pmid stringlengths 8 8 | pmcid stringlengths 8 11 ⌀ | source stringclasses 2
values | rank int64 1 9.78k | sections unknown | tokens int64 3 46.7k |
|---|---|---|---|---|---|
28176878 | PMC5296756 | pmc | 1,138 | {
"abstract": "Production of 1,3-propanediol (1,3-PDO) from glycerol is a promising route toward glycerol biorefinery. However, the yield of 1,3-PDO is limited due to the requirement of NADH regeneration via glycerol oxidation process, which generates large amounts of undesired byproducts. Glutamate fermentation by Corynebacterium glutamicum is an important oxidation process generating excess NADH. In this study, we proposed a novel strategy to couple the process of 1,3-PDO synthesis with glutamate production for cofactor regeneration. With the optimization of 1,3-PDO synthesis route, C. glutamicum can efficiently convert glycerol into 1,3-PDO with the yield of ~ 1.0 mol/mol glycerol. Co-production of 1,3-PDO and glutamate was also achieved which increased the yield of glutamate by 18% as compared to the control. Since 1,3-PDO and glutamate can be easily separated in downstream process, this study provides a potential green route for coupled production of 1,3-PDO and glutamate to enhance the economic viability of biorefinery process.",
"discussion": "Discussion Microbial production of 1,3-PDO from renewable resources has been widely investigated in recent years 4 5 11 . Due to the reduction of glycerol price 9 10 , direct conversion of glycerol into 1,3-PDO is becoming an appealing approach for 1,3-PDO production. Among the natural 1,3-PDO producers, K. pneumoniae was mostly studied due to its high productivity 12 42 . However, the industrial application of K. pneumoniae is limited since it is an opportunistic pathogen. Thus, development of a safe 1,3-PDO producer is highly desirable. E. coli as a model organism has been engineered for 1,3-PDO production from glycerol 43 . In this study, C. glutamicum , a GRAS (generally regarded as safe) strain, was engineered for the first time to produce 1,3-PDO from glycerol. C. glutamicum is a widely used industrial workhorse which is currently utilized for production of million tons of amino acids including glutamate and lysine 20 21 22 23 24 . C. glutamicum cannot oxidize glycerol 10 , thus providing a good opportunity to build a strain with high yield of 1,3-PDO. With the introduction of 1,3-PDO synthesis pathway, the engineered C. glutamicum strains can convert almost all of glycerol into 1,3-PDO (~1.0 mol PDO/mol glycerol). The conversion of glycerol to 1,3-PDO is only catalyzed by two enzymes. However, the balance of enzyme activities between the two enzymes is very important. Increasing the expression of diol dehydratase accelerated glycerol consumption rate and 1,3-PDO formation rate. However, the toxic intermediate 3-HPA was accumulated at high initial glycerol concentration which caused the abnormal cessation of fermentation. 3-HPA accumulation was also observed in Enterobacterial agglomerans and K. pneumoniae 3 37 38 39 . Imbalance of activities of diol dehydratase and 1,3-PDO dehydrogenase were also observed during 3-HPA accumulation 3 37 38 39 . Two potential reasons may cause this phenomenon. First, high glycerol concentration could enhance the rate of glycerol dehydration since the Km value of glycerol for diol dehydratase was reported to be relatively high (about 16 mM) 44 . Second, high ratio of NAD/NADH was observed at high glycerol concentration which may also result in the lower rate of 3-HPA reduction 39 . Lim et al . reported that application of UTR (untranslated region) engineering to precisely control the expression of glycerol dehydrogenase for optimal metabolic balance could significantly enhance the production of 3-HP 45 . Similar strategy could also be applied for improving the production of 1,3-PDO. When glucose was utilized as carbon source for cell growth and NADH generation, about ~50 g/L of glucose was consumed for the conversion of 30 g/L of glycerol ( Fig. 5 ). To increase the process economy, we propose that glucose can be oxidized into glutamate, another important industrial product. The cofactor can be recycled for the co-production of 1,3-PDO and glutamate. It was shown that this coupled process increased the yield of glutamate ( Fig. 7 ). Although the titers of 1,3-PDO and glutamate production were not very high, they may be further improved by combining system metabolic engineering strategies and process optimization. Since glutamate and 1,3-PDO can be easily separated in downstream processes, this study provides a promising alternative for 1,3-PDO and glutamate production with high atom economy. The same strategy may be applied for the production of other oxidized and reduced products."
} | 1,133 |
38228651 | PMC10791676 | pmc | 1,139 | {
"abstract": "The symbiogenetic origin of eukaryotes with mitochondria is considered a major evolutionary transition. The initial interactions and conditions of symbiosis, along with the phylogenetic affinity of the host, are widely debated. Here, we focus on a possible evolutionary path toward an association of individuals of two species based on unidirectional syntrophy. With the backing of a theoretical model, we hypothesize that the first step in the evolution of such symbiosis could be the appearance of a linking structure on the symbiont’s membrane, using which it forms an ectocommensalism with its host. We consider a commensalistic model based on the syntrophy hypothesis in the framework of coevolutionary dynamics and mutant invasion into a monomorphic resident system (evolutionary substitution). We investigate the ecological and evolutionary stability of the consortium (or symbiotic merger), with vertical transmissions playing a crucial role. The impact of the ‘effectiveness of vertical transmission’ on the dynamics is also analyzed. We find that the transmission of symbionts and the additional costs incurred by the mutant determine the conditions of fixation of the consortia. Additionally, we observe that small and highly metabolically active symbionts are likely to form the consortia.",
"introduction": "Introduction Examples of symbiotic interactions in which one species coexists in association with another species are frequently observed in nature and extensively studied. There is a plethora of mechanisms by which the interactions between the species form. Though definitions vary, “symbiosis” is often described as an intimate and long-term relationship of individuals of different genomes in which partners may or may not live in physical contact (often physiologically attached) and have already started to coevolve. While there is ambiguity in the definition of symbiosis and the kinds of species interactions that should fall under its purview 1 , here we adopt the above conventional definition of symbiosis 2 , which encompasses obligate symbiosis 3 , 4 . The partnership can be mutually beneficial (mutualism) or asymmetrical with or without a conflict of interest (parasitism or commensalism). Ecological interactions are often complex and can be placed on a continuum varying from fitness-reducing parasitism to fitness-increasing mutualism 5 , 6 . These host-symbiont associations have a long history of coevolution, and several mechanisms, in the long run, might contribute to determining the type of associations 3 , 4 . Establishing a permanent and obligate symbiosis between unicellular organisms that were once capable of independent existence is assumed to have played a crucial role in organellogenesis, eventually leading to eukaryogenesis. According to the endosymbiotic theory 7 , mitochondria and chloroplasts of eukaryotic cells descended from formerly free-living prokaryotes 8 – 13 . During the integration of mitochondria into an ancestor of the eukaryotic cell, a new evolutionary unit and a new level of selection emerged from free-living and independently reproducing cells, and hence is considered a major evolutionary transition 6 , 14 . We develop an ecological model that describes such interactions leading to the symbiosis between free-living (not in physical contact) and unicellular individuals to provide general conditions that eventually facilitate and stabilize such interactions leading to the formation of the symbiotic entity. In other words, we are interested in the first step in the evolutionary origin of a vertically transmitted obligate symbiosis 15 . The resident ecological system that we consider is formed by two free-living species (host and symbiont), and the existence and survival of at least one species depend on the other (commensalism). This ecological dependence between these species is crucial for theoretical modeling perspectives too. One commonly proposed ecological dependence between species is syntrophy, i.e., one species living off the metabolic products of another species 8 , 9 , 12 , 13 , 16 , 17 . It is widely assumed that interactions based on syntrophy might have played a significant part in the emergence of major endosymbiotic transitions like the origin of mitochondria as an organelle 8 , 9 , 12 , 13 , 18 . Various hypotheses of eukaryogenesis also assume syntrophic interactions (hydrogen hypothesis, sulfur hypothesis, etc.) 13 between once-independent organisms. This paper considers the initial ecological system of two species in which the host (pre-eukaryote) produces the “food” of the symbiont (ancestor of mitochondria) species. In contrast, the host is neither benefited nor harmed by the interactions. Thereby, the association can be called “syntrophic commensalism.” It is improbable that the host and symbiont, at their very first interaction and onset of their association, turned out to be immediately mutualistic with ecological compatibility. Hence, we argue that it is plausible to consider the association as a commensalistic one at the beginning, with the potential to eventually evolve into a parasitic or mutualistic one. Reciprocative metabolic exchange due to complementation (mutual or cyclic syntrophy) is assumed to be a critical factor in establishing microbial associations and forming microbial consortia, biofilms, and mats 12 . However, our model considers linear unidirectional syntrophy rather than complex mutual syntrophy. Cyclic mutual syntrophy is much more advanced, and hence, it is reasonable to assume that a relatively less sophisticated mechanism of linear and unidirectional syntrophy existed initially. For the same reason, we can assume that linear unidirectional syntrophy was more abundant, i.e., more species had this kind of relation than cyclic syntrophy. In addition, cyclic syntrophy is an example of mutualism, and our argument is that mutualism may not have evolved at the onset of interactions between two independent species. The mechanism we focus on was likely an initial step, pre-dating more complex mutualisms. Acknowledging that the initial steps in the evolution of intimate symbiotic behavior between unicellular organisms are widely debated, we narrow our scope to a single mechanism that could set off such an association. Considering that the evolution of endosymbiosis in one step is unlikely, we hypothesize that the initial event in the evolution of physically attached intimate symbiosis between free-living organisms could be a single mutation in the symbiont genome that led to the formation of a new structure that enables it to latch onto the external surface of the host (ectosymbiosis). This single mutation is likely the first step in the evolution of intimate symbiosis. In our model, we focus on this mutated symbiont that has the potential to form an ectosymbiotic linkage with the host. The formation of this consortium benefits the symbiont because it gains more straightforward access to the substrate from the host. In this scenario, we assume that during the multiplication of the host, the ectosymbiont is not lost because it is latched onto the host’s membrane; thus, vertical transmission leads to the maintenance of the association over generations. Hence, the evolution of long-term association between the host and the symbiont species with vertical transmission is a significant concern in the evolution of eukaryotes 19 , and our model formulation tries to incorporate this. However, microbial syntrophy cannot explain how one partner gets inside the other 13 . We suggest that the development of phagocytotic features could be a later step in the process rather than a trait that evolved before intimate symbiosis. The model can be extended to understand potential endosymbiosis, primarily from the integration of mitochondria. For instance, if the host predated the ectosymbiont by phagocytosis, that could have eventually opened the door for endosymbiosis (when the symbiont could survive within the host) 18 . In addition, adhesion to the exterior membrane might increase the likelihood of endosymbiosis. The main difference between the phagocytosis and syntrophy hypotheses is that, by phagocytosis, endosymbiosis evolved directly, while by syntrophy, ectosymbiosis evolved first. However, in both scenarios, the long-term association of the two species evolved with vertical transmission. We introduce a coevolutionary model and are interested in whether the mutant symbiont can invade the resident ecological system. We construct a dynamical system that follows the densities of populations corresponding to our ecological model. In addition to the free-living hosts and symbionts (resident or wild type), we consider the consortia and mutant symbionts as separate “populations.” Hence, the dimension of the dynamical system is increased by two with the introduction of a single mutation, and thereby we consider a multi-species dynamical system. The successful invasion of a stable resident system by mutant phenotypes has been discussed previously 20 – 23 . In this paper, we are interested in the outcome of evolution by which the resident phenotype of one of the species (symbiont) is replaced by the mutant phenotype of the same species. We assume that mutation is rare enough and that multiple mutant phenotypes in the same or several species do not occur simultaneously. Evolutionary substitution occurs if the mutant phenotype can invade a stable equilibrium of the monomorphic resident system. The system evolves to a stable equilibrium of the multi-species resident-mutant coevolutionary system where the species with two phenotypes has only the mutant phenotype surviving 20 , 21 , 24 . Using this concept, we investigate the ecological and evolutionary stability of inter-species consortia in the framework of an ectocommensalistic association based on syntrophy. This paper aims to obtain a set of sufficient conditions that ensure the existence and local stability of the positive equilibrium point of the coevolutionary resident-mutant system with the mutant phenotype. The successful invasion and fixation of the mutant symbiont would ensure non-zero consortia density. We provide sufficient conditions for the associations to emerge and for the symbiotic merger of the host and the symbiont to stabilize in the ecological system. In essence, we consider a commensalistic system based on the syntrophy hypothesis to study a symbiogenetic model in a multi-level selection approach.",
"discussion": "Discussion, future directions, and biological examples We found that if the ratio of cost of living to cost of reproduction of the mutant phenotype is less than that of the resident phenotype of the symbiont and the rate of addition of free-living mutant symbionts into the habitat is high (possibility of horizontal transmission), then the host-symbiont association or consortium is fixed by natural selection. We started with a commensalistic ecological interaction where the host’s metabolic product is food to the symbiont. We assume that a mutation occurs at this resident ecological equilibrium. The mutant symbionts can be joined by a novel structure to the external surface of the hosts, hence forming an ectocommensalistic association. Using a theoretical model, we analyzed the invasion of the mutant phenotype into the resident system. Even though both costs of living and reproduction of the mutant phenotype are greater than that of the resident phenotype, it is worthwhile to note that it is the ratio of the costs that matter, i.e., the additional costs of the connecting structure play a significant factor in the fixation of the consortia in the system. Eukaryogenesis, as noted in the paper, is a complex process that requires many steps. The endosymbiosis of mitochondria, the origin of the nucleus, and the evolution of a sophisticated cytoskeleton, which ultimately allows for phagocytosis, eukaryotic genome organization, etc., are all steps in this transition. Each evolutionary step needs to be advantageous; thus, mostly, we need to assume small changes. We assumed one small change: the appearance (through mutation) of a molecule that can bind a commensal species to the host. This is one small step in the evolutionary path toward eukaryogenesis. As we are agnostic on the exact order of events, the host cell could have some eukaryote-like features, but none is needed for our model to work (for example, the host could have a nucleus, which would not change our results). Furthermore, vertical transmission (when the symbionts are directly transferred from a host to its offspring) guarantees the association of the host and its symbiont in the next generation. One of the main consequences of vertical transmission is that the evolutionary success of the symbiont and its host is determined by that of the host-symbiont association 26 . Hence, the single mutation forms a new level of selection based on the fact that the symbionts can be vertically transmitted, and the consortia can be maintained, i.e., host and symbiont stick together and replicate. The main consequence of the stable physical contact between the host and symbiont species is that a new evolutionary unit and replicator type appears, namely the “consortium.” Studies show that the subsequent evolution of associations that probably started as a commensal one might lean towards parasitism in species with horizontal transmission, while vertical transmission promotes mutualism 27 , 28 . Evolution from horizontal transmission and parasitism towards vertical transmission and mutualism has also been discussed previously 29 , 30 . Thus, evidence supports the argument that the mode of transmission highly influences evolution toward mutualism or parasitism. In other words, vertical transmission opens the door to the coevolution of the host-symbiont pair in the direction of mutualism. One of the essential steps in the evolution of intimate and obligate symbiosis between primitive unicellular organisms could be a mutation that allows vertical transmission, just as we considered. Thus, the emergence of vertical transmission should be one of the critical questions in the evolution of symbiosis. Conclusively, after the fixation of the host-symbiont association in the system, the main direction for evolutionary mechanisms is that the association can evolve to be mutualistic. For example, at some point in time, excess concentrations of the metabolic waste product of the host in its proximity turn out to be toxic to itself, and the symbiont plays a significant role in reducing this concentration below a critical level, thereby helping the host in its survival. In this scenario, mutualism can emerge because each member benefits the other. This can be subject to further study with a toxicity-dependent bounding of species densities. Another possibility is the evolution of cyclic syntrophy, i.e., as a result of a mutation on the host, it can use at least one of the metabolic products of its ectosymbionts (considering the symbiont has not changed itself by other mutations). This is similar to the concept of evolutionary replacement discussed by Cressman and Garay 24 . Unlike evolutionary substitution, evolutionary replacement requires that there are mutant phenotypes in each of the species. In this scenario too, a costless mutualism can emerge. Moreover, since the host and the symbiont are different species, we can assume that the metabolic abilities of the species are quite different, and thereby the symbiont can potentially produce metabolic products that the host is incapable of producing itself and can use as food. In the case of mutualistic interactions, vertical transmission is more evolutionarily rational 27 since multi-level selection can enforce better association. In other words, such associations have a beneficial connection for both the host and its vertically transmitted symbiont; thus, the reproduction rate of this association should be higher than other possibilities. Hence, the mutations which force vertical transmission should invade the population. We assumed that the host was totally impassive toward the symbionts to make the model more straightforward. Also, we made some assumptions on the parameters of the model to maintain tractability. However, relaxing some of the assumptions in the model could be an immediate extension of the work. For instance, a responsive host that detects the presence of ectosymbionts, whose feeding is impacted by the ectosymbionts. Our model can be extended further with appropriate modifications to analyze endosymbiotic associations as well. Moreover, vertical transmission is more likely for the endosymbiont than for the ectosymbiont. For instance, suppose the ectosymbiont can move into the host cell by engulfment via phagocytosis or slowly increasing surface contact during ectosymbiotic syntrophy or by bacterial invasion 13 . After this, the “consensus” of the metabolisms of the host and its internalized symbionts can lead to the emergence of endosymbiosis, for instance, in the case of the mitochondrial ancestor 8 , 9 , 12 . Our study was motivated by the origin of eukaryotes, a major evolutionary transition, and the emergence of a new level of individuality. We try to provide a theoretical analysis of a hypothesis on the symbiogenetic origins of multi-species intimate associations, particularly a eukaryotic cell. The concept of syntrophy is prevalent among the several hypotheses trying to deepen the understanding of complex evolutionary transitions 13 , hence the need to model it mathematically. The novel growth rate introduced in the model can be utilized to structure interactions based on syntrophy and develop metabolite-dependent ecological dynamics. Utilizing the concept of mutant invasion to form an ectocommensalism in a dynamical setup to understand symbiosis is also a novel approach. The paper attempts to model (as a dynamical system) syntrophic interactions leading to the formation of symbiotic behavior between primitive unicellular organisms. Ectosymbiosis is indeed common in nature, even among prokaryotes. Recently, several ectosymbionts belonging to the Candidate Phyla Radiation clade of Bacteria and DPANN clade of Archaea have been identified in ectosymbiosis with a Bacterial or Archaeal host 31 – 33 . It demonstrates that prokaryotes are capable of being both the host and the symbiont in these associations. Nanoarchaeum equitans 34 is the most well-studied symbiont (or parasite; the exact nature of the association is still debated), which is attached to cells of Ignicoccus hospitalis as its host. Since their discovery, other nanoarchaeal ectosymbionts were identified, such as Nanoclepta minutus and its host Zestosphaera tikiterensis 35 , Nanopusillus acidilobi and its host an Acidilobus species 36 , and Nanobdella aerobiophila and its host, Metallosphaera sedula 37 . Similar symbiont-host pairs were found among the other small-celled Archaeal phylum, the Nanohaloarchaeota: Nanohalobium constans with a Halomicrobium species 38 , Micrarchaeum harzensis with Scheffleriplasma hospitalis 39 , and Nanohaloarchaeum antarcticus with its host Halorubrum lacusprofundi 40 . From the Candidate Phyla Radiation, Saccharibacteria (TM7) should be mentioned as they are ectosymbionts of various host bacteria, including Actinomyces species 41 , 42 living in the human oral microbiome. The interaction is probably initiated by the symbiont, as it is the symbiont that cannot live without its host. The host can live (can be cultured) without its ectosymbiont. Ultrastructural analysis also uncovers cell surface structures attaching to and penetrating the outer layers of the host 38 , 39 , 43 , and in many cases, a pili-like structure is observed at the attachment site 44 – 46 . Pili is an established way for archaea 45 and bacteria 44 to attach to another cell. Still, the exact way the ectosymbiont attaches to its host is unknown. Studies have implicated various proteins, such as the WD-40 protein family of Ignicoccus hospitalis 46 , SPEARE proteins of Nanohaloarchaea 40 , or proteins involved in type 4 pili formation 47 in the case of Nanoarchaeota. Proteomic analysis of the contact site of the two Archaea revealed some putative transporters and membrane proteins 48 , which other studies, mostly genetic, corroborate 49 . The attachment to the host is quite strong, and a given host cell is covered by clones derived from a single symbiotic cell 47 ; thus, the ectosymbionts proliferate on the host cell’s surface. In the end, there could be several symbionts on any host cell: around 10 Nanoarchaeum equitans on Ignicoccus hospitalis 50 , between 1 and 10 Nanopusillus acidilobi on Acidilobus species 36 , and about 4–5 (up to 17) Nanohalobium constans on Halomicrobium species. Our model is built on these observations. We hope our attempt opens a different perspective of modeling and analyzing unicellular symbiosis."
} | 5,257 |
35733669 | PMC9159792 | pmc | 1,141 | {
"abstract": "Production of biopolymers from renewable carbon sources provides a path towards a circular economy. This review compares several existing and emerging approaches for polyhydroxyalkanoate (PHA) production from soluble organic and gaseous carbon sources and considers technologies based on pure and mixed microbial cultures. While bioplastics are most often produced from soluble sources of organic carbon, the use of carbon dioxide (CO 2 ) as the carbon source for PHA production is emerging as a sustainable approach that combines CO 2 sequestration with the production of a value-added product. Techno-economic analysis suggests that the emerging approach of CO 2 conversion to carboxylic acids by microbial electrosynthesis followed by microbial PHA production could lead to a novel cost-efficient technology for production of green biopolymers.",
"conclusion": "4. Conclusion This review provided an overview of experimental and TEA studies of bioplastics production from various carbon sources. Such comparison of results in available literature showed that the technical and economic feasibility of bioplastics production depends on multiple factors including carbon source, process design, operating conditions, etc. However, the most prominent factors affecting production costs are PHA yield, productivity, and the type of carbon source selected for PHA production. In fact, carbon source selection appears to be the most important, as it affects both the productivity and the overall PHA yield per unit of carbon source consumed. By comparing different TEA studies, it was shown that by using well-established carbon sources, such as glucose and glycerol and assuming a production capacity of 9000 tons per year, PHA production costs fall within the range of $6.9–$7.5 per kg. Also, several TEA studies suggest that when using more complex carbon sources such as wastewater or agricultural biomass, in which an additional carbon source preparation step is necessary, the unit cost of PHA production can range from $5.2 to $11.0 per kg. This range of costs is similar to that obtained when using well-established carbon sources, since feedstock costs are considered to be zero, and credits can be applied to reduce production costs. Interestingly, our estimation of production costs using the emerging approach of bioplastics production through microbial electrosynthesis from CO 2 suggests that this approach could provide a feasible alternative to traditional carbon sources, such as glucose. Although a broad range of production costs was obtained ($5.71–$78) due to the uncertainties of this novel process, future studies are expected to result in significant improvements in the observed process yield and productivity. A detailed TEA study is needed to evaluate the feasibility of direct CO 2 conversion to bioplastics.",
"introduction": "1. Introduction Bio-based materials produced from renewable sources of organic carbon instead of petroleum hydrocarbons can play an important role in reducing consumption of fossil fuels and moving our society towards a circular economy. Polyhydroxyalkanoates (PHAs), which are produced for storage of carbon and energy by a large number of microorganisms within the bacterial and archaea domains, are often used for bioplastic production. The insoluble PHA granules inside the microorganisms can make up to 90% of the dry weight of the cell mass. 1 More than 150 types of PHAs have been identified. The most common form of PHA is polyhydroxybutyrate (PHB). Depending on the composition and properties of the PHA, applications can range from use in biodegradable packaging, to use as chemical additives, to usage in the fields of medicine, agriculture, wastewater treatment, and cosmetics. 2–4 In spite of multiple benefits of using biopolymers, their commercialization continues to be problematic due to high biopolymer production costs compared to polymers produced from conventional feedstock. Indeed, the price of polypropylene and polyethylene is about US $1.25–2.53 per kg, 5 while that for PHAs has been reported to be up to 16 times higher than the major petroleum-derived polymers. 6 According to a study of the global PHA market for 2018, 7 the average PHA price was US $8.0 per kg. This price varied according to the target application and quality of the PHAs. In the same study, the average price of PHAs destined for packaging and food services was calculated to be US $7.6 per kg with an estimated market size of US $25 000 000, while for biomedical applications the average price was US $11.9 per kg with a market size of US $15 000 000. In 2019, the global PHA market was estimated to be US $57 000 000 with a projected compound annual growth rate of 11.2%. 7 Integration of biopolymers into the global market can be facilitated through a thorough cost analysis and identification of technologies capable of reducing production costs, while minimizing environmental impact. Energy consumption, PHA yield, and the efficiency of the downstream processing are the most important parameters determining the cost of production. 8 Carbon sources used in pure culture microbial fermentations contribute significantly to overall environmental impact and production costs. Therefore, the use of mixed microbial communities capable of PHA production from liquid and gaseous waste streams, such as food wastes, 9 agricultural wastes, 10,11 landfill gas, carbon dioxide (CO 2 ), wastewater, 12 polystyrene waste, 13 and glycerol 14,15 is seen as a sustainable approach for bioplastics production. Biopolymers produced from CO 2 are of particular interest, as this approach also provides a sustainable method for utilization of CO 2 captured from industrial off-gases and from air. 16 In this review, PHA production technologies are described based on the type of carbon source (liquid or gaseous) used. Also, single-stage and two-stage production processes are considered. Typically, a single-stage production can be accomplished if a well-defined liquid carbon source is used, either with pure or mixed microbial cultures. If more complex carbon sources such as agro-industrial wastes or flue gases are used, the production of PHAs must be carried out in two stages, where the conversion of complex carbon sources into simple sugars or carboxylic acids is followed by a stage of PHA production. Finally, a novel approach of biopolymers production from CO 2 and electrons in a microbial electrosynthesis (MES) system is reviewed. An overview of these technologies is followed by the review of techno-economic assessments evaluating the costs of PHA production, including PHA production from CO 2 through MES. By combining a detailed review of PHA production technologies with a review of published techno-economic assessments, this study helps to select promising cost-efficient technologies for producing biopolymers from renewable carbon sources, including waste biomass and CO 2 ."
} | 1,730 |
35212604 | PMC8973982 | pmc | 1,142 | {
"abstract": "ABSTRACT In the past decades, considerable attention has been directed toward anaerobic digestion (AD), which is an effective biological process for converting diverse organic wastes into biogas, volatile fatty acids (VFAs), biohydrogen, etc. The microbial bioprocessing takes part during AD is of substantial significance, and one of the crucial approaches for the deep and adequate understanding and manipulating it toward different products is process microbiology. Due to highly complexity of AD microbiome, it is critically important to study the involved microorganisms in AD. In recent years, in addition to traditional methods, novel molecular techniques and meta-omics approaches have been developed which provide accurate details about microbial communities involved AD. Better understanding of process microbiomes could guide us in identifying and controlling various factors in both improving the AD process and diverting metabolic pathway toward production of selective bio-products. This review covers various platforms of AD process that results in different final products from microbiological point of view. The review also highlights distinctive interactions occurring among microbial communities. Furthermore, assessment of these communities existing in the anaerobic digesters is discussed to provide more insights into their structure, dynamics, and metabolic pathways. Moreover, the important factors affecting microbial communities in each platform of AD are highlighted. Finally, the review provides some recent applications of AD for the production of novel bio-products and deals with challenges and future perspectives of AD.",
"conclusion": "6. Conclusion and future perspectives Regarding worldwide concerns about climate change, widespread pollution, and scientists’ efforts for the reduction of carbon footprint, biomass conversion technologies, especially AD are promising platforms. A brilliant prospect for the AD process is conceivable as a sustainable technology, in which the biochemical energy of organic matters can be efficiently extracted. The AD process is a low-priced and multi-purpose technology, which can use various feedstocks such as food wastes, various wastewaters, agricultural residues, sludge, manure, etc. AD is a very complex biological process, and intensive efforts are still required to determine effective factors and optimum conditions for stabilization, higher yield, and productivity of new high-value products such as hydrogen and VFAs. Moreover, some impediments influence the AD process that should be overcome to achieve the great potential of this process for bioenergy and other products production. In this way, both traditional and cutting-edge methods are necessary to attain these goals. Precise prediction, process monitoring, real-time controlling, and modeling of microbial communities’ performance in the AD system would be promising and informative approaches for improved AD operation. Additionally, to investigate microbial communities’ complexity, their structure, function, activities, and interactions during AD, omics approaches have had successful impacts in recent years. These approaches may reveal an accurate vision of microbial and viral biodiversity and the spatial organization of microbial communities in different anaerobic digesters. However, all involved species in the AD process are not known and identified completely. Providing an entire dataset of all these species, their function during four stages of AD, and their metabolic capacity may guide researchers to operate anaerobic digesters under controlled situations. Comprehensively, these findings allow us to conduct the process better and shift the various phases based on the final products desired. Furthermore, the findings can bridge among microbiologists, bioinformatics scientists, and chemical engineers to discover novel microbial communities, metabolic pathways, or products of the AD process.",
"introduction": "1. Introduction The dependence of the world community on non-renewable energy resources to maintain quality of life, sustain economic development and to enable our vast transportation network is perhaps one of the most important problems facing the world today. Dwindling reserves with rapidly increasing consumption rates, combined with unstable energy prices and the environmental concerns, especially climate change demand that there is an urgent need to develop a sustainable, affordable, and environmentally friendly energy resources. Moreover, world population growth and rapid industrialization resulted in the generation of enormous wastes such as food waste, solid municipal waste, organic waste, agricultural residues, etc. that expected to increase to 70% by 2050. These wastes can be valorized into renewable energy sources and chemicals through different technologies including biological approaches [ 1 , 2 ]. Anaerobic digestion (AD) is one of the most promising alternatives to non-renewable energy resources [ 3 ]. To visualize recent distinguished work on AD, various databases were explored herein to acquire the suitable publications and data regarding this topic in 2021–2022 ( Figure 1 ).\n Figure 1. The bibliometric mapping of anaerobic digestion in 2021–2022. AD – a process in which organic materials are digested in the absence of molecular oxygen – has been employed widely for various purposes such as biogas production, waste management, and pathogen deactivation among others. In the natural environments, where oxygen content is limited, such as landfills, sediments, waterlogged soils, or intestinal tracts of ruminants, anaerobic environment exist [ 4 , 5 ]. Assyrians were the first people who employed AD to warm bathwater. In the 16th century, Persians also used AD to heat water. In the 17th century, Alessandro Volta recognized that anaerobic degradation of organic compounds could produce flammable gases such as methane during his summer trip to Lake Maggiore. However, until the 1880s, AD was not used in a full-scale application. In 1895, AD was employed in a hybrid system to treat wastewater of the city of Exeter in the United Kingdom. Two years later, in 1897, digestion plant was built up in Matunga, India, with biogas collection system. Afterwards, in the 20th century, a two-stage system (Travis Tank) was developed and subsequently this process was widely used to treat wastewaters, municipal solid wastes, sewage sludge, manures, industrial organic wastes, etc. [ 6 , 7 ]. During the 1930s, microbiologists made considerable efforts to understand the mechanisms of biogas production in the anaerobic digesters. However, the application of the AD process was limited until 1950 due to the lack of understanding of the fundamentals of AD. With a better understanding of the AD process, various small and industrial applications were developed [ 7 ]. Moreover, developments of modern techniques as well as technical equipment and also advances in the recovery of produced gases have reduced the total cost resulting in broad applications of AD. The AD process provides a precisely balanced ecological environment that successive break down of the organic macromolecules, i.e. carbohydrates, proteins, and lipids to soluble organics, which are subsequently converted into biogas by diverse groups of bacteria and archaea in the absence of oxygen [ 8–10 ]. These microorganisms work interactively in four complicated and interdependent biochemical reactions, namely hydrolysis, acidogenesis, acetogenesis, and methanogenesis that resulting in the degradation of organic materials and production of H 2 , CO 2 and CH 4 as well as H 2 S in trace amounts [ 11–14 ]. In Table 1 , some examples of various microorganisms involved in the different AD phases are shown. AD is traditionally used for waste treatment and bioenergy production, but it is being developed as a new platform in which other bio-products can be produced. However, several microbiological and operational challenges need to be resolved to achieve feasible platforms of the AD process for various final products. Extensively, microbiological aspects of AD process may differ from other industrial processes. Because in this process, sterility or pure cultures might not be necessary, while novel microbiological procedures or techniques such as whole-genome sequencing (WGS), next-generation sequencing (NGS), comparative analysis, microcosms studies, and omics approaches such as genomics, metagenomics, transcriptomics, metatranscriptomics, proteomics, metaproteomics, metabolomics, or meta-metabolomics can be more effective and applicable [ 15–17 ]. Table 1. Microorganisms involved in the various phases of the AD process [ 4 , 6 , 13 , 25 , 41 , 55 ] AD Phase Microbial Domain Microbial Genus Examples of Identified Species Hydrolysis and Fermentation Bacteria Acetivibrio, Aminobacterium, Aminomonas, Anaeromusa, Anaerosphaera Bacillus, Bacteroides, Bifidobacterium, Butyrivibrio Caldanaerobacter, Caldicellulosiruptor, Campylobacter, Cellulomonas, Clostridium Devosia Espiroquetas. Eubacterium Fervidobacterium, Fibrobacter, Fusobacterium Gelria, Gracilibacter Halocella Lactobacillus Paludibacter, Peptococcus, Peptoniphilus, Proteiniborus, Pseudomonas, Psychrobacter Ralstonia, Ruminoclostridium, Ruminococcus Selenomonas, Shewanella, Sporanaerobacter, Streptococcus, Streptomyces Thermanaerovibrio, Thermomonas, Thermomonospora, Thermotoga, Treponema, Trichococcus Pseudomonas mendocina Bacillus halodurans Clostridium hastiforme Gracilibacter thermotolerans Thermomonas haemolytica Fungi Aspergillus Humicola Penicillium Trichoderma Trichoderma reesei Acetogenesis Bacteria Acetobacterium Clostridium Desulfotignum Eubacterium Holophaga Moorella Ruminococcus Sporomusa Thermoanaerobacter, Treponema Moorella thermoacetica Desulfotignum phosphitoxidans Holophaga foetida Methanogenesis Archaea Methanobacterium, Methanobrevibacter, Methanococcus, Methanoculleus, Methanosaeta, Methanomicrobium, Methanosarcina, Methanospirillum, Methanothermobacter Methanobrevibacter smithii Methanobrevibacter arboriphilus Methanococcus vannielii \n Microbial diversity of anaerobic digesters depends on various factors such as feedstock type, seed inoculum, temperature, granulation, aeration, mixing speed, pre-treatment type, digester design, organic loading rate (OLR), solids retention time (SRT) and hydraulic retention time (HRT). For example, by increasing the temperature, the diversity of archaeal and bacterial communities decreases considerably. Moreover, short HRT and high OLR are related to Acidobacteria community while long HRT and low OLR are associated with Planctomycetes, Actinobacteria , and Alcaligenaceae [ 13 , 18 , 19 ]. Overall, most known microbial communities in the anaerobic digesters are prokaryotic ones, while some eukaryotic microorganisms take part in the digestion process such as fungi and protozoa. Anaerobic fungi (particularly the phylum Neocallimastigomycota ) found in the microbial communities are responsible for the hydrolysis process [ 13 , 20–22 ]. However, these hydrolytic fungi grow slowly, affecting their frequency in the anaerobic digesters. For example, Neocallimastix is an anaerobic fungal genus found in the rumen and produces a wide range of hydrolytic enzymes such as xylanase, cellulase, and esterase. The co-culture of this fungus with methanogens resulted in direct methane production from cellulose. Other anaerobic genera involved in the AD of cellulose belong to Anaeromyces, Caecomyces, Orpinomyces , and Piromyces [ 4 , 23 , 24 ]. Among prokaryotes, over 80% of the whole diversity is related to the domain Bacteria. The commonly identified bacterial phyla include Proteobacteria, Firmicutes , and Bacteroidetes . The latter phylum has shown hydrolytic activity, and it seems that the members of this phylum dominate under a low level of volatile fatty acids (VFAs), salts, and ammonia at mesophilic temperature in anaerobic digesters [ 25 , 26 ]. The other bacterial phyla such as Actinobacteria, Planctomycetes, Chloroflexi, Fibrobacteres, Deferribacteres, Fusobacteria, Synergistetes, Nitrospira, Acidobacteria, Tenericutes, Spirochaetes, Verrucomicrobia , and Thermotogae do sometimes exist. Regarding the domain Archaea, most identified phylum in the anaerobic digesters is the phylum Euryarchaeota . However, there are some new microorganisms that are not assigned to any microbial taxonomy and introduced as the ‘ Candidatus’ [ 4 , 8 , 13 ]. Likewise, unculturable ‘ Candidatus’ has been detected in the anaerobic digesters (e.g. OP10, BA024, OP8, TM6, EM3, OP3, and OS-K in the domain Bacteria) [ 27 ]. The literature revealed that most studies focused on the optimal conditions for biogas production and waste remediation [ 28–31 ]. Such reviews are informative, but none of them comprehensively discusses the microbiological aspects of AD platforms in which new bio-products with various biotechnological applications can be produced. Discussion on digester configurations and operating conditions of anaerobic digesters can be found in elsewhere [ 9 , 17 , 32–34 ]. This review was aimed to provide an inclusive vision of AD microbiology, the function of microbial communities and various factors involved, novel bio-products, and recent advances of the AD process. This work focused on bio-products generation from the AD process with special emphasis on process microbiology, assessment of microbial communities, and factors affecting their abundance. For this purpose, it was systematically reviewed how microbial communities function and relationships led to various bio-products production during AD. In addition, this review concluded with perspectives and challenges highlighting future research directions."
} | 3,440 |
20150502 | null | s2 | 1,143 | {
"abstract": "Microorganisms can switch from a planktonic, free-swimming life-style to a sessile, colonial state, called a biofilm, which confers resistance to environmental stress. Conversion between the motile and biofilm life-styles has been attributed to increased levels of the prokaryotic second messenger cyclic di-guanosine monophosphate (c-di-GMP), yet the signaling mechanisms mediating such a global switch are poorly understood. Here we show that the transcriptional regulator VpsT from Vibrio cholerae directly senses c-di-GMP to inversely control extracellular matrix production and motility, which identifies VpsT as a master regulator for biofilm formation. Rather than being regulated by phosphorylation, VpsT undergoes a change in oligomerization on c-di-GMP binding."
} | 192 |
24190511 | PMC3817431 | pmc | 1,145 | {
"abstract": "Artificial skin, which mimics the functions of natural skin, will be very important in the future for robots used by humans in daily life. However, combining skin's pressure sensitivity and mechanical self-healing properties in a man-made material remains a challenging task. Here, we show that graphene and polymers can be integrated into a thin film which mimics both the mechanical self-healing and pressure sensitivity behavior of natural skin without any external power supply. Its ultimate strain and tensile strength are even two and ten times larger than the corresponding values of human skin, respectively. It also demonstrates highly stable sensitivity to a very light touch (0.02 kPa), even in bending or stretching states.",
"discussion": "Discussion We have demonstrated a strong and stretchable self-healing thin film material with self-activated pressure sensitivity. This work differs from that of other groups who focus on sensitivity 37 , mechanical properties 38 39 40 , or self-healing 13 , but miss other vital functions. The strategy used for the design and fabrication of this hybrid film is the integration of nanoscale building blocks into a macroscopic structure, and is currently at the forefront of nanotechnology applications. In our design, the self-healing and pressure sensitivity properties are attributed to PDMAA and PVDF building blocks, respectively. Thus, the device demonstrates similar behaviors to those of the individual components. For example, its self-healing process occurs in air or water at different temperatures and times ( Supplementary Fig. S15 ), which is similar to the behavior of pure PDMAA and PDMAA hydrogels 32 ; owing to the response time of piezoelectric PVDF layer, the pressure-response curves of the 5S films show hysteresis ( Fig. 3 a–c ). In the magnified views of single cycle ( Supplementary Fig. S16 ), the response time obtained from the double positive peak is found to be about 0.4 s, which is close to other reported values 27 28 . The double peak shown in the single cycle illustrates the process that objects fall on the sensor: initially touching the surface of electrodes, and completely acting on the piezoelectric layer when the 3D porous electrode is pressed to form a compact structure ( Supplementary Fig. S16 ). Such detailed observations indicate that our device may have the potential to measure the subtle pressure in real life. In fact, tens of pascals, lower than the typical pressure of a gentle finger touch (~2 kPa) 41 , can be detected by the present device, which compares very favourably with the sensitivity achieved in previous artificial skins: < 0.2 kPa (ref. 1 ); < 6 kPa (ref. 2 ); < ~ 2 kPa (ref. 5 ); ~1 kPa (ref. 42 ). Moreover, the device can reliably sense pressure values ( Fig. 3a, 3e, 3f and Supplementary Fig. S17 ). Its electrical and mechanical reliabilities are comparable to those of previous works 1 5 42 . However, we notice that the voltage generated by the 5S film doesn't show a high degree of stability at a frequency of about 1.5 Hz ( Fig. 3c ). It is probably due to the incomplete recovery of 3D structural PDMAA-PVA/rGO electrodes from the compressive deformation during these intervals. As shown in Supplementary Video 1 , PDMAA recovers its shape after deformation very gradually. We further investigated the responses of the device under lower cycling frequencies ranging from 0.2 to 1 Hz for the same applied pressure. With the decrease in the testing frequency, less uneven output signals are recorded ( Supplementary Fig. S18 ). This demonstrates that the device is suitable for use at a range of low frequencies, which is common in real life 43 , for example, in walking, turning over papers and blinking eyes. The next grand challenge is to combine tactile resolution in our experiment. An interesting path forward is to integrate floating-gate transistor with 5S films. By integrating a flexible array of floating-gate transistors with pressure-sensitive rubbers, Takao Someya and his colleagues have fabricated a sensor matrix that detects the spatial distribution of applied pressure 44 . Pressure sensors, source and drain contacts, and capacitors are key components in their device. In order to exploit the self-healing function in such integrated device, one needs to find replacement for each component. The recent achievements in self-healing wires 45 and self-healing semiconductors 46 are very encouraging. By employing there materials and our 5S films in such design, a self-healing artificial skin that resolves touch profiles is hopefully realized. In conclusion, we have integrated graphene and polymers into a thin film which mimicked both the mechanical self-healing and pressure sensitivity behavior of natural skin without any external power supply. Its ultimate strain and tensile strength are even two and ten times larger than the corresponding values of human skin, respectively. It also demonstrated high sensitivity to a very light touch, even in bending or stretching states."
} | 1,255 |
29555779 | PMC5889658 | pmc | 1,146 | {
"abstract": "Significance Bacteria live in surface-attached communities known as biofilms, where spatial structure is tightly linked to community function. We have developed a genetically encoded biofilm patterning tool (“Biofilm Lithography”) by engineering bacteria such that the expression of membrane adhesion proteins responsible for surface attachment is optically regulated. Accordingly, these bacteria only form biofilm on illuminated surface regions. With this tool, we are able to use blue light to pattern Escherichia coli biofilms with 25 μm spatial resolution. We present an accompanying biophysical model to understand the mechanism behind light-regulated biofilm formation and to provide insight on related natural biofilm processes. Overall, this biofilm patterning tool can be applied to study natural microbial communities as well as to engineer living biomaterials.",
"discussion": "Discussion In conclusion, we have developed a method (Biofilm Lithography) that uses light-regulated adhesin expression (pDawn-Ag43) to quantitatively control biofilm formation and patterning with high spatial resolution. Compared with existing cell deposition and patterning approaches such as inkjet printing ( 12 , 13 ), microcontact printing ( 14 , 15 ), microfluidics ( 20 ), PDMS stenciling ( 16 ), and patterned substrate modification ( 17 – 19 ), this method can pattern on a variety of surfaces, without the need for surface prepatterning or pretreatment, within enclosed chambers, over large areas, and at high spatial resolution. Rapid prototyping of different biofilm geometries is possible with low-cost digital projectors at resolutions of 45 μ m; resolutions down to 25 μ m can be reached with photomasks and likely also with more advanced optical projector setups ( 44 , 45 ). This resolution represents an important step toward the engineering of biofilm communities, as natural biofilm microcolonies exist around this length scale ( 10 ). Our biophysical model suggests that the pDawn-Ag43 patterning mechanism works alongside natural surface adsorption/desorption in bacteria and involves the stimulation of transiently adsorbed cells on the biofilm substrate toward a more permanently attached state. This insight gives additional support to the proposed role of Ag43 as an adhesin involved in biofilm maturation as opposed to initial surface adsorption ( 46 ). Ultimately, optogenetic patterning tools such as pDawn-Ag43 can be applied toward an improved understanding of naturally existing biofilms ( 47 , 48 ), the design of synthetic microbial consortia ( 8 ), distributed metabolic engineering ( 49 ), and new types of integrated diagnostic and microfluidic devices ( 50 ), with impact and trajectory that may potentially parallel that of silicon photolithography in the semiconductor industry ( 51 , 52 )."
} | 702 |
33192236 | PMC7649386 | pmc | 1,148 | {
"abstract": "The hardware-software co-optimization of neural network architectures is a field of research that emerged with the advent of commercial neuromorphic chips, such as the IBM TrueNorth and Intel Loihi. Development of simulation and automated mapping software tools in tandem with the design of neuromorphic hardware, whilst taking into consideration the hardware constraints, will play an increasingly significant role in deployment of system-level applications. This paper illustrates the importance and benefits of co-design of convolutional neural networks (CNN) that are to be mapped onto neuromorphic hardware with a crossbar array of synapses. Toward this end, we first study which convolution techniques are more hardware friendly and propose different mapping techniques for different convolutions. We show that, for a seven-layered CNN, our proposed mapping technique can reduce the number of cores used by 4.9–13.8 times for crossbar sizes ranging from 128 × 256 to 1,024 × 1,024, and this can be compared to the toeplitz method of mapping. We next develop an iterative co-design process for the systematic design of more hardware-friendly CNNs whilst considering hardware constraints, such as core sizes. A python wrapper, developed for the mapping process, is also useful for validating hardware design and studies on traffic volume and energy consumption. Finally, a new neural network dubbed HFNet is proposed using the above co-design process; it achieves a classification accuracy of 71.3% on the IMAGENET dataset (comparable to the VGG-16) but uses 11 times less cores for neuromorphic hardware with core size of 1,024 × 1,024. We also modified the HFNet to fit onto different core sizes and report on the corresponding classification accuracies. Various aspects of the paper are patent pending.",
"introduction": "1. Introduction Over the past decade, GPUs have emerged as a major hardware resource for deep learning tasks. However, fields, such as the internet of things (IoT) and edge computing are constantly in need of more efficient neural-network-specific hardware (Basu et al., 2018 ; Deng et al., 2018 ; Alyamkin et al., 2019 ; Roy et al., 2019 ). This encourages competition among companies, such as Intel, IBM, and others to propose new hardware alternatives, leading to the emergence of commercially available deep learning accelerators (Barry et al., 2015 ; Jouppi et al., 2017 ) and neuromorphic chips (Esser et al., 2016 ; Davies et al., 2018 ; Pei et al., 2019 ). Deep learning accelerators are application specific integrated circuits (ASICs) tailored for artificial neural networks (ANN), whereas, neuromorphic chips can fall in two categories (Bose et al., 2019 ): (1) ASICs with biologically inspired spiking neural networks (SNN), which contain networks of neurons and synapses for computation and communication, or (2) ASICs with analog computing by exploiting dense non-volatile memory based crossbars to accelerate matrix-vector multiplications. Our paper is not concerned with any specific hardware but any neuromorphic architecture relying on analog crossbars for matrix-vector multiplications. A schematic of a generic crossbar-based neuromorphic chip is shown in Figure 1 . The chip has “N” number of neuromorphic cores. Network on chip (NoC) or router interfaces are not shown for illustration purposes. Each neuromorphic core contains a crossbar array of synapses as shown in the first inset of the figure. The rows and columns of the crossbar correspond to input axons and output neurons, respectively. These axons and neurons are interconnected to each other and represented in the form of blue dots at these intersections. Within each intersection of the crossbar between the word line and the bit line is a synaptic device that has memory and can perform in-memory computation (as shown in the second inset). The crossbar structure is well-suited for performing matrix-vector multiplication (MVM) (Hu et al., 2016 ) along each column in a crossbar architecture. For instance, a neuromorphic core with a core size of 256 × 256, input voltages from the respective 256 axons are fed through each word line (red horizontal lines). The bit line (yellow vertical lines) collects all of the weighted current at each synaptic node (256 × 256) and delivers to the respective output neurons for integration. The neuromorphic core size refers to the number of axons (axon size, AS) × the number of neurons (neuron size, NS) in a single neuromorphic core. The weighted current depends on the memory element at each intersection of the word line and bit line. In analog devices, using Kirchoff's current law, the total current flowing into each neuron from the respective bit lines is the sum of currents flowing through each intersection in every column. This corresponds well with how inputs in a neural network is the weighted sum of input voltages (∑( Input × Weight )). Considering such a neuromorphic chip, there are several hardware constraints: firstly, at the single device, we may have low bit precision of synaptic weights and output activations (Ji et al., 2018 ; Deng et al., 2020 ), synaptic noise and variability (Ambrogio et al., 2014a , b ). Secondly, in the chip architecture, we have a limited number of neuromorphic cores and a limitation in the core size of each neuromorphic core and the fan-in/fan-out degree of each neuron (Ji et al., 2018 ; Gopalakrishnan et al., 2019b ). Figure 1 A schematic of a neuromorphic chip with N number of neuromorphic cores. The first inset shows the crossbar array of synapses within each core. A memory device is used to implement each synapse at the crossbar intersection (as shown in the second inset). The neuromorphic chip considered in this paper is based on a crossbar architecture (Prezioso et al., 2015 ) of non-volatile memory synapses. Crossbar architecture fits well for fully connected neural networks, such as the multi-layered perceptron (MLP). Given that one of the popularly used layers in SNNs are fully connected ones (Diehl and Cook, 2015 ), crossbar architectures are a natural fit. However, with recent advancement in research related to conversion of ANNs to SNNs (Rueckauer et al., 2016 ) and training of convolutional SNNs (Wu et al., 2019 ), one of the main challenges is to efficiently map the neurons in a CNN onto the neuromorphic chip while fulfilling hardware constraints, such as core size, number of cores and fan-in/fan-out (Ji et al., 2016 ). Given existing CNNs and neuromorphic hardwares, how can we best map the CNN onto the hardware using the least number of cores? This requires understanding of the computation at each crossbar array and how best to map each convolutional layer onto the core. We may then ask what convolution layers are best suited for mapping onto neuromorphic hardwares. This is the first major contribution of this paper and these questions are addressed under section Mapping. Existing neuromorphic chips have a mapping framework which is hardware specific. IBM TrueNorth (Akopyan et al., 2015 ) uses Matlab based Corelet (Amir et al., 2013 ) which is specific to their hardware. Within Corelet, a mapping technique is implemented as a minimization problem (Akopyan et al., 2015 ). SpiNNaker and BrainScaleS use a simulator-independent Python wrapper, PyNN (Andrew et al., 2009 ). Sequential mapping is used in SpiNNaker while neural engineering framework (NEF) is used for Neurogrid (Voelker et al., 2017 ). Neutrams (Ji et al., 2016 ) implements an optimized mapping technique based on the graph partition problem: Kernighan-Lin (KL) partitioning strategy for network on chip (NoC). For mapping CNNs onto neuromorphic hardwares, an iterative process is implemented using a Python wrapper, which is also discussed in section 2.3.2. While developing deep neural networks that are to be mapped onto a neuromorphic chip, one need not in principle be aware of the underlying hardware architecture. The mapping above assumes that the CNNs and hardware constraints are given. We may however further ask how software and hardware co-design can give us both CNNs and neuromorphic hardwares that are mapped using fewer cores while achieving similar classification accuracies. Specifically, given a neuromophic hardware with square crossbar array, we would like to design a CNN that utilizes fewer cores (section Co-design). In this regard, one may take two approaches, either design the network from scratch so as to satisfy the hardware constraints (Esser et al., 2016 ) or modify an existing CNN, such as reducing the number of features (feature maps) in each convolution layer without having to split the convolution matrix among different cores (“core matrix splitting”) whereby axons and weight matrix of a particular layer are split onto multiple cores (detailed in subsection 2.2.1). This is the second major contribution of this paper, and the proposed novel hardware-friendly CNN, HFNet, is obtained by iteratively modifying the layers of existing CNNs (VGG, MobileNet, NIN, and Squeezenet; this is discussed in section 2) and the number of features (feature maps) in some layers are altered so as to fit onto cores of different sizes. This is done to avoid core matrix splitting. Finally, the different versions of HFNet are trained and their classification accuracies on the IMAGENET dataset (Deng et al., 2009 ) tested so as to study the trade-off between accuracies and core sizes (section 3.2). This work is mainly focused on mapping of feedforward deep neural networks (DNN). During the investigation of mapping techniques, we understood that designing and mapping of a CNN must be performed in close relation to each other for better hardware utilization. Hence, as a beginning, we have limited the design space to just feedforward networks instead of skipped connections. We have considered the better-performing feedforward CNN MobileNet as an initial candidate for mapping and later modified to HFNet based on mapping and traditional deep learning techniques. The trained CNNs have full precision weights and activation values for fair comparison with existing CNNs. Hardware limitations on synapses, such as low precision weights and variability issues are not within the scope of our work. The paper is organized as follows. Section 2 mainly contains two subsections, one on mapping and another on co-design. Mapping describes the computation and mapping in a crossbar array. Co-design is illustrated with the issue of core matrix splitting and then the motivation and design flow of the proposed hardware-friendly neural network, HFNet. Section 3 provides an experimental framework for the two subsections, mapping and co-design in section 2. The classification accuracy of different versions of HFNet on the IMAGENET dataset is included here. The paper is concluded in section 4 with a discussion of future works.",
"discussion": "4. Discussion and Conclusion In our work, we first study what convolutions are more hardware friendly and how to best map them onto a neuromorphic hardware. We next identify deep learning techniques to avoid which result in poor core utilization or even result in core matrix splitting. We then propose a framework for the design of more hardware-friendly CNNs, and implement it using a Python wrapper (MaD). As a result of the above, the HFNet is proposed. Different versions of the HFNet are also proposed using the framework, that have better classification accuracy with more cores used when mapped. The framework thus allows us to propose different CNNs in a more principled manner by changing design parameters. We then evaluate it by comparison with other CNNs in terms of classification accuracy on IMAGENET, number of parameters, and cores required for mapping. It is able to achieve very comparable accuracy, using about the same number of parameters as per other more hardware-friendly CNNs but with substantially fewer mapped cores. Here, we have shown results on one of the biggest and most popular datasets in visual classification, IMAGENET, our results are quite generally applicable for visual tasks which is already covering a very big application space. Second, it has been shown that networks pre-trained on IMAGENET can be used as feature extractors for spectrograms for audio analysis (Acharya and Basu, 2020 ); our method can thus potentially generalize to other types of datasets as well. We have also proposed an optimized mapping technique by considering a square shaped selection of neurons. As shown in Table 6 , we further explore how different core shapes (64 × 4,096, 128 × 2,048, 256 × 1,024, 512 × 512, 1,024 × 256, 2,048 × 128, 4,096 × 64) affect the number of cores used to map the HFNet-V3. Overall, the trend in cores used agree with the intuition provided in the proof (section 2.2.1), which is based on real numbers; while core sizes are based on integers which may lead to some discrepancy. We further note that cores required is not symmetric about 512 × 512, with larger input dimensions using less cores (e.g., 4,096 × 64 against 64 × 4,096). This is due to avoiding core matrix split while mapping. Table 6 Number of cores for different core shapes. Core shape Number of cores 64 × 4,096 111,642 128 × 2,048 58,098 256 × 1,024 26,424 512 × 512 14,263 1,024 × 256 14,069 2,048 × 128 28,240 4,096 × 64 56,480 The HFNet is a hardware-friendly CNN that is designed using an iterative process that takes into account how best it can be mapped on a neuromorphic hardware with crossbar array of synapses while achieving good accuracy. One hard constraint while mapping is avoiding core matrix splitting. As shown in section 3, a typical HFNet requires thousands of cores for mapping. This is still larger than most neuromorphic chips. For future work, we will consider how one may map the HFNet onto a neuromorphic chip with limited number of cores. While we try to optimize synaptic resources and reduce “wastage” by minimizing unused synapses, it might be possible to reuse these synapses to provide a degree of fault tolerance in the hardware by providing redundancy. Current explorations of fault tolerance mostly show reduction in performance degradation after retraining a neural network with faults (Lee et al., 2014 ; Feinberg et al., 2018 ); however, there might be scope to optimize the fault tolerance by providing some extra synapses. We feel this is an important avenue of future work. We would also consider quantized CNNs, both weight and activations, in future work, which is beyond the scope of current work. Other hardware constraints, such as synaptic noise in novel devices, will be considered in future work. Considering chip area, it maybe that only one physical neuron is implemented per core. This neuron is utilized in a time multiplexed manner to emulate all neurons within the core. If, however, there is more than one physical neuron per core, one can speed up computation by pipelining the time multiplexed neurons. Further speed-up can be achieved if the number of fan-in neurons per convolution operation is considered while increasing the physical neurons. In our current study, we only studied CNNs without skip connections. Residual networks (He et al., 2016 ) have skip connections that increase the fan-in/fan-out degree of neurons. Not only that, to map them, we would either have to save intermediate results of skip-connections at routers or in buffers at the axons. This would be interesting to consider in future work."
} | 3,873 |
29472284 | PMC5861367 | pmc | 1,149 | {
"abstract": "ABSTRACT The concept of symbiosis – defined in 1879 by de Bary as ‘the living together of unlike organisms’ – has a rich and convoluted history in biology. In part, because it questioned the concept of the individual, symbiosis fell largely outside mainstream science and has traditionally received less attention than other research disciplines. This is gradually changing. In nature organisms do not live in isolation but rather interact with, and are impacted by, diverse beings throughout their life histories. Symbiosis is now recognized as a central driver of evolution across the entire tree of life, including, for example, bacterial endosymbionts that provide insects with vital nutrients and the mitochondria that power our own cells. Symbioses between microbes and their multicellular hosts also underpin the ecological success of some of the most productive ecosystems on the planet, including hydrothermal vents and coral reefs. In November 2017, scientists working in fields spanning the life sciences came together at a Company of Biologists’ workshop to discuss the origin, maintenance, and long-term implications of symbiosis from the complementary perspectives of cell biology, ecology, evolution and genomics, taking into account both model and non-model organisms. Here, we provide a brief synthesis of the fruitful discussions that transpired.",
"introduction": "Introduction In recent years, symbiosis has gained recognition as one of the most important evolutionary processes shaping biodiversity throughout the history of life on Earth. Generally speaking, symbiosis refers to any type of intimate and long-term interaction between different organisms. A recent multidisciplinary workshop, supported by The Company of Biologists, entitled ‘Symbiosis in the microbial world: from ecology to genome evolution’ brought together researchers working at the forefront of the field to discuss symbioses involving the most numerically abundant and functionally diverse organisms on the planet, the microbes (which comprise bacteria, archaea and protists, as well as the viruses that infect them), and their interactions with multicellular hosts. These microbial symbioses range from metabolic ( McCutcheon and Moran, 2012 ) and defensive interactions ( Oliver et al., 2014 ) among free-living organisms, to the complete cellular and genomic integration that occurred during the endosymbiotic origins of mitochondria and chloroplasts in eukaryotic cells ( Embley and Martin, 2006 ; Roger et al., 2017 ). Symbiosis provides an unparalleled route to evolutionary innovation, one that underlies some of the most important transitions in the history of life. Owing to recent methodological breakthroughs, symbiosis research is undergoing a revolution. Characterising the genetic potential and metabolic capabilities of symbiotic partners has traditionally been challenging because most symbionts defy commonly applied enrichment and cultivation techniques. While many microbes are difficult to cultivate, symbionts may pose additional challenges because they often rely on interactions with other organisms in order to survive and, particularly in the case of endosymbionts, can rarely be cultivated on their own. However, the recent application of metagenomic and single-cell genomic approaches to the study of symbiosis now circumvents some of these issues by enabling the reconstruction of genomes from symbionts in their natural habitats ( Siegl et al., 2011 ; Woyke et al., 2006 ). These techniques have greatly expanded our ability to sample existing symbiotic diversity and improved our understanding of interactions among and between microbes in the environment. This flood of new data has been complemented by proteomics ( Mao and Franke, 2015 ), transfection and transformation systems enabling genetic manipulation of a wide range of organisms, and recent advances in experimental techniques such as single-cell imaging, microfluidics ( Lambert et al., 2017 ), in situ hybridization, and secondary-ion mass-spectrometry (SIMS), allowing intracellular measurement of metabolic fluxes ( Thompson et al., 2012 ). Collectively, these developments have opened up entirely new lines of symbiosis research, bringing both classical and emerging questions into the realm of tractable science. Together with increased recognition of the fundamental importance of symbiosis to many areas of biology, this growth of activity is reflected in the rapidly expanding body of literature on the subject, including over 2500 publications in 2016 alone. Current research is proceeding on multiple fronts – from ecologists studying the diversity of microbial communities over large spatial scales to cell and evolutionary biologists investigating the long-term impacts of symbiosis on cell organization and genome evolution. The breadth of approaches and perspectives being brought to bear on symbiosis is a strength, but also represents a challenge as it involves researchers from different backgrounds who need to develop a shared language.\n Workshop participants convened at Wiston House in Sussex, UK, with the aim of discussing the cellular, ecological, and evolutionary aspects of symbiosis, and its role in the history of life. Exciting new results on a broad range of symbiotic systems were presented, ranging from lab experiments on binary interactions between ciliates and their photosynthetic algal symbionts, to broad-scale analyses of complex microbial communities, such as those living in and on coral reefs. Collectively, these works employed a diversity of methodological approaches, including both traditional and cutting-edge cellular and molecular biology techniques, high-resolution imaging, molecular phylogenetics, and various ‘omics’ tools. The overall goal of the workshop was not only to stimulate fruitful discussions and to establish an integrative framework for research between all these fields, but also to identify the most important contemporary questions in the field of symbiosis research, questions that can only be tackled collaboratively by combining different tools, approaches and expertise. Here, we highlight points of consensus and controversy within and among these different fields and identify areas of opportunity for future multidisciplinary work. Symbiosis: what's in a name? While the symbiosis research community is relatively small, its practitioners work in a variety of different areas and use diverse and often non-overlapping methodological approaches to explore a myriad of organismal associations, time scales and biological problems. Symbiotic associations span a gradient that includes mutualistic, commensal and even parasitic relationships. In addition, these associations can shift over ecological and evolutionary time and in response to changes in environmental conditions and community composition. Symbioses are often cast as facultative, ‘beneficial’ metabolic interactions between organisms that can evolve into obligatory interdependencies over time. Symbioses also vary in their level of cellular and genetic integration; they include ecto- and endosymbiotic interactions, in which an organism lives on the surface or within the cell(s) of another organism, respectively. The most extreme cases of integration are the mitochondria and chloroplasts of eukaryotes, endosymbiotically-derived organelles that have long since lost their cellular autonomy ( Archibald, 2015 ; Embley and Martin, 2006 ; Roger et al., 2017 ). At the other end of the spectrum are interactions between multicellular organisms and the microbes that live on and within them. The study of symbiosis leads to a broad range of questions, only some of which are easily applied to all systems. Indeed, given its tremendous scope, it is difficult to define what symbiosis is and what it is not. To what extent is the co-evolution between animals and their microbiomes symbiotic? Does the animal microbiome and its host represent a unit of selection and can/should it be considered a holobiont ( Douglas and Werren, 2016 ; Skillings, 2016 )? Which level of metabolic interaction and/or trophic relationship constitutes a symbiosis ( Orphan, 2009 ; Schink, 2002 )? When does an endosymbiont become an organelle (and how much does it matter) ( McCutcheon and Keeling, 2014 ; Singer et al., 2017 )? These are some of the questions that symbiosis researchers continue to grapple with. Reductionist and holistic approaches to symbiosis research Some of the most spirited debates at the workshop centred on the scales at which questions about symbiosis can be most effectively addressed. These discussions were illustrative in that they made explicit certain differences in the accepted standards for evidence and methodological approaches between researchers working with tractable laboratory model systems on one hand, and those investigating the structure of complex natural communities on the other. Clearly, there are challenges in translating correlations and co-occurrence patterns reported in ecosystem and global-scale observational microbiome studies to specific, experimentally-tested functional interactions between partners. At the same time, we must also recognize that laboratory models do not necessarily fully capture the diversity and variability of symbiotic interactions that occur in nature, since the most tractable systems often involve few interacting partners. Debates between reductionism and holism are common in science, but are particularly acute in symbiosis research because the strategy used often varies depending on the system being studied. Most accounts of the evolution of tightly-integrated, inter-dependent symbioses – as exemplified by the symbiotic bacteria of many insects ( Moran et al., 2008 ) or the eukaryotic cell ( Martin et al., 2015 ) – envisage an initial weak or transient interaction between the partners that evolves to become more stable and tightly integrated over time through neutral and/or adaptive processes ( Lukeš et al., 2011 ; Szathmáry, 2015 ; Timmis et al., 2004 ; West et al., 2015 ). If this scenario is generally correct, then holistic and reductionist approaches are perhaps best suited to studying different ends of the symbiotic continuum, from a complex mix of mostly transiently interacting organisms to a much smaller set of tightly integrated partners. Top-down and bottom-up approaches to symbiosis research can be complementary: experimental work on lab models is clearly essential for providing fundamental mechanistic insight into how symbiosis works. At the same time, observational and whole-community analyses can generate hypotheses to be tested with established models, and can also suggest which new model systems need to be brought into the lab – an expensive and time-consuming process – in order to address the major outstanding questions. Microbial community stability over space and time Another nascent dimension of symbiosis research is the focus on understanding the evolutionary and ecological processes that drive the changes in patterns of symbiosis observed over short and long time scales. The first challenge is to determine how stable symbiotic microbial communities are over time and how much can be generalized from a small number of observations of natural systems that are not easily tamed in the lab. However, the study of community composition over time has revealed that some systems show high levels of variability while others are extremely stable. In order to understand these patterns, it becomes imperative to not only take into account the high-level taxonomic diversity that comprises ecological communities, but also the functional traits that are associated with each taxon. It might, therefore, also be important to model symbiont systems based on their functional traits in addition to their taxonomic composition, because cases have been described in which the former remains stable, while the latter appears to vary ( Lozupone et al., 2012 ), at least at certain levels of functional and taxonomic granularity [Douglas has referred to this as the ‘inconstant microbiome’ ( Wong et al., 2013 )]. Seen from this perspective, perhaps the most relevant unit of selection is the metabolic function performed by the interacting unit, given that similar processes can be performed by taxa (or genes) that are only distantly related: according to Doolittle and Booth, ‘it's the song, not the singer’ ( Doolittle and Booth, 2017 ). The meta-omics black box: from data to biology The inferences derived from high-throughput analyses of environmental DNA, protein sequences and/or chemical compounds are only as strong as the databases used to annotate them, and a major current roadblock is the prevalence of genes, proteins and molecules with no known function. A recent effort to define the minimal genome required for a self-replicating bacterium provided a humbling perspective: out of the 473 genes in Mycobacterium supporting a viable and free-living cell, 32% have an unknown function ( Hutchison et al., 2016 ). This highlights our very incomplete understanding of the molecular basis of vital biological processes. Currently, 50-80% of meta-omics data in hand cannot be annotated, which leads to an incomplete picture of the systems being studied by restricting the interpretation of the results to biochemical pathways and cellular processes that are already well understood. This issue is frequently encountered in symbiotic systems, which can be reservoirs of novel accessory genes due to niche-specificity ( Porter et al., 2016 ; Remigi et al., 2016 ) and the lack of cultivated representatives of numerous microbial partners. Such genes might play particularly important roles in symbiotic relationships and could thus represent important targets for future studies – if only they can be identified. Unassigned data should therefore not be dismissed, but we should instead encourage the development and use of novel analytical tools capable of delving into their coding potential and putative functions ( Hartmann et al., 2017 ). Furthermore, it is important to keep in mind that functions inferred for proteins with homologous sequences in current databases cannot be fully relied upon. For instance, even if general enzymatic properties are conserved, substrate specificity and/or reaction directionality can often not be predicted based on homology alone ( Laso-Pérez et al., 2016 ). Therefore, hypotheses on the biology of host-symbiont interactions based solely on genomic data should ideally be experimentally validated. Future directions in symbiosis research Many ecologically important host-symbiont systems cannot be easily cultivated or genetically manipulated. However, microbial isolation is returning to the spotlight and the coming years are likely to see new advances in axenic culturing techniques ( Overmann et al., 2017 ). The semantic shift from ‘unculturable’ to ‘not yet cultivated’ is a very encouraging sign and some microbes long thought to be obligate intracellular symbionts are now grown axenically ( Omsland et al., 2013 ). Metabolic pathway reconstruction of uncultured bacteria can already be used to predict their nutrient requirements and rationally design new culture conditions. In the near future, this will enable us to not only get a better understanding of the biology of organisms involved in symbiosis, but also to genetically manipulate them, which will in turn lead to greater insights into the mechanisms that regulate symbiotic interactions and host colonization. Other avenues for future research should also include the development of techniques to identify bacterial symbionts in natural communities ( Orphan, 2009 ). This could be achievable by identifying phenotypic or genomic traits that are predictive of symbiotic interactions ( Moran and Wernegreen, 2000 ) and might help to decipher how symbionts are acquired or transmitted. In addition to these technical developments, significant efforts should be made to generate high-quality reference genomes from single-celled eukaryotes, which comprise most of eukaryotic diversity. We will need such data in order to make proper sense of metagenomic and metatranscriptomic datasets generated from diverse environments, as well as to fully grasp the diversity of symbiotic relationships in nature ( Sibbald and Archibald, 2017 ). Over deeper evolutionary timescales, there is still much to learn about how and when the mitochondrial endosymbiosis occurred and its role in the origin of the eukaryotic cell ( Eme et al., 2017 ; Roger et al. 2017 ). Future sequencing and cultivation efforts will hopefully allow us to identify and study close relatives of the elusive prokaryotic ancestors of eukaryotes ( Spang et al., 2015 ; Zaremba-Niedzwiedzka et al., 2017 ), thereby allowing us to refine hypotheses on the origin of the eukaryotic cell ( Eme et al., 2017 ). Given the breadth and novelty of the work presented at the workshop, the future is undoubtedly bright for symbiosis research. Methodological advances combined with efforts to further stimulate multidisciplinary approaches will inevitably provide profound insights into microbial symbioses and unveil fundamental aspects of the complex interactions that characterize life on Earth."
} | 4,323 |
31717384 | PMC6915370 | pmc | 1,151 | {
"abstract": "Developing an eco-friendly, flexible and recyclable micro-structured dry electrode for sustainable life is essential. In this work, we have developed irregular, micro-structured sandpaper coated with graphite powder as an electrode for developing a simple, low-cost, contact-separation mode graphite-coated sandpaper-based triboelectric nanogenerator (GS-TENG) as a self-powered device and biomechanical sensor. The as-fabricated GS-TENG is a dielectric-conductor model. It is made up of a bottom layer with polytetrafluoroethylene (PTFE) as a triboelectric layer, which is attached onto a graphite-coated sandpaper-based electrode and a top layer with aluminum as another triboelectric layer as well as an electrode. The forward and reverse open-circuit voltages reach upto ~33.8 V and ~36.62 V respectively, and the forward and reverse short-circuit currents are ~2.16 µA and ~2.17µA, respectively. The output generated by GS-TENG can power 120 blue light-emitting diodes connected in series, liquid crystal display and can charge commercial capacitors along with the rectifier circuit. The capacitor of 22 µF is charged upto 5 V and is sufficient to drive digital watch as wearable electronics. Moreover, the device can track signals generated by human motion, hence it scavenges biomechanical energy. Thus, GS-TENG facilitates large-scale fabrication and has potential for future applications in wearable and portable devices.",
"conclusion": "4. Conclusion In summary, a novel micro/nano-mesh network sandpaper electrode fabrication process was reported by brush coating of graphite powder. The electrode fabricated is risk-free at the manufacturer and consumer stages, providing durable and disposable features without harming the environment. Moreover, it showed an excellent mechanical stability after integrating with triboelectric layer PTFE, without undergoing any change in electrical property. The as-fabricated GS-TENG is economical, adaptable and can be adopted with any wearable devices to harvest biomechanical energy from the surrounding environment. The maximum peak-to-peak open-circuit voltage, short-circuit current and power density obtained were 36 V, 2.17 µA and 0.941 µW·cm -2 . The output obtained could be able to charge capacitors and power LEDs, LCD and smart watches by hand tapping. Moreover, GS-TENG acts as a self-powered, active biomechanical sensor by distinguishing the output signals generated from the various human body motions. Thus, sandpaper as a substrate is one of the preferable choices to integrate over cellulose-based paper to enhance the durability of the device, when working with harsh environmental conditions. Since we proved sandpaper’s potential role in electrode fabrication, it could be one of the options to use as a synergistic material in the fabrication of low-cost TENG.",
"introduction": "1. Introduction Nowadays, wearable and portable smart electronic devices, such as flexible display [ 1 ], wearable sensors [ 2 , 3 , 4 ], flexible batteries [ 1 , 5 , 6 ] and sustainable power sources [ 7 , 8 , 9 , 10 ], have gained much attention due to their potential applications. Most of these devices depend on electricity and rechargeable batteries, to maintain the continuous and independent working of electronic devices, and the desire to replace batteries has raised great interest for researchers to find alternative power sources. Since most commonly used conventional batteries as a power source limits their applications due to their bulk size, monitoring limited life-time remains as a technological threat and has to be replaced [ 11 ]. Green energy harvesting from the living environment and utilizing it effectively to power all possible electronic devices has become an urgent need [ 12 , 13 , 14 ]. Therefore, a new power source with miniaturization, which is flexible, sustainable and maintenance-free, is greatly needed and could be used widely in health monitoring [ 15 , 16 ], wearable sensors [ 2 , 3 ], nano-robotics [ 17 ], remote environment sensors [ 18 ] and actuators [ 19 ]. As a solution, a triboelectric nanogenerator (TENG) has recently been invented as a promising power source, which provides a compelling route to convert energy from mechanical form to electrical, based on the conjunction of two physical phenomena, triboelectrification (also called contact-electrification) and static-induction [ 20 ]. The human motion-related mechanical energy such as hand tapping, tilting neck, finger motion, wrist movement, bending knees, stamping foot etc., can be effectively harvested by TENGs, converting the bio-mechanical energy into electrical energy [ 21 , 22 ]. Harvesting energy from flexible wearable TENGs by human body movements shows its unique advantage as a self-powered active motion sensor. These electronic devices are based on flexible polymers with suitable flexible electrodes, which have manufacturability, durability and capability of integration with other technologies. Electrodes are extremely essential components for TENG for its active functioning. Commonly, plastic-based electrodes are integrated with TENGs, as they are highly flexible, light-weight and convenient [ 23 , 24 ]. However, they are not suitable to work in high-temperature environments due to deformation and high thermal coefficient of expansion [ 25 ], which hinders their potential application. Most commonly, as an alternative to plastic electrodes, cellulose paper-based electrodes are found to be reliable and facile for the fabrication of TENGs due to its flexibility and fiber-like surface structures, which provide a larger surface area to bind conductive materials [ 26 , 27 , 28 ]. So far, solution-based electrodes by wet technique [ 29 ], embedding silver nanowires on paper through the infiltration technique [ 30 ], stacking metal films on paper [ 31 ] and the penciling paper technique [ 32 ] are adopted to make conductive paper. In all these methods, fiber structures on the surface of paper play a vital role to adhere any of the conductive materials onto it. Despite the considerable progress made in paper-based electrodes/TENG, coating-conducting materials on paper are usually pasting or wet solutions, which are likely to degrade and are not durable for long-term use. Therefore, it is desirable to find another strategy to develop flexible, corrosion-free, low-cost, metal-free, durable and robust electrodes/TENGs with cost-effective large-scale production. Sandpaper can be a promising candidate for electrode fabrication due to economical, flexible, durable and earth benevolent. Sandpaper is assessed according to the grit size (number of holes per linear inch in a sieve screen) [ 33 ]. Also, it inherently creates nano-micro scale roughness levels [ 34 , 35 ]. In comparison with traditional substrates for the TENG, it cut down the time-consuming and additional expensive fabrication procedures to achieve nano-micro roughness to enhance the effective contact area of the TENG. The significance of a sandpaper-based electrode depends on the risk-free manufacturing procedure, recyclable or disposable and economical, along with the necessary requirements. In this work, we demonstrate a flexible, contact-separation mode graphite-coated sandpaper-based triboelectric nanogenerator (GS-TENG) by a simple fabrication technique. In the GS-TENG, the bottom layer is made of polytetrafluoroethylene (PTFE) as a friction layer, sandpaper-based electrode, and the top layer of the GS-TENG consists of aluminum (Al) as a friction layer as well as an electrode. The sandpaper-based electrode is prepared by coating graphite powder on to the sandpaper using a brush. To ensure good conductivity and mechanical durability, the sandpaper-based electrode along with PTFE is subjected to a bending test. Under repeated cycles of bending and releasing, the resistance of the electrode is ~12.5 kΩ, remains almost constant throughout the bending cycles for more than 30 h. The as-fabricated GS-TENG is subjected to forward and reverse polarity tests for the validity of signal generation, the peak-to-peak forward and reverse open-circuit voltages up to ~33.8 V and ~36.62 V respectively, and the peak-to-peak forward and reverse short-circuit currents up to ~2.16 µA and ~2.17µA respectively, were obtained under external vibration at frequency 4 Hz. The energy conversion efficiency is found to be ~7.7% and the maximum peak power density reached upto ~0.94 µW·cm −2 at a load resistance of 30 MΩ. This is sufficient to light-up more than 120 blue light-emitting diodes (LEDs) connected in series and to power a liquid crystal display (LCD). In addition, by conducting a charging ability test of GS-TENG by charging several capacitors with different capacitance by hand tapping, the stored energy is utilized to power a smart watch. Furthermore, GS-TENG can be used as a self-powered human motion sensor, which can track signals generated by the human body, such as finger tapping, wrist movement, hand tapping and foot stepping. Thus, as-fabricated GS-TENG facilitates robust, low-cost, and ease of fabrication, which illuminates its potential for future applications as a sustainable power source and for wearable motion sensors.",
"discussion": "3. Results and Discussion 3.1. Characterizations of the Graphite-Coated Sandpaper-Based Triboelectric Nanogenerator (GS-TENG) Micro/nano-scale of silicon carbide particles as abrasives embedded on sandpaper provides an irregular surface texture. Graphite powder is uniformly spread on to the micro/nano-meshes in the sandpaper using a nylon hair acrylic paint brush. Then, the graphite-coated sandpaper/PET is hot-pressed using the roll-to-roll technique and finally, graphite-coated sandpaper/PET is used as an electrode for further device fabrication. The more detailed fabrication process is discussed in Section 2.1 . To study the surface morphology of the sandpaper-based electrode, the scanning electron microscopic (SEM) images at each stage of electrode fabrication were taken. A typical SEM image of bare sandpaper with the grit size 400, as depicted in Figure 3 a at lower and higher magnification ( Figure 3 b), shows a high density of non-uniform micro/nano-meshes. The vertical cross-sectional SEM reveals sharp edges of micro/nano-meshes’ groves ( Figure 3 c), revealing sandpaper roughness, there is a micro gap between the sandpaper base and groves tips, as indicated from the green dashed lines. This space is well utilized to coat conductive graphite powder, so that amorphous graphite powder evenly fills the textured surface of sandpaper. Moreover, a contact angle (CA) of 126° ( Figure 3 d) is formed when a water drop is placed on top of the sandpaper, owing to the presence of micro/nano-meshes, which can withstand in a harsh environment. The SEM image of graphite-coated sandpaper is shown in Figure 3 e. It can be seen that the micro/nano-groves are completely filled after graphite coating, which confirms that the graphite powder is uniformly distributed on the surface of sandpaper filling the gaps between micro/nano-groves (at higher magnification in Figure 3 f). The cross-sectional view of graphite-coated sandpaper, as shown in Figure 3 g, depicts adhesion of graphite powder on sandpaper after being subjected to roll-to-roll hot pressing. Further, Figure 3 h shows the wettability test through (CA)measurement of graphite-coated sandpaper. In general, graphite powder is hydrophobic [ 38 ], but after coating on sandpaper, the contact angle measured is 70°. This is primarily attributed to the fact that the meshes in the sandpaper lead to the asymmetrical (bumpy) layer of graphite coverage, which can be seen in the highly magnified SEM image ( Figure 3 f). Thus, showing the virtual hydrophilic nature of graphite-coated sandpaper confirms that micro/nano-meshes were completely occupied with the graphite layer. Further, PTFE film tape is stacked on the as-prepared sandpaper-based electrode, as schematically shown in Figure 2 . The cross-sectional view of the SEM image ( Figure 3 i) shows that the binding tendency of PTFE with the coated graphite on sandpaper is weak. Noticeably, there is a wide micro-gap of a few micrometers approximately between PTFE and graphite-coated sandpaper, which is due to the fact that the uneven grooves cannot stick well with PTFE. Also, a layer of graphite adhered to the PTFE fails to extensive binding with sandpaper (indicated with a circle). Hence, the stacked structure undergoes the roll-to-roll press to ensure good conductivity, as shown in the schematic in Figure 2 . The stack thickness of ~300 µm is passed between the rollers of 200 µm space, due to pressing pressure, each layer is firmly bound, which is shown in the cross-sectional SEM image ( Figure 3 j), forming a bottom layer of GS-TENG. Furthermore, the bottom layer of the GS-TENG (PTFE/graphite-coated sandpaper/PET) is subjected to a mechanical deformation test by measuring resistance under the repeated bending and releasing cycle of PTFE/graphite-coated sandpaper/PET. The resistance of the graphite-coated sandpaper electrode obtained is ~12.5 kΩ, which remained almost constant throughout the bending cycles for more than 30 h, as shown in Figure 3 k, which ensures good conductivity and mechanical stability of the electrode. Supplementary Table S1 summarizes the unique characteristics of the graphite-coated sandpaper electrode over other metallic-based electrodes. With good conducting properties, along with being economical, Al is preferred for the fabrication of GS-TENG. A thin film of Al with a thickness of 50 µm is attached to PET, forming the upper layer. Both the upper and lower layers are mounted on a PET substrate, as PET enhances the flexibility and durability of the device. The Al/PET upper layer and lower layer PTFE/graphite-coated sandpaper/PET are assembled to form GS-TENG, with an area of 45 × 45 mm 2 , as illustrated in Figure 2 and in optical image ( Figure 3 l). The commercial availability, low-cost and adaptability of the materials utilized in the fabrication process of the device provide a route for large-scale industrialization. 3.2. Working Principle of TENG The working mechanism of GS-TENG is shown in Figure 4 . The conductor-dielectric model consists of Al (conductor) as a tribo-positive layer and PTFE (dielectric) as a tribo-negative layer, according to triboelectric series [ 36 ]. When the Al layer is in complete contact with the PTFE, Al and PTFE surfaces are rendered with positive and negative charges respectively, due to triboelectrification, as shown in Figure 4 a. As the layers are separated, potential difference-developed drifts free electrons from the Al electrode to the graphite-coated sandpaper electrode through an external load ( Figure 4 b). Now, when both the layers are separated to the maximum separation distance ( Figure 4 c), electrostatic equilibrium occurs between the electrodes, and hence, no charges flow. When both the layers are contacted again by an external pressing force, charges flow back to the Al layer ( Figure 4 d). Therefore, an alternate current is generated by repeated contact and separation. Figure 4 e(i–iv) shows the comsol simulation results of the potential distribution in GS-TENG for different separation distance of 1 mm, 3 mm, 5 mm and 10 mm. The electric potential on the Al surface reaches 10 5 V when they are separated to 10 mm. The simulation results of the device show that, as the separation distance between the Al and PTFE layer increases, the potential difference also increases and reaches maximum. 3.3. Performance of GS-TENG To investigate the GS-TENGs triboelectric output performance, with an effective contact area of 45 × 45 mm 2 , contact was made periodically and separated under external vibrator excitation at a frequency of 4 Hz and constant amplitude. Initially, we conducted a switching polarity test to know the output signals truly generated from TENG. The ability of electrons to flow from the graphite-coated sandpaper electrode to Al due to the potential difference developed during contact and separation is explained in the working mechanism in detail. By connecting the positive terminal of the probe to the sandpaper-based electrode and the negative terminal to the Al electrode as forward connection mode ( Figure 5 a), the peak value open-circuit voltage and the short-circuit current were measured at ~33.8 V and ~2.16 µA, respectively. Then, the probe connections were switched reversibly to change the polarity as reverse convention mode ( Figure 5 b & c). The electrons flow from Al to the graphite-coated sandpaper electrode. Open-circuit voltage and short-circuit currents of peak values were measured at~36.62 V and ~2.17 µA, respectively. As evidence for the switching polarity, the output performance of the as-fabricated GS-TENG validates the switching polarity test and results were significant in both directions. The amount of charge transfer between electrodes with reference to the positive current peak of 15.78 nC was obtained, as shown in Figure 5 d. The output voltage current and output current of GS-TENG was measured by varying load resistance ranging from 10 to 100 MΩ. The voltage curve increases with the increasing load resistance and the current decreases with the increasing load resistance, as shown in Figure 5 e. The output power density is obtained using the equation, P = I 2 R/A , where, I is the current value at external load resistance R and A is the contact area of the TENG, respectively. Initially, power density was increased and reached the maximum at matched load resistance and further decreased as external load resistance increased, as shown in Figure 5 f. When the external load resistance matches the internal impedance of GS-TENG, the power density reaches maximum value. According to the maximum power transfer theorem [ 39 ], the maximum peak power density obtained is ~0.94 µW·cm −2 at optimal load resistance 30 MΩ, which is sufficient to drive low-power consumption electronics, proliferating the applications range of GS-TENG. In addition, energy conversion efficiency of the GS-TENG with 7.7% was calculated [ 40 ] ( Supplementary Note: 1 ). The conducting stability of the GS-TENG is examined by using the graphite-coated sandpaper electrode with different grit sizes of the sandpaper (400, 600, 800, 1000, 1200, 1500, 2000). Irrespective of grit size of the sand paper, GS-TENG performed consistent output voltage, as shown in Supplementary Figure S1 . 3.4. Application of GS-TENG for Energy Harvesting and Self-Powered Devices The practical applications of as-fabricated GS-TENG are demonstrated in Figure 6 . The output we obtain will be an alternate current. It is converted to direct current output by a rectifying circuit, as shown in Figure 6 a. The as-fabricated GS-TENG can drive more than 120 commercial blue LEDs ( Supplementary Video S1 ) connected in series, as shown in Figure 6 b, and it can also power LCD, as shown in Figure 6 c, through a rectifier circuit ( Supplementary Video S2 ). The output power generated by the as-fabricated TENG can be stored using a rectifier bridge in a capacitor or battery, as shown in Figure 6 d, which can be further used to power some electronic devices. Figure 6 e shows the charging curves of capacitors with various capacitance 1 µF, 2.2 µF, 3.3 µF, 4.7µF, 10 µF and 22 µF, under hand tapping. The charging curve of all the capacitors reveals that it can be charged to 5 V through a rectifier circuit within a short time of 18.64 s, 30.59 s, 44.72 s, 64.28 s, 143.61 s and 393.56 s respectively, by gentle hand tapping. Further, for powering some electronic devices, the stored energy in the capacitor 22 µF with 5 V is utilized to power electronic watch, as shown in Figure 6 f ( Supplementary Video S3 ). The applications demonstrated are derived from hand tapping on GS-TENG. Thus, the biomechanical energy, mainly human-related kinetic energy, can be harvested effectively to generate electricity. Additionally, GS-TENG is flexible and can be integrated easily on most of the body parts. As an active wearable electronic device, GS-TENG can detect the body motions. By harvesting biomechanical energy, movements such as finger tapping, wrist movement, hand tapping and foot stepping, onto the as-fabricated GS-TENG are responsible for generating electrical signals. The potential difference generated under each condition depends on whether the GS-TENG is triggered by finger, wrist, wrist or foot movements. As the finger is tapped ( Figure 7 a), the flexion and extension behaviors from wrist ( Figure 7 d), hand tapping ( Figure 7 g) and foot stamping ( Figure 7 j) periodically results in the electric signals of the corresponding output voltage/current of 26.83 V/14.14 µA, 39.89 V/8.86 µA, 43.84 V/17.52 µA and 55.55 V/4.93 µA respectively, as shown in Figure 7 (b/c, e/f, h/i and k/l). Since the output signals produced during each motion state are different in terms of amplitude, number of peaks, peak width, shape and time interval due to these signals depend on the type of activity which contains more information including frequency, pressure and applied force. By distinguishing the output signals, GS-TENG can be used as a self-powered, active biomechanical sensor. With the ease of fabrication, flexibility, eco-friendly, biocompatibility and integration on the human body, our sandpaper-based device will be remarkable to build smart sensors of the next generation and biomechanical energy harvesters [ 41 , 42 ]."
} | 5,391 |
33552864 | PMC7856889 | pmc | 1,152 | {
"abstract": "Abstract Real‐time detection and differentiation of diverse external stimuli with one tactile senor remains a huge challenge and largely restricts the development of electronic skins. Although different sensors have been described based on piezoresistivity, capacitance, and triboelectricity, and these devices are promising for tactile systems, there are few, if any, piezoelectric sensors to be able to distinguish diverse stimuli in real time. Here, a human skin‐inspired piezoelectric tactile sensor array constructed with a multilayer structure and row+column electrodes is reported. Integrated with a signal processor and a logical algorithm, the tactile sensor array achieves to sense and distinguish the magnitude, positions, and modes of diverse external stimuli, including gentle slipping, touching, and bending, in real time. Besides, the unique design overcomes the crosstalk issues existing in other sensors. Pressure sensing and bending sensing tests show that the proposed tactile sensor array possesses the characteristics of high sensitivity (7.7 mV kPa −1 ), long‐term durability (80 000 cycles), and rapid response time (10 ms) (less than human skin). The tactile sensor array also shows a superior scalability and ease of massive fabrication. Its ability of real‐time detection and differentiation of diverse stimuli for health monitoring, detection of animal movements, and robots is demonstrated.",
"conclusion": "3 Conclusion Sensing and distinguishing diverse external stimuli within one sensor reduces fabrication complexity and alleviates space constraints for electronic skins. The developed skin‐inspired piezoelectric tactile sensor array has the spatiotemporal detection and distinction ability of the magnitude, positions, and modes of diverse external stimuli, including gentle slipping, touching, and bending. The tactile sensor array is designed to be a multilayer architecture inspired by the structure of human skin. Specifically, two PDMS films serve as the protective layers, two PVDF films as the sensory layers, and one PDMS film as the insulative layer. The working mechanism is theoretically investigated via finite element simulation and verified by experiments. The pressure sensing experiment proves its characteristics of zero crosstalk, rapid response time, high sensitivity, and long‐term durability. The response time is 10 ms, much less than that of the human skin 15 ms. The sensitivity of the top sensory and bottom sensory layer is, respectively, 7.7 and 7.2 mV kPa −1 . The sensitivity can be tuned via changing the thickness of the protective and insulative layers. The bending sensing experiment shows that we can calculate the bending direction and bending radius using the tactile sensor array. Using the row+column electrodes, this design can be extended to more pixels using a small quantity of wires ( m + n + 2 for n × m sensing pixels), which enables large‐area scalable fabrication and improves the resolution. The real‐time detection and distinction abilities are demonstrated by monitoring the motions of the neck arterial pulse, the movements of a tiny spider, and the grasping of the robotic hand. In summary, this study reports a skin‐inspired piezoelectric tactile sensor array with real‐time differentiation ability of diverse external stimuli. Our design eases the challenging tasks in spatiotemporally distinguishing diverse stimuli within one sensor. This research provides a new strategy for tactile sensor design and would broadly benefit many fields, especially for electronic skins, health monitoring, animal movement detection, and robotics.",
"introduction": "1 Introduction Human skin is a very amazing sensor that enables simultaneously detecting the intensity and modes of diverse stimuli, including pressing, tapping, slipping, and bending. This ability mainly owes to four mechanoreceptors (SA‐I, II, and FA‐I, II) scattering in different regions of the human skin ( Figure \n 1 a ). [ \n \n 1 \n , \n 2 \n \n ] The mechanoreceptors receive the external stimuli and convert them to electronic signals. The ensemble information from these four receptors is subsequently interpreted by the human brain as body positions, object sizes, shapes, texture, etc. [ \n \n 2 \n , \n 3 \n \n ] The disabled without limbs suffer much from the lack of tactile sensing capability. This issue is relieved by the latest development of smart prosthetics integrated electronic skins, which helps the disabled to regain the human tactile system. [ \n \n 4 \n , \n 5 \n \n ] The flexible artificial electronic skins integrating with tactile sensors have a huge potential in reconstructing the human tactile system, and enable users to operate more naturally by sensing objects and grip forces. [ \n \n 6 \n , \n 7 \n , \n 8 \n \n ] \n Figure 1 Schematic of the skin‐inspired piezoelectric tactile sensor array. a) Illustration of the human skin structure (left) and the skin‐inspired tactile sensor array (right). b) Wiring connection. For an n × m sensing array ( n is the row and m is column), only ( m + n + 2) wires are needed in our design. In contrast n × m + 1 electrode wires are needed in other structures. [ \n \n 19 \n , \n 22 \n , \n 23 \n , \n 26 \n , \n 39 \n \n ] c) Schematic of the real‐time tactile sensor array system. The tactile sensor array attached to human skin can detect diverse external stimuli and transfer them to the signal processor and peripheral. After reading the signal, the peripheral integrated with a logical algorithm will recognize the positions, magnitude, and modes of the stimuli in real time. To monitor and recognize diverse mechanical stimuli for electronic skins, research communities have developed various tactile sensors. These sensors resort to different physical transduction mechanisms including resistivity (i.e., contact resistance), [ \n \n 9 \n , \n 10 \n \n ] piezoresistivity, [ \n \n 11 \n , \n 12 \n , \n 13 \n \n ] capacitance, [ \n \n 14 \n , \n 15 \n , \n 16 \n , \n 17 \n \n ] piezoelectricity, [ \n \n 18 \n , \n 19 \n , \n 20 \n , \n 21 \n \n ] and triboelectricity. [ \n \n 22 \n , \n 23 \n , \n 24 \n \n ] Moreover, efforts have been dedicated to designing microstructures, such as pyramids, [ \n \n 17 \n , \n 25 \n \n ] interlocking, [ \n \n 9 \n , \n 18 \n , \n 26 \n \n ] and hollow‐sphere [ \n \n 11 \n , \n 27 \n \n ] to remarkably improve the sensitivity and broaden detection modes of tactile sensors, including normal force, [ \n \n 13 \n , \n 14 \n , \n 15 \n , \n 16 \n , \n 17 \n , \n 18 \n , \n 19 \n , \n 20 \n , \n 21 \n , \n 22 \n , \n 23 \n , \n 28 \n \n ] shear force, [ \n \n 9 \n , \n 10 \n , \n 14 \n \n ] bending strain, [ \n \n 9 \n , \n 18 \n , \n 28 \n , \n 29 \n \n ] temperature, [ \n \n 30 \n , \n 31 \n \n ] and humidity. [ \n \n 31 \n , \n 32 \n \n ] Besides, larger‐area, [ \n \n 7 \n \n ] high‐resolution, [ \n \n 9 \n \n ] and flexibility [ \n \n 33 \n , \n 34 \n \n ] are also the desirable characteristics of tactile sensors for artificial electronic skin. Despite the notable achievements mentioned above, smart prosthetics integrating with electronic skins are not capable of completely replacing human limbs and human skin due to the single function of the tactile sensors and the disability of distinguishing different stimuli in real time. As stated by Prof. Zhenan Bao, “the sensor field has been shown to be uniform across multiple loading conditions, making it difficult to distinguish between them.” [ \n \n 6 \n , \n 14 \n \n ] Furthermore, the development of larger‐area and high‐resolution electronic devices has been restricted by the required large number of electrode wires. Even though row+column (i.e., n + m ) structures (Figure 1a ) considerably reduce the number of wires in the capacitive sensors, [ \n \n 15 \n \n ] this strategy has not been used in piezoelectric sensors because the wiring connection causes crosstalk. [ \n \n 35 \n \n ] To eliminate the crosstalk issue, every sensing element needs an individual electrode and wire connection. Consequently, an n × m sensor array needs n × m wires, which product number of wires causes many troubles in sensor array fabrication and miniaturization. Here, we propose a new piezoelectric flexible multifunctional tactile sensor array that is able to sense and differentiate the magnitude, positions, and modes of diverse external stimuli in real time (comparison between other tactile sensors and ours can be referred to Table S1, Supporting Information). We develop a simple but effective electrode topology that is crosstalk‐free and only requires n + m wires for an n × m sensor array. Figure 1a illustrates the tactile sensor design inspired by human skin. Anatomy tells that our human skin is a multilayer structure and each layer has its unique function. [ \n \n 36 \n \n ] Specifically, the epidermis forms a protective layer; the dermis serves as cushioning the body from stress and strain and sensing touch and heat. The epidermis and dermis are slightly connected by a layer called the basement membrane. Mimicking human skins, we design the tactile sensor array with two protective layers, two sensory layers, and one insulative layer (Figure 1a ). We expect to sense gentle slip stimuli, touch stimuli, and bending stimuli with high sensitivity and rapid response time. The dual‐layer comb structures of the sensory layers with row+column electrodes eliminate crosstalk and reduce the number of connection wires. The simple electrode wire topology has two advantages: 1) A large number of sensor pixels can be fabricated with a small quantity of wires; 2) individual sensor array can be assembled together with no need of rewiring. We demonstrate the skin‐inspired tactile sensor array for health monitoring, detection of animal movements, and robots.",
"discussion": "2 Results and Discussion 2.1 Design of Human Skin‐Inspired Tactile Sensor Array Inspired by the multilayer structure of human skin, we design the tactile sensor array with five layers, namely, two protective layers, two sensory layers, and one insulative layer (Figure 1a ). The protective layers made of polydimethylsiloxane (PDMS) ( E = 2.6 MPa, thickness 100 µm) mimics the epidermis of human skin to transmit external stimuli and protect the inner structure. The polyvinylidene fluoride (PVDF)‐made sensory layers ( E = 3 GPa, thickness 28 µm), which lie at different depths in the sensor array, are used to receive the stimuli from the protective layers and convert them into electrical signals. This process acts similarly as mechanoreceptors (FA‐I, II and SA‐I, II) in human skin. Analogous to human skin, we name a two‐layer structure, as the top PVDF sensory and the bottom PVDF sensory layer in the tactile sensor to represent their positions in the structure. The PDMS insulative layer (thickness 500 µm) serves in two aspects. First, it avoids the crosstalk of two PVDF layers; second, the thick middle layer guarantees the stress neutral layer of the device far away from both PVDF layers, thus enabling each PVDF layer to be only tensile or compressive under bending stimuli and enhancing the output voltage. PDMS is selected since it is a toxic‐free, biocompatible, and low‐cost material [ \n \n 37 \n \n ] and it mimics the skin tissues and dermis. PVDF has advantages such as ultratransparency, high flexibility, and lead‐free biocompatibility. [ \n \n 38 \n \n ] \n In addition to the bionic multilayer structure in our design, the layout of the PVDF layers is the key to recognize diverse stimuli modes, avoid crosstalk interferences, and reduce the number of wires. In Figure 1a , we demonstrate a 3 × 3 tactile sensor array. Each sensory layer is designed into a comb pattern. For each sensory PVDF layer, there are two electrodes attached to both sides, respectively. The electrode of one side includes three separate parts and acts as three detection channels, while that on the other side is made united and acts as a ground electrode. Each channel connects to one wire. Thus, for each channel, it avoids the crosstalk since other channels do not share the same electrodes with it. The channels of the top sensory and bottom sensory layers are arranged to be perpendicular (i.e., row+column structure). In the following, we will show this arrangement successfully achieves the differentiation of diverse external stimuli. The comb design, on the one hand, avoids the local deformation of the three channels and transfers the stress more uniformly. On the other hand, it allows the electrodes for the channels on one side to connect with each other and reduce the number of wires. As illustrated in Figure 1b , a 5 × 5 array, for example, for n × m sensing units ( n is the row and m is column), only ( m + n + 2) wires are needed in our design. In contrast, n × m + 1 electrode wires are needed in other structures. [ \n \n 19 \n , \n 22 \n , \n 23 \n , \n 26 \n , \n 39 \n \n ] Another characteristic of the proposed design is its self‐powered ability due to the piezoelectric transduction mechanism that receives much interest in the large‐area and high‐resolution electronic devices. [ \n \n 7 \n \n ] \n Furthermore, we develop a real‐time tactile system based on the tactile sensor array. As shown in Figure 1c , this system mimics the biological response process of the human being. [ \n \n 12 \n \n ] First, the PVDF films in the sensor array receive the external stimuli from the protective layers. Then, the stimuli will be converted to the electrical signal via the direct piezoelectric effect. Third, the signal is transferred to the peripheral, such as a computer, after signal processing. By reading the signal, the peripheral with the configuration of a logical algorithm will recognize the positions, magnitude, and modes of the stimuli in real time. 2.2 Working Mechanism To theoretically analyze the working mechanism of the tactile sensor array, we conduct finite element (FE) simulations. The sensor array undergoes two stimuli modes under different types of loads. In the first stimuli mode, as shown in Figure \n 2 a , the pressure is applied to a sensing pixel. Here we select the middle pixel as a representative. Both layers are thus in the compressive stress state. Figure 2b shows the stress distribution of the top sensory and bottom sensory PVDF layers. It can be observed that the maximum stress occurs at the pressure‐bearing pixel in each layer. The stress of the top sensory layer is higher than that of the bottom sensory layer. With the consideration of the stress state, the piezoelectric constitutive relations, the direction of the polarization, and circuit connection ( Figure \n 3 a ), we can predict that the output voltage of the two layers keeps opposite and the signal of the top sensory layer is higher. In the bending stimuli mode, as we mentioned above, the neutral layer lies in the PDMS insulative substrate, thus leading to the opposite stress state for the PVDF layers. Besides, under bending deformation, the output voltage generated by the PVDF layers is simultaneously determined by the bending direction and bending radius. Figure 2c defines the angle of the bending direction α and bending radius R . From Figure 2d – g , we interpret that the stress of the top sensory layer decreases with a large bending angle, while the bottom sensory layer shows an opposite trend. The stress of both layers is close to each other at 45° bending angle. The reason is that the top sensory layer bends effectively at 0°, while the most effective bending direction is 90° for the bottom sensory layer. At 45°, the structure and its applied loads are approximately symmetric, thus the stress distribution of both layers is close. As expected, a larger bending radius leads to smaller stress. Considering the stress state, the piezoelectric constitutive relations, the direction of the polarization, and circuit connection (Figure 3a ), the output voltage of the two layers in the bending stimuli mode will keep pace with each other and the voltage signal depends on both the bending angle and bending radius. We also investigate the effect of the thickness and Young's modulus of the protective and insulative layers on the stress field. Figure S1 (Supporting Information) shows the stress distribution of both top and bottom sensory layers when the thickness of two protective layers changes with the unchanged insulative layer. Under the same bending direction and bending radius, the stress with thicker protective layers is slightly higher. The stress increase comes from the larger stiffness of the device owing to the thicker protective layers. The bigger Young's modulus of protective and insulative layers will also cause the stiffer structure. Thus, the stress increases in the top and bottom sensory layers with stiffer protective layers, as illustrated in Figure S2 (Supporting Information). With the consideration of the direct piezoelectric effect, we can infer that the output voltage in these two sensory layers shows the same trend when the thickness and Young's modulus of the protective and insulative layers change. Figure 2 Working mechanism of the tactile sensor array under different stimuli modes. The finite element simulation describes the stress distribution of the sensor array under different types of external stimuli. a) Schematic of the tactile sensor array under pressure stimulus mode. b) The stress analysis shows that the signal mainly occurs at the pressure zone and the signal of the top PVDF sensory layer is larger than that of the bottom PVDF sensory layer. c) Schematic of the tactile sensor array under bending stimulus mode. d–g) The stress of the surface layer decrease with a larger bending angle, while the case of the deep layer shows the opposite trend. The stress of both layers is close to each other at 45° of bending angle. Besides, a larger bending radius leads to larger stress. Figure 3 Experimental results of the tactile sensor array under the pressure stimuli mode. The measured signal is generated by the middle sensing unit. a) The direction of the polarization of the piezoelectric sensory films and the circuit diagram of the tactile sensor array. b) The real‐time waveform of the output voltage under gentle slip stimuli. c) The real‐time waveform of the output voltage under different normal force amplitudes with a frequency of 5 Hz. d) Output peak voltage with error bar versus normal force with the amplitude from 80 to 750 kPa. The sensor array shows a high linearity, low error, and high sensitivity. e) The durability test. The top PVDF sensory responses stably after 80 000 cycles under a normal force of an amplitude of 15 N and frequency of 30 Hz. 2.3 Pressure Sensing Our skin‐inspired tactile sensor array can measure both the magnitude and positions of pressure stimuli. Figure 3 shows the direction of polarization and circuit diagram of the sensor array, and the electrical responses caused by pressure stimuli. The bottom silver electrode of the top sensory layer and the top silver electrode of the bottom sensory layer are grounded. As mentioned in Section 2.1 , the top electrode of the top sensory layer and the bottom electrode of the bottom sensory, respectively, contain three separate channels. When the pressure is applied to a pixel, one of the channels in both the top sensory layer and the bottom sensory layer will respond. The cross of two channels in the two layers represents the position applied to pressure stimulus. Excited by gentle slip stimuli (Figure 3b ), responses of both PVDF layers at this pixel are small, but the output voltage of the top PVDF sensory layer is much larger than that of the bottom sensory layer (0.1 V vs 0.02 V), attributed to the thick middle insulative PDMS layer that absorbs much energy. In this experiment, a steel bar used to stimulate the surface of the tactile sensor array is connected to the ground electrode, which eliminates the triboelectric effect on piezoelectric signal output. [ \n \n 40 \n \n ] For touch stimuli, we test the sensor with different pressure amplitudes. As shown in Figure 3b , a large pressure corresponds to the large output voltage and the real‐time waveform of the output voltage clearly shows the opposite phase of the two layers. This phenomenon is expected and has been explained previously in the FE simulations. Besides, it can be observed, again, that the signal of the top PVDF sensory layer is larger due to the existence of the insulative layer. Figure 3c shows the waveform of the output voltage under different normal force amplitudes with a frequency of 5 Hz. Figure 3d presents the relations of the peak output voltage to different amplitudes of pressure with a frequency of 5 Hz for the top and bottom sensory PVDF layers. The sensor array shows high linearity, low error and high sensitivity (7.7 and 7.2 mV kPa −1 for the top sensory layer and the bottom sensory layer, respectively). The difference in the sensitivity for both sensory layers is attributed to the insulative layer since the top sensory layer absorbs more deformation from external stimuli. The response time of the sensor array is fast (Figure S3, Supporting Information). Excited by a load of 15 N and 5 Hz, the response time is 10 ms, which has surpassed that of the human skin (about 15 ms). [ \n \n 6 \n \n ] Our sensor array also shows long‐term durability and stability. It outputs stable responses after 80 000 cycles under 15 N force (Figure 3e ). After 80000 cycles, the output voltage begins to gradually decrease (Figure S4a, Supporting Information). The SEM images of the silver electrode surface before and after long‐term durability test are shown in Figure S4b–d (Supporting Information). Before the durability test, the surface of the silver electrode is flat and unit (Figure S4b, Supporting Information). Since the top sensory layer is close to the external stimuli and the protective PDMS layer is thin, the silver electrode surface of the top sensory layer suffers damage after the 120000 cycles (Figure S4c, Supporting Information), while the that of the bottom sensory layer is still intact due to the thick insulative PDMS layer (Figure S4d, Supporting Information). 2.4 Bending Sensing The task of bending sensing is to simultaneously recognize the bending direction α and bending radius R (Figure 2c ). Here, we first experimentally obtain the electrical signal under different bending conditions and subsequently establish their relations. Figure \n 4 a – c shows the measured output voltage waveforms of the sensor array bent at 0°, 45°, and 90° with a bending radius of 15 mm. It indicates that a large bending angle lowers the output voltage of the top sensory layer but raise that of the bottom sensory layer, and the output voltage waveforms of the top sensory layer and bottom sensory layer show the same phase. The measured result is similar to the analysis in FE simulations. The trend in Figure 4d – f represents that larger output voltage occurs at a smaller bending radius. Another observation is that the top sensory layer has the highest sensitivity at 0° bending direction and 90° for the bottom sensory layer. Figure 4 Experimental results of the tactile sensor array under the bending stimuli mode. a–c) The real‐time waveform of the output voltage bending at 0°, 45°, and 90° with the bending radius of 15 mm. d–f) The output peak voltage versus bending radius. To infer the bending direction α and bending radius R , we further build the relations of the output voltage to bending direction and bending radius using the experimental results (Tables S2 and S3, Supporting Information). The fitting function is shown in Table S4 (Supporting Information) and the fitting result of the top sensory layer and bottom sensory layer are shown in Figure \n 5 a , b . As shown in Figure 5c , d , the sensor array is held by the human palm. In the first demonstration (Figure 5c ), the thumb of a tester moves and we get two voltage contours of the top sensory layer ( V \n t ) and the bottom sensory ( V \n b ). The contours have an intersection, which records the bending direction and bending radius. In this case of Figure 5c , they are 42° and 14 mm, respectively. In the second demo of Figure 5d , the tester moves his other four fingers. We interpret the bending direction and bending radius as 4° and 22 mm. Figure 5 Sensing the bending direction and bending radius. It shows that the tactile sensor array is capable of simultaneously sensing the bending direction and bending radius. a,b) 3D fitting results based on experimental data. c,d) Demonstration of the tactile sensor array to detect the hand deformation via contour lines. Based on the output voltage of the top sensory and bottom sensory layers, two contours are plotted in one plane. The bending direction and bending radius are indicated by the intersection of the contours. 2.5 Multipixel Detection and Large‐Area Scalable Design To further improve the resolution of a tactile sensor array, a large‐area design is desirable but challenging for existing sensors [ \n \n 19 \n , \n 22 \n , \n 23 \n , \n 39 \n \n ] because the increase in sensing pixels usually requires a large number of wires. In contrast, our design can be easily extended to the large‐area sensor array. It does not need numerous wires as well as rewiring. To exemplify this property, a 5 × 5 sensor array is fabricated, as shown in Figure \n 6 a , b . The number of wires used in this sensor array is 12 (compared to 25 for 5 × 5 sensor array). Pressure sensing tests are performed. As illustrated in Figure 6c , d , the location of the pressure stimuli can be easily judged in the cross of the maximum voltage signals of the top sensory layer and bottom sensory layer. Figure 6 Illustration of a large‐area scalable tactile sensor array and its ability for multipixel detection. The tactile sensor array is suitable for large‐area scalable fabrication and bulk production. a) Schematic of the 5 × 5 bionic tactile sensor array. b) Photograph of the fabricated 5 × 5 tactile sensor array with each unit of 2.5 × 2.5 mm 2 . c) Illustration of pressure stimuli and the corresponding real‐time signal from the sensor array. The sensor array accurately captures the pressure position. d,e) Detection of multiple touchpoints. The profiles of planar pressure intensity show that the sensor array accurately captures multitouch excitations. In addition to the single‐pixel stimuli illustrated in pressure and bending sensing experiments, the tactile sensor array also can identify multipixel stimuli. The working mechanism of multipixel pressure sensing is introduced as follows. First, the logical algorithm scans every channel and decides whether the channel is activated by the signal of the top sensory and bottom sensory layers. Considering the fabrication, the threshold is set to 0.4 V. The signal higher than the threshold is considered as activation. If both the top sensory and bottom sensory are activated, the voltage of this intersection is calculated by V = ( V \n t + V \n b )/2, otherwise, the voltage is represented by V = min ( V \n t , V \n b ), where V represents the external stimuli, V \n t and V \n b are the output voltage of the top sensory and bottom sensory PVDF layers, respectively. Figure 6d , e shows two examples of multitouch stimuli. In Figure 6d , a wooden stick presses the third column of the sensor array. In Figure 6e , the pressure to the top right corner of the sensor array comes from a fingertip. We can observe the magnitude and the locations of the stimuli. The real‐time waveforms of the two cases are shown in Figure S5 (Supporting Information), from which we can see the crosstalk‐free properties of the tactile sensor array. 2.6 Real‐Time Differentiation of Diverse External Stimuli Compared to the existing sensors in the literature, the developed tactile system can detect and differentiate the magnitude, locations, and modes of external stimuli in real time. A comprehensive comparison is tabulated in Table S1 (Supporting Information). The characteristics of our skin‐inspired tactile sensor are summarized as follows: i) Under the gentle slip stimuli mode, the output voltage of the bottom sensory layer is close to zero; ii) under the touch stimuli mode, the output voltage waveforms for the top sensory and bottom sensory layers show 180° phase shift; iii) under the bending mode, the output voltage waveforms for the top sensory and bottom sensory layers are in the same phase. To achieve real‐time differentiation, a logical circuit is developed in the software LabVIEW 2017. Figure S6 (Supporting Information) shows the logical flow diagram of the algorithm. All pixels are scanned by the logical algorithm. The diagram of the logical circuit based on the algorithm is shown in Figure S7 (Supporting Information). For each pixel, first, the logical circuit simultaneously decides the locations and modes of the external stimuli. When the electrical signal of the top sensory layer is higher than a threshold, which is set to 0.05 V in this study, the channel is considered in an activated state. Then, the circuit will determine whether the stimuli is a gentle slip by reading the data of the bottom sensory layer. If this signal is less than another threshold (i.e., 0.02 V), the logical circuit will consider the stimuli as a gentle slip. Otherwise, the circuit will continue to compare the sign of the signal generated by two layers. The same sign means touch stimuli mode, while the opposite sign corresponds to bending stimuli mode. After the type of external stimuli is identified, we can obtain the pressure value of every pixel using the results in Figure 3d or the bending direction and bending radius via the contours of the fitting polynomial in Figure 5 . A real‐time differentiation of gentle slip, touch, and bending stimuli is shown in Video S1 (Supporting Information). 2.7 Applications The tactile sensor array has various applications from health monitoring, detection of animal movements to robots. In Figure \n 7 a , the sensor array is used to monitor the subtle physiological signals for a healthy purpose. The sensor array is attached to the bare skin of the neck and accurately detects the weak artery pulse, of which three responses the incident wave P 1 , the tidal wave P 2 , and the diastolic wave P 3 are recorded. [ \n \n 41 \n \n ] \n Figure 7 Applications of the tactile sensor array. a) Real‐time detection of the neck arterial pulse. The peaks P 1 , P 2 , and P 3 are, respectively, the incident, tidal, and diastolic wave. b) Real‐time detection of animal motions. A spider that weighs about 5 mg is used. The voltage signal indicates the take‐off and landing positions of the spider, resting time, and duration of passage, and shows the high sensitivity of the sensor array. c) Demonstration of grasping objects using a robotic hand with a feedback module. The tactile sensor is mounted on the robotic hand and used as real‐time feedback. A piece of soft and fragile tofu keeps integrity during the movement of the robotic hand. The sensor array detects the whole of the grasping process. d) Demonstration of grasping objects using a robotic hand without the proposed sensor array. A piece of soft and fragile tofu damaged during the movement of the robotic hand. The sensor array can be used to detect the movement of small animals. For demonstration, we use a 5 mg weight spider to show the high sensitivity of our device (Figure 7b and Video S2, Supporting Information). To enhance the sensitivity to measure the extremely small pressure from the tiny spider movement, we remove all layers of the tactile sensor except the top sensory layer. Initially, the spider keeps still in the middle of the Line 1 and Line 2. It then takes off toward the wall of a plastic cup and lands on Line 1 to rest. The sensor array accurately captures all behaviors of the spider, including landing positions, resting time, and the duration of the passage. Besides, the signal of the three channels shows no interference with each other. One should note, here, that changing the thickness of the protective and insulative layers is an effective strategy to achieve the adjustment of the sensitivity for sensors. This method is also used in nature. For example, the difference in thickness between human fingertips and palms satisfies diverse sensing. The optimization of the thickness enables the sensor array to broaden the application scenarios. Finally, we perform a robotic hand experiment as an industrial application. In Figure 7c , the tactile sensor array is mounted on the robotic hand. With the aid of the sensor feedback, the robotic hand correctly interrupts the grasping process and keeps holding when the feedback signal exceeds the 4 V default threshold followed by releasing (Video S3, Supporting Information). A comparative trial without the sensor array as feedback is conducted in Figure 7d . The soft and fragile tofu is damaged without the tactile sensor (Video S4, Supporting Information)."
} | 8,248 |
36053127 | PMC9826171 | pmc | 1,153 | {
"abstract": "SUMMARY Characterizing photosynthetic productivity is necessary to understand the ecological contributions and biotechnology potential of plants, algae, and cyanobacteria. Light capture efficiency and photophysiology have long been characterized by measurements of chlorophyll fluorescence dynamics. However, these investigations typically do not consider the metabolic network downstream of light harvesting. By contrast, genome‐scale metabolic models capture species‐specific metabolic capabilities but have yet to incorporate the rapid regulation of the light harvesting apparatus. Here, we combine chlorophyll fluorescence parameters defining photosynthetic and non‐photosynthetic yield of absorbed light energy with a metabolic model of the pennate diatom Phaeodactylum tricornutum. This integration increases the model predictive accuracy regarding growth rate, intracellular oxygen production and consumption, and metabolic pathway usage. Through the quantification of excess electron transport, we uncover the sequential activation of non‐radiative energy dissipation processes, cross‐compartment electron shuttling, and non‐photochemical quenching as the rapid photoacclimation strategy in P. tricornutum. Interestingly, the photon absorption thresholds that trigger the transition between these mechanisms were consistent at low and high incident photon fluxes. We use this understanding to explore engineering strategies for rerouting cellular resources and excess light energy towards bioproducts in silico . Overall, we present a methodology for incorporating a common, informative data type into computational models of light‐driven metabolism and show its utilization within the design–build–test–learn cycle for engineering of photosynthetic organisms.",
"conclusion": "CONCLUSIONS In the present study, we combined chlorophyll fluorescence parameters defining photosynthetic and non‐photosynthetic yield of absorbed light energy with a metabolic model of P. tricornutum . This integration increased the model predictive accuracy regarding growth rate, intracellular oxygen production and consumption, and metabolic pathway usage. Through the quantification of EET, we uncovered the sequential activation of non‐radiative energy dissipation processes, cross‐compartment electron shuttling, and non‐photochemical quenching as the rapid photoacclimation strategy in P. tricornutum . The photon absorption thresholds that triggered the transition between these mechanisms were consistent at low and high incident photon fluxes, providing insights into the mechanisms that drive the ecological success of this important class of primary producers. Quantification of EET allowed us to assess engineering strategies for rerouting cellular resources and excess light energy towards bioproducts. Taken together, our results show integrating relevant measurements of photosynthetic physiology with genome‐scale models results in quantitative predictions of condition‐specific phenotypes. This paves the way for iterative design and real‐time process control of photobioproduction platforms.",
"introduction": "INTRODUCTION There is great interest in characterizing light‐driven metabolism as a result of the ecological importance and engineering potential of phototrophic microorganisms and plants. Oxygenic photosynthesis utilizes light energy to generate an oxidized protein complex capable of extracting electrons from water at photosystem II (PSII), at the same time as re‐energizing the extracted electron to reduce NADP + at photosystem I (PSI). These ‘light harvesting’ reactions drive electron transport, ATP generation, and subsequent CO 2 fixation through the Calvin–Benson–Bassham cycle (CBBC) in addition to the other energy consuming reactions throughout the cell. Light absorption by a photosynthetic cell is not constant. Because light fluxes can vary across the day and as a result of ecological or climatological features, photosynthetic microorganisms often absorb more photons than can be utilized by metabolism during these natural fluctuations. If this excess energy is not dissipated, over‐reduction of the photosynthetic electron transport chain (ETC) occurs. The resulting formation of reactive oxygen species causes damage to proteins, lipids, and nucleic acids (Dietz et al., 2016 ; Niyogi, 2000 ). Photoinhibition is a product of this damage, and it decreases photosynthetic efficiency because of damage from excess light capture. The PSII D1 subunit is the primary photoinhibition target in the photosynthetic ETC (Edelman & Mattoo, 2008 ). A complex repair cycle characterized by removal, degradation, and de novo synthesis is constitutively active to counter this damage and it is energetically expensive (Nixon et al., 2005 ). To prevent photoinhibition, excess energy can be dissipated upstream of the photosynthetic ETC complexes via a variety of mechanisms encompassing non‐photochemical quenching (NPQ), which harmlessly converts excitation energy to heat (Nicol et al., 2019 ). NPQ not only protects the photosynthetic system from oxidative stress, but also reduces the apparent efficiency of light‐biomass conversion as a smaller fraction of captured light energy enters the broader metabolic network. Here we coin the term excess electron transport (EET) as an additional important physiological feature at the intersection of photophysiology and bioengineering. This comprises several components that act either as shunts within the ETC (Jallet, Cantrell, & Peers, 2016 ; Ware et al., 2020 ) or downstream of the photosynthetic machinery within the broader metabolic network. It relieves over‐reduction of the photosynthetic ETC by dispelling electrons generated by excess light (Jallet, Cantrell, & Peers, 2016 ). Usually, these reactions are considered metabolically ‘futile’ as the electrons are deposited on elemental oxygen to generate water, for example. However, they are important in relieving photosynthetic ETC over‐reduction. There is interest in the bioengineering field to redirect these electrons away from metabolic futility towards bioproducts of interest, at the same time as maintaining the beneficial effects on ETC redox balance (Lassen et al., 2014 ; Levering et al., 2015 ). Harnessing excess reductant can convert endogenous carbon sinks, such as carbohydrates, into more energy dense products such as lipids. Additionally, recently, it was shown that engineered reductant sinks can actually increase carbon fixation and overall photosynthetic efficiency (Santos‐Merino et al., 2021 ). Thus, downregulating evolutionarily beneficial processes for photosynthetic individuals in favor of mass culture productivities offers promising avenues for increasing bioproduct and biofuel efficiency (Peers, 2014 ). Quantitative characterization of the push–pull of light capture upstream and dissipation in the metabolic network downstream of the photosynthetic ETC would enable design and optimization of these engineered reductant sinks. Properly accounting for EET facilitates this bioprocess optimization and provides insight into photoprotection strategies. Previous work in photosynthetic microorganisms (diatoms and green algae) has used photophysiology parameters derived from chlorophyll fluorescence measurements to estimate EET (Wagner et al., 2006 ). In this previous framework, EET was calculated as the difference between the total absorbed photons and the excitation energy required for biomass production and cellular maintenance. Additionally, the fraction of total absorbed light energy lost upstream of the photosynthetic ETC was estimated using chlorophyll fluorescence data, which have long been employed to assess phototrophic physiology (Krause & Weis, 1991 ). Chlorophyll fluorescence primarily quantifies the fate of absorbed light energy directed to PSII; however, there is evidence of contributions from PSI as well (Giovagnetti et al., 2015 ; Pfündel et al., 2013 ). This excitation energy has three primary fates: it can perform photochemistry at PSII, it can be dissipated as heat through NPQ processes, or it can be dissipated by other, less well characterized non‐radiative and fluorescence processes (NO). All of these can be quantified through the use of pulse amplitude modulation (PAM) chlorophyll fluorimetry (Kramer et al., 2004 ). When these values are normalized to the total excitation energy routed to PSII, they are annotated as the quantum yields Y (II), Y (NPQ), and Y (NO), respectively, the sum of which is always one. These techniques have unveiled the diverse photoprotective strategies employed by photosynthetic microorganisms to include extensive NPQ in the diatom Phaeodactylum tricornutum (Lavaud et al., 2002 ). However, these important aspects of photosynthesis have not been integrated into models of total cellular metabolism. Constraint‐based modeling coupled with flux balance analysis (FBA) has successfully been employed to characterize and engineer a wide range of biological systems (Bordbar et al., 2014 ; Küken & Nikoloski, 2019 ). Constraint‐based modeling relies on a reconstruction of the metabolic content of the organism of interest, which, when performed at the whole‐organism level, results in a genome‐scale knowledge base. Leveraging this genome‐scale reconstruction to compute cellular phenotypes via FBA results in a genome‐scale model (GEM). There have been several advances in the metabolic modeling of photosynthetic organisms to include plants (de Oliveira Dal'Molin et al., 2010 ; Poolman et al., 2013 ), cyanobacteria (Broddrick et al., 2016 ; Broddrick, Welkie, et al., 2019 ), green algae (Chang et al., 2011 ; Zuñiga et al., 2017 ), and diatoms (Levering et al., 2016 ). Typically, most models integrate light harvesting into photosynthetic GEMs by assuming a linear relationship between light absorption and photophosphorylation (Shastri & Morgan, 2005 ), However, recent modeling in the diatom P. tricornutum quantified growth rates, excitation energy partitioning between the photosystems, and cross‐compartment energetic coupling of the chloroplast and mitochondrion (Broddrick, Du, et al., 2019 ). However, that study still used simplified assumptions regarding light harvesting, possibly affecting the accuracy of absolute fluxes predicted by the model. The metabolic network that underpins GEMs is assembled from the reactant and product stoichiometry of biochemical reactions; thus, it should be feasible to couple the representation of chlorophyll fluorescence parameters as a fraction of light energy routed to PSII as a stoichiometrically balanced biochemical equation. Such a framework would enable the explicit integration of chlorophyll fluorescence data and EET as a constraint on photosynthetic metabolic processes towards an increased understanding of photoacclimation, photoprotection, and bioengineering of phototrophic metabolism.",
"discussion": "DISCUSSION Cell and photophysiology In the present study, we characterized photoautotrophic metabolism in P. tricornutum through integrated chlorophyll fluorescence measurements and genome‐scale modeling. Our observations in P. tricornutum were consistent with photophysiology under fluctuating and sinusoidal light (Wagner et al., 2006 ) and photoacclimation (Nymark et al., 2009 ). P. tricornutum exhibited efficient photoacclimation with the quanta absorbed per pigment remaining consistent between LL and HL (Figure 1b ). This efficiency was also observed when looking at the initial slope of the cell‐normalized P \n O versus QF curves (Figure 1c ) and the chlorophyll‐normalized P \n O versus PAR curves (Figure S1 ), which were consistent between both LL and HL cultures. This efficiency across a range of photoacclimation conditions likely contributes to the ecological success of diatoms in dynamic light environments (Behrenfeld et al., 2021 ). Our interpretation of photophysiology was heavily influenced by analyzing the P \n o and PAM data with QF as the independent variable. The 1 – qL versus QF curves (Figure 1d ), the shape of the chlorophyll fluorescence parameters versus QF curves (Figure 1e,f ), and the D1 content as a fraction of total protein (Table S2 ) were consistent between LL and HL. Contrasting PAM versus QF (Figure 1e,f ) with PAM versus PAR (Figure S6a,b ) illustrates how interpreting photophysiology from a QF perspective affects conclusions about photophysiology. We observed very little dissipation of excitation energy via NPQ at the experimental QF values (Figure 1e,f ). For the HL conditions, this lack of NPQ was coupled with an effective quantum yield of PSII [ Y (II)] value of 0.32 (Table 1 ), suggesting the presence of alternative dissipative mechanisms [ Y (NO)]. NPQ is an important excitation energy dissipation mechanism under dynamic light conditions (Lavaud et al., 2002 ; Olaizola et al., 1994 ; Wagner et al., 2006 ); however, our results suggest these other dissipative mechanisms are sufficient to prevent photoinhibition under stable light environments. Overall, these data suggest P. tricornutum employs a photoacclimation strategy emphasizing rapid utilization and dissipation of light energy. At times, the overall photosynthetic apparatus is under‐utilized (e.g. LL acclimated cultures). However, consistency in total PSII content per cell and the redox state of the plastoquinone pool versus quantum flux (Figure S2a,b ) suggest that P. tricornutum can immediately respond to an increase in available photon flux without the need to biosynthesize additional macromolecules. Modeling photoautotrophy and the hierarchy of constraints Translating the QF to a photon uptake constraint and P \n O into an oxygen evolution constraint in the GEM resulted in accurate predictions of photoautotrophic growth (Figure 3 ). Growth rate under the HL condition was underestimated by 12%, likely as a result of the beginning of carbon limitation in the sample during short‐term measurements of photosynthetic capacity (Figure S3a ). We chose not to spike in exogenous bicarbonate for our oxygen evolution measurements as we were interested in measuring photosynthetic parameters relevant to our culturing conditions. However, our HL growth rate underestimation suggests the rapid light curve protocol did not completely recapitulate the experimental culture carbon environment. At the same time, including several mM NaHCO 3 \n − in these assays to avoid carbon limitation, as is standard protocol, likely overestimates the available inorganic carbon for photosynthesis in an air sparged experimental culture. Future modeling efforts need to address these factors to ensure accurate simulation parameters. From our efforts to establish the hierarchy of constraints, it was clear oxygen evolution ( P \n O ) is the dominant constraint on the system, resulting in accurate growth rate predictions as the sole constraint. Water splitting at PSII is the only reaction in the model with net oxygen production and is stoichiometrically coupled to photosynthetically derived electrons. All oxygen consuming reactions in the model are linked to electron consumption either directly (e.g. oxidoreductases, epoxidases, fatty acid desaturases, etc.) or indirectly (e.g. photorespiration of 2‐phosphoglycolate, hydrogen peroxide detoxification, etc.). Thus, net cellular oxygen evolution is proportional to the amount of photosynthetically derived reductant that can form biomass macromolecular components. Additionally, the difference between gross oxygen production at PSII and net cellular oxygen evolution is proportional to EET, which was a primary motivator for refining PSII flux via inclusion of chlorophyll fluorescence parameters. Still, it should be noted that P \n O alone did not result in accurate predictions of intracellular oxygen dynamics, as seen in the modeling analysis of the MIMS data. The D1 damage constraint did not affect the predicted growth rate. This is likely a result of the excess EET in the system, which could be rerouted to nucleotide triphosphate (ATP and GTP) biosynthesis, to include mitochondrial ATP synthesis, if necessary, to satisfy the D1 degradation and biosynthetic costs. The NTP maintenance cost (GTP and ATP) was calculated to be 0.36 ± 0.07 and 1 ± 0.2 × 10 −2 mmol NTP g DW −1 h −1 at high light and low light, respectively (Table S2 ). This is likely a conservative value because it assumes 100% of the amino acids from the damaged D1 protein can be reused after proteolysis. Still, we had to increase the D1 damage constraint at high light to > 10.0 mmol NTP g DW −1 h −1 and >2.1 mmol NTP g DW −1 h −1 at low light before EET was exhausted and a decrease in the predicted growth rate was observed. However, the D1 damage constraint requires the model to deliver GTP to the plastid, which affects predictions regarding intracellular metabolic pathway activation. Thus, although this constraint may not affect predictions of growth rate, it likely increases the accuracy of metabolic reaction flux predictions. An important finding from the hierarchy was that photon uptake and chlorophyll fluorescence constraints alone accurately predict growth rate at low acclimation irradiances (Table 2 ). It is possible that, when the maximum photosynthetic rate is light limited, as is likely with our low light acclimated samples, accurately constraining the total photon input into the system results in a similar system‐level constraint as the oxygen evolution constraint. However, when the maximum photosynthetic rate is limited by metabolic reaction kinetics, such as the carbon fixation rate, photon uptake and chlorophyll fluorescence constraints no longer mirror the P \n O constraint. This may explain why this set of constraints overestimated the growth rate of our high light acclimated samples. This is a testable hypothesis and should be explored in future modeling efforts. Still, accurate predictions of biomass accumulation under light limited conditions opens the possibility for non‐invasive monitoring of culture health and productivity for biotechnology applications. A passive sampling window with integrated chlorophyll fluorescence and spectral absorption analysis, coupled with data on surface irradiance, would recreate the chlorophyll fluorescence and photon uptake constraints. Because bioproduction conditions are usually high density, resulting in low cell‐specific quantum flux, integrating non‐invasive sampling with our modeling construct could result in accurate measurements of photoautotrophic metabolism in these settings. Chlorophyll fluorescence constraints increased the accuracy of intracellular metabolic processes Constraining biomass accumulation with QF and P \n O automatically predicted relevant photosynthetic parameters in a manner similar to previous investigations (Jakob et al., 2007 ; Wagner et al., 2006 ). Importantly, GEMs also predict the optimal distribution of excitation energy between PSI and PSII, an advantage compared to previous work where it was assumed that 50% of the absorbed quanta were directed to each photosystem. Our simulations predicted a two‐fold increase in excitation energy utilized by PSII under HL compared to LL, with approximately 76% of absorbed photons directed to PSII. However, there was a similar number of charge separation events at both photosystems (Table S3 ). This is consistent with the observation that there is minimal cyclic electron flow (CEF) around PSI in P. tricornutum (Bailleul et al., 2015 ), which would result in roughly equivalent charge separations at both photosystems. Additionally, although we constrained the upper bound of CEF flux based on experimental values (Bailleul et al., 2015 ), the simulation results suggested this constraint was redundant. CEF flux at both HL and LL (0.55 and 0.67 mmol e – g DW −1 h −1 for HL and LL, respectively) were below the maximum CEF rates (1.4 and 2.2 mmol e – g DW −1 h −1 for HL and LL, respectively). It is likely that the bifurcation of excitation energy between PSII and PSI, and, thus, CEF, is captured by the QF , P \n O, and chlorophyll fluorescence constraints. Our model derived ETR differs from methodologies that assume equal excitation energy routed to both photosystems and a lower predicted quantum demand at experimental QF values (Table S3 ). For an organism such as P. tricornutum that does not employ extensive CEF (Bailleul et al., 2015 ), the advantage of this approach is diminished. However, for microalgae, cyanobacteria, or plants that dynamically reroute excitation energy between the photosystems and adjust their biomass macromolecular composition as part of their photoacclimation strategy [e.g. Chlamydomonas reinhardtii (Davis et al., 2013 ; Lucker & Kramer, 2013 ; Walker et al., 2020 )], implicit calculation of excitation routing between photosystems using our modeling framework will result in better approximations of ETR. Incorporating chlorophyll fluorescence as a GEM constraint increased the accuracy of model predictions. Capturing photon loss upstream of the photosystems recapitulated intracellular oxygen production and reductant‐mediated oxygen consumption rates. This is likely because significant excitation energy is lost as Y (NO), especially at high light. Without the chlorophyll fluorescence constraint, the model would route that excitation energy to photochemistry, increasing LET and requiring more intracellular electron and oxygen consumption via EET to not violate the oxygen evolution constraint. As a result, these new constraints resulted in accurate predictions of intracellular oxygen consumption (validated by MIMS), and affected predictions of excess reductant in the system (EET), as well as predictions regarding cross‐compartment metabolic coupling (Figure 4a,b ). The plateau in EET flux as a function of QF (Figure 4a ) was correlated with NPQ activation (Figure 1f ), suggesting that saturation of EET pathways triggers this photoprotective mechanism. Our results suggest the following model of photophysiology in P. tricornutum in a stable light environment: up to a QF of approximately 0.3 fmol photons cell −1 sec −1 , Y (NO) dissipates excess excitation energy until the steady‐state reduction of the plastoquinone pool is approximately 50%. At this point, EET is activated to facilitate re‐oxidation of the photosynthetic electron transport chain. At a QF of approximately 0.6 fmol photons cell −1 sec −1 , the EET pathways are saturated, the steady‐state reduction of the plastoquinone pool reaches 70%, and NPQ is activated to assist in dissipating captured photons. Therefore, the model accurately recreates the onset of NPQ that occurs when light absorption outpaces the ability to utilize this energy within metabolism. Interestingly, this model is consistent for cells acclimated to both LL and HL, and Phaeodactylum is known to maintain high capacity for NPQ under both light conditions (Taddei et al., 2018 ). Currently, experimentally derived photoautotrophic metabolic flux values for P. tricornutum do not exist; thus, our flux predictions are hypotheses that still require validation. Still, the approach outlined in this study is generally applicable to all phototrophic genome‐scale model simulations and previous efforts using experimentally‐derived electron transport efficiencies, as opposed to PAM, showed good agreement with 13 C metabolic flux analysis (Broddrick, Welkie, et al., 2019 ). The interest in integrating light‐driven metabolism into bioengineering and synthetic biology necessitates an iterative framework for bioprocess development. The DBTL paradigm has enabled rapid increases in bioproduct titers (Carbonell et al., 2018 ; Petzold et al., 2015 ). Computational tools are integral to these workflows. Relevant to the design step of the process, GEMs have been extensively used to rationally engineer metabolism to generate a wide variety of phenotypes (Bang et al., 2020 ; Czajka et al., 2021 ; Li et al., 2019 ). We explored implementing the modeling framework towards light‐driven bioproduct formation, as adoption of these approaches is underrepresented in phototrophic systems. Quantifying the reductant cost of cellular macromolecules provided insight into the theoretical yield of different compound classes (Table 3 ), with protein biosynthesis being the predominant energy sink. EET pathways, which were found to consume 29% of LET in HL, serve to oxidize the photosynthetic electron transport chain and resupply low energy cofactors to autotrophic metabolism. However, they are generally viewed as wasteful as this energy could be utilized to increase biomass yields (Peers, 2014 ). We utilized our model to estimate the metabolic consequences of redirecting metabolism towards important chemical precursors. The overproduction of plastid fatty acids (hexadecanoate), the shikimate pathway (chorismate), and isoprenoid precursors (isopentenyl pyrophosphate) were only partially fueled by reductant that would normally be dissipated by EET, showing there is still considerable potential to engineer primary photosynthetic metabolism to increase bioproducts yields. However, the linearity of the product formation versus fraction of redirected biomass (Figure 5 ) suggests the system is carbon limited, not reductant limited. Thus, increasing the available DIC during culturing may result in synergistic flux modes where the EET is more efficiently utilized for bioproduction without a linear decrease in biomass production. Furthermore, our modeling suggests that increasing the flux of reduced carbon to metabolic precursors of interest may require significant engineering of central carbon metabolism. For instance, increasing the production of our three selected metabolites increased the flux of 3‐phospho‐ d ‐glycerate (3 pg) through the plastid glycolytic pathway (Figure 5 ). Additionally, the increased flux of carbon to isoprenoid biosynthesis or through the shikimate pathway increased the demand for CBBC sourced biosynthetic intermediates (glyceralde‐3‐phosphate and d ‐erythrose‐4‐phosphate, respectively). Photosynthetic microorganisms, including diatoms, redirect carbon and energy to storage molecules under conditions of nutrient deprivation, such as nitrogen limitation (Alipanah et al., 2015 ; Park et al., 2015 ). This phenotype forms the basis for much of the interest in biofuel applications of these phototrophs (Levering et al., 2015 ). Thus, a reasonable strategy and potential future direction, is the diel separation of carbon fixation and bioproduct formation. Rerouting these metabolites from sugar polymer degradation via the pentose phosphate pathway or mitochondrial β‐oxidation of lipids (Jallet et al., 2020 ) may alleviate some of the pressure on the CBBC. While pathway engineering alone is often sufficient for successful bioengineering of heterotrophic microorganisms, photoautotrophic systems must account for additional design parameters, with light availability being of paramount importance. Our simulations demonstrated the utility of assessing the impact of light environment and inoculation density on overall productivity (Figure 6 ). Light ultimately drives the formation of biomass and, expectedly, modeled yields of product are lower for cells inoculated in low light versus high light. This is likely attributable to the fact the light availability cannot overcome maintenance energy requirements of dense cultures. Focusing on the high light simulations, we observed that initial inoculation density has a major effect on the yield of desired bioproduct and that this is non‐linear (Figure 6c ). Reduced yields at increasing inoculation density are due to self‐shading of the culture that reduce the amount of energy available for product formation. While our simulation held constant variables such as light level, inorganic carbon availability, and photophysiology, the modeling construct presented here can incorporate time dependent changes in these, as well as other variables, towards comprehensive bioprocess design. The next step is to build and test these strains to initiate the first iteration of the DBTL cycle. During the test phase, the model constraints provide a roadmap for relevant process parameters and a framework to evaluate process performance, including assessments on EET usage efficiency. It is important to note our design simulations do not include possible changes in photophysiology as a result of strain engineering (e.g. an increase in P \n O ). However, physiological outputs from the testing of the strain designs proposed can be re‐integrated into the modeling framework to include changes in experimental constraints. This contributes to the learn step of the DBTL cycle, closing the loop on the first iteration and enabling an updated design strategy for the next iteration."
} | 7,238 |
19627569 | PMC2727540 | pmc | 1,154 | {
"abstract": "Background Coral reef ecosystems are renowned for their diversity and beauty. Their immense ecological success is due to a symbiotic association between cnidarian hosts and unicellular dinoflagellate algae, known as zooxanthellae. These algae are photosynthetic and the cnidarian-zooxanthellae association is based on nutritional exchanges. Maintenance of such an intimate cellular partnership involves many crosstalks between the partners. To better characterize symbiotic relationships between a cnidarian host and its dinoflagellate symbionts, we conducted a large-scale EST study on a symbiotic sea anemone, Anemonia viridis , in which the two tissue layers (epiderm and gastroderm) can be easily separated. Results A single cDNA library was constructed from symbiotic tissue of sea anemones A. viridis in various environmental conditions (both normal and stressed). We generated 39,939 high quality ESTs, which were assembled into 14,504 unique sequences (UniSeqs). Sequences were analysed and sorted according to their putative origin (animal, algal or bacterial). We identified many new repeated elements in the 3'UTR of most animal genes, suggesting that these elements potentially have a biological role, especially with respect to gene expression regulation. We identified genes of animal origin that have no homolog in the non-symbiotic starlet sea anemone Nematostella vectensis genome, but in other symbiotic cnidarians, and may therefore be involved in the symbiosis relationship in A. viridis . Comparison of protein domain occurrence in A. viridis with that in N. vectensis demonstrated an increase in abundance of some molecular functions, such as protein binding or antioxidant activity, suggesting that these functions are essential for the symbiotic state and may be specific adaptations. Conclusion This large dataset of sequences provides a valuable resource for future studies on symbiotic interactions in Cnidaria. The comparison with the closest available genome, the sea anemone N. vectensis , as well as with EST datasets from other symbiotic cnidarians provided a set of candidate genes involved in symbiosis-related molecular crosstalks. Altogether, these results provide new molecular insights that could be used as a starting-point for further functional genomics studies.",
"conclusion": "Conclusion This large EST collection has provided high quality data on all aspects of a temperate symbiotic cnidarian, particularly with regard to coding sequences and regulation features. For example, we identified many novel repeated elements (RE) in 3'UTRs, suggesting an invasion of most animal sequences by some specific RE families. It will be interesting to further investigate their potential biological role, particularly on gene regulation. Phylogenetic origin and functional classification of the holobiont sequences allowed the identification of several symbiotic candidate genes. These data are now being used to develop a dedicated microarray that will provide a valuable resource for future studies on symbiotic interactions in A. viridis . Furthermore, these data also have shown the importance of Symbiodinium symbionts as well as the associated flora of the three major prokaryotic species. Some sequences will be further analysed from the new perspective of gene transfer between host and symbiont. The relatively low abundance of sequences from Symbiodinium was attributed to experimental bias; a new ongoing sequencing project should fill this gap. Finally, these data from a temperate zone cnidarian provide novel molecular insights that will complement those obtained from tropical anthozoans. This dataset is valuable resource that will be of great help for comparative genomics and evolutionary studies.",
"discussion": "Results and discussion ESTs generation and analysis A cDNA library was made from both symbiotic (gastroderm) and non-symbiotic (epiderm) tissues of the sea anemone Anemonia viridis (Additional file 1 ). In order to maximize the diversity of genes expressed under both normal and stress conditions within the symbiotic association, sea anemones were subjected to different environmental conditions before RNA extraction and cDNA library construction (light/dark sampling, thermal stress, hyperoxia conditions, see Materials & Methods). Out of 50,304 sequenced clones, a total of 41,247 readable sequences were produced, corresponding to a sequencing success of 82%, and 39,939 high quality ESTs were generated (Figure 1 and Additional file 3 ). A. viridis repetitive sequences were identified and masked using RepeatMasker before assembly. Trimmed masked ESTs were further subjected to cluster analysis using the TIGR-TGICL pipeline with default parameters [ 32 ]. A total of 30,087 ESTs were assembled into 4,652 contigs (Additional file 4 ) while 9,852 sequences remained as singletons. The mean trimmed EST sequence length was 626 bp, but sequences could be assembled into contigs of up to 4 kb (CL4Contig4, described as ribosomal protein L8, 3,887 bp). Most contigs (2,959) were composed of three ESTs or less, with an average length of 895 bp, suggesting that they covered only parts of the transcript sequences. Apart from the functional genomics resources available for the reef-building corals Acropora palmata and Montastrea faveolata [ 24 ], most cnidarian genomic studies have been performed on non-symbiotic tissues (aposymbiotic developmental stages or adults). To our knowledge, our study provides the largest EST collection from a symbiotic cnidarian. Figure 1 Flowchart of the analysis pipeline of A. viridis ESTs . EST processing and analysis pipeline used in this study. The most abundant transcripts (contigs composed of more than 100 ESTs) are listed in Table 1 . Most of these correspond to ribosomal proteins, or proteins involved in basic cellular processes (actin, elongation factor 1 alpha, CAAT enhancer binding protein). However, it should be highlighted that two of these highly abundant transcripts had no similarity with any available sequence, either in UniProt or NR databases, while two other large clusters represent proteins that might be involved in stress responses (HSP70 and ferritin). Ferritins are ubiquitously expressed proteins that have a central role in normal iron homoeostasis, as well as during oxidant stress, by reducing the participation of iron in free-radical-generating reactions [ 33 ]. Ferritin genes have also been identified as being highly expressed in Acropora EST datasets and one has been shown to be under positive selection [ 24 ]. Microarray experiments, designed to identify genes differentially expressed in response to elevated temperature in Anthopleura elegantissima [ 34 ] or A. viridis (data not shown), revealed that ferritin genes might be induced in response to increases in reactive oxygen species following thermal stress. Table 1 The 19 most abundant transcripts (>100 ESTs) Top hit Swissprot (2008.03) Cluster ID Total ESTs Length (bp) Description N. vectensis top hit (Uniprot ID) Accession Name Organism E-value CL1Contig11 828 478 CCAAT/enhancer-binding protein beta A7S4U3_NEMVE O02755 CEBPB_BOVIN Bos taurus 5E-17 CL1Contig8 482 1892 Elongation factor 1-alpha A7SSW8_NEMVE P19039 EF1A_APIME Apis mellifera 0 CL2Contig7 340 1411 Actin, cytoplasmic A7SCN8_NEMVE Q964E3 ACTC_BIOAL Biomphalaria alexandrina 0 CL3Contig4 334 764 Soma ferritin A7S8I6_NEMVE P42577 FRIS_LYMST Lymnaea stagnalis 4E-68 CL1Contig4 322 968 lipoprotein PA4545 precursor No hits found P33641 Y9F5_PSEAE Pseudomonas fluorescens 2E-31 CL4Contig4 311 3887 60S ribosomal protein L8 A7S5T9_NEMVE Q9U9L2 RL8_ANOGA Anopheles gambiae 2E-35 CL5Contig1 234 855 No hits found No hits found CL11Contig1 175 2915 Translation elongation factor 2 A7RSB9_NEMVE Q1HPK6 EF2_BOMMO Bombyx mori 0 CL3Contig3 172 1418 Pancreatic secretory granule membrane major glycoprotein GP2 A7S5P2_NEMVE P25291 GP2_CANFA Canis lupus familiaris 2E-18 CL12Contig1 170 2227 Polyadenylate-binding protein 4 A7SSV8_NEMVE Q13310 PABP4_HUMAN Homo sapiens 0 CL13Contig1 152 887 40S ribosomal protein S2 A7RJJ7_NEMVE P49154 RS2_URECA Urechis caupo 1E-114 CL2Contig1 148 1487 Actin, cytoskeletal 1A A7RU31_NEMVE P53472 ACTA_STRPU Strongylocentrotus purpuratus 0 CL7Contig2 145 1224 ADP, ATP carrier protein, mitochondrial precursor A7RN38_NEMVE P31691 ADT_ORYSJ Oryza sativa 1E-104 CL1Contig3 138 528 No hits found No hits found CL14Contig1 134 1084 60S acidic ribosomal protein P0 A7SQ36_NEMVE P47826 RLA0_CHICK Gallus gallus 1E-108 CL18Contig1 108 1100 40S ribosomal protein SA A7RKS5_NEMVE P50890 RSSA_CHICK Gallus gallus 3E-94 CL20Contig1 106 2872 40S ribosomal protein S3a A7S3J7_NEMVE P49242 RS3A_RAT Rattus norvegicus 1E-100 CL19Contig1 105 847 Heat shock 70 kDa protein cognate 4 A7SG65_NEMVE Q9U639 HSP7D_MANSE Manduca sexta 0 CL6Contig3 102 866 40S ribosomal protein S4 A7SRV8_NEMVE P47961 RS4_CRIGR Cricetulus griseus 1E-112 Repeated elements Based on reverse transcriptase domain search, 46 transposable elements (retrotransposons/retroposons) were identified (data not shown). Among these, 4 were of prokaryotic origin (without any similarity to N. vectensis sequences) and 42 were of metazoan origin. While 2 of them were almost identical to N. vectensis transposable elements, 21 only had slight similarity to N. vectensis sequences (BlastX or BlastN with E-value of 1.10 -25 to 1.10 -10 ), and 19 showed no similarity to N. vectensis sequences but had been previously identified in other Metazoa. Self BlastN analysis on our EST dataset also identified another abundant set of repeated elements. Over 1,500 repetitive sequences were identified, with 412 repeated more than 10 times in our UniSeqs dataset. A typical sequence would be around 200 bp length, harbouring 40–50 bp inverted terminal repeats at its extremities. These repeats often formed very long and stable hairpins, depending on their orientation (Figure 2 ). They were found in 19.8% of UniSeqs and were overrepresented within contigs (62% of contigs contained at least one repeat). When open reading frame could be identified, these A. viridis repeated elements were always mapped in the 3'UTRs of the protein coding genes. Using a BlastN analysis, none of these sequences were found either in the genome of N. vectensis or in the large EST datasets of Hydra magnipapillata , and search in Repbase did not show any homologs. However, members of these A. viridis repeated elements could be found in the limited EST datasets available for the symbiotic coral Acropora millepora and the sea anemone Anthopleura elegantissima . Further structural and functional characterisations of these repeated elements are currently being made. Figure 2 Representative repeated element found in many 3'UTRs of A. viridis coding genes . A: Schematic representation of the same repeated element found downstream of the ORF of the following A. viridis homologs: thioredoxin, myosin (opposite orientation) and VTA1 (truncated form). B: Calculated secondary structure (Mfold2.3 program) of the two first repeats, showing the formation of an extremely stable hairpin in the 3'UTRs of the myosin gene. Comparative analysis to related taxa Because the cDNA library was made from symbiotic tissue, we expected to find ESTs related both to the cnidarian host and to its dinoflagellate symbionts. The analysis pipeline used in this study is presented in Figure 1 . Sequences were compared with SwissProt (2008.03) and Uniprot KB (TrEMBL+Swissprot) (2008.03) databases using BlastX with a cutoff E-value of < 1.10 -10 to retrieve functional annotations. Of the assembled dataset, 6,238 UniSeqs had putative similarities while 8,266 had no similarity to any sequences in the chosen databases. Sequences were also compared with NCBI-indexed prokaryotic nucleic sequences (2007.08 release), using BlastN with a stringent E-value of < 1.10 -15 to assess the proportion of prokaryotic sequences. Finally, a specific search of the UniSeqs for virus proteins returned 7 hits. Relative contribution is shown in Figure 3 . First of all, a relatively high proportion of sequences (57%) remained that showed no significant similarities to previously described genes and were therefore considered as 'unknown'. This is somewhat comparable with results obtained from other cnidaria, Acropora palmata , Acropora millepora , Montastrea faveolata , and Nematostella vectensis [ 19 , 24 ]. Most of the UniSeqs identified in Symbiodinium sp were also of unknown origin [ 28 ]. Metazoan hits were found for 32.3% of UniSeqs (75% of annotated sequences). Among these, most of the annotated UniSeqs (4,266 out of 4685) matched with Nematostella vectensis predicted proteins. However, we also identified sequences that were clearly from the host (first Blast hits of metazoan origin with cutoff E-value < 1.10 -50 ), but these had no significant similarity to predicted proteins of N. vectensis . Three of these (a glycoprotein, a ferroxidase and an amine-oxidase) were studied in more detail (Figure 4 ). PCR and sequencing were first performed on genomic DNA from both A. viridis epiderm and in vitro cultured Symbiodinium , which confirmed the animal origin of these sequences (Figure 4A ). The first sequence studied is related to the ependymin glycoprotein family (more specifically to the Mammalian Ependymin-Related Proteins group or MERP1), which has been well described in vertebrates due to its involvement in the regeneration processes. Ependymins are secretory proteins that can bind calcium and that were found predominantly in the cerebrospinal fluid of teleost fish. A bound form has been described, associated with the extracellular matrix. Recent data demonstrated that these proteins are also present in non-vertebrate deuterostomes and protostomes [ 35 ], and that positive selection may have shaped their evolution. Figure 4B illustrates an amino acid alignment of A. viridis MERP sequences to homologous proteins found in publicly available databases (Bayesian analysis of the MERP sequence confirmed the phylogenetic relationship, not shown). The presence of ependymin proteins in basal organisms clearly indicates for the first time that this protein family is far older than previously thought (first described as chordate-specific, then deuterostome-specific, and finally also found in protostomes). In addition, there are at least 3 distinct MERP homologues in A. viridis , all of animal origin. As this is never the case in other phyla (only one homologue was found in any other species), we could hypothesize that MERP are important for symbiosis in cnidarians. The second sequence is related to a copper-dependent ferroxidase family protein (hephaestin) involved in copper detoxification, which has been described both in eukaryotes (unicellular eukaryotes and vertebrates) and prokaryotes. Nevertheless, no significant similarity was found in D. melanogaster or in C. elegans genomes, nor in any available protostome database, suggesting that this has been lost during protostome evolution. The third sequence is related to a prokaryotic amine-oxidase (permease) family protein, which has also been identified in the genome of C. elegans and other Caenorhabditidae but not in any other animal sequence database. Such species-specific occurrence could be explained by the presence of a similar prokaryote in both nematoda and A. viridis associated flora (sequence contamination) or a parallel gene transfer event during their evolution. However, molecular phylogenetic analyses (neighbor joining, maximum parsimony and maximum likelihood) showed that cnidarian and rhabditidae nematode sequences cluster to outside other prokaryote representative gene sequences (Additional file 5 ). The latter rather suggest that amine-oxidase genes were maintained in these taxa but lost in other metazoa. Figure 3 Distribution of BLAST results . Pie charts of the distribution of UniSeqs by organism. A: Distribution of all UniSeqs (unknown vs annotated sequences). B: Distribution of annotated UniSeqs. The sequences were annotated using the pipeline shown in Figure 3, and assigned a putative origin based on the Blast results. The term \"putative associated flora\" covers sequences from prokaryotes (21% of annotated sequences), fungi (0.3%) and viruses (0.1%). Figure 4 Evidence for specific metazoan genes isolated from Anemonia , but not found in the Nematostella genome . A: PCR results confirming the metazoan origin of MERP, Ferroxidase and Amine-oxidase genes. Amplifications were performed on genomic DNA from ectodermal cells (specific to A. viridis , E ), cultured zooxanthellae (specific to Symbiodinium , CZ ), or holobiont tissues (both A. viridis and Symbiodinium , H ), and water as a negative control ( C ). Three other specific genes were used to confirm the resulting profiles: S sp EF2 is the nuclear gene for Symbiodinium sp elongation factor 2, S sp psbA is the chloroplast-encoded gene for Symbiodinium sp photosystem II protein D1, and Av EF1 is the nuclear gene for A. viridis elongation factor 1 alpha. B: Amino acids alignment of several MERP sequences from Anemonia viridis (Av) Acropora millepora (Am), Mus musculus (Mm), Homo sapiens (Hs) and Carassius auratus (Ca). Signal peptides are highlighted in yellow. Quite a large proportion of UniSeqs (9%) were uniquely shared with prokaryotes, of which Proteobacteria was the most prominent bacterial group ( Pseudomonas , Bordetella , and Burkholderia ). The holobiont has already been described as a dynamic assemblage, made up of the animal host, zooxanthellae, endolithic algae and fungi, Bacteria, and Archea [ 7 ]. Such \"prokaryotic sequences\" could therefore be assigned to the sea anemone-associated flora. In addition, a small but significant number of A. viridis sequences, recognized as being of prokaryotic origin based on Blast analysis, had already been identified in the N. vectensis genome. Such genes are clearly similar in sequence to prokaryotic homologs, although they contain introns. In N. vectensis and A. millepora they were proposed as \"ancient genes\", conserved in cnidarians but lost in other animal genomes [ 19 ]. Although our results are in line with this interpretation, comparative genomic studies among cnidarians, such as A. viridis , could help to identify the most probable evolutionary scenario between maintenance of \"non-metazoan\" genes in cnidarians or lateral gene transfer events, followed by rapid intron acquisition. Surprisingly, only a small fraction of our dataset could be assigned to unicellular eukaryote sequences (putative Symbiodinium sp, 3.6% of annotated sequences). Two well accepted explanations have been proposed: i) poor representation of dinoflagellate sequences in databases, leading to wrong assignment after Blast analysis; and ii) technical bias due to the Symbiodinium cell wall impairing complete RNA extraction with standard methods, thus leading to an under-representative number of cDNAs in our library. GC content was calculated for all UniSeqs to better assign a source to the transcripts (species of origin). Figure 5 represents the distribution of GC content in the three major taxa (metazoans, unicellular eukaryotes and associated prokaryotic flora), as well as for sequences without BlastX hits (unknown). Two distinct distribution curves were observed, with the metazoan ESTs averaging 40% GC, while bacterial ESTs averaged 63%. Despite our small number of Symbiodinium sequences, their GC content (58%) agreed with that already published for dinoflagellates, which have a relatively high GC content compared with other eukaryotes: Triplett et al [ 36 ] calculated a GC content of 64% for Heterocapsa pygmaea and an EST analysis of Alexandrium tamarense estimated that coding region GC-content was 60.8%, whereas GC-content in the 3'-UTR was slightly less, at 57.6% [ 37 ]. GC-content analysis of our unknown sequences revealed that ESTs were clustered around two peaks corresponding to 35% and 58% GC content for the larger and smaller peaks respectively. These data strongly suggest that most of the unknown sequences are likely to be of animal origin. As the GC content of 3' UTRs is usually lower than that of 5' UTRs or CDSs, these results are also consistent with our hypothesis that these sequences are mainly 3' UTR (due to cloning strategy), which also explains their low level of annotation. On the other hand, the smaller peak (58% GC) could be attributed to dinoflagellate sequences, although no dinoflagellate genome is currently available to test this, and the most closely related available genomes are those of apicomplexans ( Plasmodium falciparum and Toxoplasma gondii ), which are also compact genomes of parasitic protozoans. Figure 5 GC-content distribution of unique sequences . The GC-content was calculated for all UniSeqs and compared with Blast results to assign a putative origin to the sequences. The GC-content distribution is presented by putative origin for annotated sequences ( A. viridis , Symbiodinium or associated flora). Results related to sequences of unknown function (dashed line) are discussed in the text. To gain some insight into genes potentially involved in symbiosis among cnidaria, we performed comparative Blast analyses on the subpart of A. viridis sequences that were of animal origin (12,448 sequences). They were subjected to tBlastX (cutoff E-value of < 1.10 -10 ) against cnidarian EST sequences available in NCBI-dbEST (2009.05). A parallel BlastX (cutoff E-value of < 1.10 -10 ) was performed against the N. vectensis predicted proteins dataset. Table 2 gives the number of positive Blast hits (presence vs absence) obtained from A. viridis sequences against ESTs from selected cnidarian species. This comparison highlighted two species with a high number of homologous sequences, the sea anemone M. senile and the symbiotic stony corals A. millepora and A. palmata . In the Figure 6 Venn diagram we compared A. viridis with two non symbiotic species ( N. vectensis and M. senile ), and the symbiotic Acropora species: 1,535 sequences were common to the four species, while only 129 sequences (of which 88 have no BlastX hit) were specific to the symbiotic species, which probably contain candidate genes required for the symbiosis. Figure 6 Venn diagram illustrating the distribution of positive blast hits between A. viridis and selected Cnidaria . Positive hits were identified using tBlastX (ESTs from M. senile and Acropora ) or BlastX (predicted proteins dataset of N. vectensis ), using a cutoff E-value of < 1.10 -10 . 129 genes are common to the symbiotic species, A. viridis and at least one species of Acropora . Table 2 Distribution of positive blast hits among Cnidaria tblastx (1e-10) Species Number of ESTs Number of hits Hydra magnipapillata 164325 2355 Nematostella vectensis 163314 4475 Metridium senile 29412 3226 Clytia hemisphaerica 27674 1637 Hydra vulgaris 18830 992 Acropora palmata 13945 1319 Porites astreoides 11516 1306 Aiptasia pallida 10285 1648 Acropora millepora 10247 2151 Hydractinia echinata 9460 951 Montastraea faveolata 3873 487 blastx (1e-10) Nematostella vectensis 27273 1 4266 Sequences identified as metazoan origin (12,448 sequences) were compared to cnidarian sequences available in databases. 1 predicted proteins dataset from JGI Functional classification of ESTs Finally, this set of UniSeqs was annotated using similarity searches in both nucleotide and protein databases, as well as domain searches. Gene ontology terms were then assigned automatically using customized scripts based on InterProScan search results. We also submitted N. vectensis predicted proteins to the same InterProScan analysis for a comparative approach. To homogenize annotation level we only kept the root domain, using the hierarchical domain organization available from EBI. InterPro protein domains were identified in 40% of proteins. Based on similarity searches (Figure 3 ), we assigned putative taxa ( A. viridis , Symbiodinium or associated flora) to each unique sequence. Figure 7 represents the distribution of assigned gene ontology terms. Comparison of domain occurrence in A. viridis with that in N. vectensis demonstrated that our library is highly representative of the Actinaria transcriptome. We also identified increased abundance of a number of molecular functions in A. viridis compared with N. vectensis : translation regulator activity (0.46 vs 0.23%), oxidoreductase activity (6.86 vs 4.98%), antioxidant activity (0.07 vs 0.03%), and structural molecule activity (3.24 vs 1.28%). It is noteworthy that some of these molecular functions are involved in symbiotic interactions with zooxanthellae, such as protein binding [ 24 ], nutrient transport and antioxidant activity [ 8 ]. For example, A. viridis counters the oxidative stress produced by the photosynthetic activity of its symbionts by using a great diversity of antioxidant defences [ 38 ]. Some biological processes were also increased in the A. viridis dataset compared with N. vectensis : secretion (0.81 vs 0.36%) and catabolism (0.65 vs 0.26%). The cytoplasmic components were also found to be more highly represented in A. viridis compared with N. vectensis (20.42 vs 9.59%). Figure 7 Distribution of gene ontology terms . From the A. viridis library, a total of 29% of the UniSeqs were assigned GO terms. Unique sequences were also sorted using their putative origin ( A. viridis , Symbiodinium sp or associated flora). For N. vectensis , 51% of the 27,273 predicted proteins were assigned GO terms. All these differences between the present A. viridis dataset and the N. vectensis genome may reflect the crucial role of trophic exchanges between the sea anemone and its dinoflagellate symbionts, as well as specific host adaptations."
} | 6,491 |
37316548 | PMC10267220 | pmc | 1,155 | {
"abstract": "Coral reefs are under existential threat from climate change and anthropogenic impacts. Genomic studies have enhanced our knowledge of resilience and responses of some coral species to environmental stress, but reference genomes are lacking for many coral species. The blue coral Heliopora is the only reef-building octocoral genus and exhibits optimal growth at a temperature close to the bleaching threshold of scleractinian corals. Local and high-latitude expansions of Heliopora coerulea were reported in the last decade, but little is known about the molecular mechanisms underlying its thermal resistance. We generated a draft genome of H. coerulea with an assembled size of 429.9 Mb, scaffold N50 of 1.42 Mb and BUSCO completeness of 94.9%. The genome contains 239.1 Mb repetitive sequences, 27,108 protein coding genes, 6,225 lncRNAs, and 79 miRNAs. This reference genome provides a valuable resource for in-depth studies on the adaptive mechanisms of corals under climate change and the evolution of skeleton in cnidarian."
} | 259 |
29049419 | PMC5648182 | pmc | 1,156 | {
"abstract": "The production of lignocellulosic-derived biofuels is a highly promising source of alternative energy, but it has been constrained by the lack of a microbial platform capable to efficiently degrade this recalcitrant material and cope with by-products that can be toxic to cells. Species that naturally grow in environments where carbon is mainly available as lignin are promising for finding new ways of removing the lignin that protects cellulose for improved conversion of lignin to fuel precursors. Enterobacter lignolyticus SCF1 is a facultative anaerobic Gammaproteobacteria isolated from tropical rain forest soil collected in El Yunque forest, Puerto Rico under anoxic growth conditions with lignin as sole carbon source. Whole transcriptome analysis of SCF1 during E . lignolyticus SCF1 lignin degradation was conducted on cells grown in the presence (0.1%, w/w) and the absence of lignin, where samples were taken at three different times during growth, beginning of exponential phase, mid-exponential phase and beginning of stationary phase. Lignin-amended cultures achieved twice the cell biomass as unamended cultures over three days, and in this time degraded 60% of lignin. Transcripts in early exponential phase reflected this accelerated growth. A complement of laccases, aryl-alcohol dehydrogenases, and peroxidases were most up-regulated in lignin amended conditions in mid-exponential and early stationary phases compared to unamended growth. The association of hydrogen production by way of the formate hydrogenlyase complex with lignin degradation suggests a possible value added to lignin degradation in the future.",
"conclusion": "Conclusions Several omic techniques (genomics, transcriptomics, proteomics and metabolomics) are suitable to improve our understanding of microbial communities, enzyme interactions, and how lignocellulose breakdown occurs, in both isolated cultures or microbial communities [ 69 , 70 ]. Although, there is current limited understanding of how aromatic metabolism is capable to support anaerobic lignin degradation, the set of up-regulated enzymes during lignin-amended growth could conceivably suggest their role during this process. By examining the degradation of lignin in batch culture over time, we were able to capture the transcripts involved in the initial accelerated growth and transport of glucose in the presence of lignin during early exponential phase, the diverse complement of lignin-degrading enzymes expressed during mid-exponential growth, and transporters and metabolic machinery responsible for the increased formate production observed concurrent with 60% of lignin reduced. This work will contribute to our understanding of anaerobic lignin degradation as well as the cellular machinery necessary to balance metabolism and manage toxic products formed as a result of growth in lignin amended media.",
"introduction": "Introduction One of the key challenges that environmental sciences currently face is the search for renewable energy sources that are capable to satisfy future energy demands. Among several choices, biofuels derived from biomass have emerged as one of the most promising for decreasing the harmful effects produced by the uncontrolled use of fossil fuels [ 1 , 2 ]. Lignocellulosic biomass represents close to the 90% of the dry weight of the total biomass material [ 3 , 4 ]. Lignocellulose is composed of cellulosic and hemicellulosic polysaccharides, pectic polymers, and lignin. Lignin is a phenylpropanoid-derived heteropolymer that comprises 10–30% of the lignocellulosic biomass. The stable and complex physical structure of lignin together with its chemical inhibitory nature confers rigidity to plant tissue and also serving as a protection from pathogens [ 5 , 6 ]. Due to its aromatic nature and highly branched polymer network, lignin is resistant to biological degradation and it is the primary material responsible for biomass recalcitrance [ 7 ]. Efforts aimed to advance lignin bioengineering in combination with novel lignin separation and biodegradation technologies are required to make lignocellulosic biofuels competitive with fossil fuels. Microbial lignin degradation was thought to be a negligible bioprocess in the environment until 1928 when experiments carried out by Bavendam and collaborators showed that lignin could be biologically depolymerized [ 8 ]. These initial studies were supported by several studies carried out with white-rot or brown-rot fungi [ 9 , 10 ]. In general, those microorganisms utilize extracellular peroxidases and laccases-mediated mechanisms during the aerobic degradation of lignin [ 5 ]. More recent studies have shown that lignin monomers such as vanillin, vanillic acid, ferulic acid, catechol and cinnamic acid can be metabolized under both aerobic and anaerobic conditions [ 11 – 14 ]. For instance, the phototrophic Rhodopseudomonas palustris [ 15 ], and the denitrifying Thauera aromatica [ 16 ] are well know because of their capability of anaerobic degradation of aromatic compounds. More recent experiments demonstrated that lignin can be degraded in the absence of oxygen [ 11 , 17 ]. Facultative anaerobes have been shown to be capable of using lignin as the sole carbon source under anaerobic conditions, and we have isolated numerous lignin-degrading bacteria from El Yunque experimental forest, Puerto Rico based on their ability to grow on lignin as sole C source: Klebsiella sp. strain BRL6-2 [ 18 ], Tolumonas lignolytica sp. nov [ 19 ], and Enterobacter lignolyticus SCF1 [ 20 ]. Enterobacter lignolyticus SCF1 is a facultative anaerobic Gammaproteobacteria. Its close relative Enterobacter soli strain LF7, was also isolated from tropical forest soils in Peru using lignin as sole C source [ 21 ]. Although, these soils are characterized by periods of extremely low availability of oxygen and extended periods of low redox [ 22 ], tropical forest soils have been shown to have among the fastest rates of plant litter decomposition worldwide [ 23 – 25 ]. E . lignolyticus SCF1 sparked initial attention due to its ability to grow on lignin as a sole carbon source in the absence of oxygen, as well as its high tolerance to 1-ethyl-3-methylimidazolium chloride, a toxic compound utilized as a lignocellulosic biomass pretreatment in the production of biofuels [ 26 ]. These studies employed RNA-seq, a sensitive tool used to better understand the physiological status of bacteria by quantifying genome-wide gene expression [ 27 ]. Growth of E . lignolyticus SCF1 in lignin revealed increased 4-hydroxyphenylacetate degradation pathway, catalase/peroxidase enzymes, and the glutathione biosynthesis and glutathione S-transferase (GST) proteins, which have been previously implicated in anoxic lignin degradation. Growth on lignin was also associated with increased production of NADH-quinone oxidoreductase, other electron transport chain proteins, and ATP synthase and ATP-binding cassette (ABC) transporters, suggesting that lignin-amended conditions include extra reductive capacity in addition to observed oxidative and hydrolytic lignin-reducing enzymes [ 20 ]. RNA-seq studies are often conducted as static sampling experiments [ 28 ], based on the analysis of transcriptomes of microorganisms facing conditions at a specific time-point during growth. However, microorganisms respond to environmental changes in a dynamic fashion, which was the motivation for examining lignin degradation across three time points corresponding to different stages of growth. Our current experiment also differs from previous experiments due to the different substrate utilized. In former experiments, xylose, a pentose sugar commonly present in hydrolysates from lignocellulosic biomass was utilized as main C source. Those conditions may mimic slow-growing conditions that may be often found in soil environments. However, we observed that SCF-1 grew poorly during those experiments (achieving OD 600 ~0.03) suggesting that this condition may have generated an additional stress that could obscure the global transcriptional response found when SCF-1 was incubated in the presence of lignin. Indeed, under those conditions, only the 6% (273/4599) of the genes were found to be differentially expressed when lignin was present [ 20 ]. In contrast, our current experiments were based on SCF-1 growing in glucose, which was chosen as a fast-growing condition, mimicking environmental conditions encountered in nutrient-rich ecosystems, such as those found by gastrointestinal microbiota. Both conditions differ not only in the affinity of their carbon source, but also in the quantity of energy (ATP) and reducing equivalents that the main pathways are capable to produce ( S1 Fig ). To improve our understanding of SCF1 lignin degradation over time [ 20 ], an experiment was designed to capture the gene expression profile of lignin degradation during growth (early, mid-exponential and beginning of stationary phase of growth) and in the presence and absence of lignin.",
"discussion": "Results and discussion To gain insight into physiological differences during lignin degradation, the gene transcript abundances in Enterobacter lignolyticus SCF1 were compared in cultures grown on glucose minimal media to the same media amended with lignin. Our experimental design was intended to characterize temporal dynamics during the degradation of lignin, as a way of illuminating the series of progressive biological mechanisms involved in lignin degradation. Populations were sampled for transcriptomics analysis at early exponential phase (EE), mid-exponential phase (ME) and early stationary phase (ES). Global view of the response to lignin SCF-1 achieved twice the cell density when lignin was added in the medium compared to unamended growth ( Fig 1A , S2 and S3 Figs). This difference is associated with the degradation of lignin, since in the lignin-amended cultures there was a 40% lower consumption of glucose (~8 mM) than in the unamended cultures ( Fig 1B and 1C ). It has been previously shown that SCF-1 can utilize lignin as the sole carbon source during anaerobic growth in agar [ 20 , 23 ]. Although we have been unable to detect growth of SCF-1 as sole C source in liquid medium, it has been previously reported that SCF-1 is capable to degrade lignin up to 56% within 48 hours in the presence of xylose [ 20 ]. When growing on lignin-amended LS4D liquid medium in the presence of glucose, there was a first drop in the content of lignin of about 30% (0–30 hours) at the beginning of the growth curve, that resulted in SCF-1 achieving a slightly higher cell concentration than the control. After 30 hours, the concentration of lignin was constantly decreasing until SCF-1 ultimately degraded 60% of the lignin amended in the medium. 10.1371/journal.pone.0186440.g001 Fig 1 (A) Growth curve of Enterobacter lignolyticus SCF-1 in glucose medium (open circles) and glucose medium amended with lignin (closed circles). Arrows show the times samples were harvested for transcriptomic analysis (open arrows, glucose medium and closed arrows, glucose medium amended with lignin). The concentration of lignin of amended cultures is shown in closed triangles and control (medium mas lignin, no cells) is shown in open triangles. Metabolites concentrations along growth of unamended cultures (B) and lignin amended cultures (C) . Glucose is shown in red squares, ethanol, in gray triangles, acetate, in black triangles, lactate, in darker green circles, succinate in light blue and formate, lighter green circles. One third (34%) of the 4445 predicted protein-encoding ORFs in the genome of SCF1 were differentially expressed at any time point during growth on glucose amended with lignin (mean 1533 genes +/- 275 standard error, Table 1 ). The highest number of differentially expressed genes was detected in mid-exponential phase (ME, 2465). Genes that were over expressed in more than a single time point represented ~17% of the total pool of over expressed genes, suggesting that although the degradation lignin is a dynamic process, there is a significant proportion of genes for whose differential abundance levels were sustained over time ( Fig 2A and 2B ). 10.1371/journal.pone.0186440.t001 Table 1 Total genes up and down regulated in each of the three points monitored during growth of SCF1 in lignin-amended cultures. The number in parenthesis indicates the percentage of the genes that were differentially expressed under each time. Expression level Time point Total Early Exponential (EE) Mid-Exponential (ME) Early Stationary (ES) Up regulated 733 1308 442 2483 Down regulated 511 1157 448 2116 Total 1244 (27.9%) 2465 (55.4%) 890 (20%) 4599 10.1371/journal.pone.0186440.g002 Fig 2 Venn diagrams of differentially expressed genes, where (A) green circles indicate number of total genes are up-regulated each time point, and (B) red circles indicate number of total genes were down-regulated each time point. EE indicates cells analyzed in early exponential phase, ME, mid-exponential phase and ES, early stationary phase, where actual times of sampling each cell population is indicated in the methods as well as in Fig 1A . (C) Changes in the gene profile as a result of lignin exposure in early exponential phase, mid-exponential and early stationary phase of growth. In the right side of each chart (blue), number of genes with increased relative abundance, and in the left side (red), number of genes with lower relative abundance. The genes are grouped according to functional class as defined by COG annotation. Genes with differential expression in the presence of lignin were classified under 20 categories according to their annotation function in the genome ( S1 Data ). Genes encoding hypothetical protein and proteins with unknown function (667) had the highest number of genes with greater abundance under lignin-amended growth, followed by genes encoding proteins with transporter activity (436). Genes involved in transcription and RNA metabolism (206), biosynthesis (201) and protein synthesis (180) were also represented in abundance in the differential expression data ( Fig 2C ). Transcripts reflect the progression of activities associated with enhanced batch culture growth in the presence of lignin ( Fig 3 ). Transcript production in response to lignin during early exponential phase growth was characterized by genes associated with biosynthesis and cell division. Catalase/peroxidase transcripts were over-expressed in lignin-amended compared to unamended conditions in every growth phase, but originating from different genes. These results suggest that lignin degradation is a very dynamic process that includes a wide set of genes involved in many cellular functions and takes place during the beginning of growth (early exponential) and the beginning of stationary phase, when the uptake of other substrates has stopped. By early stationary phase, waste products begin to accumulate as 60% of lignin has been degraded, and genes associated with continued expression of catalase/peroxidase enzymes and propanediol utilization related genes were found to be upregulated. The following sections describe in more detail the transcripts characterized by each growth phase. 10.1371/journal.pone.0186440.g003 Fig 3 Clustering of transcripts shows response to lignin in genes of interest in Enterobacter lignolyticus SCF1 grown in glucose with and without lignin. Fold change indicates change in relative abundance of transcript in lignin-amended compared to unamended growth. Colors to the left of the heatmap indicate functional annotations as indicated. Each row represents gene expression of cells grown in lignin and glucose, compared to cells grown in glucose alone at the same time point. Early exponential phase growth in lignin is characterized by accelerated growth and signaling During early exponential phase, many genes controlling growth and biosynthesis of new cell material were up-regulated in the presence of lignin ( Fig 2C , S2 Table ). The large and small subunits ribosomal RNAs, the rod shape-determining operon, the membrane-bound lytic murein transglycosylase, genes involved in the bacterial outer membrane biogenesis, chromosome partitioning and cell division processes were all significantly upregulated when lignin was present compared to lignin unamended conditions. Lignin degradation rates are constrained by the complexity of the substrate, and most bacterial strains regulate their metabolic pathways for utilization of carbon sources by carbon catabolite repression [ 31 ]. This process allows mixed sugars to be utilized only sequentially, and gene regulation attributed to this process was observed in SCF1 growth in the presence of lignin. Many genes related to transporters were found to be down-regulated when glucose was available in the medium, but they were switched on after glucose was no longer metabolized ( S3 Table ). Genes belonging to the phosphoenolpyruvate phosphotransferase system (PTS) related to the uptake of mannose (Entcl_3690, Entcl_3691, Entcl_3692, Entcl_3811 and Entcl_3812), xylose (Entcl_0175 and Entcl_0176) and ribose (Entcl_0174, Entcl_1204, Entcl_1206, Entcl_4081) were down-regulated in early exponential phase. However, after glucose was made unavailable due to concentrations dipping below its half-saturation constant for uptake, they were transcribed in higher abundance. The initial down regulation of genes related to the uptake of other sugars was also observed in the case of transcripts associated with degradation of cellobiose (Entcl_2546 and Entcl_3764), chitobiose (Entcl_2547 and Entcl_2548), maltose (Entcl_3261) and mannose (Entcl_4032, Entcl_4033, Entcl_4034 and Entcl_4036). Furthermore, many genes involved in the uptake of other sugars beside glucose were up-regulated after glucose uptake was stopped ( Fig 2C ) (Entcl_0166, Entcl_0167, Entcl_1205, Entcl_1207, Entcl_3382, Entcl_3383, Entcl_4082, Entcl_4174, and Entcl_4403, S4 Table ). This was presumably a response to the lack of availability of glucose, an effect that has been observed in other fermentative anaerobes [ 32 ]. Enterobacteriaceae is a family of Gram negative bacteria originally described as part of the gut microbiota. As being isolated from Puerto Rico tropical forest soils [ 23 ], SCF-1 may harbor metabolic capabilities toward survival in more oligotrophic environments, including environment-sensing mechanisms coupled to extra- and intra-cellular signal transduction pathways. For instance, SCF-1 has an overrepresentation of genes related to two-component regulatory system compared to those members of the Enterobacter genus that resides in the gut or plant microbiota ( Table 2 ). Among those, transcripts of the three genes encoding the UhpABC system (encoded by the Entcl_0034-Entcl_0037 operon), a two-component response regulator involved in the transport of hexoses as well as the transcripts of hexose phosphate transporter, UhpT (encoded by Entcl_0038), were more abundant in early exponential growth with lignin [ 33 ]. Other components of the two-component regulatory system were also up-regulated, including the sensory histidine kinase QseC (encoded by Entcl_0059), the osmolarity sensory histidine kinase, EnvZ (encoded by Entcl_0331), the sensory histidine kinase, QseC (encoded by Entcl_0729), the hybrid sensory histidine kinase, EvgA (encoded by Entcl_1061), the sensory histidine kinase, AtoS (encoded by Entcl_1498), and the Osmosensitive K+ channel histidine kinase, KdpD (encoded by Entcl_3122). Most of those genes highlight the challenge that SCF-1 faces when is surrounded with environments where lignin is present, even at small concentrations. 10.1371/journal.pone.0186440.t002 Table 2 Summary of genome attributes of Enterobacter lignolyticus SCF1 and three close relatives, E . aerogenes , E . cloacae ATCC13047, and Enterobacter sp. 638. Enterobacter aerogenes Enterobacter cloacae ATCC13047 Enterobacter sp 638 Enterobacter lignolyticus SCF1 Class Gamma-proteobacteria Gamma-proteobacteria Gamma-proteobacteria Gamma-proteobacteria Mode of life gut microbiota /opportunistic bacteria gut microbiota /opportunistic bacteria Endophytic bacteria Tropical soil bacterium Type of metabolism facultative anaerobic facultative anaerobic aerobic facultative aerobe; grows well under completely oxic and anoxic conditions Size of the genome (Mb) 5.27 5.59 4.67 4.81 Genes related to two-component regulatory system 21 20 27 33 Genes related to secretion systems 27 58 13 26 References [ 71 ] [ 71 ] [ 72 ] [ 23 ] Many genes encoding proteins involved in resistance to environmental stresses were up-regulated in the presence of lignin, including the starvation sensing protein RspA (Entcl_1608), the starvation lipoprotein Slp paralog (Entcl_1986) and the universal stress protein family 3 (Entcl_2311). Carbon starvation protein A paralog (Entcl_3779) was also up-regulated in lignin-amended cultures after 7 hours, in agreement with former studies which found cstA to be upregulated at the protein and mRNA levels [ 20 ]. CstA, is a protein involved in escaping from starvation during nutrient scavenging and transition to stationary-phase growth [ 34 ] and has been shown to be regulated by the cyclic AMP (cAMP)-cAMP receptor protein complex, a global regulator involved in sugar metabolism [ 35 ]. Although the specific role of CstA in the metabolism of aromatic compounds remains to be elucidated, it has been demonstrated that carbon starvation proteins are induced by different aromatic pollutants [ 36 ], suggesting its involvement as a mechanism of adaptive response to the presence of a more recalcitrant carbon source [ 37 ]. Similar to cstA , the gene encoding the competence protein F (Entcl_0322) was found to be upregulated in lignin-amended cultures after 7 hours. This protein has also been shown to be part of the physiological response towards nutrient depravation, conferring cells the ability of uptake exogenous DNA and use it as a source of carbon and energy, providing an advantage during competition in nutrient limiting environments [ 38 ]. Within their wide set of tools to survive to dynamic environments, bacteria have evolved a way to respond to chemical or physical changes by modulating their behavior through gene expression. In one of those mechanisms, coined as \"quorum sensing\", bacteria respond to the concentration of a chemical molecule that reflects the bacterial population density allowing multicellular communication within a population [ 39 ]. One autoinducer, called AI-2, has been shown to be produced and recognized by several species allowing both intra and inter-cellular coordination [ 40 ]. We found that many genes related to the lsr operon (Entcl_0612, Entcl_0614, Entcl_0618 and Entcl_0619), the operon responsible for the uptake and processing of AI-2, were up-regulated in the presence of lignin. Mid-exponential samples capture switch towards degradation of lignin The degradation of lignin is a dynamic process, which was evident during mid-exponential phase of growth where lignin amended culture transcripts reflected a decrease in growth rate likely caused by a switch in metabolic growth from only consuming glucose to depolymerization of lignin ( Fig 1 ). During this period, many genes encoding for the large subunits of ribosomal RNAs were in lower abundance compared to the high abundance detected in early exponential phase ( S5 Table ), likely due to the switch towards lignin degradation. Lignin degrading bacteria must overcome three main challenges towards efficient utilization of lignin as substrate. First, the mechanism of degradation should be oxidative [ 41 ] Secondly, since lignin is mainly composed by large highly insoluble molecules, lignin depolymerization enzymatic machinery has to be extracellular, or having an extracellular mediator. Third, the systems must provide a pool of enzymes with low level of specificity since lignin stereochemistry is highly variable [ 41 ]. For the sake of simplification, bacterial lignin degradation could be divided in two stages. Initially, depolymerization releases lignin monomers from the complex lignin structure mainly by the cleavage of ether linkages [ 9 , 35 ]. After monomers are released, enzymes in peripheral pathways proceed with the breakdown of these monomers into common intermediates that are further processed by the central metabolism [ 42 ]. The genome of SCF1 contains several genes encoding for enzymes that have been previously implicated in the degradation of lignin [ 20 , 23 ]. Glycoside hydrolases are common carbohydrate-active enzymes involved in plant biomass degradation, capable of hydrolyzing the glycosidic bonds between carbohydrates or between carbohydrates and lignin [ 43 ]. The genome of SCF-1 has four homologs encoding glycoside hydrolase family 1 β-glucosidases. Three of them (Entcl_0991, Entcl_1274 and Entcl_3004) were expressed in higher abundance during mid-exponential growth of lignin-amended cultures ( Table 3 ). The 6-phospho-beta-glucosidase (Entcl_0991) is located in the same operon as the cellobiose-specific form of the phosphotransferase system component (Entcl_0992), suggesting the existence of an operon devoted to the degradation of components of plant material. Previous studies have shown glycoside hydrolases were abundant in a soil-derived consortia adapted to recalcitrant carbon sources, such as complex plant polysaccharides and lignin, and also highly expressed in degradation of cellulose and hemicellulose by hindgut microbiota associated to termites [ 44 , 45 ]. 10.1371/journal.pone.0186440.t003 Table 3 Genes differentially regulated during growth related to lignin degradation. Differential expression was defined as transcripts with adjusted p-values <0.05 and absolute value of log2 fold change >1 for these comparisons. Gene ID Annotation Gene name Fold change in transcripts EE ME ES Entcl_1274 6-phospho-beta-glucosidase (EC 3.2.1.86) 0.041 0.991 -1.131 Entcl_3004 6-phospho-beta-glucosidase (EC 3.2.1.86) 0.437 2.680 -0.150 Entcl_0991 6-phospho-beta-glucosidase (EC 3.2.1.86) ascB -0.915 1.496 -0.362 Entcl_0992 PTS system, arbutin-, cellobiose-, and salicin-specific IIBC component (EC 2.7.1.69) -0.653 2.291 0.463 Entcl_0735 putative laccase (EC 1.10.3.2) 0.523 -0.921 1.354 Entcl_0736 Probable Fe-S oxidoreductase family 2 1.516 1.037 1.106 Entcl_0748 Predicted oxidoreductases (related to aryl-alcohol dehydrogenases) -0.626 1.684 0.883 Entcl_4301 Catalase (EC 1.11.1.6) / Peroxidase (EC 1.11.1.7) 0.560 3.932 0.146 Entcl_2769 Flavoprotein wrbA -1.502 3.638 0.700 Entcl_3180 Alkyl hydroperoxide reductase protein F (EC 1.6.4.-) 0.186 3.513 1.719 Entcl_3181 Alkyl hydroperoxide reductase protein C (EC 1.6.4.-) -0.010 1.468 1.126 Entcl_3797 4-hydroxyphenylacetate degradation bifunctional isomerase/decarboxylase, HpaG2 subunit 0.111 0.525 1.889 Entcl_3800 5-carboxymethyl-2-hydroxymuconate delta-isomerase (EC 5.3.3.10) 0.384 3.639 1.939 Entcl_3804 Transcriptional activator of 4-hydroxyphenylacetate 3-monooxygenase operon, XylS/AraC family 1.314 2.640 -0.058 Entcl_3805 4-hydroxyphenylacetate 3-monooxygenase (EC 1.14.13.3) 0.000 1.706 -1.529 Entcl_2233 Succinate-semialdehyde dehydrogenase [NADP+] 0.340 1.543 -0.248 Entcl_0876 Succinate-semialdehyde dehydrogenase [NADP+] (EC 1.2.1.16) -1.962 0.231 0.408 Entcl_2810 Gamma-glutamyl-aminobutyraldehyde dehydrogenase (EC 1.2.1.-) -1.292 0.011 0.519 Entcl_2233 Succinate-semialdehyde dehydrogenase [NADP+] 0.340 1.543 -0.248 During the degradation of lignin, many phenolic compounds are release into the medium. Those metabolites including acids (ferulic acid, vanillic acids and 4-hydroxybenzoic acid), alcohols (guaiacol and catechol, etc) and aldehydes (vanillin, 4-hydroxylbenzaldehyde, etc) have been reported to alter the permeability of cellular membranes at even small concentrations [ 46 ]. One of the strategies of microbes towards to cope with those stressful conditions is through the breakdown of phenols to nontoxic compounds. For instance, laccases (phenoloxidases, EC 1.10.3.2) are multicopper-containing enzymes capable of breaking phenols, aromatic and aliphatic C–C bonds by oxidation of phenolic units using molecular oxygen as the final electron acceptor [ 47 , 48 ]. Although fungal laccases have been shown to play a key role during lignin degradation in the white-rot fungi [ 49 – 51 ], little is known about their function in their bacterial counterparts. Recent research has focused on the involvement of laccases during decomposition of lignin by strains of the soil-inhabiting microorganisms Streptomyces , where the deletion of the gene encoding a laccase partially impaired lignin degradation [ 52 ]. Previous evidence has suggested that the genome of SCF1 contains two putative laccases homolgs [ 20 ]. One of those genes, encoding a multicopper oxidase type 3 (Entcl _0735), was found to be up-regulated in the presence of lignin. Its adjacent gene (Entcl_0736) that seems to be co-transcribed with, correspond to an uncharacterized enzyme that contains Radical SAM domain protein using cobalamine (Vitamine B-12) as a cofactor. Several genes involved in the synthesis of cobalamine (Entcl_1738, Entcl_1762, Entcl_1763, Entcl_1764, Entcl_1765, Entcl_1766, Entcl_1767, Entcl_1768, and Entcl_1771) were found to be up-regulated in cultures were lignin was added. Furthermore, both laccases are predicted to be extracellular and periplasmic, respectively by PSORT [ 53 ] suggesting that they may be secreted. Under anaerobic growth conditions, the role laccases during lignin degradation remains questionable, due to the fact that those enzymes employ dioxygen as electron acceptor to carry out single electron oxidations of organic compounds. To this date, it is unclear how laccases could perform oxidation with an alternative electron acceptor. Recent evidence carried out with soil bacterium Stenotrophomonas maltophilia AAP56 found that laccase activity was highly dependent on the environment and it was greatly induced by anoxic conditions. More importantly, laccase activity is often induced by addition of CuSO 4 [ 54 ]. In this example, laccase activity was greatly induced by the addition of 100 μM of CuSO 4 in the γ-Proteobacterium JB, an alkali-tolerant soil isolate [ 55 ]. Therefore, as the trace element solution utilized in the growth medium has some Cu, and our lignin contains some sulfur impurities, we cannot rule out the possibility that laccases were induced by CuSO 4 , and not by lignin. Aryl-alcohol dehydrogenases are enzymes capable of cleaving the β-aryl ether bonds, and have been shown to play a key role during depolymerization of lignin by white-rot fungi [ 56 , 57 ]. As a part of this potential mechanism in SCF-1, we found that the transcripts of an aryl-alcohol dehydrogenases (Entcl_0748) were expressed in higher abundance in lignin-amended cultures ( Table 3 ). Aryl-hydrogenases provide the hydrogen peroxide needed by ligninolytic high redox-potential peroxidases during fungal degradation of lignin [ 57 ]. Peroxidases aid in lignin degradation by removing one electron from the non-phenolic units of lignin [ 58 ]. Indeed, lignin peroxidases (EC 1.11.1.14) utilize hydrogen peroxide (H 2 O 2 ) as the co-substrate in addition to a mediator, veratryl alcohol, to degrade lignin and other phenolic compounds. H 2 O 2 is reduced to H 2 O by gaining an electron from the then oxidized LiP and then LiP release this electron to vetraryl alcohol that becomes aldehyde and gets reduced to its alcohol by gaining electrons from lignin [ 35 ]. The utilization of peroxidases in lignin degradation was previously hypothesized because SCF1 was isolated from the Luquillo LTER soils, where peroxidase activities were detected across a large area [ 20 ]. One peroxidase (encoded by Entcl_4301) and a flavoprotein disulfide reductase (encoded by Entcl_2769), an enzyme capable of sensing the local concentrations of hydrogen peroxide, were found to be up-regulated in the presence of lignin during mid-exponential phase, the period in which 60% of the lignin amended in the medium was utilized. Also, the genes encoding alkyl hydroperoxide reductases (encoded by Entcl_3180 and Entcl_3181), one of the main scavenger of H 2 O 2 inside the cells, were found to be expressed in higher abundance in the presence of lignin, during the same sampling time. Although lignin peroxidases may be important enzymes for lignin degradation in SCF-1, the way in which hydrogen peroxide is provided at a constant rate under anaerobic conditions and its role during the anaerobic degradation of lignin remains unknown. Another proposed mechanism for the degradation of lignin is via 4-hydroxyphenylacetate catabolic pathway. This pathway is composed by nine enzymes encoded by genes within a single genomic region (HpaRGEDFHIXABC) previously described as an aromatic degrading gene cluster. This cluster has been shown to be involved in the transformation of 4-hydroxyphenylacetate, mainly coming from the degradation of aromatic compounds, to intermediates of the central metabolism [ 23 ]. We found the gene encoding for the first gene of this pathway (Entcl_3797) to be up-regulated in the presence of lignin, suggesting that the pathway was turn on. Furthermore, three genes associated to this cluster (Entcl_3800, Entcl_3804 and Entcl_3805) were expressed in higher abundance when lignin was present. This pathway takes 4-hydroxyphenylacetate, a resulting compound from the fermentation of short-chain peptides and amino acids, and degrades it to pyruvate and succinate-semialdehyde [ 59 , 60 ]. Because of its toxicity, succinate-semialdehyde has to be quickly converted to succinate via NADP + -dependent succinate semialdehyde dehydrogenase previous entering into the tricarboxylic acid cycle [ 61 ]. The genome of SCF-1 contains four homologs encoding for this enzyme (Entcl_2233, Entcl_2291, Entcl_0876 and Entcl_2810) and one of them (Entcl_2233) was expressed in higher abundance in lignin-amended cultures. Other genes with a potential role during aromatic metabolism that support lignin degradation were found to be differentially expressed during lignin-amended growth. For instance, genes encoding for two anthranilate synthases (Entcl_2523 and Entcl_2524), a peripheral pathway that play an important role during anaerobic degradation of heterocyclic aromatic compounds, were found to be among top-5 upregulated genes in lignin amended cultures [ 42 ]. Also, genes encoding enzymes that belong to xanthine/uracil family permeases, such as Entcl_0083, Entcl_1586, Entcl_4435, Entcl_3516, Entcl_3019 and genes encoding for enzymes that catalyzes the further metabolization of xanthine, such as Entcl_4100, Entcl_3480, Entcl_2780, Entcl_2781, Entcl_2782, and Entcl_2795, were found to be up-regulated during lignin-amended mid-exponential growth. This set of enzymes belong to a mechanism of recycling of uric acid, xanthine and other nitrogenous products that has been previously reported to be relevant in the gut microbiome of wood-feeding beetles [ 62 ]. Lignin amendment was also associated with higher abundance of transcripts for genes associated with the uptake and metabolism of sugars. For example, the gene encoding for glucokinase (Entcl_1357), an enzyme that catalyzes the first step in glycolysis (reaction #1, S1 Fig ), was found to be upregulated during mid-exponential growth, The expression of genes encoding for glucose-6-phosphate isomerase (Entcl_3693 and Entcl_3694) (reaction #2, S1 Fig ), 6-phosphofructokinase class II (Entcl_2086 and Entcl_4346) (reaction #3, S1 Fig ) and fructose-1,6-bisphosphatase, GlpX type (Entcl_0096) (reaction #4, S1 Fig ) were also found to be expressed in higher abundance in early stationary phase. Similar behavior was observed in the genes encoding for the steps downstream in glycolysis, such as fructose-bisphosphate aldolase, class II (encoded by Entcl_4084) (reaction #5, S1 Fig ), and the NAD-dependent glyceraldehyde-3-phosphate dehydrogenase (encoded by Entcl_2022). Lignin-amended cultures produce very different metabolites by early-exponential phase E . lignolyticus SCF1, as other enterobacteria, primarily carry out heteroenzymatic fermentation using pentose and hexose sugars as carbon and energy sources, and producing acetate, ethanol, formate and lactate, as main fermentation products [ 23 ]. A high concentration level of glucose remained in the medium after SCF-1 stopped growing, 45 mM in the presence of lignin and 40 mM in the control. The inability to keep metabolizing the substrates can be explained by the increased formation of acidic metabolites, such as acetate, formate, lactate, and succinate, which destroys the cell's ability to maintain internal pH [ 63 ]. Lignin amendment resulted in a significant difference in the production of fermentation subproducts, which may be attributed to the higher quantity of C atoms that have to be funneled into common intermediates of the central metabolism (pyruvate and Acetyl-CoA). For instance, the pyruvate formate-lyase, an enzyme in charge of supplying the citric acid cycle with acetyl-CoA during anaerobic glycolysis (encoded by Entcl_0584, Entcl_2993, Entcl_4043, Entcl_4296, Entcl_4297) and also its activating enzyme (encoded by Entcl_2992, Entcl_4042, and Entcl_4288; S6 Table ) were expressed in higher abundance when lignin degradation rate was at its maximum. There was about 5 mM of more formate that was released by the cells in lignin-amended cultures ( Fig 1C ), though most of this accumulated in the first 15 hours of growth. Many genes belonging to metabolic pathways devoted to degrade other components of plant material were also regulated under lignin amendments. For instance, several genes encoding enzymes involved in the degradation of pectin, a structural polysaccharide that is key for the stability of plant cell walls, were found to be upregulated ( S7 Table ). A considerable high number of genes (48) remained upregulated across the three growth phases underlying the idea that some of those genes may represent a set of metabolic functions that have to be constitutively expressed in the presence of lignin, rather than a part of a progressive metabolic or biological mechanism towards degradation of lignin ( S8 Table ). Because of the increased formate production and transcription of formate hydrogenlyase (FHL) complex transcripts with lignin, we hypothesized that SCF-1 would produce hydrogen as an alternative to regulate pH in the lignin-amended treatments. Many members of Enterobacteriaceae , such as E . coli and Salmonella enterica , contain a membrane-associated FHL complex capable to utilize formic acid as electron donor and protons as electron acceptors during the production of H 2 and CO 2 [ 64 , 65 ]. This complex allows the accumulation of the excess of reducing equivalents produced by fermentation into hydrogen, preventing further acidification of the medium and promoting interspecies H 2 transfer. The FHL complex is composed of two main parts that catalyze hydrogen production from formate in two steps. The molybdenum-dependent formate dehydrogenase (FDHH) encoded by fdhF gene (Entcl_4087) catalyzes the reversible oxidation of formate to CO 2 with the release of protons and two electrons. A second step is catalyzed by the NiFe hydrogenase, Hyd3, encoded by Entcl_0986, that utilize these electrons to further reduce protons to H 2 [ 66 ]. Hyd3, belongs to the Group 4 [NiFe] hydrogenases, a phylogenetically distinct family of membrane-bound, cytoplasmically oriented hydrogenases that so far has been rarely characterized, and are predominantly involved in H 2 production rather than H 2 oxidation [ 67 , 68 ]. Hyd3 differs from other hydrogenases because it allows accumulation of significant concentrations H 2 , without inhibition, making SCF-1 especially well-suited to biotechnological H 2 production. Cell suspensions were capable of anaerobically producing H 2 in the presence of formate (150 mM) and lignin (0.05% w/w) ( S4 Fig ), which is evidence suggesting that the FHL complex is functional in E . lignolyticus SCF-1. However, during hydrogen production in cell suspensions, ~30% less H 2 was produced in those incubations where lignin was added compared to formate added alone. In the RNA-seq results, the two components of formate hydrogenlyase complex, the formate dehydrogenase H (Entcl_4087), the seven subunits of the formate hydrogenlyase complex (Entcl_0982-Entcl_0988), its transcriptional activator (Entcl_0976) and the neighbor NiFe hydrogenase assembly enzymes (Entcl_0977-Entcl_0982) were down-regulated during the late stages of growth. On the other hand, genes belonging to the formate dehydrogenase type N (Entcl_2347–2349), an enzyme that generally oxidizes formate taking electrons towards dissimilatory nitrate reduction, were expressed in higher abundance when lignin was present after mid-exponential growth. We also observed a higher concentration of acetate in those amended cultures ( Fig 1C ). This overproduction could be attributed to the higher expression of Phosphate acetyltransferase (Entcl_1432) (reaction number 21, S1 Fig ) and acetate kinase (Entcl_0583 and Entcl_1433) (reaction number 22, S1 Fig ) in the earlier stages of growth."
} | 10,358 |
36187553 | PMC9512660 | pmc | 1,159 | {
"abstract": "Invasive species have profound negative impacts on native ranges. Unraveling the mechanisms employed by invasive plant species is crucial to controlling invasions. One important approach that invasive plants use to outcompete native plants is to disrupt mutualistic interactions between native roots and mycorrhizal fungi. However, it remains unclear how differences in the competitive ability of invasive plants affect native plant associations with mycorrhizae. Here, we examined how a native plant, Xanthium strumarium , responds to invasive plants that differed in competitive abilities (i.e., as represented by aboveground plant biomass) by measuring changes in root nitrogen concentration (root nutrient acquisition) and mycorrhizal colonization rate. We found that both root nitrogen concentration and mycorrhizal colonization rate in the native plant were reduced by invasive plants. The change in mycorrhizal colonization rate of the native plant was negatively correlated with both aboveground plant biomass of the invasive plants and the change in aboveground plant biomass of the native plant in monocultures relative to mixed plantings. In contrast, the change in root nitrogen concentration of the native plant was positively correlated with aboveground plant biomass of the invasive plants and the change in aboveground plant biomass of the native plant. When we compared the changes in mycorrhizal colonization rate and root nitrogen concentration in the native plant grown in monocultures with those of native plants grown with invasive plants, we observed a significant tradeoff. Our study shows that invasive plants can suppress native plants by reducing root nutrient acquisition rather than by disrupting symbiotic mycorrhizal associations, a novel finding likely attributable to a low dependence of the native plant on mycorrhizal fungi.",
"conclusion": "5 Conclusions In this study, we examined how nutrient acquisition strategies of native plants responded to size differences in invasive plants. We found that invasive plants suppressed root nutrient acquisition rather than mycorrhizal colonization of native plants. To our knowledge, we are the first to show a tradeoff between the changes in root nutrient acquisition and mycorrhizal colonization rate in a native plant in response to exotic plant invasion. This tradeoff suggests that larger invasive plants can outcompete native plants through their greater influence on the root nutrient acquisition in native plants, whereas smaller invasive plants likely lean toward allelopathic effects on mutualistic mycorrhizal interactions. Finally, it should be noted that this study was conducted in the temperate zone where plants are usually not limited by soil phosphorus (P). Future studies could test generalizability of this study by selecting more invasive and native species across sites differing in soil P content and plant dependence on mycorrhizal fungi.",
"introduction": "1 Introduction Biological invasions greatly reduce biodiversity globally and have substantial economic costs ( Van Kleunen et al., 2015 ; Pyšek et al., 2017 ; Pathak et al., 2019 ; Bellard et al., 2021 ; Rodriguez et al., 2021 ). Understanding the mechanisms by which invasive plants outcompete their native counterparts and spread through new ranges may help control invasions. Invasive plant species often outcompete native plants because invasive plants have a high capacity for resource acquisition. For example, invasive plants usually shift leaf nitrogen (N) allocated for defense to photosynthesis for a greater competitive advantage over other plants ( Feng et al., 2009 ; Miao et al., 2014 ; Battini et al., 2021 ; Harkin and Stewart, 2021 ; Lin et al., 2021 ; Wang et al., 2022 ). In addition, invasive plants are increasingly reported to suppress native plants by disrupting symbiotic associations between native plants and other organisms, such as mycorrhizal fungi ( Callaway and Ridenour, 2004 ; Pringle et al., 2009 ; Wilson et al., 2012 ; Grove et al., 2017 ; Liu et al., 2020 ). Most terrestrial plants form symbiotic relationships with mycorrhizal fungi for nutrient acquisition ( Kohler et al., 2015 ; Martin et al., 2017 ; Steidinger et al., 2019 ; Tedersoo et al., 2020 ; Meng et al., 2021 ; Shi et al., 2021 ). In this symbiosis, plants help mycorrhizal fungi meet energy requirements by supplying carbon; in return, the fungi provide nutrients to plants through their enormous extrametrical hyphae ( Bonfante and Genre, 2010 ; Delavaux et al., 2017 ; Jiang et al., 2017 ; Bisseling and Geurts, 2020 ; Delaux and Schornack, 2021 ). A previous study has reported that invasive plants can suppress native plants by reducing associations between the roots of native plants and mycorrhizal fungi ( Stinson et al., 2006 ). This is an important mechanism by which invasive plants gain a foothold into new ranges, and it has received a great deal of attention ( Vogelsang and Bever, 2009 ; Lau and Schultheis, 2015 ; Dickie et al., 2017 ; He et al., 2018 ; Yu and He, 2021 ). Despite the importance of this mechanism, it is still yet fully understood ( Jordan et al., 2011 ; Gaggini et al., 2018 ). For example, mounting evidence has shown a complementary relationship between roots and mycorrhizal fungi in nutrient acquisition ( Chen et al., 2016 ; Duchene et al., 2017 ; Brundrett and Tedersoo, 2018 ; Fabianska et al., 2019 ; Tao et al., 2019 ). It is likely that native plants can increase root activity when their associations with mycorrhizal fungi are suppressed by invasive plants. However, we know little about whether invasive species with different competitive abilities differ in their influence on mycorrhizal association of the native roots. In this study, we aimed to test whether larger invasive plants are more effective than smaller plants at disrupting mutualistic associations between a native plant and mycorrhizae. We also tested whether native plants counteract reductions in mycorrhizal associations by investing more in root nutrient acquisition. For this purpose, we examined how nutrient acquisition strategies changed in a native plant, Xanthium strumarium L., in response to five invasive plant species of different sizes. Specifically, we measured changes in root N concentration and mycorrhizal colonization rates in X. strumarium grown with five invasive plant species given that the two traits are closely related to nutrient acquisition by roots ( Reich et al., 2008 ; Kong and Fridley, 2019 ) and mycorrhizal fungi ( Kong et al., 2014 ; Bergmann et al., 2020 ), respectively.",
"discussion": "4 Discussion Plants acquire soil nutrients through the root and mutualistic associations with mycorrhizae. Invasive plants have been previously shown to suppress native plant growth by reducing native plant associations with mycorrhizae ( Stinson et al., 2006 ; Tang et al., 2020 ). In this study, we found that invasive plants inhibit nutrient acquisition of a native plant, Xanthium strumarium , by interfering with roots more than by disrupting mycorrhizal interactions. In our experiments, aboveground biomass of the native species was negatively correlated with the change of root mycorrhizal colonization rate, but positively correlated with changes in root N concentration. These findings differ greatly from previous studies that invasive plants suppress native plant growth via a reduction in mycorrhizal interactions ( Nasto et al., 2017 ; Ostonen et al., 2017 ; Png et al., 2017 ; Freund et al., 2018 ; Clausing et al., 2021 ; de Vries et al., 2021 ). The native plant used in our study ( X. strumarium ) has a low mycorrhizal colonization rate (about 16%), and relies mostly on the root for nutrient acquisition. Thus, our contrasting results may be partially explained by the low dependence of X. strumarium on mutualistic mycorrhizal interactions. Interestingly, we found that the effect of each of the five invasive species on root nutrient acquisition and mycorrhizal colonization of the native species varied. For example, although invasive plants reduced both root nutrient acquisition and mycorrhizal colonization rate in the native plant, the extent to which invasive plants inhibited mycorrhizal colonization rate was negatively correlated with the size of invasive plants. Conversely, the extent to which invasive plants inhibited root nutrient acquisition was positively correlated with the size of invasive plants ( Fig. 2 ). These findings do not support our original hypothesis that larger invasive plants cause greater reductions in native plant mycorrhizal colonization rate. The rationale for our initial hypothesis was that larger invasive plants are usually able to outcompete native plants for resources; this great competitive capacity of larger invasive plants reduces native plant growth, causing a reduction in carbon supply, which lowers the ability of native plants to maintain mycorrhizal associations. However, larger invasive plants in our study caused a smaller rather than a larger reduction in mycorrhizal colonization in the native plants. The unexpected result of our study could be due to interactions between root nutrient acquisition and mutualistic mycorrhizal interactions, which are generally reported to be complementary ( Nasto et al., 2017 ; Png et al., 2017 ; Soper et al., 2018 ; Qiu et al., 2021 ). For example, plants that rely more on roots for nutrient acquisition will reduce their dependence on mycorrhizal fungi ( Liu et al., 2015a ; Chen et al., 2016 ; Bialic-Murphy et al., 2021 ). Although no such tradeoff was found in our study in the native plant ( Fig. 4 a), we did observe a tradeoff between root nutrient acquisition and mycorrhizal interactions when we compared changes in root N concentration and mycorrhizal colonization rate of the native plant in monocultures with those of native plants grown with invasive species ( Fig. 4 b). Specifically, a larger invasive plant can cause a greater reduction in root N concentration but only a small reduction in mycorrhizal colonization rate in the native plant. We speculate that the smaller reduction in mycorrhizal colonization rate by larger invasive plants may be beneficial to the native plant. For example, despite a relatively weak dependence of the native plant on mycorrhizal interactions, the smaller reduction in mycorrhizal colonization may partially meet nutrient demand of the native plant when nutrient acquisition by the root is greatly suppressed by larger invasive plants. In contrast, when the native plant was grown with a smaller invasive plant, invasive plants suppressed mycorrhizal colonization ( Fig. 2 a) more than they suppressed root nutrient acquisition ( Fig. 2 b). It is likely that these smaller invasive plants interfere with mycorrhizal colonization through allelopathy ( Kato-Noguchi, 2020 ; Kalisz et al., 2021 ; Zhu et al., 2021 ). If so, the allelopathic effects of smaller invasive plants warrant further investigation."
} | 2,744 |
28989680 | PMC5627351 | pmc | 1,160 | {
"abstract": "A chemical approach for the regulation of oil (under water) and water (in air) wettability. The super-wetting properties are highly durable at harsh physical/chemical settings.",
"conclusion": "Conclusions Here, we have developed a multilayer construction by strategically exploiting a simple covalent-LbL deposition technique in combination with a robust 1,4-conjugate addition reaction between acrylate and amine groups for the controlled and extreme regulation of water and oil wettability both in air and under water respectively. The growth of multilayers of the NCs (that formed when mixing BPEI and 5Acl in methanol) was comparatively rapid and exponential in the presence of salt, and provided appropriate topography to display extreme liquid wettability properties including superhydrophobicity/superhydrophilicity in air and superoleophobicity/superoleophilicity under water. Moreover, consecutive LbL deposition allowed facile control over the topography of the multilayer coatings, to adopt special adhesive anti-fouling properties through control of the fraction of the contact area of the solid with the beaded oil droplets. On the other hand, residual acrylate groups in the multilayers provided a convenient basis for the precise control of the fraction of the solid contact area with the beaded liquid phases both under water and in air through suitable and appropriate post chemical modification of the multilayers of the NCs with small molecules via the 1,4-conjugate addition reaction, and the appropriately post-functionalized multilayers of the NCs were eventually embedded with various special water (in air) and oil (under water) wettability properties. The current design allowed a comprehensive and comparative investigation on the homogeneous/heterogeneous wettability of both water and oil in air and under water respectively using a single material—that is the ‘reactive’ and covalent multilayer coatings of the NCs. The extreme and special wettability of both the water (in air) and the oil phases (under water) was successfully adopted by tailoring the fraction of the solid contact area with the beaded liquid droplets through both the change in chemical functionality and topography of the material. However, the same extent of change in the chemistry and topography in the material has an independent implication on the water (in air) and oil (under water) wettability. Moreover, the synthesized multilayers of the NCs were able to sustain various physical and chemical insults without compromising the anti-fouling property of the materials, and the current strategy also provided a facile avenue to decorate a wide variety of substrates with the desired anti-fouling properties. Thus, this material has the ability to be general and would be useful in various relevant applications including the synthesis of patterned interfaces through site-specific functionalization of the ‘reactive’ multilayers of the NCs, and could be useful in various advanced applications. 67",
"introduction": "Introduction Bio-inspired heterogeneous wettability of liquids (oil/water) 1 – 5 on solid surfaces is instrumental in developing various advanced materials that are useful in widespread applications such as oil/water separation, drug delivery, protein crystallization, underwater robotics, water harvesting, open microfluidics etc. \n 6 – 20 The Cassie–Baxter model explains this extreme repellency of different liquids (oil and water) on solid surfaces by hypothesizing the presence of another phase (either air or water) on the solid surface, 21 , 22 which is responsible for the heterogeneous wettability of the liquids (oil/water). As an example, the metastable trapped air in the lotus leaf makes the solid surface extremely water repellent, where the water droplets bead with an advancing water contact angle (WCA) above 150°, and the droplets readily roll off on tilting the surface below 10°. Later, this anti-fouling property was recognized as superhydrophobicity. 23 , 24 Several top down and bottom up approaches 25 , 26 have been introduced in the literature to fabricate artificial superhydrophobic surfaces by following a general principle, where (1) hydrophilic building blocks are exploited following various processes to adopt the essential hierarchical topography (appropriate combination of micro/nano features)—which is (2) again coated with inert and low surface energy chemicals (composed of either fluorinated molecules or those with a long hydrocarbon tail), generally through a chemical vapor deposition process. However, the lack of a strong interaction between the hydrophilic hierarchical structures and inert low surface energy coating makes the material inherently fragile, and so it is susceptible to the loss of the anti-wetting property on exposure to harsh physical (bending, twisting, scratching etc. ) or chemical (extremes of pH, salt etc. ) insults. 27 – 31 Recently, a few elegant approaches have been introduced to develop durable superhydrophobic materials, yet the synthesis of durable superhydrophobic surfaces still remains a challenging and interesting research topic for fundamental and applied contexts. 32 – 37 \n In 2009, a seminal report by Jiang and coworkers revealed the existence of another anti-fouling property of fish scales—which display extreme oil-repellency under water with an advancing oil contact angle (OCA) above 150° and a contact angle hysteresis below 10°. 3 Early demonstrations related to this anti-oil-fouling property revealed its prospective potential in the separation of oil/water emulsions, underwater robotics, anti-biofouling coatings, and in developing smart microfluidic devices etc. \n 7 , 9 – 11 , 14 , 18 Generally, metal oxides, 38 – 41 polymeric gels 3 , 14 , 18 , 42 – 45 and electrostatic multilayers 46 , 47 have been explored in developing appropriate topography and surface chemistry—which confer the desired heterogeneous oil-wettability under water through the impregnation and confinement of a water layer within the hierarchical topography of the coatings. These polymer gel, electrostatic multilayer and metal oxide approaches are useful for understanding the fundamentals related to this anti-oil-fouling property, but most often, these synthesized materials are liable to lose the embedded artificial anti-fouling property on exposing the materials to either harsh physical insults and/or complex chemical environments as (1) polymeric hydrogels are highly susceptible to easy deformation under regular physical/chemical manipulations and (2) metal oxide and electrostatic multilayers are known to corrode away under harsh chemical treatments. 48 \n Nevertheless, these two completely distinct anti-fouling properties are undifferentiated in terms of the mode of wettability—a common heterogeneous (Cassie–Baxter state) liquid wettability is encountered in the respective materials, however, the entrapped third media are unambiguously different: (1) air for superhydrophobicity and (2) water for superoleophobicity under water. 2 , 3 , 21 , 22 As a consequence, the general requirements to stabilize the respective trapped media (air/water) in the respective materials are significantly different: (a) a hierarchical topography with a low surface energy is the basis to achieve metastable trapped air for superhydrophobicity, 1 , 2 , 23 whereas (b) superhydrophilic materials in air with appropriate micro/nano features are efficient for displaying extreme oil repellency underwater. 3 – 5 Here, in this article, we have introduced a single material that displays multiple durable super/special liquid wettability properties including superhydrophobicity in air and both superoleophobicity and superoleophilicity underwater through facile and precise control over the fraction of the solid contact area with beaded liquid phases both in air and under oil. In our current design, a facile and robust 1,4-conjugate addition reaction between acrylate and amine groups is strategically exploited in the rapid construction of chemically ‘reactive’ thick coatings with the desired topography and chemical functionalities—which are essential parameters for adopting the desired special/super-liquid wettability in the synthesized material. First, amine ‘reactive’ nano-complexes (NCs) were synthesized from a BPEI/5Acl mixture, and constituted a multilayer coating by following a salt-assisted covalent and rapid layer-by-layer (LbL) deposition of the NCs in combination with the BPEI polymer. Then, the essential surface chemistry was adopted in the coating by strategic post chemical modification of the residual acrylate moieties in the multilayers of the NCs. Furthermore, the synthesized multilayers (9 bilayers of NC/BPEI) were explored for the gradual and extreme tailoring of both water wettability in air and oil wettability under water by controlling the fraction of the solid contact area with the respective beaded liquid phases through a facile change in both the chemical functionality and the topography of the material. Moreover, the materials displayed impeccable physical and chemical durability, and were able to withstand various kinds of insults without compromising the super-wetting properties of the materials. Furthermore, the current approach allows the decoration of various rigid/flexible substrates including wood, fabric, Al-foil etc. with desired super-anti-wetting properties. An example of such a single material—which is capable of coating various objects with desired physically/chemically durable super/special-liquid (water and oil) wettability properties both in air and under water—is unprecedented in the literature.",
"discussion": "Results & discussion Fabrication of multilayers of the nano-complexes and post chemical modification In the past, the 1,4-conjugate addition reaction was strategically exploited in both the chemical modification (dendritic amplification of functional groups or selective modification of polymeric microstructures) 49 , 50 and the development of complex nanostructures (multilayer coatings and porous and moldable gels). 51 , 52 In relevance to our current design, in 2012, Lynn and Bechler explored this 1,4-conjugate addition reaction in the fabrication of a reactive and covalent multilayer, which was constructed by a mutual reaction between a polymer (branched polyethyleneimine, BPEI) and small molecule (dipentaerythritol penta-acrylate, 5Acl), and after 80 consecutive BPEI/5Acl layer depositions and appropriate post chemical modification, the material became only hydrophobic with a water contact angle ∼138°. 52 Recently, we have reported the synthesis of a ‘reactive’ and moldable porous polymer gel by exploiting the facile Michael addition reaction 49 – 52 between the acrylate and amine groups of BPEI and 5Acl respectively, to develop a self-standing and three dimensional superhydrophobic shapeable monolith, where the polymeric gel material was formed via formation of the nano-complex (NC), and the formation of the gel was expedited in the presence of salt (NaCl). 53 In this current study, a stable nano-complex (NC) solution was first prepared by mixing BPEI/5Acl in methanol ( Scheme 1B ) both in the presence and absence of salt (0.5 mg ml –1 ), which was later strategically embedded in the multilayer construction ( Scheme 1D ) in combination with the amine containing branched polymer (BPEI) by adopting a covalent LbL deposition process ( Scheme 1C ) through the 1,4-conjugate addition reaction ( Scheme 1A ) between the residual acrylate groups (in the NCs) and amine groups (in the BPEI polymer). During the study of the LbL deposition process, with an increasing number of LbL deposition cycles, an expedited growth in the size of the NCs was observed with the dipping solution of a BPEI/5Acl mixture which was doped with NaCl salt (0.5 mg ml –1 ), as confirmed by a DLS study as shown in Fig. 1A . Later, the multilayer growth of the NCs ( Scheme 1D ) that were prepared both in the presence and absence of salt was monitored by measuring the thickness of the multilayers at regular intervals as shown in Fig. 1B . The growth of the multilayer of the NCs in the presence of salt was noticed to be exponential and yielded a thick (3.15 μm) multilayer coating after 9 bilayer (a pair of NC and BPEI layers are together designated as a single bilayer) depositions. However, the growth of the multilayers of the NCs that were prepared in the absence of salt was noticed to be sluggish and the thickness of the coating (9 bilayers of NC/BPEI) was measured to be only 830 nm. In comparison, the multilayers of the polymer that were directly constructed from BPEI and 5Acl solutions provided a coating with 700 nm thickness after 80 BPEI/5Acl layer depositions. 50 Thus, the strategic and rapid incorporation of the same components (BPEI and 5Acl) in the form of NCs in the covalent multilayers through the 1,4-conjugate addition 47 – 50 reaction allowed for high throughput synthesis of a thick and reactive polymeric coating. Scheme 1 (A) Schematic representation of the 1,4-conjugate addition reaction (Michael addition) between primary amine and acrylate moieties. (B) Structural formula of dipentaerythritol penta-acrylate (5Acl, top panel) and branched polyethyleneimine (BPEI, bottom panel). (C) Illustrating the formation of the ‘reactive’ nano-complexes (NCs) on mixing of BPEI and 5Acl in methanol, and their multilayer construction through the covalent layer-by-layer (LbL) deposition process in combination with the solution of BPEI. (D) Schematic representation of the multilayers of the NCs with residual acrylate groups. (E and F) Strategic post-chemical modifications of the multilayers of the NCs with appropriate small molecules (octadecylamine and glucamine) would provide extremes of liquid wettability both in air (superhydrophobicity, (E)) and under water (superoleophobicity, (F)). Fig. 1 (A) Illustrating the growth (size) of the nano-complexes (NCs) in the presence (black line, 0.5 mg ml –1 ) and absence (red line) of NaCl salt in the LbL dipping solution of a BPEI/NC mixture during the construction of the multilayers. (B) A comparison of the growth (thickness) of the multilayers of the NCs with LbL deposition cycles both in the presence (black line) and absence (red line) of salt in the NC solution (one of the dipping solutions in the LbL deposition process). (C–H) Field emission scanning electron microscope (FESEM) images of the multilayers (9 bilayers) of the NCs in low ((C and F); scale bar = 20 μm), medium ((D and G); scale bar = 15 μm) and high ((E and H); scale bar = 500 nm) magnifications, which were constructed in the presence (C–E) and absence (F–H) of salt in BPEI/5Acl mixtures. (I) Fourier transform infrared (FTIR) spectra accounting for the chemical functionality in the multilayers of the NCs (that were constructed in the presence of salt) before (black line) and after post chemical modifications of the multilayers with strategically selected hydrophilic (glucamine, red line) and hydrophobic (octadecylamine, blue line) small molecules. Next, the topography of the coatings was examined with field emission scanning electron microscope (FESEM) imaging. At low magnification, the multilayers of the NCs that were fabricated in the presence of salt were found to have a prominent and random microstructure all around the coatings as shown in Fig. 1C and D , whereas, in the absence of salt, such dominated microstructures were not observed in the multilayers of the NCs, and rather, comparatively premature and smaller micro-domains appeared as evident from the images in Fig. 1F and G . These micro-domains in both of these multilayer coatings (that are prepared in the presence and absence of salt) are developed due to random arrangements of granular NC structures as evident from the FESEM images at higher magnification in Fig. 1E and H . As expected, the size of the granules are comparatively smaller in the multilayers of the NCs that are prepared in the absence of salt as shown in Fig. 1H , which is consistent with the DLS study in Fig. 1A where the growth of the NCs was observed to be rapid in the presence of salt, however the exact role of salt behind this expedited growth of NCs is not understood yet. This accelerated growth of the NCs and the difference in the size of the granular structures would have eventually influenced the growth of the micro-domains in the respective multilayers. These multilayers of the NCs were further characterized by Fourier transform infrared (FTIR) spectroscopy, to investigate the available chemical functionalities in the multilayer construction. A characteristic IR peak (black curve in Fig. 1I ) at 1413 cm –1 for the symmetric deformation of the C–H bond for the β carbon of the vinyl group was noticed in the multilayer of the NCs that was synthesized through consecutive LbL deposition of BPEI and NCs in the presence of salt. This IR signature confirmed the presence of residual acrylate groups in the multilayer of the NCs, and another IR peak at 1736 cm –1 revealed the presence of an ester carbonyl stretching in the multilayer construction. 49 , 50 Based on the past reports on the Michael addition reaction, 47 – 50 this carbonyl stretching most likely appeared due to β-amino ester-type cross-links in the multilayers through the repetitive 1,4-conjugate addition reaction between the amine and acrylate from BPEI and 5Acl respectively. Furthermore, the residual acrylate groups in the multilayers are suitable for post reaction with primary amine containing small molecules (glucamine and octadecylamine) through a 1,4-conjugate addition reaction, irrespective of their chemical functionalities (hydrophilic or hydrophobic). The successful post chemical modifications of the multilayers of the NCs with primary amine containing small molecules were analyzed by further FTIR study, where the characteristic peak at 1413 cm –1 for the residual acrylate groups was noticed to deplete significantly after post chemical modification as shown in Fig. 1I (the red and blue curves are for glucamine and ODA treatments, respectively), and all spectra are normalized with respect to the ester carbonyl functionality (at 1736 cm –1 ) which is likely to remain unperturbed during the post chemical modification process. A very similar result was noticed with the multilayers of the NCs that are prepared in the absence of salt (see the ESI for more details, Fig. S2 † ). Characterization of extreme liquid wettability in the multilayers of the NCs The synthesized multilayers (9 bilayers) of the NCs, which are prepared in the presence and absence of NaCl salt, are noticed to be moderately hydrophobic in air with a CA of ∼92° and 86° ( Fig. 2A and D ) respectively. However, after post chemical modification with ODA molecules (having a long hydrocarbon tail), the multilayer of the NCs which was constructed in the presence of salt became superhydrophobic with an advancing WCA of ∼167° ( Fig. 2B ) and with a roll-off angle of 3° ( Fig. 2K–N and S4G–J † ). Moreover, a jet of water was noticed to immediately bounce away from the superhydrophobic surface as shown in Fig. 2O . In contrast, the hydrophobicity of the multilayer of the NCs that was prepared in the absence of salt was improved slightly and the water droplet had a WCA of ∼104° after post chemical modification with the same ODA molecules as shown in Fig. 2E . Furthermore, with a gradual increase in the LbL deposition cycles from 2 to 9 bilayers, the change in water wettability in the post-modified (ODA) multilayers of the NCs (both in the presence and absence of salt) and BPEI polymer was examined in detail. The water repellency of the post modified (ODA treated) multilayer of the NCs was gradually increased with the number of LbL deposition cycles, the WCA was increased from 88° (2 bilayers) to 112° (7 bilayers), and suddenly, after 8 bilayer depositions, the hydrophobicity of the multilayer of the NCs was significantly improved with a static water contact angle ( θ \n stat ) of ∼148°, and the multilayer of the NCs was embedded with superhydrophobicity ( θ \n stat ∼153°) after 9 bilayer depositions (black curve in Fig. 2J ). However, the change in the WCA of the post modified multilayers of the NCs that were prepared in the absence of salt was noticed to be completely different, and the WCA was measured to be 104° even after 9 bilayer depositions (red curve in Fig. 2J ). In another control study, the multilayer (9 bilayers) of BPEI polymer that was prepared in the presence of salt was found to be moderately hydrophobic with a θ \n stat ∼81° after post chemical modification with ODA molecules as shown in Fig. 2H . Fig. 2 (A–I) Contact angle measurements (A and B, D and E, and G and H) and digital images (C, F and I) of beaded water droplets (the red color aids visual inspection) on both the multilayers of the NCs (that were constructed in the presence (A–C) and absence (D–F) of salt) and BPEI polymer ((G–I); which was built in the presence of salt) before (A, D and G) and after (B and C, E and F, H and I) post chemical modification with ODA molecules. (J) A plot showing the change in the static water contact angle (WCA) of the beaded water droplets (4.5 μL) on ODA-treated multilayers of the NCs (built in the presence (black line) and absence (red line) of NaCl salt) and BPEI (in the presence of salt, blue line) in air with an increase in the number of bilayer deposition cycles. (K–N) Contact angle images showing the rolling of a water droplet (5 μL) on an ODA-treated multilayer of the NCs (which was synthesized in the presence of salt) that was tilted at 3° in air. (O) Digital image of a bounced water jet on the superhydrophobic surface. On the other hand, the strategic post modification of the same multilayer (9 bilayers) of the NCs (that was fabricated in the presence of salt, and having an inherent static oil contact angle (OCA) ∼31°, Fig. 3A ) with hydrophilic small molecules ( i.e. ; glucamine) yielded a completely different anti-fouling property, which is recognized as underwater superoleophobicity, where the oil droplet beaded with an advancing oil contact angle of ∼169° underwater as shown in Fig. 3B , and the oil droplet rolled off on tilting the surface at 4° (Fig. S4K–O † ). Moreover, an 11 μL DCM (model oil) droplet was observed to bounce on the surface as shown in Fig. 3K–O . In comparison, the multilayer (9 bilayers) of the NCs (having an inherent static OCA ∼36° ( Fig. 3D )) which was built in the absence of salt was observed to not display extreme oil-repellency under water even after post chemical modification with the same glucamine molecules ( Fig. 3E ). The oil droplet was beaded on the modified multilayer with a static OCA ∼140° ( Fig. 3E ). Furthermore, the effect of the gradual increase in the number of bilayers on the oil-wettability from 2 to 9 bilayers was examined in the post modified (glucamine) multilayers of both NCs (that are fabricated in the presence/absence of salt) and polymer (BPEI) ( Fig. 3J ). The change in oil-wettability was observed to be more sluggish for the post modified (glucamine) multilayers of the polymer (compared to BPEI/5Acl)—the oil wettability was increased by only ∼13° ( Fig. 3J , blue curve) on increasing the bilayer deposition from 2 bilayers ( θ \n stat ∼106°) to 9 bilayers ( θ \n stat ∼119°; Fig. 3H and I ), and the static oil contact angle was similarly increased (by ∼15°) from ∼125° (2 bilayers) to ∼140° (9 bilayers) in the post modified multilayers of the NCs that were prepared in the absence of salt. However, the change in oil-wettability was noticed to be completely different in the modified multilayers of the NCs that were constructed in the presence of salt, and the static oil contact angle was increased significantly from ∼125° (2 bilayers) to ∼162° (9 bilayers) as shown in Fig. 3J (black curve). These significant changes in the pattern of liquid (water/oil) wettability in air and under water most likely arise from the difference in topography of the multilayers of the NCs that were fabricated in the presence and absence of NaCl salt. The change in micro-domains in the multilayers was already discussed in the previous section and the morphology of the multilayers of BPEI under FESEM was observed to be mostly featureless as shown in Fig. S3E and F. † \n Fig. 3 (A–I) Contact angle images (A and B, D and E, and G and H) and digital images (C, F and I) show the underwater oil (the red color aids visual inspections) wettability on the multilayers (which were prepared in the absence (D–F) and presence (A–C and G–I) of NaCl salt) of both the NC/BPEI and BPEI/5Acl before (A, D and G) and after (B and C, E and F, and H and I) post chemical modification with glucamine molecules. (J) The change in static oil contact angle (OCA) on the glucamine treated multilayers of the NCs (black and red curves represent the multilayers which were prepared in the presence and absence of salt respectively) and polymer (BPEI, blue curve) with an increase in the number of deposition cycles is shown in the plot. (K–O) The bouncing of the oil droplet (DCM, 11 μL) on the glucamine treated multilayer (9 bilayers, and which was built in the presence of salt) of the NCs under water. Controlled and extreme tailoring of liquid wettability In the recent past, a few approaches have been introduced in the literature to control the liquid wettability through a change in topography 58 – 62 and chemistry 54 – 57 of the coatings. However, in most of the designs, either the topography was tailored by adopting a complex fabrication process 59 – 62 or different chemical functionality was introduced through delicate chemistry 55 – 57 ( i.e. ; metal–sulfur bond and metal–ion interaction). Here, (1) the reactive and covalent LbL deposition process inherently provides a simple avenue to control the topography of the multilayers, and (2) the amine reactive residual acrylate groups in the multilayers of the NCs further allows the modification of the polymeric coatings with a wide range of chemical functionalities through the facile 1,4-conjugate addition reaction. The multilayers of the NCs (from now, multilayers of the NCs will refer to the coatings which were built in the presence of salt) which initially had a WCA ∼92° and an OCA ∼31° displayed superhydrophilicity in air with a WCA of 0° and superoleophobicity under water with an OCA ∼162° after post chemical modification with glucamine molecules (hydrophilic small molecules). However, on post modification of the same multilayers of the NCs with strategically selected primary amine containing small molecules—where the hydrocarbon tail length of the reacted small molecules was judiciously and progressively increased from –C 3 H 7 (propylamine) to –C 18 H 37 (octadecylamine)—the WCA (in air) of the beaded water droplet on the multilayers of the NCs was gradually increased from ∼75° (propylamine) to 153° (octadecylamine) ( Fig. 4A red curve), whereas the OCA of the beaded oil droplet under water on the multilayer coating was continuously depleted from ∼158° to 0° ( Fig. 4A black curve), and eventually the oil wettability in the multilayer coating was transformed from superoleophobic (extreme repellency to oil) to superoleophilic (extreme attraction to oil) as shown in Fig. 4A . Such continuous and extreme regulation of both oil and water wettability using a single polymeric coating through facile and direct covalent bonding (the Michael addition reaction) is yet to be introduced in the literature. Fig. 4 (A) The plot is illustrating the controlled and extreme change in both the WCA (red line) in air and the OCA (black line) under water through facile post chemical modification with strategically selected primary amine containing small molecules. The post chemical modification of the multilayers with glucamine and ODA molecules provided extremes of oil and water wettability both in air (inset, red boxes) and under water (inset, black boxes). (B–G) Under water advancing (B, D and F) and receding (C, E and G) OCA of the beaded oil droplets on the multilayers of the NCs that were post chemically modified with propylamine (B and C), octylamine (D and E) and decylamine (F and G). (H) The plot shows the change in advancing OCA (black) and contact angle hysteresis (grey) under water after post chemical treatment of the multilayers with selected (hydrophilic, less hydrophobic and more hydrophobic) small molecules. Furthermore, other special oil-wettability under water was discovered in the multilayer coating during this strategic post chemical modification of the multilayers of the NCs with amine containing small molecules—which have various hydrocarbon tails. On all occasions, the oil droplets were beaded on the differently post modified multilayers with advancing OCAs above 155°. However, OCA hysteresis was gradually enhanced from ∼21° (propylamine) to 53° (decylamine) with increasing hydrocarbon tail length from –C 3 H 7 (propylamine; Fig. 4B and C ) to –C 10 H 21 (decylamine; Fig. 4F and G ) in the reacted small molecules as shown in Fig. 4H . This unambiguously revealed the gradual increment in adhesive interaction between the beaded oil droplets and multilayer coatings, even though the beaded droplets were extremely repelled by the coatings with an θ \n adv. above 155° ( Fig. 4B, D, F and H ; black curve), and eventually resulted in more deformation of the beaded oil droplets at the liquid/solid interface during the recession of the oil droplets as shown in Fig. 4C, E and G , and a gradual change in CA hysteresis was observed as noted in Fig. 4H (grey curve). However, post modification of the same multilayers of the NCs with small molecules having a much longer hydrocarbon tail (ODA, –C 18 H 37 ) provided a completely different super-wettability—called underwater superoleophilicity—which soaked oil with a 0° contact angle (as discussed in the previous section, Fig. 4A ). Thus, the current approach strategically exploited the 1,4-conjugate addition reaction in tailoring the liquid (water and oil) wettability including extremes of wettability and special wettability (adhesive, but with extreme liquid repellency) on solid surfaces both in air (for water wettability) and under water (for oil wettability) through the controlled modulation of chemical functionality in the reactive multilayers of the NCs. The topography of the coating is another important criteria that confers the heterogeneous wettability of liquids (oil and water) on solid surfaces both in air and under water. In the past various different approaches—which are generally either complex or require specialized equipment— 59 – 62 have been introduced to tailor the topography of the materials for adopting adhesive, but extreme, liquid repellent interfaces both in air and under water. The current approach provides a simple avenue to tune the appropriate topography in the multilayers of the NCs for adopting desired special wettability of liquids through facile and controlled regulation on the LbL deposition cycles as mentioned before. The ODA-modified multilayers of the NCs display extreme water repellency with an advancing WCA ∼167° after 8 bilayer depositions, and the receding WCA was measured to be ∼159° as shown in Fig. 5C and D . However, after 9 bilayer depositions, the same multilayer coating that was constructed from the NC and BPEI was observed to be non-adhesive superhydrophobic with advancing and receding WCAs ∼168° and ∼165° respectively as shown in Fig. 5E and F . Similarly, the oil-wettability under water was modulated in the glucamine treated multilayers of the NCs by simple control in the LbL deposition cycles. The glucamine treated multilayers of the NCs displayed adhesive superoleophobicity with an advancing OCA ∼166° and with a contact angle hysteresis ∼34° after 7 bilayer depositions, and the shape of the beaded oil droplets was significantly deformed at the oil/solid interface as shown in Fig. 5G and H . Whereas, after 8 bilayer depositions, the same multilayer of the NCs was found to be non-adhesive with an advancing OCA ∼167° ( Fig. 5I ) and with a contact angle hysteresis ∼6°, and this multilayer of the NCs became even more non-adhesive superoleophobic with an advancing OCA ∼168° and a contact angle hysteresis ∼4° after 9 bilayer depositions ( Fig. 5K and L ). The whole amount of beaded oil completely receded back without a noticeable deformation of the shape of the beaded oil droplets at the oil/solid interfaces as shown in Fig. 5J and L . Fig. 5 (A–L) Advancing (A, C, E, G, I and K) and receding (B, D, F, H, J and L) contact angle images of both the beaded water (A–F, in air) and oil (G–L, under water) phases on the multilayers (7 bilayers: A and B, and G and H; 8 bilayers: C and D, and I and J; 9 bilayers: E and F, and K and L) of the NCs that were post chemically modified with ODA (A–F) and glucamine (G–L). Thus, the current design of a ‘reactive’ multilayer construction through covalent-LbL deposition allowed us to examine independently the role of the essential parameters—(1) topography and (2) chemical functionality—which confer the heterogeneous liquid wettability on a solid surface, without perturbing the other parameters at the same time. For example, the effect of the change in chemical functionality could be studied independently without perturbing the topography of the material, and vice versa . The heterogeneous wettability of water and oil in air and under water respectively has often been described by adopting the Cassie–Baxter model, 4 where either the metastable trapped air (for water wettability in air) or the confined liquid water layer (for oil-wettability under water) within the featured surface of the material were attributed to the heterogeneous wettability for both water and oil respectively through limited access to the contact area between the solid surface and respective beaded liquid phases. The fraction of the contact area between the beaded liquid (water and oil) phase and multilayer of the NCs (that were either modified with ODA or glucamine) was estimated using eqn (1) and (2), where θ and θ \n r are the liquid (water/oil) contact angles on a multilayer of BPEI/5Acl (smooth surface, 9 bilayers) and multilayer of the NCs in the presence of salt (featured surface; 9 bilayers) respectively. 1 cos θ r = f 1 cos θ – f 2 \n 2 f 1 + f 2 = 1 \n The fraction of the solid contact area and either the trapped air or confined water contact area with the respective beaded liquid (water and oil) phases are labelled as f \n 1 and f \n 2 , respectively. The fraction of the solid contact area with the beaded water droplet on the strategically post modified multilayers of the NCs (in the presence of salt) in air was noticed to gradually decrease with the increasing hydrocarbon tail length of the amine containing small molecules from 0.844 (–C 3 H 7 : propylamine) to 0.091 (–C 18 H 37 : octadecylamine) as shown in Fig. 6J (red curve). As a consequence, the hydrophobicity in the multilayers of the NCs was progressively increased, and eventually, the multilayers of the NCs were embedded with superhydrophobicity ( Fig. 6C ) after ODA treatment ( Fig. 6A ). However, on increasing the hydrophobic tail length of the reacted small molecules in the post modified multilayers of the NCs from –C 3 H 7 (propylamine) to –C 10 H 21 (decylamine), the change in the fraction of the solid contact area for the beaded oil phase on the multilayer coatings was comparatively more sluggish under water, and the solid contact area was only increased from 0.067 to 0.276 as shown in Fig. 6J (black curve). This controlled and sluggish change in the fraction of the solid contact area with the beaded oil phase effectively controls the adhesive interaction between the multilayers and beaded oil phase—but the multilayer surfaces continued to display extreme oil repellency with an advancing OCA above 155° as the fraction of the solid contact area with the beaded liquid phase remained low in general, even after post chemical modification with decylamine molecules. However, the fraction of the solid contact area suddenly increased to 1 on the modification of the multilayers with ODA molecules—which essentially transformed the material to be underwater superoleophilic ( Fig. 6B ). Thus, the same multilayers of the NCs (9 bilayers) with an identical topography but a strategic post chemical modification ( Fig. 6A, D and G ) with appropriate small molecules essentially controlled the contact area access between the solid surface (multilayers of the NCs) and beaded liquid (water/oil) phases both under water (for oil droplets; Fig. 6B, E and H ) and in air (for water droplets; Fig. 6C, F and I ), and eventually tailored the liquid wettability both in air and under water. However, the trends in the change of the fraction of the solid contact area with beaded liquid phases both in air and under water were noted to be considerably different, which again suggested that the stability of the impregnated phases (water and oil)—which are hypothesized to control the homogeneous/heterogeneous wettability of the respective liquids in air and under water—might be different even though the multilayers of the NCs with identical topography were post modified with the same chemically functional molecules. As a consequence, with increasing hydrocarbon tail length from –C 3 H 7 (propylamine) to –C 10 H 21 (decylamine) in the reacted small molecules in the multilayer of the NCs, the hydrophobicity (in air) was increased from ∼76° to ∼121°. In comparison, the multilayers of the NCs were proficient in displaying superoleophobicity (with an advancing OCA above 155°) under water, and the adhesive property was increased with increasing hydrocarbon tail length in the reacted small molecules (Fig. S5 † ). Fig. 6 (A, D and G) Schematic representation of the multilayers (9 bilayers, in the presence of salt) of the NCs that were chemically post-modified with octadecylamine (A), octylamine (D) and glucamine (G). (B–I) Schematic illustrations of the wettability of oil (under water (B, E and H)) and water (in air (C, F and I)) phases on the multilayers that are chemically functionalized with ODA (B and C), octylamine (E and F) and glucamine (H and I) molecules. (J and K) Graphical overview of the change in the fraction of the solid surface area with the beaded water (red line) and oil (black line) phases both in air (red line) and under water (black line) through the modulation of chemical functionality (J) in the multilayers and controlling the deposition cycles (K) during multilayer construction. Next, the effect of consecutive LbL deposition (or the topography of multilayers) of the NCs on the fraction of the solid contact area with the respective liquid (either water or oil) phase was examined both in air and under water, where the chemical functionalities (ODA-treatment ( Fig. 6A ) for water wettability in air and glucamine-treatment ( Fig. 6G ) for oil wettability under water) of the multilayers were kept unaltered during the entire study. The fraction of the solid contact area of the beaded water droplet on the multilayer of the NCs (which was post modified with ODA molecules) in air was initially decreased slowly from 0.88 (2 bilayers) to 0.54 (7 bilayers) with increasing the number of LbL deposition cycles, and after 7 bilayer depositions, the fraction of the solid contact area had significantly dropped as shown in Fig. 6K (red curve), and eventually provided superhydrophobicity to the multilayers. However, in the case of the underwater beaded oil droplet on the multilayer of the NCs, that was post functionalized with glucamine molecules, the fraction of solid contact area was decreased following a ‘linear-like’ trend from 0.78 to 0.09 on increasing the number of consecutive LbL dipping cycles as observed in Fig. 6K (black curve). Thus, this analysis revealed that both the chemical functionality and topography of the material actively contribute to the heterogeneous wettability of both water and oil in air and under water respectively through the appropriate change in the solid contact area with beaded liquid droplets. Physical and chemical durability of the anti-fouling properties In general, several reported anti-fouling materials (having either superhydrophobicity in air or superoleophobicity underwater), which are developed from metal oxides, 38 – 41 polymeric hydrogels 3 , 14 , 18 , 42 – 44 or other organic substances, 45 – 47 are highly delicate and susceptible to perturbation of the chemical functionality and/or the topography that exist in the material, after exposure to common physical manipulations or harsh chemical environments. 27 – 31 , 48 So, the synthesized and optically semitransparent multilayers that are embedded with super-anti-fouling properties were exposed to various harsh chemical and physical insults to investigate the durability of the embedded anti-liquid-wettability properties in the strategically post-modified multilayers of the NCs. First, the multilayers of the NCs were placed under a continuous stream of sand (60 g) that was dropped from a 20 cm height, (Fig. S7H and I † ) and at the end, the materials were recovered with intact anti-fouling properties. Both the water and the oil droplets were beaded on the materials with advancing CAs ≥160° and CA hysteresis ≤10° both in air and under water. A very similar result was also noticed after performing another conventional abrasive test, recognized as the adhesive tape test (Fig. S7A † ), where adhesive tape was manually placed on the multilayers of the NCs, and a load applied (50 g) to facilitate the uniform contact between the adhesive surface and multilayer coatings, prior to peeling off the adhesive tape from the multilayers of the NCs ( Table 1 ). Next, the multilayer coatings were exposed to complex and corrosive aqueous solutions including extremes of pH (pH 1 and pH 11), SDS surfactant (model detergent, 1 mM), artificial sea water (having large salt concentration) and river water (having various bio-contaminants). However, the multilayer coatings retained the ability to repel both water (in air) and oil (under water) droplets with advancing CAs >160° and CA hysteresis <10° as listed in Table 1 and Fig. S6. † This impeccable durability of the liquid wettability in the synthesized materials might arise from the covalent crosslinks in the multilayers through the 1,4-conjugate addition reaction. Table 1 Both the advancing liquid (oil/water) contact angle and respective contact angle hysteresis in air and under water after exposing the multilayers to different physical conditions (sand drop test and adhesive tape test) and chemical (extremes of pH (pH 1, 11), river water, sea water and surfactant (SDS 1 mM)) contaminated aqueous environments Physical and chemical insults Underwater superoleophobicity \n In air superhydrophobicity \n \n θ \n adv. (°) \n θ \n hys. (°) \n θ \n adv. (°) \n θ \n hys. (°) Adhesive tape 165.9 ± 1.7 5.8 ± 2.4 160.5 ± 1.1 9.3 ± 1.6 Sand drop 165.2 ± 0.2 5.1 ± 0.7 164.8 ± 1.9 9.7 ± 1.4 Acid water 165.7 ± 0.7 5.3 ± 0.8 164.4 ± 0.4 4.1 ± 0.6 River water 166.6 ± 0.3 8.1 ± 0.5 161.0 ± 0.3 2.1 ± 1.3 Alkaline water 165.9 ± 0.7 6.1 ± 2.2 164.9 ± 1.5 5.3 ± 2.1 SDS water 165.7 ± 0.5 3.0 ± 1.1 165.7 ± 0.6 1.8 ± 1.4 Sea water 166.3 ± 1.0 2.5 ± 1.4 166.9 ± 0.8 6.0 ± 0.8 Multilayer coatings on various substrates There are few reported approaches 47 , 63 – 66 in the literature that can coat various substrates irrespective of the chemical composition (metallic, polymeric, hydrophilic, hydrophobic etc. ), geometry and physical states (rigid, flexible, smooth, fibrous etc. ) of the substrates. Here, the current covalent and ‘reactive’ LbL-deposition approach allows the coating of various substrates including smooth (glass or Al-foil), fibrous (synthetic fabric) and underwater oleophobic/hydrophobic (wood surface) substrates with multilayers of the NCs to decorate the substrates with the desired anti-wettability properties in air and under water through facile and appropriate post chemical modification. The uncoated synthetic fabric, which soaked both water and oil in air and under water respectively with a CA of 0° ( Fig. 7A and J ), was observed to display extreme liquid (water/oil) repellency both in air and under water with an advancing contact angle above 160° ( Fig. 7D and M ), and the respective liquid droplets (red color aids visual inspection) were beaded on the appropriately post functionalized multilayer coatings with spherical shapes ( Fig. 7G and P ), which further revealed the existence of both superhydrophobicity (in air) and superoleophobicity (under water) in the multilayer coatings. Similarly, a smooth metal surface, Al-foil, which is inherently oleophobic (OCA ∼138°; under water; Fig. 7B ) and hydrophilic (WCA ∼64° in air; Fig. 7K ), was decorated with under water superoleophobicity (an advancing OCA ∼165°, Fig. 7E and H ) and superhydrophobicity (an advancing OCA ∼168°; Fig. 7N and Q ) in air by exploiting the current design. Another model substrate—wood—which moderately repels both oil (CA ∼136° under water; Fig. 7C ) and water (CA ∼114° in air; Fig. 7L ) became superoleophobic under water and superhydrophobic in air after coating the substrate with multilayers of the NCs that were post-chemically modified with appropriate small molecules through the facile Michael addition reaction. The stability of the coating on these different substrates was investigated by the adhesive tape peeling test, and the polymeric coating was found to have uninterrupted anti-fouling properties even after the adhesive tape peeling test. Thus, the current strategy could be highly useful for marine and other relevant applications including prevention of oil contamination, oil/water separation, protein crystallization, developing smart fabrics etc. . Moreover, this approach would enable the coating of a wide range of other useful substrates with desired anti-fouling properties for a specific or wide variety of applications related to energy, health and the environment. Fig. 7 (A–I) Contact angle images (A–F) and digital images (G–I) of the beaded oil droplets (red color aids visual inspection) on both the uncoated (A–C) and multilayer (9 bilayers, treated with glucamine) coated (D–I) flexible ((A, D and G) fibrous cotton substrate and (B, E and H) Al-foil) and rigid ((C, F and I) wood) substrates under water. (J–R) Contact angle images (J–O) and digital images (P–R) showing the change in water wettability (in air) before (J–L) and after (M–R) decorating the cotton fabric (J, M and P), Al-foil (K, N and Q) and wood (L, O and R) with multilayers of the NCs (9 bilayers, post treated with ODA)."
} | 11,738 |
34568692 | PMC8459429 | pmc | 1,161 | {
"abstract": "In this work, a durable\nsuperhydrophobic fabric was fabricated\nby a facile covalent surface modification strategy, in which the anchoring\nof 10-undecenoyl chloride (UC) onto the fabric through the esterification\nreaction and covalent grafting of n -dodecyl-thiol\n(DT) via thiol-ene click chemistry were integrated into one step.\nFourier transform infrared spectroscopy (FTIR), X-ray photoelectron\nspectroscopy (XPS), and scanning electron microscopy (SEM) measurement\nresults demonstrated that UC and DT were covalently grafted onto the\nfabric surface. The formed gully-like rough structure by the grafted\nUC and DT on the fabric surface together with the inherent microfiber\nstructure, combined with the grafted low-surface-energy materials\nof UC and DT, gave the resultant modified DT–UC@fabric superhydrophobic\nperformance. The superhydrophobic DT–UC@fabric was used for\nseparation of oil–water mixtures; it exhibited high separation\nefficiency of more than 98%. In addition, it presented excellent durability\nagainst mechanical damage; even after 100 cyclic tape-peeling and\nabrasion tests, the DT–UC@fabric could preserve superhydrophobic\nperformance, which was ascribed to the formed covalent interactions\nbetween the fabric surface and the grafted UC and DT. Therefore, this\nwork provided a facile, efficient strategy for fabricating superhydrophobic\ncomposites with excellent durability, which exhibited a promising\nprospect in the application of self-cleaning and oil–water\nseparation.",
"conclusion": "3 Conclusions A non-fluorinated durable superhydrophobic DT–UC@fabric\nwas prepared via one-step covalent surface modification. It exhibited\nhigh oil–water separation efficiency of larger than 98% for\nvarious organic solvent/water mixtures. Meanwhile, the DT–UC@fabric\npresented excellent durability and could preserve superhydrophobic\nperformance against mechanical damage. This was associated with strong\ncovalent interactions between the grafted UC and DT and fabric fibers\nby the one-step surface modification method. Hence, this work provides\na facile, high-efficiency strategy to fabricate a durable superhydrophobic\nfabric.",
"introduction": "1 Introduction arge\namount of oily wastewater produced by frequent oil spills\nand leaks of organic compounds is one of the significant threats to\nhuman’s living environment and health, 1 , 2 which\nimpels the development of novel approaches and materials for efficient\noil–water separation, considered to be a great issue of concern. 3 , 4 A variety of techniques and materials have been employed for cleaning\noil-contaminated water, including physical absorption, 5 in situ burning, 6 bioremediation, 7 etc. However, these traditional cleanup methods\nconsume substantial time and human, material, and financial resources\nand have low efficiencies. Since special superhydrophobic/superoleophilic\nmaterials were reported for oil–water separation, 8 diverse types of superhydrophobic/superoleophilic\nmaterials including porous sponges 9 , 10 and fabrics 11 , 12 were developed. Particularly, cellulose-based fabrics like cotton\nand woven have gained tremendous attention owing to their superiority\nin the aspects of light weight, low cost, recycling, biodegradability,\nand excellent mechanical strength. However, the fabric surface needs\nto be modified to achieve superhydrophobic properties prior to application.\nRecently, various strategies including dip-coating, 13 chemical vapor deposition, 14 and plasma treatment 15 have been developed\nto fabricate superhydrophobic fabrics. These methods commonly involve\ndepositing nanoparticles or low-surface-energy substances onto the\nfabric surface. Unfortunately, superhydrophobic surfaces are vulnerable\nto mechanical damage. To enhance durability, it is common to establish\ncovalent bonds between fabrics and nanoparticles or low-surface-energy\nsubstances. 16 Up to date, as a novel\nchemical modification method, photo-induced\nthiol-ene click chemistry has been applied to prepare superhydrophobic\nfabrics owing to its highly efficient covalent grafting of ene-terminated\nmolecules on thiol-terminated surfaces under UV light. 17 − 19 For instance, Hou et al. grafted methacryl-heptaisobutyl polyhedral\noligomeric silsesquioxane (MA-POSS) onto a sulfhydryl pretreated cotton\nfabric via thiol-ene click chemistry. The formed rough structure and\nlow surface energy of MA-POSS promoted fabric superhydrophobicity. 20 Xue et al. used a similar method to graft dodecafluoroheptyl\nmethacrylate onto sulfhydryl-pretreated fabrics to obtain superhydrophobic\nfabrics. 21 These methods provide the fabric\nexcellent mechanical durability arising from the covalent bond interaction\nbetween fabrics and low-surface-energy compounds. However, these methods\nsuffer from cumbersome steps, such as alkali etching or sulfhydryl\ntreatment on the fabric in advance, as well as involvement of environment\nunfriendly low-surface-energy substances like fluoride. Inspired\nby previous studies, we present a facile, one-step, fluorine-free\nhighly efficient strategy to prepare a durable superhydrophobic fabric,\nwhich is schematically illustrated in Figure 1 . The anchoring of high chemical activity\nof 10-undecenoyl chloride (UC) onto fabrics through the esterification\nreaction and covalent grafting of n -dodecyl-thiol\n(DT) via thiol-ene click chemistry are combined into one step with\nthe aid of UV light. The bridging effect of UC in the reaction process\nensures the optimum attachment of DT on the fabric surface. The combination\nof the formed rough structure and low-surface-energy materials promoted\nthe fabric to be superhydrophobic and could be used for oil–water\nseparation. More interestingly, the as-prepared fabric exhibited excellent\ndurability against mechanical damages like abrasion and tape peeling.\nThis was associated with the strong covalent interactions between\nthe grafted UC and DT and fabric fibers. Hence, this work provided\na facile, high-efficiency strategy to fabricate a durable superhydrophobic\nfabric, exhibiting promising application in the field of oil recovery\nand water purification. Figure 1 Schematic preparation of a superhydrophobic\nDT–UC@fabric.",
"discussion": "2 Results\nand Discussion 2.1 Preparation of the Superhydophobic\nDT–UC@Fabric In the current work, a cellulous fabric\nwas used as the starting\nmaterial. A facile one-step dip-coating strategy was developed to\nprepare a superhydrophobic fabric, which is schematically illustrated\nin Figure 1 . The cellulose-based\nfabric was immersed into a solution containing 10-undecenoyl chloride\n(UC), n -dodecylthiol (DT), and the photo-initiator\nHMPF. The naturally existent hydroxyl groups in the fabric readily\nconverted to esters by the treatment of the fabric with UC; 22 thus, UC was grafted onto the fabric surface\nthrough the esterification reaction with hydroxyl groups due to its\nhigh chemical activity. Accompanying with the esterification reaction,\nthe thiol-ene click reaction also occurred between ene-containing\nUC and thiol-containing DT with the assistance of UV irradiation with\na wavelength of 365 nm, resulting in DT molecules covalently anchored\nonto the fabric surface. Upon the completion of the reaction, the\nfabric was washed with water, ethanol, and acetone to remove impurities\nfollowed by drying in an oven, and the target fabric was obtained\n(designated as DT–UC@fabric). The key parameters affecting\nthe fabric surface including UC and DT concentrations were optimized\nto get the best characteristics. Considering thiol-ene click chemistry\nis a highly efficient reaction, in which the theoretical molar ratio\nof thiol and ene groups is 1:1, the molar ratio of UC and DT was fixed\nat 1:1. To investigate the effect of UC and DT concentrations on the\nhydrophobicity of the DT–UC@fabric, the mass concentration\nof UC was adopted for convenience, and the corresponding results are\nplotted in Figure 2 a. When the UC concentration was increased from 2 to 8%, the water\ncontact angle (WCA) of the DT–UC@fabric increased from 135\nto 156°, indicating that the DT–UC@fabric was superhydrophobic.\nAfterward, the WCA reached the plateau at 156° even when the\nUC concentration was further increased to 10%. This might be explained\nas follows. When the UC concentration reached 8%, all hydroxyl groups\non the fabric surface were converted to ester groups completely and\nthe UC loading amount reached a maximum. The further increase of the\nUC concentration could not bring about an increase in the UC loading\namount; thus, the WCA did not increase. Therefore, the UC concentration\nwas optimized as 8%, and the corresponding fabric was utilized for\nfurther discussion. Figure 2 (a) Effect of the UC concentration on hydrophobicity of\nthe DT–UC@fabric.\n(b) FTIR and (c) XPS spectra of pristine and DT–UC@fabrics.\nHigh-resolution C 1s spectra of (d) pristine and (e) DT–UC@fabrics. The FTIR spectra of UC, DT, pristine, and DT–UC@fabrics\nare shown in Figure 2 b. Compared with the pristine fabric, two evident peaks were observed\nin the DT–UC@fabric. First, the stretching vibration peaks\nof C–H at 2920 and 2896 cm –1 23 in the DT–UC@fabric are attributed to\nthe alkyl groups in UC and DT. Furthermore, the peaks at 1626 and\n2540 cm –1 observed in the spectra of UC and DT are\nattributed to C=C 24 and S–H, 25 respectively. However, these two peaks were\nnot detected in the DT–UC@fabric, indicating that the C=C\nand S–H bonds had undergone the thiol-ene click reaction. To\nfurther understand the chemical evolution before and after modification,\nXPS was employed to monitor the fabric surface, and the XPS survey\nspectrum is shown in Figure 2 c. The pristine fabric was only composed of C and O elements,\nwith the C 1s peak and O 1s peak at binding energies of 533 and 285\neV, 26 respectively. However, for the DT–UC@fabric,\nbesides the original C and O peaks, new peaks of S 2p and S 2s were\nalso detected at binding energies of 166.6 and 230.1 eV, 19 , 21 which could be from the S element in DT. Furthermore, in their high-resolution\nC 1s spectra, three obvious peaks located at 287.6 eV (C=O),\n286.4 eV (C–O), and 284.6 eV (C–C) were observed in\nthe pristine fabric ( Figure 2 d), while a new peak was detected at 286.1 eV in Figure 2 e, which should be\nascribed to the C–S bond. 27 Moreover,\nthe C–O peak was weakened, while the C–C peak was significantly\nincreased in Figure 2 e compared with the pristine fabric, indicating that the surface\nof DT–UC@fabric was dominated by a large amount of alkyl groups.\nThis further verified that UC and DT had grafted onto the fabric surface\nthrough the one-step covalent surface modification method. 2.2 Surface Analysis of the Superhydophobic DT–UC@Fabric The surface morphologies of pristine and DT–UC@fabrics are\ncharacterized by SEM and shown in Figure 3 a–f, respectively. Both the fibers\nin pristine and DT–UC@fabrics are in ordered arrangement (see Figure 3 a,d), but it should\nto be noted that the surface on the pristine fabric is very smooth\nexcept some inherent texture (see Figure 3 b,c) due to the wicking effect of the porous\nstructure, 28 , 29 whereas the surface on the DT–UC@fabric\nis pretty rough, where tortuous gully-like structures are clearly\nobserved on the microfiber surface (see Figure 3 e,f). Figure 3 SEM images of (a–c) the pristine\nfabric and (d–f)\nthe DT–UC@fabric. AFM images of (g–i) the pristine fabric\nand (j–l) the DT–UC@fabric. Besides the SEM images, the surface morphologies of pristine and\nmodified fabrics are further characterized by AFM, as shown in Figure 3 g–l, respectively.\nCompared with the pristine fabric, the 3D surface morphology of the\nmodified DT–UC@fabric became rougher than that of the pristine\nfabric (see Figure 3 g,j), the root mean square roughness (Rq) of the fabric increasure\nfrom 14.649 to 149.856 nm, and the height difference was increased\nfrom 51.81 to 351.95 nm from the analysis chart (see Figure 3 i,l). The structural evolution\non the fabric surface should be a result of the covalently grafted\nUC and DT during the one-step modification. More evidence was found\nin the EDS spectra and element mapping of the DT–UC@fabric\n(see Figure S1 ), in which new occurrences\nof element S was detected; meanwhile, the element content of C increased\nbut the O element decreased after modification, which verified large\namounts of long alkyl chains in UC and DT had grafted on the fabric\nsurface, resulting in the fabric fiber become rough due to the wrinkle\neffect. 30 2.3 Wetting\nBehavior and Oil–Water Separation\nPerformance of the DT–UC@Fabric As depicted in Figure 4 a,b, the photographs\nrecorded the sitting status of water, coffee, and ink droplets on\npristine and DT–UC@fabrics, respectively. When these droplets\nwere dropped onto the pristine fabric, they spread quickly and then\nwere absorbed. While when water, coffee, and ink droplets were dropped\nonto the DT–UC@fabric surface, these droplets could stand and\nformed a round shape with WCA ∼156° and maintained the\nsame shape even after 30 min, indicating that the DT–UC@fabric\nfeatured stable superhydrophobic performance. While when n -hexane was dropped on the DT–UC@fabric surface, it was absorbed\nimmediately. Additionally, when the DT–UC@fabric was completely\nimmersed into water, a mirror-like phenomenon was observed ( Figure S2 ), where an air bubble layer was formed\non the surface of the fabric. This phenomenon could be attributed\nto the entrapped air cushion between water and the modified fabric\nsurface, which was formed due to the micro–nano hierarchical\nstructure in which air molecules could retain. This air cushion acted\nas a shield to effectively inhibit water penetration into the fabric\nsurface even in full immersion. After being taken out of water, the\nsurface of the fabric was not wet, implying that the as-prepared DT–UC@fabric\ndisplayed durable superhydrophobicity. Furthermore, the superhydrophobic\nDT–UC@fabric presented a self-cleaning effect when it was contaminated\nwith dust. When water droplets were continuously dropped onto the\ntilted fabric surface, they would roll off the surface at once and\nflush away the dust layer, and the dust-contaminated fabric would\nreturn to be clean ( Figure S3 and Video S1 ). Based on its special oil-absorbing\nwater-repelling feature, it could absorb dichloromethane and remove\nthem from the dichloromethane/water mixture ( Figure S4 and Video S2 ). Considering its\ninherent high porosity, the superhydrophobic DT–UC@fabric could\nbe utilized for continuous separation of oil–water mixture.\nA simple equipment was set up in Figure 4 c, where the superhydrophobic DT–UC@fabric\nwas sandwiched between two double-pass pipes. Pouring the dichloromethane/water\nmixture into the double-pass pipe, dichloromethane would settle to\nthe bottom due to its large density. Once dichloromethane approached\nthe DT–UC@fabric, it was absorbed immediately due to the superoleophilicity,\nwhich could be owing to the tortuous gully-like structures and the\ngrafted hydrophobic long alkyl chain; thus, a layer of dichloromethane\nwas formed on the fabric, which increased the contact area with oil\ndroplets. This was based on the theoretical intrusion pressure formula\nas follows 31 , 32 where γ\nis the surface tension, R is the radius of the meniscus,\nand L and A are the circumference\nand the area of the fiber pore,\nrespectively. θ is the advancing angle on the fabric. According\nto this equation, it can be seen that for oil, the fabric cannot bear\nany pressure due to Δ p < 0; thus, oil can\neasily pass through the fabric, while the water can be supported because\nΔ p > 0. As time elapsed, the absorbed dichloromethane\ngradually penetrated into the fabric and flows slowly into the beaker\nunder gravity, while water remained on the upper layer throughout\nthe process (see Video S3 ). As a result,\ndichloromethane was successfully separated from the dichloromethane/water\nmixture. Moreover, separation efficiency was taken into account to\nevaluate its separation performance. The separation efficiency of\nthe DT–UC@fabric for chloroform, dichloromethane, hexane, toluene,\ngasoline, acetonitrile, and ethyl acetate was above 98% in every case,\nas shown in Figure 4 d, indicating the DT–UC@fabric possessed excellent separation\nperformance. Figure 4 Digital photographs of water, coffee, and ink droplets\non (a) pristine\nand (b) DT–UC@fabrics. (c) Separation process of the dichloromethane/water\nmixture (DCM was dyed with Sudan II, water was dyed with CuSO 4 ). (d) Separation efficiency of the DT–UC@fabric for\nvarious organic solvent/water mixtures. 2.4 Durability of the Superhydrophobic DT–UC@Fabric In general, most superhydrophobic surfaces feature poor durability\nand are vulnerable to mechanical damage, corrosive solution environment,\nand high-temperature exposure; thus, it is essential to investigate\ntheir durability under different conditions. 33 − 36 Here, the durability of the prepared\nsuperhydrophobic DT–UC@fabric was evaluated by water washing,\nas well as acid and alkaline conditions and organic solvent exposures.\nFor the stability of washing tests, the fabric’s WCA can be\nmaintained over 150° even with washing that lasts for 20 h, as\nshown in Figure 5 a.\nWhen encountering acid and alkaline water droplets in a wide pH range\n(2–12), the fabric surface can preserve its superhydrophobic\nproperty, as shown in Figure 5 b. Additionally, even after a long time of immersion in an\norganic reagent, the fabric does not lose its superhydrophobic performance,\nas shown in Figure 5 c. Figure 5 WCA variation of the DT–UC@fabric at (a) various washing\ntimes, (b) water droplets with various pH values, and (c) soaking\nwith various organic solvents. (d) WCA variations and (e, f) SEM images\nof the DT–UC@fabric during the tape-peeling test. (g) WCA variations\nand (h, i) SEM images of the DT–UC@fabric during the abrasion\ntest. Furthermore, the durability of\nthe DT–UC@fabric against\nmechanical damage was evaluated by tape-peeling and abrasion tests.\nThe effect of tape-peeling times on WCA is shown in Figure 5 d. The WCA of the DT–UC@fabric\nonly decreased slightly with the increase in tape-peeling tests; even\nwhen tape-peeling cycles increased to 100, the fabric still preserved\nits superhydrophobicity (WCA about 152°). Figure 5 e,f revealed the SEM images of DT–UC@fabric\nafter the tape-peeling test. When compared with the sample before\nthe test, the overall fibers remained in a similar state, except for\nbecoming floating and loose, not compact as before (see Figures 5 e and 3 d). In the magnified SEM images, no obvious destruction of the rough\nstructure was observed, but minor changes like tortuous gully-like\nstructures disappeared (see Figures 5 f and 3 e). This could be due\nto the adhesive effect of the tape, which also makes the low-surface-energy\nmaterials leave the fiber surface; herein, the WCA of the DT–UC@fabric\ndecreased slightly but remained larger than 150°. The influence\nof abrasion on WCA variations of DT–UC@fabric\nis shown in Figure 5 g. With the increase in cyclic abrasion, the WCA of the DT–UC@fabric\nshowed a trend of slight decrease but remained about 151° after\n100 abrasion cycles. Figure 5 h showed that the overall morphology of the DT–UC@fabric\nsurface did not change significantly; only the fiber arrangement and\ndistribution became more compact compared with that before test (see Figure 3 d), which should\nbe associated with the abrasion pressure. In a magnified image (see Figure 5 i), the fiber surface\nstructure remained clearly rough. One should note that the previous\ntortuous gully-like structures transformed into a wrinkled state,\nimplying some grafted materials were desquamated from the fabric surface,\nthus leading to a slight decrease in WCA. The abovementioned tests\nindicated the DT–UC@fabric had excellent mechanical durability\nagainst damage; this should be derived from the formation of strong\nchemical bond between UC and DT and the fabric."
} | 4,991 |
40109105 | PMC11969388 | pmc | 1,162 | {
"abstract": "Division of labour (DOL) plays a key role across all scales of biological organization, but how its expression varies across contexts is still poorly understood. Here, we measure DOL in a crucial task, colony defence, in a social insect that affords precise experimental control over individual and colony traits, the clonal raider ant ( Ooceraea biroi ). We find that DOL in defence behaviour emerges within colonies of near-identical workers, likely reflecting variation in individual response thresholds, and that it increases with colony size. Additionally, colonies with pupae show higher defence levels than those without brood. However, we do not find evidence for a behavioural syndrome linking defence with exploration and activity, as previously reported in other systems. By showing how colony composition and size affect group response to potential threats, our findings highlight the role of the social context in shaping DOL. This article is part of the theme issue ‘Division of labour as key driver of social evolution’.",
"introduction": "1 . \n Introduction Division of labour (DOL), whereby members of a group specialize in distinct tasks, is a key feature of social systems from microorganisms [ 1 ] to insects [ 2 , 3 ] and humans [ 4 ]. DOL plays a key role at all scales of biological organization [ 5 ] and the emergence of new forms of DOL underlies all the major evolutionary transitions, such as those from prokaryotic to eukaryotic cells, from unicellular to multicellular organisms and from solitary to eusocial life [ 6 – 8 ]. DOL often increases with within-group heterogeneity, e.g. in age [ 2 , 9 ], genetic background [ 10 , 11 ] or morphology [ 12 ], which is believed to create the conditions necessary for individuals to specialize in different tasks. Social insects (ants, termites, bees and wasps) are among the most ecologically successful animals [ 13 , 14 ] and display some of the most extreme and elaborate forms of DOL [ 13 , 15 , 16 ]. Typical social insect colonies have reproductive DOL between one or a few queens, which monopolize reproduction, and (functionally) sterile workers, which perform all other tasks needed for colony maintenance [ 17 ]. Additionally, there is (non-reproductive) DOL among the workers, which specialize in a subset of maintenance tasks (e.g. foraging, nursing, defence). Thus, like a multicellular organism, a social insect colony is composed of related units (individuals in insect colonies, cells in multicellular organisms) that are specialized in different essential tasks, like reproduction (queens in social insect colonies, germline cells in multicellular organisms) or energy storage (repletes in some insect colonies, adipocytes in multicellular organisms) and therefore depend on each other for survival. As a consequence, social insect colonies are often conceptualized as ‘superorganisms’ [ 18 ]. A striking example of specialization in social insect workers is in colony defence. Most social insects live in stable nests, which must be defended against various external threats, including conspecific and allospecific intruders and competitors, predators and parasites [ 19 – 21 ]. Analogous to the immune cells that specialize in defending multicellular organisms against external threats like pathogens, some members of social insect colonies often specialize in colony defence [ 22 , 23 ]. The best-documented and most iconic cases are the morphologically specialized soldier castes of some termites [ 24 ], ants [ 25 ] and bees [ 26 ]. Age polyethism in defence has also been reported, whereby older workers engage in defence more readily than young workers [ 12 , 27 , 28 ]. However, it remains unclear whether DOL in defence can emerge in the absence of inter-individual differences in morphology or age. In social insects and other animals, inter-individual variation in aggressive behaviour has often been described as part of a ‘behavioural syndrome’ linking exploration, aggression and boldness across individuals [ 29 – 33 ]. For example, in great tits ( Parus major ), juvenile males that are more exploratory are also more aggressive (initiate more fights with conspecifics) [ 34 ]. Similarly, individuals belonging to the patroller caste of Myrmica ants show higher aggression towards allospecifics, increased exploration of a novel environment and higher activity in response to alarm pheromones [ 35 ]. Empirical studies on behavioural syndromes in social insects typically measure behavioural traits in isolated individuals (i.e. outside the colony context). Studies conducted in a social setting are rare and results are equivocal [ 36 – 38 ]. Furthermore, individual variation in exploration and defence behaviours is often associated with variation in age and genetic background [ 28 , 39 – 41 ], and it remains unclear whether behavioural syndromes manifest among workers of social insect colonies when these factors are controlled [ 29 ]. Here, we investigate DOL in colony defence against allospecific intruders in the clonal raider ant, Ooceraea biroi . In this species, colonies are queenless and composed of workers that reproduce asexually and synchronously, so that genetically near-identical adults are produced in discrete age cohorts [ 42 ]. This allows us to investigate DOL in colony defence in the absence of variation in age or genetic background between individuals and between colonies. We first ask whether there is DOL in colony defence. Then, we test whether two aspects of the social environment, colony size and brood treatment, affect DOL and efficiency in colony defence. Group size has been shown to increase DOL in other tasks (intranidal versus extranidal tasks) in the clonal raider ant [ 43 ]. We expected the presence and type of brood to affect DOL in defence in the clonal raider ant because the brood is known to affect worker physiology and behaviour in this species [ 42 , 44 ]. Finally, we ask whether DOL in colony defence is linked to a behavioural syndrome by testing for an association between individual exploratory behaviour, activity and aggressive behaviour in a colony defence context.",
"discussion": "4 . \n Discussion We found DOL in colony defence in O. biroi , demonstrated by inter-individual variation and individual consistency in defence behaviour. Although all ants used here had near-identical ages and genotypes and were reared in the same controlled environment, they differed consistently in defence behaviour, suggesting that DOL in colony defence can arise in small, homogeneous social groups. Because we measured the performance of a task in response to a controlled experimental stimulus and corrected for variation in encounter rates with the stimulus, the observed differences in behaviour are likely to reflect individual variation in response thresholds. Variation in response thresholds is one of the theoretically best-studied mechanisms by which DOL can emerge [ 57 , 58 ], but empirically measured variation in response thresholds is still rare. Here, we detect consistent variation in response thresholds that is independent of age or genetic background (as shown by inter-trial consistency in defence scores). As shown in the same species for other behaviours [ 43 ], DOL in defence increased with colony size. Although our conclusions are based on the study of a single clonal lineage (B), previous work has shown qualitatively similar patterns of DOL across two commonly used genotypes (A and B) [ 43 ], suggesting DOL in colony defence exists in other clonal lineages as well. The type of brood present in colonies affected defence efficiency, with colonies containing pupae displaying more efficient defence than colonies without brood. This pattern could reflect a context-dependent defence strategy of the colony [ 59 , 60 ]. Pupae used in this study were the product of ca 20 days of brood care, a substantial time and resource investment. In addition, late-stage ant pupae have nutritional value for the colony: the moulting fluid they produce is rich in nutrients, hormones and neuroactive substances which are consumed by adults [ 61 ]. These factors together may increase the defensive investment in pupae. We did not observe an increase in DOL in colonies with pupae, suggesting that the enhanced defensive investment may be a general colony response rather than the result of increased defensive behaviour in a subset of specialized workers. We failed to detect a behavioural syndrome linking exploration or activity with aggression as joint manifestations of boldness that correlate across individuals, as has been reported in various species [ 30 – 33 ], including ants [ 29 , 35 , 36 ]. Here, while ants with higher explorative behaviour encountered intruders more frequently (as expected), they were not more (or less) likely to engage in defence behaviour upon encounter. The discrepancy may stem from the context in which exploratory behaviour and aggression are measured. Previous studies on exploration and aggression in ants often focused on isolated individuals [ 35 , 36 ]. While studying isolated individuals allows the control of social influences and can reveal intrinsic behavioural tendencies, the relationship between exploration and aggression can differ between isolated and social contexts [ 60 ]. For instance, ants have been shown to exhibit lower levels of aggression when alone versus in a group owing to the need for colony defence [ 62 ] or the presence of social cues [ 60 , 63 ]. In a colony setting, the social environment can amplify individual defence based on the social information individuals perceive (e.g. via alarm pheromones [ 64 , 65 ] or non-nestmate cues [ 66 ]). Additionally, past work on social insects involved colonies containing individuals of varying ages, genetic backgrounds and/or morphology. In contrast, we controlled for age and genetic background in O. biroi , a monomorphic species, and found no correlation between exploratory and defensive behaviours. This suggests that the behavioural syndrome reported in social insects might stem in part from underlying inter-individual variation in factors such as genotype [ 67 ], morphological caste [ 35 ] or age [ 68 ]. How consistent differences in behaviour arise among genetically identical individuals reared in the same environment is an important, unanswered question. Studies on naturally clonal species (including fish [ 69 ], bacteria [ 70 ] and insects [ 71 ]), as well as animals from isogenic laboratory lines [ 72 , 73 ], and human monozygotic twins [ 74 ], consistently show stable inter-individual differences in behaviour, but the developmental and neurological bases of such variation are only starting to be elucidated [ 72 , 75 ]. Social insects are notorious for their extreme developmental phenotypic plasticity. In clonal raider ants, differences in adult behavioural tendencies might arise from small (potentially stochastic) differences in development (e.g. temperature and nutrition at the larval stage). Inherent differences in adult behavioural tendencies may then be plastically amplified by several factors, including group size and composition (via e.g. threshold responses), individual experience [ 76 ] or social interactions, to generate stable DOL among group members."
} | 2,816 |
33902725 | PMC8067651 | pmc | 1,163 | {
"abstract": "Background Sequencing of 16S rRNA genes has become a powerful technique to study microbial communities and their responses towards changing environmental conditions in various ecosystems. Several tools have been developed for the prediction of functional profiles from 16S rRNA gene sequencing data, because numerous questions in ecosystem ecology require knowledge of community functions in addition to taxonomic composition. However, the accuracy of these tools relies on functional information derived from genomes available in public databases, which are often not representative of the microorganisms present in the studied ecosystem. In addition, there is also a lack of tools to predict functional gene redundancy in microbial communities. Results To address these challenges, we developed Tax4Fun2, an R package for the prediction of functional profiles and functional gene redundancies of prokaryotic communities from 16S rRNA gene sequences. We demonstrate that functional profiles predicted by Tax4Fun2 are highly correlated to functional profiles derived from metagenomes of the same samples. We further show that Tax4Fun2 has higher accuracies than PICRUSt and Tax4Fun. By incorporating user-defined, habitat-specific genomic information, the accuracy and robustness of predicted functional profiles is substantially enhanced. In addition, functional gene redundancies predicted with Tax4Fun2 are highly correlated to functional gene redundancies determined for simulated microbial communities. Conclusions Tax4Fun2 provides researchers with a unique tool to predict and investigate functional profiles of prokaryotic communities based on 16S rRNA gene sequencing data. It is easy-to-use, platform-independent and highly memory-efficient, thus enabling researchers without extensive bioinformatics knowledge or access to high-performance clusters to predict functional profiles. Another unique feature of Tax4Fun2 is that it allows researchers to calculate the redundancy of specific functions, which is a potentially important measure of how resilient a community will be to environmental perturbation. Tax4Fun2 is implemented in R and freely available at https://github.com/bwemheu/Tax4Fun2 .",
"conclusion": "Conclusion With Tax4Fun2, we provide an easy-to-use, platform-independent R package, which enables researchers to predict and investigate functional profiles of prokaryotic communities based on 16S rRNA gene data. We demonstrate the high predictive power of Tax4Fun2, providing superior results to any other established tool. The key strength of Tax4Fun2 is the incorporation of user-defined and habitat-specific data, which further enhances the accuracy of the predictions. Another unique feature of Tax4Fun2 is that it enables researchers to calculate functionial redundancies, which is a relevant parameter for ecosystem monitoring and the development of management strategies to safeguard optimal ecosystem functionality. Nonetheless, functional profiles and functional redundancies are predictions only and should be treated with caution. Availability and requirements Project name: Tax4Fun2. Project homepage: https://github.com/bwemheu/Tax4Fun2 Operating system(s): Platform-independent. Programming language: R. Other requirements: BLAST+ 2.7.1 or later, R packages ape and seqinr. License: GNU General Public License v3.0. Any restrictions to use by non-academics: no.",
"discussion": "Results and discussion Tax4Fun2 evaluation We first applied Tax4Fun2 in comparison to Tax4Fun [ 23 ] and PICRUSt [ 22 ] using the same paired samples (16S rRNA gene and metagenome data), which were used to validate both tools, i.e. samples derived from the human microbiome, mammalian guts, soil and from a hypersaline microbial mat (Table 1 ), in addition to ten marine seawater samples taken in the North Sea [ 7 ] and 90 kelp-associated samples collected within the Marine Microbes Framework Data Initiative ( http://www.bioplatforms.com/marine-microbes ).\n Table 1 Accession numbers of samples/studies used to validate Tax4Fun2 Origin Sample number Accession numbers Human Microbiome 41 SRS011271, SRS011452, SRS011529, SRS011584, SRS011586, SRS013234, SRS013252, SRS013258, SRS013506, SRS013687, SRS013711, SRS013723, SRS014235, SRS014287, SRS014343, SRS014613, SRS014629, SRS014923, SRS015133, SRS015190, SRS015425, SRS015450, SRS015574, SRS015578, SRS015762, SRS015782, SRS015854, SRS015960, SRS016002, SRS016018, SRS016095, SRS016111, SRS016203, SRS016225, SRS016331, SRS016335, SRS016349, SRS016434, SRS016533, SRS016553, SRS016559 Mammalian Gut 56 4,461,284–301, 4,461,341–55, 4,461,357–58, 4,461,360–80, 4,461,383 (MG-RAST) Microbial Mat 10 4,440,963–71 (MG-RAST) Soil 14 4,477,803–5, 4,477,807, 4,477,872–7, 4,477,899, 4,477,902–4 (MG-RAST) Water 10 SRA060677 Kelp 90 57,884–936, 57,938–56, 87,958–74, 58,019–20 (https://data.bioplatforms.com/organization/about/australian-microbiome).) Functional profiles were predicted using the default workflows. For PICRUSt, processed sequences were clustered using QIIME version 1.8 [ 55 ] by closed-reference picking against the Greengenes database (version 13_5; [ 56 ]) and normalized prior to functional prediction. For Tax4Fun, OTU sequences were taxonomically classified by BLAST search [ 57 ] against the SILVA database (SILVA_123_SSURef_Nr99) [ 46 ]. We evaluated the predictive power of each tool by comparing the functional profiles predicted from the 16S rRNA data to functional profiles generated directly from the metagenomes using Spearman rank correlations. Comparing the profiles predicted with PICRUSt, Tax4Fun and Tax4Fun2 with metagenome-derived profiles clearly showed that Tax4Fun2 outperforms PICRUSt and Tax4Fun across all six tested datasets (Fig. 3 ). In addition, Tax4Fun2 was more than 20 times faster than Tax4Fun due to the smaller reference database.\n Fig. 3 Correlations between functional profiles obtained from metagenomic datasets and those predicted from 16 s rRNA data. Predictions were made with PICRUSt, Tax4Fun, and Tax4Fun2. Predictions with Tax4Fun2 were made using both supplied default reference datasets (Ref99NR and Ref100NR). Note that PICRUSt did not generate any prediction for the kelp data It should be noted that a direct comparison of functional profiles predicted with all three tools is difficult due to several changes in the KEGG Orthology since PICRUSt and Tax4Fun were developed (deprecated and new functional orthologs). Hence, predicted functional profiles as well as those obtained by metagenomic shotgun sequencing were converted to relative abundances prior to comparison. Only functions present in the metagenomic profile and in the predicted profile were considered in each comparison. On average, more than 95% of all functions in the human microbiome as well as the marine, soil and kelp-derived metagenomes were affiliated to these shared KOs. In the microbial mat and the mammalian gut samples, more than 74 and 78% of all predicted functions were affiliated to these KOs, respectively. Using user-defined data increased the accuracy and reduced the FTU Following the assumption that any additional genomic information specific for the investigated habitat further enhances the predictive power of Tax4Fun2, we used 68 metagenome-assembled genomes (MAGs) generated from the 90 kelp-associated metagenomes to build a kelp-specific reference dataset. These genomes were selected because at least one 16S rRNA gene sequence was identified in each genome. Using the default data, the median Spearman correlation coefficient was 0.72. Incorporating the kelp-specific data substantially increased the power of the functional prediction (median Spearman correlation coefficient = 0.86) and reduced the fraction of the sequences not used in the prediction (Fig. 4 ), showing that a lack of suitable reference genomes did initially limit Tax4Fun2’s performance. Moreover, using the kelp-specific dataset enabled us to predict functional profiles for samples, which failed when using only the default reference data, because next neighbour search resulted in no close matches. These results demonstrate the benefits of incorporating user-defined, habitat-specific reference databases, which distinguishes Tax4Fun2 from all other published tools.\n Fig. 4 Correlations between functional profiles retrieved from 90 kelp metagenomes and those predicted with Tax4Fun2 without and with user data added and the fraction of zOTUs and sequences unused in the prediction Functional redundancy index The simultaneousness assessment of multiple functions is usually very time-consuming [ 20 ]. Previous studies have focused on a limited number of community functions (e.g., [ 58 – 60 ]). However, the degree of functional redundancy in any given system depends on the functions considered [ 14 ]. In addition, it is difficult to draw conclusions about functional redundancy based on a single environmental situation, as species that are functionally redundant in one environment might be pivotal in another [ 61 ]. Hence, the extent of functional redundancy change as the ecological contribution of a species might change between different environments. Nonetheless, a contemporary concern for the conservation of biodiversity and the development of management strategies is that decision makers require quantitative measures as part of science-based negotiations and communications. In order to provide those measures when assessing natural or human-induced impacts on an ecosystem, we introduced the FRI with respect to multiple functions in Tax4Fun2. A high FRI indicates that a specific function is almost ubiquitous in all community members, whereas a low FRI suggests that the function is present in a few closely related species. A FRI of 0 indicates that a function has been detected in only one community member or is not present at all. Consequently, the lower the FRI the higher the probability that a function gets lost after community shifts or perturbations. To test the accuracy, we simulated 1000 microbial communities and calculated FRI values based on 16S rRNA gene data using Tax4Fun2. The FRIs calculated for each function were subsequently compared to the FRIs calculated directly from the genomes of each simulation by Spearman rank correlation. The comparison revealed that Tax4Fun2 provides a good estimate of the functional redundancy present in the microbial community (Spearman rank correlation > 90%) (Fig. 5 a).\n Fig. 5 Correlation between predicted and genome-based functional redundancy indices ( a ) and functional redundancy indices inside and outside a phytoplankton bloom ( b ). A log ratio greater than 0 indicates that a function is more redundant outside the bloom. All predictions were made using a 97% similarity cut off We further calculated FRIs using the marine seawater samples. Six of these samples were taken inside a phytoplankton bloom and three samples were taken outside the bloom [ 7 ]. Nearly 7000 functions displayed a higher functional redundancy index outside the bloom, whereas only 1468 functions had higher redundancies inside the bloom (Fig. 5 b). This indicates that the functional redundancy greatly shifts during the phytoplankton bloom. Phytoplankton blooms are usually characterized by a substrate-controlled succession, i.e. distinct bacterial clades dominate the bacterioplankton community at different stages during and shortly after the bloom [ 62 ]. Consequently, community members involved in the turnover of certain substrates at a specific stage are predominant. For instance, the SAR92 clade, the Roseobacter RCA cluster and the genus Polaribacter were more abundant in bloom samples [ 3 ]. Because we did not observe significant differences in the phylogenetic diversity of bacterioplankton communities derived from bloom and reference samples, functions predominantly associated with dominant community members are more redundant in the bloom whereas all other functions display higher redundancies in the reference samples."
} | 2,998 |
27878983 | PMC5299523 | pmc | 1,164 | {
"abstract": "Abstract Enzymatic catalysis is an ecofriendly strategy for the production of high‐value low‐molecular‐weight aromatic compounds from lignin. Although well‐definable aromatic monomers have been obtained from synthetic lignin‐model dimers, enzymatic‐selective synthesis of platform monomers from natural lignin has not been accomplished. In this study, we successfully achieved highly specific synthesis of aromatic monomers with a phenylpropane structure directly from natural lignin using a cascade reaction of β‐O‐4‐cleaving bacterial enzymes in one pot. Guaiacylhydroxylpropanone (GHP) and the GHP/syringylhydroxylpropanone (SHP) mixture are exclusive monomers from lignin isolated from softwood ( Cryptomeria japonica ) and hardwood ( Eucalyptus globulus ). The intermediate products in the enzymatic reactions show the capacity to accommodate highly heterologous substrates at the substrate‐binding sites of the enzymes. To demonstrate the applicability of GHP as a platform chemical for bio‐based industries, we chemically generate value‐added GHP derivatives for bio‐based polymers. Together with these chemical conversions for the valorization of lignin‐derived phenylpropanone monomers, the specific and enzymatic production of the monomers directly from natural lignin is expected to provide a new stream in “white biotechnology” for sustainable biorefineries.",
"conclusion": "Conclusions An enzymatic strategy for the highly selective production of aromatic monomers with phenylpronanone structure from MWL isolated from hardwood and softwood was demonstrated using five enzymes in one pot. The recognition capacity of the broad substrate range by these enzymes may shed light on the mechanism by which the enzymes cleave β‐O‐4 linkages in natural lignin preparations that show high structural heterogeneity. The selectivity of the enzymatic one‐pot reaction under mild conditions is appealing. Still, the process presents problems to overcome, including the costs of the enzyme and cofactor production, wastewater treatment, and utilization of the remaining lignin fraction. All the derivatives obtained in this study retain the aromatic rings, phenolic hydroxyl groups, and methoxy groups in their structure. This finding suggests that these compounds have various applications as lignin‐derived materials. The chemical conversions in this study demonstrated that the enzymatically‐produced phenylpropanone monomers will be the platform aromatic chemicals for sustainable industries based on woody biomasses.",
"introduction": "Introduction Biorefineries for obtaining sustainable energy and chemical production from renewable biomass have been the focus of intensive research and development owing to the depletion of world petroleum reserves together with global warming due to anthropogenic greenhouse gas emissions. 1 In particular, lignocellulosic biomass such as woody feedstock, agricultural waste, and perennial grass can be used as raw materials because they are abundant and not used for food and feed production. 2 , 3 , 4 \n The main constituents of lignocellulosic biomasses are cellulose, hemicellulose, and lignin. 5 Processes for converting cellulose and hemicellulose to bioethanol and chemicals have been developed, and they play a central role in commercial practices worldwide. 4 , 6 Although lignin is considered to be the major renewable source of high‐value aromatic compounds because of its intrinsic polyaromatic chemical structure, its use has been mostly limited to low‐value applications such as solid fuels and admixtures for concrete. 7 , 8 Many efforts have been made to develop thermochemical, catalytic, and enzymatic strategies for the efficient production of high‐value low‐molecular‐weight aromatic compounds from lignin. 8 , 9 , 10 , 11 , 12 Although substantial yields have been achieved using synthetic lignin‐model compounds, the yield and product distribution obtained from natural lignin depend highly on the specific lignin structure, which is often modified and repolymerized in complex ways during preparation. 13 The products of this process are almost always highly heterogeneous and thereby hinder lignin valorization. 14 \n For example, Lancefield et al. 15 reported an isolation method for simple aromatic monomers with a phenylpropanone structure via selective oxidation of benzylic alcohol at the Cα position, followed by reductive cleavage of the β‐O‐4 linkages using zinc as the reductant. Rahimi et al. 16 reported that the formic‐acid‐induced preoxidized lignin depolymerization produced aromatic monomers with structural variations, including diketones, aldehydes, and carboxylic acids. Although these methods offer some technological applications in definable aromatics production, they require toxic compounds, heavy metals, and high‐temperature reactions. Thus, more environmentally friendly methods need to be developed. An alternative for the specific production of aromatic monomers from lignin is enzymatic catalysis. Several enzymes can catalyze the selective cleavage of β‐O‐4 linkages in lignin‐model dimers. 17 Aromatic monomers were produced from lignin‐model synthetic dimers via a cascade reaction comprising three steps, which involved the use of multiple enzymes derived from the Sphingobium sp. strain SYK‐6. The cascade required six reactions for the conversion of racemic lignin‐model dimers with two chiral carbons at the α and β positions into respective monomers. In addition to the enzymes from the strain SYK‐6, a homology‐based amino‐acid database search for β‐O‐4‐cleaving enzymes were identified from two Novosphingobium genomes. 18 , 19 All six enzymatic reactions in the cascade were strictly stereospecific. 20 , 21 , 22 Prior to our study, no enzyme capable of efficiently reacting with the β( S )‐isomer in the last step has been identified. Owing to the lack of genetic information regarding the enzyme responsible for the conversion of the β( S )‐intermediates into aromatic monomers, a recombinant enzymatic process capable of producing aromatic monomers could not be developed. 23 Currently, aromatic monomer production from lignin preparations using enzymes has not been successful and has resulted in only a trace amount of aromatics, such as ferulic acid and vanillin from alkali lignin. 23 \n Recently, we reported that a combination of six enzymes produced using genes from a marine Novosphingobium strain, which was isolated from sunken wood in Suruga Bay, Japan, 24 entirely converted a racemic lignin‐ model dimer into its respective monomer in three steps (Scheme 1 ). 17 , 25 Two short‐chain dehydrogenase/reductases (SDRs; EC 1.1.1.–) (SDR3, SDR5) and two glutathione S ‐transferases (GSTs; EC 2.5.1.18) with β‐etherase activity (GST4, GST5) that catalyze the first and second steps, respectively, were strictly stereospecific. Surprisingly, GST3 catalyzed the third step by efficiently removing glutathione from both compounds 5 and 6 (Scheme 1 ), which were produced from a racemic lignin‐model dimer (Scheme 1 a mixture of compounds 2 and 3 ). The discovery of the nonstereoselective enzyme GST3 paved the way to convert all stereoisomers of compound 1 to their respective monomers for the first time. However, a feasible method of processing natural lignin of biomass origins using these five enzymes remains unknown in terms of reactivity and selectivity for highly heterologous natural lignin substructures. Scheme 1 Enzymatic cascade for GHP (compound 7 ) synthesis from lignin‐model dimers (GGGE, compound 1 ) via MPHPV (compound 2 , 3 ). The responsible enzymes and their required cofactors are shown. Abbreviations: SDR—short‐chain dehydrogenase–reductase, GST—glutathione S ‐transferase; GS‐GHPgst4 (compound 5 )—glutathione conjugate of GHP produced by β‐O‐4 bond cleavage and removal of guaiacol (compound 4 ) by GST4; GS‐GHPgst5(compound 6 )—glutathione conjugate of GHP produced by GST5; NAD + —oxidized form of nicotinamide adenine dinucleotide (NAD); GSH—reduced form of glutathione; and GSSG—oxidized form of glutathione. Protein accessions in the DDBJ/EMBL/GenBank database: SDR3 (GAM05523), SDR5 (GAM05547), GST3 (GAM05529), GST4 (GAM05530), and GST5 (GAM05531). In this study, we investigated the optimal reaction conditions for GST3. Using the five enzymes shown in Scheme 1 , we enzymatically produced phenylpropanone monomers from isolated natural lignin. To elucidate the catalytic mechanism of the β‐O‐4 cleavage of natural lignin, we analyzed the intermediates produced from lignin. Finally, we assessed the future prospects of these phenylpropanone monomers as new platform chemicals derived from natural lignin.",
"discussion": "Results and Discussion Optimal pH and temperature for GST3 activity We identified the optimal pH values and temperatures for the activity and kinetic parameters of SDRs (SDR3, SDR5) and GSTs (GST4, GST5). 25 In contrast, GST3 characterization has not been directly performed because the substrates for GST3, glutathione‐conjugated guaiacylhydroxylpropanones (GS‐GHPs) (Scheme 1 , compounds 5 and 6 ), have not been commercially available. Here, we enzymatically produced GS‐GHPs from racemic 3‐hydroxy‐1‐(4‐hydroxy‐3‐methoxyphenyl)‐2‐(2‐methoxyphenoxy)‐1‐propanone [(2‐methoxyphenoxy)hydroxypropiovanillone; MPHPV] (Scheme 1 , compounds 2 and 3 ) using GST4 and GST5. The reaction products were designated as GS‐GHPgst4 and GS‐GHPgst5 (Scheme 1 , compounds 5 and 6 ), respectively. Then, we purified each of the GS‐GHPs. Using purified each of the GS‐GHPs as a substrate, we investigated the effect of pH value on GST3 activity. The optimum pH value and temperatures for GST3 activity were approximately 8.0 (Figure 1 a, b) and 25–30 °C (Figure 1 c, d), respectively, both for GS‐GHPgst4 and GS‐GHPgst5. The specific activities in 0.1 M N ‐Tris(hydroxymethyl)methyl‐3‐aminopropanesulfonic acid buffer (TAPS) at pH 8.5 and 15 °C were 3.9 and 9.1 U mg −1 protein for GS‐GHPgst4 and GS‐GHPgst5, respectively.\n Figure 1 pH–activity curves of the purified GST3 using (a) GS‐GHPgst4 and (b) GS‐GHPgst5 as the substrates. The buffers used were 0.1 m 2‐( N ‐morpholino)ethanesulfonic acid (MES; pH 5.5–7.0; •), 3‐[ N ‐morpholino]propanesulfonic acid (MOPS; pH 7.0–8.0; □), TAPS (pH 8.0–9.0; ▴), N ‐cyclohexyl‐2‐aminoethanesulfonic acid (CHES; pH 9.0–10.0; ○), N ‐cyclohexyl‐3‐aminopropanesulfonic acid (CAPS; pH 10.0–11.0, ▪). The activity was measured in the buffers including 25 °C. The values are shown as percentages of the maximal activity of GST3 observed at pH 7 for GS‐GHPgst4 and pH 8 for GS‐GHPgst5, which are taken as 100 %. The temperature–activity curves of the purified GST3 using (c) GS‐GHPgst4 and (d) GS‐GHPgst5 as the substrates are also shown. The values are shown as percentages of the maximal activity of GST3 observed at 25 °C for GS‐GHPgst4 and 30 °C for GS‐GHPgst5, which are taken as 100 %. GHP (Scheme 1 , compound 7) synthesis via enzymatic cascade reactions from milled wood lignins Each enzyme retained more than 80 % of the respective maximal activity at pH 8.5. 25 The optimal temperature of SDR3 was 15 °C, lower than those of the other enzymes, indicating the instability of SDR3 at a higher temperature. SDR3 was thought to be the limiting reaction step in the enzymatic cascade reactions. Accordingly, the following enzyme reactions were conducted under optimal conditions for SDR3 activity (0.1 M TAPS at pH 8.5 and 15 °C) using the enzymes simultaneously in one pot. Milled wood lignins (MWLs) from softwood and hardwood were prepared from Japanese cedar ( Cryptomeria japonica ) and Eucalyptus globulus wood and are here designated C‐MWL and E‐MWL, respectively The main product from C‐MWL was determined to be GHP according to the retention time ( t \n R ) on the liquid chromatography–mass spectrometry (LC–MS) chromatograph, observed molecular mass, calculated elemental composition, fragmentation pattern on mass spectra, and UV spectrum using authentic GHP as a reference (Figure 2 a, c and Figure S5a, c in the Supporting Information). The yield of GHP was 2.4±0.005 wt % [24±0.05 mg g lignin \n −1 ]. The main products from E‐MWL were determined to be GHP and syringylhydroxylpropanone (SHP) using authentic GHP and SHP as references (Figure 2 b, c and Figure S5 b, c in the Supporting Information). The yields of the two compounds were 1.9±0.012 and 4.7±0.025 wt % [19±0.12 and 47±0.25 mg g lignin \n −1 ], respectively. To the best of our knowledge, this is the first report of enzymatic production of phenylpropanone monomers from MWL.\n Figure 2 One‐pot enzymatic production of GHP and SHP from (a) C‐MWL and (b) E‐MWL. Total ion chromatograms obtained from LC–MS analysis of the reaction of five enzymes (SDR3, SDR5, and GST3‐5) with the MWLs and cofactors are shown (top). The reactions were conducted under the same conditions without enzymes but with cofactors to assess the non‐enzymatic production of GHP/SHP (bottom). (c) Authentic GHP (top) and SHP (bottom) analyzed under the same conditions. There have been numerous studies on the production of aromatic monomers from lignin. 7 , 8 , 9 For example, pyrolysis and cracking processes have been extensively studied for decades. The major products from different types of pyrolysis processes are complex mixtures of deoxygenated aromatic monomers such as vinylphenols, toluene, xylene, and many other aromatics along with gaseous products, that is, hydrocarbons of low molecular weight. For the cracking process, a high yield of approximately 80 wt % was recorded for a mono‐/oligomeric phenolic mixture. The common drawback of pyrolysis and cracking processes is the required harsh operating conditions (a high temperature ranging between 300–650 °C under high pressures) and side reactions that are difficult to control. To overcome these problems, hydrogenolysis and hydrolysis under subcritical/supercritical conditions with various catalysts operating at temperatures below 300 °C have been attempted and have drawn attention for their potential to be viable new methods. 9 As an example of a significant study in this field, nickel‐based catalysts were used for the selective production of propylguaiacol and propylsyringol from sawdust with approximately 50 wt % of lignin conversion at 200 °C. 26 The supply of hydrogen and recycling of the catalysts is a major issue in the process that has to be addressed. Another excellent contribution in chemical catalysis was reported by Lancefield et al. 15 The production of the phenolic monomers (GHP and SHP) was conducted at 80 °C using a chlorinated oxidizer and zinc as a reductant under atmospheric pressure in the presence of air. The production yields of GHP and SHP from a birch lignin preparation were 0.46 and 4.6 wt %, respectively. Although these values depend strongly on the biomass source and process of lignin preparation, 13 the yields of enzymatically produced GHP and SHP in this study were comparable to those obtained by chemical catalysis using toxic halogenated aromatics and heavy metals. The theoretical yield of these monomers (GHP plus SHP) was calculated to be approximately 12 wt % based on the content of releasable β‐O‐4 linkages. 15 The experimental yields obtained by Lancefield et al. and by us reported herein were similar and approximately half of the theoretical value. The yield limitation in the chemical catalytic process was estimated to be owed to the formation of by‐products that include unknown repolymerized products. We believe that in our ongoing and future studies it should be possible to increase the yields obtained in our study by improving the catalytic property of the enzymes, for example, by improving their catalytic efficiencies and substrate recognition capacities as well as improving their protein stabilities through protein engineering. Furthermore, conducting the enzymatic process under mild conditions shows promise for avoiding repolymerization, which is likely the cause for the formation of undesired by‐products. Enzymatic production of aromatic monomers is an attractive alternative to thermal and chemical conversion of lignin. Many microbial enzymes have been reported to have the ability to produce lignin‐related aromatic monomers. Although fungal and bacterial peroxidases can break the linkages of inter lignin monomers using extracellular radical molecules, the enzyme activity and specificity is low and can even cause repolymerization during the enzymatic reaction. Other bacterial enzymes have been examined to produce aromatic monomers including vanillin, cafeic acid 4‐vinylguaiacol, coniferyl alcohol, and GHP using ferulic acid, 27 eugenol, 28 and lignin‐model dimers 22 , 23 as starting materials; however, these enzymes have not successfully been used for natural lignin preparations. In contrast, there have been a few reports demonstrating microbial production of aromatic monomer directly from lignin. For example, metabolically engineered Rhodococcus jostii RHA1 reportedly produced vanillin up to 96 mg mL −1 after a 6‐day cultivation in the media containing lignocellulose. 29 Although this study demonstrated the possibility of biological production of the aromatic monomer, vanillin is already produced chemically from lignosulfonate at industrial scale. 9 In our study, the enzyme reactions were conducted with MWLs suspended in an aqueous buffer with a small amount of organic solvent. At present, we use N , N ‐dimethylformamide to dissolve MWL. However, in a scaled‐up production we propose that a safer and greener solvent should be used instead. Biomass‐based solvents with lower boiling points than the products, such as acetone, butanol, or tetrahydrofuran, 30 are potential candidates to increase the sustainability of the enzymatic production of GHP and SHP. In general, the productivity must reach 1–3 g L −1 h −1 for the products to reach economic viability for bio‐based chemicals. 31 To meet this goal, the productivity (approximately 5 mg L −1 h −1 of the products in the current experimental setting) must be increased by 2–3 orders of magnitude by operating at higher volumetric concentrations. In addition, it is necessary to enhance the enzymatic activity and stability by protein engineering and stoichiometric optimization of the multiple reactions 32 to reduce the loadings on the enzymes and improve the yields. In addition, recycling of the cofactors of NAD + and GSH as well as an eco‐friendly pretreatment for biomasses suitable for enzymatic reactions are needed for the cascade presented in this study to become economically sound. Product recovery was conducted in this study by liquid‐phase extraction using ethyl acetate as an organic solvent. The other practical product recovery options need to be assessed and optimized because the product recovery processes accounts for a major part of a total cost and is critical for achieving sustainable production of bio‐based chemicals. 33 In addition, the residual insoluble and soluble lignin fractions have to be recovered and analyzed for efficient utilization for entire lignin valorization. Although there are many problems to overcome, further optimization of each process in the upstream and downstream processes are possible and provide grounds for further study. Analysis of substrate recognition tolerance of the enzymes for MWL substructures The bacterial enzymes involved in the cleavage of β‐O‐4 ether linkages have been considered to show substantial activity only on dimeric lignin‐model synthetic compounds, with low activity on polymeric substrates. 18 However, our enzyme system produced phenylpropanone monomers. To elucidate the reaction mechanism of the enzymes displaying substrate recognition tolerance for unknown substructures present in MWL, we analyzed the reaction products from C‐ and E‐MWL by the four enzymes (SDR3, SDR5, and GST4‐5). The glutathione conjugates produced by the enzymes were analyzed by LC–MS (Figure S6 a, b in the Supporting Information). The glutathione conjugates obtained from authentic guaiacylglycerol‐β‐guaiacyl ether (GGGE, Scheme 1 , compound 1 ) by the same enzyme set were analyzed under the same conditions as the references (Figure S6 c in the Supporting Information). The observed molecular masses in the LC–MS analyses and calculated elemental compositions of the main two reaction products ( t \n R ; 1.69 and 1.97 min) from C‐MWL were the same, m / z 500.13 (Figure S7a in the Supporting Information) and C 20 H 26 N 3 O 10 S, in agreement with those of the deprotonated parent molecular ion of glutathione‐conjugated GHP [GS (elemental composition, C 10 H 16 N 3 O 6 S)‐C 10 H 11 O 4 ]. The t \n R on LC–MS chromatograms, observed molecular mass, and fragmentation patterns on the mass spectra of the major products from C‐MLW agreed well with those of the products from the enzyme reactions using GGGE as a substrate (Figure S7 c in the Supporting Information). From these results, the two peaks were inferred to be isomers. Four major products were produced in the case of E‐MWL (Figure S8b in the Supporting Information). MS data from two peaks ( t \n R ; 1.64 and 1.96 min) were the same as those from C‐MWL, and the others ( t \n R ; 1.86 and 2.03 min) were molecules with m / z 530.14 and calculated compositions of C 21 H 28 N 3 O 11 S, in agreement with those of the deprotonated parent ion of glutathione‐conjugated SHP (GS‐C 11 H 13 O 5 ). The MS spectra contained several fragment ions, such as m / z 272.09 and 254.08 (Figure S8 a–c in the Supporting Information), in common with the MS spectrum obtained from authentic glutathione (Figure S7 d in the Supporting Information). These fragment ions were speculated to have been derived from the conjugated glutathione. LC–MS/MS analyses revealed that the reactions using C‐ and E‐MWL as the substrates produced numerous molecules with different molecular size, providing deprotonated molecular ions ranging from m / z 500.13 to 964.30 (Figure S8 in the Supporting Information). The observed molecular masses, fragment ions, and calculated elemental compositions are listed in Table 1 . These results suggested that these enzymes could accommodate not only lignin‐model dimers but also lignin oligomers consisting of approximately 10–30 carbon atoms as their substrates and cleave the β‐O‐4 linkages in natural lignin with highly heterologous substructures.\n Table 1 Glutathione‐conjugated intermediates obtained from MWLs produced by enzymatic reaction (SDR3, SDR5, GST4, and GST5) detected by LC–MS/MS. Detected parent ion [ M ‐H] − ; [HRMS, m / z ] [a] \n Source Retention time [b] [min] Detected fragment ion [c] [ m / z ] Calculated elemental composition [d] \n 500.1342 C‐MWL 1.59, 1.94 482, 470, 464, 272 , 254 , 210 , 179 , 143 , 128 \n C 20 H 27 N 3 O 10 S [GS‐C 10 H 11 O 4 ] 500.1318 E‐MWL 1.52, 1.93 482, 470, 464, 272 , 254 , 210 , 179 , 143 , 128 \n C 20 H 27 N 3 O 10 S [GS‐C 10 H 11 O 4 ] 530.1464 E‐MWL 1.95 512, 500, 494, 383, 272 , 254 , 239, 210 , 179 , 143 , 128 \n C 21 H 29 N 3 O 11 S [GS‐C 11 H 13 O 5 ] 636.1857 C‐MWL 2.31 363, 306 , 288 , 272 , 137 ND 678.1962 C‐MWL 2.58 660, 648, 642, 272 \n C 30 H 37 N 3 O 13 S [GS‐C 20 H 21 O 7 ] 678.1972 E‐MWL 2.57 660, 648, 272 \n 706.1913 E‐MWL 2.51 688, 419, 272 , 254 \n ND 738.2160 E‐MWL 2.43 720, 306 , 272 , 254 \n ND 856.2609 C‐MWL 2.92 855, 838, 306 \n ND 874.2701 C‐MWL 2.56 856, 306 , 272 \n C 40 H 49 N 3 O 17 S [GS‐C 30 H 33 O 11 ] 874.2723 E‐MWL 2.56 856, 843, 838, 802, 782, 519, 466, 321, 272 , 207, 143 , 140 C 40 H 49 N 3 O 17 S [GS‐C 30 H 33 O 11 ] 932.3058 E‐MWL 3.06 \n 306 , 305, 272 \n ND 964.3091 E‐MWL 2.55 963, 946, 782, 536, 306 , 272 \n ND [a] HRMS: high‐resolution MS. [b] The MS chromatograms obtained from LC–MS/MS analysis after enzymatic reactions with C‐ and E‐MWLs are shown in Figure S9 in the Supporting Information. [c] The fragment ions derived from glutathione (Figure S8 d in the Supporting Information) are underlined. [d] ND: not determined; GS: conjugated glutathione (elemental composition: C 10 H 16 N 3 O 6 S−). Wiley‐VCH Verlag GmbH & Co. KGaA GHP transformation into various derivatives We investigated the chemical conversion of GHP to show the feasibility of effective lignin use with a wide range of applications (Scheme 2 ). Scheme 2 Chemical synthesis that yields functional chemicals from GHP ( 7 ), a key platform chemical. Reaction conditions: (a) NaBH 4 , 26 °C, 24 h, 79 % for GPD (compound 8 ); (b) methanesulfonic acid, 65 °C, 22.5 h, 49 % for BGP (compound 9 ); (c) isopropanol and trimethylamine, 85–90 °C, 48 h, 79 % for coniferyl alcohol (compound 10 ); (d) HCl, 90 °C, 96 % for 3‐chloro‐1‐(4‐hydroxy‐3‐methoxyphenyl)‐1‐propanone (compound 11 ); and (e) sodium ethoxide, RT, 1.5 h, 61 % for GVK (compound 12 ). The simple and convenient 1‐guaiacyl‐1,3‐propanediol (GPD, 1‐(4‐hydroxy‐3‐methoxyphenyl)‐1,3‐propanediol, Scheme 2 , compound 8 ) synthesis from GHP was achieved by the reaction of GHP with NaBH 4 (Scheme 2 ). The reaction proceeded completely at room temperature in aqueous NaOH to afford GPD in 79 %. The pH adjustment to approximately 6–9 was important for the efficient separation of GPD from the reaction mixture. Some methods for synthesizing GPD have been reported, but multi‐step processes are required. 34 , 35 \n 3,3‐Bisguaiacyl‐1‐propanol (BGP, 3,3‐bis(4‐hydroxy‐3‐methoxyphenyl)‐1‐propanol,Scheme 2 , compound 9 ) was synthesized from GPD under acidic conditions. We found that GPD shows high reactivity with both acids and alkalis. In the presence of an acid catalyst, its reactivity leads to the formation of BGP (Scheme 2 b, Figure S2 a, b in the Supporting Information), a bisphenol that may be used as a raw material for epoxy resins; as a hardener for epoxy and urethane resins; and as a raw material for new functional polyesters and polycarbonates. In the earlier study, 36 the reaction of GPD with an excess of phenol was performed using hydrochloric acid to give 3‐(hydroxyphenyl)‐3‐(4‐hydroxy‐3‐methoxyphenyl)‐1‐propanol, an unsymmetry compound and possibly a mixture, in view of the two reaction sites of phenol. Coniferyl alcohol (Scheme 2 , compound 10 ) synthesis from GPD under basic conditions. Coniferyl alcohol is a monolignol, which is a precursor of lignin biosynthesis, 37 and has important applications in research areas as an artificial lignin. Coniferyl alcohol can be converted to valuable aromatic compounds, such as medicines for cancers and diabetes; functional foods and cosmetics with antioxidizing activity; and pesticides. 38 We achieved the formation of coniferyl alcohol from GPD under heating conditions using triethylamine as the base and isopropanol as the solvent (Scheme 2 c). The choice of base and solvent was crucial because coniferyl alcohol tends to polymerize under basic conditions, particularly in aqueous solvents. The yield was higher with an organic base such as triethylamine. Methanol or ethanol as the solvent led to the formation of 3‐(4‐hydroxy‐3‐methoxyphenyl)‐3‐alkoxy‐1‐propanol (an alkoxylated compound) as the main product. Guaiacylvinylketone (GVK, 1‐(4‐hydroxy‐3‐methoxyphenyl)‐2‐propene‐1‐one, Scheme 2 , compound 12 ) synthesis was performed from GHP according to a method used for 1‐(4‐hydroxyphenyl)‐2‐methyl‐2‐propene‐1‐one synthesis from 1‐(4‐hydroxyphenyl)‐3‐hydroxy‐2‐methyl‐1‐propanone. 39 The reaction of GHP with hydrochloric acid was conducted at approximately 70 °C to provide a high yield of 3‐chloro‐1‐(4‐hydroxy‐3‐methoxyphenyl)‐1‐propanone (96 %)(Scheme 2 d, compound 11 ). The chlorinated product was treated with sodium ethoxide in ethanol to GVK in 61 % (Scheme 2 e, compound 12 ). We found that this dehydrochlorination reaction also proceeded in NaOH aqueous solution under mild conditions, between room temperature and approximately 60 °C. In our preliminary polymerization experiment of GVK using azobisisobutyronitrile and tetrahydrofuran as the radical initiator and solvent, respectively, a solid was obtained. The 1 H‐NMR spectrum of the compound confirmed the presence of aliphatic methylene and methine protons and the absence of vinyl protons (Figure S4 in the Supporting Information), indicating polymerization via the vinyl group."
} | 7,079 |
38167520 | PMC10762816 | pmc | 1,165 | {
"abstract": "Background The anaerobic digestion process degrades organic matter into simpler compounds and occurs in strictly anaerobic and microaerophilic environments. The process is carried out by a diverse community of microorganisms where each species has a unique role and it has relevant biotechnological applications since it is used for biogas production. Some aspects of the microbiome, including its interaction with phages, remains still unclear: a better comprehension of the community composition and role of each species is crucial for a cured understanding of the carbon cycle in anaerobic systems and improving biogas production. Results The primary objective of this study was to expand our understanding on the anaerobic digestion microbiome by jointly analyzing its prokaryotic and viral components. By integrating 192 additional datasets into a previous metagenomic database, the binning process generated 11,831 metagenome-assembled genomes from 314 metagenome samples published between 2014 and 2022, belonging to 4,568 non-redundant species based on ANI calculation and quality verification. CRISPR analysis on these genomes identified 76 archaeal genomes with active phage interactions. Moreover, single-nucleotide variants further pointed to archaea as the most critical members of the community. Among the MAGs, two methanogenic archaea, Methanothrix sp. 43zhSC_152 and Methanoculleus sp. 52maCN_3230, had the highest number of SNVs, with the latter having almost double the density of most other MAGs. Conclusions This study offers a more comprehensive understanding of microbial community structures that thrive at different temperatures. The findings revealed that the fraction of archaeal species characterized at the genome level and reported in public databases is higher than that of bacteria, although still quite limited. The identification of shared spacers between phages and microbes implies a history of phage-bacterial interactions, and specifically lysogenic infections. A significant number of SNVs were identified, primarily comprising synonymous and nonsynonymous variants. Together, the findings indicate that methanogenic archaea are subject to intense selective pressure and suggest that genomic variants play a critical role in the anaerobic digestion process. Overall, this study provides a more balanced and diverse representation of the anaerobic digestion microbiota in terms of geographic location, temperature range and feedstock utilization. Supplementary Information The online version contains supplementary material available at 10.1186/s40793-023-00545-2.",
"conclusion": "Conclusions In this study, the metagenomic characterization of the microbial species involved in the AD process has been expanded through the analysis of a large number of different reactor types operated under a range of conditions. In addition to expanding the number of species reported in the previous version of the Biogas Microbiome database by almost three times, the analysis was focused on archaea, one of the crucial components of the microbiome. The investigation of gene composition has led to a better characterization of archaea and their methanogenic metabolism; however there still remains a degree of uncertainty in the automatic association between gene composition and phenotype, which will require the development of new investigation methods based on gene expression or machine learning. Despite this huge increase in the number of catalogued species, the great diversity of this biotechnological niche has yet to be fully explored, especially with regard to bacterial species. Inspection of the phage repertoire provided a first overview and, through the analysis of the CRISPR elements, a first characterization of phage-bacterial interactions and co-evolution. This analysis allowed to build the first version of the viral Biogas Microbiome database, currently represented only by DNA phages. The RNA phage fraction still remains to be identified and will be one of the next targets. While the viral component characterisation proved to be extremely complex, the presence of a combined database of prokaryotes and phages will certainly allow in the future a better tracking of their interactions with prokaryotes, also via means of cross-linking techniques and co-occurrence. Abundance of microbial species competing with methanogens for H 2 utilization, such as SRBs, has highlighted how a well-characterized MAG database allows to better understand the impact of the microbiome in reducing the performance in terms of biogas production. Finally, the investigation of SNVs impacting the genes involved in key functional processes has laid the foundations to study the evolution of AD microbiome and its role in reactor performance, suggesting that methanogenic archaea are under strong selective pressure.",
"introduction": "Introduction The increasing demand for energy and the depletion of fossil fuels have shifted attention towards alternative energy production processes. One such process is anaerobic digestion (AD), which has a lower environmental impact, while supporting the concept of circular economy by utilizing a variety of end-products including food, agricultural, industrial and municipal wastes. AD is a natural process that breaks down complex organic matter into simpler compounds, occurring in ecological niches with low oxygen or strictly anaerobic conditions, such as bogs, sediments and the guts of herbivores [ 1 ]. This process is also utilised at industrial-scale activities for biogas production. The breakdown is carried out by a microbial community that can range in complexity from a few species to an extremely complex microbiome consisting of thousands of species [ 2 – 4 ]. The biogas obtained from industrial reactors typically contains a mixture of methane (CH 4 ) and carbon dioxide (CO 2 ), with trace levels of hydrogen sulfide (H 2 S), ammonia (NH 4 + ), hydrogen (H 2 ), and various volatile organic compounds, depending on the feedstock and on the functional activity of the microbiota [ 5 ]. From a biotechnological perspective, CH 4 is the most significant constituent of the biogas generated during the methanogenesis step of the AD process and is produced by methanogenic Archaea [ 6 , 7 ]. AD is carried out by a diverse community of microorganisms, where each species has a unique role in highly specialized and complex microbiomes [ 8 ]. First, complex organic matter is transformed by hydrolytic bacteria into soluble organic compounds. Second, acidogenic bacteria transform the latter into intermediates (e.g., volatile fatty acids), with syntrophic acetogenesis playing an essential role in breaking compounds down into the simplest molecules. Third, methanogenesis is performed by archaea using acetate, methylamine and/or CO 2 and H 2 to produce CH 4 . Understanding the composition and role of each species in the community is crucial for a better understanding of the carbon cycle in anaerobic systems and improving biogas production. However, the isolation of many species using classical microbiological techniques can be challenging, making metagenomics an ideal alternative for characterizing the community complexity. In addition to the microbiome, the efficiency of AD is dependent on several interconnected factors, including feedstock composition, temperature, organic loading rate, hydraulic retention time, and other physicochemical parameters [ 9 , 10 ]. Controlling these parameters can offer unique opportunities for microbial selection or manipulation to improve the process efficiency. The rise of metagenomics in the last two decades has revealed the important role viruses play in shaping microbial communities [ 11 ]. In many environments viruses are responsible for selective pressure, lateral gene transfer, and nutrients recycling, all of which impact the microbiota. Understanding their role at community level is crucial and offers an opportunity to fine-tune the AD process. Despite the important role viruses play, the viral community of AD has received little attention [ 12 , 13 ]. This gap is due to difficulties associated with viral metagenomics compared to the prokaryote investigation. To separate the viral fraction of the community from the microbial is challenging, as viruses represent a small fraction of the total genetic material. Therefore, viruses are more understudied than bacteria and archaea. However, research has shown that phages, in particular, are involved in microbiome dynamics and process stability, regulating microbial abundance and diversity in full-scale biogas units [ 14 ]. Metagenome sequencing has become a valuable tool to gain insights into the genetic repertoire of non-cultivable biogas community members [ 15 ]. Advances in sequencing throughput and computational techniques nowadays allow the recovery of Metagenome-Assembled Genomes (MAGs) from highly diverse environments [ 16 ]. These MAGs are obtained through binning together assembled contigs with similar sequence composition, depth of coverage, and taxonomic affiliations [ 17 – 19 ]. Despite limitations regarding completeness and contamination [ 20 ], MAGs are useful proxies for studying microbial and viral genomes present in the system, providing insights into taxonomy, functional properties, and dynamics of the microbiome. Several studies have attempted to gain insights into the AD microbiome for biogas production [ 15 , 21 ], but the limited geographic distribution of the samples and the lack of a global analysis on the viral fraction have prevented a complete characterization of the microbiome. To address this gap, a global meta-analysis study was performed to tentatively assess the impact of physicochemical parameters, microbiota, and reactor characteristics on the AD process. The aim of this study was to complement and consolidate previous results and establish a more comprehensive reference database of microbial and viral genomes. In addition, strain-resolved metagenomics was applied to reveal fine-scale evolutionary mechanisms, functional dynamics, and strain-level metabolic variation, potentially contributing to the selection within a microbial community.",
"discussion": "Results and discussion Expansion of the global anaerobic digestion microbiome database To create a more balanced representation of MAGs from various countries, the previous iteration of the Biogas Microbiome database featured sequences from 18 studies, containing 123 samples [ 15 ]. However, it was biased towards samples collected from Denmark, which accounted for 68% of the database, mostly deriving from laboratory-scale biogas reactors and batch tests. In the present version, we added 24 studies to the database, resulting in a total of 314 samples. Notably, a significant number of samples and MAGs, derived from Chinese biogas plants, were included in a recent study and implemented in the new database [ 21 ]. This expanded Biogas Microbiome database now encompasses samples collected in eleven countries (Fig. 1 , Additional file 1 ), with China and Denmark being the primary contributors, accounting for 75% of the total number of samples. Samples were categorized according to reactor type, with 111 described as “lab scale” or “batches” operated in six countries, 104 described as “full-scale biogas plants”, and 99 as “(semi)continuous lab reactors” or “CSTR” (Additional file 1 ). Fig. 1 Geographic and microbial diversity of the expanded Biogas Microbiome database. The expanded anaerobic digestion microbiome database includes 314 samples distributed across 11 countries, with Denmark and China being the origin of most samples. Information about all the samples can be visualized in Additional file 1 To gather data, only basic statistics were collected from the metadata of different publications. The data were evenly distributed across different temperatures, with 4% being psychrophilic, 51% mesophilic, and 38% thermophilic; 22 samples had no clearly reported operating temperature or reactor type. This study provides a more comprehensive representation of microbiota structure and species growing over a wider range of temperatures, compared to previous investigations [ 15 ]. Temperature is an essential factor in determining the diversity of the community, as it was found to be inversely correlated with Chao1 and Shannon alpha-diversity indices (Pearson’s r = − 0.4, p = 1·10 –12 ) (Additional file 2 : Fig. 1). The three samples with the highest diversity were collected from reactors operated at the lowest temperature, such as Ma_2021_BGP_27, 29, and 34, with temperatures ranging from 14 to 17 °C and having Chao1 ∼4,566 and Shannon ∼6.37 (Fig. 1 ). Previous studies have reported that psychrophilic temperatures have more diversity of methanogenic Archaea [ 54 ] and fluctuation in rare biosphere taxa [ 55 ] than mesophilic reactors, which represented the majority of the samples in this study. However, psychrophilic temperatures generally slow down the metabolic activity of anaerobic microorganisms. This can result in a prolonged retention time for organic matter in the digester and lead to a decrease in methane yield per unit of organic matter compared to mesophilic or thermophilic conditions [ 56 , 57 ]. Prokaryotic community composition The binning process generated 11,831 MAGs, which, when combined with the database by Ma and colleagues [ 21 ], resulted in a total of 26,612 MAGs. To reduce redundancy, 4,568 MAGs were selected for clustering based on ANI calculation and quality analysis with CheckM. Of these, 2,217 (48.5%) were classified as high quality (completeness > = 90%; contamination < = 10%) while 2,351 (51.5%) were of medium quality (90% > completeness > = 50%; contamination < = 10%) (see Availability of Data and Materials). These results were validated with CheckM2 and presented a high correlation with those of CheckM (R 2 = 0.7). This concordance is especially relevant as it demonstrates the accuracy of CheckM2 on a large dataset, while still validating quality assessments obtained with CheckM. Specifically, the estimated completeness was higher for 59% of the MAGs, lower for 40.8%, and 0.24% showed the same value as CheckM. Notably, out of the 387 Candidatus MAGs, 293 (75.7%) had higher completeness scores with CheckM2, and 57 (14.7%) had completeness scores that were two times higher than CheckM. Additionally, 76% of the samples had more than 50% mapped reads, indicating that this version of the Biogas Microbiome database predominantly characterizes the microbiome. The batch experiment enriched culture with cellulose (Jia_2018_Batch_12 and 13) [ 58 ] had the highest proportion of mapped reads, with about 94% of bacterial MAGs mapped. Taxonomic investigation on the prokaryotic community revealed that members of Firmicutes dominate, comprising 38% of the MAGs in the database, followed by Proteobacteria (12%) and Bacteroidetes (10%). Archaea are represented by 198 species (4.3%), mainly from the Euryarchaeota phylum (76.3%), while Candidatus Bathyarchaeota and Candidatus Diapherotrites are represented by 13 and 12 members, respectively. Previous metataxonomy-driven investigations [ 15 ] identified 53 MAGs belonging to the Euryarchaeota phylum while the other Candidatus phyla were not represented. Moreover, among the archeal population, Euryarchaeota were similarly identified as the predominant phylum in another large-scale study examining the microbiome of 80 anaerobic digesters [ 59 ]. Among all the MAGs identified, only a small number (3.7%) were assigned at species level (152 Bacteria and 19 Archaea), while 19.3% were assigned at genus level (794 Bacteria; 87 Archaea). These results confirm that the fraction of archaeal species already characterized at genome level and reported in public repositories of microbial genomes is higher (9.6%) than that of bacteria (3.5%), but still very low. Analysis performed on MAG relative abundance in all the samples provided insight into the distribution of microbial species in the database. By calculating the number of samples in which each MAG abundance was greater than 0.01, 0.1, and 1%, Firmicutes, Euryarchaeota, Bacteroidetes, Proteobacteria, and Synergistetes were identified to be more widespread, while others, such as Fibrobacteres, Ignavibacteriae, Crenarchaeota, and many Candidatus phyla had a more scattered distribution (Additional file 2 : Fig. 2). Similar results were obtained using 0.1% or 0.01% as the relative abundance threshold to define the presence of a MAG (see Availability of Data and Materials). Strikingly, all the four bacterial strains enriched in previously characterized biofilm communities developed on the gas injection systems of biogas upgrading reactors were among the top widespread MAGs across samples at all relative abundance thresholds [ 60 ]. These MAGs include Firmicutes 50dbBF_058, Firmicutes 50dbBF_049, Firmicutes 50dbBF_057, and Synergistaceae sp. 24abBP_148. In particular, the most widely detected MAG Firmicutes 50dbBF_058 is alternatively classified as Limnochordia DTU010 and was previously found to possess the glycine cleavage system and the glycine synthase-reductase pathway for CO 2 reduction, which are considered important for establishing syntrophic interactions with methanogens [ 60 ]. Other 45 less widespread MAGs were here assigned to Limnochordia DTU010 and several species of the same order have also recently been identified across a numerous independent set of full-scale plants in sensible abundance [ 61 ], further raising our interest in this uncharacterized taxon. Besides, MAGs belonging to the phylum Candidatus Atribacteria were widespread but with a low relative abundance (less than 0.1%). This finding suggests that methanogenic Archaea and Synergistetes, for which only a small number of MAGs have been identified in the database, play crucial roles and are very flexible, being able to adapt to a variety of environmental conditions. The two most widespread Archaea, Methanothrix sp. 43zhSC_152 and Candidatus Methanoculleus thermohydrogenotrophicum 31mySI_10, were detected in 49 and 42 samples, respectively, with relative abundance ≥ 1%. Similarly, the two most common Synergistetes MAGs, Synergistaceae sp. 24abBP_148 and Acetomicrobium flavidum 43zhSC_162, were present in 35 and 15 samples at relative abundance ≥ 1%. In fact, members of the Synergistaceae taxon were previously found to show acetate-oxidizing ability, which may work in syntrophy with hydrogenotrophic methanogens for methane production [ 62 , 63 ]. This crucial cooperation activity may explain the similar common widespread distribution of these two phyla. On the other hand, some Candidate phyla and Planctomycetes have a scattered distribution. Genome size is correlated with the number of genes and, on average, Planctomycetes have the second-highest genome size. This suggests that the ability to colonize many different samples is not necessarily related to species gene content. The highest archaea/bacteria ratio was found in a thermophilic batch reactor fed with a synthetic medium containing methanol (Yan_2020_Batch_2, Additional file 1 ). The Archaea in Yan and colleagues [ 62 ] is represented by one MAG from the Euryarchaeota phylum with 100% completeness which could not be taxonomically assigned with higher specificity, indicating the need for a more detailed taxonomic investigation on some Archaea branches. Phage-prokaryote interactions indicate shared adaptation strategies A total of 79,922 viral genomes were identified after dereplication, and 32% of them were successfully classified. The composition of the viral community was largely dominated by tailed bacteriophages of the class Caudoviricetes , as established by previous works [ 12 , 13 , 64 , 65 ], comprising 97.6% of all viruses. Families Straboviridae , Peduoviridae , and Herelleviridae were the most represented in the dataset, with 2,927, 2,887, and 2,825 members, respectively. However, the family with the greatest coverage in the 268 samples was Peduoviridae , followed by Straboviridae and Drexlerviridae (Additional file 2 : Fig. 3B, see Availability of Data and Materials for full statistics). The virome of the 314 samples analyzed represented between 8 × 10 –4 and 0.86% of the total community. As expected, 11 samples from an enrichment of viromes experiment [ 12 ] had a higher proportion of phages (0.35–0.86%). However, the virus with the highest proportion, virMAG MPIJXP02F1_92297, was found at low abundance in these samples, while thermophilic reactors fed with exhausted gasses and carbohydrates had a higher abundance of this virus. Although PhaGCN was not able to taxonomically identify this phage, the presence of enzymes related to DNA binding domains (pfam13443), transposase related to bacteriophages (COG5421), and DNA-binding transcriptional regulator (COG3655) confirmed its classification as a viral genome. Permanova analysis revealed a significant correlation ( p < 0.01) between the temperature of the reactor and the presence of different phages (Additional file 2 : Fig. 3A). Considering taxonomically classified phages with relative abundance over 0.3%, 27 out of 47 families were influenced by temperature. Rudiviridae predominated in reactors below 25 degrees, while Suoliviridae (n = 527) and Winoviridae (n = 256) were most represented among the 10 dominant families in mesophilic environments. In thermophilic environments, Peduoviridae , Herelleviridae , Casjensviridae , and Drexlerviridae had the highest representation among the 16 dominant families. Detection of some evidence regarding previous infections can be used to study virus-host interactions. CRISPR spacers in particular were used to identify 2901 interactions between 546 MAGs and 1,822 viruses (Fig. 2 A). The presence of shared spacers between phages and bacteria suggests a history of phage-bacterial interactions, and specifically lysogenic infections [ 66 ]. Most phages had only one host, but virMAG 384269_AS06 ( Herelleviridae family) shared 40 different spacers with one bacterial MAG, Candidatus Anammoximicrobium sp. 06rmzA_251. Both have similar GC percentages (62.61–64.4) and are found exclusively in one experiment [ 67 ], suggesting they shared a long period of co-evolution, with the incorporation of new spacers in the bacteria and the ability to evade the recognition of CRISPR system by phages. In contrast, virMAG 27752_AS02 (family Straboviridae ) was found in spacers of 8 different MAGs belonging to Firmicutes Phyla, order Clostridiales, with the exception of one MAG of Tissierellia class. This phage has a broader range of infection and can be present in different conditions, from samples with only blood agar (BA) medium and casein (Zhu_2019_CSTR_2) to complex feedstock such as raw municipal biowaste (Tsapekos_2021_CSTR_5). One of the most abundant phages (virMAG MPIBJC05F1_637440, family Vertoviridae ) is present in 174 different samples from 9 countries and only two hosts were identified: Firmicutes sp. 28xzH2_85 and Clostridiaceae sp. 31mySI_51, with similar GC% and taxonomically close. Despite the specificity of this virus, the host’s widespread distribution allowed its spread in the AD system across different environmental conditions. Fig. 2 Microbial diversity of the expanded Biogas Microbiome database. Microbial community similarity across samples expressed by beta-diversity. Outer circles indicate the operational temperature and the community alpha-diversity as expressed by Chao1 and Shannon index In the majority of the analyzed experiments, the archaeal abundance was lower than bacteria, which limited the detection of CRISPR. Despite this limitation, 76 archaeal genomes were identified to possess CRISPR arrays with more than 4 spacers. Further analysis of these genomes revealed that the genera Methanothermobacter and Methanosarcina were the most infected, accounting for 70% of all recorded virus-archaea interactions (Fig. 2 B). Phages infecting Methanothermobacter and Methanosarcina archaea do not belong to the same family. In fact, those with the highest abundance infecting Methanothermobacter are found in a set of batch experiments (Kougias_2016_Batch) [ 68 ], while phages infecting Methanosarcina are less dominant but distributed across a greater number of samples, including batch experiments (Ma_2021_BGP_52) and biogas plants fed with sweet potato [ 21 ] and manure [ 12 ]. Functional analysis of the microbial community To understand the biological drivers of AD, functional annotation was integrated with taxonomic abundance, stratifying KEGG modules at different taxonomic levels. In order for a microbial consortium to perform a metabolic process, the simultaneous presence of all its functional units is required, though some gaps may exist in MAGs. As a result, only MAGs with complete and one-block-missing modules were considered (Additional file 3 for Archaea, Additional file 4 for Bacteria). Some functional modules are of particular interest in AD communities (Materials and methods, Taxonomic and functional prediction. Besides the obvious role of methanogenesis,beta-oxidation is relevant for fatty acid degradation when feedstocks are particularly rich in lipids. Additionally, modules involved in anaerobic carbon metabolismnitrogen metabolism, and sulfate reduction were also considered due to their influence on the AD process efficiency. and on the interactions between Bacteria and Archaea. For example, ammonia nitrification influences methane production since NH 4 can strongly affect both the methanogenesis process, and the growth of methanogens, including for example Methanobacterium , Methanosarcina, and Methanospirillum spp. [ 69 , 70 ]. Archaeal community Out of 198 archaeal MAGs, 70 (35%) have either complete or one missing block for methane production from acetate (M00357; acetoclastic methanogens), with most belonging to the class Methanomicrobia (72%), followed by Methanobacteria (24%). Considering methane production from CO 2 (M00567; hydrogenotrophic methanogenesis), the proportion of Methanomicrobia is even higher, with 55 MAGs out of 72 (76%). Out of the total 89 MAGs with either one of the methane production modules, 72% have both methane production modules, of which 69% belong to the Methanomicrobia class and 19% to Methanobacteria. Both classes are reported in literature as hydrogenotrophic methanogens, which is the most widespread pathway in Archaea [ 71 ]. Except for Candidatus Methanoculleus thermohydrogenotrophicum 31mySI10, which has a well-defined taxonomic affiliation, all the Candidatus archaeal MAGs harbor incomplete (or two block missing) methanogenic modules (Fig. 3 A), indicating the need for a more detailed functional investigation. Actually, a significant portion of genes found in archaea, ranging from approximately 30% to as high as 80%, code for proteins labeled as 'hypothetical proteins'. This is primarily due to the challenges in isolating and culturing most archaea in the laboratory, which makes experimental characterization of their gene repertoire difficult [ 72 ]. In contrast, the three most abundant archaeal MAGs in all the samples ( Candidatus Methanoculleus thermohydrogenotrophicum 31mySI10, Methanotrix sp. 43zhSC152, and Methanothermobacter wolfeii 31mySI 58) have both modules complete. However, only Methanothrix exhibits both functional methane pathways in vivo experiments [ 73 ]. Methanoculleus and Methanothermobacter are known hydrogenotrophs [ 74 ], and there is no evidence that they use acetoclastic pathways. This emphasizes the necessity for specific archaea annotations to ensure more precise genomes characterization (Fig. 4 ). Fig. 3 Phylogenetic tree and virome interaction. A The tree is represented in an inverted orientation, with branches color-coded according to phylum taxonomy. Legend is clockwise oriented starting from Euryarchaeota phylum The outermost ring exhibits the phylogenetic classification of the various taxa (Phylum), followed by subsequent rings that display the log-normalized coverage for each MAG (Coverage), the percentage of guanine-citosine (GC%), and the number of interactions with phages (Phages). Red asterisks indicate MAGs with more than 32 phage interaction signals. Phages that infect more than two MAGs are indicated by inner lines. B Relationships between different archaeal genera and the phages that infect them, the numbers indicate the total number of genomic sequences Fig. 4 Methanogenesis modules in archaeal MAGs. A Phylophlan tree with all archaeal MAGs encoding the methanogenesis modules. MAGs quality (completeness and contamination) and their cumulative abundance (%) are displayed in the three innermost circles. The MAGs more frequently identified in the samples of the AD database are highlighted in light purple. B Heatmap reporting the samples with more than 10% relative abundance of complete and one block missing methanogenesis modules. The barplot represents the relative abundance of each genus in the samples and was calculated by taking into account only MAGs with complete modules, and those with one block missing. M00357: methanogenesis, acetate = > methane; M00567: methanogenesis, CO 2 = > methane The estimation of the relative abundance of each KEGG module in each sample was obtained by considering the complete modules and those with one block missing. According to this calculation, samples “Kouzuma_2017_CSTR_4” and “Shi_2021_Batch_1” had the highest values for methanogenesis KEGG modules (M00357, M00567) (Fig. 3 B). In the “Shi_2021_Batch” experiment, hydrochar was used to enrich methanogenic species including Methanobacterium , Methanolinea , and Methanothrix genera, and increase methane production. Conversely, “Kouzuma_2017_CSTR_2 and 4” and most samples of “Zhu_2020_CSTR” had the highest abundance of acetoclastic functions, represented by Methanosarcina , and Methanothermobacter taxa. The presence of a diverse range of feedstock compositions can be observed in samples where methane production modules exhibit a relative abundance of over 10%, including acetate [ 4 , 74 – 76 ], casein [ 75 ], glucose [ 75 ], oleate [ 67 , 77 ], manure [ 67 , 77 – 79 ], lactate [ 4 ], butyrate [ 4 ], propionate [ 4 , 79 ], sludge [ 80 ], and municipal biowaste [ 81 ]. However, the most abundant samples with both modules [ 4 , 74 , 79 ] were fed with acetate and propionate. The observed variations in methanogenesis and acetoclastic functions across different biogas reactors indicate a significant influence on the microbial community composition on biogas production. The enrichment of methanogenic species in certain samples, facilitated by specific experimental conditions such as the use of hydrochar, suggests a potential for increased methane production. The confirmation that reactors fed with organic acids, particularly acetate, enrich specific microbial species, irrespective of temperature conditions, suggests a strategic approach for optimizing biogas production. These findings emphasize the importance of tailoring substrate selection and reactor conditions to enhance the performance of microbial communities, providing valuable insights for improving overall biogas yields in anaerobic digestion processes. Bacterial community The results obtained from the functional analysis indicate that three KEGG modules are widespread in the AD bacterial community: beta-oxidation (M00087), dissimilatory nitrate reduction to ammonium (M00530), and assimilatory sulfate reduction to sulfide (M00176) (Additional file 2 : Fig. 4). The presence of the “dissimilatory nitrate reduction to ammonium” function in 71% of phyla suggests that denitrification is widespread in the AD community, and according to this, the two-step process of nitrate conversion to ammonia and finally to nitrogen through denitrification of nitrate to nitrogen module (M00529) is very common. Despite other studies reported that the bacterial dissimilatory nitrate reduction to ammonium pathway only dominates under low nitrate availability and in sulfide-free environments [ 82 , 83 ], the present data suggest that species using nitrate (rather than oxygen) as electron acceptor are dominant in the AD microbiome. A previous study used a CSTR to remove linear alkylbenzene sulfonate (LAS) present in commercial laundry wastewater (Delforno_2020_CSTR_1); in this investigation a higher abundance of dissimilatory nitrate reduction to ammonium, denitrification, and beta-oxidation modules were identified (Additional file 2 : Fig. 4). Beta-oxidation is the second step of LAS degradation and is mainly performed by Synergistes and Syntrophus , which are widespread genera in LAS degradation reactors [ 84 – 86 ]. Wastewater frequently contains several nitrogen-rich compounds, including nitrate, nitrite, and ammonia [ 87 ], which may have contributed to the enrichment of bacteria capable of converting nitrogen to different oxidation states. Nitrogen metabolism was frequently identified in some taxa including the candidate Zixibacteria division and Acidobacteria. For example, half of the MAGs assigned to the candidate Zixibacteria division have the complete function of nitrification (M00528) and denitrification (M00529), reflecting their metabolic versatility [ 88 ]. Regarding the Acidobacteria phylum, nitrification (M00528) and complete nitrification (M00804) modules were identified in 50% and 25% of the MAGs, respectively; despite this finding the nitrification ability was not proven in isolates of the Acidobacteria phylum [ 89 , 90 ]. The Acidobacteria nitrification function was more abundant in thermophilic manure-supplemented biogas plants with high biogas production and low pH (Campanaro_2018_BGP_3). AD biogas production is heavily influenced by the organic substrates used. Some of these substrates may contain inhibitors such as sulfides, which can negatively affect the microbiome and decrease the AD process efficiency [ 91 ]. For example, sulfate-reducing bacteria (SRB) encoding proteins involved in the assimilatory sulfate reduction to sulfide function can compete with hydrogenotrophic archaea for hydrogen, and generate H 2 S as final product [ 65 , 66 ]. In particular, a batch reactor fed with cellulosic and xylan biomass (Jia_2018_Batch) showed the highest representation of assimilatory sulfate reduction to sulfide across MAGs. Although the H 2 S concentration was not reported, all the batch experiments produced low concentration of CH 4 (0.2–1.5 mM), suggesting that the process predominantly shifted to sulfate reduction [ 55 ]. Indeed, 68% of “Jia_2018_Batch” library read counts mapped to the Clostridium butyricum 37jiCB_291 MAG, a known SRB [ 83 ]. Overall, a global Biogas Microbiome database can be useful to infer putative inhibitory in the AD process by analyzing the pathways of the MAGs identified. In contrast, key carbon metabolism modules including the Arnon-Buchanan cycle, the WL pathway, and the acetogenesis are complete or have only one block missing in less than 25% of the total MAGs suggesting these are shell modules in the AD system [ 92 ]. Nevertheless, these metabolic routes for carbon were identified in a range of phyla including Actinobacteria, Chloroflexi, Firmicutes, Ignavibacteriae, Proteobacteria, and Spirochaetes. The Arnon-Buchanan cycle module, involving a reverse citric acid cycle for CO 2 fixation, was the most abundant in samples inoculated with sludge (Macedo_2020_CSTR and Zhang_2020_Batch), while the WL and acetogen modules were more abundant in a CSTR sample for the AD of saccharides with a feedstock of volatile fatty acid mixture (Zhu_2019_CSTR_5). Microbial replication rates are linked to their functional capabilities To characterize microbial dynamics across AD systems, we calculated the peak-to-trough ratio (PTR) of the MAGs coverage, which estimates DNA synthesis and generation rate [ 47 ]. Sample-specific PTRs were obtained for 782 MAGs in the dataset, corresponding to those genomes meeting the minimal coverage requirement in at least one sample. As a result, PTR values were determined in fewer than 20 samples for most MAGs, while for others, e.g. Methanothrix sp. 43zhSC_152, Synergistaceae sp. 24abBP_148, Bacteroidales sp. 28xzH2_30, and various members of Firmicutes, PTRs were estimated for over a hundred samples, reflecting their widespread abundance. Resulting PTR values generally have a long-tail distribution, with a median of 0.37 and exceeding 2 in some cases, with varying spread across different taxonomic groups (Additional file 2 : Fig. 5). Within these distributions, both coarse- and fine-grain trends matching previous knowledge were observed. Firstly, acetoclastic and hydrogenotrophic methanogens exhibited a low mean PTR, consistently with direct measures of the replication rate in isolates (Additional file 2 : Fig. 6). Secondly, slightly lower mean PTR was observed in batch reactors, consistently with the limited duration and efficiency of the processes therein compared to CSTRs and full-scale plants. When considering individual microorganisms, significant PTR trends for MAGs classified at genus or species level are in some cases in agreement with documented growth temperature preferences, where robust PTR distributions could be recovered (Fig. 5 ) [ 93 , 94 ]. In particular, microorganisms of the Methanoculleus genus are cultivated at a temperature between 28° and 37 °C, and Methanoculleus sp. 52maCN_3230 has an estimated optimal growth temperature in approximately the same interval. Porphyromonadaceae sp. 02xzSI_42 reaches the highest PTR values around 40 °C, which is its recommended cultivation temperature [ 93 ]. Similarly, Methanosarcina thermophila 28xzH2_79 has detectable non-null PTR values predominantly close to 55 °C, even though some outliers were found in the mesophilic range. PTR distribution for other species, including Methanobacterium sp. 29adLB_146, Aminobacterium sp. 23ysBP_18, and Syntrophorhabdus sp. 42zhAM_214, show their maximum values in the 25–40 °C range, in agreement with the largely mesophilic representation of these genera [ 94 ]. Although replication rate is generally influenced by a variety of factors, together, these patterns support the overall soundness of obtained distributions. Fig. 5 Relationship between PTR and temperature for individual MAGs. Shown values refer to significant relationships as assessed by quadratic polynomial fits, based on a FDR-adjusted p -value threshold of 0.1 Besides, other less obvious trends were found. While Verrucomicrobia, Fibrobacteres, and Planctomycetes are among the phyla with the largest mean PTR, subsets of Proteobacteria and Bacteroidetes show a more noticeable long tail around large values (Additional file 2 : Fig. 5). More specifically, MAGs classified as Syntrophobacterales seem to have fast replication rates in several conditions. This taxon harbors sulfate reducers, which can efficiently metabolize substrates such as pyruvate, methanol, and glucose and outcompete methanogens when sufficient sulfate is available [ 95 , 96 ]. These results are confirmed by the long tail of PTRs associated with assimilatory sulfate reduction (Additional file 2 : Fig. 6), and thus support the rapid proliferation capability of these microbes. Other of such fast-growing microbial groups are Bacteroidales (e.g. Proteiniphilum sp. 29adLB_192), with relevant proteolytic role [ 97 ], and microbes involved in fatty acid oxidation (Additional file 2 : Fig. 7). These groups could represent species with high energy generation capability from the breakdown of such macromolecules. Moreover, archaeal phyla also have a wide range of replication rates, with some MAGs exhibiting high PTRs. In particular, Methanomicrobiales sp. 21ysBP_11 and Methanosarcina flavescens 22ysBP_46 are among the fast-growing archaea with PTR consistently above 2, making Methanomicrobiales sp. 21ysBP_11 a potential promising candidate for cultivation and isolation. In general, most MAGs exhibit their highest replication rates within a 5 °C range (preferential temperature range), indicating that temperature tightly controls replication efficiency (Additional file 2 : Fig. 7). In fact, mesophilic and thermophilic communities tend to present distinct composition and diversity, as also seen above. Yet, a relevant number of bacterial and archaeal MAGs show their highest PTRs over a 20 °C window, suggesting that some species are able to better adapt across temperature regimes. Numerous genomic variants delineate microbial population heterogeneity To date, there is a lack of information in literature regarding the genetic heterogeneity of AD-relevant species [ 74 ]. The presence of the same microbial species in reactors characterized by different process parameters can allow to identify variants impacting the adaptation process, as well as a more detailed characterization of species at strain level. Variant identification was performed by aligning shotgun reads of each experiment back to the MAGs obtained from the binning process, and this approach led to the identification of 10.5 millions single nucleotide variants (SNVs). The high number of variants revealed a high genetic heterogeneity in the microbial population, and variants characterization allowed their classification as synonymous (60.5%), nonsynonymous (28.9%), intergenic (10.4%), and multigenic (0.02%). Of the 3,050 MAGs containing SNVs, only eight exhibited more than 500,000 SNVs, two of which were the methanogenic archaea Methanothrix sp. 43zhSC_152 and Methanoculleus sp. 52maCN_3230, which harbor 938,000 and 661,000 SNVs, respectively (see Availability of Data and Materials). The high number of variants observed in Methanothrix sp. 43zhSC_152 could be due to the fact that this MAG is widespread, being present in 96 of the examined samples; possibly, the presence of this species in many reactors with highly different conditions have led to the differentiation of a large number of strains harboring genetic variants. On the other hand, the presence of Methanoculleus sp. 52maCN_3230 was limited to 23 mesophilic samples, and thus, the high number of genetic variants observed in this MAG could be due to different factors in comparison to Methanotrix . In order to obtain a more reliable representation of the genomic variability, the number of SNV per MAG was normalized, both according to the genome length and to the number of samples where the MAG was identified with coverage higher than 1 (Fig. 6 B). This analysis highlighted Methanoculleus sp. 52maCN_3230 and Methanomicrobiales sp. 19jrsB_18 as outliers with more than 10,000 SNVs/Mbp per sample (Fig. 6 B). This finding is of particular interest because, while Methanoculleus sp. 52maCN_3230 is quite common in the AD samples, Methanomicrobiales sp. 19jrsB_18 was identified in two samples only. The genomic location of nonsynonymous variants identified in both methanogens allowed linking them to the gene and to the functional pathways. Interestingly, 2.5% of the Methanomicrobiales sp. 19jrsB_18 variants and 9.1% of the Methanoculleus sp. 52maCN_3230 variants were associated with core genes of the hydrogenotrophic methanogenesis, including heterodisulfide reductase ( hdr ), methyl-coenzyme M reductase ( mcr ) and formylmethanofuran dehydrogenase ( fwd ). Overall, these results suggest that variants with a crucial role in the adaptation of methanogenic archaea to reactors operated in different conditions. Fig. 6 Overview of variants distribution among different taxonomic groups. A Number of MAGs associated with each phylum. B Number of SNVs/Mbp in each phylum for MAGs with more than 100 SNVs; each dot represents a MAG. Results are reported for each phylum with more than five MAGs, while all the others are reported as “Other”. C Statistical analysis comparing the median SNV density calculated for each phylum and the average value reported for the global Biogas Microbiome database. Statistically significant results are marked with asterisks: “*” p < = 0.05, “**” p < = 0.01, “***” p < = 0.001 and “****” p < = 0.0001 In order to determine if some phyla were statistically more impacted by variants, the taxonomic assignment of each MAG at phylum level was considered along with the number of SNVs/Mbp. Results of the Mann–Whitney U-test on SNVs/Mbp distribution showed that eight phyla had a significant enrichment in the number of variants ( p -value < 0.05) (Fig. 6 C), with Candidatus Cloacimonetes ( p = 0.0003), Euryarchaeota ( p = 0.0025), Atribacterota ( p = 0.0307) and Synergistetes ( p = 0,0023), having a number higher than expected, and Firmicutes ( p = 0.0380), Planctomycetes ( p = 0.0145), Verrucomicrobia ( p = 0.0043) and Lentisphaerae ( p = 0.0015), having a lower number. In general, the results obtained for Euryarchaeota suggest that methanogenic archaea are under a strong selective pressure and harbor a large amount of genetic variability. The metabolic roles of species belonging to Candidatus Cloacimonetes and Synergistetes are still not completely clear, however, it was previously reported that some of them are characterized by acetogenesis [ 98 ], or can compete for acetate utilization with Methanosaeta [ 99 ]. To get a first glimpse on the strains composition, a phylogenetic analysis was conducted on Methanothrix sp. 43zhSC_152, Methanothermobacter wolfeii 31mySI 58 and Candidatus Methanocullus thermohydrogenotrophicum 31mySI_10, the three archeal species with the highest number of MAGs. These three species have different properties in the database: Methanotrix has a high number of MAGs and a worldwide distribution, the other two have less MAGs and are more abundant in European reactors (see Availability of Data and Materials). Phylogenetic analysis revealed a regional distribution of MAGs for Methanotrix with a distinct cluster of MAGs deriving from Chinese biogas plants. Interestingly, all the MAGs recovered for Methanothrix sp. 43zhSC_152 are mesophilic, and all the MAGs of M. wolfeii 31mySI 58 are thermophilic, while Candidatus M. thermohydrogenotrophicum 31mySI_10 is more flexible and is present in some mesophilic and thermophilic samples (Additional file 2 : Fig. 8). This finding contradicts previous data reporting this species only in thermophilic conditions [ 100 ], and suggests a different habit for this methanogen, which is able to adapt to mesophilic conditions as well. The temperature is not the main driver of strains differentiation for Candidatus M. thermohydrogenotrophicum, while the high phylogenetic distance among some M. wolfeii and Ca. M. thermohydrogenotrophicum MAGs suggest a possible impact of environmental conditions on their strains differentiation. Two M. wolfeii MAGs identified in samples subjected to high H 2 concentrations, 15tlH2_55 and 50dbBF_040, appeared quite distant from the others suggesting a selective pressure determined by environmental conditions."
} | 11,805 |
34065848 | PMC8151373 | pmc | 1,167 | {
"abstract": "Terrestrial plants evolution occurred in the presence of microbes, the phytomicrobiome. The rhizosphere microbial community is the most abundant and diverse subset of the phytomicrobiome and can include both beneficial and parasitic/pathogenic microbes. Prokaryotes of the phytomicrobiome have evolved relationships with plants that range from non-dependent interactions to dependent endosymbionts. The most extreme endosymbiotic examples are the chloroplasts and mitochondria, which have become organelles and integral parts of the plant, leading to some similarity in DNA sequence between plant tissues and cyanobacteria, the prokaryotic symbiont of ancestral plants. Microbes were associated with the precursors of land plants, green algae, and helped algae transition from aquatic to terrestrial environments. In the terrestrial setting the phytomicrobiome contributes to plant growth and development by (1) establishing symbiotic relationships between plant growth-promoting microbes, including rhizobacteria and mycorrhizal fungi, (2) conferring biotic stress resistance by producing antibiotic compounds, and (3) secreting microbe-to-plant signal compounds, such as phytohormones or their analogues, that regulate aspects of plant physiology, including stress resistance. As plants have evolved, they recruited microbes to assist in the adaptation to available growing environments. Microbes serve themselves by promoting plant growth, which in turn provides microbes with nutrition (root exudates, a source of reduced carbon) and a desirable habitat (the rhizosphere or within plant tissues). The outcome of this coevolution is the diverse and metabolically rich microbial community that now exists in the rhizosphere of terrestrial plants. The holobiont, the unit made up of the phytomicrobiome and the plant host, results from this wide range of coevolved relationships. We are just beginning to appreciate the many ways in which this complex and subtle coevolution acts in agricultural systems.",
"conclusion": "4. Conclusions Plant biologists have begun to understand that the origin, development, and ultimate success of plants is closely linked to plant interactions with the phytomicrobiome. Early evolution studies have indicated that the first terrestrial organisms evolved from the endosymbiosis of cyanobacteria allowing for the derivation of algae, non-vascular plants, and finally vascular plants. Recent explorations of plant-microbe interactions have demonstrated that plants and microbes are inseparable, having coevolved since the advent of the first plants. Knowledge of how early plants survived the challenges associated with colonizing the terrestrial environment and subsequently evolved goes back to the earlier theory of land sterility during the Precambrian period [ 108 ], which underestimated both the past and future role of microbes. Microbes have played a pivotal role in the evolution of plants and there is no doubt that microbes still hold the key to a better future for plant survival in the face climate change and a rapidly growing global agricultural sector; interest in this research area continues to grow. A greater understanding of the remarkable functions performed by the phytomicrobiome and their assistance to plant survival is being revealed and acknowledged. Considering plants and their associated phytomicrobiome as one unit, a holobiont, is a necessary step to sustainability of crop production. Considering plants as holobionts in breeding systems could yield sustainable results when compared to single or few genes improvements whose character may face environmental/evolutionary challenges, with rapidly fading efficacy. As such today’s domestication and improvement of plants, including consideration of advantages associated with the a holobiont approach is certain to reap the benefits of plant-microbe interactions [ 109 ], leading to more ecologically sustainable agriculture. Synthesizing the developments made in understanding coevolution of plants and microbes is essential in the continuing efforts to establish and advance potential areas of research that could increase sustainability in agriculture, and allow sustainable responses as evolving challenges to agricultural sustainability arise. Evidence suggests that from the earliest stages of evolution there has been substantial diversity in the microbial communities associated with plants, at any point in time. The current understanding of plants as part of holobionts is yielding more complex scientific questions which require more holistic approaches to arrive at the best hypothesis. This review is a continued effort to emphasize that microbes are always working with (beneficial) or against (pathogenic) their host, which has the potential to lead to a more resilient holobiont.",
"introduction": "1. Introduction The sophisticated and complex association between plants and microorganisms, including bacteria and fungi, have existed since the early stages of life on Earth. Colonization of terrestrial habitats began with plants, followed by animals, and was possible only when specific genes from terrestrial bacteria were transferred to algae, in order to increase tolerance to abiotic and biotic stresses present on land [ 1 ]. The relationship of cyanobacteria, a prokaryote, with eukaryotes that eventually developed into algae, was a pivotal step in the progression of this evolution [ 2 ]. Cyanobacteria played a pivotal role in formation of algae through endosymbiosis by which a cyanobacterium was incorporated into a heterotrophic eukaryote ancestor where it was retained and specialized into an organelle, thus bringing about photosynthetic eukaryotes [ 3 ]. Molecular multi-gene phylogeny has clearly indicated that cyanobacteria became the primary plastid in green and red algae, and glaucophytes; there has also been a series of secondary endosymbiosis with other eukaryote ancestors [ 4 ]. The colonization of land plants by fungal and bacterial symbionts was a critical stage to bringing about evolution of terrestrial ecosystems, but how the members of early communities interacted and influenced one another is still relatively unexplored [ 5 ]. An expanding body of fossil evidence shows that interactions among early terrestrial communities included bacteria, fungi, algae, lichens, and bryophytes—the ecosystem services provided by these organisms include the weathering of parent rock material, soil formation, stabilization of sediments, and the productivity of ecosystems [ 6 , 7 ]. When plants moved onto the land, the role of the microorganisms became clearer, including improving plant tolerance to biotic and abiotic stresses. Mutualistic interactions have been reported among microbes (e.g., plant growth promoting bacteria (PGPB) and mycorrhizal fungi) which enhance nutrient acquisition, control elements of plant development and physiology through signal compounds and phytohormones/hormone analogues that trigger stress resistance in the host plant, and by producing compounds with antagonistic activity toward plant pathogens [ 8 , 9 , 10 ]. In turn the associated plant provides habitat and releases exudates into rhizosphere, including reduced carbon as an energy source for the growth of phytomicrobiome members. Through the interactions between microorganisms and associated plants, the two together form the holobiont [ 11 ]. In this paper we focus on the role of microorganisms in the evolution of plants. We first discuss the endosymbiosis of microbes which lead to the evolution of critical plant organelles. Then, we review the fundamental roles of microbes in plant development and survival, from mutualistic to parasitic/pathogenic interactions. Finally, we propose that the holobiont concept should be incorporated into thinking around agricultural systems based on the role of microbial communities in agro-ecosystems."
} | 1,965 |
20846434 | PMC2954876 | pmc | 1,168 | {
"abstract": "Background Methylotrophic (methanol-utilizing) bacteria offer great potential as cell factories in the production of numerous products from biomass-derived methanol. Bio-methanol is essentially a non-food substrate, an advantage over sugar-utilizing cell factories. Low-value products as well as fine chemicals and advanced materials are envisageable from methanol. For example, several methylotrophic bacteria, including Methylobacterium extorquens , can produce large quantities of the biodegradable polyester polyhydroxybutyric acid (PHB), the best known polyhydroxyalkanoate (PHA). With the purpose of producing second-generation PHAs with increased value, we have explored the feasibility of using M. extorquens for producing functionalized PHAs containing C-C double bonds, thus, making them amenable to future chemical/biochemical modifications for high value applications. Results Our proprietary M. extorquens ATCC 55366 was found unable to yield functionalized PHAs when fed methanol and selected unsaturated carboxylic acids as secondary substrates. However, cloning of either the phaC1 or the phaC2 gene from P. fluorescens GK13, using an inducible and regulated expression system based on cumate as inducer (the cumate switch), yielded recombinant M. extorquens strains capable of incorporating modest quantities of C-C double bonds into PHA, starting from either C6= and/or C8=. The two recombinant strains gave poor results with C11=. The strain containing the phaC2 gene was better at using C8= and at incorporating C-C double bonds into PHA. Solvent fractioning indicated that the produced polymers were PHA blends that consequently originated from independent actions of the native and the recombinant PHA synthases. Conclusions This work constitutes an example of metabolic engineering applied to the construction of a methanol-utilizing bacterium capable of producing functionalized PHAs containing C-C double bonds. In this regard, the PhaC2 synthase appeared superior to the PhaC1 synthase at utilizing C8= as source of C-C double bonds and at incorporating C-C double bonds into PHA from either C6= or C8=. The M. ex-phaC2 strain is, therefore, a promising biocatalyst for generating advanced (functionalized) PHAs for future high value applications in various fields.",
"conclusion": "Conclusions It was found that the wildtype M. extorquens ATCC 55366 strain could not produce functionalized PHAs starting from a methanol + unsaturated carboxylic acid mixture. As a consequence, a metabolic engineering approach was used to convert the wildtype strain into a \"cell factory\" capable of producing functionalized PHAs containing C-C double bonds. The presence of the C-C double bonds in the PHA side chains was confirmed by NMR. It was also found that the M. extorquens cell factory harboring the phaC2 gene appeared superior at utilizing unsaturated carboxylic acids and at incorporating C-C double bonds into PHA starting from either C6= or C8=. Our two M. extorquens cell factories were able to produce functionalized short-chain-length/mediumchain-length PHAs (SCL/MCL-PHAs), which combine desirable functionality of some MCL-PHAs with, reportedly, superior thermo-mechanical properties of inert SCL/MCL-PHAs. Double bonds may be used to induce side chain crosslinking and to covalently bind small or macromolecules upon chemical or enzymatic modifications. To date, the advantage of exhibiting functionality has been exclusively reserved for MCL- and LCL-PHAs.",
"discussion": "Results and Discussion Production of unsaturated PHAs using the wild-type strain The potential of the pink facultative methylotroph M. extorquens ATCC 55366 was tested to utilize various fatty acids including fatty acids with C-C double bonds. The assays were performed in shake flasks using medium CHOI4 or medium CHOI5. Nearly 100 shake flask assays were conducted. The M. extorquens cultures were grown first on methanol as sole carbon and energy source and the fatty acid of interest was added at some point to the growing cultures, generally at a final concentration in the 0.1-0.3% range. A summary of the substrates tested and of the results obtained is presented in Table 1 . Only the homopolymer PHB or the copolymer PHBV was detected in the cultures as no other peaks apart from 3-hydroxybutyrate (3HB) and 3-hydroxypentanoate (3HP = 3HV) were seen in GC chromatograms. As can be seen, feeding of C5-fatty acids led to accumulation of the copolymer PHBV solely and, unfortunately, no trace of C-C double bounds could be detected in the PHAs produced upon feeding C5-fatty acids containing a C-C double bond. Control of pH is a key factor when cultures are fed free carboxylic acids. In the present study, efforts were made to minimize pH effects by regular, manual addition of 1 M KOH in order to maintain pH within a reasonable range favorable for growth and substrate utilization. Given the extent of our screening efforts, it may be reasonably concluded that the wildtype M. extorquens ATCC 55366 strain was unable to accumulate functionalized PHAs containing C-C double bonds. As a consequence, the next step consisted in developing M. extorquens strains harboring heterologous PHA synthase genes allowing for biosynthesis of the wanted functionalized PHAs. Table 1 Utilization of various fatty acids, unsaturated or saturated, for growth and PHA accumulation by M. extorquens ATCC 55366. Co-substrate Monomeric composition of the produced PHA 1 C4:0 [mol%] C5:0 [mol%] C5 30.4 69.6 C5= 2 100 - C5= > 99 tr 3 t -C5= > 99 tr t -C5= 87.9 12.1 C6 > 99 tr C6 > 99 tr C6 100 tr C6= 4 100 - C8= 98.7 1.2 C8= 5 100 - C11= 6 > 99 tr C11= 7 100 - 1 CX:Y = double bond at position Y of monomeric chain length X; 2 Three (3) similar samples were analyzed; 3 tr = detected in trace quantity; 4 Six (6) similar samples were analyzed; 5 Three (3) similar samples were analyzed; 6 Eight (8) similar samples were analyzed; 7 Three (3) similar samples were analyzed. Metabolic engineering of M. extorquens for modifying its PHA-synthesizing machinery The inability of the wildtype M. extorquens strain to yield the desired PHAs was assumed to be due to a too narrow substrate specificity of the wildtype PHA synthase enzyme. As done by several research groups [ 37 - 44 ], we then decided to engineer M. extorquens strains harboring different PHA synthase ( phaC ) genes known or assumed to code for PHA synthases of broader substrate specificity. After careful review of the literature, the phaC1 and phaC2 genes present in the P. fluorescens GK13 strain were selected because the PHA synthesis machinery of this bacterium is able to also produce longer PHA chains [ 23 , 45 ], an indication of broader substrate specificity. In a first step, the phaC1 and phaC2 genes were successfully isolated using PCR and the isolated genes were identical to the reported sequences [ 45 ]. In a second step, the two phaC genes were cloned into the pAll plasmid (Figure 1 ) and inducible expression was driven by the cloned methanol dehydrogenase (PmxaF) of M. extorquens under control of regulatory elements of P. putida F1. This new technology for inducible and regulated gene expression was developed by our group for M. extorquens [ 28 , 29 ]. It may be described as follows: (A) In the absence of the chemical inducer (p-isopropylbenzoate = cumate), the repressor protein cymR is bound to the operator site upstream of the gene of interest and transcription is blocked; (B) addition of cumate as inducer leads to formation of the cymR-cumate complex, followed by detachment of cymR from the operator and activation of the transcription of the downstream gene. The two recombinant M. extorquens strains, M. ex-pha C1 and M. ex-pha C2 , as well as the M. extorquens wildtype strain were grown in CHOI medium containing methanol, plus tetracycline in the case of the two recombinant strains. As shown in Figure 2 , addition of cumate, the inducer, led to expression of both the phaC1 gene (lane 4) and of the phaC2 gene (lane 6). Both protein bands indicated a molecular mass of approximately 62 kDa. Interestingly, a strong band, with an apparent molecular mass of 40 kDA, was present in the samples for noninduced M. ex-pha C1 cells (lane 3) or M. ex-pha C2 cells (lane 5). The same band was quasi inexistent in induced cells of the same recombinant strains. This additional band is certainly not related to the insert ( phaC gene) since it has been observed with other recombinant M. extorquens strains harboring different and various inserts. Ideally, the two recombinant phaC genes should have been introduced into a phaC -minus mutant of M. extorquens . Unfortunately, although we were able to obtain such mutants, the mutants grew very poorly in the culture media containing methanol as sole carbon source. A similar observation on the growth of strain AMI of M. extorquens on C1 and C2 compounds was also reported by Korotkova and Lidstrom [ 46 ]. Consequently, because of the presence of the native PHA synthase, PHA blends comprised of PHB and functionalized PHAs could be expected. Figure 2 SDS-PAGE (4-12%) of M. extorquens cell-free extracts from shake flask cultures . Lane M, standard marker mixture; lane 1, wildtype M. extorquens before induction; lane 2, wildtype M. extorquens after induction with cumate; lane 3, M. ex-pha C1 before induction; lane 4, M. ex-pha C1 after induction; lane 5, M. ex-pha C2 before induction; lane 6, M. ex-pha C2 after induction. The arrows indicate the putative presence of recombinant PhaC1 or PhaC2 enzyme, which showed an approximate molecular mass of 62 kDa. Growth of the two recombinant M. extorquens strains As illustrated in Figure 3 , the two recombinant strains grew well on methanol but maximal growth of the two strains was inferior to that of the wildtype strain. At 70-72 h, addition of a mixture of methanol + 10-undecenoic acid (C11=) resulted in further growth for the three strains but growth, as measured by optical density, rapidly reached a plateau in all three cases. The M. ex-pha C2 strain appeared to grow better than the M. ex-pha C1 strain under the conditions used. Phase contrast microscopy of the cultures showed the presence of granules, tentatively identified as PHB (or PHA) granules, in cells from the three M. extorquens strains and the granules increased in size over time. The appearance of the granules occurred between 68 h and 74 h and it coincided with the addition of the methanol + C11= mixture (Figure 3 , arrow). Figure 3 Growth of M. extorquens strains . Growth of the wildtype M. extorquens strain and of the two recombinant M. extorquens strains on methanol (Phase 1, 0-68 h) and on methanol + 10-undecenoic acid (C11=) (Phase 2, 68 h-end). Addition of the mixture is indicated by the arrow. (◆) wildtype strain, n = 3; (■) M. ex-pha C1 strain, n = 3; (▲) M. ex-pha C2 strain, n = 2. The medium initially contained 1 vol% methanol. At 68 h, methanol and C11= were added to give a final concentration of 2 vol% and 0.2 vol% of methanol and C11=, respectively. Many shake flask studies were conducted using various feeding strategies to verify the growth behavior in the presence of alkenoic acids. Cultures were started using methanol as main substrate and, at desired times, the selected unsaturated carboxylic acid was added to favor accumulation of a functionalized PHA. To minimize pH effects, attempts were made to maintain pH by manual addition of 1 M KOH. The growth results for strain M. ex-pha C1 are shown in Figure 4 while those for strain M. expha C2 are shown in Figure 5 . Pulse addition of the carboxylic acid is indicated by arrows. The two M. extorquens strains (Figure 4A and Figure 5A ) continued to grow, although more slowly, upon addition of 5-hexenoic acid (C6=) and \"respectable\" biomass levels (OD600nm) were obtained, between 8 and 11, with all flasks except for one case. The pH profile varied with each flask but pH values were always between 6.2 and 7.7. The growth results on 7-octenoic acid (C8=) are illustrated in Figure 4B ( M. ex-pha C1 ) and in Figure 5B ( M. ex-pha C2 ). The results were similar to those obtained with 5-hexenoic acid (C6=) except that the OD600nm profiles varied significantly more between flasks. As with 5-hexenoic acid, the pH profile varied with each flask but pH values were always between 6.2 and 7.6. With 10-undecenoic acid (C11=), however, growth appeared even more \"erratic\" as there was significant variation in the OD600nm profile between flasks (Figure 4C and Figure 5C for strains M. expha C1 and M. ex-pha C2 , respectively). Maximal OD600nm values varied between only 4 and almost 10 (Figure 4C ). Overall, most of the biomass was obtained from methanol, i.e. , before addition of the respective carboxylic acid, due to the well-known toxicity of such acids. Probably due to the lack of acceptable pH control, significant differences in optical density profiles were observed between shake flasks, especially when C11= was fed (Figures 4C and 5C ). Interestingly, in several cases, growth continued after addition of the \"toxic\" carboxylic acid. Several examples are depicted in Figures 4 and 5 . The optical density values obtained are very comparable to values found in the recent literature for similar work [ 15 , 17 , 37 - 39 , 47 ]. In this study, as in many others, emphasis was on strain development, not on process optimization. With C11= as co-substrate, manual maintenance of the pH was also more problematic, pH values lower than 6 were recorded with both strains, reaching even near 4.5 at one point with strain M. ex-pha C2 . Figure 4 Growth of the M. ex-pha C1 strain on selected unsaturated carboxylic acids . Cultures were grown first on methanol, then, the selected unsaturated carboxylic acid was added pulse-wise (indicated by arrows). Partial control of pH was done using 1 M KOH to prevent intolerable extremes in pH. Closed symbols: OD at 600 nm, open symbols: pH. (A) 5-hexenoic acid (C6=); (B) 7-octenoic acid (C8=); (C) 10-undecenoic acid (C11=). Figure 5 Growth of the M. ex-pha C2 strain on selected unsaturated carboxylic acids . Cultures were grown first on methanol, then, the selected unsaturated carboxylic acid was added pulse-wise (indicated by arrows). Partial control of pH was done using 1 M KOH to prevent intolerable extremes in pH. Closed symbols: OD at 600 nm, open symbols: pH. (A) 5-hexenoic acid (C6=); (B) 7-octenoic acid (C8=); (C) 10-undecenoic acid (C11=). Determination of the monomeric composition of polyhydroxyalkanoates Selected PHA samples were chosen as representatives and submitted to 1D and 2D NMR analysis for obtaining proof that C-C double bonds were present and that at least some of the C-C double bonds were located in the PHA side chains. Analysis of the NMR results confirmed both the presence of unsaturated PHA components in the samples and the presence of C-C double bonds in the side chains. Consequently, the representative PHA samples, including a PHA sample produced by the wildtype M. extorquens strain, were submitted to 1D and 2D NMR analyses (see Materials and Methods) to further identify which unsaturated components might be present. The results are summarized in Table 2 . The two samples derived from C11= yielded poor quality spectra. As representative example, the results for a PHA that was produced by M. ex-phaC2 on methanol and 7-octenoic acid will be presented in more detail: Table 2 1D and 2D NMR analysis of selected PHA samples 1 . Co-substrate 1D Analysis 2D Analysis C6= 2 3HB as major peak (> 90%), 3HHx= 3 Not performed. C8= 4 3HB as major peak (> 96%) At least 5 minor components: 3HP, 3HHx, 3HHx=, 3HO and 3HO=. C11= 5 Poor quality spectra. - 3HB, 3HP (very minor) No detectable unsaturated bonds 1 These samples were chosen as representatives due to availability of enough material and equipment; 2 Three (3) similar samples were analyzed; 3 Since C6= was the co-substrate, it was concluded that the detected double bonds belonged to 3HHx=; 4 Three (3) similar samples were analyzed; 5 Two (2) similar samples were analyzed. 1. The 1 H 1D spectrum indicated that 3HB was the major component comprising over 96% of the total signal (spectrum not shown). Using natural abundance 2D 1 H- 13 C correlation spectra (Figure 6 ), it was possible to identify at least 5 minor components present which were assigned to 3HP, 3HHx, 3HO, 3HHx= and 3HO=. Chemical shift assignments for these components are shown in Table 3 . Finally, at very low contour levels, we were able to detect and assign resonances at the terminus of polymer (these are too weak to detect in Figure 6 but assignments are listed in Table 3 ); Figure 6 Overlay of the 2D spectra for a PHA sample derived from methanol and 7-octenoic acid metabolism . The 2D 1 H- 13 C HSQC, HMBC and H2BC spectra are shown as black, red and blue contour plots, respectively. Single and multiple bond correlations for HP (3HP), HH (3HHx), HH= (3HHx=), HO (3HO) and HO= (3HO=) are indicated according to the following nomenclature. Multiple bond (HMBC) and 2-bond (H2BC) correlation cross peaks are identified by two numbers: the first corresponding to the 1 H atom and the second relating to the 13 C atom, e.g., HH6,4 describes the cross peak between H-6 and C-4 of 3-hydroxyhexanoate. Single correlation (HSQC) cross peaks are denoted by a single number (e.g., HP5, corresponds to the correlation of H-5 and C-5 for 3-hydroxypentanoate). Detailed analysis are provided as Appendix. Table 3 Summary of the 13 C-NMR shift values for a PHA sample derived from methanol and 7-octenoic acid metabolism. Carbon 3HB 3HP 3HHx 3HHx= 3HO= C-terminus 1 169.2 169.3 169.2 1 169.2 1 169.2 1 - 2 40.7 38.9 39.3 39.5-40.8 1 39.5-40.8 1 43.2 (CH 2 ) 3 67.7 71.9 70.7 70.06 70.8 64.5(CH) 4 19.7 26.9 35.9 38.2 33.3 22.5(CH 3 ) 5 - 9.4 18.4 132.7 24.2 - 6 - - 13.8 118.9 33.2 - 7 - - - - 138.3 - 8 - - - - 115.2 - 1 Actual value is uncertain due to overlap from major component (3HB). 2. A notable feature of the 2D HSQC spectrum was the presence of high frequency 1 H (5.0 ppm and 5.7 ppm) and 13 C (115 ppm and 135 ppm) resonances indicative of double bonds. These were assigned to the 3HHx= and 3HO= monomers. Further analysis of the NMR observations indicated that the double bond for these two monomers must be at the end of the alkyl chain; 3. Using the C3 resonances in the HSQC, we obtained a ratio of 3HB:3HO= (1:0.06), 3HB:3HHx= (1:0.05), 3HB:3HP (1:0.02), 3HB:3HHx (1:0.008, comparing methyls), 3HB:3HO (1:0.008, comparing methyls). Unfortunately, 2D NMR could not be performed with every sample due to cost and equipment availability. However for GC-FID analysis, it was now obvious that 3-hydroxyalkanoate monomers derived from 5-hexenoic acid would result in peaks for 3HHx= and 3HHx, while peaks for 3HP, 3HHx=, 3HHx, 3HO= and 3HO could be expected from metabolism with 7-octenoic acid. Consequently, GC-FID was routinely used thereafter to determine the composition of the PHAs produced and to look for the presence of C-C double bonds. Peaks for 3HB, 3HP and 3HHx were identified from corresponding methanolized PHA standards (see Materials and Methods). Since methyl esters resulting from 3-hydroxyalkenoates have a lower molecular weight compared to their saturated analogs, the peak that was recorded shortly prior to 3HHx was identified as 3HHx=. In addition to thereby related peaks for 3HB, 3HP, 3HHx= and 3HHx, GC chromatograms of PHA samples derived from methanol and 7-octenoic acid metabolism showed two more peaks that eluted at later time. From 2D NMR analysis it was concluded that these remaining peaks had to correspond to 3HO= and, eluted immediately thereafter, to 3HO. Peak areas were used to calculate the proportional monomeric compositions. Potential of the two recombinant strains for producing functionalized PHAs From the start, our intention was to develop M. extorquens strains capable of accumulating functionalized PHAs harboring C-C double bonds within the side chains. As done by others, the production of such PHAs requires feeding an unsaturated carboxylic acid in the hope that a portion of the acid will be incorporated into the PHA molecule by the recombinant PhaC1 or PhaC2 enzyme, thus, leading to the potential presence of C-C double bonds in the PHA side chains. Numerous shake flask studies were conducted using various unsaturated carboxylic acids, various feeding regimes and process conditions in the hope of identifying initial conditions that could lead to maximization of functionalized PHA production. Resulting PHA samples were generated and analyzed according to NMR and/or GC-FID analysis. As Table 4 indicates, the M. ex-pha C1 strain was able to produce PHA containing a small percentage, 0.75% +/- 0.57%, of C6:5 bond-containing material starting from C6=. The same strain, however, was very poor at incorporating C-C double bonds when either C8= or C11= was fed; only trace amounts of C-C double bonds were detected. Interestingly, the M. ex-pha C1 strain was able to incorporate significant quantities of a C5 unit starting from C8=. In this regard, C8= appeared much superior as chemical donor. Due to the high toxicity of the carboxylic acids used, cell densities were always low, generally between 1 and 2 g/L, on a dry weight basis. Table 4 Chemical composition of the PHAs extracted from recombinant M. extorquens cells grown on methanol + various unsaturated carboxylic acids 1 . Monomeric composition of the produced PHA 2 Strain Substrate [mol/L] DCW [g/L] C4:0 [mol%] C5:0 [mol%] C6:5 [mol%] C6:0 [mol%] C8:7 [mol%] C8:0 [mol%] phaC1 C6 = 3 16.82 1.59 +/- 0.23 98.11 +/- 0.89 - 0.75 +/- 0.57 1.14 +/- 0.34 - - C8 = 4 6.61 1.65 +/- 0.31 97.35 +/- 1.69 2.65 +/- 1.68 tr5 tr tr tr C11 = 6 3.43 1.11 +/- 0.28 100 - - - - - phaC2 C6 = 3 16.82 1.64 +/- 0.46 91.24 +/- 5.23 - 4.76 +/- 2.94 4.00 +/- 2.30 - - C8 = 4 6.61 1.65 +/- 0.31 84.69 +/- 4.87 5.28 +/- 1.13 2.66 +/- 0.98 2.36 +/- 0.98 5.01 +/- 1.82 tr C11 = 6 3.43 0.91 +/- 0.07 100 - - - - - 1 The PHA samples presented here were obtained from experiments shown in Figures 4 and 5; 2 CX:Y = double bond at position Y of monomeric chain length X; 3 Average +/- standard deviation, n = 3; 4 Average +/- standard deviation, n = 3; 5 tr = detected in trace quantity; 6 Average +/- standard deviation, n = 2. As illustrated in Table 4 , the M. ex-pha C2 strain appeared better than the M. ex-pha C1 strain at incorporating C6:5 bonds into PHA (C-C double bond at position 5 of C6 monomer chain length). An average of 4.76% +/- 2.94% of the PHA was made up of C6:5 bond-containing units upon feeding of C6=. The same strain appeared also better at incorporating a C6 monomer from either C6= or C8=. When C8= was fed, four of the five samples showed the obvious presence of C8:7 bond-containing units. The concentrations of C6:5 material and those of a C6 monomer were also significantly higher. A C5 monomer was detected in the samples derived from C6= but it was much more prevalent when C8= was fed, an observation valid for both recombinant strains. Feeding of C11=, as with the first strain, did not lead to production of PHA containing significant percentages of C-C double bonds. Cell densities, again, were quite low. Table 4 indicates that PHA containing small quantities of C6:5 material could be obtained with both strains after feeding C6= but only the M. ex- pha C2 strain could generate both C6:5 material and C8:7 material when fed C8=. The results listed in Table 4 show that PHA material containing modest concentrations of unsaturated C-C bonds could be obtained with the two strains. Three strong conclusions may be drawn from the results. (1) Strain M. ex-pha C2 was better at producing PHA material containing C-C double bonds; (2) the same strain was better at producing PHA material containing C-C bonds from C8=; (3) both strains gave very poor results with C11=. Several groups have observed or hypothesized that PhaC2 synthases have a lower substrate specificity than PhaC1 synthases [ 37 , 38 , 40 ]. Other groups have suggested that PhaC1 and PhaC2 synthases have different functions [ 43 , 44 , 48 ]. Contrary to observations made with other bacteria, our M. extorquens strains gave poor results with C11= and, therefore, only C6= or C8= proved to be adequate secondary substrates. Initial bioreactor studies: validation of shake flask results Growth of recombinant M. ex-phaC2 and production of PHAs in shake flasks were successfully validated in a 20 L-bioreactor (14 L as working volume). A total of 11 mL of 7-octenoic acid was added throughout the fermentation in four individual additions starting at 29 h of running time. After 96 h the final dry cell weight (DCW) reached 5 g/L with a 37% PHA content on a dry cell weight basis (data not shown). Polymers were extracted from biomass with chloroform and subjected to routine structural analysis. The monomeric composition reflected the results from shake flask experiments when 7-octenoic acid was fed (Tables 4 and 5 ). As discussed previously, PHAs produced by recombinant M. extorquens strains with two different PHA synthases were most likely blends of homo- and copolymers. To test for this hypothesis, extracted polymers were soaked in acetone to possibly separate SCL-PHAs from MCL-3HA-containing PHAs. As opposed to MCL-PHAs, poly(3-hydroxybutyrate) (PHB) is only soluble in acetone when it is amorphous [ 49 , 50 ]. PHB separated from biomass, however, is crystalline and insoluble in acetone. As listed in Table 5 , Soxhlet extraction was accompanied by a loss in PHA material. The remaining polymers were resolved into two different fractions by hot acetone. As expected, M. ex-phaC2 produced blends made up of SCL-PHAs and SCL/MCL-PHAs. Since the native PhaC enzyme is not capable of incorporating monomers other than SCL-3HA, the production of SCL/MCL-PHAs must have resulted from the action of the recombinant PhaC2 enzyme. Random abundance of the different monomers is likely the reason why some MCL-3HA units remained in the acetone-insoluble fraction. It is suggested here that the MCL-3HA content was very small in some copolymeric chains so that they could not dissolve in acetone. Table 5 Solvent fractionation to separate SCL/MCL-PHAs from SCL-PHAs. Monomeric composition of PHA 1 PHA sample Portion 2 [%] C4:0 [mol%] C5:0 [mol%] C6:5 [mol%] C6:0 [mol%] C8:7 [mol%] C8:0 [mol%] Blend 3 - 91.36 +/- 1.18 0.83 +/- 0.02 3.26 +/- 1.00 1.31 +/- 0.05 3.24 +/- 0.11 tr4 Acetone(-) 5 90.84 96.55 0.48 1.47 0.47 1.03 - Acetone(+) 6 4.63 55.85 +/- 6.43 3.33 +/- 0.70 21.72 +/- 7.81 6.98 +/- 1.56 12.12 +/- 3.48 tr Loss 7 4.52 - - - - - - 1 CX:Y = double bond at position Y of monomeric chain length X; 2 Portion after Soxhlet fractionation; 3 Chloroform extract from biomass (n = 2); 4 tr = detected in trace quantity; 5 Acetone-insoluble fraction (n = 1); 6 Acetone-soluble fraction (n = 3); 7 PHA loss during Soxhlet extraction. Our experimental PHA samples from shake flask and fermentation experiments exhibited mainly SCL-3HA monomers (3HB, C4 and 3HP = 3HV, C5). Due to the presence of two unrelated phaC genes in our recombinant strains, we conclude that a major portion of the materials consisted of PHB or PHBV resulting from the native PHA synthase. Since the PHA synthase coded by the phaC2 gene of P. fluorescens GK13 was shown to have a broad substrate specificity [ 23 ], we conclude that the remaining biopolyesters of the polymer blends belonged to the highly valuable class of short-chain-length/mediumchain-length PHAs (SCL/MCL-PHAs, 4 ≤ C ≤ 14) that have been proposed for many applications due to their desirable thermo-mechanical properties [ 23 ]. For this reason, PHBHx with a 3HHx content of 20% and less is currently under investigation for its potential to function as tissue engineering material ([ 14 ] and reference therein). Based on our results, we have successfully added functionality to this highly regarded class of PHAs by incorporating MCL-3HA units bearing terminal double bonds."
} | 7,062 |
32017312 | null | s2 | 1,170 | {
"abstract": "Inspired by biological motor proteins, that efficiently convert chemical fuel to unidirectional motion, there has been considerable interest in developing synthetic analogues. Among the synthetic motors created thus far, DNA motors that undertake discrete steps on RNA tracks have shown the greatest promise. Nonetheless, DNA nanomotors lack intrinsic directionality, are low speed and take a limited number of steps prior to stalling or dissociation. Herein, we report the first example of a highly tunable DNA origami motor that moves linearly over micron distances at an average speed of 40 nm/min. Importantly, nanomotors move unidirectionally without intervention through an external force field or a patterned track. Because DNA origami enables precise testing of nanoscale structure-function relationships, we were able to experimentally study the role of motor shape, chassis flexibility, leg distribution, and total number of legs in tuning performance. An anisotropic rigid chassis coupled with a high density of legs maximizes nanomotor speed and endurance."
} | 267 |
36307984 | PMC9700206 | pmc | 1,172 | {
"abstract": "Wood is an abundant and renewable feedstock for the production of pulp, fuels, and biobased materials. However, wood is recalcitrant toward deconstruction into cellulose and simple sugars, mainly because of the presence of lignin, an aromatic polymer that shields cell-wall polysaccharides. Hence, numerous research efforts have focused on engineering lignin amount and composition to improve wood processability. Here, we focus on results that have been obtained by engineering the lignin biosynthesis and branching pathways in forest trees to reduce cell-wall recalcitrance, including the introduction of exotic lignin monomers. In addition, we draw general conclusions from over 20 years of field trial research with trees engineered to produce less or altered lignin. We discuss possible causes and solutions for the yield penalty that is often associated with lignin engineering in trees. Finally, we discuss how conventional and new breeding strategies can be combined to develop elite clones with desired lignin properties. We conclude this review with priorities for the development of commercially relevant lignin-engineered trees.",
"conclusion": "Concluding remarks and perspectives Modification of lignin to improve wood processing has been successfully achieved in several forest tree species. However, commercialization of trees with reduced lignin levels or modified lignin is hindered by a number of factors. First, lignin engineering often results in lower biomass yields. Future efforts should therefore be made to address this issue, e.g., by tuning enzyme activity to obtain more subtle shifts in lignin levels, by reducing lignin content only in fibers, and by mutating genes located upstream of C4H (thus avoiding the accumulation of phenylpropanoids to potentially toxic levels) or genes encoding transcription factors that affect the entire lignin biosynthetic pathway. A second limitation is the lack of data on field-grown lignin-modified trees. Field trials of the most promising lignin-engineered lines are essential for the translation of basic research into applications. Importantly, in multi-year field trials, gradual fading of the reduction in wood lignin content of trees RNAi-engineered to produce less lignin has often been observed. Field trials of CRISPR-Cas-engineered trees are needed to further investigate whether this is a general phenomenon or whether confounding factors, such as variations in the level of downregulation caused by RNAi, are at play. A third hurdle is the lack of data on lignin engineering in elite lines. Here, establishing transformation protocols for elite lines would be helpful. In addition, protocols for gene editing that avoid the integration of the CRISPR-Cas-encoding DNA would further facilitate field studies and commercialization of new varieties. Furthermore, lignin engineering would benefit from evaluating the effects of stacking different lignin traits and stacking lignin traits with other desirable forestry traits (e.g., yield, wood density, pest resistance). Therefore, multiplex gene-editing strategies in which edits in different genes are stacked in a single elite line, followed by high-throughput screening of the engineered lines for combinations of edits, would further accelerate the development of lignin-engineered commercial forest trees.",
"introduction": "Introduction Fossil resources are currently the main feedstock for energy and organic compound synthesis. However, considering the high carbon footprint of the petrochemical industry, the development of renewable and carbon-neutral feedstocks is a critical societal challenge ( Vanholme et al., 2013b ; Galán-Martín et al., 2021 ; Huang et al., 2021 ; Yang et al., 2021 ). Lignocellulosic biomass is the most abundant renewable carbon source on earth and therefore a promising candidate. Wood is an important source of lignocellulosic biomass. It is mainly composed of secondary cell walls that are primarily made from cellulose, hemicellulose, and lignin. Cellulose can be used in the pulp and paper industry or can, together with hemicellulose, be hydrolyzed into monosaccharides that can be further converted into high-value compounds such as liquid biofuels, building blocks for bioplastics, cosmetics, etc. ( Bozell et al., 2007 ; Corma et al., 2007 ; Vanholme et al., 2013b ; Isikgor and Becer, 2015 ; de Vries et al., 2021b ). The additional valorization of lignin has recently been recognized as essential to enable the economic viability of lignocellulosic biorefining ( Liao et al., 2020 ; Abu-Omar et al., 2021 ). Lignin in wood of forest trees is predominantly built from monolignols that are biosynthesized by a series of enzymatic conversions starting from the amino acid phenylalanine ( Figure 1 ) ( Freudenberg, 1959 ; Boerjan et al., 2003 ; Vanholme et al., 2019 ). After their biosynthesis in the cytosol, monolignols are translocated to the cell wall ( Pesquet et al., 2010 , 2013 ; Derbyshire et al., 2015 ; Serk et al., 2015 ; Vermaas et al., 2019 ; Blaschek et al., 2020 ; Perkins et al., 2022 ). Notably, monolignols are translocated not only to the cell walls of the cells that produced them but also to those of neighboring cells ( Hosokawa et al., 2001 ; Tokunaga et al., 2005 ; Pesquet et al., 2013 ; Smith et al., 2013 , 2017 ; De Meester et al., 2018 , 2021 ). In the cell wall, monolignols are oxidized by peroxidases (PERs) and/or laccases (LACs) and subsequently polymerized into the lignin polymer through radical coupling, thereby generating a range of chemical bonds such as aryl ether (β-O-4), resinol (β-β), phenylcoumaran (β-5), biphenyl ether (5-O-4), spirodienone (β-1), and dibenzodioxocin (5-5/4-O-β) ( Berthet et al., 2011 , 2012 ; Lu et al., 2013 ; Ralph et al., 2019 ; Tobimatsu and Schuetz, 2019 ). Lignin amount and composition vary greatly within and among plant species, tissues, cell types, and cell-wall layers, and are influenced by developmental and environmental factors ( Vanholme et al., 2019 ). Hardwood (angiosperm) lignin is mainly built from the monolignols coniferyl, sinapyl, and (traces of) p -coumaryl alcohol, producing guaiacyl (G), syringyl (S), and p -hydroxyphenyl (H) units, respectively. By contrast, softwood (gymnosperm) lignin is composed mostly of G units with a minor fraction of H units. Depending on the plant species, other (less traditional) phenolic metabolites also act as monomers for lignification ( Figure 1 ) ( Ralph et al., 2019 ; Vanholme et al., 2019 ; del Río et al., 2022 ). For instance, poplar lignin typically incorporates coniferyl acetate, sinapyl acetate, and sinapyl p -hydroxybenzoate, as well as traces of coniferyl p -hydroxybenzoate, sinapyl benzoate and coniferyl ferulate ( Figure 1 ) ( Morreel et al., 2004a , 2004b ; Lu et al., 2004 ; Karlen et al., 2016 ; Zhao et al., 2021b ; de Vries et al., 2022 ). Figure 1 Metabolic pathways leading to lignin monomers. Only the route toward lignin monomers is shown; side branches of the pathway that result in soluble metabolites are not shown, with the exception of ferulic acid conjugates (see footnote 2 ). As enzymes may have different substrates and pathway perturbation may induce branching pathways, the figure includes only metabolic conversions for which strong evidence exists of involvement in biosynthesis of the shown end products. A question mark “?” means that the enzymes catalyzing the metabolic conversion(s) are not known. C3H, p -COUMARATE 3-HYDROXYLASE; CHI, CHALCONE ISOMERASE; other abbreviations can be found in the manuscript text. 1 The exact structures of the 4-dihydroxybenzoate conjugates are not known. 2 Only a small portion of the ferulic acid pool is integrated into the lignin polymer; the majority of the pool is metabolized to soluble ferulic acid conjugates. 3 The conversion of coniferin to coniferyl alcohol (which subsequently serves as a monomer for lignification) and glucose occurs in the apoplast of gymnosperm trees. Studies in Ginkgo biloba , Pinus contorta , Picea abies (Norway spruce), and Chamaecyparis obtusa (Japanese cypress) indicate that coniferin is a transport form of coniferyl alcohol in these species ( Samuels et al., 2002 ; Tsuyama et al., 2013 ; Aoki et al., 2016 ; Väisänen et al., 2020 ). Currently, narrow profit margins limit the economic feasibility of employing fast-growing woody feedstocks, such as poplar, willow, and eucalyptus, as dedicated energy and specialty chemical crops at an industrial scale ( Mahon and Mansfield, 2019 ). Because lignin is the main contributor to the recalcitrance of lignocellulosic biomass toward deconstruction, numerous research efforts have focused on altering lignin amount and/or composition, aiming to increase wood processing efficiency ( Chanoca et al., 2019 ; Bryant et al., 2020 ). However, the in planta effects of pathway perturbations are often unpredictable because the flux through the lignin biosynthesis pathway is regulated at multiple levels ( Figure 2 ): (i) at the transcriptional level ( Zhong et al., 2010 ; Bonawitz et al., 2014 ; Ohtani and Demura, 2019 ); (ii) by post-translational protein modifications ( Wang et al., 2015 ); (iii) by the abundance of enzyme-inhibiting pathway intermediates ( Wang et al., 2014 ; Eudes et al., 2016 ; Yokoyama et al., 2021 ); (iv) by the compartmentalization of substrates ( Widhalm et al., 2015 ; Eudes et al., 2016 ; Guo et al., 2018 ; Vanholme et al., 2019 ); (v) by the different activities and spatial expression patterns of gene family members ( Li et al., 2015 ; Vanholme et al., 2019 ); (vi) by the formation of protein complexes ( Chen et al., 2011 ; Widhalm et al., 2015 ; Gou et al., 2018 ; Yan et al., 2019 ); (vii) by the availability of co-factors and -substrates ( Tang et al., 2014 ; Vanholme et al., 2019 ; Hu et al., 2022 ); (viii) by pathway intermediate-triggered proteolysis of pathway enzymes ( Guan et al., 2022 ); (ix) by branching pathways diverting the flux away from monolignols toward other sinks and products, such as flavonoids, benzenoids, phenylpropanoids, and monomer-glucosides ( Vanholme et al. (2012) ; and (x) by the different parallel routes through which the pathway flux can flow via enzymes that often accept multiple substrates ( Vanholme et al., 2019 ; Tsai et al., 2020 ). In addition, the outcome of the same pathway perturbation(s) in different species can differ significantly ( Vanholme et al., 2019 ); e.g., CAFFEOYL SHIKIMATE ESTERASE ( CSE )-knockout Arabidopsis , Medicago , and poplar display huge differences in the frequency of H units incorporated into their lignin and the severity of the associated yield penalty ( Vanholme et al., 2013c ; Ha et al., 2016 ; de Vries et al., 2021a ). Figure 2 Schematic overview of multiple levels of lignin pathway regulation in trees. Arrows with filled arrowheads represent metabolic conversions, and arrows with hollow arrowheads represent translocation, protein modification, or signaling. Pac-man-like spheres represent enzymes, and polygons represent metabolites. For poplar, a mathematical model was generated that estimates how changing the expression of pathway genes affects protein abundance, metabolite concentrations, metabolic flux, and various wood traits such as lignin amount and composition, biomass yield, and saccharification efficiency ( Wang et al., 2018 ). Even so, this model also has its limitations; to generate the model, the analysis was restricted to one specific developmental stage, environmental condition, and poplar accession. Moreover, some factors that affect flux through the pathway, such as the formation of C3′H-C4H protein complexes, spatial distribution of substrates and enzymes (all xylary cell types were pooled to generate the model), and further metabolization of compounds (e.g., glycosylation), were not (yet) taken into account. It will therefore be crucial to assess the value of a lignin-engineering strategy and evaluate it under relevant forestry practices in economically relevant biomass germplasm such as (elite) poplar or eucalyptus clones. In this review, we give an overview of the results obtained by engineering the levels of endogenous and exotic lignin building blocks in forest trees. In addition, based on data derived from greenhouse- and field-grown lignin-engineered trees, we discuss the link between lignin reduction and growth phenotypes on the one hand and between lignin reduction and the age of the tree and the environment in which it was grown on the other hand. We explore the possible causes of the frequently observed yield penalty in lignin-engineered trees and discuss strategies to avoid this yield penalty. We also address breeding strategies for the modification of lignin in forest trees, including traditional breeding with genomic selection and modern breeding techniques such as transgene expression and gene editing. We conclude this article with perspectives on the development and use of commercially relevant lignin-engineered trees."
} | 3,254 |
33621334 | PMC8022974 | pmc | 1,173 | {
"abstract": "Abstract Corals of the family Acroporidae are key structural components of reefs that support the most diverse marine ecosystems. Due to increasing anthropogenic stresses, coral reefs are in decline. Along the coast of Okinawa, Japan, three different color morphs of Acropora tenuis have been recognized for decades. These include brown (N morph), yellow green (G), and purple (P) forms. The tips of axial polyps of each morph exhibit specific fluorescence spectra. This attribute is inherited asexually, and color morphs do not change seasonally. In Okinawa Prefecture, during the summer of 2017, N and P morphs experienced bleaching, in which many N morphs died. Dinoflagellates (Symbiodiniaceae) are essential partners of scleractinian corals, and photosynthetic activity of symbionts was reduced in N and P morphs. In contrast, G morphs successfully withstood the stress. Examination of the clade and type of Symbiodiniaceae indicated that the three color-morphs host similar sets of Clade-C symbionts, suggesting that beaching of N and P morphs is unlikely attributable to differences in the clade of Symbiodiniaceae the color morphs hosted. Fluorescent proteins play pivotal roles in physiological regulation of corals. Since the A. tenuis genome has been decoded, we identified five genes for green fluorescent proteins (GFPs), two for cyan fluorescent proteins (CFPs), three for red fluorescent proteins (RFPs), and seven genes for chromoprotein (ChrP). A summer survey of gene expression profiles under outdoor aquarium conditions demonstrated that (a) expression of CFP and REP was quite low during the summer in all three morphs, (b) P morphs expressed higher levels of ChrP than N and G morphs, (c) both N and G morphs expressed GFP more highly than P morphs, and (d) GFP expression in N morphs was reduced during summer whereas G morphs maintained high levels of GFP expression throughout the summer. Although further studies are required to understand the biological significance of these color morphs of A. tenuis , our results suggest that thermal stress resistance is modified by genetic mechanisms that coincidentally lead to diversification of color morphs of this coral.",
"introduction": "Introduction One of the most critical issues facing the human race is warming of our planet. Anthropogenic activities have harmed the environment in various ways, and coral reefs have been especially impacted ( Hoegh-Guldberg et al. 2007 ; Hughes et al. 2017 ). In spite of the fact that coral reefs occupy only 0.2% of the ocean area, they are estimated to harbor about one-third of all described marine species ( Knowlton et al. 2010 ; Fisher et al. 2015 ), suggesting that coral reefs are the most diverse marine ecosystems on Earth ( Wilkinson 2008 ). Scleractinian corals, a keystone component of calcium-carbonate based reefs, form obligate endosymbioses with photosynthetic dinoflagellates of the family Symbiodiniaceae, which supply the vast majority of their photosynthetic products to the host corals ( Yellowlees et al. 2008 ). However, corals now face a variety of environmental stresses, including increasing surface seawater temperatures, decimation by outbreaks of crown-of-thorns starfish, and acidification and pollution ofoceans ( Hoegh-Guldberg et al. 2007 ; Uthicke et al. 2009 ; Burke et al. 2011 ; Uthicke et al. 2015 ; Hughes et al. 2017 ). Vivid coloration of corals has been attributed to the emission of a family of fluorescent proteins (FPs), green fluorescent proteins (GFPs), red fluorescent proteins (RFPs), cyan fluorescent proteins (CFPs), and nonfluorescent blue/purple chromoprotein (ChrP) ( Dove et al. 2001 ; Kelmanson and Matz 2003 ; Salih et al. 2000 ; Smith et al. 2013 ). Corals exhibit FP-mediated color polymorphism. FPs and ChrP are thought to contribute to their acclimatization potential for the following reasons ( Kelmanson and Matz 2003 ; Dove 2004 ; Paley 2014 ; Gittins et al. 2015 ; Jarett et al. 2017 ; Takahashi-Kariyazono et al. 2018 ). First, expression of FP genes is modified in relation to environmental changes. For example, FP gene expression in adult Acropora individuals is influenced by external stimuli, such as light, heat, and injury ( Takahashi-Kariyazono et al. 2018 ). Expression levels of CFP , GFP, RFP , and ChrP increase according to light intensity ( D’Angelo et al. 2008 ; Roth et al. 2010 ). CFP expression is down-regulated in dark stress ( DeSalvo et al. 2012 ) as well as heat stress ( Roth and Deheyn 2013 ), although it is up-regulated in response to injury ( D’Angelo et al. 2012 ). Second, roles of FPs in acclimatization have been suggested. For example, FPs protect corals and their symbionts by absorbing high-energy ultraviolet radiation and re-emitting it as lower energy visible light ( Bollati et al. 2020 ). FPs also reduce oxidative stress to corals as well as to their dinoflagellate symbionts ( Salih et al. 2000 ). To survive corals must somehow adapt to increasingly stressful environments ( Skelly et al. 2007 ). \n Acropora tenuis is one of the major scleractinian corals along the coast of Okinawa, Japan ( Omori et al. 2016 ). Most A. tenuis appear brownish ( Figure 1, A–D ), reflecting the color of Symbiodiniaceae that they host. Acropora tenuis exhibits faster growth than other Acropora species, forming colonies approximately 30 cm in diameter within 3–5 years ( Figure 1, A and B ; Iwao et al. 2010 ). In 1998, along the Okinawa coast, various Acropora species suffered extensive bleaching and many died extensive bleaching and many died. After several years, they gradually recovered ( Kimura et al. 2014 ). While color morphs of A. tenuis have been described ( Nishihira and Veron 1995 ), divers in Okinawa noticed the re-appearance of yellowish green ( Figure 1, E and F ) and purple A. tenuis morphs ( Figure 1G ) in addition to brownish ones. We thought that the color polymorphism of A. tenuis might be associated with its potential to resist to stress. Figure 1 Color morphs of A. tenuis along the Okinawa coast. (A and B) Colonies with different color morphs are growing in shallow water along the coast of Okinawa. (A) Healthy condition of corals and (B) bleaching of some colonies during summer season. (C and D) Two “wild-type” morphs, NO (C) and NB (D). Both appear generally brownish while the tips of axial polyps are orange in NO (Cʹ) and blue in NB (Dʹ). (E and F) Two greenish morphs, GO (E) and GB (F ) . Tips of axial polyps are orange in GO (Eʹ) and blue in GB (Fʹ). (G) A purple colony (P) and tips of its axial polyps (Gʹ). As mentioned above, there are many studies on the expression and function of FP genes in response to environmental changes. However, few studies have addressed this question comprehensively at the genomic level. This is partly because decoding of coral genomes has only occurred recently, first in Acropora digitifera in Shinzato et al. (2011) , followed by several other species ( e.g. , Cunning et al. 2018 ; Ying et al. 2019 ). A genome-wide survey of FP members in corals has been reported only in the A. digitifera genome (Shinzazo et al. 2012; Takahashi-Kariyazono et al. 2018 ). The genome of A. tenuis has just been decoded ( Shinzato et al. 2021 ). Therefore, we attempted to determine the relationships of color morphs with their potential for environmental stress response by characterizing all FP genes present in the A. tenuis genome. We first examined whether the three-color morphs show different bleaching response to stresses, including higher seawater temperature. Then, we examined whether differences in the clade of Symbiodiniaceae are associated with differences in bleaching grade of the three-color morphs. Finally, we surveyed all candidate genes for FPs in the A. tenuis genome, and examined their expression profiles in the three-color morphs in response to rising summer surface seawater temperatures.",
"discussion": "Discussion There are three major color morphs of A. tenuis along the Okinawa coast, brown, green, and purple ( Nishihira and Veron 1995 ; Figure 1 ). They exhibit different profiles of fluorescence in the axial polyps when excited at 405 nm ( Figure 4 ). We expect that differences in fluorescence are associated with coloration of this coral, although detailed molecular, biochemical, and biophysical mechanisms underlying these differences should be addressed in future studies. Based on observations for a decade, these color morphs are stable, since color polymorphism is asexually heritable and color morphs show no seasonal variation, although the brightness of color changes. These characters have also been shown in Acropora millepora ( Paley 2014 ). Although it remains to be determined whether it is also sexually heritable, color polymorphism is likely caused by genetic variation. Field studies in 2017 showed that N morphs underwent extensive bleaching ( Figure 2, A and B ), and 10–15% of NB morphs eventually died ( Figure 2B ). In contrast, G morphs did not show bleaching ( Figure 2, C and D ). These results indicate that sensitivity to summer environmental stress differs among the three-color morphs and that N is the most sensitive while G is resistant. Along the Okinawa coast, the N color morph is most abundant ( Figure 1 and Supplementary Figure S1), while G and P are not as common. This suggests that the N morph is probably the ancestral or wild type, while G and P color morphs are more recent. If this feature is heritable, it is tempting to speculate that in Okinawa, A. tenuis has acquired the capacity to resist summer environmental stresses by developing new color morphs. In other words, polyp color polymorphism may be a strategy to survive severe summer environmental stresses, provided that the current rate of temperature change does not overwhelm the coral’s capacity to adapt. Bleaching responses are complicated, varying among colonies, taxa, and events ( van Woesik et al. 2011 ). Scleractinian corals form obligate endosymbioses with photosynthetic dinoflagellates of the family Symbiodiniaceae, and host-symbiont interactions contribute to differential bleaching susceptibility ( Enriquez et al. 2005 ; Hawkins et al. 2014 ; Wooldridge 2014 ). Acropora corals lost a gene for cystathionine ß-synthase, an essential enzyme for cysteine biosynthesis, and they depend upon symbionts to produce cysteine ( Shinzato et al. 2011 , 2021 ), partially explaining the higher sensitivity of Acropora to bleaching than other coral taxa. Because bleaching is usually caused by escape and/or death of symbionts, declining photosynthetic activity of Symbiodiniaceae may portend bleaching, as shown by previous studies ( Smith et al. 2013 ; Gittins et al. 2015 ). Indeed, decreased photosynthetic activity was detected in July in NB and P morphs ( Figure 3 ). In contrast, GO and GB morphs maintained photosynthetic activity comparable to that in other months. Since this suggests that G color morphs provide a more suitable physiological environment for symbionts than N and P morphs during times of high thermal stress. Genetic and molecular mechanisms involved in this relationship are intriguing and should be addressed in future studies. The family Symbiodiniaceae has recently been reorganized into nine clades or genera (A–I), according to new analyses of combinatorial data ( LaJeunesse et al. 2018 ). It has been suggested that in general, corals hosting clade D ( Durusdinium ) are more heat-resistant than those hosting clade C ( Cladocopium ) or that specific Symbiodinium phylotypes, such as D1, D1–4, C15, and A3 are exceptionally thermotolerant, while others ( e.g. , C3, C7, B17, and A13) are thermosensitive ( Silverstein et al. 2011 ; Tonk et al. 2014 ). Therefore, it seemed possible that differing capacities of A. tenuis color morphs to resist higher summer seawater temperatures might be due to different zooxanthella hosted by the three-color morphs. However, this is not the case in Okinawa A. tenuis , because all three host a very similar repertoires of clade C Symbiodinium , especially C1 and C5 types ( Figure 5 ). \n Acropora tenuis is useful as an experimental system because its approximately 400-Mb genome has been decoded and is thought to contain 22,802 protein-coding genes ( Shinzato et al. 2021 ). Moreover, approximately 95% of the gene models are confirmed with corresponding mRNAs. Therefore, we were able to investigate not only genes for GFP, CFP and RFP, and ChrP, but also their expression profiles in colonies with different color morphs during the summer of 2017. We found that all three morphs possess an identical array of fluorescence-related genes and that expression of these genes was confirmed in all morphs using RNA-seq. This indicates that color polymorphism is not caused by mutations in the genes themselves. Instead, molecular and genic mechanisms that control transcriptional activity or quantitative regulation of gene products probably produce the three-color morphs. Genetic and genic mechanisms underlying FP-mediated color polymorphisms and adaptation potential to variable environmental conditions have been studied in Acropora millepora ( D’Angelo et al. 2008 ; Smith et al. 2013 ; Gittins et al. 2015 ). In this species, an RFP gene, amiFP597 , is essential for color polymorphism. amiFP597 exists in multiple copies with a particular promoter type in the genome, and the number of gene copies is strongly correlated with the level of gene expression. Higher levels of gene expression are found in more intensely red morphs, which show higher resistance to strong insolation ( Gittins et al. 2015 ). This suggests the presence of variable genetic mechanisms in color polymorphism, depending on coral species. Interestingly, the P morph of A. tenuis exhibits a gene expression profile different from those of the N and G morphs. Specifically, the P morph shows higher ChrP gene expression, especially s0096.g4 ( Figure 7 ). In addition, the N and G morphs differ in expression levels of GFP genes, especially s0297.g29 and s02897.g27 ( Figure 7 ). Higher gene expression was evident in the G morph compared to the N morph ( Figure 7 ). Because the G morph is the most resistant to summer environmental stress, it is likely that stress resistance varies with expression levels of GFP genes. We expect that different control mechanisms of GFP gene expression in N and G morphs explain their different thermal tolerance. We will address this question in future studies. Coral bleaching is caused by multiple, complex environmental stresses. Except for outbreaks of crown-of-thorns starfish, typhoons, and diseases, the most deleterious influence is surface seawater temperature warming. Strong solar radiation also likely contributes to environmental stress. Seawater in Okinawa Prefecture is very transparent; thus, UV irradiation acts in concert with higher sea temperatures to cause bleaching. FPs and chromoproteins absorb UV radiation, rendering corals and their symbionts more resistant to solar stress ( Dove et al. 2001 ; Dove 2004 ; Banaszak and Lesser 2009 ; Salih et al. 2010). We need to better understand the relationship between solar stress and different expression profiles of genes for GFP and ChrP. Divers and marine researchers have noticed that some corals, sometimes called “super corals” and/or “super coral reefs” ( Dance 2019 ), are able to survive severe environmental stresses. Nonetheless, almost nothing is known about how such super corals have appeared in the reefs. Our present results regarding different stress tolerances of three A. tenuis color morphs may help to explain the resilience of these super corals."
} | 3,954 |
31474959 | PMC6706786 | pmc | 1,174 | {
"abstract": "Metabolic flexibility in aerobic methane oxidizing bacteria (methanotrophs) enhances cell growth and survival in instances where resources are variable or limiting. Examples include the production of intracellular compounds (such as glycogen or polyhydroxyalkanoates) in response to unbalanced growth conditions and the use of some energy substrates, besides methane, when available. Indeed, recent studies show that verrucomicrobial methanotrophs can grow mixotrophically through oxidation of hydrogen and methane gases via respiratory membrane-bound group 1d [NiFe] hydrogenases and methane monooxygenases, respectively. Hydrogen metabolism is particularly important for adaptation to methane and oxygen limitation, suggesting this metabolic flexibility may confer growth and survival advantages. In this work, we provide evidence that, in adopting a mixotrophic growth strategy, the thermoacidophilic methanotroph, Methylacidiphilum sp. RTK17.1 changes its growth rate, biomass yields and the production of intracellular glycogen reservoirs. Under nitrogen-fixing conditions, removal of hydrogen from the feed-gas resulted in a 14% reduction in observed growth rates and a 144% increase in cellular glycogen content. Concomitant with increases in glycogen content, the total protein content of biomass decreased following the removal of hydrogen. Transcriptome analysis of Methylacidiphilum sp. RTK17.1 revealed a 3.5-fold upregulation of the Group 1d [NiFe] hydrogenase in response to oxygen limitation and a 4-fold upregulation of nitrogenase encoding genes ( nifHDKENX ) in response to nitrogen limitation. Genes associated with glycogen synthesis and degradation were expressed constitutively and did not display evidence of transcriptional regulation. Collectively these data further challenge the belief that hydrogen metabolism in methanotrophic bacteria is primarily associated with energy conservation during nitrogen fixation and suggests its utilization provides a competitive growth advantage within hypoxic habitats.",
"introduction": "Introduction Aerobic methane oxidizing bacteria (methanotrophs) serve as the primary biological sink for the potent greenhouse gas methane (CH 4 ) ( Kirschke et al., 2013 ). Methanotrophs grow by oxidizing CH 4 to methanol with a particulate or soluble methane monooxygenase enzyme (pMMO/sMMO) and subsequently yield reducing equivalents (e.g., NADH) for cellular respiration and biosynthesis through the oxidation of methanol to carbon dioxide (CO 2 ). The gammaproteobacterial (Type I) and alphaproteobacterial (Type II) methanotrophs generate biomass by assimilating the intermediates formaldehyde or formate via the ribulose monophosphate (RuMp) or serine pathways ( Hanson and Hanson, 1996 ) respectively, whereas the verrucomicrobial methanotrophs oxidize methanol directly to formate ( Keltjens et al., 2014 ) and generate biomass by fixing inorganic carbon (CO 2 ) via the Calvin–Benson–Bassham cycle ( Khadem et al., 2012b ). Despite the apparent restriction of most methanotrophs to grow on one carbon compounds (C1), they thrive at the interface of various oxic/anoxic habitats (e.g., peat bogs, forest soils, wetlands, rice paddies and geothermal environments) ( Dunfield et al., 2007 ; Singh et al., 2010 ; Knief, 2015 ), where the availability of oxidant (O 2 ), energy and carbon resources for growth is likely to fluctuate. Given the methane monooxygenase reaction (CH 4 + O 2 + [NAD(P)H + H + ]/QH 2 CH 3 OH + NAD(P) + /Q + H 2 O) and the aerobic respiratory chain require a continual source of reductant and oxidant, methanotrophic bacteria must regulate their carbon, energy and resource allocation to fulfill metabolic demands for cellular growth and persistence ( Hanson and Hanson, 1996 ). Many bacterial species, including methanotrophs, accumulate biopolymers (e.g., glycogen, polyhydroxyalkanoates), phospholipids, and intracellular osmolytes (e.g., ectoine, sucrose) ( Strong et al., 2016 ) in response to unbalanced growth conditions. This allows resources to be strategically conserved for assistance in times of starvation. The biosynthesis of glycogen, a highly branched polysaccharide consisting of α-1,4 bonded glucose residues with additional α-1,6 branched sidechains, is a common metabolic strategy for carbon storage that is shared among evolutionarily distant species ( Wilson et al., 2010 ). Glycogen production has been widely described within Type I methanotroph species ( Linton and Cripps, 1978 ; Eshinimaev et al., 2002 ) and the production of this compound has recently been reported in the verrucomicrobial methanotroph, Methylacidiphilum fumarolicum SolV ( Khadem et al., 2012a ). The physiological role of glycogen production in methanotrophs is not precisely understood, although it is believed to serve a role in environmental survival during periods of starvation and has been implicated to symbiotic performance, colonization and virulence ( Bonafonte et al., 2000 ; McMeechan et al., 2005 ; Bourassa and Camilli, 2009 ; Wilson et al., 2010 ). Although the accumulation of intracellular glycogen may occur optimally during exponential growth ( Gibbons and Kapsimalis, 1963 ; Eidels and Preiss, 1970 ), its synthesis is typically associated with entry into stationary phase when growth is limited due to the limitation of some critical nutrient (i.e., nitrogen, phosphate) or in the presence of excess carbon ( Wilson et al., 2010 ). In bacteria, the biosynthesis of glycogen occurs by utilizing ADP-glucose as the glycosyl donor for polymer extension ( Preiss, 1984 ). The precise mechanisms governing glycogen biosynthesis in bacteria, however, remain obscure. It is likely energy availability and redox status play a primary role in regulating glycogen biosynthesis, as ATP acts as substrate for the ADP-glucose producing reaction catalyzed by glucose-1-phosphate adenylyltransferase ( Preiss, 1984 ). To remain competitive within dynamic environments ( Knief et al., 2003 ; Tavormina et al., 2010 ), some methanotrophs supplement CH 4 usage with other energy-yielding strategies ( Dedysh and Dunfield, 2010 ). Several recent studies have revealed a few strains, notably Methylocella silvestris , utilize a suite of carbon and energy substrates, including simple organic acids, alcohols and short-chain alkane gases ( Dedysh et al., 2005 ; Crombie and Murrell, 2014 ). Aerobic H 2 metabolism has also been shown in a range of methanotrophs ( Chen and Yoch, 1987 ; Shah et al., 1995 ; Hanczar et al., 2002 ) and a wide range of hydrogenases have been shown to be distributed in methanotroph genomes ( Greening et al., 2016 ). While H 2 oxidation was originally implicated in energy conservation in response to N 2 fixation ( Takeda, 1988 ), more recent findings indicate that H 2 serves a multifaceted role in the growth and survival of these bacteria. Of the verrucomicrobial methanotrophs, the activity of respiratory-linked group 1d hydrogenases can provide sufficient energy to sustain chemolithoautotrophic growth on H 2 alone ( Mohammadi et al., 2016 ; Carere et al., 2017 ). Further, mixotrophic growth (H 2 and CH 4 ) in the thermoacidophile Methylacidiphilum sp. RTK17.1 has been observed under O 2 -limiting conditions and is proposed to provide a competitive advantage over obligate methanotrophy at oxic/anoxic soil boundaries within geothermal environments ( Carere et al., 2017 ). This suggests that the additional energetic input of H 2 may counter the effect of otherwise unbalanced growth conditions. The influence of H 2 metabolism on the production of intracellular energy reservoirs, commonly associated with unbalanced growth, within methanotrophic bacteria has yet to be elucidated. In this work, we investigate the effect of H 2 metabolism on glycogen production within the methanotroph, Methylacidiphilum sp. RTK17.1. Chemostat cultivation was performed during O 2 -replete and O 2 -limited cultivation, in the presence of NH 4 + or N 2 , with or without H 2 in the headspace, to determine the influence of H 2 metabolism on observed growth rates, biomass production characteristics and transcriptional regulation. We show that cellular growth rates, molar growth yields, and the allocation of resources between protein and glycogen production vary depending on the supply of H 2 , O 2 , and nitrogen (as NH 4 + or N 2 ). Transcriptome data provided a basis of findings, showing significant differential regulation of operons encoding the group 1d [NiFe]-hydrogenase, methane monooxygenases, and nitrogenase between the conditions. In turn, these findings enhance understanding of the physiological strategies that methanotrophs use to grow and survive in different environments.",
"discussion": "Discussion The accumulation and storage of carbon and energy as polymeric reserves is a common strategy employed by microorganisms during unbalanced growth to fortify them against periods of environmental starvation ( Wilson et al., 2010 ). Nitrogen limitation is often cited as triggering the accumulation of carbon-rich reserve polymers ( Wanner and Egli, 1990 ), but there is a lack of detailed understanding of the underlying mechanisms responsible for their production. As with many heterotrophic species, methanotrophs often produce carbon-rich polymers; however despite the prevalence of glycogen production within the Gammaproteobacteria methanotrophs ( Pieja et al., 2011a ), research has primarily focused on the physiology of polyhydroxybutyrate storage in the Alphaproteobacteria methanotrophs ( Pieja et al., 2011a , b ; Sundstrom and Criddle, 2015 ). The requirement for organic carbon compounds to provide both the respiratory energy and carbon necessary for anabolic processes for methanotrophs makes it difficult to untangle the roles of nitrogen, carbon, and energy availability (e.g., ATP) in the production of intracellular glycogen. However, as verrucomicrobial methanotrophs fix CO 2 for carbon and supplement their energy requirements via the oxidation of H 2 ( Mohammadi et al., 2016 ; Carere et al., 2017 ), this affords an opportunity (obfuscated by the Type I and II methanotrophs) to investigate the influence of nitrogen, carbon, and energy availability independently. Our findings show that H 2 oxidation influences the production of intracellular glycogen reservoirs in the thermoacidophilic methanotroph, Methylacidiphilum sp. RTK17.1. During chemostat experiments, the maximum glycogen content of Methylacidiphilum sp. RTK17.1 occurred within cells grown in the absence of H 2 , under nitrogen- and O 2 - limiting growth conditions [48.86 (± 4.32)%; Table 1 and Figure 1A ]. With respect to other studies reporting on the production of carbon storage polymers in methanotroph species, the glycogen content values observed for Methylacidiphilum sp. RTK17.1 are generally congruent. In the closely related thermoacidophile, Methylacidiphilum fumarolicum SolV, a maximum glycogen content of 36% (w/w) was observed ( Khadem et al., 2012a ) in nitrogen-limited, batch grown cells and a similar value (33% w/w) has been reported in the halotolerant methanotroph Methylotuvimicrobium alkaliphilum 20Z [formerly Methylomicrobium alkaliphilum 20Z ( Khmelenina et al., 1999 ; Orata et al., 2018 )]. A maximum of 42.8 (± 17.5)% (w/w) glycogen has been reported in the industrially promising methanotroph, Methylotuvimicrobium buryatense 5GB1 [formerly Methylomicrobium buryatense 5GB1 ( Orata et al., 2018 )], during batch-growth on methanol, with up to 13.1 (± 4.0)% (w/w) glycogen reported during O 2 -limited chemostat growth on methane ( Gilman et al., 2015 ). The variability in reported glycogen content within methanotrophs (intraspecies) and between experimental trials (interspecies) is almost certainly a consequence of both underlying physiological characteristics and the inherent challenges associated with characterizing dynamic batch growth environments. We therefore sought to perform a series of carbon-excess (CO 2 and CH 4 ) steady-state experiments to gain insight into the mechanisms governing glycogen production in Methylacidiphilum sp. RTK17.1. With respect to O 2 and nitrogen limitation, rates of growth (inferred from observed biomass productivity rates), and changes to glycogen, and to a lesser extent total protein and amino acid contents, were consistent with a cell’s expected response to unbalanced growth conditions. While previous studies have reported significant changes to the amino acid composition of Staphylococcus aureus cultures in response to variable environmental conditions ( Alreshidi et al., 2015 , 2016 ), we observed no change to the relative abundance of specific amino acid residues in Methylacidiphilum sp. RTK17.1 cultures under the conditions tested. Nevertheless, a 144% increase in glycogen content was observed during unbalanced growth following the removal of H 2 gas supply. Depriving Methylacidiphilum sp. RTK17.1 cultures of the respiratory energy gains afforded from H 2 gas oxidation is demonstrative of how this strain dynamically allocates carbon, nitrogen and energy resources. The observation that Methylacidiphilum sp. RTK17.1 cells produce glycogen and grow more slowly in response to oxygen limitation is consistent with the occurrence of glycogen within the obligate chemolithoautotroph Hydrogenovibrio marinus when grown on H 2 and CO 2 under O 2 -limiting conditions ( Nishihara et al., 1993 ). Similarly, production of polyhydroxybutyrate (PHB) has been reported within heterotrophically grown cultures of Azotobacter beijerinckii in response to oxygen limitation ( Senior et al., 1972 ). We speculate that in the absence of sufficient oxygen, both glycogen and PHB reserves likely serve to not only store carbon and energy, but to maintain intracellular redox state. It is also plausible that the anabolic activities required for cell division (i.e., protein, DNA and RNA synthesis) were constrained by ATP availability under O 2 -limiting conditions. Given protein synthesis requires approximately 19 times more ATP (mmol ATP (g macromolecule) –1 ) than for saccharide polymerization ( Stouthamer, 1979 ; Russell and Cook, 1995 ; Russell, 2007 ), even considering the ATP requirements of CO 2 fixation, glycogen biosynthesis likely represents an energetic ‘cost’ savings for Methylacidiphilum sp. RTK171 compared to the ATP-demands of cell growth. These modest increases to intracellular glycogen content in response to O 2 limitation are unlikely to negatively impact biomass yields (Y ATP ; Russell and Cook, 1995 ) while also benefiting cell survivability during periods of starvation. The additional burden imposed by nitrogen limitation not only created unbalanced growth conditions with respect to carbon and nitrogen, but also increased the cell’s ATP requirement via the nitrogenase reaction (N 2 + 8H + + 16ATP → 2NH 3 + H 2 + 16ADP). Under these growth conditions, glycogen accounted for nearly half of Methylacidiphilum sp. RTK17.1 cell mass. Supplementing CH 4 oxidation with an alternative source of respiratory energy (H 2 ), however, was sufficient to offset the ATP burden imposed by N 2 fixation and consequently the production of intracellular glycogen was reduced and growth rates increased. An alternative explanation for our chemostat observations is that synthesis of glycogen during energy-limiting conditions serves as a strategy for ‘metabolic anticipation’. The combined conditions of low O 2 , nitrogen, and H 2 availability are highly limiting for a cell and further resource deprivation is likely to trigger a transition from growth to persistence. Thus, disproportionately allocating biomass into storage compounds under this condition may serve as a ‘bet-hedging’ strategy to enable longer-term survival when conditions worsen. Indeed, the synthesis and storage of intracellular carbon polymers is commonly associated with an increase in viability during periods of environmental starvation. As with PHB, glycogen catabolism supplies reduced electron carriers (e.g., NADH) into the respiratory chain, thereby enabling the continuation of metabolic processes in the absence of an exogenous energy supply (e.g., CH 4 or H 2 ). A reduced lag phase following CH 4 starvation has previously been linked to the catabolism of glycogen reservoirs within the methanotroph M. fumarolicum SolV ( Khadem et al., 2012a ). Likewise, in Methylacidiphilum sp. RTK17.1 cultures, we interpret depletions in cellular glycogen content throughout prolonged incubations at 4°C (in the absence of CH 4 ) as evidence it was being consumed to promote survival ( Supplementary Figure S1 ). Within oxic/anoxic habitats, it seems evident that Methylacidiphilum sp. RTK17.1 distributes carbon, energy and nitrogen resources during methanotrophic or mixotrophic growth to fulfill the metabolic demands imposed for cell persistence and/or proliferation ( Hanson and Hanson, 1996 ). Similar phenomena of metabolic anticipation have been observed in other species, for example mycobacteria, which accumulate storage compounds such as triacylglycerols during the early hypoxic response ( Daniel et al., 2004 ; Eoh et al., 2017 ). While hydrogenase, methane monooxygenase and nitrogenase all displayed evidence of significant transcriptional regulation in response to O 2 and nitrogen limitation, the genes associated with glycogen metabolism were constitutively expressed. These results are consistent with previous findings within the verrucomicrobial methanotrophs ( Khadem et al., 2010 ; Khadem et al., 2012a ; Mohammadi et al., 2016 ; Carere et al., 2017 ) and suggests the enzymes associated with glycogen metabolism may be allosterically regulated in response to high carbon (i.e., fructose 1,6-bisphosphate) and/or energy contents (i.e., ATP/AMP), as described in other bacterial species ( Wilson et al., 2010 ). Consistent with other verrucomicrobial methanotrophs ( Dunfield et al., 2007 ; Pol et al., 2007 ; Op den Camp et al., 2009 ), Methylacidiphilum sp. RTK17.1 also possesses three phylogenetically distinct pmoCAB operons. Based on observed ratios of non-synonymous versus synonymous substitution rates in pmoA orthologs, it has been proposed that the pMMOs encoded in Methylacidiphilum spp . serve functionally distinct roles ( Op den Camp et al., 2009 ). The observation that Methylacidiphilum sp. RTK17.1 transcriptionally regulates pMMO expression in response to oxygen availability therefore supports this hypothesis and is congruent with reports of differential expression in response to oxygen limitation ( Khadem et al., 2012a ) and during growth on methanol ( Erikstad et al., 2012 ). Finally, it is noteworthy to include that the transcriptional upregulation of the Group 1d [NiFe] hydrogenase occurred in response to O 2 limitation; whereas nitrogenase upregulation was induced by nitrogen availability. The transcriptional decoupling of these two enzymes is further evidence that the physiological role of H 2 oxidation in methanotrophs ( Mohammadi et al., 2016 ; Carere et al., 2017 ) is distinct from recycling H 2 produced during the nitrogen fixation reaction ( Bont, 1976 ; Dixon, 1976 ; Chen and Yoch, 1987 ). Collectively, these findings indicate while H 2 oxidation is sufficient to partially offset the energetic costs associated with N 2 fixation, the regulation of this enzyme is transcriptionally uncoupled from nitrogen availability."
} | 4,874 |
29402898 | PMC5799258 | pmc | 1,175 | {
"abstract": "Coral reefs are significant ecosystems. The ecological success of coral reefs relies on not only coral-algal symbiosis but also coral-microbial partnership. However, microbiome assemblages in the South China Sea corals remain largely unexplored. Here, we compared the microbiome assemblages of reef-building corals Galaxea ( G. fascicularis ) and Montipora ( M. venosa , M. peltiformis , M. monasteriata ) collected from five different locations in the South China Sea using massively-parallel sequencing of 16S rRNA gene and multivariate analysis. The results indicated that microbiome assemblages for each coral species were unique regardless of location and were different from the corresponding seawater. Host type appeared to drive the coral microbiome assemblages rather than location and seawater. Network analysis was employed to explore coral microbiome co-occurrence patterns, which revealed 61 and 80 co-occurring microbial species assembling the Galaxea and Montipora microbiomes, respectively. Most of these co-occurring microbial species were commonly found in corals and were inferred to play potential roles in host nutrient metabolism; carbon, nitrogen, sulfur cycles; host detoxification; and climate change. These findings suggest that the co-occurring microbial species explored might be essential to maintain the critical coral-microbial partnership. The present study provides new insights into coral microbiome assemblages in the South China Sea.",
"conclusion": "Conclusions In the present study, we examined the coral microbiome assemblages associated with Galaxea and Montipora collected from five locations in three biogeographic regions of the South China Sea. The highly dominant Proteobacteria and less abundant Bacteroidetes , Cyanobacteria , Firmicutes , and Actinobacteria are the main phyla in structuring the coral microbiomes. OTU-based multivariate statistical analysis demonstrated that each host species from each location was significantly different from any other host species and the corresponding seawater in the coral microbiome assemblages. Host type appeared to drive the coral microbiome assemblages rather than location and seawater. The network analysis explored a set of co-occurring microbial species assembling the coral microbiomes, which might play potential roles in host nutrient metabolism; carbon, nitrogen, sulfur cycles; host detoxification; and climate change. The findings of this study form a baseline for assessing coral microbiome development in the South China Sea, and for microbiome assemblage comparison across regions. The present study extends our knowledge of coral microbiome assemblages in the South China Sea and facilitates our understanding of coral-microbial partnership.",
"introduction": "Introduction Coral reef ecosystems are considered as the tropical rainforests of the sea, nurturing the highest biodiversity of marine life and providing vital ecosystem goods and services 1 , 2 . However, coral reefs around the world have suffered from declines and extinction risks largely due to bleaching events and emerging/reemerging diseases induced by climate change and anthropogenic disturbances 3 , 4 . The ecological success of coral reefs relies on coral-algal symbiosis, and recent studies using 16S rRNA gene amplicon pyrosequencing have revealed highly diverse and abundant microbes in individual coral colonies 5 – 9 . It is believed that some of these microbes can form partnerships with coral hosts and help them with possible access to those unavailable nutrients and metabolic pathways 10 . Compared with the well-understood coral-algal symbiosis, current understanding of coral-microbial partnership is rather limited. Although members of Endozoicomonas 6 and Prosthecochloris 11 have been shown to form potential symbioses with corals, no obligate coral-microbial symbiosis has been addressed to date. A recent advance in calcification for coral symbiotic algae demonstrated that free-living Symbiodinium in culture could form algal-microbial partnership, which facilitated the Symbiodinium calcification 12 . Taken these evidences, potentially complex coral-algal-microbial interactions in holobionts might facilitate the ecological success of coral reefs. The coral mucus, tissue, and skeleton all contain large populations of associated microbes from three domains of life, including Eukarya , Archaea , and Bacteria , as well as many viruses 13 , 14 . The term coral microbiome is employed herein to collectively refer to the bacteria and archaea in coral holobionts. The characterization of coral core microbiomes reveals that two most abundant bacterial phylotypes affiliating to the genera Propionibacterium and Ralstonia are co-localized specifically with the host’s endosymbiotic algae which are likely to facilitate the success of coral-algal symbiosis 10 . Coral microbiomes are highly complex and dynamic, usually changing with environmental conditions, host types, and tempo-spatial gradients 5 , 15 . However, we assumed that there would be co-occurring microbial species assembling coral microbiomes which might be fundamental for coral holobionts to maintain the critical coral-microbial partnership. To explore these potential microbial species, co-occurrence network analysis was conducted in this study. Network-based approaches have been widely applied in microbial ecology studies 16 such as soil 17 , coral 18 , human gut 19 , marine sediment 20 , stream biofilm 21 , activated sludge 22 , marine biofilm 23 , and other natural and man-made environments. In recent years, there has been a growing interest in coral microbiome studies 24 – 28 , but very limited knowledge has been gained in understanding the microbiome assemblages for the South China Sea corals. We do not yet clearly understand how coral microbiomes are assembled in the South China Sea and potential co-occurring microbial species making up the coral microbiomes. The aims of this study were (i) to compare the microbiome assemblages using congeneric corals collected from different locations of the South China Sea, and (ii) to explore potential co-occurring microbial species through co-occurrence network analysis. We believe the present study facilitates our understanding of coral-microbial partnership in the South China Sea and serves as a useful reference for comparison with other regional coral microbiome assemblages.",
"discussion": "Discussion In recent years, the concept of “core microbiome” has been introduced in coral microbial ecology studies 10 , 26 , 37 , while the counterpart “stable microbiome” has also been used in related studies 26 , 37 , 38 . According to the statistical methods used for data analysis, coral-associated microbes can be assigned into “core/stable microbiome” and “transient/sporadic microbiome” 37 . The “core microbiome” is described as the stable and consistent components across complex microbiome assemblages from similar habitats which can be determined with one of the five described variations, that is, a core based on shared presence, shared abundance, shared composition, phylogenetic information, or interaction 37 , 39 . Co-occurring microbial species explored here through co-occurrence network analysis are exactly equal to the “core microbiome” based on interaction, as illustrated in related studies 37 , 39 . In the present study, we collected coral samples with a large set of biological replicates from the five locations in the South China Sea, allowing us to compare the complex microbiome assemblages comprehensively and to explore co-occurrence patterns through network analysis. We further discussed the effects of host type and location in coral microbiome assemblages (the first section below) and potential roles of co-occurring microbial species explored including those functional microbes (the second section below) and those mostly found microbes (the third section below). Effects of host type and location in coral microbiome assemblages Corals harbor highly diverse microbes that might be assembled similarly or dissimilarly among host types, locations and other potential determinants. The study on Caribbean corals shows that different hosts harbor distinct host-specific microbes and that specificity varies by host type and location 15 . The study on Red Sea corals reveals that the microbiome assemblages vary largely with location and are shaped by host type 5 . Certain microbes are reported to form species-specific associations with corals 40 , suggesting the roles of host played in driving coral microbiome assemblages. The study on corals from the Great Barrier Reef shows that certain microbes are associated with corals specifically and that microbiome assemblages differ with the location but not the host type 41 . These findings support the present study that host type rather than location drives the coral microbiome assemblages, as shown in Fig. 3 . There is a limitation that three species from the genus Montipora were collected for the present study because the shared species were not found in the selected locations except for Galaxea fascicularis . The study using three species of the genus Acropora found that they harbored conserved microbes and suggested that closely related corals of the same genus were assembled with similar microbes 41 . So it can be assumed that if one of the three studied Montipora species were shared and used for the present study, the general distribution pattern of microbiome assemblages for this species from different locations might be less likely affected in the NMDS ordination. The distribution pattern of the three Montipora species shown in Fig. 3 further supports that host type rather than location drives the coral microbiome assemblages because LHT-Mo and SB-Mo from the same host species Montipora monasteriata were clustered closely and were relatively distinct from LI-Mo and CB-Mo (from Montipora venosa and Montipora peltiformis respectively). Potential roles of certain functional co-occurring microbial species To achieve a better understanding of co-occurring microbial species, potential ecological roles and associations with coral hosts were further explored. As shown in Fig. 6 , the listed co-occurring microbial species might be involved in several important biological and ecological processes: nutrient metabolism; carbon, nitrogen, and sulfur cycles; host detoxification; and climate change. Potential nitrogen-fixing bacteria such as Brevundimonas spp. and Phyllobacterium spp. might fix atmospheric nitrogen gas to ammonia. Potential ammonia-oxidizing archaea like Nitrosopumilus spp. might oxidize the holobiont ammonia to nitrite, and possible nitrite-oxidizing bacteria like Nitrospira sp. might oxidize the holobiont nitrite to nitrate. Finally, potential denitrifying bacteria such as Methylophilus sp. and Nitratireductor sp. might convert the holobiont nitrate to nitrogen gas. These co-occurring microbial species might thus be involved in the complete nitrogen cycle and changes in their population and activity might affect certain process of the nitrogen metabolism in coral holobionts 42 . It is suggested that they might serve as nitrogen regulators to keep the balance of holobiont nitrogen through providing sufficient bioavailable nitrogen and removing unneeded nitrogen. Carbon dioxide might be fixed into carbohydrates by Symbiodinium and Cyanobacteria through photosynthesis and by the co-occurring microbial species explored such as Nitrosopumilus spp., Thioalkalivibrio sp., and Thioprofundum spp. through chemosynthesis. Like Symbiodinium , the synthesized carbohydrates by the co-occurring microbial species might also serve as the host nutrients or food sources. The study has demonstrated that free-living Symbiodinium can calcify with the aid of microbial partners 12 , implying that potential co-occurring microbial species explored might also facilitate coral calcification directly or indirectly. Hydrogen sulfide is toxic to a wide range of eukaryotic organisms, including marine invertebrates like coral and sponge, by inhibiting cytochrome c oxidase and a type of catalase 43 , 44 . In addition, hydrogen sulfide generated by sulfate-reducing bacteria might lead to the initiation of coral black band disease, which has been consistently found in lab induction and field observation 45 , 46 . Co-occurring microbial species Thioalkalivibrio sp. and Thioprofundum spp., serving as potential sulfur-oxidizing bacteria, might oxidize holobiont accumulated hydrogen sulfide to sulfate, thus contributing to coral health through detoxification of reduced sulfur compounds. Both DMSP (dimethylsulfoniopropionate) and DMS (dimethylsulfide) are important compounds in the global sulfur cycle as they are closely related to cloud formation and global climate change 47 . Both reef-building corals and free-living Symbiodinium have been documented to be high producers of DMSP 48 , 49 . The generated DMSP might be degraded into climate-active gas DMS via the bacterial cleavage pathway by co-occurring microbial species explored such as Hoeflea sp., Loktanella sp., Phaeobacter sp., Roseovarius sp., Shewanella sp., and Vibrio sp. 47 . In summary, co-occurring microbial species explored might play potential roles in host nutrient metabolism; carbon, nitrogen, sulfur cycles; host detoxification; and climate change and might be essential to maintain the critical coral-microbial partnership. However, the real functions need to be tested in future studies. For example, microbial genome recovery using culture-independent methods such as genome binning and single-cell genomics is promising to validate the specific functions on a genome scale. Figure 6 A schematic representation illustrating potential roles of certain co-occurring microbial species in corals. Potential roles might be played in host nutrient metabolism; carbon, nitrogen, sulfur cycles; host detoxification; and climate change. Those co-occurring microbial species shown in red were mapped onto the well-known ecological processes. Potential roles of the mostly found co-occurring microbial species Acinetobacter spp. and Endozoicomonas spp Among the 114 co-occurring microbial species assembling the Galaxea and Montipora microbiomes, 16 Acinetobacter spp. and 10 Endozoicomonas spp. were found, of which 6 Acinetobacter spp. and 3 Endozoicomonas spp. were shared between the two corals. This finding corroborates the high occurrence of Acinetobacter spp. and Endozoicomonas spp. in coral microbiomes found in different coral species and regions (Table S3 ). However, the roles of Endozoicomonas spp. and Acinetobacter spp. are poorly understood to date. One study demonstrated that Acinetobacter spp. were dominant in the microbiomes of bleached corals but were not detected in healthy corals 50 . Acinetobacter spp. are not likely to serve as bleaching causing agents because they can also be detected in high abundance in healthy corals, as observed in the present and the other studies 7 , 8 , 15 , 34 . It is important to note, however, that Acinetobacter baumannii is a highly troublesome human pathogen worldwide due to its resistance to antibiotics 51 . This implies that potential Acinetobacter sp. might be a threat to coral hosts. An early study demonstrated that Acinetobacter guillouiae strain 20B is a DMS oxidizer 52 , suggesting that Acinetobacter spp. might participate in holobiont DMS metabolism. However, more and more recent studies have demonstrated that Acinetobacter spp. have the denitrification function, for example, Acinetobacter baumannii H1 53 , Acinetobacter johnonii DBP-3 54 , Acinetobacter sp. HA2 55 , Acinetobacter sp. YZS-X1–1 56 , and Acinetobacter sp. SZ28 57 . Besides the present study, Endozoicomonas spp. have been reported to be dominant in the microbiome assemblages of multiple coral species such as Porites astreoides 15 , Stylophora pistillata 6 , Acropora millepora 31 , Seriatopora hystrix 32 , and Coelastrea aspera 58 . Several Endozoicomonas species have been isolated from different marine invertebrates including a sponge 59 , a sea anemone 60 , a sea slug 61 , a comb pen shell 62 , and an octocoral 63 , and a stony coral 64 . Endozoicomonas spp. might play certain roles in their hosts because they are associated with a broad range of marine invertebrates. The roles of Endozoicomonas spp. in corals have been inferred in several studies 31 , 32 , 65 . Here PICRUSt prediction using 16S sequences assigned to Endozoicomonas spp. revealed an unreported function that they have a complete denitrification pathway including genes napAB , nirK , norBC , and nosZ responsible for converting nitrate to nitrite, nitrite to nitric oxide, nitric oxide to nitrous oxide, and nitrous oxide to nitrogen, respectively. However, further experiments are needed to elucidate their real roles in corals. In addition, Endozoicomonas spp. might be developed as coral health indicators because they are distributed extensively in corals and other marine invertebrates and no negative effects on corals have been reported to date."
} | 4,306 |
27725793 | PMC5048046 | pmc | 1,176 | {
"abstract": "Inhibition by ammonium at concentrations above 1000 mgN/L is known to harm the methanogenesis phase of anaerobic digestion. We anaerobically digested swine waste and achieved steady state COD-removal efficiency of around 52% with no fatty-acid or H 2 accumulation. As the anaerobic microbial community adapted to the gradual increase of total ammonia-N (NH 3 -N) from 890 ± 295 to 2040 ± 30 mg/L, the Bacterial and Archaeal communities became less diverse. Phylotypes most closely related to hydrogenotrophic Methanoculleus (36.4%) and Methanobrevibacter (11.6%), along with acetoclastic Methanosaeta (29.3%), became the most abundant Archaeal sequences during acclimation. This was accompanied by a sharp increase in the relative abundances of phylotypes most closely related to acetogens and fatty-acid producers ( Clostridium , Coprococcus , and Sphaerochaeta ) and syntrophic fatty-acid Bacteria ( Syntrophomonas , Clostridium , Clostridiaceae species, and Cloacamonaceae species) that have metabolic capabilities for butyrate and propionate fermentation, as well as for reverse acetogenesis. Our results provide evidence countering a prevailing theory that acetoclastic methanogens are selectively inhibited when the total ammonia-N concentration is greater than ~1000 mgN/L. Instead, acetoclastic and hydrogenotrophic methanogens coexisted in the presence of total ammonia-N of ~2000 mgN/L by establishing syntrophic relationships with fatty-acid fermenters, as well as homoacetogens able to carry out forward and reverse acetogenesis.",
"conclusion": "4. Conclusions Successful operation of an anaerobic reactor treating swine manure proved that Bacterial and Archaeal communities could acclimate to a steady increase in total NH 3 -N concentration up to 2040 ± 30 mg NH 3 -N L −1 . Both communities became less diverse over time. NH 3 -N tolerant phylotypes that were enriched include (1) acetoclastic methanogens ( Methanosaeta ); (2) Clostridia known to do forward and reverse acetogenesis ( Clostridium and Clostridiaceae spp.); (3) fatty-acid producers ( Coprococcus and Sphaerochaeta ); (4) hydrogenotrophic methanogens ( Methanoculleus , Methanobrevibacter , and Methanogenium ); and (5) syntrophic fatty-acid fermenters ( Syntrophomonas , Clostridium , Clostridiaceae spp., and possibly Cloacamonaceae species). Our results suggest that the gradual increase in the NH 3 -N concentration led to a microbial community acclimated to the high total NH 3 concentrations associated with anaerobic digestion of animal wastes. As summarized in Figure 7 , acetoclastic and hydrogenotrophic methanogens could coexist in the presence of NH 3 -N concentrations ~2000 mg L −1 by establishing syntrophic relationships with propionate and butyrate-fermenters, as well as homoacetogens able to carry out forward and reverse acetogenesis.",
"introduction": "1. Introduction Animal wastes contribute more than half of the biomass-based wastes generated in the United States [ 1 , 2 ]. The organic carbon in animal wastes could be a major source of renewable energy if it were captured as methane gas. Many animal wastes, including swine waste, also are rich in organic nitrogen (N) due to the high protein content in the animals' diet. Anaerobic hydrolysis and fermentation convert the organic N into ammonia-N (NH 3 -N). While typical NH 3 -N (i.e., unionized NH 3 and NH 4 \n + ) concentrations in anaerobic digesters treating domestic wastewater sludge are 650–1100 mg L −1 [ 3 ], concentrations in swine manure are as high as 8000 mg L −1 [ 4 – 6 ]. A challenge arises for treating these wastes (and ultimately capturing the organic carbon as energy source) as NH 3 -N above 1000 mg L −1 is toxic to many groups of microorganisms [ 7 , 8 ], including methanogenic Archaea [ 9 , 10 ]. In anaerobic systems without inhibition by NH 3 -N, organic acids produced from acidogenesis are fermented to acetate and H 2 , and the typical distribution of the electron flow to methane is 67% through acetate and 33% through H 2 [ 11 , 12 ]. Correspondingly, acetoclastic methanogens usually predominate in anaerobic digesters with <1000 mg NH 3 -N L −1 [ 13 , 14 ]. In contrast, the vast majority of studies on methanogenesis from swine waste report that the dominant methanogenesis pathway switches from acetoclastic to hydrogenotrophic. Several studies [ 15 – 19 ] reported that the methane production in anaerobic reactors treating wastes with high NH 3 -N occurs mainly via hydrogenotrophic methanogenesis, since acetoclastic methanogens are inhibited and washed out. The loss of acetoclastic methanogens in the bioreactors raises questions about the fate of acetate generated by fermentation. It was previously postulated that the loss of acetoclastic methanogens was compensated by syntrophic “acetate oxidation” to CO 2 and H 2 (more accurately termed reverse acetogenesis) coupled with hydrogenotrophic methanogenesis [ 20 ]. Specifically, acetate generated by fermentation is converted into H 2 and CO 2 by reverse acetogenesis, and the H 2 and CO 2 are utilized by hydrogenotrophic methanogenesis. The reactions involved in the syntrophy of reverse acetogenesis coupled with H 2 conversion into methane with their corresponding Δ G °′ are illustrated in Table 1 , equations (1a) and (2), respectively. Reverse acetogenesis might not be the only mechanism that allows acetate conversion into methane when treating high-ammonium wastes. Methanogenesis is possible if some acetoclastic methanogens are able to adapt to high-ammonium concentrations and avoid being washed out. In fact, Westerholm et al. detected acetoclastic methanogens in methanogenic reactors operating at increasing NH 3 -N concentrations (from 800 to 6900 mg L −1 ) [ 21 ]. Thus, an acclimation period may be crucial for developing a microbial community that has acetoclastic methanogens capable of tolerating high NH 3 -N. Another key aspect of microbial community acclimation is the scavenging of H 2 produced in fermentation and reverse acetogenesis. It is well known that fermentation of propionate (equation (1b) in Table 1 ) and butyrate (equation (1c)) is coupled with hydrogenotrophic methanogenesis [ 22 – 24 ]. The highly endergonic nature of these fatty-acid fermentation instances means that they can occur only at very low partial pressures of H 2 (<10 −4 atm), which requires a tight syntrophic partnership with a H 2 scavenger, such as a hydrogenotrophic methanogen, to maintain a negative Δ G °′ for fermentation to proceed [ 25 ]. The microbial ecology of anaerobic reactors treating high NH 3 -N wastes has received limited attention [ 15 , 26 – 28 ], and the possible syntrophies among different Archaea, fermenting Bacteria, and homoacetogens are yet poorly understood. High-throughput sequencing in combination with statistical analysis can illuminate microbial dynamics with high NH 3 -N concentrations by identifying key Archaea and Bacteria involved in syntrophic fatty-acid conversion into methane. In this study, we used high-throughput sequencing, parametric correlation, and qPCR (quantitative polymerase chain reaction) to analyze shifts in the Archaeal and Bacterial communities during the startup phase (first 105 days) of an anaerobic digester successfully treating swine manure to generate methane and without significant accumulation of acetate or H 2 . Contrary to previous explanations of the effects of NH 3 -N concentration higher than 1000 mg L −1 , acetoclastic methanogens played a major role in methane production. Our results point to syntrophies that involved acetoclastic methanogens, hydrogenotrophic methanogens, homoacetogens, and other syntrophic acid-fermenting Bacteria developed during the startup of a methanogenic bioreactor able to function well with NH 3 -N greater than ~2000 mg NH 3 -N L −1 .",
"discussion": "3. Results and Discussion 3.1. COD Was Converted into Methane during Bioreactor Operation despite NH 3 -N > 2000 mg/L We monitored total and soluble COD, NH 3 -N and total N, and methane and biogas production rates at regular intervals during the startup phase of the methanogenic reactor treating swine waste. The results are summarized in Figures 1 and 2 and S1, in Supplementary Material available online at http://dx.doi.org/10.1155/2016/4089684 ; Table S1 documents good COD mass-balance closure at four sampling times. Based on COD removed as CH 4 , the performance of the reactor approached a pseudo-steady state in cycle 1 of semicontinuous operation, with approximately 52% conversion. CH 4 was 70% ± 16% of the biogas, and H 2 concentrations were below detectable levels (<0.5% v/v) throughout the experiment. During semicontinuous operation, effluent soluble COD (which is composed of relatively small, biodegradable molecules [ 39 ]) represented only 4.5% ± 0.3% of the influent TCOD or ≤2.6 ± 0.6 g COD L −1 . This low SCOD concentration implies that short-chain fatty acids (including propionate, butyrate, and acetate), which mainly comprise the SCOD, did not accumulate because they were consumed by microbial metabolism leading to methane production and biomass synthesis. \n Figure 2(a) shows NH 3 -N increased from 890 ± 295 mg NH 3 -N L −1 during batch operation to 2040 ± 30 mg NH 3 -N L −1 during semicontinuous operation. The concentration of soluble total N (the sum of NH 3 -N and organic N) paralleled that of NH 3 -N and was about 50% higher than total NH 3 -N. This increasing release of organic N and NH 3 -N indicates that hydrolysis and fermentation of the protein fraction of the animal wastes increased after startup and stabilized in cycle 2. Figure 2(b) shows that methanogenesis continued to increase into cycle 3, even though hydrolysis and fermentation of protein stabilized. Higher methane generation was possible in cycle 3 because the input TCOD increased with the batch of swine waste used in that cycle; the input N did not increase in parallel with input TCOD because the feed collected swine manure was not uniform throughout the experimental period. 3.2. The Diversity of Archaea and Bacteria Decreased with Increasing NH 3 -N Concentration \n Table 2 summarizes the coefficients of the two metrics used to analyze diversity within the samples. PD-whole-tree, which is based on the phylogenetic tree, uses the branch lengths as a measure of diversity; the observed-species metric counts all unique OTUs in the sample [ 40 ]. Both metrics had similar decreasing trends for Archaea and Bacteria over time as the NH 3 -N concentration rose from 684 mg NH 3 -N L −1 at the startup of the reactor to 890 ± 295 mg NH 3 -N L −1 in batch operation and to 2040 ± 30 mg NH 3 -N L −1 for continuous operation. These significant decreases suggest selective enrichment of microorganisms tolerant to NH 3 -N. 3.3. Hydrogenotrophic and Acetoclastic Methanogens Were Abundant at ~2000 mg NH 3 -N L −1 \n We analyzed the Archaeal microbial community during reactor startup using high-throughput sequencing in order to evaluate how well acetoclastic methanogens (grouped under Methanosarcinales) or hydrogenotrophic methanogens (grouped under Methanobacteriales, Methanomicrobiales, Methanococcales, and the E2 order) were tolerant to NH 3 -N concentrations of ~2000 mg L −1 . Figure 3 compares the Archaeal communities in the inoculum with the communities after batch and semicontinuous operation. Phylotypes within the phylum Euryarchaeota (i.e., Methanomicrobiales, Methanobacteriales, E2 group, and Methanosarcinales) and those of the order pGrfc26 (within Crenarchaeota) [ 41 ] increased during the semicontinuous operation at 2040 ± 30 mg NH 3 -N L −1 , while the unidentified phylotypes decreased from 38.6 to 1.5%. These results confirm enrichment during the gradual acclimation to high and increasing total-ammonium concentrations. The relative abundance of hydrogenotrophic and acetoclastic methanogens increased over time. This agrees with increasing methane production, low effluent SCOD (including acetate in the measurement), and no detection of H 2 during semicontinuous operation. Figure 3(b) shows 8 different genera of hydrogenotrophic methanogens identified, including Methanoculleus , Methanogenium, and Methanobrevibacter . Although the genus Methanosaeta was the sole acetoclastic methanogen identified by high-throughput sequencing, its relative abundance increased from 14% in batch operation to 25% during semicontinuous operation. This increase in relative abundance could be due to an increase in the absolute abundance of Methanosaeta or a decrease in the abundance of other phenotypes. qPCR results (summarized in Figure S2) are consistent with Figure 3 and suggest that Methanomicrobiales were the most abundant methanogens followed by Methanosaetaceae and lastly by Methanobacteriales. It is likely that the first 35 d of batch operation, with <1100 mg N-NH 3 L −1 , provided an acclimation period for acetoclastic methanogens, and ~2000 mg NH 3 -N L −1 was not high enough to inhibit Methanosaeta , which were selected by the third cycle, relative to other Archaea. This corresponds to the finding by Schnürer and Nordberg that it took 3000 mg NH 3 -N L −1 to inhibit NH 3 -N-acclimated Methanosaeta spp. [ 18 ] and that this family was not detected in anaerobic digesters treating chicken wastes with above 3400 mg NH 3 -N L −1 [ 42 ]. However, a recent study reported that NH 3 -N-acclimated Methanosaetaceae spp. were the most abundant acetoclastic methanogens identified in laboratory-scale anaerobic digesters operated at increasing NH 3 -N concentrations of up to 4000 mg L −1 [ 43 ]. 3.4. Acetogens Played a Key Role in Methane Production at ~2000 mg NH 3 -N L −1 \n The most abundant Bacterial phylotypes at the family and genus levels are summarized in Figure 4 . At the genus level, phylotypes representative of producers of key short-chain fatty-acid (e.g., acetate, propionate, and butyrate) were enriched during exposure to high NH 3 -N. These Bacteria are summarized in panel (b). The most abundant phylotypes were (i) Coprococcus , a butyrate- and acetate-producer within the Lachnospiraceae family [ 44 ], (ii) Sphaerochaeta (within the Spirochaetaceae family), an acetate-, formate-, and ethanol-producer [ 45 ], (iii) Treponema , acetogenic microorganisms within the Spirochaetaceae family, and (iv) unidentified phylotypes in the Bacteroidales and Clostridiales orders (represented in purple and blue, resp.), known to harbor microorganisms capable of fermenting carbohydrates and proteins to short-chain fatty-acids [ 46 – 48 ]. To elaborate on the acetogens, we used qPCR to quantify the highly conserved formyl tetrahydrofolate synthetase (FTHFS) gene of homoacetogens [ 49 , 50 ] and reverse-acetogens (sometimes called “syntrophic acetate oxidizers”) [ 6 , 21 ]. Figure 5 shows a steady increase of the FTHFS gene; quantified FTHFS genes increased about two orders of magnitude from the beginning of the experiment to day 105 despite relative constant numbers of Bacterial 16S rRNA genes. This indicates that there was not only an increase in homoacetogenic Bacteria but also a relative increase in their proportion of the Bacterial community, revealing an enrichment of homoacetogens with increasing NH 3 -N concentrations. Low SCOD in the effluent, no detection of H 2 , and the high amount of acetoclastic and hydrogenotrophic methanogens mean that acetate and H 2 were efficiently scavenged to produce CH 4 . The presence of homoacetogens means that the sink for acetate during semicontinuous operation could have been either one of the methanogenic pathways or both. Homoacetogens could either have been doing forward acetogenesis, in which case acetoclastic methanogens scavenge acetate, or have been doing reverse acetogenesis, in which hydrogenotrophic methanogens feed on H 2 and CO 2 generated from acetate. Syntrophic acetate-utilizers were represented by phylotypes associated with the genus Clostridium within the Clostridiaceae family. Their relative abundance increased from 0.72% in the inoculum to 3.5 and 5.6% during batch and semicontinuous operation, respectively. Some strains (including C. ultunense [ 51 ] and strains similar to C. botulinum , C. sticklandii , and C. beijerinckii [ 52 ]) have been reported to perform reverse acetogenesis in methanogenic communities [ 51 , 52 ]. With high NH 3 -N concentration, these acetate-utilizers have been commonly found in syntrophy with hydrogenotrophic methanogens such as Methanoculleus [ 13 , 53 ], the most abundant methanogen identified in our reactor. Thus, it is possible that several Clostridium spp. in our reactor contributed to methane production through reverse acetogenesis coupled with hydrogenotrophic methanogenesis. 3.5. Syntrophic Fatty-Acid Fermenters Thrived at ~2000 mg NH 3 -N L −1 \n In addition to syntrophic acetate-utilizers among the Clostridia, propionate- and butyrate-fermenters that grow in syntrophy with hydrogenotrophic methanogens also were detected at relative abundances between 1.2 and 4.5% during semicontinuous operation. The detected phylotypes at the genus level were Syntrophomonas and W22 and W5. Syntrophomonas (within the Syntrophomonadaceae) ferment butyrate to acetate and H 2 in syntrophic association with hydrogenotrophic methanogens and sulfate-reducers [ 22 , 54 ]. Cloacamonas , a representative genus of Cloacamonaceae, obtains its energy from the fermentation of amino acids and can ferment propionate to acetate, H 2 , and CO 2 in syntrophy with H 2 and acetate consumers [ 55 , 56 ]. This syntroph has been named Candidatus Cloacamonas acidaminovorans , and although it has not been cultivated, its genome has been reconstructed by metagenomics [ 55 ]. Comparing the sequences (300 bp) associated with the W22 and W5 genera to available sequences (NCBI, BLAST) reveals that these genera share up to 96% similarity with Candidatus C. acidaminovorans . Therefore, it is possible that Cloacamonaceae and Syntrophomonadaceae contributed to methane production with 2040 ± 30 mg NH 3 -N L −1 by fermenting butyrate and propionate to H 2 , CO 2 , and acetate in syntrophy with hydrogenotrophic and acetoclastic methanogens. 3.6. Correlation Analysis Supports Syntrophies between Acetoclastic Methanogens with Acetogens and Hydrogenotrophic Methanogens with Syntrophic Fatty-Acid Fermenters In order to understand the effect of NH 3 -N and total N concentrations on methane production and the microbial community structure, we calculated Pearson's R coefficient among methane production rates, NH 3 -N and total N concentrations, and fermenters, syntrophs, and hydrogenotrophic and acetoclastic methanogens identified at four time points during operation of the anaerobic reactor treating swine waste. The results of the parametric correlation analysis are summarized in Figure 6 . Hydrogenotrophic methanogens ( Methanoculleus , Methanogenium , Methanobrevibacter, and an unidentified genus within Methanomicrobiales) and acetoclastic Methanosaeta were positively correlated with total N and NH 3 -N concentrations (in some cases, correlations were significant at the 0.05 level). These positive correlations suggest that hydrogenotrophic and acetoclastic methanogens thrived with increasing NH 3 -N concentration up to 2040 ± 30 mg NH 3 -N L −1 when the pH was 6.9–7.6 and the HRT was 35 d. Moreover, acetogens/fermenters Coprococcus , Sphaerochaeta , and an unidentified genus within Porphyromonadaceae were positively correlated with NH 3 -N and almost all methanogens and, consequently, correlation with methane production was positive. These positive correlations underscore the important role of acetogens for methane production at high NH 3 -N. Unidentified Clostridiaceae, which comprises several short-chain fatty-acid producers, and Clostridium had positive correlations with acetoclastic Methanosaeta , and this supports acetate generation by acetogens and homoacetogens. However, unidentified Clostridiaceae also showed a positive correlation with unidentified hydrogenotrophic Methanomicrobiales and with hydrogenotrophs Methanogenium and Methanoculleus. This supports that homoacetogens among the Clostridiaceae were possibly carrying out reverse acetogenesis. Thus, parametric analysis supports an important role for homoacetogens, but it cannot determine whether they were performing forward or reverse acetogenesis."
} | 5,110 |
23888127 | PMC3715687 | pmc | 1,178 | {
"abstract": "Following a strategy similar to that used in baker’s yeast (Herrgård et al. Nat Biotechnol 26:1155–1160, 2008 ). A consensus yeast metabolic network obtained from a community approach to systems biology (Herrgård et al. 2008 ; Dobson et al. BMC Syst Biol 4:145, 2010 ). Further developments towards a genome-scale metabolic model of yeast (Dobson et al. 2010 ; Heavner et al. BMC Syst Biol 6:55, 2012 ). Yeast 5—an expanded reconstruction of the Saccharomyces cerevisiae metabolic network (Heavner et al. 2012 ) and in Salmonella typhimurium (Thiele et al. BMC Syst Biol 5:8, 2011 ). A community effort towards a knowledge-base and mathematical model of the human pathogen Salmonella \n typhimurium LT2 (Thiele et al. 2011 ), a recent paper (Thiele et al. Nat Biotechnol 31:419–425, 2013 ). A community-driven global reconstruction of human metabolism (Thiele et al. 2013 ) described a much improved ‘community consensus’ reconstruction of the human metabolic network, called Recon 2, and the authors (that include the present ones) have made it freely available via a database at http://humanmetabolism.org/ and in SBML format at Biomodels ( http://identifiers.org/biomodels.db/MODEL1109130000 ). This short analysis summarises the main findings, and suggests some approaches that will be able to exploit the availability of this model to advantage.",
"conclusion": "Concluding remarks The availability of Recon 2 allows a great many computational analyses to be performed. We have purposely rehearsed these at a rather general level, since particular analyses, that might be relevant to particular diseases, for instance, are simply implementations of the more general approaches. One new approach that will depend on the existence of such a network as a necessary resource is personalised medicine (Hood and Flores 2012 ). There one will develop models of metabolism calibrated for each specific individual, in large part using metabolomics methods, to be used as bases for diagnostics and decisions on course of treatment. Recon 2 is a very significant step towards such a map, where such measurements have to be anchored for various types of modelling that will underpin personalised treatment decisions. The development of tissue- and condition-specific models has been demonstrated with Recon 2 and its predecessors (Jerby et al. 2010 ; Frezza et al. 2011 ; Wang et al. 2012 ). As subsequent iterations of Recon 2 develop, it is hoped that the scope of the knowledgebase, and the predictive power of derived models, will increase to keep pace with advancements in the community knowledge of human metabolism, many of which will be driven by the discipline of metabolomics."
} | 669 |
30417967 | null | s2 | 1,179 | {
"abstract": "Silk-elastin-like-protein polymers (SELPs) are genetically engineered recombinant protein sequences consisting of repeating units of silk-like and elastin-like blocks. By combining these entities, it is shown that both the characteristic strength of silk and the temperature-dependent responsiveness of elastin can be leveraged to create an enhanced stimuli-responsive material. It is hypothesized that SELP behavior can be influenced by varying the silk-to-elastin ratio. If the responsiveness of the material at different ratios is significantly different, this would allow for the design of materials with specific temperature-based swelling and mechanical properties. This study demonstrates that SELP fiber properties can be controlled via a temperature transition dependent on the ratio of silk-to-elastin in the material. SELP fibers are experimentally wet spun from polymers with different ratios of silk-to-elastin and conditioned in either a below or above transition temperature (T "
} | 248 |
39730616 | PMC11680959 | pmc | 1,180 | {
"abstract": "Reservoir computing is a machine learning framework that exploits nonlinear dynamics, exhibiting significant computational capabilities. One of the defining characteristics of reservoir computing is that only linear output, given by a linear combination of reservoir variables, is trained. Inspired by recent mathematical studies of generalized synchronization, we propose a novel reservoir computing framework with a generalized readout, including a nonlinear combination of reservoir variables. Learning prediction tasks can be formulated as an approximation problem of a target map that provides true prediction values. Analysis of the map suggests an interpretation that the linear readout corresponds to a linearization of the map, and further that the generalized readout corresponds to a higher-order approximation of the map. Numerical study shows that introducing a generalized readout, corresponding to the quadratic and cubic approximation of the map, leads to a significant improvement in accuracy and an unexpected enhancement in robustness in the short- and long-term prediction of Lorenz and Rössler chaos. Towards applications of physical reservoir computing, we particularly focus on how the generalized readout effectively exploits low-dimensional reservoir dynamics.",
"introduction": "Introduction Reservoir computing (RC) is a machine learning framework that exploits dynamical systems and has remarkable computational capabilities 1 – 3 . For example, RC using random networks, called echo state networks (ESNs), can efficiently predict chaotic time series 4 . Adding closed-loop makes an RC system autonomous and capable of replicating chaotic attractors, which are utilized to estimate Lyapunov exponents 5 . Furthermore, recent studies have shown that such ‘autonomous’ RC systems can reproduce true dynamical properties more accurately than those computed from limited training data and extrapolate true dynamical structures such as bifurcation outside the training data 6 – 8 . Another branch of research, physical RC, harnesses various physical dynamics and demonstrates high information processing capability 9 – 13 , 14 . Why does RC work so well with untrained random networks and physical systems? This is a central open problem in RC research, and more broadly in machine learning and neuroscience. Partial answers to this problem have been provided using dynamical systems theory 7 , 15 – 17 . In particular, Grigoryeva, Hart, and Ortega 17 rigorously proved the existence of a continuously differentiable synchronization map under certain conditions and explicitly showed what the RC learns when predicting chaotic dynamics. In other words, they provided a formal expression of a map, which is \\documentclass[12pt]{minimal}\n\t\t\t\t\\usepackage{amsmath}\n\t\t\t\t\\usepackage{wasysym} \n\t\t\t\t\\usepackage{amsfonts} \n\t\t\t\t\\usepackage{amssymb} \n\t\t\t\t\\usepackage{amsbsy}\n\t\t\t\t\\usepackage{mathrsfs}\n\t\t\t\t\\usepackage{upgreek}\n\t\t\t\t\\setlength{\\oddsidemargin}{-69pt}\n\t\t\t\t\\begin{document}$$\\textbf{h}$$\\end{document} , as explained later in Eq. ( 8 ), that RC approximates for prediction. Hara and Kokubu 7 uncovered a key mathematical structure for learning with RC, i.e. a smooth conjugacy between target and reservoir dynamics based on observations from the numerical study of the logistic map. Inspired by these seminal studies 7 , 17 , we propose a novel method of RC with a generalized readout. Based on generalized synchronization, the Taylor expansion of the map \\documentclass[12pt]{minimal}\n\t\t\t\t\\usepackage{amsmath}\n\t\t\t\t\\usepackage{wasysym} \n\t\t\t\t\\usepackage{amsfonts} \n\t\t\t\t\\usepackage{amssymb} \n\t\t\t\t\\usepackage{amsbsy}\n\t\t\t\t\\usepackage{mathrsfs}\n\t\t\t\t\\usepackage{upgreek}\n\t\t\t\t\\setlength{\\oddsidemargin}{-69pt}\n\t\t\t\t\\begin{document}$$\\textbf{h}$$\\end{document} , the Eq. ( 10 ), may give an interpretation of the conventional RC as a linearization of \\documentclass[12pt]{minimal}\n\t\t\t\t\\usepackage{amsmath}\n\t\t\t\t\\usepackage{wasysym} \n\t\t\t\t\\usepackage{amsfonts} \n\t\t\t\t\\usepackage{amssymb} \n\t\t\t\t\\usepackage{amsbsy}\n\t\t\t\t\\usepackage{mathrsfs}\n\t\t\t\t\\usepackage{upgreek}\n\t\t\t\t\\setlength{\\oddsidemargin}{-69pt}\n\t\t\t\t\\begin{document}$$\\textbf{h}$$\\end{document} . Moreover, it implies that the computational capabilities of RC with a generalized readout are superior to those of conventional RC. Remark that a specific type of nonlinear readout, such as \\documentclass[12pt]{minimal}\n\t\t\t\t\\usepackage{amsmath}\n\t\t\t\t\\usepackage{wasysym} \n\t\t\t\t\\usepackage{amsfonts} \n\t\t\t\t\\usepackage{amssymb} \n\t\t\t\t\\usepackage{amsbsy}\n\t\t\t\t\\usepackage{mathrsfs}\n\t\t\t\t\\usepackage{upgreek}\n\t\t\t\t\\setlength{\\oddsidemargin}{-69pt}\n\t\t\t\t\\begin{document}$$\\sum _{i} W_{i}(\\textbf{r}_{t})^{2}_{i}$$\\end{document} in the notation introduced later, has already been used in previous research 5 , 18 , 19 . We emphasize that our theoretical framework comprehensively explains the reason why these nonlinear readouts are so effective, and, furthermore, presents a new direction of research, including cubic readouts, bridging rigorous mathematics 7 , 17 with practical applications. Indeed, numerical studies on Lorenz and Rössler chaos prediction as a benchmark problem strongly support this; i.e., for both short- and long-term predictions, we reveal the significant computational capabilities of RC with generalized readout compared to conventional RC. Moreover, for long-term prediction, the autonomous RC system with generalized readout acquires notable robustness, in contrast to the lack of robustness of conventional RC.",
"discussion": "Conclusion and discussion Inspired by the seminal works on the mathematical analysis of RC 7 , 17 , we have proposed a novel method of RC with generalized readout with a theoretical guarantee of its high computational capabilities based on generalized synchronization. Numerical studies on the Lorenz and Rössler chaos have uncovered significant short- (Figs. 2 and 3 ) and long-term prediction and reconstruction abilities with improved robustness (Figs. 4 , 5 , 6 and 7 ) of the \\documentclass[12pt]{minimal}\n\t\t\t\t\\usepackage{amsmath}\n\t\t\t\t\\usepackage{wasysym} \n\t\t\t\t\\usepackage{amsfonts} \n\t\t\t\t\\usepackage{amssymb} \n\t\t\t\t\\usepackage{amsbsy}\n\t\t\t\t\\usepackage{mathrsfs}\n\t\t\t\t\\usepackage{upgreek}\n\t\t\t\t\\setlength{\\oddsidemargin}{-69pt}\n\t\t\t\t\\begin{document}$${{\\mathcal {Q}}}$$\\end{document} -ESN. The MCE \\documentclass[12pt]{minimal}\n\t\t\t\t\\usepackage{amsmath}\n\t\t\t\t\\usepackage{wasysym} \n\t\t\t\t\\usepackage{amsfonts} \n\t\t\t\t\\usepackage{amssymb} \n\t\t\t\t\\usepackage{amsbsy}\n\t\t\t\t\\usepackage{mathrsfs}\n\t\t\t\t\\usepackage{upgreek}\n\t\t\t\t\\setlength{\\oddsidemargin}{-69pt}\n\t\t\t\t\\begin{document}$$\\bar{{\\mathcal {E}}}^c$$\\end{document} and KLD \\documentclass[12pt]{minimal}\n\t\t\t\t\\usepackage{amsmath}\n\t\t\t\t\\usepackage{wasysym} \n\t\t\t\t\\usepackage{amsfonts} \n\t\t\t\t\\usepackage{amssymb} \n\t\t\t\t\\usepackage{amsbsy}\n\t\t\t\t\\usepackage{mathrsfs}\n\t\t\t\t\\usepackage{upgreek}\n\t\t\t\t\\setlength{\\oddsidemargin}{-69pt}\n\t\t\t\t\\begin{document}$$D_{\\text {KL}}$$\\end{document} have quantified these properties complementarily, i.e. from the notions of orbit and distribution. By including the higher-order approximation, we have revealed “hierarchical” improvement in reconstruction ability and robustness; i.e. the \\documentclass[12pt]{minimal}\n\t\t\t\t\\usepackage{amsmath}\n\t\t\t\t\\usepackage{wasysym} \n\t\t\t\t\\usepackage{amsfonts} \n\t\t\t\t\\usepackage{amssymb} \n\t\t\t\t\\usepackage{amsbsy}\n\t\t\t\t\\usepackage{mathrsfs}\n\t\t\t\t\\usepackage{upgreek}\n\t\t\t\t\\setlength{\\oddsidemargin}{-69pt}\n\t\t\t\t\\begin{document}$${{\\mathcal {C}}}$$\\end{document} -ESN is superior to the \\documentclass[12pt]{minimal}\n\t\t\t\t\\usepackage{amsmath}\n\t\t\t\t\\usepackage{wasysym} \n\t\t\t\t\\usepackage{amsfonts} \n\t\t\t\t\\usepackage{amssymb} \n\t\t\t\t\\usepackage{amsbsy}\n\t\t\t\t\\usepackage{mathrsfs}\n\t\t\t\t\\usepackage{upgreek}\n\t\t\t\t\\setlength{\\oddsidemargin}{-69pt}\n\t\t\t\t\\begin{document}$${{\\mathcal {Q}}}$$\\end{document} -ESN, which is superior to the \\documentclass[12pt]{minimal}\n\t\t\t\t\\usepackage{amsmath}\n\t\t\t\t\\usepackage{wasysym} \n\t\t\t\t\\usepackage{amsfonts} \n\t\t\t\t\\usepackage{amssymb} \n\t\t\t\t\\usepackage{amsbsy}\n\t\t\t\t\\usepackage{mathrsfs}\n\t\t\t\t\\usepackage{upgreek}\n\t\t\t\t\\setlength{\\oddsidemargin}{-69pt}\n\t\t\t\t\\begin{document}$${{\\mathcal {L}}}$$\\end{document} -ESN (Fig. 7 ). As the future extensions based on the present work, we discuss the following three directions: mathematical analysis, machine learning, and physical implementation. From the mathematical analysis of RC 7 , 17 , it may be natural that introducing the generalized readout improves prediction ability. However, we unexpectedly observed an improvement in the robustness of the reconstruction ability. Further analysis of the reservoir dynamics is crucial; unveiling fundamental properties such as the topological conjugacy and the mechanism behind the enhanced robustness will have major implications for several fields, including machine learning, where stabilizing the dynamics of neural networks by adding noise and normalization is one of the critical issues 20 . One of the key applications of the generalized readout is the physical RC; in many physical systems, such as photonic integrated circuit 10 , only small physical degrees of freedom are available 9 – 13 . The hard challenge is to find a way to exploit these low-dimensional dynamics for computation. We emphasize that our generalized readout paves the way, and this is why we focus on small networks in the numerical study. For future work along the line of the research 21 , 22 , combining linear physical systems with the generalized readout may be effective. The apparent drawback of using the generalized readout is the large number of parameters to be trained, still within the linear learning framework. Therefore, based on the hierarchical improvement in accuracy with increasing parameters (Fig. 7 ), the balance between accuracy and learning cost should be determined for each application. For the large number of parameters to be trained, transfer learning 23 , 24 may be efficient. Once linear regression is used, the trained parameters, i.e. the generalized readout weights, can be reused with a minor correction for similar tasks, e.g. predicting chaotic dynamics that are structurally stable. The autonomous RC with generalized readout achieves accurate predictions, e.g. longer than 8 Lyapunov times (Fig. 4 ); however, such a prediction eventually fails due to the orbital instability. Toward practical predictions of, for instance, fluid turbulence 25 , combining the autonomous RC with generalized readout and data assimilation may be essential in future work. Also, we have assumed that observational data of all state variables are available for training. It is important to investigate the effectiveness of the generalized readout in the case where only the partial observation data are available, as studied in 26 . In practice, the prediction of high-dimensional dynamical systems, such as fluid turbulence 23 , 27 , is crucial. We have found that the RC with generalized readout is effective for some high-dimensional chaos, which will be reported elsewhere. The concept of generalized readout does not require the RC framework, but rather, may be essential in a more general machine learning context, e.g. training recurrent neural networks. Although the main claim of this paper is to propose the mathematical framework and the generalized readout method, a systematic comparison across a variety of neural networks, e.g., with deep architectures 28 , in a more general task may be valuable and is left for future study. Studies of neural connections similar to \\documentclass[12pt]{minimal}\n\t\t\t\t\\usepackage{amsmath}\n\t\t\t\t\\usepackage{wasysym} \n\t\t\t\t\\usepackage{amsfonts} \n\t\t\t\t\\usepackage{amssymb} \n\t\t\t\t\\usepackage{amsbsy}\n\t\t\t\t\\usepackage{mathrsfs}\n\t\t\t\t\\usepackage{upgreek}\n\t\t\t\t\\setlength{\\oddsidemargin}{-69pt}\n\t\t\t\t\\begin{document}$${{\\mathcal {Q}}}$$\\end{document} - and \\documentclass[12pt]{minimal}\n\t\t\t\t\\usepackage{amsmath}\n\t\t\t\t\\usepackage{wasysym} \n\t\t\t\t\\usepackage{amsfonts} \n\t\t\t\t\\usepackage{amssymb} \n\t\t\t\t\\usepackage{amsbsy}\n\t\t\t\t\\usepackage{mathrsfs}\n\t\t\t\t\\usepackage{upgreek}\n\t\t\t\t\\setlength{\\oddsidemargin}{-69pt}\n\t\t\t\t\\begin{document}$${{\\mathcal {C}}}$$\\end{document} -ESN may also be interesting in the context of a learning mechanism in biological brains. Whatever the direction, the concepts from dynamical system theory used in the above discussion, such as synchronization, orbital instability, and conjugacy, will shed light on a guiding principle for future studies."
} | 3,099 |
28045456 | PMC5322298 | pmc | 1,182 | {
"abstract": "Geobacter sulfurreducens uses at least two different pathways to transport electrons out of the inner membrane quinone pool before reducing acceptors beyond the outer membrane. When growing on electrodes poised at oxidizing potentials, the CbcL-dependent pathway operates at or below redox potentials of –0.10 V vs the standard hydrogen electrode, whereas the ImcH-dependent pathway operates only above this value. Here, we provide evidence that G. sulfurreducens also requires different electron transfer proteins for reduction of a wide range of Fe(III)- and Mn(IV)-(oxyhydr)oxides, and must transition from a high- to low-potential pathway during reduction of commonly studied soluble and insoluble metal electron acceptors. Freshly precipitated Fe(III)-(oxyhydr)oxides could not be reduced by mutants lacking the high-potential pathway. Aging these minerals by autoclaving did not change their powder X-ray diffraction pattern, but restored reduction by mutants lacking the high-potential pathway. Mutants lacking the low-potential, CbcL-dependent pathway had higher growth yields with both soluble and insoluble Fe(III). Together, these data suggest that the ImcH-dependent pathway exists to harvest additional energy when conditions permit, and CbcL switches on to allow respiration closer to thermodynamic equilibrium conditions. With evidence of multiple pathways within a single organism, the study of extracellular respiration should consider not only the crystal structure or solubility of a mineral electron acceptor, but rather the redox potential, as this variable determines the energetic reward affecting reduction rates, extents, and final microbial growth yields in the environment.",
"introduction": "Introduction Fe(III)-(oxyhydr)oxides can exist in at least 15 mineral forms with variable physiochemical properties that span a wide range of formal oxidation-reduction (redox) midpoint potentials ( Nealson and Saffarini, 1994 ; Schwertmann and Cornell, 2000 ; Thamdrup, 2000 ; Majzlan et al. , 2004 ; Majzlan, 2011 , 2012 ). The electron-accepting potential of any given Fe(III)-(oxyhydr)oxide structure is not a fixed value, and becomes less favorable with increasing crystallinity, particle size, pH or ambient Fe(II) concentration ( Thamdrup, 2000 ; Navrotsky et al. , 2008 ; Majzlan, 2012 ; Sander et al. , 2015 ). Differences in effective redox potential alters the energy available to be captured by bacteria able to couple the oxidation of an electron donor to reduction of these minerals ( Thauer et al. , 1977 ; Flynn et al. , 2014 ). As a result of this structural and environmental diversity, organisms able to reduce metals could sense the energy available in their electron acceptors and utilize different electron transfer pathways, similar to how E scherichia \n coli uses distinct terminal oxidases in response to levels of available oxygen ( Russell and Cook, 1995 ; Green and Paget, 2004 ). One well-studied dissimilatory metal-reducing organism, Geobacter sulfurreducens , can grow via reduction of metal (oxyhydr)oxides ranging in predicted midpoint redox potential from +0.35 V vs the standard hydrogen electrode (SHE; for example, birnessite, ca Na x Mn 2-x (IV)Mn(III) x O 4 , x~0.4) to –0.17 V vs SHE (for example, goethite, α-FeOOH) ( Post and Veblen, 1990 ; Nealson and Myers, 1992 ; Caccavo et al. , 1994 ; Thamdrup, 2000 ; Majzlan et al. , 2004 ; Majzlan, 2012 ; Orsetti et al. , 2013 ). Recent work examining electron transfer from G. sulfurreducens to poised graphite electrodes demonstrated that this organism uses at least two different inner membrane electron transfer pathways, known as the CbcL- and ImcH-dependent pathways ( Levar, et al. , 2014 ; Zacharoff, et al. , 2016 ). The ‘low' potential, CbcL-dependent pathway, is required for growth with electrodes at or below potentials of –0.10 V vs SHE, whereas the ‘high' potential, ImcH-dependent pathway, is essential when electrodes are poised at redox potentials above this value. As mutants deficient in either pathway only grow with electrodes poised above or below these thresholds, cells lacking CbcL or ImcH may also be able to be used as ‘sensors' to report the redox potential of an extracellular electron acceptor such as a metal (oxyhydr)oxide. As it is difficult to determine the effective redox potential of these minerals ( Sander et al. , 2015 ), such information could aid laboratory characterization, and provide evidence that bacteria use different pathways for different metals in the environment. The discovery of multiple inner membrane pathways is based on work with electrodes held at constant potentials, but poses many questions regarding G. sulfurreducens' interactions with more complex mineral substrates. Does the organism transition from one pathway to the other as simple environmental factors such as Fe(II):Fe(III) ratios alter redox potentials? Do manipulations known to alter redox potential, such as pH or particle size, also influence the pathway utilized? As this ±0.5 V redox potential range represents nearly 50 kJ per mol of electrons transferred, do different electron transfer mechanisms allow higher yield of this microbe? Here, we demonstrate that G. sulfurreducens requires both the CbcL- and the ImcH-dependent electron transfer pathways for complete reduction of a variety of Fe and Mn minerals. By using mutants only able to function above or below specific redox potentials, we show that minerals often used for study of extracellular electron transfer begin as ‘high' potential electron acceptors reducible by the ImcH-dependent pathway, but transition during reduction to ‘low' potential electron acceptors requiring the CbcL-dependent pathway. Simple variations in mineral handling such as autoclaving, or pH changes that alter redox potential by 30 mV, can alter the electron transfer pathway required by the organism, further showing that bacteria respond to subtle changes in mineral redox potentials. These data highlight the complexity of studying growth with minerals, and suggests that the proteins used for electron flow out of G. sulfurreducens are more strongly influenced by redox potential than the crystal structure, aggregation behavior or even solubility of the terminal acceptor.",
"discussion": "Discussion Fe(III)-(oxyhydr)oxide minerals in nature exist as a complex continuum of potential energies. As such, it is unsurprising that bacteria would evolve equally complex mechanisms able to harness the energy available during respiration to these metals. Although a single pathway essential for reduction of all metals emerged in the model metal-reducing Shewanella spp., a similarly simple solution remained elusive in Geobacter spp. The data reported here demonstrate that even under the most common laboratory conditions where Fe(III)-(oxyhydr)oxide is precipitated and added to medium, G. sulfurreducens will utilize multiple electron transfer pathways to accomplish what is considered to be wild-type levels of Fe(III) reduction. For Fe(III)-(oxyhydr)oxides that are predicted to have redox potentials around 0 V, such as schwertmannite, akaganeite and ferrihydrite, the electron-accepting potential decreases as Fe(II) accumulates, and triggers utilization of electron transfer pathways that support lower cell yield and slower growth rate, but still allow some respiration. The ability of G. sulfurreducens to respond to the changing redox potential of its environment likely allows this organism to make the most efficient use of the provided mineral. Based on proteomic studies, G. sulfurreducens expresses at least 78 multiheme c- type cytochromes under different laboratory conditions, yet mutant analyses were initially unable to link particular cytochromes to reduction of specific extracellular electron acceptors ( Leang et al. , 2003 , 2010 ; Shi et al. , 2007 ; Rollefson et al. , 2009 , 2011 ; Aklujkar et al. , 2013 ). One hypothesis for cytochrome diversity that drives such -omic comparisons is that separate pathways will exist based on solubility (chelated vs oxide) or metal (Fe vs Mn). However, despite their many physiochemical differences, Fe(III)-citrate and Mn(IV)-(oxyhydr)oxides both represent high-potential acceptors, from the viewpoint of the inner membrane, and our data suggest that this variable influences at least one portion of the electron transfer pathway. As we detect evidence for at least one other pathway in this work, and the G. sulfurreducens genome appears to encode at least four other inner membrane quinone oxidoreductases, redox potential differences may help to explain additional complexity of Geobacter. Possessing protein machinery able to extract the full advantage from minerals as they descend through the redox tower also helps explain the dominance of these metal-reducing organisms in contaminated subsurface environments undergoing rapid redox potential changes. In order to characterize and biochemically dissect extracellular electron transfer in dissimilatory metal-reducing organisms such as G. sulfurreducens , closer attention may be needed to be paid not just to crystal structure, but the actual redox potential experienced by the organisms. Commonly synthesized Fe-(oxyhydr)oxides prepared by precipitation of Fe(III) can produce very different rates of reduction or mutant phenotypes depending on the age of the material and the length of time one is willing to incubate cells. A distinction between initial rates of Fe(III)-(oxyhydr)oxide reduction, where higher-potential conditions exist, and final extent of reduction achieved by slower but lower potential pathways may help separate these confounding effects. As Fe(II) begins to accumulate during the course of Fe(III)-(oxyhydr)oxide reduction by metal-reducing microorganisms, the redox potential will change as the ratio of Fe(II)/Fe(III) changes and as mineral structure is altered in response in increased Fe(II) concentration. One would also expect significant differences in washed cell suspensions, where the degree of respiratory coupling to growth is not part of the assay, as electron flow through the ImcH- and CbcL-dependent pathways support different growth rates and yields. It now appears that in G. sulfurreducens a transition from an ImcH-dependent electron transfer pathway to a CbcL-dependent pathway occurs both when electrodes and commonly used Fe(III)-(oxyhydr)oxides are provided as sole terminal electron acceptors. The initial rate and total extent of Fe(III)-(oxyhydr)oxide reduction may also be the result of a series of redox potential dependent processes in environmental samples, with each step supporting different cell yields. Competition in the environment could occur along any part of this continuum, rewarding those capable of rapid growth when the Fe(III)-(oxyhydr)oxide is high potential, or those able to survive with electron acceptors at relatively low redox potentials. The prevalence of ImcH and CbcL homologs in other dissimilatory metal-reducing organisms and in metagenomics data from sites undergoing active metal reduction ( Levar et al. , 2014 ; Zacharoff et al. , 2016 ) suggests that this type of redox discrimination could also occur in diverse organisms, and provides opportunities for further exploration of redox potential dependent respiration in these organisms and their environments."
} | 2,846 |
31015394 | PMC6478937 | pmc | 1,183 | {
"abstract": "Large reservoirs of natural gas in the oceanic subsurface sustain complex communities of anaerobic microbes, including archaeal lineages with potential to mediate oxidation of hydrocarbons such as methane and butane. Here we describe a previously unknown archaeal phylum, Helarchaeota, belonging to the Asgard superphylum and with the potential for hydrocarbon oxidation. We reconstruct Helarchaeota genomes from metagenomic data derived from hydrothermal deep-sea sediments in the hydrocarbon-rich Guaymas Basin. The genomes encode methyl-CoM reductase-like enzymes that are similar to those found in butane-oxidizing archaea, as well as several enzymes potentially involved in alkyl-CoA oxidation and the Wood-Ljungdahl pathway. We suggest that members of the Helarchaeota have the potential to activate and subsequently anaerobically oxidize hydrothermally generated short-chain hydrocarbons.",
"introduction": "Introduction Short-chain alkanes, such as methane and butane, are abundant in marine sediments and play an important role in carbon cycling with methane concentrations of ~1 Gt being processed globally through anoxic microbial communities 1 – 3 . Until recently, archaeal methane cycling was thought to be limited to Euryarchaeota 4 . However, additional archaeal phyla, including Bathyarchaeota 5 and Verstraetarchaeota 6 , have been shown to contain proteins with homology to the activating enzyme methyl-coenzyme M reductase (Mcr) and corresponding pathways for methane utilization. Furthermore, lineages within the Euryarchaeota belonging to Candidatus Syntrophoarchaeum spp., have been shown to use methyl-CoM reductase-like enzymes for anaerobic butane oxidation 7 . Similar to methane oxidation in many ANME-1 archaea, butane oxidation in Syntrophoarchaeum is proposed to be enabled through a syntrophic interaction with sulfur-reducing bacteria 7 . Metagenomic reconstructions of genomes recovered from deep-sea sediments from near 2000 m depth in Guaymas Basin (GB) in the Gulf of California have revealed the presence of additional uncharacterized alkyl methyl-CoM reductase-like enzymes in metagenome-assembled genomes within the Methanosarcinales (Gom-Arc1) 8 . GB is characterized by hydrothermal alterations that transform large amounts of organic carbon into methane, polycyclic aromatic hydrocarbons, low-molecular weight alkanes and organic acids allowing for diverse microbial communities to thrive (Supplementary Table 1 ) 8 – 11 . Recently, genomes of a clade of uncultured archaea, referred to as the Asgard superphylum that includes the closest archaeal relatives of eukaryotes, have been recovered from anoxic environments around the world 12 – 14 . Diversity surveys in anoxic marine sediments show that Asgard archaea appear to be globally distributed 12 , 14 – 16 . Based on phylogenomic analyses, Asgard archaea have been divided into four distinct phyla: Lokiarchaeota, Thorarchaeota, Odinarchaeota, and Heimdallarchaeota, with the latter possibly representing the closest relatives of eukaryotes 12 . Supporting their close relationship to eukaryotes, Asgard archaea possess a wide repertoire of proteins previously thought to be unique to eukaryotes known as eukaryotic signature proteins (ESPs) 17 . These ESPs include homologs of eukaryotic proteins, which in eukaryotes are involved in ubiquitin-mediated protein recycling, vesicle formation and trafficking, endosomal sorting complexes required for transport-mediated multivesicular body formation, as well as cytokinetic abscission and cytoskeleton formation 18 . Asgard archaea have been suggested to possess heterotrophic lifestyles and are proposed to play a role in carbon degradation in sediments; however, several members of the Asgard archaea also have genes that code for a complete Wood–Ljungdahl pathway and are therefore interesting with regard to carbon cycling in sediments 14 , 19 . Here, we present metagenome-assembled genomes (MAGs), recovered from GB deep-sea hydrothermal sediments, which represent an undescribed Asgard phylum with the metabolic potential to perform anaerobic hydrocarbon degradation using a methyl-CoM reductase-like homolog.",
"discussion": "Discussion Historically methanogenesis and anaerobic methane oxidation were regarded as the only examples of anaerobic archaeal short-chain alkane metabolism. The enzymes acting in these pathways were considered to be biochemically and phylogenetically unique and limited to lineages within the Euryarchaeota 4 . This study represents the discovery of the previously unknown phylum referred to as Helarchaeota, whose members encode a mcr-like gene cluster. This opens the possibility that some representatives of the Asgard archaea may have the potential for anaerobic short-chain alkane oxidation. Since the presence of these mcr genes is restricted to Helarchaeota among the known Asgard archaea 19 , these genes were likely transferred to Helarchaeota and do not constitute an ancestral trait within the Asgard superphylum. Based on current phylogenetic analysis, the Helarchaeota mcr gene cluster may have been horizontally acquired from either Bathyarchaeota or Ca . Syntrophoarchaeum (Fig. 1b , Supplementary Figure 3 ). Due to this close relationship, we based our analysis of Helarchaeota’s ability to perform anaerobic short-chain hydrocarbon oxidation on the pathway proposed for Ca. Syntrophoarchaeum. Helarchaeota probably utilize a similar short-chain alkane as a substrate in lieu of methane, but given the low-butane concentrations at our site it may not be the only substrate. Our comparison to Ca . S. butanivorans shows a consistent presence in genes necessary for this metabolism including a complete Wood–Ljungdahl pathway, acyl oxidation pathway, and internal electron transferring systems. Some of these electron-transferring systems are essential housekeeping components that may act as electron carriers for oxidation reactions. Interestingly, in the Wood–Ljungdahl pathway identified in Ca . S. butanivorans , the bacterial enzyme 5,10-methylene-tetrahydrofolate reductase (met) is thought to be substituting for the missing 5,10-methylene-tetrahydromethanopterin reductase (mer) 7 . In contrast, Helarchaeota encode the canonical archaeal-type mer. To render anaerobic butane oxidation energetically favorable, it must be coupled to the reduction of an electron acceptor such as nitrate, sulfate or iron 7 , 26 , 27 . In ANME archaeum that lack genes for internal electron acceptors, methane oxidation is enabled through the transfer of electrons to a syntrophic partner organism. In Syntrophoarchaeum, syntrophic butane oxidation is thought to occur through the exchange of electrons via pili and/or cytochromes with sulfate-reducing bacteria 7 . Helarchaeota do not appear to encode any of the systems traditionally associated with syntrophy and no partner was identified in this study. Thus, further research is needed to identify possible bacterial partners. Furthermore, the hypothesis that Helarchaeota have the ability to utilize short-chain alkanes remains to be confirmed as the genomes of members of this group do not encode canonical routes for electron transfer to a partner bacterium. However, we identified potential enzymes that may be involved in transfer of electrons. Some methanogenic archaea use formate for syntrophic energy transfer to a syntrophic partner; therefore, the reverse reaction has been speculated to be energetically feasible for methane oxidation 27 . If this is true, the presence of a membrane-bound formate dehydrogenase in the Helarchaeota genomes may support this electron-transferring mechanism, however, to our knowledge this has never been shown for an ANME archaea so far. Alternatively, the type 3 NiFe-hydrogenases encoded by Helarchaeota may be involved in transfer of hydrogen to a partner organism. For example, we identified a protein complex distantly related to the mvh–hdr of methanogens for electron transfer ( Supplementary Methods ). Mvh–hdr structures have been proposed to be potentially used by facultative hydrogenotrophic methanogens for energy transfer, but the directionality of hydrogen exchange could easily be reversed 2 . These methanogens form syntrophic associations with fermenting, H 2 -producing bacteria, lack dedicated cytochromes or pili and use the mvh–hdr for electron bifurcation 2 . The detection of a hydrophobic region in the mvh–hdr complex led to the suggestion that this complex could be membrane bound and act as mechanism for electron transfer across the membrane; however, a transmembrane association has never been successfully shown 2 . While the membrane association of the disulfide reductase/FlpD needs to be confirmed, we were able to detect several other transmembrane motifs in the associated proteins that could potentially allow electron transfer in form of hydrogen to an external partner. Thus, while we propose that the most likely explanation for anaerobic short-chain alkane oxidation in Helarchaeota is via a syntrophic interaction with a partner, additional experiments are needed to confirm this working hypothesis. The discovery of alkane-oxidizing pathways and possible syntrophic interactions in a phylum of Asgard archaea indicates a much wider phylogenetic range for hydrocarbon utilization. Based on phylogenetic analyses it seems most likely that the Helarchaeota mcr operon may have been horizontally transferred from either Bathyarchaeota or Syntrophoarchaea. However, the preservation of a horizontally transferred pathway is indicative of a competitive advantage; it follows that gene transfers among different archaeal phyla reflect alkane oxidation as a desirable metabolic trait. The discovery of the alkyl-CoM reductases and alkane-oxidizing pathways among the Asgard archaea indicates ecological roles for these still cryptic organisms, and opens up a wider perspective on the evolution and expansion of hydrocarbon-oxidizing pathways throughout the archaeal domain."
} | 2,495 |
39858512 | PMC11764364 | pmc | 1,186 | {
"abstract": "Legumes play a pivotal role in addressing global challenges of food and nutrition security by offering a sustainable source of protein and bioactive compounds. The capacity of legumes to establish symbiotic relationships with rhizobia bacteria enables biological nitrogen fixation (BNF), reducing the dependence on chemical fertilizers while enhancing soil health. However, the efficiency of this symbiosis is significantly influenced by environmental factors, such as soil acidity, salinity, temperature, moisture content, light intensity, and nutrient availability. These factors affect key processes, including rhizobia survival, nodule formation, and nitrogenase activity, ultimately determining the growth and productivity of legumes. This review summarizes current knowledge on legume-rhizobia interactions under varying abiotic conditions. It highlights the impact of salinity and acidity in limiting nodule development, soil temperature in regulating microbial community dynamics, and moisture availability in modulating metabolic and hormonal responses during drought and waterlogging. Moreover, the role of essential nutrients, including nitrogen, phosphorus, potassium, and trace elements such as iron, molybdenum, and boron, in optimizing symbiosis is critically analyzed.",
"conclusion": "8. Conclusions The review highlights the multifaceted role of legumes in sustainable agriculture and their ability to mitigate environmental challenges while providing high-quality protein sources. The symbiosis between legumes and rhizobia bacteria is crucial for biological nitrogen fixation (BNF), contributing significantly to soil fertility and reducing the dependency on synthetic nitrogen fertilizers. However, the efficiency of this process is highly influenced by environmental factors, including soil salinity, acidity, temperature, moisture, light availability, and nutrient content. Addressing these challenges requires a comprehensive understanding of the mechanisms underlying legume–rhizobia interactions and their response to abiotic stresses. Innovations in seed inoculation, soil management practices, and nutrient supplementation can improve BNF efficiency and crop yields. The economic and environmental benefits of legumes underscore their importance in sustainable agricultural systems, where they enhance soil health, support biodiversity, and reduce greenhouse gas emissions. Future research should focus on breeding stress-tolerant legume varieties and developing advanced inoculants to maximize BNF under diverse environmental conditions. This holistic approach will enable the full utilization of the potential of legumes, contributing to global food security and environmental sustainability.",
"introduction": "1. Introduction In solving global food and nutrition security problems, the cultivation of legumes as the main source of highly nutritious protein resources is gaining strategic importance. It is well known that protein, as an important nutrient, plays a key role in the physiological and biochemical reactions in the organism necessary to ensure the body’s growth and development [ 1 , 2 ]. In the structure of protein resources used for nutrition, proteins of animal origin play a significant role. However, according to the results of scientific research, the presence of a large amount of meat in the human diet can lead to obesity and disorders of the functioning of the cardiovascular system related to it. On the other hand, the so-called hidden hunger caused by insufficient micronutrients in the daily diet leads to many health problems, especially for children. They can be manifested in stunted growth, poor weight gain, cognitive impairment, and mental disorders [ 3 , 4 ]. One of the components that can ensure a healthy diet for the population of both developed and developing countries can be highly nutritious plant products made from legume seeds [ 5 ]. The main constituent components of the legume seeds are protein, complex carbohydrates, dietary fiber, vitamins, minerals, and biologically active compounds. Legume seeds are characterized by the absence of cholesterol and, with the exception of soybeans, chickpeas, white lupine, and peanuts, contain a low amount of lipids [ 6 ]. Biologically active compounds of legumes (including phytoestrogens, saponins, phenolic compounds, oligosaccharides, and alkaloids) are characterized by antithrombotic, antioxidant, opioid-like activity, have cytomodulatory or immunomodulatory effects, and contribute to the improvement of mineral bioavailability [ 7 , 8 , 9 ]. Pharmaceuticals based on their raw materials are widely used in the prevention of occurrence and medical therapy of cardiovascular, cancer diseases, diabetes, and obesity [ 10 , 11 ]. At the same time, legumes are a valuable source of highly nutritious feed resources used in livestock farming [ 12 , 13 ]. Cultivation of legumes improves the biological, physical and chemical characteristics of the soil [ 14 , 15 ], contributes to the reduction in greenhouse gasses in the atmosphere [ 16 ], and limits the occurrence of diseases, pests, diseases, and nematodes by disrupting the biological cycle of their development [ 17 ]. Legumes have a unique biological property, which consists of establishing symbiotic relationships with nitrogen fixing bacteria of the genera Rhizobium . Kebebe [ 18 ] notes that in the process of legume–rhizobia symbiosis, legumes can fix about 100 to 300 kg ha −1 of atmospheric nitrogen (N 2 ) annually, thus providing about 139–175 million tons of nitrogen to the soil. This, in turn, reduces the cost of applying about 80–90 tons of mineral nitrogen fertilizers annually [ 19 ]. Nitrogen (N) is an element that occurs mainly in the atmosphere, lithosphere, and biosphere in various degrees of oxidation [ 20 , 21 ]. Only about 2% of this element is found in living organisms and their biocenoses, with the remainder occurring in inorganic form [ 22 ]. Nitrogen in the gaseous form (it makes up about 78% of the atmospheric composition) is very persistent due to the triple covalent bond between N atoms, but in this form, it is unabsorbable by most higher organisms and unreactive [ 23 ]. In the bioavailable form, N is found in soil in the form of organic and inorganic compounds such as nitrates –NO 3 , nitrites –NO 2 , and ammonium ions –NH 4 [ 24 , 25 ]. Biological nitrogen fixation (BNF) is one of the most important processes on Earth. It involves the conversion of N 2 , which is not assimilated by plants, into ammonia, the form available to plants [ 26 , 27 ]. This process is a highly energetic transformation, requiring reducing forces, i.e., electrons and protons (H + ), as well as large amounts of energy (ATP) [ 28 ]. The symbiotic relationship between rhizobia and legumes plays a key role in sustainable agriculture by facilitating access to N 2 , improving soil fertility and reducing the need for chemical fertilizers. After the mutualistic relationship between the two symbiosis partners is established, the roots of plants form structures known as root nodules, in which MgATP-dependent N 2 reduction by the nitrogenase complex occurs ( Figure 1 ). The produced ammonia is used by plants as a source of N in the synthesis of organic compounds. To support the bacterial endosymbionts’ metabolism, photosynthetic products are supplied from the plant to nodules [ 29 ]. The establishment of the legume–rhizobia symbiosis as an N 2 -fixing system is interconnected with the physiological status of the plant host and is determined by the influence of environmental factors, such as soil type, temperature, moisture, pH, salinity, micro, and macronutrient content [ 19 , 30 , 31 ]. These factors can affect various aspects of symbiosis, including rhizobia survival in the soil, infection process, legume development, nodule function, and, indirectly, the growth of the host plant [ 32 ]. Due to symbiosis, legumes cover their nutritional needs in relation to N and, therefore, have lower requirements for mineral fertilization, which is limited to the necessary minimum, generally a small pre-sowing (starter) dose. This allows deficiencies of the nutrient to be made up in the initial period of plant growth and development, when the BNF process has not yet begun [ 33 ]. Individual legume species can only coexist with a specific bacterial species, for example, beans ( Phaseolus vulgaris L.) have the ability to symbiosis with Rhizobium leguminosarum bv. phaseoli , and soybean ( Glycine max (L.) Merr.) with Bradyrhizobium japonicum ( Table 1 ). Bacteria coexisting with legumes do not always occur naturally in the soils in which these species are grown, so it is sometimes necessary to inoculate legume seeds prior to sowing with appropriately selected bacterial strains to enable the rapid colonization of the rhizosphere and ensure effective nodulation, thereby improving the BNF process and maximizing yield [ 34 , 35 , 36 ]. However, the majority of soils used in agriculture contain rhizobia populations that compete in the plant host nodulation process with strains introduced by inoculation. The aim of this review is to summarize the current state of knowledge regarding the symbiosis between legumes and rhizobia bacteria and to identify key environmental factors affecting the efficiency of biological nitrogen fixation (BNF). Particular focus is placed on the impact of factors such as soil salinity, acidity, temperature, moisture, light intensity, and the availability of macro- and micronutrients. This review seeks to highlight the challenges and opportunities associated with optimizing the symbiosis, which can enhance legume productivity and support sustainable agricultural practices."
} | 2,426 |
37232997 | PMC10217856 | pmc | 1,187 | {
"abstract": "The persistent challenge of removing viscous oil on water surfaces continues to pose a major concern and requires immediate attention. Here, a novel solution has been introduced in the form of a superhydrophobic/superoleophilic PDMS/SiO 2 aerogel fabric gathering device (SFGD). The SFGD is based on the adhesive and kinematic viscosity properties of oil, enabling self-driven collection of floating oil on the water surface. The SFGD is able to spontaneously capture the floating oil, selectively filter it, and sustainably collect it into its porous fabric interior through the synergistic effects of surface tension, gravity, and liquid pressure. This eliminates the need for auxiliary operations such as pumping, pouring, or squeezing. The SFGD demonstrates exceptional average recovery efficiencies of 94% for oils with viscosities ranging from 10 to 1000 mPa·s at room temperature, including dimethylsilicone oil, soybean oil, and machine oil. With its facile design, ease of fabrication, high recovery efficiency, excellent reclaiming capabilities, and scalability for multiple oil mixtures, the SFGD represents a significant advancement in the separation of immiscible oil/water mixtures of various viscosities and brings the separation process one step closer to practical application.",
"conclusion": "3. Conclusions In summary, the burlap fabric coated by a layer of superhydrophobic/superoleophilic PDMS/SiO 2 composite aerogel coating were initially and creatively utilized for self-driven floating viscous oil collection. The SFGD could spontaneously attract, filter, and collect the floating oils under the synergetic effect of surface tension, gravity, and liquid pressure, requiring no extra operations such as pumping, oil/water mixture collecting, and squeezing. The superhydrophobic/superoleophilic burlap sack with excellent mechanical robustness and functionality could be easily scaled by a flexible dip-coating method and utilized for oil/water separation, displaying excellent oil recovery efficiency even after 50 cycles of usage. Furthermore, during long-time water resistance detection, the SFGD filled with the collected oil remained floating on water for 15 days with oil recovery efficiency no less than 95.8%, showing excellent endurance and practicability. It is worth stating that the SFGD could be scaled for the efficient elimination of large multiple oil mixtures on water, demonstrating the versatility for oil removal. We firmly believe that such a smartly designed oil-collection system could provide a unique perspective for dealing with floating viscous oil pollution.",
"introduction": "1. Introduction With the development of the social economy and industrialization, oil pollution such as waste edible oil, mechanical abandoned oil, and industrial spilled oil increases rapidly, which results in a serious threat to the ecosystem and human health [ 1 , 2 ]. Conventional methods and technologies such as flotation, separators, centrifugation, oil containment booms, and skimmers have been developed for oil removal but are not effective for totally eliminating oil from water, especially oil with viscosity, thus making the separation incomplete with oil residual in water [ 3 , 4 ]. Moreover, these methods usually involve low selectivity, tedious operations, energy-consuming processes, low separation efficiency, and so on, which severely blocks the practical usage of these approaches. Therefore, novel materials with special selectivity, good mechanical stability, excellent separation efficiency, and reliable recyclability are urgently needed for the separation of oil/water mixtures [ 5 , 6 ]. Recently, various separation materials with special wettability such as membrane films, porous materials, and gelation have been widely developed [ 7 , 8 , 9 , 10 , 11 , 12 ]. Superhydrophilic materials with hierarchical structures and water-binding affinity could adsorb water and make water become trapped in the rough structures once contacting water during the separation process [ 13 , 14 ]. The adsorbed water forms a hydration layer, which definitely reduces the contact area between the oil and sample surface and thus decreases the oil adhesive property. During the separation process, these materials could easily attract and filter water content from oil/water mixtures under gravity or external force, often displaying outstanding oil/water separation efficiency [ 15 , 16 ]. However, besides the drawbacks of the rigorous fabrication process, complicated operation steps, and the process being energy-consuming, having low recyclability, and sometimes requiring pump assistance, the large-scale usage of these materials for oil/water separation still remains limited because of their separation style: the water content of the viscous-oil/water mixtures passes through the materials instead of gathering floating viscous oil content from the mixtures, requiring oil/water mixtures to be gathered first [ 17 ]. Meanwhile, most reported materials are mainly focused on organic solvents or light oils (e.g., gasoline and diesel) and are still easily polluted and fouled by sticky oils [ 18 , 19 , 20 ]. Porous absorbent materials with water repellency such as sponges, foam materials, rubber, carbon-based materials, and chemosynthesis adsorbents have drawn much attention to dealing with floating oil assigned to their prominent adsorption characteristics and extrusion property [ 21 , 22 , 23 ]. The materials could spontaneously and selectively adsorb oil from the water surface due to their lipophilicity and water resistance, usually possessing excellent adsorption capacity and exceptional separation efficiency. Polydimethylsiloxane (PDMS) sorbent has been widely used for oil–water separation due to its high selectivity and recovery rates for oil types, as well as its ease of use and cost-effectiveness compared to other sample preparation methods [ 24 , 25 , 26 ]. However, its use also has several disadvantages, including higher costs than other sample preparation materials, environmental impact due to the difficulty of biodegradation of the synthetic material, the need for reusability, and limited capacity for large sample volumes. Graphene/PDMS sponge has gained attention as a promising material for oil–water separation due to its high surface area and selective affinity for different types of oil. In addition, its reusability and durability make it a cost-effective solution for oil–water separation [ 27 , 28 , 29 ]. However, the synthesis and preparation of graphene/PDMS sponge can be challenging and requires advanced methods, resulting in high production costs. Furthermore, the stability of the material may not be consistent over time, leading to degradation and decreased efficiency in oil–water separation. However, it is of concern that these adsorbent materials, once used for viscous oil adsorption and removal, would show a dramatic decline in adsorption properties, leading to an inevitable decrease in recovery efficiencies after a limited number of uses [ 21 ]. The primary reason for the decline is due to the stickiness and kinematic viscosity of the oils, which would lead to accumulated contaminants, severe pore fouling, and an irreversible decrease in their adsorption properties. In the meantime, artificial squeezing or pump-driven procedures may require tedious separation operations and high energy consumption. Therefore, novel devices and methods with easy operation, self-driven property, good oil recovery, and stable recyclability should be designed for floating viscous oil collection and removal. Herein, a superhydrophobic/superoleophilic PDMS/SiO 2 aerogel fabric gathering device (SFGD) is designed for collection and removal of floating viscous oil spills. Firstly, a layer of superhydrophobic/superoleophilic PDMS/SiO 2 composite aerogel coating is prepared on the surface of the burlap fabric using PDMS and nano-silica aerogel particles [ 30 , 31 ]. Subsequently, the burlap fabric is combined with porous plastic balls (i.e., by filling the porous plastic balls as an internal support frame inside the burlap sack) to form the SFGD. The device could spontaneously gather floating oil on a water surface and sustainably collect it into a porous sack relying on the effects of surface tension, gravity, and liquid pressure, requiring no auxiliary operations (e.g., pumping, pouring, and squeezing). During the separation process, the partially submerged fabric surface remains superhydrophobic even underwater, which mainly results from the adhered waterproof viscous oil, hierarchical surface structures, and the intrinsic water repellency of the textile surface. Therefore, an interesting phenomenon appears that the water tightly wraps the submerged part of the sack surface while being totally forbidden to pass through the textured surface; simultaneously, the oil could easily flow into the sack while being inevitably locked into the sack by the wrapped water. Finally, the floating viscous oil is collected into a container and the SGFD could be reutilized for the oil/water separation, thus demanding no clean-up treatment. Moreover, the fabrics display good water repellence, excellent wear resistance detected by an oscillating abrasion tester, and prominent reusability via a recycling experiment. The facile prepared SFGD with scalable fabrication, high recovery efficiency, and prominent reclamation for the separation of immiscible oil/water mixtures of various oil viscosities utilizing a self-driven approach is displayed to illustrate the necessity of the unique device designed herein.",
"discussion": "2. Results and Discussion 2.1. Mechanism and Surface Wettability Polydimethylsiloxane has been widely applied for the construction of superhydrophobic/superoleophilic surfaces due to its low surface energy and adhesive properties [ 32 ].Nano-silica aerogel particles were prepared by the sol-gel method and used as the surface coating composition for providing additional nanoscale surface roughness and enhancing surface [ 30 ]. As shown in Figure 1 a, PDMS prepolymer was dissolved into n-hexane solution under stirring and then the added nano-silica aerogel particles were dispersed under ultrasonic treatment because of the aggregation of the particles. After a simple dip-coating procedure and a subsequent drying process, the burlap fabric was coated with a uniform superhydrophobic PDMS/SiO 2 composite aerogel coating as exhibited in the SEM images ( Figure 1 b). During the preparation of the PDMS/SiO 2 composite aerogel coating on the fabric surface, the adhesive properties of the PDMS played a crucial role for both the fabric substrate and silica. The strong adhesion of the PDMS ensured firm and continuous bonding between the coating and the fabric surface, which was essential for the long-term stability and durability of the superhydrophobic coating. The adhesive properties of the PDMS also prevented the coating from peeling or flaking off the fabric substrate, ensuring that the coating remained intact even under high stress or wear conditions. In addition, the addition of nano-silica aerogel particles not only built a rough structure for the overall superhydrophobic coating, but also increased its mechanical stability and durability. The adhesive properties of the PDMS also helped to bond the silica particles to the fabric, forming a strong and uniform coating that could resist wear and tear. Therefore, the combination of the PDMS and silica enhanced the overall performance of the coating and made it more superhydrophobic. With the synergistic effect of hybridization and composite formation between PDMS and nano-silica aerogel particles, a layer of superhydrophobic/superoleophilic PDMS/SiO 2 composite aerogel coating is formed on the surface of burlap fabric. Generally, surfaces with low surface energy showed a stronger affinity to oil than water [ 33 , 34 ]. The wettability of the coated burlap fabric was measured in air conditions with water contact angles (WCAs) of 156° ( Figure 1 c top) and an oil contact angle of 0° ( Figure 1 c bottom), respectively, showing excellent water repellency and superoleophilicity. Once it contacted the as-prepared sample, the oil droplet would spread on the modified surface (see the Supplementary Information, Figure S2 ). As demonstrated in Figure 1 d, the sack coated by PDMS/SiO 2 composite aerogel coating displayed excellent water impact resistance ( Figure 1 d left) and outstanding water-holding properties ( Figure 1 d right). As reported, the modified Young’s equation could be not only applicable to analyze the wettability of an oil droplet on a solid surface underwater but also valid to a water droplet on a surface in oil [ 35 , 36 ]. The modified formula of water contact angle on an ideal smooth surface underoil ( θ W O ) is displayed in Equation (1). According to the equation,\n (1) cos θ W O ′ = r cos θ W O \nwhere θ W O ′ represents the water contact angle of a water droplet on the rough surface and r is the roughness of the surface. the value of surface roughness ( r ) is greater than 1, illustrating that for material with underoil WCA ( θ W O ) more than 90°, the real value of θ W O ′ increase with the strengthening of surface roughness. Thus, the nano-silica aerogel-particles play an essential role in the construction of underoil superhydrophobicity. The wettability of the as-prepared rough surfaces with superhydrophobicity was explored via a contact angle meter ( Figure 2 and Figure S3 ). In oil, the superoleophilic PDMS/SiO 2 composite aerogel coating could quickly absorb oil and the oil could be tightly adhered to the textured surface and trapped in the rough structures, which was attributed to the excellent surface affinity to oil. The oil-coated fabric surface demonstrated remarkable superhydrophobicity with a WCA of 154.1° under n-hexane ( Figure 2 g). As shown in Figure 2 a–d, when a water droplet came into contact with the oil/solid composite interface ( Figure 2 b) under n-hexane and was pressed downward ( Figure 2 c), the droplet was compressed but did not wet the interface, retaining its intact shape. When the water droplet was lifted, it remained non-adherent to the interface ( Figure 2 d). Additionally, when the water droplet was brought into contact with the interface and pressed down a certain distance, upon moving the droplet to the right ( Figure 2 e) or left ( Figure 2 f), there was no lag or adhesion between the droplet and the interface. Furthermore, as shown in Figure 2 h, the WCA on the textured surface under dimethylsilicone oil was 165.8°, providing a solid theoretical foundation for the collection and removal of viscous oils (the WCAs on the composite surface in the other viscous oils can be seen in Figure S3 ). In summary, the burlap fabric coated by PDMS/SiO 2 composite aerogel coating possesses predominantly excellent oil affinity and underoil water repellency. 2.2. Surface Chemical Component Analysis As displayed in Figure 3 , the chemical composition of the prepared burlap fabric was examined by X-ray photoelectron spectroscopy (XPS) ( Figure 3 a) and FT-IR spectroscopy ( Figure 3 b). Table S1 summarizes the quantitative data obtained for pristine burlap fabric, burlap fabric coated with PDMS only, and PDMS/SiO 2 composite aerogel coating. The XPS graphic of the burlap fabric exhibited C1s, O1s, Si2p, and Si2s peaks as shown in Figure 3 a. Comparing the results, there was a significant increase in Si peaks after the fabrics were coated with PDMS and PDMS/SiO 2 . Meanwhile, there was an apparent relatively higher ratio of O to C (43.3/24.1) or Si to C (32.6/24.1) content on the PDMS/SiO 2 -coated fabric compared to the PDMS coated fabric (O to C = 27.6/48.8; Si to C = 23.6/48.8), which could be attributed to the nano-silica aerogel particles [ 37 ]. The contents increase could also be qualified through EDX as shown in Figure S4 . Figure 3 b showed that the absorption peaks at 1050 cm −1 and 1262 cm −1 were assigned to Si-O-Si stretching vibrations attributed to silicon dioxide and silicon rubber. The occurrence of prominent bands at 796 cm −1 represented the Si-C vibrations conforming that the PDMS adhered to the fabric surface. In summary, Figure 3 demonstrated that silicon dioxide and silicon rubber were deposited on the superhydrophobic burlap fabric surface. 2.3. Mechanical Robustness The stability of micro- and nano-scale rough structures is an essential part of preparing durable special wettability surfaces which deeply affects the practicability of the materials [ 38 ]. The stickiness of the PDMS could coat and fix the nanoparticles on the fabric surfaces and the elasticity of the dried PDMS will disperse the forces once subjected to mechanical forces, which could protect the coated nano silica from facing the external force directly. To confirm the mechanical stability of the PDMS/SiO 2 composite aerogel coating on the fabric surface, the as-prepared burlap fabric was investigated by an oscillating abrasion tester as shown in Figure 4 a. The fabric side of the sample was placed into the oscillating abrasion tester equipment with grit covering the whole sample. As exhibited in Figure 4 b, the WCA of the sample remained above 152° even after 1500 abrasion cycles showing excellent durability. The stable hydrophobicity of the burlap fabric was primarily attributed to the remaining PDMS/SiO 2 rough structures ( Figure 4 b inserted images). Moreover, a blue-dyed water droplet could easily roll off the fabric surface after being treated for 1500 cycles (see the supplementary Information, Figure S6 ). 2.4. Separation of Viscous Oil/Water Mixture For understanding the separation mechanism of the immiscible oil/water mixture via the SFGD, a schematic illustration for viscous oil collection and removal was provided as displayed in Figure 5 . The SFGD was assembled by a superhydrophobic/superoleophilic burlap sack and a porous plastic ball ( Figure 5 (a,b 1 )). The porous hollow ball was selected as an internal prop that could keep the sack owning a steady inner space. When dropping into a beaker containing viscous oil/water mixtures ( Figure 5 (b 2 )), the SFGD would partially submerge in water due to gravity but not sink because of the buoyancy meaning that the average density of the device is lower than water. Subsequently, due to the superhydrophobic/superoleophilic properties of the SFGD surface fabric, the viscous oil on the water surface rapidly wetted the SFGD surface and slowly and self-drivenly passed through the outer fabric of the SFGD under the joint gravity of the SFGD and the oil as well as the liquid pressure, while the water was selectively blocked on the outer side due to the superhydrophobicity of the burlap fabric and the superhydrophobicity under the oil, finally completing the collection of the viscous oil on the water surface ( Figure 5 (b 3 )). Thus, an interesting phenomenon occurs that the water tightly wraps the submerged part of the sack surface while being totally prevented from penetrating the textured surface; however, the oil could be easily attracted and flow into the sack while being inevitably locked into the sack by the wrapped water. As shown in Figure 5 (b 4 ), when equilibrium (zero resultant force) was achieved, the oil-filled SFGD still partially submerged in water and remained unsinkable ascribed to its lower density than the water, and it maintained water repellency because of the water resistance of the collected oil and surface superhydrophobicity. Subsequently, the device was taken out and the oil was poured into a container. The leakage rate of the collected oil was rather low due to the relatively high kinematic viscosity of the inherent oil property, which could guarantee the oil recovery. Finally, after the pouring process, the SFGD was directly reused for viscous oil collection and removal without any wash treatment ( Figure 5 (b 1 ,b 5 )). Figure 6 a shows that floating silicon oil (100 ± 8 mPa·s) was collected and removed from the blue-dyed water surface via the SFGD following the steps in Figure 5 b. It required 2 h for managing the oil collection. Obviously, on first use, the SFGD was gradually adsorbed with silicon oil during the gathering process which led to the mass increase of the oil-poured empty SFGD. Therefore, after the oil removal process, the weight of the uncleaned empty SFGD increased from 14.67 g to 38.09 g, and the weight difference of the device before and after usage was 23.42 g as shown in Figure 6 b. The mass and recover efficiency of the collected oil was 68.9 g and 71.5%, respectively. However, when repeating the above separation via an uncleaned device, the weight difference of the SFGD could be controlled to ±2 g with the average recovery of 94.81% after 50 cycles ( Figure 6 b). Furthermore, during the long-time water resistance detection (see the Supplementary Information, Figure S6 ), the SFGD filled with the collected oil remained floating on the water for 15 days with an oil recovery efficiency of no less than 95.8%, showing excellent practicability. The SFGD could be applied to collect various viscous oils, such as dimethylsilicone oil (viscosity 100 ± 8 mPa·s, 500 ± 8 mPa·s, 1000 ± 8 m mPa·s), soybean oil, lubricant oil, anti-wear hydraulic oil, and gasoline engine oil, with recovery efficiency above 94% as exhibited in Figure 6 c. To evaluate the practical usage of the device (burlap sack: 15 × 20 cm, porous plastic ball diameter: 8 cm), 300 mL of viscous oil mixture containing dimethylsilicone oil (viscosity 100 ± 8 mPa·s), soybean oil, lubricant oil, anti-wear hydraulic oil, and gasoline engine oil at the ratio of 1:1:1:1:1 was added to in a 2000 mL beaker filled with 1500 mL of water ( Figure 7 a). The special wettability of the fabric provided the key foundation of the self-driven, gravity and liquid pressure aided, floating viscous oil collection device as illustrated. When dropped into the beaker, the enlarged SFGD with oil binding affinity was gradually wetted and adhered to by the oil mixture as shown in Figure 7 b,c. The floating viscous oil mixture was driven to be filtered by the fabric surface and collected in the inner surface of the sack ( Figure 7 d,e) under the effectiveness of the surface tension, gravity, and liquid pressure and simultaneously. Though repelled by the hydrophobic fabric, the water seamlessly wrapped the device due to liquid pressure and indeed prevented the oil mixture from leaking from the sack. When taken out, the device could retain the oil mixture without any leakage, which mainly resulted from the adhesive properties and kinematic viscosity of the oil. Thus, the oil mixture was completely collected and removed from water ( Figure 7 f), exhibiting that this device with remarkable functionality could be easily scalable. Thus, the self-driven oil collection SFGD was successfully applied for the separation and removal of immiscible viscous oil/water mixtures with excellent recyclability and practicability."
} | 5,769 |
38166159 | PMC10805104 | pmc | 1,188 | {
"abstract": "Complex and fluid\nbacterial community compositions are\ncritical\nto diversity, stability, and function. However, quantitative and mechanistic\ndescriptions of the dynamics of such compositions are still lacking.\nHere, we develop a modularized design framework that allows for bottom-up\nconstruction and the study of synthetic bacterial consortia with different\ntopologies. We showcase the microbial consortia design and building\nprocess by constructing amensalism and competition consortia using\nonly genetic circuit modules to engineer different strains to form\nthe community. Functions of modules and hosting strains are validated\nand quantified to calibrate dynamic parameters, which are then directly\nfed into a full mechanistic model to accurately predict consortia\ncomposition dynamics for both amensalism and competition without further\nfitting. More importantly, such quantitative understanding successfully\nidentifies the experimental conditions to achieve coexistence composition\ndynamics. These results illustrate the process of both computationally\nand experimentally building up bacteria consortia complexity and hence\nachieve robust control of such fluid systems.",
"introduction": "1 Introduction Microbes mostly live in\ncommunities to form often “invisible”\necosystems in nature. The existence of such communities is almost\nuniversal, and they have a significant impact on the “visible”\necosystems, human body, industrial complexes, and even the society. 1 − 11 Scientists have discovered that soil microbial consortium affects\nagriculture by fixing nitrogen and stimulating plant growth. 12 , 13 Human microbiome, especially human gut microbiota, plays important\nroles in human physiology, immune system, and even psychology. 1 , 7 , 14 − 17 In industrial production, a multiple-strain-based\nfermentation method has been widely used to break the limits of traditional\nsingle-strain fermentation. 18 , 19 Studying microbial\nconsortia not only improves the understanding of these influences\nbut also enables us to engineer microbial consortia as desired. However, understanding of the formation, dynamics, and behavior\nof such communities remain elusive due to difficulties in culturing,\nmeasuring, and controlling such communities. 20 − 23 Great efforts using bioinformatics\nstrategies have been made to identify and analyze the composition\nof natural microbial consortia, aiming to find out the relationships\nbetween these consortia and the corresponding environments. 10 , 24 , 25 Though informative, this strategy\nlacks the ability to establish the causality, analyze the dynamics,\nand predict the behavior of typical consortia. 26 , 27 On the contrary, synthetic biology approaches, through bottom-up\ndesigning of synthetic microbial consortia with well-defined components\nand strains, enable us to systematically study these fundamental questions\nof the microbial consortia. 28 Recently,\nscientists have been using synthetic microbial consortia to study\nthe dynamics of artificial consortia, improve fermentation yield and\nefficiency, and investigate colony patterning. 20 , 29 − 34 However, most of these studies typically focus on one ad-hoc interspecies\ninteraction using specific species, which are not expandable or systematic. 35 − 37 For example, the cross-protection mutualism method lacks the ability\nto generate negative interactions, which limits the topology of study\nto be within mutualism, commensalism, and no-interaction. 35 Here, we demonstrate a design and construction\nworkflow of microbial\nconsortia that allows us to study microbial consortia with different\ntopologies under the same framework. We combine a bacterial toxin-antitoxin\nsystem with quorum-sensing systems to construct microbial consortia\nwith different interstrain interactions. While most of the genetic\nelements are kept unchanged across all engineered systems, the strains\nin consortia have almost identical genetic and physiological backgrounds,\nenabling more accurate comparisons among different consortia constructed\nfollowing the same principle. Furthermore, with regulatory elements\nintentionally integrated into these systems, the controllability is\nimproved and hence makes predictive mathematical modeling of system\ndynamics attenable. As a proof of concept, we constructed amensalism\nand competition consortia and analyzed the population dynamics using\ncoculture experiments. Next, we calibrated the mathematical model\nand showed that most of the parameters are inheritable across different\ndesigns, suggesting the robustness and portability of genetic elements\nand modules. Finally, we mathematically analyze the consortia and\ndemonstrate the predictive power.",
"discussion": "3 Discussion Design,\nconstruction, and\ncontrol of synthetic microbial consortia\nremain challenging despite recent better understanding of microbial\nconsortia. 20 − 23 A wide range of factors, such as gene circuit topologies, intracellular\ninteractions, intercellular communications, and cell growth and competition,\nneed to be considered. In this study, we developed a design framework\nto facilitate the design and construction of E. coli-based microbial\nconsortia with common and portable genetic elements. Following the\nframework, we assembled the amensalism and competition consortia and\nshow that our synthetic consortia indeed have typical population dynamics\nthat are similar to those of analogous natural ecological systems\nwith the same topology. Furthermore, we demonstrated the feasibility\nof building up consortia complexity while keeping common parts and\nparameters constant from simple to complex systems. At last, we mathematically\nanalyzed the system and showed how we obtained a systematic understanding\nof population behavior and the prediction power of our model. Previous studies have developed methods of constructing synthetic\nmicrobial consortia. 20 , 29 − 34 Many fundamental rules for designing these systems have been developed.\nFor example, many researchers use cross-feeding auxotrophic strains,\ncross-protection against antibiotics, or niche differentiation to\nmaintain cosurvival. 35 , 36 , 69 Although current methods are capable of designing microbial consortia\nsystems for ad-hoc purposes, 30 , 35 − 37 these systems are often not tunable or expandable. In this study,\nwe move one step forward and provide a method enabling systematic\ndesign, construction, and study of synthetic microbial consortia.\nOur proposed designs can be easily induced by common drugs and easily\nextended to various topologies with parts that are widely used, well-characterized,\nand standardized. Furthermore, with more quorum-sensing systems added\nto our design, we can expand our current design to multiple strain\nconsortia with combinatorially more complex topologies. With these\nfeatures, our design principle can serve as a framework under which\nresearchers can easily generate their desired consortia and provide\ncomparable results across different studies. We also demonstrate our\nsystems are reliable with experimental and data-based theoretical\nmethods in a stepwise manner from modules to strains and then consortia.\nThrough analysis of the model, we show that we can predict the behavior\nof the system under conditions and time that are hard to reach due\nto limitations of experiment approaches. Therefore, we believe that\ncomplex systems built within our framework mimic natural counterparts,\nallowing us to construct prototype biological models and conduct experimental\nanalysis under conditions that are hard to access in natural settings."
} | 1,882 |
32325690 | PMC7231361 | pmc | 1,189 | {
"abstract": "Conductance quantization (QC) phenomena occurring in metal oxide based memristors demonstrate great potential for high-density data storage through multilevel switching, and analog synaptic weight update for effective training of the artificial neural networks. Continuous, linear and symmetrical modulation of the device conductance is a critical issue in QC behavior of memristors. In this contribution, we employ the scanning probe microscope (SPM) assisted electrode engineering strategy to control the ion migration process to construct single conductive filaments in Pt/HfO x /Pt devices. Upon deliberate tuning and evolution of the filament, 32 half integer quantized conductance states in the 16 G 0 to 0.5 G 0 range with enhanced distribution uniformity was achieved. Simulation results revealed that the numbers of the available QC states and fluctuation of the conductance at each state play an important role in determining the overall performance of the neural networks. The 32-state QC behavior of the hafnium oxide device shows improved recognition accuracy approaching 90% for handwritten digits, based on analog type operation of the multilayer perception (MLP) neural network.",
"conclusion": "4. Conclusion In this work, we demonstrate a reliable hafnium oxide based memristor device that displays homogenized conductance quantization characteristics with 32 half-integer QC states. Through deliberate design of the electrode/switching matrix interface with the assistance of a SPM tip, a single conductive filament can be generated inside the memristive layer, which suppresses the evolution randomness of the multiple CFs and enhances the uniformity of the device conductance effectively. Simulation results indicate that with the linearly and symmetrically modulated QC behavior with conductance states of 32 levels, improved pattern recognition accuracy approaching 90% can be achieved through analog type operation of the multilayer perception neural network with the present memristor devices.",
"introduction": "1. Introduction Resembling the operating principle of human brains that transmit and process information through huge amounts of interconnected neurons and synapses [ 1 , 2 , 3 , 4 ], neuromorphic computing paradigm based on new solid-state electronic devices have demonstrated advantages of high efficiency, low power consumption, and parallel processing ability when handling big-data analysis tasks [ 5 , 6 , 7 ]. Plenty of research efforts have been devoted to emulating the electrical functions of biological synapses and neurons with ferroelectric, magnetic, phase-change, and resistive switching devices [ 8 , 9 , 10 ], wherein the resistive switching memristors distinguish themselves as promising post-Moore era candidates with their simple device structure, crossbar array, and 3D stacking capability for very large scale integration [ 11 , 12 , 13 , 14 , 15 , 16 , 17 ]. In particular, the ion migration and filamentary conduction mechanism make the memristor devices extremely scalable, enabling them to easily approach the lithographic limitations [ 18 , 19 , 20 , 21 ]. It was demonstrated that the computing and energy efficiency of memristor-based in-memory computation was comparable to that of the complementary metal-oxide-semiconductor (CMOS) platforms [ 22 , 23 ]. Shrinking the dimension of the conductive filaments (CFs) into the atomic scale of quantum point contact allows memristor ballistic electron transport without scattering and quantized conductance (QC) characteristics in analog domains [ 24 , 25 , 26 ]. It not only significantly increases the data storage capacity of the devices, but also allows stronger information processing ability in neuromorphic systems. Generally, multiple conductance states of the memristor devices should be achieved in ways that are as simple as possible for practical usage [ 27 , 28 , 29 , 30 ]. Nonetheless, most of the studies until now were implemented with complex programming schemes to achieve multi-conductance characteristics, in which the involvement of varying current compliance and voltages leads to a heavy burden on the system circuit design [ 31 , 32 , 33 , 34 ]. Advancements are highly desired to simplify the operating philosophy and realize reliable and analog type conductance quantization behavior in resistive switching memristors. In this contribution, we report an effective approach to regulate the multilevel quantized conductance characteristics through pre- and customized formation of the conductive filament in a controllable manner in metal oxide based memristor devices. By inducing the directional migration of oxygen anions under the stress of a concentrated electric field through a conductive scanning probe microscopic (SPM) tip, indentation and consequently protrusion of the metal electrode into the switching matrix can cause the construction of a single conductive filament at a fixed position with better controllability. Thirty-two continuous quantized conductance states can be obtained by deliberately manipulating the as-formed CF, which gives rise to ≈20% enhancement in the uniformity of conductance value distribution in each QC state. More importantly, enhanced recognition accuracy approaching 90% can be achieved by multi-layer perception, employing HfO x memristor with homogenized, analog type conductance quantization for both handwritten digit patterns.",
"discussion": "3. Results and Discussion Variation of the switching parameters, including the programming voltages and device resistances in different QC states, is the main cause of performance deterioration in terms of operating reliability and recognition accuracy of the memristor based neural networks. It can be ascribed to the random ion migration in the polycrystalline metal oxide switching matrix, and consequently the stochastic nature of the branch-shaped multiple conductive filament formation, disruption, and regeneration during cyclic operations [ 35 , 36 ]. In order to achieve a more stable and adjustable conductance state for memristor devices, we performed pre-treatment of the HfO x switching layer using the scanning probe microscope based electrochemical lithography technique [ 37 , 38 ]. By stressing voltages onto the hafnium oxide layer through a conductive SPM tip, the highly localized electric field formed under the tip pinpoint region can induce directional migration of the mobile oxygen vacancies towards the tip position, resulting in loss of the local mass, and therefore formation of a concave structured indentation. As demonstrated in our previous study [ 38 ], capping the top Pt electrode formed a downward pointing metal protrusion into the switching layer, which acted as a microelectrode that concentrated the internal electric field distribution and led to the formation of a single conductive filament in the memristor device. The evolution of the as-formed CF is more controllable, so the reliability of the device can be greatly improved. As shown in Figure 2 a, scanning the pre-treated Pt/HfO x /Pt device between −1.5 V and +2 V with a current compliance preset of 10 mA can produce resistive switching characteristics, showing promising cycling uniformity. For over hundreds of switching cycles, both the ON and OFF state resistances were distributed in a narrow range ( Figure 2 b). A high ON/OFF ratio exceeding 10 3 can be maintained reliably, allowing a wide regulation window for the achievement of conductance quantization, with sufficient resolution for differentiating the adjacent QC states for multilevel operations. Theoretically, the ON and OFF device resistances of 35,000 Ω and 300 Ω corresponds to the conductance modulation range of 0.5 G 0 –50 G 0 , which is even broader than that used in synaptic weight updating of the reported memristive neural networks [ 39 , 40 ]. Herein, G 0 is the integer unit of quantized conductance, 77.5 μS. Similar pre-treatment can be performed with the nano-imprint lithography (NIL) technique during large scale fabrications. Conductance quantization can be realized by either controlling the current compliance in the set processes or changing the cut-off voltages of the reset processed [ 41 , 42 ]. Since the positive feedback of the set procedure usually led to uncontrolled overgrowth of the conductive filament with overshooting device conductance, or even the absence of the QC states [ 37 ], we employed a relatively more moderate reset process to modulate the evolution of the conductive filament. As depicted in Figure 2 c, resetting the Pt/HfO x /Pt device with increasing stopping voltages of −0.6 V to −1.6 V during direct current (DC) scanning can consecutively decrease the device currents. Replotting in the conductance vs. number of scanning curve, or the conductance vs. number of pulse stressing curve, reveals a continuous modulation of device conductance from 16 G 0 to 0.5 G 0 in a half-integer QC step ( Figure 2 d). The pulse-mode measurement was conducted by applying voltage pulses with the width of 10 ms and increasing amplitudes from −0.6 V to −1.6 V. For both the DC scanning and pulse stressing operations, the ramping steps of the voltages are −0.02 V in the −0.6 V to −0.84 V range, with an increase to −0.04 V in the −0.84 V to −1.6 V range. This, nevertheless, is consistent with the negative feedback characteristics of the reset process, wherein the shrinking of the conductive filament dimension with smaller device currents will slow the Joule heating related modulating process. In total, 32 quantized conductance states were obtained in a stepwise manner in the DC scanning or pulse stressing mode. When the conductance goes beyond this range, the conductive filament becomes so thick that significant Joule heating with large device currents can annihilate the CF easily, resulting in a mutant reset process with partial absence of the QC states with larger conductance values. On the other hand, as the device conductance decreases to less than 1 G 0 , the atomic point contact gets completely disconnected to the metal electrodes, and the quantum conductance effect does not exist any longer. Nevertheless, it is noteworthy that the one directional update of the synaptic weight in the reset process may lead to additional complexity in the operating methodology or circuit design during practical applications. For instance, potentiation of the synaptic weight can be only achieved by a combination operation of binary switching of the device to a high conductance state, and subsequent depression to the desired level in the negatively biased reset process [ 38 , 39 ]. Therefore, symmetrical conductance modulation and fast blind weight updating are more desired in warranting the computing efficiency of the neural network. Although the present electrode engineering strategy does not guarantee that each step of the modulation will strictly undergo a 0.5 G 0 change during the cyclic operations, linear evolution of the device conductance benefiting the synaptic weight update for training the neural network can still be received, as shown in Figure 3 a. The device conductance was modulated from 16 G 0 to 0.5 G 0 continuously in pulse-mode operation and set by the positively biased voltage scanning from 0 V to 1 V with a current compliance of 5 mA to reprogram the Pt/HfO x /Pt memristor to ON state. Afterward, the negative voltage pulse modulation was reconducted for a total of 12 times to give the data plotted in Figure 3 a. All of the 32 QC states can be obtained repeatedly, however, the nominal conductance values varied in a small range (e.g., 33.26% for 16 G 0 ) during cyclic modulation. For control sample B, without pre-treatment of the HfO x layer or top electrode protrusion, a linear but more discrete distribution of the device conductance was observed at each stage of the pulse-mode modulations ( Figure 3 b). The variation of sample B’s nominal conductance approached 55.95% for 16 G 0 . Therefore, it was more likely to reach the desired quantized conductance within the pre-treated devices, wherein the formation of protruding microelectrodes can effectively regulate the evolution of a single conductive filament (rather than multiple CFs) in a more homogenized manner. Increasing the sampling number of both kind of devices can reveal this feature more obviously. For instance, at the fifth pulse modulation state, the FWHM (full wave at half maximum) of the Gaussian curve fitting of device A’s conductance distribution around the nominal value, 14 G 0 , was 4.4 ( Figure 3 c); whereas the respective number increased to 5.5 for sample B ( Figure 3 d), suggesting that a 25% enhancement in the uniformity of the conductance value can be made possible by the present SPM pre-treatment method. Similarly, the statistics collected at the 29th pulse modulation state (for conductance of 2 G 0 ) shows FWHMs of 1.2 for sample A, and a much larger value of 2.3 for sample B. Further analysis of the collected data for both samples A and B with the normal distribution function (NDF) allows us to receive an equation,\n (1) G = ( 1 − k n ) × f ( μ , σ 2 ) , \nto mathematically predict the device conductance at certain pulse modulation stages, where G is the varying device conductance, n represents the number of the pulses applied, k , σ , and μ reflect the variation range, variability, and standard value of the device conductance at each modulation state, and f ( μ , σ 2 ) is a general number generator that obeys the NDF’s law. Upon setting k = 0.03, σ = 1, and μ = 16.5 − 0.5 n for sample A, and k = 0.03, σ = 4, and μ = 16.5 − 0.5 n for sample B, faithful reproduction of the device conductance for 12 continuous pulse modulation cycles can be obtained through the mathematical simulation shown in Figure 4 . All the simulated datasets fall in the experimental range depicted in Figure 3 a,b, suggesting that modeling with the above equation and parameters can well resemble evolution processes of the Pt/HfO x /Pt memristor devices, with either microelectrode protrusion or flattening of electrode/oxide interfaces. As such, considering the deviation of the atomic point contact’s composition and geometry from pure metallic hafnium-based cones, which is probably the case during device operation by atomic exchange with the surrounding HfO x matrix under the concentration gradient, fractional conductance quantization can also be modelled accordingly. This may offer additional guidance for deliberate tuning of the device electrical performance to receive more (e.g., 64) levels of linearly and symmetrically modulated conductance, which can lead to analog type operation and enhanced accuracy for pattern recognition with the memristive neural networks. For demonstration, simulation of supervised learning consisted of the offline training of the multilayer perception (MLP) neural network, synaptic weight updating in the memristor array, and tests of the handwritten digit pattern recognition, and was carried out using the experimental quantized conductance characteristics with back propagation (BP) algorithm ( Figure 5 a). Seven hundred and eighty four 28 × 28 pixel images of handwritten digits are obtained from the Keras database and utilized for training and recognition tests [ 43 ]. Accordingly, the as-constructed MLP network contained 784 input neurons, two-layer arrangement of 16 × 16 (256) hidden neurons, and 10 output neurons, with respect to the digits of 0 to 9 ( Figure 5 b). During the training courses, quantized conductance characteristics with 4, 16, 32, and 64 states were employed to renew the synaptic weights. Actual conductance values were considered and transferred to the synapse array upon normalization. Figure 5 c displays the variation of recognition accuracy, along with the increased training epochs and numbers, of the available QC states. As shown, the less effective weight updating of the MLP network with four levels of QC characteristics can only lead to ≈10% recognition of the handwritten digits, while increasing the numbers of the available QC states to 16 may improve the accuracy to 71.2%. Nevertheless, it was observed that as the training courses continue, the recognition rate suddenly drops to ≈10% after six epochs. This can be ascribed to the fact that small numbers of the available QC states may result in lesser amounts of conductance levels available for weight updating and thus larger learning gradients and undesired convergence of the training loss function at the local minimum, which in turn makes the neural network unable to perform the multiplication-and-accumulation (MAC) operations any longer. As a consequence, the training process fails. When the numbers of the QC states reach 32 and 64, the analog manner of the MLP network operation shows promising pattern recognition accuracy of 86.8% and 93.5% after 10 training epochs, respectively, again suggesting the numbers of conductance states will directly affect the final recognition accuracy of the neural network. Only when the conductance levels are higher than 32, the accuracy of the QC based MLP network can be steadily improved with the number of trainings. It was also noteworthy that in the present study the employment of 1-bit operation technically limits the overall computing accuracy of the network. In the case that multi-bit inputs can be projected onto multiple memristor cells, greatly enhanced accuracy can be possible [ 44 , 45 , 46 ]. To further confirm the influence of electrode engineering on the recognition accuracy of the MLP network, we use the 32-state QC characteristics of samples A and B for the supervised learning simulation. The normal distribution formula was adopted to simulate the conductance value fluctuation during the synaptic weight updating procedure. As shown in Figure 5 d, the black curve plotted with idealized conductance values at each QC state ( σ = 0) displays a recognition accuracy of ≈86.1% for handwritten digits. Minor deviation from the ideality does not affect the performance of the network significantly, and the red curve simulated with the memristive characteristics of sample A gives a comparable recognition rate of ≈86.0%. For the case of sample B without electrode engineering, although the accuracy rises rapidly as the training course continues (blue curve), there was still an obvious performance gap of ≈2% after 10 epochs, when in comparison with the network constructed from the ideal or sample A memristor devices. Therefore, the fluctuation occurring during device conductance modulation, beyond the promising linearity in synaptic weight updating, plays an important role in improving the overall performance of the memristive neural networks. Reduced fluctuation of the switching parameters can project the trained network onto the memristor array with higher learning accuracy, which in turn guarantees the fidelity of MAC operation and BP algorithm more reliably."
} | 4,740 |
40264852 | PMC12012298 | pmc | 1,190 | {
"abstract": "In recent years, the rapid progression of artificial intelligence and the Internet of Things has led to a significant increase in the demand for advanced computing capabilities and more robust data storage solutions. In light of these challenges, neuromorphic computing, inspired by human brain’s architecture and operation principle, has surfaced as a promising answer to the growing technological demands. This novel methodology emulates the biological synaptic mechanisms for information processing, enabling efficient data transmission and computation at the identical position. Two-dimensional (2D) materials, distinguished by their atomic thickness and tunable physical properties, exhibit substantial potential in emulating synaptic plasticity and find broad applications in neuromorphic computing. With respect to device architecture, memory devices based on floating-gate (FG) structures demonstrate robust data retention capabilities and have been widely used in the realm of flash memory. This review begins with a succinct introduction to 2D materials and FG transistors, followed by an in-depth discussion on remarkable research progress in the integration of 2D materials with FG transistors for applications in neuromorphic computing and memory. This paper offers a thorough review of the existing research landscape, encapsulating the notable progress in swiftly expanding field. In conclusion, it addresses the constraints encountered by FG transistors using 2D materials and delineates potential future trajectories for investigation and innovation within this area.",
"introduction": "Introduction Utilizing metal–oxide–semiconductor field-effect transistors (MOSFETs) as a critical component within the framework of integrated circuits, computer technology has firmly established a foundation for the contemporary-information-driven society. This has led to groundbreaking advancements across a multitude of technological domains. Nevertheless, the distinct separation of the memory’s physical structure and the central processing unit (CPU) necessitates that computers execute tasks in a linear sequence. When handling extensive amounts of data, the information transfer between the processor and the memory results in substantial power consumption and decreased computational speed, which substantially impedes the computational capability (i.e., the von Neumann bottleneck) (Fig. 1 A) [ 1 – 3 ]. To augment the performance of traditional computers, it is imperative to integrate a greater number of processors and memories into a single chip within a specified area or volume. However, the scaling challenges associated with processors and memories made from silicon transistors are becoming increasingly pronounced because of the constraints imposed by Moore’s law. To realize reduced power consumption and enhanced data processing efficiency, it is imperative to design novel computing architectures leveraging innovative materials and fundamental principles. Fig. 1. Contrast between the von Neumann architecture and neuromorphic computing. (A) von Neumann architecture. (B) A diagram illustrating the fundamental organizational principles of the brain [ 8 ]. (C) A state-of-the-art Si computing system [ 8 ]. (D) Neuromorphic computing. Neuromorphic computing, inspired by the human brain’s capacity for parallel computation and adaptive learning, has been raised as a solution to the inherent von Neumann bottleneck in traditional computing frameworks, leading to a series of technology breakthroughs [ 4 ]. The human brain’s nervous system, composed of approximately 10 10 to 10 12 neurons, is intricately linked through synapses with each connected to hundreds of others, thereby enabling the transmission and reception of signals. The brain, with a power consumption of only ~20 W, is responsible for categorizing stimuli, predicting outcomes, generating thoughts, and ensuring vital life functions [ 5 ]. Conversely, although contemporary digital computers are proficient at executing high-precision calculations, they necessitate substantial energy consumption when performing cognitive tasks, which are the domain of the brain’s expertise. For instance, the energy required to train the most advanced natural language processing models on contemporary supercomputers amounts to 1,000 kWh of energy [ 6 ], equivalent to the energy that the human brain would expend over 6 years performing all tasks. While our understanding of the brain remains incomplete, its remarkable capabilities can be ascribed to 3 key aspects explored in neuroscience: the vast connectivity, the organized functional hierarchy, and the temporally dependent functions of neurons and synapses (Fig. 1 B) [ 7 ]. Contemporary deep learning networks are fundamentally inspired by the brain’s hierarchical structure and synaptic framework, comprising multiple layers or transformations, each representing distinct latent features within the input. The operation of these neural networks is facilitated by hardware computing systems, fundamentally reliant on silicon-based FETs. The digital logic used for large-scale computing is composed of over billions of transistors, all integrated onto a single silicon chip. Various stages of silicon-based transistor computation are strategically layered to optimize data transfer patterns (Fig. 1 C) [ 8 ]. Although this computing model bears a superficial resemblance to neuromorphic computing, it fundamentally operates on the conventional von Neumann architecture. The physical separation between processing units and the memory devices intensifies the “memory wall bottleneck” [ 9 ], leading to inefficiencies and increased power consumption due to the CPU’s idle waiting for data from slower memory components. Traditional software or hardware strategies to realizing neuromorphic computing have challenges of high power consumption, low integration density, and insufficient reliability, which urgently necessitates to further advance neuromorphic/brain-like computing with new materials or architectures. In 1990, Mead [ 10 ] first used the term “neuromorphic” to characterize the transmission and processing of neural information in high efficiency, which differs from traditional digital computing system. Neuromorphic computing simulates the information processing behaviors in biological synapses (core physical mechanism: ion migration through synapses), enabling efficient information transmission and processing at the same position. As shown in Fig. 1 D, neuromorphic computing enables dual functions of information storage and computation at the same position, eliminating the need for additional data transmission. Its advantages include lower energy dissipation (in nanojoules) and faster processing speed (in nanoseconds) [ 11 ]. Moreover, neuromorphic computing features distributed memory and computational components [ 12 ], which can facilitate efficient, large-scale parallel computing to overcome the von Neumann bottleneck. To propel brain-like neuromorphic computing through material innovation, researchers have identified 2-dimensional (2D) materials as a pivotal candidate. These materials exhibit exceptional potential due to their atomic thickness, clean surfaces devoid of dangling bonds, unique physical properties, electrical tunability [ 13 – 15 ], and high integrability. Such intrinsic advantages position 2D materials as key feasible solutions for neuromorphic computing. In addition, the inherent high surface-area-to-volume ratio, along with sensitivity to charge transfer and electrostatic modulation at interfaces, renders 2D materials an optimal platform for fabricating synaptic devices with reliable operational principles [ 16 , 17 ]. In terms of device architecture, floating-gate (FG) transistors, commonly used as the basic structure in flash memory devices such as NOR and NAND flash [ 18 , 19 ], have attracted attention due to their advantages in miniaturization, low energy consumption, and high-efficiency data storage, which effectively address the challenges related to high-density integration and large-capacity data processing. Meanwhile, FG memory devices based on the principle of charge storage have the potential to emulate synaptic plasticity, potentially breaking through the von Neumann computing architecture and serving as artificial synaptic units inspired by the human brain [ 20 – 29 ]. Traditional FG memory is based on the MOSFET structure, where the storage function is facilitated by incorporating an FG layer to capture and release carriers [ 30 – 33 ]. When a proper voltage is applied to the gate of the FG transistor, charge injection/removal occurs in the FG layer, known as programming/erasing operations. On the basis of the charge capture/release process in the FG layer, nonvolatile memory behavior can be achieved even in the absence of applied electric field [ 34 ]. However, with increased process nodes in computer and decreased transistor size, traditional FG memory devices face issues such as increased leakage current, serious charge leakage, and reduced reliability [ 35 ]. For instance, when the thickness of the tunnel oxide layer in FG devices decreases, the stored charges in the FG layer are more prone to leakage, resulting in data instability and reduced memory reliability [ 36 ]. Given the atomic-level thickness, high electron mobility, and tunability of 2D materials [ 37 – 44 ], FG transistors incorporating 2D materials exhibit significance potential for neuromorphic computing and advanced memory applications [ 45 – 51 ]. Here, we present a comprehensive overview on the integration of 2D materials with FG transistors for neuromorphic computing and memory applications. To facilitate a deeper understanding of their architectures, we first briefly introduce commonly used 2D materials, including graphene, transition metal dichalcogenides (TMDCs), black phosphorus (BP), and hexagonal boron nitride (h-BN), and analyze their structural differences and unique physicochemical properties such as electron mobility, bandgap, flexibility, and mechanical characteristics. The potential application of these materials in FG transistors is also explored. Then, we introduce the charge tunneling mechanism of FG transistors and summarize the development of FG transistors classified by 5 common structures. In addition, we systematically review various 2D FG devices and their applications in memory devices and neuromorphic computation in terms of materials, devices, and applications (Fig. 2 ). Finally, we raise the issues of 2D FG transistors related to material preparation, device structure design, device stability control, and integration and outline the potential development trends for exploration and innovation within this domain. Fig. 2. An overview of FG devices integrating 2D materials and their neuromorphic computing applications. MLG, multilayer graphene; PSC, postsynaptic current; BG, bottom gate; FLG, few-layer graphene."
} | 2,744 |
34603732 | PMC8477758 | pmc | 1,192 | {
"abstract": "Porous membranes with special wetting properties have attracted great interest due to their various functions and wide applications, including water filtration, selective oil/water separation and oil skimming. Special wetting properties such as superhydrophobicity can be achieved by controlling the surface chemistry as well as the surface topography of a substrate. Three-dimensional (3D) printing is a promising method for the fast and easy generation of various structures. The most common method for 3D printing of superhydrophobic materials is a two-step fabrication process: 3D printing of user-defined topographies, such as surface structures or bulk porosity, followed by a chemical post-processing with low-surface energy chemicals such as fluorinated silanes. Another common method is using a hydrophobic polymer ink to print intricate surface structures. However, the resolution of most common printers is not sufficient to produce nano-/microstructured textures, moreover, the resulting delicate surface micro- or nanostructures are very prone to abrasion. Herein, we report a simple approach for 3D printing of superhydrophobic micro-/nanoporous membranes in a single step, combining the required topography and chemistry. The bulk porosity of this material, which we term “Fluoropor”, makes it insensitive to abrasion. To achieve this, a photocurable fluorinated resin is mixed with a porogen mixture and 3D printed using a stereolithography (SLA) printing process. This way, micro-/nanoporous membranes with superhydrophobic properties with static contact angles of 164° are fabricated. The pore size of the membranes can be adjusted from 30 nm to 300 nm by only changing the porogen ratio in the mixture. We show the applicability of the printed membranes for oil/water separation and the formation of Salvinia layers which are of great interest for drag reduction in maritime transportation and fouling prevention.",
"conclusion": "Conclusion In summary, we have shown a convenient methodology to produce porous membranes in a highly fluorinated polymer with adjustable submicron porosity via 3D printing using a commercially available SLA printer. The pore size can be simply adjusted by varying the amount of the non-solvent in the polymer mixture. The as-printed membranes manifest hydrophobic wetting properties with static contact angles in the range of 126°. Due to their bulk porosity and hydrophobicity, the printed membranes with an average pore size above 100 nm achieved an excellent oil–water separation efficiency of over 99% for chloroform/water mixtures. Besides, thin superhydrophobic membranes with adjustable porosity were successfully fabricated by peeling thin layers from a bulk print with a special design that we developed to facilitate layer separation. The peeled membranes showed superhydrophobic wetting properties with a static contact angle in the range of 164°. While submerging the thin superhydrophobic membranes in water, a Salvinia layer was formed trapping an air film between the micro-/nanostructure of the membrane and water. The presented membrane fabrication enables the facile generation of customized membranes for applications such as, e.g. , oil/water separation. Through their porous nature these superhydrophobic thin membranes can be potentially used for the stabilization and regeneration of a Salvinia layers in dynamic modus.",
"introduction": "Introduction Porous membranes with special wetting properties, such as superhydrophobicity have lately attracted considerable attention. These surfaces are described by a static water contact angle greater than 150°, a roll-off angle smaller than 10° and are widely used for anti-fouling coatings, 1 anti-icing surfaces 2 or oil/water separation. 3,4 Manufacturing of such membranes requires structuring of surfaces on a nano- to microscale. The micro-/nanostructure or roughness promotes the formation of residual air pockets ( Salvinia layer, i.e. Cassie wetting state), that cause the decrease in the roll-off angle, and adhesive force. 5–9 Unfortunately, this delicate surface structuring makes most superhydrophobic surfaces very sensitive to abrasion and renders them impractical for real-life applications. The most common process for producing such structures is a top-down approach, where defined micro-/nano-scale surface topographies are coated with a low-surface energy chemical such as fluorinated silanes. 10–12 In practice, often porous supports, typically mesh substrates are coated by hydrophobic materials. 13,14 Common problems associated with this method are the poor adhesion between the mesh and the built-up hydrophobic layer leading to delamination and abrasion, and the clogging of the mesh pores during the post chemical modification of the surface. Another fabrication strategy is bottom-up structuring of a material with low surface energy using processes such as, e.g. , chemical vapor deposition, sol–gel methods and layer-by-layer deposition. 15–19 The most common class of materials used are fluorinated polymers due to their unique chemical and physical properties such as low polarizability, low adhesion as well as high chemical resilience, e.g. , against corrosive liquids. Moreover, fluorine-rich polymers possess low surface energies, which endow their surfaces with excellent water and oil repellency. 20–22 Due to its versatility and ease in generating complex shapes with geometrical features, three-dimensional printing (3D printing) has emerged as a promising technology to prepare porous structures. 23–29 However, due to the limitations in resolution and adjustability of pore size, the fabrication of micro/nanostructured membranes via 3D printing remains challenging. In 3D printing processes, top-down approaches are commonly employed, where the patterned geometry is printed and the resultant surface structure is then chemically modified with low surface energy compounds. Fused deposition modelling (FDM), 30 selective laser sintering (SLS) 31 and direct inkjet writing (DIW) 32 are the most common methods for preparing superhydrophobic porous membranes. Recently, Lv et al. reported the fabrication of superhydrophobic porous membranes for oil/water separation by 3D printing an ordered mesh structure using an ink of polydimethylsiloxane (PDMS) and hydrophobic silica particles with pore sizes in the sub-millimeter range. 32 Yuan et al. reported the fabrication of polysulfone membranes via SLS and their application for oil/water separation. The printed porous membranes were subsequently coated with candle soot to achieve superhydrophobicity. 33 They also reported the fabrication of a micro-/nanostructural surface by designing a unique 3D multiscale zeolitic imidazolate framework (ZIF-L) on a 3D printed membrane for a superhydrophobic and underwater superoleophobic surface. 34 However, these reported methods enable only the formation of large pores in the micrometer range, lacking the required combined micro-/nanostructure and still requiring additional surface modifications to achieve superhydrophobicity. Besides, the pore size plays an important role in membrane applications in various fields from separation, adsorption to tissue engineering. 14,31–33,35–38 Thus, a method for adjusting the porous structure is of high interest. Polymerization induced phase separation (PIPS) has emerged as a promising method to address these limitations as it offers the possibility to create materials with inherent, self-organized porous structure with adjustable porosity. 20,39–42 As of late, it has been effectively applied in the fabrication of micro/-nanoporous materials, such as glass and polymers, using SLA. 43–47 In this paper, we report the SLA 3D printing of thin superhydrophobic membranes (100–400 μm) with adjustable porosity in the submicron range with 30 nm to 300 nm pores. The process combines both, required topography and surface chemistry for achieving superhydrophobicity in a single step. For achieving this, a photocurable fluorinated resin is mixed with a non-solvent to create a micro-/nanostructure throughout the bulk by phase separation during polymerization. The resin is then printed into a special “staircase” design, which with help of the layer-by-layer nature of the SLA printing process, enables exposing the micro-/nanostructure of the membranes by peeling off thin layers from the bulk. Superhydrophobic membranes with a static water contact angle of 164° are thus obtained with no need of further surface modifications. The printed membranes were successfully applied as oil/water separators, oil absorbers/skimmers and for the detection and potential regeneration of Salvinia layers.",
"discussion": "Results and discussion Fabrication of directly printed membranes for oil/water separation For direct 3D printing of micro-/nanostructured membranes, two different mixtures for generating Fluoropor fluorinated polymer foams were formulated. With these resin mixtures, a micro-/nanostructure can be achieved throughout the printed material, due to a phase-separation of the resin during the printing process. This way, a much smaller pore size of the membranes can be achieved compared to the feature resolution commonly achieved by SLA. By introducing different amounts of non-solvent and stabilizing the monomer/non-solvent mixture by an emulsifying agent, two stable printing resins mixtures were generated, termed Fluoropor 15 and Fluoropor 25 (see Table 1 ). The required amounts of the reactants for the preparation of the resin mixtures Fluoropor 15 resin mix Fluoropor 25 resin mix Monomer MD700 (wt%) 50 50 Emulsifying agent (wt%) 35 25 Non-solvent (wt%) 15 25 Initiator a (mg ml −1 ) 7 7 Absorber a (mg ml −1 ) 3.3 3.3 a Concentration in Fluoropor mix. Using these resins, an SLA printing process was used to produce custom-shaped membranes. Using the Asiga Pico 2, membranes disc of 35 mm diameter, consisting of several printed layers with an overall thickness of 500 μm were printed. These “membrane stacks” possess different pore volumes, as indicated by the different optical appearance of the membranes, ranging from translucent in case of Fluoropor 15 to white and non-transparent in case of Fluoropor 25 ( Fig. 2 ). A higher content of non-solvent generates a larger overall pore volume that causes the material's appearance to turn from translucent to white and non-transparent. Due to the removal of porogens and material de-swelling, the production process was accompanied by a linear shrinkage of 20% and 14% in the x – y direction, 32% and 26% in the z direction for Fluoropor 15 and Fluoropor 25, respectively. The wetting properties of the membranes were investigated by contact angle measurements. Fluoropor 15 and Fluoropor 25 show a static contact angle of 123 ± 2 °and 126 ± 4°, respectively (see ESI Fig. S1 † ). Fig. 2 3D printed disc shaped membrane stacks with a total thickness of 500 μm after printing and after porogen removal. (A) 3D printed Fluoropor 15 membrane stack; (left) directly after printing and (right) shrunken dried membrane stack after porogen removal. (B) 3D printed Fluoropor 25 membrane stack; (left) directly after printing and (right) shrunken dried membrane stack after porogen removal. To investigate the structure of the membranes, SEM measurements of the cross-section as well as the top and bottom surface were taken. SEM images of the top- and bottom surfaces of both membranes are shown in Fig. 3 . As can be seen, the micro-/nanostructure is not visible on the surfaces but clearly visible in the cross-sections of the membranes (see Fig. 3 ). In the given SLA setup, the printing process is carried out in a constrained volume between a transparent window (through which the DLP illuminates the build volume) and the build platform. Therefore, the lack of porosity on the surface is a result of the direct contact of the printing resin with the build platform (bottom surface) and the transparent window (top surface). The porosity within the bulk of the material is shown in the cross-section in z -direction (see Fig. 3E and F ). With an increased amount of the non-solvent, the phase separation occurs faster enabling the formation of thicker polymer networks, and vice versa . Fig. 3 SEM images of the top surface ( i.e. , the surface in contact with the printer's transparent window) and bottom surface ( i.e. , the surface in contact with the printer's build platform) of Fluoropor 15 and 25 membranes and the cross-sections. (A) Bottom surface of Fluoropor 15 membrane, (B) top surface of Fluoropor 15 membrane, (C) bottom surface of Fluoropor 25 membrane, (D) top surface of Fluoropor 25 membrane. (E) Cross-section and close-up of the cross-section of Fluoropor 15 membrane, showing the porosity of the material, (F) cross-section and close-up of Fluoropor 25 membrane. The porosity of both membranes was measured from the cross-sectional SEM images via image analysis by using a thresholding function and analysis of the pore's diameters using Image J (see ESI Fig. S2 † ). The pore size distribution is shown in Fig. 4 . Fluoropor 15 membrane displays a median pore size of 30 nm whereas Fluoropor 25 membrane displays a median pore size of 300 nm. Fig. 4 Plotted pore size distribution of the 3D printed Fluoropor membranes from the cross-sectional SEM images. (A) Fluoropor 15, (B) Fluoropor 25. Oil/water separation To investigate, whether the printed membranes are suitable for oil/water separation the membranes were first tested for their capacity to absorb organic solvents (see Fig. 5 ). Droplets of chloroform and cyclohexane were readily absorbed by Fluoropor 25 membrane, whereas a water droplet was repelled as expected. To test the absorption capacity of both membranes, droplets of different oils and a droplet of water were placed on Fluoropor 15 and Fluoropor 25. Due to its larger pore fraction, Fluoropor 25 absorbed both chloroform and cyclohexane directly, whereas the Fluoropor 15 membrane only partially absorbed the fluids. Thus, for all further separation experiments, Fluoropor 25 membranes were chosen. Disc-shaped membranes with a 45 mm diameter were 3D printed using the Phrozen shuffle 4K and placed into a filter funnel. A mixture of chloroform and water or cyclohexane and water (1 : 1, v/v) was poured into the funnel ( Fig. 5B ). A low vacuum pressure was applied to the setup to achieve the separation. The separation efficiency of a chloroform–water mixture and a cyclohexane–water mixture using Fluoropor 25 membrane was over 99 ± 1% and 79 ± 2%, respectively (see Fig. 5C ). Thus, an efficient separation of chloroform/water mixtures was achieved using the Fluoropor 25 membrane. In addition, the printed membrane can be easily cleaned with 2-propanol or acetone, dried and reused. The membrane stability and efficiency was tested up to 5 cycles with a small reduction of the efficiency within the error of the measurement: separation efficiency for chloroform of 99 ± 1% in cycle 1 and 98 ± 1% in cycle 5 and for cyclohexane of 79 ± 2% in cycle 1 and 74.5 ± 4% in cycle 5 was measured (see Fig. 5D ). The collected separated oils were tested for residual water content using an anhydrous copper sulfate color test. Both collected oils did not show any traces of water within the detection limit of 0.01 vol% (see ESI Fig. S3 † ). Fig. 5 Liquid absorption by Fluoropor membranes. (A) Photograph of water droplet and different oil droplets on Fluoropor 15 membrane (left) and on Fluoropor 25 membrane (right). (B) The customized lab set-up for the oil/water separation experiment, showing the metallic filter funnel and supporting mesh. (C) The separation efficiency of the Fluoropor 25 membrane for mixtures of water–chloroform and water–cyclohexane. (D) Separation efficiency over 5 separation cycles for water–chloroform and water–cyclohexane mixtures. The membrane was washed after each cycle and reuse. The data show small reduction of efficiency within the error of the measurement (E) absorption of crude oil drop in water. The crude oil was totally absorbed after 3 hours using Fluoropor 25 membrane. The Fluoropor 25 membrane was also tested for its capacity to absorb crude oil. The membrane was immersed in a vial containing water with crude oil floating on the surface. The membrane started absorbing the crude oil right after being immersed in the vial and successfully absorbed all of it after 3 h (see Fig. 5E ). This indicates that Fluoropor 25 membrane can potentially be used for oily wastewater treatments by absorbing the oil spills on the water surface. Fabrication of thin superhydrophobic membranes Via the direct printing process, only membranes with thicknesses higher than 200 μm could be achieved. Due to the shrinkage during drying, thinner membranes were not self-supporting and lost their porosity during the drying process. Thus, another approach was developed, to enable the formation of thin membranes. The layer-by-layer nature of the SLA printing process allows for peeling off individual layers from a bulk printed block, thus generating thin individual membranes. A special printing design was introduced, which facilitates the pulling of individual layers by a slight offset (staircase design) of the printed layers ( Fig. 6A and B ). By printing Fluoropor 15 and Fluoropor 25 stacks using this design, individual thin membranes of 100 μm and multiple membrane layers of 400 μm thickness could be peeled from the bulk of Fluoropor 15 and Fluoropor 25 block, respectively, after the drying process ( Fig. 6C and D ). This process enables the production of 20 membranes in 10 min by peeling individual membranes from a 5 mm stack. Contact angle measurements of the peeled membranes showed that the membranes have superhydrophobic wetting properties with static contact angles of 161 ± 2° and 164 ± 7° for Fluoropor 15 and Fluoropor 25 membranes, respectively (see ESI Fig. S4 † ). A water droplet is thus fully repelled on the membranes (see Fig. 6C and D ). Fig. 6 The 3D printed micro-/nanoporous Fluoropor stacks for membrane preparation. (A) Printed block of Fluoropor 15, (B) printed block of Fluoropor 25. (C) and (D) Peeled-off thin layers of a Fluoropor 15 membrane (C) and Fluoropor 25 membrane (D). A water droplet (dyed blue) is fully repelled by the membrane, showing a static contact angle of 161° for a Fluoropor 15 membrane and 164° for a Fluoropor 25 membrane. The superhydrophobic properties suggest an exposed micro-/nanostructure on the peeled membrane's surface. To analyze the bulk structure of the printed block, SEM images of the cross section of several printed layers were taken (see Fig. 7A and B ), showing the desired micro-/nanostructure throughout the bulk material. To investigate the peeled membrane top and bottom structure, further SEM images were taken ( Fig. 7C–F ), clearly showing the micro-/nanostructure of the material. A slight increase in the polymer network features sizes are observed for Fluoropor 25. Fig. 7 SEM images of the bulk cross-sections and the peeled Fluoropor membranes' top and bottom surfaces. (A) Fluoropor 15 cross-section and close-up, (B) Fluoropor 25 cross-section and close-up. The membranes show bulk porosity as well as on the surface, rendering them superhydrophobic. (C) Fluoropor 15 membrane top surface, (D) Fluoropor 15 membrane bottom surface, (E) Fluoropor 25 membrane top surface, (F) Fluoropor 25 membrane bottom surface. Mechanical stability of the thin superhydrophobic membranes The mechanical stability of the thin superhydrophobic membranes was evaluated by applying mechanical stress (stretching and bending tests). A stress–strain test was performed to evaluate the maximal strain for both membranes before breaking. For Fluoropor 15 membranes and for Fluropor 25 membranes the maximum strain is 22% and 32% respectively (see ESI Fig. S5 † ). To assess the influence of mechanical stress on the performance of the membranes, the static contact angle of water on the membranes was measured after 50, 100, 150, 300 and 500 stretching and bending cycles. Both membranes retain their superhydrophobic properties after 500 bending and stretching–releasing cycles with a strain of 20% for Fluoropor 15 and 25% for Fluoropor 25 membranes (see Fig. 8 ). This high stability is a promising result for using such membranes under real-world conditions. Fig. 8 Mechanical stability tests of the peeled superhydrophobic membranes under stretching and bending. (A) Static water contact angles (WCA) of the membranes after 500 stretching–releasing cycles show an unaffected membrane performance, with the membranes retaining a water contact angle of >162°. (B) WCAs of the membranes after 500 bending cycles show an unaffected membrane performance, with the membranes retaining a water contact angle of >163°. Detection of the Salvinia layer Thin, superhydrophobic membranes can be exploited for generating a stable air layer under water ( Salvinia effect) under static conditions. This air retention capability is technologically interesting as the retained air layer reduces drag force and prevents biofouling. To test the formation of a Salvinia layer on the membranes, a peeled Fluoropor 15 membrane, which exposes the micro/-nanostructure and a directly printed Fluoropor 15 membrane, which doesn't expose the micro/-nanostructure were glued on a glass slide. When the membranes were immersed under water, an air film was trapped in the micro-/nanostructure of the membrane forming a Salvinia layer, which was recognized by the silvery reflective layer on the peeled superhydrophobic membrane. In contrast, no change was noticed on the directly printed hydrophobic membrane when submerged under water (see Fig. 9A ). The volume of the different membranes was calculated by the Archimedes principle (see Fig. 9B ). The volume of the retained air layer ( Salvinia layer) varies in dependence of the surface area of the membrane in a near linear increase. The larger the surface area of the membrane is the higher the trapped air volume (see Fig. 9C ). The data suggests an air volume of ∼6 mm 3 per 7 cm 2 of surface area, which corresponds to a Salvinia layer of ∼9 ± 1 μm thickness. Fig. 9 Visualization and quantification of the Salvinia layer formation on Fluoropor 15. (A) A directly printed Fluoropor 15 membrane and a peeled-off Fluoropor 15 membrane were glued on a glass slide and partially submerged under water; the submerged area of the peeled membrane shows a silvery layer which indicated the trapped Salvinia layer whereas the color of the non-submerged area did not change. (B) The volume of the Salvinia layer was calculated by the Archimedes principle – a schematic of both directly printed and peeled Fluoropor 15 membrane under water shows the difference in displayed water volume by the Salvinia layer. (C) Change of the volume of Salvinia layer in dependence of the membrane surface area. The data suggests an air volume of ∼6 mm 3 per 7 cm 2 of surface area, which corresponds to a Salvinia layer of ∼9 ± 1 μm thickness."
} | 5,775 |
37068768 | PMC10166085 | pmc | 1,193 | {
"abstract": "As one of the most amazing aspects of life, all living\norganisms\nare formed by self-assembly, a fundamental biological design process\nin which ordered nanostructures are assembled from small parts. For\nexample, most of the biological tissues contain structurally soft\nand hard parts that are usually hierarchically organized at nano or\nmicro levels to achieve specific functions. Hydrogels are one of the\nmost promising soft materials owing to their potential applications\nin building of biological tissues and stretchable sensors. In this\nwork, a series of hydrogels are synthesized through the co-self-assembly\nof two types of amphiphiles in their aqueous solution prior to polymerization.\nSoft and hard parts with nanostructures of different order parameters\nare incorporated into the hydrogels. The hydrophilic segment (as soft\nphases) of the polymer network provides water absorption, fluid flow,\nand softness, whereas the hydrophobic segment (as hard phases) provides\nstrength and tearing and fracture resistance. Appropriate soft/hard\nnanostructures and their interfaces allow for the tailoring of the\ndesired morphological and mechanical properties, including a different\nwetting ability, toughness, energy dissipation, self-recovery, and\nfracture resistance arising from their nanostructures. This work provides\ninsights into the design of nanostructured anisotropic hydrogels with\ncontrolled morphological and mechanical properties.",
"conclusion": "3 Conclusions In summary, we reported\nthe design and synthesis of hydrogels with\ngradually changed nanostructures arising from the co-self-assembly\nof two amphiphilic molecules in their precursor solutions. Their nanostructure-dependent\nmorphological and mechanical properties were explored, including surface\nwettability, fracture stress, energy dissipation, self-recovery, and\ncrack resistance. It was found that a modulation in the fraction ratio\nof ionic to nonionic amphiphiles in the precursor solutions significantly\nchanges the assembled structures in the hydrogels from anisotropic\nto isotropic after polymerization, tailoring their physio-mechanical\nproperties, despite the molecular/chemical compositions of the precursor\nsolutions being almost the same. The hard/soft nanostructures and\ninterfacial bonding contribute significantly to their surface morphology\nand mechanical properties. This study provided insights regarding\nthe fabrication of high-performance soft materials through nanostructure\nengineering.",
"introduction": "1 Introduction Biological tissues are\noften complex composite materials with specific\nstructures or functions that have arisen over hundreds of million\nyears of natural evolution. The self-assembled structures in biological\ntissues can be mainly divided into two types of phases: soft and hard.\nThe main components in the soft phase are proteins (e.g., collagen,\nchitin, keratin, or elastin), starting with many basic building blocks\n(e.g., 20 natural amino acids) from the molecular level up to larger\nmolecules, such as polypeptides or polysaccharides. Hard phases, by\ncontrast, are mainly composed of minerals that nucleate and grow in\na biological environment. The mechanical properties of biological\ntissues containing both soft and hard phases are generally outstanding,\ndespite the weak and simple constituents from which they are assembled.\nFor example, natural nacre, consisting of a composite of inorganic\nplates (aragonite calcium carbonate as a hard phase at 95% by volume)\nand organic biopolymers (proteins and polysaccharides as a soft phase\nat 5% by volume), exhibit a unique combination of strength and toughness. 1 Strong but brittle inorganic plates, along with\norganic biopolymers having good ductility but low strength, have been\nsuccessfully integrated into a hierarchical structure. This has impacted\ncomposite materials having a high fracture toughness (1.5 kJ/m 2 ) approximately 3000 times higher than that of pure aragonite\ncalcium carbonate (∼0.5 J/m 2 ). 2 , 3 Similarly, natural bone has a unique combination of mechanical\nproperties (e.g., stiffness, strength, and toughness), which also\narises from the architectural design of its soft and hard phases with\nprecisely designed interfaces from the nano to macroscopic scale (bone\nconsists of 70% calcium phosphate crystals as a hard phase aligned\nin a staggered pattern in a 20–30% collagen matrix with some\nwater as a soft phase). 4 Many researchers\nhave tried to control the mechanical properties of artificial materials\nthrough the design of various nanostructures (e.g., the inclusion\nof nanofibers to reinforce the polymer networks) mimicking natural\nnanocomposites. The formation of such nanocomposites with soft and\nhard phases in nature is closely related to the self-assembly process\nof small composites and careful interface engineering in the hierarchical\nstructures. 5 The natural principles also\ninspire researchers to create strong interfaces or adhesions between\nsoft and hard, organic and inorganic, and living and non-living. 6 , 7 Owing to their biocompatibility, hydrophilicity, and tissue-like\nhigh water content, hydrogels are ideal soft materials having potential\napplications in biological tissues. 8 , 9 In addition,\nby incorporating ionic- or electronically conductive materials to\nhydrogels, they are also widely explored as conductive and stretchable\nmaterials with potential applications in smart skins, implantable\ndevices, and wearable electronic sensors. 10 − 13 However, conventional hydrogels\nare mechanically weak owing to their heterogeneously crosslinked polymer\nnetworks and lack of energy dissipation mechanisms. For example, common\npolyacrylamide (PAAm) or poly(acrylic acid) hydrogels exhibit high\ntransparence, high water content but weak mechanical properties. 14 , 15 An appropriate structural design can impart PAAm hydrogels with\nvarious attractive properties, including the desired morphologies,\noptical, improved tensile toughness, and strength or fracture resistance,\nwhile retaining their high water content. 16 Mimicking natural nanocomposites, applying a structural design at\nthe nanoscale is an effective strategy toward the synthesis of artificial\nsoft materials with the desired physical and mechanical properties. 17 However, the synthesis of nanocomposite\nhydrogels is still a challenge\nbecause their typically isotropic high water content excludes the\nuse of lithographic techniques in their three-dimensional (3D) polymer\nnetworks. On the other hand, using a biomimetic self-assembly, it\nis relatively easy to construct nanostructures from aqueous solutions\nof monomers prior to polymerization. Herein, we report the design\nand synthesis of hydrogels with different\nnanostructures arising from the co-self-assembly of two types of amphiphilic\nmolecules. Their nanostructure-dependent morphological and mechanical\nproperties are explored, including their tensile stress, energy-dissipation,\nself-recovery, and fracture resistance. It was found that a slight\nchange in the fraction ratio of ionic to nonionic amphiphiles in their\nprecursor solutions significantly changes the assembled structures\nin the hydrogels, from anisotropic to isotropic, after polymerization,\nresulting in a modulation of their mechanical properties, despite\nthe chemical compositions of the precursor solutions being 99.3% similar.\nThe hydrophilic segments (as a soft phase) in the hydrogels provide\nwater absorption, fluid flow, and soft properties, whereas the hydrophobic\nsegments (as a hard phase) contribute to the strength, tear, and fracture\nresistance. Rational hydrophobic/hydrophilic (hard/soft) interfaces\ncontribute to a tailoring of the desired surface morphologies and\nexcellent mechanical properties. This study provides insights into\nthe fabrication of high-performance soft materials with controllable\nmorphological and mechanical properties through nanostructure engineering.",
"discussion": "2 Results and Discussion 2.1 Co-Self-Assembly of Mixed Amphiphiles Modulating the order parameter of nanostructures in soft materials\nis not an easy task. In this work, a gradual structural change in\na hydrogel was achieved by adjusting the nanostructures from the co-self-assembly\nof two amphiphilic molecules in their precursor solutions. Co-self-assembly\nof a mixed amphiphile is a process by which multiple building blocks\nare coordinated to form nanostructures. The behavior can be understood\nin terms of a minimization of free energy, which comprises enthalpy\nand entropy terms. 18 Here, the nanostructure\ntransition from bilayer aggregates to curved bilayers and spherical\nmicelles was achieved by modifying the solution conditions from the\nconcentration of sodium dodecyl sulfate (SDS). It is well known\nthat some nonionic amphiphilic molecules can self-assemble into bilayer\nstructures in water when the concentration exceeds its critical micelle\nconcentration and the temperature increases to the Krafft point of\nthe molecules. 19 , 20 This self-assembly process of\namphiphiles has been used to mimic biological systems, such as the\nassembly of lipids and proteins, to develop more complex, hierarchical\nnanostructures. For example, n -dodecyl glyceryl itaconate\n(DGI) is a type of polymerizable amphiphilic molecules that can self-assemble\ninto bilayer membranes (lamellar liquid crystal) in water. 21 They maintain a bilayer spacing distance\nof up to 100 nm in the\npresence of small amounts of SDS (with a molar fraction of SDS to\nDGI of α = 0.026%). Owing to the existence of a layered structure,\nthe precursor solutions exhibit a slight blue color. More SDS molecules\nare then added to the above solution (increasing α to 0.5 and\n4%). The bilayers become curved and onion-like structures. 22 The entanglements among the bilayers largely\nincrease with a decrease in the layer distance. This further increases\nthe viscosity of the precursor solution. When the concentration of\nSDS is increased to α = 14% (i.e., 14.2 mM), however, the viscosity\nof the precursor solutions significantly decreases. The samples exhibit\na viscosity as low as that of water, which indicates that the layered\nstructures of DGI aggregates change into spherical micelles because\nthe concentration exceeds the critical micelle concentration of SDS\n(7.9 mM). These observations are consistent with the previous work\nby Tsujii and co-workers. 22 They measured\nthe viscosity and characterized the nanostructure variation with SDS\nin water. However, the synthesis of solid-state materials with such\nnanostructure transition and their nanostructure-dependent mechanical\nproperties have yet to be explored. 2.2 Series of Hydrogels Synthesized with Different\nOrder Degrees We then synthesized hydrogels composed of different\nnanostructures (labeled phase 1 to phase 4) from the co-self-assembly\nof amphiphiles in their precursor solutions ( Figure 1 a). After free radical polymerization, these\nnanostructures were maintained in 3D PAAm polymer networks. With the\nincrease of molar fraction of SDS to DGI (α change from 0.026\nto 14%), the structures in the corresponding hydrogels changed from\nordered to curved bilayers and spherical micelles ( Figure 1 b). Although their chemical\ncompositions are almost the same (99.3–99.98%), the prepared\nanisotropic/isotropic hydrogels exhibit dramatically different swellings,\nstructures, morphologies, and colors. Such differences occur even\nif the molar fraction variation of SDS to DGI is small, and DGI has\nan extremely small molar ratio (<5%) compared to the entire PAAm\npolymer network. Figure 1 Photographs and structural illustrations of nanocomposite\nhydrogels\nfrom the co-self-assembly of amphiphilic molecules. (a) Photographs\nof gels with different structural colors. (b) Illustrations of the\nstructural variation from highly orientated layers (phase 1) to curved\nlayers (phase 2 and phase 3) and micelle structures (phase 4) arising\nfrom the increase in the SDS ratio relative to DGI. After polymerization, the swelling behaviors of\nthe hydrogels in\nwater were then characterized ( Figure S1 ). The gels prepared from a molar fraction of SDS to DGI of α\n= 0.026% (phase 1), exhibit one-dimensional swelling in the direction\nperpendicular to the layers. The hydrogels with a curved layered structure\n(i.e., α = 0.5%, phase 2) show less anisotropic swellings, among\nwhich the swelling in the direction perpendicular to the layers is\n1.5-fold higher than that parallel to the layers. However, the hydrogels\nwith randomly curved bilayers (i.e., α = 4.0%, phase 3) or a\nmicelle structure (i.e., α = 14%, phase 4) exhibit a quasi-isotropic\nswelling. The different swelling behaviors of the gels further indicate\ndifferent nanostructures embedded in the polymer networks. As\nshown in Figure 1 a,\nthe gels exhibit different colors due to the nanostructures that\nreflect or scatter light. An ordered or low-order degree of layered\nstructure causes Bragg diffraction to produce structural colors. 23 The gels with randomly curved layers exhibited\ntranslucency which means the light is randomly scattered due to an\ninhomogeneous structure. The gels with spherical micelles were transparent,\nindicating that the light goes through gels easily. The reflection\nspectra of the hydrogels were measured after swelling. Although gels\nwith a curved layered structure also diffract light, the reflection\nintensity is much lower than that of a gel with a perfect layered\nstructure, indicating fluctuations of the layered structure in the\nhydrogels ( Figure S2 ). 2.3 Structural Characterizations and Order Parameters These nanostructures of the swollen hydrogels were then characterized\nby scanning electron microscopy (SEM). Cross-sectional SEM images\nreveal the aligned nature of the structures in these hydrogels. As\nshown in Figure 2 a–d,\nfor the phase 1 samples, they showed a closed packing orientation.\nFor phase 2 samples, there is still a clear view of the curved layered\nstructure. For the phase 3 samples, there are many interconnected\nnetworks with an average pore size of 10 μm. However, for phase\n4 samples, no typical structures were observed, and a larger pore\nsize of approximately 20 μm occurs. Figure 2 Structural and morphological\ncharacterizations. Cross-sectional\nSEM images of freeze-drying gels (after swelling) with an increase\nin the concentration of SDS in (a) phase 1 (0.026% mole ratio relative\nto DGI), (b) phase 2 (0.5%), (c) phase 3 (4%), and (d) phase 4 (14%).\n(e–h) Corresponding polarized light microscopy images for the\ncross sections. In addition, polarized optical microscopy was applied\nas a tool\nfor determining the self-assembled alignment of the structures in\nthe cross sections ( Figure 2 e–h). When the gels were oriented at 45° to the\npolarizers, a maximum light transmission occurred, whereas there was\nalmost no light transmission when the gels were parallel to one of\nthe polarizer axes. Phase 1 shows the highest birefringence, which\ngradually decreased and finally disappeared in phase 4, indicating\na variation from a highly aligned anisotropic structure to a nearly\nisotropic one. The orientation parameter degree of the nanostructures\nin the hydrogels\nwas quantitatively evaluated using small-angle X-ray scattering (SAXS)\nobserved at their cross sections ( Figure 3 ). The two-dimensional (2D) SAXS pattern\nof phase 1 shows two strong diffuse spots in the equatorial direction.\nThe corresponding azimuthal angle (φ) plot has two sharp peaks\nat φ = 90 and 270°. The orientation order parameter ( f ) is calculated to be 0.96, indicating that the layers\nare perfectly aligned in one direction ( Figure 3 a). For gels with a curved layered structure,\nthe 2D SAXS pattern has much less anisotropic scattering with two\nbroader peaks at φ = 90 and 270°. The value of f is decreased to 0.79, which is much smaller than the one\nwith ordered layered structures ( Figure 3 b). By contrast, phase 3 exhibits an almost\nisotropic scattering pattern, having a nearly flat φ plot with\nan f value close to zero, indicating a random distribution\nof the layered structures ( Figure 3 c). These results further demonstrate that a series\nof nanostructured gels were successfully prepared using the co-self-assembly\nof two types of amphiphilic molecules in their precursor solutions. Figure 3 Nanostructure-dependent\norder parameters (a–c) 2D SAXS images\nof gels with the molar fraction of SDS to DGI (a) α = 0.026%,\n(b) α = 0.5%, and (c) α = 4.0%, and their corresponding\nazimuthal angle (φ) plots. The solid curves represent fittings\nwith the obtained data. 2.4 Surface Wetting/Morphologies of Gels with\nDifferent Nanostructures These bilayers are hard and hydrophobic\nsegments embedded in soft and hydrophilic PAAm polymer networks. Such\nhydrophobic/hydrophilic packing structures change not only the nanostructures\nincorporated inside the polymer networks but also their surface morphologies\nand wetting behaviors. 24 We measured the\nstatic water contact angles of water drops on their surfaces and found\nthat these gels exhibit a different surface-wetting behavior ( Figure S3 ). Interestingly, they were able to\nobtain static water contact angles ranging from 38 to 110°. Such\nlarge tunable wetting states were achieved by modulating the order\ndegree parameters of the hydrophobic/hydrophilic nanostructures. The\nhigh order degree of a hydrophobic layered structure produces a high\nhydrophobic surface with a contact angle of up to 110°. When\nthe order degree decreases, the surface hydrophobicity gradually decreases\nand finally becomes hydrophilic with a contact angle of 38°.\nCompared with pure PAAm gels with a measured contact angle of 12°,\nthe surface-wetting behavior has been successfully tuned through the\nnanostructure engineering of these hydrogels. To further explore this,\nthe surface morphologies of gels with different nanostructures were\ncharacterized through 3D laser scanning microscopy. The surface morphologies\nof the gels have been gradually tuned from a bumpy surface to a flat\none, thus allowing the design of gel surfaces with controllable wetting\nproperties ( Figure S4 ). 2.5 Nanostructure-Dependent Tensile Strength The prepared nanocomposite gels were then cut into dumbbell shapes\nand the stress–strain tensile performances of them were tested\n( Figure 4 a,b). Phase\n1 showed a high Young’s modulus ( E ) of 0.06\nMPa, a high fracture stress (σ) of 0.35 MPa, and a fracture\nstrain (ε) of 1450%. There is a well-defined yielding point\nat a strain of approximately 30%, from which the unidirectionally\naligned bilayers/interfaces start to rupture. The phase 2 samples\nexhibit a lower modulus along the bilayer directions and a lower fracture\nstress of ∼0.2 MPa, as well as a fracture strain of 1200%.\nThe phase 3 samples show a significant decrease in the fracture stress\n(σ = 0.085 MPa, ε = 1100%) owing to the disappearance\nof an ordered layered structure. Phase 4 with micelles’ structures\nexhibit a further weaker fracture stress (σ = 0.067 MPa) despite\na similar chemical composition as that of phase 1 or phase 2. These\nnanostructured hydrogels (phases 1–4) all have a higher fracture\nstress and strain compared with those of the pure PAAm gel, which\nare only 0.023 MPa and 640%, respectively. The compressive stress–strain\ncurves for these nanocomposite gels were also measured. The phase\n1 gel holds up a stress of 1.17 MPa, where the compressive strain\nis ∼85%. The compressive stress gradually decreases from phase\n1 to phase 4. All the compressive stresses of nanocomposite gels are\nhigher than that of PAAm gel compared at the same strain ( Figure S5 ). Figure 4 Stress–strain curves of gels with\ndifferent nanostructures\nand their interfacial bonding. (a) Photograph of the gels. (b) Stress–strain\ncurves of the gels. (c) Different formation states of interfacial\nbonding are proposed for gels with different nanostructures. The molar ratio of the hard (PDGI) to soft (PAAm)\nparts is fixed\nat 5% for all gels in their as-prepared states; however, the volume\nratio of the hard parts decreases owing to the larger swelling of\nthe soft and hydrophilic PAAm segments after immersion in water. The\nvolume of the hard parts is close to 2% of the gels, as calculated\nfrom the layer thickness of the composite from the reflection spectrum.\nUnder uniaxial elongation, the hard PDGI bilayers are primarily under\ntension, whereas the soft PAAm matrix transfers the load between neighboring\nhard layers through shearing. To transfer the force, the formation\nof strong interfaces between the soft/hard parts is essential. These interfaces have no covalent bonds, but hydrogen bonds (H-bonds)\nare easily formed in the −OH or −COOH groups of bilayers\nwith the amid group (−CONH 2 ) of PAAm. 25 , 26 To certify this, we treated the gels with alkaline solution (1.0\nM) and urea solution (1.0 M), respectively, and found that some layers\nof phase 1 and phase 2 gels peeled after sufficient swelling in alkaline/water\nby breaking interfacial bonds ( Figure S6 ). However, such a peeling phenomenon was not observed in urea solution,\nsuggesting that urea cannot sufficiently break these interfacial bonds\nin these gels. The dramatic differences in their mechanical\nproperties suggest\nthat there are significant differences in these interfacial bonded\npolymer networks. As shown in Figure 4 c, different formation states of the interfacial H-bonds\nare proposed in these nanostructured gels, i.e., cooperative, less\ncooperative, and non-cooperative H-bonds. For phase 1, there are approximately\n4000 layers calculated from the Bragg diffraction, indicating that\nthousands of interfaces appeared between alternating PDGI and PAAm\nlayers. These H-bonds are aligned in one dimension, which works cohesively\nto achieve high fracture stress and large strain. For phase 2, the\nlayered structure becomes curved, therefore the interfaces are not\nwell aligned, and they cannot work cohesively in one direction to\nachieve the highest fracture stress. When the nanostructure changes\nto more curved or even spherical shapes, such cohesively working interfaces\nare lost, and the H-bonds can no longer work efficiently. Despite\nthis, they have an enhanced tensile strength of approximately 3.7-\nand 2.9-fold greater than that of the pure PAAm gel, respectively.\nIn addition, the increase in the SDS fraction ratios also decreases\nthe number of H-bonds because SDS has no functional groups to form\nH-bonds with PAAm polymer networks. The H-bonds at the interfaces\noperate as sacrificial bonds that transfer the stress and improve\nthe mechanical properties of the soft materials. 27 , 28 Similar to polymer-based nanocomposites, H-bonds also exist\nin\nthe organic–inorganic interfaces, increasing the tensile strength.\nFor example, amine groups at the end of the silane-modified Al 2 O 3 platelets are expected to form H-bonds with\nthe oxygen atoms of chitosan, increasing the adhesion between inorganic\nplatelets and an organic matrix. 29 Therefore,\nbased on the tensile strength and toughness, the optimal design raises\nthe total soft/hard interfaces or the interfacial bonding to the highest\npossible level without breaking the nanolayered structures. These\nexperimental results also indicate that amphiphilic-formed nanostructures\nplay an important role in toughening the gel matrix mainly owing to\nthe energy dissipation from the breakage of the interfacial bonding\nat soft/hard interfaces. 2.6 Nanostructure-Dependent Energy Dissipation Common isotropic hydrogels, such as PAAm, are mechanically weak\nowing to their chemically crosslinked networks and lack of energy\ndissipation mechanisms. As shown above, the nanoscale structures and\nhard/soft interfaces in polymer networks contribute significantly\nto their mechanical properties. These anisotropic gels exhibit\ndifferent amounts of energy dissipation under tensile elongation.\nThe area between the loading and unloading curves represents the energy\ndissipated per unit volume. As shown in Figure 5 a, when the tensile elongation is 100%, phase\n1 with a perfect layered structure has the highest energy dissipation\nof 25.52 kJ/m 3 (with a pronounced hysteresis). Phase 2\nwith a curved layered structure has a much lower energy dissipation\nof 5.63 kJ/m 3 ( Figure 5 b), whereas phases 3 and 4 exhibit negligible energy\ndissipation (0.2–0.9 kJ/m 3 ) ( Figure 5 c,d). Figure 5 Nanostructure-dependent energy dissipation\nand anisotropic modulus.\n(a–d) Tensile loading and unloading curves of gels with different\nnanostructures. (e) Young’s modulus and dissipated energy for\nthe gels. (f) Anisotropic modulus in the direction parallel and perpendicular\nto the layers for the gels with different nanostructures. Here, the energy dissipation in these soft materials\nmainly arises\nfrom the dissociation of the H-bonds at the interfaces. When the elongation\nis in the direction parallel to the layers, the gels with cooperative\nH-bonds exhibit the largest energy dissipation owing to the large\ndissociation of the H-bonds at thousands of layered interfaces. For\nsamples with curved layers, however, the energy dissipation decreases,\nindicating that the total energy required to break the H-bonds decreases.\nFor the gels with spherical micelles, non-cooperative H-bonds contribute\na negligible amount of energy dissipation. The H-bonds at the interfaces\ncan be considered a glue holding the layered nanocomposites together. 2.7 Nanostructure-Dependent Anisotropic Modulus Young’s modulus, expressing the stiffness of these hydrogels\nat the start of the tensile test, showed a clear decrease from phase\n1 to phase 4 ( Figure 5 e). The samples in phase 1 exhibit a much higher Young’s modulus\nthan the others, mainly arising from the ordered hard layers in the\nsoft matrix. Owing to the initial force loading up to a point where\nthe layered structure stretches (but does not deform), the hard layers\nare initially under greater tension than the soft layers. The hydrophobic\ninteractions among amphiphilic molecules in the hard layers may contribute\nmostly to Young’s modulus. These layered structures also\nmake the gels intrinsically anisotropic, in which the modulus depends\non the direction of the force vector. The anisotropy of the modulus\nfor gels with different nanostructures was evaluated based on tensile\ntests conducted in the direction parallel to the layers and through\ncompressive tests applied in the direction perpendicular to the layers.\nAs shown in Figure 5 f, the modulus in the direction along the layers ( E ∥ ) was approximately 5 times higher than that perpendicular\nto the layers ( E ⊥ ) for the gels\nwith a high order degree (phase 1). The anisotropic ratio ( E ∥ / E ⊥ ) was decreased to 2.1 for phase 2 and finally reached close to 1\n(isotropic) for phase 4 when the structure changed into spherical\nmicelles. This anisotropic ratio of the modulus in two normal directions\nis close to the values found in bone (1.7–2.1), facilitating\ntheir adaptation to environmental force loading during their evolution. 30 2.8 Nanostructure-Dependent Self-Recovery As mentioned above, the dissociation of H-bonds dissipates different\namounts of energy in different anisotropic gels. After the force is\nrelaxed, the H-bonds between the interfaces can reform and the original\nstrength of the gels is thereby recovered. The dissociation\nof H-bonds is reversible, which impacts these anisotropic hydrogels\nwith a self-recovery property. A good self-recovery behavior of hydrogels\nis important for extending their service life. Traditional double\nnetwork hydrogels have two chemically crosslinked polymer networks\nand usually cannot self-recover their stiffness after the first loading\nowing to the permanent damage of covalent bonds in the first brittle\npolymer networks. 31 Thus far, to design\nand understand the inside molecular mechanisms of a hydrogel that\ncombines both mechanical strength with long elongation and good self-recovery\nhas been a challenge. These hydrogels designed with different nanostructures\nbut similar chemical compositions were used to understand their structure-dependent\nself-recovery behaviors. As shown in Figure S7a , when a second\ncyclic loading was conducted on phase 2 samples without any waiting\nperiod (0 min), the hysteresis in the curve was smaller than the original,\nindicating that the internal molecular state does not immediately\nrecover to the original state. However, the third loading–unloading\nloop obtained after a 5 min waiting time approaches that of the first\none, with the hysteresis ratio reaching 100%, indicating a complete\nself-recovery ability of the samples. A relatively long time is needed\nfor the samples with perfect layered structures because the recovery\nof all cooperative H-bonding at the interfaces is a time-dependent\nprocess. As shown in Figure S7b , all gels\nexhibit a self-recovery behavior. Phases 1 and 2 exhibit a slow self-recovery,\nwhereas phases 3 and 4 exhibit a quick self-recovery owing to the\npresence of different states of interfacial bonds. The force–time\ndiagram of repetitive cyclic tensile tests shows that all the gels\ncan endure at least 100 stretching cycles with 100% strain applied\n( Figure S8 ). 2.9 Nanostructure-Dependent Viscoelasticity Rheological frequency sweep tests were conducted to study the dynamic\ncharacteristics of polymer networks with different nanostructures\n( Figure S9 ). For all samples, the storage\nmodulus ( G ′) is higher than the corresponding\nloss modulus ( G ″), showing the solid-like,\nelastic nature of the gels. The values of G ′\nand G ″ were improved with an increase in the\norder parameter of the nanostructures, which was mainly due to increases\nin the interfacial H-bonds. In addition, the hydrogels with a high\norder parameter showed a more significant frequency dependence. For\nexample, G ′ of the phase 1 and phase 2 gels\nincreased with increasing frequency, which is a consequence of the\nviscoelastic nature of the gels. That is, the gels are more elastic\nat a higher frequency. The loss tangent (tan δ) decreased with\nthe frequency, with a more significant decrease at a frequency of\n<0.1 rad/s, indicating more relaxation of the gels at a lower frequency.\nNot only these hydrophobic nanostructures but also nanomaterials in\nhydrogels can significantly affect the viscoelasticity of the composites.\nFor example, the addition of small amounts of carbon nanotubes or\nnanocelluloses to hydrogels can effectively improve the viscoelastic\nproperties of the composites. 32 − 35 2.10 Nanostructure-Dependent Crack Resistance Crack generation and extension mechanisms in soft materials have\nattracted considerable scientific interest owing to the lack of comprehensive\ninformation on the development of these processes under stress loading.\nHow the inherent nanostructures affect the crack propagation is also\nan interesting topic for soft materials. In conventional single-network\nhydrogels, the crack propagates extremely quickly owing to the lack\nof internal mechanisms for mechanical energy dissipation. For example,\nfor pure PAAm hydrogels, they are intrinsically weak and quite prone\nto fracture. When embedding high orientationally layers in PAAm gels\n(phase 1), they can be stretched normal to the plane of the crack\nfor more than 3.4 times the original strain even with a precut notch.\nAfter relaxation of the tensile stress, no obvious increase in the\nnotch length occurs ( Figure 6 a). The notch front becomes significantly blunt to prevent\ncrack propagation during stress loading. 36 For the gels (phase 2) with a curved layered structure, however,\nthe notch also becomes blunt and then starts to run when the strain\nundergoes a threshold of approximately 2 ( Figure 6 b). This strain value further decreases to\n1 for the phase 3 gels with a low layer orientation. The speed of\nthe crack propagation is much faster than that in phase 2 ( Figure 6 c). For the gel with\na spherical micelle structure (phase 4), the crack runs quickly with\na further decrease in the threshold strain of 0.5 ( Figure 6 d). Except for phase 1, a similar\nphenomenon occurs for the other samples where a sharp crack is first\nblunted at the tip of the cut and then propagates with a U-shaped\nprofile. However, the propagating crack travels continuously (speeding)\nthrough the whole specimen in phase 4 but incrementally in phase 2.\nThe onset of a fracture thus depends highly on the nanostructures\nof the gels, despite the chemical compositions of the samples being\nalmost the same ( Figure 6 e–h). Figure 6 Fracture tests in nanocomposite hydrogels. Each sample\nwas glued\nto two rigid clamps and precut with a sharp crack. (a) No obvious\nrupture in the phase 1 samples with a highly order degree of the layered\nstructure even at a large strain of 3.4. (b) Rupture of phase 2 after\nbeing stressed perpendicular to the crack direction at a strain of\n2.1. (c) Phase 3 samples showing a crack that rapidly extends at a\nstrain of 1.1. (d) Phase 4 samples showing a crack that rapidly extends\nat a strain of 0.58. (e–h) Schematic illustration of different\nnanostructures (from left to right: phase 1 to phase 4) to prevent\ncrack propagation as seen from their cross sections. The specimens\nhad a length of 20 mm and a width of 10 mm. An initial crack length\nof 5 mm was cut using a sharp blade. 2.11 Nanostructure-Dependent Fracture Energy Fracture energy describes the energy required to continue a fracture\nthrough the network, which is the energy required to create a unit\narea of crack growth (J m –2 ). The fracture energy\ncan be characterized by different methods, such as a pure shear test,\nsimple extension test, single-edge notch test, and trouser tear test.\nTo quantitatively characterize these nanostructured hydrogels, here\nthe fracture energy was estimated from the tearing fracture energy\nthrough trouser tear tests, where the crack is deformed by an out-of-plane\nshear loading. Typically, an initial crack is created at the\nsame position of the gels with the same size. The legs of the gels\nare then fixed by their two clamps on a tensile machine. During the\ntest, the machine maintains a slow tensile speed and the force ( F ) is recorded until the sample fails. The tearing fracture\nenergy of the gels can be calculated as T c = 2 F / t , where t is the thickness of the sample. 37 The\nelastic deformation of the two arms is not considered here. Moreover,\ntearing tests were conducted under quasi-static loading at an extension\nrate of 5 mm/min. The force–extension curves for the\nsamples are shown in Figure 7 . As the figure shows,\ngels with a high order degree of the layered structure exhibit a high\nloading force and a high tearing fracture energy ( T c ) which is the amount of energy required to advance a\nfracture plane by one unit area. A T c of\n678.0 J m –2 was measured for the phase 1 samples.\nIn comparison, T c of phases 2 and 3 gradually\ndecrease to 155.6 and 47.5 J m –2 , respectively,\nand further decrease to 30.3 J m –2 for phase 4 with\nspherical micelles ( Figures 7 a–d, S10 ). By contrast,\nfor PAAm single-polymer-network gels, a tearing fracture energy of\nonly 4.5 J m –2 is measured, through which the crack\nwould be able to propagate freely without barriers. Thus, different\nnanostructures have improved the tearing fracture energy by approximately\n151, 34.6, 10.6, and 6.7 times that of the neat gel, respectively. Figure 7 Load-extension\nresponse during trouser tear tests for (a) phase\n1, (b) phase 2, (c) phase 3, and (d) phase 4 gels with different nanostructures. The fracture energy in stretchable soft materials\ncan mainly occur\nfrom two aspects: the intrinsic fracture energy of the polymer network\nahead of the crack tip and the viscoelastic mechanical energy dissipated\nin the regions around the crack. 38 , 39 The intrinsic\nfracture energy is the energy required to break the polymer chains\nlying on the plane of the crack, which is related to the volume fraction\nof the polymer networks, the bond number per unit, and the bond energy.\nFor the single PAAm polymer network, the fracture energy arises mainly\nfrom the intrinsic fracture energy, which is quite small. For the\nnanostructured hydrogels, the interfacial H-bonds increase both the\nintrinsic fracture energy and the viscoelastic mechanical energy dissipation.\nIncluding a nanostructure to dissipate more energy is an efficient\nway to create a crack-resistant hydrogel. The cooperative H-bonds\ncan largely increase the intrinsic fracture energy and the viscoelastic\nenergy dissipation around the crack, hindering the crack propagation.\nIn addition, through an appropriate hard/soft nanostructure arrangement\nor volume ratios between two parts, the stress concentration around\nthe crack tip can be reduced owing to a more efficient force redistribution.\nThe H-bonds at the soft and hard interfaces are also crucial for redistributing\nthe load and delocalizing the deformation."
} | 9,208 |
26755434 | PMC4709512 | pmc | 1,194 | {
"abstract": "Spider major ampullate gland silks (MAS) vary greatly in material properties among species but, this variation is shown here to be confined to evolutionary shifts along a single universal performance trajectory. This reveals an underlying design principle that is maintained across large changes in both spider ecology and silk chemistry. Persistence of this design principle becomes apparent after the material properties are defined relative to the true alignment parameter, which describes the orientation and stretching of the protein chains in the silk fiber. Our results show that the mechanical behavior of all Entelegynae major ampullate silk fibers, under any conditions, are described by this single parameter that connects the sequential action of three deformation micromechanisms during stretching: stressing of protein-protein hydrogen bonds, rotation of the β-nanocrystals and growth of the ordered fraction. Conservation of these traits for over 230 million years is an indication of the optimal design of the material and gives valuable clues for the production of biomimetic counterparts based on major ampullate spider silk."
} | 285 |
28344645 | PMC5360037 | pmc | 1,195 | {
"abstract": "Background The feasibility of heterotrophic–phototrophic symbioses was tested via pairing of yeast strains Cryptococcus curvatus , Rhodotorula glutinis , or Saccharomyces cerevisiae with a sucrose-secreting cyanobacterium Synechococcus elongatus . Results The phototroph S. elongatus showed no growth in standard BG-11 medium with yeast extract, but grew well in BG-11 medium alone or supplemented with yeast nitrogen base without amino acids (YNB w/o aa). Among three yeast species, C. curvatus and R. glutinis adapted well to the BG-11 medium supplemented with YNB w/o aa, sucrose, and various concentrations of NaCl needed to maintain sucrose secretion from S. elongatus , while growth of S. cerevisiae was highly dependent on sucrose levels. R. glutinis and C. curvatus grew efficiently and utilized sucrose produced by the partner in co-culture. Co-cultures of S. elongatus and R. glutinis were sustained over 1 month in both batch and in semi-continuous culture, with the final biomass and overall lipid yields in the batch co-culture 40 to 60% higher compared to batch mono-cultures of S. elongatus. The co-cultures showed enhanced levels of palmitoleic and linoleic acids. Furthermore, cyanobacterial growth in co-culture with R. glutinis was significantly superior to axenic growth, as S. elongatus was unable to grow in the absence of the yeast partner when cultivated at lower densities in liquid medium. Accumulated reactive oxygen species was observed to severely inhibit axenic growth of cyanobacteria, which was efficiently alleviated through catalase supply and even more effectively with co-cultures of R. glutinis . Conclusions The pairing of a cyanobacterium and eukaryotic heterotroph in the artificial lichen of this study demonstrates the importance of mutual interactions between phototrophs and heterotrophs, e.g., phototrophs provide a carbon source to heterotrophs, and heterotrophs assist phototrophic growth and survival by removing/eliminating oxidative stress. Our results establish a potential stable production platform that combines the metabolic capability of photoautotrophs to capture inorganic carbon with the channeling of the resulting organic carbon directly to a robust heterotroph partner for producing biofuel and other chemical precursors. Electronic supplementary material The online version of this article (doi:10.1186/s13068-017-0736-x) contains supplementary material, which is available to authorized users.",
"conclusion": "Conclusions Culture media formulations were obtained by evaluating the impact of medium components on axenic growth of the three yeast strains and the sucrose-secreting cyanobacterium S. elongatus . The yeasts R. glutinis and C. curvatus could efficiently take up sucrose secreted from S. elongatus and expand their cell numbers in co-culture, while S. cerevisiae cells failed to grow in co-culture with S. elongatus likely due to inefficient sucrose incorporation [ 21 ]. Furthermore, robust co-cultures of R. glutinis and S. elongatus could last for more than 1 month, with higher biomass and lipid yield than the axenic culture of S. elongatus in batch culture. These findings demonstrate the potential of a stable sustainable phototrophic-heterotrophic co-culture system as an emerging bioenergy platform, taking the advantage of synthesis of organic carbon from CO 2 by photobionts and channeling the carbon to oleaginous heterotrophs without the need for exogenous supplies of carbon. Equally important, the phototroph and heterotroph both exhibited symbiotic interactions with their respective partner species. The cyanobacteria provide essential organic carbon nutrients to the yeast and perhaps other components, such as oxygen. Concomitantly, the yeast cells efficiently limit the generation of toxic ROS in the culture which enable and encourage survival and expansion of S. elongatus at low cell densities, significantly facilitating the growth of S. elongatus in co-cultures. Dealing with ROS may be a common feature defining community interaction between phototrophs and heterotrophs [ 45 ]. The finding that cyanobacterial growth in co-culture with yeast was superior to cultures containing catalase suggests additional factors gained from the heterotroph in co-culture. Two of the major constituents of some lichens include cyanobacteria and eukaryotic species [ 1 , 2 ]. The results from the co-culture of cyanobacteria and fungal partner constructed in this study provide strong evidence that mutual interactions can be achieved between phototrophs and eukaryotic heterotrophs. While the phototroph provides an organic carbon source to heterotroph, the heterotroph can play an equally important role by assisting in phototroph survival and robust growth by protecting the cultures against oxidative stress. In this way, we have established a successful artificial lichen system composed of photobionts and eukaryotic heterotrophs yielding symbiotic and synergistic interactions that represent a potential sustainable production platform for biofuels from sunlight and CO 2 .",
"discussion": "Results and discussion Axenic growth of different yeast strains Previously, sucrose permease, cscB, expressing S. elongatus cells were engineered to efficiently secrete sucrose into the culture medium. The sucrose production is dependent upon osmotic pressure, alkaline environment, and supply of IPTG inducer [ 20 ]. To examine if various yeasts could grow well with sucrose as the carbon source, and under the same media conditions in which S. elongatus cells grow and produce sucrose, yeast strains R. glutinis , C. curvatus , and S. cerevisiae were cultured individually in modified BG-11 medium. The yeast C. curvatus could not grow in BG-11[co] medium supplemented with sucrose only (Fig. 1 a). Alternatively, R. glutinis and S. cerevisiae cells grew in these conditions only after a long lag phase (Fig. 1 b, c). However, all the three yeast strains performed well in 2 g/L of sucrose complemented BG-11[co] medium containing either yeast extract (YE) or yeast nitrogen base without amino acids (YNB w/o aa). YE yielded the best growth of each strain, with the highest cell density over the duration of the culture. Fig. 1 Axenic growth of different yeast strains. a – c Effect of medium components on monoculture of yeasts; “BG” indicates BG-11 added with 2 g/L sucrose, 4 mM ammonium chloride, 1 mM IPTG, and 100 mM NaCl; “BG + YNB w/o aa” indicates “BG” supplied with YNB w/o aa; “BG + YE” indicates “BG” supplied with YE. d – f Effect of glucose/sucrose and light/dark on monoculture of yeasts; Medium used here was BG-11[co] supplied with 2 g/L sucrose or 2 g/L glucose, as indicated in the legends. g – i Effect of sucrose concentration on monoculture of yeasts; Medium used here was BG-11[co] supplied with various concentration of sucrose, as indicated in the legends. j – l Effect of NaCl concentration on monoculture of yeasts. Medium here was BG-11[co] supplied with 2 g/L sucrose, but with adjusted NaCl concentration in each condition. a , d , g , j for C. curvatus ; b , e , h , k for R. glutinis ; c , f , i , l for S. cerevisiae . Light condition was used if no specific statement. All data are averages of biological triplicates ± standard deviation \n Typically, wild-type yeast cells require YE for growth, which provides nitrogenous compounds, carbon, sulfur, trace nutrients, vitamins, and other important growth factors. However, the proposed partner strain S. elongatus showed no growth in BG-11 medium with added YE (see Additional file 1 ), but could grow well in BG-11 medium supplemented with or without YNB w/o aa. The growth of a diverse array of cyanobacteria is inhibited by exogenous metabolites such as amino acids [ 33 – 35 ], although endogenous synthesis is obligatory. For example, glutamine, lysine, as well as histidine inhibited the growth of Synechocystis PCC6803 [ 35 ]. Therefore, YNB w/o aa instead of YE was adopted as component supplement in BG-11[co] medium for yeast growth. The above results in Fig. 1 a–c indicated that the three wild-type yeast strains can consume sucrose as a sole carbon source, and no sucrose remained in the final culture medium (data not shown). As glucose is the most ubiquitous carbon source for yeast in laboratories, growth on sucrose was compared to this carbon source. C. curvatus exhibited similar growth with sucrose as a substrate compared with glucose; while R. glutinis and S. cerevisiae exhibited a prolonged lag phase with sucrose (Fig. 1 d–f). Additionally, illumination had little influence on the growth of the three yeast strains, which indicates compatibility of yeast strains with phototrophs under light condition. To further evaluate if yeast cells could grow on a sucrose substrate with a phototrophic partner in a co-culture system, the effect of varying sucrose concentrations was investigated using pure cultures of the three yeast strains in BG-11[co] medium. Growth of S. cerevisiae demonstrated dependence on sucrose concentration in the medium as previously reported [ 21 , 22 ] (Fig. 1 i). However, both C. curvatus and R. glutinis showed similar growth profiles regardless of the sucrose concentrations, although R. glutinis exhibited an overall longer lag phase (Fig. 1 g, h). A similar delay in the growth phase was observed for the R. glutinis when comparing growth on sucrose to growth on glucose (Fig. 1 e). The sucrose production by cscB \n + \n S. elongatus cells can be tuned via the osmotic pressure imposed on the culture system, with up to ~80% of total mass generated as sucrose at higher salinities [ 20 ]. Therefore, we investigated yeast growth with increasing NaCl concentrations (0, 100, 150 and 200 mM) in BG-11[co] medium (Fig. 1 j–l). The three yeast strains exhibited similar growth patterns in which the lag phases were extended and the growth rates delayed with increasing salt concentrations (see Additional file 2 ). Lag phases appeared approximately two times longer for growth of the three yeast strains with 200 mM NaCl compared with no supplemental NaCl supply. To ensure robust growth of the three yeast strains, 100 mM NaCl was used with BG-11[co] medium in co-culture with S. elongatus in subsequent studies. Axenic growth of S. elongatus Sucrose export by cscB \n + \n S. elongatus is dependent upon medium pH, with maximal sucrose export in an alkaline environment [ 20 ]. To ensure that sucrose was continuously and efficiently secreted, three different Good’s buffers, HEPES, HEPPSO, and TAPS, were tested for their capability in buffering the axenic growth of cyanobacteria S. elongatus cells. As shown in Fig. 2 , S. elongatus cells did not grow and gradually died in BG-11[co] medium without any buffering agent, in which the pH dropped from an initial value of 8.8 to below 6 after 4 days. For the pH buffered groups, however, the S. elongatus cells exhibited similar growth curves and exported the sucrose at comparable rates, with the sucrose level in the medium gradually increasing up until 9 days. Additionally, the three buffers displayed similar pH buffering capabilities and could all maintain pH values above 7.2 over the culture duration. Based on these findings, HEPES buffer was chosen as the buffering agent to be used in following co-culture experiments. Fig. 2 Axenic growth of S. elongatus with different pH buffer in the medium. a Cell numbers (per mL) of S. elongatus ; b Sucrose (mg/L) secreted by S. elongatus ; c pH value. 3 g/L of different buffers were used in the medium. All data are averages of biological triplicates ± standard deviation \n Co-culture of S. elongatus with different yeast strains Following the initial mono-culture experiments, the CscB-expressing S. elongatus was next grown in co-culture medium BG-11[co] with the different yeast strains. Shown in Fig. 3 are the cell counts for S. elongatus and the three yeast strains, along with sucrose levels. The S. elongatus cells grown in co-culture with yeast cells exhibited slightly higher cell densities, on average, than cells in mono-culture in the middle to late exponential phase (6 and 9 days). Among yeast strains, wild-type S. cerevisiae displayed no growth in the co-culture consistent with previous findings [ 22 ], and actually decreased in cell numbers over the time period, even though the initial CFU numbers inoculated were 20 times larger compared with the other two yeast species. C. curvatus showed a modest enhancement in cell numbers from 1.8 × 10 5 to 5.7 × 10 6 CFU over the 11 day culture duration. Finally, R. glutinis cells grew efficiently in co-culture, reaching cell densities five times higher than the other two cultures. Surprisingly, R. glutinis and C. curvatus cells utilized similar amounts of sucrose with sucrose levels kept below 200 mg/L in the medium compared to levels above 500 mg/L in the mono-culture of S. elongatus . Sucrose levels in the co-culture medium of S. cerevisiae with S. elongatus showed an accumulation pattern more similar to the mono-culture of S. elongatus , with final sucrose levels around 500 mg/L after 11 days, indicating the yeast S. cerevisae used in this study could not utilize sucrose at these levels, and likely require higher levels of sucrose for expansion as indicated in Fig. 1 and as demonstrated in the literature [ 21 , 22 ]. Fig. 3 Co-culture of S. elongatus with different yeast strains. a Cell numbers of S. elongatus ; b CFU numbers of yeasts; c Sucrose secreted by S. elongatus . All data are averages of biological triplicates ± standard deviation \n In order to evaluate sustainability of a co-culture system composed of cyanobacteria and yeast, long-term batch and semi-continuous cultures of the cyanobacteria S. elongatus together with the yeast R. glutinis were carried out. After 4 days in batch, the growth of S. elongatus plateaued but remained stable for more than 1 month with a marginal increase in densities (Fig. 4 a). Meanwhile, the CFU of the yeast R. glutinis increased rapidly for 8 days and then increased more gradually in the batch cultures (Fig. 4 b). In semi-continuous culture, cyanobacterial and yeast cell numbers were reduced by half through replacement of the original cultures with fresh BG-11[co] medium at 10 and 21 days, respectively, shown in Fig. 4 as arrows. Nonetheless, cell numbers rapidly recovered and eventually attained similar levels to the batch culture. Phototrophic cyanobacteria and heterotrophic yeasts possess different resource acquisition traits, so less competition is observed between members compared with either co-cultures of phototrophs or co-cultures of heterotrophs. For example, acetate-secreting cyanobacterium Synechococcus sp. overtook a culture when growing together with the acetate-consuming algae Chlamydomonas reinhardtii , and the cyanobacterial cells needed to be encapsulated in alginate beads to slow their growth [ 36 ]. Fig. 4 Long-term batch and semi-continuous co-culture of S. elongatus and R. glutinis . a Cell numbers of S. elongatus ; b CFU numbers of yeast R. glutinis . Arrows indicate the time when 50 mL of original culture was withdrawn and 50 mL of fresh BG-11[co] medium was added. All data are averages of biological triplicates ± standard deviation \n Interestingly, the final biomass DW in the batch co-culture was much higher than the batch mono-culture of S. elongatus (Table 1 ). We also examined lipid content in the co-culture versus monoculture. The total fatty acids (TFA) in the batch co-culture was 54% higher than in mono-culture, in part because of the increased biomass from the yeast partner. In semi-continuous culture, the biomass DW and TFA in the final harvested biomass were 650 and 35 mg/L, respectively. With addition of the 50 mL of culture that was removed both at 10 and 21 days, respectively, the total biomass yield finally in semi-continuous culture was 721 mg/L of medium. The TFA was 39 mg/L, which was about 60% above the levels for S. elongatus mono-culture. Also, the cultures could be manipulated by nitrogen or other stresses for even higher levels of lipids as desired. Table 1 Biomass DW and total fatty acids (TFA) in batch mono-culture, batch co-culture, and semi-continuous co-culture Mono-culture (mg/L) Co-culture (mg/L) Semi-continuous culture Final harvest (mg/L) 10st day withdraw (mg in 50 mL) 21st day withdraw (mg in 50 mL) Total harvest (mg/L) DW 570.37 ± 3.49 792.36 ± 0.98 650.00 ± 3.93 28.90 ± 0.52 50.30 ± 0.33 721.00 ± 1.02 TFA 24.61 ± 1.49 37.86 ± 0.77 34.66 ± 3.77 1.64 ± 0.02 2.69 ± 0.07 39.02 ± 2.13 All data are averages of biological triplicates ± standard deviation \n A fatty acid profile analysis revealed that the principal fatty acids produced in the mono- and co-cultures were palmitic (C16:0), palmitoleic (C16:1), stearic (C18:0), oleic (C18:1), and linoleic acid (C18:2) (Fig. 5 ). While the mono-cultures of S. elongatus contained significantly higher levels of palmitic, stearic, and oleic acids, the co-cultures of both batch and semi-continuous culture showed enhanced levels of palmitoleic and linoleic acids. These oils could serve as potential oil feedstocks for biodiesel production that require long-chain fatty acids with 16 and 18 carbon atoms [ 37 ]. Fig. 5 Comparison of fatty acid profiles in different cultures. All data are averages of biological triplicates ± standard deviation. *Significant difference in paired comparisons ( p < 0.05) \n Chemical interaction involved in cultures of cyanobacteria with/without yeast While the yeast R. glutinis is clearly dependent on the cyanobacterium for a supply of sucrose for growth, we wanted to explore any impact the yeast may have on the growth and maintenance of S. elongatus in HEPES buffered BG-11[co] medium. Hydrogen peroxide (HOOH) and related reactive oxygen species (ROS) can arise from a number of sources, including light exposure [ 38 , 39 ]. Further, chemicals present in cultural systems are known to produce HOOH [ 40 ]. Previously, HEPES with 1-10 mM concentration in culture could generate enough HOOH to kill the axenic phytoplankter Prochlorococcus strain [ 40 ]. Indeed, in our cultures, BG-11[co] medium with various concentrations of HEPES produced increasing levels of HOOH (Fig. 6 ), with HOOH concentrations approaching an asymptotic level after several days. S. elongatus cells at various inoculum concentrations were tested to check the influence of HOOH on their axenic growth and to explore any role of yeast heterotrophs on the growth of the cyanobacterium under these conditions. At higher cell inoculum concentrations (1.6 × 10 7 and 3.2 × 10 6 cells/mL), the axenic growth of S. elongatus was robust (Fig. 7 a, b), with negligible fractions of dead cells (data not shown). With a lower inoculum concentration of 1.6 × 10 6 cells/mL, however, no growth of axenic S. elongatus culture was observed, and the fraction of live cells rapidly dropped to less than 10% after 2 days, and eventually all cells died (Fig. 7 c, d). Correspondingly, the HOOH levels were kept below 1 µM in the cultures with higher cell inoculum concentrations (data not shown), but rapidly increased in the cultures with the lowest inoculum concentration, and reached 30 µM at the end of culture, showing a similar value as in BG-11[co] medium (Fig. 8 ). Although the unicellular cyanobacterium S. elongatus PCC 7942 possesses enzymes including ascorbate peroxidase and catalase to help to breakdown HOOH [ 41 ], the S. elongatus culture at the lowest cell inoculum was unable to prevent HOOH accumulation, likely resulting in the death of all S. elongatus cells. Fig. 6 HOOH formation in illuminated BG-11[co] medium containing various concentration of HEPES. All data are averages of biological triplicates ± standard deviation \n Fig. 7 Growth of S. elongatus with/without yeast R. glutinis . a Initial cell number of S. elongatus was 1.6 × 10 7 /mL; b Initial cell number of S. elongatus was 3.2 × 10 6 /mL; c Initial cell number of S. elongatus was 1.6 × 10 6 /mL; d Live fraction of S. elongatus cells with initial cell number of 1.6 × 10 6 /mL. All data are averages of biological triplicates ± standard deviation \n Fig. 8 HOOH formation in HEPES buffered culture. Initial cell number of S. elongatus and R. glutinis was 1.6 × 10 6 and 2 × 10 5 /mL, respectively. 1.25 mg/L of catalase was daily supplied. All data are averages of biological triplicates ± standard deviation \n Catalase was added in order to limit the impact of HOOH in pure cultures of S. elongatus . Contrary to BG-11[co] medium and in the pure cultures of S. elongatus , HOOH was kept at minimal concentrations following addition of catalase (Fig. 8 ). Importantly, S. elongatus grew better in the cultures with catalase, especially at an initial inoculum of 1.6 × 10 6 cells/mL, while maintaining high viabilities (Fig. 7 c, d). Also, catalase supplied to the axenic cultures of S. elongatus with higher inoculum concentrations of 1.6 × 10 7 and 3.2 × 10 6 cells/mL, exhibited a slightly enhancement in the growth and cell density during experimental growth. Next, the impact of co-culturing with R. glutinis on S. elongatus growth and HOOH was investigated. The presence of the yeast R. glutinis partner clearly facilitated the growth of S. elongatus in co-cultures consistent with effects seen in dilute co-culture of cyanobacteria with other heterotrophs including E. coli, B. subtilis, and S. cerevisiae [ 22 ]. Cell numbers of S. elongatus with higher inoculum concentrations were significantly augmented in co-culture with R. glutinis compared with the axenic culture (Fig. 7 a, b), even more than catalase alone. The growth enhancement was superior to the addition of catalase alone, suggesting that R. glutinis may be providing factors beyond HOOH reduction to the S. elongatus in co-culture. Even at the lowest inoculum concentration, the S. elongatus gradually grew in co-culture, albeit with a prolonged lag phase (Fig. 7 c). Initially, the live fraction of S. elongatus cells decreased to approximately 60% by 3 days, but then gradually recovered to nearly 100% (Fig. 7 d). At the same time, the HOOH content was progressively reduced and efficiently minimized to a similar low level as in BG-11[co] medium with catalase (Fig. 8 ). The ability of pure cultures of S. elongatus to grow was strongly dependent upon the initial cell density, yet there was effectively no density dependence when catalase was supplied. These results along with the corresponding HOOH measurements indicated that HOOH accumulation in the illuminated cultures is likely to be a major cause of the toxicity effect during axenic growth of S. elongatus . However, S. elongatus in co-culture with R. glutinis could grow even at a dilute initial cell concentration, and the HOOH accumulation in the culture was efficiently reduced, suggesting that the eukaryotic heterotroph R. glutinis plays a key role in enabling its phototrophic partner’s growth by scavenging extracellular HOOH. Previous results have shown that dilute cultures of the marine cyanobacterium Prochlorococcus can be supported when co-cultured with certain marine heterotrophic bacteria [ 42 , 43 ]. Furthermore, R. glutinis significantly boosts the growth of S. elongatus at high initial cell densities (Fig. 7 ), and the augmentation was higher than that obtained with addition of catalase. This increase indicates that the eukaryotic heterotroph may provide additional benefits to phototrophic partners in addition to HOOH elimination in order to support the growth of cyanobacteria even more. Previous studies have shown that the addition of thiosulfate, vitamin B12, biotin, and thiamine can be used to enhance and elevate the growth of the cyanobacterium Synechococcus leopoliensis CCAP1405/1 [ 44 ]."
} | 6,012 |
32555454 | PMC7332420 | pmc | 1,196 | {
"abstract": "Many corals harbour symbiotic dinoflagellate algae. The algae live inside coral cells in a specialized membrane compartment known as the symbiosome, which shares the photosynthetically fixed carbon with coral host cells while host cells provide inorganic carbon to the algae for photosynthesis 1 . This endosymbiosis—which is critical for the maintenance of coral reef ecosystems—is increasingly threatened by environmental stressors that lead to coral bleaching (that is, the disruption of endosymbiosis), which in turn leads to coral death and the degradation of marine ecosystems 2 . The molecular pathways that orchestrate the recognition, uptake and maintenance of algae in coral cells remain poorly understood. Here we report the chromosome-level genome assembly of a Xenia species of fast-growing soft coral 3 , and use this species as a model to investigate coral–alga endosymbiosis. Single-cell RNA sequencing identified 16 cell clusters, including gastrodermal cells and cnidocytes, in Xenia sp. We identified the endosymbiotic cell type, which expresses a distinct set of genes that are implicated in the recognition, phagocytosis and/or endocytosis, and maintenance of algae, as well as in the immune modulation of host coral cells. By coupling Xenia sp. regeneration and single-cell RNA sequencing, we observed a dynamic lineage progression of the endosymbiotic cells. The conserved genes associated with endosymbiosis that are reported here may help to reveal common principles by which different corals take up or lose their endosymbionts."
} | 390 |
38206123 | PMC10832554 | pmc | 1,198 | {
"abstract": "Abstract Anthranilate and its derivatives are important basic chemicals for the synthesis of polyurethanes as well as various dyes and food additives. Today, anthranilate is mainly chemically produced from petroleum‐derived xylene, but this shikimate pathway intermediate could be also obtained biotechnologically. In this study, Corynebacterium glutamicum was engineered for the microbial production of anthranilate from a carbon source mixture of glucose and xylose. First, a feedback‐resistant 3‐deoxy‐arabinoheptulosonate‐7‐phosphate synthase from Escherichia coli , catalysing the first step of the shikimate pathway, was functionally introduced into C. glutamicum to enable anthranilate production. Modulation of the translation efficiency of the genes for the shikimate kinase ( aroK ) and the anthranilate phosphoribosyltransferase ( trpD ) improved product formation. Deletion of two genes, one for a putative phosphatase ( nagD ) and one for a quinate/shikimate dehydrogenase ( qsuD ), abolished by‐product formation of glycerol and quinate. However, the introduction of an engineered anthranilate synthase (TrpEG) unresponsive to feedback inhibition by tryptophan had the most pronounced effect on anthranilate production. Component I of this enzyme (TrpE) was engineered using a biosensor‐based in vivo screening strategy for identifying variants with increased feedback resistance in a semi‐rational library of TrpE muteins. The final strain accumulated up to 5.9 g/L (43 mM) anthranilate in a defined CGXII medium from a mixture of glucose and xylose in bioreactor cultivations. We believe that the constructed C. glutamicum variants are not only limited to anthranilate production but could also be suitable for the synthesis of other biotechnologically interesting shikimate pathway intermediates or any other aromatic compound derived thereof.",
"introduction": "INTRODUCTION Aromatic compounds have a wide range of applications in the chemical, pharmaceutical or food industries where they serve as polymer building blocks, dyes, food additives or antibiotics (Balderas‐Hernández et al., 2009 ; Chung, 2016 ; Noreen et al., 2016 ). However, the chemical production of these compounds at an industrial scale is typically based on benzene, toluene or xylene (BTX) derived from crude oil (Haveren et al., 2007 ). One of the key aromatic platform chemicals is aniline, a building block for polyurethane (PU) synthesis utilized in the manufacturing of foam, elastomers, paints, adhesives or artificial leather, with a global market volume of 9.4 million tons in 2022, which is expected to increase to more than 12 million tons by 2028 with an annual growth rate of 5.2% (Driessen et al., 2017 ; Nakajima‐Kambe et al., 1999 ). Aniline is mainly produced chemically from benzene, which is initially converted to nitrobenzene with nitric acid (Bolden et al., 2015 ). However, aniline could be also produced in a more sustainable manner using the metabolic capabilities of microorganisms. This biotechnological process involves the microbial production of anthranilate (ANT), which can be subsequently converted to aniline by chemical decarboxylation. Besides aniline, ANT is also the precursor of other aromatic compounds of biotechnological interest such as methyl anthranilate (MANT) conveying the typical scent and taste of the Concord grape ( Vitis labrusca ) (Chambers et al., 2013 ; Fuleki, 1972 ; Wang & De Luca, 2005 ). Another derivative is menthyl anthranilate, which serves as a UV‐A quencher in sunscreens (Beeby & Jones, 2000 ). ANT is an intermediate of the shikimate (SA) pathway, which is responsible for the synthesis of the three aromatic amino acids (TYR, PHE, TRP), as well as secondary metabolites in plants, fungi and many microorganisms (Lee & Wendisch, 2017 ) (Figure 1 ). FIGURE 1 Biosynthesis of anthranilate from glucose and xylose in C. glutamicum . acetyl‐CoA, acetyl‐coenzyme A; ANS, anthranilate synthase; ANT, anthranilate; CHO, chorismate; DAHPS, 3‐deoxy‐arabinoheptulosonate‐7‐phosphate (DAHP) synthase; DHAP, dihydroxyacetone phosphate; DHQ, 3‐dehydroquinate; DHS, 3‐dehydroshikimate; DQD, 3‐dehydroquinate dehydratase; DQS, 3‐dehydroquinate synthase; E4P, erythrose‐4‐phosphate; F1, 6bP, fructose‐1, 6‐bisphosphate; F6P, fructose‐6‐phosphate; G3P, glyceraldehyde‐3‐phosphate; G6P, glucose‐6‐phosphate; PEP, phosphoenolpyruvate; PHE, phenylalanine; QA/SA DH, quinate/shikimate dehydrogenase; R5P, ribose‐5‐phosphate; Ru5P, ribulose‐5‐phosphate; SA, shikimate; SAK, shikimate kinase; S3P, shikimate‐3‐phosphate; S7P, sedoheptulose‐7‐phosphate; TAL, transaldolase; TCA cycle, tricarboxylic acid cycle; TKT, transketolase; TRP, tryptophan; TYR, tyrosine; X5P, xylulose‐5‐phosphate. The first step in the synthesis of aromatic compounds via the SA pathway is the condensation of erythrose‐4‐phosphate (E4P), a metabolite of the non‐oxidative part of the pentose phosphate pathway (PPP), and phosphoenolpyruvate (PEP), an intermediate of glycolysis, yielding 3‐deoxy‐arabinoheptulosonate‐7‐phosphate (DAHP) (Herrmann & Weaver, 1999 ). The reaction is catalysed by DAHP synthase, which is allosterically regulated (Liao et al., 2001 ; Shumilin et al., 1996 ). After seven enzymatic conversions, the SA pathway bifurcates at the stage of chorismate, eventually leading to PHE, TYR and TRP. The first committed step towards TRP biosynthesis is catalysed by the anthranilate synthase (ANS) converting chorismate to ANT and pyruvate, using GLN as an amino donor (yielding GLU) (Romero et al., 1995 ). The microbial production of ANT was established in a few bacterial species. A Pseudomonas putida KT2440 strain accumulated 1.54 g/L (11.23 mM) ANT in a defined medium with glucose as a carbon source using a fed‐batch process (Kuepper et al., 2015 ). In addition, a plasmid‐free strain of P. putida was constructed, which produced 3.8 mM ANT in shake flasks from glucose without the supplementation of additives (Fernández‐Cabezón, 2022 ). Higher ANT titers were only achieved by using expensive complex media. A Bacillus subtilis strain produced 3.5 g/L (25 mM) ANT using yeast extract and glucose as substrates within 60 h, whereas an engineered Escherichia coli variant accumulated 14 g/L (102 mM) ANT after 34 h in M9 minimal medium supplemented with 10 g/L of yeast extract and 30 g/L glucose followed by two pulses containing the same carbon source composition (Balderas‐Hernández et al., 2009 ; Cooper et al., 1995 ). Moreover, a titre of 567.9 mg/L (4.1 mM) ANT was achieved with Saccharomyces cerevisiae using a complex SCD medium, which is the highest ANT titre reached using a eukaryotic host (Kuivanen et al., 2021 ). Furthermore, MANT was produced from ANT with C. glutamicum using fed‐batch cultures and a two‐phase cultivation mode with a maximum titre of 5.74 g/L (38 mM) MANT (Luo et al., 2019 ). In the context of this study, the first ANT production with C. glutamicum was described. The best variant accumulated up to 26.4 g/L (192.5 mM) ANT. However, to achieve such high product titers, the ANT‐producing C. glutamicum strain was engineered to be auxotrophic for TRP and thus required TRP supplementation every 24 h during cultivation. ANT, similar to most other aromatics, is known to have antimicrobial properties, potentially limiting the microbial production of this compound (Li et al., 2017 ). C. glutamicum is used for the industrial production of proteinogenic amino acids such as glu and lys at a scale of more than 5 million tons per year (Eggeling & Bott, 2005 ). This gram‐positive soil bacterium is free of endotoxins and products derived from C. glutamicum are therefore generally recognized as safe (Zha et al., 2018 ). In addition, several C. glutamicum variants engineered for aromatic compounds such as hydroxybenzoic acids, phenylpropanoids or plant polyphenols are available demonstrating that C. glutamicum is a robust host system capable of enduring increased concentrations of cytotoxic aromatics (Kallscheuer et al., 2016 ; Kallscheuer & Marienhagen, 2018 ; Milke et al., 2019 ; Son et al., 2021 ). In addition to a wide product range, C. glutamicum can metabolize various cheap carbon sources. Frequently, glucose and fructose are used as carbon and energy sources for applied purposes (Blombach & Seibold, 2010 ). However, as the utilization of these sugars in biotechnological processes competes with food production, it is highly desirable to resort to other substrates, preferably from waste streams, such as cellobiose, arabinose and xylose (Jiang & Zhang, 2016 ; Kawaguchi et al., 2006 ; Kotrba et al., 2003 ; Schneider et al., 2011 ). Since C. glutamicum cannot utilize xylose naturally, several different pathways such as the isomerase pathway or the Weimberg pathway were implemented into the metabolism of this bacterium (Kawaguchi et al., 2006 ; Meiswinkel et al., 2013 ; Radek et al., 2014 ). In this context, C. glutamicum was engineered to produce several compounds such as protocatechuate, succinate or α‐ketoglutarate from xylose or glucose/xylose mixtures (Brüsseler et al., 2019 ; Labib et al., 2021 ; Tenhaef et al., 2021 ). Noteworthy, xylose utilization via the isomerase pathway has the advantage of providing the SA pathway precursor E4P (Kogure et al., 2016 ). This study aims to establishing microbial ANT production with C. glutamicum from different carbon sources (and mixtures) by combining rational metabolic engineering strategies with biosensor‐guided semi‐rational enzyme engineering.",
"discussion": "DISCUSSION In this study, the microbial production of ANT with C. glutamicum was established, employing rational and semi‐rational engineering strategies. Initial experiments on product toxicity were performed due to known antimicrobial properties of ANT (Csonka, 1989 ; Li et al., 2017 ). However, only a minor effect of ANT on the growth of C. glutamicum was observed, making this organism a more suitable host for the production of ANT than P. putida , which cannot grow at ANT concentrations exceeding 10 g/L (72.92 mM) (Kuepper et al., 2015 ). In order to avoid the formation of possible by‐products, the genes nagD and qsuD were deleted, whose gene products are involved in the synthesis of glycerol and QA (Jojima et al., 2012 ; Teramoto et al., 2009 ). Single deletion of nagD or qsuD had no impact on ANT production, only the double deletion increased the product titre significantly. Deletion of nagD might lead to increased availability of NADPH, which in turn could be used by qsuD ‐encoded NADPH‐dependent QA/SA dehydrogenase to form QA, resulting in no increased ANT titre (Jojima et al., 2012 ). Glycerol and QA both could not be detected in the culture supernatant by HPLC although the formation of glycerol has been already described during SA‐synthesis with C. glutamicum (Kogure et al., 2016 ). However, in the context of this particular study, glycerol accumulation was only detectable during cultivations in which SA titers exceeding 100 g/L were obtained. Hence, glycerol and QA concentrations present in the culture supernatants of the C. glutamicum strains constructed in this study might have been below the detection limit. Initially, low ANT titers obtained in cultivations with glucose as the sole carbon and energy source could be markedly increased when using glucose/xylose mixtures since xylose is rapidly converted to E4P via the isomerase pathway and reactions of the PPP (Radek et al., 2014 ). This in turn increased the carbon flux through the SA pathway, which was also evident from the observed accumulation of the central SA pathway intermediate SA. Utilization of this mixed carbon substrate allowed for the production of ANT at g/L scale, outperforming ANT synthesis with P. putida (1.5 g/L (10 mM) ANT) (Kuepper et al., 2015 ). The maximum ANT titre reached in pulsed‐fed‐batch bioreactors with C. glutamicum was 5.9 ± 1.4 g/L (43.3 mM), which exceeded the ANT titre determined for B. subtilis (3.5 g/L (26 mM)), but was lower compared to previous experiments with E. coli (14 g/L (102 mM)) (Balderas‐Hernández et al., 2009 ; Cooper et al., 1995 ). However, these high product titers were achieved in complex media limiting the comparability of the obtained results. Noteworthy, we observed an increased product formation in cultivations with glucose as the sole carbon source when genes for the heterologous isomerase pathway for xylose utilization were episomally expressed. In general, xylose isomerases are known to have a rather broad substrate spectrum also accepting glucose as a substrate, which is isomerized to fructose (Blow et al., 1992 ; Tsumura & Sato, 1965 ). Subsequently, fructose is exported and re‐imported by the fructose‐specific phosphotransferase system yielding fructose‐1‐phosphate, which then enters glycolysis after phosphorylation to fructose‐1,6‐bisphosphate (Dominguez & Lindley, 1996 ; Ikeda, 2012 ). Thus, the xylose isomerase side activity might increase glucose utilization via glycolysis in the performed cultivations. During previous engineering of E. coli , P. putida and C. glutamicum for ANT overproduction, trpD was deleted to increase ANT accumulation (Balderas‐Hernández et al., 2009 ; Kuepper et al., 2015 ; Luo et al., 2019 ). However, we favoured the replacement of the ATG start codon of trpD by GTG to reduce the translation efficiency to avoid any requirement for TRP supplementation, which is inevitable when deleting this essential gene. As was the case with the DAHP synthase, removing the feedback inhibition of the key enzyme ANS was one of the decisive factors in improving product formation. A similar strategy was pursued regarding the microbial synthesis of ANT with P. putida , expressing trpE \n \n S40 \n \n G from E. coli , which was achieved via plasmid‐based expression (Kuepper et al., 2015 ). The performed biosensor‐based screening of focused libraries for ANS variants with further reduced feedback inhibition showed that the obtained S38A/G substitutions had the same effect as the previously described S38R substitution, which confirms the assumption that the hydroxyl group of the native SER residue is essential for TRP binding (Matsui et al., 1987 ). The other amino acid substitution leading to relieved TRP‐mediated feedback inhibition was C461G. In combination with S38A or S38G, this novel amino acid substitution had a positive additive effect on ANT accumulation, indicating that the degree of feedback resistance can be gradually adjusted by substitutions of more than one key residue. Batch and pulsed‐fed‐batch cultivations of C. glutamicum ANT6 showed an interesting diauxic growth/production behaviour during which glucose was first metabolized and used for biomass formation. Subsequently, xylose was consumed and channelled towards ANT production. Noteworthy, C. glutamicum is known for its ability to cometabolize glucose with acetate, lactate or fructose among others (Cocaign et al., 1993 ; Dominguez et al., 1997 , 1998 ). The only known examples of diauxic substrate consumption of C. glutamicum are the sequential metabolism of glucose before glutamate and glucose before ethanol (Arndt et al., 2008 ; Krämer et al., 1990 ). However, in case of the engineered catabolization of the xylose via the isomerase pathway, a preferential utilization of glucose over xylose has been already observed several times in different C. glutamicum species (Kawaguchi et al., 2006 ; Radek et al., 2014 ). A possible reason for this could be a side activity of the xylose isomerases already discussed in the context of the improved ANT production in cultivations with glucose when the genes for the isomerase pathway were episomally expressed. In the case of the preferential glucose utilization observed during bioreactor cultivations, a competitive competition of glucose and xylose for the binding sites of the xylose isomerases might delay efficient xylose utilization. Noteworthy, the formation of the by‐products such as SA pathway intermediates DAHP and SA, but also xylonate and xylitol was observed during batch and fed‐batch cultivations. Previously, the endogenous myo ‐inositol dehydrogenase IolG was identified to catalyse the oxidation of xylose to xylonate and could be responsible for the observed xylonate accumulation (Tenhaef et al., 2018 ). In case of xylitol, the enzyme(s) with xylose reductase activity converting xylose to xylitol is still unknown (Radek et al., 2016 ). Taken together, by combining rational strain engineering and biosensor‐guided semi‐rational enzyme engineering approaches, an attractive C. glutamicum strain for ANT production from glucose/xylose mixtures could be constructed. This variant represents a promising starting point for the construction of other production strains for a broader spectrum of aromatic compounds of commercial interest."
} | 4,236 |
30820745 | PMC6718569 | pmc | 1,199 | {
"abstract": "The ubiquitous chlorophyll a (Chl a ) pigment absorbs both blue and red light. Yet, in contrast to green algae and higher plants, most cyanobacteria have much lower photosynthetic rates in blue than in red light. A plausible but not yet well-supported hypothesis is that blue light results in limited energy transfer to photosystem II (PSII), because cyanobacteria invest most Chl a in photosystem I (PSI), whereas their phycobilisomes (PBS) are mostly associated with PSII but do not absorb blue photons. In this paper, we compare the photosynthetic performance in blue and orange-red light of wildtype Synechocystis sp. PCC 6803 and a PBS-deficient mutant. Our results show that the wildtype had much lower biomass, Chl a content, PSI:PSII ratio and O 2 production rate per PSII in blue light than in orange-red light, whereas the PBS-deficient mutant had a low biomass, Chl a content, PSI:PSII ratio, and O 2 production rate per PSII in both light colors. More specifically, the wildtype displayed a similar low photosynthetic efficiency in blue light as the PBS-deficient mutant in both light colors. Our results demonstrate that the absorption of light energy by PBS and subsequent transfer to PSII are crucial for efficient photosynthesis in cyanobacteria, which may explain both the low photosynthetic efficiency of PBS-containing cyanobacteria and the evolutionary success of chlorophyll-based light-harvesting antennae in environments dominated by blue light.",
"introduction": "Introduction Cyanobacteria are the oldest known oxygen-producing photosynthetic organisms on our planet and their photosynthetic activity is widely held responsible for oxygenation of the Earth’s atmosphere about 2.3 billion years ago (Holland 2002 ; Schirrmeister et al. 2015 ). The ubiquitous chlorophyll a (Chl a ) pigment in the photosystems of oxygenic phototrophs absorbs both blue (440 nm) and red light (678 nm), and hence one would expect that these two colors of light are used at approximately equal efficiency. Yet, in contrast to green algae and higher plants, cyanobacteria appear to have much lower oxygen (O 2 ) production and growth rates in blue light than in red light (e.g., Lemasson et al. 1973 ; Pulich and van Baalen 1974 ; Wilde et al. 1997 ; Tyystjärvi et al. 2002 ; Wang et al. 2007 ; Singh et al. 2009 ; Luimstra et al. 2018 ). Why cyanobacteria perform less well in blue light is not yet fully resolved. It is often hypothesized that the distribution of absorbed light energy between photosystem I (PSI) and photosystem II (PSII) plays a key role (Fujita 1997 ; El Bissati and Kirilovsky 2001 ; Wang et al. 2007 ; Singh et al. 2009 ; Solhaug et al. 2014 ; Kirilovsky 2015 ; Luimstra et al. 2018 ). To optimize absorption of light energy, cyanobacteria use phycobilisomes (PBS) as light-harvesting antennae, which transfer absorbed light energy to the photosystems (Grossman et al. 1993 ; Tandeau de Marsac 2003 ). Most of the PBS are typically associated with PSII (state 1), but PBS can be relocated to PSI (state 2) at time scales of seconds to minutes, or they can be detached from the photosystems in which case they do not contribute to the photosynthetic activity of the cells (Joshua et al. 2005 ; Mullineaux 2014 ; Kirilovsky 2015 ). Blue light ≤ 450 nm is very poorly absorbed by PBS (Tandeau de Marsac 2003 ; Six et al. 2007 ), however, and hence PBS cannot distribute the absorbed light energy over the two photosystems. Instead, blue light is directly absorbed by chlorophyll and carotenoids associated with the two photosystems. Cyanobacteria contain ~ 100 molecules of Chl a per PSI (Jordan et al. 2001 ) but only ~ 35 molecules of Chl a per PSII (Umena et al. 2011 ). Moreover, the PSI:PSII ratio of cyanobacteria usually ranges between 5:1 and 2:1, depending on the growth conditions, which is higher than the approximately 1:1 ratio often found in eukaryotic phototrophs (Shen et al. 1993 ; Fujita 1997 ; Olive et al. 1997 ; Singh et al. 2009 ; Allahverdiyeva et al. 2014 ). Since cyanobacteria usually invest most of their Chl a in PSI (e.g., Myers et al. 1980 ; Fujita 1997 ; Luimstra et al. 2018 ) and, in cyanobacteria, only the carotenoids in PSI are involved in light harvesting (Stamatakis et al. 2014 ), PSI will absorb more blue photons than PSII. This implies that blue light is likely to cause an excitation imbalance between both photosystems, with low O 2 production by PSII and limited linear electron transport (Fujita 1997 ; Solhaug et al. 2014 ; Kirilovsky 2015 ), which then explains the low photosynthetic efficiency and growth rate of cyanobacteria in blue light (Luimstra et al. 2018 ). However, although many observations support this hypothesis, conclusive evidence is still limited. In this paper, we aim to contribute to a further understanding of why cyanobacteria display a low photosynthetic efficiency in blue light. Therefore, we compare the photosynthetic efficiency in blue and orange-red light of Synechocystis sp. PCC 6803 and a mutant strain lacking PBS known as the PAL mutant (Ajlani and Vernotte 1998 ). The PAL mutant does not have light-harvesting antennae that are able to redistribute light energy over both photosystems. Consequently, our expectation is that the PAL mutant will have limited energy transfer to PSII irrespective of the light color and hence will display a similarly low photosynthetic efficiency in both orange-red and blue light as the wildtype in blue light. An advanced understanding of the light-color dependence of cyanobacterial photosynthesis may contribute to an improved design of successful culture conditions in biotechnological applications, and to further clarification of the ecological distributions of cyanobacteria in waters dominated by different wavelengths.",
"discussion": "Discussion What happens if phycobilisomes cannot be used? Our results show that Synechocystis sp. PCC 6803 wildtype displays much lower O 2 and biomass production rates in blue light than in orange-red light, in agreement with previous studies (e.g., Wilde et al. 1997 ; El Bissati and Kirilovsky 2001 ; Tyystjärvi et al. 2002 ; Singh et al. 2009 ; Bland and Angenent 2016 ; Luimstra et al. 2018 ). We tested the hypothesis that cyanobacteria have a low photosynthetic efficiency in blue light because PBS do not absorb wavelengths ≤ 450 nm, and hence blue light mainly excites PSI while fewer photons excite the PSII reaction center (Myers et al. 1980 ; Fujita 1997 ; Solhaug et al. 2014 ; Kirilovsky 2015 ). This results in an excitation imbalance between the two photosystems in blue light (Luimstra et al. 2018 ). The PAL mutant lacks PBS. Therefore, if our hypothesis is correct, then the PAL mutant should have (i) a similar low biomass and O 2 production rate and (ii) a similar low PSI:PSII ratio in both orange-red and blue light as the wildtype in blue light. This reasoning is confirmed by our experiments, in which the PAL mutant indeed displayed a similar low biomass (Fig. 1 c), cell production (Fig. 1 d), Chl a content (Fig. 1 e), PSI:PSII ratio (Fig. 2 c,d; Table 1 ), and O 2 production rate per PSII (Fig. 3 ) in both orange-red and blue light as the wildtype in blue light. In other words, the poor photosynthetic performance of PBS-containing cyanobacteria in blue light was similar to the performance of a PBS-deficient mutant in both light colors. We note that our results also show some differences between the wildtype in blue light and the PAL mutant. In particular, the photophysiology of the PAL mutant shows enhanced absorption in the 400–500 nm range (Fig. 2 b), indicative of enhanced carotenoid production (see also Ajlani and Vernotte 1998 ; Kwon et al. 2013 ) and an even lower PSI:PSII ratio than the wildtype in blue light (Fig. 2 d). Hence, in a sense, the PAL mutant lacking PBS might be interpreted as a more extreme phenotype in comparison to the wildtype that contains PBS but cannot use them in blue light. The low growth rate and low PSI:PSII ratio of the PAL mutant have also been described by many other studies (Ajlani and Vernotte 1998 ; El Bissati and Kirilovsky 2001 ; Bernát et al. 2009 ; Stadnichuk et al. 2009 ; Collins et al. 2012 ; Kwon et al. 2013 ; Liberton et al. 2017 ). Comparison of different antenna mutants revealed that the PSI:PSII ratio decreases with a decrease in antenna size of the PBS, as a response to the diminished energy transfer from PBS to PSII when the antenna size is gradually reduced (Ajlani et al. 1995 ; Olive et al. 1997 ; Bernát et al. 2009 ; Stadnichuk et al. 2009 ; Collins et al. 2012 ; Kwon et al. 2013 ; Liberton et al. 2017 ). Likewise, Page et al. ( 2012 ) reported that growth rates were highest in the Synechocystis PCC 6803 wildtype, and decreased progressively with a decrease in antenna size of a series of antenna mutants. We also note the presence of the fluorescence peak at 685 nm in the 77 K fluorescence spectra of the PAL mutant (Fig. 2 d), which has also been found in previously published 77 K spectra of the PAL mutant (Ajlani and Vernotte 1998 ; Bernát et al. 2009 ; Stadnichuk et al. 2009 ; Collins et al. 2012 ; Kwon et al. 2013 ). This peak, which is commonly attributed to the chlorophyll-binding protein CP43 in PSII (Andrizhiyevskaya et al. 2005 ; Wilson et al. 2007 ; Brecht et al. 2014 ), is also present in the wildtype in blue light (Fig. 2 c). Hence, our findings with the PAL mutant are well in line with many previous studies. The novelty of our results is the recognition that, in many respects, the phenotype of the PAL mutant resembles the phenotype of the wildtype strain in blue light. This is further confirmed by our analysis of the oxygen production rates of the wildtype and PAL mutant, as measured by MIMS (Figs. 3 , 4 ). It has been shown that deletion of PBS in the PAL mutant is accompanied by profound changes in the proteome in comparison to the wildtype (Kwon et al. 2013 ; Liberton et al. 2017 ). Kwon and colleagues report a major increase of PSII proteins accompanied by a slight decrease of PSI proteins in the PAL mutant. Furthermore, they found that the PAL mutant has increased abundances of proteins involved in stress responses to high light and in carbohydrate metabolism (mostly gluconeogenesis). More recent proteome analyses by Liberton et al. ( 2017 ) support these findings, and in addition these authors found a decrease in bicarbonate transport proteins and an increase in transport proteins involved in nitrate uptake in the PAL mutant. Interestingly, the physiology of the wildtype is also completely changed when transferred from white to blue light (Singh et al. 2009 ). Transcriptome analysis by Singh et al. revealed that Synechocystis sp. PCC 6803 wildtype cells grown in blue light showed, among others, an increased expression of genes encoding PSII subunits and proteins involved in stress responses to high light, carbohydrate metabolism, and nitrate uptake, which resembles several of the proteome results in the PAL mutant (Kwon et al. 2013 ; Liberton et al. 2017 ). This indicates again that the PAL mutant and wildtype in blue light share many similarities, not only in their photophysiological traits but also in terms of their cellular metabolism. Implications for biotechnology Several studies have argued that truncation of light-harvesting antennae may be an advantage in crop and algal biomass production for, e.g., biotechnological applications (see, e.g., Ort and Melis 2011 ; Work et al. 2012 ; de Mooij et al. 2015 ; Kirst et al. 2017 ). In dense algal cultures or plant canopies, cells near the surface absorb excessive amounts of light energy that they largely dissipate as heat, while cells deeper down in the culture receive insufficient light for photosynthesis (Melis 2009 ; Blankenship and Chen 2013 ). As a consequence, much of the absorbed light energy is wasted rather than invested in biomass production. Truncation of the light-harvesting antennae reduces absorption per individual cell and thereby distributes light more evenly among the cells in dense algal cultures. It has indeed been demonstrated that antenna truncation can increase the biomass production of green algae (e.g., Polle et al. 2002 ; Perrine et al. 2012 ; de Mooij et al. 2015 ; Shin et al. 2016 ). Our results indicate that truncation of the PBS of cyanobacteria has other effects than the truncation of the chlorophyll-based light-harvesting complexes of green algae. Green algae generally maintain a PSI:PSII ratio of around 1:1, in contrast to cyanobacteria which usually have a much higher PSI:PSII ratio (Shen et al. 1993 ; Singh et al. 2009 ; Kirilovsky 2015 ) and mainly use their PBS to direct sufficient light energy to PSII. Therefore, truncation of PBS in cyanobacteria will have a stronger effect on light redistribution between the two photosystems than truncation of the light-harvesting complex in green algae. In particular, our findings show that complete deletion of the PBS of cyanobacteria, in the PAL mutant, yields an excitation imbalance between the two photosystems, which results in a low PSII activity (Fig. 3 ) and low biomass production (Fig. 1 ) in both blue and orange-red light. This matches earlier results of, e.g., Ajlani and Vernotte ( 1998 ), Bernát et al. ( 2009 ), Page et al. ( 2012 ), and Kwon et al. ( 2013 ), who also found that the PAL mutant has a lower biomass production than the wildtype. Possibly, partial truncation of the antennae by shortening or removing the rods of the PBS might improve the biomass production of cyanobacteria, as the PBS can then still distribute the absorbed light energy over the two photosystems while overall absorbance per cell is lower (Lea-Smith et al. 2014 ). Partial truncation of the PBS likely implies that less light will be transferred to PSII, which must be compensated by a decreased PSI:PSII ratio, to maintain the balance between the two photosystems. Several cyanobacterial studies show that partial truncation mutants indeed decreased their PSI:PSII ratio progressively and proportionally to a decrease in antenna size (Ajlani et al. 1995 ; Olive et al. 1997 ; Bernát et al. 2009 ; Stadnichuk et al. 2009 ; Collins et al. 2012 ; Kwon et al. 2013 ; Liberton et al. 2017 ). Although most of these mutants did not show a higher productivity than the wildtype, removal of the phycocyanin rods but not the allophycocyanin core in the Olive mutant (Rögner et al. 1990 ) resulted in an enhanced cyanobacterial productivity at high light intensities (Kwon et al. 2013 ) and low-carbon conditions (Lea-Smith et al. 2014 ). Implications for cyanobacterial photosynthesis in the blue ocean Marine cyanobacteria of the Synechococcus genus often use the phycobilin pigments phycoerythrobilin (PEB) and phycourobilin (PUB) to absorb wavelengths in the green part (525–575 nm) and blue-green part (475–515 nm) of the light spectrum (Six et al. 2007 ; Grébert et al. 2018 ). These phycobilin pigments do not absorb photons in the violet-blue part of the spectrum (≤ 450 nm), and hence, similar to Synechocystis PCC 6803, marine Synechococcus will be unable to use their PBS to harvest blue photons ≤ 450 nm for PSII (Luimstra et al. 2018 ). Consequently, Synechococcus usually thrives in the near-surface layers of the oceans and in coastal waters, where green and blue-green light is available (Partensky et al. 1999 ; Scanlan and West 2002 ; Stomp et al. 2007a ). Cyanobacteria of the genus Prochlorococcus abound in the subtropical ocean gyres and usually their populations extend deeper down in the water column than Synechococcus (Partensky et al. 1999 ; Scanlan and West 2002 ). They use a different light-harvesting strategy. Prochlorococcus does not deploy PBS, but has evolved light-harvesting antennae that effectively absorb blue light using divinyl-Chl a and b (Chisholm et al. 1992 ; Ting et al. 2002 ; Stomp et al. 2007b ). Prochlorococcus likely evolved from a PBS-containing cyanobacterium (Kettler et al. 2007 ; Scanlan et al. 2009 ), and some Prochlorococcus strains still have genes for the synthesis of phycoerythrin (Hess et al. 1996 ). Interestingly, our results show that the loss of PBS does not have major negative fitness consequences in habitats dominated by blue light, as the PAL mutant had a similarly low biomass and cell production in blue light as the wildtype. Furthermore, we note that the 685-nm peak in our 77 K fluorescence spectra, indicative of the chlorophyll-binding protein CP43, was strongly induced by blue light (Fig. 2 c) and in the PBS-deficient mutant (Fig. 2 d; see also Ajlani and Vernotte 1998 ; Kwon et al. 2013 ). The chlorophyll-binding proteins (Pcb) in the light-harvesting antennae of Prochlorococcus are closely related to CP43 (Chen and Bibby 2005 ). Our results therefore suggest that the loss of PBS and transformation from an ancestral chlorophyll-binding protein of the CP43 family to Pcb caused a major evolutionary transition, from the poor photosynthetic performance in blue light by PBS-based cyanobacteria to efficient utilization of blue light by the chlorophyll-based light-harvesting antennae of Prochlorococcus . Harnessing of blue light will have provided a major selective advantage for Prochlorococcus over PBS-containing cyanobacteria in the deep blue waters of the open ocean."
} | 4,359 |
26710854 | PMC4758251 | pmc | 1,200 | {
"abstract": "Microbial communities can show astonishing ecological and phylogenetic diversity. What is the role of pervasive horizontal gene transfer (HGT) in shaping this diversity in the presence of clonally expanding “killer strains”? Does HGT of antibiotic production and resistance genes erase phylogenetic structure? To answer these questions, we study a spatial eco-evolutionary model of prokaryotes, inspired by recent findings on antagonistic interactions in Vibrionaceae populations. We find toxin genes evolve to be highly mobile, whereas resistance genes minimize mobility. This differential gene mobility is a requirement to maintain a diverse and dynamic ecosystem. The resistance gene repertoire acts as a core genome that corresponds to the phylogeny of cells, whereas toxin genes do not follow this phylogeny and have a patchy distribution. We also show that interstrain HGT makes the emergent phylogenetic structure robust to selective sweeps. Finally, in this evolved ecosystem we observe antagonistic interactions between, rather than within, spatially structure subpopulations, as has been previously observed for prokaryotes in soils and oceans. In contrast to ascribing the diversification and evolution of microbial communities to clonal dynamics, we show that multilevel evolution can elegantly explain the observed phylogenetic structure and ecosystem diversity.",
"introduction": "Introduction Recent metagenomics studies have led to the realization that the commonly observed phenotypic diversity within microbial ecosystems is just the tip of the iceberg. High-throughput sequencing has revealed an immense variation at the level of DNA sequences ( Ochman et al. 2000 ) and ecological adaptation ( Preheim et al. 2011 ), even in narrow taxonomic groups. How this diversity comes about, and moreover how it is maintained, is still an open question. An important source of diversity in prokaryotes is the remarkable flexibility of their genome content ( Kolstø 1997 ; Snel et al. 2002 ; Koonin and Wolf 2008 ). This flexibility is not equal for all parts of the genome, as some genes are more dispensable than others. The well-conserved core genes generally encode basal cellular functions, and are complemented with a variable and more niche-specific set of accessory genes. The core genome tends to have a phylogeny that is consistent with that of the species. On the other hand, the accessory genome has a phylogeny that differs from this vertical inheritance pattern (e.g., Shapiro et al. 2012 ), showing that horizontal gene transfer (HGT) is an important process in shaping the flexibility and diversity of prokaryotes ( Fraser et al. 2009 ). Indeed, recent results have shown that the rate at which HGT happens is up to 2 orders of magnitude higher than per-gene point mutations ( Puigbò et al. 2014 ). These results indicate that the gene content of microbes can change in relatively short timescales. Populations of bacteria do not only display high diversity in gene repertoires but also at the phylogenetic level, as was recently shown for Vibrionaceae ( Preheim et al. 2011 ) and Streptomyces sp. ( Vetsigian et al. 2011 ). Strikingly, this phylogenetic diversity is maintained despite the presence of “superkiller” strains ( Vetsigian et al. 2011 ; Cordero et al. 2012 ). Vetsigian et al. (2011) explained this superkiller paradox in a model assuming high rates of de novo gene gain (e.g., HGT from a distant source), which creates an evolutionary arms race. However, the resulting clonal diversity and short evolutionary life span of lineages appear to fall short in explaining the observed phylogenetic diversity in wild populations. Surprisingly, the studies also show that killing happens less frequently between related individuals, that is, close kin are spared from killing. The frequency of killing is low up to a certain genetic threshold, after which it sharply increases. This observation once again suggests high levels of phylogenetic structure. The combination of high phenotypic diversity and high phylogenetic diversity suggests that adaptive genes sweep the population on their own, rather than resulting in genome-wide, clonal sweeps ( Shapiro et al. 2012 ). Furthermore, as de novo gene discovery fails to reproduce these patterns, internal HGT of locally present DNA is a likely candidate for the flexibility and diversity of prokaryotic populations. In Vibrio ordalii , the toxin production gene is part of a mobile genetic element (MGE) ( Cordero et al. 2012 ). Surprisingly, this toxin-production gene was not linked to the corresponding resistance factor, and can thus only transfer to individuals that already have the resistance factor in their genetic background. The resistance of V. ordalii was indeed shown to be ancestral, and other resistant-only individuals show no trace of the MGE. These results indicate that the resistance gene is part of the core genome, whereas the toxin gene is a mobile accessory gene which is frequently transferred within and potentially between populations of Vibrionaceae. This genome structure appears to prevent clonal dynamics, and impacts the observed diversity and interaction dynamics by means of gene-level selection. We here ask the question how local HGT impacts population diversity and the differential gene mobility as observed for the toxin and resistance genes in Vibrio’s. Furthermore, we wish to understand what selection pressures underlie this process, and how it shapes core-, accessory-, and pan-genomes. We investigate these questions by using a minimal model of microbial evolution, where cells undergo HGT of locally present DNA. The model includes three different levels: Genes, cells, and spatial patterns. Cells compete locally for space by means of antagonistic interactions of toxin and resistance genes. Cells stochastically take up genes from the local extracellular DNA (eDNA) pool. Genes encode their own mobility parameter, which represents the likelihood of integration into the genome after uptake. Antagonistic interactions take place between spatially patterned subpopulations. We focus on the evolution of gene mobility, and how this influences diversity and population structure. The aim is to reveal the intrinsic evolutionary pressure of gene mobility, abstracting away from the mechanisms underlying this process (e.g., uptake sequences [ Mell and Redfield 2014 ] or illegitimate recombination [ de Vries and Wackernagel 2002 ]). Our minimal model shows how HGT and the resulting interplay of gene- and cell-level selection can explain the observed ecosystem diversity and the different phylogenetic patterns of core and accessory genomes of natural prokaryotic populations.",
"discussion": "Discussion We showed that differential gene mobility evolves in a simple model where HGT occurs by incorporation of locally present eDNA. Toxin genes evolve high mobility, whereas resistance genes are mostly vertically inherited. This differential gene mobility is furthermore required to maintain a diverse population, as diversity collapses when mobility is equal for toxin and resistance genes. Moreover, for parameter values which do not evolve differential gene mobility ( supplementary fig. S3 , Supplementary Material online), the high diversity as observed in vivo does not occur. Importantly, the differential gene mobility appears to generate the within-species population structure and interaction dynamics as observed in vivo ( Vetsigian et al. 2011 ; Cordero et al. 2012 ; Cui et al. 2015 ). Our model shows that many observations on microbial ecosystems can be explained by local, internal HGT and the resulting interplay of gene- and cell-level evolution. Closed Ecosystems with High Diversity Interspecies HGT challenges our concept of species ( Zhaxybayeva and Doolittle 2011 ). On the other hand, within-species HGT on a more local scale is now challenging our expectations on the eco-evolutionary dynamics of prokaryotes ( Shapiro et al. 2012 ; Overballe-Petersen and Willerslev 2014 ; Takeuchi et al. 2014 ). We have shown in figure 8 that the dynamics of a closed ecosystem, that is, internal HGT of with a local pool of eDNA, is very different from open ecosystems. The closed ecosystem is very congruent with the observation of core-like resistomes in natural ecosystems ( Forsberg et al. 2014 ). Furthermore, despite the closed and relatively small ecosystem, adaptations such as toxin production can be present because of gene-level selection. This shows that diversity and the presence of superkillers ( Vetsigian et al. 2011 ; Cordero et al. 2012 ) do not necessarily contradict one another. Moreover, the diversity and dynamics of the closed ecosystem do not depend on high rates of de novo gene discovery, as in other modelling approaches ( Vetsigian et al. 2011 ; Cui et al. 2015 ), but on local HGT. Although observations in our model are strikingly similar to in vivo observations, bacteria in agricultural soil ( Heuer et al. 2011 ) or hospital wastewater ( Naiemi et al. 2005 ) might have the dynamics of an open ecosystem (e.g., as a result of anthropogenic influences). We have shown that increasing the external influx of genes results in very rapid dynamics of clonal expansions driven by the continuous discovery of novel genes ( fig. 8 and supplementary movie S3, Supplementary Material online). Although superficially this ongoing de novo gene discovery results in higher strain diversity, figure 8 shows that repeated genome-wide selective sweeps actually negatively impact phylogenetic diversity. In other words, our results suggest that, counterintuitive, genetic diversity is highest when the majority of available adaptations originates from the population itself. Earlier research has also challenged the idea diversity is reduced by genome-wide selective sweeps of beneficial adaptations. Shapiro et al. (2012) propose that HGT can actually result in gene-specific selective sweeps, and how the diversity of the population is not necessarily purged by this process. Indeed, Takeuchi et al. (2015) have studied a model of gene-specific sweeps of a generic beneficial adaptation through negative frequency-dependent selection. Here, we have also shown gene-specific sweeps, but now in a context in which the benefit of a gene depends on the local neighborhood and the genetic background of the host cell. As such, newly discovered toxin and resistance genes spread through the population on their own, maintaining phylogenetic diversity ( supplementary fig. S5 , Supplementary Material online). The evolutionary impact of HGT, and evolution of the other HGT-related processes, can be very different depending on the scale, the donor and recipient species, and the environment. Together with recent studies, our results show that microbial evolution is not driven by clonal dynamics, and that even small, closed niches can be expected to harbor a lot of genetic and phenotypic diversity. Rarely Beneficial Genes Evolve High Mobility The question of how and when HGT could be favorable is in most studies focused on the potential benefits of the cell ( Martinez 2009 ; Heuer et al. 2011 ). However, in our model we take into account both gene- and the cell-level selection, and show that it is actually the least beneficial genes, toxin genes, that evolve high mobility. As shown in figure 3 , as highly resistant individuals evolve, only toxin genes with mobility higher than their corresponding resistance gene are able to survive. We interpreted the low mobility of resistance in closed ecosystems as reduction of fitness cost of redundant genes, as individuals in the vicinity are already resistant. We have indeed shown that taking up resistance genes is rarely beneficial ( fig. 5 ). Congruent with these interpretations is the fact that that high diffusion of eDNA leads to increased gene mobility for both types of genes ( supplementary fig. S1 , Supplementary Material online), once again revealing how locality impacts the process of HGT. Despite the sparse presence of toxin producing individuals, resistance is sufficiently selected on the cell-level, whereas toxin genes need to evolve high gene mobility to survive. We tested the notion that less beneficial genes evolve higher gene mobility in a simple model, showing that there is indeed a significant anticorrelation between gene mobility and the frequency of selection ( P value ≪ 2e-3, R 2 = 0.4611, see supplementary material , Supplementary Material online, for implementation). As can be seen from figure 6 , the process described above results in core-like resistance repertoires with a patchy distribution of toxin genes. A similar pattern is observed in core- and accessory genomes in vivo as accessory genes are only occasionally useful (i.e., pathogenicity or secondary metabolism [ Hacker and Carniel 2001 ; Norman et al. 2009 ]). This pattern could be explained by the dominant role of loss of genes ( Wolf and Koonin 2013 ), but this does not explain the discrepancy between the phylogeny of genes and cells. We show that HGT plays an important role in shaping this observed genome mosaicism ( Zhaxybayeva et al. 2004 ). Who Wants to Take Up eDNA? In our model, we assumed that the uptake of eDNA by cells is fixed. Although not all routes of HGT can be prevented by the cell, for example, tranduction, many bacterial species are known to have evolved regulation of DNA-uptake ( Seitz and Blokesch 2013 ). This suggests that uptake of eDNA is only occasionally useful. Indeed, when the uptake of eDNA is made evolvable in our model, cells avoid uptake under many conditions except high gene-loss (data not shown). What conditions make the uptake of eDNA favorable? Although DNA can serve as a nutrient ( Finkel and Kolter 2001 ), which might lead to transformation as a side-effect ( Macfadyen et al. 2001 ), more direct benefits of natural competence are poorly understood. Contrasting the regulated uptake of eDNA that is observed for most bacteria, the human pathogens of Neiseria sp. and Helicobacter pylori both constitutively take up eDNA ( Seitz and Blokesch 2013 ). The strong environmental pressure by the immune system might directly select for diversity, potentially explaining the constitutive uptake of eDNA. Indeed, earlier studies have shown that the uptake of eDNA can prevent the loss of genomic information ( Vogan and Higgs 2011 ; Johnston et al. 2013 ; Takeuchi et al. 2014 ). Indeed, by generating and maintaining diversity HGT could contribute to the evolvability of prokaryotes ( Hindré et al. 2012 ), shaping the standing variation on which natural selection acts. In other words, an important entry point for future research would be the evolution of natural competence in relation to the environmental variability, and the signals used by the cells to induce this process. Genes in our model are not only taken up with a fixed rate but also strictly independent of each other. In other words, two genes cannot get transferred together in a single event, as we assume no genetic linkage between genes. Although such decoupled behavior is observed in the Vibrio’s discussed by Cordero et al. (2012) , other research commonly observes both genes together on a plasmid or MGE ( Alonso et al. 2000 ; Van Melderen and Saavedra De Bast 2009 ). Linking genes can circumvent some of the risks of HGT, and could even feedback on the earlier questions regarding the evolution of natural transformation. What conditions favor independent toxin and resistance genes, and when does the alternative solution of, for example, toxin–antitoxin systems evolve? To investigate this, the model can be extended to include the evolution of genome structure, where HGT of fragmented DNA or plasmids could include multiple genes in a single event. Answering the aforementioned questions will further increase our understanding on multilevel evolution and the selection pressures that acts on these different biological levels. We conclude that HGT and the resulting interplay between gene-level and cell-level selection can structure genomes and populations of microbial communities as observed in vivo."
} | 4,039 |
35423676 | PMC8693398 | pmc | 1,201 | {
"abstract": "Piezoelectric polymers have aroused tremendous attention in self-powered flexible wearable electronics. However, conventional plate-like piezoelectric devices demonstrated insufficient output power and monotonous piezoelectric energy harvesting performance only in one direction, failing to accommodate the complex multi-dimensional stress field. In this study, a poly(vinylidene fluoride) tube featuring all-directional piezoelectric energy harvesting performance with excellent rebound resilience was prepared via a fast industrial extrusion. Benefitting from the isotropic hollow tubular structure, the piezoelectric tube experienced large deformation at a small external load and thus exhibited a strong load-amplifying function to generate the optimized output power at multi-stresses from any direction, with sensitive piezoelectric performance, quick response and mechanical robustness. Finally, the practical potential as a robust energy harvester was evaluated by harvesting small energy from irregular human motions. This study provides a facile structure designing strategy for the preparation of functional piezoelectric devices for multi-directional mechanical energy harvesting applications.",
"conclusion": "4. Conclusions In this study, the isotropic tubular structure of PVDF–HFP was designed for energy harvesting at a multi-dimensional stress field via melt extrusion. The FEM and experimental results showed that the tubular structure exhibited the load-amplifying function, leading to a smaller piezoelectric loading threshold of 0.3 N, which was far lower than 1.25 N for the plate-like sample. This stress amplification effect and isotropic nature of the tubular structure ensured the equal accommodation for multi-stresses from any direction with a quick response and mechanical robustness. Finally, the as-prepared piezoelectric tube was successfully applied to harvest small energy from irregular human motions, demonstrating that the tubular structure is a promising candidate for preparing a high-performance piezoelectric device.",
"introduction": "1. Introduction Wearable electronics, including electronic watches, smart sneakers, and electronic glasses, offer great convenience and good experience to our lives for their predominance in portability and in-time feedback of human health information such as steps, state of motions, heart rates and some other functions of human bodies. 1–4 However, the power source is still from traditional batteries, the utilization of which is inevitably restricted by limited capacity, inconvenient recharging and intractable problems of recycling. 5,6 In order to solve the issues, piezoelectric energy harvesting devices are proposed to integrate with wearable electronics to achieve the self-powering by converting wasted mechanical energy from human motions into electrical power. 7–9 The application of self-powered devices in wearable electronics was put forward by Wang in 2006, 10 where the conception of nanogenerators was first realized via the nanoarray-structural design of zinc oxide. The current piezoelectric energy harvesting devices are made from inorganic piezoelectric materials such as PZT, 11 BaTiO 3 , 12 and ZnSnO 3 . 13 The brittle nature makes them hardly biocompatible to irregular motions of human bodies, including bending and twisting, leading to structural failure and deteriorated piezoelectric performance. 14 On the other hand, piezoelectric polymers are characterized by flexibility, mechanical robustness and ease of processing, and thus are considered as promising alternatives in flexible wearable self-powered devices, 8,15,16 but a large load is often required to achieve a high power output due to the low piezoelectric constants relative to inorganic ceramics. 17,18 Therefore, the reconciliation balance between highly-efficient electromechanical conversion and mechanical durability still remains to be unsolved. The piezoelectric performance of a given material is not only determined by the intrinsic piezoelectric attributes of polymer materials but also depends on the level of the imposed stress, which is, in turn, governed by the geometry of the corresponding device. 10,19,20 Compared to the traditional piezoelectric film with flat-surface, some unique geometries, such as porous, 21 array 22 and pyramid-shaped structures, 23 exhibited strong strain-amplifying function to maximize mechanical deformation, generating optimized output power. In fact, the first piezoelectric device composed of ZnO nanoarrays was a typical case of piezoelectric output enhancement via the structured design of the lever principle. However, the construction of special structures often involved complex processing techniques such as template growing, 24 chemical vapor deposition, 25 or photolithography, 26 with low efficient production. Recently, Jung et al. demonstrated that curved geometry can intensify the applied stress on a piezoelectric device to enable a higher voltage output than the flat-type one. 27 When external force was acted on the curved surface, large deformation could be generated at the direction of imposed force due to the concentrated stress. Yang et al. successfully utilized the curved structure to obtain increased piezoelectric output from ∼1 V to ∼2 V. 28 Though effective in stress amplification and piezoelectric enhancement, the preparation of a curved structure is a multi-step process with low efficiency: first, the piezoelectric film was prepared and then prestrained to maintain the curved structure. 20,29,30 More importantly, this curved structure can gradually return to the plain structure at increased loading cycles and only works at vertical loading perpendicular to the curved surface, unable to accommodate the complex stress field from all directions in practice applications. It is well-known that a tubular structure is naturally curved in any direction and can be formed directly via industrial continuous extrusion featuring highly-efficient production. Accordingly, the piezoelectric tube is expected to exhibit promising potential in piezoelectric conversion. Herein, we chose poly(vinylidene fluoride- co -hexafluoropropene) (P(VDF–HFP)) as a model polymer because the presence of hexafluoropropylene exerted a large steric hindrance to the PVDF chain to facilitate the formation of all-trans conformation and obtain electroactive β-crystals in normal melt-processing. Then, the conventional extrusion of the polymer tube was adopted to prepare a tubular piezoelectric device. The piezoelectric performance of tubular devices and working mechanisms were systematically investigated based on experiment and finite element simulation analysis. Compared to traditional plate-like structures, the tubular device possessed a smaller load threshold of piezoelectric output and was successfully utilized for harvesting energy from different human motions. This study provides a structure designing strategy for the preparation of highly-efficient piezoelectric devices.",
"discussion": "3. Results and discussions 3.1. Stress and piezoelectric behaviors at different compressing loads It is well known that the piezoelectric output highly relies on the applied stress as determined by structural features. The piezoelectric loading threshold triggering piezoelectric output generation reflects the sensitivity of a piezoelectric device to an external load and is a key factor for determining the successful application for harvesting low-frequency mechanical energy from the daily human motion. In order to verify the amplification effect of the tubular structure on the piezoelectric output, a compression load was imposed on the plate- (FS) and tube-like (TS) samples, and the corresponding compression displacement with stress and piezoelectric signals were recorded synchronously, as shown in Fig. 1 . Although the two samples had similar instantaneous remnant polarization (the corresponding information is available in the ESI, Fig. S1 † ), the different piezoelectric properties were investigated. For the plate-like sample, no distinct piezoelectric signals appeared at an external load less than 1 N, except for some noise from static electricity in the air; 32 the load threshold was 1.25 N, where the mutational piezoelectric signal with an open-circuit voltage of ∼0.3 V, was generated. On the contrary, the tubular sample exhibited a lower load threshold of 0.3 N, with a high piezoelectric output of ∼0.4 V. One can notice that the two samples had similar compression displacement at the critical loads: 0.45% for the plate-like sample and 0.52% for the tubular sample. This sufficiently demonstrated the stress amplification effect of the tubular structure, allowing for large deformation generated at a smaller load. This amplification mechanism should be analogous to the reported curved structure. 33 Fig. 1 Study on the critical loading threshold of piezoelectric signals: schematic of the piezoelectric output test (a); vertical compression model of FS (b 0 ) and TS (c 0 ); the stress–strain curves of FS (b 1 ) and TS (c 1 ) under different lateral compression loads; the corresponding piezoelectric performance of FS (b 2 ) and TS (c 2 ); load–displacement curves (d 0 ) of PVDF tube, wherein dark dots from the experimental data were consistent with the simulated value (red line); two-dimensional stress distribution pattern (d 1 ), the stress as a function of angle in a quarter area (d 2 ) and statistical results of stress distribution (d 3 ) at the corresponding load threshold of 0.3 N. Further, the finite element analysis (FEA) was carried out to reveal the structural dependence of mechanical deformation. Under a compression load, the plate-like sample experienced homogeneous deformation with part stress concentration at the edge and the stress concentrated 8–9 kPa at the load threshold of 1.25 N, which was consistent with the average value of 9 kPa calculated based on the simple stress equation, σ = F / S (the corresponding information is available in the ESI, Fig. S2 † ). On the contrary, the tubular sample exhibited different stress characteristics. It can be seen from Fig. 1d 0 that with the increase in the external load, the PVDF tube experienced elastic and plastic deformation. As the plastic deformation affected the dimensional stability and the final application, the experimental results deviated from the simulation. However, the load–displacement relationship during the elastic deformation stage of 12% displacement fitted well with the simulated result, verifying the validity of FEA. When a compression load was imposed on the tubular sample, asymmetric stress distribution was generated along the hoop direction: maximum stress appeared at the top and bottom sides, followed by a lower value at the left and right sides with minimum deformation at ±45° and ±225° off the meridian, as reflected by the stress distribution in a quarter area as presented in Fig. 1d 1 and 2 . The statistical results of stress distribution at the load threshold of 0.3 N indicated that ∼40% area of the tubular sample was subjected to a higher stress of ≥9 kPa, which was critical stress initiating piezoelectric output, as shown in Fig. 1d 3 . This amplification of local deformation should be responsible for the enhanced piezoelectric sensitivity at a small external load, beneficial for harvesting mechanical energy from human motions. Fig. 2 Studies on piezoelectric features of TS: visualization of stress at the vertical, oblique and lateral direction with a piezoelectric output at the corresponding compression direction (a); the piezoelectric performance after 500 cycles with the inset of the piezoelectric output after one cycle (b). 3.2. Piezoelectric features of tubular devices Compared to a single working mode of the conventional curved piezoelectric device limited under vertical load, a PVDF tube can equally accommodate multi-stresses from any direction in practice application due to the isotropic nature of the tubular structure. As presented in Fig. 2a , the stress value and distributions were the same under different loads along with the vertical, oblique and lateral directions. The adaptability of multi-directional stress ensured the PVDF tube to effectively harvest energy in a complex stress field, as referred to by the same piezoelectric signals in all directions. Moreover, one can notice that the tubular harvesting device presented a rapid response towards external force with stable output due to well rebound resilience. Inset in Fig. 2b showed the piezoelectric output after one cycle. The response time to the piezoelectric signal was calculated to be ∼81 ms based on the time span from 10% to 90% of the piezoelectric peak, which was comparable to the value obtained from reported sensitive piezoelectric structures. 28,34 Due to the rapid response towards external force at small loading, the PVDF tube was mechanically robust and could sustain stable output performance for over 500 cycles of repeated mechanical deformation, suggesting the potential for a long-term application. Accordingly, it can be concluded safely that benefitting from the unique hollow tubular structure the PVDF tube is characterized by a sensitive piezoelectric performance at low load, quick response at any direction, mechanical robustness, providing the possibility of robust energy harvesting application from complex stress occasion. 3.3 Energy harvesting applications Finally, the practical potential of the PVDF tube as a robust energy harvester was evaluated by harvesting small energy from irregular motions of different body regions. Accordingly, the PVDF tube was attached to different parts of human bodies, as shown in Fig. 3a . It can be noticed from Fig. 3b–d that the distinct piezoelectric outputs were generated by tensile and compression loads from wrist bending, finger bending and walking, successfully verifying that the as-prepared PVDF tube indeed adapted to various loads and generated effective piezoelectric outputs. In addition, a capacitor of 1 μF was used as an energy storage system and the charging energy was studied by deforming the piezoelectric tube ( Fig. 3e ). By periodically bending of finger or wrist and walking, the capacitor was charged to ∼1.2 V by wrist bending, 0.7 V by finger bending or walking for 100 s. These charging characteristics demonstrated that tubular devices possessed a promising prospect for energy harvesting in wearable devices and energy harvesting at complex external load. Fig. 3 Piezoelectric energy harvesting application of TS on the human body: potential mechanical source generated by movement (a); piezoelectric performance at finger bending (b), wrist joint bending (c), walking (d) and their charging curves (e)."
} | 3,698 |
28626671 | PMC5466132 | pmc | 1,202 | {
"abstract": "Methanotrophs are a biological resource as they degrade the greenhouse gas methane and various organic contaminants. Several non-methanotrophic bacteria have shown potential to stimulate growth of methanotrophs when co-cultured, and however, the ecology is largely unknown. Effects of Sphingopyxis sp. NM1 on methanotrophic activity and growth of Methylocystis sp. M6 were investigated in this study. M6 and NM1 were mixed at mixing ratios of 9:1, 1:1, and 1:9 (v/v), using cell suspensions of 7.5 × 10 11 cells L −1 . Methane oxidation of M6 was monitored, and M6 population was estimated using fluorescence in situ hybridization (FISH). Real-time PCR was applied to quantify rRNA and expression of transcripts for three enzymes involved in the methane oxidation pathway. NM1 had a positive effect on M6 growth at a 1:9 ratio ( p < 0.05), while no significant effects were observed at 9:1 and 1:1 ratios. NM1 enhanced the methane oxidation 1.34-fold at the 1:9 ratio. NM1 increased the population density and relative rRNA level of M6 by 2.4-fold and 5.4-fold at the 1:9 ratio, indicating that NM1 stimulated the population growth of M6. NM1 increased the relative transcriptional expression of all mRNA targets only at the 1:9 ratio. These results demonstrated that NM1 enhanced the methanotrophic activity and growth of M6, which was dependent on the proportion of NM1 present in the culture. This stimulation can be used as management and enhancement strategies for methanotrophic biotechnological processes.",
"conclusion": "4 Conclusions This is the first study to report that the non-methanotrophic bacterium Sphingopyxis enhances the activity of the type II methanotroph Methylocystis . We demonstrated that NM1 enhanced the population growth of M6 as well as the expression of the genes involved in the methane oxidation pathway in a density-dependent manner. These results can be used to develop and guide management and enhancement strategies for methanotrophic biotechnological processes. For instance, this stimulation can be used for accelerating start-up in methanotrophic systems.",
"introduction": "1 Introduction Methanotrophic bacteria utilize CH 4 as their sole carbon and energy source, and thus are important in the global carbon cycle [25] . They are highly diverse and found in a wide range of environments [9] , [25] . Most of the known methanotrophic bacteria belong to the Alphaproteobacteria and Gammaproteobacteria , and some Verrucomicrobia isolates are known to be methanotrophs [25] . They transform CH 4 to CO 2 , with methanol, formaldehyde and formate as intermediates [9] . In the field of biotechnology, methanotrophs are a valuable biological resource because they can degrade the greenhouse gas methane, and co-metabolize various organic compounds [25] , [27] . Therefore, methanotrophs are used in environmental engineering systems to mitigate methane emission and to remove recalcitrant contaminants ( e.g. , trichloroethylene) [7] , [20] , [23] . Various abiotic and biotic factors can affect the growth and activity of methanotrophs [1] , [26] , [30] . Previous studies largely focused on abiotic factors such as oxygen, nutrients, moisture, and temperature, etc. to enhance methanotrophic activity [9] , [25] . However, recent studies have indicated that methanotrophs interact significantly with other bacteria in different ways. Stable isotope probing (SIP) revealed metabolic interaction between methanotrophs and non-methanotrophic bacteria in a natural environment [12] . Iguchi et al. [13] recently found that isolates of Rhizobium, Sinorhizobium, Mesorhizobium, Xanthobacter , and Flavobacterium enhanced the methanotrophic activity of Methylovulum miyakonense (belonging to Gammaproteobacteria ), and that the Rhizobium isolate stimulated the methanotrophic activities of other Gammaproteobacteria methanotrophs belonging to Methylococcaceae , Methylomonas , and Methylobacter by producing an extracellular compound. Similarly, Stock etal. [26] reported that several heterotrophic bacterial isolates increased the biomass of co-cultures with methanotrophs. In addition, Ho et al. [10] reported that richness of heterotrophic bacteria was an important factor in stimulating methanotrophic activity. Microorganisms other than those isolates may also be able to enhance growth and/or activity of methanotrophs. These non-methanotrophic organisms could potentially be used as biological stimulators in methanotrophic engineering systems. To enhance methanotrophic systems using a biological stimulator, the interaction of the stimulator with methanotrophs should be elucidated. For instance, it should be determined if this type of biological stimulation is a density-dependent process. We obtained a stable methanotrophic consortium from soil, which had been maintained with methane as sole carbon and energy for more than a year. We found that Methylocystis (belonging to Alphaproteobacteria ) comprised 73% of the community, followed by Sphingopyxis , a common soil heterotrophic bacterium [25%] when examining the community using ribosomal tag pyrosequencing (unpublished data). Therefore, we hypothesized that Sphingopyxis interacts positively with Methylocystis . The main objectives of this study were to determine if Sphingopyxis enhances the methane oxidation of Methylocystis , if Sphingopyxis stimulates the population growth and/or activity (methane oxidation enzymes) of Methylocystis , and if this biological stimulation is a density-dependent process. To address these questions, Methylocystis and Sphingopyxis were mixed at different mixing ratios. Methane oxidation rate was calculated at each ratio. Population density and rRNA expression were quantified using FISH and real-time PCR. mRNA expression levels of genes involved in the methane oxidation pathway were also quantified.",
"discussion": "3 Results and discussion 3.1 Morphological characteristics of the two organisms TEM micrographs of M6 and NM1 are shown in Fig. 1 . M6 is 1.89 ± 0.27 μm in length and 1.12 ± 0.20 μm in diameter, and NM1 is 1.01 ± 0.23 μm in length and 0.57 ± 0.06 μm in diameter. The cell masses of M6 and NM1 were estimated to be 612.1 × 10 −15 and 114.7 × 10 −15 g, respectively. Cell mass of M6 is 5.3-fold greater than that of NM1. M6 is cocci-rod in shape and has well developed intracytoplasmic membranes (ICM). ICM has been observed in other methanotrophs; it is hypothesized that these membranes are related to the enzymatic process of methane oxidation [6] , [18] , [24] . NM1 has a rod-like shape and a multilayer cell wall with no flagella. Lee et al. [17] reported that Sphingopyxis sp. Gsoil 250 T is motile and rod-shaped (0.2–0.3 mm in diameter and 1.0–1.2 mm in length) with a single flagellum. Fig. 1 Micrographs of Methylocystis sp. M6 (a) and Sphingopyxis sp. NM1 (b) obtained by transmission electron microscopy. Scale bars represent 0.5 μm. 3.2 Effects of NM1 on methane oxidation activity of M6 NM1showed no negative effect on methane oxidation ( Fig. 2 ). Methane oxidation rate (MOR) of M6 increased with the number of methane spikes in all cultures, regardless of whether NM1 was added or not ( p < 0.05). MOR increased 2-fold with the second spike and 3-fold with the third spike. This increase was likely due to the population growth of M6 over time, because methane oxidation is dependent on the biomass of methanotrophs [14] . Addition of NM1 significantly increased the MOR at the 1:9 ratio of M6:NM1 ( p < 0.05), but not at the other two ratios ( p > 0.05). Thus, NM1 could enhance the methane oxidation when it was more populated than M6. Fig. 2 Methane oxidation rates of co-cultures. Methylocystis sp. M6 and Sphingopyxis sp. NM1 were mixed at 9:1 (a), 1:1 (b), and 1:9 (c) ( n = 5). The symbol * indicates a significant difference at p < 0.05. 3.3 Effects of NM1 on population growth of M6 FISH results indicated that the presence of NM1 appeared to stimulate the population growth of M6 ( Fig. 3 ). The effect of NM1 was statistically significant at the 1:9 ratio ( p < 0.05) while not significant at the 9:1 and 1:1 ratios ( p > 0.05). Ribosomal RNA is essential for protein synthesis in organisms as a component of the ribosome [2] , and its synthesis rate can reflect the cell growth rate [8] , [28] . Relative rRNA levels (treatment to control) were estimated to determine if NM1 induces cell growth of M6 ( Fig. 4 ). The added NM1 increased the relative rRNA level at all ratios; however, the effect was only significant at the 1:9 ratio of M6:NM1 ( p < 0.05), consistent with the population results. The relative rRNA levels were 1.05 ± 0.26, 1.03 ± 0.10 and 5.39 ± 1.44 at the 9:1, 1:1 and 1:9 ratios of M6:NM1, respectively. Both results indicated that NM1 stimulated the population growth of M6 in a density-dependent manner. This population increase is one mechanism by which NM1 can increase MOR because methane oxidation activity is positively correlated with the cell number of methanotrophs in a system [4] , [13] , [14] . Fig. 3 Populations of Methylocystis sp. M6 per focal area. Methylocystis sp. M6 and Sphingopyxis sp. NM1 were mixed at 9:1, 1:1, and 1:9 ratios ( n = 5). For population measurement, 20 focal spots were randomly selected and M6 cells were directly counted. The symbol * indicates a significant difference at p < 0.05. Fig. 4 Relative transcriptional expression levels of ribosomal RNA and the particulate methane monooxygenase, methanol dehydrogenase, and formaldehyde dehydrogenase genes (treatment to control). Methylocystis sp. M6 to Sphingopyxis sp. NM1 were mixed at 9:1 (a), 1:1 (b), and 1:9 (c) ratios ( n = 5). 3.4 Effects of NM1 on transcriptional expression of pMMO, MDH, and FADH A previous study showed that non-methanotrophs stimulated methanotrophic growth in the co-cultures [13] . However, it is not known whether this is due to induction of methane oxidation pathways or not. We therefore measured transcriptional expression of pMMO, MDH, and FADH, which are involved in methane oxidation. Fig. 4 shows the relative mRNA expression levels of the pMMO, MDH and FADH genes. The relative mRNA expression levels of pMMO at the 9:1, 1:1, and 1:9 ratios of M6:NM1 were 0.34 ± 0.08, 0.85 ± 0.13, and 2.67 ± 1.31, those of MDH were 0.31 ± 0.13, 0.54 ± 0.21, and 2.40 ± 0.94, and those of FADH were 0.25 ± 0.10, 0.41 ± 0.17, and 1.26 ± 0.24, respectively. The relative expression levels of all genes were less than 0.5 at the 9:1 ratio of M6:NM1 and less than 1 at the 1:1 ratio. Interestingly, relative expression ratio was at least 1.3-fold higher at the 1:9 ratio of M6:NM1. These results indicated that NM1 enhanced the transcriptional expression of the genes involved in methane oxidation when NM1 was more abundant than M6, consistent with the population and methane oxidation rate results. Relative expression of FADH was about 2-fold lower than the expression levels of the pMMO and MDH genes. We speculate that some of the formaldehyde produced was used for biosynthesis because formaldehyde has a central role as an intermediate in catabolism and anabolism [9] . Increased transcriptional expression of these genes was likely responsible for the increased oxidation rate measured at the 1:9 ratio of M6:NM1. Similarly, [11] reported that the amount of mRNA copies was correlated with the activity in the reactor. We demonstrated that NM1 concurrently enhanced the population growth of M6 and the expression of the methane-oxidation genes in a density-dependent manner. The two types of bacterial cells were mixed on the basis of cell number. Because the mass of NM1 cells is 5.7-fold less than that of M6, the mass-ratio of NM1 to M6 was estimated to be 0.02, 0.19, and 1.68 at the 9:1, 1:1, and 1:9 ratios of M6:NM1. NM1 only had significant effects on the activity and growth of M6 at the 1:9 ratio of M6:NM1, indicating that NM1 had a significant effect on M6 only when it was present at higher mass than M6. Previous studies have shown that a few vitamins and organic acids can enhance methanotrophic growth [31] . For instance, [13] reported that cobalamin (vitamin B 12 ) produced by Rhizobium stimulated the growth and activity of several methanotrophs, including Methylomonas and Methylovulum . Xing et al. [31] reported that riboflavin and organic acids (maleate, succinate, malate, and citrate) induced the population growth of Methylosinus. Thus, we hypothesize that extracellular substances from NM1 enhanced the population growth of M6 as well as the expression of the methane-oxidation enzymes in M6. Further investigations of the metabolic interactions between these two organisms are warranted. Our results also imply that methane oxidizers may commonly interact with other organisms in natural environments."
} | 3,210 |
31758008 | PMC6874547 | pmc | 1,203 | {
"abstract": "Biogenic volatile organic compounds (BVOCs) influence organism fitness by promoting stress resistance and regulating trophic interactions. Studies examining BVOC emissions have predominantly focussed on terrestrial ecosystems and atmospheric chemistry – surprisingly, highly productive marine ecosystems remain largely overlooked. Here we examined the volatilome (total BVOCs) of the microalgal endosymbionts of reef invertebrates, Symbiodiniaceae. We used GC-MS to characterise five species ( Symbiodinium linucheae , Breviolum psygmophilum , Durusdinium trenchii , Effrenium voratum , Fugacium kawagutii ) under steady-state growth. A diverse range of 32 BVOCs were detected (from 12 in D . trenchii to 27 in S . linucheae ) with halogenated hydrocarbons, alkanes and esters the most common chemical functional groups. A thermal stress experiment on thermally-sensitive Cladocopium goreaui and thermally-tolerant D . trenchii significantly affected the volatilomes of both species. More BVOCs were detected in D . trenchii following thermal stress (32 °C), while fewer BVOCs were recorded in stressed C . goreaui . The onset of stress caused dramatic increases of dimethyl-disulfide (98.52%) in C . goreaui and nonanoic acid (99.85%) in D . trenchii . This first volatilome analysis of Symbiodiniaceae reveals that both species-specificity and environmental factors govern the composition of BVOC emissions among the Symbiodiniaceae, which potentially have, as yet unexplored, physiological and ecological importance in shaping coral reef community functioning.",
"introduction": "Introduction Photosynthetic organisms, ranging from complex vascular plants to single celled microalgae, are major producers of biogenic volatile organic compounds (BVOCs), which can represent up to 10% of fixed carbon and play numerous physiological and ecological roles 1 , 2 . BVOCs are exceptionally diverse and highly reactive, with chemical lifetimes ranging from minutes (e.g. β-caryophyllene) to months (e.g. acetone) 3 , and can be synthesised as by-products of metabolic pathways 4 or be produced to maintain metabolic homeostasis 3 . The roles played by these compounds are multifaceted, from protection against abiotic stress 5 , 6 and pathogens 7 , 8 , to chemical signalling 9 – 12 . Terrestrial tropical ecosystems are well recognised “hotspots” of global BVOC emissions 13 , but growing evidence also highlights the role of tropical marine ecosystems, such as coral reefs 14 , 15 as major sources of BVOC emissions. Reef-building corals emit the highest recorded concentrations of the sulphur gas dimethyl sulfide (DMS; up to 18.7 µM reported in coral mucus) 16 , 17 , a compound involved in climate regulation 18 , coral stress response 5 , 19 and functioning as an infochemical 20 , 21 . Furthermore, the presence of acetone and dichloromethane was recently identified in reef seawater samples, indicating BVOCs produced in coral reefs are likely to be a complex mixture 22 . In addition, cultures of the dinoflagellate endosymbionts of reef-building corals (Family: Symbiodiniaceae) can produce DMS 19 , 23 , 24 and isoprene 25 . However, these observations are derived from targeted quantifications of specific BVOCs and likely represent only a small fraction of the compounds that Symbiodiniaceae can produce. Emission of BVOCs from corals are thought to largely originate from Symbiodiniaceae due to the large quantities of carbon and metabolites they translocate 26 , 27 . Metabolic coupling between coral host and algal endosymbionts is critical for viable coral reef functioning – whereby the Symbiodiniaceae drive coral productivity, but can also govern susceptibility of their host to stressors by controlling the exchange of nutrients 28 . Recent studies have moved towards identifying the functional diversity amongst the Symbiodiniaceae to determine key traits of resistance to stress 29 , 30 , which is critical to understand their susceptibility to rising seawater temperatures. Whilst different metabolic traits and hence diagnostic metabolites appear central in governing this functional diversity 31 , the role of BVOCs remains largely unexplored. The assortment of BVOCs produced by an organism has recently been termed the ‘volatilome’ 32 , 33 . Assessments of volatilomes, or volatilomics, is widely used in medical research, where the approach has been used to diagnose patient health via bio-markers in exhaled breath 34 . More recently, volatilomics has also been used to quickly distinguish plant strains and to successfully monitor long-term plant health 35 , with this approach showing promise for the development of new non-invasive monitoring tools of taxonomic and functional diversity in aquatic ecosystems 33 . Indeed, volatilomics may be particularly useful in connecting ecological processes and biogeochemical cycles in the ocean, since emissions of BVOCs from photosynthetic organisms can have a strong influence on atmospheric chemistry by increasing local cloud albedo or the residence time of greenhouse gases 36 , 37 . Within this context, characterising the volatilome of highly productive, habitat forming marine organisms should enable more accurate modelling of the impact of tropical BVOCs on climate regulation 2 , 38 . Here we provide the first characterisation of the volatilome of different Symbiodiniaceae species (spanning five genera) to determine whether and how volatile metabolite signatures are conserved across divergent taxa, and to identify volatiles that may be involved in previously overlooked physiological processes, ecosystem interactions and atmospheric cycling. In addition to this screening across-strains, we also examined the volatilomes of two Symbiodiniaceae strains with different thermal tolerance thresholds under conditions of heat stress, to investigate the extent of thermally-induced changes in the volatilome and identify BVOCs potentially involved in Symbiodiniaceae thermal tolerance.",
"discussion": "Discussion The multifaceted biological and ecological functions of BVOCs can influence ecosystem resilience 2 , and therefore understanding the roles of these compounds in shaping ‘healthy’ functioning of threatened ecosystems such as coral reefs is particularly important 33 . To contribute to an enhanced understanding of coral reef volatilomics, we performed the first characterisation of the volatilome of the coral endosymbionts Symbiodiniaceae, which revealed a substantial diversity of BVOC production. Our study screened species spanning a wide range of Symbiodiniaceae genera 39 , demonstrating that key species produce far more BVOCs than the iconic dimethyl sulfide. This diverse pool of volatiles included halogenated hydrocarbons, alkanes and esters among the most commonly detected compounds. By screening five different species (spanning five genera), S . linuchae , B . psygmophilum , D . trenchii , E . voratum and F . kawagutii , we detected 32 different BVOCs. This diversity of BVOCs is consistent with a recent screening of three species of marine macroalgae ( Ulva prolifera , Ulva linza and Monostroma nitidum ), which identified 41 volatile compounds 40 . Similarly to our study, alkanes, alkenes, ketones, aldehydes, sulphur compounds, alcohols and esters were detected in the macroalgae studied 40 . However, unlike macroalgae, Symbiodiniaceae also produced halogenated hydrocarbons, and only benzaldehyde (known to be involved in signalling amongst insects 41 ) was identified in both Symbiodiniaceae and macroalgae. Of the 32 volatile compounds detected in this study, six were present in all five Symbiodiniaceae species and were therefore defined as core volatiles. Recognising the ubiquitous nature of certain volatiles is a first step in understanding the potential ecological relevance of compounds produced by this important family of microalgae. We defined DMS, 2,3-dimethyl hexane, 6-methyl octadecane, an UC halogenated hydrocarbon, UC organosulfur and UC40.63 as core volatiles. Two of these compounds, 2,3-dimethyl hexane and 6-methyl octadecane, are both short chain alkanes, which may have originated from the oxidation of fatty acids. The inclusion of DMS in the core set of volatiles produced by all Symbiodiniaceae genera examined here adds weight to the importance and ubiquitous nature of this compound in these organisms. However, it is notable that DMS relative abundance varied substantially between Symbiodiniaceae species, with significantly more DMS detected in S . linucheae and D . trenchii than all other species, while significantly less DMS was detected in F . kawagutii compared to all other species. Both S . linucheae and D . trenchii are known to be more heat resistant species 29 , 30 , 42 , potentially suggesting that the high amounts of DMS detected under steady state conditions positively influence their resilience, allowing these species to have a larger existing pool of antioxidants 19 , 43 . In addition to DMS, differences between Symbiodiniaceae volatilomes were largely driven by the relative abundance of UC40.45, methyl jasmonate, styrene and 4-fluoro-3-trifluoromethylbenzoic acid eicosyl ester. Methyl jasmonate is ubiquitous in higher plants and is an important signalling molecule that regulates plant development and also plays a role in defence against biotic (e.g. herbivory) and abiotic (e.g. heat) stresses 44 . Methyl jasmonate was detected in all replicates of S . linucheae , and given previous observations in higher plants, this BVOC might be involved in the thermal resistance of this Symbiodiniaceae species. The function of the other two molecules, styrene and 4-fluoro-3-trifluoromethylbenzoic acid eicosyl ester, remain uncertain but styrene has previously been reported in higher plants 45 . This study identified multiple halogenated compounds however, only three were fully classified (1,2-dichloropropane, 3-trifluoroacetoxypentadecane & 4-fluoro-3-trifluoromethylbenzoic acid eicosyl ester). Halogenated compounds are of particular importance as they can degrade atmospheric ozone 46 , they are known to be produced naturally 47 and have previously been reported from marine microalgae 48 . For example, three tropical microalgae ( Amphora sp., Synechococcus sp. & Parachlorella sp.) can produce a range of iodinated and brominated compounds (methyl iodide, bromoform, dibromomethane, dibromochloromethane, and chloroform), with production shown to be species-specific and growth-phase dependent 48 , highlighting the importance of the physiological state of the cell for volatile emissions. Other examinations of a suite of common phytoplankton ( Calcidiscus leptoporus , Emiliania huxleyi , Phaeodactylum tricornutum , Chaetoceros neogracilis and Dunaliella tertiolecta ) demonstrated that all tested species emitted chloromethane, bromoform, bromomethane, chlorobenzene and dichlorobenzene 49 . Our understanding of the function of halogenated compounds remains in its infancy, with current hypotheses suggesting the tight coupling of their production with oxidative processes (potentially formed as side products during the breakdown of reactive oxygen species) 50 , 51 . Halogenated compounds are also thought to sometimes function as ‘infochemicals’, with studies on macroalgae demonstrating that bromoform supports the alga’s defence by functioning as an antimicrobial 52 – 54 . We observed that two halogenated compounds, including 3-trifluoroacetoxypentadecane and another unclassified halogenated BVOC, differed significantly between Symbiodiniaceae species and notably seem to co-occur (Correlation: 0.73, P Value = 0.002, Pearson R, Metaboanalyst4.0 55 ), with significantly higher levels of both compounds present in S . linucheae compared to all other species tested. 3-trifluoroacetoxypentadecane has previously been shown to have antimicrobial properties 56 , however, far more work is needed to accurately define the function of these halogenated compounds in Symbiodiniaceae. This initial screening of a range of Symbiodiniaceae species has demonstrated the diversity and species-specific nature of the volatilome. Despite this diversity, a consistent core emerged, suggesting that some compounds have a conserved role across species. While appreciating these potential roles is important, most natural systems rarely remain in steady-state conditions for prolonged amounts of time. Currently, marine systems are experiencing an increase in the frequency and severity of harmful thermal stress events. How corals respond to heat-wave induced bleaching and mortality is often influenced by the thermal tolerance of their endosymbiotic algae (Symbiodiniaceae) 28 . A number of physiological traits appear to differentiate stress tolerant-versus-susceptible species (or genetic variants) of Symbiodiniaceae, including maintaining integrity of photosynthetic constituents 30 , 57 and upregulation of pathways to ‘detoxify’ organelles 58 . However, the effect of thermal stress on the Symbiodiniaceae volatilome had not been explored until now. With the onset of thermal stress, a larger number of BVOCs were detected in the thermally tolerant D . trenchii , while the thermally sensitive C . goreaui produced fewer compounds relative to control conditions. The ability to synthesise specific compounds under stress could be involved in the thermal tolerance of D . trenchii , as ‘ de novo’ BVOC synthesis has been previously observed in higher plants during abiotic stresses 4 , 6 , 59 . During thermal stress, concomitantly with a decrease in cell health, a dramatic increase in nonanoic acid was recorded in D . trenchii . Nonanoic acid is a fatty acid, a group of compounds that can potentially result from increased cell membrane lysis 60 , 61 . Furthermore, significantly less DMS was detected in D . trenchii during stress, which was the only compound to significantly decrease in this species. Previous work targeting DMS production in other Cladocopium and Durusdinium Symbiodiniaceae strains also observed a decrease in DMS with the onset of heat stress 19 . Lower levels of DMS under thermal stress may either indicate that Symbiodiniaceae decrease their production, or that DMS degradation increases due to reactions with harmful molecules in response to thermal stress. Interestingly, in our study, we detected higher amounts of another sulphur compound, dimethyl disulfide (DMDS) in C . goreaui during stress. DMDS can be formed from the photo-oxidation of methanethiol 62 . These results indicate that we need to consider the full suite of volatile sulphur compounds to fully elucidate the role that these chemicals play in stress response and trophic interactions in coral reefs. An additional 17 compounds significantly increased during thermal stress in C . goreaui , with 1,3-dimethoxy-2-propanol, UC40.78 and DMDS the main drivers of the differentiation between treatments. Increases in 1,3-dimethoxy-2-propanol may have resulted from lipid peroxidation 63 , 64 . Whether these BVOCs are released prior to other visual (e.g. bleaching) or physiological 30 stress indicators remains unknown and will be important to assess for the potential use of specific BVOCs to diagnose early stress responses. The sheer quantity of detected and unidentified compounds that differed significantly across species and under heat stress highlights a critical need to robustly quantify and classify these BVOCs. However, lack of a comprehensive marine BVOC database severely limits our interpretation of volatilomic data, an issue that similarly limits progress for other metabolomic approaches 33 . Terrestrial BVOC studies have already identified ~30,000 volatile compounds to date 65 . The unidentifiable chemical diversity highlighted here teases at the potential unexplored roles of BVOCs in biological and ecological interactions in marine systems. These unidentified compounds should therefore not be discarded from future analyses as they may play key roles in stress response or could function as useful stress biomarkers that can be measured non-invasively. Volatile databases are continually improving and as such we have made available our mass spectra files (MSV000084436; MassIVE; 10.25345/C5RD5C) for future studies as more comprehensive databases become available. Numerous BVOCs are known to induce the formation of secondary organic aerosols that enhance cloud formation and albedo (eg. DMS, benzaldehyde, toluene & styrene 66 , 67 ), while other compounds can result in increased formation and residence time of crucial greenhouse gases such as ozone 68 . Marine BVOC emissions are often overlooked when it comes to global modelling of BVOC emissions, largely due to the lack of data from these ecosystems. However, marine ecosystems have a large influence on atmospheric chemistry. Tropical areas are known to have stronger convection forces that lead to greater transport of emitted BVOCs to the troposphere and stratosphere 69 . It is therefore essential to understand current tropical baseline emissions if we wish to accurately model future climate scenarios. Examining the Symbiodiniaceae volatilome is a key step towards understanding the prevalence and function of tropical marine BVOCs. However, further work is needed to clarify how free living Symbiodiniaceae BVOC production varies from endosymbiotic Symbiodiniaceae. Here we demonstrate for the first time that volatile metabolites produced by Symbiodiniaceae are not only composed of a broad spectrum of BVOCs, but that their production can be influenced by stressful – suboptimal – conditions, suggesting that these overlooked BVOCs likely operate as key constituents regulating metabolic competency. We detected six BVOCs with putative signalling and antioxidant functions that were ubiquitous across the five Symbiodiniaceae genera investigated. Many of the BVOCs reported here are as yet uncharacterised, highlighting an urgent need to develop marine-specific annotation pipelines and to further identify new and abundant compounds. This is of particular relevance given that some of these compounds might play currently uncharacterised roles in resistance and survival of corals during thermal stress events. This work provides direction for future studies to start unravelling the complex functions of volatile metabolites in coral reefs. Corals are some of the most complex symbiotic metaorganisms and the many microbial partners they harbour are likely to contribute to their BVOC emissions."
} | 4,635 |
35425300 | PMC8979045 | pmc | 1,204 | {
"abstract": "Triboelectric nanogenerators (TENGs) based on ferroelectric organic materials have advantages of high flexibility, biocompatibility, controllable ferroelectric properties, etc. However, this has limited the electrical output performance due to their lower ferroelectric characteristics than those of inorganic ferroelectric materials. A lot of effort has been made to improve the organic ferroelectric characteristics through composites, surface modifications, structures, etc. Herein, we report TENGs made of ferroelectric composite materials consisting of poly(vinylidene fluoride- co -trifluoroethylene) (PVDF-TrFE) and poly(3,4-ethylenedioxythiophene):poly(styrenesulfonate) (PEDOT:PSS). The composite was prepared by simply blending PVDF-TrFE and PEDOT:PSS with a weight ratio from 0% to 60%. When the ratio was 20%, the ferroelectric-crystalline phase was enhanced and the highest dielectric constant was observed. Accordingly, the TENGs consisting of 20% composite film and polyimide exhibited the best output performance: the maximum open circuit voltage and short circuit current were ∼15 V and ∼2.3 μA at 1 Hz oscillation, respectively. These results indicate that the ferroelectric characteristics of PVDF-TrFE can be enhanced by adding PEDOT:PSS as a nanofiller.",
"conclusion": "4 Conclusions In summary, we attempted to create all-organic triboelectric composite films with PVDF-TrFE and PEDOT:PSS. The PVDF-TrFE/PEDOT:PSS composite films maintained the characteristics of organic materials while they produced a high-electrical output. PEDOT:PSS was used as a nanofiller and was stably blended with PVDF-TrFE/PEDOT:PSS composite solutions. The effects of PEDOT:PSS at various weight ratios in PVDF-TrFE were investigated using XRD, FT-IR, and DSC. The XRD and FT-IR results showed that the electroactive phase of the composite films was preserved after the addition of different PEDOT:PSS concentrations in PVDF-TrFE. Additionally, the DSC analysis results showed an improvement Χ c at different PEDOT:PSS concentrations compared with the PVDF-TrFE films. It seems that PEDOT:PSS affects the conformation of the ferroelectric characteristics. The electrical output performance reached a maximum at a PEDOT:PSS composite film content of 20 wt%, showing an open-circuit voltage of ∼15 V, and a closed-circuit current of ∼2.3 μA for oscillations at 1 Hz. The power density of the PVDF-TrFE/20 wt% PEDOT:PSS composite films was 12.8 mW m −2 at an R L of 2 MΩ, which was shifted to an R L of 6 MΩ for the PVDF-TrFE films. The dielectric constant value increased up to the PVDF-TrFE/30 wt% PEDOT:PSS composite films compared with the PVDF-TrFE films. It can be inferred that PEDOT:PSS has a leverage effect on the ferroelectric characteristics of PVDF-TrFE. Therefore, this led to an increase in the electrical output performance by improving the accumulation of triboelectric charges. These results indicate that PVDF-TrFE/20 wt% PEDOT:PSS composite films can be applied to FTENGs at a small-scale air flow. They have the potential as a power source for a variety of sensors.",
"introduction": "1 Introduction Triboelectric nanogenerators (TENGs) based on ferroelectric polymers, self-powered energy harvesters, have attracted considerable attention because they can convert low-frequency and irregular mechanical energy to electrical energy. Also, they can be flexible, light-weight, bio-compatible, low-temperature, and solution-processable. 1–3 They provide opportunities for application to self-powered electronics, such as the Internet-of-Things, sensors, and wearable and healthcare devices. 4–7 TENGs have many harvesting applications, such as mechanical, thermal, magnetic, and solar. 8,9 Specifically, a flutter-driven TENG (FTENG) can use small-scale wind energy by utilizing a fluctuating object in the air flow, and it directly converts it to electric energy based on triboelectrification and electrostatic induction. 10,11 Moreover, the electrical output performance of TENGs is affected by the ferroelectric characteristics and surface charges. Thus, TENGs of ferroelectric polymer materials afford an improvement through the use of various methods, such as composite materials, structure, relative permittivity, and surface modifications. 12,13 Among ferroelectric polymer materials, such as polyvinylidene fluoride (PVDF) and its copolymers, are extensively used for TENGs and piezoelectric nanogenerator active films because of their outstanding chemical resistivity, flexibility, biocompatibility, and mechanical and controllable ferroelectric properties. 14,15 PVDF has five phases, including α, β, γ, δ, and ε. In particular, the β and γ phases can affect the ferroelectric characteristics and promote the surface charge density. Thus, that phases need to be improved. 16,17 The electrical output performance of PVDF is limited owing to its low-ferroelectric properties, 18 and PVDF is used to enhance the ferroelectric properties by using its copolymer. Among the copolymers, poly(vinylidene fluoride- co -trifluoroethylene) (PVDF-TrFE), the ferroelectric properties and dielectric are higher than those of PVDF and the others owing to the TrFE monomers. 19,20 Thus, they are good candidates as host materials for TENGs. When it comes to a contact-separation operational mode in TENGs, it can lead to deformation and can cause the wear of the surface during energy harvesting. This can lead to a negative influence on the electrical output performance. The additive materials with nanoscale dimensions adding to the host material that called nanofillers. Through the adequate nanofillers added to the host materials, it can be enhanced physical properties of the host materials. Moreover, this approach involves modifying the bulk properties of materials based on the functionalization of the chemical structure. 21,22 Therefore, ferroelectric materials with nanofillers can increase the physical properties and the relative dielectric constant, which plays a significant role in the generation of triboelectric charges. 23,24 Inorganics, metal oxide nanoparticles, and conducting materials such as BaTiO 3 , 18 Au, 23 AlO x , 25 Ag, 26 carbon nanotubes, 27 polyaniline, 28 and so on, 29,30 have been utilized as nanofillers in the ferroelectric polymers to enhance their electrical output. Among the conducting polymers, poly(3,4-ethylenedioxythiophene)–poly(styrenesulfonate) (PEDOT:PSS) has the advantages of a flexible, highly stable, and low-intrinsic thermal conductivity, and can be used in a solution process. 31,32 Additionally, it has a higher electrical conductivity than other conducting polymers. 33 Hence, PEDOT:PSS has been extensively used for the top and bottom electrodes of devices, buffer layer of optoelectronic devices, and as thermoelectric materials because of the aforementioned properties. 34–38 There are some studies has been conducted to fabricate the PEDOT:PSS coated on PVDF-TrFE films, but the characteristics of the composite films have yet to researched. 39 We expect that TENG composite films that PEDOT:PSS serves as a nanofiller in PVDF-TrFE. In this study, we fabricated fully organic materials for TENG films and proposed PEDOT:PSS as a nanofiller in PVDF-TrFE composite films to increase the electrical output performance of the films. Using X-ray diffraction (XRD), Fourier transform infrared (FT-IR), and differential scanning calorimetry (DSC) characterization, we investigated the effect of PEDOT:PSS at different concentrations on PVDF-TrFE composite films, in terms of the ferroelectric crystallization of the phase. We then optimized the PVDF-TrFE/PEDOT:PSS composite by varying the PEDOT:PSS concentration. The electrical output enhancement of the TENG films with PVDF-TrFE/PEDOT:PSS composite films was then measured. The dielectric constants of the PVDF-TrFE films and PVDF-TrFE/PEDOT:PSS composite films were evaluated. Furthermore, we demonstrated an FTENG using PVDF-TrFE/PEDOT:PSS composite films in the presence of air flow.",
"discussion": "3 Results and discussions We prepared PVDF-TrFE/PEDOT:PSS composite films with weight ratios in the range of 0–40%. PVDF-TrFE is insoluble in water, and PEDOT:PSS is a water-based solution; thus, when PEDOT:PSS is added at a weight ratio greater than 40%, the solution is aggregated. However, the composite solution with a ratio smaller than 40% was well dispersed, and the film made from the composite solution was semi-transparent ( Fig. 1b ). The morphology of PVDF-TrFE/PEDOT:PSS composite films was analyzed by FE-SEM as shown in Fig. S2. † It consisted of the surface of rod-like shape. To investigate the effect of PEDOT:PSS on the crystalline phase of PVDF-TrFE, we measured the XRD spectra of the composite films in the 2 θ range of 15–25° ( Fig. 2a and S3a † ). All films yielded a peak at 19.8°, which is associated with the β-phase at the (110) and (200) planes. 40,41 These results indicate that the crystallinity of PVDF-TrFE did not significantly affect by adding PEDOT:PSS. Fig. 2 The PVDF-TrFE/PEDOT:PSS composite films with different PEDOT:PSS contents of 0% and 20%: (a) XRD analysis, (b) FT-IR spectra from 400 cm −1 to 1600 cm −1 , and (c) DSC scan results heat flow vs. temperature during heating. Using FT-IR spectroscopy, we investigated the films for the transformation of phases and chain orientation within a wavenumber range of 400–1600 cm −1 . The spectra reveal important information about PVDF-TrFE and PVDF-TrFE/PEDOT:PSS composite films on the basis of the presence of the crystalline phase ( Fig. 2b and S3b † ). Crystallization of the β + γ phase was observed in all the films in the absorption bands at 840 cm −1 . 42 The strong peak was classified as a characteristic of the β + γ phase, which is a characteristic of CF 2 symmetric stretching. In addition, the peaks at 1430 and 1288 cm −1 is attributed to the β phase. The presence of the γ phase was observed at 1235 cm −1 . The bands at 763 cm −1 were associated with the α-phase. 43,44 Evaluation of the electroactive phase (β + γ phase) fraction ( F EA ) in the samples was calculated using the Lambert–Beer law eqn (1) : 45 1 Herein, I α and I EA are the absorbance intensities of the FT-IR peak at 763 cm −1 (α phase) and 840 cm −1 (β + γ phase), respectively. K α and K EA are the absorption coefficients at 763 cm −1 and 840 cm −1 , respectively; K α = 6.1 × 10 4 cm 2 mol −1 , K EA = 7.7 × 10 4 cm 2 mol −1 . 44,46 The calculation indicates that the F EA values decreased up to 62.66% with higher PEDOT:PSS content ( Table 1 ). The α-phase of the PVDF-TrFE/PEDOT:PSS composite films increased at 763 cm −1 because the C–S–C peak of PEDOT:PSS appears at approximately 761 cm −1 (Fig. S3c † ). 47 The F EA values could be reduced by increasing the PEDOT:PSS concentration. To confirm the ferroelectric crystallinity of PVDF-TrFE/PEDOT:PSS composite films, DSC was conducted for thermal analysis between 25 °C and 200 °C at a heating rate of 10 °C min −1 . Fig. 2c and S3d † shows the DSC results. The two peaks in the endothermic heat flow curves during heating were identified, wherein the first peak ( T c1 ) corresponded to the ferroelectric to paraelectric transition and the second peak represented the melting transition ( T m ). 48,49 The crystallinity ( Χ c ) of PVDF-TrFE/PEDOT:PSS (0% to 40%) composite films was calculated on the basis of DSC scans, according to eqn (2) . 50 Respective values are listed in Table 1 . 2 where Δ H m is the melting enthalpy from the area of T m peak and Δ H 0 is the enthalpy of the 100% crystalline PVDF-TrFE (91.45 J g −1 ) that was obtained from the literature. 51 Increasing the PEDOT:PSS concentration in the films by 20 wt%, the Χ c was observed to increase 37.11%. At 30% and 40% PEDOT:PSS concentration, the Χ c reduced to the similar value to that of PVDF-TrFE. It seems that excessive PEDOT:PSS disturbs the form of ferroelectric crystallinity. These DSC results indicate that PEDOT:PSS is an effective filler used to improve the ferroelectric characteristics. Additionally, this demonstrated that the PVDF-TrFE doped with 20 wt% PEDOT:PSS reached the highest ferroelectric phase. In addition, to investigate the dielectric characteristics, we tested the capacitance of the PVDF-TrFE films and PVDF-TrFE/PEDOT:PSS composite films. It was measured at 300 K in the frequency range of 100 Hz to 300 MHz, and the dielectric constant was transformed, as shown in Fig. 3 . The dielectric constant ( ε r ) was defined according to eqn (3) , 52 3 where C is the capacitance, d is the thickness of the films ( Table 2 ), ε 0 is the permittivity of air (8.85 × 10 −12 F m −1 ), and A is the active area of the films (0.0009 m 2 ). The results show that the dielectric constant value increases for contents up to 20 wt% PEDOT:PSS. In fact, the PVDF-TrFE/20 wt% PEDOT:PSS composite films yielded a higher dielectric constant (1.70 at 1 kHz) compared with those of PVDF-TrFE films (0.31 at 1 kHz) ( Table 2 ). For contents higher than 20 wt% PEDOT:PSS, the dielectric constant decreased, which seems to be attributed to the decrease in the ferroelectric crystallinity and the electrical nature of PEDOT:PSS. We expect that this is because of the formation of microcapacitors in the nanocomposite because of the presence of nanofillers. 21 Consequently, the ferroelectric characteristics of the composite films improved and it has the capability to enhance the triboelectric charges accumulate. 23 Relative fraction of the electroactive phase ( F EA ) and the degree of ferroelectric crystallinity ( Χ C ) of the poly(vinylidene fluoride- co -trifluoroethyle)/poly(3,4-ethylenedioxythiophene):poly(styrenesulfonate) (PVDF-TrFE/PEDOT:PSS) composite films at various PEDOT:PSS concentrations from 0% to 40% based on Fourier transform infrared spectroscopy and differential scanning calorimetry, respectively PVDF-TrFE 10 wt% PEDOT:PSS 20 wt% PEDOT:PSS 30 wt% PEDOT:PSS 40 wt% PEDOT:PSS \n F \n EA (%) 76.46 73.32 70.72 66.86 62.66 \n Χ \n C (%) 30.10 30.82 37.11 29.08 31.30 Fig. 3 Comparison of the dielectric constant for the PVDF-TrFE/PEDOT:PSS composite films with different PEDOT:PSS concentration from 0% to 40% at 300 K. Thickness and the dielectric constant values of the PVDF-TrFE/PEDOT:PSS composite films at various PEDOT:PSS concentrations from 0% to 40% PVDF-TrFE 10 wt% PEDOT:PSS 20 wt% PEDOT:PSS 30 wt% PEDOT:PSS 40 wt% PEDOT:PSS Thickness (μm) 17.95 17.03 16.94 16.65 16.69 Dielectric constant (at 1 kHz) 0.31 0.87 1.70 1.25 0.30 The working mechanism of the fabricated TENG, which includes triboelectrification and electrostatic induction, is shown in Fig. 4 . As shown in Fig. 4a , when the composite films and the PI tapes are in contact, the opposite charges are equally generated between the composite and PI tapes of the contact surfaces. The composite films generated more electrically negative charges than the PI tapes. 53,54 Releasing the two films, the potential difference drives electrons from the ITO electrode to the Cu electrode ( Fig. 4b )The electron flows continually until the maximum value of V o is reached as the composite films fully separate from the original states ( Fig. 4c ). In sequence, the electrons are driven from the Cu electrode back to the ITO electrode, decreasing the amount of induced charges when the composite films approach the PI tapes ( Fig. 4d ). Consequently, an alternating current is generated by the contact and separation modes between the composite films and PI tapes. Fig. 4 Working mechanism of TENG with the PVDF-TrFE/PEDOT:PSS composite films: (a) pressed, (b) releasing, (c) released and (d) presssing. To assess the electrical output performance of the PVDF-TrFE and PVDF-TrFE/PEDOT:PSS composite films, a wave driver was used for the periodic contact and separation modes. The applied tapping force was ∼8 N with 1 Hz oscillation, and the maximum distance of the contact separation was set to 7 mm. Fig. 5a and b shows the open-circuit voltage and closed-circuit current of PVDF-TrFE and PVDF-TrFE/PEDOT:PSS composite piezoelectric films containing different concentrations of PEDOT:PSS. PVDF-TrFE films generated an open-circuit voltage ( V o ) equal to 4 V and a short-circuit current ( I sc ) equal to 1.16 μA. As the concentration of PEDOT:PSS increased, V o and I sc were observed to increase up to 20 wt% PEDOT:PSS to respective values equal to 15.6 V and 2.32 μA, respectively. For PEDOT:PSS contents above 20 wt%, V o and I sc decreased. It seems that the predominance of PEDOT:PSS can be attributed to the reduced ferroelectric characteristics. These trends appear similar to the results of the preceding analysis. Fig. 5 Generated electrical output performance of TENG with the PVDF-TrFE/PEDOT:PSS composite at various PEDOT:PSS concentrations from 0% to 40%: Measured (a) open-circuit voltage and (b) closed-circuit current. (c) Output voltage and current of the PVDF-TrFE/20 wt% PEDOT:PSS composite films with the external load resistance varying from 80 kΩ to 960 MΩ. (d) Calculated the power density with at difference external load resistances of the PVDF-TrFE films and the PVDF-TrFE/20 wt% PEDOT:PSS composite films. The TENG can be explained by an inherent capacitive behavior, and we expected the quantities of triboelectric charges with electrical potential according to eqn (4) : 55,56 4 where Q is the surface charge, C is the capacitance, and V Tribo is the triboelectric voltage. Q depends on V Tribo , and the dielectric constant is defined in accordance with eqn (4) . Herein, the values of V Tribo changed with the concentration of PEDOT:PSS; as a result, the surface charge also changed. In addition, we calculated the triboelectric charge density with the relationship Q , V Tribo , and dielectric constant by using eqn (5) : 57 5 where σ is the surface charge density. The calculated values of σ for all samples were 0.52, 6.05, 13.90, 6.25, and 0.67 μC m −2 for PVDF-TrFE/PEDOT:PSS composite films at the concentrations of 0% to 40%, respectively. As a result, this shows the same tendency as those of the experimental and calculated results. We predict that PEDOT:PSS enhances the surface charge density by improving the ferroelectric characteristics. As shown in Fig. 5c , to estimate the external load resistance ( R L ) on the electrical output performance of PVDF-TrFE/20 wt% PEDOT:PSS composite films, V o and I sc were measured at different R L values from 80 kΩ to 960 MΩ. With the increase of R L , the V o value of PVDF-TrFE/20 wt% PEDOT:PSS composite films increased from 0.2 to 15.6 V, and the I sc value decreased from 2.3 μA to 160 nA. In addition, the power density was 12.8 mW m −2 at an R L of 2 MΩ ( Fig. 5d ). Compared with the power density of the PVDF-TrFE films, the maximum power density was shifted by 6 MΩ. These trends were observed at the other PEDOT:PSS contents (Fig. S4, Table S1 † ). It is thus shown that PEDOT:PSS has an effect on the internal resistance in PVDF-TrFE owing to its conductivity. The free mobile electrons in the conducting polymers have an effect on the movement of electrical charges in PVDF-TrFE. 28 Thus, it can be inferred that the composite films can generate more triboelectric charges. However, for contents above 20 wt% PEDOT:PSS, the power density decreased owing to the deterioration of the properties of the ferroelectric. As the frequency of the PVDF-TrFE/20 wt% PEDOT:PSS composite films in the contact-separation mode varied from 1 Hz to 20 Hz ( Fig. 6a and b ), the corresponding V o and I sc increased from 15 V to 68 V and from ∼2 μA to ∼23 μA, respectively. The V o of the TENG should theoretically be constant at various frequencies. We used the wave driver to change the stroke distance between the PVDF-TrFE/PEDOT:PSS composite films and PI tapes with increasing frequency. 58 In addition, the capacitance changed and the charge accumulated on the composite films as the contact-separation frequencies increased; this increased the electrical output performance of the TENG. 59 The TENG using PVDF-TrFE/20 wt% PEDOT:PSS composite films was used to charge a capacitor from 1 to 1000 μF. The electrical output of the TENG was rectified by a full-bridge diode. The 1 μF capacitor which charged up to 3 V within 25 s ( Fig. 6c ). To utilize the wind energy, we demonstrated the FTENG. Fig. 7a schematically illustrates the FTENG. The FTENG is shown in Fig. 7b and Video S1. † The V o and I sc of FTENG using PVDF-TrFE/20 wt% PEDOT:PSS composite films generated 3.7 V and 450 nA, respectively, at an air flow rate of 6.5 m s −1 ( Fig. 7c ). The electrical output performance of FTENG using PVDF-TrFE films generated a V o of 1.3 V and an I sc of 210 nA (Fig. S5 † ). The FTENG was connected to a full-bridge rectifier, and nine commercial green LEDs which were connected in series instantly lit on via the FTENG ( Fig. 7d, e and Video S2 † ). Additional studies in this area are needed to affect the wind velocity, size of the FTENG, and flutter materials. Fig. 6 PVDF-TrFE/20 wt% PEDOT:PSS composite films at different oscillations from 1 Hz to 20 Hz: (a) voltage and (b) current. (c) Charging profiles of capacitors at an oscillation of 20 Hz. Fig. 7 (a) Schematic of flutter-driven TENG (FTENG). (b) FTENG acts according to flow. (c) The electrical output voltage and current of FTENG by electric fan. (d) Schematic illustration of the full-bridge circuit diagram. (e) Photograph of nine commercial green light-emitting diodes lit on by FTENG. The TENG performances of the PVDF-based composite films with nanofillers reported in the previous literatures are listed in Table S2. † The TENG performances would be a change in properties by the host materials with nanofillers or contacting the opposite materials. It can be seen that the PEDOT:PSS is effective nanofillers to enhance the triboelectric properties of PVDF-TrFE. Furthermore, it is expected that it will have the opportunity to be applied as a power supply for a variety of sensors."
} | 5,522 |
34550968 | PMC8489724 | pmc | 1,207 | {
"abstract": "The increased complexity of synthetic microbial biocircuits highlights the need for distributed cell functionality due to concomitant increases in metabolic and regulatory burdens imposed on single-strain topologies. Distributed systems, however, introduce additional challenges since consortium composition and spatiotemporal dynamics of constituent strains must be robustly controlled to achieve desired circuit behaviors. Here, we address these challenges with a modeling-based investigation of emergent spatiotemporal population dynamics using cell-length control in monolayer, two-strain bacterial consortia. We demonstrate that with dynamic control of a strain’s division length, nematic cell alignment in close-packed monolayers can be destabilized. We find that this destabilization confers an emergent, competitive advantage to smaller-length strains—but by mechanisms that differ depending on the spatial patterns of the population. We used complementary modeling approaches to elucidate underlying mechanisms: an agent-based model to simulate detailed mechanical and signaling interactions between the competing strains, and a reductive, stochastic lattice model to represent cell-cell interactions with a single rotational parameter. Our modeling suggests that spatial strain-fraction oscillations can be generated when cell-length control is coupled to quorum-sensing signaling in negative feedback topologies. Our research employs novel methods of population control and points the way to programming strain fraction dynamics in consortial synthetic biology.",
"conclusion": "Conclusions Understanding and controlling the behavior of distributed microbial systems is essential for engineering information exchange between constituent strains [ 21 ], ecological dynamics [ 22 – 24 ], metabolic resource allocation [ 25 , 26 ], and microbial social interactions [ 27 ]. The active control of cell morphology provides a way to engineer the spatiotemporal dynamics in synthetic systems and to understand the patterns that emerge in microbial communities in nature.",
"introduction": "Introduction Understanding and designing microbial consortia with distributed functionality is of increasing interest in synthetic biology [ 1 – 7 ]. Assigning different functions to genetically distinct strains in a bacterial consortium reduces the metabolic load on each strain and thus allows for more complex functionality and greater robustness [ 8 – 12 ]. Synthetic gene circuits previously engineered in single strains, such as feedback oscillators and toggle switches [ 13 , 14 ], have recently been implemented in consortia [ 10 , 12 , 15 , 16 ]. However, we still lack the mathematical and computational tools that would allow us to engineer such systems in a principled way. Synthetic microbial consortia intrinsically require balance and control of population strain fractions to achieve desired genetic circuit functionality. Important population control studies of synthetic bacterial collectives in the literature have employed both theoretical and experimental approaches, and a number of different control mechanisms have been introduced. Such approaches include predator-prey systems [ 17 ], cross-feeding auxotrophs [ 18 ], toxin-antitoxin [ 19 ] and “ortholysis” [ 1 ] mechanisms, and external switching control [ 20 ]. Although many of these methods employ regulating feedback, population dynamics achieved via toxic agents released by a cell must be tightly controlled to prevent unwanted expression and non-robust population behaviors. Studies that focus on distributed microbial systems have also been wide-ranging and include those of information exchange between constituent strains [ 21 ], ecological dynamics [ 22 – 24 ], metabolic resource allocation [ 25 , 26 ], microbial social interactions [ 27 ], and the human microbiome [ 28 , 29 ]. In each of these examples, balance and control of constituent parts is central to robust functionality [ 30 , 31 ]. In contrast to population control mechanisms employing a toxin, here we suggest that the strain fraction in a microbial consortium can be controlled by changing the average division length of cells within each strain. Our approach draws inspiration from two active research areas: The first area focuses on how cell aspect ratio (cell length divided by cell width) affects cell ordering in close-packed environments [ 32 – 37 ]. Previous studies have combined experimental, theoretical, and computational approaches to demonstrate that decreasing cell length generally decreases cell nematic ordering in spatially confined environments such as monolayer microfluidic devices (the term nematic order is a measure of the local alignment of apolar, rod-like shapes such as bacteria [ 32 , 35 , 36 , 38 , 39 ] and is borrowed from the study of liquid-crystals [ 40 , 41 ]). The second research area concerns programming bacterial cell aspect ratio by modulating expression of the cell-division proteins MreB and FtsZ [ 42 – 44 ]. Our modeling approach explores a synthesis of these two lines of research by proposing that cell division length can be modulated dynamically in a single experiment. We simulated the growth, mechanical interactions, and intercellular signaling of two microbial strains with different average cell lengths in a spatially-extended, monolayer microfluidic device. We considered populations of rod-shaped bacteria whose axial growth leads to emergent columnar population structures [ 33 , 45 ], and found that decreasing a cell’s average length can alter population dynamics by destabilizing this emergent columnar organization. Using an agent-based model (ABM), we found that this mechanism gave shorter cells a competitive advantage (“mechanical fitness”) in close-packed microfluidic trap simulations. By mechanical fitness we mean the advantage a strain has over others due to the physical properties of the cells, and physical interactions with its environment. In particular, we propose here that a shorter average cell length can provide a competitive advantage to a strain co-cultured with longer cells by allowing shorter cells to displace other cells from a domain, thus increasing their ability to compete [ 46 ] due to mechanical properties. To better understand the essential dynamical mechanisms behind the emergent dynamics, we also developed a complementary lattice model (LM) based on the key features of microbial growth and interactions [ 33 ]. We mapped the cell division length to a single parameter, the probability of cell rotation upon division, which served as proxy for cell size. Simulations of the LM confirmed that rotation probability controlled the level of disorder in the population and allowed us to confirm the mechanism by which the more disordered strain displaces its counterpart. Finally, we used the ABM with cell-length control coupled to quorum-sensing signaling to generate spatiotemporal population oscillations.",
"discussion": "General discussion Controlling the spatiotemporal population dynamics of distributed, microbial systems is essential for optimizing their functionality. Here we demonstrated how bacterial cell morphology can be used to control spatiotemporal strain dynamics in the close-packed, monolayer environment of a microfluidic trap. We used both an agent-based model (ABM) and lattice model (LM) to show that reducing average cell division-length in a two-strain consortium confers a “mechanical fitness” advantage to the shorter-length strain. We also demonstrated theoretically how multi-strain consortia could alter strain fraction temporally using a quorum sensing network topology. Such strain fraction control in the ABM is described succinctly by a relaxation oscillator (Eqs 2 – 4 and Fig 6 ). We thus approached our problem by using different levels of granularity in the modeling and analysis. Fig 7 summarizes the relationship between the ABM, LM, and the model relaxation oscillator. Such multi-leveled analysis using different formalisms can help our understanding of biological systems whose dynamics evolve on multiple temporal and spatial scales [ 51 ]. 10.1371/journal.pcbi.1009381.g007 Fig 7 A schematic showing how the three models discussed here are related. The ABM (agent-based model) comprises a mechanical interaction physics-based model [ 48 ] coupled with a diffusion solver for realistic modeling of fast diffusion of intercellular QS signaling. Local rules for bacterial agents lead to emergent spatio-temporal dynamics as outlined in Results . The LM (lattice model) reduces modeling complexity by reducing the dimensionality of the parameter space, which simplifies explanations and testing of hypotheses. A continuum ODE (ordinary differential equation) model identifies key parameters of our oscillator model and abstracts the dynamics to a coupled 2D system with separated timescales. Using our ABM, we showed strains in a consortium that achieved a competitive advantage via two distinct mechanisms that depended on the initial configuration of the strains. With a random seeding of strain types that resulted in mixed bands of each strain, we found that the emergence of horizontally-oriented cells can lead to invasion of adjacent columns of the opposite strain. In contrast, when two strains occupied different portions of the trap, the smaller strain was more disordered in a bulk population, which led to a lateral force that acted cooperatively to increase the number of disordered cells. Both strain configurations led to the ejection of the WT strain, by mechanical action of the smaller length mutant strain, to the microfluidic trap open boundaries. The LM showed that both mechanisms can be explained by changes in a single parameter, the rotation probability, p rot . This parameter determined both the probability of invasion of a neighboring, opposite-strain column when the strains were intermingled, as well as the average lateral bulk force one strain exercised on the other, when the two strains occupied different parts of the trap. Our ABM strain fraction oscillator and effective relaxation oscillation description showed that population control can be achieved using a negative-feedback, quorum-sensing signaling topology in an experimentally relevant genetic circuit [ 12 ], and we suggest that such a design could be extendable. With a small modification to our circuit, stable strain fractions could be programmed by adding exogenous inducer to, for example, repress the division length protein expression in one of the strains. Similarly, by removing the division-length circuit altogether in one of the strains, a maximum population size could result from single-ended control in some range of initial conditions. We suggest that desired strain fractions could be robustly generated under a wide range of initial population distributions. Such control of consortium composition is a fundamental problem in synthetic biology [ 10 , 20 ], and our simulations suggest a method to relax dependence on necessary exogenous control [ 52 , 53 ]. Limitations Are the mechanisms for controlling spatiotemporal dynamics of microbial communities biologically plausible? Zheng et al [ 43 ] controlled cell width and length independently by varying the expression levels of the cell division-regulatory proteins MreB and FtsZ, respectively. Moreover, there is numerical [ 42 , 44 ] and experimental [ 54 ] evidence that cell shape affects spatial structures in crowded microbial colonies. However, we are not aware of examples of synthetic bacteria actively regulating their cell morphology in response to QS signals. Since genes that regulate cell length and width have been identified, and multicellular co-repressive circuits in which multiple strains communicate via QS signals have been built [ 12 ], we suggest the future construction of communities of synthetic organisms that regulate cell shape via quorum sensing signals. For simplicity, we have not included some factors that would impact the spatial organization of and inter-cellular communication within growing bacterial colonies in natural environments. For instance, we have ignored the mechanical interactions between cells and the extracellular matrix, which can impact the structures that emerge in biofilms [ 54 , 55 ]. We have also not included the impact of fluid flow at the colony boundaries, which can affect the concentration of QS signals in a colony [ 56 ]. However, such factors can be controlled in synthetic communities of E. coli growing in microfluidic devices [ 37 , 57 , 58 ], creating conditions close to the ones in our models. We made several assumptions to present the interactions within and across strains, and we explained the impact of these interactions in the simplest possible setting. However, these assumptions are not unreasonable: Our choice of the domain and boundary conditions was motivated by rectangular microfluidic traps used routinely by synthetic biologists [ 15 ]. The prevalent nematic ordering of cells in close-packed environments that is central to the emergence of spatio-temporal patterns has been observed experimentally [ 32 , 37 , 45 , 59 ], and has been examined numerically [ 32 , 48 , 59 ] and analytically [ 32 – 34 , 47 , 48 ]. Although such nematic order is most likely to emerge in monolayer traps, it has been observed in colonies growing in three dimensions [ 54 ]. Thus, we can expect the type of close-packed interactions we described to occur—at least at the local level—between strains or even in lineages that differ in average division length. Experiments have demonstrated that the total volume of two strains dividing at different average cell length can incrase at approximately the same rate [ 43 ]. However, in the case of strong nematic ordering in steady-state, differences in growth rates will not affect our results since the faster growing strain will just exit the trap at a higher rate in an aligned column (although initial conditions with different growth rates on trap-filling could significantly alter starting population fractions). If the cells are not strongly ordered (isotropic growth), we expect a significantly faster-growing strain would displace a slower growing strain due to differences in volume expansion rate. However, the range of relative growth rates valid for the interactions we have described here is yet to be determined."
} | 3,600 |
29556533 | PMC5850073 | pmc | 1,208 | {
"abstract": "An enduring theme in microbial ecology is the interdependence of microbial community members. Interactions between community members include provision of cofactors, establishment of redox gradients, and turnover of key nutrients to drive biogeochemical cycles."
} | 65 |
23278485 | null | s2 | 1,209 | {
"abstract": "Most microorganisms live in complex communities, where they interact both synergistically and competitively. To explore the relationship between environmental heterogeneity and the spatial structure of well-defined biofilms, single- and mixed-species biofilms of Pseudomonas aeruginosa PAO1 and Flavobacterium sp. CDC-65 was grown in a planar flow cell under highly controlled flow gradients. Both organisms behaved differently in mixed cultures than in single-species cultures due to inter-species interactions, and these interactions were significantly affected by external flow conditions. Pseudomonas and Flavobacterium showed a competitive relationship under slow inflow conditions, where the supply of growth medium was limited. Under such competitive conditions, the faster- specific growth rate of Flavobacterium allowed it to secure access to favorable regions of the biofilm by overgrowing Pseudomonas. In contrast, Pseudomonas was restricted to nutritionally depleted habitat near the base of the biofilm, and its growth was significantly inhibited. Conversely, under higher inflow conditions providing greater influx of growth medium, both organisms accumulated greater biomass in mixed biofilms than in single-species biofilms. Spatial segregation of the two organisms within the biofilms contributed to enhanced overall exploitation of available nutrients and substrates, while morphological changes favored better adherence to the surface under high hydrodynamic shear. These results indicate that synergy and competition in biofilms vary with flow conditions. Limited resource replenishment favors competition under low-flow conditions, while high flow reduces competition and favors synergy by providing greater resources and simultaneously imposing increased hydrodynamic shear that makes it more difficult to accumulate biomass on the surface. Ecological interactions that produce mechanically stronger and more robust biofilms will support more extensive growth on surfaces subject to high hydrodynamic shear, but these interactions are difficult to predict from observations of the behavior of individual organisms."
} | 534 |
38907526 | PMC11369814 | pmc | 1,210 | {
"abstract": "Abstract Cyanobacteria play a key role in primary production in both oceans and fresh waters and hold great potential for sustainable production of a large number of commodities. During their life, cyanobacteria cells need to acclimate to a multitude of challenges, including shifts in intensity and quality of incident light. Despite our increasing understanding of metabolic regulation under various light regimes, detailed insight into fitness advantages and limitations under shifting light quality remains underexplored. Here, we study photo-physiological acclimation in the cyanobacterium Synechocystis sp. PCC 6803 throughout the photosynthetically active radiation (PAR) range. Using light emitting diodes (LEDs) with qualitatively different narrow spectra, we describe wavelength dependence of light capture, electron transport and energy transduction to main cellular pools. In addition, we describe processes that fine-tune light capture, such as state transitions, or the efficiency of energy transfer from phycobilisomes to photosystems (PS). We show that growth was the most limited under blue light due to inefficient light harvesting, and that many cellular processes are tightly linked to the redox state of the plastoquinone (PQ) pool, which was the most reduced under red light. The PSI-to-PSII ratio was low under blue photons, however, it was not the main growth-limiting factor, since it was even more reduced under violet and near far-red lights, where Synechocystis grew faster compared to blue light. Our results provide insight into the spectral dependence of phototrophic growth and can provide the foundation for future studies of molecular mechanisms underlying light acclimation in cyanobacteria, leading to light optimization in controlled cultivations.",
"introduction": "Introduction Photoautotrophic production both in natural aquatic systems and in controlled cultivations largely depends on the ability of cells to acclimate to the surrounding environment. The factors affecting the fitness of phototrophic microorganisms include temperature, nutrient levels, pH, as well as both intensity and quality of the incident light. Among these, the impact of light quality has been for long the least studied phenomenon. However, as evident from recent studies, the shape of the underwater spectrum is one of the crucial factors that drive worldwide phytoplankton distribution ( Grébert et al. 2018 , Holtrop et al. 2021 ). Moreover, light quality has been shown to affect phototrophic production of commodities such as isoprene ( Rodrigues et al. 2023 ). Microalgae and cyanobacteria have developed numerous mechanisms to optimize light harvesting and energy transduction in conditions of either excessive or limited light availability. The short-term light acclimation includes state transitions (ST) ( Calzadilla and Kirilovsky 2020 ), decoupling of light-harvesting antenna ( Tamary et al. 2012 ), non-photochemical quenching (NPQ) ( Kirilovsky and Kerfeld 2016 ), heat dissipation ( Demmig‐Adams and Adams 2006 ), activation of futile cycles, such as cycling of inorganic carbon between cells and the environment ( Tchernov et al. 2003 , Müller et al. 2019 ), scavenging of reactive oxygen species ( Pospíšil 2012 ) or shifts in transcriptome ( Luimstra et al. 2020 ). The long-term acclimation involves complex changes in proteome ( Jahn et al. 2018 , Zavřel et al. 2019 ) including ratio of photosystem I (PSI) to photosystem II (PSII) ( Luimstra et al. 2020 ), regulation of synthesis of energy-storing molecules, such as glycogen ( Cano et al. 2018 ) or lipids ( Zavřel et al. 2018c ) and modifications of photosynthetic antenna during chromatic acclimation (CA). The latter process provides a great advantage in the spectrally limited underwater environments; while many strains are restricted to certain spectral niches, the chromatic acclimators can efficiently harvest a broader spectrum of available light and, therefore, significantly increase the area and depth of suitable habitats ( Sanfilippo et al. 2019 ). Up to date, eight types of CA have been recognized (CA0-7). CA1, present in Synechocystis sp. PCC 6803 (hereafter referred to as Synechocystis ), is described below. During CA2, phycoerythrin (PE) is upregulated under green light. CA3 combines complementary upregulation of PE and phycocyanin (PC) under green and red lights, respectively. During CA4, the amounts of phycourobilin and phycoerythrobilin chromophores within phycobilisomes (PBS) are shifted under blue and green lights, without affecting PBS proteins. CA5 and CA6 are related to red/far-red light acclimation. During CA5 the light harvesting is secured by chlorophyll d incorporated in the thylakoid membrane instead of PC-containing PBS that is absent. During CA6 far-red light triggers complex changes such as the induction of far-red shifted allophycocyanin and alternative photosystem proteins with a shift from chlorophyll a (Chl a ) to chlorophyll d / f ( Sanfilippo et al. 2019 ). CA0 and CA7 optimize yellow-green light absorption by regulating rod-shaped PBS (CA0) and phycoerythrocyanin (CA7) ( Hirose et al. 2019 ). While the mechanisms underlying various types of CA are known to great detail, less attention has been paid to the implications of light quality shift on phytoplankton energetics and metabolism. The summary of metabolic changes in response to light signals is called photomorphogenesis ( Montgomery 2016 , Bernát et al. 2021 ) or photoacclimation. Even though, in cyanobacteria, the response to individual wavelengths has been studied to some extent, only a few studies provided a complex understanding to cellular energetics over the entire photosynthetically active radiation (PAR) range ( Bernát et al. 2021 ). This is surprising, since the ability of phytoplankton to effectively harvest available light over a range of spectral niches ultimately determines its abundance in the environment ( Holtrop et al. 2021 ). This study provides a detailed report on light quality acclimation of Synechocystis , a type 1 chromatic acclimator. CA1 possess a specific (i) green light-dependent redistribution of light energy between photosystems through the linker CpcL that binds PBS preferentially to PSI, and (ii) a red light-dependent binding of the PC rods to the allophycocyanin (APC) core through a CpcG1 linker that helps to form canonical PBS (for more details, see review Sanfilippo et al. 2019 ). The red/green cyanobacterial chromatic acclimation sensor (CcaS) that regulates also CpcL is common for type CA1-3 ( Wang and Chen 2022 ), whereas the CpcL linker is present in many strains that are not even classified as CA ( Hirose et al. 2019 ). The light acclimation in Synechocystis thus partially reveals photoacclimation strategies of other phytoplankton species that do not possess CA. One example is a close Synechocystis relative Cyanobium gracile that adopted both common and unique photoacclimation traits ( Bernát et al. 2021 ). In general, the most favorable wavelengths for Synechocystis and many other cyanobacteria are those around the absorption maxima of the PBS (orange/red photons), whereas the least favorable are blue photons. The poor growth under blue light is a result of excitonic imbalance between PSII and PSI. The majority of blue photons is absorbed by PSI, which is typically several times more abundant than PSII ( Murakami et al. 1997 , Luimstra et al. 2018 , Bernát et al. 2021 ) and which binds more chlorophyll molecules in its structure ( Umena et al. 2011 , Netzer-El et al. 2019 ). The under-excitation of PSII limits oxygen evolution and linear electron transport rate ( Luimstra et al. 2018 ), which further limits NADPH production, carbon fixation and ultimately cell division ( Pfennig et al. 2023 ). On the other hand, orange-red light is harvested by PBS that transfer the excitation energy efficiently to PSII ( Li et al. 2021 ) but less efficiently to PSI. Despite the detailed understanding of blue- and red-light driven shifts in transcriptome or photosynthesis efficiency ( Singh et al. 2010 , Luimstra et al. 2018 , 2020 ) far less is known about acclimation to other wavelengths. Recent works reported shifts in photosynthetic efficiency and pigmentation under near UV, green and near far-red wavelengths ( Bernát et al. 2021 , Rodrigues et al. 2023 ). However, reports systematically comparing light acclimation across the whole PAR spectrum ( Pfennig et al. 2023 ) are still scarce. This study provides insight into the regulation of light capture and energy transduction in a CA1 performing cyanobacterium across the visible light spectrum (435–687 nm). The results extend our understanding of phytoplankton limitation in natural environments and can shape the design of new strains as well as cultivation strategies in the controlled cultivation systems.",
"discussion": "Discussion In this work, the light quality acclimation in the model cyanobacterium Synechocystis was studied. The experimental setup covering the whole spectrum of visible light ( Fig. 1 ) represents a significant improvement over many previous studies, where typically only two or three wavelengths were compared. Indeed, in case of PC-containing cyanobacteria such as Synechocystis , cultivation under blue and red light induces two fundamentally distinct light acclimation states (for details, see Singh et al. 2009 , Luimstra et al. 2018 , 2020 ), which was also confirmed here. However, the current study shows that although some parameters shifted between blue and red cultivation lights almost linearly, the wavelength dependency of many other parameters was highly non-linear. In addition, the range of previously studied growth wavelengths was widened by addressing acclimations to violet and near far-red illumination. Both these cultivation lights were utilized better than blue light in Synechocystis , and both led to specific responses distinct from other wavelengths (such as the temporal limitation of linear electron flow on the acceptor site). The relatively fast growth under near far-red light was presumably a result of a relatively broad range of the ‘687 nm’ LED, starting in this particular case at 625 nm ( Fig. 1 ). Absorption of true far-red light (>700 nm) is quite inefficient in Synechocystis , and growth under this light was found as slow as under blue light ( Wilde et al. 1997 ). Moreover, far-red light induces complex transcription shift that generally resembles stress response ( Hübschmann et al. 2005 ) which is also common during blue light acclimation ( Singh et al. 2009 , Luimstra et al. 2020 ). On the other hand, the increased growth rate under violet light (435 nm) compared to blue light (465 nm) was likely a result of higher PUR ( Fig. 1C ). Under violet light, the higher growth rate was accompanied by slightly increased PSI-CEF but not LEF ( Fig. 1D and E ), showing the importance of ATP generation for growth optimization. Both violet and near far-red light resulted in increased abundance of Chl–PSII and PBS–PSII ( Fig. 3A , Supplementary Figure S6 ) that, however, did not lead to higher LEF ( Fig. 1D ). This can be related to reduced capabilities to fine-tune light harvesting ( Fig. 4 ), likely as a result of high PSII/PSI ratio ( Fig. 3E ). In addition, synthesis of xanthophylls but not β-carotene was upregulated under both violet and near far-red lights ( Supplementary Figure S2 ) suggesting alterations in thylakoid membrane structure under both conditions ( Zakar et al. 2017 , Rodrigues et al. 2023 ). Blue light is known to induce a system-wide acclimation response in Synechocystis ( Singh et al. 2009 , Luimstra et al. 2018 , 2020 ). The growth limitation under blue light is a result of the inability to efficiently capture and transfer blue light to PSII, which leads to low LEF and results in insufficient production of reducing equivalents such as NADPH and/or Fd red ( Singh et al. 2009 ) as well as of ATP ( Luimstra et al. 2020 ). The low LEF under blue photons was also found here ( Fig. 1D ) and it was likely a key limiting factor for growth as well as for the accumulation of lipids and carbohydrates ( Fig. 2D and E ). Furthermore, under blue light we found strongly downregulated APC levels ( Supplementary Figure S12 ), which further reduced efficiency of the captured light transfer to both PSII and PSI ( Fig. 3 ). Transcript levels of APC genes were found upregulated under blue light in previous works, together with transcription of PBS degradation genes ( Singh et al. 2009 , Luimstra et al. 2020 ). It is therefore likely that the APC downregulation under blue light found here was related to post-transcriptional or (post-)translational regulation. The APC shortage can also explain the low kinetic rates of ST under blue and green cultivation light ( Fig. 4 ). ST become relevant for light energy distribution only when PBS are assembled in its canonical form [i.e. CpcG1 form ( Kondo et al. 2009 )], which unlikely occurred with APC shortage under blue cultivation light ( Supplementary Figure S12 ). Analogously, the APC downregulation can also explain reduced NPQ under blue light ( Fig. 4 ). The fact that the emission spectra of the 495 nm and 520 nm LEDs overlapped to a big extent, was a likely reason of similar values of many photosynthetic parameters recorded under these bluish lights, including growth rate. Above these growth wavelengths, many parameters changed dynamically. Compared to the ‘495 nm’ and ‘520 nm’ light, we found a pronounced shift under green (555 nm) light in Chl a and carotenoid levels ( Fig. 2 ), in particular, in that of zeaxanthin, echinenone and synechoxanthin ( Supplementary Figure S2 ). Since neither PSI nor PSII was upregulated under green light ( Supplementary Figure S6 ), the most likely explanation for the increased content of Chl a was its higher turnover rate, possibly compensating for the low absorbance of wavelengths around 550 nm ( Fig. 1 ). Since free chlorophyll produces oxygen radicals ( Krieger-Liszkay 2004 ), such increased turnover rate expectedly triggers an overexpression of xanthophyll but not β-carotene, in line with our results ( Supplementary Figure S2 ). We note that the increased xanthophyll levels under 687 nm cultivation light probably had a different origin, since Chl a level did not increase under that light ( Fig. 2 ). The increase in growth rate under 555 nm light, compared to the 495 nm and 520 nm light, can have multiple causes. First, due to its bandwidth, the emission spectrum of the 555 nm LEDs also contained some orange photons that are absorbed by PBS effectively and thus favor growth ( Fig. 1 ). Second, the APC shortage measured under blue light was not present under green light ( Supplementary Figure S12 ). Third, the green light induces a formation of specific PBS form named CpcL-PBS which is able to bind to PSI through a CpcL linker ( Kondo et al. 2009 ). Shifting from blue to green light, the PBS were thus able to (1) absorb more light and (2) transfer the absorbed energy to both PSII and also to PSI via CpcL-PBS, resulting in higher LEF and PSI-CEF, as shown in Fig. 1 . The fastest growth of Synechocystis was observed under orange/red lights (peak wavelengths 633 nm and 663 nm), where LEF, PSI-CEF and REF allowed to generate sufficient amount of ATP and reducing equivalents ( Fig. 1 ). Our results show that this fast growth was achieved not only by high number of photosystems and light-harvesting antenna, but rather by a delicate interplay of many processes. These include, besides the abovementioned factors, also regulation of state transitions ( Fig. 4 ) and synthesis of carbohydrates or lipids ( Fig. 2 ). Indeed, the efficient electron flow through thylakoid membranes under red light led to an increased accumulation of PQH 2 over PQ, relative to other tested wavelengths ( Fig. 5 ). The more reduced state of the PQ pool was likely related to increased ST ( Calzadilla and Kirilovsky 2020 ) as well as to the upregulation of genes of photosynthetic electron transport rate and other compartments ( Foyer et al. 2012 ) as described in great detail previously ( Singh et al. 2009 , Wiltbank and Kehoe 2019 , Luimstra et al. 2020 ). It has to be noted that the obtained results are partly reliant on growth conditions, and the wavelength dependency of growth can change with shifts in temperature, salinity and irradiance intensity ( Zavřel et al. 2017 ), or with the use of ammonia instead of nitrate as N source ( Singh et al. 2009 ). By addressing spectral dependency of light harvesting, electron transport and cellular energy storage, this work provides insight into spectral limitations of PC-rich cyanobacteria of CA1-type under controlled conditions. The results can navigate the design of new strains toward improved light utilization for the synthesis of targeted products. A call for such optimization has been announced ( Klaus et al. 2022 ), and a study addressing the effect of light quality on the synthesis of bulk chemicals has been provided recently ( Rodrigues et al. 2023 )."
} | 4,279 |
34930457 | PMC8690404 | pmc | 1,211 | {
"abstract": "Through connecting genomic and metabolic information, metaproteomics is an essential approach for understanding how microbiomes function in space and time. The international metaproteomics community is delighted to announce the launch of the Metaproteomics Initiative (www.metaproteomics.org), the goal of which is to promote dissemination of metaproteomics fundamentals, advancements, and applications through collaborative networking in microbiome research. The Initiative aims to be the central information hub and open meeting place where newcomers and experts interact to communicate, standardize, and accelerate experimental and bioinformatic methodologies in this field. We invite the entire microbiome community to join and discuss potential synergies at the interfaces with other disciplines, and to collectively promote innovative approaches to gain deeper insights into microbiome functions and dynamics. \n Video Abstract . Supplementary Information The online version contains supplementary material available at 10.1186/s40168-021-01176-w.",
"conclusion": "Conclusions Ambitious long-term plans such as the Human Genome Project, the European Green Deal, and the Human Brain Project have pushed the boundaries of science. The multidisciplinary giant Tara Oceans project aimed at deciphering the complexity of ocean life by means of planetary exploration sampling and huge metagenomics efforts is an emblematic model for environmental microbiology [ 14 ]. The International Human Microbiome Standards Project is also a pivotal project for improving data quality and optimizing omics comparability in the human microbiome field. Such long-range frameworks offer challenging objectives for many researchers, cutting-edge research infrastructure to advance methodologies and knowledge, and a visibility that is attractive for stakeholders, media, and the public. The alliance of the Metaproteomics Initiative and other microbiome disciplines could become the cornerstone of such a long-term, cutting-edge project on the comprehensive molecular analysis of microbiome ecosystem functioning and adaptation. Through its unique positioning, we are confident that the Metaproteomics Initiative will contribute to the best possible returns on investment for microbiome research.",
"introduction": "Introduction of the metaproteome initiative While metagenomic and metatranscriptomic approaches share a common technical measurement platform that is fairly uniform and standardized, the experimental protocol for metaproteomics is much more broad and under active development. This presents challenges for assessing and comparing results and capabilities. In addition, metaproteomics currently benefits greatly by the sequencing, assembly, and annotation of metagenomes for accurate interpretation of the data. Thus, to propel metaproteomics to an unprecedented level of microbiome functional analyses requires an expanded and concerted stage of development where all possible synergies between analytical scientists, protein biochemists, bioinformaticians, statisticians, microbiologists, and other multi-omics specialists should be exploited. The current positive momentum of metaproteomics is particularly ripe for such a shift in gear. Importantly, the microbiome research community needs knowledge dissemination via improved educational resources so that community members are informed of the possibilities metaproteomics can offer and have an understanding of the realistic status of the field. Ideally, the microbiome and metaproteomics community should have easier access to the most recent updates of methodologies through specialized open platforms equipped with state-of-the-art instrumentation and strong expertise in data analysis. The full impact of active research to further improve metaproteomics in terms of methodology and data analysis will only be achieved if we improve its standards, educate the whole community, and foster new conceptual advances, thus facilitating the applicability of metaproteomics to solve a wide range of key biological questions. Clearly, metaproteomics has now matured sufficiently to be a viable component of experimental design in current and future microbiome research projects. A growing community of researchers using metaproteomics had the opportunity to meet at several international symposia (Magdeburg, Germany in February 2016, Alghero, Italy, in June 2017, Leipzig, Germany in December 2018, an online edition in June 2021, and Luxembourg City, Luxembourg in September 2021) and launched specific actions such as training sessions and interlaboratory comparisons. Over the past year, members of this community have regularly discussed a variety of subjects, such as the need for a place where newcomers and experts can discuss and initiate new projects, the desire to forge metaproteomics as a critical methodology of choice for functional analyses, and the need to establish and disseminate the best options and standards for this methodology. This has enabled the development of a roadmap in order to assess the possibilities of tackling more complex projects, such as meta-omics projects and spatio-temporal studies of hundreds of differing conditions involving a large number of samples and replicates to be analyzed at increasing depth and throughput. This international community, which at the time of writing brings together over 120 members from 48 research groups from 15 countries, has proposed a new initiative under the aegis of the European Proteomics Association (EuPA, https://eupa.org), the federation of European national proteomics societies, with a clear mission and vision statement: “the Metaproteomics Initiative promotes dissemination of metaproteomics fundamentals, advancements, and applications through collaborative networking in microbiome research. We aim to be the central information hub and open meeting place where newcomers and experts interact to communicate, standardize, and accelerate experimental and bioinformatic methodologies in this field”. This will be achieved initially through a dedicated website (https://metaproteomics.org) and Twitter account (@MetaP_Init), presentations and other educational resources, an online communication channel, inter-laboratory comparisons, and regular symposia. We are delighted to announce that as of February 2021 the Metaproteomics Initiative was officially launched and supported by EuPA. We are confident that this Initiative will benefit from an open structure, facilitating exchanges and interactions with other disciplines, and favouring additional synergies at the interfaces with other omics, data science, and microbiology. Through this commentary, we invite the entire microbiome community to join this Initiative and promote collective innovative and holistic approaches to understand how microbiomes function. The formulation and execution of a community-driven, multi-lab comparison of established metaproteomic workflows, the “Critical Assessment of MetaProteome Investigation” (CAMPI) study, is the first tangible action of the Metaproteomics Initiative [ 12 ]. CAMPI has compared sample preparation, mass spectrometry analysis, and bioinformatics workflows for data analysis. This benchmarking effort has led to findings which have highlighted the importance of each step of the analytical pipeline. Such a benchmarking effort calls for discussion not only on the needs of standardized experimental procedures and data treatment workflows, but also specific guidelines in data reporting and data sharing. The CAMPI effort has also paved the way for subsequent inter-laboratory assays where advanced methodological questions will be tackled by the strength and multiple expertises of the whole community. In this regard, it is clear that the definition of appropriate standardized biological samples or benchmark datasets presents important tasks for the community. Among the other major activities already initiated by the Metaproteomics Initiative are (i) an inaugural online symposium in June 2021, (ii) its promotion among the microbiome community by means of active communication channels, (iii) gathering tutorials and lectures on metaproteomics for offering simple and free access to high-standard educational resources, and (iv) establishment of a consensual framework to stimulate, initiate and coordinate open working groups conducting projects with defined objectives and outputs. To achieve its objectives, the Metaproteomics Initiative will favor the organization of training courses for newcomers and advanced users, further promote its vision in regions currently under-represented, regularly organize new CAMPI editions open to all volunteers, propose brainstorming sessions and hackathons for solving specific bottlenecks, arrange specific round-tables for improving guidelines, and recommending metrics for high-quality results, and establish a platform where scientists with key scientific questions can meet metaproteomics specialists for smooth and efficient collaboration. As pointed out above, the benefits of a united and stronger metaproteomics community and its successful insertion at the crossroads of many disciplines are huge. Although under the EuPA umbrella, the Metaproteomics Initiative is not limited to a given geographical area and welcomes all positive ideas and feedback. As the puzzling complexity of microbiome samples has been an important driver of methodological improvements in meta-omic sciences, open challenges in metaproteomics are expected to further stimulate significant advancements in mass spectrometry-based proteome analysis and proteome bioinformatics [ 12 , 13 ]. Due to the myriad of possible applications, metaproteomic approaches are increasingly used and the field is becoming attractive to young scientists. Indeed, a new generation of scientists trained in one or more of the interrelated areas of meta-omics, data science, modeling, and microbiology is emerging. Engaging, inspiring, and educating early-career researchers are undoubtedly a rewarding prospect for the microbiome community."
} | 2,517 |
39018406 | PMC466958 | pmc | 1,212 | {
"abstract": "Recent progress in the development of synthetic polymer networks has enabled the next generation of hydrogel-based machines and devices. The ability to mimic the mechanical and electrical properties of human tissue gives great potential toward the fields of bioelectronics and soft robotics. However, fabricating hydrogel devices that display high ionic conductivity while maintaining high stretchability and softness remains unmet. Here, we synthesize supramolecular poly(ionic) networks, which display high stretchability (>1500%), compressibility (>90%), and rapid self-recovery (<30 s), while achieving ionic conductivities of up to 0.1 S cm −1 . Dynamic cross-links give rise to inter-layer adhesion and a stable interface is formed on account of ultrahigh binding affinities (>10 13 M −2 ). Superior adherence between layers enabled the fabrication of an intrinsically stretchable hydrogel power source, paving the way for the next generation of multi-layer tissue mimetic devices.",
"introduction": "INTRODUCTION Hydrogel machines and devices have begun to emerge at the forefront of bioelectronics and soft robotics, on account of their tissue-like mechanical and electrical properties ( 1 – 6 ). While conventional electronics use rigid metallic materials with electrons as charge carriers, hydrogel devices use soft water-infiltrated polymer networks with ions as charge carriers, analogous to biological living systems ( 7 – 10 ). To date, these devices have been demonstrated for use as sensors ( 11 , 12 ), actuators ( 13 , 14 ), diodes ( 15 – 17 ), and power sources ( 18 , 19 ). A major challenge in designing hydrogel devices for bioelectronic and soft robotics applications is to maintain desirable mechanical properties (high stretchability, compressibility, and tissue-like modulus), while simultaneously achieving high ionic conductivities. High ionic conductivities enable improved signal transduction for bioelectronic recording or stimulation, and increase output currents in hydrogel-based devices ( 18 , 20 ). Concurrently, superior mechanical properties improve conformability, durability, and biocompatibility when interfacing with dynamic human tissues and when used in soft robotics ( 21 , 22 ). Supramolecular polymer networks (SPNs) are particularly promising for fabricating hydrogel devices due to the dynamic nature of their cross-links ( 23 , 24 ). The reversible association and dissociation of cross-links can dissipate energy within the SPN and endow the materials with superior mechanical properties including high stretchability, compressibility, injectability, and self-healing ability ( 25 , 26 ). Notably, cucurbit[8]uril (CB[8])–based host-guest interactions have been widely used as supramolecular cross-links owing to the range of accessible binding affinities (10 3 to 10 13 M −2 ) ( 27 ), allowing for highly tunable mechanical and viscoelastic properties in physiological conditions ( 28 – 31 ). For example, we recently reported highly compressible glass-like networks based on CB[8]-enhanced polar-π interactions ( 32 ). On account of the substantial deformability and rapid self-recovery, a wearable pressure sensor was fabricated, which showed high sensitivity across a wide range of pressures, paving the way for the use of CB[8]-based SPNs in bioelectronic skins. However, the intrinsic ionic conductivity of reported SPNs remains low (∼10 −3 S cm −1 ) ( 21 , 32 – 34 ), which has limited their application in hydrogel devices for bioelectronics and soft robotics. The limitation is a result of a lack of mobile ionic species within the network, as most are made using neutral acrylamide-based monomers ( 30 , 32 ). To increase ion density and retention within the SPN, charged monomers can be used. However, as supramolecular cross-links rely on electrostatic interactions, they are sensitive toward highly ionic mixtures and charge screening effects limit the use of such monomers ( 27 , 35 , 36 ). Here, we report a variety of supramolecular poly(ionic) networks (SPINs) that maintain the superior mechanical properties of SPNs, while achieving ionic conductivities up to as high as 0.1 S cm −1 ( Fig. 1A ). Furthermore, the dynamic cross-links enable a stable interface to form between distinct hydrogel layers, which has been shown to be critical for the performance and autonomous self-healing of multi-layer stretchable devices ( 37 ). The formation of a stable interface enables access to a fully stretchable all-hydrogel power source, which can be stretched up to physiologically relevant strains (10 to 50%) while maintaining a stable voltage output. Fig. 1. Design of supramolecular polyionic networks. ( A ) Schematic showing intra- and inter-layer supramolecular cross-linking between adjacent poly(ionic) networks on account of the reversible ternary CB[8] complexes. ( B ) Comparison plot of stretchability and ionic conductivity of the SPINs to previous reports.",
"discussion": "RESULTS AND DISCUSSION We recently reported the synthesis of the molecule BPyVI, which showed high binding affinities with CB[8] ( K \n 1 K \n 2 = 2.3 × 10 13 M −2 ) and could be incorporated with neutral acrylamide-based polymer backbones ( 36 ). We hypothesized that such high binding affinity would enable the integration of supramolecular cross-links with ionic polymer materials. To incorporate the supramolecular cross-links within poly(ionic) chains, BPyVI was complexed with CB[8] and copolymerized in situ with ionic monomers via photoinitiated free-radical polymerization ( Fig. 1A ). The two cationic moieties on the BPyVI guest result in a strong ion-dipole attraction with the CB[8] portal, alleviating charge screening effects from ionic monomers and improving the solubility of the complex ( 27 , 35 ). The enhanced solubility and ultrahigh binding strength between BPyVI and CB[8] resulted in homogeneous incorporation of the CB[8] cross-links within the network across all concentrations studied (1 to 3 M; figs. S2 and S3 and table S1) ( 36 ). Conversely, when a singly charged guest monomer BVI was incorporated, an inhomogeneous cloudy network was formed with CB[8] precipitating out of solution (fig. S3). Free radical copolymerization of the 2BPyVI-CB[8] ternary complex was carried out with a library of different ionic monomers, including anionic [3-sulfopropyl acrylate potassium salt (SPAPS)], zwitterionic {[2-(methacryloyloxy)ethyl]dimethyl-(3-sulfopropyl)ammonium hydroxide (MAS)}, cationic {[2-(methacryloyloxy) ethyl] trimethylammonium chloride (DMAEA)}, and acidic [2-acrylamido-2-methyl-1-propanesulfonic acid (AMPS)] monomers ( Fig. 1A ). This strategy enables a library of backbones with different charges, yet universal 2BPyVI-CB[8] dynamic cross-links, to achieve a “plug-and-play” system for the fabrication of a variety of hydrogel devices. Pendant BPyVI guest molecules on adjacent polymer backbones bind within the cavity of the CB[8] host molecule in a 2:1 molar ratio, forming an interfacial cross-link between the hydrogel layers. The dynamic CB[8] host-guest interactions between polymer chains enables high stretchability and compressibility as well as energy dissipation through reversible association and dissociation of the host-guest complexes. Impressively, the strain at break was >15× the original length for each of the SPINs formed, while a broad range of ionic conductivities were observed. Compared to existing reports in the literature, this opens up a combination of mechanical and electrical properties within a single material, previously inaccessible by traditional covalently cross-linked networks ( Fig. 1B ) ( 38 – 48 ). Increasing the level of ionic conductivity without sacrificing the stretchability of SPNs improves performance and durability of hydrogel devices, for example, when interfacing between stretchable dynamic human tissues and rigid electrodes, or for soft robotic actuators where materials undergo high local strains ( 49 , 50 ). The viscoelastic material properties of the SPINs formed were investigated using oscillatory rheology ( Fig. 2A ). The zwitterionic (MAS) and cationic (DMAEA) networks show lower moduli compared to anionic and acidic polymers (SPAPS and AMPS). This may be due to ion-dipole interactions between the cationic-containing polymer backbones and the negative dipole at the CB[8] portal, affecting the dynamics of the cross-linking ( 27 ). To assess the contribution of supramolecular cross-links on the viscoelastic properties of the SPINs, a series of control experiments were performed using the SPAPS monomer to form a model network ( Fig. 2B ). First, to remove cross-linking between polymer chains, the smaller CB[7] homolog was used as this can only bind one guest within its cavity (fig. S2A). Second, the polymerization was performed in the absence of any CB[n]. The viscoelastic properties of these networks were then studied to elucidate the impact of CB[8] cross-links. Fig. 2. Rheological characterization of the SPINs. ( A ) Histogram plot of G ′ and G ″ values for the four SPINs. ( B ) Frequency-sweep measurements at 20°C comparing the supramolecular networks with CB[8], CB[7], and no CB[n]. ( C ) Schematic demonstrating the ternary complexes acting as sacrificial bonds that rupture under deformation and dissipate energy, which can be further reformed, resulting in the self-healing supramolecular network. ( D ) Continuous step-strain measurements at 20°C of the SPAPS networks. ( E ) Master curve of time-temperature superposition for the SPAPS network with dynamic cross-links. The SPIN formed with CB[8] showed clear gel-like behavior, with considerably higher storage ( G ′) and loss ( G ″) moduli compared to both of the controls, suggesting that CB[8]-mediated cross-linking greatly enhances the strength of the network. Networks formed with CB[7] or without CB[n] demonstrated sol-like behavior at lower frequencies ( G ″ > G ′). The network formed without CB[n] showed a higher modulus than the CB[7] network. By adding CB[7], weak cross-linking through π − π stacking of the phenyl group in the BPyVI guest is disrupted on account of 1:1 binding (fig. S2) ( 30 , 51 , 52 ). The disruption of π − π stacking results in a lower G ′ and G ″ and a shift toward sol-like behavior with a crossover point at higher frequencies. A range of storage and loss moduli could be accessed by tailoring the SPAPS ionic monomer concentration between 1 and 3 M, while maintaining the cross-linking ratio ( X ) at 2.5 mol% (fig. S5). The SPINs showed a variety of additional tissue-mimetic properties, including energy dissipation and self-recovery ( Fig. 2C ). Subjecting the networks to multiple step-strain cycles (1 and 500% strain at 1 rad s −1 ) led to a reversible sol-gel transition and rapid self-recovery ( Fig. 2D ). The networks maintained a stable G ′ and G ″ over successive strain cycles, somewhat analogous to the self-healing characteristics of human tissue. Energy dissipation within the poly(ionic) networks was investigated through time temperature superposition experiments to obtain the activation energy ( E a ) for local chain motion ( Fig. 2E ). Fitting the temperature data to a horizontal shift parameter ( a T ) resulted in an E a of 15.2 kcal mol −1 (fig. S6), comparable to the unfolding barrier of the I27 domain of the human muscle protein titin ( E a = 17.0 kcal mol −1 ) ( 30 , 53 ). The ionic conductivity of the networks was investigated using electrochemical impedance spectroscopy (EIS), in which the SPINs were placed between two electrodes and an AC voltage was applied at varying frequencies ( Fig. 3, A and B ). It was found that the zwitterionic polymer (MAS) showed the lowest conductivity, likely due to the lack of a mobile counterion, which is present in the other networks. SPAPS and DMAEA showed a similar ionic conductivity of ∼0.04 S cm −1 , while AMPS showed the highest ionic conductivity (∼0.1 S cm −1 at a frequency of 1 kHz), possibly due to the smaller and more mobile hydrogen ion. The effect of self-healing on ionic conductivity was also investigated. Each of the SPINs was cut in half perpendicular to the direction of ionic current and given 30 s to self-heal. Through impedance spectroscopy measurements, no notable difference was observed in ionic conductivity before and after self-healing, suggesting that the self-healed interface does not interfere with ionic mobility across each of the SPINs ( Fig. 3A ). Fig. 3. Electrical and mechanical characterization of the SPINs. ( A ) Bode impedance plot comparison of each of the SPINs, showing both pristine and self-healed specimens. ( B ) Ionic conductivity comparison of the four networks, showing pristine (blue) and self-healed (gray) specimens. ( C ) Tensile stress-strain curves of the four SPINs; inset: photographs of a typical SPIN under strain. ( D ) Compressive stress-strain curves for the SPINs over three consecutive cycles. ( E ) Impedance of the SPAPS and DMAEA at 1 kHz during tensile deformation. ( F ) Stress-strain curves of the original and self-healed SPAPS SPIN after different healing times at 25°C. ( G ) Indirect demonstration of self-healing via confocal microscopy of the SPAPS SPIN, one chemically labeled with sulforhodamine B (red), the other with fluorescein (green). We further probed the mechanical properties of the SPINs, through both tensile and compressive testing ( Fig. 3, C and D ). For both bioelectronic and soft robotic applications, the networks should be mechanically compatible with biological tissues, i.e., compliant and stretchable, as well as robust and tough enough to withstand everyday movement ( 3 , 54 ). A similar trend in modulus was observed for tensile testing as rheological measurements, with the poly(cationic) chains having a lower modulus, again possibly due to ion-dipole interactions between CB[8] portal and poly(cationic) backbones. All of the networks showed a high degree of extensibility before fracture (<15× initial length), with the cationic SPAPS network displaying more than 20× extensibility of the initial length. Furthermore, to assess the suitability of the networks in a stretchable device, the impedance of the anionic and cationic gels were also tested under tensile deformations up to 200% strain ( Fig. 3E ). The networks showed a typical impedance response for an ionic conductor, remaining below that of the theoretical Poisson ratio ( 1 ). The networks also showed impressive compressibility and subsequent rapid self-recovery ( Fig. 3D ). The anionic, cationic, and acidic polymer networks showed no fracture at 90% compression and fully recovered within 30 s over three cycles of compression. However, in the case of the zwitterionic polymer network (MAS), water was expelled out of the hydrogel during the compressive test. This may be due to the intra- and interchain ionic interactions from opposing charges on the zwitterionic backbone, driving the polymers toward a coiled structure ( 55 ), and pushing the water out from the network. On account of the dynamic and reversible cross-links, the SPINs also showed exemplary self-healing properties tested by cutting a dumbbell-shaped hydrogel sample in half, re-attaching each side, and performing a typical tensile test ( Fig. 3F ). It was found that within 2 min, the stretchability of the healed hydrogel was already markedly higher than that of human tissue (>500%). Following 2 hours, the hydrogel recovered more than 90% of stretchability compared to the pristine state. While the materials show efficient self-healing ability and high stretchability, there is an intrinsic trade-off with elasticity, which can be observed from extension and retraction tensile curves. A hysteresis loop can be observed, especially at low rates of extension (figs. S9 and S10). The self-healing properties of the hydrogel were also indirectly probed by observing the diffusion of dyes across the healing interface ( Fig. 3G ). Separate SPAPS gels loaded with either sulforhodamine B (red) or fluorescein (green) were placed in contact, and the diffusion of the dyes at the interface was examined over time using confocal microscopy. It was observed that a yellow color formed at the juncture owing to the diffusion of dyes across the interface, indicating effective self-healing within the gel sample. To demonstrate the applicability of the SPINs in hydrogel devices, we developed a soft power source, which utilizes the poly(anionic) and poly(cationic) networks as cationic and anionic selective membranes. Inspired by the electric eel, it has been demonstrated that by interfacing a high-salinity hydrogel, a cation-selective gel, a low-salinity gel, and an anion-selective gel in sequence, a voltage can be formed via reverse electrodialysis ( 18 , 56 – 58 ). The salt concentration gradient between the high- and low-salinity gels allows for Cl − ions to pass through the anion-selective gel and Na + ions to pass through the cation-selective gel, resulting in an imbalance of charges across the cell ( Fig. 4A ). To date, such hydrogel power sources have only used covalent networks and have not been demonstrated under strain on account of poor interfacial stability between layers ( 18 , 19 , 59 ). We hypothesized that dynamic networks would demonstrate superior interfacial stability through the formation of inter-layer cross-links, achieving a fully stretchable, conformable all-hydrogel power source. Fig. 4. Demonstration of the SPINs in a multi-layer hydrogel power source. ( A ) Design of the soft power source, including a high-salinity hydrogel, a cation-selective gel, a low-salinity gel, and an anion-selective gel. ( B ) Fluorescence microscopic images of the interface between the SPAPS and AAm supramolecular polymer networks, chemically labeled with sulforhodamine B (red) and fluorescein (green). ( C ) Adhesion testing of the SPINs compared to covalently cross-linked networks, showing a bar plot of the fracture stress. ( D ) Adhesion testing of the SPINs compared to covalently cross-linked networks, showing a bar plot of the elongation at break. ( E ) Plot of current and voltage in response to various external loads for one gel cell (black) and two gel cells (blue) in series. ( F ) Photographs and voltage measurements of the hydrogel power source under applied strain. To form the high- and low-salinity gels, acrylamide monomer (AAm, 2 M) was copolymerized with 2.5 mol% 2BPyVI-CB[8] monomer complex in the presence of high (2.5 M) and low (0.015 M) concentrations of NaCl (fig. S12). To ensure even strain application across the hydrogel device with stress, additional acrylamide (1 M) was copolymerized with the DMAEA ionic monomer (2 M) to increase the strength of the anionic-selective hydrogel (fig. S13). The interfaces formed between different polyionic networks was probed using dye-diffusion experiments as well as adhesion testing ( Fig. 4, B to D ). Similar to the self-healing test, fluorescein and sulforhodamine B dyes were loaded into polyacrylamide (AAm) and SPAPS gels, respectively, and the diffusion was monitored via confocal microscopy. Faster diffusion was observed compared to just the SPAPS gel alone and, within 3 hours, the dyes were completely mixed at the observation window boundary ( Fig. 4B ). The faster diffusion of dyes is likely due to ions diffusing from the SPAPS to AAm gel. To probe the interfacial toughness that supramolecular cross-links can impart between adjacent hydrogel layers, adhesion testing was conducted and compared to covalently cross-linked networks. Testing was performed between the cation-selective hydrogel (SPAPS), and high- and low-salt AAm hydrogels ( Fig. 4, C and D ). Specimens were cut into rectangular pieces and then brought into contact with the alternative hydrogel, and the force required to detach the two hydrogels at the interface was then measured. It was found that the fracture stress required for the supramolecular gels was more than five times higher than that of their covalent counterpart, suggesting a stronger interface in the supramolecular gel on account of the interfacial association of cross-links between hydrogel layers. The supramolecular cross-links enabled the layered specimens to be elongated by more than 10 times their initial length before fracturing at the interface ( Fig. 4D ). The adhesion of SPAPS to the low-salinity gel was higher than to the high-salinity gel. This is likely due to the effect of salts on the binding dynamics of the 2BPyVI-CB[8] complex (fig. S11). Through 1 H nuclear magnetic resonance (NMR) spectroscopy, it was found that at higher salt concentrations, the peaks associated with the guest begin to broaden and shift downfield, suggesting a change in complexation dynamics, which can be attributed to charge screening effects or competitive binding with ionic species. Interfacing each of the component hydrogels in series, a voltage of ∼115 to 125 mV open circuit was measured, in line with values observed for previously reported covalent eel-inspired hydrogel power sources ( 18 ). We constructed voltage-power curves by connecting a series of known-load resistances to the power source while monitoring the voltage across the load ( Fig. 4E ). Currents of up to 2 μA were achieved when low resistances were connected in parallel. In addition, by connecting two cells in series, it was observed that the open circuit voltage scaled linearly with the number of cells, increasing to ∼248 mV. The voltage of the power source was also tested up to physiologically relevant strains (e.g., 50%), which can be experienced by bioelectronic devices under everyday use ( Fig. 4F ) ( 60 ). In the case of the SPINS, the voltage remained stable (within 10% of initial value), and the interface was preserved between the hydrogel layers. In contrast, the covalent gel showed an interfacial fracture at only 10% strain, leading to a voltage drop down to zero. In summary, we have successfully developed a variety of SPINs based on anionic, cationic, zwitterionic, and acidic polymer backbones. By incorporating supramolecular ternary complexes with ultrahigh binding strengths (>10 13 M −2 ), the reversible association and dissociation kinetics endowed the networks with highly tunable viscoelastic properties, as well as tissue-like self-recovery and energy dissipation. Each of the networks also showed high stretchability (>1500%), while simultaneously achieving high ionic conductivities of up to 0.1 S cm −1 . Furthermore, the dynamic nature of the cross-links enabled a stable interface to form between different ionic polymer networks. As a result, we developed a fully stretchable, multi-layer all-hydrogel power source that could maintain a stable voltage up to 50% strain. We anticipate that such a combination of electrical and mechanical properties will enable the next generation of multi-layer hydrogel devices for integrated bioelectronic platforms including actuators, organic electrochemical transistors, or electrochromic displays in bioelectronics and soft robotics."
} | 5,766 |
39241231 | PMC11420864 | pmc | 1,213 | {
"abstract": "Two-dimensional graphene and graphene-based materials\nare attracting\nincreasing interest in neuromorphic computing applications by the\nimplementation of memristive architectures that enable the closest\nsolid-state equivalent to biological synapses and neurons. However,\nthe state-of-the-art fabrication methodology involves routine use\nof high-temperature processes and multistepped chemical synthesis,\noften on a rigid substrate constraining the experimental exploration\nin the field to high-tech facilities. Here, we demonstrate the use\nof a one-step process using a commercial laser to fabricate laser-induced\ngraphene (LIG) memristors directly on a flexible polyimide substrate.\nFor the first time, a volatile resistive switching phenomenon is reported\nin the LIG without using any additional materials. The absence of\nany precursor or patterning mask greatly simplifies the process while\nreducing the cost and providing greater controllability. The fabricated\nmemristors show multilevel resistance-switching characteristics with\nhigh endurance and tunable timing characteristics. The recovery time\nand the trigger pulse-dependent state change are shown to be highly\nsuitable for its use as a synaptic element and in the realization\nof leaky-integrate and fire neuron in neuromorphic circuits.",
"conclusion": "Conclusion A cost-effective patterning and transfer-free\nlaser-induced graphene\n(LIG) process was used to implement memristive devices directly on\na commercial flexible polyimide substrate. For the first time, volatile\nresistance switching was demonstrated in a LIG-only memristor without\nthe incorporation of additional materials. The fabricated prototypes\n(VLM) were electrically characterized, generating reliable resistance-switching\ncharacteristics. Compared to the previous laser-fabricated graphene-based\nmemristors reported in the literature, the VLM featured an improved\nresistance switching ratio of 20 at a comparable switching voltage\nof ±3 V and an I max / I off ratio around 10 3 with distinguishable resistance\nswitching for 100 cycles. Although promising results were obtained,\nthe VLM exhibited reasonable variability, which suggests that further\noptimization of the contacts and fabrication process is required.\nBased on the clockwise hysteresis shown by the VLMs, corroborated\nwith numerical device simulations, the potential mechanism giving\nrise to the device switching was attributed to the built-in electric\nfield originated by the redistribution of defects and ions. The measured I – V characteristics and the dependence\nof the resistance values on the recovery time, which were found to\nbe a function of the device length, agreed well with the proposed\nmechanism. The dynamic response of the VLM showing the input-pulse-dependent\nstate change was shown to emulate short-term plasticity schemes such\nas paired-pulse depression (PPD). Finally, the possibility to achieve\ngreater control over the material properties by directly tuning the\nlaser power was demonstrated through the implementation of an all-LIG\ncircuit to reverse the switching direction. A simple circuit designed\nto obtain a standard anticlockwise hysteresis shows the potential\nto implement different learning rules and components akin to biological\nsynapses.",
"introduction": "Introduction In recent years, there exist an increasing\ninterest in promoting\na novel computing paradigm able to mimic the complex behavior of biological\nbrains. 1 − 3 At the leading edge of this revolutionary change\nare Spiking Neural Networks (SNNs) which attempt to efficiently translate\nthe biological neural functioning into an artificial hardware. 4 , 5 The existing implementations of SNNs, however, usually involve a\nlarge number of active CMOS components preventing these architectures\nfrom the inherent high-density integration of the real biological\nnetworks. 6 − 8 In order to overcome the CMOS intrinsic limitations,\nseveral emerging devices based on different physical mechanisms have\nbeen proposed, including phase change materials or magneto-electric-controlled\nferromagnets, among others. 2 However, these\nstructures face severe limitations, such as stringent operating conditions\nand unsuitable switching properties, compromising the emulation of\na realistic neural network. In this landscape, resistive switching\nmemristors have emerged as a favorable choice for realizing SNNs with\nminimum footprint. 9 , 10 In particular, two-terminal devices\nusing a wide range of metallic electrodes sandwiching a bulk insulator\nhave already demonstrated both nonvolatile and volatile switching\nbehavior, 11 , 12 emulating, respectively, the synapse retentive\nlong-term plasticity (LTP) 13 , 14 and dynamic short-term\nplasticity (STP). 10 , 15 Recent studies have revealed\nthat layered two-dimensional (2D)\nmaterials, such as graphene, transition metal dichalcogenides (TMDCs),\nand insulating 2D materials, are able to reproduce and enhance the\nfunctionality of memristive devices. 16 , 17 A paradigmatic\nadvantage of 2D materials is found with their deposition on flexible\nsubstrates, 18 − 20 that could find a direct application on future neural\ninterface technologies, enabling, e.g., to connect and interact with\nliving neuronal networks or to recover the processing capabilities\nlost by neurodegeneration. 21 The\nstate-of-the-art fabrication methodology for 2D materials involves,\nnevertheless, routine use of high-temperature, time-consuming processes,\nand multistepped chemical synthesis, as exemplified by many of the\ngraphene- 17 , 22 − 24 and other 2D materials-based 25 , 26 memristors reported in the literature, constraining the experimental\nexploration in this field to high-tech facilities. In this context,\nLaser-Induced Graphene (LIG) has emerged as a cost-effective alternative\ntechnique for the production and patterning of graphene films using\ncommercial laser machines on flexible substrates. 27 , 28 LIG has already shown high electrical conductivity (25 S·cm –1 ), and good thermal stability (>900 °C). 28 , 29 Additionally, patterning- and lithography-free fabrication process\nis allowed on a flexible substrate, with a remarkable control of the\nporosity, defect density, and geometry by adjusting the laser type\nand parameters such as the applied power and speed. 28 , 30 The unique features of LIG have already been employed to develop\nnumerous applications in the area of flexible electronics and energy\nstorage devices such as sensors and supercapacitors. 31 , 32 The implementation of memristors on a flexible substrate both with\nLIG and laser-reduced graphene oxide (rGO) has also been reported. 33 − 36 In particular, in the works by Tian et al. 33 and Fatima et al. 34 laser scribing was\nused to generate rGO from a precursor GO film. In both studies, rGO\nwas used only as one of the electrodes, while the active resistive\nswitching material was HfO 2 and MXene, respectively. Interestingly,\nthe laser-fabricated rGO in ref ( 33 ) has been shown to not exhibit resistive switching\nbehavior on its own. On the other hand, the work by Romero et al. 36 used a laser to reduce the GO film and it did\ndemonstrate resistive switching directly on rGO. However, any work\nusing laser rGO, as the aforementioned, requires additional preprocessing\nsteps to prepare the precursor GO film, not needed in the LIG process.\nAnother study by Enaganti et al. 35 used\nlaser-fabricated graphene by directly scribing a polymer film to demonstrate\na memristor. The memristor reported in this later work showed similar\nnonvolatile bipolar resistive switching behavior as the ones with\nrGO. While, the nonvolatile memristors have a diverse set of applications\nsuch as memories, the volatile self-recovering memristors (as the\none presented in our work) have gained special attention aiming at\ndifferent specialized applications such as selectors for memristive\ncrossbar arrays, 37 artificial neurons in\nneuromorphic computing, 38 , 39 and true random number\ngenerators (TRGN), 40 based on their recovery\ntime. 41 In this work, we employ a\none-step process exploiting a commercial\nlaser to implement for the first time a Volatile-Laser-induced graphene\nMemristor (VLM) on a flexible polyimide substrate. The absence of\nany precursor and patterning mask enables a substantial simplification\nof the process, considerably reducing the manufacturing time and cost.\nThe following is the organization of the paper: First, the structural\nproperties of the VLM are described, followed by a discussion on the\nelectrical performance. The physical mechanism governing the resistive\nswitching is elucidated next by utilizing numerical device simulations\nand experimental measurements. Afterward, the potential of the fabricated\ndevices to emulate some relevant features of biological synapses is\nshown, and finally, the main conclusions of the work are presented.",
"discussion": "Results and Discussion Structural Properties As schematically depicted in Figure 1 a, LIG is generated\nby laser engraving the polyimide film. The chemical structure of the\nemployed polyimide, which is insulating in its pristine form, is a\npolymer chain of sp 3 and sp 2 bonded carbon atoms\nwith C–O, C=O, and C–N bonds (see Figure S1 ). The laser delivers localized heating\nto the substrate, originating the breaking of the carbon–oxygen\nand carbon–nitrogen bonds and giving rise to a rearrangement\nof the residual carbon chains and release of the gaseous byproduct. 30 Figure 1 b, 1 c, respectively, show close-up optical\nmicroscopy and Atomic Force Microscopy (AFM) images of the scribed\nregion corresponding to the dark zone on the flexible polyimide substrate.\nHere, L and W stand for the length\nand width of the VLM, respectively. The laser scriber is arranged\nto engrave a series of closely packed lines on the substrate, as shown\nin the zoomed-in view of Figure 1 b. Figure 1 (a) Schematic representation of the laser engraving process\nresulting\nin closely packed lines on the polymer substrate and the use of SMU\nprobes as contacts. (b) Image of a series of fabricated volatile LIG\nmemristors (VLM). The zoomed image shows a close-up of one of the\nVLM with length L and width W . (c)\nAtomic force microscopy (AFM) image of the engraved surface. (d) Raman\nspectroscopy result of two VLMs obtained from two different laser\nintensity and speed settings. (e) Measured current–voltage\n( I – V ) characteristics of\nthe VLM, starting with positive polarity for the voltage sweep corresponding\nto the forming loop (red line) and subsequent working loop (blue line).\n(f) I – V characteristics of\nthe VLM, starting with negative polarity for the voltage sweep of\nthe forming loop (red line) and subsequent working loop (blue line).\n(g) Repeated I – V measurements\nfor 27 cycles (gray); red line shows the median data. The Raman spectra of the resulting material are\nrepresented in Figure 1 d for two different\nscenarios: first one corresponding to a high speed and low applied\npower (red line) and a second one (blue line) with lower speed and\nhigher power. Both combinations depict the characteristic D, G, and\n2D peaks. 28 The G peak appears at 1590\ncm –1 when the polymer is exposed to low laser intensity,\nand becomes narrower and shifts to 1580 cm –1 for\nhigher power, signifying the re-establishment of the sp 2 carbon bonding and the reduction in the oxygen content. 42 The occurrence of a broader D peak at 1350 cm –1 , associated with the presence of structural defects\nin graphene, is also reduced in intensity and width with increasing\nlaser power. 43 , 44 The most prominent feature of\nthe graphene sheet, the 2D peak, is highly incremented with an increasing\nlaser power, and its appearance at 2700 cm –1 indicates\nthe presence of few layers of graphene. 42 According to the selected laser power and its engraving speed, it\nis therefore possible to tune the material properties and, in principle,\nthe electrical conductivity of the exposed region. Electrical Performance The measured current–voltage\n( I – V ) characteristics of\ntwo VLMs with square cross sections W = L = 250 μm are shown in Figure 1 e,f. First, a triangular voltage sweep with different\npolarity is applied to the pristine sample, initially positive (from\n0 to 6 V and back) for Figure 1 e and negative (from 0 to −6 V and back) for Figure 1 f, respectively.\nThe measured current (dashed red line) shows hysteresis, with its\nvalue switching between two clearly distinguishable resistance states.\nThis first hysteresis loop is characterized by a switching voltage\n( V switch ) of ≈5.2 and −6.2\nV, with a maximum current value of around 74 and 69 mA, respectively,\nin Figure 1 e, 1 f. Note that V switch is defined here as the voltage where the current decreases by 10%\nof the maximum value. After the initial I–V loop, which resembles\nthe forming process usually observed in metal oxide memristors, 12 subsequent triangular biasing signals (with\namplitude ranging from 5 to −5 V) are applied. The resulting I – V cycles (termed as the working\nloops here and plotted as solid blue lines in Figure 1 e, 1 f) present reduced V switch and current with a rather symmetric response.\nIn more detail, V switch ∼±\n3 and ∼±2.5 V and maximum current values of 5 and 4 mA\nare observed in the working loops of Figure 1 e, 1 f, respectively.\nThe off-current of the device, I off , measured\nat V read = 0.1 V, is in the range of 7–0.1\nμA, which provides an I max / I off ratio around 10 3 . These high\noperating currents due to the wide cross-sectional area could, in\nprinciple, be reduced by improving the laser resolution. Notably,\nthe direction of the hysteresis of the VLMs, depicted by the numbered\narrows in Figure 1 e, 1 f, is reversed (i.e., clockwise) as compared to\nthe usual volatile memristors. 25 The VLMs\nstart in a low resistance state (LRS) and switch to the high resistance\nstate (HRS) at V switch . After a recovery\ntime on the removal of the bias, the VLM returns to the LRS, showing\nvolatile and diffusive ionic characteristics. The switching behavior\nis similar for both voltage polarities and is also independent of\nthe initial forming voltage polarity. Figure 1 g shows 27 I – V sweeps of the VLM, demonstrating a stable switching behavior\nwith a resistance ratio R HRS / R LRS ≈ 20. Note that the R HRS / R LRS is estimated for the median I – V data, represented by the red\nsolid line in Figure 1 g. Compared to other laser-fabricated graphene-based memristors reported\nin the literature, the VLM exhibits a superior resistance-switching\nratio with a comparable V switch and the\nsame maximum DC endurance (as shown later). Table 1 compares the performance of the VLM with\nsome previously reported laser-fabricated graphene-based memristors.\nIt should be highlighted that the VLM demonstrates volatile dynamic\nswitching, whereas those reported in the literature in Table 1 are nonvolatile memristors. Table 1 Table Comparing the Parameter of the\nVLM with Some of the Laser-Fabricated Graphene-based Memristors in\nthe Literature device V switch R HRS / R LRS max. DC endurance reference Ag/HfO x /laser-scribed graphene ≈±1.6 10 100 ( 23 ) Ag/laser fabricated GO/Ag ≈±2.5 6 100 ( 36 ) Cu/LIG/Cu ±2 1.5 100 ( 35 ) Cu/LIG-MnO 2 /Cu ±4 1.4 100 ( 35 ) Al/LIG/Al ±3 20 100 this work Memristive Mechanism Unlike high-quality CVD or exfoliated\ngraphene, LIG does not result in a monolayer crystalline sp 2 lattice of carbon atoms. The fabricated LIG contains structural\ndefects and residual ions of oxygen or oxygen-containing species. 30 , 45 The narrowing of the G peak and the reduction in the intensity of\nthe D peak observed in the LIG Raman spectra ( Figure 1 d) suggest a reduction of both the structural\ndefect and the oxygen content with increasing laser power. Therefore,\nto assess the role of the intrinsic defects in the observed resistive\nswitching, I – V measurements\nwere performed for VLM samples fabricated with different engraving\nspeeds in Figure 2 a.\nThe slower one (30 mm/s) results in high conductivity and a lack of\nresistive switching, while the VLM prepared with a higher engraving\nspeed (45 mm/s) shows much lower conductivity and resistive switching.\nThis result supports the crucial role of residual defects and ions\nin generating the resistance-switching phenomena. Figure 2 (a) Measured I – V characteristic\nof the LIG for two different laser engraving speeds: fast (45 mm/s;\nblue dashed line) shows resistive switching, and slower (30 mm/s;\nred line) does not show resistive switching but a much higher conductivity\n(current is limited to 100 mA to avoid structural damages). Switching\ncharacteristics of the VLMs as a function of time where (b) shows\nthe switching from LRS to HRS and (c) showing the recovery of the\nVLMs toward LRS when voltage is kept constant at 0.1 V. (d) Measured\ncurrent (blue line) and voltage (red dashed line) of the LIG sample\nas a function of time with 3 differentiated stages and the corresponding\ndistribution of the ions, extracted from device simulations in (e)\nstage 1, (f) stage 2, and (g) stage 3. Here L is\nthe length, and T LIG is the thickness\nof the VLM. As seen earlier, the VLM exhibits a switch from\nLRS to HRS, indicating\na nonfilamentary mode of operation. 12 , 46 The VLM internal\ndynamics oppose the current flow, leading to a switch from high conductivity\nto low conductivity. Such clockwise/clockwise direction of switching\nfor both voltage polarities could result from the presence of traps\nas well as ions modifying the internal electric field of the device. 46 − 48 To distinguish between the role of traps and the ions in generating\nresistive switching, the dependence of the recovery time on the length\nof the VLM is analyzed. Figure 2 b shows the voltage pulses applied to VLMs of increasing lengths\n( L = 250, 750, and 1000 μm). During the rising\nvoltage sweep, the VLMs are in LRS and a high current flows through\nthe VLM, before switching to HRS. After this switching, the voltage\nlevel is reduced and maintained at a low constant value of 0.1 V to\nmonitor the current flowing through the VLMs without affecting the\ninternal mechanism. Hence, Figure 2 c depicts the current transient evolution for this\nconstant voltage value of 0.1 V after the switching, proving the direct\ncorrelation between the recovery time ( t rec ) and the VLM length ( L ). This observed dependency\ncorroborates the role of drift and diffusion of ions in generating\nresistive switching and rules out the possibility that interface or\nbulk traps cause memristive switching. Additionally, Figure 2 b suggests that t rec can be tuned by adjusting the VLM length, which may\nbe conveniently reduced by scaling down the devices. From the\nabove observations, it is likely that the memristive effect\nis caused by a built-in electric field, 46 generated due to the redistribution of residual defects or ions\nwhen an external bias is applied. To understand and visualize the\nion dynamics during the switching process, additional device simulations\nare performed. An in-house developed numerical tool is employed that\nsolves the electrostatics and time-dependent electronic and ionic\ntransport self-consistently. The system of equations includes the\nPoisson equation and the time-dependent continuity equation for electrons\nand ions in the device. The details of the simulator can be found\nin the Supporting Information . In\nthe simulations, the negative ions (anions) are considered mobile,\nwhile the positive ions (cations) are stationary. To optimize computational\ntime and due to limited information on material parameters, they were\nadjusted to align with the normalized trend of the experimentally\nobtained I – V (refer to Supporting Information ). While not matching the\nexperimental I – V shape exactly,\nthe simulated I – V was able\nto replicate the observed clockwise switching of the resistance state.\nIt is worth noting that reversing the polarity of the mobile species\nalso results in the same switching trend. Based on the internal dynamics,\nthe switching mechanism can be understood as follows. The experimentally\nmeasured transient current under the application of a triangular voltage\nsweep is shown in Figure 2 d, while the corresponding anion distributions obtained from\nthe device simulations at each applied bias are shown in Figure 2 e–g, respectively.\nNote that Figure 2 e–g\nshows the anion difference density with respect to their initial concentration\nat zero applied bias. During the rising voltage sweep (stage 1), the\nVLM is in an LRS and a high current flows through the LIG sample;\nhowever, simultaneously, mobile charged ions drift in the opposite\ndirection of the applied external field. This results in a redistribution\nof the ions and the appearance of a built-in electric field inside\nthe VLM, which opposes the external electric field (stage 2). Therefore,\nthe net electric field perceived by the electrons diminishes and the\nVLM switches to an HRS. As a consequence, the current flowing through\nthe VLM decreases, corresponding to stage 2 in Figure 2 d, where the current is reduced by a factor\nof 10. When the external bias is diminished, during the falling voltage\nramp (stage 3), the external electric field drops, and the ions start\ndiffusing back, as can be seen from the shifting of the Δ anion\npeak in Figure 2 g.\nHowever, since the ion diffusion is slow, the internal field drop\nis not instantaneous; a substantial nonuniform anion density persists,\ncausing the electrons to keep moving under a reduced net electric\nfield. Thus, the I – V characteristics\nfeature a hysteretic behavior. On removal of the bias, ions eventually\ndiffuse back to their initial positions after a recovery time and\nthe VLM returns to the original LRS. Different experiments were\ncarried out to further support the aforementioned\nmechanism. Accordingly, the R LRS value\nshould get reinstated if enough time is provided for the ions to diffuse\nback to their original state, i.e., the recovery time. Then, if repeated I – V sweeps are carried out at time\nintervals shorter than the recovery time, a reduction of R LRS will result after each sweep (in correspondence to\nthe ions not recovering the equilibrium state). This is observed in Figure 3 a: here, the off\ntime (time interval between each voltage sweep) is kept to be less\nthan the recovery time. Similarly, the change in the effective electric\nfield and therefore the R HRS value depend\non the ions drift during the rising voltage sweep. In Figure 3 b, repeated voltage sweeps\nwith peak values of 3.5 4, and 4.5 V, applied with no off time, show\nthat, if the peak voltage applied to the VLM is properly modulated,\ndifferent values of the R HRS can be achieved,\nalso in line with the above mechanism. Additionally, the modulation\nof the I – V characteristics\nwith varying voltage scan rates is observed in Figure 3 c (additional measurements in Figure S3 ), which further corroborates the mechanism.\nAs the voltage scan rate increases, ions are unable to follow the\napplied signal, leading to a reduced switching ratio. Eventually,\nthe hysteresis disappears at sufficiently high scan rates. Figure 3 Repeated I – V measurement\nof the VLM with (a) rest time between subsequent sweeps smaller than\nthe recovery time for the ions thereby reducing R LRS in each cycle, (b) increasing maximum applied voltage\n(red: 3.5 V, blue: 4 V and green: 4.5 V) and hence maximum applied\nexternal electric field thereby increasing R HRS in each cycle and (c) different voltage scan rates (red:\n304 mV/s, blue: 922 mV/s and green: 2.7 V/s). Statistical Measurements Multiple I – V cycles were measured to evaluate the statistical\nvariation of the fabricated VLMs. Figure 4 a shows the device-to-device V switch variation for the first forming loop (FL) and the\nworking loop (WL). Each point in Figure 4 a corresponds to measurements on a different\nfresh device with the boxplot enclosing V switch . Figure 4 b shows\nthe cycle-to-cycle variability of V switch in the WL for two different VLM devices after multiple consecutive I – V sweeps. Although the variability\nis reasonable in device D2, device D1 shows a higher statistical spread\nof the V switch . Figure 4 (a) Device-to-device\nvariability of the VLMs for the forming loop\n(FL) and working loop (WL). Each point corresponds to a different\nfresh device. (b) Cycle-to-cycle variability of two different VLM\ndevices for multiple consecutive I – V sweeps. (c) Endurance characteristics of the VLM showing\ndistinguishable switching for 100 cycles. Further, the endurance is tested by extracting\nthe resistance value\nfrom the full I – V sweep measurements\nperformed on device D1, at a read voltage of 0.5 V. R HRS (blue circles) and R LRS (red squares) for 100 consecutive cycles are displayed in Figure 4 c. Although the resistance\nvalues tend to increase with the cycling process, the VLM shows distinguishable\nresistance states, with a resistance-switching ratio of ≈20\nover the 100 cycles. The full I – V curve for each sweep of the endurance test is provided in Figure S4 . Interestingly, along with the resistance\nvalues, V switch also changes from cycle\nto cycle, resulting into the high variability of D1 observed in Figure 4 b. Circuit Design Application Aiming at testing the applicability\nof the VLMs as components of different circuits, they were connected\nby aluminum sheets. Figure 5 a depicts the I – V characteristics of a VLM device with and without contacts, showing\nthat the device retains its resistance-switching behavior. Figure 5 (a) Demonstration\nof the unaltered I – V characteristics\nafter the use of aluminum sheet as contacts.\nInset shows the fabricated device. Results of the pulse voltage stress\nmeasurements showing the experimental demonstration of PPD for a pulse\nperiod of (b) 1100 ms and (c) 38 ms, respectively. (d) Resistor network\n(inset) employed to obtain the I – V characteristics that shows the possibility of obtaining both, clockwise\nand anticlockwise hysteresis. The dynamic response of the device is analyzed\nby applying a pulse\nvoltage stress (PVS), consisting of a series of consecutive triangular\npulses with different periods as trigger input. Figure 5 b shows the VLM current response when it\nis exposed to a PVS of amplitude 3.5 V and a period of 1100 ms. The\noutput current through the VLM presents a pulse-dependent gradual\nchange, suitable to emulate the relevant paired-pulse depression (PPD)\nfeatured by biological synapses. 10 , 49 After each\nset of 12 (later 8) paired pulse trains, a 550 ms pause is set (equivalent\nto skipping half a pulse) to evidence that the VLM retains its state\nand resumes from the same maximum current value. The VLM can thus\nefficiently emulate the dynamics and characteristic time scale of\nthe STP behavior of biological synapses, where the strength changes\ntemporarily on a time scale ranging from milliseconds to minutes. 15 Moreover, as shown in Figure 3 b, the recovery time depends on the length\nof the device, which, unlike the filament-based memristors, provides\na direct design knob to achieve a higher control of the time scales\nas both faster and slower recovery times can be achieved. In Figure 5 b, the current changes\nfrom 10 mA at t = 0 s to 8 mA at the end of the first\nset of 12 pulses and eventually reduces to 5 mA after a total of 52\npulses. Figure 5 c shows\nthe VLM response for a significantly increased pulse rate where the\nperiod of the trigger pulses is 38 ms. The output current of the VLM\nstill shows a pulse-dependent gradual change (from 7.5 mA at t = 0 s to 6.5 mA after 26 pulses). Both, Figure 5 b,c further confirm the memristive\nmechanism of the VLM, providing additional support to the observations\ndepicted in Figure 2 f, 2 g. Further, the time scale and the\n(pulse-dependent) state change\nof the VLM are highly suitable for the implementation of the LIF neuron, 38 however, it naturally switches from LRS to HRS,\ni.e., in the clockwise direction. This is the opposite to the usual\nbehavior of memristors used to implement the LIF neuron, which have\nanticlockwise switching. The anticlockwise switching in the VLM can\nbe obtained with an all-LIG-based resistive network schematized in\nthe inset of Figure 5 d, which consists of two fixed-value resistors (R1 and R2) in addition\nto the switchable VLM. Notably, both the VLM and the resistors are\nfabricated from LIG, making use of the same procedure described in Results and Discussion section. The laser power\nand velocity, together with the length of the engraved regions, are\nadjusted to achieve either the desired value of the fixed resistor\nor the VLM, using aluminum sheets to interconnect them (see inset\nof Figure 5 d). This\nimplementation further exemplifies the remarkable flexibility and\ncontrol of the LIG fabrication process. If a high current flows through\nthe VLM ( I VLM ) in the LRS, then, the current\nthrough the parallel resistor R2 ( I R2 )\nremains low, while, when the VLM switches to a HRS and I VLM has a lower value, I R2 increases to a higher one. The values of R1 and R2 were carefully\nselected such that the total current in the circuit ( I R1 = I VLM + I R2 ) remains nearly unaltered during the switching. Figure 5 d shows the current\nthrough R2, with the sought anticlockwise hysteresis. This topology\ncan also be combined with a nonvolatile element to implement long-term\nplasticity and the other desired learning rules. 10"
} | 7,382 |
31507164 | null | s2 | 1,217 | {
"abstract": "Stimuli-responsive color-changing hydrogels, commonly colored using embedded photonic crystals (PCs), have potential applications ranging from chemical sensing to camouflage and anti-counterfeiting. A major limitation in these PC hydrogels is that they require significant deformation (>20%) in order to change the PC lattice constant and generate an observable chromatic shift (∼100 nm). By analyzing the mechanism of how chameleon skin changes color, we developed a strain-accommodating smart skin (SASS), which maintains near-constant size during chromatic shifting. SASS is composed of two types of hydrogels: a stimuli-responsive, PC-containing hydrogel that is patterned within a second hydrogel with robust mechanical properties, which permits strain accommodation. In contrast to conventional \"accordion\"-type PC responsive hydrogels, SASS maintains near-constant volume during chromatic shifting. Importantly, SASS materials are stretchable (strain ∼150%), amenable to patterning, spectrally tunable, and responsive to both heat and natural sunlight. We demonstrate examples of using SASS for biomimicry. Our strategy, to embed responsive materials within a mechanically matched scaffolding polymer, provides a general framework to guide the future design of artificial smart skins."
} | 322 |
34850119 | PMC8643692 | pmc | 1,218 | {
"abstract": "Abstract Recent advances in DNA nanotechnology led the fabrication and utilization of various DNA assemblies, but the development of a method to control their global shapes and mechanical flexibilities with high efficiency and repeatability is one of the remaining challenges for the realization of the molecular machines with on-demand functionalities. DNA-binding molecules with intercalation and groove binding modes are known to induce the perturbation on the geometrical and mechanical characteristics of DNA at the strand level, which might be effective in structured DNA assemblies as well. Here, we demonstrate that the chemo-mechanical response of DNA strands with binding ligands can change the global shape and stiffness of DNA origami nanostructures, thereby enabling the systematic modulation of them by selecting a proper ligand and its concentration. Multiple DNA-binding drugs and fluorophores were applied to straight and curved DNA origami bundles, which demonstrated a fast, recoverable, and controllable alteration of the bending persistence length and the radius of curvature of DNA nanostructures. This chemo-mechanical modulation of DNA nanostructures would provide a powerful tool for reconfigurable and dynamic actuation of DNA machineries.",
"introduction": "INTRODUCTION The understanding and application of geometrical and mechanical responses of DNA with binding ligands have been one of the major challenges for decades ( 1 , 2 ). Small molecules non-covalently interacting with DNA have multiple binding mechanisms including intercalation and groove binding ( 3 ), and have been used for pharmaceutical ( 4 ), biological ( 5 ) and fluorescence applications ( 6 ). The underlying binding mechanisms of such ligands and corresponding nanomechanical characteristics of DNA have been widely studied by tweezer-based force spectroscopy ( 7–14 ) and atomic force microscope (AFM) imaging ( 15–17 ). However, most studies so far mainly focused on the response of an unstructured long DNA strand ( 7–13 , 15 ), while the remarkable recent advance of structural DNA nanotechnology enabled the utilization of structured DNA assemblies ( 18 ). A number of DNA nanostructures with planar ( 19 ) to 3D ( 20 ) shapes have been constructed and used as intracellular delivery carriers by containing DNA-binding drugs ( 21–23 ). Chemically-activated dynamic and reconfigurable mechanisms have also been utilized as nanomechanical components ( 24–26 ), molecular sensors ( 27 ) and connectors for higher-order assemblies ( 28–30 ). While chemo-mechanical actuation can provide a fast (up to seconds at strand level ( 2 )) and repeatable geometric response to DNA assemblies, it is difficult to select a proper DNA-binding ligand and predict its working range for a targeted function of DNA nanostructures. In this study, we used three intercalators, ethidium bromide (EtBr), doxorubicin (DOX) and dimeric cyanine dye oxazole yellow (YOYO-1), and two groove binders, 4′,6-diamidino-2-phenylindole (DAPI) and bisbenzimide H 33258 (H33258), to investigate how these DNA-binding molecules affect the mechanical and geometrical characteristics of structured DNA assemblies. Increased mechanical flexibility of the bundles in bending was commonly observed for all binders, but the softening effect was different depending on their intrinsic binding modes. EtBr and DAPI showed a substantial recovery of modulated flexibility by a simple buffer exchange process, thereby enabling repeatable chemo-mechanical actuation. Also, their ability to control the structural deformation was demonstrated for bent structures. Furthermore, we found that the bending stiffness of DNA origami bundles could be systematically varied by using ethidium intercalation together with engineered defects ( 31 ) (via short single-stranded gaps). Computational studies revealed a potential mechanism of how ethidium intercalation make the structured DNA bundles flexible.",
"discussion": "RESULTS AND DISCUSSION Molecules interacting with double-stranded DNA (dsDNA) can change the geometric and mechanical characteristics of structured DNA assemblies, which usually have a complex strand pathway and structural motifs (Figure 1 ). For example, scaffolded DNA origami, one of the most widely used techniques for folding DNA into a nanostructure, typically uses an M13mp18 circular single-stranded DNA (ssDNA) consisting of 7249 nucleotides (nt) as a scaffold and up to 200 short oligonucleotides (staples) which have complementary sequences to the scaffold ( 19 , 40 ). There exist multiple Holliday junctions (crossovers) and backbone breaks (nicks) in order to arrange DNA strands in a given pathway with inter-helical crosslinks to form a structure. Various curved and twisted structures can be constructed through designing lattice-packing rules, crossover intervals, and cross-sectional shapes ( 20 , 42 , 43 ). In this study, a simple six-helix-bundle (6HB) designed on a honeycomb lattice was selected to investigate the chemo-mechanical response to various DNA-binding molecules. Its geometric and mechanical characteristics have been well studied ( Supplementary Figure S1 ) ( 31 , 44 , 45 ), and its slender profile is appropriate to measure the equilibrated shape and calculate the bending flexibility ( 31 ). First of all, we tested five representative DNA-binding molecules having different binding mechanisms to study the mechanical response of the 6HB reference (6HB-Ref) design (Figure 2 ). Assembled 6HB structures were diluted to 0.5 nM with folding buffer containing 20 mM MgCl 2 and targeted binder concentration (Materials and Methods). Images of individual 6HB monomers incubated with different binders and concentrations were collected using AFM, and their mechanical flexibilities were characterized by measuring the bending persistence length from the extracted contours (Figure 2A – F Supplementary Figures S2–S44 , and Supplementary Tables S1 and S2 ). Kurtosis values for representative 15 cases were converged to theoretical value of 3 implying that the deposited structures were equilibrated in 2D ( Supplementary Figure S45 ) ( 33 , 46 ). Figure 1. Schematic illustration of the chemo-mechanical response of a DNA origami nanostructure with DNA-binding ligands. DNA-binding molecules such as ethidium can induce the deformation and the softening of DNA duplexes, leading to the changes in the flexibility and the geometry of an assembled structure. Gray and pink boxes indicate the equilibrated dsDNA without and with ethidium intercalation, respectively. Figure 2. Mechanical response of DNA origami bundles with different DNA binders. Chemical structure of binders, representative AFM images, and 120 aligned monomer contours of 6HB-Ref bundles incubated at the different concentrations with ( A ) EtBr, ( B ) DOX, ( C ) YOYO-1, ( D ) DAPI and ( E ) H33258. Refer to Supplementary Figures S2–S44 and Supplementary Table S1 for the results of the entire concentration range. Scale bars in the AFM images are 200 nm. Scale bars and ticks in the monomer contour graphs are 100 nm. ( F ) Calculated bending persistence lengths of the 6HB structures with different binders and concentrations. The molecular concentration of 6HB monomers was maintained at 0.5 nM for all cases. The dotted lines indicate the exponential fitting curves of experimental data for each case. Error bars indicate the standard deviation. ( G ) Measured monomer contour lengths of the 6HB-Ref bundles with different binders and concentrations. Error bars indicate the standard deviation. ( H ) Recovery test using buffer exchange method. Refer to Materials and Methods for details. In order to confirm that the samples in our measurement condition are fully equilibrated in 2D substrate, we calculated the kurtosis, which is the ratio between the fourth moment and the square of the second moment for the angle distribution ( 46 ). Kurtosis around the value of 3 indicates that the angle distribution is Gaussian, therefore the structure is fully equilibrated in 2D. We performed this analysis to 15 cases with different DNA-binders and concentrations, and concluded that the kurtosis values were generally around 3 within a contour length range. EtBr is a common DNA-binding fluorophore and known to increase the contour length, unwind the twist angle and decrease the persistence length of dsDNA by intercalation ( 7 , 10 , 47 , 48 ). We observed a gradual reduction in the bending stiffness of the 6HB-Ref design throughout the experimental ranges, and the maximum stiffness reduction was calculated as 67.3% in 32 μM concentration (Figure 2A and F , and Supplementary Figures S2–S11 ). The persistence length of DOX-intercalated 6HB-Ref showed a very similar trend with EtBr-intercalated cases (Figure 2B and F , and Supplementary Figures S12–S20 ). DOX is a type of anthracycline antibiotic and anticancer drugs, and known as an intercalating-dominant DNA binder. Geometrical perturbation by its intercalation is similar to that with EtBr ( 47 ), even though a complex force-dependent behavior regarding the persistence length was reported ( 49 ). Besides the mono-intercalators, green-fluorescent DNA staining dye YOYO-1 was chosen as an example of bis-intercalating molecules ( 6 ). Due to the characteristics of bis-intercalation, its effect on elongation and unwinding of DNA are known to be stronger than that of mono-intercalators although contradictory results were reported regarding its effect on the persistence length ( 13 , 50 ). We observed a more drastic reduction in the bending stiffness up to 82.4% in only 1 μM YOYO-1 concentration. This is possibly due to the higher binding affinity of YOYO-1 than EtBr, and more significant extension and unwinding of dsDNA strands induced by bis-intercalation (Figure 2C and F , and Supplementary Figures S21–S25 ) ( 2 , 13 ). Highly stacked configurations by forming long chains were appeared in 2 μM concentration. Intermolecular crosslinking induced by YOYO-1 might be the reason for polymerization since the ratio of dye to DNA bp (∼1:1.8 in 2 μM concentration) largely exceeds the known maximum value (1:4) ( 6 , 51 ). Lastly, the effect of minor groove binders was investigated. DAPI and H33258 are reported to preferentially interact with AT-rich regions, but the intercalative binding mode is also known to exist with GC-rich or mixed sequences ( 52–56 ). In general, minor groove binding mode of ligands is known to induce no significant geometric perturbation, but additional intercalating mode at different sites can influence the changes in geometrical and mechanical characteristics of dsDNA ( 3 ). To investigate the response by such effects at the DNA structure level, DAPI and H33258 were treated to the 6HB-Ref design. The results showed that the persistence length of DAPI-bound 6HB decreased slowly until 4 μM concentration, and then rapidly dropped at higher concentration levels (Figure 2D and F , and Supplementary Figures S26–S35 ). Unlike the intercalator-incubated samples, some of the structures incubated with DAPI at a higher concentration (>8 μM) have locally curved regions, which contribute largely to a drastic decrease of the bending persistence length (Figure 2D and Supplementary Figure S26 ). Meanwhile, H33258-bound 6HB maintained a similar level of persistence length until 8 μM, and then was getting slowly softer at higher concentrations (Figures 2E , F, and Supplementary Figures S36–S44 ). Interestingly, EtBr, DAPI and H33258 ligands have almost no effect on the contour length of the bundle across the tested concentration range, whereas DOX and YOYO-1 showed its increase up to 19% and 22% in 32 and 1 μM, respectively (Figure 2G and Supplementary Table S1 ). Though it has been known that mono-intercalator EtBr tends to increase the contour length of dsDNA ( 10 , 47 ), that of our 6HB structure was not significantly changed by it. More confined configuration of dsDNA in the structured assembly might cause this difference. The reversibility of chemo-mechanical modulation was tested for four DNA-binding molecules (Figure 2H and Supplementary Figure S46 ). Binder-incubated sample solutions were exchanged to pure folding buffer by three times during the centrifugal filtration (Materials and methods) ( 32 ). The recovery efficiency defined as the ratio between the average persistence lengths of initial and recovered 6HB monomers was measured. EtBr and DAPI showed a relatively high recovery rate (78.6% and 91.1%, respectively), whereas DOX and YOYO-1 showed an incomplete, partial recovery (36.1% and 47.5%, respectively). The recovery rate might be closely related to the binding affinity with DNA. It was reported that DOX-intercalated DNA origami structures could contain DOX up to few hours ( 21 ), and YOYO-1-DNA complex could maintain stable binding ( 13 , 51 ). One thing to note here is that the binding affinity and corresponding mechanical characteristics of DNA with binding ligands are reported to be influenced by surrounding buffer conditions as well ( 13 ). In contrast to many single-molecule experiments that use buffers with monovalent cations such as sodium chloride and sodium phosphate, divalent cations such as magnesium are typically used in DNA origami within 10 to 20 mM concentration range ( 40 ). Therefore, a careful selection of the type and concentration of binding ligands might be necessary depending on the purpose and working condition of DNA nanostructures. In addition to the mechanical flexibility, the effect of DNA-binders on the geometry of structured DNA assemblies was investigated by using a curved 6HB (6HB-Arc) design (Figure 3 ). The semi-circular geometry of the 6HB-Arc structure was designed by insertions and deletions of bps to the straight 6HB-Ref design ( Supplementary Table S3 ) ( 43 ). When incubated with EtBr, structures with a smaller radius of curvature were observed in 0.5 μM concentration (Figure 3A and E ). As the EtBr concentration increases (>1 μM), however, most structures were reconfigured into non-circular helical coils. It is because the gradual unwinding of constituting dsDNA strands due to the ethidium intercalation induces a cumulative out-of-plane rotation of the entire structure (Figure 3A and D , and Supplementary Figure S47 ). On the other hand, 6HB-Arc structures incubated with groove binders showed quite distinctive behaviors. The radius of curvature of DAPI-incubated 6HB-Arc gradually reduced up to 32% in 8 μM concentration without noticeable out-of-plane deformation, resulting in the formation of closed circles as shown in AFM images (Figure 3B and D – E , and Supplementary Figure S48 ). Similar to the decrement in the bending persistence length of DAPI-incubated straight 6HB structures, we attribute this behavior to the additional intercalating mode of DAPI that appears in higher molecular concentrations as well ( 57 ). Increased portion of the non-circular structures observed in 32 μM concentration supports this hypothesis that an increased amount of intercalation might have a negative impact on the structure to maintain its planar geometry ( Supplementary Figure S48 ). The decreased radius of curvature of DAPI-incubated 6HB-Arc structures was successfully reverted by buffer exchange, though some of the highly distorted monomers seemed to fail to recover their original shapes ( Supplementary Figure S49 ). On the contrary, H33258 did not induce the change in the radius of curvature until 8 μM concentration, as similar to its effect on the bending persistence length of the straight 6HB-Ref structure (Figures 3C – E and Supplementary Figure S50 ). But a large amount of non-circular structures was also observed in 32 μM concentration ( Supplementary Figure S50 ). Figure 3. Geometrical response of curved DNA origami structures with different DNA binders. Schematic illustration of the initial (gray) and deformed (colored) shapes and representative AFM images of 6HB-Arc structures incubated at the different concentrations with ( A ) EtBr, ( B ) DAPI and ( C ) H33258. Scale bars in the AFM images are 200 nm. ( D ) The ratio of circular shape structures with different binder types and concentrations. ( E ) Measured radii of curvature of the 6HB-Arc structures with DAPI and H33258 binders. The molecular concentration of 6HB-Arc monomers was maintained at 0.5 nM for all cases. Error bars indicate the standard deviation. Next, we explored the possibility of modulating the mechanical stiffness of these bundles by both chemical (via binding molecules) and physical (via engineered defects ( 31 )) modifications (Figure 4 , Supplementary Figures S1 and S51–S74 , and Supplementary Tables S2 and S4 ). In addition to the 6HB-Ref structure, four-helix-bundle designed on a square lattice (4HB-Ref) was tested ( Supplementary Figure S1 ). Flexible versions of these structures achieved using multiple 5-nt-long single-stranded gaps (or engineered defects ( 31 )) were included in the test (indicated as 6HB-Gap and 4HB-Gap) (Figure 4A ). As we increased the concentration of EtBr, a similar trend of stiffness reduction was universally observed for both regular and defect-engineered structures, indicating that chemical and physical modulations can be used simultaneously (Figures 4B – E ). Normalized bending stiffness, defined as the ratio of the persistence length to that of the unmodified structure, showed a similar rate of stiffness reduction for all cases (Figure 4F ). Combined with engineered defects and EtBr intercalation, the maximum amounts of reduction in the bending stiffness were 90.8%, and 85.6% for 6HB and 4HB, respectively ( Supplementary Table S1 ). These values are above the range of maximum softening effect that could be achieved using only engineering defects (67.5% and 70% in 6HB and 4HB, respectively) ( 31 ). Figure 4. Systematic modulation of bending persistence length using EtBr with different cross-sections and defect-engineered bundles. ( A ) Schematic illustration of defect-engineering. For 6HB-Gap and 4HB-Gap design, 5-nt-long ssDNA gaps were inserted in every nick position. Refer to Supplementary Figure S1 for details. ( B ) Representative 120 aligned monomer contours of 6HB-Gap bundles with different EtBr concentration. Ticks and scale bars: 100 nm. ( C ) Representative 120 aligned monomer contours of 4HB-Gap bundles with different EtBr concentration. Ticks and scale bars: 100 nm. ( D ) Calculated bending persistence lengths of 6HB-Ref and 6HB-Gap structures with different EtBr concentration. Error bars indicate the standard deviation. Refer to Supplementary Figures S51–S58 for detailed results. ( E ) Calculated bending persistence lengths of 4HB-Ref and 4HB-Gap structures with different EtBr concentration. Error bars indicate the standard deviation. Refer to Supplementary Figures S59–S74 for detailed results. ( F ) Normalized bending persistence lengths of all cases showing the decreasing behavior of bending persistence length as the increase of EtBr concentration. To understand the structural effect of DNA-binding molecules on structured DNA assemblies, we performed the MD simulation for 105-bp-long 6HB structure (Figures 5A – F and Supplementary Figures S75–S77 ). Considering the uncertainty of ethidium-binding positions on the DNA origami bundle, 60 ethidium molecules were initially distributed into the bundle structure in three different ways; all ethidiums to dsDNA region (Et-dsDNA), all ethidiums to crossover region (Et-Cross), and 30 ethidiums to dsDNA and 30 ethidiums to crossovers (Et-Mixed) (Figure 5A ). According to the previous report ( 24 ), this corresponds to approximately 3 μM EtBr concentration when 0.5 nM DNA bundles exist in solution. During the equilibrium process, nine, four, and five ethidium molecules fell out from their initial positions in Et-dsDNA, Et-Cross, and Et-Mixed cases, respectively. The structure without intercalated ethidium was also simulated (indicated as Ref in Figure 5 ). For all cases, molecular trajectories of final 100 ns were used in analysis. Figure 5. MD simulation and computational analysis of ethidium-intercalated DNA origami bundles. ( A ) Cadnano ( 34 ) diagrams of 105-bp-long 6HB designs with different ethidium positions. For each case, ethidium molecules are initially located in dsDNA regions, crossovers, or evenly distributed in both regions, respectively. Refer to Supplementary Figures S75–S77 for MD simulation snapshots. ( B ) The time-average cross-sectional shape of five planes of 6HB structures with and without ethidium intercalators. The colored regions indicate the bp coordinates at each vertex. Refer to Supplementary Methods for the detailed process (ticks: 20 Å). ( C ) Root-mean-square deviation (RMSD) of MD trajectories for the whole simulation time. (D) Normalized bending persistence lengths of the ethidium-intercalated 6HBs calculated from PCA of MD trajectories. Refer to Supplementary Methods and Supplementary Figure S78 for more information. ( E ) Distribution of time-averaged inter-plane distance values. ( F ) Distributions of the time-averaged rotation angle between plane 1 and plane 5. A negative value indicates the counterclockwise rotation. ( G ) Schematic illustration of the first bending mode shape of the untwisted and twisted 6HB model. 210-bp-long 6HB was used for FE analysis. ( H ) The normalized persistence length of 6HB with respective to the normalized twist angle of dsDNA. ( I ) Schematic illustration of the first bending mode shape of the twisted 6HB models with different dsDNA bending stiffnesses. ( J ) The normalized persistence length of 6HB with respect to the normalized persistence length of dsDNA. P.L.: persistence length. Tw: twist angle. The structural integrity of the bundles was analyzed from the coordinates of bps located at five cross-sectional planes with 21-bp-interval (Figures 5A , B). With ethidium intercalation, an increased amount of fluctuations was observed possibly due to the weakened basepairing interaction. Also, the root-mean-square deviation (RMSD) of ethidium-intercalated cases was higher than that of the reference (Figure 5C ). The normalized bending persistence length was calculated using the eigenfrequency of the first bending mode from principal component analysis (Figure 5D and Supplementary Figure S78 ). Et-dsDNA, Et-Cross and Et-Mixed structures were 51.7%, 33.6% and 35.4% less stiff in bending, respectively, than the reference one without EtBr. From experiments, the corresponding reduction values ranged from 43% to 48% in 2–4 μM EtBr concentration, suggesting that ethidium might be bound to both dsDNA and crossover regions. According to the equilibrated shapes at the final time step, the inter-plane distance was increased for all three ethidium-intercalaed cases (Figure 5E ). However, Et-dsDNA structure was left-twisted whereas Et-Cross design showed a right-handed twist (Figure 5F ). Et-Mixed case showed almost no cross-sectional rotation, as the same number of ethidiums were intercalated to dsDNA and crossover regions in simulation. Since the portion of normal dsDNA regions is usually larger than that of crossovers in DNA origami designs, left-handed twist and coiling of ethidium-intercalated DNA origami bundles were reported in other experiments ( 24 , 58 ). Coarse-grained simulation based on FE analysis was performed to see the effect of global twist of the bundle on its bending persistence length (Figure 5G – J ) ( 41 ). First, we calculated the bending persistence length by changing the intrinsic twist angle of dsDNA to induce various levels of global twist while all the other geometric and mechanical properties were maintained. Overall, the bending stiffness decreased with the intrinsic twist angle of dsDNA (Figure 5G and H ), implying that the global twist reduces the bending stiffness of the bundle. In addition, we changed the bending rigidity of dsDNA. Its effect on the softening of the bundle was bigger for more twisted bundles (Figure 5I and J ). This result indicates that two effects of ethidium intercalation on dsDNA, geometric unwinding and mechanical softening, might synergistically increase the flexibility of structured DNA assemblies. In summary, we investigated the chemo-mechanical response of structured DNA assemblies to DNA binding molecules, providing a simple and effective method to chemo-mechanically control the geometric and mechanical characteristics of DNA nanostructures. All intercalative ligands generally induced the softening of DNA origami structures in bending, while bis-intercalator YOYO-1 exhibited a more drastic effect. Minor groove binders showed a relatively irregular decreasing pattern than intercalators. In terms of controlling the curved shape, DAPI could be used to decrease the radius of curvature with suppressed out-of-plane deformation. Since EtBr and DAPI showed a good recovery rate of bending stiffness after buffer exchange, systematic and repeatable modulation of the shape and flexibility of structured DNA assemblies could be achieved by using them. Also, by exploiting the compatibility of chemo-mechanical modulation with the physical defect design, the controllable range of mechanical flexibility could be further expanded. We expect that many other molecules including drugs, metal ions, and peptides that can interact with DNA and change its geometrical and mechanical characteristics, could be utilized as chemo-mechanical modulators of DNA nanostructures as well. Understanding and utilization of chemo-mechanical response of structured DNA assemblies may advance the method for the design and actuation of molecular machines and chemo-responsive nanomechanical components. Structural deformation and stiffness modulation induced by DNA-binding molecules can realize the unique working mechanism distinguished from DNA strand-based actuation system. For example, we recently demonstrated the threshold-based chemo-mechanical reconfiguration of the DNA origami ring structure ( 58 ). Cumulative internal stress induced by DNA-binding molecules were stored within a structure, and the structure was transformed to the other programmed state when the stress is over the threshold level. The results shown in this work can be utilized to develop more advanced reconfiguration mechanisms and control the working range of them. Also, our findings can provide an important information for the design of DNA-based molecular containers or DNA hydrogels that can be utilized to various biological and medical applications."
} | 6,657 |
37631767 | PMC10459513 | pmc | 1,220 | {
"abstract": "A spiking neural network (SNN) is a type of artificial neural network that operates based on discrete spikes to process timing information, similar to the manner in which the human brain processes real-world problems. In this paper, we propose a new spiking neural network (SNN) based on conventional, biologically plausible paradigms, such as the leaky integrate-and-fire model, spike timing-dependent plasticity, and the adaptive spiking threshold, by suggesting new biological models; that is, dynamic inhibition weight change, a synaptic wiring method, and Bayesian inference. The proposed network is designed for image recognition tasks, which are frequently used to evaluate the performance of conventional deep neural networks. To manifest the bio-realistic neural architecture, the learning is unsupervised, and the inhibition weight is dynamically changed; this, in turn, affects the synaptic wiring method based on Hebbian learning and the neuronal population. In the inference phase, Bayesian inference successfully classifies the input digits by counting the spikes from the responding neurons. The experimental results demonstrate that the proposed biological model ensures a performance improvement compared with other biologically plausible SNN models.",
"conclusion": "5. Conclusions In this work, we introduced a biologically plausible SNN model based on the synaptic wiring method and Bayesian inference. The former approach was implemented by dynamically changing the inhibition weight update, and the latter was based on the statistical properties of neuronal spike firing. The SNN model was designed mainly based on the neuroscientific phenomena of the human brain. The algorithms we developed improved the biological motivation for the lateral inhibition in the inhibitory neuron layer, as well as the inference, when compared to the conventional biological SNN model. In the experiment, the proposed SNN structure achieved high accuracy, outperforming other biologically plausible SNN models. The formation of a neuronal population through the inhibitory weight update process induced population firing, which led to a specific output neuron being more responsive to a specific input pattern. Furthermore, we proposed a Bayesian inference method inspired by biological mechanisms. The conventional inference method just uses the summation of spikes from each output neuron. The proposed inference method is a more probabilistic approach compared to the conventional inference method. In conclusion, the two proposed algorithms, synaptic wiring with inhibitory weight updating and Bayesian inference, led to performance improvement. Our model also showed superior performance with both a small number of neurons and a large number of neurons.",
"introduction": "1. Introduction Neural networks that adopt biologically plausible models have recently become popular in numerous pattern recognition studies [ 1 ]. Popular machine learning (ML) models, such as deep neural networks (DNNs) [ 2 , 3 , 4 ], are inspired by biological neural networks, which mimic the information transmission among neurons in the human brain. However, a significant gap remains between the DNNs and the neuronal learning mechanisms found in biology in terms of the operation of the information transmission [ 5 ]. In contrast to the DNNs, which oversimplify the brain’s mechanism of transmitting signals to the adjacent neurons, the spiking neural networks (SNNs) mimic a variety of biological mechanisms based on the structural similarities between the brain and nervous system [ 6 ]. As fundamental spike-firing models for SNNs, the Hodgkin–Huxley [ 7 ] and Izhikevich [ 8 ] models formulate the biological mechanism of the neurons well. The leaky integrate-and-fire model (LIF) was proposed by Burkitt et al. [ 9 ] with a simpler biological architecture than the aforementioned models for better practical applicability. As a biologically plausible learning method, unsupervised spike timing-dependent plasticity (STDP) was proposed by Bi et al. [ 10 , 11 ] and adopted as a learning rule in several SNN architectures. Although it cannot exhibit biologically realistic long-term potentiation (LTP) or depression (LTD) in Bienenstock–Cooper–Munro behavior [ 10 ], encoded spike trains based on biological mechanisms, such as temporal or rate encoding, can reduce the difference between the classical LTP/LTD mechanisms and STDP [ 11 ]. The rate encoding mechanism is widely used in various SNN architectures, converting information into spikes based on the mean spike-firing rate of the stimulus. In temporal encoding, information is encoded with the precise spike timing corresponding to the stimulus intensity [ 12 ]. Both biological encoding methods demonstrate satisfactory classification performance with the STDP-based spiking neural networks [ 13 ]. Although several SNNs designed based on biological models have been recently proposed, hybrid models of SNNs with DNNs have been suggested for better performance [ 14 , 15 , 16 ]. For example, the structure of convolutional neural networks (CNNs) [ 17 ] has been applied to SNNs, creating “Spiking CNNs” [ 18 , 19 ]. Even though these models had improved classification performance, they were slightly distant from bio-realistic neural networks. Moreover, other biophysical methods rely on supervised learning, where the neurons are provided with label information for classification problems [ 20 ]. Spike-Prop [ 21 ], which first used the backpropagation algorithm, showed that a feedforward SNN using error backpropagation could solve the nonlinear classification problem by solving the extended XOR problem in a time-coded network. Several studies have reported that error minimization based on label information is biologically plausible [ 22 , 23 , 24 ]. In contrast, other studies have argued that the mechanism of neurons learning input stimuli via minimizing the error between the predicted output and the correct answer does not seem to be biologically plausible [ 5 , 25 , 26 ]. For this reason, biologically plausible SNN models inspired by biological mechanisms have been proposed. Khacef et al. [ 27 ] proposed an unsupervised SNN based on a self-organizing map algorithm [ 28 ] mimicking the learning process of the visual cortex. Its network structure determines the firing neuron by selecting the smallest distance between the input stimulus and the weight of the neurons. A heterogeneous spiking neural network (H-SNN) proposed by Xueyuan et al. [ 29 ] analyzes the patterns of long-term and short-term memory using an STDP-based learning process. It performs the prediction of a moving object based on feedforward connections without a recursive structure. Compared with supervision-based SNNs, it yielded similar or better performance. In this paper, we present a new SNN that imitates biologically plausible mechanisms in an unsupervised manner. The learning process in the human brain can be explained as follows: when a group of neurons are stimulated corresponding to an input, their synapses become strengthened as the learning proceeds. Thus, if the same type of stimulus is received repeatedly, the population of neurons that respond to the stimulus work together to fire spikes [ 30 , 31 , 32 , 33 ]. In the proposed SNN algorithm, the group of neurons creating the spikes corresponding to the specific stimulus are wired based on the synaptic wiring method. When a new input from the same category is fed into the network, the synaptically wired neurons create spikes together, which is how the algorithm learns a dataset. The basic architecture of the SNN is inspired by Diehl et al. [ 34 ], whose work is updated with the addition of the following models: inhibition weight updating of the inhibitory neurons, synaptic wiring among the neurons, and Bayesian inference based on a biologically plausible concept. The proposed SNN in Figure 1 comprises three layers: the input neuron layer, first neuron layer, and second neuron layer. First, for the input data, handwritten digit images from the MNIST dataset [ 35 ] or handwritten letter images from the EMNIST dataset [ 36 ] are fed into the input neuron layer in the form of spikes using Poisson encoding. Next, the Poisson spike trains from the input neuron layer are delivered into the first neuron layer, where STDP works as a learning rule in a fully connected fashion. Third, the inhibition weight update method is applied between the neurons in the first and the second neuron layers by changing the values of the inhibition weight dynamically. Finally, the synaptic wiring of the neurons in the first neuron layer is performed. The neurons of the first and second neuron layers act either as presynaptic or postsynaptic neurons. If one neuron stimulates the following neuron, the first is called the presynaptic neuron and the second is called the postsynaptic neuron. A single neuron could have thousands of synaptic connections with many different postsynaptic neurons. Further, the neuronal connections are not static but vary over time [ 37 ]. The neurons that actively respond to the input data learn to fire spikes together once they are grouped together. The suggested SNN model aimed to improve classification performance with the MNIST and EMNIST datasets. After the learning process, Bayesian inference is applied to the final phase of the classification problem, which is also biologically plausible [ 38 , 39 , 40 , 41 , 42 , 43 ]. The remainder of this paper is structured as follows. Section 2 describes the architecture: the proposed layers and neuronal connections. Section 3 details the simulation results in comparison with another bio-plausible unsupervised SNN architecture. Section 4 presents an evaluation of the biological reasoning of our models. Section 5 concludes this paper.",
"discussion": "4. Discussion The proposed two-layer SNN architecture was carefully designed based on the biological plausibility. The following sections address the biological interpretation of the SNN model proposed in this study. 4.1. Inhibitory Neurons In biological neural networks, the excitatory effect of a neuron is offset by the inhibitory neurons by dynamically changing the inhibition weights, thereby ensuring that the excitation of the neurons is controlled [ 68 ]. In contrast to the conventional SNN model proposed by Diehl et al. [ 34 ], our model induces spike firing from the excitatory neurons belonging to the same group using the synaptic wiring algorithm, applying the biological concept of the neuronal population activity to the inhibitory neuron layer. As a result, the two different forces between the excitatory and inhibitory neurons work together to create a rhythmic force. This force between the neurons is crucial for the brain to function smoothly as an information carrier [ 68 ]. Therefore, the dynamically changing inhibition weight between the inhibitory neuron and the excitatory neuron is critically important to produce the aforementioned behavior [ 69 ]. The proposed algorithm models the biological function with the implementation of the dynamic update of the inhibition weight. 4.2. Synapse To interpret the creation of the synaptic connections among neurons, Hebb’s postulate, which states that “cells that repeatedly fire together, wire together”, is widely accepted in the field of neuroscience and related fields [ 70 ]. The proposed SNN model is based on this postulate, as it implements synaptic wiring among neurons. In examining Hebb’s postulate, two important aspects of Hebb’s theory have been noted [ 61 ]. First, the learning process for the neurons takes place “locally”. For example, if two neurons, A and B , are connected with each other, the other neurons may not be considered. Thus, the proposed model only considers the neuronal connection between two neurons; that is, the target neuron and the maximum spiking neuron in the first neuron layer. Second, the neurons involved in the connection must fire simultaneously. The proposed architecture also models the synaptic connection of the neurons firing simultaneously, given the data. Based on these aspects, it can be claimed that the proposed SNN is biologically plausible with regard to mimicking the human brain. 4.3. Biological Inference During the testing process, synaptically connected neurons actively fire together, and the maximum spiking neuron is represented by a group of neurons rather than a single neuron. For the final inference of this neuron group, an additional biologically inspired inference method called Bayesian inference was applied. When humans and other animals are put in a situation where they need to make decisions and control movement with insufficient information, their brains tend to rely on prior knowledge in addition to current observations [ 39 , 40 , 41 , 42 , 43 ]. To implement this model, an appropriate PDF was first chosen (see Equation ( 5 )). In the biological context, Poisson distribution is frequently used to explain the irregular timing of neuron spike firings [ 71 ]. Thus, we applied the Poisson distribution in calculating the probability of the spike firings of each neuron in relation to a given input. To verify this method, it was compared to another popular probability distribution; that is, the Gaussian distribution. The comparison of the two models is shown in Table 5 , where the results clearly demonstrate that the Poisson PDF-based model outperformed the other. As presumed, the Poisson PDF is suitable for the prediction of spike events occurring in a given interval of time [ 66 ]. Figure 7 illustrates an example of spiking responses from a single neuron trained with the class-one dataset. The x-axis represents the number of spikes, and the y-axis represents the probability of the occurrence of each spike. Note that the overall spike firing distribution resembles the Poisson distribution."
} | 3,470 |
35064082 | PMC8795499 | pmc | 1,221 | {
"abstract": "Significance It is well known that c-di-GMP concentration rises in surface-sensing bacteria and functions as a “molecular switch” for biofilm formation. Here, we provide an important recasting of this picture: Intracellular c-di-GMP signals do not just increase in surface-sensing bacteria; such signals are cooperatively broadcast across multiple generations of cells in a lineage with oscillations that undergo both amplitude and frequency modulation, which are controlled by the coupling between pili appendages and c-di-GMP synthesis machinery. The right “tuning” of these signals in terms of frequency and amplitude correlates ultimately to surface commitment. Amplitude and frequency modulation of c-di-GMP signals allows encoding of more complex instructions. Thus, our work provides a more nuanced understanding of how c-di-GMP signaling drives surface commitment.",
"discussion": "Discussion Our findings here paint a more complex and nuanced understanding of cell surface commitment in response to c-di-GMP signaling. By using dimensionality reduction techniques, such as PCA and FA, we can correlate different parameters that characterize c-di-GMP oscillations, including the mean, amplitude, and period, to various quantitative indicators of surface commitment, such as how much of a bacterial lineage persists on the surface. We find that surface commitment in young biofilms is correlated with lineages with elevated c-di-GMP levels and lower surface motility, in agreement with numerous other studies ( 24 , 37 , 38 ). However, we also find that swift surface commitment for the first 10 to 20 generations also correlates with multigenerational oscillating c-di-GMP levels with short (∼4 to 6 h) rather than long (∼10+ h) oscillation periods. Finally, we find the surprising result that the behaviors exhibited by different individual branches of the family tree are not independent from one another but rather correlated with the collective behavior of the entire multibranched lineage. That is, “family matters” in biofilm formation and appears to facilitate cooperativity in surface commitment behavior. In the time progression of c-di-GMP signals, we note that oscillations clearly appear even when they are averaged across different branches of a lineage, rather than being “washed out” by averaging. This suggests that at least part of the c-di-GMP signal in different branches of a lineage initiated by a single progenitor cell are approximately in phase and thereby contribute to emergence of cooperative surface commitment behavior in a statistical manner ( 25 ). The c-di-GMP signal can in principle be impacted by differences in local environmental conditions (such as surface inhomogeneities) experienced by different branches of the lineage. This makes the observation of a persistent common oscillatory c-di-GMP signal across branches of a lineage that propagates from progenitor to progeny all the more striking. In fact, “decoherence effects” can be seen as oscillations gradually become out of phase in different branches, in that cells in more widely separated branches of the family tree tend to behave more differently. An example of this decoherence can be seen in SI Appendix , Fig. S1 . In choosing reporter systems for observing real-time behavior in live single cells, we had tried to balance two key considerations in the experimental design. One is the response time of the reporter, which is related to the time lag between the readout (e.g., fluorescence) and the actual activity being reported (e.g., intracellular c-di-GMP levels). The other is the half-life of the fluorophore, which is related to the time lag between the signal’s activation and the signal’s decay. Both can introduce systematic errors by broadening and distorting the temporal features of the measured fluorescence behavior. Response time contributes to a systematic error in the form of time lag between the readout and the reported activity, while half-life contributes to a systematic error in the form of a time-dependent background signal and a higher baseline level. In general, the longer the response time and/or half-life, the more the errors contribute. For example, by not decaying quickly, the use of a long half-life reporter will enhance the apparent rate of fluorescence signal increase at the beginning of an oscillation and reduce the apparent rate of fluorescence decrease at the end of an oscillation. The transcriptional reporter used in our experiments works via c-di-GMP binding to FleQ, a transcription factor for the cdrA gene; c-di-GMP binding to FleQ causes this transcription factor to dissociate from the cdrA promotor and thereby facilitate derepression of gfp transcription. One contribution to the noted time lag is the maturation time of the fluorescent protein, which is about 4 min ( 39 ). Another contribution to reporter time lag is that associated with c-di-GMP binding to FleQ and derepression of gfp transcription. The dynamics of these processes are not as well known but are expected to be significantly less than a division time. We estimate the reporter time lag to be about 10 to 30 min, which is smaller than the time scale of the oscillations, which have a period on the order of about 5 h for WT (with some mutants exhibiting smaller oscillation periods of at least 2 h). We note that GFP half-life can be quite variable from system to system, since protease activity can depend on many factors, such as growth conditions and strain identity ( 40 ). * We are currently in the process of obtaining more precise measurements of the reporter time lag, which will likely be subject to the intrinsic stochasticity of GFP transcription and translation. The half-life of the GFP used in the c-di-GMP reporter deserves comment. Our calibration experiments show an estimated reporter half-life of 3.6 ± 1.3 h ( n = 5) in P. aeruginosa PA14 ( SI Appendix , Fig. S5 ), which is shorter than reported values in Escherichia coli ( 40 ). We chose to use a fluorescent protein with a longer half-life so that we could have a higher signal-to-noise ratio and a more stringent test of the existence of oscillations. The long-half-life reporters will tend to be stable longer in the cells and therefore contribute to the signal longer, resulting in a higher signal-to-noise ratio but at the cost of having an increased signal baseline. Moreover, long-half-life reporters will also tend to obscure decreases in c-di-GMP levels. That we observe oscillations in reporter intensity, with clear increases and decreases in signal level, is a strong indication that oscillations in c-di-GMP levels exist. The trade-off is that we expect potentially distortions to the precise line shape of the oscillations. Finally, we consider how the specific choice of a cdrA -based transcriptional reporter influences our results. In our experimental design, we chose to use the plasmid-based cdrA transcriptional reporter as an indicator of the presence of c-di-GMP to facilitate comparison with other experiments, in part because it is one of the most widely used reporter systems. This choice of using a cdrA transcriptional reporter involved some trade-offs. CdrA is a c-di-GMP–regulated adhesin involved in biofilm formation, and its expression level is unlikely to be constant in time ( 41 ). This implies that there can be a changing landscape of competition for c-di-GMP binding between the sensor and natural cellular processes, due to induced changes in cdrA expression during biofilm formation, especially those triggered by interactions between bacteria and EPSs, such as Psl ( 42 ). Via this competition, we expect monotonic increases in cdrA expression to weaken the fluorescence readout rather than generate oscillations. In the extreme scenario of oscillatory cdrA expression that is large enough to drive oscillations in the reporter signal via such binding competition, we note that cdrA expression is itself regulated by c-di-GMP, so oscillations in cdrA expression imply the existence of oscillations in c-di-GMP. Although these potential distortions are difficult to completely eliminate, we have tried to minimize these types of effects by using P. aeruginosa PA14 rather than PAO1, which does not leave Psl EPS trails on the surface to interact with other bacteria and thus does not have potential modulation of cdrA expression from Psl-induced up-regulation of DGCs and c-di-GMP ( 42 ). It is possible in the future to use different reporter systems: Fluorescence-based reporters are still the main tools for probing real-time dynamics of small molecules in live single cells ( 43 ). Besides plasmid-based reporters, other types of reporters include riboswitch- and fluorescence resonance energy transfer (FRET)-based reporters. Riboswitches are regulatory elements of untranslated RNA that can bind ligands such as c-di-GMP and control gene expression. FRET reporters emit a fluorescence signal when the ligand binds and causes a conformational change in the receptor. At present, all of these reporters have trade-offs between response time, signal-to-noise ratio, and dynamic range. Riboswitch and plasmid reporters are expected to have similar issues regarding the time lag of gene expression, while FRET reporters generally have low signal levels and limited dynamic range. Nevertheless, we plan to investigate multigenerational c-di-GMP oscillations using these and new biosensors on the horizon. The existence of c-di-GMP oscillations that can be amplitude- and frequency-modulated imply that intracellular c-di-GMP levels are maintained between maximal and minimal limits for many generations, likely with ongoing corrections from the cell’s c-di-GMP regulatory networks, which in turn depends on how quickly intracellular c-di-GMP levels transduce to c-di-GMP consequences relative to the oscillation period: If the former is slow compared with the latter, then the oscillations will be averaged out and a single uniform population will result. On the other hand, if response to changing c-di-GMP is fast relative to the oscillation period, it is tempting to hypothesize that c-di-GMP oscillation between high and low values can in principle result in a diverse rather than uniform population, guided by a kind of “division of labor” ( 9 ) between cells “born” at different parts of a lineage, some with behaviors associated with low c-di-GMP, such as increased motility, and some with behaviors associated with high c-di-GMP, such as increased biosynthesis of EPS. Interestingly, we note a positive statistical correlation between the amplitude of the c-di-GMP oscillation and the mean c-di-GMP level, so that lineages with elevated average c-di-GMP levels tend to have larger oscillation amplitudes, and vice versa, reaching both higher and lower values depending on the size of the oscillation amplitude relative to the average level. This correlation implies that in some lineages there can be forms of amplitude modulation of the c-di-GMP signal that leads to c-di-GMP decreases and forms of c-di-GMP increases that lead to uncontrolled amplitude modulation, both of which can result in potential instability in the biofilm-forming trajectory. Large amplitude oscillations in the c-di-GMP signal can result in extended periods of low c-di-GMP levels even if the mean of that signal is high. These large fluctuations in c-di-GMP levels can be mitigated via another key parameter for surface commitment in the form of the oscillation period: Lineages with elevated c-di-GMP levels and larger oscillation amplitudes can in principle be brought under control by having the c-di-GMP signal be frequency modulated to have a faster feedback response in the cell’s regulatory networks governing c-di-GMP levels. Furthermore, unlike the amplitude, the oscillation period is relatively uncorrelated to other parameters, such as the mean c-di-GMP level, so that frequency modulation provides an additional independent level of control on top of amplitude modulation for surface commitment. Indeed, this phenomenon can be observed in the subset of lineages that successfully commit to the surface, which tend to have short oscillation periods in the intracellular c-di-GMP levels. Short c-di-GMP oscillation periods imply that the signal can approach the average quickly no matter what its current value is, thus allowing the cell to respond to large c-di-GMP fluctuations and to have more opportunities to be at the average value. Moreover, having short oscillation periods implies a network fast enough to rapidly respond to environmental or other cues. Our analysis is also consistent with the observations of heterogenous levels of c-di-GMP observed across the population of a single strain when observed at a single-cell level ( 4 , 9 , 30 , 31 , 43 – 46 ). Consistent with the points raised above, we observe that mean c-di-GMP levels and surface motility are not always inversely correlated, which is one of the guiding principles of the present c-di-GMP signaling paradigm (i.e., low mean c-di-GMP and increased surface motility are correlated with lower surface commitment, and vice versa) ( 24 , 37 , 38 ). We see evidence of this phenomenon in the PilO–SadC interaction mutants, which have conflicting independent surface commitment outcomes, such as low mean c-di-GMP and decreased surface motility for the HI PilO–SadC interaction mutants, or high mean c-di-GMP and increased surface motility for the LO mutant. Furthermore, comparing these data with bulk biofilm assays can give some insight into the relative importance of these correlations to surface commitment. Compared with WT, the HI mutants have decreased bulk biofilm formation, while the LO mutant does not have significantly decreased bulk biofilm formation. This comparison suggests that mean c-di-GMP is more important than surface motility for surface commitment, which is consistent with mean c-di-GMP’s explaining a larger portion of the total variance of the data. Thus, here we demonstrate not just the importance of c-di-GMP oscillations for surface commitment (statistically ∼70% as important as c-di-GMP increases) but also how such oscillations can be controlled via the important PilO–SadC coupling between T4P and a key hub DGC, as well as how too little or too much PilO–SadC coupling can lead to dysregulation of c-di-GMP amplitude and frequency modulation and thereby impact surface commitment. Our findings are most striking in the case of SadC, given that the Δ sadC mutants tend to have lineages with the highest c-di-GMP levels but with long oscillation periods and large amplitudes that allow prolonged exposure of cells to both high and low c-di-GMP levels. Thus, while some cells in the sadC mutant population show high c-di-GMP levels, the dysregulation of the other contributing factors, long oscillation periods and large amplitudes, as well as increased surface motility, contribute to the bulk observation that the Δ sadC mutant has a strong biofilm formation defect and lower bulk c-di-GMP levels ( 31 ). Finally, our data suggest that the high c-di-GMP levels observed in some Δ sadC mutant cells is apparently due to RoeA’s overcompensating for the absence of SadC and thereby leads to c-di-GMP signaling dysregulation. In the Δ sadC Δ roeA double mutant we now observe the expected results of lineages with a drastic decrease in c-di-GMP levels and reduced period and surface motility. Of note, there are likely to be other DGC proteins involved in early signaling, as we do not observe complete absence of c-di-GMP levels even in the double-null mutant. In a more general compass, these results suggest that successful surface commitment and eventual biofilm formation entails precise control of c-di-GMP levels during surface sensing via a complex feedback regulatory network of proteins that include the entire repertoire of DGCs and PDEs. In support of our findings here, there are multiple reports illustrating examples where behaviors typically associated with increased c-di-GMP levels (i.e., biofilm formation) occur when levels of this dinucleotide remain unchanged or even decrease, and vice versa. For example, studies from the Sauer laboratory showed a transient increase in c-d-GMP levels triggered by the DGC enzyme NicD during dispersion in response to glutamate treatment ( 47 ); while c-di-GMP levels eventually drop during dispersion, the initial increase in c-di-GMP is unexpected. The Sauer group also showed that, despite the low level of c-di-GMP found in a strain mutated for the PA3177 gene, the PA3177 mutant strain displays a WT biofilm architecture ( 48 ). As another example, a study by Gomelsky and coworkers showed that, as expected, high levels of c-di-GMP promote EPS production, but this results in reduced biofilm formation on plastic ( 49 ). Furthermore, the dual domain protein FimX has a degenerate DGC and contested PDE activity; interestingly, a recent publication found that the Δ fimX mutant of P. aeruginosa produces more EPS but less biofilm biomass than WT over 48 h in a flow cell system ( 50 ). With regard to motility, the DGC DgcA is required for gliding motility by Bdellovibrio bacteriovorus ; the Δ dgcA mutant shows decreased global c-di-GMP and is nonmotile because it lacks flagella ( 51 ). Furthermore, Alexandre and coworkers showed that a chemotaxis receptor in Azobacter enhances motility in response to temporary increases in c-di-GMP. When bound to c-di-GMP, the receptor promotes persistent motility by both increasing swimming velocity and by decreasing swimming reversal frequency ( 52 ). Another example can be found in E. coli , where c-di-GMP binding to the receptor YcgR, which then binds to the FliG component of the flagellar switch complex, promoting smooth swimming by, again, decreasing swimming reversal frequency ( 53 ). Finally, and analogous to our findings here, Sauer and coworkers found that GcbA, a DGC, contributes to reduction of swimming motility, likely via suppression of flagellar reversals without an associated increase in biofilm formation ( 54 ). Of course, as we further explore the mechanistic basis of these observations other explanations may come to light, including local-versus-global pools of this second messenger or as-yet-unknown c-di-GMP signaling processes, but we posit that our augmented paradigm for understanding the influence of c-di-GMP present here may explain some of these observations."
} | 4,625 |
27704470 | PMC5050174 | pmc | 1,222 | {
"abstract": "Biological technologies for recycling rare metals, which are essential for high-tech products, have attracted much attention because they could prove to be more environmentally friendly and energy-saving than other methods. We have developed biological recycling technologies by cell surface engineering for the selective recovery of toxic heavy metal ions and rare metal ions from aqueous wastes. In this study, we aimed to construct a unique biological technique to recover rare metals ‘in solid’ form by reducing rare metal ions, leading to a practical next-generation recovery system. Sulfate-reducing bacteria (SRB) can reduce Pt(II) to Pt(0), and hydrogenases of SRB contribute to the reduction. Therefore, we constructed yeasts displaying their hydrogenases on the ‘cell membrane’, and reduction experiments were performed under anaerobic conditions without any electron donors. As a result, hydrogenase-displaying yeasts produced black precipitates in PtCl 4 \n 2− solution. Based on X-ray photoelectron spectroscopy (XPS) and transmission electron microscopy (TEM) observations, the constructed yeasts were found to successfully produce the precipitates of Pt(0) through the reduction of Pt(II). Interestingly, the precipitates of Pt(0) were formed as nanoparticles, suitable for industrial usage.",
"introduction": "Introduction Rare metals are essential for high-tech products such as cell phones and automobiles, leading to ever-growing demands today. However, a stable supply of rare metals is difficult on a global scale because rare metals are expensive and localized, and their abundances are extremely low. Therefore, technologies for recovering rare metals from waste products, industrial wastewater, and sea are required. Recently, various physical, chemical, and biological methods have been studied to prepare metallic forms for the recycling of rare metals (Du et al. 2008 ; Fredrickson et al. 2008 ; Cueto et al. 2011 ). However, physical and chemical methods have some disadvantages. For example, they involve toxic solvents and consume high energy. On the other hand, biological methods can produce metallic forms in a more environmentally friendly and energy-saving manner than other methods, because they only require microbes as the basic constituent. Therefore, biological methods have currently attracted much attention, and various microbes that produce metallic forms have been screened on the earth (Klaus-Joerger et al. 2001 ; Konishi et al. 2006 ). Platinum, especially, is a rare precious metal used in catalytic converters or fuel cells. However, there are growing concerns about platinum exhaustion by 2050 because its crustal abundance is extremely low. Additionally, the Republic of South Africa accounts for about 70 % of platinum production. Therefore, developing efficient technologies to accumulate and recycle platinum is desired to ensure a stable and continuous supply of platinum. It has been reported that sulfate-reducing bacteria (SRB) contribute to the bioremediation of toxic metal ions such as chromate or uranium by reducing them (Michel et al. 2001 ; Payne et al. 2002 ). SRB are anaerobic prokaryotes found ubiquitously in nature and many studies of SRB have been conducted on the genus Desulfovibrio . It became clear afterwards that SRB can reduce tetrachloroplatinate (PtCl 4 \n 2− ) in aqueous solutions to elemental platinum [Pt(0)] and that hydrogenases on the cell membrane of SRB contribute to the reduction reaction (Riddin et al. 2009 ). We have developed biological recycling technologies using cell surface engineering for the selective recovery of rare metals and toxic heavy metals causing environmental pollution (Ueda and Tanaka 2000 ; Kuroda and Ueda 2010 , 2011 ; Nishitani et al. 2010 ; Kuroda et al. 2012 ). In cell surface engineering, any metal-binding proteins/peptides can be displayed on the yeast cell surface with maintenance of their activities. Furthermore, ‘cell membrane’ display of target proteins (Hara et al. 2012 ) has been achieved by fusing the target proteins to plasma membrane proteins such as Yps1p (Gagnon-Arsenault et al. 2006 ). The adsorbed metal ions can be easily recovered from yeast cells, because they are adsorbed on the yeast cell surface and a metal recovery process does not need breaking down the yeast cells. Therefore, we can repeatedly use the surface-engineered yeast cells as bioadsorbents in the next reaction. In this study, we attempted to construct yeasts that display hydrogenase on their ‘cell membrane’ and use the constructed yeasts to recover platinum in ‘solid form’ by reducing Pt(II) to Pt(0). Yeast cells grow faster than SRB and are easily handled, because of the simple medium required to culture yeast cells and their safe biological features as compared to those of SRB. Therefore, it is likely that hydrogenase-displaying yeasts could recover platinum more effectively than SRB. In addition, platinum is mainly used as nanoparticles in catalytic converters of automobiles. Our method to recover platinum as nanoparticles, not simple solid forms, using hydrogenase-displaying yeasts could be useful as a next-generation technology for efficient metal recovery.",
"discussion": "Discussion We constructed yeasts displaying hydrogenase on their cell membrane and achieved the recovery of Pt(II) as nanoparticles using the constructed yeasts. So far, the concentration recovery of metal ions from aqueous solutions has been performed by the surface-engineered yeasts displaying the proteins/peptides that bind metal ions (Kuroda and Ueda 2010 , 2011 ). In addition to this strategy, recovery of metals in a solid state from an aqueous solution is a remarkable strategy for practical use, in terms of the usability of the recovered metals. Especially, platinum is mainly used as nanoparticles in industries such as those producing catalytic converters of automobiles. In the case of platinum recovery as ions, an additional process for reducing the ions to nanoparticles is necessary. Therefore, our biological technique used to recover Pt(II) as nanoparticles can save cost and time, and is ecological because any electron donors are unnecessary in this system. According to the result of XPS analysis (Fig. 4 ), the black precipitates produced by hydrogenase-displaying yeasts included Pt(0) with binding energies of 70.7–71.3 eV, suggesting that the hydrogenases displayed on yeast cell membrane contributed to the Pt(II) reduction to Pt(0). In addition, Pt(0) produced by hydrogenase-displaying yeasts formed nanoparticles with lattice fringes and the size of the platinum nanoparticles was about a few nanometers in diameter (Fig. 5 b). Although XPS analysis revealed that hydrogenase-displaying yeasts produced Pt(0), there is a possibility that black precipitates produced by hydrogenase-displaying yeasts contained other platinum compounds such as PtO or Pt(OH) 2 . Although further analysis is required to demonstrate the composition of the black precipitates, the produced black precipitates could be used as material in fuel cells or catalytic converters. Platinum was used as a reduction target metal in this study. However, there are some reports that SRB could recover Au(III) as Au (0), in addition to Pt(II) (Creamer et al. 2006 ). If hydrogenases of SRB contribute to the reduction of Au(III), hydrogenase-displaying yeasts also could reduce Au(III). Furthermore, there are five kinds of metals bearing chemical and physical properties similar to those of platinum [ruthenium (Ru), rhodium (Rh), palladium (Pd), osmium (Os), and iridium (Ir)]. These five metals and platinum are together called platinum group metals (PGMs), and are important because of their catalytic properties (Kettler 2003 ). Owing to the similarity in their properties, hydrogenase-displaying yeasts could reduce these metal ions. Therefore, the constructed hydrogenase-displaying yeasts could be available as an environmentally friendly technology for the recovery of platinum group metal ions."
} | 1,998 |
38980052 | PMC11334424 | pmc | 1,224 | {
"abstract": "ABSTRACT Metagenomic sequencing has advanced our understanding of biogeochemical processes by providing an unprecedented view into the microbial composition of different ecosystems. While the amount of metagenomic data has grown rapidly, simple-to-use methods to analyze and compare across studies have lagged behind. Thus, tools expressing the metabolic traits of a community are needed to broaden the utility of existing data. Gene abundance profiles are a relatively low-dimensional embedding of a metagenome’s functional potential and are, thus, tractable for comparison across many samples. Here, we compare the abundance of KEGG Ortholog Groups (KOs) from 6,539 metagenomes from the Joint Genome Institute’s Integrated Microbial Genomes and Metagenomes (JGI IMG/M) database. We find that samples cluster into terrestrial, aquatic, and anaerobic ecosystems with marker KOs reflecting adaptations to these environments. For instance, functional clusters were differentiated by the metabolism of antibiotics, photosynthesis, methanogenesis, and surprisingly GC content. Using this functional gene approach, we reveal the broad-scale patterns shaping microbial communities and demonstrate the utility of ortholog abundance profiles for representing a rapidly expanding body of metagenomic data. IMPORTANCE Metagenomics, or the sequencing of DNA from complex microbiomes, provides a view into the microbial composition of different environments. Metagenome databases were created to compile sequencing data across studies, but it remains challenging to compare and gain insight from these large data sets. Consequently, there is a need to develop accessible approaches to extract knowledge across metagenomes. The abundance of different orthologs (i.e., genes that perform a similar function across species) provides a simplified representation of a metagenome’s metabolic potential that can easily be compared with others. In this study, we cluster the ortholog abundance profiles of thousands of metagenomes from diverse environments and uncover the traits that distinguish them. This work provides a simple to use framework for functional comparison and advances our understanding of how the environment shapes microbial communities.",
"introduction": "INTRODUCTION Environmental metagenomics has uncovered the diversity of microbial life driving nutrient cycling in different ecosystems. An overarching goal of the field has been to connect environmental conditions to genomic characteristics to identify “who’s there?” and “what are they doing?”. As DNA sequencing technology has become more accessible, the amount of metagenomic data has ballooned. For example, the Joint Genome Institute’s Integrated Microbial Genomes and Metagenomes (JGI IMG/M) database contains more than 30,000 metagenomes ( 1 ) and the European Nucleotide Archive (ENA) contains upwards of 50,000 metagenomes ( 2 ). Large cross-study analyses of this data can increase our understanding of the patterns in microbiome function across environments. However, these efforts are resource intensive given the size and complexity of the data. To illustrate this, consider the hierarchical nature of a processed metagenome consisting of sequencing reads assembled into contigs which contain genes that are assigned a function, taxonomy, and relative abundance. Furthermore, contigs can be binned into metagenome-assembled genomes (MAGs) based on characteristics such as tetranucleotide frequency and read coverage ( 3 , 4 ). Genome-centric analysis of metagenomes can shed light on the metabolic potential of uncultivated microorganisms. For instance, MAGs have expanded our view of nitrogen fixation in the surface ocean ( 5 ), methane metabolism in archaea ( 6 – 8 ), and carbon degradation in permafrost ( 9 ). Multiple large collections of MAGs have been compiled using metagenomes from the human gut ( 10 ), anaerobic digestors ( 11 ), and a mix of different environments ( 12 ). While these collections provide a valuable resource for the exploration of the vast functional and taxonomic diversity of microbiomes, they do not reveal the functional composition of the whole community or enable facile comparison across samples. To this end, a gene-centric approach may be preferred. Early metagenomic studies, using limited sample sets, observed how ortholog abundance differed across environments according to the geochemical context ( 13 , 14 ). Studies of specific ecosystems have linked functional variation to parameters such as temperature, pH, and latitude within topsoil ( 15 ), permafrost soil ( 16 ), and ocean microbiomes ( 17 , 18 ). Together, these studies demonstrate how ortholog abundance is a valuable metric for representing a metagenome and linking it to its source environment. In addition, abundance profiles have a simple data structure that allows for large-scale comparisons and the placement of new data in the context of existing samples. In this study, we evaluate the characteristics of KEGG Ortholog Group (KO) abundance profiles across a large and diverse set of metagenomes from JGI IMG/M. We use unsupervised clustering to show how functional potential stratifies samples by environment and identify the marker KOs that differentiate sample groupings. Interestingly, we note differences in GC content between functional clusters and explore the KOs associated with this trait. The compiled data set can be explored and expanded upon with new metagenomes. Overall, this study provides insight into how environment shapes the functional composition of microbial communities and supports the use of ortholog abundance profiles for understanding metagenomic data.",
"discussion": "DISCUSSION Metagenomics is an increasingly valuable tool for interrogating the structure and function of microbiomes, nonetheless analysis of this complex data across large sample sets remains challenging. Here, we focus on KO abundance profiles as a consistent and low-dimensional embedding of functional potential to enable comparison across a large set of metagenomes. This dimension reduction results in our entire data set occupying a few hundred megabytes of storage, while the associated raw sequencing data likely occupies hundreds of terabytes. Some of the earliest metagenomics studies made similar observations of ecosystem-specific clustering of function using only a few samples ( 13 , 14 ). We extend these findings to a much larger and more diverse data set and analyze cluster-specific marker KOs. Splitting the markers into their associated KEGG pathway categories highlighted which metabolic functions defined each cluster, and by extension their respective environments. Broadly, we found that the availability of oxygen for respiration and light for photosynthesis were strong drivers of differentiation (aquatic vs terrestrial vs sludge). Furthermore, KOs related to domain-specific cellular functioning were enriched in different clusters (i.e., archaea in Cluster 3, eukaryotes in Cluster 2). Finally, the top cluster-specific marker KOs pointed at several interesting adaptations to different environments. Cluster 1 was dominated by soils and characterized by marker genes related to the degradation of a diverse set of aromatic compounds as well as antibiotic synthesis and resistance. There were marker genes associated with the breakdown of vanillate, syringate, catechol, and anthranilate which are all key intermediates in the bacterial degradation of lignin ( 28 ). Lignin is an important component of plant cell walls that accounts for approximately 30% of global organic carbon stocks ( 29 ), as such it represents a key carbon source in soils ( 28 , 30 ). Interestingly, the prm propane monooxygenase was one of the most differentially abundant genes in Cluster 1. In addition to propane oxidation ( 31 ), studies have associated this enzyme with N -nitrosodimethylamine degradation ( 32 ) and 1,4-dioxane degradation ( 33 ), suggesting that it may participate in the breakdown of other recalcitrant compounds in the environment. The apparent importance of this enzyme complex in soils warrants a more detailed exploration of its diversity and function. Previous reports have identified soils as rich reservoirs of both antibiotics ( 13 , 34 ) and antibiotic resistance ( 35 , 36 ), and our results reinforce that this as a defining functional characteristic of soils compared to other ecosystems. Lastly, sub-clustering suggested soils could be grouped based on fungal vs bacterial abundance, possibly driven by pH, which has been shown to be one of the strongest drivers of microbial community assembly in soils ( 15 , 37 ). Cluster 2 contained almost exclusively freshwater and marine samples and was clearly distinguished by genes related to photosynthesis and associated pigments. This cluster had the highest proportion of eukaryotic contigs and had marker KOs associated with eukaryotic cell biology presumably from an abundance of algae. Aside from photosynthesis, dimethylsulfoniopropionate (DMSP) demethylase was one of the top marker genes. DMSP is an important osmolyte for phytoplankton that can account for as much as 10% of their fixed carbon and also acts as a key carbon source for bacteria ( 38 ). The identification of the DMSP demethylase as one of the top marker genes illustrates the extent to which phototrophs shape the food web in aquatic systems. Sub-clustering divided freshwater and marine metagenomes based on the metabolism of compatible solutes and transport of inorganic ions, highlighting the unique adaptations of microbial life to aquatic environments of differing ionic strength. Furthermore, these observations illustrate how the clustering method can reveal finer scale differences in functional potential when the broader environmental context is held constant. Cluster 3 consisted of a variety of anaerobic environments from aquatic sediments to anaerobic digestors and gut microbiomes. This points to the conservation of anaerobic metabolisms (i.e., methanogenesis, sulfur reduction, and fermentation) in spite of the broader environmental context. Indeed, an evolutionary study of the CODH/ACS complex, a key enzyme in the Wood-Ljungdahl pathway of carbon fixation and the top marker KO for Cluster 3, revealed an astounding level of conservation across many anaerobic bacterial and archaeal lineages ( 39 ). Cluster 3 harbored numerous marker genes specific to archaeal cell functioning and had the highest proportion of archaeal contigs, highlighting the unique dominance and adaptation of this domain of life in anaerobic settings. We also found many CRISPR-associated genes as markers for Cluster 3, which is in alignment with a study of isolate genomes that found CRISPRs were more prevalent in anaerobes compared to aerobes ( 40 ). This suggests that defense against phage predation is especially important in anaerobic settings in contrast to Cluster 1 where genes for antibiotic production and resistance were uniquely abundant. Cluster 3 had one sub-cluster associated with digestive system samples and spore formation. These environments may experience fluctuating conditions that require sporulation for the survival of anaerobes compared to the anaerobic sediments, found in the other sub-cluster, that have more steady environmental conditions. We observed significant differences in GC content across functional clusters, which corroborates previous reports of metagenome GC content partitioning across environments ( 41 , 42 ). This phenomenon has been discussed from a variety of viewpoints, such as differences in carbon to nitrogen ratios, oxygen requirements, genome size, and DNA repair pathway ( 43 ). Using our functional gene approach, we found that NHEJ-related genes were correlated with high GC content while polV and a cytosine methyltransferase were correlated with low GC content. The role that NHEJ and polV play in shaping GC content has been explored in isolate genomes ( 43 , 44 ). In addition, methylated cytosine bases are mutational hotspots ( 45 ) that have been suggested to play a role in the evolution of GC content in vertebrate genomes ( 46 ). Our results suggest that differences in DNA repair and methylation play a role in shaping GC content even at the whole community level. This reinforces the notion that environmental conditions shape community functional potential, which, in turn, may contribute to the evolution of distinct genomic traits, such as GC content. While this approach provides an insightful glimpse into the functional potential of microbial communities across ecosystems, it is still limited by our ability to comprehensively assemble and annotate short metagenomic sequencing reads. The sequence assembly difficulties are particularly acute in soil samples due to their immense diversity. Furthermore, we are only able to assign KOs to roughly 40% of genes in this data set. A key question becomes, is the composition of microbial dark matter ( 47 ), what we cannot assemble or annotate, vastly different from what we can assemble and annotate? Given the functional redundancy of microbial communities, it is possible that the elucidation of the “dark matter” will not significantly change our overall view of community function. Lastly, it is worth noting that metagenomes only tell us about functional potential and should not be misconstrued as proxies of actual activity, for which proteomics, transcriptomics, or stable-isotope probing would be more appropriate. Overall, this study demonstrates the value of ortholog abundance profiles for representing microbial communities and supports their use in modeling community function. Ortholog abundance-based models have shown promise, for example, in explaining geochemical patterns of nutrient cycling in the ocean ( 17 , 48 , 49 ) and in predicting metabolic preference for sugars vs amino and organic acids in a collection of isolates ( 50 ). This approach takes advantage of the functional redundancy of microbial communities, that is the fact that functional traits are more stable than taxonomy ( 51 , 52 ). Focusing on function simplifies the comparison across microbial communities as evidenced by this study, which includes thousands of metagenomes. The data set provides a valuable resource for the exploration of the functional gene landscape across a range of ecosystems. It is amendable to expansion with additional samples and metadata describing key parameters like pH, temperature, and oxygen availability. Integrating ortholog abundance profiles with quality metadata will bring us closer to a predictive understanding of how microbial communities assemble and function in the face of varying environmental conditions."
} | 3,674 |
35539772 | PMC9082551 | pmc | 1,226 | {
"abstract": "Sediment microbial fuel cells (SMFCs) is a promising technology for bioremediation, environmental monitoring and remote power supply in various water environments. Optimizing the anode/cathode surface area ratio (SAR a/c ) is important to enhance the power and decrease the cost of SMFCs. However, in fact, little information has been reported to optimize the SAR a/c of SMFCs in individual or stacked mode. This study comparatively analyzed the effects of electrode surface areas on the performance of single SMFCs and serial SMFC-stacks under separated- and connected-hydraulic conditions. The results suggested an optimal SAR a/c of 1 to 1.33 for both single and serial stacked SMFCs. Voltage reversal occurred in serial SMFC stacks with unoptimal SAR a/c but not in optimized stacks. The more the SAR a/c deviated from the optimal SAR a/c , the more easily the voltage reversal occurred ( i.e. lower reversal current). Compared to a separated-hydraulic environment, a connected-hydraulic environment showed no effect on the power generation of anode-limiting SMFC stacks but decreased the power generation and reversal current of cathode-limiting SMFCs, probably due to larger parasitic current. The results are important for the scale-up and application of SMFCs.",
"conclusion": "4. Conclusions Optimization of electrode surface area is essential in scale-up and application of SMFCs. Our results suggested an optimal SAR a/c between 1 to 1.33 for both single and serial-stacked SMFCs, which indicated a small difference in the reaction rates between SMFC anodes and cathodes. Voltage reversal occurred in serial SMFC stacks with unoptimal SAR a/c but not in the optimized stacks, and the more the SAR a/c deviated from the optimal value, the more easily the voltage reversal occurred. Compared to a separated-hydraulic environment, a connected-hydraulic environment showed no effect on the power generation of anode-limiting SMFC stacks but decreased the power generation and reversal current of cathode-limiting SMFCs.",
"introduction": "1. Introduction Plenty of organic matters and contaminants accumulate in sediments due to various hydrobiological metabolisms and geochemical processes in water environments. An energy density of 12.2–67.1 kJ L −1 can be generated if the organic matters in sediments (0.2–2.2%) are oxidized completely. 1 Therefore, sediment is not only a contaminant resource but also an untapped energy reserve on the earth. Sediment microbial fuel cells (SMFCs) can convert chemical energy stored in sediment organic matters into electricity via microbial extracellular electron transfer, and have been used as promising tools in bioremediation, environment monitoring and remote power supply in laboratory or practical water environments. 1–4 Compared with other types of microbial fuel cells (MFCs), SMFCs generally generate much lower power densities. 5 Increasing the power output is one of the key challenges before the wide application of SMFCs. Slow chemical diffusion in sediments is considered to be main reason for the low power generation of SMFCs. 6 Therefore, it was suggested that the surface area of the anode in sediments should be larger than that of the cathode. Several studies used multiple-anodes to increase anode surface areas and balance the low diffusion efficiency in sediments. 7–9 However, in fact, little information of the effect of anode to cathode surface area ratio (SAR a/c ) on SMFC power generation has been reported. Since the power output of SMFCs does not always increase with the electrode surface area, 10 optimization of SAR a/c will be essential for assembling cost-efficient SMFCs. It has been reported that SMFCs stacked in series or parallel could efficiently increase the power output compared to a single SMFC. 11,12 Voltage reversal, caused by imbalanced reaction rates between connected SMFCs units, is a critical problem in serial MFC stacks. Recent studies showed that voltage reversal could not only decrease power output but also cause unrecoverable damages on biofilms or even electrode materials. 13 To avoid or postpone voltage reversal, the electrode surface areas of an SMFC in a serial stack should also be optimized to match the electrodes of the neighbour SMFC, which could be different to the electrode optimization of a single SMFC. Moreover, most reported SMFCs or SMFC-stacks were operated in separated reactors. However, multiple SMFCs applied in practical environments will share the same water environment which and perform differently with the separated SMFCs due to possible ion cross-transfer. 12 Therefore, the effects of electrode surface area on SMFC-stacks should be examined under both shared- and separate-hydraulic conditions. To test the effects of electrode surface area on the performances of single and serially-stacked SMFCs, this study assembled SMFCs with both anode and cathode surface areas varying from 146 to 646 cm 2 . The performances of the SMFCs were analysed in single or serially stacked operation mode. Moreover, the effects of shared-catholyte on voltage reversal were also analysed compared to that under separated-catholyte condition. The results suggested an equal or slightly larger anode surface area relative to cathode for either single or serially-stacked SMFCs.",
"discussion": "3. Results and discussion 3.1 Effects of SAR a/c on the power generation of single SMFCs To evaluate the effects of electrode surface area, other possible differences such as microbial communities, substrate compositions and concentrations among different SMFCs must be avoid. Therefore, all SMFCs were initiated with the same electrode material and surface areas under the same operation condition. After a 15 days operation, all SMFCs showed similar power generation capabilities (0.35 ± 0.01 V) ( Fig. 2A ). The voltages of the SMFCs differed largely upon the cut of the electrodes. As expected, SMFC voltages deceased with the electrode surface area. According to polarization curve of each SMFC, the MP of SMFCs showed consistent trend ( Fig. 2B ) with their voltage curves ( Fig. 2A ). SMFC 646 generated an MP of 0.28 mW and a maximum power density of 4.3 mW m −2 (normalized to cathode surface area), which was comparable to several SMFCs that operated in lab or in practical fields. 15–18 The MP of the anode-limiting SMFCs increased from 0.04 to 0.25 mW when the anode increased from 166 to 486 cm 2 ( Fig. 2B ). For the cathode-limiting SMFCs, the MP increased from 0.083 to 0.28 mW when the cathode increased from 166 to 486 cm 2 . By comparing the MP of SMFC c486 and SMFC 646 , an increase of cathode surface area from 486 to 646 cm 2 did not increase MP, suggesting an optimal SAR a/c of 1.33 for a single SMFC, i.e. the surface area of anode should be larger than that of the cathode. Fig. 2 also showed that the cathode-limiting SMFC always generated higher power than the anode-limiting SMFCs with the same electrode surface area, e.g. the MP of SMFC c166 was 0.09 mW while that of SMFC a166 was 0.05 mW, which also suggested higher electron transfer capability of cathode than the anode. All SMFCs showed no significant difference in MP when we switch the separated-hydraulic environment to shared-hydraulic environment. Fig. 2 Power generation of SMFCs with different electrode areas. (A) Voltage generation before and after cutting down electrode; (B) MP of different SMFCs, the red text indicate the SAR a/c of different SMFCs. Our results supported the suggestion that the anode electron transfer rate was lower than that of cathode in SMFCs. 6 However, the surface area difference between anode and cathode could be smaller than general consideration. Electrode generally accounts for the main capital cost of SMFCs as membrane separator was not needed in SMFCs. 19 Although the optimal SAR a/c may vary according to the electrode materials and environments, optimization of the electrode surface area is always needed to increase the cost-effectiveness of SMFCs, especially for the scale-up and field-application of SMFCs. 3.2 Effects of anode surface area on the power outputs of serial SMFC stacks To analyse the effect of anode surface area on the serial stacks of SMFCs, we stacked SMFC 646 with the anode-limiting SMFCs in which the surface areas of the limiting anode increased from 166 to 486 cm 2 . Under separated-hydraulic condition, the MP of those anode-limiting SMFCs stacks increased from 0.2 to 0.47 mW when the anode surface area increased from 166 to 486 cm 2 ( Fig. 3A ), indicating that anodes were still the key limit in the power generation of those SMFC stacks. The MP of all the anode-limiting stacks were lower than the sum of the MP of the SMFCs before stack. For example, the MP of SMFC 646 –SMFC a166 stack was expected to be 0.32 mW as the according to the MP of the two SMFCs before stack (0.04 and 0.28 mW). However, the MP of SMFC 646 –SMFC a166 stack was 0.2 mW, i.e. 37.5% of the power lost after stack. Voltage reversal was observed in all anode-limiting stacks which could explain the power loss ( Fig. 3B ). In contrast, SMFC 646 –SMFC 646 showed no voltage reversal and thus no power loss after stack. Electrode polarization showed more rapid potential increase of the limiting anodes relative to the cathode potentials, suggesting that anodes with smaller surface areas were the main reason for the voltage reversal in those SMFC stacks. Since all SMFCs were operated under the same condition, differences in the electrode surface area was supposed to be the only reason of the reversal. To rule out the possibility that lacking in electron donor caused reversal, we injected thiosulfate and acetate, two common electron donors in sediments, into the sediments which did not affect or even improved the voltage reversal. Polarization curves showed that compared to SMFC 646 –SMFC 646 , cell potentials fell more rapidly in the anode-limiting stacks, especially in the high-current area ( Fig. 3A ), suggesting higher potential loss in anode-limiting stacks caused by higher activation resistance, ohmic resistance and especially diffusion resistance. 18 Fig. 3B further showed that the diffusion limitation mainly occurred on the limiting anodes, which was consistence with the low chemical diffusion rate in sediments. The reversal current increased from 0.33 to 0.85, 0.97 mA when the surface area of the limiting anode increased from 166 to 326, 486 cm 2 , respectively. The results suggested that the performance of a SMFC stack was determined by the limiting-electrode, and improved limiting-electrode performance could postpone or eliminate the voltage reversal. Fig. 3 Polarization and power curves of the anode limiting stacks. (A) Polarization and power curves of the anode limiting stacks in separated-hydraulic environments; (B) electrode potential variation during polarization in separated-hydraulic environments; (C) polarization and power curves of the anode limiting stacks in connected-hydraulic environments; (D) electrode potential variation during polarization in connected-hydraulic environments. Arrows indicate reversal current. In addition to voltage reversal, ion-conduction is another possible reason that may cause voltage or power loss of MFCs in connected-hydraulic environments. When we switched the separated-hydraulic environment to shared-environment, the open circuit potential (OCP) of each stack decreased by 0.1–0.2 V, indicating ion-conduction occurred between the stacked SMFC units. However, the MP and voltage reversal current showed no significant difference before and after switching to shared-environment ( Fig. 3C and D ). Previous reports showed consistent results that ion-conduction could cause OCP loss in MFC stacks. 12,20,21 However, inconsistent effects of ion-conduction on power generation were reported. In line with our result, Dekker et al. reported that shared-electrolyte had no effect on the power of serial MFC-stack. 22 In contrast, Zhuang and Zhou reported that the ion-conduction could decrease power generation. 20 In fact, many factors other than ion-conduction can affect the power generation of MFC stacks, e.g. internal resistance, voltage reversal. Therefore, the occurrence of ion-conduction in serial SMFCs does not ensure a power loss when the other factors facilitate power generation, which may explain the inconsistent effects of ion-conduction on power generation reported in different literatures. 3.3 Effects of cathode surface area on the power outputs of serial SMFC stacks Similar to the anode-limiting stacks, the MP of cathode-limiting stacks increased with the cathode surface area ( Fig. 4A ), indicating cathode as the limiting factor in those stacks. Under separated-hydraulic condition, each cathode-limiting stack showed higher MP than that of the anode-limiting stack with the same electrode surface area, which was consistent with the higher MP of cathode-limiting SMFCs before stack. Voltage reversals were observed in the stacks of SMFC 646 –SMFC c166 and SMFC 646 –SMFC c326 , but not in SMFC 646 –SMFC c486 and SMFC 646 –SMFC c646 ( Fig. 4B ), indicating the optimal SAR a/c of single SMFCs could also postpone voltage reversal when they were serially stacked. Electrode polarization curves showed that the rapid cathode potential decease was the main reason for the reversal and the reversal current of SMFC 646 –SMFC c166 was smaller (0.57 mA) than that of SMFC 646 –SMFC c326 (0.8 mA). Both anode- and cathode-limiting stacks showed consistent trend that the limiting electrode was always the bottle neck in SMFC stacks and the worse the limiting electrode performed, the earlier the reversal occurred ( i.e. the reversal current decreased with the surface area of the limiting electrode). In contrast to the anodic diffusion-dominated potential loss in anode-limiting stacks, cathodic ohmic resistance play a more important role in the potential loss of cathode-limiting stacks ( Fig. 4B and D ). Fig. 4 Polarization and power curves of the cathode-limiting stacks. (A), Polarization and power curves in separated-hydraulic environments; (B), electrode potential variation during polarization in separated-hydraulic environments; (C), polarization and power curves in connected-hydraulic environments; (D), electrode potential variation during polarization in connected-hydraulic environments. When we switched separated cathode environments to shared environment, the OCP decreased by 0.3–0.4 V, indicated more OCP loss caused by ion-conduction in those cathode-limiting stacks relative to anode-limiting stacks (0.1–0.2 V). The MP of all cathode-limiting stacks showed significant decrease ( Fig. 4C ), different to the minor MP loss for in anode-limiting stacks ( Fig. 3C ). The cathode potential curves decreased while anode potential curves increased after switch to shared environment ( Fig. 4D ), indicating potential loss on both anode and cathode. Moreover, cathode potentials of SMFC c166 and SMFC c326 showed more rapidly decreased with current increase and showed voltage reversal at 0.29 and 0.65 mA respectively, which were lower than those (0.6 and 0.79 mA) under separated environments. SMFCs will be deployed in hydraulic-connected environments in practical application. Power loss caused by ion-conduction must be considered if multiple SMFCs are deployed. Enlarging the distance between SMFC units has been suggested to decrease the ion-conduction. 20,21 However, the internal resistance and wire length of a SMFC stack may also increase with the distance. It is still unclear why the shared-environment decreased the power MP and reversal current of the cathode-limiting stacks but not of the anode-limiting stacks. One possible explanation is that ( Fig. 5 ): in cathode-limiting stacks, the electron and proton generations on anodes were more rapid than the consumption by cathode due to the insufficient cathode surface area, resulting excess electrons and protons that contribute to parasitic cells. 20,21 In anode-limiting stacks, the electron and proton generation of the anode was slower than the consumption by cathode, and thus less electron and proton can be used in forming parasitic cells. As a result, the cathode-limiting stacks can generate more power than the anode-limiting stacks with the same electrode surface area under separated environments because of higher reaction rate on cathode than anode. On the other hand, cathode-limiting stacks lost more power due to higher parasitic current loss under shared-hydraulic environments. Fig. 5 Different electron/proton losses caused by anode-limiting (A) and cathode-limiting SMFC stacks (B) in hydraulic-connected environments. Red electrodes indicating the limiting-electrodes. Solid black lines indicate wires that transfer electrons for power generation; dashed blue arrows indicated electrons lost as parasitic current. Arrow width indicates the amount of electron loss from anode."
} | 4,239 |
37063728 | PMC10103077 | pmc | 1,228 | {
"abstract": "The corrosion of materials severely limits the application scenarios of triboelectric nanogenerators (TENGs), especially in laboratories, chemical plants and other fields where leakage of chemically corrosive solutions is common. Here, we demonstrate a chemical-resistant triboelectric nanogenerator (CR-TENG) based on polysulfonamide (PSA) and polytetrafluoroethylene (PTFE) non-woven fabrics. The CR-TENG can stably harvest biological motion energy and perform intelligent safety protection monitoring in a strong corrosive environment. After treatment with strong acid and alkali solution for 7 days, the fabric morphology, diameter, tensile properties and output of CR-TENG are not affected, showing high reliability. CR-TENG integrated into protective equipment can detect the working status of protective equipment in real time, monitor whether it is damaged, and provide protection for wearers working in high-risk situations. In addition, the nonwoven-based CR-TENG has better wearing comfort and is promising for self-powered sensing in harsh environments.",
"conclusion": "4. Conclusion In conclusion, a chemical-resistant TENG was developed using PSA non-woven fabrics and S-PVA–PTFE membranes. The stability of the device in harsh environment and the reliability of output under different degrees of corrosion were verified. The designed TENG has three functions: acid and alkali corrosion resistance, dynamic monitoring of human movement and automatic safety protection monitoring. The protective glove based on CR-TENG can monitor the integrity of itself in real time, ensuring the personal safety of the staff. The developed TENG can be further laminated on the easily bendable and damaged parts of the chemical protective clothing, and response to joint movement and safety. This chemical-resistant TENG can realize self-powered intelligent sensing while meeting the requirements of wearing comfort, and has broad application prospects in harsh high-risk corrosion environment.",
"introduction": "1. Introduction Human industrial activity drives the modernization process, but it is also prone to accidents in high-risk work environments such as chemical plants, 1 laboratories 2 and special workshops. 3 Of these, leakage or splashing of corrosive solutions such as acid and alkali is the most common event. 1,4 Protective clothing can effectively protect workers in high acid-alkali-hazardous environments, but the common coated protective clothing on the market has problems such as poor air permeability and poor wearing comfort. 5,6 Moreover, when protective clothing is damaged, workers are often unable to perceive it in time, which poses a serious hidden danger to the safety and health of operators. 5,7,8 The wearable triboelectric nanogenerators (TENGs) in self-powered mode can be used for real-time sensing detection, and respond in time when insecurity occurs to prevent danger. 9–13 However, most of the current materials and electrodes of TENGs cannot resist the corrosion of chemical solutions, which may severely limit the work and application scenarios of TENGs. 14–18 At present, there are few reports about corrosion-resistant TENG. Among them, platinum film 19 and copper–nickel alloy conductive tape 20 have poor air permeability and comfort, so it is not easy to wear or integrate into protective equipment. The polytetrafluoroethylene (PTFE) composite yarn developed by Ma et al. solved the problem of electrostatic shedding and successfully applied PTFE material to protective textiles. 1 However, the fabric obtained by the warp and weft weaving method has large pores, and there is a risk of penetration of corrosive chemicals. Therefore, the corrosion-resistant TENG, which is comfortable to wear, has good safety performance and can be integrated into protective equipment for hazard warning and risk reduction, and still needs further researches. As a high-performance fabric, polysulfonamide (PSA) is wear-resistant, durable, 21 flame-retardant, 22–24 excellent in waterproof performance, 25 and has good chemical stability to resist corrosion. 26–28 Therefore, this material is suitable for protective equipment such as protective clothing and protective gloves in special environments. 21 Here, we prepared PSA nonwovens by using needle-free wire electrode electrospinning equipment. PSA materials obtained by electrostatic spinning technology not only retain its own advantages of hydrophobic, corrosion resistant and chemical stability, but also wear comfort, have good air permeability, which had the potential to produce or integrate into protective equipment. In the preparation process, needle-free wire electrode electrospinning equipment was used, that can not only provides sufficient high voltage to ensure the stability of the spinning process, but also overcomes the problem of low efficiency of traditional wet spinning, 29 and can efficiently produce materials for industrial and commercial applications. In addition, pure PSA spinning solution has the difficulty of poor spinnability, and we solved this problem by using a small amount of thermoplastic polyurethane (TPU) as the auxiliary polymer of the solution. In this work, we combined PSA fabrics with nanogenerators and developed a chemically stable frictional electric nanogenerator (CR-TENG). The device had a short-circuit current of 250 nA and an open-circuit voltage of 26 V. After testing, TENG's material can resist corrosion of strong acid and strong alkali solution for up to 7 days. Meanwhile, the TENG output remained stable after various degrees of corrosion (from pH = 3 to pH = 11) and under different temperature working environment. Moreover, CR-TENG can monitor human motion status while working stably in harsh environments, and can be integrated into protective equipment for self-powered safety monitoring. This TENG overcomes the impact of environmental corrosion on the material and enables self-powered sensing applications in locations with harsh operating environments, risk and hidden dangers such as security monitoring.",
"discussion": "3. Result and discussion 3.1 CR-TENG In this work, we developed a chemical-resistant CR-TENG. This single-electrode working mode TENG is composed of PSA, sintered PVA–PTFE (S-PVA–PTFE) and carbon nanotube coating electrode with excellent chemical stability and electrical conductivity, and its structure is shown in Fig. 1b . Among them, PSA material (contact angle 131°) and S-PVA–PTFE material (contact angle 159°) have excellent waterproof performance. In addition, after sintering, the acid and alkali resistance and electrical performance of PVA–PTFE material are improved. 30 The carbon nanotube coating as the electrode not only has excellent flexibility, 31 but also ensures the stability of the electrical output. 32 As shown in Fig. 1a , TENG can be integrated into human joints ( e.g. elbows and knees) and gloves. The CR-TENG fixed on the elbow and knee joints, as shown in Fig. 1c , continuously contacts and separates with the movement of the joints, and the electrical signals are outputted to monitor human movement dynamically. In addition, the smart glove composed of CR-TENG can not only collect hand motion energy to detect hand movement status, but also automatically cause an alarm when the working state of the glove is in danger in real time, as shown in Fig. 1d . This self-powered protective glove with safety monitoring feature helps protect operators working in high-risk areas. To prepare the friction layer material, we used an electrospinning device with a needle-free wire electrode ( Fig. 1e ). The equipment can manufacture fiber membranes on a large scale, and can realize the industrial production and application of electrospinning fabrics. 3.2 Working mechanism of CR-TENG In the single-electrode TENG, the PSA loses electrons better than S-PVA–PTFE, so after contact, the surface of PSA has a positive potential, while the surface of S-PVA–PTFE loses electrons and has a negative potential. The working mechanism of CR-TENG is shown in Fig. 2a . Due to the contact electrification effect, when PSA fabric and S-PVA–PTFE material contact under the action of external force, as shown in Fig. 2a(i) , charge transfer occurs on the surfaces of the two non-woven fabrics, and equal and opposite charges are induced. When the external force is unloaded, as shown in Fig. 2a(ii) , the two contact surfaces are separated from each other. In order to balance the static electricity, the charge is directed from the carbon nanotube electrode to the ground. As shown in Fig. 2a(iii) , when the two materials are separated to the maximum extent, the charge will stop flowing. At this time, the resulting potential difference and the corresponding current reach the peak. When PSA approaches S-PVA–PTFE again, the charge moves from the ground to the carbon nanotubes, and the whole process is reversed (as shown in Fig. 2a(iv) ). The two fabrics repeat the process of contact and separation, continuously outputting alternating current. Fig. 2 Triboelectric performance of CR-TENG. (a) Schematic illustration of working mechanism of CR-TENG. Output of CR-TENG at different pressures (b), distances (c) and frequencies (d). (e) Durability and stability testing of CR-TENG. 3.3 Electrical output performance of CR-TENG To study the triboelectric properties, a CR-TENG composed of PSA (40 mm × 50 mm) material and S-PVA–PTFE (30 mm × 30 mm) material was designed to harvest mechanical energy in the environment. When the temperature was 24 °C and the humidity was 55%, the method of controlling variables was used to evaluate the adaptability of TENG to different environments. As shown in Fig. 2b , when the impact force changes, the greater the force, the more contact between the two friction layers, the current and the voltage increase accordingly. When the distance between the friction layers increases, as shown in Fig. 2c , the electrical output also rises. When the impact force and distance are constant, the current output of TENG increases with the increase of the impact frequency, as shown in Fig. 2d . This is mainly due to the shorter duration of the current peak at higher impulse frequencies, resulting in larger short-circuit currents. Accordingly, the open-circuit voltage improves as the frequency increases. The maximum current (250 nA) and voltage (26 V) were obtained when the impact frequency was 2 Hz, the impact force was 30 N and the distance between the two friction layers was 30 mm. Durability and long-term stability are important indexes to judge whether TENG can be practically applied. At constant frequency, pressure, and friction spacing, 4000 contact-separation cycles were tested on CR-TENG. During the cycling process, the current signal does not have obvious deletion and reduction, as shown in Fig. 2e , and the CR-TENG exhibits excellent persistence and output stability. 3.4 Performance of CR-TENG after acid and alkali-base treatments The hydrophobicity of PSA and S-PVA–PTFE was tested by dipping method. The two non-woven fabrics were immersed in methyl orange solution for a certain period of time and then taken out ( Fig. 3a ). The surfaces of the two materials were not dyed and discolored, and no liquid remained on the surface. Both friction materials have excellent water resistance. In order to test the anti-corrosion performance of CR-TENG, the possibility of its application in severe environments, especially in strong acid and alkali conditions, was confirmed. First, after immersing the PSA composite membrane and the S-PVA–PTFE fiber membrane in strong acid and strong alkali solutions for 10 min, it was found that the surfaces of the materials were not damaged ( Fig. 3b ). To study the chemical durability of PSA and S-PVA–PTFE. We treated the two materials in 36% concentrated hydrochloric acid and 1 mol L −1 sodium hydroxide solution for 7 days. After taking out and drying, the microstructure was characterized by scanning electron microscope, and the change in the microstructure of the comparative materials were observed and the fabric diameter transformations were analyzed. Fig. 3 Macroscopic and microscopic comparison after acid and alkali treatment. (a) The process diagram of PSA composite membrane and S-PVA–PTFE composite membrane immersed in methyl orange solution. (b) Optical photos of PSA and S-PVA–PTFE before and after acid–base treatment. (c) SEM image of the PSA material untreated and treated in 36% hydrochloric acid and 1 mol L −1 sodium hydroxide solution for 7 days, and the fiber diameter distribution before and after treatment. (d) SEM image of S-PVA–PTFE material untreated and treated with concentrated hydrochloric acid and sodium hydroxide solution for 7 days, and the fiber diameter distribution of fibers before and after treatment. (e) and (f) are FTIR analysis of PSA and S-PVA–PTFE materials before and after corrosion, respectively. (g) Air permeability test of PSA materials before and after chemical treatment. (h) Softness test of the materials. The SEM images in Fig. 3c(i)–(iii) show that the PSA films immersed in two corrosive solutions for a long time did not suffer from dissolution and curling, and the fibers did not find any abnormality such as breakage. Fig. 3c(iv) shows the fiber diameter distribution before and after corrosion of PSA non-woven fabrics. The fiber diameters of PSA fabrics without chemical corrosion are mainly (51.28%) in the 125–160 nm range, while 56.28% and 54.45% of PSA fibers treated with strong acid and strong alkali are distributed between 125–160 nm, respectively. There was no reduction in the fiber diameter. Similarly, it can be found that the S-PVA–PTFE nonwovens soaked in 36% concentrated hydrochloric acid and 1 mol L −1 sodium hydroxide for 7 days did not dissolve or partially dissolve, nor did they break, bend, or pit, as shown in Fig. 3d(i)–(iii) . As shown in Fig. 3d (iv) , the particle size distribution shows that the diameter distribution range of S-PVA–PTFE fibers before and after corrosion is mainly concentrated between 210–300 nm (72.56% untreated, 75% after strong acid treatment, 69% after strong alkali treatment). There was no obvious change in fiber diameter before and after material treatment, showing excellent corrosion resistance. FTIR spectra were tested and analyzed to further support the chemical resistance of the materials. As shown in Fig. 3e , PSA materials before and after corrosion all have stretching vibration absorption peaks of N–H bond in amide at 3337 cm −1 , which is one of the characteristic peaks of PSA. And they all have the stretching vibration peak of –C \n \n\n<svg xmlns=\"http://www.w3.org/2000/svg\" version=\"1.0\" width=\"13.200000pt\" height=\"16.000000pt\" viewBox=\"0 0 13.200000 16.000000\" preserveAspectRatio=\"xMidYMid meet\"><metadata>\nCreated by potrace 1.16, written by Peter Selinger 2001-2019\n</metadata><g transform=\"translate(1.000000,15.000000) scale(0.017500,-0.017500)\" fill=\"currentColor\" stroke=\"none\"><path d=\"M0 440 l0 -40 320 0 320 0 0 40 0 40 -320 0 -320 0 0 -40z M0 280 l0 -40 320 0 320 0 0 40 0 40 -320 0 -320 0 0 -40z\"/></g></svg>\n\n C– of benzene ring at 1530 cm −1 and the symmetric stretch peak of sulfone group (–SO 2 –) at 1150 cm −1 . 33 For S-PVA–PTFE, Fig. 3f shows the characteristic absorption peaks of C–F at 617 cm −1 and C–O at 1125 cm −1 existed before and after chemical treatment. 34 FTIR analysis proved that neither of the two materials showed shrinkage or deviation of characteristic peak value after chemical corrosion. In addition, air permeability affects the wearing comfort of materials. We tested the permeability of PSA materials before and after corrosion. As shown in Fig. 3g , the PSA material can achieve air permeability of 18.23 mm s −1 . At the same time, the air permeability of PSA material after strong acid and base treatment remained stable, and did not decrease. Through the softness test, it was found that both materials showed good flexibility, as shown in the Fig. 3h . Good air permeability and flexibility can ensure the wearing comfort. Through the above observation and research, both materials were found to have excellent chemical stability and corrosion resistance. Based on –SO 2 – groups, PSA materials have excellent chemical stability. During the sintering process of the PVA–PTFE film, the PTFE dissolves and completely wraps the surface of the PVA fiber. 30 The S-PVA–PTFE composite film has excellent acid and alkali resistance and corrosion resistance. The chemical structures of PSA and S-PVA–PTFE materials are shown in Fig. 4a . Fig. 4b shows that the tensile strain of PSA changed slightly after being etched by strong acid solution (42% to 45%); the tensile strain of the PSA nonwoven treated with strong alkali solution is almost unchanged, and the stress decreases slightly. The tensile stress–strain of the sintered PVA–PTFE material changed slightly after being treated with two corrosive solutions, as shown in Fig. 4c . It can be seen that the two fabric materials after severe corrosion still maintain a certain stability, tensile ability and deformation ability. Fig. 4 Performance testing in harsh environment. (a) Chemical structures of PSA and S-PVA–PTFE. (b) Tensile stress–strain curves of the PSA film before and after corrosion. (c) Tensile stress–strain curves of S-PVA–PTFE membrane before and after corrosion. (d) Schematic diagram of the treatment of the film in an acid–base environment. (e) Short current and (f) voltage of CR-TENG treated with solutions of different pH values. (g) and (h) are 4000 cycles of current of CR-TENG after strong acid and strong alkali treatment respectively. (i) Output current of the TENG at different temperatures. In order to study the resistance of CR-TENG to different degrees of corrosion, PSA and sintered PVA–PTFE were treated with solutions of different pH values (pH = 3, pH = 5, pH = 7, pH = 9, and pH = 11), as shown in Fig. 4d . When controlling variables, the test compared the changes of current and voltage before and after corrosion. Fig. 4e shows that the current output of TENG after corrosion in different pH solutions is almost unchanged, and only slightly decreases at pH = 3 and pH = 11. The voltage of TENG did not change significantly under different degrees of corrosion, as shown in Fig. 4f . After being treated with highly acidic (pH = 3, as shown in Fig. 4g ) and alkaline (pH = 11, as shown in Fig. 4h ) solutions, the output of CR-TENG is stable during the 4000 operating cycles with no significant loss of signal. The electrical performance of CR-TENG remains stable under different degrees of corrosion, and TENG has strong environmental adaptability and stability. In addition, the CR-TENG was subjected to different temperature operating environments in order to study the sensitivity of its output to temperature. As shown in Fig. 4i , under operating conditions of 25 °C, 35 °C, 45 °C and high temperature 55 °C, the current output of the developed TENG remains stable and does not rise or fall. The temperature adaptability of the TENG provides conditions for its steady working under high temperature environment. 3.5 CR-TENG for biological motion monitoring An arched CR-TENG was designed as shown in Fig. 5a(i) , and integrated it into a glove to develop a smart glove. As shown in Fig. 5a , when the finger touches the CR-TENG on the palm, a negative voltage signal is generated; when the finger is removed, a positive output is generated; when the finger is stationary, no signal is generated. CR-TENG has the function of reflecting its own working state, as shown in Fig. 5b . When touched, an intact CR-TENG that is not damaged generated obvious electrical signals; however, after the PSA friction layer was destroyed, the TENG output signal decreased. In addition, nitrile gloves fixed with CR-TENG realized the intelligent self-powered detection of hand movement. This may be because the damaged friction layer hinders the transfer of electrons. We designed different gestures and described these gestures with different numbers, as shown in Fig. 5c . Fig. 5d shows the real-time voltage signals collected by bending different fingers sequentially. When hand gestures of “4-0-1-2-3-4” were carried out in sequence as shown in the above picture, finger contact caused contact and separation of the corresponding CR-TENG, and corresponding electrical signals were generated to record finger movements trajectory. TENG exhibits excellent gesture detection and recognition capabilities. The CR-TENG also showed a sensitive response to major joint movements, such as at the elbows and knees. As shown in Fig. 5e , as the elbow is straightened and bent at a certain rate, the PSA and S-PVA–PTFE continuously friction with each other to produce a stable and repeatable voltage output. In addition, the CR-TENG can also be fixed on the wearer's knee to further record the activities of the knee, as shown in Fig. 5f . The developed CR-TENG which can monitor its own working state has outstanding biological motion sensing ability and stable signal output. CR-TENG realizes accurate detection of human joint motion by identifying sensing signals, and it can be further superimposed on the flexible and easily bendable parts of the chemical protective equipment to track and detect the movement trajectory of workers in high-risk environment. Fig. 5 CR-TENG detects human motion. (a) Physical image of the CR-TENG at the joint and its photograph for monitoring hand motion. (b) Output voltage of intact and damaged CR-TENG. (c) Different gestures and their corresponding numbers. (d) The output signal in sequence when the finger moves “4-0-1-2-3-4”. (e) Elbow movement and output during movement. (f) Knee activity and output during activity. 3.6 Intelligent security protection monitoring In order to enable the corrosion-resistant TENG with self-powered sensing mode for further applications in high-risk environment. We designed a smart protective glove, the outer layer of the glove is PSA non-woven fabric and the inner layer is S-PVA–PTFE material coated with CNTs, as shown in Fig. 6a . Wearing protective gloves with corrosion resistance and real-time motion monitoring functions, as shown in Fig. 6b , can effectively prevent the corrosion and injury of splashed or leaked acid–base liquids to the staff. Moreover, the continuous friction between the two materials realizes the dynamic monitoring of the movement of the worker's hand during the operation. Fig. 6 Application of CR-TENG in the intelligent chemical protective clothing system. (a) Structural drawings of the protective glove. (b) Application scenario diagram. (c) Voltage output before and after glove breakage. (d) Current output before and after glove breakage. The CR-TENG in the glove can also detect and sense the working state of the protective glove in real time through the change of the output. Through research and comparison, it can be found that, as shown in Fig. 6c , the output voltage of the intact glove is 37 V during use, and the output drops to 27 V when damaged. Likewise, the intact glove produced a current of 159 nA, while the broken glove current output showed a clear downward trend, as shown in Fig. 6d . When the electrical signal drops below a certain value, an alarm will be triggered to indicate that the gloves are damaged, and it should be replaced in time. Observing the change of output during the use of the gloves can help us clearly realize the working status of the gloves. This smart protective glove can detect the integrity of the chemical protective clothing in real time and remind users to change it in time to prevent chemical injuring to their hands. CR-TENG is expected to be further combined with protective gear such as protective clothing for intelligent safety monitoring in high-risk environments."
} | 5,981 |
24332540 | null | s2 | 1,229 | {
"abstract": "Bacteria frequently live in densely populated surface-bound communities, termed biofilms [1-4]. Biofilm-dwelling cells rely on secretion of extracellular substances to construct their communities and to capture nutrients from the environment [5]. Some secreted factors behave as cooperative public goods: they can be exploited by nonproducing cells [6-11]. The means by which public-good-producing bacteria avert exploitation in biofilm environments are largely unknown. Using experiments with Vibrio cholerae, which secretes extracellular enzymes to digest its primary food source, the solid polymer chitin, we show that the public goods dilemma may be solved by two very different mechanisms: cells can produce thick biofilms that confine the goods to producers, or fluid flow can remove soluble products of chitin digestion, denying access to nonproducers. Both processes are unified by limiting the distance over which enzyme-secreting cells provide benefits to neighbors, resulting in preferential benefit to nearby clonemates and allowing kin selection to favor public good production. Our results demonstrate new mechanisms by which the physical conditions of natural habitats can interact with bacterial physiology to promote the evolution of cooperation."
} | 315 |
27043584 | PMC4847037 | pmc | 1,230 | {
"abstract": "This study investigated the mechanism and key factors influencing concurrent phosphorus (P) recovery and energy generation in microbial fuel cells (MFC) during wastewater treatment. Using a mediator-less dual chamber microbial fuel cell operated for 120 days; P was shown to precipitate as struvite when ammonium and magnesium chloride solutions were added to the cathode chamber. Monitoring data for chemical oxygen demand (COD), pH, oxidation reduction potential (ORP) and aeration flow rate showed that a maximum 38% P recovery was achieved; and this corresponds to 1.5 g/L, pH > 8, −550 ± 10 mV and 50 mL/min respectively, for COD, pH cathode , ORP and cathode aeration flow rate. More importantly, COD and aeration flow rate were shown to be the key influencing factors for the P recovery and energy generation. Results further show that the maximum P recovery corresponds to 72 mW/m 2 power density. However, the energy generated at maximum P recovery was not the optimum; this shows that whilst P recovery and energy generation can be concurrently achieved in a microbial fuel cell, neither can be at the optimal value.",
"conclusion": "4. Conclusions A mediator-less dual chamber microbial fuel cell was used to investigate concurrent P recovery and energy generation. Due to the high pH around the cathode, the MFC forms precipitates containing phosphorus in the catholyte and on the cathode surface. The main component of the precipitate was determined by X-ray diffraction analysis to be magnesium ammonium phosphate hexahydrate. P recovery at the cathode ranged from 1% to 38%; and correlation analysis indicated that COD and aeration flow rate are the key factors influencing P recovery and energy generation. Further studies are recommended to minimize the energy loss due to the precipitates on the cathode surface. This could be focused on a movable cathode that can be replaced once the voltage starts to decrease, or finding a specific material that prevents any particle attachment on the cathode surface.",
"introduction": "1. Introduction Microbial fuel cells (MFCs) are systems that convert chemical energy in an organic substrate in wastewater into electrical energy. Phosphorus (P) is one of the key pollutants that can be removed during wastewater treatment in MFC with the concurrent generation of energy. P is essential for crop growth, and approximately 60% of P used around the world comes from non-renewable sources [ 1 ]. Therefore, there is increasing interest in finding new sustainable sources of P. Wastewater is a rich source of nutrients that can be used as a sustainable source of P. Magnesium ammonium phosphate hexahydrate (MAP or struvite) is an efficient slow release fertilizer [ 2 ]. The mechanism of struvite precipitation is highly dependent on solution pH (pH > 8). The precipitation occurs in equimolecular concentration of magnesium (Mg), ammonium (NH 4 ) and P; and these combine with water to form struvite [ 3 ]. P can be recovered as struvite from different waste streams including reject wastewater and digester effluent; and it can be achieved using several methods, including: chemical addition, carbon dioxide stripping or electrolysis [ 4 , 5 , 6 ]. However, the chemical addition is a costly process and the chemicals used to raise the pH can account for up to 97% of struvite cost [ 6 , 7 ]. Recent research findings on MFCs have demonstrated their potential to recover P as struvite [ 7 , 8 , 9 ]. P can be recovered exclusively through precipitation in MFC because P compounds are not involved in electron transfer via reduction—oxidation (REDOX) reactions [ 10 ]. P recovery as struvite using MFCs involves cathode reactions, whereby water is consumed and hydroxide is generated due to electrons transferred from the anode to the cathode (Equations (1) and (2)):\n (1) Anode chamber : CH 3 COO - + 4 H 2 O → 2 HC O 3 − + 9 H + + 8 e - \n (2) Cathode chamber : 2 H 2 O + O 2 + 4 e − → 4 OH − The generated hydroxide leads to increase in the pH around the cathode, and P starts to precipitate. P recovery from different wastewater sources has been widely studied using single chamber MFCs [ 5 , 8 , 9 ]. Ichihashi and Hirooka [ 9 ] reported 27% P recovery as struvite from swine wastewater. It was noted that precipitates formed around the cathode, which indicates that the pH around the cathode is higher than at any other place in the MFC. In addition, synthetic wastewater was used in a single chamber MFC to identify the effect of Mg and NH 4 concentration on P recovery as struvite [ 8 ]. Dosing ammonium and magnesium for struvite precipitation is essential to form struvite; as the concentrations of NH 4 and Mg increases, more P is precipitated. In another study, a 3-stage single chamber MFC was used to recover 78% P as struvite from human urine [ 4 ]. Using urine as a substrate increased the recovered P because urea hydrolysis increased the pH and created an alkaline environment which accelerated the precipitation. From these studies, it has been shown that MFC has the potential to recover P as struvite from different wastewater sources. However, most of these studies were conducted in single chamber MFCs, where pH buffering may occur due to the accumulation of protons and hydroxide ions in the same electrolyte; this could limit the P recovery. To overcome this limitation, dual chamber MFCs are increasingly being used. They have better separation between the anode and cathode chambers; this leads to pH splitting which creates an alkaline environment around the cathode to improve the P recovery process [ 10 ]. Such a 2-stage dual chamber MFC was operated in this study to recover P in the form of struvite. The specific objectives of the study were: (i) to understand the mechanisms of P recovery as struvite in dual-chamber MFC; and (ii) to identify the influencing factors for the P recovery.",
"discussion": "3. Results and Discussion 3.1. Phosphorus Recovery and Electricity Generation in Dual Chamber MFC After the start-up period, stable electricity production was achieved in the system. At the beginning of each batch (batch duration = 48 h), electricity production peaked with a current density of (324 mA/m 2 ) and decreased with time until the end of each batch. The maximum power density was achieved at COD = 0.7 g/L, and the system achieved a maximum power density of 198 mW/m 2 with a current density of 0.49 mA/m 2 . Due to the degradation of organic matter at the anode, electrons are released and transferred through the external resistance to the cathode. Current is generated during the transfer of electrons from anode to cathode. At the cathode, water is consumed and hydroxide is generated; this increases the pH around the cathode electrode (pH > 8). Precipitates around the cathode were observed after a short period of adding Mg and NH 4 solution to the cathode chamber. Cathode effluent was filtered at the end of each batch and the precipitations were analysed using XRD (see Figure 2 ). P was recovered in the dual chamber MFC as magnesium ammonium phosphate hexahydrate (struvite) at the cathode chamber, by dosing 8 mM of NH 4 Cl and MgCl 2 (0.76 g/L and 0.42 g/L, respectively) to achieve a molar ratio (Mg: NH 4 : P: 1:1:1). It was observed that the greater the precipitates at the cathode, the less energy generated by the MFC. This suggests that precipitates at the cathode obstruct the mass transfer of ions and oxygen [ 8 ]. Current density was negatively affected by P precipitates at the cathode chamber, where low current density (62 mA/m 2 ) was observed with high P precipitates ( Figure 3 ). During phase 1 (COD = 0.7 g/L), the maximum current density achieved was 370 mA/m 2 and average cathode pH was 7.4. Small amount of struvite was precipitated in the cathode chamber (2%–8% of P). In phase 2 (COD = 1 g/L), the current density decreased significantly as compared to phase 1, where average cathode pH increased to 7.8 and struvite precipitate increased from 8% to 26%. In phase 3 (COD = 1.5 g/L), the decrease in current density was significantly more as a result of the increased amount of precipitates on the cathode electrode ( Figure 4 ). Average cathode pH increased and peaked at 8.3. The maximum P recovery was achieved in phase 3 with a maximum recovery of 38%. These results implies that the greater the precipitates at the cathode, the less energy generated by the MFC. 3.2. COD Removal and Coulombic Efficiency The system was operated at three different COD concentrations (0.7, 1.0 and 1.5 g/L) to simulate COD concentrations of reject wastewater. The anode COD removal efficiency ranged from 70% to 90%. At low COD concentration, the average removal efficiency was 90% and at high COD concentration, the average removal efficiency was 70%. Increasing influent COD concentration decreased COD removal efficiency. The average coulombic efficiency was 10% and it decreased with increasing COD concentration. Similar findings were reported by Sleutels et al. [ 12 ], where it was noted that increasing substrate concentration leads to decreases in the coulombic efficiency. Furthermore, increasing cathode pH due to hydroxide generation leads to a decrease in electricity generation, and this then decreases the coulombic efficiency [ 13 ]. This result demonstrates that COD concentration is an important factor for obtaining high coulombic efficiency and energy generation in MFCs. The coulombic efficiency can be improved by using a better configuration of dual chamber MFC. The type used in the current study is the H-type which is known to be a low-efficiency system due to the limited surface area for ion exchange membranes, and the long distance between electrodes [ 14 ]. In addition, the diffusion of oxygen from the cathode and from the pH and ORP probes port decrease the coulombic efficiency showing that the oxygen could work as an electron acceptor in the anode chamber. 3.3. Effect of Different Operational Parameters The performance of MFCs can be influenced by different operational parameters such as: COD concentration, chamber volume and catholyte aeration flow rate. It is therefore very important to understand the impact of these parameters on MFC performance. The influence of COD concentration which acts as fuel for the bacteria, electrolyte volume and cathode air flow rate on MFCs electricity generation have been studied [ 13 , 15 , 16 , 17 ]. In general, the studies show that increasing COD concentration and cathode aeration leads to increased energy generation; however, the influence of these parameters on P recovery has not been investigated. 3.3.1. P recovery at Different Substrate Concentrations Different COD concentrations were used to identify the impacts of COD concentration on P recovery. The COD concentration at the anode chamber varied from 0.7 to 1.5 g/L (organic loading rate = 0.35 to 0.75 g COD/L/day). It was observed that increasing the anolyte COD leads to increase in the recovered P at the cathode ( Figure 5 ). As the COD concentration increased from 0.7 g/L to 1.5 g/L, the recovered P increased from 7% to 38%. Similarly, as influent COD concentration increased at the anode, anode oxidation reactions increased ( Figure 6 ). This implies that organic matter degradation increased due to substrate availability for the microorganism. As a result of the organic matter degradation, electrons are liberated and transferred from the anode to the cathode [ 11 ]. Increasing COD concentration leads to increase in the transfer of electrons from the anode to the cathode chamber. This implies that more electrons are available at the cathode for oxygen reduction reactions; this, in turn facilitates struvite formation. Increased oxygen reduction reactions at the cathode leads to increases in the generation of hydroxide ions; and this increased the pH to >8 at the cathode ( Figure 6 and Figure 7 ). It was noted that cathode pH increased as COD concentration increased ( Figure 7 ). The average cathode pH increased from 7.4 at COD = 0.7 g/L to 8.3 at COD = 1.5 g/L. Energy generation was negatively affected by P precipitation in the cathode chamber. As COD concentration increased, more precipitates on the cathode were observed. The current density decreased with increasing COD concentration due to the amount of precipitates on the cathode surface ( Figure 6 ). The precipitates on the cathode covered the surface area of the cathode; and this obstructed the mass transfer of ions and oxygen. This finding is consistent with the finding of Hirooka and Ichihashi [ 8 ], which showed that the electricity generated by MFCs with precipitate was lower than that of MFCs without precipitate. 3.3.2. P Recovery at Different Anode and Cathode Volumes Electrolyte volume is another important parameter for P recovery. Three different volumes (180, 225, and 270 mL) were used to investigate the impact of anode and cathode volumes on P recovery and energy generation. The volumes were chosen based on the total volume of the chamber and the electrode surface area. Increasing the electrolyte volume leads to increases in the P precipitated. The greater the volume of synthetic wastewater available at the anode and the cathode, the more P that can be recovered from the solution. By increasing anode chamber volume, more organic matter is offered for the organism at the anode for oxidation, and consequently; more reduction reactions occurred at the cathode. Moreover, the impact of changing chamber volume on cathode pH was not significant. The average power density output (COD = 0.7 g/L) of the MFC under three different volumes (180, 225, 270 mL) were 25, 20 and 18 mW/m 2 , respectively. Increasing anode chamber volume from 180 mL to 270 mL leads to decreases in the average power density from 25 mW/m 2 to 18 mW/m 2 . The decrease in power density occurred due to the amount of P precipitates in 270 mL which was greater than 180 mL. Anode and cathode volumes have an impact on energy generation, as increasing the volume leads to decreases in the power density ( Figure 8 ). Achieving high power density and high P recovery concurrently in the system would require further research and development as P precipitation on the cathode leads to deteriorating electricity production in the MFC; and this reduces the generated power. 3.3.3. P Recovery at Different Aeration Flow Rates at the Cathode Cathode aeration is an important parameter for system optimization. The microbial fuel cells were operated under controlled dissolved oxygen concentrations; different aeration flow rates were used to examine the impact of aeration on both energy generation and P recovery. Identifying the impacts of aeration flow rates on MFC performance is important for scaling up the system and assessing the operational cost. Previous studies have shown that cathode aeration has a great impact on energy generation, where the current increased as the DO concentration increased. These results show the importance of supplying the optimal amount of oxygen to obtain optimal performance of the MFC [ 18 ]. The average current density of the MFC under three different aeration flow rates (no aeration, 50 and 150 mL/min) were 92, 127 and 155 mA/m 2 , respectively. Applying aeration in the cathode chamber with a flow rate of (150 mL/min) increased the average current density by more than 40% (from 92 to 155 mA/m 2 ) ( Figure 9 ). In addition, the generated power was increased, and these were in agreement with similar findings by Mashkour and Rahimnejad [ 18 ]. With regard to P recovery, it was found that increasing aeration flow rate at the cathode leads to increase in cathode pH, where oxygen is being used as an electron acceptor and hydroxide ions are produced [ 10 ]. With (50 mL/min) flow rate, cathode pH reached 8.5, which was optimal for P recovery as struvite; whereas with no aeration, cathode pH reached 7.5 ( Figure 10 ). The difference in P recovery between the flow rates of 50 mL/min and 150 mL/min was not significant; however, there was an increase of 25% in power density when the aeration flow rate increased from 50 to 150 mL/min. These results show that passive or low aeration cannot supply enough oxygen for the transferred protons from the anode chamber. These findings also show that cathode aeration is an influencing factor for energy generation and P recovery. 3.4. Statistical Analyses The correlations between the monitored parameters and MFC performance was studied to identify the influencing factors on MFC performance for P recovery and energy generation. Table 1 shows that COD significantly affected anode pH, cathode pH and current density. The more substrate in the anode chamber, the higher the pH at the cathode. Moreover, cathode aeration has an impact on cathode pH where supplying the optimal amount of oxygen to the cathode leads to an increase in the pH. High cathode pH leads to P precipitation as struvite and deterioration in electricity generation due to the precipitates on the cathode electrode. The correlation analysis suggests that COD and cathode aeration are the key influencing factors for P recovery and energy generation. Designing the system based on these findings can help to achieve the optimal condition for energy generation and P recovery."
} | 4,313 |
36103401 | PMC10653274 | pmc | 1,231 | {
"abstract": "Memtransistors that combine the properties of transistor\nand memristor\nhold significant promise for in-memory computing. While superior data\nstorage capability is achieved in memtransistors through gate voltage-induced\nconductance modulation, the lateral device configuration would not\nonly result in high write bias, which compromises the power efficiency,\nbut also suffers from unsuccessful memory reset that leads to reliability\nconcerns. To circumvent such performance limitations, an advanced\nphysics-based model is required to uncover the dynamic resistive switching\nbehavior and deduce the key driving parameters for the switching process.\nThis work demonstrates a self-consistent physics-based model which\nincorporates the often-overlooked effects of lattice temperature,\nvacancy dynamics, and channel electrostatics to accurately solve the\ninteraction between gate potential, ions, and carriers on the memristive\nswitching mechanism. The completed model is carefully calibrated with\nan ambipolar WSe 2 memtransistor and hence enables the investigation\nof the carrier polarity effect (electrons vs holes) on vacancy transport.\nNevertheless, the validity of the model can be extended to different\nmaterials by a simple material-dependent parameter modification. Building\nupon the existing understanding of Schottky barrier height modulation,\nour study reveals three key insights—leveraging threshold voltage\nshifts to lower write bias; optimizing lattice temperature distribution\nand read bias polarity to achieve successful memory state recovery;\nengineering contact work function to overcome the detrimental parasitic\ncurrent flow in short channel ambipolar memtransistors. Therefore,\nunderstanding the significant correlation between the switching mechanisms,\ndifferent material systems, and device structures allows performance\noptimization of operating modes and device designs for future memtransistors-based\ncomputing systems.",
"conclusion": "Conclusions In summary, the fundamental understanding\nof the field-driven vacancy\ntransport and the origin of resistive switching in memtransistor is\ndiscussed using an accurate physics-based model, where nonlinear equations\nfor lattice temperature, vacancy dynamics, and channel electrostatics\nare modeled self-consistently. The model overcomes the limitations\nof previously reported methods that not only overlook the effects\nof lattice temperature and vacancy kinetics on RS but also lack insights\non device downscaling implications. Our findings suggest that channel\nelectrostatic potential and threshold voltage variations due to the\nspatial distribution of chalcogen vacancy distribution are dominant\nfactors contributing to resistive switching. The observation is attributed\nto the interplay between lattice self-heating and electric fields,\nwhich create an asymmetric effect on vacancy distribution. Additionally,\nit is found that the tuning of the lattice temperature distribution\nthrough the gate-source and gate-drain voltage, with the same read\nvoltage polarity, is the knob to achieving memory reset operation.\nWe also show that the model is able to explain the observed anomalous\nconductance crossover in short channel devices to be the result of\ndrain voltage-induced parasitic current. This understanding led to\nthe proposal of contact work function engineering to suppress the\nleakage current flow. We believe that the critical insights on the\nswitching mechanism and identification of the parameters that affect\nthe RS mechanism allow predictions on reliable operation, material\nselection, and device optimization for future computing applications.",
"discussion": "Results and Discussion Physics-Based Model for Defect-Engineered Memtransistor The memtransistor has a planar three-terminal structure comprising\nof source and drain with the vacancy-rich channel layer, electrostatically\ncoupled to the gate electrode. Figure 1 a shows the device architecture employed in this study,\nwhere the RS layer is a multilayer WSe 2 with Se defects\ncreated by a controlled process of argon plasma treatment. While the\nWSe 2 switching layer is modeled as a bulk semiconductor\nof 10 nm thickness with known material properties such as dielectric\nconstant, 19 effective mass, 20 band gap, 21 , 22 etc., the Se vacancy\ndefects are modeled as positively ionized dopants. The assumption\nof Se vacancy defects as positively ionized dopants complies with\nthe reported first-principles density functional theory (DFT) calculations,\nthat the Se vacancies in WSe 2 form a deep-trap level, located\nat an energy level of 0.1 eV from the conduction band. 23 It is to be noted that such vacancy migration\nmechanisms have been widely investigated in conventional two-terminal\nmemristors. 24 , 25 Specifically, in interface-type\nmetal oxide-based memristors, the RS mechanism is attributed to the\nchanges in the SBH due to the concentration difference of oxygen vacancies\nnear the contact. However, the memory window gets undermined in interface-type\nswitching leading to difficulty in distinguishing between different\nstates. On the other hand, in a memtransistor, the defects are employed\nas dopants in the channel, and an additional gate terminal could electrostatically\nmodulate dopant distribution, which would result in much larger conductance\nchanges than the SBH or contact resistance changes. Figure 1 Self-consistent physics-based\nmodel for memtransistors. (a) Schematic\ndiagram of WSe 2 memtransistor with Se defective channel\nlayer. C1 is the cut line along the channel. (b) Calibrated initial\nvacancy distribution profile ( n d ( t = 0 s)) across C1. (c) Flowchart illustrating the parameters\nand equations self-consistently solved in two numerical solvers. To quantify the concentration of Se vacancies ( n d ) formed and its impact on the transport properties,\nthe memtransistor I d -V g characteristics before and after Ar plasma treatment\nare analyzed as explained in Supporting Information 1 . The vacancy distribution profile with a peak concentration\nof 1 × 10 18 cm –3 as calibrated from\nthe V t shift after Ar plasma treatment\nis employed as the initial condition for the n d profile ( Figure 1 b). We also tested the output characteristics of the non-Ar\nplasma-treated devices to understand the effect of interface trap\ncharges at the WSe 2 /SiO 2 interface ( Supporting Information 2 ). With the observed\nminimal hysteresis, one can rule out the role of interface traps on\nthe RS mechanism. The systematic investigation of the RS mechanism\nof memtransistors involves solving both transistor electrostatics\nand vacancy dynamics self-consistently. The complete resistive switching\nbehavior for the hole and electron-dominated transport regimes is\ncaptured by solving five partial differential equations (PDEs)—(1)\nPoisson’s equation, (2) current continuity equation for electrons,\nand (3) holes, (4) lattice heat flow equation, and (5) continuity\nequation for vacancies considering drift/diffusion migration. The\ncomplete iterative simulation workflow is depicted in Figure 1 c. The thermodynamic model\nis employed in a semiclassical device simulator, where the drift-diffusion\napproximation is extended to account for electro-thermal properties\nunder the assumption that the charge carriers are in thermal equilibrium\nwith the lattice. 26 Accordingly, the Poisson\nequation, current continuity equation, and lattice heat equation are\nself-consistently solved to deduce the spatial distribution of electrostatic\npotential and lattice temperature across the device. Thus, the effect\nof lattice self-heating on vacancy migration and conductance modulation\nis also investigated in this work. However, the drift/diffusion\ntransport of vacancies is described\nusing the ion hopping model proposed by Mott–Gurney. 13 In this approach, the vacancy transport is modeled\nby considering the vacancy migration induced by three factors: lateral\nelectric field, vacancy concentration gradient, and lattice temperature\ngradient. As shown by eq 5 in Figure 1 c, the application of an electric field causes defects\nto migrate in a direction that is determined by the ionic charge and\nfield direction. Here the term v.n d represents\nthe drift flux due to the electric field, where v is the vacancy mobility, and Fick’s diffusion flux ( D∇n d ) is due to the vacancy concentration\ngradient. The Soret diffusion flux D.S n d ∇ T indicates the movement of vacancies along\nthe temperature gradient due to Joule heating. We consider only the\nin-plane thermal conductivity of WSe 2 in the lattice heat\nflow equation, since the cross-plane thermal conductivity of WSe 2 is reported to be at least 30× smaller than the in-plane\ncomponent. 27 , 28 Additionally, owing to the high\ncross-plane activation energy of vacancies due to the van der Waals\ngap, the vacancy distribution may not be significantly modified by\nthe thermal conductivity in the cross-plane direction. For each iterative\nstep, the vacancy dynamics are evaluated from the continuity equation\nincluding three forces such as diffusive, thermal, and electric field\nby using the solution of the potential and lattice temperature derived\nfrom the earlier step. An adiabatic approximation is adopted here,\nas we assume that the vacancy and carrier dynamics work in different\ntime scales and can be modeled independently. The details of the different\nparameters used in the physics-based model are discussed in the methods section , and the assumptions used in the\nmodel are described in Supporting Information 3 . Effect of Lattice Temperature and Electric Field on Vacancy\nMigration Understanding the correlation between the role\nof electric field and lattice temperature on vacancy migration under\ndifferent carrier types is essential to capture the complete vacancy\ndynamics in a memtransistor. The carrier transport mechanism in an\nambipolar memtransistor under static vacancy distribution is described\nin Supporting Information 4 , where V tp is defined as the threshold voltage for holes,\nwhich corresponds to the gate voltage at which the current conduction\ntransitions from thermionic to the tunneling-dominated regime. While\nelectric field-induced vacancy migration has been well-studied, 7 the role of lattice self-heating on vacancy dynamics\nis yet to be investigated. The interaction between electric field-accelerated\nfree carriers and the crystal lattice heats up the transistor during\nits operation. 29 Given that the vacancy\ndiffusivity and mobility accelerate with lattice temperature, the\ninclusion of self-heating effects is critical for the accurate prediction\nof memory states. The simulated spatial distribution of the electric\nfield and lattice temperature under the hole-dominated conduction\nregimes is shown in Figure 2 a,b, respectively. The lattice temperature profiles are consistent\nwith the previously reported self-heating effects on two-dimensional\nmaterial-based transistors. 30 , 31 However, it is to be\nnoted that the experimental demonstration of lattice temperature on\nsub-1 μm channel length demands innovative techniques in thermal\nmeasurements and characterization to overcome the diffraction-limited\nspatial resolution. 32 , 33 As shown in Figure 2 a, when the device is operated\nas a p-channel field-effect transistor (p-FET) ( V g is negative), for smaller − V d , the potential difference between the gate terminal\nand left contact is larger than that at right contact, resulting in\na higher electric field. As − V d bias increases, the electric field at the right contact gradually\nincreases ( Supporting Information 5 ). This\ncan be attributed to the larger depletion region at the right contact\nas shown in Supporting Information 6 . As\nthe − V d potential is further increased\nbeyond pinch-off voltage (| V d | > (| V g | – |V tp |)), the channel potential does not vary, and all the additional\npotential is dropped across the more resistive region near the right\ncontact (Figure S7a in Supporting Information 7 ). Consequently, the electric field peak near the right contact\nexceeds that at the left contact at V d = −7 V ( Figure S5b ). The pinch-off\ncondition can be inferred from the current saturation seen in the\noutput characteristics as shown in Supporting Information 8 (Figure S8a). Figure 2 Effect of lattice temperature and electric\nfield on vacancy migration.\nSpatial distribution of (a) electric field, (b) lattice temperature\nas a function of V d for the hole-dominated\nregime ( V g = −20 V), spatial distribution\nof (c) electric field, (d) lattice temperature as a function of V d for the electron-dominated regime ( V g = 5 V), (e) vacancy distribution profile during\nthe hole-dominated regime (“After + V d ” describes the state after application of a positive\nbias at the right terminal of the device), (f) vacancy distribution\nprofile during the electron-dominated regime after application of\npositive bias. In order to examine the effect of electrons as\nthe majority carrier\ntype on vacancy transport, we investigate the spatial distribution\nof electric field and lattice temperature under the electron-dominated\nconduction regime at V g = 5 V ( Figure 2 c,d). For V d > 0, a higher electric field is observed\nat\nthe left contact, owing to the larger potential difference between\nthe left contact and gate terminal ( Figure 2 c and Figure S7b ). However, due to the Se vacancy-induced n -doping,\nthe threshold voltage for electrons ( V tn ) is a large negative value. Hence even at the maximum applied V d potential of 7 V, the condition V d >|V g – V tn | does not arise, and the\nchannel pinch-off\ndoes not happen for the electron-dominated regime as shown in the\noutput characteristics in Supporting Information 8 (Figure S8b). The consequence of this behavior is that the\nelectric field intensity remains always high at the left contact,\neven at high V d , contrary to the electric\nfield spatial distribution observed in hole-dominated conduction.\nThe vacancy distribution profile for electron and hole-conduction\nregimes is shown in Figure 2 e,f, respectively. The spatial distribution of temperature\nand potential used to generate the vacancy distribution is added in Supporting Information 9 . The positively charged\nSe vacancies move in the direction of the electric field and toward\nlow-temperature regions. While the vacancy migration due to the electric\nfield as well as lattice temperature is along the same direction for\nhole conduction, the driving forces oppose each other during the electron\nconduction regime ( Supporting Information 10 ). Thus, the vacancy migration strength is stronger for the hole\nconduction regime, whose implication for the conductance modulation\nwill be discussed in the next section. By leveraging the ambipolar\nconductance behavior of the WSe 2 memtransistor, the physical\ninsights on the vacancy interaction between different types of carriers\nallow for additional functionalities in future memristive computing\napplications. Resistive Switching Mechanism in Hole-Dominated Conduction Regime Unlike a transistor, a memtransistor requires the application of\nhigh drain bias and resultant high electric field to promote vacancy\ntransport. Consequently, the channel electrostatics is expected to\nbe largely affected by the high drain bias, particularly when the\nchannel thickness is small (∼10 nm). Here, we show that one\nsuch implication of high drain bias is the current rectification observed\nin memtransistors ( Figure 3 a). Although such characteristics have been reported previously, 7 , 15 the underlying physical mechanism has not been investigated thoroughly.\nA sound understanding of the rectification effect is critical because\nthe magnitude of current affects the lattice temperature of the device\nas discussed in the previous section, which in turn modulates the\nvacancy mobility. Figure 3 Resistive switching mechanism in hole-dominated conduction\nregime.\n(a) Output characteristics of memtransistor operated in hole conduction\nregime, (b) Se vacancy distribution for Path 1 and Path 2, (c) energy-band\ndiagram for Path 1 and Path 2, (d) Se vacancy distribution profile\nfor Path 3 and Path 4, (e) energy-band diagram for Path 3 and Path\n4, (f) resistance change from State 0 to 2 illustrating nonrecovery\nof memory states, where State 1 is read at V d = −0.5 V and State 2 is read at V d = +0.5 V. Our study shows that the origin of the observed\ncurrent rectifying\ncharacteristics is the drain voltage-induced tunneling barrier width\nmodulation at the carrier injection contact, as a function of the\npolarity of V g and V d . For the hole-dominated conduction regime, a forward current\nrectifying characteristic is observed, that is, a larger current for\npositive drain bias than negative bias ( Figure 3 a). To evaluate the role of drain bias polarity\non the current transport behavior, we carefully analyzed the dependency\nof the carrier injection terminal and the bias application terminal\non the Schottky barrier (SB) width modulation ( Supporting Information 11 ). It is to be noted that, during\npositive drain bias operation, the hole injection terminal (right\ncontact) is the same as the bias applied terminal. Due to the high\napplied bias at the hole injection terminal, the resulting high electric\nfield modulates the hole barrier at the contact by reducing the barrier\nwidth, leading to a larger tunneling current ( Figure S11b ). On the other hand, for negative drain bias,\nthe potential difference and thus the electric field is smaller at\nthe hole injection (left) terminal, since it is grounded. Thus, the\ndrain bias does not alter the tunnel barrier width, resulting in a\nsignificantly smaller current ( Figure S11a ). The observations shed light on how the current transport behavior\nis influenced under high-bias operation, where the key is the strong\ndependence of drain bias on the SB width, and the current rectification\nis purely a “contact” effect. To investigate the\nRS behavior of the WSe 2 memtransistor\nunder a hole-rich environment, we analyze the device’s response\nto voltage pulses of different polarities at V g = −20 V. The responses are modeled by describing the\nvacancy dynamics and the resultant electrostatics variations along\nthe channel and near the contacts. The sequence of bias voltage applications\nis as follows: Path 1:0 to −7 V, Path 2: −7 V to 0 V,\nPath 3:0 to 7 V, Path 4:7 to 0 V. The electrostatic potential and\nlattice temperature profile when the device is operated for the different\npolarities of V d bias is shown in Supporting Information 12 . As discussed in the\nprevious section, there is an asymmetric distribution of electrostatic\npotential and lattice temperature at V g = −20 V, V d = −7 V. This\nimplies that the electric field is high at the contact/channel interface\nand significantly reduces toward the center of the channel. The resultant\nvacancy distribution profile before (black) and after (red) application\nof V d = −7 V is shown in Figure 3 b, wherein the positively\ncharged Se vacancies accelerate in the direction of the electric field,\nleading to a vacancy-depleted region at the left contact (hole-injection\nterminal). It is to be noted that all the snapshots of vacancy distribution\nshown are taken at the maximum drain bias condition ( V d = +7 V and −7 V), where significant vacancy redistribution\nhappens. The experimental demonstration of such an in-plane vacancy\ndistribution profile on long-channel devices has been reported. 7 , 15 Although the experimental demonstration of n d profile is not reported in this work, the simulated vacancy\nprofile in response to the electric field is consistent with the prior\nreport from Jakub et al. 15 To study how\nvacancy redistribution affects the lateral electrostatic potential,\nthe energy-band diagram along the channel length is carefully analyzed\nas shown in Figure 3 c. It can be observed that the depletion of vacancies at the left\nterminal (hole injection terminal) translates to a source-to-channel\nbarrier of 0.07 eV for holes. Thus, the threshold voltage for the\nholes increases with a reduction in the drain current for the same\nread gate voltage (Path 1 to 2 transition). With the reversal\nof voltage polarity (positive drain bias), the\nright contact becomes the hole injection point. The Se vacancies are\nthen driven away from the right electrode in response to the electric\nfield direction and lattice temperature ( Figure 3 d). The resulting vacancy depletion leads\nto a larger potential barrier of 0.32 eV for holes at the source-channel\njunction, as shown in Figure 3 e. The larger potential barrier is attributed to the 40×\nincrease in drain current for positive bias and the corresponding\n∼60 K increase in the lattice temperature resulting in a 500×\nvacancy change near the carrier injection terminal. Thus, the threshold\nvoltage for holes increases, and the conductance switches from high\nto low, in agreement with the experiment data (Path 3 to 4 transition).\nIt is noteworthy that the observed larger memory window for positive\ndrain bias stems from the larger vacancy migration due to higher lattice\ntemperature. As explained earlier, it is the drain voltage-induced\ncurrent rectification that results in a larger current and self-heating\nfor positive drain bias. The model is able to capture the changes\nin conductance consistent with the experiment data as shown in Figure 3 a. In conclusion,\nwe found that the accumulation or depletion of vacancy\nconcentration along the memtransistor channel length represents a\n“barrier” to the carrier flow from source to drain,\nwhich affects the threshold voltage ( V t ) of the device and has superior control over the conductance modulation.\nWe also investigated the impact of a nonuniform vacancy distribution\nprofile along the thickness of the WSe 2 memtransistor on\nthe RS mechanism ( Supporting Information 13 ). It can be inferred that, despite the nonuniformity in vacancy\ndistribution along the thickness of the WSe 2 channel, n d redistribution happens laterally in response\nto the electric field and lattice temperature, resulting in vacancy\ndepletion near the carrier injection terminal and corresponding barrier\nfor carriers. Here we identify channel electrostatics and V t variations as additional key factors for memtransistor\nRS modulation apart from the widely investigated SBH modulation. All\nthe above analyses suggest the importance of incorporating the potential\nvariations along the whole channel (rather than merely at the contacts)\nin order to capture both accurate conductance transition as well as\nthe degree of conductance modulation. Optimization of the Gate and Drain Potential for Memory State\nRecovery Owing to the asymmetric vacancy distribution profile,\nthe memtransistor would require optimization of the lattice temperature\nand electric field profile through the gate-source and gate-drain\nvoltage tuning to achieve a true memory reset. This can be better\nunderstood by analyzing the resistance change versus voltage sweep\n( Figure 3 f), where\nthe device switches from high conductance to low conductance irrespective\nof drain bias polarity reinstating the fact that modulation of drain\nbias polarity alone does not necessarily result in a memory state\nrecovery ( Supporting Information 14 ). Such\nunipolar characteristics can be attributed to the fact that vacancy\ndepletion always occurs at the carrier injection terminal. During\nDC characterization of the memtransistor, there is no provision to\ntune the carrier injection point, as the write bias polarity determines\nthe carrier injection terminal ( Figure S14c ). We explain here the procedure for memory state recovery by accurate\ntuning of the gate and drain potential by the application of AC pulses,\nwhich circuit designers can rely on. Figure 4 a–c shows the sequence of the gate\nand drain bias application for successful memory state recovery. As\nanalyzed in Figure 3 and Supporting Information 12 , the application\nof positive bias results in a high electric field and temperature\nat the right contact. Consequently, the vacancies are drifted away\nfrom the right contact in the direction of the electric field, leaving\na vacancy-depleted area, that raises the barrier for holes ( Figure 4 a). Thus, conductance\nswitches from high (State 0) to low (State 1) as shown in Figure 4 d. Note that it is\nalready found from the earlier discussion that the sole reversal of\ndrain bias polarity does not suffice to recover the original n d profile. Given that conductance of the memtransistor\nlargely depends on the n d profile and\nthe carrier injection terminal, we attempt to restore the previous n d profile by identifying the appropriate bias\nconditions, while not changing the injection terminal. Hence a write\npotential of V g = 5 V and V d = −7 V is applied, and the conductance state\nis read at the same read voltage of V d = 0.5 V. With the reversed electric field polarity, the vacancies\nnow drift toward the right contact, thus restoring the previous state\nof high conductance (State 0) as shown in Figure 4 b. The scenario is now similar to the formation\nand dissolution of conductive filament in a filamentary ReRAM, by\nmodulating the gate potential without changing the polarity of the\nread voltage. Following this, the application of positive bias at V g = −15 V results in a larger barrier\nfor holes, leading to a low conductance (recovered to State 1) as\nshown in Figure 4 c.\nA schematic representation of how memory state recovery is feasible\nby reading the conductance state from the same carrier injection terminal\nusing AC bias is shown in Supporting Information 15 . Our analysis suggests that vacancy profile recovery is\nfeasible via tuning of drain and gate potential to render reproducible\nswitching of memory states, under the same read voltage polarity ( Figure 4 d). Figure 4 Procedure for memory\nstate recovery. (a) n d profile from State\n0 to State 1 after V g = −15 V, V d = 7 V, (b) V g =\n5 V, V d = −7\nV, (c) V g = −15 V, V d = 7 V, (d) resistance change from State 0 to 3 implying\nsuccessful memory state recovery, where all the resistance states\nare read at V d = +0.5 V. All the spatial\ndistribution is across the cut line C1 as shown in Figure 1 a. Switching Power Efficiency Gains with Threshold Voltage Modulation After having established that the channel electrostatics and lattice\ntemperature have superior control on conductance modulation, we verify\nif there are any switching energy gains by incorporating threshold\nvoltage modulation. As the switching energy per bit is the integral\nof switching power over the pulse width, a lower programming voltage\nwould lead to a reduction in switching power and thus energy. In conventional\nSBH-modulated resistive switching, the vacancies need to accelerate\nuntil the opposite electrode to vary the SB width. In contrast, we\nshow here that accounting for vacancy modulation near the carrier\ninjection contact edge would suffice to alter the source-to-channel\nbarrier and thus the threshold voltage of the device as discussed\nin Figure 5 . Figure 5 Switching power\nefficiency gains. (a) Spatial distribution of vacancies\nwith a uniform vacancy profile as the initial condition. i initial, ii after application of V d = 3 V, and iii after application of V d = 4 V, (b) simulated transfer characteristics\nafter application of V d = 3 V and V d = 4 V for the corresponding vacancy profile\nshown in (a), (c) memory window vs gate voltage for the vacancy distribution\nshown in (a), (d) spatial distribution of vacancies with nonuniform\nvacancy profile as the initial condition. i initial, ii after application of V d = 3\nV, and iii after application of V d = 4 V, (e) transfer characteristics after application of V d = 3 V and V d =\n4 V for the corresponding vacancy profile shown in (d), (f) memory\nwindow vs gate voltage for the vacancy distribution shown in (d),\n(g) schematic representation illustrating how vacancy modulation at\nthe contact-channel interface results in improved power efficiency,\n(h) benchmark plot between write potential and memory window while\nincorporating V t modulation in comparison\nto other reported memtransistor works. We simulate here two kinds of vacancy distribution\n(uniform vs\nnonuniform along the length of the channel) as the initial condition\nto verify the hypothesis. Figure 5 a illustrates the vacancy distribution after the application\nof V d = 3 V and V d = 4 V with uniform vacancy distribution as the initial condition.\nFor V d = 3 V, vacancies are found to deplete\nnear the right contact and accumulate near the source-channel interface\n(ii in Figure 5 a).\nThe conductance change is due to V t shift,\nand hence conductance changes from high to low as the depleted vacancies\nact as a barrier for holes ( Figure 5 b). Furthermore, with the application of a larger write\nvoltage of V d = 4 V, vacancies now travel\nalong the channel length and are accumulated beneath the left contact\n(iii in Figure 5 a).\nIn this case, the conductance change is due to contact resistance,\nand hence conductance changes from low to high as the vacancies at\nthe hole injection point reduce contact resistance ( Figure 5 b). We can observe that the\nmemory window is larger with the application of V d = 3 V, resulting in an improvement of switching power\nefficiency by 25%. A similar reduction in write bias is observed for\nthe case where the initial vacancy distribution is nonuniform ( Figure 5 d–f). The\ndependence of the memory window on gate voltage ( V g ) can be explained with respect to the tuning of channel\nresistance in the memtransistor as a function of V g ( Figure 5 c,f). As the read V g is varied from the\nsubthreshold ( V g < V t ) to the superthreshold ( V g > V t ) conduction regime, there is\na\nsignificant drop in channel resistance due to the large accumulation\nof charge carriers in the channel. Thus, the contribution from vacancy\ndistribution-induced channel electrostatics variations (i.e., source-to-channel\nbarrier variations) reduces leading to a reduction in the memristive\nwindow with increasing V g . A schematic\nrepresentation of how V t shift can be\nleveraged for improved power efficiency is shown in Figure 5 g. Figure 5 h shows a benchmark plot between memory state\nwrite potential and memory window while incorporating V t modulation in comparison to other reported memtransistor\nworks. 7 , 15 , 34 Resistive Switching Mechanism in the Electron-Dominated Conduction\nRegime To further evaluate the influence of an electron-rich\nenvironment on RS behavior, we investigate the memtransistor response\nby appropriately biasing the device with a gate bias of V g = 0 V. For the electron-dominated conduction regime,\na reverse rectifying characteristic is observed, where negative drain\nbias results in a larger current than positive bias ( Figure 6 a). As shown in Supporting Information 16 , the high drain voltage-induced\nreduction in tunnel barrier width happens for negative drain bias,\nresulting in a larger current. Figure 6 Resistive switching mechanism in electron-dominated\nconduction\nregime. (a) The output characteristics of memtransistor operated in\nelectron conduction regime, (b) Se vacancy distribution profile before\n(Path 1) and after (Path 2) application of negative V d , (c) Se vacancy distribution profile before (Path 3)\nand after (Path 4) application of positive V d , (d) comparison of the lateral electric field with increasing\ngate potential for electron-dominated conduction regime, (e) comparison\nof electron concentration with increasing gate potential for electron-dominated\nconduction regime, (f) comparison of the lateral electric field with\nincreasing gate potential for hole-dominated conduction regime, (g)\ncomparison of hole concentration with increasing gate potential for\nhole-dominated conduction regime, (h) schematic representation of\nthe key parameters affecting the conductance changes in memtransistor. The sequence of V d application\nis as\nfollows: Path 1: 0 to −7 V, Path 2: −7 to 0 V, Path\n3: 0 to 7 V, Path 4: 7 to 0 V, and the response of vacancies is studied\nas a function of bias voltage polarity. The electrostatic potential\nand lattice temperature profile when the device is operated for the\ndifferent polarities of V d bias is shown\nin Supporting Information 17 . During Path\n1 (negative drain bias), the vacancies are found to deplete from the\nright terminal, in response to the electric field direction ( Figure 6 b). However, it can\nbe observed that the vacancy migration is less significant, resulting\nin smaller hysteresis. On the other hand, for positive drain bias,\nthe electron current is low, owing to the drain voltage-induced current\nrectification behavior as discussed in the previous section. Due to\nlow current and smaller lattice temperature, vacancy distribution\nis hardly altered ( Figure 6 c), resulting in minimal hysteresis. The simulated hysteric\noutput curve at V g = 0 V is also consistent\nwith the experiment as shown in Figure 6 a. Note that, despite the high lattice temperature,\nthere is substantially less memory hysteresis for negative drain bias\n(Paths 1 and 2). In order to unveil other contributing factors to\nthe RS memory window, we first analyze the influence of carrier concentration\nand carrier screening on vacancy dynamics. As opposed to a hole-rich\nenvironment, the positive vacancies in the channel could be screened\nby the electrons, where the vacancy potential decays exponentially\n( Supporting Information 18 ). The decay\nconstant is the characteristic length called Debye length, which is\nrelated to electron density and lattice temperature. 35 The effect of carrier screening on the applied electric\nfield is studied by comparing the lateral electric field ( E x ) variation during both hole and electron accumulation\nregimes ( Figure 6 d,f). Figure 6 e,g shows the electron\nand hole concentrations, respectively, in the channel as a function\nof different V g . It can be noted that,\nfor an electron-rich regime, the magnitude of E x decreases with increased positive V g near the right contact and at larger electron concentrations,\nthe E x reduction ceases ( Figure 6 d). On the other hand, under\nthe hole accumulated regime, there is no significant variation in E x as a function of hole concentration ( Figure 6 f). We attribute\nthis behavior to the carrier-induced screening of potential, which\ndepends strongly on the carrier concentration. Apart from carrier-induced\nscreening, we identify the read gate\nvoltage to be an additional key factor contributing to the memory\nwindow. As the subthreshold conduction regime ( V g < V t ) is governed by thermionic\nemission from the contacts and the source-to-channel barrier, the\nlarger conductance difference due to threshold voltage changes can\nbe capitalized by reading the conductance state at a V g near-threshold regime. As discussed in the previous\nsection, as the read gate voltage is transitioned from a subthreshold\nto a near-threshold regime, the conductance change at a particular\ngate read voltage reduces, resulting in the closing of the memory\nwindow. It is the interplay between the threshold voltage and contact\nresistance-dominated regime that is resulting in the tuning of the\nmemory window. Due to the n-type doping from the Se vacancies, it\nis much easier for the device to attain the contact resistance-dominated\nregime compared to the hole-dominated conduction. The key parameters\naffecting the conductance changes in the memtransistor are summarized\nin Figure 6 h. The reason\nfor observed small hysteresis is a combination of factors such as\ncarrier screening, smaller current, and saturation regime of V g operation. Impact of Channel Length Scaling on Memtransistor RS Behavior Scaling of the memory device is necessary to improve the memory\ndata capacity, and hence the investigation of the impact of channel\nlength scaling deserves keen attention. 36 Here we investigate the RS mechanism on a 100 nm long memtransistor.\nThe DC RS behavior is shown in Figure 7 a, and the sequence of bias voltage applications is\nas follows: Path 1: 0 to −7 V, Path 2: −7 to 0 V, Path\n3: 0 to 7 V, and Path 4: 7 to 0 V. As shown in Figure 7 a, the output characteristics show slope\nchanges as a function of negative V d ,\nresulting in a conductance crossover point at room temperature. To\nunderstand this anomalous behavior, we thoroughly investigated the\nresponse of the device to negative drain bias. Owing to the reduced\nchannel length, the vacancies now experience a larger electric field\nstrength, and the corresponding vacancy distribution is shown in Figure 7 b. Figure 7 c shows that the simulated I d is consistent with the experiment. The total\ncurrent can be deconvoluted to investigate the independent contribution\nfrom electrons and holes, as shown in Figure 7 d ( Supporting Information 19 ). Our analysis suggests the presence of additional electron\ntunneling current from the opposite (right) contact, despite the device\nbeing operated as a p-FET with holes as the dominant carrier. The\nhigh bias at the right contact (electron injection terminal) modulates\nthe barrier width for electrons, resulting in tunneling for V d > −4 V. On the other hand, the hole\ncurrent saturates once it reaches the thermionic emission limit. Figure 7 Impact\nof channel length on memtransistor resistive switching behavior.\n(a) Resistive switching characteristic of a 100 nm channel length\nWSe 2 memtransistor for V g =\n−20 V. A conductance crossover is observed for negative drain\nbias at room temperature, (b) lateral vacancy distribution profile\nbefore and after application of negative bias at the right contact,\n(c) output characteristics comparison between experiment and simulation\nshowing slope changes due to parasitic electron current for negative V d , (d) total I d is\ndeconvoluted into individual contributions from hole and electron\ncurrent, (e) output characteristics showing suppression of electron\ncurrent with contact work function engineering, (f) memtransistor\ndevice schematic with a high work function metal at the right contact,\n(g) output characteristics of MoS 2 memtransistor with unipolar\nconduction, (h) total I d is deconvoluted\ninto individual contributions from hole and electron current for MoS 2 memtransistor. Hole current is smaller than electron current\nfor the entire range of gate voltages, (i) schematic diagram of MoS 2 memtransistor showing the absence of parasitic hole current. To examine the role of gate voltage in modulating\nthe conductance\ncrossover point, the electron current modulation as a function of\nnegative gate voltage is investigated. As shown in Supporting Information 20 , the onset of parasitic electron\ncurrent is dependent on the gate voltage. Specifically, as negative V g increases, a larger negative V d is required to achieve the same electron current. These\ncharacteristics can be explained by examining the energy-band profile\nat the right contact where electron injection happens ( Figure S20b ). With higher negative V g , the barrier for electrons increases, and thus a larger\nnegative V d is required to reduce the\nSB width and initiate a tunneling current. Note that such parasitic\nelectron current is undesirable, as it reduces the memory window and\nwould also result in a read error. We identify a few control knobs\nto alleviate this issue. Reducing the operating write voltage is one\nsolution to circumvent this issue. However, vacancy migration could\nbe compromised due to reduced driving force. It should be noted that\nthe anomalous parasitic electron current arises only for negative\nwrite voltage consistent with the experiment. This is because, during\npositive bias, there is no parasitic electron current, as V d does not modulate the injection barrier. Thus,\nreplacing the right contact electrode with a higher work function\nmetal that results in larger SBH for electrons can eliminate the parasitic\ncurrent ( Figure 7 e,f). On the other hand, we found that a unipolar memtransistor based\non the MoS 2 switching layer is less susceptible to such\nparasitic drain voltage-induced current. Figure 7 g shows the output characteristics of the\nMoS 2 memtransistor for increasing gate voltage. Unlike\nthe WSe 2 memtransistor discussed previously, there is no\nslope change in the conductance characteristics as a function of drain\nvoltage. The total current is deconvoluted into electron and hole\ncurrent, to investigate the independent contribution as shown in Figure 7 h. Although the presence\nof parasitic hole current is observed, the magnitude never surpassed\nthe total electron current, for the entire range of gate voltages.\nThe behavior is attributed to the smaller SBH for electrons and larger\nSBH for holes that restricts the parasitic hole tunneling current\n( Figure 7 i). Table 1 summarizes\nthe suitability of different channel materials for memtransistor applications\nagainst several features. The ease of defect creation without compromising\non the transport attributes is a critical factor. While the formation\nenergy for X in MX 2 (M-metal, X-chalcogenide) is the smallest\nfor Te in MoTe 2 , the comparable formation energy between\nMo and Te will hinder the selectivity for defect formation. On the\nother hand, with a larger formation energy difference between M and\nX in MoS 2 and WSe 2 , X defect creation is guaranteed. 14 Being a functional material with ambipolar conductance,\nWSe 2 allows for additional dominant charge carrier tunability\nagainst unipolar devices such as MoS 2 . While the ambipolar\nnature of the WSe 2 memtransistor would be suited to reduce\nthe device count that could relax routing congestions and process\ncomplexity constraints in advanced scaled technologies, 41 − 43 the associated parasitic conduction needs to be taken care of using\ndevice architectural innovation such as the proposed contact work\nfunction engineering. In the case of metal oxide semiconductors, such\nas the widely used InGaZnO, the high vacancy formation energy 39 and lack of carrier tunability limits its utility\nfor memtransistors. The performance challenges and opportunities of\nmemtransistors with the solutions proposed in the work are summarized\nin Figure 8 . Table 1 Comparison of Different Memristive\nChannel Materials for Their Suitability for Memtransistor Application RS layer vacancy formation\nEnergy 14 carrier mobility\n(cm 2 / (V s)) suitability\nfor defect creation ambipolar\nconduction implication\non short channel devices WSe 2 W-2.7 eV; Se-5 eV 230 4 Yes Yes Parasitic current flow due\nto drain voltage-induced barrier width reduction can be suppressed\nby contact work function engineering. MoS 2 Mo-7.5 eV; S-2.9 eV 200 37 Yes No Immune against the parasitic\ncurrent flow. MoTe 2 Mo-3.75 eV; Te-2.5 eV 80 38 No Yes Not applicable due to challenges\nin defect creation. Semiconductor Metal Oxides O-6.6 eV 39 70 40 No No Not applicable since the\nvacancy formation energy is too high. Figure 8 Schematic representation of the performance challenges faced by\nmemtransistors and the solutions proposed in this work to overcome\nthem."
} | 10,833 |
26500622 | PMC4594100 | pmc | 1,233 | {
"abstract": "Aerobic methane-oxidizing bacteria (MOB) are a diverse group of microorganisms that are ubiquitous in natural environments. Along with anaerobic MOB and archaea, aerobic methanotrophs are critical for attenuating emission of methane to the atmosphere. Clearly, nitrogen availability in the form of ammonium and nitrite have strong effects on methanotrophic activity and their natural community structures. Previous findings show that nitrite amendment inhibits the activity of some cultivated methanotrophs; however, the physiological pathways that allow some strains to transform nitrite, expression of gene inventories, as well as the electron sources that support this activity remain largely uncharacterized. Here we show that Methylomicrobium album strain BG8 utilizes methane, methanol, formaldehyde, formate, ethane, ethanol, and ammonia to support denitrification activity under hypoxia only in the presence of nitrite. We also demonstrate that transcript abundance of putative denitrification genes, nirS and one of two norB genes, increased in response to nitrite. Furthermore, we found that transcript abundance of pxmA , encoding the alpha subunit of a putative copper-containing monooxygenase, increased in response to both nitrite and hypoxia. Our results suggest that expression of denitrification genes, found widely within genomes of aerobic methanotrophs, allow the coupling of substrate oxidation to the reduction of nitrogen oxide terminal electron acceptors under oxygen limitation. The present study expands current knowledge of the metabolic flexibility of methanotrophs by revealing that a diverse array of electron donors support nitrite reduction to nitrous oxide under hypoxia.",
"conclusion": "Conclusion The present study demonstrates that an aerobic methanotroph – M. album strain BG8 – couples the oxidation of C 1 (CH 4 , CH 3 OH, CH 2 O, HCO 2 H), C 2 (C 2 H 6 , C 2 H 6 O), and inorganic (NH 3 ) substrates to NO 2 - reduction under O 2 limitation resulting in release of the potent greenhouse gas N 2 O. The ability to couple C 1 , C 2 , and inorganic energy sources to O 2 respiration and denitrification gives M. album strain BG8 considerable metabolic flexibility. We propose a model for methane driven denitrification in M. album strain BG8 ( Figure 6 ). This discovery has implications for the environmental role of methanotrophic bacteria in the global nitrogen cycle in both N 2 O emissions and N-loss. Comparing the genome and physiology of the NO 2 - respiring M. album strain BG8 to NO 3 - respiring M. denitrificans FJG1 suggests that the inability of M. album strain BG8 to reduce NO 3 - to N 2 O is likely due to the absence of a dissimilatory nitrate reductase in the genome, but that expression of predicted denitrification genes, nirS and norB1 , enable this aerobic methanotroph to respire NO 2 - . FIGURE 6 Proposed model for NO 2 - respiration and central metabolism in Methylomicrobium album strain BG8 . During hypoxia, M. album strain BG8 utilizes electrons from aerobic CH 4 oxidation to respire NO 2 - . Abbreviations: pMMO, particulate methane monooxygenase; mdh, methanol dehydrogenase; Cyt, cytochrome; nor, nitric oxide reductase; nir, nitrite reductase; ndh, NAD(P)H dehydrogenase complex; Q, coenzyme Q; bc 1 , cytochrome bc 1 complex.",
"introduction": "Introduction Aerobic methane-oxidizing bacteria (MOB) form an important bridge between the global carbon and nitrogen cycles, a relationship impacted by the global use of nitrogenous fertilizers ( Bodelier and Steenbergh, 2014 ). Ammonia (NH 3 ) and nitrate (NO 3 - ) can stimulate the activity of methanotrophs by acting as a nitrogen source for growth and biomass production ( Bodelier et al., 2000 ; Bodelier and Laanbroek, 2004 ). Further, some methanotrophs such as Methylomonas denitrificans utilize NO 3 - as an oxidant for respiration under hypoxia ( Kits et al., 2015 ). Evidently, denitrification in aerobic methanotrophs functions to conserve energy during oxygen (O 2 ) limitation ( Kits et al., 2015 ). Alternatively, NH 3 and nitrite (NO 2 - ) can act as significant inhibitors of methanotrophic bacteria ( King and Schnell, 1994 ). NH 3 is a competitive inhibitor of the methane monooxygenase enzyme and NO 2 - , produced by methanotrophs that can oxidize NH 3 to NO 2 - , is a toxin with bacteriostatic properties that is known to inhibit the methanotroph formate dehydrogenase enzyme that is essential for the oxidation of formate to carbon dioxide ( Dunfield and Knowles, 1995 ; Cammack et al., 1999 ; Nyerges et al., 2010 ). In spite of the recent discovery that aerobic methanotrophs can denitrify, the energy sources, genetic modules, and environmental factors that govern denitrification in MOB are still poorly understood. M. denitrificans FJG1 respires NO 3 - using methane as an electron donor to conserve energy. However, it is not known whether C 1 energy sources other than CH 4 (methanol, formaldehyde, and formate) can directly support denitrification. Another possibility, which has not yet been investigated, is that C 2 compounds (such as ethane and ethanol) and inorganic reduced nitrogen sources (NH 3 ) support methanotrophic denitrification. Previous work shows that several obligate methanotrophs, including Methylomicrobium album strain BG8, oxidize ethane (C 2 H 6 ) and ethanol (C 2 H 6 O) using particulate methane monooxygenase (pMMO) and methanol dehydrogenase (MDH), respectively, even though neither substrate supports growth ( Whittenbury et al., 1970 ; Dalton, 1980 ; Mountfort, 1990 ). NH 3 may be able to support methanotrophic denitrification because many aerobic methanotrophs are capable of oxidizing NH 3 to NO 2 - : a process facilitated by the presence of a copper-containing monooxygenase (CuMMO) enzyme and, in some methanotrophs, a hydroxylamine dehydrogenase homolog ( Poret-Peterson et al., 2008 ). The ability to utilize alternative energy sources to support denitrification would augment the metabolic flexibility of methanotrophs and enable them to sustain respiration in the absence of CH 4 and/or O 2 . Methylomicrobium album strain BG8 is an aerobic methanotroph that belongs to the phylum Gammaproteo bacteria; the genome lacks a soluble methane monooxygenase but does contain one particulate methane monooxygenase operon ( pmoCAB – METAL_RS17430, 17425, 17420) and one operon encoding a putative copper monooxygenase ( pxmABC – METAL_RS06980, 06975, 06970) with no known function. The genome also contains gene modules for import and assimilation of NH 4 + ( amtB – METAL_RS11045 /gdhB – METAL_RS11695 /glnA – METAL_RS11070 /ald – METAL_RS11565), assimilation of NO 3 - ( nasA – METAL_RS06040 /nirB – METAL_RS15330, nirD – METAL_RS15325), oxidation of NH 2 OH to NO 2 - ( haoA – METAL_RS13275), as well as putative denitrification genes – cytochrome cd 1 nitrite reductase ( nirS – METAL_RS10995), and two copies of cytochrome c -dependent nitric oxide reductase ( norB1 – METAL_RS03925, norC1 – METAL_RS03930 /norB2 – METAL_RS13345). The recent release of several genome sequences of aerobic methanotrophs, including M. album strain BG8, points to the frequent presence of putative nitrite and nitric oxide reductases, while only three cultivated methanotrophs possess a respiratory nitrate reductase ( Stein and Klotz, 2011 ; Stein et al., 2011 ; Svenning et al., 2011 ; Khadem et al., 2012b ; Vuilleumier et al., 2012 ; Kits et al., 2013 ). It is also unclear whether methanotrophs that lack a respiratory nitrate reductase but possess dissimilatory nitrite and nitric oxide reductases are still capable of denitrification from NO 2 - . Moreover, due to the significant divergence of the methanotroph nirS from known sequences, it is not known, whether nirS is the operational nitrite reductase in the methanotrophs that lack a nirK ( Wei et al., 2015 ). While the genome of the nitrate respiring M. denitrificans FJG1 encodes both nirS and nirK nitrite reductases, transcript levels of only nirK increased in response to denitrifying conditions ( Kits et al., 2015 ). The goal of the present study was to test whether a variety of C 1 , C 2 , and inorganic energy sources can directly support denitrification, characterize the environmental factors that regulate NO 2 - -dependent N 2 O production in M. album strain BG8 and to assess the expression of its putative denitrification inventory.",
"discussion": "Discussion Methylomicrobium album Strain BG8 Produces N 2 O Only as a Function of Hypoxia and NO 2 - Batch cultivation of M. album BG8 clearly revealed that both NO 2 - and low O 2 were required for denitrification, as measured by N 2 O production. Although batch cultures of M. album strain BG8 have been shown to produce N 2 O previously in end-point assays ( Nyerges et al., 2010 ), the mechanism and required conditions for denitrification by this strain were not determined until now. N 2 O production by M. denitrificans FJG1 was also shown to be dependent on hypoxia ( Kits et al., 2015 ); however, this strain was able to respire NO 3 - in addition to NO 2 - likely due to the presence of a narGHJI dissimilatory nitrate reductase that is absent in the genome of M. album strain BG8. The genome of M. album strain BG8 encodes putative dissimilatory nitrite ( nirS ) and nitric oxide ( norB ) reductases ( Kits et al., 2013 ) like M. denitrificans FJG1; hence, it is likely that N 2 O by M. album strain BG8 is from the enzymatic reduction of NO 2 - to N 2 O via the intermediate NO. The correlation between N 2 O production and low O 2 tension is similar to two other microbial processes, aerobic denitrification in heterotrophic bacteria such as Paracoccus denitrificans and nitrifier denitrification in ammonia-oxidizing bacteria ( Richardson et al., 2001 ; Kozlowski et al., 2014 ). Aerobic denitrification in chemoorganoheterotrophs and nitrifier-denitrification in ammonia-oxidizing bacteria is a tactic used to maximize respiration during O 2 limitation or to expend surplus reductant ( Richardson et al., 2001 ; Stein, 2011 ). Utilization of NO 2 - in combination with or instead of O 2 in the respiratory chain of M. album strain BG8 would reduce the overall cellular O 2 demand, thus conserving O 2 for additional CH 4 oxidation. Thus, it is possible that M. album strain BG8 uses NO 2 - as a terminal electron acceptor under O 2 limitation to maximize total respiration. The N 2 O yield percentage from NO 2 - by M. album strain BG8 (5.1 ± 0.2%) is similar to that of Nitrosomonas europaea ATCC 19718 (ca. 4.8%) and one order of magnitude higher than that of Nitrosospira multiformis ATCC 25196 (0.27 ± 0.05%; Kozlowski et al., 2014 ; Stieglmeier et al., 2014 ). Denitrification by M. album Strain BG8 is Enzymatically Supported by Diverse Reductant Sources Resting cells of M. album strain BG8 reduced NO 2 - to N 2 O at the expense of any of four tested C 1 substrates (CH 4 , CH 3 OH, CH 2 O, HCO 2 H), the two C 2 substrates (C 2 H 6 , C 2 H 6 O), and NH 4 Cl. These data show that intermediates of the methanotrophic pathway and co-substrates of pMMO, MDH, and likely hydroxylamine dehydrogenase support respiratory denitrification. These results agree with previous work on the methanotroph Methylocystis sp. strain SC2, which couples CH 3 OH oxidation to denitrification under anoxia ( Dam et al., 2013 ). Remarkably, both C 2 compounds we tested – C 2 H 6 and C 2 H 6 O – supported denitrification. The ability of C 2 compounds to support denitrification in methanotrophs may have environmental significance as natural gas consists of ∼1.8–5.1% (vol%) C 2 H 6 ( Demirbas, 2010 ). Further, C 2 H 6 O is a significant product of fermentation by primary fermenters during anoxic decomposition of organic compounds ( Reith et al., 2002 ). The results also demonstrate that electrons derived from the oxidation of NH 3 to NO 2 - were effectively utilized by nitrite and nitric oxide reductases in M. album strain BG8, which represents yet another pathway for methanotrophic N 2 O production that is not directly dependent on single-carbon metabolism, provided that the methane monooxygenase can access endogenous reductant ( Dalton, 1977 ; King and Schnell, 1994 ; Stein and Klotz, 2011 ). Instantaneous O 2 consumption and N 2 O production measurements ( Figures 2 – 4 ) provide strong support that catabolism of C 1 – C 2 substrates and ammonia is directly coupled to NO 2 - reduction under hypoxia in M. album strain BG8. Some aerobic methanotrophs ferment CH 4 and excrete organic compounds such as citrate, acetate, succinate, and lactate ( Kalyuzhnaya et al., 2013 ). Some studies also suggest that methanotrophs only support denitrification within CH 4 -fed consortia by supplying these excreted organics to denitrifying bacteria, since methanotrophs were thought incapable of denitrification by themselves ( Costa et al., 2000 ; Knowles, 2005 ; Liu et al., 2014 ). Although M. album strain BG8 may excrete organic compounds under hypoxia when provided with CH 4 , the ability of CH 3 OH, CH 2 O, HCO 2 H, C 2 H 6 , C 2 H 6 O, or NH 3 oxidation to support denitrification unequivocally demonstrates the linkage between methanotroph-specific enzymology and denitrifying activity within a single organism. Transcription of Predicted Denitrification Genes, nirS and norB1 , Increased in Response to NO 2 - but not Hypoxia The expression of a nirS homolog in an aerobic methanotroph has been investigated so far only in the NO 3 - respiring M. denitrificans FJG1 ( Kits et al., 2015 ). Interestingly, the genome of M. denitrificans FJG1 encodes both the copper-containing ( nirK ) and cytochrome cd 1 containing ( nirS ) nitrite reductases and only the steady state mRNA levels of nirK increased in this strain in response to simultaneous O 2 limitation and NO 3 - availability ( Kits et al., 2015 ). In the case of M. album strain BG8, which only possesses a nirS homolog, we showed that the abundance of this nirS transcript responded positively to NO 2 - treatment but not to O 2 limitation. This suggests that NO 2 - availability alone elicits the expression of nirS , even though hypoxia was required for NO 2 - reduction to occur. The cytochrome c dependent nitric oxide reductase ( norB ) is widely found in the genomes of aerobic methanotrophs ( Stein and Klotz, 2011 ). This may in part be due to the need to detoxify NO that is produced during aerobic ammonia oxidation by reducing it to N 2 O ( Sutka et al., 2003 ). The expression of norB in Methylococcus capsulatus strain Bath increased 4.8-fold after treatment with 0.5 mM sodium nitroprusside, a NO releasing compound ( Campbell et al., 2011 ). It is possible that the NorB protein is involved in detoxification of NO during NH 3 oxidation in M. capsulatus strain Bath, since the genome lacks a dissimilatory nitrite reductase. More recently, it was demonstrated in M. fumariolicum strain SolV that transcription of norB was upregulated during O 2 limitation during chemostat growth ( Khadem et al., 2012a ); however, it is unknown whether M. fumariolicum strain SolV can consume NO 2 - or NO. The transcription of norB in M. denitrificans FJG1 increased 2.8-fold in response to NO 3 - and hypoxia ( Kits et al., 2015 ). While the genome of M. album strain BG8 encodes two copies of the norB gene, only one copy ( norB1 ) is followed by norC – the essential cytochrome c -containing subunit ( Mesa et al., 2002 ). Although some organisms like Cupriavidus necator possess two independent functional nitric oxide reductases ( Cramm et al., 1997 ), the present work illustrates that expression of only norB1 in M. album strain BG8 is responsive to NO 2 - treatment. Although the function of NorB may differ between M. album strain BG8 and M. capsulatus strain Bath, both bacteria show a similar transcriptional response of norB genes to NO 2 - ( Campbell et al., 2011 ). Transcript Abundance of pxmA Significantly Increased in Response to both NO 2 - and Hypoxia Genomes of some aerobic methanotrophs belonging to the phylum Gammaproteobacteria have been shown to encode a sequence divergent CuMMO protein complex, pXMO ( Tavormina et al., 2011 ). The function and substrate of the putative pXMO protein encoded by the pxm operon remains unknown. Previous studies on the pxm operon have shown that it is expressed at low levels during growth in Methylomonas sp. strain LW13 as well as in freshwater peat bog and creek sediment ( Tavormina et al., 2011 ). Metagenomic sequencing of the SIP-labeled active community in an oilsands tailings pond revealed that pxmA sequences were present in the active methanotroph community ( Saidi-Mehrabad et al., 2013 ). Analysis of the transcriptome of M. denitrificans FJG1 revealed that steady state mRNA levels of the pxmABC operon increased ∼10-fold in response to denitrifying conditions ( Kits et al., 2015 ). We now demonstrate that expression of pxmA in M. album strain BG8 is significantly increased in response to both NO 2 - and hypoxia. We did not observe any increase in the expression of pxmA in O 2 limited NMS-only cultures where denitrification was not occurring, suggesting that hypoxia alone is not sufficient to illicit an increase in the steady state mRNA levels. This study adds further support to the observation that expression of pxmA is responsive to denitrifying conditions. However, it must be noted that at 72 h in the NO 2 - amended media, absolute transcript abundance of pxmA (1 × 10 3 copies pxmA /1 × 10 9 copies 16s rRNA) was three orders of magnitude lower than absolute transcript abundance of pmoA (1 × 10 6 copies pxmA /1 × 10 9 copies 16s rRNA)."
} | 4,460 |
39869783 | PMC11833320 | pmc | 1,234 | {
"abstract": "Abstract Iron plays a pivotal role in regulating ocean primary productivity. Iron is supplied from diverse sources such as the atmosphere and the geosphere, and hence iron biogeochemical research has focused on identifying and quantifying such sources of “new” iron. However, the recycling of this new iron fuels up to 90% of the productivity in vast oceanic regions. Evidence points to the key role of microbes in mediating this recycling, referred to as the “ferrous wheel”, that remobilises iron initially supplied to ocean biota. In the iron-limited subantarctic waters of the Southern Ocean, iron uptake is dominated by microbes smaller than 2 μm and exhibits seasonal and depth-related variations. The microbial community within the <2 μm size fraction comprises heterotrophic bacteria and picophytoplankton, both competing for iron. Here, we dissect the demand component of the ferrous wheel by separately assessing iron uptake by heterotrophic bacteria and photoautotrophic picophytoplankton. To explore the seasonal and depth-related variability in iron uptake, the influence of light on iron uptake in both bacterial and phytoplankton communities was examined. We observed that picoeukaryote phytoplankton demonstrated iron uptake rates 10 times greater than those observed in bacteria when normalized to biomass. Light was shown to stimulate iron uptake by 8- to 16-fold in phytoplankton and by 4- to 8-fold in heterotrophic bacteria. These results highlight the unexpectedly significant role of picoeukaryotic phytoplankton in driving the speed of the ferrous wheel, with implications for iron recycling across diurnal cycles, different oceanic depths, and seasonally.",
"conclusion": "Conclusions Prior comparisons of the relative contributions of recycled and new Fe have focused on the Fe-ratio [ 7 , 9 ]. However, the rate of recycling over the daily light:dark cycle has never been factored into the relative contributions of new and regenerated Fe. Our findings show that rates of Fe uptake, and presumably then recycling, differ between picoeukaryotic phytoplankton and heterotrophic bacteria: picoeukaryotes are a small biogenic Fe pool that spins very quickly compared to the larger but slower turning Fe pool of heterotrophic bacteria. This finding – of two different ferrous cogs – has potential, but unknown, implications for the rate and possibly fates of recycled Fe [ 8 , 9 ]. Furthermore, our findings suggest that light can increase the rate at which Fe is shuttled through the ferrous wheel relative to that in the dark. The diel expression of microbial Fe transporter transcripts reinforces the idea of diurnal patterns in the supply term of the ferrous wheel [ 53 ]. In addition to the diurnal cycle, the influence of these light-mediated processes has wider implications for rates of Fe recycling with season. The trends reported in Fig. 2 capture both changes in ambient light over the season and the relationship between the depth of the euphotic zone. As such they place our summertime measurements (an upper bound for Fe uptake over the season) in a wider context and point to the need for additional metrics. Specifically, we need seasonal Fe uptake versus irradiance curves for heterotrophic bacteria and phytoplankton (normalized to their respective biomasses). To determine the relationship between Fe and C biogeochemistry, we also need complimentary measurements of NPP. Our summertime findings suggest that prior observations of relative constancy in Fe uptake with depth [ 13 ] are not applicable across the annual cycle. We estimated that irradiance in the mixed layer was 4.5- and 14-fold greater in the summer compared to the spring and fall, respectively ( Table S1 ). Is there a threshold in mean water column irradiance that triggers a transition in Fe uptake for the phytoplankton, and is the increase linked to the 10-fold increase in NPP? Furthermore, how are these changes in phytoplankton NPP and Fe uptake influenced by mixotrophy and how do they influence the daily and seasonal bacterial demands for Fe and C [ 15 , 54–56 ]? While our summertime findings of light-mediated increases in Fe uptake by phytoplankton are not surprising [ 18 , 40 ], our new observation of light-stimulated Fe uptake in heterotrophic bacteria raises questions about the underlying mechanism(s). Taken together our findings point to future research priorities that focus on how the Fe uptake components of the ferrous wheel function across the diel cycle. If as we propose that light can increase the rate at which Fe is shuttled through the ferrous wheel relative to that in the dark and does so differently for heterotrophic bacteria compared to phytoplankton, then there may be a fundamental shift in the balance of Fe uptake vs recycling over the diel cycle. This will be superimposed on cycles of particle production and destruction driven by other ecological processes [ 57 ]. Understanding this superimposition will require the intersection of ferrous wheel dynamics with the known sequence of processes within the microbial loop over the diurnal cycle [ 58 , 59 ].",
"introduction": "Introduction The trace element iron (Fe) plays a central role in the biogeochemical cycles of key elements such as carbon, oxygen, nitrogen, sulfur, and phosphorus through its regulation of key microbial physiological rates including photosynthesis and respiration [ 1 ]. Despite being one of the most abundant elements in the Earth’s crust, Fe is present at vanishingly low concentrations over much of the global ocean due to both its limited solubility and from being extremely particle reactive [ 2 ]. Such reactivity means that Fe is readily removed from the water column in nearshore waters [ 3 ]. Despite dissolved Fe being present at picomolar concentrations in much of the surface ocean, these concentrations are typically higher than those set by the solubility limit of inorganic Fe in seawater due to organic complexing agents, called ligands, which are often produced by microbes and enhance Fe solubility [ 2 ]. Ligands bind over 99% of dissolved Fe in seawater and largely control the bioavailability of Fe to microorganisms. The bioavailability of Fe is further affected by photochemical reactions that make Fe complexed to ligands more available for uptake [ 2 ]. The pivotal role of Fe in cellular physiology along with the low availability of dissolved Fe means that it is an element that is “heavily trafficked” by ocean biota [ 2 ]. Fe is supplied by wide-ranging mechanisms such as aerosol dust, ocean upwelling and melting icebergs [ 4 ]. Most research on the marine biogeochemistry of Fe since the 1990s has focused on quantifying these “new” sources of Fe for example hydrothermal inputs [ 5 ]. In contrast, the role of recycled Fe is understudied. Iron recycling is known to play a central role in the terrestrial biosphere where it is termed the “ferrous wheel” [ 6 ]. In the ocean, the ferrous wheel is comprised of the key microbes—bacteria, phytoplankton, viruses, and grazers—contributing to Fe demand and supply [ 4 ]. Only a few studies have quantified the relative contribution of new versus recycled Fe. The major contribution of recycled Fe results in a very low Fe-ratio (i.e. new Fe uptake/uptake of new and recycled Fe) [ 7 , 8 ]. Process studies show seasonality in the Fe-ratio: ~0.1 (i.e. 90% of Fe for biota comes from recycling) during summer in low Fe High Nitrate Low Chlorophyll (HNLC) subantarctic waters [ 7 ]; ~0.6 to 0.1 during a subtropical springtime bloom [ 9 , 10 ]. The patterns observed in these two short-term process studies were extended to estimate the relative contributions of new and recycled Fe to the Southern Ocean over the seasonal cycle [ 5 ]. Subsequent modelling has suggested that Fe recycling is central to driving annual primary productivity but there are currently few data available to improve model parameterizations [ 11 , 12 ]. Previous results from the subantarctic Southern Ocean Time Series (SOTS) site in the fall showed that cells <2.0 μm in diameter take up most of the Fe [ 13 ]. This size class comprises heterotrophs (bacteria) and photoautotrophs (phytoplankton) that compete for this limiting resource [ 14 , 15 ]. Therefore, we sought to answer the following question: how much Fe do heterotrophic bacteria take up compared to phytoplankton? Answering this question is important because the fate of Fe (recycled versus exported) and its efficiency in fuelling primary production depend on whether bacteria or phytoplankton take it up [ 16 ]. However, competition is just one aspect of this narrative, as bacteria also rely on labile dissolved organic carbon (DOC) released by phytoplankton as a carbon (C) source and microbes release Fe-binding ligands that can facilitate Fe uptake by both bacteria and phytoplankton [ 17–21 ]. The biogeochemical cycles of Fe and C are linked by the Fe:C ratio—as an elemental or, in our study, uptake ratio—in phytoplankton and heterotrophic bacteria. We have observed that Fe uptake is less sensitive to changes in underwater irradiance than inorganic C uptake is (i.e. net primary productivity, NPP) [ 13 , 22 ]. This suggests that either (i) phytoplankton Fe uptake is less sensitive to changes in irradiance than is NPP; or (ii) a process less sensitive to light, such as heterotrophic bacterial assimilation, is responsible for most of the Fe uptake. We thus asked how light affects Fe uptake in bacteria and phytoplankton. Possible light-mediated influences include photochemical reactions that make Fe complexed to ligands more available for uptake in the light, and/or the regulation of physiological processes that influence Fe uptake. In the latter case, light might affect Fe uptake directly by supplying reductant and/or energy produced photosynthetically for extracellular Fe reduction and uptake, or indirectly, such as through altering the supply of DOC to fuel the growth of heterotrophic bacteria ( Fig. 1 ). Figure 1 Conceptual illustration of the ferrous wheel size-fractionation experimental design. Treatments were unfiltered and pre-incubation <0.8 μm filtrations (Bact PRE ) incubated in the light and in the dark. Pre-incubation size fractionation (Bact PRE ) isolated heterotrophic bacteria from the ferrous wheel. Comparison with post-incubation size fractionation allowed us to compare the effects of isolating heterotrophic bacteria from the ferrous wheel and to compare the effects of light and dark treatments on Fe uptake by both phytoplankton and heterotrophic bacteria. Dashed circles represent the <0.8 μm (0.2–0.8 μm) fraction isolated from the unfiltered treatments during post-incubation size fractionation. Labelled boxes denote processes or groups affected by the treatments. Here, we dissect the spokes of the ferrous wheel associated with Fe demand by quantifying the uptake rates of heterotrophic bacteria and phytoplankton in the subantarctic Southern Ocean during summer and situate these findings within a seasonal context. To do so, we conducted bioassays in which the effects of light on Fe photochemistry and uptake physiology were studied by comparing light and dark incubations, and the effects of DOC supply and competition between phytoplankton and heterotrophic bacteria were examined by isolating bacteria from the larger members of the ferrous wheel by pre-incubation size fractionation ( Fig. 1 ).",
"discussion": "Discussion Small cells dominate Fe uptake and are therefore key players in the ferrous wheel ( Fig. 2 ). Our experiment isolated the heterotrophic bacteria from picophytoplankton that are similarly sized but primarily differ in trophic mode and Fe acquisition strategies [ 34–36 ]. Heterotrophic bacteria have significant Fe requirements, possibly exceeding those of phytoplankton [ 37–39 ]. It has been hypothesized that bacteria are superior competitors for nutrient acquisition based on their small size and high surface area:volume (SA:V) ratio [ 22 , 37 ]. In field studies, it has been further assumed that the majority of Fe uptake in the <2 μm fraction is by bacteria [ 13 , 16 ]. Our findings challenge this notion and provide insights into the drivers of Fe uptake by both picoeukaryotic phytoplankton and heterotrophic bacteria. Overall, in the ferrous wheel experiment phytoplankton took up ~79% of Fe, higher than the partitioning observed previously (e.g. 60–70% in the Northeast Pacific [ 14 , 37 ]; 50% at SOTS in fall [ 15 ]). Our most arresting result was that picoeukaryotes take up the majority of Fe. They accounted for 53% of phytoplankton Fe uptake, and 42% of total Fe uptake (phytoplankton + bacteria). In contrast, heterotrophic bacteria accounted for only 21% of total Fe uptake. Picoeukaryotes are comparable to heterotrophic bacteria in size, and hence in their SA:V ratios ( Table S4 ). But, when normalized to C biomass, the Fe uptake rates of the slightly larger picoeukaryotes were 10-fold greater than those of heterotrophic bacteria ( Fig. 4D ). Our finding that the Fe uptake rates of picoeukaryotes greatly exceed those of heterotrophic bacteria challenges the notion that bacteria require more Fe per unit of C biomass [ 16 ]. To provide context for the higher Fe uptake rates in picoeukaryotes compared to heterotrophic bacteria observed in our study, we calculated the surface area-normalized dissolved Fe uptake rate constant, k in-app /SA [ 40 ] ( Table S5 ). Picoeukaryotes had a k in-app /SA 6-fold higher than the global mean (of 560 single cell k in-app /SA values from 31 low-Fe stations) [ 41 ]. In marked contrast, the k in-app /SA of heterotrophic bacteria was 6-fold lower than the global mean. We calculated a similarly low k in-app /SA for Fe-limited heterotrophic bacteria at SOTS in the fall [ 15 ]. These calculations not only challenge the notion that heterotrophic bacteria are superior competitors for Fe acquisition but also reveal an unexpectedly effective Fe acquisition apparatus in picoeukaryotes. Small cells can dominate photosynthetic biomass and primary production in many marine ecosystems [ 42 ] including the Southern Ocean [ 43 ]. Who are the Southern Ocean picoeukaryotes? In short, we do not know. Their small size and lack of morphological features hinders traditional taxonomic identification [ 44 ]. Indeed, they are referred to as URGOs (Unidentified Round Green Objects) [ 44 ] or UNANs (Unidentified NANoflagellates) [ 45 ]. They are most likely haptophytes, prasinophytes, and chlorophytes based on HPLC pigment analyses [ 45 ]. Improved tools, such as more and better annotations of marine eukaryotic genomes, are needed to better characterize the taxonomy and ecological functions of these organisms. An open question arising from our study is whether the fate of picoeukaryotes within the ferrous wheel differs from that of heterotrophic bacteria. The most sophisticated biogeochemical models (e.g. PISCES-NEMO) represent diatoms, nanoflagellates, and heterotrophic bacteria but not this potentially important picoeukaryote group [ 12 , 46 ]. Light influences several processes that affect both Fe and C biogeochemistry ( Fig. 1 ). The simulated in situ depth-profile incubations revealed less change in Fe uptake than in NPP in the <2 μm fraction, particularly in the spring and fall ( Fig. 2 ). We assumed that this trend was due to heterotrophic bacteria that take up Fe but negligible amounts of inorganic C. What we observed was more nuanced. As expected, light had a greater influence on the NPP of photosynthetic phytoplankton (>0.8 μm) than on that of heterotrophic bacteria. However, what was unexpected was the major influence that light had on Fe uptake by both phytoplankton and heterotrophic bacteria; an ~12-fold stimulation of phytoplankton (>0.8 μm), but also a ~ 5-fold stimulation of the heterotrophic bacteria (0.2–0.8 μm; ~85% bacterial biovolume (F pop ), ~92% bacterial C biomass). Our study also shows that this light stimulation of Fe uptake occurs for heterotrophic bacteria in the absence of a “new” source of DOC (i.e. potentially a source of natural ligands) from phytoplankton ( Fig. 5 ) and in the presence of the photosynthetic electron transport inhibitor DCMU ( Fig. S7 ). A 15-fold light stimulation of Fe uptake from in situ ligands in a cyanobacteria-dominated subantarctic community has been reported previously [ 18 ], a result that has been observed elsewhere [ 40 , 47 ]. In contrast, only a 2-fold light stimulation of Fe uptake is evident when using photolabile “model” ligands [ 18 , 40 , 47 ]. These observations suggest that natural in situ ligands have unknown photochemical properties and interactions that are not captured well by model ligands with respect to stimulating Fe uptake. However, there are some compelling clues in the literature to explain these observations. Vibrioferrin, a bacterially produced Fe-binding ligand (i.e. a siderophore) is photochemically reactive and increases Fe uptake in both phytoplankton and heterotrophic bacteria but differentially (>20-fold vs 1.7-fold, respectively) [ 20 ]. This mutualistic relationship is hypothesized to be effective within the diffusive boundary layer (i.e. the phycosphere) of larger cells. The effectiveness of this bacterial–algal mutualism in picoeukaryotes is unclear because as the phycosphere decreases with cell size it becomes indistinguishable from surrounding seawater [ 48 ]. Alternatively, a pool of weak but abundant ligands such as polysaccharides (a constituent of the labile DOC pool) may facilitate Fe retention and uptake if its size increases in the light via increased exudation by phytoplankton [ 49 ]. In addition to photochemistry, light may stimulate microbial Fe uptake at the physiological level. Light induces a physiological response that increases Fe uptake in marine diatoms [ 50 ]. However, high-light grown diatoms retain their enhanced ability to take up Fe in the dark (for at least 4 h) [ 50 ], which would likely mitigate the differences in light vs dark Fe uptake by phytoplankton that we observed ( Fig. 5 ; Fig. S6 ). Photoheterotrophy is a widespread mode of bacterial metabolism, notably in the oligotrophic surface ocean, where microbes experience chronic nutrient limitation. One especially widespread form of photoheterotrophy is based on proteorhodopsin, which uses light to generate proton motive force that can drive ATP synthesis or nutrient uptake [ 51 ]. However, the role of proteorhodopsin in bacterial Fe metabolism is unclear [ 52 ]. To date, there are no reports on the physiological effect of light on bacterial Fe uptake. Cumulatively, these results shed light on how little is known about the environmental drivers of Fe uptake, the identity of Fe-binding ligands, and the effects of ligands on Fe bioavailability. There is no smoking gun; it is likely that a combination of mechanisms contributes to the light stimulation of Fe uptake that we observed. Pre-incubation size fractionation isolated heterotrophic bacteria from the ferrous wheel ( Fig. 1 ). Comparison with post-incubation size fractionation allowed us to potentially demarcate the effects of isolating heterotrophic bacteria from the ferrous wheel and the effect of light on Fe uptake by both phytoplankton and heterotrophic bacteria. We found that in the absence of phytoplankton, higher Fe uptake rates by bacteria were observed in the light after 24 hours incubation but not after 48 hours ( Fig. 5 ; Fig. S6 ). Isolating bacteria not only removed competition with phytoplankton for Fe but also shut off the immediate source of labile DOC (i.e. algal exudation), which may play dual roles as a C source and a pool of Fe-binding ligands. Our results suggest that altering both factors (DOC supply and Fe competition) was a zero-sum game under our experimental conditions. The transient increase we observed could be due to the presence of a residual labile DOC pool and/or a reduction in competition for Fe. After 48 hours the light-modulated advantage of bacteria isolated from the ferrous wheel disappeared, and there was no difference in Fe uptake rates by bacteria in the presence or absence of phytoplankton. Our findings point to tight coupling across both Fe and C and different trophic levels in the microbially mediated ferrous wheel."
} | 5,064 |
39824840 | PMC11742687 | pmc | 1,235 | {
"abstract": "Biological neural systems seamlessly integrate perception and action, a feat not efficiently replicated in current physically separated designs of neural-imitating electronics. This segregation hinders coordination and functionality within the neuromorphic system. Here, we present a flexible device tailored for neuromorphic computation and muscle actuation. Each individual device component emulates essential synaptic functions for neural computing, while the collective ensemble replicates muscle actuation in response to efferent neuromuscular commands. These properties stem from densely-packed, hydrophilic nanometer-sized channels, and the erection of a high-entropy, intricately silver nanowires to capture and store of hydrated cations. Leveraging the remarkable deformation effect, we demonstrate hazard detection-avoidance robot, and multidimensional integration for arbitrary programmed shapes like 360° panoramic information capture and soft-bodied biological deformations wherein localized responses to stimuli are harmoniously integrated to achieve arbitrary coordinated motion. These results provide a significant avenue for the development of future flexible electronics and bio-inspired systems.",
"introduction": "Introduction The intricate organisation and functions, e.g. sensation and movement, observed in biological entities currently surpass the capabilities of electronic and machinery devices in various aspects 1 – 4 . Understanding and harnessing the 1 → N differentiation-integration logic inherent in biological functionalities, and drawing inspiration from this, purposefully expanding and integrating of numerous simple, identical subunits result in the formation of a collectively-characterised entity, presents an opportunity to introduce innovative concepts into the design of flexible electronic devices and systems. A neuromuscular junction (NMJ) is a type of synapse that connects the end of a motor nerve to a skeletal muscle to control its movements. Thousands of NMJs and muscle fibres coordinate and integrate with each other to enable an organism to accomplish a variety of movements and deformations 5 . The intricate control mechanism by using a variety of NMJ combinations could be referential to the development of soft robots with natural motion and environmental adaptability, endowing these robots with awareness of surrounding changes, and intelligently responsive motions. To date, materials and devices that emulates singular synapse 6 – 10 and neuromorphic system 4 , 11 – 15 , and artificial muscles driven by heat 16 , electricity 17 , light 18 , and other means 19 , 20 have been developed. An artificial neuromuscular system composed of artificial synapses and artificial muscles is attracting tremendous attention (Supplementary Table 1 ). However, due to differences in materials and device structures, these components have remained physically separated in previous reports. This separation increases system complexity, requires additional manufacturing process and reduced the reliability and efficiency of the system. A single materials and device system is desired that combines and integrates synaptic information processing with effectors, which is conducive to the further development and integration of neuromorphic systems. Herein, we demonstrate a device that integrates the functions of neural computing and mechanical actuation in the “differentiated” and “integrated” forms, respectively. The working mechanism is realised by well-confined sub-nanoscale channels in membranes of perfluorosulfonic acid ionomer (PFSA) modified by the addition of polyvinyl alcohol (PVA). This unique architecture permits interactions with differently-sized cations at the molecular level. The placement of an interwoven layer of silver nanowires (Ag-NWs) at one terminus of the nanochannels facilitates the capture and storage of hydrated cations to achieve threshold opening, sensitisation, and desensitisation, which are essential characteristics of sensory receptors. The device exhibits exceptional flexibility and demonstrates precise deformation actuation. This property facilitates the emulation of snail stalk eyes, enabling 360° panoramic information capture. Moreover, we have intricately integrated each micro-unit through patterned design, achieving cross-dimensional control of complex deformations while simultaneously performing neural computing functions. By integrating neural sensing and actuation functionalities in a unified manner, the devices enhance the integration level and functionality of soft electronics, thereby unveiling promising application prospects in domains such as edge intelligence devices, neuromorphic soft robots, and bioinspired electronic systems.",
"discussion": "Discussion We have presented a synapse-motor coupler device (SMCD), which is a new type of electronic device in which the basic units perform neuromorphic computation, whereas an ensemble of basic units works as an artificial muscle that responds to neuromuscular actuation. The synapse-muscle coupler device can mimic a natural muscle that is innervated by arbitrarily-distributed neuromuscular junctions. These properties originate from special nanoscale ion-transport channels that form in PVA-modified PFSA membrane. An Ag-nanowire forest at one end of these nanoporous structures captured ions, and thus enabled a series of important neuroplasticity behaviours. A functional electronic system that uses this device successfully emulates the threshold, relaxation, and sensitisation of nociceptors. As a proof-of-concept, it successfully replicated the closing movement of the Venus flytrap and achieved multidimensional motion combinations of arbitrary shapes and hazard detection-avoidance robot. SMCD possesses integrated functions of artificial synapses and effectors, providing a unique and simplified strategy for the development of the next generation of neuromorphic viable electronic devices and soft robots, with illustrative examples including soft-bodied bionic hands, crawling agents for pipelines, and similar applications, humanoid neural reflex arcs, and neural prostheses."
} | 1,530 |
35857470 | PMC9258956 | pmc | 1,236 | {
"abstract": "Bacteria commonly form aggregates in a range of coral species [termed coral-associated microbial aggregates (CAMAs)], although these structures remain poorly characterized despite extensive efforts studying the coral microbiome. Here, we comprehensively characterize CAMAs associated with Stylophora pistillata and quantify their cell abundance. Our analysis reveals that multiple Endozoicomonas phylotypes coexist inside a single CAMA. Nanoscale secondary ion mass spectrometry imaging revealed that the Endozoicomonas cells were enriched with phosphorus, with the elemental compositions of CAMAs different from coral tissues and endosymbiotic Symbiodiniaceae, highlighting a role in sequestering and cycling phosphate between coral holobiont partners. Consensus metagenome-assembled genomes of the two dominant Endozoicomonas phylotypes confirmed their metabolic potential for polyphosphate accumulation along with genomic signatures including type VI secretion systems allowing host association. Our findings provide unprecedented insights into Endozoicomonas -dominated CAMAs and the first direct physiological and genomic linked evidence of their biological role in the coral holobiont.",
"introduction": "INTRODUCTION Endosymbiotic microbial aggregates are a common feature within tissues of many animals and often demonstrate tight mutualistically beneficial symbiotic roles. These structures (named bacteriocytes, bacteriomes, trophosomes, “bacterial aggregates,” “cyst-like aggregates,” or “intracellular colonies of bacteria”) have been widely reported in terrestrial eukaryotes such as plants [e.g., ( 1 , 2 )], insects [e.g., ( 3 – 5 )], and freshwater single-celled eukaryotes ( 6 ) in addition to many marine phyla such as gutless oligochaetes ( 7 ), deep-sea tubeworms ( 8 ), sponges ( 9 ), sea anemones ( 10 ), ascidians ( 11 ), mollusks ( 12 , 13 ), and corals ( 14 ). Many of these microbial symbionts live in specific compartments of their host, maintaining an obligate relationship that facilitates holobiont fitness through metabolic interactions, nutrient exchange, and defense mechanisms ( 1 – 7 , 9 , 15 , 16 ). However, for many marine invertebrates such as corals, sea anemones, ascidians, and mollusks, the identity and function of the bacteria within these aggregates are currently poorly understood despite being prevalent. Numerous 16 S ribosomal RNA (rRNA) amplicon-based studies have reported Endozoicomonas (Gammaproteobacteria)–affiliated taxa as the dominant members of the coral microbiome, with diverse phylotypes associated with individual species and across taxonomically dispersed hosts ( 11 , 17 – 20 ). Endozoicomonas strains have been isolated from corals, and their metabolic potential was inferred from derived genomic studies on the isolates and metagenomes assembled from culture-independent studies, suggesting integrated metabolic links facilitating nutrient acquisition and provision and therefore potentially important roles in host health ( 14 , 21 – 24 ). Some Endozoicomonas have been reported to locate within aggregates of the coral tissues ( 14 ) and other marine invertebrates ( 10 , 11 , 13 ), although other bacterial taxa including Rickettsiales (Alphaproteobacteria) and Chlamydia (Chlamydiae) have also been reported to form aggregations in corals ( 23 , 25 ) and mollusks ( 26 , 27 ); hence, these coral-associated microbial aggregates (CAMAs) could be constructed with polymorphous types of bacteria ( 28 , 29 ). However, the spatial localization of these symbionts is still largely overlooked, particularly with respect to individual phylotypes and populations across bacterial lineages. In addition, the linkages between bacterial physiology and metabolic roles within the coral holobiont are unknown. Here, we provide an unprecedented characterization of CAMAs in the coral Stylophora pistillata , visualizing the in situ distribution of CAMAs in the three-dimensional (3D) space of the coral polyp and estimating the numbers of bacteria in respective CAMAs. Furthermore, we identified and characterized bacterial phylotypes within CAMAs from coral samples from two geographic locations and confirmed their phylogenetic relationships and putative functions through metagenomic approaches. Last, we mapped elemental distribution within the CAMAs and elucidated the putative ecological functions of CAMAs in the coral holobiont (see more details in the workflow of our research in fig. S1).",
"discussion": "DISCUSSION This study provides an unprecedented, high-resolution characterization of CAMAs within the polyps of S. pistillata , revealing an average bacterial density of 0.77 cells/μm 3 within the structures and an estimated 0.67 × 10 6 bacterial cells housed within the aggregates in a single polyp. CAMAs were dominated by Endozoicomonas , although individual CAMAs have multiple phylotypes and interrogation of recovered MAGs, linked with cell-level elemental mapping, and indicated a capability of sequestering phosphate and potentially cycling within the coral holobiont. Phosphorus enrichment was higher in CAMAs than in the surrounding coral tissues and Symbiodiniaceae and, at the cell level, could be observed in the intracellular spaces of Endozoicomonas cells within the CAMAs. Coral tissues fluctuate in the levels of oxygen with oxygenic photosynthesis driven by the Symbiodiniaceae creating hyperoxic conditions during the day and night time respiration, resulting in low or even anaerobic conditions ( 33 ). Consistent with the phosphorus enrichments at the single-cell level by NanoSIMS imaging, the reconstructed MAGs identified the capability of Endozoicomonas to synthesize polyphosphate via the putative polyphosphate kinase ( PPK ) protein (with aerobic conditions during daytime) and release the phosphate via the polyphosphatase enzyme PPX (with low oxygen or anaerobic conditions during the night), with associated transporters (high- and low-affinity phosphate transporters) corresponding to phosphate uptake and release. Phosphate enrichment can reduce Symbiodiniaceae photosynthesis, which, in turn, lowers calcification processes and subsequent coral growth [e.g., ( 34 , 35 )]. Sequestering of polyphosphate within the CAMAs may therefore represent a buffering mechanism to modulate phosphate internally within the coral tissues, facilitating efficient photosynthesis by the Symbiodiniaceae endosymbionts and optimizing coral growth during daytime. Previous studies have suggested that the frequency of CAMAs in coral tissues could be influenced by dissolved nutrients ( 29 ), and therefore, high P levels in surrounding seawater could facilitate the development of CAMAs in the coral tissues. A role for bacteria in phosphate cycling within coral reef marine invertebrate hosts might be common. Zhang et al. ( 36 ) detected a group of uncharacterized cyanobacterial symbionts in three sponges that enabled polyphosphate synthesis and suggested that those bacteria play a role in phosphorus sequestration and recycling in the environment. An intriguing genomic feature of the two recovered MAGs in this study was the near-complete set of T6SS structure genes, in contrast to other genomes of Endozoicomonas retrieved from corals that did not carry either any or only one of the T6SS-like genes, although they are commonly detected in Endozoicomonas species from noncoral hosts. Bacterial T6SS is a common secretion system, detected in approximately 25% of all Gram-negative bacteria ( 37 ) and important for virulence traits in pathogenesis ( 38 ), bacterial communication ( 39 ), cell aggregate formations ( 40 ), and biofilm formation ( 41 ). Many plant pathogenic bacteria process the T6SS, producing the virulent effectors to eliminate their antagonists and suppress plant defenses, thereby allowing bacterial colonization in host tissues ( 42 ). This process resembles the symbiosis establishment of Vibrio fischeri in the marine bobtail squid, in which T6SS is used to eradicate their competitors ( 43 ). In the plant, recent evidence demonstrated a beneficial role for the rhizobial T6SS, promoting the formation of root nodules and subsequent plant growth ( 44 ). Given these findings in other hosts, the T6SS system in Endozoicomonas may facilitate and maintain CAMA formation in the host coral. The predicted T6SS proteins showed phylogenetic relatedness to members within the families Oceanospirillaceae, Vibrionaceae, and several families from the order Alteromonadales from marine ecosystems, which may have been acquired horizontally. A recent comparative genomic analysis of culturable coral-associated bacteria has revealed that the Vibrionales genomes have a greater number of T6SS-related domains than other bacterial taxa, and pathogenic Vibrio strains in addition have more than nonpathogenic Vibrio strains ( 20 ). The specific roles of T6SS-related domains in the Endozoicomonas are not known, although potentially important in establishing their dominance in CAMAs. Whether the Endozoicomonas are parasitic, commensal, or mutualistic is also still to be resolved. CAMAs derived from two sampling locations (Okinawa, Japan and Kenting, Taiwan) were characterized via 16 S rRNA gene sequencing of bulk tissues and laser dissection excision of individual CAMAs ( n = 67), with Okinawa samples particularly dominated by one Endozoicomonas phylotype (OTU 2). Kenting samples displayed higher dispersion in Endozoicomonas phylotypes, and although OTU 2 was still the most dominant, three other Endozoicomonas OTUs were characterized, although OTUs 5 and 18 were never retrieved with the OTU 16 phylotype across any of the samples, which may indicate competition exclusion. Closely related organisms that occupy overlapping niches have been documented to compete for resources ( 45 ), which can allow for symbiotic bacteria to monopolize a host niche ( 46 , 47 ). However, why a single phylotype (OTU 2) is found in CAMAs of samples from Okinawa while multiple related Endozoicomonas phylotypes coexist in Kenting samples is still to be elucidated. The Endozoicomonas genus displays high taxonomic diversity being a host generalist, although specific subclades have been identified across coral taxa ( 48 ). Previous studies have also identified that S. pistillata harbors geographically distinct genotypes of Endozoicomonas ( 14 ), and taxonomic comparisons from our study demonstrated phylotypes detected here clustered into this S. pistillata –specific subclade. These sequences also shared a basal clade with P. damicornis –associated Endozoicomonas , which along with S. pistillata belong in the Pocilloporidae family and are commonly recognized as also having prevalent CAMAs within their tissues ( 14 , 25 , 49 ). These close host-bacterial phylogenetic relationships provide further evidence for patterns of phylosymbiosis across coral species and the potential for cophylogeny ( 48 , 50 , 51 ). Endosymbiotic microbial aggregates are critical to the biological success of many higher organisms. CAMAs have been widely documented, although understanding their distribution in coral polyps, geographic specificity, genetic divergence, and functions that underpin coral health has remained elusive. This study provides unprecedented insights into these tissue structures and identifies dominant Endozoicomonas phylotypes and their genomic and metabolic signatures that are integrated into the coral holobiont."
} | 2,870 |
29129968 | null | s2 | 1,237 | {
"abstract": "Caribbean coral reefs are declining due to a mosaic of local and global stresses, including climate change-induced thermal stress. Species and assemblage responses differ due to factors that are not easily identifiable or quantifiable. We calculated a novel species-specific metric of coral bleaching response, taxon-α and -β, which relates the response of a species to that of its assemblages for 16 species over 18 assemblages. By contextualizing species responses within the response of their assemblages, the effects of environmental factors are removed and intrinsic differences among taxa are revealed. Most corals experience either a saturation response, overly-sensitive to weak stress (α > 0) but under-responsive compared to assemblage bleaching (β < 1), or a threshold response, insensitive to weak stress (α < 0) but over-responsive compared to assemblage bleaching (β > 1). This metric may help reveal key factors of bleaching susceptibility and identify species as targets for conservation."
} | 251 |
21594298 | null | s2 | 1,240 | {
"abstract": "The spatially controlled positioning of functional materials by self-assembly is one of the fundamental visions of nanotechnology. Major steps towards this goal have been achieved using DNA as a programmable building block. This tutorial review will focus on one of the most promising methods: DNA origami. The basic design principles, organization of a variety of functional materials and recent implementation of DNA robotics are discussed together with future challenges and opportunities."
} | 123 |
38862519 | PMC11166942 | pmc | 1,241 | {
"abstract": "It has been previously shown that devices based on microbial biofilms can generate hydrovoltaic energy from water evaporation. However, the potential of hydrovoltaic energy as an energy source for microbial growth has remained unexplored. Here, we show that the electroautotrophic bacterium Rhodopseudomonas palustris can directly utilize evaporation-induced hydrovoltaic electrons for growth within biofilms through extracellular electron uptake, with a strong reliance on carbon fixation coupled with nitrate reduction. We obtained similar results with two other electroautotrophic bacterial species. Although the energy conversion efficiency for microbial growth based on hydrovoltaic energy is low compared to other processes such as photosynthesis, we hypothesize that hydrovoltaic energy may potentially contribute to microbial survival and growth in energy-limited environments, given the ubiquity of microbial biofilms and water evaporation conditions.",
"introduction": "Introduction Phototrophy and chemotrophy represent two widely recognized forms of microbial metabolism 1 – 3 . Phototrophs, including algae and cyanobacteria, utilize sunlight to convert carbon dioxide and water into energy-rich organic compounds through photosynthesis 4 – 6 . In contrast, chemotrophs harness their energy by breaking down various organic and inorganic compounds via enzymatic reactions 2 , 7 . Certain microorganisms possess dual metabolic capabilities, manifesting phototrophic and chemotrophic modes of metabolism 8 , 9 . This metabolic versatility grants them the adaptability to thrive in diverse environments, contributing significantly to ecological diversity and resilience. However, microbial communities inhabiting extreme environments, characterized by high temperatures, high salinity, extreme acidity or alkalinity, anaerobic conditions, extreme aridity, and even the early Earth, encounter challenges in accessing conventional energy sources due to the lack of photosynthetic structures and limited availability of electron donors 7 , 10 . Consequently, there is an increasing interest in exploring alternative energy sources capable of sustaining microbial life within these challenging environments. Electroactive microorganisms (EAMs) have attracted significant interest for their ability to directly exchange electrons with solid materials in the past few decades. This includes both donating to and uptaking electrons from solid materials. Various microorganisms, such as bacteria, eukaryotes, and archaea, have been extensively studied for their electron exchange capabilities 11 – 14 . The discovery of EAMs has unveiled new pathways in microbial metabolism and energy acquisition, driven by their unique ability to uptake extracellular electrons. For instance, Lu et al. 1 demonstrated that non-phototrophic chemoautotrophic and heterotrophic EAMs like Alcaligenes faecalis and Acidithiobacillus ferrooxidans could directly uptake extracellular electrons released from the photocatalysis of metal sulfides and oxides to support their growth 1 . In addition, Yang et al. 15 reported that EAMs such as the nonphotosynthetic CO 2 -reducing bacterium Moorella thermoacetica can uptake extracellular electrons from cadmium sulfide nanoparticles to sustain cellular metabolism under light exposure 15 . More recently, a discovery by Yamamoto et al. 16 elucidated the ability of electroautotrophic microorganisms to thrive on extracellular geoelectrical electrons in deep-sea hydrothermal environments 16 . These findings underscore the potential of naturally occurring extracellular electrons within microbial surroundings as a viable alternative energy source for cellular growth and metabolism. This offers a compelling supplement to the reliance on light or redox reactions between chemicals. Water evaporation is a widespread and natural physical process that occurs within the global water cycle 5 , 17 . It amounts to approximately 5.2 × 10 11 m 3 /yr, which is 57 times greater than the global water supply available to mankind 18 , 19 . The process involves heat transport and dynamic mass phenomena, constituting a significant energy flow 20 . In this context, a revolutionary hydrovoltaic energy generation process was proposed to harness the interaction between water molecules and hydrophilic groups of artificial materials during the natural process of water evaporation, which converted ambient heat energy into sustainable hydrovoltaic electrical energy 21 – 23 . Water evaporation through the utilization of hydrophilic materials, such as carbon, Al 2 O 3 , and silicon nanowires, has shown potential for generating sustainable hydrovoltaic energy 22 , 24 , 25 . Recent advancements have developed this concept by investigating the hydrovoltaic process of natural biomaterials 26 , 27 , aiming to explore the presence of this distinct process within natural environments. Remarkably, it was observed that hydrovoltaic energy generated from water evaporation through microbial biofilms could achieve voltages of ~ 0.4 V with currents around 5 μA 28 – 30 . This electrical energy output surpasses the minimum energy requirements for microbial growth, generally around 0.1 V and 0.02 μA 1 , 8 , 11 . These revelations open new avenues for exploring the potential interactions between ubiquitous water evaporation and microorganisms, particularly in the realm of energy conversion and metabolism. Yet, the feasibility of directly harnessing water evaporation-induced hydrovoltaic electrons (WE-HE) within microbial biofilms as a novel energy source to sustain and promote microbial growth remains a frontier to be fully explored and delineated. Here, we present experimental evidence for a previously unreported microbial metabolism process. This process involves the generation of hydrovoltaic electrons through natural water evaporation in a biofilm, which not only supports the survival of microorganisms but also enables their growth through extracellular electron uptake. To prove this concept, we selected a typical electroautotrophy strain 8 , Rhodopseudomonas palustris , to construct the hydrovoltaic biofilm. In addition, we examined other microorganisms including electroautotrophic and typical heterotrophic EAM strains, as well as non-EAM strains, for comparative purposes. The growth pattern of R. palustris was dependent on the intensity of water evaporation, correlating with sustainable carbon (CO 2 ) fixation and nitrate (NO 3 − ) reduction. Stable-isotope probing and transcriptomic analyses revealed that CO 2 fixation into cellular biomass relied on the WE-HE. Our calculations about the efficiency of water evaporation-to-cellular biomass conversion indicated that the WE-HE pathway was less efficient than traditional photosynthesis, but it represents a potential source of energy due to the widespread occurrence of water evaporation and its substantial energy potential.",
"discussion": "Discussion Water evaporation is a widespread occurrence on Earth, occurring on various surfaces such as rivers, lakes, wetlands, soils, oceans, and even plant leaves 17 , 53 . Microbial biofilms, a form of collective life with emergent properties and a much higher level of organization than single microbial cells, dominate various Earth habitats, accounting for 40~80% of bacterial and archaeal cells (1.2 × 10 30 cells) 54 , 55 . These biofilms inherently align with water evaporation 56 , 57 . There is speculation about the possibility of the widespread occurrence of hydrovoltaic energy generation through water evaporation from biofilms. It was found that the generated electrons effectively supported biomass synthesis in microorganisms with a considerable energy conversion efficiency. Over 50 days, 30.99 kJ of ambient heat energy (∆E w ) was absorbed through water evaporation, generating 0.22 J of electric energy (termed as hydrovoltaic energy, ∆E e ). This corresponds to an energy conversion efficiency of 7.1 × 10 −3 ‰ from ambient heat energy to hydrovoltaic energy (Fig. 5g and Supplementary Table 1 ). Impressively, 96% of this hydrovoltaic energy could be further converted into chemical energy (∆E b ) for biosynthesis in microorganisms (Fig. 5g and Supplementary Table 2 ). However, by considering the energy flux of each sub-process, the overall energy conversion efficiency for microbial growth was found to be 6.7 × 10 −3 ‰. It is worth noting that the overall energy conversion efficiency was primarily limited by the process of converting ambient heat energy to hydrovoltaic energy through water evaporation. This particular step relies heavily on the utilization of the ambient heat energy for the phase transition of water from a liquid to a gas state 17 . Although natural water evaporation for microbial growth displays lower conversion efficiency compared to the other natural process such as natural photocatalysis based on light-induced and mineral-mediated pathways (13 × 10 −2 ‰) and natural photosynthesis (1%) 1 , 5 , 15 , it remains a plentiful and widely available energy source due to the prevalence of microbial biofilms and water evaporation conditions. This research demonstrates that hydrovoltaic electrons can function as an energy source for driving microbial growth through natural water evaporation, presenting a novel mechanism. Although some initial mechanisms have been proposed to elucidate the origins of electrons and protons in biofilms through water evaporation, further research is necessary to thoroughly investigate these sources in future studies. Particularly, in laboratory conditions, artificial conductive metal wires are utilized to establish an external electrical circuit that aids in the movement of electrons from the lower to the upper layers of the biofilm. Interestingly, natural environments host microbial nanowires and conductive minerals present, suggesting the potential for naturally occurring self-contained electrical pathways within biofilms 58 , 59 , enhancing electron transfer in the biofilm. Hydrovoltaic electrons are continuously delivered to EAMs for intracellular carbon fixation and the reduction of compounds such as NO 3 − , and methyl orange. Microorganisms are abundantly found in various ancient environments 60 . It is speculated that, during the early stages of anoxic Earth, hydrovoltaic electrons may have served as an energy source, sustaining microbial growth. This hypothesis provides a new means of interpreting the evolution and early development of life. In modern environments, microorganisms have been also discovered in extreme settings, ranging from deserts to subterranean rivers 10 , 61 . These locales often lack organic matter, typically produced by chemolithotrophs. However, the ubiquitous process of water evaporation at the water-solid interface generates hydrovoltaic electrons. This phenomenon may directly or indirectly facilitate microbial metabolism and support the growth of other microorganisms, such as heterotrophs and chemoautotrophs. This study advances our understanding of the potential impact of water evaporation on biological processes and energy cycling in the natural world. In conclusion, this study investigated the utilization of hydrovoltaic electrons generated through water evaporation in microbial biofilms as an energy source for microbial growth. The research primarily focused on R. palustris as a model organism and examined its growth patterns and metabolic processes in the presence of the WE-HE. The results demonstrated that the WE-HE effectively supported the cellular growth of R. palustris by coupling the CO 2 fixation and nitrate reduction processes. Furthermore, the study explored the potential of other microorganisms to utilize hydrovoltaic energy through extracellular electron uptake, noting that other EAMs were able to harness hydrovoltaic electrons for growth and/or metabolism. Overall, the findings highlight the prospective role of hydrovoltaic energy as an alternative and sustainable energy source within microbial ecosystems, unveiling a novel mechanism elucidating how microbial growth and energy cycles operate within natural environments."
} | 3,035 |
38284057 | PMC10809701 | pmc | 1,242 | {
"abstract": "The demand for clean-energy\ncollection has gradually increased\nin recent years, making triboelectric nanogenerators a promising research\nfield, because of their advantages in convenient manufacturing, diversified\nmaterials, and diverse synthesis and modification possibilities. However,\nrecent studies indicate that charge decay, a major limiting factor\nin the triboelectric output, prevents the induced charge from combining\nwith the bottom electrode, leading to charge loss. The use of charge-trapping\nsites to retain the induced charge generated during the friction process\nis an important solution in the field of triboelectric nanogenerator\nresearch. This study proposes the use of an elastic ink with macroscopic\nmagnetism as trapping sites by coating the ink as dots between the\npolytetrafluoroethylene (PTFE) dielectric layer and the electrode\nlayer. Nickel particles in the magnetic ink are doped into the system\nas microcapacitors, which prevent the combination of the friction\nlayer and induced charges on the back electrode. Because the nickel\nitself can be used as a charge-potential trap to capture the charge\nintroduced by the charge-injection process, the charge can be maintained\nfor a long time and achieve a long-term high-output state. The output\nvoltage was more than 6 times that of the reference group without\nthe magnetic-ink coating after 3 h. The results provide a reference\ndirection for research on preventing charge decay and trapping charges\nin triboelectric nanogenerators.",
"conclusion": "Conclusions In summary, elastomer ink droplets containing magnetic nickel were\ncoated on the back of the dielectric layer of PTFE and the nylon films.\nThe magnetic substances in the elastomer were used as trapping sites\nto maintain the surface charge caused by charge injection and to prevent\nthe combination of induced and friction layer charges. Thus, the maximum\nsurface-charge density was preserved for a longer period and the output\nvoltage increased. The number of magnetic particles and the N-PM,\nNM-P, and NM-PM structures were compared as variables. After the charge-injection\nprocess, compared with the basic control group N-P, the N-PM group\nexhibited a higher output voltage (260 V) after 4000 cycles of friction\nand showed no downward trend, maintaining 59% (259 to 153 V) of the\nvoltage just after injection and even after 3 h of rest, which was\nmore than 5.5 times that of the reference group (28 V). The results\nprove that the particles in magnetic ink provide trapping properties\nand effectively maintain the high charge level caused by charge injection\nover a long time. The TENG was used as a general sensor to detect\ndifferent movement behaviors of the human body, thereby proving its\npractical-application significance.",
"introduction": "Introduction Owing to the increasingly significant\ngreenhouse effect and various\nenvironmental problems caused by traditional fossil-fuel energy sources,\nthe collection of clean energy has received increasing attention. 1 − 3 As a solution for clean energy, triboelectricity has gradually become\na promising research field for converting electricity using the charged\nelectrostatic-induction coupling effect. Owing to their simple structure,\nflexible and diverse materials, and low cost, triboelectric nanogenerators\n(TENG) are widely used in biosensing, artificial intelligence, and\npower supply for small equipment. 4 , 5 The surface-charge\ndensity is positively correlated with the output of a TENG and the\ncharge decay induced by the combination process directly leads to\na lower output efficiency, thereby limiting its application. 6 To improve the charge density, recent studies\nhave proposed doping particles with a high dielectric constant, increasing\nthe frictional contact area, and charge injection. 7 − 10 Charge injection can significantly\nincrease the charge density in a short time; however, the injected\ncharge is still lost by being combined with the charge of the back\nelectrode. 11 The charge injected onto the\nsurface is naturally lost to the air because of the lack of trapping\nsites. 12 Previous studies have maintained\ncharge by increasing the number of trapping sites through chemical\nmodification, adding complex internal structures to increase the trapping\narea, and doping particles with trapping properties in the friction\nlayer. 13 , 14 The printing process and the all-printed\nTENG have also been developed to improve the output performance. 15 However, these methods involve complex manufacturing\nprocesses. Thus, developing methods to easily and quickly enhance\nthe trapping properties of TENGs is necessary to maintain a high output\nperformance for a long time. The introduction of magnetic substances\nsuch as nickel into the\nmanufacturing process of the friction layer of TENGs has been reported\nin the literature, 16 , 17 while the hysteretic behavior\nof contact force response in triboelectric nanogenerator has been\nreported as an earliest related study. 18 However, the previous magnetic body was used as a macroauxiliary\nbody to drive the TENG to complete its friction motion. Alternatively,\nit could be regarded as a large capacitor with multiple microcapacitors\ncombined in parallel and series when the magnetic-metal material is\nintroduced into the friction layer in the form of fine particles using\nits own high capacitance. This enhances the capacitance of the TENG,\nthereby increasing the maximum charge it can hold and improving the\noutput performance. 19 No previous reports\nhave investigated magnetic substances as charge-trapping sites and\ntheir insertion between the friction layer and the electrode as a\ntrapping layer. Thus, developing a simple magnetic-trapping layer\nfor TENG devices is necessary. The introduction of the trapping properties\nof the TENG device should be researched by adding a magnetic substance\nto maintain a high equilibrium surface charge for a long time. In this study, magnetic ink was used as the trapping site, and\nthe simple method of adding a trapping layer between the friction\nlayer and the electrode was developed to maintain a high-output state\nfor a long time. Elastomer ink droplets containing magnetic nickel\nwere coated on the back of the dielectric layer of polytetrafluoroethylene\n(PTFE) and nylon films. After drying, half spheres (approximately\n0.5 × 0.5 mm) were formed, which served as trapping sites for\npreserving the surface charge introduced by the charge-injection process.\nThe number of spheres was 4 × 4, 6 × 6, and 8 × 8.\nFour structures—nylon–PTFE (N-P) (reference group),\nnylon and magnetic particle–PTFE (NM-P), nylon–PTFE\nand magnetic particle (N-PM), and nylon, magnetic particle–PTFE\nand magnetic particle (NM-PM)—were measured to investigate\nthe differences in the trapping properties of positive/negative materials.\nAfter the charge-injection process, the N-PM group showed the highest\noutput voltage and trapping properties compared with the basic control\ngroup, N-P. Even after 3 h, it maintained 58% of the output voltage\nafter injection, indicating good trapping performance. Finally, the\nTENG was used as a general sensor to detect human-movement behavior,\nwhich demonstrated the application significance of the display. In\nthis study, a simple and fast method is proposed to enhance the properties\nof TENGs, which provides a feasible reference for the development\nof TENG devices that maintain high-performance output for a long time.",
"discussion": "Results and Discussion Nylon11, obtained by the casting method, was used as the positive\ntriboelectric material, and commercially available PTFE tape was used\nas the negative triboelectric material to assemble the TENG. Magnetic\nink purchased from the MAGRON company was directly deposited on the\nsurfaces of the nylon and PTFE films using a dropper. The dropped\nviscoelastic polymer ink contained metallic nickel particles. As a\nviscoelastic polymer falls naturally under gravity, flat droplets\nform on the film surface. After the polymer was dried at high temperatures,\nthe fabricated electrodes were affixed to the back of the film and\nassembled. First, a variable design was used for the number of magnetic\ndroplets as follows: 4 (horizontal) × 4 (vertical), 6 ×\n6, and 8 × 8 dots. This was used to determine the influence of\nthe number of magnets on the output voltage and trapping performance\nafter charge injection (as shown in Figures S2 and S3 , the pictures of the push–pull gauge and injection\ndevice are attached). To investigate the difference in the trapping\nproperties of the\npositive and negative materials after the addition of magnetic particles,\nthe output voltages of the N-P, NM-P, N-PM, and NM-PM structures were\nmeasured. An overview of the experiment is shown in Figure 1 ( Figure 1 c was taken by one of the authors). Figure 1 a shows a conceptual\ndiagram\nof the TENG structure, where the magnet wrapped with a viscoelastic\npolymer forms a flat droplet shape. This caused a certain height difference\nto the surface of the friction layer, which can be regarded as an\nincrease in roughness. 18 In the charge-injection\nprocess shown in Figure 1 b, a high voltage was applied to the tip of a needle and a strong\nelectric field was generated when the back electrode was grounded.\nThe molecules in the air are ionized, which acquire positive or negative\ncharges and move toward the film. 20 The\ncharged ions were trapped on the surface of the film, significantly\nincreasing the surface-charge density of the friction layer. 21 Owing to the properties of the magnetic polymer\nelastomer itself, it compressed to a certain extent under pressure;\nthus, the contact area increased when certain magnetic particles are\nadded. However,\nif the magnetic particle droplets are over the certain amount, the\nair will fill the space between the friction layer and the electrode\non the back, reducing the overall dielectric constant and capacitance\nlevel, also, because of the thickness of the droplets themselves,\nwhen the number of droplets increases, leading to an overall thicker\nstructure, eq 1 shows\nthat the surface charge density gradually decreases. 22 Figure 1 c shows a photograph of the TENG structure. The left side shows PTFE\ncoated with magnetic droplets, and the right side shows a nylon film\ncoated with magnetic droplets. Nylon is relatively brittle; thus,\nafter the back is coated with liquid droplets, damage inevitably occurs\nduring friction, which also affects the output performance. 1 where σ is\nthe surface\ncharge density, ε 0 is the vacuum dielectric constant,\nε r is the relative dielectric constant of the friction\nlayer, V tri is the triboelectric voltage,\nand d is the thickness of the dielectric layer. Figure 1 (a) Schematic\nof the structure of PTFE–magnetic droplets–electrode.\n(b) Schematic of the charge injection process and the trapping effect\nof magnetic droplets. (c) Photograph of the PTFE–magnetic dot–electrode\nand nylon–magnetic dot–electrode. The operating mechanism of the TENG is illustrated in Figure 2 a. The frictional\ncharge generated by a TENG is typically attributed to coupled triboelectrification\nand electrostatic-induction effects. When the nylon and PTFE films\nwere in full contact with each other, the charge on the nylon surface\nmoved to the PTFE because its Fermi level was considerably higher\nthan that of the PTFE. This resulted in a negative charge on the PTFE\nand a positive charge on the nylon surface. When the two friction\nlayers were separated to produce an interval, the back electrode produced\nopposite charges induced by the charge on the friction layer owing\nto the electrostatic-induction effect. Electrons flowed from the PTFE-side\nelectrode to the nylon-side electrode, which provided a negative charge,\nwhereas the PTFE back electrode provided a positive charge. When the\ncharge reached equilibrium, PTFE and nylon were completely separated.\nWhen pressure was applied again to bring the nylon and PTFE into contact,\na reverse current flowed from the PTFE-side electrode to the nylon-side\nelectrode. When the TENG contacts separated, the device alternately\ngenerated current owing to the coupling of the contact-charged cation\nand electrostatic induction. 23 Figure 2 b shows a schematic\nof the trapping properties of the devices containing magnetic particles\ncompared with the reference group. The top friction layer caused charge\nflow, which produced a surface charge owing to contact. Owing to the\ncharacteristics of magnetic particles and dielectric polymers, the\nparticles can act as a trapping site for free electrons and generate\ndipoles near themselves. This increased the surface-charge density\nand the flow of electrons through the external circuit, thereby enhancing\nthe electrical-output characteristics. Figure 2 (a) Working mechanism\nof the magnetic particle contained TENG.\n(b) Trapping mechanism of the trapping effect induced by the magnetic\nparticles. Figure 3 depicts\nthe trapping role played by the magnetic dots in the TENG structure\nin more detail. The TENG models with and without magnetic particles\nwere further explained to describe the practical effect of the trapping\nproperties on the voltage output. When the friction layers were in\ncontact with each other, the surface charge generated by the contact\nprocess induced an opposite charge on the bottom electrode. In addition\nto the natural dissipation of charge in air, the back-electrode-induced\ncharge also decreased the surface-charge density. When magnetic particles\nprovide trapping sites, they effectively capture free electrons and\nprevent them from binding to the induced charge on the electrode,\nthereby preventing charge loss. In the process of charge injection,\nthe magnetic particles acting as trapping sites can effectively capture\nthe injected charge compared to the reference group without a trapping\nlayer. Increasing the amount of captured charge can directly increase\nthe surface-charge density and prevent it from combining with the\ncharge on the electrode to avoid charge loss. After injection, the\ncharge will inevitably dissipate into the air, resulting in loss.\nIn this process, the surface charge continues to combine with the\ninduced charge at the bottom electrode; thus, the charge level almost\nreturns to its original level before injection after a few hours,\nresulting in a low output voltage. Because samples containing trapping\nsites effectively capture charges, they can slow the dissipation of\ncharges to the air to a certain extent and block the combination of\nthe surface and induced charges on the back electrode, thereby maintaining\na high-voltage output after a few hours. 24 Figure 3 Charge\ntransport and trapping mechanisms with/without magnetic\nparticles. Figure 4 a shows\nthe output voltage of the nylon dielectric layer coated with magnetic\nparticles at the bottom (NM-P group); i shows the N-P group, and ii,\niii, and iv show the 4 × 4, 6 × 6, and 8 × 8 dots of\nthe NM-P groups, respectively (the output voltages of the TENG under\ndifferent injection conditions are shown in Figures S4–S6 for 4 × 4, 6 × 6, and 8 × 8 dots,\nrespectively). Notably, as shown in Figure 1 c, nylon itself has brittle mechanical properties,\nand when relatively hard magnetic particles are added to the back\nside, the part close to the particles will incur damage during friction,\nthereby affecting the overall output voltage. In addition, nylon has\npoor trapping properties, and previous studies have indicated that\nthe surface-charging performance of nylon after charge injection is\nworse than that of PTFE. 25 After injecting\na voltage of +7 kV, little difference was observed between the 4 ×\n4 dot combination, 6 × 6 dot group, and the reference group,\nwhile the 8 × 8 dot group showed the lowest output voltage. This\nmay be because the increase in the overall particle number inevitably\ncaused air to fill the space between the friction layer and the electrode\non the back, and the breakdown effect reduced the overall dielectric\nconstant and capacitance level. 26 In addition,\nthe excessive increase in the overall thickness of the TENG is a factor\nthat cannot be ignored. As mentioned, an excessively strong magnetic\nfield leads to the penetration of the electric field, which has a\nmore adverse effect on the voltage output. 27 , 28 However, after undergoing charge injection, the three groups of\nsamples containing magnetic particles showed better trapping properties\nthan the reference group in 4000 friction cycles and retained a high-output\nstate for a relatively long time. However, even with the addition\nof magnetic particles, the injection charge on the nylon surface could\nnot maintain a high output state for a long time, and the 8 ×\n8 dots group showed the best trapping performance of 63% after 5000\nfriction cycles (63 V decayed to 40 V). Figure 4 (a) Output measurement\ndata of the (i) N-P and (ii–iv) NM-P\nTENG. (measured by cycles). (b) Schematic of the nylon–magnetic\ndot–electrode TENG. (c) The summary graph of the 4 samples\nis to compare the output performance. The next measurement was of the NM-PM structure, as shown in Figure 5 a; i shows the N-P\ngroup, and ii, iii, and iv show the 4 × 4, 6 × 6, and 8\n× 8 dots of the NM-PM groups, respectively. In this experiment,\ncharge injections of +7 and −7 kV were performed on the surfaces\nof the positive and negative friction layers, respectively, and the\nfriction experiment was conducted for 5000 consecutive cycles. The\nresults showed that the output performance of the reference group\nwithout magnetic particles was superior. This result can be attributed\nto the obvious disadvantage of this structure: the positive and negative\nfriction layers contain magnetic particle droplets simultaneously,\nand this part has an extra height. In this case, only the raised part\nof the magnetic particle droplets will experience friction, where\nthe contact area changed to magnetic particle droplets, magnetic particle\ndroplets from magnetic particle droplets opposite film, greatly reducing\nthe friction contact area; also when the magnetic particle droplets\nare added on both sides at the same time, the thickness of the overall\nTENG will inevitably increase, which will also adversely affect the\noutput voltage. However, its output performance continued to decline\nby approximately 10% (from 201 to 179 V) after 4000 cycles of friction.\nAs more particles were introduced, the trapping property exhibited\nan increasing trend, and the output voltage of the 8 × 8 dot\ngroup was almost the same (reduced from −54 to −53 V)\nafter the same 5000 cycles of friction. Although the overall output\nvoltages of all groups with trapping layers were lower than those\nof the reference group, an increase in the trapping properties could\nbe verified. Figure 5 (a) Measurement data of the (i) N-P and (ii–iv)\nNM-PM TENG.\n(measured by cycles). (b) Schematic of the NM-PM TENG. (c) Summary\ngraph of the 4 samples to compare the output performance. Finally, the N-PM structure was analyzed; Figure 6 (i) shows the N-P group, and\nii, iii, and\niv show the 4 × 4, 6 × 6, and 8 × 8 dots of the N-PM\ngroups, respectively. According to previous studies, PTFE has good\ntrapping properties and can maintain a high surface charge for a certain\nperiod without the addition of trapping layers. The test results showed\nthat when compared to the reference group, the output voltage increased\nbut later declined as the number of magnetic particles increased.\nWe estimated that the reason for the decrease in the output voltage\nin the 8 × 8 dot group was the same as that in the NM-P group.\nAs shown in Figure 6 a,b, the 4 × 4 dot group demonstrated a high output voltage\nof 280 V after charge injection, 1.4 times that of the reference group\nof 200 V. Additionally, it maintained nearly the same output voltage\nafter 5000 friction cycles, demonstrating high output performance\nover a long period. Figure 6 c,e shows the voltage test results measured once every hour\nwithout continuous friction. The reason for applying this measurement\nis that in a state of continuous friction, new surface charges are\nconstantly generated owing to electrostatic-induction effects; therefore,\nobserving the phenomenon of charge decay is difficult [as shown in Figure 6 a(i), even after\n4000 cycles of friction, the charge is only reduced by 10%]. To better\nobserve the charge-decay phenomenon, the test time was set to once\nevery hour. The test results showed that the output voltage of the\nsample without a trapping layer rapidly decreased by 75% (from 164\nto 46 V) after 1 h of injection and almost decreased to its original\nvalue before injection after 3 h of injection (28 V). With an increase\nin the number of magnetic particles, the trapping properties improved.\nAfter 3 h of injection, the 4 × 4, 6 × 6, and 8 × 8\ndots maintained outputs of 33% (207 to 69 V), 59% (259 to 153 V),\nand 58% (193 to 112 V) of the highest output voltage after charge\ninjection, respectively. In addition, the output voltage 3 h after\nthe injection process showed that the N-PM 6 × 6 dot group maintained\na considerably higher output level (153 V) than that of the reference\nN-P group (28 V). This realized the goal of maintaining a high output\nfor a long time. Figure 6 (a) Measurement data of the (i) N-P and (ii–iv)\nN-PM TENG\n(measured by cycles). (b) Summary graph of the 4 samples to compare\nthe output performance. (c) Measurement data of the (i) N-P and (ii–iv)\nN-PM TENG (measured by hours). (d) Schematic of the N-PM TENG. (e)\nSummary graph of the 4 samples to compare the output performance (measured\nby hours). Figure 7 shows the\nperformance and practical applications of the TENG device after the\ncharge injection. Figure 7 a shows a circuit diagram of the TENG when load resistance\nwas applied for testing. The output voltage and current were measured\nusing changes in applied resistance to calculate the actual output\npower of the TENG. As shown in Figure 7 b, after the charge injection, the output voltage gradually\nincreased from 12 to 220 V with an increase in resistance, whereas\nthe output current decreased from 2.8 to 0.1 mA. The actual output\npower of the entire TENG reached its highest value of 167 mW/m 2 at 2 × 10 –7 Ω. As part of the\npractical application, we chose a sample of the N-PM structure and\nfirst performed charge injection. Figure 7 d shows the sensitivity of the TENG as a\nsensor. By applying different forces, holding them for approximately\n1 s, and releasing them again, the measurement results showed that\nthe sensor maintained the same proportional growth when the stress\nincreased, indicating good sensitivity. Figure 7 e,f shows the performance of the sensor when\nstrapped to the wrist of a person and placed under the insole, respectively.\nWith an increase in the degree of wrist bending, the output voltage\nshowed a gradually increasing trend, and the shape of the output signal\nwas maintained, showing a stable sensing performance. When the sensor\nwas pressed under the insole with different actions, such as “gentle\nstep, mid step, and hard step”, the frequency and size of the\nsignal changed correspondingly, which confirmed the good sensitivity\nof the sensor. (The photograph of the finger bending sensor and the\ninsole walking sensor is shown in Figures S7 and S8 . The video of the finger bending sensor is shown in Video S1 .) Figure 7 (a) Schematic of the circuit of the load\nresistance measurement.\n(b) Voltage and current changes depend on the load resistance. (c)\nPower calculated by the load resistance measurement. (d) Sensitivity\nmeasurement under different pressure. (e) Performance of the wrist\nbending sensor. (f) Performance of step sensor."
} | 5,912 |
29899510 | null | s2 | 1,243 | {
"abstract": "Quorum sensing (QS), where bacteria secrete and respond to chemical signals to coordinate population-wide behaviors, has revealed that bacteria are highly social. Here, we investigate how diversity in QS signals and receptors can modify social interactions controlled by the QS system regulating bacteriocin secretion in Streptococcus pneumoniae, encoded by the blp operon (bacteriocin-like peptide). Analysis of 4096 pneumococcal genomes detected nine blp QS signals (BlpC) and five QS receptor groups (BlpH). Imperfect concordance between signals and receptors suggested widespread social interactions between cells, specifically eavesdropping (where cells respond to signals that they do not produce) and crosstalk (where cells produce signals that non-clones detect). This was confirmed in vitro by measuring the response of reporter strains containing six different blp QS receptors to cognate and non-cognate peptides. Assays between pneumococcal colonies grown adjacent to one another provided further evidence that crosstalk and eavesdropping occur at endogenous levels of signal secretion. Finally, simulations of QS strains producing bacteriocins revealed that eavesdropping can be evolutionarily beneficial even when the affinity for non-cognate signals is very weak. Our results highlight that social interactions can mediate intraspecific competition among bacteria and reveal that competitive interactions can be modified by polymorphic QS systems."
} | 365 |
35363428 | PMC8973260 | pmc | 1,244 | {
"abstract": "Abstract The one‐pot synthesis of 2,5‐furandicarboxylic acid from 2‐furoic acid with a yield of 57 % was achieved for the first time using a Pd‐catalyzed bromination‐hydroxycarbonylation tandem reaction in HOAc‐NaOAc buffer. This synthetic protocol shows major improvements compared to previously reported methods, such as using biomass‐based 2‐furoic acid as low‐cost raw material, one‐pot synthesis without isolation of intermediate products, and no need for an acidification procedure. Experiments indicate that the involved Xantphos‐modified Pd‐catalyst and the buffer solution play significant promoting roles for each individual reaction whereas Br 2 (as the brominating reagent) had a negative effect on the second hydroxycarbonylation step, while CO was deleterious for the first bromination step. Hence, in this practical one‐pot synthesis, Br 2 should be consumed in the first bromination step as fully as possible, and CO is introduced after the first bromination step has been completed.",
"conclusion": "Conclusion This work presents the one‐pot bromination‐hydroxycarbonylation tandem reaction for synthesis of 2,5‐FDCA from 2‐FCA for the first time. Under the optimal conditions, 2,5‐FDCA was obtained in the yield of 57 % without the need of acidification procedure. The use of HOAc‐NaOAc buffer with pH 3.58 resolves the conflicting problem for bromination‐hydroxycarbonylation sequence. It has been found that the involved Xantphos‐modified Pd catalyst and HOAc‐NaOAc buffer solution played significant promoting role for each individual reaction whereas Br 2 (as the brominating reagent) had negative effect on the second‐step hydroxycarbonylation, and CO was a poison for the first‐step bromination. Hence, in this practical one‐pot synthesis, Br 2 should be consumed in the first‐step bromination as flly as possible, and CO is charged after the first bromination step has been completed.",
"introduction": "Introduction The production of bio‐based 2,5‐furandicarboxylic acid (2,5‐FDCA) has attracted much attention in the recent decade due to its versatility as an important bio‐based platform chemical as well as the concern to decrease the dependence on fossil feedback to produce chemicals.[ \n 1 \n , \n 2 \n ] 2,5‐FDCA has proved itself a suitable monomer for the synthesis of polymers, such as poly(ethylene furandicarboxylate) (PEF), a potential polymer used in soft drink bottling, \n [3] \n which might be a replacement for the fossil‐derived poly(ethylene terephthalate) (PET). In the literature, there are three main routes for the synthesis of 2,5‐FDCA as summarized in Scheme 1 . Route‐I involves oxidation of 5‐(hydroxymethyl) furfural (HMF) to 2,5‐FDCA through a number of thermal‐catalytic,[ \n 4 \n , \n 5 \n , \n 6 \n , \n 7 \n , \n 8 \n , \n 9 \n , \n 10 \n ] electrocatalytic, \n [11] \n photocatalytic \n [12] \n and biocatalytic[ \n 13 \n , \n 14 \n , \n 15 \n ] processes. Thereinto, sequential oxidation of the two functional groups (formyl and hydroxyl) bound to the furan ring leads to the formation of 2,5‐FDCA. Although this reaction route has been widely studied, the low selectivity of 2,5‐FDCA is still a concerning problem due to the formation of many unstable intermediate products as well as the lability of HMF itself.[ \n 1 \n , \n 2 \n , \n 16 \n ] Moreover, HMF, generally obtained from edible fructose in conflict with food industry, is required to be of very high purity but is quite difficult to be purified as a solid compound. Route‐II and Route‐III pertain to furfural or its downstream product being converted into 2,5‐FDCA. The industrial production of furfural from non‐edible pentose (derived from hemicellulose) is a mature process. The production of 2‐furancarboxylic acid (2‐FCA) from furfural oxidation using several heterogeneous catalysts is also well established.[ \n 17 \n , \n 18 \n , \n 19 \n , \n 20 \n , \n 21 \n ] Hence, 2‐FCA is very suitable to replace HMF, and Route‐II, carboxylation of 2‐FCA by CO 2 into 2,5‐FDCA, becomes an interesting protocol along with the increased concern on CO 2 utilization. This reaction only occurred from 200 to 350 °C with need of strongly alkaline salts in order to deprotonate the very weakly acidic C−H bond (p K \n a >40) to react with CO 2 , but the reaction mechanism remains speculative.[ \n 16 \n , \n 22 \n , \n 23 \n , \n 24 \n , \n 25 \n ] In Route‐III, carbonylation of 5‐bromofuroic acid (as the derivative of 2‐FCA) with CO and H 2 O, was reported by Yin's group, wherein 2,5‐FDCA was obtained with excellent selectivity under mild conditions using homogenous and heterogenous Pd‐ catalysts, respectively.[ \n 26 \n , \n 27 \n , \n 28 \n , \n 29 \n ] In addition, the Henkel‐type disproportion[ \n 30 \n , \n 31 \n , \n 32 \n ] of potassium‐2‐furoate at 260 °C in the presence of Lewis‐acidic catalysts like CdI 2 or ZnCl 2 to form 2,5‐FDCA and furan was firstly reported by D. S. van Es. \n [33] \n \n Scheme 1 Different pathways for the synthesis of 2,5‐FDCA. The use of fructose as the starting material instead of HMF in the synthesis of 2,5‐FDCA through stepwise oxidation processes further enhanced the reaction complexity, which required more elaborately designed catalytic systems and engineering techniques.[ \n 34 \n , \n 35 \n , \n 36 \n , \n 37 \n ] Accordingly, not only the applied catalysts in the mode of conventional heating or electro‐/photo‐initiation, but also the selection of the bio‐derived raw materials determines the practicability and diversity in the synthesis of 2,5‐FDCA. Inspired by Yin's work as well as our continuous interest in carbonylation using ligand‐modified Pd catalysts, in this work, we explored the one‐pot bromination‐hydroxycarbonylation tandem reaction for the synthesis of 2,5‐FDCA acid from 2‐FCA, which is an unprecedented transformation in the literature. In the reaction sequence, the bromination of 2‐FAC by brominating reagent such as liquid bromine (Br 2 ) goes by the very well established acid‐catalyzed electrophilic aromatic substitution mechanism, whereas the hydroxycarbonylation of 5‐bromofuroic acid ( 1 ) with CO and water occurs in the presence of a phosphine‐modified Pd catalyst with presence of base. Evidently, in this one‐pot sequence, the acid is required for the first bromination step and the base is required for second step, which are totally conflicting conditions. In order to solve this problem, a buffer solution composed of HOAc and NaOAc with pH range of 3.0 to 4.0 was applied in this work, which proved workable not only in the way to enable each individual reaction performing efficiently but to also directly yield the product, 2,5‐FDCA, without the need for acidification.",
"discussion": "Results and Discussion We initiated our studies by carrying out the optimization of reaction conditions for one‐pot bromination‐carbonylation tandem reaction for the synthesis of 2,5‐FDCA ( 2 ) from 2‐FCA, which was catalyzed by Xantphos‐modified PdCl 2 (Table 1 ). PdCl 2 was identified as the best choice of catalyst precursor. The mild reaction temperature of 90 °C was required as otherwise many unidentified sideproducts coming from ring‐opening reactions of 2‐FCA and the products ( 1 and 2 ) were formed.\n Table 1 Palladium‐catalyzed bromination‐hydroxycarbonylation tandem reaction for synthesis of 2,5‐FDCA ( 2 ) from 2‐FCA. [a] \n \n \n \n \n \n Entry \n \n Time 1 [h] \n \n Time 2 [h] \n \n Amount of brominating reagent [mmol] \n \n H 2 O [mmol] \n \n Solvent \n \n Ligand \n \n Conv. [%] [b] \n \n \n Sel. of 1 [%] [b] \n \n \n Sel. of 2 [%] [b] \n \n \n Yield of 2 [%] [b] \n \n \n 1 \n \n 12 \n \n 12 \n \n 3 (Br 2 ) \n \n 50 \n \n DMF \n \n Xantphos \n \n 67 \n \n 2 \n \n 42 \n \n 28 \n \n 2 \n \n 12 \n \n 12 \n \n 4 (Br 2 ) \n \n 50 \n \n DMF \n \n Xantphos \n \n 90 \n \n 1 \n \n 50 \n \n 45 \n \n 3 [c] \n \n \n 12 \n \n 12 \n \n 4 (Br 2 ) \n \n 50 \n \n DMF \n \n Xantphos \n \n 77 \n \n 0 \n \n 17 \n \n 13 \n \n 4 \n \n 10 \n \n 12 \n \n 4 (Br 2 ) \n \n 50 \n \n DMF \n \n Xantphos \n \n 88 \n \n 1 \n \n 60 \n \n 53 \n \n 5 \n \n 8 \n \n 12 \n \n 4 (Br 2 ) \n \n 50 \n \n DMF \n \n Xantphos \n \n 83 \n \n 1 \n \n 49 \n \n 43 \n \n 6 \n \n 10 \n \n 4 \n \n 4 (Br 2 ) \n \n 50 \n \n DMF \n \n Xantphos \n \n 88 \n \n 1 \n \n 57 \n \n 50 \n \n 7 \n \n 10 \n \n 2 \n \n 4 (Br 2 ) \n \n 50 \n \n DMF \n \n Xantphos \n \n 73 \n \n 0 \n \n 67 \n \n 49 \n \n 8 \n \n 10 \n \n 2 \n \n 4 (Br 2 ) \n \n 5 \n \n DMF \n \n Xantphos \n \n 77 \n \n 2 \n \n 74 \n \n 57 \n \n 9 \n \n 10 \n \n 2 \n \n 4 (Br 2 ) \n \n 5 \n \n THF \n \n Xantphos \n \n 81 \n \n 1 \n \n 31 \n \n 25 \n \n 10 \n \n 10 \n \n 2 \n \n 4 (Br 2 ) \n \n 5 \n \n NMP \n \n Xantphos \n \n 80 \n \n 2 \n \n 31 \n \n 25 \n \n 11 \n \n 10 \n \n 2 \n \n 4 (Br 2 ) \n \n 5 \n \n – \n \n Xantphos \n \n 85 \n \n 1 \n \n 18 \n \n 15 \n \n 12 [d] \n \n \n 10 \n \n 2 \n \n 4 (Br 2 ) \n \n 5 \n \n DMF \n \n Xantphos \n \n 77 \n \n 2 \n \n 48 \n \n 37 \n \n 13 [e] \n \n \n 10 \n \n 2 \n \n 4 (Br 2 ) \n \n 5 \n \n DMF \n \n \n L1 \n \n \n 79 \n \n 11 \n \n 70 \n \n 55 \n \n 14 [f] \n \n \n 10 \n \n 2 \n \n 4 (Br 2 ) \n \n 5 \n \n DMF \n \n PPh 3 \n \n \n 72 \n \n 3 \n \n 50 \n \n 36 \n \n 15 [g] \n \n \n 10 \n \n 2 \n \n 4 (Br 2 ) \n \n 5 \n \n DMF \n \n Xantphos \n \n 71 \n \n 3 \n \n 44 \n \n 31 \n \n 16 \n \n 10 \n \n 2 \n \n 4 (NaBrO) \n \n 5 \n \n DMF \n \n Xantphos \n \n 69 \n \n 5 \n \n 81 \n \n 56 \n \n 17 [h] \n \n \n 10 \n \n 2 \n \n 4 (NBS) \n \n 5 \n \n DMF \n \n Xantphos \n \n 86 \n \n 0 \n \n 5 \n \n 4 \n \n 18 [i] \n \n \n 10 \n \n 2 \n \n 4 (TBABr 3 ) \n \n 5 \n \n DMF \n \n Xantphos \n \n 78 \n \n 1 \n \n 7 \n \n 6 \n \n 19 [j] \n \n \n 10 \n \n 2 \n \n 4 (Br 2 ) \n \n 5 \n \n DMF \n \n Xantphos \n \n 50 \n \n 0 \n \n 38 \n \n 19 \n \n 20 [k] \n \n \n 10 \n \n 2 \n \n 4 (Br 2 ) \n \n 5 \n \n DMF \n \n Xantphos \n \n 57 \n \n 2 \n \n 47 \n \n 27 \n \n 21 [l] \n \n \n 10 \n \n 2 \n \n 4 (Br 2 ) \n \n 5 \n \n DNF \n \n Xantphos \n \n 75 \n \n 1 \n \n 75 \n \n 56 \n [a] 2‐FCA 2 mmol, PdCl 2 5 mol %, ligand 5 mol %, NaOAc 6 mmol, HOAc 87 mmol, solvent 5 mL, 90 °C. N 2 (0.1 MPa) was charged initially for the first step reaction and CO (1.0 MPa) was then charged after time 1 for the second step reaction; [b] determined by HPLC. Many kinds of ring‐opening side products were found but could not be identified; [c] CO (1.0 MPa) was always present during the overall tandem reaction; [d] PdCl 2 2 mol %, Xantphos 2 mol %; [e] L1 10 mol %; [f] PPh 3 10 mol %; [g] Xantphos 10 mol %; [h] NBS, N ‐Bromosuccinimide; [i] TBABr 3 , tetrabutylammonium tribromide; [j] Pd(OAc) 2 5 mol %; [k] Pd(CH 3 CN) 2 Cl 2 5 mol %; [l] sodium 2‐furoate (2 mmol) applied instead of 2‐FCA. Wiley‐VCH GmbH The excess Br 2 of two equivalents was required to guarantee the first bromination step of 2‐FCA towards 5‐bromofuran‐2‐carboxylic acid ( 1 ) (Entry 2 vs. 1). It was found that the presence of CO (1.0 MPa) dramatically reduced the yield of target product 2 due to the inhibited bromination of 2‐FCA in the CO atmosphere (Entry 3 vs. 2). Hence, an N 2 atmosphere (0.1 MPa) was applied for the first bromination step, and CO (1.0 MPa) was followingly introduced for the second carbonylation step. The long reaction time of 10 h was needed for this N 2 ‐protected bromination and a shorter time of 2 h was found sufficient for the second carbonylation step (Entries 2, 4, 5 vs. 6 and 7). Longer reaction times of up to 12 h just resulted in the prevailing of the ring‐opening reaction of 1 , leading to a decreased selectivity for 2 in the overall tandem process (Entry 4 vs. 2). Less water resulted in higher selectivity for 2 (Entry 8 vs. 7). A solvent screening showed that DMF was the best solvent, corresponding to the highest yield of 2 (Entry 8, 57 %). The use of THF, NMP or no solvent led to comparable conversions of 2‐FCA but at much lower selectivity for 2 (Entries 8 vs. 9–11). A decreased concentration of the Pd catalyst led to the obviously diminished selectivity of 2 (Entry 8 vs. 12), implying that the involved Xantphos‐modified PdCl 2 catalyst also contributed to the first bromination step of 2‐FCA since just the success of this bromination could warrant the subsequent carbonylation ensuing smoothly. With these established optimal reaction conditions, the ligand effect on the reaction efficiency was investigated. When L1 with the same xanthenyl‐skeleton but as a mono‐phosphine was applied instead of the bisphosphine Xantphos, a comparable yield of 2 of 55 % was obtained (Entry 13 vs. 8) whereas the use of the mono‐phosphine PPh 3 just corresponded to a much lower yield of 36 % (Entry 14). However, an increased amount of Xantphos led to a drop in selectivity for 2 despite no dramatic effect on 2‐FCA conversion (Entry 15). Using NaBrO instead of Br 2 as brominating reagent led to a comparable yield of 2 with an improved selectivity of 81 % despite a relatively lower conversion of 2‐FCA (Entry 16 vs. 8). Besides, the use of NBS ( N ‐bromosuccinimide) or TBABr 3 (tetrabutylammonium tribromide) corresponded to an extremely low selectivity for 2 due to the prevailing furan‐ring‐opening side‐reactions (Entries 17 and 18). In comparison to Pd(OAc) 2 and Pd(CH 3 CN)Cl 2 as catalyst precursors, PdCl 2 resulted in a much higher yield of 2 (Entries 8 vs. 19 and 20), presumably due to the in situ formed acidic HCl derived from the reaction of PdCl 2 and HOAc favoring the first bromination step. Additionally, when sodium 2‐furoate was applied as the staring material, a nearly identical outcome was obtained in comparison to 2‐FCA (Entry 21 vs. 8). Since buffer solution composed of NaOAc (6 mmol) and (HOAc 87 mmol) was involved in excess, 2‐FCA and its sodium salt always co‐existed in this reaction system regardless of whether 2‐FCA (2 mmol) or sodium 2‐furoate (2 mmol) were used as the starting materials. In general, an acid is required as the catalyst for classical bromination of aromatic compounds. A base is required as the scavenger for hydroxycarbonylation of aryl bromides. Hence, in this studied bromination‐hydroxycarbonylation sequence, the sole presence of acid (HOAc) or base (NaOAc) nearly shut down the tandem reaction (Entries 1 and 2 of Table 2 ). The use of a buffer solution composed of HOAc and NaOAc with pH 3.58 was proven to be an effective strategy to resolve the conflicting problem for this sequence. As indicated in Table 2 , the selectivity towards 2 was sensitive to the pH value of the involved buffer. Increasing the pH value of HOAc‐NaOAc buffer from 3.58 up to 5.00 dramatically decreased the selectivity towards 2 , leading to lowered yields (Entry 3 vs. 4–6). However, the decrease of the pH value of the HOAc‐NaOAc buffer from 3.58 down to 3.00 also reduced the selectivity for 2 due to the prevalence of the competitive furan‐ring‐opening reactions under the stronger acidic conditions (Entry 3 vs. 7). The use of an HOAc‐NH 4 OAc buffer in place of HOAc‐NaOAc with pH 3.50 led to a similar yield for 2 (Entry 3 vs. 8).\n Table 2 The effect of pH value of the buffer on Pd‐catalyzed bromination‐hydroxycarbonylation tandem reaction for synthesis of 2,5‐FDCA ( 2 ) from 2‐FCA. [a] \n \n \n \n \n \n Entry \n \n Buffer \n \n Conv. [%] [b] \n \n \n Sel. of 1 [%] [b] \n \n \n Sel. of 2 [%] [b] \n \n \n Yield of 2 [%] [b] \n \n \n Acid (amount [mmol]) \n \n Salt (amount [mmol]) \n \n pH value \n \n 1 \n \n – \n \n NaOAc (6) \n \n 9.88 [c] \n \n \n 61 \n \n 8 \n \n 13 \n \n 8 \n \n 2 \n \n HOAc (87) \n \n – \n \n 2.50 [d] \n \n \n 70 \n \n 6 \n \n 21 \n \n 15 \n \n 3 \n \n HOAc (87) \n \n NaOAc (6) \n \n 3.58 [e] \n \n \n 73 \n \n 0 \n \n 67 \n \n 49 \n \n 4 \n \n HOAc (33) \n \n NaOAc (6) \n \n 4.00 [e] \n \n \n 77 \n \n 1 \n \n 32 \n \n 25 \n \n 5 \n \n HOAc (10) \n \n NaOAc (6) \n \n 4.50 [e] \n \n \n 69 \n \n 0 \n \n 25 \n \n 17 \n \n 6 \n \n HOAc (3.3) \n \n NaOAc (6) \n \n 5.00 [e] \n \n \n 68 \n \n 0 \n \n 19 \n \n 13 \n \n 7 \n \n HOAc (330) \n \n NaOAc (6) \n \n 3.00 [c] \n \n \n 83 \n \n 1 \n \n 55 \n \n 46 \n \n 8 \n \n HOAc (87) \n \n NH 4 OAc (6) \n \n 3.50 [c] \n \n \n 77 \n \n 1 \n \n 66 \n \n 51 \n [a] 2‐FCA 2 mmol, PdCl 2 5 mol %, Xantphos 5 mol %, Br 2 4 mmol, H 2 O 50 mmol, DMF 5 mL, 90 °C, time 1 =10 h, time 2 =2 h. N 2 (0.1 MPa) was charged initially for the first reaction step and CO (1.0 MPa) was then charged after 10 h for the second reaction step; [b] fetermined by HPLC. Many kinds of ring‐opening side products were found but could not be identified; [c] in saturated aqueous solution of NaOAc (25 °C); [d] in pure HOAc (25 °C); [e] the pH value is calculated according to the equation of pH=p K \n a +lg[C salt /C acid ] (p K \n a =4.74, representing the p K \n a value of HOAc at 25 °C; C salt is defined as the concertation of the salt; C acid is defined as the concertation of the acid. Wiley‐VCH GmbH In order to demonstrate the (positive/negative) influences of the involved Br 2 , CO, and Xantphos‐PdCl 2 catalyst on each individual reaction, the bromination of 2‐FCA and the hydroxycarbonylation of 1 (5‐bromofuran‐2‐carboxylic acid) with CO and H 2 O were separately carried out under controlled conditions (Scheme 2 ). When the bromination of 2‐FCA by Br 2 was performed under an N 2 ‐atmosphere with presence of HOAc‐NaOAc buffer in DMF, only 13 % yield of the brominated product 1 was obtained. The consumed 2‐FCA was mostly converted into products related to the furan‐ring‐opening reaction of 2‐FCA [Scheme 2 ‐(1)]. In comparison, under the same conditions, when the catalyst of Xantphos‐PdCl 2 presumed to only be responsible for second hydroxycarbonylation step of 1 was involved, the efficiency of the bromination of 2‐FCA by Br 2 was dramatically improved, resulting in 75 % selectivity for 1 with 53 % yield [Scheme 2 ‐(2)]. This outcome indicated that the Xantphos‐PdCl 2 catalyst also exhibited significant activity in the bromination of 2‐FCA along with the acidic HOAc‐NaOAc buffer. The control experiment in Scheme 2 ‐(3) demonstrated that the presence of water in small amounts (2.5 equiv. with respect to 2‐FCA), which was required as the substrate in the hydroxycarbonylation, had a negligible effect on the bromination of 2‐FCA towards 1 . However, the presence of CO (1.0 MPa) almost stopped the bromination to deliver 1 [Scheme 2 ‐(4)], possibly due to the in situ coordination of CO to the Pd center to quench the catalytic performance of PdCl 2 as a Lewis‐acidic catalyst, which was consistent with the result in entry 3 of Table 1 . Hence, in the practical bromination‐hydroxycarbonylation tandem reaction, the first bromination step should be performed without a CO‐atmosphere as shown in Tables 1 and 2 . On the other hand, the independent hydroxycarbonylation of 1 with CO and H 2 O was carried out as shown in Scheme 2 ‐(5–8) under different conditions. It was indicated that the hydroxycarbonylation of 1 in the presence of HOAc‐NaOAc as an acid‐scavenger afforded 2 in the yield of only 73 % despite 99 % conversion of 1 , wherein the complicated and unidentified side products were found due to furanyl‐ring‐opening reactions of 1 and 2 [Scheme 2 ‐(5)]. In comparison, the use of NaOAc instead of HOAc‐NaOAc to repeat the reaction led to 93 % yield of 2 along with 99 % conversion of 1 [Scheme 2 ‐(6)], which implied that the acidity of HOAc‐NaOAc indeed would drive the furan‐ring‐opening reaction to happen in parallel. However, the use of NaOH, a stronger base, in place of NaOAc stopped the expected hydroxycarbonylation due to the prevailing hydrolysis of furan‐ring of 1 [Scheme 2 ‐(7)]. Similarly, the presence of Br 2 also shut down the expected carbonylation due to the severe oxidative degradation of 1 by Br 2 [Scheme 2 ‐(8)]. Hence, it was suggested that the involved Br 2 should be exhausted completely in the first‐step bromination of 2‐FCA in order to rule out its negative effect on the second‐step hydroxycarbonylation of 1 . Scheme 2 Control experiments (reaction conditions: substrate 2 mmol, PdCl 2 5 mol %, Xantphos 5 mol %, Br 2 4 mmol, NaOAc 6 mmol, HOAc 87.4 mmol, H 2 O 5 mmol, DMF 5 mL; N 2 pressure 0.1 MPa, CO pressure 1.0 MPa; Conv. % and Sel. % were determined by HPLC.)"
} | 4,877 |
24347684 | null | s2 | 1,245 | {
"abstract": "Self-healing polymeric materials are systems that after damage can revert to their original state with full or partial recovery of mechanical strength. Using scaling theory we study a simple model of autonomic self-healing of unentangled polymer networks. In this model one of the two end monomers of each polymer chain is fixed in space mimicking dangling chains attachment to a polymer network, while the sticky monomer at the other end of each chain can form pairwise reversible bond with the sticky end of another chain. We study the reaction kinetics of reversible bonds in this simple model and analyze the different stages in the self-repair process. The formation of bridges and the recovery of the material strength across the fractured interface during the healing period occur appreciably faster after shorter waiting time, during which the fractured surfaces are kept apart. We observe the slowest formation of bridges for self-adhesion after bringing into contact two bare surfaces with equilibrium (very low) density of open stickers in comparison with self-healing. The primary role of anomalous diffusion in material self-repair for short waiting times is established, while at long waiting times the recovery of bonds across fractured interface is due to hopping diffusion of stickers between different bonded partners. Acceleration in bridge formation for self-healing compared to self-adhesion is due to excess non-equilibrium concentration of open stickers. Full recovery of reversible bonds across fractured interface (formation of bridges) occurs after appreciably longer time than the equilibration time of the concentration of reversible bonds in the bulk."
} | 420 |
34505915 | PMC9436828 | pmc | 1,246 | {
"abstract": "Rhizosphere microbiomes have received growing attention in recent years for their role in plant health, stress tolerance, soil nutrition, and invasion. Still, relatively little is known about how these microbial communities are altered under plant competition, and even less about whether these shifts are tied to competitive outcomes between native and invasive plants. We investigated the structure and diversity of rhizosphere bacterial and fungal microbiomes of native annual forbs and invasive annual grasses grown in a shade-house both individually and in competition using high-throughput amplicon sequencing of the bacterial 16S rRNA gene and the fungal ITS region. We assessed how differentially abundant microbial families correlate to plant biomass under competition. We find that bacterial diversity and structure differ between native forbs and invasive grasses, but fungal diversity and structure do not. Furthermore, bacterial community structures under competition are distinct from individual bacterial community structures. We also identified five bacterial families that varied in normalized abundance between treatments and that were correlated with plant biomass under competition. We speculate that invasive grass dominance over these natives may be partially due to effects on the rhizosphere community, with changes in specific bacterial families potentially benefiting invaders at the expense of natives. Supplementary Information The online version contains supplementary material available at 10.1007/s00248-021-01853-1.",
"conclusion": "Conclusions While our study was limited in scope, our results revealed that these six native annual forbs were host to a community of microbes distinct from, and more diverse than, those of their invasive competitors, but these close associations may be disrupted by invasive grasses. The joint bacterial rhizosphere of invasive grasses and native forbs differed from those of plants grown alone, with grasses contributing more to the abundance of ASVs from families that were linked to decreases in forb performance and/or increased in grass performance. Moreover, three of these families were higher in abundance in joint microbiomes relative to those of forbs alone. Forbs on the other hand contributed more to the abundance of ASVs from only one family that was associated with increased forb performance and decreased grass performance, but this family declined in abundance in the joint rhizosphere relative to forbs. Given the correlative nature of our study, more research is needed to understand the role of the microbiome in invasive grass dominance over native forbs and to clarify whether the observed changes in plant performance were indeed driven by these candidate bacterial families. Regardless, our study highlights the importance of considering the microbiome in ecosystems facing dominance by invasive species.",
"introduction": "Introduction Plants and their associated soil microbial communities have coevolved complex dynamic relationships through time. The soil directly surrounding plant roots, known as the rhizosphere, is home to a diversity of microbes that play vital roles in plant health [ 1 ], nutrition [ 2 ], and stress tolerance [ 3 , 4 ] and is distinct from the surrounding bulk soil [ 1 , 5 – 7 ]. Though rhizosphere composition is driven predominantly by abiotic factors [ 8 – 10 ], these communities can be highly host plant-specific [ 11 , 12 ], with plants exhibiting strong local adaptations to their home microbial community [ 13 – 15 ]. Furthermore, there is growing evidence that microbial communities vary in the presence of certain neighbors [ 16 ], and can even affect competitive outcomes between plants [ 17 – 19 ] with the potential to drive either coexistence or exclusion [ 20 , 21 ]. Microbes can affect plant-host competition both directly via resource availability and indirectly through community interactions. For example, microbes can make nitrogen more readily available to plants [ 22 , 23 ], which could preferentially benefit hosts with faster uptake rates [ 24 , 25 ] thereby increasing their competitive success [ 26 ]. Plants may also indirectly compete for beneficial plant-growth-promoting microbes [ 7 , 27 , 28 ] or come into contact with novel pathogens through competition with a novel plant, such as an invader [ 29 ]. These interactions can lead to the loss of specialized microbes in the inferior competitor [ 30 ] and drive the resulting microbial community to resemble that of the dominant competitor [ 18 , 19 ]. Shifts in the microbiome are a potentially important contributor to the dominance of invasive plants. Although there is strong evidence for local adaptation between plants and their home microbiome, invasive plants are relatively novel players, causing disruptions in important plant–microbe interactions by increasing soil microbial activity [ 31 ], reducing microbial biomass [ 32 ] and diversity [ 33 ], increasing nitrification rates [ 24 ], and changing microbiome composition [ 34 , 35 ]. These plant-soil feedbacks can benefit the invader [ 35 , 36 ], selectively harm natives [ 37 ], and/or benefit other invaders [ 38 ]. While it is known that invaders and natives can harbor distinct microbial communities, and that competition between host plants can affect microbiome composition (and vice versa), little is known of how changes in the abundance of specific microbial taxa may be tied to the competitive dominance of invaders and/or the competitive inferiority of natives. Serpentine soils in California annual grasslands are characterized by high heavy metal content and low levels of essential plant nutrients, and as a result, are home to a unique community of native forbs. While some of these soils are rocky and shallow, other serpentine areas are deeper and more finely textured (“lush” serpentine) and therefore hold more water and nitrogen [ 39 ], making them more favorable to invasive annual grass establishment and growth. These fast-growing invasive grasses outcompete the less abundant and competitively inferior native forbs, contributing to their declines across California grasslands [ 40 – 42 ]. In native-dominated serpentine areas, microbes have been shown to increase seedling survival [ 43 ] and facilitate heavy metal tolerance [ 44 ]. With the invasion of grasses, microbiome changes may strengthen the advantage grasses have over natives if they recruit beneficial microbes away from natives or harbor novel microbes harmful to natives. Alternatively, locally adapted microbes may be harming invasive grasses, helping native forbs to persist in these areas. We sought to understand how competition between native forbs and invasive grasses affects the bacterial and fungal rhizosphere microbiomes, and how microbes may be shaping competitive outcomes in this community. Our main questions were as follows: Do microbiomes differ between invasive grasses and native forbs? Do microbiomes of grass-forb pairs (i.e., under competition) differ from the microbiomes of (a) invasive grasses and (b) native forbs? If so, are microbiomes of grass-forb pairs sourced more from invasive grasses, native forbs, or equally from both? Given compositional changes in the microbiome in grass-forb pairs, which specific microbial families are driving these changes? Are abundances of these microbial families correlated with plant performance in grass-forb pairs? To answer these questions, we conducted a manipulative shade-house experiment in which forbs and grasses were grown in field-collected soil both individually and in grass-forb pairs. For each question, we looked at effects in both the bacterial and the fungal rhizosphere measured through high throughput amplicon sequencing, focusing our analysis on group-level results (i.e., forb, grass, and grass-forb pairs).",
"discussion": "Discussion In our experiment, the rhizosphere bacterial community of native forbs was structurally distinct from those of invasive grasses and had higher alpha diversity. Interactions between grasses and forbs here correlated with shifts in the bacterial rhizosphere towards communities dissimilar to both individual forb and grass microbiomes, with marginally higher diversity than grasses alone. Studies in similar annual grasslands have found little change in bacterial structure with increased invasive grass abundance [ 61 – 63 ] but invader dominance has been linked to decreased bacterial diversity in other systems [ 33 , 64 ]. While we expected the microbiomes of grass-forb pairs to be a mixed community sourced from both forbs and grasses, we also expected the joint rhizosphere to be dominated by the microbes of the dominant competitor as in Hortal et al. [ 18 ] and Lozano et al. [ 19 ]. In partial support of this, we found that four out of the six families that varied both between forbs and grasses and between individuals and pairs came predominantly from grasses and increased in abundance relative to forbs, while the one family that was majority sourced from forbs saw a decrease in abundance. At the community level, however, we found that the novel assemblage of bacteria in pairs was sourced nearly equally from both groups. These results suggest that the invasions of grasses into native forb habitats may be associated with microbiome shifts in both groups; however, these changes may have more negative consequences for native forbs than invasive grasses. Consistent with invasive annual grasses aiding in native forb declines across California annual grasslands, we found that grasses outperformed forbs in paired pots despite no differences when grown alone. Plants were well-watered and harvested before shading could become problematic, therefore differences in performance likely resulted from below-ground interactions. Given the large role played by microbes in below-ground interactions, the lack of a relationship between grass and forb performance in pairs, and the relationship between microbial abundance and plant performance in pairs, we speculate that invasive grass dominance over native forbs in our experiment was partially mediated by microbes. Invaded grassland soils tend to be depleted in plant available nitrogen due to the nutrient-demanding nature of invasive grasses [ 32 , 62 , 63 ], but with increased microbial activity and faster nitrogen cycling [ 24 , 31 , 63 ]. These microbe-driven changes may selectively benefit fast-growing grasses over natives, especially in nutrient-poor serpentine soils where added nitrogen has been found to preferentially aid invasives [ 51 , 65 ]. Invasive grasses may further benefit from decreased microbial diversity, which has been linked to both decreased forb nitrogen uptake and increased grass nitrogen uptake [ 66 ]. Beyond competition for nutrients, grasses and forbs may also indirectly compete through recruitment of microbes harmful to natives [ 67 ] and/or helpful to invasives [ 36 ]. Other studies in serpentine grasslands have shown that soil primed with invasive grass negatively affected native forb growth due to changes in the microbial community [ 68 , 69 ]. Although future work should be conducted to tease apart the role of plant density on microbial shifts and to understand the degree to which microbial shifts drive competitive outcomes in this system, our results add to a body of evidence that the invasion of grasses alters the microbiome and suggests that their dominance may be linked to these shifts. Plant performance in pairs could not be explained by the performance of their neighbor. Instead, plant performance in pairs was tied to the abundance of key bacterial families. We found links between bacterial regularized log transformed counts (i.e., normalized abundance) and plant biomass in grass-forb pairs for five of the six main differentially abundant families: three that were correlated with decreased forb performance: Burkholderiacae (marginal), Clostridacaeae_1 (marginal), and Fibrobacteraceae, three that were correlated with increased grass performance: Burkholderiaceae, Clostridiaceae_1 (marginal), and Veillonellaceae (marginal), and one that was correlated with both decreased grass performance and increased forb performance: Methylophilaceae. Methylophilaceae have been found in soils with high heavy metal content and include microbes critical for heavy metal attenuation and immobilization [ 70 , 71 ]. Members of Methylophilaceae also form beneficial symbioses with plants [ 72 – 74 ]. This family may include important taxa for serpentine-adapted species with roles in heavy metal attenuation and plant growth [ 73 ]. Forbs contributed a higher percentage of ASVs from this family to the joint grass-forb rhizosphere microbiome, further supporting close associations between native forbs and this group. If interactions with invasive grasses are driving declines in this family as our results suggest, grass dominance in this system may be partially due to a decrease in locally adapted microbes similar to Cavalieri et al. [ 30 ]. Grasses were likely driving higher normalized abundances of Fibrobacteraceae, Veillonellaceae, and Clostridiaceae_1, as they contributed a higher percentage of ASVs from these families to the joint rhizosphere microbiome in pairs. Although Fibrobacteraceae is typically lower in abundance in soils with high heavy metal content [ 71 ], this family is also positively associated with invasive grass dominance [ 75 ]. Furthermore, all three families are known for their cellulose-degrading properties [ 76 – 78 ], and therefore may contain important taxa for organic matter degradation [ 79 ]. Both Fibrobacteraceae and Clostridiaceae_1 also contain known nitrogen-fixers [ 80 – 82 ]. If taxa in these families are increasing nutrient availability, they could be disproportionately helping fast-growing invasive annual grasses [ 51 , 65 ]. Burkholderiaceae are considered keystone members in grasslands as endophytes or pathogens [ 83 – 86 ] and are known for their antimicrobial properties [ 87 – 89 ], nitrogen-fixing abilities [ 80 , 90 ], and competitive dominance, especially in N-limited environments [ 91 ]. Some members are even known to suppress fungal pathogens [ 92 , 93 ]. It is possible that the Burkholderiacae observed in pairs here are either providing positive functional benefits to grasses or are less harmful to grasses than forbs as pathogens. In contrast to our bacterial findings, we found no significant differences in fungal communities across treatments, either in terms of structure or diversity. This suggests loose fungal associations with plant hosts in our study, possibly due to functional redundancy of fungi across large geographic scales [ 94 ]. Within annual grasslands, however, there is evidence of invader-driven fungal shifts [ 61 ] and increases in fungal relative abundance [ 95 ]. Our limited fungal findings may indicate that fungi are generally less important than bacteria for competitive outcomes between plants [ 96 ]. Alternatively, fungi may be important, but complex fungal-fungal interactions at the community level may cancel out fitness effects on plants by beneficial and detrimental fungi [ 97 ]. In addition, experimental constraints such as unaccounted fungal regional source pools (such as airborne local spores or local watersheds [ 94 ]), plant age and development at time of sampling [ 16 ], and methodology (e.g., sampling rhizosphere here instead of root endophytes (e.g., Emam et al. [ 98 ]) or fungi in the rhizoplane (e.g., Edwards et al. [ 1 ])), may have prevented the observation of interactions between the fungal community and plant competition."
} | 3,907 |
35548758 | PMC9085526 | pmc | 1,247 | {
"abstract": "Gamma (γ)-terpinene, a monoterpene compound, which is generally used in the pharmaceutical and cosmetics industries, due to its physical and chemical properties, is expected to become one of the more influential compounds used as an alternative biofuel in the future. It is necessary to seek more sustainable technologies such as microbial engineering for γ-terpinene production. In this study, we metabolically engineered Escherichia coli to produce γ-terpinene by introducing a heterologous mevalonate (MVA) pathway combined with the geranyl diphosphate synthase gene and γ-terpinene synthase gene. Subsequently, the culture medium and process conditions were optimised with a titre of 19.42 mg L −1 obtained. Additionally, in-depth analysis at translation level for the engineered strain and intermediate metabolites were detected for further analysis. Finally, the fed-batch fermentation of γ-terpinene was evaluated, where a maximum concentration of 275.41 mg L −1 with a maintainable feedstock of glycerol was achieved.",
"conclusion": "4. Conclusions In summary, we have demonstrated the feasibility of producing γ-terpinene in metabolically engineered E . coli in fermentation conditions using glycerol as a carbon source. In this study, γ-terpinene was successfully produced by assembling a biosynthetic pathway using the methylerythritol 4-phosphate or heterologous MVA pathway and by combining the GPPS2 gene and TPS2 gene in E . coli . Subsequently, the culture medium and process conditions were optimised to further enhance the titre of γ-terpinene production, yielding 19.42 mg L −1 . Finally, we also evaluated the fed-batch fermentation of γ-terpinene using the optimised culture medium and process conditions, which successfully, when scaled up, reached a total production of 275.1 mg L −1 of γ-terpinene, 1.9 mmol mol −1 glycerol to γ-terpinene. This study has provided a sustainable strategy to produce γ-terpinene using glycerol as carbon source and has laid the foundation for future industrial production of monoterpenes with glycerol fermentation platforms.",
"introduction": "1. Introduction Monoterpenes (C 10 H 16 ) are compounds that are part of the terpene class, consisting of two isoprene units. They are known to form a remarkably sophisticated defence system in plants in the fight against predators such as insects and herbivores, 1,2 as well as having a number of other production purposes such as a hydrocarbon antioxidants, 3–5 in fragrances, 6 as fine chemicals and in medicinal products. 7–11 Recently, an increased interest in the production of transportation fuels from renewable resources has catalysed many research endeavours focussing on developing microbial systems for production of these natural resources. 12–14 In particular, monoterpenes are considered to be the best candidates as precursors in the production of alternative aviation fuels. Previously, it has been reported that the hydrogenated monoterpenes, limonene and pinene, enhanced cold weather performance of jet fuel mixtures. It was shown that limonene could match the volumetric energy, whereas pinene could match the net heat of combustion characteristics of the aviation turbine fuel JP-10. 15,16 Gamma (γ)-terpinene (CAS: 99-85-4), is also a known potential biofuel alternative; is synthesised from the cyclisation of the geranyl diphosphate (GPP); 17–22 and is produced from either the methylerythritol 4-phosphate pathway (MEP) or the mevalonate (MVA) pathway ( Fig. 1 ). 23 The current isolation technique of γ-terpinene from plants and citrus fruit is inefficient, requiring substantial expenditure of natural resources as well as the use of a number of environmentally hazardous chemicals. 24,25 Escherichia coli has well established metabolic pathways, simple genetic manipulation technology, and its fermentation technology is approaching maturity. 26 Therefore, a more sustainable technology using E . coli for the production of γ-terpinene is required to allow it to be considered as a viable bio-based advanced biofuel in the future. Fig. 1 γ-Terpinene biosynthesis pathway. γ-Terpinene was produced from the MEP pathway and MVA pathway, respectively. Pathway intermediates: GAP, glyceraldehyde 3-phosphate; PYR, pyruvate; DXP, 1-deoxy- d -xylulose 5-phosphate; MEP, 2 C -methyl-D-erythritol 4-phosphate; CDP-ME, 4-diphosphocytidyl-2 C -methyl-D-erythritol; CDP-MEP, 2-phospho-4-(cytidine-5′-di-phospho)-2 C -methyl-D-erythritol; MEcPP, 2 C -methyl-D-erythritol 2,4-cyclodiphosphate; HMBPP, 1-hydroxy-2-methyl-2( E )-butenyl 4-diphosphate; IPP, isopentenyl diphosphate; DMAPP, dimethylallyl diphosphate; GPP, geranyl diphosphate; A-CoA, acetyl-CoA; AA-CoA, acetoacetyl-CoA; HMG-CoA, 3-hydroxy-3-methylglutaryl-CoA; MevP, mevalonate 5-phosphate; MevPP, mevalonate 5-diphosphate. Enzymes in MEP pathway: DXS, DXP synthase; DXR, DXP reductoisomerase; MCT, CDP-ME synthase; CMK, CDP-ME kinase; MDS, ME-cPP synthase; HDS, HMBPP synthase; HDR, HMBPP reductase; IDI, IPP isomerase. Enzymes in MVA pathway: MvaE, acetyl-CoA acetyltransferase/HMG-CoA reductase; MvaS, HMG-CoA synthase; ERG12, mevalonate kinase; ERG8, phosphomevalonate kinase; ERG19, mevalonate pyrophosphate decarboxylase; IDI1, IPP isomerase. γ-Terpinene synthase (TPS) belongs to the monoterpene cyclases family, and is the catalysis responsible for the conversion of GPP into γ-terpinene. 27,28 Currently, TPS extraction has been reported from Eucalyptus , Origanum vulgare , Citrus limon , Citrus unshiu , Thymus vulgaris , and Thymus caespititius . Furthermore, it has been shown that TPS extracted specifically from Coriandrum sativum , C . limon , and O . vulgare was able to catalyse the conversion of GPP into γ-terpinene as well as a number of other products, such as α-thujene, myrcene, sabinene, limonene, pinene, linalool, α-phellandrene, ocimene, cymene, α-terpinolene, α-thujene and α-terpinene. 29–31 Moreover, the microbial production of GPP was available in E . coli . Therefore, when TPS was introduced into E . coli , the microbial production technique for γ-terpinene was able to be constructed. Additionally, to ensure sustainable production of γ-terpinene by a microorganism like E. coli , the ability to use a renewable carbon source such as glycerol is necessary, along with a more developed and optimised fermentation medium and process conditions. 32–34 Based on our previously work, 33,34 in this study γ-terpinene was produced by assembling a biosynthetic pathway in an engineered E. coli strain, using the heterologous MVA pathway combined with the addition of the acetyl-CoA acetyltransferase/HMG-CoA reductase gene ( mvaE ), HMG-CoA synthase gene ( mvaS ) from Enterococcus faecalis , geranyl diphosphate synthase gene ( GPPS2 ) gene from Abies grandis and TPS2 gene from T . vulgaris , due to its catalytic specificity. Additionally, on the chromosome of this particular E . coli strain the mevalonate kinase gene ( MK / ERG12 ), phosphomevalonate kinase gene ( PMK / ERG8 ), mevalonate pyrophosphate decarboxylase gene ( PMD / ERG19 ) and IPP isomerase gene ( idi / IDI1 ) were present. 35 Finally, fed-batch fermentation of γ-terpinene was evaluated using the optimised culture medium and process conditions. This study has begun the necessary foundations needed for a more sustainable future method of γ-terpinene production.",
"discussion": "3. Results and discussion 3.1 The effect of different strains In order to confirm that the observed difference in γ-terpinene production was indeed related to the heterogeneous expression of the TPS2 gene, a number of strains were constructed in this study ( Table 1 ). We used the amino acid sequence of the TPS2 gene from T . vulgaris to optimise the gene for the preferred codon usage of E . coli and had the gene synthesised by Sangon Biotech (Shanghai, China). Initial testing of the pACYC– TPS2 plasmid (pHW1) in the E. coli strain BL21 (DE3) for the production of γ-terpinene only yielded trace amounts with the MEP pathway. In order to begin the improvement of the production of γ-terpinene, the GPPS2 gene from A . grandis was added into the plasmid to improve the efficiency of the conversion of DMAPP to GPP and generate the plasmid pACYC– GPPS2 – TPS2 (pHW2) and to form the strain HW2. Following on from this strain, the mvaE and mvaS genes from E . faecalis were added to form pACYC– mvaE – mvaS – GPPS2 – TPS2 (pHW3) and to build the strain HW3. HW3 resulted in a 2-fold increase in the production of γ-terpinene when compared with the previous strain, HW2. Subsequently, to generate HW4, the plasmids pHW3 and pTrc-low 34 which contains the MVA pathway genes ERG12 , ERG8 , ERG19 , and IDI1 from Saccharomyces cerevisiae were used. The yield of the HW4 strain increased significantly, with it being 15 times higher than that of the HW3 strain. Finally, the E . coli strain CIBTS1756 ( ref. 35 ) was transformed with pHW3 to construct the final strain, HW5. As illustrated in Fig. 2 , the production of γ-terpinene in HW4 and HW5 was clearly higher than that of the other strains HW1, HW2, and HW3. Furthermore, the HW4 strain, which also contained the plasmid pTrc-low, increased the production of γ-terpinene more than 15-fold compared with HW3. The pTrc-low plasmid had low stability in the HW4 strain, therefore to avoid the loss of the pTrc-low plasmid and to also reduce the use of the antibiotics, we integrated the pTrc-low plasmid into the E . coli chromosome of the HW5 strain. This integrated strain resulted in a 9.0% increase in γ-terpinene production compared to the previous strains. Fig. 2 γ-Terpinene production in the varying E. coli strains. HW1, HW2, and HW3 strains contained the plasmid pHW1, pHW2, and pHW3, respectively. HW4 contained two plasmids, pTrc-low and pHW3. HW5 was the CIBTS1756 strain with the plasmid of pHW3. Each strain for γ-terpinene production was cultured in a shake-flask containing 20 g L −1 glycerol, 5 g L −1 yeast extract, 9.8 g L −1 K 2 HPO 4 , 0.3 g L −1 ferric ammonium citrate, 2.1 g L −1 citric acid monohydrate, 0.1 g L −1 MgSO 4 and 1 mL L −1 trace element solution. Protein production was induced with 0.1 mM IPTG and 100 μg mL −1 ampicillin, 34 μg mL −1 chloramphenicol was added when required. 3.2 The effect of different carbon sources Carbon is the main feedstock in most fermentation media and is known to be quite influential in the process, therefore finding an efficient, as well as relatively sustainable carbon source for the production of γ-terpinene was crucial. 37 In this study, we investigated the use of 40 g L −1 glycerol (C 3 H 8 O 3 ) or 20 g L −1 glucose (C 6 H 12 O 6 ) to culture the strain HW5. Previously, it was been established that the conversion of glycerol to acetoacetyl-CoA is different from that of glucose, with both generating a number of acidic products such as formate, lactate, and acetate during fermentation leading to changes in the pH of the fermentation media. 36,38 Therefore, further study to develop a better understanding of the effects of the different carbon sources on the pH during the production of γ-terpinene was required. The effects of the different carbon sources on the pH of the media are shown in Fig. 3A . Fig. 3 Effects of fermentation source and culture conditions on γ-terpinene production. (A) The effect of varying carbon sources, glycerol and glucose, on both pH and γ-terpinene production in the strain HW5. (B) The effect of varying concentration of magnesium. (C) The effect of the concentration of IPTG on γ-terpinene production. When the OD 600 of the culture reached 0.6–0.8, the fermentation broth was induced for 24 h using varying concentrations of IPTG in a shake-flask. At pH 6.0, the production of γ-terpinene with glycerol as the carbon source during fermentation yielded 3.26 ± 0.02 mg L −1 a 3.64-fold increase on that yielded by fermentation with glucose as the carbon source. However, at pH 7.5, the production of γ-terpinene with glucose as the carbon source during fermentation yielded 1.13 ± 0.01 mg L −1 , 3.03-fold higher than that produced with glycerol as the carbon source under the same conditions. These results indicate that the use of glycerol during fermentation is more suitable at the lower pH when compared to glucose and therefore was used in future work. These results clearly highlight the different effects of using glycerol or glucose as the sole carbon source during fermentation. 3.3 The effect of Mg 2+ Divalent cations especially magnesium ions are necessary for supporting catalysis of the γ-terpinene synthetase and for also activating several glycolytic enzymes. The appropriate concentration of magnesium is highly important in the activity of the enzyme. In this case, Mg 2+ is needed to ensure the conversion of sugar to γ-terpinene during the fermentation process. 28,39,40 Moreover, Mg 2+ is essential to both physiological and biochemical functions in microorganisms, including growth, cell division, synthesis of essential fatty acids, regulation of cellular ionic levels, and maintaining membrane integrity and permeability. 41 To investigate the influence of varying concentrations of Mg 2+ on the production of γ-terpinene, various concentrations of Mg 2+ were added into the fermentation medium. As shown in Fig. 3B , the production of γ-terpinene was inversely proportional to the Mg concentration, with decreasing yield with increasing concentrations of Mg 2+ . The most suitable concentration of Mg 2+ in the production of γ-terpinene was 70 mg L −1 . However, this concentration of Mg 2+ was less than what has been established previously. This may be due to having an effect on the fermentation media pH, rather than just affecting microbial cell growth and the activity of the enzymes. 42 3.4 The effect of IPTG Various concentrations of IPTG (0.05 mM, 0.1 mM, 0.2 mM, 0.3 mM, and 0.4 mM) were added into the fermentation media, to develop the most efficient induction of γ-terpinene production. Once the OD 600 reached 0.6–0.8, the culture temperature was decreased from 37 °C to 30 °C and incubated for 24 h. A gas sample from the headspace of the sealed cultures was then used in quantification. The induction of γ-terpinene production at the various concentrations of IPTG is shown at Fig. 3C . We found that at low concentrations of IPTG, HW5 was most efficient at producing γ-terpinene, with a yield of 5.13 mg L −1 with an optimised IPTG concentration of 0.1 mM. However, it is known that induction with IPTG can have an impact on the cell biomass, expression yields, as well as affecting plasmid stability, 43,44 therefore future work should focus on using a different inducer to lower the stress response of the cell. 45 3.5 Toxicity of commercial γ-terpinene to E . coli Throughout the induction and subsequent production of γ-terpinene, as its concentration increased, an effect on the growth of the E. coli production strain was noted. In order to understand the effect of the concentration of γ-terpinene on the E. coli production strain further, we measured the OD 600 every 3 h for 36 h in 50 mL of LB medium in sealed shake-flask bottles, with various concentrations of commercially available γ-terpinene (0.5 g L −1 , 1 g L −1 , 2.5 g L −1 , and 5 g L −1 ). As seen in Fig. 4 , HW5 was found to grow more slowly at the 2.5 and 5 g L −1 concentrations of commercially available γ-terpinene, with an inhibition rate (IR) at the 12 th hour of growth of 74.9% and 78.3%, respectively. These results indicate that γ-terpinene has the ability to slow down and even inhibit the growth of HW5. Therefore, the toxicity caused by the presence of γ-terpinene, at various concentrations on the E. coli production strain is a key factor to address in the future to ensure efficient production. Fig. 4 The growth of strain HW5 in LB medium with different concentration of commercially available γ-terpinene. The growth (OD 600 ) was monitored every 3 h for 36 h. Concentrations of γ-terpinene were added to the LB medium as follows: 0 g L −1 (■), 0.5 g L −1 (●), 1 g L −1 (▲), 2.5 g L −1 (▼) and 5 g L −1 (◀). The toxicity of overexpressed proteins has attracted much research attention recently, with it having a major effect on the host biomass as well as the production of certain biochemicals. 33,46,47 Research has focussed on increasing product tolerance in the host microbe by understanding membrane fluidity and function, 48,49 the function of efflux pumps 50,51 and the significance of changed expression of particular genes of those strains that are more tolerant of protein production. 52 Furthermore, it may be a better strategy to improve the hydrocarbon biofuels tolerance in the microbial host. 53 3.6 Fermentation for γ-terpinene production in both shake-flask and fed-batch cultures The optimised carbon source, the Mg 2+ concentration IPTG and pH culture conditions yielded 5.13 mg L −1 of γ-terpinene. Following on from this, we attempted to control the concentration of glycerol in the fed-batch fermentation (0, 20, 40, 80 and 100 g L −1 ) to possibly improve the production of γ-terpinene even more so. As we hypothesised, the concentration of glycerol in the culture, had a large effect on the yield, increasing the production of γ-terpinene to 19.42 mg L −1 with 100 g L −1 glycerol ( Fig. 5 ). Due to the positive results obtained in fed-batch fermentation, we performed scaled-up experiments. In a 5 L fermenter with a working volume of 2 L in batch mode, fed-batch medium (containing glycerol) was continuously added at a 3% flow rate once protein production was induced with 0.1 mM IPTG, at an OD 600 of 20. The changes in biomass and γ-terpinene production over the course of fermentation are shown in Fig. 6 . After 3 days of fermentation, the 2 L culture reached an OD 600 of 100, with an accumulated γ-terpinene yield of 275.41 mg L −1 . Fig. 5 The effect of glycerol concentration on γ-terpinene production. Fig. 6 A time course of γ-terpinene production in the strain HW5 in fed-batch fermentation. A 5 L fermenter containing 2 L of fermentation medium was used, and the temperature was maintained at 37 °C. Once the culture reached an OD 600 of 20, protein expression was induced with 0.1 mM IPTG and the temperature was changed to 30 °C. From this process, 275.41 mg L −1 of γ-terpinene was produced. The HW5 strain enables the microbial synthesis of γ-terpinene, but future work should focus on how to modify the metabolism further and/or improve extraction technologies to get higher production. For example, formate is the main by-product of γ-terpinene glycerol fermentation, therefore if we were to knockout the pfl gene, we would then allow the metabolic flux to favour acetyl-CoA production, meaning less toxic products. 35 Also, the yield was found to be proportional to the concentration of glycerol in the fed media, however, the low conversion efficiency of this as the carbon source during the fermentation process was hard to overcome. In future work, it may possible that the glycerol conversion efficiency could be increased with the presence of additional genes in the system including sldAB , gldA , dhaKLM , glpK , and glpD . 54–57 Furthermore, further study of glycerol-tolerant strains is needed to improve the balance of glycerol fermentation and yield or protein. 58 Finally, the extraction methods used in the isolation of γ-terpinene may be altered to include dodecane or poly-α-olefin in situ extraction technology to improve the yield of γ-terpinene in fermentation production. 59,60 3.7 Translation level analysis with SDS-PAGE In order to more depth analysis of the perturbation of the engineered strain to provide data for further improvement in the efficiency of our engineered strain for the production of γ-terpinene. SDS-PAGE was chosen to analyze the strain at the protein level. As shown in Fig. 7 , the proteins TPS2 (63.3 kDa), GPPS2 (41.3 kDa), mvaE (87.5 kDa), mvaS (43.4 kDa), ERG12 (48.4 kDa), ERG8 (50.5 kDa), ERG19 (44.0 kDa) and IDI1 (33.4 kDa) were successfully expressed. The results indicated that the expression levels of four proteins (ERG12, ERG8, ERG19 and IDI1) in the MVA downstream pathway were relatively high in the HW5 strain. However, protein expression levels of GPPS2 and TPS2 were found to be relatively low. It is known that the high expression of GPPS2 and TPS2 proteins is needed for the efficient production of C10 compound γ-terpinene. 61 Therefore, future experiments will focus on optimizing the GPPS2 and TPS2 genes to increase the efficiency of the strain for the production of γ-terpinene. Fig. 7 SDS-PAGE analysis of protein products. M: the protein marker; CK: negative control BL21 (DE3) with pACYCDuet1; HW5: CIBTS1756 with pACYC– mvaE – mvaS – GPPS2 – TPS2 . 3.8 Analysis of intermediate metabolites The monitoring of intermediate metabolites is helpful in further engineering of the engineered strains. In this study, the metabolites of all the engineered strains were detected by HPLC. As shown in Table 2 , succinic, lactic, acetic, mevalonic and butanediol were the main by-products. The results indicated that GPPS2 gene in the strain HW2 could cause the increment of lactic and butanediol compared to the strain HW1, which might cause by carbon flowing to the anaerobic respiratory chain. When the two upstream genes ( mvaE and mvaS ) were introduced into the strain HW2 to form the strain HW3. The mevalonate produced by strain HW3 was nearly five times more than that of strain HW2. The accumulation of mevalonate showed that the two genes are able to play a crucial role in the MVA pathway. Subsequently, the downstream genes of MVA pathway ( ERG12 , ERG8 , ERG19 and IDI1 ) enhanced (HW4) to convert the accumulated mevalonate into γ-terpinene. Finally, to make the strain more stabile and reduce the use of antibiotics, the four downstream genes were integrated into the E . coli chromosomes (HW5). As shown in Table 2 , the accumulation of mevalonate was decreased, meanwhile the butanediol increase notably. Butanediol is a reductive by-product, which indicated that the MVA pathway is more energetically friendly. Just like the production of isoprene in E . coli through MVA pathway, redundant NAD(P)H might accumulate in the cell, which might cause metabolic disturbance to the host. 62 Therefore, how to equilibrate the redox balance in E . coli is vitally for further engineering of the engineered strain. Analysis of intermediate metabolites Succinic acid (mg L −1 ) Lactic (mg L −1 ) Acetic (mg L −1 ) Mevalonate (mg L −1 ) Butanediol (mg L −1 ) HW1 437.01 ± 6.62 400.67 ± 12.09 889.54 ± 35.16 125.23 ± 2.63 4.11 ± 0.35 HW2 397.18 ± 27.09 765.06 ± 46.66 629.55 ± 84.91 122.58 ± 3.41 9.25 ± 6.16 HW3 324.47 ± 17.13 847.58 ± 47.62 316.58 ± 10.56 603.95 ± 22.44 6.46 ± 2.03 HW4 11.57 ± 0.48 0.00 ± 0.00 479.63 ± 20.14 119.38 ± 10.32 36.30 ± 1.80 HW5 63.00 ± 16.93 0.00 ± 0.00 909.37 ± 36.48 119.60 ± 2.53 51.25 ± 1.57"
} | 5,763 |
29745651 | null | s2 | 1,248 | {
"abstract": "Biology often provides the inspiration for functional soft matter, but biology can do more: it can provide the raw materials and mechanisms for hierarchical assembly. Biology uses polymers to perform various functions, and biologically derived polymers can serve as sustainable, self-assembling, and high-performance materials platforms for life-science applications. Biology employs enzymes for site-specific reactions that are used to both disassemble and assemble biopolymers both to and from component parts. By exploiting protein engineering methodologies, proteins can be modified to make them more susceptible to biology's native enzymatic activities. They can be engineered with fusion tags that provide (short sequences of amino acids at the C- and/or N- termini) that provide the accessible residues for the assembling enzymes to recognize and react with. This \"biobased\" fabrication not only allows biology's nanoscale components (i.e., proteins) to be engineered, but also provides the means to organize these components into the hierarchical structures that are prevalent in life."
} | 273 |
35695425 | PMC9426491 | pmc | 1,249 | {
"abstract": "ABSTRACT Improving the availability of representative isolates from the coral microbiome is essential for investigating symbiotic mechanisms and applying beneficial microorganisms to improve coral health. However, few studies have explored the diversity of bacteria which can be isolated from a single species. Here, we isolated a total of 395 bacterial strains affiliated with 49 families across nine classes from the coral Pocillopora damicornis . Identification results showed that most of the strains represent potential novel bacterial species or genera. We also sequenced and assembled the genomes of 118 of these isolates, and then the putative functions of these isolates were identified based on genetic signatures derived from the genomes and this information was combined with isolate-specific phenotypic data. Genomic information derived from the isolates identified putative functions including nitrification and denitrification, dimethylsulfoniopropionate transformation, and supply of fixed carbon, amino acids, and B vitamins which may support their eukaryotic partners. Furthermore, the isolates contained genes associated with chemotaxis, biofilm formation, quorum sensing, membrane transport, signal transduction, and eukaryote-like repeat-containing and cell-cell attachment proteins, all of which potentially help the bacterium establish association with the coral host. Our work expands on the existing culture collection of coral-associated bacteria and provides important information on the metabolic potential of these isolates which can be used to refine understanding of the role of bacteria in coral health and are now available to be applied to novel strategies aimed at improving coral resilience through microbiome manipulation. IMPORTANCE Microbes underpin the health of corals which are the building blocks of diverse and productive reef ecosystems. Studying the culturable fraction of coral-associated bacteria has received less attention in recent times than using culture-independent molecular methods. However, the genomic and phenotypic characterization of isolated strains allows assessment of their functional role in underpinning coral health and identification of beneficial microbes for microbiome manipulation. Here, we isolated 395 bacterial strains from tissues of Pocillopora damicornis with many representing potentially novel taxa and therefore providing a significant contribution to coral microbiology through greatly enlarging the existing cultured coral-associated bacterial bank. Through analysis of the genomes obtained in this study for the coral-associated bacteria and coral host, we elucidate putative metabolic linkages and symbiotic establishment. The results of this study will help to elucidate the role of specific isolates in coral health and provide beneficial microbes for efforts aimed at improving coral health.",
"introduction": "INTRODUCTION Genetic and genomic approaches have provided an in-depth understanding of the coral-associated microbiome. However, despite the diversity of the coral microbiome being well described ( 1 ), the functions these communities play regarding host fitness and health, or indeed the role of single members, remain less well understood ( 2 – 4 ). To this end, high-quality assembled reference genomes derived from metagenomic analyses are now desirable, thus enabling a better understanding of the integrated metabolic links of the coral holobiont ( 5 ). However, the often-close symbiotic relationships bacteria have with their coral hosts make it difficult to eliminate contaminating coral DNA for subsequent analyses ( 6 ). Studies now, however, are starting to overcome these issues, with recent work outlining 102 metagenome-assembled or single cell-sequenced bacterial genomes, along with 103 genomes derived from the cultured fraction of coral-associated bacteria (summarized in Data Set S2 at http://data.scsio.ac.cn/metaData-detail/1514896851117494272 and http://isolates.reefgenomics.org/download2/ ). Although culture-independent molecular methods are useful in describing the diversity of any given host, the isolation and characterization of cultured fractions are still vital to elucidate strain characteristics. A previous report ( 7 ) highlighted the wealth of information which can be mined from such collections, identifying over 3,000 coral-associated bacteria held in laboratories around the world, 1,045 of which had full-length 16S rRNA gene sequences available. The study also increased the number of genomes available for these coral associates to 74 at the time of publication ( 7 ). However, there remains little understanding of culturable bacteria from a given coral species at a single time point ( 7 ). In this study, we investigate in detail the cultural fraction of Pocillopora damicornis , describing novel bacterial taxa associated with this coral species. Furthermore, through sequencing and analysis of genomes of these isolates and the coral host, we provide insight into the putative function these isolates offer to their host.",
"discussion": "DISCUSSION Diverse and novel bacteria cultured from P. damicornis . In this study, we isolated 395 bacterial strains from a single coral species, P. damicornis . There were 80 potential novel species and 17 potential novel genera in this collection, which greatly increases the currently available cultured bacteria derived from P. damicornis and corals in general. For example, in a recent effort to describe the cultured fraction of coral-associated bacteria 3,055 isolates were identified spread across 52 studies or held in laboratories in private collections ( 7 ). Only 15 of these (0.5%) were from the coral P. damicornis , and these were affiliated with seven genera ( 7 ). Further, that meta-analytical study identified 12 putatively novel genera ( 7 ). In this study, our isolates span 36 formally described genera and 17 potential novel genera (see Fig. S4 in the supplemental material). And our study focused on just one coral species, therefore further demonstrating the high species diversity of the coral microbiome (not including Symbiodiniaceae) and highlighting that cultivation approaches are underutilized in exploring the coral microbiome. 10.1128/msystems.00327-22.4 FIG S4 Cultured coral-associated bacterial genera. Bacterial genera in the light blue circle were newly obtained in this study, and those in the light pink circle were collated in the previous study ( 7 ). Genera in the overlapping circle were common to both studies. Download FIG S4, TIF file, 0.3 MB . Copyright © 2022 Li et al. 2022 Li et al. https://creativecommons.org/licenses/by/4.0/ This content is distributed under the terms of the Creative Commons Attribution 4.0 International license . In addition to the increase in available cultured coral-associated bacterial strains and 16S rRNA gene data, we also added 118 genomes of coral-associated bacteria to data banks, representing 111 bacterial species. Our efforts effectively more than double the number of coral-associated bacterial genomes now available to researchers around the world for open-source use. That said, even with this significant contribution the number of genomes available for coral-associated bacteria remains limited, and there is an urgent need to increase this resource, which would facilitate a greater understanding of the putative mechanisms for establishment, maintenance, and function of bacteria associated with their coral hosts. ELR-containing and cell-cell attachment proteins involved in symbiosis maintenance. Across the taxa of strains for which genome assembly was conducted, we observed that ELR-containing proteins were more prevalent in Flavobacteriaceae than in Actinobacteria and Alphaproteobacteria , implying potentially different degrees of association within the coral host. The two Aquimarina strains had the overall highest number of ELRs, which is somewhat similar to Aquimarina megaterium EL33, an octocoral associate ( 7 ). Four Alteromonadales strains had high numbers of WD40 and tetratricopeptide repeats, which again confirms previous findings for other coral-associated Alteromonadales (i.e., >250 tetratricopeptide and 29 to 142 WD40 repeats [ 7 ]). The same was found for vibrios in this and previous ( 7 ) studies, in that no or only a few ARPs were detected. In contrast to the findings of Sweet et al. ( 7 ) and a metagenomic binning study by Robbins et al. ( 5 ), we did not find high numbers of ankyrin and WD40 repeat sequence signatures in our Endozoicomonadaceae strain (SCSIO 12664). This was surprising as Endozoicomonas has been proposed to be a true symbiont in many coral species, though it likely also has a free-living life stage ( 23 , 24 ). The lower abundance or complete absence of ELRs in our novel genus may therefore imply that SCSIO 12664 is only loosely related to the coral host and/or that the bacterium utilizes alternative mechanisms to maintain a stable association with the host. That said, although the ELRs are indeed found to be prevalent in corals and other marine invertebrate bacterial symbionts, the mechanism by which ELRs actually interact with host cells and the outcomes of these interactions still remain largely unknown. Understanding symbiosis is one key goal, and analyzing both the genomes of the corals and their associated bacteria allows us to explore integrated links that may support a symbiotic lifestyle. For example, eukaryotic-like proteins are known to be used by some bacteria to manipulate host cellular processes ( 12 – 14 ), as well as mechanisms for attachment to the host tissue ( 14 , 16 ). For P. damicornis we were able to detect fibronectin type II- and III-encoding genes in the host genome, while FnBPs were found in only three of the bacterial isolates. We noticed most of our isolates actually possessed fibronectin type III-encoding genes. Fibronectin III domains are the bacterial binding sites and mediate self-interaction in eukaryotes but have also been found in bacterial FnBPs and may interact with the self-binding region of fibronectin ( 16 , 25 ). The fibronectin III domains may thus play an important role in bacterium-bacterium adhesion, and the isolates containing fibronectin III-encoding genes may also possess distinctive mechanisms allowing for the binding of host fibronectin. Further investigation is therefore warranted in this regard and would ascertain the functions of these eukaryote-like and binding proteins, ultimately unlocking the key mechanisms involved in coral-bacterial symbiont interactions. Genomic and phenotypic support for the contribution to carbon and nitrogen cycling. Besides acquiring carbon through zooplankton predation (heterotrophic feeding) and photosynthetic fixation by the endosymbiotic algae Symbiodiniaceae, several essential genes for carbon fixation pathways have been identified in the metagenome-assembled genomes of coral-associated bacteria and archaea ( 5 ). Here, we have also identified six carbon fixation pathways, and four isolates (SCSIO 12425, SCSIO 12644, SCSIO 12654, and SCSIO 12727) in particular were found to possess the complete sets of genes encoding the Calvin-Benson cycle. Our results imply the presence of bacterial carbon fixers in the holobiont (in addition to the Symbiodiniaceae). Of specific interest are the aerobic anoxygenic phototrophic bacteria (AAPB) observed in this study, which possess both pufL and pufM genes. To our knowledge, this is the first study reporting the potential presence of AAPB associated with coral based on the presence of pufLM genes, though expression of relevant gene pathways is still required to confirm whether this finding has any functional relevance to the coral holobiont. We hypothesize that these bacteria likely utilize light as an additional source of energy for metabolism ( 26 ), much the same way as the symbiotic algae, and therefore indirectly reduce the consumption of organic carbon. Additionally, light appears to stimulate and promote the cell activity of AAPB, thus promoting the carbon cycle process, which will have a knock-on effect for the host and other members of the coral microbiome ( 27 ). It should be noted that although we did not isolate any diazotrophic microbes from P. damicornis tissues, this was likely due to the methodology utilized and does not imply they are not present. Similarly, our culture approaches would not retrieve the archaeal lineages ( Thaumarchaeota ), which have been implicated in metabolism of urea/ammonia to nitrate in corals ( 5 ). That said, we were able to identify several nitrifying and denitrifying bacteria which are likely contributing to the nitrogen cycling process within the coral holobiont. While the function of coral-associated bacteria was largely inferred based on our genomic data, we also conducted biochemical assays on some of the novel isolates, and three (SCSIO 12603, SCSIO 12610, and SCSIO 12664) revealed nitrate-reducing capabilities. Interestingly, there were some inconsistencies between the phenotypic and genomic evidence for some of these isolates, which highlights the necessity of functional verification when describing the roles of key coral-associated bacteria. CAZy family and carbon metabolism provide cues for microniche adaptation. Differences in the chemical composition of labile organic matter exuded by coral holobionts influence the community and growth of associated bacteria, which can subsequently differentially enrich bacterial lineages ( 28 ). Microbial GDEs from GH23, GH73, GH103, and CE4 target the glycosidic linkage of N -acetylmuramoyl or N -acetylglucosaminyl residues in peptidoglycan, which makes up a significant proportion of the chemical composition of coral mucus ( 29 ). GH16 enzymes are active on various algal structural polysaccharides including laminarin, agar, and carrageenan ( 30 ). Glycoside hydrolases from the GH3 family are exo-acting β- d -glucosidases, α- l -arabinofuranosidases, β- d -xylopyranosidases, N -acetyl-β- d -glucosaminidases, and N -acetyl-β- d -glucosaminide phosphorylases with broad substrate specificities that carry out a range of functions including plant and bacterial cell wall remodeling, cellulosic biomass degradation, and pathogen defense ( 31 , 32 ). Our isolates revealed enriched GDEs known to target carbohydrates originating from algae and coral mucus. Notably, GDEs targeting plant cell wall carbohydrates in Bacteroidetes were found to be more abundant than those in other clades, while the frequency of sequences for GDEs targeting peptidoglycan was higher in Proteobacteria ( Fig. 4B ). Accordingly, we hypothesize distinct specializations among Bacteroidetes and Proteobacteria for their utilization of alga- and mucus-derived organic matter, respectively. Our results demonstrate that the specific potential for carbon metabolism facilitates the adaptation of the coral-associated bacteria to ecological niches with supplied carbohydrate substrates, and the carbohydrate composition of these microniches is likely to constitute one of the major driving forces that shape the local microbial community structures ( 33 ). It is now well known that there are many genes involved in several intermediary metabolic pathways which tend to be reduced in endosymbiont genomes ( 34 ). In this study, we identified two of our isolates (SCSIO 12583 and SCSIO 12664) which had low GDE frequencies, for example. Low potential for carbohydrate utilization implies a relatively minor role in cycling reef organic carbon and preference for an obligate endosymbiotic lifestyle. Further evidence of this is the presence of N -acetyl-β-glucosaminidase in strain SCSIO 12664. As coral mucins could be N -glycosylated ( 29 , 35 ), a bacterium with N -acetyl-β-glucosaminidase could therefore penetrate the host mucus layer (the coral’s first line of defense against invading bacteria [ 36 ]) and then subsequently establish endosymbiosis ( 35 ). Source of amino acids and polyamines in the coral holobiont. The ability to synthesize amino acids varies among corals, with many essential amino acids acquired from Symbiodiniaceae and other exogenous sources ( 5 , 37 ). The host genome in this study revealed that several amino acids could not be synthesized by the coral itself. Instead, P. damicornis may rely on its bacterial symbionts to assimilate these amino acids, in combination with its Symbiodiniaceae and heterotrophic feeding. Further, several isolates in this study were shown to potentially synthesize polyamines, such as putrescine and spermidine. Polyamines are involved in bacterial swarming motility and invasion ability ( 38 , 39 ), biofilm formation and disassembly ( 40 , 41 ), virulence ( 42 – 44 ), proper cell division ( 45 , 46 ), and bacterial ( 47 ) and interkingdom ( 44 ) cell-cell communications. Interestingly, polyamines have also been found enriched in diseased coral colonies ( 48 ) and appear to play important roles in abiotic stress responses ( 49 ). At present, the relevance of polyamines produced by coral-associated bacteria to coral health is unknown though certainly worth further investigation. Finally, we were also able to identify several isolates which have the potential of biosynthesizing GABA from putrescine, and a Labrenzia strain (SCSIO 12622) which may be able to synthesize GABA from arginine. GABA shows multiple physiological functions, such as acting as an inhibitory neurotransmitter in the central nervous system of animals ( 50 ), gut modulation ( 51 ), protection against stress ( 52 ), and the regulation of settlement and metamorphosis of marine invertebrate larvae ( 53 – 55 ). Our results therefore highlight potential interactions mediated by bioactive compounds produced by these coral-associated bacteria. Putative opportunistic vibrio. A previous meta-analytical study associated with culture collections of coral-associated bacteria identified that pathogenic and nonpathogenic Vibrio strains show functional separation based on their genome data ( 7 ). This offers a potentially useful tool to ascertain pathogenicity of members of this genus from genomic data and in the absence of laboratory inoculation trials. The functional profiles of the five vibrios cultured in this study demonstrated that they grouped with known pathogens (see Fig. S2 in the supplemental material), although the corals from which these bacteria were sampled showed no signs of compromised health. Interestingly, the abundant virulence factors present in these isolates also implied they are potentially opportunistic bacteria. But we still should interpret these findings with caution as the results are based only on the predicted functional profiles of these five vibrios, and further experiments are required for verifying their roles. Furthermore, we also found these five vibrios possessed multiple (between 45 and 60) putative antibiotic resistance genes, including msbA , msrB , rpsJ , and tet homologues ( 35 ). The true causal agents of the majority of coral diseases remain elusive ( 1 , 56 ), though recently amoxicillin paste has been successfully used to treat corals affected by stony coral tissue loss disease ( 57 , 58 ). The information about antibiotic resistance genes carried by the coral-associated bacteria obtained in this study may provide clues for the selection of antibiotics for both susceptible and resistant putative pathogens alongside potential beneficial communities. To conclude, in this study we have increased the number of described cultured coral-associated bacteria by 395 isolates, 80 of which appear to be novel species and 17 from novel genera. Further, we sequenced 118 genomes of these isolates along with the host, P. damicornis . Although we recognize that this is not an exhaustive list of the cultural fraction from a coral, it likely represents the most in-depth study to date (accounting for ~8% of the total bacterial fraction of the host microbiome). That said, there are many taxa detected in 16S rRNA gene amplicon libraries which remain unculturable at this time, including many Endozoicomonas species ( 8 ). We did, however, culture a novel lineage of Endozoicomonadaceae (SCSIO 12664). This study therefore provides a strong base for future genetic studies that will improve the resolution of metagenomic analyses of the coral-associated microbiome. Moreover, pure cultures of the coral-associated microbiome acquired in this study will help to elucidate the role of bacteria in the coral holobiont ( Fig. 5 ), assess the impact of specific isolates on coral health, and identify and provide beneficial microbes for efforts aimed at restoring reefs and improving coral health ( 3 , 59 ). FIG 5 Schematic representing the proposed contributions of bacterial members to the Pocillopora damicornis holobiont."
} | 5,225 |
34835306 | PMC8623284 | pmc | 1,250 | {
"abstract": "Stony coral tissue loss disease (SCTLD) is an emergent and often lethal coral disease that was first reported near Miami, FL (USA) in 2014. Our objective was to determine if coral colonies showing signs of SCTLD possess a specific microbial signature across five susceptible species sampled in Florida’s Coral Reef. Three sample types were collected: lesion tissue and apparently unaffected tissue of diseased colonies, and tissue of apparently healthy colonies. Using 16S rRNA high-throughput gene sequencing, our results show that, for every species, the microbial community composition of lesion tissue was significantly different from healthy colony tissue and from the unaffected tissue of diseased colonies. The lesion tissue of all but one species ( Siderastrea siderea ) had higher relative abundances of the order Rhodobacterales compared with other types of tissue samples, which may partly explain why S. siderea lesions often differed in appearance compared to other species. The order Clostridiales was also present at relatively high abundances in the lesion tissue of three species compared to healthy and unaffected tissues. Stress often leads to the dysbiosis of coral microbiomes and increases the abundance of opportunistic pathogens. The present study suggests that Rhodobacterales and Clostridiales likely play an important role in SCTLD.",
"conclusion": "5. Conclusions The goal of this study was to determine if the lesions of coral colonies showing signs of stony coral tissue loss disease (SCTLD) had a consistent microbial signature across five different coral species: Colpophyllia natans , Pseudodiploria strigosa , Montastraea cavernosa , Orbicella faveolata , and Siderastrea siderea . Diversity indices revealed that the lesion (DL) tissue of all but one species ( P. strigosa ) had a higher microbial beta diversity dispersion than apparently healthy (AH) tissue. In three species ( M. cavernosa , O. faveolata , and S. siderea ), DL tissue had a higher beta diversity dispersion than both AH tissue and unaffected (DU) tissue. Furthermore, DL tissue of every species had a higher species richness compared to AH tissue. In three species ( C. natans , M. cavernosa , and S. siderea ), the species richness of DL tissue was also higher than DU tissue. DL tissue consistently had higher relative abundances of the order Rhodobacterales compared to AH and DU tissues, except for S. siderea , a finding that has been observed in other coral species [ 39 , 40 , 51 , 52 ]. In addition, order Clostridiales was enriched in the DL tissue of three of the five species investigated in the present study, suggesting that Clostridiales may also play an important role in SCTLD. The beta diversity dispersion and species richness of water samples did not differ between the vulnerable and epidemic zones, suggesting that a bacterial signature of SCTLD was not detected in the water column, potentially due to sampling location within the sites. The results presented herein expand our understanding of SCTLD as it relates to the coral microbiome of critical reef-building species along Florida’s Coral Reef. Understanding how SCTLD, coupled with changing environmental conditions, can affect host-microbe interactions is an important step towards developing practical and effective disease mitigation strategies.",
"introduction": "1. Introduction Corals host a diversity of microorganisms composed of viruses, fungi, archaea, endolithic algae, protozoa, bacteria, and algal symbionts [ 1 , 2 , 3 , 4 , 5 ]. Microbes living in or on a coral can be beneficial to the coral host by playing important roles in the cycling and recycling of nutrients [ 6 , 7 , 8 ], the production of amino acids [ 8 , 9 ], protection against pathogens [ 10 , 11 ], and enhancing larval settlement and metamorphosis [ 12 ]. Disturbances, such as those caused by climate change, can alter coral–microbe interactions and lead to bleaching, disease, and the mortality of the coral host [ 13 , 14 , 15 ]. Climate change, overfishing, and pollution are among the many stressors contributing to the decline of coral reef health worldwide [ 16 , 17 ]. As a result of these stressors and others (e.g., sedimentation), there has been an increase in widespread bleaching events, disease incidence, and subsequent mortality among coral communities in recent decades [ 18 , 19 , 20 , 21 , 22 ]. In 2014, a novel coral disease was documented off Southeast Florida, USA, and has since spread through Florida’s Coral Reef (FCR) and across much of the Caribbean region [ 23 , 24 ]. Stony coral tissue loss disease (SCTLD) is believed to affect more than 23 species of scleractinian corals [ 25 ]. Signs of active SCTLD include focal or multifocal lesions moving at chronic to acute rates, followed by partial or whole colony tissue loss, often resulting in complete mortality of coral colonies [ 23 , 26 , 27 , 28 , 29 ]. While much remains unknown about SCTLD, advances in our understanding of the etiology [ 30 ], spatial epidemiology [ 31 , 32 , 33 , 34 ], diagnostics [ 35 ], and treatment [ 29 , 36 , 37 ] of SCTLD have been achieved. Corals, especially many of the major Caribbean reef-building species, exposed to SCTLD experience high mortality [ 23 , 38 ]. For example, >97% of Meandrina meandrites and Dichocoenia stokesii colonies monitored off Miami-Dade County, FL, USA, died in the year following the onset of the SCTLD outbreak [ 23 ]. SCTLD is presumed to spread through waterborne transmission and by direct contact [ 26 ], and barotropic oceanic currents correlate with the spatio-temporal progress of the disease throughout FCR [ 31 ]. The spatio-temporal dynamics indicate that SCTLD follows a contagion model over both large [ 32 ] and small [ 34 ] spatial scales, again potentially indicating a novel pathogen driving disease dynamics. A causative agent of SCTLD has not yet been identified, making it challenging to identify possible vectors or intermediate hosts. However, a recent study isolated Vibrio coralliilyticus from some active SCTLD lesions and revealed that V. coralliilyticus may play an opportunistic role in exacerbating the disease [ 35 ]. Thus, bacteria play an important, possibly secondary, role in lesion advancement in SCTLD [ 30 ], and identifying the microbes associated with SCTLD lesions is critical for identifying the key microbiota likely involved in disease progression. Recent studies have used high-throughput 16S rRNA sequencing to describe the microbiomes of active SCTLD lesions. Meyer et al. [ 39 ] compared the microbial community compositions of four species of coral, each displaying signs of active SCTLD: Montastraea cavernosa , Orbicella faveolata , Diploria labyrinthiformis , and Dichocoenia stokesii . The following bacterial orders were enriched in SCTLD lesions of all but one coral species ( O. faveolata ): Flavobacteriales, Clostridiales, Rhodobacterales, Alteromonadales, and Vibrionales. Rosales et al. [ 40 ] identified two bacterial orders, Rhodobacterales and Rhizobiales, in the active lesions of Stephanocoenia intersepta , Diploria labyrinthiformis , Dichocoenia stokesii , and Meandrina meandrites that were more prevalent compared to the microbiomes of the tissue of apparently healthy colonies and the unaffected tissues of diseased colonies. Water samples collected at sites with SCTLD also had relatively high abundances of Rhodobacterales compared to sites with no signs of SCTLD, and both water and sediments shared sequences with lesions of diseased corals [ 40 ]. The objectives of the present study were to determine if a specific microbial signature exists in the lesions of corals experiencing SCTLD by characterizing the microbial community (1) between disease outbreak zones (epidemic and vulnerable zones), (2) among sites within each zone, and (3) among coral sample tissue types (lesion and unaffected tissues from diseased colonies, and tissue from apparently healthy colonies). To evaluate the microbial diversity and composition of corals affected by SCTLD, tissues were collected from five coral species; Colpophyllia natans , Pseudodiploria strigosa , Montastraea cavernosa , O. faveolata , and Siderastrea siderea . Samples were collected from sites in the Middle Florida Keys (the epidemic zone) and the Lower Florida Keys (the vulnerable zone). The SCTLD microbial signatures of M. cavernosa and O. faveolata have been characterized previously for samples collected from the endemic zone in Southeast Florida [ 39 ], allowing for species-specific comparisons across FCR. Additionally, we collected water samples from our study sites, allowing for a comparison of our results to a previous study which examined the SCTLD microbial signatures in water [ 40 ].",
"discussion": "4. Discussion The stony coral tissue loss disease (SCTLD) outbreak has caused the widespread mortality of important reef-building species in Florida and in the Caribbean Region. To date, the identity of the presumed pathogen(s) responsible for SCTLD remains unknown, despite many studies on microbial communities associated with SCTLD [ 39 , 40 , 51 , 52 ]. In the present study, the microbiomes of five coral species ( Colpophyllia natans , Pseudodiploria strigosa , Montastraea cavernosa , Orbicella faveolata , and Siderastrea siderea ) were analyzed to determine if a specific microbial signature exists across different species of corals with active SCTLD lesions. The lesions of four of these five coral species (except S. siderea ) had significantly higher abundances of Rhodobacterales, a finding that is consistent with studies of other susceptible coral species ( Table 2 ). In addition to Rhodobacterales, Clostridiales was also a significant and ubiquitous member of the lesion microbial community, especially for C. natans , M. cavernosa , and P. strigosa . 4.1. Diversity Indices In general, the microbial beta diversity dispersion between zones (i.e., vulnerable and epidemic) and within zones was not significant in either water or apparently healthy (AH) coral samples, except for those from P. strigosa , which were more dispersed in the vulnerable zone than the epidemic zone. These results, with the exception of P. strigosa , are similar to Rosales et al. [ 40 ], who found no difference in the AH coral microbial beta diversity dispersion between the vulnerable and epidemic zones. Coral microbiomes are sensitive to environmental perturbations (e.g., thermal stress, nutrient pollution), which can cause an increase in the microbial beta diversity dispersion [ 53 ]. It is possible that AH P. strigosa colonies in the vulnerable zone had recently been exposed to an environmental stressor, which caused an increase in the beta diversity dispersion. However, longitudinal studies are needed to parse out the influence of environmental conditions on beta diversity dispersion through time. The present study did not find a difference in the dispersion of microbial communities in the water column between zones. In Rosales et al. [ 40 ], the microbial beta diversity dispersion of water samples was significantly different between zones, with greater dispersion observed within the epidemic zone compared to the vulnerable zone. Even though the beta diversity dispersion of environmental samples (i.e., water samples) did not differ between zones in the present study, differences in environmental conditions between zones should not be ruled out as a possible factor in driving these differences. Additionally, the water sample sizes in the present study were lower than those used in Rosales et al. [ 40 ], potentially limiting the ability to observe meaningful differences between the zones in this study. This discrepancy between findings may also be attributed to differences in how the water samples were gathered: in the present study, water sample bottles were held approximately 0.5 m above the benthos and were not specifically gathered above coral colonies. In Rosales et al. [ 40 ], water sample bottles were held directly over the benthos, gathering water approximately 20 cm or less from the bottom, and in the epidemic zone, water samples were collected directly over colonies with SCTLD. The distance at which water samples are collected above the bottom influences the microbial signature of the water samples, as illustrated by Weber et al. [ 54 ]. Water samples collected within the coral ecosphere, or the environment immediately surrounding an individual coral colony (e.g., water < 30 cm above the colony), can have a different microbial signature compared to water collected >1 m above the reef. Therefore, bacterial beta diversity dispersion may be related to the sampling location (i.e., distance above the benthos or individual coral colonies) in the water column. Among the tissue sample types of most coral species, the epidemic zone had a dispersed microbial community. While dispersion was not significantly different among the three tissue sample types in P. strigosa , as previously identified in Diploria labyrinthiformis , Dichocoenia stokesii , and Meandrina meandrites [ 40 ], the lesion (DL) tissue of the other coral species in this study were more dispersed than AH (in four coral species) and unaffected (DU) tissue (in three coral species, except C. natans ) tissues. In Meyer et al. [ 39 ], microbial dispersion was similar between the DL tissue and DU tissue of diseased colonies from three susceptible species, Montastraea cavernosa , Diploria labyrinthiformis , and Dichocoenia stokesii . However, in M. cavernosa , the DL and DU tissues did have a higher dispersion compared to AH tissue. Similarly, in the present study, there were no significant differences between the DL tissue and the DU tissue of C. natans . Therefore, colonies showing signs of SCTLD may have a disrupted microbiome even far away from the lesion, suggesting a systemic effect [ 55 ]. It has been reported in other studies that stress, such as heat stress, often has a stochastic effect on the microbial community composition that can result in an increase in the beta diversity [ 56 ]. In Rosales et al. [ 57 ], for example, Acropora cervicornis exposed to diseased ramets had a higher beta diversity compared to control corals not exposed to disease. Eaton et al. [ 55 ] also showed that visibly unaffected areas (DU tissue) on diseased corals later showed signs of tissue loss after coral fragments were separated and isolated from the active disease border on the parent colony, and Landsberg et al. [ 30 ] found lytic necrosis characteristic of SCTLD lesions within some samples of DU tissue, again suggesting that SCTLD may be systemic within coral colonies. In the epidemic zone, there were also significant differences in the species richness among tissue sample types, with DL tissues having a higher species richness compared to both DU and AH tissues in three species ( C. natans , M. cavernosa , and S. siderea ) and compared to the AH tissue of two species ( O. faveolata and P. strigosa ). This difference in alpha diversity may be a result of microbial dysbiosis, or an imbalance in the natural microbiome that can disrupt coral–microbe interactions and lead to disease [ 8 , 58 ]. The microbiomes of corals exposed to stressful environmental conditions (e.g., acidification, and increased temperature) often experience a shift in microbial community composition and, consequently, an increase in the species richness [ 8 , 59 ]. Microbial shifts may be attributed to a loss in beneficial bacteria; thus, freeing up niche space for putative pathogens to inhabit [ 13 , 60 ]. In a study conducted by MacKnight et al. [ 58 ], disease-resistant corals exposed to white plague disease (WPD) had a higher dysbiosis threshold compared to corals that developed WPD lesions. The authors hypothesized that certain bacteria may be helping to prevent pathogens from colonizing disease-resistant corals; thus, also preventing dysbiosis and the onset of WPD. In the present study, the higher microbial alpha diversity of coral DL tissues was potentially due to the decreased stability of the coral microbiomes and the hosts’ inability to prevent pathogenic infection [ 59 ], or due to an increased propensity for diseased tissues and surface mucus to become colonized by diverse opportunistic bacteria, including Rhodobacteraceae [ 61 , 62 ]. 4.2. Differences Were Detected among Sites and Tissue Sample Types Bacterial communities of all but one species ( O. faveolata ) were different among at least three sites of the epidemic zone ( Table S1 ). Even though sites within the epidemic zone were similarly dispersed, there were significant groupings among sites, suggesting a site-level effect on the bacterial signature of corals within the epidemic zone. There are likely several factors driving these site-wide differences. In a previous study, Williams et al. [ 34 ] showed that the coral species diversity, coral cover, and size of coral colonies affect SCTLD prevalence and severity. Sites with higher abundances of M. cavernosa and O. faveolata compared to four other susceptible species ( C. natans , P. strigosa , Diploria labyrinthiformis , and Dichocoenia stokesii ) had a greater disease prevalence. In addition, colonies that ultimately became diseased were significantly larger than colonies that did not display signs of SCTLD over the course of the study. This finding was also observed by Sharp et al. [ 63 ]. Both studies [ 34 , 63 ] observed that the coral density did not likely play a role in the spatio-temporal dynamics of SCTLD. While coral density may not be a factor, differences in the coral species diversity and size of colonies among sites in the present study may explain the site-level differences observed in microbial communities. Even though the microbial communities of these corals are site-specific ( Table S1 ), there were consistent signatures within the DL tissues of corals among sites. 4.3. Rhodobacterales and Clostridiales in SCTLD Lesions The DL tissues of four species in this study ( C. natans , P. strigosa , O. faveolata , and M. cavernosa ) had significantly higher relative abundances of Rhodobacterales compared to both DU and AH tissues. In previous studies, Rhodobacterales was also differentiated in the lesions of Stephanocoenia intersepta , Diploria labyrinthiformis , Dichocoenia stokesii , Meandrina meandrites, and Montastraea cavernosa in Florida [ 39 , 40 ], and, although not significant, Rhodobacteraceae was enriched in the lesion tissues of Meandrina meandrites and O. franksi in the U.S. Virgin Islands [ 52 ]. Other studies have reported high abundances of Rhodobacterales, specifically Rhodobacter , in disease tissue of black band disease, white plague, and white band disease [ 64 , 65 , 66 , 67 , 68 , 69 ]. This group of bacteria play an important role in colonizing submerged marine surfaces and is often considered the primary and most common colonizer [ 70 , 71 ]. The fast-growing nature of Rhodobacterales allows members of this group to thrive in areas that are rich in amino acids and other nutrients [ 72 ]. These bacteria also have the ability to produce antibiotic compounds [ 73 ]. Under stressful conditions, Rhodobacterales appear to be typical and abundant opportunistic bacteria associated with corals [ 61 , 62 , 74 ]. In this study, an exception to the pattern of higher relative abundance of Rhodobacterales in DL compared to both DU and AH tissues was the coral S. siderea . The DU tissues of S. siderea had a lower abundance of Rhodobacterales compared to both DL and AH tissues, but the latter tissue sample types had statistically similar abundances. Siderastrea siderea often shows signs of SCTLD that are distinct from other susceptible species, including areas of pinkish tissue discoloration and mucus strands [ 30 ], leading to some speculation about whether this species has the same disease or a generalized stress response [ 25 ]. However, the characteristic hallmark lytic necrosis of SCTLD as described in the lesion tissues of other susceptible species has been reported in S. siderea tissue [ 30 ], so differences in how the tissue loss progresses through the colony in S. siderea (generally originating in polyp mouths instead of moving across the colony [ 30 ]) or its species-specific holobiont response may affect the gross presentation of the disease and the relative abundances of bacteria across tissue sample types compared with other species. However, it should be noted that the method of sampling for S. siderea (predominantly coring and scraping versus scraping only) may have influenced the composition of the microbial community in the samples examined, especially if less tissue and mucus were potentially sampled. Surface mucus on corals has a diverse microbial community, the composition of which changes during the course of disease [ 61 ]. Scraping with a coring device may have obtained slightly deeper tissue samples which may have a different microbial composition. Studies to evaluate the vertical and horizontal distribution and abundances of the microbial flora in SCTLD-affected colonies over time in relation to lesion progression, coral species, sampling method, and mucus quantity may resolve this question. In the present study, Clostridiales was another significantly differentiated group of bacteria in lesion tissue. ASV50 specifically was enriched in the lesion tissue of C. natans , M. cavernosa , and P. strigosa . Of these corals, C. natans and P. strigosa are highly susceptible to SCTLD, and exhibit acute to subacute tissue loss [ 25 , 30 ]; M. cavernosa is considered moderately susceptible, but the rate of lesion progression varies widely by colony, and most of the diseased colonies sampled in this study were experiencing subacute tissue loss (authors’ pers. obs.). It is possible that Clostridiales is a signature of faster lesion progression, given that this order has been documented in other studies in M. cavernosa and C. natans [ 39 , 52 ] in addition to two other highly susceptible species: Dichocoenia stokesii and Diploria labyrinthiformis [ 39 ]. However, Rosales et al. [ 40 ] did not observe significantly differentiated Clostridiales in DL tissue of Dichocoenia stokesii or Diploria labyrinthiformis , but only within DL tissues of Stephanocoenia intersepta , another moderately susceptible species [ 25 ]. Similar to Rhodobacterales, Clostridiales is commonly found in the lesion tissue of other coral diseases, such as black band disease, white plague disease, and white syndrome [ 67 , 68 , 75 ]. When colonizing, these opportunistic anaerobic bacteria can necrotize host tissue. For example, humans and animals exposed to clostridial spores (e.g., through contaminated drinking water) can contract clostridial myonecrosis (gas gangrene) which is a lethal infection that causes severe necrosis of muscle and soft tissue [ 76 ]. The initial appearance of SCTLD lesions in deeper basal body wall tissues may indicate a role for anaerobic bacterial pathogenesis, although thus far, no histological evidence has been found for co-occurring bacteria in lesion initiation [ 30 ]. However, the exact role, if any, of Clostridiales in lesion progression and tissue necrosis of coral colonies with SCTLD, should be examined among highly susceptible species."
} | 5,846 |
21555213 | null | s2 | 1,252 | {
"abstract": "The conversion of biomass to CH4 (biomethanation) involves an anaerobic microbial food chain composed of at least three metabolic groups of which the first two decompose the complex biomass primarily to acetate, formate, and H2. The thermodynamics of these conversions are unfavorable requiring a symbiosis with the CH4-producing group (methanogens) that metabolize the decomposition products to favorable concentrations. The methanogens produce CH4 by two major pathways, conversion of the methyl group of acetate and reduction of CO2 coupled to the oxidation of formate or H2. This review covers recent advances in the fundamental understanding of both methanogenic pathways with the view of stimulating research towards improving the rate and reliability of the overall biomethanation process."
} | 199 |
24256871 | null | s2 | 1,253 | {
"abstract": "Bacteria deploy a range of chemistries to regulate their behaviour and respond to their environment. Quorum sensing is one method by which bacteria use chemical reactions to modulate pre-infection behaviour such as surface attachment. Polymers that can interfere with bacterial adhesion or the chemical reactions used for quorum sensing are therefore a potential means to control bacterial population responses. Here, we report how polymeric 'bacteria sequestrants', designed to bind to bacteria through electrostatic interactions and therefore inhibit bacterial adhesion to surfaces, induce the expression of quorum-sensing-controlled phenotypes as a consequence of cell clustering. A combination of polymer and analytical chemistry, biological assays and computational modelling has been used to characterize the feedback between bacteria clustering and quorum sensing signalling. We have also derived design principles and chemical strategies for controlling bacterial behaviour at the population level."
} | 251 |
28819314 | PMC5561142 | pmc | 1,257 | {
"abstract": "Living organisms constantly maintain their structural and biochemical integrity by the critical means of response, healing, and regeneration. Inanimate objects, on the other hand, are axiomatically considered incapable of responding to damage and healing it, leading to the profound negative environmental impact of their continuous manufacturing and trashing. Objects with such biological properties would be a significant step towards sustainable technology. In this work we present a feasible strategy for driving regeneration in fabric by means of integration with a bacterial biofilm to obtain a symbiotic-like hybrid - the fabric provides structural framework to the biofilm and supports its growth, whereas the biofilm responds to mechanical tear by synthesizing a silk protein engineered to self-assemble upon secretion from the cells. We propose the term crossbiosis to describe this and other hybrid systems combining organism and object. Our strategy could be implemented in other systems and drive sensing of integrity and response by regeneration in other materials as well.",
"introduction": "Introduction In contrast to living organisms, most everyday objects we use lack the ability to respond to damage and self-heal or regenerate. In cases where the cost of fixing an object is higher than the cost of replacing it, the latter option is usually preferred, enabled by a global industry that makes mass quantities of things such as electrical appliances, furniture, clothes, cars, and other similar mass-manufactured consumer products. This reliance may have very negative long-term environmental impact. It is therefore both interesting and important to hypothesize, whether objects can be made or programmed to respond and adapt like living organisms do. As an example, fabric - prominent objects in human culture and technology - are flexible woven materials consisting of a network of natural or artificial fibres, made using diverse processes resulting in fabrics with a wide range of properties for various applications. The production of one kilogram of cotton for textile requires up to 20,000 litres of water 1 , and in addition consumes energy and chemicals (pesticides, fertilizers, etc.) leading to CO 2 emission into the atmosphere and water pollution. It is estimated that over 11 million tons of textiles are trashed annually in the U.S alone 2 . However, adaptive fabrics, or adaptive clothing, would self-heal in response to tear, and stretch or shrink as required; they could potentially also self-clean, remain protected from moisture, light, heat, and chemicals; and could potentially change their structure to allow physical protection (armor), evaporation, appearance, and more - all features exhibited by living organisms. Such fabrics could be the basis for a truly sustainable textile technology. Self-healing and regeneration are highly complex properties based on two coupled phases: a sensory phase, in which a certain threshold of breach of structural integrity is detected by the system; and a synthetic phase, in which synthesis of new material is driven in response to the breach. While sensing per se may be easy to achieve (e.g. by integrating conducting fibers into textile such that tearing would result in a measureable change in the resistivity of a fabric segment), synthesis and its coupling to sensing is more challenging, especially if the purpose is achieving continuous, low-maintenance ability to self-heal, that does not require constant refilling and tuning of fabric monomer reservoirs. Although self-healing and regeneration are critical components of sustainable textiles, reports on achieving them in artificial textile systems is extremely scarce. A recent study reported a polyelectrolyte layer-by-layer film coupled to squid ring proteins as a textile capable of suturing tears 3 . Several reports demonstrated fabrics capable of restoring their protective hydrophobic coating 4 , 5 . A conducting fiber-containing, yarn-based supercapacitor has been shown in which magnetic attraction restores lost connectivity between electrodes 6 . While clearly exhibiting homeostatic behaviors, these reports lack other critical capabilities such as making new material de novo; most utilize artificial mechanisms deviating from biological strategies. In this work we designed and studied a preliminary draft towards true biological regeneration in fabric, based on integrating a bacterial biofilm into fabric, creating a structural hybrid between object and organism. Our first step was to study the feasibility of hybridizing the fabric with the biofilm such that the latter remains viable and metabolically active. For this work we chose Bacillus subtilis for being widely-studied, easy to work with and to genetically engineer, their biolfilms are characterized and can be reproducibly made. We cultured B. subtilis biofilms embedded inside pieces of fabric, screening a range of fabrics that included various materials (animal, plant, mineral, synthetic) and weaving patterns (fiber diameter, fiber density) (Fig. 1A , Supplementary note 1 ). All fabrics were compatible with biofilm growth and maintenance, with minor detected differences in viability or activity between groups ( Fig. 1A \n ) . Interestingly, some fabrics significantly improved the biofilm growth, while others inhibited it (Fig. 1A , panels marked with asterisks). Additionally, there was a clear correlation between fabric architecture and biofilm appearance; hybrids with less dense fabrics exhibited rough-surfaced, disordered biofilms, and ones within denser fabrics exhibiting the opposite phenotype (Fig. 1B , Supplementary note 1 ). This dependency suggests that the fabric serves as a structural framework or scaffold for the biofilm, and highlights the possibility of designing specific fabrics to achieve desired biofilm phenotypes and growth patterns. Figure 1 Fabric and hybrid analysis. ( A ) Fabric architectures (1st and 5th columns) were first examined without biofilms and unfixed using scanning electron microscope (SEM, thick scale bars = 300 µ m), and analyzed by FiberMetric (2nd and 6th columns showing fiber diameter distribution. Asterisks denote fabrics of various types which improved biofilm growth: *synthetic fabric made of polyester and cotton with large-sized threads and dense weave, **natural fabric made of cotton with small-sized threads and plain weave, ***synthetic fabric made of polyester with large fibres and loose weave. Fabric-biofilm hybrids were fixed after 3 days in culture and visualized in SEM (3rd and 7th columns, thin scale bars = 100 µ m). Biofilm viability was scored each day on a score of 0–3 (0 = no growth and clear medium, 3 = full growth and cloudy medium, 4th and 8th columns). Control biofilm without fabric was scored 0–1–2–3. Results representative of 3 independent repeats. ( B ) Image analysis showing simulated roughness of biofilm in each of the three representative fabrics (* , ** , ***, from left to right), demonstrating the correlation between weaving density and biofilm structure (x and y axes, location pixels; z axis, pixel intensity on a scale of 0–255). \n In order to configure the response/synthesis role of the biofilm, the response of the biofilm to mechanical tear was mapped. While other responses have been previously reported 7 – 10 , the specific response to mechanical strain and tear, a likely natural scenario in bacterial evolution, has not. For this, total RNA was extracted from B. subtilis biofilms 5 min after subjecting them to mechanical tear, sequenced (Supplementary note 2 ) and analyzed to obtain the transcriptome response and identify tear-responsive elements. Rather than a single tear across a biofilm, and In order to maximize the signal, ~1,000 of small lateral tears were induced in the entire biofilm (average of ~3 tears per mm 2 biofilm) using a custom-built array of metal needles positioned at high density and movable on the XYZ axes. Transcriptome analysis highlighted specific pathways involved in the response to tearing, particularly cell wall remodeling (teichuronic acid and peptidoglycan biosynthesis) and cell division (phosphate uptake, nucleotide and aminoacyl-tRNA biosynthesis) (Fig. 2A , Supplementary note 3 & 4 ). Tearing also induced activation of the sigma M regulon, which has been shown to operate in response to cell wall stress induced by antibiotics and other chemical stimuli 11 . Interestingly, population control genes such as skf (sporulation killing factor) and sdpA/B (sporulation delaying proteins) were inhibited, suggesting a potential disinhibition for purposes of populations regrowth. These patterns were highly reproducible in independent experiments. Based on these findings, 5 promoters were identified and selected as candidate tearing-induced drivers of fabric synthesis (Supplementary note 5 ). All 5 showed at least 8-fold increase in expression upon stimulus while maintaining minimal expression unstimulated. Test drivers were constructed in which each of the 5 promoters was placed to control expression of the lux ABCDE operon. Biofilms transformed with a selected promoter, pst_sigA, responded well to tearing ( Fig. 2B,C \n ) . Figure 2 \n B. subtilis biofilm transcriptome response to mechanical tear and subsequent mounting of a reporter response. ( A ) graph representation of the most up- or down-regulated genes, in Log(2) of the fold change compared with untreated biofilm. Mean p-value for all changes is 0.007 ± 0.014. ( B ) A representative response of B. subtilis biofilms to tear, measured as relative luminescence units (RLU) derived from expression of the lux ABCDE operon under the control of pst -sigA promoter. This graph shows maximal signal achieved after 30 min. Data labels show mean values. P -value constitutive vs. +tear < 0.05; P -value -tear vs. +tear < 0.05. ( C ) A representative kinetic measurement of the response to tear driven by pst -sigA. Blue line represents pst -sigA reporter strain and orange line represents negative control (wildtype strain). P -value < 0.05 in all points (n = 3). \n Next, we turned to designing the synthetic part of the system. The choice of genes for fabric synthesis was guided by mechanistic simplicity: a single gene, and the ability to self-assemble into a functional fiber under specific conditions. Arthropod silks have been known for millennia and are still considered industrial benchmarks today 12 . However, silks from spider species or from the silkworm Bombyx mori require complex weaving organs, making them unsuitable for the purposes of the present design 13 , 14 . For this reason, silk from other sources was examined. Raspy crickets ( Gryllacrididae ) produce silk for building leaf shelters 15 . Recently, several genes encoding cricket silk were cloned from the cricket labial glands, and their partial sequences include alanine/glycine/serine-rich repeats typical of silk proteins from other species 16 . In order to evaluate the suitability of these proteins for the synthetic module, segments of these protein sequences (termed spseg I/II/III/IV/V) were selected and fused to histidine tags for expression vector construction (Supplementary note 6 ). Isoelectric points of the protein segments were calculated, with an interesting distribution into two groups, an ‘acidic’ group containing 3 proteins with pI at ~5.0, and an ‘alkaline’ group containing the remaining 2 proteins with pI at ~8.0 (Fig. 3A ). The proteins were expressed in both insect and bacterial systems (Supplementary note 7 ), in both cases showing efficient assembly upon cumulative acidification and dehydration (increasing protein concentration) (Fig. 3B ) into fibers of a mean diameter of 10 um, with an elastic modulus of 4.54 GPa and tensile strength of 617 MPa. Figure 3 Silk protein analysis and self-assembly. ( A ) amino acid distribution in all 5 silk protein segments. ( B ) calculated net charge on protein surface. ( C ) SEM images showing assembled protein fibers following concentration by dehydration and acidification to approximately the calculated isoelectric point. Size bars = left, 50 μ m; right, 100 μ m. \n Finally, a selected protein segment was placed under the control of a selected promoter, and the fabric-biofilm hybrids were prepared using B. subtilis transformed with the constructed vector. Five min following tearing of the hybrid without submerging or lifting it from the well, and under acetic acid-mediated slight acidification of the medium (pH ~6.0), we observed assembled single fibers, originating within ±20 um from the rim of the tear region (Fig. 4A ; Supplementary movie S12 ). It is noteworthy that the time periods we used are short for a secreted protein to accumulate in significant amounts outside the cell, however the observed rapid assembly is caused by cell lysis due to impact with the acid. Interestingly, silk protein assembly occurred mostly along and around fabric fibers, suggesting that the assembly process is more efficient on the fiber surface, hence the fabric serves as scaffold or guide for the process. Hybrids made with wildtype biofilms were torn as well, without any apparent response (Fig. 4B,C ). The assembled silk fibers showed a strong (~30%) nitrogen band, while the scaffold fabrics showed only carbon and oxygen bands (~70% and 30%, respectively). Figure 4 Regenerating fabric system. ( A ) Representative SEM images of regenerating fabric at t = ~15 minutes following tear of fabric-biofilm hybrid. Yellow arrowheads point at newly-made and assembled silk fibers. Red stars point at metal net discs used for imaging. Orange circles point at fabric fibers. ( B ) images of torn regenerating hybrids (top 3 panels) vs. torn hybrids made with wildtype biofilms (bottom 3 panels), the latter showing no newly-made fibers. All images were taken at tear region. ( C ) Quantitation of imaging fields visualized by light microscopy (sham, wildtype biofilms; vector, engineered biofilms). All size bars = 100 μ m. \n The aim of the present study was to demonstrate plausibility, rather than yield an optimal design. Much work is still needed, mainly in optimizing the expression and assembly efficiency in order to achieve full coverage of the torn region, the fabric material and structure, and the fabric-biofilm interface. It is hypothesized that specific fabrics could be designed, woven, and tested for desired outcomes. Examples include having the fabric retaining liquid medium for an extended period, or the fabric itself being coated by a layer of food source such as starch, which the bacteria can metabolize. An effect of fabric type on biofilm growth was observed in this study, with potential reasons being retention or concentration of nutrients in the fibers. In addition, our observations suggest that the fabric serves as a structural scaffold to both the biofilm and protein fiber assembly, a function that can also be encoded in the architecture of designed fabrics. The draft concept we describe highlights interesting questions and potential solutions on the technical and conceptual levels. For example, the compliance of individuals to wear or use a bacterially-loaded fabric could be improved by engineering a better interface in which the bacteria are contained within the fabric fibers and do not form a separate layer. Another example is that the biofilm is not expected to survive laundry. A potential solution to this issue could be to add dried bacteria or spores as a post-laundry or pre-wearing reagent, effectively reloading the fabric with regeneration capacity. Particularly, current progress in synthetic biology encourages expanding the present concept even further: the bacteria within the fabric could be engineered to produce scent, pigments, or even antibiotics against skin-infecting bacteria. Lastly, the coupling of genes encoding sensing of environmental cues with genes encoding material synthesis highlights a novel category of machines, robots, and useful tools. From adhesives and metals to fractal materials and optic fibers, biological organisms fabricate an astonishing variety of materials currently available to us only by complex engineering processes, and do so in their natural environment under relatively ambient conditions. These possibilities are discussed in Supplementary note 8 . This work strengthens the possibility of a novel class of hybrid systems - objects with fully integrated biological components. Recently, a similar system was reported that employes hygroscopic properties of bacteria to design a ventilating fabric 17 . We propose the term crossbiosis to describe these systems (after the term symbiosis), to stress the fact that a living organism and a nonliving object are combined to create a new entity with synergistic properties."
} | 4,216 |
32730250 | PMC7392205 | pmc | 1,258 | {
"abstract": "In the animal kingdom, various forms of swarming enable groups of autonomous individuals to transform uncertain information into unified decisions which are probabilistically beneficial. Crossing scales from individual to group decisions requires dynamically accumulating signals among individuals. In striking parallel, the mammalian immune system is also a group of decentralized autonomous units (i.e. cells) which collectively navigate uncertainty with the help of dynamically accumulating signals (i.e. cytokines). Therefore, we apply techniques of understanding swarm behavior to a decision-making problem in the mammalian immune system, namely effector choice among CD4+ T helper (Th) cells. We find that incorporating dynamic cytokine signaling into a simple model of Th differentiation comprehensively explains divergent observations of this process. The plasticity and heterogeneity of individual Th cells, the tunable mixtures of effector types that can be generated in vitro , and the polarized yet updateable group effector commitment often observed in vivo are all explained by the same set of underlying molecular rules. These rules reveal that Th cells harness dynamic cytokine signaling to implement a system of quorum sensing. Quorum sensing, in turn, may confer adaptive advantages on the mammalian immune system, especially during coinfection and during coevolution with manipulative parasites. This highlights a new way of understanding the mammalian immune system as a cellular swarm, and it underscores the power of collectives throughout nature.",
"introduction": "Introduction Collective behavior–the coordinated action of many autonomous individuals–can accomplish sophisticated information-processing tasks that may be impossible for lone individuals. This has led to the repeated evolution of swarming across various taxa [ 1 ]. For example, honeybee swarms leverage multiple types of interactions among individuals to choose the best nesting site among several options [ 2 , 3 ]. Ant swarms leverage variability in chemical signaling among individuals to dynamically track moving food sources [ 4 , 5 ]. Bacterial swarms use quorum sensing–a special class of collective behavior in which different group decisions emerge depending on the density of constituent individuals–to measure and respond in unison to fluctuating environmental conditions [ 6 , 7 ]. In each example, collective behavior allows swarms to integrate conflicting, changing and otherwise uncertain information into unified decisions which are dynamically updated and probabilistically beneficial. Although swarms are typically considered to comprise distinct organisms, collective behavior can also arise from cells within an organism. In particular, the mammalian immune system embodies many aspects of collective behavior. Immune cells are decentralized and autonomous individuals that together make coherent decisions despite substantial uncertainty [ 8 , 9 ]. For example, CD4+ helper T (Th) cells collectively decide whether a foreign invader warrants an immune response, and which immune effectors to deploy. Understanding how such decisions emerge requires understanding how cells collectively coordinate their behavior [ 10 , 11 ]. Just as in insect swarms, this communication involves complex feedbacks within and among Th cells [ 12 , 13 ], which vary drastically in their signaling outputs [ 14 , 15 , 16 , 17 ]. Thus, we propose that the lens of collective behavior may reveal novel insights into how the immune system processes uncertain, conflicting, and changing information [ 8 , 9 ]. We apply that lens here to study Th effector choice. This process begins when sentinels called antigen-presenting cells (APCs) enter lymph nodes, bearing fragments of parasites called antigens. Th cells that recognize and bind these antigens form immunological synapses with APCs, through which they receive instructions to proliferate and differentiate into a given effector type (e.g. Th1, Th2, Th17) [ 18 ]. These types correspond to different classes of infection; for example, Th1 cells combat intracellular microparasites, while Th2 cells combat extracellular macroparasites [ 18 ]. Each Th cell broadcasts its type to its neighbors via diffusible signaling molecules called cytokines, influencing their effector differentiation [ 19 , 20 ]. At the Th group scale (e.g., across a lymph node), accurate differentiation into the effector type best matched to the current threat is critical for host survival [ 21 , 22 , 23 ]. Effector choice is difficult for several reasons. First, information is limited: APCs are rare, such that each Th cell has a low probability of receiving effector instruction from an APC on a per-hour basis [ 24 ]. After APC contact, a Th cell resists further contact for up to 72 hours, precluding ongoing APC instruction [ 25 ]. Second, information may be conflicting: because mammalian hosts in nature are constantly coinfected with parasites requiring different effector responses [ 26 ], APCs that have encountered different types of parasites will instruct for different effector types. Third, information may be changing and deceptive: many parasites manipulate APCs to instruct for inaccurate effector types in order to escape clearance [ 27 , 28 , 29 ]. At first glance, cytokine signaling among Th cells may only amplify this uncertainty. It is unclear how Th cells process conflicting and potentially untrustworthy information. Furthermore, Th cells seem to process information differently in different settings. In vivo , Th cells often make strictly polarized decisions. Coinfections with parasites requiring different effector responses often elicit unified commitment to one effector type exclusively [ 30 , 31 , 32 ]. On the other hand, in vitro , Th cells given conflicting effector stimulation adopt mixed effector types, simultaneously secreting cytokines characteristic of different effector types [ 17 , 33 , 34 ]. These contradicting results are difficult to reconcile. Given that Th cells can plastically switch effector types [ 13 , 35 , 36 , 37 ] and display broad cell-to-cell variability in their cytokine expression [ 14 , 15 , 16 , 17 ], the strictly polarized decisions that arise despite conflicting APC instruction in vivo seem especially difficult to explain. Here, we solve this immunological puzzle by understanding Th cells as a swarm. Through collective behavior, group consensus and commitment can arise despite conflicting environmental cues via dynamic signals that cross scales from the individual to the group. Whether these signals are autoinducers secreted by bacteria [ 6 ], pheromones deposited by ants [ 4 ], or even startle responses among schooling fish [ 38 ], what matters is that they dynamically accumulate. By analogy, Th cells are individuals, all Th cells in a lymph node form a group, APCs are (possibly) conflicting environmental signals, and cytokines are the dynamically accumulating signals among individuals. With this motivation, we modified a well-studied model of the gene expression motif driving Th1 vs. Th2 differentiation [ 33 , 39 , 40 , 41 ] to include the key cytokines in this process. We then addressed four Questions: Does the model explain how mixed Th effector types arise in vitro ? Does the same model explain how polarized group effector decisions arise in vivo ? When does dynamic cytokine signaling matter in vivo , given the presence of APCs? What advantages could collective coordination via dynamic cytokine signaling provide? We find that polarized group effector decisions emerge only above a threshold cell density, simultaneously explaining in vitro and in vivo observations, and epitomizing quorum sensing. Moreover, our model predicts that quorum sensing operates even in the presence of APCs and leverages cell-to-cell variability in cytokine signaling to discern true from deceptive information. Indeed, it has recently been suggested that Th cells use quorum sensing to make other immune decisions [ 42 , 43 ], and one group has provided empirical evidence that Th cell density modulates the rate of memory differentiation [ 44 ]. Empirical studies have also demonstrated that quorum sensing regulates key processes in other closely related immune cells, such as CD8+ killer T cell proliferation [ 45 ] and B cell motility [ 46 ]. Here, we provide the first comprehensive explanation of how a quorum emerges from a group of Th cells, and why Th quorum sensing might adaptively benefit the host organism. Model development Although Th effector differentiation is the result of a complex gene expression program, the simplifying assumption that the transcription factor T-bet primarily drives Th1 differentiation and the transcription factor GATA3 primarily drives Th2 differentiation is both common and empirically supported [ 33 , 39 , 40 , 41 ]. Both transcription factors induce their own expression (“self-activation”) and diminish each other’s expression (“cross-inhibition”), forming a previously analyzed transcription factor motif [ 33 , 39 , 40 , 41 ]. While these transcription factors are confined within Th cells, Th1 and Th2 cells also secrete cytokines–IFNγ and IL-4, respectively–that diffuse through the extracellular space and similarly self-activate and cross-inhibit [ 13 ]. We searched the immunological literature for all known molecular interactions among these transcription factors and cytokines. Together, they form a system of four ordinary differential equations (“ODEs”) ( Fig 1 and S1 Table , see S1 Text regarding mathematical forms). The parameter values assigned to each interaction were also grounded in the immunological literature ( S1 Table ); nonetheless, sensitivity analyses show that errors in these parameter estimates do not qualitatively alter our results ( S2 Text ). Thus, these ODEs and parameter values describing average molecular expression across a group of Th cells are the foundation of our analyses. d T F 1 = [ ( b + p 1 T F 1 h p P 1 h p + T F 1 h p ) ( X 2 h x X 2 h x + T F 2 h x ) + ( s 1 C Y 1 h s S 1 h s + C Y 1 h s ) ( Z 2 h z Z 2 h z + C Y 2 h z ) − d TF 1 T F 1 ] d t Eq 1 d T F 2 = [ ( b + p 2 T F 2 h p P 2 h p + T F 2 h p ) ( X 1 h x X 1 h x + T F 1 h x ) + ( s 2 C Y 2 h s S 2 h s + C Y 2 h s ) ( Z 1 h z Z 1 h z + C Y 1 h z ) − d TF 2 T F 2 ] d t Eq 2 d C Y 1 = [ ( a 1 T F 1 h a A 1 h a + T F 1 h a ) ( R 2 h r R 2 h r + T F 2 h r ) ( U 2 h u U 2 h u + C Y 2 h u ) − d CY 1 C Y 1 ] d t Eq 3 d C Y 2 = [ ( a 2 T F 2 h a A 2 h a + T F 2 h a ) ( R 1 h r R 1 h r + T F 1 h r ) ( U 1 h u U 1 h u + C Y 1 h u ) − d CY 2 C Y 2 ] d t Eq 4 10.1371/journal.pcbi.1008051.g001 Fig 1 Model schematic. T-bet and GATA3 are the master transcription factors controlling Th1 and Th2 differentiation, respectively, and are confined within Th cells. IFNγ and IL-4 are the master cytokines controlling Th1 and Th2 differentiation, respectively, and are free to diffuse through the extracellular space. Together, expression of these four molecules are the four state variables of the dynamic model. Each of these four molecules can upregulate (arrow-head interactions) or downregulate (T-head interactions) the expression of the other molecules in the model. References to the immunological literature supporting the existence of the depicted interactions and their assigned parameter values are provided in S1 Table . We made several modifications to this basic ODE model in order to address the four Questions outlined above. First, Question 1 requires studying not just the mean but the full distribution of molecular expression across a group of Th cells. Thus, we extended our model into a system of four stochastic differential equations (“SDEs”), by adding to each equation the differentials of independent Brownian motion processes (+[ n TF TF 1 ] dW TF 1 for Eq 1 , and analogously for Eqs 2 – 4 , see S3 Text ). Because mammalian cells express proteins in nearly discrete bursts [ 47 , 48 ] such that any given cell fluctuates across the entire distribution of expression through time [ 49 , 50 ], SDEs are an appropriate mathematical tool [ 51 ]. In fact, the specific form of our stochastic term appropriately models lognormal fluctuations in molecular expression, because distributions of T-bet, GATA3, IFNγ, and IL-4 expression among Th cells span several orders of magnitude with large positive skew [ 34 , 52 ]. Therefore, many simulated sample paths of these SDEs together approximate the distribution of molecular expression in a group of Th cells. Second, Question 2 requires comparing Th cells in vitro vs. in vivo . A major difference between these settings is cell density: in vitro Th culture requires ~10 6 cells/mL [ 33 , 34 ], whereas Th cells exist in vivo in lymph nodes at ~10 9 cells/mL [ 20 ]. Thus, we extended our model to accommodate this range of cell densities, by identifying which parameters depend on cell density. While intracellular transcription factors are measured as the number of copies per cell and therefore do not depend on cell density, extracellular cytokines are measured in terms of concentration in the extracellular space and do depend on cell density. As cell density increases, the proportion of extracellular space decreases, compacting secreted cytokines into smaller volumes. Thus, in terms of extracellular concentration, both cytokine production ( a 1 , 2 ) and removal ( d CY1 , 2 ) rates increase with increasing cell density ( S1 Fig and S1 Text ). Moreover, production is driven only by cellular secretion, but removal is driven by both cellular consumption and free decay (which is often fast for molecules involved in cytokine regulation [ 53 ]). Therefore, over the range of cell densities we studied, a 1 , 2 scales more steeply with cell density than does d CY1 , 2 . Consequently, cytokines dynamically turn over faster and accumulate to higher levels as cell density increases ( S1 Fig ). See S1 Text for a full explanation of the units and cell density dependencies in this model. Third, both Questions 3 and 4 require accounting for instruction by APCs. Biologically, APCs provide Th1 or Th2 instruction by secreting Th1- or Th2-driving cytokines directly onto a Th cell surface via an immunological synapse [ 18 , 54 , 55 , 56 ]. Therefore, we included APC instruction by augmenting every appearance of CY 1 and CY 2 (except decay) in the model equations with constants APC 1 and APC 2 , whose values depend on the frequencies of Type 1 and 2 APCs, respectively (see S1 Table ). For example, ( s 1 C Y 1 h s S 1 h s + C Y 1 h s ) becomes ( s 1 ( C Y 1 + A P C 1 ) h s S 1 h s + ( C Y 1 + A P C 1 ) h s ) in Eq 1 , and so forth. This allows APCs to influence Th decision-making without following the same rules of production and removal as cytokines.",
"discussion": "Discussion Quorum sensing, and other forms of swarming, have repeatedly evolved across various taxa to allow groups of organisms to collectively navigate their environments [ 1 ]. Swarming is particularly useful when information is limited [ 2 , 3 ], changing [ 4 , 5 ], or otherwise uncertain [ 7 , 38 ]. All three qualifiers describe the information regarding effector choice that Th cells receive from APCs, which are rare, mutable in effector type, and even subject to sabotage from parasites [e.g. 27 , 28 , 29 ]. Moreover, Th cells possess a well-known mechanism of dynamic signal accumulation–a requirement for quorum sensing [ 6 ] and other forms of swarming–in the form of cytokine secretion and consumption. Finally, Th cells are well-suited to swarming in an evolutionary sense. In most swarms, fitness is measured at the individual level (e.g. one fish in a school), such that the benefit of information-sharing to an individual must outweigh the individual cost of helping conspecific competitors, if swarming is to evolve. To the contrary, the evolutionarily relevant fitness of Th cells is measured at the group scale (i.e. the entire host organism) and is unconstrained by costs to individual Th cells [ 8 ]. For all these reasons, a system of quorum sensing among Th cells may be logically expected. Following this expectation, we modeled the quantitative system by which Th cells integrate potentially conflicting and uncertain information from various sources, including each other and APCs, to make effector choices. We label this system quorum sensing because it requires both dynamic signaling across scales and sufficient cell density. At artificially low cell density, as in cell culture, with signaling among Th cells prohibited, individual Th cells polarize toward Th1 or Th2, but not as a unified collective ( Fig 2C and Fig 7A ). At artificially low cell density with signaling among Th cells permitted, cytokines promote stable mixtures of Th1, Th2, and Th1-Th2 hybrid T cells ( Fig 2C and Fig 7B ). Both conclusions resonate with experimental data ( Fig 2A and 2B , and Fig 3 ). Only at biologically realistic cell density with signaling among Th cells permitted do unified group decisions between Th1 and Th2 effector types emerge ( Fig 4A and Fig 7C ). This too resonates with various experimental observations [e.g. 30 , 31 , 32 ]. The necessary ingredients for this committed effector choice–dynamic cross-scale signaling among cells and sufficient cell density–define a quorum sensing process. 10.1371/journal.pcbi.1008051.g007 Fig 7 Cartoon of major conclusions. Orange circles represent Th1 cells, blue circles represent Th2 cells, and other shades represent Th1-Th2 hybrid cells. The large gray shapes represent APCs, or experimentally provided effector stimulation. (a) At 10 6 cells/mL with no dynamic cytokine signaling, individual neighboring Th cells adopt oppositely polarized effector types. (b) At 10 6 cells/mL with dynamic cytokine signaling, oppositely polarized Th cells cause each other to become Th1-Th2 hybrids. (c) At 10 9 cells/mL with dynamic cytokine signaling, mixed effector types resolve into fully polarized Th1 or Th2 groups, via quorum sensing. Initial polarization by APCs, effector hybrid formation as cytokines dynamically accumulate, and quorum emergence as cytokines accumulate further and APCs are ignored, may define 3 phases of Th effector differentiation in vivo . Because this progression of scenarios also implies a progression of the extracellular cytokine concentration, Th quorum sensing can be understood as a series of phases in time ( Fig 7C ). Th cells deciding between Th1 and Th2 effector types begin by collecting binary information from APCs. As secreted cytokines accumulate, they invoke molecular feedbacks by which Th cells share information and tune the Th1-Th2 balance of their neighbors. Cytokines continue to dynamically accumulate until they surpass a threshold that is only attainable at biological cell density, precipitating a unified group-level decision between Th1 and Th2 effector types. This series of phases discounts APC instruction over time, such that the Th quorum decision is eventually irreversible by APC instruction ( Fig 5A and 5B ). This is consistent with the idea that information gathered by the immune system early in an infection is most trustworthy [ 54 ]. Many parasites are capable of manipulating APCs into instructing for the incorrect effector type [ 28 , 29 ]; for example, during infection with Leishmania spp., APCs provide appropriate Th1 instruction early in an infection, but later succumb to sporadic manipulation events that alter their effector instructions [ 27 ]. Because parasitic manipulation is predicted to influence the evolved structure of immune systems [ 8 , 57 , 58 ], it may be that quorum sensing is an adaptive parry. If a Th quorum has ceased obeying APC instruction by the time manipulation occurs, then the quality of the host immune response is not compromised, providing robustness in the face of sabotage. Nonetheless, quorum commitment could be maladaptive when a switch in effector type is truly required. Our model predicts that such switches are possible, when opposing APC instruction is coupled with stochastic variability in molecular expression ( Fig 6A–6E ). The underlying mechanism, stochastic attractor switching, is observed in other natural systems of collective decision-making, for example by allowing insect swarms to respond to dynamically changing environments [ 4 , 5 ]. Because stochastic attractor switching is a probabilistic phenomenon, the cumulative probability of a transition between states increases with the length of the time window under consideration. Importantly, this allows a Th quorum to discern sustained and legitimate changes in APC instruction from transient and manipulated perturbations to APC instruction. We find that discernment operates best when cytokines, rather than transcription factors, are subject to large cell-to-cell expression variability ( Fig 6F ). Combined with the observation that exaggerated cell-to-cell variability in cytokine expression is a conserved trait across mammalian species [ 59 ], this raises the tantalizing possibility that cytokine expression variability is an adaptive feature of immune signaling [ 8 ]. Indeed, signaling variability in other biological swarms, such as house-hunting ants, has already been postulated as an adaptive mechanism to mitigate the speed vs. accuracy tradeoff inherent to decision-making processes with uncertain information [ 60 ]. It is possible that natural selection has converged on swarming as a common solution to such problems of uncertainty, both among individual organisms in groups as well as among individual cells within organisms. If the evolution of the mammalian immune system can be understood in this way, then more insights into its organization and functioning may emerge as analogies with other biological swarms are explored further [ 8 , 9 ]. Despite offering a quantitative explanation of Th effector choice that reconciles disparate observations by conceptually unifying collective behavior and immunology, our model does have limitations. For example, our model explains why unified effector choices can emerge in vivo but not in vitro , and yet unified effector choices are not always observed in vivo [e.g. 61 ]. While our model predicts the long-term equilibrium outcome of Th effector choice, immunity in vivo is a non-equilibrium process: cellular birth, migration, and death, parasite replication and death, metabolic inputs and constraints, stochastic events, and a plethora of other factors constantly perturb the immune system. Not every data set will conform to equilibrium predictions, but equilibrium predictions can help explain broad patterns that emerge from the balance of many studies. Additionally, while the model only addresses Th1 vs. Th2 differentiation, many other effector types exist [ 18 ]. In fact, Th differentiation choices between Th17 and iTreg are driven by self-promoting and cross-inhibiting molecular interactions similar to those in this model [ 62 ]. Just as this model assumes that a single master transcription factor underlies Th1 and Th2 effector types (T-bet and GATA3, respectively), so too has this “master regulator” assumption been applied to other effector types (e.g. RORγt for Th17, Foxp3 for iTreg) in other mathematical studies to insightfully recover experimentally observed patterns of Th effector differentiation [ 63 , 64 ]. Thus, this model could likely be adapted to represent different or additional Th effector types without changing its basic predictions. The model also simplifies several details of T cell biology. First, while we have assumed that Th cells exist at roughly 10 9 cells/mL inside lymph nodes, not all these Th cells actively participate in immunity. In fact, early during infection, only 1 in every 10 5 or fewer Th cells are activated by any given antigen [ 65 ], such that the density of activated Th cells is quite small. However, these activated Th cells proliferate, increasing their numbers by several orders of magnitude [ 65 , 66 ]. Moreover, bystander Th cells which have not been activated by the antigen still participate in effector choice [ 31 ]. These processes greatly increase local cell density in the lymph node, likely surpassing the threshold density for quorum sensing. Accounting for Th proliferation might lengthen the information collection and information sharing phases identified by our model, tuning the amount of time until the onset of the decision-making phase ( Fig 7C ), but it should not preclude quorum sensing altogether. Second, Th cells given consistent effector instruction for long periods of time may undergo epigenetic modifications to commit irreversibly to an effector type, losing plasticity [ 67 ]. While our model does not include epigenetic entrenchment, this phenomenon likely requires over a week of stimulation [ 33 ] and therefore does not interfere with any of the results we present. Finally, while our mathematical approach highlights key design principles embedded in the vast complexity of mammalian immunity, direct empirical evidence of quorum sensing in the Th effector choice process remains to be collected. Though technically challenging, experiments that track the effector commitment of individual Th cells over extended time periods given conflicting or fluctuating instructions are needed to test the predictions of this model further. Although unified effector commitment among Th groups may benefit hosts who have coevolved with deceptive parasites, it can also be detrimental. For example, helminth infection can establish an organ-scale commitment to Th2 immunity that prevents vaccines from eliciting proper Th1 memory against deadly intracellular pathogens [ 31 ]. Indeed, there is evidence that pre-existing chronic infections consistently diminish vaccine efficacy [ 68 ]. Ultimately, we expect that parallel mechanistic and evolutionary understandings of emergent immune phenomena can suggest new ways to manipulate our immune systems, and when it is wise to do so. In turn, successful application of such cross-disciplinary thinking to immunological problems can highlight the power and importance of collectives throughout the natural world."
} | 6,557 |
36529189 | null | s2 | 1,261 | {
"abstract": "No abstract available"
} | 5 |
29445119 | PMC5943517 | pmc | 1,263 | {
"abstract": "Spiders produce multiple silks with different physical properties that allow them to occupy a diverse range of ecological niches, including the underwater environment. Despite this functional diversity, past molecular analyses show a high degree of amino acid sequence similarity between C-terminal regions of silk genes that appear to be independent of the physical properties of the resulting silks; instead, this domain is crucial to the formation of silk fibers. Here, we present an analysis of the C-terminal domain of all known types of spider silk and include silk sequences from the spider Argyroneta aquatica , which spins the majority of its silk underwater. Our work indicates that spiders have retained a highly conserved mechanism of silk assembly, despite the extraordinary diversification of species, silk types and applications of silk over 350 million years. Sequence analysis of the silk C-terminal domain across the entire gene family shows the conservation of two uncommon amino acids that are implicated in the formation of a salt bridge, a functional bond essential to protein assembly. This conservation extends to the novel sequences isolated from A. aquatica . This finding is relevant to research regarding the artificial synthesis of spider silk, suggesting that synthesis of all silk types will be possible using a single process.",
"conclusion": "Conclusion In silk research, the C-terminus is a key component of the minimal sequence used for recombinant spider silk production, as without this region fibers will not form (Stark et al. 2007 ). What we find is an overall model for all silk types irrespective of physical traits that illustrates the range of environments in which this single protein family may be utilised. Our results allow us to identify conserved amino acid residues essential to the correct formation of silk proteins, thereby enabling the identification of less essential residues that may be chemically functionalised in artificially synthesised silk using techniques such as click-chemistry (Harvey et al. 2016 ). This study will aid researchers in selecting suitable sequences without repetitive testing in vivo whilst predicting sites which may be suitable for modification, as seen in Harvey et al. ( 2016 ), and build a repository of spidroin parts that could be combined to achieve novel and custom characteristics.",
"introduction": "Introduction Spider silk proteins form multiple types of materials, including fibers and glues, each with specific mechanical properties and functions (Blamires et al. 2017 ; Eisoldt et al. 2011 ; Garb 2013 ; Hormiga and Griswold 2014 ). Gene duplication, recombination and diversification are thought to have generated this huge diversity of different silk types, each having a different ecological function (Clarke et al. 2015 ). These functions include safety lines, the structural frameworks of webs, external layers of protection around egg sacs, and securing prey (Hinman et al. 2000 ; Rising and Johansson 2015 ). In the case of Argyroneta aquatica (Clerck 1757 ) (Araneae: Cybaeidae (Catalog 2016 )), silk is used in a typical fashion to construct a web in which the spider resides and performs a number of actions, from feeding to mating and storing eggs (Schütz and Taborsky 2003 , 2005 , 2011 ; Schutz et al. 2007 ; Seymour and Hetz 2011 ). What is unusual about this spider is that it is the only known species to spin silk whilst submersed in water. The subsequent sheet web is then inflated with air drawn down from the surface and is utilised as an air reservoir. The silken “diving bell” allows for oxygen diffusion to occur, allowing the spider to avoid surfacing for extended periods of time (Seymour and Hetz 2011 ). Cybaeus angustiarum , a fellow cybaeid, is a terrestrial species found in dense forests (often on north facing, scree slopes) in areas under stones or in decaying wood with humidity close to 100%. In all spiders, liquid silk dope is passed through specialised, elongated glands within the body of the spider and extruded through spinnerets (Askarieh et al. 2010 ; Rising and Johansson 2015 ; Vollrath and Knight 2001 ). All silk genes have three components; a type-specific repetitive region flanked by conserved N-terminal (Motriuk-Smith et al. 2005 ; Rising et al. 2006 ) and C-terminal (Challis et al. 2006 ; Collin et al. 2016 ; Gnesa et al. 2012 ; Hagn et al. 2010 ) domains. Whilst the repetitive region determines the mechanical properties of each silk product (Hayashi and Lewis 1998 ; Hayashi et al. 1999 ), the terminal domains work together ensuring that individual proteins assemble correctly and that the fibre forms at the correct stage of the spinning process (Andersson et al. 2014 ; Andersson et al. 2017 ; Rising and Johansson 2015 ). The N-terminal domain restricts the formation of silk fibers to a precise point in the silk duct, preventing silk proteins stored in the silk gland from agglutinating (Askarieh et al. 2010 ). The C-terminal domain drives spontaneous fibre formation, likely through use of a pH-sensitive “salt bridge” (Ittah et al. 2006 ; Stark et al. 2007 ), where noncovalent interactions between one basic and one acidic residue are disrupted at low pH because the acidic residue becomes protonated and is no longer charged. Salt bridges, either individual or paired, have been proposed to explain the dimeric bundling of 4 or 5 alpha helices in the C-termini of one particular silk, the major ampullate (Hagn et al. 2010 ; Sponner et al. 2004a , 2005 ). Whilst the miniature spidroin 4RepCT, formed of four copies of a MaSp repetitive region and one C-terminus, has been sufficient to produce self-assembling silk fibers (Stark et al. 2007 ), it has recently been shown in the minispidroin NT2RepCT that including the N-terminal region achieves greater efficiency in the production of synthetic fibers (Andersson et al. 2017 ). However, studies have shown that the level of solubility demonstrated by the terminal regions differs between spider species, leading to further questions around their function and the suitability of individual terminal regions for use in silk protein synthesis (Andersson et al. 2014 , 2017 ; Askarieh et al. 2010 ). A degree of amino acid sequence conservation has been observed from studies of small numbers of different silks (Beckwitt and Arcidiacono 1994 ; Challis et al. 2006 ; Collin et al. 2016 ; Gnesa et al. 2012 ; Hagn et al. 2010 ; Sponner et al. 2004b , 2005 ). What is not clear is the extent to which this conservation is maintained, particularly where silk proteins exhibit vastly different properties (e.g., the glue-like aggregate and piriform silk versus the superior strength of major ampullate and aciniform silks). Additionally, the diving bell spider Argyroneta aquatica spins silk whilst completely submersed in fresh water; given how silk proteins are dehydrated as part of the spinning process, does this “extreme” environment necessitate variation within the silk protein and spinning process? Here, we analyse the genetic sequences silks of all known types from spiders of as many groups as are available in GenBank. We investigate whether the salt bridge structure is conserved across all species and silk types and ask if changes in biophysical properties or the utilisation of silk in an “extreme” environment necessitates a change in how silk proteins are formed.",
"discussion": "Discussion This analysis encompasses 19 families from within the Araneae and shows conservation of a pH-sensitive salt bridge, typically composed of an arginine-glutamic acid pairing, within all the silk types and species examined. Our finding of this degree of conservation confirms an essential role for this feature in the correct assembly of silk fibers. The extended coverage in terms of species and phylogenetic diversity suggests this feature has been conserved for the entire evolutionary history of the group—some 360 million years. Moreover, this conservation has persisted despite the diversification of other spider traits. For example, terrestrial spiders such as the mygalomorphs produce two to three silk proteins from a large, undifferentiated silk gland whereas the Orbiculariae have evolved several specialised glands from which different types of silk are extruded (Blackledge and Hayashi 2006 ; Blamires et al. 2017 ; Clarke et al. 2017 ; Garb 2013 ; Garb et al. 2010 ; Perez-Rigueiro et al. 2010 ; Rising et al. 2007 ; Stark et al. 2007 ). Conservation of C-terminus in all A. aquatica silk genes Argyroneta aquatica predominantly spins its silk in water, although will occasionally leave a dragline when walking in a terrestrial environment. Comparatively, individual Cybaeus did not appear to produce a dragline but given time would spin silken webs in which they resided, which appear visually similar to the dehydrated diving bells of A. aquatica . In sequences obtained from both these species, and also from the first fully-annotated spider genome of Nephila clavipes (Babb et al. 2017 ) in which a number of new silk sequences were identified, the residue pairing within the C-terminal domain has been conserved. This suggests that the formation of silk fibers by A. aquatica occurs using the same method as in terrestrial spiders. This leads to questions around how this spider is adapted to spinning silk underwater and the subsequent mechanical properties of A. aquatica silk in such an “extreme” environment. This is particularly relevant as studies of silk in humid conditions suggest the structure may be temporarily affected by the level of humidity; the degree to which major ampullate silk supercontracts depends on the species, whereas minor ampullate silk does not supercontract (Agnarsson et al. 2009 ; Blackledge et al. 2009 ; Boutry and Blackledge 2010 ). Conservation of a physical structure The residues responsible for the formation of the pH-sensitive salt bridge(s) are both uncommon and conserved within the restricted number of sequences studied thus far (Challis et al. 2006 ; Hagn et al. 2010 ). Analysis of the expanded range of sequences used in this study shows this trait is maintained, with >90% of the C-terminal domain typically composed of hydrophobic and hydrophilic residues. Of the charged residues present, at least one acidic and one basic residue is conserved in all sequences and as such inferred to be crucial to the formation of a salt bridge and therefore the correct folding of silk proteins into a fibre. The conservation of an arginine-glutamic acid pair of residues suggests this may be the optimum pairing for a salt bridge in silk, but the presence of lysine and aspartic acid in some sequences implies this is not always essential, although the effect of these substitutions on the final protein structure and its physical properties is currently unknown. The higher level of hydrophilic residues around the basic residue and hydrophobic residues around the acidic residue may be necessary to ensure the correct formation of the salt bridge during the protein folding stage and allowing its regulation by smaller changes in pH due to the local environment created by the protein. Where two pairs of residues are conserved in major ampullate, minor ampullate, piriform, and some aciniform sequences it is likely that there are two salt bridges present in the C-terminus, although further structural analysis would confirm that the second is not merely sequence duplication or an evolutionary artefact suggesting the presence of a second bridge."
} | 2,889 |
21297863 | PMC3027613 | pmc | 1,264 | {
"abstract": "Biogas production from renewable resources is attracting increased attention as an alternative energy source due to the limited availability of traditional fossil fuels. Many countries are promoting the use of alternative energy sources for sustainable energy production. In this study, a metagenome from a production-scale biogas fermenter was analysed employing Roche's GS FLX Titanium technology and compared to a previous dataset obtained from the same community DNA sample that was sequenced on the GS FLX platform. Taxonomic profiling based on 16S rRNA-specific sequences and an Environmental Gene Tag (EGT) analysis employing CARMA demonstrated that both approaches benefit from the longer read lengths obtained on the Titanium platform. Results confirmed Clostridia as the most prevalent taxonomic class, whereas species of the order Methanomicrobiales are dominant among methanogenic Archaea . However, the analyses also identified additional taxa that were missed by the previous study, including members of the genera Streptococcus , Acetivibrio , Garciella , Tissierella , and Gelria , which might also play a role in the fermentation process leading to the formation of methane. Taking advantage of the CARMA feature to correlate taxonomic information of sequences with their assigned functions, it appeared that Firmicutes , followed by Bacteroidetes and Proteobacteria , dominate within the functional context of polysaccharide degradation whereas Methanomicrobiales represent the most abundant taxonomic group responsible for methane production. Clostridia is the most important class involved in the reductive CoA pathway (Wood-Ljungdahl pathway) that is characteristic for acetogenesis. Based on binning of 16S rRNA-specific sequences allocated to the dominant genus Methanoculleus , it could be shown that this genus is represented by several different species. Phylogenetic analysis of these sequences placed them in close proximity to the hydrogenotrophic methanogen Methanoculleus bourgensis . While rarefaction analyses still indicate incomplete coverage, examination of the GS FLX Titanium dataset resulted in the identification of additional genera and functional elements, providing a far more complete coverage of the community involved in anaerobic fermentative pathways leading to methane formation.",
"introduction": "Introduction The fraction of renewable energy forms for energy supply is constantly increasing since fossil fuels are running short and energy production from fossil fuels brings about emissions of the greenhouse gas carbon dioxide which has implications on the climate. In this context the production of biogas by means of fermentation of biomass becomes more and more important because biogas is regarded as a clean, renewable and environmentally compatible energy source [1] , [2] . Moreover, generation of energy from biogas relies on a balanced carbon dioxide cycle. In Germany biogas is mainly produced from energy crops such as maize and liquid manure in medium-sized agricultural biogas plants [1] . The microbiology of biogas formation from organic matter is complex and involves interaction of different microorganisms. In the first step of the digestion process, organic polymers of the substrate such as cellulose, other carbohydrates, proteins and lipids are hydrolysed to low-molecular weight compounds [3] – [5] . Cellulolytic Clostridia and Bacilli are among other bacteria important for this step. Subsequently, fermentative bacteria convert low-molecular weight metabolites into volatile fatty acids, alcohols, and other compounds which are then predominantly metabolised to acetate, carbon dioxide and hydrogen by syntrophic bacteria [6] – [11] . These latter compounds are in fact the substrates for methane synthesis which is accomplished by methanogenic Archaea \n [12] , [13] . Hydrogenotrophic Archaea are able to reduce carbon dioxide to methane using hydrogen as an electron donor, whereas aceticlastic Archaea convert acetate to methane [14] – [18] . The biochemistry and enzymology of methanogenesis is well known for model organisms, but the functioning of biogas-producing microbial communities on the whole is insufficiently explored. Community structures of biogas-producing microbial consortia were analysed for different systems and settings including a thermophilic municipal biogas plant [19] , a thermophilic anaerobic municipal solid-waste digester [20] , thermophilic upflow anaerobic filter reactors [21] , a completely stirred tank reactor fed with fodder beet silage [22] , a two-phase biogas reactor system operated with plant biomass [23] , an anaerobic sludge digester [24] , mesophilic anaerobic chemostats [25] , [26] , a packed-bed reactor degrading organic solid waste [27] and many other habitats. Most of these studies were based on the construction of 16S rRNA clone libraries and subsequent sequencing of individual 16S rRNA clones. The resulting nucleotide sequences were then taxonomically and phylogenetically classified to deduce the structure of the underlying community. Also, mcrA clone libraries were used to elucidate methanogenic archaeal communities of different habitats [28] – [32] . The mcrA gene encodes the alpha subunit of methyl-coenzyme M reductase representing the final enzyme in the methanogenesis pathway. Since mcrA is present in all methanogenic Archaea analysed so far, it serves as a phylogenetic marker for this group of Archaea . Usually, analyses of mcrA and 16S rRNA clone libraries do not cover the whole complexity of the respective habitats since sequencing can only be done for limited numbers of clones. Moreover, results of clone library analyses are always biased by the choice of primers that are used for amplification of marker gene fragments and cloning efficiencies. In recent years, microbial communities have been studied on the basis of their metagenomes which became accessible by applying high-throughput sequencing technologies. Recently, the first metagenome sequencing approach for a biogas-producing community was described [33] . Community DNA isolated from a production-scale biogas plant fed with maize silage, green rye and low amounts of chicken manure was sequenced on the Genome Sequencer FLX platform which resulted in 142 million base pairs of sequence information. Bioinformatic methods were employed to deduce the taxonomic composition and functional characteristics of the intrinsic biogas community [34] . Analysis of the community revealed Clostridia as the most prevalent phylogenetic class, whereas species of the order Methanomicrobiales are dominant among methanogenic Archaea . Similar results were obtained by parallel construction of 16S rRNA and mcrA amplicon libraries and subsequent sequencing of cloned fragments [35] . Moreover, bioinformatics results indicated that Methanoculleus species play a dominant role in methanogenesis and that Clostridia are important for hydrolysis of plant biomass in the analysed fermentation sample. Rarefaction analysis of the metagenome data showed that the sequencing approach was not carried out to saturation. Sufficient coverage of non-abundant microbial groups in the fermentation sample would require deeper sequencing. Therefore, the available total community DNA preparation from the biogas fermentation sample was additionally sequenced on the GS FLX Titanium platform, which provides longer read lengths and increased throughput compared to the GS FLX platform. This paper describes an integrated analysis of the GS FLX and the GS FLX Titanium datasets with the objective to deepen the knowledge on the taxonomic structure and composition of a microbial community involved in biogas production within an agricultural, production-scale biogas plant. Moreover, the described analysis intends to elucidate the metabolic capacity of the community, functional roles of specific microorganisms and key organisms for the biogas production process.",
"discussion": "Discussion In a recent study, the metagenome of a biogas-producing microbial community was sequenced employing the GS FLX pyrosequencing platform. Since analysis results showed insufficient coverage of the community, the study was complemented by additional sequencing of the same DNA preparation on the GS FLX Titanium platform. During the analysis, a previously unreported GC bias in pyrosequencing data was identified, which affected sequences from both sequencing runs, thus indicating the importance of thorough data screening and filtering to avoid a contortion of results. However, sequence data generated using the GS FLX Titanium chemistry was only marginally affected by this issue and only a very small fraction of reads had to be excluded from further analysis. These differences result from improvements in the Titanium chemistry, leading to less bias compared to the previous GS FLX technology. Meanwhile, a new emPCR kit containing a specific additive has been launched by Roche Applied Science. Initial studies based on microbial genome sequencing revealed an almost bias free sequencing of even very high GC content regions (data not shown). The composition of the microbial community was deduced from both the taxonomic classification of 16S rRNA fragments as well as the assignment of Environmental Gene Tags (EGTs) on different taxonomic ranks Obtained results essentially confirmed taxonomic profiles of the previous study. However, less abundant taxa could be identified by analyzing the GS FLX Titanium dataset thus justifying the additional sequencing effort. The fact that only a very small fraction of metagenomic reads actually contains fragments of the 16S rRNA gene emphasizes the advantages of using software such as the CARMA pipeline, which accurately classifies gene fragments detected in metagenome sequence data. A rarefaction analysis was performed to estimate the coverage of the microbial community in both sequencing datasets; as expected, sequencing data from the GS FLX Titanium platform provides a far more complete view of the underlying community, while the GS FLX sequencing run was not carried out to saturation. During the functional characterization of the community, members of the phylum Firmicutes could be confirmed to represent the dominant organisms involved in the breakdown of polysaccharides together with Bacteroidetes . Beyond that, the novel analyses showed that Proteobacteria also play an important role in polysaccharide degradation. Clostridia were found to dominate within the functional context ‘acetogenesis’, as deduced by mapping of bacterial taxa to metagenome hits representing the Wood-Ljungdahl pathway which is also known as the reductive acetyl-CoA pathway. Methanomicrobiales are the most abundant order involved in methanogenesis using CO as a carbon source, while acetate only seems to play a minor role, as indicated by a low fraction of Methanosarcinales . Based on the identification of 16S rRNA fragments from the Methanoculleus genus and subsequent assembly, the presence of several Methanoculleus species closely related to Methanoculleus bourgensis in the studied biogas fermenter could be demonstrated. A rough characterization of the genomic content of these Methanoculleus species was conducted by mapping the metagenome sequence reads to the published genome of Methanoculleus marisnigri JR1. Comparison of the genomic content of dominant Methanoculleus species within the analysed sample and the reference species M. marisnigri JR1 revealed that there are several differences mainly concerning genes that might have been acquired by horizontal gene transfer. Metagenome reads assigned to the genus Methanoculleus represent inter alia methanogenesis and membrane-bound hydrogenase genes predicted to be of importance for the pathway leading to the formation of methane within biogas. The close relationship of the Methanoculleus species in the studied biogas fermenter makes reconstruction of the genomic sequence of one of the dominant Methanoculleus species from the metagenomic sequence reads rather unlikely, since a reliable distinction between the most abundant strains can not be assured. In comparison, the GS FLX Titanium data offers a far more complete view of the analysed fermenter, even though analysis results give evidence that the available sequence data still does not fully cover the microbial community."
} | 3,114 |
35754893 | PMC9168828 | pmc | 1,265 | {
"abstract": "Superamphiphobic coatings may significantly change the wettability of a substrate, and so are attractive for applications in aero/marine engineering, biotechnology, and heat transfer. However, the coatings are caught in a double bind when their durability is considered, as they are vulnerable to mechanical abrasion. Meanwhile, the wide use of organic solvents for preparing the coatings generates environmental pollution. Here, we present a waterborne superamphiphobic coating through one-step spraying that repels a wide range of liquids. By tailoring the repellence of the nano-silica to waterborne resin, a network structure is constructed to protect the embedded nano-silica from damage. Thus, the coatings are durable against 725 cycles of friction tester abrasion under a load of 250 g, showing a significant improvement in the mechanical durability by 3–25 times. Moreover, our coating also shows excellent comprehensive durability, including resistance to oil-flow erosion, falling sand impact, chemical attack, thermal treatment, etc. This strategy can be introduced to various waterborne resins, demonstrating its universality, and may offer a new insight to design sustainable superamphiphobic coatings for long-term practical applications.",
"conclusion": "4 Conclusions In summary, we demonstrate a facile, scalable, and sustainable approach to creating waterborne superamphiphobic coatings. By controlling the repellence of fluorinated nano-silica to waterborne resin, the coatings consist of the network structure for resisting damage and the embedded fluorinated nano-silica for repelling a wide range of liquids. Under the protection of network structure, the coatings preserve their superamphiphobicity after 725 cycles of friction tester abrasion under a load of 250 g, boosting mechanical durability by 3–25 times, compared to those conventional coatings. Meanwhile, the coating demonstrates superior mechanical stability and long-term durability towards comprehensive tests under various harsh environments. Moreover, this strategy can be applied to various waterborne resins to improve the universality of superamphiphobic coatings. The findings provide a new design to strengthen the mechanical stability of waterborne superamphiphobic coatings and show the wide application prospects in fuel transportation, anti-fouling, architecture, and other fields, especially for those that require durability under harsh conditions.",
"introduction": "1. Introduction Superamphiphobic coatings with the ability to repel liquids are of interest for fundamental research and practical applications in anti-fouling, 1–4 petrochemical engineering, 5–7 corrosion protection, 8,9 and heat transfer. 10–13 Characterizing with rough textures and low-surface-energy chemistry, these coatings enable a liquid drop on their surface to bead up and roll off readily. 14–19 Truly harnessing these preferred functions also calls for simultaneously preserving their mechanical durability; however, displaying such two features appears mutually exclusive. 20,21 This is because the liquid repellence requires a minimized interfacial contact area which typically takes advantage of trapping air pockets within refined structures, but makes a trade-off in mechanical durability, especially under abrasion conditions. 22,23 Meanwhile, the exposure of the underlying layer by abrasion may change the chemistry property of the surface, for example-the wettability, thereby leading to the adhesion of liquids to the surface. 24,25 Diverse strategies have been developed to enhance the robustness of superamphiphobic surfaces. For example, adhesives, such as organic polymer or aluminum dihydrogen phosphate, can be introduced to improve the adhesion between the functionalized nanoparticle coating and the substrate. 26–28 However, the coatings lose their performance once the upper layer is abraded off, thereby leading to a moderate enhancement in durability. Alternatively, the strategy of the self-similar structure can achieve durability by sacrificing the upper layers, 29–31 which in turn compromises the lipophobicity. Currently, constructing a rigid armour can bear the mechanical load, 24,32–34 but poses a strict requirement on the specific substrate and advanced microfabrication technologies, which are low-throughput, time-consuming, and cost-ineffective. Besides, the preparation of most superamphiphobic surfaces depends on organic solvents to dissolve low-surface-energy substances, 30,35 which leads to environmental pollution. Although water is a green and safe choice as the solvent, 36 the incompatibility between the water and low-surface-energy substances impedes the preparation of waterborne superamphiphobic coatings. Consequently, most efforts have been focused on the preparation of the waterborne superamphiphobic coatings, 37–39 but rarely on the breakthrough in the durability. To date, it remains challenging to fabricate coatings while achieving both superamphiphobicity and durability by simple fabrication processes, especially for waterborne coatings. 40 Here, water and waterborne resins are selected to prepare the superamphiphobic coatings. Although the repellence of the fluorinated nano-silica to waterborne resin was regarded as the barrier to preparing waterborne superamphiphobic coatings, we utilize and control this property to the optimized one to promote the shrinkage of waterborne resin ( Fig. 1 ). Then, the frame unit forms with cavities. After layer-by-layer spraying, the network structure is constructed with the embedment of fluorinated nano-silica. On the one hand, the embedded fluorinated nano-silica provides superamphiphobicity. On the other hand, the network structure can bear the external load to protect the embedded nano-silica, which promises the mechanical and environmental durability of superamphiphobic coatings. The new design principle for waterborne superamphiphobic coatings features the following advantages: (1) the coatings can be prepared by various waterborne resins in a facile and sustainable manner. (2) The coatings repel various liquids. (3) The coatings show a significant improvement in mechanical durability by 3–25 times and can withstand comprehensive harsh conditions. Fig. 1 Schematic illustration of network coating formation.",
"discussion": "3 Results and discussion 3.1 Design of network coatings We first tailored the wettability of fluorinated nano-silica to obtain the optimized repellence to waterborne resin ( Fig. 2a ). The fluorinated nano-silica turned hydrophilic due to the fluorocarbon surfactant (determined by the hydrophilic carboxyl groups), which was used to help water dissolve low-surface-energy substances (Fig. S1a † ). This altered surface chemistry, which has been widely regarded as the failure in superamphiphobicity, was investigated and then utilized to tailor the wettability of fluorinated nano-silica by controlling the decomposition of fluorocarbon surfactant (Fig. S1b † ). Thus, the fluorinated nano-silica showed totally different wettability after heat treatment at different temperatures (Fig. S1c † ). Fig. 2 Design of network coatings. (a) Wettability control of the fluorinated nano-silica by heat treatment at different temperatures. (b) TEM image showing the shrinkage of the waterborne resin driven by the fluorinated nano-silica. (c) Surface morphology of the network coating. The inset of the SEM image shows the embedment of fluorinated nano-silica in the network structure. (d) Element maps of F and Si of the network coating. (e) The contact angles and roll-off angles of liquids with various surface tensions. The inset of the optical photograph shows the repellence of the network coating to various liquids. The fluorinated nano-silica with a water contact angle of ∼125° ( i.e. , treated at 160 °C) repelled the waterborne FEVE resin and promoted its shrinkage, which is therefore the reasonable choice ( Fig. 2b ). The shrunken waterborne resin served as the skeleton to surround the fluorinated nano-silica and gradually constructed the network structure by layer-by-layer crosslinking, which was evidenced by the increasing roughness (Fig. S2 † ). Finally, the network structure formed with a stable roughness after heat treatment at 180 °C ( Fig. 2c ). Amounts of cavities with a diameter of 1–15 μm were uniformly distributed on the coating, and several small cavities were contained within the large cavities, showing an interconnected frame, which seemed like a network. Note that, the continuous interconnected frame would be broken if further increasing the coating treatment temperature, thereby leading to the crack formation on the network structure (Fig. S3 † ). Moreover, it can be observed that the fluorinated nano-silica was embedded in the network structure. Due to the repellence, the fluorinated nano-silica avoided the excessive package by waterborne resin, thereby contributing to the roughness. However, it was impossible for constructing such a structure by using the fluorinated nano-silica with either insufficient or extreme repellence (Fig. S4 † ). For example, the fluorinated nano-silica treated at 100 °C possessed a water contact angle of ∼58°, showing an affinity to waterborne resin. The resultant coating showed a smooth structure with a few cavities. Conversely, the fluorinated nano-silica treated at 180 °C achieved superamphiphobicity, meaning the excessive repellence to waterborne resin. The incompatibility led to the difficulty in mixing them. Thus, the coating showed a loose structure without the continuously interconnected bridges ( i.e. , network structure). For comparison, the layered coating was prepared according to the conventional strategy (Fig. S5a † ). The layered coating exhibited a relatively smooth structure (Fig. S5b † ). Such a difference is evidenced by the porosity analyses, that the network coating possessed much more cavities than the layered coating, especially the micro-scale ones (Fig. S5c and d † ). Note that, the cavities below 10 nm were formed by the stack of nano-silica. \n Fig. 2d and S6a † show the surface chemistry of the network coating. The corresponding distributions of F and Si elements demonstrated the successful graft of FDTS on nano-silica, which provides low surface energy. Besides, the black regions in the maps referring to the cavities can also be observed. The FTIR spectra further verified the successful graft, that the peak intensity of –OH groups became weaker and a new peak of –CF 2 – groups emerged after nano-silica fluorination (Fig. S6b † ). In addition, the peak of C \n \n\n<svg xmlns=\"http://www.w3.org/2000/svg\" version=\"1.0\" width=\"13.200000pt\" height=\"16.000000pt\" viewBox=\"0 0 13.200000 16.000000\" preserveAspectRatio=\"xMidYMid meet\"><metadata>\nCreated by potrace 1.16, written by Peter Selinger 2001-2019\n</metadata><g transform=\"translate(1.000000,15.000000) scale(0.017500,-0.017500)\" fill=\"currentColor\" stroke=\"none\"><path d=\"M0 440 l0 -40 320 0 320 0 0 40 0 40 -320 0 -320 0 0 -40z M0 280 l0 -40 320 0 320 0 0 40 0 40 -320 0 -320 0 0 -40z\"/></g></svg>\n\n O groups (corresponding to the hydrophilic carboxyl groups as evidenced in Fig. S1a † ) disappeared on the fluorinated silica after heat treatment at 180 °C, thereby removing the influence of the fluorocarbon surfactant on the superamphiphobicity. The entire coating was covered by –CF x groups with low surface energy. Under the synergistic contribution of rough texture and low surface energy, the network coating, which was thicker than 45 μm, was super-repellent ( i.e. , contact angle >150° and roll-off angle <10°) to a wide range of liquids, even the decane with a low surface tension of 23.8 mN m −1 ( Fig. 2e , S7 and Table S1 † ). These drops of water, olive oil, hexadecane, ethanol solution (60 wt%), and decane sitting on the network coating were nearly spherical (inset in Fig. 2e ). Moreover, our coating not only realized the superamphiphobicity in air, which allowed the falling decane droplet to bounce off the surface without adhesion, but also repelled the water in oil (Fig. S8 † ). Note that, due to the sufficient temperature of 180 °C to totally remove the fluorocarbon surfactant, the network coatings treated at higher temperatures showed a similar superamphiphobicity to that treated at 180 °C, despite crack formation on their network structure (Fig. S9 † ). Significantly, the reported strategies of the self-similar structure were achieved by mixing superamphiphobic particles and adhesive/matrix. It leads to the excessive package of superamphiphobic particles by adhesive/matrix, which sacrifices the oleophobicity. However, our coating utilizes the controlled repellence of fluorinated nano-silica to waterborne resin, thereby mitigating the challenge and exhibiting a similar superamphiphobicity to the layered coating, whose fluorinated nano-silica was sprayed separately from the adhesive. The new design concept of tailoring the repellence of fluorinated nano-silica also endowed our coating with universality. Our strategy could be introduced to various waterborne resins, whether oleophilic or oleophobic, and flexible or rigid ( Fig. 3a and Table S2 † ). For example, the network coating achieved superamphiphobicity even using the waterborne epoxy resin with a water contact angle of ∼34.7° and an olive-oil contact angle of ∼15.8°. As shown in Fig. 3b , the spherical droplets of ethanol solution (60 wt%) stood on various network coatings. However, most reported durable superamphiphobic coatings were prepared by the specific or optimized matrix/adhesive after amounts of experiments. 27,28,31 From an engineering perspective, our coating could be applied to the propeller with a complex shape and a size of 0.55 m, demonstrating the simple and large-scale fabrication ( Fig. 3c ). The coated propeller was endowed with the self-cleaning property. The dust on the coating surface was removed by the ethanol solution flow and left a clean surface as described in Fig. S10. † Moreover, the versatility was demonstrated by preparing the network coating onto various substrates, such as rigid, flexible, smooth, and porous substrates, proposing the potential in various applications (Fig. S11 † ). We also fabricated the network bulk and film for friendly use, which could be attached to the substrates when required ( Fig. 3d and S12 † ). The spherical droplets of ethanol solution stood on the upper and inner surface of the bulk, showing the all-dimensional superamphiphobicity. Besides, the network bulk and film with excellent flexibility could tolerate more than 100 cycles of bending without deformation. Fig. 3 Universality of network coatings. (a) The contact angles of olive oil on various waterborne resins and the corresponding network coatings. The network coating achieves superamphiphobicity even using the waterborne resin with an olive-oil contact angle of ∼16°. (b) Optical photographs of network coatings prepared by various waterborne resins, including fluorocarbon, epoxy, polyurethane, and acrylic resin. (c) Optical photograph of the network coating on the propeller with the complex shape. (d) Optical photograph of the network film with excellent flexibility. Fig. 4 Comprehensive durability of network coatings. (a and b) Water and olive-oil repellence of different coatings after friction tester abrasion. The inset of the schematic illustration shows the friction tester. (c) Comparison of the abrasion resistance of the network coating with that of the existing reports. (d) Olive-oil repellence of different coatings after oil stirring. (e) Comprehensive durability of the network coating. (f) The repairable property of the network coating. 3.2 Comprehensive durability The durability of superamphiphobic coatings is crucial to practical applications. For example, repeated abrasion is common in daily life. Here, we first performed the friction tester abrasion under a load of 250 g. Consistent with our prediction, the network coating using both 180 °C heat treatment and fluorinated nano-silica treated at 160 °C showed the best abrasion resistance, which was used to examine the durability (Fig. S13 † ). As shown in Fig. 4a and b , our coating with a thickness of ∼80 μm exhibited a stable superamphiphobicity during the abrasion. The contact angles and roll-off angles of water and olive oil were still greater than 150° and less than 10°, respectively, even after 600 cycles of abrasion. Further increasing abrasion to 1000 cycles, water droplets can still easily roll off. In sharp contrast, the layered coating with the thickness of ∼100 μm showed a much faster deterioration of superamphiphobicity than the network coating and effortlessly lost the superamphiphobicity within 100 cycles of abrasion, thereby leading to the liquid adhesion. It is well known that coating durability relates to the coating thickness. We then determined the critical abrasion cycles—that is, the maximum number of abrasion cycles that the coating could withstand without a loss of performance, and showed the index of the network coatings with different thicknesses (Fig. S14 † ). The anti-abrasion capacity was enhanced with the thickness increment. From an engineering perspective, the relatively thin coating with a thickness of ∼80 μm was a reasonable choice, which achieved both satisfactory superamphiphobicity and durability. Therefore, we compared this index with various reported durable coatings, 26,27,29,42,43 whose minimum thickness was ∼100 μm ( Fig. 4c ). The critical abrasion cycles of our coating reached 725, which was 3–25 times higher than that of the reported coatings, demonstrating a significant improvement in mechanical durability. Note that, such an improvement effect was also applicable to other kinds of waterborne resins, meaning that the universality of our strategy covered both superamphiphobicity and mechanical stability (Fig. S15 † ). We also performed oil stirring tests to demonstrate the durability of our coatings against soft mechanical damage. As shown in Fig. 4d , the network coating could tolerate the erosion of olive-oil flow for 96 h, which was 8 times longer than the layered coating. More comprehensive durability tests were conducted to simulate various environments in practical applications, including Taber abrasion for mechanical abrasion (400 cycles under a 250 g load), the sand impact for mechanical impact (40 g min −1 for 60 min), the sandy water stirring for flow erosion (15 m s −1 for 120 h), the chemical attack (immersion in 3.5 wt% NaCl solution for 14 days) and the heat treatment (250 °C for 7 days) ( Fig. 4e ).In addition, we also carried out the chemical attack tests induced by various corrosive solutions (pH value of 1–13 for 48 h) and the thermal stability tests at temperatures ranging from −196 to 350 °C for 2 h (Fig. S16 † ). Our coatings could withstand all these extremely harsh conditions, maintaining the roll-off angles of olive oil less than 10°. Significantly, the excellent durability of the network coatings could be realized not only on the glass substrate but also on the steel substrate (Fig. S17 † ). Note that, coating adhesion is also an important factor for durability. The cross-hatch test was performed on the coating followed by tape peeling (ASTM D3359-17 standard). Almost no removal of coating can be observed, demonstrating the strong adhesion of the network coating (Fig. S18 † ). Moreover, our coating was endowed with the repairable property, which further improved the reliability for practical applications. As shown in Fig. 4f , the network coating turned oleophilic after oxygen plasma treatment. Then the superamphiphobicity was regenerated by sandpaper abrasion. The procedure of oxygen plasma treatment and abrasion was counted as one cycle. Our coating could be repaired over 10 cycles, without deterioration of superamphiphobicity. 3.3 Mechanical stability mechanism The superior durability of our coating is attributed to the network structure. As shown in Fig. 5a , during abrasion, the network structure tolerates the external force and prevents the embedded nano-silica from removal. Such protection could be evidenced by the surface morphology of the abraded network coating ( Fig. 5b ). Several cavities (5–75 μm) could be observed, which are much smaller than the abradants. Although the skeletons (superficial structure) underwent the severe damage, the nano-silica remained intact within the cavities to provide superamphiphobicity. The AFM image also shows the rough structure within the cavities, further demonstrating the protection ( Fig. 5c ). However, the adhesive of the layered coating only plays a role in binding the nano-silica. Without the structure for protection, the nano-silica is vulnerable to mechanical abrasion ( Fig. 5d ). Therefore, the layered coating turned smooth after abrasion, and the nano-silica was almost completely removed, leading to the absence of roughness and low-surface-energy chemistry ( Fig. 5e and f ). We also used a micro-scratching test to visualize the behavior of the coating surface under scratching (Fig. S19a † ). With a tip size of 5 μm, the indenter applied a 400 MPa load to the coating surface, which is much more severe than daily abrasion. Even so, the network structure tolerated the major stress and kept the nano-silica on the surface (Fig. S19b † ). In contrast, the upper layer of the layered coating was entirely scratched off, leaving the adhesive layer alone (Fig. S19c † ). Fig. 5 Mechanical stability mechanism. (a) Schematic illustration showing the protection of the network structure for embedded nano-silica. (b and c) SEM and 3D-view AFM images of the network coating after abrasion. (d) Schematic illustration of the easy removal of fluorinated nano-silica on the layered coating. (e and f) SEM and 3D-view AFM images of the layered coating after abrasion. (g) Wide-scan XPS spectra of the surface elements of the network coating and layered coating before and after abrasion. (h) The change in coating roughness along with abrasion. All coatings were abraded by 600 cycles of friction tester abrasion. Such differences can be demonstrated from both aspects of chemical component and roughness. For the network coating, the reduction of Si content, which represents fluorinated nano-silica, was only 3% even after 600 cycles of abrasion ( Fig. 5g ). This is a negligible contribution to the degradation of superamphiphobicity. However, the loss of nano-silica on the layered coating reached 33%, indicating 11 times higher than that on the network coating. Meanwhile, the roughness change of the network coating was also much slower than that of the layered coating ( Fig. 5h ). After 600 cycles of abrasion, the network coating still kept its roughness larger than 100 nm, indicating strong structural stability. The result posed a sharp contrast to that of the layered coating, whose roughness was rapidly reduced to only ∼30 nm within 200 cycles of abrasion. Together, these microscopic results demonstrate the notable resistance of network structure that significantly enhances mechanical durability."
} | 5,785 |
23269911 | PMC3529302 | pmc | 1,266 | {
"abstract": "This paper presents a digital silicon neuronal network which simulates the nerve system in creatures and has the ability to execute intelligent tasks, such as associative memory. Two essential elements, the mathematical-structure-based digital spiking silicon neuron (DSSN) and the transmitter release based silicon synapse, allow us to tune the excitability of silicon neurons and are computationally efficient for hardware implementation. We adopt mixed pipeline and parallel structure and shift operations to design a sufficient large and complex network without excessive hardware resource cost. The network with 256 full-connected neurons is built on a Digilent Atlys board equipped with a Xilinx Spartan-6 LX45 FPGA. Besides, a memory control block and USB control block are designed to accomplish the task of data communication between the network and the host PC. This paper also describes the mechanism of associative memory performed in the silicon neuronal network. The network is capable of retrieving stored patterns if the inputs contain enough information of them. The retrieving probability increases with the similarity between the input and the stored pattern increasing. Synchronization of neurons is observed when the successful stored pattern retrieval occurs.",
"conclusion": "4 Conclusion We have reported our silicon neuronal network based on the digital operational circuit, which can be efficiently implemented in a FPGA device. Our silicon neuron is implemented by using the DSSN model where neuronal behaviors are abstracted using mathematical techniques so that it is capable of realizing behaviors in both Class I and II neurons with small number of multipliers to reduce the hardware resource requirement. Because the state variables are not reset in the spiking dynamics, it is expected that this model can describe the dependence of spike waveform on the stimulus far more effectively than the LIF and the IZH models where resetting of the variables is one of the points that reduce the complexity in their models and their implementations. It restricts dynamics in the spikes by assuming their maximum values are uniform. The IZH model actually can realize various neuronal activities including that of Class II neurons, which the LIF model cannot, though its spikes have a very similar waveform (Figure 3 C). Meanwhile, simplicity in our model is maintained by reducing the number of multiplications, which is effective to realize compact digital circuit implementations. By utilizing the techniques of the phase plane and the bifurcation analyses, we successfully found the parameters for Class I and II models where the coefficients are selected in powers of two or sums of two such numbers. It allows us to replace the multiplications in our model by shift and add operations except for calculation of the square of variable v . Thus, our model can be implemented with single multiplier, which is the same to in (Cassidy and Andreou, 2008 ). Their circuit cannot realize the graded response of Class II neurons because they are implementing the IZH model. In addition, our silicon neuron circuit can be expanded to a 3-variable version with no additional multiplier, which can produce autonomous bursting similar to in the IZH model (Kobayashi et al., 2011 ). It was also shown that our model can reproduce very complex neuronal behaviors including chaotic ones by the fixed point representation (Kohno and Aihara, 2007 ) that requires less hardware resource than the floating point representation, which are used in (Thomas and Luk, 2009 ). Our silicon synapse model qualitatively describes process of the transmitter release, receptor activation, and generation of synaptic current described in the kinetic models (Destexhe et al., 1998 ). It was implemented without using multipliers by selecting Δ t α and Δ t β in power of two numbers. The salient feature of the synaptic output in our model is the time course of rise and decay that is dependent on the spike width. However it is usually neglected in other silicon neuronal networks (Cassidy and Andreou, 2008 ; Thomas and Luk, 2009 ; Arthur et al., 2012 ). We constructed a fully connected network of 256 neurons on a Digilent Atlys FPGA board equipped with a Xilinx Spartan-6 LX45 FPGA. Calculating one step of a neuron needs 257 multiplications and 267 additions. This large amount of calculation was solved by the pipelined and parallel structure based on the tradeoff between hardware resource requirement and updating speed of the network. The functionality of our silicon neuronal network and the significance of the Class II model in our silicon neuron were demonstrated by an auto-associative memory task. Its performance was evaluated by storing 4 patterns and applying inputs similar to them but including errors. The result shows that our silicon neuronal network has potential of retrieving the stored pattern even when the input pattern contains error and the neurons fire synchronously in case of successful retrieving. The Class II mode network has higher retrieving probability than the Class I mode network which is caused by differences of dynamical properties in Class I and II neurons. We can expect that the retrieving probability of our network is better than (Arthur et al., 2012 ) because it can only realize Class I neurons. It is known that one of the major difference between Class I and II neurons is the dependence of the spike form on the input strength. In Class II neurons, it depends strongly on the input, whereas it is almost constant in Class I neurons. We expect that this difference is playing at least a partial role in the performance of the auto-associative task, which will be elucidated in our future works. In this paper, our silicon neuronal network was tested only in its fully connected network topology without any adaptive learning rules. Because any connecting topology can be realized by disabling appropriate connections in the fully connected one, our system is ready to be tested in any settings. Under such restrictions, our results support that our silicon neuronal network can execute one of the most fundamental tasks for the neuronal networks and its distinctive feature of realizing Class II neurons can improve retrieving performance. The custom neuromorphic chips investigated by both FACETs and NeuroGrid are the compact analog circuits which are composed of the silicon neurons based on the detailed neuron models (Bruderle et al., 2010 ; Choudhary et al., 2012 ). Their circuit is compact and consumes lower power but is generally sensitive to the noise and the fabrication mismatch. SpiNNaker simulates the detailed neuronal model which runs in the software on the embedded ARM processors with a high speed clock (Jin et al., 2010 ). Such systems generally consume higher power in comparison to our silicon neuronal network which is based on the optimized for implementation model and implemented by the compact and dedicated hardware running with a low speed clock. In SyNAPSE, the digital circuit implementation of the LIF model is used (Arthur et al., 2012 ), which we already have mentioned above. For real-time operation as an artificial nerve system, our silicon neuronal network requires very low clock frequency around several thousands of kilo hertzs. Each silicon neuron has 256 synaptic inputs, which will be increased up to about 10,000, which is a typical number of synaptic connection of a neuronal cell in the neocortex. In such case, the clock frequency will be about a hundred mega hertz (about 40 times faster than current frequency). Digital circuits with such range of clock frequency can be implemented by the near-threshold logic technology which consumes very low power. And it also consumes less power in cheap FPGAs. Thus our system can be suitably applied to robot controllers and compact intelligent sensor devices. For example, there is a possibility that our silicon neuronal network is connected to the event-based biomimetic sensors via additional silicon synapses dedicated to external inputs and realize an intelligent sensor such as retina-like image sensors. On the other hand, our silicon neuronal network can operate much more faster than the nerve system even in entry-level FPGA devices. Actually, our system can operate with 100 MHz system clock, which 40 times accelerated in comparison to the real-time operation. Thus, our system can be applied as a high speed simulator of neuronal networks composed of the qualitative neuron models, which is utilized as an important tool for the connectionists. Compared with the event-based network (Chicca et al., 2004 ), our network is expected to catch sensitive event information for its high speed operation and low power consumption. In our future works, we will evaluate performance of our silicon neuronal network in the auto-associative task more in detail from theoretical viewpoints. It includes the comparison of the performance between our DSSN model’s and the IZH model’s networks as well as the evaluation of the memory capacity and the effect of introducing the STDP learning rules into both of the networks. This will elucidate more clearly how the neuron classes affect the performance of the auto-associative memory task. The large-scale network will also be pursued that can be implemented in a single FPGA chip, which will be applied to realizing intelligent sensors including retina-like image sensors. We expect possibility that the selectivity of neuron classes in our silicon neuronal network can improve such devices.",
"introduction": "1 Introduction The nervous system transmits signals by cooperation between neurons and synapses. The neuron generates an overshoot of its membrane potential (spike) when stimulated by a sufficient large current. The waveform is distributed to the synapse and causes neuronal transmitters to be released. The information processing in the nerve system is autonomous, flexible, and robust against various signal distortions. The silicon neuronal network is designed to reproduce activities of the nerve system in real-time. Compared to the current computers, the silicon neuronal network is based on the parallel and distributed processing mechanism rather than the serial centralized framework. This distinctive computational style is expected to allow real-time and large-scale processing of advanced task similar to that in the nerve system (Mallik et al., 2005 ; Mitra et al., 2009 ). Besides, the hybrid network constructed with the silicon and the biological neurons is investigated to learn complex behaviors in neurons (Le Masson et al., 2002 ). A silicon half-center oscillator composed of silicon neurons is proposed for application as an embedded biomedical device and a motion controller (Simoni and DeWeerth, 2007 ). The ionic-conductance-based model of a neuronal cell describes its dynamics of ions and ionic channels as exactly as possible. Though equations of this type of models are generally complex, it can reproduce neuronal dynamics considerably precisely. Success of the first one, the Hodgkin-Huxley model (Hodgkin and Huxley, 1952 ), gave rise to various neuron models of this type. Silicon neurons that implement this type of models can reproduce the complex neuronal behaviors, including bursting, tonic firing, and so on (Mahowald and Douglas, 1991 ; Simoni et al., 2004 ; Yu and Cauwenberghs, 2010 ). The integrate-and-fire (IF) model aims to describe the spike generation in neurons with simple equations without taking ionic dynamics into account (Lapicque, 1907 ). Later, a leakage term was incorporated to describe attracting nature of the resting state, which formulated the leaky IF (LIF) model. It is an efficient and compact model but with the tradeoff of dynamics. Some of neuromorphic chips that implement neuronal networks with LIF neurons are low power for real-time simulation and conveniently applicable to various applications, optimization, recognition, and memory (Chicca et al., 2007 ; Chakrabartty, 2010 ; Arthur et al., 2012 ). Several efforts to reduce the limitation in the dynamics of the LIF model resulted to expanded LIF models including generalized (Jolivet et al., 2004 ), exponential (Brette and Gerstner, 2005 ), and quadratic (Izhikevich, 2006 ) IF models. They were implemented to realize simple silicon neurons that can produce variety of neuronal activities such as spike-frequency adaptation and autonomous bursting (Rubin et al., 2004 ; Indiveri et al., 2010 ; van Schaik et al., 2010 ). However, the limited structure in the LIF model prevents realizing the property of Class II neurons in the Hodgkin’s classification. A quadratic IF model proposed by Izhikevich (IZH) successfully simulates a wide variety of neuronal activities by combination of a two-variable differential equation and reset of the state variables. Whereas most of the above silicon neurons are realized by the analog electronic circuit technology, there are several digital circuit implementations of the LIF (Indiveri et al., 2011 ) and the IZH (Cassidy and Andreou, 2008 ; Thomas and Luk, 2009 ) models. One of them succeeded to realize a large-scale network with 1024 neurons on a single FPGA chip. In these implementations of the IZH model, the equations are solved using the floating point operation. One of the important points in realizing digital silicon neurons that can simulate various neuronal activities with compact and simple circuits, is to select a neuron model with such capability and find a suitable circuit for its implementation. The IZH model is considerably a good selection because its non-linearity is only second order and it can be implemented with fewer multipliers than other models with the similar capability. This model, however, is not fully capable of realizing the graded responses to the stimulus of Class II neurons. This is because the IZH model approximate the spike process by reset of the state variables, which leads to very similar spikes in response to various stimulus. For example, the maximum membrane potential values in spikes are uniform (30 mV). Another neuron model named a mathematical-structure-based Digital Spiking Silicon Neuron (DSSN) model was proposed (Kohno and Aihara, 2007 ). This model was designed to simulate several classes of neurons by simple digital arithmetic circuits. It was demonstrated that complex behaviors similar to those in a brain area can be reproduced by an implementation by fixed point operation circuits, which is expected to reduce the hardware resource requirement in circuit implementation. Because this model does not approximate the spike process by the reset of the state variables, it can realize more effectively the graded responses of Class II neurons than the IZH model. Because the transmitter release at the chemical synapses is controlled by the membrane potential at the axon terminal, the graded response property of the neuronal cells is reflected to the amount of synaptic transmitter release, which is modeled in our silicon synapse as illustrated in Figure 3 . With the DSSN model in Class II mode, the information of input signal is more directly reflected in the transmitter release than the other 2 models. With the neuronal models with resetting of the state variables including the IZH model (Figure 3 C), this property is almost ignored although there is a possibility that it plays some roles in the information processing in the nerve system. We implemented a network of the DSSNs and silicon synapses on a FPGA device. We have developed a silicon synapse model based on the kinetic ones in (Destexhe et al., 1998 ) that describes the transmitter release in the presynapse and information of duration of a spike. To demonstrate that our implementation is operating appropriately, we executed an auto-associative memory task, retrieving a memorized pattern by its fragments, which has been widely investigated theoretically (Hopfield, 1982 ; Knoblauch and Palm, 2001 ; Sudo et al., 2009 ). Behaviors of the associative memory in our network are evaluated by an overlap index (Domany and Orland, 1987 ; Aoyagi, 1995 ). Synchrony of neurons is also investigated by another index, the phase synchronization index (PSI; Rosenblum et al., 2001 ). Similar retrieving ability is also shown in a network (Arthur et al., 2012 ) which is composed by the LIF model based silicon neurons. The LIF model can realize only Class I neurons whereas the DSSN model in our network can simulate both Class I and II neurons by selecting appropriate parameters. In this paper, we report the comparison between the performance of auto-associative tasks in the networks composed of Class I and II neurons. This paper is organized as follows: In the second section, the model of our silicon neuron and its bifurcation structure are introduced firstly. Then the model of our silicon synapse is presented. We explain the architecture of the implementation of our silicon neuronal network thirdly, including its pipeline structure that improves the efficiency of circuit area occupation. Fourthly, we discuss our FPGA implementation and blocks of bidirectional data transfer with a PC. The experiment results and their analysis are followed as the third section. The conclusion section follows where summary, discussion, and views of our future work are presented."
} | 4,349 |
36134138 | PMC9417048 | pmc | 1,267 | {
"abstract": "Artificial synapses based on electrolyte gated transistors with conductance modulation characteristics have demonstrated their great potential in emulating the memory functions in the human brain for neuromorphic computing. While previous studies are mostly focused on the emulation of the basic memory functions of homo-synapses using single-gate transistors, multi-gate transistors offer opportunities for the mimicry of more complex and advanced memory formation behaviors in biological hetero-synapses. In this work, we demonstrate an artificial hetero-synapse based on a dual-gate electrolyte transistor that can implement in situ spatiotemporal information integration and storage. We show that electric pulses applied on a single gate or unsynchronized electric pulses applied on dual gates only induce volatile conductance modulation for short-term memory emulation. In contrast, the device integrates the electric pulses coincidently applied on the dual gates in a supralinear manner and exhibits nonvolatile conductance modulation, enabling long-term memory emulation. Further studies prove that artificial neural networks based on such hetero-synaptic transistors can autonomously filter the random noise signals in the dual-gate inputs during spatiotemporal integration, facilitating the formation of accurate and stable memory. Compared to the single-gate synaptic transistor, the classification accuracy of MNIST handwritten digits using the hetero-synaptic transistor is improved from 89.3% to 99.0%. These findings demonstrate the great potential of multi-gate hetero-synaptic transistors in simulating complex spatiotemporal information processing functions and provide new platforms for the design of advanced neuromorphic computing systems.",
"conclusion": "Conclusion In conclusion, we report a flexible artificial hetero-synapse based on dual-gate electrolyte gated transistors with spatiotemporal information integration and storage capability. The device can effectively capture and store the signals applied to the dual gates that occurred in synchronization, performing in situ coincident detection and memory function of biological hetero-synapses. Such a device can autonomously filter the random noise signals in the multiple inputs and promote the formation of accurate and stable memory, leading to enhanced classification accuracy of the objects. These results open up new avenues for the efficient implementation of complex biological hetero-synaptic functions in physical devices and can possibly be used in flexible intelligent wearable systems.",
"introduction": "Introduction The human brain is highly efficient in executing various complex cognitive tasks such as learning, memory and decision making, outperforming state-of-the-art digital computers. 1,2 Synapses are the fundamental building blocks of the human brain for learning and memory, and their strength (synaptic weight) can adaptively adjust in response to neural spikes for information processing. 3–6 For decades, great efforts have been devoted to building electronic devices for the emulation of various memory formation behaviors in synapses, dreaming of physically realizing artificial intelligence systems with learning abilities close to the human brain. 1,7,8 An electrolyte gated transistor is a typical device that consists of an ionic gating dielectric and channel layer made of semiconductors. 5,9,10 The voltage pulses applied on the gate terminal can tune the concentration of mobile ions around the channel layer, which modulates the channel conductance through ionic electrostatic or/and chemical doping effects. 8,11,12 Extensive studies have demonstrated the feasibility of the electrolyte gated transistor to emulate the memory formation of biological synapses, such as short-term and long-term memory effects, with excellent performances including good operation stability, low cycle to cycle variations, and high energy efficiency. 3,13–16 However, previous studies are mainly limited to the emulation of simple memory functions in homo-synapses using single-gate transistors, where the memory effects are solely dependent on the electric pulses applied on the single gate. 4,17–20 In contrast, biological hetero-synapses having sophisticated morphologies can effectively integrate the inputs delivered by multiple presynaptic terminals during memory formation, and the memory behaviors are determined by the spatiotemporal correlations among the stimuli. Although experimental demonstration of hetero-synaptic functions has been realized in multi-terminal memristors, the inputs applied on the additional terminals mainly play a role of modulation for homo-synaptic devices, rather than being used for information processing. 21,22 In this case, an artificial hetero-synapse to reproduce spatiotemporal information processing functions remains largely unexplored. Recent studies showed that a multi-gate electrolyte gated transistor can be obtained by incorporating additional gate terminals, 5,17,23–28 offering opportunities for the implementations of complex memory functionalities of hetero-synapses, such as in situ noise filters and storage that are critical for accurate and stable memory storage. In this work, we experimentally demonstrate an artificial hetero-synapse based on a dual-gate electrolyte gated transistor that is capable of computing and memorizing spatiotemporal information of multiple inputs. We show that electric pulses applied on a single gate or unsynchronized electric pulses applied on dual gates induce volatile conductance modulation and exhibit a short-term memory effect. In contrast, the application of electric pulses onto the two gate terminals in synchronization can lead to a super-linear summation of the device current and result in long-term memory formation, performing coincident detection computing and memory operations. Additional simulation results suggest that an artificial neural network built with such synaptic devices can effectively filter the random noise signals during information integration, helping to form accurate and stable memory. Compared to the single-gate synaptic transistor, the classification accuracy of MNIST handwritten digits is increased from 89.3% to 99.0%. These studies highlight the potential of the multi-terminal electrolyte gated transistor in emulating complex spatiotemporal information processing functions for efficient neuromorphic computing.",
"discussion": "Results and discussion \n Fig. 1a illustrates the schematic of a hetero-synapse with two presynaptic terminals and a postsynaptic terminal. The presynaptic terminals stimulated by neural spikes release neurotransmitters that bind to receptors on the postsynaptic terminal, generating an excitatory postsynaptic current (EPSC) and inducing memory formation. 29,30 Fig. 1b illustrates a dual-gate electrolyte gated transistor inspired by the biological counterpart, where the two gates and the drain/source channel correspond to the presynaptic terminals and postsynaptic terminal respectively. The voltage pulses applied on the gate and the drain/source current correspond to presynaptic spikes and postsynaptic current. The ionic transistor reported here employs [EMI][TFSA] as the ion-gel, where the mobile ions [TFSA] − play the role of neurotransmitters whose spatial distribution in the device is controlled by the electric stimulation conditions. The organic semiconductor P3HT is chosen as the channel material due to its good chemical stability, easy fabrication and compatibility with the ion gel. An optical image of the fabricated transistor device is illustrated in Fig. 1c . The fabrication process of the device is depicted in Fig. S1 † in the ESI. † Fig. 1d shows the transfer characteristic curve of the synaptic transistor, obtained by monitoring the drain current during the sweep of the gate voltage (1.0 V → −1.0 V → 1.0 V). A hysteretic window in the I D – V G curve at negative gate voltages was observed, and the conductance ratio reaches ∼1000, suggesting the effective modulation of the channel conductance. The output characteristic curve of the ionic transistor is systematically investigated, as shown in Fig. 1e . With the negative gate voltage applied, the drain current increases with the drain voltage, and the increasing rate is proportional to the gate voltage amplitude. These results proved the reliable modulation of conductance in the electrolyte gated transistor transistors. Fig. 1f shows that the characteristic peak positions in the Raman spectrum of the P3HT channel shift to lower values after device gating, indicating that the conductance modulation stems from the formation of positive polarons during the voltage gating induced electrostatic/electrochemical doping of the P3HT films. 31–34 Fig. 1 (a) Schematic of a biological hetero-synapse with two presynaptic terminals and a postsynaptic terminal. (b) Schematic of a dual-gate electrolyte gated transistor. (c) Optical image of the P3HT based dual-gate electrolyte gated transistor. The channel length is 100 μm. Scale bar: 500 μm. (d) Transfer characteristic curve of the transistor measured during gate voltage sweeping. (e) Output characteristic curves of the transistor at different gate voltages. (f) Raman spectra of the P3HT channel film before and after gate voltage application. (g) Scheme showing a bent electrolyte gated transistor. (h) The channel current of the device in response to an electric pulse with varying amplitude at different bending states. (i) The channel current as a function of the pulse width for the device at different bending states. Notably, owing to the intrinsic flexibility of the P3HT and [EMI][TFSA] ion gel as flexible organic materials, the ionic transistor device can be deposited on flexible substrates such as PET (polyethylene terephthalate). Fig. 1g shows the channel current of the device in response to gate pulses when the PET flexible substrate is flattened ( R = ∞), with a bending radius of 1 cm ( R = 1 cm) and a bending radius of 0.5 cm ( R = 0.5 cm) respectively. It shows that under different pulse stimulation conditions (amplitude and pulse width), the device shows negligible conductance variations under different bending conditions, indicating its potential for flexible electronic application ( Fig. 1h and i ). To evaluate the performance of the electrolyte gated transistor for memory function emulation, we studied the evolution of channel current with the gate pulses applied onto the gates. Fig. 2a shows the schematic of applying electric pulses on a single gate of the device, and a typical current response (EPSC) triggered by the gate pulse (−1.2 V and 100 ms) is shown in Fig. 2b . During pulse stimulation, the postsynaptic current increases from 8.9 nA to 190 nA quickly and then decays rapidly to the resting state within 500 ms. This result indicates that the voltage pulse drives [TFSA] − ions towards the P3HT interface during its application and these ions spontaneously diffuse back after the pulse is removed. This process emulates the short-term memory formation and can be explained by the formation of a volatile electric double layer at the P3HT surface that causes electrostatic doping effects. 35 The short-term dynamics enable the device to emulate the paired pulse facilitation (PPF) of synapses, where the second pulse (−1.2 V and 100 ms) closely following the first identical pulse can excite a higher device current ( Fig. 2c ). The inset in Fig. 2c plots the relationship between the PPF index and the time interval, where the PPF index refers to the ratio between A 2 and A 1 , and A 1 and A 2 represent the device current during the first and second pulse stimulations respectively. The PPF index decreases quickly with the increase of the time interval in the paired pulses, owing to the increased time interval that allows the back diffusion of more ions and prevents ion accumulation at the P3HT interface. The decay trend can be well fitted by an exponential decay equation illustrated in the inset in Fig. 2c . The characteristic decay time constant τ 1 is 203.83 ms, on the same order of magnitude as that in biological synapses. 30,36,37 Fig. 2d shows the EPSC current obtained in a device stimulated by pulses with different pulse frequencies (2.8 Hz, 3.3 Hz, 4.0 Hz, and 5.0 Hz). The peak EPSC current increases from 555.81 nA to 802.3 nA when the frequency increases from 2.8 Hz to 5.0 Hz ( Fig. 2e ). Likewise, with the pulse frequency fixed, the device current increases with the number of applied pulses ( Fig. 2f ). Fig. 2g shows that the peak EPSC current increases from 712.1 nA to 1690.1 nA when the pulse number increases from 15 to 50. To this end, the device possesses an intrinsic short-memory effect when programmed by the given electric pulses applied on a single gate. Notably, increasing the pulse amplitude can lead to a long-term memory effect, which was avoided in the implementations (Fig. S2, ESI † ). Fig. 2 (a) Schematic showing the electrostatic ion modulation of the channel layer in the electrolyte gated transistor under single-terminal voltage gating. (b) A typical EPSC obtained in the electrolyte gated transistor triggered by an electric pulse (−1.2 V and 100 ms). (c) EPSCs triggered by a pair of electric pulses with a time interval of 500 ms. A 1 and A 2 represent the current intensity after the first and second pulse stimulations, respectively. Inset: PPF index as a function of pulse interval (Δ t ); the curve is fitted with an exponential decay function. (d) EPSCs excited by five electric pulses (−1.2 V and 100 ms) with different frequencies. (e) Peak postsynaptic current and peak variation ( A t – A 1 ) as a function of the pulse frequency in (d). (f) EPSCs excited by electric pulses (−1.4 V, 0.5 s) with different pulse numbers. (g) Peak current and peak current difference ( A t – A 1 ) as a function of pulse number in (f). In biological neural networks, a postsynaptic terminal can receive excitatory synaptic inputs from multiple presynaptic terminals clustered by a dendritic tree. The strength of the inputs contributed by a single presynaptic terminal is usually insufficient to drive the membrane potential of the postsynaptic terminal to the threshold for stable memory formation. 23,24,30 However, the postsynaptic terminal can effectively accumulate and amplify the inputs from multiple presynaptic terminals, performing supralinear spatiotemporal signal integration. A schematic depicting this process is illustrated in Fig. 3a , showing the EPSC triggered by the two synchronized spikes separately delivered by two presynaptic terminals (red solid line), much higher than that produced by either one (pink dotted line). To explore the feasibility of implementing these effects in a dual-gate synaptic transistor, we studied the response of the synaptic transistor to pulse stimuli applied onto two gates. Fig. 3c illustrates the EPSCs excited by electric pulses (−1.2 V and 100 ms) applied to the presynaptic terminals, i.e. gate 1 ( G 1 ) and gate 2 ( G 2 ), respectively. Specifically, when the electric pulse was applied to G 1 or G 2 at different moments, each stimulation event excited an EPSC of ∼0.3 μA. In contrast, for the electric pulses applied to G 1 and G 2 simultaneously, we can detect a much higher EPSC (∼1.1 μA). Moreover, the difference in the current intensity increases with the pulse number ( Fig. 3d ). Fig. 3e plots the measured EPSC as a function of the expected sum for different pulse amplitudes (0.6 V, 0.8 V, 1.0 V, 1.2 V and 1.4 V). The expected sum (ES) is defined as the arithmetic sum of the EPSC when G 1 and G 2 were triggered separately (ES = A 1 + A 2 ), and the measured sum (MS) is the EPSC when the G 1 and G 2 were triggered simultaneously. It is apparent that the ES is much higher than the MS, indicating the supralinear integration of the synchronized inputs delivered by multiple gates. Importantly, we note that after experiencing dual-gate stimulation, the device shows a nonvolatile conductance modulation effect, exhibiting a long-term memory effect, as demonstrated in Fig. 3f . The observation is likely caused by the accumulation of abundant ions near the channel that penetrate into the P3HT layer to induce the electrochemical doping effect ( Fig. 3b ). The strong chemical bonding between the ion and channel layer prevents the back diffusion of ions to the gate dielectric after the removal of the applied pulses, which account for the improved retention performance, thus enabling long-term memory formation. 2,36 This process emulates the spatiotemporal integration of spike inputs of hetero-synapses during memory formation which promotes accurate and stable memory formation. Fig. 3 (a) Illustration of a hetero-synapse excited by electric inputs transmitted by two presynaptic terminals and the resulting EPSC as an output. (b) Schematic showing the electrochemical ion modulation of the channel layer in the transistor under dual-terminal voltage gating. (c) EPSCs excited by the presynaptic spike (−1.2 V and 100 ms) separately/simultaneously applied on G 1 and G 2 . (d) EPSCs excited by five presynaptic spikes (−1.2 V, 100 ms, and 3.3 Hz) separately/simultaneously applied on G 1 and G 2 . (e) The measured sum (MS) current plotted as a function of the expected sum (ES) current. (f) Time-dependent device current under different stimuli conditions. In Fig. 4 , we simulated an artificial neural network using such a multi-gate electrolyte gated transistor for MNIST (Modified National Institute of Standards and Technology) handwritten digits training and recognition. A fully connected artificial neural network (784 × 10) is established for the training and inference ( Fig. 4d ). The 784 input neurons correspond to a digit (0–9) image with 28 × 28 pixels, and the 10 output neurons correspond to the 10 digits. During the implementations, both gates act as the presynaptic terminals to transmit the signals delivered by the same neuron. Random noises are created and superimposed with the images transmitted by each presynaptic terminal. Two typical digit patterns with noises applied to the presynaptic terminals are illustrated in Fig. 4b . As the noise signals transmitted by each path are random and uncorrelated, they are mostly unsynchronized in spatial and temporal domains. The simulation result shows that when these digit patterns are delivered to the device, the noise components can be autonomously filtered with the overlapped patterns retained, producing a high-quality digit image with enhanced feature contrast ( Fig. 4c ). Fig. 4e shows the recognition rate as a function of the training epochs, for the dual-gate transistor base artificial neural network. A recognition rate of ∼99.0% is reached after ∼90 training epochs. A control experiment using the single-gate synaptic device for the training and inference was also executed. Owing to the increased probability of learning the noises, the accuracy of the trained network in classifying the digit patterns drops to only 89.3%, and needs more training epochs (∼120). These results prove that the dual-gate electrolyte gated transistor is more tolerant to noise disturbances than the conventional single-gate electrolyte gated transistor, highlighting the superiority of employing the multi-gate electrolyte gated transistor for complex hetero-synaptic function emulation and neuromorphic computing. For practical implementations using such proof-of-concept devices, continued efforts will be needed for further device minimization and high-density integration. Fig. 4 (a) Illustration of a hetero-synapse performing spatiotemporal signal integration and memorization. Noise signals are introduced during the transmission of the original signals. (b) Examples of two noisy images separately delivered to the dual-gate hetero-synaptic transistor for processing. (c) The image stored in the dual-gate hetero-synaptic transistor after processing. (d) Illustration of a neural network based on the hetero-synaptic transistor for MNIST image recognition. (e) Comparison of the classification accuracy for the networks based on the dual-gate hetero-synaptic transistor and the single-gate homo-synaptic transistor."
} | 5,080 |
35475552 | PMC9545801 | pmc | 1,268 | {
"abstract": "Abstract The global impacts of climate change are evident in every marine ecosystem. On coral reefs, mass coral bleaching and mortality have emerged as ubiquitous responses to ocean warming, yet one of the greatest challenges of this epiphenomenon is linking information across scientific disciplines and spatial and temporal scales. Here we review some of the seminal and recent coral‐bleaching discoveries from an ecological, physiological, and molecular perspective. We also evaluate which data and processes can improve predictive models and provide a conceptual framework that integrates measurements across biological scales. Taking an integrative approach across biological and spatial scales, using for example hierarchical models to estimate major coral‐reef processes, will not only rapidly advance coral‐reef science but will also provide necessary information to guide decision‐making and conservation efforts. To conserve reefs, we encourage implementing mesoscale sanctuaries (thousands of km 2 ) that transcend national boundaries. Such networks of protected reefs will provide reef connectivity, through larval dispersal that transverse thermal environments, and genotypic repositories that may become essential units of selection for environmentally diverse locations. Together, multinational networks may be the best chance corals have to persist through climate change, while humanity struggles to reduce emissions of greenhouse gases to net zero.",
"conclusion": "8 CONCLUDING REMARKS Climate change is increasing the frequency and intensity of coral‐bleaching events and is changing the composition, architectural complexity, and functioning of coral reefs. Under this reality, the future of coral reefs may appear grim. Nonetheless, and despite global declines, it seems that many coral reefs still host enough genetic diversity for adaptation and for perhaps recovery in some form. The best way to support the resilience, adaptation, and recovery of coral reefs is to urgently reduce global emissions of greenhouse gases while working cooperatively to create both local and mesoscale coral‐reef sanctuaries. It is imperative to know which coral species and which reefs to prioritize for protection, based on their adaptive potential and innate resilience. Taking a broad transdisciplinary approach to investigate coral bleaching will improve predictive models, help mitigate the risks, and bolster management and conservation efforts to preserve coral reefs through climate change. Alongside the urgent global need to reduce emissions of greenhouse gases, all possible local and multinational actions should be made to conserve coral reefs—one of the most wondrous ecosystems on the planet—into the future.",
"introduction": "1 INTRODUCTION The relationship between scleractinian corals and their photosynthetic microalgal symbionts has allowed corals to build coral reefs for millions of years. In recent decades however thermal‐stress events have increased in frequency and intensity resulting in widespread coral bleaching (Glynn, 1996 ; Heron et al., 2016 ; Figure 1 ). Coral bleaching represents the breakdown of a long co‐evolutionary relationship between the coral host and its photosynthetic symbionts (Coles & Jokiel, 1977 ; Gates et al., 1992 ; Goreau, 1964 ; Hoegh‐Guldberg, 1999 ; LaJeunesse et al., 2018 ; Rädecker et al., 2021 ; Weis, 2008 ). This breakdown leads to the visual whitening of corals through the loss of intracellular microalgal symbionts (Box 1 ), which can result in coral mortality and changes in reef communities over large regions (Hughes et al., 2018 ; McClanahan et al., 2020 ; Stuart‐Smith et al., 2018 ). Such changes reduce the goods and services that reefs provide, including their capacity to keep up with sea‐level rise (Perry et al., 2013 ; van Woesik & Cacciapaglia, 2021 ), and thereby protect coastal communities from storm waves (Ferrario et al., 2014 ). Yet, our understanding of coral bleaching resulting from thermal stress and its cascading consequences on coral reefs is incomplete. We are just beginning to understand the role of molecular, genetic, and phenotypic traits in determining which individuals, species, and populations of corals are likely to survive. FIGURE 1 Global coral bleaching from 1980 to 2020. Coral bleaching was calculated as a percentage of the coral colonies that were bleached at the time of survey, from 11,068 sites in 89 countries ( n = 23,298; data from van Woesik & Kratochwill, 2022 ) Although numerous studies collect coral‐bleaching data (Box 2 ), these studies are rarely integrated across biological levels of organization. Still, understanding variation in thermal tolerance among individuals at the molecular and physiological levels is essential for elucidating a population's vulnerability at the ecological level. Similarly, variable responses to thermal stress may provide insight into physiological and molecular mechanisms and highlight adaptive potential at different geographic scales. Therefore, there is a need for a conceptual framework that connects environmental conditions to coral‐bleaching responses across biological, spatial, and temporal scales. Such a framework should integrate individual‐based molecular and physiological responses with coral populations and communities. Here, we explore potential links across biological disciplines that may increase our understanding of coral‐bleaching responses and their impact on coral‐reef ecosystems. Elucidating coral‐bleaching responses is critical in determining which habitats and oceanic regions might serve as climate‐change refugia and how we could manage them. Building on past reviews on coral bleaching (Glynn, 1996 ; Brown, 1997 ; Baker et al., 2008 ; Suggett & Smith, 2020 ), we update the relevant information and synthesize recent coral‐bleaching discoveries from an ecological, physiological, and molecular perspective. This synthesis aims towards providing a conceptual framework that integrates coral‐bleaching responses across biological scales in a way that sheds light on how best to identify climate‐change refugia and possible management strategies. We also recommend which data, metadata, and processes are critical for: (i) improving predictive models of coral resilience (Box 2 ) and (ii) refining the capacity to detect inter‐connected networks of reefs for the establishment of mesoscale sanctuaries (thousands of km 2 ) in which high levels of coral genetic diversity, phenotypic adaptation, thermal tolerance, and resilience are most likely. Incorporating models of coral resilience and refugia into sanctuary planning will help to improve the ecosystem‐level resilience of local marine reserves and expand multinational mesoscale sanctuaries. Multinational mesoscale sanctuaries have the potential to simultaneously protect coral reefs from local and regional scale stressors, bridge territorial boundaries, and give corals and coral reefs a fighting chance of coping with climate change because they can preserve standing genetic diversity while maintaining connectivity between reefs through larval dispersal across contrasting environmental and thermal gradients. BOX 1 The coral bleaching phenomenon The term ‘bleaching’ was coined to describe the visible paling of the coral surface, as the white skeleton becomes visible through the animal's translucent tissue that has lost pigmentation and symbionts (Glynn, 1983 ; Goreau, 1964 ). Coral bleaching is synonymous with the breakdown of the symbiotic relationship between the coral host and its microalgal symbionts (family Symbiodiniaceae). Yet, bleaching does not imply a physiological pathology per se because bleaching can result from multiple biotic and abiotic stressors, including freshwater, disease, pollutants, UV radiation, and suboptimal seawater temperatures (Glynn, 1996 ; Goreau, 1964 ). Most recent coral‐bleaching events are caused by anomalously high seawater temperatures (Glynn, 1996 ; Hoegh‐Guldberg, 1999 ; Hughes et al., 2018 ; Sully et al., 2019 ) resulting from climate change. Under extreme temperature stress, coral tissue deterioration or detachment from the skeleton can occur, often leading to mortality (Gates et al., 1992 ; Leggat et al., 2019 ). In many cases, however, the coral tissue remains intact during bleaching, albeit severely compromised and deprived of a nutrition source due to the loss of symbiotic microalgae that provide energy from photosynthesis to the coral host. Without sufficient heterotrophic compensation (Grottoli et al., 2006 ; Levas et al., 2016 ) or high enough energy reserves (Anthony et al., 2009 ), the bleached coral may eventually die from starvation. Processes involved in the establishment and maintenance of the coral‐microalgal symbiosis provide insights into bleaching mechanisms. On establishment of symbiosis, the microalgal cells are taken into the coral gastrodermis, coral immune responses are repressed (Voolstra et al., 2009 ), and the cells are incorporated into the symbiosome (Davy et al., 2012 ). There the microalgal symbionts are maintained in a vegetative, immobile, nutrient‐limited state that stimulates the release of excess photosynthetic carbon for the coral host to harvest (Barott et al., 2015 ; Jokiel et al., 1994 ). Recognition and signaling between both partners are crucial for maintaining nutritional demands (Rädecker et al., 2021 ). Under environmental stress, translocation of microalgal photosynthates to the coral host slows (Hughes et al., 2010 ), reducing the coral's primary source of organic carbon, signaling starvation and amino acid digestion (Rädecker et al., 2021 ). The onset of physiological stress and bleaching is strongly dependent on the rate of heating, the accumulated thermal stress, and the maximum temperature (Middlebrook et al., 2010 ; Savary et al., 2021 ; Voolstra et al., 2020 ). Thermal stress affects multiple processes in both partners, resulting in direct impairment of key cellular functions, homeostasis, and nutrition (Rädecker et al., 2021 ; Roach et al., 2021a ). In the microalgal symbiont, several factors have been implicated as pressure points during thermal stress, especially when combined with high irradiance. These factors include photosystem II repair, thylakoid membrane stability, and both photosynthetic and heterotrophic carbon assimilation pathways (Hughes et al., 2010 ; Iglesias‐Prieto et al., 1992 ; Tchernov et al., 2004 ; Warner et al., 1999 ). Disrupted photosynthetic and mitochondrial electron flow leads to elevated reactive‐oxygen and nitrogen species. These disruptions alter redox homeostasis, create oxidative stress (Brown et al., 2002 ; Krueger et al., 2014 ; Lesser, 1997 ), and trigger a coral‐innate‐immune response. Transcriptomic analyses show that oxidative stress disrupts calcium (Ca 2+ ) homeostasis, which leads to altered cytoskeletal and cell‐adhesion, reduced calcification, and expression of stress‐response genes (DeSalvo et al., 2008 ; Rodriguez‐Lanetty et al., 2009 )—however, gene‐network analyses indicate that additional, but less understood, mechanisms are also involved in coral bleaching (Dixon et al., 2020 ; Rose et al., 2016 ). Persistent disruption to cellular homeostasis under thermal stress can cause both the coral and the microalgal symbiont to undergo necrotic and apoptotic cell death (Dunn et al., 2012 ; Lesser & Farrell, 2004 ) and can lead to fatal coral bleaching."
} | 2,863 |
23889801 | PMC3750612 | pmc | 1,269 | {
"abstract": "Background Contemporary coral reef research has firmly established that a genomic approach is urgently needed to better understand the effects of anthropogenic environmental stress and global climate change on coral holobiont interactions. Here we present KEGG orthology-based annotation of the complete genome sequence of the scleractinian coral Acropora digitifera and provide the first comprehensive view of the genome of a reef-building coral by applying advanced bioinformatics. Description Sequences from the KEGG database of protein function were used to construct hidden Markov models. These models were used to search the predicted proteome of A. digitifera to establish complete genomic annotation. The annotated dataset is published in ZoophyteBase, an open access format with different options for searching the data. A particularly useful feature is the ability to use a Google-like search engine that links query words to protein attributes. We present features of the annotation that underpin the molecular structure of key processes of coral physiology that include (1) regulatory proteins of symbiosis, (2) planula and early developmental proteins, (3) neural messengers, receptors and sensory proteins, (4) calcification and Ca 2+ -signalling proteins, (5) plant-derived proteins, (6) proteins of nitrogen metabolism, (7) DNA repair proteins, (8) stress response proteins, (9) antioxidant and redox-protective proteins, (10) proteins of cellular apoptosis, (11) microbial symbioses and pathogenicity proteins, (12) proteins of viral pathogenicity, (13) toxins and venom, (14) proteins of the chemical defensome and (15) coral epigenetics. Conclusions We advocate that providing annotation in an open-access searchable database available to the public domain will give an unprecedented foundation to interrogate the fundamental molecular structure and interactions of coral symbiosis and allow critical questions to be addressed at the genomic level based on combined aspects of evolutionary, developmental, metabolic, and environmental perspectives.",
"conclusion": "Conclusions We offer ZoophyteBase as an unprecedented foundation to interrogate the molecular structure of the predicted A. digitifera proteome. Some key findings include proteins with relevance to host-symbiont function, dysfunction and recovery including those that direct vacuolar trafficking and proteins linking symbiont photosynthesis to coral calcification. An extensive catalogue of mammalian-like proteins essential to neural function and venoms related to distant animal phyla suggests their origins lie deep in early eumetazoan evolution. Homologues of prokaryotic genes that have not been described previously in any eukaryote genome such as flagella proteins, proteins essential for nitrogen fixation and photosynthesis point towards lateral gene transfer, perhaps mediated by viruses, that may lead to “shared” metabolic adaptations of symbiosis, and provide corals with limited ability for gene-encoded adaptation to a changing global environment. It is anticipated that understanding how the genome of a coral hosts interacts with that of its vast array of symbionts, and how it may regulate its metabolic quotient, for example through biochemical or epigenetic modification, will rapidly accelerate our ability to predict the fate of coral reefs.",
"discussion": "Discussion Regulatory proteins of symbiosis Metabolic cooperation is a key feature of coral-algal symbiosis that allows reef-building corals to inhabit the often nutrient-poor waters of tropical oceans [ 54 ]. In this phototropic symbiosis, fixed carbon produced by resident algae is released to the host for nutrition, and the algal symbionts benefit by acquiring the inorganic nutrient wastes of host metabolism [ 2 , 55 ]. The symbiotic dinoflagellates reside and proliferate within a specialised phagosome (the symbiosome) maintained within host gastrodermal cells. This arrangement requires complex biochemical coordination by the coral at various metabolic stages that includes endocytosis (phagocytosis) by post-settlement polyps to acquire algal symbionts, accord symbiosome recognition to arrest phagosomal maturation for sustained organelle homeostasis, activate symbiophagy or exocytosis to eliminate damaged symbionts [ 56 , 57 ], and regulate apoptotic or exocytotic pathways to remove excess or impaired populations, all of which have long been recognised as essential to preserve the stability of coral symbiosis [ 58 ]. Although these processes are poorly understood in corals, it has been realised from studies of the sea anemone Aiptasia pulchella , a related anthozoan also containing Symbiodinium sp. endosymbionts, that the persistence of algal-containing symbiosomes in Cnidaria relies on the exclusion or retention of small Rab GTPase family proteins that are key regulatory components of vesicular trafficking and membrane fusion in eukaryotic cells [ 59 ]. Significantly, ApRab3 and ApRab4 accumulate in the biogenesis of maturing symbiosomes of A. pulchella [ 60 , 61 ], and mature symbiosomes enveloping healthy dinoflagellates have tethered ApRab5 [ 62 ], a checkpoint antagonist of downstream ApRab7 and ApRab11 proteins that would otherwise direct autophagy of the symbiont cargo [ 63 , 64 ]. Our annotation of the A. digitifera genome reveals sequences encoding putative Rab homologues of the Ras superfamily of proteins (Table 1 ). In a comparison of cnidarian Rab proteins, eight proteins of A. digitifera matched homologues of Aiptasia pulchella , twenty-nine matched proteins encoded by the aposymbiotic freshwater H. magnipapillata and the aposymbiotic anemone N. vectensis genomes, while seven Rab and Rab-interacting proteins of A. digitifera did not match other cnidarian proteins (Table 2 ). Significantly, the eight homologues of A. digitifera that matched exclusively Rab proteins of A. pulchella included homologues of the aforementioned ApRab3, ApRab4 and ApRab5 proteins attributed to the maintenance of healthy symbiosomes in Aiptasia , while homologues of the autophagic ApRab7 and ApRab11 proteins are found also in N. vectensis . While Rab GTPase proteins and their effector proteins coordinate consecutive stages of endocytic vesicular transport [ 65 , 66 ], soluble N-ethylmaleimide-sensitive factor attachment receptor (SNARE) proteins are essential for Rab assembly to complete endosomal fusion of vesicle membranes [ 67 ], a process by which Rab proteins impart specificity by binding distinct Rab and SNARE partner proteins prior to membrane fusion [ 68 ]. Genes encoding syntaxin-like SNARE proteins have been unambiguously identified [ 69 ] from coral EST database libraries constructed from expressed mRNA isolated from various early life stages of Acropora aspera , A. millepora , A. palmata and Orbicella faveolata (= Monastraea faveolata ), as well as from the genome of the sea anemone N. vectensis [ 70 ]. In metazoans, vacuolar r-SNARE receptor proteins comprise the syntaxin, synaptobrevin and VAMP family proteins, of which there are eight syntaxin and syntaxin-binding proteins (plus two plant-like syntaxins). Additionally, there are one t-SNARE target protein to direct vacuolar morphogenesis, two synaptosomal proteins, one synaptosomal complex ZIP1 protein (yeast homologue), one synaptobrevin membrane protein of secretory vesicles, ten vesicle-associated membrane proteins (VAMPs), a vacuolar protein-8 regulator of autophagy, four vacuolar-sorting proteins and two SEC22 vesicle trafficking protein encoded in the genome of A. digitifera (Table 1 ), many of which may interact to provide metabolic transport between the endoplasmic reticulum and Golgi apparatus [ 71 ]. Included in this vast but yet unexplored repertoire of vacuolar-acting proteins are the syntaxin-binding amisyn and tomosyn regulators of SNARE complex assembly and disassembly [ 72 , 73 ], which may control membrane fusion in the phagocytic establishment and dis-sociation of coral symbiosis. Table 1 Regulatory proteins of symbiosis in the predicted proteome of A. digitifera Gene sequence KEGG Orthology Encoded protein description v1.06849 K06110 Exocyst complex component 3 v1.00063; v1.01826 K06111 Exocyst complex component 4 v1.06336; v1.06337; v1.15354 K07195 Exocyst complex component 7 v1.04340 [+ 4 other sequence copies] K14966 Host cell factor v1.01629; v1.19166 K12481 Rabenosyn-5 v1.18447 [+ 26 other sequence copies] K07976 Rab family, other (similar to Rab-6B) v1.02380 K12480 Rab GTPase-binding effector protein-1 v1.01032 K13883 Rab-interacting lysosomal protein v1.14682; v1.03256; v1.07709 K12484 Rab11 family-interacting protein-1/2/5 v1.13055; v1.13176; v1.16348 K12485 Rab11 family-interacting protein-3/4 v1.01275 K07932 Rab-like protein-2B v1.17629 [+ 13 other sequence copies] K07933 Rab-like protein-3 v1.03299; v1.09653 K07934 Rab-like protein-4 v1.08498 K07935 Rab-like protein-5 v1.16155 [+5 other sequence copies K07874 Ras-related protein Rab-1A v1.09098 K07875 Ras-related protein Rab-1B v1.13558; v1.08983 K07877 Ras-related protein Rab-2A v1.14260 K07878 Ras-related protein Rab-2B v1.07500; v1.20532; v1.07498 K07884 Ras-related protein Rab-3D v1.21242; v1.07502 K07880 Ras-related protein Rab-4B v1.01341; v1.05619 K07888 Ras-related protein Rab-5B v1.07125 K07889 Ras-related protein Rab-5C v1.09239 K07893 Ras-related protein Rab-6A v1.10443; v1.13335 K07897 Ras-related protein Rab-7A v1.03086; v1.17122; v1.07231 K07916 Ras-related protein Rab-7 L1 v1.02275 [+ 4 other sequence copies] K07901 Ras-related protein Rab-8A v1.24612 K07899 Ras-related protein Rab-9A v1.00411 K07900 Ras-related protein Rab-9B v1.10697; v1.01515 K07903 Ras-related protein Rab-10 v1.22278; v1.04408; v1.12528 K07905 Ras-related protein Rab-11B v1.07033; v1.23028 K07881 Ras-related protein Rab-14 v1.02275 K07908 Ras-related protein Rab-15 v1.16455; v1.14911; v1.14959 K07910 Ras-related protein Rab-18 v1.04714 K07911 Ras-related protein Rab-20 v1.01878; v1.12184 K07890 Ras-related protein Rab-21 v1.09930 K06234 Ras-related protein Rab-23 v1.13579; v1.12841 K07912 Ras-related protein Rab-24 v1.10183 K07913 Ras-related protein Rab-26 v1.08199 K07885 Ras-related protein Rab-27A v1.13978; v1.18893 K07917 Ras-related protein Rab-30 v1.03085; v1.06007; v1.07729 K07918 Ras-related protein Rab-32 v1.24721 K07919 Ras-related protein Rab-33A v1.18892 K07920 Ras-related protein Rab-33B v1.16060 K07876 Ras-related protein Rab-35 v1.15894 K07922 Ras-related protein Rab-36 v1.03080 K07923 Ras-related protein Rab-38 v1.21391 K07924 Ras-related protein Rab-39A v1.14786 K07928 Ras-related protein Rab-40 v1.05611 [+ 13 other sequence copies] K08502 Regulator of vacuolar morphogenesis (t-SNARE domain) v1.18253 K08520 SEC22 vesicle trafficking protein A/C v1.15499 K13814 t-SNARE domain-containing protein 1 v1.05749 K08516 Synaptobrevin homologue YKT6 v1.13229 K12768 Synaptonemal complex protein ZIP1 v1.16533; v1.17141 K08508 Synaptosomal-associated protein, 23 kDa v1.05301 K08509 Synaptosomal-associated protein, 29 kDa v1.19071 K04560 Syntaxin 1A v1.04614; v1.22747 K08486 Syntaxin 1B/2/3 v1.16462 K08490 Syntaxin 5 v1.20758; v1.21534 K08498 Syntaxin 6 v1.22836; v1.15499 K08488 Syntaxin 7 v1.01959; v1.24227 K08501 Syntaxin 8 v1.02007; v1.06683; v1.12727 K08491 Syntaxin 17 v1.21308; v1.11830; v1.01582 K08492 Syntaxin 18 v1.22100; v1.09457 K08518 Syntaxin binding protein 5 (tomosyn) v1.18555 K08519 Syntaxin binding protein 6 (amisyn) v1.12938 K08500 Syntaxin of plants SYP6 v1.06575 K08506 Syntaxin of plants SYP7 v1.14699 K08507 Unconventional SNARE in the endoplasmic reticulum protein 1 v1.23782 [+ 38 other sequence copies] K08332 Vacuolar protein 8 v1.15282; v1.24603; v1.01672 K12196 Vacuolar protein-sorting-associated protein 4 v1.17791 [+ 4 other sequence copies] K12479 Vacuolar protein sorting-associated protein 45 v1.20907 K11664 Vacuolar protein sorting-associated protein 72 v1.15996 [+ 5 other sequence copies] K12199 Vacuolar protein sorting-associated protein VTA1 v1.15614 K08510 Vesicle-associated membrane protein 1 (synaptobrevin) v1.13353 K13504 Vesicle-associated membrane protein 2 (synaptobrevin) v1.12458; v1.07528 K13505 Vesicle-associated membrane protein 3 (cellubrevin) v1.19735; v1.21831; v1.07186 K08513 Vesicle-associated membrane protein 4 (Golgi transport) v1.05299 K08514 Vesicle-associated membrane protein 5 (exocytosis) v1.13557; v1.24610 K08515 Vesicle-associated membrane protein 7 (exocytosis) v1.12279 K08512 Vesicle-associated membrane protein 8 (endobrevin) v1.00261; v1.08699; v1.04334 K06096 Vesicle-associated membrane protein A v1.20177 K10707 Vesicle-associated membrane protein B v1.15472; v1.03568 K06027 Vesicle-fusing ATPase v1.11431; v1.10487 K08517 Vesicle transport protein SEC22 v1.06393; v1.13003; v1.08735; v1.04261 K08493 Vesicle transport interaction with t-SNAREs 1 Table 2 Distribution of Rab homologues of Aiptasia puchella , Hydra magnipapillata and Nematostella vectensis in the predicted proteome of A. digitifera A. digitifera Rab protein Cnidarian encoding Rab homologue Rab-like protein- 2B, Rab-2B Rab-3D, Rab-4B, Rab-5B, Rab-26, Rab-32, Rab-38 A. puchella Rab-like protein-3, Rab-36 N. vectensis Rab-2A, Rab-23 A. puchella , H. magnipapillata Rab-like protein-6B, Rab-6A, Rab-7 L1, Rab-10, Rab11B, Rab-30, Rab-33B A. puchella , N. vectensis Rab effector protein-1, Rab11-interacting protein-3/4 H. magnipapillata , N. vectensis Rab-like protein-4, Rab-like protein-5, Rab-1A, Rab5C, Rab-7A, Rab-8A, Rab-9A, Rab-14, Rab-18, Rab-20, Rab-21, Rab-24, Rab-27A, Rab-35 A. puchella , H. magnipapillata, N. vectensis Rab-interacting lysomal protein, Rab11-interacting protein-1/2/5, Rab-1B, Rab-9B, Rab-3A, Rab-39A, Rab-40 No match In the final step of exocytosis there is a cytosolic influx of calcium which binds to synaptotagmin to actuate completion of membrane SNARE protein assembly with exocytic docking to form the conducting channel for trans-membrane vesicular transport on activation by vesicle-fusing ATPase [ 74 ]. As synaptotagmin proteins are not included in the KEGG database, Zoophytebase was used for BLAST searches with all known synaptotagamin sequences [ 27 ]. Synaptotagamin proteins from A. digitifera were found having similarity to homologues from diverse invertebrate and vertebrate organisms, including one from the human genome (Table 3 ). Other Ca 2+ -sensing proteins of A. digitifera , such as calmodulin and the calcium binding protein CML, are given with calcification and Ca 2+ -signalling proteins. Table 3 Synaptotagmin proteins in the predicted proteome of A. digitifera Gene sequence GenBank Accession Genome encoded homologue v1.08623 GI:268530614 Caenorhabditis briggsae : XP_002630433 (worm) v1.20682; v1.10560; v1.02080; v1.10015 GI:150416761 Platynereis dumerilii : ABR68850 (worm) v1.10269; v1.04412 GI:288869516 Nasonia vitripennis : NP_001165865 (wasp) v1.01508 GI:29378331 Lymnaea stagnalis : AA093847 (snail) v1.18613 GI:391339919 Metaseiulus occidentalis : XP_003744294 (mite) v1.07402 GI:260834895 Branchiostoma floridae : XP_002612445 (lancelet) v1.01542 GI:149067023 Rattus norvegicus : EDM16756 (rat) v1.20683 GI:383860584 Megachile rotundata : XP_003705769 (bee) v1.17688 GI:48529130 Oreochromis niloticus ; XP_003452067 (fish) v1.15777; v1.14902 GI:269785031 Saccoglossus kowalevskii: NP_001161667 (worm) v1.17175; v1.11521 GI:11559313 Halocynthia roretzi : BAB18864 (ascidian) v.1.03344; v1.03345 GI:12658419 Manduca sexta ; AF331039 (moth) v1.16152 GI:395729192 Pongo abelii : XP_003780414 (orangutan) v1.10268 GI:327283049 Anolis carolinensis : XP_003226254 (lizard) v10.2778 GI:125984480 Drosophila pseudoobscura XP_001356004.1 (fly) v1.02083; v1.02777 GI:226490194 Schistosoma japonicum : CAX69339.1 (fluke) v1.04326 GI:167744962 Homo sapiens : 2R83_A (human) v1.14682; v1.04180 GI:241704658 Ixodes scapularis : XP_002411967 (tick) Intriguingly, annotation of the A. digitifera genome reveals a host cell factor (K14966), but this is not related to the elusive “host factor” of symbiosis demonstrated to be present in tissue homogenates of corals and other marine invertebrates that harbor Symbiodinium spp. endosymbionts [ 75 - 77 ]. Instead, this mammalian transcriptional coactivator host cell factor (HFC-1) is known to mediate the enhancer-promoter assemblies of herpes simplex (HSV) and varicella zoster (VZV) viruses for activation of the latent state for replication [ 78 ], such that the coral HCF homologue may have similar relevance as a viral checkpoint transcriptional coactivator of virulence in A. digitifera . HCF-1 expression is coupled also to chromatin modification [ 79 , 80 ] suggesting that the coral protein homologue may have an additional role in epigenetic reprogramming of the chromatin histone-DNA complex at different stages of development. Planula and early developmental proteins In this section we discuss predicted proteins encoded in the A. digitifera genome having functional homology to known proteins are specific to early embryonic development, planula larvae function and morphogenesis, which are given in Table 4 . Annotation of the coral genome reveals a large set of homeobox proteins involved in the regulation of anatomical development during morphogenesis. The homeobox is a highly conserved DNA sequence (homeodomain) within genes that binds to DNA in a sequence-specific manner [ 81 ] often at the promoter region of their target gene to affect transcription in the developing embryo. Amonst these transcriptional regulators, Hox genes are essential to metazoan development as their expressed proteins differentiate embryonic regions along the anterior-posterior axis (the Hox code) and are recognised for their contribution to the evolution of morphological diversity [ 82 ]. Hox genes are well characterised in cnidarians and, given their importance in embryonic development, it is not surprising that molecular evidence from the Cnidaria reveal that the genetic origins of Hox genes predate the cnidarian-bilaterian divergence [ 83 - 85 ] yet had evolved after divergence of the sponge and eumetazoan lineages [ 86 ]. Hox genes of cnidarians are typically located in a conserved genomic collinear cluster, which is apparent also for A. digitifera , whereby the order of the genes on the chromosome is the same as that of gene expression in the developing embryo. Included in our annotation are genes encoding two LIM homeobox proteins and a LIM homeobox transcription factor (Lhx) having conserved roles in neuronal development [ 87 ], which in N. vectensis are responsible for the development of neural networks in developing larvae and juvenile polyps [ 88 ]. Unlike N. vectensis [ 89 ], the coral genome expresses a homeobox BarH-like protein that in vertebrates directs neurogenesis [ 90 ]. Distinct from homeodomain proteins, but serving similar functions, are various protein activators, regulators and receptors of cellular morphogenesis. Annotation of the coral genome has revealed multiple sequence alignments to a protein homologue of the dishevelled-associated activator of morphogenesis 1 (Daam1) that initiates cytoskeleton formation via the control of actin assembly. Daam1 was found crucial for gastrulation in Xenopus [ 91 ], wherein Daam1 mutants of Drosophilia exhibit trachea defects [ 92 ], and in mammals Daam1 is highly expressed in multiple developing organs and is deemed essential for cardiac morphogenesis [ 93 ]. Similar morphogenetic genes express regulatory proteins that are necessary for vacuole biogenesis in yeasts [ 94 ]. Others express bone morphogenetic proteins (and their BMP receptors), which are potent multi-functional growth activators that belong to the transforming growth factor beta (TGFbeta) cytokine superfamily of proteins that in humans have various functions during embryogenesis, skeletal formation, neurogenesis and haematopoiesis [ 95 ]. However, since many of the homeobox and morgenetic proteins (Table 4 ) are homologues of proteins with functions ascribed to higher organisms, their precise function in A. digitifera cannot be ascertained by KEGG orthology alone. Table 4 Planula and early developmental proteins in the predicted proteome of A. digitifera Gene sequence KEGG Orthology Encoded protein description v1.09797; v1.11180; v1.08414 K03776 Aerotaxis receptor (oxygen sensing) v1.07838 [+5 other sequence copies] K07822 Archaeal flagellar protein FlaC v1.14039; v1.11310; v1.11309 K05502 Bone morphogenetic protein 1 v1.01025; v1.17008; v1.15796; v1.23658 K04662 Bone morphogenetic protein 2/4 v1.02299; v1.07696; v1.10675 K04663 Bone morphogenetic protein 5/6/7/8 v1.06335; v1.01763 K04673 Bone morphogenetic protein receptor type-1A v1.13481 K13578 Bone morphogenetic protein receptor type-1B v1.10550 [+4 other sequence copies] K04671 Bone morphogenetic protein receptor type-2 v1.00912 [+4 other sequence copies] K13579 Bone morphogenetic protein receptor type-1, invertebrate v1.19370 K14624 C-C motif chemokine 2 v1.23163 K12499 C-C motif chemokine 5 v1.08576 K05511 C-C motif chemokine 15/23 v1.09229 K05512 C-C motif chemokine 19/21 v1.09305 K08373 C-C chemokine receptor-like 2 v1.04942 K04179 C-C chemokine receptor type 4 v1.02658 K04245 Chemokine-like receptor 1 v1.21300 K12671 C-X-C motif chemokine 10 v1.16396; v1.21991 K10035 C-X-C motif chemokine 16 v1.23712 K11522 Chemotaxis family two-component system response regulator PixG v1.09435 K13490 Chemotaxis family, histidine kinase sensor response regulator (WspE-like) v1.14142; v1.05300 K05874 Chemotaxis protein I, serine sensor receptor (MCP family) v1.07361 K05877 Chemotaxis protein IV, peptide sensor receptor (MCP family) v1.17411 K03414 Chemotaxis protein CheZ v1.16104 K00575 Chemotaxis protein methyltransferase CheR v1.15537 [+ 7 other sequence copies] K08482 Circadian clock protein KaiC v1.14925 [+ 4 other sequence copies] K02223 Circadian locomoter output cycles kaput protein v1.06432 [+ 9 other sequence copies] K04512 Dishevelled associated activator of morphogenesis v1.17637 [+ 70 other sequence copies] K10408 Dynein heavy chain, axonemal v1.00202 [+5 other sequence copies] K10409 Dynein intermediate chain 1, axonemal v1.04986; v1.09649; v1.23645 K11143 Dynein intermediate chain 2, axonemal v1.08695; v1.09481; v1.23153 K10411 Dynein light chain 1, axonemal v1.11684 K10412 Dynein light chain 4, axonemal v1.23322; v1.01131; v1.04207 K10410 Dynein light intermediate chain, axonemal v1.14083 K02401 Flagellar biosynthetic protein FlhB v1.16997 K02420 Flagellar biosynthetic protein FliQ v1.02867 K02396 Flagellar hook-associated protein 1 FlgK v1.18101; v1.13427 K02408 Flagellar hook-basal body complex protein FliE v1.04339; v1.07633 K06603 Flagellar protein FlaG v1.17895[+5 other sequence copies] K02383 Flagellar protein FlbB v1.21111 K02413 Flagellar protein FliJ v1.17651 [+ 13 other sequence copies] K02415 Flagellar protein FliL v1.01971 [+ 6 other sequence copies] K02418 Flagellar protein FliO/FliZ v1.14031 K02423 Flagellar protein FliT v1.08025 K02394 Flagellar P-ring protein precursor FlgI v1.02396; v1.15777 K02409 Flagellar M-ring protein FliF v1.20693 K09451 Homeobox protein aristaless-like 4 v1.24732 [+5 other sequence copies] K09452 Homeobox protein aristaless-related v1.15788; v1.19334; v1.04164 K09313 Homeobox protein cut-like v1.01801 K09319 Homeobox protein engrailed v1.16835; v1.06323 K09320 Homeobox even-skipped homologue protein v1.0412; v1.054771 K09354 Homeobox protein expressed in ES cells 1 v1.13604 K09324 Homeobox protein goosecoid v1.06346; v1.08163 K09325 Homeobox protein goosecoid-like v1.17295; v1.17294 K09361 Homeobox protein, BarH-like (vertebrate neurogenesis) v1.07457 K09316 Homeobox protein DLX, invertebrate v1.11157; v1.08573; v1.15250 K09317 Homeobox protein EMX v1.01800 K09321 Homeobox protein GBX v1.10929; v1.06346; v1.05443; v1.07458 K09310 Homeobox protein GSH v1.13684; v1.24444 K08025 Homeobox protein HB9 v1.16254; v1.16064 K08024 Homeobox protein HEX v1.07458; v1.06706; v1.06705 K09339 Homeobox protein HLX1 v1.06347; v1.06348; v1.17294 K09302 Homeobox protein HoxA/B2 v1.06125 K09306 Homeobox protein HoxA/B/C6 v1.19818 K09304 Homeobox protein HoxA/B/C/D4 v1.06706 K09301 Homeobox protein HoxA/B/D1 v1.02056 K09353 Homeobox protein LBX v1.06347; v1.06348 K09328 Homeobox protein Unc-4 v1.24342; v1.04552 K09318 Homeobox protein ventral anterior v1.03823; v1.10070; v1.04435 K09309 Homeobox protein Nkx-1 v1.12852 [+ 4 other sequence copies] K08029 Homeobox protein Nkx-2.2 v1.21630 K09345 Homeobox protein Nkx-2.5 v1.10625 K09347 Homeobox protein Nkx-2.8 v1.10625; v1.13865; v1.05476 K09348 Homeobox protein Nkx-3.1 v1.21628; v1.05475; v1.05477 K09995 Homeobox protein Nkx-3.2 v1.06135; v1.10071 K09349 Homeobox protein Nkx-5 v1.14702 K08030 Homeobox protein Nkx-6.1 v1.14917; v1.11907 K09350 Homeobox protein Nkx-6.2 v1.00777; v1.21453 K09322 Homeobox protein MOX v1.00602 [+ 6 other sequence copies] K09326 Homeobox protein OTX v1.16722; v1.12785 K09374 LIM homeobox protein 3/4 v1.11281; v1.05135 K09375 LIM homeobox protein 6/8 v1.07988; v1.22037 K09371 LIM homeobox transcription factor 1 v1.09328 [+ 5 other sequence copies] K10394 Kinesin family member 3/17 v1.09196; v1.12479 K11525 Methyl-accepting chemotaxis protein PixJ (MCP family) v1.17028; v1.13473 K08473 Nematode chemoreceptor v1.13159; v1.00655 K09330 Paired mesoderm homeobox protein 2 v1.15178; v1.10962; v1.16587; v1.01557 K02633 Period circadian protein v1.23288; v1.13857 K04627 Pheromone a factor receptor v1.22464; v1.17135 K11213 Pheromone alpha factor receptor v1.05611 [+ 13 other sequence copies] K08502 Regulator of vacuolar morphogenesis v1.04431 K09333 Retina and anterior neural fold homeobox-like protein v1.17636 K09331 Short stature homeobox protein v1.14704 K09340 T-cell leukemia homeobox protein v1.11765 NA 1 Tektin v1.04154 K02669 Twitching motility protein PilT 1 NA KEGG orthology designation not assigned. Another protein encoded in the A. digitifera genome is a retina and anterior neural fold homeobox-like (RAX) protein that may activate the development of primitive coral photoreceptors [ 96 , 97 ], including a blue light-sensing, cryptochrome photoreceptor that in A. millepora is implicated in the detection of light from the lunar cycle of night time illumination to signal synchronous coral spawning [ 98 , 99 ]. Photosensitive behaviours and the circadian rhythms of corals are well described, and diurnal cycles of gene transcription that regulate circadian biological processes in the coral A. millepora have been reported [ 100 ]. Such traits in A. millepora appear regulated by an endogenous biological clock entrained to daily cycles of solar illumination [ 101 ]. Annotation of the A. digitifera genome reveals a circadian timekeeper protein KaiC [ 102 ] that in cyanobacteria is activated during the diurnal phosphorylation rhythm [ 103 , 104 ]. In Synechococcus elongatus , KaiC regulates the rhythmic expression of all other proteins encoded in the genome [ 105 ], yet no homologue of any of the prokaryotic clustered circadian kiaABC genes has been identified in eukaryotes [ 106 ]. In Drosophila , KaiC together with a homologue of the eukaryotic period (Per) circadian protein drives circadian rhythms in eclosion (hatching) and locomotor activity [ 107 ]. Nevertheless, a circadian locomotor output cycles kaput (CLOCK) homologue (Table 4 ) was found in our annotation. Since CLOCK proteins serve as an essential activator of downstream elements in pathways critical to the regulation of circadian rhythms in eukaryotes [ 108 ], it would be worthy to examine how transcription of the RAX-like homeobox protein in this coral contributes to the development of circadian functions by activation of kaiC , per and Clock genes. Such a study might reveal that components of the animal circadian clock are more ancient than data previously suggested [ 109 ]. Broadcast-spawning corals, such as A. digitifera , release gametes, and the fertilised eggs develop into planula larvae within the water column until they have reached settlement competency, find a suitable hard substrate, attach and develop into the polyp on metamorphosis. Coral sperm and planula larvae achieve motility using flagella (sperm) or cilia (larvae) as their locomotor organelles. The eukaryotic axonemal proteins of cilia and flagella are composed of a dynein ATPase protein to provide mechanochemical energy transduction together with the principle structural proteins of the ciliary/flagellar microtubules [ 110 ]. The flagellar/ciliary microtubules consist of filaments composed of α- and β-tubulins, microtubule-stabilising tektins and kinesin motor proteins [ 111 - 113 ]. The coral genome encodes members of the dynein axonemal (flagella and cilia) proteins (Table 4 ) and many of the dynein cytoplasmic proteins (not tabulated), the latter being involved in intracellular organelle transport and centrosome assembly. The coral genome encodes α- and β-tubulins and members of the eukaryotic kinesin superfamily proteins (not tabulated). Amongst the many kinesin proteins encoded in the coral genome is the kinesin family member 3/17 protein, which is a direct homologue of the kinesin-II intraflagellar transport protein FLA10 essential for flagella assembly in the alga Chlamydomonas [ 114 ]. The microtubule-stabilising tektin protein, which is required for cilia and flagella assembly [ 113 ], is also encoded in the coral genome [note: there is no KEGG orthology identifier assigned to this protein]. It was a surprise, however, to find a large complement of prokaryotic flagellar proteins encoded in the coral genome consisting of archaeal flagellar (FlaC and FlaG), bacterial filament (FibB, FlliE, FliF, FliJ, FliK, FliO/FliZ, FliQ and FliT) homologue components (Table 4 ). Included also are the prokaryote homologues FlgN and FlbB that regulate transcriptional activation of flagellar assembly [ 115 , 116 ] and FlhB which controls the substrate specificity of the entire prokaryotic flagellar apparatus [ 117 ]. Encoded in the coral genome is a flagella-independent Type IV twitching mobility protein PilT that affords social gliding translocation in many prokaryotic organisms controlled by complex signal transduction systems that include two-component sensor regulators [ 118 ]. It is unlikely that these genes are derived from contamination from bacterial DNA. Such contamination would manifest itself by the random occurrence of bacterial genes from the whole genome including many housekeeping genes. In this case, the genes occur as members of groups with specialised functions, suggesting that multiple horizontal gene transfers between bacteria and the coral genome have occurred [ 119 ]. Their precise function in A. digitifera remains unknown; homologues of these prokaryotic genes have not been described previously in any other eukaryote genome. Linked closely with flagellar/ciliary proteins are the sensory receptors that signal chemoattraction or avoidance to direct cellular motility. The coral genome reveals a variety of genes that encode chemoreceptor and chemotaxis proteins (Table 4 ). The chemoreceptor proteins of A. digitifera include an oxygen-sensing aerotaxis receptor that in bacteria invokes an avoidance response to anoxic micro-environments [ 120 ]. Encoded also are a nematode sensory chemoreceptor homologue [ 121 ], two homologous pheromone factor receptor proteins that in fungi activate a species-specific mating response [ 122 ], three chemotaxis protein sensor receptors belonging to the methyl-accepting chemotaxis family of proteins (MCPs) in bacteria and archaea [ 123 ], and two proteins (CheZ and CheR) and two regulators (PixG and WspE) of the two-component signal transduction (TCST) system for activation of gene expression. In bacteria and archea, as well as some plants, fungi and protozoa [ 124 ], TCST systems mediate many cellular processes that respond to a broad range of environmental stimuli via activation of a specific histidine (or serine) kinase sensor and its cognate response regulator [ 125 ]. There are 77 sequence matches to various elements of the TCST family of proteins in the A digitifera genome (data not tabulated). Included also are genes encoding members of the chemotactic cytokine (chemokine) family of sensory proteins that on secretion directs chemotaxis in nearby responsive cells by stimulating target chemokine receptors; both chemokine and chemokine receptor proteins are encoded in the coral genome. Significantly, sensory chemokines/chemokine receptors are found in all vertebrates, some viruses and some groups of bacteria, but none have been described previously for invertebrates [ 126 ]. Neural messengers, receptors and sensory proteins Corals and other cnidarians are the earliest extant group of organisms to have a primitive nervous system network [ 127 ] thought to be evolved from a eumetazoan ancestor prior to the divergence of Cnidaria and the Bilateria [ 128 , 129 ]. Unlike marine sponges (Porifera) that predate synaptic innovation [ 130 ], cnidarians possess a homogenous nerve net that, although lacking any form of cephalization, accommodates fundamental neurosensory transmission across the nerve net to end in a motoneural junction to coordinate tentacle movement required for feeding and predator avoidance [ 131 ]. The nervous systems of cnidarians consist of both ectodermal sensory cells and their effector cells and endodermal multipolar ganglions capable of neurotransmission [ 132 ]. At the functional level, synaptic transmission in cnidarians relies on fast neurotransmitters (glutamate, GABA, glycine) and slow neurotransmitters (catecholamine, serotonin, neuropeptides) for sensory-signal conduction [ 133 ]. At the ultrastructural level, many cnidarian neurons have multifunctional traits of sensory, neurosecretory and stimulatory attributes [ 134 ]. Significantly, the genome of A. digitifera encodes the expression of a ciliary neurotrophic factor, which is a polypeptide hormone and nerve growth factor that promotes neurotransmitter synthesis, neurite outgrowth and regeneration [ 135 ]. Additionally, the coral genome encodes nerve growth factor and neurotrophic kinase receptors, a survival motor neuron protein, a survival neuron splicing factor, the neural outgrowth protein neurotrimin, and a neurotrophin growth factor attributed to signalling neuron survival, differentiation and growth (Table 5 ). Encoded for neuron regulation and development are several neuron cation-gated channels, a neuronal guanine nucleotide exchange factor, a neurotransmitter Na + symporter, several neurogenic differentiation proteins, a neuronal PAS domain transcription factor for activation of neurogenesis, the axon guidance protein neurophilin-2, a neural crest protein of embryonic neural development, neural ELAV-like transcription proteins of neurogenesis, a Notch protein (79 sequence domain matches) and a neutralized protein subset of the Notch signalling pathway that promotes neuron proliferation in early neurogenic development. Structural elements of the coral nerve net include neurofilament polypeptides and neuronal adhesion proteins. Table 5 Neuronal and sensory proteins in the predicted proteome of A. digitifera Gene sequence KEGG Orthology Encoded protein description v1.01918 [+ 5 other sequence copies] K01049 Acetylcholinesterase v1.18087; v1.14516 K04136 Adrenergic receptor alpha-1B v1.06394 K04137 Adrenergic receptor alpha-1D v1.09628; v1.15688; v1.00966 K04140 Adrenergic receptor alpha-2C v1.19831; v1.20450 K04142 Adrenergic receptor beta-2 v.17293 K00910 beta-Adrenergic-receptor kinase v1.13740 [+ 5 other sequence copies] K04828 Amiloride-sensitive cation channel 1, neuronal (degenerin) v1.23541 [+ 6 other sequence copies] K04829 Amiloride-sensitive cation channel 2, neuronal v1.09323 [+ 4 other sequence copies] K04439 beta-Arrestin v1.07723; v1.22465 K04641 Bacteriorhodopsin v1.08062 K05420 Ciliary neurotrophic factor v1.03288 [+ 5 other sequence copies] K02295 Cryptochrome v1.20011; v1.20036; v1.20084; v1.18607 K04948 Cyclic nucleotide gated channel alpha 1 v1.21470 K04951 Cyclic nucleotide gated channel alpha 4 v1.21783; v1.01466; v1.01466; v1.01466 K05326 Cyclic nucleotide gated channel, invertebrate v1.03645 K05391 Cyclic nucleotide gated channel, other eukaryote v1.21256 K08762 Diazepam-binding inhibitor (GABA receptor, acyl-CoA-binding protein) v1.22156 [+ 6 other sequence copies] K00503 Dopamine beta-monooxygenase v1.21775: v1.15989 K04148 Dopamine D1-like receptor v1.14160; v1.01697 K04144 Dopamine receptor D1 v1.05089; v1.20018 K04145 Dopamine receptor D2 v1.14030; v1.23273 K04146 Dopamine receptor D3 v1.20536 K13088 ELAV-like protein 1 v1.18658 [+ 5 other sequence copies] K13208 ELAV-like protein 2/3/4 v1.05774 [+ 18 other sequence copies] K04313 G protein-coupled receptor 6 v1.00572; v1.18152 K08404 G protein-coupled receptor 17 v1.23842 K04316 G protein-coupled receptor 19 v1.03948 K08411 G protein-coupled receptor 26 v1.09271 K08383 G protein-coupled receptor 34 v1.05595 K04243 G protein-coupled receptor 37 (endothelin receptor type B-like) v1.04019 K08409 G protein-coupled receptor 45 v1.19913; v1.09821; v1.04291 K08450 G protein-coupled receptor 56 v1.05404 K04321 G protein-coupled receptor 63 v1.02179; v1.10397 K08451 G protein-coupled receptor 64 v1.23269 [+ 5 other sequence copies] K08408 G protein-coupled receptor 68 v1.21091 K08421 G protein-coupled receptor 84 v1.11008 K04302 G protein-coupled receptor 85 v1.21884; v1.01951 K08452 G protein-coupled receptor 97 v1.03243 [+ 13 other sequence copies] K08378 G protein-coupled receptor 103 v1.13790; v1.18939 K08453 G protein-coupled receptor 110 v1.09442; v1.14019 K08455 G protein-coupled receptor 112 v1.24009 K08456 G protein-coupled receptor 113 v1.04290 K08459 G protein-coupled receptor 114 v1.06608; v1.24223 K08457 G protein-coupled receptor 115 v1.10800 [+ 6 other sequence copies] K08458 G protein-coupled receptor 116 v1.07662 [+ 6 other sequence copies] K08462 G protein-coupled receptor 125 v1.09663; v1.08981 K08463 G protein-coupled receptor 126 v1.24252 K08464 G protein-coupled receptor 128 v1.02750 [+ 26 other sequence copies] K08465 G protein-coupled receptor 133 v1.05774 [+ 11 other sequence copies] K08466 G protein-coupled receptor 144 v1.05497; v1.13272; v1.01323 K08436 G protein-coupled receptor 152 v1.08653 [+ 5 other sequence copies] K08467 G protein-coupled receptor 157 v1.11807; v1.10392; v1.10394 K08469 G protein-coupled receptor 158 v1.07294; v1.00247 K08439 G protein-coupled receptor 161 v1.05167 K08442 G protein-coupled receptor 176 v1.08677; v1.23465; v1.19865; v1.06986 K12762 G protein-coupled receptor GPR1 v1.13395 K08291 G protein-coupled receptor kinase v1.18529; v1.07599; v1.05558 K12487 G protein-coupled receptor kinase interactor 2 v1.02481 K04619 G protein-coupled receptor family C group 5 member B v1.22242 K04622 G protein-coupled receptor family C group 6 member A v1.08625; v1.13650; v1.13048; v1.18694 K04599 G protein-coupled receptor Mth (Methuselah protein) v1.07465; v1.10540 K08341 GABA(A) receptor-associated protein (autophagy-related protein 8) v1.09831 [+ 30 other sequence copies] K05270 Gamma-aminobutyric acid (GABA) receptor, invertebrate v1.18702; v1.11701 K05183 Gamma-aminobutyric acid (GABA) A receptor beta-3 v1.04252 [+ 6 other sequence copies] K05185 Gamma-aminobutyric acid (GABA) A receptor epsilon v1.06325 K05186 Gamma-aminobutyric acid (GABA) A receptor gamma-1 v1.00048 K05188 Gamma-aminobutyric acid (GABA) A receptor gamma-3 v1.07506 [+ 6 other sequence copies] K04615 Gamma-aminobutyric acid (GABA) B receptor 1 v1.07506 [+ 24 other sequence copies] K04616 Gamma-aminobutyric acid (GABA) B receptor 2 v1.06426; v1.10563; v1.01138 K05192 Gamma-aminobutyric acid (GABA) receptor theta v1.15485 K05198 Glutamate receptor, ionotropic, AMPA 2 v1.09807 K05200 Glutamate receptor, ionotropic, AMPA 4 v1.04764 K05207 Glutamate receptor, ionotropic, delta 2 v1.15247 [+ 12 other sequence copies] K05313 Glutamate receptor, ionotropic, invertebrate v1.15247 [+ 7 other sequence copies] K05202 Glutamate receptor, ionotropic, kainate 2 v1.00617 K05203 Glutamate receptor, ionotropic, kainate 3 v1.09688 [+ 6 other sequence copies] K05208 Glutamate receptor, ionotropic, N-methyl D-aspartate 1 v1.21204 [+ 4 other sequence copies] K05212 Glutamate receptor, ionotropic, N-methyl-D-aspartate 2D v1.01622 K05214 Glutamate receptor, ionotropic, N-methyl-D-aspartate 3B v1.01418 [+ 5 other sequence copies] K05387 Glutamate receptor, ionotropic, other eukaryote v1.04275 K05194 Glycine receptor alpha-2 v1.10737; v1.06885 K05195 Glycine receptor alpha-3 v1.05488 K05271 Glycine receptor alpha-4 v1.08900; v1.06885 K05196 Glycine receptor beta v1.18634 K05397 Glycine receptor, invertebrate v1.14569; v1.14570 K09071 Heart-and neural crest derivatives-expressed protein v1.16783 [+ 4 other sequence copies] K02168 High-affinity choline transport protein v1.13837 K07608 Internexin neuronal intermediate filament protein, alpha v1.01671 K04309 Leucine-rich repeat-containing G protein-coupled receptor 4 v1.09480; v1.05605 K04308 Leucine-rich repeat-containing G protein-coupled receptor 5 v1.15300 [+ 8 other sequence copies] K08399 Leucine-rich repeat-containing G protein-coupled receptor 6 v1.17524 [+ 14 other sequence copies] K04306 Leucine-rich repeat-containing G protein-coupled receptor 7 v1.21700; v1.03578; v1.17196 K04307 Leucine-rich repeat-containing G protein-coupled receptor 8 v1.16104 K08396 Mas-related G protein-coupled receptor member X v1.08718; v1.02042; v1.02042 K04604 Metabotropic glutamate receptor 1/5 v1.22794 [+ 7 other sequence copies] K04605 Metabotropic glutamate receptor 2/3 v1.15331 K04607 Metabotropic glutamate receptor 4 v1.01418 K04608 Metabotropic glutamate receptor 6/7/8 v1.21698; v1.04544; v1.21739 K14636 MFS transporter, solute carrier family 18 (acetylcholine transporter) 3 v1.05751; v1.19720; v1.22165; v1.02336 K04134 Muscarinic acetylcholine receptor v1.11550 K04129 Muscarinic acetylcholine receptor M1 v1.01913 [+ 4 other sequence copies] K04131 Muscarinic acetylcholine receptor M3 v1.18723 K04132 Muscarinic acetylcholine receptor M4 v1.08171 K04133 Muscarinic acetylcholine receptor M5 v1.07408 [+ 34 other sequence copies] K02583 Nerve growth factor receptor (TNFR superfamily member 16) v1.15265 [+ 91 other sequence copies] K06491 Neural cell adhesion molecule v1.13789; v1.24010; v1.03980 K09038 Neural retina-specific leucine zipper protein v1.24586; v1.16386; v1.16387 K08052 Neurofibromin 1 v1.05520; v1.15407; v1.07950 K04572 Neurofilament light polypeptide v1.19724 K04573 Neurofilament medium polypeptide (neurofilament 3) v1.15787 [+ 4 other sequence copies] K09081 Neurogenin 1 (neurogenic differentiation protein) v1.00345; v1.05338; v1.10997 K08033 Neurogenic differentiation factor 1 v1.07355; v1.14517 K09078 Neurogenic differentiation factor 2 v1.08832 K09079 Neurogenic differentiation factor 4 v1.06678; v1.06677 K01393 Neurolysin v1.16238 [+ 19 other sequence copies] K06756 Neuronal cell adhesion molecule v1.20460; v1.16967 K06757 Neurofascin NFASC (cell adhesion molecule CAMs) v1.22060; v1.03561 K07525 Neuronal guanine nucleotide exchange factor v1.03908 K09098 Neuronal PAS domain-containing protein 1/3 v1.00089 K05247 Neuropeptide FF-amide peptide v1.21565 K08375 Neuropeptide FF receptor 2 v1.06392 [+ 11 other sequence copies] K04209 Neuropeptide Y receptor, invertebrate v1.08609 [+ 31 other sequence copies] K06819 Neuropilin 2 v1.11492 [+ 5 other sequence copies] K03308 Neurotransmitter:Na+ symporter, NSS family v1.16744 [+ 8 other sequence copies] K06774 Neurotrimin v1.05353 K03176 Neurotrophic tyrosine kinase receptor type 1 v1.20055 K04360 Neurotrophic tyrosine kinase receptor type 2 v1.03803 K04356 Neurotrophin 3 v1.09523 K04803 Nicotinic acetylcholine receptor alpha-1 (muscle) v1.11940 K04806 Nicotinic acetylcholine receptor alpha-4 v1.01548 K04808 Nicotinic acetylcholine receptor alpha-6 v1.05056; v1.12097 K04809 Nicotinic acetylcholine receptor alpha-7 v1.07222; v1.11069 K04810 Nicotinic acetylcholine receptor alpha-9 v1.18231 [+ 32 other sequence copies] K05312 Nicotinic acetylcholine receptor, invertebrate v1.24404 K04813 Nicotinic acetylcholine receptor beta-2 (neuronal) v1.06514; v1.23640 K04815 Nicotinic acetylcholine receptor beta-4 v1.18634 K04816 Nicotinic acetylcholine receptor delta v1.18231 [+ 32 other sequence copies] K05312 Nicotinic acetylcholine receptor, invertebrate v1.05293 [+ 78 other sequence copies] K02599 Notch protein v1.15348 [+ 4 other sequence copies] K04256 c-Opsin protein v1.01972 K08385 G0-Opsin protein v1.13345 [+ 5 other sequence copies] K04255 r-Opsin protein v1.00749; v1.03435 K00504 Peptidylglycine monooxygenase v1.12323 [+ 11 other sequence copies] K00678 Phosphatidylcholine-retinol O-acyltransferase v1.18340 [+ 6 other sequence copies] K09624 Protease, serine, 12 (neurotrypsin, motopsin) v1.08030 [+ 9 other sequence copies] K01931 Protein neuralized v1.04431 K09333 Retina and anterior neural fold homeobox-like protein v1.01789; v1.06542 K00061 Retinol dehydrogenase v1.05804 [+ 6 other sequence copies] K11150 Retinol dehydrogenase 8 v1.22340; v1.14029 K11151 Retinol dehydrogenase 10 v1.24399; v1.07017 K11154 Retinol dehydrogenase 16 v1.19667; v1.16885; v1.24371 K00909 Rhodopsin kinase v1.12432; v1.15302; v1.07505 K09516 all- trans -Retinol 13,14-reductase v1.09104 [+ 6 other sequence copies] K05613 Solute carrier family 1 (glial high affinity glutamate transporter), member 2 v1.19779; v1.08769; v1.22032 K05617 Solute carrier family 1 (high affinity Asp/glutamate transporter), member 6 v1.19293; v1.19292 K14387 Solute carrier family 5 (high affinity choline transporter), member 7 v1.10901; v1.19493 K05336 Solute carrier family 6 (neurotransmitter transporter), invertebrate v1.24615 [+ 10 other sequence copies] K05034 Solute carrier family 6 (neurotransmitter transporter, GABA) member 1 v1.07932 K05046 Solute carrier family 6 (neurotransmitter transporter, GABA) member 13 v1.01817 K05036 Solute carrier family 6 (neurotransmitter transporter, dopamine) member 3 v1.20691; v1.16333; v1.15484; v1.02123 K05038 Solute carrier family 6 (neurotransmitter transporter, glycine) member 5 v1.15484; v1.15484 K05042 Solute carrier family 6 (neurotransmitter transporter, glycine) member 9 v1.18461; v1.09068; v1.02237; v1.20880 K05333 Solute carrier family 6 (neurotransmitter transporter) member 18 v1.02239; v1.13836; v1.09067 K05334 Solute carrier family 6 (neurotransmitter transporter) member19 v1.21997 [+ 5 other sequence copies] K12839 Survival of motor neuron-related-splicing factor member 30 v1.21997 [+ 6 other sequence copies] K13129 Survival motor neuron protein Cnidarians differentiate highly specialised sensory and mechanoreceptor cells involved in the capture of prey and for defence against predators. Their stinging cells, termed nematocysts or cnidocytes, are stimulated by adjacent chemosensory cells. Nematocysts trigger the release of a stinging barb (cnidae tubule) via ultra-fast exocytosis on physical contact with ciliary mechanoreceptors of the cnidocyte to deliver the discharge of its venom [ 136 ]. Despite considerable advances in the sensory biology of cnidarians, knowledge of the specific receptor genes that regulate cnidocyte function remains incomplete. In Hydra , and perhaps other cnidarians, cnidocyte discharge is controlled by an ancient light-activated, opsin-mediated phototransduction pathway [ 137 ] that precedes the evolution of cubozoan (box jellyfish) eyes [ 138 ]; cubozoans are the most basal of animals to have eyes containing a lens and ciliary-type visual cells similar to that of vertebrate eyes [ 139 ]. These G-coupled opsin photoreceptors of the retinylidene-forming protein family encoded in the genome of A. digitifera include rhodopsin, bacteriorhodopsin, c-opsin, r-opsin and G 0 -opsin (Table 5 ), but not the Gs-subfamily of opsin receptors reported to be present in sea anemones, hydra and jellyfish [ 140 ], that together with cyclic nucleotide-gated (CNG) ion channel proteins, arrestin (β-adrenergic receptor inhibitor) and other retino-protein receptors, are usual components of the bilaterian phototransduction cascade. Present also are genes to express rhodopsin kinase and β-adrenergic receptor kinase which are related members of the serine/threonine kinase family of proteins that specifically initiate deactivation of G-protein coupled receptors. Additional proteins of retinol metabolism of the phototransduction pathway encoded in the A. digitifera genome are retinol dehydrogenase, all- trans -retinol 13,14 reductase and phosphatidylcholine (lichthin)-retinol O-acyltransferase, a neural retina-specific leucine zipper protein that is an intrinsic regulator of photoreceptor development and function, and a retina and anterior neural fold homeobox-like protein that modulates the expression of photoreceptor genes within the rhodopsin promoter. The genome of A. digitifera encodes also a blue light-sensing, cryptochrome photoreceptor thought to signal synchronous coral spawning by detecting illumination from the lunar cycle [ 98 , 99 ]. The A. digitifera genome reveals genes to express a broad array of neurotransmitter receptor proteins (Table 5 ), including glycine and glutamate neuroreceptors, adrenergic receptors that target non-dopamine catecholamines (i.e., epinephrine and norepinephrine), dopamine, muscarinic and nicotinic acetylcholine receptors, sensory G protein-coupled receptors and γ-aminobutyric acid (GABA) ligand-gated ion channel and G protein-coupled receptors (and inhibitors), several of which are encoded in high copy numbers. Cellular trafficking of neurotransmitters to presynaptic terminals is essential for neurotransmission, and significantly the genome of A. digitifera encodes a wide range of solute carrier neurotransmitter transporters, including a high affinity choline transporter and an acetylcholine-specific protein belonging to the major facilitator superfamily (MFS) of secondary transporters. Encoded also is dopamine β-monooxygenase that catalyses the conversion of dopamine to norepinephrine in the catecholamine biosynthetic pathway, which is necessary for cross-activation of adrenergic neuroreceptors [ 141 ]. Notably, the A. digitifera genome encodes acetylcholinesterase that is expressed at neuromuscular junctions and cholinergic synapses where its protease activity serves to terminate synaptic transmission. The primitive nervous networks of cnidarians are strongly peptidergic with at least 35 neuropeptides identified from different cnidarian classes [ 142 ]. Our annotation of the sequenced A. digitifera genome, however, revealed only the neuropeptide FF-amide neurotransmitter, a RF amide related peptide, and its neuropeptide FF and Y receptors (Table 5 ). Neuropeptides are usually expressed as large precursor proteins which comprise multiple copies of “immature” neuropeptides. Our annotation did not readily reveal these precursor neuropeptide proteins, but we did find enzymes required for their processing, for example, a variety of carboxypeptidase enzymes (not tabulated) that remove propeptide carboxyl residues at basic peptidase sites, and the mature peptide neurotransmitters that are finished by consecutive modification by peptidylglycine (α-hydroxylating) monooxidase (PHM) and peptidyl α-hydroxyglycine α-amidating lyase (PAL) enzymes, both of which are commonly expressed in mammals as a single bifunctional peptidylglycine monooxygenase (K00504/EC 1.14.17.3) [ 143 ]. Our extensive catalogue of animal-like neural and sensory proteins revealed by genome annotation is testament that essential neurobiological features were developed in the primitive neural networks of early eumetazoan evolution. Calcification and Ca 2+ -signalling proteins The massive structures of coral reefs evident today are a construction of aggregated calcium carbonate deposited over long geological time by scleractinian corals and other calcifying organisms, yet our understanding of the molecular processes that regulate the biological processes of coral calcification is limited [ 144 ]. Ca 2+ transfer from seawater to the calicoblastic site of coral calcification occurs by passive diffusion through the gastrovascular cavity [ 145 ] and by active calcium transport [ 146 ]. Active entry of Ca 2+ through the oral epithelial layer is regulated by voltage-dependent calcium channels, such as demonstrated by the L-type alpha protein cloned from the reef-building coral Stylophora pistillata [ 147 ]. Ca 2+ transport across the calioblastic ectoderm to the extracellular calcifying site is facilitated by the plasma-membrane ATP-dependent calcium pump that in S. pistillata resemble the Ca 2+ -ATPase family of mammalian proteins [ 148 ]. By 2H + /Ca 2+ -exchange at the calioblastic membrane, Ca 2+ -ATPase removes H + (from the net reaction Ca 2+ + CO 2 + H 2 O ⇒ CaCO 3 + 2H + ) thereby increasing the saturation state of CaCO 3 to sustain calcium precipitation [ 146 ]. Importantly, located also at the calicoblastic membrane is carbonic anhydrase [ 149 ] which is required to catalyse the intermediate step of calcification by the reversible hydration of carbon dioxide (CO 2 + H 2 O ⇒ HCO 3 - + H + ). In coral phototrophic symbiosis, despite numerous studies describing the well-known phenomenon of light-enhanced calcification, the relationship linking symbiont photosynthesis to coral calcification has been elusive [ 150 , 151 ]. Nonetheless, efforts to better understand the calcifying response of scleractinian corals to environmental change and ocean acidification are gaining traction [ 149 , 152 , 153 ]. Voltage-gated calcium channels (VGCCs) have been examined extensively in mammalian physiology for converting membrane potential into intracellular Ca 2+ transients for signalling transduction pathways (reviewed in [ 154 ]). VGCC signalling affects cellular processes to include muscle contraction, neuronal excitation, gene transcription, fertilisation, cell differentiation and development, proliferation, hormone release, activation of calcium-dependent protein kinases, cell death via necrosis and apoptosis pathways, phagocytosis and endo/exocytosis. Remarkably, annotation of the genome of A. digitifera reveals sequences encoding homologues of all the VGCC (α, αδ, β, and γ) subunits of the molecular (L, N, P/Q and R) phenotypes expressed in mammalian physiology (Table 6 ). There are multiple sequences encoding three variants of Ca 2+ -transporting ATPase, of which at least one is necessary for coral calcification. There is only one sequence match for expressing carbonic anhydrase in the genome of A. digitifera , which may reflect the high catalytic efficiency of this calcifying enzyme [ 155 ], although a BLAST search of ZoophyteBase does reveal scaffolds with low e-values which on future experimental inspection might uncover multiple copies of this enzyme essential for calcification. There are multiple sequences that express solute carrier Na + /Ca 2+ - and Na + /K + /Ca 2+ -exchange families of transport proteins that with expression of the coral Ca 2+ /H + -antiporter may regulate cellular pH and Ca 2+ homeostasis. Table 6 Calcification and Ca 2+ -signalling proteins in the predicted proteome of A. digitifera Gene sequence KEGG Orthology Encoded protein description v1.06452; v1.06451; v1.24424; v1.16923 K07300 Ca2+:H+ antiporter v1.01669 [+ 9 other sequence copies] K01537 Ca2+−transporting ATPase v1.22367; v1.22366; v1.22365 K05850 Ca2+ transporting ATPase, plasma membrane v1.19074 K05853 Ca2+ transporting ATPase, sarcoplasmic/endoplasmic reticulum v1.22416; v1.22417; v1.15682; v1.00750 K14757 Calbindin D28 v1.24568 [+ 9 other sequence copies] K01672 Carbonic anhydrase v1.09241 K08272 Calcium binding protein 39 v1.02323 [+ 39 other sequence copies] K13448 Calcium-binding protein CML v1.05162 [+ 21 other sequence copies] K13412 Calcium-dependent protein kinase v1.09352 K07359 Calcium/calmodulin-dependent protein kinase kinase v1.06475; v1.07555;v1.00945; v1.00159; v1.21122 K08794 Calcium/calmodulin-dependent protein kinase I v1.06475; v1.01061; v1.21150; v1.22443 K04515 Calcium/calmodulin-dependent protein kinase II v1.00159 K05869 Calcium/calmodulin-dependent protein kinase IV v1.21927; v1.01218; v1.22226; v1.06623; v1.13703 K06103 Calcium/calmodulin-dependent serine protein kinase v1.13460 K08284 Calcium channel MID1 v1.20738; v1.01401 K12841 Calcium homeostasis endoplasmic reticulum protein v1.22794 [+ 11 other sequence copies] K04612 Calcium-sensing receptor v1.10079 [+ 17 other sequence copies] K02183 Calmodulin v1.10994 K14734 S100 calcium binding protein G v1.02488 [+ 14 other sequence copies] K05849 Solute carrier family 8 (sodium/calcium exchanger) v1.23153 [+ 9 other sequence copies] K13749 Solute carrier family 24 (sodium/potassium/calcium exchanger) v1.14863 K12304 Soluble calcium-activated nucleotidase 1 v1.18656 [+ 13 other sequence copies] K04858 Voltage-dependent calcium channel alpha-2/delta-1 v1.13222 K04860 Voltage-dependent calcium channel alpha-2/delta-3 v1.08078 [+ 9 other sequence copies] K05315 Voltage-dependent calcium channel alpha 1, invertebrate v1.03896 [+ 6 other sequence copies] K05316 Voltage-dependent calcium channel alpha-2/delta, invertebrate v1.04798 K05317 Voltage-dependent calcium channel beta, invertebrate v1.22788 K04863 Voltage-dependent calcium channel beta-2 v1.09999 K04872 Voltage-dependent calcium channel gamma-7 v1.02505 K04873 Voltage-dependent calcium channel gamma-8 v1.03648[+ 6 other sequence copies] K04850 Voltage-dependent calcium channel L type alpha-1C v1.03648; v1.17267 K04851 Voltage-dependent calcium channel L type alpha-1D v1.03648; v1.13219; v1.21895 K04857 Voltage-dependent calcium channel L type alpha-1S v1.06313; v1.01656; v1.23096 K04344 Voltage-dependent calcium channel P/Q type alpha-1A v1.08078 [+ 10 other sequence copies] K04849 Voltage-dependent calcium channel N type alpha-1B v1.07968 K04852 Voltage-dependent calcium channel R type alpha-1E v1.01364; v1.13467; v1.08705 K04854 Voltage-dependent calcium channel T type alpha-1G v1.15414; v1.14241; v1.09595 K04855 Voltage-dependent calcium channel T type alpha-1H Implicit to coral calcification is Ca 2+ regulation that affects signalling of other vital cellular functions. Cellular Ca 2+ is mediated by the calcium-sensing receptor calmodulin (18 sequence matches) and other messenger calcium-binding effectors (Table 6 ), including the calcium-binding protein CML (40 protein domain sequence matches). Calcium/calmodulin-protein kinase proteins are arguably key to Ca 2+ -signalling in coral symbiosis but, with the exception of activation of sperm flagellar motility [ 156 ], their precise role has not been elaborated. Plant-derived proteins Endosymbiosis has contributed greatly to eukaryotic evolution, most notably to the genesis of plastids and mitochondria derived from prokaryotic antecedents. Genetic integration by endosymbiont-to-host transfer (EGT) or replacement (EGR) has been a significant force in early metazoan innovation, whereby nuclear transferred genes may even adopt novel functions in the host cell or replace existing versions of the protein that they encode [ 157 ]. Prokaryote-to-eukaryotic gene transfer has been widespread in evolution, but examples of genetic exchange between unrelated eukaryotes, such as between algal symbionts and their multicellular eukaryote host, are considered rare (reviewed by [ 158 , 159 ]). One such example is aroB (3-dehydroquinate synthase) transferred to the genome of the sea anemone N. vectensis , which sequence best fits that of the dinoflagellate Oxyrrhis marina [ 119 ]. Close inspection of the amino acid sequence of the aroB gene product, as reported by Shinzato et al. [ 45 ], clearly shows this protein to be 2- epi -5- epi -valiolone synthase (EVS), a sugar phosphate cyclase orthologue that catalyses the conversion of sedoheptulose 7-phosphate to 2- epi -5- epi -valiolone found to be a precursor of the mycosporine-like amino acid (MAA) sunscreen shinorine in the cyanobacterium Anabaena variabilis [ 160 ]. Additionally, the EVS gene of N. vectensis has a distinctive O -methytransferase fusion that is identical in O. marina [ 161 ]. The shikimate pathway is essential to apicomplexan parasites of the genera Plasmodium , Toxoplasma and Cryptosporidium and of Tetrahymena ciliates to express a pentafunctional aroM gene similar to that of Ascomycetes, which is thought to have been conveyed by fungal gene transfer to a common ancestral progenitor [ 162 ]. In a separate example, H. viridis expresses a plant-like ascorbate peroxidase gene ( HvAPX1 ) during oogenesis in both symbiotic and aposymbiotic individuals [ 163 ], whereby peroxidase activity is coincident with oogenesis and embryogenesis that in Hydra acts as a ROS scavenger to protect the oocyte from apoptotic degradation [ 164 ]. The sacoglossan (sea slug) molluscs Elysia chlorotica and E. viridis (Plakobranchidae) acquire plastids on ingestion of the siphonaceous alga Voucherea litorea (termed “kleptoplasty”) and, by maintaining sequestered plastids in an active photosynthetic state, has emerged as a model organism for the transfer of nuclear-encoded plant genes from algal symbiont to its animal host [ 165 ]. In this symbiosis, the family of light-harvesting genes psbO , prk (phosphoribokinase) and chlorophyll synthase ( chlG ) are entrained in the genome of Elysia chlorotica (reviewed in [ 166 , 167 ]), although there is debate whether these genes are transcriptionally expressed (compare [ 168 ] and [ 169 ]). Also, phylogenomic analysis of the predicted proteins of the aposymbiotic unicellular choano-flagellate Monosiga brevicollis , considered to be a stem progenitor of the animal kingdom [ 170 , 171 ], reveals 103 genes having strong algal affiliations arising from multiple phototrophic donors [ 172 ]. Such notable examples illustrate the transfer of algal genes to animal recipients. KEGG orthology-based annotation of the predicted proteome of A. digitifera reveals a plethora of sequences presumed to be of algal origin (Table 7 ). Like E. chlorotica , the coral genome has encoded the photosystem II (PSII) protein PsbO of the oxygen-evolving complex of photosynthesis, as well as the PSII light-harvesting complex protein PsbL that is important in protecting PSII from photo-inactivation [ 173 ]. Encoded also are the photosystem I subunit proteins PsaI and PsaO. Additionally encoded are the photosystem P840 reaction center cytochrome c551 (PscC) protein and the photosynthetic reaction center M subunit protein, the light-harvesting proteins complex 1 alpha (PufA), the complex II chlorophyll a / b binding protein 6 (LHCB6), the cyanobacterial phycobilisome proteins AcpF and AcpG, the phycocyanin-associated antenna protein CpcD, the phycocyanobilin lyase protein CpcF and the phycoerythrin-associated linker protein CpeS. Like E. chlorotica , the coral genome encodes chlorophyll synthase (ChlG), a chlorophyll transporter protein PucC, a light-independent nitrogenase-like protochlorophyllide reductase enzyme that is sensitive to oxygen [ 174 ] and a red chlorophyll reductase essential to the detoxification of photodynamic chlorophyll catabolites arising from plant/algal senescence [ 175 ]. Three chlorosome proteins of the photosynthetic antenna complex of green sulphur bacteria, a bacteriochlorophyll methyltransferase involved in BChl c biosynthesis [ 176 ] and the retinylidene bacteriorhodopsin of phototrophic Archaea are also encoded in the coral genome. Present are genes encoding subunit 6 of the cytochrome B 6 f complex that links PSII and PSI via the plastoquinone pool, together with chloroplast ferredoxin-like NapH and NapG proteins and their 2Fe-2S cluster protein. The coral genome, however, encodes sequences for NAD + -ferredoxin reductase (HcaD; not tablulated), rather than the required NADP + -ferredoxin reductase of photosynthesis. Annotation of the A. digitifera genome revealed genes unexpectedly encoding ferredoxin hydrogenase [EC:1.12.7.2] and that of its small subunit protein (Table 7 ) involved in light-dependent production of molecular hydrogen having its [Fe-Fe]-cluster coupled to the photosynthetic transport chain via a charge-transfer complex with ferredoxin (see [ 177 ]). Table 7 Plant-derived proteins in the predicted proteome of A. digitifera Gene sequence KEGG Orthology Encoded protein description v1.14452 K09843 (+)-Abscisic acid 8′-hydroxylase v1.18868 K14496 Abscisic acid receptor PYR/PYL family (PYL) v1.21983; v1.05890 K03342 p-Aminobenzoate synthetase / 4-amino-4-deoxychorismate lyase (PabBC) v1.15436 K02822 Ascorbate-specific IIB component, PTS system (PTS-Ula-EiiB) v1.11187; v1.13966 K00423 L-Ascorbate oxidase v1.20081; v1.22465 K13604 Bacteriochlorophyll C20 methyltransferase (BchU) v1.07723 K04641 Bacteriorhodopsin (BoP) v1.21858 K04040 Chlorophyll synthase (ChlG) v1.01742 K08945 Chlorosome envelope protein A (CsmA) v1.04797; v1.14208 K08946 Chlorosome envelope protein B (CsmB) v1.18698 K08948 Chlorosome envelope protein D (CamD) v1.18637 K02642 Cytochrome b 6 f complex subunit 6 (PetL) v1.21101; v1.14192; v1.14548 K01735 3-Dehydroquinate synthase (AroB) v1.05796 K10210 4,4′-Diaponeurosporene oxidase (carotenoid biosynthesis; CrtP) v1.11730 K04755 Ferredoxin, 2Fe-2S (FdX) v1.19154; v1.00014 K00532 Ferredoxin hydrogenase v1.00014 K00534 Ferredoxin hydrogenase small subunit v1.17698; v1.06031; v1.16647 K02574 Ferredoxin-type protein (NapH) v1.23058 K02573 Ferredoxin-type protein (NapG) v1.08414 K08926 Light-harvesting complex 1 alpha chain (PufA) v1.21458 K08917 Light-harvesting complex II chlorophyll a/b binding protein 6 (LHCB6) v1.03743 K08226 MFS transporter, BCD family, chlorophyll transporter (PucC) v1.13030; v1.08678 K13413 Mitogen-activated protein kinase kinase 4/5, plant ((MKK4_5P) v1.02429; v1.10744; v1.03340 K08929 Photosynthetic reaction center M subunit (PufM) v1.03631 K02696 Photosystem I subunit VIII (PsaI) v1.11432 K14332 Photosystem I subunit (PsaO) v1.17422 K02713 Photosystem II protein (PsbL) v1.18303 K02716 Photosystem II oxygen-evolving enhancer protein 1 (PsbO) v1.12300; v1.21136 K08942 Photosystem P840 reaction center cytochrome c551 ((PscC) v1.00280 K02097 Phycobilisome core component 9 (AcpF) v1.10967 K02290 Phycobilisome rod-core linker protein (AcpG) v1.02166 K02287 Phycocyanin-associated, rod protein (CpcD) v1.19642; v1.07305; v1.19572; v1.01248 K02289 Phycocyanobilin lyase beta subunit (CpcF) v1.10441 K05382 Phycoerythrin-associated linker protein (CpeS) v1.13406 K10027 Phytoene dehydrogenase (desaturase; CrtI) v1.18809; v1.06199 K02291 Phytoene synthase (CrtB) v1.20411; v1.02037; v1.14064; v1.21095 K09060 Plant G-box-binding factor (GBF) v1.10035 K00218 Protochlorophyllide reductase [NifEN-like; Por] v1.21846 K05358 Quinate dehydrogenase (QuiA) v1.03127 K13545 Red chlorophyll catabolite reductase (ACD2) v1.05899 K00891 Shikimate kinase (AroK, AroL) v1.21101; v1.14192; v1.05899 K13829 Shikimate kinase / 3-dehydroquinate synthase (AroKB) v1.12938 K08500 Syntaxin of plants (SYP6) v1.06575 K08506 Syntaxin of plants (SYP7) v1.04929 K09834 Tocopherol cyclase (VTE1, SXD1) v1.01022 K05928 Tocopherol O -methyltransferase v1.05457 K09838 Zeaxanthin epoxidase (ZEP, ABA1) Like N. vectensis and the dinoflagellate Oxyrrhis marina , the genome of A. digitifera encodes an O -methyltransferase which is immediately downstream of EVS, but the two genes are not fused. Using a ZoophyteBase BlastP search, the O -methyltransferase showed little sequence homology with the corresponding protein of A. variabilis (e-value of 6.972E -2 and Bit score of 34.27), whereas the EVS protein shared 87% absolute sequence identity to the A. variabilis EVS protein. What role, if any, these two genes play in mycosporine-like amino acid (MAA) biosynthesis in A. digitifera has yet to be determined, although it has been suggested from the transcriptome of Acropora microphthalma that MAA biosynthesis proceeds from a branch point at 3-dehydroquinate of the shikimic acid pathway as a shared metabolic adaptation between the coral host and its symbiotic zooxanthellae [ 40 ]. The 3-dehydroquinate synthase enzyme of the shikimic acid pathway, thought to be a key intermediate in an alternative MAA biosynthetic pathway in A. variabilis [ 178 ], is instead encoded by the fused aroKB gene of A. digitifera (Table 7 ). Additional shikimate proteins of the predicted proteome, although not limited to phototrophs, are shikimate kinase (AroK), quinate dehydrogenase (QuiA) and the conjoined p -aminobenzoate synthase and 4-amino-4-deoxychlorismate lysate (PabBC) enzyme necessary for folate biosynthesis [ 179 ]. Other plant-related gene homologues include the phytohormone abscisic acid receptor protein (PabBC) and its cytochrome P450 monooxygenase abscisic acid 8′-hydroxylase, L-ascorbate oxidase and PTS system degrading enzymes, the unique SYP6 and SYP7 syntaxins of plant vesicular transport, tocopherol cyclase and a tocopherol O -methyltransferase enzyme that converts γ-tocopherol to α-tocopherol. Essential for carotene biosynthesis are phytoene synthase (CrtB) and phytoene dehydrogenase (CrtI) enzymes. Significantly, encoded within the coral genome is zeaxanthin epoxidase that is essential for abscisic acid biosynthesis and is a key enzyme in the xanthophyll cycle of plants and algae to impart oxidative stress tolerance. Given that viruses often mediate gene transfer processes, it is intriguing that certain bacteriophages of marine Synechococcus and Prochlorococcus cyanobacteria are reported to carry genes encoding the photosynthesis D1 ( psbA ), and D2 ( psbD ) proteins, a high-light inducible protein (HLIP) [ 180 , 181 ] and the photosynthetic electron transport plastocyanin ( petE ) and ferredoxin ( petF ) proteins thought to enhance the photosynthetic fitness of their host [ 182 - 184 ]. Accordingly, it has been suggested that the transfer of psbA by viruses associated with Symbiodinium could lessen the severity of thermal impairment to PSII and the response of corals to thermal bleaching [ 185 ]. It is yet unknown if phages or dinoflagellate-infecting viruses [ 186 ], particularly those of Symbiodinium [ 187 ], may affect gene transfer leading to complementary (or “shared”) metabolic adaptations of symbiosis [ 119 , 188 ]. Proteins of nitrogen metabolism It is well accepted that intracellular Symbiodinium spp. provide reduced carbon for coral heterotrophic metabolism by photosynthetic carbon fixation. Because of this metabolic relationship, light is a critical feature in the bioenergetics of coral symbiosis [ 189 ]. The algal photosynthate translocated to corals, however, is deficient in nitrogen at levels necessary to sustain autotrophic growth. While corals can assimilate fixed nitrogen from surrounding seawater [ 190 ], “recycled” nitrogen within the symbiosis may account for as much as 90% of the photosynthetic nitrogen demand [ 191 ]. It would not be surprising then that light would have a strong influence on the uptake and retention of ammonium by symbiotic corals. Consequently, corals excrete excess ammonium in darkness [ 192 ], and in light excretion is induced by treatment with the photosynthetic electron transport inhibitor 3-(3,4-diclorophenyl)-1,1-dimethylurea (DCMU) [ 193 ]. Since ammonia is the product of nitrogen fixation, these observations suggest that the coral holobiont may fix nitrogen in the dark, or when photosynthesis is repressed, during which coral tissues are hypoxic [ 194 ], and nitrogenase activity is not inactivated by molecular oxygen [ 195 ]. Tropical coral reefs are typically surrounded by low-nutrient oceanic waters of low productivity but, paradoxically, the waters of coral reefs often have elevated levels of inorganic nitrogen [ 196 , 197 ] attributed to high rates of nitrogen fixation. While nitrogen fixation from diazotrophic epiphytes of the coral reef substrata and sediments [ 197 , 198 ] and diazotrophic bacterioplankton of the coral reef lagoon [ 199 ] provide substantial quantities of fixed nitrogen for assimilation by the coral reef, mass-balance estimates show this input to be less than the community’s annual nitrogen demand [ 200 ]. Endolithic nitrogen-fixingbacteria are abundant in the skeleton of living corals where they benefit from organic carbon excreted by overlaying coral tissues to provide a ready source of energy for dinitrogen reduction [ 201 ]. Additionally, intracellular nitrogen-fixing cyanobacteria are reported to coexist with dinoflagellate symbionts in the tissues of Monastraea cavernosa and to functionally express nitrogenase activity [ 202 ]. Corals also harbour a diverse assemblage of heterotrophic microorganisms in their skele-ton, tissues and lipid-rich mucus (reviewed in [ 203 ]), and these communities include large populations of diazo-trophic bacteria [ 204 , 205 ], and archaea [ 206 ]. Apart from nitrogen fixation, the coral microbiota contributes to other nitrogen-cycling processes, such as nitrification, ammonification and denitrification [ 207 , 208 ]. We were surprised to find several nitrogen fixation and cycling proteins encoded in the genome of A. digitifera (Table 8 ), notably a nitrogen fixation NifU-like protein, the Nif-specific regulatory protein (NifA), the regulatory NAD(+)-dinitrogen-reductase ADP-D-ribosylastransferase protein, a nitrifying ammonia monooxy-genase enzyme and nitrate reductase, which are usually expressed only by prokaryotic microorganisms. Table 8 Proteins of nitrogen metabolism in the predicted proteome of A. digitifera Gene sequence KEGG Orthology Encoded protein description v1.23444; v1.09133; v1.23443 K05521 ADP-ribosylglycohydrolase (DraG) v1.09202 K10944 Ammonia monooxygenase subunit A v1.03645 [+ 8 other sequence copies] K03320 Ammonium transporter, Amt family v1.12268; v1.12269 K06580 Ammonium transporter Rh v1.02406 K01954 Carbamoyl-phosphate synthase (CPS) v1.01524; v1.18283; v1.18284 K01948 Carbamoyl-phosphate synthase (CPS, ammonia) v1.01615 K04016 Formate-dependent nitrite reductase (NrfA) v1.16277; v1.23483; v1.13667; v1.22675 K00261 Glutamate dehydrogenase (NAD(P)+) v1.17166; v1.11089 K01745 Histidine ammonia-lyase v1.22825; v1.08034;v1.o8520 K05123 Integration host cell factor (INF) subunit beta v1.11343 K05951 NAD+−dinitrogen-reductase ADP-D-ribosyltransferase (DraT) v1.00547 K02584 Nif-specific regulatory protein (NifA) v1.18869 K00371 Nitrate reductase 1, beta subunit v1.06763 K08346 Nitrate reductase 2, beta subunit v1.14858; v1.00685; v1.23148 K05916 Nitric oxide dioxygenase v1.16954 [+ 5 other sequence copies] K02448 Nitric oxide reductase NorD protein v1.06115 K02164 Nitric oxide reductase NorE protein v1.17629 [+ 12 other sequence copies] K04748 Nitric oxide reductase NorQ protein v1.24077 [+ 4 other sequence copies] K13125 Nitric oxide synthase-interacting protein v1.21801; v1.05719; v1.23577; v1.19464 K13253 Nitric-oxide synthase, invertebrate v1.05980 K00363 Nitrite reductase (NAD(P)H) small subunit v1.00101 K02598 Nitrite transporter NirC v1.02355; v1.18772 K04488 Nitrogen fixation protein NifU v1.17812 K02589 Nitrogen regulatory protein PII 1 v1.09150 K02570 Periplasmic nitrate reductase NapD v1.01560 K02571 Periplasmic nitrate reductase NapE v1.10035 K00218 Protochlorophyllide reductase [NifEN-like] v1.08939 K03737 Pyruvate-flavodoxin reductase (NifJ) v1.17373 K00365 Urate oxidase v1.13217 K01427 Urease v1.16409 [+ 5 other sequence copies] K03187 Urease accessory protein v1.13217 K01429 Urease subunit beta v1.13217 K14048 Urease subunit gamma/beta v1.12211 [+ 4 other sequence copies] K00106 Xanthine dehydrogenase/oxidase v1.12212 K13481 Xanthine dehydrogenase small subunit (Excluding amino acid and pyrimidine/purine nucleotide synthesis or metabolism). The presence of genes encoding proteins involved in nitrogen fixation raises speculation that corals may contribute directly to, or perhaps co-regulate, certain processes that catalyse the reduction of dinitrogen (N 2 ) to ammonia (NH 3 ) by the enzyme nitrogenase reductase (NifH). The functional NifH enzyme is a binary protein composed of a molybdenum-iron (MoFe) protein (NifB/NifDK), or its NifEN homologue, fused with a FeMo-cofactor (FeMoco) protein [ 209 ]. While genes encoding NifB, NifDK (or NifEN) and their FeMo-cofactor do not appear in the genome of A. digitifera , a gene encoding the NifEN-like protein protochlorophyllide oxidoreductase (POR) is present (Table 8 ). POR has all three subunits with high similarity to the assembled MoFe nitrogenase [ 210 ], but this homologue is unlikely to be effective in nitrogen reduction [ 211 , 212 ] since its activity is light dependent [ 213 ] when tissues are highly oxic [ 193 ]. The NifU protein encoded in the coral genome preassembles the metallocatalytic Fe-S clusters for maturation of nitrogenase [ 214 ], but its assemblage without NifS, a cysteine desulfurase needed for [Fe-S] cluster assembly [ 215 ], would be incomplete, and its pre-nitrogenase receptor is also missing. Yet, the coral does have the nifJ gene that encodes pyruvate:flavodoxin oxidoreductase required for electron transport in nitrogenase reduction [ 216 ]. The regulatory NifA protein encoded in the coral genome might activate, on stimulation by the integration host factor (INF), transcription of nitrogen fixation ( nif ) operons of RNA polymerase [ 217 ], and both of these proteins are encoded in the coral genome. Additional to this transcriptional control, post-translational nitrogenase activity is controlled by reversible ADP-ribosylation of a specific arginine residue in the nitrogenase complex [ 218 ]. NAD(+)-dinitrogen-reductase ADP-D-ribosyltransferase (DraT) inactivates the nitrogenase complex while ADP-ribosylgly-cohydrolase (DraG) removes the ADP-ribose moiety to restore nitrogenase activity, and both of these enzymes are encoded in the coral genome. Given that genes encoding essential constituent proteins of nitrogenase assembly appear incomplete, corals are unlikely to fix nitrogen per se , but co-opted elements of the coral genome to regulate processes of nitrogen fixation by its diazotrophic consortia is a prospect worthy of exploration [ 219 ]. Nitrofying/nitrifying bacteria and archaea express the enzyme ammonia monooxygenase that converts fixed ammonia to nitrite (via hydroxylamine) and the enzyme nitrite (oxido)reductase completes the oxidation of nitrite to nitrate, and both of these enzymes are entrained in the genome of A. digitifera (Table 8 ). The ammonia monooxygenase subunit A (amoA ) of archaeal consorts has been described in nine species of coral from four reef locations [ 220 ], but the presence of amoA in the coral genome, together with encoded ammonium transport proteins, was not anticipated. Another protein of prokaryotic origin encoded in the coral genome is nitrate reductase (periplasmic, assimilatory and respiratory), the latter being required for anaerobic respiration by bacteria [ 221 ], and unlike the nitrate reductase family of sulphite oxidase enzymes in eukaryotes, the nitrate reductases of prokaryotes (K00363) belong to the DMSO reductase family of enzymes. Also encoded in the coral genome are a nitrite transporter (NirC) and a formate-dependent nitrite reductase (NrfA) required for nitrite ammonification [ 222 ]. In addition to nitrite reduction, NrfA reduces nitric oxide, hydroxylamine, nitrous oxide and sulphite, the last providing a metabolic link between nitrogen and sulphur cycling in coral metabolism. Other enzymes of nitrogen metabolism encoded in the coral genome are the carbamoyl-phosphate synthase family of enzymes [ 223 ] that catalyses the ATP-dependent synthesis of carbamoyl phosphate used for the production of urea (ornithine cycle) to provide a ready store of fixed-N in the urea-nitrogen metabolism of corals [ 224 ]. Another nitrogen source comes from glutamate dehydrogenase (GDH) that reversibly converts glutamate to α-ketoglutarate with liberation of ammonia, and as expected [ 225 ], this enzyme is encoded in the coral genome, together with the prokaryotic nitrogen regulatory protein PII of glutamine synthase, which in bacteria is activated in response to nitrogen availability. Encoded also is histidine ammonia-lyase (histidase) that liberates ammonia (and urocanic acid) from cytosolic stores of histidine. It is now accepted that uric acid deposits accumulated by symbiotic algae provide a significant store of nitrogen for the coral holobiont [ 226 ], so it is noteworthy that the coral genome encodes urate oxidase (uricase) to catalyse uric acid oxidation to allanotoin from which urea and ureidoglycolate are produced in a reaction catalysed by allantoicase (allantoate amidinohydrolase), both of which known isoforms are present in the coral genome. Encoded in the coral genome is also urease to catalyse the hydrolysis of urea, presumably excreted by its algal symbionts, with the release of carbon dioxide and ammonia to meet the nitrogen demand of the coral holobiont during periods of low nitrogen availability. Similarly, xanthine dehydrogenase (xanthine: NAD + -oxidoreductase) acts by oxidation on a variety of purines, including hypoxanthine, to yield urate for the recycling of nitrogen in coral nutrition. Many of the aforementioned proteins of nitrogen metabolism, including Nif proteins, have been detected in the proteome of an endosymbiont-enriched fraction of the coral S. pistillata [ 39 ]. Notwithstanding consideration of the rapid diffusion rate of nitric oxide (NO) or its apparent short biological half-life [ 227 ], there is debate about the provenance of endogenously produced NO in signalling the bleaching of corals in response to environmental stress. Elevated nitric oxide synthase (NOS) activity and NO production in algal symbionts has been attributed to the thermal stress response of corals [ 228 , 229 ], whereas the host is ascribed to be the major source of NO during exposure to elevated temperature [ 230 , 231 ]. While our annotation may not resolve this dispute, we show (Table 8 ) that nitric oxide synthase enzymes (Nor D, Nor E, Nor Q and an invertebrate NOS protein) are encoded in the genome of A. digitifera , together with a nitric oxide-interacting protein (NOIP) that in higher animals regulates neuronal NOS activity [ 232 ]. Nitric oxide is an intermediate of nitrite reduction catalysed by nitrite reductase (NIR), which by further reduction produces ammonia. The coral genome also encodes nitric oxide dioxygenase (NOD) that converts nitric oxide to nitrate. Accordingly, enhanced expression of NIR (NO reduction) or NOD (NO oxidation) could ameliorate the NO-signalling response of coral bleaching presumed activated by environmental stress. DNA repair Cellular DNA is prone to damage caused by the products of normal metabolism and by exogenous agents. Damage to DNA from metabolic processes include the oxidation of nucleobases and strand interruptions by the production of reactive oxygen species (ROS), from alkylation of nucleotide bases, from the hydrolysis of bases causing deamination, depurination and depyrimidination, and from the mismatch of base pairs from errors in DNA replication. Damage affected by external agents include exposure to UV light causing pyrimidine dimerization and free radical-induced damage, exposure to ionising radiation causing DNA strand breaks, thermal disruption causing hydrolytic depurination and single-strand breaks, and by xenobiotic contamination to cause DNA adduct formation, nucleobase oxidation and DNA crosslinking. Most of these lesions affect structural changes to DNA that alter or prevent replication and gene transcription at the site of DNA damage. Thus, recognition and repair of DNA abnormalities are vital processes essential to maintain the genetic integrity of the coral genome. Since there are multiple pathways causing DNA damage at diverse molecular sites, there are likewise diverse and overlapping processes available to repair cellular DNA damage. Of the many nuclear repair processes, photoreactivation (photolyase), base excision repair and nucleotide excision repair are the main elements for the repair of cellular DNA damage. Exposure to sunlight is an absolute requirement for phototrophic symbiosis, but excessive exposure of corals to solar ultraviolet radiation can inflict direct damage to DNA by pyrimidine dimerization and 6-4 photoadduct formation and cause indirect damage by the production of ROS to initiate free-radical damage. While there have been abundant studies on the sensitivity of corals to solar ultraviolet radiation, only a few have examined the effects of solar UV to cause DNA damage. Photoreactivation has been shown to be an important repair pathway for reversing UV-activated DNA damage in adult coral [ 233 ] and coral planulae [ 234 ]. UV damage to DNA was first demonstrated by the detection of unrepaired cyclobutane pyrimidine dimers (CPDs) in the host tissues and algal symbionts of the coral Porities porites , in which CPDs had increased in a UV dose-dependent manner [ 235 ], whereas CPDs and 6-4 pyrimidine-pyrimidone photoadducts in the coral Montipora verrucosa holobiont were correlated inversely with levels of coral “sunscreen” protection [ 236 ]. The effects of solar UV radiation causing DNA lesions in coral have been determined by use of the comet assay [ 237 ], and UV-induced DNA damage and repair has been examined in the symbiotic anemone Aiptasia pallida [ 238 ]. The comet assay showed also that DNA lesions in coral planulae had increased on acquiring algal symbionts, presumably from greater ROS production resulting as a by-product of photosynthesis [ 239 ]. Iron-induced oxidative stress was found likewise to enhance DNA damage in the coral Pocillopora damicornis as determined by the occurrence of DNA apurinic/apyrimidinic sites caused by hydrolytic lesions [ 240 ]. Significantly, DNA damage in the host and algal symbionts of the coral Montastraea faveo-lata was found to occur simultaneously during thermal “bleaching” stress, and DNA damage is further enhanced on exposure to greater irradiances of solar radiation [ 241 ]. Nevertheless, despite the serious risk of unrepaired DNA damage to coral survival, the DNA repair processes of corals to mitigate the detrimental effects of environmental stress have not been adequately characterised at the transcriptome level of expression [ 29 , 242 ]. Our annotation of the sequenced genome of A. digitifera has revealed genes encoding a large repertoire of DNA repairing enzymes and their adaptor proteins (Table 9 ). Given strong evidence for DNA photoreactivation in corals having been reported [ 233 , 234 ], it was surprising to find only one gene in single copy that encodes a sole photolyase enzyme for reversing pyrimidine dimer and 6-4 photoadduct formation. Notably, we found genes encoding 6 members of the ERCC family of nucleotide excision repair enzymes, together with the UV excision repair protein RAD23, for the repair of UV-induced DNA damage. More abundant are the DNA mismatch repair enzymes from the MLH, MSH, Mut and PMS protein families and related glycosylase/lyase proteins for repairing erroneous insertion, deletion and mis-incorporation of bases to arise during DNA replication and recombination. There is additionally a specific gene that encodes a 3′-endonuclease protein that has a preference to correct mispaired nucleotide sequences. Abundant also are other members of the RAD-family of DNA repair proteins, including 28 sequence copies of a gene encoding the RAD50 protein for DNA double-strand break repair that, together with members of the MRE, Rec, REV, Swi5/Sae3, XRCC and XRS families of recombination and polymerase proteins, have complementary roles in DNA repair. Apparent also in the genome are the DNA helicase proteins, including RuvB–like proteins, which are primarily involved in DNA replication and transcription, but assist also in the repair of DNA damage by separating double strands at affected sites of DNA damage to facilitate repair. Of the multiple families of ATP-dependent DNA helicase proteins encoded in the coral genome, RecQ and helicase Q predominate. Encoded in the coral genome are 5 homologues of the DNA repair alkB proteins that reverse damage to DNA from alkylation caused by chemical agents by removing methyl groups from 1-methyl adenine and 3-methyl cytosine products in single-stand DNA. Annotated also are genes encoding DNA ligase 3 for repairing single-strand breaks, DNA ligase 4 to repair double-strand breaks, and a DNA cross-link repair 1C protein with single-strand specific endonuclease activity that may serve in a proofreading function for DNA polymerase. Taken together, expressing this arsenal of DNA protection may provide corals with limited ability to transcribe gene-encoded adaptation to a changing global environment. Table 9 DNA repair proteins in the predicted proteome of A. digitifera Gene sequence KEGG Orthology Encoded protein description v1.02961; v1.13402 K03575 A/G-specific adenine glycosylase (MutY) v1.11766 K03919 Alkylated DNA repair protein v1.04821 K10765 Alkylated DNA repair protein alkB homologue 1 v1.02479 K10766 Alkylated DNA repair protein alkB homologue 4 v1.20302 K10767 Alkylated DNA repair protein alkB homologue 5 v1.24450 K10768 Alkylated DNA repair protein alkB homologue 6 v1.02766; v1.09413 K10770 Alkylated DNA repair protein alkB homologue 8 v1.01590 [+ 4 other sequence copies] K10884 ATP-dependent DNA helicase 2 subunit 1 v1.18810; v1.03166; v1.08449 K10885 ATP-dependent DNA helicase 2 subunit 2 v1.08013 K03722 ATP-dependent DNA helicase DinG v1.03542 K14635 ATP-dependent DNA helicase MPH1 v1.06737 [+ 5 other sequence copies] K15255 ATP-dependent DNA helicase PIF1 v1.17360; v1.21235 K10899 ATP-dependent DNA helicase Q1 v1.01081 [+ 8 other sequence copies] K10730 ATP-dependent DNA helicase Q4 v1.16859 K10902 ATP-dependent DNA helicase Q5 v1.11661 [+ 19 other sequence copies] K03654 ATP-dependent DNA helicase RecQ v1.20397 K03656 ATP-dependent DNA helicase Rep v1.18049; v1.07731; v1.05830 K10905 ATR interacting protein v1.01679 K01669 Deoxyribodipyrimidine photo-lyase v1.03410; v1.12968; v1.00865; v1.16876 K10887 DNA cross-link repair 1C protein v1.07474; v1.07473; v1.01809 K10610 DNA damage-binding protein 1 v1.13116; v1.03378; v1.16328 K10140 DNA damage-binding protein 2 v1.17099 [+ 5 other sequence copies] K11885 DNA damage-inducible protein 1 v1.05469 K06663 DNA damage checkpoint protein v1.02859; v1.14719; v1.21030; v1.10920 K04452 DNA damage-inducible transcript 3 v1.02191 K10844 DNA excision repair protein ERCC-2 v1.19108 [+ 5 other sequence copies] K10843 DNA excision repair protein ERCC-3 v1.22267 [+ 4 other sequence copies] K10848 DNA excision repair protein ERCC-4 v1.15137 [+ 5 other sequence copies] K10846 DNA excision repair protein ERCC-5 v1.18550; v1.02606; v1.14935; v1.08831 K10841 DNA excision repair protein ERCC-6 v1.20045; v1.01844; v1.11724; v1.03203 K10570 DNA excision repair protein ERCC-8 v1.15430; v1.03058 K03658 DNA helicase IV v1.00228 [+ 4 other sequence copies] K11665 DNA helicase INO80 v1.00136; v1.0678; v1.21529 K10776 DNA ligase 3 v1.23293; v1.19418; v1.23430; v1.15721 K10777 DNA ligase 4 v1.19248 K07458 DNA mismatch endonuclease, patch repair protein v1.19011 K08739 DNA mismatch repair protein MLH3 v1.11513; v1.11449 K08735 DNA mismatch repair protein MSH2 v1.14781 K08736 DNA mismatch repair protein MSH3 v1.05696; v1.22444; v1.19162 K08740 DNA mismatch repair protein MSH4 v1.04904 K08741 DNA mismatch repair protein MSH5 v1.15360; v1.19426; v1.08585 K08737 DNA mismatch repair protein MSH6 v1.02429 [+ 8 other sequence copies] K03572 DNA mismatch repair protein MutL v1.03990 K03555 DNA mismatch repair protein MutS v1.14015 K07456 DNA mismatch repair protein MutS2 v1.08443 K10864 DNA mismatch repair protein PMS1 v1.15229 K10858 DNA mismatch repair protein PMS2 v1.08658; v1.14152; v1.01681 K15082 DNA repair protein RAD7 v1.16407 [+ 27 other sequence copies] K10866 DNA repair protein RAD50 v1.22193 K04482 DNA repair protein RAD51 v1.02646; v1.22076 K10958 DNA repair protein RAD57 v1.15671 [+ 4 other sequence copies] K04483 DNA repair protein RadA v1.16193; v1.19033 K04485 DNA repair protein RadA/Sms v1.16079; v1.07685 K04484 DNA repair protein RadB v1.21363; v1.22360; v1.02900 K03584 DNA repair protein RecO (recombination protein O) v1.18390 K03515 DNA repair protein REV1 v1.04705 K10991 DNA repair protein Swi5/Sae3 v1.13920; v1.03800; v1.16133 K10803 DNA repair protein XRCC1 v1.15052 K10879 DNA repair protein XRCC2 v1.09315 [+ 4 other sequence copies] K10886 DNA repair protein XRCC4 v1.02733; v1.24592 K10868 DNA repair protein XRS2 v1.14551; v1.23176 K10873 DNA repair and recombination protein RAD52 v1.20503 [+ 4 other sequence copies] K10875 DNA repair and recombination protein RAD54 v1.23173; v1.16050 K10877 DNA repair and recombination protein RAD54B v1.07227; v1.08907; v1.09439; v1.02644 K10847 DNA repair protein complementing XP-A cells v1.11534 [+ 5 other sequence copies] K10865 Double-strand break repair protein MRE11 v1.07939 K03660 N-glycosylase/DNA lyase v1.16163 K03652 3-Methyladenine DNA glycosylase v1.07231 K10726 Replicative DNA helicase Mcm v1.05482 K04499 RuvB-like protein 1 (pontin 52) v1.19813 K11338 RuvB-like protein 2 v1.06890 K15080 Single-strand annealing weakened protein 1 v1.17193; v1.14087 K03111 Single-strand DNA-binding protein v1.15575 K10800 Single-strand monofunctional uracil DNA glycosylase v1.07134 K10992 Swi5-dependent recombination DNA repair protein 1 v1.13860 K03649 TDG/mug DNA glycosylase family protein v1.14423; v1.14399; v1.05070 K03648 Uracil-DNA glycosylase v1.23838 K10791 Three prime repair exonuclease 2 v1.19522 K10839 UV excision repair protein RAD23 Stress response proteins Annotation of the A. digitifera genome reveals a wide assortment of thermal shock proteins, molecular chaperones and other stress response elements that are given in (Table 10 ), excluding antioxidant and redox-protective proteins which are described in the next section. Heat shock proteins 70 kDa, 90 kDa, 110kDA, HspQ and HspX (the last two proteins being homologues of the bacterial heat shock factor sigma32 and α-crystallin, respectively) are encoded in the coral genome, together with several HSP gene transcription factors. HSPs play a role in various cellular functioning such as protein folding, intracellular protein trafficking and resistance to protein denaturation. HSP expression is usually increased on exposure to elevated temperatures and other conditions of biotic and abiotic stress that include infection, inflammation, metabolic hyperactivity, exposure to environmental toxicants, ultraviolet light exposure, starvation, hypoxia and desiccation [ 243 ]. HSPs and chaperones are transcriptionally regulated and are induced by heat shock transcription factors [ 244 ], of which there are several encoded in the coral genome. Since HSPs are found in virtually all living organisms, it is not surprising that cnidarian hsp transcription and protein expression (HSP60, HSP70 and HSP90) have been profiled as a stress determinant [ 245 - 250 ] and early warning indicator of coral bleaching [ 251 - 254 ]. The coral genome reveals also a cold shock protein encoded by the cspA gene family, but profiling its expression with other stress response proteins activated by sub-optimum cold temperatures [ 255 ] has not been reported. Additionally, the coral genome encodes transcription of a homologue of the universal stress protein A (UspA), a member of an ancient and conserved group of stress-response proteins [ 256 , 257 ], which have been studied mostly in bacteria [ 258 ] but have been described also in several plants [ 259 ] and animals, including members of the Cnidaria [ 260 ]. Usp transcripts have been quantified in the thermal stress response of the coral Montastraea faveolata [ 261 ] and its aposymbiotic embryos [ 262 ]. Another gene product of potential interest is a homologue of the oxidative-stress responsive protein 1 (OXSR1) that belongs to the Ser/Thr kinase family of proteins, as do other mitogen-stress activated protein kinases (MAPKs), that regulate downstream kinases in response to environmental stress [ 263 ] by interacting with the Hsp70 subfamily of proteins [ 264 ]. Another significant response protein encoded in the coral genome (Table 10 ) is a homologue of the stress-induced phosphoprotein 1 (30 domain sequence alignments), known also as the Hsp70-Hsp90 organising protein (HOP) belonging to the stress inducible (STI1) family of proteins, which is a principle adaptor protein that mediates the functional cooperation of molecular chaperones Hsp70 and Hsp90 [ 265 , 266 ]. It is yet to be determined if Hop1 transcription may serve as a primary indicator of environmental stress in corals. Table 10 Stress response proteins in the predicted proteome of A. digitifera Gene sequence KEGG Orthology Encoded protein description v1.04616; v1.06277 K03694 ATP-dependent Clp protease subunit ClpA v1.04617; v1.23486; v1.23484; v1.10207 K03695 ATP-dependent Clp protease subunit ClpB v1.13464 K03697 ATP-dependent Clp protease subunit ClpE v1.06903; v1.11461 K06891 ATP-dependent Clp protease adaptor protein ClpS v1.12577; v1.09531; v1.17184 K03544 ATP-dependent Clp protease subunit ClpX v1.09407 K08054 Calnexin (protein-folding chaperone) v1.16781 K08057 Calreticulin (Ca 2+ -binding chaperone) v1.04005 K10098 Calreticulin 3 (Ca 2+ -binding chaperone) v1.02702[+ 5 other sequence copies] K03704 Cold shock protein (beta-ribbon, CspA family) v1.01907; v1.18998 K07213 Copper chaperone v1.23457; v1.01713; v1.19228 K04569 Copper chaperone for superoxide dismutase v1.08719; v1.19128 K09502 DnaJ homologue subfamily A member 1 v1.08719; v1.18432 K09503 DnaJ homologue subfamily A member 2 v1.16210; v1.22054 K09504 DnaJ homologue subfamily A member 3 v1.19128 K09505 DnaJ homologue subfamily A member 4 v1.04818 [+ 6 other sequence copies] K09506 DnaJ homologue subfamily A member 5 v1.02841; v1.02842 K09507 DnaJ homologue subfamily B member 1 v1.00368; v1.13308; v1.16977; v1.03340 K09508 DnaJ homologue subfamily B member 2 v1.11537; v1.09205; v1.08628; v1.02840 K09511 DnaJ homologue subfamily B member 5 v1.24549 [+ 9 other sequence copies] K09512 DnaJ homologue subfamily B member 6 v1.01573 K09513 DnaJ homologue subfamily B member 7 v1.00352; v1.09196; v1.06645 K09514 DnaJ homologue subfamily B member 8 v1.18536 [+ 4 other sequence copies] K09515 DnaJ homologue subfamily B member 9 v1.14710 K09517 DnaJ homologue subfamily B member 11 v1.14959 K09518 DnaJ homologue subfamily B member 12 v1.09205 K09519 DnaJ homologue subfamily B member 13 v1.16242 K09520 DnaJ homologue subfamily B member 14 v1.20109; v1.03468 K09521 DnaJ homologue subfamily C member 1 v1.07111 [+ 5 other sequence copies] K09522 DnaJ homologue subfamily C member 2 v1.21077 [+ 13 other sequence copies] K09523 DnaJ homologue subfamily C member 3 v1.07739; v1.22910 K09524 DnaJ homologue subfamily C member 4 v1.01239 [+ 13 other sequence copies] K09525 DnaJ homologue subfamily C member 5 v1.17629 [+ 29 other sequence copies] K09527 DnaJ homologue subfamily C member 7 v1.18619; v1.08300; v1.23789 K09528 DnaJ homologue subfamily C member 8 v1.13575; v1.04213 K09529 DnaJ homologue subfamily C member 9 v1.05956; v1.05955; v1.21265; v1.21205 K09530 DnaJ homologue subfamily C member 10 v1.13525; v1.04120 K09531 DnaJ homologue subfamily C member 11 v1.09496 [+ 4 other sequence copies] K09533 DnaJ homologue subfamily C member 13 v1.24546 K09534 DnaJ homologue subfamily C member 14 v1.05866 K09536 DnaJ homologue subfamily C member 16 v1.16151; v1.08307; v1.14980 K09537 DnaJ homologue subfamily C member 17 v1.16309 K09539 DnaJ homologue subfamily C member 19 v1.05241; v1.22999; v1.17372 K14258 Facilitated trehalose transporter (anhydrobiosis) v1.12967; v1.19789 K14590 FtsJ methyltransferase [heat shock protein] v1.02247 K09414 Heat shock transcription factor 1 v1.24112 K09416 Heat shock transcription factor 3 v1.05839 K09419 Heat shock transcription factor, other eukaryote v1.12890 [+ 10 other sequence copies] K03283 Heat shock 70 kDa protein 1/8 v1.07996 K09489 Heat shock 70 kDa protein 4 v1.02854; v1.07452; v1.01623 K09490 Heat shock 70 kDa protein 5 v1.14149; v1.14150 K09487 Heat shock protein 90 kDa beta v1.07995; v1.07996; v1.16399; v1.11283 K09485 Heat shock protein 110 kDa v1.08943; v1.05577 K11940 Heat shock protein HspQ v1.00537; v1.00043 K03799 Heat shock protein HtpX v1.01623 K04046 Hypothetical chaperone protein v1.16216 K08268 Hypoxia-inducible factor 1 alpha v1.08869; v1.15120 K09097 Hypoxia-inducible factor 1 beta v1.22724 K09095 Hypoxia-inducible factor 2 alpha v1.23698 [+ 16 other sequence copies] K06711 Hypoxia-inducible factor prolyl 4-hydroxylase v1.16737; v1.22345 K09486 Hypoxia up-regulated 1 (heat shock protein 70 family) v1.10188 K08900 Mitochondrial chaperone BCS1 v1.17197; v1.04394 K04445 Mitogen-stress activated protein kinases v1.16301; v1.21224; v1.19344 K04043 Molecular chaperone DnaK v1.09682; v1.16748; v1.07471; v1.13624 K03687 Molecular chaperone GrpE v1.01621; v1.04945; v1.15919 K04044 Molecular chaperone HscA v1.18210 K04083 Molecular chaperone Hsp33 v1.17478; v1.16977; v1.10289; v1.19907 K04079 Molecular chaperone HtpG v1.08895; v1.18099 K11416 Mono-ADP-ribosyltransferase sirtuin 6 v1.02024 K11411 NAD-dependent deacetylase sirtuin 1 v1.04813 K11412 NAD-dependent deacetylase sirtuin 2 v1.22049; v1.22211; v1.02221 K11413 NAD-dependent deacetylase sirtuin 3 v1.11849; v1.02221 K11414 NAD-dependent deacetylase sirtuin 4 v1.05495 K11415 NAD-dependent deacetylase sirtuin 5 v1.04868 K11417 NAD-dependent deacetylase sirtuin 7 v1.15070 [+ 4 other sequence copies] K08835 Oxidative-stress responsive protein 1 (OXSR1) v1.04503 K11875 Proteasome assembly chaperone 1 v1.01531 K11878 Proteasome assembly chaperone 4 v1.01210 K11879 Proteasome chaperone 1 v1.18611 K11880 Proteasome chaperone 2 v1.00599 [+ 29 other sequence copies] K09553 Stress-induced-phosphoprotein 1 (HOP1) v1.08830 K13057 Trehalose synthase (anhydrobiosis) v1.22042 K03533 TorA specific chaperone v1.16986 [+ 7 other sequence copies] K06149 Universal stress protein A Molecular chaperones are a diverse family of proteins expressed by both prokaryotic and eukaryotic organisms that serve to maintain correct protein folding in a 3-dimensional functional state, assist in multiprotein complex assembly and protect proteins from irreversible aggregation at synthesis and during conditions of cellular stress [ 267 ]. Additionally, heat shock proteins and their co-chaperones may regulate cell death pathways by inhibition of apoptosis [ 268 ]. The coral genome encodes a large number of DnaJ subfamily (J-domain) chaperones (Hsp40) that with co-chaperone GrpE (Table 10 ) regulates the ATPase activity of Hsp70 (DnaK in bacteria) to enable correct protein folding [ 269 ]. The coral genome encodes homologues of the molecular chaperones HscA (specialised Hsp70), the redox-regulated chaperone Hsp33, HtpG (high temperature protein G), members of the calnexin/calreticulin chaperone system of the endoplasmic reticulum, a mitochondrial chaperone BCS1 protein necessary for the assembly of the respiratory chain complex III and a specific chaperone of trimethyl N-oxide reductase (TorA). The coral genome also encodes hypoxia-inducible factors (HIFs) that moderate the deleterious effects of hypoxia on cellular metabolism (reviewed in [ 270 ]). In the HIF signalling cascade, the alpha subunits of HIF are hydroxylated at conserved proline residues by HIF prolyl-hydroxylases allowing their recognition for pro-teasomal degradation, which occurs during normoxic conditions but is repressed by oxygen depletion. Hypoxia-stabilised HIF1 upregulates the expression of enzymes principally of the oxygen-independent glycolysis pathway, and in higher animals promotes vascularisation, whereas the mammalian HIF2 paralogue regulates erythropoietin control of hepatic erythrocyte production in response to hypoxic stress [ 271 ]. The roles of HIF1 and HIF2 homologues in corals have been established, with HIF1 regulation of glycolysis critical to metabolic function during the dark diurnal anoxic state of coral respiration [ 193 , 272 ]. Heat shock proteins that repair unfolded or misfolded protein have a complementary function to the ubiquitin-proteasome system (ubiquitins not tabulated) that selects damaged protein for degradation [ 273 ], such that HSP chaperones and the proteasome act jointly to preserve cellular proteostasis [ 274 , 275 ]. Thus, several proteasome chaperones and assembly chaperones are encoded in the A. digitifera genome (Table 10 ). While proteasome chaperones serve to target aberrant proteins for ubiquination, the proteasome chaperones facilitates 20S assembly for biogenesis of the multiunit 26S proteasome that is activated in response to stress [ 276 , 277 ], possibly by FtsJ (aka RrmJ), a well-conserved heat shock protein having novel ribosomal methyltransferase activity that targets methylation of 26S rRNA under heat shock control [ 278 , 279 ]. The HspQ protein encoded in the coral genome, although studied almost exclusively in bacteria, is known to stimulate degradation of denatured proteins caused by hyperthermal stress, particularly DnaA that initiates DNA replication in prokaryotes [ 280 ]. Specifically, HspQ (heat shock factor sigma32) regulates the expression of Clp ATPase-dependent protease family enzymes [ 281 , 282 ], of which ClpA, ClpB, ClipE, the protease adaptor protein ClpS [ 283 ] and the unfoldase ClpX protein [ 284 ] are encoded in the coral genome (Table 10 ). HspX is a small 16 kDa α-crystallin chaperone (Acr) protein belonging to the Hsp20 family of proteins [ 285 ] that suppresses thermal denaturation and aggregation of proteins [ 285 ]. Significantly, Acr proteins are known to bind with carbonic anhydrase [ 286 ] and may have importance in moderating stress-induced loss of calcium deposition. Thus, HspX/Acr expression may account for differences in the thermal sensitivity of corals to calcification that varies among genera [ 287 ]. In a different context, HspX is attracting considerable attention for its potential to elicit long-term protective immunity against human Mycobacterium tuberculosis infection by chaperoning a host-protective antigen [ 288 ] that by extension, but yet untested, may likewise repress virulence in the initiation and progression of microbial coral disease [ 289 , 290 ]. The coral genome encodes complete membership of the human sirtuin (SIRT1-7) family of NAD(+)-dependent protein deacetylases and ADP-ribosyltransferases. Mammalian SIRT1 (a homologue of yeast Sir2) is an important regulator of metabolism, cell differentiation, stress response transcription and pathways of cellular senescence (reviewed in [ 291 ]). SIRT proteins regulate chromatin function through deacetylation of histones that promote subsequent alterations in the methylation of histones and DNA to affect, via deactivation of nuclear transcription factors and co-regulators, epigenetic control of nuclear transcription. As NAD + -dependent enzymes, SIRT1 can regulate gene expression in response to cellular NAD + /NADH redox status providing a metabolic template for epigenetic transcriptome reprogramming [ 292 , 293 ]. In the human genome repertoire, SIRT1 modulates cellular responses to hypoxia by deacetylation of HIF1α [ 294 ] and inhibits nitric oxide synthesis by suppression of the nuclear factor-kappaB (NF-κB) signalling pathway [ 295 ], SIRT2 promotes oxidative stress resistance by deacetylation of forkhead box O (FOXO) proteins [ 296 ], SIRT3 decreases ROS production in adipocytes [ 297 ], SIRT4 regulates fatty acid metabolism and stress-response elements of mitochondrial gene expression [ 298 ], SIRT5 is a protein lysine desuccinylase and demalonylase of unknown function [ 299 ], SIRT6 activates base-excision repair [ 300 ] and SIRT7 inhibits apoptosis induced by oxidative stress by deacylation of p53 [ 301 , 302 ]. The significance of coral SIRT proteins, by analogy, to exert stress tolerance is yet to be examined. Metallochaperones are an important class of enzymes that transport co-factor metal ions to specific proteins [ 303 ]. The copper chaperone protein ATX1 (human ATOX1) delivers cytosolic copper to Cu-ATPase proteins and serves as a metal homeostasis factor to prevent Fenton-type production of highly reactive hydroxyl radicals. ATX1, which is strongly induced by molecular oxygen, functions additionally as an antioxidant to protect cells against the toxicity of both the superoxide anion and hydrogen peroxide [ 304 ]. Encoded also is a specific copper chaperone essential to the activation of Cu/Zn superoxide dismutase [ 305 , 306 ] that is enhanced by photooxidative stress in scleractionian corals [ 307 ], although reported to be less pronounced in the host than in symbiotic algae [ 308 ]. In addition to high light exposure, reef-building corals of shallow reef flats are occasionally exposed to the atmosphere for periods that can last several hours during extreme low tides. Hence, species that are adapted to withstand acute desiccation (anhydrobiosis) have a better chance of surviving such conditions. The disaccharide trehalose is an osmolyte that in some plants and animals allows them to survive prolonged periods of desiccation [ 309 ]. The hydrated sugar has high water retention that forms a gel phase when cells dehydrate, which on rehydration allows normal cellular activity to resume without damage that would otherwise follow a dehydration/rehydration cycle. Furthermore, trehalose is highly effective in protecting enzymes in their native state from inactivation from thermal denaturation [ 310 ]. Given that A. digitifera is endemic on shallow reef flats prone to exposure at low tides [ 311 ], it is not surprising that the coral genome encodes trehalose synthase and a facilitated trehalose transporter for protection against dehydration. Antioxidant and redox-protective proteins Oxygen is vital for life, but it can also cause damage to cells, particularly at elevated levels. In coral symbiosis, the photosynthetic endosymbionts of corals typically produce more oxygen than the holobiont is able to consume by respiration, so that coral tissues are hyperoxic with tissue p O 2 levels often exceeding 250% of air saturation during daylight illumination [ 193 ]. Furthermore, because algal symbionts reside within the endodermal cells of their host, coral tissues must be transparent to facilitate the penetration of downwelling light required for photosynthesis by their algal consorts. In clear shallow waters this entails concurrent exposure to vulnerable molecular sites of both partners to damaging wavelengths of ultraviolet radiation. The synergistic effects of tissue hyperoxia and UV exposure can cause oxidative damage to the symbiosis via the photochemical production of cytotoxic oxygen species [ 312 ] that are produced also during normal mitochondrial function [ 313 ]. Consequently, protective proteins (antioxidant enzymes) are expressed to maintain the fine balance between oxygen metabolism and the production of potentially toxic reactive oxygen species (ROS). If this balance is not maintained by regulation of oxidative and reductive processes (redox regulation), oxidative stress occurs by the generation of excess ROS, causing damage to DNA, proteins, and lipids. Corals elaborate a variety of molecular defences that including the production of UV-protective sunscreens, (MAAS), antioxidants, antioxidant enzymes, chaperones and heat shock proteins, which are often inducible under conditions of enhanced oxidative stress [ 307 ], including conditions that elicit coral bleaching [ 314 , 315 ]. An excellent review on the formation of ROS and the role of antioxidants and antioxidant enzymes in the field of redox biology is given by Halliwell [ 316 ]. Annotation of the A. digitifera genome reveals sequences encoding two isoforms of the antioxidant enzyme superoxide dismutase (SOD) from both the Cu/Zn and Fe/Mn families of SOD (Table 11 ). These metalloprotein enzymes catalyse the dismutation of superoxide to yield molecular oxygen and hydrogen peroxide, the latter being less harmful than superoxide. Superoxide can oxidize proteins, denature enzymes, oxidize lipids and fragment DNA. By removing superoxide, SOD protects also against the production of reactive peroxynitrite formed by the combination of superoxide and nitric oxide, which is a precursor reactant for production of the supra-reactive hydroxyl radical. Hydrogen peroxide per se is a mild oxidant, but it readily oxidises free cellular ferrous iron to ferric iron with production of hydroxyl radicals via the Fenton reaction. Accordingly, both the removal of hydrogen peroxide and the expression of proteins, such as transferrin, (bacterio)ferritins and metallothioneins, that bind reactive (transition) metal ions is important to protect cellular components from acute oxidative damage. Oddly, only a metallothionein expression activator was found encoded in the coral genome without finding a sequence to activate transcription of the actual metallothionein protein gene. Table 11 Antioxidant and redox-protective proteins in the predicted proteome of A. digitifera Gene sequence KEGG Orthology Encoded protein description v1.10918 K04756 Alkyl hydroperoxide reductase subunit D v1.11551 K03387 Alkyl hydroperoxide reductase subunit F v1.07812 K03594 Bacterioferritin v1.21362 [+ 4 other sequence copies] K00429 Catalase (bacterial) v1.17525 [+ 4 other sequence copies] K03781 Catalase (peroxisonal) v1.23457; v1.01713; v1.19228 K04569 Copper chaperone for superoxide dismutase v1.20153; v1.20154 K10528 Hydroperoxide lyase v1.19687; v1.19688; v1.18796; v1.18795 K00522 Ferritin heavy chain v1.06441 K03674 Glutaredoxin 1 v1.19449 K03675 Glutaredoxin 2 v1.14929 [+ 5 other sequence copies] K03676 Glutaredoxin 3 v1.13285; v1.03722; v1.03688; v1.10496 K00432 Glutathione peroxidase v1.13174; v1.13775; v1.05473 K00383 Glutathione reductase (NADPH) v1.14344; v1.19399; v1.01421 K01920 Glutathione synthase v1.02173 K09238 Metallothionein expression activator v1.09719; v1.16134; v1.18608 K07390 Monothiol glutaredoxin v1.14890; v1.17685 K07305 Peptide-methionine (R)-S-oxide reductase v1.14909 K00435 Peroxiredoxin v1.14106 K13279 Peroxiredoxin 1 v1.08691 K11187 Peroxiredoxin 5, atypical 2-Cys peroxiredoxin v1.01410 K11188 Peroxiredoxin 6, 1-Cys peroxiredoxin v1.03688 K05361 Phospholipid-hydroperoxide glutathione peroxidase v1.05148 K05905 Protein-disulfide reductase v1.02922; v1.22772; v1.24164 K05360 Protein-disulfide reductase (glutathione) v1.06810 K12260 Sulfiredoxin v1.01713 [+ 4 other sequence copies] K04565 Superoxide dismutase, Cu/Zn family v1.09974; v1.20324 K04564 Superoxide dismutase, Fe/Mn family v1.02378 K11065 Thiol peroxidase, atypical 2-Cys peroxiredoxin v1.22324 [+ 7 other sequence copies] K03671 Thioredoxin 1 v1.05148; v1.03230; v1.20699 K03672 Thioredoxin 2 v1.17881 [+ 5 other sequence copies] K13984 Thioredoxin domain-containing protein 5 v1.04532; v1.24501 K09585 Thioredoxin domain-containing protein 10 v1.11551; v1.19049 K00384 Thioredoxin reductase (NADPH) v1.10930 K14736 Transferrin As expected from the foregoing, the genome of A. digitifera encodes the antioxidant enzyme catalase (CAT) that is highly efficient in decomposing hydrogen peroxide to yield molecular oxygen and water. Two isoforms of CAT are encoded at multiple sites. One is a peroxisomal eukaryotic CAT enzyme that targets the removal of hydrogen peroxide formed as a by-product of oxidase enzymes, and the other is a related catalase domain-containing protein presumed also to decompose hydrogen peroxide. Glutathione peroxidise (GPx) reduces both hydrogen peroxide and lipid hydroperoxides, the latter of which are formed by radical-induced lipid autoxidation. Phototrophic organisms, including higher plants, utilise ascorbate peroxidase (APx) as a primary catalyst for the reduction of hydrogen peroxide and lipid hydroperoxides. However, unlike the freshwater cnidarian H. viridis [ 164 ], there is no evidence for transfer of APx-encoding genes to A. digitifera . The antioxidant enzymes SOD, CAT, GPx and APx are well characterised in the algal and animal partners of coral symbiosis (reviewed in [ 317 ]). Additionally, the coral genome has sequences encoding alkyl hydroperoxide reductase, hydroperoxide lyase, phospholipid-hydroperoxide glutathione peroxidase, thiol peroxidase and multiple isoforms of peroxiredoxin, all of which function in the detoxification of organo-hydroperoxides that are produced as a by-product of aerobic metabolism. Additionally, sulfiredoxin (Table 11 ) repairs peroxiredoxins when these enzymes are inhibited by over-oxidation [ 318 ]. Thioredoxins and glutaredoxins have important secondary roles in regulating multiple pathways in many biological processes, including redox signalling of apoptotic pathways, which have been attributed to processes involved in coral bleaching [ 56 ]. Other enzymes that regulate cellular thiol-disulfide homeostasis in this coral are monothiol glutaredoxin and protein-disulfide reductase. The coral genome encodes the ubiquitous thioredoxin system of antioxidant proteins (Table 11 ) that act as electron donors to peroxidases and ribonucleotide reductase (the latter not tabulated). By cysteine thiol-disulfide exchange, thioredoxins function as a protein thiol-disulfide oxidoreductase [ 319 ]. In the thioredoxin system, thioredoxins are maintained in their reduced state by NADPH-dependent, flavoenzyme thioredoxin reductase [ 320 ]. Peptide-methionine (R)-S-oxide reductase can additionally rescue thioredoxin from oxidative inactivation by disulfide reduction. Related glutaredoxins share many of the functions of thioredoxins but are reduced directly by glutathione, rather than by a specific reducing enzyme, while in turn glutathione is kept in its native state by NADPH: glutathione reductase. In recent years there has been a particular focus on the role of ROS in coral bleaching, fuelled by dire prediction of future catastrophic episodes caused by environmental change affected by global warming [ 321 ]. Early predictions of coral bleaching were based principally on physical environmental parameters, rather than on the determination of the physiological state of coral populations to such conditions. While gene expression markers are being developed to monitor sub-bleaching levels of stress in situ (e.g., [ 261 ]), Kenkel et al. [ 322 ] opined that the current challenge for implementing expression-based methods lies in identifying coral genes demonstrating the most pronounced and consistent stress response, preferably with a large dynamic range to enable reliable quantification. To this end, we offer in Table 11 the annotation of novel redox-related genes for examination as potential candidate biomarkers to monitor the physiological response of A. digitifera to environmental stress. Proteins of cellular apoptosis Apoptosis is the signalling of programmed cell death (PCD) that occurs in multicellular organisms in response to cellular injury. A key feature of apoptosis is the activation of endogenous endonucleases causing nuclear fragmentation, chromatin condensation and chromosomal DNA fragmentation, which typically presents in affected cells by the morphological appearance of plasma membrane blebbing and cell shrinkage. Caspases and related family member proteases are described as “executioners” of apoptosis that on post-translational activation degrade the regulatory proteins that prevent DNA degradation. Fragmentation of nuclear DNA is one of the hallmarks of apoptotic cell death that occurs by PCD stimuli in a wide variety of proliferating cells. NF-κB is a protein complex that controls the transcription of DNA that can induce the expression of nitric oxide synthesis (NOS) to produce NO that is a well-known promoter of the of the pro-apoptotic transcription factor p53 cell-cycle gatekeeper of the caspase cascade. In contrast to necrosis, which is the outcome of PCD, apoptosis mediates the fragmentation of damaged cells, which by phagocytosis are removed or degraded in phagolysosomes to spare surviving cells from the uncontrolled release of cytotoxic agents. Proteins of the caspase-mediated apoptotic cascade are regarded as products of constituent housekeeping genes that are necessary to maintain healthy multicellular function [ 323 ]. In the progression of cnidarian bleaching, apoptotic pathways are activated [ 322 - 325 ], but not all corals that suffer bleaching are destined to die [ 326 , 327 ]. Coral survival has been attributed to having a high level of apoptotic protection at the onset of coral bleaching [ 328 ] and during post-bleaching recovery [ 329 ] by specific activation of anti-apoptotic Bcl-2 proteins in surviving cells [ 330 ]. Cnidarians have a complex apoptotic protein network that has exceptional ancestral complexity and is comparable to that of higher vertebrates [ 331 , 332 ]. Cnidarian metamorphosis is tightly coupled with caspase-dependent apoptosis [ 333 ] and subsequent host-symbiont selection by post-phagocytic winnowing of Symbiodinium genotypes during the establishment of coral-dinoflagellate mutualism [ 334 ]. As expected, the coral genome of A. digitifera encodes multiple isoforms of genes that transcribe the caspase family of apoptotic effectors (Table 12 ). Included in this signalling pathway are the pro- and anti-apoptotic Bax/Bcl regulators and Bcl-2 athanogene (DNA-binding) activators of apoptosis. Notable in our annotation dataset are multiple genes that encode the protein domains of the apoptotic protease-activating factor (Apaf) that triggers assembly of the apoptosome leading to caspase activation [ 335 ]. Additional to this arsenal of cell cycle regulators are the death associated protein-6 (DAXX), a Fas-binding adaptor of c-Jun N-terminal kinase (JNK) activation [ 336 ], death-associated protein kinase (DAPK), a mediator of calcium/calmodulin-regulated Ser/Thr kinase [ 337 ], and the programmed cell death 6-interacting protein (PDCD6IP), which binds to PDCD-6 for execution of apoptosis via the caspase-3 pathway [ 338 ]. PDCD6IP activation of apoptosis is an enigma since PDCD-6 is not encoded in the coral genome, nor is caspase-3. Other cell cycle regulators are the p53 binding and p53-associated parkin-like proteins, and the activating TP53 regulating kinase protein and TP53 apoptosis effector of TP53 gene expression. Table 12 Proteins of cellular apoptosis in the predicted proteome of A. digitifera Gene sequence KEGG Orthology Encoded protein description v1.17521; v1.02505; v1.20702; v1.05077 K02159 Apoptosis regulator BAX (BCL2-associated) v1.05086; v1.20659 K02161 Apoptosis regulator BCL-2 v1.17522; v1.00181; v1.10817; v1.20703 K02163 Apoptosis regulator BCL-W v1.05147 [+ 6 other sequence copies] K12875 Apoptotic chromatin condensation inducer v1.22264 [+ 72 other sequence copies] K02084 Apoptotic protease-activating factor (Apaf) v1.17326; v1.20305; v1.11586 K09555 BCL2-associated athanogene 1 v1.08601 K09558 BCL2-associated athanogene 4 v1.02839 K09559 BCL2-associated athanogene 5 v1.01518 K13087 BCL2-associated transcription factor 1 v1.20278; v1.00172; v1.07858 K14021 BCL-2 homologueous antagonist/killer v1.09624 K02561 BCL2-related (ovarian) killer protein v1.17749 K08573 Calpain-3 v1.00595; v1.14671; v1.00040 K08574 Calpain-5 v1.00040 K08575 Calpain-6 v1.19153; v1.17749 K08576 Calpain-7 v1.15226 K04740 Calpain-12 v1.02951 K08582 Calpain-15 v1.11167; v1.06681; v1.20230; v1.01376 K08585 Calpain, invertebrate v1.0312 7 [+ 6 other sequence copies] K08583 Calpain, small subunit 1 v1.17229; v1.00023; v1.09976 K02186 Caspase 2 v1.11989 [+ 5 other sequence copies] K04397 Caspase 7 v1.02756 [+ 27 other sequence copies] K04398 Caspase 8 v1.01818 K04399 Caspase 9 v1.00817 [+ 4 other sequence copies] K04400 Caspase 10 v1.02005 K04741 Caspase 12 v1.00818 [+ 11 other sequence copies] K04489 Caspase apoptosis-related cysteine protease v1.13260 K07367 Caspase recruitment domain-containing protein 11 v1.06297 [+ 44 other sequence copies] K02832 CASP2 and RIPK1 adaptor with death domain v1.21531 K02308 Death-associated protein 6 (DAXX) v1.09448; v1.15529; v1.20164 K08803 Death-associated protein kinase (DAPK) v1.23110; v1.14222; v1.03658 K12366 Engulfment and motility protein 1 (phagocytosis/apoptosis) v1.18448 [+ 78 other sequence copies] K02373 Fas (TNFRSF6)-associated via death domain (FADD) v1.24288 [+ 66 other sequence copies] K10130 Leucine-rich repeats and death domain-containing protein v1.20620 K04734 NF-kappa-B inhibitor alpha v1.01706 K14214 NF-kappa-B inhibitor delta v1.10378; v1.10729; 1.05609; v1.05609 K05872 NF-kappa-B inhibitor epsilon v1.17893; v1.22419; v1.00700; v1.08415 K09256 NF-kappa-B inhibitor-like protein 1 v1.04158 [+ 211 other sequence copies] K09257 NF-kappa-B inhibitor-like protein 2 v1.05320; v1.06979; v1.04467; v1.21371 K02580 Nuclear factor NF-kappa-B p105 subunit v1.20334; v1.22743 K11970 p53-Associated parkin-like cytoplasmic protein v1.14920; v1.11864; v1.15271; v1.11865 K06643 p53-Binding protein v1.04289 K06708 Programmed cell death 1 ligand 2 v1.05882 [+ 7 other sequence copies] K12200 Programmed cell death 6-interacting protein (PDCD6!P) v1.10959; v1.04994 K04727 Programmed cell death 8 apoptosis-inducing factor v1.16714 K06875 Programmed cell death protein 5 (PDCD-5) v1.13112 K03171 Tnfrsf1a-associated via death domain v1.24655; v1.12385 K10136 TP53 apoptosis effector v1.09087 K08851 TP53 regulating kinase v1.05030; v1.07044 K11859 Tumor necrosis factor, alpha-induced protein 3 v1.22799 K04389 Tumor necrosis factor ligand superfamily member 6 v1.05776 K05470 Tumor necrosis factor ligand superfamily member 7 v1.13754 K05472 Tumor necrosis factor ligand superfamily member 9 v1.21776 [+ 6 other sequence copies] K04721 Tumor necrosis factor ligand superfamily member 10 v1.04001 K05473 Tumor necrosis factor ligand superfamily member 11 v1.19776 K05474 Tumor necrosis factor ligand superfamily member 12 v1.09015; v1.14041 K03158 Tumor necrosis factor receptor superfamily member 1A v1.07010 K05141 Tumor necrosis factor receptor superfamily member 1B v1.19735 K05142 Tumor necrosis factor receptor superfamily member 4 v1.13754 K03160 Tumor necrosis factor receptor superfamily member 5 v1.22577 K05143 Tumor necrosis factor receptor superfamily member 6B v1.20003 K05144 Tumor necrosis factor receptor superfamily member 7 v1.23750; v1.17970; v1.19022 K05146 Tumor necrosis factor receptor superfamily member 9 v1.07527 K05148 Tumor necrosis factor receptor superfamily member 11B v1.10221 K05151 Tumor necrosis factor receptor superfamily member 13C v1.14826; v1.01054 K05152 Tumor necrosis factor receptor superfamily member 14 v1.09514 K05156 Tumor necrosis factor receptor superfamily member 19 v1.01640 K05161 Tumor necrosis factor receptor superfamily member 26 v1.08207; v1.16237; v1.14824 K10133 Tumor protein p53-inducible protein 3 Our genome annotation reveals 73 sequence matches for expressing the Apaf protein domain that, in conjunction with a high copy number for expressing caspase-8 (28 protein sequence matches), may enhance coral survival during embryogenesis by suppressing receptor-induced protein kinase (45 sequence matches) during early development [ 339 ]. The most conserved function of the CAPS2/RIPK adaptor (45 sequence matches) encoded in the coral genome is its essential regulation of apoptosis [ 340 ]. We find a wide repertoire of genes that additionally encode proteins that mediate apoptosis (Table 12 ). Amongst these are the calpain Ca +2 -sensing family of proteins that initiate the signalling of apoptotic pathways [ 341 ]. There are 79 matches to sequences that encode the tumor necrosis Fas superfamily member 6 (TNFRSF6) receptor, which coupled with the death domain (FADD) protein is a cell signalling mediator for recruitment of caspase-8 that activates the apoptotic cysteine protease cascade. Coincident in the genome are 67 sequences encoding the leucine-rich repeat and death domain-containing (LRDD) adaptor that, by interacting with other p53-inducible death domain-containing (PIDD) proteins such as FADD, induces the caspase-2 pathway of apoptosis in response to DNA damage [ 342 ]. Elements of the NF-κB signalling pathway of cnidarians are highly conserved traits [ 343 ], which includes the caspase cascade and the pro-apoptotic and anti-apoptotic Bcl-2 family of proteins [ 344 ]. The coral genome of A. digitifera encodes the pleiotropic nuclear factor NF-κB p105 subunit, and astonishingly there are 212 sequence matches to the NF-κB inhibitor-like protein 2 domain with fewer matches to the NF-κB inhibitor-like protein 1 and NF-κB family inhibitors alpha, delta and epsilon. Evident in our genome annotation is the tumor necrosis factor-alpha induced protein 3 (TNFAIP3), a cytokine produced by activated (inflammatory) macrophages. Although TNF cytokines are a major extrinsic mediator of cellular apoptotic pathways, the precise function of the superfamily members of TNF ligands and receptors (Table 12 ) remains elusive in coral symbiology. Microbial symbiosis and pathogenicity It is well established that corals associate with a vast consortia of microbes, including phototrophic symbionts ( Symbiodinium spp.) and other eukaryotic microbionts, cyanophytes, heterotrophic bacteria, archaea and viruses [ 345 ]. Corals harbour diverse and abundant prokaryotic communities with distinct populations residing in separate habitats of the host skeleton, tissues and surface mucus layer (reviewed in [ 203 ]). Microbial populations are dominated by a few coral-specific taxonomic traits [ 346 ], but the majority of the population comprises a high number of taxonomically diverse, low-abundance ribotypes [ 347 ] with much of the diversity within the coral microbiome belonging to the “rare” biosphere [ 348 , 349 ]. The coral microbiome is vital to the nutrition and health of the holobiont [ 350 ] and contributes significantly to the protection of coral reef ecosystems against the detrimental effects of organic enrichment [ 351 , 352 ]. One emerging threat to coral reefs is the outbreak of infectious diseases (reviewed in [ 353 ]). Although highly subjective and with little experimental evidence to date, the coral probiotic hypothesis [ 354 ] suggests that the coral prokaryotic microbiome can adapt to changing environmental conditions by selective microbial reorganisation to impart greater resistance to disease and pathogen-mediated bleaching [ 355 ]. Whether the coral microbiome can respond to changing environmental conditions more rapidly than by host genetic mutation and selection based on contemporary phenotypic evolution on ecological time-scales [ 356 ], is a topic of current debate [ 357 ]. Corals, like other invertebrates, have an innate immune system based on self-histocompatibility recognition (reviewed in [ 358 ]), but to date few adaptive components have been identified [ 359 ]. Corals do not produce antibodies and thus lack a true adaptive immune system. Nonetheless, corals once susceptible to infection and bleaching caused by a specific bacterial agent can become immune to the invading pathogen by a phenomenon termed “experience-mediated tolerance”, a precept of the hologenome theory of evolution [ 360 ], although how this process occurs is largely unknown. In our annotation of the genome sequence of A. digitifera we uncovered genes encoding the expression of disease resistance proteins (Table 13 ), two of which match the plant RPM1 and RPS2 pathogen resistance proteins that guard against disease by binding with pathogen avirulence receptors [ 360 , 361 ]. Significant also is a gene to express the pathogenesis-related protein PR-1 (29 sequence domain matches) that is inducible in plants for systemic acquired resistance to pathogenic invasion [ 362 ]. We uncovered also multiple genes encoding the expression of myeloperoxidase (MPO) enzymes. MPOs produce hypochlorous acid from hydrogen peroxide and chloride ion (requiring heme as a cofactor), and it oxidizes tyrosine to the tyrosyl radical using hydrogen peroxide as an oxidizing agent. Hypochlorous acid and tyrosyl radicals are strong cytotoxic agents that in higher organisms are used as a primary defence by neutrophils to protect against invading pathogens. Phenoloxidase (tyrosinase) activity is reported to contribute to the innate defence system of A. millepora and Porites sp. [ 363 ] via activation of the melanin-signalling pathway that is induced in response to coral bleaching and localised disease [ 364 , 365 ]. Three genes of A. digitifera encode tyrosinase enzymes (data not tabulated) to account for the phenoloxidase activity reported in corals. Table 13 Microbial symbiosis and pathogenicity proteins in the predicted proteome of A. digitifera Gene sequence KEGG Orthology Encoded protein description v1.06126 K13061 Acyl homoserine lactone synthase v1.19990 K01372 Bleomycin hydrolase v1.00209; v1.06178 K03587 Cell division protein FtsI (penicillin-binding protein 3) v1.18860 K13458 Disease resistance protein v1.16231; v1.00374; v1.08191 K13457 Disease resistance protein RPM1 v1.13482 [+ 4 other sequence copies] K13459 Disease resistance protein RPS2 v1.07889 K12090 Cag pathogenicity island protein 5 v1.24345 K12091 Cag pathogenicity island protein 6 v1.18924; v1.17622 K12093 Cag pathogenicity island protein 8 v1.05278 K12096 Cag pathogenicity island protein 11 v1.02083 K12104 Cag pathogenicity island protein 19 v1.12907 K12109 Cag pathogenicity island protein 24 v1.00209; v1.06178 K03587 Cell division protein FtsI (penicillin-binding protein 3) v1.13874 K07259 Carboxy/endopeptidase (penicillin-binding protein 4) v1.12514; v1.09758 K04127 Isopenicillin-N epimerase v1.21332 K04126 Isopenicillin-N synthase v1.07742 K02547 Methicillin resistance protein v1.17478; v1.16977; v1.10289; v1.19907 K04079 Molecular chaperone HtpG (anti-bacterial) v1.08255 K13651 Motility quorum-sensing regulator, GCU-specific toxin v1.14792 [+ 7 other sequence copies] K10789 Myeloperoxidase v1.02333 [+ 26 other sequence copies] K13449 Pathogenesis-related protein 1 v1.05017 K03693 Penicillin-binding protein v1.17507 K12556 Penicillin-binding protein 2X v1.13874 K07259 Penicillin-binding protein 4 v1.16655 K02171 Penicillinase repressor v1.14688 K15126 Type III secretion system cytotoxic effector protein v1.20647 K03980 Virulence factor, integral membrane protein v1.18964 K03810 Virulence factor, oxidoreductase domain The genome of A. digitifera also reveals homologues of genes that promote bacterial pathogenicity (Table 13 ), including virulence factors that are expressed and excreted by invading pathogens (bacteria, viruses, fungi and protozoa) to inhibit certain protective functions of the host. Such are the bacterial Type III cytotoxic effector protein and multiple Type IV Cag pathogenicity island proteins encoded in the coral genome. Many Gram-negative bacteria utilize Type III secretion proteins, which are regulated by quorum sensing, to deliver cytotoxic effector proteins into eukaryote host cells during infection. Cag (cytotoxin-associated) pathogenicity island (PAI) proteins are encoded by mobile genetic elements of the Type IV system secreting both proteins and large nucleoprotein complexes [ 366 ] that may be transferred between prokaryotes to enhance selected traits of virulence [ 367 ]. Our annotation reveals genes encoding six pathogenicity island proteins (Table 13 ) with similarity to the Cag PAI proteins of the human Heliobacter pylori , an infectious bacterium causing peptic ulcers that may lead to the development of stomach cancer. While many properties of Type III and IV secretion system proteins have been well characterized in bacteria, the functional purpose of homologous genes in A. digitifera , if expressed, are unknown. The genome of A. digitifera contains genes of bacterial origin that encode the motility quorum-sensing regulator of the GCU-specific mRNA interferase toxin and acyl homoserine lactone synthesis used for the communication of quorum sensing between bacteria to enable the coordination of group behaviour based on collective population density. Apparent in our annotation (Table 13 ) is a wide array of microbial penicillin-binding proteins (PBPs) that have an affinity for β-lactam antibiotics that by binding to PBPs prevent bacteria from constructing a cell wall. There are genes also to enhance antibiotic resistance, including potential expression of a penicillinase repressor, a methicillin resistance protein and bleomycin hydrolase (cysteine peptidase). Additionally, isopenicillin-N synthase and an isopenicillin-N epimerase, both of which catalyse key steps in the biosynthesis of penicillin and cephalosporin antibiotics, are encoded in the coral genome. Taken as a whole, we demonstrate an extensive presence of ancient non-metazoan genes that are maintained in the genome of A. digitifera , as is reported in the genomes of A. millepora and the anemone N. vectensis [ 368 ]. Recent thought on genome evolution places these ancestral conserved domains as ‘orphan’ or ‘taxonomically restricted’ genes [ 352 , 369 , 370 ], rather than acquired later by horizontal gene transfer. There is, of course, little knowledge of how or when, if at all, these non-metazoan genes are expressed or even their function to mediate pathogenicity in the coral holobiont. Proteins of viral pathogenicity Marine viruses were of minor interest until 1989, when it was realised that virus-like particles (VLPs) are the most abundant biological entities to occupy aquatic environments with variable numbers reaching ~10 8 VLPs ml -1 [ 371 ]. Typically, VLPs surpass the number of marine bacteria by an order of magnitude in coastal waters [ 372 ]; their diversity is extremely high and many are specific to the marine environment [ 373 , 374 ]. Significant VLP numbers are reported from the surrounding waters of oceanic coral reef atolls [ 375 ], in waters flowing across the reef substratum [ 376 ] and in samples taken within the close vicinity of coral colonies [ 377 , 378 ]. The viral load within the surface microlayer of scleractinian corals is enumerated as being 10 7 -10 8 VLPs mL -1 [ 379 ] and, based on VLP morphological diversity, is attributed to infecting various microbial hosts (bacteria, archaea, cyanobacteria, fungi and algae) residing within the coral mucus [ 380 ]. VLPs have been observed in the epidermal and gastrodermal tissues of corals and occasionally occur in the mesogloea [ 381 ]. Latent viruses were found to infect Symbiodinium isolated from several scleractinian corals [ 382 - 384 ] with a preponderance of eukaryotic algae-infecting phycodnaviruses suggested [ 385 ]. A wide range of bacteriophage and eukaryotic virus families have been identified within scleractinians using metagenomic analyses [ 207 , 386 - 388 ], with bacteriophages being by far the most abundant entities (Wood-Charlson EM, Weynberg KD, Suttle CA, Roux S, van Oppen MJH: Methodological biases in coral viromics , submitted). The importance of the coral-virus interactome in bleaching and disease (reviewed in [ 185 , 389 ]) is founded on reports showing that VLP abundances are higher in the seawater immediately surrounding diseased compared to that of healthy corals [ 378 ], that latent viruses are induced by heat stress in symbiotic dinoflagellates of the sea anemone Anemonia virdis [ 382 ] and the coral Pavona danai [ 383 ], and that UV exposure induces a latent virus-like infection in cultured Symbiodinium [ 187 ]. Quantitative 454 pyrosequence analysis of the coral Porites compressa on exposure to reduced pH, elevated nutrients or thermal stress showed that the abundance of its viral consortia varied across treatments, but notably a novel herpes-like virus increased by up to 6 orders of magnitude on exposure to abiotic stress [ 387 ], although some caution may be warranted in assessing the reliability of such determinations [Wood-Charlson et al. , submitted]. Unexpectedly, the proteome of an endosymbiont-enriched fraction of the coral Stylophora pistillata showed a significant 114-fold increase in a viral replication protein on thermal bleaching [ 39 ], which is consistent with the finding of VLP induction in P. compressa by similar treatment [ 387 ]. General aspects of histocompatibility [ 390 - 393 ] and the genetic structure of innate immune receptors of the Cnidaria [ 363 , 394 - 401 ], including the immune response effected by coral disease and bleaching [ 364 , 402 ], have been examined extensively, hence further elaboration here is unnecessary. Instead, we focus on proteins that directly regulate the pathogenicity of coral-associated microbes and viruses. The A. digitifera genome encodes protein homologues having either putative antiviral and virus-promoting activities (Table 14 ). These homologues include the antiviral “superkiller” helicase SKI2 protein that acts by blocking viral mRNA translation [ 403 ] and, together with the superkiller proteins SKI3 (69 sequence alignments) and SKI8 of the exosome complex, function in a 3′-mRNA degradation pathway [ 404 ]. The coral genome encodes also three exoribonuclease (RNase) enzymes (XRN, XRN2 and RNB) with antiviral RNA-degrading properties [ 405 , 406 ]. Annotation of the coral genome reveals homologues to four interferon proteins (IFNB, IFNG, IFNW1 and IFNT1). Interferons are potent and selective antiviral cytokines [ 407 ], which are induced by viral infection or by sensing dsRNA, a by-product of viral replication, leading to the transcription of interferon-stimulated genes whose products have antiviral activities and others having antimicrobial, antiprolifera-tive/antitumor or immumomodulatory effects [ 408 , 409 ]. Included in the coral antivirus defence system are three members of the interferon regulatory transcription factor (IRF1, IRF2 and IRF8) family proteins. IRF1 and IRF2 are transcriptional activators of cytokines and other target genes [ 410 ]; IRF1 is known to trans-activate the tumor suppressor protein p53 [ 411 ] while IRF2 regulates post-transcriptional induction of NO synthase [ 412 ]. Conversely, IRF8 is an interferon consensus sequence-binding protein that is a negative (interference) regulator of enhancer elements common to interferon-inducible genes [ 413 ]. The coral genome additionally includes an interferon-stimulated 20 kDa protein (ISG20) RNase specific to deactivation of singled-stranded RNA viruses [ 414 ]. The coral genome encodes several interferon-inducible proteins, notably interferon gamma induced GTPase (IGTP) that accumulates in response to IFNB [ 415 ], the interferon-induced GTP-binding protein Mx1 that is a key element of host antiviral defence [ 416 ], the interferon-induced helicase C domain-containing protein1 (aka MDA-5), which is an immune receptor that senses viral dsRNA to activate the interferon antiviral-response cascade [ 417 ] and the interferon-induced transmembrane protein (IFITM1) that suppresses cell growth [ 418 ]. The coral genome encodes the interferon-gamma receptor 2 (IFNGR2) transmembrane protein that activates downstream signal transduction cascades that control cell proliferation and apoptosis [ 419 ]. Encoded also is a homologue of the human bone marrow stromal cell antigen 2 (BST2) that inhibits retrovirus infection by preventing VLP release from infected cells [ 420 ]. Additionally encoded is a mitochondrial antiviral-signalling protein (MAVS) that triggers the host immune response by activation of the nuclear transcription factor NF-kB and the interferon regulatory transcription factor IRF3 which coordinates the expression of type-1 interferons such as IFNB [ 421 ]. Table 14 Regulatory and related proteins of viral pathogenicity in the predicted proteome of A. digitifera Gene sequence KEGG Orthology Encoded protein description v1.20647; v1.06188; v1.21287 K12599 Antiviral helicase SKI2 v1.18443 [+ 40 other sequence copies] K12807 Baculoviral IAP repeat-containing protein 1 (BIRC1) v1.06263 [+ 6 other sequence copies] K04725 Baculoviral IAP repeat-containing protein 2/3/4 (BIRC2/3/4) v1.14355 K08731 Baculoviral IAP repeat-containing protein 5 (BIRC5) v1.04171 [+ 7 other sequence copies] K10586 Baculoviral IAP repeat-containing protein 6 (BIRC6) v1.12348; v1.01945; v1.16612 K06731 Bone marrow stromal cell antigen 2 (antiviral BST2) v1.01539 [+ 7 other sequence copies] K04012 Complement component receptor 2 (CR2) v1.17305 K04462 Ecotropic virus integration site 1 protein (EVI1) v1.1496 [+ 4 other sequence copies] K12618 5′-3′ Exoribonuclease 1 (antiviral XRN1) v1.22746; v1.19002; v1.12850; v1.21216 K12619 5′-3′ Exoribonuclease 2 (antiviral XRN2) v1.09005 K01147 Exoribonuclease II (antiviral RNB) v1.22793; v1.12978; v1.19008; v1.20838 K09239 HIV virus type I enhancer-binding protein (HIVEP) v1.02776 [+ 7 other sequence copies] K15046 Influenza virus NS1A-binding protein (NS1A-BP) v1.09829; v1.13077 K05415 Interferon beta (IFNB) v1.11946; v1.21512; v1.11221; v1.11927 K04687 Interferon gamma (IFNG) v1.21512 K14140 Interferon gamma induced GTPase (ITGP) v1.11946 K05133 Interferon gamma receptor 2 (IFNGR2) v1.01539 [+ 4 other sequence copies] K04012 Interferon-induced GTP-binding protein Mx1 v1.10782; v1.23797; v1.17119; v1.03221 K12647 Interferon-induced helicase C domain-containing protein 1 v1.06274; v1.15849; v1.05943 K06566 Interferon induced transmembrane protein (IFITM1) v1.21327; v1.24081 K05440 Interferon, omega 1 (IFNW1) v1.11817 K09444 Interferon regulatory factor 1 (IRF1) v1.11816; v1.07639 K10153 Interferon regulatory factor 2 (IRF2) v1.11421 K10155 Interferon regulatory factor 8 (IRF8) v1.02158 K12579 Interferon-stimulated gene 20 kDa protein (ISG20) v1.15947 K05442 Interferon tau-1 (IFNT1) v1.22825; v1.08034; v1.08520 K05788 Integration host factor subunit beta (IHFB) v1.14899 K08220 MFS transporter, FLVCR family virus subgroup C receptor v1.04514; v1.04513; v1.16929 K12648 Mitochondrial antiviral-signalling protein (MAVS) v1.17718; v1.08002; v1.08001; v1.22382 K06081 Poliovirus receptor-related protein 1 (PVRL1) v1.21413; v1.06637 K06531 Poliovirus receptor-related protein 2 (PVRL2) v1.11740; v1.21467; v1.11410; v1.17135 K06592 Poliovirus receptor-related protein 3 (PVRL3) v1.15077 K06593 Poliovirus receptor-related protein 4 (PVRL4) v1.04158 [+ 68 other sequence copies] K12600 Superkiller protein 3 (antiviral SHI3) v1.18238 [+ 4 other sequence copies] K12601 Superkiller protein 8 (antiviral SHI8) The coral genome encodes a full set of baculoviral IAP repeat-containing proteins BIRC 1-6 (Table 14 ). The IAP (inhibitor of apoptosis) family proteins were first identified secreted by baculovirus to protect infected cells from death in the progression of viral replication [ 422 ]. Expressed by most eukaryotic organisms (reviewed in [ 423 ]), their IAP function is presumably conserved in corals. The coral genome encodes a full set of poliovirus receptor-related proteins (PVRL1-4) of the immunoglobulin superfamily, which bind and transport herpesviruses at the cellular membrane in the establishment of latent infections (reviewed in [ 424 ]). Encoded also is a complement component (3d/Epstein Barr virus) receptor 2 (CR2) protein that binds to the Epstein-Barr virus Herpes viridae with antigenic activity for disease prevention [ 425 ]. Another encoded protein is a homologue of the human immunodeficiency virus type 1 (HIV-1) enhancer-binding protein (HIVEP; aka EBP1) that attaches to the HIV long terminal repeat (LTR) region to activate transcription via the HIV LTR [ 426 ]. Present in the coral genome is also a homologue of the influenza virus non-structural binding protein NS1A-BP that interacts with the NS1 virulence factor of the influenza A virus Orthomyxoviridae to interfere with NS1-inhibition of pre-mRNA splicing within the host nucleosome [ 427 ]. NS1A-BP inhibits NS1A-mediated disruption of the host immune response caused by restricting interferon production and the antiviral effects of IFN-induced proteins [ 428 ]. The genome of A. digitifera encodes an integration host factor subunit beta (IHFB), first discovered as a host factor for bacteriophage λ integration of mobile genetic elements, that in E. coli is involved in multiple processes of DNA replication, site-specific recombination and gene expression [ 429 ]. A homologue of the MFS transporter feline leukemia virus subgroup C receptor (FLVCR) cell surface protein is encoded in the coral genome, which in cats confers susceptibility to FeLV-C infection [ 430 ]. Encoded also is a viral integration site 1 (EVI1) that in humans is an oncogenic transcription factor, often activated by viral infection, to cause proliferation of invasive tumours [ 431 ]. Arguably, these homologue proteins typically expressed in such distantly related species may have similar relevance in viral interactions of the coral holobiome. How these regulatory proteins and viral receptors interact and respond to viral infection in corals is yet to be realised. The absence of virion-specific sequences (e.g. for nucleic acid replication or capsid structure) suggests that proviral DNA is absent from the coral genome, or it may be an artefact of the limited number of marine viral sequences deposited in public databases. Discovery of viral activity through proteomics [ 39 ] may, therefore, suggest that viral proteins are synthesised from a lytic infection, but this requires confirmation. Toxins and venom A review of protein sequences deposited in the UniProt database in October 2012 shows that there are 150 known cnidarian toxins. These toxins have diverse biological activities (neurotoxins, pore-forming cytolysins and venom phospholipases) used to capture prey and for protection against predators [ 432 ] that are best characterised in sea anemones (Actiniaria) with 141 sequences deposited [ 433 , 434 ]. The cytotoxin MCTx-1 isolated from the Net Fire Coral Millepora dichotoma is the only toxin from a coral deposited in Uniprot (accession number A8QZJ5). However, our initial examination of the predicted proteome of A. digitifera shows 18 proteins with similarity to bacterial toxins and associated regulatory proteins (Table 15 ). Unlike reports from proteomic examination of the coral S. pistillata [ 39 ] and nematocysts (stinging organelles) of the jellyfish Olindias samba-quiensis [ 435 ], Tamoya haplonema , Chiropsalmus quadrumanus , Chrysaora lactea (PF Long et al ., pers comm), by sea anemones [ 434 ] and by the highly dangerous box jellyfish Chironex fleckeri [ 436 , 437 ], no venoms typical of higher animals were found in the A. digitifera genome. This was because our annotation was carried out using the KEGG database (release v58 [ 53 ]) to relate A. digitifera protein sequences to KEGG orthologues. The KEGG database is a collection of proteins from well characterised and ubiquitous biochemical pathways. Animal venoms, however, are highly specialised proteins for which this release of the KEGG database does not contain any described orthologues. Table 15 Proteins homologous to bacterial toxins in the predicted proteome of A. digitifera Gene sequence KEGG Orthology Encoded protein description v1.20214 K11029 Anthrax edema toxin adenylate cyclase (CyaA) v1.17686 K10921 Cholera toxin transcriptional activator (ToxR) v1.13017 K11020 Exotoxin A (ToxA) v1.23507 K13655 HTH-type transcriptional regulator (MsqA) antitoxin for MqsR v1.21184 K11009 Murine toxin (Ymt) v1.04313 K11033 Non-hemolytic enterotoxin A (NheA) v1.08011 K11034 Non-hemolytic enterotoxin B/C (NheBC) v1.08255 K13651 Motility quorum-sensing regulator (MqsR) interferase toxin v1.15986 K11059 Probable enterotoxin A (EntA) v1.13046 K04392 Ras-related C3 botulinum toxin substrate 1 (Rac1) v1.13966 K11007 Shiga toxin subunit B (StxB) v1.23958 K11063 Toxin A/B (TcdAB) v1.21174 K10930 Toxin co-regulated pilin (TCP) v1.05802 K10961 Toxin co-regulated pilus biosynthesis protein I (TcpI) v1.21783 K10964 Toxin co-regulated pilus biosynthesis protein S (TcpS) v1.14688 K15126 Type III secretion system cytotoxic effector protein (BteA) v1.05520 K11028 Vacuolating cytotoxin (VacA) v1.06590 K10954 Zona occludens toxin (Zot) KEGG orthology-based annotation of the A. digitifera genome reveals genes encoding protein homologues of 10 bacterial toxins, 7 regulatory toxin proteins and a botulinum protein substrate (Table 15 ). Of the 9 toxin homologues, one with similarity to anthrax edema factor (EF) adenylate cyclase (CyaA) is one of three proteins that comprise the anthrax toxin of Bacillus anthracis , the other two being a protective antigen (PA) and lethal factor (LF). Without the LF protein, anthrax CyaA has no known toxic effects in animals [ 438 ], although the EF protein does play an important role in disabling cellular functions vital for microbial host defences [ 439 ]. The A. digitifera genome encodes a secretion virulence factor exotoxin A-like protein produced by Pseudomonas aeruginosa , which for this bacterium affects local tissue damage, bacterial invasion and immunosuppression within their eukaryote host [ 440 ] with pathogenicity similar to that of the diphtheria toxin [ 441 ]. Another encoded protein is a murine-like toxin (Ymt) produced by the enterobacterium Yersinia pestis , which is the causative agent responsible for transmission of the notorious bubonic plague [ 442 ]. Additionally, two hemolytic enterotoxins similar to NheA and NheBC produced by Bacillus cereus [ 443 ], an enterotoxin (EntA) similar to that of Staphylococcus aureus [ 444 ], a Shiga-like enterotoxin (StxB) produced by Shigella dysenteria , the diarrhoea-causing toxin A/B (TcdAB) such as that secreted by Clostridium difficile [ 445 ], and a protein similar to the zonula occludens (tight junction) enterotoxin (Zot) secreted by Vibrio cholera [ 446 ] are encoded in the A. digitifera genome. Within the predicted proteome is also a homologue of the vacuolating cytotoxin (VacA) produced by Helicobacter pylori that colonises the gastric mucosa of the human stomach epithelium [ 447 ]. Although a direct homologue of the cholera toxin (CT) was not found encoded in the A. digitifera genome (Table 15 ), a protein similar to its transcriptional activator ToxR was. ToxR not only controls the expression of CT in Vibrio cholera [ 448 ], but also a co-regulated pilin (TcpA) protein that is under control of the ToxR regulon cascade [ 449 ]. Bacterial TcpA protein is assembled into toxin-coregulated pili that induce the transfer of DNA by horizontal exchange of genetic material during conjugation [ 450 ]. TcpA and two toxin co-regulated biosynthetic proteins (TcpI and Tcps) of the bacterial virulence-associated pilus appendage [ 451 ] are encoded in the coral genome. Entrained also are the motility quorum-sensing interference regulator MsqR and its transcriptional regulator MsqA that in Eschericia coli controls biofilm formation by inhibiting quorum-sensing motility, and together the MqsR/MqsA complex represses the lethal cold shock-like protein cspD gene [ 452 ] that on expression impairs DNA replication [ 453 ]. The A. digitifera genome likewise encodes a Type III secretion system T3SS cytotoxic effector (BteA) protein [ 454 ] that in Gram-negative invasive bacteria is translocated into host cells to suppress innate immunity to enhance virulence [ 455 , 456 ]. However, the ecophysiological significance of these toxigenic proteins and allied regulators, if indeed expressed by the coral genome, is unknown. In addition to using the KEGG database, we undertook a BLAST search of the predicted proteome of A. digitifera against peptide sequences for all animal venoms using the annotated UniProtKB/Swiss-Prot Tox-Prot program [ 457 ]. This search revealed a large number of accession hits from the predicted proteome, although these are unlikely to be true multiple copies given that the genome sequence has yet to be completely assembled. However, just taking a single accession number from each annotation reveals a complex array of 83 toxins that represents the predicted venom of A. digitifera (Table 16 ); UniProt BLAST E-values are given in Additional file 1 : Table S16b. These venoms are highly diverse and are significantly homologous to toxins from a wide variety of venomous marine and terrestrial creatures such as fish, reptiles, other cnidarians, cone-snails, stinging insects and even a venomous mammal (Shrew), covering the complete range of pharmacological properties known in venoms, including cytolytic, neurotoxic, haemotoxic, phospholipase, proteinase and proteinase inhibitor activities. Both the number of toxins predicted in the venom of A. digitifera and the degree of homology to such widely divergent phyla is remarkable. Accordingly, cnidarian venoms may possess unique biological properties that might generate new leads in the discovery of novel pharmacologically active drugs. Gene duplication followed by mutation and natural selection is widely held as the key mechanism whereby the large diversity of toxins found within a single venom could have evolved [ 458 , 459 ]. Conversely, primary mRNA splicing patterns have been shown to account for the diversity of metallopro-teinases in the pit viper Bothrops neuwiedi [ 460 ]. Variations in peptide processing have also been shown by proteomics and transcriptomics to explain how a limited set of genes transcripts could generate thousands of toxins in a single species of cone snail [ 461 ]. Despite these various processes that could account for the evolution of toxin diversity, it has never been demonstrated how gene duplications or variations in transcript or peptide processing could have radiated across the very different poisonous creatures found on Earth. Our data (Table 16 ) reveal that the predicted toxins of A. digitifera venom are orthologues to all of the most important superfamilies of peptide/protein venoms found in diverse taxa. We posit that the origins of toxins in the venoms of higher organisms may have arisen from deep eumetazoan innovations and that the molecular evolution of these venom super gene families can now be addressed taking an integrated venomics approach using Cnidaria such as the jellyfish as model systems [ 462 ]. Table 16 UniProt-predicted homologues of animal venom proteins in the predicted proteome of A. digitifera Gene sequence UniProt toxin accession Animal with closest homology v1.01916 [+ 5 other sequence copies] Q92035; Acetylcholinesterase Bungarus fasciatus (Banded Krait) v1.06761; v1.08075; v1.09840; v1.20323 Q9IAM1; Agkisacutacin (subunit anticoagulant protease) Deinagkistrodon acutus (Sharp-nosed Viper) v1.04809 A8QL52; L-Amino acid oxidase Bungarus fasciatus (Banded Krait) v1.06380 Q4JHE1; L-Amino acid oxidase Pseudechis australis (Mulga Snake) v1.10291 P81383; L-Amino acid oxidase Ophiophagus hannah (King Cobra) v1.14412 A6MFL0; L-Amino acid oxidase Demansia vestigiata (Lesser Black Whipsnake) v1.16469 P81383; L-Amino acid oxidase Ophiophagus hannah (King Cobra) v1.23477 P81382; L-Amino acid oxidase Calloselasma rhodostoma (Malayan Pit Viper) v1.16440 C5NSL2; Bandaporin (haemolysin) Anthopleura asiatica (Sea Anemone) v1.16571 [+ 10 other sequence copies] Q76B45 ; Blarina toxin (vasoactive protease) Blarina brevicauda (Northern Short-Tailed Shrew) v1.06055 [+ 20 other sequence copies] Q593B6; Coagulation factor V Pseudonaja textilis (Eastern Brown Snake) v1.07831; v1.10094 ; v1.20732 P14530; Coagulation factor IX Protobothrops flavoviridis (Okinawa Habu Snake) v1.01708 [+ 5 other sequence copies] Q4QXT9; Coagulation factor X Tropidechis carinatus (Rough-Scaled Snake) v1.09601; v1.10410 Q93109; Equinatoxin-5 (cytolysin) Actinia equina (Beadlet Anemone) v1.06821 Q08169 ; Hyaluronidase Apis mellifera (European Honey Bee) v1.08924 I0CME7; Hyaluronidase, Conohyal-Cn1 Conus consors (Singed Cone) v1.06189 [+ 112 other sequence copies] Q9XZC0; α-Latrocrustotoxin Lt1a (neurotoxin) Latrodectus tredecimguttatus (Mediterranean Black Widow Spider) v1.02942 [+ 8 other sequence copies] G0LXV8; α-Latrocrustotoxin Lh1a (neurotoxin) Latrodectus hasseltii (Australian Redback Spider) v1.00644 [+ 32 other sequence copies] Q25338; Δ- Latroinsectotoxin Lt1a (neurotoxin) Latrodectus tredecimguttatus (Mediterranean Black Widow Spider) v1.07446 A7X3X3; Lectin, Lectoxin Enh4 (platelet binding) Enhydris polylepis (Macleay’s Water Snake) v1.20653 A7X3Y6; Lectin, Lectoxin Enh7 (platelet binding) Enhydris polylepis (Macleay’s Water Snake) v1.02561,v1.11493; v1.16681 A7X3Z4; Lectin, Lectoxin Lio1 (platelet binding) Liophis poecilogyrus (Water Snake) v1.13597; v1.08696; v1.10757; v1.20654 A7X3Z7; Lectin, Lectoxin Lio2 (platelet binding) Liophis poecilogyrus (Water Snake) v1.18386, v1.15479 A7X413; Lectin, Lectoxin Lio3 (platelet binding) Liophis poecilogyrus (Water Snake) v1.06094 A7X406; Lectin, Lectoxin Phi1 (platelet binding) Philodryas olfersii (Green Cobra) v1.06416; v1.16248; v1.23712 A7X3Z0; Lectin, Lectoxin Thr1 (platelet binding) Thrasops jacksonii (Black Tree Snake) v1.17681 Q6TPG9; Lectin, Mucrocetin (platelet binding) Protobothrops mucrosquamatus (Brown Spotted Pit Viper) v1.00077 [+ 14 other sequence copies] Q66S03; Lectin, Nattectin (platelet binding) Thalassophryne nattereri (Toad Fish) v1.12241; v1.02332; v1.12298 Q71RQ1; Lectin, Stejaggregin-A (platelet binding) Trimeresurus stejnegeri (Bamboo Viper) v1.02245 [+ 19 other sequence copies] A0FKN6; Metalloprotease, Astacin-like toxin Loxosceles intermedia (Recluse Spider) v1.03638; v1.14772 Q90391; Metalloprotease, Atrolysin Crotalus atrox (Western Diamondback Rattlesnake) v1.13106 D3TTC2; Metalloproteinase, Atragin Naja atra (Chinese Cobra) v1.11132 Q7T1T4; Metalloproteinase, BjussuMP-2 Bothrops jararacussu (Jararacussu Pit Viper) v1.02168 O73795; Metalloproteinase, Disintegrin Gloydius brevicaudus (Chinese Mamushi Snake) v1.06910 Q7SZE0; Metalloproteinase, Disintegrin Gloydius saxatilis (Rock Mamushi Snake) v1.22282 P14530; Metalloproteinase, Disintegrin Protobothrops flavoviridis (Okinawa Habu Snake) v1.03804 Q2UXQ5; Metalloproteinase, EoVMP2 Echis ocellatus (West African Carpet Viper) v1.02016 Q91511; Mucrofibrase-5, Hypotensive serine protease Protobothrops mucrosquamatus (Brown Spotted Pit Viper) v1.09026 Q7ZZN8; Natrin-2 (neurotoxin) Naja atra (Chinese Cobra) v1.04153; v1.04595; v1.12730; v1.04157 A0ZSK3; Neoverrucotoxin (haemolysin) Synanceia verrucosa (Reef Stone Fish) v1.12433 [+ 5 other sequence copies] A2VBC4; Phospholipase A1 Polybia paulista (Neotropical Social Wasp) v1.00019; v1.13757 Q06478; Phospholipase A1 1 Dolichovespula maculata (Bald-Faced Hornet) v1.09322; v1.09961; v1.13629 P0CH47; Phospholipase A1, Magnifin Vespa magnifica (Giant Hornet) v1.03556 P53357; Phospholipase A1 2 Dolichovespula maculata (Bald-Faced Hornet) v1.13015; v1.16921 D2X8K2; Phospholipase A2 Condylactis gigantean (Giant Caribbean Sea Anemone) v1.18628 Q9TWL9; Phospholipase A2, Conodipine-M Conus magus (Magical Cone) v1.11796 Q9PUH9; Phospholipase A2, Acidic S9-53 F Austrelaps superbus (Lowland Copperhead Snake) v1.09883 Q8AXW7; Phospholipase A2, Basic Micrurus corallinus (Painted Coral Snake) v1.14874 Q90WA8; Phospholipase A2, Basic 2 Bungarus fasciatus (Banded Krait) v1.11797 P20256; Phospholipase A2, Basic PA-12C Pseudechis australis (Mulga Snake) v1.07278 [+ 34 other sequence copies] Q7SZN0; Prothrombin activator Pseutarin-C Pseudonaja textilis (Eastern Brown Snake) v1.11045 P83370; Prothrombin activator Hopsarin-D Hoplocephalus stephensii (Stephen’s Branded Snake) v1.04104 [+ 5 other sequence copies] Q58L94; Prothrombin activator Notecarin D2 Notechis scutatus (Tiger Snake) v1.00387 [+ 9 other sequence copies] Q58L90; Prothrombin activator Omicarin C Oxyuranus microlepidotus (Inland Taipan ) v1.02137 [+ 38 other sequence copies] Q58L91; Prothrombin activator Omicarin C Oxyuranus scutellatus (Coastal Taipan) v1.00618 [+ 10 other sequence copies] Q58L93; Prothrombin activator Porpharin D Pseudechis porphyriacus (Red-Bellied Black Snake) v1.09896 P81428; Prothrombin activator Trocarin D Tropidechis carinatus (Rough-Scaled Snake) v1.13726 A6MFK7; Prothrombin activator Vestarin D1 Demansia vestigiata (Lesser Black Whipsnake) v1.02129; v1.05362; v1.20273 Q6T269; Protease inhibitor, Bitisilin-3 (neurotoxic) Bitis gabonica (Gaboon Viper) v1.06980; v1.09028 Q3SB05; Pseudechetoxin (neurotoxin) Pseudonaja textilis (Eastern Brown Snake) v1.21284 [+ 5 other sequence copies] D8VNS7; Ryncolin-1 (haemostasis inhibitor) Cerberus rynchops (Dog-Faced Water Snake) v1.18895 [+ 20 other sequence copies] D8VNS8; Ryncolin-2 (haemostasis inhibitor) Cerberus rynchops (Dog-Faced Water Snake) v1.14251; v1.10489; v1.14254 D8VNS9; Ryncolin-3 (haemostasis inhibitor) Cerberus rynchops (Dog-Faced Water Snake) v1.06759 [+ 7 other sequence copies] D8VNT0; Ryncolin-4 (haemostasis inhinitor) Cerberus rynchops (Dog-Faced Water Snake) v1.01273 Q9YGN4; Salmorin toxin (haemostasis inhibitor) Gloydius brevicaudus (Chinese Mamushi Snake) v1.09855; v1.09856 B2DCR8; SE-Cephalotoxin Sepia esculenta (Golden Cuttlefish) v1.16247 O13060; Serine protease, 2A Trimeresurus gramineus (Bamboo Viper) v1.08397; v1.09733 Q9DF66; Serine protease, 3 (haemostasis inhibitor) Protobothrops jerdonii (Jerdon’s Pit Viper) v1.03275 Q9DG84; Serine protease, Serpentokallikrein-2 (haemostasis inhibitor) Protobothrops mucrosquamatus (Brown Spotted Pit Viper) v1.16638 Q7SYF1; Serine protease, Cerastocytin (platelet binding) Cerastes cerastes (Saharan Horned Viper) v1.22320 P0C5B4; Serine protease, Gloshedobin (platelet binding) Gloydius shedaoensis (Shedao Pit Viper) v1.15074 [+ 4 other sequence copies] B2D0J4; Serine protease, Venom dipeptidyl peptidase 4 Apis mellifera (European Honey Bee) v1.05361 B6RLX2; Serine protease inhibitor, TCI (neurotoxin) Ophiophagus hannah (King Cobra) v1.10994 B7S4N9; Serine protease inhibitor, Taicatoxin (neurotoxin) Oxyuranus scutellatus (Coastal Taipan) v1.11218; v1.23374 Q90WA0; Serine protease inhibitor, Textilinin-2 (thrombin inhibitor) Pseudonaja textilis (Eastern Brown Snake) v1.17856; v1.22256 Q8T3S7; Serine protease inhibitor, U1-aranetoxin-Av1a (neurotoxin) Araneus ventricosus (Devil Spider) v1.04154 [+ 4 other sequence copies] Q98989; Stonustoxin (haemostasis inhibitor) Synanceia horrida (Estuarine Stonefish) v1.09427; v1.16619; v1.19446 Q76DT2; Toxin AvTX-60A (cytolysin) Actineria villosa (Okinawan Sea Anemone) v1.12311 Q9GV72; Toxin CrTX-A (haemolysin) Carybdea rastonii (Jimble Jellyfish) v1.07546 [+ 5 other sequence copies] P58911; Toxin PsTX-60 (haemolysin) Phyllodiscus semoni (Night Anemone) v1.11270; v1.14265 E2IYB3; Veficolin-1 (complement activator) Varanus komodoensis (Komodo Dragon) v1.02115 Q98993; Verrucotoxin (cytolysin) Synanceia verrucosa (Reef Stonefish) Detoxification proteins of the chemical defensome There have been considerable advancements made to better understand the effects of pollution on coral reef habitats. The three main categories of environmental pollutants from anthropogenic sources are nutrient enrichment (eutrophication), hydrocarbon pollution and heavy metal contamination. Eutrophication from terrestrial inputs are a significant threat to coral reefs stemming from the discharge of treated sewage, the runoff of agricultural fertilizers (plus herbicides and pesticides), and by sedimentation caused by the erosion of organic-rich soils [ 463 ]. Notwithstanding that eutrophication can shift coral reef communities towards macroalgae domination [ 19 ], nitrogen and phosphorus enrichment can diminish coral growth and affect the photosynthetic performance of their algal symbionts [ 464 ]. Nutrient enhancement alters multiple pathways of primary metabolism that in coral is complicated by the photosynthetic demands of its symbiotic partners. While corals respond to hypertrophic levels of nutrients by activating general stress-response proteins [ 465 ], there are no specific proteins known to mitigate the cellular effects of nutrient enrichment on corals per se , and we have not attempted to identify such in this study. Gene families and their regulators that defend against chemical stressors comprise the chemical defensome encoding a network of detoxifying proteins that allows an organism to sense, transform and eliminate potentially toxic endogenous metabolites and xenobiotic contaminants [ 466 ]. Expressed proteins of the chemical defensome include the biotransformation cytochrome P450 (CYP) family of enzymes, conjugating enzymes, efflux transporters, heavy metal membrane pump exporters and their transcriptional activators. Annotation of the genome of A. digitifera reveals multiple genes encoding 20 hemoproteins belonging to the Phase II cytochrome P450 superfamily of monooxidase enzymes that catalyse the oxidation of diverse organic substances (Table 17 ). The substrates of CYP enzymes include intermediates of lipid metabolism and sterol/steroid biosynthesis, and include the detoxification of exogenous xenobiotics. Of significance are the CYP1A-type (aryl hydrocarbon hydroxylase) enzymes that have been studied widely in the hepatic response of fishes to polycyclic aromatic hydrocarbon (PAH) contamination (from crude or fuel oil) and exposure to polychlorinated biphenyl and dibenzodioxin toxicants (reviewed in [ 467 ]). CYP450 activity has been detected in the corals Favia fragum [ 468 ], Siderastrea siderea [ 469 ], Montastraea faveolata [ 470 ] and Pocillopora damicornis , [ 471 ]. Furthermore, CYP encoding sequences have been extracted from the genome of N. vectensis [ 472 ] and the transcriptome of A. millepora [ 29 ]. As well as providing chemical defence, mixed-function CYPs perform multiple endogenous tasks that are often taxon-specific. Hence, the orthology and substrate specificity of coral CYP enzymes cannot be predicted solely on homology to CYPs of known function assigned to higher metazoans. Similar to the function of CPY enzymes, there are genes encoding p-hydroxybenzoate 3-monooxygenase, an oxidoreductase catalyzing aryl oxidation and the soluble and microsomal forms of epoxide hydrolase that converts epoxides, formed by the degradation of aromatic compounds, to trans-diols that by conjugation are readily excreted. Conjugating enzymes to eliminate hydroxylated substrates are the detoxifying UDP-glucuronosyltransferase and sulfotransferase families of enzymes. Estrone sulfotransferase is significant for inactivation of exogenous (contraceptive) estrogens [ 473 ] and similar endocrine-disruptive contaminants released from treated wastewater [ 474 ]; their occurrence in marine waters are known to disrupt the reproduction and development of fish [ 475 ] and corals [ 476 ]. Glutathione S-transferase (GST) enzymes catalyse the addition of reduced glutathione to the reactive sites of electrophilic toxins [ 477 ]. Surprisingly, only two isoforms of GST were detected in the A. digitifera genome (Table 17 ), whereas 18 distinct GST-encoding genes (6 classes + 1 fungal-type) were classified from genome sequences of N. vectensis [ 472 ]. This unexpected genome reduction of GST elaboration in A. digitifera begs further examination. Table 17 Proteins of the chemical defensome in the predicted proteome of A. digitifera Gene sequence KEGG Orthology Encoded protein description v1.06127; v1.06128 K01015 Alcohol sulfotransferase v1.09267 K00537 Arsenate reductase v1.24496; v1.24495; v1.03953 K03893 Arsenical pump membrane protein v1.10691 K07755 Arsenite methyltransferase v1.20443 K11811 Arsenical resistance protein ArsH v1.14972 K01551 Arsenite-transporting ATPase v1.17644; v1.00480; v1.08150; v1.22865 K01014 Aryl sulfotransferase v1.21535; v1.11835; v1.02456 K01534 Cd 2+/ Zn 2+ -exporting ATPase v1.03485; v1.21926; v1.05686 K01533 Cu 2+ -exporting ATPase v1.22646 [+ 8 other sequence copies] K07408 Cytochrome P450, family 1, subfamily A, polypeptide 1 v1.01284 K07421 Cytochrome P450, family 2, subfamily T v1.10544; v1.02314, v1.17490 K07422 Cytochrome P450, family 2, subfamily U v1.23039 [+ 13 other sequence copies] K07422 Cytochrome P450, family 3, subfamily A v1.07750 K07425 Cytochrome P450, family 4, subfamily A v1.22798; v1.23000 K07426 Cytochrome P450, family 4, subfamily B v1.02020 [+ 4 other sequence copies] K07427 Cytochrome P450, family 4, subfamily V v1.19495 K07428 Cytochrome P450, family 4, subfamily X v1.15382 K15002 Cytochrome P450, family 6 v1.16427 K07430 Cytochrome P450, family 7, subfamily B v1.17631 K00498 Cytochrome P450, family 11, subfamily A v1.08074 [+ 4 other sequence copies] K15004 Cytochrome P450, family 12 v1.02478 [+ 5 other sequence copies] K00512 Cytochrome P450, family 17, subfamily A v1.06713 K07435 Cytochrome P450, family 20, subfamily A v1.22414 [+ 5 other sequence copies] K07436 Cytochrome P450, family 24, subfamily A v1.20153 K12665 Cytochrome P450, family 26, subfamily C v1.08074 [+ 6 other sequence copies] K00488 Cytochrome P450, family 27, subfamily A v1.06537 K07439 Cytochrome P450, family 39, subfamily A v1.22302 [+ 5 other sequence copies] K07440 Cytochrome P450, family 46, subfamily A v1.16335 K09832 Cytochrome P450, family 710, subfamily A v1.18439; v1.02594; v1.02593 K01016 Estrone sulfotransferase v1.07758 [+ 5 other sequence copies] K00699 Glucuronosyltransferase v1.00764 K13299 Glutathione S-transferase kappa 1 v1.17188 K00799 Glutathione S-transferase v1.04140 K07239 Heavy-metal exporter, HME family v1.10181 K00481 p-Hydroxybenzoate 3-monooxygenase v1.16748; v1.07471 K08365 MerR family transcriptional regulator, mercuric resistance v1.04382; v1.24424 K13638 MerR family transcriptional regulator, Zn(II)-responsive v1.12760 K08363 Mercuric ion transport protein v1.04179; v1.01891; v1.00145 K03284 Metal ion transporter, MIT family v1.21500 [+ 5 other sequence copies] K01253 Microsomal epoxide hydrolase v1.08005 K08970 Nickel/cobalt exporter v1.03484 K08364 Periplasmic mercuric ion binding protein v1.05406 K07245 Putative copper resistance protein D v1.14635 K08726 Soluble epoxide hydrolase v1.01929; v1.19296 K05794 Tellurite resistance protein TerC v1.10880; v1.15709; v1.12348 K07803 Zinc resistance-associated protein Many toxicological studies on the effects of pollution on cnidarian fitness have focused on their response to heavy metal contamination, including copper, cadmium, mercury and zinc [ 478 , 479 ]. In scleractinian corals the uptake and toxic effects of copper [ 480 - 483 ], cadmium [ 482 ] and mercury [ 484 , 485 ] have been studied at the metabolic level with specific studies to examine the effects of heavy metal toxicity on coral fertilisation [ 486 - 488 ], settlement [ 487 ], metamorphosis [ 486 ] and in coral bleaching [ 489 ]. Yet, the identification of molecular markers to monitor the response of Cnidaria to sub-lethal levels of heavy metal exposure has been elusive [ 490 ]. We were delighted to uncover in our annotation a wide range of genes to express metal-specific (arsenic, copper, mercury, nickel/cobalt and tellurium) resistance, transportation and membrane pump exporting proteins that, together with non-specific heavy metal ion export proteins (Table 17 ), might prove useful for monitoring the environmental response of A. digitifera to heavy metal contamination. Included in the heavy metal defensome are the Mer-family of transcriptional regulators of Hg- and Zn-resistance proteins and a periplasmic ion-binding protein attributed to the Hg detoxification system of bacteria [ 491 ]. Enzymes specific for arsenic detoxification are an arsenate oxidoreductase for conversion of arsenate to arsenite [ 492 ] and arsenite methyltransferase for conversion of arsenite to the less toxic dimethylarsenite that is amenable to excretion [ 493 ]. Such processes may enhance the resilience of corals exposed to natural [ 494 ] and site-affected [ 495 ] levels of arsenic contamination. In contrast, there were no (organo)cyanide detoxification genes apparent in the A. digitifera genome, but one sequence (v1.01601; K10814) encodes for hydrogen cyanide synthase of unknown metabolic purpose (data not tabulated). Ancillary evidence suggests that the expression of HCN synthase could be linked to quorum sensing [ 496 ] for regulating microbial densities of the coral holobiont community. Epigenetic and DNA-remodelling proteins In all Kingdoms of life, DNA methylation and chromatin remodelling is pivotal to the regulation of gene transcription independent of underlying allelic variation. One such process mediated by epigenetic changes in eukaryotic biology is the all-important cellular differentiation during morphogenetic development. Epigenetic modifications cause the activation, regulation or silencing of certain genes without changing the basic DNA code. Changes in epigenetic regulation can persist during cell division and across multiple generations [ 497 ]. In addition, cytosine methylation may be associated with a higher mutation rate, because deamination of the methylated base produces thymine resulting in C/T mutations, which on reproduction may be transmitted by the germline to subsequent generations in selective processes of evolution [ 498 ]. On the other hand, environmentally induced destabilisation of the epigenome can produce epigenetic gene variants (epialleles) that activate transcription and mobilization of DNA transposable elements, which may subsequently lead to stable heritable traits of environmental adaptation, as does occur by genetic imprinting in plants [ 499 ]. Transposition has thus the potential to direct increased frequencies of permanent genetic mutations for selective adaptation. One way by which genes are regulated at the epigenome is through the remodelling of the chromatin histone-DNA complex (the nucleosome), which by post-translational modification changes the template structure of DNA associated histone proteins. These modifications are affected by histone-lysine (and histone-arginine) N-methyltransferase enzymes (Table 18 ) by which these proteins may be further modified by acetylation, ADP-ribosylation, ubiquination, and phosphorylation (annotation not tabulated). The methylation pattern of histone lysine residues is highly predictive of the gene expression states of transcriptional activation and repression [ 500 ]. Necessary epigenomic reprogramming of histone modification at different stages of cell development is affected by the activation of histone and lysine-specific demethylase enzymes (Table 18 ). Determinants for recognition of the histone code are being revealed by a growing body of experimental data providing valuable information on the molecular tractability of binding sites involved in epigenetic signalling [ 501 ], which will enhance further insight to epigenetic function. Table 18 Epigenetic and DNA-remodelling proteins in the predicted proteome of A. digitifera Gene sequence KEGG Orthology Encoded protein description v1.04426; v1.02042 K02528 16S rRNA (adenine1518-N6/1519-N6)-dimethyltransferase v1.22358; v1.00249 K14191 18S rRNA (adenine1779-N6/1780-N6)-dimethyltransferase v1.19400; v1.04238 K00561 23S rRNA (adenine2085-N6)-dimethyltransferase v1.05107; v1.05242 K01488 Adenosine deaminase v1.04152; v1.09790 K14857 AdoMet-dependent rRNA methyltransferase SPB1 v1.00197 K13530 AraC family transcriptional regulator DNA methyltransferase v1.12967; v1.19789; v1.07763 K14589 Cap-specific mRNA (nucleoside-2′-O-)-methyltransferase 1 v1.24281 K01489 Cytidine deaminase v1.16211; v1.14952; v1.01094; v1.06983 K00558 DNA (cytosine-5-)-methyltransferase v1.19683; v1.05688; v1.04223 K11324 DNA methyltransferase 1-associated protein 1 v1.14033; v1.19860; v1.19081; v1.04188 K11420 Euchromatic histone-lysine N-methyltransferase v1.02068 K01487 Guanine deaminase v1.02920 K05931 Histone-arginine methyltransferase CARM1 v1.17589 [+ 7 other sequence copies] K11446 Histone demethylase JARID1 v1.07640 K06101 Histone-lysine N-methyltransferase ASH1L v1.13515; v1.18577; v1.20187; v1.19182 K09186 Histone-lysine N-methyltransferase MLL1 v1.08381 K09187 Histone-lysine N-methyltransferase MLL2 v1.24258; v1.19182 K09188 Histone-lysine N-methyltransferase MLL3 v1.07992; v1.10302; v1.13829 K09189 Histone-lysine N-methyltransferase MLL5 v1.06939; v1.15255; v1.15254 K11424 Histone-lysine N-methyltransferase NSD1/2 v1.05552 K11422 Histone-lysine N-methyltransferase SETD1 v1.07744 K11423 Histone-lysine N-methyltransferase SETD2 v1.03190 K11431 Histone-lysine N-methyltransferase SETD7 v1.21867 K11428 Histone-lysine N-methyltransferase SETD8 v1.18700 [+ 8 other sequence copies] K11421 Histone-lysine N-methyltransferase SETDB v1.07557; v1.11409 K11419 Histone-lysine N-methyltransferase SUV39H v1.24733; v1.13497 K11429 Histone-lysine N-methyltransferase SUV420H v1.15405; v1.10291; v1.17601; v1.02845; v1.08629 K11450 Lysine-specific histone demethylase 1 v1.23155; v1.09394; v1.17624; v1.05370 K14835 Ribosomal RNA methyltransferase Nop2 v1.18460 [+ 6 other sequence copies] K03500 Ribosomal RNA small subunit methyltransferase B v1.07407; v1.03110 K08316 Ribosomal RNA small subunit methyltransferase D v1.12193 K02427 Ribosomal RNA large subunit methyltransferase E v1.11499 K11392 Ribosomal RNA small subunit methyltransferase F v1.16053; v1.12676 K03437 RNA methyltransferase, TrmH family v1.12453; v1.05459 K13097 Methylcytosine dioxygenase v1.07692 K07451 5-Methylcytosine-specific restriction enzyme A v1.21815; v1.17113 K00565 mRNA (guanine-N7-)-methyltransferase v1.06363; v1.03360; v1.21218 K05925 mRNA (2′-O-methyladenosine-N6-)-methyltransferase v1.09661 K07442 tRNA (adenine-N1-)-methyltransferase catalytic subunit v1.08094; v1.04036; v1.18614 K03256 tRNA (adenine-N(1)-)-methyltransferase non-catalytic subunit v1.11456; v1.00738; v1.04577 K03439 tRNA (guanine-N7-)-methyltransferase v1.08042 K14864 tRNA methyltransferase v1.20501 K00557 tRNA (uracil-5-)-methyltransferase v1.15147 K14964 Set1/Ash2 histone methyltransferase subunit ASH2 v1.08925 K00571 Site-specific DNA-methyltransferase (adenine-specific) Direct epigenetic modification of DNA (or mRNA) occurs by methylation of cytosine, and to a lesser extent adenosine and guanine, by nucleobase-specific DNA methyltranferases (Table 18 ) to give 5-methylcytosine (5-meC), 3-methyladenosine (3-meA) and 3-methylguanine (3-meG) nucleotides, respectively. The principal modification product, 5-methylcytosine behaves much like regular cytosine by pairing with guanine, but in areas of high cytosine methylation, genome transcription is strongly repressed (reviewed in [ 502 ]), together with the repression of other chromatin-dependent processes, including the incorporation of transposable elements [ 503 ]. Alteration in the methylation status of the entire genome, individual chromosomes or at specific gene sites is essential for normal cellular function, but processes for reprogramming methylated DNA at different stages of cell development, unlike the reversal of histone modifications, is poorly defined [ 504 ]. While there are abundant enzymes to repair DNA damage caused by spurious N-alkylation, direct nucleotide C-demethylation (via the hypothetical “DNA demethylase” [ 505 ]) is thermodynamically infeasible. Instead, removal of epigenetic C-methylated nucleobases occurs by several base-repair pathways involving DNA excision or mismatch repair enzymes. The genome of A. digitifera encodes expression of a specific DNA glycosylase enzyme [ 506 ] for excision of 3-meA, but there are no such enzymes encoded for the excision of 5-meC and 3-meG, although there is encoded a 5-methylcytosine-specific restriction enzyme. Another pathway for DNA demethylation requires base-specific deamination by the AID/Apobec family of deaminase enzymes that, for example, converts 5-meC to thymine that is replaced subsequently by cytosine by C/T mismatch repair enzymes. These methylated nucleobases are recognized for deamination by the cytosine, adenosine and guanine deaminase enzymes [ 507 ] that are encoded in the A. digitifera genome, and their deaminated bases are subsequently removed by DNA mismatch repair enzymes. Additionally, the genome of A. digitifera encodes a methlycytosine dioxygenase enzyme that converts 5-methylcytosine to 5-hydroxymethycytosine (5-hmC), which is recognized for removal by the base excision repair pathway [ 508 ] or via its 5-hmC deaminated intermediate [ 507 ]. Combined, these DNA demethylation pathways are able to remodel epigenetic modifications at different stages of cell development. Most current knowledge on DNA and protein methylation comes from studies of mammals and plants, while our understanding of the extent and roles of DNA methylation in invertebrates, marine invertebrates in particular, is still limited [ 509 ]. Little is known about the epigenetic potential of corals to acclimatize and adapt to the thermal and synergistic stressors that cause wide-spread coral “bleaching” [ 510 ]. Yet, given that acclimatization occurs via the generation of epiallele variants that can in some instances lead to stable heritable traits of environmental adaptation, there is growing interest in the prospect that epigenetic modifications in corals or their algal symbionts [ 511 ] may drive adaptation to defend against the damaging threat imposed by rising temperatures from global climate change. It is anticipated that this field of study will rapidly accelerate with the need to better understand epigenetic processes that may contribute to the persistence of coral reefs."
} | 46,651 |
34740965 | PMC8609437 | pmc | 1,270 | {
"abstract": "Significance Life on Earth depends on ecologically driven nutrient cycles to regenerate resources. Understanding how nutrient cycles emerge from a complex web of ecological processes is a central challenge in ecology. However, we lack model ecosystems that can be replicated, manipulated, and quantified in the laboratory, making it challenging to determine how changes in composition and the environment impact cycling. Enabled by a new high-precision method to quantify carbon cycling, we show that materially closed microbial ecosystems (CES) provided with only light self-organize to robustly cycle carbon. Studying replicate CES that support a carbon cycle reveals variable community composition but a conserved set of metabolic capabilities. Our study helps establish CES as model biospheres for studying how ecosystems persistently cycle nutrients.",
"discussion": "Discussion The primary results of our study are the demonstration that CES can be powerful model microbial ecosystems for studying nutrient cycling and the development of a high-resolution method for quantifying cycling in closed communities. Model systems have proved essential for advancing every area of biology, including from gene expression ( 44 ), to development ( 45 ), to evolution ( 46 ). However, we lack model systems to serve the same purpose at the level of the community or ecosystem ( 18 ). Since CES are closed, nutrient cycling is required for persistence. Therefore, CES constitute model systems for studying nutrient cycling at the level of the collective with the key property of permitting control of community composition, nutrient, and energy availability. Given this tractability, CES constitute model biospheres for understanding how communities are organized to satisfy the constraints placed on them by nutrient cycling and for learning how evolutionary processes impact this organization. One of the main limitations in the field was a lack of precise, long-term, in situ measurements of nutrient cycling. We have overcome this limitation and demonstrated that CES are amenable to quantitative measurements of nutrient cycling while interrogating community structure at the taxonomic and metabolic levels. Our taxonomic and metabolic characterization of replicate CES showed that carbon cycling in CES can be sustained by diverse bacterial consortia that exhibit a conserved set of metabolic capabilities. The result points to the idea that the emergent functional property of carbon-cycling microbial communities is likely a conserved set of metabolic capabilities ( 39 ) that are robust to variation in the taxonomic structure of the system. However, some aspects of community function that we have not measured may depend on the taxonomic structure of the community, such as phosphorous sequestration. Ultimately, the functional aspects of the community that can be performed by diverse taxa likely depend on the phylogenetic conservation of the associated phenotypic traits. In this context, our data suggest that the conserved properties of carbon-cycling CES are likely carbon utilization pathways and the taxonomic diversity in our CES potentially reflects the weak phylogenetic conservation of carbon utilization phenotypes ( 47 ). It will be interesting to extend this study to understand the role of this taxonomic variability and metabolic convergence in determining the robustness of nutrient cycling to environmental perturbations such as changes in temperature or light levels. While previous studies have considered functional robustness in communities ( 48 ), our CES offer the advantages of real-time measurements of community function for many replicate consortia in the laboratory. The fact that CES are hermetically sealed means that they differ markedly from natural communities where immigration can change the makeup of the community. Despite this difference, we propose that CES can act as model systems for understanding how nutrient cycling constrains the structure of a community. While immigrations can and do alter the taxonomic structure of communities in the wild, it is frequently observed that metagenomic structure is tightly coupled to abiotic factors ( 39 ), suggesting that the assembly of functional communities may be deterministic given specific environmental contexts ( 8 , 26 ). In this case, provided a CES is initialized with sufficient metabolic diversity to satisfy the constraints on the system set by cycling, the final functional structure of the community may not depend strongly on whether or not immigrations are allowed to occur, a hypothesis that could be tested by opening CES and introducing invaders. Nutrient cycling in wild microbial communities often involves recycling of a single nonsubstitutable nutrient such as sulfur ( 5 ) or carbon ( 49 ), with other essential nutrients available in excess. This is in contrast to CES where no nutrients are supplied exogenously and biomass generation requires cycling all nutrients at once. In our CES it remains unclear to what extent nutrients other than carbon are cycled, such as those primarily involved in anabolism (N, P, Fe). It may be that the generation times in our CES are long, yielding few cell divisions in the course of the experiment. In this case carbon exchange could be utilized by algae and bacteria for maintenance energy. In this situation the cycling of nutrients such as N or P would be slow. In contrast, if generation times are short and many generations occur over the course of an experiment, nutrients such as N and P would need to be rapidly cycled to sustain cell division ( 50 ). In addition, quantifying abundance dynamics and metabolite exchanges in our CES would reveal how specific ecological interactions endow these communities with stable cycling capabilities. Detailed data on abundance dynamics would also permit comparison between our experiments and the substantial existing body of theoretical work on closed ecosystems ( 51 – 54 ). In particular, because the energy available to the system is readily varied by changing light intensity, CES could be used to test the proposed role of energetics in determining community structure ( 55 ). CES have a key role to play in future work understanding evolution at the level of the community. Simulations and directed evolution approaches have been used to ask whether and how ecosystem-level traits can be selected ( 56 , 57 ). As with directed evolution in proteins or organisms, the target of adaptation by the ecosystem is typically stipulated by the experiment. For example, communities might be selected for the production ( 58 ) or degradation ( 56 ) of a particular compound. Prior work in this field has faced two problems. First, it has been challenging to perform selection in the laboratory on a community-level trait that cannot be optimized by adaptation of an individual member of the community ( 57 , 59 ). For example, selecting a community for fast degradation of a compound can result in simply selecting the strain that degrades that compound most rapidly. Second, community-level evolution requires a notion of heritability, whereby successive generations of a community retain emergent traits of the parent community. However, theoretical work suggests a way to circumvent these obstacles: When selection acts on interaction-dependent properties of the ecosystem, such as metabolite exchange between strains, individual traits evolve to improve community heredity ( 60 ). Consistent with this expectation is the proposal that communities mediated by competition or exchange of resources can behave as cohesive units exhibiting emergent traits that are transmitted between generations ( 61 ). However, experimentally selecting a community on the basis of an emergent function that relies on interactions between constituents is a challenge. Nutrient-cycling closed ecosystems would appear to be ideal systems to address this problem since carbon cycling requires cooperative metabolic processes. Moreover, nutrient cycles in CES are frequently observed to be self-regulating ( 19 ). For example, autotroph senescence could provide additional carbon to heterotrophs and thus increase respiration, enhancing autotroph growth, a process that likely occurs in our CES. Self-regulation removes the requirement that the experimenter fine tune parameters to maintain functional stability ( 58 ). Therefore, the dual properties of obligate metabolic interactions to ensure persistence and self-regulation suggest that CES would be natural candidates for ecosystem breeding or long-term experimental evolution. Such efforts could yield insights into the coevolutionary dynamics governing symbioses in natural communities and potentially the eco-evolutionary dynamics of genome streamlining ( 62 ). Given these possibilities, we propose that CES, coupled with careful measurements of metabolite dynamics like those made here, constitute powerful model systems for the quantitative study of emergent nutrient cycling in communities."
} | 2,254 |
26837573 | PMC4797299 | pmc | 1,271 | {
"abstract": "Lignin-derived (e.g. phenolic) compounds can compromise the bioconversion of lignocellulosic biomass to fuels and chemicals due to their toxicity and recalcitrance. The lipid-accumulating bacterium Rhodococcus opacus PD630 has recently emerged as a promising microbial host for lignocellulose conversion to value-added products due to its natural ability to tolerate and utilize phenolics. To gain a better understanding of its phenolic tolerance and utilization mechanisms, we adaptively evolved R. opacus over 40 passages using phenol as its sole carbon source (up to 373% growth improvement over wild-type), and extensively characterized two strains from passages 33 and 40. The two adapted strains showed higher phenol consumption rates (∼20 mg/l/h) and ∼2-fold higher lipid production from phenol than the wild-type strain. Whole-genome sequencing and comparative transcriptomics identified highly-upregulated degradation pathways and putative transporters for phenol in both adapted strains, highlighting the important linkage between mechanisms of regulated phenol uptake, utilization, and evolved tolerance. Our study shows that the R. opacus mutants are likely to use their transporters to import phenol rather than export them, suggesting a new aromatic tolerance mechanism. The identified tolerance genes and pathways are promising candidates for future metabolic engineering in R. opacus for improved lignin conversion to lipid-based products.",
"introduction": "INTRODUCTION Lignocellulosic biomass, comprised of cellulose, hemicellulose and lignin ( 1 , 2 ), remains an underutilized substrate in sustainable microbial production of fuels and chemicals ( 3 – 6 ). One main challenge is that current biorefinery pretreatment approaches release diverse toxic degradation compounds from lignin during conversion of lignocellulosic biomass to fermentable sugars ( 7 ). These lignin degradation compounds include a wide array of phenolics that can severely inhibit microbial production of fuels or chemicals, leading to lower yields ( 8 ). Currently, unconverted lignin is typically burned to provide thermal energy onsite, but the amount of waste lignin is predicted to escalate as lignocellulose-based biorefinery output increases ( 4 , 9 , 10 ). Lignin, the second most abundant terrestrial polymer, constitutes ∼15–30% of lignocellulose ( 11 ) and is more energy dense than cellulose and hemicellulose due to its higher carbon-to-oxygen ratio. Unfortunately, lignin is much more difficult to depolymerize due to its complex molecular structure. Structural heterogeneity also leads to a broad spectrum of breakdown products, substantially compromising the efficiency of chemical catalysis approaches for product synthesis and purification. Some bacteria and fungi can consume lignin breakdown products and utilize them as carbon sources ( 12 ), potentiating fuel and chemical production via lignin consolidated bioprocessing ( 13 – 15 ). One such bacterium, Rhodococcus opacus , is a promising microbial host for converting lignocellulose to useful products due to its naturally robust lipid production and ability to both tolerate and metabolize diverse phenolic compounds ( 14 , 16 – 18 ). Rhodococcus strains are found in diverse environments ( 19 , 20 ) and can tolerate environmental stresses such as desiccation and high salinity ( 21 , 22 ) as well as chemical stresses such as high concentrations of butanol ( 23 , 24 ). Often isolated from polluted or contaminated environmental samples ( 25 , 26 ), R. opacus strains have a strong innate tolerance to benzene, toluene and lignocellulosic hydrolysates from different sources ( 27 , 28 ) and can metabolize aromatic compounds ( 14 , 16 ). Rhodococcus species can convert aromatics to acetyl-CoA and succinyl-CoA ( 12 , 29 ), which are important precursors for converting phenolics to bioproducts ( 30 ). Originally isolated from soil at a gas works plant, R. opacus PD630 (hereafter R. opacus unless specified) is known to accumulate large amounts of the biodiesel precursors triacylglycerols (TAGs, up to 76% of cell dry weight) when using sugars as a carbon source. Thus, R. opacus has been a target strain for commercial-scale lipid production using sugars derived from lignocellulose ( 18 , 31 – 34 ). Growth inhibition by toxic compounds (either end-product or feedstock) is a major limiting factor for commercialization of biochemical processes ( 35 , 36 ). Developing production hosts with natural tolerance to toxic inhibitors may significantly reduce time and efforts for host optimization. The tolerance capabilities of R. opacus are hypothesized to come from its highly hydrophobic cell wall ( 22 ) and/or its ability to consume a diverse array of compounds ( 20 ), but few studies have directly investigated phenolic tolerance mechanisms in this organism. While we have recently explored the central metabolism in wild-type (WT) R. opacus and found simultaneous utilization of glucose and phenol (i.e. no catabolite repression) using 13 C-metabolite fingerprinting ( 37 ), more work is necessary to thoroughly characterize the catabolic pathways of aromatic compounds and global metabolism to elucidate tolerance mechanisms in R. opacus . In this study, we adaptively evolved increased phenol tolerance of R. opacus by sequentially sub-culturing in phenol as a sole carbon source and screening fast-growing mutants over 40 passages. We selected phenol as a model lignin degradation product to avoid the confounding effects of many compounds present in heterogeneous lignin degradation product streams. Phenol has a shared substructure to many of the compounds that can be derived from lignin ( 38 ), and it also has a similar level of toxicity to that of other compounds derived from lignin ( 39 , 40 ). Even though previous studies have demonstrated bioconversion of lignocellulose-derived compounds into lipids by R. opacus ( 14 , 27 , 40 – 44 ), the connection between the tolerance phenotype and specific cellular mechanisms remains elusive. Here, we present a combined adaptive evolution/omics approach leveraging multiple phenol-adapted strains to identify possible mechanisms for phenolic tolerance and utilization in R. opacus (Supplementary Figure S1). This approach builds on previous studies by examining multiple paths to phenolic tolerance from the same strain background and resolving these tolerance strategies by comparing the transcriptomic response in different growth conditions. We performed adaptive evolution of R. opacus on increasing concentrations of phenol to select for increasingly tolerant strains and identified two high-performing strains through in-depth phenotyping. Next, we performed whole-genome sequencing of the selected strains to identify genomic alterations during adaptive evolution and employed comparative transcriptomics to identify transcriptional changes between strains in different growth conditions (i.e. in glucose and different concentrations of phenol). This approach proved to be effective in gaining insights into tolerance mechanisms and identifying promising gene candidates that can facilitate future metabolic engineering efforts in Rhodococcus to produce fuels and chemicals directly from lignin breakdown products or from lignocellulosic hydrolysates rich in toxic lignin degradation compounds.",
"discussion": "DISCUSSION In this work, we analyzed genomic and transcriptomic changes between adaptively evolved R. opacus mutants and the WT strain grown in phenol. Laboratory adaptive evolution has been used for many different bacteria to improve growth on diverse compounds ( 68 ). We used phenol as a sole carbon source in our adaptive evolution process and identified strains with higher tolerance to the toxic chemical by measuring their total biomass accumulation, lag phase and growth rate (Figure 1 and Supplementary Tables S1–S3). After ∼200 generations, adapted strains contained 3–4 SNPs in the genomes, which was more rapid accumulation of SNPs than reported by Lenski et al. in their long-term evolution experiment and other labs (8.9 × 10 −11 per base-pair per generation; 69 ). Strong selective pressure by the toxic compound being the sole carbon source probably contributed to the faster rate of SNP accumulation compared to the E. coli experiment, although the ‘normal’ rate of SNP accumulation during adaptive evolution in R. opacus has not been determined. Unfortunately, we did not identify ‘obvious’ SNPs that contributed to the striking transcriptomic profile difference between WT and evolved strains. The two adapted strains had improved growth profiles with increased biomass accumulation and shorter lag phases that could be due to altered transport of phenol in the mutants, leading to changes in intracellular concentrations of phenol and thus expression of phenol-responsive genes for phenol detoxification and degradation. Differences were also observed between WT and adapted strains in terms of lipid accumulation. From RNA-Seq expression data, phenol appears to be mainly degraded into acetyl-CoA and succinyl-CoA via the β-ketoadipate pathway ( 12 ). Acetyl-CoA can be integrated directly into fatty acid biosynthesis pathways, and succinyl-CoA can enter the TCA cycle to generate reducing equivalents and ATP for cell growth and lipid synthesis ( 70 ). Compared to WT, the adapted strains maintained the ability to synthesize and accumulate lipids when grown in glucose, but they both accumulated significantly more lipids than WT when grown in phenol (Figure 2 , Supplementary Tables S4 and S5). The adapted strains in this study demonstrated high phenol consumption capabilities that match or surpass other phenol-degrading bacterial strains and mixed cultures. Pseudomonas putida MTCC1194, which was adapted for growth in phenol as a sole carbon source, had maximum phenol degradation rates of ∼12 mg phenol/l/h during growth at its maximum phenol concentration of 1 g/l ( 71 ). Another approach used mixed cultures to increase phenol degradation capacity, with tolerances up to 0.8 g/l phenol and degradation rates of 15.4 mg/l/h ( 72 ). Another approach is to isolate strains directly from environments such as contaminated wastewater. Bacillus brevis , isolated from an industrial wastewater, showed the highest phenol tolerance and utilization at concentrations up to 1.75 g/l phenol with degradation rates of ∼20 mg/l/h ( 73 ). Our adapted strains showed higher tolerance than B. brevis with growth in 2 g/l phenol, and experiments at 1.5 g/l showed phenol degradation rates of 22 and 21 mg/l/h for evol33 and evol40, respectively (Supplementary Figure S3). In summary, evol33 and evol40 demonstrated comparative phenol degradation rates and tolerances to the best-performing environmental isolates. Our transcriptomic data suggested that phenol-adapted strains express lower levels of stress-response genes in the conditions tested (low nitrogen, minimal media and phenol as a sole carbon source). Phenol is known to cause oxidative stress response in bacteria ( 74 ). In our study, R. opacus upregulated ‘DNA protection during starvation protein’ LPD04106, or Dps, which is thought to be involved in oxidative stress resistance ( 75 , 76 ). It seems likely that low nutrient condition can induce the gene, but the adapted strain, especially evol33, showed significantly lower upregulation of this gene compared to WT in phenol, suggesting that evolved strains sensed less oxidative stress in our experiment. A similar trend was observed with the major housekeeping chaperone genes. DnaK is a major bacterial Hsp70 (70 kDa heat shock protein) and functions with co-chaperones GrpE and DnaJ to prevent aggregation of denatured proteins ( 77 , 78 ). R. opacus expressed 2 copies of the dnaK operons (LPD2079–02081 and LPD03845–03847) in the conditions we tested, and the major operon appeared to be LPD02079 and the following 3 genes. All DnaK and co-chaperone genes were upregulated more in WT than in the phenol-adapted strains in low phenol, indicating that the low-nutrient, phenol medium condition encountered by WT may cause proteins to misfold, and that the phenol-adapted strains encounter lower level of cellular stress than WT at the same concentration of phenol in the media. We observed clear patterns of altered gene regulation in phenol-adapted strains compared to WT R. opacus in response to phenol. In WT, one of the two copies of phenol hydroxylase genes was dominantly upregulated, but in adapted strains the second copy (LPD06740 and LPD06741) was also upregulated to the same or even higher level (Figure 3 , Table 2 and Supplementary Figure S8). Both SNP analysis and RNA-Seq results suggest that transport of phenol or compounds related to its metabolism is altered in the adapted strains. In fact, phenol was consumed at faster rates by the adapted strains than the WT strain (Supplementary Figure S3). evol33, which additionally had a putative transporter gene with a single point mutation compared to evol40 (Table 1 ), upregulated phenol degradation genes more strongly than evol40 (Figure 3 ). It is well documented that many identified transcriptional regulators for aromatic degradation and catabolism are AraC/XylS family ( 79 – 81 ). AraC/XylS-family transcriptional regulators bind to ligands (e.g. arabinose), and upregulation of effector genes is tunable by adjusting the concentration of ligands ( 82 – 84 ). Therefore, it is possible that adapted strains allow more phenol molecules to be transported into the cell, which in turn activate the second phenol degradation operons to the degree observed (6307-fold upregulation in evol33, 2624-fold upregulation in evo40 and 657-fold upregulation in WT; Figure 3 and Supplementary Table S6). Higher upregulation of β-ketoadipate pathway genes in adapted strains could also be influenced by higher intracellular concentrations of phenol or its downstream metabolite, catechol, which has been observed with aromatic degradation pathways in other organisms ( 79 , 80 ). There are 46 genes annotated as shikimate transporters in the complete genome of R. opacus . Of the 46, LPD06699 was uniquely co-upregulated with phenol degradation genes, at 353- and 103-fold in evol33 and evol40 comparing low phenol to glucose, and 937- and 572-fold in evol33 and evol40 comparing high phenol to glucose, respectively. In WT, LPD06699 was upregulated 5-fold in low phenol compared to glucose. Except for LPD07505, which was upregulated in phenol only in adapted strains, none of the other genes annotated as a shikimate transporter were regulated in the same manner (Supplementary Table S6). Furthermore, predicted shikimate utilization genes LPD05461, LPD06023, LPD17011 and LPD05485 were not highly transcribed or upregulated in phenol (Supplementary Table S6), suggesting that LPD06699 and potentially LPD07505 may be involved in phenol transport. Another interesting finding was the upregulation of a putative YRNA and Rsr ( 85 ) in adapted strains in phenol. YRNAs were recently found to exist in a wide variety of bacterial species, including R. opacus ( 86 ). R. opacus Rsr homolog, LPD06269, caught our attention because of its phenol-responsive upregulation in evolved strains, even though it was annotated as a hypothetical protein. LPD06269 was constitutively expressed at a medium-low level in all strains in glucose, but upregulated 52-fold in low phenol compared to glucose and 115-fold in high phenol compared to glucose in evol33 (8.3-fold and 17-fold in evol40), and upregulation of the predicted YRNA region showed the same pattern (Supplementary Figure S14). In WT, transcript levels of LPD06269 slightly decreased in low phenol (0.87-fold), and no upregulation of YRNA region was observed, suggesting that this response was phenol-dependent and specific to the adapted strains. We are in a process of focusing on small noncoding RNAs in R. opacus , since we removed the majority of small RNAs from the total RNA samples in order to obtain maximal amount of mRNA in this study. Nonetheless, our results indicate that small RNA-mediated global response may also be involved in adaptive evolution of R. opacus against phenol as a sole carbon source, and it may explain the ability of R. opacus to rapidly adapt to diverse environments without accumulating substantial genomic changes. This work suggests that characterizing funneling pathways (which convert lignin-derived compounds into pathway metabolites such as catechol; 87 ) and transporters for phenolics is important for conversion of phenolics into fuels and chemicals. Genes converting phenol to catechol were some of the most highly upregulated genes when strains were grown in phenol compared to glucose, suggesting that phenol-to-catechol conversion may be a limiting step for growth on phenol. Under the condition that we tested (low nitrogen, minimal medium and aerobic condition), phenol seems to be degraded mainly by ortho-cleavage via the catechol branch of the β-ketoadipate pathway into succinyl-CoA and acetyl-CoA, consistent with our previous result obtained from WT cultures grown in 0.5 g/l phenol supplemented with 13 C glucose ( 37 ). Other phenolic compounds would require other funneling enzymes to convert them into precursors such as catechol and protocatechuic acid for degradation ( 87 ). This study also identified putative transporters for phenol, and transport of many different phenolic compounds would require either promiscuity for transporters or different transporters for each compound. Note that while previous studies on transporters of toxic chemicals focused on efflux pumps to minimize the intracellular concentrations of these compounds ( 88 – 90 ), R. opacus is likely to use its transporters to import toxic phenolics, suggesting a new mechanism of bacterial aromatic tolerance. Increased phenolic compound flux into the cell requires balancing between transport and degradation to prevent accumulation of toxic compounds within the cell. A recent study demonstrated that adaptive evolution can increase R. opacus tolerance to lignocellulose-derived inhibitors ( 40 ). While they used media supplemented with glucose, our evolution process is different in that phenol was used as a sole carbon source to obtain highly tolerant mutants to phenol. Consequently, we were able to decouple the effects of other potentially available carbon sources in lignocellulose-derived feedstock on changes in the genome and transcriptome. We also demonstrated that phenol as a sole carbon source can support R. opacus for both growth and accumulation of lipids (biodiesel precursors) and that adaptive evolution can increase lipid productivity, an important factor for current lignin conversion processes ( 15 ). Combining adaptive evolution and omics analyses, our approach provides insights into the tolerance mechanism of the promising production host, facilitating future lignocellulose, specifically lignin, valorization efforts."
} | 4,757 |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.