui 1996 ), is used. At the time of Leighton's original work, no stacks of volumetric data were presented. One reason might have been that low-vacuum SEMs (also called atmospheric [ Robinson 1975 ; Danilatos 1980 ] or environmental SEMs; for a recent review see Donald [2003] ), which can be used to image nonconducting specimens (see below), were not widely available. In the original report ( Leighton 1981 ), the sample had to be removed from the SEM chamber for coating with a conducting layer. Furthermore, digital image acquisition, storage, and processing were in their infancy and quite limited in capacity. In this paper we show, first, that by using backscattering contrast and low-vacuum operation we can obtain high-contrast images from cut block faces and, second, that in combination with a custom-designed microtome we can obtain 3D ultrastructural data with a resolution and from volumes appropriate for the 3D reconstruction of local neural circuits. Results As a first step, we explored whether sufficient contrast and stable images can be obtained from uncoated block faces of plastic-embedded tissue, prepared essentially as for TEM. There are a number of techniques known that allow imaging of nonconducting specimens in the SEM. One technique relies on choosing an accelerating voltage at which the introduction of charge by the electron beam is just balanced by the sum of backscattered and secondary electrons (for a review see Joy and Joy [1996] ). We first tried the charge-balance method (electron energies around 600 eV) but found that the secondary-electron signal provides, not surprisingly, unsatisfactory contrast (data not shown). The secondary-electron signal is the only contrast available because low-energy backscattered electrons (BSEs) cannot be easily detected at such low voltages. Furthermore, residual charging, which may not be strong enough to preclude the formation of an image, can still lead to slight shifts between images by electrostatic beam deflection. This then destroys the alignment between successive images, which is one of the main reasons to establish serial block-face imaging. Another set of techniques relies on providing ions in the specimen chamber to neutralize charge on the sample rather than avoiding net charge introduction by the electron beam. This can, for example, be achieved by maintaining a low concentration of gas in the chamber ( Robinson 1975 ; Moncrieff et al. 1978 ) at a pressure (10–100 Pa) that is low enough for a substantial fraction of the beam's electrons to reach the sample unscattered and thus be able to form a tight focus. Ions needed for discharging the sample are generated by those electrons that do strike gas molecules or atoms. The residual-gas method (also called low-vacuum or variable-pressure mode) can be used at higher accelerating voltages and is, therefore, compatible with BSE contrast. Since scattering of electrons is strongly dependent on the charge of the atomic nucleus, the BSE signal provides a clear distinction between heavy-metal-stained and unstained structures ( Figures 1 and S1 ) that, after contrast reversal, look very similar to traditional transmission electron micrographs. Images of uncoated block faces ( Figure 1 ) resemble scanning electron micrographs obtained from block faces of similar specimens coated with a very thin conducting film and imaged under high-vacuum conditions ( Richards and ap Gwynn 1996 ). Our BSE images also demonstrate that the resolution under these conditions is sufficient to resolve structures such as synaptic vesicles ( Figure 1 A) or the array of actin filaments characteristic of skeletal muscle ( Figure 1 B). Synaptic contacts in cortical tissue can be clearly identified as well ( Figure 1 D and 1 E). Figure 1 Resolution and Contrast Using the Backscattered Electron Signal (A and B) Presynaptic vesicles (SV) and postsynaptic folds (SF) are clearly visible (A) in a motor endplate preparation embedded in Spurr's resin. Similarly, the hexagonal array of actin filaments (AA) can be clearly resolved (B) in a different region from the same image (both images were smoothed using the ImageJ “smooth” command). Imaging conditions for (A) and (B): electron energy, 7.5 keV; spot, 3.5; chamber pressure, 30 Pa (H 2 O); pixel dwell time, 30 μs. The scanning resolution was 6.7 nm/pixel. (C) The effect of beam exposure on the block surface. Note the increased brightness and the lack of chatter in the central region (inside the dashed rectangle), from which a stack was acquired at higher resolution before taking the image shown. The tissue was rat neocortex embedded in Spurr's resin. Imaging conditions for (C): electron energy, 7.5 keV; spot, 3; digital resolution for stack acquisition, 26.7 nm/pixel; dwell time, 30 μs. (D and E) Cortical tissue embedded in Epon. Synapses (SD) are clearly discernable (E). Imaging conditions for (D) and (E): electron energy, 7.5 keV beam current; spot, 3; chamber pressure, 30Pa (H 2 O); pixel dwell time, 30 μs. The scanning resolution was 9.5 nm/pixel. Note that more backscattering corresponds to darker pixels in (A), (B), (D), and (E) but to brighter pixels in (C). These results show that we can obtain sufficient contrast from uncoated block faces. On this basis we decided to construct an ultra-microtome appropriate for sequential sectioning inside the sample chamber of an SEM. There are a number of special requirements for such a microtome. (1) The position of the block has to remain stable or be returned to the same location after each image is taken in order to ensure alignment of subsequent images, which is one of the primary objectives of serial block-face imaging. (2) The distance between the block face and the SEM objective lens has to be sufficiently small to allow high-resolution imaging and the efficient collection of backscattered electrons under low-vacuum conditions. (3) The block face has to be perpendicular to the optical axis of the electron optical column to keep the whole block face in focus; for all commercial SEM instruments this means horizontal orientation. (4) The operation of the microtome must be remotely controllable to allow automation. (5) The microtome has to be compatible with low vacuum. Finally, (6) the range of the fine advance mechanism has to be at least several hundred microns to permit continuous sectioning of volumes large enough to be of interest in the study of local neuronal circuits. Conditions 1, 2, and 3 rule out the modification of a conventional ultra-microtome, in which the sample block, its face oriented vertically, is moved against a stationary knife. Condition 2, in addition, rules out the use of a conventionally mounted diamond knife, since the boat would protrude upwards, colliding with the BSE detector and even with the objective lens. We, therefore, decided to construct a microtome de novo ( Figure 2 ; for details see Materials and Methods ), which does share some design features with the instrument built by Leighton (1981) . A custom knife that could fit under the detector was designed in cooperation with the Diatome company and fabricated by Diatome. To meet condition 1 above, the knife is moved for cutting. In the z-direction, sample advance rather than knife advance is used in order to keep the cutting plane, and hence the location of the block surface, constant. This removes the need for refocusing the SEM, thereby contributing to image stability and registration. To avoid sliding friction, crossed leave flexures were used for the knife arm and for the sample advance mechanism. To provide advance motion with sufficiently fine resolution, we used a lever mechanism that scaled the motion of a motorized micrometer down by a factor of roughly 1/9. To minimize heat production, which can lead to thermal drift and is hard to dissipate in vacuum, we used a direct-current motor drive. Figure 2 SEM Microtomy (A) Principle of SBFSEM operation: (1) a SEM image is taken of the surface of the plastic-embedded tissue preparation (amber trapezoid). (2) Then with a diamond knife (blue) an ultrathin slice is cut off the top of the block. (3) After retraction of the knife, the next picture is taken. The pictures shown are from an actual stack (cerebellar cortex) but are not successive slices; rather, they are spaced by five images (about 315 nm) to make the changes more apparent. (B) Usually cut-off slices pile up on the top of the knife. Protruding into the picture from the right is a puffer pipette, occasionally used to remove debris from the knife. (C and D) The mechanical design for the in-chamber microtome is shown in an overview (C) and a close-up of knife and sample (D) in renderings from the computer-aided design software. Most parts are nonmagnetic stainless steel (grey). A large-motion leveraged piezo actuator (green part on the left) drives the knife holder back and forth. The custom diamond knife (light blue) is clamped in a special holder. The sample (amber) advance is driven via a lever by a direct-current-motor-driven micrometer (dark blue). The retraction during the backwards knife motion is again piezo actuated (green cylinder in the lower right of [C]). Bearing springs are brown. The BSE detector (red) is depicted schematically above the sample. Not shown is the lateral positioning mechanism. Preparation of samples for the serial block-face SEM (SBFSEM) was done using procedures common for preparing samples for viewing in the TEM (see, for example, Hayat 2000 ). Since SBFSEM does not allow contrast enhancement after cutting, which is often used in the TEM, we tested several methods of enhancing the contrast of the whole block and found that exposing the tissue to uranyl acetate (see Materials and Methods and also Hayat [2000] , p. 342ff) leads to excellent contrast in BSE mode (see Figure 1 ). We could reliably cut serial sections thicker than about 50 nm, the exact lower thickness limit depending on the embedding material and the beam exposure (an upper thickness limit was not established). When trying to cut thinner sections the actual thickness became uneven or the knife would skip every other cut, as recognized by the lack of change in structural detail. Another indication of uneven cutting is alternating brightness of the whole images or image regions. The reason for this is that the backscattering efficiency generally increases after beam exposure (see Figure 1 C), which may be because a selective ablation of the embedding matrix leads to an effective increase in the heavy-metal concentration. Usually the cut-off slices pile up on the top surface of the knife (see Figure 2 B). Only rarely does a sliver get deposited on the block surface. When this happens it results in the loss of only one image since the sliver is reliably removed by the next passage of the knife. During the cutting of the stack used for Figure 3 A and 3 B, three instances of deposited debris occurred (white horizontal streaks in Figure 3 B). Currently efforts are underway to devise methods (such as using a brief puff of gas from a glass pipette, such as can be seen in Figure 2 B) to reliably remove the debris from the knife after each cut. Figure 3 3D Datasets: Five Slices (A) Five slices from a stack containing a total of 365 slices at 63-nm section thickness; same tissue as in Figure 1 A and 1 B. Note beginning of myelin sheath (MS) in the lowest slice. Imaging conditions as in Figures 1 A and 1 B except resolution is 13.4 nm/pixel. (B) Reslice of the same dataset along the red line show in the top image of (A). Arrows point to the presynaptic ending (PS) and the z-band (ZB). Image corners in (A) touch the reslice (B) at the depths at which they were taken. (C and D) Cerebellar tissue displayed at low (C) and high (D) resolution; note nuclear envelope (NM). Note differences in lateral and vertical resolution. Imaging conditions for (C) and (D): electron energy, 7.5 keV; spot size, 3.5; digital resolution, 12.7 nm/pixel. We found, somewhat surprisingly, that cutting quality was often better in the area scanned by the electron beam than outside (see Figure 1 C), presumably because of a modification of the mechanical properties of the embedding material. We have successfully cut a number of different embedding materials (Araldite, Epon, and Spurr) and tissue types (mammalian muscle, cortex, cerebellum, retina, zebrafish brain, and fly brain) using our SBFSEM setup. Figure 3 shows examples of 3D datasets. Since lateral registration between successive slices is crucial for the applicability of the SBFSEM technique for automated 3D reconstruction, we quantified the registration using cross correlation. Figure 4 shows the registration precision throughout a stack of images from muscle tissue. The fluctuations are mostly around 10 nm in size (standard deviation). Large fluctuations (greater than 60 nm) are seen occasionally but can be traced to temperature drifts or are spurious values caused by debris on the block face. Some of the apparent position shifts and fluctuations do not, furthermore, represent registration errors but are caused by systematic or random pattern shifts. An example is the shift of the central cloud of points in Figure 4 B, which is caused by the tilt of the actin fibers from the block-face normal. Figure 4 The Alignment of Successive Images in a Stack Shifts between images were quantified using the positions of the peaks of the cross correlation (see Materials and Methods ); same dataset as in Figure 3 A and 3 B. (A) The peak shifts in x-direction are shown for five different subregions distributed over the field of view. Four of the regions have a size of 256 × 256 pixels, one has a size of 512 × 512 (black trace). The peaks around slices 59 and 202 are caused by slice debris on the block face (see also streaks in Figure 3 B). (B) The x/y displacement for the 512 × 512 region is shown in a scatter plot. For the central cluster the standard deviations are 10.9 nm and 11.8 nm for x and y, respectively. The depth resolution depends, of course, on the section thickness but also on the depth below the sample surface from which the BSE signal originates, which, in turn, scales with the electron penetration (see chapter 3 of Goldstein et al. [2003] ). For our early experiments we had used the electron energy of 7.5 keV because lower voltages gave unsatisfactory signals. At 7.5 keV the resolution along the z-axis is not very good, as is apparent from Figure 3 D. To improve the z-resolution one could conceivably use deconvolution techniques because the point spread function contains a sharp edge and hence high spatial frequencies along the z-axis, even if the electron penetration into the sample is large. The reason is that while a structural detail can contribute to the signal if buried in the sample, the contribution from that detail drops immediately to zero as soon as the slice containing the detail is removed. Electron penetration depends strongly ( Kanaya and Okayama 1972 ) on the energy (scaling as E 1.67 ). This means, for example, that at 4 keV the depth is reduced to about one-third of that at 7.5 keV. The actual value for the depth penetration is difficult to estimate for a composite material such as plastic-embedded heavy-metal-stained tissue. For a detailed discussion of the issues surrounding depth penetration see, for example, chapter 3 of Goldstein et al. (2003) . We, therefore, decided to operate the SBFESM at lower voltages. In order to obtain sufficient signal at the lower electron energy, where the efficiency of semiconductor-diode BSE detectors declines steeply ( Funsten et al. 1997 ), we replaced the detector amplifier with an instrument originally designed to detect single ion-channel currents (commonly called a patch-clamp amplifier [ Hamill et al. 1981 ]) and found that we could now get satisfactory images for electron energies of 4 keV. In Figure 5 the depth resolutions at 7.5 and 4 keV are directly compared for the same sample. It is particularly encouraging that at 4 keV it is possible to recognize clearly membranes that are parallel to the block surface (en-face membranes), which is crucial for the successful reconstruction of cellular topology. Figure 5 Energy Dependence of the Depth Resolution The lateral resolution does not change very much as the electron energy is reduced from 7.5 keV (A) to 4 keV (B), but the resolution along the z-direction is dramatically different (C). The lines between (A), (B), and (C) indicate the z-positions in the stack from which (A) and (B) were taken. In the high-resolution view (lower panel in [C]) membranes that were en-face (EM) in the original slices can be clearly recognized. The nominal slice thickness was 55 nm. Tissue is rat cortex. Imaging conditions: dwell time, 25 μs; spot sizes, 3.5 and 2.8 for 4 and 7.5 keV, respectively. Discussion We have shown that fully automated acquisition of truly 3D datasets at nanoscopic resolution is possible with serial block-face imaging in the SEM. The lateral resolution is sufficient to recognize most cellular organelles and synaptic specializations and does not appear to be limited by the beam spot diameter but by the spread of the beam in the sample ( Kanaya and Okayama 1972 ; see also Joy and Joy 1996 ). This spread could be further reduced by using lower beam energies, provided that more sensitive detectors for such low electron energies can be obtained ( Funsten et al. 1997 ). Lower beam energies also would further improve the z-resolution, even though at some point elemental contrast in the BSE signal begins to decline ( Joy and Joy 1996 ). The resolution is, in practice, also limited by the radiation dose that a specimen can sustain. Since at lower electron energies the beam's energy is deposited into a rapidly shrinking volume, heating effects could become an issue. Nonetheless, it appears conceivable that nearly isotropic resolution can be achieved if the sectioning thickness can be reduced to around 20 nm (see below), with imaging then done at primary electron energies of around 2 keV. Because the registration between successive slices (see Figure 4 ) is mostly better than the resolution, the 3D data stacks can be used without further alignment and distortion correction (for the complete set of raw data see Datasets S1–S20 ; slices through this dataset are shown in Figures S2–S4 ) . This is crucial if automated reconstruction is to be performed, since alignment of serial sections often requires manually identified landmarks. It would be highly desirable to reliably cut sections thinner than 50 nm, which appears to be the current limit in our hands. The exact reason for this thickness limit is not clear as yet. Much thinner sections have been cut from plastic blocks ( Frosch et al. 1985 ), albeit under atmospheric pressure with the sections floated onto water. An important difference between standard ultra-microtomy and SBFSEM is that, because of the inevitable exposure of the block face to the electron beam, the mechanical properties of the block surface are changed, affecting cutting of subsequent sections. While at very high doses these effects are deleterious, they can be helpful at intermediate levels for suppressing chatter (see Figure 1 C). To suppress chatter and to achieve thinner sectioning, we have begun to use the oscillating-knife technique ( Studer and Gnaegi 2000 ) (see Figures 5 and S2–S4 ). It might also help to reduce the sample temperature, which, in addition to affecting cutting properties directly by changing the polymer stiffness, might substantially improve the resistance to radiation damage ( Lamvik 1991 ). SBFSEM might ultimately allow a smaller section thickness than is routinely possible with serial sectioning for the TEM, since there the mechanical integrity of the cut sections has to be ensured. Alternatively, one might use nonmechanical surface-ablation techniques such as plasma etching, reactive or focused ion beam etching, or ion milling to achieve smaller depth increments. Nonmechanical techniques do, however, suffer from a dependence of the ablation rate on the composition, which is, by the very nature of embedded tissue, rather inhomogeneous. In spite of the similarity in appearance, it is important to keep in mind a number of differences between transmission imaging of sections and backscattered SBFSEM. First, in the SEM, a large fraction of the entire beam energy is deposited inside the sample, leading to much increased radiation damage per primary electron. This is partly counteracted by the fact that BSE imaging is a “dark field” technique and, therefore, each detected electron carries more information. Furthermore, because the primary-electron energies can be lower by a large factor in the SEM (4 keV) than in the TEM (typically about 80 keV but routinely as high as 300 keV and up to 3 MeV in special cases), scattering cross sections are much larger in the SEM. This may, however, not always translate into improved contrast, since in the SEM only electrons scattered by angles larger than 90° can contribute to the image, while rather small deflections are enough to remove an electron from the beam in the TEM. A major practical advantage of SEM block-face imaging is that there is no need to manually handle sections, and no occlusion occurs by EM-grid bars. On the other hand, it is, unlike in TEM, impossible to reimage a section at higher resolution. Finally, it is unlikely that SEM surface imaging will ever reach the lateral resolution that can be achieved by imaging thin sections in the TEM. The total volume that can be reconstructed by SBFSEM is currently limited in the lateral dimension by the digital resolution available on commercial SEM instruments. It should, however, be straightforward to increase the digital resolution by modifying the scan and data-collection cir-cuitry. Eventually, off-axis electron-optical aberrations will become significant, but the total field of view can then be increased by mechanical translation of the microtome and tiling of multiple images. In the z-direction the stack height that can be cut continuously is, first, limited by the travel of the fine advance. Second, and more fundamentally, the cutting pyramid becomes unstable and too compliant if it is too high. However, a single step of image realignment after retrimming and repositioning will allow the continuation in the z-direction for almost unlimited distances. A major practical limitation is, of course, the acquisition speed of the SEM, which is ultimately dependent on the required signal-to-noise ratio (R SN ). We can estimate the minimally achievable voxel dwell time (τ d ) using the backscattering coefficient (η) as follows: where I B is the beam current and e is the elementary charge. The available beam current depends on the electron-gun type, the desired resolution, and the electron energy, but a value of I B = 1 nA is realistic for our purposes. We set η to 0.1, which is the value for carbon and thus a lower limit since backscattering is, of course, higher for the heavy-metal-stained areas ( Drescher et al. 1970 ; see also chapter 3 in Goldstein et al. [2003] ). To get R SN =100 we then need a dwell time of 16 μs, allowing data rates slightly above 50,000 voxels/s. A cube 200 μm on a side imaged at a resolution of 10 nm × 10 nm × 50 nm, which corresponds to 1.6 teravoxels, would then require a total scan time of 25,600,000 s, which is on the order of one year. However, the tracing of axons probably can, with the appropriate staining, be achieved at 20-nm lateral resolution and with an R SN of ten or less. This reduces the estimated time by a factor of 400, to less than one day. While in some cases the resolution provided by conventional (tungsten filament) electron guns may be sufficient, we feel that the long-term stability of field-emission emitters is essential for imaging large volumes. The approach described here will speed up the collection of medium-to-high resolution 3D electron microscopic datasets by several orders of magnitude. The acquired data will, in addition, not require time-consuming and error-prone alignment and distortion correction. This should have wide-ranging applications not only in biomedical research but also in materials characterization ( Alkemper and Voorhees 2001 ). We are particularly interested in the complete reconstruction of local neural circuits such as those that underlie the detection of motion in the retina ( Barlow et al. 1964 ; Euler et al. 2002 ). This requires the imaging of volumes containing at least one complete dendritic tree and is therefore virtually impossible using conventional electron microscopic methods. The data quality using current staining techniques is good enough to manually trace neuronal processes in many cases (see Figure S5 ). One of the major challenges will be the automation of data analysis, which could well require the development of novel or the further optimization of existing staining techniques that highlight the plasma membranes or the extracellular space ( Brightman and Reese 1969 ). It is quite likely that scanning-probe methods such as atomic-force or near-field microscopy could be used instead of SEM to image the block face. It would be particularly interesting if multiple fluorescent dyes could be detected and discriminated in this way since they still provide unmatched specificity in labeling biological samples. Materials and Methods Imaging All data shown were taken on a environmental SEM with a field-emission electron gun (QuantaFEG 200, FEI, Eindhoven, the Netherlands) mostly at an electron energy of 7.5 keV (exceptions as noted, see Figures 5 and S2–S4 , Datasets S1–S20 , and Video 2 ). For data in Figures 5 and S2–S4 , Datasets S1–S20 , and Video 2 , a highly sensitive current amplifier was used (Axopatch200B, Axon Instruments, Union City, California , United States) to amplify the current from the solid-state backscattered electron detector (type L2, K. E. Developments, Cambridge, United Kingdom). The beam current values for the parameters used were roughly interpolated from manufacturer's data, yielding estimated beam currents of 190, 430, and 880 pA for spot sizes of 2.5, 3, and 3.5, respectively, at a beam energy of 7.5 keV, and 150, 330, and 645 pA for spot sizes of 2.5, 3, and 3.5, respectively, at a beam energy of 4 keV. Video 2 Cortical Tissue A movie sequence through a stack of block-face images from cortical tissue (same dataset as in Figure 5 ). The beginning and the second half of the stack were taken at 4 keV electron energy, the middle part at 7.5 keV (see also Figure 5 C). The dimensions of this volume are 11.64 μm laterally and 9 μm vertically (see also Figure S5 ). Video 1 Neuromuscular Junction A movie sequence through a stack of block-face images from muscle tissue (same dataset as in Figures 1 A, 1 B, 3 A, and 3 B). Specimen preparation Muscle tissue was prepared as described in Schwarz et al. (2000) . For the preparation of rodent brain tissue the animals were perfused transcardially first with 30 ml of phosphate-buffered saline and then with 40 ml of fixative solution (4% paraformaldehyde in 0.1M PBS [pH 7.4]). The brain tissue was then removed and kept in fixative over night at 4 °C. After being washed twice in PBS, tissue slices (0.2 to 1.5 mm thick) were cut on a vibratome (752 M Vibroslice, Campden Instruments, Leichester, United Kingdom) and kept for 24 h in PBS at 4 °C. Pieces about 1.5 mm in size were then excised and washed three times for 30 min each in cacodylate buffer at pH 7.4.The tissue was postfixed for 2 h in 2% osmium tetroxide/1.5% potassium ferric cyanide in aqueous solution at room temperature. Then the tissue was subjected to a contrast enhancement step by soaking it over night in a solution of 4% uranyl acetate in a 25% methanol/75% water mixture ( Stempak and Ward 1964 ) at room temperature. After that the tissue was dehydrated in a methanol sequence (25%, 70%, 90%, and 100% for 30 min each) followed by infiltration of the epoxy (Spurr, Epon 812, or Araldite, all from Serva, Heidelberg, Germany) monomer (epoxy/methanol 1:1, for 3 h rotation at room temperature; epoxy/methanol 3:1, overnight at 4°C; pure epoxy, 3 h rotating at room temperature). Polymerization was 48 h at 60 °C for Epon and at 70 °C for Spurr and Araldite. The block face was trimmed to a width of several hundred microns and a length of about 500 μm using either a conventional microtome or a sharp knife. SEM images of the untrimmed block face can be used to select the desired field of view before the final trimming step producing the desired small cutting pyramid. Data analysis Reslicing of image stacks (see Figures 2 B– 2 D and 5C , as well as S2–S4 ) was done using the ImageJ reslicing command, which interpolates the z-axis data so that the digital resolution matches that of the lateral direction. For calibration of the z-axis, see below. Image shifts (see Figure 4 ) were measured by cross-correlating subregions (256 × 256 or 512 × 512) of subsequent slices. The peak positions were determined by first normalizing the cross correlations, then raising the values to the 64th power, and finally calculating the barycenter. All calculations were performed using ImageJ (version 1.32g). The neurite reconstruction shown in Figure S5 was done using Amira 3.1 (Mercury Computer Systems, TGS Unit, Düsseldorf, Germany). Microtome The microtome was constructed using mostly custom-machined parts made from nonmagnetic stainless steel. The leave springs were made out of bronze. The specimen advance contained a lever mechanism that reduced the motion of the motorized micrometer (M227.10 with controller C862, Physik Instrumente, Karlsruhe, Germany) by a factor of 0.11. This scaled the position uncertainty of the motorized micrometer (50 nm) down to a value 5.5 nm. Retraction of the sample during reverse motion of the diamond knife was driven by a closed-loop piezo element (P841.10, with controller E-610.S0, Physik Instrumente). The cutting motion was driven by a large-displacement piezo element (PX-1500, Piezojena, Jena, Germany). Suspension of the microtome on steel balls sliding on sapphire plates allowed lateral positioning driven by piezo-actuated slip-stick motors (Picomotor 8321-V, New Focus, San Jose, California, United States). The microtome was controlled using a computer interface designed originally for electrophysiology applications (1401 power, CED, Cambridge, United Kingdom) using scripts written in the Spike2 programming environment (CED). Analog voltages that were generated by the computer interface drove the cutting and the retraction piezos. The specimen-advance motor was controlled via serial interface. During automated stack acquisition the QuantaFEG 200 microscope was controlled using a modified keyboard that allowed simulated key pressings via a serial interface. Both serial interface connections were driven by commands in Spike2 software scripts. The diamond knife used was custom made by Diatome, derived from their ultra 35° type. The cutting edge was 1.5 mm long. Unlike in the standard knife, in which the diamond is soldered to a piece of hard metal—increasing the clearance height necessary above the cutting edge and thereby increasing the working distance—our diamond was directly clamped in a custom-made stainless-steel holder. The clearance angle was fixed at 6°. The section thickness values quoted were calculated using the nominal micrometer position change and the mechanical reduction ratio of the advance mechanism. For some of the data ( Figures 5 and S2–S5 ; Datasets S1–S20 ) the knife was oscillated (about 300 nm pp at 12 kHz) along the line of the cutting edge using a modified, piezo-driven knife holder. Supporting Information Dataset S1 Cortical Tissue Slices 1–99 (248.1 MB ZIP). Click here for additional data file. Dataset S2 Cortical Tissue Slices 100–199 (252.6 MB ZIP). Click here for additional data file. Dataset S3 Cortical Tissue Slices 200–299 (252.7 MB ZIP). Click here for additional data file. Dataset S4 Cortical Tissue Slices 300–399 (252.6 MB ZIP). Click here for additional data file. Dataset S5 Cortical Tissue Slices 400–499 (251.9 MB ZIP). Click here for additional data file. Dataset S6 Cortical Tissue Slices 500–599 (252.2 MB ZIP). Click here for additional data file. Dataset S7 Cortical Tissue Slices 600–699 (253.7 MB ZIP). Click here for additional data file. Dataset S8 Cortical Tissue Slices 700–799 (255.9 MB ZIP). Click here for additional data file. Dataset S9 Cortical Tissue Slices 800–899 (256.1 MB ZIP). Click here for additional data file. Dataset S10 Cortical Tissue Slices 900–999 (253.8 MB ZIP). Click here for additional data file. Dataset S11 Cortical Tissue Slices 1,000–1,099 (252.6 MB ZIP). Click here for additional data file. Dataset S12 Cortical Tissue Slices 1,100–1,199 (252.6 MB ZIP). Click here for additional data file. Dataset S13 Cortical Tissue Slices 1,200–1,299 (251.9 MB ZIP). Click here for additional data file. Dataset S14 Cortical Tissue Slices 1,300–1,399 (251.8 MB ZIP). Click here for additional data file. Dataset S15 Cortical Tissue Slices 1,400–1,499 (250.7 MB ZIP). Click here for additional data file. Dataset S16 Cortical Tissue Slices 1,500–1,599 (251.4 MB ZIP). Click here for additional data file. Dataset S17 Cortical Tissue Slices 1,600–1,699 (252.7 MB ZIP). Click here for additional data file. Dataset S18 Cortical Tissue Slices 1,700–1,799 (250.5 MB ZIP). Click here for additional data file. Dataset S19 Cortical Tissue Slices 1,800–1,899 (253.4 MB ZIP). Click here for additional data file. Dataset S20 Cortical Tissue Slices 1,900–2,000 (254.9 MB ZIP). Click here for additional data file. Figure S1 Muscle Tissue Complete field of view for dataset underlying Figure 1A and 1B . Grayscale is inverted from data taken; no smoothing or contrast enhancement was applied. (10.8 MB TIF). Click here for additional data file. Figure S2 Large Volume of Cortical Tissue Bottom slice of a stack of 2,000 images taken at 4 keV. Slice thickness was 55 nm. Spotsize was 3.4. Pixel size is 26.7 nm. The pixels correspond to the original data. The area shown corresponds to 54.8 × 47.3 μm. The raw data can be found as numbered TIF images in Datasets S1–S20 . (10.6 MB TIF). Click here for additional data file. Figure S3 Large Volume of Cortical Tissue: X-Resliced Stack Same volume data as used for Figure S2 . The stack was resliced along the horizontal dotted line shown in Figure S2 . In the vertical direction the data were interpolated so that each slice now corresponds to slightly more that two pixels. Horizontal white lines are slices with deposited debris. The total stack height was 110 μm. (8.2 MB TIF). Click here for additional data file. Figure S4 Large Volume of Cortical Tissue: Y-Resliced Stack Same as Figure S3 but now resliced along the vertical dotted line in Figure S2 . (7.1 MB TIF). Click here for additional data file. Figure S5 Neurite Reconstruction Manual reconstruction of selected processes in cortical tissue (data from Video 2 and Figure 5 ). Blue, portion of proximal apical dendrite; green, secondary dendrite with three synaptically connected axons (yellow, ocher, and mauve). Insets show the synaptic contacts. Also shown is a passing axon that is not synaptically connected within the volume analyzed. Only the lower part of the stack, which was taken at 4 keV electron energy, was used. (17.5 MB TIF). Click here for additional data file.	/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC524270.xml
544547	A Golden Age of Brain Exploration	The Allen Brain Atlas of gene expression in the mouse brain is poised to serve as an outstanding resource to neuroscience	Armed with billions of cells, elaborate circuitry, and a seemingly animate anatomy, capable of growing as it learns, the brain is a marvelously enigmatic organ. Much to the chagrin of those that study it, the brain remains perhaps too mysterious. Although genetic information exploded out of the Human Genome Project, it has been of little consequence to neuroscience—a discipline still grappling with the boundaries and names for distinct brain regions. According to United States National Institute of Mental Health Director Thomas Insel, over 99% of the neuroscience literature focuses on only 1% of the estimated 15,000–16,000 genes expressed in the brain. David Van Essen, a neurobiologist at Washington University in St. Louis, Missouri, likens the current genetic map of the brain to a 17th century map of Earth. A voyage around the Earth had already proven it was round, but landmass resolution was still vague at best. Magellan's benefactors, though, never bankrolled a technical advance quite like the Allen Brain Atlas. Neuroscience's unlikely sugar daddy, Microsoft cofounder and the world's fifth wealthiest man, Paul Allen, created the $100 million dollar Allen Brain Institute in Seattle, Washington, two years ago. The first explicit goal of the institute was to create an open-access, visual, searchable online map of genes expressed in the brain, as well as of brain circuitry and cell location. Roughly one petabyte of data—equal to the memory necessary to hold the information held in about 50 Libraries of Congress—will be produced as a result. In mid-December, the first 2,000 genes were uploaded. By 2006, the Allen team plans to have as many as 24,000 genes online. While the ultimate goal is to map the human brain, the atlas ushers in a new era of neurogenetics—an attempt to make connections between anatomical, genetic, and behavioral observations. Of Mice and Men The initial effort will focus on the standard, inbred lab mouse strain known affectionately as C57BL6. Like it or not, mice are remarkably similar to humans—sharing 99% of our genes. Humans, at this time, provide too many hurdles—not the least of which is a lack of willing brain donors that are the same age. Since the C57BL6 strain is inbred, the mice are also much more uniform than humans—a key to constructing the most accurate representative map possible of one species' adult brain. In situ hybridization—A cross-section through the mouse brain shows gene expression (black dots) in specific brain regions (Copyright: David Anderson) With a map of mouse genes in hand, scientists will be able to develop informed hypotheses about genes that may affect human brain function and dysfunction. “People will be able to look first in the mouse atlas, then more selectively focus on human cases,” says Gregor Eichele, director of the Max Planck Institute for Experimental Endocrinology in Hannover, Germany. Indeed, researchers may do well to focus their efforts on those specific cells in which homologous genes are expressed in the mouse. Insel suggests an initial mental health application: find dyspyndin , a gene linked to schizophrenia. Insel is banking on the atlas to locate genes linked to conditions including bipolar disorder, schizophrenia, and autism. Once identified, such critical genes can be examined in detail and used in studies aimed at disease cures or drug screens. Presumably, the atlas will be a boon for drug discovery and development by providing information on drug targets present within brain cells. Eichele points out that the Allen Brain Atlas will eliminate time wasted testing fruitless hypotheses. “If you think gene X may be involved in [hippocampal-dependent] memory, but it's not expressed in the hippocampus, you shouldn't bother following that line of research,” he says. Insel agrees that where genes are expressed in the brain will be most telling. “In the brain, more than any other organ, function follows form,” he says. Cellular resolution of expression patterns will prove necessary to uncover as yet unknown relationships between circuitry, cell type, and gene expression in the brain, says Arthur Toga, a neuroscientist at the University of California, Los Angeles, and Allen Brain Atlas advisor. Ed Lein, a neuroscientist at the Allen Brain Institute, thinks that mapping at the cellular scale will also redefine anatomy. Traditionally, neuroanatomists have delineated brain regions pretty much by eye, identifying clusters of cells and patterns of connections that look the same. “We're starting to redefine boundaries of regions by cell type,” says Lein, defining cell type by gene expression pattern. “Robot” used for high-throughput in situ hybridization, developed by Gregor Eichele and colleagues at the Max Planck Institute for Experimental Endocrinology (Photo: Gregor Eichele) The difficulty, according to Allen Brain Institute scientific advisor David Anderson, a neuroscientist at the California Institute of Technology in Pasadena, California, is in understanding how a gene mutation affects behavior. “It is impossible without a knowledge of the circuits in which a gene is expressed,” he says. Anderson himself studies innate behaviors such as fear responses. Specifically, he's most interested in understanding the function of different regions of the almond-shaped amygdala. Once he finds genes expressed in neurons within regions of the amgydala implicated in fear, he plans to determine neuron function by creating transgenic animals in which specific neuron activity has been silenced. Thus far, Anderson's laboratory has endured the slog of using microarray techniques to identify genes with expression patterns linked to fear behavior, followed by in situ hybridization, which traditionally involves twenty-odd complex, error-prone steps, to find just some of the genes expressed in different parts of the amygdala. “The Allen Brain Atlas will identify a whole set of genes that we would have had to spend several years to find,” he says. Indeed, locating where even one gene of interest is expressed in the brain eats up valuable research time, especially when there are so many potentially interesting genes. In most cells, 10,000 genes can be expressed. Eichele improved the slow error-prone process into an automated, fast, “high throughput” method that caught the attention of the Allen team because it was capable of meeting the needs of such an ambitious project. In situ hybridization uses labeled probes for specific messenger RNA sequences, allowing scientists to test individual brain tissue samples for gene expression. Automated in situ hybridization data will be generated for the entire mouse transcriptome—the full complement of activated genes in a particular tissue at a particular time—on a genome-wide scale. The mouse is a young adult at 56 days old, free from the confounding factors of development. After it is sacrificed, the mouse brain is immediately frozen, then sliced very thinly—to get forty sections from each millimeter of thickness—so that the probes for hybridization can expose gene expression in individual cells. Mining the Mind for Riches The in situ data will be matched in a three-dimensional framework to the reference atlas developed by Allen's team. The resulting images will be turned into a virtual microscope, allowing users to focus down on genes expressed in regions of interest. While the Allen Brain Atlas is somewhat like other genomics projects in scale, it is unique. “No one's gone into a 3-D structure like a tissue and examined it in a systematic way,” says Allan Jones, senior director of Allen Brain Atlas Operations. In that way, he adds, it's a much richer dataset than the Human Genome Project. “As we're ramping up—fully by spring of next year—we'll be generating about 1,000 microscope slides a day with four mouse brain sections on each slide,” says Jones. Each day, those sections are scanned and stitched together electronically into 300-megabyte batches. “Scaling up a lab process is currently the biggest challenge,” says Jones. “In effect, we're turning an art form into something that gives high-quality data day in and day out,” says Jones. “If you are off slightly when cutting a 2-D brain slice, it becomes very difficult to map back into a 3-D context.” While it is a logical starting point, a spatial understanding of gene expression is just one way to mine the brain. Other avenues currently being pursued explore individual variability, development, and comparisons between species (See Box 1 ). The sheer comprehensive nature of the Allen Brain Atlas will be its crowing achievement, and its complement to the other ongoing atlas efforts. However, the magnitude of the project also poses its greatest hurdle—about which onlookers have expressed some concern. Box 1. A Bevy of Brain Databases The Allen database will provide a spatial map of neurogenetic data, specifying where the 99% of shared mammalian genes are expressed in the brain. Other online databases are striving to provide alternative axes of information that will detail individual variability, species comparisons, and changes during development. Going deep on the genetic variability axis of understanding is Web QTL, a collection of images from 800 brains of 35 different strains of mice. “We're interested in genetic sources of variation,” says Rob Williams, a neurobiologist at the University of Tennessee in Memphis. “We study many strains of mice and map the upstream modulators that control expression differences.” Van Essen's online atlas strives to map the structural and functional areas of the cerebral cortex, believed to be the seat of thought, learning, emotion, sensation, and movement, for humans, macaques, rats, and mice. In constructing the SuMS (for “surface management system”) database, they've put most of their effort into comparing datasets between species. Finally, GenSAT (Gene Expression Nervous System Atlas) follows gene expression as it changes through the development of an organism. Using a method that manipulates “bacterial artificial chromosomes” to insert, change, or delete parts of large gene sequences, one transgenic mouse is created for each gene. When reporter genes are added to the bacterial artificial chromosomes, cells with selective gene activity glow. This advance takes a step towards relating gene expression patterns to connectivity between brain regions. “It's clear that they want to do a first-rate job of working with all this information, but I'm not sure they entirely appreciate how hard it's going to be to manage the staggering amount of information they're getting,” says Van Essen. As manager of his own comparative anatomy database, he understands the magnitude of the task. He points out that properly digitizing the data in electronic form and registering one particular slice to a standardized reference with meaningful coordinates is not trivial. “In an atlas, there will be considerable variability from one individual to the next—even in inbred mouse strains,” says Van Essen. Even if it's only 20% variability, it is still going to pose challenges for managing experimental data. Van Essen acknowledges that this isn't just an impediment for Allen's team, but for neuroscience as a whole. “These tools, in general, don't emerge from a vacuum. They emerge best when rich, challenging datasets are staring people in the face,” he says. An even larger problem is communicating the data effectively. Figuring out how to navigate the tremendous morass of data will be a bioinformatic stumbling block. Lein acknowledges that figuring out how to best annotate the data is one of the bigger challenges for the future—particularly in cases where gene expression doesn't match the agreed-upon boundaries of anatomical regions. For now, users will be able to mine the data by gene name only. The initial release consists of an image viewer to view the in situ hybridization data for one or several genes at a time, along with a reference atlas to determine the structures in which genes are expressed. Future releases will allow the user to conduct more sophisticated searches, such as by anatomical structures. “Not only is there this 3-D structure, but there are lots of studies where people are trying to understand what drives the turning on and off of genes,” says Jones. “At the end, if the atlas has a big impact, it will be in providing the precise coordinates for those people to tease apart what specific DNA elements drive expression within regions or structures.” And while the Allen Brain Atlas will provide a fine level of detail, there are limitations. Much like the Human Genome Project, the information will be the starting point and not the end point of understanding brain function. “It won't change strategy for doing experiments,” says Nobel laureate and Columbia University neuroscientist Eric Kandel. “The atlas will be a catalyst rather than a direction setter.” The Final Frontier Overall, neuroscience is entering a new era. Insel notes that recent work has proven the brain to be extraordinarily dynamic, birthing neurons throughout a lifespan. Brain functions seem more modular than global. And there is no real separation between the mind and brain. “Mental disorders are brain disorders,” he says. Over the next 5–10 years, neurogenomics will fuel a golden age of discovery in neuroscience. In fact, scientists may even reach an overarching goal—understanding how the wild card of environment impacts brain function. “We want predictive genetics able to accommodate environmental differences,” says Rob Williams, neurobiologist at the University of Tennessee in Memphis. For Kandel, one thing is certain. “Most of the mysteries of the brain lie ahead of us.” Using WebCaret and SumsDB to visualize functional magnetic resonance imaging activations and visual areas on a flat map of the human “Colin” cortical atlas (Image: David Van Essen; dataset with publication sources available: http://sumsdb.wustl.edu:8081/sums/directory.do?dir_id=702541 )	/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC544547.xml
544584	DNA vaccines: designing strategies against parasitic infections	The complexity of parasitic infections requires novel approaches to vaccine design. The versatility of DNA vaccination provides new perspectives. This review discusses the use of prime-boost immunizations, genetic adjuvants, multivalent vaccines and codon optimization for optimal DNA vaccine design against parasites.	Introduction DNA vaccination was introduced in 1990 by a study that demonstrated the induction of protein expression upon direct intramuscular injection of plasmid DNA in myocytes [ 1 ]. DNA vaccines are new types of sub-unit vaccines allowing protein expression in mammalian cells after introduction of plasmid or recombinant viral vectors encoding the selected protective antigen. Protective immunity conferred by DNA vaccines has been shown in many animal models of various diseases including HIV, tuberculosis and cancer [ 2 - 4 ]. DNA vaccines induce strong humoral and cellular immunity and have the potential to increase immunogenicity through modifications of the vector or incorporation of adjuvant-like cytokine genes. Successful vaccines should be able to induce strong immune responses which are long-lasting and in most cases providing protection against different strains of the same pathogen. Progress has been made towards development of DNA vaccines against viral and bacterial pathogens showing protection and lasting immunity [ 5 ]. Application of this new vaccination technology with regard to parasitic infection provides new hope for significant advances in anti-parasitic vaccine research. An important consideration in developing vaccines against parasites is the complexity of parasitic diseases. Parasite infections, unlike most viral and bacterial infections, tend to be chronic and associated with immunodepression or inappropriate immune responses [ 6 ]. Parasites have complex life cycles and host immunity to stage-specific antigens may not overlap with other later stages or vector-borne stages. Antigenic variation and other immune evasion mechanisms also complicate the development of vaccines against parasites. However, with recombinant DNA technology and the versatility of DNA vaccination, it is now possible to take rational parasite specific strategies to vaccine design and overcome the obstacles presented by parasitic diseases. Improving DNA vaccine efficacy against parasitic disease can be achieved by: prime-boost immunizations, genetic adjuvants, multivalent vaccines or codon optimization. This review describes the application of these strategies, using specific parasites as examples, to improve DNA vaccine efficacy (see Table 1 [ 7 - 19 ]). Table 1 Summary of DNA vaccine optimization in parasites Optimization Method Parasite Specific Modifications and Improved Responses Reference Genetic Adjuvant Malaria Co-immunization of merozoite surface protein-1 (MSP1) of P. yoelii with IL-12 in A/J mice elicited strong Th1 type responses characterized by high levels of IFN-γ. Parasite specific antibodies also protected against parasite infection. [7] Construction of DNA plasmid encoding C-terminal region of MSP1 ( P. falciparum ) was tested with plamids expressing GM-CSF or recombinant GM-CSF protein in monkeys. Co-immunization with GM-CSF protein lead to higher Ab titers and higher response to boosting with MSP1. [8] MuStDO5 is a multivalent vaccine composed of 5 plasmids encoding P. falciparum proteins and GM-CSF. When tested for safety in mice and rabbits via i.m/i.d. injections, the vaccine was determined safe and well tolerated without development of autoimmunity. [9] Leishmania Vaccination with plasmids encoding L. amazonensis P4 nuclease, HSp70 or murine IL-12 was tested in the susceptible Balb/c mouse model. Co-immunization with P4 nuclease and IL-12 protected mice against parasite challenge as determined by 4 log reduction in parasite burden and increased levels of IFN-γ and TNF-α. [10] Following p36/LACK prime-boost immunization with a combination of DNA vectors expressing IL-12 and IL-18 in mice, highest protection was observed compared to controls. [11] Schistosoma Co-administration of DNA plasmids encoding IL-18 and S. mansoni glutathione S-transferase elicited 30 fold increase in antigen specific IFN-γ secreting cells, 28% reduction in egg laying and 23% reduction in worm burden in mice. [12] Multivalent vaccine Malaria Prime boost regimen with vectors encoding functional domains of TRAP and CS antigens of P. cynomogli was more effective at reducing peak parasitemia in rhesus monkeys. [13] A multistage P. knowlesi vaccine with plasmids encoding 2 pre-erythrocytic, 2 blood stage antigens and GM-CSF was administered to rhesus monkeys followed by a boost with a pox virus encoding all 4 antigens. Monkeys developped Abs against sporozoites, infected erythrocytes and CPS protein. [14] Six pre-erythrocytic antigens linked together to produce a polyprotein in a DNA vaccine and either MVA or FP9 were tested in mice against P. falciparum . Greater responses were seen when a heterologous viral regimen was used, producing multispecific T cells. [15] Leishmania L. major TSA and LmST11 antigens were expressed either as single genes or as digene construct and tested in the susceptible Balb/c model. Administration of the genes in either constructs lead to protection via polyspecific immune responses. [16] Schistosoma Three doses of 4 plasmids encoding S. japonicum antigens, Sj62, Sj28, Sj23 and Sj14 3-3-, induced high levels of IFN-γ and partial protection from challenge infection when administered in mice. [17] Entamoeba DNA plasmids encoding either Entamoeba histolytica cysteine protease 112 or adhesin 112 were co-administered to hamsters, leading to protection against liver abscess formation. No protection was observed with either plasmid alone. [18] Codon optimization Malaria P. falciparum erythrocyte binding protein and MSP1 antigens were codon optimized for expression in mammals. 10 to 100 fold less optimized plasmid DNA was required to induce high Ab titers in mice. [19] Prime-Boost Immunizations Current sub-unit vaccines predominantly induce strong antibody responses and weak cellular immunity. DNA vaccines in animal models can induce both strong humoral and cellular mediated responses, but although safe in humans, DNA vaccines do not produce the same magnitude of cellular immunity [ 20 ]. In cases where the pathogen is intracellular, an antibody response is not sufficient for protection and cell-mediated immunity is required. This is the case with malaria, where the parasite infects hepatocytes and erythrocytes, and cytotoxic T cells play an important role in protection. Therefore, it is important to devise vaccination strategies that enhance T cell immunogenicity and confer a protective cellular immune response to intracellular pathogens. A novel approach to increase T cell responses to vaccination is the heterologous prime-boost immunization strategy [ 21 ]. This method consists of priming and boosting with different vectors encoding the same antigen. The principle of the strategy is to first prime some T cells to be antigen-specific and then boost to induce rapid T cell expansion upon repeated exposure to the specific antigen. DNA plasmids are good priming agents since they are internalized by antigen presenting cells and can induce antigen presentation via MHC class I or class II. DNA plasmid backbones are immunogenic due to the presence of stimulatory unmethylated CpG motifs that readily induce Th1 cytokine expression, leading to cellular mediated immunity. Recombinant viral vectors, which are non-replicating and safe, are excellent for boosting. Viral vectors induce high protein expression and presentation via MHC class I which leads to greater antigen specific T cell expansion [ 22 ]. Common boosting vectors in vaccine trials include modified Vaccinia virus Ankara (MVA), recombinant Vaccinia virus (rVv), attenuated adenoviruses, and attenuated pox viruses like fowl pox (FP9). These viruses are highly attenuated and non-replicating but still able to produce proteins. The MVA vector, for example, was developed by over 500 serial passages in chicken embryo fibroblasts and has acquired a replication defect in late stage virion assembly. This vector was used for smallpox vaccinations in 1970 and is known to be safe as well as highly immunogenic. Viral vectors induce strong production of proinflammatory cytokines, which generate greater levels of cell-mediated immunity. Overall the immunogenicity of viruses is greater than that of plasmid DNA, however when administered alone the immune response is generally targeted to vector components. For this reason heterologous vaccination, priming and boosting with different vectors, promotes antigen-specific responses rather than vector-specific responses. The resulting effect when using the heterologous prime-boost technique is the generation of memory T cells to the antigen by priming then amplification of these cells by boosting. This approach has been used extensively to create effective immunizations against malaria, and in a variety of parasites [ 23 - 32 ] (see Table 2 ). Table 2 Prime-boost immunization trials against parasites Parasite Antigen Priming agent Boosting agent Response Reference Malaria Circumsporozoite protein of P. berghei Attenuated fowlpox virus or DNA MVA Potent CD8+ T cell responses were elicited in mice with FPV/MVA vaccination. Novel regimen was more protective against challenge than DNA-MVA immunizations. [7] P. falciparum surface protein (Pfs25) DNA Recombinant protein Intramuscular injections in rhesus monkeys showed significant increase in transmission blocking antibodies. [8] Circumsporozoite protein of P. yoelii DNA Pox virus Immunized neonatal mice showed 93% protection which was CD8+ T cell dependent. [9] P. falciparum erythrocyte binding protein DNA Recombinant protein Higher antibody titers and the ability to reduce parasitemia without drug intervention in Aotus monkeys. [10] Circumsporozoite protein of P. falciparum DNA RTS, S/ASOZA Malaria volunteers develop P. falciparum specific Abs and Th1 specific CD4+ and CD8+ T cells upon vaccination. [11] Leishmania Leishmania infantum LACK DNA Recombinant vaccinia virus 60% protection, associated with cell mediated responses, was observed in dogs after challenge compared to controls. [12] p36/LACK DNA Recombinant vaccinia virus Vaccination in mice resulted in 70% reduction in lesion size and 1000-fold reduction in parasite loads. [13] L. infantum acidic ribosomal protein PO (LiPO) DNA Recombinant protein Boosting elicited stronger IgG2a titers but could not protect against challenge compared to DNA alone. [14] Schistosome Cu/Zn cytosolic superoxide dismutase (SOD), signal peptide SOD and glutathione peroxidase (GP) DNA MVA DNA vaccines were tested against S. masoni challenge in mice. Boosting with MVA for the same genes had no increased effect expect for mutated GP antigen were boosting lead to 85 % protection. [15] To further improve the efficacy of a Plasmodium yoelii DNA vaccine, mice were primed intramuscularly with DNA vaccine and granulocyte/macrophage colony stimulating factor (GM-CSF) plasmid and boosted with rVv encoding the same circumsporozoite protein (CSP) [ 33 ]. This combined strategy of genetic adjuvant and prime-boost immunization elicited improved responses and protection while also reducing the dose of initial DNA vaccine required. In chimpanzees, a DNA-prime and MVA-boost regimen encoding thrombosin-related adhesion protein (TRAP) with GM-CSF protein as adjuvant induced specific T cell and antibody response that was long lasting against P. falciparum [ 34 ]. Complete protection against P. berghei challenge characterized by strong CD8+ T cell responses was observed in mice after intradermal adenovirus-prime-MVA-boost encoding CSP [ 35 ]. These studies led to the assessment of prime-boost immunizations in humans in both naive volunteers and field trials in endemic areas. DNA-prime-MVA-boost vaccines encoding a polyepitope string fused to P. falciparum pre-erythrocytic TRAP antigen were administered via gene-gun to healthy volunteers with no adverse effects [ 36 ]. The polyepitope in the vaccine encodes a single polypeptide, which constitutes of a string of T and B cell epitopes from different sources, including tetanus toxin and BCG. In fact, this heterologous prime-boost immunization elicited interferon-γ (IFN-γ) secreting, antigen-specific T cells in humans, which were significantly higher than responses observed with either vector alone [ 37 ]. Furthermore, this study demonstrated partial protection, measured by delayed parasitemia, after challenge with a different strain of P. falciparum . Another group demonstrated that priming with DNA vaccine for P. falciparum CSP and boosting with a recombinant protein vaccine in adjuvant (RTS, S/AS02A) induced the production of significant antibody and T cell responses in healthy volunteers [ 38 ]. Phase I clinical trials in The Gambia in semi-immune adults have demonstrated that heterologous DNA-prime-MVA-boost regimen encoding P. falciparum TRAP antigen is safe, well tolerated and induces responses greater than those observed in naive volunteers [ 39 ]. Boosting with the MVA vaccine 12 months after the initial prime-boost immunization in this clinical trial was successful in re-expanding the T cell population and demonstrated the safe use of MVA to boost at different periods to maintain T cell immunity. Genetic Adjuvants Adjuvants are used to strengthen the immune response to a vaccine and have been critical in modern vaccine development. Genetic adjuvants are expression vectors encoding biologically active molecules such as cytokines, chemokines and co-stimulatory molecules. These adjuvants can be encoded on the same vector as the antigen or expressed on a separate vector and co-injected with the vaccine. This method provides adjuvant activity at the site of antigen production, with lasting effect from transfected cells. Cytokines are chosen as genetic adjuvants because they regulate cells involved in host defense and can be used to modulate immune responses. Co-delivery of cytokines in DNA vaccine formulation has been used extensively for a wide range of infectious and parasitic diseases (see Table 2 ) to enhance the T cell subset responses known to be protective. Vaccine development against schistosomiasis has been hindered by a lack of consensus on the type of immune response that would be protective. However, it is generally believed that the best strategy for an anti-pathology vaccine is immune deviation. Pathology in schistosomiasis is associated with egg-induced granuloma formation for which there is evidence for a role for Th2 cytokines. The strategy here is to use genetic adjuvants of the Th1 cytokine subset, like interleukin-12 (IL-12), to skew the immune response and provide protection [ 40 ]. Therefore immune deviation is attained with the use of selected genetic adjuvants. Siddiqui et al . [ 41 ] generated DNA vaccines encoding Schistosoma mansoni large subunit of calpain (Sm-p80) and either mouse GM-CSF or IL-4 to determine their adjuvant effect in mice. GM-CSF may work as adjuvant through its activating effect on dendritic cells and macrophages. Intramuscular vaccination with Sm-p80 alone provided 39% protection and this protection was significantly increased to 44% with GM-CSF co-administration and 42% with IL-4. The addition of GM-CSF led to an increase in total IgG and IgG1 while Th1 type IgG2a antibody titers remained high in protected animals [ 42 ]. Since protection was associated with Th1 type antibodies, the Sm-p80 DNA vaccine was further enhanced with co-delivery of plasmids encoding mouse IL-2 or IL-12 [ 43 ]. Greater protection was observed with IL-2 and modest but significantly higher protection was provided by IL-12 co-delivery. Both IL-2 and IL-12 are key cytokines in Th1 cell differentiation. The co-delivery of these cytokines increased IgG2a antibody levels and decreased IgG1 levels, indicating that these genetic adjuvants were successful as Th1 enhancers. Other studies reported no enhancement of protection or immune responses when IL-12 was co-injected, but these differences may be attributed to the nature of the vaccine antigen [ 44 ]. Multivalent Vaccines Another advantage of DNA vaccines is the possibility to integrate several antigens into the plasmid or to administer a mixture of plasmid vectors. The development of multivalent vaccines consisting of several antigens is a novel approach to create broad range protection against different parasite strains and parasite life cycle stages (see Table 2 ). Parasites are complex organisms with multiple life cycle stages and antigenic variation mechanisms to evade immune system recognition. Furthermore, not all individuals respond to the same antigens in natural infections. Multivalent vaccines have a greater amount of protective epitopes and could be effective in a greater proportion of the population. However, in multivalent vaccines, the optimal association or combination of antigens must be assessed to obtain synergistic effects. Vaccination studies against leishmaniasis in mice have identified various parasite antigens with varying degrees of protection as protein vaccines. When combined into multivalent DNA vaccines these antigens have the ability to confer complete or enhanced protection. In fact, a DNA vaccine including a mixture of plasmids encoding three antigens, Leishmania major -activated C kinase (LACK), thio-specific antioxidant (TSA), and L. major stress-inducible protein (LmST11) was able to induce complete and long lasting protection after parasite challenge in mice compared to killed Leishmania parasites and rIL-12 [ 45 ]. This protection was characterized by reduced parasite load and the recruitment of CD8+ and CD4+ T cells to the site of infection. The same group tested the combination of these antigens and the route of administration to optimize the results of the previous study [ 46 ]. It was determined that a cocktail vaccine composed of all three antigens was more effective than LACK alone or LmSt11 and TSA combined. Furthermore, intradermal injection of the plasmid mixture was more effective than intramuscular or subcutaneous injections, reducing the dose of vaccine required five-fold. Another study also demonstrated that prime-boost co-injection of plasmids encoding two different L. major cysteine proteinase genes (Cpa/Cpb) was protective and characterized by IFN-γ production by spleen cells, while separate injections were not protective [ 47 ]. The cysteine proteinases are expressed at different levels during parasite development and are thought to be involved in modulation of the host response for parasite survival. In this study the cysteine proteinases, only when combined, had the capacity to induce long lasting immunity of the Th1 type. Comparative evaluations of potential protective antigens is necessary to determine optimal DNA vaccine design [ 48 ] as the nature of antigens can have important effects on vaccine efficacy. Codon Optimization Interspecies differences in codon usage are a major obstacle in DNA vaccine development. This is due to the fact that DNA vaccines use host cells for transcription and translation of proteins. Every species has a codon bias for which most genes are encoded and this use of selected codons is related to gene expression efficiency. Closely related species use similar codons. However, in cases where there is a great difference in codon usage between the pathogen and mammals, codon optimization may be required. This strategy involves the modification of codon usage for the genes encoded in a DNA vaccine to a suitable codon bias for increased expression in mammals. This method has proved effective in many systems [ 19 , 49 , 50 ], increasing protein expression in vitro and antigen specific responses in vaccinated animals. In our laboratory, we have developed a codon-optimized DNA vaccine encoding a portion of the Entamoeba histolytica Gal-lectin [ 51 ]. E. histolytica genes are rich in A:T codons, whereas mammalian codons are more G:C rich. Protein expression of the E. histolytica Gal-lectin protein using the wild type sequence was difficult and stable clones were difficult to obtain in mammalian cells. Codon optimization was performed to ultimately increase protein expression in gerbils, a model for experimental amoebiasis; therefore gerbil codon usage was used to re-write the Gal-lectin Hg1l gene. Transfection of Cos-7 cells with the optimized vaccine construct produced a protein which was immunoreactive with a Gal-lectin specific monoclonal antibody (3F4), demonstrating successful expression of this amoebic protein. Upon vaccination with this codon optimized DNA plasmid, mice developed antigen specific antibodies of the Th1 isotype and Gal-lectin specific cellular immune responses. Conclusions In this review, strategies for increased DNA vaccine efficacy against parasitic diseases to date, i.e. prime-boost immunizations, genetic adjuvants, multivalent vaccines and codon optimization, have been discussed. DNA vaccine technology provides the versatility required to separate protective components of immunity from counter-protective responses. As seen with genetic adjuvants, DNA vaccines can focus on the protective cytokines involved and include antigens that stimulate the production of specific cytokines. This allows designing vaccination strategies that are tailored to a particular infection or even a specific stage of infection. Parasitic diseases are complex, involving changes in immunological responses during the course of infection and changes in immunity to stage specific antigens. The advent of optimization strategies with DNA vaccines presents researchers with the tools to design effective vaccines with specific purposes. It is possible to enhance DNA vaccine efficacy, thus increasing immune responses and protection, through the use of these methodologies. However, it is important to note that these strategies need to be adjusted to the parasite system in order to provide the greatest benefit upon vaccination. For example, Sedegah et al . [ 52 ] reported reduced immunogenicity of multistage P. falciparum DNA vaccines when administered as a mixture of plasmids compared to single plasmid injections. Another study, however, demonstrated that a mixture of three plasmids encoding P. falciparum blood-stage antigens had no reduction in immunogenecity when co-injected [ 53 ]. Therefore many aspects of a DNA vaccine can contribute to its efficacy, and each must be evaluate to understand the interactions between vaccine components. In fact, it is clear that other factors are important in vaccine design, such as the nature of the antigen, the presence of immunostimulatory CpG motifs in the plasmid backbone, the vaccine delivery system or the site of injection [ 8 , 9 , 16 , 24 , 32 , 54 ]. The method of vaccine delivery is an important variable in vaccination design. DNA vaccination has been successful through a variety of injection routes, including intradermal, intramuscular, and intranasal. Although intramuscular injections are most common and give consistent responses, alternative routes of delivery may be desired depending on the disease model. Mucosal DNA vaccine immunizations, against intestinal parasites for example, are effective to generate mucosal immune responses at the site of infection. For leishmaniasis, where the disease manifests itself as cutaneous lesions, an intradermal injection targeting Langerhans' cells may be optimal [ 46 ]. The gene-gun is a unique method of DNA vaccine delivery, which has been used successfully against a variety of parasites. The gene-gun accelerates plasmid-coated gold particles to supersonic speed with helium gas and delivers them to the outer layers of the skin. In reality this vaccine delivery system is being tested alongside intramuscular injections in The Gambia field trials for malaria vaccines in humans. The Powderject ® XR1 is a needle-free powder injection system that delivers fine gold particles coated with the vaccine vectors directly into epidermal cells, specifically dendritic cells. This vaccination method is advantageous since it eliminates the cold chain requirement and reduces the chances of needle-borne contamination. Moreover, the gene-gun method is safe and seems as immunogenic as intramuscular injections in these trials [ 36 ]. The greatest challenge in designing DNA vaccines against parasites is making the vaccine suitable for humans while providing strong, long lasting immune responses. Many studies in laboratory animals are successful but the results cannot be replicated in humans. The prime-boost strategy has shown the most success as a delivery technique in larger animals or humans. Field trials with prime-boost malaria vaccines are ongoing and will provide experts with insight with regards to the safety and the immune responses required for protection in humans. Meanwhile, other groups are reporting improved responses in mice or larger mammals with other vaccines, suggesting that this vaccination strategy may be applicable to many other parasitic diseases [ 3 , 29 ]. A variety of combinations of other enhancement strategies with prime-boost immunization have been explored, including the use of genetic adjuvants or multivalent plasmids [ 11 , 14 , 15 ]. Prime-boost immunizations against Leishmania parasites in mice have improved cellular immune responses when plasmids expressing IL-12 and IL-18 are co-injected [ 11 ]. A DNA prime-protein boost vaccine in monkeys encoding two P. cynomolgi antigens (CSP/TRAP) resulted in lower peak parasitemia and higher antibody and cellular responses than controls [ 13 ]. Taken together, the techniques described above will allow parasitologists to develop effective DNA vaccines that are designed to target a specific immune response during parasitic infection. The optimized approach provided by DNA vaccine technology will produce vaccines ready for clinical and practical applications, as well as providing a greater understanding of the underlying complexity of immunity in parasitic infections. Competing interests The author(s) declare that they have no competing interests. Author's contribution CI and KC produced the manuscript together. All authors read and approved the final manuscript.	/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC544584.xml
552325	Umbilical metastases: current viewpoint	Background Umbilical metastases from a malignant neoplasm, also termed Sister Mary Joseph's nodule, are not commonly reported in the English literature, and they have usually been considered as a sign of a poor prognosis for the patient. The present article reports on the current view point on umbilical metastasis besides discussing the epidemiology, clinical presentation, pathophysiology and treatment. Method A search of Pubmed was carried out using the term 'umblic*' and 'metastases' or metastasis' revealed no references. Another search was made using the term "Sister Joseph's nodule" or sister Joseph nodule" that revealed 99 references. Of these there were 14 review articles, however when the search was limited to English language it yielded only 20 articles. Articles selected from these form the basis of this report along with cross references. Results The primary lesions usually arise from gastrointestinal or genitourinary tract malignancies and may be the presenting symptom or sign of a primary tumour in an unknown site. Conclusion A careful evaluation of all umbilical lesions, including an early biopsy if appropriate, is recommended. Recent studies suggest an aggressive surgical approach combined with chemotherapy for such patients may improve survival.	Background Cutaneous metastases localised to the umbilicus are named "Sister Mary Joseph's nodules". In 1949 Sir Hamilton Bailey initially used this eponym in his book "Physical Signs in Clinical Surgery" to describe umbilical metastases, in honour of Sister Mary Joseph, the superintendent nurse and surgical assistant of Dr. William Mayo at St. Mary's Hospital in Rochester (presently the Mayo Clinic in Minnesota, USA). Sister Mary Joseph was the first to note the link between umbilical nodules and intra abdominal malignancy [ 1 - 4 ]. Epidemiology The occurrence of cutaneous metastases from malignant neoplasms occurs in from 1% to up to 9% of individuals, as determined at autopsy. Those metastases to the umbilicus are uncommon and represent only 10% of all secondary tumours which have spread to the skin [ 5 , 6 ]. Epidemiological studies showed that this condition predominates in females [ 7 ]. From a review of the literature, umbilical neoplastic nodules can be due to a primary tumour in 38% of cases, due to endometriosis in 32% of individuals, and in 30% are actually secondary tumour deposits from a primary tumour elsewhere [ 8 ]. If these nodules are secondary tumour deposits then the source of the primary tumour may be from the gastrointestinal (35–65%) and genitourinary (12–35%) tract. In addition, in 3–6% of cases it originates from haematological malignancies, or lung and breast cancers. In 15% to 30% of patients the source of the primary site of the tumour remains unknown [ 9 - 11 ]. Clinical findings Sister Mary Joseph's nodules usually present as a painful lump on the anterior abdominal wall. It has irregular margins and a hard fibrous consistency. The surface may be ulcerated and necrotic, with either blood, serous, purulent, or mucous discharge from it. The size of the nodule usually ranges from 0.5 to 2 cm, although some nodules may reach up to 10 cm in size [ 8 ]. High-resolution ultrasound (US) helps to clarify the clinical findings by detecting solid umbilical nodules, even if the diagnosis is difficult to make clinically. Moreover, careful examination and imaging of the abdominal contents may also point to the diagnosis [ 12 ]. Pathophysiology A full understanding of the mechanisms whereby the tumour spreads to the umbilicus remains unclear. However, following anatomical criteria, and several hypothesis have been proposed. The umbilical ring is a scar invaginated on the abdominal wall between the transversalis fascia and peritoneum. After birth, the foetal cord structures develop into ligaments or peritoneal folds: 1) median umbilical ligament secondary to the obliterating urachus, 2) medial umbilical ligaments (which are obliterated umbilical arteries); 3) ligamentum teres (obliterated left umbilical vein) that continues into 4) the falciform ligament. On the lateral umbilical folds the inferior epigastric vessels and, sometimes, a vestigial vitelline duct connecting the umbilicus to the ileum can be recognised. The umbilical region shows a rich arterial supply that includes the inferior epigastric and deep circumflex iliac branches of the external iliac artery, and the superior epigastric branch of the internal mammary artery. The venous drainage includes several anastomotic branches, coming from cranially the axillary vein, through the internal mammary vein, and caudally, the femoral vein through the superficial epigastric vein. In addition, the umbilicus may be connected with the portal system, through small umbilical veins. The lymphatic system connects the umbilical region to the axillary, inguinal, and para-aortic lymph nodes. The deep lymphatic system passes along the falciform ligament, pierces the diaphragm and enters the anterior mediastinum or courses to the nodes around the iliac arteries [ 12 - 15 ]. All these systems (arterial, venous and lymphatic) as described represent possible routes by which metastatic tumour cells could implant into the umbilical region. It is reasonable to suggest that direct extension of tumour through the peritoneum is the preferred route for gastrointestinal tumours. Furthermore, the common association between hepatic and umbilical metastases might suggests that the hypothesis that the tumour spreads from the primary tumour to the liver through the portal system and then through the lymphatic and/or venous channels, they spread to the umbilicus. It is still unclear if the umbilical tumour spread precedes the hepatic spread or vice versa . Renal cell carcinoma typically spreads via extra-renal extension, lymphatic dissemination or venous invasion by the tumour. Intraperitoneal spread may occur as a result of disruption of the renal capsule [ 16 ]. The dissemination of neoplastic cell through the urachus is assumed to be the mechanism for the bladder cancers. Haematogenous, lymphatic and venous spread all represent valid mechanisms of tumour spread from gynaecological cancers [ 7 , 14 ]. Prognosis and therapy Usually the presence of an umbilical metastasis indicates a poor prognosis, is a sign of advanced neoplastic disease, and may not be amenable to surgery. The survival of these patients without treatment has been reported to range from 2 to 11 months from the time of initial diagnosis [ 17 - 19 ]. However, recent studies have suggested that there are several factors which are able to influence the prognosis of such patients. Certain data has shown a better survival (mean 9.7 months) in patients who detect an umbilical metastasis before definitive treatment of the primary tumour. In contrast, when the lesion appears after the primary tumour has been treated then the survival for these patients does not exceed the 7.6 months [ 16 , 17 ]. Moreover, the aetiology of the primary malignancy determines the prognosis. For example, a better survival rate for patients with primary ovarian carcinoma has been reported previously [ 14 ]. Finally, the type of treatment seems able to influence the patient's prognosis. Despite some authors proposing only palliative treatment because of these patients poor prognosis [ 10 , 17 , 20 ], recent studies have demonstrated that there is a better survival (21 months) for patients if they are treated with a combination of surgery and adjuvant therapy instead of surgery alone (7.4 months) or chemotherapy alone (10.3 months) [ 9 , 11 , 14 , 18 ]. Obviously, the appropriateness of such an aggressive treatment approach is determined by the clinical state of the patient. Conclusion A careful examination of all umbilical lesions is recommended, especially in those patients with gastrointestinal and genitourinary tract malignancies. All umbilical mass lesions should be biopsied to determinate the pathological nature of the lesion. Clinical experience suggests that, whenever it is possible, an aggressive surgical approach combined with chemotherapy treatment may be considered to offer the patient the best survival probability. Competing interests The author(s) declare that they have no competing interests. Authors' contributions RG contributed to conception, design, literature search and preparation of manuscript. MC contributed to conception, design, literature search and preparation of manuscript. FE contributed to conception, design, literature search and preparation of manuscript. MB contributed to conception, design, literature search and preparation of manuscript. All authors read and approved the manuscript	/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC552325.xml
535341	Missing paternal demographics: A novel indicator for identifying high risk population of adverse pregnancy outcomes	Background One of every 6 United Status birth certificates contains no information on fathers. There might be important differences in the pregnancy outcomes between mothers with versus those without partner information. The object of this study was to assess whether and to what extent outcomes in pregnant women who did not have partner information differ from those who had. Methods We carried out a population-based retrospective cohort study based on the registry data in the United States for the period of 1995–1997, which was a matched multiple birth file (only twins were included in the current analysis). We divided the study subjects into three groups according to the availability of partner information: available, partly missing, and totally missing. We compared the distribution of maternal characteristics, maternal morbidity, labor and delivery complications, obstetric interventions, preterm birth, fetal growth restriction, low birth weight, congenital anomalies, fetal death, neonatal death, post-neonatal death, and neonatal morbidity among three study groups. Results There were 304466 twins included in our study. Mothers whose partner's information was partly missing and (especially) totally missing tended to be younger, of black race, unmarried, with less education, smoking cigarette during pregnancy, and with inadequate prenatal care. The rates of preterm birth, fetal growth restriction, low birth weight, Apgar score <7, fetal mortality, neonatal mortality, and post-neonatal mortality were significantly increased in mothers whose partner's information was partly or (especially) totally missing. Conclusions Mothers whose partner's information was partly and (especially) totally missing are at higher risk of adverse pregnant outcomes, and clinicians and public health workers should be alerted to this important social factor.	Background Pregnancy outcomes are not only important measures of health status of the mothers and infants, but are important measures of socioeconomic development in a society. For example, infant mortality has been considered the single most comprehensive measure of the health and wealth in a society [ 1 , 2 ]. Both maternal and paternal characteristics, such as age, race, marital status, education, and cigarette smoking [ 3 - 5 ], are important determinants of pregnancy outcomes. However, partner information is often missing in routine vital statistics data. For example, about 17% of United Status birth certificates contain no paternal information and more than 40% of the babies born to adolescent women have no information on the paternal age in the birth certificate [ 6 ]. There might be important differences in the pregnancy outcomes between mothers with versus those without paternal information. Detailed description of these differences may benefit the clinicians and public health workers who serve high risk pregnant women. We conducted extensive search of the literature and have not been able to locate a single study on this issue. Epidemiologic studies based on a data with missing information on exposure, outcomes, and potential confounders can bias the study results, unless the missing occurred randomly [ 7 ]. Previous studies on missing data [ 8 - 11 ] have focused on the impact of missing information on study results and the methods to treat the missing variables. The main objective of the current study is to describe the distribution of maternal characteristics and pregnancy outcomes in mothers whose partner's information was partly or totally missing. We also attempted to assess the potential source of missing partner information and its impact on outcomes, by exploring the reasons of differences between partly missing versus totally missing. We used a large twin registry data in the United States to examine this issue. This data has been used in several peer-reviewed studies [ 12 - 14 ]. The reasons of using data on twins were two folds. First, adverse pregnancy outcomes are more common in twins than in singletons. Therefore, it is easier to detect the difference in adverse pregnancy outcomes in twins than in singletons. Second, because linkage for multiple births requires an examination and processing of the recorded variables, it will increase the validity of the administrative data. Methods We used the matched multiple birth file created by the National Center for Health Statistics (NCHS), Centers for Disease Control and Prevention [ 15 ]. The multiple birth file in 1995–1997 in the United States were linked infant deaths information according to uniform coding specifications, which have undergone vigorous editing and reviewing during the process of the record linkage. Sets of multiples in the 1995–1997 birth file were matched by plurality, state and county of occurrence of delivery, mother's date of birth, date of last menstrual period (LMP), number of prenatal visits, level of education, weight gain during pregnancy, and date of delivery. The matching was successful for 98% of the multiple sets [ 15 ]. Available study variables in the database include socio-demographic information of the parents, maternal life-style factors such as smoking during pregnancy, obstetric history, complications of the pregnancy, labor and delivery, birth weight, gestational age, and other infant outcome variables. Only complete twin sets were included in the current study, with twin set as the unit of analysis in comparison of maternal characteristics and outcomes, and randomly selected one twin in each twin sets as the unit of analysis for fetal and neonatal outcomes. There were two variables for the partners in the data: partner's age and race. The eligible study subjects were divided into three groups according to the availability of the two variables of the partners: both available (group 1), one missing (group 2), and both missing (group 3). We first described the distribution of the maternal characteristics (age, race, marital status, education, cigarette smoking, place of birth, and prenatal care service) of the three study groups. The prenatal care service was defined as adequate, intermediate, and inadequate according to the time of prenatal care visit initiation and the number of prenatal visit using the method described by Kessner et al [ 16 ]. If the first prenatal care initiated at third trimester, or the times of prenatal visit less than 5 in 34 gestation weeks, 4 in 32–33 weeks, 3 in 30–31 weeks, or 2 in 22–29 weeks, the prenatal care service was defined as inadequate; if the first prenatal care initiated at first trimester, and the times of prenatal visit more than 8 in 36 gestation weeks, 7 in 34–35 weeks, 6 in 32–33 weeks, 5 in 30–31 weeks, 4 in 26–29 weeks, 3 in 22–25 weeks, or 2 in 18–21 weeks, the prenatal care service was defined as adequate; and the remainders were defined as intermediate. We compared the rates of maternal medical complications (anemia, cardiac disease, acute or chronic lung disease, diabetes, genital herpes, hydramnios, oligohydramnios, hemoglobinopathy, chronic hypertension, pregnancy-associated hypertension, eclampsia, incompetent cervix, renal disease, RH sensitization, and uterine bleeding), labor and delivery complications (febrile (any temperature reading >38°C), meconium, premature rupture of membrane (>12 hours), abruption placenta, plancenta previa, seizures during labor, precipitous labor (2 nd stage <3 hours), prolonged labor (2 nd stage >20 hours), dysfunctional labor, breech, malpresentation, cephalopelvic disproportion, cord prolapse, anesthetic complications, and fetal distress), and obstetric interventions (cesarean section, vacuum / forceps, and labor inductions) among the three study groups. Finally we compared the rates of fetal and neonatal mortality and morbidity among the three study groups. We used the Chi-square test to test the difference of all rates among the three study groups, used the normal approximate method to calculate the 95% confidence interval of the rates of maternal medical complications, labor and delivery complications, obstetric interventions, and the fetal and neonatal mortality and morbidity. We calculated the relative risk in group 2 and group 3 compared with group 1 for all the outcomes mentioned above. Fetal death was defined as stillbirth weighing 500 grams or more, or if weight was unknown, 20 completed gestation weeks or more. Neonatal death was defined as live-born infant died at 0–27 days of age, and post-neonatal death was defined as those died at 28–364 days of age. Preterm birth (PTB) was defined as gestational age < 37 weeks. Fetal growth restriction (FGR) was defined as less than 10th percentile of birth weight-for-gestational-age z score. The birth weight-for-gestational-age z score was calculated using the following formula: Z = (observed birth weight - mean of birth weight)/SD Where mean and SD was based on all infants in the database, stratified by gender and gestational week. Low birth weight (LBW) was defined as birth weight <2500 grams. Low Apgar score was defined as five minute Apgar score < 7. Anemia was defined as hemoglobin <13. Other fetal and infant outcomes examined included birth injury, hyaline membrance disease, meconium aspiration syndrome, assisted ventilation (<30 minutes and > = 30 minutes), seizures, central nervous system anomalies, circulatory / respiratory anomalies, digestive system anomalies, urogenital system anomalies, musculoskeletal / integumental anomalies, and chromosomal anomalies. Results There were 304466 twins (152233 twin pairs) in the database. Of which, father's age and race were recorded in 259302 twins (129651 twin pairs, 85.17%, group 1), father's age or race was missing in 6036 twins (3018 twin pairs, 1.98%, group 2), and both father's age and race were missing in 39128 twins (19564 twin pairs, 12.85%, group 3). Table 1 describes the distribution of maternal characteristics and prenatal care service among the three study groups. The proportions of teenage, blacks, unmarried, lower education, cigarette smoking during pregnancy, and inadequate perinatal care service were significantly higher in group 2 and (especially) group 3 as compared with group 1 (Table 1 ). Table 1 The distribution (%) of baseline maternal characteristic among three study groups Partner information available (N = 129651) Partner information partly missing (N = 3018) Partner information totally missing (N = 19564) Age* <20 y 5.1 13.7 21.4 20 y- 75.9 72.2 69.4 35 y- 19.0 14.1 9.2 Race* white 84.0 69.0 46.1 Black 12.1 26.3 51.3 Other 3.9 4.7 2.6 Marital status* Married 82.6 39.0 4.1 Unmarried 17.4 61.0 95.9 Education* <12 y 13.1 28.4 36.1 12 y 30.0 31.2 39.9 13–15 y 24.1 15.2 15.9 16 y 19.6 5.8 2.9 >16 y 13.2 19.4 5.2 Cigarette smoking* Yes 8.1 14.9 19.1 No 70.5 55.8 69.5 Not available# 21.3 29.4 11.4 Prenatal care service* Adequate 79.4 55.5 54.1 Mediate 17.8 37.0 34.6 Inadequate 2.7 7.5 11.3 Place of birth* Native 84.5 72.7 88.1 Foreign 15.5 27.3 11.9 Compared group 2 and group 3 with group 1 with Chi-square test, p < 0.001 # The state of California did not send data on smoking, and all births to residents in California were classified as not available. Table 2 compares the rate of maternal morbidity, labor and delivery complications, and obstetric interventions among the three study groups. The rates of anemia, acute or chronic lung disease, hemoglobinopathy, chronic hypertension, eclampsia, renal disease, meconium, premature rupture of membrane, abruption placenta, and precipitous labor were significantly higher, and the rates of cardiac disease, diabetes, RH sensitization, placenta previa, dysfunctional labor, cephalopelvic disproportion, cesarean section, vacuum / forceps delivery, and induction were significantly lower in group 2 and (especially) group 3 than in group 1 (Table 2 ). Table 2 Comparison of maternal medical complication, labor complication, and obstetric intervention among three study groups (%, 95%CI) Partner information available (N = 129651) Partner information partly missing (N = 3018) Partner information totally missing (N = 19564) Medical risk factors Anemia 2.81 (2.72 – 2.90) 4.27 (3.55 – 5.00) 4.67 (4.37 – 4.96)* Cardiac disease 0.64 (0.60 – 0.69) 0.40 (0.17 – 0.62) 0.44 (0.35 – 0.53)* Acute or chronic lung disease 0.90 (0.85 – 0.95) 1.36 (0.95 – 1.77)* 1.64 (1.46 – 1.82)* Diabetes 3.46 (3.36 – 3.56) 3.02 (2.40 – 3.63) 2.15 (1.94 – 2.35)* Genital herpes 0.76 (0.71 – 0.80) 1.06 (0.69 – 1.43) 0.82 (0.70 – 0.95) Hydramnios / Oligohydramnios 1.87 (1.80 – 1.94) 2.95 (2.35 – 3.55)* 2.08 (1.88 – 2.28) Hemoglobinopathy 0.07 (0.06 – 0.08) 0.17 (0.02 – 0.31) 0.19 (0.13 – 0.26)* Chronic hypertension 0.90 (0.85 – 0.95) 0.89 (0.56 – 1.23) 1.20 (1.04 – 1.35)* Pregnancy-associated hypertension 7.61 (7.46 – 7.75) 6.39 (5.52 – 7.27)* 7.38 (7.01 – 7.75) Eclampsia 0.94 (0.89 – 0.99) 1.39 (0.97 – 1.81) 1.22 (1.07 – 1.38)* Incompetent cervix 0.81 (0.76 – 0.85) 0.80 (0.48 – 1.11) 0.72 (0.60 – 0.84) Renal disease 0.28 (0.25 – 0.31) 0.56 (0.30 – 0.83) 0.40 (0.31 – 0.49)* RH sensitization 069 (0.64 – 0.73) 0.53 (0.27 – 0.79) 0.49 (0.39 – 0.58)* Uterine bleeding 1.10 (1.04 – 1.16) 0.83 (0.50 – 1.15) 0.97 (0.83 – 1.10) Complication of labor Febrile 1.36 (1.30 – 1.43) 1.36 (0.95 – 1.77) 1.48 (1.31 – 1.65) Meconium 1.37 (1.31 – 1.43) 2.09 (1.58 – 2.69)* 1.92 (1.73 – 2.11)* Premature rupture of membrane (>12 h) 6.34 (6.21 – 6.47) 7.19 (6.27 – 8.11) 7.91 (7.53 – 8.29)* Abruptio placenta 1.16 (1.10 – 1.22) 1.03 (0.67 – 1.39) 1.44 (1.27 – 1.61)* Placenta previa 0.47 (0.43 – 0.51) 0.50 (0.25 – 0.75) 0.30 (0.22 – 0.38)* Seizures during labor 0.05 (0.04 – 0.06) 0.07 (0.03 – 0.16) 0.08 (0.04 – 0.12) Precipitous labor (<3 hours) 1.42 (1.36 – 1.49) 1.29 (0.89 – 1.70) 1.86 (1.67 – 2.04)* Prolonged labor (>20 hours) 0.56 (0.52 – 0.60) 0.63 (0.35 – 0.91) 0.52 (0.42 – 0.62) Dysfunctional labor 2.34 (2.26 – 2.42) 1.52 (1.09 – 1.96)* 1.97 (1.77 – 2.16)* Breech / malpresentation 21.3 (21.0 – 21.5) 21.4 (20.0 – 22.9) 21.4 (20.8 – 22.0) Cephalopelvic disproportion 0.90 (0.85 – 0.95) 0.76 (0.45 – 1.07) 0.52 (0.42 – 0.62)* Cord prolapse 0.55 (0.51 – 0.59) 0.56 (0.30 – 0.88) 0.65 (0.54 – 0.77) Anesthetic complications 0.07 (0.06 – 0.09) 0.10 (0.01 – 0.21) 0.09 (0.05 – 0.13) Fetal distress 3.20 (3.10 – 3.29) 3.45 (2.79 – 4.10) 3.44 (3.18 – 3.70) Obstetric intervention Cesarean 52.0 (51.7 – 52.3) 46.1 (44.7 – 48.2)* 47.5 (46.8 – 48.2)* Vacuum / Forceps 6.76 (6.62 – 6.89) 5.40 (4.59 – 6.21)* 4.46 (4.17 – 4.75)* Induction 13.0 (12.8 – 13.2) 11.9 (10.7 – 13.0) 10.2 (9.8 – 10.6)* P < 0.05 for group 2 or group 3 compared with group 1 Table 3 compares the fetal and neonatal mortality and morbidity among the three study groups. The rates of PTB, FGR, LBW, Apgar score <7, fetal mortality, neonatal mortality, and post-neonatal mortality were significantly higher in group 2 and group 3 than in group 1; the RRs of PTB, FGR, LBW, Apgar score <7, fetal mortality, neonatal mortality, and post-neonatal mortality ranged from 1.08 to 3.86 in group 2 and group 3 compared with group 1 (Table 3 ). The rates of hyaline membrane disease and assisted ventilation (< or > = 30 minutes) were also significantly higher in group 3 than in group 1 (Table 3 ). However there were no statistically significant differences in the rates of congenital anomalies among the three study groups (Table 3 ). Table 3 Comparison of fetal and neonatal mortality and morbidity among three study groups in twins in United States Partner information available (N = 129651) Partner information partly missing (N = 3018) Partner information totally missing (N = 19564) % (95%CI) % (95%CI) RR# % (95%CI) RR# General condition PTB (<37W) 53.5 (53.2 – 53.8) 57.8 (56.0 – 59.7) 1.08 59.6 (58.9 – 60.3)* 1.11 FGR (<10 percentiles z score) 8.26 (8.10 – 8.41) 10.9 (9.72 – 12.0)* 1.32 12.9 (12.4 – 13.4)* 1.56 LBW (<2500 g) 51.4 (51.1 – 51.7) 60.0 (58.2 – 61.7)* 1.17 64.6 (63.9 – 65.3)* 1.26 Apgar score (<7) 3.23 (3.13 – 3.32) 5.14 (4.35 – 5.92)* 1.59 6.45 (6.11 – 6.79)* 2.00 Fetal mortality 0.85 (0.81 – 0.89) 3.28 (2.65 – 3.90)* 3.86 1.59 (1.43 – 1.75)* 1.87 Neonatal mortality 1.92 (1.85 – 2.00) 3.83 (3.12 – 4.53)* 1.99 3.90 (3.62 – 4.17)* 2.03 Post-neonatal mortality 0.46 (0.43 – 0.50) 0.95 (0.59 – 1.32)* 2.07 1.12 (0.97 – 1.27)* 2.43 Abnormal conditions of the newborn Anemia (HGB.<13) 0.40 (0.36 – 0.43) 0.35 (0.13 – 0.57) 0.88 0.46 (0.36 – 0.55) 1.15 Birth injury 0.16 (0.14 – 0.18) 0.18 (0.02 – 0.33) 1.13 0.17 (0.11 – 0.23) 1.06 Hyaline membrane disease 3.28 (3.18 – 3.38) 2.88 (2.27 – 3.49) 0.88 3.87 (3.60 – 4.14)* 1.18 Meconium aspiration syndrome 0.09 (0.07 – 0.11) 0.11 (0.01 – 0.22) 1.22 0.09 (0.05 – 0.13) 1.00 Assisted ventilation (<30 minutes) 3.45 (3.35 – 3.55) 2.67 (2.08 – 3.26)* 0.77 3.87 (3.60 – 4.14)* 1.12 Assisted ventilation (> = 30 minutes) 3.82 (3.71 – 3.92) 3.23 (2.58 – 3.88) 0.85 5.04 (4.73 – 5.35)* 1.32 Seizures 0.07 (0.05 – 0.08) 0.18 (0.02 – 0.33) 2.57 0.10 (0.06 – 0.15) 1.43 Congenital anomalies Central nervous system anomalies 0.15 (0.13 – 0.18) 0.20 (0.04 – 0.36) 1.33 0.14 (0.09 – 0.19) 0.93 Circulatory / respiratory anomalies 0.43 (0.39 – 0.46) 0.60 (0.32 – 0.87) 1.40 0.45 (0.36 – 0.54) 1.05 Digestive system anomalies 0.11 (0.09 – 0.13) 0.17 (0.02 – 0.31) 1.55 0.11 (0.07 – 0.16) 1.00 Urogenital system anomalies 0.21 (0.19 – 0.24) 0.23 (0.06 – 0.40) 1.10 0.17 (0.11 – 0.23) 0.81 Musculoskeletal / integumental anomalies 0.38 (0.34 – 0.41) 0.43 (0.20 – 0.66) 1.13 0.40 (0.31 – 0.49) 1.05 Chromosomal anomalies 0.10 (0.09 – 0.12) 0.17 (0.02 – 0.31) 1.70 0.08 (0.04 – 0.12) 0.80 ** P < 0.05 for group 2 or group 3 compared with group 1; #: relative risk compared with group 1. Discussion Our large population based study found that the risks of fetal and infant mortality, low birth weight, preterm birth, fetal growth restriction, Apgar score <7, and the need for mechanical ventilation were increased in infants born to mothers with missing information on partner's age or race, especially when both variables of the partner were missing. The women reported no information (or reported only part of the information) on partners appear to have higher prevalence / incidence of most of the maternal morbidity and obstetric complications than in mothers with available information on partners. The poorer pregnancy outcomes observed in women with partner's information been partly and (especially) totally missing than in women with available information on partners were due largely to known important socioeconomic factors for adverse pregnancy outcomes. Many of these women were teenagers, black race, unmarried, with low education, high frequency of cigarette smoking, and inadequate prenatal care services. Teenagers [ 17 ], black race [ 18 ], unmarried [ 19 ], low education [ 3 ], cigarette smoking [ 4 ], and inadequate prenatal care services [ 20 ] are known risk factors of adverse pregnancy outcomes. Our findings that the rates of FGR, LBW, Apgar score <7, hyaline membrane disease, and assisted ventilation were significantly higher in the group with paternal information totally missing than the group of only partly missing suggest that the totally missing group is a particularly high risk group, whereas in the partly missing group, the missing may have occurred in randomly. Exceptions did occur, however. For example, the occurrences of several maternal medical conditions, such as cardiac disorders, diabetes, and RH sensitization, were actually lower in women with no available information on partners (group 2 and 3) than women with available information on partners (group 1). This may, in part, be due to the younger maternal age in groups 2 and 3 than in group 1. The risk of cardiac disorders and diabetes increases rapidly with advancing age [ 21 , 22 ]. Younger mothers are less likely to have had previous pregnancy and therefore less chance for RH sensitization [ 23 ]. The rates of obstetric interventions, including cesarean section, vacuum / forceps, and induction delivery, were lower in groups 2 and 3 than in group 1, which could be explained by the younger age of this group as cesarean section is less common in young women [ 24 ]. The rates of placenta previa and cephalopelvic disproportion were lower in women in groups 2 and 3 than in group 1, this could be explained by the younger age and lower parity of these groups as placenta previa and cephalopelvic disproportion are less common in younger and lower parity women [ 25 , 26 ]. We do not know why the rates of various congenital anomalies were not higher in infants born to mothers with no available information on fathers, despite their much higher risks of fetal and infant mortality and neonatal morbidity. Our study was based on birth certificates, and the diagnosis of many birth defects may be difficult at birth, especially when there is a lack of access to quality perinatal care services such as high resolution ultrasound [ 27 ], although differences in maternal age may play a role here. Our study examined the difference of adverse pregnancy outcomes among three study groups in twins only. It is biologically plausible to apply the study findings to singletons and higher order of multiple pregnancies as well, although the magnitude of the differences may be smaller in singletons and larger in higher order of multiple pregnancies than in twins. Since in the database, there was no indication why a particular variable was not recorded, we can only speculate the reason of missing partner's information. Corresponding variables (i.e., age and race) for the mothers, including those mothers with missing information on partners, were completely recorded. These women may have no relationship or only remote relationship with their partners, as suggested by the extraordinarily high unmarried rates (61% for women missing one variable on their partners and 96% for women missing both variables on their partners). As a result, they may have difficulty or for unknown reasons they may be reluctant to provides partner's information. The paternal information missing may be a comprehensive measure and may be related with age, education, socioeconomic status, race, religion background, or even the reporting quality in different place. Our study only described the association between the paternal information missing and the adverse pregnancy outcomes. Exploration of the reasons of missing partner information and its relationship with adverse pregnancy outcomes require in-depth analysis of data with relevant information. Conclusions Women with no available information (or partly) on partners appear to have higher risks of developing adverse pregnancy outcomes. For this reason, extra attention for these women and their offspring by health care providers and public health workers is needed. Competing interests The authors declare that they have no competing interests. Authors' contributions Hongzhuan Tan carried out the data analysis and drafted the manuscript. Shi Wu Wen designed the study and helped in results interpretation. Mark Walker participated in the sequence alignment and participated in the results interpretation. Kitaw Demissie participated in the study design and results interpretation. Pre-publication history The pre-publication history for this paper can be accessed here:	/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC535341.xml
539235	Allele specific synthetic lethality between priC and dnaAts alleles at the permissive temperature of 30°C in E. coli K-12	Background DnaA is an essential protein in the regulation and initiation of DNA replication in many bacteria. It forms a protein-DNA complex at oriC to which DnaC loads DnaB. DNA replication forks initiated at oriC by DnaA can collapse on route to the terminus for a variety of reasons. PriA, PriB, PriC, DnaT, Rep and DnaC form multiple pathways to restart repaired replication forks. DnaC809 and dnaC809 , 820 are suppressors of priA2::kan mutant phenotypes. The former requires PriC and Rep while the latter is independent of them. RnhA339::cat mutations allow DnaA-independent initiation of DNA replication. Results It is shown herein that a priC303::kan mutation is synthetically lethal with either a dnaA46 or dnaA508 temperature sensitive mutation at the permissive temperature of 30°C. The priC - dnaA lethality is specific for the dnaA allele. The priC303::kan mutant was viable when placed in combination with either dnaA5 , dnaA167 , dnaA204 or dnaA602 . The priC - dnaA508 and priC - dnaA46 lethality could be suppressed by rnhA339::cat . The priC - dnaA508 lethality could be suppressed by a dnaC809 , 820 mutation, but not dnaC809 . Neither of the dnaC mutations could suppress the priC - dnaA46 lethality. Conclusions A hitherto unknown function for either DnaA in replication restart or PriC in initiation of DNA replication that occurs in certain dnaA temperature sensitive mutant strains at the permissive temperature of 30°C has been documented. Models considering roles for PriC during initiation of DNA replication and roles for DnaA in replication restart were tested and found not to decisively explain the data. Other roles of dnaA in transcription and nucleoid structure are additionally considered.	Background The loading of the DnaB replicative helicase at the E. coli origin of DNA replication ( oriC ) is a highly regulated and is thought to be a key step in the assembly of the replisome. DnaB makes important contacts with the τ-subunit of DNA Polymerase III Holoenzyme and DNA primase [ 1 ]. DnaB loading at oriC during initiation of DNA replication is a sequence specific, cell cycle regulated event dependent on the DnaA and DnaC proteins (reviewed in [ 2 - 4 ]). In vitro , in the presence of several other accessory proteins ( i.e. RNA polymerase, DNA gyrase, HU protein), multiple DnaA proteins bind to four asymmetric 9 bp DnaA binding sites in the 225 bp oriC region allowing formation of a protein/DNA complex [ 5 - 7 ]. This in turn causes melting of a nearby AT rich sequence. A complex of DnaB 6 -DnaC 6 then interacts with the DnaA/ oriC complex to load DnaB at the AT rich sequence. It is thought that DnaA may have other roles in the cells in addition to initiation. These additional roles stem from the fact that there are many DnaA binding sites in the chromosome outside the oriC region [ 8 , 9 ] and that DnaA binding to these asymmetric 9 bp sites, can bend the DNA [ 10 ]. It can be hypothesized based on large number of potential DnaA binding sites in the chromosome and the ability of DnaA to bend DNA that it may influence the structure of nucleoid. It has been shown that if a DnaA binding site falls within a promoter region that mutations in dnaA can affect the level of transcription from that promoter [ 11 - 14 ]. Thus, mutations in dnaA may have global effects in gene expression and nucleoid structure as well as affecting initiation of DNA replication at oriC . In E. coli , dnaA is an essential gene. Several different dnaA temperature sensitive mutant alleles have been isolated. Many of these share the property that they are double mutants ( dnaA5 , dnaA46 , dnaA508 and dnaA602 – Table 1 and [ 15 ]). Several of the double mutants share a mutation: a change at codon 184 that replaces an alanine with a valine. Additionally, a dnaA850:: Tn 10 mutant has been isolated. This is only viable in strains that have an alternate, oriC -independent method of initiation of DNA replication [ 16 ]. The loading of DnaB by DnaC in E. coli can occur away from oriC at a repaired replication fork. This is governed by the Replication Restart Proteins (RRPs): PriA, PriB, PriC, DnaT and Rep. The genes coding for these products form multiple pathways to identify the proper substrate and then help DnaC load DnaB (Figure 1 and [ 17 ]). Since replication restart is thought to be an essential process, the poor viability (versus complete inviability) of priA mutants suggested the availability of alternate pathways. The PriA-independent pathway depends on PriC and Rep [ 17 ]. Two types of priA suppressor mutations have been found and both map to the dnaC gene. The first typified by dnaC809 (E176G) and is dependent on the genes in the PriA-independent pathway of replication restart, rep and priC [ 17 ]. The second type of priA suppressor has an additional mutation (K178N) in dnaC relative to dnaC809 and makes this protein's suppression of priA mutant phenotypes independent of priC and rep . This dnaC allele is called dnaC809 , 820 [ 17 , 18 ]. The multiplicity of replication restart pathways may be a general property in Bacteria since Bacillus subtilis also has a similar arrangement of PriA-dependent and PriA-independent pathways [ 19 , 20 ]. As mentioned above, DnaA-dependent initiation of DNA replication at oriC and replication restart share several properties. The most important of these is that both strive to make protein-DNA complexes that are recognized by the DnaC protein so that DnaB can be loaded. Previous work by Kogoma and colleagues showed that oriC -DnaA-independent initiation of DNA replication could take place in an rnhA mutant strain [ 16 , 21 ]. This type of initiation of DNA replication was termed Constitutive Stable DNA Replication (cSDR) and is dependent on both recombination and replication restart functions (reviewed in [ 22 ] and Sandler, submitted). To begin testing the roles of priB and priC in cSDR (reviewed in [ 22 ]) required the construction of dnaA ts rnhA priB and dnaA ts rnhA priC triple mutant strains. However when trying to construct these strains, we found that priC was required for growth in two different dnaA ts strains, dnaA46 and dnaA508 , at the permissive temperature of 30°C. When priC303::kan was tested with other dnaA ts alleles, the synthetic lethality was found to be allele specific. Two different mutations were found to suppress the priC - dnaA lethality. One was rnhA339::cat , a non-allele specific suppressor of dnaA mutants. The other was dnaC809 , 820 , a suppressor of priC and priA mutations. The first type suppressed both the dnaA508 - priC and dnaA46 - priC lethality while the latter only suppressed the dnaA508 - priC lethality. These studies suggest that priC may have a role in initiation of DNA replication at oriC with certain dnaA alleles and or that dnaA may have an additional role in the cell important for replication restart. Results PriC , but not priB , is required for growth in a dnaA508 mutant at the permissive temperature of 30°C We began this study by asking if priB and priC are required for cSDR. During this study it was found that certain dnaA ts could be introduced into a strain containing a priB mutation, but not into a strain containing a priC mutation. The standard P1 transductional cross used for the introduction of the dnaA ts mutant alleles is shown in Figure 2 . Table 2 shows that the co-transduction frequency between tnaA300:: Tn 10 and dnaA508 in a wild type strain is 92% (49/53). Using a priB mutant strain as the recipient, the co-transduction frequency between tnaA300::Tn10 and dnaA508 was approximately the same as when the wild type strain was used as the recipient (data not shown). Surprisingly, when a priC mutant was used as a recipient, the co-transduction frequency was 0/72 or 0% (Table 2 ). This suggested that the priC303::kan and dnaA508 mutation may be synthetically lethal at the permissive temperature of 30°C. It is also formally possible that priC303::kan suppressed the temperature sensitive nature of dnaA508 . These two possibilities are tested below. Since it is known that the absence of other gene products (i.e., rnhA [ 23 ] and trxA [ 24 ]) can suppress the essentiality of dnaA , it is possible that the absence of priC might also suppress the temperature sensitivity of the dnaA ts allele. If so, then one should be able to detect the presence of the dnaA mutation on the chromosome of the 42°C resistant transductants. To test this possibility, Tet R transductants selected at 30°C were screened for a 42°C R phenotype. These were then further screened for the presence of a restriction site polymorphism (a Eco NI site) created by the dnaA508 mutation. To do this, the dnaA region from the 42°C R Tet R transductants was amplified by standard Polymerase Chain Reaction methods using the primers, prSJS480 and prSJS481 (Table 3 ). The amplified DNA was then restricted with Eco NI. Examination of eight independent 42°C R Tet R transductants, constructed using JC12390 as the donor, revealed no case in which a restriction pattern was consistent with the presence of the temperature sensitive allele (data not shown). From these results, it is concluded that the priC303::kan mutation does not suppress the absence of dnaA508 and is synthetically lethal with it. PriC303::kan is synthetically lethal with dnaA46 and dnaA508 , but not with dnaA5 , dnaA167 , dnaA204 or dnaA602 It is possible that either the dnaA508 - priC303::kan synthetic lethality at the permissive temperature of 30°C is allele specific or occurs with all dnaA ts alleles. To distinguish between these two possibilities, several other dnaA ts alleles were tested. Selection of a diverse collection of mutations to test was aided by an already large repertoire of characterized dnaA ts alleles [ 15 ] and the recent elucidation of the crystal structure of DnaA from Aquifex aeolicus [ 25 ]. This allowed the selection of several temperature sensitive dnaA alleles that had different amino acids substitutions in different parts of the protein (Table 1 ). Hence, it was attempted to introduce dnaA5 , dnaA46 , dnaA167 , dnaA204 and dnaA602 into the priC303::kan strain (SS145) using the selectable marker tnaA300::Tn10 as before. Table 2 shows that the synthetic lethality only occurred additionally with dnaA46 , but not with dnaA5 , dnaA167 , dnaA204 or dnaA602 . It is concluded that the synthetic lethality between priC303::kan and dnaA508 and dnaA46 at 30°C is allele specific. The priC303::kan dnaA ts synthetic lethality is solely due to the absence of only priC and the presence of the dnaA ts mutation Since both the dnaA and priC genes are in operons, it is possible that the synthetic lethality seen above is due to not just the mutation in priC or dnaA , but also due to that mutation and or polar effects on downstream genes within their respective operons. While the potential polar effects of a priC303::kan insertion mutation are easily envisioned, the potential polar effects of a dnaA missense mutation are less obvious. This is tested here as well because, as introduced above and discussed below, dnaA mutations can have effects on the level of transcription of promoters in which there are DnaA binding sites. Since dnaA binds to its own promoter and autoregulates its own expression [ 12 , 26 ], it is possible that dnaA mutation may effect transcription from its own promoter and subsequently effect dnaN and recF expression. It was first tested if the synthetic lethality between the dnaA508 and priC303::kan was dependent solely on the priC gene. This was necessary to determine because priC303::kan is an insertion mutation and could be polar on the downstream gene, ybaM . This was tested by cloning priC into a plasmid (pTH1, see Methods) and seeing if the priC plasmid could complement the synthetically lethal phenotype. pTH1, containing just the priC promoter and gene, in the priC303::kan mutant strain (SS145), allowed the dnaA508 allele to be introduced into that strain at the wild type co-transduction frequency of 90% (data not shown). It was then tested if the synthetic lethality was due to polar effects of dnaA508 or dnaA46 on downstream dnaN - recF expression. This was tested in a similar way. A plasmid, pAB3 [ 27 ], that expresses the dnaA gene in trans was introduced into the priC303::kan mutant strain (SS145). This strain was then used as a recipient in a cross with either ALO450 ( dnaA46 tnaA300:: Tn 10 ) or JC12390 ( dnaA508 tnaA300:: Tn 10 ). Tet R transductants were selected at 30°C. In each case, several transductants were selected and screened for the presence of the dnaA ts mutation by backcrosses to JC13509. The temperature sensitive phenotype associated with dnaA508 and dnaA46 was detected in each case (data not shown). It is concluded that the synthetic lethality seen between priC303::kan and dnaA508 or dnaA46 is due to solely the absence of priC and the presence of the dnaA ts mutation. RnhA mutations suppress the priC-dnaA synthetic lethality It has been shown that rnhA mutations are non-allele specific suppressors of both dnaA ts and dnaA insertion mutations [ 16 ]. The mechanism of suppression is thought to be the stabilization of R-loops on the chromosome [ 22 ]. To determine if the priC - dnaA synthetic lethality is suppressed by a mutation in rnhA , it was attempted to introduce dnaA508 and dnaA46 into a priC303::kan rnhA339::cat (SS1531) double mutant strain. It was found that this dnaA ts rnhA priC triple mutant combination was viable in each case (see SS1543 and SS3032 in Table 4 ). This suggested that the cause of the priC - dnaA lethality was a defect in the mutant dnaA protein's ability to initiation of DNA replication and that priC has some role in initiation of DNA replication in the dnaA508 and dnaA46 mutant strains. DnaC809 , 820 , but not dnaC809 , suppresses the absence of priC in the dnaA508 mutant at 30°C The above experiment suggested that PriC has a role in initiation of DNA replication in certain dnaA mutants. If so, then suppressors of priC 's role in replication restart should not suppress the priC - dnaA synthetic lethality. Two types of replication restart suppressor mutations are known and were tested [ 18 , 28 ]. DnaC809 suppresses the phenotypes of priA2::kan and dnaT82 2 [ 28 , 29 ]. In vitro , DnaC809 can suppress the absence of all RRPs on several different substrates [ 30 ]. In vivo however, priA2::kan suppression requires priC and rep [ 17 ]. DnaC809 can be additionally mutated to make the priA suppression both priC and rep independent [ 17 ]. This additionally mutated dnaC allele is called dnaC809,820 [ 18 ]. To test the above hypothesis, priC303::kan dnaC809 (SS1099) and priC303::kan dnaC809,820 (SS1100) strains were constructed (Table 4 ) and used as recipients in crosses with the donor P1 from either JC12390 ( dnaA508 ) or AL0450 ( dnaA46 ). Table 2 shows that when dnaC809 , 820 was used as the recipient and JC12390 as the donor that 41/48 or 83% of the Tet R transductants were also temperature sensitive (they inherited the dnaA508 allele). However, when only dnaC809 was used as the recipient, 0/63 Tet R transductants inherited the temperature sensitive phenotype. It is concluded that dnaC809 , 820 can suppress the absence of priC in the dnaA508 mutant and dnaC809 cannot. The dnaA46 allele was additionally tested and was not suppressed by either dnaC809 or dnaC809 , 820 (Table 3 ). From this it is concluded that dnaC809 , 820 is able to suppress the absence of some dnaA ts allele. In contradiction to the suggestion of the above rnhA experiment, this result suggests that dnaA may have some role in replication restart necessary in a priC mutant. Discussion This paper shows that priC , a gene involved in both the PriA-dependent and PriA-independent pathways for replication restart, is also required for cell viability in two of six dnaA ts mutants at the permissive growth temperature of 30°C. These results were surprising on at least two accounts. The first is that in vitro systems for either replication restart or initiation of DNA replication at oriC posed no requirement for the DnaA or PriC protein respectively. The second is that priC has no known role in vivo in initiation of DNA replication (the only reported role is in replication restart [ 17 ]) and that dnaA has no known role in replication restart. One way to answer the question of why the mutations are synthetically lethal is to see what types of mutations may suppress the lethality. RnhA339::cat , a non-allele specific suppressor of dnaA mutants role in initiation of DNA replication, could suppress the priC - dnaA synthetic lethality for both dnaA ts mutant alleles. Such suppression is strong evidence that priC and dnaA may both be missing a function needed during initiation of DNA replication. It was further observed, however, that dnaC809 , 820 (but not dnaC809 ) could suppress the absence of priC in the dnaA508 mutant. Neither dnaC809 nor dnaC809 , 820 could suppress the dnaA46 - priC synthetic lethality. DnaC809 , 820 is a PriC-independent suppressor of several genes required in replication restart. This suppression implicates dnaA in replication restart. Thus the inferences from the two types of suppressors seem to contradict one another. What function(s) in dnaA ts strains are missing for initiation of DNA replication that make the strain dependent on priC at 30°C? A structure-function analysis of DnaA would help to answer this question. Briefly, based on alignments of DnaA proteins, the X-ray crystal structure of the DnaA protein from Aquifex aeolicus and much research on the genetics and biochemistry of DnaA, the DnaA protein can be divided into four domains with four proposed functions: Domain I) DnaB recruitment, Domain II) Linker region, Domain III) ATP binding and Domain IV) DNA binding [ 25 , 35 , 36 ]. Table 1 shows that the six mutations tested substitute amino acids spread throughout DnaA. The two mutants that show a requirement for priC have mutations in Domains I (DnaB recruitment) and III (ATP hydrolysis). However, several of the mutations not requiring priC also affect Domain III. An interesting aspect to dnaA genetics is that many temperature sensitive mutants have two mutations (Table 1 and [ 2 ]). The dnaA5 , dnaA46 and dnaA602 all have mutations in Domain III near the ATP binding region. Their second mutations cause amino acid replacements in other domains. Unfortunately, the positions of the second changes yield no clues about what might make dnaA508 and dnaA46 mutants require priC for growth at 30°C and why the other four dnaA mutants do not. What might PriC be doing to help initiation of DNA replication in a dnaA ts strain? One idea is that the dnaA ts protein is defective in its ability to create a region of ssDNA at oriC . Since the RRPs are also thought to help create regions of ssDNA (away from oriC ) so that DnaC binds and loads DnaB, it is possible that PriC may help the DnaA mutants in this endeavor. Another type of model that is formally possible is that PriC may somehow stabilize the dnaA ts protein. This seems unlikely, however, given that dnaC809 , 820 can rescue the synthetic lethality of priC303::kan and dnaA508 . Other models may also be possible. One needs to consider if DnaA may be involved in replication restart. In considering this, one needs to remember that DnaA has the ability to bind DNA at specific sites and bend it. It has been shown that different dnaA ts allele can differentially influence the rate of initiation of transcription in some promoters with DnaA binding sites (see below). This in turn can influence replication restart in two ways. First changes in the level of gene expression of a single gene (or groups of genes) may indirectly influence the replication restart process. Second, the ability of DnaA to bind to many sites on the chromosome may influence structure of the nucleoid and the sites at which replication restart may occur. There are many examples where several phenotypes had been tested systematically for several dnaA alleles (Table 1 and [ 31 - 33 ]). The only other system that seems to have some similarity to the data here is one in which the ability to replicate λ P + plasmids was investigated [ 27 , 34 ]. Table 1 shows that dnaA46 , dnaA204 and dnaA508 all fail to replicate these plasmids while dnaA5 , dnaA167 and dna602 can. The model proposed to explain this phenomenon suggests that DnaA is required to activate transcription at the λ P R promoter and that the dnaA46 , dnaA204 and dnaA508 mutations decrease this ability[ 27 ]. With the exception of dnaA204 , the inability to replicate these plasmids mirrors the ability of the dnaA ts mutants to grow in the absence of priC . The results presented in this paper do not allow one to definitively know whether the synthetic lethality studied here is due to a failure in initiation of replication at oriC or is it due to a failure in replication restart based on the study of the suppressors. Since dnaA mutations can affect more than just initiation of DNA replication, it is tempting to speculate that some other dnaA function: transcription of a particular gene or set of genes or the shape of nucleoid in the priC mutant may contribute or be the cause of the synthetic lethality. Understanding the molecular mechanism underlying the dnaA ts - priC synthetic lethality may require appreciation of these other aspects of dnaA biology. Conclusions A hitherto unknown function for either DnaA in replication restart or PriC in initiation of DNA replication that occurs in certain dnaA temperature sensitive strains at the permissive temperature of 30°C has been documented. Models considering roles for PriC during initiation of DNA replication and roles for DnaA in replication restart were tested and found not to decisively explain the data. Other roles of dnaA in transcription and nucleoid structure are additionally considered. Methods Bacterial strains All bacterial strains used in this work are derivatives of E. coli K-12 and are described in Table 4 . The protocol for P1 transduction has been described elsewhere [ 37 ]. All P1 transduction were selected on 2% agar plates containing either minimal or rich media and either tetracycline 10 μg/ml or kanamycin 50 μg/ml final concentration. All transductants were first purified on the same type of media on which they were selected. Tests for temperature sensitivity were then done by replica plating patches of the purified transductants at 30°C and 42°C on solid rich media without any antibiotics. Growth was scored by either the presence or absence of a patch after 24 hours. Cloning of the priC gene Wildtype chromosome DNA was used as the template in a standard PCR reaction using prSJS283 and prSJS284 (Table 3 ) as the priming oligonulceotides. The amplified PCR fragment (that includes the putative promoter) was purified by gel electrophoresis and cloned into the pCR 2.1 using the TOPO-TA cloning system from Invitrogen. The priC containing plasmid was called pTH1. Authors contributions TH carried out the initial part of the molecular genetic studies. These were completed by SJS. SJS conceived of the study and wrote the manuscript.	/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC539235.xml
552331	Australia's international health relations in 2003	A survey for the year 2003 of significant developments in Australia's official international health relations, and their domestic ramifications, is presented. The discussion is set within the broader context of Australian foreign policy. Sources include official documents, media reports and consultations with officers of the Department of Health and Ageing responsible for international linkages.	Australia's health relations with other nations in the field of health constitute an important sub-set of health policy not only because of the intrinsic significance of bi-lateral and multilateral linkages, but also because of their ramifications for health policy at the domestic level. In broad terms, these health relations encompass a range of interactions with consequences for health, including: membership of global and regional bodies; the negotiation of international agreements; action to counter particular external threats to health; assistance to developing countries; and international trade and investment in health-related goods and services. In 2003 there were continuing developments in all these areas within a wider foreign affairs context overshadowed by official policy concerns about global and regional security, the deployment of the Australian armed forces in various theatres of service, and renewed fears of the human and economic costs of infectious diseases. Balancing these concerns with national defence were renewed efforts to forge bi-lateral trade links in global trade environment characterized by the emergence of trade blocs centred in North America, Europe and Southeast Asia. Although consultation occurs with states and territories, it is the Australian Government that is constitutionally responsible for conducting Australia's international relations. These responsibilities include appointing representatives to international bodies and organizations, such as the United Nations and its various agencies, including the World Health Organization and assenting to agreements and regulations promulgated by international agencies. The formulation and implementation of policy with direct or indirect international health ramifications is not centralized, but is usually the result of consultations between various relevant government departments and statutory authorities. An important element of the Australian Government's foreign affairs powers relates to international treaties. While a degree of consultation with state and territory governments and with the public occurs, and the national parliament is able to scrutinize and comment upon international treaties, it is the executive that has the final decision on such agreements. WHO and other international agencies concerned with health In 2003 Australia continued to play a strategically important and respected role in international organizations concerned with health, especially the World Health Organization. At the World Health Assembly, the governing body of the WHO, the Australian delegation supported resolutions concerned with strengthening nursing and midwifery and child and adolescent health. In the wake of SARS, Australia also supported the review of the International Health Regulations and is likely to subscribe to them [ 1 ]. The voluntary nature of WHO standards and regulations, which can be accepted or rejected by member states, is well illustrated by the International Health Regulations since Australia and Papua New Guinea declined to accept them when they were last promulgated. Australia should be better placed to influence developments in WHO in the next three years as a result of being nominated for a term on the Executive Board. The Department of Health and Ageing was closely involved with international comparative health data projects including WHO's World Health Survey and the health systems performance survey of the Organization for Economic Cooperation and Development. Australia's participation in the health mandate of the Commonwealth of Nations was illustrated by the Therapeutic Goods Administration's provision of a secretariat for the Clearing House of Commonwealth Agencies for Chemical Safety. Australia also participated in the meeting of Commonwealth health ministers on the eve of the annual World Health Assembly of WHO in Geneva [ 2 ]. International agreements In December Australia signed the Framework Convention on Tobacco Control (FCTC) the first multilateral treaty negotiated under the auspices of the World Health Organization. For the first time, nations were invited to implement control measures covering such issues as health warnings, advertising, packaging and labelling, sales, and smuggling. They were also called upon to embrace policy measures designed to counter the global tobacco epidemic [ 3 ]. The FCTC provided an impetus to the domestic policies of many countries with limited progress on tobacco control and also allowed for the transnational activities of tobacco corporations to be countered with global policy action. The FCTC has limited potential to further Australian domestic policy, which is in advance of that in most countries. If necessary, the Australian Government could call upon its "external affairs" to assert constitutional primacy over this policy area. However, this is unlikely in the context of close cooperation between various levels of government in Australia in establishing national tobacco control policies. Australian leadership was evident in WHO's formulation of the FCTC, having been nominated by the Western Pacific region as vice-chair of the Bureau for the Negotiating Body. A reciprocal health care agreement with Norway was signed, further expanding the rights of Australian residents to immediate and necessary treatment in the national health systems of countries with which Australia has reciprocal treaties. These include New Zealand, UK, Italy, Malta, Holland, Sweden, Finland, and the Republic of Ireland. These arrangements are "cost neutral" and do not include costly accounting or administrative procedures. In terms of domestic policy, the continuing "internationalisation" of Medicare (pioneered by the Hawke Labor Party ministry at the time of Medicare's introduction) by the Liberal-National Party Coalition is paradoxical since local citizens are being encouraged to opt out of public hospital treatment through a rebate on private health insurance and penalties for higher income earners who do not insure privately. Whilst these agreements have cemented closer diplomatic ties, their potential benefits to international travellers, especially those subject to punitive insurance premiums or the refusal of insurance due to old age or infirmity, remain inadequately publicized. Treaties are also being negotiated with Denmark and Belgium. Following years of negotiations and planning, a treaty was signed with New Zealand establishing a single joint therapeutic goods agency. This body, due to commence operations in 2005, will regulate prescription and retail drugs, therapeutic devices and also complementary medicines. It will replace the Australian Therapeutic Goods Administration and its New Zealand counterpart. To a large extent, the two regulatory systems will have been integrated, although there are still areas of disagreement (e.g. policies on the advertising of PBS medicines) which will need to be negotiated. This joint agency creates a model in international health relations which other states could profitably emulate where they share common concerns and have similar health systems. In December 2002 the two countries finalized treaty arrangements establishing both a joint standards code and a joint statutory authority, Food Standards Australia New Zealand [ 4 ]. These arrangements parallel bi-lateral developments for the joint regulation of food standards. These developments have furthered Australian foreign policy concerned with establishing trans-Tasman free trade, commenced some two decades ago with the negotiation of the Closer Economic Relations agreement with New Zealand. The new regulatory arrangements have created a virtual trans-Tasman free market in food (subject to plant and animal quarantine considerations) and therapeutic drugs. While not having the legal status of a treaty, for some years the Department of Health and Ageing has had memoranda of understanding with its counterparts in China, Indonesia, Thailand and Japan. In 2003 further activities were undertaken under the auspices of these agreements. During the state visit of China's president Hu Jinta, a plan of action was signed between the two health ministries. The Indonesian relationship continued with the inclusion of a health delegation to the Sixth Australia Indonesia Ministerial Forum in Jakarta in March, preceded by two rounds of meetings between officials of the Indonesian and Australian health departments. The Australia-Japan Partnership in Health and Family Services formed the basis for negotiations for joint research on mental health and an international conference on suicide prevention [ 5 ]. In a related development, the Department of Foreign Affairs and Trade promoted aged care expertise as an export service through the Australia Japan Conference. In the course of 2003 Australia finalized free trade agreements (in reality, preferential trade agreements) with Singapore and Thailand and continued negotiations with the USA [ 6 ]. From the perspective of the Australian health industry, the agreement with Singapore offered tariff-free trade in pharmaceuticals and other therapeutic goods and the gradual removal of tariffs in the case of Thailand. All countries imposed reservations on free trade in the sensitive areas of health services, although traditional Thai massage exponents will be permitted to operate in Australia. Domestically, these agreements required intersectoral policy collaboration in the interests of health. Policy makers in the Department of Health needed to intensify their understanding of the dynamics of international trade, while those making foreign policy had to consider the health dimensions of ostensibly commercial arrangements. The free trade agreement with the USA raised controversies about attempts to include the Pharmaceutical Benefits Scheme (PBS) in concessions demanded by US negotiators. These issues have been outlined in the account of developments in the PBS elsewhere in this series of review articles. SARS The emergence and rapid spread of Severe Acute Respiratory Syndrome (SARS) to several countries in East and Southeast Asia and to Canada revived popular atavistic fears of pandemics and damaged the tourism and travel industry, as well as some Australian suppliers of goods and services to Asia. WHO issued a global alert on the disease in March, and the last reported case of international occurred in July. So serious was the threat of SARS to the economies of some countries that a special meeting of health ministers, attended by Australia, was organized by Asia-Pacific Economic Cooperation (APEC) to discuss the situation. A task force was subsequently established by APEC to deal with SARS. An example of the economic costs of the disease was the decision by the Governments of Singapore and Australia to postpone negotiations on a greater share of the Sydney-Los Angeles air route (dominated by QANTAS) for the Singapore carrier, due to uncertainty about demand. In Australia SARS was declared a quarantinable disease under the Quarantine Act 1908 and policy guidelines for health professionals, airline and border control staff and the general public were developed by the Department of Health and Ageing, which also led an inter-departmental task force to monitor world developments. Until July 2003, when the WHO announced that no country was still considered SARS-affected, international aircraft arriving at Australian airports were required to obtain "SARS-free" clearance, nurses were posted at airports and restrictions on elective surgery were placed on travellers returning from affected countries [ 7 ]. During the period of WHO's alert, Australia had reported only five probable, and one laboratory-confirmed, case of SARS. Health and foreign aid Australia's official international development assistance programme is an important foreign policy tool, especially in the Asia-Pacific region. Some $225 m. (of a total of $ 1.8 b.) was allocated to health-related international development assistance in 2003–4 budget of the Australian Agency for International Development (AusAID). However, while Australia's contribution to HIV/AIDS control and its regional advisory role associated with SARS were acknowledged by the Foreign Affairs Minister in his report to Parliament, health assistance received little prominence. Security, good governance and counter-terrorism were emphasized as the focus for the official foreign aid programme. Support for essential services in Papua New Guinea continued as a major imperative [ 8 ]. The lower priority of health was further underscored by a decision to no longer appoint designated health advisors to the permanent staff of AusAID. It should be noted, however, that the emergence of SARS served to reinforce health as an important element on the international assistance agenda. Global health workforce mobility The fact that the domestic health workforce is now part of a global market for skilled workers was further demonstrated by continuing efforts to recruit nurses from overseas, the decision of the Australian Health Ministers Conference to sanction dentists from selected Commonwealth countries to work in public clinics. In addition, a scheme to recruit overseas-trained medical practitioners was included in the Australian Government's Medicare Plus policy initiatives [ 9 ]. It is intended that these doctors will work in rural and remote areas officially designated as having medical workforce shortages and also in positions within Aboriginal Controlled Community Health Services. Yet, metropolitan hospitals are have also become reliant upon overseas-trained doctors for their staffing. This policy, accompanied by the liberalization of immigration arrangements for medical doctors, has represented a volte face from previous policies deliberately designed to discourage foreign doctors from immigrating in the belief that controlling the number of doctors would contribute to cost-containments of Medicare. It also continues to raise the ethical danger of Australia contributing to a "brain drain" of medical staff from countries that are themselves short of such expertise. In 2002 the Commonwealth of Nations had agreed to a code of practice for the international recruitment of health workers to help safeguard the interests of developing nations. Australia has endorsed the code. The Australian Government will need to handle policies associated with the recruitment of overseas-trained health personnel with care due to professional sensitivities and the need for legislation at the state level to regularize the status of some professions. Concluding observations This brief review of Australia's international health relations in 2003 has demonstrated that health must be seen as an integral part of trade and security within the wider foreign policy context. The protection of health in free trade arrangements is important for their domestic legitimacy. It is vital that those involved in health policy are aware of its potential international dimensions, while those responsible for foreign affairs include health in their approach. Official health linkages have served to promote good will in some otherwise difficult relationships, as has been the case with Indonesia. They have also helped to promote a positive international image for Australia. Note The opinions expressed in this article are the sole responsibility of the author.	/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC552331.xml
554973	Food selection associated with sense of coherence in adults	Background Favorable dietary habits promote health, whereas unfavorable habits link to various chronic diseases. An individual's " sense of coherence" (SOC) is reported to correlate with prevalence of some diseases to which dietary habits are linked. However, understanding what determines an individual's dietary preferences and how to change his/her behavior remains limited. The aim of the present study was to evaluate associations between dietary intake and SOC in adults. Methods Diet intake was recorded by an 84-item semi-quantitative food frequency questionnaire and SOC was measured by the 13-item Antonovsky questionnaire in 2,446 men and 2,545 women (25–74 years old) from the population based northern Sweden MONICA screening in 1999. Results Intakes of energy, total and saturated fat, ascorbic acid, sucrose, and servings of fruits, vegetables, cereals, and sweets correlated with SOC among women, whereas intakes of total and saturated fat, ascorbic acid, fiber, and alcohol, and servings of fruits, vegetables, bread, bread and cereals, fish, and potatoes correlated with SOC among men. With a few exceptions, intakes of these nutrients/foods were significantly explained by SOC quartile scores in linear GLM models. Both women and men classified into the highest SOC quartile had significantly higher age-BMI-education standardized mean intakes of vegetables than those in the lowest quartiles. Women in the highest SOC quartile also had higher intake of fruits but lower intakes of energy, total and saturated fat, sucrose, and sweets. Projection to latent structures (PLS) multivariate modeling of intakes of the 84 food items and food aggregates simultaneously on SOC scores supported low SOC to coincide with a presumably less health promoting dietary preference, e.g. intake of pizza, soft drinks, candies, sausages for main course, hamburgers, mashed potato, chips and other snacks, potato salad, French fries, whereas men and women with high SOC scores were characterized by e.g. high intake of rye crisp whole meal bread, boiled potato, vegetables, berries, and fruits. Conclusion Both men and women in the highest, as compared with the lowest, SOC score quartile reported more "healthy" food choices. Dietary habits for individuals in the lowest SOC quartile therefore may render a higher risk for various endemic diseases.	Background Favorable dietary habits promote health, whereas unfavorable habits are linked to development of various chronic diseases, such as cardiovascular diseases, type-2 diabetes and dental caries. An excess intake of energy, excess refined sugars and saturated fatty acids, and too little fibers and antioxidants can contribute to the development of chronic diseases [ 1 - 4 ]. In general a favorable dietary pattern is characterized by a rich content of fruits, vegetables and fiber-rich cereal products and a low content of fat and refined sugar. A diet rich in saturated fat and sugar, but low in fiber, fruit and vegetables is considered unfavorable [ 5 ]. The public health message during the latest decades, i.e. to reduce total fat intake, especially saturated fat, and to eat more vegetables, fruits and whole meal cereals, has been constant [ 6 , 7 ]. An individual's dietary pattern is largely set by cultural traditions and availability, but both physiological and psychological influences have been described [ 8 ]. However, understanding what determines an individual's dietary preferences and how to change his/her behavior is limited. The medical sociologist, Aaron Antonovsky [ 9 ] was of the opinion that approaches to health and disease can be either salutogenic (origins of health) or pathogenic (origins of disease). He showed health to be connected to an individual's " Sense of coherence" (SOC), and " generalized resistance resources " (GRRs), such as income, education, ego strength, knowledge, which would provide energy to combat various stressors, and thus influential factors in the salutogenic model. Central to sense of coherence are comprehensibility (the cognitive component), manageability (the instrumental component), and meaningfulness (the emotional component). Individuals with low SOC scores are reported to have a higher frequency of various diseases [ 10 - 13 ], some of which are also linked to dietary habits [ 14 - 17 ]. An association between SOC and food selection/eating pattern is indicated by the reported ( i ) higher sucrose intake in adolescents with low SOC scores [ 14 ], ( ii ) lower ability to change dietary habits and lose weight in over-weight individuals with moderate to low SOC scores [ 15 ], and ( iii ) better blood sugar control in type-2 diabetics with high SOC scores, whereas those with low scores have poorer sugar balance [ 16 ]. The aim of the present study was to evaluate the association between dietary intake and SOC in adults. The hypothesis was that low SOC scores were associated with less favorable habits and vice versa. Methods Study population The Northern Sweden MONICA Project was performed in Västerbotten and Norrbotten, the two most northerly counties in Sweden, with a total population of around half a million inhabitants and a high prevalence of cardiovascular disease. Surveys were performed in 1986, 1990, 1994 and 1999 [ 18 ]. In each of the age groups 25–34, 35–44, 45–54 and 55–64 years 250 men and 250 women were randomly selected and invited to participate. In 1994 and 1999, the age group 65–74 years was added. Every person selected was invited to an examination at the nearest health care centre. In the 1999 survey, all participants from the three previous cohorts (n = 5,129), as well as a new cross-sectional sample of 2,500 randomly selected people, were invited. In total 6,000 individuals (71.8 %) participated. For the present paper all individuals aged 75 years or older (n = 206) were excluded. In the MONICA project various cardiovascular risk factors, including weight, height and education level, were monitored. Education was grouped into three levels: primary school (≤ 9 years), secondary school (10–12 years) and university education (≥ 13 years). Body mass index (BMI) was calculated as the ratio between weight (kg) and height 2 (m). BMI was grouped as <20 underweight, 20–24.9 normal weight, 25 – 26.9 moderate overweight, 27 – 29.9 overweight, and ≥ 30 obese. Special forms for questions about food intake and sense of coherence were added. Recording of sense of coherence (SOC) Sense of coherence was monitored using the 13-item questionnaire by Antonovsky [ 9 , 19 ]. Participants who had not answered all 13 questions were excluded. The SOC scores, with a theoretical range from 13 to 91, were calculated as described previously [ 9 , 10 , 19 ]. High scores denote strong SOC. Recording of dietary intake The subjects were requested to complete a self-administered, semi-quantitative and optically readable food frequency questionnaire [ 20 ]. Frequencies of consumption of 84 food items were reported on an increasing, nine-level scale, including never, maximum once a month, 1–3 times per month, once a week, 2–3 times a week, 4–6 times a week, once a day, 2–3 times a day, and 4 or more times a day. The questionnaire included eight questions on various types of fats, nine on milk and other dairy products, eight on bread and cereals, ten on fruit, greens and root vegetables, and nine on soft drinks and sugar-containing snacks, and five questions on spirits, wine and beer. Twenty nine of the remaining 35 questions recorded intake of potatoes, rice, pasta, meat and fish, and six varied items, such as salty snacks, coffee, tea and juice. The respondents indicated their average portion of a) potatoes/pasta/rice, b)vegetables and c) meat/ground meat/sausages/fish by comparison of four color photos illustrating four plates with increasing portion sizes of potato, vegetables and meat. For the other food items, we assumed gender- and age-standard portion sizes [ 20 ]. The reported frequencies of consumption were converted to number of intakes per day, and energy and nutrient intakes were calculated by multiplying these frequencies by portion size and energy or nutrient content from a food composition database from the Swedish National Food Administration. The energy and nutrient contents were calculated using the software MAT's (Rudans Lättdata, Sweden). Participants who had more than 10% missing answers were excluded. Single missing answers in sections where normally only one of several options is consumed frequently, such as type of milk, were not grounds for exclusion. Nutrient intake could not be estimated if portion size indication was missing. Statistical methods Data were analyzed separately for men and women using the SAS System for Windows (Release 8.02, SAS Institute, Cary, NC). Dietary variables were logarithmically transformed to improve normality. Univariate Pearson correlation coefficients were calculated between ln-transformed diet measures and SOC score. Differences between means for men and women were tested with Student's t- tests, and differences between more groups, i.e., age, BMI and education level groups, by ANOVA. P-values ≤ 0.05 were considered significant. Dietary intake and reporting thereof are highly influenced by gender, age, BMI, and educational level [ 21 ]. Therefore, mean intakes were standardized for age, BMI and educational level using general linear model (proc GLM). Significance of food item/nutrient predictors was based on ( i ) type I and ( ii ) type-III sum of square estimates, where ( i ) corresponds to a univariate regression, i.e., only the SOC score had entered the model, and ( ii ) corresponds to a multiple regression, i.e., all independent variables (SOC, gender, age, BMI, and educational level) were included in the model. To evaluate food selection pattern in individuals in relation to SOC score, PLS multivariate projection to latent structures was applied [ 22 ]. PLS is a method to relate two data matrices X and Y to each other by linear multivariate modeling. In contrast to traditional linear modeling, co-varying variables may be included. The PLS parameters carrying information about the x- and y- variables, e.g. R 2 , Q 2 and VIP, were generated as previously described [ 22 ]. The R 2 - and Q 2 -values give the capacity of the X matrix to explain (R 2 ) and predict (Q 2 ; equals cross-validated R 2 ) the variance of the Y matrix. The relative importance of each x-variable for the correlation structure among X and Y is given as a VIP-value ( V ariables of I mportance in the P rojection); VIP-values >1.0 are influential and VIP-values ≥ 1.5 highly influential. Results In total 2,446 men and 2,545 women answered the SOC- and food-frequency-questionnaires in such a way that the inclusion criteria were met. This corresponded to 86.6% of the 25–74 year old participants in the northern Sweden MONICA 1999 study. In the present cohort SOC scores varied from 23 to 91 points (Fig. 1 ). Means in quartile groups based on SOC distribution increased by approximately 9 points per quartile among both men and women (see Additional file 1 ). Mean scores neither differed significantly between men and women, nor between normal weight, overweight or obese individuals, nor between groups with different length of education. However, mean scores increased significantly with increasing age in both sexes (see Additional file 1 ). Univariate correlation analyses between SOC scores and daily intake of nutrients or servings per week in food groups revealed significant associations among women for intakes of energy (kCal/day), total and saturated fat (g/week), ascorbic acid (mg/week), sucrose (g/week), and servings/week of fruits, vegetables, cereals, and sweets (data not shown). Among men significant correlations were seen for daily intake of total and saturated fat (g/day), ascorbic acid (mg/day), fiber (g/day), and alcohol (g/day), and servings of fruits, vegetables, bread, bread and cereals, fish, and potatoes (data not shown). Except for ascorbic acid and cereal intake in women, and bread, bread and cereals, and potato intake in men, variations in intakes of these foods/nutrients were significantly explained by SOC quartile scores in linear GLM models (Type I SS; Tables 1 and 2 ). After standardization for age, BMI and education level SOC quartile scores still contributed in explaining intake of saturated fat, vegetables, sucrose and sweets in women and vegetables and alcohol in men (Type III SS; Tables 1 and 2 ). In accordance, mean (age, BMI and education standardized) intakes of vegetables were significantly higher in women and men (1.1 and 0.8 servings more per week, respectively) in the highest, compared to the lowest, quartiles (Tables 2 and 3 ). Women in the highest quartile also had higher intakes of fruits (+1.0 serving/week), but lower intake of energy (-365 kCal/week), total and saturated fat (-15 and -8 g/week, respectively), sucrose (-14 g/week), and sweets (-1.3 servings/week) (Table 1 ). PLS multivariate modeling of intakes of the 84 single food items and food aggregates simultaneously as an X-matrix on SOC-scores (Y-matrix) rendered models which supported low SOC to coincide with a less health promoting dietary preference and vice versa. Thus, in both men and women low SOC scores coincided with high intake of pizza, soft drinks, candies, sausages for main course, hamburgers, mashed potato, chips and other snacks, potato salad, French fries, and traditional broth soaked bread and potato dumplings (Table 3 , items in bold). In contrast, both men and women with high SOC scores were characterized by a high intake of rye crisp bread (whole meal), boiled potato, and vegetables (Table 3 , items in bold). In addition, some more gender specific characteristic food selections are seen in Table 3 . Discussion The present study shows that a high sense of coherence (SOC) is associated with health-promoting food choices, whereas low scores are associated with a less favorable food pattern. SOC score independently contributed in explaining variations in intakes of vegetables in both men and women, and in intakes of saturated fat, sucrose, sweets, and fruits among women, and alcohol intake among men. Correlations for other nutrient/food intakes seen in the initial univariate evaluations, disappeared when tested together with age, BMI and educational level in multiple linear GLM modeling. This indicates that some dietary preferences and food selections are linked to socioeconomic status, whereas others are more bound to the individual's sense of coherence, i.e. his/her personal way of grasping and handling life situations. Although the present results support earlier findings that associate high SOC scores with a more healthy diet [ 14 - 17 ], the interpretations should be considered in relation to methodological strengths and limitations of the study. The Northern Sweden MONICA study cohort of 1999 consisted of 6,000 participants, corresponding to 71.8 % of the invited subjects. Of these 86% could be included in the present evaluation. All subjects were primarily invited to the study by random from a continuously updated population ledger. Taken together, this allows for generalization of the results to a population level, at least to a population similar to that in northern Sweden. Within the MONICA project, extensive quality assessments of various methods and measures have been performed [ 23 ], and both the SOC and food frequency questionnaires were printed in an optically readable format to minimize errors due to faulty entry of data. Furthermore, the food frequency questionnaire has been found valid in a random sub-sample of representative adults [ 20 ], and the Cronbach alpha value [ 24 ] among SOC answers, was 0.81, which is in accordance with other studies [ 10 , 13 , 15 ], indicates reliability and internal consistency among the answers [ 19 ]. However, the cross-sectional design of the study, where the SOC scores were measured at the same time as the food preferences, precludes the possibility of interpreting the identified associations as causally related. The sense of coherence instrument has been constructed to measure an individual's capacity to cope in a salutogenic way [ 9 ]. In accordance with this the present study shows that the SOC level is associated with a "healthy" eating pattern. It has previously been shown that physically active individuals have higher SOC scores than inactive persons [ 25 , 26 ]. The hypothesis that SOC links to life style in a wider sense is therefore supported, even though we are aware that it cannot be ruled out that a healthy life style per se may influence SOC. The four cross-sectional MONICA screenings in 1986, 1990, 1994 and 1999 have demonstrated a distinct time trend in dietary intake in northern Sweden [ 27 ]. The present results therefore raise some questions. Are individuals with higher SOC more prone to change eating pattern? Are those changes more likely to follow dietary recommendations? Is a low SOC level a marker for a less healthy life style in general? Since all previous participants were re-called in 1999 answers may be searched for within the frame work of the MONICA project. Notably, both men and women in the highest, as compared with the lowest SOC score quartile, reported more food choices from the contemporary public message of a healthy life style [ 6 , 7 ]. In contrast, food selection in those in the lowest SOC-score quartile may render these individuals at a higher risk for various chronic diseases. Further knowledge of the influence of an individual's sense of coherence on his/her willingness and ability to change dietary habits may be of advantage when designing preventive programs. Conclusion Both men and women in the highest, as compared with the lowest, SOC score quartile reported more "healthy" food choices. Dietary habits for individuals in the lowest SOC quartile therefore may render a higher risk for various endemic diseases. Competing interests The author(s) declare that they have no competing interests Authors' contributions All participants have contributed to study design and manuscript preparation; IJ has been main author, UL has done data processing, BN has been responsible for SOC, BL for general behavior aspects, and BS is PI for the Northern Sweden MONICA study. Supplementary Material Additional File 1 SOC scores in gender, age, BMI, education level and SOC-quartile groups. Differences between means for men and women are tested with t-test, and among more groups, i.e. age groups, BMI education level, with ANOVA. ns for p > 0.05 Click here for file	/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC554973.xml
548382	Clash of kingdoms or why Drosophila larvae positively respond to fungal competitors	Background Competition with filamentous fungi has been demonstrated to be an important cause of mortality for the vast group of insects that depend on ephemeral resources (e.g. fruit, dung, carrion). Recent data suggest that the well-known aggregation of Drosophila larvae across decaying fruit yields a competitive advantage over mould, by which the larvae achieve a higher survival probability in larger groups compared with smaller ones. Feeding and locomotor behaviour of larger larval groups is assumed to cause disruption of fungal hyphae, leading to suppression of fungal growth, which in turn improves the chances of larval survival to the adult stage. Given the relationship between larval density, mould suppression and larval survival, the present study has tested whether fungal-infected food patches elicit communal foraging behaviour on mould-infected sites by which larvae might hamper mould growth more efficiently. Results Based on laboratory experiments in which Drosophila larvae were offered the choice between fungal-infected and uninfected food patches, larvae significantly aggregated on patches containing young fungal colonies. Grouping behaviour was also visible when larvae were offered only fungal-infected or only uninfected patches; however, larval aggregation was less strong under these conditions than in a heterogeneous environment (infected and uninfected patches). Conclusion Because filamentous fungi can be deadly competitors for insect larvae on ephemeral resources, social attraction of Drosophila larvae to fungal-infected sites leading to suppression of mould growth may reflect an adaptive behavioural response that increases insect larval fitness and can thus be discussed as an anti-competitor behaviour. These observations support the hypothesis that adverse environmental conditions operate in favour of social behaviour. In a search for the underlying mechanisms of communal behaviour in Drosophila , this study highlights the necessity of investigating the role of inter-kingdom competition as a potential driving force in the evolution of spatial behaviour in insects.	Background A common idea in animal ecology is that adverse or stressful environmental conditions facilitate the evolution of social behaviour [ 1 ]. The formation of groups across a huge number of animal taxa is thus considered to have broad implications for the benefit of individuals, including mate finding, the efficient location and use of resources, thermoregulation, energetic benefits and defence against natural enemies or competitors [ 2 , 3 ]. Basic proximate prerequisites for communal behaviour are cues indicating the location of conspecifics and the ability to receive and process information regarding these cues, which in turn induce inter-individual attraction [ 3 ]. Because the costs and benefits of communal behaviour typically vary with environmental conditions, the degree to which individuals are mutually attracted is regulated by signals indicating the presences of predators, food availability, etc. [ 4 ]. In the vast group of insects that depend on ephemeral resources, such as decaying plant tissues, dung and carrion, aggregation in the immature stages across resource patches is the result of the choice of a female to lay batches of eggs and/or to aggregate with conspecifics [ 5 - 8 ]. In studies of Drosophila as an ecological model system, one benefit that females flies seem to achieve by this spatial aggregation is that larval survival probability to the adult stage is highest at intermediate densities [ 9 , 10 ], indicating the existence of so-called Allee effects [ 11 ]. Competing filamentous fungi co-occurring with Drosophila larvae on the same patches have been demonstrated to cause high rates of mortality when larvae feed solitarily or in small groups, whereas larger groups are able to hamper mould growth [ 12 ] (Fig. 1 ), which in turn increases larval survival [ 9 , 13 , 14 ]. Although the mechanisms leading to mould suppression are not fully understood, physical damage of the fungal tissue from the feeding (shovelling food with the mouth hooks) and locomotor (crawling and digging) behaviour of the fly larvae [ 15 ] seems to be the major cause of the repression of mould growth [ 12 , 14 ]. Figure 1 The effect of larval density on mould growth. The effect of Drosophila larval density (a. one larva, b. 5 larvae, c. 10 larvae) on the growth of Aspergillus niger . Patches (2.5 cm diameter) contained standard Drosophila rearing medium. Photographs were taken 10 days after infection with fungal spores. Spores and fly larvae were simultaneously transferred to the patches. Whereas one larvae did not significantly hamper mould development (a), five and ten larvae caused a substantial reduction in fungal growth (b) or even entirely suppressed fungal development (c). (unpublished study) Given the relationship between spatial oviposition patterns, Allee effects and the suppression of mould, spatial aggregation in Drosophila can be interpreted as an adaptive behaviour against competing fungi on larval feeding sites in order to enhance offspring survival. These ecological interrelationships might set conditions for facilitating social behaviour in the fly larvae because, at the level of larval behaviour, a more efficient strategy that might control the rapid establishment of noxious fungi would be to exert physical stress directly on fungal colonies. Thus, larvae should display an assortative behaviour on the site on which fungi are growing, rather than moving randomly and independently of each other across a resource patch, by which the fungal tissues might only incidentally be destroyed. In the present study, I have provided groups of Drosophila melanogaster Meigen (Diptera, Drosophilidae) larvae with fungal-infected (2-day-old colonies of Aspergillus niger van Tieghem) and uninfected (control patches) (F-C treatment) food patches and examined whether the distribution of larvae across the patches is driven by fungal infection. In comparison with this naturally occurring heterogeneity in patch quality, I have also studied the distribution of fly larvae when they were offered only infected (F-F treatment) or uninfected (C-C treatment) food patches in order to test for the existence of grouping behaviour in two types of homogenous larval environment. If grouping is irrelevant under the given experimental setting, no deviation from the regular larval distribution across the food patches would be expected, i.e larvae should distribute themselves across patches in order to minimise larval competition for food [ 16 ]. Although Drosophila is a thoroughly studied model organism in foraging biology [ 17 , 18 ], knowledge about social interactions between the insect larvae is surprisingly limited. This is intriguing because drosophilids are also model organisms in spatial ecology in which Drosophila communities are characterised by strong intraspecific aggregation across patchily distributed substrates (e.g. decaying plant tissues) [ 19 - 21 ]. The lack of knowledge concerning social interactions among larvae and its possible role in competition with filamentous fungi have provided the specific impetus of the present study. Results Larval aggregation in the F-C treatment (Δ pl ) The proportion of larvae on fungal-infected patches minus the proportion on uninfected patches, Δ pl , was used as a measure of the way in which Drosophila larvae distributed themselves between the two types of food patches in the F-C treatment (see method section for details). The number of larvae in both food patches (LARVAE) and the experimental day (DAY) did not influence Δ pl (Table 1 ). The estimated intercept for Δ pl was significantly different from zero (GLM d.f. = 1, mean square = 4.475, F = 20.13, P < 0.0001, N = 35), the positive value for Δ pl (Fig. 2a ; intercept estimate : 0.3576 ± 0.0797, t = 4.49, P < 0.001) indicating the aggregation of larvae on fungal-infected sites (see method section). Table 1 Effect of LARVAE and DAY on larval aggregation in the F-C treatment. Analysis of variance for the effect of the number of larvae in both food patches (LARVAE) and experimental day (DAY) on Drosophila larval distribution between fungal-infected and uninfected food patches (F-C treatment). Explanatory variable d.f. Mean square F -value P LARVAE 1 0.0361 0.15 0.7042 DAY 3 0.0470 0.19 0.9018 Error 30 0.2462 Figure 2 Larval aggregation in the heterogeneous (F-C) and two types of homogeneous (F-F and C-C) larval environment. (a) Δ pl (where Δ pl = proportion of larvae from the fungal-infected patch – proportion of larvae from the uninfected patch) as a measure of larval aggregation in the F-C treatment (Δ pl = 0: no effect of fungal-infected patches on larval distribution behaviour; Δ pl > 0: aggregation of larvae on fungal infected patches; Δ pl < 0: larvae avoid fungal colonies). (b) \|Δ pl \| as a measure of the general tendency of Drosophila larvae to aggregate with conspecifics in the heterogeneous environment (F-C) and two types of homogeneous environment (F-F and C-C). Because larval aggregation in the F-C treatment was measured independently of the patch type (see Methods), \|Δ pl \| is larger than Δ pl (2a). (F: fungal-infected patches, C: uninfected control patches) Comparison of larval aggregation in the F-C, F-F and C-C treatment (\|Δ pl \|) \|Δ pl \|, the absolute value of Δ pl , was used as a measure of the general tendency of Drosophila larvae to aggregate with conspecifics in the heterogeneous environment (F-C) and the two types of homogenous environment (F-F or C-C). By using \|Δ pl \|, aggregation in the F-C treatment was quantified independently of whether a food patch was infected with fungi or not. With regard to all three larval environments, the estimated intercepts for \|Δ pl \| were significantly different from zero, and hence indicate larval aggregation (Table 2 ). Within each treatment LARVAE and DAY had no effect on \|Δ pl \| (Table 3 ). In comparison with the homogenous environments (F-F and C-C treatment), the F-C treatment induced stronger larval aggregation (Fig. 2b , Table 4 ). Moreover, there is a statistical trend of LARVAE influencing fly larval aggregation (Table 4 ). This was due to differences in LARVAE as a function of TREATMENT (GLM d.f. = 2, mean square = 0.0134, F = 3.34, P = 0.0393, N = 105). Significantly fewer larvae were found to be feeding in both patches in the C-C treatment (8.89 ± 1.64 SE) than in the F-C (9.43 ± 0.95 SE) or the F-F treatment (9.46 ± 0.61 SE). However, LARVAE within one type of environment had no effect on larval aggregation (Table 3 ). Table 2 The general tendency to aggregate with conspecifics (\|Δ pl \|) in the heterogeneous (F-C) and two types of homogeneous (F-F and C-C) larval environment. Test of the effect of intercept as the only explanatory variable for the general tendency of Drosophila larvae to aggregate with conspecifics (measured as \|Δ pl \|, see text for details) in three types of larval environment (F-C, F-F or C-C). Whereas \|Δ pl \| = 0 and no explanatory power of intercept would indicate a regular distribution of larvae across the food patches, \|Δ pl \| > 0 and a significant effect of intercept indicates larval aggregation in one of the experimental food patches (see also Fig. 2b). Note that, in contrast to Δ pl (Fig. 2a), \|Δ pl \| measures larval aggregation in the F-C treatment independently of whether a food patches was fungal-infected or not. For each type of larval environment an individual test was performed, with N = 35 for each treatment. Parameter estimate Larval environment Intercept ± SE t -value P F-C 0.51 ± 0.05 10.55 <0.0001 F-F 0.37 ± 0.05 8.13 <0.0001 C-C 0.35 ± 0.04 9.24 <0.0001 Table 3 The effect of LARVAE and DAY on the general tendency of Drosophila larval aggregation (\|Δ pl \|) under three environmental conditions. Analysis of variance for the effect of LARVAE and DAY on Drosophila larval distribution between food patches in three different larval environments (F-C, F-F or C-C). Larval environment Explanatory variable d.f. Mean square F -value P F-C LARVAE 1 0.2120 2.63 0.1154 DAY 3 0.0931 1.15 0.3433 Error 30 0.0807 F-F LARVAE 1 0.0682 1.03 0.3177 DAY 3 0.1132 1.71 0.1855 Error 30 0.0661 C-C LARVAE 1 0.0329 0.66 0.4236 DAY 3 0.2090 4.18 0.4998 Error 30 0.0400 Table 4 The effect of the larval environment on the general tendency of Drosophila larvae to aggregate with conspecifics (\|Δ pl \|). Mixed model analysis of variance for the tendency to aggregate with conspecific larvae in D. melanogaster in three types of larval environment (F-C, F-F or C-C). Larval aggregation was measured as \|Δ pl \|, the absolute value of Δ pl (see Methods). TREATMENT (F-C, F-F or C-C) and LARVAE were fixed main effects, whereas experimental day (DAY) was a random factor. DAY is nested within TREATMENT and was used as the error term in testing the effect of TREATMENT. Explanatory variable d.f. Mean square F -value P TREATMENT 2 0.4168 5.08 0.0302 LARVAE 1 0.2237 3.44 0.0670 TREATMENT (DAY) 9 0.0831 1.28 0.2603 Error 92 0.0651 Discussion On the background of ecological interactions between insects and filamentous fungi on ephemeral resources, the experiment presented in this study was designed to test for social attraction in Drosophila larvae, an attraction that I hypothesised to be advantageous when larvae are confronted with noxious moulds. The results demonstrate that the fly larvae significantly aggregated on food patches on which young fungal colonies were growing (F-C treatment, Fig. 2a ). Moreover, when provided with a homogeneous environment (F-F or C-C treatment), larvae displayed significant aggregation across the two food patches (Fig. 2b ). In comparison, however, aggregation was significantly enhanced when larvae had the choice between a mould-free and a mould-infected site (Fig. 2b ). Thus, the results suggest that grouping behaviour in Drosophila larvae involves both mutual attraction between group members [ 2 ] and the attraction of individuals to the same environmental stimulus [ 22 ], i.e. cues emitted by the fungi. Group formation in eusocial insects and those living in groups for part or most of their lives is often mediated by pheromones, e.g. cuticular hydrocarbons that induce attraction between individuals [ 22 , 23 ]. Chemical communication is also widespread in drosophilid behaviour, including those associated with spatial aggregation in adult flies [ 24 - 26 ]. Because several receptors on the cephalic lobe of Drosophila larvae have gustatory, mechanosensory and olfactory functions [ 27 ], both chemical and physical cues (e.g. substrate vibrations caused by larval movements) might be involved in mutual attraction. However, the mechanisms leading to grouping and group cohesion in Drosophila larvae are unknown. Interestingly, strong social attraction (communal digging) between Drosophila larvae is present in third-instar larvae, a behaviour regulated by a peptide neuromodulator ( Drosophila neuropeptide F, dNPF) [ 28 ]; this shows striking similarities to the correlation of social feeding and the expression of a neuropeptide Y receptor homologue (NRP-1) in Caenorhabditis elegans [ 29 ]. Moreover, neurons that detect aversive environmental stimuli have been demonstrated to induce social feeding in C. elegans [ 30 ], thus providing support for the proposed relationship between environmental stress and social behaviour [ 31 ]. In contrast to C. elegans , the intimate communal digging behaviour in old third-instar Drosophila larvae does not occur in the context of food foraging behaviour but is part of the post-feeding phase prior to pupation [ 28 ]. Whereas downregulation of dNPF expression coincides with social behaviour in old and non-feeding Drosophila larvae, higher levels of dNPF expression in younger larvae seem to suppress strong larval aggregation and communal digging behaviour [ 28 ]. Therefore, it remains to be seen whether similar neural regulatory mechanisms are involved in earlier developmental stages of Drosophila larvae with respect to social affinity related to foraging for food and attraction to fungal competitors. With regard to the proximate causes of attraction to mould-infected sites, fungal-borne volatiles such as CO 2 or ethylene [ 32 ] might be perceived by fly larvae and might guide them to young mould colonies. A general reason for the formation of social groups seems to be an adaptive response to stressful environmental conditions [ 1 , 31 ]. As outlined in the introduction, filamentous fungi co-occurring with drosophilids on larval feeding sites can impede fly larval development; indeed, larval aggregations have been shown successfully to suppress mould growth [ 12 ] (Fig. 1 ). Consequently, the presence of fungi may indicate stressful ecological conditions that initiate attraction towards fungal patches and enhance the mutual attraction of Drosophila larvae (Fig. 2 ). In connection with the benefits that accrue from mould suppression, the present study demonstrates that the larval-driven inhibition of fungal development is not a mere by-product attributable to the maternal decision to aggregate eggs across patches but is the consequence of a positive response of individuals to conspecifics. Because adult density-dependent oviposition choices influence a larva's food quality and its susceptibility to natural enemies [ 33 ] or abiotic stress, as well as its probability of coming into contact with intra- and interspecific competitors, the present study demonstrates the possibility of adaptive behavioural relationships between the well-known adult gregariousness in Drosophila and communal behaviour in the immature stages. Further investigating this kind of behavioural adult-offspring correlations would strongly contribute to our understanding of the evolutionary costs and benefits of spatial aggregation in insect communities [ 34 ]. Conclusion The study presented here demonstrates that fungal-infected food patches (1) attract first-instar Drosophila larvae and (2) enhance group foraging behaviour. Because the larval-driven reduction in mould growth [ 12 ] has been shown to improve the chance of larval survival to the adult stage [ 9 , 13 , 14 ], social attraction to fungal-infected sites may reflect an adaptive behavioural response. In connection with the maternal behaviour of aggregating eggs across substrate patches, the condition-dependent mutual attraction of larvae can be discussed as a communal defence behaviour against competing mould. Given that filamentous fungi seriously deteriorate the developmental conditions for insect larvae, mould may constitute one important ecological factor that has, at least in the larval stages, facilitated social attraction in Drosophila . Thus, this study highlights the largely unappreciated role of inter-kingdom competition as a potentially important driving force in the evolution of insect behavioural traits. In general, the group formation of Drosophila larvae in response to well-defined ecological conditions might be an interesting model system for the study of proximate and ultimate aspects of social biology. Methods Experimental set-up I experimentally analysed the grouping behaviour of Drosophila larvae using a D. melanogaster strain that originated from wild animals caught in 2003 near Kiel, northern Germany (approx. 54° N, 10° E). Flies had been reared for 18 generations on standard Drosophila medium (30 gram corn meal, 30 gram sugar, 30 gram brewer's yeast extract (Leiber, Germany), and Nipagin) under constant environmental conditions (photoperiod of 16 hours and a temperature of 22°C). In order to obtain first instar larvae, a population of approx. 300 individuals (five to seven days old) were offered a 10 cm Petri dish containing a hard Agar medium (22 gram Agar, 90 cm 3 sugar beet syrup and 9.5 cm 3 Nipagin (10% in 95% ethanol) per 500 cm 3 water), on which they were allowed to oviposit for a period of 16 to 18 hours, including a period of darkness. Subsequently, flies were removed and the eggs were incubated at 22°C over a 16-hour photoperiod. After 24 hours, almost all larvae had hatched from the eggs and were then isolated from the medium by washing them off the Agar plate onto fine-meshed gauze with water. These larvae were used in the experiments. I used the same medium as for fly rearing (without adding the antifungal agent Nipagin) to simulate an uninfected or a fungal-infected larval environment as follows: aliquots of 3.5 cm 3 hot medium were transferred to each of two small pots (10 mm diameter, 5 mm high) that were glued to the bottom of a Petri dish (45 mm diameter, 13 mm high). Within one Petri dish, the larval food patches were placed at a distance of approx. 10 mm and surrounded by an Agar layer (5 mm high), so that the surfaces of the Agar and the food patches were at the same level. After the food patches had cooled down, one patch was provided with 1 μl of water containing approx. 800 conidiospores of the fungus A. niger (F treatment) and, as a control, the second patch was provided with spore-free water (C treatment). In addition to this F-C treatment, I simultaneously prepared arenas in which both patches were infected with spores (F-F) or both patches remained uninfected (C-C). The arenas were immediately sealed with lids and incubated under the aforementioned conditions. Two days later, the fungal spores had germinated and tiny translucent hyphal colonies were visible. A group of ten Drosophila larvae were transferred, with a fine brush, to each arena at a distance of approx. 10 mm from each of the two food patches. Subsequently, the arenas were sealed with lids and stored at 22°C in an illuminated incubator. In order to avoid any systematic effects on larval distribution behaviour that might be caused by the position of the light tubes, the arenas of all three treatments were randomly arranged in the incubator. After six hours, the number of larvae in each food patch was recorded. Preliminary experiments had shown that this time period was sufficient to obtain final larval distribution patterns across the two patches, which remained nearly constant until the next day. Five to ten replicates for each treatment were simultaneously prepared at four different days. Statistical analysis Based on the number of larvae that were found to be feeding in both patches, I calculated the proportion of larvae in each patch. Larvae that were not found on any of the food patches were ignored. However, I tested for the effect of the number of larvae in both patches (LARVAE) on the degree to which larvae aggregated across the food patches (see below). To obtain a measure of the degree of larval aggregation across the two food patches in the F-C treatment, the proportion of larvae on the fungal-infected patch in each arena was reduced by the proportion of larvae on the uninfected patch, yielding Δ pl . Δ pl = 0 would indicate no effect of the presence of fungal-infected and uninfected food sites on larval distribution patterns. Whereas Δ pl > 0 would indicate aggregation on fungal-infected sites and thus larval attraction to fungal colonies, Δ pl < 0 is expected if larvae avoid fungal colonies and aggregate on uninfected patches. Subsequently, I used the absolute values of Δ pl , \|Δ pl \|, that were obtained in all three types of treatment (F-C, F-F and C-C) in order to compare the tendency to aggregate with conspecific larvae in the heterogeneous larval environment (F-C) with larval aggregation in two types of homogeneous environment (F-F or C-C). Note that, because the absolute value of Δ pl can only be equal to or larger than zero, \|Δ pl \| measures larval aggregation in the F-C treatment independently of whether a food patch was fungal-infected or not. I applied the GLM procedure provided by SAS version 8.2 to test if Drosophila larvae aggregated on fungal infected food patches, i.e. if Δ pl is significantly larger than 0 (see above). For this only the intercept was tested as an effect in the statistical model [ 35 ]. The result of the parameter estimate for the intercept are given. Before this test, I verified that LARVAE and experimental DAY did not affect Δ pl (Table 1 ), which justifies the removal of these variables from the full model (backward elimination of non-significant variables) [ 36 ]. The same procedure was applied to test for the general tendency of larval aggregation (measured as \|Δ pl\| ) under different environmental conditions (see Table 2 to 4 ). To analyse the effect of TREATMENT (F-C, F-F or C-C) and LARVAE on the general propensity of Drosophila larvae to aggregate across the experimental food patches (\|Δ pl \|), I used the aforementioned GLM procedure with a RANDOM statement to account for possible effects of experimental DAY on larval distribution patterns. In this model, TREATMENT and LARVAE were fixed main effects. Since five to ten replicates for each treatment were prepared at four different days, DAY was considered as a categorical random factor. DAY is nested within TREATMENT and was used as the error term in testing for the effect of TREATMENT [ 37 ]. The results of the tests of hypotheses for mixed model analysis of variance are shown in Table 4 .	/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC548382.xml
520599	Molecular Signatures of Proliferation and Quiescence in Hematopoietic Stem Cells	Stem cells resident in adult tissues are principally quiescent, yet harbor enormous capacity for proliferation to achieve self renewal and to replenish their tissue constituents. Although a single hematopoietic stem cell (HSC) can generate sufficient primitive progeny to repopulate many recipients, little is known about the molecular mechanisms that maintain their potency or regulate their self renewal. Here we have examined the gene expression changes that occur over a time course when HSCs are induced to proliferate and return to quiescence in vivo. These data were compared to data representing differences between naturally proliferating fetal HSCs and their quiescent adult counterparts. Bioinformatic strategies were used to group time-ordered gene expression profiles generated from microarrays into signatures of quiescent and dividing stem cells. A novel method for calculating statistically significant enrichments in Gene Ontology groupings for our gene lists revealed elemental subgroups within the signatures that underlie HSC behavior, and allowed us to build a molecular model of the HSC activation cycle. Initially, quiescent HSCs evince a state of readiness. The proliferative signal induces a preparative state, which is followed by active proliferation divisible into early and late phases. Re-induction of quiescence involves changes in migratory molecule expression, prior to reestablishment of homeostasis. We also identified two genes that increase in both gene and protein expression during activation, and potentially represent new markers for proliferating stem cells. These data will be of use in attempts to recapitulate the HSC self renewal process for therapeutic expansion of stem cells, and our model may correlate with acquisition of self renewal characteristics by cancer stem cells.	Introduction Hematopoietic stem cells (HSCs) are the best-described adult stem cell population at a phenotypic and functional level. Recent attempts have been made to characterize their molecular regulation by comparing their gene expression profiles with those of other stem cell populations ( Ivanova et al. 2002 ; Ramalho-Santos et al. 2002 ; Fortunel et al. 2003 ). These analyses of normal steady-state stem cells revealed so-called stem cell signatures, but the overlap of genes that universally defined “stemness” was extremely limited ( Fortunel et al. 2003 ). Here, we have focused on HSCs alone in order to systematically examine one process, that of HSC self renewal, comprising a cycle of quiescence, proliferation, and reinduction of a dormant state. In a normal adult, HSCs reside in the bone marrow, where they are relatively inactive. Long-term HSCs divide infrequently to produce more proliferative short-term HSCs, which in turn generate the lineage-committed progenitors that manufacture the billions of differentiated hematopoietic cells that daily enter the peripheral blood. One hallmark of HSCs is their ability to rapidly proliferate in response to stressors such as myelosuppressive chemotherapy or bone marrow transplantation in order to quickly generate work-horse progenitors as well as additional stem cells, which then return to quiescence ( Dixon and Rosendaal 1981 ). While this expansion of HSCs occurs naturally in vivo, there is as yet little understanding of the genes that control this process. A full appreciation of the molecular regulation of stem cell self renewal could illuminate the development of cancers ( Sherr 1996 ) as well as potentially inform strategies for in vitro stem cell expansion, which would have enormous clinical advantages. Thus, we sought to understand the molecular mechanisms by which HSCs respond to an activating trigger, initiate a program of cell division, and resume quiescence by suppression of cell division. Our approach was to examine the transcriptional profiles of purified adult HSCs throughout a time course of induced proliferation, and compare the gene expression in these cells to that of naturally dividing fetal liver HSCs (FL-HSCs). Normal adult HSCs are largely nondividing, with around 1%–3% in cycle and approximately 90% in G0 ( Morrison and Weissman 1994 ; Goodell et al. 1996 ; Bradford et al. 1997 ; Cheshier et al. 1999 ). A single injection of the pyrimidine analog 5-fluorouracil (5FU) kills cycling hematopoietic cells, bringing the spared quiescent HSCs into cycle to repopulate the depleted bone marrow ( Van Zant 1984 ; Harrison and Lerner 1991 ; Randall and Weissman 1997 ). HSC proliferation proceeds in a time-dependent manner, peaking 5 to 6 d after treatment, with approximately 20% of HSCs in cycle, before returning to normal around day 10 ( Figures 1 A and S1 ; Randall and Weissman 1997 ). Changes in the cell surface profile concomitant with cell cycle activation have been observed. The receptor tyrosine kinase c-Kit, normally expressed at high levels in quiescent HSCs, is down-regulated after 5FU treatment ( Randall and Weissman 1997 ). Conversely, the markers Mac1 and AA4.1, absent on normal HSCs, are expressed at low levels after 5FU treatment ( Szilvassy and Cory 1993 ; Randall and Weissman 1997 ). Figure 1 P-Sig and Q-Sig Defined by Gene Expression Levels in HSCs in Different Stages of Cell Cycle (A) Graphic depicting the changes in bone marrow cellularity and number of HSCs in cell cycle following 5FU treatment (adapted from Harrison and Lerner 1991 ; Randall and Weissman 1997 ) . (B) Schematic of 5FU-HSC time course analysis. The genes that change over the time course can be split into two groups based on the day of maximum expression (TOM). (C) Schematic of pair-wise comparison between quiescent adult HSCs and FL-HSCs, showing groups of genes either up-regulated in the quiescent adult cells or up-regulated in the cycling FL-HSCs. (D) Genes that were both up-regulated in FL-HSCs and were in the proliferation group composed the P-sig. The P-sig shows 94% overlap with the group of genes that were up-regulated in FL-HSCs and changed over the time course. (E) Genes that were both in the quiescence group and up-regulated in adult HSCs were termed the Q-sig. The Q-sig overlaps 96% with the set of genes that were up-regulated in adult HSCs and changed over the time course. (F) Overlap of the ST-HSC signature with P-sig revealed 73% in common, defining the common P-sig. (G) Overlap of the LT-HSC signature with Q-sig revealed 58% in common and defined the common Q-sig. This figure is interactive online, and provides contextual access to Tables S1–S11 . Use your mouse to highight animated areas of the graphic. Click on these areas to link to related files. During the latter part of mammalian embryonic development, HSCs reside in the fetal liver, where they undergo a massive expansion prior to entering the bone marrow. Approximately 30% of murine FL-HSCs are in cycle ( Morrison et al. 1995 ), and similar to 5FU-activated HSCs (5FU-HSCs), they express AA4.1 and Mac1 ( Jordan et al. 1990 ; Morrison et al. 1995 ). Given the similarities between 5FU-activated HSCs and FL-HSCs, we hypothesized that they would share similar gene expression profiles vis-à-vis proliferation and that simultaneous comparison of FL-HSCs, adult quiescent HSCs, and 5FU-HSCs would define genes specifically involved with HSC proliferation. Indeed, we defined both proliferation and quiescence signatures for HSCs, validated these groupings using Gene Ontology (GO) classifications, and created a model of the HSC proliferation cycle. Results Experimental Design Our overall approach was to isolate highly purified HSCs in the states described above, obtain their gene expression profiles using Affymetrix microarrays, and apply statistical and bioinformatics methods to facilitate comparisons among the samples. To construct a profile of the time-dependent induction of HSC proliferation, 5FU-HSCs were isolated at days 0, 1, 2, 3, 6, 10, and 30 after treatment. Adult quiescent HSCs and 5FU-HSCs were isolated according to Hoechst 33342 efflux, termed the side population (SP) and Sca1 + characteristics ( Goodell et al. 1996 ) ( Table 1 ; Figure S2 A). Further analysis of these populations revealed them to be highly homogeneous with more than 97% having Sca1 + /Lineage − characteristics ( Figure S3 ). Transplantation into lethally irradiated hosts, performed for both quiescent and 5FU-treated SP/Sca1 + cells, confirmed their stem cell activity (data not shown). FL-HSCs were isolated by FACS for AA4.1 + , c-Kit + , Sca1 + , and Lineage − characteristics from embryos 13.5–14.5 d postcoitus ( Jordan et al. 1990 ) ( Table 1 ; Figure S2 B.) RNA probes were prepared from HSCs using two rounds of in vitro transcription and applied to Affymetrix MGU74Av2 microarrays. Hybridization, scanning, and production of raw data files were performed according to standard protocols. To correct raw intensity values for systemic variables such as fragmentation efficiency, hybridization conditions, and scanner effects, microarrays were normalized before intensity values were converted to gene expression measures. Normalization and model-based expression measurements were performed with GeneChip Robust Multichip Analysis ( Wu et al. 2003 ), which is more precise and accurate in estimating fold changes than Affymetrix MAS 5.0 and the recently published Robust Multichip Analysis method ( Irizarry et al. 2003 ), and is available as part of the open-source Bioconductor project ( http://www.bioconductor.org ). Further statistical analysis was performed in R ( http://www.r-project.org ). Quality control was performed both pre- and postnormalization. Briefly, chips were inspected for spatial defects, intensity outliers, and amplification bias. After screening, the two chips representing biological replicates with the highest correlation (R 2 = 0.97–0.99, average = 0.98) in each group or time point were selected for further analysis. Raw data and normalized expression data are available for download from Gene Expression Omnibus ( http://ncbi.nlm.nih.gov/geo ) or http://franklin.imgen.bcm.tmc.edu/SCGAP/downloads/SPTimecourse . Normalized expression data along with all filtering criteria used to obtain our gene lists are available in Table S46 . A gene-by-gene query tool is available at http://franklin.imgen.bcm.tmc.edu/PLoS . Table 1 Comparison of Phenotypic and Functional Characteristics of HSC Populations a Goodell et al. (1996) b Randall and Weissman (1997) c Szilvassy and Cory (1993) d Jordan et al. (1990) e Morrison and Weissman (1994) f Morrison et al. (1995) g Uchida (2004) d.p.c., days postcoitus, N.D., not determined Time of Maximum Grouping Reveals Strong Time Ordering to Expression Data We began our analysis of the 5FU time course by identifying genes that varied over time. This was accomplished by fitting smooth curves to the expression profiles using regression analysis with time as a continuous variable. ANOVA on these profiles revealed 1,488 genes that showed a significant change over the time course ( p < 0.05). Principle component analysis revealed that the time course data consisted of two major groups: genes that up-regulated and genes that down-regulated over the time course (data not shown). We further explored the expression data with unsupervised (k-means) clustering and observed that when the number of predefined groups was low (2–3), only the overall pattern of up- or down-regulation was discernable; however, as we increased the number of groups (4–8), more complex patterns with peaks early or late in the time course were visible (data not shown). Since the 5FU treatment consisted of a single dose administered at time zero, we speculated that the downstream effects of 5FU treatment would be represented by groups of genes whose gene expression profiles showed time-ordered peaks propagating through the time course. The expression profile of groups created by k-means clustering supported this hypothesis. Therefore, to more directly delineate these peaking subsets, we sorted the genes into groups by their time of maximum expression (TOM). Strikingly, these groups had two predominant patterns over the time course: one group up-regulated with 5FU treatment with a TOM at day 2, 3, or 6, and one group down-regulated, exhibiting TOM at day 0, 1, 10, or 30 ( Figure 1 ). By correlating these patterns to HSC cell cycle status after 5FU treatment ( Figure 1 A), we assigned the up-regulated genes to the “proliferation” group (680 genes) and the down-regulated genes to the “quiescence” group (808 genes) ( Figure 1 B). To validate these time-dependent expression-pattern-based gene groupings, we compared our quiescence and proliferation groups to the genes differentially expressed between quiescent adult HSCs and FL-HSCs. The latter were identified in a pair-wise comparison between adult HSCs and FL-HSCs that revealed 1,772 genes that were at least 2-fold differentially expressed ( Figure 1 C). Since FL-HSCs are in cycle, as are 5FU-HSCs, a list of genes expressed in common between the time-course-defined proliferation group and those up-regulated in FL-HSCs should specifically contain genes involved in HSC proliferation, eliminating genes involved in interacting with their very different source environments. We designated this list of 338 genes our “proliferation signature” (P-sig; Figure 1 D). Likewise, the 298 genes in common between the time-course-defined quiescence group and those up-regulated in adult HSCs relative to FL-HSCs defined our “quiescence signature” (Q-sig; Figure 1 E). In Figure 2 B and 2 D, each gene within the P-sig and Q-sig is represented by a single line, and its relative expression along the time course is represented by the intensity of the colors on the heat map. Genes discussed later in the text are highlighted. To examine whether similar signatures could be generated without the TOM groupings (which could potentially introduce a bias), we examined the list of genes overlapping between the set of those up-regulated in FL-HSCs and the entire set of genes that change during the time course (see Figure 1 D). A striking 94% of the P-sig overlaps with these genes. Similarly, 96% of genes in the Q-sig overlap with the set of genes that are up-regulated in adult HSCs and change over the entire time course (see Figure 1 E). In other words, overlapping the pair-wise comparison with our expression-pattern-based groups, i.e., TOM groupings, identified essentially the same genes as did overlapping the time course with quiescent adult HSC and FL-HSC data, thus correlating, at the gene level, the TOM groupings to populations with known biological differences. Figure 2 P-Sig and Q-Sig Show Patterns of Activation and Down-Regulation with Respect to Cell Cycle Status (A) Averaged pattern of P-sig gene expression over the 5FU time course plotted in solid lines, with the contributing TOM subgroups plotted in dashed lines. (B) Heat map of each gene in P-sig over the 5FU time course showing TOM subgroups in brackets. (C) Averaged pattern of Q-sig gene expression over the 5FU time course plotted in solid lines, with the contributing TOM subgroups plotted in dashed lines. (D) Heat map of each gene in Q-sig over the 5FU time course showing TOM subgroups in brackets. For both heat maps, relative expression levels are displayed according to color intensity, blue (lowest) to yellow (highest). This figure is interactive online, and provides contextual access to Tables S12–S18 . Use your mouse to highight animated areas of the graphic. Click on these areas to link to related files. We then plotted the average pattern for the P-sig and Q-sig and examined their component TOM groups (see Figure 2 ). The patterns of genes in the TOM subgroups of the P-sig were very similar, with an overall off-on-off pattern that corresponded to the number of HSCs in cycle after 5FU treatment (see Figures 1 A, 2 A, and 2 B). Although mutually exclusive gene lists, TOM 0 and 30 were almost identical in pattern and were highly similar at the functional level (see below). Genes in TOM 1 and 10 shared the overall pattern of down-regulation with the Q-sig, but showed early and late peaks, respectively, the significance of which is discussed below. Overall we found the individual TOM groups to be highly coherent with a high degree of correlation between the individual genes and the mean profile of each group ( Table S47 ). Q-sig and P-sig Overlap with Published Data to Give “Common” Signature Encouraged by these results, we performed a parallel analysis on a raw dataset from Akashi et al. (2003) , who compared the transcriptional profiles of adult long-term HSCs (LT-HSCs) and short-term HSCs (ST-HSCs). Although isolated by different methods, the Rho low KTSL cells isolated by Akashi et al. and our quiescent adult HSCs are functionally equivalent ( Wolf et al. 1993 ; Goodell et al. 1996 ). ST-HSCs have the ability, as do LT-HSCs, to contribute to all lineages of the hematopoietic system, but are not able to maintain long-term engraftment in irradiated hosts. They are also more in cycle than LT-HSCs and express low levels of Mac1 ( Table 1 ) ( Morrison and Weissman 1994 ; Cheshier et al. 1999 ). We therefore suspected that the genes 2-fold differentially expressed between LT-HSCs and ST-HSCs, approximately 300 and 600 genes, respectively, would be enriched for quiescence and proliferation genes, respectively. When we compared these lists with the list of genes changing after 5FU treatment, we observed that almost all the genes in common between LT-HSC and time course lists were in the quiescence group list. Similarly, most of the genes in common between the ST-HSC and the time course lists were in the proliferation group list. This confirmed that many of the gene expression changes that occur between LT-HSCs and ST-HSCs are the same changes that occur after activation of HSC with 5FU, and we designated these list intersections as the LT-HSC signature and ST-HSC signature, respectively. A natural question was whether the Q-sig and P-sig described above would have any overlap with the LT-HSC signature and ST-HSC signature groups. Remarkably, 58% of the genes were in common between LT-HSC signature and the Q-sig, and 73% of the genes were in common between ST-HSC signature and the P-sig. We named these highly selected lists (53 and 118 genes, respectively) the “common quiescence signature” (cQ-sig) and “common proliferation signature” (cP-sig) ( Figure 1 F and 1 G). As we show below, these “common” signatures derived from the three-way intersection of 5FU-HSC data, adult-HSC-versus-FL-HSC data, and LT-HSC-versus-ST-HSC data were highly enriched for genes related to HSC proliferation. Novel Uses of Gene Ontologies Allowed Functional Validation of Gene Groupings To investigate the biological significance of the groupings described above, we developed novel methods for utilizing the GO annotations ( Ashburner et al. 2000 ) ( http://www.geneontology.org ) to analyze the content of gene lists. The GO is a controlled vocabulary that describes gene functions in their cellular context and is arranged in a quasi-hierarchical structure from more general to more specific. Since the vocabulary of annotations is fixed, it allows for functional comparisons of mutually exclusive gene lists, such at the TOM groups. We began by mapping each gene in the lists being analyzed to the GO tree structure. This allowed us to count the number of times each gene hit at or below any particular node in the GO structure. Once the lists were mapped, we were able (a) to calculate a measure of similarity (distance) between the lists using the distributions of each list across the various levels of the GO tree and (b) to calculate the enrichment of the various GO categories in each list ( Figure 3 A– 3 C). Figure 3 GO Analysis and Chromosomal Clustering (A) Dendrogram of gene lists clustered solely according to their similarity in GO content. (B) Bar graph showing enrichments of selected GO groups in the Q-sig and P-sig. Fold changes are relative to whole microarray ( p < 0.05). Asterisk marks groups in which no genes were found (complete depletion). (C) Percentage of genes within each list that are in the GO groups “cell cycle” or “cell–cell adhesion.” (D) Distribution of hits within Q-sig and P-sig on each chromosome normalized for number of expected hits for whole microarray. Pound sign denotes significant differences between Q-sig and P-sig ( p < 0.05). This figure is interactive online, and provides contextual access to Tables S19–S45 . Use your mouse to highight animated areas of the graphic. Click on these areas to link to related files. We clustered the gene lists based on this distance metric ( Figure 3 A). As can be seen, GO-based clustering recapitulated the previous expression-pattern-based groupings: TOM days 0, 1, 10, and 30 clustered with the list of genes up-regulated in adult HSCs; and TOM days 3 and 6 clustered with those genes up-regulated in FL-HSCs. We calculated a probability of 0.003 that we could arrive at the grouping pattern shown by chance. Importantly, this indicated that the content of these clusters, as defined by their biological process using GO, was highly similar despite the nonoverlapping nature of the TOM groups. Although recapitulating the expression-pattern-based groupings, our GO-based clustering also revealed that TOM 1 has a unique signature amongst the quiescence cluster, suggesting a distinctive role for the genes in this group in governing HSC quiescence ( Figure 3 A). Our strategy for mapping gene lists to the GO structure also allowed us to calculate statistically significant enrichments of particular GO categories within our gene lists. We achieved this by mapping the whole microarray (approximately 12,000 genes) onto the GO structure and then calculating fold enrichments for each GO category in our lists relative to the microarray. We expected to find differences between the Q-sig and P-sig in the frequencies of antiproliferative and proproliferative genes, and verification of this served as proof-of-principle for our experimental design. Indeed, we found the GO category “regulation of cell cycle” (containing genes like the antiproliferative genes p21 [cyclin-dependent kinase inhibitor 1A] and GADD45β [Growth arrest and DNA-damage-inducible 45, beta] ) to be 2.1-fold increased in the Q-sig over the total array ( Figure 3 B). Moreover, the category “DNA replication” was about 5-fold greater in the P-sig, while this category was absent in Q-sig ( Figure 3 B). Intriguingly, the GO group “defense response,” containing many of the H2 genes of the MHC class I family, was slightly enriched in the Q-sig, but was depleted by over 5-fold in the P-sig ( Figure 3 B). Signal transduction molecules such as those in the GO groups “protein kinase cascade” were enriched 4.3-fold in the Q-sig ( Figure 3 B). The GO group “ATP synthesis coupled electron transport” was enriched almost 21-fold in the P-sig, which correlates with the high energy requirements of cell division ( Figure 3 B). As discussed above, our results and the data from Akashi et al. (2003) have remarkable overlap at the gene level. Using the common signature lists, we observed further refinements in key GO categories. For example, “cell cycle” genes were less than 4% of all genes on the chip, yet they represented 21% of the genes in the common P-sig ( Figure 3 C). Progressive enrichment in “cell–cell adhesion” was also observed ( Figure 3 C). Although almost 19% of the genes in our “common” signatures have no previously defined biological process, given the remarkable enrichment of proliferation-related genes in our common signatures, we can infer that they also may be involved in HSC proliferation. TOM Analysis Uncovered Orderly Progression of HSC Activation We further utilized the GO-based analysis of the TOM groups within the Q-sig and P-sig to gain insight into the biological activities of HSCs at these time points. Because of the high similarity of TOM 0 and 30 in both expression pattern and GO categorization, we treated them as a single group. “Regulation of transcription” was enriched 1.5-fold in TOM 0 and 30 and comprised 16 genes, including several key transcriptional regulators of cell cycle such as the oncogenes c-fos and c-maf, as well as the global transcriptional repressor histone deacetylase 5 . The GO categories “regulation of cell cycle,” “cell–cell adhesion,” and “defense response” were specifically enriched in TOM 1 (approximately 4-fold each). Many genes in these groups are negative regulators of cell cycle, such as p21, Tob 1/APRO6 (Transducer of ErbB2.1-1), Btg3/APRO4 (B-cell translocation gene 3), cyclin G1, GADD45β, and melanoma antigen, family D, 1 . Prior experiments have shown a decrease in the number of HSCs in cycle during the first day after 5FU treatment as compared to untreated HSCs (see Figure 1 A; Randall and Weissman 1997 ). We therefore concluded that many of the genes in TOM 1 are responsible for this momentary pause in cell cycle, and this explained why these genes were initially up-regulated and then sharply down-regulated as rapid HSC proliferation began (see Figure 1 A and 2 C). In the P-sig, TOM 3 and TOM 6 showed astonishingly different GO contents despite their similar expression patterns (see Figure 2 A). Genes in the GO category “cell cycle” identified in the P-sig are concentrated in TOM 3. Specifically, genes in both “DNA replication” and “M phase” were enriched about 18-fold and 10-fold, respectively, indicating a preparation for cell division. TOM 6 was enriched almost 3-fold in genes involved with biosynthesis of many essential cellular components, such as ATP (8.8-fold), nucleotides (5.6-fold), and proteins (2.4-fold). These data suggest early and late phases of proliferation, represented by the genes in TOM 3 and TOM 6, respectively. As discussed above, by day 10 after 5FU treatment the number of HSCs in cycle is reduced to near pretreatment levels (see Figure 1 A). Although the signals responsible for restoring quiescence remain elusive, we believe that this process may be mediated by JAK/STAT and other signaling pathways. Overall, the GO category “signal transduction” showed approximately 2-fold enrichment in the TOM day 10 list. SOCS3 (Suppressor of cytokine signaling 3), whose product suppresses responses to growth factors in part by inhibiting JAK/STAT signaling, was most highly expressed at day 10, along with STAT3 and STAT6 . JAK/STAT signaling has been implicated in regulation of proliferation and differentiation of various hematopoietic cell types. Chromosomes 2, 7, 12, and 17 Contain HSC Proliferation Control Regions Our expression data can be combined with data from the mouse genome projects to correlate gene expression changes observed after 5FU treatment with higher order genome-wide regulation. For example, we analyzed the contents of Q-sig and P-sig for clustering on particular chromosomes. Four chromosomes exhibited significant enrichment between the two signatures: Chromosomes 12 and 17 were enriched in the Q-sig, and Chromosomes 2 and 7 were enriched in the P-sig (see Figure 3 D). Earlier work identified quantitative trait loci (QTL) on Chromosomes 17 and 7 associated with the control of HSC frequency and proliferation of hematopoietic progenitors, respectively ( Phillips et al. 1992 ; Geiger et al. 2001 ). p21, a prototypic member of the Q-sig, was specifically found within a QTL on Chromosome 17 associated with regulating HSC frequency. This region is syntenic with human Chromosome 6p21, a known hot spot for translocations linked to leukemias and lymphomas ( Huret et al. 1986 ; Johansson et al. 2002 ). Microarray Gene Expression Changes Reflect Changes in Protein Expression and HSC Behavior In order to determine whether some of the observed gene expression changes were accompanied by measurable differences in protein expression, we identified two genes whose expression changed over time and whose product could be tracked using flow cytometry. Gene expression of Sca1, a known marker of HSCs, showed significant increase after 5FU treatment despite having a high starting level ( Figure 4 A). Flow cytometric analysis showed that Sca1 antigen expression was also distinctly higher after 5FU ( Figure 4 B). Sca1 -null mice have a defect in HSC self renewal that has been interpreted as a loss of proliferative capacity ( Ito et al. 2003 ). Our data support this finding since maximal expression of Sca1 both at the gene expression and protein level was at day 6/7 post 5FU treatment. We also analyzed CD48, a cell adhesion molecule previously associated with T-cell activation and proliferation ( Kato et al. 1992 ; Chavin et al. 1994 ; Gonzalez-Cabrero et al. 1999 ), which peaked in gene expression 6 d after 5FU treatment ( Figure 4 A). By flow cytometry, CD48 antigen was detected on quiescent HSCs, but exhibited a substantially higher level of expression at the height of HSC proliferation ( Figure 4 B). To determine whether high levels of CD48 antigen on HSCs coordinated with proliferation in a similar fashion as on T-cells, we performed cell cycle analysis of CD48 + and CD48 − HSCs. Further characterization of CD48 + HSCs 6 d post 5FU revealed a greater than 3-fold enrichment in the number of cells in cycle over CD48 − HSCs ( Figure 4 C). This finding is the first report of a marker that enriches for cycling HSCs. Figure 4 Gene Expression Profiles Correlate with Protein Expression on HSCs (A) Gene expression over time. The actual observed values of each replicate at each time point are shown in red, and the line connects the predicted expression value at each time point based on our regression analysis. (B) Antigen expression on HSCs measured by flow cytometry. Gray lines represent negative control, red lines represent protein expression at day 0, and blue lines represent protein expression at day 7. (C) Cell cycle analysis of CD48 − and CD48 + HSCs isolated 6 d post 5FU treatment. Discussion Here we have identified proliferation and quiescence signatures of HSCs. Our experimental design utilized a combination of pair-wise comparisons and time course microarray experiments. The pair-wise analysis allowed us to find the genes different between quiescent and cycling HSCs, while the time course data allowed us to order these genes in a time-dependent manner. The power of our overall methodology is reflected in the remarkable overlaps between the gene lists presented and those extracted from published data ( Akashi et al. 2003 ), in particular the common P-sig and common Q-sig. Applying a novel approach to utilizing the GO annotations, we calculated the statistical significance of the enrichment of particular GO categories in our lists. We also devised a new method for calculating the distance between gene lists based on the GO structure. This allows one to assess the functional similarity, in “GO space,” of gene lists that may not have any actual genes in common (such as our TOM groups). Furthermore, since the GO vocabulary is not specific to any species, this method allows for cross-species and cross-platform comparisons of gene lists. Re-analysis of data from previous studies may reveal a functional stem cell signature in GO space that was not evident at the gene level ( Ivanova et al. 2002 ; Ramalho-Santos et al. 2002 ; Fortunel et al. 2003 ). Applying GO analysis to the TOM groupings revealed elemental subgroups within the signature lists that allowed us to construct a molecular model of the HSC activation cycle. The majority of unperturbed HSCs reside in a quiescence niche and express receptors, for example the metabolism- and ageing-associated receptor IGF1R and the receptor tyrosine kinase Tie1, that allow them to respond to multiple mitogenic signals ( Figure 5 A). They also express high levels of transcription factors, such as c-fos and GATA-2, that enable swift activation of HSCs. This expression profile, found in the TOM 0 and 30 groups, suggests that although adult HSCs are quiescent, they are in a “state of readiness” to react to changes in their environment. Figure 5 Model of HSC Activation Cycle (A) Normal HSCs reside in a quiescent niche in a “state of readiness” exemplified by the indicated genes. (B) Upon stress (5FU treatment), HSCs “pause” by remaining quiescent and in their niche while they “prepare” to proliferate. HSCs receive signals from proinflammatory cytokines at this point. The signals induce a proliferative state that is divisible into early (C) and late (D) phases. (C) “Early proliferation” is marked by an increase in expression of genes involved in DNA replication, repair, and cell migration molecules that allow movement of HSCs from the quiescence niche to the proliferative zone. (D) “Late proliferation” is marked by expression of many cell cycle genes as well as many energy pathway molecules. (E) Re-induction of quiescence involves changes in migratory molecule expression, which leads to return of cells to their quiescence niche, as well the expression of antiproliferative genes. Immediately after activation is triggered (here by 5FU), HSCs enter a superquiescent “pause.” This state, found at TOM 1 and also observed by cell cycle analysis ( Randall and Weissman 1997 ), is mediated by antiproliferative genes such as Tob1, p21, and Btg3 ( Figure 5 B). Interestingly, p21 -null mice have defects in HSC self renewal ( Cheng et al. 2000 ). We observed up-regulation of TIMP3 and the serine proteinase inhibitor A-3 g, which inhibit cell migration ( Qi et al. 2003 ). At least six interferon-γ-induced genes were also up-regulated at this point, suggesting that HSCs are responding to proinflammatory signals. We speculate that the pause in HSC proliferation and migration allows HSCs to survive 5FU cytotoxicity while the cells simultaneously “prepare” to proliferate and repopulate the bone marrow to ensure survival of the animal. In the early phase of proliferation starting at day 3, when increased numbers of HSCs in cell cycle are first detected ( Randall and Weissman 1997 ), HSCs have committed to cell division, as can be seen by the maximal expression of genes involved in DNA replication and repair ( Figure 5 C). At day 6, the late phase of proliferation, when the greatest number of HSCs are in cycle, we see expression of genes involved with energy production, indicating an overall increase in metabolic activity in the HSCs ( Figure 5 D). Prior work has linked HSC mobilization with proliferation ( Wright et al. 2001 ; Heissig et al. 2002 ), and our data indicate that the opposite is also true: to proliferate, HSCs need to move out of their quiescence niche and into a proliferative zone ( Figure 5 C and 5 D). We see the up-regulation of α4-integrin at day 3 followed by a dramatic decrease at day 6 post 5FU treatment. Experiments that block α4-integrin function by blocking antibodies or via knockout technology have previously shown that down-regulation induces increased mobilization and delays recovery after 5FU treatment ( Craddock et al. 1997 ; Scott et al. 2003 ). The gene expression pattern displayed by α4-integrin predicts that down-regulation of α4-integrin is necessary for 5FU-induced proliferation. As stated above, down-regulation of α4-integrin is sufficient to alter recovery of bone marrow progenitors after 5FU treatment, supporting the link between HSC proliferation and migration in our model. Down-regulation of c-Kit has also been linked to mobilization of HSCs ( Heissig et al. 2002 ), and its expression is lowest at day 6 post treatment. In order to “reset quiescence,” HSCs need to return to their niche ( Figure 5 E). This process begins at day 10, when the number of cycling HSCs falls and HSCs express the high levels of specific antiproliferative genes such as Btg1 and several components of the JAK/STAT signal transduction pathway. Both SOCS3 ( Soriano et al. 2002 ) and STAT3 ( Levy and Lee 2002 ) have been associated with both positive and negative regulation of proliferation and differentiation of various hematopoietic cell types. Simultaneous expression of SOCS3, STAT3, and STAT6 suggests a complex regulation of HSC quiescence, but earlier work examining STAT signaling in other stem cell populations gave us insight into the role of JAK/STAT signaling in HSCs. Expression of STATs has been shown to establish and maintain stem cell pluripotency in embryonic stem cells ( Raz et al. 1999 ). However in Drosophila testes, JAK/STAT activation is crucial for stem cell self renewal; perturbations by both loss and increase in expression lead to dramatic changes in the stem cell compartment ( Kiger et al. 2001 ). Notably, activation of the JAK/STAT pathway by PKD1 induces cell cycle arrest through p21-dependent mechanisms ( Bhunia et al. 2002 ). This supports our hypothesis that JAK/STAT signaling is important for inducing quiescence at day 10, since we have shown that p21 is likely involved in HSC cell cycle arrest. The involvement of JAK/STAT signaling in both stem cell pluripotency and HSC quiescence suggests that these processes may be linked in HSCs. Endoglin, also found in the TOM 10 group, is known to be expressed on both murine ( Chen et al. 2002 ) and human ( Pierelli et al. 2001 ) HSCs, and has been shown to decrease cell migration by increasing cell–cell adhesion ( Liu et al. 2002 ). Its expression pattern was negatively correlated with 5FU-HSC proliferation: it was lowest at day 6 after 5FU treatment, and highest at day 10. Our data suggest that HSC proliferation requires mobilization from the niche, and that restoration of quiescence is accompanied by a return to the niche. Endoglin's expression pattern makes it an ideal candidate for mediating HSC-to-niche homing and long-term association. Our model derived from gene expression profiles correlates well with the literature on HSC cell cycle and mobilization. Although other models of HSC mobilization and 5FU treatment have previously been proposed ( Heissig et al. 2002 ), our data allow association of specific genes with particular stages of HSC activation and recovery. Our model predicted that CD48 might preferentially mark cycling HSCs, and our cell cycle analysis of CD48 + and CD48 − HSCs confirmed this prediction. Our model also postulates the presence of “quiescence” and “proliferative” zones in the bone marrow; osteoblasts may be a component of this quiescence niche ( Calvi et al. 2003 ; Zhang et al. 2003 ). In summary, we present proliferation and quiescence signatures for HSCs that show remarkable overlap with published literature. In addition, this study revealed new, uncharacterized genes whose role in HSC self renewal needs to be explored: some of these genes may play as yet undiscovered roles in the development of cancer or may aid in the manipulation of stem cells for therapeutic uses. Finally, harnessing the GO using novel bioinformatics approaches to analyze our data at a global level allowed us to propose a model of the HSC activation cycle. Materials and Methods Flow cytometry For quiescent adult HSCs and 5FU-HSCs, whole bone marrow (WBM) was collected from the femurs and tibias of ten to fifteen 8- to12-wk-old normal or 5FU-treated C57Bl/6 mice. For 5FU treatment, mice were injected intravenously with a single dose of 5FU (150 mg/kg body weight; Sigma, St. Louis, Missouri, United States) and killed at day 0, 1, 2, 3, 6, 10, or 30 after injection. Day 0 mice were untreated, and day 1 WBM was isolated 17–19 h after injection; all subsequent days were in 24-h increments. WBM was stained with Hoechst 33342 to identify the SP cells ( Goodell et al. 1996 ) and then magnetically enriched for Sca1 + cells (autoMACS; Miltenyi Biotec, Sunnyvale, California, United States). Cells were stained with a biotinylated Sca1 antibody (clone E13–161.7; BD Pharmingen, San Diego, California, United States) and visualized with strepavidin-PE (Molecular Probes, Eugene, Oregon, United States). Sca1-enriched WBM was sorted for the SP profile and Sca1 positivity on a MoFlo (Cytomation, Fort Collins, Colorado, United States). Representative flow diagrams of cell sorting can be found in Figure S2 A. Phenotypic purity was typically 95% or greater. Regarding functional purity of the sorted populations, evidence from multiple sources in our lab and others suggests that both normal bone marrow and 5FU-treated SP cells are very highly enriched for HSCs. The whole SP contains both LT-HSCs and ST-HSCs, but has very limited contamination from committed progenitors or differentiated hematopoietic cells. For FL-HSCs, fetal livers were removed from embryos 13.5–14.5 d postcoitus and dissociated ( Jordan et al. 1990 ). Fetal liver cells were magnetically enriched for c-Kit + cells using c-Kit-biotin (clone 2B8, BD Pharmingen) and visualized with strepavidin-APC (Molecular Probes). The c-Kit-enriched cells were stained with a lineage cocktail consisting of cychrome-conjugated CD4 (L3T4), CD8 (53–6.7), B220 (RA3–6B2), GR1 (RB6–8C5), and Ter119 (Ter119) as well as Sca1-PE, and AA4.1-FITC (all antibodies from BD Pharmingen). FL-HSCs were identified as negative for the lineage markers and positive for Sca1, c-Kit, and AA4.1 (see Figure S2 B). Percentage of enriched cells was between 0.02% and 0.04% of total cells, with a purity of approximately 90%. For protein expression validation, SP cells from days 0 and 7 post 5FU treatment were analyzed for expression of Sca1-FITC, c-Kit-APC, and CD48-PE (HM48–1, BD Pharmingen) by flow cytometry. RNA isolation and amplification Total RNA was isolated from approximately 35,000–70,000 sorted HSCs using the RNaqeuous kit (Ambion, Austin, Texas, United States). All samples were then digested with DNaseI to eliminate residual genomic DNA, and extracted with phenol:chloroform. Total RNA was then subjected to two rounds of linear amplification using T7-based in vitro transcription (IVT) (MessageAmp, Ambion). Briefly, total RNA was reverse transcribed with an oligo-dT primer containing a T7 promoter sequence at the 5′ end (oligo-dT-T7 primer). To prime second-strand synthesis, RNA–cDNA hybrids were digested with RNaseH, producing patches of single-stranded cDNA. The second strand was filled in by DNA polymerase. The double-stranded cDNA served as a template for T7 RNA polymerase-driven IVT, which yielded up to 100× the starting mRNA pool. RNA probes were labeled in the second round of IVT with biotinylated nucleotides (Enzo Biotech, Farmington, Connecticut, United States). The second round of amplification was performed using random primers for first-strand synthesis and the oligo-dT-T7 primer to prime second-strand synthesis. Overall amplification was estimated to be 10,000-fold or greater ( Gallardo et al. 2003 ). Microarray hybridization Affymetrix (Santa Clara, California, United States) MG-U74Av2 chips were hybridized with fragmented, biotinylated aRNA according to standard protocols. Chips were then washed and counterstained using PE-conjugated strepavidin. Signal was amplified using the Affymetrix protocol for antibody amplification. The raw image (.DAT) and intensity (.CEL) files were generated using MAS 5.0 software ( http://www.affymetrix.com ). Microarray analysis Chip quality was assessed using various parameters outputted by a combination of the following software packages: MAS 5.0 ( http://www.affymetrix.com , BRB Array tools ( http://linus.nci.nih.gov/BRB-ArrayTools.html , and Bioconductor version 1.2 ( http://www.bioconductor.org ). Twenty-one chips were hybridized and analyzed, but only 16 (approximately 75%) passed our quality control standards (scale factor ≤ 10, 3′-to-5′ ratio ≤ 25, R 2 ≥ 0.97). Normalization and model-based expression values were calculated using the GeneChip Robust Multichip Analysis method ( Wu et al. 2003 ), available as part of the Bioconductor package. Statistical analysis Time-dependent expression profiles for each gene were analyzed by regressing the normalized expression values using polynomial least squares regression. ANOVA was performed on the coefficients of regression to identify genes with significant time patterns ( p < 0.05). The smooth curve fitting assumed that the expression trajectory for each gene followed a continuous time pattern. The class of fifth-degree polynomials was chosen for the fits, because it was the highest degree polynomial that did not interpolate the time point means. Analysis was performed in R 1.7.1 ( http://www.r-project.org ) using the Bioconductor suite of R packages. Source code for the analysis, including the curve fitting procedure, is available in Protocol S1 . GO analysis GO analysis was performed using the 1 October 2003 build of the gene ontologies ( http://www.geneontology.org ) and the GO annotations for each probe set on the MGU74Av2 chip, provided by Affymetrix ( http://www.affymetrix.com , downloaded 8 October 2003). The GO vocabulary structure was then instantiated as a directed acyclic graph and traversed to obtain hit counts for the genes in our lists that mapped at or below each node in the GO structure. To assess the significance of gene counts at each term, the annotations for the entire array were mapped to the GO structure, and counts for the whole array were obtained at each GO term. The significance of counts in particular categories was obtained via a sampling-without-replacement statistical model for the gene counts in each GO category. The probability of a count of k genes to a GO node at some level of the GO hierarchy was modeled according to the hypergeometric probability law: In the formula, B(x,y) is the binomial coefficient for x choose y . The value C is the total number of genes annotated to the GO node under consideration for the entire gene set. The value of L is the number of genes annotated to all nodes at the same level of the GO hierarchy, again considering the entire arrayed gene set. The value n is the number of genes annotated to terms at the same GO level for the gene list under consideration. The p value (one sided) for the node under consideration is obtained by summing probabilities as determined by the formula for all values of X from k to n . A web-based tool to perform this analysis on any gene list is available at http://franklin.imgen.bcm.tmc.edu/OntologyTraverser . The list distance metric was determined from the estimated joint distribution of probe counts across the GO structure for each gene list. This joint distribution was estimated by obtaining the counts at each GO node at each level. Only those nodes with non-zero counts in at least one list were included in the calculations. Relative frequencies at each GO node at each GO level were obtained by normalizing to the total counts at each level for each list. Once the frequency distribution at each level was determined, a Kullback–Leibler-like distance metric was constructed. Briefly, the distance metric is a weighted average of Kullback–Leibler distances at each level of the GO. The formula for computing distance between a pair of lists is The weights α k were normalized to sum to one and were drawn from the Poisson mass function with a mean of four. Since the GO levels are ordered in terms of increasing specificity, the contribution of each level was weighted differently: positive weight was applied to the middle of the GO hierarchy (levels 3–8), and weights for levels lower than 3 and higher than 8 were set to 0. The indices i and j in the formula indicate the lists being compared. The index k indicates the level of the GO under consideration, and the index γ considers each GO node at the level k. To compute the significance of our list dendrogram we determined the probability that we could arrive at the grouping pattern by chance. We determined the number of dendrograms with the “two group” pattern divided by the total number of labeled dendrograms. For our case, our “two group” dendrogram consisted of two subtrees with three and five arms, respectively. The total number of labeled dendrograms was the product of the number of labeled three-leaf dendrograms (three) and the number of labeled five-leaf dendrograms (105), which is 415. We divided this number by the total number of eight-leaf dendrograms (135,135) to attain the 0.003 probability. An R function for making this calculation is contained in the R script provided in Protocol S1 . Chromosome analysis Gene hits per chromosome were counted for Q-sig and P-sig as well as the total MGU74Av2 chip. Number of hits in our signatures was centered to the expected frequency of the number of hits on the total chip using the following equation. The number of hits above/below expected equals X − nP i , where X equals the number of genes in list on chromosome i, n equals the total number of genes in list, and P i equals the frequency of chromosome i hits on total chip (which equals the number of genes on total chip on chromosome i divided by the number of genes on total chip with known chromosome position). To determine the significance of enrichments and depletions of gene hits on each chromosome, we calculated a Z -score with the following equation. Chromosome enrichment or depletion between signatures was considered significant if the additive Z -score of Q-sig and P-sig was significant to 0.02 < α < 0.05. Supporting Information Figure S1 Cell Cycle Analysis of HSCs Cell cycle analysis of bone marrow SP cells before (left) and 6 d post (right) 5FU treatment. Before treatment, approximately 2% of adult quiescent HSCs are in cycle; 6 d after 5FU treatment, approximately 22% of HSCs are in cycle. (2.0 MB EPS). Click here for additional data file. Figure S2 FACS Isolation of HSCs (A) Representative flow cytometry plots of bone marrow enriched for Sca1 + cells at each time point. The indicated regions contain the SP cells. The table shows prevalence and purity from several isolations. (B) Representative flow cytometry analysis of fetal liver enriched for c-Kit + cells. (2.9 MB EPS). Click here for additional data file. Figure S3 Homogeneity of SP Cells SP profile of adult HSCs and 5FU-HSCs 6 d post 5FU treatment. Arrows point to analysis in SP cells of Sca1 and lineage marker expression showing greater than 97% homogeneity for Sca1 + and Lineage − expression. For analysis of adult HSCs on day 0, the lineage markers used were Mac1, CD4, CD8, B220, GR1, and Ter119. For analysis of 5FU-HSCs on day 6, all of the above markers were used except for Mac1, because of its low level expression on HSCs after 5FU treatment. (2.2 MB EPS). Click here for additional data file. Protocol S1 R Script for Constructing Gene Lists (8 KB TXT). Click here for additional data file. Table S1 Genes Up-Regulated in FL-HSCs (800 KB HTML). Click here for additional data file. Table S2 Genes in Proliferation Group (666 KB HTML). Click here for additional data file. Table S3 Genes That Change over 5FU Treatment Time Course (1.4 MB HTML). Click here for additional data file. Table S4 Genes in Quiescence Group (790 KB HTML). Click here for additional data file. Table S5 Genes Up-Regulated in Adult HSCs (933 KB HTML). Click here for additional data file. Table S6 Genes in P-Sig (336 KB HTML). Click here for additional data file. Table S7 Genes in Q-Sig (296 KB HTML). Click here for additional data file. Table S8 Genes in ST-HSC Signature (164 KB HTML). Click here for additional data file. Table S9 Genes in LT-HSC Signature (77 KB HTML). Click here for additional data file. Table S10 Genes in Common P-Sig (122 KB HTML). Click here for additional data file. Table S11 Genes in Common Q-Sig (59 KB HTML). Click here for additional data file. Table S12 Genes in TOM0 Group of Q-Sig (62 KB HTML). Click here for additional data file. Table S13 Genes in TOM1 Group of Q-Sig (70 KB HTML). Click here for additional data file. Table S14 Genes in TOM10 Group of Q-Sig (169 KB HTML). Click here for additional data file. Table S15 Genes in TOM30 Group of Q-Sig (45 KB HTML). Click here for additional data file. Table S16 Genes in TOM3 Group of P-Sig (52 KB HTML). Click here for additional data file. Table S17 Genes in TOM2 Group of P-Sig (18 KB HTML). Click here for additional data file. Table S18 Genes in TOM6 Group of P-Sig (280 KB HTML). Click here for additional data file. Table S19 Lists of GO Groups Enriched in Adult HSCs, FL-HSCs, and TOM Groups (2 KB HTML). Click here for additional data file. Table S20 Genes in Q-Sig, P-Sig, Common Q-Sig, and Common P-Sig in the GO Category “Cell Cycle” (7 KB HTML). Click here for additional data file. Table S21 Genes in Q-Sig, P-Sig, Common Q-Sig, and Common P-Sig in the GO Category “Cell–Cell Adhesion” (3 KB HTML). Click here for additional data file. Table S22 Genes in P-Sig in the GO Category “ATP-Synthesis-Coupled Electron Transport” (6 KB HTML). Click here for additional data file. Table S23 Genes in P-Sig in the GO Category “DNA Replication” (8 KB HTML). Click here for additional data file. Table S24 Genes in P-Sig in the GO Category “Cell Cycle Checkpoint” (6 KB HTML). Click here for additional data file. Table S25 Genes in P-Sig in the GO Category “Hydrogen Transport” (7 KB HTML). Click here for additional data file. Table S26 Genes in Q-Sig in the GO Category “Regulation of Cell Cycle” (7 KB HTML). Click here for additional data file. Table S27 Genes in Q-Sig in the GO Category “Defense Response” (9 KB HTML). Click here for additional data file. Table S28 Genes in Q-Sig in the GO Category “Protein Kinase Cascade” (7 KB HTML). Click here for additional data file. Table S29 Genes in Q-Sig in the GO Category “Cell–Cell Adhesion” (6 KB HTML). Click here for additional data file. Table S30 TOM1 Genes within GO Categories and the Fold Enrichment of Each Category (92 KB HTML). Click here for additional data file. Table S31 TOM1 Genes within GO Categories That Were Significantly Enriched (31 KB HTML). Click here for additional data file. Table S32 TOM10 Genes within GO Categories and the Fold Enrichment of Each Category (136 KB HTML). Click here for additional data file. Table S33 TOM10 Genes within GO Categories That Were Significantly Enriched (45 KB HTML). Click here for additional data file. Table S34 Genes Up-Regulated in Adult HSCs within GO Categories and the Fold Enrichment of Each Category (244 KB HTML). Click here for additional data file. Table S35 Genes Up-Regulated in Adult HSCs within GO Categories That Were Significantly Enriched (14 KB HTML). Click here for additional data file. Table S36 TOM0 Genes within GO Categories and the Fold Enrichment of Each Category (52 KB HTML). Click here for additional data file. Table S37 TOM0 Genes within GO Categories That Were Significantly Enriched (8 KB HTML). Click here for additional data file. Table S38 TOM30 Genes within GO Categories and the Fold Enrichment of Each Category (61 KB HTML). Click here for additional data file. Table S39 TOM30 Genes within GO Categories That Were Significantly Enriched (14 KB HTML). Click here for additional data file. Table S40 TOM3 Genes within GO Categories and the Fold Enrichment of Each Category (65 KB HTML). Click here for additional data file. Table S41 TOM3 Genes within GO Categories That Were Significantly Enriched (32 KB HTML). Click here for additional data file. Table S42 TOM6 Genes within GO Categories and the Fold Enrichment of Each Category (215 KB HTML). Click here for additional data file. Table S43 Lists of TOM6 Genes within GO Categories That Were Significantly Enriched (199 KB HTML). Click here for additional data file. Table S44 Lists of Genes Up-Regulated in FL-HSCs within GO Categories and the Fold Enrichment of Each Category (244 KB HTML). Click here for additional data file. Table S45 Genes Up-Regulated in FL-HSCs within GO Categories That Were Significantly Enriched (14 KB HTML). Click here for additional data file. Table S46 GeneChip Robust Multichip Analysis Normalized Data and Filtering Information (9.7 MB XLS). Click here for additional data file. Table S47 Goodness of Fit within Each TOM Group This table gives the 0.25, 0.5, and 0.75 quartile of the gene correlations (Pearson's) to their TOM group mean shown in Figure 2 A and 2 C. (27 KB DOC). Click here for additional data file. Accession Numbers The LocusLink ( http://www.ncbi.nlm.nih.gov/LocusLink/ ) accession numbers for the genes and gene products discussed in this paper are BTG1 (Locuslink 12226), Btg3/APRO4 (Locuslink 12228), CD48 (Locuslink 12506), c-fos (Locuslink 14281), c-maf (Locuslink 17134), cyclin G1 (Locuslink 12450), Endoglin (Locuslink 13805), GADD45β (Locuslink 17873), GATA-2 (Locuslink 14461), histone deacetylase 5 (Locuslink 15184), IGF1R (Locuslink 16001), melanoma antigen, family D, 1 (Locuslink 94275), p21 (Locuslink12575), receptor tyrosine kinase Tie1 (Locuslink 21846), serine proteinase inhibitor A-3 g (Locuslink 20715), SOCS3 (Locuslink 12702), STAT3 (Locuslink 20848), STAT6 (Locuslink 20852), Suppressor of cytokine signaling 3 (Locuslink 12702), TIMP3 (Locuslink 21859), Tob 1/APRO6 (Locuslink 22057), and α4-integrin (Locuslink 16401). The GEO ( www.ncbi.nlm.nih.gov/geo ) accession numbers for microarrays discussed in this paper are GSM26734-GSM26749.	/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC520599.xml
521491	Using email reminders to engage physicians in an Internet-based CME intervention	Background Engaging practicing physicians in educational strategies that reinforce guideline adoption and improve the quality of healthcare may be difficult. Push technologies such as email offer new opportunities to engage physicians in online educational reinforcing strategies. The objectives are to investigate 1) the effectiveness of email announcements in engaging recruited community-based primary care physicians in an online guideline reinforcement strategy designed to promote Chlamydia screening, 2) the characteristics of physicians who respond to email announcements, as well as 3) how quickly and when they respond to email announcements. Methods Over a 45-week period, 445 recruited physicians received up to 33 email contacts announcing and reminding them of an online women's health guideline reinforcing CME activity. Participation was defined as physician log-on at least once to the website. Data were analyzed to determine participation, to compare characteristics of participants with recruited physicians who did not participate, and to determine at what point and when participants logged on. Results Of 445 recruited physicians with accurate email addresses, 47.2% logged on and completed at least one module. There were no significant differences by age, race, or specialty between participants and non-participants. Female physicians, US medical graduates and MDs had higher participation rates than male physicians, international medical graduates and DOs. Physicians with higher baseline screening rates were significantly more likely to log on to the course. The first 10 emails were the most effective in engaging community-based physicians to complete the intervention. Physicians were more likely to log on in the afternoon and evening and on Monday or Thursday. Conclusions Email course reminders may enhance recruitment of physicians to interventions designed to reinforce guideline adoption; physicians' response to email reminders may vary by gender, degree, and country of medical training. Repetition of email communications contributes to physician online participation.	Background Many clinical practice guidelines have been disseminated through print publications and the Internet to improve the quality of healthcare. Complex physician behavior-change interventions have been created and applied to clinical practice guidelines to determine if the guidelines are being practiced. Research indicates that the adoption of clinical practice guidelines continues to lag behind their production and dissemination [ 1 ]. Adoption of such guidelines has been improved by the use of secondary strategies, which involve education, audit and feedback, benchmarking, office system interventions, and multifaceted interventions [ 1 ]. Continuing medical education (CME) providers and others interested in improving the quality of health care have had difficulty actively engaging practicing physicians in secondary strategies that may lead to guideline adoption [ 2 ], since CME courses and other educational interventions must compete with the physician learner's multiple priorities [ 3 ]. Over the past ten years, accessibility to educational opportunities has increased with advancements in technology, particularly the Internet. The Internet offers a delivery system for educational interventions and research activities that are more convenient to the physician learner than traditional live large group lectures and seminars [ 4 ]. The Internet generally functions in a passive manner, or as a "pull" technology, allowing the user to determine when, where and how to seek information. Engaging physicians in online educational activities may require the use of "push" technologies rather than pull technologies. "Push" technologies allow information to be delivered to the user rather than requiring the user to actively search for the desired in f ormation; they require minimal effort on the part of the recipient, which greatly supports their adoption [ 5 ]. Email was the first type of online push technology [ 5 ]. Other forms of push technologies are actively used such as pop-ups, list-serves, and screen savers. Based on preliminary research presented at the Proceedings of the American Medical Informatics Association's 2002 Annual Symposium , screen savers have been shown to be effective as a reminder system in engaging physicians in bioterrorism CME activity, although the use of email has exceeded all other Internet applications [ 5 ]. Flanagan and colleagues demonstrated promise for email as a means of engaging physicians in online activities designed to reinforce guideline use. They found physician response rate to email solicitation to be 50% over a 14-month study period [ 6 ]. Beyond this, little is known about the effectiveness of email in engaging physicians in online educational interventions designed to improve the quality of healthcare [ 7 ]. The purposes of this study were to investigate 1) the effectiveness of email reminders in engaging recruited community-based primary care physicians in an online guideline reinforcement strategy that was designed to promote women's health, 2) the characteristics of physicians who respond to email announcements, as well as 3) how quickly and when they respond to email announcements. Methods This study is a subanalysis of data from a randomized controlled trial (RCT), described in detail elsewhere [ 8 ]. The goal of the RCT was to improve Chlamydia screening rates of at-risk young women by community-based primary care physicians by using a multi-faceted online guideline reinforcement intervention. Primary care physicians were recruited from a national sample of 923 eligible offices identified by a large managed care organization's administrative data. Eligible offices had at least 20 young women (ages 16–26) at risk for chlamydial infection and at least one primary care physician (Internal Medicine, Family Medicine/General Practice, Pediatrics) with Internet access. In Phase I of recruitment occurred at the office level and Phase II at the physician level. In Phase I, all eligible offices were invited to participate via facsimile; an office was designated as "recruited" when one of its physicians declared intent to participate. In Phase II, an active link to the Intervention module was e-mailed to physicians recruited in Phase I. Recruited physicians were assigned at the time of logon to either a control or Internet CME intervention arm designed to improve Chlamydia screening rates; physicians within the same office were assigned to the same arm of the study. We designated physicians as "participating" when they first engaged the Internet intervention. The study was conducted between February 1, 2002 and December 31, 2002. The intervention included a series of four modules, feedback of performance data and a quality improvement toolbox. The education modules were based upon adult learning and behavior change theory [ 8 ]. Our driving principles included case-based learning [ 9 , 10 ] making programs interactive by adapting to the program learners' readiness-to-change stage [ 11 ] and performance feedback for behavioral motivation and reinforcement [ 12 ] The control condition was a series of four text-based modules on topics unrelated to Chlamydia. Physicians receive 1 category 1 CME credit per module for their participation. The main outcome measure was Chlamydia screening rates. The subanalysis of this trial focuses on the use of email reminders to engage physicians in the online intervention. Physician recruitment occurred from November of 2001 to January of 2002. Following recruitment, the intervention was initiated in February of 2002 via email broadcast to recruited physicians. During 2002, four separate educational modules were offered. Each module was introduced with a series of email announcements followed by email reminders that contained tailored subject line (i.e. Dear Dr. John Smith) and text message containing the URL that would connect the physician directly to the educational module. The first three modules, emphasized: (1) young sexually active women are at high risk for asymptomatic infection that may lead to future serious health consequences; (2) newer urine-based screening allows diagnosis without a pelvic examination; and (3) infection may be effectively treated with a one-dose antibiotic. The fourth module reviewed previously introduced concepts. A total of 33 announcements and reminders were sent to recruited physicians between February 1, 2002 and December 31, 2002, which represents one reminder overall for each 1.5 weeks of the study duration. Initially, we were concerned that frequent email reminders would be considered intrusive. After a trial period, we found that we could increase the frequency of reminders without difficulty, thus fewer email reminders were sent during the initial phases of the study. For the subanalysis of this trial, the principal outcome measure is participation, defined as physician log-on at least once to the website. Data was collected electronically when physicians logged on to the intervention. In addition, several evaluation questions were used at the conclusion of the online educational activities to explore CME preferences. Descriptive analyses of patient and physician demographics are used to compare baseline characteristics of patients and physicians in the email participants versus the recruited non-participants. Statistical significance is determined with tests appropriate to the distribution of the data (chi-square for categorical variables and student-t test for continuous variable). Two-tailed tests are used for all analyses. Results and discussion Four hundred eighty physicians were recruited to participate, representing 380 physician offices. Of the 463 recruited physicians, 445 were successfully contacted by email using the addresses they had furnished at the time of recruitment. Of these 445 recruited physicians, 210 (47.2%) physicians from 190 offices logged on to at least one of the educational modules. Of the 210 physicians who logged on at least once during the 45-week study period, one hundred twenty-four (59%) returned again to log-on for Module 2, eighty-seven (41%) logged on for Module 3, and forty-four (21%) logged on to Module 4. Two hundred and ten physicians of 445 logged on at least once to the website, leaving 235 physicians as non-participants. Figure 1 represents total physician log-on by week. Analysis of log-on days indicated that participants were most likely to log-on on Monday or Thursday (see Figure 2 ). Log-on times were also examined and findings indicate that physicians were most likely to log-on to a module between the hours of 3 P.M. and 7 P.M (15:00–19:00). Other common times for participants to log-on included times earlier in the day, or between the hours of 8 P.M. and midnight (20:00–24:00) (see Figure 3 ). Figure 1 Cumulative internet engagement by physician over time Figure 2 Number of logins by day of the week. Figure 3 Number of logins by time of day. The participant characteristics (n = 210) were compared with those of recruited but non-participating physicians (n = 235) (Table 1 ). Age was not significant, however there were significant differences by gender, degree, and country of medical training. By race, 81.5% of the participants were Caucasian, 8.8% were Asian, 3.0% were African American, 3.5% were Hispanic and 3.0% were listed as other with no significant differences. The largest percentage of recruited physicians was family practitioners (41.3%), followed by general internists (29.4%), pediatricians (9.6%), and general practitioners (1.5%), with no significant differences. Female physicians were significantly more likely to participate than males (p = . 0001), Medical Doctors (MDs) were significantly more likely than Doctors of Osteopathy (DOs) to participate (p = . 01), and graduates of U.S. medical schools significantly more likely than graduates of international medical schools to participate (p = . 01). In addition to demographic characteristics, chlamydia screening rates of participant and non-participant physicians were compared. Baseline screening rates of nonparticipants were significantly lower than those of participants. Non-participant chlamydia screening rates were 14.6% at baseline, compared to 17.4% for participants (p < .006). Table 1 Participant* versus recruited** but non-participant characteristics Recruited Non-participants Participants p-value N 235 210 Mean Age 44.3 45.2 .143 Gender Male 60.0% 40.0% .0001 Female 40.0% 60.0% Degree DO 65.3% 34.7% .01 MD 51.9% 48.0% Medical Training International 67.5% 32.5% .01 USA 52.7% 47.3% Participated physician engaged in at least one module *Eligible offices had at least 1 eligible physicians with at least 20 female patients who were candidates for Chlamydia screening according to the Health Employers Data and Information Set Technical Specifications, 2000. From follow-up evaluation question data, 100% of responding DOs felt that the course email reminders were effective in reminding them about educational modules compared to 92.6% of MDs; 95.9% of US medical graduates rated the reminders as useful, while only 77.8% of graduates from non-US medical schools found them useful. Concerning preferences for delivery of CME, no DOs reported web-based activities as their preferred method for lifelong learning; they preferred local (50.0%) and national meetings (50.0%). MDs were more likely to prefer web-based CME (37.0%) to local (35.1%) or national (16.6%) meetings. Female physicians reported a preference for web-based CME (50.0%), whereas male physicians preferred local (35.0%) and national (27.5%) meetings over web-based CME (25.0%). The results of this study indicate that email reminders may be useful in engaging nearly half of a group of practicing primary care physicians recruited to participate in an online women's health educational series. This data is consistent with McMahon et al.'s findings in comparing the use of email, fax and mail, finding that email reminders were more useful to increase response rates [ 12 ]. The study participation rate, (47.2%) is also consistent with the work of Flanagan et. al.'s study of participation in web decision support tools for the management of pneumonia [ 6 ]. However, the gap of up to 3 months between recruitment and the initiation of the online educational activities may have contributed to a lower participation rate. It is possible that by decreasing the gap in time between recruitment and announcement of the availability of the online educational activity, participation rates could be increased. Future study designs using email reminders should consider beginning the intervention immediately following agreement to participate or shortly thereafter. Two current studies using email reminders to promote educational courses in the prevention of glucocorticoid-induced osteoporosis and in the secondary prevention of cardiovascular disease in patients following a myocardial infarction have been designed to deliver the intervention immediately following agreement to participate [ 13 ]. Findings from follow-up evaluation question data indicate that CME providers interested in targeting specific groups of physicians may benefit from using alternative methods of CME recruitment and delivery. CME providers targeting DOs may want to explore ways to engage DOs in web-based learning activities or consider focusing activities that target DOs to local or national meetings. Providers of CME who are interested in engaging male physicians and graduates of U.S. medical schools may find email reminders useful, but they may also want to explore additional methods of recruitment. Persistent email reminders did increase physicians' response rates to online education, but response rate decreased with the number of reminders. The first three reminders produced the largest responses, with decline after the 10 th reminder. Based on our experience, it would seem reasonable that providers and researchers with limited resources consider focusing their announcements/reminders on the first 3–10 encounters. Data from time of log-on underscores the advantages of asynchronous online interventions for busy clinicians. Traditional "live" online symposia scheduled for the middle of the day might appeal to the physicians who logged on between 10 AM and 2 PM (10:00–14:00), but data from this intervention suggest that many physicians have more available time later in the day for online educational activity. The investigators findings of Monday being a frequent day for log-on was unexpected, but may offer an opportunity for future study designs to include weekend email broadcasting rather than a Thursday broadcast. The topic of the online educational activity may influence response to email announcements and reminders. While baseline Chlamydia screening rates were relatively low in both groups, the significant difference between the groups may indicate that those who are likely to perform better according to clinical practice guidelines are more likely to participate in online educational strategies that reinforce their use. Or the higher screening rates may be associated with a higher degree of interest in the overall topic area of women's health. Also related to the topic addressed in this online study, previous studies of preventive practices of female physicians have indicated they are more likely than males to promote preventive practices [ 14 ] and screening [ 15 ] including Chlamydia screening among their female patients than are male physicians. The advertised educational topic for this study was an online women's health course. More female physicians than male physicians responded to the email course reminders, but the topic may have had more appeal to female physicians than to male physicians, leaving the issue of whether there are gender differences among physicians in response to email reminders unresolved. Comparisons of characteristics of participant physicians and non-participant recruited physicians may be useful to those designing online recruitment and engagement strategies for future studies. Those using email reminders to communicate with physician populations including large numbers of DOs, however, may benefit from considering blended methods of CME recruitment and delivery. Using various methods of reaching providers, may also enhance DO participation. CME providers targeting DOs may want to explore additional ways to engage DOs in web-based learning activities or consider focusing activities that target DOs at local or national meetings. Conclusions Physicians' online clinical information seeking and engagement in online education continues to grow [ 12 ]. Researchers of online interventions who are attempting to improve the quality of healthcare and physician performance should continue to study and evaluate physician online behavior. Knowing when, where, and how physicians seek information on the Internet, and how they respond to receiving specific information pushed toward them, will prove to be very useful for targeting future quality improvement interventions. Reminding physicians often via email about online educational opportunities appears to increase engagement in a community-based primary care physician audience. The early and consistent implementation of this push technology may increase physicians' utilization of interventions designed to improve practice. Authors' contributions MA participated in the statistical analysis and drafted the manuscript. BCC participated in the statistical analysis and drafting of the manuscript. LC participated in all phases of the project. TW participated in the draft of the manuscript. CS participated in the design and coordination of the study and the drafting of the manuscript. MNR participated in the coordination of the study. NWW participated in the design and implementation of the study. JJA conceived of the study and participated in all phases of the study. All authors read and approved the final manuscript. Competing interests The authors declare that they have no competing interests. Pre-publication history The pre-publication history for this paper can be accessed here:	/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC521491.xml
553991	Volume-based non-continuum modeling of bone functional adaptation	Background Bone adapts to mechanical strain by rearranging the trabecular geometry and bone density. The common finite element methods used to simulate this adaptation have inconsistencies regarding material properties at each node and are computationally demanding. Here, a volume-based, non-continuum formulation is proposed as an alternative. Adaptive processes corresponding to various external mechanical loading conditions are simulated for the femur. Results Bone adaptations were modeled for one-legged stance, abduction and adduction. One-legged stance generally results in higher bone densities than the other two loading cases. The femoral head and neck are the regions where densities change most drastically under different loading conditions while the distal area always contains the lowest densities regardless of the loading conditions. In the proposed formulation, the inconsistency of material densities or strain energy densities, which is a common problem to finite element based approaches, is eliminated. The computational task is alleviated through introduction of the quasi-binary connectivity matrix and linearization operations in the Jacobian matrix and is therefore computationally less demanding. Conclusion The results demonstrated the viability of the proposed formulation to study bone functional adaptation under mechanical loading.	Background Much research effort has been devoted to understanding the functional adaptation of bone under physiological loading ever since the idea of bone functional adaptation was proposed by Wolff more than one hundred years ago [ 1 - 14 ]. Various computational models have been put forward in the past decades and the methods describing the changing rate of bone density corresponding to strain energy density, with finite element implementation, have become the most popular of them [ 6 , 15 - 29 ]. The common finite element approach is to take the element densities as the state variables and define elements with either constant or varying densities, then update the material densities for the next step of computation according to the computed strain energy density [ 22 , 23 , 26 , 27 ]. With more and more powerful desktop computers and commercial finite element analysis software available, this approach is widely used today. Yet some specific problems of this approach are not well addressed so far, although decades have passed, and the numerical results are inevitably affected. One common problem is the inconsistency of material densities on element boundaries [ 30 ]. During the updating of material densities in each step, different elements may take different densities due to strain energy density, thus often leading to conflicting material properties at the boundaries shared by more then one element. Since this conflict affects the integration points, which always come from the element boundaries, the errors are carried forward and cannot be eliminated by smoothing techniques. So it is not surprising that, if the program is allowed to run for a certain time, most of the elements tend to become either saturated or completely resorbed, leading to checker-board patterns especially in the proximal area of the femur [ 30 ]. Some effort has been made trying to solve this problem. For example, a node-based variant of the finite element method was tried with focus on the densities of the nodes rather than densities of the whole elements [ 21 , 30 ]. The node densities are then interpolated across the whole elements before the next step of computing begins. The results are improved, but the stress and strain quantities are still conflicting at the nodes. Other previous work has used Voronoi structures [ 31 ] to study the effects of crack growth on trabecular bone, tapered strut models [ 32 ] to study the ageing effect through a parametric approach or continuum FEM [ 33 ] to compute the strain energy density in order to overcome individual drawbacks of the common method described. Their potential impact on the formula proposed here is discussed below. The long existing problems and the limitations of assuming a continuum drive this new effort to explore the possibility of a non-continuum formulation of bone functional adaptations through nodal analysis in the hope of eliminating the errors present in the previous approaches. In the proposed non-continuum formulation, neighboring nodes are connected by struts that are defined with invariant material densities with respect to time but strut volumes are defined as state variables indicating different configurations of bone structure. The updating of strut volume will depend on the strain energy density in the strut in the previous step. As a result, there is no conflict either in density or in strain energy density. The shift of state variables from bone densities to bone volumes not only eliminates the errors inherent to the density-based finite element approaches but also transforms the continuum formulation to a non-continuum formulation [ 34 ]. The advantages of a volume-based non-continuum formulation may be appreciated by looking at the bone volume ratios in osteoporotic bones. The ever-increasing resolution of modern imaging techniques now allows to take a much closer look at the trabecular structure of the bone. In the trabecular network, trabeculae with different lengths and thicknesses are connected with each other to form a scaffold serving both mechanical and biological functions [ 35 , 36 ]. They are well connected in normal bones but poorly connected in osteoporotic bones in addition to reduced thickness. To characterize the trabecular structure, two terms are often used: bone volume/tissue volume (BV/TV) ratio and bone material orientation [ 15 , 25 ]. Although the cortical bone is densely packed with mineralized material, the trabecular bone dominates the inside space of the bone, highly exposed to bone marrow, highly distributed in volume, and highly involved in bone remodeling. The ratio of trabecular bone volume over tissue volume can be below 30% in osteoporotic bones, which means most space is taken up by void or bone marrow and this questions the appropriateness of a continuum formulation [ 35 ]. Besides elimination of the errors mentioned earlier, the small physiological range of bone deformation during normal activities allows linearization operations in the volume-based non-continuum formulation. This saves computation time and alleviates the high demand on hardware resources. The proposed volume-based non-continuum formulation shows computational advantages in modeling bone functional adaptations and has much potential for clinical applications in this field. Results Changes in trabecular structure Fig. 1 , 2 , 3 and 4 show the node-based representation of bone adaptation results according to loading cases of one-legged stance, abduction, adduction and the combined loading case respectively. Figure 1 Adaptation Results: one-legged stance. Results of bone functional adaptation. The color bar shows percentage of actual bone density against maximum bone density, which is 1.74 g/cm 3 . The density of bone structure is not indicated by the number of sample nodes selected in that region, but by the density (converted from volume) of each node, which is expressed as degree of "red" in this illustration. Figure 2 Adaptation Results: abduction. Results of bone functional adaptation. The color bar shows percentage of actual bone density against maximum bone density, which is 1.74 g/cm 3 . The density of bone structure is not indicated by the number of sample nodes selected in that region, but by the density (converted from volume) of each node, which is expressed as degree of "red" in this illustration. Figure 3 Adaptation Results: adduction. Results of bone functional adaptation. The color bar shows percentage of actual bone density against maximum bone density, which is 1.74 g/cm 3 . The density of bone structure is not indicated by the number of sample nodes selected in that region, but by the density (converted from volume) of each node, which is expressed as degree of "red" in this illustration. Figure 4 Adaptation Results: the combined loading case. Results of bone functional adaptation. The color bar shows percentage of actual bone density against maximum bone density, which is 1.74 g/cm 3 . The density of bone structure is not indicated by the number of sample nodes selected in that region, but by the density (converted from volume) of each node, which is expressed as degree of "red" in this illustration. The nodal density is a percentage relative to its maximum value (0~100%). It is 'converted' based on volume information (BV/TV) for the purpose of easy visual inspection. As suggested by Zhu [ 34 ], the value of 1.74 g/cm 3 is used as the maximum value in the present formulation. In the combined loading case and one-legged stance, the nodal densities are generally high in the proximal area where a considerable number of nodes reach the highest density due to the relatively high load, while in the large distal area, the densities become lower. In the case of abduction, the external load is the smallest out of the three loading cases and the nodal density in this case seldom reaches the maximum. Although the high densities also appear in the proximal area of femoral head, the densities are lower than those of other loading cases. In the large distal area, the node densities generally range at very low levels. In the case of adduction, the highest densities still appear in the proximal area of femoral head, but in the large diaphyseal area, the densities range at very low levels. In summary, one-legged stance and the combined loading case generally result in higher bone densities than the other two loading cases due to higher mechanical loading. The femoral head and neck are the regions where densities change most drastically under different loading conditions while the distal area always contains the lowest densities regardless of the loading conditions. Program convergence and bone fracture probability Starting from a uniform material distribution, the program computes the strain energy density, adjusts the strut configurations, and continues on to the next iteration. Fig. 5 shows the convergence of the model, which demonstrates the adapting process of the model evaluated with the normalized error residue against the final solution. The solution vector starts with a state of uniform material distribution, it then moves toward the final state with uniform strain energy density in the solution space. As the residue of solution drops, the residue of bone volume/tissue volume ratio also drops toward trivial while the program progresses. Figure 5 Program Convergence. Program convergence evaluated by residues between successive solutions. A prediction of bone fracture risk has been proposed using stress range levels [ 37 ] from which analysis of the material fatigue strength under given loading conditions can be derived. The material is subjected to a range of stress levels due to external load, then the stress leads to fatigue damage and finally leads to collapse of the material. This prediction method was used to estimate the fracture probability of the simulation described here. The BV/TV ratios, stress levels and fracture probability are shown in Fig. 6 , 7 and 8 respectively. While the program progresses toward the final solution, the volume of bone material used, indicated by the BV/TV ratio, increases by a few percents then slowly decreases again; meanwhile, the stress level moves down to a low level in the final phase and accordingly, the estimated fracture probability drops from around 90% to around 2% in the final step. It is safe to say that the program finally simulates a reasonable configuration of bone internal structure as a result of the physiological adaptation process. Figure 6 BV/TV Evolution During Adaptation Process. BV/TV evolution during adaptation process. Figure 7 Evolution of Apparent Normal Stress Level Within Bone Tissue. Apparent stress level evolution during adaptation process. Figure 8 Fracture Probability. Bone facture probability during adaptation process. Discussion A volume-based non-continuum formulation has been developed that describes the adaptation of bone to various mechanical loading situations. In the finite element approach to bone adaptation simulation, the integration of the entries in the coefficient matrix can be a heavy computing task [ 22 , 23 , 26 , 27 ]. In the model proposed here, this is alleviated through the introduction of the simpler connectivity matrix and Jacobian matrix [ 38 ]. In daily physiological activities, bone deformations are small and the Jacobian matrix can be linearized. As mentioned in the introduction, one common problem is the inconsistency of material densities or strain energy densities on element boundaries [ 30 ]. Since all the integration points come from the boundaries, the resulting errors will essentially affect the computation. In the improved node-based implementation of the finite element method, the stress and strain quantities are still conflicting at the nodes. In the volume-based non-continuum formulation proposed here, this conflict is eliminated. The material density and strain energy density are all consistent in each individual strut, and the computing becomes less demanding. As described above, previous work [ 31 - 33 ] addresses some of the weaknesses of the common FEM model. Makiyama [ 31 ] employed the "Voronoi structure" to study the effects of crack growth on trabecular bone. The method for generating a Voronoi structure could be quite useful when it calibrates the artificially constructed structure against the physical trabecular structure scanned from a patient. This might then serve as the starting state of bone configuration before adaptation begins. Moore [ 39 ] proposed a model to replace the partially damaged trabecula with another trabecula reduced in thickness. If this concept is combined with that of the strut structure, one may also derive the model proposed here, that is, a strut model with either varied modulus due to bone mineralization or adaptive cross-section/volume or even tapered struts as proposed by Kim [ 32 ]. Hip fracture is one typical manifestation of osteoporosis, and the results obtained by the simulation indicate that considerable changes of bone structure take place in the regions of femoral head and neck, where the stress level is normally higher than that of distal regions. The variations in stress level as shown in Fig. 7 reflect the adaptive process of the bone internal structure and different structural configurations will yield different stress levels in spite of little change in bone volume / tissue volume ratio. In the current literature, the time scale for adaptive processes is not very well defined. This general lack of knowledge poses a problem for any experimental proof of concept – while the numbers of strain repetition can be predefined, they must be done within biologically suitable time frames. If a given strain comes too sudden, the bone may break instead of remodel; if the strain is applied over too long a period, it may not be a sufficient stimulus to activate adaptive processes. The lack of well-defined temporal constraints, however, is common. Kim's approach [ 32 ] is very interesting in as much as it may allow to integrate the effect of time in the model proposed here. At present, however, there is not sufficient data available to allow to integrate time effects of the clinically interesting mid-range scale, i.e. weeks to months. Kim's model looks at the process of ageing of 35 years and more. The simulation model presented here may, beyond theoretical calculations, be applied to look at two clinical questions. Firstly, the simulation can be adjusted so that a realistic density distribution is the starting point, and outcomes following certain loading conditions, such as a predefined number of load cycles can then be predicted. Secondly, the program can integrate the measured bone density of a given patient to estimate the fracture risk based on stress level calculations. Conclusion By eliminating the common inconsistencies at each node, the formulation presented here shows good numerical performance and successfully predicts reasonable bone structure changes under different loading conditions. It is viable to serve as an alternative method apart from the traditional finite element based approached to study bone adaptations. In conclusion, the volume based non-continuum formulation is a new approach to bone adaptation study and has its own advantages. Methods Volume-based representation of the trabecular bone structure In the volume-based non-continuum formulation used here, the trabecular structure is represented by a connected strut system and each strut can take different sizes according to the mechanical loading requirements, that is, strain energy density. The strut representation is shown in Fig. 9 , which resembles a small volume of the trabecular structure. In this setting, the BV/TV ratio can be directly obtained from the ratio of the strut volumes over the unit volume, and material orientation can be obtained though the resultant of the vectorial material components of the struts, as described by equation (1) and (2). Figure 9 Bone Structure Decomposition. Representation of bone material by struts. Each strut can assume different geometric dimensions and material properties. Apparent mechanical property of bone material is based on the strut configuration. where v i is the volume of the i -th strut in the j -th basic unit, V j is the volume of the j -th basic unit, R j is the material orientation of the j -th basic unit, is the orientation of the i -th strut in the j -th basic unit and N is the number of struts in the j -th basic unit. For the formulation proposed here, the bone structure is decomposed into and represented by a connected network of struts. These struts are a mathematical abstraction of the physical bone structure, from which the BV/TV ratio can be derived. The thickness of a strut is adapted during the bone adaptation process in mimicry of the physiological processes. Volume adaptation under mechanical loading The bone mass will vary under mechanical loading. In engineering, the general relationship between varying mass, density and volume is described as: Since the density here is taken as constant regarding time, the second term on the right hand side, , simply vanishes and the mass variation is realized through volume variation under mechanical loading. Based on the density-based adaptation proposed by Zhu X. et al (39), which is stated as: the volume-based adaptation can thus be stated as: where β i = U i / ρ i k , which is a comparative coefficient describing the comparison of a given mechanical stimulus in each sensor cell with the reference value k , and U i represents the strain energy density for the I -th sensor unit; N is the number of sensor cells and f i ( x ) is the spatial influence function; B ( t ) is a remodeling coefficient; α indicates the remodeling power of strain energy density [ 34 ]. Non-Continuum formulation With the whole bone represented by a volume-based strut system, the non-continuum formulation can be noted as follows: where A is a connectivity matrix describing the connecting relationship between the struts, α is the linearized Jacobian matrix, is the nodal displacement vector to be solved for and is the loading vector derived from the external mechanical load. The generalized conjugate residue method is used to solve this formulation [ 38 , 40 ]. Connectivity matrix A is the matrix to show the relationship between connected struts with the entries of 1, -1 or 0. A strut starts from the node with the index corresponding to the entry 1 and ends at the node with the index corresponding to -1. It is further illustrated in Fig. 10 and equation (7). Figure 10 Connectivity of Struts. Physical connectivity relationship between struts is indicated by the connectivity matrix. Finally, the different loading conditions to be applied are shown in Fig. 11 . Figure 11 Loading cases. Quantitative information of different loading cases. Four loading cases are considered: one-legged stance, abduction, adduction and the combined loading case (weighted based on their respective daily occurrence cycles). Competing interests The author(s) declare that they have no competing interests. Authors' contributions Zhengyuan Wang developed the formulation and partly prepared the manuscript, Adrian Mondry participated in the adaptation controls and partly prepared the manuscript.	/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC553991.xml
549077	Association study with Wegener granulomatosis of the human phospholipase Cγ2 gene	Background Wegener Granulomatosis (WG) is a multifactorial disease of yet unknown aetiology characterized by granulomata of the respiratory tract and systemic necrotizing vasculitis. Analyses of candidate genes revealed several associations, e.g. with α (1)-antitrypsin, proteinase 3 and with the HLA-DPB1 locus. A mutation in the abnormal limb mutant 5 (ALI5) mouse in the region coding for the hydrophobic ridge loop 3 (HRL3) of the phospholipaseCγ2 ( PLCγ-2 ) gene, corresponding to human PLCγ -2 exon 27, leads to acute and chronic inflammation and granulomatosis. For that reason, we screened exons 11, 12 and 13 coding for the hydrophobic ridge loop 1 and 2 (HRL1 and 2, respectively) and exon 27 of the PLCγ-2 protein by single strand conformation polymorphism (SSCP), sequencing and PCR/ restriction fragment length polymorphism (RFLP) analyses. In addition, we screened indirectly for disease association via 4 microsatellites with pooled DNA in the PLCγ -2 gene. Results Although a few polymorphisms in these distinct exons were observed, significant differences in allele frequencies were not identified between WG patients and respective controls. In addition, the microsatellite analyses did not reveal a significant difference between our patient and control cohort. Conclusion This report does not reveal any hints for an involvement of the PLCγ -2 gene in the pathogenesis of WG in our case-control study.	Background Wegener granulomatosis (WG) is a systemic inflammatory disease of unknown aetiology characterized by granulomata of the respiratory tract and systemic necrotizing vasculitis [ 1 ]. There is a strong and specific association with presence of anti-neutrophil cytoplasmatic antibodies to a defined target antigen, proteinase 3 (PR3-ANCA), which is present within primary azurophil granules of neutrophils (PMN) and lysozymes of monocytes [ 2 ]. Upon cytokine priming of PMNs, this enzyme translocates to the cell surface, where PR3-ANCAs can interact with their antigens and activate PMNs [ 3 ]. It has been shown that PMNs from patients with active WG expressing PR3 on their surfaces produce respiratory burst and release proteolytic enzymes after activation with PR3-ANCA [ 4 ]. The consequence is a self-sustaining chronifying inflammatory process. WG appears as a multifactorial disease, and environmental influences still remain elusive. Several factors, such as bacterial infections, have been proposed as probable initiators of the disease [ 5 ]. It has been reported that chronic carrier status of Staphylococcus aureus is a risk factor for disease exacerbation in WG [ 6 ]. Recently a strong association of WG with distinct HLA-DPB1 alleles or rather an extended haplotype, respectively, in the MHC class II region has been reported [ 7 ]. In addition, analyses of candidate genes revealed several associations, e.g. with α (1)-antitrypsin, and proteinase 3 [[ 8 ] and [ 9 ]]. A primary candidate gene, PLCγ -2, was proposed on the basis of a novel animal model system. A mutation in the abnormal limb mutant 5 (ALI5) mouse [ 10 ] causes a phenotype comparable to human autoimmunity disease like WG: inflammations, granulomatosis, affected organs (lung, kidney with glomerulonephritis, eye and skin) and ANCAs were detected in the ALI5 mouse initially. Yet, this result has to be confirmed in further studies. The ALI5 mutation is located in the genomic region coding for the hydrophobic ridge loop 3. In mouse PLCγ-2 mutation leads to acute and chronic inflammations with a phenotype comparable to WG. Therefore, human PLCγ-2 is a good candidate gene for seeking predisposing genetic factors for WG. The human gene is a member of the PLC family comprising 12 closely related molecules involved in signal transduction from numerous receptors [ 11 ]. The protein is activated by cytoplasmatic tyrosine kinases (Lyn, Syk, and Btk) which are induced by engagement of B-cell receptors. In turn, the PLCγ-2 activation leads to the generation of diacylglycerol (DAG) and inositol 1,4,5-trisphosphat (IP3). While DAG activates protein kinase C (PKC, [ 12 ]), IP3 mediates Ca 2+ mobilization, which is required for activation of B-cells [ 13 ]. PLCγ-2 itself is expressed mainly in B-cell, hematopoetic cells, macrophages, granulocytes, testis, sperm, skin and brain [ 14 ]. The protein consists of two catalytic domains, which are separated by two SH2 and one SH3 domains [ 15 ]. It was established that PLCγ-2 mediates the coupling of G-protein-coupled receptors (GPCRs) and Ca 2+ entry in cell lines [ 16 ]. The human PLCγ-2 gene is localized on chromosome 16 (16q24.1) spanning 179 kb of genomic DNA. The 4.2 kb mRNA consists of 33 Exon coding for a M r 140,000 protein with 1265 amino acids [ 17 ]. PLCγ-2 is highly polymorphic [ 18 ]. Here, we report on an indirect association screen by microsatellite analysis and pooling of DNA in a case-control study design. Furthermore, we screened exons 11, 12 and 13, partly coding for the hydrophobic ridge loop 1 and 2 of PLCγ-2 protein, by single stranded conformation polymorphism (SSCP), sequencing and the PCR/RFLP method. As the ALI5 mutation is located in the region corresponding to the human exon 27, this exon was also screened by SSCP. Results Analyses of microsatellites In the present study we analysed 4 microsatellites intra- or juxta-genic of the PLCγ-2 gene with pooled DNAs from WG patients compared with those from controls (table 1 ; figure 1 ). All markers exhibited at least 3 alleles and did not show any intra-subgroup differences. The microsatellite analyses did not reveal significantly different allele distributions between WG patient and control pools (figure 1 ). Table 1 Primer sequences and information about microsatellites used in study Primer sequence No. sense 1 antisense nucleotide marker 1 CGCACATGTATCCAGAACT AGAGGTGGACCCATGCTTA di (AT) 2 CAAAGAAGATAAGGGCAGGC CCTAGGCGACTCAGTGAGACT tetra (TTTA) 3 AGGAGTTCGAGAAGAGCCTG TGCCACTACACCCAGATGAT di (AC) 4 TGATCTGTGTCTGGGCTTTC AGTTGTGACCCTAACATTGCA *di (AC) 1 The fluorescence labelled tail (5-Fam-CATCGCTGATTCGCACAT) was added to the 5'end of each sense primer Figure 1 Schematic representation of the human PLCγ-2 gene with relative localization of exons and investigated microsatellites. P values were generated by contingency tables (for details see "materials and methods"). Vertical lines: exons; No1-6: investigated microsatellites 1-6; Ex: exon. SSCP, sequencing and PCR/RFLP analyses SSCP analysis was chosen to identify mutations or polymorphisms, respectively, in exons 11, 12, 13 and 27 of the PLCγ-2 gene. DNA's of patients and controls with altered migration behaviour were sequenced. Identified variations were genotyped individually by SSCP and/or PCR/RFLP in individual patient and control samples (see table 2 ). Exon 12 revealed a low frequent SNP (1122G>A), which represents the first base pair of the exon. In exon 13 two previously identified SNPs (1293T>C and 1296T>C) were detected [ 18 ]. Both SNPs showed similar frequencies as reported [ 18 ]. All variations were found not significantly different in WG patients compared to the control group, and they were not associated with amino acid substitutions. In one patient our analyses revealed a single base substitution (3030G>A) in exon 27. Table 2 Summary of found variations in the human PLCγ-2 gene in exons 11, 12, 13 and 27 variation frequency of alleles exon bp 1 codon WG patients controls P value restriction enzymes 11 - - - - - - 12 1122G>A T329T 3/262 3/162 0.55 BanI 13 1293T>C 2 F382F 7/350 4/324 0.68 PfIFI 1296T>C 2 D383D 143/262 96/186 0.78 TaqI 27 3030G>A T961T 1/166 3 0/180 3 0.51 - 1 Numbering according to BC007565 (UCSC); 2 Previously reported SNPs; 3 Frequencies determined by SSCP analyses Discussion As recently reported a mutation in the PLCγ -2 gene in the ALI5 mouse leads to acute and chronic inflammation and granulomatosis. In addition, the ALI5 mice show similar symptoms as WG patients. For this reason screening of PLCγ -2 appears as a logical consequence in seeking candidate genes for WG. In the present study PLCγ -2 was analysed by an indirect microsatellite approach using pooled DNA from WG patients with a defined PR3-ANCA + status and a matched control cohort as reported before [ 7 ]. Furthermore, exons partly representing the catalytic domains of the PLCγ-2 protein, HRL1, 2 and 3, were analysed by SSCP, sequencing and by the PCR/RFLP method. We did not find any hints of an involvement of the PLCγ -2 gene in the pathophysiology of WG. Hence, we exclude at this time that the PLCγ -2 gene predisposes for WG in our cohort. Our study revealed 4 single base substitutions, 2 of which were reported before [ 18 ]. A further silent and low frequent SNP was detected in exon 12 which did not differ significantly between patient and control cohorts after PCR/RFLP analyses. As this SNP is the first basepair (bp) after the 3'-splice site one might hypothesise an influence on splicing but to present knowledge this SNP does not change a consensus sequence required for splicing. The ALI5 mutation is located in the HRL3 domain partly corresponding to the human exon 27 of PLCγ -2. Our analysis revealed a SNP (G>A) at position 3030 in one WG patient. The indirect microsatellite based approach did not reveal any association of PLCγ -2 with WG. Altogether 4 microsatellites spread in the PLCγ -2 gene were analysed. The approach using pooled DNA and ad hoc designed markers intragenic or in the immediate vicinity of a distinct gene has proven to be a reliable and efficient method to detected predisposing loci in WG before [ 7 ]. Here, none of the markers did show a significant allele distribution between the patient and control group. Conclusion In conclusion, our analysis of the human PLCγ -2 gene did not reveal an association of PLCγ -2 with WG. In contrast to ALI5 mice, where a single mutation leads to distinct symptoms of inflammatory autoimmunity, human WG depends on a more complex genetic background. Further analysis of all exons of PLCγ -2 might yield an association with WG but our microsatellite analysis strongly suggests that predisposition for WG is not due to variations in the PLCγ -2 gene. Material and Methods Patients and controls 175 well-characterized patients of German origin with a clinical diagnosis of WG and a defined PR3-ANCA + status were included in present study. Diagnosis of WG was established according international standards [[ 19 ] and [ 20 ]]. All patients were biopsy-proven. Biopsies were seen in German reference centre for vasculitis (Department of Pathology, University of Schleswig-Holstein Campus Luebeck, Germany) by 2 different observers. A group of 165 healthy individuals of German origin were used as controls. All persons gave informed consent. Microsatellite analysis Pooling of DNA was performed as reported before [ 7 ]. Patient (n = 150) and healthy control (n = 100) individuals from the abovementioned groups were divided into 3 and 2 sub-pools, respectively, containing 50 persons each. In this study 3 intragenic microsatellites as well as one in the immediate vicinity of the gene were included (table 1 ; see also UCSC Database, June 2002 Freeze; URL). Oligonucleotides were designed by Primer Express 2.0 software (ABI) adjusted to an annealing temperature of 55°C. For PCR we used the 'tailed primer PCR' as described before [ 7 ]. For amplification three oligonucleotides were used: 1. tailed sense primer (tailed F), 2. anti-sense primer and 3. labeled primer (labeled F) corresponding to the 5'-tail sequence of tailed F. PCR conditions were as follows: 1 × PCR buffer (Qiagen), 1.5 pmol labeled F, 0.2 mM each dNTP, 3 mM MgCl 2 , 0.2 pmol tailed F, 1.5 pmol reverse primer, 0.25 U Qiagen Hot Start Taq (Qiagen) and 50 ng DNA. Electrophoreses were run using ABI standard protocols. Raw data were analyzed by the Genotyper software (ABI) producing a marker-specific allele image profile (AIP, see [ 21 ] and [ 7 ]). AIP consists of series of peaks with different heights reflecting the allele frequency within each analyzed DNA pool. Statistics for comparisons of allele frequencies of patients and controls was performed as described before [[ 21 ] and [ 7 ]]. Case and control distributions were compared statistically by means of contingency tables. SSCP, sequencing and PCR/RFLP analyses Exons 11, 12, 13 and 27 were analysed by the SSCP method. PCR was performed using oligonucleotides reported before ([ 18 ], exon 27 corresponding to exon 26). Heat-denaturated fragments were then separated by polyacrylamide gel electrophoresis under non-denaturating conditions yielding specific band patterns for each of the alleles. Results were visualized by autoradiography. Alleles of representative probes were determined by direct DNA sequence analysis on a 377 ABI automatic sequencer (ABI). Afterwards, variations were genotyped individually by PCR/RFLP method with restriction enzymes specific for the respective change (for details see table 2 ). The variation in exon 27 was individually genotyped by the SSCP method. Authors' contribution PJ and SW carried out the molecular genetic studies, performed the statistical analysis and drafted the manuscript. PY participated in study design and helped to draft the manuscript. EC and WLG provided the samples and performed diagnostics of the patient group. JTE conceived of the study, and participated in its design and coordination and helped to draft the manuscript. All authors read and approved the final manuscript.	/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC549077.xml
553985	Determinants of health insurance ownership among South African women	Background Studies conducted in developed countries using economic models show that individual- and household- level variables are important determinants of health insurance ownership. There is however a dearth of such studies in sub-Saharan Africa. The objective of this study was to examine the relationship between health insurance ownership and the demographic, economic and educational characteristics of South African women. Methods The analysis was based on data from a cross-sectional national household sample derived from the South African Health Inequalities Survey (SANHIS). The study subjects consisted of 3,489 women, aged between 16 and 64 years. It was a non-interventional, qualitative response econometric study. The outcome measure was the probability of a respondent's ownership of a health insurance policy. Results The χ 2 test for goodness of fit indicated satisfactory prediction of the estimated logit model. The coefficients of the covariates for area of residence, income, education, environment rating, age, smoking and marital status were positive, and all statistically significant at p ≤ 0.05. Women who had standard 10 education and above (secondary), high incomes and lived in affluent provinces and permanent accommodations, had a higher likelihood of being insured. Conclusion Poverty reduction programmes aimed at increasing women's incomes in poor provinces; improving living environment (e.g. potable water supplies, sanitation, electricity and housing) for women in urban informal settlements; enhancing women's access to education; reducing unemployment among women; and increasing effective coverage of family planning services, will empower South African women to reach a higher standard of living and in doing so increase their economic access to health insurance policies and the associated health services.	Background A health system in any country performs instrumental functions of stewardship (oversight), creation of resources (investment and training), delivering services (provision), and financing (collecting, pooling and purchasing) [ 1 ]. Ultimately, the effectiveness and efficiency with which these functions are executed determine the extent to which a health system achieves its intrinsic goals of improving health, responding to people's non-medical expectations, and fairness of financial contributions. For instance, the extent to which South Africa's post-apartheid government will be able to attain its vision of creating a caring and humane society in which all citizens have access to affordable good quality health care will depend on the performance of its national health system [ 2 ]. Prior to the 1994 democratic elections in South Africa, the health system was built on the apartheid ideology, which was characterized by racially segregated health services, geographical disparities, fragmentation, duplication and specialized hospital-centred services favouring the urban populations [ 3 ]. The health system's functions were inefficiently and inequitably performed. The health system was not effective in: improving the health status of the majority of the formerly disenfranchised South Africans; responding to their legitimate non-medical expectations; and ensuring their social protection from the impoverishing catastrophic health expenditures [ 4 ]. Consequently, provision of primary health services was for a long time neglected and inequitable [ 34 ]. The worsening poverty situation in some provinces has made the primary health services inaccessible to the majority of the population [ 35 ]. In an attempt to overcome the above-mentioned public health challenges, the post-apartheid democratic governments have introduced several reforms in the health sector: creation of a quasi-federal structure with one national and nine provincial departments of health (provincial legislatures and bureaucracies); establishment of a district health system, including expansion and upgrading of the primary care infrastructure; and health care financing [ 5 ]. The latter included: use of a population and need-based resource allocation mechanism by the National Department of Health up to 1995 (after which each province receives a block grant directly from the national treasury); the removal of public sector fees for pregnant and lactating women, children under six years of age and all those who use the public primary health care system; and the enactment of a Medical Insurance Schemes Act aimed at regulating the medical insurance schemes industry more effectively. In spite of the abovementioned policy interventions, primary health care still remains under funded. For example, in 2003/2004 financial year, South Africa spent a total of 36.9 billion Rands (1 Rand = US$6) on health: 61.3% was spent on hospitals curative care, 16.1% on primary health care, 1.8% on HIV/AIDS treatment, 2.3% on nutrition, 4.0% on emergency services, 4.3% on administration, 2.5% on training, 1.6% on support services and 6.1% on other services [ 34 ]. According to Doherty and McLeod [ 6 ] and McIntyre et al. [ 7 ], the primary objectives of the Medical Insurance Schemes Act of 1998 are to: (i) increase the number of people covered; (ii) improve health-related cross subsidization within individual medical insurance schemes and curtail exclusion of high-risk groups, e.g. the elderly; (iii) prevent 'dumping' of medical insurance scheme members on public hospitals by requiring the schemes to cover a prescribed minimum package of hospital services for all members; and (iv) ensure effective health care cost containment. In 2001, there were 146 registered medical schemes (i.e. those falling fully under the Act) and eight Bargaining Council schemes (i.e. those granted exemption from certain provisions of the Act), all covering a total of 7 million people [ 6 ], i.e. less than 20% of the South African population. An additional 2 million people were covered by private insurance or industry-specific health services [ 7 ]. Unfortunately, there has been no significant increase in the size of the population covered by the medical schemes since the implementation of the Act. For instance, in 2003 South Africa had uninsured population of 38.6 million people: 15.1% were from Eastern Cape province, 6.0% from Free State, 17.8% from Gauteng, 22.1% from Kwazulu Natal, 12.8% from Limpopo, 7.2% from Mpumalanga, 1.7% from Northern Cape, 8.5% from North West and 8.7% from Western Cape [ 34 ]. The objective of this study was to examine the relationship between individuals' demographic, economic and educational characteristics, and their likelihood of being insured. Methods Conceptual framework There are two kinds of risks involved in health care: (i) the risk of becoming ill, with the accompanying loss in the quality of life, cost of medical care, loss of productive time during illness and, in more serious cases, death; and (ii) the risk of total or incomplete or delayed recovery [ 8 ]. Welfare economics of uncertainty predicts that individuals would like to insure against both forms of risks. The theory of expected utility, on which this study is based, assumes that each individual strives to maximize the expected value of a utility function; individuals are normally risk-averse, meaning that they have a diminishing marginal utility of income; and health risks for different individuals are basically independent, so that pooling them reduces the risk to the insurer to relatively small proportions [ 9 ]. In the South African National Health Inequalities Survey (SANHIS) [ 10 ], the respondents were asked the following question: "Does anyone in this household belong to a medical aid or health insurance scheme? 1 = Yes, 2 = No". Given the dichotomous nature of this question, we shall assume that the potential health insurance consumer faces the choice between purchase of some or no insurance. The consumer chooses between the two prospects on the basis of the utility expected from each. The potential consumers of insurance are assumed to make decisions based on the magnitude of the perceived difference between the level of expected utility with insurance (EU 1 ) and expected utility without insurance (EU 2 ). We need to analyse the effect of changes in the independent variables on the difference in the level of the expected utility of the two prospects, i.e. EU 1 minus EU 2 . If the difference were equal to zero, the consumer would be expected to be indifferent between the two prospects. However, if the difference were greater than 0, then the risk-averse consumer would be expected to opt for insurance [ 11 ]. The empirical model used in the analysis of individual household's choice between having no health insurance and having health insurance is presented in the Appendix . The variables included in the model are defined in Table 1 . Table 1 Definition of variables Variable Variable description Health insurance ownership 1 = if the respondent has health insurance; 0 otherwise Health rating 1 = if self-evaluated health status is excellent, very good or good; 0 otherwise Environment rating 1 = if the respondent feels that the environment she lives in is good, very good or excellent; 0 otherwise Residence 1 = if the respondent resides in either metro formal area, metro transitional area, smaller city/town formal area, smaller city/town transitional area, or rural white farms; 0 = metro informal area, smaller city/town informal area, or rural – "homeland" Income Total monthly gross income in Rand (US$≈6 Rand) Education Respondent's education level: 1 = matriculation (standard 10 or secondary school) and above; 0 = below matriculation Age Respondent's age in years Age squared Respondent's age squared Race 1 = if respondent is white; 0 if person of colour Household size Total number of persons in a household Occupation 1 = if a white-collar worker; 0 otherwise Employment status 1 = if unemployed and looking; 0 = otherwise Smoking 1 = if the respondent smokes cigarettes; 0 otherwise Alcohol use 1 = if the respondent drinks alcohol; 0 otherwise Contraceptive use 1 = if respondent uses a contraceptive; 0 = otherwise Marital status 1 = if married; 0 = single, separated or divorced Since economic theory does not provide much guidance on model specification, the choice of explanatory variables in the current study were guided by the past health insurance demand studies undertaken in the U.S.A. [ 12 - 15 ], Europe [ 16 - 18 ] and Israel [ 19 , 20 ]. Once again, based on past health insurance choice analysis studies, the coefficients of the variables included in equation 2 in the Appendix would, a priori be expected to assume the signs indicated in Table 2 . Table 2 Hypothesized relationships between the dependent variable (insurance ownership) and independent variables Independent Variables Variable coefficient Expected sign Studies from which the hypothesized signs are based Health rating B 1 Negative Trujillo [23], Coasta and Garcia [28] Environment rating B 2 Indeterminate Residence β 3 Positive Liu and Chen [31] Income β 4 Positive Deb et al [22], Trujillo [23], Vera-Hernandez [26], Coasta and Garcia [28], Besley et al [32] Education β 5 Positive Deb et al [22], Trujillo [23], Vera-Hernandez [26], Coasta and Garcia [28], Besley et al [32], Liu and Chen [31] Age β 6 Positive Trujillo [23], Liu and Chen [31], Grossman [12] Age squared β 7 Negative Trujillo [23], Grossman [12] Race β 8 Indeterminate Household size β 9 Negative Deb et al [22], Vera-Hernandez [26], Besley et al [32] Occupation β 10 Positive Vera-Hernandez [26] Employment status β 11 Negative Vera-Hernandez [26], Liu and Chen [31] Smoking β 12 Indeterminate Alcohol use β 13 Indeterminate Contraceptive use β 14 Indeterminate Marital status β 15 Positive Rhine et al [21], Trujillo [23], Liu and Chen [31] A Chi-Squared test (χ 2 ) for independence was undertaken to test the relationship between health insurance ownership and the individual independent variables. The null hypothesis is that the two variables are independent of one another (because there is a significantly large difference between the observed data and the calculated expected data). The alternate hypothesis is that the two variables are dependent on each other (because there is not a significant difference). Thus, for health rating, the hypotheses are as follows: (i) H 0 : A person's health insurance ownership status and their health rating are independent or unrelated; and (ii) H A : A person's health insurance ownership status and their health rating are related (dependent on each other). Data The data for empirical analysis were taken from the 1994 South African National Health Inequalities Survey, a household survey of a randomly selected sample of the South African population between the ages of 16 and 64 years [ 10 ]. The full sample was 3,796 persons, out of which 3,489 were women. Our analysis focused on the latter. The data set was rich with economic, demographic, social and health characteristics of the respondents. Results Descriptive statistics Table 3 presents the frequency and percentage distribution of the dependent and independent variables. Overall, 30% of the women in the sample said that they had a household member with a health insurance policy. Ninety five percent of the respondents who reported to have got a household member belonging to a medical aid or health insurance scheme lived in formal city dwellings and/or on farms owned by white South Africans. Table 3 Frequencies and percentages for explanatory variables Variables With insurance: frequency (%)[N = 1044) Without insurance: frequency (%) [N = 2445] Chi-square (p-value) Health rating: 1 = Excellent/very good/good 435 (41.67) 1238 (50.63) 23.57 ( p < 0.0001) 0 = Fair and poor 609 (58.33) 1207 (49.37) Environment rating: 1 = Good/very good/Excellent living environment 491 (47.03) 1580 (64.62) 93.843959 (P < 0.0001) 0 = Fair or poor 553 (52.97) 865 (35.38) Residence: 1 = Formal city dwellings + white farms 995 (95.31) 1625 (66.46) 325.45 (P < 0.0001) 0 = Informal dwellings +"former homelands" 49 (4.69) 820 (33.54) Income (in Rand): No regular income 137 (13.12) 275 (11.25) 1238.93 (P < 0.0001) 1 – 950 100 (9.58) 1497 (61.23) 951 – 1900 179 (17.15) 435 (17.79) 1901 – 3800 317 (30.36) 191 (7.81) 3801 – 7600 223 (21.36) 38 (1.55) 7600 + 88 (8.43) 9 (0.37) Education: 1 = Matriculation (standard 10) and above 566 (54.21) 2205 (90.18) 579.15 (P < 0.0001) 0 = Below matriculation 478 (45.79) 240 (9.82) Age (in years): 16 – 25 128 (12.26) 328 (13.42) 16.53 (P = 0.0024) 26 – 35 299 (28.64) 666 (27.24) 36 – 45 284 (27.20) 586 (23.97) 46 – 55 189 (18.10) 404 (16.52) 56 and above 143 (13.70) 461 (18.85) Race: 1 = African, Coloured & Indian 947 (90.71) 1981 (81.02) 50.87 (P < 0.0001) 0 = White 97 (9.29) 464 (18.98) Household size: 1 – 4 household members 647 (61.97) 1043 (42.66) 123.30 (P < 0.0001) 5 – 8 351 (33.62) 1114 (45.56) 9 – 12 41 (3.93) 240 (9.82) 13 and above 5 (0.48) 48 (1.96) Occupation: 0 = White-collar worker 326 (31.23) 168 (6.87) 357.05 (P < 0.0001) 1 = Blue-collar worker 718 (68.77) 2277 (93.13) Employment status: 1 = Involuntarily unemployed 967 (92.62) 1933 (79.06) 95.94 (P < 0.0001) 0 = Voluntarily unemployed or employed 77 (7.38) 512 (20.94) Smoking: 1 = If a cigarette smoker 292 (27.97) 607 (24.83) 3.78 (P = 0.0519) 0 = Not a cigarette smoker 752 (72.03) 1838 (75.17) Alcohol use: 1 = Alcohol drinker 148 (14.18) 276 (11.29) 5.71 (P = 0.0168) 0 = Not alcohol drinker 896 (85.82) 2169 (88.71) Contraceptive use: 0 = Uses contraceptives 213 (20.40) 399 (16.32) 8.43 (P = 0.0037) 1 = Does not use contraceptives 831 (79.60) 2046 (83.68) Marital status: 1 = Married 627 (60.06) 1144 (46.79) 51.53 (P < 0.0001) 0 = Single, separated, divorced 417 (39.94) 1301 (53.21) In the overall sample, the median income was 849.5 Rand and the mean was 1546.4 Rand, with a standard deviation (STD) of 1998.2 Rand. The sub-sample of the households without health insurance had an average income of 853 Rand (STD = 1073) compared to 3172 Rand (STD = 2623) in the group with insurance. Seventy-seven percent of the households with at least one member with health insurance had a monthly income of more than 951 South African Rand, compared to only 28% among the group without health insurance. The average and median household size in the overall sample was five members, with a standard deviation of about 3. The group with health insurance had an average household size of four people (STD = 2.1) compared to 5 people (STD = 2.7) in the group without insurance. Fifty-seven percent of the sub-sample without health insurance had a household size of more than five members compared to 38% among the group with health insurance. The group with health insurance had an average age of 40 years (STD = 12.5) compared to 41 years (STD = 14.1) in the group without insurance. The average age in the whole sample was 40.9 years (STD = 13.6) and the median was 39 years. As alluded to in the methods section, a Chi-Squared test for independence was undertaken to test the relationship between health insurance ownership and the individual independent variables. The results are presented in the last column of Table 3 . Since the computed Chi-Squared values for health rating, environment rating, residence, income, education, age, race, household size, occupation, employment status, alcohol use, contraceptive use and marital status were greater than their respective critical Chi-Squared values (at 5% significance level) we reject the null hypotheses and conclude that the row and column variables in Table 3 are not independent. Thus, for example in the case of health rating, we would accept the alternative hypothesis (H A ) that a person's health insurance ownership status and their health rating are related (dependent on each other). Regression analysis Table 4 provides odds ratios, 'p' values, coefficients, and 't' test values. The t-test is used to test the hypothesis (i.e. H 0 : β = 0) about individual regression slope coefficients. The 't' values for individual variables are obtained by dividing their coefficients (e.g. β INCOME ) by their standard errors (e.g. SE INCOME ). For example, the coefficient for income is 0.0004727 and the standard error is 0.0000339, and given H 0 : β = 0, the relevant t-value is indeed 13.946 as specified in Table 4 : Table 4 Logistic model regression results Explanatory variables Odds ratios [95% confidence interval] Coefficients 't' Health rating 0 0.000009-0.0005 -9.676 -9.537* Environment rating 26.76 12.43–57.60 3.287 8.404* Residence 6.969 4.93–9.84 1.942 11.020* Income 1.001 1.00-1.00 0.0005 13.946* Education 2.315 1.80–2.97 0.84 6.600* Age 1.148 1.09–1.20 0.138 5.751* Age squared 0.999 0.99–1.00 -0.0008 -3.401* Race 0.787 0.59–1.04 -0.239 -1.691 Household size 0.891 0.86–0.93 -0.115 -5.519* Occupation 0.733 0.54–0.99 -0.311 -1.971 Employment status 0.518 0.38–0.69 -0.657 -4.308* Smoking 1.633 1.29–2.07 0.49 4.052* Alcohol use 0.617 0.45–0.84 -0.483 -3.033* Contraceptives use 0.372 0.26–0.52 -0.988 -5.700* Marital status 1.841 1.49–2.27 0.611 5.765* Constant - - -4.385 -8.755 Sample size 3489 χ 2 (15) 1438.62 Prob > χ 2 0 Pseudo-R 2 0.3379 Log likelihood -1409.7041 Note: * Indicates the coefficients are statistically significant at 95% confidence level, based on a two-tailed test. On the basis of chi-square test of the log-likelihood ratio, the joint effects of estimated logistic model are statistically significant at the 0.1% level. t INCOME = (β INCOME )/(SE INCOME ) = (0.0004727)/(0.0000339) = 13.946 The decision rule is: reject the null hypothesis (H 0 : β = 0) if the calculated t-value, t k , is greater than the critical t-value, t c , as long as the sign of t k is the same as the sign of the coefficient implied in the alternative hypothesis (H A : β≠0). Otherwise, accept the null hypothesis (H 0 ) that the estimated regression coefficient in question is not significantly different from zero. In the above example, the coefficient of income is statistically significant at 95% level of confidence, based on a two-sided test, since the computed t-value (13.946) is greater than the critical t-value (1.960). The coefficient (β) of the estimated binary logit model measures the impact of a one-unit change in an explanatory variable (R i ) on the log of odds of a health insurance policy ownership, holding other explanatory variables constant. The coefficients for environment rating, residence, income, education, age, smoking and marital status are statistically significant at 95% level of confidence, and have positive signs. The latter result implies that an increase in any of these variables spontaneously impacts positively on the log of odds of health insurance policy ownership, holding other factors constant. Contrastingly, the coefficients for household size, alcohol, contraceptive use, age-squared, occupation, employment and health rating are statistically significant, and have got a negative effect on the log of odds of health insurance policy ownership. Discussion Health rating Social scientists define health as a product of life expectancy (measured in years) and health-related quality of life (i.e. mobility, activities of daily living, social participation, pain, anxiety/depression, energy) [ 22 , 23 ]. The SANHIS data set [ 10 ] contained only self-evaluated categorical health status data. The respondents rated their current health status as either excellent, very good, good, fair or poor. The variable was re-coded into a dichotomous variable: 1 = excellent, very good, or good; and 0 if fair or poor. An individual's stock of health determines the total amount of time he/she can spend producing commodities and money earnings [ 24 ]. Health status is an important determinant of both earnings and capacity for enjoying life. A decline in the death rate at working ages may improve earning prospects by extending the period during which earnings are received [ 25 ]. The coefficient for the health status variable took a negative sign, implying that the demand for health insurance was likely to be low among individuals who were in excellent, very good or good health. In the current study, 58.3% of the respondents in the sub-sample with a health insurance policy assessed their health status as either fair or poor. This may be a case of adverse selection [ 26 ], which results in insurance having the greatest appeal to individuals who are more likely to fall sick [ 9 , 27 ]. Adverse selection, depending on its extent, could jeopardize the economic viability of a health insurance scheme. The adverse selection problem can be curbed in two main ways without compromising equity objectives, namely: (i) compulsory social health insurance (or a national health service) for a defined population through legislation – not an option for private health insurance which is voluntary by definition; and (ii) the government could step in and provide health insurance to all those at exceptionally high risk (e.g. the elderly and those with chronic diseases) and the poor who cannot afford the premiums. Commercial health insurance firms often curb adverse selection by introducing experience rating, i.e. linking insurance premium to the degree of assessed risk of falling sick (this action may have negative equity implications); and/or subjecting all those who apply for insurance cover to a thorough medical examination (this could potentially lead to cream-skimming, excluding all those with high risks of falling sick). Economic factors There were three economic variables, namely, income (+), occupation (-) and employment (-), where (.) is the hypothesized sign of the coefficients. The coefficients of the three variables were statistically significant and had the expected signs. High incomes, white-collar occupations and being gainfully employed are significant predictors of health insurance ownership. The proportion of people with health insurance rises considerably as one moves up the household income distribution ladder, with the coverage going from 6.3% of those in the income range 1–950 Rand to over 90.7% among those earning 7600 Rand and above per month. The trend is similar to that reported by Harmon and Nolan [ 16 ] in Ireland and Propper [ 17 ] in England and Wales. This implies that any macroeconomic interventions aimed at decreasing involuntary unemployment and boosting disposable incomes among households will spontaneously increase the probability of health insurance ownership. Thus, the post-apartheid South African government's economic programmes of black empowerment, small-scale micro financing programmes and land (and other assets) redistribution programmes are likely to increase the number of households with the ability to purchase health insurance policies. Demographic factors The demographic factors include: age (+), age squared (-) and household size (-), where (.) is the postulated sign of the coefficient. The coefficients for age and age squared were statistically significant at the 5% level. Economic theory predicts that as individuals advance in age, their inherited health stock depreciates at an increasing rate (a manifestation of the biological process of ageing) and they tend to increase investments in health (including health insurance) in an attempt to decrease the rate of depreciation. This is consistent with Grossman's findings [ 24 ] that because the health stock depreciation rate rises with age, it is not unlikely that unhealthy (old) people will make larger gross investments in health than healthy (young) people. The coefficient for household size variable had a statistically significant negative effect on the likelihood of health insurance policy ownership. This finding is intuitively sensible since any increase in the household size, while holding the income constant, reduces the per capita income. Social factors The social factors include: education (+) and marital status (+). The coefficient for education was statistically significant, and had the expected positive sign. Respondents with at least a matriculation (secondary) level of education were two times more likely to be in possession of a health insurance policy than those with a lower level of education. This could be attributed to a positive relationship between a person's: (i) educational level and propensity to acquire skills; (ii) stock of knowledge and his/her market and non-market productivity [ 24 ] and earnings; and (iii) education and knowledge about the advantage of making regular small insurance payments to avoid the risk of catastrophic medical expenditures [ 28 ]. Marital status had a statistically significant positive effect on health insurance ownership. Married persons are more likely to have insurance cover than those who are single, separated or divorced. This finding is consistent with the result obtained by Harmon and Nolan [ 16 ], Rhine et al. [ 12 ], Trujillo [ 14 ] and Liu and Chen [ 28 ]. Married couples may have a higher demand for health insurance due to: (i) the need to protect their children [ 16 ]; (ii) higher combined income; and (iii) being more averse to the risk of catastrophic health expenditures than those who are single, separated or divorced. Spatial and environmental factors The spatial and environmental factors – residence (+) and environment rating (+) – had a statistically significant effect on health insurance ownership. The respondents living in formal urban settlements or rural white-owned farms had a seven times higher odds of owning a health insurance policy than those living in informal urban settlements or former rural homelands. This could partly be a reflection of the economic well being of the former group. The respondents who felt that the environment they lived in was good, very good or excellent were twenty-seven times more likely to have a health insurance cover than those who lived in fair or poor environments. This phenomenon may be a reflection of a better socio-economic status of those living in relatively affluent, formal (and cleaner) residential areas vis-à-vis informal settlements (which are relatively deprived of all kinds of social amenities). Behavioural factors The behavioural factors included in the analysis were: contraceptive use (-), alcohol use (-) and smoking (+). The three had a statistically significant effect on the demand for health insurance. The coefficient for contraceptives assumed a negative sign. This implies that the use of contraceptives may not necessarily be linked with individuals' attitudes toward risk. The coefficient for alcohol use also took a negative sign, which implies that those people who drank alcohol were less likely to purchase health insurance. On the contrary, the parameter for smoking assumed a positive sign, meaning that being a cigarette smoker, the probability of a person demanding health insurance coverage increases. In the context of the health insurance market, the latter finding could be a source of concern if it were a signal for the presence of moral hazard. Moral hazard is a potential cost of insurance in which the presence of insurance increases the tendency for losses to occur through careless, irresponsible or perhaps illegal behaviour [ 26 ]. For example, the fully insured individuals may embark on risky behaviours, such as smoking, and by so doing expose themselves to a high risk of developing various forms of cancer (throat cancer, lung cancer, etc.). Insurers attempt to control moral hazard by careful underwriting of applicants for insurance and by various policy provisions, such as deductibles (which requires an individual to pay for a certain amount of health care received before the insurance comes into effect) and co-insurance/co-payment (which requires the insured person to pay a certain percentage of eligible medical expenses in excess of the deductibles, with the insurer paying the remainder) [ 15 , 29 ]. This study used consumption of contraceptives, alcoholic drinks and cigarettes as proxies for consumers' attitudes toward health risks. The three may not be ideal proxies for risk attitudes; however, there were no better alternatives in the data set. If this were a study primarily designed to analyse the demand for health insurance, it would have been preferable to either proxy risk attitudes using a qualitative scale variable ranging from 1 (extremely risk-averse) to 10 (risk-lover) [ 11 ], directly estimate the revealed risk-aversion using experimental data [ 29 ], or ask the respondent to report whether he/she would consider paying for private health insurance at the point of demand [ 30 ]. Limitation of the study The main weakness of the current study is that since the data set upon which the analysis was based was gathered for a different purpose (i.e. it was not dedicated to health insurance), it did not contain insurance-specific attributes, e.g. premiums, co-payments, deductibles, benefits covered and the quality of care in the health facilities where the insured sought care. Thus, we had a situation where important explanatory variables were left out of the estimated regression equation 2 (in the Appendix ), leading to specification bias or omitted variable bias [ 21 ]. The omission of a relevant independent variable can change the estimated coefficient away from the true value of the population coefficient. Further research In sub-Saharan Africa, there is need for studies on the following: • The determinants of private and social health insurance policy ownership, which include both health insurance programme attributes (e.g. premiums, co-payments, deductibles) and household socio-economic characteristics (including attitudes toward risk). • The willingness and ability to pay for various forms of health insurance, including voluntary and non-voluntary insurance schemes [ 31 ]. • The economic viability of various forms of health insurance, e.g. social health insurance and community-based prepaid schemes. • Design of innovative health insurance schemes; for example, within farmers' cooperative societies [ 36 ], savings and credit societies, agricultural estates, women/men developmental groups [ 37 ], civil service, etc. • Whether the expansion of private health insurance under the current health care delivery system would yield significant public sector cost savings, and improved targeting of subsidies for the poor and preventive services [ 32 ]. • Optimal ways of curbing health insurance problems of moral hazard and adverse selection. • Benefit-incidence analysis of alternative health insurance arrangements. Conclusion The environment rating, residence, income, education, age, smoking and marital status variables were all found to have a statistically significant (at 95% confidence level) positive relationship with ownership of health insurance schemes. Contrastingly, the other covariates, namely: health rating, age squared, household size, occupation, employment, alcohol use and contraceptive use had a significantly negative relationship with health insurance ownership. There are a number of policy implications of this study: • High incomes, white-collar occupations and being gainfully employed are significant predictors of health insurance ownership. Thus, economic development (or poverty reduction) programmes geared at: (i) improving incomes of the vulnerable segments of the South African population; (ii) reducing involuntary unemployment; and (iii) creating white-collar job opportunities will empower South African women to reach a higher standard of living and in doing so boost their economic access to health insurance policies and the relevant health services. • Policies aimed at ensuring that the majority of South Africans attain a matriculation (i.e. secondary) education level will increase by almost two-fold, the probability of acquiring health insurance. • The self-assessed health status was found to have a statistically negative effect on the demand for health insurance. This implies existence of adverse selection. However, this problem can be reduced through compulsory social health insurance [ 33 ]; or through state insurance for high-risk groups, particularly the poor. Since 53% of the South African population lives below the income poverty line of US$2 per day [ 34 ], implementation of the social health insurance programme [ 7 ] would increase access to basic health services by poor. Competing interests The author(s) declare that they have no competing interests. Authors' contributions JMK recoded the raw data and participated in the development of the conceptual framework, analysis and drafting of sections of the document. LGS, BN, GMM, RC and TM participated in the development of the conceptual framework and drafting of sections of the document. All authors read and approved the final manuscript. Appendix: Empirical model A binary logit model was used in the analysis of individual household's choice between having no health insurance and having health insurance. We assumed that the expected utility associated with each health insurance option is a function of a vector of its attributes (X i ) and a vector of a household's socioeconomic characteristics (R i ), plus a stochastic error term (ε). The latter component captures errors in model specification (e.g. omission of relevant variables) and errors in data measurement. Algebraically, a household's decision process can be expressed as: EU ij = g ( X ij , R i ) + ε .................................... (1) ; where: EU ij is the utility that i th household expects to derive from choosing j th health insurance option; j = 1 if a household has health insurance; j = 2 if a household has no health insurance; and X, R and ∈ are as defined above. The basic assumption is that the i th household opts for 'health insurance' if EU i1 > EU i2 , prefers 'no health insurance' if EU i1 < EU i2, and is indifferent between the two options if EU i1 = EU i2 . Thus, the probability that i th household prefers to have health insurance is: P i1 = P(EU i1 > EU i2 ). And, conversely, the probability that i th household prefers not to have health insurance is: P i2 = P(EU i1 < EU i2 ). To determine the probability of health insurance ownership, the following model was estimated: P ij = ( α + β 1 HEALTH_RATING + β 2 ENVIRONMENT_RATING + β 3 RESIDENCE + β 4 INCOME + β 5 EDUCATION + β 6 AGE + β 7 AGE_SQUARED + β 8 RACE + β 9 HOUSEHOLD_SIZE + β 10 OCCUPATION + β 11 EMPLOYMENT + β 12 SMOKING + β 13 ALCOHOL_USE + β 14 CONTRACEPTIVES_USE + β 15 MARITAL_STATUS + ε i ........................ (2) where: P ij = 1 if individual 'i' owns insurance (j = 1) and equals zero otherwise (j = 0); ( α ) is the intercept term; (β's) are the estimated coefficients; and ε i is the stochastic error term. The explanatory variables included in the model are defined in Table 1 . Because of the limitations associated with linear probability model, the logit version of equation 2 was estimated, using maximum Likelihood Method. Pre-publication history The pre-publication history for this paper can be accessed here:	/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC553985.xml
544553	The Global HIV/AIDS Vaccine Enterprise: Scientific Strategic Plan	The development of an HIV vaccine remains one of the most difficult challenges confronting biomedical research today. A new international collaboration shares its plan to address the challenge	Introduction In June 2003, an international group of scientists proposed the creation of a Global HIV Vaccine Enterprise [ 1 ]. The authors invited discussion of this proposal, and challenged scientists to identify new strategies and mechanisms to accelerate the global effort to develop a safe and effective HIV vaccine. This paper describes the processes that led to agreement on the major roadblocks in HIV vaccine development, summarizes current scientific priorities, and describes an initial strategic approach to address those priorities. Specific research is not prescribed. Rather, the intent is to stimulate both researchers and funders to explore new, more collaborative, cooperative, and transparent approaches to address the major obstacles in HIV vaccine development identified in the plan, in addition to continuing the productive, high-quality programs already underway. The major difficulties encountered in the development of an HIV vaccine are scientific, not organizational. The motivation behind the proposal for a Global HIV/AIDS Vaccine Enterprise was the recognition that development of an HIV vaccine remains one of the most difficult challenges confronting biomedical research today [ 2 , 3 ]. Fortunately, scientific progress has created new opportunities that could be harnessed more effectively through global coordination and collaboration. These new opportunities include an expanded HIV vaccine candidate pipeline, improvements in animal models, a growing database from clinical trials, and the availability of new quantitative laboratory tools that make comparisons among vaccine studies feasible. Confronting major roadblocks and harnessing these new opportunities requires an effort of a magnitude, intensity, and design without precedent in biomedical research, with the Human Genome Project as a potentially useful model [ 4 ]. More specifically, the critical scientific insights generated by the creativity of individual investigators, as well as small groups and individual networks, could be significantly augmented by a properly organized, managed, and systematized international effort targeted on the design and clinical evaluation of novel HIV immunogens. An international collaborative effort that addresses a shared scientific plan, provides information exchange among groups, links clinical trials with standardized laboratory assays and evaluation in animal models, applies new knowledge to improvements in vaccine design in an iterative manner, and supports a transparent process for decision making in all aspects of vaccine discovery, design, development, and clinical testing will prove critical to success. The Global HIV/AIDS Vaccine Enterprise represents a novel paradigm to seek and identify international agreement on the critical roadblocks for developing an HIV vaccine and on creating a shared scientific plan that addresses those roadblocks (see Box 1 ). The Enterprise proposes to coordinate efforts at a global level, facilitate use of common tools and technologies, and help ensure access to optimized resources. Furthermore, the Enterprise approach is a way of behaving as a global community of problem-solvers, more openly sharing information, ensuring that the shared scientific plan is implemented, and basing decisions on evidence rather than advocacy. Box 1. Key Points in the Scientific Strategic Plan • More new HIV infections and AIDS deaths occurred in 2004 than in any prior year ( Figures 1 – 3 ). A vaccine is critical for the control of the pandemic. • Development of an HIV vaccine is one of the world's most difficult and important biomedical challenges. • Harnessing new scientific opportunities for HIV vaccine development will require an effort of a magnitude, intensity, and design without precedent in biomedical research. • The Global HIV Vaccine Enterprise is an alliance of independent organizations committed to accelerating the development of a preventive HIV/AIDS vaccine based on a shared scientific plan. • The scientific strategic plan was developed with the collaboration of over 140 scientists and other participants from 17 countries and several international organizations. • The plan identifies critical unanswered scientific questions along the critical path for vaccine discovery, from antigen design to the conduct of clinical trials. • Novel vaccine candidates need to be designed to induce high levels of broadly reactive and persistent immune responses against HIV strains circulating in different parts of the world. • Standardization and validation of high-throughput laboratory assays conducted under GLP will allow comparison of results from different vaccines, which is a linchpin of rational decision making in vaccine development. • The Enterprise will encourage decision makers to establish clear and transparent processes to identify and prioritize the most promising vaccine candidates. • The Enterprise will seek to engage the best researchers who are willing to work in a highly collaborative manner and to dedicate the majority of their efforts to solve the fundamental roadblocks in HIV vaccine development. • To mount an accelerated global search for a safe and effective HIV/AIDS vaccine, annual funding for such research should double—to US$1.2 billion per year. • Several founding partners of the Enterprise have already committed, or are planning to commit, new funding to support the proposed Enterprise activities, and to create a culture of mutual accountability for the effective implementation of the scientific strategic plan. • Enterprise activities are guided by an international Coordinating Committee, supported by different technical expert groups, including representatives from funders and implementers of HIV vaccine R&D. It must be emphasized, however, that the major difficulties encountered in the development of an HIV vaccine are scientific, not organizational, and arise directly from the complexities of HIV and AIDS. “Small science” should not be replaced with “big science.” Both approaches must be undertaken. Creation of research environments that support the creativity both of individual investigators and of larger, collaborative efforts will accelerate the scientific breakthroughs needed to successfully develop a safe and effective HIV vaccine. Scientific Priorities Prioritization process In August 2003, the authors of the Enterprise proposal invited a group of leading scientists, public health experts, and policy makers to meet at the Airlie House in Warrenton, Virginia, United States, to refine the vision for the Enterprise. The Airlie group agreed that the Global HIV/AIDS Vaccine Enterprise should be developed as an alliance of independent organizations committed to accelerating the development of a preventive vaccine for HIV/AIDS through implementation of a shared scientific strategic plan, mobilization of additional resources, and greater collaboration among HIV vaccine researchers worldwide [ 5 ]. The subsequent initial planning phase of the Enterprise involved leading government research agencies, private industry, non-governmental organizations, and funders involved in HIV vaccine research and development (R&D) activities, including the Bill & Melinda Gates Foundation (BMGF), the International AIDS Vaccine Initiative (IAVI), the National Agency for Research on AIDS of France (ANRS), the United States National Institutes of Health (NIH), the United Nations Joint Programme on HIV/AIDS (UNAIDS), the World Health Organization (WHO), and the Wellcome Trust. The Enterprise is expected to grow with time and include additional organizations and research groups willing to contribute to the implementation of its scientific strategic plan. A Steering Committee composed of representatives from several of the founding organizations provided guidance and coordination, with the BMGF serving as interim Secretariat. Six Working Groups involving more than 120 participants from 15 countries, the WHO, and UNAIDS were formed to develop the scientific plan of the Enterprise. These Working Groups met from January to April 2004, identified critical unanswered questions, and proposed actions to address them. In May 2004, the Steering Committee of the Enterprise analyzed the recommendations from the Working Groups and identified the scientific priorities for initial action. Several common themes emerged from the Working Groups. There was clear agreement on the key scientific challenges, as well as strong consensus that the HIV vaccine field has progressed to a point where it should be possible to answer some of the persistent questions more definitively. To meet these challenges, the Working Groups called for enhanced access to reagents and technologies, adequate resources, and strengthened human capacity in several key areas, especially in developing countries, where clinical trials need to be conducted. There was also agreement that the present way of doing business, which centers primarily on individually led research groups or networks, needs to be supplemented by establishing focused, collaborative structures and providing access to common standards and technologies, which would enable comparison of data and candidate vaccines. This would, in turn, support a rational process for decision making to advance candidate vaccines through the different phases of evaluation. Vaccine discovery One immediate goal is to design HIV candidate vaccines that consistently induce potent, broadly reactive, persistent neutralizing antibodies, as well as memory T cells that suppress viral replication and prevent escape of virus from immune control [ 6 , 7 ]. Additional research is also needed to identify how mucosal [ 8 ] and innate [ 9 , 10 ] immunity could be harnessed to develop effective HIV vaccines. The ability to develop effective vaccines would be greatly enhanced by an understanding of what specific immune response or responses correlate with vaccine-induced protection [ 11 ]. The current state of the art suggests a two-pronged strategy to accelerate the development of a safe and effective HIV vaccine. One component should center on candidate vaccines already in the pipeline, nearly all of which are designed primarily to induce T cell responses. In some animal models these T-cell-inducing candidate vaccines suppress post-infection viremia and prevent or delay HIV disease, rather than prevent infection [ 12 , 13 ]. In studies of individuals infected with HIV, viral load correlates with efficiency of transmission [ 14 ], suggesting that a vaccine capable of suppressing viral load might reduce HIV transmission. The second component should address critical gaps in scientific knowledge through carefully designed, focused, coordinated, and well-supported approaches. The fruits of this work will be a clearer understanding of what properties are needed for a successful vaccine and how to design candidates that incorporate those properties. Scientific areas in which a more collaborative and organized Enterprise approach will be beneficial include the following: vaccine design based on the characteristics of recently transmitted viruses, evaluation of immune correlates of protection in animal models, and design of novel candidates vaccines that induce neutralizing antibodies and T cell immune responses. Identifying which T cell candidate vaccine is most promising has become an urgent priority. Vaccine design Strategically, vaccines that are designed based on recently transmitted viruses hold the best hope of inducing relevant immune responses against currently circulating strains. Recent data suggest that the subset of viral strains that are sexually transmitted has unique genetic and antigenic properties, including greater susceptibility to neutralization than the bulk of circulating virus [ 15 ]. While such observations require confirmation, newly transmitted viruses are nonetheless the crucial targets of vaccine-induced immunity. Therefore, virological and immunological characterization of acute/early HIV infection should inform the design of vaccines and also guide the design of trials capable of determining whether immunization impacts virus levels and the course of HIV infection. To address these issues, a representative number of virus strains derived from recently infected individuals representing those populations who will participate in vaccine efficacy trials, including populations in developing countries, should be obtained. These virus isolates should be subjected to a comprehensive genetic and biologic characterization, together with an analysis of host immune responses and the genetic background of those populations participating in the clinical trials. Figure 1 Adults and Children Estimated to Be Living With HIV as of the End of 2004 (Total: 39.4 [35.9–44.3] million) (Map: UNAIDS/WHO) This continuous and ongoing effort will require a multidisciplinary global approach, linking investigators who are conducting epidemiological and cohort studies (to allow for detection of acute/early infections), laboratory scientists working on the virology and immunology of acute/early infection and on the genetic characterization of affected human populations, vaccine designers and manufacturers, and clinical trialists. In addition, systems for data management and analysis that will facilitate the rapid translation of new information into improved vaccine designs need to be developed. Immune correlates Nonhuman primate models of AIDS offer opportunities to evaluate potential correlates of immune protection. While a particular immunization strategy that works in animal models may or may not predict protection in humans, important insights into potential immunologic mediators of protection would result from such studies. Several experimental vaccines induce varying degrees of protection against simian immunodeficiency virus (SIV) or chimeric simian/human immunodeficiency virus in rhesus macaques. In particular, studies using models in which a very high level of protection from acquisition of infection was achieved are needed, i.e., immunization with live attenuated SIV and attenuation of SIV infection by short-term antiretroviral treatment administered immediately after SIV inoculation [ 16 , 17 ]. To facilitate this process, assays for many different immune responses to SIV and chimeric simian/human immunodeficiency virus need to be standardized, validated, and made available to different research groups. Likewise, agreements need to be reached on those monkey challenge models that most closely resemble HIV transmission and infection in humans. Large numbers of animals will be needed to achieve statistical significance for experimental findings [ 18 ], which in turn will require expanded primate breeding and housing capability. A multidisciplinary approach that links virologists, immunologists, vaccine developers, primatologists, data and project managers, and others will be needed. Neutralizing antibodies There is increasing agreement that a successful vaccine needs to induce both humoral and cell-mediated immunity. Development of immunogens capable of inducing antibodies that neutralize primary HIV isolates from all genetic subtypes and regions of the world remains the most difficult challenge in the field of HIV vaccinology [ 19 , 20 ]. Success will likely require a deeper understanding of the structural motifs of the HIV envelope protein that interact with cellular receptors and/or that are recognized by broadly neutralizing antibodies [ 19 ]. This strategy will require numerous well-characterized, broadly neutralizing monoclonal antibodies, the application of peptide and carbohydrate chemistry, structural biology, and genetic engineering approaches to immunogen design, and the use of iterative approaches guided by the immunogenicity of new designs. Given the importance of these endeavors and the uncertainty as to what path will lead to success, multiple intersecting approaches need to be explored, including, for example, the design, production, and evaluation of (1) envelope proteins that stably reveal neutralization epitopes that may be only transiently exposed during viral entry into target cells, (2) immunogens that contain rigid, stable epitopes that mimic the portion or portions of envelope recognized by broadly neutralizing monoclonal antibodies, (3) modified envelope proteins that better expose existing relevant epitopes, and (4) molecules that resemble a stabilized version of the mature envelope trimer on the virion surface. These are examples of current approaches being explored, some or all of which may prove ineffective. Additional novel ideas need to be proposed and explored. To achieve the above objectives, new tools and technologies such as those able to detect rare, broadly neutralizing monoclonal antibodies through large-scale screening of human sera will have to be developed. In addition, the very limited existing capacity to translate structural information into stable immunogen products needs to be expanded. T cell vaccines Nearly all current vaccine candidates in the clinical pipeline are T-cell-inducing vaccines, e.g., poxvirus recombinant vectors, adenoviral vectors, DNA constructs with or without adjuvants, and lipopeptides. The ongoing effort to evaluate these products and to develop new ones is considerable [ 21 ]. Identifying which T cell candidate vaccine or vaccines are most promising has become an urgent priority. However, these evaluations are being conducted within separate preclinical research groups and, to a lesser extent, separate clinical trial networks, with the result that candidate vaccines may not be optimally compared preclinically or clinically. This approach may result in delays in identifying the most promising candidates, and it risks devoting time and resources to inferior products, although it is recognized that the specific immune responses needed for a successful vaccine remain unknown. The identification and optimization of promising candidates will require (1) defining clear, transparent processes for decision making, (2) establishing agreement on vaccine characteristics upon which decisions should be based, (3) developing and using validated assays to assess those parameters, to allow for preclinical and clinical comparison among candidates, and (4) establishing closer coordination and data-sharing among product developers, which will accelerate the availability of critical information needed to identify and further develop the most promising candidates. Development of an HIV vaccine remains one of the most difficult challenges confronting biomedical research today. Research is also needed to develop improved novel T-cell-inducing candidate vaccines, especially those that avoid or otherwise circumvent anti-vector immune responses [ 22 ], and those that induce persisting high levels of immunity, especially mucosal immunity. In addition, a thorough, systematic exploration of adjuvants that markedly enhance the quantity, quality, and durability of immune responses to HIV vaccines is needed. Laboratory standardization Comparison of results from preclinical and clinical studies is the linchpin of rational decision making regarding further development of vaccine candidates. Therefore, the initiation of approaches that will permit valid comparisons is crucial. Progress to standardize and validate a limited number of T cell assays has been made within the laboratories of vaccine developers and within some partnering research networks. This approach now needs to be more broadly applied and extended to the analysis of neutralizing antibody responses. A robust infrastructure that develops, expands, and ensures broad access to quality assay technologies will allow valid comparison of data across trials and networks worldwide. In order to achieve this goal, the following are required: (1) a decision-making process to select a set of robust assays, standardized and validated across laboratories, for measuring vaccine-induced immune responses in humans and animals; (2) wide availability of common reagents (such as peptides, control sera, and virus panels); (3) capacity for developing novel assays and reagents of potential value and for their translation to preclinical and clinical settings; (4) “core” laboratories that run selected assays and serve as a reference laboratory for satellite laboratories (clinical and preclinical work would take place in separate facilities, and clinical studies would require Good Laboratory Practices [GLP] conditions); (5) satellite laboratories located at or very near clinical trial sites to carry out a range of activities such as processing blood, storing and shipping specimens, and conducting basic immunological evaluation, and to participate in other Enterprise-organized activities such as acute/early infection studies; (6) an ongoing global quality assurance function encompassing all participating core and satellite laboratories and covering both routine safety as well as immunologic and virologic assessments; and (7) transfer of research assays and, when and where feasible, validated endpoint assays to satellite labs, including the necessary training activities. In addition, new assay development has failed to keep pace with current understanding of the biology of the immune system and recent advances in technology. A more active program of applied research and assay development is needed to explore new concepts that would advance technical abilities and provide a better understanding of the immune responses generated by HIV vaccines. Cellular immunity Two assays are currently used for the primary evaluation and enumeration of antigen-specific T cells: Interferon-γ ELISPOT and multiparameter flow cytometry. The ELISPOT assay was initially developed to measure CD8+ T cell responses. Several observations in both mice and humans have indicated that protective immune responses will likely require stimulation of both CD4+ and CD8+ T cell effector and memory functions; it is unlikely that induction of Interferon-γ-secreting T cells alone correlates with protective immunity [ 11 ]. Therefore, additional laboratory assays measuring multiple HIV-specific cell types as well as functional capabilities will be needed to thoroughly evaluate vaccine-induced immune responses. These assays should also permit rapid assessment of the magnitude and breadth of immune responses, and enumerate the specific epitopes that are recognized. Figure 2 Estimated Number of Adults and Children Newly Infected with HIV during 2004 (Total: 4.9 [4.3–6.4] million) (Map: UNAIDS/WHO) Humoral immunity Different laboratories use different assays to measure antibodies that neutralize HIV and related viruses, SIV and chimeric simian/human immunodeficiency virus. These assays vary technically, but the most widely accepted assays measure reduction in virus infectivity in cells that express the receptors necessary for virus entry. Assays that offer the greatest value are those that are validated, amenable to high throughput, low in cost, readily transferable, and that can be performed according to GLP guidelines. The ability to measure the magnitude and breadth of neutralization against diverse HIV strains is essential to evaluating responses generated by candidate HIV vaccines. Only with multiple strains of virus can neutralization breadth be ascertained in a meaningful way. Standard panels of HIV strains are in early stages of development. Expansion or extension of current standardization and validation activities, production and provision of necessary reagents, and access to quality assurance programs are needed to ensure worldwide comparability of assay results [ 23 ]. The strains of virus incorporated into a worldwide panel need to be carefully selected to reflect the current epidemic and should include early isolates from individuals at potential vaccine trial sites [ 24 ]. Molecular epidemiological studies and elucidation of the role of genetic factors and immune responses of the host in the transmission of HIV at the population level will also help guide vaccine design and evaluation [ 25 , 26 ]. Another specific priority is an assessment of the neutralizing antibody response generated in the recently completed Phase III trials of HIV envelope glycoprotein 120 candidate vaccines using a global virus panel. The results would establish a baseline level of neutralization potency and breadth that is non-protective, which would be extremely valuable in reaching informed decisions about advancing future antibody-based candidate vaccines. As more HIV candidate vaccines enter the pipeline, current capacity will be rapidly exhausted. A major obstacle to designing a suitable global virus panel is the paucity of information on neutralization serotypes. There is general agreement that if a reasonably small number of neutralization serotypes exist, their identification would guide the creation of an optimal panel of isolates for neutralizing antibody assays and the design of polyvalent immunogens. Although there is some controversy as to whether HIV-1 neutralization serotypes exist, the magnitude of benefit that would result if serotypes were identified warrants establishment of a neutralization serotype discovery program that employs the latest technologies. Product development and manufacturing Manufacture of vaccine candidates for large clinical trials and to meet eventual worldwide demand requires the development of processes for producing consistent, active vaccine batches on a large scale. Development of these bioprocesses must be integrated with analytical work (e.g., toxicity and stability testing), incorporate validated assays, and be applicable to the manufacture of sufficient vaccine to meet global needs after licensure. These processes are typically individually developed as a candidate vaccine advances from early clinical testing to late-stage evaluation and licensure. Worldwide expertise and capacity for this bioprocess development work is already limiting and exists almost exclusively in the private sector. As more HIV candidate vaccines enter the pipeline, current capacity will be rapidly exhausted. The initial priority is to identify or establish one or more dedicated HIV vaccine bioprocess and analytical development groups that bring together the skill set and capacity to manufacture different promising candidates for clinical trials. The bioprocess development groups would also help train people and transfer manufacturing skills in whole or in part to manufacturing sites around the world. This training program would address the acute shortage of bioprocess experts. At a later stage, building, acquiring, or contracting facilities to carry out bioprocess and analytical work and to produce several different types of candidate vaccines should be considered. Such facilities would further assist in transferring manufacturing technology to other production facilities, preferably in one or more developing countries. Decisions about which candidates a facility undertakes would be made through a well-defined, comprehensive evaluation process. The facilities could eventually be expanded to provide production capacity to launch a vaccine for public health use, should no manufacturer be available to produce the vaccine quickly upon licensure. Clinical trials capacity As a growing number of HIV candidate vaccines begin to move through the clinical trials pipeline, the gap between existing global capacity and future requirements for conducting large efficacy trials has grown in magnitude and urgency, especially in developing countries. This gap in developing countries must be addressed through (1) increasing the quantity and quality of research staff, (2) establishing sustainable research facilities to support trials, and (3) expanding access to large, well-defined populations of uninfected people at high risk of HIV infection. The acute shortage of qualified personnel is a major bottleneck to the conduct of clinical trials in developing countries. The recommended solutions take a long-term view and are aimed at building site capacity rather than preparing for specific trials. Sites should not be confined to conducting HIV vaccine trials but should be positioned to contribute to other research of public health importance to the community and the country, including, for example, other areas of HIV research (e.g., microbicides and treatment) and/or other diseases. Additional field trial sites must be developed to be able to conduct planned and anticipated efficacy trials. Sites should be selected in a strategic, data-driven manner, and should demonstrate the ability to recruit and retain large numbers of HIV-negative volunteers from populations with substantial HIV incidence. New efficacy trial sites should be developed in regions with emerging epidemics rather than only in areas with already-established disease. “Early-warning systems” must be available to identify these newly emerging sub-epidemics. Defining optimal methods for collection of HIV incidence data from populations at potential efficacy trial sites is essential. Whenever possible, efficacy trial sites should be linked to (1) academic medical centers to enhance research capacity and help train clinical researchers, (2) accredited local and regional laboratory facilities to provide infection endpoint and safety assessments, and (3) centers that can provide appropriate care and treatment to trial participants. Figure 3 Estimated Adult and Child Deaths from AIDS during 2004 (Total: 3.1 [2.8–3.5] million) (Map: UNAIDS/WHO) The acute shortage of qualified personnel is a major bottleneck to the conduct of clinical trials in developing countries with severe or rapidly emerging HIV epidemics. Development of intellectual capacity at these sites should emphasize (1) expanding research training opportunities for personnel in the broad range of topics required to conduct high-quality clinical research, (2) establishing and adequately supporting long-term career paths for such individuals, and (3) fostering political and social environments locally and nationally that support the conduct of clinical research. Building HIV scientific and operational expertise at clinical trial sites should be linked to other HIV/AIDS research activities (e.g., identifying and characterizing incident/early HIV infections, collecting newly transmitted strains, and measuring incidence in high-risk populations). Site development must include strategies to develop or enhance existing capacity to deliver health care, including HIV prevention, care, and treatment, to the local community participating in clinical trials. Provision of, or referral to, basic clinical services such as voluntary counseling and testing and diagnosis and treatment of sexually transmitted infections will be essential. In addition, site development should include building skills that are ancillary but critical to the actual conduct of clinical trials, such as educating communities, building community partnerships, managing site finances, and piloting applications through regulatory decision-making processes. Regulatory considerations The Enterprise must address a number of problems that currently impact the review of HIV vaccine trial protocols and that could delay future decisions regarding product licensure in developing countries. Most regulatory challenges arise from the fact that regulatory approvals are granted at the national level, but many developing countries lack the expertise, well-defined processes, clear delineation of authority, and/or other system components needed to make regulatory decisions expeditiously. As a result, new products are often licensed in these regions based on prior approval in the US or Europe and/or endorsement by the WHO. Under these circumstances, data specific to developing country populations (e.g., disease burden or childhood vaccination schedules) often do not enter into the decision making. The absence of defined pathways to approve products targeting a country's needs when a product is not also submitted to regulators in the US or Europe remains another obstacle. The Enterprise process has identified these action-item priorities: (1) harmonize and exchange information needed by regulatory bodies within the differing legal frameworks of different countries, (2) facilitate regulatory decision making, possibly using regional approaches for conducting reviews and making recommendations, (3) build regulatory capacity, (4) perform risk/benefit evaluations in the context of differing epidemic dynamics and country needs and resources, (5) identify and remove potential scientific impediments to rapid regulatory decision making, and (6) address ethical issues that interface with regulatory decision making, such as ensuring informed consent and defining the degree to which trial participants should receive a standard of care that is higher than others in their community. Intellectual property issues Given the Enterprise focus on stronger collaboration, data sharing, and use of common materials and reagents, an intellectual property (IP) framework that facilitates this “enabling environment” is crucial for success. While IP issues may arise throughout the vaccine development process, at present the top priority is to stimulate early stage research and vaccine design by increasing scientific freedom to operate and sharing of data and biological materials. Specific areas for further consideration include: (1) minimizing restrictions on freedom of operation, perhaps by early stage covenants not to litigate and followed by later stage agreements based on true valuations of IP; (2) sharing of information (including clinical trial data), materials, expertise, trade secrets, and platform technologies in a protected and secure manner while also remaining in compliance with national laws devised to prevent monopolies and insider trading; (3) recognizing the contribution of different countries to HIV vaccine development through approaches that assure affordable access to successful vaccines; and (4) maximizing access to essential technologies and inventions. Scientific Plan Scientific activities On October 21, 2004, a group of participants from 16 countries, the European Commission, UNAIDS, and the WHO met to finalize the scientific plan and to discuss how to formulate specific actions. Participants noted that the structure of an activity should depend on several factors, including, for example, the degree to which the activity can be predefined, the degree to which the creativity of academic researchers needs to be harnessed, and the mechanisms available to the funding organization. A number of options were discussed, with consensus as to those that would fit various scientific priorities. First, networks of focused consortia and real or virtual centers are well suited to systematically address many of the major scientific roadblocks identified in this plan. These consortia or centers would link to each other to ensure a comprehensive, systematic approach, sharing information so that each can be as productive as possible, and also to share reagents and procedures so that data among groups can be compared and, where possible, merged for analysis ( Figure 4 ). The specific scientific areas that could be supported by consortia or centers include (1) addressing fundamental scientific problems, such as the definition of correlates of immune protection in selected animal models and the characterization of acute/early infection in potential vaccine trial sites; (2) designing and evaluating novel vaccines, such as immunogens that neutralize primary isolates, and improved T cell vaccines that avoid immunological escape and/or that induce persisting mucosal or persisting systemic responses; and (3) providing for a systematic evaluation of potential adjuvants. The success of consortia or virtual centers will depend on engaging the best researchers, getting them to work collaboratively and dedicate the majority of their effort to HIV vaccine research, resolving IP issues, obtaining support for researchers from their institutions, and keeping the group focused on specific, well-defined questions. More than one consortium may be needed for systematic coverage of vaccine design research (e.g., monoclonal-antibody-identified epitopes, native envelope, and modified envelope). Figure 4 A Possible Model to Address Key Scientific Questions through an Appropriate Organizational Infrastructure (Courtesy of John Mascola; illustration: Giovanni Maki) Second, a global system of central laboratories linked to satellite laboratories that work together (using GLP) would provide a range of standardized functions, help ensure the quality of clinical research, and enable comparison of data from different trials ( Figure 5 ). Together this system could (1) conduct preclinical or clinical assays, particularly critical endpoint assays that require standardization and/or validation; (2) develop, optimize, and validate new assays and platforms; (3) transfer assays from central labs to satellite labs; (4) develop and implement a global quality control/quality assurance program and proficiency testing for assays performed at central and satellite laboratories; (5) implement vaccine-related research that requires validated assays and close cooperation and collaboration among labs globally, such as a Virus Neutralization Serotype Discovery Program, and the characterization of recently transmitted HIV isolates; and (6) contribute to the development of technological infrastructure in developing countries. Figure 5 A Possible Model for a Comprehensive Global Laboratory Network for the Standardized Assessment of Humoral Immune Responses (Courtesy of David Montefiori; illustration: Giovanni Maki) Third, a number of contract laboratories capable of developing, acquiring, storing, and distributing common reagents will prove critical to the success of collaborative research and development projects, and to ensuring reagent quality. These reagents could include (1) peptides, antisera/antibodies, and viral isolates for immune assays, including a standard panel of virus strains and sera representative of the global genetic and immunologic variability of HIV, and (2) additional broadly neutralizing monoclonal antibodies, especially from non-clade B viruses, to facilitate elucidation of the motif or motifs they recognize. These contract laboratories would be expected to work very closely with and enable the work of Enterprise consortia, centers, immune assessment laboratories, and clinical sites. Fourth, a network of Clinical Research Training Centers in developing countries could work collaboratively to ensure development of quality trial sites. These centers would (1) conduct or facilitate training of trial site personnel in activities that are generic to the conduct of clinical trials, as well as those specific for HIV vaccine trials, for example, an HIV vaccine fellowship program for developing country scientists; (2) coordinate and work together with other Enterprise consortia or centers, such as those established to characterize acute/early infection in developing country settings or to prepare a standard panel of HIV strains representative of currently circulating viruses; and (3) share standard operating procedures, vaccine development plans, and strategies for engaging and ensuring community and political support. Fifth, a network of individuals and companies with manufacturing experience, particularly process development expertise, could link to consortia, centers, and others involved in vaccine development to provide development and manufacturing expertise to facilitate the advancement of improved HIV vaccine candidates. The above structures are proposed to address the initial Enterprise scientific priorities. Additional consultative groups, reference and centralized facilities, and other mechanisms may be needed to facilitate collaborative work and strengthen the global capacity for the conduct of HIV vaccine research and development as the field progresses. Different implementing and funding agencies will need to work in close collaboration to ensure harmonious implementation of the scientific plan. Initial actions should focus on the areas of vaccine discovery and standardization of laboratory assays, which are considered critical for the success of the Enterprise and the eventual development of a safe and effective HIV vaccine. Activities to address recommendations in the areas of product development and manufacturing, clinical trials capacity, regulatory considerations, and IP issues should be launched after these initial components of the plan are under way. Regardless of timing, each scientific endeavor needs to outline specific strategies to ensure information exchange and capacity building among the collaborating partners and institutions. The funding mechanisms employed (i.e., contracts, grants, interagency agreements, etc.) will depend on the task to be accomplished and the needs and capabilities of each funding organization. In the spirit of coordination, collaboration, and transparency promoted by the Enterprise, two or more partners may jointly support one or more activities, taking care to avoid duplication in the use of their respective resources. When a research area is jointly funded, all communication regarding goals, research plans, progress, obstacles, etc., should be openly and transparently shared among all stakeholders—funders, project managers, and researchers. Guiding principles As an alliance of independent entities, the Global HIV/AIDS Vaccine Enterprise will be challenged to carry out three essential functions. One is to continue regular scientific assessments. The scientific priorities outlined in this paper will need to be monitored, re-evaluated, and updated. An evolving scientific plan must reflect lessons learned, new opportunities, and the influence of new scientific findings and new technologies. Revised versions of the scientific plan must be made fully and publicly available. The second essential function is to establish global processes. To optimize progress across a large and complex set of activities at the global level, standards, performance criteria, and processes for data sharing, communication, and convening must be established. The Enterprise will convene fora to address policy issues such IP, clinical trials, site development, and regulatory hurdles. And the third essential function is shared accountability. The partners in this alliance will need to create a culture of mutual accountability for the effective implementation of the scientific strategic plan. Since the Enterprise is not a single organization, a shared “way of doing business” is one of its most important defining traits. Articulating an explicit set of “working principles” is therefore crucial to the identity and smooth functioning of the Enterprise. For the Enterprise as a whole the following conditions apply: (1) the central task is to develop and implement an ambitious scientific plan with the necessary scale, balance and sequence of activities, and structure to carry it out; (2) the plan must focus on critical roadblocks that would benefit substantially from global collaboration while fostering continued R&D by individuals, small groups, and individual networks; (3) the incentives holding the alliance together will include collaborative arrangements and structures that give people the resources, necessary critical mass, centralized facilities, common reagents, assays and technologies, and data they need to effectively remove critical roadblocks; (4) all activities will reflect the commitment to create an environment that maximizes the ability of participants to share data and biological materials, e.g., through the use of common standards for measurements and appropriate IP arrangements; and (5) the Enterprise also commits to working for rapid global access to a successful vaccine. For participating investigators and organizations, key principles include (1) the willingness and desire to work in an open, collaborative fashion, sharing data and reagents in a collegial fashion, with the appropriate balance between productive competition and effective collaboration, and (2) the willingness and ability to devote the majority of their time to tackling these problems within a focused environment, completely committing to solve the problems at hand. Organizational structure of the Enterprise The implementation of the scientific plan of the Enterprise will be overseen and supported by the organizational structure described in Figure 6 . Figure 6 Proposed Organizational Structure of the Global HIV/AIDS Vaccine Enterprise (Illustration: Giovanni Maki) The Coordinating Committee will facilitate all aspects of the Enterprise's activities. This committee consists of representatives of the Enterprise founders as well as additional scientific leaders selected from inside and outside the field of HIV vaccine research and development. The committee will develop procedures for term rotation and inclusion of new members, to ensure appropriate representation of all relevant partners, and will engage external stakeholders for advice, expertise, and assistance, appointing technical expert groups as needed. A Secretariat will provide logistical and administrative support to the Coordinating Committee and Enterprise partners. The BMGF will serve as Interim Secretariat until a permanent Secretariat is established. The road to success will be a bumpy one. The Funders Forum will be an open forum of sovereign, independent funding organizations, starting with a nucleus of those who already embrace the principles of the Enterprise and who are actively supporting or intend to support and fund HIV vaccine research and development. Members of the Funders Forum will be high-level decision makers within the ranks of funding organizations and governments, as close as possible to the source of resources. Since the Enterprise is not a discrete organization with a pool of money, funders will support specific areas using their own mechanisms, according to their own practices and policies, and following Enterprise principles. The scientific plan will provide guidance that may help funders better align existing resources but, more importantly, will facilitate the efficient and focused application of new resources as they become available. Multiple funders who wish to support a single Enterprise-defined project could form collaborative agreements, memoranda of understanding, or other forms of written agreement among themselves to outline their respective roles and responsibilities; address IP, program management, oversight, and other issues; and establish mechanisms for communication and conflict resolution. The funders with greatest flexibility could provide incentives for sharing reagents and data, and linking projects together, e.g., by supporting the additional work that nationally or regionally funded laboratories would need to undertake in order to participate in a global network, or by supporting a program to develop and share reagents. In some cases, funders may wish to support an implementing organization that will take responsibility for managing the project and reporting back to the funder and other stakeholders. In other cases, funders may have the capability and capacity to play a substantial role in facilitating the project. In still other cases, funders may have the capability to assume a leadership role in overseeing the conduct of the activity, particularly in cases where the activity is well defined in advance. In addition, an Annual Stakeholders Forum will be organized to bring together the broader community of scientists, policy makers, public health officials, and community representatives involved in the search for an HIV/AIDS vaccine. This meeting will serve as a forum to (1) update the broader community on Enterprise activities and progress, and (2) provide the community with a mechanism for feedback and dialog. Funding issues Global expenditures on HIV vaccine research and development in 2002 were tentatively estimated to be on the order of US$624–670 million, the large majority (67.3%) provided by the public sector, followed by the philanthropic sector (17.4%) and industry (15.3%). An analysis of how those funds have been invested revealed that the large majority (43.1%) is being used in preclinical research activities, followed by clinical trials (28.2%), basic research (20.7%), cohort development and clinical trial infrastructure (6.5%), and vaccine education, advocacy, and policy development (1.4%) [ 27 ]. The largest funder of HIV vaccine research and development activities has been the NIH, with almost US$350 million in 2002. The NIH budget for HIV vaccine research has grown from less than US$50 million in 1996, to an estimated US$514.6 million for 2005, corresponding to 17.6% of the NIH total HIV-related research budget for 2005. The Enterprise Coordinating Committee will analyze the additional financial requirements to fully implement the scientific plan of the Enterprise, and the Enterprise Secretariat will explore options to leverage these funds from the public and private sector. Initial estimates by Enterprise partners suggest that US$1.2 billion per year, or double the current expenditures on HIV vaccine research and development, will be needed. Although this amount may appear unrealistic at present, it would represent only a fraction of the total global expenditures in response to the AIDS pandemic and a very reasonable investment in view of the enormous social, political, and economic consequences of the pandemic. However, it is essential that the proposed increase in funding for HIV vaccine R&D be additional to existing AIDS expenditures, and not at the expense of current prevention, treatment, and care efforts. The founding partners of the Enterprise, including the NIH, the BMGF, and the Wellcome Trust have already committed, or are considering committing, resources towards new initiatives that will begin to enact portions of the Enterprise scientific plan over the next six to nine months. Each funder will utilize their own funding processes and will align the design, scope, and scale of programs to those laid out in this plan. For example, the NIH National Institute of Allergy and Infectious Diseases will establish the Center for HIV Vaccine Immunology, which will target several scientific priorities identified here. Political support As a sign of global recognition of the importance of better, more strategic coordination in the search for an HIV vaccine, the “Group of Eight” leading industrialized nations in June 2004 endorsed the goals of the Enterprise and agreed to review progress in implementation at its 2005 summit meeting in the United Kingdom [ 28 ]. Likewise, on October 19, 2004, Ministers of Health from seven European countries (France, Germany, Italy, the Netherlands, Spain, Sweden, and the United Kingdom) adopted a statement of intent to coordinate efforts to accelerate research for an HIV vaccine within the context of the global effort. Next Steps With almost 5 million new HIV infections and 3 million AIDS deaths occurring every year worldwide, the development of a safe, effective, and accessible HIV vaccine represents one of the most urgent global public health needs. This global emergency led to the proposal to harness the power of science to find a definitive solution to one of the most catastrophic health problems of our time. The Global HIV/AIDS Vaccine Enterprise has evolved over the past 18 months from a concept proposed in a scientific journal by a cadre of researchers to a global consensus concerning the major scientific roadblocks facing HIV vaccine development, a strategic approach to address those roadblocks, and guiding principles for the plan's implementation in a manner and degree commensurate with the challenges at hand. Several organizations have already embraced the Enterprise concept and are moving to tackle portions of the scientific plan. Still, much more remains to be done. The road to success will be a bumpy one requiring the energy, commitment, and action of a wide number of government and non-governmental organizations globally. Recognizing the enormity of the roadblocks as well as the potential benefits of a safe and effective HIV vaccine, it is essential that many more organizations and agencies contribute additional expertise and resources and work together as a global community in a cooperative, collaborative, and transparent manner to fully implement the Enterprise scientific plan.	/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC544553.xml
549063	Computational prediction of human metabolic pathways from the complete human genome	A computation pathway analysis of the human genome is presented that assigns enzymes encoded by the genome to predicted metabolic pathways. This analysis provides a genome-based view of human nutrition.	Background The human genome is a blueprint, but for what machinery? One approach to understanding the complex processes encoded by the human genome is to assign its enzyme products to biochemical pathways that define regulated sequences of biochemical transformations. Pathway and interaction assignments place genes in their larger biological context, and enable causal inferences about the likely effects of mutations, drug interventions and changes in gene regulation. They are a first step toward quantitative modeling of metabolism. Assignment of genes to pathways also permits a validation of the human genome annotation because patterns of pathway assignments spotlight likely false-positive and false-negative genome annotations. For example, false-negative assignments appear as pathway holes: missing enzymes within a pathway that are likely to be hiding in the genome. SRI's Bioinformatics Research Group has developed a pathway-bioinformatics technology called a pathway/genome database (PGDB), which describes the genome, the proteome, the reactome and the metabolome of an organism. A PGDB describes the replicons of an organism (chromosome(s) or plasmid(s)), its genes, the product of each gene, the biochemical reaction(s), if any, catalyzed by each gene product, the substrates of each reaction, and the organization of those reactions into pathways. Pathway Tools is a reusable software environment for constructing and managing PGDBs [ 1 ]. It supports many operations on PGDBs including PGDB creation, querying and visualization, analysis, interactive editing, web publishing, and prediction of the metabolic-pathway complement of an organism. The power of Pathway Tools is derived from both its database schema, and its software components. Both were originally developed for the EcoCyc project [ 2 , 3 ]. A PGDB can be thought of as a symbolic computational theory of a species' metabolic functions and genetic interactions [ 4 ], encoding knowledge in a manner suitable for computational analysis. Indeed, once an organism's genome and biochemical network are encoded within the schema of a PGDB, new possibilities for symbolic computational analysis arise, because many important semantic relationships are described in a computable fashion. PathoLogic is one of the Pathway Tools software components. Its primary function is to generate a new PGDB from an organism's annotated genome. PathoLogic predicts the metabolic pathways of the organism, providing new global insights about its biochemistry, and generates reports that summarize the evidence for the presence of each predicted metabolic pathway. We used PathoLogic to generate HumanCyc, a PGDB for Homo sapiens , from the annotated human genome. The genome data used as input to PathoLogic combined data from the Ensembl database [ 5 ], the LocusLink database [ 6 ] and GenBank [ 7 ]. Our analysis assigns 2,709 human enzymes to 135 predicted metabolic pathways. It provides a genome-based view of human nutrition that associates the essential dietary requirements of humans that were previously derived mainly from animal and tissue extract studies to a set of metabolic pathways whose existence is derived from the human genome. The analysis also identifies probable omissions in the human-genome annotation in the form of pathway holes (missing enzymes within the predicted pathways); we have identified putative genes to fill some of those pathway holes. This paper describes the generation of HumanCyc, and presents an analysis of the human metabolic map. The computationally predicted pathways are consistent with known human dietary requirements. We compare the predicted human metabolic pathway complement to the pathways of Escherichia coli and Arabidopsis thaliana and identify 35 pathways that are shared among all three organisms, and therefore define an upper bound on a potential set of universally occurring metabolic pathways. Results Prediction of human metabolic pathways We applied PathoLogic to the input files containing the H. sapiens annotated genome, as described in Materials and methods, generating HumanCyc. Table 1 shows the results of PathoLogic's enzyme matching during the PGDB automated build. This computational matching process found more than 2,300 matches between gene products in the annotated genome and reactions in MetaCyc. Both the ambiguous matches (row 3 in Table 1 ) and the proteins labeled as 'probable enzymes' by PathoLogic (row 5) were examined manually; about half of them were manually matched to enzymes, as explained in Materials and methods. Sometimes one gene product is matched to more than one reaction, as happens with multifunctional enzymes (for example, the gene product shown in Figure 1 would be matched to two different reactions). So the number of matches is higher than the number of proteins matched. The 'Unmatched' row includes human proteins that are not enzymes. Table 2 shows statistics from version 7.5 of HumanCyc (released in August 2003), after manual refinement of the PGDB was completed. The 2,742 enzyme genes in HumanCyc correspond to 9.5% of the human genome, and can be subdivided into 1,653 metabolic enzymes, plus 1,089 nonmetabolic enzymes (including enzymes whose substrates are macromolecules, such as protein kinases and DNA polymerases). Our best estimate of the total number of human metabolic enzymes is the sum of the 1,653 known enzymes plus the 203 pathway holes, for a total of approximately 6.5% of the human genome allocated to small-molecule metabolism (compared to 16% of the E. coli genome). Of the 1,653 metabolic enzymes, 622 are assigned to a pathway in HumanCyc, and the remainder are not assigned to any pathway; we expect that in the future some of the latter group of enzymes will be assigned to some known human pathways not yet in HumanCyc, and to some human pathways that remain to be discovered. Of the metabolic enzymes, 343 are multifunctional. The number of enzymes is less than the number of enzyme genes because, in many cases, the products of multiple genes are required to form one active enzyme complex. Table 3 shows all pathways present in HumanCyc, arranged according to the MetaCyc pathway taxonomy. Only the top two levels in the taxonomy are shown for the sake of brevity. The 135 metabolic pathways in HumanCyc is a lower bound on the total number of human metabolic pathways; this number excludes the 10 HumanCyc superpathways that are defined as linked clusters of pathways. The average length of HumanCyc pathways is 5.4 reaction steps. Example HumanCyc pathways are shown in Figures 2 and 3 . All HumanCyc pathways can be accessed online from the HumanCyc Pathways page [ 8 ]. HumanCyc 7.5 contains 1,093 biochemical reactions, 896 of which have been assigned to one or more of the 2,709 enzymes in HumanCyc. There are more enzymes than reactions because of the existence of isozymes in the human genome. This leaves 203 reactions that have no assigned enzyme. These reactions correspond to the above-mentioned pathway holes for the HumanCyc pathways. Of the 896 reactions that have assigned enzymes, 428 have multiple isozymes assigned. Filling holes in HumanCyc pathways The PathoLogic-based analysis of the annotated human genome inferred 135 metabolic pathways. A total of 203 pathway holes (missing enzymes) were present across 99 of these pathways; that is, 38 pathways were complete. Using our hole-filling algorithm [ 9 ], no candidate enzymes were found for 115 of the 203 pathway holes. For the remaining 88 pathway holes, candidates were obtained and evaluated. In 25 of these 88 cases putative enzymes were identified with sufficiently strong support that the enzyme and pathway annotations within HumanCyc have been updated to reflect these findings. See the HumanCyc release note history [ 10 ] for a list of these 25 hole fillers added to HumanCyc version 7.6. The original annotations of the human proteins that were identified as candidate hole fillers fell into several classes: A description of each class is presented below, with examples included for some. Open reading frames (ORFs) with no assigned function (6 candidates) Putative enzymes were identified, for example, for the N -acetylneuraminate lyase (LocusLink ID 80896), aldose 1-epimerase (LocusLink ID 130589) and imidazolonepropionase (LocusLink ID 144193) reactions. In each of these cases, the function of the protein was previously unknown. Proteins assigned a nonspecific function (7 candidates) The pathway hole filler assigned an enzyme previously annotated with a general function. For example, 'amine oxidase (flavin-containing) B' (LocusLink ID 4129), was assigned to a more specific reaction, putrescine oxidase. A 'fatty acid synthase' (LocusLink ID 54995) was identified to fill the 3-oxoacyl-ACP synthase reaction. Proteins assigned a single function but which our analysis indicates are multifunctional (9 candidates) In these cases the program is postulating an additional function for a gene that already has an assigned function. The pathway hole filler identified the enoyl-CoA hydratase enzyme (LocusLink ID 1892) as a potential hole filler for the 3-hydroxybutyryl-CoA dehydratase reaction in the lysine degradation and tryptophan degradation pathways. The dihydrofolate synthase hole in formylTHF biosynthesis was filled by the enzyme (LocusLink ID 2356) catalyzing the folylpolyglutamate synthase reaction. Proteins that may have been assigned an incorrect specific function Although our analyses of other pathway/genome databases have revealed examples we consider to have been assigned an incorrect function in the original annotation, our analysis of the 25 HumanCyc pathway holes that we filled revealed no candidates in this category. The pathway hole filler not only identifies candidate proteins for each pathway hole, but also determines the probability that each candidate has the desired function. Table 4 displays the homology-based features used by the pathway hole filler to compute this probability. The table shows three example reactions, each with two candidate enzymes and the data gathered for each. The columns in the table display the computed probability that the candidate has the desired function; the number of query sequences that hit the candidate (number of hits); the E-value for the best alignment between the candidate and a query sequence (best E-value); the average rank of the candidate in the lists of BLAST hits; and the average percentage of each query sequence that aligns with the candidate. In the first example, 28 imidazolonepropionase sequences from other organisms were retrieved from Swiss-Prot and the Protein Information Resource (PIR). Using BLAST, each sequence was used to query the human genome for candidate enzymes. Protein A was found in all of the 28 lists of BLAST hits. From the numbers in the table, it is fairly obvious that protein A is more likely to catalyze the imidazolonepropionase reaction than is protein B. In the second example, given the best E-value (1e-110) it is again not surprising that the computed probability that protein C has N -acetylglucosamine-6-phosphate deacetylase activity approaches 1.0. In the last example, both proteins have excellent BLAST E-values; in fact, the E-value for protein F indicates a better match with the query sequences than the E-value for protein E. In this case, protein E is found in 19 lists of BLAST hits versus four for protein F, and on average aligns with a much larger fraction of each query sequence. When examined in more detail, we discover that the four query sequences that identified candidate F in their BLAST output are multifunctional proteins with both aldose-1-epimerase activity and UDP-glucose 4-epimerase activity. Protein F aligns with the amino-terminal region of each of the four query sequences, and has no detected similarity in the carboxy-terminal regions. The UDP-glucose 4-epimerase activity lies in the amino-terminal region of each multifunctional query protein. Nutritional analysis of the human metabolic network Nutritional requirements and their genetic and biochemical basis are thought to have evolved principally in prokaryotes, over billions of years [ 11 ]. Specific nutritional challenges have driven the evolution of metabolic pathways and the functional capabilities mediated by them. Indeed, eukaryotic life acquired the basic building blocks of metabolism, that is, sets of genes encoding enzymes that mediate specific metabolic pathways, from prokaryotic ancestors. One may define a metabolic pathway as a conserved set of genes that endow an organism with specific nutritional/metabolic capabilities, for example, the ability to grow in the absence of phenylalanine because of the ability to synthesize phenylalanine. Current knowledge of human nutrition based on metabolic pathways is derived from various sources. One is clinical observation of inherited human metabolic diseases and nutrient deficiency states. For some pathways, like oxidative phosphorylation and the TCA cycle, direct studies of human tissues, such as human muscle biopsies, have been made. Nuclear magnetic resonance (NMR) has been used directly on humans to study aspects of carbohydrate and energy metabolism. Stable isotopes have been used to trace human metabolism, from which inferences about nutrition have been made. Dietary studies have been made in experimental mammals such as rats and mice and metabolic pathways experimentally elucidated in model organisms. Here we compare previously accepted human nutritional requirements with pathways derived from the human genome to evaluate their agreement. For example, biosynthetic pathways for essential human nutrients, that is, substances that must be provided in the diet such as the essential amino acids and vitamins, would not be expected to occur in the human genome. Integration of human genome data with clinical, biochemical, physiological and other data obtained both directly from humans and indirectly from model organisms should, over time, lead to a deeper understanding of human metabolism and its nutritional implications in health and disease. When the genome sequences of individuals are available, it may be possible to address questions about the variation in optimal nutrition from person to person. Explicit identification of specific areas of inconsistency will serve to focus ongoing experimental efforts to elucidate the molecular basis of human nutrition and metabolism. For all of the nine amino acids essential for humans, PathoLogic did not predict the presence of a corresponding biosynthetic pathway (see Table 5 ) [ 12 ]. And for all of the 11 nonessential amino acids, PathoLogic did predict the presence of a corresponding biosynthetic pathway. For 12 of 13 essential human vitamins, PathoLogic did not predict the presence of a corresponding metabolic pathway (note that PathoLogic could not have predicted such a pathway for six of those vitamins because MetaCyc does not contain such a pathway). PathoLogic did predict the presence of a pathway called 'pantothenate and coenzyme A biosynthesis pathway', which is not expected given that pantothenate is an essential human nutrient. However, examination of the predicted pathway reveals that no enzymes in the first part of the pathway (biosynthesis of pantothenate) are present; all enzymes are in the portion of the pathway that synthesizes coenzyme A from pantothenate. Thus, this false-positive prediction can be attributed to the fact that MetaCyc does not draw a boundary between what should probably be considered two distinct pathways. No hard-and-fast rules are generally accepted as to how to draw boundaries between metabolic pathways; therefore the PathoLogic method cannot produce objective and well accepted pathway boundaries (nor can any other known algorithm). Comparative analysis of the metabolic networks of human, E. coli and Arabidopsis Table 6 indicates whether or not each HumanCyc pathway is present in the EcoCyc E. coli PGDB and in the AraCyc PGDB for A. thaliana [ 13 ]. More precisely, we say a pathway is shared among multiple PGDBs if the same MetaCyc pathway has been predicted to be present in each PGDB; that is, if the pathway has exactly the same set of reactions in the PGDBs (the unique identifier of the MetaCyc pathway is reused in any PGDB to which the pathway is copied). The comparison does not consider how many pathway holes are in the PGDBs, but relies on the PathoLogic prediction (plus subsequent manual review) that the pathway is present; that is, if PathoLogic determines that the pathway is present despite its holes, the comparison considers it to be present. Note that we do not count the presence of related pathway variants; that is, if organism A contains pathway P and organism B contains a variant of P, we do not score this case as a shared pathway. Some shared pathways will include pathway holes. Figure 4 shows how the three metabolic networks intersect by means of a Venn diagram, depicting each PGDB's pathway complement as a circle. The number within a given intersecting area denotes the number of pathways shared by the corresponding combination of PGDBs. For example, HumanCyc has 55 pathways in common with EcoCyc, as well as 67 with AraCyc, while EcoCyc and AraCyc share 69 pathways. Thirty-five pathways are common to all three databases, and are shown in Table 6 . The 35 pathways include significant numbers of pathways from all the pathway classes (biosynthesis, catabolism and energy metabolism), and constitute a significant fraction of the pathway complements of both E. coli (20.1% of the 174 pathways in EcoCyc) and H. sapiens (25.7% of the 135 pathways in HumanCyc). Those 35 pathways therefore constitute a likely upper bound on the number of universally and exactly conserved metabolic pathways. It is an upper bound in the sense that as more organisms are considered, the list of universal pathways cannot grow larger. We propose that the cofactor biosynthesis pathways shared among all three organisms have been conserved because first, they produce complex molecules that are not available from the environments of these organisms; second, these molecules are used as cofactors in so many reactions within the metabolic networks that the requirement for them is absolute; and third, no other pathway to accomplish the synthesis of that molecule has evolved. A study by Ouzounis and Karp surveyed global properties of the E. coli metabolic network, including the most frequently used substrates and cofactors [ 14 ]. Together, the two pyridine nucleotides NAD and NADP are the third most common substrate in the E. coli metabolic network (after water and ATP): removing the ability to synthesize NAD would disable so many reactions as to be insurmountable. Pyridoxal-5'-phosphate is the second most common cofactor (after Mg 2+ ). Coenzyme A and acetyl-CoA together constitute the seventh most common substrate in E. coli , formylTHF constitutes the 23rd most common substrate, and thioredoxin and glutathione constitute the 40th and 41st most common substrates. Discussion Pathway variants The level of metabolic pathway variation in the biosphere remains to be determined. Metabolic pathways have been experimentally elucidated in a small number of model prokaryotic and eukaryotic organisms. Despite the relatively small number of carefully studied organisms, significant pathway variation has been observed both between distinct organisms and within a given organism. For example, at least four variants of the 'glycolytic pathway' have been described [ 15 ]. Sets of variant pathways for glycolysis [ 16 ], leucine degradation [ 17 ], and NAD biosynthesis [ 18 ] can be viewed through the MetaCyc website. In MetaCyc, variant pathways are named with roman numerals; for example, 'NAD biosynthesis I' and 'NAD biosynthesis II'. Metabolic pathways appear to have diverged in a manner analogous to the divergence of biological sequences. The demonstrated existence of pathway variants and ongoing uncertainty as to the full extent of such variation has significant implications for ongoing efforts to predict biochemical pathways from incomplete genomic data. First, it means that the precision with which we can infer pathways in one organism from another solely on the basis of genomic data remains to be determined, because when genomic evidence is found in organism O k for the presence of a pathway, P j , that was experimentally elucidated in an organism O j , this alone does not constitute conclusive evidence for the presence of P j in O k , since a variant of P j (P k ) may be present in O k (note that any such P k variant may not even have been experimentally characterized). Second, for those pathways with known closely related variants (for example, two pathways differing by only a single step with one step differing from the other only at the level of co-reactants/products, such as one using NADP/NADPH the other using NAD/NADH as cosubstrates/products) it is often impossible to choose among these variants on the sole basis of genomic data because of the limited resolution of sequence analysis. We have developed a general approach to representing the presence of pathway variants. MetaCyc pathways (approximately 500) have been grouped using a function-based classification system. Each grouping defines a 'pathway family', each member of which (a pathway variant) endows an organism with one or more specific functional capabilities, for example, the ability to grow in the absence of phenylalanine. Within a given pathway family, variants may be clustered into one or more subfamilies. Subfamilies are groups of pathways that show significant overlap/similarity in terms of individual reaction steps (reactants and products of each reaction); enzymatic activities that catalyze these steps; and genes encoding the enzymes that mediate these activities The similarity between variants within a given subfamily suggests they evolved from a common ancestor pathway. For example, at least four variants of the glycolytic pathway have been described; all enable the conversion of glucose to pyruvate and show significant overlap/similarity in their component reactions. Other pathway variants have been observed that show little similarity to each other (for example, some of the amino-acid degradation pathways) and these are therefore believed to have evolved from distinct ancestor pathways. The existence of pathway subfamilies indicates that multiple pathways have coevolved to meet common nutritional/metabolic challenges. For the purposes of this paper, genomic evidence for the presence of a specific biochemical pathway, P 1 , in humans is taken as evidence that P 1 and/or other members of the pathway family to which P 1 belongs are likely to be present in humans (including those not yet included in MetaCyc and/or experimentally elucidated). Indeed, PathoLogic sometimes inferred the presence of multiple variant pathways in humans. This occurred because when evidence was found in the genome for the presence of one member of a pathway family/subfamily, this evidence often also supported the presence of other members of this family/subfamily. In these cases, all inferred variants were included in HumanCyc. Of course, the specific members of a given pathway family actually present in humans may include one or more of those inferred from MetaCyc or other members of this pathway family not yet described in MetaCyc and/or not yet experimentally elucidated from any organism. It is attractive to think that multiple variant pathways might refer to metabolically differentiated tissues in the body, or to different regulatory states available to the same tissue. An example of the latter would be the liver; at different times of day it either synthesizes glycogen, taking glucose from the blood, or it degrades glycogen to maintain the blood glucose level. HumanCyc as a tool The HumanCyc PGDB is freely available for use by the scientific community from the SRI website [ 19 ]. Basic queries to HumanCyc can be issued through the BioCyc Query Page [ 20 ]. This page supports a number of query types. For text searches through the DB, for example, enter 'tryptophan' next to the 'Query All (by name)' box, and then click 'submit' to retrieve a list of all enzymes, pathways, compounds, and reactions whose name includes that word. Click on 'Choose from a list of pathways' to generate a list of all pathways within HumanCyc. Click 'submit' near 'Browse Ontology' to browse one of several possible classification hierarchies, such as the ontology that classifies metabolic pathways according to their physiological role. When viewing a HumanCyc pathway display, be aware that the software omits enzyme names for pathway holes. That is, when no human protein has been identified that catalyzes a reaction within a pathway, no enzyme name is drawn next to that reaction. The cellular overview [ 21 ], a full human metabolic pathway map, provides a tool for analysis of high-throughput datasets. The Omics Viewer [ 22 ] allows the user to visualize gene-expression data, protein-expression data, metabolomics measurements, or reaction flux data in a pathway context by painting reaction steps in the metabolic overview with colors that represent the expression levels of the genes coding for enzymes that catalyze those steps. The PathoLogic summary page for H. sapiens includes a report that lists the evidence for each predicted pathway in HumanCyc, with pathways sorted according to the MetaCyc pathway ontology [ 23 ]. HumanCyc is currently available for local installation at your institution [ 24 ] as a set of flat files, and as part of an executable program running on Windows/Intel, Linux/Intel and Solaris/Sun. The latter configuration can function as both a desktop application, and as a local mirror of the HumanCyc website on the user's intranet. Related work Related databases of human metabolic pathways include the GenMAPP [ 25 ] collection of 14 drawings of human metabolic pathways that are available through the web; the KEGG [ 26 ] metabolic pathway website, which allows coloring of a set of reference pathway maps to indicate which enzymatic steps have corresponding human enzymes; and the Reactome (Cold Spring Harbor Laboratory, European Bioinformatics Institute (EBI) and the Gene Ontology (GO) Consortium) project, which has curated human metabolic pathways in a web-accessible database [ 27 ]. PathoLogic was the first program for automated inference of genome-scale pathway reconstructions from genome data [ 28 ]. Additional methods for prediction of metabolic pathways include methods that infer 'extreme pathways' and 'elementary modes' from flux-balance models of reaction networks [ 29 ], which have been applied to the human mitochondrion [ 30 ]. Rather than recognizing known, historically defined pathways in a large reaction network, as does PathoLogic, these methods infer pathways from the stoichiometric matrix representing a biochemical network by convex analysis. The biological utility and meaning of these pathways, and their correspondence to known metabolic pathways that have been established through experimental studies, remains to be demonstrated. One benefit of pathway analysis by projecting previously known pathways onto a genome using PathoLogic is the fact that projection of known pathways tells us what reactions to expect to occur in another organism, leading to the identification of pathway holes, and the directed search for their fillers. In fact, the occurrence of hundreds of pathway holes in every genome we have analyzed raises the question of how the extreme pathways method can work given that the reaction networks it relies on must omit hundreds of reactions (the pathway holes) as a result of the incompleteness of genome annotations. The following additional pathway prediction methods are related to this work, but have not been applied on a genome scale and have received little validation. Zien and colleagues use gene-expression data to score pathways that are enumerated combinatorially from a reaction database. Van Helden and colleagues use clustering of gene-expression data to generate seed reactions that are used to combinatorially enumerate pathways [ 31 , 32 ]. McShan et al . infer novel biotransformation rules for xenobiotics from the molecular graphs of compounds [ 33 ]. A recent review discusses the potential for reconstruction of metabolic networks using metabolomics technology [ 34 ]. Limitations of HumanCyc and future work The user of HumanCyc should be aware of several potential limitations that influence the interpretation of the DB contents. First, HumanCyc is incomplete in the sense that some known (and unknown) human metabolic pathways are not present in it. It will require a significant database curation effort to enter all known pathways. Second, HumanCyc probably contains false-positive pathway predictions. Our approach is to err on the side of being more inclusive in our pathway predictions so that all potential pathways are brought to the attention of the scientific community for evaluation. For example, HumanCyc sometimes contains multiple pathway variants that we currently lack the evidence to choose between, and in other cases the actual human pathway may be a variant of the pathway present in HumanCyc. Third, HumanCyc does not encode information about the location (compartment(s), cell type(s) or tissue(s)) in which a pathway occurs. Each pathway is defined using all identified human enzymes, meaning that location-specific versions of a pathway are not specifically identified. Furthermore, the presence of a pathway could be incorrectly inferred because the enzymes that make up a pathway are found in different locations and are never expressed in a common location, meaning that the pathway can never occur in its entirety. Fourth, HumanCyc does not contain nucleotide or amino-acid sequences. Our future work will address the first three of these limitations, and will include a significant effort to manually refine and update HumanCyc with pathway and enzyme information from the biomedical literature. To date, three pathways have been added to HumanCyc through manual curation. Experimentally elucidated pathways in HumanCyc will be annotated as such with evidence codes; pathways predicted by PathoLogic are annotated as computationally inferred. HumanCyc can be used to develop an agenda for experimental refinement of the human metabolic map. Predicted pathways should be experimentally verified, with particular attention to choosing among multiple pathway variants. Candidate 'fillers' for the 178 unresolved pathway holes identified herein should be identified. Conclusions PGDBs endow genomic information with an extended dimension that allows researchers to analyze an organism's genome with respect to the causal relationships inherent to a metabolic network. In this sense, HumanCyc provides the opportunity to look at human metabolic processes within the context of the annotated human genome, and vice versa . The computationally predicted metabolic network provided a rational framework for understanding the genetic basis of some well-characterized human dietary requirements, that is, 11 nonessential amino acids and 22 essential nutrients (9 essential amino-acids and 13 essential vitamins). PathoLogic's pathway prediction process provides a reasonably accurate picture of the metabolic network, and Pathway Tools provides the user with extensive capabilities for refining the DB to reflect improvements in our understanding of the human metabolic network. The query and visualization capabilities of the Pathway Tools software (such as the visualization of gene-expression data superimposed on the metabolic map) will facilitate novel approaches for analyzing the complexity of functional relationships within the human genome. Materials and methods Data gathering and preparation The PathoLogic program generates a new PGDB starting from the annotated genome of an organism, meaning a complete genome sequence (closed or gapped) for which gene prediction and sequence analyses have already identified the locations of likely coding regions, and have predicted the functions of these genes. We used Ensembl Build 31 as our main data source for the annotated human genome, and complemented that information with data from LocusLink, the National Center of Biotechnology Information (NCBI) database of genetic loci. We used GenBank as our source for the mitochondrion genome, as this information was not included in Ensembl. LocusLink mitochondrial loci were used to complement this information when applicable. The information from these different sources required special preprocessing to make it available to PathoLogic. That preprocessing addressed three needs: first, to convert the disparate data formats used by these sources into a format parsable by PathoLogic; second, to extract information useful to PathoLogic; and third, to remove redundancy among the sources. In the case of Ensembl, the standard Ensembl flat files included just a subset of the information needed for the PGDB generation process. We therefore used the EnsMart facility provided by Ensembl [ 5 ] to generate files with the required data. EnsMart allows the user to select subsets of the genome (from small regions to entire chromosomes) and output desired data about each gene in different tabular formats. Additional data file 1 lists the genetic element files generated using EnsMart. One file was generated for each human chromosome and contig, with the file names corresponding to the chromosome and contig names in Ensembl. When the corresponding chromosome for a given contig was known, the contig's file name would be constructed by prepending the chromosome's name to the contig's name, otherwise 'Un', for unknown, was prepended to the contig's name. We selected LocusLink file LL3_030319 (Version from 19 March, 2003), which contained all data required by PathoLogic (the LocusLink organism-specific files, like the Ensembl files, did not include all the needed database fields). Table 7 shows the types of information extracted from each source when available. For example, we included a large number of comments extracted from LocusLink in HumanCyc; each such comment in HumanCyc ends with a citation to LocusLink to properly attribute its source. Function descriptions for gene products are very important for PathoLogic, as they will be matched against enzyme activity names stored in MetaCyc, a multi-organism database of experimentally determined metabolic pathways and enzymes [ 15 ]. In Ensembl, such functional information had to be parsed from the 'function' field in EnsMart data. This field sometimes includes long and complex descriptions of the gene product's function, mostly extracted by Ensembl from SwissProt [ 35 ]. These descriptions could include synonyms, EC numbers, and multiple functions (see Figure 1 ). All that information had to be extracted and presented to PathoLogic in a structured form. When generating the input files for PathoLogic, information from Ensembl and LocusLink was combined only when the Ensembl record for a gene included a cross reference to a LocusLink ID. In such a case, data from the LocusLink entry was merged with that of the Ensembl gene, meaning that when both databases provide an attribute such as a gene name, the Ensembl data is preferred. This approach created gene objects for HumanCyc that include information from both the Ensembl and the LocusLink entries. Those HumanCyc gene objects use the Ensembl ID as their unique identifier, and have database links to the corresponding LocusLink entries. For the mitochondrial data, the number of LocusLink mitochondrial loci was small enough so that they were easily checked for matches with corresponding genes in the GenBank files. LocusLink loci that had no direct counterpart in either the Ensembl data or the GenBank mitochondrial data (for example, LocusLink loci that were not directly referenced in any Ensembl gene record) were assigned their own record in the PathoLogic input files. The LocusLink ID was used as the unique identifier for these gene objects. Only loci corresponding to 'real' genes (not models, phenotypes or pseudogenes) were included in HumanCyc. We must point out that records corresponding only to LocusLink loci lack gene position information, so they cannot be precisely placed on a chromosome map. We were aware that adding these LocusLink-only-based records to HumanCyc would produce some redundancy in the database, but this was accepted for the sake of completeness. Manual analysis of similar Ensembl and LocusLink-based gene objects after building HumanCyc led to the fusion of gene objects corresponding to the same gene. The number of LocusLink gene objects that were not merged to corresponding gene objects from Ensembl or GenBank is shown in the 'nonredundant' column of Table 7 . It is readily apparent from Table 7 that HumanCyc, thanks to the combination of the Ensembl and LocusLink data sources, has excellent cross-reference coverage to many other biological databases (including Ensembl and LocusLink themselves). In addition to the databases mentioned in Table 7 , we added links to the GeneCards genomic database for those genes with known HUGO IDs. Seventy-six PathoLogic input files were generated from the preceding data sources: 24 for the human chromosomes, one for the mitochondrion, 50 for the different contigs not yet integrated to the chromosomal sequences, and one called 'unknown' for all the loci that had no chromosome information. A replicon object was created in HumanCyc for each of these files. Prediction of human metabolic pathways using PathoLogic This section summarizes the PathoLogic algorithm (for a more detailed description of the method see [ 36 , 37 ]). For an evaluation of the accuracy of PathoLogic see [ 36 ]. After initializing the schema of the new PGDB, a database object is created for each replicon and contig, and for each gene and its corresponding gene product. PathoLogic then tries to determine the metabolic reaction catalyzed (if any) by each gene product in the organism by using its EC number, if provided in the annotation, and by matching the name of each gene product against the extensive dictionary of enzyme names within the MetaCyc DB [ 15 ]. Finally, the list of reactions now known to be catalyzed by the organism is matched against all the pathways in MetaCyc. For pathways with significant numbers of matches (see [ 36 ] for a detailed description of this algorithm), PathoLogic imports the pathway and its associated reactions and substrates from MetaCyc into the new PGDB. This method of pathway prediction is analogous to predicting the function of a protein based on sequence similarity to a protein of known function, in that both methods recognize the presence of something known (a known pathway versus a known protein function) based on a similarity between patterns (a pattern of enzymes present versus a sequence pattern). The two methods share similar limitations: just as sequence similarity cannot predict protein functions that are not in the sequence database, PathoLogic cannot predict pathways that are not in MetaCyc. As mentioned above, PathoLogic will assign reactions for those enzymes that have an exact EC number match or name match against MetaCyc. A gene product name may not exactly match that of any enzyme in MetaCyc. Some enzyme names will produce ambiguous matches or no match at all. PathoLogic assembles a list of 'probable enzymes' that includes both ambiguous matches and nonmatched proteins whose names suggest enzymatic activity. This list is examined manually through a PathoLogic module that helps the user evaluate possible matching candidates within MetaCyc and assign probable enzymes to the correct reaction, if possible. An alternative to our strategy of using the existing EC number and function assignments from LocusLink and Ensembl would be for us to discard those assignments and to reanalyze the genome using sequence analysis methods to produce new assignments. We rejected this approach for two reasons: first, it would discard some experimentally derived function assignments in place of less reliable computational assignments; and second, we consider the Ensembl function predictions to be of high quality, and we are aware of no evidence that our group, or any other group, has a sequence analysis methodology that will produce function assignments that are substantially more accurate than those of Ensembl. Finally, PathoLogic generates reports summarizing the amount of evidence supporting each pathway in the new PGDB, and listing the pathway holes, that is, the enzymes missing from each predicted pathway. This information helps the user identify pathways that should be deleted from the PGDB, such as variant pathways and false-positive predictions made by PathoLogic. For example, MetaCyc includes eight variants of the TCA cycle. Several of these might appear in a newly predicted PGDB. False-positive pathways, some of which are predicted because they share reactions with other pathways in MetaCyc, should be removed from the PGDB. Once variant or false-positive pathways are eliminated by the user, PGDB generation has been completed. Filling holes in HumanCyc pathways To determine the function of a protein sequence, researchers typically use a single sequence to search for potential homologs in a large public database. To identify sequences to fill pathway holes, we have, in effect, reversed this search process. We search the genome for a sequence that will provide the enzymatic function needed to fill each pathway hole. Our method uses multiple isozyme sequences (retrieved via MetaCyc from Swiss-Prot and PIR) to search a genome for similar sequences (hole-filler candidates). We then evaluate each candidate to determine the probability that the sequence has the desired function based on homology and pathway-based data. A hole-filler tool that implements this pathway-driven gene-finding methodology has been developed [ 9 ]. Analysis of HumanCyc metabolic network Once HumanCyc was built and manually refined, as explained above, we examined the metabolic network within HumanCyc in order to make a preliminary assessment of the quality of PathoLogic's predictions and to check for pathways not previously thought to occur in humans. We also compared the metabolic network of HumanCyc to that of two of our curated PGDBs, corresponding to a bacterium and a plant. These PGDBs are EcoCyc ( E. coli ) and AraCyc ( A. thaliana ). Additional data files The following additional data are available with the online version of this article. Additional data file 1 contains a table listing the data file names generated from EnsMart for each human chromosome or contig and provided as input to PathoLogic, thus indicating which contigs are associated with which chromosomes. Supplementary Material Additional data file 1 A table listing the data file names generated from EnsMart for each human chromosome or contig and provided as input to PathoLogic, thus indicating which contigs are associated with which chromosomes Click here for additional data file	/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC549063.xml
551610	Low adherence with antihypertensives in actual practice: the association with social participation – a multilevel analysis	Background Low adherence is a key factor in explaining impaired effectiveness and efficiency in the pharmacological treatment of hypertension. However, little is known about which factors determine low adherence in actual practice. The purpose of this study is to examine whether low social participation is associated with low adherence with antihypertensive medication, and if this association is modified by the municipality of residence. Methods 1288 users of antihypertensive medication were identified from The Health Survey in Scania 2000, Sweden. The outcome was low adherence with antihypertensives during the last two weeks. Multilevel logistic regression with participants at the first level and municipalities at the second level was used for analyses of the data. Results Low social participation was associated with low adherence with antihypertensives during the last two weeks (OR = 2.05, 95% CI: 1.05–3.99), independently of low educational level. However, after additional adjustment for poor self-rated health and poor psychological health, the association between low social participation and low adherence with antihypertensives during the last two weeks remained but was not conclusive (OR = 1.80, 95% CI: 0.90–3.61). Furthermore, the association between low social participation and low adherence with antihypertensives during the last two weeks varied among municipalities in Scania (i.e., cross-level interaction). Conclusion Low social participation seems to be associated with low adherence with antihypertensives during the last two weeks, and this association may be modified by the municipality of residence. Future studies aimed at investigating health-related behaviours in general and low adherence with medication in particular might benefit if they consider area of residence.	Background The effectiveness [ 1 , 2 ] and efficiency [ 3 ] of antihypertensives may be questioned, as adherence with antihypertensives may be as low as 50% [ 1 , 4 ]. The efficacy of antihypertensives has been evaluated in randomised clinical trials (RCTs). The RCTs are often of short duration, the study population is usually carefully selected and patients with co-morbidity or advanced ages are often excluded from these trials [ 5 - 8 ]. Furthermore, even though drop-outs and lost to follow up occur in RCTs, adherence with medication treatment is often actively supported. In actual practice, however, many patients, who would be excluded from RCTs, receive medication [ 2 ] for a long time and may not be as adherent with medication as those included in RCTs [ 5 ]. Low adherence is one important cause of uncontrolled hypertension [ 9 , 10 ]. Yet, low adherence is sometimes unrecognised [ 11 ] and is often interpreted as treatment resistance [ 10 , 12 , 13 ]. However, little is known about which factors determine low adherence in actual practice [ 14 , 15 ]. The purpose of this study is to examine whether low social participation is associated with low adherence with antihypertensives during the last two weeks, and if this association is modified by the municipality of residence. Social participation is an important concept for understanding the influence of social factors on individual health and behaviour, and can be viewed as a feature of individual social networks [ 16 ]. Good social networks have been suggested to influence health behaviours, possibly through information exchange and establishment of health-related group norms [ 17 ]. Accordingly, a high level of social participation may facilitate adherence [ 16 ] and this association could be modified by the area of residence [ 18 , 19 ]. Multilevel analysis handles both the micro-scale of people and the macro-scale of context simultaneously within one model [ 20 ]. This analytic approach has been suggested as an interesting tool in pharmacoepidemiology [ 21 ]. The first aim of this study is to examine whether low social participation is associated with low adherence with antihypertensives during the last two weeks, independently of low educational level and health status (i.e., poor self-rated health and poor psychological health). The second aim is to analyse whether the hypothesised association between low social participation and low adherence with antihypertensives during the last two weeks varies between municipalities in Scania. Methods Participants The Health Survey in Scania 2000 (HSS-2000) was a postal self-administered questionnaire sent out to a random sample of 23 437 individuals born from 1919 to 1981 living in Scania. The purpose of the HSS-2000 was to obtain information about health conditions and different types of health hazards among the inhabitants of Scania [ 22 ]. The province of Scania in southern Sweden has a population of about 1.2 million inhabitants and is divided into 33 municipalities. In total, 59% participated, of which 98% had complete information about medication use. The present study focused on those 9.6% who indicated use of antihypertensives during the last year and who had complete information about social participation (n = 1288). The Ethical Committee at the Medical Faculty of Lund University approved the study proposal of The HSS-2000, and all of the participants received written information about the survey. Outcome variable Use of antihypertensives was based on an affirmative answer to the question "Have you during the last year used medicine, which was bought at the pharmacy...?" and indicating "Medication for high blood pressure" Low adherence with antihypertensives during the last two weeks (dichotomised) was based on the question "Have you used (this) medicine during the last year, but not during the last 2 weeks?" Those participants who answered yes were considered to have low adherence. Explanatory variables Age was categorised in five groups: <35 (reference), 35–44, 45–54, 55–64 and ≥ 65 years. Low social participation (dichotomised) was assessed after the respondent stated involvement in three or fewer activities (lowest quartile) of 13 formal or informal activities (study circle/course at work place, other study circle/course, union meeting, meeting of other organisations, theatre/cinema, arts exhibition, church, sports event, letter to editor of a newspaper/journal, demonstration, night club/entertainment, large gathering of relatives, private party), which the respondent might have participated in during the previous 12 months [ 23 ]. Low educational level (dichotomised) was defined as having nine years of education or less. Poor self-rated health (dichotomised) was defined as a value of ≤ 3 on an ordinal self-rated health scale ranging from 1 ("Very bad") to 7 ("Very good") [ 24 ]. Poor psychological health (dichotomised) was determined by giving three or more affirmative answers to the 12 items composing the Standardised General Health Questionnaire (GHQ-12) [ 25 ]. Statistical analysis Because of the hierarchy in the data, with individuals nested in municipalities, we used multilevel logistic regression [ 26 ] with individuals at the first level and municipalities at the second level. The area of residence might affect a person's social participation [ 27 ], and, consequently, it may be possible that the influence of low social participation on low adherence with antihypertensives during the last two weeks may vary between municipalities. Therefore, we let the slopes of the association between low social participation and low adherence vary at the municipality level. This random slopes analysis gives information about whether the association between low social participation and low adherence is different in different municipalities. The first model i was created to study the influence of low social participation on low adherence with antihypertensives during the last two weeks, adjusting for age and sex. The second model ii was extended to also include low educational level, because low educational level could be a confounder in the association between low social participation and low adherence with antihypertensives during the last two weeks. The third model iii additionally contained poor self-rated health and poor psychological health. Impaired health may affect both low social participation and low adherence with antihypertensives. Fixed effects The results are shown as odds ratios (OR) with 95% confidence intervals (CI) Random effects We calculated the second level variance (variation between municipalities) regarding prevalence of low adherence with antihypertensives during the last two weeks (i.e., the intercepts in the multilevel regression), and the second level variance regarding the association between low social participation and low adherence with antihypertensives during the last two weeks (i.e., the slope variance in the multilevel regression). We also calculated the covariance between intercept and slope residuals. The covariance gives information about whether the association between low social participation and low adherence with antihypertensives during the last two weeks depends on the prevalence of low adherence in the different municipalities (i.e., cross-level interaction). Parameters were estimated using the Restricted Iterative Generalized Least Squares (RIGLS) and penalised quasilikelihood (PQL). Extra-binomial variation was explored systematically in all models and we found no evidence for under- or over-dispersion. The MLwiN, Version 1.1 software package [ 28 ] was used for the analyses. Results Low adherence with antihypertensives during the last two weeks was found among 11% (145/1 288) of the participants and 49% (635/1 288) were classified as having low social participation. The participants mean age was 63 years. Those participants classified as having low social participation more often reported low adherence with antihypertensives during the last two weeks, low educational level, poor self-rated health and poor psychological health than those who were not classified as having low social participation (Table 1 ). Table 1 Characteristics of the participants (n = 1288) according to individual low social participation. Low social participation Yes (n = 635) No. (%) No (n = 653) No. (%) Total (n= 1288) No. (%) Age (mean years) 65 60 63 Women 340 (54) 346 (53) 686 (53) Low adherence with antihypertensives 96 (15) 49 (8) 145 (11) Low educational level 417 (70) 284 (45) 701 (57) Poor self-rated health 112 (19) 64 (10) 176 (14) Poor psychological health 124 (21) 95 (15) 219 (18) Fixed effects Participants with low social participation had on average a more than twofold higher probability of reporting low adherence with antihypertensives during the last two weeks than those who did not have low social participation (OR = 2.28, 95% CI: 1.16–4.49) (Table 2 ). This association between low social participation and low adherence with antihypertensives during the last two weeks persisted after adjusting for low educational level (OR = 2.05, 95% CI: 1.05–3.99). However, after additional adjustment for poor self-rated health and poor psychological health, the association between low social participation and low adherence with antihypertensives during the last two weeks was not conclusive using a 95% confidence interval (OR = 1.80, 95% CI: 0.90–3.61). Table 2 Municipality variance and age adjusted odds ratios (95% confidence intervals) of low adherence with antihypertensives during the last two weeks in relation to sex, low social participation, low educational level, poor self-rated health and poor psychological health. Model i Model ii Model iii OR 95% CI OR 95% CI OR 95% CI Women vs. men 1.21 (0.82–1.78) 1.15 (0.78–1.69) 1.17 (0.78–1.77) Low social participation (yes vs. no) 2.28 (1.16–4.49) 2.05 (1.05–3.99) 1.80 (0.90–3.61) Low educational level (yes vs. no) 1.87 (1.18–2.96) 1.76 (1.09–2.84) Poor self-rated health (yes vs. no) 1.45 (0.83–2.54) Poor psychological health (yes vs. no) 1.54 (0.92–2.59) Variance SE Variance SE Variance SE Municipality variance in low adherence with antihypertensives (intercept variance) 0.801 (0.461) 0.776 (0.450) 0.793 (0.467) Municipality variance of the association between low social participation and low adherence with antihypertensives (slope variance) 1.812 (0.889) 1.720 (0.857) 1.799 (0.916) Municipality covariance between intercepts and slopes -1.163 (0.609) -1.116 (0.591) -1.175 (0.625) Random effects We found a variance between the municipalities in both low adherence with antihypertensives during the last two weeks (intercept variance) and in the association between low social participation and low adherence with antihypertensives during the last two weeks (slope variance) (Table 2 and Figure 1 ). The negative covariance between intercepts and slopes (Table 2 ) suggested that the associations between low social participation and low adherence with antihypertensives during the last two weeks (slopes) in the 33 municipalities depended on different prevalence of low adherence with antihypertensives during the last two weeks in the different municipalities. The association between low social participation and low adherence with antihypertensives during the last two weeks (slope) was weaker in municipalities with higher prevalence than in municipalities with lower prevalence of low adherence with antihypertensives during the last two weeks (i.e., cross-level interaction). Figure 1 Slope variance in the association between low social participation and low adherence with antihypertensives during the last two weeks among 33 municipalities in Scania, Sweden. Discussion Main findings Our results suggest that low social participation is associated with low adherence with antihypertensives during the last two weeks, independently of low educational level. In other words, the association between low social participation and low adherence withstood adjustment for socio-economic position (expressed by educational level) in our analyses. Social participation might therefore be considered as a real construct, and not only a proxy for socio-economic position. However, the association between low social participation and low adherence with antihypertensives during the last two weeks was weakened after additional adjustment for poor self-rated health and poor psychological health. Furthermore, the association between low social participation and low adherence with antihypertensives during the last two weeks may vary between municipalities in Scania. The weakening of the association between low social participation and low adherence with antihypertensives during the last two weeks after we adjusted for poor self-rated health and poor psychological health may be an expression for confounding. Impaired health may negatively affect both social participation and adherence with antihypertensives. On the other hand, the observed reduction of the association may instead be telling us that physical and mental health are in the pathway between low social participation and low adherence with antihypertensives during the last two weeks [ 29 ]. Social participation may be considered as an early factor in the causal pathway that determines individual health-related behaviour, such as low adherence with medication. Social networks, which are connected to social participation, may promote shared norms around health behaviours, as treatment adherence, which could explain the pathway between social networks and impaired health [ 16 ]. Our present finding that the association between low social participation and low adherence with antihypertensives during the last two weeks may vary between municipalities in Scania gives empirical support to the existence of cross-level interactions (i.e., between municipality and individual) associated with health-related behaviours, such as low adherence with medication. These behaviours may be a result of the interaction between a person and his or her area of residence [ 18 ]. In a previous analysis, we observed that both individual and neighbourhood social participation are associated with individual impaired health and with use of hormone replacement therapy in women [ 18 ]. The present study shows that multilevel regression analysis can be used for investigation of geographical disparities in health and health-related behaviour (e.g, adherence with medication), without analysing any specific area characteristic [ 30 ]. In multilevel analysis, area effects can be investigated by measures of variance and by examining how area boundaries modify individual level associations [ 31 ]. Limitations of the study The rather low participation rate (59%) may increase the risk of selection bias and reduce the ability to generalise the results and compare them to other populations. Nevertheless, the participation rate for participants aged 51–80 was about 65% and the participants using antihypertensives had a mean age of 63 years. The low adherence question we used was stated in a non-threatening manner, which might facilitate for participants to give an honest response and not underreport low adherence [ 32 ]. Self-reported adherence has been reported to correlate with clinical measures of disease activity and control [ 4 ]. Moreover, self-report offers a convenient and non-invasive estimate of adherence behaviour. Nevertheless, the procedure of measuring adherence is controversial. Self-report can be subject to self-presentational and recall biases. People may overestimate their adherence and their memory may be inaccurate [ 32 ]. We might have reduced memory bias in this study by restricting the recall time to two weeks. However, the prevalence of current low adherence (11%) in this study is lower than low adherence reported in a longer period of time, which may be as high as 50% [ 1 , 4 ]. Therefore, our results may underestimate the association between social participation and low adherence with antihypertensives. It is possible that some participants with high adherence in the last two weeks had low adherence in other periods of the year. If this kind of misclassification would be more frequent among participants with low social participation, there would be differential misclassification, and the association between low social participation and adherence with antihypertensives could be underestimated. Non-differential misclassification would also underestimate the association between low social participation and low adherence with antihypertensives. Other ways of measuring adherence may be more appropriate, such as Morisky's four-item scale [ 33 ], which will be used in the Health Survey for Scania 2004. People with low social participation and low adherence with antihypertensives during the last two weeks may have been less inclined to respond to the HSS-2000 questionnaire. This possible selection bias could lead to an underestimation of the association between low social participation and low adherence with antihypertensives during the last two weeks. Conclusion Our results suggest that low social participation is associated with low adherence with antihypertensives during the last two weeks, independently of low educational level. In addition, the association between low social participation and low adherence with antihypertensives during the last two weeks seems to vary between the municipalities in Scania, which gives empirical support to the existence of cross-level interactions (i.e., between municipality and individual) associated with health-related behaviours, such as low adherence with medication. We have recently showed that factors related to the area of residence influence the individual blood pressure level, especially in people using antihypertensive medication [ 34 ], which is in concordance with the results of this present study. Future studies aimed at investigating health-related behaviours in general and low adherence with medication in particular might benefit if they consider that area of residence may modify associations between individual variables. Competing interests The author(s) declare that they have no competing interests. Authors' contributions JM and KJ developed the original idea, participated in the design of the study, performed the statistical analyses and drafted the manuscript. LR and JS participated in the design of the study and revised the manuscript. TL participated in the design of the study, helped to collect the data and revised the manuscript. All authors read and approved the final manuscript. Pre-publication history The pre-publication history for this paper can be accessed here: http://www.biomedcentral.com/1471-2458/5/17/prepub	/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC551610.xml
509309	A Protein Required for Fruitflies to Dispatch Wasp Parasites	null	For a three-millimeter invertebrate, the Drosophila fruitfly has a remarkably sophisticated immune system. Granted, it can't customize an immune response by grooming cells to “remember” and target specific pathogens. But it can rally the less specialized tools of innate immunity to fight disease and infection, and in so doing draws on several aspects of blood cell development (called hematopoiesis)—the foundation of the cellular immune response—also found in vertebrates. In fruitflies, as in vertebrates, hematopoiesis occurs in distinct stages and locations, with nascent cell populations migrating to establish new hematopoietic frontiers. A population of progenitor cells generates all the blood cell types in the organism. These cells arise in two distinct waves and in two distinct locations, with their progeny differentiating into the specialized tissues and organs of the hematopoietic system. In mammals, these organs include the liver and bone marrow, an ongoing source of blood cells after embryogenesis. In fruitflies, the definitive hematopoietic organ is the lymph gland, which churns out three types of blood cells: plasmatocytes and crystal cells, which are also produced by embryonic hematopoietic precursors, and lamellocytes. Plasmatocytes account for up to 95% of circulating fruitfly blood cells and act much like their mammalian counterpart, the macrophage, engulfing substances deemed foreign and dangerous. Crystal cells account for most of the rest and are involved in melanization reactions, which trigger mechanisms involved in containing and killing invading microbes. Unlike plasmatocytes and crystal cells, lamellocytes appear en masse and only under certain conditions, such as the unwelcome appearance of parasitic wasp eggs. Lamellocytes encapsulate and neutralize the invader. Cellular immune response to parasitization in Drosophila requires the EBF ortholog Collier Molecular factors involved in determining the fate of hematopoietic cells have been identified for plasmatocytes and crystal cells but not for lamellocytes—until now. In a search for genes that might precipitate lamellocyte differentiation, Marie Meister, Alain Vincent, and colleagues homed in on a protein, called Collier (Col), that is expressed in lymph glands at the end of embryogenesis. Col is quite similar to a mammalian protein, called Early B-cell Factor (EBF), that controls B-cell development in mice. Both proteins are transcription factors, exerting control by initiating gene transcription. To investigate Col's part in lamellocyte development, the researchers had to get around the fact that mutations that render Col nonfunctional eventually kill the embryo. Using tricks of the genetics trade, Meister and colleagues generated fly larvae that survive loss-of-function mutations in the gene that encodes Col, allowing them to study its role in hematopoiesis. The mutants had normal amounts of circulating plasmatocytes and crystal cells, but when exposed to parasitic wasp eggs, could not muster the requisite response: lamellocyte production. With no lamellocytes, fly larvae had no means of protection against encroaching wasp eggs, which, uncontested, developed into parasitic larvae. The flies with normal Col levels had no such problem, producing considerable numbers of lamellocytes. In these flies, Col expression was restricted to a lymph region called the posterior signaling center (PSC). Col's influence on lamellocyte fate was strong enough that forcing Col expression in precursor blood cells induced lamellocyte differentiation even in the absence of wasp infestation. Based on these findings, Meister and colleagues propose a model for Col-mediated lamellocyte differentiation in which wasp infestation activates Col-expressing cells in the PSC, which then instructs immature blood cells in the lymph gland to become lamellocytes and dispatch the gathering threat. Col's role in fruitfly hematopoiesis closely parallels that of its mammalian ortholog in white blood cell development, EBF. Both are required to generate specialized populations of cells in response to a particular immune threat, and both confer an extra line of defense when faced with special circumstances—key features of vertebrate adaptive immunity. Could it be that building blocks of adaptive immunity were already in place some 550 million years ago, when flies and vertebrates parted ways? Researchers will have to investigate the molecular agents of immunity in intervening species to find out.	/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC509309.xml
551605	Selection of reference genes for gene expression studies in human neutrophils by real-time PCR	Background Reference genes, which are often referred to housekeeping genes, are frequently used to normalize mRNA levels between different samples. However the expression level of these genes may vary among tissues or cells, and may change under certain circumstances. Thus the selection of reference gene(s) is critical for gene expression studies. For this purpose, 10 commonly used housekeeping genes were investigated in isolated human neutrophils. Results Initial screening of the expression pattern demonstrated that 3 of the 10 genes were expressed at very low levels in neutrophils and were excluded from further analysis. The range of expression stability of the other 7 genes was (from most stable to least stable): GNB2L1 (Guanine nucleotide binding protein, beta polypeptide 2-like 1), HPRT1 (Hypoxanthine phosphoribosyl transferase 1), RPL32 (ribosomal protein L32), ACTB (beta-actin), B2M (beta-2-microglobulin), GAPD (glyceraldehyde-3-phosphate dehydrogenase) and TBP (TATA-binding protein). Relative expression levels of the genes (from high to low) were: B2M, ACTB, GAPD, RPL32, GNB2L1, TBP, and HPRT1. Conclusion Our data suggest that GNB2L1, HPRT1, RPL32, ACTB, and B2M may be suitable reference genes in gene expression studies of neutrophils.	Background Neutrophils are the most numerous granulocytes in blood and are responsible for the first line of host defence. However, neutrophils have frequently been implicated in the pathogenesis of many diseases because they can produce various cytokines, chemokines and other proinflammatory mediators [ 1 , 2 ]. Numerous studies have been performed on the mechanisms that regulate the bioactivity of neutrophils. Understanding patterns of expressed genes may provide insight into complex regulatory networks and help to identify genes implicated in diseases. Quantitative real time PCR is one of the most powerful quantification methods for gene expression analysis. Similar to other methods used in expression studies, data from samples are usually required to be normalized against a set of data or references to correct for the difference in the amount of starting materials. The genes used as references are often referred to as housekeeping genes, assuming that those genes are constitutively expressed in certain tissues and under certain circumstances. However, the literature shows that the expression levels of the so called "housekeeping genes" may vary in different tissues, different cell types, and different disease stages [ 3 - 6 ]. Therefore, the selection of the reference genes is critical for the interpretation of the expression data. In this study, we investigated 10 commonly used housekeeping genes (Table 1 ), and found 5 genes could be preferential reference genes for gene expression studies in human neutrophils. Results RNA quality and quantity RNA analysis by an Agilent 2100 Bioanalyzer provided the size profiles and the concentration of the samples. All the RNA samples used in this study were of good quality despite the long neutrophil isolation procedure. Intact rRNA subunits of 28S and 18S were observed on both the gel electrophoresis and electrophotogram, indicating that the degradation of the RNA was minimal (Figure 1 ). Expression patterns of the candidate genes in neutrophils Initial screening for the gene expression pattern suggested that the 10 candidate housekeeping genes were differentially expressed in neutrophils (data not shown). Based on the band intensity of the PCR products, the two lowest expressed genes, two medium expressed genes and the three highest expressed genes were chosen for real-time PCR analysis. ABL1, PBGD and TUBB were excluded from further evaluation due to their extremely low expression level. Standard curve and real-time PCR Standard curves were generated by using copy number vs. the threshold cycle (Ct). The linear correlation coefficient (R 2 ) of all the seven genes ranged from 0.976 to 0.999. Based on the slopes of the standard curves, the amplification efficiencies of the standards were from 91%~100%, which were derived from the formula E = 10 1/-slope -1. The Ct values of all the 7 genes in all the unknown samples were within 15.9 to 33.5 cycles, covered by the range of the standard curves. Electrophoresis analysis of all the amplified products from real-time PCR showed a single band with the expected sizes, and no primer dimer was observed. The dissociation plots provided by the ABI Prism 7900HT also indicated a single peak in all the reactions. The stability and expression level of reference genes in the neutrophils The gene expression levels were measured by real-time PCR, and the expression stabilities were evaluated by the M value of GeNorm. The ranking of the expression stability in these genes was (from the most stable to the least stable): GNB2L1, HPRT1, RPL32, ACTB, B2M, GAPD and TBP (Figure 2 ). The M values of GNB2L1, HPRT1, RPL32, ACTB, and B2M were lower than 0.5, and therefore these genes were concluded to be stably expressed housekeeping genes in neutrophils. A normalization factor (NF) was calculated based on the geometric mean of the copy numbers of these 5 selected reference genes in each sample. After normalization against the NF, the ranking of the relative expression levels was (from high to low): B2M, ACTB, GAPD, RPL32, GNB2L1, TBP, and HPRT1 (Figure 3 ). Based on both the expression stability and expression level, our data suggested that B2M and ACTB can be used as a reference gene for high abundance gene transcripts, RPL32 and GNB2L1 for medium abundance transcripts, and HPRT1 for low abundance transcripts in gene expression studies. Discussion Real-time PCR is one of the most sensitive and flexible quantification methods for gene expression analysis. It provides simultaneous measurement of gene expression in many different samples for a number of genes. However, many factors in real-time PCR may affect the results, including the selection of the reference genes. An ideal reference gene should be expressed at a constant level among different tissues of an organism, at all stages of development, and should be unaffected by the experimental treatment. However, no one single gene is expressed at such a constant level in all these situations [ 4 , 7 ]. For example, ACTB, GAPD, 18S and 28S rRNA are the most commonly used reference genes, but a number of studies have provided solid evidence that their transcription levels vary significantly between different individuals, different cell types, different developmental stages, and different experimental conditions [ 3 - 6 ]. Therefore, thorough validation of candidate reference genes is critical for accurate analysis of gene expression. It is also well known that RNA quality and quantity are critical for successful gene expression analysis. Degraded and inaccurately quantified RNA would give misleading results. In this study, the total RNA was extracted from isolated human neutrophils, and usually it takes 2–3 hours from drawing the blood to obtaining the pure neutrophils. RNA degradation is frequently observed. For this reason we performed careful RNA analysis by using an Agilent 2100 Bioanalyzer (Agilent Technologies) before the gene expression study. The results indicated our RNA samples were of good quality. Other quantification methods which need a microgram-level of RNA were not practical for our study because the amount of RNA extracted from the neutrophils from 10 ml blood was very limited (around 3–5 μg). DNA contamination is another important factor that affects the accuracy of gene expression analysis. In this study, the following steps were taken to prevent and monitor DNA contamination: (1) RNase-free DNase I treatment on all the RNA samples; (2) The primers were designed to be able to distinguish the PCR product derived from mRNA or genomic DNA (Table 2 ); ( 3 ) Dissociation analysis by ABI Prism 7900HT; (4) Gel electrophoresis of all the amplified PCR products. With all these precautions in place we were confident that there was no detectable DNA contamination. The signal from SYBR I was specifically from the desired amplicons, not from artefacts (primer dimers or genomic DNA contamination). For the reasons discussed above, we have confidence that our gene expression results were accurate and reliable, and we further analyzed the expression stability and expression level. The principle that the expression ratio of two ideal reference genes should be identical in all samples is well established. Based on this principle we found GNB2L1, HPRT1, RPL32, ACTB, and B2M were stably expressed in the neutrophils, and they were used for the calculation of a normalization factor (NF). After normalization we found B2M was the most highly expressed, followed by ACTB, RPL32, GNB2L1, and HPRT1 was the lowest expressed. As the expression level of the reference genes may be an additional factor for consideration in the process of reference gene selection, this ranking of the relative expression level of the candidate reference genes may be informative for future gene expression studies in neutrophils. Conclusion To our knowledge, this is the first detailed study of the stability and level of reference gene expression in neutrophils. We found GNB2L1, HPRT1, RPL32, ACTB, and B2M are good choices for reference gene(s) selection. B2M and ACTB can be used for high-abundance mRNA, RPL32 and GNB2L1 for medium-abundance mRNA, and HPRT1 for low-abundance mRNA in expression studies of neutrophils. For more accurate normalization, as suggested by other authors [ 8 ], we recommend a combination of the stably expressed genes GNB2L1, HPRT1, RPL32, ACTB, and B2M as a panel of reference genes for the normalization. Methods Candidate genes for expression studies Ten housekeeping genes were selected from commonly used reference genes (ABL1, ACTB, B2M, GAPD, GNB2L1, HRPT1, PBGD, RPL32, TBP, and TUBB). Gene symbols and their full names, gene accession numbers as well as functions are listed in Table 1 . These genes were chosen because they have different functions in order to avoid genes belonging to the same biological pathways that may be co-regulated. In selecting the genes to be analyzed, preference was given to pseudogene-free genes in the NCBI linked database (Table 1 ). All the primers were designed by the software, Primer 3, . Hairpin structure and primer dimerization were analyzed by NetPrimer. Primers spanning at least one intron were chosen to minimize inaccuracies due to genomic DNA contamination. The length of the primers was from 18-mer to 22-mer, GC content was from 45% to 60%, and the expected PCR products range from 114 bp to 318 bp. If the genes have pseudogenes, primers were chosen according to the alignment results between the genes and the pseudogenes, so that the primers were unique to the genes and different from the pseudogenes (Table 2 ). Subjects and sample preparation A total of 15 volunteers were recruited (Table 3 ). All participants signed an informed consent document. 20 ml of peripheral blood was taken into heparinized tubes. Neutrophil isolation was performed by a Dextran-Ficoll sedimentation and centrifugation method [ 9 ]. Briefly, 20 ml blood was mixed with 5% Dextran (100,000–20,000 k Da; Sigma) in RPMI (9:2). After 40 min sedimentation, the white blood cell rich plasma was transferred onto the top of 10 ml Ficoll (Pharmacia), and centrifuged at 2500 rpm for 15 min. The cell pellet contained the neutrophils. The contaminating erythrocytes were removed by hypotonic lysis. The isolated neutrophils were subject to Kimra staining and microscopic cell differential count. The purity of the neutrophils was calculated. Samples with more than 2% eosinophils were excluded from the study. Half of the isolated neutrophils were used for RNA isolation. RNA extraction and RT-PCR Total RNA was isolated using RNeasy Mini Kit (Qiagen) as described by the manufacturer. Genomic DNA was eliminated by RNase-free DNase I digestion (Qiagen) during the isolation procedure. Isolated total RNA was analyzed on an Agilent 2100 Bioanalyzer using the RNA 6000 pico labchip Kit (Agilent Technologies). First strand cDNA synthesis was carried out with SuperScript RNase H - Reverse Transcriptase (Invitrogen) and random primers (Invitrogen) in a total volume of 20 μl. Reverse transcription was performed at 37°C for 1 hour followed by 72°C for 15 min. Amplification of gene transcripts To screen the basal expression patterns of the candidate genes in neutrophils, three randomly selected samples were tested by PCR with the ten primer pairs (Table 2 ). The expression study was performed using a 384 well plate on an ABI Prism 7900HT Sequence Detection System (Applied Biosystems) with QuantiTect SYBR Green PCR Kit (Qiagen). The reactions were performed according to the manufacturer's instructions with minor modifications. The PCR program was initiated at 95°C for 10 min to activate Taq DNA polymerase, followed by 45 thermal cycles of 15 seconds at 94°C, 30 seconds at 58°C and 30 seconds at 72°C. Size analysis of the PCR products (dissociation analysis or meting curve analysis) was performed immediately after the real-time PCR. The temperature range used for the melting curve generation was from 60°C to 95°C. Each sample was analyzed in triplicate wells. In addition, all the reactions were further subject to electrophoresis on 2.5% agarose gels stained with ethidium bromide to confirm the expected PCR products. Standard curves The amplified fragments from each primer pair were purified with QIAquick PCR purification Kit (Qiagen), and confirmed by DNA sequencing (University of British Columbia, NAPS Unit). The concentrations of the PCR products were quantified by a spectrophotometer (Perkin-Elmer Lambda 2 UV/VIS Spectrometer), which were further transformed to copy numbers based on the length and base composition of the PCR products. A ten-fold series dilution was made and 10 to 1,000,000 copies were used for generating standard curves in the real-time PCR, plotted as Ct values (cycle numbers of threshold or crossing points) versus logarithms of the given concentrations of the DNA templates. Determination of Gene stability and expression levels in human neutrophils Gene stability was also evaluated using the geNorm software program [ 8 ]. Briefly, this approach relies on the principle that the expression ratio of two perfect reference genes would be identical in all samples in all experimental conditions or cell types. Variation in the expression ratios between different samples reflects the fact that one or both of the genes are not stably expressed. Therefore, increasing variation in this ratio corresponds to decreasing expression stability. The geNorm program can be used to calculate the gene expression stability measure ( M ), which is the mean pair-wise variation for a gene compared with all other tested control genes. Genes with higher M values have greater variation in expression. The stepwise exclusion of the gene with the highest M value allows the ranking of the tested genes according to their expression stability. The proposed threshold for eliminating a gene as unstable was M ≥ 0.5. In the final analysis, genes with M value lower than 0.5 were considered as stably expressed genes, and were used for normalization factor (NF) calculation. Using the NF we calculated and ranked the expression level of all the seven genes in our samples. Abbreviations ACTB, beta-actin; ALB1, Abelson murine leukemia viral oncogene homolog 1; B2M, beta-2-microglobulin; cDNA, complementary DNA; GAPDH, glyceraldehyde-3-phosphate dehydrogenase; GNB2L1 (Guanine nucleotide binding protein, beta polypeptide 2-like 1; HPRT1, Hypoxanthine phosphoribosyltransferase 1; NCBI, National Center for Biotechnology Information; PCR, polymerase chain reactions; PBGD, porphobilinogen deaminase; RPL32, ribosomal protein L32; RT-PCR, reverse transcription-polymerase chain reactions; TBP, TATA-binding protein; TUBB, beta-tubulin Authors' contributions XZ performed all the experimental procedures and was the primary author of the manuscript. LD participated in the study design and data analysis. AS conceived of the study, participated in the study design and coordination. All authors read and approved the final manuscript.	/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC551605.xml
497045	Identification of astrocytoma associated genes including cell surface markers	Background Despite intense effort the treatment options for the invasive astrocytic tumors are still limited to surgery and radiation therapy, with chemotherapy showing little or no increase in survival. The generation of Serial Analysis of Gene Expression (SAGE) profiles is expected to aid in the identification of astrocytoma-associated genes and highly expressed cell surface genes as molecular therapeutic targets. SAGE tag counts can be easily added to public expression databases and quickly disseminated to research efforts worldwide. Methods We generated and analyzed the SAGE transcription profiles of 25 primary grade II, III and IV astrocytomas [ 1 ]. These profiles were produced as part of the Cancer Genome Anatomy Project's SAGE Genie [ 2 ], and were used in an in silico search for candidate therapeutic targets by comparing astrocytoma to normal brain transcription. Real-time PCR and immunohistochemistry were used for the validation of selected candidate target genes in 2 independent sets of primary tumors. Results A restricted set of tumor-associated genes was identified for each grade that included genes not previously associated with astrocytomas (e.g. VCAM1, SMOC1, and thymidylate synthetase), with a high percentage of cell surface genes. Two genes with available antibodies, Aquaporin 1 and Topoisomerase 2A, showed protein expression consistent with transcript level predictions. Conclusions This survey of transcription in malignant and normal brain tissues reveals a small subset of human genes that are activated in malignant astrocytomas. In addition to providing insights into pathway biology, we have revealed and quantified expression for a significant portion of cell surface and extra-cellular astrocytoma genes.	Background Astrocytomas are the most frequent malignant primary brain tumors in adults. Clinically, this group of tumors can be divided into four World Health Organization (WHO) grades. Pilocytic astrocytomas (WHO grade I) are generally slow growing and non-infiltrative pediatric tumors, which are rarely fatal. For the infiltrating astrocytomas, survival decreases with increasing grade. Grade II astrocytomas patients survive an average of over 5 years, but survival drops to 3 years for anaplastic astrocytomas (grade III). Grade IV astrocytomas (glioblastoma multiforme or GBM) account for about half of all astrocytic tumors, with a median survival of less than a year. Effective treatment options for the invasive grade II to IV tumors are still limited to surgery and radiation therapy, with most chemotherapy regimens showing little or no increase in survival. Several recent gene expression profiling studies of human astrocytomas have been able to distinguish between various grades of astrocytomas and between astrocytomas and other human glial tumors, and to identify new molecular classes within histological grade [ 3 - 7 ]. This enhanced molecular classification based on expression patterns of genes and pathways holds promise for better diagnostic and prognostic tools. Candidate glioblastoma associated genes have also been identified using expression profiling [ 8 - 11 ]. While these studies in brain cancers have produced leads for potential therapeutic targets, a systematic and comprehensive evaluation of gene expression in malignant astrocytomas is not readily and freely available for the scientific community. In this study, we sought to create a public and comprehensive gene expression resource for astrocytomas, with the primary intention of aiding searches for new therapeutic targets in malignant astrocytomas. For this purpose we produced and analyzed in detail 25 gene expression profiles of primary astrocytic tumors (grade II, III and IV) using Serial Analysis of Gene Expression (SAGE) [ 12 ]. Complete expression profiles are posted for the scientific community at the CGAP SAGE Genie website [ 2 ]. The utility of the resource was validated by extensive comparisons of tumor with normal tissue. SAGE profiles on normal brain and other tissues created by the Cancer Genome Anatomy Project [ 8 ] were used to subtract out the genes normally expressed in adult brain, leaving a small and specific set of astrocytoma associated genes for each class, and revealing cell surface or extra-cellular matrix related genes highly expressed in the tumors when compared to their expression in normal tissue. A subset of the tumor-associated genes was validated in an independent set of tumors at both the transcript and protein level. In summary we have identified several novel tumor-associated markers in astrocytic tumors as well as various cell surface markers highly expressed in the most aggressive tumor types. Methods Tissue and RNA Astrocytic tumor samples from 21 adults and 4 children were obtained from the Duke Brain Tumor Bank. All samples were classified based on histology according to the World Heath Organization grading criteria. Pediatric normal cortex (15 months) was a gift of Dr. Rachel Myerowitz and normal pediatric cerebellum was from the Maryland Brain Bank. Normal adult cortex and cerebellum were rapid autopsy samples obtained from the Duke Alzheimer's Brain Bank. Total RNA from normal substantia nigra was obtained from Clontech (Palo Alto, CA). PolyA+ RNA from normal adult leukocytes was obtained from Stratagene (La Jolla, CA), as noted for each library's information at the SAGE Genie Website [ 2 ]. RNA integrity was confirmed by gel electrophoresis prior to SAGE library construction. SAGE libraries and informatics SAGE libraries from 25 selected astrocytomas: 8 grade II astrocytomas, 10 grade III anaplastic astrocytomas and 7 grade IV glioblastomas (primary GBM) were constructed using Nla III as the anchoring enzyme and BsmF I as the tagging enzyme using a micro-SAGE protocol. The SAGE library clones were partially arrayed at Lawrence Livermore National Laboratories and inserts were purified and sequenced at the BC Cancer Agency Genome Sciences Centre or arrayed and sequenced at Agencourt Bioscience Corporation. The SAGE 2000 software version 4.12 (available at ) was used to extract SAGE tags from the original sequence files, remove duplicate ditags, remove linker sequences, remove one base pair variations of linker sequences and tabulate the occurrence of each tag. Tag sequences, tag counts and gene associations were stored in a Microsoft Access relational database for subsequent selection of tags with a particular profile. A total of 2,605,122 tags were obtained with an average of 102,988 tags per library. Normal neural tissue tags included a total of 443,560 tags from normal brain [ 8 , 13 ]. These SAGE normal brain libraries have the following unique identifiers: cortex_B_BB542; cortex_B_pool6; thalamus_B_1; cerebellum_B_1; cerebellum_B_BB542; peds_cortex_B_H1571 and substantia_nigra_B_1. Tags totaling 48,039 were also included from normal leukocytes (SAGE_Leukocytes_normal_B_1). Detailed library information and tag counts for each tissue are located at CGAP's SAGE Genie [ 2 ]. Tag counts were normalized to 100,000 tags per library. Real-time PCR Total RNA extraction, cDNA synthesis and quantitative PCR were performed as previously described [ 1 ]. Gene expression levels were normalized to 3 genes; GAPDH, ribosomal protein RPS27 and low molecular mass ubiquinone-binding protein (QP_C). Both RPS27 and QP-C showed a relatively even expression level across the libraries as assed by SAGE analysis. Relative expression levels were calculated in comparison to the levels in nine normal neural tissues, including normal brain (3x), cerebellum (2x), thalamus, gray matter, caudate nucleus and pediatric cortex according to Saha et al ., 2001 [ 14 ]. A list of the PCR primers used for each gene is available upon request. Immunohistochemistry Formalin-fixed 5-μm paraffin-embedded sections were stained with various antibodies using the biotin/streptavidin RTU Vectastain Universal Quick kit (Vector laboratories, Burlingame, CA) as previously described [ 1 ]. Shortly, sections were deparaffinized in HemoD and re-hydrated through descending alcohols. Endogenous peroxidase was quenched by incubating the slides in methanol/5% H 2 O 2 at room temperature for 10 min. Non-enzymatic antigen retrieval was performed using the Antigen retrieval solution AR or Citra (Biogenex, San Ramon, CA) in combination with microwave treatment. Sections were then blocked in PBS/0.5% Triton X-100 containing 2.5% normal horse serum for at least 30 min at room temperature and incubated with primary antibody overnight at 4°C. The primary antibodies used were mouse monoclonal anti-aquaporin1, at a 1:60 dilution (Abcam Limited, Cambridgeshire, United Kingdom) and mouse monoclonal anti-Topoisomerase II Alpha used at a 1:30 dilution (Novocatra Laboratories, New Castle, United Kingdom). Sections were developed with a DAB (Sigma, St. Louis, MO) substrate, counterstained with hematoxylin and mounted with Cytoseal-60. Tissue micro arrays The tissue micro array used in these studies contained cores from 20 glioblastomas, 20 anaplastic astrocytomas, 20 infiltrating astrocytomas and 20 oligodendroglial lesions and was prepared according to methods described by Kononen and Kallioniemi [ 15 ]. Microscopic examination of the array confirmed that the appearance of the tumor tissue cores corresponded to that in the donor blocks. A neuropathologist (CGE) selected the tumor areas sampled in each case and examined the resulting arrays to ensure they accurately represented the donor cases. Results and Discussion SAGE gene expression profiles, selection and confirmation of tumor-associated genes This report describes the comprehensive generation of expression profiles of three astrocytic tumor grades based on Serial Analysis of Gene Expression (SAGE). The main goals were to identify genes not expressed in normal brain tissue and genes highly expressed in the more aggressive astrocytomas encoding cell surface or extra-cellular matrix related proteins that could be of potential therapeutic interest. We generated SAGE profiles on 8 infiltrating astrocytomas, 10 anaplastic astrocytomas and 7 glioblastoma samples. Combined with two glioblastoma profiles [ 8 ], previously deposited on SAGE Genie, this study analyzed 2,734,106 astrocytoma SAGE tags. On average we could distinguish over 27,000 unique tags in each tumor grade after excluding those with single counts (Table 1 ). Complete expression profiles and library information are posted for the scientific community at the Cancer Genome Anatomy Project (CGAP) SAGE Genie website , where the libraries can be downloaded or viewed online using SAGE Genie tools [ 2 ]. In order to identify tumor-associated genes, we sought highly expressed transcripts in each grade of astrocytoma that were not expressed in 7 normal brain tissues. This also helped control for contaminating normal cells within the tumor sample. Tags expressed in a normal leukocyte SAGE library were also included in the analysis so they could be subtracted to reduce the chances of identifying transcripts from white blood cells that frequently infiltrate these tumors. Initially we selected for tags with an average expression of at least 3 per 100,000 tags. Subsequently, we selected for tags expressed at less than 2 counts per 100,000 in each of the 8 normal libraries (7 normal neural tissues & one leukocyte), reducing the number of tags to less than 100 per tumor type (Table 1 ). We further narrowed down the list of candidates by including only those tag sequences that could be matched with a full-length cDNA sequence. From these lists of genes we selected respectively 8, 16 and 10 genes for real-time PCR analysis in an independent set of 14 to 17 grade II, grade III and grade IV primary tumors. Only 6 genes could not be confirmed by real-time PCR analysis, 3 of 8 (grade II selection) and 3 of 16 (grade III selection). Another 8 genes were confirmed in only 20 to 25 % of the tumors. Table 2 lists those genes with a 5-fold or more over-expression by real time PCR in at least 30% of the tumor samples from the corresponding grade, when compared to an average of the normal neural tissue expression. The results show that the in silico selection using SAGE profiles revealed tumor-associated genes that can be found in a different set of primary tumors implying a possible role for these genes in tumor development and increasing their value as putative therapeutic targets. The limited availability of high quality antibodies for the identified tumor-associated genes (Table 2 ) narrowed our study at the protein level to those genes that had previously been implicated in astrocytic tumors. Monoclonal antibodies for TOP2A and AQ1 were used to analyze the expression at the protein level in individual glioblastoma sections and in a tissue micro-array. Strong nuclear staining in 5 of 8 individual glioblastomas tested was found for TOP2A (Figure 1A,1C and 1D ). The GBMs showed intensely staining cells with the percentage of positive cells varying between 3 and 10%. Similar results were obtained when staining a tissue micro-array containing a different set of 20 GBMs. In six cases we found 5 to 10% of the cells expressing TOP2A. The percentage of strong positive cells among the 20 anaplastic astrocytomas present on the tissue micro array was lower and estimated to be 2 – 3% in 10 cases. Normal neural brain tissues like cortex, white matter, spinal cord and hippocampus where negative for TOP2A (Figure 1B ). Cytoplasmatic staining was observed when anti-aquaporin 1 monoclonal antibodies were used. Three of 5 individual GBM sections where positive, with examples shown in Figures 1F and 1G . These results emphasize that the high transcript levels of TOP2A and Aquaporin 1 correlate well with a high protein level in a third independent set of tumors, implying that this might also be the case for the other identified tumor-associated genes as listed in Table 2 and Table 3 . Potential therapeutic targets in astrocytic tumors Encouraged by the confirmation of our initial in silico analysis for tumor-associated genes we formulated a slightly simplified approach to find cell surface, extra-cellular matrix and cell adhesion related genes. We selected for transcript tags with at least a ten fold over-expression when compared to the average expression level in normal neural brain tissues. Next we applied a filter that would include only those tags with an average expression of at least 5 or 10 counts per 100,000 tags in 30% or more of the tumors, respectively for anaplastic astrocytomas and glioblastomas. The generated lists of transcript tags were mapped to the corresponding gene using SAGE Genie, where after the gene ontology information (if available) was used as a final filter to identify membrane, cell surface and cell adhesion related genes (Table 3 ). Interestingly, almost 50% of the genes identified as highly expressed in both tumor types have not previously been implicated in astrocytomas and are potential new therapeutic targets. Intracellular proteins that contribute to the fusion of the vesicles with the plasma membrane during exocytosis include synaptosomal protein and vesicle-associated membrane proteins (VAMP). Both anaplastic astrocytomas and glioblastomas show high expression of VAPB and anaplastic astrocytomas express caveolin1 (Table 3 ). It has been shown that caveolae require intact VAMP for targeted transport in endothelial cells. Caveolae and associated proteins might be targeted in cancer as recently suggested [ 16 ]. One of the other genes expressed in common between the three astrocytomas is chitinase 3-like 2 ( CHI3L2 ) or YKL-39 (Table 2 ). CHI3L2 is a chondrocyte growth related gene and is an antigen found in rheumatoid arthritis [ 17 ] and osteoarthritis and a possible immunotherapy target. Another commonly over-expressed gene is Neuromedin B . This neuropeptide has been implicated as an autocrine growth factor in lung cancer cells [ 18 ] that binds to a G protein-coupled receptor on the cell surface and might have a similar role in astrocytic tumors. It is tempting to speculate that a specific neuropeptide antagonist or neutralizing antibodies might reduce astrocytoma growth. Neuromedin B had previously been described as a GBM marker [ 9 ], and was included along with ABCC3 in the real-time PCR analysis as a positive control. Another relatively unknown gene is the recently characterized SMOC-1 [ 19 ], which was identified as a grade II and III tumor-associated gene (Table 2 ). This gene is related to SPARC/osteonectin, which was reported to participate in angiogenesis and tumor formation of human melanomas. Another extra cellular gene, Matrix Gla protein [ 20 ] had increased expression levels in higher-grade astrocytomas. MGP helps regulate the calcification of the extra cellular matrix [ 20 ]. Aquaporin 1 is an integral membrane protein important in the regulation of water transport in various epithelial and endothelial cell types [ 21 ]. The over-expression of AQP1 in human brain tumors was described in a limited array study of 4 Glioblastomas [ 11 ] and it has been suggested that the protein might play a role in brain tumor edema in a similar way as the closely related aquaporin-4 [ 22 ]. Although the specific role of AQP1 in brain tumors is still unknown, our demonstration that AQP1 is consistently expressed in GBM may prompt other studies. Thymidilate synthetase and Topoisomarese 2A were over-expressed in glioblastoma as well as in our previous study of medulloblastoma [ 1 ]. Considering the role of Top2A as a molecular target of various anticancer drugs, and its identification as a survival marker in astrocytomas [ 23 ], its over-expression at the protein level in multiple brain tumors and the development of TOP2A inhibitors [ 24 ] makes the molecular targeting of TOP2A worthy of further investigation. In summary we have identified a number of new tumor-associated genes for three different grades of astrocytic tumors, and helped re-confirm in a larger set of samples several previously known astrocytoma genes. Despite the high heterogeneity among gliomas, a small set of genes is consistently observed at high levels in more than a third of each grade of astrocytoma studied. Many other cell surface, extra-cellular matrix or cell adhesion genes have been identify as potential targets for cancer therapy in astrocytic tumors. Although the therapeutic value of these markers is speculative at this point, by integrating this data onto the commonly used gene expression resource, SAGE Genie , this data can be used as a standard to determine gene expression in astrocytomas. Further evaluation by in vitro and in vivo studies will be necessary to establish the role of these over-expressed genes in brain tumor development and progression. Authors' contributions KB performed computational analyses, generated experimental data, participated in the design of the study and drafted the manuscript. JBE generated experimental data. CBE and GJR participated in the study design and manuscript editing. All authors read and approved the final manuscript. Pre-publication history The pre-publication history for this paper can be accessed here:	/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC497045.xml
554768	Increased p53 immunopositivity in anaplastic medulloblastoma and supratentorial PNET is not caused by JC virus	Background p53 mutations are relatively uncommon in medulloblastoma, but abnormalities in this cell cycle pathway have been associated with anaplasia and worse clinical outcomes. We correlated p53 protein expression with pathological subtype and clinical outcome in 75 embryonal brain tumors. The presence of JC virus, which results in p53 protein accumulation, was also examined. Methods p53 protein levels were evaluated semi-quantitatively in 64 medulloblastomas, 3 atypical teratoid rhabdoid tumors (ATRT), and 8 supratentorial primitive neuroectodermal tumors (sPNET) using immunohistochemistry. JC viral sequences were analyzed in DNA extracted from 33 frozen medulloblastoma and PNET samples using quantitative polymerase chain reaction. Results p53 expression was detected in 18% of non-anaplastic medulloblastomas, 45% of anaplastic medulloblastomas, 67% of ATRT, and 88% of sPNET. The increased p53 immunoreactivity in anaplastic medulloblastoma, ATRT, and sPNET was statistically significant. Log rank analysis of clinical outcome revealed significantly shorter survival in patients with p53 immunopositive embryonal tumors. No JC virus was identified in the embryonal brain tumor samples, while an endogenous human retrovirus (ERV-3) was readily detected. Conclusion Immunoreactivity for p53 protein is more common in anaplastic medulloblastomas, ATRT and sPNET than in non-anaplastic tumors, and is associated with worse clinical outcomes. However, JC virus infection is not responsible for increased levels of p53 protein.	Background The current World Health Organization classification for tumors of the nervous system includes medulloblastoma, medulloepithelioma, ependymoblastoma, supratentorial primitive neuroectodermal tumor (sPNET) and atypical teratoid/rhabdoid tumor (ATRT) in the category of embryonal brain neoplasms [ 1 ]. These tumors are united by their primitive cytological appearance and the ability to differentiate into multiple cell types. However, recent studies indicate that these lesions are genetically, and to some extent clinically, separable. ATRT are defined by the presence of rhabdoid cells, contain INI1 mutations, and cause particularly grim clinical outcomes [ 2 ]. Medulloblastomas commonly contain isochromosome 17q, but this chromosomal alteration is rarely detected in sPNET or ATRT [ 3 ]. Global gene expression profiles also suggest that medulloblastoma, sPNET and ATRT are distinct entities [ 4 ]. Little is known about the differences in p53 expression and function among the various embryonal brain tumor subtypes. Initial reports on the p53 tumor suppressor gene suggested it was mutated in 10% or less of medulloblastomas [ 5 - 8 ]. However, Frank and colleagues have recently shown that the p53 pathway is inactivated by mutation of p53, methylation of p14ARF, or deletion of INK4/ARF in 21% of medulloblastomas [ 9 ]. In their study, 5 of the 6 medulloblastomas with alterations abrogating p53 function had significant anaplasia. Large cell/anaplastic changes in medulloblastoma are prognostic of significantly worse clinical outcomes [ 10 - 12 ]. Interestingly, p53 protein accumulation, which is often associated with loss of functionality, has been found by some [ 13 - 15 ], but not others [ 16 , 17 ], to predict shorter survival in medulloblastoma patients as well. Several other lines of evidence also support a role for the p53 pathway in medulloblastomas. First, medulloblastomas sometimes arise in the context of Li Fraumeni syndrome, in which p53 germline mutations predispose patients to a wide range of neoplasms [ 18 ]. Second, inactivation of p53 accelerates the formation of medulloblastomas in transgenic mouse models [ 19 ]. Finally, inactivation of p53 and Rb simultaneously, either through genetic disruption or overexpression of viral T antigen, results in medulloblastomas in rodents [ 20 - 22 ]. It has been suggested that viral infection of human CNS tissues could promote formation of brain tumors by inhibiting p53 and Rb activity [ 23 ]. Some researchers have reported the presence of JC virus or other oncogenic polyomaviruses in human brain tumor specimens, including medulloblastomas [ 24 , 25 ]. Large T antigen expressed by these viruses binds and inactivates p53 [ 26 ]. This process results in the accumulation and immunohistochemical detection of p53 protein. In human neural tissue this is best demonstrated in progressive multifocal leukoencephalopathy, in which JC virus infected oligodendroglial cells are strongly p53 immunopositive [ 27 , 28 ]. It is therefore possible that the accumulation of p53 protein in some human medulloblastomas is caused by viral infection. In order to confirm the association between p53 immunopositivity, clinical outcome, and embryonal tumor subtype, we stained a tissue array containing representative cores from 80 embryonal brain tumors for p53. We also investigated JC virus infection as a possible mechanism for accumulation of p53 protein by searching for viral sequences using a highly sensitive quantitative real time polymerase chain reaction (PCR) assay. We found an association between p53 immunoreactivity, clinical outcome, and tumor subtype, but did not detect JC virus in medulloblastoma or supratentorial PNET. Methods Clinical material Medulloblastomas and other embryonal brain tumors diagnosed at the Johns Hopkins University Department of Pathology were identified through review of departmental records. Classic, desmoplastic/nodular and large cell/anaplastic medulloblastomas were classified using World Health Organization guidelines [ 1 ]. Nuclear size, cell morphology and the frequency of mitosis and apoptosis were used as previously described to grade anaplasia [ 11 ]. 80 Tumors from 78 patients were used to create a tissue array as previously described [ 29 ]. Patients ranged from 8 months to 55 years of age, with a median age of 9 years. Microscopic examination of the array confirmed that the appearance of tumor tissue cores corresponded to donor blocks. Frozen tumor tissue obtained from medulloblastomas resected at the Johns Hopkins Hospital was snap-frozen in liquid nitrogen and stored at minus 80°C prior to nucleic acid extraction. DNA was extracted using Trizol and further purified using a DNeasy column (Qiagen, Valencia, CA) according to the manufacturer's instructions. This study was approved by the Johns Hopkins University Institutional Review Board. Immunohistochemistry The tissue array was sectioned at four microns, deparaffinized, and subjected to antigen retrieval by steaming (20 minutes at 80°C). Slides were then incubated at room temperature for 45 minutes with monoclonal antibody directed towards p53 (1:2000, clone DO-7, DAKO, Carpinteria, CA). Primary antibody was detected using the avidin-biotin complex (ABC) method with diaminobenzadine serving as the chromagen. We semiquantitatively graded staining intensity as negative, weak, or strong. Carcinomas with mutations leading to p53 stabilization were used as positive controls. No staining was seen in the absence of primary antibody (negative control). Detection of virus by real time PCR JC virus sequences were amplified from DNA using the forward primer PEP-1 (5'-AGT CTT TAG GGT CTT CTA CC-3') and reverse primer PEP-2 (5'-GCC AAC CTA TGG AAC AG-3') [ 30 ]. Additional specificity for detection of JC virus was achieved using the FAM/Black Hole Quencher-1 (FAM/BHQ-1) labeled TaqMan probe (5-/56-FAM/ CCA ACA CTC TAC CCC ACC T /3BHQ_1/-3) [ 31 ]. This probe does not cross-react to the closely related human BK polyomavirus, or simian SV40 polyomavirus. Fifty microliter reaction volumes were used, comprised of 1X universal master mix (Applied Biosystems, Foster City, CA), 0.05 μM probe, 0.4 μM PEP-1, 0.4 μM PEP-2, and 16–288 ng tumor DNA. Amplifications were performed using a Biorad ICycler with the following thermal profile: 95°C for 10 minutes, then 50 cycles at 95°C for 15 seconds and 57°C for 1 minute. Quantitation of JC virus in tumors was determined using linear regression with an external standard curve included on each plate containing a 5-fold dilution series of known input JC virus plasmid diluted in a constant background of human placental DNA (70 ng/μl). To normalize for sampling variability, we quantitated the total cell equivalents in each sample by amplifying a human endogenous retrovirus gene (ERV-3) in an equal amount of tumor DNA [ 32 ]. Conditions for ERV-3 amplification were: 0.4 μM forward primer (PHP10-F: 5'-CAT GGG AAG CAA GGG AAC TAA TG-3'), 0.4 μM reverse primer (PHP10-R: 5'-CCC AGC GAG CAA TAC AGA ATT T-3'), and 0.25 μM TaqMan probe labeled with FAM and BHQ (PHP-P505/ERV-3 Probe: 5'-/56-FAM/TCT TCC CTC GAA CCT GCA CCA TCA AGT CA/3BHQ_1/-3'). ERV-3 amplification was performed on an ABI 5700 using the following thermal profile: 95°C for 10 minutes, followed by 50 cycles of amplification at 95°C for 15 seconds and 60°C for 30 seconds. DNA extracted from the diploid cell line ATCC CCL 205 diluted in a constant background of 50 ng/μl salmon sperm DNA was used to construct the ERV-3 standard dilution curve. Statistical analysis p53 statistical analyses were performed using GraphPad PRISM4 software (GraphPad Software, San Diego, CA). A two tailed Fishers Exact test was used to compare immunohistochemical staining profiles between groups. Significance of survival differences was assessed using log-rank analysis of Kaplan-Meier curves. The formula 1-(alpha) 1/N was used to calculate the fiducial (exact) 95% upper bound of JC virus prevalence. Results p53 protein levels are increased in anaplastic medulloblastoma, ATRT and sPNET We used immunohistochemistry to examine p53 protein expression in representative cores from 80 embryonal brain tumors on a tissue array. Cores from 5 of the cases could not be evaluated due to crush artefact, cautery, or low cellularity. Of the remaining 75 tumors, 64 were medulloblastomas; 31 of these medulloblastoma were of the classic subtype, 13 were nodular/desmoplastic, and 20 were large cell/anaplastic. The tissue array used in this study also contained 11 other CNS embryonal tumors, including 8 sPNET and 3 ATRT. Among the sPNET were 3 with the long epithelial surfaces characteristic of medulloepithelioma. Immunoreactivity for p53 was present in a minority (35%) of the 75 embryonal tumors, and a relatively small subset of cells stained in most of the positive cases. In all positive cases the majority of the immunoreactivity was in tumor cell nuclei. An example of a nodular medulloblastoma with weak, scattered p53 immunoreactivity is shown in Figure 1A . A medulloblastomas and a sPNET with strong immunoreactivity are shown in Figure 1B and 1C , respectively. Only 16% (5/31) of classic medulloblastomas and 23% (3 /13) of nodular medulloblastomas were immunopositive for p53, and staining was weak in all of these cases. In contrast, 45% (9/20) of anaplastic medulloblastomas, 67% (2/3) of ATRT, and 88% (7/8) of sPNET were positive for p53. The intensity of staining in the anaplastic medulloblastomas and extracerebellar embryonal tumors was strong in many cases (Figure 2 ). The increase in p53 immunoreactivity in anaplastic medulloblastomas was statistically significant when compared to non-anaplastic ones, including classic and nodular lesions (P = 0.03, Fisher's Exact test). Other embryonal tumors (sPNET and ATRT) also had a significant increase in p53 expression compared to non-anaplastic medulloblastoma (P = 0.0001, Fisher's exact test). p53 expression was often widespread in the severely anaplastic medulloblastoma, ATRT, and sPNET groups, with 6 of 12 tumors showing immunoreactivity in over 25% of cells, while classic and nodular medulloblastomas always had fewer than 25% immunopositive cells (Table 1 ). Staining for p53 was distributed evenly in most lesions, rather than being concentrated in focal groups of tumor cells. p53 Immunopositivity is significantly associated with worse clinical outcomes We next examined whether p53 immunopositivity was associated with worse clinical outcomes in embryonal brain tumor patients. Of the 75 tumors from 73 patients scored on the array, survival data was available for 66 individuals. 61 percent of patients were alive at last contact, with follow up times ranging from 3 to 215 months (median 47 months). When survival of the entire embryonal brain tumor cohort was analyzed, 67% (31/46) of patients with p53 immunonegative tumors were alive, as compared to 45% (9/20) of patients with immunopositive tumors. Log rank analysis of Kaplan Meier survival curves confirmed the significance of this difference (P = 0.02). When only the 56 medulloblastoma patients with clinical follow up were analyzed, 71% of individuals with p53 negative tumors survived, compared to 58% of individuals with p53 positive tumors. However, these differences were not significant on log rank analysis (P = 0.25). The 6 patients with intensely immunoreactive tumors (3 anaplastic medulloblastoma, 2 sPNET, 1 ATRT) all died from their disease in less than two years. JC virus infection does not account for increased levels of p53 protein To address the possibility that p53 accumulation in medulloblastoma and other embryonal brain tumors is due to viral infection, we used quantitative RT PCR to search for viral sequences in tumor DNA. Of the cases used on the tissue array, 19 had material frozen suitable for high-quality DNA extraction. Four of these cases were p53 positive. Also available in our frozen tumor bank was tissue from 14 additional embryonal lesions (4 classic medulloblastoma, 5 anaplastic medulloblastoma, 1 nodular medulloblastoma and 4 sPNET). We did not detect JC virus sequences in any of these 33 tumors. We repeated the analysis on DNA extracted from a different tumor tissue fragment in 5 cases, but these spatially distinct regions also failed to contain viral DNA. ERV-3, an endogenous human retroviral element, served as a control and was easily amplified from tumor DNA samples (Figure 3 ). ERV-3 copy numbers ranged from ~2 × 10 3 - 2 × 10 5 per PCR assay (mean 3.4 × 10 4 ). Because two copies of ERV-3 are present per human genome, DNA from at least 1000 cells was assayed in each PCR reaction. In contrast, all of the specimens tested were completely negative for JC virus. The quantitative Taqman assay was able to detect a minimum of ~10 copies of JC virus per PCR reaction as determined using a standard dilution curve. Thus the number of JC virus genomes should not exceed 10 per 1,000 tumor cells. Given our 0% detection rate, the 95% confidence limit for the largest value of the 'true' underlying JC virus prevalence is no more than 7.6% (1-(0.05) 1/38 ). Discussion To investigate the prognostic potential and pathological role of p53 expression in embryonal brain tumors, we analyzed this protein in 75 medulloblastoma, ATRT and sPNET using immunohistochemistry. Overall, we found significant increases in p53 immunoreactivity in anaplastic medulloblastomas (45% positive) as compared to non-anaplastic ones (18% positive). The percentage of p53 immunopositive medulloblastomas in our study (27%) fell within the previously reported range of 3% to 53% [ 13 , 16 , 17 , 33 , 34 ]. This wide variation in published values is likely due to differences in antibodies used, their dilutions, and antigen retrieval protocols. Cuttoffs for a calling a tumor "positive" also varied among previous investigators, with some scoring only intensely positive lesions. Only 5% of our medulloblastomas fell into this strong staining category, and all of these were anaplastic. Mutation of the p53 gene often results in a stabilized protein of altered functionality that accumulates in the nucleus of tumor cells [ 35 ]. Our data are thus consistent with those recently reported by Frank and colleagues, who found that the TP53 pathway was frequently disrupted in large cell/anaplastic medulloblastomas [ 9 ]. Interestingly, supratentorial PNET and ATRT were also more commonly p53 immunopositive than non-anaplastic medulloblastoma in our study. While extracerebellar PNET were included in several earlier studies, p53 immunoreactivity was not reported separately for these lesions [ 14 , 15 ]. Ho and colleagues documented p53 mutations in 6 of 14 sPNET but did not examine protein expression [ 36 ]. In another relatively large series, only 1 of 12 sPNET contained a mutation in the p53 gene [ 37 ]. Finally, Postovsky and colleagues described an unusual p53 mutation in a case report of a sPNET [ 38 ]. With regard to rhabdoid lesions, Berrak and colleagues documented faint p53 immunoreactivity in 6 of the 7 ATRT of the CNS they examined [ 39 ]. Malignant rhabdoid tumors arising outside the CNS are also commonly p53 immunopositive, and mutations predicted to inactivate p53 function have been documented in some [ 40 ]. Thus while the number of sPNET and ATRT we examined was relatively low, our data, combined with earlier reports, suggests that p53 function may be commonly altered in embryonal tumors arising outside the cerebellum. Clinical outcomes were significantly worse for embryonal brain tumor patients in our study whose lesions were p53 immunopositive. However, cases most commonly positive for p53 (anaplastic medulloblastomas, sPNET, and ATRT) are all more clinically aggressive than non-anaplastic medulloblastomas, making it difficult to infer causality resulting from p53 accumulation. p53 expression did not predict outcome within the group of medulloblastoma patients, although all 3 strongly p53 immunopositive tumors were severely anaplastic and associated with quite short survival. Interestingly, in a recently published report, Ray and colleagues found that p53 immunoreactivity was the only biological marker predictive of poor outcome on both univariate and multivariate analyses in a group of 112 medulloblastoma patients [ 41 ]. We also examined the potential role of viral infection as a mechanism for p53 protein stabilization in medulloblastoma. None of the four p53-positive cases from our tissue array with frozen material available contained viral sequences. We also tested nine additional samples of histological subtypes that were more commonly associated with p53 immunopositive staining (anaplastic medulloblastomas and sPNETs) where p53 staining was not performed. None of these tumors were JCV positive. While sample availability precluded complete testing of all confirmed p53 positive tumors for JCV, the available data do not support a causal role for JCV infection in p53 accumulation and development of embryonal brain tumors. The impact of viruses on medulloblastoma pathogenesis is controversial. The most commonly implicated agents are the polyomavirus family members JC virus, BK virus and SV40 virus. Of these, JC virus, which infects approximately 80% of the pediatric population and can cause CNS disease in immunosuppressed patients, has been most studied. It was shown decades ago that inoculation of JC virus into rodents resulted in formation of medulloblastoma-like cerebellar tumors [ 20 , 22 , 42 , 43 ]. Transgenic mice containing JC virus early region sequences also develop medulloblastomas [ 44 ]. Del Valle and colleagues isolated JC virus large T antigen sequences from 11 of 23 human medulloblastomas they examined, and suggested inactivation of p53 and Rb by viral T antigen could be important in the pathogenesis of human embryonal brain tumors [ 24 ]. A second gene (Agno) from JC virus was later reported to be present in 11 of 16 medulloblastoma samples by the same group [ 25 ]. We failed to detect JC virus sequences in the 33 cases examined, using a sensitive and specific technique which should identify as few as 10 viral genomes per PCR reaction. This suggests the p53 immunopositivity we observe is not caused by JC virus infection in the majority of cases. Our data are also not consistent with the hypothesis that ongoing JC virus infection is common in medulloblastoma. Other recent studies have also reported a lack of JC virus DNA in medulloblastomas. Hayashi and colleagues failed to identify JC viral sequences in 13 medulloblastomas [ 45 ]. Kim and colleagues similarly failed to identify JC virus in 15 medulloblastomas, 5 sPNET and 2 medulloblastoma cell lines [ 46 ]. Rollison and colleagues examined 225 brain tumors, including 20 medulloblastomas, for JC, BK and SV40 viruses at two different laboratories [ 47 ]. No tumor tested positive in both laboratories. Finally, Weggen and colleagues failed to detect JC virus sequences in any of 116 medulloblastomas analyzed, although 2 of these cases were SV40 positive [ 48 ]. These reports do not rule out the possibility of a "hit and run" process in which virus participates initially in the formation of a lesion and is then lost. They do strongly suggest that ongoing JC virus infection is not common in human medulloblastomas, as was initially suggested. Interestingly, it has recently been shown that the putative SV40 infection of human mesotheliomas can be accounted for by contamination of samples with small amounts of common laboratory plasmids containing regions of the T-antigen gene [ 49 ], calling into question the true association of SV40 with human cancers, including brain tumors. Conclusion In summary, we find significantly increased p53 protein levels in anaplastic medulloblastomas, sPNET, and ATRT as compared to classic and nodular medulloblastoma. JC virus was not detected in the 33 tumors examined, suggesting that T-antigen binding does not appear to be an ongoing factor in the pathobiology or p53 protein accumulation of embryonal brain tumors. Competing interests The author(s) declare that they have no competing interests. Authors' contributions Drs. Eberhart, Shah and Gravitt planned the study and wrote the initial manuscript draft. Dr. Eberhart, Ms. Chaudhry and Ms. Khaki collected samples, isolated DNA, and performed immunohistochemical staining. Dr. Gravitt supervised and R Daniel performed the JC virus copy number analysis in tumors. All authors reviewed and commented on the final manuscript. Pre-publication history The pre-publication history for this paper can be accessed here:	/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC554768.xml
521485	Simple statistical models predict C-to-U edited sites in plant mitochondrial RNA	Background RNA editing is the process whereby an RNA sequence is modified from the sequence of the corresponding DNA template. In the mitochondria of land plants, some cytidines are converted to uridines before translation. Despite substantial study, the molecular biological mechanism by which C-to-U RNA editing proceeds remains relatively obscure, although several experimental studies have implicated a role for cis -recognition. A highly non-random distribution of nucleotides is observed in the immediate vicinity of edited sites (within 20 nucleotides 5' and 3'), but no precise consensus motif has been identified. Results Data for analysis were derived from the the complete mitochondrial genomes of Arabidopsis thaliana , Brassica napus , and Oryza sativa ; additionally, a combined data set of observations across all three genomes was generated. We selected datasets based on the 20 nucleotides 5' and the 20 nucleotides 3' of edited sites and an equivalently sized and appropriately constructed null-set of non-edited sites. We used tree-based statistical methods and random forests to generate models of C-to-U RNA editing based on the nucleotides surrounding the edited/non-edited sites and on the estimated folding energies of those regions. Tree-based statistical methods based on primary sequence data surrounding edited/non-edited sites and estimates of free energy of folding yield models with optimistic re-substitution-based estimates of ~0.71 accuracy, ~0.64 sensitivity, and ~0.88 specificity. Random forest analysis yielded better models and more exact performance estimates with ~0.74 accuracy, ~0.72 sensitivity, and ~0.81 specificity for the combined observations. Conclusions Simple models do moderately well in predicting which cytidines will be edited to uridines, and provide the first quantitative predictive models for RNA edited sites in plant mitochondria. Our analysis shows that the identity of the nucleotide -1 to the edited C and the estimated free energy of folding for a 41 nt region surrounding the edited C are the most important variables that distinguish most edited from non-edited sites. However, the results suggest that primary sequence data and simple free energy of folding calculations alone are insufficient to make highly accurate predictions.	Background RNA editing is the process whereby an RNA sequence is modified from the sequence corresponding to the DNA template. A particular form of RNA editing in plant mitochondria, by which some cytidines are converted to uridines before translation, occurs in many land plant lineages. Although cytidine to uridine conversion is most common, the reverse conversion is sometimes observed [ 1 - 4 ]. In plants, the phenomenon is best studied, albeit still poorly understood, in the mitochondria and plastids of angiosperms [ 5 - 8 ]. The majority of plant mitochondrial RNA editing occurs in coding sequences, and editing frequently changes codons, resulting in changes of amino acids, or, in some cases, creation of entirely new open reading frames [ 1 , 9 , 10 ]. These changes often result in an increase in similarity with respect to homologous protein sequences among different organisms (such as in wheat [ 11 ]), and Gray has postulated that the RNA editing process functions as a repair mechanism to correct otherwise-deleterious genomic mutations [ 12 ]. RNA editing has also been detected in introns, where it is conjectured to improve splicing efficiency [ 13 ]. The precise biochemical basis for C-to-U editing in plant mitochondria is unknown, although experimental evidence suggests a deamination reaction [ 14 - 18 ]. Despite substantial study, the molecular biological mechanism by which C-to-U RNA editing proceeds remains relatively obscure, although several experimental studies have implicated a role for cis -recognition [ 19 - 21 ]. The mechanism by which edited sites are recognized is also still poorly understood, but the importance of surrounding nucleotides has been noted [ 22 ]. A highly non-random distribution of nucleotides in the immediate vicinity of edited sites (within 10–20 nucleotides 5' and 3') is observed, but no precise consensus motif has been identified [ 9 , 16 ]. Additionally, previous studies suggest that inferred secondary structure is not important in site recognition for C-to-U conversion [ 16 , 19 ]. Identifying edited sites thus remains an open problem, one to which we have applied tree-based statistical models and an extension of such models. When applied to a similar problem (predicting peptide binding to major histocompatibility complex (MHC) class I molecules [ 23 ]), tree-based statistical methods generated very accurate models, identifying specific important residues when no precise sequence motif had previously been identified. Therefore, we were motivated to apply tree-based statistical models and an extension, random forests, to the problem of C-to-U RNA editing in angiosperm mitochondria using complete mitochondrial genome data for three species: Arabidopsis thaliana , Brassica napus and Oryza sativa . The objective for the current research was to identify sequence features that may provide insights into C-to-U editing of plant mitochondrial RNA. We address the following specific questions. Is there evidence that sufficient information exists within sequence regions flanking edited sites to accurately predict editing? Is there an association between estimated free energy of folding for short sequence regions containing edited sites and C-to-U editing? We report tree-based statistical analysis of three complete mitochondrial genomes and show that relatively simple models provide moderately accurate prediction of C-to-U edited sites. Results Tree-based statistical models Analysis of each of the three species-specific mitochondrial genome data sets yielded substantially similar results (Table 1 ). Using flanking nucleotides and estimates of folding energy as predictor variables, the optimistic re-substitution-based estimates for cross-validated pruned models had a mean correct classification rate of 0.705 (sensitivity [the proportion of observations correctly identified as edited] = 0.640, and specificity [the proportion of observations correctly identified as non-edited] = 0.883) across the three species. As an additional classification tree analysis, we examined a dataset generated by combining the data from the three species. These results were generally similar to those described above for the mean of the individual genome datasets. The classification tree model is shown in Figure 1 ; the partition is defined based on the nucleotide immediately 5' (-1 position) of the edited/non-edited site. Of the 1972 observations with pyrimidine at the -1 position, 1262 (0.64) are edited and 710 (0.36) are non-edited sites. Of the 722 observations with purine at the -1 position, 85 (0.12) are edited and 637 (0.88) are non-edited sites. Random forests Results from random forests (Table 2 ) were very similar to those obtained with classification trees and were somewhat more accurate. In single-species analyses, the mean accuracy rate was 0.744 (sensitivity = 0.717, specificity = 0.809). Analysis of the larger, combined data set yielded a model better than any of the single genome models with an accuracy of 0.848 (Table 2 ). Analysis of variable importance showed that the -1 position is overwhelmingly the most important factor in determining editing status. Other variables of lesser predictive value include estimated free energy of folding, and the -2 and +1 positions relative to the edited/non-edited site (Figure 2 ). Discussion Despite their simplicity, the tree-based statistical models derived here performed moderately well, with mean accuracies across species generally ~0.71. Single trees were improved upon by constructing models based on ensembles of tree-based models (random forests) each of which was built using random subsamples of the data. This sub-sampling has the effect of reducing the variance through averaging and also reducing the correlation among models. One of the advantages that random forests have over single classification trees is that they provide quantitative measures of variable importance, whereas with a simple classification tree, one is primarily limited to inferring variable importance from the frequency and location of the occurrence of variables in the model. One measure of variable importance is the decrease in the Gini index (a measure of impurity of observations at a particular node) induced by splitting on the variable, averaged over all trees [ 24 ]. In order to infer the relative importance of the predictor variables, we considered the measure of variable importance produced during the random forest run on the combined dataset, which is the most broadly representative dataset considered here. A plot of the variable importance measure for this dataset is shown in Figure 2 ; more important variables are shown as higher bars. The measure strongly indicates that the residue immediately 5' of the edited site (-1 position) is very important. These variable importance results are in agreement with previous work on C-to-U editing in mitochondria of Arabidopsis thaliana, which noted the -1, and -2 positions had highly non-random nucleotide distributions [ 9 ]. However, the results here differ from the past study of Arabidopsis in that we find no indication that the -17 position has much importance in edited site recognition. Also previously noted was that for 93.1% of the time [ 9 ], the -1 position contained a pyrimidine, which is the data partition found by the classification trees. The free energy results contrast with previous studies indicating that secondary structure was not important in edited site recognition [ 16 , 19 ]. Our results show free energy is a relatively important variable in the random forest analyses. These results therefore indicate that secondary structure, as measured by free energy of folding for the 41 nt region centered on an edited/non-edited site, does help in distinguishing edited from non-edited sites. Previous studies determined putative secondary structures for mRNA regions containing edited sites and looked for conserved structural motifs. In contrast, we used estimates of free energy of folding, which are much easier to compare quantitatively. It may be that secondary or tertiary structure is even more important in determining edited sites than shown here; however, secondary structure may not be effectively represented by the calculated estimates of free energy of folding analyzed. Conclusions Simple models based on nucleotides surrounding edited/non-edited sites and on estimated folding energies of those regions provide moderately accurate prediction of C-to-U RNA edited sites. More nuanced representation of secondary or higher-order structure in combination with variables based on the nucleotide positions found important here might improve models. Overall, the results strongly suggest that the C-to-U editing mechanism in plant mitochondria does not depend exclusively on the primary sequence immediately in the vicinity of the edited site. Methods Data sources We obtained complete mitochondrial genome sequences and information regarding edited sites from GenBank [ 25 ] for three species: Arabidopsis thaliana (L.) Heynh. (mouse-ear cress), 455 edited sites, GenBank accession number NC_001284 [ 9 ]; Brassica napus L. (rapeseed), 425 edited sites, GenBank accession number AP006444 [ 26 ]; and Oryza sativa L (rice), 486 edited sites, GenBank accession numbers AB076665 and AB076666 [ 27 ]. None of the GenBank entries noted U-to-C RNA edited sites. Variable selection Incomplete annotations in the GenBank sequences required us to algorithmically determine on which strand an edited site fell (the GenBank files sometimes supplied only a position number, with no strand information). The algorithm, implemented in a Perl script, scanned the entire GenBank file and built an in-memory representation of the layout of all genes and coding sequence regions in the genome. The strand with which an edited site was associated could then be determined by consulting the resultant genome map and checking which strand at the edited site contained a gene region. In no case were genes on both strands at an edited site, so strand localization was always unambiguous. In a few cases, however, a gene containing an edited site could not be located, or a site marked as a C-to-U edit did not contain a C in either strand. In these cases, the supposed edited site was eliminated from further consideration. Final numbers of included sites were as follows: Arabidopsis , 444; Brassica , 422; Oryza , 481. In total, 19 edited sites in the GenBank files were not included across all three species. We also constructed a set of null observations of cytidines that are not edited to uridines. In constructing a null-set, it is important to ensure that the observations are as alike as possible to the edited observations (differing only in the trait to be measured), or the resulting model may be fictive. Here, our null-set observations were non-edited cytidines chosen at random from within gene regions of the genome. Additionally, we chose cytidines such that the null set had exactly the same distribution of codon positions as did the edited set, because the distribution of edited sites within the three possible positions of a codon is highly non-random with a bias to the first two positions [ 9 ] (Table 3 ). For each observation, we recorded 40 nucleotide state variables: one variable for each of the 20 nucleotides sites 5' and 3' of the edited C (on the same strand). We chose a value of 20 for the number of nucleotides 5' and 3' so as to encompass the entire range of semi-conserved positions previously suggested, the most extreme of which occurs 17 bases 5' of the edited site [ 9 ]. In some cases other edited sites occurred within the 20 nucleotides 5' and 3' of the edited site used as a response variable. In these cases the edited sites as predictor variables were recorded as C. The low frequency of these sites at a particular position with respect to other edited sites results in non-significant effects, independent of how these sites are handled. In those cases where a full 20 nucleotides were not included within an annotated mRNA, the missing nucleotides were treated as unknown. Additionally, we included two variables based on free energy expressed in units of kcal/mole at 20°C: the estimated free energy of folding for each 41-nucleotide sequence (20 bases 5', the edited/non-edited base, 20 bases 3') and the change in free energy of folding between the non-edited and edited versions of the 41-nucleotide sequence. Free energies of folding were calculated using mfold [ 28 , 29 ] version 3.1 with program parameters except temperature at default values. Finally, we included codon position as a variable, even though the null set had been chosen so non-edited sites had the same distribution of codon position as the edited sites, as shown in Table 3 . Including codon position as a predictor variable allows for possible interactions with other variables. Finally, we created a combined data set to use alongside the species-specific datasets. The combined dataset is the result of combining all edited sites from all three species (there were no observations identical in all predictor variables), and then randomly selecting negative examples from the set of those already chosen for the three individual datasets. Negative examples were chosen to exactly match the positive examples in distribution over both species and codon position. The combined dataset comprises 2,694 observations. Data analysis Tree-based statistical models We used the R language for statistical computing [ 30 ], version 1.7.1 to conduct our analyses. Analyses included tree-based statistical models using rpart [ 31 ] and random forests using the FORTRAN implementation of random forest version 3.1 [ 24 , 32 ]. Tree-based statistical models [ 33 ], also known as classification and regression trees (CART) [ 34 ], are generated by recursively creating binary partitions of a dataset. Each partition is based on the value of a single predictor variable chosen to best produce homogeneous collections of a nominal or ordinal response variable (classification) or to best separate low and high values of a continuous response variable (regression). More precisely, the partitions may be considered as questions of the following form: Is the observation x i ∈ A ? Where A is a region of the variable space defined by some criterion of a single predictor variable. Answering such a question for all observations produces two groups: those observations for which the answer is yes (those in region A ) and those for which the answer is no (x i ∉ A , those in ). Subsequent binary partitioning continues until stopping criteria (variously defined) are met [ 34 ]. The result is a classification or a regression tree: a hierarchical series of data bifurcations that depicts the partition definitions and describes the resulting data subsets defined by each partition. To address concerns about possible over-fitting models to the data we used 10-fold cross-validation and pruned trees to the shortest within 1- SE of the best tree. We assessed the significance of our tree-based statistical models through permutation where the predictor variables are randomized with respect to the response variable [ 35 ]. The frequency of observing a result value equal to or better than the observed value in 1 × 10 4 permutations is the estimate of the probability associated with the observed result. Random forests If one tree-based statistical model is good, then an ensemble (forest) of appropriately constructed tree models should be even better, which is the principal idea of random forests. A random forest attempts to improve upon a simple tree-based statistical model by generating a collection of such models and using them in aggregate [ 24 , 32 ]. Each model in a random forest is generated from a bootstrap sample of the original dataset, and at each node in each model a search for the best possible split is through a subset of variables selected at random from the bootstrap sample of predictor variables. These randomization steps decrease prediction error through variance reduction resulting from averaging and by decreasing the correlation between individual models in the ensemble [ 36 , 37 ]. Each of our random forest analyses comprised 1 × 10 4 individual models constructed by sub-sampling seven predictor variables at each node. Several model summary statistics were calculated, including sensitivity, which is the proportion of observations correctly identified as edited, specificity, which is the proportion of observations correctly identified as non-edited, and accuracy, which is the total proportion of observations correctly identified. More formally, these definitions are: sensitivity = true positives/ ( true positives + false negatives ); specificity = true negatives/ ( true negatives + false positives ); and accuracy = ( true positives + true negatives )/ total . Authors' contributions MPC conceived, designed and coordinated the study. DSM carried out the programming and statistical analyses. Both authors wrote and approved the final manuscript. Supplementary Material Additional File 1 Arabidopsis thaliana data file File is plain text, space delimited. First row is column headings with variable names: edit; + site is edited, - site is not edited; -20 through 20, nucleotide position relative to edited site; cp, codon position; fe, estimated folding energy; dfe, difference in estimated folding energy between pre-edited and edited sequences; and loc, location of focus site (position 0) in GenBank file. Each subsequent line represents a observation. Click here for file Additional File 2 Brassica napus data file File is plain text, space delimited. First row is column headings with variable names: edit; + site is edited, - site is not edited; -20 through 20, nucleotide position relative to edited site; cp, codon position; fe, estimated folding energy; dfe, difference in estimated folding energy between pre-edited and edited sequences; and loc, location of focus site (position 0) in GenBank file. Each subsequent line represents a observation. Click here for file Additional File 3 Oryza sativa data file File is plain text, space delimited. First row is column headings with variable names: edit; + site is edited, - site is not edited; -20 through 20, nucleotide position relative to edited site; cp, codon position; fe, estimated folding energy; dfe, difference in estimated folding energy between pre-edited and edited sequences; and loc, location of focus site (position 0) in GenBank file. Each subsequent line represents a observation. Click here for file Additional File 4 Combined data file File is plain text, space delimited. First row is column headings with variable names: edit; + site is edited, - site is not edited; -20 through 20, nucleotide position relative to edited site; cp, codon position; fe, estimated folding energy; dfe, difference in estimated folding energy between pre-edited and edited sequences; and loc, location of focus site (position 0) in GenBank file. Each subsequent line represents a observation. Click here for file	/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC521485.xml
534089	Anti-Ri antibodies associated with short-term memory deficits and a mature cystic teratoma of the ovary	Background The IgG autoantibody ANNA-2 (anti-Ri) is a type 2 antineuronal antibody that has been found to bind to highly conserved and widely distributed adult brain proteins encoded by the Nova-1 and Nova-2 genes. Anti-Ri antibodies are typically detected in the serum and cerebrospinal fluids of patients with neurological disorders such as opsoclonus/myoclonus and cerebellar ataxia and in association with gynecologic and breast malignancies. Case Presentation This report describes an unusual example of a 33-year-old female patient who developed short-term memory deficits over a 3-month period. An extensive neurological work-up, including a panel of paraneoplastic markers was negative with the exception of a high titer serum Anti-Ri (1:15,3600). A large left ovarian mass was palpated, surgically resected and eventually diagnosed as a mature cystic teratoma. Post-operatively, memory deficits had disappeared within 1 month and serum Anti-Ri titers had decreased significantly to 1:256. An extensive diagnostic work-up for other malignancies was negative. Conclusion Although, Anti-Ri antibodies are typically associated with malignancies, this case illustrates the potential association between benign tumors and this autoantibody.	Background Paraneoplastic syndromes (PNS) are signs or symptoms attributable to tissue damage at sites that are remote from a primary malignancy or their metastases. PNS involving virtually every level of the neuro-muscular system have been described [ 1 - 3 ]. Typically, the paraneoplastic syndromes causing neurologic disorders precede the diagnosis of the neoplasm and are often the main reasons that medical attention is sought [ 2 , 3 ]. Recent attention has thus centered on discovering novel serum or cerebrospinal fluid markers that can specifically identify not only the presence of a malignancy, but the type of malignancy involved if present. A wide variety of anti-neuronal antibodies have been associated with the many PNS-associated neurologic disorders [ 2 ]. These antibodies have varying degrees of sensitivity and specificity for the underlying types of malignancies. However, in the right setting, the presence of a given auto-antibody in combination with specific signs and symptoms, may reasonably predict the primary site and tumor type. The vast majority of tumors associated with PNS-neurological disorders are malignant [ 2 ]. Notable exceptions include the identification of antibodies to voltage gated potassium channels (VGKC) in patients with thymomas [ 4 ]. We describe in this report the finding of a common anti-neuronal antibody (anti-Ri) in a patient with a benign neoplasm: mature cystic teratoma of the ovary, and whose neurologic symptoms, short-term memory deficits, was apparently associated with her tumor. Clinical history A 33-year-old nulligravid female with no significant past medical history presented to her physician with complaints of a short-term memory loss of approximately three months' duration. This included an inability to remember details about 24-hour old events. There had been no major socio-economic or personal changes in her life over this period. A detailed neurologic examination was notable only for her presenting complaint. Routine laboratory work-up, including a lumbar puncture were all within normal limits. A physical examination revealed a large right adnexal mass, which upon ultrasonographic assessment showed internal features suggestive of a malignancy. A panel of serum paraneoplastic autoantibodies was then requested, including anti-Hu, anti-Yo, anti-Ri, anti-Tr and anti-Ma1/2. All were normal with the exception of IgG anti-Ri, measured at 1:15,360 by an indirect immunofluorescence method. An extensive diagnostic work-up failed to reveal any malignancies. The patient subsequently underwent a right salpingo-oophorectomy, and her adnexal mass was diagnosed as a benign mature cystic teratoma of the ovary. Almost immediately following her surgery, the patient expressed a subjective improvement in her symptoms. Within a month, the serum anti-Ri had decreased to 1:256, and a detailed neurologic examination revealed resolution of her symptoms. She has not experienced any relapse in her symptoms in the 1 year since her surgery. Discussion In 1988 and 1991, Budde-Steffen at al [ 5 ] and Luque et al [ 6 ] described a subpopulation of patients with opsoclonus and a history of breast cancer in whose serum and CSF were identified an antibody that reacted against 55 kD and 80 kD proteins that were designated Anti-Ri (also known as ANNA-2). It has since been shown that these antigens are highly conserved but widely distributed central nervous system neuronal proteins which are encoded by the Nova-1 and Nova-2 genes and which may play a role in neuronal maturation and homeostasis [ 7 - 9 ]. As those seminal reports indicate, Anti-Ri was initially associated with opsoclonus/myoclonus and cerebellar ataxia symptomatology in patients with breast and gynecologic cancers. They have however subsequently been associated with a wide variety of malignancies that have included lung, gastric and bladder carcinomas [ 10 ]. Indeed, in one study, 32% and 36% of cancers associated with anti-Ri in patients with suspected PNS were breast and lung carcinomas respectively [ 10 ]. The spectrum of associated neurologic symptoms has also expanded considerably and now includes vertigo, muscle weakness, dysarthria, dysphagia, dementia, deafness, myelopathy, opthalmoplegia, encephalomyelitis, rigidity, nausea, myelopathy, sensorimotor neuropathy [ 10 ]. What these cases illustrate is that with a few exceptions [ 4 ], the vast majority of tumors associated with PNS-neurological disorders are malignant. Our case thus highlights the potential association between a benign neoplasm and the presence of these antibodies and neuronal symptomatology. Is it possible that the presence of the autoantibody and the ovarian tumor are completely fortuitous?. Given the temporal relationship between the resolution of her symptoms and the sharp decrease in her anti-Ri titer following her surgery, we believe this is unlikely. However, the precise mechanistic basis for this association as well as potential influence of outside factors remains to be elucidated. It should also be noted that high titers of Anti-Ri have been identified in patients with a history of ovarian cancer but without any evidence of a PNS, suggesting caution in assessing the specificity of this auto-antibody for PNS. In summary, we describe in this report an association of a common anti-neuronal antibody (anti-Ri) in a patient with a benign neoplasm, mature cystic teratoma of the ovary, and whose neurologic symptoms, short-term memory deficits, was apparently associated with her tumor. Although, Anti-Ri antibodies are typically associated with malignancies, this case illustrates the potential association between benign tumors and this autoantibody. Competing interests The authors declare that they have no competing interests. Authors' contributions Dr Fadare wrote the initial version of the manuscript. Dr Hart managed the patient, provided clinical information, and revised the manuscript.	/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC534089.xml
554998	Salivary changes and dental caries as potential oral markers of autoimmune salivary gland dysfunction in primary Sjögren's syndrome	Background the classification criteria for primary Sjögren's syndrome (pSS) include a number of oral components. In this study we evaluated if salivary flow and composition as well as dental caries are oral markers of disease severity in pSS. Methods in 20 patients fulfilling the American-European Consensus criteria for pSS and 20 age-matched healthy controls whole and parotid saliva flow rates and composition, measures of oral dryness, scores of decayed, missing and filled tooth surfaces (DMFS), periodontal indices, oral hygiene, and dietary habits were examined. Results in pSS, salivary flow rates, pH, and buffer capacities were lower, and DMFS, salivary sodium and chloride concentrations higher than in the healthy controls. DMFS also correlated inversely to salivary flow rates and positively to oral dryness. Apart from slightly increased gingival index, and more frequent dental visits in pSS, the periodontal condition, oral hygiene or sugar intake did not differ between these two groups. In pSS, findings were correlated to labial salivary gland focus score (FS) and presence of serum-autoantibodies to SSA/SSB (AB). The patients having both presence of AB and the highest FS (>2) also had the highest salivary sodium and chloride concentrations, the lowest salivary phosphate concentrations, lowest salivary flow rates, and highest DMFS compared to those with normal salivary concentrations of sodium and chloride at a given flow rate. Conclusion the salivary changes observed in some pSS patients reflect impaired ductal salt reabsorption, but unaffected acinar transport mechanisms, despite low salivary secretion. Our results suggest that changes in salivary flow and composition as well as dental caries may serve as potential markers of the extent of autoimmune-mediated salivary gland dysfunction in pSS. The study also indicates that the ductal epithelium is functionally affected in some pSS patients, which calls for future pathophysiological studies on the mechanisms underlying this impaired salt reabsorption.	Background Primary Sjögren's syndrome (pSS) is a common systemic autoimmune disease that mainly affects middle-aged women. The exact aetiology remains elusive, and appears multifactorial. The disease is characterised by a chronic periductal lymphocytic infiltration of the exocrine glands, predominantly of the lacrimal and salivary glands. The chronic inflammation is assumed to lead to glandular tissue destruction, resulting in keratoconjunctivitis sicca and hyposalivation and associated symptoms of ocular and oral dryness [ 1 , 2 ]. The salivary dysfunction in pSS is of significant clinical importance and may cause chronic oral discomfort as well as compromised oropharyngeal functions [ 3 - 5 ]. Subsequently, the patients' well-being and oral health-related quality of life are also impaired [ 5 , 6 ]. Cross-sectional studies have shown that patients with pSS have a high caries experience with lesions usually located on the cervical surfaces of the teeth, as well as early dental loss due to caries [ 3 , 7 - 10 ], despite good oral hygiene [ 11 ]. Dental caries, which is a local, but also a multifactorial disease, arises from a concerted action of the aetiological factor (dental plaque) and determinants including salivary factors, medication, systemic disease as well as behavioural factors [ 12 ]. However, the selective significance of salivary factors on dental caries in pSS remains an open question. In light of the impaired saliva formation in pSS it is noteworthy that an exact cut-off value for salivary flow and/or separate salivary constituents predicting the risk of developing dental caries has not yet been agreed on. In the diagnosis of Sjögren's syndrome (SS), an unstimulated whole saliva flow rate of ≤ 1.5 ml/15 min is currently considered pathological [ 13 ], but at this point the caries process may have been going on for years revealed by the presence of a high number of decayed, missed and filled teeth [ 3 , 7 - 10 ]. The American-European Consensus Classification Criteria for pSS include evaluation of salivary gland dysfunction by means of parotid sialography, salivary gland scintigraphy, whole saliva sialometry and labial salivary gland biopsy as well as questionnaire concerning symptoms of oral dryness [ 13 ]. The purpose of this study was to evaluate parotid flow rate, whole and parotid saliva composition and dental caries as potential oral markers of disease severity in patients with pSS. Thus we hypothesised that patients with pSS exhibit specific changes in salivary flow and composition due to presence of focal lymphocytic infiltration in their labial salivary glands (FS) and/or serum autoantibodies to Ro (SSA) and/or La (SSB) antigens (AB) as compared to age- and gender-matched healthy controls with normal salivary gland function. We further hypothesised that in patients with pSS dental caries is related to impaired salivary flow and associated changes in saliva composition. Methods The study included 20 randomly selected patients with pSS, who attended the Department of Oral Medicine, University of Copenhagen, for routine follow-up examinations. They were diagnosed according to the Copenhagen criteria, but also met the American-European Consensus Classification Criteria for pSS [ 13 , 14 ]. All patients had complaints of dry eyes and dry mouth according to the questionnaire of the American-European classification criteria [ 13 ] as well as evidence of keratoconjunctivitis sicca. Eighteen patients had unstimulated whole saliva flow rate (UWS) ≤ 1.5 ml/15 min. Two patients had UWS >1.5 ml/15 min, but otherwise fulfilled the criteria. Fourteen patients had a positive labial salivary gland biopsy with focus score ≥ 1 as well as presence of AB. Among the remaining 6 patients, 2 had FS, but not presence of AB, and 4 patients had presence of AB, but not FS. The labial salivary glands of the latter 4 patients displayed chronic sialoadenitis where inflammatory cells were observed but did not form any foci. In all 20 patients, estimation of the presence of serum autoantibodies was performed within the same timeframe as the saliva collections and the clinical assessments. In 16 of these patients labial salivary gland biopsies were performed at the same time. The remaining 4 patients refused a labial salivary gland rebiopsy, as they had their diagnosis of pSS for 25, 23, 16 and 10 years, respectively. In order to understand the underlying salivary pathophysiology in pSS, the study also included 20 healthy controls selected among persons attending the Institute of Odontology for dental treatment to obtain reference values of normalcy. The selection was based on the requirements that controls had no present or past medical history of systemic disease, did not take any medication including contraceptives and hormonal replacement, had no current use of tobacco products, and matched the pSS patients with regard to age. Clinical examination One examiner (AMLP) conducted the whole saliva collections, interviews, and oral clinical examinations. About 2 hours after collection of whole saliva, another examiner (AB) conducted collections of parotid saliva. In order to minimize the influence of circadian cycle on salivary secretion and composition all procedures were carried out in the same order and at a fixed time of the day, i.e. between 9.00 and 11.45 a.m. The local Danish Ethical Committees approved the study protocol, and all participants completed an informed consent form according to the Declaration of Helsinki. The number of decayed, missing and filled surfaces (DMFS) was recorded excluding third molars as previously described in details [ 5 ]. Plaque and gingival indices and periodontal probing pocket depth were determined at four sites per tooth on six index teeth (16, 21, 24, 36, 41, and 44) [ 8 ]. Each subject underwent a standardised interview including inquiries on medical disease symptoms, dental history, dental visits, medication, tobacco smoking, dietary (especially regarding daily sugar intake) and oral hygiene habits. Furthermore, symptoms of oral dryness were assessed by means of a categorised questionnaire (based on Beck's inventory scale, item 9 [ 15 ]) with four degrees of severity (scores 0–3), as previously described [ 8 ]). Collection of whole and parotid saliva Unstimulated whole saliva (UWS) and paraffin-stimulated whole saliva (SWS) were sampled over a 15-min and 5-min period, respectively, as described previously [ 8 ]. Unstimulated parotid saliva (UPS) and stimulated parotid saliva (SPS) were sampled over an average period of 20 min and 5 min, respectively. A saliva collection set-up impermeable to CO 2 was used [ 16 ]. Stimulation of the parotid saliva was initiated by applying 1 ml of 1% citric acid to the dorsal part of the tongue every 15 sec. The flow rate, given with 2 decimal places, was determined by weighing the saliva-collecting cup, tube, and syringe before and after saliva collection, and expressed as ml/min [ 8 ]. The subjects were instructed to refrain from eating, drinking, smoking, and any oral hygiene for 2 h preceding the saliva sampling. Sialochemical analysis Concentrations (mM) of sodium, potassium and total calcium were determined by atomic absorption spectroscopy [ 16 , 17 ]. Concentrations (mM) of chloride were measured by coulometric titration and total phosphate (mM) by the molybdic reaction [ 16 , 17 ]. Total protein (μg/ml) was measured by the Coomassie reaction. Amylase activity (stated as the catalytic activity of the enzyme, i.e., kat = mole/s and given in μkat/l, where one μkat/l corresponds to 60 U/l) was measured by means of the Phadebas™ test kit. Saliva was diluted 1000 times more than usually recommended for plasma by the test kit. The volume of saliva required for complete sialochemical analysis was 240 μl corresponding to an average collection period for UPS of less than 20 min and for SPS less than 5 min for the pSS patients. Parotid saliva pH, P CO2 , buffer capacity, and saturation with regard to hydroxyapatite Parotid saliva pH and P CO2 were determined on an ABL 605 blood gas analyser (Radiometer™), and the HCO 3 - concentration was calculated from the pH and Pco 2 values as previously described [ 17 ]. The parotid saliva buffer capacity (β), which originated from the HCO 3 - and phosphate buffer systems, was calculated individually for each saliva sample at the pH value of the sample [ 18 ]. Briefly, the buffer capacity for each buffer system was calculated as: 2.3 [C] * ([C B ]/[C]) * (1-([C B ]/[C])) Where ([C]) states the total concentration of the buffer system and ([C B ]) states the concentration of the base in the buffer system. The concentrations of the bases was calculated as previously described [ 16 ] and the pK values used for the bicarbonate and phosphate buffer systems was 6.1 for carbonic acid [ 16 , 17 ] and 6.8 for H 2 PO 4 - [ 19 ]. The sums of calculated β from both the HCO 3 - and phosphate buffer systems are in this study denoted as saliva β. At pH values above 5 (counting for all except one sample) the contribution from the protein buffer system, relative to the HCO 3 - and phosphate buffer systems, was considered negligible [ 16 ] and therefore not included in the saliva β. The degree of saturation (DS HAP ) of saliva with respect to hydroxyapatite, i.e. the calcium phosphate salt and main mineral of the tooth tissues, was calculated according to the method described by Schmidt-Nielsen [ 19 ]. The negative logarithm of the ionic product of hydroxyapatite (pI HAP ) was calculated from the saliva p [Ca 2+ ], p [HPO 4 2- ], and p [H + ]. The negative logarithm of the solubility product of hydroxyapatite (pSP HAP ) was calculated by the constant given by Schmidt-Nielsen [ 19 ] and the ionic strength of each saliva sample. Given the relatively low ionic strength of the saliva samples analysed (0.039 ± 0.020) we found this simple method for calculation of DS HAP acceptable. Finally, the critical pH of parotid saliva, i.e. the pH value at which saliva is saturated (i.e., neither super- nor undersaturated) with respect to hydroxyapatite (pI HAP equal to pSP HAP ) was calculated. Statistical analysis Differences in the salivary and clinical parameters between the patient and control groups were analysed by Wilcoxon rank-sum test (categorised variables) and two-sampled t -test (numerical variables). Fisher's test (less than ten in one category) and the chi-squared test (more than ten) were used for analysis of distributions between the two groups. Associations between variables were analysed by the Spearman rank order correlation analysis ( r s ). In order to find the best predictor for a given outcome, multiple regression analysis was used with stepwise backward elimination with the adjusted R-squared values given. Significance was selected at a level of P ≤ 0.05. Results Clinical findings Table 1 summarises some of the anamnestic data and results of self-assessed oral dryness of the 20 female patients with pSS and 20 age-matched healthy females included in this study. Sixteen pSS patients were taking prescribed medicines on a regularly daily basis. Seven of these patients were taking medicines known to impair saliva secretion and/or cause compositional changes of saliva [ 20 ]. These included antihistamines and antidepressants that cause reduction in salivary flow due to inhibition of the muscarinic cholinergic receptors [ 21 ], and blockers of β-adrenoceptors that lead to impaired protein secretion [ 22 ]. However, none of these medicines are known to cause specific changes in the inorganic salivary composition. The 7 patients taking "xerogenic" medicines all had symptoms of oral dryness and/or were diagnosed as pSS with hyposalivation (n = 5) before their current medication was commenced. We therefore found no significant association between the intake of medicines, including the number of medicines, the salivary flow rates, focus scores, and serum autoantibodies. All the pSS patients complained of dry mouth. In pSS, a significant inverse correlation was found between scores of oral dryness and the SWS ( r s = -0.50, P = 0.025), but only a tendency toward a relationship between the oral dryness and UWS, UPS and SPS. On the other hand, in both groups as a whole, displaying large variations in oral dryness and salivary flow rates, the subjective measures were significantly inversely correlated to UWS ( r s = -0.83, P < 0.001), SWS ( r s = -0.81, P < 0.001), UPS ( r s = -0.74, P < 0.001), and SPS ( r s = -0.77, P < 0.001). Table 1 Anamnestic data and results of self-assessed oral dryness of the female patients with primary Sjögren's syndrome (pSS) and the female healthy controls. Results are given as number of patients or scores (yes/no) and as means ± SD. pSS ( n = 20) Healthy controls ( n = 20) P-value Age (years) 60 ± 15 56 ± 13 0.437 a Duration of disease (years)* 6 ± 7 0 NA Duration of symptoms (years) 10 ± 7 0 NA Xerogenic medicines (yes/no) † 7/13 0/20 NA Smokers (yes/no) § 5/15 0/20 NA Cigarettes per day (smokers only) § 7 (2–20) 0 NA Tooth brushing (times per day) 3 ± 1 2 ± 1 0.001 a Dental floss and/or toothpicks daily (yes/no) 16/4 15/5 1.000 b Dental visits per year (number) 3 ± 1 2 ± 1 0.001 a Oral dryness (score 0/1/2/3) 0/3/4/13 20/0/0/0 <0.001 b * The time from established diagnosis to present examination. † Patients were interviewed regarding their intake of medicines. Their intake of xerogenic medicines included antidepressants, antihistamines and β-blocking agents [20]. § Patients were also interviewed regarding their smoking habits, since cigarette smoking has been shown to reduce the extent of labial salivary gland lymphocytic infiltration [23] and to have a negative impact on periodontal status. Smoking, however, has no impact on salivary flow or composition [24]. P-values obtained by a two-sample t-test (a) and Fisher exact test (b). NA; not analyzed. The results of the clinical examination are shown in Table 2 . In spite of the fact, that the patients brushed their teeth with fluoride-containing toothpaste and visited their dentist more frequently than the healthy individuals (Table 1 ), they had a significantly higher number of decayed, missed and filled teeth and a higher gingival index. No significant differences were found between the two groups in terms of plaque index (frequency distribution of scores), the periodontal probing pocket depth or the regular daily use of dental floss and/or toothpicks. Apart from avoiding acidic, spicy, crunchy and dry foods the patients' dietary habits (including sugar intake) did not differ from those of the healthy controls. Neither the time since pSS diagnosis, nor the duration of disease symptoms or the age, were correlated to the flow rates. Moreover, plaque- and gingival indices as well as periodontal probing pocket depth were not correlated to the salivary flow rates, time since pSS diagnosis, duration of disease symptoms, DMFS, or age. In the groups as a whole, however, a positive association between the gingival index and age was found ( r s = 0.33, P = 0.04). Table 2 Oral findings in the patients with primary Sjögren's syndrome (pSS) and the healthy controls. Results are given as number of patients or scores (yes/no) and in medians (ranges). pSS ( n = 20) Healthy controls ( n = 20) P-value No. of teeth 22 (6–28) 28 (0–28) 0.011 a DMFS* 83.0 (25–140) 43.0 (10–140) 0.001 a Distribution of D/M/FS 23/745/1020 14/305/662 <0.001 b Subjects with dentures (yes/no) 4/16 7/13 0.480 c Distribution of PI (no. of 0/1/2/3 scores) 245/179/54/2 244/152/60/0 0.272 b Distribution of GI (no. of 0/1/2/3 scores) 346/116/16/2 330/95/31/0 0.034 b Probing pocket depth (PPD, mm) 2 (1–4) 2 (1–3) 0.523 a D/M/FS, decayed, missed, filled surfaces; PI, plaque index; GI, gingival index. P-values obtained by Wilcoxon rank sum test (a), Chi 2 - test (b), and Fisher exact test (c). Salivary flow and composition As shown in Tables 3 and 4 , the pSS patients had significantly lower UWS, SWS, UPS and SPS than the healthy controls. All pSS patients and one single healthy individual (without symptoms or clinical signs of local or systemic disease) had with UWS ≤ 0.10 ml/min and/or SWS ≤ 0.70 ml/min, which are the cut-off values for hyposalivation [ 13 , 25 ]. In 8, 11 and 5 patients, UWS, UPS and SPS, respectively, were close to or equal to 0 ml/min. In both the patient and control group UWS and SWS were mutually correlated ( r s = 0.79, P < 0.001 and r s = 0.45, P < 0.05, respectively), but only in pSS a mutual correlation was obtained between UPS and SPS ( r s = 0.74, P < 0.001). Table 3 Unstimulated (UWS) and stimulated (SWS) whole salivary flow rate and composition in the patients with primary Sjögren's syndrome (pSS, n = 20) and the healthy controls ( n = 20). Results are given in median (range). UWS SWS pSS Healthy controls P-value pSS Healthy controls P-value Flow rate (ml/min) 0.02 (0.00–0.23) ( n = 20) 0.39 (0.06–1.10) ( n = 20) <0.001 0.14 (0.01–1.66) ( n = 20) 1.40 (0.54–2.82) ( n = 20) <0.001 Na + (mM) 12.0 (8.0–47.0) ( n = 11) 8.0 (4.5–17.0) ( n = 20) 0.008 16.0 (8.0–59.0) ( n = 12) 10.3 (5.5–30.0) ( n = 20) 0.020 K + (mM) 22.2 (5.1–48.0) ( n = 11) 21.9 (6.8–33.4) ( n = 20) 0.536 21.6 (15.2–34.6) ( n = 12) 22.0 (11.5–26.5) ( n = 20) 0.533 Total calcium (mM) 2.0 (1.3–3.2) ( n = 7) 1.7 (0.5–2.8) ( n = 19) 0.285 1.3 (0.7–1.9) ( n = 12) 1.3 (0.8–2.5) ( n = 20) 0.558 Cl - (mM) 25.8 (16.0–62.0) ( n = 10) 18.3 (5.2–26.0) (n = 19) 0.003 22.6 (13.5–58.6) ( n = 11) 16.1 (10.5–28.6) ( n = 20) 0.037 Total phosphate (mM) 6.0 (1.2–15.0) ( n = 9) 6.8 (2.4–11.7) ( n = 19) 0.961 4.4 (2.4–8.5) ( n = 11) 4.1 (1.6–11.5) ( n = 20) 0.606 Total protein (mg/ml) 3.00 (1.52–8.84) ( n = 8) 3.29 (1.36–6.25) (n = 19) 0.560 3.50 (0.421–8.93) ( n = 12) 3.04 (1.23–5.54) ( n = 20) 0.491 Total protein output (mg/min) 0.210 (0.06–0.80) ( n = 8) 1.09 (0.50–2.38) ( n = 19) <0.001 0.95 (0.11–4.33) (n = 12) 3.88 (2.35–8.06) ( n = 20) <0.001 Amylase activity (μkat/l) 365 (0–4500) ( n = 10) 980 (0–8940) ( n = 19) 0.155 1490 (0–4080) ( n = 12) 2280 (0–5890) ( n = 20) 0.199 P-value obtained by Wilcoxon's rank sum test. Table 4 Unstimulated (UPS) and stimulated (SPS) parotid salivary flow rates, composition including pH, P CO2 , buffer capacity, degree of saturation with regard to hydroxyapatite, and critical pH in the patients with primary Sjögren's syndrome (pSS, ( n = 20)) and the healthy controls ( n = 20). Results are given in medians (range). UPS SPS pSS Healthy controls P-value pSS Healthy controls P-value Flow rate (ml/min/gland) 0.00 (0.00–0.04) ( n = 20) 0.04 (0.00–0.13) ( n = 20) <0.001 0.10 (0.00–0.87) ( n = 20) 0.72 (0.17–1.57) ( n = 20) <0.001 pH 5.5 (4.9–6.2) ( n = 7) 6.1 (5.4–6.6) ( n = 14) 0.008 6.8 (5.3–7.6) ( n = 13) 7.1 (6.8–8.0) ( n = 19) 0.020 P CO2 (KPa) 2.5 (2.1–4.8) ( n = 7) 3.3 (0.5–7.9) ( n = 14) 0.455 3.5 (1.4–7.5) ( n = 13) 5.5 (0.3–10.0) ( n = 19) 0.021 Na + (mM) 7.0 (0.0–111.0) ( n = 9) 0.0 (0.0–8.5) ( n = 19) 0.002 15.5 (3.5–93.0) ( n = 15) 7.5 (0.0–65.0) ( n = 20) 0.034 K + (mM) 24.5 (4.5–52.5) ( n = 9) 29.0 (17.5–62.0) ( n = 19) 0.209 26.5 (12.0–53.0) ( n = 15) 23.0 (18.0–40.0) ( n = 20) 0.689 Total calcium (mM) 1.0 (0.4–2.4) ( n = 9) 1.0 (0.6–4.1) ( n = 16) 0.887 1.1 (0.6–1.8) ( n = 13) 0.8 (0.4–1.4) ( n = 20) 0.008 Cl - (mM) 22.5 (5.0–100.5) ( n = 9) 19.0 (12.0–43.0) ( n = 19) 0.588 27.0 (12.5–94.0) ( n = 14) 15.5 (9.0–59.0) ( n = 20) 0.037 Bicarbonate (mM) 0.2 (0.0–0.8) ( n = 8) 0.8 (0.1–3.5) ( n = 14) 0.017 3.6 (0.1–26.8) ( n = 12) 11.3 (3.8–43.9) ( n = 19) 0.025 Total phosphate (mM) 5.6 (2.3–15.0) ( n = 7) 9.3 (4.1–18.9) ( n = 16) 0.066 6.1 (2.1–8.2) ( n = 11) 5.2 (3.2–16.9) ( n = 20) 0.577 Total protein (mg/ml) 1.11 (0.47–8.97) ( n = 9) 2.02 (0.64–5.34) ( n = 18) 0.316 1.33 (0.62–6.66) ( n = 14) 1.60 (0.84–4.19) ( n = 20) 0.589 Total protein output (mg/min) 0 (0–0.18) ( n = 9) 0.07 (0–0.30) ( n = 18) <0.001 0.13 (0–1.38) ( n = 14) 1.30 (0.29–5.42) ( n = 20) <0.001 Amylase activity (μkat/l) 7760 (3700–21200) ( n = 7) 5815 (1670–27140) ( n = 18) 0.193 5220 (560–17240) ( n = 13) 5520 (710–14710) ( n = 20) 0.624 β(mmol H + /l . pH unit) 1.2 (0.1–2.9) ( n = 7) 4.0 (2.3–8.9) ( n = 13) <0.001 4.4 (0.5–6.7) ( n = 10) 5.2 (0.7–10.1) ( n = 19) 0.171 DS HAP 1.4 (0.2–3.6) ( n = 9) 3.1 (2.5–10.4) ( n = 16) 0.001 7.3 (0.6–21.4) ( n = 13) 10.0 (4.5–22.0) ( n = 19) 0.372 Critical pH 5.4 (5.1–6.2) ( n = 7) 5.3 (5.0–5.5) ( n = 16) 0.413 5.5 (5.2–6.2) ( n = 11) 5.7 (5.2–5.8) ( n = 19) 0.611 The buffer capacity (β) was calculated from the bicarbonate and phosphate concentrations individually for each saliva sample at its pH value, DS HAP denotes the degree of saturation with regard to hydroxyapatite. P-value obtained by a Wilcoxon's rank sum test. * corresponds to mean value below 0.01 ml/min. In 10, 9, 13 and 7 cases, respectively, volume of UWS, SWS, UPS and SPS samples from pSS patients were insufficient for full sialochemical assessment. In pSS, the concentrations of sodium were significantly higher in both whole and parotid saliva as compared to the healthy controls. Similar results were obtained for chloride concentrations (except for the UPS) (Tables 3 and 4 ). Regarding the concentrations of the other electrolytes (potassium, total calcium and total phosphate) as well as the organic components (total protein and amylase activity) no statistically significant differences were found between the two groups. In order to test whether the changes in sodium and chloride concentrations were merely related to changes in salivary flow rates, the patients and healthy controls were matched by their UPS and SPS flow rates within the range of 0.01–1.00 ml/min. Figures 1A and 1B illustrate that the pSS patients may be divided into two subgroups. In one group of patients the salt reabsorption pattern was similar to that of the healthy controls, whereas another group displayed high concentrations of parotid sodium and chloride despite low parotid flow rates. Figure 1 shows the saliva composition as a function of the parotid saliva flow rate (both unstimulated and stimulated) in patients with pSS (bold circles) and healthy age-matched controls (open circles). Lines were fitted by linear (a and b) or non-linear regression (c-h) depending on the best obtainable fit judged from the R-squared values with unbroken lines representing pSS and dotted lines the controls. Figures 1a and b illustrates that the pSS patients can be divided into two subgroups with regard to salt reabsorption. One group of patients follow the reabsorption pattern of the healthy controls, whereas the other group display high sodium and chloride concentrations despite low parotid flow rates indicating differences in the degree of glandular affection by the disease of the patients studied. The pH and bicarbonate concentration in parotid saliva were significantly lower in the pSS patients as compared to the healthy controls (Table 4 ). Subsequently, the pSS patients also had a significantly reduced buffer capacity (β) of UPS and a lower degree of saturation with regard to hydroxyapatite (DS HAP ) in their UPS than the controls. In pSS, the pH in UPS was in fact only 0.1 pH units from their critical pH. Overall, the salivary β and DS HAP were reduced due to low pH, bicarbonate and phosphate concentrations. In pSS, P CO2 of SPS was lower than in that of the healthy controls ( P = 0.02). Nevertheless, at corresponding parotid flow rates, be it stimulated or not, the relations between the flow rates, the saliva pH and the bicarbonate concentrations were similar to those of the healthy controls (Fig. 1C and 1D ). Also the concentrations of potassium, total phosphate, total calcium and total protein (Fig. 1E, F, G , and 1H ) were similar to those of the controls. Determinants of DMFS As shown in Table 2 the pSS patients had significantly higher DMFS than the healthy controls (p < 0.001) due to higher number of filled (51 ± 23) and missed (37 ± 35) tooth surfaces than the controls (33 ± 19 and 15 ± 35, respectively) ( P < 0.05). Also the prevalence of dental treatment, i.e. the number of filled surfaces relative to the total number of present tooth surfaces, was significantly higher in the patients (55 ± 27%) than in the controls (25 ± 14%) ( P < 0.001). However, no significant differences were obtained for the number of decayed surfaces between the patients and controls. Table 5 shows the explanatory power of the two sets of variables exhibiting the highest explanatory power on DMFS in the pooled patient and control groups, namely complaints of oral dryness and the actual salivary flow rates. The estimate states the theoretical expected increase or decrease in DMFS upon an integer increase of the explaining variables, i.e. one score of subjective oral dryness or one ml of saliva. Furthermore, the subjects' age were added to the analyses due to the well-known effect of this variable upon the number of missing surfaces, which was also found to be significant in this study ( r s = 0.63, P < 0.001). As shown DMFS increases with increasing age (per year) and increasing complaints of oral dryness, but decreases with increasing salivary flow rates. Table 5 Explanatory power of age as well as subjective and objective salivary variables on DMFS in the patient group and control group all together ( n = 40). Estimate for DMFS SD for DMFS estimate Adjusted R-squared P-value Subjective model Age 1.4 0.4 0.26 <0.001 + oral dryness questionnaire 11.1 2.3 0.53 <0.001 Objective model Age 1.4 0.4 0.26 <0.001 + UWS -57.4 20.7 0.37 <0.001 + SWS -28.9 9.4 0.49 <0.001 + UPS -499.2 173.0 0.57 <0.001 Estimate, SD, adjusted R-squared, and P-value obtained by multiple regression analysis. The estimate states the increase or decrease in DMFS upon addition of the explaining variables (age + oral dryness, or + UWS, SWS, and UPS) by an integer (one year, score, or ml). Since complaints of oral dryness and salivary flow demonstrated strong interrelations they could not be included in the same analysis. As shown age in combination with complaints of oral dryness explained 53% of the variance in DMFS, and age combined with salivary flow rates explained 57%. In the pSS patients, regardless of their age, the complaint of oral dryness, but not the actual salivary flow rate, was associated with dental status (DMFS). The lack of significant correlation between flow rate and DMFS may be ascribed to the narrow window of flow rates in pSS. Scores of oral dryness correlated significantly with the total DMFS ( r s = 0.53, P < 0.05) and the prevalence of dental treatment ( r s = 0.51, P < 0.05). In addition, the prevalence of dental treatment increased with the duration of pSS ( r s = 0.49, P < 0.05), independently of the patient's age. Salivary gland focus score, serum autoantibodies, salivary flow and composition In the pSS group, the focus scores (which ranged from 0 to 4) were significantly inversely correlated to UWS ( r s = -0.53, P < 0.05), SWS ( r s = -0.68, P < 0.01), SPS ( r s = -0.57, P < 0.01), but not to UPS. Focus scores were not correlated to oral dryness, time since pSS diagnosis, and duration of disease symptoms. It is noteworthy, however, that the patient group exhibited large variability with regard to time since diagnosis and age (Table 1 ). The number of focus scores and presence of serum autoantibodies were positively correlated to salivary concentrations of sodium and chloride in SPS ( r s = 0.61, P < 0.05 and r s = 0.71, P < 0.01). Although highly increased sodium and chloride concentrations were found in UPS as well, the number of observations was too small to generate statistical significance due to the sample volumes that did not allow for sialochemistry (Table 4 ). Results further revealed that pSS patients with the lowest UWS (i.e., 0–0.05 ml/min) also had the highest focus scores (mean focus score 2.3 vs . 0.7, respectively, P < 0.05), presence of serum autoantibodies (n = 14), highest concentrations of salivary sodium and chloride, but lowest salivary concentrations of total phosphate compared to patients with UWS ≥ 0.05 ml/min. Pooling data of UPS and SPS for the pSS group allowed us to compare profiles of the patients with high sodium and chloride concentrations to those with low ones. The results of this analysis revealed that pSS with more than 40 mM sodium in their parotid saliva also had significantly higher focus scores as well as presence of serum autoantibodies, than pSS patients below this value ( P < 0.01). The former also tended to have lower parotid flow rates, more decayed tooth surfaces and to be younger than the other pSS patients. Discussion Oral dryness and reduced salivary flow rates are some of the most predominant and troubling oral sequelae of pSS. This discussion focuses on the hypothesis if impaired salivary flow and/or composition and high caries experience may serve as potential markers of disease severity in pSS. Caries experience, oral hygiene, and salivary flow rates The present study confirms the results of previous cross-sectional studies showing that pSS patients have a significantly higher DMFS compared to healthy controls, despite the fact that these patients often have a good oral hygiene and frequent dental follow-up visits [ 5 , 9 , 10 ]. When comparing SS patients with patients suffering from other immune diseases and patients with xerostomia of other origins, Boutsi et al . [ 26 ] found no significant differences in the number of decayed, missed or filled teeth. In accordance with our results, the SS patients were characterised by having lower salivary flow rates, better oral hygiene habits, slightly higher gingival scores, but similar plaque scores, compared to the other groups [ 26 ]. Regarding the other periodontal measures, our results support those of previous studies showing that presence of periodontal disease is not substantially increased in pSS [ 2 , 26 ]. The slightly increased gingival scores could be explained by altered inflammatory response probably due to hormonal changes [ 27 ]. It has been reported that pSS patients harbour lower numbers of periopathogenic microorganisms than healthy controls [ 28 ]. In contrast, and despite a good oral hygiene, pSS patients appear to harbour higher numbers of cariogenic and acidophilic microorganisms such as Streptococcus mutans and Lactobacillus species than healthy controls [ 11 ]. Similar relationships between low salivary flow rates and increased Lactobacillus counts have also been observed in patients with hyposalivation of other origins [ 29 ]. Thus, impaired salivary flow and changes in saliva composition as seen in pSS are assumed to favour a more aciduric oral microflora manifested by an increased incidence of caries and fungal infections [ 28 ]. At low flow rates, the bicarbonate concentration, pH, and buffer capacity as well as the clearance of microorganisms and dietary sugars in the oral cavity generally decrease [ 29 ], thereby promoting an environment dominated by these oral pathogens and prolonged exposure of dietary sugars to the teeth. In pSS patients with severely reduced salivary flow rates the shift in the oral microflora appears to occur despite a good oral hygiene. In addition, the high number of microbial retention sites generated by dental restorations such as fillings, crowns and bridges found in pSS patients may contribute to the shift in oral microflora [ 28 ] and complicate the maintenance of sufficient oral hygiene procedures. In accordance with previous studies we found that the DMFS score is inversely correlated to salivary flow rates and especially the unstimulated whole saliva flow rate [ 8 , 10 ]. However, it has not yet been possible to identify a cut-off value for salivary flow rate that can predict the risk of developing dental caries, as caries is a multifactorial disease. In this study, 90% of the patients had UWS flow rates below 0.10 ml/min, but as this value is merely a part of the classification criteria for pSS, it does not necessarily reflect the point at which caries is likely to develop. It has been suggested that salivary flow rates up to 0.16 ml/min, which include most pSS patients, may result in oral candidiasis [ 31 ] and increased development of experimental caries [ 29 ]. Meanwhile, the onset of increased caries activity in pSS remains unclear. It has been stated that in patients with SS, loss of teeth due to caries precede the first symptom of xerostomia by on average 9 years [ 10 ]. Changes in salivary flow and composition, and subsequently development of caries, appear to precede the symptom of oral dryness by several years. Increased caries activity that cannot be explained by changes in habits related to oral hygiene or diet may therefore represent a useful clinical feature to suspect early pSS in women without complaints of oral dryness, or intake of xerogenic medications. In contrast to the uncertainty pertaining to the onset of pSS and the initial consequences in terms of increased caries lesions, the future perspectives on the patients' dental health seem gloomy without professional preventive intervention. In accordance with the observation that past caries experience is one of the best predictors for future caries [ 32 , 33 ], the pSS patients in this study have a much higher risk of developing future caries and ultimately to loose teeth than the control group. Accordingly, these patients should receive an individual dental care programme in terms of oral hygiene instructions, professional oral hygiene regimens, fluoride treatment, dietary supervision and frequent dental follow-up visits in order to prevent accelerated caries development. Self-assessment of oral dryness In healthy subjects, the sensation of oral dryness usually occurs when whole saliva flow rate is reduced with more than 50% [ 34 ]. In this study, all patients had complaints of dry mouth and also substantially decreased salivary flow rates, whereas the healthy controls had no dry mouth complaints and in general, salivary flow rates within the normal range [ 35 ]. Nonetheless, in pSS, the scores of oral dryness were only inversely correlated to SWS, which may reflect a larger span of SWS values than of UWS. We have previously found an inverse correlation between scores of oral dryness and UWS [ 8 ]. In the present study, the patients had symptoms of oral dryness median 10 years prior to the diagnosis of pSS. Symptoms of oral dryness in combination with the participants' age could explain about half of the variance in DMFS (Table 5 ). The questionnaire may therefore not only be helpful in assessing the intensity of oral dryness but also in identifying patients with high DMFS. Saliva pH and buffer capacity The ability of human saliva to buffer acids is essential for maintaining pH values in the oral environment above the critical pH for hydroxyapatite (HAP), thereby protecting the teeth against demineralisation. The buffer systems responsible for the human saliva buffer capacity include the bicarbonate, phosphate and protein systems [ 17 , 36 - 38 ]. In normalcy, where the pH ranges from 6.0 to 7.5, the bicarbonate and phosphate buffer systems are by far the dominant ones having optimal buffering capacity at their pK values of 6.1 and 6.8, respectively [ 16 , 19 ], whereas the proteins have some effect on the buffer capacity at acidic pH values below 5 [ 16 , 38 ]. In this study, the parotid saliva buffer capacity was calculated individually for each pSS patient and healthy control based on the saliva bicarbonate and phosphate concentrations at the respective saliva pH values. This calculation gives an estimate of the buffer capacity of the saliva at the time it is secreted from the parotid gland. In the pSS patients, the buffer capacity of UPS was significantly lower than in the healthy controls, mainly due to the low pH value and bicarbonate concentration in saliva caused by the low flow rates. Moreover, in pSS most of the pH values of UPS were below the relevant pK values of both the bicarbonate and phosphate buffer systems. As compared to healthy controls, the pSS patients will therefore experience far more abundant pH drops in their saliva if exposed to acidic challenges leading to a higher risk of tooth demineralisation. Apart from a low buffer capacity, and thereby impaired ability to maintain a non-acidic saliva pH, the pSS patients also had a significantly lower degree of saturation with respect to HAP in their saliva (Table 4 ). Thus, in pSS, the mean pH of UPS was only one tenth of a pH unit above their mean critical pH. This implies that even a minor drop in pH will lead to undersaturation of their saliva with respect to HAP and result in either caries lesions or erosive damage to the teeth depending on the origin of the acidic challenge. Salivary bicarbonate and phosphate Despite the differences between pSS patients and healthy controls in sodium and chloride concentrations at comparable low flow rates, stimulated or not, the HCO 3 - concentration and saliva pH did not differ between the two groups (Fig. 1A–D ). The transport of bicarbonate in the salivary glands is believed to occur via chloride/bicarbonate exchange mechanisms [ 39 ]. The concentration of bicarbonate in saliva is a consequence of the metabolic CO 2 -turnover in the salivary glands. CO 2 freely diffuses across the epithelial boundaries, and due to the presence of carbonic anhydrase, the partial pressure for CO 2 and pH in the glandular compartments governs how much of bicarbonate buffer system is present in form of HCO 3 - in the saliva. The duct epithelium has dual functions with respect to bicarbonate transport, since it can both reabsorb (at low secretion rates) and secrete (at high secretion rates due to increased metabolic turnover of the gland). This study demonstrated a tendency towards lower total phosphate concentrations in pSS patients than in healthy controls. Previous studies have found significantly reduced phosphate concentrations in stimulated parotid and SM/SL saliva of patients with SS compared to healthy controls and patients with conditions resembling SS [ 40 - 42 ]. It should be stressed that the mechanism behind phosphate transport in human salivary gland tissues has not yet been fully characterised. It probably includes an acinar secretion and/or a ductal reabsorption via sodium-phosphate co-transport mechanism as that observed in the renal proximal tubules [ 43 ]. Salivary sodium and chloride Our sialochemical results are interpreted within the frame of the classical two-stage model of saliva formation [ 44 ]. Under normal physiological conditions, and in response to nervous stimuli, the acinar cells produce primary saliva, which has an ionic composition resembling that of plasma. As the primary saliva passes through the duct system it becomes modified by reabsorption of sodium and chloride (but without water due to the low water permeability) whereby the final saliva secreted into the oral cavity becomes hypotonic with sodium and chloride concentrations much below that of the original primary saliva. The composition of the final saliva secreted into the oral cavity strongly depends on the secretion rate in such a way that at low flow rates the saliva contains low sodium and chloride concentrations and as the flow rates increase the concentrations of sodium and chloride will rise. This normal physiological relation between parotid flow rate and sodium and chloride concentrations was seen in the healthy controls and in some of the pSS patients (0.01–1.00 ml/min) as well (Fig. 1A and 1B ). However, for other pSS patients with flow rates within the same frame, another picture emerged, since their sodium and chloride concentrations were remarkably higher than normally whether stimulated or not. Thus, on a group basis the pSS patients have significantly higher concentrations of sodium and chloride than the healthy controls (Table 4 ). This finding is in accordance with several previous studies on whole saliva, parotid and submandibular/sublingual saliva (SM/SL) [ 5 , 40 - 42 , 45 , 46 ]. The concentrations of sodium and chloride have also been shown to be higher in SM/SL of patients with pSS and secondary SS compared to patients with clinical conditions resembling SS, i.e., sialoadenosis, sodium retention dysfunction syndrome and medication-induced xerostomia [ 42 ]. Overall, these compositional changes appear to be unique for some pSS patients. It has been stated that SM/SL glands are affected earlier by SS than the parotid glands due to an average reduction of stimulated SM/SL flow rate preceding that of the stimulated parotid flow rate [ 40 , 42 ]. On the other hand, cut-off values for sodium, chloride and phosphate in SPS and stimulated SM/SL being predictive for SS demonstrated almost similar specificity (69 and 71%, respectively) and sensitivity (81%) [ 32 ]. Other salivary constituents Despite the low salivary flow rates seen in the pSS, the acinar transport mechanisms involved in the formation of primary saliva seem to be unaffected by the glandular lymphocytic infiltration. Accordingly, concentrations of potassium, total calcium, total protein and amylase activity in whole and parotid saliva did not differ from those of the healthy controls, which is in agreement with previous reports [ 40 , 45 ]. Furthermore, it has been shown that the output of statherin and acidic proline-rich proteins, which reflect the secretion of selected parotid proteins, did not differ between pSS patients and healthy controls [ 5 ]. Normal concentrations of total calcium, total protein and levels of amylase activity indicate that the remaining functional acinar cells are capable of synthesis and secretion of primary saliva with normal composition despite the marked lymphocytic infiltration and structural changes. Impaired in ductal salt reabsorption reflects disease severity in pSS The changes in salivary composition indicate that the duct epithelium, and mainly the striated duct epithelium, cannot effectively reabsorb the high concentrations of sodium and chloride of the primary saliva in some of the pSS patients, despite low salivary flow rates. The pSS patients who exhibited high concentrations of sodium and chloride were also characterised by having the lowest flow rates, the highest focus score (FS) and highest concentrations of serum autoantibodies (AB) in addition to a tendency of being younger than the pSS patients with normal salivary concentrations of sodium and chloride. Overall, the subgroup of pSS patients with the highest concentrations of salivary sodium and chloride and lowest total phosphate concentrations appeared to be more severely affected by pSS having more exocrine and non-exocrine disease manifestations than those without these salivary changes supporting previous observations [ 5 ]. It should be stressed that there was no relationship between reduced salivary flow and compositional changes and intake of medications. The question is whether the subgroup of patients with normal salivary concentrations of sodium and chloride and some preserved salivary gland function has a late onset of pSS or a "milder" glandular response to autoimmunity compared to the subgroup with high salivary sodium and chloride concentrations or salivary secretion close to or equal to 0 ml/min. Another relevant question is whether the latter subgroup of patients also has a greater risk of developing malignant lymphoma. The results of our study therefore need to be tested in a large prospective cohort study and compared with group of patients with non-immunological destruction of their salivary glands in order to validate the use of specific salivary changes and dental caries as markers of salivary gland dysfunctional severity/disease severity in pSS. At present there is no international expert consensus regarding measures for assessment of disease activity, severity, damage or outcome in pSS that can be used in the evaluation of clinical trials of new therapies and longitudinal observational studies. Recent reports, however, indicate that the development and evaluation of such measures has begun [ 47 , 48 ]. In pSS, the salivary gland histopathology is characterised by periductal lymphocytic infiltration and acinar destruction. The duct epithelium, however, appears relatively unaffected by the lymphocytic infiltration, which contrasts the observed salivary changes in pSS. On the other hand, the discrepancy may be explained by morphological differences between the labial salivary glands, which predominantly consist of mucous acini and have a very short duct system, and the parotid glands, which comprise serous acini and have long duct system. Nevertheless, it has been stated that the histopathological changes of the labial salivary glands mimic those of the parotid glands [ 49 , 50 ]. It has been suggested that Bcl-2 positive basal cells of striated/excretory ducts possess an extensive capacity for pluridirectional morphogenetic differentiation [ 51 ]. On this basis, it could be speculated that the duct cells in some pSS patients possess antigenic properties, which initiate an autoimmune response that could be associated with morphogenesis and cell differentiation of the salivary gland tissues. It has not yet been possible to identify a specific anti-salivary duct antibody in pSS that is capable of inhibiting cell differentiation. Our finding of more pronounced acinar and ductal tissue functional impairment in pSS patients with both FS and AB, than in patients with FS or AB, supports the idea that circulating autoantibodies or inflammatory mediators produced locally by the inflammatory cells, interfere with the neural release of neurotransmitter substances or interact with the binding of neurotransmitters to receptors on the cell surface, thereby impairing the acinar secretion and/or the ductal reabsorptive modification of saliva [ 2 ]. Along this line, an in vitro study on isolated human acini and duct segments from pSS patients have shown that these cells possess functional receptor systems and normal response in changes in the intracellular free calcium concentration upon maximal secretagogue stimulation [ 52 ]. The fact that the ductal uptake of sodium via amiloride-sensitive epithelial sodium channel (ENaC) is regulated by circulating adrenal mineralocorticoids, e.g. aldosterone [ 53 ] could also indicate that some pSS patients have low levels of aldosterone. However, this still needs to be clarified. Conclusion The results of this study indicate that specific sialometric and sialochemical glandular changes, particularly changes in sodium and chloride concentrations, may serve as oral markers of disease severity in pSS. The question arises whether pSS represents a continuum of patients with different stages of disease/affection of salivary glands or different diseases affecting the salivary gland tissues. There is great need of disease assessment and outcome markers in pSS. The hypotheses generated from our results on changes in salivary flow and composition as well as high caries experience as potential markers of the extent of autoimmune-mediated salivary gland dysfunction in pSS therefore need to be tested in a large prospective cohort study including patients with early to long-standing disease. In addition, the mechanisms underlying the impaired ductal salt reabsorption observed in the pSS patients with presence of both labial salivary gland focus score and serum-autoantibodies need to be further elucidated in future pathophysiological studies. Abbreviations SS = Sjögren's syndrome; pSS = primary Sjögren's syndrome; FS = focal lymphocytic infiltration in the labial salivary glands; UWS = unstimulated whole saliva flow rate; SWS = stimulated whole saliva flow rate; UPS = unstimulated parotid saliva flow rate; SPS = stimulated parotid saliva flow rate; AB = presence of serum autoantibodies to Ro (SSA) and/or La (SSB) antigens; DMFS = decayed, missing and filled tooth surfaces; PI and GI = plaque and gingival indices; PPD = periodontal probing pocket depth; HAP = hydroxyapatite. Competing interests The author(s) declare that they have no competing interests. Authors' contributions All three authors participated in the design of the study and the statistical analysis. AMLP attended coordination of the study and conducted the whole saliva collections, interviews, oral clinical examinations and labial salivary gland biopsies. AB performed the parotid saliva collections and sialochemical analyses. All authors read and approved the final manuscript. Pre-publication history The pre-publication history for this paper can be accessed here:	/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC554998.xml
509308	Finding Mutations That Disrupt Cortical Development	null	As the presumed “seat of consciousness,” the cerebral cortex mediates the higher-level cognitive processing—such as abstract thought—that humans like to think distinguishes them from other animals. The cerebral cortex is, in fact, significantly larger in the human brain and has far more “columns” than it does in other mammals, particularly compared to the rat, a traditional model for brain study. Neurons in these cortical columns have similar response properties and form fundamental units of brain processing. It is the larger brain surface area, accommodating a greater number of cortical columns, that gives humans the computational edge. By studying the genetic and molecular agents of cortex development, scientists hope to understand the nature and extent of cortical cognitive function and identify effective therapies to repair brain injury and disease. Analyses of mutant mice have provided insights into the mechanisms controlling cerebral cortex development, but many details of cortical development remain to be revealed. Identifying individual molecules and genes involved in discrete brain processes is particularly difficult given the complexity of brain structure and function. Geneticists have traditionally linked genes to specific biological pathways by first screening large numbers of individuals of a species for unusual physical traits (phenotypes) and then determining the genetic makeup of these mutants to home in on the faulty gene. This approach, called forward genetics, typically requires large numbers of individuals to find unusual phenotypes and so has traditionally focused on fast-breeding organisms like zebrafish and fruitflies. But zebrafish and fruitflies are unlikely to reveal the secrets of higher consciousness. Now Andrew Peterson and colleagues have updated the forward genetic screen and added a new resource to the neuroscientist's “brain dissection” toolkit. Their approach, which labels specific populations of neurons with protein “reporters,” offers a novel way to find mutations in developing neurons and to identify mutations that interfere with cerebral cortex development. The reporter used here highlights mutations that disrupt interneuron migration into the cortex as well as those that affect cortex growth and morphology. (Interneurons are one of the primary types of cortical neurons.) Distribution of GABAergic interneurons in wild-type (top) and mutant (bottom) cortex Like most forward genetic approaches, the researchers started with a genetically well-characterized breed, then used a chemical mutagen to damage the organism's DNA. After two or three rounds of breeding, the researchers looked for cortical-related defects in the developing embryos. In this case, however, the region of the mouse brain that gives rise to developing interneurons was labeled in the original mice. Thus, when the researchers screened for mutants with defects associated with forebrain development and interneuron migration, they could easily find the cells and genes involved. The screen identified thirteen mutations affecting cortical development and interneuron migration. (Developing neurons travel along the fibers of other brain cells before reaching their ultimate destination in the brain.) Three mutations are variants of genes known to play a role in cortical development; nine mutations were in genes that had not been linked to cortical development before. The screen described here takes advantage of a chemical mutagen (called ENU) that induces mutations at a single base pair, or nucleotide, in DNA. These mutations tend to have quite selective effects on protein function—by changing the composition of a single domain—which can provide information on how the protein should normally function and highlight its role in a particular process. The selective nature of ENU, the authors argue, offers new information about how the mutated genes identified here function in cortical development and how these putative roles might be tested. Altogether, the results suggest that this type of focused screening, so long a resource in fly genetics, can be a powerful tool in mammalian biology as well. That the strategy outlined here could identify novel mutations in a process as complicated as cerebral cortex development suggests that it could do the same for a broad range of biological processes. If the success of other model systems moving in this direction is any indication, this new strategy in the mouse offers researchers a powerful resource for identifying the genetic underpinnings of living systems.	/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC509308.xml
522826	Validity and reproducibility of an interviewer-administered food frequency questionnaire for healthy French-Canadian men and women	Objective To evaluate the validity (study 1) and the reproducibility (study 2) of an interviewer-administered food frequency questionnaire (FFQ). Method The FFQ was designed at Laval University and contains 91 items and 33 subquestions. Study 1 : The FFQ was compared against a 3-day food record (2 week-days and 1 weekend-day), at week 0, 6 and 12 of a nutritional intervention. Study 2 : In order to evaluate the reproducibility of the FFQ, 2 registered dietitians administered the FFQ 4-weeks apart among subjects who were not part of the nutritional intervention. Results Study 1 : Mean values for intake of most nutrients assessed by the FFQ and by the 3-day food record were not statistically different. Energy-adjusted correlation coefficients for major macronutrients ranged from 0.36 for proteins to 0.60 for carbohydrates (p ≤ 0.01). Agreement analysis revealed that on average, 35% of the subjects were classified in the same quartile when nutrients were assessed by either the 3-day food record or the FFQ. Study 2 : Significant associations were observed between dietary measurements derived from the two FFQs administered 4 weeks apart. Correlation coefficients for the reproducibility of macronutrients ranged from 0.66 for carbohydrates to 0.83 for lipids after energy adjustment. On average, 46% of the subjects were classified in the same quartile when nutrient intakes were assessed by either FFQ. Conclusion These data indicated that the FFQ developed has a good validity and is reproducible.	Background There is increasing evidence that nutrients may be important in the development of chronic diseases such as coronary heart disease (CHD) and type 2 diabetes. In the late 60s, the Mediterranean diet became a topic of interest primarily because of results of the Seven Countries Study, which demonstrated that the 15-y mortality rate from CHD in Southern Europe, was two to threefold lower than in Northern Europe or United States [ 1 ]. More recently, results from The Lyon Diet Heart Study showed that a Mediterranean alpha-linolenic acid-rich diet prevented the recurrence of cardiovascular events more than did the usual prudent Western diet in men [ 2 - 4 ]. Reliable instruments for diet measurements are necessary to identify which components of the Mediterranean diet are the best candidates to explain, in part, such a protective effect. Accurate assessment of dietary intakes, when based on self-report in free-living populations poses significant scientific challenges. All standard dietary assessment methods including food records, dietary recalls and list-type methods such as food frequency questionnaires (FFQ), are subjected to considerable error and bias, and none of these can be considered as a 'gold standard' measure [ 5 ]. FFQ has become a common way to estimate usual food intake because it usually requires less than thirty minutes to complete [ 6 ]. It also imposes less burden on subjects than most of the other dietary assessment methods. However, disadvantages of the FFQ have been identified. In fact, it may be difficult cognitive task for respondent to recall frequencies of intakes over a given period of time. Also, the precision in quantifying intakes is not possible with a FFQ. Dietary habits vary not only from country to country but also from region to region. Specific FFQs must be validated to assess nutritional habits conducted in geographically and/or culturally distinct regions [ 6 ]. It is also important in nutritional intervention to consider the sensitivity of the method over the duration of a study, especially in study that is testing the effects of dietary changes [ 7 ]. In Québec, no validated FFQ was adapted for the needs of our nutritional intervention design. In fact, in the context of our nutritional intervention we wanted to use a FFQ to evaluate a Mediterranean food pattern in a North-American context that would contain foods available in Québec and also foods characteristic of the Mediterranean diet. In order to improve the precision in quantifying reported intakes we decided to use and interviewer-administered FFQ to facilitate the determination of portion size using food models. Thus, we decided to design an interviewer-administered FFQ to assess the dietary changes among a French-Canadian population in a nutritional intervention promoting the Mediterranean food pattern. The first purpose of the present study was therefore to test the validity of this interviewer-administered FFQ. In order to reach this objective, nutrient intakes derived from the FFQ were compared to intakes obtained by the 3-day food record. Comparisons were performed for baseline as well as for post-intervention values. This allowed us to determine whether our FFQ would permit to identify similar changes in nutrient intake in response to the intervention as the ones measured by the 3-day food record. As a second objective, we also wanted to estimate the short-term reproducibility of this FFQ in a control group who did not receive the nutritional intervention. Subjects and methods Subjects This paper reports results of two studies: 1) a study on validity of the FFQ tested against a 3-day food record during a nutritional intervention program promoting the Mediterranean food pattern and 2) a study on reproducibility of the FFQ in a control population. For the validation study, women from the Québec City metropolitan area were recruited through the Laval University newspaper during the summer of 2001. Women included in the study were aged between 30 to 65 years [ 8 ]. To be eligible, women had to be free from metabolic disorders requiring treatment, to have stable body weight for at least 3 months prior to the start of the study and to be in charge of food purchases and meal preparation most of the time. One hundred and twenty six women were invited to a screening visit for an evaluation of their food habits. Among this initial group of women, 94 were found to be eligible according to the above criteria and 77 women agreed to take part to the study. Three women left the study during the 12-week intervention for personal reasons. One participant did not complete the FFQ at week 12 and 2 other participants did not complete all three food records. Therefore, 71 women were included for the FFQ validation analysis. For the reproducibility study, 20 men and 19 women from the Québec City metropolitan area were also recruited through the Laval University newspaper during the summer of 2002. Men and women included in the study were aged between 25 to 70 years. Three men and 4 women did not complete the second FFQ for personal reasons. Therefore, 17 men and 15 women were included for the reproducibility analyses. The study was approved by the Ethics Committees of Laval University. Food frequency questionnaire The interviewer-administered FFQ developed inquired on the food habits during the last month and is based on typical food items, which are available in Québec. It contains 91 items among which 27 had between 1 and 3 subquestions. The FFQ was structured to reflect Quebecers' food habits and food items were listed in food groups (vegetables; fruits; legumes, nuts and seeds; cereals and grain products; milk and dairy products; meat/processed meat; poultry; fish; eggs; sweets; oils and fats; fast foods and drinks). Because of the nature of our nutritional intervention, our FFQ was designed to make sure to document with enough details consumption of typical items of the Mediterranean diet such as type of oils, whole grain products and legumes. The 30-minutes FFQ was administered face-to-face by one of the 3 registered dietitians involved in the study. During the interview, the dietitians used food models for a better estimation of the real portion consumed by the subject. Participants were questioned about frequency of intake for different foods during the last month and were asked to report the frequency of these intakes in terms of day, week or month. The subquestions allowed a better definition of the food items consumed. For example, following the question "How often do you eat yogurt?" subjects were asked about the fat percentage and brand of the yogurt consumed. An open question at the end of the FFQ allowed individuals to report any other frequency eaten foods not listed in the FFQ and provide details about usual recipes used in order to quantify better intakes of individual food items. Cut-off value to evaluate energy intake Estimates of basal metabolic rate (BMR) were calculated from the Harris-Benedict formulas based on height, weight, age and sex [ 9 ]. Energy intake reported from FFQs and from 3-day food records were compared with estimates of BMR to calculate the number of participants who may have underreported their energy intake [ 10 ]. It is suggested that a ratio between energy intake and estimate BMR of less than 1.35 might not represent long term habitual intake in a non dieting population [ 10 ]. Study 1: Validation Intervention and FFQ The methodology of the nutritional intervention has been described previously [ 8 ]. Briefly, the study was conducted in 2 phases. Each phase was conducted using a similar 12-week intervention design. The FFQ was administered at screening (t = 0) and then at weeks 6 and 12. The intervention included 2 group sessions. Individual sessions took place during the 1 st , the 6 th and 12 th weeks of the intervention in order to evaluate the changes and to select further objectives for increasing the adherence to the Mediterranean food pattern. Three registered dietitians were trained to provide a standardized intervention. 3-day food record Each participant completed a 3-day food record, 2-week days and 1-weekend day, at week 0, 6 and 12. At screening of the nutritional intervention, a dietitian provided 15 minutes of instructions to each participant on how to complete the food records. Written copies of record examples were provided to each subject. Also, participants were encouraged to consume usual amounts of typical foods and drinks for the completion of the food record. Participants were not required to weight foods but were asked to measure the volume of foods consumed with household measurements (cups, tablespoons) or to indicate the weight of commercial products when it was possible to assess portion sizes. The food record included a section for recording information recipes. After completing the food record, participants met with the dietitian to review all the information for record accuracy and completeness and portion size of individual items on the food record were reviewed when needed by using food models. Anthropometry At weeks 0, 6 and 12, body weight and height were measured according to the procedures recommended at the Airlie Conference on the Standardization of anthropometric measurements [ 11 ] and body mass index (BMI) was calculated. Study 2: Reproducibility Study design The 32 participants of the reproducibility study were distributed into two groups. In the first group, dietitian #1 administered the FFQ and 4 weeks later dietitian #2 administered the FFQ for the second time. Inversely, in the second group of participants, dietitian #2 administered the first FFQ and 4 weeks later dietitian #1 administered the FFQ for the second time. An interval of one month was chosen to reduce any training effect and memory influence of the method. Both dietitians were taught to use the FFQ similarly, using the same examples of portion size, and asking similar questions. Anthropometry At the first visit, body weight and height were measured according to standard procedures [ 11 ] and BMI was calculated. Analysis Nutritional analysis Evaluation of nutrient intakes derived from the FFQs and food records was performed using the Nutrition Data System for Research (NDS-R) software version 4.03, developed by the Nutrition Coordination Center, University of Minnesota, Minneapolis, MN, Food and Nutrient Database 31, released in November 2000 [ 12 ]. This database includes more than 16 000 food items for which the complete nutritional value of 112 nutrients is included. For the purpose of our study, intakes of selected nutrients susceptible to affect the CHD risk profile were analysed: energy, proteins, carbohydrates (CHO), lipids, saturated fatty acids (SFA), monounsaturated fatty acids (MUFA), polyunsaturated fatty acids (PUFA), trans fatty acids, eicosapentaenoic acid (EPA), docosahexaenoic acid (DHA), total as well as insoluble and soluble dietary fibers, vitamin C, folate, iron and calcium. Intakes of vitamin and mineral supplements were not included in the present analysis, which focused on dietary nutrients only. Statistical analysis In the validation study, means and standard deviations for nutrient intakes were calculated from the FFQs and from the food records. Student t-tests were performed to determine the differences between nutrient intakes assessed by the FFQ and by the 3-day dietary record. Since many variables were not normally distributed, Spearman correlations were used to put into relationship nutrient intakes from FFQ with those from the food record. Student t-tests were also performed to determine the differences between changes in nutrient intakes in response to the nutritional intervention assessed by the FFQ and by the 3-day dietary record. Agreement analyses were performed to verify the concordance of different nutritional variables among quartiles of the distribution between FFQ and the 3-day food record. In the reproducibility study, Student t test was performed to assess the differences between both FFQs. Spearman correlations were used to put into relationship the nutrients reported in the first and the second FFQ. Agreement analyses were assessed to verify the degree of concordance in classifying subjects among quartiles of the distribution between both FFQs. For both studies, the dietary variables were log transformed when necessary to achieve a normal distribution, and the formula log(x + 1) was used for alcohol because some subjects had a value of 0 g for alcohol intake. In addition, to make the comparisons based on absolute nutrient intakes, correlations were also made using energy-adjusted variables. Adjustment for total energy intake was achieved by using the residual method proposed by Willett and Stampfer [ 13 ]. Residuals are computed from regression models, with total energy intake as the independent variable and nutrient intakes as the dependent variable. Values were considered as being very well correlated for correlations ranging between 0.7 to 0.9, well correlated for correlations ranging between 0.5 to 0.7 and moderately well correlated for correlations between 0.3 to 0.5, as suggested by Rimm et al [ 14 ]. Because age and BMI may have influenced the manner in which subjects answered the FFQ or completed the 3-day food record, partial correlations for age and BMI were also computed. Also, Student t-tests were performed to verify variation in nutrients intakes obtained from the two FFQ administered by the two interviewers. All analyses were performed with the SAS statistical package version 8.02 (SAS Institute, Cary, N.C., USA). Results Study 1: Validation Women had a mean BMI of 25.8 ± 3.9 kg/m 2 and a mean age of 46.8 ± 7.9 y. Average daily nutrient intakes derived from FFQ and the 3-day food records are shown in Table 1 . Total energy intake measured by the FFQ was not different from the intake assessed by the 3-day food record (difference below 5%) at week 0 of the dietary intervention (Table 1 ). For FFQ and the 3-day food record respectively, 34% (n = 24) and 38% (n = 27) of subjects had at baseline a ratio between energy intake and estimated basal metabolic rate at or below 1.35. Also, the intakes of proteins, CHO, SFA, PUFA, trans fatty acids, cholesterol, alcohol, vitamin C, folate, calcium and iron measured by the FFQ and by the 3-day food record at week 0 did not differ significantly. Measurement of total dietary fibers and soluble fibers differed significantly between the FFQ and the food-record but differences were below 10%. Mean nutrient intakes measured with FFQ were within 10% of values obtained with the 3-day food record for 14 of the 19 nutrients measured at week 0. For total lipids and MUFA the differences were significant but below 15%. Adjusting for energy intake did not alter these observations. Similar observations were noted at week 12 except for intakes of lipids, total dietary fibers and soluble fibers that were not anymore significantly different between the two methods and for PUFA intake that was estimated as being significantly higher with the FFQ (data not shown). Table 1 Mean values of daily intakes of nutrients and Spearman correlation coefficients between values derived from the 3-day food record and the FFQ at week 0 of the nutritional intervention (n = 71). Dietary record FFQ Difference (%) a Unadjusted Energy-adjusted Energy (kcal) b 2055 ± 521 2143 ± 568 4.3 0.29** Protein (g) 81.3 ± 16.4 82.6 ± 23.3 1.6 0.27* 0.36 CHO (g) 245.0 ± 58.9 242.2 ± 60.5 -1.1 0.40 0.60*** Lipids (g) b 80.1 ± 34.4 90.0 ± 34.8* 12.4 0.29* 0.56* SFA (g) b 27.1 ± 14.0 30.0 ± 12.7 10.7 0.30 0.56* MUFA (g) b 33.9 ± 14.3 39.0 ± 16.3 15.0 0.26* 0.48* PUFA (g) b 13.1 ± 6.0 14.4 ± 7.0 9.9 0.38 0.46* EPA (g) b 0.06 ± 0.07 0.05 ± 0.04 -16.7 0.33 0.33 DHA (g) b 0.17 ± 0.23 0.12 ± 0.08 -29.4 0.30 0.30 Trans fat (g) b 3.4 ± 2.0 3.6 ± 2.6 5.9 0.45* 0.56* Cholesterol (mg) b 280.1 ± 139.8 300 ± 116 7.5 0.30 0.36** Dietary fiber (g) 21.9 ± 6.4 19.7 ± 5.0* -10.0 0.32 0.38 Soluble fiber (g) 7.3 ± 2.0 6.6 ± 1.7** -9.6 0.22 0.27* Insoluble fiber (g) 14.4 ± 4.7 13.2 ± 3.5 -8.3 0.32 0.37 Alchool (g) b 11.4 ± 11.2 10.9 ± 10.7 -4.4 0.61* 0.66* Vitamin C (mg) 138.5 ± 57.3 137.0 ± 62.2 -1.1 0.19 0.19 Folate (mcg) 394.6 ± 115.4 383.2 ± 107.1 -2.9 0.32 0.39 Calcium (mg) 898.4 ± 320.5 962.3 ± 399.8 7.1 0.49* 0.56* Iron (mg) b 15.1 ± 4.3 14.1 ± 4.1 -6.6 0.43 0.53* Value are means ± SD a (value derived from FFQ-value derived from 3-day food record)/(value derived from 3-day food record) × 100 b Analyses were performed on log transformed values Significant difference between the two methods p ≤ 0.05; p ≤ 0.01; *p ≤ 0.0001 Correlations between FFQ and 3-day food record were statistically significant at p ≤ 0.05; p ≤ 0.01; * p ≤ 0.0001 Spearman correlations between values of nutrient intakes measured by the 3-day food record and those assessed by the FFQ at week 0 of the nutritional intervention are shown in Table 1 . Analyses were performed on unadjusted as well as on energy-adjusted values. The average correlation coefficient for the nutrients presented in Table 1 was 0.44 at week 0 (energy-adjusted). Values derived from the FFQ and the 3-day food record were well correlated (energy-adjusted) for CHO, lipids, SFA, trans fatty acids, alcohol, calcium and iron (0.5 < r < 0.7) and moderately well correlated for proteins, MUFA, PUFA, cholesterol, dietary fibers, insoluble fibers, EPA and DHA (0.3 < r < 0.5). Further adjustment for age and BMI did not materially modify these correlations. Correlations between FFQ and dietary food record at week 12 were slightly higher than correlations at week 0 (data not shown). Table 2 presents changes observed in response to the 12-week nutritional intervention derived from the 3-day food record and from the FFQ. For the majority of the nutrients analyses changes observed by the 3-day food record were similar to those observed by the FFQ except for vitamin C intake for which a higher increase was noted when dietary changes were assessed by the FFQ. Table 2 Changes in daily energy and selected nutrients intakes derived from the 3-day food record and from the FFQ in response to the 12-week nutritional intervention (n = 71). Dietary record FFQ Energy (kcal) -197 ± 464 -200 ± 500 Protein (g) -1.4 ± 15.8 -0.5 ± 18.8 CHO (g) -13.9 ± 60.8 -7.4 ± 59.2 Lipids (g) -12.3 ± 32.5 -16.1 ± 28.3 SFA (g) -6.2 ± 13.1 -9.6 ± 9.8 MUFA (g) -4.2 ± 14.7 -4.7 ± 15.3 PUFA (g) -1.1 ± 6.2 -0.8 ± 6.2 EPA (g) 0.06 ± 0.13 0.06 ± 0.06 DHA (g) 0.10 ± 0.40 0.11 ± 0.11 Trans fat (g) -1.4 ± 2.1 -1.6 ± 2.4 Cholesterol (mg) -51.9 ± 142.6 -65.8 ± 92.2 Dietary fiber (g) 3.0 ± 8.0 4.7 ± 6.9 Soluble fiber (g) 0.7 ± 2.7 1.2 ± 2.3 Insoluble fiber (g) 2.3 ± 5.7 3.5 ± 4.8 Alcohol (g) -2.8 ± 10.4 -1.9 ± 9.4 Vitamin C (mg) 3.6 ± 65.9 29.4 ± 62.9* Folate (mcg) -7.8 ± 131.9 18.1 ± 113.8 Calcium (mg) 3.8 ± 304.1 33.6 ± 331.9 Iron (mg) -0.3 ± 5.3 1.0 ± 4.3 Value are means ± SD Significant difference between the two methods p ≤ 0.05 Agreement between quartile classification of FFQ and 3-day food record is show in Table 3 . Percentage of agreement varied from 29.4% for proteins to 64.7% for trans fatty acids for the lowest intakes (1 st tertile) and from 27.8% for CHO to 55.6% for alcohol for the highest intakes (4 th tertile). When considering all nutrients studied, it was found that, on average, 35.1% of subjects were categorized exactly in the same quartile by the FFQ and by the 3-day food record and 5.1% of the subjects were misclassified in extreme quartiles i.e. subject in the first quartile according to one method and in the fourth quartile according to the other (not shown). Table 3 Percentage of agreement for the classification into quartiles of the distribution of selected dietary variables using either the 3-day food record or the FFQ at week 0. Lowest quartile with 3-day food record and FFQ (%) Highest quartile with 3-day food record and FFQ (%) Exact agreement across quartiles (%) Energy (kcal) 47.1 38.9 39.4 Protein (g) 29.4 38.9 28.2 CHO (g) 41.2 27.8 26.8 Lipids (g) 35.3 38.9 33.8 SFA (g) 41.2 33.3 29.6 MUFA (g) 35.3 36.8 33.8 PUFA (g) 47.1 38.9 35.2 Trans fat (g) 64.7 38.9 47.9 Cholesterol (mg) 41.2 44.4 36.6 Dietary fiber (g) 47.1 44.4 31.0 Alcohol (g) 52.9 55.6 43.7 Study 2: Reproducibility The 32 subjects in study 2 had a mean BMI of 23.9 ± 3.6 kg/m 2 for women and 27.6 ± 5.1 kg/m 2 for men. The mean age was 42.5 ± 10.4 y for women and 41.2 ± 11.9 y for men. Table 4 shows average daily nutrient intakes derived from the two FFQs (FFQ1, FFQ2) administered 4 weeks apart by two different dietitians. Measurement of total energy intake was not different between the two FFQs (difference of 6.8%). Also, the intakes of proteins, CHO, lipids, SFA, MUFA PUFA, trans fatty acids, cholesterol, alcohol and micronutrients measured by FFQ1 and by FFQ2 did not differ significantly. Adjustment for total energy intake did not alter these observations. Similar results were observed when analyses were performed within each gender separately. Subjects in the first group (in which dietitian #1 administered the 1 st FFQ) showed similar variation in nutrients intakes between FFQ1 and FFQ2 than subjects from the 2 nd group (in which dietitian #2 administered the first FFQ). Table 4 Mean values of daily intakes of nutrients from two FFQ a and Spearman correlation coefficients between values derived from the two FFQs a . FFQ1 FFQ2 Difference (%) c Unadjusted d Energy-adjusted d Energy (kcal) 2283 ± 584 2128 ± 480 -6.8 0.73 Protein (g) b 89.4 ± 28.4 86.8 ± 24.4 -2.9 0.65* 0.83* CHO (g) 286.3 ± 88.2 262.2 ± 80.6 -8.4 0.79* 0.66*** Lipids (g) b 84.6 ± 25.1 78.9 ± 19.8 -6.7 0.47* 0.82*** SFA (g) b 28.8 ± 10.1 27.1 ± 8.6 -5.9 0.51* 0.81*** MUFA (g) b 34.9 ± 11.0 32.6 ± 8.4 -6.6 0.52* 0.82* PUFA (g) b 14.3 ± 6.2 13.1 ± 4.7 -8.4 0.60 0.74* EPA (g) b 0.06 ± 0.05 0.07 ± 0.06 16.7 0.55 0.55 DHA (g) b 0.12 ± 0.09 0.15 ± 0.12 25.0 0.55 0.56 Trans fat (g) 3.5 ± 1.7 3.2 ± 1.4 -8.6 0.60 0.79*** Cholesterol (mg) 278.2 ± 102.8 269.3 ± 75.1 -3.2 0.47* 0.73* Dietary fiber (g) 22.9 ± 7.9 20.8 ± 6.3 -9.2 0.76* 0.75* Soluble fiber (g) 7.8 ± 2.4 7.1 ± 2.0 -9.0 0.82* 0.87* Insoluble fiber (g) 15.0 ± 5.5 13.7 ± 4.4 -8.7 0.73* 0.81* Alchool (g) 9.6 ± 6.7 9.4 ± 8.0 -2.1 0.76* 0.70* Vitamin C (mg) 201.6 ± 103.5 165.2 ± 77.0 -18.1 0.81* 0.62* Folate (mcg) b 442.0 ± 147.4 394.5 ± 128.5 -10.7 0.71* 0.69* Calcium (mg) b 1153.5 ± 453.5 1085.1 ± 456.2 -5.9 0.79* 0.86* Iron (mg) b 15.8 ± 5.7 14.5 ± 4.7 -8.2 0.64* 0.79*** Values are means ± SD a 2 FFQ were administered 4 weeks apart b Student t test analyses were performed on log transformed values c (value derived from FFQ2 - value derived from FFQ1)/(value derived from FFQ1) × 100 d Correlations between FFQ1 and FFQ2 were statistically significant at * p ≤ 0.05; p ≤ 0.01; * p ≤ 0.0001 Table 4 presents Spearman correlations between nutrient intakes measured by FFQ1 and FFQ2. The average correlation coefficient for the nutrients presented in Table 4 is 0.74. Values derived from the two FFQs after adjustment for energy intake were generally very well correlated for proteins, lipids, SFA, MUFA, PUFA, trans fatty acids, cholesterol, dietary fibers, soluble and insoluble dietary fibers, alcohol, calcium and iron (0.7 < r < 0.9) and well correlated for CHO, EPA, DHA, vitamin C and folate (0.6 < r < 0.7). Partial correlations between nutrients derived from FFQ1 and those derived from FFQ2 were unchanged after adjusting for age, BMI and for the dietitian who administered the 1 st FFQ. In addition, similar correlations were observed when analyses were computed within each gender separately (not shown). Table 5 shows agreement between quartile classification when FFQ1 was compared to FFQ2. Percentage of agreement varied from 28.6% for proteins to 75.0% for energy and alcohol for the lowest intakes (1 st quartile) and from 37.5% for SFA and MUFA to 75.0% for energy, proteins, CHO and alcohol for the highest intakes (4 th quartile). When considering all the nutrients studied, it was found that on average, 45.7% of subjects were categorized exactly in the same quartile by both FFQs and 2.6% of the subjects were misclassified in extreme quartile (not shown). Table 5 Percentage of agreement for the classification into quartiles of the distribution of selected dietary variables using either FFQ1 or FFQ2. Lowest quartile with FFQ1 and FFQ2 (%) Highest quartile with FFQ1 and FFQ2 (%) Exact agreement across quartiles (%) Energy (kcal) 75.0 75.0 62.5 Protein (g) 28.6 75.0 37.5 CHO (g) 50.0 75.0 59.4 Lipids (g) 50.0 62.5 40.6 SFA (g) 50.0 37.5 34.4 MUFA (g) 62.5 37.5 40.6 PUFA (g) 50.0 62.5 46.7 Trans fat (g) 62.5 50.0 40.6 Cholesterol (mg) 37.5 50.0 40.6 Dietary fiber (g) 62.5 50.0 43.8 Alcohol (g) 75.0 75.0 56.3 Discussion Accurate assessment of dietary intakes and dietary changes plays a central role in nutritional studies. Each tool used to evaluate dietary intakes has some strengths and limitations. Also, all standard dietary assessment methods are subjected to bias such as underreporting [ 6 ]. In our study, we developed a 91-items interviewer-administered FFQ that was sufficiently accurate to measure intakes of nutrients in the habitual diet of subjects from the Québec City metropolitan area and changes in nutrient intakes following a 12 week intervention promoting the Mediterranean food pattern. In the validation study, coefficients of correlation between values derived from the FFQ and those obtained by the 3-day food record ranged from 0.30 to 0.60 for macronutrients and from 0.19 to 0.56 for micronutrients at week 0. It has been previously reported that correlation coefficients for validation studies ranged from 0.4 to 0.7, similar to our results after energy adjustment [ 6 , 15 ]. Also, our interviewer-administered FFQ did not significantly overestimate energy intake compared to a 3-day food record. The fact that the interviewer used food models to facilitate the estimation of portion size can contribute to explain this findings. It has been shown that FFQ can both under- and overestimate intakes of specific nutrients. In fact, many validation studies have reported that FFQ, as compared to food-record or 24-hour recall overestimate nutrients intakes as well as energy intake [ 16 - 20 ]. In contrast, other studies have reported that FFQ did not systematically overestimate energy and nutrients intakes [ 14 , 21 - 23 ]. Despite the fact that we obtained similar values for energy intake with both dietary methods, we can not exclude the possibility that both tools are subjected to underreporting and therefore underestimate usual dietary intakes. It has been previously suggested that subjects may tend to underreport actual food intake by as much as 20% when completing a weighted dietary record [ 24 ]. It has been argued that subjects who complete 3-day food record may change their nutritional food habits in order to simplify the recording of food intakes or to impress the dietitian. Also, errors in 3-day food records can be attributable to interpretation of the dietitian encoding the records. In our study, the same dietitian verified all the food records to make sure that dietary data were coded similarly for all participants. In the present study, 38% of subjects included in the validation study at week 0 had a ratio between energy intake to estimate BMR below 1.35. Considering that they had to be weight stable to be included in the study it is likely that these women were underestimating their habitual diet. Black et al concluded in a review that underreporting was observed in a great majority of nutritional surveys independently of the method used [ 25 ]. Earlier studies conducted in lean women demonstrated that underreporting was mainly explained by undereating [ 26 ] or underreporting snack foods [ 27 ] whereas in obese subjects underreporting could be explained by an underestimation in recording portion size and to social desirability. In addition, underreporting occurs more often among foods considered 'bad' or 'unhealthy' [ 28 ]. In our validation study, there were no significant differences between BMI of women who were considered as underreporters and women who did not underreport (not shown). In a nutritional intervention, interpretation of the study outcomes with regard to dietary changes will depend not only of the validity and the reproducibility of the method used but also of the sensibility of the method to detect dietary changes in response to the intervention. In our nutritional intervention study, conducted in a sample of healthy women, both diet assessment methods detected similar dietary changes over the duration of the intervention. These findings suggested that our FFQ is sensitive to dietary changes in response to our intervention and could be used to assess dietary changes during a nutritional intervention. Our results are in agreement with study that showed that in response to a nutritional intervention a FFQ measured similar dietary changes as compared to 24-hour recalls [ 29 ] or 4-day food records [ 7 ]. The major differences between the two methods in our study were noted for total lipids and MUFA intakes. Our FFQ was designed to assess precisely lipid intake and many questions were asked about types of fat used to spread or to cook. The more important differences between FFQ and 3-day food record for MUFA and lipids could therefore be explained by the fact that it was difficult for participants to report precisely their lipid consumption when completing the FFQ. It has also been reported in obese men that underreporting of food record is usually specific to lipid intake [ 30 ] and it is thus possible that some women did not record all fats or foods high in fat consumed when completing their 3-day food record. Therefore, it is difficult to determine whether our FFQ tended to overestimate lipid or whether the 3-day food record tended to underestimate it. Also, dietary changes for these nutrients were in the same magnitude in response to our intervention with both methods. On the other hand, Mediterranean diet is usually considered high in MUFA. In North America, MUFA are mostly provided by partially hydrogenated vegetable oils and animals products [ 31 ]. In that context, MUFA to SFA ratio could be considered as a better indicator of a Mediterranean diet. In our study, we noted that this ratio was not different between the two methods at baseline and changes observed in response to the nutritional intervention did not differ significantly (not shown). The agreement in quartile classification was acceptable for selected nutrients with a mean of 35.1% of subjects who were in exact agreement and 5.1% who were misclassified in extreme quartiles. This finding is similar to previous observations [ 16 , 19 , 32 , 33 ]. In many studies, classification in the same segment of the distribution using two different methods is found in 30% to 40% of subjects [ 16 , 32 , 34 ]. When analyses were performed at week 6 and 12 after the beginning of the nutritional intervention, coefficients of correlation were slightly higher than at week 0. We suggest that this finding be partly explained by the intervention effect. As previously reported [ 35 ] subjects could be influenced by a learning effect. In fact, subjects could be influenced by the first FFQ experience and be more adequately prepared for the second FFQ. The nutritional intervention may have also influenced the manner in which subjects were completing their 3-day food records during the study. In a nutritional intervention, it is important to use a reproducible method to insure that dietary changes observed are due to the intervention effects and not to the instrument error. Our study suggests that the FFQ presents a good degree of reproducibility. In fact, in reproducibility studies the coefficients of correlation generally ranged from 0.5 to 0.7 [ 6 ]. In our study, coefficients of correlation ranged, after energy adjustment, from 0.62 for vitamin C to 0.83 for protein intakes. These values are similar to correlations reported by others [ 14 , 18 , 19 , 21 , 22 , 33 , 35 - 37 ]. In our reproducibility study, lower mean energy intake and nutrient intakes were found at the second administration of the FFQ as compared to the first FFQ (difference of approximately 10%). However, relatively high and uniform correlation coefficients for values derived from the two FFQs were observed. Riley et al [ 35 ] also reported with an administered FFQ that energy intake was 10% lower at the second FFQ administration and this reduction was uniform for all nutrients studied. In our study, intakes of most nutrients were systematically higher when measured with the first FFQ compared to the second one, except for alcohol consumption, which remained the same. Seasonal variation can not explain this difference because both FFQs were administered during the same season. The fact that subjects estimated a lower frequency of intake during the second administration of the FFQ may be explained by their earlier experience in completing the FFQ. Better general knowledge of dietary intakes could lead to a readjustment in estimation of intakes after the first administration of the FFQ and therefore changes in estimated energy intake. Conclusions In conclusion, the FFQ that we developed to estimate usual average daily energy and nutrient intakes in subjects from the Québec City metropolitan area is valid and reproducible. Mean energy and nutrient intakes were estimated accurately by our FFQ compared to the 3-day food record. The fact that our FFQ was administered by a dietitian trained to insure a standardized administration of FFQ was important to optimize the validity and reproducibility of the method. We also showed that both methods appeared to underestimate energy intake in a large proportion of subjects. Usually, food records are considered as the gold standard method to assess dietary intakes. It is however important to recognize that food records also have their own limitations. In nutritional studies, an interviewer-administered FFQ, as the one we used in the present study, can be used to assess energy and nutrient intakes and requires less time to compute dietary informations than food records. FFQ also decreases the possibility of interpretation by the coding person. Finally, there is still a need to develop other efficient methods to measure dietary intakes that will permit to more closely match habitual dietary intakes of individuals in their living environment. List of abbreviations used CHD: coronary heart disease FFQ: Food Frequency Questionnaire BMR: basal metabolic rate (BMR) BMI: body mass index NDS-R: Nutrition Data System for Research CHO: carbohydrates SFA: saturated fatty acids MUFA: monounsaturated fatty acids PUFA: polyunsaturated fatty acids EPA: eicosapentaenoic acid DHA: docosahexaenoic acid Competing interests None declared. Authors' contributions JG participated to data collection, performed data analysis and drafted the manuscript. GN and AL participated to data collection. BL and SL conceived the study, and participated in its design and coordination. All authors read and approved the final version of the manuscript.	/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC522826.xml
497044	Irinotecan plus folinic acid/continuous 5-fluorouracil as simplified bimonthly FOLFIRI regimen for first-line therapy of metastatic colorectal cancer	Background Combination therapy of irinotecan, folinic acid (FA) and 5-fluorouracil (5-FU) has been proven to be highly effective for the treatment of metastatic colorectal cancer. However, in light of safety and efficacy concerns, the best combination regimen for first-line therapy still needs to be defined. The current study reports on the bimonthly FOLFIRI protocol consisting of irinotecan with continuous FA/5-FU in five German outpatient clinics, with emphasis on the safety and efficiency, quality of life, management of delayed diarrhea, and secondary resection of regressive liver metastases. Methods A total of 35 patients were treated for metastatic colorectal cancer. All patients received first-line treatment according to the FOLFIRI regimen, consisting of irinotecan (180 mg/m 2 ), L-FA (200 mg/m 2 ) and 5-FU bolus (400 mg/m 2 ) on day 1, followed by a 46-h continuous infusion 5-FU (2400 mg/m 2 ). One cycle contained three fortnightly administrations. Staging was performed after 2 cycles. Dosage was reduced at any time if toxicity NCI CTC grade III/IV was observed. Chemotherapy was administered only to diarrhea-free patients. Results The FOLFIRI regimen was generally well tolerated. It was postponed for one-week in 51 of 415 applications (12.3%). Dose reduction was necessary in ten patients. Grade III/IV toxicity was rare, with diarrhea (14%), nausea/vomiting (12%), leucopenia (3%), neutropenia (9%) and mucositis (3%). The overall response rate was 31% (4 CR and 7 PR), with disease control in 74%. After primary chemotherapy, resection of liver metastases was achieved in three patients. In one patient, the CR was confirmed pathologically. Median progression-free and overall survival were seven and 17 months, respectively. Conclusions The FOLFIRI regimen proved to be safe and efficient. Outpatient treatment was well tolerated. Since downstaging was possible, combinations of irinotecan and continuous FA/5-FU should further be investigated in neoadjuvant protocols.	Background Patients with advanced C olo R ectal C ancer (CRC) have been demonstrated to benefit from chemotherapeutic treatment in terms of both quality and duration of life [ 1 ]. Of these treatments, fluoropyrimidines are the most often used and best investigated drugs [ 2 , 3 ]. 5-Fluorouracil (5-FU)-based chemotherapy, – usually biomodulated with folinic acid (FA) to increase its affinity for thymidylate synthase [ 4 ] -, was associated with an approximate doubling of the median survival compared to routine standard of care [ 5 ]. Additionally, recent phase III studies suggested that combinations of FA/5-FU with irinotecan or oxaliplatin resulted in improved response rates and prolonged survival [ 6 - 9 ]. These encouraging results prompted the use of combination therapy of irinotecan and FA/5-FU as a first line chemotherapy in CRC. Irinotecan (CPT-11), a potent inhibitor of the enzyme topoisomerase I, has demonstrated anti-tumorogenic activity in metastatic colorectal cancer, when used alone or in combinations with FA/5-FU, as adjuvant or palliative treatment (for review see 10 or 11). In randomized phase III clinical trials, second-line therapy with irinotecan significantly improved survival compared to supportive care [ 12 ] or to infusional FA/5-FU [ 13 ]. In the first-line setting of metastatic colorectal cancer, two randomized multi-center phase III clinical trials demonstrated synergistic activity of irinotecan with both bolus and infusional FA/5-FU regimens [ 6 , 7 ]. In both studies, combinations of irinotecan and FA/5-FU were superior to the control arms, irinotecan alone or FA/5-FU, specifically in regard to response rate, progression-free, and overall survival. However, the best regimen of irinotecan and FA/5-FU has yet to be defined. Altogether, irinotecan combined with continuous FA/5-FU infusions seemed to be superiorly active and less toxic than combination with FA/5-FU bolus regimens [ 6 , 7 ]. Recently, irinotecan was investigated in a bi-monthly protocol with bolus FA/5-FU and a continuous 48 h infusion 5-FU (simplified LV5-FU2 regimen; FOLFIRI) [ 14 - 16 ]. In a consecutive phase III clinical trial, a response rate of 56% was achieved with FOLFIRI, as compared to 39% achieved with the original Saltz protocol [ 6 , 14 ]. Furthermore, irinotecan schedules of weekly and of once every two or three weeks demonstrated similar efficacy and quality of life, as well as significantly lower incidences of severe diarrhea in patients with 5-FU-refractory, metastatic colorectal cancer [ 15 - 17 ]. In contrast to the irinotecan and bolus FA/5-FU regimen, which has attracted criticism due to unexpectedly high early death rates due to gastrointestinal toxicity and thromboembolic events observed in two subsequent trials [ 18 , 19 ], no increased 60-day mortality rate was found in two recent trials, each with continuous 5-FU treatment arms [ 14 , 17 , 20 , 21 ]. To date, little data is available regarding irinotecan combined with the simplified bi-monthly LV5-FU2 regimen as a first-line treatment in patients with metastatic colorectal cancer [ 14 ]. Therefore, we initiated this prospective open-label, multi-center phase IV clinical trial to evaluate its toxicity and efficacy in German outpatient clinics. We were especially interested in the safety of this regimen in an outpatient setting, with particular emphasis placed on the prompt and aggressive management of delayed diarrhea with loperamide, hospitalization and parenteral rehydration in case of refractory diarrhea lasting more than 48 hours [ 22 , 23 ]. Furthermore, all patients were closely monitored for the possibility of resection of liver metastases after successful response. Methods Accrual and eligibility After approval by the local ethical committee, patients were consecutively recruited from five German outpatient clinics (one university hospital, one community hospital, and three general practices). Eligibility requirements included (1) histologically documented adenocarcinoma of the colon or rectum and progressive measurable metastatic disease, (2) minimum life expectancy of three months, (3) Karnofsky performance status ≥ 60, (4) adequate hematologic, hepatic, and renal function, and (4) no prior chemotherapy for metastatic disease. Participants needed to be between 18 and 75 years of age. This study required that previous adjuvant 5-FU-based therapy with or without radiation therapy be completed at least 6 months prior to entry. Patients with CNS metastases, bowel obstruction, or ileus were excluded from the study. The study was approved by the ethics commission board responsible for all participating institutions. Prior to treatment, all patients gave written informed consent. Treatment and management of toxicity As previously described [ 14 - 16 ], treatment consisted of the bi-monthly combination of irinotecan 180 mg/qm given as a 90-min intravenous infusion, day (d)1, FA 200 mg/m 2 d1, 5-FU bolus 400 mg/m 2 d1, followed by 5-FU 46 h continuous infusion 2400 mg/m 2 (simplified LV5-FU2 schedule). To prevent expected toxicities, patients were carefully informed of the potential risk of delayed diarrhea and neutropenia and the need for early intervention with loperamide [ 22 ] and metoclopramide, prophylactic antibiotics, or hospitalization and parenteral rehydration in case of refractory diarrhea lasting more than 48 hours. Patients with loperamide-resistant diarrhea defined as loose stools persisting for more than 24 hours despite adequate treatment with loperamide would receive a trial of the oral steroid budesonide (9 mg per day for a maximum of 4 days) [ 23 ]. Atropine was given for irinotecan-related cholinergic symptoms if needed [ 25 ]. Antiemetic treatment was performed using metoclopramide or HT5 antagonists in a sequential manor. The prophylactic use of colony-stimulating factors was not permitted. Treatment was continued until one of the following occurred: disease progression, unacceptable adverse effects, or withdrawal of consent by the patient. Assessments Primary measures of the study were the overall objective response rate (ORR, complete and partial responses), overall survival, and quality of life. Secondary measures included the disease control rate (ORR + stable disease), time to progression, and frequency and severity of toxicities. Quality of life was assessed after inclusion into the study and as often as possible during the course of treatment, using the EORTC QLQ-C30 (version 2) questionnaire [ 24 ]. Safety assessments and complete blood counts were performed weekly. Toxicity was graded according to National Cancer Institute Common Toxicity Criteria (NCI CTC). Toxicities not defined by NCI CTC criteria were classified as grade 0 (none), grade I (minor), grade II (moderate), grade III (severe), and grade IV (life-threatening). In case of any toxicity grade II, with the exception of hand-foot syndrome or alopecia, the next scheduled doses of irinotecan, folinic acid and 5-FU were delayed for a maximum of 1 week (or resolution of diarrhea for at least five days). In case of toxicity grade III/IV or if improvement from grade II to I (or resolution of diarrhea) was not achieved by two weeks, the following chemotherapy doses were reduced by 20 percent. If grade III/IV toxicity did not improve by 2 weeks, treatment was discontinued. Dose reductions were mandatory from the first cycle of chemotherapy in case of toxicity higher than grade II, and chemotherapy was resumed only after complete recovery from diarrhea. Tumor response was assessed according to World Health Organization (WHO) criteria. Tumor reassessment by the same imaging method used to establish baseline tumor measurement was generally performed after every two courses of therapy until progression. A complete response (CR) was defined as complete disappearance of evidence of cancer. A partial response (PR) was defined as a reduction in the sum of the products of the bi-perpendicular diameters of all measurable lesions by at least 50%. Progressive disease (PD) was defined as an increase in the sum of the products of the greatest bi-perpendicular diameters of all measurable lesions by at least 25% or the appearance of new lesions. Stable disease was defined as any reduction or increase in measurable lesions which did not meet the criteria for PR or PD. Confirmed objective responses were those for which a follow-up scan obtained at least four weeks later demonstrated the persistence of the response. The assessment of response and progression was based on investigator-reported measurements. Statistical analysis Statistical analysis including survival analysis according to Kaplan-Meier was performed with the SPSS software package. The deadline for data evaluation was March 15, 2004. Survival was measured from the time of diagnosis to the date of death or last follow-up. Progression-free survival was calculated from treatment onset to the time of progression, study withdrawal or death of any cause. Patients who received at least one dose of the treatment regimen were evaluated for toxicity, and patients who completed at least two chemotherapy cycles were evaluated for response. Results Between 10/2001 and 5/2003, 35 consecutive patients (25 male, 10 female) with metastatic colorectal cancer were enrolled into the study. The median age of these patients was 62 years, ranging from 38 to 73. Baseline characteristics of all patients are summarized in table 1 . Most patients were in good overall physical condition, although 60% had at least two metastatic sites. All patients had undergone surgery prior to chemotherapeutic treatment. Six patients previously had received adjuvant chemotherapy, one of them in combination with radiation therapy. All patients were evaluated for toxicity, for response, and survival. The 35 patients received a total of 151 chemotherapy cycles (mean 4,3 per patient), consisting of 451 administrations. Overall, 51 (12,3 %) of all administrations had to be delayed for one week. During the complete study period, 19 patients had a delay of therapy for a median of nine days and in 10 patients (29%) a dose reduction was necessary at some point during the treatment period. The most common cause for discontinuation of study treatment was disease progression (18 patients, 51%). In case of discontinuation, 16 (46%) patients received a second line treatment with either oxaliplatin plus a FA/5-FU consisting regimen or an epidermal growth factor receptor antagonist. Hematologic toxicity was mild to moderate in the majority of patients (table 2 ). Only one patient (3 %) had a grade III/IV leucopenia, three patients (9 %) had a grade III neutropenia, and grade III or IV anemia or thrombocytopenia were not observed. The predominant non-hematologic toxicities were nausea/vomiting and delayed diarrhea, which affected a total of 21 (60%) and 10 (29%) patients, respectively. However, grade III/IV of these side effects were only observed in 4 (11%) and 2 (6%) patients, respectively (table 3 ). Other non-hematological toxicities were predominantly mild, including mucositis (I°, 5 patients, 14%), fever (I°/II°, 5 patients, 14%), cholinergic syndrome (I°, 2 patients, 6%), constipation (I°, 8 patients, 23%), alopecia (I°, 9 patients, 26%, II°, 1 patient, 3%), asthenia (I°, 2 patients, 6%). Regarding the irinotecan induced delayed diarrhea, 11 patients received at least one course of loperamide [ 22 ] and 1 patient received budesonide for loperamide refractory diarrhea [ 23 ]. In two patients treatment-related hospital admissions as a result of III/IV leucopoenia and grade III diarrhea were required. Other adverse events in three patients included a bowel obstruction due to local recurrence, unexplained vertigo, and renal failure due to urethral obstruction. Pulmonary embolism did not occur in any patients during treatment. With regard to response, four complete (CR) and seven partial responses were seen, and thus an overall response rate of 31% was observed (table 4 ). In addition, 15 patients (43%) had stable disease (disease control rate, 74%). Disease progression occurred in nine patients (26%). Resectability of metastases was achieved in three patients. In one patient CR, was pathologically confirmed. Median progression-free survival was seven months and overall survival was 17 months (95% confidence intervall: 9–25 months, figure 1 ). Quality of life data were obtained before and at least once during treatment from 13 patients [ 24 ]. The 13 patients evaluated for quality of life did not differ in their pattern of response to chemotherapy from the total population of all evaluated patients. Global health status improved slightly during treatment compared to pre-therapy values (figure 2 ). In addition, patients treated with the FOLFIRI regimen had a small increase in emotional and physical wellbeing compared to a previously reported cohort of untreated patients. No remarkable changes in the other items of the questionaire were seen during treatment, especially with regard to therapy-dependent symptoms such as nausea and vomiting, diarrhea and pain. Slightly increased nausea and fatigue were observed in our patients. However, a clear trend could not be concluded from our data. Discussion In the current phase IV study we evaluated toxicity and efficacy of the FOLFIRI combination of irinotecan with FA/5-FU as first-line chemotherapy for metastatic colorectal cancer. This combination was established within a phase I clinical trial with a recommended irinotecan dose of 180 mg/qm [ 15 , 16 ]. Irinotecan-containing regimens have been the most commonly used chemotherapy protocols for metastatic colorectal cancer in North America since the publications of Saltz et al. [ 3 , 6 , 7 , 12 , 17 ]. After unexpectedly high early death rates, due to gastrointestinal toxicity and thromboembolic events, which were reported in two subsequent trials due to gastrointestinal toxicity and thromboembolic events, the safety of the Saltz regimen became a subject of considerable debate [ 18 , 19 ]. However, in our previous experience with this regimen we did not observe any major complications [ 25 ]. Moreover, additional comprehensive data showed that combinations of irinotecan with continuous FA/5-FU, either weekly or bi-monthly, resulted in higher response rates and better survival. Therefore, we investigated the bi-monthly FOLFIRI schedule in an outpatient setting for its safety and clinical efficacy. In the majority of our patients the FOLFIRI regimen was well tolerated. Gastrointestinal toxicity or thrombembolic events were never fatal. Most hospitalizations were for prevention rather than treatment of life-threatening conditions. Delayed diarrhea, a well known side effect of irinotecan [ 22 ], was generally managed in the outpatient setting using loperamide, which was administered to approximately one third of the patients. Budesonide, which has demonstrated activity in loperamide-refractory diarrhea was required in only one of our patients (3%). Overall, we observed relatively low toxicity in our study, with NCI CTC grade III leucopenia amounting to 3%, diarrhea to 14% and nausea/vomiting to 11%. The toxicity observed in our study was lower than that reported by Douillard et al. in the pivotal European first-line trial in the patient group receiving weekly irinotecan (80 mg/qm), 24-h HD-5-FU (2300 mg/qm) preceded by 2-h FA 500 mg/qm [ 7 ]. In this patient group, grade III/IV toxicities were reported for leucopenia 20.4%, diarrhea 44.4% and nausea 7.4%. The lower toxicity in our study might be due to the lower per day doses of 5-FU (2400 mg/qm administered over 48 h) and L-FA (200 mg/qm) used. In the EORTC phase III study 40986, comparing first-line AIO schedule alone with irinotecan (80 mg/qm), FA 500 mg/qm and continuous FA/5-FU (2300 mg/qm), the 5-FU-dose had to be reduced to 2000 mg/qm because of initially high toxicity in an interim analysis [ 26 ]. In addition, lower toxicity in our study may have been more limited because of the early and rigorous dose reductions according to our protocol. Furthermore, we observed improved physical and emotional status and an increase in global health status during treatment [ 24 ]. This is in concordance with our previously reported data [ 25 ]. Tournigand et al. demonstrated an increase in weight of at least 5% in 35% and an improved physical status in 35% of the irinotecan/FA/5-FU treated patients, respectively [ 14 ]. Koehne et al. reported a significantly better quality of life in the irinotecan/FA/5-FU group compared to FA/5-FU [ 26 ]. The response rate achieved in our study (31%) was quite comperable tp previously published data [ 6 , 7 , 14 , 26 ]. In these studies, response rates were 40–56 % with time to progression (TTP) of 6–8 months. Tumor control (CR+PR+SD) was achieved in 74% of our patients, which is similar to other reports. Median progression-free and overall survival, was comparable, but slighty less than 8,5 and 21,5 months reported by Tournigand [ 14 ]. Three reasons may account for these differences in survival between the studies. The most important reason may be that a higher percentage of patients (21 patients, 60% compared to 41% and 10% reported by Tournigand and Saltz, respectively) in our study had two or more metastatic sites, indicating a larger tumor burden and consequently a worse prognosis regarding survival [ 6 , 14 ]. Second, our study included patients with carcinoma of the rectum (12 patients, 34% versus 15 %). And finally, as much as five (14%) of our patients had previously received adjuvant FA/5-FU-containing chemotherapy or radiotherapy compared with 10% of the patients in the other study [ 6 ]. It appears particularly noteworthy that after chemotherapy three of our patients achieved surgical resectability of their metastases. To our knowledge these results are the first ever reported to suggest a potential role for the FOLFIRI regimen in the neoadjuvant setting. Thus far, studies of regional chemotherapy for initially unresectable colorectal liver metastases could demonstrate some success with secondary curative surgery. In two recently published retrospective studies chronomodulated chemotherapy with oxaliplatin and FA/5-FU was used as neoadjuvant treatment for patients with unresectable colorectal liver metastases [ 27 , 28 ]. Therefore, combination regimens of irinotecan or oxaliplatin with FA/5-FU should be strongly considered as standard first-line chemotherapy for metastatic colorectal cancer. Through multidisciplinary efforts involving both surgeons and medical oncologists, it should be possible, to translate the antitumour activity of the new first-line regimens into long-term survival benefit for patients with initially unresectable colorectal liver metastases. Conclusions The FOLFIRI regimen, consisting of irinotecan with continuous FA/5-FU over 48 h, given on an outpatient basis was safe and well tolerated in our study. The rate of severe side effects was comparably low with this bi-monthly regimen. As tumor control was achieved in about 75% and downstaging of metastatic disease was possible in some cases, combinations of irinotecan plus continuous FA/5-FU should be further investigated in neoadjuvant protocols for initially unresectable liver metastases. List of abbreviations CRC Colorectal Cancer FOLFIRI chemotherapeutic regimen consisting of irinotecan combined with continuous FA/5-FU infusions FA Folinic Acid 5-FU 5-Fluorouracil HD-5-FU high dose 5-FU NCI CTC National Cancer Institute Common Toxicity Criteria CR Complete remission PR Partial response SD Stable disease PD Progressive disease TTP Time to progression OS Overall survival Author's contributions MM, SS and AT drafted the manuscript. MM and MH initiated the study. JS, CZ, HH, BA, MS, OK, T, PG and MH were involved in the patient's treatment as well as the documentation of response and side effects. Pre-publication history The pre-publication history for this paper can be accessed here:	/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC497044.xml
555463	Treatment of osteoporosis in an older home care population	Background Previous research indicates that many patients with fractures indicative of underlying osteoporosis are not receiving appropriate diagnostic follow-up and therapy. We assessed osteoporosis treatment coverage in older home care clients with a diagnosis of osteoporosis and/or prevalent fracture. Methods Subjects included 330 home care clients, aged 65+, participating in a longitudinal study of medication adherence and health-related outcomes. Data on clients' demographic, health and functional status and service utilization patterns were collected using the Minimum Data Set for Home Care (MDS-HC). A medication review included prescribed and over-the-counter medications taken in the past 7 days. Criteria for indications for osteoporosis therapy included diagnosis of osteoporosis or a recent fracture. Coverage for treatment was examined for anti-osteoporotic therapies approved for use in 2000. Results Of the 330 home care clients, 78 (24%) had a diagnosis of osteoporosis (n = 47) and/or had sustained a recent fracture (n = 34). Drug data were available for 77/78 subjects. Among the subjects with osteoporosis or a recent fracture, 45.5% were receiving treatment for osteoporosis; 14% were receiving only calcium and vitamin D, and an additional 31% were receiving drug therapy (bisphosphonate or hormone replacement therapy). The remaining 54.5% of subjects were not receiving any approved osteoporosis therapy. Conclusions The high prevalence of undertreatment among a population of older adults with relatively high access to health care services raises concern regarding the adequacy of diagnosis and treatment of osteoporosis in the community.	Background Osteoporosis is defined as ' . . .a skeletal disorder characterized by compromised bone strength predisposing a person to an increased risk of fracture. Bone strength reflects the integration of two main features; bone density and bone quality' [ 1 ]. The presence of a vertebral or other fragility fracture is a strong predictor of risk of future fracture, and is a major indicator of the presence of osteoporosis [ 2 - 4 ]. A fragility fracture is defined as a low trauma fracture (eg. a fall from a standing height or less) [ 3 ]. Many studies indicate that fewer than 30% of patients with fragility fractures are being treated for osteoporosis. Rates of treatment appear particularly low following hospital discharge (<10%) [ 5 - 7 ], suggesting that treatment recommendations are not being made to family physicians. Assessment of treatment rates 6 months to 2 years after fracture are somewhat higher (in the 20–38% range) [ 8 - 13 ]. Kiebzak et al. followed patients for 1–5 years after hip fracture. At follow-up, they found that 71% of women were receiving treatment, but the proportion of men receiving therapy remained relatively low [ 14 ]. Four Canadian studies of osteoporosis treatment following fragility fracture were identified. Two studies examined osteoporosis treatment among hip fracture patients after hospital discharge, and in both, treatment rates were relatively low (< 10%) [ 6 , 15 ]. In Ontario, Hajcsar et al. interviewed patients one year after fragility fracture, and found that the use of approved pharmaceutical agents such as bisphosphonates and hormone replacement therapy (HRT) remained low (7.4% and 16%, respectively) [ 13 ]. Khan et al. reported considerably higher treatment rates among patients after wrist fracture, with approximately 38% of patients receiving either a bisphosphonate or HRT at follow-up [ 9 ]. The purpose of this study was to examine osteoporosis treatment among older home care clients with a diagnosis of osteoporosis and/or prevalent fracture. As this population has relatively frequent contact with health care providers, we expected that osteoporosis treatment rates should be higher than previously reported. Methods Participants were 330 older home care clients participating in a longitudinal study of medication adherence and related health outcomes. Details of the primary study can be found elsewhere [ 16 ]. Briefly, between April and June of 2000, 585 home care clients residing in two southern Alberta health regions were identified by random samples stratified by rural/urban residence. Rural clients represented those living on a farm, acreage or in a village or town (with a population less than 10,000) and residing greater than 35 km from a major urban centre (cities of Calgary or Lethbridge). Inclusion criteria were: currently receiving home care services under the jurisdiction of their respective health region, age 65 or greater and provision of informed consent from either the subject or a legal guardian. Of the original random sample of 585, 10 subjects had died and 10 had moved or were unavailable at the time of study recruitment. Further exclusions included subjects who were in hospital, transferred to long term or palliative care, or mentally incompetent without a legal guardian (n = 40), who required a translator (n = 6) or who posed a safety concern for the study nurses (n = 7). Of the remaining 512 subjects, 46 (9.0%) refused to participate and 136 (26.6%) were not contacted after the predetermined sample size (based on estimation of differences in adherence rates between rural and urban subjects) [ 16 ] had been achieved. Data on demographics, health and functional status and service utilization were collected with a standardized assessment tool, the Minimum Data Set for Home Care (MDS-HC) [ 17 , 18 ], and several supplemental questions regarding medication use, smoking and tobacco use. Information on therapeutic substances was recorded from container labels for all substances used in the past 7 days. Therapeutic substances included prescribed, over-the-counter and complementary or alternative products. All drug data were entered into a database using the Anatomical Therapeutic Chemical classification system (ATC codes). Assessment of medication adherence was based on self-report data [ 16 ]. Four study nurses, trained in the administration of the MDS-HC and medication assessment, collected data during in-home interviews lasting approximately 1.5 hours. This study received ethics approval from the Health Research Ethics Board of the University of Calgary and the Ethics Review Committee of the Chinook Health Region. Fracture prevalence and diagnosis of osteoporosis were determined from the disease diagnoses section of the MDS-HC (Section J). The MDS-HC includes specific categories for charting of: 'hip fracture', 'other fracture', and 'osteoporosis'. Prevalent fractures were defined as those that had resulted in a hospitalization (in last 90 days), currently required treatment and/or symptom management or were being monitored by a home care professional. Coverage for treatment was examined for anti-osteoporotic therapies approved for use in 2000 [ 19 ]. These included calcium with vitamin D, and the following pharmaceutical agents: etidronate, alendronate, hormone replacement therapy, raloxifene and calcitonin. Prevalence estimates of treatment for osteoporosis among subjects with a prevalent fracture and/or diagnosis of osteoporosis were reported. Descriptive bivariate comparisons were also conducted to examine potential variations between osteoporotic subjects who received treatment and those who did not. Fisher's exact test was reported for the bivariate comparisons. Due to the limited number of subjects with osteoporosis or fracture, multivariable analyses were not feasible. Results Demographic and health status variables are summarized in Table 1 . The mean age of the total sample (n = 330) was 82 years (standard deviation = 7.8, range 65–101) and most clients were female (78.5%) and not married (70.6%). Almost half (41.8%) had completed at least high school education and a substantial proportion of subjects lived in a seniors' lodge (36.7%). A total of 78 subjects (23.6%) had at least one potential indication for osteoporosis treatment. The demographic characteristics of these subjects were similar to the sample as a whole (Table 1 ), with the exception that this group included a greater proportion of women (93.6%). Demographic data were only available for a sub-group of all non-respondents. Analyses of this sub-sample showed no significant differences between non-respondents and respondents in relation to age and sex. Table 1 Summary of demographic characteristics and health status among older home care clients. Variable Total Study Population N = 330 % (n) Subjects with diagnosis of osteoporosis or fracture N = 78 % (n) Demographics Age Mean (SD) 82 (7.8) 83 (8.3) (≥ 85) 43.3 (143) 42.3 (33) Female 78.5 (259) 93.6 (73) Not married 70.6 (233) 79.5 (62) Education (≥ High School) 41.8 (138) 44.9 (35) Living arrangements Seniors' Lodge* 36.7 (121) 34.6 (27) Health Status Cognitive impairment (CPS** score >1) 23.3 (77) 24.4 (19) # Comorbid conditions (>2) 71.8 (237) 71.8 (237) * Versus private home ** MDS-HC Cognitive Performance Scale The prevalence of osteoporosis diagnoses and fractures are summarized in Table 2 . Forty-seven subjects (14.2%) had a diagnosis of osteoporosis, three of whom had at least one prevalent fracture. Prevalence of at least one fracture without diagnosis of osteoporosis was recorded for 31 (9.4%) subjects; 13 hip fractures (1 with 'other fracture'), and 18 with 'other' fractures. One client had only unlabelled medication bottles, thus medication data were available for 77 of these subjects. Approximately 50% of subjects with a diagnosis of osteoporosis or prevalent hip fracture received treatment, compared to only 28% of subjects with 'other' fractures (p = 0.11). Table 2 Indications for treatment of osteoporosis among older home care clients. Diagnoses N = 330 Receiving treatment* N = 77 Indication % (n) % (n) Diagnosis of osteoporosis** 14.2 (47) 50.0 (23/46 † ) Hip fracture ‡ 3.9 (13) 53.8 (7/13) 'Other' fracture ?? 5.5 (18) 27.8 (5/18) Total 23.6 (78) 45.5 (35/77) * Treatments included: calcium with vitamin D and/or drug therapy (bisphosphonate or hormone replacement therapy). ** Three subjects also had at least one fracture. † Medications unlabelled for one subject. ‡ One subject also had 'other fracture' ?? Type of fracture was available for 10/18 subjects: 3 wrist, 1 humerus, 2 ankle, 2 vertebral, 1 shoulder and 1 rib fracture. Specific therapeutic interventions for osteoporosis among subjects with a diagnosis of osteoporosis and/or fracture are shown in Figure 1 . No subjects were using calcitonin or raloxifene. In total, only 35 subjects (45%) were receiving at least minimal treatment for osteoporosis (i.e. at least calcium with vitamin D). Only 30 subjects (39%) were receiving therapy with an approved prescription drug (bisphosphonate and/or hormone replacement therapy), with 5 subjects (6%) being treated only with calcium and vitamin D. Six subjects receiving prescription drug therapy were also taking calcium and vitamin D (not shown in Figure). Of the 42 subjects not receiving minimal treatment (i.e. drug therapy and/or calcium with vitamin D), 3 were taking calcium only, and 10 were taking a multivitamin that may have contained calcium and/or vitamin D. Bivariate comparisons of potential factors that may influence receipt of treatment are summarized in Table 3 . A greater proportion of nonadherent subjects were observed in the group that did not receive treatment (50% versus 31%). The proportion of subjects with lower education levels was also relatively higher among the non-treated group. None of the 5 men among the osteoporotic group were receiving treatment. Table 3 Comparison of osteoporotic subjects by treatment status. (n = 77) Variable No Treatment n = 42 % (n) Treatment n = 35 % (n) Fisher's exact p-value Age (≥ 85) 42.8 (18) 42.8 (15) 1.00 Sex (male) 11.9 (5) 0 0.06 Education (≥ High School) 38.1 (16) 54.3 (19) 0.18 Residence (seniors' lodge) 40.4 (17) 28.6 (10) 0.34 Cognitively impaired 23.8 (10) 22.8 (8) 0.60 Depression 14.3 (6) 14.3 (5) 0.75 Confined to wheelchair/bed 21.4 (9) 20.0 (7) 1.00 Fall(s)in past 90 days 31.0 (13) 22.9 (8) 0.45 On steroid medication 4.76 (2) 11.4 (4) 0.40 Nonadherent (with any prescribed medication) 50.0 (21) 31.4 (11) 0.11 Figure 1 Osteoporosis therapy among subjects with a diagnosis of osteoporosis and/or fracture. (HRT = hormone replacement therapy) No treatment (n = 29) Calcium only (n = 3) Multivitamin (n = 10) Bisphosphonate (n = 21) HRT (n = 7) Bisphosphonate + HRT (n = 2) Calcium + vitamin D only (n = 5) Discussion Home care clients are expected to have relatively frequent contact with health care providers. Among those with a diagnosis of osteoporosis or prevalent fracture (n = 78), 30% had weekly contact with nursing staff, 42% had at least one emergency or emergent care visit or hospitalization in the past 3 months, and the majority (67%) had a recent (<6 months) review of their total medication regimen. Thus, more opportunity for intervention following fracture, and higher rates of treatment would be expected among this population. Although the proportion of home care clients receiving treatment was somewhat higher than reported in most previous studies (45% versus <39%, respectively) [ 5 - 8 ], [ 10 - 13 ], the majority of subjects with a diagnosis of osteoporosis or fracture indicative of osteoporosis were still not receiving even minimal therapy (i.e. calcium and vitamin D). Consistent with previous reports [ 14 ], we also found that men were not receiving therapy following fracture. In this home care sample, none of the 5 men with indication for osteoporosis treatment were receiving therapy. Three of these men had fractures at sights other than at the hip, where lower intervention rates were observed even for women. However, 2 of these men had a charted diagnosis of osteoporosis and were still not receiving treatment. Our data also indicate that subjects who fracture at sites other than the hip may be less likely to receive treatment. While many of the fractures observed in this study (Table 2 ) were typical of osteoporotic fractures with respect to fracture site [ 2 , 4 ], our data are limited in that the MDS-HC does not include an assessment of the cause of the fracture. If some fractures were the result of trauma, our findings may overestimate the deficit in osteoporosis therapy. Conversely, the prevalence of osteoporosis may be underestimated in our study population, as the diagnosis was based on self-report information. The prevalence of osteoporosis in Canadian women aged 50+ years, based on BMD, has been estimated at 15% [ 20 ]. If diagnosis were based on BMD alone, a higher prevalence would be expected in the older population in this study. However, physicians' diagnoses of osteoporosis may derive from factors other than BMD, and our findings are consistent with previous work that suggests that many patients with osteoporosis are not receiving appropriate therapy. We also examined several factors potentially associated with undertreatment. However, the small numbers of subjects with osteoporosis or prevalent fracture limit the interpretation of our findings. Although some trends are apparent, such as undertreatment of men and those with lower education, we cannot comment conclusively on these observations due to the limited sample size and cross-sectional nature of the study. Further evaluation of these factors utilizing a larger sample and appropriate multivariable analysis may provide further insight regarding the potential impact of specific determinants. Speculation regarding reasons for the apparent undertreatment of osteoporosis has focused on physician oversight. However, several factors may play a role in the choice to initiate treatment. A recent study indicated that some physicians have concerns about the proven efficacy of osteoporosis treatment among older populations and may be exercising judgment with respect to minimizing polypharmacy in this population [ 21 ]. Further, physicians are not the only factor involved in decisions to initiate, and/or continue with therapy. In a study conducted by McKercher et al., physicians reported that the choice to initiate osteoporosis treatment was also dependent upon acceptance of the therapy by patients and/or family members [ 21 ]. Interviews with patients following fragility fracture also indicate that some patients choose not to use osteoporosis therapy [ 8 , 13 ] (e.g. concerns about side effects). The higher rate of nonadherence with medications among those not receiving therapy in this study, also suggests that lack of therapy may be due to patient choices. However, the assessment of adherence by self-report, as used in this study, only provided a measure of adherence with overall drug regimens, not specific medications. Collection of more detailed information on adherence with specific osteoporosis therapies in future studies may clarify this association. Despite limitations, our findings highlight the growing concern that many patients with osteoporosis are not receiving appropriate therapeutic interventions. The lower treatment rate among subjects with fractures at sites other than the hip suggests that physicians may not be recognizing the probability of underlying osteoporosis. While the diagnosis of osteoporosis by BMD measurement has been widely publicized, recent clinical practice guidelines have emphasized the importance of a history of fragility fracture in the identification of patients with osteoporosis [ 3 ]. Our findings suggest that improving recognition of osteoporosis among older persons presenting with fractures may be an important educational objective for practicing physicians. Conclusions Four previous studies conducted in Canada have examined osteoporosis diagnosis and treatment interventions following fracture. Our data confirm low treatment rates among patients with fracture, and also indicate that even patients with a documented diagnosis may not be receiving therapy. The reasons for lack of treatment of osteoporosis are not yet clear. The reports of patient and physician concerns regarding side effects and polypharmacy warrant further investigation, and suggest that nonpharmacologic interventions may be more acceptable in certain patient populations. Considering the complexity of issues involved in decisions to initiate and continue with treatment, future studies focusing more on evaluation of physician and patient awareness of osteoporosis and factors influencing treatment decisions are needed. Evaluation of other patient-related factors, such as adherence and persistence with osteoporosis therapy, and their impact on health outcomes (e.g. fracture), are also relevant. Such studies may provide insight regarding specific interventions needed to reduce risk of morbidity and mortality associated with osteoporosis. Competing interests Shelly Vik and Colleen Maxwell have no competing interests. David Hanley's competing interests include consultancies with, honoraria for speaking from, or involvement in research with, the following companies or organizations: Amgen, Astra-Zeneca, Aventis, the Dairy Farmers of Canada, Eli Lilly, Merck, Novartis, NPS Pharmaceuticals, Pfizer, Procter and Gamble, Roche, and Wyeth. Authors' contributions Analysis of data, interpretation and the original draft were completed by Shelly Vik. Colleen Maxwell and David Hanley contributed to conception and interpretation, and provided critical evaluation of clinical and methodological content, and subsequent revisions. Pre-publication history The pre-publication history for this paper can be accessed here:	/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC555463.xml
497050	Echocardiographic assessment and percutaneous closure of multiple atrial septal defects	Atrial septal defect closure is now routinely performed using a percutaneous approach under echocardiographic guidance. Centrally located, secundum defects are ideal for device closure but there is considerable morphological variation in size and location of the defects. A small proportion of atrial septal defects may have multiple fenestrations and these are often considered unsuitable for device closure. We report three cases of multiple atrial septal defects successfully closed with two Amplatzer septal occluders.	Introduction Atrial septal defect (ASD) closure is now commonly performed using a transcatheter, percutaneous approach and with the Amplatzer septal occluder, large defects can be safely closed [ 1 , 2 ]. Device deployment requires a rim of atrial septal tissue surrounding the defect to allow effective capture of the septum by the occluder. The rim of tissue is also important to separate the septal occluder from important structures including the inferior vena cava, coronary sinus and the atrioventricular valves. The majority of patients require a single device for closure of the ASD but a small proportion of patients may have more than one defect in the atrial septum. This can be difficult to diagnose using transthoracic echocardiography (TTE) as abnormal colour flow obscures the origins of the shunt, particularly if the second defect is situated inferiorly. We report three cases of patients referred for ASD closures that were found to have multiple ASDs and the techniques used to close these defects. Case 1 A 34-year old woman was referred for consideration of percutaneous ASD closure. The ASD had been diagnosed when the patient was 12 years old and TTE had suggested that the right ventricle was dilating. At cardiac catheterisation there were mildly elevated right ventricular systolic pressures and a pulmonary to systemic flow ratio of over two. The secundum ASD was estimated to be 15 mm wide using TTE with aneurysmal formation of the interatrial septum. The patient was admitted for percutaneous ASD closure and underwent uncomplicated placement of a 17 mm Amplatzer septal occluder. Transesophageal echocardiography (TEE) during the procedure revealed the presence of a second ASD near the inferior vena cava and a small post-procedure shunt. The septal occluder did not completely cover both defects. Equivalent chest x-ray radiation dose (assuming a single posteroanterior projection chest x-ray is eight centi-Gray/cm 2 ) was 400. Repeat TTE continued to demonstrate left to right shunting and the patient was readmitted for a further device closure six months later. At cardiac catheterisation, the second defect was identified low in the secundum septum between the fossa ovalis and the mouth of the coronary sinus. The defect was successfully closed with a 9 mm Amplatzer septal occluder with no evidence of obstructed flow in either the coronary sinus or the inferior vena cava. Equivalent chest x-ray radiation dose for the second procedure was 187. The patient remained well with no evidence of residual shunt six months following the procedure. Case 2 A 31-year old woman was found to have a secundum ASD during investigations for breathlessness. The defect was estimated to be 10 mm wide on TTE with evidence of right atrial and right ventricular dilatation. Left atrial size was normal. She was referred was further investigation and treatment. Cardiac catheterisation demonstrated a pulmonary to systemic flow shunt of four to one. Peri-procedure TEE revealed that there were two defects; one measuring 24 mm located inferiorly between the inferior vena cava and coronary sinus os with the second defect situated superiorly and measuring 30 mm. The inferior defect was closed using a 24 mm Amplatzer septal occluder. Equivalent chest x-ray radiation dose was 403. The patient was discharged the following day and readmitted three months later for closure of the superior defect. This was performed using a 30 mm Amplatzer septal occluder. The procedure was technical difficult as after deployment of the left sided disc, the device initially lay obliquely in the defect. The final position was satisfactory with no evidence of intra-cardiac shunting (figure 1 ). There was no interference between the two devices and mitral valve function remained normal. Equivalent chest x-ray radiation dose for the second procedure was 727. The patient remained well and at six month follow-up there was no evidence of residual shunt. Figure 1 Stored fluoroscopy following placement of the two Amplatzer septal occluders (ASOs). TEE – Transesophageal echocardiography probe. Case 3 A 30-year old woman was investigated for palpitations and a secundum ASD of approximately 20 mm was diagnosed. TEE suggested that the atrial septum was fenestrated with a small inferior rim of tissue and she was referred for device closure. At cardiac catheterisation it became clear that there were two principal defects, one in the fossa ovalis and the other situated inferior and posterior between the fossa ovalis and the coronary sinus. The superior hole was closed with a 16 mm Amplatzer septal occluder but failed to cover the inferior defect. A 15 mm Amplatzer septal occluder device was subsequently placed successfully across the inferior defect with a stable position (figure 2 ). Equivalent chest x-ray radiation dose for the procedure was 124. Subsequent follow-up revealed a small left to right shunt between the two septal occluders but further intervention was not considered necessary. Figure 2 Transesophageal echocardiography four-chamber image following deployment of the two Amplatzer septal occluders (ASO). LA – left atrium, LV – left ventricle, RA – right atrium, RV – right ventricle. Discussion There is considerable morphological variation of secundum-type ASDs. Podnar et al reported the echocardiographic findings of 190 patients with isolated secundum ASDs referred for device closure [ 3 ]. Twenty four per cent had centrally placed defects but the remaining 144 patients had morphological variations. A deficient superior anterior rim was seen in 42%, a deficient inferior posterior rim in 10%, perforated aneurysm of the interatrial septum was seen in 7.9% and 7.3% of patients had multiple septal defects. Experience of multiple ASDs closure using more than one Amplatzer septal occluder remains limited [ 4 , 5 ]. In the worldwide report of use of the Amplatzer septal occluder, 3460 patients received a single device but only 45 patients received two devices for multiple ASDs [ 1 ]. Cao et al reported a series of 22 patients who had two septal occluders implanted simultaneously for multiple ASDs [ 6 ]. Closure rate was 97.7% with one device embolisation. In closely positioned multiple defects the septal occluder should be implanted in the largest defect aiming to cover any smaller defects but in widely separated defects more than one device is required. Echocardiographic studies have suggested that patients with multiple ASDs should have a rim of tissue of more than seven millimetres between defects to allow the deployment of two septal occluders [ 6 ]. Continuous echocardiographic monitoring is required for device positioning. When TEE is used, patients usually require a general anaesthetic due to the prolonged oesophageal intubation. The development of intracardiac echocardiography now provides an alternative to TEE for device closure. Benefits include more detailed imaging, a reduced need for general anaesthesia, and reduced radiation exposure [ 7 ]. In particular, use of intracardiac echocardiography allows clearer visualisation of the inferior atrial septum. Three-dimensional echocardiography may allow more detailed assessment of multiple ASD anatomy and septal occluder positioning. One question that remains unclear is whether multiple septal occluders should be deployed simultaneously or implanted as staged procedures. Serious complications during single septal occluder implantation is a rare occurrence (less than 0.3% of cases) but it likely that simultaneous deployment will increase the procedure risk [ 1 , 6 ]. When implanting large devices (greater than 20 mm) with little septal separation it is favourable to deploy the larger device and bring the patient back for a further procedure once the device has stabilised. Conclusions There is considerable variation in atrial septal defect anatomy. A small proportion of patients with an ASD have more than one defect and these can be closed using conventional septal occluders under transoesophageal echocardiography guidance. The use of intracardiac echocardiography should allow more accurate device positioning, particularly defects located low in the atrial septum.	/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC497050.xml
543572	Application of microarray technology in pulmonary diseases	Microarrays are a powerful tool that have multiple applications both in clinical and cell biology arenas of common lung diseases. To exemplify how this tool can be useful, in this review, we will provide an overview of the application of microarray technology in research relevant to common lung diseases and present some of the future perspectives.	Introduction Microarray technology is rapidly becoming a standard technology used in research laboratories all across the world. Since its first application in the mid 1990s [ 1 ] microarray technology has been successfully applied to almost every aspect of biomedical research [ 2 - 7 ] with over 250 papers in respiratory research alone. Research conducted the last ten years has elevated the status of microarray technology from poorly understood and doubtfully applied in the fields of medicine to one that requires attention when the examination of clusters of genes in a single experiment is considered. Far more progress has been made toward an understanding of the pivotal role of microarrays in respiratory research by providing the scientists well-established knowledge concerning numerous genes that can be used as potential drug targets, mediators and inflammatory molecules with important cellular functions, evidence that captured the interest of both clinicians and researchers and caused a consecutive year by year rise of the applications of microarrays in experiments designed to study pulmonary diseases. Thus microarray technology since its first application [ 8 ] in the field of respiratory medicine has already been used the past five years in almost every aspect of respiratory research with an increased rate of application resulting to an overall of approximately 250 published papers until now (Figures 1 , 2 ). Though the majority of experiments using microarray platforms was designed to study lung cancer (Figure 2 ) we excluded from this review this data, because considering the number of published papers, the enormous data derived from these experiments could compile a separate review article on its own. The scope of this review was based on the fact that although there are numerous original published papers using this pioneering method, the number of review articles summarizing the importance of microarrays in the research field relevant with pulmonary diseases still remains inadequately small. Figure 1 Diagram showing the number of published papers using microarray technology in respiratory research the last ten years since 1995 when microarrays were first applied in clinical medicine. Figure 2 Diagram showing the number of published papers in research relevant to common lung diseases such as asthma, COPD, IPF, ALI/PE, SARC, SCL, lung cancer, the last ten years since 1995 when microarrays were first applied in clinical medicine. Applications of microarrays in medicine DNA microarrays, microscopic arrays of large sets of DNA sequences immobilized on solid substrates, are valuable tools in areas of research that require the identification or quantitation of many specific DNA sequences in complex nucleic acid samples [ 8 ]. They are ordered samples of DNA and each sample represents a particular gene. These arrays can then be assayed for changes in the expression patterns of the representative genes after different treatments, different conditions or tissue sources. There are numerous ways to measure gene expression including northern blotting, differential display, serial analysis of gene expression and dot-blot analysis. The problem with all these techniques is that they are unsuitable for the parallel testing of multiple genes' expression. Microarrays, based on Southern's method of nucleotide hybridization, contain multiple DNA sequences (probes) spotted or synthetized on a relatively small surface. This feature of microarrays allows the simultaneous monitoring of the expression of thousands of genes, thus providing a functional aspect to sequence information, in a given sample [ 9 ]. Currently, genomic microarrays are used in medicine for the following purposes: [ 10 - 13 ] 1. Determination of transcriptional programs of cells for a given cellular function (e.g., cell function, cell differentiation, etc.) or when they are exposed to certain conditions leading to activation, inhibition or apoptosis. 2. Compare and contrast transcriptional programs to aid diagnosis of diseases, predict therapeutic response and provide class discovery and sub-classification of diseases. 3. Identification of genome-wide binding sites for transcriptional factors that regulate the transcription of genes. 4. Prediction of gene function. 5. Identification of new therapeutic targets (target identification, target validation, and drug toxicity). 6. Development of public databases that will help us understand the functioning of complex biological systems. 7. Genetics of gene expression: Although this is a relatively new study field, it is advancing rapidly with major implications in complex clinical traits by the identification of promising candidate genes. Thus, we briefly review the current implementations of this novel approach highlighting its necessity in the research field. Treating mRNA transcript abundances as quantitative traits and mapping gene expression quantitative trait loci for these traits has been pursued in gene-specific ways. Unlike classical quantitative traits, the genetic linkages associated with transcript abundance permits a more precise look at cellular biochemical processes. Schadt et al. [ 14 ] described comprehensive genetic screens of three specific transcriptomes by considering gene expression values as quantitative traits. Authors treated the gene expression levels derived by a microarray analysis in mice liver tissues as quantitative traits in a standard linkage analysis using evenly spaced autosomal markers. Interestingly they found that a substantial portion of these genes had at least one significant gene expression quantitative trait locus (eQTL) depending on the LOD (log odds ratios) scores. Since transcript abundances are increasingly used as surrogates for clinical traits, knowledge about their genetic control can help dissect the genetics of complex traits. In the same study investigators revealed the importance of LOD scores to differentiate whether the expression levels of the genes under study is regulated by variations within the gene itself (cis) or at a separate locus (trans). They found that eQTL with LOD scores are cis acting (gene affects transcription of the gene itself) in most cases, whereas moderately significant eQTL are trans acting (genes acting on the transcription of other genes). Furthermore this study undertook an investigation on how the heritability of gene expression can be studied within and between families and demonstrated that a significant portion of differentially expressed genes derived from reference families had a detectable genetic component. The latter finding suggests that this group of genes may serve as novel therapeutic targets for complex human diseases, given that their degree of genetic control was so readily identifiable in a small number of families. Microarray technologies DNA microarrays are used to estimate the levels of mRNA in the cell. The process can be described in three steps: 1) Array construction : Currently, there are two widely used microarray technologies: • In situ synthetized oligonucleotide (20–25 mers) microarrays-mainly oligonucleotides synthetized by photolithography or ink-jet technology on a glass surface. • Spotted, in glass or nylon membrane matrices, microarrays-mostly created by robotic printing of pre-prepared cDNAs or oligonucleotides (polymerase chain reaction-PCR-generated products from cDNA libraries or clone collections) [ 9 , 15 ]. 2) Sample preparation and array hybridization : The next step in the microarray experiment is to prepare the material that will be hybridized with the microarray. Gene expression is measured by the amount of mRNA therefore it must be extracted from the sample cells or tissues. For high-density microarrays one has to convert mRNA into cRNA, whereas for spotted arrays one can use mRNA, cDNA, or cRNA. The RNA needs to be labeled so that the detecting machinery can measure the quantity of RNA present. In oligonucleotide microarrays mRNA is extracted from experimental samples and is labeled with a fluorescent oligonucleotide (biotin). The biotin labeled cRNA and each sample is hybridized to a separate array, the array is scanned and absolute expression levels are obtained for each probe by using a dedicated laser scanner. In contrast, in spotted microarrays, mRNA is extracted from a sample and a control and one is labeled with cy-5 (red fluorescent dye) and the other with cy-3 (green fluorescent dye). Expression values are based on the competitive hybridization of the two samples being directly compared on a single array. Conventionally, in radioactive nylon membrane arrays RNA probes are labeled with P 33 or P 32 dCTP during a reverse transcription reaction. The great advantage of nylon microarrays is that they require relatively small amounts of PCR products because radioactive targets have a higher intrinsic detectability whereas in glass arrays the quality of RNA is not as critical as it is in nylon arrays. Therefore nylon arrays are mainly used when sample material is scarce or a small number of genes need to be assayed [ 16 , 17 ]. 3) Image analysis and data acquisition: The laser causes excitation of fluorescently or radioactively labeled cDNA probes. Only the spots representing mRNAs in the sample give emission signals. The emission is measured using a scanning confocal laser microscope for fluorescently labeled arrays or a flat-bed scanner for radioactive nylon arrays and finally data is analyzed by appropriate software. In spotted microarrays using fluorescent probes if particular mRNA from the sample is in abundance, the spot with a complementary probe will be red; if the concentration of the particular mRNA is higher in the control, the spot will be green. If both samples contain the same amount of a given mRNA, the spot will be yellow (Figure 3 ). In nylon membrane microarrays using radioactive probes acquisition of phosphor-representations of radioactive hybridizations is performed with a high resolution digital autoradiography system displaying in real time the quantitative image of radioisotopes deposited on biological samples [ 18 ]. In high density oligonucleotide microarrays the absence of the fluorescence of the specific spots means that complementary mRNA is not present in the sample. If the fluorescence is present, the intensity of the signal is a measure of the level of particular mRNAs in the examined cell population [ 11 - 19 ]. Figure 3 Image from laser scanning confocal microscope of a DNA microarray slide. mRNA, is extracted from a sample and a control and after its transcription into more stable cDNA, one is labeled with cy-5 (red fluorescent dye) and the other with cy-3 (green fluorescent dye). The two cDNA populations are allowed to hybridize to the same microarray slide. If particular mRNA from the sample is in abundance, the spot with a complementary probe will be red (induction of gene expression in sample condition); if the concentration of the particular mRNA is higher in the control, the spot will be green (induction of gene expression in control condition). If both samples contain the same amount of a given mRNA the spot will be yellow (equal gene expression in both conditions). (Adapted with permission of Dr Karameris Andreas.) Application of microarrays in pulmonary diseases 1. Microarrays in idiopathic pulmonary fibrosis (Table 1 ) Table 1 Studies utilizing microarray technology to analyze IPF Investigator Microarray type Species/Sample size Summary/Key findings Normalization procedure Number of genes Type of tissue Replications per data point Zuo et al. 22 Oligonucleotide 8.400 genes 5 patients with IPF Lung tissue specimens Gene expression analysis reveals matrilysin as a key regulator of PF in mice and humans. Gene expression levels normalized by a scaling factor multiplied to the average of differences of probe pairs (matched-mismatched) / 2 replicates Kaminski et al. 23 Oligonucleotide 6.000 genes 30 mice Lung tissue specimens Global analysis of gene expression in PF reveals distinct programs regulating lung inflammation and fibrosis. Mean hybridization intensities of all probe sets on each array were scaled to an arbitrary, fixed level/4 replicates Katsuma et al. 24 cDNA 4.224 genes 22 mice Lung tissue samples Molecular monitoring of bleomycin-induced pulmonary fibrosis by cDNA microarray-based gene expression profiling. Quantified signal intensities were converted by logarithms of base two 4 replicates Chambers et al. 25 Oligonucleotide 6.000 genes Human lung fibroblasts Global expression profiling of fibroblast responses to transforming growth factor-beta1 reveals the induction of ID1. Gene expression levels normalized by a scaling factor multiplied to the average of differences of probe pairs (matched-mismatched)/ 2 replicates Liu et al. 26 cDNA 10.000 genes 12 rats Lung tissue specimens FIZZ1 stimulation of myofibroblast differentiation. Average median ratios Cy3/Cy5 values normalized to 1.0/ 4 replicates Abbreviations: ID: Inhibitor of Differentiation, IPF: Idiopathic Pulmonary Fibrosis, PF: Pulmonary Fibrosis Idiopathic pulmonary fibrosis (IPF) is a refractory and lethal interstitial lung disease characterized by fibroblast proliferation, extracellular matrix (ECM) deposition and progressive lung scarring. The incidence of IPF is estimated at 15–40 cases per 100.000 per year, and the mean survival from the time of diagnosis is 3–5 yr regardless of treatment [ 20 , 21 ]. The etiology of IPF has remained elusive and the molecular mechanisms are poorly understood. To elucidate the molecular mechanisms that lead to end-stage human pulmonary fibrosis Zuo et al. [ 22 ] analyzed lung biopsy samples from five patients with clinically, radiologically and histologically proven pulmonary fibrosis (usual interstitial pneumonia-UIP) and compared to samples from three resected lungs with normal histologic findings, by using oligonucleotide microarrays. Using a combined set of scoring systems they determined that matrilysin (matrix metalloproteinase-MMP-7), a metalloprotease not previously associated with pulmonary fibrosis, was the most informative increased gene in their data set. Immunochemistry demonstrated increased expression of matrilysin protein in fibrotic lungs derived from different patients. Furthermore, in a separate set of experiments matrilysin knockout mice were dramatically protected from pulmonary fibrosis in response to intratracheal bleomycin. Their results identify matrilysin as a mediator of pulmonary fibrosis and a potential therapeutic target. Nevertheless potential limitations of the study include the small sample size and the relative inability of microarray analysis using whole lung homogenates to assess the exact cells that were overexpressing the informative genes. The application of analytic microarray approaches using gene expression signatures of specific cell types coupled with advanced data mining computational tools will ameliorate this hardship. In another study Kaminski et al. [ 23 ] used oligonucleotide microarrays to analyze the gene expression programs that underlie pulmonary fibrosis in response to bleomycin, in two strains of susceptible mice. Microarray analysis performed by different investigators and at different time points demonstrated a considerable overlap between genes induced by bleomycin in these two distinct strains of mice. Differential gene expression in response to bleomycin included upregulation of genes known to be associated with bleomycin-induced lung injury and fibrosis such as transforming growth factor-β1 (TGF-β1), as well as genes not previously associated with the disease. Confirmational studies performed and further verified a portion of the microarray data. Surprising insights were derived from comparing gene expression patterns in response to bleomycin of mice homozygous for a null mutation of the integrin β6 subunit gene (β6 -/- ), thus protected from pulmonary fibrosis, and wild type mice. Interestingly a simple hierarchical cluster analysis identified most of the known TGF-β1 inducible genes preferentially induced in wild type mice. The latter finding provides support for the hypothesis that β6 knockout mice are protected from pulmonary fibrosis as a consequence of failure to activate TGF-β. The great importance of this study results from the identification and the global availability of several genes that are likely to be directly relevant to the fibrotic process. However the inability of microarray technology to detect genes that are not included in the array, identify critical proteins that participate in biological responses and ascribe changes in gene expression in specific cellular types limit the scientific rigidity of the data derived and highlights the necessity for combined application of novel approaches. Furthermore driven by the same perspective idea to investigate the gene expression pattern in bleomycin-induced pulmonary fibrosis, Katsuma et al. [ 24 ] constructed a lung chip derived from a normalized lung cDNA. They performed large-scale analyses of gene expression and illuminated a time-dependent change in the expression profile of genes related to the inflammatory and fibrotic responses in this model of pulmonary fibrosis, similar to that observed by Kaminski et al [ 18 ]. Using cluster analysis they classified genes into groups based on a time-dependent gene expression. Interestingly this profile was well correlated with observed histopathological changes and data was confirmed with real time-(RT)-PCR methods. Nevertheless apparent inconsistencies with the gene expression pattern revealed by Kaminski et al. [ 23 ] highlight the inability of microarray approach in distinguishing changes in transcriptional regulation from changes in cellular composition of the organ being studied. Thus, it is likely these discrepancies between findings of the two studies to represent differences in cellular composition, rather than differences in transcriptional regulation. One of the most informative studies scrutinizing the global gene expression profile of fibroblasts in response to one of their most potent activators, TGF-β1, has been published by Chambers et al. [ 25 ] Gene expression analysis of human lung fibroblasts treated with TGF-β1 has led investigators to uncover novel TGF-β1-inducible genes including genes encoding inhibitors of differentiation (ID) as well as genes that are usually expressed by highly differentiated smooth muscle cells. The induction of these genes was further confirmed at the mRNA level (Northern blot analysis) and the protein level (Western blot analysis) for primary cultures of adult lung fibroblasts. The potential relevance of these observations in vivo was established in a separate set of confirmational experiments in rats where it was revealed an overexpression of the ID by myofibroblasts within the fibrotic regions. These novel suggestions have major impact on our understanding of the crucial role of TGF-β1 as a fibroblast differentiation factor in response to fibrogenic stimuli. Nonetheless the present study does not determine the precise role of ID in regulating fibroblast responses to TGF-β1. To do so this study should be coupled with independent methods using ID blocking strategies. Moreover the use of incomplete arrays that detect only the included genes, the variability of cellular composition of tissues studied even from the same organ and the inability of this technology to distinguish changes in cellular composition from transcriptional changes pose major limitations to the global application of these results. Furthermore these observations offer plausible explanations for the lack of similar effects of TGF-β1 in previous studies. Moreover, bleomycin-induced pulmonary fibrosis rat model was extensively studied by Liu et al. [ 26 ] with the use of rat cDNA microarray platforms, in an attempt to highlight genes that may be involved in fibrosis. Interestingly a novel and unexpected finding of this microarray analysis was the identification of FIZZ1 as the most highly and prominently induced gene in bleomycin-treated lungs, evidence consistent with the RT-PCR results. Further analysis of its protein product, demonstrated a unique pattern of localization primarily to alveolar epithelial cells (AECs) derived from bleomycin-injured lungs. To illuminate the exact role of FIZZ1 in inflammation and fibrosis, the effects of co-culturing FIZZ1-expressing AECs on fibroblasts were examined. These analyses demonstrated the significant higher stimulation of normal lung fibroblasts, by high FIZZ1-expressing-AECs, as compared to that observed by control AECs. The major contribution of this new molecule in the differentiation of fibroblasts to myofibroblasts was suggested and verified by its transfection into normal lung fibroblasts which provoked their stimulation independently of TGF-β activation. These preliminary results suggest that this technology could identify unexpected molecular participants in IPF and might help in the development of novel targets for improved treatment. The method may also allow molecular fingerprinting that could improve the ability to identify subclassifications of pulmonary fibrosis that might be more informative than the current classification based primarily on histologic and radiographic patterns [ 27 ]. Nonetheless these studies characterized as "fishing expeditions" are limited by the inability of microarrays to detect the final expression product (protein), identify genes that are not included in the array and ascribe changes in gene expression in specific cellular types. However our view is that there is nothing wrong with a "fishing expedition" if what you are after is "fish", such as new genes involved in a pathway, potential drug targets or expression markers that can be used in predictive or diagnostic fashion. Hence, these observations are not to diminish their value for understanding basic biological processes and even for understanding, predicting and eventually treating human disease (Table 1 ). 2. Microarrays in asthma (Tables 2 and 3 ) Table 2 Studies utilizing cDNA microarray technology to study asthma Investigator Microarray type Species/Sample size Summary/Key findings Normalization procedure Number of genes Type of tissue Replications per data point Zou et al. 32 cDNA 40.000 elements 10 monkeys Lung tissue samples Microarray profile of differentially expressed genes in a monkey model of allergic asthma. Ratios of Cy5/Cy3 multiplied to the balance coefficient of the microarray / 3 replicates Brutsche et al. 33 cDNA 600 genes 40 subjects Mononuclear cells CAGE score for atopy and asthma. Absolute difference of the expression of CAGE scored genes 1 replicate Sayama et al. 38 cDNA 14.000 genes human umbilical cord mast cells Transcriptional response of human mast cells stimulated via the Fc (epsilon) RI and identification of mast cells as a source of IL-11. Array-specific normalization coefficient was calculated by centering in log base 2 space a dataset consisting of all elements with an I/D> 3-fold / 2 replicates Brutsche et al. 41 cDNA 600 genes 40 subjects Mononuclear cells Apoptosis signals in atopy and asthma measured with cDNA arrays G.I was normalized to the housekeeping G.I / 1 replicate Syed et al. 42 cDNA 12.228 genes Human CD4 + T cells CCR7 down-regulation in asthma Median G.I of each filter normalized any differences in cDNA probe activity between filters/ 1 replicate Banerjee et al. 43 cDNA 1.176 genes 18 mice Lung tissue samples Gene expression profiling in inflammatory airway disease associated with elevated adenosine G.I was normalized to the housekeeping G.I / 2 replicates Abbreviations: CAGE: Composite atopy gene expression, CCR7: Chemokine receptor 7, G.I: Gene Intensity, I/D: Intensity/Background ratio, Th: T helper, RI: Immunoglobulin receptor Table 3 Studies utilizing oligonucleotide microarray technology to study asthma Investigator Microarray type Species/Sample size Summary/Key findings Normalization procedure Number of genes Type of tissue Replications per data point Lee et al. 34 Oligonucleotide 6.500 genes Human airway cells IL-13 induces dramatically different transcriptional programs in three human airway cell types. Gene expression levels normalized by a scaling factor multiplied to the average of differences of probe pairs (matched-mismatched)/ 1 replicate Temple et al. 35 Oligonucleotide 6800 genes Human eosinophils Microarray analysis of eosinophils reveals a number of candidate survival and apoptosis genes. Geometric mean of the scaling (standard experiment) factor served as normalization factor/ 2 replicates Hakonarson et al. 36 Oligonucleotide 5.000 genes Rabbit and human ASM Association between IL-1beta/TNF-alpha-induced glucocorticoid-sensitive changes in multiple gene expression and altered responsiveness in ASM. Gene expression levels normalized by a scaling factor multiplied to the average of differences of probe pairs (matched-mismatched) / 2 replicates Laprise et al. 37 Oligonucleotide 12.000 probe sets 8 subjects Lung tissue samples Functional classes of bronchial mucosa genes that are differentially expressed in asthma. Mean hybridization intensities of all probe sets on each array were scaled to a fixed level/ 2 replicates Nakajima et al. 39 Oligonucleotide 12.000 genes Human MCs and eosinophils Gene expression screening of human mast cells and eosinophils using high-density oligonucleotide probe arrays: abundant expression of MBP in MCs Mean hybridization intensities of all probe sets on each array were scaled to an arbitrary, fixed level / 1 replicate Abbreviations: ASM: Airway Smooth Muscle cells, MBP: Major Basic Protein, MCs: Mast cells, TNF: Tumor Necrosis Factor Asthma is one of the most serious allergic diseases associated with both genetic and environmental factors such as allergens, respiratory tract infections, and atmospheric pollutants. Most asthma is associated with atopy, a predisposition to generate immunoglobulin (Ig)-E against environmental allergens [ 28 ]. However, only a proportion of atopic individuals develop lower airways symptoms consistent with an asthmatic phenotype. It is therefore tempting to speculate that the development of asthma requires combined inheritance of genes which alter the immune cell response to the environment, and at the same time, render the airways structural and neural regulation susceptible to injury caused by inflammation. Several asthma/atopy associated genes have been identified from linkage and association studies within families and revealed that there are multiple chromosomal regions, containing potential candidate genes, associated with various asthma phenotypes [ 29 - 31 ]. Microarray technology offers a new opportunity to gain insight into global gene expression profiles in asthma, leading to the identification of asthma associated genes. Several experimental models have been used for this purpose although no animal disease model is identical to human disease. Zou et al.[ 32 ] were the first attempted to produce an allergen-induced gene expression profile in the lung of a non-human primate using genomics tools such as microarrays and real time-(RT)-PCR in an independent way. Microarray data generated from this study and validated by RT-PCR using same lung samples, revealed a differential gene expression pattern between control and challenged animals. Furthermore investigators established that genes identified by microarray technology represented genes truly regulated by inhalation antigen challenge. This was done by determining that the regulated expression levels identified by microarray assay from a single animal were confirmed by RT-PCR studies using multiple similarly treated animals. Potential limitations of this study include the time-limited gene expression profile tested which may not reflect the chronic aspect of asthma and the absence of evidence that the antigens used would produce the same allergic reaction in humans. Brutsche et al. [ 33 ] designed an array based composite atopy gene expression (CAGE) score to evaluate the diagnosis of atopy and asthma and assess disease activity in order to guide therapeutic decisions. The CAGE score was determined by using 10 genes dysregulated atopic individuals according to a specific algorithm. The application of this score in a group of asthmatic patients revealed that this approach had a better sensitivity and specificity than total IgE in differentiating atopic from non-atopic subjects. Correlation between CAGE score and total IgE was found, and there was a trend for correlation with asthma severity. It is noteworthy that the CAGE score was able to quantify phenotype-specific alteration in gene expression of atopic individuals. Perspectively the CAGE score can be further improved through a better reproducibility of microarray systems compared with the filter arrays and the possibility of a better selection of genes. Therefore it may be used as a prognostic and diagnostic tool or to monitor the effects and side-effects of asthmatic therapy in the not distant future. Several morphologic changes in the airways of patients with asthma have been attributed to the Th-2 produced cytokines such as IL-13 and IL-5. However the molecular mechanisms underlying the contributions of these cytokines to asthma remain largely unknown. Towards this direction Lee et al. [ 34 ] applied oligonucleotide microarray technology in primary cultures of three human airway cell types (epithelial, smooth muscle cells and lung fibroblasts) to elucidate the effects of IL-13 in these cell types. Interestingly, the results of this study demonstrated that despite initiation of an identical signaling pathway (STAT6), IL-13 induced highly distinct transcriptional programs in each of the three cell types suggesting a coordinate and distinct contribution to asthma pathogenesis by each of the cell types examined. Although the quality of the genechip analysis was estimated and validated by RT-PCR methods applied in a small number of selective genes, however there are important limitations in this study including the possible differences between transcriptional responses and gene expression profile of a cell type in vivo and in vitro . One of the greatest disadvantages of microarrays and at the same time challenges for most of the investigators is the objective difficulty dealing with the results of the experiments resulting from the large quantities of information. Currently, the hurdle faced is the routine interpretation of this information to identify among thousands of dysregulated genes, those who are informative, causal and specific to the phenotypic change of interest. Thus, Temple et al. [ 35 ] compared the results derived from the application of oligonucleotide microarray technology in eosinophils isolated from human peripheral blood before and after treatment with IL-5 and in an alternative cellular model, TF1.8 cells, whose survival was known to be dependent on IL-5. Comparison of these two models facilitated the identification of the genes that rule the apoptosis and survivability of eosinophils and demonstrated a small group of genes whose regulation was similarly coordinated in both systems. Authors combined different cellular models focused on the same experimental paradigm and looked for common changes. This approach helped the scientists to focus attention on a subset of genes most likely to be causal and relevant to the phenotypic change of interest and filter out non-specific gene expression change. Combination of this method with proteomics approaches and tissue distribution analysis can add another filter for genes of interest and generate data of sufficient scientific rigidity. Microarrays apart from their remarkable effectiveness in identifying novel gene expression patterns can also be used to clarify physiological mechanisms underlying the actions of numerous drugs, such as those applied in the management of patients with asthma. Several studies have utilized microarray technology to assess the gene expression profile of cells and tissues before and after treatment with commonly applied drugs such as corticosteroids. Two of them are reviewed here. Recently, Hakonarson et al. [ 36 ] in addition with the role of two pleiotropic cytokines, IL-1β and tumor necrosis factor (TNF)-α, in the pathophysiology of asthma, have reported the effectiveness of dexamethasone in the treatment of asthma with the use of oligonucleotide microarrays. The accumulation of the two cytokines in a medium where human airway smooth muscle-(ASM) – cells were cultured, elicited an overexpression of several proinflammatory genes known to regulate smooth muscle contractility and relaxation. The latter finding noted in four separate microarray experiments was consistent with the increased responsiveness of rabbit cytokine-treated tissues in acetylcholine. The administration of dexamethasone provoked a repression to the majority of the microarray-study derived genes as well as to the contractility of the cytokine-treated ASM. Collectively these findings suggest a crucial role of ASM in expressing a host of glucocorticoid-sensitive proinflammatory gene patterns that affect the structure and the function of the airways. However these observations were based on studies using rabbit and human ASM cells. Hence the issue of potential species differences warrants consideration. Further studies using samples derived from homogeneous species and validation techniques are necessary to streamline these observations. Microarray analysis performed by Laprise et al. [ 37 ] indicated a differential gene expression pattern in bronchial tissues from healthy and asthmatic individuals, a profile that included not only genes previously implicated in the pathogenesis of asthma but also new potential candidates. The remarkable ascertainment of this study, conducted with bronchial tissues which are known as a primary site for airway inflammation and remodeling, was that the expression of one third of the genes was partially or completely corrected by inhaled corticosteroid treatment. The latter evidence further illuminates the true impact of first line therapy offered to asthmatic patients. However application of this technology may be limited by the disease's spatial and temporal heterogeneity due to differences in cellular composition between asthmatic and control tissue. Ultimately the results obtained using microarrays need to be verified firstly with confirmational studies (RT-PCR and in situ hybridization) and secondly with separate experiments. Mast cells represent key cells in the initiation and progression of asthma, releasing several mediators of inflammation, such as certain cytokines and chemokines. The past few years several studies have been focused on the identification of new mast cell products through the gene expression analysis. In one of them published by Sayama et al. [ 38 ] application of cDNA microarrays in only two populations of stimulated human mast cells exhibited among other genes a significant upregulation of the gene encoding IL-11. The latter finding was further confirmed by a separate set of experiments where an increased secretion of IL-11 by activated human mast cells was noted. However further microarray analyses coupled with functional approaches and independent studies examining the potential role of IL-11 in the pathogenetic mechanisms of asthma as well as in the alterations of mast cell proliferation and survival, are required. Furthermore, Nakajima et al. [ 39 ] in their attempt to evaluate the significance of protein products present in mast cells applied oligonucleotide microarray technology in human mast cells derived from different sources and in eosinophils. The most impressive finding of this study was the abundant expression of major basic protein (MBP) among the transcripts for expected mast cells specific proteins such as tryptase. Authors also confirmed in independent studies using RT-PCR and flow cytometry that MBP was expressed at both the transcript and protein levels in various types of mast cells. While this result is really intriguing and opposes to the already known data which indicates the unique presence of MBP to eosinophils [ 40 ], it is incomplete and unable to determine the biologic significance of MBP present in mast cells. Microarrays using nylon membrane radioactive cDNAs have already been applied in the research field of asthma and much good work has been done with this technology. Brutsche et al. [ 41 ] applied nylon membrane cDNA microarray technology in blood samples derived from atopic asthmatic and nonasthmatic patients, and healthy control subjects to investigate the systemic activation of apoptosis pathways of inflammatory cells in lung tissue. They identified significantly altered expression of several apoptosis-related genes in atopy and asthma compared with the healthy subjects suggesting that these alterations could be due to genetic or environmental factors. Several verification experiments have been used to further validate a considerable amount of differentially expressed genes on mRNA level (RT-PCR), as well as protein (ELISA) and receptor level (double stained fluorescence methods). The profile of altered gene expression did not show a definite pattern that was suggestive of survival or apoptosis. Potential criticisms of this approach include the large amounts of data variability derived from the heterogeneity of the studied samples and the lack of proteomics analyses. Thus this study is unable to give any statement on the activity of the apoptotic pathways. To do so the data should be combined with the proteomics analysis of proteins involved in apoptosis. The latter will contribute to the characterization of protein patterns and will allow for the assessment of overall changes in the protein content associated with apoptosis. One of the first gene-profiling studies highlighting the potential role of chemokines and their receptors in the pathogenesis of asthma was conducted by Syed et al. [ 42 ] They used nylon membrane radioactive arrays compiled from mixed biological samples to determine the gene expression pattern of T-cells from patients with atopic and non-atopic asthma and found altered gene expression profile for CCR7 (chemokine receptor 7) between patients and controls, findings that were confirmed by RNA dot plot analysis. Data derived from this analysis is indicative of a possible role of this molecule in the progression of asthma. However the small number of patients recruited in this study and the lack of functional genomic analysis allow us to make only speculations on the exact role of this factor in the disease process. In another study Banerjee et al. [ 43 ] in their attempt to identify and characterize biological roles for adenosine-regulated genes applied radioactive cDNA microarray technology in lung specimens derived from normal and adenosine deaminase (ADA)-deficient mice. The results of this study profile the differential expression of a vast number of genes, that may be regulated by adenosine and hence play a pivotal role in modulating the underlying lung pathology. The reliability of the results derived from the microarray approach was also confirmed with gene specific RT-PCR analysis. Moreover authors used a separate set of experiments and demonstrated both with microarray analysis and protein localization that therapy with ADA in the deficient group of mice regulated expression of several genes modulating pulmonary inflammation and cell adhesion. The consistency of findings derived from the two experiments provides to microarray analysis a high degree of confidence. However critical limitations of this study originate from the disability of gene expression analysis to distinguish changes in transcriptional regulation from changes in cellular composition. Hence a more in-depth analysis is required to quantify the gene expression and establish a direct regulation of these genes by adenosine signaling. Accumulated evidence from these analyses revealed that microarray analysis can be a powerful tool for identifying mediators of allergic asthmatic disease through a genomic-based strategy using non-human primates and provide us a novel large scale of differentially expressed genes. Additionally, authors compared different cellular models sharing similar experimental paradigm to focus on the most likely informative genes and filter out the bystanders. The application of this approach further streamlined the pivotal role of microarrays in determining transcriptional responses of genes to several inflammatory cytokines and in identifying gene expression patterns and important mediators associated with the initiation and the progression of asthma. Moreover, microarrays coupled with separate set of experiments have provided the investigators with useful knowledge concerning the efficacy of the already applied drugs in the treatment of asthma and helped them to understand their anti-inflammatory role in terms of physiology and molecular biology. Nevertheless, most of the studies exhibited essential weaknesses generated by the heterogeneity of samples studied and compared and by the disability of microarrays to quantify gene expression and yield information about transcriptional responses and post-translational protein modifications. Therefore more in depth analysis of the microarray results is needed in combination with novel approaches that will help us focus on the specific genes and elucidate their role in the cellular function and the pathogenesis of asthma (Tables 2 , 3 ). 3. Microarrays in Chronic Obstructive Pulmonary Disease (Table 4 ) Table 4 Studies utilizing microarray technology to study COPD Investigator Microarray type Species/Sample size Summary/Key findings Normalization procedure Number of genes Type of tissue Replications per data point Koike et al. 45 cDNA 450 genes Rats AM cDNA microarray analysis of gene expression in rat alveolar macrophages in response to organic extract of diesel exhausts particles. G.I was normalized to the housekeeping G.I 2 replicates Yamanaka et al. 46 cDNA 18.432 genes Human AEC Gene expression profiles of human small airway epithelial cells treated with low doses of 14- and 16-membered macrolides. G.I was normalized to the housekeeping G.I 3 replicates Fuke et al. 47 cDNA 77 genes 30 patients Lung tissue specimens Chemokines in bronchiolar epithelium in the development of chronic obstructive pulmonary disease. Signal normalized to a given gene transcript 3 replicates Vuillemenot et al. 48 Oligonucleotide 12.000 genes 10 mice Lung tissue specimens Lymphoid tissue and emphysema in the lungs of transgenic mice inducibly expressing tumor necrosis factor-alpha. Signal normalized to internal control 2 replicates Hackett et al. 50 cDNA 4.600 genes 22 individuals AEC Variability of antioxidant-related gene expression in the airway epithelium of cigarette smokers. Mean hybridization intensities of all probe sets on each array were scaled to an arbitrary, fixed level / 2 replicates Morris et al. 52 Oligonucleotide 6.500 genes Mice Lung tissue samples Loss of integrin alpha (v) beta6-mediated TGF-beta activation causes MMP-12-dependent emphysema. Mean hybridization intensities of all probe sets on each array were scaled to an arbitrary, fixed level / 2 replicates Golpon et al. 53 Oligonucleotide 6.500 genes Human/mice Lung tissue samples HOX genes in human lung: altered expression in primary pulmonary hypertension and emphysema. Gene expression levels normalized by a scaling factor multiplied to the average of differences of probe pairs/ 3 replicates Abbreviations: AEC: Airway Epithelial Cells, AM: Alveolar Macrophages, COPD: Chronic Obstructive Pulmonary Disease, G.I: Gene Intensity, TGF-b: Transforming Growth Factor-beta, MMP: Metalloproteinase Chronic obstructive pulmonary disease (COPD) is a chronic disease characterized by progressive airflow obstruction, chronic cough and dyspnea in advanced stages, caused by smoking, environmental, and hereditary factors. It is associated with two clinical entities, chronic bronchitis and emphysema. In nowadays, the invention and application of microarray technology offers scientists the opportunity to gain a better understanding on the pathophysiology of COPD through the identification of novel gene expression patterns, leading to illumination of genes candidates for modern therapeutical approaches [ 44 ]. It is already known that chronic bronchitis can be induced by several types of environmental pollutants such as diesel exhaust particles (DEP). Though recently a microarray study has been published by Koike et al. [ 45 ] addressing the effect of such pollutants on the gene expression profiles of alveolar macrophages, however a complete analysis including the transcriptome and proteome, is needed to elucidate the toxic effect of air pollutants on pulmonary cells. Microarray approach is already being applied in respiratory clinical pharmacology with the identification of genes {Yamanaka et al. [ 46 ]} that can serve as potential molecular targets of common drugs applied in the management of patients with chronic bronchitis. However, studies being published in the field of respiratory pharmacogenomics lack of scientific rigidity primarily due to incomplete available arrays that will help scientists to determine much larger numbers of pharmacologically relevant genotypes. Far more progress should be made towards this direction. One of the major limitations in our attempt to elucidate the exact role of specific cell types in the pathogenesis of COPD is the compact anatomy of the lung which makes unraveling specific cell type gene expression changes difficult, requiring immunoelectron microscopy or laser capture microdissection. The first study to perform quantitative cell type-specific gene expression analysis using the pioneering technology of laser capture microdissection in human tissue samples coupled with RT-PCR and cDNA approach was recently published by Fuke et al. [ 47 ] Authors performed individual analyses and revealed a specific cell type upregulation of three inflammatory chemokines reportedly relevant to the pathogenesis of COPD emphasizing the pivotal role of these cells and their products in driving the inflammation. Although data was not fully confirmed by microarray analysis, however discrepancies between methods illustrate more the potential danger of depending solely on array approach rather than limit the scientific consistency of these results. Further research investigating the functional consequences of these changes is required. Several inflammatory cytokines have been implicated in the pathogenesis of emphysema, including TNFa, a molecule with versatile pathogenetic mechanisms. As a means to investigate some of them that culminate to lung-pathology, Vuillemenot et al. [ 48 ] applied oligonucleotide microarray technology coupled with independent studies (histologic and immunohistochemical analyses) in an experimental model they developed. Results derived both by microarray approach and independent studies revealed a direct correlation between TNFa and emphysema. However functional approaches should be applied in combination with gene expression analysis to shed further light in the mechanisms by which TNF promotes airspace enlargement. Despite the clear link of smoking to the risk for chronic bronchitis, only 15–20% of cigarette smokers develop COPD, [ 49 ] suggesting that there must be risk factors other than smoking that contribute to the susceptibility to this disease. To address this issue Hackett et al. [ 50 ] implemented microarray technology in human airway epithelial cells of smokers and non-smokers and demonstrated differential anti-oxidant related gene expression between the two groups of volunteers. One of the most intriguing aspects of this study is that individual assessment by hierarchical clustering of anti-oxidant related gene expression in response to smoking displayed a remarkable variability suggesting variability in the responses of different individuals to the chronic oxidant stress of smoking. However the extent of this variability may be explained by the nature of the microarray assay and the large amounts of data variation derived from these studies. Thus it is of high risk to speculate that these genes may serve as useful genetic markers in future epidemiologic studies determining susceptibility to smoking induced COPD. Further studies applying high-tech computational clustering tools coupled with independent validation tests are required. One of the most exciting aspects of microarrays is their use as tools for actively introducing serendipity to one's research [ 51 ]. In experiments designed to identify global transcriptional programs responsible for regulating lung inflammation and pulmonary fibrosis, as described previously, [ 23 ] microarray experiments were performed by Morris et al. [ 52 ] on lung tissue from wild-type mice and mice lacking a member of the integrin family (avβ6) known to be involved in activation of latent TGF-β. In addition to identifying distinct cluster of genes involved in these processes, these studies combined with RT-PCR validation tests and independent experiments led to the identification of novel pathways by which TGF-β can regulate emphysema through the upregulation of Mmp-12, the most highly induced gene in the lungs of β6 knockout mice. The role of Mmp-12 in the pathogenesis of emphysema was verified in an independent cohort where β6 knockout mice deficient in the expression of Mmp-12 displayed no alveolar enlargement. Although these results do not eliminate the possibility that other proteases may interact with Mmp-12 in the development of emphysema, however suggest that abnormalities in any of the steps in this pathway of TGF-β activation may contribute to genetic or acquired susceptibility to emphysema in humans. Presently, very few studies dealing with the role of HOX genes in the adult respiratory system are available in the literature. Golpon et al. [ 53 ] investigated the expression pattern of HOX genes, in fetal and diseased lung specimens (emphysema, primary pulmonary hypertension), by applying two microarray survey techniques and their analysis reflects one of the most detailed and informative studies in this field. They compared the HOX gene expression pattern in human and mouse lungs and found that HOX genes are selectively expressed in the human lung. This study also yielded an altered HOX-gene expression pattern among fetal, adult and lung specimens with emphysema and pulmonary hypertension, by identifying different types of HOX genes overexpressed in each of these conditions indicating differential HOX gene expression as a potential factor that contributes to the development of certain pulmonary diseases. Though the overall size of tissue samples studied was small data from this study comprises evidence with high degree of confidence, validated and confirmed both in an independent cohort (degenerate RT-PCR) and by alternative methods (quantitative RT-PCR and in situ hybridization). Possible limitations include small number of tissues studied, incomplete microarray survey techniques and minor discrepancies between the findings generated from validation studies. Collectively these results suggest that microarray analysis with its ability to highlight gene expression profiles on a large scale and coupled with progressive technologies and independently validated data has led researchers to shed further light into transcriptional programs regulating emphysema and to the identification of common mediators and molecular pathways involved in the pathogenesis of both COPD and pulmonary fibrosis, indicating novel targets for therapeutic interventions and useful genetic markers assessing susceptibility to COPD. Although limitations such as inconsistency between findings derived by microarray approach and independent studies, lack of functional changes assessment and significant data variability can be detectable in these studies, however evidence derived from these analyses is valuable and heavily informative (Table 4 ). 4. Microarrays in acute lung injury and pulmonary edema (Table 5 ) Table 5 Studies utilizing microarray technology to study ALI/PE Investigator Microarray type Species/Sample size Summary/Key findings Normalization procedure Number of genes Type of tissue Replications per data point McDowell et al. 56 cDNA 8.374 genes 6 mice Lung tissue samples Differential gene expression in the initiation and progression of nickel-induced ALI. Ratios of Cy5/Cy3 multiplied to the balance coefficient of each microarray / 5 replicates Olman et al. 58 cDNA 588 genes 36 patients Pulmonary edema Lung fibroblasts Microarray analysis indicates that pulmonary edema fluid from patients with ALI mediates inflammation, mitogen gene expression, and fibroblast proliferation through bioactive IL-1. G.I normalized to the housekeeping G.I 2 replicates Kupfner et al. 61 Oligonucleotide Mice Lung neutrophils Role of NF-κB in endotoxemia-induced alterations of lung neutrophil apoptosis. Gene expression levels normalized by a scaling factor multiplied to the average of differences of probe pairs / 3 replicates Cher et al. 62 Oligonucleotide 8.800 genes 21 rats Whole lungs Pulmonary inflammation and edema induced by Phospholipase A 2. Each gene was divided by the median of its values in all samples / 3 replicates Sabbadini et al. 63 Oligonucleotide 12.600 genes 14 rabbits Lung tissue samples Gene expression analysis in interstitial lung edema induced by saline infusion. Gene expression levels normalized by a scaling factor multiplied to the average of differences of probe pairs / 2 replicates Perkowski et al. 64 cDNA 8.374 genes 20 mice Lung tissue samples Gene expression profiling of the early pulmonary response to hyperoxia in mice. Difference between observed log-ratio and corresponding fitted ratio/ 5 replicates Ward et al. 65 cDNA 7.398 genes 6 rats Lung and other organ samples Molecular signatures of sepsis: multiorgan gene expression profiles of systemic inflammation. Gene expression levels normalized by a scaling factor multiplied to the average of differences of probe pairs / 4 replicates Abbreviations: ALI: Acute Lung Injury, G.I: Gene Intensity, PE: Pulmonary Edema, NF-κB: Nucleus Factor-κB Acute lung injury (ALI), a severe respiratory syndrome, develops in response to numerous insults. This syndrome that responds poorly in therapeutic interventions and has a poor prognosis has been associated with a myriad of mediators including cytokines, reactive oxygen and nitrogen species, growth factors and proteolytic enzymes [ 54 , 55 ]. Despite extensive research since the initial description of ALI over 30 yr ago, questions remain about the basic pathophysiologic mechanisms that are critical to the diminished survival and their relationship to therapeutic strategies. McDowell et al. [ 56 ] in their attempt to determine the interactions between the great amount of factors that have been associated with the development of ALI, analyzed 8,374 murine cDNAs for temporal changes and functional relationships throughout the initiation and progression of ALI in mice exposed to particulate NiSO 4 . Novel interactions between factors (antioxidant genes) previously associated with ALI and factors (surfactant proteins) previously not associated with ALI emerged from the application of functional genomics during nickel-induced ALI. Data derived from this experiment and partially confirmed by Northern blot analysis and nuclease protection assays is valuable and consistent with the ongoing attempts to treat ALI with exogenous surfactant-associated proteins [ 57 ] in combination with antioxidant therapy and may determine new therapeutical interventions. This study reveals the great importance of functional genomics not simply to provide a catalogue of all the genes and information about their functions, but to help scientists to understand the possible interplay of components contributing to lung injury. Although the fibroproliferative response to lung injury occurs in high frequency in patients with ALI, the mechanisms of this response are largely unknown. One of the most meaningful and informative studies addressing this important issue was recently published by Olman et al [ 58 ]. Authors applied radioactively nylon membrane arrays in human lung fibroblasts exposed either to ALI or hydrostatic pulmonary edema fluids and revealed a potential mitogenic activity of IL-1β and its importance as a modulator of fibroblast proliferation. Data derived from this study of a considerable number of patients and replicated both in an independent cohort (use of IL-1 antagonist receptor) and by alternative laboratory methods (RT-PCR, Northern blot analysis) is of fundamental value and may provide scientists with several independent lines of evidence that IL-1 amplifies the inflammatory and fibroproliferative process through regulation of fibroblast mitogenesis and gene expression. A major limitation mentioned in this study was the pooling of pulmonary edema samples due to their limited volume that did not allow authors to perform clinical-molecular correlations in an individual way. ALI is frequently associated with endotoxemia and is characterized by the accumulation in the lungs of large populations of neutrophils activated to produce proinflammatory mediators. Many studies had demonstrated a critical role of the endotoxemia-induced activation of NF-κB in neutrophils in the development of ALI [ 59 , 60 ]. One of the most important studies focused on the role of NF-κB activation in lung neutrophils apoptosis after endotoxemia was conducted by Kupfner et al. [ 61 ]. Though gene expression analysis revealed a significant role of NF-κB as a modulator of neutrophil apoptosis and data was confirmed by proteomics analysis there were major discrepancies between these findings and the results derived from individual experiments utilizing an inhibitor of nuclear translocation of NF-κB. The latter demonstrated no significant alterations in the percentage of the endotoxemia -induced apoptotic lung neutrophils in mice treated with the inhibitor evidence that reflects anti-apoptotic mechanisms not solely dependent on NF-κB. Although these findings can be informative indicating novel anti-apoptotic pathways for modern therapeutic approaches they should be re-evaluated in the context of new microarrays analyses of considerable sample size coupled with confident validation steps and independent experiments. To gain a more comprehensive understanding on the complex interplay between several inflammatory cytokines involved in the pathogenesis of pulmonary edema, several group of investigators {Cher et al. [ 62 ], Sabbadini et al.[ 63 ]} applied oligonucleotide microarray techniques in different experimental models of lung edema. Interestingly, these analyses revealed differential expression patterns for many inflammatory genes implicating some of them in the pathogenesis of pulmonary edema. Though these studies applied several confirmational tests (RT-PCR, Northern blot, Western blot) to validate results derived form microarray experiments, there are substantial weaknesses and concerns such as discrepancies between findings determined by these techniques that pose major limitations and limit their scientific rigidity. The study of Perkowski et al. [ 64 ] has been on of the most extensive and informative studies of the effect of 100% oxygen on the mouse lung. The authors used the cDNA microarray approach to evaluate the molecular profiling occurring during the early response to hyperoxia in mice. Among the vast amount of data derived from two different array sets they distinguished a cluster of genes of great interest (antioxidant enzymes, cell cycle progression regulators, endothelial cell and ECM genes) whose expression was substantially altered in response to hyperoxic stimuli. Authors were encouraged to observe that, of those genes where array data was compared with RT-PCR, changes occurred in same direction and were of similar magnitude. Further validation of the data with Northern-blot analysis for some of these genes reassuringly confirmed these alterations whereas for only one gene investigators performed functional activity assessment demonstrating similar notable findings. Though the results of this microarray study were double validated by standard molecular biology techniques there are still a number of caveats that should be kept in mind, including the post-transcriptional modifications that cannot be readily detected by gene expression changes and the large amounts of data variability that limits the careful analysis of every gene and highlights the necessity for further confirmatory tests and studies utilizing human lung samples. One of the first gene-profiling studies to address an important disease process, such as sepsis at a multisystem level, was that of Ward et al. [ 65 ] Using DNA microarray platforms, authors examined the sepsis-induced gene expression patterns at a multiorgan level, in mice. One of the most intriguing aspects of this study is the identification of genes (many of them not previously associated with sepsis response) that have a distinctive organ-specific expression profile as well as of genes with a relatively universal response to sepsis indicating interesting associations between organs. Validation of the data was performed by Northern blot analysis in only four selected genes and similar quantitative concordance between the two analyses was achieved. Although the microarray analysis provides an informative insight in the pathogenetic mechanisms of a complex disease process the lack of sufficient confirmatory tests applied in all studied genes and the absence of functional genomics that will help us understand the exact role of the newly characterized genes in the septic response pose major limitations in the study and illuminates the need for further characterization of the sepsis-induced gene expression profiles. In summary, these studies exhibit the crucial role of a novel molecular technology in discovering, through global analysis of gene expression, genes previously identified only by their DNA sequence. Although the array analysis provides in some studies a comprehensive overview of gene expression in the lung during ALI [ 56 , 58 , 61 , 64 ], and sepsis [ 65 ] and after hyperoxia [ 64 ], however there are numerous concerns arising from the large amounts of data variability, the lack of proteomics approaches in most of them and the controversial findings of microarray analysis and confirmational techniques. Unfortunately only three studies [ 58 , 61 , 64 ] used independent methodological criteria to validate a relatively small portion of their results, evidence that highlight the necessity for further more widespread evaluation of these findings. With the use of these approaches, more precise diagnosis and risk assessment of ALI based on expression profiles can be achievable in the next ten years, leading to more accurate determination of prognosis and new therapeutical interventions (Table 5 ). 5. Microarrays in sarcoidosis Sarcoidosis is a chronic systemic disorder characterized by the presence of non-caseating granulomas and accumulation of T-lymphocytes and macrophages in multiple organs [ 66 ]. The mechanisms leading to the persistent accumulation of inflammatory cells are not fully understood. Apoptosis, a dynamic process involved in the control of the "tissue load" of immune effecter cells at inflamed sites, limits inflammatory tissue injury and promotes resolution of inflammation [ 67 ]. Whether or not reduced apoptosis is involved in the pathogenesis of sarcoidosis is unclear. Rutherford et al. [ 68 ] in their attempt to shed further light on apoptosis signals in the peripheral blood of sarcoidosis patients with self limited and progressive disease in comparison with healthy controls used high-density probe arrays containing 12.626 genes. Though this study demonstrated significant differences in the expression of apoptosis-related genes in peripheral blood of patients with acute onset sarcoidosis compared to controls, ultimately did not manage to show a definite profile that was suggestive of survival or apoptosis. Although authors applied functional genomics a potential criticism of their approach is that they cannot give any statement on the activity of the apoptotic pathways. To do so the data should be combined with the proteomics analysis [ 69 ] of proteins involved in apoptosis. The latter will contribute to the characterization of protein patterns and will allow for the assessment of overall changes in the protein content associated with apoptosis. Collectively these findings not only reveal the importance of the microarray platforms in identifying gene expression patterns that give the scientists the opportunity to elucidate the pathophysiological processes of complex diseases, such as sarcoidosis but also illuminate some of their origin disadvantages. Future directions, challenges and limitations of microarray technology The last five years has seen the emergence of a novel technology applied in almost every aspect of respiratory research, a technology that has also great future perspectives and may provide scientists with numerous avenues of investigation that have clinical implications. Since nowadays, microarray technology has been successfully used for the identification of potential target genes for therapeutic intervention in IPF,[ 22 ] mechanistic studies in animal models of asthma [ 32 , 45 ] and IPF,[ 23 , 24 ] and helped the investigators to shed further light into the transcriptional programs involved in cytokine signaling [ 34 , 36 , 42 ] and apoptosis [ 35 , 41 , 61 ]. Furthermore, this technology increased our hopes in the field of diagnosis and clinical assessment of complex diseases such as asthma [ 33 ] and revealed modern approaches in therapeutic interventions in asthma [ 37 , 43 ] as well as COPD [ 46 ]. Finally, it provided scientists with useful information in their attempt to gain better understanding in molecular mechanisms regulating several pathological processes such as IPF, [ 22 - 26 ] asthma, [ 33 , 35 - 39 , 42 ] COPD, [ 45 - 48 , 50 , 52 ] lung fibrosis in acute lung injury [ 56 , 61 , 64 , 65 ] and pulmonary edema [ 58 , 63 , 64 ]. However, the feeling of excitement arising from the relative ease of producing the results of microarray experiments comes to contrast with the confusion arising from the objective difficulty of dealing with the results. Managing and mining the huge amount of data generated by microarray experiments still remains a major challenge and limitation for most investigators. Whether gene expression changes are considered primary vs. reactive for a given disease is a complicated issue, and one can only begin to judge that if the methods and approaches used to generate the data are of sufficient scientific rigidity. The diversity and scope of the data require the creation of multidisciplinary teams consisting of physicians, biologists and bioinformaticians (mathematicians, computational biologists and database managers) [ 10 ]. Thus, we can conclude, that these diverse experimental schemes pose diverse computational requirements, such as advanced data mining, clustering and analysis tools, including interpreting patterns of gene expression with self-organizing maps. Because of the statistical issues raised by microarray technology, it is necessary for any meaningful interpretation that the data is replicated using independent methodological criteria, preferably with separate samples rather than the tissue or RNA used to derive the original targets. A rapid high through-put method for confirmation of microarray data is quantitative RT-PCR. Alternatively, Northern blots or ribonuclease protection assays provide the benefit of direct quantification. So far some studies have started to adapt these approaches (Figure 4 ) but there are still limitations. Because a microarray analysis may reveal putative changes in the expression of tens or hundreds of genes, it is practically impossible to validate all of the data. However, it is incumbent upon investigators to evaluate a reasonable number of biased in their selection genes. Thus genomics and gene expression experiments are sometimes derided as "fishing expeditions". Hence it is necessary that these conventional techniques should be coupled with advanced data mining tools to help the scientists to face the greatest challenge, namely the extraction of biological meaning from microarray data and the prioritization of candidate genes for follow up [ 11 ]. Another challenge is to gain a holistic view of the human genome and biology, by applying genomic microarray in combination with the proteomic microarray to overpass the origin disadvantage of microarray technology that gives the users the view of inducible genes only. Sequence and gene expression analysis alone is insufficient to fully inform the investigator on the cell state and function. To the best of our knowledge only few of the studies utilizing microarray platforms in respiratory research has taken advantage of this approach and in a limited number of genes [ 69 ]. This combination would be crucial to better understand the functional aspects of disease and to bridge the long way between genotype and phenotype due to environment-gene interactions [ 12 , 70 ]. Additionally, it will be critical to develop improved methods for unbiased amplification of small RNA samples so that meaningful data can be obtained by applying microarrays on small tissue samples and pure cell populations, such as are samples obtained by microdissection of tissue sections [ 47 ]. This approach will solve the problem created by dramatic differences in cellular composition of affected and unaffected tissue and by spatial and temporal heterogeneity of disease that limits the optimal application of microarrays to the study of diseases [ 27 ]. Finally it is noteworthy to be mentioned that microarrays like other new diagnostic and research tools are highly cost-intensive. Considering the high costs of microarray based experiments it is important to say that this disadvantage will inevitably limit the speed with which they are introduced into clinical practice and restrict their application in university hospitals and other medical institutes [ 71 ]. Therefore it is crucial that academic centers and other specialized units should understand that joint ventures with biotechnological and pharmaceutical companies are critical to overlap all the financial obstacles so that microarray technology will be likely to reach most large hospitals with huge potential gain in clinically relevant information for individual patients and their diseases. Figure 4 Diagram showing the number of studies cited in this review article that validated the data derived from the microarray analysis either by confirmational studies (RT-PCR, Northern blot analysis, or both) or independent experiments (protein analysis, in situ hybridization, transgenic mice etc) in comparison with the total number of studies reviewed in this article. The majority of the studies cited in this review manuscript have used at least one confirmational test to replicate the microarray findings. Conclusion Currently, the use of microarray technology in respiratory research is limited by the tissue sample, incomplete available arrays and the analysis of data generated from this technology. Clinical use of microarrays technology is still in its infancy and remains exploratory. For these array-based methods to become truly revolutionary, they must become an integral part of the daily activities of the typical molecular biology laboratory and biomedical institute. There is plenty of room for technical improvements, further development, and more widespread acceptance and accessibility. Optimistically thinking we expect that over the next few years the pattern of development and use of microarrays will be on a similar trajectory to that seen for computers and other high-tech electronic devices, which started out as exotic and expensive tools in the hands of the few developers and then moved quickly to become easier to use, more available and less expensive. Alternatively authors believe that so far microarrays have not added much to our understanding and the possibility to live up to the great 'hype' that was generated belongs to the distant future. Our view is that the application of these approaches will improve dramatically the effectiveness and reliability of microarray technology in studies of diseases of complex organs like the lung, and will have a major impact on our understanding of molecular pathogenesis for the foreseeable future. Whether our hopes will be fulfilled or disproved remains to be seen. List of abbreviations ADA: Adenosine Deaminase ALI: Acute Lung Injury ASM: Airway Smooth Muscle a-SMA: a-Smooth Muscle Actin BAL: Bronchoalveolar lavage CAGE: Composite Atopy Gene Expression COPD: Chronic Obstructive Pulmonary Disease CCR7: Chemokine Receptor 7 DEP: Diesel Exhaust Particles ECM: Extracellular Matrix eQTL: expression quantitative trait locus G.I: Gene Intensity Ig: Immunoglobulin ID: Inhibitor of Differentiation I/D: Intensity/Background ratio IPF: Idiopathic Pulmonary Fibrosis MBP: Major Basic Protein MCs: Mast Cells MCP: Monocyte Chemoattractant Protein MMP: Matrix Metalloproteinase NF-κB: Nuclear Factor-Kb nsPLA 2 : snake venom phospholipase A 2 RT-PCR: real time-polymerase chain reaction TGF-b: Transforming Growth Factor-b TNF: Tumor Necrosis Factor UIP: Usual Interstitial Pneumonia VEGF: Vascular Endothelial Growth Factor Competing interests The authors declare that they have no competing interests. Authors' contributions AT, GP and DB were involved with the study conception. AT performed the data acquisition and interpretation. DB and GP were involved in revising the article for important intellectual content. All authors read and approved the final manuscript.	/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC543572.xml
521490	Methods of assessment of patients for Nd:YAG laser capsulotomy that correlate with final visual improvement	Background This paper attempts to clarify the usefulness of various simple pre-operative measures in estimating the potential for a visually successful capsulotomy. Methods 24 patients attending for capsulotomy had pre-operative measures of glare with BAT tester, visibility of posterior pole and grading of posterior capsular pearls and fibrosis seen at slit lamp. Visual function was measured before and after standardised capsulotomy. Correlations of the various preoperative measures with eventual visual function improvements were calculated. Results Pearls at slit lamp and poor posterior pole visualisation were all correlated with improvements in visual acuity and contrast sensitivity after capsulotomy. Amount of fibrosis visible at slit lamp and glare assessment were not correlated with vision improvements after laser. Conclusion Of the various measures that are taken prior to Nd : YAG capsulotomy, some correlate with eventual visual improvement but for others no clinical utility was found. Practitioners should note these findings as they are especially of use in more questionable or high-risk cases to help determine whether referral for PCO treatment by Nd: YAG capsulotomy is likely to benefit the patient.	Background Posterior capsular opacification (PCO) remains one of the most common post operative morbidities in modern day cataract surgery [ 1 , 2 ] and Nd:YAG posterior capsulotomy is one of the most commonly performed surgical procedures. However, the Nd: YAG capsulotomy procedure has been associated with complications such as damage to intraocular lenses [ 3 ], post operative intraocular pressure increases [ 4 ], cystoid macular oedema [ 4 ], disruption of the anterior vitreous face [ 5 ] and increased incidence of retinal detachment [ 6 ]. Until recently Nd:YAG laser treatments have cost the U.S healthcare system up to $250 million annually [ 7 ]. Apart from exposing a patient to unnecessary risk, unqualified capsulotomies worsen this burden to the developed and developing world [ 8 ]. PCO is an extremely common development in patients after cataract extraction and in many mild cases it may not be immediately obvious whether it is visually significant. Patients may have reduced vision from other undetermined causes or have some measures of visual function that are not reduced at all. Information is needed that will help the practitioners to decide on the visual significance of a patient's PCO using tools that are easily available. The specific aim of this study was to assess the correlation between preoperative measures that are easily performed (fundus visibility, capsule opacity grading and initial glare testing) with eventual visual function improvements after Nd: YAG capsulotomy. This information would be of use to practitioners when deciding whether a patient had PCO that was clinically significant. In other words, we aimed to test the clinical utility of these preoperative tests for Nd:YAG laser capsulotomy. Methods Local research ethical committee approval for the study was obtained and the study was performed in accordance with the treaty of Helsinki. 26 consecutive patients who had been referred for Nd: YAG capsulotomies for posterior capsular opacification were recruited and full informed consent was obtained. Patients excluded were those that had media opacities other than PCO or were not suitable for capsulotomy treatment. Patients in whom macular or disc pathologies were present were excluded from posterior pole assessment arm of the study as grading of visualisation would not be valid. Patients were roughly equally divided between silicone and acrylic IOLs. Prior to Nd:YAG capsulotomy, the patients visual function was assessed with their normal physiological pupil state in terms of best corrected distance and near visual acuity using Bailey Lovie logMAR charts and contrast sensitivity with and without glare, using a Pelli-Robson chart under standard illumination levels at 1 meter. Glare was tested using a Mentor brightness acuity tester (BAT) instrument (set at medium illumination levels) and recorded as the level of contrast sensitivity chart read when exposed to the BAT. Pupils were dilated with topical 1% tropicamide. After a minimum of 20 minutes, the size of the pupil was recorded using a millimetre ruler and the PCO was graded according to the slit-lamp appearance (grade 0 to 4). Use of such slit lamp grading has been established in scientific literature for PCO assessment [ 9 - 16 ] There are of course many other systems available for assessment of PCO [ 17 ] We chose a system based on slit lamp grading as it is the most commonly performed method of PCO assessment in practice. Many such slit lamp grading systems exist, but none have been proven to be gold standard in terms of clinical utility. After our own studies on the effect of PCO on visual function [ 18 ], we decided to use a protocol based on that described by Sellman and Lindstrom [ 19 ], recording pearls and fibrosis separately, which we feel is likely to be as clinically valid as any system that is commonly used in pre-assessment for Nd;YAG capsulotomy. All these tests were performed by one practitioner (T.A.), who was masked to fundus gradings, using the following scale, for pearls and fibrosis separately; 0 None visible at all 1 Visible but none reaching to IOL edge 2 At IOL edge 3 Well Inside IOL edge but visual axis clear 4 Across visual axis Visualisation of the posterior pole was then assessed by examining the optic disc and macula using a Volk 90D lens. Visualisation of the optic disc was subjectively graded according to the following scale (adapted from the Madurai Intraocular lens study IV [ 20 ]): 0 Clear view of optic disc margin, blood vessels at the optic disc and nerve fibre layer (NFL examined using the red-free filter) 1 Clear view of optic disc margin, but disc blood vessels and/or nerve fibre layer are not clearly seen 2 Optic disc margin, as well as disc blood vessels and nerve fibre layer are not clearly seen Visualisation of the macula was subjectively graded according to the following scale: 0 Clear view of foveal reflex, peri-foveal blood vessels and nerve fibre layer 1 Diminished foveal reflex, but clear view of peri-foveal blood vessels and nerve fibre layer 2 Blurred foveal reflex, peri-foveal blood vessels and/or nerve fibre layer The totals for the visualisations of the optic disc and the macula were combined to produce a total posterior pole visualisation score ( PolVS ), ranging from 0 to 4 in order of decreasing visualisation. All examinations of the posterior pole were carried out by the same examiner (NP). In order to be masked as to whether the patient was pre- or post- Nd:YAG capsulotomy, NP examined the fundus with the lenses already placed in front of the eye, thus obscuring any anterior segment view. The patients then had Nd:YAG capsulotomy by the same surgeon (TA). This involved initial setting of 1 mJ and subsequent rises of 0.5 mJ as necessary to pierce the posterior capsule. The laser treatment was initiated off axis in a horizontal line across centre, followed by a line in the vertical axis to form a cross. Any obvious lines of capsule tear were treated with laser if deemed beneficial and overall energy used was kept to a minimum. Treatments in this manner produced small capsulotomies of size 2–3 mm diameter. Size of capsulotomy was dictated by ease of making openings, and concerns over energy used. In general the aim was to create an opening using minimum energy, which might be small, but which could be enlarged at follow up visits if deemed necessary. Four weeks post Nd: YAG capsulotomy, the patients were reassessed again in terms of visual function as described earlier, by the same practitioner (TA). The pupils were dilated using 1% tropicamide and after a minimum 20-minute interval, the size of the pupil was recorded. The PolVS was again graded according to the same scales as before, by the same examiner (NP), again masked as to the state of the posterior capsule. Four weeks post-capsulotomy was chosen for re-examination, on the basis of evidence suggesting that capsulotomies enlarged progressively up to one month after Nd: YAG laser and then stabilised thereafter [ 21 ] On inspection, the data was found to be of skewed distribution. All correlation calculations were therefore performed using the Spearman's rank correlation coefficients. Means were compared using Wilcoxon signed ranks test for non-normally distributed data. Significance was at the p < 0.05 level. Because the visualisation scale was composed of a combination of sub-scales (2 items for PolVS), internal validity was determined by the Cronbach test of reliability19,20. This is a commonly used test statistic to determine the degree with which constituent items within a scale correlate with each other. An alpha coefficient of = 0.7 is considered necessary for a composite of measurements to be considered a scale. Statistical analysis was performed using SPSS for Windows (version 8.0) for all calculations. Results A total of 26 eyes of 26 patients were recruited into the study. Mean age was 75.2 years (range 52 to 90). There were 14 females and 12 males. 2 patients only had one examination pre-Nd:YAG capsulotomy, and declined any further examinations. Therefore, 24 patients were seen pre- and post- Nd:YAG capsulotomy. The PolVS scale was found to be internally reliable and consistent (α coefficient = 0.7824). Mean "pearl" grading score was 3.3 (± 1.3) and mean "fibrosis" grading score was 2.3 (± 1.4). 5 patients were omitted from PolVS score as macular or disc pathology prevented objective grading. Mean improvement of PolVS (n = 19) was 1.95 (S.D. ± 1.31) (95% C.I. 1.32 to 2.58) (p < 0.0001, Wilcoxon signed ranks test). Improvements in the PolVS score after Nd: YAG capsulotomies are shown in fig. 1 . Mean improvement of distance logMAR visual acuity was 0.32 (± 0.29) (95% CI 0.20 to 0.44) and near logMAR visual acuity was 0.32 (± 0.29) (95% CI 0.20 to 0.44). Mean improvement of log contrast sensitivity was -0.41 (± 0.39) (95% CI 0.25 to 0.57) and glare testing vision was -0.35 (S.D. ± 0.41) (95% CI 0.18 to 0.52). (Improved contrast sensitivity and glare testing visual function are associated with lower grading scores.) For all the above analyses, p < 0.0001, except for mean improvement in glare testing (p = 0.001) (Wilcoxon signed ranks test). The main aim of the study was to determine which pre-operative assessments on patients attending for Nd: YAG capsulotomy would correlate with eventual visual function outcome of the patient, and thus be useful clinical measures. The results of the various correlation coefficient analyses are shown in table 1 . The table demonstrates that a patient's improvement in visual function after laser capsulotomy is correlated to measurements made before treatment of pearls at slit lamp and posterior pole visibility score. Measurements of fibrosis at slit lamp and of glare do not correlate with eventual improvements in vision after capsulotomy. Discussion If it is decided to offer a patient Nd; YAG capsulotomy treatment they need to be told of the likely benefits of the treatment as well as risks. Information on potential visual function improvements for each specific patient would be especially welcome in decisions in uncertain cases and higher risk patients such as high myopes. Practitioners should have some evidence for the potential benefit to visual function before suggesting treatment. Retroilluminated photograph analysis by computer have been correlated with improvements in vision [ 22 ] but these tools are not available to most ophthalmologists in the examination room. Practitioners are able to subjectively estimate the amount of PCO on slit lamp examination of posterior capsule but often with more subtle PCO it is difficult to assess clinical relevance. It is thought by some [ 23 ] that contrast sensitivity with glare testing are likely to be particularly sensitive to visual loss from PCO, but of course these visual parameters are also open to influence from many other ocular states. Visibility of discs has been used as an assessment of the amount of PCO [ 20 ] but without convincing evidence of validity. Although all of the above tests have been used to try to assess how much benefit a particular patient may gain from capsulotomy, there has previously been little empirical evidence supporting their use. This study provides some evidence for the use and avoidance of various preoperative tests in assessing potential benefit from Nd; YAG laser capsulotomy. Preoperative glare measures There has been some controversy as to the relative usefulness of the different measures of visual function for assessing PCO severity [ 23 - 27 ], with some suggestion that glare assessment might be particularly useful for such anterior segment disorders [ 23 , 27 ] This study suggests that initial glare measures with the BAT did not correlate with eventual improvements in any visual function. Our results do, however, show a definite improvement in glare following Nd:YAG capsulotomy, all of whom underwent a small capsulotomy (2–3 mm). It appears that glare assessment with BAT is of no clinical utility when assessing the potential improvement a patient may gain from Nd:YAG capsulotomy. This may reflect the difficulties of BAT glare measurement which has been found to be inconsistent in some studies [ 28 ]. Indeed, glare assessment by any means is not simple to perform reliably and with clinical utility. Slit lamp assessment of PCO by practitioner There were two types of PCO assessed, pearls and fibrosis, both graded 0–4 with identical criteria for severity. The study shows very clearly that the fibrosis score was not significantly correlated to any improvements in visual function. The pearl score however was significant in correlating with the eventual improvement of all visual functions. This agrees with previous work using computerised analysis of retro-illumination photographs, which showed that only very central fibrosis affected vision whereas pearls could be detrimental even in para-axial locations [ 18 ] Other studies have also shown that pearls have a greater effect on vision than fibrosis [ 29 ]. The assessment of fibrosis at slit lamp is commonly performed before Nd:YAG capsulotomy, but according to this study does not correlate with eventual visual outcomes. This finding was perhaps the most poignant due to the common clinical use of amount of fibrosis at slit lamp when deciding whether a patient would benefit from capsulotomy. The finding deserves further investigation, but it may be that antero-posterior thickness of the fibrosis as well as density of opacification are important contributory variables to any visual loss. This thickness of PCO would not be easily assessed accurately and routinely at slit lamp, and was not incorporated into the grading scale used for this experiment. In contrast, clinical utility of pearl assessment is demonstrated, by its correlation with eventual improvement of visual function. Visualisation of posterior pole Visualisation of posterior pole according to the scale above correlated with improvements in visual acuity and contrast sensitivity. Using the specific criteria as described may have increased the objectivity of the test. This confirms expected agreement between the difficulty of visualization of the posterior pole and the patient's own vision of the outside world. An assessment of the posterior pole along the suggested guidelines is shown to have significant clinical utility. It may be that some of the correlations found were due to some criteria measured being closely related, for example, pearls at sit lamp and visualisation of fundus. The numbers in this analysis were unfortunately insufficient for stepwise multiple linear regression analysis, which could be used in future studies to determine the relative importance of the different measures. However the principle aims of the current study were realised with the presented experimental protocol and statistical analysis. Conclusions This study confirms the clinical value of pre-YAG measurements of pearls graded at slit lamp and posterior pole visualisation. However, other measures were not significantly correlated with improvements in visual function, namely pre-YAG testing of glare with BAT and pre-YAG grading of fibrosis seen at slit lamp. If a practitioner is uncertain as to the potential visual benefit of an Nd:YAG capsulotomy, an assessment of posterior pole visualisation or pearls at slit lamp should be useful. There is little clinical utility shown from this study in attempting glare assessment or assessment of fibrosis at slit lamp and these assessments may lead to erroneous expectations from treatment. This information is of practical value to practitioners faced with the common problem of assessing suitability of patients for Nd: YAG capsulotomy. It should especially be of use in more uncertain or high-risk cases to determine whether referral to PCO treatment by Nd: YAG capsulotomy is clinically appropriate. Competing interests The author(s) declare that they have no competing interests. Authors' contributions Both authors were involved in planning, design, execution and writing of this paper. Both read and approved the final manuscript. Pre-publication history The pre-publication history for this paper can be accessed here:	/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC521490.xml
544552	Fish Need You	null	There is little in biology that compares in beauty and limpidity to the development of a zebrafish embryo as viewed through a light microscope. The transparent eggshell and embryo tissues expose the minutest details of cell migrations and organ assembly to the curious viewer. Within a day, distinct vertebrate features emerge: a distinct head with the outlines of two large eyes, a quickly pumping heart, a notochord, and a growing array of somites—the bone and muscle precursors—stretching from trunk into tapering tail. The transparent zebrafish embryo has allowed geneticists to discover a large number of mutants with anomalies in the development of external and internal organs. Seven mutations, collectively known as “You-class,” turn the pointed, chevron-like somites into shallow, rounded arcs (“You” stands for “U-shaped”). Ian Woods and William Talbot now show that the You mutation disrupts a new modulator of Hedgehog signaling. Hedgehog is an extracellular signaling protein that can impose various fates on target cells at close proximity or over longer distances. Much research is focused on understanding the factors that promote or limit Hedgehog's activity and range. Woods and Talbot propose that the You protein acts in the extracellular environment to promote Hedgehog signaling. Hedgehog was originally named for mutations that cause excess brush-like denticles to grow on the surface of fruitfly embryos, but it is now known to direct countless developmental decisions in invertebrates and vertebrates alike. In addition, several cancers are known to result from inappropriate Hedgehog signaling. In fish, Hedgehog's best-documented role is in muscle development. In the absence of Hedgehog signaling, cells destined to become slow muscle fibers fail to differentiate properly. A subset of these slow muscle cells—the muscle pioneers—congregate near the dorso-ventral midline of the embryo, where the dorsal and ventral halves of somites converge. When these specialized cells are absent, abnormal somite assembly leads to the U-shaped phenotype. The authors found that You mutants showed many telltale signs of reduced Hedgehog signaling. Proteins that are normally expressed at certain times during the development of slow muscle cells were not activated in You mutants, indicating that these cells did not form. Mutant embryos also displayed reduced expression of the Hedgehog receptor Patched, a universal reporter of Hedgehog signaling activity. In addition, You mutants had specific ventral spinal chord defects that are shared by known Hedgehog pathway mutants. Yet You mutants expressed Hedgehog normally. Moreover, Hedgehog targets could still be activated in You mutants in response to excess Hedgehog signaling, suggesting that the signaling cascade is left intact. The authors concluded that the You protein was a facilitator rather than a crucial transmitter in Hedgehog signaling, likely acting at a step upstream of a cell's response to Hedgehog. Normal muscle pioneers could form in chimeric embryos (embryos made of wild-type and You mutant cells) regardless of which cells—the Hedgehog-producing cells or Hedgehog-responding muscle precursors—expressed You. This made it most likely that the You protein acted outside the cells, perhaps as a cell matrix component. The authors mapped the You mutation and found that it disrupted the coding region of a gene encoding a putative secreted protein. The predicted You protein is closely related to members of the mouse SCUBE family, a group of proteins that are defined by characteristic extracellular motifs (although these proteins have not yet been linked to Hedgehog signaling). This observation strengthens the hypothesis that the You protein has extracellular functions, and the researchers' experimental evidence supports a role for You in transport or stabilization of Hedgehog. At later stages You could also participate in other signaling pathways, as its expression does not always coincide with that of Hedgehog during zebrafish development.	/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC544552.xml
552319	The small heat-shock proteins IbpA and IbpB reduce the stress load of recombinant Escherichia coli and delay degradation of inclusion bodies	Background The permanently impaired protein folding during recombinant protein production resembles the stress encountered at extreme temperatures, under which condition the putative holding chaperones, IbpA/IbpB, play an important role. We evaluated the impact of ibpAB deletion or overexpression on stress responses and the inclusion body metabolism during production of yeast α-glucosidase in Escherichia coli . Results Deletion of ibpAB , which is innocuous under physiological conditions, impaired culture growth during α-glucosidase production. At higher temperatures, accumulation of stress proteins including disaggregation chaperones (DnaK and ClpB) and components of the RNA degradosome, enolase and PNP, was intensified. Overexpression of ibpAB , conversely, suppressed the heat-shock response under these conditions. Inclusion bodies of α-glucosidase started to disaggregate after arrest of protein synthesis in a ClpB and DnaK dependent manner, followed by degradation or reactivation. IbpA/IbpB decelerated disaggregation and degradation at higher temperatures, but did hardly influence the disaggregation kinetics at 15°C. Overexpression of ibpAB concomitant to production at 42°C increased the yield of α-glucosidase activity during reactivation. Conclusions IbpA/IbpB attenuate the accumulation of stress proteins, and – at high temperatures – save disaggregated proteins from degradation, at the cost, however, of delayed removal of aggregates. Without ibpAB , inclusion body removal is faster, but cells encounter more intense stress and growth impairment. IbpA/IbpB thus exert a major function in cell protection during stressful situations.	Background Production of recombinant proteins as inclusion bodies in Escherichia coli can induce several stress responses [ 1 - 3 ] and may interfere with primary metabolism and cellular protein synthesis [ 4 - 6 ]. The metabolic burden of the synthesis of recombinant and stress proteins [ 7 , 8 ] may result in stop of the synthesis of components of the protein synthesising machinery, and ribosomes may be actively degraded [ 4 , 9 , 10 ]. Most commonly, the heat-shock response is induced, as the recombinant proteins – unable to fold properly – overload the chaperone network [ 11 , 12 ]. While appropriate execution of stress responses is of vital importance for the host Escherichia coli [ 13 ], the high metabolic load of strong foreign protein production can prevent proper adaptation [ 14 , 15 ]. The stress situation can become permanent, and only few cellular proteins except heat-shock proteins are synthesised. Analogously, incubation at extreme temperatures is characterised by near exclusive heat-shock protein synthesis as a consequence of extensive aggregation of cellular proteins [ 18 , 19 ]. To deal with this challenge, the chaperone DnaK changes its function: Under physiological conditions, DnaK promotes folding of kinetically trapped intermediates by ATP-dependent partial unfolding, but at high temperatures it becomes a holding chaperone and binds its client proteins permanently, unable to release them in a folding-competent conformation [ 20 ]. The small heat-shock proteins (sHsps) are intrinsic holding chaperones of multimeric structure. The multimers dissociate at high temperatures into smaller entities for exposure of client binding sites. After binding of misfolded client proteins, the sHsps form large globular assemblies [ 21 ]. Thereby, they prevent irreversible aggregation, and under permissive conditions transfer the client proteins to ATP-dependent chaperones for refolding or degradation [ 22 , 23 ]. Also the E. coli sHsps, IbpA and IbpB, reversibly dissociate upon prolonged incubation at 50°C from multimers of 2 – 3 MDa to monomers in the case of IbpA and to oligomers of 650 – 700 kDa in the case of IbpB, exposing hydrophobic sites [ 24 , 25 ]. IbpA and IbpB are highly homologous [ 26 ], but show distinct features. While IbpB is soluble when overexpressed alone, prevents inactivation or aggregation of proteins upon heat and oxidative stress in vitro and facilitates their subsequent refolding by DnaK [ 23 - 25 ], a role for IbpA in mediating the transfer of IbpB and other proteins into the insoluble cell fraction has been proposed [ 19 , 27 ]. IbpA and IbpB have been found associated with thermally aggregated cellular proteins and with inclusion bodies [ 26 , 28 ]. The resolubilisation of protein aggregates, which in vitro and in vivo is mediated by ClpB together with DnaK [ 28 - 31 ], can be retarded both in strains overexpressing ibpAB as well as in some cases in strains lacking ibpAB [ 19 , 32 , 33 ]. IbpA and IbpB might control the reversible transition between soluble and aggregated forms that account for the plasticity of inclusion bodies [ 27 , 34 , 35 ]. The role of IbpA/IbpB in cell protection is less clear. Overexpression of ibpAB confers resistance to heat and other stresses [ 37 ], but deletion of the ibpAB operon does not affect growth even at high temperatures. Only recovery from prolonged exposure to extreme temperatures such as 50°C are impaired in an ibpAB mutant, and growth of ibpAB dnaK double mutants at high temperatures [ 19 , 36 ]. In this study we exploit the similarities between incubation at extreme temperature and production of aggregation-protein proteins to examine the functions of the small heat-shock proteins IbpA/IbpB in inclusion body metabolism and stress relief during α-glucosidase production. The enzyme α-glucosidase from Saccharomyces cerevisiae produced in recombinant E. coli is a suitable model system. It accumulates as inclusion bodies and induces the heat-shock response including prolonged expression of the ibpAB operon [ 17 ]. It is, however, also able to fold into its native conformation in E. coli if produced at low temperature [ 38 ], enabling the study of in-vivo-reactivation kinetics. Results IbpA/IbpB levels during α-glucosidase production with and without IbpAB coexpression The concentrations of IbpA and IbpB were estimated from the proteomes of cells producing α-glucosidase for two hours. The ibpAB operon was induced already during α-glucosidase production at 30°C, with moderately higher levels at elevated temperatures (Table 1 ), whereas no IbpA/IbpB was detected in the plasmid-free strain. The ibpAB operon including its own promoter was inserted immediately downstream of the α-glucosidase gene into the plasmid pKK177-3/GLUCP1, expecting overexpression of ibpAB from its heat-shock promoter and as a tricistronic operon together with α-glucosidase. This plasmid gave ten and four times higher concentrations of IbpA and IbpB, respectively, at 30°C, and the sHsp level increased further in proportion to temperature (Table 1 ). With ibpAB overexpression, the IbpA accumulation was intensified more strongly and the ratio of IbpA to IbpB rose from 2–3 to 6–8. As the two sHsp may exert different functions, this change need to be accounted for. Depending on the temperature, overexpression increased the concentrations of IbpA and IbpB relative to their concentrations without overexpression four to more than 25 times (Table 1 ). This wide range of concentrations served to investigate the impact of IbpA/IbpB on α-glucosidase production and in-vivo-reactivation. Elevated levels of IbpA/IbpB increase α-glucosidase accumulation α-Glucosidase was produced in E. coli MC4100 with and without ibpAB overexpression and in an otherwise isogenic ibpAB deletion mutant. α-Glucosidase accumulation was enhanced at elevated IbpA/IbpB level and reduced in the deletion strain in a temperature dependent manner; the effect was strongest at high temperature (Table 2 ). At temperatures below 30°C, the IbpA/IbpB levels did not influence the amount of α-glucosidase, which accumulated mainly in the soluble cell fraction in active form (Table 2 ). Post-induction growth of the ibpAB deletion mutant was retarded at higher temperatures (Fig. 1 ), although there was no impact of the ibpAB deletion on growth without induction of α-glucosidase synthesis (not shown) or concomitant to production at lower temperature. Thus, the sHsp are effective only under conditions that favour inclusion body formation. IbpA/IbpB attenuate accumulation of stress proteins during α-glucosidase production To study the impact of IbpA/IbpB on cellular reactions during α-glucosidase production, the proteomes of cultures producing α-glucosidase for two hours at various temperatures were separated by two-dimensional gel electrophoresis (shown for 42°C in Fig. 2 ). While chaperones assisting de-novo protein folding such as GroEL and GroES were not significantly influenced by IbpA/IbpB level (Fig. 2 ), the DnaK concentration showed a negative correlation with IbpA/IbpB availability at 42°C (Fig. 3 ). The disaggregating chaperone ClpB reached higher concentrations in the ibpAB deletion mutant at all temperatures, in addition to the induction at elevated temperatures (Fig. 3 ). ClpB accumulation was repressed, however, in the wildtype strain at 30°C, and with overexpression of ibpAB both at 30 and 37°C. As the overall level of IbpA/IbpB was low at these temperatures, and other cellular proteins were not affected, this repression is unlikely to be a passive effect of the metabolic burden of ibpAB expression. A similar pattern was found for enolase, which showed higher levels at elevated temperatures only in the deletion mutant (Fig. 3 ). As enolase is a glycolytic enzyme as well as a putative structural component of the RNA degradosome [ 39 ], we examined another metabolic enzyme, glyceraldehyde 3-phosphate dehydrogenase (GAPDH), and polynucleotide phosphorylase (PNP), another RNA degradosome member. The positions on the 2D gels of further degradosome components, RNAse E and RhlB, are not listed in the database. While PNP accumulated to higher concentrations at elevated temperatures and reached highest levels in the deletion mutant (Fig. 3 ), GAPDH concentration was not significantly affected by temperature and had a slightly lower concentration in the ibpAB deletion mutant at 42°C (Fig. 3 ). Thus, enolase seems to be induced as a part of a stress response rather than as a metabolic enzyme. A typical house-keeping protein, the translational elongation factor EF-Tu, had a two to three times higher level at 42°C than at lower temperatures, but the concentrations were not effected by IbpA/IbpB level (Fig. 3 ). Thus, while housekeeping proteins are not influenced by IbpA/IbpB levels, the accumulation of ClpB and components of the RNA degradosome was intensified in the ibpAB deletion mutant and attenuated with ibpAB overexpression. This indicates that IbpA/IbpB can relieve the cells of the stress encountered during α-glucosidase production. Impact of chaperones on in-vivo-reactivation and degradation of α-glucosidase from inclusion bodies The impact of chaperones on the metabolism of α-glucosidase inclusion bodies was examined in wildtype E. coli MC4100 with or without overexpression of ibpAB and in isogenic strains carrying deletions of ibpAB , clpB , or dnaK . After production of α-glucosidase at 37°C for two hours, the cultures were treated with tetracycline to arrest protein synthesis and transferred to 30°C to initiate disaggregation of inclusion bodies as reported for other proteins [ 34 ]. In the wildtype and ibpAB strains, disintegration of inclusion bodies, increase of α-glucosidase activity, and higher concentrations of α-glucosidase in the soluble cell fraction could be detected already within 20 min; the disaggregation was accompanied by degradation (cf. below). While the concentration of α-glucosidase was 10–25% higher in dnaK and clpB deletion mutants than in the wildtype strain, the specific activities obtained during the production phase were less than 0.1 Umg -1 , compared to about 0.3 Umg -1 in the other strains (data not shown). Importantly, neither disaggregation, nor reactivation nor degradation took place in these mutants within two hours, and the concentration of α-glucosidase inclusion bodies remained unchanged. The kinetics of disaggregation, reactivation and degradation of inclusion bodies at modified levels of IbpA/IbpB, i.e. with deletion or overexpression of ibpAB , were investigated as a function of temperature. After production of α-glucosidase at 37°C, the cultures were transferred to various temperatures. As protein synthesis was arrested before the transfer, the initial IbpA/IbpB level did not depend on the disaggregation temperature. The α-glucosidase concentration in the insoluble cell fractions of all cultures decreased with time, but with different kinetics. In the ibpAB deletion strain, the disaggregation kinetics were not influenced by the incubation temperature (Fig. 4A–D ). With overexpression of ibpAB , the removal of α-glucosidase from the inclusion bodies was fastest at 15°C (Fig. 4A ), proceeding with similar rates as in the other strains. The disaggregation was, however, progressively slower at 30 and 37°C (Fig. 4C,D ). At 37°C, α-glucosidase was largely preserved in the insoluble cell fraction of the ibpAB overexpressing strain (Fig. 4D ). Thus, disaggregation was retarded at higher temperatures in an IbpA/IbpB dependent manner. Disaggregation of α-glucosidase inclusion bodies was accompanied by an increase of the specific α-glucosidase activities in the cell extracts. The reactivation showed a clear temperature optimum in the range of 24 to 30°C, most evident with wildtype levels of IbpA/IbpB (Fig. 4F,G ). The specific α-glucosidase activities increased sevenfold up to 2 Umg -1 within 3 h, whereas at both higher and lower temperatures of 37°C and 15°C, only 0.5–0.75 Umg -1 were obtained (Fig. 4E,H ). Both, ibpAB overexpression and deletion, did not improve the maximum reactivation yields. The resolubilisation, measured as an increase of the α-glucosidase concentrations in the soluble cell fraction, showed very similar kinetics as the reactivation (data not shown). A large part of the disaggregated α-glucosidase was subjected to proteolysis: within three hours after arrest of protein synthesis, 30 – 40% of α-glucosidase were lost due to degradation at 15°C in all three strains (Fig. 4I ), and in the ibpAB deletion mutant at all temperatures (Fig. 4I–L ). Similar to the effect on disaggregation, higher ibpAB expression levels or elevated temperatures synergistically retarded degradation, reducing the loss to 10% with ibpAB overexpression at 37°C (Fig. 4L ). IbpA/IbpB thus influenced disaggregation, resolubilisation, reactivation and degradation in a temperature-dependent manner. IbpA/IbpB favour reactivation of α-glucosidase produced at high temperature As IbpA/IbpB are designed for action at extreme temperature, α-glucosidase inclusion bodies were produced at 42°C, followed by arrest of protein synthesis and incubation at 30°C. The production temperature did not influence the initial disaggregation kinetics (data not shown). About one third of the α-glucosidase produced concomitant to ibpAB overexpression was resistant to disaggregation after incubation over night, while it was eliminated apart from traces in the other strains (Fig. 5A ). Also the removal of other components of the inclusion bodies, which appear as smear in the SDS PAGE of the insoluble cell fractions, were impaired in the overexpressing strain (Fig. 5A ), resembling the core inclusion bodies that remain with other proteins [ 34 ], whereas the inclusion bodies produced at low IbpA/IbpB levels were completely dissolved after prolonged incubation. Only outer membrane proteins (indicated by arrows) that result from coprecipitation of cell debris during inclusion body preparation [ 40 ] remain at constant level in the insoluble cell fraction (Fig. 5A ). Thus, the delay of disaggregation by IbpA/IbpB is not specific to α-glucosidase, but effects also cellular proteins in a similar way. Production at 37°C resulted in degradation of 20% of α-glucosidase during incubation at 30°C for 3 h in the wildtype strain (Fig. 4K ). After production at 42°C, however, there was less than 10% of α-glucosidase lost during reactivation at 30°C with or without ibpAB overexpression (Fig. 6 ). Thus, degradation of α-glucosidase after arrest of protein synthesis was predetermined by the preceding production temperature. The protection was dependent on IbpA/IbpB availability, as in the deletion mutant the loss of α-glucosidase was substantial regardless of the production temperature (Fig. 4K , 6 ). Consequently, disaggregated α-glucosidase was quantitatively transferred to the soluble cell fraction after production at 42°C concomitant to ibpAB overexpression, and became the most prominent band in the gel. The band intensity increased from 2 to 7% of the total cellular protein (Fig. 5B ). In contrast, significantly less α-glucosidase could be detected in the soluble fraction of the deletion mutant even after prolonged incubation, finally corresponding to only 4% of cellular protein (Fig. 5B ). The α-glucosidase activity increased during the initial reactivation phase with similar slopes with and without ibpAB overexpression (Fig. 7B ). Stationary level of activity was reached earlier, however, in the ibpAB deletion mutant and in the wildtype strain without overexpression. Hence the final activities in these strains were independent of the production temperature (Fig. 7A,B ). In contrast, the activity continued to increase for about 12 h after arrest of protein synthesis subsequent to α-glucosidase production at 42°C with ibpAB overexpression (Fig. 7B ), resulting in higher activity yields than obtained in the wild type strain. The final α-glucosidase activities after production with overexpression of ibpAB at 42°C were two times higher than after production at 37°C (Fig. 7A,B ). IbpA/IbpB can thus ameliorate reactivation of inclusion bodies produced at high temperature. Discussion Intensified stress response and growth impairment in ibpAB deletion strain Impairment of cell growth by ibpAB deletion is observed only under conditions characterised by depletion of free DnaK, such as under extremely high temperature or artificially controlled low level [ 19 , 33 , 36 , 37 ]. Accumulation of α-glucosidase as inclusion bodies mimics the aspect of permanently impaired protein folding, and growth impairment during α-glucosidase production was more intense in the ibpAB mutant. Moreover, DnaK controls the expression of heat-shock genes. Consequently, chromosomal ibpAB and other heat-shock genes were induced during α-glucosidase production. Especially in the deletion strain, lack of IbpA/IbpB was compensated by enhanced accumulation of the disaggregating chaperone ClpB. Also the levels of two components of the RNA degradosome, enolase and PNP, were elevated in the ibpAB deletion strain at 42°C. Enolase homologues have been reported to be induced by heat shock in a variety of species, including Bacillus subtilis [ 41 ], but there was no indication that the E. coli eno gene can be induced by heat-shock or high temperature [ 42 ]. Both ClpB and DnaK are essential for disaggregation and de novo protein folding ClpB and DnaK accumulated to higher levels in the ibpAB deletion strain. These chaperones cooperate with IbpA/IbpB in a "disaggregation triad" [ 33 ]. Each chaperone, ClpB and DnaK, was essential for reactivation as well as degradation of α-glucosidase, indicating that both processes operate after disaggregation of the inclusion bodies. With other protein aggregates, varying dependencies on DnaK and ClpB for degradation or reactivation were observed [ 29 , 30 , 32 , 43 - 45 ]. To the aggregate-specific factors influencing the chaperone demand for disaggregation in vitro belong their sizes and accessible surface areas [ 31 , 46 , 47 ]. Also the substrate specificity of chaperones may influence the protein-specific requirements. The impaired disaggregation in clpB and dnaK mutants reduced the accumulation of active α-glucosidase already during production, and a negative impact of the clpB deletion on de novo folding has been observed also with a number of other proteins [ 48 ]. Resolubilisation from aggregates might thus be a general route to active protein, in accordance with a quantitative model of inclusion body formation kinetics [ 16 ]. This constitutes a major difference to the irreversible aggregation at extreme temperatures, when the client release of DnaK is retarded [ 20 ]. IbpA/IbpB disburden DnaK and reduce degradation Ongoing attempts to fold α-glucosidase will finally result in its degradation. IbpA/IbpB, which are postulated to recognise similar traits of aggregation prone proteins as DnaK [ 19 ], will withdraw α-glucosidase from the futile cycle of DnaK binding and release. Thereby, IbpA/IbpB on the one hand disburden the DnaK system to make resources available to repress stress responses, and on the other hand protect α-glucosidase from degradation. These functions might be related, as the heat-shock response includes also proteases that may accumulate to lower levels with ibpAB overexpression. Deletion of ibpAB , conversely, lowers the amount of α-glucosidase and other recombinant proteins [ 27 , 49 ]. Consequently, unlike to incubation at 50°C, ibpAB deletion did not increase aggregation during α-glucosidase production, rather reduced accumulation of insoluble α-glucosidase. Functional replacement of DnaK by IbpA/IbpB is also reflected by the amount of DnaK associated with inclusion bodies and aggregated cellular proteins, that is increased in the ibpAB deletion mutant and decreased by ibpAB overexpression [[ 19 ], unpublished results]. IbpA/IbpB prevent degradation in a temperature-dependent manner Disaggregation and degradation of inclusion bodies took place after arrest of protein synthesis. The disaggregation at and above 30°C was retarded by high levels of IbpA/IbpB. In wildtype cells, IbpA/IbpB accumulation is therefore restricted to requiring conditions, e.g. by degradation of IbpA/IbpB after dissociation from protein aggregates [ 33 ] or repression of ibpAB expression despite an otherwise active heat-shock response [ 14 ]. Temperatures during both, formation and disaggregation of inclusion bodies, were important for the rate of disaggregation and yield of reactivation. Irreversible structural changes of IbpA/IbpB at low temperatures [ 25 ] may impair the protective function, making the disaggregation kinetics at 15°C independent of the IbpA/IbpB level. Hence with other systems, depending on the conditions, no influence of the ibpAB deletion on disaggregation kinetics or even delayed disaggregation in the mutant were observed [ 19 , 33 , 34 , 50 ]. The reactivation rate is determined by the difference of disaggregation and degradation rates, which are influenced in parallel by the availability of the sHsps. Hence reactivation was hardly effected by ibpAB deletion or overexpression, in accordance with observation for other proteins [ 32 , 34 ]. With high temperature during production, degradation during the subsequent disaggregation phase was nearly completely avoided in ibpAB proficient strains. In the ibpAB deletion strain, on the contrary, production temperature did not ameliorate the recovery of soluble α-glucosidase. Consequently, the reactivation phase was prolonged with higher IbpA/IbpB concentrations, and the final α-glucosidase activities were increased from 4 in the wildtype cells to 5 Umg -1 with ibpAB overexpression. As the yield of reactivated protein is more important for heat-stress survival than the rate of disaggregation [ 56 ], the reduction of degradation may also contribute to the cell protective function of the sHsps. Conclusions Deletion of ibpAB renders E. coli more susceptible to growth impairment by production of α-glucosidase. Without recombinant protein production, ibpAB deletion affects growth only at extreme temperatures or in combination with mutations of dnaK . Thus, production of an aggregation prone protein constitutes a good model system to study the functions of IbpA/IbpB under non-lethal conditions. Two major functions were found for IbpA/IbpB: First, the sHsps protect cells during α-glucosidase production and disburden the DnaK system. In the ibpAB deletion strain, insufficient free DnaK was available to control ClpB accumulation, which was strongly intensified. Also enolase and PNP showed a similar pattern. Second, IbpA/IbpB reduce degradation of α-glucosidase during disaggregation from inclusion bodies. The IbpA/IbpB function depends on elevated temperature. Improved reactivation and protection from degradation were found after production at higher temperature, in accordance with the known heat-activation of sHsps. Also during disaggregation, a higher temperature was beneficial, whereas no protection from degradation was found at 15°C. In the ibpAB deletion strain, however, removal of inclusion bodies was accelerated, possibly due to the higher levels of ClpB and DnaK that were found to be essential for reactivation and degradation of α-glucosidase. Thus, the sHsp level is in a delicate balance: IbpA/IbpB are required to protect the cell, but abundant IbpA/IbpB impede the necessary adaptation and retards removal of aggregates. Methods Strains and Plasmids Escherichia coli MC4100 araD139 Δ (argF-lac)U169 rpsL150 relA1 flbB5301 deoC1 ptsF25 rbsR and JGT17 MC4100 Δ ibpAB [ 36 ] were used to produce the yeast α-glucosidase encoded on the plasmid pKK177-3/GLUCP1 [ 38 ]. Plasmid pUBS520, supplying the minor argU tRNA and carrying the lacI repressor gene [ 51 ] was cotransformed for improving the production of α-glucosidase [ 52 ]. The sequence from plasmid pIbp [ 36 ] between the two Hind III restriction sides, coding for the ibpA and ibpB genes including their native promoter, was cloned into the Hind III site of the plasmid pKK177-3/GLUCP1 [ 38 ], located between the α-glucosidase GLUCPI gene and the 5ST1T2 terminator, to give the new plasmid named pKK177-3/GLUCP1_ibpAB, in which the ibpAB operon is under the control of its native promoter and of the tac -promoter upstream of the α-glucosidase gene. Proper orientation of the insert was tested by Ear I digestion. Culture conditions The cultures were incubated on Luria-Bertani (LB) medium, supplemented with appropriate antibiotics (ampicilline 100 μg mL -1 , chloramphenicol 50 μg mL -1 ), on a rotary shaker at 37°C to an OD 600 of 0.5, induced with 1 mM IPTG and transferred to different temperatures as indicated in the results section. For reactivation experiments, tetracycline was added to a final concentration of 25 μgmL -1 two hours after induction to arrest protein synthesis, and the cultures were transferred to the indicated temperatures. Product analysis For SDS-PAGE analysis, cell pellets resuspended in phosphate buffer pH 7 were incubated on ice with 33 mgL -1 lysozyme for 30 min and disrupted by sonication for 20 s. Soluble and insoluble fractions were separated by centrifugation for 20 min at 13,000 rpm. The insoluble fractions were washed with phosphate buffer twice. SDS-PAGE analysis was performed according to the method of Laemmli [ 53 ]. Levels of α-glucosidase were estimated by densitometry of Coomassie-stained gels. α-glucosidase activity was measured as described [ 54 ]. For the calculation of specific activities, the protein concentration was determined according to Bradford [ 55 ] with bovine serum albumin as standard. Specific activities are expressed as units of α-glucosidase per milligram of total protein (U mg -1 ). Two-dimensional gel electrophoresis Cell pellets from a culture volume of 8/OD 600 mL were dissolved in 200 μL lysis buffer containing 8 M urea, 4% CHAPS, 60 mM DDT, 2% Pharmalyte 3–10 and 0.002% bromophenol blue and incubated for one hour at room temperature. The first dimension was carried out on an IPGphor using 24 cm pH gradient 3–10 strips (amersham pharmacia biotech), with the voltage profile recommended by the manufacturer. The second dimension was run in an Ettan Dalt six (amersham pharmacia biotech) on 12.5% polyacrylamide gels. Gels were stained with Coomassie brilliant blue for quantification or by silver staining for visualisation. Gels for cultures incubated at 42°C were performed in quadruplicate, for other temperatures once. Protein spots were identified by comparison with the E. coli 2D database . Spot identities were verified by N-terminal sequencing for ClpB (RLDRLTN) and enolase (SKIVKII). α-glucosidase was not dissolved from the inclusion bodies and is not visible in the gels.	/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC552319.xml
521484	Complete mitochondrial genome sequence of Urechis caupo, a representative of the phylum Echiura	Background Mitochondria contain small genomes that are physically separate from those of nuclei. Their comparison serves as a model system for understanding the processes of genome evolution. Although hundreds of these genome sequences have been reported, the taxonomic sampling is highly biased toward vertebrates and arthropods, with many whole phyla remaining unstudied. This is the first description of a complete mitochondrial genome sequence of a representative of the phylum Echiura, that of the fat innkeeper worm, Urechis caupo . Results This mtDNA is 15,113 nts in length and 62% A+T. It contains the 37 genes that are typical for animal mtDNAs in an arrangement somewhat similar to that of annelid worms. All genes are encoded by the same DNA strand which is rich in A and C relative to the opposite strand. Codons ending with the dinucleotide GG are more frequent than would be expected from apparent mutational biases. The largest non-coding region is only 282 nts long, is 71% A+T, and has potential for secondary structures. Conclusions Urechis caupo mtDNA shares many features with those of the few studied annelids, including the common usage of ATG start codons, unusual among animal mtDNAs, as well as gene arrangements, tRNA structures, and codon usage biases.	Background Mitochondrial genomes are physically separate from the nuclear genome. For animals, they are typically circular with a compact arrangement of an identical set of 37 genes [ 1 ]. For some animals, all genes are on the same strand, whereas for others they are divided between the two. The arrangement of these genes can remain stable for long periods of time; for example, human [ 2 ] and shark [ 3 ] mtDNAs have the same gene arrangement, and do those of fruit fly [ 4 ] and shrimp [ 5 ]. However, in other lineages, rearrangements have occurred much more rapidly. Many of the same processes that occur in large and complex nuclear genomes also take place in these diminutive genomes, so comparisons among mtDNAs can address general questions of genome evolution, but in a model system that is much more tractable for a large number of taxa. Toward this end, this article describes the complete mtDNA sequence of the fat innkeeper worm, Urechis caupo , the first example from the phylum Echiura. Echiurans comprise about 150 species and are commonly called spoon worms because of the shape of their extensible proboscis. Unlike annelids, they have no overt segmentation, but they develop via trochophore larvae, very similar to those of annelids. U. caupo is a pink, sausage shaped worm that lives in U-shaped burrows in the mud or sand of the intertidal or subtidal zones. Unlike other echiurans, it feeds on plankton by filtering using an elaborate mucus net. Results and discussion Gene content and organization The mtDNA of Urechis caupo is 15,113 nts in length (GenBank accession number AY619711 ) and contains the same 37 genes found for nearly all animal mtDNAs [see ref. [ 1 ]]. All genes are transcribed from the same strand (Fig. 1 ), as is the case for the two studied annelid mtDNAs, the polychaete Platynereis dumerilii [ 6 ] and the oligochaete Lumbricus terrestris [ 7 ] and for several other animal mtDNAs. The arrangement of the genes is substantially similar to those of the two annelids, and shares short regions of similarity with several other mtDNAs, as can be seen in Table 1 . Figure 1 Mitochondrial gene map of the echiuran Urechis caupo. All genes are transcribed from the same DNA strand. Scaling is only approximate. Genes are designated by standard nomenclature except for tRNAs, which are identified only by the one-letter code for the corresponding amino acid, with the two serine and two leucine tRNAs differentiated by numeral as identified in Fig. 3. "nc" indicates the largest non-coding regions; it may be that transcription initiates here, but this is not known. Table 1 Mitochondrial gene arrangement identities found in pairwise comparisons between Urechis caupo and various animals. Full taxon names are given here for the annelids Lumbricus terrestris and Platynereis dumerilii , the mollusks Katharina tunicata , Loligo bleekeri , Cepaea nemoralis , and Mytilus edulis , the brachiopods Terebratulina retusa and Terebratalia transversa , the platyhelminths Fasciola hepatica , Taenia crassiceps , Echinococcus multilocularis , and Hymenolepis diminuta , the arthropods Drosophila yakuba , Anopheles gambiae , Artemia franciscana , Daphnia pulex , Apis mellifera , Locusta migratoria , Ixodes hexagonus , Rhiphicephalus sanguineus , Limulus polyphemus and Lithobius forficatus , the nematodes Trichinella spirallis , Onchocerca volvulus , Meloidogyne javanica , Ascaris suum , and Caenorhabditis elegans , the echinoderms Arabacia lixula , Asterina pectinifera , Paracentrotus lividus , Strongylocentrotus purpuratus , and Florometra serratissima , the hemichordate Balanoglossus carnosus , and the chordate Branchiostoma floridae along with the gene order most typical for vertebrates. Complete citations can be found in Boore (1999) or updated by following the "Evolutionary Genomics" link at . Contiguous gene arrangements are separated by a comma; a slash indicates a gap containing one or more unrelated genes. L. terrestris and P. dumerilii cox3 , trnQ , nad6 , cob , trnW , atp6 , trnR , trnH , nad5 , trnF / trnL2 , nad1 , trnI , trnK , nad3 / trnT , nad4L , nad4 L. terrestris but not P. dumerilii trnL1 , trnA , trnS2 , trnL2 / trnD , atp8 K. tunicata trnL2 , nad1 / nad4L , nad4 / trnH , nad5 , trnF L. bleekeri nad6 , cob / nad4L , nad4 / nad5 , trnF / trnD , atp8 C. nemoralis trnL1 , trnA M. edulis trnL2 , nad1 / trnT , nad4L T. retusa trnL2 , nad1 / nad4L , nad4 / trnH , nad5 , trnF / trnL1 , trnA / trnD , atp8 / cox1 , cox2 T. transversa trnP , trnD F. hepatica , T. crassiceps , E. multilocularis trnI , trnK , nad3 / nad4L , nad4 / trnY , trnL1 / trnS2 , trnL2 H. diminuta trnI , trnK , nad3 / nad4L , nad4 / trnY , trnL1 D. yakuba , A. gambiae , A. franciscana , D. pulex nad6 , cob / nad4L , nad4 / trnH , nad5 , trnF / trnD , atp8 A. mellifera , L. migratoria nad6 , cob / nad4L , nad4 / trnH , nad5 , trnF I. hexagonus , R. sanguineus , L. polyphemus , L. forficatus nad6 , cob / trnL2 , nad1 / nad4L , nad4 / trnH , nad5 , trnF / trnD , atp8 / cox1 , cox2 T. spirallis nad6 , cob / nad4L , nad4 / trnH , nad5 , trnF / trnR , trnH / trnL1 , trnA / trnD , atp8 / cox1 , cox2 O. volvulus trnP , trnD M. javanica trnN , trnG A. suum NONE C. elegans NONE A. lixula , A. pectinifera , P. lividus and S. purpuratus trnL2 , nad1 , trnI F. serratissima nad1 , trnI / nad2 , trnY B. carnosus trnL2 , nad1 / nad4L , nad4 B. floridae and the typical vertebrate arrangement trnL2 , nad1 , trnI / nad4L , nad4 Base composition and codon usage The U. caupo mtDNA is 62% A+T, about the same as for annelid mtDNAs (64% and 62% for P. dumerilii and L. terrestris , respectively). As is typical, all homodinucleotides and homotrinucleotides are greatly over represented relative to a random distribution and CG is the least frequent dinucleotide, both in absolute number and in the ratio of observed to expected. The gene-coding strand has a strong skew against G vs. C but about equal amounts of A vs. T; G-skew ([G-C]/[G+C]) is – 0.24 and T-skew ([T-A]/[T+A]) is – 0.016 [ 8 ]. These values show no striking variation across the length of the mtDNA. Codon usage (Table 2 ) reflects these values, with those ending in A or T being most frequent. In all cases except for two, where they are synonymous, NNC codons are in greater abundance than NNG codons, as is consistent with the coding strand being rich in C relative to G. The two exceptions are CGG and GGG codons, which are each in greater abundance than their respective synonyms, CGC and GGC. This invites the speculation that there is something favored about the GG dinucleotide created when G appears in the second codon position. However, this is not consistently seen, since in the remaining case, AGC codons outnumber AGG codons two-to-one. This effect has been shown to be very strong for codon usage pattern of the mtDNA of the brachiopod Terebratalia transversa [ 9 ]. Table 2 Codon usage in the 13 protein-encoding genes of the Urechis caupo mitochondrial genome. The total number of codons is 3722. The anticodon of the corresponding tRNA gene is shown in parentheses below each amino acid designation. Stop codons are not included in this analysis. Amino acid Codon N % Amino acid Codon N % Phe (F) TTT 161 4.3% Ser (S2) TCT 108 2.9% (GAA) TTC 115 3.1% (TGA) TCC 65 1.7% Leu (L2) TTA 146 3.9% TCA 74 2.0% (TAA) TTG 10 0.3% TCG 3 0.1% Tyr (Y) TAT 42 1.1% Cys (C) TGT 18 0.5% (GTA) TAC 65 1.7% (GCA) TGC 11 0.3% TER TAA --- --- Trp (W) TGA 77 2.1% TAG --- --- (TCA) TGG 21 0.6% Leu (L1) CTT 105 2.8% Pro (P) CCT 72 1.9% (TAG) CTC 62 1.7% (TGG) CCC 44 1.2% CTA 224 6.0% CCA 80 2.1% CTG 38 1.0% CCG 6 0.2% His (H) CAT 33 0.9% Arg (R) CGT 6 0.2% (GTG) CAC 55 1.5% (TCG) CGC 5 0.1% Gln (Q) CAA 84 2.3% CGA 46 1.2% (TTG) CAG 12 0.3% CGG 9 0.2% Ile (I) ATT 200 5.4% Thr (T) ACT 76 2.0% (GAT) ATC 100 2.7% (TGT) ACC 91 2.4% Met (M) ATA 171 4.6% ACA 93 2.5% (CAT) ATG 52 1.4% ACG 5 0.1% Asn (N) AAT 61 1.6% Ser (S1) AGT 7 0.2% (GTT) AAC 65 1.7% (TCT) AGC 16 0.4% Lys (K) AAA 78 2.1% AGA 62 1.7% (TTT) AAG 14 0.4% AGG 8 0.2% Val (V) GTT 49 1.3% Ala (A) GCT 75 2.0% (TAC) GTC 28 0.8% (TGC) GCC 75 2.0% GTA 99 2.7% GCA 122 3.3% GTG 18 0.5% GCG 14 0.4% Asp (D) GAT 23 0.6% Gly (G) GGT 14 0.4% (GTC) GAC 37 1.0% (TCC) GGC 27 0.7% Glu (E) GAA 73 2.0% GGA 127 3.4% (TTC) GAG 10 0.3% GGG 36 1.0% Gene initiation and termination Mitochondrial genes commonly use several alternatives to ATG as start codons. However, 11 of the 13 proteins coding genes of U. caupo mtDNA use ATG. The only exceptions are cox1 , which uses GTG and nad3 which uses ATC. In the case of cox1 , there is an in frame stop only three codons upstream and neither of the intervening codons is ATG. Also, this inference of starting on GTG specifies a set of amino acids well matched to those at the beginning of other Cox1 proteins. The situation for nad3 is nearly identical, with an in frame stop only four codons upstream and no intervening ATG codons. However, downstream of the inferred start are several ATA codons that can not be ruled out as alternatives. The commonality of using ATG as a start codon has also been noted for mitochondrial genes of four annelids, Platynereis dumerilii [ 6 ], Lumbricus terrestris [ 7 ], Helobdella robusta and Galathealinum brachiosum (previously considered to be of the phylum Pogonophora) [ 10 ] and a sipunculid, Phascolopsis gouldii [ 11 ]. A complete stop codon without overlap of the downstream gene is found for all except cox2 , nad1 , nad2 , cob , and nad5 (Fig. 2 ). In each of these cases, it appears that an abbreviated stop codon is generated by cleavage of a downstream tRNA from the polycistronic transcript, which is then completed to a TAA stop codon by polyadenylation. However, in two of these cases ( nad2 and cob ), a complete stop codon could be formed by including only the next two nucleotides, and two other cases ( nad1 and cox2 ), there is an in frame stop codon just one or two codons downstream, respectively. It is not clear how gene overlaps could be resolved from a polycistronic transcript (assuming that the genes of this mtDNA are expressed in this way), but the presence of these stop codons seems beyond coincidence. It could be that they serve as a "back up" in case translation should begin in the absence of transcript cleavage. Figure 2 A greatly abbreviated schematic of the sequence of Urechis caupo mtDNA. In the interest of brevity, the middle portion of each large gene is omitted and replaced by a numeral indicating the number of nucleotides removed. Since all mitochondrial proteins are thought to initiate with formyl-methionine, an M is placed in parentheses at the first codon position of cox1 (GTG) and nad3 (ATC) to indicate nonconformity to the genetic code. Asterisks indicate inferred stop codons whether complete or abbreviated and plus symbols mark nucleotides that would form the first in frame, complete stop codon if genes instead overlap. Transfer RNAs Twenty-two regions can be folded into the typical cloverleaf structures of the expected set of tRNAs (Fig. 3 ). There are several mismatched nucleotide pairs within stems; nearly all of these are flanked by multiple G-C pairs, suggesting that they may provide compensatory stability for these arms. T precedes the anticodon and a purine follows it for all tRNAs. The two serine tRNAs lack potential for folding a DHU arm, as has been found for a number of other animal mtDNAs. There is an alternative folding possible for tRNA(S2) with a six-member anticodon stem and only one nucleotide separating the acceptor and DHU stems; this unusual folding has been found for the homologous genes of some mammals. tRNA(R) also does not have potential for a normally paired DHU arm, although there are three potential nucleotide pairs if two (rather than one) nucleotides were between the DHU and anticodon stems. However, this potential pairing could, alternatively, be a coincidence, with the DHU arm having no paired stem for this tRNA. Those with paired DHU arms have stems of three to five nucleotide pairs and loops of three to eight nts. All tRNAs have potential for stems of three to six nucleotide pairs for their TΨC arms with loops of three to seven nts. One of the tRNAs for serine has the anticodon TCT; although this is often found, the alternative of GCT is otherwise common. Figure 3 The 22 inferred tRNA genes folded into the typical cloverleafstructures. Nomenclature for tRNA substructures is indicated on tRNA(V). Ribosomal RNAs As has been the case for all studied animal mtDNAs to date, two rRNA genes are identified, one for each of the small and large mitochondrial ribosomal subunits. Determining the precise ends of the rRNA transcript requires experimentation, but if it's assumed that they extend to the boundaries of the adjacent genes, then rrnS is 903 nucleotides and rrnL is 1266 nucleotides in length. These genes are arranged sequentially, but without an intervening tRNA gene as is otherwise commonly found. Non-coding regions The largest non-coding region is only 282 nts long. The region is 71% A+T and contains one palindrome of an 11 nt sequence (TCAAAAGGGGT/ACCCCTTTTGA, with a slash indicating the center), but otherwise no large repeat elements. Obviously, this has potential for forming a stem-loop structure, and it may be significant that a short sequence a few nucleotides upstream, TCAAAA, has the potential for competing with this to form a short hairpin with the TTTTGA at the end of the palindrome. There has been previous speculation that regions with potential for competing, mutually exclusive hairpins may play a role in regulating transcription and/or replication [e.g. ref. [ 7 ]]. There are four other potential hairpins in this region with stems of 5–6 bp and loops of 5–17 nt. All four nucleotides occur in homopolymers with much greater frequency than expected by chance, often in runs of four or five. The second largest non-coding region is 43 nt between trnS1 and cox3 . This has no repeat elements and the base composition is unremarkable. What role, if any, these sequences have in the regulation of transcription and/or replication awaits further study. Aside from these 282 and 43 nt regions, there are only 36 total intergenic nucleotides scattered among 14 regions. In seven cases these are 2–6 nts long (CCAAA, AT, TCCC, TAAA, CATAAA, AT, and ACACCT). For the other seven cases, genes are separated by a single nucleotide, and in six of these, that nucleotide is a C. (The remaining case is a T.) The prevalence of C is consistent with the measured G-skew between the strands, although it is possible that this otherwise indicates some function of these nucleotides. Conclusions This is the first description of a complete mitochondrial genome sequence of a representative of the phylum Echiura. The genome contains the same 37 genes most commonly found in animal mtDNAs. Many features are most similar to those found for annelid mtDNAs, including A+T content, use of protein initiation codons, size and potential secondary structures of the largest non-coding region, and the relative arrangement of many genes. As in annelids examined to date, all genes are found on the same DNA strand. As noted for brachiopod mtDNA, there is a preference for G nucleotides to appear in tandem, without obvious explanation. Further description and comparison of complete mtDNA sequences will continue to produce a picture of genome evolution, particularly once sampling includes representatives of each animal phylum. Methods Molecular techniques A preparation of Urechis caupo total DNA was the kind gift of Eric Rosenthal. The entire mtDNA sequence was obtained using techniques detailed in [ 9 ]. Briefly, small fragments (450–710 nt) were amplified from cox1 , cob , and rrnS using primer pairs HCO 2198/LCO 1490 [ 12 ], CytbF/CytbR [ 10 ], and 16SARL/16SBRH [ 13 ], respectively. The sequences of these fragments were determined using dye-terminator chemistry (PE Biosystems) on an ABI 377 automated DNA sequencer. Primers were then designed facing "out" from these fragments to amplify the intervening regions (~2.9 to ~8 Kb) using long-PCR protocols with rTth-XL polymerase (PE Biosystems) as in [ 9 ]. Sequences were determined from the ends of these long-PCR fragments, then internally by "primer walking". To ensure quality, all sequences were determined on both strands and base calls for all chromatograms were verified by eye. Gene annotation Genes encoding rRNAs and proteins were identified by matching nucleotide or inferred amino acid sequences to those of Lumbricus terrestris mtDNA [ 7 ]. Since it is not possible to precisely determine the ends of rRNA genes by sequence data alone, they were assumed to extend to the boundaries of flanking genes. Each protein gene start was inferred as the eligible initiation codon nearest to the beginning of its alignment with homologous genes that does not cause overlap with the preceding gene. In five cases, an abbreviated stop codon was inferred where cleavage of a downstream tRNA from the transcript would leave a partial codon of T or TA, such that subsequent mRNA polyadenylation could generate a TAA stop codon. In each case an extension of this gene to the first in frame stop codon would cause overlap with the downstream tRNA. Genes for tRNAs were identified generically by their ability to fold into a cloverleaf structure and specifically by anticodon sequence. Abbreviations cox1 , cox2 , cox3 , cytochrome oxidase subunit I, II, and III protein genes; cob , cytochrome b gene; atp6 , atp8 , ATP synthase subunit 6 and 8 genes; nad1 , nad2 , nad3 , nad4 , nad4L , nad5 , nad6 , NADH dehydrogenase subunit 1–6, 4L genes; trnA , trnC , trnD , trnE , trnF , trnG , trnH , trnI , trnK , trnL1 , trnL2 , trnM , trnN , trnP , trnQ , trnR , trnS1 , trnS2 , trnT , trnV , trnW , trnY , transfer RNA genes designated by the one-letter code for the specified amino acid, with numerals differentiating cases where there are two tRNAs for the same amino acid.	/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC521484.xml
551604	Identification of a new family of putative PD-(D/E)XK nucleases with unusual phylogenomic distribution and a new type of the active site	Background Prediction of structure and function for uncharacterized protein families by identification of evolutionary links to characterized families and known structures is one of the cornerstones of genomics. Theoretical assignment of three-dimensional folds and prediction of protein function even at a very general level can facilitate the experimental determination of the molecular mechanism of action and the role that members of a given protein family fulfill in the cell. Here, we predict the three-dimensional fold and study the phylogenomic distribution of members of a large family of uncharacterized proteins classified in the Clusters of Orthologous Groups database as COG4636. Results Using protein fold-recognition we found that members of COG4636 are remotely related to Holliday junction resolvases and other nucleases from the PD-(D/E)XK superfamily. Structure modeling and sequence analyses suggest that most members of COG4636 exhibit a new, unusual variant of the putative active site, in which the catalytic Lys residue migrated in the sequence, but retained similar spatial position with respect to other functionally important residues. Sequence analyses revealed that members of COG4636 and their homologs are found mainly in Cyanobacteria, but also in other bacterial phyla. They undergo horizontal transfer and extensive proliferation in the colonized genomes; for instance in Gloeobacter violaceus PCC 7421 they comprise over 2% of all protein-encoding genes. Thus, members of COG4636 appear to be a new type of selfish genetic elements, which may fulfill an important role in the genome dynamics of Cyanobacteria and other species they invaded. Our analyses provide a platform for experimental determination of the molecular and cellular function of members of this large protein family. Conclusion After submission of this manuscript, a crystal structure of one of the COG4636 members was released in the Protein Data Bank (code 1wdj; Idaka, M., Wada, T., Murayama, K., Terada, T., Kuramitsu, S., Shirouzu, M., Yokoyama, S.: Crystal structure of Tt1808 from Thermus thermophilus Hb8, to be published ). Our analysis of the Tt1808 structure reveals that we correctly predicted all functionally important features of the COG4636 family, including the membership in the PD-(D/E)xK superfamily of nucleases, the three-dimensional fold, the putative catalytic residues, and the unusual configuration of the active site.	Background The PD-(D/E)XK domain is ubiquitously found in enzymes involved in metabolism of nucleic acids, mostly in nucleases with diverse biological functions. The first structurally characterized members of the PD-(D/E)XK superfamily were restriction enzymes (REases) (reviews: [ 1 , 2 ]). Crystallographic studies revealed that this superfamily groups together many nucleases with different cellular functions, including: phage λ exonuclease [ 3 ], bacterial enzymes exerting ssDNA nicking in the context of methyl-directed and very-short-patch DNA repair: MutH [ 4 ] and Vsr [ 5 ], Tn7 transposase TnsA [ 6 ], a family of archaeal Holliday junction resolvases (Hjc and Hje) from different species of Archaea [ 7 - 9 ], a Holliday junction resolvase (endonuclease I) from phage T7 [ 10 ], and an archaeal XPF/Rad1/Mus81 family nuclease that cleaves branched structures generated during DNA repair, replication, and recombination [ 11 ]. All members of the PD-(D/E)XK superfamily share a common structural core, comprising a mixed β-sheet of 4 or 5 strands flanked on both sides by α-helices [ 1 , 2 , 12 ]. These secondary structures are often embedded in very different peripheral elements, which sometimes constitute the majority of the protein. The common β-sheet serves as a scaffold for a weakly conserved active site, typically comprising two or three acidic residues (Asp or Glu) and one Lys residue, which together form the hallmark bipartite catalytic motif (P) D ...X n ...( D/E )X K (where X is any amino acid). The Lys residue serves to position a water molecule for an in-line attack on the scissile phosphodiester bond, while the carboxylate residues coordinate a Mg 2+ ion, which acts as a cofactor. Despite the wealth of structural and biochemical data, obtained mainly for REases (summarized in a collection of reviews: [ 13 ]), there is still controversy over the exact catalytic mechanism and the number of metal ions required (1, 2, or 3) by PD-(D/E)XK nucleases [ 14 , 15 ]. Moreover, it was found that some members of the PD-(D/E)XK superfamily developed different variants of the active site. In Vsr and its homologs, the (D/E)XK half-motif was replaced by "FxH" and an additional, unique catalytic His residue appeared in another part of the common three-dimensional fold [ 5 ]. In some REases, the acidic residue from the (D/E)XK half-motif was found to have "migrated" to another region of the polypeptide in a way that the position of the carboxylate group in the active site is generally maintained as in the "orthodox" members of the PD-(D/E)XK superfamily, despite the side chain is attached to another place in the backbone [ 16 - 19 ]. In a few enzymes, the conserved Lys was found to be replaced by a Glu, Gln, or Asn residue [ 20 - 22 ]. Crystallographic analyses have also revealed the PD-(D/E)XK fold in proteins that do not function as deoxyribonucleases at all and exhibit no conservation of the active site with the above-mentioned enzymes. The structure of the C-terminal catalytic domain of tRNA splicing endoribonuclease (RNase) EndA is identical to the minimal core of the PD-(D/E)XK fold [ 23 ], yet this protein lacks the Mg 2+ binding site common to its cousins that cleave phosphodiester bonds in DNA. Remarkably, on the opposite side of the common fold, EndA developed a different active site, whose geometric configuration is very similar to that of a His-Tyr-Lys triad in structurally unrelated RNase A [ 24 ]. Finally, the N-terminal domain (NTD) of the RPB5 subunit of RNA polymerase from Saccharomyces cerevisiae exhibits perfect conservation of the restriction enzyme-like structure, but lacks any catalytic residues – it is postulated that it functions as a nucleic acid binding domain devoid of any catalytic activity [ 25 ]. The divergence exhibited by the members of the PD-(D/E)XK superfamily is remarkable. Even enzymes with very similar biological functions, such as REases that recognize and cleave the same substrate, can exhibit little or no significant sequence similarity. Thus, most of the afore-mentioned enzymes were considered unrelated until the corresponding crystal structures were solved. Only in a few cases the membership in the PD-(D/E)XK superfamily was successfully predicted using bioinformatics (in some cases backed up by mutagenesis of hypothetical catalytic residues) before the actual structures were determined [ 26 - 29 ]. The catalogue of members of the PD-(D/E)XK superfamily is therefore far from being complete and it is expected that new lineages will be discovered as new sequences appear in the databases. Here, we predict that a large uncharacterized protein family with an unusual phylogenetic distribution is likely to represent a new branch of PD-(D/E)XK nucleases. Results Sequence analysis of COG4636 reveals remote similarity to PD-(D/E)XK nucleases In the course of analyses of proteins with unknown structures, we came across a family of sequences grouped together in the Clusters of Orthologous Groups (COG) database [ 30 ] as COG4636 and annotated as "uncharacterized protein conserved in Cyanobacteria". Analyses of cross-references to other databases revealed no functional information about any member of this family. Nonetheless, preliminary analysis of sequence conservation combined with secondary structure prediction revealed a characteristic pattern of α-helices and β-strands associated with conserved carboxylate residues (review: [ 31 ]), which suggested that members of COG4636 may belong to the PD-(D/E)XK superfamily (Figure 1 ). The multiple sequence alignment revealed nearly perfect conservation of a "PD" half-motif, but only partial conservation of the "(D/E)XK" half-motif. Specifically, instead of the Lys residue most members of COG4636 possessed a hydrophobic amino-acid, such as Leu or Val. This suggested that the apparent similarity to the pattern of catalytic residues typical for the PD-(D/E)XK superfamily may be either spurious or indicate a new family of enzymes with an active site devoid of the otherwise conserved residue. We searched for homologs of the analyzed family, beyond sequences from complete genomes grouped together in COG4636, by carrying PSI-BLAST searches of the nr database. Altogether, we collected 435 sequences with significant similarity to COG4636, which will be hereafter referred to as "COG4636+". No statistically significant sequence similarity was detected to any protein with an experimentally determined function. Figure 1 Multiple sequence alignment of selected representatives of the extended COG4636+ family. The selection of representative sequences includes the modeled protein from Nostoc (motif H-PD-EXX-K, members from G. violaceus with different order of putative catalytic residues (Gv1: H-PD-EXK; Gv2: S-PD-EXD-K; Gv3: H-PD-EXD; Gv4: H-PD-EXX-N; Gv5: Q-PD-EXX-K), and members of mono-phyletic clusters from D. hafniense , C. aurantiacus , S. coelicolor , T. thermophilus , and G. violaceus ). The positions of putative catalytic residues are labeled with "". The variable termini, which could not be confidently aligned, are not shown; the number of omitted residues is indicated. A complete alignment of full-length sequences is available for download from . Amino acids are colored according to the physico-chemical properties of their side-chains (negatively charged: red, positively charged: blue, polar: magenta, hydrophobic: green). Conserved residues are highlighted. Elements of predicted secondary structure (helices and strands) are indicated by tubes and arrows, respectively. In order to test the hypothesis of the evolutionary connection between COG4636 and the PD-(D/E)XK superfamily we carried out the fold-recognition analysis, which allows to predict the three-dimensional fold of the target protein by matching its sequence with the available protein structures and assessing the sequence-structure compatibility using a combination of criteria, such as sequence similarity, match of secondary structure elements, compatibility of residue-residue contacts, etc. (review: [ 32 ]). Sequences of individual members of COG4636 were therefore submitted to the GeneSilico protein fold-recognition metaserver [ 33 ]. Disappointingly, no methods reported statistically significant matches between these sequences and proteins with known structures. Only a few threading methods that explicitly use the structural information from the templates (FUGUE, INBGU, mGenTHREADER, SAM-T02, and 3DPSSM) reported, in some cases, matches to structures of PD-(D/E)XK nucleases, but never at the first position of the ranking. However, in the course of CASP-5 protein structure prediction contest we found that the fold-recognition operation for strongly diverged proteins can be greatly improved by limiting the analysis to the conserved core, i.e. omission of strongly diverged regions and non-conserved insertions, as well as using a refined multiple sequence alignment rather than allowing the servers to build their own sequence profiles from unrefined PSI-BLAST results [ 34 ]. Thus, we modified the multiple sequence alignment of the COG4636+ family by removing strongly diverged termini that could not be reliably aligned, and submitted to the meta-server only the core section, comprising ca. 110 aa. This time, as expected, fold-recognition analysis of a well-defined protein core gave unambiguous results: mGenTHREADER, SPARKS, and FUGUE reported structures of Holliday junction resolvases Hjc and Hje, members of the PD-(D/E)XK fold [ 7 - 9 ], at the first positions of their rankings, with significant scores (0.45, -2.08, and 3.46, respectively). Results obtained from the primary servers have been supported by the consensus server Pcons [ 35 ], which reported the Hjc and Hje enzymes at the first four position of its ranking, with scores 1.38-1.20, compared to the insignificant score 0.61 for the subsequent fold in the ranking. Modeling and model-based identification of a putative active site In order to identify the putative active site of newly predicted members of the PD-(D/E)XK superfamily, we modeled the structure of one of the COG4636+ members, whose sequence was close to the consensus calculated for the whole family (hypothetical protein all3650 from Nostoc sp. PCC 7120, GI: 17231142) and used it as a platform to study the three-dimensional arrangement of conserved residues. A homology model of all3650 was constructed using the "FRankenstein's Monster" approach (see Methods and ref. [ 34 ]), starting with the unrefined alignments between the consensus sequence and the structures of Hjc and Hje enzymes (1gef, 1hh1, and 1ob8) reported by threading methods. Initially, the model of the protein core was constructed by iterating the homology modeling procedure, evaluation of the sequence-structure fit by VERIFY3D, merging of fragments with best scores, and local realignment in poorly scored regions. Local realignments were constrained to maintain the overlap between the secondary structure elements found in the template structures, and those predicted for the target. This procedure was stopped when the regions in the protein core (helices and strands) obtained acceptable VERIFY3D score (>0.3) or their score could not be improved by any manipulations, while the average VERIFY3D score for the whole model could not be improved. The final alignment between all3650 and the three structures used as templates is shown in Figure 2 . The final model of the core, comprising residues 39–188, obtained a poor average VERIFY3D score of 0.13 due to low scores in the variable loops that could not be modeled with confidence. However, the secondary structure elements (with the exception of the C-terminal helix), obtained an acceptable average score of 0.37. It is important to note that all catalytic residues of the PD-(D/E)XK fold are found in the stable regions of regular secondary structure rather than in loops [ 36 ]. The variable N-terminus, which could not be modeled because of the strong divergence and the lack of appropriate template structures, was added "de novo" using the fragment insertion method ROSETTA [ 37 ]. The coordinates of the final, full-length model (Figure 3 ) are available as supplementary material [see Additional file 1 ] and on-line at Figure 2 Fold-recognition alignment between all3650 and structures of Hjc and Hje. Amino acids are colored according to the physico-chemical properties of their side-chains. Conserved residues are highlighted. Secondary structure elements experimentally identified in Hjc and Hje and predicted for all3650 are shown between the target and the template sequences. Known and predicted catalytic residues are indicated by "" (above the alignment for the target, below the alignment for the templates). Figure 3 Homology model of all3650. Helices and strands are shown in green and yellow, respectively. The predicted catalytic residues are shown in the wireframe representation and labeled. The termini are indicated. The model of all3650 reveals a typical PD-(D/E)XK nuclease-like spatial arrangement of one Lys ε-amino group (from the residue K127) and two carboxylate groups (from residues D83 and E111) (Figure 4 ). The modeled structure suggest also an additional highly conserved His residue (H43) that could be a part of the metal ion-binding site or be involved in substrate-binding. Strikingly, in all3650 as well as in the great majority of sequences from the COG4636+ family, the conserved Lys (K127 in all3650) is found not in the common position in the same β-strand as the conserved Glu residue (E111 in all3650), but in a spatially adjacent α-helix. Thus, the predicted active site is formed by a "PD-EXX-K" sequence motif. This "migration" of the presumptive catalytic Lys residue and retention of the original position of the spatially adjacent carboxylate in COG4636+ members resembles the situation reported for a number of restriction enzymes such that as Cfr10I, NgoMIV, Ecl18kI, SsoII, and PspGI [ 16 , 18 , 19 , 38 ]. In the latter enzymes, however, it is the carboxylate that is relocated and the original position of the Lys residue is retained, such the active site is formed by a "PD-XXK-E" sequence motif (Figure 4 ). Figure 4 Spatial conservation of the PD-(D/E)XK active site in all3650, Hjc, and NgoMVI. A) The predicted structure of all360 is shown in the same orientation as the crystal structures of the bona fide PD-(D/E)XK nucleases: B) Holliday junction resolvase Hje (1ob8 in PDB [9]) and C) REase NgoMIV (1fiu in PDB [78] to illustrate the spatial conservation of side-chains in the active site (the carboxylate residues in red and the Lys residue in blue), despite the lack of their conservation in the PD-EXX-K, PD-DXK, and PD-XXK-E variants of the sequence motif. Only the common core is shown, terminal regions and insertions have been omitted for clarity of the presentation. Inspection of the multiple sequence alignment reveals that only two carboxylates (corresponding to D83 and E111 in all3650) are practically invariant in the COG4636+ family, while all the others undergo various substitutions (Figure 1 ). In a small group of sequences (represented by a hypothetical protein gll0909 from G. violaceus , GI: 37520478) the Lys residue is present both at the "classical" and alternative position, thereby forming a "PD-EXK-K" variant of the active site. This arrangement resembles a putative evolutionary intermediate between the "classical" active site and the newly discovered rearranged variant. In another lineage of the COG4636+ family, an Asp residue appears in the position normally occupied by Lys in the C-terminal half-motif. Some of the members of this lineage (exemplified by glr2344 from G. violaceus , GI: 37521913) exhibit therefore the "PD-EXD-K" motif, but the majority (exemplified by hypothetical protein glr1284 from G. violaceus , GI: 37520853) lack the Lys residue and exhibit only the "PD-EXD" variant. In another lineage (represented by gll1896 from G. violaceus , GI: 37521465) the Lys residue is replaced by Asn to form the "PD-EXX-N" variant of the predicted active site. The conserved His residue (H43 in all3650) is present in most members of the COG4636+ family, with the exception of a small lineage of closely related proteins (represented by gll1896 from G. violaceus , GI: 37520579) in which it is substituted by Gln, and a larger group of more diversified sequences, in which it is substituted by Thr or Ser. Most members of the latter group possess a Lys or Arg residue in the "catalytic" position and hence exhibit "PD-EXK-K" (see above) or "PD-EXR-K" variants of the active site. It will be very interesting to determine experimentally, which of those residues in different configurations are involved in catalysis, and which are only auxiliary. In particular, it would be interesting to find if both or either of the Lys residues present in the potential "intermediate" versions of the active site are required for catalysis. Phylogenomic analysis of the COG4636+ family Sequence searches of the nr database at the NCBI revealed that the great majority of members of the COG4636+ family (382 of total 435) originate from Cyanobacteria; of these, 84% were found in just 6 genomes ( G. violaceus PCC 7421, Nostoc punctiforme PCC 73102, Crocosphaera watsonii WH 8501, Nostoc sp. PCC 7120, Anabaena variabilis ATCC 29413, Synechocystis sp. PCC 6803). It is astonishing that members of COG4636+ represent over 2% of all protein-encoding genes of G. violaceus PCC 7421 (95 of 4430 total [ 39 ]), other completely sequenced genomes of Cyanobacteria are completely devoid of them or encode only 1 or 2 sequences from this family. We were not able to identify any members of the COG4636+ family in the sequences derived from seawater samples collected from the Sargasso Sea [ 40 ] and deposited in the "environmental samples" database at the NCBI. Since the prevalent Cyanobacteria found in the Sargasso Sea are Synechococcus and Prochlorococcus , the lack of COG4636+ members in the environmental samples is in good agreement with the paucity of these genes in the fully sequenced genomes of these species. In order to reconstruct the evolutionary history of the COG4636+ family, we calculated the phylogenetic tree, based on the same reliable section of the multiple sequence alignment that was used for protein structure prediction (see Methods). Unfortunately, in all trees obtained with different methods and parameters, the majority of deep branches received very low bootstrap support (data not shown), hence the relationships within the whole family must be regarded as unresolved. We were able, however, to identify a number of branches with bootstrap support >90%. Many of such branches comprise members from one species only. This situation is characteristic for sequences found in a few non-Cyanobacterial species; for instance 8 sequences from D. hafniense DCB-2 (Firmicutes), 7 sequences from C. aurantiacus (Chloroflexales), and 6 from S. coelicolor (Actinobacteria) each form a separate species branch on the phylogenetic tree, while 14 sequences from T. thermophilus HB27 (Deinococcus-Thermus lineage) form three separate branches. Several monophyletic groups of closely related sequences are also observed in G. violaceus (e.g. a sub-family comprising 7 sequences with GI numbers: 37522824, 37520777, 37522646, 37521452, 37522233, 37520151, 37522558). There is also one branch comprising 6 closely related sequences in C. watsonii , GI numbers: 45527153, 45527776, 45524526, 45527777, 45527775, 45527774). Other statistically significant branches, however, comprise members from different species, suggesting that they were either formed prior to speciation or that their members were transmitted horizontally between different genomes of already existing species. To identify if members of COG4636+ are encoded by any known mobile genetics elements or if they are preferentially associated with any other proteins, we analyzed the genomic neighborhood of all members of the family. Although we carefully examined annotations of predicted open reading frames (ORFs) in the range of 3000 bp upstream and downstream, we weren't able to identify any recurrent type of proteins, either with respect to the molecular or cellular function or the predicted three-dimensional fold (data not shown). Also no preference for occurrence of COG4636+ family members within or near any apparent mobile genetic elements (putative prophages etc.) was observed. Thus, insertion of the genes encoding putative COG4636+ nucleases seems virtually random. The only notable exception is a neighborhood of another member of COG4636+, suggesting tandem duplication. We identified one instance of 4 consecutively arranged genes in the genome of C. watsonii WH8501, all from the above-mentioned branch of 6 closely related sequences (the other two relatives are located elsewhere on the chromosome). We also found a few tandem duplications: 9 in C. watsonii WH8501 and 5 in G. violaceus PCC7421, 5 in Nostoc sp. PCC6803, 2 in N. punctiforme PCC73102, 2 in A. variabilis ATCC 29413, 2 in Synechocystis sp. PCC6803, 2 in T. thermophilus HB27, 1 in T. erythraeum IMS101 and 1 in M. magnetotatcticum MS-1. In general, however, tandem duplications are rare and the distribution of COG4636+ family members along the chromosomes of Cyanobacteria with completed genomes seems completely erratic (Figure 5 ). Figure 5 Localization of COG4636+ family members in the chromosomes of Cyanobacteria with completed genomes. Circular chromosome maps of genomes with at least three genes encoding COG4636+ members (indicated by dots). Genes shown in dark blue are transcribed clockwise (positive reading frame) and those in red are transcribed anticlockwise (negative reading frame). Dots plotted inside the circle indicate that more than one gene is localized in the same region of the map (1/360 of the genome length). Discussion Our results suggest that functionally uncharacterized proteins grouped together in COG4636 are a branch of the PD-(D/E)XK superfamily, which has not been identified to date due to a presence of an unusual variant of the active site, which lacks the conserved Lys residue at the typical position in the primary sequence. That the catalytic Lys can migrate in the framework of the active site of PD-(D/E)XK nucleases has been suggested earlier, based on the sequence analysis of another nuclease domain found in site-specific, non-long terminal repeat retrotransposable elements [ 2 ], but to date no molecular model was offered to suggest the alternative point for the attachment of the side chain to the protein backbone. Our sequence analysis of the COG4636+ family and the structural model of one of its members explain the problems with identification of the PD-(D/E)XK motif on the sequence level and provide a platform for further studies. Specifically, our analysis points at the most interesting members of the family, which display previously not observed variants of the PD-(D/E)XK active site. Experimental analyses of these proteins and determination of the role of individual amino acids in the evolutionary context may help to better understand the plasticity of the PD-(D/E)XK active site and may settle down the controversy in the field of nucleases regarding the mechanism(s) of the reaction. Phylogenomic analyses show that putative nucleases grouped in the COG4636+ family are exceptionally abundant in genomes of certain Cyanobacteria, but absent in others. They are typically abundant in the sequenced genomes of freshwater species, but scarce in the genomes of marine species, with the exception of C. watsonii WH 8501, which was isolated from tropical waters of the Western Atlantic and Pacific oceans. It is remarkable that members of COG4636+ are almost absent from the genomes of Synechococcus and Prochlorococcus species thriving in the Sargasso sea, as well as in the environmental samples isolated from that region. On the other hand, in G. violaceus PCC 7421 they comprise over 2% of all protein-encoding genes. This phylogenetic distribution resembles that of mobile genetic elements such as introns or insertion sequences (reviews: [ 41 , 42 ]) and suggests that the contemporary COG4636+ family originated from a few predecessors that underwent extensive horizontal gene transfer and massive proliferation in certain genomes. Monophyly of COG4636+ sequences in non-Cyanobacterial species strongly suggests that proliferation occurred in each of these species independently, following a single event of colonization by horizontal transfer from a Cyanobacterium (or in the case of T. thermophilus – three independent successful colonizations). We hypothesize that the mechanism by which these putative nucleases induce their proliferation in a genome is similar to that displayed by homing nucleases and restriction enzymes [ 43 ], namely to incise the DNA by introducing nicks or double-strand breaks, which stimulates recombination and may lead to tandem duplications and a variety of genomic rearrangements [ 44 - 47 ]. Frequent cleavage of the genomic DNA would be lethal for the cell, therefore if members of COG4636+ are indeed active as nucleases, then they should target rare sequences (in a manner similar to homing endonucleases; review: [ 48 ]) or unusual structures in the DNA (similarly to the structure-specific Holliday junction resolvases), or their activity would have to be somehow regulated (inhibited) by interactions with other proteins or cellular processes (for instance by DNA modification). There are known examples of Holliday junction resolvases carried on defective lambdoid prophages [ 49 ]. Unfortunately, analysis of the genomic neighborhood shows no preferred association of COG4636+ members with any mobile genetic elements or particular gene families that could give us hints about the cellular processes they could be part of or suggest how their predicted nuclease activity could be inhibited or regulated. Especially, we found no correlation with the presence of known or putative methyltransferases. This suggests that despite sharing the common PD-(D/E)XK fold with REases, COG4636+ members are unlikely to serve as parts of restriction-modification systems, which are known to be abundant in Cyanobacteria [ 50 , 51 ]. It must be noted, however, that multiple solitary DNA methyltransferases were reported in Anabaena PCC 7120 [ 51 ], and these enzymes could potentially provide protection against the cleavage of the chromosomal DNA by at least some of the COG4636+ members found in this organism. One possibility is that COG4636+ members serve as a part of the restriction barrier, similarly to the unrelated NucA family of extracellular nucleases found in Cyanobacteria, e.g. Anabaena sp. PCC 7120 [ 52 ] and Microcystis sp. [ 53 ]. They could also fulfill a role in maintenance of the identity of the species by controlling the flow of incoming DNA, as recently suggested for restriction-modification systems [ 54 ]. From the genomic analyses it appears, however, that the primary function of COG4636+ members is to spread and multiply, and their cellular roles may be merely side-effects of this selfish expansion. It is very likely that their nuclease activity is recombinogenic and may increase the frequency of genomic rearrangements. Moreover, the multiplication of closely related COG4636+ members in certain genomes leads to an abundance of dispersed related DNA sequences, which by themselves may increase the frequency of genome rearrangements by homologous recombination. It was suggested that in the marine Cyanobacteria the factors that increase the genome plasticity might not be promoted by natural selection due to the homeostatic environment of the open ocean [ 55 ]. Conversely, the unstable environment of fresh waters might promote the spreading of factors that destabilize the genome by increasing the frequency of recombination and thereby increase the diversity of the population. This is in good agreement with our finding of prevalence of COG4636+ members in Cyanobacteria that thrive in fresh waters and their paucity in marine species (with the exception of C. watsonii WH 8501). Summarizing, it is plausible that members of COG4636+ fulfill an important role in the genome dynamics of Cyanobacteria and other species they colonize. We hope that our predictive study will facilitate experimental determination of the molecular and cellular function of members of this intriguing protein family. Methods Sequence analysis Searches of the non-redundant (nr) database were carried out at the NCBI using PSI-BLAST [ 56 ] with default parameters, using different sequences from COG4636 as queries. Significantly similar sequences were retrieved from all searches and pooled together. Identical sequences from the same organism were removed. A multiple sequence alignment was generated using MUSCLE [ 57 ] with default parameters and subsequently adjusted manually, based on the analysis results of secondary structure prediction (see below), to ensure that no unwarranted gaps are introduced within α-helices and β-strands. Phylogenetic inference was carried out using the reliable central section of the multiple sequence alignment. The matrix of pairwise distances was calculated from sequences according to the JTT model [ 58 ] with gaps treated as missing data. The neighbor-joining (NJ) tree was inferred according to the method of Saitou and Nei [ 59 ]. Phylogenomic analysis The Eutils module from the Biopython package was used as an interface to access remotely the NCBI databases [ 60 ]. The Gene Identification numbers of proteins included in the final multiple alignment sequences were used to identify the corresponding GenPept entries, which were downloaded into a local Barkeley database using an in-house developed parser based on the SAX package . The "coded_by" field from each GenPept file was used to identify the corresponding DNA sequence, which were also downloaded into the database. The sequence in the range of 3000 bp upstream or downstream from the region encoding a COG4636+ member were scanned for the presence of annotated Open Reading Frames (ORFs). Initially, the functional categorization of these ORFs was carried out based on the automatic assignment into the PFAM and COG families. In the absence of any recurrent function, the annotations of all ORFs were carefully re-analyzed visually and in uncertain cases, additional searches against the CDD database were carried out [ 61 ]. The distribution of COG4636+ members on the chromosome maps was visualized using a program developed in-house specifically for that purpose. Protein structure prediction Secondary structure prediction and tertiary fold-recognition was carried out via the GeneSilico meta-server gateway at [ 33 ]. Secondary structure was predicted using PSIPRED [ 62 ], PROFsec [ 63 ], PROF [ 64 ], SABLE [ 65 ], JNET [ 66 ], JUFO [ 67 ], and SAM-T02 [ 68 ]. Solvent accessibility for the individual residues was predicted with SABLE [ 65 ] and JPRED [ 66 ]. The fold-recognition analysis (attempt to match the query sequence to known protein structures) was carried out using FFAS03 [ 69 ], SAM-T02 [ 68 ], 3DPSSM [ 70 ], BIOINBGU [ 71 ], FUGUE [ 72 ], mGENTHREADER [ 73 ], and SPARKS [ 74 ]. Fold-recognition alignments reported by these methods were compared, evaluated, and ranked by the Pcons server [ 35 ]. Homology modeling Fold-recognition alignments to the structures of selected templates were used as a starting point for homology modeling using the "FRankenstein's Monster" approach [ 34 ], comprising cycles of model building, evaluation, realignment in poorly scored regions and merging of best scoring fragments. The positions of predicted catalytic residues and secondary structure elements were used as spatial restraints. Briefly, preliminary models were generated based on the alignments to various template structures returned by the FR servers. The sequence-structure fit in these models was assessed using VERIFY3D [ 75 ] and visualized using the COLORADO3D server [ 76 ]. The most common and best-scoring fragments were merged to produce a hybrid model, in which the sequence-structure was re-evaluated. In the poorly scoring fragments the alignment was locally modified by shifting the sequences within the limits of predicted secondary structures and a next generation of models corresponding to different alignments was generated. The cycles of evaluation of models, generation of hybrids and local re-alignment in problematic regions continued until the global VERIFY3D score could not be improved. Regions, which could not be modeled because of the lack of the appropriate template structure, were added "de novo" using the fragment insertion method ROSETTA [ 37 ]. Note added in Proof After submission of this manuscript, a crystal structure of one of the COG4636+ members was released in the Protein Data Bank (code 1wdj; Idaka, M., Wada, T., Murayama, K., Terada, T., Kuramitsu, S., Shirouzu, M., Yokoyama, S.: Crystal Structure of Tt1808 from Thermus thermophilus Hb8 To be Published). Our analysis of the Tt1808 structure and its comparison with the model of all3650 confirms our predictions. Tt1808 does indeed exhibit the PD-(D/E)xK fold: the DALI [ 77 ] search of the the Protein Data Bank (PDB) database with 1wdj revealed that its 8 closest structural matches with Z-scores in a range of 5.3-3.7 are members of the PD-(D/E)xK superfamily, including the Holliday junction resolvases we used as templates to model the all365 protein. Analysis of the Tt1808 structure (Figure 6 ) reveals that we correctly predicted the topology of the catalytic domain in all365. We only mispredicted an α-helix in the C-terminus of all365; in Tt1808 this element is replaced by a β-hairpin. We have also successfully modeled the structure of the N-terminal subdomain but failed to predict the interaction between this part and two loops of the catalytic domain (compare Figure 3 and Figure 6 ). It is important to note that these errors concern regions that do not influence any of our functional interpretations based on the all3650 model. Most importantly, the identity of presumed catalytic residues of all365 was predicted correctly, including the postulated unusual position of the Lys residue (in our model of all365 the side chain of K127 has a different orientation than K130 in Tt1808, but such details are irrelevant to our functional interpretations). It is interesting to note that Tt1808 has the S-PD-EXR-K variant of the active site, and that the side chain of the R118 residue, which replaced the "classical" catalytic Lys, points away from other catalytic residues, on the opposite side of the loop between the "EXR" and "K" elements. Summarizing, we correctly predicted all functionally important features of the COG4636+ family, including the membership in the PD-(D/E)xK superfamily of nucleases, the three-dimensional fold, the putative catalytic residues, and the unusual configuration of the active site. Figure 6 The crystal structure of Tt1808 (1wdj in PDB). Tt1808 is shown in the same orientation and is colored and labeled in the same way as the homology model of all3650 on Figure 3. Two regions of differences between Tt1808 and the model of all3650 are indicated: the N-terminal subdomain has a similar fold, but different orientation (magenta line) and the C-terminal region folds as a β-harpin (cyan line) rather than as an α-helix. List of abbreviations aa, amino acid(s); bp, base pair(s); nt, nucleotide; e, expectation; REase, restriction endonuclease; ORF, product of an open reading frame, Authors' contributions MF carried out all sequence analyses and structure predictions using fold-recognition methods and ROSETTA. JMB built the homology model, analyzed spatial vs. sequential conservation of the putative active site, and wrote the manuscript. Both authors have read and accepted the final version of the manuscript. Table 1 Distribution of COG4636+ family members among different bacteria. organism / genome phylum habitat data source COG4636+ members total disrupted Gloeobacter violaceus PCC 7421 Cyanobacteria calcareous rock C 95 1 Nostoc punctiforme PCC 73102 Cyanobacteria cycad (endosymbiont) WGS 71 7 Crocosphaera watsonii WH 8501 Cyanobacteria marine water WGS 62 1 Nostoc sp. PCC 7120 Cyanobacteria fresh water C 58 1 Anabaena variabilis ATCC 29413 Cyanobacteria fresh water WGS 45 5 Synechocystis sp. PCC 73102 Cyanobacteria fresh water C 36 1 Thermus thermophilus HB27 Deinococcus-Thermus thermal environment C 14 - Trichodesmium erythraeum IMS101 Cyanobacteria marine water WGS 10 3 Desulfitobacterium hafniense DCB-2 Firmicutes sewage sludge WGS 8 - Chloroflexus aurantiacus Chloroflexi fresh water (hot springs) WGS 7 - Streptomyces coelicolor A3(2) Actinobacteria soil C 6 - Rhodopirellula baltica SH 1 Planctomycetes marine water C 5 1 Moorella thermoacetica ATCC 29413 Firmicutes fresh water (ponds) WGS 3 - Deinococcus radiodurans R1 Deinococcus-Thermus unknown C 3 - Magnetospirillum magnetotacticum MS-1 Proteobacteria fresh water (ponds) WGS 2 1 Synechococcus elongatus PCC 73102 Cyanobacteria fresh water WGS 2 - Aquifex aeolicus VF5 Aquificae fresh water (hot springs) C 2 - Kineococcus radiotolerans SRS30216 Actinobacteria unknown (isolated from radioactive work area) WGS 2 - Caulobacter crescentus CB15 Proteobacteria fresh water C 1 - Thermosynechococcus elongatus BP-1 Cyanobacteria fresh water (hot springs) C 1 - Synechococcus sp. PCC 73102 Cyanobacteria brackish (euryhaline) and/or marine water UGS 1 - Microcystis aeruginosa Cyanobacteria fresh water (lakes, ponds and rivers) NR 1 - Prochlorococcus marinus str. MIT9313 Cyanobacteria marine water C - - Prochlorococcus marinus subsp. marinus CCMP1375 Cyanobacteria marine water C - - Prochlorococcus marinus subsp. pastoris CCMP1986 Cyanobacteria marine water C - - Synechococcus sp. WH 8102 Cyanobacteria marine water C - - C – Completed genomic sequence, WGS – Whole Genome Shotgun, UGS – Unfinished Genomic Sequence, NR – non-redundant database (NCBI). ORFs were regarded as "disrupted" if they bear frameshift mutations or stop codons. Supplementary Material Additional File 1 The additional data file all3650.pdb contains the coordinates of the original all3650 model (obtained before the Tt1808 structure was published) in the PDB format. Click here for file	/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC551604.xml
509320	Deterministic Tumor Evolution	null	The essential difference between cancer cells and normal cells is that cancer cells evolve. Most cancers arise from a single cell through a sequential evolutionary process of mutation and selection. Cancer cells harbor mutations in a number of critical genes that, at various stages during the evolution of the tumor, provide those cells with a selective advantage. Many of the phenotypes, or physical outcomes, conferred by these mutant genes are subverted from a normal cell's repertoire, including proliferation, invasion, migration, loss of differentiation, and loss of apoptosis (programmed cell death); other phenotypes, such as immortalization, are novel. Tumor evolution is thought to adhere to Darwinian principles, with mutations arising randomly within an individual cell, followed by selection for mutant clones with favorable traits. Support for this idea stems from the observation that end-stage tumors have mutations in a number of genes. But linking mutations in particular genes with defined stages is difficult for most human cancers and especially so for the last and deadliest stage: when cancer disseminates throughout the body during metastasis. It's unclear, for example, whether there are mutations in particular genes or sets of genes that enhance metastasis. That is, do metastatic lesions develop through a continuation of Darwinian evolution, or is metastasis an intrinsic property of the primary tumor, meaning that further genetic evolution is not required? It is also unclear whether there is a “preferred” sequence of mutations, such that selective pressure for particular mutations depends on preexisting mutations. Squamous cell carcinoma invading muscle layer from ARF -null mouse Since the earliest days of research on oncogenes—genes that can cause a cell to become cancerous—it has been known that certain oncogenic mutations cooperate to transform normal cells into cancer cells. For example, an “activating” mutation in the oncogene Ras and the loss of the tumor suppressor p53 cooperate to transform cells. The paper by Christopher Kemp and his colleagues at the Fred Hutchinson Cancer Research Center sheds light on some of these questions, and many of the issues center around the most notorious oncogene: Ras . Using a well characterized mouse model of squamous cell carcinogenesis, which generates a form of skin cancer, the authors examine both the functional and evolutionary relationships between three cancer genes that play major roles in most human cancers: Ras and the tumor suppressors Arf and p53 . Two seminal early observations set the stage: mutational activation of Ras is the initiating genetic event in this cancer model, while mutation of p53 occurs later, during the benign to malignant transition; and expression of mutant Ras in cells activates p53 via signaling through the protein encoded by Arf . Kemp et al. confirm that this pathway is active in “autochthonous” tumors—which grow and develop where they are initiated—by showing that p53 expression in tumors with Ras mutations is dependent on the presence of Arf . Thus, during the early benign stages of tumor growth, Ras activates Arf , which in turn activates p53 , thereby inhibiting tumor progression. This provides strong selective pressure in favor of cells with mutations in either Arf or p53 , and these mutations are indeed observed as the tumors progress to malignancy. That Arf and p53 function as tumor suppressors was confirmed by demonstrated accelerated tumor progression in mice lacking either Arf or p53 . This answers a longstanding question concerning the nature of the signal that activates p53 during autochthonous tumor development: Mutation of Ras not only initiates tumor development but, through its intracellular signaling through Arf and p53 , directly influences the subsequent evolutionary trajectory of the tumors. In this view, secondary evolutionary events are determined by the preexisting genetic lesion, as a result of direct signaling interactions. The authors go on to show that tumors lacking Arf or p53 show accelerated metastatic dissemination, a phenomenon rarely seen in mouse squamous cell cancer models. Thus both benign and malignant tumors lacking these tumor suppressors are at high risk for metastasis. As Ras is well known to confer many phenotypes required for the metastatic process, it appears that Ras , together with loss of its inhibitors, Arf and p53 , may be sufficient to drive this process. More direct evidence that metastasis does or does not require further genetic evolution awaits.	/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC509320.xml
552318	FlyPNS, a database of the Drosophila embryonic and larval peripheral nervous system	Background The embryonic and larval peripheral nervous system of Drosophila melanogaster is extensively studied as a very powerful model of developmental biology. One main advantage of this system is the ability to study the origin and development of individual sensory cells. However, there remain several discrepancies regarding the organization of sensory organs in each abdominal segment A1-A7. Description We have constructed a web site called FlyPNS (for Fly Peripheral Nervous System) that consolidates a wide range of published and unpublished information about the embryonic and larval sensory organs. It communicates (1) a PNS pattern that solves the discrepancies that have been found in the recent literature, (2) the correspondence between the different nomenclatures that have been used so far, (3) a comprehensive description of each sensory organ, and (4) a list of both published and unpublished markers to reliably identify each PNS cell. Conclusions The FlyPNS database integrates disparate data and nomenclature and thus helps understanding the conflicting observations that have been published recently. Furthermore, it is designed to provide assistance in the identification and study of individual sensory cells. We think it will be a useful resource for any researcher with interest in Drosophila sensory organs.	Background The Drosophila abdominal larval PNS is composed of a constant number of neurons and associated cells, whose characteristics and positions are reproducible between individuals [ 1 - 5 ]. Its stereotyped pattern has made it an ideal system to study many aspects of developmental biology such as cell determination, asymmetric cell divisions, cell lineages, development and remodeling of axons and dendrites, cell migration, and cell death. Despite the extraordinary utility of the PNS for understanding these diverse problems, a prominent current limitation is the lack of agreement in the literature of the exact number, position and nomenclature of sensory neurons. For example, many different names have been used to describe abdominal neurons, reflecting the number of cells within a cluster, their position [ 6 , 7 ], their dendritic arborization [ 8 , 19 ] or their cell lineage origin [ 9 ]; and the correspondence between names of cells has remained vague. Furthermore, there has been no consensus on the exact number of neurons located in the dorsal cluster, which is one of the most extensively studied regions of the PNS. First studies indicated that 10 [ 4 ], 11 [ 2 ], or 12 [ 4 , 6 ] neurons are present. Following these studies, this cluster was usually thought to contain 12 neurons. A single compilation of the different nomenclatures that accounts for all peripheral neurons would not only benefit those within the field, but should also make it easier for researchers outside of the immediate field to comprehend the literature. Construction and content We have constructed a website, named FlyPNS, that consolidates a wide range of published and unpublished information regarding the embryonic and larval sensory organs and their associated glial cells. Motoneurons and other glial cells have not been included. The FlyPNS web site is arranged in 6 sections: General references, Nomenclature, Abdominal PNS organization, Sensory organ description, Antibodies and enhancer-traps, Gal4 lines, and a Search option. The General references section provides a list of papers on various aspects of sensory organ morphology and development. Subcategories include general descriptions, descriptions of dendritic arborizations, axonal pathways, developmental changes, and functions of multidendritic neurons. Links are provided to PubMed entries for access to abstracts or full-texts of papers. In the Abdominal PNS organization section, we have represented the PNS pattern that we have observed (Fig. 1 ), with the notable presence of 13 neurons in the dorsal cluster (as also mentioned recently in [ 10 , 11 ]). In previous representations that indicated a smaller number of cells in the cluster, either the anterior-located Cut-negative neuron that we have named dmd1 (Fig. 1 ) or one of the Cut-positive multidendritic neurons may have been missed. Since we found that the precise position of sensory cells may vary with segment, embryo and developmental stage (unpublished data), the most reliable criteria for identification of Drosophila embryonic and larval sensory organs are marker expression and cell morphology, and these should be used whenever possible. Our study of the cell markers Cut, Collier, E7-2-36, E7-3-49, Elav, Engrailed and Nubbin/Pdm1 [ 11 - 13 ] and unpublished data), as well as neuron morphology [ 8 , 13 ], revealed that distinctive features can be attributed to each neuron, thereby allowing the unambiguous identification of every PNS neuron. The sections entitled Sensory organ description, Antibodies and enhancer-traps, and Gal4 lines relate these observations. It should also be noted that the precise geometry of the PNS pattern presented in the web site might vary somewhat depending on the stage examined and the techniques used to dissect and mount the preparation. Figure 1 Diagram of the sensory cells of an embryonic or larval abdominal hemisegment (A1-A7) . Three types of sensory organs are found: (1) external sensory organs, composed of three accessory cells (oval shape) and one or several neurons (circular shape), (2) chordotonal organs, composed of accessory cells (oval shape) and neurons (elongated triangular shape), (3) multidendritic neurons (diamond shape). The most conspicuous sensory dendrites are represented as a straight line. Note that the dmd1/dda1 neuron presents a unique dendritic arborization different from the other da neurons. Anterior is left and dorsal is up. The section on Nomenclature , presented in table form, lists all nomenclatures that have been used in the literature, drawing explicit connections between the various names wherever possible (Table 1 ). This table is aimed to resolve inconsistencies in the literature, especially when particular cells have been omitted or subtle denominations interchanged. Many different names have been used for both the external sensory organs and the multidendritic neurons. The multidendritic neurons have been classified and named based on their gross morphology and the substrate upon which they extend their dendrites [ 14 ]. They were therefore called tracheal dendrite (md-td), bipolar dendrite (md-bd) and dendritic arborization (md-da) neurons. In recent papers, the "md" has often been dropped and they have been referred to as td, bd, or da neurons. Whereas the bd and td neurons usually reside singly and their identification is usually not problematic, the da neurons are found grouped as clusters of cells and it has been more difficult to assign individual identities. Early papers named these clusters according to the number of da neurons they contain. The dorsal cluster was therefore called dmd5 (for dorsal multidendritic cluster of 5) [ 15 ] or dmd6 [ 6 ]. Merritt and Whitington were the first to name nearly all da neurons as individuals [ 7 ] by providing each with an alphabetic designation reflecting their ventral-to-dorsal position. However, their dorsal cluster contained only 5 da neurons, and since we now know that this cluster typically contains an additional, sixth da neuron, it has not been possible to assign their specific names to the cluster. When Grueber et al. characterized the morphology of all da neurons [ 8 ], and Sweeney et al. the morphology of the dorsal cluster neurons [ 19 ], each group chose to assign names in the same vein as Merritt and Whitington. However, the typical relative positions of some neurons were not resolved because specific markers were lacking. Cut antibodies were subsequently used to discriminate the different neurons, and their typical relative positions were established [ 13 ]. Table 1 Correspondence between the nomenclatures previously used to name sensory organs or neurons. [5] [16] [3] (names es organs) [4] (names es and ch neurons) [6] (names es, ch and md neurons) [7] (names es, ch and md neurons) [1] (names sensory organs) [8] (names md da neurons) other (Only the first paper using this other nomenclature is indicated) external sensory (es) organs - - p1 vesA vesA vesA vp1 - vc1 [17] - - p2 vesB vesB vesB vp2 - vc2 [17] - - p3 vesC vesC vesC vp3 - vc3 [17] - - p4 v'esA v'esA v'esA vp4 - vc4 [17] - - p5 v'esB v'esB v'esB vp4a - vc4a [17] - - p6 v'es2 v'es2 v'es2 vp5 - vc5 [17] b sensory hair C h1 lesA lesA lesA lh2 - - - - p7 lesB lesB lesB lp2, lc1 - - H sensory hair C h2 lesC lesC lesC lh1 - - b+st sensory hair F h4 desB des2 desA2 dh2 - h3 desA desB desB dh1 - - s - p8 desC desC desC dp1, dc1 - - b - p9 desD desD desD dp2, dc2 - - chordotonal (ch) organs - - - vchA vchA vchA vch1 - - - - - vchB vchB vchB vch2 - - - - - lch5 lch5 lch5 lch5 - - - - - v'ch1 v'ch1 v'ch1 lch1 - - multidendritic (md) neurons - - - - vbd vbd - vmd3 [9] - - - - vdaC vdaC vmd2 [9] - - - - vdaA vdaD vmd1 [9] - - - - vmd5 vdaD vdaA-D vdaA vmd4 [9] - - - - vdaB vdaB vmd1a [9] - - - - vpda vpda vdap vpda - - - - - v'ada v'ada vdaa v'ada vmd4a [9] - - - - - v'td1 [this paper] - - - - v'td2 v'td2 vtd1/2 - v'td2 [this paper] - - - - v'pda v'pda v'dap v'pda - - - - - ldaA ldaA lda ldaA - - - - - lbd - isbp - ldb, lbd [18] - - - - ltd ltd istd - - - - - - ldaB ldaB ltd ldaB - - - - - dbd dbd dbp - - - - - - ddaA ddaA the ddaA-B-C-D-E cluster has also been called - - - - ddaB or ddaC? ddaA-E ddaB dmd5 [15] - - - - dmd6 ddaB? ddaC - - - - - ddaD ddaF [19] - - - - ddaD? ddaE - - - - ddaD? ddaF ddaD [19] - - - - - ddaE - - dda1 [18] dmd1 [13] Each nomenclature presented in Table 1 has its own advantages, and no single nomenclature is yet uniformly agreed upon by all researchers. The nomenclature presented on some FlyPNS pages, such as the "Sensory organ description" page, simply reflects the names that we use in our own work and that have also been found in several recent publications. We hope and think that with time researchers will agree on a common constructive nomenclature that also allows wide comprehension of the wealth of data already available on PNS development. The Sensory organ description section gives an extensive description for each individual sensory organ: position, name(s), markers, morphology, development and related references. A few unpublished observations relating to maker expression and morphology are also mentioned. A typical sensory organ description is shown in Fig. 2 . Links to other sensory organ descriptions, antibodies, enhancer-trap markers and references are available. Clicking on the various cells depicted in the adjacent diagram also allows navigating between sensory organ descriptions. The extensive cross-linkage of the various pages of the web site is designed to facilitate the comprehension of the available data. Figure 2 Description of the sensory organ v'ada . Links to other sensory organ descriptions, antibodies, enhancer-trap markers and references are shown. Clicking on a cell depicted in the diagram links to its sensory organ description. The Antibodies, enhancer-traps and Gal4 lines pages display a collection of published and unpublished data on expression patterns of antibodies, enhancer-traps and Gal4 lines, with links to the gene/insertion information contained in Flybase when available. Finally, the Search option provides a convenient link to the information about any of the organs/markers. Utility, discussion and conclusions Much of the discordance in the literature regarding the larval and embryonic PNS pattern seems to be rooted in variation in the precise position of sensory cells with segment, embryo, and time, and a lack of specific markers for some organs. These factors might explain why, for example, some dorsal cluster neurons were often missed in previous publications. We have attempted to resolve remaining ambiguities by (1) identifying the total number of sense organs and neurons in each abdominal segment A1-A7, (2) resolving different nomenclatures used for each organ, and (3) providing a collection of molecular markers and descriptions that provide an unambiguous guide for cell identification. This data compilation will hopefully also increase the accessibility of the fly PNS literature to non- Drosophila investigators. Availability The FlyPNS web site is available from the link under Miscellaneous in the FlyBase Allied Data section . Comments and additions are welcomed by e-mail to VO. List of abbreviations PNS: peripheral nervous system da neuron: neuron displaying a dendritic arborization Authors' contributions The PNS pattern was established based on the observations of WBG and VO. VO created the FlyPNS web site and WBG made significant improvements and corrections to the web site. Both VO and WBG wrote the manuscript.	/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC552318.xml
544546	Turning to Science	A review of John Brockman's new book Curious Minds: How a Child Becomes a Scientist.	Do you want your child to grow up to be a scientist? If so, then by all means follow the roadmaps set down by the 27 scientists who contributed their personal histories to John Brockman's Curious Minds: How a Child Becomes a Scientist . You could do worse than start with a family history rich in science, like Nicholas Humphrey whose grandfather, Nobel laureate A. V. Hill, remembers being startled by a solar eclipse while out rabbit hunting as a boy in England and having the presence of mind to smear the rabbit's blood over the glass from his pocket watch in order to watch the phenomenon in safety. Or, if you don't have a healthy lineup of scientific superstars in your family history, you could educate yourself about the wonders of science, like Murray Gell-Man's autodidactic father, and insist that your son at least try majoring in physics before settling on linguistics or archaeology. If you cannot be inspired to teach yourself quantum physics, you might also try the more commonplace achievement of a bad marriage to create an unstable emotional life for your daughter, so she might retreat into “solitary and bookish tendencies,” as the evolutionist Lynn Margulis did. Perhaps you cannot bring yourself to actively steer your child into science, but hope this book might guide you to at least watch for the signs of future genius. You might watch for a young Freeman Dyson in his crib, bored by the stuffed animals and mobiles, occupying his time by adding infinite series of fractions. Or you might watch for your daughter sneaking into your study to read your medical reference books on the sly, as Janna Levin did before ultimately becoming a professor of physics. Then again, you could watch for your son to emulate David Buss by reaching high for a C+ grade point average in school, indulging in recreational drugs, and taking a night shift job at a truck stop, before a scientifically minded girlfriend and a fortuitous lottery-based scholarship to the University of Texas turned him to the pursuit of evolutionary psychology. The pattern is clear: there is no pattern. And the book's strong contingent of psychologists is not shy about commenting on the dubious reliability of self-narrative. Steven Pinker writes, “Don't believe a word of what you read in this essay on the childhood influences that led me to become a scientist. Don't believe a word of what you read in the other essays, either….Recounting childhood influences is a mental process no less subject to quirks and errors than falling for the visual illusions on the back of a cereal box….None of us has taken part in the experiments that would isolate the causes of our choices in life.” Or as Nicholas Humphrey puts it more simply, “Each of us is who we are, and we must each have had some sort of childhood. Who's to say whether any particular factor carried the weight that our self-narrative now likes to attribute to it.” This book, then, is not a guide to the prescientific childhood, but a set of memoirs bound by their common conclusion. Many of them are highly entertaining, some of them are self-conscious and pedantic. All of them highlight the passion of childish curiosity extending into adulthood, either by family tradition or because a mentor appeared at the right place and time to ensure that this curiosity was not abandoned. As the developmental psychologist Allison Gopnik notes, “I suspect that there are few reports of scientists with a childhood fascination for babies, because most of those children turned into nursery school teachers or children's librarians or just stay at home mothers….It seems to me now that I was destined to become either a psychologically minded philosopher or a philosophically minded pscyhologist. But given slightly different contingencies, I might have become a frustrated preschool teacher or faculty wife.” We wish we could find prescriptions for guiding our children towards success, but the nature versus nurture debate has long since been replaced by the more sophisticated understanding of just how inextricably intertwined these two influences are in creating even a lowly zebrafish, let alone anything as behaviorally complex as a human being. This understanding can lead to frustration, as simple causes and cures become more elusive. Selling Mozart recordings for expectant mothers to play to their wombs can help assuage the fears and ignite the hopes of overachieving, yuppie parents, but there is little evidence that such interventions create the sought after intellectual advantage for the newborn. Mostly what we see in these essays are stories about people—stories that people tell about themselves. Stories of childhoods interrupted by war, of mentors persecuted by McCarthyism. Would a book tracing the self-styled childhood influences of outstanding politicians or judges have been so different? Possibly fewer socially awkward children and less gadget-building in the mix, possibly not. The contributors to this book are all outstanding, motivated scientific leaders who have chosen or stumbled into their intellectual paths. But I am not convinced that reading their brief reflections compares with the unique opportunity of interviewing Albert Einstein that John Brockman identified as one inspiration for this book—for turning what began as dinner party conversation into a methodical attempt at biographical anthology. Much of these pages are in fact dinner party conversation, whether it is Robert Sapolsky's horror of the white Southern gentlemen of Harvard's intellectual elite playing cards and drinking while dividing the spoils of sociobiology, or Allison Gopnik's personal relief that there were no prohibitions against sleeping with teaching assistants when she was a nubile college student. Some of it is heart-rendingly personal, as when Jaron Lanier talks about his futile attempt at age 11 to attract friends to his homemade electronic Halloween haunted house. All it really tells us is that there are as many ways of forming a scientist as there are of forming a human being.	/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC544546.xml
524259	Only Connect: The Functional Architecture of Brain Connectivity	null	Imagine three cities, A, B, and C, splayed across the landscape to form a triangle, with each connected to the other two by two-lane roads. Such an arrangement of cities and roads constitutes a structural network. On any given day traffic may flow, say, only from A to B to C, or in both directions between A and B but from C only to A, or in both directions between all three, or any one of ten other arrangements. Within this structural network, then, there are 13 possible functional networks. If these cities are embedded within a larger network of routes and destinations, their particular triangular traffic pattern represents a “motif” of connectivity, akin to a recurring musical motif within a larger symphony. Such connectivity networks are central to information processing in the brain, and understanding the recurring structural and functional motifs they contain is one way to begin to dissect how the symphony of brain function is composed. In this issue, Olaf Sporns and Rolf Kötter identify several common motifs in real brain networks, and show that brains tend to maximize the number of functional motifs while keeping the number of structural motifs relatively low. Figure 1 The authors began with the frequency of motifs of different sizes (two, three, four, or five nodes) found in the visual cortex and whole cortex of the macaque monkey, the cat cortex, and the nervous system of the nematode Caenorhabditis elegans . For comparison, they generated matrices that contained an equivalent number of components (nodes and connections), but whose connections were either random or lattice-like, in which all nearest neighbors were connected. They found that, compared to the artificial networks, the biological ones were relatively low in structural diversity. For instance, macaque visual cortex contained instances of 3,697 different motifs with five nodes, versus 8,887 for equivalent random networks. Functionally, however, unlike the artificial systems, the biological systems were maximally diverse, with the maximum functional motif diversity (e.g., 13 for three vertices and 9,364 for five vertices) observed in all motif sizes they investigated. The researchers also found some intriguing patterns within this maze of connectivity. For instance, not all motifs were found in equal numbers. A common functional motif for three vertices was for both A and C to communicate back and forth with B, but not with each other. This structure allows B to function as an integrator of signals from A and C, while keeping the activities of A and C distinct from one another. This kind of structure is widespread throughout the nervous system. The authors then ran an evolutionary algorithm on their artificial networks. They showed that by selecting for maximal functional motif number, the structure of the artificial systems quickly came to resemble the structure of the real ones, with dense local connections and relatively fewer long-distance ones. Such a structure, termed “small world” connectivity, promotes cooperation between functional units, and efficient information exchange. Taken together, these results suggest that one factor that may drive the evolution of neural architecture is the maximization of functional connectivity within a network of relatively few neural actors.	/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC524259.xml
549076	Omega-3 fatty acids decreased irritability of patients with bipolar disorder in an add-on, open label study	This is a report on a 37-patient continuation study of the open ended, Omega-3 Fatty Acid (O-3FA) add-on study. Subjects consisted of the original 19 patients, along with 18 new patients recruited and followed in the same fashion as the first nineteen. Subjects carried a DSM-IV-TR diagnosis of Bipolar Disorder and were visiting a Mood Disorder Clinic regularly through the length of the study. At each visit, patients' clinical status was monitored using the Clinical Monitoring Form. Subjects reported on the frequency and severity of irritability experienced during the preceding ten days; frequency was measured by way of percentage of days in which subjects experienced irritability, while severity of that irritability was rated on a Likert scale of 1 – 4 (if present). The irritability component of Young Mania Rating Scale (YMRS) was also recorded quarterly on 13 of the 39 patients consistently. Patients had persistent irritability despite their ongoing pharmacologic and psychotherapy. Omega-3 Fatty Acid intake helped with the irritability component of patients suffering from bipolar disorder with a significant presenting sign of irritability. Low dose (1 to 2 grams per day), add-on O-3FA may also help with the irritability component of different clinical conditions, such as schizophrenia, borderline personality disorder and other psychiatric conditions with a common presenting sign of irritability.	Introduction According to the United States National Institute of Mental Health (NIMH), Bipolar Disorder (BPD), also known as manic-depressive illness, is a serious medical illness that causes shifts in a person's mood, energy, and ability to function. Different from the normal ups and downs that everyone goes through, the symptoms of bipolar disorder are severe. Bipolar disorder is a complex, chronic condition associated with considerable morbidity and mortality, including a high rate of suicide. Bipolar disorder causes dramatic mood swings from overly "high" and/or irritable to sad and hopeless, and then back again, often with periods of normal mood in between. Severe changes in energy and behavior go along with these changes in mood. The periods of highs and lows are called episodes of mania and depression. Most people with bipolar disorder can achieve substantial stabilization of their mood swings and related symptoms over time with proper treatment. A strategy that combines medication and psychosocial treatment is optimal for managing the disorder over time. Background Omega-3 fatty acids (0-3FA) may have a beneficial effect on irritable mood. Low O-3FA levels in red blood cell membranes of depressed patients hint that O-3FA may be helpful in treating mood disorders [ 1 ]. A recent article has given an excellent review of O-3FA and studies showing their effectiveness in depression, bipolar disorder and aggression [ 2 ]. In this article, two published studies are discussed that have reported on similar therapeutic effects of O-3FA [ 2 ]. One placebo-controlled study of 20 patients revealed that ethyl ester of eicosapentaenoic acid (E-EPA) was effective in stabilizing the moods of depressed patients [ 3 ]. Another report, a double-blind, placebo controlled study (N = 22/19), measured the effect of O-3FA docosahexaenoic acid (DHA) on the aggressive tendencies of college students. The O-3FA DHA group (1.5–1.8 g O-FA DHA/day) did not display any increase in aggressive tendencies when external stressors peaked, while the placebo group displayed a significant increase in their aggressive tendencies under similar circumstances [ 4 ]. In a recent study, 25% of 111 patients with bipolar-I disorder who met criteria for a DSM-IV major depressive episode also experienced substantial irritability in the absence of associated symptoms of mania. These findings suggest that abnormal irritability is not limited to mania or mixed states [ 5 ]. However, recent studies give caution that at a 6 gram per day average daily dose, as a single agent, omega 3 fatty acids may not be as effective as an antidepressant [ 6 - 9 ]. O-3FA may also help with the irritability component of different clinical conditions, such as depression, mania, schizophrenia, borderline personality disorder and other psychiatric conditions with a common presenting sign of irritability. Numerous other conditions have an irritability component, including Borderline Personality Disorder, Alzheimer's disease, Premenstrual Dysphoric Disorder, to name a few [ 10 - 12 ]. There is one report suggesting beneficial effect of Omega-3 Fatty acid treatment for Borderline Personality Disorder. This double-blind, placebo-controlled pilot study specifically showed that EPA may influence both aggression and depression [ 12 ]. Although attention-deficit/hyperactivity disorder (ADHD) also has an irritability component, recent publications bring doubt to the O3FA connection in ADHD [ 13 , 14 ]. A recent, open ended, O-3FA add-on study has shown beneficial effect of O-3FA on irritability in 19 patients with mood disorders [ 15 ]. These patients had already been receiving different combinations of pharmacotherapy and talk therapy. Despite their treatment, the irritability component of their illness was still causing social, occupational and other life disturbances. Hence, they were chosen for the O-3FA add-on component of the study. In the nineteen-patient phase of the study, bipolar patients of every subtype, ages 18 to 65 years, with significant irritability were studied. All patients received a systematic assessment battery at entry and were treated by a psychiatrist, trained to deliver care and measure outcomes in patients with bipolar disorder, consistent with expert recommendations. At every follow-up visit, the treating psychiatrist completed a standardized assessment and assigns a clinical status based on DSM-IV criteria. Patients had independent evaluations at regular intervals throughout the study and remain under the care of the same treating psychiatrist while receiving variable medications and talk therapy, depending on their need [ 15 ]. In the 19-patient study, a paired sample t-test revealed a large decrease in the percent of days irritable after O-3FA was administered. Before treatment, the mean irritability percentage was 81.05 (SD = 23.31) and after treatment the mean irritability percentage dropped to 30.00 (SD = 36.67). Despite the small number of patients in the study (n = 19), the difference between means was statistically significant (t (18) 4.512, p < .001). Using a paired sample t-test, a significant difference was also found between the highest irritability score (mean = 2.79; SD = 0.92) and the last recorded irritability (mean = 0.79; SD = 0.85) while taking O-3FA (t(18) = 8.270; p < .001) [ 15 ]. Methods This is a report on a 37-patient continuation phase of the open ended, O-3FA add-on study. Subjects consisted of the original 19 patients, in addition to the 18 new patients recruited and followed in the same fashion as the first nineteen [ 15 ]. Subjects carried a DSM-IV-TR [ 16 ] diagnosis of Bipolar Disorder and were visiting a Mood Disorder Clinic regularly throughout the length of the study. At each visit, patients' clinical status was monitored using the Clinical Monitoring Form [ 17 ]. Subjects reported on the frequency and severity of irritability experienced during the preceding ten days; frequency was measured by way of percentage of days in which subjects experienced irritability, while severity of that irritability was rated on a Likert scale of 1 – 4 (if present). The irritability component of Young Mania Rating Scale [ 18 ] (YMRS) was also recorded quarterly on 13 of the 39 patients consistently. The patients were asked about general dietary omega-3 intake before the fish oil was added on, and basic nutritional guidance was given to subjects at the clinic. Patients in general were not heavy fish/product consumers. Dosage Starting dose and last maintenance dose were available for 37 subjects (Table 1 ). Subjects self-medicated, and therefore, the last maintenance dose of O-3FA was chosen by each subject. The mean starting dose was 1824.32 mg (SD 1075.07), and the mean for the last maintenance dose was considerably higher at 2878.38 mg (SD 2011.79). The increase was statistically significant using a paired sample t-test (t = -3.44, 36df, p = .001). Statistical Results Percentage of Irritable (Days) The initial mean was 63.51 (SD 34.17), indicating that on average, subjects were irritable for about six of the previous ten days. The mean for the last recorded percentage was less than half of the initial score: 30.27 (SD 34.03). The decrease was found to be statistically significant using a paired sample t-test (t = 4.36, 36 df, p < .001). The difference between the distributions was examined using the non-parametric sign test. The number of negative differences (25) significantly exceeded positive differences (7); there were five ties, and the pre/post distributions were significantly different (p < .003). YMRS Irritability Sub-score Thirty four subjects had initial and last recorded YMRS irritability sub-scores. As with the above means there was a sizable decrease. The initial mean score was 3.18 (SD 1.09). The mean for the last recorded percentage was 1.68 (SD 1.89). The decrease was found to be statistically significant using a paired sample t-test (t = 4.21, 33 df, p < .001). YMRS Total Score Starting and last recorded YMRS scores were available for 34 subjects. The mean starting score 10.71 (SD 6.77), and the mean for the last recorded score was 4.85 (SD 5.63). The decrease found to be statistically significant using a paired sample t-test (t = 4.14, 33 df, p < .001). Severity Thirty six subjects had initial and last recorded severity scores on the ADE. Again, a decrease was found. The initial mean score was 2.14 (SD 1.22). The mean for the last recorded score was 0.94 (SD 0.92). This decrease was found to be statistically significant using a paired sample t-test (t = 5.23, 35 df, p < .001). Composite: Severity and Irritability As an exploratory measure, a composite score was created by multiplying the ADE severity score, which has a maximum of 4 points, by the percentage of the ten days prior to measurement which the patient was rated as irritable. The initial mean on this composite was 159.72. As with other measures, there was wide variation: SD = 122.92. The mean for this measure on the last recorded scores was percentage was about one-fourth of the initial score: 43.89 (SD 64.38). The decrease was found to be statistically significant using a paired sample t-test (t = 5.00, 35 df, p < .001). Last Recorded Maintenance Dose and Percentage of Irritability After Because of apparent wide variation on these two measures and a concern that outliers may have affected some results, the last recorded irritability scores were plotted against the maintenance dose. This revealed a rather bimodal pattern, in which relatively lower irritability measures (≤ 50%) clustered in the quadrant with lower dosage levels (≤ 4,000 mg). Duration and YMRS Total In response to a similar observation regarding wide variation in the last recorded values (84 days to 5.5 years) the values were also plotted. A clearly bimodal pattern appeared in which 11 subjects (about one-third of study participants) clustered in the quadrant representing short duration (<500 days) and higher YMRS totals (>7). The remaining two-thirds of subjects clustered in the quadrant representing short duration and lower YMRS totals (<6). Subject Weight The mean start weight was 176.97 lbs (SD 43.13), and the mean for the last weight recorded was slightly higher at 178.59 lbs (SD 43.24). The increase was not statistically significant. Follow-up Subjects Follow-up information, recorded after the collection of the "last" scores for most of the above variables, was available for 13 of the 37 subjects. Final YMRS total or scale scores were not available for this sub-group. Omega 3 Duration The final date recorded for the duration of O3 was derived based from an O3 start date and a "final" date recorded for O3. The time period ranged 84 days to 1995 days (5.46 years). The mean duration of O3 for this group was 439.62 days (SD = 487.46). Dosage For these subjects, the mean starting dose was 1807.69 mg (SD 990.34), and the mean for the last maintenance dose was higher at 2615.38 mg (SD 1894.66). The increase was not significant. Percentage Irritable (Days) The initial mean was 82.31 (SD 20.88). The mean for the last recorded percentage was dramatically lower: 25.38 (SD 32.04). The decrease was found to be statistically significant using a paired sample t-test (t = 6.52 12 df, p < .001). The difference between the distributions was examined using a sign test. The number of negative differences (12) significantly exceeded positive differences (0); there was one tie, and the pre/post distributions were significantly different (p < .001). Severity The initial mean score for the 13 subjects with final scores was 2.69 (SD 0.95). The mean for the final score was 0.77 (SD 0.83). This decrease was found to be statistically significant using a paired sample t-test (t = 6.22, 12 df, p < .001). Composite: Severity and Irritability An exploratory composite score, described above, was also created for the subjects with final scores. For these subjects, the initial mean was higher than that of the total group, 223.08. Again, there was wide variation: SD = 104.19. The mean for this measure on the last recorded scores was percentage was much lower that the initial score: 33.08 (SD 39.87). The decrease was found to be statistically significant using a paired sample t-test (t = 6.70, 12 df, p < .001). Weight For these 13 subjects, the mean start weight was 166.23 lbs (SD 35.68), and the mean for the final weight recorded was also slightly higher at 168.23 lbs (33.62). As with the previous finding regarding weight, the increase was not statistically significant. Results Omega-3 Fatty Acids added onto the existing treatment helped with the irritability component of a significant percentage of patients suffering from bipolar disorder with a persistent sign of irritability. Discussion As seen from the standard deviations of several of the variables discussed here, measures ranged widely. This creates difficulty in using descriptive data, such as means, to adequately portray subject attributes and performance. Using data reduction techniques or grouping subjects according to high and low scores on various attributes may be one way to increase the descriptiveness, which would be possible and more reasonable with a larger pool of subjects. A potential limitation or interpretive consideration merits discussion. For many of the variables discussed above, noticeable differences in measures were observed between the "starting" versus "last recorded" group (n = 37) and the "starting" versus "final" measures group (n = 13). Given these differences and the smaller number of subjects in the second set of comparisons, "starting" versus "final" comparisons should be interpreted with caution until differences inherent in this "final" subgroup (n = 13) are more clearly understood. This is clearly seen in the results of sign tests, in which the apparent magnitude of the "final" effects is pronounced. Statistically significant within-subjects differences were found in several independent variables. This is especially notable given the small number of subjects. The preliminary findings suggest that a rigorously designed study tailored especially to the examination of the effects of O-3FA is warranted. The majority of data were collected within an ongoing "best-practice, outpatient bipolar disorder study" that involved medications and talk therapy which we have not reported or discussed herein. Results must, therefore, be interpreted with caution. There are several mechanisms through which O-3FA are theorized to help with mood, irritability, aggression etc. Suggested theories of mechanism converge on the theory of nerve cell membrane stabilization. A recent study has come closest to showing physical proof of effectiveness of O-3FA through indirect demonstration of greater membrane fluidity, as detected by reductions in Tesla-2 (T2) values in MRI scans [ 19 ]. The overlapping beneficial effects of antipsychotics, antidepressants, anticonvulsants, O-3FA, and nonpsychoactive cannabinoids, as they relate to pain, stroke, schizophrenia, psychoneuroimmunology, Alzheimer's disease, and stress, may be because of their common effects at protein kinases, thus affecting the structure and function of the cell membrane and the cell [ 20 ]. These changes should help the cell operate within an optimal level of excitation, which may be related to emerging evidence that these therapeutic agents have neuroprotective value [ 20 ]. A recent randomized placebo controlled double blind intervention study suggests an adaptogenic role for O-3FA in stress [ 21 ]. We would like to discuss briefly the issue of daily dosing of O-3FA for nutrition and medicinal purpose: Recent studies give caution that at a 6 gram per day average daily dose, as a single agent, omega 3 fatty acids may not be as effective as an antidepressant [ 6 - 9 ]. However, these studies may have given too high of a dose of O-3FA, above 6 grams daily, with possibly beyond a therapeutic window of effectiveness for O-3FA. Our scatter plots indicate that the optimum effective dose for irritability is at 1–2 gram of EPA plus DHA per day, which would be the dosing we suggest. A recent exploratory dose study of O-3FA for schizophrenic patients showed that 2 g/day EPA-treated patients had lower symptom scores, and needed less medication greatest. In this study, there was a positive relationship between improvement on rating scales and rise in red blood cell arachidonic acid concentration as well [ 22 ]. The United States (US) accounts for more than 51% of the 430.3 billion dollar expended on pharmaceutical products worldwide each year [ 23 ]. World healthcare society first needs access to low-cost, nontoxic, non-expert-dependent interventions to ensure basic health outcomes. Food may represent the most cost-effective means of promoting public health [ 23 ]. The American Heart Association recommends consumption of two servings of fish per week for persons with no history of coronary heart disease and at least one serving of fish daily for those with known coronary heart disease [ 24 ]. Approximately 1 g per day of EPA acid plus DHA acid is recommended for cardioprotection [ 24 ]. Higher dosages of omega-3 fatty acids are required to reduce elevated triglyceride levels (2 to 4 g per day) and to reduce morning stiffness and the number of tender joints in patients with rheumatoid arthritis (at least 3 g per day) [ 24 ]. We conclude that it is beneficial in many ways to establish a regular intake of 1–2 g per day EPA acid plus DHA, similar to daily intake of vitamins with minerals. Dietary interventions to remedy omega-3 deficiency is necessary [ 23 ]. It is time for more aggressive funding for research into medicinal foods, such as omega-three fatty acids [ 23 ].	/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC549076.xml
554769	A robust method for the amplification of RNA in the sense orientation	Background Small quantities of RNA (1–4 μg total RNA) available from biological samples frequently require a single round of amplification prior to analysis, but current amplification strategies have limitations that may restrict their usefulness in downstream genomic applications. The Eberwine amplification method has been extensively validated but is limited by its ability to produce only antisense RNA. Alternatives lack extensive validation and are often confounded by problems with bias or yield attributable to their greater biological and technical complexity. Results To overcome these limitations, we have developed a straightforward and robust protocol for amplification of RNA in the sense orientation. This protocol is based upon Eberwine's method but incorporates elements of more recent amplification techniques while avoiding their complexities. Our technique yields greater than 100-fold amplification, generates long transcript, and produces mRNA that is well suited for use with microarray applications. Microarrays performed with RNA amplified using this protocol demonstrate minimal amplification bias and high reproducibility. Conclusion The protocol we describe here is readily adaptable for the production of sense or antisense, labeled or unlabeled RNA from intact or partially-degraded prokaryotic or eukaryotic total RNA. The method outperforms several commercial RNA amplification kits and can be used in conjunction with a variety of microarray platforms, such as cDNA arrays, oligonucleotide arrays, and Affymetrix GeneChip™ arrays.	Background The increased use of microarray expression profiling to study both the molecular biology of cancer and the cellular physiology of difficult-to-isolate cell types has led to a growing need for methods that allow the use of limiting quantities of RNA. Small surgical biopsies, fine needle aspirates, cyto-lavages, punch biopsies and blood samples often yield only 1–4 μg quantities of RNA as starting material for expression profiling. This limitation has prompted the development of amplification methods that produce the quantities of RNA required for microarray analysis. Changing requirements for the type and quantity of amplified RNA, driven by evolving microarray technologies, have led to the development of novel amplification strategies. While current methods are capable of delivering high-yield RNA amplification, this is often only achieved after complex priming strategies (for example, involving 4 or more primers) are coupled with multiple rounds of PCR and/or in vitro transcription, resulting in time consuming and costly protocols. Here, we present an overview of RNA amplification strategies, identify key limitations to existing techniques, and describe a simple, robust, and cost-effective strategy for single round amplification of RNA in the sense orientation. RNA amplification methods Early attempts to amplify RNA employed a strategy based upon the Polymerase Chain Reaction (PCR) [ 1 - 4 ]. These methods relied on the terminal transferase activity of reverse transcriptase to allow addition of primer sites to the 3' end of reverse-transcribed, first-strand cDNA. Multiple rounds of PCR primed from this site and from the poly-(A) + sequence on the second-strand cDNA could then be used to facilitate amplification. These methods were confounded by differential amplification of cDNA and by introduction of errors by Taq polymerase. This problem prompted the development of a linear, T7-based in vitro transcription (IVT) method by Van Gelder and Eberwine [ 5 - 7 ]. In what has now become known as the "Eberwine Method," RNA templates are primed with an oligo(dT) primer that has been 5' modified to contain a promoter for the T7 RNA polymerase and are subsequently reverse transcribed into first-strand cDNA. The RNA-cDNA hybrid is then treated with E. coli RNAse H, and priming for second-strand cDNA synthesis occurs by either RNA nicking and priming or cDNA hairpinning 6 . Second-strand cDNA synthesis is carried out with E. coli DNA polymerase and E. coli DNA ligase followed by blunt-ending with T4 DNA polymerase. Transcription and amplification are then accomplished using the T7 RNA polymerase, which binds to the T7 promoter introduced during first-strand cDNA synthesis, producing antisense RNA (aRNA). Technical revisions of the Eberwine method have included changes in first-strand primer concentration to minimize the appearance of non-sequence dependent RNA in the amplified product [ 8 ], supplementation of second-strand priming with random primers to improve its efficiency, and modifications that allow multiple rounds of IVT to augment yield [ 6 , 9 , 10 ]. Concerns regarding the fidelity of amplification with these methods stem from the 3' bias introduced by the use of the promoter-modified oligo(dT) primer during first-strand cDNA synthesis, and questions remain over the degree to which this amplified RNA reflects the true transcriptome of the unamplified sample. To correct for this potential bias, three alternatives have been developed to the Eberwine protocol. One such alternative [ 11 ] is based upon the Eberwine approach, but second and subsequent rounds of amplification are primed with random nonamer primers modified by the addition of an upstream T3 promoter sequence (T3N9 primer). IVT from this T3 promoter prevents serially compounding the 3' bias introduced by the oligo(dT) primer across multiple rounds of amplification. The T3N9 primer has also been used to prime the initial round of reverse transcription, a modification that is useful for amplifying partially-degraded samples of RNA [ 12 ]. In this case, the method sacrifices the ability to selectively amplify mRNA for the versatility generated by the random priming and subsequent amplification of any RNA sequence present in the sample. A second alternative to the Eberwine method is the "template-switching" (TS) strategy [ 13 ]. This technique centers on the observation that the Moloney murine leukemia virus reverse transcriptase (MMLV-RT) adds an oligo(dC) region to the 3' end of first-strand cDNA after reaching the terminal end of an RNA template [ 14 ]. When an oligo(G) primer is added to the second strand RT reaction, it will hybridize with this oligo(dC) sequence, and the MMLV-RT will switch strands (the "template-switch") and continue the reverse transcription reaction. This strategy can be used to append a T7 promoter to the 5' end of the oligo(G) primer[ 13 , 15 ], facilitating RNA amplification by IVT. Yield can be further improved by combining this technique with PCR amplification after cDNA synthesis [ 16 ] for bulk production of amplified RNA [ 17 ]. Several variations on this theme involving changes in the primers and in the details of the PCR amplification have been described, all of which rely on a combination of TS primers and PCR-based amplification to produce large amounts of amplified RNA [ 18 , 19 ]. A third recently introduced alternative is Ginsberg's Terminal Continuation (TC) technique [ 20 ]. In this approach, the initial reverse transcription reaction is primed with a mixture of an oligo(dT) primer and a modified TC primer. The former primes the reverse transcription of mRNA, while the latter is essentially a T7 promoter-containing GC-rich sequence that primes second strand cDNA synthesis. According to Ginsberg, the "likely mechanism (for this incorporation) is that the TC primer binds preferentially to GC-rich CpG islands flanking 5' regions of DNA that contain promoter sequences." [ 21 ] Initial reports of the results of TC amplification show promise for linear amplification of high-quality RNA, but extensive validations of this method have yet to be conducted. Validity of amplified RNA in microarray applications The degree to which the pool of amplified RNA generated by these methods reflects the unamplified sample from which it is derived is an obvious concern for microarray applications and other downstream analyses, and the history of attempts to validate RNA amplification methods is summarized by Nygaard et al. [ 22 ]. Briefly, they note that a combination of Northern blotting, dot blot differential screening of cDNA probes synthesized from aRNA, internal RNA standards, hierarchical clustering, qRT-PCR, subgroup analysis, and ratio-intensity (RI) plot analysis has lead to the conclusions that relative expression levels are well-preserved after 1–3 rounds of amplification, that important over- or under-expressed genes are detectable after amplification, and that amplification may actually improve detection of RNA present in low copy number. However, they noted that few studies quantitatively compare total RNA and amplified RNA, that the true sources of amplification biases have not been thoroughly investigated, and that the degree to which "noise" affects the differences in these profiles has not been adequately assessed. Accordingly, they conducted multiple hypothesis testing based upon t-tests and ANOVA analysis of several technical parameters to study the nature and the magnitude of biases and variability associated with the use of amplified RNA in microarray expression profiling. They observed that approximately 10% of genes showed statistically significant differences in relative expression level between amplified and their unamplified counterparts and noted that neither technical replication of amplifications nor molecular characteristics of the sample were the likely cause of these observed differences. Despite these differences, they stated that the increased quantity and purity of mRNA hybridized in studies using amplified RNA increases overall fluorescence intensity and improves detection of low-abundance messages. Because they noted that more that 50% of the amplified RNA showed log 2 (ratio) differences within ± 0.5 of the unamplified RNA, they concluded that RNA amplification is useful in expression profiling and is likely to assist in the measurement of low-abundance RNA. Despite the methodological validation provided by these studies, only recently have statistically-based analyses been used to compare RNA amplification methods in a critical, head-to-head fashion. A study by Zhao [ 23 ] examined differences between T7-based amplification protocols, including the Eberwine method and the template switch (TS) technique, where these methods were used to amplify RNA extracted from tumor samples rather than from the idealized situation of cell lines. They observed that the use of TS strategies does not improve the fidelity of RNA amplification versus the Eberwine method, that there is good correlation between samples after amplification, that the overall bias introduced by T7-based amplification strategies is uniformly low, and that the results are reproducible. Their overall conclusion was that T7-based amplification protocols generate high-fidelity, amplified antisense RNA that is suitable for use in cDNA microarray analysis. Limitations of current amplification protocols It is currently accepted that RNA amplification strategies based on the Eberwine method maintain relative mRNA levels between samples when amplifying from either total RNA or poly(A) + RNA [ 24 , 25 ], are useful with low microgram quantities of starting RNA [ 26 ], and are capable of preserving differential expression profiles when used in conjunction with microarray analysis [ 13 ]. Accordingly, efforts have become increasingly focused upon developing an optimized protocol that minimizes amplification bias, provides versatility, and reduces technical complexity. While current commercial kits and published protocols are making progress to these ends, we feel that several important dimensions have been overlooked, limiting their utility. First, most commercial protocols produce RNA only in the antisense orientation. With the emergence of spotted oligonucleotide arrays and mixed cDNA/oligo arrays [ 27 ], appropriate orientation of amplified RNA is an important experimental consideration. While TS and TC protocols can be modified to produce sense RNA, this generally comes at the cost of increased technical and biological complexity but adds no demonstrable benefit over the Eberwine strategy. Second, many current protocols rely on multiple rounds of IVT and/or PCR to produce amplified RNA, and even a few rounds of this amplification has the potential to introduce sequence error [ 6 ] or systematic bias [ 28 ]. While some applications, such as laser capture microdissection, may produce only picogram quantities of RNA and thus require extensive amplification, it should be noted that even small biopsy samples frequently yield low microgram quantities of RNA and thus may require only a single round of amplification. Finally, the increasing complexity of many amplification protocols creates logistical problems when implementing them for large-scale projects. Protocols that require multiple rounds of amplification can accumulate material and labor costs that quickly exceed those associated with the actual microarray analysis, and amplification can become a "rate-limiting" (or "cost-limiting") factor in study design. Additionally, the number of enzymes, reagents, and custom primers continues to increase as protocols become more complex, a situation which is equally undesirable. We believe that a protocol that addresses these limitations and provides a versatile and robust method for RNA amplification is needed. Goals for a revised RNA amplification protocol Our goal was to develop a strategy based upon the Eberwine method but with the ability to produce sense RNA from small quantities of total or poly-(A) + RNA extracted from both ideal samples (e.g. cell line RNA) and "real-world" samples (e.g. tumors or tissues). This protocol should avoid the need for PCR steps and should require a minimum number of primers (two). Additionally, the protocol should be cost-effective, efficient, and technically simple to conduct. Finally, the method should give results consistent with similar amplification techniques when used with subsequent microarray analysis. We believe these criteria have been met with our protocol, which consequently will be useful in a variety of laboratory applications. Results Priming efficiency and cDNA yield Quantitation of first strand cDNA after hydrolysis of the cDNA/RNA hybrid shows that priming with the T7-Oligo(dT) promoter/primer followed by reverse transcription produces cDNA that totals 5–10% by mass of the original mass of total RNA (data not shown). This is consistent with the average percentage of mRNA in the total RNA sample and therefore suggests that priming proceeds efficiently and that first strand cDNA synthesis successfully copies the population of mRNA in the original total RNA sample. Quantitation of second strand cDNA after priming with the T3N9 promoter/primer indicates a 120% increase in mass versus the mass of the first strand cDNA (data not shown). This suggests that second strand priming proceeds efficiently and that the first strand cDNA is successfully converted to double stranded cDNA by the second strand synthesis reaction. The extra mass (in excess of 100%) is attributable to the use of random nonamer primers which initiates synthesis at multiple positions along each first-strand cDNA template. Median transcript length While inherent constraints of reverse transcription limit the amount of truly "full-length transcript" produced during any RNA amplification scheme, it is nonetheless desirable for an amplification method to favor the production of long transcript. RNA amplified using our method has a median length of 794 nucleotides (range: 70–9000 nt) suggesting that this method is capable of preserving long transcript during amplification (Table 2 ). Table 2 Yield and transcript length of amplified RNA produced by three amplification methods. Values in parentheses represent p -values from t -test comparing the value against that of the study protocol. Study Protocol RiboAmp Protocol ( p -value vs. Study Protocol) BDSMart Protocol ( p -value vs. Study Protocol) Mean yield from 4 μg total RNA ( , μg) 26.29 71.56 (<0.04) 4.87 (<<0.0001) Median amplified RNA length (nt) 794 507 (<0.003) 764 (0.88) Quantitative yield Our method produced an average ( ) of 26.3 μg of amplified sense mRNA from 4 μg of total RNA (Table 2 ). Assuming that 5% of the original sample of total RNA was mRNA, this reflects a 130-fold amplification. The protocol has been tested with as little as 1 μg of starting total RNA and produces sufficient mRNA for microarray analysis (~2 μg) under these conditions (data not shown). Amplification bias Any amplification will introduce some degree of bias into the population of amplified RNA. While recent reports suggest that the magnitude of this bias may be relatively unimportant as long as the bias is highly reproducible [ 29 ], we feel that aggressively limiting any source of experimental bias improves the quality of subsequent microarray data. We systematically evaluated the performance of RNA amplified using this protocol against identical hybridizations from unamplified RNA. Log 2 (ratio) plots of fluorescence intensity in amplified versus unamplified samples were constructed and analyzed for correlation as described above. Our results show that the amplification bias introduced by this method is small, as reflected in the high average correlation coefficient ( ) between expression profiles from amplified and unamplified RNA samples ( = 0.8009, Table 3 , Figure 2 ). Table 3 Statistical analysis of microarray data generated by comparing hybridizations of amplified RNA samples to identical hybridizations of unamplified (total RNA) samples. Values in parentheses represent P -values from t -test comparing the value to that of the study protocol. Study Protocol RiboAmp Protocol ( p -value vs. Study Protocol) BDSmart Protocol ( p -value vs. Study Protocol) Mean Correlation Coefficient ( ) 0.8009 0.7202 (<0.01) 0.7679 (0.23) Mean % of elements with absolute difference of ± 2 log 2 units 0.34 1.34 (<0.02) 0.60 (0.56) Mean % of elements with absolute difference of ± 1.5 log 2 units 1.31 4.05 (<0.05) 1.63 (0.67) Mean % of elements with absolute difference of ± 2 log 2 units 5.44 12.54 (<0.02) 6.52 (0.61) Mean % of amplified array elements within ± 1 log 2 unit 94.56 87.46 (<0.02) 93.48 (0.61) Figure 2 Correlation between expression profiles of RNA samples amplified using the study protocol and corresponding unamplified RNA samples demonstrates a low degree of amplification bias. Reproducibility We evaluated the reproducibility of this method and the consistency of the propagation of the amplification bias by comparing the expression profiles of hybridizations of three independently amplified RNA samples. Log 2 (ratio) plots of fluorescence intensity were constructed and analyzed as described above. Our results demonstrate a high degree of correlation between independent replicates ( =.9446, Figure 3 ) and a narrow dispersion (σ 2 = 0.00023, CV = 1.60%), suggesting that the amplification is highly reproducible and that the amplification bias is introduced consistently when the protocol is repeated. Figure 3 Correlation between independent replicates of expression profiles of RNA samples amplified using the study protocol demonstrates the high degree of reproducibility of the method. Comparison with commercial amplification kits We compared these results to results from the RiboAmp mRNA amplification kit (Arcturus), which produces antisense RNA using a version of the Eberwine method, and from the BDSmart mRNA amplification kit (BD Biosciences), which produces sense RNA using a template-switching strategy. Both protocols were carried out according to the manufacturer's specifications using 4 μg of the pooled cell line RNA and the glioblastoma RNA as the starting material for amplification. The final yield of each method was measured using the NanoDrop spectrophotometer, the median size of the amplified mRNA from each kit was assayed using the Bioanalyzer 2100 (Agilent), and the amplification bias of each method was assessed using microarray analysis (as described above). Three independent replicates were conducted for each method, and the reproducibility of the amplification was assessed as described above. The mean correlation coefficient ( ) and the dispersion of each group (σ 2 , CV) were calculated, demonstrating a high degree of reproducibility for both the RiboAmp ( = 0.950, σ 2 = 0.0004, CV = 2.09%) and the BDSmart ( = 0.895, σ 2 = 0.0010, CV = 3.47%) protocols. These results suggest that subsequent comparisons of our amplification method to these commercial techniques are not confounded by individual variability among replicates. Additionally, the strong correlation between independent replicates of our method (see above) and of these commercial techniques suggests that performing additional independent replicates would not add significant statistical power to subsequent comparisons. The average yield from our method ( = 23.6 μg) was less than that of the Arcturus RiboAmp method ( = 71.6 μg, p < 0.04) but greater than that of the BDSmart method ( = 4.87 μg, p = 1.7 × 10 -14 ). The decreased yield versus the Arcturus RiboAmp method is expected, because IVT with the T7 promoter (production of antisense RNA, Arcturus method) proceeds more efficiently than IVT with the T3 promoter (production of sense RNA, our method). The improved yield over the BDSmart method is notable considering that both of these methods involve production of amplified sense RNA (Table 2 ). The median length of the amplified RNA from our method (794 nt) was greater than that of the Arcturus RiboAmp method (507 nt, p < 0.003) and comparable to that of the BDSmart amplification method (764 nt, p = 0.88) (Table 2 ). The correlation between expression profiles of amplified and unamplified RNA was 10.6% better following amplification with our method ( = 0.8009) versus amplification with the Arcturus RiboAmp method ( = 0.7202, p < 0.01) and was comparable to the correlation after amplification with the BDSmart method ( = 0.7679, p = 0.15) (Table 3 ). Discussion The emergence of methods to study the transcriptome initially necessitated the production of relatively large quantities of RNA from experimental systems, and Eberwine and Van Gelder's T7-based system of In vitro Transcription 5–6 became the first standard protocol for RNA amplification designed to fulfill these requirements. The evolution of microarray technology and other RNA techniques have imposed requirements on the type, length, and/or orientation of the starting RNA. Additionally, the growing interest in medical and microbial genomics now requires that data be collected from samples that are becoming progressively smaller and more difficult to acquire. This changing role for RNA amplification has catalyzed the development of multiple institutional and commercial amplification protocols, all of which claim to be high-yield, low-bias techniques capable of amplifying RNA with the specific characteristics required for downstream applications. Template switching, terminal continuation, and other novel techniques for RNA amplification that have been developed over the past five years have become increasingly technically complex, and we feel that this is a significant disadvantage. In particular, they stray from the Eberwine approach, which has been extensively validated, and often rely on the use of proprietary (i.e. unspecified) enzymes, primers, and reaction components, which limit the utility of the methods (as well as protocol modification and fine-tuning). While proponents of these methods argue that the limitations are offset by improved yield, production of long transcript, reproducibility, and reduced amplification bias, we have found in practice that no single protocol is able to perform consistently in all four areas. The technique that we have described affords the advantages of all of the aforementioned protocols while eliminating the major limitations and controlling the technical complexity. Several aspects of this protocol make it a robust tool that we believe will be useful in conjunction with a variety of experimental systems and downstream applications. First, the method we present is based upon the Eberwine technique. The methods are both straightforward and validated and do not require the use of custom enzymes, multiple (more than 2) proprietary primers, or PCR steps to complete the amplification. Second, this new method produces amplified mRNA in the sense orientation and can be used for production of aRNA with minor modification. This corrects for the downstream limitations imposed by antisense amplification in both the original Eberwine method and the Arcturus RiboAmp kit and allows amplified mRNA produced by our method to be used in conjunction with both cDNA (sense and antisense orientation) and long oligonucleotide (sense orientation) arrays. Third, our method requires only one round of amplification to produce RNA in the sense orientation. We believe that this represents a significant advantage over previously described, T7-based amplification strategies (most notably the method of Xiang [ 11 ]), where the modifications necessary for producing sense RNA require a minimum of two rounds of amplification. Fourth, our method produces sufficient amplified sense RNA for multiple microarray analyses after a single round of amplification from as little as 1 μg of total RNA. This yield is significantly higher than the BDSmart amplification kit and we have demonstrated that our method produces amplified mRNA that performs at least as well as the BDSmart amplification kit and superior to the Arcturus RiboAmp protocol in microarray applications. Another advantage of our protocol is that it is readily adaptable to a variety of experimental systems. The approach that we have described synthesizes amplified RNA from eukaryotic total RNA in the sense orientation. However, because this method adds unique promoters to both the first and second cDNA strand (T7 and T3, respectively), either orientation of RNA can be produced from the same population of cDNA simply by selecting the appropriate RNA polymerase for IVT. Moreover, changing the two primers allows the protocol to be adapted to produce sense or antisense RNA from either prokaryotic or eukaryotic total RNA without any additional adjustments. The protocol can also be modified for use with either T3 or T7 IVT systems by changing the promoter sequence of the promoter-modified primer, it can be used to salvage partially-degraded RNA (as described by Xiang [ 11 ]), and it can be used for indirect RNA labeling protocols by substituting modified ribonucleotide bases into the IVT mix. Table 4 summarizes the primer pairs used for these alternate strategies, and Table 1 gives the necessary primer sequences. We have successfully tested a variety of these strategies and have achieved similar degrees of bias, reproducibility, and yield (data not shown). We believe that the ease with which this protocol can be adapted to a variety of experimental systems is a major advantage. Table 1 Primer sequences for study protocol and its variations. Primer Name Sequence Concentration (ng/μL) T3N9 5'-GCGCGAAATTAACCCTCACTAAAGGGAGANNNNNNNNN-3' 100 T7N9 5'-GGCCAGTGAATTGTAATACGACTCACTATAGGGAGGCGGNNNNNNNNN-3' 100 Oligo(dT) 24 5'-TTTTTTTTTTTTTTTTTTTTTTTT-3' 100 T3-Oligo(dT) 24 5'-GCGCGAAATTAACCCTCACTAAAGGGAGATTTTTTTTTTTTTTTTTTTTTTTT-3' 100 T7-Oligo(dT) 24 5'-GGCCAGTGAATTGTAATACGACTCACTATAGGGAGGCGGTTTTTTTTTTTTTTTTTTTTTTTT-3' 100 Random Hexamer 5'-NNNNNN-3' 3000 Table 4 Primer strategies for alternate versions of the study protocol. Type of Total RNA First Primer Second Primer IVT System † Type of Product Orientation of Product Eukaryotic Oligo(dT) 24 , T7-Oligo(dT) 24 T3N9 T3 mRNA Sense T3-Oligo(dT) 24 Random Hexamer T3 mRNA Antisense* Prokaryotic (or Partially Degraded*) Random Hexamer T3N9 T3 Total RNA Sense T3N9 Random Hexamer T3 Total RNA Antisense † T7 IVT systems may be used by changing the promoter sequence of the modified primer to a T7 binding sequence Modified Eberwine Amplification ** As described by Xiang 11 In clinical and research settings using laser capture microdissection or other strategies where very small amounts of starting RNA (10–100 ng) are available, multiple rounds of amplification may still be necessary. A final advantage of our protocol, as discussed previously, is that a second round of amplification can be easily incorporated and can be used to produce additional RNA in either the sense or antisense orientation. Future testing of this protocol will focus on the quantity and fidelity of the RNA population after multiple rounds of amplification and will include appropriate comparisons to commercial, multi-round techniques. However, the purpose of the present technique is to minimize bias while maintaining yield after a single round of amplification. In this setting, our results show that our method outperforms two commercial, single-step protocols when using identical quantities of initial RNA. We believe the performance and versatility of this method make it particularly beneficial to investigators attempting to conduct microarray analysis from 1–4 μg of initial RNA. Neither this nor any other current strategy for RNA amplification is optimally suited to every experimental setting, and each technique has its own advantages and limitations. The challenge of selecting an amplification strategy is choosing a method that provides sufficient amplification and appropriate RNA orientation for downstream requirements yet minimizes the labor, cost, and bias introduced by the process. The method that we have described has been designed and validated for use in experiments where microarray analysis is to be performed with approximately 1 μg of starting total RNA. In this setting, our method produces a population of amplified RNA having minimal and consistent amplification bias, orientation compatible with both spotted cDNA and oligonucleotide arrays, and quantity sufficient for several downstream assays from a single amplification. Additionally, the protocol we present here can be easily adapted for use in a variety of experimental systems. This includes production of either sense or antisense RNA in prokaryotic or eukaryotic systems (simply changing one or both primers). None of these variations require multiple rounds of amplification to produce RNA of a specific orientation, but the protocol can easily be modified to enable multiple rounds when necessary. Finally, we would like to note that there are some limitations to the method we have outlined here. Readers may note that several amplification strategies have been reported to produce greater than 1,000-fold increases in RNA yield, one aspect of our method that may be viewed as a limitation is its ~130-fold amplification. However, most higher-yield amplification strategies use multiple rounds of amplification and are subject to additional amplification bias. While we recognize that researchers using submicrogram quantities of starting material may consider strategies incorporating multiple rounds of amplification, in our experience even small tissue biopsies yield more than 1 μg of total RNA. The ability to amplify this quantity of RNA efficiently and effectively was one of our major design considerations. In this setting, 130-fold amplification produces more than enough RNA for multiple downstream microarray analyses and confers the benefit of reduced bias inherent in a single-round amplification protocol. Conclusion We have developed a robust method for amplification of mRNA in the sense orientation. This protocol is based upon the Eberwine method and incorporates the advantages of more recent amplification techniques while eliminating many of the limitations of these strategies. Our method allows the production of sense mRNA with one round of IVT, yields 130-fold amplification, preserves long transcript, and produces mRNA that is well suited for downstream microarray applications. Microarray assays performed with RNA amplified using our protocol demonstrate that the method results in low amplification bias and is highly reproducible. Additionally, our method is readily adaptable for the production of sense or antisense, labeled or unlabeled RNA from intact or partially-degraded samples of prokaryotic or eukaryotic total RNA. The method outperforms several commercial RNA amplification kits and can be used in conjunction with a variety of downstream microarray platforms (cDNA arrays, oligonucleotide arrays, Affymetrix GeneChip™ assays). We feel that these advantages make our method a robust tool with the potential for application in a variety of research settings. Methods Overview of the method The protocol that we have designed is based on the Eberwine method but uses novel priming strategies to produce amplified RNA in the sense orientation. First-strand cDNA synthesis from total RNA is primed with an Oligo(dT) 24 primer followed by heat denaturing and then cooling to facilitate primer annealing. Reverse transcription produces cDNA-mRNA hybrids which are then subjected to alkali hydrolysis to remove template mRNA. First-strand cDNA (fs-cDNA) is separated from residual enzymes, nucleotides, and mRNA fragments using a spin column technique. Second-strand cDNA synthesis is primed with random nonamers containing an upstream T3 promoter sequence (T3N9). The sample is heated to denature the fs-cDNA and to eliminate secondary structure. The temperature is rapidly dropped to the upper limit of the annealing range and then ramped more slowly to a final temperature of 4°C. We believe that this strategy minimizes "self-priming" of the fs-cDNA and, in the absence of competitive priming from RNA fragments, facilitates optimal annealing of the second-strand primer. Second-strand cDNA (ss-cDNA) synthesis is carried out using E. coli DNA polymerase and ligase followed by blunt-ending with T4 DNA Polymerase. The double-stranded cDNA is purified using a spin column technique, and in vitro transcription (IVT) from the T3 promoter sequences incorporated into the ss-cDNA produces amplified RNA in the sense orientation (Figure 1 ). Figure 1 Overview of study method for amplification of RNA in the sense orientation Cell line RNA Total RNA used in this study was derived from three cell lines: PA-1 ovarian teratocarcinoma, CaOV-3 ovarian adenocarcinoma, and U118MG glioblastoma. All cell lines were grown in a monolayer under Dulbecco's minimum essential medium (DMEM) with 10% fetal bovine serum (FBS) and 1% penicillin/streptomycin. Cell lines were held at 37°C in an atmosphere of 4% CO 2 and maintained accordingly. When culture plates achieved ~90% confluence, RNA was extracted and purified following TIGR standard operating procedures based on the Trizol method (Invitrogen) [ 30 ]. RNA extracted from all three cell lines was pooled to create a RNA mixture with a final relative composition of 42.56% CaOV-3 RNA, 29.89% PA-1 RNA, and 27.55% U118MG RNA. Purity was verified by measurement of OD 260/280 and OD 230/260 in TE (pH 8.0) using a NanoDrop spectrophotometer, and RNA integrity was assessed by measurement of the 28S/18S ratio with the Agilent Bioanalyzer 2100 using the Agilent Total RNA Nano chip assay. The final values for the RNA pool were: OD 260/280 = 2.04, OD 230/260 = 2.20, 28S/18S = 2.13, indicating high purity, non-degraded RNA. Superas-In RNAse inhibitor (Ambion) was added to a final concentration of 1U/μL according to manufacturer's recommendations. RNA was stored at -80°C until it was used in the RNA amplification process. RNA amplification protocol All temperature controlled steps in this protocol were conducted using a Peltier thermal cycler (PTC225 DNA Engine Tetrad, MJ research) in thin-walled, nuclease-free PCR tubes (BioRad). All enzymes, buffers, and other reaction components were purchased from Invitrogen unless otherwise noted. Four (4) μg of pooled total cellular RNA (described above) was as starting material for the amplifications. Priming for first strand cDNA synthesis is carried out by combining the total RNA with 1 μL of T7-Oligo(dT) primers (100 ng/μL, Ambion) and 1 μL of Superas-In RNAse inhibitor (20U/μL, Ambion). The mixture is diluted to 17.8 μL in DEPC treated water, heated at 70°C for 10 minutes, and cooled to 4°C for five minutes. First strand cDNA synthesis is accomplished by adding 6 μL of 5× First Strand Buffer, 3 μL 0.1M DTT, 1.2 μL of dNTP mix (50 mM), and 2 μL of SuperScript II reverse transcriptase followed by incubation at 42°C for 2.5 hours. After incubation, the reaction is cooled to 4°C for 2 minutes. Template RNA is hydrolyzed by adding 10 μL of 1N NaOH (Sigma) and 10 μL of 0.5M EDTA (Sigma) followed by incubation at 65°C for 15 minutes. The pH is subsequently neutralized by adding 10 μL of 1N HCl (Sigma). Hydrolyzed RNA and residual dNTPs are removed using the Minelute reaction cleanup kit (Qiagen) according to the manufacturer's protocol. Two rounds of elution from the column, each in 10 μL of Buffer EB (Qiagen) are performed to improve recovery. Priming for second strand cDNA synthesis is accomplished by combining the eluted first-strand cDNA with 2 μL of random nonamer primers modified by the addition of a T3 promoter sequence at the 5' end (T3N9 primer, 100 ng/μL, Table 1 ) 11 (Operon) followed by incubation at 95°C for 3 minutes. Immediately following this incubation, the samples temperature is decreased as quickly as possible to 50°C followed by a temperature ramp from 50°C to 4°C at the rate of 0.4°C/s (~120s ramp time) and then a hold at 4°C for 2 minutes. Second strand cDNA synthesis is accomplished by adding 7μL of 5× Second Strand Buffer, 2 μL of dNTP mix (50 mM), 4 μL of E. coli DNA polymerase I (10U/μL), and 1 μL of E. coli DNA ligase (10U/μL) followed by incubation at 16°C for 2 hours. After incubation, blunt-ending is carried out by adding 2 μL of T4 DNA polymerase (5U/μL) and incubating at 16°C for 5 minutes. The reaction is terminated by adding 3.5 μL of 0.5 M EDTA (Sigma). Double stranded cDNA is purified using the Minelute reaction cleanup kit (Qiagen) following the manufacturer's protocol. Two rounds of elution with Buffer EB (Qiagen) are carried out to improve yield. The first elution is in 10 μL of Buffer EB, and the second is in 7 μL of Buffer EB. This provides the eluted cDNA in a final volume of 16 μL (corrected for 1 μL of retention on the column), which is the appropriate volume for use in the subsequent IVT reaction. In vitro transcription is achieved using the T3 MegaScript kit (Ambion). A double reaction is used to improve yield, so 16 μL of NTP mix, 4 μL of 10× Reaction Buffer, and 4 μL of T3 MegaScript Enzyme Mix are added to the sample. The reaction is incubated at 37°C for 14 hours (overnight). The amplified RNA is purified using either the RNEasy Mini or the RNEasy Minelute kit (Qiagen) according to the manufacturer's protocol. Superas-In RNAse inhibitor (Ambion) is added to the amplified RNA to a final concentration of 1U/μL according to the manufacturer's recommendations. Assessment of products and intermediates cDNA and RNA quantity and purity were assessed by measurement of OD 260/280 and OD 230/260 in the corresponding elution buffers (pH 8.0) using a NanoDrop spectrophotometer. RNA length distribution and integrity were assessed with the Agilent Bioanalyzer 2100 using the Agilent Total RNA Nano chip assay. Agilent's Bioanalyzer Expert 2100 software was used to quantify transcript length, and the software's automatic integration tool was used to determine the area under the curve. The median transcript length for each method was taken as the point at which half of the integrated area was achieved. Microarray analyses Microarray analysis was conducted to determine the magnitude of amplification bias. Pooled human cell line RNA (as described above) as well as RNA extracted from one human glioblastoma were amplified as described. The amplified sense RNA (2 μg) was used as the starting material for synthesis of cDNA target, and a spiking control consisting of RNA transcripts of 10 genes from the Arabidopsis thaliana genome was added to each sample prior to cDNA synthesis in order to provide a consistent positive reference in subsequent hybridizations 28 . cDNA target synthesized from the cell line RNA was indirectly labeled with Cy-5 while target derived from the glioblastoma was labeled with Cy-3. The labeled cDNA from the cell line pool and from the glioblastoma were cohybridized to human 32,000 element spotted cDNA arrays. The array production, cDNA synthesis, target labeling, and hybridization were conducted following TIGR standard operating procedures [ 31 ]. Independent replicates were conducted for all amplifications and hybridizations. In order to compare our method to commercially-available RNA amplification kits, microarray analysis was conducted using identical pooled cell line and glioblastoma RNA that was amplified with two commercial products (RiboAmp kit, Arcturus; and BDSmart mRNA Amplification kit, BD Biosciences). Amplification of 4 μg of total RNA was carried out according to the manufacturer's protocol for both kits, and 2μg of amplified RNA from each sample was used for microarray analysis as described above. Independent replicates were conducted for all amplifications and hybridizations. Unamplified total RNA from both the pooled cell line and the glioblastoma was used as the starting material for the control arrays. Ten (10) μg of total RNA from each sample was used for cDNA synthesis, and the remainder of the protocol was carried out as described for the amplified samples. Data analysis Microarray data was analyzed using the TIGR TM4 software package for microarray analysis [ 32 ]. TIGR Spotfinder was used to isolate spots on the arrays, correct for background, and assess the reliability of spots for downstream analysis. TIGR MIDAS was then used to adjust the data by applying LOWESS normalization [ 33 , 34 ] followed by standard deviation regularization [ 33 , 35 ]. The Cy5 channel (pooled cell line sample) was taken as the reference for all transformations. Normalized values of fluorescence intensity (FI) were log 2 transformed, and was calculated for each array element. Assessments of bias were conducted by comparing the log 2 (ratio) values for each element in the amplified sample arrays to their counterparts in the unamplified control arrays. Measurements of reproducibility were conducted by comparing the log 2 (ratio) values of corresponding elements in independent replicates. Statistical methods The degree of bias introduced by an amplification method was assessed by plotting against the and calculating the correlation coefficient ( r ) for each independent replicate. The mean correlation coefficient ( ) was calculated for each method, and differences between for each amplification method were tested for statistical significance. Welch's T-Test [ 36 ] was used for all comparisons to control for the possible heteroscedastic nature of the original array data. The degree of reproducibility of our amplification method was assessed by plotting the log 2 (ratio) values for paired hybridizations derived from independent amplifications of the same starting RNA, calculating the correlation coefficient ( r ) for a series of independent replicates, and computing the mean correlation coefficient ( ). Reproducibility was further investigated by calculating the variance (σ 2 ) and the coefficient of variation for (CV), which describe the dispersion of the values contributing to . Authors' contributions This method is based on an idea by Norman Lee. John Quackenbush and Nicholas Marko further developed the approach. Marko reduced this to practice in the laboratory, greatly expanding on the method and creating a robust protocol. Bryan Frank contributed significantly to the development, implementation, and optimization of the protocol. Marko conducted the data analysis and was largely responsible for the present manuscript.	/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC554769.xml
543573	Population genetic structure of Anopheles gambiae mosquitoes on Lake Victoria islands, west Kenya	Background Understanding the genetic structure of island Anopheles gambiae populations is important for the current tactics in mosquito control and for the proposed strategy using genetically-modified mosquitoes (GMM). Genetically-isolated mosquito populations on islands are a potential site for testing GMM. The objective of this study was to determine the genetic structure of A. gambiae populations on the islands in Lake Victoria, western Kenya. Methods The genetic diversity and the population genetic structures of 13 A. gambiae populations from five islands on Lake Victoria and six villages from the surrounding mainland area in the Suba District were examined using six microsatellite markers. The distance range of sampling sites varied between 2.5 and 35.1 km. Results A similar level of genetic diversity between island mosquito populations and adjacent mainland populations was found. The average number of alleles per locus was 7.3 for the island populations and 6.8 for the mainland populations. The average observed heterozygosity was 0.32 and 0.28 for the island and mainland populations, respectively. A low but statistically significant genetic structure was detected among the island populations (F ST = 0.019) and between the island and mainland populations (F ST = 0.003). A total of 12 private alleles were found, and nine of them were from the island populations. Conclusion A level of genetic differentiation between the island and mainland populations was found. Large extent of gene flow between the island and mainland mosquito populations may result from wind- or human-assisted dispersal. Should the islands on Lake Victoria be used as a trial site for the release program of GMM, mosquito dispersal between the islands and between the island and the mainland should be vigorously monitored.	Background Despite 50 years of malaria vector control efforts, malaria remains a major public health threat in tropical and subtropical countries [ 1 - 3 ]. In recent years, malaria has caused increased human mortality and morbidity as malaria epidemics have spread to areas where it was previously rare [ 4 , 5 ]. The current strategies for malaria control involve the treatment of infected individuals with antimalarial drugs to kill the parasites and vector management to reduce human-vector contacts via residual spraying and the use of insecticide-impregnated bednets. As demonstrated in multisite trials throughout Africa, the large-scale use of insecticide-treated bednets can reduce overall mortality by up to 30% [ 6 ] and morbidity in young children [ 7 ]. The emergence of insecticide resistance in mosquito vectors [ 8 ] and antimalarial drug resistance in Plasmodium [ 9 ] has significantly reduced the viability of many malaria control programs. An efficacious malaria vaccine will not be available in the near future [ 10 ]. One potential alternative malaria control strategy is based on the genetic disruption of mosquito vector competence [ 11 - 13 ]. This genetic control approach requires identification and cloning of parasite-inhibiting genes in the mosquito vectors, development of stable and efficient mosquito transformation tools and the development of strategies for spreading the parasite-inhibiting genes. Over the past several years, remarkable progress has been made in the development of mosquito germline transformation and in the identification of parasite-inhibiting molecules. For example, A. gambiae cell lines were successfully transformed with the Hermes element [ 14 , 15 ], and the Minos transposable element bearing an exogenous gene was efficiently integrated into the genome of Anopheles stephensi [ 16 , 17 ]. Genetic linkage maps have been constructed for A. gambiae [ 18 ], and genes conferring mosquito refractoriness to malaria parasites have been mapped [ 19 ]. Availability of complete A. gambiae genome sequences will greatly facilitate identification and cloning of parasite-inhibiting genes [ 20 ]. The success of the transgenic mosquito approach depends on the spread and even fixation of parasite-inhibiting genes into natural populations. Presently, releasing transgenic mosquitoes to the field is premature. Isolated islands have been suggested as an ideal natural site for testing transgenic mosquito release strategies and spatial spreading of transgenes [ 13 , 21 , 22 ]. Information on mosquito population genetic structure and gene flow on islands and the surrounding mainland area is critical. Using microsatellite markers, the A. gambiae population genetic structure in the African continent has been examined [ 23 - 28 ]. These studies revealed that the Great Rift Valley in East Africa is a substantial gene flow barrier for A. gambiae ; however, no significant genetic structure was detected for mosquito populations between western Kenya and West Africa. The minimum area associated for a deme of A. gambiae in western or coastal Kenya is larger than 50 km [ 24 ]. Simard et al . [ 29 ] found a high degree of genetic differentiation of the Anopheles arabiensis populations from the high plateau of Madagascar and those from Réunion and Mauritius islands (F ST ranges from 0.080 to 0.215). Population substructure was also detected on the island of São Tomé, West Africa [ 22 ]. The present study examined the genetic diversity and the population genetic structures of A. gambiae mosquitoes from five islands on Lake Victoria and the surrounding mainland in western Kenya. This information is valuable for selecting field sites to test transgene release strategies and evaluating the spread of transgenes in nature. Materials and Methods Study sites and mosquito collection Anopheline female mosquitoes were collected from seven villages on five islands in Lake Victoria and from six villages in the mainland Suba District, western Kenya (Fig. 1 ). The sampled islands were Kibuogi, Mfangano (Sena village), Ngodhe, Takawiri and Rushinga. Mosquitoes were collected from three villages (Kamsengere, Utajo and Wanyama) on Rushinga Island and one village on each of the other islands. Mfangano Island is the largest and the most offshore (about 10 km away from the nearest mainland village). Rushinga Island is the most populated among the five islands and is connected to the mainland by a walkway. The islands are about 2.5–21.0 km apart. Also, five mainland villages (Ragwe, Roo, Gingo, Mbita and Kasunga) along the shore of Lake Victoria and one inland village (Ruri) about 11 km away from the lakeshore were selected. The distance between the islands and the mainland sites ranges from 4.9 to 35.1 km. Malaria on these islands and the mainland area is holoendemic, and A. gambiae mosquitoes are the major malaria vectors in this region [ 30 ]. Figure 1 Map of study area showing the distribution of Anopheles gambiae populations on the Lake Victoria islands and the surrounding mainland area in Suba District, western Kenya. At least 170 anopheline mosquitoes were collected from four to 28 houses within each village using the pyrethrum spray collection method [ 31 ]. Mosquitoes from Mbita, Kasgunga and Ruri were sampled in May 1997; collection in other villages was conducted in April and May 1999. A. gambiae sensu lato ( s.l .) specimens were separated from other anophelines according to the identification key provided by Gillies and Coetzee [ 32 ] and then preserved in 95% ethanol and kept at -20°C until further analyzed. PCR assay for species identification PCR analysis was conducted for species identification using the rDNA-PCR method because individual species within the A. gambiae species complex cannot be identified by morphology alone [ 33 ]. About 100 A. gambiae s.l . females per village were tested. If the initial PCR testing failed to amplify for a sample, then the PCR analysis was repeated once or twice until successful amplification was achieved. If a sample could not be identified after three PCR amplifications, it was scored as unknown. Microsatellite loci and genotyping Six microsatellite markers were used for specimen genotyping, including AGXH1D1 and AGXH131 of Chromosome X, AG2H46 and AG2H79 of Chromosome 2, and AG3H29C and AG3H33C of Chromosome 3 [ 17 , 18 , 22 , 23 ]. Microsatellite analyses were conducted on 51–70 individuals per village (See the additional file : A table of sample size, allelic number, heterozygosities and breeding coefficient of 13 A. gambiae populations from the Lake Victoria islands and the surrounding mainland in western Kenya). A Li-Cor Model 4200 Automated DNA Analyzer (Li-Cor Inc., Lincoln, NE) was used for gel electrophoresis. For the apparatus to detect PCR products, one primer in every pair of microsatellite primers must be fluorescently labelled. To reduce the cost associated with synthesis of fluorescently labelled primers, we used the "tailed primer" method [ 34 , 35 ]; that is, the forward primer for each microsatellite locus was synthesized with an additional 19 bp sequence (5' CACGACGTTGTAAAACGAC 3') added to the 5' end of the primer. A third primer with the same 19 bp sequence was directly labelled with the fluorescence and was used as the sole type of labelled primer for the detection of all microsatellite alleles. The tailed primer method reduced the cost of oligonucleotide synthesis by >80%. The 10 μl PCR reaction contained 1X Taq buffer, 0.2 mM dNTPs, 1.5 pmol forward and reverse primers, 1.5 pmol fluorescently labelled 19 bp sequences, 1.5 mM MgCl 2 , 1.0 μg BSA, 1.0 unit Taq polymerase and about 20 ng genomic DNA. Cycling conditions in a MJ Research PTC-220 thermocycler were 35–40 cycles of 94°C for 30 seconds, 55°C for 30 seconds and 72°C for 45 seconds. Allele sizes were determined using Gene ImagIR computer software [ 36 ]. The allele sizes used in the analysis were true allele sizes that have been adjusted for the 19 bp tail in the forward primer. Data analysis Microsatellite polymorphism was measured by the number of alleles and heterozygosity at each locus. Using the probability test available in the GENEPOP computer program [ 37 ], conformance with Hardy-Weinberg Equilibrium (HWE) was tested for each locus and population, and the Bonferroni correction was applied for multiple comparisons. The F IS statistics and probability test were used to determine whether distortion from HWE resulted from heterozygosity deficiency or excess using. Because the probability test is robust to low allele frequencies, rare alleles were not pooled. Variations in heterozygosity among the populations were analyzed following Weir's method [ 38 ], using the analysis of variance (ANOVA) with subpopulations, individuals, loci and interactions of loci, and individuals as factors. All factors were treated as random effects except loci. The Fisher exact test was performed to detect linkage disequilibrium for pair-wise loci in each population and the pooled population. Population genetic structure was examined with Wright's F-statistics (F ST ) using FSTAT 2.8 [ 39 ]. F ST statistic appears to be more sensitive to detect intraspecific differentiation than R ST [ 40 , 41 ]. The standard deviations of the F-statistics were obtained for each locus by a jackknife procedure over all the alleles and were used to test the statistical significance. Nei's unbiased genetic distances [ 42 ] were calculated for all pairs of populations based on microsatellite allele frequencies at six loci using TFPGA [ 43 ]. A dendrogram was created based on the pair-wise genetic distances using the unweighted pair group method with arithmetic mean (UPGMA). The bootstrap confidence values were generated by 1,000 permutations. The isolation-by-distance model of population genetic structure was tested by linear regression of pair-wise F ST /(1 - F ST ) against the natural logarithm of straight-line geographical distance between population pairs [ 44 ]. Statistical significance of the regression was tested using the Mantel test with 10,000 permutations [ 45 ]. Results Population genetic variability A moderate to high level of polymorphism was found in six loci across the 13 populations (See the additional file ). The three populations from Rushinga Island had a similar number of alleles per locus (ANOVA, F = 0.02, df = 2, P > 0.05) and observed heterozygosities (F = 0.029, df = 2, P > 0.05). Among the island populations, the average observed number of alleles per locus was not significantly different (F = 0.08, df = 6, P > 0.05), but observed heterozygosity varied significantly (F = 4.52, df = 6, P < 0.01). The three populations on Rushinga Island, Kamsengere, Utajo and Wanyama, showed significantly lower heterozygosity than other islands. Similarly, the six mainland populations did not differ in the number of alleles per locus (F = 0.29, df = 5, P > 0.05), but they varied significantly in the observed heterozygosities (F = 5.45, df = 5, P < 0.01). In particular, the Ruri population had the highest observed heterozygosity (0.343), about two-fold higher than the Mbita population (See the additional file ). Overall, there was no significant difference between the island and mainland populations in the number of alleles per locus (7.3 vs. 6.8; t = 0.67, df = 74, P > 0.05) and observed heterozygosities (0.32 vs. 0.28; t = 1.82, df = 74, P > 0.05). A total of 12 private alleles were identified, nine of them from the island populations. A total of 14.1% loci (11 out of 78 tests) showed significant departure from Hardy-Weinberg equilibrium, all due to heterozygote deficiency. This was caused entirely by heterozygote deficiency in the locus AG2H46, a locus known for the presence of null alleles in western Kenyan A. gambiae populations [ 46 ]. The Fisher exact test revealed linkage disequilibrium in 13 out of 195 pairs of loci (6.7%; data not shown), suggesting a low level of linkage disequilibrium among the six loci scored. Population genetic structure A low, but significant, genetic structure was detected among the seven island and the six mainland populations (Table 1 ). The genetic differentiation in the seven island populations (F ST = 0.019, P < 0.001) was almost twice as high as the six mainland populations (F ST = 0.010, P < 0.001). Genetic differentiation between island and mainland populations was also small (F ST = 0.010, P < 0.001). Pair-wise comparisons between all populations revealed that only seven pairs (Kibougo/Kamsengere, Kasgunga/Kamsengere, Takawiri/Ruri, Sena/Ruri, Utajo/Ruri, Ngodhe/Ruri and Ngodhe/Gingo) exhibited significant F ST values, and six of them were between an island and a mainland population. Table 1 F ST estimates of Anopheles gambiae populations on the islands of Lake Victoria and from surrounding mainland sites in western Kenya Locus Among seven island populations Among six mainland populations Between island and mainland areas Among all populations AGXH1D1 0.082* 0.022* 0.002* 0.042* AGXH131 0.000 0.009* 0.005* 0.006* AG2H46 0.008* 0.006 0.006* 0.009* AG2H79 0.011* 0.011* 0.000 0.010* AG3H29C 0.026* 0.008* 0.000 0.007* AG3H33C 0.003* 0.009* 0.000 0.005* Overall 0.019* 0.010* 0.003* 0.012*** * P < 0.05, P < 0.01, * P < 0.001. The Mantel test revealed a significant correlation between geographic distance and pair-wise F ST /(1 - F ST ) (P < 0.001), suggesting that the population genetic structure of A. gambiae populations from the island and mainland is consistent with the isolation-by-distance model. When the Ruri population is removed from the analysis, the correlation was still statistically significant (P = 0.015). Therefore, the population genetic structure of our study populations is consistent with the isolation-by-distance model. The cluster analysis revealed that the Ruri population, located farther inland from the other populations, was out-grouped from other populations with a significant bootstrap value, while other mainland and island populations were intermixed with non-significant branch bootstrap values (Fig. 2 ). Figure 2 A UPGMA tree based on Nei's unbiased distance showing genetic divergence among Anopheles gambiae populations. The numbers above branches indicate those with >50% bootstrap support. The populations marked with an asterisk are the mainland populations. Discussion The present study demonstrated a similar level of genetic diversity between the island A. gambiae populations in the Lake Victoria and adjacent mainland populations in the Suba District, western Kenya. For the seven island populations, the average number of alleles at six microsatellite loci was 7.3 and the observed heterozygosity was 0.32. For the six mainland populations, the average number of alleles was 6.8 and the observed heterozygosity was 0.28. The population genetic diversity at most loci in this study was similar to other western Kenyan populations [ 23 , 47 - 49 ]. Compared with West Africa populations [ 23 , 27 , 48 , 49 ], lower heterozygosities, particularly at loci AG2H46, AG2H79 and AG3H33, were reported in this study, caused by fewer alleles detected in the studied populations. The comparable level of genetic diversity between island and mainland populations suggests that the island mosquito populations have a similar effective population size as the mainland populations, and they have not suffered severe genetic bottleneck during the previous vector control efforts. For each population, all loci except the AG2H46 locus did not show a significant deviation from Hardy-Weinberg equilibrium, suggesting that the microsatellite markers used in the study are not under strong selection and mosquito populations are in random mating. A heterozygote deficit at the locus AG2H46 was observed for all populations in this study. Heterozygote deficiency at the locus AG2H46 was also demonstrated in other western Kenya populations by Lehmann et al . [ 23 , 24 ]; the presence of null alleles as a result of mutations in the primer-annealing region was the cause. A small but statistically significant genetic structure was detected for A. gambiae populations among the five islands in Lake Victoria (F ST = 0.019) and among the six villages in the mainland in an area of approximately 40 × 20 km 2 (F ST = 0.010). The degree of genetic differentiation between the island populations in this study was less than for the island A. gambiae populations of São Tomé, western Africa (F ST = 0.032) [ 22 ]. The lower F ST estimates in the populations in this study were probably caused by shorter distance between islands (3–15 vs. 23–38 km) [ 22 ] and a lack of mountainous topography as gene flow barriers. The F ST estimates for the mainland populations in this study were comparable to other studies on the western Kenya populations (F ST = 0.0033) [ 24 , 27 ]. The genetic differentiation between island and mainland populations was small but statistically significant (F ST = 0.003). Thus, there is a very small degree of genetic isolation between island and mainland populations. This estimation is consistent with the private allele distribution in the studied populations, in which nine of the 12 private alleles were from the island populations. Further evidence for a small degree of genetic differentiation between island and mainland populations is from pair-wise population comparisons in which six out of the seven pairs that exhibited significant genetic differentiation were between an island population and a mainland population. The low level of genetic differentiation between island and mainland mosquito populations implies large gene flow between the two areas (83.1 migrants per generation). The normal flight range of A. gambiae is usually less than 1 km [ 50 ]. The distance to the lake shore of the mainland from the islands ranges from 2.5 to 15 km, farther than the normal flight range of the mosquitoes. Thus, mosquito migration is likely assisted by wind. Lindsay et al . [ 51 ] found that the spatial distribution of A. gambiae mosquitoes was related to the predominant wind direction at night, suggesting that wind assisted the dispersal of mosquitoes from their breeding site. A. gambiae have been shown to fly up to 7 km with the assistance of wind [ 52 , 53 ]. This distance is in the range for mosquitoes to disperse between the closest islands and between islands and their closest mainland in this study area. Mosquitoes may also use one island as a stepping-stone to extend their dispersal distance. Mosquito migration may also be assisted by human activities. A study on Aedes polynesiensis populations from islands found no significant effect of geographic distance on the population genetic structure, but detected a significant correlation between gene flow and commercial traffic by planes and/or boats between islands [ 54 ]. The introduction of A. arabiensis to the Mascarene islands and Madagascar was thought to be caused by human transportation by steamship lines [ 55 , 56 ]. In Lake Victoria, small wooden boats may transport mosquito larvae between the islands and the mainland. A. gambiae larvae were collected at the bottom of a wooden fishing boat [ 57 ]. Rushinga Island in the study area was connected to the mainland by a walkway, and the island mosquito larvae could be moved to the mainland by vehicle transportation. The results of this study of the population genetic stricture of island and mainland A. gambiae populations have implications for the ecological safety evaluation of the transgenic mosquito release program. During the initial field test of environmental safety and public health consequences by transgenic mosquito release, ideal sites would be islands that are totally genetically isolated from other islands and the mainland, with a sufficient number of human inhabitants and active malaria transmission on the island. Such an island may be extremely difficult to find, so islands with some genetic isolation from the mainland may have to be chosen. If so, the Lake Victoria islands could be used as field test sites; however, due to potential gene flow between the islands and between the islands and the mainland, mosquito dispersal between the islands and between the islands and the mainland should be vigorously monitored. After the release of the genetically modified mosquitoes, long-term monitoring programs should be launched to evaluate the spread of the transgenes to any unintended areas. In addition, methods to minimize the negative effects of transgene leak need to be developed prior to the field trial of transgene release [ 58 ]. Conclusions This study showed that a low level of genetic differentiation existed between the island and mainland populations and no any genetically-isolated population was found among the 13 mosquito populations. If the islands on Lake Victoria were used as a trial site for the program to release genetically-modified mosquitoes, short-term and long-term mosquito dispersal between the islands and between the island and the mainland should be vigorously monitored. Authors' Contributions HC conducted species identification using PCR, microsatellite analyses and drafting the manuscript. NM was responsible for sample collection, and participated in species identification and drafting the manuscript. JB and GY supervised the study, and assisted data analysis and manuscript preparations. Supplementary Material Additional File 1 A table of sample size, allelic number, heterozygosities and breeding coefficient of 13 A. gambiae populations from the Lake Victoria islands and the surrounding mainland in western Kenya. Click here for file	/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC543573.xml
552330	Inhibition of human immunodeficiency virus type-1 (HIV-1) glycoprotein-mediated cell-cell fusion by immunor (IM28)	Background Immunor (IM28), an analog of dehydroepiandrosterone (DHEA), inhibits human immunodeficiency virus type-1 (HIV-1) by inhibiting reverse transcriptase. We assessed the ability of IM28 to inhibit the cell-cell fusion mediated by HIV envelope glycoprotein in an in vitro system. For this purpose, we co-cultured TF228.1.16, a T-cell line expressing stably HIV-1 glycoprotein envelopes, with an equal number of 293/CD4+, another T cell line expressing CD4, and with the SupT1 cell line with or without IM28. Results In the absence of IM28, TF228.1.16 fused with 293/CD4+, inducing numerous large syncytia. Syncytia appeared more rapidly when TF228.1.16 was co-cultured with SupT1 cells than when it was co-cultured with the 293/CD4+ cell line. IM28 (1.6 – 45 μg/ml) completely inhibits cell-cell fusion. IM28 also prevented the development of new syncytia in infected cells and protected naive SupT1 cells from HIV-1 infection. Evaluation of 50% inhibitory dose (IC50) of IM28 revealed a decrease in HIV-1 replication with an IC50 of 22 mM and 50% cytotoxicity dose (CC50) as determined on MT2 cells was 75 mM giving a selectivity index of 3.4 Conclusions These findings suggest that IM28 exerts an inhibitory action on the env proteins that mediate cell-cell fusion between infected and healthy cells. They also suggest that IM28 interferes with biochemical processes to stop the progression of existing syncytia. This property may lead to the development of a new class of therapeutic drug.	Background The human immunodeficiency virus type-1 (HIV-1) envelope glycoprotein is composed of two subunits: a surface glycoprotein (gp120) and a trans-membrane glycoprotein (gp41). These two subunits interact with each other in a non covalent manner. Gp120 is critical for attachment to host cell CD4 receptors, whereas gp41 contains the fusion sequence. HIV and simian immunodeficiency virus (SIV) require a co-receptor in addition to CD4 for entry into cells. Primary HIV can use a broad range of co-receptor molecules, including CCR1, CCR2b, CCR3, CCR4 and CXCR4 [ 1 - 3 ]. However, expression of a co-receptor together with CD4 on some cell types does not confer susceptibility to infection [ 1 ]. Not all human cell types that express an appropriate co-receptor support virus replication, indicating that other factors that affect viral tropism are present. HIV-1 viral entry is inhibited in the presence of the ligands to these chemokine receptors. RANTES, MIP-1α and MIP-1β, all of which are ligands for CCR5, inhibit macrophage-tropic isolates, whereas SDF-1, the specific ligand for CXCR4, inhibits entry by T-cell-tropic isolates [ 4 - 6 ]. The ability of HIV-1 envelope glycoproteins to induce cell-cell fusion is an interesting property because molecules that inhibit the fusion process are possible antiviral drugs and may lead to the identification of important functional regions either on the viral glycoprotein or on cell membranes. A hydrophobic, 25-amino acid, conserved segment located at the N-terminus of gp41 and gp120/41 has been shown to be involved in the fusion reaction between the viral envelope and the host cell plasma membrane [ 7 , 8 ]. There is evidence suggesting that this sequence is also involved in the cytopathic process underlying HIV-1 infection of target cells [ 9 , 10 ]. Exposure of this hydrophobic peptide to the aqueous environment in the vicinity of the target cell initially depends on gp120/41 function [ 11 ]. This protein is activated after interacting with primary receptor CD4. This activation requires the presence of human co-factors [ 12 , 13 ]. According to this model, further interaction of the fusion peptide to bind membrane lipid with the cell membrane depends mainly on the ability of the peptide to bind membrane lipid components. Hence, drugs that are able to interfere with membrane proteins became relevant for the therapy of HIV, even though it is still important to inhibit virus replication. We have previously shown that IM28 can inhibit HIV-1 reverse transcriptase activity [ 14 ]. Here, we assessed its capacity to inhibit the fusion of HIV-1-infected cells to naive cells. We found that IM28 was able to inhibit cell-cell fusion in an in vitro system. We showed that IM28 significantly blocks HIV-1 glycoprotein-mediated cell-cell fusion. Results We determined the concentrations of various drugs required to inhibit and to partially inhibit the fusion of TF228.1.16 and 293/CD4+ (Table 1 ). All these drugs decreased the percentage of surface covered by syncytia. The concentration of IM28 (6.43 μg/ml) that inhibited the formation of syncytia was similar to that of DXSF 500 000 (3.52 μg/ml) (Table 1 ). There were no statistical differences between the inhibitory concentrations of any of the drugs tested and IM28. To confirm these observations, we used SupT1 cells because fusion takes place more rapidly in these cells. These cells were mixed with TF22.1.16 cells in the presence or absence of dexamethasone or IM28 and fusion was examined by light microscopy after various periods of co-cultivation. In the absence of dexamethasone or IM28, TF228.1.16 cells fused with SupT1 cells, forming aggregates (Figure 1a ). Infected cells were spindle-shaped with large syncytia after overnight culture (Figure 1b ). Table 1 Effect of drugs on fusion of TF228.1.16 cells to 293/CD4+ cells Treatment Effect § None F F F IM28 F (0.60) P (1.83) I (6.43) Dexamethasone F (0.48) P (1.67) I (5.20) Con A F (0.09) P (0.22) I (0.79) Heparin F (2.70) P (7.00) I (22.0) Suramin F (1.57) P (3.90) I (15.0) Dextran Sulfate 10,000 F (0.02) P (0.06) I (0.20) Dextran Sulfate 500,000 F (0.37) P (1.15) I (3.52) § F = 50–60% of the surface is covered by syncytia; P = partial inhibition of fusion: < 10% of the surface is covered by syncytia; I = inhibition of syncytia formation. TF228.1.16 cells were mixed with 293/CD4 + cells (1:1 cell ratio) and transferred to a 24-well plate (10 5 cells per well in 200 μl of culture medium). TF228.1.16 cells and 293/CD4 + cells were incubated in the presence or absence of the drug (the final concentration in μg/ml is indicated in parenthesis) for 18 h. Following co-culture, three random fields of cells were photographed (not shown) and percentage fusion was determined as previously described [10]. Figure 1 Photomicrograph of SupT1 cells co-cultivated with TF 228.1.16 cells. Cell forming syncytia are aggregated (A). In the presence of dexamethasone (B) cells are mainly exploded vs. in the presence of IM28. In the presence of dexamethasone (Figure 2 ) or IM28 (Figure 3 ), the fusion of TF228.1.16 and SupT1 cells was completely inhibited in a dose-dependent manner. Indeed, in the presence of 0.5 μg/ml dexamethasone or IM28, time of incubation had no effect on syncytia formation. This concentration of dexamethasone or IM28 did not result in the lysis of existing syncytia but stopped the fusion reaction and the appearance of new syncytia (Figure 3 ). The time of incubation did not affect the inhibition of syncytia in the presence of dexamethasone, but did have an effect for 0.5 μg/ml IM28. In addition, the highest concentration (> 0.5 μg/ml) of both drugs completely inhibited syncytia formation. At this concentration of dexamethasone, the inhibition of syncytia was accompanied by cell death bursting (Figure 4 ), whereas the same concentration of IM28 did not lead to the burst (Figure 4 ). Figure 2 Effect of dexamethasone on fusion of TF228.1.16 cells with SupT1 cells. TF228.1.16 cells were mixed with SupT11 cells (1:1 cell ratio) and transferred to a 24-well plate (105 cells per well in 200 ml of cultured medium). After 24 h of co-culture in the presence or absence of dexamethasone (10 mg/ml), three random fields of cells were photographed and the percentage fusion was determined as described in Table 1. Figure 3 Effect of IM28 on fusion of TF228.1.16 cells with SupT1 cells. TF228.1.16 cells were mixed with SupT11 cells (1:1 cell ratio) and transferred to a 24-well plate (10 5 cells per well in 200 μl of cultured medium). After 24 h of co-culture in the presence or absence of corticosteroids (dexamethasone or IM28) (10 μg/ml), three random fields of cells were photographed and the percentage fusion was determined as described in Table 1. Figure 4 Effect of IM28 and dexamethasone on SupT1 cells co-cultured with TF228.1.16. Zoom of negative photomicrograph of SupT1 cultures co-cultivated with TF 228.1.16 cells (A) in the presence of dexamethasone (B) and IM28. Note the evident syncytia in (A) with an apparent slender shape of infected cells. Cells treated with dexamethasone were atrophic and sometimes exploded whereas cells incubated with IM28 were round. To further characterize the biological effect of the drug, the 50% inhibitory dose (IC50) and the cytotoxic dose (CC50) of IM28 were evaluated and the selectivity index which is the CC50/IC50 ratio was determined. The decrease in HIV-1 replication was obtained with an IC50 of 22 mM and the CC50 as determined on MT2 cells was 75 mM giving a selectivity index of 3.4. Discussion IM28 is a potent new derivative of DHEA that can stop the replication of HIV-1 by inhibiting its reverse transcriptase activity [ 14 ]. Here, we show that IM28 can also prevent and inhibit the fusion of infected cells (TF228.1.16 cells) to naïve cells including 293/CD4+ cells, which are stably transfected with human CD4 and highly susceptible to HIV-1 infection, and SupT1 cells [ 15 , 16 ]. The fusion of 293/CD4+ cells with TF228.1.16 cells was completely inhibited by a lower dose of IM28 than was the fusion of SupT1 cells with TF228.1.16 cells (data not shown). The fusion of TF228.1.16 cells to H4CD4+ (CD4 positive glial cell line) cells obtained by transfection of human neuroglioma cells [ 17 ] is also inhibited by IM28 (not shown). Therefore, IM28 and dexamethasone may inhibit cell-cell fusion and recombination-induced fusion mediated by the HIV env protein. Although the precise site at which IM28 acts to inhibit cell-cell fusion remains unknown, our results suggest that IM28 fights the HIV-1 virus at a new site. It is possible that this drug interacts with phospholipase A2 (PLA2), which plays an important role in the entry of HIV virus in the host cell [ 18 , 19 ]. Indeed, dexamethasone, a glucocorticoid, can inhibit the HIV-1, HIV-2 and SIVmac251 envelope glycoproteins and activate PLA2. PLA2 is activated when the envelope glycoprotein interacts with CD4. Due to its local membrane-destabilizing effect, PLA2 may play an important role in preparing the cell membrane for fusion with the viral particle. Activated PLA2 hydrolyzes membrane phospholipids in the sn-2 position, producing arachidonic acid and lysophospholipids [ 20 ]. These biochemical events also have downstream effects; the membrane is destabilized locally [ 21 , 22 ], and arachidonic acid and lysophospholipids are generated. They are potent detergents and may favor fusion [ 23 ]. In addition, arachidonic acid is the precursor of eicosanoids, prostanoids, leukotrienes and lipoxins, which may mediate further activation [ 24 ] and PLA2-induced hydrolysis of ether lipids gives rise to paf-acether [ 25 ]. It is possible that the interaction between gp120 and CD4 specifically modifies the cell membrane locally, preparing it for fusion. We hypothesize that the gp120-CD4-co-receptor complex activates PLA2 through protein kinase C (PKC) and plays a critical role in the fusion of the membrane phospholipids of the host cells and gp41 before viral entry. Indeed, the complex formed by CD4 and p56lck acts as the major receptor for HIV-1, HIV-2 and SIV, delivering intracellular activating signals. This complex binds to the viral envelope glycoprotein gp120. Following this binding, chemokine engagement appears to be required to generate the fusion active form of the envelope protein. This may involve the formation of a gp120-CD4-chemokine receptor complex, in which engagement of the chemokine receptor is dependent on a CD4-induced conformational change in env gp120 [ 26 - 28 ] as previously defined for the number of parameters contributing to fusion, i.e., fusion glycoproteins and the host-cell receptors [ 29 ]. However, further investigations are required to determine the real binding site of IM28. It is possible that IM28 acts on virus replication to inhibit existing syncytia, as previously reported [ 14 ]. Therefore, although the biochemical basis of this phenomenon remains to be discovered, IM28 prevents and inhibits the cell-cell fusion induced by HIV-1, giving it additional beneficial effects. Since differential ability to incorporate or maintain envelope on the virion might account for the differences in cell-to-cell versus cell-free infections in primary isolates, further studies with a more quantitative assay available for determining fusion inhibition as previously described [ 33 , 34 ] may also provide us with a greater understanding of the HIV-1 envelope structure and the HIV entry process. Conclusion In conclusion, our data show that IM28, a potent new analog of DHEA, is able to prevent and to inhibit cell-cell fusion, an important step at the beginning of HIV infection of naive cells, this drug seems to display the required properties for an anti-HIV drug. Methods Cell lines Three cell lines were used: TF228.1.16, which is a BJAB cell line that stably produces functionally active HIV-1 envelope protein (BH-10 clone of HIV-1 LAI) [ 30 ]. 293/CD4+ (human embryonic kidney 293 cells which over express human CD4), obtained through the AIDS Research and Reference Reagent Program; and SupT1 cells, purchased from the American Type Culture Collection (Rockville, MD, USA). Reagents DHEA, dextran-sulfate (DXSF), dexamethasone, suramin, heparin, the mannose-specific lectin concanavalin A (ConA) and Rowell Park Memorial Institute (RPMI)-1640 medium were purchased from Sigma-Aldrich (St Quentin-Fallavier, France). Cells were cultured in complete medium containing L-gltamine, penicillin, streptomycin and fetal calf serum. All these reagents were purchased from Invitrogen (Eragny, France). IM28 was produced from DHEA as specified in its data sheet (INPI 0990847; Fr2792201; Wo0106666; CRPH, Gabon). Fusion and syncytia assays Cultured 293/CD4+ cells in complete medium were harvested by trypsinization. These cells (5 × 10 4 ) were combined with an equal number of TF228.1.16 cells in a 24-well plate and incubated overnight at 37°C in a humidified incubator with 5% carbon dioxide as described by Moore et al. 1993 [ 11 ]. Adherent cells were fixed and stained with diff-quick (Sigma-Aldrich) and then observed under a Leitz microscope. To examine the effect of IM28 on HIV-1 envelope glycoprotein-mediated fusion, 293/CD4+ cells were mixed with TF228.1.16 cells in the presence of IM28. As a positive control for fusion inhibition, cells were incubated in parallel with dexamethasone, ConA, heparin, suramin and dextran sulfate 10 000 or 500 000, compounds known to interfere with mannose residues of envelope glycoprotein on HIV infectivity and HIV and measles virus-induced cell fusion [ 31 , 32 ]. The inhibitory activity of IM28 on fusion of 293/CD4+ cells with TF228.1.16 cells is expressed as a function of concentration and was compared with the inhibitory activity of the above mentioned compounds that interact with the HIV envelope protein. Fusion was examined by light microscopy after co-cultivation for 32 h. The percentage fusion is the ratio of cell surface involved in syncytia to the total cell surface. Syncytia were defined as giant cells, with diameters more than four times bigger than those of single cells. Percentage fusion was divided into three classes: 56–00% of the surface covered by syncytial = fusion; partial inhibition of fusion: < 10% of the surface is covered by syncytial = P; inhibition of syncytia formation = I. Statistical analysis Data were analyzed by one-way analysis of variance (ANOVA) followed by Dunnett' test. All analyses were performed using the Graph-Pad Prism ® computer program. Only P < 0.05 was considered significant. Authors' contributions D M coordinated and participated in the design of the study, statistical analysis and the drafting of the manuscript. V P-M carried out and participated in the biological tests. M-Y A carried out and participated in the biological tests. B O carried out and participated in the biological tests. E M participated in the design of the study, carried out the biological tests and participated in the drafting of the manuscript.	/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC552330.xml
535354	Supracubital perineurioma misdiagnosed as carpal tunnel syndrome: case report	Background Perineuriomas have been defined as tumorous lesions of the peripheral nerves which derive from perineurial cell proliferation and may be associated with abnormalities on chromosome 22. Case presentation Three years after a painful cubital vein procaine injection, a 33 year-old man developed a median nerve lesion, initially diagnosed as carpal tunnel syndrome. Symptoms progressed despite appropriate surgery. Clinical and electrophysiological re-evaluation revealed a fusiform mass at the distal upper arm, confirmed by MRI. Immunohistochemical studies classified the tumor as a mixed perineurioma and neuroma. Conclusions Perineurioma mixed with neuroma may potentially caused by the previous trauma or cytotoxic effects of procaine.	Background In the past years, there has been much confusion concerning the definition of the entity of rare focal lesions of the peripheral nerves since terms such as perineurioma, localized hypertrophic neuropathy, or hypertrophic neuritis have been used as synonyms [ 1 - 5 ]. According to the revised World Health Organization classification of tumors of the nervous system, however, perineuriomas have been defined as tumorous lesions of the peripheral nerves which derive from perineurial cell proliferation. They show strong immunoreactivity for the epithelial membrane antigen and may be associated with abnormalities on chromosome 22 [ 6 - 9 ]. In contrast, localized hypertrophic neuropathy has been defined as a distinct entity which is comprised of Schwann cell-onion-bulb formations, immunohistochemically stains strongly for protein S100, and is epithelial membrane antigen negative [ 9 - 11 ]. Localized hypertrophic neuropathy may be caused by a non-neoplastic undefined stimulus [ 12 , 13 ]. Case presentation Three years after a cubital procaine-HCL 0.05% injection had caused acute severe local pain radiating to his forearm and wrist, a 33-year old man complained about pain in his right hand. Two months later, he also suffered from numbness in the distribution of the right median nerve and wasting of thenar muscles. Furthermore a muscular atrophy in his right forearm was noted. Nerve conduction studies showed an increased distal latency in the median nerve, a very low thenar compound muscle action potential on median nerve stimulation, a reduced sensory conduction velocity of 39 m/s in the median nerve on thumb stimulation and no sensory nerve action potential on stimulation of the 2 nd and 3 rd fingers. Electromyography of the right biceps brachii muscle was normal. In the fibrillations, an increased duration and amplitude of motor unit potentials and reduced recruitment pattern were found in the right abductor pollicis brevis muscle. Forearm muscles were not investigated. A carpal tunnel syndrome was diagnosed and treated by surgery. Ten months later the patient was admitted to our department because of persistence of his symptoms. Nerve conduction studies showed no response over the thenar muscle on median nerve stimulation at the wrist and elbow. Fibrillations and sparse motor unit potentials with increased duration and amplitude (up to 10 mV) were found in flexor digitorum superficialis muscle, in contrast to the normal Electromyography-findings of the right flexor carpi ulnaris. On median nerve stimulation at the elbow a low compound muscle action potential with an increased distal latency was recorded over flexor digitorum superficialis muscle in the right side (right: amplitude 1 mV, latency 9.0 ms, left: amplitude 7 mV, latency 3.2 ms). No sensory nerve action potential was recorded at the wrist on stimulation of the 1st, 2nd or 3 rd finger. Somatosensory evoked potentials of the left median and right radial and ulnar nerve were normal. No potential was recorded at Erb's point in the supraclavicular fossa, at the sixth and second cervical vertebra and the contralateral cortex on median nerve stimulation on the right. Proximal median nerve lesion was suggested. Palpation along the median nerve revealed a fusiform mass at the distal third of the right upper arm, which could be confirmed by MRI (see figure 1 and 2 ). Figure 1 Coronal T2-weighted MRI reveals a slightly hyperintense fusiforme tumorous lesion of the median nerve approximately 5 cm above the right elbow (arrows). Figure 2 Axial fat-suppressed T2-weighted MRI shows a tumorous lesion of the median nerve with a fascicular pattern (arrow). Surgical resection of a 7 cm long segment of the median nerve with the tumorous lesion and replacement with a sural nerve graft was undertaken. Resection was necessary because of severe involvement of fascicles without any possibility to separate the tumour from the median nerve by micro-surgery. Histological investigation revealed marked peri- and endoneurial fibrosis, severe axonal loss as well as proliferation of concentric whorl-like formations resulting in multicompartment arrangement. These pseudo onion-bulb formations showed strong immunoreactivity for the epithelial membrane antigen and were predominantly negative for S-100 protein, suggesting a proliferation of perineurial cells rather than Schwann cells. In addition, small groups of regenerating axonal sprouts surrounded by perineurial ensheathment were visible, indicating some neuroma-like component (see figure 3 and 4 ). Figure 3 Multicompartment arrangement of concentric whirl-like formations showing strong immunoreactivity with the monoclonal antibody against the epithelial membrane antigen (arrows). Figure 4 Immunohistochemistry using the antibody against protein S-100: The pseudo-onion bulb formations are S-100-negative. In addition, small groups of regenerating axonal sprouts (arrows) surrounded by a perineurial ensheathment are visible indicating some neuroma-like component. The tumour was diagnosed as perineurioma and additional neuroma by histopathology. On clinical follow-up four years later a partial recovery of forearm muscle strength could be noted and the patient was free from pain. Conclusions Carpal tunnel syndrome was falsely diagnosed in this case because of the increased distal latency in the median nerve, although during the first investigation an atrophy of the forearm muscle had already been noticed. Measurement of the median nerve latency to the atrophic flexor digitorum superficialis muscle would have disclosed the site of the lesion at that time. Morphological and immunohistochemical studies classified the lesion as a mixed tumorous lesion with perineurioma and components of neuroma. Neuromas are benign non-neoplastic lesions of the peripheral nerves which develop after disconnection of a nerve or a single fascicle. The history of severe pain with radiation to forearm and wrist immediately after procaine injection indicates that mechanical trauma by the needle and/or toxic effects of the local anesthetic may have caused nerve damage [ 14 ]. In perineuriomas, the underlying etiology still remains unclear. It has been recently suggested that perineuriomas are clonal neoplasms which may be associated with abnormalities on chromosome 22 [ 8 ]. In our case, however, hyperplastic reaction to the preceding nerve damage by trauma or toxic effect may have contributed to the pathogenesis of perineurioma and neuroma-like components, similar to mechanisms supposed to cause localized hypertrophic neuropathy [ 10 , 13 ]. Strong immunoreactivity for epithelial membrane antigen and virtually negative staining for S-100 unequivocally characterized the patient's tumor as a perineurioma and excluded localized hypertrophic neuropathy. Perineurioma and localized hypertrophic neuropathy are characterized clinically by slowly progressive motor mononeuropathy without significant pain or numbness [ 2 , 10 , 15 ]. Neuromas, however, are painful. Pain and numbness in our patient were possibly caused by the neuroma-like component of the tumor. Authors' contributions CS carried out the first neurological examination, study of literature and participated in writing and design of the manuscript. JEA made enquiries in Switzerland in order to get information about first injection and participated in writing of the manuscript. ENJ carried out the immunohistological investigations and prepared the figures 3 and 4 . GS carried out the radiological investigations and prepared figures 1 and 2 . LS performed the clinical follow-up, reviewed and corrected the manuscript. GA performed the nerve conduction studies and electromyography study and signed responsible for the description of this investigation. Competing interests The author(s) declared that they have no competing interests. Pre-publication history The pre-publication history for this paper can be accessed here:	/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC535354.xml
553984	Statistical model of natural stimuli predicts edge-like pooling of spatial frequency channels in V2	Background It has been shown that the classical receptive fields of simple and complex cells in the primary visual cortex emerge from the statistical properties of natural images by forcing the cell responses to be maximally sparse or independent. We investigate how to learn features beyond the primary visual cortex from the statistical properties of modelled complex-cell outputs. In previous work, we showed that a new model, non-negative sparse coding, led to the emergence of features which code for contours of a given spatial frequency band. Results We applied ordinary independent component analysis to modelled outputs of complex cells that span different frequency bands. The analysis led to the emergence of features which pool spatially coherent across-frequency activity in the modelled primary visual cortex. Thus, the statistically optimal way of processing complex-cell outputs abandons separate frequency channels, while preserving and even enhancing orientation tuning and spatial localization. As a technical aside, we found that the non-negativity constraint is not necessary: ordinary independent component analysis produces essentially the same results as our previous work. Conclusion We propose that the pooling that emerges allows the features to code for realistic low-level image features related to step edges. Further, the results prove the viability of statistical modelling of natural images as a framework that produces quantitative predictions of visual processing.	Background A number of models approach the computational modelling of primary visual cortex by using two processing stages. First, there is a linear filtering with filters that are bandpass, oriented, and spatially localized. In some models, the outputs of the linear filters are half-wave rectified, but this difference is inessential because a rectification is done in the second stage anyway. The second stage then consists of pooling together rectified outputs of the first stage, so that cells that have the same orientation and frequency, as well as similar spatial locations, are pooled together. This pooling is then essentially a summation of rectified outputs of filters of different phases. These two processing steps are assumed to roughly correspond to simple and complex cells in V1, respectively. While there is controversy of the validity of such models, see e.g. [ 1 - 3 ], this is probably the simplest and most succesful approach. Recent research has seen a number of models that attempt to explain these processing stages based on statistical modelling of natural images (ecologically valid input). First, application of independent component analysis (ICA) [ 4 ] or sparse coding [ 5 ] shows that the statistically optimal linear features of natural images are very similar to those computed in simple cells in V1 [ 6 - 12 ]. Second, application of a variant of ICA in which some pooling is done in a second stage leads to processing that is similar to what is done in complex cells [ 13 ]. Thus, models based on natural image statistics have been able to succesfully reproduce the above-mentioned two stages, and many well-known observations on V1. It would be most useful if we could use this modelling endeavour in a predictive manner, so that we would be able to predict properties of cells in the visual cortex, in cases where the properties have not yet been demonstrated experimentally. This would give testable, quantitative hypotheses that might lead to great advances, especially in the research in extrastriate areas such as V2, whose function is not well understood at this point. Here, we attempt to accomplish such predictive modelling in order to predict properties of a third processing step, following the two described above. Previously, we have applied a modification of the ICA / sparse coding model on the outputs of modelled complex cells whose input consisted of natural images [ 14 ]. The modification consisted of assuming that the coefficients in the generative decomposition, as well as the values of the higher-order features, were all non-negative. We extend our previous results in two ways. The complex cells in our previous work were all constrained to have the same frequency, which was done in order to reduce the computational load. Here, we first report a technical advance: it is not necessary to make the assumptions of nonnegativity as in [ 14 ]. Thus, we are able to use the conventional, computationally optimized ICA algorithms, in particular the FastICA algorithm [ 15 ]. We are then easily able to incorporate complex cells of different frequencies in the input without exceeding available computational resources. This enables us to study whether some kind of interaction between different frequencies emerges in the statistically optimal higher-order representation. Results Experiment 1: Using ordinary ICA with no constraints As described in Methods, we input a large number of natural image patches into model complex cells that computed the sum of squares of outputs of two simple cells, one odd-symmetric and the other even-symmetric. Then, we performed independent component analysis of the complex cell outputs using the FastICA algorithm. In the first experiment, we used only the output from complex cells in a single frequency band, f 2 in Figure 1 . The purpose was to show that the results in [ 14 ] can be replicated using ordinary ICA methods. The higher-order features are represented by their basis vectors a i which show the contribution of the third-stage feature of index i on the activities of complex cells. A collection of the obtained basis vectors is shown in Figure 2 for the nonlinearity g 1 (see Table 1 ), visualized in the same way as in [ 14 ], see Methods. We can see the same kind of emergence of collinear features as in [ 14 ]. That is, the higher-order features code for the simultaneous activation of complex cells that together form something similar to a straight line segment. Those coefficients that are clearly different from zero have almost always the same sign in a single basis vector. Defining the sign as explained in Methods, this means that the coefficients are essentially non-negative. We thus see that the constraint of non-negativity of the basis vectors imposed in [ 14 ] has little impact on the results: even without this constraint, the system learns basis vectors which are mainly non-negative. Other FastICA nonlinearities led to similar basis vectors. However, some led to a larger number of longer contours. Figure 3 shows the distribution of lengths for different nonlinearities. The nonlinearity g 4 (robust skewness) seems to lead to the largest number of long contours. Experiment 2: Emergence of pooling over frequencies In the second experiment, the complex-cell set was expanded to include cells of three different preferred frequencies. In total, there were now 432 complex cells. We performed ICA on the complex-cell outputs when their input consisted of natural images. Thus, we obtained 432 higher-order basis vectors (features) a i with corresponding activities s i . We visualized a random selection of higher-order features learned from natural images in Figure 4 . The visualization shows that the features tend to be spatially localized and oriented, and show collinearity as in Experiment 1. What is remarkable in these results is that many cells pool responses over different frequencies. The pooling is coherent in the sense that the complex cells that are pooled together have similar locations and orientations. A smaller number of cells is shown in more detail in Figure 5 , where the coefficients in all orientation bands are shown separately. We computed the frequency pooling measure P i in Equation (4) of Methods for the learned basis vectors. The distribution of this measure for natural image input and white Gaussian noise input is shown in Figure 6 . The figure shows that frequency pooling according to this measure was essentially nonexistent for white Gaussian noise input, but relatively strong for many basis vectors when the input consisted of natural images. To express this more quantitatively, we computed the 99% quantile for the white Gaussian noise input. Then, 59% of the basis vectors for natural image input had a pooling index P i that was larger than this quantile. (For the 95% quantile the proportion was 63%.) Thus, we can say that more than half of the higher-order basis vectors, when learned from natural images, have a pooling over frequencies that is significantly above chance level. To show that the pooling measure is valid, and to further visualize the frequency pooling in the higher-order features, we chose randomly basis vectors learned from natural images that have pooling significantly over chance level ( P i above its 99% quantile for white Gaussian noise). These are plotted in Figure 7 . Visual inspection shows that in this subset, all basis vectors exhibit pooling over frequencies that respects the orientation tuning and collinearity properties. The corresponding results when the input is white Gaussian noise are shown in Figure 8 , for a smaller number of higher-order cells. (To make the comparison fair, these were randomly chosen among the 59% that had higher pooling measures, the same percentage as in Figure 7 .) Pooling over frequencies as well as collinearity are minimal. Some weak reflections of these properties can be seen, presumably due to the small overlap of the filters in space and frequency, which leads to weak statistical correlations between complex cells that are spatially close to each other or in neighbouring frequency bands. We also examined quantitatively whether the higher-order features are tuned to orientation. We investigated which complex cell has the maximum weight in a i for each i in each frequency band. When the input consisted of natural images, in 86% of the cells the maximally weighted complex cells were found to be located at the hot-spot ( x i , y i )* (i.e., point of maximum activity, see Methods for exact definition) and tuned to the preferred orientation of the higher-order feature for every frequency f . This shows how the higher-order features are largely selective to a single orientation. When the input consisted of Gaussian white noise, only 34% of the cells were found to be orientation-selective according to this criterion. Finally, we synthesized images from higher-order feature activities to further visualize the higher-order features (see Methods). Figure 9 shows a slice orthogonal to the preferred orientation of one higher-order basis vector (H209 in Figure 5 ). The intensity of the synthesized image shows no side-lobes (unnecessary oscillations), while representing a sharp, localized edge. In contrast, synthesis in the white Gaussian noise case (also shown in Figure 9 ) gives curves that have either side-lobes like the underlying Gabor filters, or do not give a sharp localized edge. Thus, the curve obtained from synthesis of the features learned from natural images corresponds better to the notion of an edge. We propose that the utility of pooling over frequencies is due to the broadband nature of real-world edges. Typical edges in natural images are probably not very similar to typical band-pass Gabor functions (or V1 receptive fields) which have oscillations. A proper representation of such broad-band edges would seem to require pooling over different frequencies. Discussion Frequency channels and edges What is the functional meaning of the pooling we have found? We propose that this spatially coherent pooling of multiple frequencies leads to representation of an edge that is more realistic than the band-pass edges given by typical Gabor filters [ 16 ]. Presumably, this is largely due to the fact that natural images contain many sharp, step-like edges that are not contained in a single frequency band. Thus, representation of such edges is difficult unless information from different frequency bands is combined. In terms of frequency channels, the model predicts that frequency channels should be pooled together after complex cell processing. Models based on frequency channels and related concepts have been most prominent in image coding literature in recent years, both in biological and computer vision circles. The utility of frequency channels in the initial processing stages is widely acknowledged, and it is not put into question by our results – in fact, the statistical modelling framework does show that using band-pass simple and complex cells is statistically optimal [ 6 , 13 ]. However, the question of when the frequency channels should be pooled or otherwise combined has received little attention [ 17 , 18 ]. Our results point out that a statistically optimal way is to pool them together right after the complex cell "stage", and this pooling should be done among cells of a given orientation which form a local, collinear configuration. Related work Several investigators have looked at the connection between natural image statistics, Gestalt grouping rules, and local interactions in the visual cortex [ 14 , 19 - 21 ]. However, few has considered the statistical relations between features of different frequencies so far. It should be noted that some related work on interactions of different frequencies does exist in the models of contrast gain control [ 22 ]. Compared to our own previous work [ 14 ], the main difference seems to be in the frequency tuning of the model complex cells. In [ 14 ], the complex cells were all constrained to have the same spatial frequency tuning – just as in Experiment 1 of the present paper. Therefore, it was impossible to obtain results related to frequency pooling. It seems that any differences in the results are not due to differences in the statistical analysis of the complex-cell outputs or the natural image data set used, because in Experiment 1 of the present paper, we essentially replicated the results of [ 14 ]. The statistical model for analyzing the outputs of complex cells was somewhat different in our earlier work: the components s i and the coefficients a ki were constrained to be non-negative, following proposals by [ 23 , 24 ]. However, this constraint seems to be immaterial, because even without imposing the constraint, the coefficients turned out to be essentially non-negative (after defining the global sign as described in Methods). Recent measurements from cat area 18 (somewhat analogous to V2) emphasize responses to "second-order" or "non-Fourier" stimuli, typically sine-wave gratings whose amplitudes are modulated [ 17 , 25 ]. These results and the proposed models are related to our results and predictions, yet fundamentally different. In the model in [ 25 ], a higher-order cell pools outputs of complex cells in the same frequency band to find contours that are defined by texture-like cues instead of luminance. The same cell also receives direct input from simple cells of a different frequency, which enables the cell to combine luminance and second-order cues. This is in stark contrast to higher-order cells in our model, which pool outputs of complex cells of different frequencies. They can hardly find contours defined by second-order cues; instead they seem to be good for coding broad-band contours. Furthermore, in [ 17 , 25 ], any collinearity of pooling seems to be absent. This naturally leads to the question: Why are our predictions so different from these results from area 18? We suspect this is because it is customary to think of visual processing in terms of division into frequency channels – "second-order" stimuli are just an extension of this conceptualization. Therefore, not much attempt has been made to find cells that break the division into frequency channels according to our prediction. On the other hand, one can presume that the cells found in area 18 in [ 17 , 25 ] are different from our predictions because they use a different coding strategy from the one used in our model, perhaps related to the temporal aspects of natural image sequences [ 26 , 27 ]. Another closely related line of work is by Zetzsche and coworkers [ 28 , 29 ] who emphasize the importance of decomposing the image information to local phase and amplitude information. The local amplitude is basically given by complex-cell outputs, whereas the physiological coding of the local phases is not known. An important question for future work is how to incorporate phase information in the higher-order units. Some models by Zetzsche et al actually predict some kind of pooling over frequencies, but rather directly after the simple cell stage (see Fig. 16 in [ 29 ]). Towards predictive modelling The present results are an instance of predictive modelling, where we attempt to predict properties of cells and cell assemblies that have not yet been observed in experiments. To be precise, the prediction is that in V2 (or some related area) there should be cells whose optimal stimulus is a broad-band edge that has no sidelobes while being relatively sharp, i.e. the optimal stimulus is closer to a step-edge than the band-pass edges that tend to be optimal for V1 simple and complex cells. The optimal stimulus should also be more elongated [ 30 , 31 ] than what is usually observed in V1, while being highly selective for orientation. Statistical models of natural images offer a framework that lends itself to predictive modelling of the visual cortex. First, they offer a framework where we often see emergence of new kinds of feature detectors – sometimes very different from what was expected when the model was formulated. Second, the framework is highly constrained and data-driven. The rigorous theory of statistical estimation makes it rather difficult to insert the theorist's subjective expectations in the model, and therefore the results are strongly determined by the data. Third, the framework is very constructive. From just a couple of simple theoretical specifications, e.g. non-Gaussianity, natural images lead to the emergence of complex phenomena. We hope that the present work as well as future results in the same direction will serve as a basis for a new kind of synergy between theoretical and experimental neuroscience. Conclusion We have shown that pooling over complex cells of different frequency preferences emerges when we model the statistical properties of natural images. This is accomplished by applying ordinary ICA on a set of modelled complex cells with multiple frequencies, and inputting natural images to the complex cells. The resulting independent components, as represented by the corresponding basis vectors, code for simultaneous activation of complex cells that have similar orientations, form a collinear configuration, and span multiple frequencies. Thus, statistical modelling of natural stimuli leads to an interesting hypothesis on the existence of a new kind of cells in the visual cortex. Methods Data and statistical analysis The natural images were 1008 gray-scale images of size 1024 × 1536 pixels from van Hateren's database, available at (category "deblurred") [ 8 ]. We manually chose natural images in the narrower sense, i.e. only wildlife scenes. From the source images, 50,000 image patches of size 24 × 24 pixels were randomly extracted. The mean grey value of each image patch was subtracted and the pixel values were rescaled to unit variance. The resulting image patch will be denoted by I ( x , y ). The complex-cell model was similar to our previous work [ 14 ]. The filter bank consisted of a number of complex cells arranged on a 6 × 6 grid. Complex-cell responses x k to natural images were modelled with a classical energy model: where and are even- and odd-symmetric Gabor receptive fields whose energies are pooled together in the complex cell. The complex cells had 6 × 6 = 36 different spatial locations, and at each location, four different preferred orientations and three different frequency bands. The aspect ratio was fixed to 1.5 and frequency bandwidth to 1.5 octaves, which implied an orientation bandwidth of 37°, according to the definitions in [ 8 ]. The frequency tiling of the Gabor filters is shown in Figure 1 , in which all the filters W were normalized to unit norm for visualization purposes. The actual normalization we used in the experiments consisted of standardizing the variances of the complex cell outputs so that they were equal to unity for natural image input. The number of complex cells totalled K = 36 × 4 × 3 = 432. Note, however, that in Experiment 1 we only used a single frequency band. Independent component analysis (ICA) was performed on the vector x = ( x 1 ,..., x K ) using the FastICA algorithm [ 15 ]. The orthogonalization approach was symmetric. Different nonlinearities g were used, see Table 1 . Thus we learned (estimated) a linear decomposition of the form or in vector form where the vector a i = ( a 1 i ,..., a ki ) gives a higher-order basis vector. The s i define the values of the higher-order features in the third cortical processing stage. Note that the signs of the basis vectors are not defined by the ICA model [ 4 ], i.e. the model does not distinguish between a i and - a i because any change in sign of the basis vector can be cancelled by changing the sign of s i accordingly. Here, we defined the sign for each vector a i so that the sign of the element with the maximal absolute value was positive. To obtain a baseline with which to compare our results, and to show which part of the results is due to the statistical properties of natural images instead of some intrinsic properties of our filterbank and analysis methods, we did exactly the same kind of analysis for 24 × 24 image patches that consisted of white Gaussian noise, i.e. the gray-scale value in each pixel was randomly and independently drawn from a Gaussian distribution of zero mean and unit variance. The white Gaussian noise input provides a "chance level" for any quantities computed from the ICA results. Analysis of the ICA results We visualized the resulting higher-order basis vectors a i following [ 14 ] by plotting an ellipse at each centrepoint of complex cells. The orientation of the ellipse is the orientation of the complex cell k , and the brightness of the ellipse is proportional to the a ki coefficient of the basis vector a i , using a gray-scale coding of coefficient values. In Experiment 1, i.e. the case with a single frequency band, we used this method directly to visualize each higher-order basis vector in a single display. In Experiment 2, i.e. the multifrequency case, we visualized each frequency band separately. In Experiment 2, we are interested in the frequency pooling of complex cells in different higher-order features. We quantified the pooling over frequencies using a simple measure defined as follows. Let us denote by a i ( x , y , θ , f n ) the coefficient in the higher-order basis vector a i that corresponds to the complex cell with spatial location ( x , y ), orientation θ and preferred frequency f n . We computed a quantity which is similar to the sums of correlations of the coefficients over the three frequency bands, but normalized in a slightly different way. This measure P i was defined as follows: where the normalization constant C m is defined as and likewise for C n . For further analysis of the estimated basis vectors, we defined the preferred orientation of a higher-order feature. First, let us define for a higher-order feature of index i the hot-spot ( x i , y i )* as the centre location ( x , y ) of complex cells where the higher-order component s i generates the maximum amount of activity. That is, we sum the elements of a i that correspond to a single spatial location, and choose the largest sum. This allows us to define the tuning to a given orientation of a higher-order feature i by summing over the elements of a i that correspond to the spatial hotspot and a given orientation; the preferred orientation is the orientation for which this sum is maximized. We also computed the length of a higher-order feature as described in [ 14 ]. It is also possible to perform an image synthesis from a higher-order basis vector. However, the mapping from image to complex-cell outputs is not one-to-one. This means that the generation of the image is not uniquely defined given the activities of higher-order features alone. A unique definition can be achieved by constraining the phases of the complex cells. We assume that only odd-symmetric Gabor filters are active. Furthermore, we make the simplifying assumptions that the receptive fields W in simple cells are equal to the corresponding basis vectors, and that all the elements in the higher-order basis vector are non-negative (or small enough to be ignored). Then, the synthesized image for higher-order basis vector a i is given by where the square root cancels the squaring operation in the computation of complex-cell responses, and H denotes the set of indices that correspond to complex cells of the preferred orientation at the hotspot. Negative values of a ki were set to zero in this synthesis formula. Authors' contributions A.H. conceived the basic idea and the principles of the experimental set-up, and wrote the paper. M.G. performed the experiments and elaborated the experimental set-up. P.O.H. assisted in the experiments and the writing.	/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC553984.xml
544591	Mitral valve prolapse associated with celiac artery stenosis: a new ultrasonographic syndrome?	Background Celiac artery stenosis (CAS) may be caused by atherosclerotic degeneration or compression exerted by the arched ligament of the diaphragm. Mitral valve prolapse (MVP) is the most common valvular disorder. There are no reports on an association between CAS and MVP. Methods 1560 (41%) out of 3780 consecutive patients undergoing echocardiographic assessment of MVP, had Doppler sonography of the celiac tract to detect CAS. Results CAS was found in 57 (3.7%) subjects (23 males and 34 females) none of whom complained of symptoms related to visceral ischemia. MVP was observed in 47 (82.4%) subjects with and 118 (7.9%) without CAS (p < 0.001). The agreement between MVP and CAS was 39% (95% CI 32–49%). PSV (Peak Systolic Velocity) was the only predictor of CAS in MPV patients (OR 0.24, 95% CI 0.08–0.69) as selected in a multivariate logistic model. Conclusion CAS and MVP seem to be significantly associated in patients undergoing consecutive ultrasonographic screening.	Background Celiac artery stenosis (CAS) may be caused by atherosclerotic degeneration, as observed in different vascular districts, or by extrinsic compression (ECCA) usually exerted by an abnormally developed arched ligament of the diaphragm [ 1 - 7 ]. ECCA may cause abdominal or epigastric pain triggered by meals (angina abdominis), a syndrome called Celiac Artery Compression Syndrome (CACS) [ 1 , 2 ]. Resolution of symptoms after surgical resection of the arched ligament has been frequently reported. However, some aspects, such as the vascular etiology of symptoms, are still controversial, as the improvements observed after surgery may be attributed to the concurrent destruction of the celiac ganglion rather than to the severance of the arched ligament. Mitral Valve Prolapse (MVP) is the most common valvular disorder, with a prevalence of about 6–7% in the general population [ 8 , 9 ]. It consists of systolic displacement of one or both valvular leaflets into the left atrium, with or without mitral regurgitation. Patients with MVP may also be affected by other cardiac anomalies, such as prolapse of the aortic valve and/or of the tricuspid valve, ostium secundum interatrial sect defect, atrio-ventricular left accesses, and/or extra-cardiac anomalies, such as pectus excavatum and loss of the physiological curvature of the thoracic spine, with greater frequency with respect to the general population. An increased incidence of MVP has been demonstrated in connective tissue disorders, especially in Marfan's syndrome [ 10 ]. Particular genic anomalies have been associated with MVP [ 11 - 13 ]. The association between MVP and CAS has not been extensively investigated. Accordingly, this study was aimed at verifying this hypothesis based on a possible common origin of the two conditions, i.e. the abnormal development of connective tissue causing both ECCA exerted by the arched ligament of the diaphragm and exuberance of valvular mitral tissue. Methods Study population The study population consisted of 1560 out of 3780 (41%) consecutive patients undergoing echocardiographic assessment of MVP, who also had celiac artery Doppler ultrasound performed between November 1999 to September 2004. Echocardiographic examination MVP was defined as clear-cut billowing of one or both mitral leaflets across the mitral annular plane in 2-dimensional parasternal long axis recording or >2 mm late systolic posterior displacement of mitral leaflets by M-mode. A >4 mm displacement defined a moderate-severe prolapse. Ultrasonography (HP Sonos 1000, Tohsiba Corevision, Esaote Caris, Esaote Caris plus, Kontron Iris) was performed without intestinal preparation to limit meteorism. The digestive phase was taken into account, because it may influence visceral arterial flow. The evaluation of arterial flow in the celiac tract was recorded by continuous Doppler (CW) and, whenever possible, with the aid of the color-Doppler signal. No corrections of flow velocity by evaluation of the cosine of the insonorisation axis and the axis of the vessel were used: in this way, flow velocity can never be overestimated. The Doppler signal was collected in apnea during inspiration. During expiration, an increase in the compression of the celiac tripod by the arched ligament occurs in the cases of CAS caused by ECCA, with concurrent increase in flow velocity with respect to the velocity during inspiration [ 14 ]. An example is provided in fig. 1 , 2 , 3 . Figures 4 5 6 7 8 9 are examples of patients with CAS and MVP. Figure 1 Continuous Doppler of Celiac Artery Stenosis, deep expiration. Figure 2 Continuous Doppler of Celiac Artery Stenosis, modeste inspiration. Figure 3 Continuous Doppler of Celiac Artery Stenosis, deep inspiration. CAS was defined as severe in case of Peak Systolic Velocity (PSV) flow velocity in the upper celiac tract greater than 2.0 metres/second [ 15 , 16 ] and End Diastolic Velocity (EDV) greater than 0.5 metres/second. PSV between 2.0 and 2.5 metres/second defined moderate CAS, whilst those > 2.6 metres/second identified severe CAS. Statistical analysis Continuous variables are expressed as means ± 1 standard deviation (SD). Differences between groups were compared using Student's t-test and chi-square test, as appropriate. Association between CAS and MPV was performed using the Kappa statistics, estimated with 95% confidence interval. Univariate odds ratio (OR) along with their corresponding 95% confidence intervals were computed for describing association of clinical variables with MPV in CAS patients, using a logistic regression model. Selection of variables significantly associated with MPV in CAS patients was made using a multivariate logistic model. Selection criterion was the Akaike Information Criterion, applied backward to the multivariate logistic model. The statistical significance was settled at a p value < 0.05. The R (release 1.9) statistical package and the Harrell's Design and Hmisc libraries were used for analysis. Results The incidence of MVP in patients with (10.6%) and without (9.3%) assessment of CAS did not differ significantly. Among those undergoing both examinations, CAS was found in 57/1560 (3.6%) patients. The clinical characteristics of patients with and without CAS are reported in Table 1 . The incidence of CAS was significantly higher among subjects with as compared to those without MVP (28% vs. 0.7%, OR 5.11 95% CI 3.43 to 7.61; p < 0.001). Likely, the incidence of MVP was significantly higher among subjects with as compared to those without CAS (82.4% vs 7.9 %, OR 55.11 95% CI 27.17 to 111.97; p < 0.001). Thus, the overall concordance between CAS and MVP was 39% (95% CI 0.28 to 0.49). MPV as indicator of CAS has a high sensitivity (0.82 95% CI 0.71 to 0.90) and specificity (0.92 95% CI 0.91 to 0.92). Table 1 Clinical characteristics of patients with and without CAS. CAS (n= 57) No CAS (n= 1503) P-value Age 40 ± 21 46 ± 14 0.002 Male 23 (40%) 691 (46%) 0.400 MVP 47 (82.4%) 118 (7.9%) <0.001 No patient with CAS was complaining of symptoms related to visceral ischemia, i.e. abdominal or gastric pain in concurrence with food ingestion. Thirteen (22.8 %) of the 57 subjects with CAS were suffering from cardiovascular disorder. In particular, a macrovascular atherosclerotic disorder (myocardial infarction, carotid plaques, abdominal aorta aneurism, peripheral arterial disease involving the lower limbs) was ascertained in 7 subjects (6 male and 1 female, mean age 71+14 years), whilst no sign of macrovascular atherosclerosis was found in the remaining 50 subjects (16 males and 34 females, mean age 35+12 years). In patients with CAS, factors associated with MPV are age, sex, PSV and BMI. At multivariate analysis only PSV resulted as an independent factor associated with MPV (OR 0.24 95% CI 0.08 to 0.69). Discussion The incidence of CAS is not clearly established, the majority of the literature consisting of isolated reports [ 17 - 21 ]. Surprisingly, one study [ 22 ] evaluating the incidence of CAS in 400 asymptomatic subjects undergoing angiographic examination because of hepatic neoplasm, reported a 7.3% incidence of hemodinamically significant CAS. Its origin ECCA in 55% and atherosclerosis in 10%, whilst it remained not determined in 35% of cases. The incidence of CAS and MVP in the present study dealing with an unselected population was 3.7% and 10.6%, respectively. In addition, this is the first report, to our knowledge, to demonstrate a strong association between the two conditions. CAS may be caused by ECCA or atherosclerosis. Unfortunately, the ultrasound technique has not enough resolution to allow an etiological discrimination that may be difficult to get even with angiographic examination [ 22 ]. However, clinical information may be of help to hypotesize the etiology of CAS in our population. Indeed, CAS was likely due to atherosclerotic disease in the 7 subjects with macrovascular atherosclerotic disorders, whilst it could be secondary to ECCA in the 50 subjects with no feature of atherosclerosis. Of interest, MVP was found in no subject of the first group, whilst it was present in 47/50 (94%) of the second group. Characteristics of CAS subjects according to the presence of atherosclerotic markers are reported In Table 2 . On the basis of these findings, it is logical to hypothesize that ECCA represents the factor explaining the association between CAS and MVP. The two pathological conditions may share a common genetic disorder causing a similar defect of the connective tissue at the two anatomic sites. Nevertheless, no correlation was found between CAS severity and degree of MVP. Table 2 Characteristics of CAS patients according to the presence of atherosclerotic disorders. Univariate OR (95% CI) are computed for the association of each single variable with MPV. N No MPV (N = 10) MPV (N = 47) Combined (N = 57) OR (95% CI) Sex 57 70% (7) 34% (16) 40% (23) 0.22 (0.005,0.97) Age 57 41.7/66.0/72.7 20.0/33.0/47.5 23.0/36.0/55.0 0.15 (0.04, 0.57) PSV 53 2.6/2.8/3.5 2.1/2.4/2.8 2.2/2.5/2.9 0.24 (0.08, 0.69) EDV 53 0.9/1.0/1.2 0.7/0.85/1.0 0.7/0.9/1.1 0.43 (0.18, 1.05) BMI 48 23.8/27.4/30.3 19.0/21.5/23.0 18.9/21.2/23.2 0.10 (0.01, 0.69) Conclusions The results of this study demonstrate that CAS is a relatively frequent finding among patients undergoing Doppler sonography of the celiac artery and is frequently associated to MVP. On the basis of these findings, further investigations are warranted aimed at determining the exact incidence of ECCA associated with MVP, the family distribution of the association between MVP and ECCA, or the prognostic implication of ECCA in subjects with MVP. Additionally, genetic studies could be advisable in subjects presenting with MVP associated to ECCA. List of abbreviations CAS Celiac Artery Stenosis MVP Mitral Valve Prolapse ECCA Extrinsic Compression of Celiac Artery CACS Celiac Artery Compression Syndrome PSV Peak Systolic Velocity EDV End Diastolic Velocity Figure 4 Continuous Doppler of Celiac Artery Stenosis Figure 5 Mitral valve Prolapse, M-mode, patient of fig. 4 Figure 6 Color Doppler of aorta: celiac artery stenosis (longitudinal proiection) Figure 7 Color Doppler of aorta : celiac artery stenosis (longitudinal proiection) Figure 8 Continuous Doppler of Celiac Artery Stenosis. Figure 9 Mitral valve Prolapse, M-mode patient of fig. 8	/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC544591.xml
553990	Colorectal cancer metastasis: in the surgeon's hands?	Background Lymphovascular ligation before tumour manipulation during colorectal cancer resection is termed the ' no-touch isolation ' technique. It aims to reduce the intra-operative dissemination of colorectal cancer cells. Recently, the detection of circulating tumour cells has been enhanced by molecular biology techniques. This paper reviews the evidence for the no-touch isolation technique in light of the recent developments in circulating tumour cell detection. Methods Studies investigating the effect of colorectal cancer surgery on circulating tumour cells were identified by a Medline search using the subject headings colorectal neoplasms and neoplasm circulating cells together with the map term 'no-touch isolation technique'. Further references were obtained from key articles. Results Molecular biological techniques have improved the detection of circulating colorectal cancer cells. There is a trend towards reduced tumour cell dissemination with the no-touch technique compared with the conventional method. However the benefit in terms of improved patient survival remains unproven. Conclusion The no-touch isolation technique reduces circulating tumour cell dissemination but further work is needed to determine the significance of this with regards to patient survival.	Background Of the patients with colorectal cancer (CRC) undergoing surgery for resectable disease, 30–50% will subsequently develop metastases [ 1 ]. Dissemination of tumour cells is therefore thought to occur early on in the disease process. The principle of early lymphovascular ligation before manipulation of the tumour during the surgical resection of a CRC has been termed the ' no-touch isolation ' technique. This was proposed by Barnes [ 2 ] as a way of reducing the incidence of liver metastases by diminishing the intra-operative dissemination of CRC cells. An early proponent of the technique was Turnbull et al [ 3 ], but their findings have not been confirmed in other studies [ 4 ]. Animal studies suggest tumour cells are shed into the circulation during resection of a primary tumour, increasing the likelihood of metastases [ 5 ]. However, the evidence in humans is not clear. This may be because the early techniques used to detect circulating tumour cells (CTC) were not sufficiently sensitive to detect the small number of cancer cells present within the blood. These detection methods relied on cytological examination of blood smears, and allowed the detection of 1 tumour cell within 100 normal cells [ 6 ]. In recent years, the question of tumour cell dissemination during surgical resection has been re-examined following the introduction of molecular biology techniques to detect CTC. These are much more sensitive, and it is now possible to detect one tumour cell in a sample of 10 7 normal cells [ 7 ]. This paper reviews the evidence for the no-touch isolation technique for the resection of a CRC in light of the recent developments in molecular biological techniques for CTC detection. Methods In performing this review, the studies looking at the effect of CRC surgery on CTC were identified by a Medline search using the subject headings colorectal neoplasms and neoplasm circulating cells together with the map term 'no-touch isolation technique' [ 8 - 16 ]. Further references were obtained from key articles [ 17 - 20 ]. We decided to review this topic by addressing three areas of uncertainty. Firstly, does the surgical manipulation of a CRC increase the level of tumour cells within the blood? Secondly, what is the biological significance of CTC in CRC? Thirdly, is the no-touch isolation technique of CRC resection associated with an improved patient survival? To investigate the effect of tumour manipulation on CTC, we have reported the conversion, for each reviewed article, of a negative preoperative blood sample (for CTC) to a positive intra- or post-operative sample within the same patient. We have only noted conversions within the same sample source (systemic venous (SV) or portal venous (PV) blood). Where conversion rates cannot be determined, as data for individual patients are not provided, a comparison between overall preoperative and intra/post-operative positivity for CTC is given. Unless otherwise stated in the reviewed article, we have presumed the conventional method was utilised (tumour manipulation before lymphovascular ligation). Results A) Does the surgical manipulation of a CRC increase the level of tumour cells within the blood? A number of experimental models have been used to investigate the effect of surgical manipulation on CTC. Nishitaki et al [ 5 ] studied 2 groups of rabbits with surgically-inoculated liver tumours. The first group had a laparotomy and manipulation of the hepatic tumour 14 days after inoculation, together with resection of the tumour. The second group had a laparotomy and resection of the tumour but without manipulation. Two weeks later the rabbits were sacrificed, and pathological examination showed significantly more hepatic venous invasion by cancer cells in the tumour manipulation group, together with significantly more lung metastases and shorter mean survival. Romsdahl et al [ 21 ] measured the number of tumour cells by cytological examination in blood samples from the inferior vena cava of rats during manipulation of a thigh tumour. There was a manifold increase in the number of CTC during manipulation, which decreased rapidly, so that 4 minutes after manipulation 93–96% of the cells had disappeared from the circulation. In a similar experiment with mice, Liotta et al [ 22 ] showed a 10-fold increase in tumour cells in the venous effluent of a thigh tumour during manipulation. In order for the no-touch technique to be effective in humans, early ligation of the main pedicle must eliminate passage of liberated tumour cells into the portal and systemic circulations. However, Salsbury et al [ 23 ] found the number of tumour cells in the common iliac vein dramatically increased after ligation of the inferior mesenteric vein, suggesting there was a shift of the draining venous blood into the systemic circulation following ligation. In an experimental model using dogs, Ackerman [ 24 ] showed clamping of the mesenteric veins produced an increased intestinal lymphatic and venous outflow, thereby enhancing the possibility of tumour cell dissemination. He concluded ligation of the major arterial supply, followed by the marginal vessels, is the most effective technique in minimising outflow. On theoretical grounds, the no-touch technique would appear unsuitable for low rectal cancers, as it is necessary to mobilise tumours in the rectum before all the draining veins are controlled. In addition, a large proportion of the venous drainage occurs into the iliac system, thereby negating the effect of mesenteric pedicle ligation. Results of reviewed articles There have been numerous studies investigating the effect of tumour manipulation on CTC in human tumours. This review focuses on studies involving CRC. The overall conversion rates for each study (ie conversion from negative preoperatively for CTC to positive intra or post-operatively taking into account all study patients) are listed in Additional Files 1 and 2 . The conversion rates taking in to account only those patients negative preoperatively for CTC are also listed. The only study to compare directly the conventional and no-touch techniques for the resection of a CRC involved the non-random assignment of 27 patients into conventional and no-touch groups [ 8 ]. The technique of mutant-allele-specific amplification (MASA) was used to identify mutations of K-ras or p53 genes in the primary tumour of each patient. Cells containing the same mutations were then examined for in blood samples taken before, during, and after surgery. In the conventional group, 11 patients had K-ras or p53 mutations in their primary tumour. None of these patients had preoperative blood samples that contained cells exhibiting similar mutations. However, 8 (73%) did have cells containing similar mutations in the blood taken intra-operatively following tumour manipulation. In the no-touch group, 7 patients had mutations in their primary tumour. None of these patients showed similar mutations preoperatively, and only 1 patient (14%) became positive intra-operatively. From these findings, the authors conclude the no-touch technique is effective at reducing the intra-operative dissemination of colorectal tumour cells. The other studies involving CRC have used either the conventional or no-touch surgical techniques, but have not directly compared both. Weitz et al [ 9 ] and Sales et al [ 10 ] employed the no-touch technique and found an overall conversion rate of 16% and 9% respectively. Both studies conclude the no-touch technique is effective at reducing tumour cell dissemination. Bessa et al [ 17 ] also used the no-touch technique in 50 patients who were randomly assigned to undergo open or laparoscopically-assisted colectomy. They found an overall conversion rate of 8% and 12% respectively in SV samples. Studies employing the conventional technique of CRC resection have found overall conversion rates of 0 – 80% in SV samples [ 11 - 15 , 18 , 19 ] (see Additional file 2 ). Funaki et al [ 20 ] demonstrated the highest conversion, with 4 out of a study sample of 5 patients converting intra-operatively. This is much higher than in any other study, and although the sample size is small, the authors suggest this is evidence of tumour cell dissemination as a result of surgical manipulation. It is interesting to note 80% of these patients had rectal cancers, which require significant mobilisation before control of the venous output is obtained. Griffiths et al [ 18 ] did not detail individual patients and so the conversion rate cannot be determined, but they found 4% of patients were positive preoperatively, 50% were positive intra-operatively, and 8% postoperatively. Similarly, Ito et al [ 15 ] found significantly more patients were positive for CTC following surgery compared with preoperatively (51% v 38%, p = 0.003). Most authors conclude their results demonstrate evidence of tumour cell dissemination. However, Garcia-Olmo et al [ 11 ], who used the conventional technique and did not demonstrate a conversion in any patients, concluded the no-touch isolation technique was unnecessary. As a number of patients were positive for CTC before surgery, a more direct assessment of the effect of tumour manipulation might be to analyse the conversion rates only for those patients negative for CTC preoperatively. This reveals a much greater conversion rate for most studies (see Additional Files 1 and 2 ). For example, using the conventional method, Funaki et al [ 20 ] found all 4 patients negative preoperatively converted intra-operatively. Similarly, Tien et al [ 13 ] demonstrated a conversion of 49% for SV, and 52% for PV samples. Conversely, Garcia-Olmo et al [ 11 ] used the conventional method and failed to show conversion in any patients. For the no-touch technique, the conversion rates are also greater if only the patients negative preoperatively are considered. Bessa et al [ 17 ] demonstrated the highest conversion rate, with 50% and 33% of patients converting following laparoscopically-assisted colectomy within SV and PV samples respectively. Limitations and variations of studies under review 1) Choice of technique used to detect CTC Before the advent of molecular biology, the detection of tumour cells within the blood depended on cytological examination of a blood smear. The incidence of circulating tumour cells during CRC resection detected by cytological examination of blood smears has been reported as 25–67% [ 25 , 26 ]. Griffiths et al [ 18 ] state it is easy to detect CTC within a blood sample, but this method of detection underestimates the true number of circulating cells as there is probably a considerable loss in the preparation of blood cell concentrates. Also, difficulties in interpretation arise when other large mononuclear cells mimicking cancer cells are present. Immunocytochemistry has been used to detect CTC in a number of tumour types, including colorectal cancer [ 27 ]. This technique labels tumour cells with antibodies against specific cell components that are not expressed in haematopoietic cells. It has the advantages of preserving cellular morphology and allowing identification of cell clusters, an important drawback of molecular detection techniques [ 28 ]. It also permits simultaneous assessment of other cellular details such as proliferation activity and oncogene expression. However, it is labour intensive due to the low concentration of CTC. The sensitivity has also been questioned [ 29 ], although this can be enhanced with flow cytometry. In recent years, molecular biological techniques have been developed for use in tumour cell detection. The polymerase chain reaction (PCR) involves detection and amplification of certain DNA mutations. A copy of the DNA containing the target sequence is made and the process is repeated many times, producing multiple copies and thereby increasing the sensitivity of detection. Mutations in the k-ras gene have often been used for detection of CTC in CRC. However, the detection of these sequences within the blood does not necessarily imply there are circulating tumour cells, as DNA is relatively stable and may represent fragments of tumour DNA released by cell necrosis or apoptosis and persisting in the blood for some time [ 30 ]. Also, this method of detection relies on the presence of specific DNA mutations within the primary tumour, and the genetic changes associated with CRC are known to be heterogenous [ 28 ]. An alternative technique, now more commonly employed, is reverse transcriptase-PCR (RT-PCR). This detects messenger RNA (mRNA), which is used as an indicator of tissue-specific gene transcription. The technique is rapid, more easily automated than immunocytochemistry, and is capable of detecting one tumour cell in 10 7 normal cells [ 19 ]. RNA is rapidly degraded and so its presence in the blood suggests active expression by circulating tumour cells [ 13 ]. However, there is a lack of standardisation of RT-PCR techniques, resulting in different sensitivities and specificities between centres [ 31 ]. Illegitimate transcription (when tissue-specific genes are transcribed in non-specific cells) can occur, and the presence of non-malignant epithelial cells in the blood (for example, following venepuncture [ 32 ]) reduces the specificity. RT-PCR detects the number of copies of mRNA, not the number of CTC, so it may not accurately assess an increase in tumour cells secondary to mobilisation [ 13 ]; also it may be difficult to derive the prognostic value of CTC if the levels cannot be accurately determined in individual patients. False negatives can occur with RT-PCR, affecting the technique's sensitivity. The marker of interest may not be expressed because of tumour cell heterogeneity or because there is a poorly-differentiated subclone that has lost the ability to express the tissue-specific marker [ 33 ]. PCR inhibitors may be present within some body tissues, and it is important to note the in-vitro sensitivity reported in the reviewed articles will be higher than the true in-vivo sensitivity, as the cell lines used to determine the sensitivity are known to express strongly the marker of interest and are not affected by the presence of in-vivo PCR inhibitors. Most of the reviewed articles that state the specificity of their technique claim a rate of 100%, indicating all the tumour samples were positive for the marker of interest, whereas none of the control blood samples (from healthy volunteers or patients undergoing resection for benign disease) were positive. However, Patel et al [ 14 ] examined 143 control subjects and found 7 (5%) were positive when a single blood sample was assessed for CTC. 2) Choice of marker used to detect CTC There are no tumour-specific genes for CRC, and it is known malignant cells continue to express markers that are characteristic of their tissue of origin. Therefore, a range of epithelial cell markers, which are normally absent from the blood, have been used as RT-PCR targets for cancer cells derived from epithelial tissues. The cytokeratin (CK) polypeptides are found within the cytoplasm of epithelial cells, and several CK markers have been used to detect CTC. The most common of these is CK 20, which is expressed in gastrointestinal epithelia, urothelium, Merkel cells, and tumours derived from these tissues [ 9 ]. Carcinoembryonic antigen (CEA) is a glycoprotein found on the surface of epithelial cells. It is overexpressed in nearly all colorectal cancers [ 34 ], as well as other tumours such as breast and non-small cell lung carcinoma, and has been widely used to detect CTC. Guanylyl cyclase C (GCC) is another tissue-specific marker only expressed in normal intestinal mucosa. It is a member of the guanylyl cyclase family of receptors, and its expression persists once the epithelial cell has undergone neoplastic transformation [ 13 ]. Cell surface sialylated carbohydrates, such as sialyl Le a and sialyl Le x , are associated with CRC formation and progression [ 35 ], and have been used to detect disseminated tumour cells as well as predict prognosis. Concerns regarding marker specificity have been raised. CK 8, 18 and 19 have all been shown to be expressed in normal blood [ 29 ]. CK 20 and CEA are more specific, although they have also been seen in control samples [ 31 , 36 ]. The CEA family includes homologous genes expressed in granulocytes [ 37 ], and CEA may be induced by surgical stress [ 15 ] as well as occurring in the blood of patients with inflammatory conditions such as colitis [ 38 ]. GCC appears to be the most specific marker investigated. Bustin et al [ 31 ] found CK 20 transcripts in all 21 healthy volunteers studied, whereas GCC was found in only a single control. 3) Choice of blood sample used to detect CTC The source of blood for analysis appears to be important. Most studies have investigated systemic venous samples for the presence of CTC. Animal studies suggest tumour cells become trapped in the capillary bed of the first organ encountered [ 39 , 40 ], and so for CRC would be seen in the liver but not in peripheral blood. Theoretically, CRC cells must pass through the liver, lungs and the microcirculation of the other tissues of the body before they pass into the systemic venous circulation. Also, any CTC should be diluted by the larger blood volume of the peripheral circulation [ 12 ]. There should be many more circulating colorectal tumour cells in the portal circulation than the systemic circulation, therefore. Griffiths et al [ 18 ] found 57% of patients were positive for CTC in PV blood during resection a colorectal tumour, as opposed to 50% of patients who were positive in their SV samples. Tien et al [ 13 ] found more PV samples were positive before, during and after surgery compared with SV samples. However, 3 of the 17 who were positive in the peripheral blood were negative in all portal samples, suggesting tumour cells were bypassing the portal circulation and entering the systemic circulation directly. It is interesting to note these 3 patients had Dukes stage C disease and it may be that tumour cells were passing directly from the lymphatics into the peripheral circulation. Bessa et al [ 17 ] found there was concordance between SV and PV samples for CEA mRNA in only 65% of cases. They also found no patients demonstrated a conversion in PV samples following open colectomy, however 8% converted in SV blood. Due to the discrepancy in sample sources, we have only noted patients who converted within the same venous compartment. Several studies detected CTC in draining vein blood intra-operatively, and compared it against a reference SV sample taken before manipulation. It is difficult to be certain of conversion in these patients, however, as it is possible the PV samples were positive preoperatively. It is important for subsequent studies to measure CTC in both systemic venous and draining vein blood, before, during and after tumour manipulation to investigate further the discrepancy in CTC detection between the venous compartments. The importance of multiple sampling should be considered. Glaves et al [ 41 ] suggests cancer cells are intermittently shed in to the blood, so sampling errors may occur if single samples are taken. Tien et al [ 13 ] sampled SV and PV blood twice during tumour mobilisation and found that of the 14 patients who were positive in portal blood during mobilisation, 5 (36%) were only positive in the second sample, and so would have been considered negative if a single sample was taken. In the SV samples, 3 out of 17 (18%) patients were only positive in the second sample. Mori et al [ 19 ] also performed sequential sampling intra-operatively, and found of the 5 patients positive for CTC at any of the 4 sampling time points, 2 (40%) were positive at a single time point only. Wharton et al [ 42 ] showed by increasing the sampling frequency (from once to twice), the detection of CTC is significantly increased. Weitz et al [ 9 ] suggest another factor affecting the detection of CTC may be the degree of intra-operative blood loss and subsequent administration of intravenous fluids, thereby diluting the blood volume and reducing the likelihood of tumour cell detection. They showed the chance of intra-operative detection of tumour cells was halved by a blood loss of 0.5 l, and so excluded all patients with an intra-operative blood loss of more than 1 l from further statistical analysis. They found of the 7 patients who were positive pre- or post- but not intra-operatively, 5 (71%) had a blood loss of more than 1 litre, suggesting a possible dilution effect during the operation. In the articles reviewed, the degree of blood loss was only reported in the study by Weitz et al. B) What is the biological significance of CTC in CRC? The detachment of malignant cells from the primary tumour is an early step in the formation of metastases [ 5 ]. The release of tumour cells is a continuous process [ 43 ], however the metastatic process is inefficient and the presence of tumour cells within the blood does not necessarily imply the subsequent development of metastases [ 44 ]. It is poorly understood which steps in the metastatic process are responsible for the inefficiency of tumour cells to form overt metastases. Early studies suggested fewer than 0.1% of circulating tumour cells survive in the circulation [ 39 ], and only 0.01% form metastases [ 45 ]. It was thought the majority of CTC are removed from the circulation within 24 hours [ 39 ], by elimination in the first capillary bed they encounter [ 46 ]. However, recent work using cytoplasmically-labelled tumour cells (as opposed to nuclear-labelled cells which are more vulnerable to destruction) has found the majority of cells survive in the circulation for several days following injection, and the inefficient part of the metastatic process appears to be the variable growth of the cancer cell at the secondary site [ 47 ]. The metastatic process may be enhanced by the surgical procedure itself. It is known the entrapment of tumour cells in the microcirculation of a target organ is facilitated by the presence of fibrin and platelets [ 48 , 49 ]. The activation of coagulation that occurs during an operation may therefore augment this process [ 50 ]. Also, surgical stress has been shown to induce immune suppression [ 51 ], thereby increasing the metastatic efficiency. The prognostic value of CTC in CRC has yet to be fully determined. Fujita et al [ 52 ] found patients negative preoperatively for CK-19 or CK-20 had significantly fewer recurrences and better 5-year disease free survival. Nakagoe et al [ 53 ] found a high sialyl Le (x) antigen or CEA in blood draining the tumour was an independent prognostic indicator of poor survival, and Yamaguchi et al [ 12 ] found a similar finding in patients positive for both CEA and CK-20. Other studies show conflicting results, however. Bessa et al [ 54 ] found a preoperative peripheral blood sample positive for CEA did not predict prognosis, and in a comprehensive review, Tsavellas et al [ 29 ] conclude, at present, the presence of CTC cannot be considered to be a reliable indicator of prognosis in any common solid malignancy because of the lack of large standardised trials with sufficient follow up. C) Is the no-touch isolation technique of CRC resection associated with an improved patient survival? There are few studies directly comparing the outcomes following conventional and no-touch techniques. Turnbull et al [ 3 ] retrospectively compared 664 patients who underwent CRC resection using the no-touch technique against 232 patients with similar histological stage operated on by different surgeons employing the conventional method. They found the overall 5-year survival rate for the no-touch group was 51%, compared with 35% for the conventional group. They state the difference in mortality was due to an increased incidence of hepatic metastases in the conventional group. Subsequent criticism [ 4 ] of Turnbull's findings suggested patient selection may not have been random, and the basis of the no-touch technique incorporates an extended lymphadenectomy. The study also excluded cancers of the rectum. Wiggers et al [ 4 ] randomly assigned 236 patients with tumours of the colon or rectum to a no-touch or conventional surgery group, and found no significant difference between the groups in terms of recurrence or survival. However, there was a trend towards fewer and later onset of liver recurrences in the no-touch group. Their recommendation was that the no-touch technique should be used for tumours where it is easily applicable. The largest series was reported recently by Slanetz [ 55 ], who retrospectively reviewed 1863 cases of colorectal cancer resection over a period of 24 years. In 1050 cases, tumour mobilisation had occurred before regional mesenteric vessel ligation, whereas 813 cases had vessel ligation performed initially. The extent of mesenteric resection and tumour differentiation was reportedly similar between the two groups. The author reports the sequence of vessel ligation had little impact on the incidence of cancer-related deaths at 5 and 10 years, with no significant difference in survival rates between the early vessel ligation and conventional groups for colonic or rectal tumours. However, the sequence of vessel ligation did have a significant effect on the distribution of metastases. The early vessel ligation group was associated with fewer liver metastases but more systemic metastases compared with the conventional group. These findings may support the theory that after mesenteric pedicle clamping, there is a shift in draining blood into the systemic circulation [ 23 ]. There was also a significant increase in the local recurrence rate with the no-touch compared with the conventional technique (22.6% v 14.6%; p = 0.0001). Another component of the no-touch isolation technique investigated by Slanetz is the control of intraluminal spread of malignant cells by applying bowel clamps or ligatures before tumour manipulation. He reviewed the results of 599 CRC resections in which bowel ligation prior to tumour mobilisation was used, and compared the data against 1416 resections in which bowel ligation was not performed. He found the application of bowel ligatures before tumour mobilisation significantly improved the 5-year cancer-related death rate for colon cancer (20% v 25%, p = 0.02), but not for rectal cancer. When considering colon and rectal cancer combined, bowel ligation prior to mobilisation significantly reduced the local (12% v 19%, p = 0.02) and distant (liver: 10% v 15%, systemic: 13% v 18%, p < 0.001) recurrence rates compared to resection without prior bowel ligation. These findings support the earlier work by Cole et al [ 56 ], who showed that ligatures around the bowel controlled the rate of local lymphatic and intraluminal dissemination of malignant cells. Conclusion It is difficult to draw any firm conclusions from the articles studied due to the lack of standardisation of sample source, CTC detection method and the small sample sizes. However, the only study directly comparing conventional and no-touch surgical techniques found a conversion rate of 73% for conventional surgery and 14% for the no-touch technique, suggesting a benefit of vascular clamping before tumour mobilisation. The overall conversion rates for the other studies employing the no-touch technique were 0–16% (see Additional file 1 ), compared with 0–80% for studies utilising conventional surgery (see Additional file 2 ). Therefore, these data suggest there is a trend towards reduced tumour cell dissemination for the no-touch isolation method. The benefit of this in terms of improved patient survival, however, remains unproven. On purely theoretical grounds, the presence of alternate lymphovascular pathways for most colorectal tumours ensures complete isolation of the tumour-bearing segment is hard to achieve, limiting the technique's efficacy. In addition, the extent of mesenteric resection, and not the surgical technique employed, is probably the most important determinant of patient outcome following CRC surgery [ 55 ]. Due to the lack of consensus regarding the best technique for detection, the biological importance of CTC has not been fully determined. With the introduction of molecular biological techniques, the sensitivity for detection has improved considerably. The hope is that accurate detection of occult circulating malignant cells would help to stage the disease and predict prognosis, as well as monitor the response to therapy and highlight early the possibility of recurrence. However due to the differences in methodology and conclusions of studies investigating this area, the detection of CTC cannot, at present, be used to dictate therapeutic strategy. Despite the methodological heterogeneity of the articles reviewed, we believe the introduction of highly sensitive methods of CTC detection has forced a re-analysis of the role of the no-touch isolation technique in CRC surgery, and the stages of CRC resection using the no-touch technique would appear to offer a sensible, systematic approach to the surgical management of large bowel tumours. However, further work needs to be done to investigate the remaining areas of uncertainty. In particular, future developments in CTC detection must ensure more automation and greater standardisation of techniques between centres. RT-PCR appears to offer the greatest sensitivity, and techniques detecting multiple markers that are more tumour-specific should be investigated [ 57 , 58 ]. The true biological importance of CTC in CRC needs to be assessed further, and finally, the actual benefit of the no-touch technique can only be determined by large scale randomised clinical trials utilising multiple venous sampling in patients matched for age and stage of disease. Competing interests The author(s) declare that they have no competing interests. Authors' contributions GA researched the topic and prepared the manuscript, IM conceived the article and AC critically analysed and formatted the manuscript. Supplementary Material Additional File 1 Conversion rates for studies employing the no-touch isolation technique of CRC resection Click here for file Additional File 2 Conversion rates for studies employing the conventional technique of CRC resection Click here for file	/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC553990.xml
552324	Diagnosis and treatment of carotid body paraganglioma: 21 years of experience at a clinical center of Serbia	Background The carotid body paraganglioma (chemodectoma) is a relatively rare neoplasm of obscure origin. These are usually benign and commonly present as asymptomatic cervical mass. Patients and methods Records of 12 patients (9 female and 3 male) with carotid body tumors treated between 1982 and 2003, treated at our center were retrospectively reviewed. Data on classification, clinical presentation, and surgical treatment were extracted from the case records. Surgical complications and treatment outcome were noted and survival was calculated by actuarial method. The literature on carotid body paraganglioma was reviewed. Results The average age of the patients was 52 years (range 30–78 years). Eight of these cases presented as a large asymptomatic non-tender neck mass, and two each presented with dysphagia, and hoarseness of voice. As per Shamblin classification seven of tumors were type II and 5 were types III. In 7 cases subadventitial tumor excision was performed, while in 5 associated resection of both external and internal carotid arteries was carried out. The artery was repaired by end-to-end anastomosis in one case, with Dacron graft in one case, and with saphenous vein graft in 3 cases. There was no operative mortality. After a mean follow-up of 6.2 years (range 6 months to 20 years), there were no signs of tumor recurrence in any of the cases. Conclusions Surgical excision is the treatment of choice for carotid body paragangliomas although radiation therapy is an option for patients who are not ideal candidates for surgery. For the tumors that are in intimate contact with carotid arteries, the treatment by vascular surgeon is recommended.	Background Paraganglioma arising from the carotid body are relatively rare tumors but constitute majority of head and neck paragangliomas (60–70%) [ 1 - 6 ]. The term paraganglia was first used by Kohn in the early twentieth century and is the most appropriate nomenclature from an embryologic standpoint [ 3 - 5 ]. Other terms such as carotid body tumor, glomus tumor, chemodectomas, and nonchromaffin tumor are less accurate terms and therefore should be best avoided [ 7 - 13 ]. The neoplasm present as asymptomatic neck mass. We report our experience with surgically treated carotid body tumors. Patients and methods Between 1982 and 2003, 12 patients (9 female and 3 male) with carotid body paraganglioma were surgically treated at the Institute for Cardiovascular Diseases, Serbian Clinical Centre. Mean age of the patients was 52 years (range 30–78 years). The records of these patients were retrospectively reviewed for clinical presentation, diagnostic work-up, surgical treatment, and outcome. Descriptive data was presented as frequency and percentage. Survival was calculated by actuarial method. World literature on carotid body paraganglioma was reviewed. The articles were extracted using the key words carotid body and paraganglioma. All the patients were followed-up periodically every 6 months for the first year, yearly for next 5 years, thereafter only select patients were followed. The patients who had undergone carotid artery repair were followed-up with yearly duplex scanning; two patients were followed by a computerized tomography (CT) scan, and one by regular magnetic resonance (MR) imaging. Results Eight cases presented as large non-tender neck masses located just anterior to the sternocleidomastoid muscle, two patients presented with dysphagia due to hypoglossal nerve compression, while two other had hoarseness of voice. The duplex ultrasonography and selective carotid angiography were used for diagnosis in the eleven cases, CT in five, and MR imaging in three cases (Figure 1 and 2 ). In one case the diagnosis was established on intraoperative exploration. Figure 1 Selective carotid angiography showed carotid body paraganglioma. The typical separation (''lyre sign") of external and internal carotid arteries, are presented. Figure 2 Selective carotid angiography showed hypervascularization of the carotid body paraganglioma mostly from the external carotid artery. Intraoperatively on exploration of the neck seven of the cases showed a medium size tumor intimately associated and compressing carotid vessels (Shamblin II), and a large tumor involving carotid vessels in five cases (Shamblin III). In 7 cases of Shamblin II carotid body paraganglioma a subadventitial tumor excision was performed while in other 5 cases both external and internal carotid arteries were resected. One of these was repaired by end-to-end anastomosis, one with interposition of Dacron ® graft, and other 3 were repaired with reversed saphenous vein graft. The histological examination showed no signs of malignancy in any of the tumors. In two cases transient hypoglossal nerve palsy was noticed. Another patient had unexpected postoperative hoarseness of voice due to the transient vagus nerve palsy. All these three cases subsequently recovered. There was no operative mortality. The patients were followed-up from the 6 months to 20 years (mean 6.2 year) no local, regional or distant metastasis was noticed. The actuarial survival was 100%. Discussion The carotid body was first described by von Haller in 1743 [ 14 ]. It is highly specialized organ located at the common carotid artery bifurcation. Its feeding vessels run primarily from the external carotid artery. The function of the carotid body is related to autonomic control of the respiratory and cardiovascular systems, as well as blood temperature [ 3 , 10 , 12 , 15 - 23 ]. Paraganglioma is a relatively rare neoplasm occurring in carotid body [ 1 - 6 ]. The carotid body paraganglioma is more common in women [ 2 - 5 , 20 , 25 - 28 ]. The incidence of bilateral carotid body lesions is approximately 10%. Most of these lesions are benign however malignant behavior is often encountered. For diagnosis of malignant carotid body paraganlioma there are no clear histological characteristics that differentiate it from benign lesions. This diagnosis is reserved for the tumors with local, regional and distant metastasis. The rate of malignancy is reported to be 6–12.5% of all cases [ 3 - 5 , 9 , 11 , 29 - 35 ]. The 7–9% of the cases are hereditary [ 2 , 4 , 20 , 25 - 28 , 36 ]. None of our cases were bilateral or hereditary. Carotid body paraganglioma often present as slow growing, non-tender neck masses located just anterior to the sternocleidomastoid muscle at the level of the hyoid. The tumor is mobile in the lateral plane but its mobility is limited in the cephalocaudal direction [ 3 - 5 , 13 - 15 , 21 , 26 , 36 - 38 ]. Occasionally the tumor mass may transmit the carotid pulse or demonstrate a bruit or thrill [ 39 ]. Because of its location in close approximation to carotid vessels and X-XII cranial nerves, tumors enlargement causes progressive symptoms such as dysphagia (two of our cases), odynophagia, hoarseness of voice (two of our cases) or other cranial nerve deficits [ 2 - 5 , 14 , 26 , 27 , 32 , 37 , 40 ]. The patients may give a history suggestive of symptoms associated with catecholamine production such as fluctuating hypertension, blushing, obstructive sleep apnea and palpitations [ 3 - 5 , 10 , 14 , 15 , 21 - 23 , 37 ]. Size of the tumor has a great importance not only for its clinical manifestations but also for treatment. In 1971, Shamblin introduced a classification system based on tumors size [ 41 ]. They classified small tumors that could be easily dissected away from the vessels as group I. Group II (7 of our cases) included paragangliomas of medium size that were intimately associated and compressed carotid vessels, but could be separated with careful subadventitial dissection. Group III consisted of (5 of our cases) tumors that were large and typically encased the carotid artery requiring partial or complete vessel resection and replacement. Histologically, carotid body paraganglioma resemble the normal architecture of the carotid body. The tumors are highly vascular, and between the many capillaries are clusters of cells called Zellballen [ 41 ]. The carotid angiography is the most useful diagnostic test for paragangliomas. The angiography demonstrates tumor blood supply and widening of the carotid bifurcation by a well-defined tumor blush ("lyre sign"), which is classic pathognomonic angiographic finding [ 5 , 8 , 37 - 39 , 42 , 43 ]. MR and contrast CT are more effective non-invasive imagining modalities comparing with duplex ultrasonography, especially for small tumors [ 3 , 37 - 39 , 42 - 45 ]. Radioimmunodetection of carotid body paraganglioma by 111 In labeled anti-CEA antibody is also described in literature [ 9 , 46 ]. The differential diagnosis includes other tumors in this area, carotid artery aneurysms and elongation. For this reason using of precutaneous fine-needle aspiration for preoperative diagnosis of carotid body paraganglioma, can be very dangerous [ 47 ]. Resection of carotid body paraganglioma carries inherent risks of injury to the cranial nerves, carotid arteries as well excessive blood loss. Reigner first attempted resection of a carotid body paraganglioma in 1880, but the patients did not survive [ 48 ]. Maydel was the first to remove a carotid body paraganglioma successfully in 1886, but the patient became aphasic and hemiplegics due to internal carotid artery ligature [ 49 ]. In 1903, Scudder performed the first successful removal of carotid body paraganglioma [ 50 ]. The surgical excision with careful subadventitial dissection is treatment of choice for most carotid body paragangliomas (Shamblin I and II) [ 2 - 6 , 14 - 18 , 34 , 37 - 40 , 43 ]. The Shamblin III of carotid body paraganglioma requires resection of the external and/or internal carotid artery. If the internal carotid is encased in tumor or damaged during resection, immediate repair/replacement should be performed [ 15 , 37 , 39 , 40 , 42 , 43 , 51 , 52 ]. The second problem during tumor excision is bleeding, which sometimes can be massive. In such cases clamping of all carotid arteries is useful, with placement of internal carotid shunt [ 18 , 35 , 37 ]. Having in mind our experience with surgical treatment of both carotid body gangliomas as well as carotid stenosis, we recommend Pruitt-Inahara double balloon occlusive internal carotid shunt [ 37 ]. The placement of this shunt through incision on the common carotid artery contributes to the adequate bleeding control from the common and internal carotid arteries, as well as brain protection. This procedure gave a clean and dry operative field during tumor removal [ 3 , 37 ]. Some other articles recommend angiographic embolization preoperatively [ 3 , 23 , 37 , 42 , 53 - 55 ]. The Preoperative embolization of a carotid body paraganglioma can be performed by ethanol or polyvinyl alcohol. The finally result is a complete devascularization [ 55 ]. Earlier the carotid body paragangliomas were considered radioresistant [ 34 ]. However, more recent studies indicate good responses to radiation therapy [ 11 , 30 ]. Most authors recommend radiotherapy for giant and recurrent carotid body paragangliomas, and with malignant carotid body paragangliomas metastatic to the regional lymph nodes [ 8 , 33 - 36 ]. The modern surgical techniques have reduced the risk of postoperative stroke in carotid body paraganglioma resection to less than 5% [ 37 , 40 , 56 ]. However, the incidence of cranial nerve injury remains strikingly high, ranging from 20% to 40% [ 37 , 38 , 48 , 56 , 57 ]. In 20% of patients the neurological deficits is permanent. We found two (18%) transient hypoglossal, and one transient vagus nerve damage. recurrence after complete resection occurs in approximately 6% of patients [ 15 , 37 , 39 , 40 , 42 , 43 , 51 , 52 ]. In our study however, there were no recurrences. The patients with internal carotid artery reconstruction should undergo duplex scanning periodically to identify graft stenosis. Conclusion Early operative management is warranted to avoid the possibility of eventual metastasis and progressive local invasion to the point of inoperability. In case of tumors intimately contact with carotid arteries, the treatment by vascular surgeon is recommended. Competing interests The author(s) declare that they have no competing interests. Funding source Nil Authors' contributions LBD: Preperation of draft manuscript VBD: Literature search, data collection DMV: Study design, data analysis, interpretation, preparation of draft RPS: Study coordination, data interpretation, manuscript preperation SND: Manuscript editing, preparation of final manuscript for publication All authors read and approved the manuscript.	/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC552324.xml
524503	Association of changes in health-related quality of life in coronary heart disease with coronary procedures and sociodemographic characteristics	Background Few studies have focused on the association between the sociodemographic characteristics of a patient with the change in health-related quality of life (HRQOL) following invasive coronary procedures, and the results remain inconclusive. The objective of the present study was to measure the temporal changes in HRQOL of patients with coronary heart disease, and assess how these changes are associated with invasive coronary procedures and sociodemographic characteristics. Methods This was a prospective study of 254 patients with angina pectoris and 90 patients with acute coronary syndrome. HRQOL was assessed with the multi-item scales and summary components of the SF-36, both 6 weeks and 2 years after baseline hospitalization in 1998. Paired t -tests and multiple regression analyses were used to assess temporal changes in HRQOL and to identify the associated factors. Results Physical components of HRQOL had improved most during the 2 years following invasive coronary procedures. Our findings indicated that patients with angina pectoris who were younger, male, and more educated were most likely to increase their HRQOL following invasive coronary procedures. When adjusting for baseline HRQOL scores, invasive coronary procedures and sociodemographic characteristics did not explain temporal changes in patients with acute coronary syndrome, possibly due to higher comorbidity. Conclusion Sociodemographic characteristics should be taken into account when comparing and interpreting changes in HRQOL scores in patients with and without invasive coronary procedures.	Background In the medical treatment of patients with coronary heart disease (CHD), invasive coronary procedures – such as percutaneous coronary intervention – are effective in reducing mortality and morbidity [ 1 , 2 ]. An important supplementary outcome of medical interventions and the processes of health care is health-related quality of life (HRQOL) [ 3 , 4 ]. There is a strong interest in differences in the care and outcomes between sociodemographic groups to optimize population health. Differential temporal changes in HRQOL between diverse sociodemographic groups may be of interest in secondary prevention programs to maximize the benefit from treatment for CHD. Most studies investigating the association between sociodemographic characteristics and HRQOL in patients with CHD focus on cross-sectional group comparisons [ 5 - 11 ]. Longitudinal studies on the association between sociodemographic characteristics and HRQOL have indicated that being female [ 12 - 14 ] and lower socioeconomic status [ 15 ] are associated with less temporal improvement in HRQOL. These studies, however, lacked information on medical interventions and focused on short-term changes lasting only up to 1 year. Studies assessing the effect of invasive coronary procedures on HRQOL, have shown that HRQOL improves after intervention [ 16 - 20 ] to levels similar to population norms [ 21 - 23 ]. Some of these studies were clinical trials and involved very selective populations. Only few studies have focused on the association of sociodemographic variables with temporal changes in HRQOL following invasive coronary procedures. A recent observational study indicated that higher income is associated with greater improvement in physical HRQOL following invasive coronary procedures [ 24 ]. Improvements in physical HRQOL appear to be unrelated to the age of patients [ 25 ], whereas elderly patients exhibit a stronger improvement in mental HRQOL after medical intervention [ 24 , 26 ]. However, the association of sex and educational attainment with changes in HRQOL following invasive coronary procedures remains inconclusive. In the present study of patients with CHD, we aimed to (i) describe the effect of invasive coronary procedures on different domains of HRQOL at both 6 weeks and 2 years after baseline hospitalization, (ii) assess the association between sociodemographic characteristics and temporal changes in HRQOL. Methods Baseline data in the present study were derived from the Norwegian study on outcomes research and quality improvement (RESQUA), a cross-sectional postal survey of HRQOL and the experiences of patients receiving hospital care [ 27 ]. All patients from surgical and internal medicine wards at 17 hospitals (4 teaching hospitals, 6 central hospitals, and 7 local hospitals) between October and December 1998 were sent a questionnaire 6 weeks after hospital discharge. No response within 4 weeks triggered one reminder. Patients younger than 16 years and those registered as dead at discharge were excluded from the study. Participants with CHD discharged from internal medicine wards were selected for a follow-up postal survey approximately 2 years later, in October and November 2000. We used information on primary and secondary diagnoses from the patient-administration systems of the hospitals, and included patients with acute coronary syndrome (ICD-9 410.xx and 411.xx) and angina pectoris (ICD-9 413.xx). Patients with chronic heart failure (ICD-9 428.xx) as the primary diagnosis were classified as angina pectoris or acute coronary syndrome dependent on their secondary diagnoses. Measures Health-related quality of life HRQOL was measured by the Norwegian version 1.2 of the Short-Form 36 (SF-36), a widely used generic health status measure that enables comparison with normative scores [ 28 , 29 ]. The scales and items of the SF-36 have satisfactory reliability, validity, and responsiveness, also in patients with CHD [ 3 , 10 , 30 , 31 ]. Single items of the SF-36 are transformed and aggregated into eight multi-item scales: Physical Functioning, Physical Role Limitations, Bodily Pain, General Health Perceptions, Vitality, Social Functioning, Emotional Role Limitations, and Mental Health. The resulting summated rating scales range from 0 to 100, with higher scores indicating better health. To estimate the potential impact of CHD on HRQOL, we compared the SF-36 scores from the patients in our study with normative data from the Norwegian general population [ 32 ]. Norm scores for the eight multi-item scales were adjusted to reflect age and sex distributions similar to those of the patients in the present study. These adjusted norm data for the eight multi-item scales were used to calculate the standardized Physical (PCS) and Mental (MCS) Component Summary scores [ 33 ]. Procedures, sociodemographics, and comorbidity Invasive coronary procedures referred to diagnostic and therapeutic procedures, such as coronary angiography, which contributes to diagnosis of potential coronary artery disease, and when followed by angioplasty, it can contribute to a relief from chest-pain as well as improve the prognosis in high-risk patients. Procedure codes were derived from the patient-administration systems of the hospitals in 1998 (Classification of Operations; version 3, 1995). We defined invasive coronary procedures as a dichotomous variable, differentiating between patients with and without invasive coronary procedures during baseline hospitalization in 1998. The age and sex data were also derived from the administration systems in 1998. Information about the highest level of educational attainment was obtained from self-reported data in the 1998 postal survey. This variable was heavily positively skewed, and we therefore created two groups: (1) below and equal to, and (2) above the median value in our cohort. As a crude estimate of the degree of comorbidity for each patient, we used the total number of secondary diagnoses registered in the administration databases during baseline hospitalization in 1998. Statistical analyses Changes in HRQOL were only assessed in patients who had valid scores on all multi-item scales both in 1998 and 2000. We used χ 2 -statistics or the t -test for independent samples to analyze the extent of selective attrition, and differences in the use of invasive coronary procedures across characteristics of respondents. Temporal changes in HRQOL were analyzed with paired-samples t -tests. As a measure of the minimally important difference in intra-individual scores, we calculated the standardized response mean, a distribution-based approach that compares temporal change by the standard deviation of change [ 34 ]. Standardized response means of 0.2–0.5, 0.5–0.8, and > 0.80 are regarded as small, moderate, and large, respectively [ 35 ]. Additionally, we applied multivariate linear regression analyses to determine the association of invasive coronary procedures and sociodemographic factors with PCS and MCS scores 2 years after baseline hospitalization. By including baseline PCS and MCS scores in the regression model, the regression coefficients of invasive coronary procedures and sociodemographic factors indicate the one unit increase in 2-year PCS and MCS scores, provided that baseline scores are held constant. All analyses were performed separately in patients with angina pectoris and acute coronary syndrome. We chose a 5% statistical significance level. The Regional Medical Research Ethics Committee, the Data Inspectorate, and the Norwegian Board of Health approved the study. Results A total of 1,534 patients with CHD were sent a questionnaire in 1998, and 1,059 (69%) responded. In 2000, 700 patients with valid HRQOL scores were sent a follow-up questionnaire (and, where necessary, one reminder), and 473 patients responded. After excluding 38 patients who had recently died and 9 patients with an unknown address, the adjusted response rate in the follow-up study was 72% (Figure 1 ). Figure 1 Flow chart describing attrition in the cohort of patients with coronary heart disease A total of 254 patients with angina pectoris and 90 patients with acute coronary syndrome had valid MCS and PCS scores both in 1998 and 2000, of which 108 patients with angina pectoris and 41 patients with acute coronary syndrome underwent an invasive coronary procedure. The majority underwent catheterization (N = 79) or percutaneous coronary intervention (N = 45). Twenty-one patients underwent coronary bypass surgery and the remaining patients (N = 4) underwent other medical procedures related to the cardiovascular system. Compared to the original cohort of patients, patients with valid HRQOL scores both in 1998 and 2000, more often had undergone invasive coronary procedures, were male, younger, and had lower comorbidity (Table 1 ). Compared to the cohort with valid responses in 1998, attrition was associated with age, educational attainment and comorbidity. Angina pectoris patients who responded both in 1998 and in 2000 had higher PCS scores in 1998 compared to nonrespondents to the follow-up survey (mean HRQOL score 42 versus 39; P < 0.001). Among patients with valid HRQOL scores both in 1998 and 2000, women, elderly patients, and patients with higher comorbidity had fewer invasive coronary procedures during the baseline hospitalization (Table 2 ). Educational attainment was not associated with invasive coronary procedures. Table 1 Baseline characteristics of respondents and non-respondents I. Total II. Non-respondents at 6 weeks III. Respondents at 6 weeks a IV. Respondents at 6 weeks & 2 years a Comparison Column IV and total nonresponse ( P -value) Comparison Column IV and nonresponse at 6 weeks ( P -value) N = 1534 N = 475 N = 700 N = 344 Age, mean (SD) 69 (12) 71 (12) 65 (11) 64 (10) <0.001 0.001 Gender (% women) 34 44 29 27 0.001 0.2 > 10 years education (%) 44 49 0.014 > 1 diagnosis (%) 65 72 60 56 0.001 0.039 Emergency admission (%) 70 80 62 58 <0.001 0.07 Acute Coronary Syndrome (%) 31 32 28 26 0.047 0.4 Invasive Coronary Procedure (%) 33 22 40 43 <0.001 0.06 Teaching Hospital (%) 54 48 57 59 Central Hospital (%) 26 27 25 22 0.042 0.123 Local Hospital (%) 20 25 18 19 a Respondents to all multi-item scales of the SF-36 Table 2 Invasive coronary procedures (ICP) according to characteristics of baseline hospitalization and sociodemographic characteristics in patients with angina pectoris and acute coronary syndrome Angina pectoris Acute coronary syndrome No ICP ICP No ICP ICP Baseline characteristics N = 146 N = 108 N = 49 N = 41 Sex Men (%) 70 77 73 76 Women (%) 30 23 27 24 Education ≤ 10 years (%) 53 52 53 44 > 10 years (%) 47 48 47 56 Age (mean; (SD)) 64 (11) 61 (9) 68 (9) 62 (10) Type of hospital Teaching (%) 34 98 31 90 Central (%) 33 2 36 10 Local (%) 33 0 34 0 Length of hospital stay (mean (SD)) 3.8 (3.1) 3.9 (3.7) 8.5 (4.8) 7.1 (7.3) No. of diagnoses (mean (SD)) 1.9 (1.1) 1.6 (0.9) 2.9 (1.2) 2.2 (1.0) Six weeks after hospitalization, patients with angina pectoris and acute coronary syndrome had lower scores compared to the Norwegian norm data in all domains of HRQOL (Table 3 [see Additional file 1 ]). Patients without invasive coronary procedures exhibited the largest differences, particularly in domains referring to physical aspects of HRQOL: Physical Role Limitations, Emotional Role Limitations, and Bodily Pain; but also in General Health Perceptions. Two years after the baseline hospitalization in 1998, scores on all multi-item scales were still below the scores of the norm population. Over the 2 years analyzed, Physical Role Limitations ( P = 0.001) and Social Functioning ( P = 0.003) improved in angina pectoris patients without invasive coronary procedures, corresponding to a small effect size (Table 3). Physical Functioning ( P = 0.03), Physical Role Limitations ( P < 0.001), and Bodily Pain ( P = 0.03) improved in patients with angina pectoris undergoing invasive coronary procedures. The change in Physical Role Limitations corresponded to a moderate effect size. A significant deterioration was found in General Health Perceptions ( P = 0.04). In patients with acute coronary syndrome without invasive coronary procedures, Physical Role Limitations ( P = 0.003), Social Functioning ( P = 0.005), and Emotional Role Limitations ( P = 0.009) significantly improved. Physical Role Limitations ( P = 0.001) improved in patients with acute coronary syndrome undergoing invasive coronary procedures. Patients with invasive coronary procedures showed a small improvement in PCS scores ( P = 0.034 for angina pectoris and P = 0.015 for patients with acute coronary syndrome). MCS scores remained stable during the 2 years of follow-up; only patients with angina pectoris without an invasive coronary procedure experienced a small improvement in MCS scores ( P = 0.007). Multiple linear regression analyses revealed that, after taking baseline PCS scores into account, invasive coronary procedures and being younger, male, and more educated were significantly associated with higher PCS scores in 2000 in patients with angina pectoris (Table 4 [see Additional file 2 ]). For these patients, being older was significantly associated with higher MCS scores in 2000. In patients with acute coronary syndrome, PCS scores and MCS scores in 2000 were significantly associated only with baseline scores, and not with invasive coronary procedures or sociodemographic characteristics. Discussion In the present study, most improvement was found in the physical components of HRQOL 2 years following invasive coronary procedures. These results support the findings of Krumholz et. al. [ 21 ] that the SF-36 scale for Physical Role Limitations was most responsive after elective coronary angioplasty. Furthermore, in patients with angina pectoris, PCS scores improved more among those who were male, younger, and more educated, independently of invasive coronary procedures. One explanation for this, as suggested by some previous studies, is related to differences in disease severity: women and patients from disadvantaged socioeconomic strata may have more extensive coronary disease at the onset of symptoms [ 12 , 13 , 36 ]. Additionally, undesirable events and adverse experiences might have stronger negative emotional consequences in this group [ 37 ], suggesting worse adaptation to the long-lasting physical limitations of CHD and a greater risk of recurrent events [ 36 ]. When adjusting for baseline scores, invasive coronary procedures and sociodemographic characteristics did not explain any additional variation in PCS and MCS scores 2 years after hospitalization in patients with acute coronary syndrome. This may be due to the relatively small sample size. An alternative explanation is that patients with acute coronary syndrome exhibited higher comorbidity that could limit the effect of invasive coronary procedures on HRQOL, and accordingly, the sensitivity of the SF-36 in detecting differences [ 20 ]. Our results demonstrated that invasive coronary procedures and sociodemographic characteristics were weakly associated with MCS scores and indicated small deviations from the population norm, which corresponds to previous findings in patients with CHD [ 6 , 38 ]. This may be attributable to health care having less impact on mental health than on physical health. An alternative explanation refers to the construction of the SF-36 MCS and PCS measures. The scores of these component scales are calculated using all eight multi-item scales with factor score coefficients derived from factor analysis with orthogonal rotation, thereby defining that PCS and MCS are uncorrelated. Mean scores on the multi-item scales that are below the population mean will contribute to component scores opposite to the direction defined by the factor score coefficient [ 39 ]. In our study, the low scores of Physical Role Limitations contributed negatively to PCS and positively to MCS. Hence, MCS scores were probably inflated by poor physical health. The RAND-36 has been suggested as an alternative method for computing PCS and MCS scores that avoids the orthogonal approach of the SF-36 [ 40 , 41 ]. Other factors may have influenced our results, for example selective attrition. In our study, the respondents to both surveys represent a survivor cohort, and hence attrition may have reduced the temporal changes in SF-36 scores and possibly lead to underestimation of the associations with invasive coronary procedures and sociodemographic factors. Moreover, the use of self-administered and postal questionnaires may have contributed to missing SF-36 items, especially in elderly subjects who are associated with a higher frequency of missing values for items used to score physical and emotional role functioning [ 24 ]. The appropriateness of the SF-36 for use in elderly populations with expected low response rates, reduced cognitive functioning, and shifts in conceptualizations of subjective health, has been discussed previously [ 32 ]. Consequently, caution should be exercised when employing norms among people aged 70 years and older. Another limitation of our study is the lack of HRQOL data before the baseline hospitalization in 1998, which prevented us from assessing the full impact of invasive coronary procedures on subsequent HRQOL and its association with sociodemographic characteristics. We also did not examine the influence of use of medical services after the baseline hospitalization. Finally, coronary patients and invasive procedures were defined by registry data from the patient-administration systems of hospitals, which might be inaccurate and mask some of the underlying clinical differences that could influence the HRQOL results. Our findings indicated that patients with angina pectoris who were younger, male, and more educated were most likely to increase their HRQOL following invasive coronary procedures. In patients hospitalized for acute coronary syndrome, temporal change in HRQOL was not associated with invasive coronary procedures and sociodemographic characteristics, possibly due to higher comorbidity. In a usual care setting the occurrence of invasive coronary procedures varies with sociodemographic characteristics [ 42 , 43 ]. The association of both sociodemographic variables and invasive coronary procedures with HRQOL outcomes makes it imperative to take these into account when comparing and interpreting change scores to reduce the risk of spurious findings. Authors' contributions MV carried out the follow-up study, analyzed the data, and drafted the manuscript. KIP performed the baseline survey. AR participated in the design of the study. KS participated in the design and coordination of the study. All authors read and approved the final manuscript. Supplementary Material Additional File 1 Table 3: SF-36 multi-item and summary scales 6 weeks and 2 years after baseline hospitalization in patients with angina pectoris and acute coronary syndrome according to invasive coronary procedures (ICP). Norwegian norm data and 2-year change scores Click here for file Additional File 2 Table 4 Predictors of Physical Component Summary (PCS) and Mental Component Summary (MCS) scores 2 years following the baseline hospitalization in patients with angina pectoris and acute coronary syndrome. Multivariate linear regression analysis; unstandardized regression coefficients ( B ) and 95% Confidence Intervals (CI) Click here for file	/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC524503.xml
551611	Cα-H···O=C hydrogen bonds contribute to the specificity of RGD cell-adhesion interactions	Background The Arg-Gly-Asp (RGD) cell adhesion sequence occurs in several extracellular matrix molecules known to interact with integrin cell-surface receptors. Recently published crystal structures of the extracellular regions of two integrins in complex with peptides containing or mimicking the RGD sequence have identified the Arg and Asp residues as key specificity determinants for integrin recognition, through hydrogen bonding and metal coordination interactions. The central Gly residue also appears to be in close contact with the integrin surface in these structures. Results When hydrogen atoms are modelled on the central Gly residue with standard stereochemistry, the interaction between this residue and a carbonyl group in the integrin surface shows all the hallmarks of Cα-H···O=C hydrogen bonding, as seen in the collagen triple helix and in many crystal structures of small organic molecules. Moreover, molecular dynamic simulations of the docking of RGD-containing fragments on integrin surfaces support the occurrence of these interactions. There appears to be an array of four weak and conventional hydrogen bonds lining up the RGD residues with main chain carbonyl groups in the integrin surface. Conclusions The occurrence of weak Cα-H···O=C hydrogen bonds in the RGD-integrin interaction highlights the importance of the conserved Gly residue in the RGD motif and its contribution to integrin-ligand binding specificity. Our analysis shows how weak hydrogen bonds may also play important biological roles by contributing to the specificity of macromolecular recognition.	Background The Arg-Gly-Asp (RGD) sequence is one of the most easily recognised motifs in molecular biology [ 1 ]. Discovered in fibronectin in 1984 [ 2 ], this tripeptide appears to be conserved in the cell attachment sites of many proteins from the extracellular matrix (ECM). The later discovery that RGD is recognised by members of the integrin family of cell surface receptors [ 3 ], confirmed the central role of RGD and suggested that its presence in a protein sequence might be indicative of cell-adhesion functionality [ 4 ]. Integrins are ubiquitously expressed heterodimer cell surface molecules that act as receptors for ECM molecules and other cell-surface adhesins. Through these cell-matrix and cell-cell interactions integrins control diverse cell functions such as adhesion, shape, growth, differentiation and mobility, and therefore contribute to important physiological processes such as development, immune responses and cancer [ 5 ]. Integrins are complex signalling engines: their extracellular domains interact with the ECM while their cytoplasmic tails interact with the cytoskeleton and other intracellular signalling molecules. Current hypotheses suggest that conformational changes resulting from these interactions enable integrins to transmit signals across the membrane in both directions. Recent advances in the structural biology of several integrin domains and their interactions with ligands have begun to define possible working scenarios for the signalling mechanisms [ 6 - 13 ]. As a consequence of their role in so many fundamental processes, integrin defects have been implicated in many common diseases, from cancer to pathogen invasion. An ability to block a particular integrin-ligand interaction may be a possible route to the control of certain pathological states, hence it is not surprising that some integrins have become attractive targets for drug design. Understanding the molecular bases of the interaction of integrins with their ligands is therefore essential for effective protein-based design of inhibitors or activators of their function. A milestone was reached in 2002 with the determination of the crystal structure of the extracellular segment of αVβ3 integrin in complex with a cyclic peptide containing the prototypical RGD sequence [ 8 ]. In that structure, the amino acids defining the RGD sequence are seen to establish specific interactions with corresponding residues in the integrin heterodimer surface, spanning the interface between the αV and β3 subunits (Figure 1a ). Very recently, another landmark paper has reported several crystal structures of the extracellular region of the fibrinogen-binding integrin αIIbβ3 [ 12 ]. In addition to providing an improved picture of the allosteric basis of integrin signal transmission, this new set of structures shows the molecular details of the interaction between the αIIbβ3 RGD-binding site and various ligand mimetics (Figure 1b ). These interactions are remarkably consistent with those previously observed in the complex between the αVβ3 integrin fragment and the cyclic RGD peptide ( c RGD) [ 8 ]. At first glance, two interactions consistently seen in these crystal structures appear to be key in defining the specific molecular recognition between the RGD sequence in an integrin ligand and the surface of its integrin receptor: the Asp residue of the RGD triad completes the coordination of a divalent metal ion bound to the β subunit, while the Arg side chain extends in the opposite direction to form salt-bridge hydrogen bonds with one or two Asp residues in the α subunit. These two specific interactions or their equivalent are seen both in the c RGD-αVβ3 structure and in the structures of αIIbβ3 in complex with ligand-mimetics (Figure 1 ). There are no significant hydrophobic "pockets" or exosites contributing to the binding specificity. For example, a large fraction of the c RGD peptide does not make any contact with the αVβ3 integrin surface (Figure 1a ). In the broader context of RGD-containing ligands and their integrin receptors, it would seem that these interactions are mainly electrostatic and that the two charged residues in the RGD sequence are necessary and sufficient for attachment [ 14 ]. What about the central Gly residue? In their analysis of the c RGD-αVβ3 crystal structure, Xiong et al. report that the Gly central residue makes several hydrophobic interactions with the integrin surface, including a contact with the carbonyl oxygen of residue Arg216 in the β3 integrin subunit [ 8 ]. Such contact between the Gly methylene group and a main-chain carbonyl oxygen is also observed in the crystal structure of αIIbβ3 in complex with the peptidomimetic eptifibatide (EFB) [ 12 ] (Figure 1b ), which is a cyclic heptapeptide containing a homo Arg-Gly-Asp sequence. The particular geometry of these contacts is strikingly reminiscent of a motif previously described in the collagen triple helix: a hydrogen bonding arrangement where the α-carbon of the Gly residue acts as hydrogen bonding donor in Cα-H···O=C interactions (Figure 2 ) [ 15 ]. So-called "weak" hydrogen bonds, such as those between carbon and oxygen atoms, have been traditionally neglected in descriptions of three-dimensional structures of macromolecules. Yet, C-H···O hydrogen bonds are ubiquitous in protein structures: virtually every conventional N-H···O=C hydrogen bond in every β-sheet in every determined protein structure carries a companion Cα-H···O=C interaction [ 16 , 17 ]. This applies to both parallel and antiparallel β-sheets, and exactly the same topology is also observed in the collagen triple helix [ 15 ]. For collagen and the β-sheet structures, the occurrence of Cα-H···O=C interactions is indicative of a very tight fit between the molecules involved, a close-packed structure in which all groups participate in some form of hydrogen bonding interaction. How important are Cα-H···O=C and other weak hydrogen bonds in shaping the three-dimensional structure of proteins and macromolecular complexes? The subject has stimulated considerable debate (see [ 18 ] and [ 19 ] for reviews), although theoretical studies leave no doubt about the cohesive nature of these interactions [ 20 - 23 ]. With a strength approximately one-half of that from conventional hydrogen bonds, it seems reasonable to assume that the large numbers of weak hydrogen bonds detected in proteins may contribute to their stability. Furthermore, several biochemical functions have been linked to specific C-H···O hydrogen bonds, where position is more important that numbers. One example is the Gly-X-X-X-Gly motif, known to favour helix-helix interactions in membrane [ 24 ] and soluble proteins [ 25 ] via position-specific Cα-H···O=C hydrogen bonds. Another is the proposed role of C-H···O hydrogen bonds from cytosine and thymine bases to amino acid side chains during DNA-protein recognition [ 26 ]. Weak C-H···O hydrogen bonds have also been surveyed at protein-protein interfaces [ 27 ], and have been reported to play specific roles in catalysis [ 28 ], and in substrate and inhibitor recognition [ 29 - 32 ]. Recently, a server to identify weak hydrogen bonding interactions in protein structures has been made publicly available [ 33 ]. The functional occurrence of weak C-H···O hydrogen bonds in protein-ligand, protein-protein, and protein-DNA recognition suggests that their presence should be examined in detail in the structures of macromolecules with biomedical or biotechnological interest. Their potential should not be neglected in rational drug design approaches [ 31 ]. With this in mind, we present here an analysis of possible Gly-Cα-H···O=C interactions between RGD motifs and the RGD-binding sites from the αVβ3 and αIIbβ3 crystal structures. We conclude that the mutual geometry of the interaction is consistent with Cα-H···O=C hydrogen bonding. We discuss the implications of these hydrogen bonds for the cell adhesion interactions between integrins and their RGD-containing ligands. Results and discussion Building standard-geometry Hα atoms on the Gly central residues of the c RGD and EFB peptides produces the geometric arrangements shown in Figure 2 , clearly reminiscent of the hydrogen bonding pattern previously described in the collagen triple helix (Figure 2c ). The metrics of these Gly-carbonyl contacts are shown in Table 1 . The Cα···O distances in the c RGD-αVβ3 and EFB-αIIbβ3 structures appear to be longer than the mean Cα···O distance in collagen, but are well within the observed range in crystal structures of small organic molecules (see below). Both Hα atoms from the collagen Gly residues are in hydrogen bonding position (Cα–Hα···O > 90°), and their Cα-Hα···O=C hydrogen bonds adopt a three-centred and bifurcated configuration (Figure 2c and [ 15 ]), that is not seen in the integrin structures. Nevertheless, the central Gly residues in the c RGD and EFB peptides appear to have one and two Hα atoms respectively in hydrogen bonding position to the carbonyl group of Arg216, a residue on the surface of the β3 subunit and directly at the interface with the αV and αIIb subunits. An obvious caveat to this analysis comes from the moderate resolution of the c RGD-αVβ3 and EFB-αIIbβ3 crystal structures (3.2 Å and 2.9 Å respectively). Positional errors inevitable at that resolution may affect the precision of the fitting of the c RGD and EFB peptides and the accuracy of the hydrogen bonding geometries for both weak and strong hydrogen bonds. For example, a close look at the salt-bridge interactions between the Arg guanidinium group from the c RGD peptide and two Asp side chains on the αV integrin surface (Asp150 and Asp218), shows less than "ideal" hydrogen bonding orientation, especially for Asp150 (not shown). Yet, the accumulated knowledge of hydrogen bonding geometries in high-resolution crystal structures and their significant variability leaves no doubt about the existence of these strong hydrogen bonds and their contribution to the specificity of binding. A similar level of confidence can be achieved for the Gly-Cα-Hα···O=C hydrogen bond by analysing the metrics of equivalent interactions in high-resolution crystal structures of small organic and organometallic molecules. Figure 3 shows the two fragment probes used in a statistical search for Gly-Cα-Hα···O=C nonbonded interactions in the Cambridge Structural Database (see Methods). Figure 4 shows that single hydrogen bonding (only one angle Cα-Hα···O ≥ 90°) predominates over the bifurcated case, and that a broad maximum in the Cα···O distribution occurs at about 3.4 Å, which can be taken as the "hydrogen bonding distance" for this type of interaction. This value is consistent with the theoretical value of 3.34 Å for the Gly-Cα-Hα···OH 2 hydrogen bond from ab initio quantum calculations [ 22 ]. Figure 5 shows the distributions of Hα···O distances and Cα-Hα···O angles for the single hydrogen bond. The Hα···O distribution shows a broad maximum around 2.7 Å, whereas the angular distribution is very broad with maxima around 110° and 140°. Average parameters for single and double Gly-Cα-Hα···O=C hydrogen bonding (Table 1 ) are perfectly compatible with those calculated for the c RGD-αVβ3 and EFB-αIIbβ3 structures respectively, even though the accuracy of the values shown in Table 1 is clearly overestimated with respect to the resolution of these crystal structures. Thus, strictly from a geometrical point of view, the contacts between the Gly residues in the c RGD and EFB peptides and the main chain carbonyl group from Arg216 in the integrin surface bear all the characteristics of Cα-Hα···O=C hydrogen bonding. This observation is consistent with the exceptionally high frequency of intermolecular Gly-Cα-H···O=C hydrogen bonds recently reported in high resolution crystal structures of protein-ligand complexes [ 32 ]. A simple molecular docking analysis further supports the occurrence of Gly-Cα-H···O=C hydrogen bonds between RGD-containing ligands and integrin binding sites. In a first set of calculations, an RGD tripeptide was docked into the binding sites of both αVβ3 and αIIbβ3 integrins using constrained molecular dynamics (MD). Two constraints were imposed in the docking calculations: the carboxyl group from the Asp residue had to complete the metal coordination on the β3 subunit, and the Arg side chain had to form a salt bridge with appropriate Asp residues in the αV and αIIb subunits (see Methods), as observed in the crystal structures of αVβ3 and αIIbβ3 with different ligand-mimetics. Ten slightly different RGD models were obtained from the NMR structures of the adhesion domain of fibronectin [ 34 ], and were placed about 10 Å away from the integrin surface. Then these RGD models were subject to MD simulations until they docked into the integrin binding sites. In a second set of calculations, a longer peptide fragment with sequence VTGRGDSPAS from the adhesion domain of fibronectin was also docked into the binding sites of the two integrins (Figure 6 ). Again, ten different models for this peptide were obtained from fibronectin NMR structures [ 34 ]. Most of the simulations converged to models with Cα···O contact distances between the central Gly residue and the carbonyl of Arg216 in the 2.7–3.7 Å range (Figure 6b ), with either one or two Gly-Hα atoms in hydrogen bonding orientation. These models were also the most favourable energetically (Table 2 ). From these calculations it seems to emerge that the RGD binding sites of the αVβ3 and αIIbβ3 integrins are primed to place the central Gly residue in the RGD triad directly above the carbonyl group of Arg216 of the β3 subunit (as observed in the c RGD-αVβ3 and EFB-αIIbβ3 crystal structures), forming one or two Cα-H···O hydrogen bonds that complement the main metal-coordination and salt-bridge interactions from the Asp and Arg side chains. How important are these weak C-H···O hydrogen bonds in stabilising the c RGD-αVβ3 and EFB-αIIbβ3 complexes? A quantitative analysis of C-H···O hydrogen bonding at protein-protein interfaces has shown that they have an important contribution to the association and stability of protein complexes, accounting for about one third of the total hydrogen bonding interaction energy [ 27 ]. In fact, some of the hydrophobic or van der Waals interactions usually invoked to explain stabilising close contacts between molecules can be described better as weak C-H···O hydrogen bonds. These occupy a middle ground between the highly directional, conventional hydrogen bonds, and the directionless van der Waals interactions [ 32 ]. The recurrent appearance of some weak hydrogen bonding topologies in many structures of proteins and at protein-protein interfaces also reinforces the notion that they have a significant contribution to macromolecular stability. The most common occurrence of C-H···O hydrogen bonds in protein structures is a widespread Cα-H···O=C hydrogen bond N -terminal to the conventional N-H···O=C hydrogen bond in β-sheets [ 16 , 17 ] and in the collagen triple helix [ 15 ]. In this structural motif (Figure 7a ), the Cα-H donor group is in the residue immediately N -terminal to the one carrying the N-H donor group, and both share the same C = O group as acceptor, an arrangement sometimes referred as "bifurcated" hydrogen bond [ 17 , 18 , 27 ]. This bifurcated hydrogen bonding motif is also the most common occurrence of C-H···O hydrogen bond at protein-protein interfaces [ 27 ]. The situation in Figure 7b occurs when the residue N -terminal to the one carrying the N-H donor group is Gly, with one or two Hα from Gly being in hydrogen bonding position. The bifurcated hydrogen bond scenario also occurs in the c RGD-αVβ3 and EFB-αIIbβ3 structures, where the N-H group from the Asp residue in the RGD peptide donates a hydrogen bond to the main chain carbonyl group from Arg216 of β3. This hydrogen bond has very bad geometry in the c RGD-αVβ3 structure (distance H···O 2.69 Å, angle N-H···O 133°), but looks better in the EFB-αIIbβ3 structure (distance H···O 2.39 Å, angle N-H···O 144°). These deviations from ideal hydrogen bonding geometry might be consequence of the resolution of the crystal structures, but all the MD docking simulations described above result in N-H···O=C hydrogen bonds that are slightly longer (typical H···O distances 2.4–2.5 Å) and slightly less linear (typical N-H···O angles 140°-150°) than the average hydrogen bonds between peptide groups in protein secondary structures. Automatic computational docking calculations of known integrin ligands on structural models of αVβ3 and αVβ5 consistently predict this N-H···O=C hydrogen bond to occur whenever an N-H group is present in the proximity of the carboxylate moiety [ 35 , 36 ]. Had these simulations included all nonpolar hydrogens, the companion Cα-H···O=C hydrogen bonds from the Gly residues would also have been observed. A quick inventory of hydrophobic interactions between the c RGD and EFB peptides and the αVβ3 and αIIbβ3 surfaces suggests an additional candidate for classification as C-H···O=C hydrogen bond, between the Cβ-Hβ group from the Asp residue of the c RGD and EFB peptides and the main chain C = O group from Asn215 in the β3 subunit. This weak hydrogen bond is adjacent to the stronger, conventional hydrogen bond between the main chain N-H group from Asn215 and Oδ2 from the Asp residue in the RGD motif. Thus, a total of four hydrogen bonds, weak and conventional, aligns the bottom of the c RGD and EFB peptides against the integrin surfaces (Figure 8 ), and complements the main interactions from the Asp carboxyl and Arg guanidinium groups to provide a higher binding specificity. It is clear from the c RGD-αVβ3 and EFB-αIIbβ3 structures that any side chain other than Gly in the RGD triad would not allow it to fit snugly within the integrin binding site, with the resulting weakening of hydrogen bonding and van der Waals interactions. Furthermore the main chain conformation for the central Gly residue in the c RGD-αVβ3 structure falls in a region of the Ramachandran map that is not allowed to any L-amino acid residue. Thus, Gly residues at the centre of the RGD motif are essential for being small, for being able to adopt specific main chain conformations, and for being able to interact closely with the integrin surface via Cα-H···O=C hydrogen bonds. All three characteristics contribute to the integrin-binding specificity of Gly residues at the centre of RGD motifs. Inasmuch as the c RGD-αVβ3 and EFB-αIIbβ3 structures remain valid models for the structural basis of integrin-RGD ligand-binding specificity, it is reasonable to assume that the weak Cα/β-H···O=C hydrogen bonds depicted in Figure 8 will also occur in RGD-based cell-adhesion interactions. A special feature of the integrin surface at the RGD-binding site is the presence of two main chain carbonyl groups exposed to the solvent in the β3 subunit: Asn215 and Arg216. In absence of ligands these groups will probably interact with water molecules through conventional hydrogen bonding interactions (as seen for example in the crystal structure of the cacodylate-bound form of αIIbβ3, PDB accession code 1TXV [ 12 ]). Upon ligand binding, the RGD residues will displace these waters and place one amide and two methylene groups in hydrogen bonding position to carbonyl groups, increasing the specificity of the RGD-integrin interaction through multipoint recognition (Figure 8 ). This strategy will obviously be exploited by many competitive inhibitors for the integrin RGD-binding site. For example it is possible to substitute the weaker Cα-H donors from the Gly residue by a conventional N-H group (Figure 7c ). This strategy has been exploited already in the design of aza-peptide and azacarba-peptide RGD mimetics [ 37 - 39 ], several of them with nanomolar activity. Molecular modelling of the interaction of these peptides with αVβ3 and αVβ5 RGD binding sites predicts the hydrogen bonding topology shown in Figure 7c [ 36 ]. It is interesting to notice that even in the absence of a conventional hydrogen bonding donor, the carbonyl group Arg216 in the β3 subunit still may be acceptor for weak hydrogen bonds. In the crystal structure of αIIbβ3 in complex with tirofiban [ 12 ], a non-peptidomimetic inhibitor derived from L-tyrosine, the Cδ1 atom from the substituted Tyr ring is some 3.01 Å away from the carbonyl oxygen of the very same Arg216. If a hydrogen atom is built with standard geometry on Cδ1, the calculated Hδ1···O distance is 2.01 Å and the Cδ1-Hδ1···O angle is 172°, again hydrogen bonding-like metrics. How should this Cδ1···O=C contact be called? We think that a description in terms of weak C-H···O hydrogen bonding is in this case more accurate than referring to this interaction as simply hydrophobic. Conclusions We have analysed in detail recently published structural data on the interaction between the extracellular regions of two integrins and peptides containing or mimicking the RGD sequence [ 8 , 12 ]. From this analysis we conclude that Cα-H···O=C hydrogen bonds from the central Gly residue also contribute to the specificity of binding. Weak hydrogen bonds are traditionally overlooked when describing protein structures, although they probably contribute to their stability. We think that our analysis provides one of the most interesting examples of C-H···O hydrogen bonds playing an important biological role, and may contribute to reverse the current trend of neglect of these interactions. In a recent paper, Sarkhel and Desiraju suggest that Nature may take advantage of the weaker C-H···O hydrogen bonds to optimise the efficiency of protein-ligand interactions, with a larger number of interactions coming into play even at the expense of the strength of the individual interactions [ 32 ]. By using more interactions, they suggest, specificity of recognition is increased, and because individual interactions are weaker, reversibility is possible. Our analysis of the interaction between the c RGD and EFB peptides and the αVβ3 and αIIbβ3 integrin surfaces would seem to corroborate this suggestion. Methods Integrin binding sites and hydrogen building The following crystal structure coordinates were downloaded from the Protein Data Bank [ 40 ]: αVβ3 integrin in complex with a cyclic RGD peptide ( c RGD), PDB accession code 1L5G [ 8 ]; αIIbβ3 integrin structure at 2.7 Å resolution, PDB accession code 1TXV [ 12 ]; αIIbβ3 in complex with eptifibatide (EFB), PDB accession code 1TY6 [ 12 ]; αIIbβ3 in complex with tirofiban, PDB accession code 1TY5. Models for integrin RGD-binding sites on αVβ3 and αIIbβ3 were obtained by selecting coordinates from integrin residues within 10 Å from the bound peptides. For the αIIbβ3 binding site, coordinates of the corresponding residues in the 1TXV structure were used, as this crystal structure has a better resolution. Coordinates for metal ions and structural waters present in the RGD-binding sites but not interfering with the binding of c RGD or EFB were also maintained. Hydrogen atoms were built with standard stereochemistry for the c RGD and EFB peptides and for the integrin RGD-binding sites as defined above, using the program REDUCE [ 41 ]. For the purpose of the analysis presented here all hydrogen atoms discussed in this paper could be positioned with satisfactory accuracy and predictable orientation. Molecular docking calculations For the molecular docking calculations, conformational models for RGD and VTGRGDSPAS peptides were obtained from the NMR structures of the adhesion domain in fibronectin [ 34 ]. Ten conformational models were used for each peptide. Each model was first manually docked approximately into the coordinates of the binding sites of αVβ3 and αIIbβ3 integrins, using the c RGD-αVβ3 and EFB-αIIbβ3 structures for guidance. Then each docked model was pulled away to about 10 Å from the integrin surfaces, and was docked back into the integrin binding site via molecular dynamics (MD) simulations using the program CNS [ 42 ]. Five simulations were run for each model, to a total of 50 MD simulations for each peptide-integrin pairing. A set of distance restraints was applied to the docking MD simulations, as observed on the c RGD-αVβ3 and EFB-αIIbβ3 structures. The side chain of the Asp residue group was restrained to coordinate the bound metal ion in the RGD-binding sites and to receive a hydrogen bond from the amide group of Asn215, in the β3 subunit. The side chain of the Arg residue was restrained to form hydrogen bonds with residues Asp150 and Asp218 on the αV subunit or residue Asp224 on the αIIb subunit. Additional restraints were imposed in the MD simulations with the VTGRGDSPAS peptide: the ring of Pro172 was restrained to hydrophobic contact with the side chain of Lys125, in the β3 subunit, and the Cα atoms of the N - and C -terminal residues in the peptide model were restrained not to separate more than 5 Å from each other. The coordinates of the integrin binding sites were kept fixed in all the simulations, and only the peptides were allowed to refine by restrained MD and energy minimisation. All molecular models were analysed with the program CHAIN [ 43 ] in a Silicon Graphics workstation. Analysis of hydrogen bonding geometry in crystal structures of organic molecules A survey in the Cambridge Structural Database [ 44 ] (July 2003 release), was carried out for Cα···O contacts between glycine-like fragments and carbonyl groups (Figure 3 ). List of abbreviations ECM, extracellular matrix; RGD, Arg-Gly-Asp sequence; c RGD, cyclic pentapeptide with sequence Arg-Gly-Asp-D-Phe-N(Me)-Val; EFB, eptifibatide; PDB, Protein Data Bank; CSD, Cambridge Structural Database; MD, molecular dynamics. Authors' contributions J.B. conceived the study and carried out the analysis of the structural data and molecular docking calculations. Both authors participated in the design, coordination and writing of the manuscript. Both authors read and approved the final manuscript.	/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC551611.xml
535340	Cross-species hybridisation of human and bovine orthologous genes on high density cDNA microarrays	Background Cross-species gene-expression comparison is a powerful tool for the discovery of evolutionarily conserved mechanisms and pathways of expression control. The usefulness of cDNA microarrays in this context is that broad areas of homology are compared and hybridization probes are sufficiently large that small inter-species differences in nucleotide sequence would not affect the analytical results. This comparative genomics approach would allow a common set of genes within a specific developmental, metabolic, or disease-related gene pathway to be evaluated in experimental models of human diseases. The objective of this study was to investigate the feasibility and reproducibility of cross-species analysis employing a human cDNA microarray as probe. Results As a proof of principle, total RNA derived from human and bovine fetal brains was used as a source of labelled targets for hybridisation onto a human cDNA microarray composed of 349 characterised genes. Each gene was spotted 20 times representing 6,980 data points thus enabling highly reproducible spot quantification. Employing high stringency hybridisation and washing conditions, followed by data analysis, revealed slight differences in the expression levels and reproducibility of the signals between the two species. We also assigned each of the genes into three expression level categories- i.e. high, medium and low. The correlation co-efficient of cross hybridisation between the orthologous genes was 0.94. Verification of the array data by semi-quantitative RT-PCR using common primer sequences enabled co-amplification of both human and bovine transcripts. Finally, we were able to assign gene names to previously uncharacterised bovine ESTs. Conclusions Results of our study demonstrate the harnessing and utilisation power of comparative genomics and prove the feasibility of using human microarrays to facilitate the identification of co-expressed orthologous genes in common tissues derived from different species.	Background Microarrays are routinely used for large scale transcriptome analyses and have been widely and successfully employed for simultaneously monitoring the expression of a potentially unlimited number of genes in parallel, thus providing the basis for identifying genes differentially expressed in distinct cell-types, developmental stages, disease states and cells subjected to exogenous reagents [ 1 ]. The rapid and significant improvements of cDNA-chip technologies and the availability of multi-species gene catalogues within the various data bases have made possible the comparison of gene expression levels within a single mammalian organism and across different organisms on a large-scale. The advantages of cross-species hybridisation are two-fold. First, cross-species gene-expression comparison is a powerful tool for the discovery of evolutionarily conserved mechanisms and pathways of expression control. The advantage of cDNA microarrays in this context is that broad areas of homology are compared and hybridization probes are sufficiently large so that small inter-species differences in nucleotide sequence would not affect the analytical results. This comparative genomics approach would allow a common set of genes within a specific developmental, metabolic, or disease-related gene pathway to be evaluated in experimental models of human diseases. Second, the use of microarrays in studies in mammalian species other than human and rodents, for example nonhuman primates, bovine, sheep and porcine may advance our understanding of human health and disease, for example the use of animal models in drug target validation. However, the inavailability of adequate sequence data and commercial cDNA and oligonucleotide microarrays keeps this technology beyond the reach of investigators working on economically and scientifically important large domestic species such as cattle, pigs and sheep. A potential solution to this problem is the use of cross-species hybridisations, i.e, human sequence-based arrays as tools for undertaking comparative genome expression studies. Such analyses have been performed using ape brain RNA as target on a human oligonucleotide array [ 2 ] and pig, mouse and Atlantic salmon RNA on human nylon arrays- [ 3 - 7 ]. These types of studies represent critical areas of research directly related to the understanding of human diseases because nonhuman primates, bovine, sheep and porcine play a crucial role in biomedicine, such as, organ transplantation, vaccine development, viral pathogenesis, gene therapy and a host of other human health-related technologies. A crucial step employing domestic animals in biomedicine is genetic modification which requires extensive embryo and embryo-related technologies, such as in vitro production of embryos for stem cell derivation and somatic nuclear transfer cloning. Employing the bovine model and sensitive RT-PCR assays, it has been shown that the majority of embryos derived from such sources display distinct mRNA expression patterns in a variety of developmentally important genes compared to their in vivo derived counterparts [ 8 ]. Some of these aberrations lead to "Large offspring syndrome", a complex of multiple pathologies observed in offspring derived from in vitro production and/or nuclear transfer of which significant oversize is a predominant feature [ 9 ]. Analysis of mRNA expression patterns in early embryos via cDNA microarray technology would provide insights into the function of gene regulatory networks and would thus be a major step forward in unravelling molecular mechanisms underlying developmental abnormalities. The technology to amplify the minute amounts of mRNA in early embryos without significantly altering the ratio of the various mRNAs in the original cell has recently been described [ 10 , 11 ] and a prototype mouse cDNA-macroarray enriched in embryonic sequences has been developed [ 11 ]. Important criteria for evaluating any microarray system include the reproducibility of the data generated, the specificity of detection of the targeted gene, and the validity of the results that identify and establish differential gene expression. The experiments described here show the systematic validation of cross-species microarray analysis, with emphasis on the reproducibility and statistical analysis of generated data using standard microarray data analysis tools. Specifically, we investigated the feasibility and reproducibility of cross-species hybridisation of orthologous genes within a defined developmental and metabolic pathway using as a test case and the first description of its kind, human and bovine fetal brain RNA as Cy-dye labelled targets on a human cDNA microarray. The microarray is composed of 349 genes each spotted 20 times to ensure reproducible validation by independent technologies such as semi-quantitative RT-PCR as carried out in this study, or alternatively, Real-Time PCR and Northern blot analysis. Results Array fabrication and gene annotation The human cDNA microarray used in this study consisted of 349 fully sequenced and annotated cDNAs as described in the Supplemental Table 1 . Thirty-five spots containing only spotting solution (3x SSC / 1.5 M betaine) served as negative controls. In addition, in a separate control plate, the housekeeping genes HPRT and β- ACTIN (made up of a dilution series of 25-, 50-, 100-, 150- and 200 ng/μl) were employed as endogenous guide dots to enable accurate grid placement prior to image analysis. These spots can be seen as intense yellow signals at the periphery of each block as portrayed in Figure 1A . The embryonic-specific gene, OCT-4 and Arabidopsis cDNAs ( Cab - T97312; Cwlp - T02614; Lhb1B1 - T21965; Ohp - T22679) were included as negative controls to monitor hybridisation-specificity. The transcription factor OCT -4 is expressed in human embryonic stem cells and primordial germ cells and is down-regulated upon differentiation [ 10 , 12 , 13 ]. Table 1 Summary of primer sequences, annealing temperatures and size of amplicons. Derivation of primer sequences: H and B denote human and bovine respectively. The nucleotide positions of the primer sequences are in parenthesis. Nucleotides highlighted in bold denote differences in the orthologous gene sequences. Genes Primer sequences (5' – 3') Annealing temperature ( ° C) Fragment size (bp) Accession numbers CDC27 (1226–1250) H TATTACATCTCCCCCAAACGCACTG 54 311 NM_001256 H CB170694 B (1512–1536) H CCATTTCACGAAGAAGGCTCATCAA CDC6 (89–113) B TCCCAAGCGGGTTGGT G TTATTCAC 54 208 NM_001254 H BF600055 B (272–296) B GCGACAGACT G TACTGTAGGCTTCA PCNA (195–217) B GAGGCGC T TAAGGATCTCA T CAA 54 382 AF527838 H CB531519 B (554–576) B ATTCACCAGAAGGCATCTTTACT SLC11A3 (38–61) B ACCCCTGGAGGGAACTCATCTAAT 57 276 AF215636.1 H BG689712 B (290–313) B GCTGATGCTCCCATCAAAATACTG VDAC2 (652–675) B CCT C GGTTGTGATGTTGACTTTGA 57 448 NM_003375.1 H TC152811 B (1076–1099) B GTGGCCTCCAGCATTAATGCTCTT CALNEXIN (1054–1078): H GCTGGTTAGATGATGAGCCTGAGTA 56 237 NM_001746.1 H BM431118 B (1266–1290): H TCC TG GGTTTCCAGATTCCCTGGTA β-ACTIN 441–461) H GTTGCTATCCAGGCTGTGCT 60 469 NM_001101 H NM_173979 B (890–910) H CGGATGTCCACGTCACACTT Figure 1 Array fabrication and gene annotation (A) False-colour image generated from a bovine-Cy5, human-Cy3 hybridisation. False colour images were generated using the programme ScanAlyze, version 2.44 . The full array (24 × 25 spotting pattern) consists of 16 blocks with each gene spotted 5 times per block therefore 80 potential data points present for expression analysis per gene. A blow-up of 4 blocks illustrating the Chip design is presented. The ACTIN cDNAs acting as guide-dots can be seen as intense yellow spots demarcating each block. In addition, the majority of spots appear yellow due to similar expression levels of the orthologous genes. (B) Functional annotation of the 349 genes as set out by the Gene Ontology Consortium .The proportion of these genes within each functional / biological annotation is represented on the Y-axis and the annotation on the X-axis. (C) Chromosomal distribution of the 349 genes. The number of genes within each chromosome is represented on the Y-axis and the chromosome numbers on the X-axis. Classification of the genes according to Gene Ontology annotations (molecular function) and chromosomal location (Figures 1B and 1C ) demonstrates that the selected genes encompass a range of twelve different functional classes and are located on all except the Y-chromosome. This implies that there is no obvious bias towards biological characteristics and the selected gene set can be viewed as representative for the current study. Global data characteristics Our experimental design incorporated dye-swaps and four replicated hybridisations. Three hundred and forty nine cDNAs, each representing a unique gene were spotted 20 times on each microarray. Four independent experiments were performed with human and bovine brain, respectively. Within each experiment the 20 replicates were averaged to yield a reliable signal for the respective probe. In the next step, the replicated signals from the different experiments were averaged to compute overall characteristics. The overall correlation of expression of the human and bovine genes is shown in Figure 2A . The regression line bears a slope close to one (1.13) thus indicating similar expression of human and bovine genes across the replicated experiments with a slight increase of expression in the human brain. The overall correlation coefficient is 0.94. Figure 2 Global data characteristics (A) Global correlation of bovine (X-axis) and human (Y-axis) experiments. The plot shows a log-log (base 2) plot of the mean signal intensities from the four independent experiments for each gene. The red line shows the regression line, the box displays the parameters of the regression line (intercept, slope) and the overall correlation. Four genes that are significantly differentially expressed as judged by using the Wilcoxon's rank-sum test are denoted by black diamond shapes. (B) MA-plot of a single experiment. The X-axis show the log (base 2) of the squared product of Cy5-red and Cy3-green intensities of each gene in the same experiment. Y-axis shows the log-ratio (base 2) of red and green intensity. The horizontal lines mark the two-fold over-expressed genes (Top line; bovine over-expression and bottom line; human over-expression). (C). Classification of detection levels of the 349 genes. The genes were sub-divided into three classes of expression levels. Class 1 (BG-tag < 0.8) – low (not detected). Class 2 (BG-tag between 0.8 and 0.9) – medium (boarderline detection) and Class 3 (BG-tag > 0.9) – high expression (readily detectable) The number of genes (X-axis) within each class (Y-axis) is also shown in the figure (black = bovine, white = human). Additionally, we computed MA-plots [ 14 ] in order to detect artificial dependencies of the log-ratio across the signal range. Figure 2B illustrates a typical result from a specific hybridisation experiment. The horizontal lines mark the two-fold levels of over-expression in bovine and human, respectively. A clear observation is that a few of the 349 genes under investigation fall outside these thresholds with the expression of 5 genes being more than 2-fold over-expressed in bovine and 16 genes more than 2-fold over-expressed in human. However, there is no increase or decrease in fold-change for the vast majority of genes and this is fairly stable across the signal range. In order to measure whether a given gene was significantly expressed, we compared its cDNA's signal to a signal distribution derived from negative controls represented by approx. 2,500 empty spot positions on the array. After quantification of each array a low non-zero intensity is assigned to each of these empty spots reflecting the amount of background signal on the array. Since these positions are spread uniformly over the array, the distribution of these signals reflects the distribution for signal noise and is an indicator whether signals are at the background level or reflect reliable expression levels. For each cDNA we counted the relative proportion of empty positions on the array that are smaller than the actual observed intensity (BG-tag). BG-tags from replicated experiments for the same cDNA were averaged. Thus, high values (close to one) indicate that the cDNA is expressed in the respective tissue whereas low values reflect noise. The limit of visual detection of a spot corresponds to a BG-tag level of 0.9, however there is a grey zone around this value. Comparison with RT-PCR analyses showed [ 15 ] that this level is consistent with the limit of detection at the 25 th cycle of a standard PCR reaction. cDNAs were considered as "detected" when their average BG-tag was above 0.9. We grouped the 349 genes into three classes of expression levels with respect to their intensity values above background levels (i.e, BG-tag) with class 1 as low (not-detected), class 2 as medium (possibly detected) and class 3 as high level of expression (detected), respectively. Within class 1, are 19 bovine and 17 human genes. Class 2, comprises 26 bovine and 17 human genes and finally class 3 has 304 bovine and 315 human genes, respectively. The genes within expression class 1 have signal intensities below the detection level of our microarray analysis platform and as such these genes can be designated as "Absent" (See Figure 2C for a graphical illustration). A list of all the intensity values is given (see Additional file 1 ). Data reproducibility and species variability An essential criterion for the applicability of cross-species experiments is the data reproducibility. For example, if human probes hybridise to bovine mRNA with far less reproducibility than to human mRNA high changes in expression of bovine tissues can be observed that are solely due to technical variability. In order to test whether the variability in gene expression levels is conserved within both species or in contrast, is higher in the bovine than in the human hybridisations, we calculated the coefficient of variation (CV) across the replicated experiments. Histograms are shown in Figure 3A . We then defined four classes of CVs for the genes (highly reproducible signals – CV < 0.25, good reproducibility – 0.25 < CV < 0.5, medium reproducibility – 0.5 < CV < 0.75 and poor reproducibility CV > 0.75). Note that a CV of 1.0 indicates that the signal standard deviation is in the order of the signal itself and therefore no meaningful statement on the measurement can be made. Figure 3 Data reproducibility and species variability (A) Histograms of CVs for human and bovine hybridisations. (B) Classification of signal variability in measured gene expression intensities. Human (white), bovine (grey) and mixed samples (black). CVs (co-efficient of variation) were calculated for repeated signal intensities of the four independent hybridisations and then sub-divided into four classes: Class 1 – CV < 0.25 high reproducibility, Class 2 – CV [0.25 – 0.5] good reproducibility, Class 3 – CV [0.5 – 0.75] medium reproducibility Class 4 – CV > 0.75 poor reproducibility. Figure 3B depicts the number of genes that fall within the respective CV-classes when analysing all experiments (human and bovine- black), human only (white) and bovine only (grey). There is a 10% decrease in the number of reproducible genes when comparing human and bovine but the overall effect is similar. Approximately 305 genes show a CV<0.5 with human brain (87%) compared to 280 with bovine brain (80%). Only 15 genes (4%) show high variability across the experiments for human brain compared to 20 genes (5.7%) for bovine brain. Figure 3B illustrates that there is a slight but not dramatic increase in variability when performing cross-species hybridisation (the number of variable genes (class 3 and 4) increase by a factor of 1.5). However, for the vast majority of the genes the reproducibility is similar. The observed effect is not due to exact cutoff values set for the different CV-classes since a slight shifting of the class borders leads to the similar results. For example, shifting the borders about a slight factor of ϑ = ± 0.05 to the left/right respectively gives approximately the same factor of increase for variable bovine to variable human genes. Differential gene expression in bovine and human fetal brain Three statistical tests that judge the significance of differences in the levels of gene expression in human and bovine fetal brains were employed (Student's t-test, Welch-test and, Wilcoxon's rank-sum test) as described previously [ 16 ]. Tests were carried out with the four independent experiments in bovine and human brain, respectively. Note that with independent experiments we mean the independent technical replicates since we do not employ different biological replicates in our study. Messenger RNA levels of seven genes were significantly (p < 0.05 with Student's t-test and Welch test) different between the two species, whilst four genes were found to be differentially expressed only with the Wilcoxon test (p < 0.05). This emphasizes that the Wilcoxon test is more conservative than the parametric tests. In contrast to the two parametric tests that assume a specific parametric signal distribution (Gaussian distribution) for the underlying signal series, the Wilcoxon test is non-parametric. Thus, P-values calculated by this test are valid in a more general set-up, i.e. for larger classes of probability distributions, than with the other tests. Furthermore, since the Wilcoxon test is based on ranks rather than on the underlying signals it is a more robust procedure in the sense that it is less sensitive against outliers from the model assumptions. These four genes are, ZNF278 , APOARGC , KIAA1609 and MGC12904 – highlighted as diamond shapes in Figure 2A . The expression levels of the vast majority (98%) of the genes remained unchanged. Furthermore, taking into account corrections for multiple testing, no gene is differentially expressed at the global experiment significance level of 0.05. For example, the lowest P-value of the Welch test is 1.98e-03, ( APOARGC ). Thus, Holm's setp-wise correction would start with an adjusted experiment level P-value of p = 0.05/349 = 1.43e-04. Similarly, measuring the false discovery rate by qvalues [ 17 ] results in no significant differential gene expression. Thus, we conclude that the level of expression of the individual 349 genes under investigation within human and bovine brain is roughly the same. Nucleotide sequence alignments of bovine and human transcripts In order to characterise some of the vast number of unknown bovine ESTs within the various databases we screened for orthologs to the human genes [ 18 ]. The 349 known human genes were screened against the TIGR Bos Taurus gene index (BtGI Release 8.0). Using high-quality matches (>85% identity, >100 bp overlap, E-value < 1.0e-15) we were able to ascertain the expression of 316 orthologous genes (see Additional file 1 ). Figure 4 shows the quality of the matches. Of these genes, 16 had sequence identities of greater than 95%, 137 genes with identities between 90% and 95%, 120 genes with identities of 85% to 90%. The remaining genes did not meet the criterion for the assignment as orthologs [ 18 ]. Forty genes had identities of 80% to 85%, 3 genes had identities between 75% and 80% and finally 33 genes did not have a significant BLAST hit. These matches are considered to be insignificant and therefore implying that these 33 human genes (assigned values (0) in Additional file 1 ) do not as yet have their bovine homologs present in the current bovine databases. Figure 4 Nucleotide sequence alignments of bovine and human transcripts All 349 human genes were matched against the TIGR Bos taurus gene index, BtGI Release 8.0, which contains 87,257 unique sequences. (A) Histogram of %-identities with the best match. (B) Histogram of bp overlap with the best match. Employing a comparative genomics approach we have functionally annotated previously uncharacterised bovine ESTs. These genes are depicted in the Supplemental Table as bovine ESTs lacking a gene name or description in the TIGR Bos Taurus gene index. As an example, the gene SLC11A3 which encodes a protein which functions as a solute carrier has 92% nucleotide sequence identity with an overlap of 500 bp with its bovine orthologue. Additionally, we have demonstrated co-amplification of this transcript in both human and bovine fetal brain RNA using primers derived from the bovine sequence – Figure 5A and Table 1 . Figure 5 Independent verification of array results by semi-quantitative RT-PCR (A) Ethidium bromide stained gel illustrating independent verification of differential gene expression by semi-quantitative RT-PCR. PCR reactions were carried out in 50 μl volumes, loaded in the order H- (human), B- (bovine) and VE- (water as negative control) then resolved on a 2 % agarose gel as described in Materials and Methods. Each panel consist of genes of the same functional category – listed on the left hand side. A 100 bp ladder (Invitrogen) was used to confirm the sizes (Table 1) of the PCR products. (B) A graphical illustration of the comparison of the expression levels of the orthologous genes as deduced by the microarray analysis and semiquantitative RT-PCR. Verification of expression levels of cross-hybridised orthologous genes by semi-quantitative RT-PCR Semi-quantitative RT-PCR was used to confirm the deduced expression data generated by the high stringency cross-species hybridisation. We selected a set of genes belonging to distinct families based on their published functional annotation, for example, cell cycle ( CDC27 , CDC6 and PCNA ), solute transport- ( SLC11A3 ), protein assembly- ( CALNEXIN ), anion channel- ( VDAC2 ) and as an endogenous reference, the housekeeping gene, β- ACTIN . Primer sequences were designed to co-amplify the orthologous genes (Table 1 ). The RT-PCR analyses (Figure 5A ) using a common set of gene-specific primers clearly demonstrate co-amplifcation of the orthologous transcripts and, in addition, differences in expression levels between genes within the same species are discernable. For example, the house keeping gene β- ACTIN and CALNEXIN which is involved in protein assembly are more abundant than for examples genes involved in cell cycle control. Most importantly, the trends in the deduced ratios (see Additional file 1 and Figure 5B ) of expression levels of orthologous genes from the hybridisation analysis are confirmed by the RT-PCR assay. Discrepancies in the bovine-human ratio with array data and the semi-quantitative RT-PCR data are also observable. Furthermore, some of the genes (for example CDC6 and VDAC2 ) are expressed at a low level (expression class 1, compare Figure 2C ). Here, microarray measurements are less reliable than PCR-based measurements and one comes close to the borderline of the detection limit of microarrays. Discussion Customised and focussed microarrays containing orthologous genes related to function, tissue, or pathways are becoming widely adopted for studying mRNA expression patterns. In addition, the level of cross hybridisation between genes with high sequence identity is also of interest because arrays are not always available for mammalian species other than human and rodents so cross-species hybridisations are often carried out [ 19 , 20 ]. In view of this, it is crucial to know whether the hybridisation conditions (i.e. the stringency used for the hybridisation and subsequent washes) would enable identification of altered gene expression across species. Here we have demonstrated cross-hybridisation of orthologous transcripts by adopting a high stringency hybridisation and wash protocol. The overall correlation co-efficient of gene expression in the fetal brains of human and bovine was 0.94, with the regression line at a slope of 1.13 (Figure 2A ). This suggests that the 349 genes under investigation have rather similar expression levels as judged by the intensity values of the human and bovine genes across the replicated experiments. The use of replicate experiments is clearly essential in order to ascertain true expression change. For example, judging single experiments we find a couple of genes escaping the 2-fold bounds. The MA-plot [ 14 ] shown in Figure 2B suggests that of the 349 genes, 5 are over-expressed more than 2-fold in bovine and 16 are more than 2-fold over-expressed in human. However, most of the changes are due to experimental variability of this specific experiment. An extremely important aspect of cross-species hybridisations is data reproducibility. A poor hybridisation experiment would lead to a high variability in the respective replicate experiments in human and bovine. With regards to this aspect, we identified fifteen genes (4%) in human and twenty genes (5.7%) in bovine with variability in expression levels in the four hybridisation experiments. In addition, approximately 305 genes have a CV of less than 0.5 with human brain (87%) compared to 280 with bovine brain (80%). However, this level of variability is low and can be expected due to the fact that the probe set used are of human origin. Moreover, the slight decrease in reproducibility due to species and nucleotide sequence differences can be compensated for by increasing the number of independent repetitions (biological and technical replicates) of the experiments. The unexpectedly low level of variability in gene expression levels between human and bovine fetal brain can also be attributed to the 80 data points examined per gene and the four replicate hybridisations (technical replicates) carried out for each species. Our finding further emphasises the need for sufficient technical and biological replicates in all microarray experiments. In assessing differential gene expression in bovine and human fetal brain, we observed that the expression levels of the individual orthologous genes were roughly in the same broad range. For example, if one considers genes involved in pathways of cell cycle control, the microarray deduced ratios (bovine:human) were for CDC27 : 0.95, CDC6 : 1.36, PCNA : 1.07, respectively, whereas the deduced ratios derived from the semi-quantitative RT-PCR assay were 0.59 for CDC27 , 0.35 for CDC6 and 0.87 for PCNA , respectively. Similarly the comparative ratios for SLC11A3 and CALNEXIN are 0.65 vs. 0.98 and 0.88 vs. 0.97. An illustration of co-amplification with common primers and also the confirmation of the array data for a selection of genes can be seen in Figures 5A and 5B . The expression ratios of all the other genes under investigation are given (see Additional file 1 ). Though co-amplification is possible by semi-quantitative RT-PCR, there is amplification bias related to primer mismatch (See Table 1 ). This effect was further highlighted when primer efficiencies were calculated using comparative Real-Time PCR (data not included). An important difference between the two techniques is that cDNA array hybridisation is less sensitive to minor mismatches than PCR primer annealing. The bias in amplification specificity is not a drawback in this study and species specific primers would obviously be used to compare gene expression levels in an experiment in which a developmental/metabolic pathway or disease model is investigated. The similar expression ratio of the orthologous genes implicated in cell cycle control and initiation of eukaryotic genome replication which is conserved in all eukaryotes highlights the feasibility and importance of our study in examining conserved pathways operative within the fetal brains of human and bovine and of course across other species using human cDNA microarrays. A similar finding has also been confirmed when comparing yeast co-regulated genes against the archaeal and bacterial operons. This implies that the components of the protein translation process are conserved across organisms at the expression level with minor specific differences [ 21 ]. We also identified four significantly differentially expressed genes as judged by the Wilcoxon test. The Wilcoxon test is known to be more conservative than the Student's t-test and the Welch-test, however, we have more confidence in this test since it is distribution-free, in particular it does not depend on the partly unrealistic assumption of an underlying Gaussian distribution. The genes ZNF278 and APOARGC were 1.45 – and 1.66 -fold over-expressed in human whereas KIAA1609 and MGC12904 were 1.60 – and 1.37-fold over-expressed in bovine. ZNF278 encodes a zinc finger-containing transcription factor that acts as a transcriptional repressor and also implicated in small round cell tumours [ 22 ]. The gene APOARGC encodes a protein with hydrolase activity. MGC12904 and KIAA1609 are uncharacterised ESTs so comments cannot be made with regard to their function. Orthologous genes between human and mouse and between human and rat both have a mean of approximately 85% sequence identity [ 18 , 23 ]. In two independent and unrelated studies carried out on cDNA and 50-mer oligonucleotide microarrays, cross-hybridisation was only observed with genes with 70%-80% and 50%-75% overall sequence identity, respectively [ 24 , 25 ]. With respect to these studies, our data unequivocally confirm the feasibility and reproducibility of cross-species hybridisation of orthologous genes within defined developmental and metabolic pathway(s) operative in human and bovine fetal brains. In addition, we have been able to assign gene names to previously uncharacterised bovine ESTs, thus, highlighting the importance of comparative genomics in identifying orthologous genes across species. Furthermore, using this protocol of cross-species hybridisation, we have compared gene expression in bovine unfertilised oocytes and blastocyst using a larger microarray (The Human Ensembl Chip) comprising 15,500 fully sequenced and annotated genes and ESTs. The correlation co-efficient between the RNA samples is 0.26, thus reflecting the diversity between oocyte derived maternal transcripts and embryonic transcripts derived from the blastocyst (Adjaye et al., unpublished). Conclusions In summary, this study highlights the significance and utilisation power of comparative genomics and also demonstrates the feasibility of using human cDNA microarrays to facilitate the identification of differentially expressed genes in human and bovine fetal brain. Our results indicate that cross-species hybridisation is not only a useful short-term solution for studying species for which gene maps, cDNA or oligo microarrays are not yet available, but also possesses tremendous power in enabling the unravelling of common evolutionary evolved mechanisms in different species. Methods Microarray fabrication For the generation of probes for spotting, cDNA inserts from a single 384-well plate of non-redundant, fully sequenced, annotated human cDNA collection (Human Ensembl set RZPD1.1) were employed- (BenKahla et al., unpublished) and a second partial plate consisting of control genes and empty wells to be used for data normalisation were amplified by PCR in a 384-well format. Bacterial clones were transferred to the PCR plates using 384-well replicators (Genetix Ltd, New Hampshire, UK). PCR amplifications were carried out in a total volume of 25 μl consisting of 1x PCR buffer, 1.5 M Betaine (Sigma, Germany), 1U Taq polymerase, 200 mM of each dNTP (Invitek, Germany), 25 pmoles of M13-forward (5'GTAAAACGACGGCCA3') and M13-reverse (5'CAGGAAACAGCTATGAC3') primers respectively. Cycling parameters typically consisted of an initial denaturation step at 95°C for 5 min followed by 35 cycles of denaturation at 95°C for 1 min, annealing at 55°C for 1 min, elongation at 72°C for 1 min, then a final elongation step at 72°C for 10 min. All amplified products were analysed by agarose gel electrophoresis and in all cases single bands were obtained per gene. Purification was carried out using isopropanol precipitations. In brief, 24 μl of the PCR product was transferred into 384-well Genetix plates (Genetix Ltd, New Hampshire, UK) previously filled with 18 μl of isopropanol per well. The samples were mixed by gentle vortexing after sealing the plates with transparent tape. After precipitation overnight at -20°C, the plates were spun at 3,500 rpm for 2 hrs at 21°C, the isopropanol decanted off and the plates dried briefly in a SpeedVac without heating. The pellets were resuspended in 6 μl of spotting solution (3x SSC /1.5 M Betaine (N, N, N-trimethylglycine; Sigma, Germany)). Random sampling of cDNAs revealed concentrations ranging between 200 to 300 ng/μl. Finally the samples in each well were spotted twenty times (for signal averaging) with four slit pins arranged in a 2x 2y printhead (Chipmaker 2 pins, Telechem, Sunnyvale, CA, USA). Prior to every spotting run the performance of all components of the robot is carefully tested – in particular the performance of the washing station in test runs designed to detect possible sample carry over. A modified Genetix Q-array (Genetix Ltd, New Hampshire, UK), controlled by a novel in-house software was used for the arraying of the samples on SuperAmine™ aminosilane-coated microscope slides (Telechem, Sunnyvale, Ca, USA). Post-processing of the arrays was performed using 1,2-dichloroethane and the acylating catalyst N-methylimidazole following protocols described previously [ 26 ]. Isolation of total RNA Human brain RNA was isolated from a medically terminated fetus at 10 weeks gestation. This was provided by the MRC-funded Human Embryonic Tissue Bank maintained at the Institute of Child Health and University College London, England. Bovine fetal brain was obtained from a 3–4 months old bovine fetus at a local abattoir near Mariensee, Germany. Immediately after slaughter of the pregnant female, the tissue was plunged into liquid nitrogen and transported to the laboratory. Brain tissues were homogenised using a Dounce homogeniser, total RNA isolated using TRIzol reagent (Invitrogen) and further purified using phenol/chloroform extractions and precipitation with ethanol after DNase 1 (Promega) treatment. All procedures were as described by the manufacturers. RNA Purity, integrity and concentrations were evaluated on the Agilent 2100 Bioanalyzer. High quality RNAs with A260/A280 ratio over 1.8 with intact ribosomal 28S and 18S RNA bands were utilised for subsequent labelling reactions. Direct labelling of RNA and hybridisations Four independent labelling (including dye-swaps) reactions per species were carried out using 25 μg total RNA from both human and bovine per labelling reaction. Direct incorporation of Cy3 and Cy5 during reverse transcription was carried out in a 20 μl reaction volume using 1 μg of anchored oligo-dT primer. The RNA/primer mix was incubated at 70 ° C for 5 min., left at room temperature for 10 min and then cooled on ice for 2 min. The following reagents were then added: 4 μl first-strand buffer, 2 μl 0.1 M DTT, 0.5 μl dNTP mix (25 mM each of dATP, dGTP, dCTP and 10 mM dTTP), 1 μl of 1.0 mM Cy5- or Cy3-dUTP (Amersham Pharmacia) and 1 μl (200 U/μl) Superscript II (Invitrogen). The labelling reaction was carried out at 42 ° C for 1.5 hrs. The reaction was stopped with 4 μl of 0.5 M EDTA. The input RNA was hydrolysed by the addition of 2 μl of 2.5 M NaOH and incubated at 37 ° C for 15 mins. followed by neutralization with 10 μl HEPES free acid (2 M, pH 5.5). Labelled cDNAs (four replicates per species which included dye-swaps) were purified from unincorporated Cy-dyes using Microcon YM-30 purification columns (Millipore). All labelled cDNAs were routinely analysed on a Fuji scanner (FL8-8000) and by ethidium bromide staining to ascertain dye incorporation and size-range of synthesized cDNAs. After concentrating cDNAs by evaporation in a SpeedVac, labelled targets were resuspended in 20 μl of hybridisation buffer (10 μg polydA and 20 μg Human Cot1 DNA,-Invitrogen; DIG-Easy Hybridisation mix -Roche). After thorough resuspension, the cDNA was denatured by heating at 95 ° C for 5 mins followed by 20 mins. at 42 ° C to enable annealing of the blocking reagents to repetitive sequences within the target cDNAs. The hybridisation mixture consisting of Cy3-labelled bovine and Cy5-labelled human RNA and vice versa was placed on the blocked array under a 24 × 40 mm coverslip (Menzel-Glaser, Germany). To maintain humidity inside the chamber, 20 μl of 3x SSC was added to the reservoir wells. The chamber was then tightly sealed and slides incubated at 42 ° C for 18 hrs in a waterbath. Slides were washed twice in 0.2x SSC / 0.1% SDS and then twice in 0.2x SSC. Washes were carried out at room temperature with 10 min durations per wash. Finally, the slides were dried by centrifugation at 1100 rpm for 10 min. Image acquisition and data analysis Fluorescence images were captured using an Affymetrix 428 scanner (Affymetrix, Santa Clara, CA) with appropriate gains on the photomultiplier tube (PMT) to obtain the highest intensity without saturation. A 16 bit TIFF image was generated for each channel for subsequent image analyses. Image analysis was carried out by placing the centre of each spot manually (grid-finding step) using the software AIDA (Raytest-Germany) and then by quantifying in a pre-defined neighbourhood around this spot centre using a two-dimensional Gaussian distribution (quantification step). Data analysis comprised two distinct parts. In the first part, data were normalised to eliminate extrinsic influencing factors and artefacts not attributable to the probe-target interaction. For each tissue (human and bovine brain) under analysis the whole batch of experimental replicates was normalized simultaneously. Normalisation should eliminate multiplicative technical bias between the different experiments and result in the same median signal level for each experiment. In a first step, the local background of each spot was subtracted from the spot's signal intensity. Then, for each experiment j the median signal, med j , was computed and a multiplicative factor was calculated according to med ref /med j , where med ref is the global derived from the average intensity values of the cDNAs across the full batch of experiments. The multiplicative factor was used to adjust the signal of the i th cDNA in the j th experiment, x ij , by x ij * med ref /med j . In the second part, a number of numerical characteristics were calculated in order to quantify the cross-species comparison. These characteristics were signal detection value, signal reproducibility and statistical significance of differential expression. Signal detection was judged by a number of empty positions that were spread across the array. For each experiment, the proportion of signals of empty positions lower than the actual spot signal was calculated. Then across all repetitions, the average proportion was kept as the signal detection value. Employing this procedure, a signal strength was quantified for each cDNA. Signal reproducibility was judged by calculating the co-efficient of variation (CV) for each cDNA across all experiments. In order to judge differential expression of a gene in human and bovine brain we calculated three statistical tests: Student's t-test, Welch-test and Wilcoxon's rank sum test for each cDNA based on the signal series derived from human and bovine experimental repetitions. The three tests make different assumptions on the signal series. Whereas the first two assume that the series are Gaussian distributed, the Wilcoxon test is distribution-free. Low P-values calculated by these tests indicate significant differences in signal intensity within the human and bovine samples [ 15 ]. Semi-quantitative RT-PCR analysis The list of primers, annealing temperatures and genes under investigation is shown in Table 1 . For primer annealing and reverse transcription, 1.0 μl (2.5μg/μl) of DNase1 treated human and bovine total RNA was added to 2.0 μl (50 μM) Oligo-dT primer plus 9.0 μl of RNAse free water. The mixture was spun briefly and heated to 70 ° C for 5 mins and cooled on ice. Thereafter, the following components were added sequentially, 4.0 μl of 5x RT. Buffer (Invitrogen), 2.0 μl of 0.1 M DTT (Invitrogen) 1.0 μl of (10 mM) dNTP (Amersham) and 1.0 μl (200 U/μl) Superscript II (Invitrogen). After pulse spinning, incubation was carried out at 42 ° C for 1.5 hrs. Gene-specific PCR amplifications were carried out in a total volume of 50 μl consisting of 5 μl of 10x PCR buffer, 0.2 μl (5 U/μl) Taq Polymerase, 1.0 μl (10 mM) dNTP, 2.5 μl of each (20 μM) primer, 1.5 μl (50 mM) MgCl 2 , 2.0 μl of the first strand cDNA (equivalent to 125 ng of input RNA) and distilled water to 50 μl. All reagents were purchased from Invitrogen. Cycling parameters consisted of an initial denaturation step at 95°C for 5 min followed by 30 cycles of denaturation at 95°C for 30 sec, annealing for 30 sec, elongation at 72°C for 30 sec, then a final elongation step at 72°C for 5 min. Primer sequences, annealing temperatures and predicted size of cDNA products are shown in Table 1 . The amplification reaction was carried out in a PTC 200 PCR machine (MJ Research). After PCR amplification, 50 μl of the reaction products were resolved on 2.0 percent agarose gel containing 0.2 μg/ml ethidium bromide. The gel was placed on a U.V. transilluminator (UVP Model TFM-20, Ultra-Violet Products Ltd., Cambridge, U.K.) with 25-watt UV tubes for high fluorescence and high sensitivity on stained gels), and imaged with a Photometrics Quantix 1401E, 12-bit cooled CCD camera (Roper Scientific GmbH, Ottobrunn, Germany). Densitometric analysis of the digitized image was performed with the IPLab-Gel program from Scanalytics (Fairfax, VA, U.S.A.). The semi-quantitative RT-PCR assay provides sensitive and reliable results, (26). The linear range of amplification was determined by product quantification after different cycle numbers. Identical amounts of bovine and human mRNA were amplified with the same number of PCR cycles and were always placed next to each other on the gel. Each RT-PCR was repeated at least three times for each gene. The housekeeping gene β- ACTIN was used as endogenous control. Authors' contribution J.A. conceived the study's idea, designed and optimised the protocols, carried out the hybridisations and was pivotal in developing the analysis plan and writing the manuscript. R.H. developed the analysis plan, analysed the data and was pivotal in writing the manuscript. D.H and M.N designed the cross-species primers and carried out the semi-quantitative RT-PCR. W.W. performed the initial image analysis. A.BK. selected the clones used as probes for creating the Chip and carried out the assignment of orthologous genes. T.B. assisted in the semi-quantitative RT-PCR analysis J.W.C. was involved in the conceptualisation and writing C.H. was instrumental in the Chip design and production. H.N. is the Head of the Department of Biotechnology at the Institute for Animal Science and was involved in the conceptualisation and writing. H.L. is the Head of the Department of Vertebrate Genomics at the Max Planck Institute for Molecular Genetics. Supplementary Material Additional File 1 Summary of the expression data and the corresponding gene annotations . The terms used under each column are explained viz; IMAGE ID: Gene specific identifier assigned by the IMAGE consortium Gene: HUGO gene name Chromosome: Chromosomal location of gene Description: Gene description Molecular Function: Assigned function by the Gene Ontology Consortium BG-tag-bovine: Signal detection probability in the bovine samples BG-tag-human: Signal detection probability in the human samples CV-bovine: Co-efficient of variation between replicate signal intensities CV-human: Co-efficient of variation between replicate signal intensities Mean-bovine: Mean signal across all hybridisation experiments using bovine RNA log-Mean-bovine: Logarithm (base 2) of the mean value Mean-human: Mean signal across all hybridisation experiments using human RNA log-Mean-human: Logarithm (base 2) of the mean value Ratio(bovine vs human): Ratio of the mean values log2(Ratio): Log-ratio (base 2) of the mean values P-value-Student's-t-test: P-value of Student's t-test using the replicate experiments P-value-Welch-test: P-value of the Welch test using the replicate experiments P-value-Wilcoxon-test: P-value of Wilcoxon's test using the replicate experiments Match: A BLAST hit with E-value < 1.0e-15 was found (1) or not (0) Homology: Classification of sequence homology between human gene sequence and bovine ESTs. "ortholog" = best match in both directions, "paralog" = bovine EST has its best match with another human sequence. TIGR-BTGI0-51503- Matches: Best results of the BLAST matches using the Ensembl- annotated gene sequence with the TIGR- database Identity: %-identity of best match Overlap: Overlap in base pairs Click here for file	/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC535340.xml
549062	Microarray-based resequencing of multiple Bacillus anthracis isolates	Custom-designed resequencing arrays were used to generate 3.1 Mb of genomic sequence from a panel of 56 Bacillus anthracis strains. Sequence quality was shown to be very high by replication and by comparison to independently generated shotgun sequence	Background Population genomics, the study of genome-wide patterns of genetic variation in a large number of organisms, is emerging as a vigorous new field of study [ 1 - 3 ]. Rapid, accurate and inexpensive resequencing could enable a variety of potential applications and studies. For the biowarfare (BW) pathogen, Bacillus anthracis , genomic sequences from multiple strains and non-pathogenic close relatives could aid studies that definitively identify B. anthracis in environmental and clinical samples, determine forensic attribution and phylogenetic relationships of strains, and uncover the genetic basis of phenotypic variation in traits such as mammalian virulence. Moreover, first recognizing the presence of a novel pathogen, and then attempting the difficult task of discerning between novel naturally occurring pathogenic organisms (for instance Bacillus cereus G9241 [ 4 ]) and artificially enhanced bacterial pathogens, requires a thorough knowledge of extant patterns and levels of genetic variation in natural populations. Unusual patterns of genetic variation may serve as evidence aiding the detection of these unusual types of pathogens. The current technological model for genome sequencing employs high-throughput shotgun sequencing at large centers. This highly successful enterprise has completed about 200 bacterial genomes with more than 500 ongoing as of July 2004 [ 5 ]. The genome sequences of the B. anthracis Ames chromosome (5.2 Mb, NC_003997) and plasmids pXO1 (181.6 kilobases (kb), NC_001496) and pXO2 (96.2 kb, NC_002146) have been determined [ 6 - 8 ], as have the genomes of three near neighbors, B. cereus ATCC 14579 [ 9 ], B. cereus ATCC 10987 [ 10 ] and B. cereus G9241 [ 4 ]. A strain of B. anthracis Ames strain isolated from a victim of the autumn 2001 bioterror attack in Florida was also sequenced to a high level of coverage using the random shotgun method and compared to the Ames sequence to identify 60 new markers that included single nucleotide polymorphisms (SNPs), inserted or deleted sequences, and tandem repeats [ 11 ]. The success of this effort has led to an extensive phylogeny-based whole-genome shotgun resequencing effort in B. anthracis (reported by [ 12 ]). Whole-genome shotgun studies are increasingly being used to explore variation among more closely related bacterial strains [ 13 - 15 ]. However, the relatively high costs of these efforts have limited the extent of their application. Numerous molecular methods for genotyping B. anthracis and near neighbors of the Bacillus cereus sensu lato group [ 16 ] have been developed and successfully employed in a wide variety of studies. These include DNA sequence surveys from one or a few number of loci [ 17 - 21 ], repetitive element polymorphism-PCR [ 22 , 23 ] and amplified fragment length polymorphisms (AFLP) [ 24 - 27 ]. However, because of the relative paucity of genetic variation between isolates [ 28 ], the most effective method for subtyping B. anthracis has employed multiple locus variable number of tandem repeats analysis (MLVA) [ 29 - 31 ]. Similar to the mammalian short tandem repeat methodology, MLVA determines strain phylogenetic relationships based on a relatively few, highly variable genomic repeat regions. While being relatively rapid and inexpensive, a key limitiation of MLVA lies in its exclusive focus on loci with common alleles that are differentiated by size. Because of the relatively rapid mutational process generating variation at these loci, similarly sized markers may have different evolutionary origins. Clearly, a method for rapid, inexpensive genome resequencing of bacterial strains would be of great benefit for genotyping, forensics and studies of the genetic basis of strain phenotypic variation. Developing DNA-based biodetection assays depends upon prior knowledge of patterns of genetic variation within and between bacterial species. It would be ideal to enable technologies that could combine the high information content of whole-genome resequencing of strains while also being rapid and inexpensive like MLVA, AFLP and multi-locus sequence typing (MLST). Furthermore, while conventional strain typing methodologies have focused on the utility of common variants, rare variants may prove to be especially informative for forensic applications. High-density oligonucleotide resequencing microarrays are a highly parallel technology that can enable the rapid identification of DNA sequence variants with minimal laboratory effort and infrastructure [ 32 , 33 ]. Previous applications of microarrays on bacterial genomes [ 34 , 35 ] or small eukaryotic genomes like yeast [ 36 , 37 ], focused on methods that scanned specific genes or a genomic region for genetic variants. Initial high-throughput microarray applications in the human genome for SNP discovery [ 38 - 40 ] were successful, but also reported that between 12% and 45% of the detected variants were false. Subsequent experimental improvements and the development of the ABACUS algorithm/software package [ 32 ] significantly reduced SNP false-positive ascertainment, radically improved genotype calling and automatically assigned quality scores to each genotype call. These fundamental advances enabled rapid resequencing of 40 human genomic regions [ 32 , 41 ] and ABACUS is now the standard application for microarray-based resequencing. Here we present the first microarray-based high-throughput resequencing of a large collection of B. anthracis isolates. Our study first reaffirms, and then directly demonstrates that the quality of microarray-generated DNA sequence data is directly comparable to that produced by conventional shotgun sequencing. We then estimate the levels of genetic variation in the annotated genomic regions we resequenced, characterize the frequency spectrum of DNA sequence variants we observe, and finally explore patterns of linkage disequilibrium and recombination among those variants. Because of the scalability and minimal effort associated with microarray-based resequencing, our work demonstrates the possibility of a rapid and cost-effective method of genome resequencing that could be applied to both environmental, and ultimately clinical specimens. Results Resequencing B. anthracis with microarrays A panel of 56 B. anthracis strains from the Biological Defense Research Directorate's strain collection (see Additional data file 1) was resequenced using Affymetrix resequencing arrays (RAs) and base calls determined using the ABACUS software package [ 32 ]. Each RA was capable of resequencing 29,212 base-pairs (bp) or about 0.5% of the B. anthracis genome from a single isolate sample (see Additional data file 2). Long PCR sample preparation and chip processing was conducted for 118 RAs. Analysis of these 118 RAs with the ABACUS software package shows that 115 are successful (97.5%). Experimental failure occurs when less than 60% of the total possible bases fail to achieve quality scores exceeding the ABACUS user-defined threshold. For this study, the total threshold was set at 31 and a strand minimum of -2 [ 32 ], as determined from analysis of a replication experiment described below. The 115 successful RAs call 92.6% of the possible bases (3,109,539 bp out of a total possible of 3,359,380 bp). Figure 1 shows the distribution of quality scores across all 3,359,380 base calls. Amplicon failure, typically arising from long PCR (LPCR) failure, accounts for 1.1% of the uncalled bases. The remaining base-calling failure (6.3%) consists of features on the RAs that fail to generate quality scores exceeding the experimental threshold. Previous results demonstrated that base-calling failure was concentrated among RA oligonucleotide probes containing multiple purines. Purine-rich probes were observed to have lower hybridization intensities at identical positions across multiple RAs. Guanine-rich probes, in particular, showed the greatest reduction in hybridization intensity (see Figure 6 in [ 32 ]). Consequently, total quality scores at these sites frequently failed to exceed the quality-score threshold and they remained uncalled. To determine if probe sequence composition, specifically purine and guanine content, contributed to the 6.3% of bases not called, the sequence composition of the purine-rich oligonucleotide probes at 4,209 sites successfully called on all 115 RAs (484,035 total sites) was compared to that at the 886 sites that failed to be called on any RA (101,890 total sites). These failed sites account for 3.0% of the total base calling failure in the experiment. Uncalled sites are composed of oligonucleotide probes with a significantly higher purine composition ( P < 10 -22 ). A similar pattern is detected if we limit our analysis to guanine-rich probes ( P < 10 -9 ). This latter result is surprising given that the B. anthracis genomic sequences we examined have a low G+C content (~34%). Nevertheless, these analyses demonstrate that both purine-rich and guanine-rich oligonucleotide probes are significantly more likely to fail to generate quality scores exceeding the experimental threshold. Assessing microarray resequencing data quality Building on the recognition of the importance of automated algorithms to assess data quality [ 42 , 43 ], we used two methods to assess the quality of microarray resequencing data [ 32 ]. The first consisted of a replicate experiment where 51 samples were independently hybridized on 102 RAs. A parameter search that optimized the percentage of called bases, while minimizing the number of discrepancies between replicates was then performed. A total of 1,489,812 bases could have been called in each replicate experiment. At the optimal parameter values (total threshold of 31, strand minimum of -2 see Cutler et al. [ 32 ]), 90.6% (1,349,178) of sites are called in both replicates. Other parameter values provide similar levels of base calling and discrepancy rates. The optimal parameter values are similar to those previously used by Cutler et al. [ 32 ]. Of the bases called in both replicates, 1,349,177 are called identically. Only one site is called differently. This corresponds to a replication discrepancy rate of 7.4 × 10 -7 (Table 1 ). If repeatability could be related to accuracy, then this level of repeatability would correspond to a phred score of at least 61 [ 42 , 43 ]. This calculation assumes that the discrepancy rate corresponds to a binomial error probability of P , where phred = -10 log 10 P . These replication levels and discrepancy rates are consistent with those previously reported [ 32 ], providing further evidence for the ability of RAs analyzed with ABACUS to produce highly replicable data. While RA data is highly replicable, repeated systematic errors would not be detected in a replicate experiment. To obtain an independent estimate of RA sequence accuracy, we compared the sequence data from 30 RAs where the same B. anthracis strain had been sequenced using the random shotgun approach and deposited in GenBank ( B. anthracis : strain Ames, NC_003997 [ 8 ], Vollum, NZ_AAEP00000000, 4 June 2004 update, strain Australia 94, NZ_AAES00000000, 7 June 2004 update, strain Kruger B NZ_AAEQ00000000, 7 June 2004 update (J Ravel, DA Rasko, MF Shumway, L Jiang, RZ Cer, NB Federova, M Wilson, S Stanley, S Decker, TD Read, et al. , unpublished work). In a comparison of 398,467 bp of RA- and shotgun-generated sequence, we observed 15 discrepancies occurring at six sites. This corresponds to a discrepancy rate of 3.8 × 10 -5 . If we make the conservative assumption that all discrepancies lie in the RA-generated sequence, this level of accuracy would correspond to a phred score of at least 44. To determine if this conservative assumption is warranted, we examined in greater detail the nature of the RA/shotgun sequence discrepancies. Five of the discrepant sites, accounting for 10 discrepancies total (twofold RA replication at each site), were found in Kruger B strain sequences. The one remaining site, accounting for five discrepancies (fivefold RA replication at this site), was found in Vollum strain sequences. At all 15 discrepancies, the RA called a base identical to the Ames reference sequence [ 8 ], while the Kruger B/Vollum shotgun sequence called a new SNP. The fact that the shotgun sequence called a SNP at every discrepancy was surprising, leading us to examine more closely the level of shotgun coverage and assembly at each discrepant site. A comparison of the latest shotgun assembly of the Kruger B strain (J Ravel, et al. , unpublished work) with the RA Kruger B strain base calls agreed with the RA base calls. The latest Vollum shotgun assembly (J Ravel, et al. , unpublished work) still disagreed at the one site (five discrepancies total), but this discrepancy was based on a single shotgun sequencing read with a phred score of 7 at the discrepant base. Clearly, the shotgun coverage lacks sufficient depth at this site to make a reliable base call and it seems far more likely that the fivefold RA base call is correct. Hence, the RA sequence data has less than one discrepancy per 398,467 bases called, or a discrepancy rate of < 2.5 × 10 -6 (Table 1 ). This observed level of sequencing accuracy corresponds to a phred score of 56. These data demonstrate that our conservative assumption is not warranted. Resequencing array data quality from a single experiment matches, and in some cases perhaps exceeds, that obtained by multiple DNA sequencing reads using conventional DNA sequencing technologies [ 42 , 43 ]. Patterns and levels of genetic variation in B. anthracis We identify 37 SNPs among 56 B. anthracis strains. The SNP location, base-call, and position relative to the respective GenBank reference sequences [ 6 - 8 ] are contained in Additional data file 3. Twenty-four of the 37 SNPs, including two singletons, were independently confirmed in identical strains where whole-genome random shotgun sequence was available (A0039, A4088 and A0442 in Additional data file 1 (J Ravel, et al. , unpublished work)). Of the remaining 13 SNPs not independently verified by The Institute of Genomic Research (TIGR), 11 were seen only once in our collection of strains and two SNPs were seen three times. Population genetic inference typically assumes that study samples are selected without prior knowledge of their patterns of genetic variation. For this study, we selected diverse strains from widely distant geographic regions in an attempt to sample the full extent of genetic variation in B. anthracis . The number of SNPs identified, the amount of sequence generated and the nucleotide diversity [ 44 ] of the 56 strains is contained in Table 2 . We performed analyses for sequences comprising the total dataset, for each genomic region separately, and for the total dataset with each resequenced base assigned into an annotated SNP class. We report three main findings. First, the total average level of DNA sequence variation in B. anthracis is very low. This finding is in agreement with previous studies [ 11 , 28 ]. This level of genetic variation is much lower than that seen in commonly studied bacterial species [ 14 ], roughly half of that observed in the human genome and 25-fold lower than that observed in D. melanogaster [ 38 , 39 , 45 - 48 ]. Second, the B. anthracis chromosome appears less variable than either the pXO1 or pXO2 plasmids, although this difference is not statistically significant. Third, the patterns of genetic variation by SNP class (see Table 2 and Additional data file 4) are similar to that seen in other well studied bacterial [ 14 ] and eukaryotic genomes [ 45 ]. Silent sites, those sites that when mutated do not alter the protein primary structure, are significantly more variable than are amino acid altering replacement sites ( P = 0.0011). Intergenic regions are observed to have intermediate levels of genetic variation, whereas replacement sites, those sites that when mutated alter the protein primary structure, are the least variable. Replacement sites are marginally significantly less variable than intergenic sites ( P = 0.039) whereas silent sites are not significantly more variable than intergenic sites ( P = 0.22). The neutral theory of molecular evolution predicts a characteristic frequency spectrum of SNPs, or segregating sites, for populations at equilibrium [ 49 ]. Deviations from this expected distribution are observed when an experimental population sample contains an excess of low frequency, rare SNPs, or an excess of high frequency, common SNPs, relative to the neutral expectation. These deviations can arise as a consequence of demographic history and/or the action of natural selection [ 50 ]. Figure 2 compares the observed and expected percent of SNPs in four allele-frequency classes. The data suggest an observed excess of rare SNPs as compared to that expected under the neutral theory. For example, while the neutral theory predicts that approximately 60% of SNPs should have minor allele frequencies less than or equal to 0.25, we observe that more that 92% of the B. anthracis SNPs we discovered have minor allele frequencies that fall into this class, a statistically significant difference (Figure 2 ). We used the Tajima's D statistic [ 50 ] to further assess this pattern for the entire dataset, for SNPs from each genomic region and for each SNP class (Table 2 ). Tajima's D is a summary statistic for the site (or SNP) frequency spectrum, whose value is negative when there is an excess of rare variants and positive when there is an excess of common variants, relative to the neutral expectation. The test statistic is calculated from two different estimates of levels of genetic variation, the number of segregating sites [ 44 ] and the average number of nucleotide differences estimated from pairwise comparisons [ 50 ]. We observe that Tajima's D is negative for SNPs comprising the total dataset, each genomic region and each SNP class. While none of the individual test statistics is statistically significant, they collectively suggest an excess of rare variants in B. anthracis . If we scale our variation estimates drawn from the 0.5% resequenced in 56 B. anthracis genomes, we can estimate a range around the total number of SNPs that one would detect upon sequencing two random B. anthracis isolates, sampled in the same fashion as isolates in this study were chosen. Our results indicate that we should expect to find, on average, between 944 (standard deviation (SD) 454) [ 50 ] and 1,586 SNPs (SD 762) [ 44 ]. A substantial proportion of these SNPs, probably more than expected under the neutral theory, would be rare. Using multiple sequence alignments of 17 genes from B. anthracis (NC_003997, Ames) and B. cereus (NC_004722, ATCC 14579 [ 9 ] and NC_003909, ATCC 10987 [ 10 ]) the patterns of genetic polymorphism and divergence at silent and replacement sites was assessed. The raw counts are presented in Table 3 . It is striking that two B. cereus strains exhibit more polymorphism at silent and replacement sites than divergence from B. anthracis . This result confirms, at the DNA sequence level, previous results suggesting that the B. cereus species group is diverse and polyphyletic in origin. B. anthracis then appears to be a clonal lineage derived from, and nested within, a diverse species. In other words, the species names do not encompass or reflect the evolutionary history of the species [ 10 , 51 , 52 ]. No evidence for recombination in B. anthracis chromosome The 37 SNPs discovered on the B. anthracis chromosome and plasmids pXO1 and pXO2, possess in total, 636 pairs of sites where two alleles are observed. In principle, the alleles at each pair of sites could form four distinct haplotypes. Plasmid transfer between different B. anthracis strains would affect physically unlinked site pairs resulting in four distinct haplotypes. Homologous recombination or gene conversion between physically linked site pairs is also expected to produce all four haplotypes. The straightforward counting of the number of haplotypes that one detects in a large population sample, such as the one used in this study, is often referred to as the four-gamete test [ 53 ]. Among the 636 site pairs in our sample, we observe 26 pairs of sites with two haplotypes, 610 pairs of sites with three haplotypes, and no pairs of sites with four haplotypes. This striking result implies that the value of D', the standardized measure of linkage disequilibrium (LD) [ 54 ], is equal to 1, its maximum value, for all site pairs that we observe. Among the 137 site pairs where we could have detected statistically significant LD at P < 10 -3 , we observe that 52 site pairs exhibit statistically significant LD. Four of the six site pairs showing significant LD on the B. anthracis main chromosome are over 500 kb apart. Correlation of RA resequencing data with MLVA typing Because of the low level of genetic variation in B. anthracis ([ 28 , 29 ] and this study), determining the phylogenetic relationship among B. anthracis strains has proven difficult. Twenty-four B. anthracis strains characterized with a single fluorescent AFLP primer combination were reported to be monomorphic [ 27 ]. One recent MLST study sequenced seven housekeeping genes (approximately 3 kb total) in 5 B. anthracis strains and reported that the strains were monomorphic at the sites examined. Another recent MLST study sequenced seven genes (approximately 3 kb total) in 11 diverse B. anthracis strains finding three polymorphic nucleotides [ 55 ]. Neither the AFLP nor the MLST studies discover and genotype sufficient genetic variation to distinguish between B. anthracis strains. The most successful marker-based approach used to date, MLVA, determined the genotypes at eight VNTR loci in 426 B. anthracis isolates, enabling the construction of a phylogenetic tree of B. anthracis strains [ 29 ]. We sought to determine if our resequencing of 0.5% of each of 56 B. anthracis genomes is capable of confirming the major phylogenetic groupings determined by MLVA. To test this, we concatenated the 37 variant positions for all strains in this study, calculated a distance matrix using a simple Kimura substitution model, and generated an Unweighted Pair Group Method Arithmetic Mean (UPGMA) tree (see methods [ 56 ]; Figure 3 ). The strains group together in a manner broadly similar to that found by Keim et al. [ 29 ] with B strains forming an outgroup and most A strains being found together in the same subgroups (Figure 3 ). There are exceptions: one group in Figure 3 contains a mix of A3a, A1a, A1b and A2 strains. This anomaly is probably due to the relatively few SNPs that effectively distinguish these groups when only 0.5% of the genome is sampled. All B. anthracis Ames strains but ASC394 correctly cluster in an A3b group. B. anthracis ASC394 may be a case of an originally mistyped or mislabeled strain. Nevertheless, our data suggest that limited, random resequencing of 0.5% of the 56 B. anthracis genomes discovers and genotypes sufficient genetic variation to determine the major phylogenetic relationships among B. anthracis strains. Discussion Population genomics requires the random sampling of genome-wide patterns of DNA sequence variation in a large number of organisms. Such studies require high-throughput, highly accurate, cost-effective resequencing technologies. While the conventional industrial-scale shotgun-sequencing model is clearly the best technology available for de novo generation of genomic sequence, it may not be the best approach for resequencing large numbers of strains. RAs, as originally applied for human genome resequencing [ 32 ], offer one competing technology that can rapidly produce very high-quality data with limited personnel and infrastructure requirements. Our application of RAs to resequence multiple genomic regions in the biowarfare pathogen, Bacillus anthracis , further supports this perspective. Studies of DNA sequence variation are most informative when both rare and common variants are identified. While the limited ascertainment of selected common variants can be employed to identify broad evolutionary relationships among bacterial genomes, and in fact underlies most bacterial strain typing methodologies, the ultimate forensic application of resequencing lies in the ascertainment of rare, presumably newly arising variants, that may allow more precise determination of a strain's origin. Rare variants may be particularly informative since they are likely to be restricted to specific strains (substrains/isolates). Strain genotyping of common variants provides an incomplete description of genomic patterns of DNA sequence variation, while obtaining most or all of the genomic sequence from multiple strains allows a maximally informative analyses of DNA sequence variation, its function, and ultimately, the evolutionary history of the organisms. The ability to rapidly, accurately and inexpensively resequence entire bacterial genomes should also contribute to an understanding of a variety of important phenotypic traits in B. anthracis and other bacterial pathogens [ 57 - 62 ]. Our study demonstrates that microarray-based resequencing is technologically robust and generates highly replicable and accurate data when compared to alternative sequence technologies (Table 1 ). In this experiment, 115 RAs, or 97.5% of the total attempted, were processed successfully obtaining an average high-quality base-calling rate of 92.6%. Called bases are shown to be highly replicable (discrepancy rate of 7.4 × 10 -7 ) and accurate when compared to conventional shotgun sequence (discrepancy rate of < 2.5 × 10 -6 ). Clearly, RA-generated resequencing data from a single experiment is comparable, in terms of data quality, to DNA sequence generated from multiple shotgun reads by a DNA sequencing center. The major technical challenge facing RA-based resequencing is to increase overall call rates while not compromising data quality. Modifications of RA synthesis, experimental protocols and the ABACUS software algorithm could all contribute to improved base-calling rates. While it is possible to increase call rates while sacrificing data quality, there is a need to focus on generating very high-quality data at virtually all sites. If this is absent, the second-best outcome is to call all bases in an environment in which we understand the nature of probable errors. In diverse fields where RAs might be widely used as a first-stage screening tool, such as BW agent identification or human clinical testing, the imperative is to use highly sensitive technologies that minimize the false-negative rate. False-positive findings could be confirmed later in a second-stage screen with an alternative technology such as conventional dideoxy chain termination sequencing. Microarray-based resequencing identifies and genotypes SNPs in a single experiment. No prior knowledge of the variability of a site is required - only a reference genomic sequence. Microarray design and applications are flexible. It is, however, important to note that the use of RAs in this study is not as a SNP typing technology. Thus, problems in interpreting the inferred phylogenetic relationships between strains that arise from SNP typing schema are avoided [ 63 ]. RA-based resequencing resembles MLST methodology used for bacterial strains [ 52 , 55 , 64 ]. MLST attempts to choose the most informative genomic regions to resequence, largely because of the costs associated and technological limitations in generating enough DNA sequence data on a large collection of variant strains. While a typical MLST approach might resequence between 3 and 4 kb, in organisms like B. anthracis that have low levels of genetic variation ([ 28 , 51 , 55 ] and this study), this amount of generated sequence is insufficient. Clearly, RAs, such as those used in this study that can resequence approximately 29 kb, could rapidly increase this amount and be used for MLST studies. Furthermore, manufacturing improvements that reduce RA feature sizes enable the resequencing of greater quantities of genomic sequence per microarray. Ongoing work at NMRC/BDRD is evaluating RAs that can resequence 300 kb per chip. At that RA feature density, when combined with whole-genome amplification protocols, a single technician in two days could resequence the entire B. anthracis genome on approximately 15 RAs. Our data provides the first population genetic estimation of the levels and patterns of DNA sequence variation in B. anthracis . We report three main findings. First, among B. anthracis isolates sampled in the same fashion as in this study we would expect two randomly selected B. anthracis strains to differ, on average, at between 944 (SD 454) and 1,586 SNPs (SD 762). The variance surrounding these expectations is large, and any two isolates may differ from the expectation. Closely related, nonrandomly sampled isolates, such as those sequenced in [ 11 ], will have far fewer SNPs than that expected for samples drawn from a worldwide collection. Nevertheless, our data suggest that were it possible to rapidly resequence entire B. anthracis genomes, sufficient genetic variation is likely to be found to make very fine-level discrimination of strain collections. Resequencing offers the best chance to identify newly arising, rare, strain-specific variants that will discriminate between very closely related strains, since we expect identical genotypes at the known common genetic variants [ 11 ]. We also observe, that as seen in eukaryotic genomes [ 45 ], the amount of silent variation per site within genes is much higher than that seen at replacement sites. Intergenic regions are seen to have intermediate levels of polymorphism. This pattern is expected to arise if noncoding intergenic regions possess variants visible to natural selection. If SNPs in intergenic regions were purely neutral, then we would expect to see levels of polymorphism similar to that at silent sites, which are undoubtedly under less stringent selective forces. Second, the neutral theory of molecular evolution predicts that in a population at equilibrium, a significant proportion of the observed genetic variation will consist of rare genetic variants [ 49 ]. We observe a significant excess of rare SNPs as compared to that expected under the neutral theory (Table 2 ). This pattern of variation classically has at least two possible causes. The first consists of a recent population expansion from a small founder population. The second consists of the action of natural selection on genetic variants [ 65 - 67 ]. Resequencing technologies will be of particular use in populations of organisms exhibiting this pattern of genetic variation. Finally, we see no evidence for plasmid exchange or recombination altering the patterns of DNA sequence variation among B. anthracis strains in the regions that we resequenced. Some of the regions that we resequenced contain genes whose function influences B. anthracis pathogenicity or surrounded the bacterial origin of replication. In other bacterial species, these types of regions are the most likely to exhibit recombination [ 14 ]. The fact that we observe no evidence of plasmid exchange or recombination among physically linked markers in the regions that we resequenced, is striking. The simplest interpretation of this observation is that the B. anthracis strains that we examined are ultimately derived from a single clonal ancestor and that the exchange of plasmids and recombination between strains during the course of their evolution is either very rare or nonexistent. While models of natural selection could also account for the patterns that we see, we think a simple demographic model of recent, rapid clonal expansion is parsimonious and best supported by our data. Hence, our findings suggest that B. anthracis populations consist of multiple closely related clones whose life histories prevent the opportunity for homologous recombination between different strains. We note, however, that while we resequenced 0.5% of the B. anthracis genome, including regions where we expected to detect recombination, further data collection from multiple genomic regions, or the entire genome, would allow a more thorough analysis of this pattern. Sequencing a larger percentage of the genome in a similar-sized or larger sample of isolates would provide greater power to detect rare recombination events. We are undertaking such a project to test the validity of our inference and to better determine if recombination is rare or absent among B. anthracis strains. The absence of recombination in B. anthracis , a potential biowarfare agent, suggests a novel approach to identifying a newly arising or a genetically engineered strain. A recombination event could arise through rare natural genetic exchange or as a consequence of genetic engineering. Irrespective of the cause, the discovery of a B. anthracis strain possessing evidence of genetic recombination would warrant close examination and probably demand immediate further phenotypic and genomic characterization. Taken together, the findings of a low number of differences between strains, a preponderance of rare variants, and an absence of recombination all point to a scenario where the current world population of B. anthracis has expanded recently from a single clone derived from, and nested within a diverse species, B. cereus . Other bacterial pathogens, such as the potential biowarfare agent Yersinia pestis , possess a similar recent pattern of rapid expansion [ 15 ]. However, the patterns of genetic variation in Y. pestis are quite different from that seen in B. anthracis , for instance in the much more active role of insertion sequences in Yersinia . We speculate that the B. anthracis history of clonal expansion could arise as a consequence of the life history of a highly pathogenic sporulating mammalian pathogen. Exploring the population biology of less virulent members of the B. cereus group could directly test this. These population genomics studies could determine if clonal clusters of B. cereus strains exhibit similar population dynamics and patterns of genetic variation, or whether the picture of B. anthracis emerging from studies such as this is as unusual as the level of pathogenicity of the species itself. Conclusions Microarray-based resequencing can rapidly generate very high quality data, enabling population genomics studies in bacteria. We find no evidence for plasmid exchange or recombination altering the patterns of DNA sequence variation among B. anthracis strains in the regions that we resequenced The patterns of genetic variation in the B. anthracis regions resequenced are consistent with that expected for a bacterial species that has undergone a rapid, historically recent expansion from a single clone. Detecting plasmid exchange or recombination between B. anthracis genetic variants could act as an indicator of a newly emerging or genetically engineered strain. Materials and methods B. anthracis strains surveyed We selected a geographically diverse panel of 56 B. anthracis strains from the Biological Defense Research Directorate collection (see Additional data file 1). Twenty-four of the strains originated from the Louisiana State University collection [ 29 ]. These have been typed by MLVA [ 29 ] and in order to sample diversity, we chose a group that had representatives of the A1a, A1b, A2, A3a, A3b, A3d, A4, B1 and B2 lineages. The remaining 35 strains originate from a UK collection and were chosen to represent geographical variation as well as unusual phenotypes such as gamma phage and penicillin resistance. Six of the UK strains were reisolates of the Ames strain [ 11 ], which allowed us to test the reproducibility of resequencing. Resequencing array design Unique genomic sequences were identified using Miropeats [ 68 ] at the default thresholds from among the B. anthracis Ames chromosome (5.2 megabase-pair (Mb), NC_003997) and plasmids pXO1 (181.6 kb, NC_001496) and pXO2 (96.2 kb, NC_002146). The genomic regions that we resequenced included at least one gene of interest (pXO1: toxin lethal factor precursor lef , toxin moiety, protective antigen pagA ; pXO2: encapsulation protein gene CapC ; Ames chromosome: vrrA , DNA-directed RNA polymerase rpoB , yfhp protein), but also included many surrounding loci (see Additional data file 4 for complete listing). The total chip design consisted of 6,191 bp from pXO1, 6,725 bp from pXO2, and 16,584 bp from the Ames chromosome (total submitted sequence 29,500 bp). From these unique sequences, a single 20 × 25 μm RA design capable of resequencing 29,212 bp or 0.5% of the B. anthracis genome was fabricated by Affymetrix (see Additional data file 3). The final sequences submitted for RA design are contained in Additional data file 5. B. anthracis strain genomic DNA isolation Five milliliters of brain heart infusion (BHI) was inoculated and grown 12-16 h at 37°C. One-ml aliquots of cells were centrifuged for 10 min at 5,000-7,500 g . Pellets were resuspended in 720 μl enzymatic lysis buffer (20 mM Tris-Cl pH 8.0, 2 mM EDTA, 1.2% Triton X-100, 20 mg/ml lysozyme) and incubated at 37°C for 1 h. After incubation 100 ml of Proteinase K was added along with 800 ml of Qiagen buffer AL, and incubated at 70°C for an additional 30 min. Then, 800 ml of 100% ethanol was added and this was split onto four of the Qiagen DNAeasy tissue kit. The DNA was then washed and eluted according to the Qiagen protocol. After the DNA was eluted, it was passed through a 0.22 mm filter. Sterility was confirmed by plating 10% DNA preparation directly on SBA plates with a second 10% inoculated into a 5 ml broth culture. The plate and the broth were allowed to incubate for 7 days. Two hundred milliliters of the broth culture was subcultured onto SBA at day 4. If there was no growth on any of these cultures the DNA was considered sterile and removed from the BSL-3 lab for subsequence analyses. Sample preparation and RA hybridization Genomic DNA was amplified using Long PCR (LPCR) protocols described in Cutler et al . [ 32 ]. The primers that amplified each RA fragment are shown in Additional data file 3. The primer sequences were: ant8 AAAAAGACGAGATGCGTCAACATCCCGTCCCA, ant9 TCAACTAAATCCGCACCTAGGGTTGCTGTAAG, ant10 ATTACTTTGAGTGGTCCCGTCTTTATCCCCCT, ant11 ACATTAGCAGGCAAGGACAGTGGTGTTGGAGA, ant14 ATTCACGCTCTCCCACCCAGATATTCCTACAT, ant15 GTCCTAATATCGGTGAGCAACGCAGGGTAGTT, ant20 GAGAAGAACCCCTACTACACGCATTGATACTG, ant21 TTTAGTAGCGAGGGTACAGGCGCGTTTATACC, ant26 TGGAAGCAGGCTTCGTAAGTGTAGGCGACGTT, ant27 GTTGCATGTTCGCTCCCATAAGTGCGCGGTTA, ant 32 AATGGGTGTATAGGGGTGATCTGTTGTGATGG, ant33 TCCATGTTCGGCCATCTGATTCCGTCACTACT. Long PCR product concentration was determined by using Pico-Green (Molecular Probes, Inc.) with lambda DNA standards (Invitrogen). The LPCR products were then pooled, DNAse digested, biotin endlabelled and hybridized to individual RAs overnight following established protocols contained in [ 32 ]. Subsequent washes and stains were carried out as described in Cutler et al. [ 32 ] and were only washed and not antibody stained. RAs were scanned at 570 nm, with a pixel size of 3 μm/pixel averaged over two scans. Automated grid alignment and base calling was performed for the .DAT files on a Mac G5 computer with the ABACUS software suite. RA sequence determination An ABACUS parameter search was employed to determine those parameters that called the maximal number of bases while minimizing discrepancies [ 32 ]. This total experiment consisted of 118 RAs, of which three failed (< 60% base calling). Of the remaining 115 RAs, 8 were used to sequence individual strains once. Of the remaining 107 RAs, 96 were used to replicate hybridize 48 B. anthracis strains, while the remaining 11 RAs were used as additional multiple replicates of these same strains. In total, sequence data was generated from 56 unique B. anthracis strains (see Additional data file 1 for strain listing). In order to obtain the most complete data possible, for those strains with replicate RA sequences, a single composite strain sequence was generated for subsequent population genetic analyses. The current version of ABACUS algorithm is not designed to detect insertion/deletion variation. The effect of oligonucleotide probe composition was determined by choosing for each base, the probe with the most purines or the most guanines. The number of times that a given base was called was tabulated across all 115 successful RAs. The mean purine and guanine composition was determined for the classes that were called in all 115 RAs and uncalled in all 115 RAs. A Student's t test with unequal variances was used to test for difference in mean sequence composition (purines/guanines) between the always called and never called classes. The DNA sequence files for the 115 RAs and the original RA image files (.DAT files) are available from the authors and will be made available through the NCBI Trace Archive. Population genetic analyses All population genetic analyses were calculated using the popgen_fasta2.0.c code (Cutler DJ, unpublished work) on the collection of 56 sets of B. anthracis fasta files. The fasta files were analyzed in total and separately for the main chromosome and plasmids pXO1 and pXO2. The identification of genes was taken from publicly available annotation contained in the relevant GenBank refseq files ( B. anthracis str. Ames NC_003997; pXO1, NC_001496; pXO2 NC_002146). The statistical significance of linkage disequilibrium between site pairs was performed by using the Fisher's Exact Test at P < 10 -3 [ 69 ]. Estimating levels of genetic variation To account for missing data, θ is estimated by [Σ n (S n /a n )]/L, where S n is the number of observed segregating sites at positions with exactly n alleles sequenced (n is a maximum of 56, fewer with missing data), a n = Σ i = 1.. n -1 1/i, and L is the total length of the sequence examined. Var{θ} is estimated by [Σ n (L n θ/a n + (L n ) 2 b n θ 2 /(a n ) 2 ]/L 2 , where L n is the number of sites with data from exactly n alleles, and b n = Σ i = 1.. n -1 1/i 2 . With missing data π is estimated by [Σ i 2p i q i n i /(n i - 1)]/L, where the sum is taken over all sites i, p i and q i are the allele frequencies at site i, and n i is the number of alleles sequenced at site i. To determine if the estimates of theta between SNP types (silent, replacement, intergenic) are significantly different, we used the number of samples sequenced, the number of segregating sites, and the length of the region to find a maximum-likelihood estimate of theta per site for each SNP type using equations 11 and 12 in Hudson [ 70 ]. We compared all possible SNP types against each other (silent vs replacement, silent vs intergenic, replacement vs. intergenic). For a given pair of SNP types, we first determined the maximum-likelihood estimator of theta for each type individually. We then determined the maximum-likelihood estimator of theta, assuming both types had identical theta per site. We ask whether the model with different thetas for each type fits significantly better than the model with a single theta through a likelihood ratio test. Reported significances are the p -values from the likelihood ratio test. Site frequency spectrum Comparing the observed site frequency spectrum with that expected under the neutral theory is a powerful approach to detect unusual patterns of genetic diversity. We employed two different approaches for this analysis. First, we calculated the expected number of sites with minor allele frequency i as Σ n θL n [1/i + 1/(n - i)] and from this determine the expected percent of sites under the neutral expectation. This is directly compared with the observed percent of SNPs in Figure 2 . Confidence intervals for the sample proportion of each SNP minor allele frequency classes as where N is the number of SNPs observed for each class, is their observed frequency, and . As a second method, we employed Tajima's D statistic [ 50 ], estimated as (π - θ)/Var(π - θ). Under the neutral model, π and θ have the same expectation, hence Tajima's D is expected to be 0. Since π is a function of site heterozygosities and θ is a function of the total number of segregating sites, Tajima's D is negative (positive) with an excess (deficit) of rare sites. We use our estimated values of π [ 51 ] and θ [ 45 ], multiplied by the total genome B. anthracis genome length (5,505,178), to determine the expected number of SNPs that we would expect to observe among two B. anthracis strains sampled in same random fashion as isolates in this study were chosen. Using Equations 6-9 in [ 51 ], we calculated the variance of and θ estimators. The one standard deviation (SD) that we report is the square root of this variance. Phylogenetic tree inference The 37 variable positions identified in this study were concatenated together to create artificial sequence types. A DNA distance matrix was created using DNADIST, plotted as a UPGMA tree using NEIGHBOR and the tree plotted using DRAWGRAM [ 57 ]. Additional data files The following additional data are available with the online version of this article. Additional data file 1 lists B. anthracis strains from the Biological Defense Research Directorate (BDRD) strain collection resequenced in this study. Additional data file 2 lists the BDRD-01 RA fragment names, the GenBank reference sequence from which they are derived, the length of the unique genomic sequences submitted to RA design, the length of the unique genomic sequences capable of being queried, and the LPCR primer pairs used to amplify the RA fragments. Additional data file 3 lists the B. anthracis SNPs identified in this study. The data include the BDRD SNP ID, the GenBank reference sequence and RA fragment containing the SNP, the SNP position relative to the GenBank reference sequence and the RA sequence, the SNP frequency, and the listing of the base calls in all strains at sites harboring SNPs. Additional data file 4 lists the 31 B. anthracis genes partially or wholly resequenced in this study. The observed number SNPs by SNP type (silent vs replacement) for each gene are provided. Finally, Additional data file 5 shows the genomic sequences submitted to RA design for BDRD-01. Supplementary Material Additional data file 1 B. anthracis strains from the Biological Defense Research Directorate (BDRD) strain collection resequenced in this study Click here for additional data file Additional data file 2 The BDRD-01 RA fragment names, the GenBank reference sequence from which they are derived, the length of the unique genomic sequences submitted to RA design, the length of the unique genomic sequences capable of being queried, and the LPCR primer pairs used to amplify the RA fragments Click here for additional data file Additional data file 3 The B. anthracis SNPs identified in this study. The data include the BDRD SNP ID, the GenBank reference sequence and RA fragment containing the SNP, the SNP position relative to the GenBank reference sequence and the RA sequence, the SNP frequency, and the listing of the base calls in all strains at sites harboring SNPs Click here for additional data file Additional data file 4 The 31 B. anthracis genes partially or wholly resequenced in this study Click here for additional data file Additional data file 5 The genomic sequences submitted to RA design for BDRD-01 Click here for additional data file	/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC549062.xml
554782	Atlas – a data warehouse for integrative bioinformatics	Background We present a biological data warehouse called Atlas that locally stores and integrates biological sequences, molecular interactions, homology information, functional annotations of genes, and biological ontologies. The goal of the system is to provide data, as well as a software infrastructure for bioinformatics research and development. Description The Atlas system is based on relational data models that we developed for each of the source data types. Data stored within these relational models are managed through Structured Query Language (SQL) calls that are implemented in a set of Application Programming Interfaces (APIs). The APIs include three languages: C++, Java, and Perl. The methods in these API libraries are used to construct a set of loader applications, which parse and load the source datasets into the Atlas database, and a set of toolbox applications which facilitate data retrieval. Atlas stores and integrates local instances of GenBank, RefSeq, UniProt, Human Protein Reference Database (HPRD), Biomolecular Interaction Network Database (BIND), Database of Interacting Proteins (DIP), Molecular Interactions Database (MINT), IntAct, NCBI Taxonomy, Gene Ontology (GO), Online Mendelian Inheritance in Man (OMIM), LocusLink, Entrez Gene and HomoloGene. The retrieval APIs and toolbox applications are critical components that offer end-users flexible, easy, integrated access to this data. We present use cases that use Atlas to integrate these sources for genome annotation, inference of molecular interactions across species, and gene-disease associations. Conclusion The Atlas biological data warehouse serves as data infrastructure for bioinformatics research and development. It forms the backbone of the research activities in our laboratory and facilitates the integration of disparate, heterogeneous biological sources of data enabling new scientific inferences. Atlas achieves integration of diverse data sets at two levels. First, Atlas stores data of similar types using common data models, enforcing the relationships between data types. Second, integration is achieved through a combination of APIs, ontology, and tools. The Atlas software is freely available under the GNU General Public License at:	Background One important goal in bioinformatics is to integrate data from disparate sources of heterogeneous biological information. Data integration allows us to assemble targeted data reagents for bioinformatics analyses, and to discover scientific relationships between data. Most public repositories of biological data focus on deriving and providing one particular type of data, be it biological sequences (e.g., GenBank [ 1 ], UniProt [ 2 ]), molecular interactions (The Biomolecular Interaction Network Database (BIND) [ 3 - 5 ], The Human Protein Reference Database (HPRD) [ 6 ]), or gene expression (The Stanford microarray database [ 7 ]). Integrating these disparate sources of data enables researchers to discover new associations between the data, or validate existing hypotheses. Several recent studies have demonstrated the power of integrative bioinformatics. Using data from genomic sequences and annotations, mRNA expression, and subcellular localization, Mootha et al were able to use bioinformatics approaches to identify one of the disease genes responsible for Leigh syndrome [ 8 ]. In another example of an integrative bioinformatics approach, Stuart et al used existing publicly available data to generate hypotheses about the functional roles of gene sets [ 9 ]. These two examples illustrate the potential of querying integrated public data to reveal novel relationships. However, working with publicly available biological data can be challenging due to the volume and complexity of the data types. With the proliferation of massive, publicly available data sets, researchers need a way to readily access this data. Querying distributed data has inherent limitations such as the server resource restrictions of the remote resource, concerns of secure data transmission over the internet, and of course the actual logistics of querying distributed resources. In such an environment, the distributed search space is difficult to process in a high-throughput way, and requires complex queries to tie together the heterogeneous data. Consequently, there is a need for a data integration solution that facilitates search and retrieval in an efficient, flexible, high-throughput manner. Several active solutions are available that attempt to integrate data and that provide the tools to retrieve that data. We have grouped these existing systems into three major categories, based on how the data is stored and integrated: full record, SQL-based, and distributed. Full record systems like SRS [ 10 ] and Entrez [ 11 ] store the intact record in a table and extract specific fields to index and cross-reference. SeqHound [ 12 ] is a powerful system that stores Entrez information (fully annotated sequence and structure information) locally and can be accessed programmatically through application programming interfaces APIs. Much like Entrez and SRS, fully intact records are stored in SeqHound, with specific fields indexed. The major advantages of SeqHound over Entrez is that it is locally installable and provides API access to the data. SeqHound highlights the power and utility of a locally installable warehouse. SQL-based systems implement relational models to store data. This allows SQL-level access to specific parts of the data model, enabling detailed queries on the data for greater specificity of results. The data in relational models are stored as primitive data types as opposed to storing fully intact records that need parsing or processing to access the parts therein. For example, sequences and their annotated biological features can be stored in their own fields in the database, permitting 'substring' operations to extract parts of the sequence that span a particular feature type using SQL. Systems like EnsMart [ 13 ] and DBGET/LinkDB [ 14 ] provide data in a relational form, such that the power of SQL is at users' disposal. EnsMart's relational back-end provides users with the ability to construct intricate queries on the data by taking advantage of SQL. Distributed systems implement software to access heterogeneous databases that are dispersed over the internet. JXP4BIGI [ 15 ] have created a generalized method to access, extract, transform, and integrate distributed data. The tool acts as a middle-ware for constructing a local instance of a data warehouse. This system is customizable, versatile and uses industry standard data modeling, distribution, and presentation software. BioMOBY [ 16 ] is a semantic-based system utilizing ontologies, and a services model to support user queries. TAMBIS [ 17 ], like BioMOBY, is also a semantic-based system, and is also service-model driven. These semantic web implementations do not house the data locally, but rather query the original data provider for available services before sending queries to that particular data provider. These systems are quite powerful for interrogating disparate data sources of information. However, a disadvantage is that large queries may take a long time to return or may not be returned at all due to server resource restrictions. As well, the level of data integration is only at the services level, rather than at a field-based level which can provide much better resolution for queries. Atlas is a versatile, flexible, and extensible data warehouse that provides a solution to these challenges. Our approach establishes common relational data models enabling the reuse of each class of data model to store all data of the same type. For example, a single interaction data model is used to store information from any of the interaction data sets such as BIND, MINT, EBI IntAct [ 18 ], Database of Interacting Proteins (DIP) [ 19 ], and HPRD. Instances of these data models, once populated by the source data, can then be interrogated using the developed retrieval APIs. These APIs encapsulate the SQL calls used for fine granular access to the data. Furthermore, ontological information stored in these databases captures the relationships between the many data types. Finally, tools are developed that capitalize on the API methods, to facilitate application specific demands of end-users, ranging from simple queries of specific data types, to complex queries that infer molecular interactions across species. Atlas then, is designed for use by a wide audience from biologist to software developer. Construction and content The Atlas system is made up of five main parts: 1) the source data, 2) the ontology system, 3) the relational data models, 4) the APIs, and 5) the applications (see Figure 1 ). The following sections outline the Atlas architecture in detail. Source data We categorize the Atlas data sources into four main groups: 'sequence', 'molecular interactions', 'gene related resources', and 'ontology' (Figure 1 ). Currently, the data sources that fall into these categories are: 'sequence', GenBank, RefSeq [ 11 ], and UniProt ; 'molecular interactions', HPRD, BIND, DIP, IntAct, and MINT; 'gene related resources', Online Mendelian Inheritance in Man (OMIM) [ 20 ], LocusLink [ 11 , 21 ], Entrez Gene [ 22 ], and HomoloGene [ 11 , 23 ]; and 'ontology', NCBI Taxonomy [ 11 , 24 ], and Gene Ontology [ 25 , 26 ]. Table 1 lists each of the sources of data incorporated into Atlas, and provides URLs where those sources can be found. Note that GenBank refers to the integrated records from the International Nucleotide Sequence Database Collaboration (GenBank [ 11 ], DDBJ [ 27 ], and EMBL [ 28 ]). Relational data models (schema design) This section describes the composition of the data models of the source data included in Atlas. The data models we present here are implemented in MySQL [ 29 ], an open source relational database management system (RDBMS). As such we only provide Data Definition Language (DDL) files that are compatible with MySQL. Currently there are no plans to port these to other RDBMS systems. Ontology Ontologies serve to define the concepts and relationships both within a system and between systems. This vocabulary of concepts and relationships is representative of a given expert domain of discourse such as sequences, gene annotations, and taxonomy. In Atlas, ontologies are categorized into two classes: Atlas defined ontologies and external ontologies. The Atlas defined ontologies are used to represent the concepts and relationships found specifically within Atlas, as well as to characterize concepts and relationships implicitly defined by the GenBank Sequence Feature data model. External ontologies include such things as NCBI Taxonomy for organism classification, Gene Ontology for gene annotations enabling categorization of biological features based on function, process, and cellular component, and the Proteomics Standards Initiative Molecular Interaction Standard (PSI-MI) controlled vocabulary [ 30 ]. The Atlas internal ontologies contain definitions of terms such as identifier types like accession numbers, GI numbers, PSI-MI terms and identifiers, PubMed identifiers, file format types like XML, relationship terms, and concepts like GenBank Sequence Features and Feature Qualifiers, Sequencing Techniques. This part of the Atlas ontology consists of three tables: Ontology which include terms and definitions, Ontology_type that defines ontology source and category, and Ontology_Ontology which stores term-term relationships. Foreign key constraints are used to ensure data integrity. In contrast to these tightly integrated ontologies, two other external vocabularies are instantiated as independent MySQL databases: GO and NCBI Taxonomy. These ontologies, unlike the others, do not implement foreign key enforcements to the other database modules. As a result, when ontology terms are updated, references to deleted terms deemed to be invalid are kept in the system until such time a full data set reload is performed. The Atlas internal ontology exists largely to help describe Sequence Features as they exists in the GenBank Sequence Feature model, as this is the primary data source for features. Neither the Open Biological Ontologies (OBO) [ 31 ] relationship terms, nor the Sequence Ontology (SO) [ 32 ] relationship terms suited our needs as a feature ontology. We utilize the basic relationships similarly found in OBO and SO, such as 'is-a', 'part-of', and 'inverse-of' but we also define more specific terms such as 'is-synonym-of', 'refers to PubMed', 'feature-includes-qualifier', and 'gene-contains-promoter'. By defining these specific relationships, we simplify the ontology tree into a flatter structure that is simple to query. In addition, subject-predicate-object triples are not explicitly defined in the internal ontology, but rather are assigned at loading-time as the GenBank Sequence Feature data is parsed and stored into the database. The relationship terms are not necessarily complete, but sufficient for our needs, and as new relationships are encountered, these are added accordingly. For example, we mapped all 66 GenBank feature keys to an entry in our Ontology table, which has enabled us to do feature-level queries for any type of feature in GenBank, or genomes we annotated in-house. We caution the reader that it is generally understood that not all GenBank features have the same informational value, nor quality of information. However, to capture the maximum amount of information, we chose to extract and store all annotated features. With the locations of the features stored in Atlas, sub-sequences of features can be extracted in a high-throughput manner using SQL, the APIs, or the toolbox applications. This is particularly useful, for example, in extracting features like non-coding RNAs from complete genomes, or regions spanning a particular gene of interest. We are actively integrating selected external ontologies, and expanding our internal ontologies. Plans for ontology integration include the National Library of Medicine (NLM) MeSH term and the Microarray Gene Expression Data (MGED) ontology [ 33 ]. We are evaluating the option of adopting frame-based ontology representations, and existing ontologies such as TAMBIS Ontology (TaO) [ 17 , 34 ]. In the near future, we will release the Atlas ontology in other formats such as GO flat file, RDF/XML, and OWL. A complete list of the ontologies is available on the Atlas website, and we provide the MySQL dumps for these: . Sequence model The schema for sequences is organized into three main parts: sequence , which stores the sequence string and associated meta-data such as sequencing technique and molecule type; sequence identifiers , for which all identifiers that appear in the records are stored (see Figure 2 ); and annotated sequence features , for which feature keys, qualifier keys and values and feature locations are stored. Though output of features into General Feature Format Version 2 (GFF2) [ 35 ] is supported, the Feature table, one will note, does not explicitly contain source or type fields. This information is stored in other tables and can be pulled together dynamically as a GFF2 record is being constructed. For example, the BioID_type table contains the database source information in its db_source field and the internal Atlas Ontology table's term field which represents the feature type used in the GFF2 output. However, to reflect the fact that features in such output are now reconstructed from the Atlas system, we prefix the original source type with 'Atlas:', such as in 'Atlas:GenBank/RefSeq'. The reader will note that there are two different Ontology tables in Atlas. A more detailed explanation for the motivation for having two different kinds of Ontology tables is described in the previous Ontology section. However, in the context of sequence features, it is the internal Atlas Ontology table that is relevant. The sequence string is stored in the Sequence table. Additional fields for: sequencing technique, tech, such as expressed sequence tags (ESTs); molecule type, moltype, such as DNA, RNA, protein, and nucleic acid; sequence length, length; the NCBI taxonomy identifier, taxonid; and the definition line, defline, are also stored in the Sequence table. Fields such as taxonid, tech, and moltype can be used separately, or in combination to produce customizable queries that return highly specific sets of data. Sequence identifiers, as with all other external identifiers, are managed through a layer of abstraction by associating them with internal identifiers within Atlas, which act as primary keys. Having a single internal identifier for a sequence allows us to relate all other identifiers found in the record to each other. In addition, ontologies for all types of identifiers currently found in GenBank ASN.1 data files, as well as relationships between these identifiers are modeled in the Bioid and Bioid_Bioid tables, respectively. As mentioned above, sequence features are also modeled in Atlas. For details please refer to the Ontology section below. Molecular interactions For molecular interaction data, we developed a relational model compliant with the PSI-MI. Adopting a common interaction data model allowed us to unify data from different sources, and allows us to develop a set of common interaction retrieval APIs. Currently, HPRD, BIND, DIP, IntAct and MINT are included as interaction data sources. BIND, DIP, MINT and IntAct release their data in PSI-MI format. HPRD is releasing data in both PSI-MI standard format, and their own XML format. At the time of this publication, BIND released data as indexed flat files, ASN.1, XML, and PSI-MI format (level 2). The Atlas interaction model consists of four major entities: Interactor, Interaction, Experiments and Dbxref. Interactor holds information about one of the interacting members in an interaction, such as the interactor's name, taxonomy, sequence, molecular type, features, subcellular localizations, and external identifiers. An Interaction consists of one or more interactors, and one or more experiments. Experiment stores information about the experiments used to identify interactions. Finally, Dbxref is used to crosslink the external identifiers such PubMed id, RefSeq accession, HPRD id, BIND id, and Ontology id, for example (see Figure 2 ). As an additional note, the Feature table in the Interaction database is mainly used to store protein features involved in the interactions. We will release a PSI-MI level 2 compliant version of the Interaction model, and API upon public release of the level 2 specification. Gene related resources We integrated OMIM, LocusLink, Entrez Gene, HomoloGene and the annotation part of GO into the Atlas system in order to provide gene-related information. Conveniently, the OMIM and LocusLink data sources provide flat file tables which could be imported directly with the MySQL import function. Entrez Gene will eventually replace LocusLink, however in order to maintain a smooth transition and backward compatibility, we are maintaining populated relational models for both Entrez Gene and LocusLink until LocusLink is officially retired. Integration between HomoloGene and Sequence is achieved by relating the taxonomic, protein sequence and gene identifiers with Atlas' Bioid table. This allows us to integrate these databases and provide linkage between, for example, orthologous genes present in different interaction scenarios (see Utility of the Atlas system). Application programming interfaces There are two classes of APIs in Atlas: loader and retrieval. Components of Atlas for which we have developed our own relational models, such as the Biological Sequences component, or the Molecular Interactions component, each have their own set of loader APIs. The loader APIs used to build the loading applications, populate instances of the relational models in the Atlas databases. Though most end-users will never need to use the loader APIs, they are critical to the implementation of the Atlas loading process, and are provided to the software development community. The other class of APIs are the retrieval APIs. These APIs serve to retrieve the data stored in Atlas. They are required for developing custom retrieval applications such as the Atlas toolbox applications. The loader API for Biological Sequences is implemented in C++ as it relies heavily on the NCBI C++ Toolkit [ 36 ] to parse the ASN.1 data. The Biological Sequence retrieval API, on the other hand, is provided in all three languages: C++, Java, and Perl. The Java and Perl APIs return sequences as BioJava SimpleSequence and BioPerl Bio::Seq objects, respectively. The loader and retrieval APIs for Molecular Interactions are provided in Java. Though retrieval APIs are not supported in all languages, further development in Perl and C++ will be added if our user community requests them. Please refer to Figure 1 for a mapping of data modules to currently supported programming languages. The project is also open source and other developers are encouraged to contribute. All the transactions between the APIs and the database are specified by the numerous SQL statements which are all defined within the majority of the API methods. Application programming interface architecture The API is constructed using object-oriented methodologies, employing objects to represent everything from low-level database connections to high-level data structures, and their access methods. This is illustrated in Figure 3 . Common in the design of the C++, Java, and Perl APIs, are a set of APIs written for MySQL database connectivity which handles the opening and closing of MySQL connections, as well as managing the execution of the SQL statements themselves. All subsequent APIs that interact with the Atlas database extend from this set of APIs. Both the data loader and the retrieval utilities share a common class responsible for low-level data transformations. This class includes methods that facilitate conversions between two internal Atlas identifiers, such as bioid_id to ontology_id, or methods that convert internal Atlas identifiers to externally referenced public identifiers, such as GenBank accession numbers, or GI numbers. Inheriting this shared identifier conversion class benefits both the loader APIs and the retrieval APIs, by providing them with the necessary tools to integrate information. The Biological Sequences component of Atlas manages common identifiers, and hash maps in the Seq class. This class is inherited by both the SeqLoad class and SeqGet class, which define the loader methods, and retrieval methods, respectively. Another feature of the Biological Sequences API, is its ability to control stream output based on molecule types. API users simply specify which molecule type to filter by, through calls to higher-level retrieval methods, and SeqGet will then handle the logistics of stream management. Similarly with the Molecular Interactions component of Atlas, the InteractionDb class is inherited by the InteractionLoad class and the InteractionGet class, respectively defining the loader and retrieval methods which manipulate the data in memory. Our Java interaction APIs, for example, are tightly coupled to our interaction data model with classes representing all the major schema objects such as Interaction, Feature, Dbxref, and Experiment. The APIs are works in progress and we continue to develop and improve them. We are considering even more tightly coupled API development by using XML schema code generators such as JAXB. All the source code is provided under the GNU General Public License (GPL), and therefore any developer can model future API development on numerous functions we have already implemented. Applications Toolbox The Atlas toolbox is a collection of applications that use the C++ API to perform common sequence and feature retrieval tasks. The applications are standard Unix command-line based tools that follow a command-line option-based interface for parameter entry. These are end-user applications and do not require any programming ability to use them. We have developed toolbox applications for sequence retrieval from accession and GI numbers, retrieval of sequences from all organisms beneath a given node in the NCBI taxonomy tree, retrieval of features given accession and GI numbers, retrieval of sub-sequences corresponding to specific features identified by qualifiers and their values, and retrieval of a set of interactions associated with a molecule given the accession number of an Interactor. Besides being useful tools, the toolbox applications' source code provides good examples of application development using the APIs. Software developers wishing to use the APIs can use these toolbox applications as a starting point for their own custom applications (see Table 2 ). Data loaders Data loaders are provided in Atlas to facilitate the parsing and loading of the source datasets into their respective Atlas database tables. Two main classes of loaders are currently supplied in the Atlas package: sequence loaders and interaction loaders. Though other types of data are loaded into Atlas, their loading is trivial as MySQL database dumps of these datasets are already provided by the data providers. The first class of loaders is the sequence-based loaders. Within this class there are two applications provided: seqloader and fastaloader. The seqloader performs the majority of the sequence loading from GenBank and RefSeq datasets. These datasets have long been represented as ASN.1 (binary/text) by the NCBI [ 37 ], and are compact and well defined for storing of structured data. The seqloader was built using the NCBI C++ Software Development Toolkit [ 36 ] which was designed to specifically parse the ASN.1 sequence data, extracting such things as the sequence, associated identifiers, features of the sequence and related publications. There are, however, instances where sequence data is missing from the ASN.1 records. In these situations, we obtain the missing records from the NCBI Entrez system in the form of Fasta records. The fastaloader application is then used to update the sequence field in Atlas with the sequences from the Fasta records. The second class of loaders is interaction-based loaders. These loaders are exclusively implemented in Java. The datasets loaded by this class of loaders include BIND, HPRD, MINT, IntAct and DIP. All the interaction loaders are designed to parse the data in the way that best deals with that particular source data's structure and content (mostly XML). The interaction data is loaded using a common interaction object model, and the interaction loading APIs provide a flexible and extensible framework for future interaction data loading efforts. Currently, we are developing a PSI-MI level 2 data loader. Besides these classes of loaders, there is also a Java based loader that parses and loads UniProt sequence data. In addition, scripts are used to load datasets for which MySQL dumps, or tab-delimited database dumps are provided. This is handled using the MySQL import function, and eliminates the need to devise special parsers and loaders. GenBank and RefSeq are checked daily for incremental updates from the NCBI. Accession numbers are used to maintain the integrity of the data. New accession numbers reflect new records and will be inserted into the database. Updated sequences or records with same root accession number and patched annotations will replace existing records in the database. When new releases of GenBank/RefSeq are made available, all databases are purged and reloaded to remove retired records and to maintain referential integrity. Web tools Though we encourage the use of Atlas as an in-house repository, it can also act to serve the wider internet community. We provide a publicly available web interface to the Atlas databases to demonstrate some of its functionality. This offers basic access to GenBank, RefSeq, NCBI Taxonomy, Atlas Ontologies, BIND, HPRD, MINT, IntAct and DIP. Web interfaces to the Atlas toolbox applications: ac2gi, ac2seq, ac2tax, feat2seq, gi2ac, gi2feat, gi2seq, gi2tax, tax2seq, techtax2seq, tech2seq are available. In addition, interacting partners for proteins identified by accession numbers or GI numbers can be retrieved from any of the four interaction databases stored in Atlas. These web tools can be found at: . Utility of the Atlas system The Atlas data warehouse offers maximum flexibility of data retrieval and integration. Users can access data in Atlas at the SQL, API and end-user application levels. Routine, pre-defined queries can be accessed through the APIs in Java, C++, and PERL (see API section, above), enabling developers to incorporate these queries in their software applications. Most of these queries have been used to build the Atlas toolbox, a set of end-user applications that run on the Unix command-line (Table 2 ). Included in the toolbox are common utilities for converting GenBank ASN.1 sequences to file formats supported by the NCBI Toolkit [ 1 ] such as XML, GenBank Flat File, and FASTA. In addition, information regarding features that are annotated on sequence records can be exported as General Feature Format Version 2 (GFF2). The recently developed General Feature Format Version 3 (GFF3) is not currently supported in Atlas, to allow its specification time to stabilize. However, its support in Atlas is planned in future releases. In the following sections, we illustrate use-cases of the system at the SQL, API and toolbox levels with specific biological themes in mind. Single record queries Single record queries are the simplest use case of the system. Users can input a GenBank or RefSeq accession number and/or GI number into the ac2seq and gi2seq toolbox applications to retrieve the relevant sequence record in Fasta, GenBank or ASN.1 format. Features on a particular sequence can also be retrieved independently with GenBank or RefSeq accession numbers and/or GI numbers. The single record queries can also be performed in batch mode where the user supplies a list of accession numbers or GI numbers and all data pertinent to the list of identifiers is then retrieved. Genome annotation Atlas provides tools to generate data reagents for genome analysis as well as a data model for storing biological features that have been annotated on the sequences. Coupled with Pegasys [ 38 ] and Apollo [ 39 ], the Atlas system is an essential part of our annotation platform (see Figure 4 ). Atlas functions simultaneously as a data reagent generator for sequence alignment analysis, a storage system for annotations that are to be submitted, and a data transformation tool that can convert Apollo-compatible data to NCBI submission tool compatible data. Atlas provides users with the ability to generate custom sets of data to use as reagents. For example, using tax2seq, users can input a specific node of the NCBI taxonomy tree using its scientific name, or its NCBI taxonomy id and retrieve all nucleotide and amino acid sequences from organisms in the tree rooted at that node. This has special utility in genome analysis where specific sets of data from close relatives of the genome of interest enable comparative genomic methods for functional annotation. Furthermore, this type of taxonomy querying can be combined with the 'tech' field in the NCBI data model to produce sequences derived from different sequencing techniques such as expressed sequence tags (EST), genome survey sequence (GSS), sequence tagged sites (STS), high throughput genomic (HTG), etc. Compiling these specific data sets allows the user to perform more directed sequence similarity searches, for example, that yield more specific hits. Using the sequence data structure to model existing annotations in sequence records, Atlas can be used to store additional annotations created in Sequin [ 40 ] and Apollo [ 39 ]. We have built a GAME XML [ 41 ] loader that stores annotations exported from Apollo. When used for this purpose, Atlas serves as a holding bay for sequences that can be submitted to DDBJ, EMBL, or GenBank in a relational form that can be mined in the interim using the multi-level query system provided by the Atlas APIs (see Figure 4 ). Additionally, the annotations stored from a GAME XML [ 41 ] file are exportable in GFF2, or Sequin Feature Table Format [ 42 ] for use with NCBI submission tools like tbl2asn [ 42 ]. Inference of protein-protein interactions Deriving new associations from the information extracted from Atlas has proven to be particularly useful in developing a prototype system that infers interactions across species, detailed in "Ulysses – an Application for the Projection of Molecular Interactions across Species" (Kemmer D: in preparation, from the Wasserman and Ouellette laboratories). Given that the data for protein-protein interactions found within model organisms can be extremely sparse, Ulysses employs homology information to help bridge the gaps in the interaction data by projecting known interactions in one species onto other species for which those interactions are not known, and subsequently inferring potentially novel interactions in those species. Ulysses is able to perform its analyses and inferences, in part, by capitalizing on the integration, offered by Atlas, of HPRD, BIND, and HomoloGene. Atlas makes it possible to retrieve interactions for one species known to occur in another species, by integrating these datasets under one query space, and by providing the API and tools which make such queries simple. As an example, in both the MINT and DIP databases, protein C-C chemokine receptor type 3 (SwissProt accession number P51677) was found to interact with protein Small inducible cytokine A24 precursor (SwissProt accession number O00175) in human (MINT interaction 14962; DIP interaction 10472E). Although referenced by different publications ([ 43 ], [ 44 ]), both interactions are likely to be the same. With corroborating evidence for these seemingly synonymous interactions, it can be claimed with more certainty that two proteins do indeed interact. Furthermore, homologs for both sequences can be found in mouse and rat through HomoloGene. Though these homologs are not found to be interacting partners in either mouse or rat, it is reasonable to speculate that such interactions exist in both these organisms. Disease-gene associations The Atlas system is also being used to determine yeast orthologs of genes that are implicated in human disease (Hieter P: in preparation). The inference being that human genes for which there are yeast orthologs represent essential genes which are candidates for human disease agents. Compiling the reagents for this custom database was straightforward using the Atlas tools. It takes advantage of the linkage between sequence identifiers, Taxonomy, HomoloGene, and OMIM. Discussion We have built a data warehouse of biological information with the goal of providing high-throughput, flexible access to the data by means of SQL queries, API-level queries, and end-user application-level queries. Our goal was to create a system that serves as a platform through which information from many sources of data can be interrogated, enabling biologists and computer scientists to easily carry out queries necessary for their research. The data warehouse facilitates complex queries on local instances of GenBank, RefSeq, UniProt, HPRD, BIND, NCBI Taxonomy, HomoloGene, Gene Ontology, OMIM, Entrez Gene, and LocusLink. With previously disparate data now unified in a relational model, SQL can be used to retrieve this consolidated information at once. Though Atlas can act to serve data publicly over the internet, its simple setup enables anyone or any institution to easily serve their own customized data warehouses to their own local users. Installing Atlas in-house to serve local users gives the data provider full control over the data they serve. Giving users access to the system on a high-bandwidth internal network offers convenience and high-performance for large queries, such as retrieving all human ESTs. Such data is then more readily retrieved with lower latency and higher bandwidth than attempting to retrieve the same data over the internet. One of the important strengths of the Atlas architecture is that it allows data integration at two levels. The first level uses a common data model to integrate similar types of data from different sources (e.g., GenBank or UniProt, and BIND or HPRD). The second level uses the APIs, ontologies, and tools to cross-reference disparate types of data. For example, consider the task of retrieving all amino acid sequences, and from all organisms found within the taxonomic tree rooted at a given taxonomic node (e.g., vertebrata ), from the RefSeq database. With a single call to the taxonName2Sequences method, the user can accomplish this task. Within these API methods are SQL statements which first retrieve the taxonid from the Taxonomy database. Then using a recursive method, the taxon identifiers for all organisms beneath that given taxon node, are returned. All amino acid sequences for each of these taxon identifiers are then retrieved using taxonId2Sequences (see API documentation [ 45 ] for more details). Uniting disparate sources of data is a useful exercise that highlights the challenges that the data itself presents. Any changes to the source data structure often requires software code changes in order to properly parse the new data format. Failure to do so often leads to the inability to load at least some of the information, if not all. Furthermore, the quality of the original data may often be imperfect as much of this data is curated manually, and hence is subject to data entry errors. Everything from missing data to improperly spelled key terms can impede the loading process. For this reason, it is essential to devise a system that is robust enough to handle unforeseeable exceptions. Policies on how to handle such exceptions are important to define and implement. We try to adhere to the careful logging of incorrect entries that we find during the loading process, and to promptly report these to the data providers for remediation. This is especially important when the specifications of the data are already strictly defined, yet are not followed, or are being misinterpreted. Semantic inconsistencies may arise due to differences in the interpretation of biological concepts and data, and differences in how such information is mapped into an integration system. That is to say, two systems may contain different data for the same semantic entity. For example, two interaction databases containing localization data for the proteins stored within, may indicate conflicting localization information for a given protein if the set of experimental evidences, used to determine localization, were different between the two systems. Such conflicts between data source providers pose challenges during the integration process as decisions need to be made to resolve the conflict. We continue to evaluate methods of resolving such conflicts. One simple solution is to store the information from all sources as is, and also annotate that information with the source from which it came, so as not to have any information loss. In this way, users can decide on which source they believe and poll the data accordingly. Another solution, which is not as clear cut, would be to selectively merge data, pruning those facts we determine to be incorrect (perhaps based on some measure of consensus between multiple systems), thus leaving only one instance of a factoid in our database. However, as it would not necessarily be our goal to judge the correctness of data, this is perhaps a task better left to users of our system. Comparison with other systems Several other systems are available which have similar goals and provide good solutions to the problem of data integration. We have chosen to discuss Atlas in the context of three other systems: Entrez [ 11 ], SeqHound [ 12 ] and EnsMart [ 13 ]. The Entrez system, produced by the NCBI, provides "an integrated database retrieval system that enables text searching, using simple Boolean queries, of a diverse set of 20 databases". This web-based system is extremely extensive in the scope of data it provides, and in fact many of the Atlas data sources originate from NCBI (GenBank, RefSeq, HomoloGene, Taxonomy, OMIM, Entrez Gene, and LocusLink). The Entrez resources can be found on the NCBI website [ 46 ]. In contrast to Entrez, Atlas warehouses the data locally, obviating the need for low-throughput, internet-based queries. Also, additional data sets like HPRD, DIP, MINT and BIND, not currently available through the Entrez interface, have been added to Atlas. SeqHound [ 12 ] is a database of biological sequences and structures, developed by the Blueprint Initiative [ 47 ]. SeqHound also stores information on OMIM, LocusLink, and Gene Ontology. SeqHound and Atlas warehouse similar data types. SeqHound provides some different data than Atlas (most notably MMDB). For interaction data, SeqHound utilizes the BIND database. In contrast, Atlas stores interaction data from a number of sources including BIND, HPRD, MINT, DIP, and IntAct. Atlas then, is a more comprehensive repository of interaction data. The major difference between SeqHound and Atlas is in their architectural design. SeqHound stores full records and indexes specific fields which are extracted upon loading. In contrast, Atlas provides relational models for all data sources. This allows SQL-level access to specific parts of the data model. The data in the Atlas relational models are stored as primitive data types as opposed to storing whole records that need parsing or processing. For example, sequences and their annotated biological features can be stored in their own fields in the database, permitting 'substring' operations to extract parts of the sequence that span a particular feature type using SQL. Other systems like EnsMart [ 13 ] and the UCSC genome browser [ 48 ] have also adopted fully relational models. These systems also provide SQL access over the full data model, and allow arbitrarily complex queries similar to Atlas. EnsMart is a software system designed by EMBL-EBI [ 49 ] and the Sanger Institute [ 50 ] which produces and manages automated annotations. The focus of EnsMart is slightly different than Atlas in that its 'core' data is fully sequenced eukaryotic genomes. While information on these genomes is extremely rich in EnsMart and well-integrated using relational models, Atlas attempts to provide a much more extensive source of sequence information. This enables researchers interested in bacteria, viruses, plants or humans to access the system and sources of integrated data with equal facility. The Atlas system is designed to be locally installed and is not a data provider per se , but rather an engine that should be accessed 'in-house'. As with any locally-installable system of this nature, significant time and hardware resources are needed to make the system functional. The utility of the Atlas system will far outweigh the setup time required to get it up and running. Currently, API access to Atlas is limited to the users at the UBC Bioinformatics Centre, University of British Columbia, however the web tools are available worldwide. Future work When working with sources of data from different data providers (for example UniProt and RefSeq), it is advantageous to create mappings from one data source to the other to prevent redundancy and to make associations between proteins to map annotations from one source to the other. We are investigating the idea of an identifier consolidation that can resolve mRNAs and proteins from different sources that are referring to the same protein product to a single identifier. We will constantly monitor and adjust any change of the data sources. In the near future, we will provide support for a PSI-MI level 2 release, and complete the migration of LocusLink to Entrez Gene. In addition, we are expanding Atlas to include other sources of data. We are currently adding MEDLINE, dbSNP and pathway data to support an integrative genomics and clinical informatics initiative, currently underway in our laboratory. With Atlas in hand, we are also working on an integration project that superimposes co-expression networks derived from microarray experiments and protein-protein interaction networks, to estimate the utility of co-expression networks in inferring protein interactions. Conclusion Atlas is a data warehouse that enables high-throughput, flexible and complex queries on biological data. The system integrates sequences, molecular interactions, taxonomy and homology, and functional annotations on genes. The system functions as data infrastructure to support bioinformatics research and development. Atlas is currently being used in genome annotation projects, disease-gene association projects and inference of molecular interactions. We are releasing Atlas to the scientific community in the hope that it will foster creative ideas for how to make novel associations between disparate sources of data using existing public data sets. Availability and requirements Atlas is available from the UBC Bioinformatics Centre, University of British Columbia. The Atlas package can be downloaded from the Atlas website at: The Atlas package contains the Atlas source code and represents the core of the project. The package is distributed under the GNU General Public License. Atlas is designed to run on Unix based systems. Please consult the user manual (available with the distribution) for detailed configuration, compilation and installation instructions. Additional packages are also provided at the website listed above. These packages include a snapshot of the NCBI C++ Toolkit (CVS version 20040505), a MySQL dump of sample data, and additional documentation. The NCBI C++ Toolkit, that is provided, is required only for those users who wish to build the loader applications or for those that require the utilities that convert ASN.1 format to GBFF, EMBL, and XML formats, etc. Those setting up the database will need to install MySQL Server 4.x. Atlas has been tested, specifically, with MySQL Server versions 4.0.9, 4.0.18 and 4.0.20, running on either Linux or AIX. The Atlas sequence-related binaries (toolbox applications and loader applications) are developed in C++ and therefore a C++ compatible compiler, such as the one included with the GNU GCC suite of tools, should be installed before attempting to build these binaries. We have tested the build process with GNU GCC versions 2.95.3, 2.96, 3.1 and 3.2. In addition, MySQL Client version 4.x and particularly its runtime library, libmysqlclient.a(so), is required. MySQL Client versions 4.0.14 and 4.1.0-alpha were tested. Details on the configuration and use of this library are outlined, in more detail, in the Atlas manual. For users that require Atlas tools that are based on Java, such as the loading and retrieval tools for LocusLink, BIND, HPRD, and HomoloGene datasets, a compatible Java interpreter must be installed. The API has been tested with J2SE 1.4.1 and J2SE 1.4.2. The Atlas Java API also requires BioJava version 1.4pre, or higher. For those using the Perl based Atlas tools, a compatible Perl interpreter must be installed. BioPerl version 1.4 must also be installed. Perl version 5.6.1 has been tested. Each of the packages have their own minimum system requirements. Specific memory, hard disk space and CPU requirements for each package are listed in the manual. As a general guideline, it is essential to have a generous amount of available memory, especially if one anticipates processing large sequences in memory. Another important factor is the amount of available hard disk space. The amount of sequence data to be loaded into Atlas will largely determine your disk space requirements. The Atlas database requires a minimum of 50 GB (RefSeq), plus adequate space for satellite databases. The satellite databases include such things as GO, LocusLink, HPRD, BIND, MINT, and DIP, which are relatively smaller datasets. Note that sequence data can greatly exceed these minimum estimates and the requirements should be carefully planned. Authors' contributions SS was the architect of the system, developed the C++ APIs and wrote the first draft of this manuscript. YH was the database administrator responsible for schema design, data integrity and maintenance. TX contributed the Java APIs. MMSY contributed the PERL APIs. JL developed the C++ APIs, the toolbox and the user manual. BFFO was the principal investigator, conceived of the project and guided its development. JL, YH, MSSY and BFFO all contributed to the writing of this manuscript.	/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC554782.xml
539234	P80, the HinT interacting membrane protein, is a secreted antigen of Mycoplasma hominis	Background Mycoplasmas are cell wall-less bacteria which encode a minimal set of proteins. In Mycoplasma hominis , the genes encoding the surface-localized membrane complex P60/P80 are in an operon with a gene encoding a cytoplasmic, n ucleotide-binding protein with a characteristic H istidine triad motif (HinT). HinT is found in both procaryotes and eukaryotes and known to hydrolyze adenosine nucleotides in eukaryotes. Immuno-precipitation and BIACore analysis revealed an interaction between HinT and the P80 domain of the membrane complex. As the membrane anchored P80 carries an N-terminal uncleaved signal peptide we have proposed that the N-terminus extends into the cytoplasm and interacts with the cytosolic HinT. Results Further characterization of P80 suggested that the 4.7 kDa signal peptide is protected from cleavage only in the membrane bound form. We found several proteins were released into the supernatant of a logarithmic phase mycoplasma culture, including P80, which was reduced in size by 10 kDa. Western blot analysis of recombinant P80 mutants expressed in E. coli and differing in the N-terminal region revealed that mutation of the +1 position of the mature protein (Asn to Pro) which is important for signal peptidase I recognition resulted in reduced P80 secretion. All other P80 variants were released into the supernatant, in general as a 74 kDa protein encompassing the helical part of P80. Incubation of M. hominis cells in phosphate buffered saline supplemented with divalent cations revealed that the release of mycoplasma proteins into the supernatant was inhibited by high concentrations of calciumions. Conclusions Our model for secretion of the P80 protein of M. hominis implies a two-step process. In general the P80 protein is transported across the membrane and remains complexed to P60, surface-exposed and membrane anchored via the uncleaved signal sequence. Loss of the 4.7 kDa signal peptide seems to be a pre-requisite for P80 secretion, which is followed by a proteolytic process leading to a helical 74 kDa product. We propose that this novel form of two-step secretion is one of the solutions to a life with a reduced gene set.	Background The contact of a pathogenic bacterium with its eukaryotic host provokes a multitude of reactions. A prerequisite for successful infection with the host is the cytadhesion of the bacterium generally mediated by surface localized proteins [ 1 ]. Besides adhesion, pathogens like Listeria , Yersinia and even some of the mycoplasmas are able to invade the host cells [ 2 - 4 ]. An intracellular localization is obviously a privileged niche, as the bacteria are well protected from the immune system. Moreover, bacteria not only remain concealed, but have evolved strategies for an attack on the eukaryotic cell. In secreting virulence factors, such as antigenic or toxic proteins, bacteria can mislead the host immune response or damage the colonized tissue [ 5 - 7 ]. The large majority of exported proteins possess an N-terminal signal sequence [ 8 ]. Most signal sequences are recognized by the Sec-dependent protein translocation complex (translocase), which mediates membrane translocation of unfolded precursors [ 9 ]. The signal sequences of proteins predicted to be recognized by type I signal peptidases are composed of a short, positively charged amino-terminal region (n-region), a central hydrophobic region (h-region) and a more polar carboxyl region (c-region) containing the cleavage site [ 10 ]. The signal peptides present in pre-lipoproteins additionally contain a well-conserved lipobox with an invariant cysteine residue that is lipid-modified prior to precursor cleavage by signal peptidase II [ 11 , 12 ]. Cleavage of the signal peptide is not required for translocation of the proteins through the membrane, but is generally the final step in processing [ 13 ]. However, some precursors remain membrane bound because of an uncleaved hydrophobic signal peptide and diffuse laterally from the translocase [ 14 ]. In the last few years, computer programs such as PSORT-II, PSORT-B, ExProt and SignalP have been developed to facilitate the identification of putative secreted proteins [ 15 - 18 ]. Comparison of proteomes of Gram-negative bacteria, Gram-positive bacteria and Archaea using ExProt revealed that the fraction of putative secreted proteins ranged from 8% in the archaeal bacterium Methanococcus jannaschii to 37% in the mollicute Mycoplasma pneumoniae [ 17 ]. Analysis of the exported proteins of Bacillus subtilis found that only 50% of the secreted proteins were detected by genomic prediction, indicating that proteomic analyses of secreted proteins (the secretome) are necessary for a comprehensive definition of all secreted proteins [ 19 ]. Only a handful of mollicute genomes have been decoded, but no analyses of their secretomes have been conducted. A secreted protein, probably processed by the classical mechanism described above, has been characterized in the swine pathogen M. hyopneumoniae . P102 is encoded as a precursor protein carrying a type I signal sequence and is found exclusively in the extra-cellular milieu suggesting cleavage by signal peptidase I. The expression of secreted P102 is coupled to that of the surface-exposed cilium adhesin P97, which seems to represent a new variant of processed surface antigens. P97 is derived from a 126 kDa precursor protein by cleavage at amino acid residue (aa) 195. The cleaved 22 kDa N-terminal fragment, which carries an uncleaved type I signal sequence is found embedded in the membrane, in the cytoplasm and in a soluble form in the supernatant, whereas the mature P97, proposed to be membrane bound, is the target of complex proteolytic cleavage, which leads to the subsequent release of some fragments into the supernatant [ 20 , 21 ]. The lipoprotein MALP-404 of M. fermentans is a further example of a surface-localized protein undergoing proteolysis after reaching its target. Site-specific cleavage leads to the generation of the membrane bound immune-stimulatory lipopeptide MALP-2 and the release of the remainder (the RF fragment) in a soluble form into the supernatant [ 22 ]. In M. hominis , which is mainly found as a commensal in the urogenital tract, but has also been associated with human urogenital tract infections [ 23 , 24 ], two variants of a cell-surface protein exist, one a lipoprotein (P120) and a homologous P120' without a lipid anchor, but containing an uncleaved N-terminal signal sequence for type I signal peptidases [ 25 ]. Recently we identified a gene locus coding for a surface-localized protein complex composed of a 60 kDa lipoprotein (P60) tightly bound to an 80 kDa precursor protein (P80) with an uncleaved type I signal sequence [ 26 ]. We have proposed that the uncleaved N-terminus extends into the cytoplasm and thus mediates an interaction with the concomitantly expressed cytoplasmic HinT protein. HinT is found in prokaryotes and eukaryotes and in the latter is known to function as an adenosine nucleotidyl-hydrolase [ 27 ]. The data in this paper suggest that the signal peptidase I recognition site is protected in the membrane bound form of P80, but is accessible as the first step in the release of the helical part of P80 into the cell culture supernatant, a process which is accompanied by a decrease in size of 10 kDa. Results To confirm our hypothesis that the P60/P80 membrane complex interacts with the cytoplasmic HinT via the uncleaved N-terminus of P80 that extends into the cytoplasm, we set out to express different fragments of the P80 protein to map the domain of P80 that interacted with HinT. As mycoplasmas use TGA as a tryptophan codon, rather than a stop codon as in E. coli , the TGA codons were mutated to TGG to allow expression. However, even though we mutated all TGA codons to TGG in the hit A gene, which encodes P80, we were not able to express and purify P80 peptides with an intact N-terminus. While P80-2, the C-terminal region of P80 from AA 320 to AA 713, was stably expressed in E. coli , P80-NT, the initial 21AA of the N-terminus, and P80-1, the N-terminal portion of P80 from AA 1 to AA 326, were degraded at the N-terminus during purification (Figure 1 ). Purification of the 2.6 kDa P80-NT fused to dihydrofolate reductase (DHFR) led to complete loss of the P80 region. As P80 NT did not contain the cleavage site of the signal peptidase I this proteolysis may be due to the presence of further processing signals within the first 21 AA of P80. The isolation of P80-1 revealed rapid degradation of the 40.7 kDa peptide with loss of 5, 6 and 8 kDa portions of the protein and loss of the poly His tag. Expression of the whole rP80 polypeptide chain (AA 1–713) resulted in a protein about 10 kDa smaller than expected, which lacked the poly His tag (depicted as a gray striped box in Figure 1 ). The conjecture that N-terminal degradation occurs was supported by Western blot analyses with anti-tetra His antibodies in which the purified forms of P80-1 and rP80 were not detectable (data not shown). As conventional protease inhibitors had no effect on the P80 degradation (data not shown) and as type I signal peptidases share the unusual feature of being resistant to general inhibitors of the four other peptidase classes [ 10 ] we speculated that the degradation observed in the expressed rP80 may be in fact the result of physiological processing in E. coli . Analysis of the mechanism of rP80 processing In order to test the hypothesis that the degradation of rP80 is mediated by E. coli signal peptidase I, we characterized different mutants of the recombinant P80 protein according to their processing in E. coli . Using PCR, we constructed clones expressing the mature P80 protein without the 44 AA signal peptide ( ΔSP ) and an 89 AA N-terminal truncated P80 protein, which corresponds to the proposed helical, surface-localized part of P80 ( Helic. ). As the absence of proline at the +1 position of the mature protein is consistent with the observation that the SPase I of E.coli was inhibited by recombinant precursor proteins having a proline residue at this position [ 12 ], we generated a P80 precursor with an Asn to Pro mutation ( N/P ) at the (+1) position in the mature protein expecting a reduced or complete inhibition of signal peptide cleavage (Figure 2-A ). Western-blot analysis of the heterologous expressed P80 variants in whole E. coli lysates revealed that a marked reduction of P80 processing was indeed observed for the ( N/P ) mutant, while the sizes of secreted P80 variants corresponded to that of the helical variant (Figure 2-B ). Small amounts of rP80 precursor were only observed in the lysate after prolonged exposure in immunostaining with the P80-specific mAb LF8 (data not shown). All variants, with the exception of the N/P mutant, were transported across the inner and the outer membranes and were of nearly the same size as the secreted 74 kDa peptide of the P80 protein found in M. hominis (Fig. 2 ; FBG, supernatant). In the case of P80-( ΔSP ) and P80-( Helic. ), the two variants lacking signal sequences as translocation signals, this transport may be mediated by an alternative, Sec-independent export mechanism. P80 secretion in Mycoplasma hominis The data suggest that the recombinant P80 processing in E. coli is signal peptidase I-mediated and therefore we investigated the protection of the signal peptidase I recognition site of P80 in Mycoplasma hominis [ 26 ]. Therefore, we analyzed the cellular proteins and the proteins in the culture supernatant in a mycoplasma culture in early-, mid- and stationary growth phases. To ensure that no viable cells were left in the supernatant we checked the separation efficiency by determining the titres of mycoplasmas in the supernatant and the original culture. The titres in the supernatants were 0.01% to 0.13% of those of the original cultures, demonstrating that the Western blot analysis would not be confounded by inclusion of cellular proteins in the supernatants (data not shown). As might be expected the stationary phase culture contained several proteins that might be released from lysed cells, including P50, P60, OppA and cytoplasmic proteins such as P55 (Figure 3-A ). However, the same membrane proteins were found in the mid-logarithmic phase suggesting release of these proteins from living cells into the supernatant. Indeed, the amount of the cytoplasmic P55 was reduced in this culture suggesting that in mid-logarithmic phase few proteins from lysed cells are present. Lipoproteins, such as P50 and P60 were also found in the supernatant of a mycoplasma culture at the beginning of logarithmic growth (early), whereas OppA was absent. In general, immunostaining of P80 was weak in samples derived from the supernatant of a broth culture. This might be due to a masking of P80 by albumin, which was present in high concentrations in the culture medium and of similar size as the secreted P80 protein. Interestingly, the P80 protein was the only one of the proteins analyzed that had a decreased size in the supernatant. To exclude the possibility that the reduction in size was simply a result of the large amount of albumin in the culture medium, we purified P80 from the supernatant of a culture in mid logarithmic growth using affinity chromatography. In addition we analyzed P60 as a control. As shown in Figure 3-B , the secreted P80 peptide was decreased in size by approximately 10 kDa in comparison to the cellular form, whereas P60 remained unaltered. A degree of processing of the P80 precursor was detectable, with several distinct P80 staining bands between 80 and 74 kDa, running through distinct steps of degradation as shown for the recombinant P80-1 peptide (see Figure 1 ). As the activity of proteases often depends on the presence of divalent cations, we examined the secretion of mycoplasma proteins in the presence or absence of divalent cations. Mycoplasma cells were incubated for 1 h at 37°C in phosphate buffered saline (PBS) supplemented with up to 5 mM magnesium or calcium chloride, or chelating agents as EDTA and EGTA. The addition of divalent cations did not increase the processing of P80 (data not shown). As the presence of 5 mM calcium ions resulted in a complete disappearence of P80 from the supernatant we examined the release of mycoplasma proteins from cells following stepwise addition of calcium ions at concentrations of 0 to 5 mM (Figure 4 ). Silver staining of the cell and supernatant preparations revealed that only a few proteins were released in the supernatant whereas most proteins remained cell-embedded (Figure 4-A ). Increasing the calcium chloride concentration to more that 1.3 mM led to a dramatic reduction in secretion. All the antigens analyzed, with the exception of OppA, were found in the supernatant, even the cytoplasmic proteins P55 and elongation factor Tu (EF-Tu) (Figure 4-B ). The cytoplasmic proteins that were detected in the supernatant were not released from lysed cells as the titres of the cultures remained the same throughout the experiment (data not shown). These findings are in accordance with those of Antelmann and coworkers, who found several cytoplasmic proteins, such as elongation factor G (EF-G) and arginase (RocF) as to be components of the secretome of Bacillus subtilis [ 19 ]. Increasing the concentration of calcium ions to more than 0.6 mM CaCl 2 inhibited a release of most mycoplasma proteins from the cells (Figure 4-B , lane 4). Of the proteins analyzed, only EF-Tu was found in the supernatant of a sample treated with 1.3 mM CaCl 2 (Fig. 4-B , lane 3). Thus the surface-localized P80 protein, normally complexed with the P60 lipoprotein in the membrane, appears to be released into the extra-cellular environment as a 74 kDa cleavage product. Our model of processing describes the initial cleavage of the 4.7 kDa signal peptide by a type I signal peptidase, followed by further degradation by unknown mechanisms, leading to a stable 74 kDa peptide. Discussion The data presented here indicate that the membrane protein P80 of Mycoplasma hominis is a representative of a new group of proteins in the Mollicutes . P80 occurs as a precursor protein (complexed with the lipoprotein P60) anchored in the membrane and exposed on the surface, and is also released into the extra-cellular environment. As P80 and P60 are concomitantly expressed and both complexed in the membrane fraction it is likely that the soluble P80 variant is derived from the membrane bound precursor. Interestingly, the findings from the heterologous expression of different P80 mutants in E. coli suggest that release of P80 is initiated by cleavage of the N-terminal type I signal sequence, followed by amino-terminal proteolysis by an unknown mechanism, leading to a stable P80 peptide with a predicted helical structure in the supernatant. As the genome of M. genitalium , the smallest known self-sufficient organism, lacks a gene for the type I signal peptidase [ 30 ] it was initially speculated that this enzyme may be absent in the Mollicutes , despite the fact that several proteins with type I-signal sequences had been identified [ 10 , 25 ]. However signal peptidase I genes have been found in the genomic sequences of Mycoplasma gallisepticum (MGA_1091) and M. pulmonis (MYPU_6300) [ 31 , 32 ]. The release of mycoplasmal proteins into the extra-cellular milieu may follow different pathways to those described previously for other bacteria. While the exclusive detection of P102 of M. hyopneumoniae in the extra-cellular matrix suggests that the intrinsic N-terminal type I signal peptide is immediately cleaved after translocation of the precursor through the membrane, as would be expected from current knowledge of the classical protein secretion pathway, the precursor proteins P80 and P120' of M. hominis are attached to the membrane without being processed or released into the supernatant [ 21 , 25 , 26 ]. The later release of a surface-localized peptide is described here for the first time. Davis and Wise described site-specific proteolysis of the lipid-anchored MALP-404 of M. fermentans that leads to the generation of the immune-stimulatory lipopeptide MALP-2 and showed that the residual RF peptide was soluble and was released into the supernatant [ 22 ]. Processing of the P97 cilium adhesin precursor, P126, of M. hyopneumoniae appears to be quite complex. The precursor protein P126, which carries a type I signal sequence with a cleavage site between AA 31 and AA 32, is predominantly cleaved between AA 194 and AA 195, suggesting that the SP-I signature may be essential only for the translocation of the whole precursor protein across the membrane. The cleavage products P22 and P97 both remain closely associated with the membrane. While P97 is generally subjected to further site-directed proteolysis, leading to the release of peptides into the extra-cellular milieu, further processing of P22 occurs to be strain dependent [ 21 ]. Release of proteins into the surroundings of a mycoplasma cell should lead to an immediate alteration of the cell surface architecture and, as most membrane proteins are targets for the host immune response [ 33 , 34 ], may also interfere with the host effector response. Type II secreted proteins appeared to be typically associated with non-invasive organisms colonizing mucosal surfaces, such as Mycoplasma hominis , and are considered to be required for the establishment of an infection at these sites. Recently in Legionella pneumophila a type II protein secretion system was characterized as a virulence factor linked to an intracellular infection [ 35 ]. P80 has significant similarity with two protein sequences, a hypothetical protein of M. pulmonis (MYPU_0060; gi15828477) the gene of which is followed by P60 (MYPU_0070) and HinT (MYPU_0080) gene homologues, and a rhoptry protein of Plasmodium yoelii yoelii (gi23481286). Rhoptry proteins are released during host cell invasion by apicomplexans [ 36 ]. As P80 interacts with HinT, a cytoplasmic protein found in all kingdoms, which in eukaryotes hydrolyses AMP derivatives and in yeasts functions as a regulator of an RNA polymerase II domain [ 27 ], a further scenario is imaginable: We do not know which factors promote the processing of the surface-exposed P80 protein. However, release of the helical P80 peptide is accompanied by the retention of the signal peptide I in the membrane. Some signal peptides have been found to leak back into the cell where they bind to proteins, such as the Ca 2+ -binding calmodulin [ 37 ], or are presented to the immune system [ 38 ], where they probably have a secondary signaling function distinct from their role in targeting [ 39 ]. Our results indicate that an increase in calcium ions prevents secretion of mycoplasma proteins. Thus, an extra- or intra- cellular stimulus – such as a local variation in the concentration of Ca 2+ ions – may activate the signal peptidase induced processing of the membrane bound P80 protein. This would lead to a soluble P80-helix and the membrane anchored signal peptide. After loss of the transmembrane spanning C-terminal amino acids or due to a leakage of the whole signal sequence (as described above) the signal peptide could reach the cytoplasm where an interaction with HinT may be a further step in modulating or promoting cellular processes such as growth or activation of gene expression, a process we are currently investigating. Conclusions The data presented here clearly demonstrate that P80 is a secreted antigen of Mycoplasma hominis . This is, to our knowledge, the first description of secreted protein with a type I signal sequence that is stably embedded in the membrane as precursor protein and, as we propose as a model of secretion, is subsequently released from the membrane into the supernatant after signal peptidase I cleavage. As in other mycoplasmas several proteins have been shown to either possess an uncleaved signal sequence or to be further processed after anchoring in the membrane, this may be a common phenomenon in mycoplasmas. Because of their minimal coding capacity, mycoplasmas may have evolved a strategy to use secretory proteins in a dual role, as surface localized proteins in the membrane for cell surface architecture, and, in response to a change in the environment, as a soluble protein released from the cell surface. Methods Cloning and expression of rP80 and P80 mutants The P80 protein encoding region (ACC Z29068) of Mycoplasma hominis strain FBG was amplified by PCR [ 26 ] using oligonucleotides (MWG Biotech, Ebersberg, Germany) that change the mycoplasma TGA tryptophan codon to TGG. Fragments were either directly cloned for the expression of distinct P80 regions or fused by SOE (splicing by overlap extension)-PCR [ 40 ]. To facilitate cloning of PCR products [ 41 ], restriction sites were inserted in the primer sequence without changing the amino acid sequence. P80 variants with mutations in the N-terminal part of P80 were generated by PCR by amplifying the region encoding the helical portion of P80 ( Helic., AA 90–713), amplifying a fragment encoding the mature P80 polypeptide ( ΔSP , AA 45–713), and mutating the asparagine codon at position +1 of the mature protein to a proline codon ( N/P ). Amplicons were cut at the restriction sites within the primers and ligated in-frame into the expression vector pQE30 (Qiagen, Hilden, Germany). To express the P80 N-terminus (AA 1–21) as a fusion protein with dihydrofolate reductase and a poly His-tag, we ligated the respective Rca I/ Bgl II restricted amplicon in frame into the Nco I/ Bgl II restricted plasmid pQE60 and inserted the Bam HI/ Bgl II restricted DHFR-fragment of pQE40 in the Bgl II site of the plasmid pQE60-P80 NT. Plasmids were propagated in DH5αF' for constitutive expression [ 41 ]. Cells from a mid-logarithmic phase culture in LB broth (Gibco BRL, Life Technologies Inc., Gaithersburg, Md.) containing ampicillin (100 μg/ml) were harvested by centrifugation (20,000 × g , 30 min, 4°C). In the secretion assay, the pelleted cells and supernatant, which was concentrated by lyophilizing, were subjected to Western blot analysis. The recombinant peptides P80-NT and P80-2 were purified by Ni-NTA chelation according to the manufacturer's protocol, while rP80 and P80-1 were purified with the Sepharose-coated P80-specific antibodies LF8 and NB12 as described by Henrich et al. [ 34 ]. Sequence analysis The analysis of DNA and protein sequences and the design of oligonucleotides were facilitated by use of the software package Lasergene (DNASTAR Inc. 1996, Madison, Wisc.). The integrity of the different plasmid sequences was confirmed by analysis on an ABI sequencer using the method of Sanger [ 42 ]. Mycoplasma culture, osmotic lysis and secretion assays Mycoplasma hominis strain FBG was cultivated in PPLO broth supplemented with arginine from frozen stocks of a mid-logarithmic phase broth culture as described previously [ 43 ]. The titer of the broth culture was determined by the measuring the number of color changing units (CCU) [ 34 ]. For protein secretion analyses, the mycoplasma cells from early- to mid-logarithmic phase cultures were harvested by centrifugation (20,000 × g , 10 min, 4°C), and the cell pellets and supernatants were subjected to Western blot analysis. Alternatively, cell pellets derived by low speed centrifugation (10,000 × g , 10 min, 4°C) were re-suspended in phosphate-buffered saline (PBS; 120 mM NaCl, 5 mM KCl, 20 mM Tris-HCl, pH 7.5) containing 5.0, 2.5, 1.25, 0.63, 0.31 or 0.00 mM CaCl 2 and incubated at 37°C for 60 min. Soluble proteins were then separated from insoluble material by centrifugation (15,000 × g, 15 min, 4°C), and all samples analyzed by Silver staining [ 44 ] and Western blotting. Western blot analysis Proteins were separated in 9.5% polyacrylamide gels [ 45 ], transferred to nitrocellulose (Schleicher and Schüll, Dassel, Germany) with a semidry blotting apparatus (Phase, Mölln, Germany) [ 46 ], and immunostained as described by Henrich et al. [ 34 ] using the monoclonal antibodies BG11 (anti-OppA), BG2 (anti-P50), LF8 (anti-P80), or CG4 (anti-P60), all of which are directed against membrane proteins, and AH10 (P55) and KD2 (EF-Tu), which are directed against proteins primarily located in the cytoplasm. Authors' contributions RH began the characterization of antigens released into the supernatant of a M. hominis culture as part of her thesis. MH carried out the immunoassays and completed the secretion assays. BH carried out the molecular genetic studies, participated in the design of the study and drafted the manuscript. All authors have read and approved of the final manuscript.	/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC539234.xml
544585	Retroviral transduction of peptide stimulated t cells can generate dual t cell receptor-expressing (bifunctional) t cells reactive with two defined antigens	Background Tumors and viruses have developed many mechanisms to evade the immune system, including down-regulation of target antigens and MHC molecules. These immune escape mechanisms may be able to be circumvented by adoptively transferring T cells engineered to express two different T cell receptors, each specific for a different antigen or MHC restriction molecule. Methods PBMC from the blood of normal healthy donors were stimulated for three days with an antigenic peptide from cytomegalovirus (CMV) pp65. These CMV reactive cultures were transduced with a encoding the TIL 5 T cell receptor (TCR) that mediates recognition of the dominant epitope of the melanoma antigen MART-1. Following selection for transduced cells, the cultures were evaluated for recognition of CMV pp65 and MART-1 expressing targets. Results We were able to rapidly create bifunctional T cells capable of recognizing both CMV pp65 and MART-1 using a combination of HLA-A2 tetramer staining and intracellular staining for interferon-γ. These bifunctional T cells were sensitive to very low levels of antigen, recognize MART-1 + tumor cells, and maintained their bifunctionality for over 40 days in culture. Conclusion Bifunctional T cells can be engineered by transducing short term peptide stimulated T cell cultures. These bifunctional T cells may be more effective in treating patients with cancer or chronic virus infections because they would reduce the possibility of disease progression due to antigen and/or MHC loss variants.	Background It has long been established that tumors and viruses have multiple mechanisms for evading the immune system including the inhibition of T cell function through the release of inhibitory cytokines and factors [ 1 , 2 ], down-regulation of MHC molecules [ 2 , 3 ] and the spontaneous generation of antigen-loss variants [ 1 - 3 ]. In latter case, despite the loss of a single antigen on a tumor or virus-infected cell, there can remain functional HLA class I molecules and multiple antigens that can serve as targets for immune destruction. Therefore, immunotherapy strategies which target multiple antigens and/or multiple HLA class I molecules may be more effective than therapies targeting single antigens presented by a single HLA class I molecule. We and others have shown that it is possible to use retroviral vectors encoding TCRs isolated from tumor- or virus-reactive T cell clones to engineer human T cells to recognize any antigen [ 4 ]. While every T cell that is transduced to express a second TCR expresses its own TCR capable of recognizing some antigen, it has only recently been shown that "bifunctional" T cells capable of recognizing two known antigens can be generated [ 5 ]. Using this technology, it may be possible to treat patients with T cells bearing two functional T cell receptors (TCRs) with each TCR being specific for a different tumor-associated antigen (TAA) or viral antigen restricted by one or more HLA molecule. These bifunctional T cells would retain effectiveness against single antigen-loss variants or HLA loss variants and may have improved efficacy over monospecific T cells for the treatment of tumors or viruses. In the current study, we show that it is possible to rapidly generate T cell populations containing T cells reactive with two defined antigens, CMV pp65 and MART-1. These T cell cultures are highly avid for both antigens and retain their reactivity for at least six weeks. More importantly, this methodology could easily be adapted to closed culture systems making it more attractive for use in clinical trials. Methods Tumor Cell Lines All melanoma and renal cell carcinoma cell lines used in this study were established from surgical specimens obtained from cancer patients undergoing immunotherapy at the Surgery Branch, National Cancer Institute. Melanoma cell lines 624 MEL (HLA-A2 + , MART-1 + ), 624-28 MEL (HLA-A2 - , MART-1 + ), 1300 MEL (HLA-A2 + , MART-1 + ), and SK23 MEL (HLA-A2 + , MART-1 + ) and renal cell carcinoma cell line UOK131 (HLA-A2 + , MART-1 - ) were maintained complete medium (CM) which consisted of RPMI 1640 medium (Life Technologies, Gaithersburg, MD) supplemented with 10% heat inactivated fetal bovine serum (Life Technologies), and penicillin (100 U/ml)/streptomycin (100 μg/ml)/L-glutamine (2.92 mg/ml)(Life Technologies) as described [ 6 ]. The PG13 A7 retroviral producer cell line, the source of the TIL 5 TCR retrovirus used in this study, has been described elsewhere [ 7 ]. Retroviral production by PG13 A7 cells was carried out using an optimized protocol described by Lamers et al [ 8 ] in T175 flasks at 32°C in CM. T2 cells and COS A2 were maintained in CM as described [ 6 ]. T Cells R6C12 is an HLA-A2 restricted, gp100:209–217 reactive CTL clone that was isolated from the peripheral blood of a melanoma patient vaccinated with the gp100:209–217 210M peptide at the Surgery Branch, National Cancer Institute. R6C12 cells were expanded using 10 ng/ml anti-CD3 mAb (Ortho Biotech, Raritan, NJ) and 300 IU/mL recombinant human IL-2 (rhIL-2) (Chiron, Berkeley, CA) in T cell medium (TCM) which consisted of RPMI 1640 supplemented with 10% pooled human AB serum (Valley Biochemical, Winchester VA), HEPES (Life Technologies), 2-mercaptoethanol (Life Technologies), penicillin (100 U/ml)/streptomycin (100 μg/ml)/L-glutamine (2.92 mg/ml)(Life Technologies) as described [ 9 ]. Peripheral blood mononuclear cells (PBMC) obtained from leukapheresis of healthy donors were used as a source of T cells for establishing CMV pp65 peptide stimulated T cell cultures and feeders for T cell expansion were purchased from BRT Laboratories (Baltimore, MD). Peptides All peptides used in this study were purchased from Synthetic Biomolecules (San Diego, CA). Peptides used were MART-1:27–35 (AAGIGILTV), Influenza M1:58–66 (GILGFVFTL), CMV pp65:495–503 (NLVPMVATV), or gp100:209–217 (ITDQVPFSV). Each peptide was maintained as a concentrated stock (2–5 mg/ml) in 100% DMSO (Sigma, St. Louis, MO) and diluted in the appropriate medium prior to immediate use. CMV pp65 Expressing Targets Given that it is technically difficult to obtain CMV infected targets for immunologic assays, COS A2 cells were engineered to express a mini-gene encoding the CMV pp65:495–503 peptide epitope (COS A2 CMV). A CMV minigene was constructed using complementary synthetic oligonucleotide primers (sense primer: 5'-GGCCCGCGCAGGCAGCATGAACCTGGTGCCCATGGTGGCTACGGTTTAGTGA-3', anti-sense primer: 5'-GGCCTCACTAAACCGTAGCCACCATGGGCACCAGGTTCATGCTGCCTGCGCG-3', Integrated DNA Technologies, Coralville, IA) that encoded the CMV pp65:495–503 peptide epitope with an ATG translation initiation codon, a Kozak consensus sequence [ 10 ] and Not I compatible "sticky ends" to facilitate insertion into the Not I site of the SAMEN CMV/SRα retrovirus. Equal molar amounts of each synthetic oligonucleotude were mixed and ligated into the SAMEN CMV/SRa retrovirus using a rapid ligation protocol and transformed into DH5α competent E. coli cells (Life Technologies) as described [ 11 ]. Recombinant clones were sequenced to insure proper orientation and retroviral supernatants were produced by cotransiently transfecting 293GP cells with plasmids encoding the retroviral backbone and the vesicular stomatitis virus envelope as described [ 11 ]. COS A2 CMV cells were generated by culturing COS A2 cells overnight with retroviral supernatants supplemented with 8 μg/ml polybrene (Sigma). Peptide Stimulation and Transduction of PBMC PBMC from healthy donors were stimulated in vitro with 5 μg/ml of CMV pp65:495–503 peptide in TCM containing 300 IU/mL IL-2 for 3 days. T cell cultures were then transduced using a modified Retronectin (TaKaRa, Otsu, Japan) protocol with the A7 retrovirus as follows: 24-well plates were coated with Retronectin then were preloaded with retrovirus according to the manufacturer's instructions. 2.6 × 10 6 T cells were added to each well in 1.3 ml (2 × 10 6 cells/ml) of A7 retroviral supernantant supplemented with 300 IU/ml rhIL-2 and the plates were centrifuged for 90 min at 1000 g. The next day the medium was replaced with fresh A7 retroviral supernatant and the centrifugation was repeated. The cells were rested for 24 hours and then transduced cells were selected in 1 mg/ml of G418 (Research Products International, Mt. Prospect, IL) for five days. Cultures were assayed for antigen reactivity, cyropreserved, and/or expanded for additional assays. Transduction of T Cell Clones T cell clone R6C12 was cultured at 2 × 10 6 cells/ml in TCM supplemented with 300 IU/ml rhIL-2, and 2 μg/ml anti-CD28 mAb (Becton, Dickenson, and Company, Franklin Lakes, NJ) in 24 well tissue culture plates pre-coated overnight with 10 μg of anti-CD3 mAb (Ortho Biotech, Bridgewater, NJ) for three days prior to transduction. Transduction of R6C12 was carried out as described above for CMV peptide stimulated T cell cultures except 2 μg/ml anti-CD28 mAb was added to the medium and 10 μg anti-CD3 mAb was bound to the culture plates in addition to Retronectin. Antigen Recognition Assays The antigen reactivity of each T cell culture (TIL 5 TCR transduced and untransduced) was assayed for MART-1:27–35 and CMV pp65:495–503 or gp100:209–217 reactivity in interferon-γ release assays. 5 × 10 4 T cells were cocultured in a 1:1 ratio overnight in 0.2 ml of CM in duplicate individual wells of a 96-well plate with a panel of stimulators that included T2 cells loaded with 5 μg/mL MART-1:27–35, Influenza M1:58–66, CMV pp65:495–503, or gp100:209–217 peptide and a panel of tumor cells. The amount of interferon-γ released was measured by ELISA as described [ 6 ]. Intracellular Cytokine Release Assay and Tetramer Staining The existence of CMV pp65:495–503/MART-1:27–35 reactive bifunctional T cells was determined by first staining T cells for intracellular interferon-γ production following coculture with HLA-A2 + MART-1 + stimulator cells followed by fluorescence staining with HLA-A2/CMV pp65:495–503 tetramers. 1 × 10 5 T cells were cocultured in a 1:1 ratio peptide loaded T2 cells or tumor cells for five hours in CM supplemented with 10 μg/ml brefeldin-A. Cells were then collected and stained with PE-conjugated HLA-A2/CMV pp65:495503 tetramers (Beckman Coulter Immunomics, San Diego, CA), fixed in 1% paraformaldehyde (Sigma), permeabilized using 0.5% saponin (Sigma), and then stained with FITC-conjugated anti-interferon-γ (Biosource International, Camarillo, CA). Relative log fluorescence of 10 4 live cells was measured by flow cytometry using a FACS Scan flow cytometer (BD Biosciences, Mountain View, CA). Results Recognition of Peptides and Tumor Cells by TIL 5 TCR-transduced CMV peptide stimulated T cells Bifunctional T cells reactive with CMV and MART-1 were engineered by first stimulating donor PBMC with CMV pp65:495–503 peptide for three days then transducing the T cell cultures with a retrovirus encoding a TCR specific for MART-1:27–35 presented by HLA-A2. After five days of selection in G418, the T cells were assayed for reactivity against the CMV pp65:495–503 and MART-1:27–35 antigens in interferon-γ release assays. Significant amounts of interferon-γ were released when the TIL 5 TCR-transduced CMV peptide stimulated T cells were cocultured with CMV pp65:495–503 or MART-1:27–35 peptide-loaded T2 cells, COS cells engineered to express HLA-A2 with a CMV pp65:495–503 mini-gene, or HLA-A2 + MART-1 + tumor cells (Figure 1 ). These cells did not release interferon-γ when stimulated with T2 cells loaded with Flu M1:58–66 peptide, COS A2 (MART-1 - CMV - ), 624-28 MEL (HLA-A2 - MART-1 + CMV - ), or RCC UOK131 (HLA-A2 + MART-1 - CMV - ) cells. Untransduced CMV peptide stimulated T cells only released interferon-γ when stimulated with CMV pp65:495–503 peptide loaded T2 or COS HLA-A2 + CMV + cells. These cultures were extremely sensitive to antigen stimulation since significant amounts of interferon-γ were released when stimulated with T2 cells loaded with 5 × 10 -4 μg/mL MART-1:27–35 peptide and <5 × 10 -7 μg/mL CMV pp65:495–503 peptide (Figure 2 ). These bulk cultures also continued to be reactive to both antigens more than 40 days post transduction (Figure 3 ). These results indicate that three day peptide stimulated PBMC cultures can be activated in vitro for efficient retroviral transduction. Furthermore, the antigen reactivity of these T cells is consistent with bifunctional T cells capable of recognizing both CMV and MART-1. Figure 1 Recognition of peptides and tumor cells by TIL 5 TCR-transduced CMV-stimulated T cells. TIL 5 TCR-transduced and untransduced CMV-stimulated bulk cultures were cocultured for 24 hours in a 1:1 ratio with T2 cells loaded with 5 μg/ml of peptide, COS A2 cells with or without a CMV minigene, or A2 + , MART-1 + tumor cells (SK23 MEL, 624 MEL), A2 - , MART-1 + tumor cells (624.28 MEL), or A2 + , MART-1 - tumor cells (UOK131 RCC). Supernatants were collected and the amount of interferon-γ released was measured using ELISA. Values are the average of triplicate wells. Figure 2 Sensitivity of bifunctional cultures to low levels of recognized antigens. TIL 5 TCR-transduced cultures were cocultured for 24 hours in a 1:1 ratio with T2 cells loaded with decreasing concentrations of peptide. Supernatants were collected and the amount of interferon-γ released was measured using ELISA. Values are the average of triplicate wells. Asterisk () indicates value was greater than maximum point on standard curve. Figure 3 Long term maintenance of bifunctionality in culture. TIL 5 TCR-transduced cultures were cocultured for 24 hours in a 1:1 ratio with T2 cells pulsed with 5 μg/ml peptide at varying time points beyond transduction. Supernatants were collected and the amount of interferon-γ released was measured using ELISA. Values are the average of triplicate wells. Asterisk () indicates value was greater than maximum point on standard curve. Antigen Recognition by CMV-tetramer positive T cells While the antigen reactivity of our T cell cultures was consistent with us having engineered bifunctional T cells, it was necessary to confirm that individual T cells possess the capability to recognize both CMV pp65 and MART-1. To confirm that we had engineered bifunctional cells, each T cell culture was stained with HLA-A2/CMV pp65:495–503 tetramers for anti-CMV reactivity and with intracellular anti-interferon-γ monoclonal antibodies following stimulation with HLA-A2 + MART-1 + cells for anti-MART-1 reactivity. It should be noted that the reciprocal experiment, staining with HLA-A2/MART-1:27–35 tetramers and intracellular anti-interferon-γ staining following CMV pp65:495–503 peptide stimulation could not be performed since TIL 5 TCR expressing cells do not bind HLA-A2/MART-1:27–35 tetramers (unpublished). Cells that were double stained with tetramers and for intracellular anti-interferon-γ were considered to be reactive with both antigens and therefore bifunctional. As shown in Figure 4 , 2.7% of the TIL 5 TCR-transduced T cells were double stained following stimulation with MART-1:27–35 loaded T2 cells compared to 0.06% of the untransduced T cells. When stimulated with CMV pp65:495–503 peptide loaded cells, 28.35% of the TIL 5 TCR-transduced T cells were double stained. These results are representative of multiple cultures which routinely have approximately 10% of the CMV reactive T cells also recognizing MART-1. These results confirm that bifunctional T cells can be obtained by transducing three day peptide stimulated PBMC cultures with retroviral vectors encoding TCR genes. Figure 4 Peptide recognition by CMV tetramer + T cells. TIL 5 TCR-transduced and untransduced CMV-stimulated bulk cultures were cocultured in the presence of 10 μg/ml brefeldin A for 5 hours in a 1:1 ratio with T2 cells pulsed with 5 μg/ml of peptide. Cells were then collected, stained with PE-conjugated HLA-A2 MHC CMV tetramer, fixed, permeabilized, and stained with FITC-conjugated anti-IFN-γ mAb. Samples were analysed using two-color flow cytometry. The percentage of dual positive staining cells (upper right quadrant) is as indicated. Although the reactivity with MART-1:27–35 peptide loaded T2 cells shown in Figure 4 confirmed that we successfully engineered CMV pp65 peptide stimulated PBL-derived T cell cultures to contain bifunctional T cells, it was important to determine if these cultures could recognize the physiologic levels of antigen presented by tumor cells. When stimulated with HLA-A2 + MART-1 + tumor cells, 0.65% (1300 MEL) and 0.52% (SK23 MEL) of the T cells were HLA-A2/CMV pp65 tetramer positive and interferon-γ positive indicating that approximately 20% of the peptide reactive T cells were also tumor reactive (Figure 5 ). Tumor cells are poor antigen presenters relative to T2 cells because they often fail to express the accessory molecules required for efficient T cell recognition. Furthermore, only those T cells with sufficient TIL 5 TCR expression to yield high avidity T cells are capable of responding to the levels of processed antigen on the surface of tumor cells. This explains why a smaller fraction of bifunctional T cells are reactive with tumor cells compared to peptide-loaded T2 cells. Figure 5 Tumor cell recognition by CMV tetramer + T cells. TIL 5 TCR-transduced and untransduced CMV-stimulated bulk cultures were cocultured in the presence of 10 μg/ml brefeldin A for 5 hours in a 1:1 ratio with A2 + , MART + tumor cells (SK23 MEL, 1300 MEL) or A2 + , MART-1 - tumor cells (UOK131 RCC). Cells were then collected, stained with PE-conjugated HLA-A2 MHC CMV tetramer, fixed, permeabilized, and stained with FITC-conjugated anti-IFN-γ mAb. Samples were analyzed using two-color flow cytometry. The percentage of dual positive staining cells (upper right quadrant) is as indicated. MART-1 Recognition by a TIL 5 TCR-transduced gp100-reactive T cell clone To demonstrate that the creation of bifunctional T cells capable of recognizing two tumor antigens was possible using our transduction methods, T cell clone R6C12 cells were activated with anti-CD3 and anti-CD28 mAb then transduced to express the TIL 5 TCR. TIL 5 TCR-transduced R6C12 cells were cocultured with gp100:209–217, Influenza M1, or MART-1:27–35 peptide-loaded T2 cells. As shown in Figure 6 , TIL 5 TCR transduced R6C12 cells released interferon-γ when stimulated with MART-1:27–35 or gp100:209–217 loaded T2 cells, but not T2 cells loaded with the irrelevant influenza M1 peptide. In contrast, untransduced R6C12 cells only released interferon-γ when stimulated with gp100:209–217 loaded T 2 cells. Figure 6 Peptide recognition by TIL 5 TCR-transduced gp100-reactive T cell clone. TIL 5 TCR-transduced and untransduced R6C12 cells were cocultured for 24 hours in a 1:1 ratio with T2 cells loaded with 5 μg/ml of peptide. Supernatants were collected and the amount of interferon-γ released was measured using ELISA. Values are the average of duplicate wells. Discussion Here we report the successful engineering of T cells that are able to respond independently to two unrelated known antigens via both an endogenous and a retrovirally-transduced T cell receptor. These T cells were able to respond to low concentration of peptide, and were able to recognize antigen-positive tumor cells. By utilizing the initial antigen response as the activation for transduction, our 12-day protocol represents an efficient technique for generating bifunctional T cells from donor blood, and theoretically can be applied to any tumor or viral antigen in the context of one or more MHC restricting elements. Many previous efforts at creating TCR transductants used non-specific activation of bulk or clonal populations [ 7 , 8 , 12 ] or, for the creation of bifunctional T cells, specific activation of semi-clonal populations with peptide-loaded autologous PBMC [ 5 ]. Non-specific T cell activation fails to expand T cell populations with known reactivity hence making it virtually impossible engineer T cells reactive with two know antigens. Engineering clonal or semi-clonal populations of T cells will create T cells reactive with two known antigens [ 5 ]. However, this process necessitates the establishment of antigen reactive T cell clones or long term T cell cultures prior to transduction. Although technically feasible, the creation of bifunctional T cells from T cell clones (this study) or long term T cell cultures [ 5 ] is time consuming and in our experience has a comparatively low yield of bifunctional cells. Furthermore, it is likely that the reactivity and therapeutic efficacy of T cells are diminished with extended culturing (13). Therefore, any method capable of rapidly producing bifunctional T cells will be better suited to clinical applications. In contrast to using anti-CD3 mAb, in vitro stimulation with antigenic peptides will preferentially activate antigen reactive T cells to expand. These proliferating antigen reactive T cells can be transduced to express a second TCR. Based on our tetramer analysis, only 0.7% of the unstimulated donor PBL stained with the CMV tetramer (data not shown), compared to 44.6% of our peptide stimulated populations (data not shown). This profound expansion allowed for more efficient transduction, and 2.7% of the resulting culture was measurably bifunctional (figure 4 ). As retroviral transduction and in vitro selection for transduced T cells becomes more efficient, the frequency of bifunctional T cells in these cultures will increase to the point where it is feasible to treat patients. The combination tumor/viral bifunctional cells we have generated here may have novel uses in immunotherapy, such as bypassing tumor- or viral-induced T cell unresponsiveness. Fossati and colleagues demonstrated that naïve bifunctional T cells "preactivated" via one TCR prior to adoptive transfer would then mediate cytotoxicity via the second TCR [ 14 ]. Animal and in vitro studies have shown that peripherally-induced tolerance can be reversed, resulting in regained T cell responsiveness [ 15 , 16 ]. It may be possible to reactivate tolerized T cells in vitro or in vivo by activating a second T cell receptor specific for a non-tolerized antigen [ 16 , 17 ]. In addition, viral antigens such as those associated with influenza, trigger alternate T cell activation pathways [ 18 ] and have been shown to elicit a strong T cell immune response [ 19 ]. Redirecting the vigorous anti-viral T cells which have not been exposed to the immunologic tolerance associated with most tumor-reactive T cells may be effective in eradicating tumor burden. The substantial proliferation in response to strong immunogens such as viral antigens can also be used to improve the localization of T cells that also have anti-tumor activity. Using murine bifunctional T cells created by retroviral transfer of a chimeric immunoglobulin receptor specific for an ovarian cancer-associated tumor antigen to alloreactive T cells, Kershaw and colleagues were able to demonstrate in vivo expansion in response to alloimmunization and demonstrated anti-tumor activity [ 20 ]. It is possible that tumor/viral bifunctional cells would also behave in this way, and we are currently working on murine models with human/mouse chimeric TCRs to test this hypothesis. In addition, some current immunotherapy protocols for the treatment of metastatic melanoma involve immunodepletion prior to adoptive cell transfer [ 21 ]. Such protocols are similar to solid organ and stem cell transplantation in that the patients are temporarily immunosuppressed and at risk for reactivation of latent viruses such as CMV and Epstein-Barr virus. Tumor/viral bifunctional T cells may be particularly useful in this setting, where the anti-viral activity may help treat reactivation, and the reactivation of the virus may further boost the anti-tumor activity of the T cells by inducing additional stimulation of the bifunctional cells. Another consideration to bear in mind with the creation of bifunctional T cells is alternate pairing of the alpha and beta chains resulting in the combination of novel T cell receptors within a bifunctional cell. These T cells could have undesirable autoimmune properties. This could be circumvented by identifying T cells within a bifunctional population that have maximal expression of both the endogenous and introduced TCRs, indicating minimal cross-pairing of chains [ 6 ]. Screening for these T cells and selectively expanding them would reduce the risk of untoward autospecificity. In our experiences, it has been difficult to transduce PBL-derived T cells from normal donors that are stimulated with antigenic peptides derived from self-antigens (data not shown). This is likely due to the low precursor frequency and/or the state of immunologic tolerance of T cells reactive with antigens such as gp100 or tyrosinase [ 20 , 22 , 23 ]. These limitations do not preclude generating T cells capable of recognizing two different tumor antigens, for we have demonstrated here that a T cell clone reactive with gp100:209–217 can be engineered to also recognize MART-1. However, transducing T cell clones is more time consuming since it is first necessary to isolate the T cells clones prior to transduction. There are two potential strategies for overcoming the limitations of transducing T cells with low precursor frequencies or that are immunologically tolerant. First, is transducing actively expanding tumor infiltrating lymphocyte cultures which contain tumor antigen-reactive T cells [24]. Second, patients vaccinated against tumor associated self antigens often have increased frequencies of antigen reactive T cells in their peripheral blood [ 23 ], and these T cells may lend themselves to activation and expansion in vitro to enable efficient retroviral transduction. Conclusion The approach for generating bifunctional T cells we describe in this study may be feasible for viral infections and malignancies and may represent a powerful approach for those patients that otherwise would fail immunotherapy due to the accumulation antigen- or MHC-loss variants. Abbreviations PBMC, peripheral blood mononuclear cells; TCR, T cell receptor; TAA, tumor-associated antigen; CMV, cytomegalovirus Competing interests The authors declare that they have no competing interests. Authors' Contributions AL designed the experiments, performed the transductions, carried out the cocultures and flow cytometry and prepared the manuscript. GC engineered the CMV minigene and the CMV-expressing COS cells, and edited the manuscript. MN conceived of the study, oversaw design and execution of the experiments, and finalized the manuscript.	/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC544585.xml
548142	Asthma is a risk factor for acute chest syndrome and cerebral vascular accidents in children with sickle cell disease	Background Asthma and sickle cell disease are common conditions that both may result in pulmonary complications. We hypothesized that children with sickle cell disease with concomitant asthma have an increased incidence of vaso-occlusive crises that are complicated by episodes of acute chest syndrome. Methods A 5-year retrospective chart analysis was performed investigating 48 children ages 3–18 years with asthma and sickle cell disease and 48 children with sickle cell disease alone. Children were matched for age, gender, and type of sickle cell defect. Hospital admissions were recorded for acute chest syndrome, cerebral vascular accident, vaso-occlusive pain crises, and blood transfusions (total, exchange and chronic). Mann-Whitney test and Chi square analysis were used to assess differences between the groups. Results Children with sickle cell disease and asthma had significantly more episodes of acute chest syndrome (p = 0.03) and cerebral vascular accidents (p = 0.05) compared to children with sickle cell disease without asthma. As expected, these children received more total blood transfusions (p = 0.01) and chronic transfusions (p = 0.04). Admissions for vasoocclusive pain crises and exchange transfusions were not statistically different between cases and controls. SS disease is more severe than SC disease. Conclusions Children with concomitant asthma and sickle cell disease have increased episodes of acute chest syndrome, cerebral vascular accidents and the need for blood transfusions. Whether aggressive asthma therapy can reduce these complications in this subset of children is unknown and requires further studies.	Background Sickle cell disease is a common debilitating hematologic disease occurring in 1 in 650 African Americans. Lung disease is a major cause of cardiopulmonary disability and mortality [ 1 , 2 ]. Progressive restrictive lung disease related to recurrent episodes of acute chest syndrome may develop with advancing age [ 3 ]. Acute chest syndrome is a clinical manifestation triggered by pathological processes including infections, fat embolism, infarction and bronchospasm [ 4 ]. Nearly 70% of patients with acute chest syndrome are hypoxemic (as measured by pulse oximetry < 90% or PO 2 < 80 mmHg by blood gas analysis) [ 5 ]. Hypoxemia causes sickle hemoglobin to gel, inducing red blood cell sickling and vaso-occlusion within pulmonary blood vessels. Asthma is the most common chronic disease of childhood occurring in 6% of the general population. Moreover, the African American population has the highest prevalence (8%), emergency department (ED) visits, hospitalizations and risk of mortality of any ethnic population in the United States [ 6 ]. Reversible airway obstruction, airway inflammation and nonspecific bronchial hyperresponsiveness are the hallmarks of asthma. Exacerbations of asthma result in mucous plugging, bronchoconstriction with decreased air exchange, ventilation-perfusion mismatch and subsequently hypoxemia. Aggressive treatment with oxygen, bronchodilators and oral corticosteroids are recommended for symptomatic relief of acute episodes. The African American population is at significant risk for the occurrence of both diseases simultaneously but little is known of the affect of asthma on individuals with sickle cell anemia. In this study, we examined hospitalized children with sickle cell disease with concomitant asthma establishing whether there was an increased rate of acute chest syndrome or other complications compared to patients with sickle cell disease without asthma in those presenting with vaso-occlusive pain crises. Methods We performed a 5-year retrospective chart review of 48 children with asthma and sickle cell disease and compared them to a control group of 48 children with sickle cell disease alone. The 48 children with sickle cell and asthma represented all patients with complete medical records with both diseases that were cared for at our center. These children had no hospitalizations for sickle cell crises at other facilities. Children in the control group were matched for age, hemoglobinopathy and gender. Sickle cell disease was determined by high performance liquid chromatography and isoelectric focusing analysis and grouped into phenotypes of hemoglobin SS or hemoglobin SC. Children were included in the asthma group if they had medical record documentation of a discharge diagnosis of asthma (ICD-493) and had been prescribed asthma medications. Vaso-occlusive pain was defined as pain that could not be explained by injury or infection requiring hospital admission and treatment with intravenous pain medication. The criteria for a diagnosis of acute chest syndrome included respiratory distress, hypoxemia, and a new infiltrate on chest x-ray that required hospitalization and a transfusion of packed red blood cells. A cerebral vascular accident diagnosis was based on the new onset of an acute neurological syndrome with a focal neurological finding on examination associated with ischemic changes (images compatible with stroke) on a brain MRI or computed tomography scan. Inpatient admissions were recorded from March 1997 to March 2002 for episodes of acute chest syndrome, vasoocclusive pain crises, cerebral vascular accidents, total blood transfusions, exchange transfusions and chronic transfusions (monthly blood transfusions). Exclusion criteria included any child who had incomplete documentation of their hospital records or those who had moved into or out of the Milwaukee area during the 5-year study period. Six children were excluded because they moved from Milwaukee. The study was approved by the Investigational Review Board. Statistical Methods The Mann-Whitney test was used to evaluate differences between the incidents of vasoocclusive pain crises, acute chest syndrome, total blood transfusions and cerebral vascular accidents. Chi-squared analysis was used to evaluate the number of exchange and chronic transfusions. Statistical significance was given to p values 0.05 or less. Results All subjects in the asthma group had asthma recorded in their medical record with symptoms consistent with asthma. All had been prescribed albuterol; while 28 (58%) had been prescribed controller medications (24 inhaled corticosteroids, 12 inhaled cromolyn sodium, and 5 leukotriene modifier). Most patients had been prescribed combination therapy with more than 1 controller medication or had been switched from 1 controller to another. There was no reliable means to confirm adherence to the prescribed asthma drug regimen. Seventeen (35%) subjects had not had prior ED or hospital admissions for asthma. Of these children, 13 (76%) had been prescribed only albuterol. Twelve (25%) had at least 1 ED visit and at least 1 hospital admission for asthma. Twenty-six (54%) subjects had only hospital admissions for asthma, while 15 children had only been treated in the ED for acute asthma. Of 14 children with a severity category documented, 9 had mild intermittent asthma, 1 each with mild persistent and moderate persistent asthma, and 3 with severe persistent asthma. No patient had documented bronchoprovocation with methacholine and only 3 had documented spirometry with 1 consistent with airway obstruction, 1 with reversibility of airway obstruction, and one with poor technique and unreliable results. Two patients (age 3) were too young for spirometry. Four subjects had been seen in the Asthma and Allergy clinic for consultation and the diagnosis of asthma reaffirmed. Twenty-one (44%) subjects had a primary family member with asthma. Only 5 children were born preterm (27, 33, 34, 36, and 37 weeks gestation) and none were diagnosed with bronchopulmonary dysplasia. The cases and controls were well-matched for age, gender and type of hemoglobinopathy (Table I ). The cases (21 males and 27 females) consisted of 42 children with HgbSS and 6 with HgbSC. The control group (17 males and 31 females) included 41 children with HgbSS and 7 with HgbSC. The age of the children ranged from 3 to 18 years old with a mean age of 10.1 years (median 10 years) for the case patients and 10.3 years (median 10 years) for the control children. Table 1 Subject Demographics Patients with Sickle Cell & Asthma Patients with Sickle Cell Disease Males 21 17 Females 27 31 Mean Age 10.1 years 10.3 years HgbSS 42 41 HgbSC 6 7 Patients with sickle cell disease and asthma had significantly more episodes of acute chest syndrome, cerebral vascular accidents, blood transfusions and chronic transfusions as compared to the control group (Table 2 ). Admissions for vaso-occlusive pain crisis and exchange transfusions were not statistically significant between groups. Table 2 Results in children with sickle cell disease with and without asthma Patients with Sickle Cell & Asthma (n = 48) Patients with Sickle Cell Disease (n = 48) Statistical Significance Admissions for Acute Chest Syndrome 90 58 P = 0.03 Number of Cerebral Vascular Accidents 10 2 P = 0.05 Total Blood Transfusions 432 226 P = 0.01 Chronic Blood Transfusions 13 4 P = 0.04 Vaso-occlusive Pain Crises 248 223 P = 0.52 Exchange Blood Transfusions 9 4 P = 0.21 No patient with SC disease in either group had a history of a cerebral vascular accident. Only 1 patient with SC disease had acute chest syndrome, while 3 children in the asthma group had acute chest syndrome with one child experiencing 2 episodes. Discussion Despite the high prevalence of these two diseases that affect the African American population, there is paucity of research investigating patients with concomitant asthma and sickle cell disease. In this study, we discovered that children with both asthma and sickle cell disease are significantly more likely to develop severe complications of sickle cell disease including acute chest syndrome and cerebral vascular accidents compared to children with sickle cell disease alone. The significance of these findings relates to the hypothesis that appropriately aggressive treatment of asthma in children with sickle cell disease may diminish the frequency of pulmonary complications. Patients with reactive airway disease and sickle cell disease have a lower transcutaneous oxyhemoglobin saturation [ 7 ]. The lower transcutaneous oxyhemoglobin saturation increases sickling of red blood cells, causing subsequent gelling and vaso-occlusion in multiple organs leading to numerous complications. A high prevalence (73%) of airway hyperreactivity to cold-air challenge occurs in children with sickle cell anemia even in the absence of clinical symptoms of asthma [ 8 ]. Cold air or other provocative challenge tests had not been performed in our group of patients. Our results are in agreement with an abstract that showed 18 children with both asthma and sickle cell disease had increased hospital admissions for pain crises and acute chest episodes compared with patients with only sickle cell disease [ 9 ]. Our results are unique because we assessed for an increased frequency of transfusions and cerebral vascular accidents in a larger group of patients. Cerebral vascular accidents affect 10% of the children with sickle cell disease often with devastating long term implications. The Cooperative Study of Sickle Cell Disease reported that recent and recurrent episodes of acute chest syndrome are risk factors for cerebral vascular accidents [ 10 ]. Our findings of increased cerebral vascular accidents in patients with sickle cell disease and asthma may be due to their increased and recurrent episodes of acute chest syndrome. The mechanism linking stroke and acute chest syndrome is unknown but may be related to hypoxemia due to pulmonary disease. Hypoxemia has been reported to increase adhesion of red blood cells to endothelium [ 11 ]. Likewise, increased red blood cell adhesion to pulmonary endothelium has been associated with bone marrow vaso-occlusive crises due to hypoxia, cytokine expression and fat embolization [ 12 ]. Whether hypoxemia secondary to asthma affects red blood cell and pulmonary/cerebral endothelium adhesion characteristics is unknown. Our study demonstrates that pediatric patients with sickle cell disease and asthma are more likely to require acute and chronic blood transfusions. It is not surprising that these patients needed more frequent transfusions since treatment for cerebral vascular accidents and severe acute chest syndrome is blood transfusions. Although SS type of sickle cell anemia is more severe than SC type, the small numbers of patients with SC prevent us from making broad conclusions regarding the degree of risk based on hemaglobinopathy. Two studies have demonstrated increased airway hyperreactivity with reversibility in patients with sickle cell disease without known asthma [ 8 , 13 ]. Thirty-five percent of sickle cell patients had evidence of lower airway obstruction and 78% of these reversed with bronchodilator. Even in 30% of those with normal lung function, 30% had a positive response to bronchodilator. Therefore, even methods to determine the presence bronchial hyperresponsiveness are not sufficient to discriminate asthma from acute chest syndrome. Whether this hyperreactivity is due to asthma or is secondary to the pathophysiology of sickle cell disease is still unclear. Additional studies are needed to confirm that sickle cell disease is associated with the development of reversible airway obstruction. If the association is valid, the effects of routine use of anti-inflammatory controller agents prophylactically or therapeutically would deserve investigation. A randomized, placebo-controlled trial of 43 episodes of acute chest syndrome in 38 children revealed that intravenous dexamethasone prevented clinical deterioration in mild to moderately severe episodes of acute chest syndrome [ 14 ]. Mild and moderately severe acute chest syndrome was defined as respiratory distress, and normal mental status without pulmonary infiltrates or arterial hypoxemia. The study excluded children with an exacerbation of reactive airways disease. If patients with sickle cell disease are at increased risk of airway inflammation and obstruction, successful treatment with intravenous steroids may have been due to aggressive treatment of underlying asthma. Limitations of our study include those inherent in a retrospective analysis including selection bias, measurement bias and confounding factors. The children who received transfusions for acute chest syndrome likely represent the more severe form of disease. In contrast, those patients deemed to have asthma were primarily being treated as if they had mild to moderate asthma. Only 3 had been diagnosed with severe persistent asthma and none were on daily or every-other-day oral steroids. Importantly, even individuals with mild intermittent asthma can experience severe asthma exacerbations. Other sources of potential error that deserve recognition include the diagnosis of asthma (whether over-diagnosed or under-diagnosed), medication compliance and unknown admission to other hospitals. Although a strict definition of asthma was lacking, the history gleaned from the medical records supported an asthma diagnosis. Most patients with asthma are not cared for by asthma specialists and frequently the diagnosis is made on clinical grounds without formal pulmonary function testing. Additionally, spirometry was not routinely performed during ED and hospital admissions. Although other EDs and hospitals in the metropolitan area evaluate, treat, and admit pediatric patients with asthma exacerbations, the concomitant diagnosis of sickle cell disease likely prompted evaluation at the children's hospital. Conclusions In this small series, children with a history of asthma and sickle cell disease developed acute chest syndrome and cerebral vascular accidents more frequently than children with sickle cell disease without asthma. The implications of this retrospective study are wide ranging and should lead to further prospective investigations. Whether aggressive asthma therapy in patients with sickle cell disease and asthma reduces the incidence of serious complications is unknown. The potential gains are far reaching and could make enormous impacts on the morbidity, quality of life and mortality of many patients. Competing interests The author(s) declare that they have no competing interests. Authors' contributions MEN conceived the study, participated in its design and coordination and drafted the manuscript. JL participated in data collection. MCZ participated in coordination and manuscript preparation. JPS participated in design, coordination and manuscript preparation. KJK participated in design and coordination.	/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC548142.xml
524517	Persistent expression of chemokine and chemokine receptor RNAs at primary and latent sites of herpes simplex virus 1 infection	Inflammatory cytokines and infiltrating T cells are readily detected in herpes simplex virus (HSV) infected mouse cornea and trigeminal ganglia (TG) during the acute phase of infection, and certain cytokines continue to be expressed at lower levels in infected TG during the subsequent latent phase. Recent results have shown that HSV infection activates Toll-like receptor signaling. Thus, we hypothesized that chemokines may be broadly expressed at both primary sites and latent sites of HSV infection for prolonged periods of time. Real-time reverse transcriptase-polymrease chain reaction (RT-PCR) to quantify expression levels of transcripts encoding chemokines and their receptors in cornea and TG following corneal infection. RNAs encoding the inflammatory-type chemokine receptors CCR1, CCR2, CCR5, and CXCR3, which are highly expressed on activated T cells, macrophages and most immature dendritic cells (DC), and the more broadly expressed CCR7, were highly expressed and strongly induced in infected cornea and TG at 3 and 10 days postinfection (dpi). Elevated levels of these RNAs persisted in both cornea and TG during the latent phase at 30 dpi. RNAs for the broadly expressed CXCR4 receptor was induced at 30 dpi but less so at 3 and 10 dpi in both cornea and TG. Transcripts for CCR3 and CCR6, receptors that are not highly expressed on activated T cells or macrophages, also appeared to be induced during acute and latent phases; however, their very low expression levels were near the limit of our detection. RNAs encoding the CCR1 and CCR5 chemokine ligands MIP-1α, MIP-1β and RANTES, and the CCR2 ligand MCP-1 were also strongly induced and persisted in cornea and TG during the latent phase. These and other recent results argue that HSV antigens or DNA can stimulate expression of chemokines, perhaps through activation of Toll-like receptors, for long periods of time at both primary and latent sites of HSV infection. These chemokines recruit activated T cells and other immune cells, including DC, that express chemokine receptors to primary and secondary sites of infection. Prolonged activation of chemokine expression could provide mechanistic explanations for certain aspects of HSV biology and pathogenesis.	Introduction Acute viral infections are usually cleared from the primary site of infection by the host immune response [ 1 ], but some viruses can persist at other sites in a latent form. Herpes simplex virus (HSV), for example, causes a primary infection at a mucosal site, which is cleared within 7–10 days by the host immune response. HSV, nevertheless, enters sensory neurons and establishes a latent infection within those cells. In a mouse corneal model of HSV-1 infection, infectious virus is detected in corneal secretions and tissue for approximately 7 days [ 2 ]. Similarly, infectious virus is detected in trigeminal ganglion (TG) tissue for up to approximately 10 days [ 2 ]. Latent infection is established by 30 days postinfection (dpi) because no infectious virus can be detected in homogenates of TG tissue at that time. HSV DNA, however, is readily detected in latently infected TG for at least 150 dpi [ 3 - 5 ]. Viral gene expression is greatly attenuated during latent infection because the only abundant viral gene product detected is the latency-associated transcript or LAT [ 6 ]. Nevertheless, low levels of lytic transcripts can be detected in ganglia latently infected with HSV [ 5 ]. Evidence of viral protein expression is provided by the continued T cell infiltration [ 7 , 8 ], elevated levels of interferon γ (IFN-γ) and TNF-α transcripts and numbers of IL-6 expressing cells in the ganglia, [ 3 , 9 - 11 ]. Expression of IFN-γ and TNF-α transcripts persists in TG latently infected with HSV strains unable to replicate in neurons, indicating that neither HSV replication nor ability to reactivate are required for persistent cytokine gene expression [ 3 ]. While CD4 + T cells appear to be important in immunized mice for protection against challenge virus infection [ 12 ], CD8 + T cells appear to be important for establishment of latent infection in mice [ 7 ]; and CD8 + T cells specific for HSV persist in TG for long periods of time [ 8 ]. Thus, there is evidence for long-term immune surveillance in the ganglion during latent infection by HSV. Chemokines are critical for recruiting inflammatory cells to infected tissues. Chemokine specificity is due in large part to the cell-specific expression of their respective receptors (reviewed in [ 13 - 15 ]. Inflammatory-type receptors including CCR1, CCR2, CCR5, and CXCR3 are expressed by activated T cells, macrophages, natural killer (NK) cells, and immature ( i.e . potent for antigen capture but not antigen presentation) dendritic cells (DC), while homostatic-type receptors including CCR7 and CXCR4 are highly expressed by resting T and B cells and mature ( i.e ., antigen-presenting) DC (Table 1 ). In addition, receptors including CCR2, CCR5 and CXCR3 are expressed on cells ( e.g. Th1 cells) specific for infection-induced inflammation, while others including CCR3 and CXCR4 are on cells ( e.g. , Th2 T cells) associated with allergic inflammation. Certain receptors are expressed by specific subsets of a given cell type. For example, CCR6 is highly expressed on Langerhans-like (CD34 + ) DC that migrate to skin, but not on monocyte-derived DC that migrate to non-skin tissues (reviewed in [ 14 ]. Acute viral infection in the mouse corneal model system is known to induce the expression of cytokines and chemokines in corneal tissue. Thomas et al. [ 16 ] observed the induction of transcripts encoding N51/KC, macrophage inflammatory protein-1 β (MIP-1β), MIP-2 and monocyte chemotactic protein 1 (MCP-1) and the cytokines IL-1, IL-6, IL-12, and TNF-α. Similarly, Tumpey et al. [ 17 ] showed induction of MIP-2, MIP-1α, and MCP-1 chemokines in the cornea during acute infection. Infection of mouse fibroblast cells by HSV induces expression of IL-6 [ 18 ], and infection of macrophages by HSV induces RANTES expression directly [ 19 ]. Infection of other cell types may induce expression of other cytokines and chemokines. Less is known about chemokine expression during HSV latent infection phase. Halford et al. [ 10 ] observed RANTES RNA expression, in addition to RNAs for IL-2, TNF-α, IFN-γ, and IL-10, during latent infection. Table 1 Expression of Chemokine Receptors, Chemokines and Cytokines in Leukocyte Populations Chemokine receptors Cell type expression Chemokine ligand Proposed primary function(s) CCR1 T cells, macrophages, immature dendritic cells (DC), natural killer cells (NK) RANTES; MIP-1α; MCP-3, and 4; HCC-1, 2, and 4 Migration of DC to sites of inflammation Recruitment of T cells, macrophages and NK CCR2 T cells, natural killer cells (NK), macrophages, immature DC MCP-1, 3, and 4 Migration of effector T cells (Th1) Migration of DC progenitors to sites of inflammation CCR3 eosinophils, basophils, T cells eotaxin-1 and 2; RANTES; MCP-2, 3, and 4; HCC-2 Recruitment of eosinophils CCR5 T cells (Th1, Tc1), macrophages, immature DC RANTES; MIP-1α and 1β Migration of effector T cells (Th1) Migration of DC to sites of inflammation Recruitment of macrophages CCR6 immature DC (CD34+/Langerhans-like), T cells MIP-3α Migration of DC to skin CCR7 T cells, B cells, mature DC SLC, ELC Migration of naïve T cells to lymph nodes Migration of memory T cells to lymphoid tissue Migration of B cells Migration of DC to lymphoid tissues CXCR3 T cells (Th1, Tc1) IP-10, MIG, ITAC Migration of effector T cells (Th1) CXCR4 T cells, macrophages, DC, B cells, others including neurons SDF-1 Migration of effector T cells (Th2) Migration of B cells Migration of hematopoietic progenitors Chemokines Receptor MIP-1α T cells, NK, macrophages, others CCR1, CCR5 Chemoattract macrophages, T cells, NK, and others MIP-1β T cells, NK, macrophages, others CCR5, CCR1 (weak) Chemoattract macrophages, T cells, and others RANTES T cells, NK CCR1, CCR5, CCR3 (weak) Chemoattract T cells and others MCP-1 macrophages, others CCR2 Chemoattract macrophages, T cells, NK, and others Eotaxin-1 epithelial cells, NK, macrophages, others CCR3 Chemoattract eosinophils Cytokines Receptor IFN-γ T cells, NK IFN-γR Activation of antiviral response TNF-α macrophages, NK, others TNF-R Broad activation of antiviral and inflammatory response Recent studies have shown that HSV infection activates Toll-like signaling and chemokine synthesis [ 20 , 21 ]. Thus, we hypothesized that HSV infection might induce prolonged expression of a broad range of chemokines at sites of acute and latent infection. Real-time quantitative RT-PCR methods have facilitated studies of immune cell RNA expression in mouse models [ 22 , 23 ]. We report here the use of real-time RT-PCR to monitor RNA expression of selected chemokine receptors and their chemokine ligands during HSV infection of mouse corneal and TG tissue. Our data show that RNA encoding inflammatory-type chemokine receptors and their ligands persists in infected corneas and TG long after infectious virus can be detected, suggesting prolonged chemokine production and subsequent homing of inflammatory immune cells to these tissues. Strikingly, the data demonstrate the persistent expression of chemokines and chemokine receptor genes in the apparent absence of detectable viral productive infection transcripts in infected corneas. Results Development of TaqMan ® RT-PCR assays to measure viral and host gene expression during acute and latent infection To monitor RNA expression of viral and host genes during HSV infection of mice, we developed TaqMan ® RT-PCR assays for the quantification of transcripts from the HSV tk and ICP0 genes and from mouse genes encoding selected chemokine receptors and their ligands. In the real-time PCR assay detailed in Materials and Methods, RNA isolated from corneal and ganglionic tissue was used for synthesis of cDNA. Primers and Taqman ® probes for the viral or cellular genes (Table 2 ) were used in real-time PCR assays to measure the concentration of cDNA for each transcript. Table 2 Primer and Probe Sequences Forward Primer Reverse Primer Probe* HSV tk CGAGACAATCGCGAACATCTAC CCCCGGCCGATATCTCA CCACACAACACCGCCTCGACCA ICP0 CTGCGCTGCGACACCTT CAATTGCATCCAGGTTTTCATG TGCATGCACCGCTTCTGCATCC Chemokine receptor CCR1 GGGTGAACGGTTCTGGAAGTAC CAGCCATTTTGCCAGTGGTA ACATGCCTTTGAAACAGCTGCCGAA CCR2 ATGAGTAACTGTGTGATTGACAAGCA GCAGCAGTGTGTCATTCCAAGA CTCTGTCACCTGCATGGCCTGGTCT CCR3 ACCAGCTGTGAGCAGAGTAAACAT CACAGCAGTGGGTGTAGGCA CACCTCAGTCACCTGCATGGCCA CCR5 ACTGCTGCCTAAACCCTGTCA GTTTTCGGAAGAACACTGAGAGATAA TCCGGAACTTCTCTCCAACAAAGGCA CCR6 TTGGTGCAGGCCCAGAAC GAACACGAGAACCACAGCGAT CCAAGAGGCACAGAGCCATCCGA CCR7 CTGCTACCTCATTATCATCCGTACCT TGATCACCTTGATGGCCTTGT CTCCAGGCACGCAACTTTGAGCG CXCR3 TGTAGTTGGGCTAGCTCGAACTT ACCTGGATATATGCTGAGCTGTCA GCATCCTGGCAGCAAAGTTACGGG CXCR4 CTCCAAGGGCCACCAGAA GGCAAAGAAAGCTAGGATGAGG CGCAAGGCCCTCAAGACGACAGTC Chemokine MIP-1α TCATCGTTGACTATTTTGAAACCAG GCCGGTTTCTCTTAGTCAGGAA AGCCTTTGCTCCCAGCCAGGTGTC MIP-1β AGGGTTCTCAGCACCAATGG GCTGCCGGGAGGTGTAAGA CTCTGACCCTCCCACTTCCTGCTGTTT RANTES CTGTCATCGCTTGCTCTAGTCCTA CGGATGGAGATGCCGATTT ATCCCCTACTCCCACTCCGGTCCTG MCP-1 GCTGGGTTCAGTTTCCTTAAGC CCTAGTCTTTAGCTGTGAGACCTTCTG AGGCCTCGCTGCTCCACATCCA Eotaxin-1 CCTAAGACGTGCTCTGAGGGAAT TCCCATCTGGAACTACATGAAGC TCAGCACCAGTCGCCCAAGGACT Cytokine IFN-γ TGAGTATTGCCAAGTTTGAGGTCA GTGGACCACTCGGATGAGCT CCACAGGTCCAGCGCCAAGCA TNF-α ACAAGGCTGCCCCGACTAC CGCAGAGAGGAGGTTGACTT CCTCACCCACACCGTCAGCCG * all probes FAM-5' and 3'-TAMRA To characterize the range over which the HSV tk and ICP0 real-time PCR assays were accurate and linear, we tested 10-fold dilutions of purified HSV genomic DNA (kind gift of Jean Pesola) starting from 5.5 × 10 4 copies for tk and ICP0 gene levels. The HSV tk and ICP0 primer/probe sets gave linear amplification curves over 4 logs of template concentrations until the limit of detection within the linear range was reached at 55 DNA copies for tk and 550 copies for ICP0 (not shown). At these limits of detection, the threshold cycle (CT) value, which indicated the PCR cycle at which a significant increase in amplification was first detected, was 39.2 for tk at 55 DNA copies and 36.5 for ICP0 at 550 DNA copies. Using 2-fold dilutions of uninfected mouse TG cDNA, we observed that the primer/probe sets for host genes listed in Table 2 including GAPDH gave linear amplification curves over at least 3 and up to 7 dilutions. In all cases, CT values changed by about 1 cycle for every 2-fold change in template concentration as expected (not shown). Thus our assays matched well with previously described TaqMan ® assays [ 22 - 24 ] for linearity and sensitivity. Following corneal inoculation of mice with HSV or virus diluent (mock), we collected corneas and TG during acute (3 and 10 dpi) and latent (30 dpi) phases. To monitor viral gene expression in infected mice, we tested tissue samples for tk and ICP0 gene transcripts. In infected corneal tissue, HSV tk and ICP0 transcripts were readily detected at 3, but not at 10 or 30 dpi where CT values = 40 (indicating no measurable RNA) (Fig. 1 ). Thus we could not detect lytic transcripts in infected corneas beyond the acute phase using this assay. Figure 1 HSV tk and ICP0 RNA expression in mock and HSV-infected cornea and TG. RNA isolated from tissues harvested at 3, 10, or 30 days postinfection (d) was subjected to TaqMan RT-PCR analysis using HSV tk primers/probe (A) and HSV ICP0 primers/probe (B) as described in Materials and Methods. Mouse GAPDH RNA was measured in multiplex reactions, and used to calculate relative expression using the formula Rel Exp= 2 -(ΔΔCT) × 1000 as described in Materials and Methods. Shown below the plots are relative expression values and the CT value measured for tk (A) and ICPO (B) in each sample. The ICP0 signal detected at 10 and 30 dpi in HSV-infected TG is likely due to LAT RNA as described in the text. Results shown are for one experiment (Experiment #1) in which the number of individual mouse tissues pooled were 10 for cornea and 6 for TG. Similar results were obtained in two additional experiments (Experiment #2 and Experiment #3), except for variation in detection of tk RNA in infected TG at 30 dpi as described in the text. In infected TG, tk RNA peaked at 3 dpi then dropped precipitously (200-fold) to low but readily detectable levels by 10 dpi. At 30 dpi, we detected very low or undetectable tk RNA expression in infected TG. In the experiment shown in Fig. 1A , we measured a CT value of 38.2 for tk expression in infected TG at 30 dpi, resulting in a relative expression value of 0.0002. In an independent experiment, we measured a CT of 38.1 for tk RNA in 30 dpi TG; however, a CT value of 40 was measured in two additional experiments (not shown). CT values for all reactions without RT were 40, indicating no DNA contamination. Thus, while tk expression in latent TG was at the limit of detection for our assay, our ability to detect tk expression in some but not all latent TG was consistent with previous reports in which very sensitive RT-PCR assays were used to detect tk (and ICP0 ) gene transcripts in some but not all TG during latent infection [ 5 , 25 ]. In those previous reports, an assay that included a radioactive Southern blotting step subsequent to RT-PCR could detect single copies of tk nucleic acid per PCR reaction. Our present assay for tk transcripts is at least 50-fold less sensitive than that used by Kramer and Coen [ 5 ]. ICP0 RNA levels were similar to tk in that they peaked at 3 dpi in cornea and TG (Fig. 1B ). However, because our ICP0 probe/primer set overlaps latency-associated transcript minor (LAT) – coding sequences, the signal detected at 10 and 30 dpi in TG but not cornea may be due to minor LAT read-through RNAs. RT-PCR analysis of LAT transcripts from the TGs at 30 dpi was consistent with latent virus in infected TG (unpublished results). Chemokine and chemokine receptor expression in infected cornea and ganglia We next used TaqMan ® RT-PCR to monitor expression of a selected series of mostly T cell and macrophage-specific chemokine receptors and chemokines in mock and HSV-infected cornea and TG. We chose chemokine receptors CCR1, CCR2, CCR5, and CXCR3, which are expressed by activated T cells, macrophages, NK cells, and immature DC that would be part of the immune infiltration in response to HSV infection, and their ligands MIP-1α, MIP-1β, RANTES, and MCP-1. For comparison, we included CCR3 which is primarily expressed on granulocytes, the CCR3 ligand eotaxin-1, CCR6 which is primarily expressed on resting T cells and immature Langerhans-like ( i.e., skin homing) DCs, CCR7 which is primarily expressed on resting T and B cells and mature DCs that home back to lymphoid tissues, and CXCR4 which is broadly expressed on many immune and non-immune cell types (Table 1 ). We also tested the chemokine-inducing cytokines IFN-γ and TNF-α, whose RNA and protein have previously been shown to be expressed during both acute and latent phases of HSV infection [ 3 , 9 - 11 ]. i. Chemokine and chemokine receptor expression in infected cornea Epithelial cells of the cornea are the initial sites of replication following infection but infectious virus and viral mRNAs are not detectable past 7–10 dpi [ 26 ]. We harvested RNA from mock and HSV-infected cornea at 3, 10, and 30 dpi, and tested for chemokine receptor and chemokine RNA expression in parallel. As expected for tissues supporting active replication or having recently cleared virus, chemokine receptors CCR1, CCR2, CCR5, CCR7, CXCR3 and CXCR4, but not CCR3 or CCR6, were highly expressed and strongly induced ( i.e. , >3-fold) at 3 and 10 dpi (Fig. 2 and Table 3 ). Chemokines MIP-1α, MIP-1β, RANTES, and MCP-1, but not eotaxin-1, were also highly expressed and strongly induced in infected cornea at 3 and 10 dpi. IFN-γ and TNF-α were also induced in infected cornea as previously reported [ 16 ]. Surprisingly, induction of all host RNAs tested persisted into latent phase at 30 dpi in infected corneas. For example, CCR1, CCR2, and CCR5 exhibited similar induction and similar or only slightly reduced expression levels at 30 dpi as compared to earlier time points. Relative expression and induction of CCR7 and CXCR4 in infected cornea appeared to be biphasic in that values were high at 3, lower at 10, and higher again at 30 dpi. These results suggested that continued presentation of HSV antigens stimulates chemokine production and subsequent homing of effector cells to cornea despite the apparent clearance of infectious virus. Figure 2 Relative levels of chemokine and chemokine receptor RNA expression in mock and HSV-infected cornea. Corneas were harvested at 3 (A), 10 (B), or 30 (C) days postinfection, and relative levels of expression were determined by TaqMan RT-PCR analysis as described in Fig. 1 and Materials and Methods. Results shown are the average of relative expression values determined using cDNA from two independent experiments, with each cDNA subjected to 2 or 3 separate measurements. Dashed bars represent ranges of individual values. Each cDNA was synthesized from RNA isolated from pooled corneas (5 mice) as described in Fig. 1 and Materials and Methods. The induction ratios (HSV+ vs. mock) for individual genes are tabulated in Table 3. Table 3 Induction Ratio (HSV+/Mock) of Transcripts for Chemokine Receptors, Chemokines and Cytokines in Cornea and Trigeminal Ganglia (TG) Cornea a TG b Gene 3d 10d 30d 3d 10d 30d CCR1 11 (9.2–12) 18 (13–23) 20 (10–26) 5 (2.2–7.1) 15 (9.0–19) 4 (1.7–7.0) CCR2 14 (9.2–19) 22 (11–32) 14 (8.1–24) 3 (1.5–4.2) 15 (11–19) 3 (1.3–4.4) CCR3 2 (1.0–5.0) 3 (2.0–5.0) 3 (2.5–3.3) 2 (0.5–5.0) 8 (2.2–20) 3 (1.8–4.7) CCR5 12 (12.3–12.5) 11 (8.0–14) 20 (8.9–36) 9 (4.8–11) 57 (22–110) 9 (7.0–10) CCR6 3 (2.4–3.0) 2 (1.0–2.5) 5 (1.5–8.5) 3 (0.3–11) 3 (1.0–5.0) 14 (1.0–40) CCR7 24 (8.0–40) 5 (3.0–6.5) 17 (13–21) 13 (9.0–17) 19 (17–20) 7 (2.0–11) CXCR3 10 (5.0–18) 15 (8.0–23) 5 (2.8–6.5) 2 (1.0–4.0) 104 (54–160) 36 (14–59) CXCR4 11 (4.8–14) 3 (1.7–4.0) 45 (33–74) 0.6 (0.4–0.9) 4 (2.9–6.2) 3 (2.3–3.7) MIP-1α 69 (33–106) 394 (263–1700) 34 (16–53) 232 (80–471) 126 (80–168) 25 (13–45) MIP-1β 53 (39–67) 285 (261–310) 16 (11–21) 282 (10–595) 230 (202–245) 31 (24–37) RANTES 55 (36–73) 43 (38–48) 16 (12–18) 64 (61–66) 304 (302–306) 31 (12–50) MCP-1 54 (52–55) 64 (55–74) 12 (7.5–20) 153 (113–194) 22 (16–27) 3 (1.6–4.2) Eotaxin-1 3 (1.9–3.5) 1 (0.6–1.3) 3 (1.0–5.4) 5 (3.3–9.1) 2 (1.2–2.8) 1.5 (0.7–2.3) IFNγ Inf. c Inf. Inf. Inf. Inf. Inf. TNF-α 3 (2.9–3.0) 3 (2.6–3.8) 7 (3.9–12) Inf. Inf. Inf. a Induction ratios were calculated as relative expression in HSV-infected/relative expression in mock-infected cornea. Each value is the average of induction ratios (2 or 3 separate measurements per cDNA sample) from two independent experiments. Ranges of individual ratios are in parentheses. b Induction ratios were calculated for HSV- vs. mock-infected TG as in footnote a . Each value is the average of induction ratios (2 or 3 separate measurements per cDNA sample) from three independent experiments, with ranges in parentheses. c Inf., infinite due to relative expression = 0 in all or most mock-infected samples. ii. Chemokine and chemokine receptor expression in infected ganglia In infected TG, transcripts from the genes encoding receptors CCR1, CCR2, CCR5, CCR7, and CXCR3 were induced by HSV infection during both acute (3 and 10 dpi) and latent (30 dpi) phases (Fig. 3 and Table 3 ). Peak induction of these RNAs was at 10 dpi during the clearance phase. CXCR4 was induced at 10 and 30 dpi but not at 3 dpi. While we measured induction of CCR3 and CCR6 at 10 and 30 dpi, their very low expression was at the limit of our detection (i.e., relative expression values < 0.5) as also seen in corneas. RNAs for the MIP-1α, MIP-1β, RANTES, and MCP-1 chemokines were also strongly induced at each timepoint, particularly at 3 dpi. Eotaxin-1 was induced at 3 dpi, but much less so at 10 and 30 dpi. As seen previously [ 3 ] cytokines IFN-γ and TNF-α were strongly induced at 3 and 10 dpi, but much less so at 30 dpi. Figure 3 Relative levels of chemokine and chemokine receptor RNA expression in mock and HSV-infected TG. TG were harvested at 3 (A), 10 (B), or 30 (C) days postinfection, and RNA levels were determined by TaqMan RT-PCR analysis as described in Fig. 1, Fig. 2 and Materials and Methods. Results shown are the average of relative expression values determined using cDNA from three independent experiments, with each cDNA subjected to 2 or 3 separate measurements. Dashed bars represent ranges of individual values as described in Fig. 2. The induction ratios (HSV+ vs. mock) for individual genes are tabulated in Table 3. A striking finding in this analysis was the persistent expression of inflammatory cell RNAs during the latent phase of TG infection when detectable production of infectious virus has ceased. To determine if induction of these RNAs persisted past 30 dpi, we monitored expression of a limited number of transcipts from in TG collected at 45, 62, and 90 dpi. In previous studies [ 3 - 5 ], HSV genomic DNA was maintained at constant levels (~10 4 copies per TG) for up to 150 dpi in infected TG, indicating that latent virus persists well beyond 90 dpi in this mouse model. Induction of all RNAs in our panel persisted for at least 62 dpi; furthermore, all but CCR3 and eotaxin-1 were also induced at 90 dpi (Table 4 ). Thus chemokine receptor and ligand expression persisted long into the latent phase in infected TG. Table 4 Induction Ratio (HSV+/Mock) of Transcripts for Chemokine Receptors and Chemokines in Trigeminal Ganglia (TG) at Late Times Post-Infection Induction Ratio a Gene 45d 62d 90d CCR2 3 (3.2–3.3) 5 (1.3–8.8) 2 (2.2–2.4) CCR3 8 (5.0–12) 3 (1.0–4.4) 0.7 (0.4–1.0) CCR5 5 (5.1–5.7) 7 (4.9–9.0) 5 (2.9–6.5) CXCR3 17 (10–24) 68 (25–111) 20 (11–28) MIP-1α 10 (7.0–13) 35 (4.0–67) 4 (1.0–7.0) Eotaxin-1 3 (1.5–3.9) 2 (1.1–3.1) 1.5 (0.8–2.3) a Induction ratios were calculated as relative expression in HSV-infected/relative expression in mock-infected cornea as described in Table 3. Each value is the average induction ratio (2 separate measurements per cDNA sample) from one experiment. Ranges of individual ratios are in parentheses. Discussion Recent studies have shown that HSV infection induces Toll-like signaling and chemokine synthesis. Thus, we hypothesized that HSV infection might induce a broad range of chemokines at sites of primary and latent infection. In agreement with and extending previous studies [ 3 , 9 - 11 ], we have found evidence for persistent expression of chemokines and trafficking of inflammatory cells including activated T cells to acutely infected corneal tissue and to latently infected trigeminal ganglia. We also observed prolonged expression of chemokine and chemokine receptor gene transcripts in corneal tissue, the primary site of HSV-1 infection in this model system, long after infectious virus has been cleared. Microarray analysis of host gene expression has also demonstrated long-term alterations of host gene expression during latent infection by HSV, including alterations in expression of CXCR6 mRNA in TG [ 27 ]. These results argue for long-term persistence or expression of viral antigens or immunogens and stimulation of expression of these chemokines, even at the primary site of infection, the cornea. Recent results [ 28 ] have shown similar elevated chemokine expression in lung tissue after clearance of murine gamma herpesvirus 68. It will be of interest to determine how widespread this effect is among different virus infections or whether it is unique to viruses that persist in the host, such as the herpesviruses. Potential mechanisms for elevated expression of chemokines and chemokine receptors after viral clearance Low level expression of viral lytic transcripts in TG during latent infection has been documented [ 5 ], which could result in low level expression of viral proteins. Recent results have shown that HSV-1 can activate Toll-like receptor 2 to stimulate chemokine expression and secretion and to activate NF-κB regulated promoters [ 20 ]. Lund et al. [ 21 ] showed that infectious HSV-2 and also purified HSV-2 DNA activates signaling through DC-expressed Toll-like receptor 9, resulting in the induction of IFN-α secretion. Toll-like receptor activation by HSV-2 DNA raises the intriguing possibility that HSV DNA alone is at least partially responsible for TLR-dependent induction of chemokine expression in latent TG. Among the transcripts that we studied, we detected persistent expression of transcripts for MIP-1α, MIP-1β, and RANTES, whose expression is activated by Toll-like receptors [ 29 ]. Expression of MIP-1α and MIP-1β could recruit NK cells, which express CCR5, and immature dendritic cells, which express CCR1 and CCR5, into the site of infection. Thus, elevated expression of at least some of the chemokines could be due to Toll-like receptor activation. It is also possible that other chemokines that were not assayed in this or previous studies are also induced during latent HSV infection via Toll-like receptor dependent mechanisms. Elevated expression of chemokine receptors is likely due to the chemokine-induced trafficking of inflammatory cells to the site of infection or, in the case of 30 days postinfection or latent infection, the site of viral antigen persistence. Although we have not examined expression of IP-10, a chemokine also induced by Toll-like receptor signaling [ 29 ], we did examine the expression of transcripts for CXCR3, its receptor on activated T cells. Levels of both are elevated during latent infection in TG. Thus, stimulation of expression of this chemokine could attract activated T cells to the latently infected TG, providing a mechanism for the persistent presence of HSV-specific CD8 + T cells in latently infected TG [ 8 ]. Implications of persistent chemokine expression Long-term inflammatory responses in neural tissue could induce pathology due to damage to neuronal cells. A number of neurological diseases have been associated with HSV infection [ 30 ], and these could be associated with these long-term inflammatory responses. In addition, the possibility of other types of specific pathological effects is raised. Role of HSV in coronary heart disease Recent data have shown an association between HSV-1 seropositivity and myocardial infarction and coronary heart disease in older adults [ 31 ]. These authors hypothesized that HSV-1 reactivation from autonomic nerves that innervate the coronary arteries could cause infection of endothelial cells, endothelial injury, and the initiation of an acute thrombotic event. Similarly, based on our work, HSV infection might induce expression of MCP-1 and IL-8, which are known to cause adhesion of monocytes to vascular endothelium [ 32 ], an early step in the development of atherosclerotic lesions in mouse models (reviewed in Gerszten et al. [ 32 ]. Therefore, the induction and prolonged expression of these chemokines by HSV infection could play a role in the pathogenesis of coronary heart disease. Role of HSV in HIV transmission Considerable evidence has accumulated for the role of genital herpes infections in promoting the transmission of human immunodeficiency virus (reviewed in [ 33 ]. Although we examined HSV-1 in these studies, HSV-2 shares many biological properties with HSV-1. Thus, it is conceivable that genital herpes infections could similarly induce the expression of chemokines in the genital mucosae and the trafficking of dendritic cells and CD4 + T cells to that site. In addition to the break in the genital epithelium provided by the genital lesion, the recruitment of dendritic cells and CD4 + T cells to sites of HSV infection would provide cells to transport HIV to lymph nodes and the primary host cell, respectively, and increase the potential for HIV infection. Implications for HSV biology and vaccine design Recent studies on the persistence of CD8 + T cells in latently infected ganglia have concluded that these cells play a role in maintaining the latent infection [ 8 ]. The results presented here raise the possibility that the presence of CD8 + T cells in latently infected TG's could be the result of chemokine expression. Thus, further studies are needed to establish the causal relationship between the presence of CD8 + T cells in latently infected ganglia and maintenance of latent infection. Various HSV strains, including replication-defective mutants and amplicon vectors which do not establish neuronal latency efficiently, have been shown to induce durable immune responses [ 12 , 34 , 35 ]. These results suggest that the basis for the durable immune responses may be the persistence of antigen or continued antigen expression at sites of primary infection. Further studies are needed to determine the source of this antigen and the mechanism of the induction of chemokine expression at primary and latent sites of HSV infection. Materials and Methods Viruses, infection of mice, and tissue collection HSV-1 KOS was propagated and titered on Vero cell monolayers as described previously [ 36 ]. Seven-week-old HSD:ICR mice (Harlan, Sprague, Dawley) were anesthetized and infected with 2 × 10 6 pfu of virus or mock infected with virus diluent via corneal scarification as described [ 2 ]. At specific days post infection (dpi), cornea and TG were collected and flash-frozen on dry ice with minimal elapsed time post sacrifice [ 5 ]. Cornea and TG from each time and treatment group were pooled prior to isolation of RNA. A total of four infections were performed: in Exp. #1 cornea and TG were collected at 3, 10, and 30 dpi; in Exp. #2 TG were collected at 3, 10, and 30 dpi; in Exp. #3 TG were collected at 3, 10, 45, 62, and 90 dpi; and in Exp. #4 cornea and TG were collected at 30 dpi. Preparation of RNA and cDNA, and real-time quantitative RT-PCR Total RNA was purified from tissues using RNA STAT-60 (Tel-Test, Friendswood, TX), followed by secondary purification and DNAse I treatment using RNeasy columns (Qiagen). cDNA was synthesized using the Omniscript Reverse Transcriptase Kit (Qiagen) for Exp. #1 or TaqMan ® Reverse Transcription Reagents (Perkin Elmer) for Exps. #2, #3, and #4 following the manufacturers' suggested protocols. Design of the PCR primers and TaqMan ® probe s for mouse chemokine and chemokine receptors was done using Primer Express (Applied Biosystems) software. Primer and probe sequences are listed in Table 2 . Primers and the VIC-labeled TaqMan ® probe s for the housekeeping control genes rodent GAPDH and 18S rRNA were purchased from Applied Biosystems. Real-time quantitative RT-PCR assays were performed with reagents recommended by the manufacturer (Applied Biosystems) using an ABI PRISM 7700 Sequence Detection System instrument. Briefly, 0.5 μL (approximately 300 pg) of cDNA was added to 25μL reactions containing 12.5 μL of PCR Universal Mix (Applied Biosystems), 600 nM F primer, 600 nM R primer, 200 nM FAM-labeled TaqMan probe, 200 nM rodent GAPDH F primer, 200 nM rodent GAPDH R primer, and 100 nM rodent GAPDH TaqMan ® probe. The number of PCR cycles needed for FAM or VIC fluorescence to cross a threshold where a statistically significant increase in change in fluorescence (CT=threshold cycle) was measured using Applied Biosystems software. Relative RNA expression was determined using the formula Rel Exp= 2 -(ΔΔCt) × 1000 where ΔΔ CT= (CT gene of interest-CT rodent GAPDH in experimental sample)-(CT gene of interest-CT rodent GAPDH in a no-template control sample) (the ΔΔ CT method, Taqman ® Bulletin #2: Relative Quantitation of Gene Expression, Applied Biosystems, updated 2001, ). To assure that GAPDH RNA levels were not affected by HSV infection and thus a good control, we repeated most analyses using 18S rRNA as an internal control. In all cases tested, induction measurements (HSV+/mock) were indistinguishable whether 18S or GAPDH were used (not shown). Control reactions lacking RT were used to test for the presence of contaminating HSV or mouse DNA, and in all cases either no or low (relative to when RT was present) levels of amplification were measured (not shown). Purified HSV-1 genomic DNA was kindly provided by Jean Pesola. Competing interests The author(s) declare that they have no competing interests. Authors' Contributions W. Cook, R. Walker and T. Burwell performed the RT-PCR analyses of chemokine transcripts. M. Kramer and H. Holman performed the animal infections and provided tissues for transcript analysis. D. Coen and D. Knipe participated in the design of experiments, oversight of the conduct of the experiments, and in the interpretation of the results.	/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC524517.xml
554972	The architecture of mammalian ribosomal protein promoters	Background Mammalian ribosomes contain 79 different proteins encoded by widely scattered single copy genes. Coordinate expression of these genes at transcriptional and post-transcriptional levels is required to ensure a roughly equimolar accumulation of ribosomal proteins. To date, detailed studies of only a very few ribosomal protein (rp) promoters have been made. To elucidate the general features of rp promoter architecture, I made a detailed sequence comparison of the promoter regions of the entire set of orthologous human and mouse rp genes. Results A striking evolutionarily conserved feature of most rp genes is the separation by an intron of the sequences involved in transcriptional and translational regulation from the sequences with protein encoding function. Another conserved feature is the polypyrimidine initiator, which conforms to the consensus (Y) 2 C +1 TY(T) 2 (Y) 3 . At least 60 % of the rp promoters contain a largely conserved TATA box or A/T-rich motif, which should theoretically have TBP-binding capability. A remarkably high proportion of the promoters contain conserved binding sites for transcription factors that were previously implicated in rp gene expression, namely upstream GABP and Sp1 sites and downstream YY1 sites. Over 80 % of human and mouse rp genes contain a transposable element residue within 900 bp of 5' flanking sequence; very little sequence identity between human and mouse orthologues was evident more than 200 bp upstream of the transcriptional start point. Conclusions This analysis has provided some valuable insights into the general architecture of mammalian rp promoters and has identified parameters that might coordinately regulate the transcriptional activity of certain subsets of rp genes.	Background Ribosomes are vital organelles, which catalyze protein synthesis in all living organisms. Eukaryotic ribosomes consist of four RNA molecules (rRNAs) and 79 different proteins. The mammalian genes encoding the rRNAs are multicopy and clustered at a few loci, whereas those encoding the ribosomal proteins (rp genes) are single copy and scattered throughout the genome [ 1 ]. In addition to the functional rp genes, all of which contain introns, mammalian genomes contain many nonfunctional intronless rp pseudogenes [ 2 ]. The earliest determinations of mouse rp gene sequences and of transcriptional start points (tsp's) revealed a salient feature of rp genes, namely that transcription is initiated at a C residue within a polypyrimidine tract [ 3 - 5 ]. A recent study by Kenmochi and coworkers [ 6 ] has demonstrated that this is a general property of virtually all human rp genes. Because of this novel feature, the rp mRNAs contain a 5' terminal oligopyrimidine sequence (TOP), which is essential for their translational control [ 7 ]. Coordinated expression of the rp genes at transcriptional and post-transcriptional levels is required to ensure a roughly equimolar accumulation of ribosomal proteins. Transcriptional run-on measurements with nuclei of rapidly proliferating cells indicated equivalent loading of RNA polymerases on three unlinked mouse rp genes [ 8 ], consistent with the equal abundance of the corresponding rp mRNAs [ 9 ]. Moreover, the promoters of these genes were of comparable strength in driving the expression of a common reporter gene [ 8 ]. These results suggested that similar promoter strength and mRNA processing efficiency might provide a basis for the coordinated expression of rp genes. However, whether this concept applies to all rp genes or to distinctive subsets of genes is presently unclear. Despite the obvious importance of rp gene expression for cell viability, there have been very few experimental studies of rp promoter architecture and transcriptional regulation in higher eukaryotes. To date, efforts to identify functionally relevant cis-acting regulatory elements and transcription factor binding sites have been made for only 9 mammalian [ 8 , 10 - 23 ] and 2 amphibian [ 24 , 25 ] rp genes, less than 15 % of the total rp gene complement. These studies identified regulatory elements in the promoter-proximal regions, both upstream and downstream of the tsp. Some of these elements contained binding sites for known transcription factors, notably GABP, Sp1 and YY1. When the binding of any individual factor was eliminated by a site-specific mutation, transcriptional activity was reduced, but not abolished, indicating that the overall transcription efficiency is determined by a combinatorial effect of multiple factors. No regulatory element common to all rp genes was found, although certain elements were present in several of the genes that were studied. None of these rp genes contained a canonical TATA box in the -25 to -30 region, although some had a "TATA-like" A/T-rich sequence, which might bind the general transcription factor TBP under certain circumstances. Because the previous experimental studies were limited to a small subset of rp genes, it is important to know whether the results of those studies can be used to predict special features of promoter architecture that characterize the entire set of rp genes or that could be used to sort the rp genes into classes with similar features. This applies not only to transcription factor-binding sites, but also to the polypyrimidine tract that spans the tsp. As an initiator element, this tract is atypical because the tsp of initiators is normally an adenine residue, flanked on both sides by pyrimidines [ 26 ]. To address these issues, I have applied the principle of "phylogenetic footprinting", which holds that important regulatory sequences will have a tendency to be evolutionarily conserved and thus revealed by a sequence comparison of corresponding regions of orthologous genes [ 27 ]. Some regulatory promoter elements may escape detection by this approach, but a substantial majority will probably be recognized with confidence [ 28 ]. Although previously done on a small scale for a few selected rp genes [ 19 , 21 , 29 - 32 ], such an analysis has not heretofore been made for the entire rp gene population. I have therefore compared the promoter regions of all 79 orthologous human and mouse rp genes and have extended the comparisons to include chicken, amphibian and fish counterparts when these sequences were available. I have compared the sequence organization of the rp promoters with that of the promoters of non-rp TOP genes and other genes in the housekeeping category. This analysis has provided some useful insights into the general architecture of mammalian rp promoters and has identified parameters that might coordinately regulate the transcriptional activity of certain subsets of rp genes. Results Compilation of the promoter sequences of orthologous rp genes To make a study of the evolutionarily conserved features of ribosomal protein gene promoters, I first had to assemble a set of orthologous human and mouse rp gene sequences, in which the transcription start sites are reasonably well defined. For all of the human rp genes, annotated sequences are now available owing to the studies of Kato, Kenmochi and coworkers [ 6 , 33 ], who systematically determined the 5' termini of human rp mRNAs by the oligo-capping method. While these studies indicated some small variations in the exact 5' ends of individual rp mRNAs, the tsp's could usually be specified to within a few nucleotides. Independent determinations of the human tsp's by the oligo-capping procedure are also available on the database of transcription start sites (DBTSS) compiled by Sugano and colleagues [ 34 ]. The agreement between the two sets of oligo-capping data and the results of primer extension/nuclease protection (PE/S1) experiments is generally very good [see Additional file 1 ]. In all cases, the human tsp's selected in the present study correspond either to the most abundant oligo-capped cDNA or to an observed variant. In contrast to the situation with the human rp genes, accurate determinations of the tsp's of mouse rp genes have been made in only a few cases, and therefore the annotations of rp genes in the mouse genome databases are generally incomplete. To overcome this problem, I used the mVISTA alignment program [ 35 ] to compare mouse sequences extending several kb upstream of the coding regions with the sequences of the corresponding human orthologues. Fortunately, the strong conservation of sequences in the proximal promoter regions and in the first exons enabled me to identify the probable tsp's of most mouse rp genes with confidence. For the most part, the tsp's that were predicted by this strategy corresponded exactly to those previously determined by primer extension and/or nuclease protection assays of murine rp genes [see Additional file 1 ]. Some information on putative tsp's of mouse rp genes is available on the DBTSS database. However, the most prominent of these tsp's did not agree with the PE/S1 data. Moreover, they were not coincident with the aligned C tsp's of the human orthologues, but were frequently at adjacent or nearby T residues. Conceivably, these discrepancies are caused by a technical bias, as is sometimes encountered in the generation of genome-wide full-length cDNA libraries [ 36 ]. Therefore, relying on maximum sequence conservation and the best agreement with previous experimental data, I compiled aligned sequences of all 79 pairs of orthologous human and mouse rp genes, and used these sequences for further analysis of the promoter regions. In a preliminary analysis, I used the mVISTA sequence alignment program to examine six to ten kb of 5' flanking sequence of three pairs of orthologus rp genes ( rpS16 , rpL30 and rpL32 ). These genes were selected because the mouse orthologues have been previously studied extensively in my laboratory. Scans with a 50 bp window indicated that the conserved (> 75 % identity) flanking sequences are largely confined to regions within a few hundred base pairs of the tsp. There was very little alignment of the upstream region sequences and many gaps caused by the presence of numerous insertion elements. With this analysis, I detected only one short block of conserved sequence at -2 kb in rpL30 (77 % identity over 84 bp), one short block at -1.5 kb in rpL32 (82 % identity over 60 bp), and none in rpS16 . Neither of the conserved blocks contained any recognizable transcription factor binding sites or other remarkable characteristics. The lack of any credible long-range regulatory elements is consistent with earlier conclusions based on transient transfection experiments with these genes [ 8 , 17 , 37 ]. Given these results, I restricted the analysis of the full set of rp genes to the transcribed portions of the genes and about one kilobase of 5' flanking sequence [see Additional file 2 ]. An initial analysis of the large data set revealed the presence of repetitive sequence insertion elements (sines for human and B1/B2 for mouse) in the 5' flanks of a very high proportion of the rp genes [see Additional file 2 ]. Indeed, as seen in Figure 1 , half of the rp genes contain such an element within 500 bp of the tsp and over 80 % contain elements within 900 bp. The distribution is very similar in the human and mouse rp genes. Both ancestral elements, which are moderately conserved between mouse and human, and lineage-specific elements could be identified. As in the case of the three examples discussed above, very little sequence conservation was evident more than 200 bp upstream of the tsp. Accordingly, for the refined analysis of rp promoter architecture, I analyzed segments extending from 200 bp upstream to 100 bp downstream of the tsp. These segments were scrutinized for the quality of TATA box motifs in the -25 region, for conserved sequences in the initiator region, for transcription factor binding sites, and for the location of the AUG translation start codons. Figure 1 The location of insertion elements in the 5'-flanking regions of rp genes . Insertion (ins) elements were identified by the RepeatMasker program in scans of up to one kilobase of sequence 5' of the tsp, and the location of the element nearest to the tsp recorded for 79 human and mouse rp genes [see Additional file 2 ]. The distance from the tsp was divided into 200 bp intervals, the percentage of rp genes with an element within each interval determined, and the values plotted cumulatively against the distance from the tsp. Criteria for the annotation of rp promoter sequences To evaluate the quality of TATA box motifs, I established criteria based on rules derived from a crystallographic analysis of TBP-DNA complexes [ 38 ]. With these rules one can classify each nucleotide in the motif as being "preferred" or "acceptable" or "incompatible" with TBP binding (Figure 2a ). I considered motifs with a string of 6 or more compatible nucleotides, of which at least 5 are "preferred", as being capable of binding TBP with high affinity (+ motifs). Those with a string of 6 compatible nucleotides, 3 or 4 of which are "preferred", were considered as possible low affinity sites (+/- motifs). Sequences that do not satisfy these criteria were judged to be incapable of unaided binding to DNA (- motifs). Figure 2 Criteria for rp promoter annotation . (a). Quality of the TATA box for TBP binding based on a structural analysis of TBP-DNA complexes. Rules adopted for + and +/- ranking and some examples in rp promoters are shown at the right. (b). Criteria used for identification of potential transcription factor-binding sites by the rVISTA program for motifs aligned in human-mouse promoter sequence comparisons and with the FindPatterns program for unaligned motifs. Employing stringent criteria (Figure 2b ), I used the rVISTA program [ 39 ] to search systematically for conserved (aligned) sites that would be predicted to bind the three transcription factors (GABP, Sp1, and YY1) that were identified in earlier experimental studies of rp promoters and also for highly conserved sites that might bind other ubiquitously expressed factors. In addition, I scanned both the human and mouse rp promoter sequences for unaligned optimal sites for the above-mentioned factors and for a motif termed Box A, which was previously implicated in rp gene expression [ 19 ]. I used the results of these analyses to complete the annotation of all 79 pairs of aligned human and mouse rp promoter sequences, which can be viewed individually on pages 1 through 80 of a supplementary file [see Additional file 3 ]. Annotated comparisons of human/mouse promoter sequences Illustrative examples of four annotated promoter sequence comparisons are shown in Figure 3 . All four of these examples illustrate the strong sequence conservation of the non coding portions of exon 1, which is a general feature of the entire rp gene family [see Additional file 3 ]. The AUG translation initiation codon is located within exon 1 of rpL13a (panel d) and within exon 2 of rpL30 (panel b), whereas in rpS18 (panel a) and rpS4 (panel c), it is at the extreme 3' end of exon 1. Many rp genes have this latter feature, which may be relevant to the evolution of vertebrate rp promoters (see below). Figure 3 Examples of comparisons of human and mouse rp promoter sequences from -200 to +100 . Exon I is in uppercase letters, highlighted in yellow with the ATG translation initiation codon, if present, highlighted in gray. The sequences evaluated for TATA box quality are highlighted in green. Putative transcription factor-binding sites in aligned sequences are shown above the sequences, highlighted in fuchsia; sites in unaligned sequences are shown above the human or below the mouse sequences and enclosed in carets. An upstream sequence identified as an insertion element is in red letters. Gaps inserted by the alignment program to maximize overall sequence identity are indicated by dashes. (a) rpS18 , (b) rpL30 , (c) rpS4 , (d) rpL13a . In rpL30 the elements with known functional relevance are underlined with asterisks. The rpS18 promoter contains a conserved TATA (+) motif, while the TATA motifs of rpS4 and rpL13a were scored as (+/-) and that of rpL30 as (-). Well-conserved (aligned) consensus binding sites for GABP and/or Sp1 are located upstream of the tsp in all four examples; downstream YY1 sites are evident in rpL30 and rpL13a . The rpS18 and rpL13a promoters also contain conserved consensus binding sites for the ubiquitous AP1/ATF factors. The promoter elements that were found to have functional significance in experimental studies of the mouse rpL30 gene [ 11 ] are all conserved in the human orthologue (panel b). This includes sites for RFX1 and the Gamma Factor as well as the GABP and YY1 sites. In addition to the conserved sites, a few non conserved (unaligned) GABP and Sp1 sites of uncertain relevance are evident in rpL30 and rpL13a . The results of this type of analysis for the full set of rp promoters are tabulated in Table 1 . The identity between human and mouse sequences in the 300 bp segment (-200 to +100) that was analyzed ranged from 50 to 75 % with an average of 61 %. The location of the translation initiation codon was conserved in all except one rp gene ( rpL29 ). In 29 % of the rp genes the AUG codon is at the extreme 3' end of exon 1 and in 47 % it is in exon 2. Thus, in 3/4 of the rp genes, the genetic elements involved in transcriptional/translational regulation are spatially separated from those with protein encoding function. Table 1 Ribosomal protein promoters Quality of the TATA Box Motifs, the Number of Transcription Factor-Binding Sites and the Location of AUG Codons. RP GENE TATA Quality % H/M identity in 300 bp ‡ GABP (5' of tsp) Sp1 (5' of tsp) YY1 (3' of tsp) OTHER (5' of tsp) AUG † Hs Mm H-M H H-M H H-M H SA + + 58 1 2 1 ATF, CREB E2 S2 + + 54 1 1 E2 S3 + + 60 1 1 1 E1 S3a – – 75 2 1 AP1 E1 S4 +/– +/– 70 1 1 1 E1e S5 +/– + 62 1 E2 S6 – – 56 3 * 1 1 Box A * E1 S7 +/– + 56 4 2 Box A E2 S8 +/– +/– 60 1 E1e S9 + + 50 1 1 ATF, CREB, AP1 E2 S10 – – 64 1 1 1 1 Nrf1 E2 S11 +/– +/– 53 1 1 1 1 Box A, ATF, CREB, AP1 E1 S12 + + 62 1 Nrf1 E2 S13 – +/– 61 1 1 AP1 E1 S14 – – 62 3 * 1 AP1 E2 S15 +/– +/– 52 1 1 1 Nrf1 E1e S15a +/– +/– 59 1 2 1 Box A E2 S16 – +/– 74 2 * 2 Nrf1 E1 S17 +/– + 60 1 1 2 1 E1e S18 + + 64 3 1 ATF, CREB E1e S19 +/– +/– 72 1 1 Box A (2) E2 S20 + + 63 1 1 E1e S21 + +/– 52 3 1 E2 S23 – – 65 1 1 1 E1e S24 + + 59 1 1 2 E1e S25 – – 56 1 1 E1e S26 + + 53 1 E1e S27 + + 71 2 1 1 E1 S27a – – 57 1 1 1 E2 S28 + + 67 1 1 1 1 AP1 E1 S29 – – inc E1 S30 – – 68 4 1 ATF, CREB E2 L3 + + 56 2 1 1 1 E1e L4 + + 68 2 E1e L5 + + 62 1 1 E1e L6 +/– +/– 62 1 1 1 E2 L7 +/– +/– 69 1 1 Nrf1 E1 L7a + + 51 2 1 1 1 Box A, B E1e L8 +/– +/– 50 1 1 1 1 1 AP1 E2 L9 – – 57 1 ATF, CREB (2) E2 L10 + + 60 1 E2 L10a + + 55 Box A, Nrf1 E1 L11 +/– +/– 64 1 1 1 E2 L12 + + 73 1 Box A E1 L13 – +/– 55 1 1 E2 L13a +/– +/– 67 2 2 1 ATF, CREB, AP1 E1 L14 +/– +/– 58 2 1 Nrf1 E1e L15 – – 66 1 1 1 2 E2 L17 – – 69 1 1 1 Box A E2 L18 – – 51 1 2 ATF, CREB, AP1 E1e L18a – – 67 1 1 E1 L19 +/– +/– 65 1 1 E1 L21 – – 62 1 1 E2 L22 – – 61 E1 L23 + + 64 1 1 1 E1 L23a +/– +/– 63 1 Box A E1 L24 +/– +/– 56 1 1 E1 L26 +/– +/– 64 1 1 2 Box A E2 L27 – – 66 3 1 1 E2 L27a – – 56 1 1 1 Box A E1e L28 + +/– 56 2 1 1 1 E2 L29 +/– +/– 68 2 1 1 E1e//2 L30 – – 60 2 1 1 * RFX1* Gamma* E2 L31 – – 62 1 1 E2 L32 – – 68 1 * 1 * 1 * Gamma * E2 L34 – – 55 1 1 2 1 E2 L35 – – 59 2 4 1 E1e L35a – – 64 1 1 1 E2 L36 +/– + 51 1 1 E2 L36a + + 62 1 2 2 1 E1e L37 + – 60 2 2 1 1 2 E1e L37a +/– +/– 52 1 Nrf1 E1e L38 +/– – 62 2 E2 L39 – +/– 61 1 1 1 1 E1e L40 – – 62 2 3 ATF, CREB E2 L41 + + 54 1 1 1 ATF, CREB, AP1 E2 LP0 + + 59 1 E2 LP1 +/– +/– 62 1 E1 LP2 +/– – 55 1 E2 Bolded numbers indicate that the sites are aligned in orthologous human and mouse genes. An asterisk indicates that there is experimental evidence for the functionality of one or more of the sites. ‡ From –200 to +100 relative to the tsp. † E1, within Exon 1; E1e, at the extreme 3' end of Exon 1; E2, within Exon 2. The locations are conserved in all rp genes except rpL29 . Contrary to previous impressions based on an incomplete set of rp genes, the rp promoters cannot generally be classified as "TATA-less". Thirty-five percent of the promoters contain, in the -25 region, a TATA box that should theoretically bind TBP with high affinity. An additional 25 % have an A/T-rich tract in this region, which might bind TBP with lower affinity, and the remainder would not be predicted to bind TBP without help from other proteins. For the most part, the TATA box quality is well conserved between the two mammalian species: the assessed quality of human and mouse TATA motifs was the same in 82 % of the rp promoters. The prevalence in rp promoters of evolutionarily conserved GABP, Sp1 and YY1 binding sites is readily apparent from these results. Conserved upstream GABP and Sp1 sites are present in 54 % and 48 % of the rp promoters, respectively. Conserved downstream YY1 sites are present in 52 % of the rp promoters. The ratio of aligned to unaligned sites in the human rp promoters is approximately 2:1 for GABP, 1:1 for Sp1 and 4:1 for YY1. The occurrence of unaligned sites in the mouse rp promoters is similar to that in the human rp promoters [see Additional file 3 ]. It is noteworthy that 76 % of the rp promoters contain at least one conserved upstream GABP and/or Sp1 site within 200 bp of the tsp. If unaligned sites are included, the proportion of human rp genes with an upstream GABP and/or Sp1 site is 92 %. Conserved consensus motifs for other ubiquitous transcription factors are present in a much lower proportion of the rp promoters, namely 12 % for the Box A-binding factor, 11 % for AP1, 10 % for ATF/CREB and 8 % for Nrf1. The rp gene initiator The existence of a novel polypyrimidine initiator sequence in which the conventional A residue at the tsp is replaced by a C residue is well known. Moreover, the roles of this sequence in rp gene transcription and rp mRNA translation have been previously demonstrated experimentally for a few rp genes [ 7 , 40 - 42 ]. What is not known is whether there is a consensus initiator sequence that characterizes the entire rp gene set. To address this issue, I used the pairwise alignments of human and mouse orthologues to produce an occupancy matrix for positions -8 to +10 of the initiator [see Additional file 4 ]. With this matrix and the standard consensus rules, I determined that the rp consensus initiator sequence is (Y) 2 C +1 TY(T) 2 (Y) 3 . A striking result is that T's are strongly preferred over C's at positions +2, +4 and +5. This preference, which is graphically illustrated in Figure 4 , might be related to a transcriptional function of the initiator or a translational function of the TOP or to both functions. Figure 4 The consensus initiator sequence of mammalian rp genes . Seventy-nine pairs of orthologous human and mouse rp gene sequences were compared at positions -8 to +10 and the occurrence of each nucleotide or pair of nucleotides depicted by the height of the letters: A, G, C. T, Y = C/T, R = A/G, W = A/T, K = G/T, S = C/G, M = C/A. The tsp is the C at position +1. The extent of conservation of rp promoter features in non-mammalian vertebrates To determine the features of rp promoter architecture that are conserved over large evolutionary periods, I compared the promoter sequences of five chicken, six amphibian and five fish rp genes to their mammalian orthologues using the Clustal W multiple sequence alignment program. An example of such an alignment for rpS3 (Figure 5 ) shows strong conservation of both the coding and non-coding portions of the first exon, the TATA box motif, the initiator sequence, and a downstream YY1 site that spans the translation initiation codon. In contrast, the overlapping GABP and Sp1 sites, which are aligned in the human and mouse promoters, are not conserved in the amphibian and fish promoters. However, each of these promoters contains an unaligned consensus site for GABP or Sp1. A summary of the multi-sequence analysis for 11 rp genes (Table 2 ) indicates that the extent of conservation observed for rpS3 is fairly typical, although some small variations are evident. The conformance to the consensus initiator sequence is the same (90 %) for mammals and lower vertebrates. Yet, the particular C residue within the initiator that is used as the major tsp may differ slightly. Also the adjudged quality of the TATA box for TBP binding is not always the same. Nevertheless, except for the location of upstream transcription factor-binding sites, the general features of rp promoter architecture are usually well conserved over large evolutionary distances. Figure 5 Comparison of rpS3 promoter sequences in human (Hs), mouse (Mm), Xenopus laevis (Xl) and Fugu rubripes (Fr) . Alignment by the ClustalW program with 100% and 75% identities enclosed in boxes. The first exon is highlighted in yellow with the ATG translation initiation codon in gray. Putative binding sites for TBP, GABP, Sp1 and YY1 are highlighted in green, red, blue and pink, respectively. Table 2 Comparison of promoter features in mammals and lower vertebrates Gene Species Initiator TATA GABP Sp1 YY1 Box A AUG Acc. No. rpS3 Hs/Mm TT C CTTTCCT + 1 1 1 E1 AB061838, NM_012052 Xl GC- C ------ + † 1 E1 Z34530 Fr C--------- + † 1 E1 X97794 rpS6 Hs/Mm SC C TCTTTTY – 3 1 E1 X67309, D28348, Z54209 Xl C --- C ---C- – E1 AF020551 rpS7 Hs/Mm SY C TCTTYCT +/– + 1 E2 Z25749, AB055774, NM_011300 Xl - CC ------- – 1 E2 X71081 Fr ---------- + 1 E2 X94942 rpS15 Hs/Mm WY C TYTTCYG +/– 1 1 1 E1e M32405, AB055776, NM_001018 Gg C --C------ +/– 1 † 1 E1e D10167 rpS24 Hs/Mm TY C TCTTTTC + 1 E1e AB062069, U12202, NM_011297 Xl ----T C C--T + 1 E1e M33517 Fr -----C---T +/– 1 E1e AJ001398 rpL5 Hs/Mm GY C YTTTTCC + 1 E1e AB055762, NM_00969, mCG13589 Gg -G- C ------ – 1 E1e D10737 rpL7a Hs/Mm TT C TYTCTCC + 2 1 1 E1e X61923, X54067 Gg G CCC --T-TA +/– 1 1 E1e X62641 Fr -C------GC +/– E1e Y15171 rpL14 Hs/Mm YY C TTCTCGC +/– 2 E1e AB061822, mCG22708 Xl -- C --T--T- – E1e X06552 rpL18 Hs/Mm ST C TTTCCGG – 1 2 E1e AB061825, mCG132477 Xl TC C -----TC +/– † 1 E1e X05025 Om TC C ---T-CC – † E1e AF240376 rpL30 Hs/Mm TT C CTTTCTC – 2 1 E2 AB070559, K02928 Gg C-- C -----G – 1 E2 D14521 rpL37a Hs/Mm TC C TYTYYGG +/– 1 E1e NM_000998, NM_009084 Gg -- C -----CT +/– 1 E1e D14167 Hs/Mm: Homo sapiens/Mus musculus; Gg: Gallus gallus; Xl: Xenopus laevis; Fr: Fugu rubripes; Om: Oreochromis mossambicus (tilapia). For the initiator sequences, C represents an alleged tsp based on various types of experimental evidence. For the transcription factor binding sites, only those sites that are conserved (aligned) between human and mouse are listed. A † symbol indicates that the promoter contains an unaligned site for the factor. Designation of AUG locations as in Table 1. Comparison of rp promoter features with those of non-rp TOP genes and other housekeeping genes It is of interest to know which features of rp promoter architecture are specific and which might be common to other ubiquitously expressed genes. To this end, I examined the promoters of two additional sets of genes. One set consisted of non-rp genes that also produce translationally controlled mRNAs with 5'-terminal oligopyrimidine tracts (non-rp TOP genes). At the present, there are nine known genes in this category for which the tsp of at least one orthologue has been experimentally determined [see Additional file 5 ]. The non-rp TOP set contains genes encoding translation elongation factors (eEF1A1, eEF1B and eEF2), RNA-binding proteins (PABPc1 and hnRNPA1), a major nucleolar protein (nucleoplasmin/B25), a protein with tubulin-binding properties (TCTP/p23), and two genes that do not encode proteins, but rather have small nucleolar RNA-encoding sequences embedded within their introns ( gas5 and U17HG ). Annotated aligned sequences of the nine human and mouse non-rp TOP genes are presented on pages 1 through 9 of a supplementary file [see Additional file 6 ] and the results are summarized in Table 3 . Table 3 Non-rp top genes GENE TATA Quality % H/M identity in 400 bp * GABP SP1 YY1 Other Initiator -2 to +7 † AUG Hs Mm eEF1A1 + + 44 1, 2 TT C TTTTTC E2 eEF1B + + 68 2 1 TC C TTTTTY E1 eEF2 + + 55 4 2 KT C TCYYCC E1e PABP cl + + 66 2, 1 TT C CCCTTC E1 hnRNPA1 – – 71 2 YT C CTTTCT E1 B23 + + 73 1, 1 BoxA TT C CYTGGC E1 Tpt1 + + 67 1 1 1 GC C TTTTCC E1 gas5 + + 55 1, 2 RK C YTTTCG -- U17HG + + 43 1 ST C YYTYTW -- * From -300 to +100 relative to the tsp. † The bolded C is the tsp. Other symbols are as in Table 1. There are some notable differences in the promoter architecture of rp genes and non-rp TOP genes. First, GABP- and YY1-binding motifs, which are prevalent in the rp promoters, are rarely found in the non-rp TOP promoters (Table 4a ). Second, eight of the nine non-rp TOP promoters have conserved TATA boxes that would be expected to bind TBP with high affinity, whereas only a third of the rp promoters have such TATA boxes (Table 4b ). These differences were considered to be statistically significant when analyzed by Fisher's exact test (Table 4 ). The initiator sequences of the non-rp TOP genes resemble the rp consensus (89 % identical from -2 to +7) except for the roughly equal occurrence of C's and T's at +2. Table 4 Comparison ofthe promoters ofrp genes, non-rp top genes, and other housekeeping genes (a) Occurrence of transcription factor binding sites: proportion of genes with at least one binding site. FACTOR HUMAN RP GENES HUMAN NON-RP TOP GENES HOUSEKEEPING GENES Aligned Total Aligned Total Total GABP (5' of tsp) 54 % 68 % 22 % * 33 % 35% SP1 (5' of tsp) 48 % 70 % 89 % 89 % 90 % GABP or SP1 76 % 92 % 100 % 100 % 95 % YY1 (3' of tsp) 52 % 59 % 11 % † 11 % † 10 % † (b) Potential for TBP binding to TATA boxes in the -25 region: proportion of genes in each category. TATA BOX QUALITY RP GENES NON-RP TOP GENES HOUSEKEEPING GENES + 35 % 89 % ‡ 30 % +/– 25 % – 20 % – 39 % 11 % ‡ 50 % Based on analyses of 79 pairs of orthologous human and mouse rp genes, 9 pairs of orthologous non-rp TOP genes and 20 randomly selected human or mouse housekeeping genes. The symbols , †, and ‡ indicate the percentages considered to be significantly different from the corresponding values for rp genes according to a Fisher's two sided exact test of the three sets of data. : p = 0.086; †: p = 0.031 and 0.00075; ‡: p = 0.0029. For the rp genes, the TATA box quality was scored as: + when both orthologues are + or when one is + and the other +/-, +/– when both orthologues are +/– or when one is + and the other is –, and – when both orthologues are – or when one is - and the other is +/–. For the non-rp TOP genes, both orthologues were either + or –. The second set of ubiquitously expressed genes consisted of 20 housekeeping genes randomly selected from the eukaryotic promoter database [see Additional file 7 ]. The promoters of these genes also have an under-representation of YY1 sites compared to the rp promoters, and like both the rp and non-rp TOP genes, contain abundant motifs for Sp1 (Table 4a ). The proportion of these genes that have TATA boxes with TBP-binding capability is similar to that observed for the rp genes (Table 4b ). Discussion The foregoing analysis of 79 mammalian ribosomal protein genes has revealed several features of rp promoter architecture, some of which are largely conserved over long periods of vertebrate evolution (about 450 million years) and others that are strongly conserved only in mammals. (about 90 million years). One highly conserved feature, present in over 3/4 of the rp genes, is the separation by an intron of the sequences involved in transcriptional and/or translational regulation from the sequences with protein-encoding function. In 47 % of the rp genes, the AUG translation initiation codon is in exon 2, and in 29 % of the genes, it is at the extreme 3' end of exon 1. It would seem that at an early stage of vertebrate evolution, these regulatory sequences were appended as discrete units to the loci containing the protein-encoding sequences. The polypyrimidine tract that spans the tsp is present in all vertebrate rp genes. This tract can function as a transcriptional initiator [ 40 , 41 ] and also embraces the TOP sequence, which is essential for the translational control of rp mRNAs [ 7 , 42 ]. Based on the assignments of human and mouse tsp's used in the present study, the average lengths of the polypyrimidine tracts and TOP sequences are 12.2 bp and 8.2 bp, respectively. A compilation of conserved sequences in the -8 to +10 regions of orthologous human and mouse rp genes revealed the consensus sequence (Y) 2 C +1 TY(T) 2 (Y) 3 . Thus, in addition to the C at the tsp, there is a clear preference for T over C at positions +2, +4, and +5. This preference may reflect a structural bias for transcription, e.g., ease of strand separation, or for translation, e.g., affinity of rp mRNAs for a putative repressor. The presence of transposable element residues in the 5' flanks of the rp genes is noteworthy. Half of the rp genes contain an element (sines for humans, B1/B2 for mouse) within 500 bp of the tsp, and over 80 % of the genes contain an element within 900 bp. Some elements are moderately conserved between mouse and human, but most appear to be lineage-specific. These elements are unlikely to have any specific role in rp promoter function. They may be passively tolerated because the vast majority of conserved 5' sequence is confined to within 200 bp of the tsp. Within a segment from -200 to +100, the sequence identity between human and mouse rp orthologues ranges from 50 to 75 % with an average of 61 %, whereas the sequence match beyond -200 is of borderline significance. The observation that 35 % of the rp promoters contain a TATA box motif at -25, which would be predicted to bind TBP with high affinity, and that an additional 25 % possess A/T-rich motifs, which might bind TBP with lower affinity, was unexpected. The assessed quality of these motifs for TBP binding, made according to rules established by a detailed structural study of TBP-DNA complexes [ 38 ], was largely conserved between human and mouse rp orthologues. Some of the promoters classified as poor (-) binders, e.g., rpL32 , might bind TBP weakly without help from an additional protein [ 43 , 44 ], so that the true proportion of rp promoters with TBP-binding capability might actually be greater than 60 %. Thus, contrary to earlier views based on an analysis of a small subset of rp genes, many of the rp promoters should not be classified in the "TATA-less" category. When human and mouse rp promoter regions from -200 to +100 were scanned for conserved (aligned) transcription factor-binding sites with the rVISTA program [ 39 ], which is based on consensus sequences and matrix tables in the TRANSFAC database, three ubiquitously expressed factors that had previously been implicated in rp promoter activity predominated. Using high stringency criteria, I detected aligned GABP- and Sp1- binding sites upstream of the tsp and aligned YY1-binding sites downstream of the tsp in approximately half of the rp promoters. The occurrence of aligned motifs for other ubiquitously expressed factors is considerably less, i.e., only about 10 % for any single factor. Whereas Sp1 sites are also commonly found in the promoters of many housekeeping genes, including the non-rp TOP genes, the prevalence of GABP and YY1 sites appears to be a more prominent feature of the rp promoters. GABP is a heteromeric protein consisting of an ETS family member, which has DNA-binding capability, and an ankyrin repeat-containing subunit, which greatly improves the stability of GABP-DNA interactions [ 45 , 46 ]. Two-thirds of the human rp promoters contain one or more potential GABP-binding sites upstream of the tsp, 79 % of which are perfectly aligned in the orthologous mouse rp promoters. Previous experimental studies have implicated GABP as a positive transcriptional regulator of the mouse rpL30 and rpL32 genes [ 10 , 14 ], the human rpS14 and rpS6 genes [ 20 , 22 ], and the Xenopus rpL18 gene [ 24 ]. The GABP sites that positively contribute to transcriptional activity are generally located upstream of the tsp, although in some cases, sites overlapping the initiator or downstream of the tsp may also be relevant. In the mouse rpS16 gene, GABP binding at the initiator decreases transcription activity both in vivo and in vitro [ 12 ]. For simplicity, only upstream, presumably positively acting, GABP sites were included in the tabulation. YY1 is a zinc finger-containing protein with a variety of gene-specific functions, including transcriptional activation and repression, positioning of RNA polymerase II, and chromatin modification [ 47 , 48 ]. Fifty-nine percent of the human rp genes contain at least one YY1 site downstream of the tsp, 88 % of which are conserved in the mouse orthologues. Two functionally relevant binding sites for YY1 (originally termed "delta") were detected downstream of the tsp in mouse rpL32 [ 15 ]. In mouse rpL30 , a downstream YY1 interaction also contributed positively to transcriptional activity, but only when a strong upstream interaction with GABP was eliminated [ 11 ]. Interestingly, the vast majority of YY1 sites are located downstream of the tsp, and therefore, for simplicity, only downstream YY1 sites were included in the tabulation. The alignment of mammalian YY1 sites is frequently preserved in chickens, frogs and fish, in contrast to the GABP and Sp1 sites, which rarely have aligned counterparts in those species (Table 2 ). The aligned downstream site in Xenopus rpL18 was shown to interact with the frog homologue of YY1, but the functional significance of this interaction was not demonstrable with the reporter constructs that were used in these experiments [ 25 ]. The diverse activities of YY1 in different promoter contexts and its propensity for interactions with a wide variety of proteins [ 49 ], have led to the idea that it may have multiple mechanistic roles in transcriptional regulation. Which of these roles apply to the rp promoters remains to be established. Based on the results of the analysis of rp promoter sequences, I have sorted the different rp promoters into eight classes according to whether they possess conserved binding sites for the three prevalent transcription factors, GABP, Sp1 and YY1 (Figure 6 ). Only 10 of the rp promoters do not appear to contain conserved binding sites for any of the three factors, and considering the high stringency used for the analysis, the true number may even be lower. Moreover, 8 of these 10 promoters contain non-conserved GABP, Sp1 or YY1 sites and/or conserved sites for other ubiquitously expressed factors. TATA boxes with predicted affinity for TBP are distributed among the promoters of all eight classes. This classification might prove useful for interpreting the results of experiments in which cell-specific or physiologically induced variations in the expression of different subsets of rp genes are observed. In addition, when more extensive data on the relative transcription rates of rp genes or the relative abundance of rp mRNAs become available, this classification might help account for any as yet undetected variability. Figure 6 Classification of rp gene promoters . The data in Table 1 were used to classify the rp promoters into eight groups according to whether they contain aligned sites for GABP, Sp1, YY1, or various combinations of these factors. The promoters highlighted in green are those in which the TATA box quality of both orthologues was ranked as + or in which one ranked as + and the other as +/-. The promoters highlighted in yellow are those in which both orthologues ranked as +/- or in which one ranked as + and the other as -. Non highlighted promoters are those in which both orthologues ranked as – or in which one ranked as +/- and the other as -. It is worth noting that some ribosomal proteins can also have extraribosomal functions [ 50 , 51 ]. Among the mammalian proteins that have been demonstrated or presumed to have additional functions are rpS3, rpS4, rpL5, rpL7, rpL10, rpL13a, rpLP0 and rpLP2. While, collectively, the promoters of the genes encoding these proteins do not fall into any particular class, most contain conserved binding sites for one or more of the three prevalent transcription factors and have TATA boxes with TBP-binding capability. Recently, the protein RACK1/Asc1p, which had previously been implicated in various signal transduction processes, was shown to have the properties of an authentic 40S ribosomal protein [ 52 ]. When I analyzed the promoter structures of the orthologous human (NM_006098) and mouse (NM_088143) genes that encode this protein, I observed several features in common with the rp promoters. The tsp of the mammalian RACK1 gene is embedded in a polypyrimidine tract that conforms perfectly to the rp initiator consensus sequence. Moreover, the promoter contains a TATA box of +/- quality, an aligned upstream binding motif for Sp1 and an aligned downstream motif for YY1, but no readily detectable motifs for other ubiquitous transcription factors. Interestingly, the features of this promoter resemble those of the rpL13a gene, which also encodes a protein with apparent pleiotropic function. This analysis has highlighted features of rp promoter architecture that are shared by a high proportion of the rp genes. The evolutionary conservation of these features lends strong support to their functional relevance. Yet, superimposed on this general design are variations that confer certain idiosyncratic characteristics on each promoter. There does not seem to be a single master switch that co-regulates all rp genes at the transcriptional level. Rather, the rp promoters are tuned to respond to a combination of factors, including components of the general transcription machinery, a relatively small group of sequence-specific transcription factors, and modifiers of chromatin structure. The inherent functional redundancy and lack of dependence on any single factor are useful design features for genes that must be expressed in a broad spectrum of cell types and environmental situations. Conclusions A sequence comparison of the promoter regions of all 79 orthologous human and mouse ribosomal protein genes has revealed several evolutionarily conserved features that are characteristic of a high proportion of the rp gene set. One such feature, which is also evident in the rp genes of lower vertebrates, is the separation by an intron of the sequences involved in transcriptional and translational regulation from the sequences with protein encoding function. Another conserved feature is the polypyrimidine initiator, which in mammals conforms to the consensus (Y) 2 C +1 TY(T) 2 (Y) 3 . Contrary to previous impressions based on studies of a small subset of rp genes, the majority of rp promoters contain a TATA box or an A/T-rich motif at -25 that should theoretically have TBP-binding capability. Similarly, approximately half of the rp promoters contain conserved binding motifs for transcription factors previously implicated in rp gene expression, namely upstream GABP and Sp1 sites and downstream YY1 sites. Conserved motifs for other ubiquitous factors occurred much less frequently. Transposable element residues within 900 bp of 5'-flanking sequence were present in over 80 % of the rp genes; very little sequence conservation was evident more than 200 bp upstream of the tsp. Some of these architectural features were seen to be specific for rp promoters. From the results of this analysis, it was possible to sort the rp promoters into eight classes according to their possession of putative binding sites for GABP, Sp1 and YY1, and also to specify which promoters should have intrinsic affinity for TBP. This classification might prove useful for interpreting the results of experiments in which cell-specific or physiologically induced variations in the expression of different subsets of rp genes are observed. Methods The rp gene sequences were extracted from three database sources [see Additional file 2 ]. The vast majority of sequences were obtained from the UCSC database [ 53 ], which conveniently uses uppercase and lowercase letters to distinguish exon and flanking/intron sequences, respectively. The remaining sequences were obtained from the ncbi [ 54 ], Celera [ 55 ] and the recently available RPG [ 56 ] databases. For alignment of human and mouse rp promoter sequences I used mVISTA [ 57 ], which is based on the AVID global alignment program [ 58 ]. The locations of repetitive insertion sequence elements were determined by the RepeatMasker program supplied with mVISTA and, in many cases, corroborated by ncbi annotations. Transcription factor-binding sites were detected with the rVISTA program and the FindPatterns tool of the GCG program. The analysis of non-rp TOP genes was made similarly to that of the rp genes. A graphic representation of the rp initiator consensus sequence was obtained with the Weblogo program [ 59 ]. For alignment and viewing of three or more orthologus rp genes, I used the ClustalW program [ 60 ] and SeqVu shareware l.0.1 (Garvan Institute, Sydney Australia). List of abbreviations rRNA, ribosomal RNA; rp, ribosomal protein; tsp, transcriptional start point; TOP, terminal oligopyrimidine sequence; TBP, TATA-binding protein. Supplementary Material Additional File 1 ST1: Identification of the transcriptional start points (tsp's). Comparison of experimental determinations of the tsp's of human and mouse rp genes. Click here for file Additional File 2 ST2: Promoter region sequences of rp genes. Sources of rp sequences, amount of analyzed sequence 5' of tsp, and locations of insertion elements nearest the tsp. Click here for file Additional File 3 Aligned promoter sequences of rp genes. Annotated sequence alignments of 79 orthologous human and mouse rp promoters and a key to the annotation. Click here for file Additional File 4 ST3: Characterization of the rp initiator. Occupancy matrix for determination of the consensus sequence of the mammalian rp initiator. Click here for file Additional File 5 ST4: Promoter region sequences of non-rp TOP genes. Sources of non-rp TOP gene sequences, the amount of analyzed sequence 5' of the tsp, and location of insertion elements nearest the tsp. Click here for file Additional File 6 Aligned promoter sequences of non-rp TOP genes . Annotated sequence alignments of 9 orthologous human and mouse non-rp TOP promoters. Click here for file Additional File 7 ST5: Housekeeping gene promoters. List of 20 housekeeping gene sequences extracted from the Eukaryotic Promoter Database, which were analyzed for GABP, Sp1 and YY1 binding sites and for TATA box quality. Click here for file	/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC554972.xml
521686	Long-distance transport of L-ascorbic acid in potato	Background Following on from recent advances in plant AsA biosynthesis there is increasing interest in elucidating the factors contributing to the L -ascorbic acid (AsA) content of edible crops. One main objective is to establish whether in sink organs such as fruits and tubers, AsA is synthesised in situ from imported photoassimilates or synthesised in source tissues and translocated via the phloem. In the current work we test the hypothesis that long-distance transport is involved in AsA accumulation within the potato tuber, the most significant source of AsA in the European diet. Results Using the EDTA exudation technique we confirm the presence of AsA in the phloem of potato plants and demonstrate a correlation between changes in the AsA content of source leaves and that of phloem exudates. Comparison of carboxyflourescein and AgNO 3 staining is suggestive of symplastic unloading of AsA in developing tubers. This hypothesis was further supported by the changes in AsA distribution during tuber development which closely resembled those of imported photoassimilates. Manipulation of leaf AsA content by supply of precursors to source leaves resulted in increased AsA content of developing tubers. Conclusion Our data provide strong support to the hypothesis that long-distance transport of AsA occurs in potato. We also show that phloem AsA content and AsA accumulation in sink organs can be directly increased via manipulation of AsA content in the foliage. We are now attempting to establish the quantitative contribution of imported AsA to overall AsA accumulation in developing potato tubers via transgenic approaches.	Background L -Ascorbic acid (AsA), the reduced form of vitamin C, is an essential antioxidant for many biological systems and must be obtained via the diet by humans, primates and a few other animals which are unable to synthesise AsA endogenously [ 1 ]. The main dietary source of AsA for all these organisms are plants and insufficient intake of this micronutrient results in the onset of a debilitating disease (scurvy) and eventually death. In spite of its obvious relevance for humankind, our understanding of how plants synthesise AsA is still rudimentary. It was only in 1998 that an evidence-backed AsA biosynthetic pathway in plants was put forward [ 2 ]. The original proposal was supported by the characterisation of Arabidopsis thaliana AsA-deficient mutants ( vtc1 ) which were defective in the activity of a pathway enzyme [ 3 ]. Since then, other AsA-deficient A. thaliana mutants have been identified which do not seem to be affected in any of the known pathway genes [ 4 , 5 ]. Additionally, up-regulation of AsA accumulation in plants has been achieved via ectopic over-expression of biosynthetic genes unrelated to the proposed biosynthetic pathway [ 6 - 9 ]. This has led to proposals of additional steps, branches or alternative routes to the original pathway [ 7 , 9 , 10 ]. Many of the recent advances in plant AsA biosynthesis have been obtained through investigations of model systems such as tobacco or Arabidopsis thaliana [ 2 , 3 , 9 , 11 ]. On the other hand, our understanding of the mechanisms controlling AsA accumulation in the edible parts of crop plants (e.g. fruits and vegetables), which represent the main dietary source of vitamin C [ 12 ] remains limited. One problem is that AsA functions outside the chloroplasts are much less understood compared with those associated with photosynthetic metabolism [ 13 ]. We have also no biological or taxonomic explanation for the massive variability in AsA contents in sink organs such as fruits which can contain over 3000 mg/100 g FW in the fruits of camu camu ( Mirciaria dubia ) [ 14 ] or less than 10 mg/100 gFW as is the case for grapes, apples or plums [ 15 ]. Large variability is also found in non-green vegetables [ 16 ] whilst dry seeds are completely devoid of AsA [e.g. [ 17 ]]. These examples highlight the high degree of tolerance for AsA content in storage organs and suggest the biological feasibility for the development of AsA-rich crop products. To this end, one key question that needs to be answered is whether AsA accumulation in sink organs occurs as a result of biosynthesis in situ or import from the foliage. Recently Franceschi and Tarlyn [ 18 ] observed long distance movement of 14 C-AsA from leaves to flowers and root tips in model systems such as A. thaliana and Medicago sativa . Research in our laboratory demonstrated the occurrence of AsA in the phloem of a number of crop plants and we also observed that the plant phloem was capable of supporting active AsA biosynthesis from a number of precursors [ 19 ]. We have previously shown that rates of AsA accumulation are highest soon after organ formation in sink organs such as blackcurrant berries [ 20 ] and potato tubers [ 21 ] when sink activity is strongly induced [ 22 ]. Taken together, these findings indicate that sink-source relationships may play an important role in defining AsA accumulation in sink organs. In this work we have extended our investigation on AsA accumulation in potato tubers, the main source of vitamin C in the European diet [ 23 ]. We present here evidence for the implication of long-distance transport in AsA accumulation in potato tubers. Additionally, we show that artificial increase of AsA content in the foliage results in phloem AsA enrichment and AsA accumulation in tuberising stolons. Our findings have implications for the development of strategies for increasing the nutritional value of crop plants. Results Detection of AsA in the phloem Fig. 1 shows HPLC chromatograms of potato source leaf phloem exudates collected from severed petioles in buffer containing either EDTA or CaCl 2 . Total AsA (AsA t ; L -ascorbic acid + dehydroascorbic acid) appeared as the largest peak of absorbance at 245 nm retained by the column in both cases. When exudates were collected in the presence of CaCl 2 instead of EDTA, all exudate derived peaks were strongly reduced due to the reduction in exudation caused by callose gelation [ 24 ]. In experiments to test the stability of authentic AsA (final concentration 0.1 μM) in the presence of EDTA or CaCl 2 exudation buffers less than 5% oxidation was observed in either case over 90 min (data not shown). AsA localisation to the vascular tissue was confirmed histochemically in sections of potato stems and tubers incubated with ethanolic AgNO 3 at 3°C (Fig. 2 ). Intense deposition of metallic silver was observed in the vasculature of stems. In tubers metallic silver deposits appeared as short strands or well defined spots, sometimes in the perimedullary zone and also in the cortex. Although AgNO 3 staining and CFDA treatments could not be carried out on the same section due to interference between the treatments, a clear similarity was observed in the pattern of metallic silver deposits and the distribution of fluorescence in stem and tuber sections. No AgNO 3 staining was observed in control sections pre-incubated with 1% CuSO 4 for 18 h in order to oxidise AsA (data not shown) [ 25 ]. AsA distribution during tuber development Fig. 3 shows AsA t distribution along the axis of stolons and developing tubers. In non-swelling stolons the AsA t content was maximal in the apical section with a sharp basipetal decline such that the more basal sections contained approximately 10% of the AsA t content found in the apex. Tuberising stolons showed a 50% reduction in the AsA t content of the apical section. This was accompanied by increases in the subapical 3–10 mm sections. In developing tubers there was a very substantial increase in the AsA t content in the sub-apical region, corresponding to the swelling area. AsA transport Figure 4 shows the changes in AsA t content of source leaves of potato plants acclimatised for 14 days in cabinets with artificial diurnal dark/light cycles. The cabinets were off-phased so that plants at different stages in their lighting regime were available at any given time. The AsA t content of source leaves was monitored for 24 h in plants sampled simultaneously from the two cabinets. At the end of the dark phase, leaf AsA t content was approximately 12 mg/100 gFW and it progressively increased following artificial sunrise. After 10 h of light, the leaf AsA t content increased to approximately 30 mg/100 gFW after which it leveled off. Following artificial sunset the AsA t level gradually fell back to approximately 12 mg/100 gFW within 6–10 h. Table 1 shows the AsA t content of source leaves and tuberising stolons collected at 12.00 h from the light-phase or dark-phase plants (after collection of phloem exudate for 90 min in a prehumidified atmosphere in leaves). In the table are also reported the chromatographic AsA t peak areas from phloem exudates collected from source leaves and tuberising stolons of the same plants. The exudate data are reported as peak areas as the exact volume of exudate collected could not be established. Whilst the leaves from light-phase plants contained over twice the AsA t level of leaves from dark-phase plants, no significant difference was found in the AsA t content of tuberising stolons from the two sets of plants. The AsA t peak area in chromatograms of light-phase leaf exudates was 1.8-fold larger than that obtained from dark-phase leaves. In tuberising stolons the increase in exudate peak area was more pronounced (4.6-fold). The correlation between leaf AsA t content and the AsA t levels of phloem exudates was also investigated in glasshouse-grown plants following the supply of a range of AsA biosynthetic precursors using the flap technique [ 18 ]. Figure 5 shows the chromatographic AsA t peak area of exudates from leaves pre-treated for 24 h with 20 mM MES pH 5.5, 2 mM CaCl 2 alone (control) or containing precursors at a final concentration of 25 mM. Incubation with D -glucose ( D -Glc) resulted in a slight (10%) reduction in leaf AsA t content compared with the control but no significant change in AsA t was detected in the exudates. By contrast, supply of L -galactose ( L -Gal) or L -galactono-1,4-lactone ( L -GalL) increased the AsA t content of leaves (4.9 and 6.2-fold respectively) and, more substantially, of exudates (10.8 and 11.2-fold respectively). With all treatments, the replacement of EDTA with CaCl 2 in collection wells resulted in a significant reduction of AsA t peak area in the exudate chromatograms. An artificial increase of AsA t content in the totality of source foliage of whole plants was obtained through the supply of the direct AsA precursor L -GalL via the flap technique for 24 h to all terminal leaflets of the four lowermost nodes. At the end of the incubation period the AsA t content of source leaves was 2.2-fold higher than in control plants (Fig. 6 ). AsA t content also significantly increased in sink organs such as flowers (33%) and, more substantially in developing tubers (80%) compared with the control. No changes in AsA t levels were observed in petioles, stems or non-tuberising stolons. Discussion AsA in the potato phloem We have previously shown that AsA represented the largest peak of absorption at 245 nm retained by the column when phloem sap of potato source leaves obtained by aphid stylectomy was analysed by HPLC [ 19 ]. In the present study we obtained very similar results from phloem samples collected from different organs of the potato plant by a different approach i.e. the EDTA exudation technique [ 24 ]. Localisation of AsA to the vascular tissue of stems and developing tubers was further demonstrated histochemically exploiting the specific interaction between AsA and AgNO 3 at low temperature with ensuing formation of metallic silver deposits [ 25 ]. In particular, the distribution of metallic silver in developing tubers generated a pattern reflecting the intense phloem anastomosis typical of these storage organs [ 26 ] and resembling the phloem network involved in the unloading of labelled assimilates in plants supplied with 14 CO 2 [ 22 ]. This was further evidenced by the close similarity between the silver nitrate staining and distribution of fluorescence in tubers following supply of CFDA to source leaves. We have previously demonstrated that the distribution of CF in developing potato tubers can be used to identify symplastic phloem unloading in these organs [ 22 ], thus the similar distribution of CF and metallic silver deposits in developing tubers implies that transfer of phloem AsA to storage parenchyma cells occurs with the mass flow of assimilates. This hypothesis is further supported by the changes in AsA t distribution along the stolon axis following tuber induction which closely resemble the changes in sucrose content and radiolabelled assimilate distribution observed in plants labelled with 14 CO 2 [ 22 ]. The decline of AsA t in the apices of tuberising stolons accompanied by its accumulation in the subapical region may thus reflect the induction of symplastic phloem unloading in this region resulting in a distal migration of sink activity from the apex. However, this pattern also correlates with changes in the mitotic index along the axis as a result of tuber induction with cessation of cell division in the apical region of the stolon [ 27 ] and its activation in secondary meristems within the swelling region [ 28 ]. This may be relevant in view of the purported role of AsA in cell division [ 29 ]. Long-distance AsA transport in potato In preliminary experiments, we observed diurnal changes in the AsA t content of mature source leaves of potato. Circadian or diurnal oscillations of AsA content in photosynthetic tissues are known [e.g. [ 30 ]] as well as light-induction of AsA accumulation in leaves [e.g. [ 31 ]] and sink organs. Light-induced expression of specific AsA biosynthetic enzymes has been observed [e.g. [ 30 ]] as well as a general increase in the AsA biosynthetic flux [ 11 ]. By growing plants in cabinets with out-of-phase light/dark regimes we were able to simultaneously sample plants in the dark or light phase when maximal differences in foliage AsA t content occurred. Leaf AsA t content showed diurnal changes with the maximal levels (observed during the light-phase) 2-fold higher than the lowest levels observed in the dark. Phloem exudates obtained from source and sink organs of light phase plants always showed significantly higher AsA t levels than exudates obtained from dark phase plants. Indeed, the AsA t content of exudates from tuberising stolons of light phase plants was over 4-fold higher that of dark-phase plants. Our findings indicate that changes in source leaf AsA biosynthesis rapidly impact on the phloem AsA t content resulting in transport of de novo synthesised AsA directly to developing sinks. This was also confirmed in experiments where exogenous AsA precursors such as L -GalL or L -Gal were supplied to source leaves via the flap technique, a treatment which resulted in substantial AsA t enrichment in phloem exudates. We were also able to more than double the AsA t content of the source foliage of whole plants by "bulk" supply of exogenous L -GalL to the majority of source leaves for 24 h. This resulted in significant AsA t increases in sink organs such as flowers and developing tubers. To our knowledge this is the first demonstration that AsA accumulation in the foliage results in AsA increases in storage organs and it may explain the positive effect of light irradiation on AsA content of fruits and vegetables [ 32 ]. The rapid changes in phloem AsA concentration which reflected changes in mesophyll AsA content indicate that AsA storage in the foliage does not occur, unlike the case of assimilated carbohydrates. It is tempting to speculate that AsA produced within mesophyll cells is directly taken up by SE/CC complex for translocation, as also deduced from the results reported by Franceschi and Tarlyn [ 18 ]. The kinetics of AsA transport have been studied across a number of plant membranes such as the plasmlemma, the chloroplast membrane, the thylakoid membrane and the tonoplast [ 33 ]. However, to date no transporter has been identified although at least 12 genes encoding putative nucleobase/ascorbate transporters (NATs) have been identified in the Arabidopsis genome. NATs belong to a superfamily of integral membrane transporters, which move purines, pyrimidines or ascorbic acid across biological membranes and which also comprises mammalian transporters specific for ascorbate [ 34 ]. Functional characterisation of these transporters is in progress [ 35 ]. As the plant phloem contains the complete enzymic complement for AsA biosynthesis [ 19 ], we have already speculated that the indirect transfer of AsA from the mesophyll to the phloem could involve the transport of the non-charged intermediate L -Gal across membranes. We show here that exogenous supply of L -Gal or L -GalL to source leaves leads to AsA enrichment of phloem exudates and AsA accumulation in tuberising stolons and developing tubers. Furthermore, we have observed uptake of L -[1- 14 C]Gal by source phloem of potato (data not shown), as already demonstrated in N. benthamiana [ 19 ]. Conclusions The evidence we provide here for long-distance transport of AsA in potato plants, corroborates earlier findings in A. thaliana and M. sativa [ 18 ] and suggests that source-sink AsA translocation may be a general occurrence in plants. What remains to be established is the relative contribution of phloem-derived AsA (whether synthesised in the mesophyll or in the phloem itself) to overall AsA accumulation in potato tubers. Slices excised from developing potato tubers can synthesise AsA from a variety of substrates (Hancock and Viola, unpublished). Moreover, microtubers obtained in vitro from nodal cuttings cultured in the dark with sucrose as the sole carbon source contain similar amounts of AsA as field-grown tubers [ 21 ]. These latter findings in particular are difficult to reconcile with the hypothesis of a major role played by long-distance AsA transport in tuber AsA accumulation. The development of transgenic potato plants with tissue and organ-specific down-regulation of AsA biosynthesis will be required to address this issue. Methods Plant material and growth conditions Potato plants cv. Desiree were grown in 40 cm pots in unheated glasshouses under natural light in compost. In order to subject plants to artificial light/dark cycles, plants were transferred after 35 days to Sanyo Fitotron 1700 controlled environment cabinets and maintained for a further 14 days on 14 h–10 h light-dark cycles with day and night temperatures of 22°C and 15°C respectively. Light was provided by 60 W incandescent lamps to provide a photon flux of 900 μmol m 2 s -1 at the top of the canopy. Relative humidity was maintained at a constant 70% and plants were watered daily. In all cases experiments were performed on tuberising plants after 40–60 days from planting. Throughout the text stolons are defined as non-swelling (uniform diameter along terminal 15 mm) or tuberising (swelling 2–5 mm diameter). Swellings between 5–10 mm diameter are defined as developing tubers. Quantification of AsA in plant tissues Tissue was extracted in a mortar and pestle with ice-cold 5% metaphosphoric acid (MPA) containing 5 mM tris(2-carboxyethyl)phosphine hydrochloride TCEP (9:1 v/w). Samples were then held on ice for 60 min to allow reduction of dehydroascorbic acid to AsA therefore, all data are reported as total AsA pool (AsA t ) i.e. reduced L -ascorbic acid + dehydroascorbic acid. Samples were then centrifuged at 16000 g for 5 min at 1°C and AsA t in the supernatant quantified by HPLC according to the method of Hancock et al [ 36 ]. Briefly, 20 μl of sample supernatant were injected onto a 300 × 7.8 mm ID Coregel 64H ion exclusion column (Interaction Chromatography, San Jose, CA, USA) with a 4 × 3 mm ID carbo-H + guard cartridge (Phenomenex, Macclesfield, UK) maintained at 50°C. Mobile phase was 8 mM H 2 SO 4 at 0.6 ml min -1 and AsA t was detected at 245 nm using a Gynkotech UVD 340S diode array detector (Dionex, Camberley, UK). Detection of AsA in the phloem Phloem exudates were collected from the petiole of source leaves or tuberising stolons using an adaptation of the method developed by King and Zeevart [ 24 ]. Following excision of the organs, a portion of the petiole (5 mm) or stolon (10 mm) was removed under water, the sample was rinsed and the cut end transferred to a 0.6 ml reaction tube containing 200 μl 15 mM EDTA pH 7.5. In the case of petioles, samples were transferred to a pre-humidified atmosphere at 20°C and exudate collected for 90 min in the dark. In the case of stolons, exudates were collected from the cut end which remained attached to the plant and moist paper was wrapped around the top of the reaction tube to minimise evaporation. Control samples were run in parallel in which petioles or stolons were incubated in 5 mM CaCl 2 pH 7.5 to induce callose gellation and reduce exudation [ 24 ]. At the end of the incubation, MPA and TCEP were added to the samples to a final concentration of 5% and 5 mM respectively. Following centrifugation (16000 g, 1°C, 5 min), AsA t concentration was determined by HPLC as described above. Histochemical localization of AsA in tubers using the AgNO 3 method was carried out as previously described [ 19 ]. Briefly, tubers were hand sliced to form approximately 2 mm sections, washed in distilled water and fixed and stained in 5% (w/v) AgNO 3 dissolved in 66% (v/v) aqueous ethanol containing 5% (v/v) glacial acetic acid at 3°C in the dark for up to 24 h. The reaction was stopped by washing the tissue twice for 15 min in ethanolic ammonium hydroxide (95% (v/v) 70% ethanol, 5% (v/v) NH 4 OH ACS reagent, Sigma-Aldrich, Dorset, UK) [ 25 ]. Finally the tissue was transferred to 70% (v/v) ethanol and stored at 3°C prior to photography. Transport of carboxyfluorescein in potato plants Phloem transport through potato stems into developing tubers was visualised using the fluorescent transportable molecule carboxyflourescein (CF) as previously described [ 22 ]. Plants were labelled with 20 ml of an aqueous carboxyflourescein diacetate (CFDA) solution (1 mg ml -1 ) introduced via open stomata on the abaxial leaf surface using a plastic syringe. The acetylated compound is able to diffuse across cell membranes unlike its deacetylated derivative CF which is produced in vivo by endogenous esterases and is used as a marker for phloem strands and symplastic unloading from the phloem [ 37 ]. Plants were left to translocate CF for 5 h prior to hand sectioning (2 mm) and examination of stem and sink tissues for fluorescence using a MRC2000 confocal microscope (Bio-Rad, Hemel Hemstead, UK). Supply of precursors to leaves Leaf AsA t levels were manipulated using an adaptation of the 'flap' technique [ 18 ]. An incision (15 mm) was made either side of the midrib of terminal leaflets and the 'flap' formed was placed into a 0.6 ml eppendorf tube containing 500 μl of 20 mM MES pH 5.5, 2 mM CaCl 2 alone or with the addition of various intermediates at a final concentration of 25 mM for 24 h. At the end of the incubation period, leaflets were ground in liquid nitrogen and extracted in 5% MPA containing 5 mM TCEP (9:1 v/w) and the AsA t content measured by HPLC. For measurement of phloem exudates, treated leaflets were excised under water and placed in eppendorf tubes for collection of phloem exudates as described previously. In some experiments the terminal leaflets of the four lower nodes on all stems (between 8 and 10 per plant) were simultaneously supplied with 500 μl 20 mM MES pH 5.5, 2 mM CaCl 2 alone or containing 25 mM L -GalL for 24 h. Four independent plants were used for each treatment. At the end of the incubation individual plants were separated into flowers, source leaves, leaf petioles, plant stems, non-tuberising stolons (terminal 15 mm), or tuberising stolons (swelling portion). Tissues were immediately frozen in liquid nitrogen and lyophilised. Lyophilised tissue was ground to a powder and 3 × 1 g fractions of each tissue were extracted in 5% MPA, 5 mM TCEP (19:1 v/w) and the AsA t content of each extract analysed by HPLC to give an average value for each tissue. No developing tubers larger than 5 mm diameter were present in the plants. Authors' contributions LT undertook most of the physiological and biochemical experiments. RDH participated in the design and coordination of the study, the biochemical experimentation and the writing of the manuscript. SA participated in the physiological, biochemical and histochemical experiments. SH undertook the CFDA labeling and silver staining of potato plants and participated in production of the figures. RV conceived the study, participated in its design and coordination and drafted the manuscript.	/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC521686.xml
514601	STING Millennium Suite: integrated software for extensive analyses of 3d structures of proteins and their complexes	Background The integration of many aspects of protein/DNA structure analysis is an important requirement for software products in general area of structural bioinformatics. In fact, there are too few software packages on the internet which can be described as successful in this respect. We might say that what is still missing is publicly available, web based software for interactive analysis of the sequence/structure/function of proteins and their complexes with DNA and ligands. Some of existing software packages do have certain level of integration and do offer analysis of several structure related parameters, however not to the extent generally demanded by a user. Results We are reporting here about new Sting Millennium Suite (SMS) version which is fully accessible (including for local files at client end), web based software for molecular structure and sequence/structure/function analysis. The new SMS client version is now operational also on Linux boxes and it works with non-public pdb formatted files (structures not deposited at the RCSB/PDB), eliminating earlier requirement for the registration if SMS components were to be used with user's local files. At the same time the new SMS offers some important additions and improvements such as link to ProTherm as well as significant re-engineering of SMS component ConSSeq. Also, we have added 3 new SMS mirror sites to existing network of global SMS servers: Argentina, Japan and Spain. Conclusion SMS is already established software package and many key data base and software servers worldwide, do offer either a link to, or host the SMS. SMS ( S ting M illennium S uite) is web-based publicly available software developed to aid researches in their quest for translating information about the structures of macromolecules into knowledge. SMS allows to a user to interactively analyze molecular structures, cross-referencing visualized information with a correlated one, available across the internet. SMS is already used as a didactic tool by some universities. SMS analysis is now possible on Linux OS boxes and with no requirement for registration when using local files.	Background A need to integrate, visualize and mine large amount of protein structure data is accelerating. In order to accommodate visualization of data originating from several sources and make analysis of protein structure and structural parameters easier, we developed Sting Millennium Suite (SMS). SMS is a web-based suite of programs and databases providing visualization and a complex analysis of molecular sequence and structure for the data deposited at the Protein Data Bank (PDB) [ 1 ]. Using SMS it is possible to analyze: sequence to structure relationships, quality of the structure, nature and volume of atomic contacts of intra and inter chain type, relative conservation of amino acids at the specific sequence position based on multiple sequence alignment, indications of Folding Essential Residue (FER) based on relationship of the residue conservation to the intra-chain contacts, Cα – Cα and Cβ – Cβ distance geometry etc.. Specific emphasis in SMS is given to Interface Forming Residues (IFR) – amino acids that define interactive portion of the protein surfaces. SMS may simultaneously display and analyze previously superimposed structures. Parsing of data from relevant Data Bases (PDB [ 1 ], HSSP [ 2 , 3 ], Prosite [ 4 ]) is one of the key features of integrated SMS environment for structure/function analysis. SMS also has its own built in data bases: Contacts, Interface Contacts, Surface Accessibility, Dihedral Angles and Secondary Structure Elements [ 5 ]. This article is intended to show how Sting Millennium Suite of programs can be useful in the study of protein structure and analysis of its function, emphasizing recent improvement introduced to SMS. The program has extensive built-in instructions and detailed easy-to-use help which user is invited to consult before and during SMS use. Results and Discussion SMS overview In addition to basic macromolecular visualization, SMS is capable of identifying and visualizing the macromolecular interfaces as well as showing and analyzing previously aligned structures. SMS also does visualization of amino acids conservation based on multiple sequence alignments, in the context of three-dimensional protein structure, identification of the nature and volume of atomic contacts of intra and inter-chain type, presentation of data about the quality of a given structure etc.. SMS provides number of modules (SMS components (some of which are to be described in details separately)) to conveniently visualize large amount of physical-chemical, structural and biological information about the proteins with known structure. Variety of one-click-away renderings and color schemes helps to visualize bonding interactions and locations of residues of interest, as well as to localize patterns of evolution/conservation. The interactions which occur in the protein or between protein and its inhibitor/substrate, can be analyzed in great details with SMS. Graphical contacts SMS offers to the user a graphical presentation of inter-atomic contacts established between amino acids in form of the fan. The base point of the fan is the selected amino acid. From the base point a user can detect number of colored lines connecting to other residues (presented by single letter code). Colors of the fan lines follow SMS code of contacts. A specific HTML table displays residue name and number, its pair in contact establishing, type of the contact, distance between contacting atoms and accessibility and entropy of two contacting residues. Such contacts are divided in number of classes: hydrogen bonds, hydrogen bonds with intermediary water molecules, hydrophobic contacts, aromatic ring stacking contacts, electrostatic (attractive and repulsive) contacts and finally disulphide bridges. A special table is built for those interactions across the interface ( IFR Graphical Contacts ). Both Graphical and IFR Contacts are fully integrated with SMS so that information about any particular amino acid is highlighted in simultaneous fashion across sequence, structure and contacts window. The diagram Ramachandran Plot [ 6 ], used for checking the quality of the structure, is presented in SMS using all advantages of Java programming language. Menu options on interactive SMS Ramachandran Plot allow for coupling of data displayed in the dihedral angle window with a window showing the 3D structure of a molecule. Number of subsets among amino acids can be highlighted for better correlation of a 3D structure position and a phi-psi spot. Full integration and data coupling makes this SMS component a breed apart from the similar public domain products. A user may also produce an image in the gif format which is more appropriate for printing of publication quality figures. Again, SMS Ramachandran Plot is fully integrated with other SMS windows, allowing a user to concomitantly see structure and sequence information highlighted according to selection done in Ramachandran plot or in the sequence window. The module Scorpion provides a graphical presentation for simple statistical data on a frequency of occurrence for given amino acid and also for amino acid local environment in terms of class of amino acids surrounding given central residue. The Protein Dossier module provides a graphical report of several important structural characteristics of the PDB entry. It offers a plot from PDB cartoon annotated with color coded scales representing for each amino acid a corresponding temperature factor, solvent accessibility of the chain in isolation and in a complex with other present chains in the PDB file, sequence conservation in (HSSP derived) multiple alignment (relative sequence entropy) and histograms representing the atomic contacts (as in the Graphical contacts module), as well as IFR residue identification and IFR contacts. In addition, comparison of the Secondary Structure annotated by PDB, by DSSP [ 7 ] and by STRIDE [ 8 ] is presented. With STINGpaint it is possible to paint amino acids within multiple alignment of sequences according to two optional color schemes: STING's scheme and William Taylor [ 9 ] color scheme. This has effect on how easily the user can grasp regions of sequence identity. In addition, the user is presented with an entropy bar which facilitates even further pinpointing highly variable positions. The ConSSeq presents a sequence for a given PDB file and a consensus sequence (as found in the HSSP). A consensus sequence is obtained from the sequence alignment of the sequence-wise homologous proteins. Above those two sequences, ConSSeq shows a graphic bars colored by scale of colors according to the sequence conservation. The height of graphic bars is reflecting relative entropy. ConSSeq also offers information about residues present in other homologous sequences, with their respective frequency. For fast inspection of data, this program also generates a sequence logo. Complete interactivity with both sequence and chime-structure frame/window of the SMS is now operational, offering much better conditions for the thorough analysis of structure and sequence (alignment) interdependence. The Java Cα-Cα [Cβ-Cβ] Distance Plot is a diagram where the distances between the α [β] carbon of one residue and all α [β] carbon atoms of other residues, within a single chain of the PDB file, are represented by colored squares in a symmetrical plot. All the above mentioned modules and some others available from SMS, can be accessed either from the STING Millennium's sequence window or entering through the independent entry web page. An extensive list of links is available to increment a volume of information on a protein under the study. In this new SMS release we introduced ProTherm [ 10 ] link, exceptionally important information on protein stability/mutations, provided by the web site of Dr. Akinori Sarai group. The Sting Millennium and some of the SMS components are now capable of importing local files in PDB format. Algorithm and implementation SMS is organized in two logical layers: SMS server and SMS client. The server side is responsible for updating regularly all relevant public domain databases used by SMS. At the same time, SMS server is also responsible for calculation of a number of macromolecular properties for each PDB structure. The SMS client side provides to a user a friendly graphical interface and communicates to the SMS server, sending user's requests and receiving SMS responses. SMS interactive interface has been mostly implemented in the Java programming language, taking advantage of its object oriented design and graphical representation capabilities. Most important Java classes in SMS are dedicated to sequence and structure parameter presentation, depiction and interaction. Additional classes are used for efficient data handling utilities. As it is known, the object oriented software design is suitable especially because of its ease in code reusability and also because it provides interfaces for linking new software modules, resulting in systems easily expandable and built with extended capabilities. In addition, the Java programming language is very attractive to users for reasons of portability a key feature in today's versatile computing world. SMS also make extensive use of the C ++ programming language, mostly for complex calculation of specific parameters. SMS runs in the Netscape browser or in Microsoft Internet Explorer (for Microsoft Windows operating system) and in the Netscape/Mozila (for the Linux OS) and requires installation of Java Plugin 1.3.1 and CHIME. Some restrictions apply, so a user is invited to consult details of SMS Requirements. Users can run the SMS program by selecting a previously deposited structure in the Protein Data Bank, or using local files with pdb format. Input file format for SMS SMS accepts the PDB format files from RCSB/PDB repository and also accepts local files of the same format, at the client end. A user is able to see structure of the local file in chime/SMS structure window as well as a sequence corresponding to this particular structure. The sequence itself is presented in the separate sequence window. Additionally, some other SMS components will work fine with user's local files: Graphical Contacts, IFR Graphical Contacts, SMS Ramachandran Plot, Scorpion, Formiga, Ca-Ca and Cb-CB contacts and Protein Dossier (although the last one might not have all the usual components that it displays for public PDB files). Comparison to other software packages Increase in availability of molecular structure data during the last decade, urged the development of computer applications for sequence/structure analysis and visualization. Consequently, numerous approaches have been made to the problem of sequence/structure visualization and analysis, resulting in diverse software packages: Protein Explorer, Cn3D, Swiss PdbViewer and ProCheck [ 11 - 14 ]. Each of these products seems to have been developed primarily to accomplish specific tasks. Inevitably, these products have differential strengths in areas that they cover, making difficult the task of comparisons and definitely arbitrary to certain extent. SMS, as well as comparable software resources, come with intuitive user friendly GUIs, allowing for easy navigation through the vast amount of structural data. SMS main advantage is the clear presentation of sequence along with the structure in addition to number of visualizing tools for variety of structure describing parameters. In the input layer SMS uses data from public databases: PDB, HSSP, DSSP and SwissProt. Simultaneous display of computed features/parameters/descriptors along with available annotations from above Databanks provides a useful and reach environment, which may complement and in many cases substitute and exceed the already existing tools for sequence/structure/function analysis and visualization. Conclusions Structure analysis is a difficult task due to the large amount of possible parameters/descriptors that can be calculated and associated with the sequence and corresponding structure. The way in which structure data and structure descriptors are stored and displayed, represents a major challenge when interactivity of a user with the data dispersed among many resources is addressed. Several structure viewers already exist, each one of them better suited to different needs and research interests. SMS offers an easy to use computer environment, designed to facilitate concomitant display of as many parameters as possible, coupled in a consistent fashion to each other. Experimental data and calculated information are all embodied in a clear display that offers instantly an intuitive aspect of a given structure and a large amount of biological information at hand. Inspection of the SMS displayed information can lead to valuable conclusions and cover a wide variety of biology issues concerning entire protein families. SMS has already been applied as a didactic tool for learning details about sequence/structure/function relationship in several universities. Future plans to extend the software platform include the ability to handle ever more descriptors/parameters of protein structure with the simultaneous display and analysis including data extracted from the statistical elaboration of common features among members of certain protein families/folds. In order to achieve such goal, we count with most generic yet very usable tools: Chime viewer and JAVA programming language. In addition, we count on growing interest of other research groups in participating in this project, contributing with their data and benefiting from the resulting unification of data format and data display. Issues such as the geometrical increase in the volume of the disk space and available CPU time for updating such a large data base should be taken into account. SMS is available free and can be accessed through the web. A user has to be careful with proper configuration of IT components (Operating System, browser, Chime viewer, Java JER version, firewall warnings) so that SMS can be used to its fullest potential. The detailed online manual/help/tutorial for viewing and analyzing displayed data is available and recommended for frequent consultation. Availability and requirements Project Name: STING Millennium Suite Lab Home Page: Project Home Page: Operating System(s): Servers: Extensively tested on SGI IRIX 6.5, SUN Solaris7.0 and 8.0 and LINUX Red Hat 7.3, 8.0 Clients: MS Windows XP, NT, 2000 with Netscape 7.0 and IE 6.0 SP1, platform with Java Runtime Environment (JRE) 1.3.1 installed and Linux Rad Hat with Mozila/Netscape 7.0 and CrossOver plugin. Chime 2.6 SP3/SP4 (depending on OS and browser used) plugin is essential for structure presentation. Programming Language: JAVA, C++, Fortran, JavaScript Other requirements: Installation of JRE 1.3.1. License: Free for Academic use. Abbreviations SMS: Sting Millennium Suite IFR: Interface Forming Residues FER: Folding Essential Residue PDB: Protein Data Bank RCSB: Research Collaboratory for Structural Bioinformatics GUI: Graphical User Interface Authors' contributions RH created the Graphical User Interface and part of the data processing programming and general procedures for SMS mirror installation, RT worked on general GUI and statistics for SMS access, AJM worked on ConSSeq integration to SMS, JP worked on HTML design and SMS help pages, IO worked on SMS implementation on Linux OS, PK worked on general data interpretation and GUI suggestions, MY worked on mathematical algorithms for parameter calculation, AM carried out most of the data processing and programming, and GN coordinated the whole project, suggesting the general directions and innovating features of the application. All authors have read and accepted the final manuscript.	/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC514601.xml
516770	POSA: Perl Objects for DNA Sequencing Data Analysis	Background Capillary DNA sequencing machines allow the generation of vast amounts of data with little hands-on time. With this expansion of data generation, there is a growing need for automated data processing. Most available software solutions, however, still require user intervention or provide modules that need advanced informatics skills to allow implementation in pipelines. Results Here we present POSA, a pair of new perl objects that describe DNA sequence traces and Phrap contig assemblies in detail. Methods included in POSA include basecalling with quality scores (by Phred ), contig assembly (by Phrap ), generation of primer3 input and automated SNP annotation (by PolyPhred ). Although easily implemented by users with only limited programming experience, these objects considerabily reduce hands-on analysis time compared to using the Staden package for extracting sequence information from raw sequencing files and for SNP discovery. Conclusions The POSA objects allow a flexible and easy design, implementation and usage of perl-based pipelines to handle and analyze DNA sequencing data, while requiring only minor programming skills.	Background Today, many genetics laboratories have access to modern capillary DNA sequencing machines, such as the ABI PRISM 3100, 3700 or 3730. These machines generate vast amounts of raw sequence data with little user intervention. Consequently, the amount of data to be analyzed has expanded and the bottleneck now is the analysis capacity. Data analysis capacity can be increased by higher levels of automation. Investments in infrastructures to process the raw sequencing data in sophisticated but rigid pipelines might be justified for larger laboratories and larger projects but might be too costly for smaller laboratories. In addition, rigid pipelines are too impractical if different projects share run-time on the same machine while requiring (slightly) different analysis procedures (e.g. vector trimming is needed in plasmid sequencing, but needless when sequencing PCR products). Nucleotide sequence analysis can be performed with a variety of software tools. Although the number of console and web-based software tools has grown rapidly, the routine use of data input, output and storage may be inconvenient. Furthermore, for performing a series of analyses with different software tools, the sequence data need to be reformatted to the required data structure. Alternatively, sophisticated software suites that provide an integrated environment often are expensive. Several of the available software solutions are designed to facilitate automated DNA sequence analysis at low cost. Well-known solutions are the Staden package and Bioperl. The Staden Package contains pregap4 and gap4 , full-featured applications with an intuitive graphical user interface [ 1 ]. These programs handle a list of raw sequence reads method-by-method. The programs in the Staden Package typically require a degree of user intervention and thus hands-on time. Alternatively, Bioperl is a group of perl modules describing many genetics and genomics concepts [ 2 ]. For example, it includes the Bio::Seq::SeqWithQuality object that provides some of the basic properties of a raw sequence (i.e. its nucleotide sequence and quality values); the Bio::Tools::Primer3 object provides methods to work with primer3 input and output. However, to build custom DNA sequencing data pipelines, basic programming skills are needed to combine all these modules. Smaller laboratory sites, however, often need to implement versatile pipelines that can be adjusted for any research question that suits the project best; at the same time, they often also do not have dedicated programmers available. Although (semi-)automated procedures have been published by other groups [ 3 , 4 ], these are mostly focused on one particular pipeline and environment. Here, we present POSA, a set of two new perl objects (Read.pm and Contig.pm) that describe a raw sequence and a Phrap contig in detail and are easily implemented in perl-based pipelines. Because these objects provide building blocks for sequencing data analysis pipelines and the actual pipelines are built using perl-scripts, the POSA objects can be used in very diverse settings. Implementation The POSA source code is entirely coded in object-oriented Perl and consists of two objects: Read.pm and Contig.pm. In general, there are two important concepts associated with objects: methods (built-in procedures that can be performed on the object) and properties (describing some of the characteristics of the object). Most methods in the objects rely on the availability of other third-party programs (see Dependencies). Basically, POSA provides a wrapper around these programs and provides easy design and implementation of these programs in automated data analysis. The Read.pm object describes a DNA sequence trace and includes methods for data import from a variety of formats. It relies on Phred [ 5 , 6 ] for import and interpretation of raw sequence data. The original trace data are stored in binary ( scf ) format within the object. Other methods of Read.pm use modules of the Staden Package [ 1 ], such as qclip and vector_clip (if installed). Properties of Read.pm include e.g. the DNA sequence, quality scores, template and vector names and read direction. The Contig.pm object contains a method to assemble contigs of reads using the Phrap program [ 6 ]. The object typically is created based on a list of Read.pm objects and can be exported as alignments or screened for polymorphisms using PolyPhred [ 7 ]. Both the Read.pm and Contig.pm objects were designed with flexibility in mind. To allow a (virtually) unlimited amount of data to be processed, the perl scripts using these objects work sequence-by-sequence rather than method-by-method. Typically, these objects are called from straightforward perl scripts that outline the analysis steps to be performed. Example scripts using the objects can be accessed from the download website. An example of a script and output using the two objects to process a set of reads and annotate sequence polymorphisms from the assembled contig is given in Figure 1 and Figure 2 . Figure 1 A typical script that takes a list of ab1 files for analysis and assembly, reports the contig, and lists the putative SNP positions and SBE primers. Figure 2 Typical output as generated by the script in Figure 1. POSA was developed with perl 5.6.1 and tested on a SuSE linux 8.1 system for abi -files from the ABI PRISM 377 DNA Sequencer and 3100 Genetic Analyzer (Applied Biosystems). Phred , Phrap and PolyPhred versions were 0.000925.c, 0.990329 and 4.05, respectively. Results and Discussion Functionality POSA provides an interface to design and implement automated sequencing data analysis. Sequencing data may be used in a variety of formats and originate from a variety of sources, e.g. data in fasta, abi/ab1 or scf format retrieved from websites or from newly generated traces. In addition, new objects can be initiated from a text file or can be opened from previous stored objects. Subsequently, a variety of methods can be applied, including basecalling and assessment of quality codes (by Phred ), quality clipping, vector clipping, screening for E. coli (or other) sequence, contig assembly (by Phrap ) and analysis. The method asPrimer3 can automatically generate input for the primer3 program [ 8 ] and is available in both objects. To facilitate automated SNP discovery or typing, the SearchSnp s method will generate output as shown in Figure 2 . This method is based on the PolyPhred program and uses the 'rank' argument to set the stringency. Finally, data can be stored in objects, or in files in either exp , scf or fasta format. In addition, the data can be saved in a primer3 input file to allow automated PCR primer design, or data can be saved in MIPE format (i.e. an XML format to store information on PCR experiments; see ). Data on assembled contigs can be exported as a list of reads in a contig, as consensus sequence, as alignment, as putative SNPs, as SBE primers for SNP genotyping or as gff file for visualization in Gbrowse [ 9 ]. Combinations of the diversity of input, analysis and output options allow for a wide spectrum of possible implementations. Examples of possible analysis pipelines include (but are not limited to) BAC-end sequencing with automated PCR primer design for chromosome walking and resequencing of PCR products with SNP annotation either for SNP genotyping or for SNP discovery and SBE primer design. Examples of scripts are provided on the web site . Performance Although it represents only one of the numerous possible POSA-based pipelines, performance of POSA was validated by comparison of SNP discovery with the data after analysis using the Staden package. To do so, 5 PCR products were resequenced from a panel of 16 individuals to identify SNPs. Manual editing using the Staden Package revealed a total of 48 SNPs. Automated analysis using POSA also yielded a total of 48 SNPs with SNP ranking codes 1-3. Together, 41 SNPs were assigned with both manual editing and POSA. The remaining 7 SNPs assigned in manual editing corresponded to SNPs with ranks 4-6 in the POSA analysis. The 7 SNPs that were only assigned by POSA all originated from regions with lower quality sequence. While analysis time was reduced from several hours to a few minutes, POSA assigned SNPs in a way that was highly consistent with manual editing. This was expected because POSA provides options for an integrated analysis pipeline, but essentially is a wrapper around well-established sequence analysis tools like Phred , Phrap and PolyPhred . Intended use and benefits for users POSA is a tool that provides easy and highly automated DNA sequence and contig data analysis using popular analysis tools. Automated sequence analysis reduces analysis time from several hours to a few minutes. Pipelines can easily be expanded or adapted through perl scripts. Writing or altering the perl scripts is straightforward to do for people with only basic computer skills, although more linux/unix experience might be necessary to install the required software (e.g. Phred and Phrap ). Overall, this guaranties easy implementation of highly automated quality pipelines in combination with high flexibility in setup and design. The perl objects are released under an open source license, allowing code improvements by the user community. Conclusions POSA describes a DNA sequence read and a Phrap contig assembly in detail. These objects allow a flexible and easy setup of perl-based pipelines to handle DNA sequencing data, including generating primer3 input and automated SNP discovery, while requiring only little programming skills. Availability and requirements Project name: POSA Project home page: Operating system: platform independent Programming language: Perl 5.6.1 License: Artistic License (Open Source) Requirements • Perl modules: Carp; Statistics::Descriptive; Tie::File; IO::File; POSIX:: Storable. • Phred, Phrap, PolyPhred • Pregap4, gap4 (Staden Package (optional)) • Primer3 (optional) List of abbreviations POSA Perl objects for DNA sequencing data analysis SNP single nucleotide polymorphism abi/ab1ABI PRISM trace file format scf standard chromatogram format exp experiment file format, developed by Staden (see ) MIPE minimum information on PCR experiments (see ) BAC bacterial artificial chromosome PCR polymerase chain reaction SBE single base extension Authors' contributions JA programmed the Perl objects and participated in development of concept and architecture of the software; BJ participated in development of concept and architecture and wrote the manuscript; MG supervised the project. All authors read and approved the final manuscript.	/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC516770.xml
521692	The PedsQL™ Family Impact Module: Preliminary reliability and validity	Background The PedsQL™ Measurement Model was designed to measure health-related quality of life (HRQOL) in children and adolescents. The PedsQL™ 4.0 Generic Core Scales were developed to be integrated with the PedsQL™ Disease-Specific Modules. The newly developed PedsQL™ Family Impact Module was designed to measure the impact of pediatric chronic health conditions on parents and the family. The PedsQL™ Family Impact Module measures parent self-reported physical, emotional, social, and cognitive functioning, communication, and worry. The Module also measures parent-reported family daily activities and family relationships. Methods The 36-item PedsQL™ Family Impact Module was administered to 23 families of medically fragile children with complex chronic health conditions who either resided in a long-term care convalescent hospital or resided at home with their families. Results Internal consistency reliability was demonstrated for the PedsQL™ Family Impact Module Total Scale Score (α = 0.97), Parent HRQOL Summary Score (α = 0.96), Family Functioning Summary Score (α = 0.90), and Module Scales (average α = 0.90, range = 0.82 – 0.97). The PedsQL™ Family Impact Module distinguished between families with children in a long-term care facility and families whose children resided at home. Conclusions The results demonstrate the preliminary reliability and validity of the PedsQL™ Family Impact Module in families with children with complex chronic health conditions. The PedsQL™ Family Impact Module will be further field tested to determine the measurement properties of this new instrument with other pediatric chronic health conditions.	Background Pediatric health-related quality of life (HRQOL) is increasingly acknowledged as an important health outcome measure in clinical trials and health services research and evaluation [ 1 , 2 ]. Additionally, in pediatric chronic health conditions, the impact of disease and treatment on family functioning is a salient concern given the essential role of the family in child adaptation to disease [ 3 - 5 ]. Within this context, the impact of pediatric chronic health conditions on the family has been conceptualized within a theoretical risk and resistance framework, in which parent adjustment and the family system as a whole have been identified at increased risk [ 6 ]. Although there are a number of well-developed generic measures of family functioning, such as the Family Environment Scale [ 7 ], instruments that specifically measure the impact of pediatric chronic health conditions on parent and family functioning are less common. The two most widely utilized family impact instruments are the Impact on Family Scale and the Child Health Questionnaire (CHQ). The Impact on Family Scale-Revised is a brief unidimensional instrument that measures one factor of general negative impact on the social and familial systems and has demonstrated good reliability and validity in the samples tested [ 8 ]. The CHQ, a well validated instrument which contains scales measuring child HRQOL [ 9 ], contains a scale measuring whether the child's health or behavior limited family activities or caused family conflict. The CHQ also contains two parent self-report scales which measure the impact of the child's health on parent worry or concern and limitations in meeting their own needs. Although these two well-developed measures existed when we conceptualized the PedsQL™ Family Impact Module, after an analysis of the items and scales of the existing instruments, we felt that a PedsQL™ Family Impact Module would make a significant contribution to the literature by creating a multidimensional instrument that could stand alone, or be easily integrated into the PedsQL™ Measurement Model [ 10 ]. The PedsQL™ Measurement Model includes not only generic health-related quality of life [ 11 - 13 ] and disease-specific measurement instruments [ 14 - 18 ], but also generic measures of fatigue [ 15 , 19 ], healthcare satisfaction [ 20 , 21 ] and evaluations of the healthcare built environment [ 21 ]. Thus, we envisioned a Family Impact Module that would contribute to the literature by identifying items and scales which were not redundant with existing instruments, and which would further enhance the measurement options available through the PedsQL™ Measurement Model. In this context, the PedsQL™ Family Impact Module was developed and initially field tested in families with medically fragile children with complex chronic medical conditions as part of our evaluation of the healing environment of a Children's Convalescent Hospital [ 21 ]. In order to provide a contrast group to these children in this long-term care facility, we selected a population of children with comparable complex chronic medical conditions who were residing at home with their families. Since these children's severe medical conditions prevented them from providing self-report, the PedsQL™ Family Impact Module was designed as a parent proxy-report instrument. This study investigates the preliminary reliability and validity of the PedsQL™ Family Impact Module in medically fragile children with complex chronic health conditions. We hypothesized that the PedsQL™ Family Impact Module would distinguish between families in which the child resided at home versus those whose child resided in a long-term care facility based on the extant literature on pediatric chronic health conditions and the impact on parents and families [ 5 , 6 ]. Method Participants and Settings Participants were the parents of 23 medically fragile pediatric patients with complex chronic health conditions, such as severe cerebral palsy and birth defects. Participants from the Children's Convalescent Hospital (CCH) were parents of 12 pediatric patients who were residents of this long-term care facility. For each CCH family, the family member who completed the PedsQL™ Family Impact Module was the resident's mother. Participants from the REACH program (an outpatient program designed to reach out to families who choose to take care of their medically fragile children at home) were the parents of 11 pediatric patients. For each REACH family except one, the family member who completed the PedsQL™ was the patient's mother. PedsQL™ Family Impact Module The 36-item PedsQL™ Family Impact Module Scales encompass 6 scales measuring parent self-reported functioning: 1) Physical Functioning (6 items), 2) Emotional Functioning (5 items), 3) Social Functioning (4 items), 4) Cognitive Functioning (5 items), 5) Communication (3 items), 6) Worry (5 items), and 2 scales measuring parent-reported family functioning; 7) Daily Activities (3 items) and 8) Family Relationships (5 items). Items and scales were developed through focus groups, cognitive interviews and pre-testing measurement development protocols [ 10 , 11 ], and our prior research and clinical experiences with children with chronic health conditions and their families. Table 1 contains a general description of the scale items. Table 1 PedsQL™ Family Impact Module – general content of scales Parent Functioning # Items General Content Physical Functioning 6 Problems with physical functioning, including feeling tired, getting headaches, feeling weak, and stomach problems Emotional Functioning 5 Problems with emotional functioning, including anxiety, sadness, anger, frustration, and feeling helpless or hopeless Social Functioning 4 Problems with social functioning, including feeling isolated, difficulty getting support from others, and finding time or energy for social activities Cognitive Functioning 5 Problems with cognitive functioning, including difficulty maintaining attention, remembering things, and thinking quickly Communication 3 Problems with communication, including others not understanding the family's situation, difficulty talking about child's health condition, and communicating with health professionals Worry 5 Problems with worrying, including worrying about child's treatments and side effects, about others' reactions to child's condition, about the effect of the illness on the rest of the family, and about child's future Family Functioning # Items General Content Daily Activities 3 Problems with daily activities, including activities taking more time and effort, difficulty finding time and energy to finish household tasks Family Relationships 5 Problems with family relationships, including communication, stress, and conflicts between family members, and difficulty making decisions and solving problems as a family Total Score is computed by averaging all 36 items. Parent HRQOL Summary Score is computed by averaging 20 items in Physical, Emotional, Social, and Cognitive Functioning. Family Summary Score is computed by averaging 8 items in Daily Activities and Family Relationships. The PedsQL™ Family Impact Module was developed as a parent-report instrument. A 5-point response scale is utilized (0 = never a problem; 4 = always a problem). Items are reverse-scored and linearly transformed to a 0–100 scale (0 = 100, 1 = 75, 2 = 50, 3 = 25, 4 = 0), so that higher scores indicate better functioning (less negative impact). Scale Scores are computed as the sum of the items divided by the number of items answered (this accounts for missing data). If more than 50% of the items in the scale are missing, the Scale Score is not computed [ 22 ]. Although there are other strategies for imputing missing values, this computation is consistent with the previous PedsQL™ peer-reviewed publications, as well as other well-established HRQOL measures [ 23 , 24 ]. The PedsQL Family Impact Module Total Scale Score is the sum of all 36 items divided by the number of items answered. The Parent HRQOL Summary Score (20 items) is computed as the sum of the items divided by the number of items answered in the Physical, Emotional, Social, and Cognitive Functioning Scales. The Family Functioning Summary Score (8 items) is computed as the sum of the items divided by the number of items answered in the Daily Activities and Family Relationships Scales. Procedure The PedsQL™ Family Impact Module was mailed to families whose children were residents at the CCH and outpatients in the REACH program, along with a self-addressed stamped envelope in which to return the survey to the research team. A letter was included in the packet explaining the study, the confidentiality with which their data would be treated, and that the healthcare staff would not see this information. The protocol was approved by the Institutional Review Board at Children's Hospital and Health Center, San Diego. Statistical Analysis Scale internal consistency reliability was determined by calculating Cronbach's coefficient alpha [ 25 ]. Scales with reliabilities of 0.70 or greater are recommended for comparing patient groups, while a reliability criterion of 0.90 is recommended for analyzing individual patient scale scores [ 26 , 27 ]. Construct validity for the PedsQL™ Family Impact Module was determined utilizing the known-groups method. The known-groups method compares scale scores across groups known or expected to differ in the construct being investigated. In this study, PedsQL™ Family Impact Module scores in groups differing in residence of the child (Convalescent Hospital inpatient sample versus REACH outpatient sample) were computed [ 28 , 29 ], using independent sample t-tests. We hypothesized that families whose children were residents in the Convalescent Hospital would report significantly higher scores (less negative impact) than families whose children were being taken care of at home based on the exant literature on pediatric chronic health conditions, families, and parental adjustment [ 6 ]. In order to determine the magnitude of the differences between families, effect sizes were calculated [ 30 ]. Effect size as utilized in these analyses was calculated by taking the difference between the Convalescent Hospital sample mean and the REACH sample mean, divided by the pooled standard deviation. Effect sizes for differences in means are designated as small (.20), medium (.50), and large (.80) in magnitude [ 30 ]. Statistical analyses were conducted using SPSS for Windows. Results Means and Standard Deviations Table 2 presents the means and standard deviations of the Convalescent Hospital inpatient sample and the REACH outpatient sample. Table 2 Scale descriptives for PedsQL™ Family Impact Module: Comparisons across CCH and REACH samples CCH Sample REACH Sample Scale # Items N Mean SD N Mean SD Difference Effect Size Total Impact Score 36 12 81.00 17.06 11 62.49 17.26 18.51 1.08 Parent HRQOL Summary 20 12 83.75 15.55 11 62.94 19.83 20.81* 1.17 Physical Functioning 6 12 82.99 17.36 11 53.03 22.83 29.26* 1.45 Emotional Functioning 5 12 78.33 18.26 11 64.48 26.59 13.85 0.61 Social Functioning 4 12 85.42 17.34 11 61.93 25.99 23.49 1.07 Cognitive Functioning 5 12 88.75 12.81 11 74.09 18.95 14.66* 0.91 Communication 3 12 73.61 24.58 11 52.15 24.67 21.46* 0.87 Worry 5 12 69.17 21.09 11 56.82 25.52 12.35 0.53 Family Summary 8 12 84.27 20.47 11 68.81 24.11 15.46 0.69 Daily Activities 3 12 85.14 24.75 11 51.89 31.48 33.25*** 1.18 Family Relationships 5 12 83.75 23.07 11 78.95 27.62 4.80 0.19 Note: Higher values equal better health-related quality of life and family functioning. HRQOL = health-related quality of life; CCH = Children's Convalescent Hospital. REACH = outpatient sample. p < .05, p < .02, **p < .01; equal variances not assumed. Effect sizes are designated as small (.20), medium (.50), and large (.80). Internal Consistency Reliability Internal consistency reliability alpha coefficients for the PedsQL™ Family Impact Module Scales are presented in Table 3 . The scales exceeded the minimum reliability standard of 0.70 [ 26 ]. Most PedsQL™ Family Impact Module Scales approached or exceeded the reliability criterion of 0.90 recommended for analyzing individual patient scale scores [ 26 , 27 ]. Table 3 PedsQL™ Family Impact Module: Internal consistency reliability for total, CCH, and REACH samples Scale Total N CCH N REACH N Total Impact Score .97 23 .97 12 .95 11 Parent HRQOL Summary Score .96 23 .96 12 .95 11 Physical Functioning .91 23 .84 12 .88 11 Emotional Functioning .90 23 .83 12 .93 11 Social Functioning .88 23 .87 12 .88 11 Cognitive Functioning .93 23 .93 12 .91 11 Communication .88 23 .79 12 .95 11 Worry .82 23 .80 12 .84 11 Family Summary Score .90 23 .93 12 .89 11 Daily Activities .91 23 .95 12 .83 11 Family Relationships .97 23 .98 12 .96 11 Note: HRQOL = health-related quality of life; CCH = Children's Convalescent Hospital. REACH = outpatient sample. Construct Validity Table 2 presents the effect sizes and t-test results of the PedsQL™ Family Impact Module Scales for families with children at the CCH and REACH. The effects sizes were all in the medium to large effect size range except for one scale. Although the small sample size decreases the probability of detecting statistically significant differences, 7 of the 11 comparisons were statistically significant. Discussion This study presents the preliminary reliability and validity of the newly developed PedsQL™ Family Impact Module. All internal consistency reliabilities exceeded the recommended minimum alpha coefficient standard of 0.70 for group comparisons, with most scales approaching or exceeding an alpha of 0.90, recommended for individual patient analysis [ 26 ]. The PedsQL™ Family Impact Module Scales performed as hypothesized utilizing the known-groups method. Where statistically significant differences existed between families with children at the CCH and REACH, REACH families were lower functioning, generally confirming the hypothesis that families whose medically fragile children live in a residential facility are higher functioning than those whose children live in the home. The present findings have certain limitations. Information on nonparticipants and an accurate response rate were not available, which may limit the generalizability of the findings. The generalizability of the findings is further limited by the small sample size and the selection of medically fragile children with complex chronic medical conditions. Whether the instrument would perform well in groups of children with other chronic health conditions is a matter of empirical inquiry. Given that instrument validation is an iterative process and consistent with this paradigm, the PedsQL™ Family Impact Module will be further field tested in other pediatric chronic health conditions with larger populations of children. Conclusion The study demonstrates the preliminary reliability and validity of the PedsQL™ Family Impact Module, an instrument designed to assess the impact of pediatric chronic health conditions on parents' HRQOL and family functioning. As predicted, families of children with medically fragile conditions who resided in a children's convalescent hospital were higher functioning than families of similar children who resided at home. List of Abbreviations HRQOL Health-Related Quality of Life PedsQL™ Pediatric Quality of Life Inventory™ Authors' Contributions JWV and PD conceptualized the rationale and design of the study. JWV designed the instrument and drafted the manuscript. PED participated in the instrument design and coordination of initial data collection. SAS performed the statistical analysis and participated in study coordination, instrument development, and data collection. TMB participated in study conceptualization and design, instrument development, and data collection. All authors read and approved the final manuscript.	/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC521692.xml
314300	Gene Expression Signature of Fibroblast Serum Response Predicts Human Cancer Progression: Similarities between Tumors and Wounds	Cancer invasion and metastasis have been likened to wound healing gone awry. Despite parallels in cellular behavior between cancer progression and wound healing, the molecular relationships between these two processes and their prognostic implications are unclear. In this study, based on gene expression profiles of fibroblasts from ten anatomic sites, we identify a stereotyped gene expression program in response to serum exposure that appears to reflect the multifaceted role of fibroblasts in wound healing. The genes comprising this fibroblast common serum response are coordinately regulated in many human tumors, allowing us to identify tumors with gene expression signatures suggestive of active wounds. Genes induced in the fibroblast serum-response program are expressed in tumors by the tumor cells themselves, by tumor-associated fibroblasts, or both. The molecular features that define this wound-like phenotype are evident at an early clinical stage, persist during treatment, and predict increased risk of metastasis and death in breast, lung, and gastric carcinomas. Thus, the transcriptional signature of the response of fibroblasts to serum provides a possible link between cancer progression and wound healing, as well as a powerful predictor of the clinical course in several common carcinomas.	Introduction Since the classic observations of the many histologic similarities between the tumor microenvironment and normal wound healing, it has been proposed that tumor stroma is “normal wound healing gone awry” ( Dvorak 1986 ). During normal wound healing, coagulation of extravasated blood initiates a complex cascade of signals that recruit inflammatory cells, stimulate fibroblast and epithelial cell proliferation, direct cell migration, and induce angiogenesis to restore tissue integrity. Many of these normally reparative processes may be constitutively active in the tumor milieu and critical for tumor engraftment, local invasion, and metastasis to distant organs ( Bissell and Radisky 2001 ). Indeed, keratinocytes from the wound edge transiently exhibit many similarities to their transformed counterparts in squamous cell carcinomas ( Pedersen et al. 2003 ). Epidemiologically, chronic wound and inflammatory states are well-known risk factors for cancer development: the connection between cirrhosis and liver cancer, gastric ulcers and gastric carcinoma, and burn wounds and subsequent squa-mous cell carcinoma (so-called Majorlin's ulcer) are but a few examples. In the genetic blistering disorder recessive dystrophic epidermolysis bullosa, nearly 80% of the patients develop aggressive squamous cell carcinoma in their lifetime ( Mallipeddi 2002 ), attesting to the powerful inductive environment of wounds for cancer development. In recent years, the roles of angiogenesis, extracellular matrix remodeling, and directed cell motility in cancer progression have been intensely studied ( Bissell and Radisky 2001 ). Nonetheless, a comprehensive molecular view of wound healing and its relationship to human cancer is still lacking. Thus, there is currently no established method to quantify the risk of cancer from wounds diagnostically or to intervene therapeutically. The complete sequence of the human genome and the advent of microarray technology have spurred a revolution in the classification and diagnosis of human cancers ( Golub et al. 1999 ; Alizadeh et al. 2000 ; Perou et al. 2000 ; Sorlie et al. 2001 ; van 't Veer et al. 2002 ; Ramaswamy et al. 2003 ). By detailing the expression level of thousands of genes simultaneously in tumor cells and their surrounding stroma, gene expression profiles of tumors can provide “molecular portraits” of human cancers. The variations in gene expression patterns in human cancers are multidimensional and typically represent the contributions and interactions of numerous distinct cells and diverse physiological, regulatory, and genetic factors. Although gene expression patterns that correlate with different clinical outcomes can be identified from microarray data, the biological processes that the genes represent and thus the appropriate therapeutic interventions are generally not obvious. In this study, we explore an alternative strategy to infer physiologic mechanisms in human cancers. We began with a gene expression profile derived from a cell culture model of a physiological process. The in vitro expression profile is used to guide interpretation of publicly available gene expression data from human cancers and thereby test a specific hypothesis. In principle, this strategy allows one to connect the controlled and dynamic molecular perturbations possible in vitro with the complex biology of human clinical samples in a comprehensive and quantitative fashion. Fibroblasts are ubiquitous mesenchymal cells in the stroma of all epithelial organs and play important roles in organ development, wound healing, inflammation, and fibrosis. Fibroblasts from each anatomic site of the body are differentiated in a site-specific fashion and thus may play a key role in establishing and maintaining positional identity in tissues and organs ( Chang et al. 2002 ). Tumor-associated fibroblasts have previously been shown to promote the engraftment and metastasis of orthotopic tumor cells of many epithelial lineages ( Elenbaas and Weinberg 2001 ). We previously observed that the genomic response of foreskin fibroblasts to serum, the soluble fraction of coagulated blood, represents a broadly coordinated and multifaceted wound-healing program that includes regulation of hemostasis, cell cycle progression, epithelial cell migration, inflammation, and angiogenesis ( Iyer et al. 1999 ). We hypothesized that if one could identify a canonical gene expression signature of the fibroblast serum response, this signature might provide a molecular gauge for the presence and physiologic significance of the wound-healing process in human cancers. Results Identification of a Stereotyped Genomic Response of Fibroblasts to Serum We previously observed that the global transcriptional response of fibroblasts to serum integrates many processes involved in wound healing ( Iyer et al. 1999 ). Because fibroblasts from different anatomic sites are distinct differentiated cells with characteristic gene expression profiles ( Chang et al. 2002 ), we investigated whether the genomic responses to serum varied significantly among fibroblasts cultured from different anatomic sites. Fifty fibroblast cultures derived from ten anatomic sites were cultured asynchronously in 10% fetal bovine serum (FBS) or in media containing only 0.1% FBS. Analysis of the global gene expression patterns, using human cDNA microarrays containing approximately 36,000 genes, revealed that although fibroblasts from different sites have distinctly different gene expression programs, they share a stereotyped gene expression program in response to serum ( Figure 1 A). Selection for genes that were concordantly induced or repressed by most types of fibroblasts yielded 677 genes, represented by 772 cDNA probes, of which 611 are uniquely identified by UniGene ( http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?db=unigene ). This common genomic response to serum includes induction of genes that represent entry into and progression through the cell cycle (e.g., E2F1 , FOXM1 , PTTG1 ), induction of cell motility (e.g., CORO1C , FLNC ), extracellular matrix remodeling ( LOXL2 , PLOD2 , PLAUR ), cell–cell signaling ( SDFR1 , ESDN , MIF ), and acquisition of a myofibroblast phenotype (e.g., TAGLN , TPM2 , MYL6 ). Analysis of the public Gene Ontology (GO) annotation of the fibroblast serum response genes confirmed a significant enrichment of genes involved in cell proliferation, blood coagulation, complement activation, secretory protein synthesis, angiogenesis, and proteolysis, reflecting the diverse roles that fibroblasts may play during wound healing (Worksheet 9 in Dataset S2 ). Figure 1 Identification and Annotation of a Common Serum Response in Fibroblasts (A) The fibroblast common serum response. Genes with expression changes that demonstrate coordinate induction or repression by serum in fibroblasts from ten anatomic sites are shown. Each row represents a gene; each column represents a sample. The level of expression of each gene in each sample, relative to the mean level of expression of that gene across all the samples, is represented using a red–green color scale as shown in the key; gray indicates missing data. Representative genes with probable function in cell cycle progression (orange), matrix remodeling (blue), cytoskeletal rearrangement (red), and cell–cell signaling (black) are highlighted by colored text on the right. Three fetal lung fibroblast samples, cultured in low serum, which showed the most divergent expression patterns among these samples (in part due to altered regulation of lipid biosynthetic genes [ Chang et al. 2002 ]), are indicated by blue branches. (B) Identification of cell cycle-regulated genes in the common serum response signature. The expression pattern of each of the genes in (A) during HeLa cell cycle over 46 h after synchronization by double thymidine block is shown ( Whitfield et al. 2002 ). Transit of cells through S and M phases during the timecourse, verified by flow cytometry, is indicated below. Approximately one-quarter of genes demonstrate a periodic expression patterns and are therefore operationally annotated as cell cycle genes; the remainder of the genes are used in further analyses to define the CSR. (C) Validation of annotation by temporal expression profiles. Timecourse of gene expression changes in a foreskin fibroblast culture after shifting from 0.1% to 10% FBS is shown. Global gene expression patterns were determined using cDNA microarrays containing 36,000 genes; genes whose transcript levels changed by at least 3-fold during the timecourse and those in (A) are displayed. The cell cycle genes identified in the analysis illustrated in (B) were found to have a distinct temporal expression pattern with coordinate upregulation at 12 h. One of the most consistent and important responses of human cells to serum is proliferation. Abnormal cell proliferation is also a consistent characteristic of cancer cells, irrespective of any possible involvement of a wound-healing response. We therefore sought to eliminate the contributions of genes directly related to cell proliferation, to improve the specificity of a genomic signature of the fibroblast serum response. To identify features directly related to cell cycle progression, we examined the expression pattern of these 677 genes during the cell cycle (in HeLa cells) ( Whitfield et al. 2002 ). Despite the well-known role of serum as a mitogen, only one-quarter (165 out of 677 genes) of the fibroblast serum response genes showed periodic expression during the cell cycle ( Figure 1 B). The majority of the genes whose expression levels in fibroblasts showed the most consistent response to serum exposure do not appear simply to reflect cell growth or division; these 512 serum-responsive and cell cycle-independent genes are operationally defined as the fibroblast core serum response (CSR). Comparison of the common fibroblast serum response with a detailed analysis of the temporal program of gene expression following serum exposure in foreskin fibroblasts confirmed that the cell cycle genes and the CSR have distinct temporal profiles during serum stimulation and are thus distinguishable biological processes ( Figure 1 C). Expression of Fibroblast CSR in Human Cancers Because serum (as distinct from plasma and normal extracellular fluid) is encountered in vivo only at sites of tissue injury or remodeling and induces in fibroblasts a gene expression response suggestive of wound healing, we reasoned that expression of fibroblast CSR genes in tumors might gauge the extent to which the tumor microenvironment recapitulates normal wound healing. We examined the expression of genes comprising the fibroblast CSR in publicly available microarray data from a variety of human cancers and their corresponding normal tissues. To facilitate visualization and analysis, we organized the gene expression patterns and samples by hierarchical clustering ( Eisen et al. 1998 ). Remarkably, we observed a predominantly biphasic pattern of expression for the fibroblast CSR in diverse cancers, including breast cancers, lung cancers, gastric cancers, prostate cancers, and hepatocellular carcinoma. Expression levels of genes that were activated by serum in fibroblasts varied coordinately in tumors, and genes that were repressed by serum in fibroblasts were mostly expressed in a reciprocal pattern ( Figure 2 ). Figure 2 Survey of Fibroblast CSR Gene Expression in Human Cancers Expression patterns of available CSR genes in over 500 tumors and corresponding normal tissues were extracted, filtered as described in Materials and Methods , and organized by hierarchical clustering. The response of each gene in the fibroblast serum response is shown on the right bar (red shows activated; green shows repressed by serum). The strong clustering of the genes induced or repressed, respectively, in fibroblasts in response to serum exposure, based solely on their expression patterns in the tumor samples, highlights their coordinate regulation in tumors. The dendrograms at the top of each data display represent the similarities among the samples in their expression of the fibroblast CSR genes; tumors are indicated by black branches, normal tissue by green branches. In each of the tumor types examined, the expression pattern of the fibroblast CSR genes in normal tissues closely approximated that seen in quiescent fibroblasts cultured in the absence of serum ( Figure 2 ). In prostate and hepatocellular carcinomas, all of the normal tissue samples had the serum-repressed signature and almost all of the tumors had the serum-induced signature, albeit with varying amplitude. In breast, lung, and gastric carcinomas, the common fibroblast serum response signature was clearly evident in some of the tumors and apparently absent in others, suggesting that a “wound-healing phenotype” was a variable feature of these cancers. We therefore classified breast, lung, and gastric cancer samples based on the pattern of expression of the genes that comprise the fibroblast CSR. Link between the Gene Expression Signature of Fibroblast Serum Response and Cancer Progression To investigate the stability and consistency of the serum response signature in individual tumors and to explore its clinical implications, we examined CSR gene expression in a group of locally advanced breast cancers with extensive clinical and molecular data ( Perou et al. 2000 ; Geisler et al. 2001 ; Sorlie et al. 2001 ). As shown in Figure 3 A, the expression profiles of the CSR genes were biphasic, allowing a natural separation of these tumors into two classes. Interestingly, in 18 out of 20 paired tumor samples obtained from the same patients before and after excisional biopsy and chemotherapy, the CSR expression phenotypes were consistent between the two samples. Thus, the wound-related expression program appears to be an intrinsic property of each tumor and not easily extinguished. In a set of 51 patients with clinically matched disease and equivalent treatment ( Sorlie et al. 2001 ), primary tumors with the activated CSR signature were significantly more likely to progress to metastasis and death in a 5-y follow-up period ( p = 0.013 and 0.041, respectively) ( Figure 3 B). Using an alternative analytic approach, classifying each sample by the Pearson correlation between tumor and fibroblast expression patterns of the fibroblast CSR genes, also reproduced the identification of two classes of samples with differing clinical outcomes (Worksheet 2 in Dataset S2 ). A gene expression pattern similar to the serum-activated program of fibroblasts is thus a powerful predictor of prognosis. Other significant prognostic factors in these same patients include tumor grade, estrogen receptor status, and tumor subtype based on gene expression profile ( Geisler et al. 2001 ; Sorlie et al. 2001 ). Tumor stage, lymph-node status, and p53 status were not statistically significant predictors of survival in these patients ( p = 0.13, 0.79, 0.05, respectively). A “basal-like” subtype of breast cancer, characterized by molecular similarities of the tumor cells to basal epithelial cells of the normal mammary duct and associated with a particularly unfavorable prognosis ( Sorlie et al. 2001 ), was significantly associated with a gene expression pattern resembling the fibroblast CSR: six of seven basal-like breast cancers had the “serum-activated” gene expression signature ( p = 0.0075, Fisher's exact test). Thus, the presence or absence of the wound-like phenotype may be linked to intrinsic features of the tumor cells. Figure 3 Context, Stability, and Prognostic Value of Fibroblast CSR in Breast Cancer (A) Expression patterns of CSR genes in a group of breast carcinomas and normal breast tissue previously described in Perou et al. (2000 ). Genes and samples were organized by hierarchical clustering. The serum response of each gene is indicated on the right bar (red shows induced; green shows repressed by serum). Note the biphasic pattern of expression that allows each tumor sample to be classified as “activated” or “quiescent” based on the expression of the CSR genes. The previously identified tumor phenotype (color code) and p53 status (solid black box shows mutated; white box shows wild-type) are shown. Pairs of tumor samples from the same patient, obtained before and after surgery and chemotherapy, are connected by black lines under the dendrogram. Two primary tumor–lymph node metastasis pairs from the same patient are connected by purple lines. (B) Kaplan–Meier survival curves for the two classes of tumors. Tumors with serum-activated CSR signature had worse disease-specific survival and relapse-free survival compared to tumors with quiescent CSR signature. Similar results were obtained whether performing classification using all breast tumors in this dataset or just the 58 tumors from the same clinical trial ( Sorlie et al. 2001 ). We considered the possibility that the observed phenomenon may be simply a reflection of the number of fibroblasts in tumor samples. Perhaps tumors that are infiltrative or otherwise worrisome clinically would demand a wide margin of excision that would include more fibroblasts in the resultant samples. However, classification of breast cancers using the top 1% most highly expressed fibroblast genes (which include a number of extracellular matrix genes and have been previous observed as the “stroma signature” [ Perou et al. 2000 ]) showed no relationship between the generic fibroblast signature and clinical outcome ( p = 0.75; Worksheet 1 in Dataset S2 ). Thus, the prognostic value of the fibroblast CSR likely reflects the physiologic state of the tumor microenvironment and not just the number of fibroblasts in tumor stroma. Similarly, although the mitotic index is an established criterion of tumor grade, classification of these tumors based on expression of cell cycle genes (specifically, all S and G2/M phase genes identified by Whitfield et al. [2002]) only had moderate prognostic value ( p = 0.08; Worksheet 1 in Dataset S2 ). This result also suggests that the prognostic value of the fibroblast CSR is unlikely to be accounted for by the incomplete annotation and removal of genes representing cell growth or division. To extend and validate these results, we tested the prognostic power of the fibroblast CSR signature in independent datasets and different kinds of human cancer ( Figure 4 ). Using published DNA microarray data from a study of gene expression patterns in a group of 78 early (tumor smaller than 5 cm, stage I and IIA) breast cancer patients (van 't Veer et al. 2002), we could segregate the patients into two groups based on expression of the fibroblast CSR genes in the biopsy samples. Tumors with the serum-induced signature had a significantly increased risk of metastasis over 5 y ( p = 0.00046) ( Figure 4 A). Multivariate Cox proportional hazard analysis confirmed that the CSR classification is a significant independent predictor ( p = 0.009); the serum-induced gene expression signature was associated with a 3.3-fold relative risk of breast cancer metastasis within 5 y of diagnosis. In the two breast cancer datasets examined, approximately 50% of the CSR genes demonstrated significant differences in expression between the activated and quiescent groups of samples, but permutation and 10-fold balanced leave-one-out analyses revealed that the correct classification can be accomplished using as few as 6% of CSR genes (Worksheets 10–12 in Dataset S2 ). Thus, the expression pattern of the CSR genes provides a robust basis for predicting tumor behavior. Similarly, in analysis of published DNA microarray data from 62 patients with stage I and II lung adenocarcinomas ( Bhattacharjee et al. 2001 ), tumors with the serum-induced signature were associated with significantly higher risk of death compared to tumors with the serum-repressed signature ( p = 0.021) ( Figure 4 B). These results suggest that presence or absence of a wound-like phenotype in these cancers, with its prognostic implication for their metastatic potential, may be determined at an early stage in their development. In a second, independent group of lung adenocarcinomas of all stages ( Garber et al. 2001 ), tumors with the fibroblast serum-induced signature were associated with a significantly worse prognosis ( p = 0.0014) ( Figure 4 C). A significant correlation between advanced stage and the serum-induced signature was also apparent in this dataset. Finally, in 42 patients with stage III gastric carcinomas, all treated with gastrectomy alone ( Leung et al. 2002 ), tumors with the activated CSR signature were again associated with shorter survival ( p = 0.02) ( Figure 4 D). These results suggest that a wound-healing phenotype, reflected in the expression of a set of serum-inducible genes in fibroblasts, is strongly linked to progression of diverse human carcinomas and can provide valuable prognostic information even at an early stage in the natural history of a cancer. Figure 4 Prognostic Value of Fibroblast CSR in Epithelial Tumors Kaplan–Meier survival curves of tumors stratified into two classes using the fibroblast CSR are shown for stage I and IIA breast cancer (van 't Veer et al. 2002) (A), stage I and II lung adenocarcinoma ( Bhattacharjee et al. 2001 ) (B), lung adenocarcinoma of all stages ( Garber et al. 2001 ) (C), and stage III gastric carcinoma ( Leung et al. 2002 ) (D). For many other cancers, simple stratification based on expression of genes in the fibroblast CSR gene set is unlikely to be predictive of outcome. The dramatic differences in cellular composition and architecture among the tissues in which cancers can arise may influence the role that a wound-healing response can play in their progression. For example, lymphoma cells proliferate in the specialized microenvironment of lymph nodes and bone marrow, and the “stromal” cells in the central nervous system, predominantly astrocytes and microglia, are markedly different from those associated with most epithelial tissues. Indeed, in our initial analysis, the pattern of expression of the fibroblast CSR genes failed to stratify the outcomes in diffuse large B-cell lymphoma ( Rosenwald et al. 2002 ), medulloblastoma ( Pomeroy et al. 2002 ), and glioblastoma multiforme (M. Diehn and P. O. Brown, unpublished data). Histological Architecture of CSR Gene Expression in Tumors Both to validate the DNA microarray results and to investigate the histological architecture of CSR gene expression in tumors, we examined the expression patterns of five CSR genes implicated in extracellular matrix remodeling and cell–cell interaction, using tissue microarrays containing hundreds of breast carcinoma tissues. PLAUR , also known as urokinase-type plasminogen activator receptor, is a well-characterized receptor for matrix-degrading proteases that has been implicated in tumor cell invasion ( Blasi and Carmeliet 2002 ; Sidenius and Blasi 2003 ). LOXL2 is a member of a family of extracellular lysyl oxidases that modify and cross-link collagen and elastin fibers ( Akiri et al. 2003 ). PLOD2 is a member of the lysyl hydroxylase family that plays important roles in matrix cross-linking and fibrosis ( Van Der Slot et al. 2003 ). SDFR1 , previously named gp55 and gp65, encodes a cell surface protein of the immunglobulin superfamily that regulates cell adhesion and process outgrowth ( Clarke and Moss 1994 ; Wilson et al. 1996 ). ESDN is a neuropilin-like cell surface receptor that was also previously found to be upregulated in metastatic lung cancers ( Koshikawa et al. 2002 ). All five of these genes were included in the fibroblast CSR gene set by virtue of their induction by serum in fibroblasts (see Figure 1 ). Anti-PLAUR antibody is commercially available and served as a positive control. We prepared specific riboprobes for LOXL2 and SDFR1 and generated affinity-purified anti-peptide antibodies to PLOD2 and ESDN to detect the predicted protein products. As shown in Figure 5 , PLAUR, LOXL2, PLOD2, and ESDN were not detectably expressed in normal breast tissue; SDFR1 was expressed at a low level in normal breast epithelial cells ( n = 11). In contrast, all five genes were induced in a significant fraction of invasive ductal carcinomas of the breast. As previously reported ( Costantini et al. 1996 ), PLAUR protein is expressed in both tumor cells and peritumoral stroma (70 out of 96, 73% positive) ( Figure 5 ). PLOD2 protein and SDFR1 mRNA were detected in breast carcinoma cells and in a small but consistent fraction of peritumor stroma cells (78 out of 100, 78% positive, and 55 out of 79, 70% positive, respectively). ESDN protein was detected exclusively in breast carcinoma cells (69 out of 112, 62% positive). In contrast, LOXL2 mRNA was abundant in peritumoral fibroblasts around invasive carcinomas (45 out of 106, 42% positive). LOXL2 protein has been previously reported to be expressed in normal mammary ducts and increased in invasive breast carcinoma cells ( Akiri et al. 2003 ). Our data suggest that LOXL2 is primarily synthesized by peritumoral fibroblasts, but may act on or in the vicinity of epithelial cells during tissue remodeling. Collectively, these results suggest that the pathophysiology represented by expression of the fibroblast CSR genes in cancers represents a multicellular program in which the tumor cells themselves, tumor-associated fibroblasts, and perhaps diverse other cells in the tumor microenvironment are active participants. Figure 5 Histological Architecture of CSR Gene Expression in Breast Cancer Representative ISH of LOXL2 and SDFR1 and IHC of PLOD2, PLAUR, and ESDN are shown (magnification, 200×). Panels for LOXL2 , PLAUR, PLOD2, and ESDN represent cores of normal and invasive ductal breast carcinoma from different patients on the same tissue microarray. Panels for SDFR1 demonstrate staining in adjacent normal and carcinoma cells on the same tissue section. Arrows highlight spindle-shaped stromal cells that stain positive for SDFR1 and PLOD2. No signal was detected for the sense probe for ISH or for control IHC without the primary antibody. Discussion The remarkable ability of a single physiological fluid—serum—to promote the growth and survival of diverse normal and cancer cells in culture suggests that there may be a conserved, programmed response to the molecular signals that serum provides. In vivo, serum as a physiological signal has a very specific meaning: cells encounter serum—the soluble fraction of coagulated blood—only in the context of a local injury. In virtually any tissue, a rapid, concerted multicellular response, with distinct physiological exigencies that evolve over minutes, hours, and days, is required to preserve the integrity of the tissue and often the survival of the organism. In response to a wound, many of the normal differentiated characteristics of the cells in the wounded tissue are temporarily set aside in favor of an emergency response. In wound repair, as in cancer, cells that ordinarily divide infrequently are induced to proliferate rapidly, extracellular matrix and connective tissues are invaded and remodeled, epithelial cells and stromal cells migrate, and new blood vessels are recruited. In all these respects, a wound response—and the characteristic physiological response to serum—would appear to provide a highly favorable milieu for cancer progression. We defined a stereotyped genomic expression response of fibroblasts to serum, which reflects many features of the physiology of wound healing. When we examined the expression of these genes in human tumors, we found strong evidence that a wound-like phenotype was variably present in many common human cancers (including many that are not known to be preceded by chronic wounds) and was a remarkably powerful predictor of metastasis and death in several different carcinomas. The proposed link between the fibroblast serum response signature and cancer progression raises many questions for additional studies. Perhaps most importantly, our results do not allow us to distinguish whether the wound-like phenotype has a functionally important role in tumor progression or merely serves as a marker for the underlying propensity of a cancer to progress and metastasize. However, at least three genes induced in the fibroblast serum response, PLAUR , LOXL2 , and MIF , have been previously shown to increase cancer invasiveness or angiogenesis in animal xenograft models; each of these three genes has also been shown to play an important role in wound healing ( Akiri et al. 2003 ; Nishihira et al. 2003 ; Sidenius and Blasi 2003 ). Thus, we are inclined to believe that coordinate induction of a wound-healing program in carcinomas contributes to tumor invasion and metastasis. Several potential mechanisms might contribute to the wound-like gene expression pattern in cancers. In some cancers, ongoing local tissue injury, resulting from growth and dysfunctional behavior of the tumor cells, could continuously trigger a normal wound-healing response. The classic observation of deposited fibrin products in human tumors is consistent with this model ( Dvorak 1986 ). Inflammatory cells, presumably recruited by tissue disorder, may amplify the wound response and contribute to tumor invasion in part by expression of metalloproteinases ( Coussens et al. 2000 ; Daniel et al. 2003 ). The wound response might also be initiated directly by signals from the tumor cells ( Fukumura et al. 1998 ), whose ability to activate an inappropriate wound-healing response—favorable to cell proliferation, invasion, and angiogenesis—might be strongly selected during cancer progression. The possibility that stromal cells might play a primary role in promoting a wound-like phenotype in some cancers is raised by studies showing that tumor-associated fibroblasts can enhance tumor engraftment and metastasis in animal models ( Elenbaas and Weinberg 2001 ) and the demonstration in some cancers of genotypic abnormalities in tumor-associated fibroblasts ( Kurose et al. 2002 ). Heterotopic interaction experiments, genetic models, and cell-culture models should enable these and other possible mechanisms to be investigated. Our results illustrate the power of using gene expression data from specific cells or physiological and genetic manipulations to build an interpretive framework for the complex gene expression profiles of clinical samples ( Lamb et al. 2003 ). Several prognostic models based on gene expression patterns have previously been identified from systematic DNA microarray profiles of gene expression in human cancers. Some of these prognostic gene expression profiles appear to reflect the developmental lineage of the cancer cells ( Alizadeh et al. 2000 ; Sorlie et al. 2001 ; Pomeroy et al. 2002 ), some appear to reflect the activity of specific molecular determinants of tumor behavior (e.g., the activity of PLA2G2A in gastric cancer [ Leung et al. 2002 ]), while still others represent the mechanistically agnostic results of machine-assisted learning ( van 't Veer et al. 2002 ; Ramaswamy et al. 2003 ). Although they serve to identify many of the same tumors with unfavorable prognosis, the genes that define the fibroblast CSR overlap minimally with the genes previously used to predict outcome in the same cancers. For example, the fibroblast CSR involves only 20 out of 456 genes in an “intrinsic gene list” that can serve to segregate breast cancers into prognostically distinct groups ( Perou et al. 2000 ) and four out of 128 genes that define the general metastasis signature reported by Ramaswamy et al. (2003 ). Only 11 genes are in common between the 231 gene van't Veer poor prognosis signature for breast cancer ( van 't Veer et al. 2002 ) and the fibroblast CSR genes. The prognostic power of these different sets of genes illustrates the multidimensional variation in the gene expression programs in cancers and the complex interplay of many distinct genetic and physiological factors in determining the distinctive biology of each individual tumor. Our success in discovering a significant new determinant of cancer progression, using previously published and publicly available data, illustrates the richness of the data as a continuing source for future discoveries and the importance of unrestricted access to published research data ( Roberts et al. 2001 ). The signals and regulatory systems that normally initiate, sustain, and eventually shut down the physiological response to a wound remain to be identified and understood. Identification of the molecular control mechanisms in this pathway may pave the way to new cancer therapies or chemopreventative agents. For example, cyclooxygenase 2 is strongly induced in the response of fibroblasts to serum ( Iyer et al. 1999 ), and platelet-derived growth factor is one of the principal molecular signals and mitogenic factors in serum. Platelet-derived growth factor receptor and cyclooxygenase 2 are inhibited by imatinib mesylate and nonsteroidal anti-inflammatory agents, respectively—two drugs with established efficacy in treating or preventing cancer ( Bergers et al. 2003 ; Huls et al. 2003 ). Whether these or other small molecules might derive significant activity against cancer from their ability to inhibit a dysregulated wound-healing response will be an important question for future investigation. Materials and Methods Cells and tissue culture Human primary fibroblasts from ten anatomic sites were cultured in 0.1% versus 10% FBS, as previously described ( Chang et al. 2002 ). For the serum induction timecourse, foreskin fibroblasts CRL 2091 (American Type Culture Collection [ATCC], Manassas, Virginia, United States) were serum-starved for 48 h and harvested at the indicated timepoints after switching to media with 10% FBS, essentially as described in Iyer et al. (1999 ). Microarray procedures Construction of human cDNA microarrays containing approximately 43,000 elements, representing approximately 36,000 different genes, and array hybridizations were as previously described ( Perou et al. 2000 ). mRNA was purified using FastTrack according to the manufacturer's instructions (Invitrogen, Carlsbad, California, United States). For the serum timecourse, RNA from all of the sampled timepoints were pooled as reference RNA to compare with RNA from individual timepoints as described in Iyer et al. (1999 ). Data analysis For defining a common serum response program in fibroblasts, global gene expression patterns in 50 fibroblast cultures derived from ten anatomic sites, cultured in the presence of 10% or 0.1% FBS, were characterized by DNA microarray hybridization ( Chang et al. 2002 ). We selected for further analysis genes for which the corresponding array elements had fluorescent hybridization signals at least 1.5-fold greater than the local background fluorescence in the reference channel, and we further restricted our analyses to genes for which technically adequate data were obtained in at least 80% of experiments. These filtered genes were then analyzed by the multiclass Significance Analysis of Microarrays (SAM) algorithm ( Tusher et al. 2001 ) to select a set of genes whose expression levels had a significant correlation with the presence of serum in the medium, with a false discovery rate (FDR) of less than 0.02%. The corresponding expression patterns were organized by hierarchical clustering ( Eisen et al. 1998 ). Genes that were coordinately induced or repressed in response to serum in most samples (Pearson correlation, greater than 90%) were identified. This set of 677 genes, represented by 772 cDNA probes, of which 611 are uniquely identified by UniGene ( http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?db=unigene ), was termed the common fibroblast serum response gene set. To identify the subset of these 677 genes whose variation in expression was directly related to cell cycle progression, we compared this set of genes to a published set of genes periodically expressed during the HeLa cell cycle ( Whitfield et al. 2002 ). Because both datasets were generated using similar cDNA microarrays, we tracked genes by the IMAGE number of the cDNA clones on the microarrays. The majority of the genes in the fibroblast serum response gene set showed no evidence of periodic expression during the HeLa cell cycle. One hundred sixty-five genes, represented by 199 cDNA clones, overlapped with the cell cycle gene list; the remaining 512 genes, represented by 573 clones, of which 459 are uniquely identified in UniGene, was termed the CSR gene set. The patterns of expression in human tumors of the 512 genes of the fibroblast CSR gene set were analyzed using data from published tumor expression profiles. Detailed methods and primary datasets are available as Datasets S1 and S2 and on our Web site ( http://microarray-pubs.stanford.edu/wound ). We used the Unigene unique identifier (build 158, release date January18, 2003) to match genes represented in different microarray platforms. For cDNA microarrays, genes with fluorescent hybridization signals at least 1.5-fold greater than the local background fluorescent signal in the reference channel (Cy3) were considered adequately measured and were selected for further analyses. For Affymetrix data, signal intensity values were first transformed into ratios, using for each gene the mean values of the normalized fluorescence signals across all the samples analyzed as the denominators ( Bhattacharjee et al. 2001 ). The genes for which technically adequate measurements were obtained from at least 80% of the samples in a given dataset were centered by mean value within each dataset, and average linkage clustering was carried out using the Cluster software ( Eisen et al. 1998 ). In each set of patient samples, the samples were segregated into two classes based on the first bifurcation in the hierarchical clustering dendrogram. For the datasets shown, the clustering and reciprocal expression of serum-induced and serum-repressed genes in the tumor expression data allowed two classes to be unambiguously assigned. Samples with generally high levels of expression of the serum-induced genes and low levels of expression of the serum-repressed genes were classified as “activated”; conversely, samples with generally high levels of expression of serum-repressed genes and low levels of expression of the serum-induced genes were classified as “quiescent.” Survival analysis by a Cox–Mantel test was performed in the program Winstat (R. Fitch Software). In situ hybridization and immunohistochemistry Digoxigenin-labeled sense and antisense riboprobes for LOXL2 and SDFR1 were synthesized using T7 polymerase-directed in vitro transcription ( Iacobuzio-Donahue et al. 2002 ). Sense and antisense riboprobes for SDFR1 were made from nucleotides 51–478 of IMAGE clone 586731 (ATCC #745139), corresponding to the last 388 nucleotides of the 3′ end of the coding sequence and 39 nucleotides of the 3′ untranslated region. Sense and antisense riboprobes for LOXL2 were made from nucleotides 41–441 of IMAGE clone 882506 (ATCC #1139012), corresponding to the 3′ end of the coding sequence. In situ hybridization (ISH) results were considered to have appropriate specificity when we observed a strong, consistent pattern of hybridization of the antisense probe and little or no hybridization of the corresponding sense probe. Immunohistochemical (IHC) staining was performed using Dako (Glostrup, Denmark) Envision Plus following the manufacturer's instructions. Anti-PLAUR antibody against whole purified human uPA–receptor protein (AB8903; Chemicon, Temecula, California, United States) was used at 1:200 dilution. Affinity-purified polyclonal antibody to PLOD2 was produced by immunizing rabbits with peptides EFDTVDLSAVDVHPN, coupled to keyhole limpet hemocyanin (KLH) (Applied Genomics, Inc., Sunnyvale, California, United States); affinity-purified antiserum was used for IHC at 1:25,000 dilution. Similarly, affinity-purified polyclonal antibody to ESDN was produced by immunizing rabbits with peptide DHTGQENSWKPKKARLKK coupled to KLH (Applied Genomics, Inc.) and used for IHC at 1:12,500 dilution. High-density tissue microarrays containing tumor samples were constructed as described in Kononen et al. (1998 ). ISH ( Iacobuzio-Donahue et al. 2002 ) and IHC ( Perou et al. 2000 ) were as reported. ISH and IHC images and data were archived as described in Liu et al. (2002 ). Supporting Information Figure 1 A can be interactively explored at http://microarray-pubs.stanford.edu/wound/ . Raw datasets and all supporting data are also available at http://microarray-pubs.stanford.edu/wound/ . Dataset S1 Detailed Bioinformatic Methods Provides a description of microarray datasets, cross-platform mapping and data normalization, classification of cancers by fibroblast CSR genes and correlated clinical outcomes (Worksheets 1–8 in Dataset S2 ), the top 1% fibroblast genes in breast cancer prognosis (see Worksheet 1 in Dataset S2 ), cell cycle S and G2/M genes in breast cancer prognosis (see Worksheet 1 in Dataset S2 ), analysis of GO annotations of fibroblast serum response genes (see Worksheet 9 in Dataset S2 ), and the minimum number of CSR genes necessary for tumor classification (see Worksheets 10–12 in Dataset S2 ). (120 KB DOC). Click here for additional data file. Dataset S2 Supporting Data Excel Worksheets of clinical and microarray data, as described in Dataset S1 . (736 KB XLS). Click here for additional data file. Accession Numbers The Locus Link ( http://www.ncbi.nlm.nih.gov/LocusLink/ ) accession numbers for the genes discussed in this paper are CORO1C (Locus Link ID 23603), E2F1 (Locus Link ID 1869), ESDN (Locus Link ID 131566), FLNC (Locus Link ID 2318), FOXM1 (Locus Link ID 2305), LOXL2 (Locus Link ID 4017), MIF (Locus Link ID 4282), MYL6 (Locus Link ID 4637), PLAUR (Locus Link ID 5329), PLOD2 (Locus Link ID 5352), PTTG1 (Locus Link ID 9232), SDFR1 (Locus Link ID 27020), TAGLN (Locus Link ID 6876), and TPM2 (Locus Link ID 7169). The accession numbers of the Gene Ontology (GO) ( http://www.geneontology.org/ ) terms that appear in Dataset S1 are angiogensis (GO:0001525), blood coagulation (GO:0007596), complement activation (GO:0006956), immune response (GO:0006955), N-linked glycosylation (GO:0006487), protein translation (GO:0006445), and proteolysis and peptidolysis (GO:0006508).	/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC314300.xml
514615	Primary small-cell neuroendocrine carcinoma of the duodenum – a case report and review of literature	Background Small-cell neuroendocrine carcinoma in the duodenum is an extremely rare neoplasm with poor prognosis. Case presentation A 57-year-old man presented with sudden onset gastrointestinal bleeding and fainting attacks. Duodenoscopy and hypotonic duodenography revealed a 3 × 3 cm protruding tumor with ulcerations situated opposite the ampulla of Vater in the second part of the duodenum. Local excision of the tumor was performed, followed by adjuvant chemotherapy with 5-fluoro uracil and leucovorin. Examination of the tumor by immunohistochemistry and electron microscopy indicated it to be neuroendocrine in nature, expressing synaptophysin and AE1/AE3, and containing dense core granules. The patient showed no sign of recurrence and has been disease-free for more than 48 months after surgery. Conclusions Most cases of small-cell neuroendocrine carcinoma in the duodenum show rapid progression of the disease, and even radical surgery with or without chemotherapy do not prevent death. We report a rare subtype of small-cell neuroendocrine carcinoma. This subtype appears to have a much better prognosis, and may be amenable to local excision, if the lesion is away from the ampulla of Vater.	Background Duodenal Neuroendocrine tumors constitute 5% of all gastrointestinal neuroendocrine tumors [ 1 , 2 ]. Most of these show well-differentiated features and are classified as carcinoids or somatostatinomas [ 3 - 6 ]. Occurrence of carcinoma is rare, and carcinomas with anaplastic character, which are classified as small-cell carcinomas, are even less frequent [ 7 - 12 ]. The most common small-cell neuroendocrine carcinoma (NEC) is the small-cell undifferentiated carcinoma of the lung [ 13 , 14 ]. Although the features of these pulmonary tumors are well defined, the characteristics of their extrapulmonary counterparts are still unknown. We report a case of small-cell NEC in the duodenum that had unique morphological features and exceptionally good clinical outcome. Case presentation A 57-year-old man presented with sudden gastrointestinal tract bleeding and episode of fainting. Duodenoscopy (Figure 1a ) and hypotonic duodenography (Figure 1b ) revealed a 3 × 3 cm protruding tumor with two ulcerations located opposite the ampulla of Vater in the second part of the duodenum. Laboratory data showed no abnormalities in blood chemistry, tumor markers (CEA, CA19-9, NSE, proGRP) and endocrine markers (somatostatin, gastrin, glucagons, serotonin, VIP) except a moderate anemia (9.5 g/dl hemoglobin). No abnormal findings were observed in the chest X-ray and computed tomography (CT). Figure 1 (a) Duodenoscopy showing a 3 × 3 cm protruding tumor with two ulcerations located opposite the ampulla of Vater in the second portion of the duodenum. (b) Hypotonic duodenography showing the donuts-shape tumor in the duodenum. A laparotomy was performed. As there was no serosal invasion or regional lymphadenopathy wide local excision of the tumor was performed. On gross examination, the tumor showed two ulcerations and two different morphological components (Figure 2a and 2b ). One component (component A) was round in shape with a round ulceration on the top, and the other component (component B), which enclosed the round component, was crescent in shape with a spindle-shaped ulceration on the top. The two components showed different histopathological and immunohistochemical features (Table 1 ). The round component contained fibrous tissue, small nuclei, and clear nucleoli. Histopathologically, the crescent component had more anaplastic features typical of small-cell carcinoma, such as sheets of tightly packed anaplastic cells with round nuclei and scanty cytoplasm (Figure 2c , 2d ). Neuroendocrine differentiation was investigated using immunohistochemical and ultrastructural techniques. Both components showed neuroendocrine features, with immunochemistry identifying synaptophysin and AE1/AE3 (Figure 3a and 3b ), and electron microscopy identifying dense core granules (Figure 4 ). Immunochemistry also showed that the crescent component expressed less cytokeratin, vimentin and CD56, and more MIB-1 than the round component. Figure 2 Macroscopic and microscopic findings of the tumor. (a) Gross appearance of the tumor. The tumor was divided into two components, component A (round shape) and B (crescent shape). (b) Photomicrograph of the gross appearance of the tumor (Hematoxylin and eosin X 2). (c) Photomicrograph of the component A showing fibrous tissue, small nuclei, and clear nucleoli. (Hematoxylin and Eosin X 40). (d) Photomicrograph of the component B showing more anaplastic features typical of small-cell carcinoma, such as sheets of tightly packed anaplastic cells with round nuclei and scanty cytoplasm. (Hematoxylin and Eosin X 40). Table 1 Immunochemical characteristics of the two components of the tumor. Synaptophysin AE1/AE3 Vimentin CD56 chromogranin A MIB1 (A) Round component ++ ++ ++ + - 25% (B) Crescent component + - - - - 50% LCA, L26, UCHL1, CD3, ASMA, M-actin, desmin, CD34, NF, GFAP, and S100 were negative in both components. Figure 3 Immunostaining for AE1/AE3 showing (a) diffuse cytoplasmic positivity in the component A, and (b) no reactivity in the component B. Figure 4 Ultrastructural study showed cytoplasmic dense-core granules in the component A. The patient was discharged three weeks after operation with uneventful postoperative period. Four cycles of monthly adjuvant chemotherapy with 5-fluoro uracil (5-FU) (325 mg/m 2 ) and leucovorin (20 mg/m 2 ) were administered. The patient showed no sign of recurrence and is disease-free 48 months after surgery. Discussion Neuroendocrine carcinomas (NEC) in the duodenum are extremely rare, and are classified as either 'small-cell' or 'non small-cell' types. The small-cell NEC occurring in the duodenum and elsewhere in gastrointestinal tract are similar to the small-cell carcinoma of the lung [ 7 , 8 ]. Only eight cases of small-cell NEC in the duodenum have been reported, with the present case being the ninth (Table 2 ) [ 7 - 12 ]. Most cases occurred in middle-aged or geriatric males with lesions in the ampulla of Vater. Extra-ampullary small-cell NECs in the duodenum are extremely rare, with only two cases reported previously [ 7 , 8 ]. Table 2 Profiles of the cases of primary small-cell neuroendocrine carcinoma in the duodenum reported in the literature. Author Year Age/sex Clinical Manifestations Size (mm) Morphology Location Metastasis Surgery Chemotherapy Prognosis (month) 1 Swanson 1986 [7] 76 M Abdominal pain, anorexia, weight loss 15 ulceration adjacent to the ampulla* LN, liver biospy 5-FU, doxorubicin, mitomycin Dead (1.5) 2 Zamboni 1990 [8] 62 M jaundice, weight loss 25 polypoid the papilla of Vater LN PD - Dead (7) 3 Zamboni 1990 [8] 66 M jaundice, abdominal pain 20 ulceration the papilla of Vater LN PD - Dead (6) 4 Zamboni 1990 [8] 51 M jaundice, weight loss, abdominal pain 30 soft fungating mass the papilla of Vater LN PD - Dead (17) 5 Lee 1992 [9] 86 M jaundice, recurrent pancreatitis ? polypoid Peri ampullary ? - - (>5) 6 Sarker 1992 [10] 53 F jaundice, weight loss, back pain 35 mass with small ulceration the papilla of Vater LN PD 5-FU, TNF, interferon Recurrence + (>18) 7 Sato 1995 [11] 74 M jaundice 35 polypoid the papilla of Vater ? PPPD - ? 8 Shim 2000 [12] 54 M jaundice 30 ulceration the papilla of Vater liver PD cisplatin, etoposide, radiation Dead (8) 9 Sata 2004 present case 57 M GI Tract bleeding 30 mass with ulceration Peri ampullary - Local resection 5-FU leucovorin disease free (>48) * Hormone VIP; ? Don't know; M – male; F – Female; PD-pancreaticoduodenectomy; LN-lymph node; PPPD-pylorus preserving pancreaticoduodenectomy; 5 FU-f fluoro uracil; TNF – tumor necrosis factor The natural course of small-cell NECs in the duodenum is still not clear. Most cases reported in the literature show rapid progress of the disease, with radical surgery and/or chemotherapy not altering the clinical course, and thus a poor prognosis. In six of the previous eight cases, patients underwent pancreaticoduodenectomy for removal of the tumor, while the remaining two did not undergo surgery because of multiple liver metastasis or poor general condition. In spite of radical resection with or without adjuvant chemotherapy, most cases showed rapid recurrence and metastasis. Of the eight reported cases, one was unusual as it occurred in a middle-aged female with rapid progress of the disease but effective response to adjuvant chemotherapy using 5FU, tumor necrosis factor, and interferon this patient survived for more than eighteen months [ 10 ]. The present case was treated by a local excision of the tumor followed by adjuvant chemotherapy using 5-FU and leucovorin, and showed a distinctively unique clinical course, with the patient surviving for more than 48 months without any sign of recurrence. This case was presented with gastrointestinal bleeding, which contributed to early diagnosis, whereas the other previous cases in literature presented either with abdominal pain or jaundice. Hence, the good prognosis in present case could also be due to its earlier presentation. The lesion in the present case showed a different immunohistological character from that in other cases, such as no immunoreactivity to neuron-specific enolase (NSE) or chromogranin A. These differences too might partly explain the different character of this case. Conclusions This report identifies a new subtype of small-cell NEC in the duodenum. This subtype appears to have a much better prognosis, and may be amenable to local excision, if the lesion is away from the ampulla of Vater. Competing interests None declared. Authors' contributions NS, MT, MK, KY, KK, HN took part in the operation, performed the literature search and drafted the manuscript for submission. HN supervised the preparation of the manuscript and edited the final version for publication. TS, KS performed pathological investigations and contributed to the pathological content of the manuscript. All authors read and approved the manuscript.	/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC514615.xml
314466	HIV-1 Nef Binds the DOCK2–ELMO1 Complex to Activate Rac and Inhibit Lymphocyte Chemotaxis	The infectious cycle of primate lentiviruses is intimately linked to interactions between cells of the immune system. Nef, a potent virulence factor, alters cellular environments to increase lentiviral replication in the host, yet the mechanisms underlying these effects have remained elusive. Since Nef likely functions as an adaptor protein, we exploited a proteomic approach to directly identify molecules that Nef targets to subvert the signaling machinery in T cells. We purified to near homogeneity a major Nef-associated protein complex from T cells and identified by mass spectroscopy its subunits as DOCK2–ELMO1, a key activator of Rac in antigen- and chemokine-initiated signaling pathways, and Rac. We show that Nef activates Rac in T cell lines and in primary T cells following infection with HIV-1 in the absence of antigenic stimuli. Nef activates Rac by binding the DOCK2–ELMO1 complex, and this interaction is linked to the abilities of Nef to inhibit chemotaxis and promote T cell activation. Our data indicate that Nef targets a critical switch that regulates Rac GTPases downstream of chemokine- and antigen-initiated signaling pathways. This interaction enables Nef to influence multiple aspects of T cell function and thus provides an important mechanism by which Nef impacts pathogenesis by primate lentiviruses.	Introduction Primate lentiviruses persist in the host by active replication and can reemerge from latent reservoirs that are established in cells of the immune system ( Finzi and Siliciano 1998 ; Douek et al. 2003 ). The infectious cycle is intimately linked to interactions between circulating T cells and antigen-presenting cells ( Stevenson et al. 1990 ; Bukrinsky et al. 1991 ; Embretson et al. 1993 ; Swingler et al. 1999 ; Geijtenbeek et al. 2000 ). These interactions involve T cell migration, adhesion, and antigen-initiated signaling, processes that are dependent on cytoskeletal dynamics regulated by the Rho subfamily of small GTPases ( Hall 1998 ; Schmitz et al. 2000 ). The lentiviral accessory protein Nef is a multifunctional regulator that is important for rapid progression to AIDS ( Kestler et al. 1991 ; Piguet et al. 1999 ; Renkema and Saksela 2000 ). One key function of Nef is its ability to facilitate activation of infected cells and thus provide an environment that is conducive for viral replication ( Skowronski et al. 1993 ; Baur et al. 1994 ; Du et al. 1996 ; Schrager and Marsh 1999 ; Simmons et al. 2001 ). Another important function of Nef is its ability to promote evasion of the antiviral immune response. This is accomplished by downregulation of class I MHC complexes from the surface of infected cells, which protects against detection by cytoxic T cells specific for viral antigens ( Schwartz et al. 1996 ; Collins et al. 1998 ). The ability of Nef to facilitate T cell activation is well documented. Thymic and peripheral CD4 + T cells from transgenic mice are hypersensitive to stimulation via the T cell antigen receptor (TCR) ( Skowronski et al. 1993 ; Hanna et al. 1998 ), as are resting primary human CD4 + T cells ( Schrager and Marsh 1999 ; Wang et al. 2000 ) and cell lines transduced to express HIV-1 Nef ( Alexander et al. 1994 ; Baur et al. 1994 ). Nef was reported to associate with molecules that play important roles in antigen-initiated signaling in T cells, including elements of signaling pathways involving small GTPases. Specifically, Nef was reported to associate with Vav ( Fackler et al. 1999 ) and activate p21-activated serine–threonine kinases (PAKs), possibly though the activation of Rac or CDC42 ( Lu et al. 1996 ). Recent observations that Nef can activate Rac in a glial cell line have strengthened the connection between Nef and these pathways ( Vilhardt et al. 2002 ). The notion that effects of Nef on signaling machineries in T cells are mediated by small GTPases, their effectors, or both represents an attractive possibility, yet the exact mechanism resulting in activation of these pathways has remained elusive. Since Nef likely functions as an adaptor protein, we exploited a proteomic approach to directly identify the key molecules Nef uses to subvert the signaling machinery in T cells. Here we show that Nef targets a key activator of Rac GTPases that functions downstream of the TCR and chemokine receptors. Results Nef Binds DOCK2, ELMO1, and Rac in T Cells To identify downstream effectors of Nef in T lymphocytes, we generated CD4 + Jurkat T cells that stably express the extensively studied patient-derived HIV-1 Nef protein NA7 ( Mariani and Skowronski 1993 ) tagged at its C-terminus with HA and FLAG epitopes (NA7-hf) ( Figure 1 A). Nef and its associated proteins were purified by successive immunoprecipitations with anti-HA- and then anti-FLAG epitope antibodies, followed each time by elution with the respective peptide epitope and resolved by SDS-PAGE. Several polypeptides with apparent molecular weights ranging from approximately 20 kDa to 220 kDa copurified with HIV-1 Nef, but were absent in preparations from control cells that do not express Nef ( Figure 1 B). Gel slices containing these polypeptides were digested with trypsin, and the resulting peptides were sequenced by liquid chromatography tandem mass spectroscopy (LC/MS/MS) and subjected to database searches ( Hu et al. 2002 ). Two abundant Nef-associated proteins, DOCK2 and ELMO1, were thus identified. DOCK2 is a lymphocyte-specific CED5/DOCK180/Myoblast City (CDM) family protein that regulates the activity of Rac1 and Rac2 GTPases downstream of chemokine receptors and the TCR and is essential for lymphocyte migration and normal antigen-specific responses of T cells ( Fukui et al. 2001 ; Reif and Cyster 2002 ; Sanui et al. 2003a ). Rac GTPases are members of the Rho subfamily of small GTP-binding proteins that control several processes, including cytoskeletal rearrangements during cell motility and T cell activation ( Hall 1998 ). Recent studies showed that ELMO1 functionally cooperates with CDM family proteins to activate Rac ( Brugnera et al. 2002 ; Sanui et al. 2003b ). Significantly, our mass spectroscopic analyses of Nef-associated proteins also identified Rac2. Furthermore, in addition to Rac2-specific peptides, we also detected peptides shared by Rac1 and Rac2, raising the possibility that Rac1 also associates with HIV-1 Nef in T cells. The ubiquitously expressed Rac1 and hematopoietic cell-specific Rac2 are 95% identical, and both isoforms regulate cytoskeletal dynamics and gene expression in T lymphocytes ( Yu et al. 2001 ; Croker et al. 2002 ). Figure 1 DOCK2, ELMO1, and Rac Are Abundant Nef-Associated Proteins in T Cells (A) Schematic representation of epitope-tagged HIV-1 Nef (NA7-hf). The structured regions of Nef are boxed and the disordered regions, as determined by X-ray crystallography and NMR studies, are shown by a thin line. The locations of the N-terminal myristoyl moiety, prolines P72 and P75 in the PP-II helix, arginine R106, leucines L164 and L165 (LL164), and the C-terminal HA-FLAG epitopes are indicated. (B) DOCK2, ELMO1, and Rac2 specifically copurify with HIV-1 Nef from Jurkat T cells. Jurkat T cells (1.8 ×10 10 ) stably expressing NA7-hf (lane 3) or control Jurkat cells (lane 2) were subjected to the two-step immunopurification procedure described in the text (see Materials and Methods ). Polypeptides present in purified immune complexes were resolved by SDS-PAGE and analyzed by LC/MS/MS. We identified 58 DOCK2-specific peptides covering 869 out of 1830 total amino acid residues (47.5% coverage, expectation value 6.0 × 10 –130 ), 10 ELMO1-specific peptides covering 122 out of 727 total amino acid residues (16.8% coverage, expectation value 1.0 × 10 −10 ), and three Rac-specific (two of which were Rac2-specific) peptides covering 26 out of 192 total amino acid residues (13.5% coverage, expectation value 4.6 × 10 −4 ). Bands corresponding to DOCK2, ELMO1, Rac2 and their predicted molecular weights, NA7-hf Nef, and the FLAG peptide used for elution are indicated. Nef Binds the DOCK2–ELMO1–Rac Complex By analogy to previously described interactions among Rac, ELMO1, and CDM family proteins ( Brugnera et al. 2002 ; Sanui et al. 2003b ), our finding that DOCK2, ELMO1, and Rac2 copurified with HIV-1 Nef suggested that DOCK2 forms a ternary complex with ELMO1 and Rac2 and that Nef binds this complex. To investigate these possibilities, we attempted to reconstitute these interactions in human embryonic kidney 293 (HEK 293) cells. Although HEK 293 cells express endogeneous ELMO1, our initial studies revealed that the association of Nef with DOCK2 and Rac2 was significantly enhanced by ectopic expression of ELMO1 (data not shown). Thus, to determine whether ELMO1 and Rac2 copurify with DOCK2, DOCK2-containing complexes were purified from HEK 293 cells transiently expressing His-tagged DOCK2, Myc-tagged ELMO1, and Myc-tagged Rac2 via DOCK2 using Ni–NTA resin and eluted with imidazole. Immunoblotting revealed that ELMO1 and Rac2 copurified with DOCK2 ( Figure 2 A, lane 3), indicating that DOCK2 complexes with ELMO1 (DOCK2–ELMO1) and Rac2. Figure 2 Lentiviral Nef Binds the DOCK2–ELMO1–Rac Complex (A) HIV-1 Nef binds the DOCK2–ELMO1–Rac2 complex. His-DOCK2, Myc-ELMO1, and Myc-Rac2 alone (lanes 1, 3, and 5) or together with NA7-hf Nef (lanes 2, 4, and 6) were transiently expressed in HEK 293 cells as indicated. DOCK2 was precipitated from extracts (lanes 1 and 2) with Ni–NTA resin (lanes 3 and 4). Nef–DOCK2 was then precipitated with anti-FLAG affinity gel (lanes 5 and 6), and the epitope-tagged proteins were detected by immunoblotting and visualized by enhanced chemiluminescence. (B) Rac1 associates with HIV-1 Nef. Nef and associated proteins were isolated from extracts of HEK 293 cells transiently expressing DOCK2, ELMO1, and Rac1 either alone (lanes 1 and 4), with NA7-hf (lanes 2 and 5), or with a Nef variant containing a disrupted myristoylation signal (lanes 3 and 6). Nef and associated proteins were detected in anti-FLAG immunoprecipitates (lanes 1–3) and in extracts (lanes 4–6) by immunoblotting. (C) The interaction with DOCK2, ELMO1, and Rac2 is a conserved function of lentiviral Nef proteins. The ability of selected hf-tagged HIV-1 (lanes 1–3 and 5) and SIV mac239 (lane 4) Nef proteins to bind DOCK2, ELMO1, and Rac2 was determined as described in the legend to (B) above. The protein band in (C) indicated by the asterisk is the heavy chain of anti-FLAG mAb. Subsequently, we asked whether ELMO1 and Rac2 are subunits of DOCK2–Nef complexes. DOCK2–Nef-containing complexes were isolated from HEK 293 cells transiently expressing His-tagged DOCK2, Myc-tagged ELMO1, Myc-tagged Rac2, and HA-FLAG epitope-tagged HIV-1 Nef (NA7-hf) via DOCK2 using Ni–NTA resin and eluted with imidazole ( Figure 2 A, lane 4). DOCK2–Nef complexes were then reisolated from this eluate via Nef by anti-FLAG immunoprecipitation. It is evident that ELMO1 and Rac2 also copurified with complexes containing both Nef and DOCK2 ( Figure 2 A, lane 6), thus supporting the possibility that HIV-1 Nef binds DOCK2–ELMO1 complexes that contain Rac2. Nef Targets Rac1 and Rac 2 Isoforms Our mass spectroscopic analyses indicated that HIV-1 Nef associates with Rac2, but left open the possibility that it also targets Rac1. Therefore, we tested whether Nef can associate with Rac1 in the context of DOCK2 and ELMO1 using the same HEK 293 transient expression assay. Nef and its associated proteins were isolated from cell extracts by anti-FLAG immunoprecipitation and visualized by immunoblotting ( Figure 2 B). Nef formed readily detectable complexes incorporating Rac1 ( Figure 2 B, lane 2), while a mutant Nef protein unable to associate with membranes due to disruption of its N-terminal myristoylation signal (NA7 (G2 ∇ HA) ), and therefore functionally defective, did not associate with DOCK2, ELMO1, or Rac1 ( Figure 2 B, lane 3). These results indicate that myristoylated Nef targets the Rac1 and Rac2 isoforms. Nef proteins from well-characterized primate lentiviruses display considerable amino acid sequence variation. Therefore, we verified that Nef proteins from additional well-characterized laboratory HIV-1 strains (SF2 and NL4–3) bind DOCK2, ELMO1, and Rac2. We also tested a Nef protein from a strain of pathogenic SIV, mac239, that is important for rapid progression to AIDS in experimentally infected rhesus macaques ( Kestler et al. 1991 ). Nef and its associated proteins were isolated from HEK 293 cell extracts by anti-FLAG immunoprecipitation and visualized by immunoblotting ( Figure 2 C). Functional Nef proteins from all lentiviral strains tested associated with DOCK2, ELMO1, and Rac2. This indicates that the ability to associate with Rac and its upstream regulators is a conserved function of primate lentiviral Nef. Nef Activates Rac in Resting T Cells Since DOCK2, ELMO1, and Rac are major Nef-associated proteins in Jurkat T cells and since DOCK2 mediates Rac activation, we determined the effect of Nef on Rac activity in these cells. The active GTP-bound form of Rac (Rac GTP ) binds the p21-binding domains (PBD) of PAKs directly ( Burbelo et al. 1995 ). Hence, we used a PBD–GST fusion protein in pulldown assays to measure the fraction of activated Rac in vivo. Jurkat T cells were transduced with a lentiviral vector directing the expression of HIV-1 Nef (FUGWCNA7) or a control empty vector (FUGW). Extracts prepared from these cells were incubated with PBD–GST and the fraction of PBD-bound Rac was determined by immunoblotting ( Figure 3 A). Notably, the expression of Nef resulted in a readily detectable increase in the steady-state level of PBD-bound Rac ( Figure 3 A, lane 3), consistent with the possibility that the interaction of Nef with DOCK2–ELMO1 increases Rac activation. Figure 3 Nef Activates Rac in Resting CD4 + T Lymphocytes (A) HIV-1 Nef activates Rac in Jurkat T cells. Jurkat T cells (lane 1) were transduced with a control empty vector (FUGW; lane 2) or the same vector expressing HIV-1 NA7 Nef (FUGWCNA7; lane 3). Rac GTP was precipitated from cell extracts with recombinant PAK1 PBD–GST. PBD–GST bound Rac GTP (top), total Rac present in extracts (middle), and Nef (bottom) were detected by immunoblotting. (B) Flow cytometric analysis of Gag and CD4 expression in resting CD4 + T lymphocytes transduced with HIV-1 derived vectors in the presence of IL-7. Percentages of cells productively infected with nef -deleted H-Δ vector (boxed area in middle panel) or with HIV-1 NA7 nef containing H-NA7 vector (right panel) are shown. Results obtained with uninfected control CD4 + T cells cultured in the presence of IL-7 are also shown (left panel). (C) HIV-1 Nef specifically activates Rac in resting primary CD4 + T lymphocytes. Rac GTP and CDC42 GTP were precipitated with PAK1 PBD–GST from extracts prepared from CD4 + T lymphocytes transduced with HIV-1 derived vectors, shown in (B), and analyzed as described in (A). In nontransformed T lymphocytes, Rac activation through DOCK2 is tied to chemotactic and antigenic stimuli. To assess whether Nef can uncouple these processes, we determined the effect of Nef on Rac activation in primary CD4 + T lymphocytes in the absence of stimulation with antigen and chemokines. While resting T cells are normally refractory to productive infection by lentiviruses and lentivirus-derived vectors, a sizable fraction becomes permissive for infection when cultured in the presence of cytokines such as IL-7 ( Unutmaz et al. 1999 ). We used this procedure to infect primary resting CD4 + T lymphocytes with an HIV-1-derived vector expressing HIV-1 NA7 Nef (H-NA7) or a control nef -deleted vector (H-Δ). Since HIV-1 Env protein may activate DOCK2-controlled signaling pathways through binding to chemokine receptors such as CXCR4, env -defective, VSV-G-pseudotyped viruses were used in these experiments. We cultured 98% pure populations of CD4 + T cells isolated from the peripheral blood leukocytes of healthy donors for 5 d in the presence of IL-7 and then transduced them with Nef-expressing H-NA7 or control H-Δ virus. The purity of the infected populations and the efficiency of transduction were assessed 4 d later by flow cytometric analysis of CD4 expression on the cell surface and intracellular p24 Gag expression, respectively. As shown in Figure 3 B, between 13% and 15% of CD4 + T cells were productively infected. Notably, the unusually low level of CD4 on the surface of cells infected with H-NA7 virus was due to robust downregulation of cell surface CD4 by NA7 Nef ( Mariani and Skowronski 1993 ). Cell extracts were prepared from the infected populations, and PBD–GST pulldown assays were performed to determine the fraction of activated Rac. Strikingly, infection with H-NA7 resulted in a readily detectable increase in the steady state level of activated Rac ( Figure 3 C). Based on direct quantitations of chemiluminescent signals of total and PBD–GST bound Rac, we estimated that approximately 1.2% of the total Rac in extracts from cells transduced with H-NA7 was bound to PBD–GST as compared to 0.2% in extracts from cells transduced with H-Δ. The activation of Rac was specifically due to the expression of Nef and not other viral gene products, as infection with the otherwise isogenic H-Δ virus did not increase PBD–GST-reactive Rac. To address the specificity of Nef effect towards Rac, we then asked whether Nef affects activity of CDC42 GTPase, which also uses PAK as a downstream effector ( Burbelo et al. 1995 ). Direct quantitations of chemiluminescent signals for total and PBD–GST-bound CDC42 revealed that less than 0.2% of the total CDC42 in extracts from H-NA7 and H-Δ transduced cells was PBD–GST bound. Therefore, we concluded that Nef primarily activates Rac and not CDC42 in CD4 + T lymphocytes in the absence of antigenic stimuli. Nef Activates Rac through DOCK2–ELMO1 Next we asked whether Nef activates Rac through DOCK2–ELMO1. To determine whether ELMO1 is required for the effect of Nef, we measured Rac activation by Nef in NS1 lymphoma cells, which do not express the endogenous ELMO1 ( Sanui et al. 2003b ) and in NS1 cells in which ELMO1 expression was restored by retrovirus-mediated transfer of ELMO1 cDNA (NS1 ELMO1 ). NS1 and NS1 ELMO1 cells were infected with a lentiviral vector expressing HIV-1 NA7 Nef (FUGWCNA7) or with a control empty vector (FUGW). Cell extracts were prepared from the infected populations, and PBD–GST pulldown assays were performed to determine the fraction of activated Rac. In agreement with a previous report (Sanui at al. 2003b), NS1 cells contain a small but readily detectable pool of activated Rac in spite of the lack of detectable ELMO1 expression, which is most likely generated by ELMO1-independent mechanism(s) ( Figure 4 A, lane 1). Notably, expression of Nef in the absence of ELMO1 and expression of ELMO1 in the absence of Nef did not increase the fraction of activated Rac (compare lane 2 to lane 1 and lane 3 to lane 1, respectively, in Figure 4 A). In contrast, expression of Nef in the presence of ELMO1 induced a readily detectably increase in the pool of activated Rac in the NS1 ELMO1 cells ( Figure 4 A, lane 4). These observations indicate that Nef activates Rac through an ELMO1-dependent mechanism. Figure 4 ELMO1 and DOCK2 Mediate Rac Activation by HIV-1 Nef (A) ELMO1 is required for Rac activation by Nef in NS1 cells. Rac GTP and total Rac in the extracts prepared from ELMO1-deficient NS1 cells (lanes 1 and 2) and ELMO1-expressing NS1 cells (lanes 3 and 4) following transduction with a lentiviral vector expressing HIV-1 Nef (lanes 2 and 4) or a control empty vector (lanes 1 and 3) were visualized as described in the legend to Figure 3 . (B) Nef activates Rac through DOCK2 and ELMO1 in HEK 293 cells. Rac GTP and total Rac in the extracts prepared from HEK 293 cells coexpressing the indicated proteins were visualized as described above. Next, we studied Rac activation by Nef in HEK 293 cells, which do not express endogeneous DOCK2. Combinations of Nef, DOCK2, ELMO1, and either Rac1 or Rac2 were expressed in HEK 293 cells by transient transfection. The expression of Nef in the absence of DOCK2 had little effect on the activation of either Rac isoform ( Figure 4 B, lanes 2 and 7). Ectopic expression of DOCK2 and ELMO1 increased the fraction of activated Rac1 and Rac2 by approximately 3- to 4-fold ( Figure 4 B, lanes 3 and 8), which is in agreement with a previous report ( Sanui et al. 2003b ). Significantly, coexpression of myristoylated, but not unmyristoylated, Nef ( Figure 4 B, lanes 4 and 9 versus lanes 5 and 10) with DOCK2 and ELMO1 further enhanced the fraction of activated Rac isoforms by approximately 2-fold, and this effect of Nef was more pronounced for Rac2 (compare lane 9 with lane 4 in Figure 4 B). Together, these data indicate that myristoylated Nef stimulates Rac activation through the DOCK2–ELMO1 complex. Nef Activates Rac through Association with DOCK2 and ELMO1 We identified mutations in Nef that disrupt Rac activation. Nef was reported to associate with an active form of PAK, and this interaction was suggested to be important for Nef effects on the cytoskeleton and T cell activation ( Fackler et al. 1999 ; Arora et al. 2000 ; Wang et al. 2000 ). Since PAK is an immediate downstream effector of Rac, we asked whether the abilities of Nef to activate Rac through DOCK2–ELMO1 and to associate with activated PAK are correlated. Two different Nef mutations that were previously reported to disrupt its association with activated PAK (P72A,P75A and R106A) ( Sawai et al. 1995 ; Renkema and Saksela 2000 ), yet unlike the myristoylation signal mutation (G2 ∇ HA) did not significantly affect Nef functions in other assays, were tested for their effects on Rac activation in Jurkat T cells ( Figure 5 A) and in HEK 293 cells transiently expressing DOCK2, ELMO1, and Rac2 ( Figure 5 B). Both mutations abolished the ability of Nef to stimulate Rac activation ( Figures 5 A and 5 B, lanes 4 and 5). As expected, the same was true for mutation of the myristoylation signal in Nef ( Figures 5 A and 5 B, lane 3). In contrast, a mutation that specifically abrogates the interaction of Nef with clathrin adaptor proteins (LL164AA; Greenberg et al. 1998 ) had little disruptive effect on Rac activation ( Figures 5 A and 5 B, lane 6). Figure 5 Nef Potentiates Rac Activation through Association with DOCK2–ELMO1 (A and B) Myristoylation signal, P72,P75, and R106 in Nef are required for Rac activation. Rac GTP and total Rac in the extracts prepared from Jurkat T cells transduced with lentiviral vectors expressing no Nef (−) or the indicated Nef proteins (A) and HEK 293 cells transiently coexpressing the indicated Nef mutants together with DOCK2, ELMO1, and Rac2 (B) were visualized as described in the legend to Figure 3 and quantified by direct imaging of chemiluminescent signals. The fraction of total Rac present in the extracts that was PBD–GST bound is shown in the histograms. Data in the histogram shown in (B) are averages of three independent experiments and error bars represent two standard deviations. (C) Myristoylation signal, P72,P75, and R106 in Nef are required for association with DOCK2, ELMO1, and Rac2. The ability of selected Nef mutants to associate with DOCK2, ELMO1, and Rac2 was determined as described in Figure 2 . Next we asked whether mutations in Nef that disrupted Rac2 activation affected the association with DOCK2, ELMO1, and Rac2 ( Figure 5 C). The LL164AA mutation, which did not significantly compromise the stimulation of Rac activation, did not affect the association of Nef with these proteins ( Figure 5 C, lane 6). In contrast, mutations that reduced Rac2 activation by Nef also diminished its association with DOCK2, ELMO1, and Rac2 ( Figure 5 C, lanes 3–5). Notably, the P72A,P75A and R106A mutations completely disrupted detectable association with ELMO1 and Rac2, but only weakened that with DOCK2, suggesting that Nef associates with both DOCK2 alone and DOCK2 complexed with ELMO1 and/or Rac2 and that the P72A,P75A and R106A mutations preferentially disrupt binding to the latter complex. Thus, robust stimulation of Rac2 activation by Nef requires its association with both DOCK2 and ELMO1. Functional Consequences of Nef Interactions with DOCK2–ELMO1 and Rac The CD4 + T lymphocyte is a major target of infection by primate lentiviruses. Nef was reported to lower the threshold signal required for antigen-induced responses of T cells ( Schrager and Marsh 1999 ; Wang et al. 2000 ), and this effect was proposed to be an important component to stimulation of viral replication by Nef in vivo ( Alexander et al. 1994 ; Simmons et al. 2001 ). Since DOCK2 mediates Rac activation downstream of the TCR to modulate T cell responsiveness and downstream of chemokine receptors to mediate chemotactic responses ( Fukui et al. 2001 ; Sanui et al. 2003a ), we studied effects of Nef on these processes in T lymphocytes. Purified populations of primary resting CD4 + T cells were transduced with VSV-G-pseudotyped H-NA7 or nef -deleted H-Δ in the presence of IL-7. Cells were stimulated 4–6 d following transduction with plate-bound anti-CD3 and anti-CD28 antibodies, mixed at various ratios, for various amounts of time. Intracellular IL-2 and p24 Gag were visualized and quantified by flow cytometry to provide a measure of cellular activation of the infected cells in response to the stimulation. In the absence of anti-CD3/anti-CD28 stimulation, neither uninfected (Gag-negative) nor the productively infected (Gag-positive) cells, produced detectable amounts of IL-2 ( Figure 6 ). In contrast, stimulation through CD3 and CD28 induced readily detectable accumulation of IL-2 in both H-Δ- and H-NA7-transduced populations. Notably, a larger fraction of productively infected CD4 + T lymphocytes typically proceeded to express IL-2 than uninfected cells. This phenomenon suggests that the permissive state for HIV-1 infection induced by IL-7 is associated with an increased responsiveness to activation via CD3 and CD28. However, no significant difference in the levels of IL-2 expression in cells infected with H-Δ compared to those infected with H-NA7 was detected across a wide range of stimulation conditions. These observations contrast with results from previous studies using T cell lymphoma and nontransformed CD4 + T lymphocytes expressing Nef alone ( Schrager and Marsh 2000 ; Wang et al. 2001 ). This difference could, for example, reflect the modifying effect of other HIV-1 gene products that were not tested in the previous experiments. On the other hand, we cannot exclude the possibility that IL-7 treatment masks the effect of Nef. Since Nef deregulates DOCK2 function and DOCK2 is essential for proper signaling through the immunological synapse, further studies under a variety of conditions that modulate the formation and function of the immunological synapse may be required to reveal this effect of Nef in the context of HIV-1 infection. Figure 6 Effect of Nef on IL-2 Expression in HIV-1-Infected CD4 + T Lymphocytes Stimulated through CD3 and CD28 CD4 + T lymphocytes transduced with H-Δ and H-NA7 HIV-1-derived vectors were not stimulated (unstimulated) or stimulated with immobilized anti-CD3 and anti-CD28 mAbs (anti-CD3, anti-CD28) in the presence of Golgi-Stop for 5 h and stained for intracellular IL-2 and p24 Gag. Percentages of IL-2-positive and IL-2-negative cells in the Gag-negative and Gag-positive populations are shown. DOCK2 regulates the activation of Rac proteins during lymphocyte migration in response to chemokine gradients ( Fukui et al. 2001 ). Therefore, we also asked whether Nef affects lymphocyte chemotaxis. Jurkat T cells, which constitutively express CXCR4, a major coreceptor for T-cell tropic HIV and a receptor for stromal-derived factor 1 (SDF-1) ( Deng et al. 1996 ; Feng et al. 1996 ), were transiently transfected with a control plasmid expressing enhanced green fluorescent protein (GFP) alone or with a plasmid that coexpresses Nef and a GFP marker protein from the same bicistronic transcription unit. We then measured the chemotaxis of transfected populations to SDF-1 using a transwell migration assay. The relative frequency of control and Nef-expressing cells in the migrated populations was determined by flow cytometric measurement of GFP expression. Approximately 30% of control cells migrated regardless of the level of GFP expression, indicating that the chemotaxis in this assay was robust ( Figure 7 A). In contrast, the chemotaxis of cells coexpressing Nef and GFP was inhibited in a dose-dependent manner. Figure 7 Nef Disrupts T Cell Migration to SDF-1 (A) Migration of cell populations shown in (B) expressing GFP (open circle), ectopic CXCR4, and GFP (filled circle), HIV-1 NA7 Nef and GFP (open box), HIV-1 NA7 Nef, ectopic CXCR4 and GFP (filled box) to SDF-1 was measured in transwell assays. (B) Transient expression of ectopic CXCR4 restores CXCR4 levels on the surface of Nef-expressing cells. Flow cytometric analysis of Jurkat T cells transiently expressing GFP (panel 1) or HIV-1 NA7 Nef and GFP (panel 2) and together with ectopic CXCR4 (panels 3 and 4, respectively). Control experiments revealed that Nef caused a modest decrease in cell surface expression of CXCR4 (compare panels 1 and 2 in Figure 7 B). This observation raised the possibility that Nef-expressing cells were unresponsive to SDF-1 due to abnormally low levels of CXCR4 at the cell surface rather than due to deregulation of the DOCK2–ELMO1 complex. To address this possibility, we restored CXCR4 on the surface of Nef-expressing cells to levels equal to and even higher than those seen in control cells by transiently expressing ectopic CXCR4 receptor and then performed migration assays using these cell populations ( Figure 7 B, panels 3 and 4). Significantly, the migration of cells with restored CXCR4 levels was still inhibited by Nef expression ( Figure 7 A). We concluded that HIV-1 Nef blocks lymphocyte migration to SDF-1 principally by interfering with CXCR4-controlled signaling cascades rather than by downregulating CXCR4 from the cell surface. We then asked whether inhibition of Jurkat T cell migration by Nef correlated with its ability to potentiate Rac activation by DOCK2 and ELMO1. As expected, we found that disruption of the myristoylation signal in Nef (NA7 (G2 ∇ HA) ) abolished the inhibition of migration ( Figure 8 A and 8 B). Furthermore, the P72A,P75A, and R106A mutations diminished the ability of Nef to block migration, albeit to different extents. In contrast, the LL164AA mutation, which had little disruptive effect on enhancement of Rac activation, was fully functional in this assay. These observations suggested that deregulated activation of Rac GTPases is instrumental for the defective migration of Nef-expressing cells. To explore this possibility further, we ectopically expressed constitutively active Rac1 and Rac2 (Rac1 G12V , Rac2 G12V ) and, as controls, wild-type Rac1 and Rac2 in Jurkat T cells and measured their migration to SDF-1 ( Figure 8 C). Expression of wild-type Rac GTPases stimulated chemotaxis approximately 2- to 3-fold. In contrast, expression of the constitutively active forms of each Rac GTPase severely suppressed lymphocyte migration to SDF-1. Thus, deregulated Rac activation inhibits directional cell movement, most likely by disrupting spatially organized rearrangements of the cytoskeleton that are induced by chemokine gradients. These results further support a model in which Nef disrupts migration to SDF-1 by activating Rac through the DOCK2–ELMO1 module and thus uncoupling Rac activation from chemokine receptor signaling. Figure 8 Nef Disrupts Chemotaxis by Activating Rac through DOCK2–ELMO1 (A and B) Jurkat T cells expressing wild-type or mutant HIV-1 Nef proteins and GFP reporter were used in transwell chemotaxis assays with SDF-1. Percentage of migrated cells expressing GFP alone (open circle), or together with HIV-1 NA7 (open square), NA7 (G2 ∇ HA) (open diamond), NA7 (P72A,P75A) (open triangle), NA7 (R106A) (filled circle), and NA7 (LL164AA) (filled triangle) is shown as a function of GFP fluorescence intensity in (A) and in (B) for the single GFP fluorescence intensity interval indicated by the shaded rectangle in (A). (C) Constitutively active Rac GTPases disrupt lymphocyte migration to SDF-1. Migration of Jurkat T cells transiently expressing wild-type (Rac1, Rac2), constitutively active (Rac1 G12V , Rac2 G12V ), or as a control HIV-1 Nef were also measured. Data shown are averages of three independent experiments and error bars represent two standard deviations. The above observations predicted that Nef likely causes a general migration defect. Therefore, we also studied the effect of Nef on Jurkat T cell migration to the MIP-1β chemokine. Since the MIP-1β receptor (CCR5) is not constitutively expressed in Jurkat T cells, we transiently expressed CCR5 and GFP marker from a bicistronic vector either alone or together with HIV-1 NA7 Nef. Using this vector, CCR5 expression levels were positively correlated with GFP marker protein expression ( Figure 9 A, panels 2 and 3). Notably, flow cytometric analysis revealed that CCR5 cell-surface expression was not downregulated by Nef (compare panels 5 and 6 in Figure 9 A). We then measured the ability of these cells to migrate to MIP-1β. Migration of cells coexpressing CCR5 and HIV-1 Nef was impaired compared to control cells expressing CCR5 at comparable levels ( Figure 9 B). These data confirm that Nef induces a general defect in lymphocyte migration by targeting DOCK2–ELMO1 and Rac. Figure 9 HIV-1 Nef Disrupts CCR5-Mediated Migration (A) HIV-1 Nef does not downregulate CCR5. Flow cytometric analysis of CCR5 and GFP in Jurkat T cells transiently expressing GFP alone (panel 1) or CCR5 and GFP in the absence (panel 2) and presence (panel 3) of HIV-1 Nef. Histograms of CCR5 expression for cell populations within a single GFP fluorescence intensity interval indicated by the rectangle in panel 1 are shown in panels 4 to 6, respectively. (B) Percentage of cells migrated to MIP-1β and expressing GFP alone (open circle), GFP and CCR5 (open triangle), or GFP, CCR5 and HIV-1 Nef (filled triangle) is shown as a function of GFP fluorescence intensity. Discussion Nef is a multifunctional adaptor protein that modulates signal transduction and protein-sorting machineries. We purified to near homogeneity an abundant Nef-associated protein complex from T cells and identified by mass spectroscopy its major subunits as DOCK2 and ELMO1, a bipartite Rac activator ( Sanui et al. 2003b ). Notably, the extensive large-scale biochemical purification and sensitive proteomic analyses described in this report did not detect several cellular proteins previously reported to associate with HIV-1 Nef and mediate the effects of Nef in the cell (data not shown). This includes cellular proteins such as Vav ( Fackler et al. 1999 ), PAKs ( Sawai et al. 1995 ), inositol triphosphate receptor ( Manninen and Saksela 2002 ), Lck protein-tyrosine kinase ( Baur et al. 1997 ), clathrin adaptors ( Greenberg et al. 1997 ; Le Gall et al. 1998 ; Piguet et al. 1998 ), and others. Why were proteins previously reported to associate with Nef not detected in our studies? One possible explanation is that previously reported associations are unstable under the biochemical purification conditions used in our studies. This likely explains the apparent absence of clathrin adaptors in Nef preparations purified in our studies, since we know that Nef binds AP-2/AP-1 clathrin adaptors only weakly in the salt and pH conditions used here (data not shown). It is also possible that in some cases, epitope tags may be buried and therefore inaccessible to the monoclonal antibodies (mAbs) used for immunoaffinity purification. Moreover, some of the previously reported associations, especially those with protein kinases, such as PAKs, were best detected by an ultrasensitive in vitro kinase assay ( Sawai et al. 1995 ) and, as our data show, are likely of exceedingly low stoichiometry. (Of note, we did detect the presence of p62 phosphoprotein [PAK] by in vitro kinase assays of anti-Nef purifications [data not shown]). Finally, some of the described associations, such as that with thioesterase, are known to occur only with selected Nef variants ( Cohen et al. 2000 ), while those we report here occur with all Nef variants tested. Nonetheless, the specific isolation of the Nef–DOCK2–ELMO1–Rac complex reported here provides strong biochemical evidence to reinforce predictions from previous genetic studies that Nef functions through multiple independent interactions with different sets of downstream effector proteins. Our observations strongly argue that DOCK2–ELMO1 is the major upstream regulator used by Nef to activate Rac in T cells and that through this interaction Nef can activate Rac in CD4 + T lymphocytes even in the absence of stimulation with antigen or chemokines. These data indicate that Nef targets a critical switch, DOCK2–ELMO1, that regulates Rac GTPases downstream of chemokine receptors and the TCR and uses it to modulate the downstream processes they control. Thus, the interaction of Nef with DOCK2 and ELMO1 provides an important mechanism by which Nef may impact pathogenesis by primate lentiviruses. Our data strongly suggest that DOCK2–ELMO1 is the major activator of Rac targeted by Nef in T lymphocytes. This model is supported by the observations that Nef physically associates with a complex that contains DOCK2–ELMO1 and Rac, that specific mutations in Nef simultaneously disrupt its ability to bind this complex and to activate Rac, and that Nef fails to activate Rac in the absence of ELMO1. Although previous reports implicated Vav, a Rac1 guanine nucleotide exchange factor (GEF), as the critical downstream effector that Nef binds directly to activate PAK ( Fackler et al. 1999 ), our data do not support this possibility. Notably, we have been unable to detect the presence of Vav in anti-Nef immune complexes by both proteomic analyses and immunoblotting, indicating that these interactions are of low abundance relative to those with DOCK2 and ELMO1 and are therefore unlikely to mediate the bulk of Nef's impact on the Rac pathway. Significantly, ELMO1 is ubiquitously expressed ( Gumienny et al. 2001 ) and can associate with Nef in nonlymphoid cells in the absence of DOCK2 (data not shown). Thus, we postulate a general mechanism in which ELMO1, possibly in complex with another CDM family protein, mediates Rac activation and PAK recruitment by Nef that is observed in nonlymphoid cells ( Sawai et al. 1995 ; Fackler et al. 1999 ; Arora et al. 2000 ). Our observation that the expression of Nef from the integrated HIV-1 provirus in primary CD4 + T cells did not alter IL-2 production in standard activation protocols was unexpected, because previous genetic evidence linked Nef binding to DOCK2–ELMO1 and the ensuing activation of Rac GTPases to the the ability of Nef to facilitate T cell activation. Specifically, the same mutations that we observed to disrupt Rac activation through the DOCK2–ELMO1 complex were previously reported to disrupt the stimulatory effect of Nef on aspects of T cell activation ( Sawai et al. 1995 ; Simmons et al. 2001 ), and the observed effects were in some cases dramatic ( Wang et al. 2000 ). Moreover, Rac activation by DOCK2 facilitates T cell responsiveness to antigen, as disrupted Rac activation in DOCK2 (−/−) and Rac2 (−/−) mice is associated with defective immunological synapse formation and depressed antigen-specific responses ( Yu et al. 2001 ; Sanui et al. 2003a ). Since Nef deregulates DOCK2 function and DOCK2 is essential for proper signaling through the immunological synapse, further studies under a variety of conditions that modulate the formation and function of the immunological synapse may be required to reveal the effect of Nef in the context of HIV-1 infection. Interestingly, the chemokine receptor system plays crucial roles in infection by primate lentiviruses. Previous studies revealed its essential role for virus entry into target cells ( Deng et al. 1996 ; Feng et al. 1996 ). Primate lentiviruses also exploit this system to recruit uninfected target cells to sites of viral replication ( Weissman et al. 1997 ; Swingler et al. 1999 ). Our observations reveal an additional level of complexity in primate lentivirus–chemokine receptor system interactions. Inhibiting chemotaxis of the infected T cells likely disrupts the generation of the immune response. During generation of the immune response, T cells are initially activated in the paracortex of lymph nodes and then migrate to the edges of follicles, where they interact with antigen receptor-activated B cells ( Garside et al. 1998 ). This physical interaction is required to drive B cells towards antibody production, isotype switching, and the affinity maturation of the antibody response. Hence, the development and maturation of the immune response require the ordered migration of activated T cells to specific sites within lymphoid tissue ( Delon et al. 2002 ). Notably, recent in vivo evidence documents Nef-dependent alterations in the distribution of SIV mac239-infected CD4 + T lymphocytes in the lymph nodes of experimentally infected rhesus macaques ( Sugimoto et al. 2003 ). In lymph nodes of monkeys infected with nef- deleted SIV, most infected T cells were located in the B cell-rich follicles and in the border region between the paracortex and the follicles. In contrast, in monkeys infected with SIV harboring a functional nef gene, most productively infected T cells remained in the T cell-rich paracortex and were only infrequently present in proximity to B cell follicles. This evidence shows that Nef disrupts the ordered migration patterns of infected CD4 + T lymphocytes in vivo and reinforces the possibility that this disruption impairs the generation and maturation of the immune response to antigens. Since a large fraction of HIV-1-infected CD4 + T cells is specific for HIV-1 antigens ( Douek et al. 2002 ), this effect of Nef provides another mechanism to suppress the antiviral immune response. Taken together, the functional interaction of Nef with DOCK2, ELMO1, and Rac enables Nef to modulate multiple aspects of T cell function. Materials and Methods Construction of expression plasmids Sequences encoding variant and mutant Nef proteins tagged at their C-termini with a peptide (hf) containing the HA and FLAG epitopes (DTYRYIYANATYPYDVPDYAGDYKDDDDK) were subcloned into pBABE- puro and pCG expression plasmids ( Morgenstern and Land 1990 ; Greenberg et al. 1998 ). In NA7 (G2 ∇ HA) Nef, the myristoylation signal was disrupted by insertion of the HA epitope (ANATYPYDVPDYAG) at glycine G2. The full-length human ELMO1 cDNA (clone IMAGE:4521393; ResGen, Carlsbad, California, United States), DOCK2 cDNA (KIAA0209, clone ha04649; Kazusa DNA Research Institute, Chiba, Japan), and cDNAs encoding wild-type and mutant forms of Rac1 and Rac2 (kindly provided by Linda Van Aelst, Cold Spring Harbor Laboratory, Cold Spring Harbor, New York, United States) were subcloned into pCG plasmids containing N-terminal c-Myc (EQKLISEEDL) or polyhistidine (HHHHHHH) epitope tags, using standard techniques. pBABE ELMO1 contains an N-terminal c-Myc epitope-tagged ELMO1 cDNA subcloned into the pBABE- neo vector. H-NA7 was constructed by substituting the HIV-1 NL4–3 nef coding sequence with that of HIV-1 NA7 nef ( Mariani and Skowronski 1993 ) in pNL4–3 carrying a frameshift mutation at the KpnI site at position 6463 in env , kindly provided by Klaus Strebel (National Institute of Allergy and Infectious Diseases, National Institutes of Health, Bethesda, Maryland, United States). H-Δ is based on H-NA7 with a deletion that removes residues 1–34 of Nef and prevents expression of Nef protein. For chemotaxis assays, cDNAs encoding HIV-1 NA7 Nef, human CCR5 (kindly provided by Frank Kirchhoff, Universitätsklinikum, Ulm, Germany), and Rac proteins were subcloned into the pCGCG bicistronic vector that directs the expression of GFP from an internal ribosomal entry site element ( Lock et al. 1999 ). Generation of stable cell lines pBABE plasmids were introduced into the Phoenix amphotropic packaging cell line, kindly provided by G. P. Nolan (Stanford University Medical Center, Palo Alto, California, United States), by calcium phosphate coprecipitation, and viral supernatants were used to infect a Jurkat T cell subline ( Greenberg et al. 1998 ), provided by Dan R. Littman (New York University School of Medicine, New York, New York, United States), or NS1 cells. Transduced Jurkat cells were selected and subsequently maintained in the presence of puromycin (0.4 μg/ml) (Sigma, St. Louis, Missouri, United States). Transduced NS1 cells were selected and maintained in the presence of G418 (1.0 mg/ml) (Invitrogen, Carlsbad, California, United States). Immunoaffinity purification of epitope-tagged Nef and associated proteins Unless stated otherwise, all reactions were performed at 4°C. Approximately 1.8 × 10 10 Jurkat T cells stably expressing NA7-hf (or control cells) were lysed for 1 h in 200 ml of LB buffer (150 mM NaCl, 50 mM Tris–HCl [pH 7.5], 1% Triton X-100, 10% glycerol) (Complete Protease Inhibitors, Roche, Basel, Switzerland). Extracts were precleared with protein G–agarose (Roche) for 1 h and incubated with 12CA5 mAb crosslinked to protein G–agarose beads for 4 h. The immunoprecipitate was washed extensively with LB, and bound proteins were eluted by competition with HA peptide (0.2 mg/ml) (ANATYPYDVPDYAG; Invitrogen) in LB for 45 min at 30°C. The eluate was incubated with anti-FLAG M2 affinity gel beads (Sigma) overnight, and the immunoprecipitate was washed extensively with LB and then LB modified to contain 0.1% Triton X-100 (FEB buffer). Proteins were eluted with FEB containing FLAG peptide (0.2 mg/ml) (Sigma) for 45 min at 30°C. The eluate was concentrated on Microcon centrifugal filter devices (Millipore, Billerica, Massachusetts, United States) with a cutoff of 8 kDa. Protein identification by mass spectrometry Nef and associated proteins were separated by 8%–17% SDS-PAGE and visualized using SYPRO stain (Molecular Probes, Eugene, Oregon, United States). Visible bands were excised and processed for identification by mass spectrometry. The samples were analyzed by LC-MS/MS as described previously ( Hu et al. 2002 ). Spectra resulting from LC/MS/MS were analyzed with the SONARS software package (ProteoMetrics LLC, New York, New York, United States). Lentiviral vectors and infections FUGWCNA7 contains the wild-type HIV-1 NA7 nef coding sequence under control of the CMV promoter, subcloned downstream of the Woodchuck-responsive element in the FUGW lentiviral vector ( Lois et al. 2002 ). FUGWC vectors containing amino acid substitutions in Nef have a similar structure. Supernatants containing infectious particles were produced by calcium phosphate cotransfection of HEK 293 cells, as described previously ( Lois et al. 2002 ). For biochemical analyses of Rac activation, approximately 10 7 Jurkat T cells or NS1 cells were infected with supernatants containing approximately 10 7 infectious units of FUGW or FUGWC vectors encoding wild-type or mutant HIV-1 NA7 Nef proteins, in the absence of polycationic agents. Cell extracts were prepared 3–4 d following infection and used for PBD–GST pulldown assays. PBMCs were purified from buffy coats of healthy donors (New York Blood Bank, Hicksville, New York, United States) by density gradient separation on Lymphocyte Separation Medium (ICN Biomedicals, Inc., Irvine, California, United States), and CD4 + T lymphocytes were isolated using CD4 + T Cell Enrichment Columns (R&D Systems, Minneapolis, Minnesota, United States). Replication incompetent HIV-1 particles pseudotyped with VSV-G were produced by calcium phosphate transfection of HEK 293 cells and used to infect >98% pure populations of CD4 + T lymphocytes that were cultured in the presence of IL-7 ( Unutmaz et al. 1999 ). Cell-surface CD4 in the infected populations was revealed with FITC-conjugated SK3 mAb (Becton Dickinson, San Jose, California, United States). Cells stained for CD4 were permeabilized using Cytofix/Cytoperm Kit (BD PharMingen, San Jose, California, United States) and p24 Gag expression was revealed with PE-conjugated KC57-RD1 mAb (Beckman Coulter, Inc., San Diego, California, United States) as described elsewhere ( Mascola et al. 2002 ). CD4 and Gag expression were quantitated simultaneously using an LSR-II flow cytometer (Becton Dickinson). Cell stimulations and IL-2/Gag assays Anti-CD3 mAb and anti-CD28 mAb (MAB100 and MAB342; R&D Systems), alone or in combinations, were immobilized on 12-well microtiter plates (351143; Becton Dickinson) in PBS overnight at 4°C. Wells were washed three times with PBS, and 5 × 10 5 CD4 + T lymphocytes were added 4–5 d after they were transduced with H-Δ or H-NA7 vectors in the presence of IL-7. In some experiments, transduced cells were cultured for additional 48 h in the absence of IL-7 before stimulations. Stimulations were performed for 4 h to 16 h in the presence and absence of Golgi-Stop (Becton Dickinson). Cells were recovered from wells by vigorous pipetting, fixed, and permeabilized using Cytofix/Cytoperm Kit (BD Bioscience PharMingen). Intracellular IL-2 and p24 Gag were revealed simultaneously with PE-conjugated rat anti-human IL-2 mAb, (559334; B&D Biosciences PharMingen) and FITC-conjugated KC57-RD1 mAb (Beckman Coulter, Inc.), respectively. Transient transfections of HEK 293 cells, immunoprecipitations, and immunoblotting HEK 293 cells were transfected by calcium phosphate coprecipitation with 20 μg of pCG plasmids expressing epitope-tagged DOCK2, ELMO1, Rac2 or Rac1, and/or tagged Nef proteins and a control empty vector. Cells were lysed 48 h posttransfection in LB buffer. To isolate Nef and associated proteins, extracts were incubated overnight with anti-FLAG M2 Affinity Gel (Sigma), the immunoprecipitates were washed four times with LB buffer, once with LB containing 0.5 M LiCl, and proteins were eluted with FLAG peptide as described above. To isolate DOCK2 and associated proteins, extracts were incubated with Ni–NTA agarose (Qiagen, Valencia, California, United States) and proteins were eluted from the precipitate with 250 mM imidazole. DOCK2–Nef complexes were isolated from imidazole eluates with anti-FLAG M2 affinity gel as described above. Eluted proteins proteins were separated by 16% SDS-PAGE, electroblotted onto PVDF membrane (Millipore), and immunoblotted with the following antibodies: anti-c-Myc mAb (1:100; Oncogene, Tarrytown, New York, United States), anti-FLAG M2 mAb (1:5000; Sigma), 12CA5 mAb (1:5000), or rabbit serum raised to a DOCK2-specific peptide (GDKKTLTRKKVNQFFKTM). Immune complexes were revealed with HRP-conjugated antibodies specific for the Fc fragment of mouse or rabbit immunoglobulin G. (1:5,000; Jackson ImmunoResearch Laboratories, Inc., West Grove, Pennsylvania, United States) and ECL (Amersham, Little Chalfort, United Kingdom). Rac and CDC42 activity assays Cells were lysed in LB modified to contain 500 mM NaCl, 0.5% sodium deoxycholate, 0.1% SDS, and 10 mM MgCl 2 (RB buffer). Extracts were incubated for 2 h with 40 μg of recombinant PAK1 PBD–GST (kindly provided by Linda Van Aelst) bound to glutathione–agarose beads (8 mg/ml) (Roche), and beads were washed extensively with RB buffer. Bound proteins and aliquots of extracts were separated by 16% SDS-PAGE. Rac was immunoblotted with anti-c-Myc epitope mAb (1:100; Oncogene), or with anti-Rac mAb (1:1000; BD Transduction Laboratories), CDC42 was detected with sc-87 rabbit antibody (1:100; Santa Cruz Biotechnology, Santa Cruz, California, United States), and Nef was detected with rabbit serum raised against HIV-1 HxB3 Nef (1:300; Mariani and Skowronski 1993 ). Immune complexes were visualized by chemiluminescence using Lumi-Lite Plus (Roche). Chemiluminescent signals were imaged and quantitated using the FluorChem Imaging System and software (Alpha Innotech, Cannock, United Kingdom). Chemotaxis assays Jurkat T cells were transfected by electroporation with plasmids coexpressing Nef and GFP marker protein from a single bicistronic transcription unit (pCGCG NA7; Lock et al. 1999 ), except that in some experiments were also cotransfected with pCG CXCR4, expressing human CXCR4 receptor. Transfected populations were used 24–48 h later in transwell chemotaxis assays in the presence of 10 ng/ml of SDF-1 (R&D Research). For CCR5-dependent migration, Jurkat T cells were transfected with pCGCG CCR5, expressing human CCR5 and GFP, alone or in the presence pCG NA7, and migration to MIP-1β (10 ng/ml) was measured. Cells were allowed to migrate for 2 h, and the relative frequency of GFP-positive cells in the initial and migrated populations was determined by flow cytometry using an LSR-II flow cytometer (Becton Dickinson). Supporting Information Accession Numbers The LocusLink ( http://www.ncbi.nlm.nih.gov/LocusLink/ ) accession numbers of the genetic loci discussed in this paper are DOCK2 (LocusLink ID 1794), ELMO1 (LocusLink ID 9844), Rac1 (LocusLink ID 5879), and Rac2 (LocusLink ID 5880).	/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC314466.xml
340962	Learning to Discern Images Modifies Neural Activity	xx	The primate brain processes a remarkably diverse array of visual cues to recognize objects in dynamic settings crammed with unfamiliar objects. Not surprisingly, repeated viewing aids recognition, but how the brain orchestrates this experience-driven improvement is unclear. Visual input to the brain travels from the eye to the primary visual cortex (V1), at the back of the brain. From there, signals are sent to nearby extrastriate cortical areas, which process “early” visual cues. Both the “lower level” extrastriate cortex and “higher level” inferior temporal (IT) cortex are important for object recognition in primates. In monkeys and humans, lesions in the IT cortex severely affect the ability to recognize objects. In these higher-level cortical regions, neurons carry more information about an object after subjects learn to recognize that object. This modified neural activity is thought to reflect internal representations of specific aspects of the learned task—such as learned recognition of three-dimensional objects—and these representations often remain stable even though certain features of the visual stimulus—such as size or image degradation—change. With recent evidence suggesting that lower level brain regions like the primary visual cortex are also capable of learning-related modifications, it appears that both early and higher brain areas of the “ventral visual stream” benefit from learning. It is not clear, however, how learning modifies these discrete brain regions to coordinate this processing. By training monkeys to recognize degraded images, Gregor Rainer, Han Lee, and Nikos Logothetis of the Max Planck Institute for Biological Cybernetics in Germany have identified a subset of neurons that compensate for indistinct visual inputs by coordinating disparate regions in the brain. The monkeys' improved performance, they propose, stems from the informational enrichment of a subset of lower level neurons. Along with an increase in learning-induced firing activity, V4 neurons—extrastriate cortical neurons associated with detecting visual input of intermediate complexity—encode more information about relevant details to resolve indeterminate visual cues. V4 neurons likely interact with higher cortical levels to help the monkeys interpret the degraded indeterminate images as something recognizable. The researchers presented the monkeys with different “natural” images, including pictures of birds and humans, then subjected the images to different levels of “stimulus degradation”—making them harder to read by adding varying amounts of visual noise. Using this approach, the researchers could record the activity of the V4 neurons as the monkeys were presented with the different images. The monkeys viewed a sample image and then signaled whether a second image, presented after a brief delay, was a match or not. When Rainer et al. analyzed the activity of the V4 neurons associated with the different images, they found there was no significant change in the activity or information conveyed by V4 neurons associated with novel or undegraded familiar images. On the other hand, learning not only significantly improved the monkeys' ability to recognize degraded stimuli but also increased both the activity and informational encoding of the V4 neurons. But how did individual V4 neurons facilitate this enhanced ability to recognize degraded stimuli? After identifying a subset of neurons that showed enriched neural activity in response to degraded or indeterminate stimuli, the researchers studied the monkeys' eye movements to determine any behaviors that might explain why monkeys performed better with familiar degraded stimuli. They mapped the monkeys' eye movements while allowing them to freely view the different familiar and novel images—but this time with just two coherence levels (undegraded and 45% coherent). There was substantially more overlap, in terms of where the monkeys looked for the 45% and 100% coherent images after learning. This suggests that monkeys learned to focus their attention on particular salient features, and were thus better able to identify degraded versions of these images. Neurons in the V4 area appear to be recruited to distinguish the relevant visual signal from the visual noise, and thus play a critical role in resolving indeterminate stimuli when salient features are present. These results, together with previous studies showing the sensitivity of prefrontal cortex neurons to novel stimuli, indicate that the prefrontal cortex processes novel stimuli while the V4-rich extrastriate visual areas convey details about hard to decipher images. It may be that as the V4 neurons refine their competence through learning, they also support the ability of the prefrontal cortex to process different but similar visual cues. Vision is a dynamic process, Rainer et al. conclude, characterized by ongoing interactions between stimulus-driven brain regions and feedback from higher-order cognitive regions. Monkeys can learn to recognize degraded images	/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC340962.xml
549512	Sweet proteins – Potential replacement for artificial low calorie sweeteners	Exponential growth in the number of patients suffering from diseases caused by the consumption of sugar has become a threat to mankind's health. Artificial low calorie sweeteners available in the market may have severe side effects. It takes time to figure out the long term side effects and by the time these are established, they are replaced by a new low calorie sweetener. Saccharine has been used for centuries to sweeten foods and beverages without calories or carbohydrate. It was also used on a large scale during the sugar shortage of the two world wars but was abandoned as soon as it was linked with development of bladder cancer. Naturally occurring sweet and taste modifying proteins are being seen as potential replacements for the currently available artificial low calorie sweeteners. Interaction aspects of sweet proteins and the human sweet taste receptor are being investigated.	Sweet and taste modifying proteins The prevalence of obesity and diabetes has increased dramatically in recent years in the United States, with similar patterns seen in several other countries including India [ 1 ] as well. Diabetes mellitus is a chronic disease caused by inherited or acquired deficiency in production of insulin by the pancreas or by the ineffectiveness of the insulin produced [ 2 ]. Artificial sweeteners like Saccharin, Aspartame, Cyclamate and AcesulfameK are used world-wide as low calorie sweeteners by patients affected by diseases linked to the consumption of sugar, e.g. diabetes, hyperlipemia, caries, obesity etc. but they have side effects such as psychological problems, mental disorders, bladder cancer, heart failure and brain tumors [ 3 - 7 ]. Sweet proteins have the potential to replace these artificial sweeteners, by acting as natural, good, low calorie sweeteners, as we know that proteins do not trigger a demand for insulin in these patients whereas sucrose does. In humans, the sweet taste is mainly due to the recently discovered T1R2-T1R3 receptor [ 8 - 10 ], two of the three members of the T1R class [ 8 - 10 ] of taste-specific proteins hypothesized to function in combination as a heterodimer. The human T1R2-T1R3 receptor recognizes natural and synthetic sweetness and T1R1-T1R3 recognizes umami taste [ 11 , 12 ]. So far there are seven known sweet and taste-modifying proteins, namely Brazzein [ 13 ], Thaumatin [ 14 ], Monelin [ 15 ], Curculin [ 16 ], Mabinlin [ 17 ], Miraculin [ 18 ] and Pentadin [ 19 ]. Properties and characteristics of these proteins are illustrated in Table 1 . The key residues on the protein surface responsible for biological activity have not yet been identified with certainty for any of these proteins [ 20 ]. Monellin was found to be 100000 times sweeter than sucrose on a molar basis [ 21 ], followed by Brazzein and Thaumatin which are 500 times [ 13 ] and 3000 times sweeter then sucrose [ 14 ] respectively (the latter two on a weight basis). All of these proteins have been isolated from plants that grow in tropical rainforests. Although most of them share no sequence homology or structural similarity, Thaumatin shares extensive homology with certain non-sweet proteins found in other plants [ 15 ]. The potential industrial applications of these proteins are the low calorie sweetener industry and the cola, snacks, food and chocolate industries. Brazzein Brazzein is the smallest, most heat-stable [ 13 ] and pH-stable member of the set of proteins known to have intrinsic sweetness. The protein, consisting of 54 amino acid residues, is reported to be between 500 and 2000 times sweeter than sucrose [ 22 ] and represents an excellent alternative to available low calorie sweeteners. It was originally isolated from the fruit of an African plant Pentadiplandra brazzeana Baillon [ 23 ]. Heat and pH stability of the protein make it an ideal system for investigating the chemical and structural requirements of a sweet-tasting protein. Based on the wild-type brazzein, 25 mutants were produced to identify critical regions important for sweetness. To assess their sweetness, psychophysical experiments were carried out with 14 human subjects. First, the results suggest that residues 29–33 and 39–43, plus residue 36 between these stretches, as well as the C-terminus are involved in the sweetness [ 24 ]. Second, charge plays an important role in its interaction with the sweet taste receptor [ 24 ]. Thaumatin The thaumatins are a class of intensely sweet proteins isolated from the fruit of the tropical plant Thaumatococcus danielli . The protein crystallizes in a hexagonal lattice after a temperature shift from 293 to 277 K. The structure has been solved at 1.6 Å resolution. Its fold was found to be identical to that found in three other crystal forms grown in the presence of crystallizing agents of differing chemical natures [ 25 ]. It consists of 207 amino acid residues with eight intramolecular disulfide bonds and contains no free cysteine residues. It aggregates upon heating at pH 7.0 above 70 degrees C, whereupon its sweetness disappears [ 26 , 27 ]. The protein is approximately 10000 times sweeter than sugar on a molar basis [ 28 ]. It is a protein that tastes intensely sweet only to Old World monkeys and to higher primates, including man [ 29 ], as it has been found that the protein binds to certain elements in taste pores of Rhesus monkey foliate papillae [ 30 ]. Thaumatin has been approved for use in many countries as both a flavor enhancer and a high-intensity sweetener [ 31 ]. Monellin Monellin, a sweet protein, consists of two noncovalently associated polypeptide chains, an A chain of 44 amino acid residues and a B chain of 50 amino acid residues [ 32 ]. The protein can be purified from the fruit of Dioscoreophyllum cumminsii grown in West Africa and is approximately 100,000 times sweeter than sugar on a molar basis and several thousand times sweeter on a weight basis [ 28 ]. Single-chain monellin (SCM), which is an engineered 94-residue polypeptide, has proven to be as sweet as native two-chain monellin, and is more stable than the native monellin at high temperature and in acidic environments [ 33 ]. Native monellin is relatively sensitive to heat or acid treatment, which may cause separation of the sub-units and denaturation of the protein. Despite misgivings about the stability of the protein to heat and acid, downstream processes have been established. Its D-enantiomer has been crystallized and analyzed by X-ray crystallography at 1.8 Å resolution. Two crystal forms (I and II) were found under crystallization conditions similar, but not identical, to the crystallization conditions of natural L-monellin [ 34 ]. One NMR study of a non-sweet analog in which the Asp B7 of protein was replaced by Abu B7 (L-2-Aminobutylicacid), showed similar 3-dimensional structures of these two proteins, indicating that the lack of the beta-carboxyl group in the Abu B7 analog is responsible for the loss of sweetness [ 35 ]. Recent research on identifying binding sites on the receptor by means of structure-taste relationships, found that four monellin analogues, [AsnA16]-, [AsnA22]-, [GlnA25]-, and [AsnA26]-monellin were 7500, 750, 2500, and 5500 times as sweet as sucrose on a weight basis, respectively. Thus, among them, [AsnA22]-monellin and [GlnA25]-monellin were less sweet than the native monellin [ 36 ]. Curculin Curculin which is extracted from Curculigo latifolia acts as a good low calorie sweetener. Its maximum sweetness is equal to 0.35 M of sucrose. It has taste modifying abilities since water and sour substances elicit a sweet taste after consumption of curculin [ 37 ]. There is no other protein currently available with both sweet taste and taste modifying abilities [ 38 ]. The taste modifying activity of the protein (discussed below) remains unchanged when it is incubated at 50°C for 1 hr between pH 3 and 11 [ 39 ]. The molecular weight of Curculin was determined by low angle laser light scattering and was found to be 27800 [ 38 ]. Its three-dimensional model has been built from the X-ray coordinates of GNA, a mannose-binding lectin from snowdrop ( Galanthus nivalis ) [ 38 ]. The three mannose-binding sites present in GNA were found in curculin but were not functional. Some well exposed regions on the surface of the three-dimensional model of the said protein could act as epitopes responsible for the sweet-tasting properties of the protein [ 40 ]. The protein can be crystallized by the vapor diffusion method using polyethylene glycol 400 as a precipitant. The crystals belong to orthorhombic space group P2(1)2(1)2(1) with unit cell dimensions: a = 105 Å, b = 271 Å, c = 48.7 Å. The crystals diffract X-rays to resolution of 3.0 Å and are suitable for X-ray crystallographic studies [ 41 ]. Mabinlin Mabinlin is a sweet protein with the highest known thermostablility [ 42 ]. It is derived from Capparis masaikai and its sweetness was estimated to be around 400 times that of sucrose on weight basis. It consists of an A chain with 33 amino acid residues and a B chain composed of 72 residues. The B chain contains two intramolecular disulfide bonds and is connected to the A chain through two intermolecular disulfide bridges [ 43 ]. Its heat stability is due to the presence of these four disulfide bridges [ 44 ]. The sweetness of Mabinlin-2 is unchanged after 48 hr incubation at boiling point [ 17 ]and of Mabinlin-3 and -4 are unchanged after 1 hr at 80°C [ 45 ]. Miraculin Miraculin is a taste-modifying protein that belongs to the class of sweet proteins. It is extracted from Richadella dulcifica an evergreen shrub native of West Africa. The protein is a single polypeptide with 191 amino acid residues [ 46 ]. It modifies the sweet receptor in such a way that it can be stimulated by acid [ 47 ]. Thus, miraculin has the unusual property of modifying sour taste into sweet taste [ 46 ]. Taste-modifying protein modifies the sweet taste receptor on binding and this behavior of these proteins is responsible for modification in taste of sour substance [ 46 , 47 ]. All acids (which are normally sour) taste sweet after consumption of these proteins. The effects of these proteins exist for around half an hour after consumption and intake of any sour substance will therefore taste sweet during this period of time. The taste buds come to there normal state with time. Pentadin Pentadin is a sweet protein extracted from the plant Pentadiplandra brazzeana , a shrub found in tropical forests of a few African countries. Not much information is available about the protein despite its isolation several years ago, in 1989 [ 48 ]. The protein was reported to be around 500 times sweeter then sucrose on a weight basis. It also consists of subunits coupled by disulfide bonds [ 49 ]. Interaction of sweet proteins with their receptor Humans detect taste with taste receptor cells. These are clustered in taste buds. Each taste bud has a pore that opens out to the surface of the tongue enabling molecules and ions taken into the mouth to reach the receptor cells inside. There are five primary taste sensations salty, sour, sweet, bitter and umami. Sweet and umami (the taste of monosodium glutamate) are the main pleasant tastes in humans. T1Rs are mammalian taste receptors that assemble two heteromeric G-protein-coupled receptor complexes T1R1+T1R3, an umami sensor, and T1R2+T1R3, a sweet receptor [ 50 ]. Sweet and taste-modifying proteins interact with the T1R2-T1R3 receptor with a different mechanism compared to small molecular weight compounds [ 51 ]. Recently, it has been shown that the T1R2-T1R3 receptor has many characteristics similar to the mGluR [ 52 ], apart from some minor differences in the active site region. The major work by Kunishima et al. [ 52 ] solving the crystal structure of the N-terminal active site region of the subtype 1 of mGluR both free and complexed with glutamate has helped a lot in understanding the mechanism of interaction between ligand and T1R2-T1R3 receptor. Their structural work on mGluR and its N-terminal domain [ 52 , 53 ] showing considerable conformational change induced by the glutamate complexation. The 'Active' and 'resting' conformations of m1-LBR, an extracellular ligand binding region of mGluR, is modulated by the dimer interface. The protomer can form 'open' or 'closed' confirmations and is made up of two domains namely LB1 and LB2. The population of active conformers depends on ligand binding, i.e. the so called 'closed-open_A'. The ligand-free receptor exists as two different structures, free form I (open-open_R), the 'resting' conformation with two open protomers and free form II (closed-open_A), nearly identical to the complexed form (Figure 1 , references 52, 54). The mechanism suggested by these structures is that the receptor is in dynamic equilibrium, and that ligand binding stabilizes the 'active' dimer. There are thus two ways, in principle, to activate the receptor: first, to complexate form I with the proper ligand (glutamate for the mGluR, aspartame or any other small molecular weight sweetener for the T1R2-T1R3 receptor) and second, by shift the equilibrium between free form I and free form II in favor of free form II. The exact mechanism of interaction of sweet proteins with the T1R2-T1R3 sweet taste receptor has not yet been elucidated [ 51 ]. Low molecular mass sweeteners and sweet proteins interact with the same receptor, the human T1R2-T1R3 receptor[ 52 ]. Studies have shown that the T1R3 receptor protein is encoded by the Tas1r3 gene involved in transduction of sweet taste [ 55 ]. Recently it has been found that T1R3-independent sweet- and umami-responsive receptors and/or pathways also exist in taste cells [ 56 ]. Conclusion and scope of further work As it has been found that sweet proteins are thousands of times sweeter than sucrose and are of low calorie value, these proteins can be used as natural low calorie sweeteners by people suffering from diseases linked to consumption of sugar e.g. obesity, diabetes and hyperlipemia. Candidate proteins can be checked for biological activity with the human taste receptor. Also mutations can be induced in candidate sweet proteins to analyze changes in their physical, chemical and biological properties. The work can be taken forward by solving the structures of the proteins and taste receptors with a view to increasing the efficiency of these sweeteners.	/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC549512.xml
535803	Seroprevalence of simian immunodeficiency virus in wild and captive born Sykes' monkeys (Cercopithecus mitis) in Kenya	Background The Sykes' monkey and related forms ( Cercopithecus mitis ) make up an abundant, widespread and morphologically diverse species complex in eastern Africa that naturally harbors a distinct simian immunodeficiency virus (SIVsyk). We carried out a retrospective serological survey of SIV infection from both wild and captive Sykes' monkeys from Kenya. We compared two commercially available, cross-reactive ELISA tests using HIV antigens with a novel SIVsyk antigen-specific Western blot assay and analyzed the data by origin, subspecies, age and sex. Results The SIVsyk antigen-specific Western blot assay detected more serum samples as positive than either of the cross-reactive ELISA assays. Using this assay, we found that seroprevalence is higher than previously reported, but extremely variable in wild populations (from 0.0 to 90.9%). Females were infected more often than males in both wild and captive populations. Seropositive infants were common. However, no seropositive juveniles were identified. Conclusion We have developed a specific and sensitive Western blot assay for anti-SIVsyk antibody detection. Sykes' monkeys are commonly infected with SIVsyk, but with extremely variable prevalence in the wild. Higher infection prevalence in females suggests predominantly sexual transmission. High infection prevalence in infants, but none in juveniles, suggests maternal antibodies, but little or no vertical transmission.	Background Both human immunodeficiency virus-1 (HIV-1) and human immunodeficiency virus-2 (HIV-2) have been evolutionarily linked to zoonotic transmissions of naturally occurring simian immunodeficiency viruses (SIVs) from Central African chimpanzees ( Pan troglodytes troglodytes ) and West African sooty mangabey monkeys ( Cercocebus atys ), respectively [ 1 , 2 ]. In addition, the SIVsmm from sooty mangabeys has been transmitted in captivity to Asian macaques ( Macaca mulatta ) and is pathogenic in that species, producing an acquired immunodeficiency syndrome-like condition [ 3 - 6 ]. Evidence of infection by naturally occurring SIVs has been documented in a variety of different non-human African primate species [ 7 - 9 ] and additional types of SIV are likely to exist in other African primates as well [ 9 ]. Despite the importance of naturally occurring SIVs as reservoirs for human disease as well as infections in other primate species, few studies have addressed the distribution or prevalence of these viruses in wild host populations [ 10 ]. Research in this area has been limited by the fact that many primate species are endangered in the wild and therefore samples used for SIV studies have often been acquired from research centers and zoos, not wild animals. The patterns of transmission in captive animals may have been influenced by altered contacts or behaviors in captivity and may not reflect natural transmission patterns. Furthermore, not all SIVs grow well in cell culture and it has consequently been difficult to generate specific reagents for their study. The Sykes' monkey ( Cercopithecus mitis ) belongs to an abundant, widespread and morphologically diverse species complex found throughout East Africa [ 11 - 13 ]. Previous investigations have indicated that these monkeys are infected with a unique virus: SIVsyk. Both wild and captive C. mitis have previously been shown to have antibodies that cross-reacted with SIVsmm and SIVagm [ 14 ]. A virus isolate (SIVsyk173) was made from one SIVsyk infected animal and an infectious clone was produced, sequenced and shown to form a distinct phylogenetic group from other known SIVs [ 15 ]. Very recently, several additional, related SIVsyk sequences have also been reported [ 16 ]. These viruses apparently do not grow well in human or rhesus macaque PBMCs, but have been shown to replicate to some degree in immortalized human lymphoid cell lines [ 17 ]. SIVsyk also does not produce detectable disease in either the natural host or in rhesus macaques [ 15 , 17 ] (personal observation M. G. Otsyula), consistent with other observations that infection of many natural host species with their specific SIV is non-pathogenic [ 15 ]. Here we report the results of a retrospective serosurvey for the prevalence of SIVsyk in wild Sykes' monkey populations in comparison to captive animals. We compare the use of two different cross-reactive ELISA assays with a novel SIVsyk-specific Western blot assay based on the isolate SIVsyk173. We provide a breakdown of seroprevalence results between captive and geographically isolated wild populations, subspecies, sexes and age groups. These data support the utility of using an antigen-specific assay rather than a cross-reactive assay and indicate that the prevalence of SIVsyk infection is highly variable in wild populations and higher in adult animals than previously reported. Results Comparison of SIVsyk Western blot assay to cross-reactive ELISAs Two hundred and seventy six serum samples from both wild and captive C. mitis were screened for the presence of antibodies cross-reactive to HIV-1 and HIV-2 using two commercial ELISA tests. The 276 samples included 100 sequential (duplicate or triplicate) serum samples from the same animals taken at different times. From these sera, the Innotest HIV-1/HIV-2 antibodies test (Innogenetics, Belgium) detected a total of 132 positive samples and 144 negative samples. The Genescreen HIV-1/2 version 2 test (Biorad, Japan). detected 129 positive samples and 147 negative samples. Five samples gave disparate results using these two tests, four of these samples were positive using the Innotest kit, but negative using the Genscreen test and one sample was positive using the Genscreen test, but negative using the Innotest kit. To confirm these results and compare the sensitivity and specificity of these assays, all 276 samples were screened by an SIVsyk viral antigen-specific Western blot assay. Serum samples were considered positive by SIVsyk Western blot assay if they showed reactivity to ENV, GAG or CA proteins. The SIVsyk Western assay identified a total of 138 samples as positive and 138 as negative. Ten samples that were negative by the Genscreen assay and seven samples that were negative by the Innotest assay were identified as positive by the SIVsyk Western assay (Figure 1 ). One sample that was identified as positive by both ELISA tests was consistently negative by SIVsyk Western. Using the SIVsyk Western assay as the standard, this results in a 99.3% specificity for both ELISA assays and 93.9% and 95.8% sensitivity for the Genscreen and Innotest assays, respectively, for detection of anti-SIVsyk serum antibodies. Using McNemar's repeated measures test, the difference between the Western and the Innotest is just under the 95% confidence limit (p = 0.0703) and the difference between the Western and the Genescreen test is statistically significant (p = 0.0117). Figure 1 Representative SIVsyk Western blot results. Blank (B), Negative serum sample (-), Positive serum sample (+). Lanes 1–10, samples negative by at least one ELISA test, but positive for ENV, GAG or CA by Western. Lane 11, sample positive by both ELISA tests, but negative by Western. Serosurvey of wild animals The results from the SIVsyk Western assay were analyzed according to geographic origin, subspecies, age and sex. Figure 2 shows a map of Kenya indicating the locations of the wild collection sites as well as the approximate ranges of the three Sykes' monkey subspecies. Sequential serum samples were excluded from this analysis, reducing the number of samples considered to 176 (109 wild and 67 captive). Of the 109 individual samples from wild populations, 51 positives were detected, for a total wild seroprevalence of 46.8%. This includes both sexes and all age classes from all locations. Between individual wild populations, the seroprevalence level was quite variable, with a high of 90.9% on the coast at Lamu, to 0% at Kwale, also on the coast. Of wild populations, 25 of 38 adult highland Sykes', 11 of 13 adult lowland Sykes' and 1 of 2 adult blue Sykes' were positive, yielding prevalences of 65.8%, 84.6% and 50.0%, respectively for adults of each subspecies. Among the wild born adult animals, 17 of 26 (65.4%) males and 20 of 27 (74.1%) females were positive. Although 1 of 2 (50.0%) wild born infant males and 1 of 1 (100.0%) wild born infant females were positive, 0 of 3 (0.0%) wild born juvenile males were positive. No wild female juveniles were available for testing. Table 1 provides a list of the seroprevalence for all samples from each wild collection site, as well as a list of prevalence among adults only at each site. Tables 2 and 3 provide a breakdown of wild animal seroprevalence according to subspecies, and age and sex, respectively. A Fisher's exact (2 × 2) contingency test for sex-related differences in the total wild population yielded a p value of 0.559, indicating that sex-related differences in the wild are not statistically significant. An exact 3 × 2 contingency test for age-related differences yielded a p value of 0.069, indicating that age-related differences in the wild population are just slightly not statistically significant at the 95% confidence level. A chi squared 8 × 2 contingency test for differences between the eight wild populations yielded a p value of 0.00099, indicating that the differences between populations was highly significant, although the small size of some of the samples violates the assumptions of this test. Repeating this test using only adult animals is also statistically significant (p = 0.0293), however, the further decrease in sample size using only the data from adult animals further violates the assumptions of this test. Figure 2 Map of Kenya, East Africa. Wild population locations and approximate subspecies ranges are shown. Table 1 Anti-SIVsyk serum antibodies among wild Sykes' monkeys according to location. Data for all age groups as well as for adults only in each population is provided. Serum samples from animals with no recorded location are listed as unknown. Location Total Positive/Tested Total Percent positive Adults Positive/Tested Adults Percent positive Aberdares 5/7 71.4% 4/6 66.7% Karen 1/4 25.0% 0/3 0.0% Kibwesi 1/7 14.3% 1/2 50.0% Kwale 0/9 0.0% 0/1 0.0% Lamu 10/11 90.9% 9/9 100.0% N. Kinaangop 3/7 42.9% 0/0 - Ololua 20/34 58.8% 17/23 73.9% Thika 3/4 75.0% 3/4 75.0% Unknown 8/26 30.8% 3/5 60.0% Total 51/109 46.8% 37/53 69.8% Table 2 Anti-SIVsyk serum antibodies among wild and captive born Sykes' monkeys according to subspecies. Serum samples from wild animals of unrecorded subspecies are listed as unknown. Colony animals of either mixed breeding or unknown origin are listed as unknown. Wild born animals Captive born animals Subspecies Positive/Tested Percent positive Positive/Tested Percent positive Highland 25/38 65.8% 9/17 52.9% Lowland 11/13 84.6% 3/4 75.0% Blue 1/2 50.0% 0/0 - Unknown 14/56 25.0% 12/46 26.1% Total 51/109 46.8% 24/67 35.8% Table 3 Anti-SIVsyk serum antibodies among wild and captive born Sykes' monkeys according to age and sex. Serum samples from wild animals of undetermined age or sex are listed as unknown. Wild born animals Captive born animals Age Sex Pos./Tested Percent pos. Pos./Tested Percent pos. Adult Male 17/26 65.4% 4/8 50.0% Female 20/27 74.1% 9/13 69.2% Subtotal 37/53 69.8% 13/21 61.9% Juvenile Male 0/3 0.0% 0/6 0.0% Female 0/0 - 0/5 0.0% Subtotal 0/3 0.0% 0/11 0.0% Infant Male 1/2 50.0% 4/16 25.0% Female 1/1 100% 7/19 36.8% Subtotal 2/3 66.7% 11/35 31.4% Unknown Unknown 12/50 24.0% 0/0 - Total 51/109 46.8% 24/67 35.8% Serosurvey of captive animals Sixty seven individual serum samples were available from captive animals. The overall seroprevalence in captivity was 35.8% (24/67). This includes both sexes and all age classes. Ten of 17 captive adult highland Sykes' and 3 of 4 captive adult lowland Sykes' were positive, yielding prevalences of 59.0% and 75.0%, respectively for these two subspecies. No known captive Blue Sykes' were available for testing. Among captive adult animals, 4 of 8 (50.0%) males and 9 of 13 (69.2%) females were positive. Four of 16 (25.0%) captive infant males and 7 of 19 (36.8%) captive infant females were positive. However, no captive juveniles (0 of 6 (0.0%) males and 0 of 5 (0.0%) females) were positive. Tables 2 and 3 provide a breakdown of captive animal seroprevalence according to subspecies, and age and sex, respectively. A Fisher's exact (2 × 2) contingency test for sex-related differences in the captive population yielded a p value of 0.646, indicating that sex-related differences in captivity are not statistically significant. An exact 3 × 2 contingency test for age-related differences yielded a p value of 0.001, indicating that age-related differences in captivity are statistically significant. Serosurvey comparisons The overall captive seroprevalence was 35.8%, which is slightly lower than the overall wild seroprevalence of 46.8%. Infection prevalence was highest in adults at 69.8% for wild populations and 61.9% in captive populations. Adult females were infected at a higher prevalence than adult males in both the wild (74.1% (20/27) for wild adult females and 65.4% (17/26) for wild adult males) and captive populations (69.2% (9/13) for captive adult females and 50.0% (4/8) for captive adult males). Infants were also found to have a high seroprevalence at 66.7% (2/3) from wild populations and 31.4% (11/35) from captivity, although few wild infants were surveyed. No seropositive juveniles were identified from either wild (0/3) or captive (0/11) populations. A Fisher's exact (2 × 2) contingency test for sex-related differences in the combined wild and captive populations yielded a p value of 0.455. This indicates that while the there is a clear trend in each data set showing higher prevalence among females, the small numbers of samples do not show a statistically significant difference between infection in males and females. An exact 3 × 2 contingency test for age-related differences in the combined wild and captive populations yielded a p value of 0.00001, indicating that age-related differences in the combined populations are highly statistically significant. Discussion Previous SIV serosurveys of Cercopithecus mitis have used assays relying on cross-reactivity between antibodies against SIVsyk and other HIVs or SIVs. To assess the relative effectiveness of the use of cross-reactive tests, we compared serological results using two commercially available HIV-1/HIV-2 ELISA kits with an authentic SIVsyk antigen Western blot assay. The cross-reactive ELISA assays were very specific and in only one out of 276 samples gave a false positive result compared to the SIVsyk antigen Western assay. Curiously, this one particular captive juvenile female was repeatedly positive by both ELISAs but repeatedly negative by Western. This animal may possibly have been infected in captivity with a different SIV that produces cross-reactive antibodies to HIV-1 and HIV-2, but not to SIVsyk. On the other hand, the cross-reactive ELISAs failed to detect between 4.2 to 6.1% of the animals that were positive by the SIVsyk antigen Western assay. This result may be due to increased sensitivity using a Western blot format, but is also possibly due to increased sensitivity to the authentic SIVsyk viral antigens. Variable SIV infection prevalences have previously been reported in this species. However, information on age, sex and captive or wild status was not provided in any previous survey. Lowenstine et al, 1986 [ 18 ] found no evidence of infection in two captive Sykes' monkeys at US zoos using an HIV-1 antigen based ELISA. Ohta, et al, 1988 [ 19 ], found 4 out of 18 (22%) Sykes' monkeys from East Africa with evidence of infection, presumably by Western using SIVagm antigen, but no further details were given. Three other reports specifically tested Sykes' monkeys at the Institute for Primate Research in Kenya, but it is unknown which of the samples tested in this survey are from the same animals. Emau, et al, 1991 [ 17 ], reported 59 out of 100 (59%) positive using a radioimmuno-precipitation assay cross-reactive with SIVsmm and SIVagm. Tomonaga, et al, 1993 [ 20 ], reported 9 out of 35 (25%) lowland Sykes', 13 out of 24 (54%) highland Sykes', and 1 out of 14 (7%) blue Sykes' positive using an immunofluorescence assay with SIVagm antigen. Otsyula et al, 1996 [ 14 ], reported 12 out of 35 (34%) positive by Western blot using SIVagm antigen and 8 out of 35 (23%) positive by Western blot using SIVsmm antigen. The low incidence of infection reported by Otsyula, et al 1996 [ 14 ] using Western blots with SIVagm as well as SIVsmm antigens supports the hypothesis that the increased sensitivity in the present study is due to the use of SIVsyk antigens and is not likely to be simply due to the use of Western blot assay vs. ELISA. To avoid age-related bias, a comparison between captive and wild populations should be made using only adult animals because the captive population contained many more samples from infants and juveniles than did the wild population (see Table 3 ). Both captive and wild adults showed a much higher prevalence than previous reports. Just over 60% of captive adults and nearly 70% of wild adults were seropositive. However, in wild populations, seroprevalence varied greatly according to location (from 0 to 90.9%) and this variability was statistically significant. The differences in seroprevalence may be due to small sample sizes, however, the average troop size in this species is approximately 10 animals [ 12 ] and we have sampled between 4 and 34 animals from each individual site (some sites have more than one troop present). Alternatively, host immunological parameters, genetic differences in virus strains, troop size, individual or troop behavior, and habitat may all affect the seroprevalence of these viruses among certain populations. Although the small sample sizes did not provide statistical significance, in both captive and wild animals, a higher infection prevalence was seen in females than in males. This possibly indicates a predominantly sexual transmission route, rather than transmission by aggressive behavior between males as has been shown in mandrills [ 21 ]. Sykes' monkeys usually live in a harem arrangement where a dominant male controls sexual access to the females in the troop. This may explain the variability in seroprevalence among wild populations, since the infection status of a dominant male would heavily influence the status of the entire troop. However, other breeding strategies exist in this species and influxes of multiple solitary males into troops have been documented during periods when multiple females are sexually receptive [ 22 ]. Unfortunately, we have no data on the breeding situations in the sampled groups or which samples came from dominant versus lower status males. In addition to the adults, a high seroprevalence was also observed in infants. However, no seropositive juveniles were identified, even though 14 juvenile animals were sampled from both wild and captive populations. A possible explanation is that infants are seropositive due to the presence of maternal antibodies in breast milk, but little or no actual maternal-fetal transmission occurs either in utero , during birth or through breastfeeding. An alternative possibility to consider is that SIVsyk infections could be cleared by infants and that adult seroprevalence is due to later exposure. Conclusions In order to investigate the infection prevalence of SIVsyk in the common and widespread Sykes' monkey in East Africa, we used two commercially available ELISA assays and a novel SIVsyk antigen-specific Western blot assay to perform a retrospective serosurvey of 276 previously collected serum samples from both wild and captive animals. To develop the antigen-specific Western assay, we propagated SIVsyk strain 173 in a human CEMx174 lymphoid cell line expressing human CCR5. Comparisons between the cross-reactive ELISAs and the SIVsyk Western assay indicate that the antigen-specific Western assay is more specific and sensitive, supporting the utility of using antigen-specific assays in this and other SIV serosurveys. We report that using this Western assay, infection prevalence is higher on average than previously reported, but that the prevalence in wild populations is variable. In captivity and in the wild, females are more commonly infected than males, although these differences are not statistically significant. Positive serology is also common in infants, but no positive juveniles were identified. We interpret this as evidence of maternal antibodies infants, but a low incidence of maternal-fetal transmission. Methods Specimens The sera used in this study were obtained between 1981 and 1996 from animals that had been previously trapped in the wild or from previously sampled captive animals housed at the Institute of Primate Research in Nairobi, Kenya. Serum samples were stored continuously at -20°C. All animals were captured, housed and sampled in accordance with humane animal use guidelines established by the Kenya Institute of Primate Research. Wild born monkey samples used in this study were derived from eight locations. In general, these sites were rural or suburban areas that bordered forested areas with natural flowing water. These sites are typical of areas in which humans have encroached on natural habitats and now may come in contact with these monkeys as they forage for food. The geography of the sites varied considerably from heavily forested areas in the highlands of Kenya to drier forests along the lowland coast of Kenya. The range of the lowland Sykes' ( C. mitis kibonotensis Lonnberg) includes coastal and eastern regions of Kenya [ 11 , 13 ]. The range of the highland Sykes' ( C. mitis kolbi Neumann) includes the upland central portion of the country [ 11 , 13 ]. The range of the blue Sykes' ( C. mitis stuhlmannii Matschie) includes western Kenya and the border with Uganda [ 11 , 13 ]. Some specimens from the wild came from unknown locations and these are noted as such. Captive animals used for breeding were housed in an area composed of nine outdoor enclosures. Typically, breeding groups were composed of approximately ten individuals. Each breeding group consisted of a harem of one male and four to six females, plus several infants. Infants were left with mothers for a period of time and were usually moved to another group between the age of six months and two years to prevent inbreeding. Individuals not involved in breeding were kept in groups composed of approximately ten juvenile males and females in larger outdoor enclosures. Animals were aged according to dental inspection and according to sexual maturity. Animals believed capable of mating, usually over two years of age, were considered adults. Infants were less than six months of age. ELISA Analysis Serum samples were tested using the Innotest HIV-1/HIV-2 Antibodies test (Innogenetics, Ghent, Belgium) and Genscreen HIV1/2 version 2 (Biorad, Tokyo, Japan). The Innotest ELISA kit uses synthetic peptides representing HIV-1 subtype M envelope, HIV-2 envelope, and HIV-1 group O envelope antigens. The Genscreen uses purified gp160 envelope and p25 nucleocapsid recombinant protein of HIV-1 and a peptide that mimics the immunodominant epitope of HIV-2 envelope. Each kit was used according to the manufacturers' recommendations. Photometric measurements were performed with a Dynatech MRX plate reader at 450 and 650 nm wavelengths. Positive and negative determinations were based upon cutoff values as determined by the manufacturers' protocols. Cells and Virus CD4+ CEMx174 cells were transduced with a CCR5-encoding vector (pBABE-puro-CCR5, a gift of David Dorsky, University of Connecticut) as previously described [ 23 ]. The resulting CCR5 expressing cell line, designated CEM-R5, was maintained in RPMI 1640 with 10% FBS and 1 ug/ml puromycin. The SIVsyk-1.2 isolate (obtained from the AIDS Reference and Reagents Program) was inoculated into CEM-R5 and the CEMx174 parent cell lines. After 7–14 days in culture, widespread syncytia formation was observed in CEM-R5 cells while few syncytia were observed in the parental CEMx174 cells. Cell associated viral proteins were extracted from SIVsyk-1.2 infected CEM-R5 cells grown for 7–10 days post-inoculation. Western Blot Analysis Serum samples were screened by Western blot using SIVsyk-specific viral antigens. SIVsyk-1.2 infected cells were centrifuged for 5 min at 1,000 × g and resuspended in lysis buffer containing 1% triton, 50 mM tris pH 7.5, 2 mM EDTA. Cell debris was pelleted for 2 min. at 12,000 × g and the supernatant was added to loading buffer containing 50 mM Tris-HCl pH 6.8, 100 mM DTT, 2% SDS, 10% glycerol, 0.1% bromophenol blue. Viral antigens were separated on a 12% polyacrylamide gel and subsequently transferred to a PVDF membrane. Membrane strips were blocked with a solution of PBS-T containing 5% dry milk. Serum samples were diluted 1:500 and incubated for one hour at room temperature with individual membrane strips. Primary serum antibodies were detected with a 1:4,000 dilution of goat-anti-rhesus HRP conjugated secondary antibody (Nordic Immunology, Tilburg, Netherlands). Colorimetric detection was performed using the Immunopure Metal Enhanced DAB Substrate Kit (Pierce Endogen, USA). Competing interests The author(s) declare that the have no competing interests. Authors' contributions BRE and EM conducted the ELISA and Western blot experiments and BRE wrote the draft manuscript. DE and JR cultured the SIVsyk virus. MGO provided serum samples and oversaw the ELISA experiments. SFM assisted with and oversaw the Western blot experiments and prepared the final manuscript. All authors read and approved the final manuscript.	/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC535803.xml
503396	Resident interest and factors involved in entering a pediatric pulmonary fellowship	Background Relatively little is known about interest in pediatric pulmonology among pediatric residents. The purpose of this study, therefore, was to determine at this institution: 1) the level of pediatric resident interest in pursuing a pulmonary fellowship, 2) potential factors involved in development of such interest, 3) whether the presence of a pulmonary fellowship program affects such interest. Methods A questionnaire was distributed to all 52 pediatric residents at this institution in 1992 and to all 59 pediatric residents and 14 combined internal medicine/pediatrics residents in 2002, following development of a pulmonary fellowship program. Results Response rates were 79% in 1992 and 86% in 2002. Eight of the 43 responders in 1992 (19%) had considered doing a pulmonary fellowship compared to 7 of 63 (11%) in 2002. The highest ranked factors given by the residents who had considered a fellowship included wanting to continue one's education after residency, enjoying caring for pulmonary patients, and liking pulmonary physiology and the pulmonary faculty. Major factors listed by residents who had not considered a pulmonary fellowship included not enjoying the tracheostomy/ventilator population and chronic pulmonary patients in general, and a desire to enter general pediatrics or another fellowship. Most residents during both survey periods believed that they would be in non-academic or academic general pediatrics in 5 years. Only 1 of the 106 responding residents (~1%) anticipated becoming a pediatric pulmonologist. Conclusions Although many pediatric residents consider enrolling in a pulmonary fellowship (~10–20% here), few (~1% here) will actually pursue a career in pediatric pulmonology. The presence of a pulmonary fellowship program did not significantly alter resident interest, though other confounding factors may be involved.	Background The specialty of pediatric pulmonology is relatively new, having been recognized as a pediatric sub-specialty by the American Board of Medical Sub-specialties in 1984. In 1997, there were approximately 500 board certified pediatric pulmonologists in the United States and Canada [ 1 ]. This number has increased in recent years with 708 board certified pulmonologists being identified in 2003 [ 2 ]. It has been estimated that there is one pediatric pulmonologist for every 280,000 children in the United States [ 1 ]. There are over 50 pediatric pulmonary fellowship programs in North America with approximately 30–35 fellows graduating each year. The demand for pediatric pulmonologists has increased during the past decade, with many academic centers looking for two or more pulmonologists simultaneously [ 3 ]. A recent national survey of medical directors at children's hospitals across the country found that vacancy rates for faculty in pediatric pulmonology (25 of 136 positions vacant = 18.4%) was ranked second highest, behind only pediatric endocrinology, among over 40 pediatric subspecialties [ 3 ]. Despite this demand, little is known about overall interest in this field among pediatric residents. For this reason, this study was completed to examine interest among pediatric residents in entering a pulmonary fellowship. The specific aim of this study, therefore, was to determine at our institution: 1) the level of pediatric resident interest in pursuing a pulmonary fellowship; 2) potential factors involved in development of such interest; 3) whether the presence of an active fellowship program affected resident interest in such a program. Methods This study involved the distribution of a questionnaire to all pediatric residents. The questionnaire was initially distributed in 1992 prior to institution of a pulmonary fellowship. The questionnaire was placed in the hospital mailbox of each resident. The study was repeated (and the questionnaire redistributed) in 2002 after the fellowship, which began in 1994, had been functioning for several years. To improve the response rate, the questionnaire was distributed twice, one month apart, during each time period. The questionnaire was a three-page, 18 question form that took approximately 15 minutes to complete (copy enclosed in Appendix [see Additional file 1 ]). The questionnaire asked several epidemiological questions (e.g. year of residency, whether medical school was attended at this institution), several questions that pertained to an individual resident's interest in entering any specialty fellowship, and approximately 12 questions specifically dealing with interest in doing a pulmonary fellowship and factors that may be either positively or negatively related to such interest. The survey utilized a flow diagram, and consequently, slightly different questions were asked depending on a resident's interest or lack of interest in a pulmonary fellowship. The study was analyzed after dividing the respondents into 2 groups: those that had considered a pulmonary fellowship during their residency ("+PF") and those that had not considered a pulmonary fellowship during their residency ("-PF"). Neither Human Subjects approval nor specific resident consent was obtained. However, all questionnaires were answered anonymously and residents were in no way coerced or forced to complete the survey. Residents who completed the form were given a certificate for a free meal at Children's Hospital of Wisconsin. Several questions were based on a 4-point Likert scale and ranked from 0 ("not at all important") to 3 ("very important") in a resident's decision to consider or not consider doing a pulmonary fellowship. Numerical responses for each survey period were averaged (mean ± SD) and ranked. Fisher's exact test was used to compare categorical variables and the Student's t-test was used to compare the means of measured variables in 2 independent samples. A p value of ≤ 0.05 was considered significant. Medical Education records were also reviewed anonymously (data collated by year of residency) in 2003 to determine residents' actual career decisions. Results The questionnaire was distributed to 52 pediatric residents (including 2 chief residents) in 1992 and to 59 pediatric residents (including 3 chief residents) and 14 medicine/pediatric residents (including 1 chief resident) in 2002. A medicine-pediatric residency program did not exist during the initial distribution period. To avoid compromising confidentiality in the relatively small medicine/pediatric group, residents were not asked to list their residency program in 2002 and consequently the 2 resident groups during that period were combined. Forty-three of the 52 residents completed the survey in 1992 (79%) compared to 63 of 73 (86%) during 2002 (p = NS). Of the 43 respondents in 1992, 30 (70%) had considered doing a fellowship in any pediatric subspecialty and of those, 15 (35% of all respondents) believed they were "very likely" to do a fellowship. Of the 63 respondents in 2002, 40 (63%) had considered doing a pediatric fellowship and 16 of those (25% of all respondents) were "very likely" to continue with fellowship training. These numbers compare with 9 of the 52 residents in 1992 (17%) who actually completed fellowship training compared to 16 of the 47 graduating residents from the 2002 survey (34%) who began fellowship training in 2003 or 2004. Eight residents (19%) in 1992 had considered a pulmonary fellowship compared to 7 residents (11%) in 2002 (p = NS). Table 1 lists these responses based on survey period and year of residency (internal medicine-pediatric residency is 4 years long). Although all residents have significant exposure to pulmonary patients during their residency, no correlation existed between interest in a pulmonary fellowship and having previously taken a one-month elective in pediatric pulmonology during residency. Three of the 15 residents who had considered a pulmonary fellowship ("+PF") had taken a pulmonary elective compared to 7 of the 91 residents who had not considered such a fellowship ("-PF"; p = 0.15). Of the 10 residents who had taken a pulmonary elective, only 2 thought that the elective experience affected their decision in considering a pulmonary fellowship (one from the +PF and one from the -PF group). Table 1 Resident response based on year of survey (1992 or 2002) and year of residency (1–5) 1992: Residency Year Resident Number Number Responded % Response +PF* -PF** 1 21 20 95% 5 15 2 19 12 63% 2 10 3 10 9 90% 0 9 4 2 2 100% 1 1 1992 totals 52 43 79% 8 35 2002: Residency Year Resident Number Number Responded % Response +PF* -PF** 1 24 22 92% 3 19 2 21 20 95% 3 17 3 22 17 77% 1 16 4 5 3 60% 0 3 5 1 1 100% 0 1 2002 totals 73 63 86% 7 56 * +PF refers to those residents who had considered doing a pulmonary fellowship during their residency. ** -PF refers to those residents who had not considered doing a pulmonary fellowship during their residency. Table 2 lists and ranks the reasons given by residents who had considered doing a pulmonary fellowship. The highest ranked factors given by +PF residents included wanting to continue one's education after residency, enjoying caring for pulmonary patients in general, and enjoying both pulmonary physiology and the pulmonary faculty. The only factor that approached a statistically significant difference for +PF between the 2 survey periods was the statement "I enjoy the tracheostomy-ventilator population," with a score of 0.6 ± 0.5 in 1992 vs. 1.6 ± 1.1 in 2002 (p = 0.057), suggesting that this factor became somewhat more important to the 2002 +PF residents in their consideration of a pulmonary fellowship. Table 3 lists and ranks reasons given by residents who had not considered doing a pulmonary fellowship. The scores given by -PF residents were generally not as high as those given by +PF residents, i.e. the factors listed were not as important to the -PF vs. the +PF residents. Major factors listed by -PF residents included not enjoying the tracheostomy/ventilator population and certain pulmonary patients in general, including chronic patients, as well as a desire to enter general pediatrics or another fellowship. The significant differences for -PF between the 2 survey periods are listed in Table 3 . Table 2 Mean (± SD) scores for factors given by +PF residents (those who considered a pulmonary fellowship) Factor* 1992 Score 1992 Rank 2002 Score 2002 Rank I want to continue my education after residency 2.8 ± 0.4 1 2.4 ± 0.8 1 I enjoy pulmonary-related procedures 2.5 ± 0.8 2 2.0 ± 1.0 10 I enjoy caring for pulmonary patients 2.4 ± 0.5 3 2.4 ± 0.5 1 I like pulmonary physiology 2.4 ± 0.8 3 2.3 ± 0.8 4 I like the pulmonary faculty 2.4 ± 0.5 3 2.4 ± 0.8 1 I enjoy pulmonary inpatient coverage 2.3 ± 0.5 6 2.3 ± 0.5 4 I enjoyed working with a particular faculty member 2.1 ± 0.9 7 2.1 ± 0.7 7 I might enjoy pulmonary-related research 2.0 ± 1.2 8 1.4 ± 0.8 12 I enjoy pulmonary clinics 2.0 ± 1.0 8 2.2 ± 0.8 6 I enjoy cystic fibrosis patients 1.9 ± 1.1 10 2.1 ± 1.1 7 I enjoyed caring for a particular pulmonary patient 1.7 ± 1.3 11 2.1 ± 0.7 7 The salary would be attractive 1.1 ± 1.2 12 1.0 ± 0.8 13 I enjoy the tracheostomy/ventilator population 0.6 ± 0.5 13 1.6 ± 1.1# 11 Each factor was scored from 0 ("not at all important") to 3 ("very important") relating to a resident's decision to consider doing a pulmonary fellowship. #2002 score approached statistically significant difference vs. 1992 score (p = 0.057). Table 3 Mean (± SD) scores for factors given by -PF residents (those who did not consider a pulmonary fellowship) Factor 1992 Score 1992 Rank 2002 Score 2002 Rank I don't enjoy the trach/vent population 2.0 ± 1.1 1 1.9 ± 1.3 1 I want to enter general pediatrics 1.9 ± 1.2 2 1.8 ± 1.3 2 I don't enjoy certain pulmonary patients 1.6 ± 1.1 3 1.2 ± 1.1 4 There are too many chronic patients 1.5 ± 1.1 4 1.5 ± 1.1 3 I want to enter another fellowship 1.5 ± 1.2 4 1.2 ± 1.3 4 I don't know enough about it to decide 1.5 ± 1.1 4 1.0 ± 1.1 7 Not enough pulmonary patient experience 1.4 ± 1.1 7 0.6 ± 0.9# 9 I don't enjoy the BPD + population 1.2 ± 1.0 8 1.2 ± 1.1 4 I don't think pulmonary is very interesting 0.9 ± 0.9 9 0.6 ± 0.8 9 Some of the pulmonary patients scare me 0.7 ± 0.8 10 0.8 ± 1.0 8 I can't afford being a fellow 0.7 ± 1.1 10 0.4 ± 0.8 12 The pulmonologists work too hard 0.4 ± 0.6 12 0.4 ± 0.7 12 I don't enjoy the cystic fibrosis population 0.3 ± 0.6 13 0.6 ± 0.9 9 Too few pulmonary job openings 0.3 ± 0.5 13 0.0 ± 0.2# 17 I don't enjoy the asthma population 0.2 ± 0.5 15 0.4 ± 0.8 12 Pulmonologists don't earn enough money 0.2 ± 0.5 15 0.1 ± 0.4 16 Poor experiences with pulmonary faculty 0.2 ± 0.4 15 0.3 ± 0.8 15 Each factor was scored from 0 ("not at all important") to 3 ("very important") relating to a resident's decision to not consider doing a pulmonary fellowship. #Indicates 2002 score significantly different than 1992 score, p < 0.001. + bronchopulmonary dysplasia The last question in the survey asked residents what they thought they would be doing 5 years in the future. These responses are shown in Table 4 . The majority of residents during both survey periods believed that they would be in either non-academic or academic general pediatrics in 5 years. However, when grouped together during the 2 periods, the +PF residents were less likely to see themselves in the future as general pediatricians compared to the -PF residents (p < 0.05). Interestingly, when the actual career decisions of the 1992 residents were reviewed, 35 of the 52 residents (67%) went into non-academic general pediatrics, 5 (10%) entered academic general pediatrics, 9 (17%) completed pediatric subspecialty training, 2 (4%) began a non-pediatric medical specialty, and in 1 case (2%) the eventual career decision could not be determined. Only 1 resident (from 1992 survey) in the entire group of 106 survey responders (~1%) believed they would be in the field of pediatric pulmonology in the future. This resident did complete a pulmonary fellowship at another institution and is currently in an academic pulmonary practice. Additionally, a second resident (from 1992 survey) is currently an academic pediatric pulmonologist after having initially completed one year of a different subspecialty fellowship following residency. That resident then completed a pulmonary fellowship at this institution. Lastly, there was no significant difference in the overall level of interest in a pulmonary fellowship from 1992 compared to 2002. Table 4 Anticipated professional plans of resident 5 years following survey completion Category 1992 2002 +PF -PF +PF -PF General pediatrics, non-academic 1 14.5 2 29 Academic general pediatrics 1 5 1 8.5 Academic non-pulmonary pediatric specialty 3 12 2 16 Academic pediatric pulmonology 1 0 0 0 Non-academic non-pulmonary pediatric specialty 1 1.5 1 1.5 Non-academic pediatric pulmonology 0 0 0 0 Non-pediatric medical specialty 0 0 0 1 Non-medical vocation 0 0 0 0 Unknown 1 2 1 0 See text and footnote to Table 1 for explanation of +PF and -PF. The numbers represent actual number of residents responding. *Some residents listed 2 categories, hence their score was divided between them. Discussion This study found that a significant percentage of pediatric residents considered doing a pulmonary fellowship after their residency training, ranging from 11% in the 2002 group to 19% in the 1992 group. However, these residents also viewed themselves as less likely to enter general pediatrics perhaps suggesting that they were simply considering several pediatric subspecialties at some time during their residency training. This seems likely, as the highest scored factor by the +PF residents was the desire to continue their education after residency. Despite fairly high percentages of residents considering a pulmonary fellowship, only 1 resident in the entire group (~1%) actually believed that they would be a pediatric pulmonologist 5 years after the survey was completed. This percentage is very similar to that of first-time takers of the 1995 General Pediatrics Certifying Examination who believed they would be a pediatric pulmonologist in the future (1.1%) [ 4 ]. If one were to extrapolate this number to the entire class of graduating pediatric residents per year, about 2600 residents, approximately 30 residents would be entering the field of pediatric pulmonology each year [ 5 ]. This number is very similar to the actual number of graduating residents who enter a pediatric pulmonary fellowship each year [ 2 ]. The majority of residents who considered doing a pulmonary fellowship (8 of 15) were in their first year of residency. From personal experience, residents often consider various practice options early on in their training and frequently do not tend to narrow their choices until their second or third year of residency. This observation should be kept in mind when trying to recruit residents for pulmonary fellowship positions by seeking out those residents potentially interested in Pulmonology early on in residency rather than later. The ranking of factors that may contribute to an interest in a pulmonary fellowship were remarkably similar during the 2 time periods. The only score that approached a statistically significant difference was the statement "I enjoy the tracheostomy/ventilator population" and this score tended to increase in 2002. However, a dislike for the tracheostomy/ventilator population also received the highest score among those residents not interested in a pulmonary fellowship and was even higher than both the desire to enter general pediatrics and another fellowship program. These data may simply be a center phenomenon but might suggest a more global "disinterest" in this patient population that may need to be further studied. Two scores in the -PF resident group decreased from 1992 to 2002: not enough pulmonary patient experience to decide on a pulmonary fellowship and too few perceived pulmonary job openings. The first may be a positive reflection on the local resident experience with pulmonary patients in recent years or on the fellowship program itself though this is only speculative. The second received very low scores during both periods and may not be very relevant. However, there has been some evidence of a reversal in the prior trend of residents entering general pediatrics in recent years, suggesting greater interest among residents in fellowships [ 6 ]. On the other hand, a recent survey of practicing pediatric pulmonologists noted that 69% of respondents did not believe that there was need for additional pulmonologists in their locale [ 7 ]. Interestingly, despite recent articles relating high post-residency debt with a disinterest in entering a fellowship, this was not noted in this study with mean scores during both survey periods of less than 1.0 for the statement "I can't afford being a fellow." [ 8 , 9 ]. This study did not ask residents to state their current level of indebtedness. Other "job concern" factors including job availability and the workload of pulmonologists did not appear to be significant negative factors for the -PF residents. This study has certain shortcomings. This study asked residents to score specific factors that may or may not have been relevant to an individual resident. Although residents were given the opportunity to add personal comments, few did. In addition, certain patient populations, e.g those with respiratory infections, were not included as options in the survey and these omissions could cause study bias. Other factors that may contribute to residents' decisions regarding fellowship training were not addressed in this study. These factors include resident teaching by the faculty, resident gender, spouse occupation, mentor encouragement or other personal reasons [ 10 - 12 ]. In addition, this study involved residents in only one program and cannot be generalized to all residency programs. Although this study did not find that the presence of a pulmonary fellowship significantly affected resident interest in such a fellowship, a significant difference may have been found with a larger sample size. In fact, the study appears to be underpowered despite the high response rate. More than 250 individuals would need to have been included in the study to detect a significant difference in residents' initial interest in a pulmonary fellowship (assuming a power of 80%). In addition, other changes may have occurred within the residency program including higher loan repayments, changes in department philosophy (e.g. new department chairman), and changes in local pediatric job opportunities, which may have affected the results. Despite these limitations, this study may prove useful to those who are recruiting pediatric residents as potential pulmonary fellows. Further larger studies looking at multiple residency programs may provide more insight in the future. Lastly, studies like these help to reiterate the need for the specialty of pediatric pulmonology to "prospectively and objectively determine realistic future training needs" [ 13 ]. Conclusions Although many pediatric residents consider enrolling in a PF (~10–20% here), few (~1% here) will actually pursue a career in pediatric pulmonology. The presence of a PF program did not significantly alter resident interest, though other confounding factors may be involved. Abbreviations BPD: bronchopulmonary dysplasia PF: pulmonary fellowship +PF: those residents who considered taking a pulmonary fellowship -PF: those residents who had not considered taking a pulmonary fellowship Competing interests None declared. Pre-publication history The pre-publication history for this paper can be accessed here: Supplementary Material Additional File 1 Appendix. Resident questionnaire Click here for file	/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC503396.xml
544963	Healthy children, healthy country: the use of governing instruments in shifting the policy paradigm	The evidence on early childhood strongly suggests the need to shift child health policy from the current focus on social welfare to a socio-ecologically based approach. This paper reviews three governing instruments, exhortation, expenditure and regulation, that have been used by governments in Australia and discusses the relative effectiveness of these approaches in shifting the child health policy paradigm.	The evidence for healthy public policy for children There can be no keener revelation of a society's soul than the way it treats its children [ 1 ]. Research evidence has demonstrated that the experiences of early childhood can have a profound lifelong impact on a child's health, wellbeing and competence [ 2 ]. The importance of the early years of life in influencing future outcomes, such as crime, obesity, heart disease, mental health problems and poor school outcomes has been identified and highlighted [ 3 ]. While there are many factors found to influence rising crime rates, various researchers have identified children with manifested behaviour disorders in early childhood [ 4 ], academic difficulties and non-engagement in schooling [ 5 ], and the quality of neighbourhood supervision and support as contributing factors to criminal behaviour [ 6 ]. Education, literacy and other social determinants of health can influence the coping skills of children, which provide the basis of learning, behaviour and health throughout life [ 7 ]. Poverty, whether measured in absolute or relative terms, has a negative effect on children's health [ 8 , 9 ]. In particular, poverty is associated with developmental delay, poor school achievement and employment futures, behaviour problems, increased incidence of chronic illness, visual and hearing defects and dental problems [ 10 ]. Parental poverty and exposure to unhealthy environments (eg smoking; low levels of literacy; nutrition; emotional support) reduce a child's life chances. Studies in neurobiology, neurodevelopment and early intervention show that the time period from conception to school age is a critically important time for brain development, setting the scene for prevention of some of the identified adverse outcomes through early identification and intervention [ 11 ]. Consistent with the increasing evidence, many governments have identified support in early childhood as a lifelong determinant of health, wellbeing and competence, as a matter for policy development, initiating actions to ensure comprehensive child development strategies for their societies. This approach requires a whole of government response, integrating health, welfare, education and other relevant parts of government. The evidence suggests that healthy public policy for infants, children and their parents is dependent on understanding of the socio-ecological factors supported by integrated, multidisciplinary and intersectoral policy and programs. In Canada, Britain and the United States, targeted interventions in the antenatal period, infancy and childhood, including parenting skills programs, are recognized for their potential to support healthier families. A socio-ecological model of health is increasingly perceived to be the most appropriate approach for the early years of life agenda. Consistent with this approach, the United Kingdom program 'Sure Start' , has been positively reviewed by the UK Audit Office and is considered by many to be the standard for the whole of government approach [ 12 ]. In addition, a recently released ten year plan is attempting to significantly change the way in which children are treated throughout UK systems [ 13 ]. The Canadian experience is widely quoted as best practice [ 2 ], with both federal and provincial investment in early childhood (See, for example [ 14 - 16 ]). In the US, during the Presidents' Summit for America's Future held in April 1997, Presidents Bill Clinton, George Bush, Jimmy Carter and Gerald Ford and First Lady Nancy Reagan stressed the importance of early childhood, calling the nation to action. American policy in this area has built upon influential reports that have led to investment in early childhood in most states [ 12 ]. The adoption of healthy public policy for children based on this socio-ecological framework has been inconsistent throughout Australia. In an attempt to explore these inconsistencies, this paper reviews the use of three governing instruments, that is exhortation, expenditure and regulation, by national and state governments in Australia. Governing instruments are the major mechanisms governments use to seek compliance, support and implementation of public policy. Governing instruments range from minimum coercion by exhortation, through expenditure, taxation, regulation, to maximum coercion through public ownership [ 17 ]. The following sections describe the impact of the use of exhortation, expenditure and regulation on the implementation of healthy child policy. Consensus building – exhortation as the national instrument of choice During the 1990s the Australian Government identified the health of children and young people as a key policy area, with a series of policy documents: • The National Health Goals and Targets for Australian Children and Youth (1992) • The National Health Policy for Children and Young People (1995) and associated Implementation Plan (1996) • The National Health Policy for Young Australians (1997). These documents provided broad national goals for children and young people: • Reducing preventable premature mortality • Reducing the impact of disability • Reducing the incidence of vaccine preventable disease • Reducing the impact of conditions occurring in adulthood with their origins or early manifestation in childhood or adolescence • Enhancing family and social functioning Although the evidence supporting a broader definition of child health was strong, the focus of these National Health Goals and Targets remained heavily focused towards surveillance and the reduction of injury and illness, perhaps reflecting a comfort with current and past approaches. To date there has been little evidence of an integrated, multidisciplinary approach to child health at the national level. The 2003/04 federal budget did not provide the broad whole of government approach recommended for child health, with only a few targeted interventions, such as the National Meningococcal C Campaign, and a much greater focus on the health needs of the aging population. In 2003, the Australian of the year, Fiona Stanley suggested that the social and economic policies of the Government were not effective in tackling the issues associated with ensuring healthy children and young people [ 18 ]. The platform for a paradigm shift was established in 2001 with the appointment of the Minister for Children and Youth Affairs and the subsequent statement in 2002 of the intent to develop a National Agenda for Early Childhood. The consultation paper Towards the Development of a National Agenda for Early Childhood signaled a changing paradigm, with a whole of child and life course approach addressing promotion, prevention and early intervention for all children. The last years have seen the creation of ever more advisory groups, partnerships and inquiries with a mandate to influence child health policy. The Child and Youth Health Intergovernmental Partnership (under the auspice of the Australian Health Ministers' Advisory Council) was convened in December 2001 to develop a national child public health strategy and advise on the National Agenda for Early Childhood. Their draft strategy framework Better Child Public Health: A Strategic Approach to Building Capacity – A National Action Plan 2004–2007 has been developed and is being used in consultation and capacity building initiatives. In October 2002 the Minister for Children and Youth Affairs referred an inquiry into improving children's health and well being to the Standing Committee on Family and Community Affairs. The Australian Council for Children and Parenting (ACCAP), an advisory body to the Minister for Children and Youth Affairs, was granted a two year term from July 2003, with a focus on strategic advice in the areas of early childhood intervention and prevention, parenting and child protection, foster care and emerging early childhood initiatives, including advising about the continuing development of the National Agenda for Early Childhood. As described above, the policy approach at the national level has focused almost entirely on exhortation, the least coercive instrument, where support and compliance are sought voluntarily through persuasion and discussion. In comparison with other countries, such as Britain and Canada, the lack of a common and shared understanding of the socio-ecologic approach and its implications has made it difficult to show any significant advances in this area. In fact, the recent demise of the Child Health Unit within the Australian Government Department of Health suggests less focus on child health. Nationally, child health has not been heavily addressed through other policy instruments, such as expenditure, taxation, or regulation, although more recently the Commonwealth Government Department of Family and Community Services (DFaCS) established the 'Communities for Children ' initiative as part of the Stronger Families and Communities Strategy. Communities for Children will directly fund 35 Australian communities between $1 and 4 million over four years to support parents, neighbourhoods and the wider community to give children the healthy start they need [ 19 ]. Importantly, there was little evidence of a community development or even a consultative approach in the implementation of this program, with the perception that Communities for Children has not been set up to respond to the greatest need. It has been suggested that system change can be accomplished by motivating institutions, systems and actors to move in common directions and develop structures that sustain these efforts over time [ 20 ]. This requires a high level of trust among the participants, such that they eventually share common goals and voluntarily seek to achieve common ends. Success in using exhortation as a policy instrument requires that information not only flow from government, but also to it [ 21 ]. The strong use of exhortation at the national level may be seen as the only way to encourage change, given the shared responsibility for child welfare among the various levels of government in Australia. Yet it is precisely this divided accountability and responsibility that has been identified 'as the greatest barrier to the reform of children's services' [[ 22 ] pg. 980]. The use of exhortation may be successful at motivating common approaches but will be much less effective at ensuring the sustaining structures are developed. This is apparent in the existing committee structures, which still operate from within the government structures and are thus unable to cross the 'silos' to promote the needed whole-of-government approach. To be effective in changing the paradigm in this area, the exhortation process will require back up by more coercive governing instruments. Conflicting expenditures – potential for uncertain outcomes in Victoria In comparison with other Australian states, Victoria has been slow to provide visible translation of the socio-ecological model of health for children and young people in a coordinated and systematic way to state policy. A recent review of Victorian paediatric services suggested that Victoria needed to establish a child and young people focus ensuring appropriate mechanisms to plan, coordinate and monitor across government departments and service providers [ 23 ]. It was suggested that a structure was required to coordinate child health among of the various portfolios in the Victorian Department of Human Services – Health, Housing, Welfare, and Disability – as well as among the broader Government departments, contributing to the whole of government approach required for early childhood intervention programs. The lack of coordinated focus on child health in Victoria is perhaps the result of a lingering policy focus on health surveillance. Despite the increasing evidence that surveillance and screening programs have limited effectiveness in child health [ 24 , 25 ], it is only recently that Victoria has increased the focus on the social determinants of health [ 26 , 27 ]. Most recently, Victoria has committed to the 'Best Start' program and will pilot it as demonstration projects in 10 communities across the State with an investment of $7.6 million. Best Start is auspiced by the Departments of Human Services and Education and Training and is focused on reducing the impact of disadvantage (from any cause) and enhancing the life chances of all children by strengthening the universal preventative system [ 28 ]. The aims of Best Start are multi-level including the social, emotional and physical well-being of children, capacity building of parents and carers and communities to assist them to become more child friendly, while focusing on specific interventions for socially disadvantaged families [ 29 ]. The demonstration projects are required to follow a prescribed implementation and evaluation process attempting to measure what works, under what circumstances and for whom, to ultimately improve services elsewhere in the State. Five approved demonstration sites with a total of $7.6 million are ensuring a 'brighter future' for the children of Frankston, Hume, Shepparton, Whittlesea and Yarra Ranges, while the rest of the State's children wait in the dark. Despite the intentions of Best Start , existing government funding and reporting in the area of maternal and child health is still largely focused on surveillance [ 26 ]. The Victorian approach to policy implementation in the area of early childhood support is focused predominantly on expenditure. Public expenditure is moderately coercive, with distribution of government funds to achieve particular policy objectives. But the small expenditure allocated to 'healthy' child policy that is limited to identified demonstration sites with expectations that the program will be shown to be effective before statewide mainstream implementation is overshadowed by a much larger expenditure pool that is not focused on the socio-ecological model. While Best Start signals intent to change the child health policy paradigm, the incentives established through the broader expenditure pool suggest, for the moment, maintenance of the status quo in Victoria. Guidelines – will enforcement back the regulatory approach in NSW? In New South Wales the 'Families First' initiative targets families with children 0 to 8 years, with the aim of helping parents give their children a good start in life. Demonstrating the commitment to a whole of government approach, the Office of Children and Young People (OCYP), located within The Cabinet Office, reporting directly to the Premier, has played a lead role in the development and implementation of the Families First strategy. This evidence-based approach is delivered jointly by five NSW government agencies – Area Health Services, Community Services, Education and Training, Housing and Disability, Ageing and Home Care in partnership with parents, community organisations and local government. NSW Health supports 'the ongoing development of partnerships at policy, planning and service delivery levels to enable improved co-ordination and intersectoral collaboration in the delivery of child health services' [[ 30 ] pg. 44]. NSW has also successfully translated much of the evidence into coordinated service planning and delivery at the regional level. Paediatric networks, associated with the Area Health Services, were established in 1997 and today provide designated primary, secondary and tertiary level services for families with children aged 0 to 5 years [ 31 ]. In addition, the NSW Commission for Children and Young People focuses on increasing the participation of children and young people in decision making that affects their lives, promoting the safety and welfare of children and young people, and strengthening the important relationships in the lives of children and young people and improving their well-being [ 32 ]. The implementation of Families First has been guided by a series of policy and practice guidelines. Recently, an independent review of Families First implementation within three regions, (Orana Far West, Illawarra and South West Sydney) found that the system changes required to build and strengthen service networks for families needed more than agreement and goodwill, with considerable effort to develop structures and processes that sustain interagency collaboration [ 33 ]. This resulted in a further guide to implementing sustainable and effective child and family service networks. Regulation involves the imposition of requirements to meet specific obligations. Often regulation is seen to exist within legislation outlining strict rules of behaviour. However, in health policy guidelines are considered effective means of imposing regulation, recognising the inherent uncertainty in safe practice in health care [ 17 ]. The implementation of 'healthy' child policy in NSW suggests a strong focus on the socio-ecological approach supported by the research evidence. This is apparent in the whole of government approach with leadership from the Premier's Office, backed by regulation to effect the necessary changes in the delivery system. However, it is yet to be seen whether the policy will be supported with the necessary resources for compliance. While regulation involves shifting costs of compliance from government to other participants, enforcement and monitoring can be expensive and difficult [ 21 ]. Without adequate enforcement the potential for inequality and inequity in access to the proposed service model is high. Conclusions A variety of instruments have been used by government to change the child health policy paradigm from that focused on social welfare to 'healthy' public policy predicated on a socio-ecological foundation. NSW has chosen regulation to implement a child health policy framework that is built upon the international evidence of the effectiveness of integrated, multidisciplinary and intersectoral policy and programs. A little slower to change paradigms, Victoria has established demonstration programs through targeted expenditure, without an overarching whole-of-government child health policy framework. Nationally, there is a move to change the paradigm to a broader definition of child health almost exclusively through exhortation. Instrument choice is influenced by a variety of factors. The use of exhortation by the Commonwealth Government is a relatively risk-free easy approach, which can counteract the divided accountabilities among federal and state governments in the area of health and social services. Exhortation is easy to implement; it is the least coercive and relies on voluntary goodwill. While it may be successful in building a common understanding, and even this is debatable, an independent review of the Families First implementation found that the system changes required to build and strengthen service networks for families needed more than agreement and goodwill to develop the necessary structures and processes – a suggestion that without other approaches to structural reform, exhortation is unlikely to be successful. The expenditure policy of Victoria illustrates an approach that is compromised by the lack of an underlying agreed evidence-based policy framework. This lack of coordinated government approach is reflected in conflicting expenditure policy in this area, with the potential to confound outcomes. While the regulatory approach of NSW suggests bold steps to change the paradigm, in fact, because regulation is not subjected to the same level of scrutiny of other instruments, such as expenditure, and even exhortation [ 17 ], it is a deceptively simple mechanism to implement policy [ 21 ]. The strength of the government intent to change the paradigm will only become apparent with visible enforcing of the service delivery directions. The evidence for a new policy paradigm is strong. But the use of these different policy instruments underscores the lack of shared understanding and policy agenda. Oberklaid suggests that while there are similarities in the rhetoric throughout Australia, there has been relatively little investment in child health [ 12 ]. The change in the child health public policy paradigm will only be successful when the governing instrument or combination of instruments induces the appropriate public and private behaviour. Perhaps we should be thankful that child health policy is on the agenda, and even without a strong, coordinated approach built on the evidence, one would agree that 'these developments in early childhood services demonstrate the translation of research evidence into policy and practice, even if the implementation may be flawed, belated or under-resourced' [[ 5 ] pg.15]. Competing interests The authors declare that they have no competing interests.	/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC544963.xml
368181	Corrections: The Roles of APC and Axin Derived from Experimental and Theoretical Analysis of the Wnt Pathway	null	In PLoS Biology, volume 1, issue 1: The Roles of APC and Axin Derived from Experimental and Theoretical Analysis of the Wnt Pathway Ethan Lee, Adrian Salic, Roland Krüger, Reinhart Heinrich, Marc W. Kirschner DOI: 10.1371/journal.pbio.0000010 Table 1: In the legend, the words fluxes and flux appeared without the fl . Table 3: In the legend, the word coefficients appeared without the fi . In the table, some numbers in the section “Binding, dissociation” were marked with a ± sign that should have been a ∓. Table 4: In the legend, the word coefficients appeared without the fi . Please see the corrected legends and table below. Table 1. Numeric Values of Input Quantities of the Model for the Reference State The data are grouped into concentrations of pathway components, dissociation constants of protein complexes, concentration ratios, fluxes and flux ratios, and characteristic times of selected processes. Experimental evidence for these data is discussed in the text. From these data, the following rates and rate constants are calculated: v 12 = 0.42 nM · min −1 (rate of β-catenin synthesis), v 14 = 8.2 · 10 −5 · nM · min −1 (rate of axin synthesis), k 4 = 0.27 min −1 , k 5 = 0.13 min −1 , k 6 = 9.1 · 10 −2 nM −1 · min −1 , k −6 = 0.91 · nM −1 · min −1 , k 9 = 210 min −1 , k 10 = 210 min −1 , k 11 · 0.42 min −1 , k 13 = 2.6 · 10 −4 min −1 , k 15 = 0.17 · min −1 . See Table S2, found at http://dx.doi.org/10.1371/journal.pbio.0000010.t002 , for more precise numbers used in the calculations. Bold: Measured values, Italics: Estimated values. DOI: 10.1371/journal.pbio.0000010.t001 Table 3 Control Coefficients for the Total Concentrations of β-Catenin and Axin and Parameters Quantifying the Sensitivity and the Robustness of the Wnt/β-Catenin Pathway Table 4. Concentration Control Coefficients for the Total Concentrations of β-Catenin and Axin Relative to Changes in the Concentrations of Pathway Components The control coefficients were obtained by numerical determination of the response to a change of total concentrations by 1%. Coefficients are given for the reference state and for the standard stimulated state. DOI: 10.1371/journal.pbio.0000010.t004 The full text XML and HTML versions of the article have been corrected online. This correction note may be found at DOI: 10.1371/journal.pbio.0020089.	/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC368181.xml
555852	BioAfrica's HIV-1 Proteomics Resource: Combining protein data with bioinformatics tools	Most Internet online resources for investigating HIV biology contain either bioinformatics tools, protein information or sequence data. The objective of this study was to develop a comprehensive online proteomics resource that integrates bioinformatics with the latest information on HIV-1 protein structure, gene expression, post-transcriptional/post-translational modification, functional activity, and protein-macromolecule interactions. The BioAfrica HIV-1 Proteomics Resource is a website that contains detailed information about the HIV-1 proteome and protease cleavage sites, as well as data-mining tools that can be used to manipulate and query protein sequence data, a BLAST tool for initiating structural analyses of HIV-1 proteins, and a proteomics tools directory. The Proteome section contains extensive data on each of 19 HIV-1 proteins, including their functional properties, a sample analysis of HIV-1 HXB2 , structural models and links to other online resources. The HIV-1 Protease Cleavage Sites section provides information on the position, subtype variation and genetic evolution of Gag, Gag-Pol and Nef cleavage sites. The HIV-1 Protein Data-mining Tool includes a set of 27 group M (subtypes A through K) reference sequences that can be used to assess the influence of genetic variation on immunological and functional domains of the protein. The BLAST Structure Tool identifies proteins with similar, experimentally determined topologies, and the Tools Directory provides a categorized list of websites and relevant software programs. This combined database and software repository is designed to facilitate the capture, retrieval and analysis of HIV-1 protein data, and to convert it into clinically useful information relating to the pathogenesis, transmission and therapeutic response of different HIV-1 variants. The HIV-1 Proteomics Resource is readily accessible through the BioAfrica website at:	Background Although the HIV-1 genome contains only 9 genes, it is capable of generating more than 19 gene products. These products can be divided into three major categories: structural and enzymatic (Gag, Pol, Env); immediate-early regulatory (Tat, Rev and Nef), and late regulatory (Vif, Vpu, Vpr) proteins. Tat, Rev and Nef are synthesized from small multiply-spliced mRNAs; Env, Vif, Vpu and Vpr are generated from singly-spliced mRNAs, the Gag and Gag-Pol precursor polyproteins are synthesized from full-length mRNA. The matrix (p17), capsid (p24) and nucleocapsid (p7) proteins are produced by protease cleavage of Gag and Gag-Pol, a fusion protein derived by ribosomal frame-shifting. Cleavage of Nef generates two different protein isoforms; one myristylated, the other non-myristylated. The viral enzymes (protease, reverse transcriptase, RNase H and integrase) are formed by protease cleavage of Gag-Pol. Alternative splicing, together with co-translational and post-translational modification, leads to additional protein variability [ 1 ]. Phylogenetic analysis, on its own, provides little information about the conformational, immunological and functional properties of HIV-1 proteins, but instead, focuses on the evolution and historical significance of sequence variants. To understand the clinical significance of genetic variation, sequence analysis needs to be combined with methods that assess change in the structural and biological properties of HIV-1 proteins. At present, information and tools for the systematic analysis of HIV-1 proteins are limited, and are scattered across a wide-range of online resources [ 2 , 3 ]. To facilitate studies of the biological consequences of genetic variation, we have developed an integrated, user-friendly proteomics resource that integrates common approaches to HIV-1 protein analysis (Figure 1 ). We are currently using this resource to better understand the structure-function relationships underlying the emergence of antiretroviral drug resistance, and to examine the process of immune escape from cytotoxic T-lymphocytes (CTLs). Figure 1 Site map of BioAfrica's HIV-1 Proteomics Resource, showing the separation of Beginner's and the Advanced area of the website, along with all major subject headings. We have categorized the Proteomics Resource into the following main subject headings (Figure 2 & 3 ): Figure 2 Schematic representation of BioAfrica's HIV-1 Proteomics Resource, showing its five major components: the HIV-1 Proteome (General Overview, Domains/Folds/Motifs, Genomic Location, Protein-Macromolecule Interactions, Primary and Secondary Database Entries, and References and Recommended Readings), the HIV-1 Protease Cleavage Sites section, the HIV-1 Protein Data-mining Tool, the HIV-1 BLAST Structure Tool, and the Proteomics Tools Directory (for Beginners and Advanced investigators). Figure 3 The central webpage of BioAfrica's HIV Proteomics Resource 1. HIV Proteome – Information about structure and sequence, as well as references and tutorials, for each of the HIV-1 proteins (Figure 4 ); Figure 4 The central webpage of the HIV-1 Proteome section of the BioAfrica website . 2. HIV-1 Cleavage Sites – Information about the position and sequence of HIV-1 Gag, Pol and Nef cleavage sites (Figure 5 ); Figure 5 The HIV-1 Protease Cleavage Sites section of the BioAfrica website . 3. HIV Protein Data Mining Tool – Application for detecting the characteristics of HIV-1 M group isolate (subtype A to K) proteins using information available in public databases and tools (Figure 6 ); Figure 6 The central webpage of the HIV-1 Protein Data Mining Tool section of the BioAfrica website, where a specific HIV-1 genomic region is selected to be analyzed . 4. HIV Structure BLAST – Similarity search for analyzing HIV protein sequences with corresponding structural data (Figure 7 ); Figure 7 The BLAST HIV-1 protein structure similarity search is an online tool that searches for all protein structure data within the PDB that have an amino acid sequence similar to the query sequence . 5. Proteomics Online Tools – Directory of data resources and tools available for both protein sequence and protein structure analyses of HIV (Figure 8 & 9 ). Figure 8 The introductory listing of proteomics resources for HIV research chosen to give a general overview of online tools and databases relevant for the analysis of HIV protein data . Figure 9 The advanced listing of online tools and databases relevant for the analysis of HIV protein data . The proteome link In the HIV-1 Proteome section, each of the 19 HIV-1 proteins has a webpage that is divided into six parts: "general overview", "genomic location", "domains/folds/motifs", "protein-macromolecule interactions", "primary and secondary database entries", and "references and recommended readings" (Figure 4 ). The overview includes a description of the protein, a list of known isoforms, a representative tertiary structure animated image (GIF format) of the protein and its co-ordinates (PDB format), a link to chime tutorials, if available, and information about cleavage sites, localization, and functional activity. The genomic location section provides information on the location of the sequence relative to the reference sequence, HIV-1 HXB2 [ 4 ], sequence data (fasta format), and information about the length, molecular weight and theoretical isoelectric point (pI) of the protein. The domains/folds/motifs section contains information about functional domains and predicted motifs (glycosylation, myristoylation, amidation, phosphorylation and cell attachment sites) of HIV-1 HXB2 [ 4 ], and provides structural predictions (secondary structure, transmembrane regions, low complexity regions, and coiled-coil regions). The section on protein-macromolecule interactions includes information on protein complexes, protein-protein/DNA/RNA interactions, signal-transduction pathways, and potential interactions with other pathogens. The section on primary and secondary databases contains a list of database entries that are needed to retrieve information on protein structure, nucleotide/amino acid sequence data, protein sequence annotation, proteins with similar sequence and structure (such as Los Alamos National Laboratories HIV Sequence Database and the RCSB Protein Data Bank), as well as information on post-translational modification and protein-protein interactions. A list of key reviews and publications, used in the development of the BioAfrica HIV-1 Proteomics Resource, is provided in the references and recommended readings section. As an example, the proteome webpage for Tat, describes how this protein up-regulates HIV-1 gene expression by interacting with the long-terminal repeat (LTR) of HIV-1, promoting the elongation phase of viral transcription, allowing full-length HIV-1 mRNA transcripts to be produced [ 5 , 6 ] (Figure 10 ). The webpage also gives information on the structural organization of tat gene. The mRNA is derived from spliced exons encoded in two different open reading frames. In HIV-1 HXB2 , these reading frames are separated by a distance of 2334 nucleotides. Some HIV-1 isolates, including HIV-1 HXB2 , contain an artifact of laboratory strains consisting of a premature stop codon at position 8424 of exon 2. The presence of this stop codon leads to the synthesis of a truncated form of Tat that is 86, rather than 101 amino acids in length. The protein has two different isoforms – one translated from early-stage multiply spliced mRNA (p14); the other from singly-spliced mRNA (p16) [ 7 ]. Important functional domains include the acidic, amphipathic region (1- M E PV D P R L E PW KH P G SQ P KT A -21; the hydrophobic residues are highlighted in bold, and polar residues are italicized) at the N-terminus of the protein; the cysteine-rich disulphide bond region (22- C TN C Y C KK CC FH C QV C -37); the core, basic and glutamine-rich region (49- R KK RRQRRR AH Q NS Q TH Q ASLSK Q -72) that is important for nuclear localization and TAR-binding activity, and the RGD cell-attachment site that binds to cellular integrins. In addition to being expressed in HIV-1-infected cells, Tat is also released into the extracellular fluid where it acts as a growth factor for the development of Kaposi's Sarcoma. Additional information about Tat and its protein-protein interactions can be found on the proteome page of the BioAfrica website located at . Figure 10 A general overview of the HIV-1 Proteome section of the BioAfrica website, as exemplified by the Tat web page . Protease cleavage sites link Post-translational cleavage of the Gag, Gag-Pol and Nef precursor proteins occurs at the cell membrane during virion packaging, and is essential to the production of infectious viral particles. Drugs that inhibit this process, the protease inhibitors (PIs), are the most potent antiretroviral agents currently available. Thus it is important to collect information, not only on the sequence of protease enzymes from different HIV-1 subtypes, but also on the natural polymorphisms and resistance mutations that may effect their catalytic activities, drug responsiveness, substrate specificities, and cleavage site characteristics. Studies have shown that resistance mutations in the protease of subtype B are associated with impaired proteolytic processing and decreased enzymatic activity, and that compensatory mutations at Gag and Gag-Pol cleavage sites can partially overcome these defects [ 8 ]. These findings suggest that variation at protease cleavage sites may play an important role, not only in regulation of the viral life cycle, but also in disease progression and response to therapy. The cleavage site section of the BioAfrica webpage is the direct extension of a recent publication in the Journal of Virology describing the location and variability of protease cleavage sites [ 9 ] (Figure 5 ). Together, these two resources provide information on the structure, amino acid composition, genetic variation and evolutionary history of protease cleavage sites, and on the natural selection pressures exerted on these sites. The section also serves as a baseline for understanding the impact of natural polymorphisms and resistance mutations on the catalytic efficiency of the protease enzyme, and on its ability to recognize and cleave individual Gag, Gag-Pol and Nef substrates. Such studies are important for understanding the mechanisms underlying the emergence of PI-induced drug resistance, and for designing alternative, optimized therapies. Protein data-mining tools link The HIV-1 Protein Data-Mining Tool contains twelve sequence analysis techniques for assessing protein variability among different strains of HIV-1 (Figure 6 ). These tools allow the user to manipulate, analyze and compare published [ 9 - 12 ] and newly-acquired data in a user-friendly, hands-on manner. The analysis is initiated by selecting a particular subset of HIV-1 proteins, either from the user's database, or from the representative dataset of group M viruses (subtypes A through K). Using this dataset, the investigator can then perform a variety of protein-specific analyses. With a single click of the mouse, users can download the amino acid sequence in fasta format; obtain sequence annotations from SwissProt [ 13 ] or GenBank [ 14 ]; identify functional motifs using BLOCKS [ 15 ], PROSITE [ 16 ] or ProDom [ 17 ]; perform similarity searches using the BLAST program available at Genbank [ 18 ], conduct structural comparisons using the BioAfrica BLAST Structure program; determine amino acid composition, predict hydrophobicity and tertiary structure using the Swiss-Model homology modelling server [ 19 ], and obtain a list of potential protein-macromolecule interactions from the Database of Interacting Proteins (DIP) [ 20 ]. A representative analysis of HIV-1 Tat is shown in Additional file 1 . The selected dataset, consisting of eight reference strains – four subtype B (HXB2-1983-France, RF-1983-US, JRFL-1986-US, WEAU160-1990-US) and four subtype C (92BR025-1992-Brazil, 96BW0502-1996-Botswana, TV002c12-1998-SouthAfrica, TV001c8.5-1998-SouthAfrica) isolates – were analyzed using PROSITE [ 16 ]. As shown in Additional file 1 , all eight isolates had identical amidation, cysteine-rich and myristylation motifs at amino acid codons 47–50, 22–37 and 44–49, respectively. Three (75%) of the B isolates contained a second myristylation site at codons 42–47, as did three (75%) subtype C viruses. One (25%) of the C viruses carried an extra GNptGS myristylation motif at position 79–84. In addition, all four (100%) C isolates contained a novel myristylation motif, GSeeSK, at amino acid position 83–88, that was not present in four B viruses selected for study. However, the most striking difference between the two subtypes was the increased number of phosphorylation motifs in subtype C relative to B viruses. This increase, which occurs in cAMP/cGMP-dependent kinase, protein kinase C (PKC) and casein kinase II (CKII) phosphorylation sites, has been reported previously [ 21 ], but the significance of these findings remain to be established. The analysis also highlighted the atypical nature of the HIV-1 HXB2 isolate, which, in addition to a premature stop codon, contained no cAMP/cGMP, PKC or CKII phosphorylation sites. The blast structure tool link The HIV-1 BLAST Structure Tool facilitates the analysis of HIV-1 protein structure by allowing for rapid retrieval of archived structural data stored in the public databases (Figure 7 ). Users may input any HIV-1 amino acid sequence and obtain a list of similar HIV protein sequences for which structural data have been experimentally determined and deposited into the Protein Data Bank (PDB) [ 22 ]. After downloading the data from the PDB, subsequent structural analyses can be performed using the software programs and web-servers listed in the Proteomics Tools Directory. For example, a query using an amino acid sequence of HIV-1 Integrase protein from NCBI (gi\|15553624\|gb\|AAL01959.1) results in a list of 54 structural models (ie. PDB_ID\|1K6Y) within the PDB. Each of these structural models can be retrieved from the PDB, and the most appropriate structural model could be used for generating a homology model using the query protein sequence. The proteomics tools directory link The HIV-1 Proteomics Tools Directory is divided into two web pages. The initial webpage is a concise compilation of some of the most commonly used protein-specific Internet resources (Figure 8 ). This "beginners" page displays a short list of websites for each of the following twelve categories: "protein databases", "specialized viral-protein databases", "motif and transcription factor databases", "protein sequence similarity searches", "protein sequence alignment", "protein sequence prediction tools", "protein sequence analysis", "protein sequence manipulation", "protein structure analysis", "molecular modelling tools", "tutorials", and "downloads". In addition, the Proteomics Tools Directory has an advanced web page for users who are looking for alternative, or more specialized, protein analysis tools (Figure 9 ). The advanced webpage displays a list of more than 200 links to different websites and web-servers. These data sources contain a variety of information ranging from specialized protein sequence databases to software programs capable of performing rigid body protein-protein molecular docking simulations. Conclusion The impending rollout of antiretroviral therapy to millions of HIV-1-infected people in sub-Saharan Africa provides a unique opportunity to monitor the efficacy of non-B treatment programs from their very inception, and to obtain critical new information for the optimization of treatment strategies that are safe, affordable and appropriate for the developing world. An integral part of this massive humanitarian effort will be the collection of large amounts of clinical and laboratory data, including genetic information on viral subtype and resistance mutations, as well as routine CD4+ T-cell counts and viral load measurements. The mere collection of this data, however, does not ensure that it will be used to its maximum potential. To achieve full benefit from this explosive source of new information, the data will need to be appropriately collated, stored, analyzed and interpreted. The rapidly emerging field of Bioinformatics has the capacity to greatly enhance treatment (and vaccine) efforts by serving as a bridge between Medical Informatics and Experimental Science. By correlating genetic variation and potential changes in protein structure with clinical risk factors, disease presentation, and differential response to treatment and vaccine candidates, it may be possible to obtain valuable new insights that can be used to support and guide rationale decision-making, both at the clinical and public health levels. The HIV-1 Proteomics Resource, described in this report, is an initial first step in the development of improved methods for extracting and analyzing genomics data, converting it into biologically useful information related to the structure, function and physiology of HIV-1 proteins, and for assessing the role these proteins play in disease progression and response to therapy. The Resource, developed at the Molecular Virology and Bioinformatics Unit of the Africa Centre of Health and Population Studies, is a centralized user-friendly database that is easily accessed through the BioAfrica website at [ 23 ]. List of abbreviations used AA – Amino Acid BLAST – Basic Local Alignment Search Tool CKII – casein kinase II CTLs – cytotoxic T-lymphocytes DIP – Database of Interacting Proteins DNA – deoxyribonucleic acid Env – envelope glycoprotein Gag – group-specific antigen polyprotein GIF – Graphics Interchange Format HIV – Human Immunodeficiency Virus HIV-1 – Human Immunodeficiency Virus Type-1 HTTP – Hypertext Transfer Protocol LTR – long-terminal repeat mRNA – messenger RNA NCBI – National Center for Biotechnology Information Nef – negative factor PDB – Protein Data Bank pI – isoelectric point PIs – protease inhibitors PKC – protein kinase C Pol – polymerase polyprotein Rev – ART/TRS anti-repression transactivator protein RNA – ribonucleic acid RNase H – ribonuclease H Tat – transactivating regulatory protein Vif – virion infectivity factor Vpr – viral protein R Vpu – viral protein U Competing interests The author(s) declare that they have no competing interests. Authors' contributions RSD created and maintains BioAfrica's HIV proteomics resource, HIV proteome section, proteomics tools directory, HIV-1 protein data-mining tool and HIV structure BLAST tool; performed protein sequence and structural model analyses; and wrote the manuscript. TDO conceived and maintains the BioAfrica website, and continues to oversee its rapid expansion; created the cleavage sites section; and participated in the design and implementation of the HIV proteomics resource. CS participated in the design of the HIV proteomics resource, with an emphasis on the proteomics tools directory. SD participated in the design and creation of the HIV proteome section, with an emphasis on the HIV-1 Tat protein. MG participated in the design of the HIV proteomics resource, with an emphasis on the HIV proteome section. SC supervised the project, and participated in the design and implementation of the HIV proteomics resource. All authors read and approved the final manuscript. Supplementary Material Additional File 1 A table containing a comparative summary of potential functional motifs (cysteine-rich region, myristoylated Asparagine, amidation, cAMP- and cGMP- dependent kinase phosphorylation, Protein Kinase C phosphorylation, and Casein Kinase II phosphorylation) in the HIV-1 Tat proteins of subtypes B and C, as identified using PROSITE. Click here for file	/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC555852.xml
555846	Heat shock protein 72 expression allows permissive replication of oncolytic adenovirus dl1520 (ONYX-015) in rat glioblastoma cells	In this study we have made novel observations with regards to potentiation of the tumoricidal activity of the oncolytic adenovirus, dl1520 (ONYX-015) in rat glioblastoma cell lines expressing heat shock protein 72 (HSP72) due to permissive virus replication. ONYX-015 is a conditionally replicating adenovirus that is deleted for the E1B 55 kDA gene product whose normal function is to interact with cell-cycle regulatory proteins to permit virus replication. However, many murine and rodent cell lines are not permissive for adenovirus replication. Previously, it has been reported that the heat shock response is necessary for adenovirus replication and that induction of heat shock proteins is mediated by E1 region gene products. Therefore, we hypothesized that HSP72 expression may allow for permissive replication of ONYX-015 in previously non-permissive cells. Rat glioma cell lines 9L and RT2 were transfected with a plasmids expressing HSP72 or GFP. After infection with ONYX-015, no tumoricidal activity is observed in GFP expressing cell lines despite adequate transduction. In contrast, HSP72 transfected cells show cytopathic effects by 72 hours and greater than 75% loss of viability by 96 hours. Burst assays show active virus replication in the HSP72 expressing cell lines. Therefore, 9L-HSP72 and RT2-HSP72 are ideal models to evaluate the efficacy of ONYX-015 in an immunocompetent rat model. Our study has implications for creating rodent tumor models for pre-clinical studies with E1 region deleted conditionally replicating adenovirus.	Background Adenovirus vectors are commonly utilized in cancer gene therapy experiments. They readily infect numerous tumor cell types and are easily manipulated allowing for transgene expression[ 1 ]. In an effort to improve selectivity of these vectors for malignant cells, replication selective or conditionally replicating adenoviruses were created[ 2 ]. With greater understanding in the molecular aspects of adenovirus replication, these viruses were designed such that replication was predicated on alterations in cell cycle regulation, thus rendering only malignant cells susceptible. These adenovirus systems rely on virus replication as a means of exerting tumoricidal effect. One of the first of these conditionally replicating adenoviruses was dl1520 or ONYX-015[ 3 ]. It has been used extensively in cancer gene therapy clinical trials[ 4 , 5 ]. This virus is deleted for the E1B-55 kDA gene[ 6 ]. The E1 region genes of the common serotypes of adenovirus optimize their own replication in target cells and interact with many cell cycle associated gene products. The E1a region genes promote transition of cells into S phase through their interaction with the retinoblastoma protein (pRB)[ 7 ]. pRB in conjunction with p53 can inhibit transition of cells into the S-phase of the cell cycle. Binding of E1a to pRB prevents association of pRB to the E2F transcription factor resulting in activation of E2F[ 8 ]. E2F then functions to transition cells into the S-phase of the cell cycle which is conducive for optimal adenovirus replication. In wild type adenoviruses, the role of the E1B-55 kDa gene product is to neutralize p53[ 9 , 10 ]. p53 induction is thought to promote cell cycle arrest resulting in termination of the virus replicative lifecycle in that cell[ 11 ]. Thus it was hypothesized that deletion of the E1B-55 kDa gene product would result in p53 induced termination of virus replication in normal cells while being replication permissive in malignant cells with abnormal p53 expression or regulation[ 9 ]. Early reports with ONYX-015 indicated replication selectivity for tumor cells with abnormal p53 functioning[ 12 ]. However, the exact mechanism of replication selectivity has been brought to question with subsequent reports of ONYX-015 replication independent of p53[ 13 ] and researchers have looked at other factors associated with adenovirus replication. Various heat shock proteins have been shown to be necessary for efficient adenovirus replication. The avian adenovirus CELO requires the induction of HSP70 and HSP40 for production of viral proteins[ 14 ]. CELO mutants lacking the E1 region genes were replication incompetent in A549 cells. However, heat shock protein induction or expression allowed for permissive replication of these mutants. Similarly in a number of human cell lines, both heat shock and HSP72 expression enhanced the oncolytic effect of E1 containing adenovirus but not for E1 deleted adenovirus[ 15 ]. Adenovirus E1 region genes from serotype 5 and 12 have previously been shown to induce HSP72 expression [ 16 ]. HSP72 is a molecular chaperone protein involved in the repairing denatured proteins. Through this role, it can inhibit apoptosis downstream of caspase activation but prior to loss of mitochondrial membrane potential[ 17 ]. Stably transfected or E1 transformed cell lines show constitutive expression of the inducible HSP72 [ 18 ]. HSP72 levels are correlated with cell-cycle with increases in HSP72 during S-phase with a maximum level in the post-S-phase period [ 19 ]. In 293 cells which are transformed with the adenovirus E1 region gene, E1A mRNA accumulated just prior to HSP72 expression suggesting that E1A is responsible for the cell cycle regulation of HSP72 expression [ 18 ]. Additionally, HSP72 co-localizes with E1A gene products in infected cells [ 20 ]. HSP72 is predominately cytoplasmic but translocates to the nucleus following adenovirus infection. Double labelling experiments with E1A and HSP72 in infected cells showed co-localization of these molecules within the nucleoli and exhibited similar reticular and punctuate nuclear staining patterns depending on cell cycle. HSP72 transport to the nucleus was E1A and virus infection dependent. In 293 (E1 region transformed) cells, HSP72 and E1A co-localization was not observed until after infection with virus [ 20 ]. These data suggest a physical complex is necessary between adenovirus E1A, HSP72 and other adenovirus gene products One of the difficulties with studying E1 adenovirus deletion mutants is the lack of appropriate animal tumor models which allow for permissive growth of these viruses. Common animal tumor models for glioblastoma are the cell lines 9L and RT2. Previously, it has been shown that replication competent E1 deletion mutants can transduce 9L cells and result in transgene expression but there is a block against virus replication[ 21 ]. In this study we addressed if HSP72 expression could allow for permissive replication of the ONYX-015 adenovirus in these rat glioblastoma cells. Results Transduction Efficiency Prior to studying the efficacy of ONYX-015 to result in tumor cell lysis, we determined the optimal dose necessary for tissue transduction. Adenovirus infection efficiency is predicated on the expression of adenovirus surface receptors[ 22 ]. Using a recombinant adenovirus expressing green fluorescent protein (GFP), we transduced various glioma cell lines and analyzed them by fluorescence microscopy. Results are summarized in Table 1 . Both the rat cell lines, 9L and RT2 required a significantly high dose of adenovirus to result in effective transduction. In contrast, the human glioma cell lines, DBTRG and NB4, require much less virus for efficient transduction. Table 1 Transduction efficiency of various CNS tumor cell lines by AD-GFP Numbers represent percentage of cells transduced with a recombinant adenovirus vector expressing green fluorescent protein (AD-GFP) at the stated multiplicity of infection (MOI). Transduction was determined by fluorescence microscopy of cells on a hemacytometer and is represented as % of total cells. Multiplicity of Infection Cell Line 0 1 10 100 300 500 1000 9L 0 0 1 27 54 63 74 RT2 0 1 8 41 68 81 93 DBTRG 0 69 100 100 - - - NB4 0 74 100 100 - - - Cytopathic Effect The dramatic difference in tumoricidal effect between HSP72 transfected and control tumor cells is best observed by cytopathic effect. In Figure 1 we show the affects of virus infection on RT2-GFP and RT2-HSP72 cells 96 hours after infection with ONYX-015. Non-infected RT2-GFP and RT2-HSP72 cells showed normal glioma histological features and growth characteristics (Figure 1 , panel A and B). The cells were infected when 50% confluent. At 96 hours, RT2-GFP cells infected at MOI of 100 or 300 showed little cytopathic effect and very little difference was ascertained in comparison to non-infected cells (Figure 1 , panel C and E). In contrast, the majority of RT2-HSP72 cells were dead or dying (Figure 1 , panel D and F). The few remaining cells were larger in size with oval or round morphology. These cells showed morphological characteristics consistent with adenovirus replication as observed in 293 cells. Similar results were obtained when assessing 9L-HSP72 cells. In table 2 we show the time to full CPE in the rat glioblastoma cell lines and a variety of human controls. Given the differences in transduction efficiency, the human cell lines were infected at a lower multiplicity of infection. No CPE is detected in the 9L-GFP and RT2-GFP cell lines. In contrast, 9L-HSP72 and RT2-HSP72 show CPE in as little as 4 days comparable to human glioblastoma cells when infected at an MOI of 10 and show CPE more rapidly than the human cell lines despite a lesser transduction efficiency. Table 2 Time to Cytopathic Effect Cells grown in 96 well plates were infected with ONYX-015 at the stated multiplicity of infection per row of cells. Values represent days until 50% of wells in a row showed cytopathic effect. Multiplicity of Infection Cell Line 0 1 10 100 300 9L-GFP 0 0 0 0 0 9L-HSP72 0 0 8 5 4 RT2-GFP 0 0 0 0 0 RT2-HSP72 0 0 0 4 4 DBTRG 0 6 4 - - NB4 0 5 4 - - 293 0 2 1 - - Figure 1 Cytopathic effect of ONYX-015 on GFP and HSP72 transfected glioma cells. Photographs (100× magnification) of RT2-GFP and RT2-HSP72 cells 96 hours after ONYX-015 infection at various dosages are represented. A – RT2-GFP cells mock infected with virus. B – RT2-HSP72 cells mock infected with virus. C – RT2-GFP cells infected at MOI of 100. D – RT2-HSP72 cells infected at MOI of 100. E – RT2-GFP cells infected at MOI of 300. F – RT2-HSP72 cells infected at MOI of 300. HSP72 Expression Augments ONYX-015 Toxicity When HSP72-transfected cells were infected with a low dose (MOI 100) of ONYX-015, there is no loss of cell viability. However, comparison of cell numbers to GFP-transfected cells or AD-GFP infected cells showed a dramatic cytostatic effect (Figure 2 ). 9L-GFP and 9L-HSP72 cells were transduced with either AD-GFP or ONYX-015. There is no effect on cell proliferation over time of the 9L-GFP cells infected with either AD-GFP or ONYX-015 (data not presented). In contrast, the 9L-HSP72 cells show normal proliferation when infected with AD-GFP infected cells and their cell numbers increase over time. However, ONYX-015 infected 9L-HSP72 cells showed loss of proliferative potential with very little increase in cell number over time. After 72 hours, the AD-GFP infected cells showed a decrease in proliferation but this is likely due to overcrowding of cells in the tissue culture vessel and/or the rapid utilization of consumable energy in the media. Figure 2 Cytostatic effect of ONYX-015 infection on HSP72 transfected 9L cells. 9L-HSP72 cells were transduced with either AD-GFP (□) or ONYX-015 (■) at a multiplicity of infection of 100. Cell numbers were determined over time (X-axis) by direct counting using a Coulter Counter. Values (Y-axis) represent log of cell number as a fraction of the initial cell number plated. When assessing for cell viability, 9L-GFP (figure 3 ) and RT2-GFP (figure 4 ) cells exhibited minimal toxicity when infected with ONYX-015. In contrast, both 9L-HSP72 and RT2-HSP72 cell lines were more susceptible to ONYX-015 cell lysis. In both HSP72 transfected cell lines, loss of cell viability began as early as 24 hours after infection. Both 9L-HSP72 and RT2-HSP72 show greater than 90% loss of viability by 96 hours. Figure 3 9L Cell viability after ONYX-015 infection. Cell viability (as % viable cells; y-axis) of 9L-GFP (■) and 9L-HSP72 (●) cell lines after infection with ONYX-015 at a MOI of 300 over time (x-axis). There is statistical significance (p < 0.05) when comparing GFP and HSP72 transfected cell numbers by analysis of variance at the 72 and 96 hour time points. Figure 4 RT2 Cell viability after ONYX-015 infection. Cell viability (as % viable cells; y-axis) of RT2-GFP (■) and RT2-HSP72 (●) cell lines after infection with ONYX-015 at a MOI of 300 over time (x-axis). There is statistical significance (p < 0.05) when comparing GFP and HSP72 transfected cell numbers by analysis of variance at the 72 and 96 hour time points. Burst Assays The mechanism of cell death associated with replication competent adenoviruses has been attributed to release of virus from cells following virus replication. Having observed cytopathic effects (figure 1 ) in association with ONYX-015 infection of 9L-HSP72 and RT2-HSP72 cells, we presumed the same mechanism applied. To confirm this hypothesis we evaluated adenovirus replication in GFP and HSP72 transfected tumor cells by burst assay. In a normal burst assay, an increase in virus titer after 72 hours is indicative of virus replication. Tumor cells were also infected with the replication incompetent virus (AD-GFP) as a negative control and no significant virus replication was measured (data not presented). Since HSP72 transfected cells were susceptible to ONYX-015 mediated cell death we expected an increase in virus titer if the cell death was a result of permissive adenovirus replication. After infection of cells with ONYX-015, the 9L-HSP72 and RT2-HSP72 cells showed an increase in virus titer confirming virus replication (figure 5 ) in contrast to 9L-GFP and RT2-HSP72 transfected cells. There was a direct correlation between virus titer and cell death. Virus titer in both GFP transfected and HSP72 transfected cells increased in proportion to cell death. Figure 5 Burst assays. Burst ratio is given as the virus titer at 72 hours compared to 4 hours of various cell lines (x-axis) infected while 50% confluent with ONYX-015 at MOI of 100. Results are representative of a typical experiment of at least three performed. () represents statistical significance (p < 0.05) by Student's t-test between 9L-HSP72 and 9L-GFP. (*) represents statistical significance (p < 0.05) by Student's t-test between RT2-HSP72 and RT2-GFP. Discussion ONYX-015 is a promising new cancer gene therapy vector which exerts its tumoricidal effect by selective replication in cancer cells[ 6 ]. Replication selectivity previously was attributed to interactions of E1 gene products with p53[ 3 ]. Subsequent reports have shown that ONYX-015 replicates in cells with wild-type p53 as efficiently as in cells with mutant-p53[ 13 ]. Although the mechanism of cancer cell selectivity is still under investigation, the various hypotheses consistently involve cell-cycle regulatory proteins and their interactions within apoptosis pathways. HSP72 is a potent regulator of apoptosis [ 23 ]. Also, there is evidence that activation of the heat shock response by adenovirus early region genes is necessary for virus replication[ 14 ]. Previously it has been shown that heat shock induced HSP72 expression resulted in increased production of adenovirus proteins[ 24 ]. Additionally, the E1A adenovirus gene products necessary for virus genome replication have previously been shown to co-localize with HSP72[ 20 ]. To study the role of HSP72 in augmenting ONYX-015 replication, we chose two glioma cell lines with different derivations. 9L is a chemically induced gliosarcoma while RT2 is a virally transformed glioblastoma. Compared to many human tumor cell lines, these cell lines show relative resistance to the cytotoxic affects of ONYX-015. Resistance in part was secondary to low adenovirus transduction efficiency. At the multiplicities of infection utilized in this study, we never achieved 100% cell transduction (table 1 ). In most cases, when less than 30% of permissive HSP72-transfected cells were transduced, tumor cell growth was rapid and outpaced ONYX-015 virus replication resulting in no discernable cytotoxic or cytostatic effect after 1 week in culture (data not presented). Therefore, replicative adenovirus cytotoxicity is likely a balance between virus transduction efficiency, replication efficiency and cell growth characteristics. HSP72 transfection had no bearing on cell doubling time of either 9L or RT2 cells (data not presented) RT2 cell-doubling time is approximately 8 hours while 9L cell-doubling time is approximately 14 hours. Both cell lines after HSP72 transfection showed equal susceptibility to ONYX-015. Further evidence for the balance between transduction efficiency, replication efficiency and cell doubling time is evident in that ONYX-015 infection resulted in a predominantly cytostatic affect resulting in decreased cell division before evidence of cytotoxicity. The most dramatic affect of the role HSP72 in promotion of virus replication is evident in figure 1 . At 96 hours, the GFP transduced cells showed very little cytotoxic effect of ONYX-015 infection. In contrast, there is a distinct cytopathic effect at the same doses of ONYX-015 infection in the HSP72 transduced cells. These cells were rendered susceptible to ONYX-015 toxicity. The mechanism of cell death is related to increased virus replication. Significantly more ONYX-015 was isolated from HSP72 transduced cells than the GFP-transfected controls. The E1a gene products of adenovirus are responsible for activation of the HSP72 promoter and may affect HSP72 levels during the cell cycle[ 18 ]. Conversely, there is no evidence that HSP72 interacts with adenovirus promoters to stimulate viral transcription. There is evidence of HSP72 interactions with adenovirus structural proteins such as hexon[ 25 ] and fiber[ 26 ]. Adenovirus assembly is known to be inefficient with only a small percentage of total structural proteins eventually utilized in the production of infectious virus [ 26 ]. In conclusion, our studies demonstrated that HSP72 expression in rodent glioma tumor cells potentiated ONYX-015 replication and oncolysis of tumor cells. Further studies on the role of HSP72 in tumor types with wild-type and mutant p53 tumor suppressor genes is warranted to understand the molecular interactions of this chaperone protein in promoting virus replication. Additionally, addressing the role of HSP72 as an inhibitor of apoptosis in virus replication would further help characterize the interaction of cell-cycle regulatory pathways with virus replication. Furthermore, we intend to evaluate the role of HSP72 associated permissive replication in other animal cell lines to establish animal models for the study of ONYX-015 and similar replication competent adenoviruses. Conclusion For clinical applicability, these findings have a number of implications. Currently, clinical trials are ongoing with ONYX-015. Due to the previous association of ONYX-015 replication with p53 status, the current tumor types chosen for clinical trials are those exhibiting abnormal p53 expression in a majority of cells. Our study indicates, that tumor types with high HSP72 may also be viable candidates to evaluate ONYX-015 efficacy, as they are likely to demonstrate enhanced susceptibility to this replication competent virus. As HSP72 expression in tumors is associated with a more aggressive tumor phenotype, ONYX-015 may be ideal as adjunctive therapy for advanced disease. Additionally, one of the greatest limitations to adenovirus gene therapy is patient safety at high doses of adenovirus administration. Since, HSP72 expression resulted in oncolysis at much lower MOIs, a lower dose of virus administered to a HSP72 positive tumor may have the same benefit as a higher dose, allowing one to use lower doses to achieve the same effect. In addition, for some tumor types such as breast cancer, hyperthermia has been shown to significantly improve response rates[ 27 ]. HSP72 is readily induced with hyperthermia. Therefore, HSP72 induction by hyperthermia may also serve as an effective strategy to augment ONYX-015 oncolytic activity. We are currently in the process of studying this hypothesis. Lastly, the role of HSP72 in cell cycle regulation suggests alternate pathways for the mechanism of E1b deleted adenovirus replication in tumorigenic cells. HSP72 overexpression in non-transformed fibroblasts did not result in ONYX-015 oncolysis (data not presented). Understanding the molecular interactions of adenovirus proteins with molecular chaperones would help determine cell selectivity to replication competent adenoviruses influencing the design of more potent viruses. Methods Cell Lines 9L and RT2 cells are rat glioma cell lines with different derivations. 9L cells are chemically induced gliosarcoma cells[ 28 ] while RT2 are virally transformed and have glioblastoma histology[ 29 ]. These cell lines were obtained from Dr. Martin Graf (Medical College of Virginia, Richmond, VA). Cells were kept at 37°C, 5% CO 2 and 95% humidity in Dulbecco's modified eagle medium (Cellgro, Herndon, VA) supplemented with 10% (v/v) heat inactivated fetal bovine serum (BioWhittaker, Walkersville, MD), 2 mM L-glutamine and 100 units/ml Penicillin and 1000 ug/ml Streptomycin (Invitrogen). 9L and RT2 cells were transfected with the plasmids pEGFP-N2 (Clontech) or pHSP72 using Lipofectamine (Invitrogen) according to the manufacturers protocol followed by selection in 0.2 mg/ml G418 (Invitrogen). 293 cells were obtained from American Type Culture Collection (ATCC, Manassas, VA; CRL-1573) were used for virus propagation and purification. Plasmid Creation pHSP72 was created by ligation of the DNA Polymerase blunt ended 2.3 KB Bam H1-Hind III fragment from plasmid pH2.3 (ATCC) into the blunt ended Xho1-XbaIA vector pEGFP-N2. Adenovirus Creation of ONYX-015 has previously been described[ 6 ]. The virus was propagated on 293 cells and purified by cesium chloride gradient followed by dialysis. Virus was stored at -80°C until use. Titer was determined by tissue culture infectious dose-50 method (TCID-50). Briefly, 293 cells were plated at 10 4 cells/well in 96 well flat-bottomed tissue culture plates. Virus titer from combined supernatant and freeze-thawed cell lysate was determined by serial 10-fold dilution into different rows of the titer plate. After 10 days, wells were scored for cytopathic effect. The titer was calculated using the formula T = 10 1+d(S-0.5) where d is the Log 10 of the dilution and S is the sum of the ratios of positive wells in each row. A more lengthy description of this method is provided in the adenovirus application manual at . Adenovirus Infection Cells were infected when 50% confluent with the dose of adenovirus given in the appropriate table. The infection was performed in serum free DMEM media for 90 minutes at 37°C and 5% CO 2 . Cells were subsequently washed with phosphate buffered saline and then cultured in complete medium. Cell Viability Determination Cell viability was determined by the standard Trypan-blue exclusion test. Cells were washed with PBS and then lifted from tissue culture plates with Trypsin-EDTA solution (Invitrogen). Cells were stained with 0.2% Trypan blue solution for 5 minutes and then counted on a hemacytometer. Authors' contributions JM performed all of the experiments outlined in this study. JAK independently replicated results for validity. JAK and MRS contributed equally to the molecular cloning of the plasmids and adenovirus propagation and purification. All authors participated in maintenance of cells in culture. The hypothesis, study design and statistical analysis were all performed by MRS. MRS drafted the entire manuscript with editing and approval by all of the authors.	/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC555846.xml
406403	Small Fish, Big Science	The European Union recently awarded 12 million Euros to the ZF-MODELS research consortium to study zebrafish models for human development and disease	Francis Hamilton, the Briton who first described zebrafish ( Danio rerio ) in 1822, would be astounded to see the scientific attention now afforded to this two-inch-long native of Indian rivers. A fish with no economic worth was how he described this little creature. Yet recently, the European Union awarded 12 million Euros to the ZF-MODELS research consortium to study zebrafish models for human development and disease. When and why did zebrafish swim from home aquaria into research labs, and what can we learn about our biology from this surprising source? The Early Days It was the late 1960s when phage geneticist George Streisinger began to look for a model system in which to study the genetic basis of vertebrate neural development. His passion for tropical fish led him to the humble zebrafish. He was a ‘visionary’, remembers neurobiologist Judith Eisen (University of Oregon, Eugene, Oregon, United States), ‘who laid the groundwork for the use of zebrafish as a developmental model’. Eisen, who now heads her own research group, went to Oregon in 1983 to work on Xenopus neural development but soon became attracted to zebrafish as a model organism. By the early 1980s, she explains, Streisinger had worked out many of the genetic tricks needed to tackle zebrafish development. What's more, the fish had ‘wonderful embryology’. The embryo, which develops outside its mother, is transparent. ‘You can see different cell types, watch individual cells develop, do transplantation experiments’, Eisen enthuses, ‘and development is quick but not too quick’. Being able to watch individual neurons developing in real time opened up whole new avenues of research for Eisen and other neurobiologists. Fast Forward to the Big Screen The properties of zebrafish that attracted Eisen soon attracted people interested in other aspects of vertebrate development to the stripy tiddler ( Figure 1 ). As Eisen comments: ‘No other developmental model has risen to prominence so quickly’. These days more than 3,000 researchers are listed on ZFIN, a United States–based information resource for the zebrafish research community. Figure 1 Adult Zebrafish (Image courtesy of Lukas Roth, University College London, London, United Kingdom.) The speedy expansion was driven in great part by two genetic screens initiated in 1992–1993 by Christiane Nüsslein-Volhard in Tübingen, Germany, and Wolfgang Driever and Mark Fishman in Boston, Massachusetts. The aim of both screens was to identify genes with unique and essential functions in zebrafish development, and in 1996 an issue of the journal Development was dedicated to the mutants that had been isolated and characterised. These screens, says Ralf Dahm (Max Planck Institute for Developmental Biology, Tübingen, Germany), project manager of the ZF-MODELS consortium, ‘were the first major zebrafish projects, and they showed that zebrafish was a model organism to be reckoned with’. ‘That was a fantastic time’, says Derek Stemple, then a postdoc with Driever but now a group leader at the Wellcome Trust Sanger Institute (Cambridge, United Kingdom) and a ZF-MODELS participant. ‘From Wolfgang's lab, I was able to take the mutations that affected notochord development, and have been studying them ever since’. The notochord is an embryonic structure that forms the primitive axial skeleton of the developing embryo, and because mutations affecting notochord development result in shortened embryos, seven of the affected genes have been named after the dwarves in Snow White—zebrafish, like some other developmental models, have many imaginatively named mutants. Stemple now knows the identity of six of these mutated genes, all of which lead to disruption of basement membrane around the notochord. Many mutants from those first two screens are still used by developmental biologists, but another set of mutants has recently been isolated by Nancy Hopkins, Amgen Professor of Biology at the Massachusetts Institute of Technology (Cambridge, Massachusetts, United States). About ten years ago, Hopkins started to develop insertional mutagenesis in zebrafish ( Figure 2 ). In this approach, mutations are caused by the random insertion of viral DNA throughout the fish genome. The inserted DNA acts as a tag, making cloning of mutated genes very straightforward, although the efficiency of the initial insertional mutagenesis is much lower than that of the chemical mutagenesis used in the 1992–1993 screens. Hopkins has isolated 550 mutants in her screen, representing around 400 different genes, and has cloned more than 300 of these genes to date. Some of the fruits of this project are published in this issue of PLoS Biology. Hopkins's group is now collaborating with 25 external laboratories on the annotation of the mutant collection with funding from the National Center for Research Resources, part of the United States National Institutes of Health. Figure 2 A Zebrafish Pigment Mutant The mutant called bleached blond was produced by insertional mutagenesis. The embryos in the picture are four days old. At the top is a wild-type embryo, below is the mutant. The mutant lacks black pigment in the melanocytes because it fails to synthesise melanin properly. (Image courtesy of Adam Amsterdam, Massachusetts Institute of Technology, Boston, Massachusetts, United States.) The Tübingen researchers did another chemical mutagenesis screen between 2000 and 2001, and are now starting a third screen of 6,000 genomes as part of the ZF-MODELS project. ‘Each of our screens has built on the previous one by including more specific assays’, explains Dahm. Mutagenesis for the third screen is underway, but the assays, which include looking for defects that specifically affect adults, are still at the pilot stage; this autumn, the project's executive committee, which is headed by Nüsslein-Volhard, will decide which assays to use in the full-scale screen. ‘Just over half the 17 partners in the consortium will come to Tübingen to do screens’, predicts Dahm. ‘By bringing in expertise in different systems in this way we should greatly increase the efficiency of the screen’. What Else Will ZF-MODELS Do? The ZF-MODELS consortium, which is funded under the European Union's Sixth Framework Programme, aims to establish zebrafish models for human diseases, discover genes that will lead to the identification of new drug targets, and gain fundamental insights into human development. ‘We will mainly focus on using advanced technologies that have recently become available’, says scientific coordinator Robert Geisler (Max Planck Institute for Developmental Biology). For example, Geisler's lab will use DNA chip technology to investigate gene expression patterns in zebrafish mutants and so provide increased knowledge of the regulatory pathways that act in zebrafish development. Consortium members will also use ‘reverse genetics’ to investigate these pathways. In reverse genetics, researchers start with a gene of interest and investigate the phenotypic effect of altering its activity; by contrast, in ‘forward genetics’ the starting point is to look for a particular phenotype and then hunt out the altered gene that is causing it. Two reverse genetics approaches will be used by the consortium. Gene expression will be transiently knocked down with morpholinos, short segments of the gene that block its function. In addition, a recently developed technique known as TILLING (targeting induced local lesions in genomes) will be used to knock out gene activity permanently. The first step in TILLING is to mutagenise male zebrafish and mate them with untreated female fish, explains Stemple, whose group is one of three ZF-MODELS partners who will use this approach. Offspring are raised to adulthood, and the DNA of each individual is then genotyped for the exon of interest. The consortium already has a collection of 6,000 such individuals, and once a fish carrying a mutation in the gene of interest has been identified, it will be outcrossed to produce offspring, half of which will carry the desired mutation on one of their chromosomes. ‘It is then a matter of identifying these heterozygote fish and incrossing them to get homozygous fish in which you can see the phenotype that correlates with that mutation’, says Stemple. As well as helping to produce knock-outs for other researchers, Stemple is also using the TILLING technique to develop zebrafish models for muscular dystrophy. Among the genes that are important in notochord development are those that encode laminins. This led Stemple into studying muscular dystrophy because laminins are involved in the human disease. ‘When we used morpholinos to disrupt [the production of] dystroglycan, a laminin receptor, we got a good model for muscular dystrophy’, he explains. Now, he plans to use TILLING to disrupt up to 30 other genes known to be involved in human muscular dystrophy. ‘In particular, we will look for hypomorphic mutants, fish that are viable but on the edge of falling apart’. These mutants can be used to identify small molecules that push the fish into muscular dystrophy. Finding molecules that can cause a disease in this way ‘might give us a handle on something to fix the disease’, says Stemple. In another strand of the ZF-MODELS project, zebrafish expressing green fluorescent protein (GFP) in specific cells or tissues will be generated and characterised ( Figure 3 ). In such fish, developing structures can be easily imaged over time in the living embryo. One researcher working on this aspect of the project is Stephen Wilson, Professor of Developmental Genetics at University College London (United Kingdom). GFP lines can be made either by attaching to the GFP gene regions of DNA that control, or ‘drive’, GFP expression in selected cell types or by allowing the GFP gene to insert randomly in the genome and looking for fish with specific expression patterns. ‘There are now many lines of fish available with different GFP expression patterns’, says Wilson, ‘and it is important to catalogue their expression so that people can use the most appropriate lines for their research’. Figure 3 Zebrafish Hindbrain (Left) Dorsal view of GFP-expressing neurons in the hindbrain of a one-day-old zebrafish embryo. (Right) Antibody-labelled axons. (Image courtesy of Dave Lyons, University College London.) Wilson's own interest is in neuroanatomy. Together with Jon Clarke, another developmental neurobiology group leader at University College London, he plans to analyse GFP lines in which small groups of neurons or particular parts of neurons are labelled, and in this way start to build a detailed reconstruction of early brain neuroanatomy. This, in combination with other work on zebrafish carrying mutations affecting neural development, will give the team ‘a better picture of how a vertebrate brain is built’. A final, important aspect of the ZF-MODELS project, adds Dahm, is database construction. ‘We will be developing a set of databases that will integrate all of the project data’, he explains. ‘In addition, we hope to integrate our data with that of ZFIN in the United States to make one central zebrafish resource’. But Fish Aren't People The researchers of the ZF-MODELS consortium are understandably excited about participating in what will, says Geisler, bring an already strong European zebrafish community closer together. But zebrafish researchers in the United States are also excited by the ZF-MODELS project. ‘We need big lab models like ZF-MODELS in developmental biology’, says Hopkins, noting that the days of small groups working in isolation are long gone. This consortium, adds Howard Hughes Medical Institute Investigator Leonard Zon (Harvard Medical School, Boston, Massachusetts, United States), ‘will not only have an effect on European zebrafish science but also on how it is done in the United States’. But how much can zebrafish tell us about human development and disease? A lot, say zebrafish researchers. ‘Fish really are just little people with fins’, says Hopkins. ‘Of course, there are developmental differences between people and fish, and no one pretends that we can answer every question about human development in zebrafish’. Nevertheless, zebrafish studies can provide valuable clues to the genes involved in human diseases and to potential targets for therapeutic interventions. Hopkins provides the following illustration: ‘We have been doing “shelf screens”, in which we go back to our collection of mutants to find all those that affect the development of a single organ. When Zhaoxia Sun, a postdoc in my lab who now has an independent position at Yale Medical School, screened three-day-old fish for cystic kidney disease, she found 12 different genes. Two were known to cause human cystic kidney disease, so we knew we were in the human disease pathway somewhere, but we had no idea what the other genes were’. Hopkins and Sun have since identified the remaining genes, and these point to a single pathway being involved in the human disease. Developmental geneticist Didier Stainier (University of California, San Francisco, California, United States) is also using zebrafish to study organ development, in particular, heart development. The zebrafish heart is like the early human heart—a tube with an atrium, ventricle, and valves. ‘Everything we have found in the fish is relevant to the human heart’, says Stainier. ‘Obviously, there are additional processes involved in humans, but the basic outline of heart development in fish and people is largely similar’. Stainier has a collection of zebrafish in which valve formation is faulty. ‘Some of the genes we have found will be involved in human congenital valve defects’, he predicts. Knowing the identity of these genes will be useful diagnostically, but, in addition, zebrafish studies can reveal exactly what has gone wrong at a cellular level. The ability to follow individual cells as organs develop is key to this, says Stainier, who reported in March that fibronectin is required for heart development because, by regulating the polarisation of epithelial cells, fibronectin ensures the correct migration of myocardial cells. And in this issue of PLoS Biology, Stainier's lab have identified another zebrafish gene that is involved in heart development—cardiofunk, which encodes a special type of muscle protein. A Proliferation of Zebrafish Models of Human Disease Many researchers are now recognising the value of zebrafish models of human disease. Over the past three to four years, says Zon, this area of research has become a growth industry. The interest in disease models has grown hand-in-hand with the development of morpholinos to knock out specific genes, and the advent of TILLING, says Zon, ‘has set off a whole new fury. There are now large numbers of investigators who will try to knock out their favourite gene and come up with a model’. Zon has worked on disease models for blood ( Figure 4 ), blood vessel, and heart disorders but is currently studying zebrafish models of cancer. ‘We started by doing chemical mutagenesis and screened for cell-cycle mutants. These were embryonic lethals, but when we looked at heterozygote carriers of these mutations, some developed cancer at a high rate as adults’. Now Zon and his colleagues have returned to the cell-cycle mutant that yielded this cancer-susceptible heterozygote and are using embryos in high-throughput screening assays to look for small molecules that can suppress the cell-cycle phenotype. These molecules, reasons Zon, may have potential as anticancer drugs. Figure 4 Zebrafish kugelig Mutants The image shows live one-week-old zebrafish embryos. The embryos around the outside are wild-type fish. Those in the middle are a mutant called kugelig and have a homozygous mutation in a gene called cdx4 . Loss of the proper functioning of this gene causes the obvious trunk and tail defects but also causes a reduction in the number of haematopoietic stem cells in the embryos, which therefore become severely anaemic. Studies on this mutant might lead to the discovery of molecules that can drive stem cell differentiation, for example, or could help improve understanding of human haematological malignancies. (Image courtesy of Alan Davidson, Harvard Medical School, Boston, Massachusetts, United States.) And the Future of Zebrafish Research? Bigger and bigger seems to be the consensus. Chemical screens like Zon's for anticancer drugs can be set up for other human diseases such as muscular dystrophy. Work like Stainier's on organ development may have applications in tissue engineering. ‘If we can find out what drives differentiation in zebrafish’, he suggests, ‘we might be able to do the same for human cells’, making human tissue replacement therapy a practical possibility. And while many zebrafish researchers will continue to study development, others are now moving into the realms of physiology and behavioural studies. Geisler sums up zebrafish developmental research past, present, and future as follows: ‘No other [vertebrate] organism offers the same combination of transparent and accessible embryos, cost-effective mutagenesis screening, and, more recently, a sequenced genome, [DNA] chip, GFP, and knockout technology’. Add to that the potential of zebrafish embryos as a screening platform for small molecule libraries and the new technologies that allow forward and reverse genetics, and it is clear that zebrafish are not about to revert to being pretty pets swimming in small tanks in the corner of the living room. Where to Find Out More ZF-MODELS More details of the work included in this European Union Integrated Project can be found at http://www.zf-models.org ZFIN The ZFIN Web site, at http://z.n.org/ZFIN , provides an extensive database for the zebrafish community including genetic, genomic, and developmental information; search engines for zebrafish researchers and laboratories; listings of meetings; and links to many other zebrafish sites, including sites with movies of zebrafish development. The special issue of Development (Dec 1; 1996; 123: 1–461) on the first two mutagenesis screens contains 37 research articles and can be freely accessed at http://dev.biologists.org/content/vol123/issue1/index.shtml	/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC406403.xml
314472	Role of Saccharomyces Single-Stranded DNA-Binding Protein RPA in the Strand Invasion Step of Double-Strand Break Repair	The single-stranded DNA (ssDNA)-binding protein replication protein A (RPA) is essential for both DNA replication and recombination. Chromatin immunoprecipitation techniques were used to visualize the kinetics and extent of RPA binding following induction of a double-strand break (DSB) and during its repair by homologous recombination in yeast. RPA assembles at the HO endonuclease-cut MAT locus simultaneously with the appearance of the DSB, and binding spreads away from the DSB as 5′ to 3′ exonuclease activity creates more ssDNA. RPA binding precedes binding of the Rad51 recombination protein. The extent of RPA binding is greater when Rad51 is absent, supporting the idea that Rad51 displaces RPA from ssDNA. RPA plays an important role during RAD51 -mediated strand invasion of the MAT ssDNA into the donor sequence HML . The replication-proficient but recombination-defective rfa1-t11 (K45E) mutation in the large subunit of RPA is normal in facilitating Rad51 filament formation on ssDNA, but is unable to achieve synapsis between MAT and HML . Thus, RPA appears to play a role in strand invasion as well as in facilitating Rad51 binding to ssDNA, possibly by stabilizing the displaced ssDNA.	Introduction Repair of double-strand breaks (DSBs) by homologous recombination involves the search for homology to locate an intact donor sequence. The search is successful when the broken DNA molecule basepairs with the homologous template, termed synapsis, and forms strand invasion intermediates of recombination. In budding yeast and other higher eukaryotes, this process requires both the Rad51 strand exchange protein and the single-stranded DNA (ssDNA)-binding protein replication protein A (RPA) ( Alani et al. 1992 ; Shinohara et al. 1992 ; Ogawa et al. 1993 ; Sung 1994 ; Symington 2002 ). RPA was first discovered through its essential role in SV40 DNA replication in vitro as a ssDNA-binding protein ( Wold et al. 1989 ). The RPA complex forms a heterotrimer, which consists of three subunits of 70, 34, and 14 kDa, encoded by the RFA1 , RFA2 , and RFA3 genes, respectively ( Wold 1997 ). Deletion of any of these genes leads to lethality in yeast ( Heyer et al. 1990 ; Brill and Stillman 1991 ). The biological function of RPA was further demonstrated to be important in homologous recombination in Saccharomyces cerevisiae ( Alani et al. 1992 ; Firmenich et al. 1995 ; Umezu et al. 1998 ) and in other aspects of DNA metabolism. Cells carrying a point mutation (K45E) in the largest subunit of RPA ( rfa1-t11 ) are proficient for DNA replication, but their ability to perform mating-type ( MAT ) gene switching, single-strand annealing, and meiotic recombination is severely impaired ( Umezu et al. 1998 ; Soustelle et al. 2002 ). Because RPA is essential for DNA replication, a great deal about its role in recombination has been learned from in vitro studies of Rad51-mediated strand exchange ( Bianco et al. 1998 ; Symington 2002 ). These studies have shown that RPA facilitates the formation of continuous Rad51 filaments on ssDNA by removing inhibitory secondary structures ( Alani et al. 1992 ; Sugiyama et al. 1997 , 1998 ). A similar requirement is seen in bacteria, where the ssDNA-binding protein SSB apparently plays an analogous role to allow the Rad51 homologue RecA to polymerize across regions that contain secondary structures ( Shibata et al. 1980 ; West et al. 1981 ; Kowalczykowski and Krupp 1987 ; Kuzminov 1999 ). Rad51 further displaces RPA, while RecA displaces SSB, leading to the filament that facilitates the search for homologous double-stranded DNA (dsDNA) sequences and then catalyzes strand invasion and the formation of a displaced single strand ( Kowalczykowski et al. 1987 ; New et al. 1998 ; Eggler et al. 2002 ; Sugiyama and Kowalczykowski 2002 ). However, order-of-addition experiments have suggested that if RPA/SSB is added to ssDNA prior to Rad51/RecA, successful displacement will not occur because RPA/SSB has higher affinity for ssDNA, unless mediator proteins, such as Rad52 and Rad55/Rad57 in yeast and RecO/RecR in bacteria, are present ( Umezu et al. 1993 ; New et al. 1998 ; Shinohara and Ogawa 1998 ; Kuzminov 1999 ; Sugiyama and Kowalczykowski 2002 ; Symington 2002 ). But if Rad51/RecA has polymerized onto ssDNA first under conditions that prevent the formation of secondary structures, further addition of RPA/SSB will stimulate in vitro strand exchange in a species-specific manner ( Heyer and Kolodner 1989 ; Morrical and Cox 1990 ; Sung 1994 ; Sugiyama et al. 1997 ). Using the same in vitro system, Kantake et al. (2003 ) have examined the effects of the rfa1-t11 (Rfa1-K45E) mutation on strand exchange with Rad51. Although rfa1-t11 protein bound to ssDNA identically to wild-type and could stimulate strand exchange if Rad51 was preloaded onto ssDNA, the mutant protein exhibited delayed and less-efficient strand exchange if it was first bound to ssDNA, especially at higher concentrations, even in the presence of Rad52. This defect was explained by the slow displacement of rfa1-t11 from ssDNA by Rad51. Recently, immunostaining experiments have been carried out in S. cerevisiae as well as in higher eukaryotes to investigate the association of RPA to DSBs following γ-irradiation and during meiosis. These studies have suggested that RPA and Rad51 form subnuclear foci at sites of ssDNA after irradiation and during meiotic recombination ( Gasior et al. 1998 ; Golub et al. 1998 ; Raderschall et al. 1999 ) and that RPA is recruited to these sites prior to Rad51 ( Golub et al. 1998 ; Gasior et al. 2001 ). In order to understand better how RPA is involved in DSB repair in vivo, we have looked at its function in MAT switching in yeast, which is the well-studied example of DSB-induced homologous recombination ( Haber 2002a ). MAT switching is initiated when HO endonuclease creates a site-specific DSB at the MAT locus, which is then repaired by gene conversion using one of the two heterochromatic donor sequences, HML or HMR ( Pâques and Haber 1999 ; Haber 2000 , Haber 2002a ). By using a galactose-inducible HO endonuclease gene ( Jensen et al. 1983 ), the induction of the DSB and its repair occur synchronously in a population of cells so that the kinetics of DSB repair and the appearance of intermediates of recombination can be followed by physical monitoring of the process via Southern blot and PCR assays ( Haber 1995 , Haber 2002a , Haber 2002b ). To learn more precisely about how RPA participates in homologous recombination in vivo, we have used chromatin immunoprecipitation (ChIP) assays ( Dedon et al. 1991 ; Sugawara et al. 2003 ) to analyze the association of RPA and Rad51 to DNA as it undergoes MAT switching. At the same time, the fate of recombining DNA was analyzed by Southern blot and PCR techniques ( White and Haber 1990 ; Haber 1995 ). The combination of these approaches has enabled us to visualize the kinetics and extent of RPA binding to a DSB and the homologous template during recombination. We report that the biological function of RPA is also required during the strand invasion step of recombination. Rfa1-t11 mutant cells are not defective in Rad51 nucleoprotein filament assembly, as observed by ChIP, but are incapable of performing the strand exchange and thus the completion of DSB repair. Results Kinetics and Extent of RPA Binding at DSB Ends in the Absence of DNA Repair In wild-type yeast cells, a DSB created at the MAT locus can be repaired by gene conversion with one of the two donor sequences, HML or HMR , or the DSB can be left unrepaired in most cells by deleting these donors ( Haber 2002a ). In order to monitor RPA binding to DSB ends, we first performed ChIP analysis on strains in which both of the donor loci were deleted so that the DSB at MAT could not be repaired and 5′ to 3′ exonuclease activity would generate resected ssDNA unimpeded for many hours ( Lee et al. 1998 ). In these strains, nearly complete cutting of MAT by the galactose-induced HO endonuclease occurred within 20 min after induction (see below). In wild-type cells, after HO induction, significant RPA binding to sequences close to the HO cleavage site was seen by ChIP ( Figure 1 A and 1 B), using a pair of primers (P1 and P2) that amplify sequences 189 bp to 483 bp distal to the HO cut ( Figure 1 A). As shown in Figure 2 , RPA was recruited to DSB ends as soon as the DSB could be detected on a Southern blot (20 min after induction). The binding of RPA increased for about 2–3 h, until presumably all sequences near MAT had been rendered single-stranded ( Frank-Vaillant and Marcand 2002 ) (see Figure 1 B). At later times, one detects RPA binding at increasing distances from the cleavage site, as these regions were rendered single-stranded by 5′ to 3′ exonuclease activity ( Lee et al. 1998 ) (see Figure 1 C). Figure 1 Recruitment of RPA to a DSB in the Absence of DNA Repair A strain deleted for donors (yXW1), thus incapable of repairing a DSB by gene conversion, was pregrown in YP–lactate medium, and 2% galactose was added to the culture to induce a DSB at MAT . DNA was extracted at intervals after HO cutting, to which polyclonal antibody against Rfa1 was applied to immunoprecipitate RPA-bound chromatin. Another set of DNA samples were taken at the same time for Southern blot analysis. (A) Map of MAT showing the locations of the HO-cut site as well as the StyI restriction sites and the primers (P1 and P2), 189 bp to 483 bp distal to the DSB, used to PCR-amplify RPA-associated MAT DNA from the immunoprecipitated extract. Purified genomic DNA was digested with StyI, separated on a 1.4% native gel, and probed with a 32 P-labeled MAT distal fragment to monitor the appearance of the HO-cut fragment (see Materials and Methods). The 1-h timepoint represents 1 h after galactose induction of the HO endonuclease. (B) PCR-amplified RPA-bound MAT DNA in a wild-type strain (yXW1). As controls, primers to an independent locus, ARG5,6 (see Materials and Methods), were used to amplify DNA from the immunoprecipitated chromatin. PCR samples were run on ethidium bromide-stained gels (reverse images are shown). Quantitated signals were graphed for the wild-type strain. IP represents ratio of the MAT IP signal to ARG5,6 IP signal. Error bars show one standard deviation. (C) RPA-bound chromatin was PCR-amplified from sites located proximal and distal to the DSB and then quantitated and graphed as described in (B). The DSB is shown at 0 bp. (D) Effect of formaldehyde cross-linking on RPA binding to ssDNA. In both the noncross-linked samples and the cross-linked samples, 4 ng of single-stranded heterologous β-lactamase ( AMP ) gene DNA was added during the extract preparation step of ChIP. The amount of input genomic and heterologous DNA was measured by PCR primers specific to the ARG5,6 locus and to the AMP sequence, respectively. RPA-associated ARG5,6 and AMP DNA were analyzed from the IP samples. PCR samples were run on ethidium bromide-stained gels (reverse images are shown). Figure 2 Timing of Recruitment of RPA versus Rad51 to the DSB An unrepairable DSB was created in the wild-type strain (yXW1), and closer timepoints were harvested at 20 min and 30 min after the HO cut. DNA samples extracted at each timepoint were split. One half was applied with antibody against Rfa1 to immunoprecipitate RPA-associated DNA, while the other half was applied with anti-Rad51 antibody to immunoprecipitate Rad51-bound chromatin. RPA- or Rad51-associated MAT DNA was PCR-amplified and run on ethidium bromide-stained gels (reverse images are shown). DNA signals were quantitated and graphed as described in Figure 1 for RPA ChIP. PCR-amplified ARG5,6 signals from the input DNA were used as controls for quantitation and graphing for Rad51 ChIP (see Materials and Methods). In carrying out the ChIP measurements, we were aware that RPA is both abundant within the cell and binds strongly and cooperatively to ssDNA in vitro ( Heyer and Kolodner 1989 ). It was possible that some of the RPA binding we measured by ChIP could have arisen after the cells were broken and could be independent of formaldehyde cross-linking. Indeed, in the absence of cross-linking, we found that there was substantial binding of RPA to the HO-cut MAT locus, which was resistant to both addition of 2 mg of ssDNA (equivalent to a 1,000-fold genome excess) at the time of cell breakage and washing with 4.7 M NaCl, although it greatly reduced background binding (data not shown). However, in formaldehyde cross-linked samples, there was no such adventitious binding of RPA to ssDNA regions, apparently because the formaldehyde-treated proteins are no longer able to bind. This was shown directly by adding 4 ng of purified single-stranded β-lactamase ( AMP ) gene DNA from plasmid pBR322 at the time of cell breakage. Whereas there was substantial ChIP of the AMP sequences in noncross-linked samples, there was almost no signal in cells that had first been treated with formaldehyde (see Figure 1 D). RPA Binding Precedes the Binding of the Strand Exchange Protein Rad51 In vitro studies of the early steps of recombination have suggested that in order to make a continuous and functional nucleoprotein filament, RPA must bind before Rad51 to ssDNA to remove inhibitory secondary structures ( Sugiyama et al. 1997 ). Indirect immunofluorescence experiments in S. cerevisiae have also suggested that RPA assembles before Rad51 at DSBs after γ-irradiation ( Gasior et al. 2001 ). Therefore, the timing of recruitment of both RPA and Rad51 proteins to a DSB in vivo were compared by ChIP. RPA was detected at MAT 20 min after the HO cut, while Rad51 binding was not observed until the 30 min timepoint ( Figure 2 ). Similar results were obtained in strains that are able to carry out gene conversion (see Figure 4 ). These observations strongly support the idea that RPA binding to HO-cut DNA precedes that of Rad51. Figure 4 Localization of RPA and Rad51 to HML and MAT during DSB-Induced Gene Conversion A strain carrying an HML α donor (yXW2), thus able to repair the DSB at MAT by gene conversion, was treated with 2% galactose to induce HO endonuclease and then with 2% glucose after 1 h to repress further HO expression. DNA extracted at intervals after HO cutting was split. One half was applied with antibody against Rfa1 to immunoprecipitate RPA-associated DNA, while the other half was applied with anti-Rad51 antibody to immunoprecipitate Rad51-bound chromatin. Another set of DNA samples were taken at the same time for Southern blot analysis. (A) Diagram of MAT and HML showing the locations of the primers, 189 bp to 483 bp distal to the DSB at MAT (P1 and P2) and 189 bp to 467 bp from the uncleaved HO recognition site at HML (P1 and P3), used to PCR-amplify RPA- and Rad51-associated MAT and HML DNA from the immunoprecipitated extract. (B) Purified genomic DNA was digested with StyI, separated on a 1.4% native gel, and probed with a 32 P-labeled MAT distal fragment to monitor the appearance of the HO-cut fragment and the repaired product Yα (see Figure 1 A; see Materials and Methods). The arrowhead indicates the switched product Yα. RPA- and Rad51-bound MAT and HML DNA were PCR-amplified with primers P1 and P2 and with P1 and P3, respectively. Samples were run on ethidium bromide-stained gels. (C and D) Reverse images are shown for RPA ChIP (C) and Rad51 ChIP (D). DNA signals were quantitated and graphed as described in Figure 2 . Error bars show one standard deviation. In Vivo Competition between RPA and Rad51 for ssDNA Studies of RPA in vitro would suggest that the amount of RPA bound to ssDNA may be limited by its displacement by Rad51, through the help of Rad52, and the Rad55/57 heterodimer ( Sung 1997a , 1997b ; New et al. 1998 ; Shinohara and Ogawa 1998 ; Sugiyama and Kowalczykowski 2002 ). To test this idea, we deleted RAD51 and measured RPA binding at MAT . The extent of RPA binding was approximately 5- to 6-fold higher in the rad51 Δ strain, consistent with this expectation ( Figure 3 ). A similar result was found in a rad52 Δ strain ( Figure 3 ), supporting the hypothesis that the displacement of RPA by Rad51 depends on Rad52, which acts as a mediator between these two ssDNA-binding proteins ( Sung 1997a ; New et al. 1998 ; Song and Sung 2000 ; Sugiyama and Kowalczykowski 2002 ; Sugawara et al. 2003 ). Figure 3 Effect of rad51 Δ and rad52 Δ on the Extent of RPA Binding to an Unrepairable DSB An unrepairable DSB was created in wild-type (yXW1), rad51 Δ (ySL306), and rad52 Δ (ySL177) strains and RPA-bound chromatin was immunoprecipitated using anti-Rfa1 antibody. PCR-amplified DNA from the MAT locus was run on ethidium bromide-stained gels (reverse images are shown). DNA signals were quantitated and graphed as described in Figure 1 . Error bars show one standard deviation. RPA Is Recruited to Both the Donor and the Recipient Sequences during Gene Conversion We then examined RPA in a strain in which the DSB at MAT a could be repaired by gene conversion, using HML α as the donor ( Figure 4 A). As soon as the DSB was visible, an increase in RPA binding was seen (Figure 4 B and 4 C). RPA binding increased for about 1 h and then decreased nearly to the baseline level about the time that MAT switching was completed (Figure 4 B and 4 C). Importantly, RPA also appeared to become associated with the donor locus. This was detected by ChIP using the same primer P1 located in the Z region shared by MAT and HML and an HML sequence-specific primer P3 ( Figure 4 A). Whereas RPA could be found associated with MAT 20 min after HO induction, its association with HML was seen only after 1 h ( Figure 4 C). The association of RPA with HML came at about the same time as we saw synapsis between HML and MAT as revealed by ChIP with anti-Rad51 antibody ( Figure 4 D; also Sugawara et al. 2003 ). The extent of RPA binding to HML was substantially less than seen at MAT ( Figure 4 C), where ssDNA may extend further than the 320 bp of homology between MAT and HML ; these more distal ssDNA sequences would not be involved directly in recombination. The lower amount of RPA binding at the donor locus may also arise from a lower concentration of RPA that is needed at the sites of strand invasion or a transient presence of RPA at those loci. But the fact that cross-linked RPA can immunoprecipitate the donor locus might indicate that RPA is recruited onto the single-stranded D-loop that is created by strand invasion. This would be consistent with in vitro studies of Rad51-mediated strand exchange that suggest that strand invasion per se can occur without RPA, but that the heteroduplex DNA is unstable unless the displaced strand is bound by RPA ( Eggler et al. 2002 ). A similar requirement for SSB was suggested in RecA-mediated strand invasion ( Lavery and Kowalczykowski 1992 ). We cannot entirely rule out the possibility that the apparent association of RPA with HML resulted from the cross-linking of synapsed MAT and HML sequences directly or through Rad51-containing cross-links and where RPA was bound to ssDNA sequences distal to the 320-bp homology shared by MAT and its donor. Further evidence of a role for RPA in synapsis will be presented below. To understand better the dynamics between RPA and Rad51 during gene conversion, we also examined Rad51 recruitment to MAT and HML relative to that of RPA ( Figure 4 D). Rad51 was only detected at MAT 30 min postinduction (compared to 20 min for RPA). Rad51 binding increased for about 1 h and then remained bound for several hours. As reported previously ( Sugawara et al. 2003 ), Rad51 showed a delayed association with the donor HML , reflecting the time required to form a functional filament and to search the genome for homologous sequences. These observations provide evidence of the time at which synapsis between MAT and HML is achieved. Here, too, Rad51 association with the donor remained for several hours, beyond the time when MAT switching is completed. rfa1-t11 Mutant Cells Are Defective in the Synapsis Step of Gene Conversion To learn more about RPA function during recombination, we investigated the behavior of the rfa1-t11 (K45E) mutation in the largest subunit of RPA. This mutation has little effect on DNA replication per se, but severely impairs both gene conversion ( MAT switching) and single-strand annealing pathways of homologous recombination ( Umezu et al. 1998 ) (also Figure 6 A). Cells containing this mutation displayed hyperresection at meiotic DSB ends and defects in the repair of these DSBs ( Soustelle et al. 2002 ). In vitro biochemical studies have shown that rfa1-t11 is displaced from ssDNA by Rad51 more slowly than wild-type RPA, and as a consequence, Rad51-mediated strand exchange is inhibited when the ssDNA is complexed with the mutant RPA heterotrimer ( Kantake et al. 2003 ). Here, we examined binding of Rfa1-K45E in vivo by ChIP and also its effect on Rad51 localization to an HO-induced DSB, using the same antibodies as against wild-type RPA. Figure 6 rfa1-t11 Was Not Able to Associate with the Donor Sequence during Gene Conversion The wild-type strain carrying the HML α donor (yXW2) and an isogenic strain carrying the rfa1-t11 mutation (yXW3) were treated with 2% galactose to induce HO endonuclease and then with 2% glucose after 1 h to repress further HO expression. DNA extracted at intervals after HO cutting was split. One half was applied with antibody against Rfa1 to immunoprecipitate RPA-associated DNA, while the other half was applied with anti-Rad51 antibody to immunoprecipitate Rad51-bound chromatin. Another set of DNA samples was taken at the same time for Southern blot analysis. (A) Purified genomic DNA was digested with StyI, separated on a 1.4% native gel, and probed with a 32 P-labeled MAT distal fragment to monitor the appearance of the HO-cut fragment and the repaired product Yα (see Figure 1 A; see Materials and Methods). Arrowheads indicate the switched product Yα. (B) RPA-bound MAT and HML DNA was PCR-amplified with primers P1 and P2 and with P1 and P3, respectively (see Figure 4 A). Samples were run on ethidium bromide-stained gels (reverse images are shown). DNA signals were quantitated and graphed as described in Figure 1 . In a strain lacking HML and HMR , Rfa1-K45E binding was nearly identical to wild-type, both in a RAD51 and in a rad51 Γ background ( Figure 5 A). Moreover, binding of Rad51 to ssDNA at the HO-cut MAT locus was also comparable to that observed in wild-type cells ( Figure 5 B). Thus, the Rfa1-t11 protein is neither impaired in loading onto ssDNA, nor does it affect the loading of Rad51 in vivo. Figure 5 rfa1-t11 Mutation Does Not Affect the Recruitment of Itself or Rad51 to an Unrepairable DSB (A) An unrepairable DSB was created in wild-type (yXW1), rfa1-t11 (ySL31), rad51 Δ (ySL306), and rfa1-t11 rad51 Δ (ySL351) strains, and half of the DNA sample was immunoprecipitated with anti-Rfa1 antibody to extract rfa1-t11 -bound chromatin. (B) For wild-type (yXW1) and rfa1-t11 (ySL31) strains, the other half of the DNA sample was applied with anti-Rad51 antibody to extract Rad51-associated chromatin. PCR-amplified DNA from the MAT locus was run on ethidium bromide-stained gels (reverse images are shown). DNA signals were quantitated and graphed as described in Figure 2 . We then examined the effect of rfa1-t11 during HO-induced switching of MAT a to MAT α, using HML α as the donor. As shown previously ( Umezu et al. 1998 ), rfa1-t11 strongly impaired MAT switching, with only 15% product evident after 5 h ( Figure 6 A). RPA bound normally to the MAT locus, but unlike what occurs in wild-type strains, its binding remained undiminished at later times ( Figure 6 B). Moreover, there was no increased association of RPA with HML over background levels ( Figure 6 B). When we examined Rad51 binding in this mutant, we found that Rad51 immunoprecipitated with MAT DNA, but not with HML ( Figure 7 A). In support of this important finding, we also used a PCR assay to show that rfa1-t11 prevented the appearance of newly synthesized DNA using the 3′ end of the invading strand as a primer ( Figure 7 B). In this assay, a primer specific for the Yα region in HML (pA) can only amplify a strand invasion product with a primer specific for MAT -distal sequences (pB) if the 3′ end of the strand-invading DNA is extended by DNA polymerase at least 35 nucleotides ( White and Haber 1990 ) ( Figure 7 B). These data strongly raise the possibility that RPA is required during the process of strand invasion and synapsis and not merely to facilitate formation of a Rad51 filament, as the binding of Rad51 to ssDNA at MAT seems to be normal in both kinetics and extent. Figure 7 rfa1-t11 Mutants Are Defective in the Strand Invasion Step of Gene Conversion (A) One half of the DNA extract collected from a typical timecourse experiment as described in Figure 6 was applied with anti-Rad51 antibody to immunoprecipitate Rad51-bound chromatin. Primers P1 and P2 and P1 and P3 were used to PCR-amplify Rad51-bound MAT and HML DNA, respectively (see Figure 4 A). Samples were run on ethidium bromide-stained gels (reverse images are shown). DNA signals were quantitated and graphed as described in Figure 2 . (B) Input DNA was used to PCR-amplify strand invasion product using a unique primer distal to MAT (pB) and a primer within the Yα sequence from HML (pA) ( White and Haber 1990 ). PCR-amplified ARG5,6 signals from the input DNA were used as loading control. Discussion ChIP analysis provides a powerful tool for studying in vivo protein–DNA and protein–protein interactions. Using ChIP and related assays, we have demonstrated important roles of RPA during homologous recombination in vivo that could not have been known with certainty from in vitro studies. RPA is recruited to the DSB ends as soon as the DSB is detected on a Southern blot, and its binding precedes that of Rad51 (see Figures 2 and 4 ), which supports the idea that RPA is required to remove inhibitory secondary structures on ssDNA for Rad51 to polymerize across these regions ( Sugiyama et al. 1997 , 1998 ). This observation is also consistent with in vivo immunofluorescent staining results, suggesting that RPA foci appear earlier than Rad51 foci after irradiation ( Golub et al. 1998 ; Gasior et al. 2001 ). Rad51 apparently displaces RPA from ssDNA, with the help of Rad52 (see Figure 3 ) and perhaps the Rad55/Rad57 auxiliary proteins. We note that our results are different from those reported by Wolner et al. (2003 ), who observed initial binding of RPA only after 45 min, whereas Rad51 was detected 25 min earlier, although it is not clear whether there is a statistically significant increase in Rad51 binding at the earliest time. In that assay, RFA1 was tagged with 13 Myc epitope tags, which may have altered its behavior. We believe our results are consistent with the fact that RPA has a higher-affinity constant for ssDNA and is present in much greater abundance in the cell ( Heyer and Kolodner 1989 ; Mazin et al. 2000 ; Sugawara et al. 2003 ). We noticed that when RPA ChIP was carried out in donorless strains as well as in rfa1-t11 strains that carry the donor loci, there was a continued presence of some RPA near the ends of a DSB. This may occur for several reasons. First, it is likely that the formation and maintenance of the Rad51 filament are a dynamic process, with subunits coming off the end and perhaps being replaced by RPA before being in turn replaced by Rad51. Second, the Rad51 nucleoprotein filament may not be, in vivo, a fully continuous structure, given that there are only about 3,500 monomers of Rad51 in the cell and that Rad51 binding is not highly cooperative ( Mazin et al. 2000 ; Sugawara et al. 2003 ). Finally, the very ends of the DSB can religate and be recleaved by HO in a cycle that lasts several hours in the absence of donor sequences (and hence in the absence of homologous recombination) to repair the DSB ( Frank-Vaillant and Marcand 2002 ). Thus, a fraction of molecules will be newly generated and will show RPA binding before Rad51, as we saw for the initial DSB. The ChIP analysis presented here has shown that RPA is required for homologous recombination even after Rad51 has bound to ssDNA. First, RPA can immunoprecipitate donor sequences, the timing of which coincides with the loading of Rad51 at HML (see Figure 4 ). Second, the replication-proficient but recombination-deficient mutant of the largest subunit of RPA ( rfa1-t11 ) is able to allow Rad51 to bind to ssDNA, but is incapable of forming normal levels of strand invasion and primer extension products (see Figures 6 and 7 ). We offer two possible explanations for this unexpected finding. First, whereas Rad51 can bind to ssDNA in rfa1-t11 , it may not establish a functional filament capable of carrying out a search for homology and strand invasion, even though the association of Rad51 with ssDNA appears to be normal. But the defect in cells with rfa1-t11 seems different from that seen in cells lacking Rad55 ( Sugawara et al. 2003 ), where there was delayed and less-extensive binding of Rad51 to ssDNA; moreover, although Rad51 eventually bound, it was unable to catalyze synapsis between MAT ssDNA and HML . In rad55 Γ cells, it is likely that the Rad51 filament is discontinuous and unable to function. However, with rfa1-t11 , the loading of Rad51 onto ssDNA appears to be identical to that seen in wild-type cells (see Figures 5 B and 7 A). Alternatively, in rfa1-t11 cells, the filament may indeed be functional, but RPA is needed to stabilize the strand invasion intermediate and rfa1-t11 is unable to carry this out. RPA may be required to bind to the displaced D-loop, to prevent rapid reversal of the process, which is implicated by in vitro studies of strand exchange ( Eggler et al. 2002 ). In that study, extensive heteroduplex could be formed without RPA, as revealed by psoralen cross-linking of joint molecule DNA before removal of Rad51 by deproteinization, but without cross-linking, the deproteinized joint molecule DNA fell apart into the original single-stranded and double-stranded substrates very quickly. An analogous role for SSB has been suggested in RecA-mediated strand invasion ( Lavery and Kowalczykowski 1992 ), in which SSB prevents the reversal of DNA strand exchange by removing the displaced single strand. It is possible that the Rfa1-K45E mutation renders the mutant RPA complex unable to bind to the displaced ssDNA at HML and thus unable to carry out strand exchange, while binding to MAT ssDNA that has a free 3′ end tail is not affected. Both in vitro ( Kantake et al. 2003 ) and in vivo analyses showed that rfa1-t11 was able to bind to ssDNA very similarly to wild-type, but our in vivo data did not see any significant impairment of its Rad52-mediated displacement by Rad51. It should be noted that the inhibition of Rad51-mediated strand exchange by rfa1-t11 in vitro was carried out with saturating amounts of Rad51 (whereas the amount of Rad51 in the cell is quite limited) and that the inhibition of Rad51-mediated strand exchange was impaired primarily when RPA was present in excess ( Kantake et al. 2003 ). How these conditions relate to those prevailing in vivo remains unknown. In this regard, it is also noteworthy that in vitro studies did not see any impairment of single-strand annealing ( Kantake et al. 2003 ), whereas in vivo, single-strand annealing is nearly eliminated in rfa1-t11 strains ( Umezu et al. 1998 ). Further comparisons of in vitro and in vivo data will be valuable in understanding how the more complex environment within the cell affects processes of recombination. Materials and Methods Strains Donorless strains are isogenic derivatives of JKM139, which has the genotype of ho Δ hml Δ:: ADE1 MAT a hmr Δ:: ADE1 ura3–52 leu2–3,112 trp1 :: hisG lys5 ade1–100 ade3 :: GAL :: HO . The wild-type strain yXW1 was constructed by transforming JKM139 with pGI4 ( bar1 :: ADE3 ) ( Wach et al. 1994 ). ySL83 contains yku80 Δ:: KAN and bar1 :: TRP1 . ySL306 and ySL177 contain rad51 Δ:: URA3 and rad52 Δ:: TRP1 , respectively ( Lee et al. 1998 , 2001). ySL31 has the point mutation (K45E) in the largest subunit of RPA ( Lee et al. 1998 ), and ySL351 was derived from ySL31 and contains rad51 Δ:: LEU2. Strains capable of undergoing DSB-induced gene conversion were derived from OAy470 ( Aparicio et al. 1997 ), which has the genotype of ho Δ MAT a ura3–1 trp1–1 leu2–3,112 his3–11,15 ade2–1 can1–100 bar1 :: hisG . A galactose-inducible GAL :: HO gene was integrated at ADE3 of OAy470 using YIPade3HO constructed by L. L. Sandell ( Sandell and Zakian 1993 ) to obtain the wild-type strain yXW2. yXW3 is an isogenic derivative of yXW2, into which the point mutation (K45E) of Rfa1 was introduced by integration and excision of a YIp5 ( URA3 -containing) plasmid ( Lee et al. 1998 ). DNA analysis When cells were harvested for ChIP at intervals after induction of HO (see below), a second set of DNA samples were collected for Southern blot analysis as described before ( White and Haber 1990 ). The strand invasion/primer extension assay in Figure 7 B was previously described ( White and Haber 1990 ). The primers used were 5′-GCAGCACGGAATATGGGACT-3′ (pA) and 5′-ATGTGAACCGCATGGGCAGT-3′ (pB). ChIP ChIP was carried out as described previously with minor modifications ( Dedon et al. 1991 ; Sugawara et al. 2003 ). Cells were pregrown to a density between 5 × 10 6 and 1 × 10 7 cells/ml in YP–lactate medium and HO endonuclease was induced by addition of 2% galactose. Strains undergoing DSB-induced gene conversion were treated with 2% glucose after 1 h to repress further cutting by HO. Proteins were cross-linked by addition of 1% (final concentration) formaldehyde to 45 ml of culture for 10 min, followed by quenching with 125 mM glycine (final concentration) for 5 min. Cells were lysed with glass beads, and the extracts were sonicated to shear the DNA to an average size of 0.5 kb. Extracts were then divided into IP and input samples (12:1 ratio). IP samples were split. Half of the extract was incubated with polyclonal anti-Rfa1 antibody (kindly provided by S. Brill) for 1 h at 4°C and bound to protein G–agarose beads for 1 h at 4°C. In the ChIP experiments described in Figure 2 , Figure 4 D, Figure 5 B, Figure 6 , and Figure 7 , the other half of the extract was incubated with affinity-purified anti-Rad51 antibody (provided by P. Sung) or unpurified antibody (provided by A. Shinohara) for 1 h at 4°C and bound to protein A–agarose beads for 1 h at 4°C. The protein-bound beads were carried through a series of washes, followed by elution of the proteins and reversal of cross-linking (6 h at 65°C). Samples were treated with proteinase K followed by phenol extraction and ethanol precipitation. In the control experiments described in Figure 1 D, 4 ng of purified single-stranded β-lactamase ( AMP ) gene DNA from plasmid pBR322 was added at the time of cell breakage. IP and input samples were further subject to PCR to test the presence of the AMP sequences. PCR amplification The MAT -specific primers were 5′-TCCCCATCGTCTTGCTCT-3′ (P1) and 5′-GCATGGGCAGTTTACCTTTAC-3′ (P2), which amplifies a PCR product of 293 bp. The HML -specific primers were 5′-TCCCCATCGTCTTGCTCT-3′ (P1) and 5′-CCCAAGGCTTAGTATACACATCC-3′ (P3), which amplifies a PCR product of 280 bp. Primers used for the amplification of the sites proximal to the DSB (see Figure 1 C) were −29.8 kb, 5′-TCGTCGTCGCCATCATTTTC-3′ and 5′-GCCCAAGTTTGAGAGAGGTTGC-3′; −16.6 kb, 5′-CGTCTTCTCAGCGAACAACAGC-3′ and 5′-GCAATAACCCACGGAAACACTG-3′; −9.5 kb, 5′-TCAGGGTCTGGTGGAAGGAATG-3′ and 5′-CAAAGGTGGCAGTTGTTGAACC-3′; −5.3 kb, 5′-ATTGCGACAAGGCTTCACCC-3′ and 5′-CACATCACAGGTTTATTGGTTCCC-3′; −3.6 kb, 5′-ATTCTGCCATTCAGGGACAGCG-3′ and 5′-CGTGGGAAAAGTAATCCGATGC-3′; −1.6 kb, 5′-ATGTCCTGACTTCTTTTGACGAGG-3′ and 5′-ACGACCTATTTGTAACCGCACG-3′; and −0.2 kb, 5′-AAAGAAGAAGTTGCAAAGAAATGTGG-3′ and 5′-TGTTGCGGAAAGCTGAAACTAAAAG-3′. Oligos used for the sites distal to the DSB were 0.2 kb, 5′-CCTGGTTTTGGTTTTGTAGAGTGG-3′ and 5′-GAGCAAGACGATGGGGAGTTTC-3′; 2.1 kb, 5′-GCCTCTATGTCCCCATCTTGTCTC-3′ and 5′-GTGTTCCCGATTCAGTTTGACG-3′; 3.1 kb, 5′-TAACCAGCAATACCAAGACAGCAC-3′ and 5′-TTTTACCTACCGCACCTTCTAAGC-3′; 5.7 kb, 5′-CCAAGGAACTAATGATCTAAGCACA-3′ and 5′-ACCAGCAGTAATAAGTCGTCCTGA-3′; and 9.5 kb, 5′-TGGATCATGGACAAGGTCCTAC-3′ and 5′-GGCGAAAACAATGGCACTCT-3′. These PCR primers gave products of about 300 bp. Primers specific for the ARG5,6 locus were either 5′-AGAAAGGGGGTATTATCAATGGCTC-3′ and 5′-AGGAAAATCACGGCGCAAAA-3′, which amplifies a PCR product of 533 bp, or 5′-CAAGGATCCAGCAAAGTTGGGTGAAGTATGGTA-3′ and 5′-GAAGGATCCAAATTTGTCTAGTGTGGGAACG-3′, which amplifies a PCR product of 381 bp. Normalization using these two different pairs of primers has been shown not to affect the final quantification results. Primers used for the amplification of the AMP sequences (see Figure 1 D) were 5′-GAAGACGAAAGGGCCTCGTG-3′ and 5′-GCTGCAGGCATCGTGGTGTC-3′, which amplifies a PCR product of 750 bp. All PCR assays were accompanied by reactions using dilutions of the 0-h input sample to assess the linearity of the PCR signal and to create a calibration curve, as described before ( Sugawara et al. 2003 ). Samples were run on ethidium bromide-stained agarose gels (1.4%) and quantitated using an Innotech Alphaimager™ and Quantity One software™ (BioRad, Hercules, California, United States), which was also used to correct for minor deviations from a linear response in signal. Quantification and graphing were carried out as described previously with minor changes ( Sugawara et al. 2003 ). For RPA ChIP analysis, all IP samples were first normalized to IP signals from an independent locus ( ARG5,6 ) on chromosome V in a multiplex experiment, by using ARG5,6 and MAT or HML primers in the same PCRs. This was accomplished by dividing each MAT or HML IP signal by the corresponding ARG5,6 IP signal to correct for differing amounts of chromatin collected at each timepoint. Then MAT or HML IP signals at later timepoints were normalized and graphed to the 0-h IP signal to measure the relative increase. For Rad51 ChIP analysis, quantification and graphing were carried out as described before ( Sugawara et al. 2003 ), in which all IP samples were normalized to the ARG5,6 input signals at the respective time points. Graphing represents the average of at least three independent ChIP timecourse experiments for each strain. Supporting Information Accession Numbers The Saccharomyces Genome Database ( http://www.yeastgenome.org/ ) ID accession numbers for the entities discussed in this paper are ARG5,6 (S0000871), HML (L0000791), HMR (L0000792), MAT (L0001031), Rad51 (S0000897), Rad52 (S0004494), Rad55 (S0002483), Rad57 (S0002411), RFA1 (S0000065), RFA2 (S0005256), and RFA3 (S0003709).	/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC314472.xml
516016	The Hotdog fold: wrapping up a superfamily of thioesterases and dehydratases	Background The Hotdog fold was initially identified in the structure of Escherichia coli FabA and subsequently in 4-hydroxybenzoyl-CoA thioesterase from Pseudomonas sp. strain CBS. Since that time structural determinations have shown a number of other apparently unrelated proteins also share the Hotdog fold. Results Using sequence analysis we unify a large superfamily of HotDog domains. Membership includes numerous prokaryotic, archaeal and eukaryotic proteins involved in several related, but distinct, catalytic activities, from metabolic roles such as thioester hydrolysis in fatty acid metabolism, to degradation of phenylacetic acid and the environmental pollutant 4-chlorobenzoate. The superfamily also includes FapR, a non-catalytic bacterial homologue that is involved in transcriptional regulation of fatty acid biosynthesis. We have defined 17 subfamilies, with some characterisation. Operon analysis has revealed numerous HotDog domain-containing proteins to be fusion proteins, where two genes, once separate but adjacent open-reading frames, have been fused into one open-reading frame to give a protein with two functional domains. Finally we have generated a Hidden Markov Model library from our analysis, which can be used as a tool for predicting the occurrence of HotDog domains in any protein sequence. Conclusions The HotDog domain is both an ancient and ubiquitous motif, with members found in the three branches of life.	Background We have found the HotDog domain, as we suggest calling the Hotdog fold, to be widespread in eukaryotes, bacteria, and archaea and to be involved in a range of cellular processes, from thioester hydrolysis, to phenylacetic acid degradation and transcriptional regulation of fatty acid biosynthesis. We present the superfamily and its functional subfamilies here. The Hotdog fold was first observed in the structure of Escherichia coli β-hydroxydecanoyl thiol ester dehydratase (FabA), where Leesong et al . noticed that each subunit of this dimeric enzyme contained a mixed α + β 'hot dog' fold [ 1 ]. They described the seven-stranded antiparallel β-sheet as the 'bun', which wraps around a five-turn α-helical 'sausage', see Figure 1 . This characteristic fold has been found in a number of other enzymes, including: 4-hydroxybenzoyl-CoA thioesterase (4HBT) from Pseudomonas sp. strain CBS-3 [ 2 ] and Arthrobacter sp. strain SU [ 3 ], a novel gentisyl-CoA thioesterase from Bacillus halodurans [ 4 ] and in Escherichia coli thioesterase II [ 5 ]. Results and Discussion Although several proteins are now known to contain a Hotdog fold from structural analysis it has not to our knowledge been demonstrated that these proteins can be related to each other by sequence similarity. We have attempted to unify these structurally related proteins using a sequence analysis approach. Using sequence analysis means that we will identify additional proteins that are likely to contain a Hotdog fold. We have used the PSI-BLAST program [ 6 ] and used a representative of each Hotdog fold of known structure as a query against Swiss-Prot and TrEMBL protein database [ 7 ]. We used the sequences of the following PDB entries: 1C8U [ 5 ], 1IQ6 [ 8 ], 1LO7 [ 9 ], 1MKB [ 1 ], 1NJK [ 10 ], 1O0I [ 11 ] and 1PSU [ 12 ]. These searches have uncovered many novel members of this superfamily as well as finding links between the known structures with a Hotdog fold (see Table 1 and Additional file 1 ). The Pfam database [ 13 ] contains a Thioesterase superfamily with 697 members, each member containing a 4HBT domain (accession: PF03061) corresponding to the HotDog domain. The SCOP database [ 14 ] contains a thioesterase/thiol ester dehydrase-isomerase superfamily, divided into 5 families, namely the 4HBT-like, beta-hydroxydecanoyl thiol ester dehydrase, Thioesterase II (TesB), MaoC dehydratase and PaaI/YdiI-like families. Our searches have found a total of 1357 proteins (see Additional file 2 ) to be related to the known structures of HotDog domain proteins. We took these proteins and clustered them using single linkage clustering to define subfamilies with common functions. This clustering puts 1293 (95%) of the sequences into 85 clusters (see Additional file 3 ). The HotDog domain is found to be associated with a wide range of other domains. The various domain architectures are shown schematically in Figure 2 . We describe the 17 subfamilies (Table 1 ) that have some experimental characterisation, below. The 17 subfamilies contain 909 proteins or 67 % of the total number of HotDog domain proteins. 384 (28 %) proteins cluster into the remaining groups, which contain predominantly hypothetical proteins or proteins that have no known function. They are not discussed here but we hope that our analysis may help in identifying functions for these proteins. Finally we have generated a Hidden Markov Model (HMM) library by concatenating together the HotDog domain sequences of the 85 clusters generated in our analysis (see Additional file 4 ). This library can be used in conjunction with the HMMER program [ 15 ] to search for HotDog domain(s) in any protein of interest. Acyl-CoA thioesterase subfamily The largest subfamily represents over a hundred acyl-CoA thioesterases that are widespread throughout the prokaryotic kingdom, with members also found in eukaryotes. This group of enzymes catalyze the hydrolysis of acyl-CoA thioesters to free fatty acids and coenzyme A (CoA-SH.) [ 16 ]. The subfamily includes thioesterases with activity towards medium and long chain acyl-CoAs (medium chain acyl-CoA hydrolase and cytosolic long-chain acyl-CoA hydrolase/brain acyl-CoA hydrolase (BACH) respectively) and also cytoplasmic acetyl-CoA hydrolase (CACH), which hydrolyzes acetyl-CoA to acetate and CoA-SH. Brown-fat-inducible thioesterase (BFIT), a cold-induced protein found in brown adipose tissue (BAT) [ 17 ] is also included in this group. Both BFIT and CACH possess a StAR-related lipid-transfer (START) domain [ 18 ] that is involved in lipid binding, consistent with the role of BFIT and CACH in lipid metabolism. Duplication of the HotDog domain and recruitment of the START domain seems to be a mammalian innovation. FabZ like dehydratase subfamily Members of this subfamily are found in a wide range of bacteria and sporadically in eukaryotes. In E. coli the products of the fab operon catalyze the four sequential reactions necessary for each round of fatty acid elongation [ 19 ]. The third step in each cycle of fatty acid elongation involves the dehydration of the β-hydroxyacyl-ACP protein intermediate by β-hydroxyacyl-[acyl carrier protein] dehydratase (FabZ) to give trans -2-decenoyl-ACP. FabZ is effective at dehydrating both short-chain and long chain saturated and unsaturated pathway intermediates. This subfamily also contains a dehydratase component of the coronafacic acid (CFA) biosynthetic cluster encoded by the cfa2 gene [ 20 , 21 ]. CFA is the polyketide constituent of a phytotoxin called coronatine, which is a virulence factor of Pseudomonas syringae , a plant pathogen that causes disease in many agriculturally important plants [ 20 ]. MaoC dehydratase-like subfamily The mao C gene exists as an operon with the maoA gene in E. coli and is an enoyl-CoA hydratase involved in supplying (R)-3-hydroxyacyl-CoA from the fatty acid oxidation pathway to polyhydroxyalkanoate (PHA) biosynthetic pathways in fadB mutant E. coli strains. It was identified through its homology to P. aeruginosa (R)-specific enoyl-CoA hydratase (PhaJ1) [ 22 ]. PHAs are polyesters of (R)-hydroxyalkanoic acids, synthesized by numerous bacteria as an intracellular carbon and energy storage material in times of excess carbon sources [ 23 ], with intermediates of fatty acid metabolism such as enoyl-CoA, (S)-3-hydroxyacyl-CoA, and 3-ketoacyl-CoA acting as precursors for PHA biosynthesis [ 22 ]. The crystal structure of the (R)-specific enoyl-CoA hydratase (phaJ) from the Aeromonas caviae has shown that this enzyme also contains a Hotdog fold/domain [ 8 ]. The E. coli MaoC C-terminal HotDog domain is most likely responsible for its enoyl-CoA hydratase actvity. MaoC also contains an N-terminal short-chain dehydrogenase domain, involved in catalysing dehydrogenation of a variety of aliphatic and aromatic aldehydes using NADP as a cofactor. This subfamily also includes the human 17 β-hydroxysteroid dehydrogenase (17 β HSD) type 4, one of four different human 17 β HSDs that catalyze the redox reactions at position C17 of steroid molecules, one of the final steps in androgen and estrogen biosynthesis [ 24 , 25 ]. We also include a NodN-like sub-subfamily here that is found in another cluster containing several other MaoC proteins. Rhizobium and related species form nodules on the roots of their legume hosts, a symbiotic process that requires production of Nod factors, which are signal molecules involved in root hair deformation and meristematic cell division [ 26 ]. The nodulation gene products, including NodN, are involved in producing the Nod factors, however the role played by NodN is unclear. YbgC-like subfamily This subfamily contains a large number of proteins about which very little is known except for the YbgC protein. The YbgC protein of the tol-pal cluster in the gamma-proteobacterium Haemophilus influenzae [ 27 ] has been shown to catalyze the hydrolysis of short-chain aliphatic acyl-CoA thioesters. The tol-pal cluster is present in many Gram-negative bacteria and is important for the maintenance of cell envelope integrity [ 28 ] and this operon is well conserved across gram-negative bacteria. Therefore we hypothesize that uncharacterized members of this subfamily are thioesterases. The Asp 17 residue is conserved in YbgC from Haemophilus influenzae and Pseudomonas aeruginosa , along with the backbone amide NH of Tyr 24 , suggestive of a nucleophilic attack mechanism very similar to the Pseudomonas sp. strain CBS-3 thioesterase mechanism discussed below in the 4HBT class I section. FabA-like subfamily The dehydration of the β-hydroxyacyl-ACP protein intermediate during the third step in each cycle of fatty acid elongation can be catalyzed by β-hydroxydecanoyl-ACP dehydratase/isomerase (FabA), as well as by FabZ, to give trans -2-decenoyl-ACP. FabA is uniquely able to isomerise trans-2-decenoyl-ACP to cis-3-decenoyl ACP, initiating unsaturated fatty acid biosynthesis [ 19 ] and is specific for acyl ACPs of 9–11 carbons in length. Polyketides are a large and structurally diverse class of natural products, produced mainly by soil-dwelling bacteria such as Pseudomonas spp. and Streptomyces spp. They include clinically useful drugs such as the antibiotic erythromycin A and the immunosuppressants FK506 and rapamycin. The biosythesis of polyketides is very similar to that of fatty acids [ 21 ] and polyketide synthases (PKSs) have been classified as type I or type II according to fatty acid synthase (FAS) similarity. Most bacteria and plants use a highly conserved type II FAS system, which uses a distinct enzyme for each reaction. This is in contrast to the mammalian type I system (also used by fungi and some mycobacteria), which uses one multifunctional polypeptide to catalyze all reactions [ 29 , 30 ]. The HotDog domain is found in type II fatty acid synthesis in bacteria (FabA/FabZ), but also in a small number of bacterial polyketide synthases that are of the type I, being composed of several modules [ 31 ] such as β keto-acyl synthases and omega-3 polyunsaturated fatty acid synthase (PfaC). The marine bacteria Shewanella sp. SCRC-2738, Moritella marina strain MP-1 and Photobacterium profundum strain SS9 contain an eicosapentaenoic acid (EPA) biosynthetic cluster ( pfaA-D ), responsible for the synthesis of this omega-3 polunsaturated fatty acid (PUFA), [ 32 , 33 ]. The PfaC protein contains two HotDog domains (see Figure 2 for the domain organisation found in P. profundum ), which are also found in the eukaryotic marine protist, Schizochytrium , suggesting that the PUFA synthetic cluster has undergone lateral gene transfer [ 32 ]. This subfamily also includes several fatty acid synthase proteins from bacteria, such as Mycobacterium bovis fatty acid synthase. This multifunctional protein is capable of catalysing de novo synthesis and chain elongation of fatty acids [ 34 ] and has a very similar domain architecture to the polyunsaturated fatty acid synthases, as it contains an acyl-transferase, β-keto acyl synthase N and C-terminal domains (see Figure 2 ). The catalytic residues of FabA's bifunctional active site are His 70 and Asp 84 , His 70 is conserved in FabZ dehydratase, but Asp 84 is replaced with Glutamate. This replacement may be responsible for FabZ's inability to catalyze the isomerization reaction [ 1 ]. Fat subfamily Acyl-ACP thioesterases In plants, fatty acid synthesis occurs in the stroma of plastids, where the acyl chains are bound to the acyl carrier protein (ACP) during extension cycles [ 35 ]. Acyl-ACP thioesterases terminate fatty acid synthesis in plants by hydrolysing the thioester bond existing between an acyl moiety and the ACP [ 36 ]. In higher plants acyl-ACP thioesterases have been classified into two gene classes, fatA and fatB , based on sequence similarity and substrate specificities [ 37 , 38 ]. Arabidopsis FatA displays highest activity towards oleoyl-ACP whereas Arabidopsis FatB is most active towards palmitoyl-ACP [ 37 ]. This subfamily contains both FatA and FatB members [ 35 ]. The proteins in this subfamily range in length from 240 to 400 amino acids and therefore we hypothesized that they might contain two HotDog domains, located at the N and C teminal halves. By splitting the sequence of proteins from this subfamily into an N-terminal half and C-terminal half we were readily able to detect the relationship to other subfamilies using PSI-blast with query proteins such as Q899Q1 and Q42714, confirming our hypothesis. TesB-like subfamily This subfamily contains the E. coli medium chain length acyl-CoA thioesterase II [ 5 ] encoded by the tesB gene [ 38 ], which is a close homolog of the human thioesterase II (hTE) enzyme. hTE catalyzes the hydrolysis of palmitoyl-CoA to CoA and palmitate and was identified as a human T cell protein that binds to the myristoylated HIV-1 Nef protein, correlating with Nef-mediated CD4 down regulation [ 39 ]. hTE could regulate targeting of the cytoplasmic Nef protein to the plasma membrane, which is dependent on a lipid modification, i.e. a myristoylation anchor and recombinant hTE shows maximal activity with myristoyl-CoA [ 39 ]. However further studies have shown that hTE localizes to peroxisomes [ 40 , 41 ], dependent on a C-terminal peroxisomal targeting sequence, SKL, and coexpression of Nef and hTE results in relocation of Nef to peroxisomes, so the role of Nef and hTE during HIV infection remains unsolved. The catalytic site of E. coli thioesterase II was identified by site directed mutagenesis and involves a hydrogen-bonded triad of Asp 204 , Thr 228 , and Gln 278 , which synergistically activate a water molecule for nucleophilic attack of the carbonyl thioester carbon of medium chain length acyl-CoA substrates [ 5 ]. This is a novel reaction mechanism for a thioesterase and differs from the nucleophilic mechanisms used by β-hydroxydecanoyl dehydratase and 4HBT thioesterase in both Pseudomonas and Arthrobacter discussed below. This subfamily is found in bacteria and eukaryotes. 4HBT class II subfamily This subfamily includes 4-hydroxybenzoyl CoA thioesterase (4HBT) from Arthrobacter sp. strains SU and TM1 encoded by the fcbC gene [ 3 ]. The Pseudomonas thioesterase uses the Asp 17 residue to mediate the hydrolysis reaction as discussed below in the 4HBT class I section. Gln 58 from Arthrobacter corresponds to the Asp 17 residue in Pseudomonas but inspection of the Arthrobacter strain SU active site has revealed the catalytic base (or nucleophile) to be Glu 73 , on the opposite side of the substrate binding pocket to Asp 17 [ 3 ]. Also the Pseudomonas thioesterase dimers form a tetramer with their long α-helices facing inwards, in contrast to Arthrobacter thioesterase where the dimers form a tetramer with their long α-helices facing outwards [ 3 ]. In Pseudomonas and Arthrobacter thioesterases, the 4-hydroxyphenacyl moieties are positioned in such an orientation that the thioester C = O interacts with the α-helical N-terminus by means of hydrogen bonding to a backbone amide NH, on Tyr 24 in Pseudomonas and Gly 65 in Arthrobacter , and it is this contact that results in polarization of the C = O for nucleophilic attack [ 3 ]. While the structure of Arthrobacter sp. strain SU thioesterase displays a similar Hotdog-fold topology to the 4HBT class I Pseudomonas enzyme, the enzymes differ at the level of catalytic platform, CoA binding site and quaternary structure [ 3 , 42 ]. This is not an unexpected finding as Todd et al. have found that 12 of the 31 superfamilies they analyzed displayed positional variation for residues playing equivalent catalytic roles [ 43 ]. A surprising inclusion in this subfamily is the ComA2 protein from Bacillus subtilis . ComA is a response regulator and transcription factor [ 44 ] that together with the histidine kinase, ComP, constitutes a two-component signal transduction system required for the development of competence. The com A locus is composed of two ORFs. ComA2 is cotranscribed with ComA1, which is required for competence while ComA2 is not [ 45 ], and so the role of the HotDog domain in this protein remains a mystery. PaaI subfamily The phenylacetic acid (PA) catabolic pathway in E. coli has been characterised and found to contain 14 genes, allowing catabolism of this aromatic compound into likely Krebs cycle intermediates [ 46 ]. The paa operon in E. coli encodes PaaI, which is probably a thioesterase involved in the catabolism of PA. The catabolism of phenylacetic acid (PA) in E. coli begins with an activation step where Phenylacetyl-CoA ligase, PaaK, converts phenylacetate into Phenylacetyl-CoA. 4-chlorobenzoate-CoA ligase catalyzes a similar reaction at the first step of the 4-chlorobenzoate-degradation pathway. The thioesterase, PaaI, may be involved in a reaction similar to the last step in the degradation of 4-chlorobenzoate (see 4HBT class I below), however this remains to be demonstrated. FapR subfamily This small subfamily is restricted to firmicutes. FapR is a highly conserved transcriptional regulator found in many gram-positive organisms, including all species of Bacillus [ 47 ]. It controls expression of genes involved in type II fatty acid and phospholipid biosynthesis, by binding to a consensus promoter sequence of the fap regulon and acting as a negative regulator. Malonyl-CoA, an intermediate in the lipid biosynthetic pathway, controls FapR. The HotDog domain has likely retained its substrate specificity for malonyl-CoA, but appears to have lost its catalytic ability, in common with the ligand binding domain of other transcriptional regulators. FapR contains a helix-turn-helix motif at the N-terminus (see Figure 2 ), which is similar to the DeoR transcriptional regulator family (data not shown), consistent with its role as a DNA binding protein. 4HBT class I subfamily The crystal structure of 4HBT from the soil-inhabiting bacterium Pseudomonas sp. strain CBS-3 [ 2 ] has helped define the HotDog domain. A lot of attention has been focused on this microorganism because of its ability to survive on 4-chlorobenzoate (4CBA) as its only source of carbon [ 48 ]. 4CBA is a by-product of microbial degradation of industrial pollutants such as DDT and polychlorinated biphenyl herbicides [ 49 ] and this strain of Pseudomonas may be used as a bioremediation agent for degrading 4CBA. Pseudomonas sp. strain CBS-3 contains an fcb operon responsible for hydrolytic dechlorination of 4CBA, with 4CBA-CoA ligase (FcbA), 4CBA-CoA dehalogenase (FcbB), and 4HBT (FcbC) catalyzing sequential reactions that result in the degradation of 4CBA to 4-hydroxybenzoate. The thioesterase catalyzes the third step in the degradation pathway, which is the hydrolysis of the 4-hydroxybenzoyl-CoA thioester moiety to give 4-hydroxybenzoate and CoA [ 50 ]. 4HBT from Pseudomonas sp. strain DJ-12 [ 51 ] is also found in this subfamily. The organization of the fcb operon in strain DJ-12 is different from that observed in strain CBS-3. The fcb genes are organised as B-A-C in both strains but strain DJ-12 has three ORFs between A and C called T1, T2, and T3 that are unique to this strain. These three genes are similar to the C4-dicarboxylate transport system in Rhodobacter capsulatus , suggesting that they may encode membrane proteins involved in the uptake of 4CBA [ 51 ]. This is in contrast to the gene organisation observed in the 4HBT class II, where Arthrobacter sp. strain SU and strain TM1 have an A-B-C order [ 51 ]. There is a duplication of the cluster in strain SU, where it is found on a plasmid, whereas only one copy exists in strain TM1, where it is located chromosomally. Both operons contain a T gene located at the end of the cluster, possibly involved in 4CBA uptake. Bacillus halodurans C-125 contains a gene called BH1999, encoding a novel gentisyl-CoA thioesterase, which catalyzes the hydrolysis of gentisyl-CoA (2,5-dihydroxybenzoyl-CoA)[ 4 , 52 ] to yield gentisate (2,5 dihydroxybenzoate). BH1999 is found in a gentisate oxidation pathway gene cluster in B. halodurans . Gentisate has been implicated as an intermediate in the degradation of several industrial aromatic compounds [ 4 ]. Gentisyl-CoA thioesterase and 4HBT from Pseudomonas perform different physiological functions but remain in the same subfamily because they are highly related. The active site residues Asp 16 and Asp 31 of gentisyl-CoA thioesterase align with Asp 17 and Asp 32 of 4HBT. These are crucial residues that are proposed to function in nucleophilic catalysis and substrate binding respectively. Loss of Asp 17 in the Pseudomonas enzyme effectively halts catalysis, while loss of the corresponding Asp 16 residue to the Bacillus halodurans enzyme only reduces its catalytic rate by 230-fold, perhaps indicating that the hydrolysis reaction does not proceed through an Asp 16 -mediated nucleophilic attack mechanism previously proposed for Asp 17 [ 53 , 4 ]. Asp 17 in Pseudomonas strain CBS-3 has been suggested to participate in nucleophilic catalysis rather than general base catalysis based on the following observations. The Asp 17 carboxylate is located at a distance of 3.2 Å from the substrate C = O thioester bond, its aligned trajectory and the absence of a water molecule near the reaction centre are all suggestive of a role for Asp 17 as a catalytic nucleophile [ 9 , 53 ]. Asp 32 in Pseudomonas interacts with the benzoyl OH of 4-hydroxybenzoyl-CoA [ 9 ] and perhaps Asp 31 plays a similar role. Other subfamilies/ members In the above sections we have described the 11 subfamilies that have some functional characterization. In this section we describe the other 6 subfamilies that have no functional characterization, except they are associated with other domains or have been structurally characterized. The CBS associated subfamily contains the hypothetical protein BH3175 from Bacillus halodurans . The BH3175 protein contains two homologous copies of the CBS domain [ 54 ]. Scott et al . have recently shown that tandem pairs of CBS domains act as sensors of cellular energy status by binding AMP, ATP, or S-adenosyl methionine and mutations in CBS domains impair this binding in several hereditary disorders [ 55 ]. Although we do not know the substrate or activity of this subfamily of the HotDog superfamily, we can suggest that this step is regulated in an energy dependent manner by the CBS domains. 3-hydroxyacyl-CoA dehydrogenase is an enzyme involved in fatty acid metabolism, catalyzing the reduction of 3-hydroxyacyl-CoA to 3-oxoacyl-CoA [ 56 ]. The hydroxyacyl-CoA dehydrogenase-associated subfamily includes 3-hydroxyacyl-CoA dehydrogenase from Agrobacterium tumefaciens strain C58, which contains a HotDog domain at its C-terminus and the two domains (3HCDH_N and 3HCDH) associated with 3-hydroxyacyl-CoA dehydrogenase activity are located at the N-terminus and central portion of this protein. The combination of activities may allow substrate to be passed from one domain to the next. Other subfamilies in the superfamily include the YiiD protein from E. coli , where an acetyltransferase domain is fused. The human mesenchymal stem cell protein DSCD75 and its counterpart in mouse also contain a HotDog domain. A Structural proteomics project has shown that the conserved hypothetical E. coli protein YbaW contains a Hotdog fold [ 10 ]. Finally the Ralstonia solanacearum hypothetical protein RSp0367, containing a HotDog domain and two AMP-binding domains, found in proteins involved in ATP-dependent covalent binding of AMP to their substrate, is a member of another subfamily. Domain fusion events It has been shown that proteins that are functionally linked are occasionally found to be fused in various genomes. These fusion proteins have been termed Rosetta proteins [ 57 , 58 ] and can be used to predict the functional linkages of proteins with each other. The HotDog domain superfamily contains several rosetta proteins where the fused proteins are also found unfused in other genomes. In these cases they are adjacent to each other in known operons. The examples found in the HotDog superfamily are shown in Figure 3 and are described briefly here. Within the FabZ subfamily the LpxC deacetylase domain (UDP-3-O-acyl N-acetylglucosamine deacetylase) is fused to the FabZ-like HotDog domain in Chlorobium tepidium (see Figure 3a ). LpxC catalyzes the N-deacetylation of UDP-3-O-acyl N-acetylglucosamine deacetylase, the second and committed step in the biosynthesis of lipid A, which anchors lipopolysaccharide (LPS) in the outer membranes of most gram-negative bacteria [ 59 ]. The unfused proteins are found adjacent in operons from several species of chlamydia and cyanobacteria. In the 4HBT class II subfamily we observed the order of the operon is ligase(A)-dehalogenase(B)-thioesterase(C). In Bacteroides thetaiotaomicron there is a Rosetta protein that contans a haloacid dehalogenase-like hydrolase domain (see Figure 3b ). This domain architecture is similar to the fcb operon structure in Arthrobacter , with a dehalogenase-like hydrolase (HAD) domain and a HotDog domain (see Figure 3 ) i.e. it represents a fusion of the fcbB and fcbC gene products to form a novel protein in B. thetaiotaomicron . The final domain fusion is in the 3-hydroxyacyl-CoA dehydrogenase from Agrobacterium tumefaciens strain C58, which possesses the HotDog domain, 3HCDH_N domain (3-hydroxyacyl-CoA dehydrogenase, NAD binding) and 3HCDH (3-hydroxyacyl-CoA dehydrogenase, C-terminal domain) domain (see Figure 3c ). This may represent a fusion of the PaaC and PP3281 proteins in the gamma-proteobacterium Pseudomonas putida 2440 phenylacetic acid degradation operon. These fusion events suggest that the domain fusion process can occur in a simple scheme with two distinct phases. Firstly, two proteins are recombined into adjacent positions in an operon. Secondly, the two genes are then fused by a process of mutation that removes the stop codon at the end of the first gene and maintains reading frame through the second gene [ 60 , 61 ]. Sequence motifs The MASIA program [ 62 ] was used to search for HotDog domain motifs in the aligned sequences of the 17 subfamilies. The various motifs are found in Additional file 5 . It must also be noted that the PROSITE database release 18.29 [ 63 ] contains a consensus sequence motif (PS01328), called the 4-hydroxybenzoyl-CoA thioesterase family active site, and this is found in 29 Swiss-Prot, TrEMBL and TrEMBL-NEW entries cross-referenced with PS01328. This consensus pattern, [QR]-[IV]-x(4)-[TC]- D -x(2)-G [IV]-V-x-[HF]-x(2)-[FY], where D is the active site residue, is found in the YbgC-like subfamily and in the 4HBT-I subfamily. 19 of the 29 members are found in the YbgC-like group and 3 in the smaller 4HBT-I group. The remaining 7 proteins are scattered in various clusters consisting of hypothetical or unknown proteins. We have found, using MASIA, that this motif is found in the entire YbgC and 4HBT-I subfamilies, extending the number of proteins containing this motif to 107. We have also identified a HGG motif in the 4HBT-II and PaaI subfamilies. This motif is HGGAS-x-ALA E in the 4HBT-II subfamily and HGG-x-IF-x-LA D in PaaI members. The active site residue, Glu 73 , is known for 4-hydroxybenzoyl-CoA thioesterase from Arthrobacter sp. Strain SU, however the active site for E. coli PaaI is not known and we suggest that it is Asp 61 in the HGG motif above, which is 100 % conserved in all members of this subfamily (see Additional file 6 ). Conclusions We have defined and analyzed the HotDog domain superfamily and in our analysis of this superfamily we have found 18 different domain architectures and defined 17 subfamiles. We have also investigated the domain organisation and the role that this plays in generating functionally diverse enzymatic and nonenzymatic functions based on the HotDog fold. Domain duplication, domain recruitment and incremental mutation have been key to the evolution of this superfamily. We have also looked at gene context and operon structures and found many examples of fusion proteins, in which the HotDog domain has been fused to another protein to generate functional diversity. The large number of subfamilies we have found, the diverse range of activities these proteins participate in and the taxonomic distribution of the HotDog domain indicates an ancient superfamily that has diverged substantially to fulfil numerous roles in the cell. Our analysis may help with further experimental investigation of members of this superfamily. Some members of this superfamily, such as the P. falciparium FabZ enzyme have been proposed as a target for new anti-malarial drugs [ 64 ] as FabZ homologues are not found in humans. Finally our analysis identified hundreds of novel proteins such as human mesenchymal stem cell protein DSCD75 and the Ralstonia solanacearum hypothetical protein RSp0367 as probable enzymes potentially involved in lipid metabolism. Given that the large majority of proteins in this family are involved in bacterial lipid metabolism we suggest that the HotDog domain evolved in bacteria first and may then have been transferred to eukaryotes and archaea on several occasions. Since this time duplication and mutation has allowed it to fill a variety of roles. Methods Sequence analysis All PSI-BLAST searches were carried out using default inclusion thresholds and searched against the Swiss-Prot and TrEMBL sequence database (SWISS-PROT release 42.12 and TrEMBL release 25.12). To define subfamilies we clustered the results of an all-against-all search of the 1357 HotDog domain proteins using NCBI BLASTP and single linkage clustering at an E-value of 10 -15 . Operon analysis Gene context/operon analysis was carried out with the GeConT tool ( Ge ne Con text T ool) [ 65 ] available at the GeConT Home Page [ 66 ]. Domain analysis Protein domain analysis was carried out using Pfam [ 13 ] (release 12.0) available at the Pfam Home Page [ 67 ]. Motif analysis Consensus motif sequences were identified in the subfamily alignments using the MASIA program [ 62 ] available at the MASIA 2.0 Home Page [ 68 ]. Authors' contributions AB conceived the study and was involved in all stages of the manuscript preparation including figure preparation. SCD conducted the PSI-BLAST searches, prepared the manuscript and some of the figures. Both authors read and approved the final manuscript. Supplementary Material Additional File 1 A network showing the unification of the HotDog superfamily using PSI-BLAST searches. Each yellow oval represents a query sequence used to seed a PSI-blast search. We compared each PSI-BLAST output to all the others and connected them with a line if they shared any sequences in common. There are only two connections in the graph that were not made (Image is in EPS format). Click here for file Additional File 2 A complete list of the 1357 HotDog domain containing proteins. Click here for file Additional File 3 The HotDog domain superfamily. This xml-like file contains 1293 (95%) of the HotDog domain containing sequences, grouped into 85 clusters, which permits investigators to immediately identify HotDog domain(s) in their 'unknown' protein of interest and allow them to infer some functionality. Click here for file Additional File 4 A HotDog domain HMM library. This library can be used in conjunction with the HMMER program to search for HotDog domains in any protein sequence. Click here for file Additional File 5 A list of motifs identified in each subfamily. Motifs were identified using the MASIA program [62]. A motif starts when at least 3 of 4 consecutive positions are more than 40% conserved and extend until at least 2 amino acids in a row are less than 40% conserved [ 71 ]. Motifs corresponding to PROSITE motif PS01328, [QR]-[IV]-x(4)-[TC]-D-x(2)-G [IV]-V-x-[HF]-x(2)-[FY] are underlined and in bold. Motifs highlighted in red and green are conserved between the respective subfamilies. Click here for file Additional File 6 Subfamily alignments. These alignments were constructed using the MAFFT alignment program [72] and rendered using the CHROMA software package [73]. Known active site residues are indicated below the subfamily alignments. The highly conserved Asp residue in the PaaI subfamily is proposed as an active site residue based on motif similarities between the 4HBT-II subfamily and the PaaI subfamily. Jpred predicted consensus secondary structures are indicated above the alignments [74]. Click here for file	/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC516016.xml
423132	fMRI Beyond the Clinic: Will It Ever Be Ready for Prime Time?	Functional magnetic resonance imaging offers the promise of peeking into the human mind. What signals in the human brain can we really detect and how should the technology be used?	Functional magnetic resonance imaging—fMRI—opens a window onto the brain at work. By tracking changes in cerebral blood flow as a subject performs a mental task, fMRI shows which brain regions “light up” when making a movement, thinking of a loved one, or telling a lie. Its ability to reveal function, not merely structure, distinguishes fMRI from static neuroimaging techniques such as CT scanning, and its capacity to highlight the neural substrates of decisions, emotions, and deceptions has propelled fMRI into the popular consciousness. Discussions of the future of fMRI have conjured visions of mind-reading devices used everywhere from the front door at the airport terminal to the back room of the corporate personnel office. At least one “neuromarketing” research firm is already trying to use fMRI to probe what consumers “really” think about their clients' products. But will fMRI's utility in the real world ever match the power we currently imagine for it? Is fMRI likely to leave the clinic for widespread use in the courtroom or the boardroom? Are there neuroethical nightmares just around the corner? Or are all these vivid specters really just idle speculations that will never come to pass? 150,000 Grains of Rice To understand the potential, and the limitations, of fMRI, it's helpful to know how the technique works. The heart of the apparatus is a large donut-shaped magnet that senses changes in the electromagnetic field of any material placed in its center, in particular—when a head is scanned—the blood as it flows through the brain. When a region of the brain is activated, it receives an increased flow of oxygenated blood (the extremely rapid redirection of blood within the active brain is one of the underappreciated wonders supporting neural activity). This influx of oxygenated blood alters the strength of the local magnetic field in proportion to the increase in flow, which is detected and recorded by the imaging machinery. The resolution of the best fMRI machines—the smallest “volume picture element,” or voxel, they can distinguish and make an image of—is currently about 1.5 mm ×1.5 mm × 4 mm, the size of a grain of rice. There are approximately 150,000 of these little volumes in the typical brain, and the immense computers hooked up to the scanners record and integrate signals from all of them. In a typical experiment, a subject, lying still with his head surrounded by the magnet, does nothing for thirty seconds, then performs some task for thirty seconds, then lies still for thirty seconds. For each voxel, the signal during the task is compared to the signal at rest; those areas of the brain with stronger signals during the task are presumed to be processing the information that underlies the performance of the task ( Figure 1 ). According to Joy Hirsch, Director of the Functional Magnetic Resonance Imaging Research Center at Columbia University, fMRI represents a “quantum leap” over any previous technology for imaging the brain. “It enables us for the first time to probe the workings of a normal human brain,” she says. “It's really opening the black box.” Figure 1 The Basics of fMRI Blood oxygen level–dependent signals are measured and compared between test and resting conditions. (Image courtesy of Joy Hirsch, Columbia University.) The first caveat about fMRI's imaging power, though, and one that every neuroimager stresses, is that a voxel is a long way from a neuron. There are an estimated 100 billion neurons, so at best, an fMRI is signaling blood flow changes associated with the increased activity of tens of thousands of neurons. As a result, says Hirsch, fMRI “falls short when we want to ask about more detailed brain processes. We're not learning that much about how neurons are doing local computing.” While resolution will improve over time, it seems unlikely that fMRI will ever detect the activity of individual neurons, and so its ability to dissect the “fine structure” of thought is inherently limited. (Even should it become possible to detect and integrate the workings of every neuron in the brain, it would still be far from clear how neuronal firing patterns translate into coherent, perceived thoughts, and this gap is unlikely to be bridged by any advance in imaging technology alone.) These limitations have not prevented fMRI researchers from making some major discoveries about brain function, however. Hirsch, for instance, showed in one study that minimally conscious individuals still process human speech, and in another, that those who become bilingual as young children employ overlapping language areas in the cerebral cortex, while those who learn a second language later in life use a different area for the second language. The key strength of fMRI, she says, is that it provides the ability to test these kinds of hypotheses about structure–function relationships in the normal brain. All Sizes Do Not Fit One But the hypotheses that can be tested and the conclusions that can be drawn are still largely about group averages, not about the functionings of individual brains, and therein lies a second major caveat about the use of fMRI beyond the clinic. John Gabrieli, Associate Professor of Psychology at Stanford University, has shown that distinct activation patterns in the brains of dyslexic children normalize as they improve their reading skills ( Figure 2 ). It seems like a small leap from there to including an fMRI as part of the workup for a schoolchild struggling in the classroom. But, Gabrieli cautions, that small leap in fact traverses a huge chasm, on one side of which is the group data from which conclusions are drawn in studies, and on the other side, the application of these conclusions to the individual child. “At the moment, fMRI would be among the most useless things to do. We would love to get it to the point that it would be useful [on an individual basis],” he says, but it's not there yet. “There is no single-subject reliability,” says Gabrieli. “Where we are now, I'm not aware of any applications for which it would be responsible to interpret an individual scan [based on group data].” Figure 2 Different Activation Patterns in the Brains of Dyslexics As Compared to Normal Subjects in a Rhyming Task (Images courtesy of John Gabrieli, Stanford University.) There are similar limitations to most other applications of fMRI—while conclusions can be made about aggregated data, individual scans are for the most part too hard to interpret. There is not yet any real understanding of how brain patterns change over time in an individual, or how interindividual differences should be interpreted in relation to the conclusions that are valid for groups. This makes fMRI an unlikely tool for job screening, for instance. While one study has shown a brain signature in a group of white people that is associated with racial bias, denying a particular individual a job on the basis of such a scan would likely lead straight to a lawsuit, with experts debating whether this scan in this individual on this day does or doesn't reflect his underlying racial attitudes. On the other hand, Hirsch has used individual scans to help locate a patient's language structures that must be spared during neurosurgery. “If you are a neurosurgeon planning a resection, you don't want an average brain at all. Millimeters matter.” But her success is precisely because she is not using group data to make inferences about the individual—she is not leaping over the chasm, but instead is toiling entirely on the other side of it. “The goal is personalized medicine,” she says. A Little Guilty Knowledge Is a Dangerous Thing Even this kind of personalized approach with fMRI is fraught with problems when researchers attempt to apply it outside the clinic, because of limitations in the technology itself. One researcher with firsthand knowledge of these problems is Daniel Langleben, Assistant Professor of Psychiatry at the University of Pennsylvania School of Medicine. In 2002, Langleben showed that when subjects were hiding information in an attempt to deceive (so-called guilty knowledge), they had intense activity in five distinct brain areas not seen when they were telling the truth. In effect, Langleben used the fMRI as a lie detector. It is potentially even more powerful than a standard polygraph test, he says, because there are thousands of brain regions which can be scanned for deception-triggered variation, versus only three variables—skin conductance, respiration, and blood pressure—used in the standard polygraph. Not surprisingly, Langleben got a lot of press after he announced his results, and his experiment led directly to speculation that we might eventually see fMRIs installed at airports, scanning the brains of would-be terrorists trying to deceive security screeners, or in courtrooms, catching perjurers red-handed (or perhaps red–anterior-cingulate-gyrused?). Langleben is enthusiastic about the potential for an fMRI-based lie detector, and has even applied to the Department of Justice for a grant to develop the technology (they turned him down, saying it was too expensive). But he is also clear about how difficult it will be to get one that really works outside the highly structured confines of the research lab. “We are a long way from making a working polygraph,” he says. Even with a “Manhattan Project” type effort, he speculates it would take at least ten years. “There are still essential discoveries to make along the way,” he says, “and there's a good chance it would end in total failure.” It's not just a matter of developing the imaging technology, he stresses—“we'll need fundamental developments in semantics, too.” This is because “a lot still depends on how you ask the question”—the subtlest of differences can dramatically shift which areas of the brain respond. Given the sensitivity of the fMRI result to such seemingly minor perturbations, it's hard to imagine it could be reliably adapted to the hurly-burly of an airport security checkpoint. Even well-performed scans done in topnotch clinics may not easily find their way into the courtroom. Perhaps the least likely use of fMRI is in determining if a defendant is telling the truth, according to Hank Greely, Professor of Law at Stanford Law School, since compelling someone on trial to submit to an fMRI could be seen as a violation of the Fifth Amendment right against self incrimination, just as giving spoken testimony against oneself is. On the other hand, says Greely, DNA samples and fingerprints can be compelled—whether a brain scan is more like testifying or more like submitting to a blood test is an open question. Still, for the moment, scanning under duress simply isn't feasible, since all you have to do to ruin a good scan is move your head. Motion-correcting algorithms can be used, but they are nowhere near advanced enough to correct for large-scale movements by an unwilling subject. It's much more likely that an fMRI of a willing defendant would be introduced to convince the jury he is telling the truth, or performed before trial to rule out an innocent suspect. While to Greely's knowledge fMRI evidence hasn't yet been used in court, “it's certain to be tried,” and the barrier to its admission will fall as both the reliability and the ease of administration increase. “The easier, the cheaper, the more pleasant a technique is, the more likely it is to be used in the legal system.” Other forensic uses of fMRI are likely to arrive sooner rather than later. Could scans showing diminished impulse control—a function controlled by several regions of the brain, including the striatum and the ventromedial prefrontal cortex—be used to support more lenient sentencing, or even acquit a defendant, because he couldn't control his violent impulses? Or alternatively, will those same scans be used to argue for harsher sentences, since the defendant is clearly “hardwired” to commit similar crimes again? Courts already consider other factors, such as a history of child abuse, in an attempt to more fully understand the psychological state of the defendant. Will brain scans be seen as the ultimate “objective” look into the mind of the person on trial? Deciding all these issues of admissibility will be judges who will need to weigh competing claims from lawyers with competing interests, backed up by expert witnesses with competing theories. Here, the desire to apply the science may rush ahead of its demonstrated validity. Langleben, for one, doesn't think fMRI will be legitimately ready for the courtroom for a long time. On the other hand, he says, “if you want to abuse this technique and claim that it works, you can create tests that will produce results—I can see how it could be done. We know enough to rig it.” But still, he says, “we have all the tools we need to prevent this—there are enough people who are sufficiently honest [who would counter the premature use of fMRI in these contexts].” For now, at least, given the problems inherent in current fMRI technology, the neuroethical nightmare scenarios of widespread brain scanning seem unlikely to come to pass, at least until radical advances make it far cheaper, much less invasive, far less sensitive to subtle perturbations, and with a much more robust ability to legitimately extrapolate from a finding about a group to a prediction about an individual. Where fMRI is concerned, “a penny for your thoughts” is currently more like “a million pennies for a group-averaged hemodynamic response to highly constrained stimuli under entirely artificial conditions.” In light of this, bioethical concerns about fMRI applications should perhaps be viewed not as predictions of a certain future but rather as worstcase scenarios, a reminder of what we want to avoid. “It's a funny thing about the bioethics field,” says Greely. “The general approach is to look for bad news.” While many of these “worst cases” seem highly unlikely to come to pass, Judy Illes, of the Stanford Center for Biomedical Ethics, thinks some action is warranted now, if only to generate a better understanding of the ethical dimensions of fMRI research. She notes that “bioethicists are often viewed as the ethics police,” but she doesn't see regulations as the right path to shape the future uses of fMRI. Instead, she thinks a coalition of involved parties—scientists, lawyers, ethicists, politicians—should work together to develop guidelines that all will find acceptable. “I'm not in the business of stopping anything.” What everyone apparently already agrees on is the need for carefully designed experiments and cautious interpretation of the data. “A huge message in imaging is that you really have to look at the experimental setup at the common-sense level,” says Gabrieli, and avoid the tendency to “pick the most dramatic interpretation.” “The public needs to be reminded of the limitations of these findings,” agrees Hirsch. And as Langleben puts it, expressing his skepticism that there will ever be a one-size-fits-all, foolproof fMRI mind reader: “I don't think we'll ever be able to be stupid about it.”	/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC423132.xml
521080	Delta-9 tetrahydrocannabinol (THC) inhibits lytic replication of gamma oncogenic herpesviruses in vitro	Background The major psychoactive cannabinoid compound of marijuana, delta-9 tetrahydrocannabinol (THC), has been shown to modulate immune responses and lymphocyte function. After primary infection the viral DNA genome of gamma herpesviruses persists in lymphoid cell nuclei in a latent episomal circular form. In response to extracellular signals, the latent virus can be activated, which leads to production of infectious virus progeny. Therefore, we evaluated the potential effects of THC on gamma herpesvirus replication. Methods Tissue cultures infected with various gamma herpesviruses were cultured in the presence of increasing concentrations of THC and the amount of viral DNA or infectious virus yield was compared to those of control cultures. The effect of THC on Kaposi's Sarcoma Associated Herpesvirus (KSHV) and Epstein-Barr virus (EBV) replication was measured by the Gardella method and replication of herpesvirus saimiri (HVS) of monkeys, murine gamma herpesvirus 68 (MHV 68), and herpes simplex type 1 (HSV-1) was measured by yield reduction assays. Inhibition of the immediate early ORF 50 gene promoter activity was measured by the dual luciferase method. Results Micromolar concentrations of THC inhibit KSHV and EBV reactivation in virus infected/immortalized B cells. THC also strongly inhibits lytic replication of MHV 68 and HVS in vitro . Importantly, concentrations of THC that inhibit virus replication of gamma herpesviruses have no effect on cell growth or HSV-1 replication, indicating selectivity. THC was shown to selectively inhibit the immediate early ORF 50 gene promoter of KSHV and MHV 68. Conclusions THC specifically targets viral and/or cellular mechanisms required for replication and possibly shared by these gamma herpesviruses, and the endocannabinoid system is possibly involved in regulating gamma herpesvirus latency and lytic replication. The immediate early gene ORF 50 promoter activity was specifically inhibited by THC. These studies may also provide the foundation for the development of antiviral strategies utilizing non-psychoactive derivatives of THC.	Background The Kaposi's Sarcoma-Associated Herpesvirus (KSHV/HHV-8) is the likely cause of Kaposi's sarcoma in AIDS patients, aging individuals and organ transplant patients [for review, see [ 1 ]]. KSHV is also implicated in AIDS-associated primary effusion lymphoma (PEL) and a subset of cases of the lymphoproliferative disorder multicentric Castleman's disease. Epstein-Barr virus (EBV) belongs to the same group of herpesviruses and is also involved in human malignancies such as Burkitt's lymphoma, Hodgkin's disease, nasopharingeal carcinoma, and AIDS-associated lymphoma [for review, see [ 2 ]]. Related viruses such as herpesvirus saimiri (HVS) of monkeys and the murine gamma herpesvirus 68 (MHV-68) have been developed as animal models [ 3 - 7 ]. The DNA genomes of these oncogenic viruses persist in nuclei of lymphoid cells in a latent episomal circular form and a few of these cells can produce small amounts of infectious virus. In response to extracellular signals, the latent virus can be reactivated leading to production of more infectious virus progeny. This switch from latent to lytic infection is thought to be important in the pathogenesis of herpesviruses and the spread of infection within the organism and among individuals. Reactivation of latent virus is often initiated by extracellular signals activating through receptors and various transcription factors. Transcription factors activate the promoter of the critical viral gene open reading frame 50 (ORF 50) of KSHV, HVS, MHV 68 or its homologue Rta of EBV [ 8 - 16 ]. The ORF 50 protein is also a transcription factor and further boosts production of its own mRNA [ 10 , 13 , 16 ]. An important function of ORF50/Rta is to activate early lytic genes [ 1 , 2 ] involved in DNA synthesis. After synthesis of early proteins, the process culminates with expression of late genes leading to production of virion components, virus assembly, release of progeny virus and cell death. The major psychoactive cannabinoid compound of marijuana, Δ 9 tetrahydrocannabinol (THC), has been shown to modulate and primarily suppress immune responses against various pathogens [for review see [ 17 , 18 ]]. THC binds to and activates either one of the two cannabinoid receptors (CB1 and CB2) located on the surface of both brain and lymphoid cells [ 17 - 26 ]. CB1 and CB2 belong to the family of G protein-coupled receptors characterized by seven transmembrane loops interacting with the ligand on the outer surface of the cell. The receptors also contain an intracellular signaling domain. Several endogenous natural ligands of the CB receptor family termed endocannabinoids have been described. One example is anandamide, a lipid eicosanoid compound generated by the arachidonic acid pathway. The CB receptors are conserved through various vertebrate species including mammals and even fish. They have been shown in various tissues to modulate signaling and gene activation in response to short-lived endocannabinoid ligands or THC. However, THC has been also proposed to influence cellular function by other mechanisms due to its hydrophobic nature and likely association with lipid structures such as cell membranes [for review, see [ 17 , 18 ]]. This membrane-mediated effect is clearly less specific as all types of cells may be subject to it. Since THC is an immune modulator we hypothesized that it may have an effect on gamma herpesvirus replication and/or latency. The data presented show that THC inhibits reactivation and lytic replication of these herpesviruses, possibly through inhibition of the ORF 50 promoter. Methods Tissue culture cell lines, virus, and THC The KSHV positive primary effusion lymphoma cell line BCBL-1 isolated by Don Ganem and co-workers [ 27 ], and BC-3 isolated by Ethel Cesarman and co-workers [ 28 ], were obtained through the US National Institutes of Health AIDS Research and Reference Reagent Program (Rockville, MD, USA) and from the American Type Culture Collection (Manassas, VA, USA). The EBV positive P3HR1 cell line was from George Miller [ 29 ]. MHV 68 was obtained from Jeffrey Sample (St. Jude Children's Research Hospital, Memphis TN USA). NIH3T12 cells were from Samuel Speck (Yerkes Primate Research Center, Emory University, Atlanta, GA, USA). HVS strain 484 was isolated as described [ 30 ]. Owl monkey kidney (OMK) cells were from Danny Daniel (New England Research Primate Center, Southborough, MA, USA). THC was obtained from the National Institute on Drug Abuse, NIH (Bethesda, MD, USA). Three independent batches of synthetic THC were tested that gave very similar results. The purity of these THC preparations exceeded 99%. Antiviral assay based on the Gardella gel method Cell suspensions were loaded in wells of a vertical agarose gel as described [ 31 , 32 ]. The cell layer was then overlaid with a lysis solution containing SDS and pronase resulting in gentle cell lysis and liberation of cellular and viral DNA. The gel was subjected to electrophoresis. Latent episomal DNA migrates much more slowly in these gels than linear replicating DNA [ 31 , 32 ]. After electrophoresis, Southern blotting was performed; DNA was transferred to nitrocellulose followed by hybridization with radiolabeled cloned viral DNA as described [ 32 ]. Radioactive images were analyzed and quantitated by a phosphoimager. The antiviral effect (IC50) was quantitated by comparing the values obtained from the episomal bands with those from linear bands as described [ 32 ]. Virus yield reduction and cytotoxicity assays NIH3T12 cells were infected with MHV 68, OMK cells with HVS, and both cell types with HSV-1 at a multiplicity of infection of 2 in the presence or absence of various concentrations of THC dissolved in DMSO. Control cultures were treated with DMSO. IC50 was determined by measuring virus titers in THC treated and control samples after one cycle of freeze-thawing of the cultures. Additional controls were uninfected cells grown in the presence of THC or DMSO. THC or DMSO was present throughout the experiment. Dual luciferase promoter assays To evaluate the effect of THC on the KSHV ORF 50 promoter, a DNA fragment corresponding to the promoter region upstream of the mapped mRNAs [reference [ 9 ], nucleotides 70513–71513] was cloned into the basic firefly luciferase vector (Promega, Inc., Madison Wisconsin). To assay the effect of THC on the MHV 68 ORF 50 gene, a DNA fragment containing the 0.4 kb full length ORF 50 promoter (nucleotides 66242–66652) cloned upstream of the firefly luciferase was provided by Dr. Samuel Speck [ 46 , 47 ]. About 5 μg KSHV ORF 50 firefly plasmid DNA was transfected into BCBL-1 cells. About 5 μg MHV 68 ORF 50 firefly plasmid DNA was transfected into NIH312 cells. Control renilla luciferase expression vector (under the control of the CMV immediate early promoter) was included and co-transfected along with the ORF 50 reporter constructs. Cells were incubated for 48 h with either 5 μg/ml THC solution or an equal amount of DMSO. Luciferase activities were determined using a dual luminometer. Results THC inhibits KSHV and EBV DNA replication To evaluate the effect of THC on KSHV replication, we selected BCBL-1 and BC-3 lymphoblastoid cells because these cultures spontaneously produce small amounts of virus, and we used a previously published antiviral drug testing protocol [ 32 ] to determine whether THC induces or inhibits virus replication. Dead cells from BCBL-1 or BC-3 lymphoblastoid cell cultures were removed by Ficoll gradients. Cells were cultured in RPMI 1640 growth medium containing 10% serum for 3 days in the presence or absence of various concentrations of THC (dissolved in DMSO). Control cells were cultured in the presence of 0.1% DMSO (this concentration of DMSO was also used in THC treated cultures). Cells were suspended and loaded in wells of a vertical agarose gel, then overlaid with a lysis buffer to remove proteins from DNA. After lysis, slowly migrating episomal DNA representing the resident latent viral genome, and fast migrating linear DNA representing lytic virus replication, were separated by electrophoresis. After electrophoresis, viral DNA was visualized by Southern blotting. Figure 1 , left panel shows that control DMSO-treated BCBL-1 cells spontaneously produce a small amount of linear DNA, indicating lytic replication/reactivation in a subpopulation of the cells. Various concentrations of THC showed concentration-dependent inhibition of linear but not episomal KSHV DNA. These data were reproducible and similar observations were made in five independent experiments. Based on these experiments we calculated the 50% inhibitory concentration (IC 50 ) of THC at 1 μg/ml or about 3.3 μM. To evaluate whether another KSHV strain is also susceptible to THC, the BC-3 cell line was also tested. In these cells, THC had a similar inhibitory effect on KSHV reactivation as observed with BCBL-1 cells (not shown). Two experiments analogous to the one described for KSHV were performed with the EBV transformed B cell line P3HR1. P3HR1 cell cultures were grown in the presence of the phorbol ester TPA, because EBV did not reactivate spontaneously and TPA was essential to induce lytic EBV replication. The cultures were also supplemented with either various concentrations of THC dissolved in DMSO or with equivalent volumes of DMSO. Fig. 1 , right panel, shows a representative experiment. The autoradiogram indicates that THC inhibited EBV linear DNA synthesis. Based on quantitative analysis of the radioactive bands and extrapolation of data, the 50% inhibitory concentration (IC 50 ) for THC was estimated at around 0.9 μg/ml or about 3 μM. THC inhibits MHV 68 and HVS lytic replication in monolayer cells We examined the effects of THC on the replication of MHV68 and HVS, two rhadinoviruses related to KSHV and to lesser extent to EBV. These viruses can infect monolayer cells and are suitable for testing the effects of compounds on lytic virus replication by virus yield reduction assays, a more widely used test for evaluating antiviral drugs. NIH3T12 cells were infected with MHV68 at a multiplicity of infection of 2 in the presence or absence of various concentrations of THC dissolved in DMSO. Control cultures were treated with DMSO. THC or DMSO was present throughout the experiment. Cell cultures were incubated for 48 h. As shown in Figure 2 , a typical full cytopathic effect of MHV 68 was observed in the control infected cell culture treated with DMSO; most adherent cells become detached from the plate, and remaining loosely adhered cells were round and denser than uninfected controls. However, when infected cells were cultured in the presence of 1.25 μg/ml or higher concentrations of THC, they remained indistinguishable from uninfected control cells and remained adhered to the plate. The effect of 0.6 μg/ml of THC was intermediate between uninfected cells and the full cytopathic effect. These results indicated a protective effect of THC against destruction of host cells by the virus and suggested that THC may inhibit MHV 68 replication. Similar results were obtained with HVS in owl monkey kidney cells (not shown). To quantitatively determine antiviral effects of THC, yield reduction assays were performed. NIH3T12 cell cultures were infected with MHV 68 and incubated with THC or DMSO dilutions for 48 h. Cell-associated virus was liberated by freeze-thawing. After low speed centrifugation, virus titer was determined in the culture supernatants. Figure 3 shows that virus yield (expressed as infectious units of virus per ml) was highly significantly inhibited by THC. Inhibition of virus yield was over 300-fold at 10 μg/ml THC. The 50% inhibitory concentration (IC 50 ) was estimated at around 0.6 μg/ml, equivalent to about 1.9 μM. Similar results were obtained from two independent experiments. Similar and reproducible results were obtained with HVS in monkey kidney cells (not shown). THC is not cytotoxic to murine NIH3T12 or owl monkey kidney (OMK) cells Because the observed effects might be due to a non-specific toxic effect of THC, we tested whether THC alters cell division or morphology in NIH3T12 and owl monkey kidney (OMK) cells. Monolayers of these cells were prepared at about 25% confluency and cultured for 2 days in the presence of THC concentrations ranging 0.6–10 μg/ml. A series of photographs (Figure 4 ) show that THC treated cultures were indistinguishable from control cultures, formed confluent monolayers, and showed no altered morphology. Owl monkey kidney (OMK) cells also showed no toxicity in an analogous experiment (not shown). The effect of higher THC concentrations on cell division was also determined. Exponentially growing BCBL-1 or NIH12 cells were cultured for two days in the presence of various concentrations of THC or DMSO and cell counts were determined. The 50% cell division inhibitory concentration of THC for BCBL-1 cells was around 33 μM. NIH3T12 cells were less sensitive and the 50% inhibitory concentration was around 99 μM (not shown). THC has no comparable effect on HSV-1 lytic replication We next examined whether or not the THC inhibitory effect can be observed against other herpesviruses and tested whether drug treatment suppresses the replication of the alpha herpesvirus, HSV-1. Monolayers of NIH3T12 were prepared, infected with about 100 infectious units of HSV-1, and cultured for 3 days in the presence of THC concentrations ranging 0.6–10 μg/ml. A series of photographs shows (Figure 5 ) that typical plaques developed in THC treated cultures, indistinguishable from those in control cultures. HSV-1 replication in OMK was also unaffected by THC in an analogous experiment (not shown). We also examined the effects of THC on production of HSV-1 virus yield in NIH3T12 cells. Cells were infected with HSV-1 at a multiplicity of infection of 2 in the presence or absence of various concentrations of THC dissolved in DMSO. Control cultures were treated with DMSO. THC or DMSO was present throughout the experiment. Cell cultures were incubated for 24 h and cell-associated virus was liberated by freeze-thawing. After low speed centrifugation the HSV-1 titer was determined in the culture supernatants. Figure 6 shows that THC had no significant inhibitory effect on replication of HSV-1 in NIH3T12 cells. THC inhibits the ORF 50 promoter To evaluate the effect of THC on the MHV 68 and KSHV ORF 50 promoters, luciferase reporter assays were performed. To assay the effect of THC on the MHV 68 gene, a DNA construct published by the Speck laboratory [ 46 , 47 ] containing the 0.5 kb full length ORF 50 promoter cloned upstream of the firefly luciferase was transfected into NIH312 cells. Control renilla luciferase expression vector (under the control of the CMV immediate early promoter) was included and co-transfected with the ORF 50 reporter. Cells were incubated for 48 h with either 5 μg/ml THC or equal amount of DMSO. Luciferase activities were determined using a dual luminometer. The results show (Table 1 ) that 5 μg/ml THC suppressed the ORF 50 promoter about 7.4 fold. In contrast, the control CMV promoter activity was only reduced by 35%. Similar results were obtained with the KSHV ORF 50 promoter construct containing 1 kb of the promoter sequence. In the transfected BCBL-1 cells, 5 μg/ml THC inhibited the ORF 50-driven luciferase activity almost 4-fold. Interestingly, the co-transfected CMV promoter was slightly stimulated by THC in these cells. These results were obtained from three independent transfection experiments for both the MHV 68 and KSHV promoter assays. Cannabinoid receptor antagonists can reverse the inhibitory effect of THC on KSHV replication To investigate whether the observed antiviral effects of THC are mediated through the cannabinoid receptors, two antagonists of THC, SR141716A (which acts on CB1) and SR144528 (which inhibits CB2; a gift from Dr. Pierre Casallas, Sanofi Recherche), were tested in the standard BCBL-1 KSHV reactivation assay (Figure 7 ). BCBL-1 cells were incubated with or without THC and in the presence of these compounds. The THC inhibition of KSHV DNA synthesis was not reversed by treatment with a single receptor antagonist (not shown). The antagonists had no effect on spontaneous KSHV reactivation. However, Figure 7 shows that they reversed the inhibition of linear DNA synthesis by 1.25 μg/ml THC. Discussion To summarize the antiviral effects of THC and to compare THC with well-characterized antiviral drugs, we compiled data from the literature as well as from our own experimental results. Table 2 shows the 50% inhibitory concentrations (IC 50 ) of four known antiviral drugs and of THC. Data on the antiviral effects of acyclovir, PFA and ganciclovir on KSHV were obtained from our Gardella gel assays [ 32 ]. Usherwood et al. [ 33 ] also used a Gardella gel-based assay to determine the effect of acyclovir against MHV 68 in transformed B cell line S11. The data regarding the effects of four antiviral drugs against MHV 68 and on cell division are from the work of Neyts and De Clercq [ 34 ] using standard lytic virus replication inhibition assays on monolayers. The THC antiviral results are from the data described in this paper (Gardella gels for KSHV using BCBL-1 cells and NIH12 antiviral assays for MHV 68). These data suggest that THC is a potent and selective antiviral agent against KSHV comparable with some well-characterized anti-herpesvirus compounds. THC is even more potent and selective against MHV 68 than acyclovir, ganciclovir and foscarnet. Cidofovir appears to be most potent in these in vitro experiments; however, this drug is known to cause serious side effects and is toxic to the kidney in humans. As outlined earlier, THC can modulate and inhibit the activation of immune cells, so it is not entirely surprising that it can down-regulate the reactivation of viruses residing in lymphocytes, as shown by the data. However, the antiviral effect of THC is not cell specific, since MHV 68 and HVS replication was also strongly inhibited by THC in NIH3T12 or owl monkey kidney monolayer cell cultures. This observation suggests that THC either directly or indirectly targets a viral gene shared by these herpesviruses. The data presented in Figure 7 suggest that THC may inhibit KSHV replication through the cannabinoid receptors. When BCBL-1 cells were treated with THC the receptor antagonists partially reversed this effect, suggesting a role for the CB receptors expressed by BCBL-1 cells. However, more studies are required to evaluate whether CB1, CB2 or both receptors are involved. As discussed in the Introduction, ORF 50/Rta is a critical gene for both reactivation of latent virus and lytic replication in monolayers. Interestingly, ORF 50/Rta activation involves cAMP signaling [ 12 , 35 ]. In contrast, cannabinoid receptor binding has been shown to down-regulate the level of activated CREB through a decrease in cyclic AMP synthesis [ 17 , 18 , 26 , 36 ]. Therefore, one possible explanation is that THC inhibits cAMP signaling, leading to decrease of ORF 50/Rta-mediated transcription and block of virus replication. This antiviral mechanism of THC is supported by our data. Luciferase reporter assays showed that in the presence of THC, initiation of transcription of ORF 50 mRNA in both KSHV and MHV 68 is markedly reduced (4-fold and 7.4-fold, respectively) as compared with the CMV immediate early promoter. These data suggest selective inhibition of the ORF 50 promoter of MHV 68 and KSHV by THC. However, it is also anticipated that THC may also block other cellular and viral genes,, as this drug has been shown to cause a wide range of changes in lymphocyte gene expression [ 17 , 18 ]. Conclusions Early studies have attempted to evaluate whether THC has effects on HSV-1 and HSV-2 replication [ 37 - 44 ]. Most of these studies concluded that THC directly or indirectly enhances replication/reactivation of these viruses, although Lancz et al. [ 42 ] showed that a very high concentration (330 μM) decreases the infectivity of virions. Our data presented in this paper show no effect of THC on HSV replication at lower concentrations. To resolve these conflicting observations, investigation of this issue should continue in the light of new advances in herpesvirus molecular biology and cannabinoid research. Interestingly, statistical analysis indicates a lower incidence of Kaposi's sarcoma in HIV positive women using non-intravenous drugs [ 45 ]. About 5.4% of HIV positive women with no drug use developed KS, whereas none of the 47 women in this study who only used marijuana suffered from KS [ref. [ 45 ], and James Goedert, personal communication]. This report, however, involved relatively few individuals so further analysis of a larger cohort is warranted. We believe that studies on cannabinoids and herpesviruses are important to continue because there are obvious potential benefits. Better understanding may lead to the development of specific non-psychoactive drugs that may inhibit reactivation of oncogenic herpesviruses. Competing interests None declared. Authors' contributions MMM carried out most of these studies. TAS carried out some of the luciferase assays. TWK and HF participated in the design of the study. PGM conceived the study and participated in its design and overall coordination. All authors read and approved the final manuscript. Pre-publication history The pre-publication history for this paper can be accessed here:	/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC521080.xml

Subsets and Splits

No community queries yet

The top public SQL queries from the community will appear here once available.