pmid stringlengths 8 8 | pmcid stringlengths 8 11 ⌀ | source stringclasses 2
values | rank int64 1 9.78k | sections unknown | tokens int64 3 46.7k |
|---|---|---|---|---|---|
26122431 | null | s2 | 2,964 | {
"abstract": "Biofilm formation is a complex process involving various signaling pathways and changes in gene expression. Many of the sensory mechanisms and regulatory cascades involved have been defined for biofilms formed by diverse organisms attached to solid surfaces. By comparison, our knowledge on the basic mechanisms underlying the formation of biofilms at air-liquid interfaces, that is, pellicles, is much less complete. In particular, the roles of flagella have been studied in multiple solid-surface biofilm models but remain largely undefined for pellicles. In this work, we characterize the contributions of flagellum-based motility, chemotaxis and oxygen sensing to pellicle formation in the Gram-positive Bacillus subtilis. We confirm that flagellum-based motility is involved in, but is not absolutely essential for, B. subtilis pellicle formation. Further, we show that flagellum-based motility, chemotaxis and oxygen sensing are important for successful competition during B. subtilis pellicle formation. We report that flagellum-based motility similarly contributes to pellicle formation and fitness in competition assays in the Gram-negative Pseudomonas aeruginosa. Time-lapse imaging of static liquid cultures demonstrates that, in both B. subtilis and P. aeruginosa, a turbulent flow forms in the tube and a zone of clearing appears below the air-liquid interface just before the formation of the pellicle but only in strains that have flagella."
} | 363 |
33681975 | PMC8012112 | pmc | 2,965 | {
"abstract": "ABSTRACT Biofilm-forming bacteria have the potential to contribute to the health, physiology, behavior and ecology of the host and serve as its first line of defense against adverse conditions in the environment. While metabarcoding and metagenomic information furthers our understanding of microbiome composition, fewer studies use cultured samples to study the diverse interactions among the host and its microbiome, as cultured representatives are often lacking. This study examines the surface microbiomes cultured from three shallow-water coral species and two whale species. These unique marine animals place strong selective pressures on their microbial symbionts and contain members under similar environmental and anthropogenic stress. We developed an intense cultivation procedure, utilizing a suite of culture conditions targeting a rich assortment of biofilm-forming microorganisms. We identified 592 microbial isolates contained within 15 bacterial orders representing 50 bacterial genera, and two fungal species. Culturable bacteria from coral and whale samples paralleled taxonomic groups identified in culture-independent surveys, including 29% of all bacterial genera identified in the Megaptera novaeangliae skin microbiome through culture-independent methods. This microbial repository provides raw material and biological input for more nuanced studies which can explore how members of the microbiome both shape their micro-niche and impact host fitness.",
"conclusion": "CONCLUSION The dermis represents both a first line of defense for marine animals, and an environment uniquely colonized by phylogenetically and functionally diverse bacterial representatives. The challenge for marine microbial ecology remains in linking the phylogenetic diversity of host-associated microbes to their functional roles within the community. Progress will depend on merging data from sequencing-based studies with manipulative experiments using cultured isolates, thereby providing insight on nutrient cycling capacity, interspecies communication, defense against pathogens and parasites, host immune response, wound repair, microbial community succession and benefits to host development, health and survival (Krediet et al . 2013 ; Meyer, Paul and Teplitski 2014 ; Medina et al . 2017 ; Bell, Garland and Alford 2018 ; Longford et al . 2019 ; van Oppen and Blackall 2019 ). Moreover, culture-based methods have already proven successful in conservation efforts in terrestrial ecosystems (Daskin et al . 2014 ; Loudon et al . 2014 ) and could prove equally as valuable in marine systems (van Oppen and Blackall 2019 ). Our culturable library holds great promise for future manipulative experiments that can examine the metabolic factors and surface properties influencing complementary associations between animal host and microbe, as our study was able to capture a rich repository of marine microbes and inspires new questions that will undoubtedly fuel future research into the varied and elusive roles of marine microbes.",
"introduction": "INTRODUCTION The epidermis is an animal's first line of defense against the external environment, yet not impervious to environmental influences. The dermis of an organism can be thought of as its own ecosystem, hosting a diverse microbial milieu, where the composition of communities is driven by both endogenous host factors and the exogenous environment. In marine ecosystems, surface-associated microbes must contend with unpredictable external variables including temperature, pH, salinity fluctuations and skin/mucus shedding by the host. The animal surface microbiome plays a significant role in the health of the host by protecting the body against transient pathogenic microorganisms, selecting for antibiotic-producing commensal strains, playing critical roles in host nutrition and significantly impacting immune system development (Nelson et al . 2015 ; Bourne, Morrow and Webster 2016 ; Apprill 2017 ; Ross, Rodrigues Hoffmann and Neufeld 2019 ). Often marine animal surfaces are referred to as ‘hot spots’ of microbial diversity. These environments are dominated by unique assemblages of host-specific bacteria capable of generating specific microenvironments that promote higher-level microbial community organization, including the production of shielding biofilm matrices, antiprotozoal factors and chemical compounds that assist in protection from predators, viruses and environmental stressors (Bik et al . 2016 ; Dang and Lovell 2016 ). The uniqueness of surface-associated microbiota in both their genetic composition and functional roles in comparison to their free-living counterparts (Burke et al . 2011 ) indicates either an active role of the hosts in recruitment of epibiotic bacteria or a passive mechanism of colonization based on the eukaryotic surface and exudates (Wahl et al . 2012 ). The surface microbiomes of both cetaceans (Nelson et al . 2015 ) and coral (Rosenberg et al . 2007 ; Zaneveld et al . 2016 ) have been studied to elucidate the role of the microbial community on the health, persistence and resilience of these keystone and habitat-forming species. In corals, it is well known that their resident microalgal symbionts mediate host physiology. However, it has been suggested that bacteria, archaea and fungi participate in cycling of nutrients and organic matter in coral as well (Apprill 2017 ). While coral microbial communities are often host specific, spatially restrictive and can be highly stable across geographic and environmental conditions, oversimplification of host–microbe association via broad profiling studies can miss low abundance taxa that contribute to physiologically significant interactions with their hosts (Bourne, Morrow and Webster 2016 ). Similarly, marine mammals, particularly cetaceans, are considered sentinel species in marine ecosystems (Bossart 2011 ). There are few microbiome studies in cetaceans and still fewer that target members of the skin microbiome (Chiarello et al . 2017 ; Hooper et al . 2019 ; Apprill et al . 2020 ). Recent investigations into the skin microbiome of cetaceans have suggested the existence of species-specific assemblages that can change seasonally (Bierlich et al . 2017 ). These assemblages can also be linked to the presence of epiphytic diatoms (Hooper et al . 2019 ) and are shown to be distinct from surrounding planktonic samples (Chiarello et al . 2017 ), suggesting ecological and evolutionary forces including the unique features of an animal's epidermis have shaped cetacean skin microbiomes. The availability of sequence data is rapidly readjusting our view of how bacterial communities associated with marine organisms vary in composition depending on host type and environment (Vega Thurber et al . 2009 ; Sunagawa, Woodley and Medina 2010 ; Barott et al . 2011 ; Hentschel et al . 2012 ; Sison-Mangus et al . 2014 ; Cooper and Smith 2015 ; Rouco, Haley and Dyhrman 2016 ). However, there is a gap in microbial diversity obtained between sequencing-based techniques and culture-dependent isolation of strains. This study therefore sought to identify key features of animal surfaces and marine environments in order to recapitulate environmental conditions with the goal of isolating as diverse a complement of surface microbiome representatives possible from the dermis of two species of whale ( Delphinapterus leucas and Megaptera novaeangliae ) and the mucus and tissue from three species of coral ( Porites astreoides , Acropora palmata and Millepora alcicornis ). Six culture conditions were used in this study and chosen based on either previous success in culturing marine microbes or were developed here using knowledge of features associated with the dermis surface chemistry and microbial signaling processes. A total of three media variations used marine agar (MA), combining an agar and a commercially available marine broth base, and is a widely used media type demonstrated to yield biologically diverse microbial flora because of its nutrient richness (Mincer et al . 2002 ). One marine agar variation contained the antimicrobial enzyme, lysozyme (LYS), which is produced by cetaceans within the upper lamellae of the stratum corneum of the dermis and may select for microbes resistant to this non-specific defense mechanism (Seegers and Meyer 2004 ; Mouton and Botha 2012 ). Another marine agar variation contained both the secondary messenger cyclic adenosine monophosphate, shown to regulate bacterial metabolism, virulence gene expression, flagella mobility, surface attachment and biofilm formation (Smith, Wolfgang and Lory 2004 ; Fuchs et al . 2010 ; Ono et al . 2014 ; Dang and Lovell 2016 ), and N -(oxo-hexanoyl)-homoserine lactone, a homoserine lactone used by the majority of Gram-negative bacteria with quorum sensing ability (HSL-AMP; Bruns, Cypionka and Overmann 2002 ). Acylhomoserine lactone signals are known to mediate interspecies communication, biofilm formation and community structure (Wang et al . 2020 ). Lower nutrient media types included Actinobacteria -selecting media (AC; Okami and Hotta 1988 ) and a low nutrient media (R2A) containing chitin as both an energy source and substrata, that facilitates slower-growing, oligotrophic microbes (Reasoner and Geldreich 1985 ). The final media type selected for marine fungi (KJ; Kjer et al . 2010 ). Here we demonstrate the cultivation of 592 isolates contained within 15 bacterial orders representing 50 bacterial genera and two fungal species isolated from 25 cetacean and coral samples, indicating targeted cultivation by use of media additives and intensive microbial cultivation efforts yields high microbial diversity in comparison to other published efforts. Multivariate analyses demonstrate that microbial diversity is associated with sampled species rather than media variation, although different media variations and host surfaces were successful in culturing potentially phylogenetically distinct strains based on SSU rRNA variation. With this new microbial repository from diverse animal hosts, we can now begin to explore how interactions among microbial community members can impact host fitness.\n\nCulture-dependent design introduces bias in recovered microorganisms Many of the culturable bacteria from both whale and coral species included Alteromonas sp. (Fig. 2 ), which are globally distributed marine copiotrophic bacteria that dominate in heterotrophic blooms and outcompete other bacteria due to their rapid growth when organic nutrients are readily available (Math et al . 2012 ). The ubiquity of this genus in our whale samples could have resulted due to the nutrient-rich media base used during cultivation, which has previously led to an overestimation of their abundance in the environment due to the ease at which the cells form colonies on agar plates and the voracity with which they take advantage of sporadic inputs of organic matter (Behringer et al . 2018 ). However, our data indicate that a fraction of the 127 Alteromonas strains isolated here were only found on coral or on cetaceans, indicating potential host specificity for particular strains. Additionally, isolates that are phylogenetically similar can exhibit different biochemical profiles and functional roles (Lampert et al . 2006 ). Culture independent studies by themselves would likely fail to link varied contributions of similar strains to larger implications, such as their effects on host fitness. Future studies should integrate more diverse media types, particularly low nutrient conditions, to target more slower growing microorganisms that might have been overlooked in this study.",
"discussion": "DISCUSSION This study uses an intensive culturing effort to isolate 592 microbes comprising 50 bacterial genera within 15 orders, from the surface microbiomes of two species of whales and three species of corals, demonstrating the effectiveness of using our methodology and suite of culture conditions in capturing microbial diversity from diverse marine animals (Table S2, Supporting Information). A wealth of knowledge can be gained by using bacterial strains in manipulative experiments to understand disease origin; community dynamics relating to resistance, resilience and persistence; functional roles of microbiome members; and impacts on host health. However, investigations have been severely limited by the availability of phylogenetically diverse microbes in culture as efforts to recover and isolate marine microbes from environmental samples struggle to overcome microbial culturing challenges (Dance 2020 ). This study is also the first to develop a comprehensive collection of cultured bacteria associated with the skin of healthy cetaceans, supplementing the growing body of literature examining the culturable bacteria of cetacean respiratory and digestive systems (Buck et al . 2006 ; Venn-Watson, Smith and Jensen 2008 ; Morris et al . 2011 ; Godoy-Vitorino et al . 2017 ). Additionally, efforts outlined here yielded a cultured microbial collection, from coral tissue and mucus, greater in phylogenetic diversity and total number of strains recovered than catalogued previously (Lampert et al . 2006 ; Chimetto et al . 2008 ; Shnit-Orland and Kushmaro 2009 ; Galkiewicz et al . 2011 ). Moreover, 36% of the isolates were unique to a single animal surface, indicating our culturing technique is capable of capturing highly diverse bacterial isolates from diverse host organisms. Strain specificity is a common feature in many symbiotic relationships, and strain-level genetic differences can exist in symbionts, particularly for bacteria (Bongrand and Ruby 2019 ; Apprill 2020 ). Culture-independent analyses of coral-associated bacteria find sub-genus phylogenetic clusters represented exclusively in one clade of Scleractinian coral, including strains within the genus Ruegeria , isolated in this study (Huggett and Apprill 2018 ). The potential for strain specificity to host origin is found in many instances in this study, since OTUs grouped by 99% similarity within the same bacterial genus were isolated solely from separate hosts (Fig. 3 ). These potential host-specific strains include bacteria within the genera Ruegeria , Paracoccus , Maricaulis , Pseudomonas , Pseudoalteromonas , Idiomarina and Gramella . A total of four of these genera— Ruegeria , Paracoccus , Idiomarina and Gramella—m ay contain distinct bacterial strains isolated exclusively from different coral species in the same environment. Further examination is needed to indicate if these associations are host-driven rather than environment-driven, and understanding the factors driving host-specific strain association requires interrogation of the host microscale environmental and bacterial species-specific metabolic characteristics and exchange to more fully resolve stable host-bacterial relationships (Ziemert et al . 2014 ; Mönnich et al . 2020 ). The microbial library produced in this study is poised to investigate the potential for metabolically selective, strain-specific symbioses. Future work should include a FISH-microscopy based approach to elucidate the micro-scale interactions these cells have with their animal surfaces. Acquisition of bacteria detected in previous culture-independent efforts Cetacean skin microbes isolated here resemble the main phyla identified through culture-independent methods from Tursiops sp. blow (Bik et al . 2016 ; Nelson et al . 2019 ), skin microbiota from Tursiops truncatus and Orcinus orca (Chiarello et al . 2017 ) and the oral microbiome of Delphinus delphis, Stenella coeruleoalba and Phocoena phocoena (Soares-Castro et al . 2019 ) which are also dominated by Proteobacteria, Firmicutes and Bacteroidetes. Additionally, at the family level, our study was successful in isolating representatives from the Roseobacter clade, one of the most dominant groups of surface colonizers previously identified from the skin of killer whales (Chiarello et al . 2017 ). However, isolates that we obtained from humpback whale skin through cultivation are phylogenetically distinct from core members of the skin of this same species as identified through previous cultivation-independent surveys. Although phylogenetic analyses of humpback whale microbiomes reveal the ubiquitous and abundant presence of Tenacibaculum sp. and Psychrobacter sp. across humpback whale populations, other less-abundant microbial genera exhibit temporal shifts in presence and absence, throughout the foraging season and between whales inhabiting different geographic regions (Apprill et al . 2014 ; Bierlich et al . 2017 ). Of these less abundant genera, we did have success in cultivating 14 of the 43 (approx. 32.5%) differentially present bacterial genera found to be distinct from the core microbial members (14 of 49 (29%) of all bacterial genera identified); (Bierlich et al . 2017 ). Understanding these temporally and geographically variable bacterial–host associations can provide insight into the ecological roles of the microorganisms and the biochemical and environmental drivers of microbial community shifts. We were also able to culture 268 bacterial isolates from coral surface swabs and syringe samples (115 and 153 isolates, respectively). Of these isolates, the most frequently recovered species from coral swab and syringe samples were members in the orders Flavobacteriales, Bacillales, Rhodobacterales, Alteromonadales and Sphingomonadales (Fig. 2 ). Each of these bacterial orders have been previously identified in surveys of 16S rRNA genes from the microbiomes of Caribbean corals (Sunagawa, Woodley and Medina 2010 ; Morrow et al . 2012 ; Kimes et al . 2013 ; Ainsworth et al . 2015 ; Apprill, Weber and Santoro 2016 ). Furthermore, representatives of the family Rhodobacteraceae and the genera Bacillus, Erythrobacter, Gramella, Marinobacter, Neptuniibacter and Oceanobacillus were recovered from all three coral species studied, and each of these genera have been documented on P. astreoides and A. palmata (Fig. 3 and Table S2, Supporting Information; Sharp, Distel and Paul 2012 ; McDevitt-Irwin et al . 2017 ). Roseobacter clade-associated bacteria, members of the family Rhodobacteraceae, and the genus Marinobacter are known to associate with P. astreoides throughout the coral life cycle, and these bacteria may be passed down from coral parent to larvae (Sharp, Distel and Paul 2012 ), which could be a mechanism to retain select metabolically distinct bacterial strains. Marine Roseobacter strains are known to facilitate coral settlement (Sharp et al . 2015 ). Additionally, we were able to isolate representatives from Halomonas , Hyphomonas , Microbulbifer, Paenibacillus, Photobacterium, Pseudovibrio and Oceanicola (Table S2, Supporting Information) frequently found on both healthy and diseased P. astreoides colonies (McKew et al . 2012 ; Meyer, Paul and Teplitski 2014 ; Staley et al . 2017 ). Multivariate analyses highlight variation in cultured microbial communities The non-metric multidimensional (NMDS) ordination and perMANOVA tests of group differences indicated that cultured microbial community composition differs across sampled species rather than across culture condition (Fig. 4A ). However, the global group difference across sampled species is largely driven by the difference between M. novaeangliae and the three coral species. Significant differences between coral species are likely driven by differences in multivariate dispersion, since cultured microbial communities from A. palmata samples were relatively similar and less disperse between samples. This result suggests that variation in cultured microbial community composition is associated with differences in environment and biological characteristics of the species surface, rather than culture media composition, like nutrient availability or the presence of quorum sensing molecules, for example. Despite their use for targeted cultivation, media variations did not select for particular bacterial clades (Fig. 4A ). Additionally, the AC media, designed to select for microorganisms within the Actinobacteria phylum, did not select for any bacteria within this phylum and instead allowed for the growth of three OTUs within the order Bacillales. The Bacillales order contains many spore-producing organisms. It is possible that the heat shock may have selected for heat-resistant spores that structurally, metabolically and functionally differ from vegetative cells (Gopal et al . 2015 ). Due to time constraints, microbial biomass from D. leucas samples were cryopreserved before colony picking, likely selecting for microbes with rapid growth. The bias induced by the microbial isolation method for the D. leucas samples produced low microbial OTU richness and resulted in a less robust dataset that inhibited model convergence in multivariate analyses. This bias underscores the importance of immediate microbial isolation upon surface sampling. The NMDS ordination and perMANOVA test of group differences indicated that cultured microbial community composition differs between coral sampling method, with swab samples enriched with microbes in the Oceanospirillaceae, Rhodobaccteraceae and Alteromonadaceae families and syringe samples enriched with microbes from Bacillaceae and Staphylococcaceae families (Fig. 5 ). The syringe sampling method more likely recovers microorganisms contained in the coral mucus, while the swabbing method more likely recovers microorganisms attached to the coral surface. Previous work suggests the existence of specific coral mucus-associated bacteria that regulate mucus layer bacterial populations through antimicrobial activities (Ritchie and Smith 1997 ; Shnit-Orland and Kushmaro 2009 ), and future inquiries should investigate the mechanisms that mediate divergent microbial communities across coral surface micro-environments. Additionally, coral syringe samples maintain greater phylogenetic diversity than coral swabs, potentially because the mucosal layer is frequently colonized by bacteria carried to the animal via sediments (Apprill, Weber and Santoro 2016 ; Glasl, Herndl and Frade 2016 ). This greater phylogenetic diversity in the syringe samples (e.g. mucosal layer only) is reflected in the orders of bacteria most frequently recovered from each set of samples (Fig. 2 ). Greater microbial diversity in the coral mucus-only samples may also be due to the syringe method of collection. Coral micro-habitats sampled using this method may dislodge Symbiodinaceae from the coral surface, and therefore bacteria associated with these cells may also be represented in the data. Culture-dependent design introduces bias in recovered microorganisms Many of the culturable bacteria from both whale and coral species included Alteromonas sp. (Fig. 2 ), which are globally distributed marine copiotrophic bacteria that dominate in heterotrophic blooms and outcompete other bacteria due to their rapid growth when organic nutrients are readily available (Math et al . 2012 ). The ubiquity of this genus in our whale samples could have resulted due to the nutrient-rich media base used during cultivation, which has previously led to an overestimation of their abundance in the environment due to the ease at which the cells form colonies on agar plates and the voracity with which they take advantage of sporadic inputs of organic matter (Behringer et al . 2018 ). However, our data indicate that a fraction of the 127 Alteromonas strains isolated here were only found on coral or on cetaceans, indicating potential host specificity for particular strains. Additionally, isolates that are phylogenetically similar can exhibit different biochemical profiles and functional roles (Lampert et al . 2006 ). Culture independent studies by themselves would likely fail to link varied contributions of similar strains to larger implications, such as their effects on host fitness. Future studies should integrate more diverse media types, particularly low nutrient conditions, to target more slower growing microorganisms that might have been overlooked in this study."
} | 6,075 |
24024617 | null | s2 | 2,967 | {
"abstract": "This study provides a detailed secondary structural characterization of major ampullate dragline silk from Latrodectus hesperus (black widow) spiders. X-ray diffraction results show that the structure of black widow major ampullate silk fibers is comprised of stacked β-sheet nanocrystallites oriented parallel to the fiber axis and an amorphous region with oriented (anisotropic) and isotropic components. The combination of two-dimensional (2D) (13)C-(13)C through-space and through-bond solid-state NMR experiments provide chemical shifts that are used to determine detailed information about the amino acid motif secondary structure in black widow spider dragline silk. Individual amino acids are incorporated into different repetitive motifs that make up the majority of this protein-based biopolymer. From the solid-state NMR measurements, we assign distinct secondary conformations to each repetitive amino acid motif and, hence, to the amino acids that make up the motifs. Specifically, alanine is incorporated in β-sheet (poly(Alan) and poly(Gly-Ala)), 3(1)-helix (poly(Gly-Gly-Xaa), and α-helix (poly(Gln-Gln-Ala-Tyr)) components. Glycine is determined to be in β-sheet (poly(Gly-Ala)) and 3(1)-helical (poly(Gly-Gly-X(aa))) regions, while serine is present in β-sheet (poly(Gly-Ala-Ser)), 3(1)-helix (poly(Gly-Gly-Ser)), and β-turn (poly(Gly-Pro-Ser)) structures. These various motif-specific secondary structural elements are quantitatively correlated to the primary amino acid sequence of major ampullate spidroin 1 and 2 (MaSp1 and MaSp2) and are shown to form a self-consistent model for black widow dragline silk."
} | 407 |
36616539 | PMC9824380 | pmc | 2,968 | {
"abstract": "The aging and damage of artificial skin materials for artificial intelligence robots are technical problems that need to be solved urgently in their application. In this work, poly (vinylidene fluoride) (PVDF) fibers containing a liquid agent were fabricated directly as biomimetic microvasculars, which were mixed in a glycol–polyvinyl alcohol–gelatin network gel to form biomimetic self-healing artificial skin composites. The self-healing agent was a uniform-viscous buffer solution composed of phosphoric acid, acetic acid, and sodium carboxymethyl cellulose (CMC-Na), which was mixed under 40 °C. Microstructure analysis showed that the fiber surface was smooth and the diameter was uniform. SEM images of the fiber cross-sections showed that there were uniformly distributed voids. With the extension of time, there was no phenomenon of interface separation after the liquid agent diffused into the matrix through the fiber cavity. The entire process of self-healing was observed and determined including fiber breakage and the agent diffusion steps. XRD and FT–IR results indicated that the self-healing agent could enter the matrix material through fiber damage or release and it chemically reacted with the matrix material, thereby changing the chemical structure of the damaged matrix. Self-healing behavior analysis of the artificial skin indicated that its self-healing efficiency increased to an impressive 97.0% with the increase in temperature to 45 °C.",
"conclusion": "4. Conclusions The self-healing microvasculars can repair the skin of artificial intelligence ro-bots in times when the skin is damaged. At the same time, gland tubes in the skin can secrete oil, moisturize the skin, and delay skin aging. In this work, a self-healing glycol–PVC–gelatin network gel artificial skin was successfully fabricated composited with PVDF fibers containing a liquid agent. This self-healing artificial skin can realize self-healing and anti-aging functions, and it is expected to be applied to artificial intelligence robot skins to prolong the service life of the artificial skin and help it to better cope with various use scenarios. It was found that the artificial skin could achieve both self-healing and anti-aging functions, providing reduced cost and a new direction for the practicality of artificial skin. From the mentioned preliminary results, the following conclusions can be drawn: (1) The microstructure of the fiber/artificial skin was studied by various methods in this study. It was found that the inner and outer surfaces of the fibers were smooth without defects. The cross-section of the fiber is a regular circle. There are many small holes in the fiber wall material with uniform distribution and pore diameter. After fracturing under different conditions, the hollow fiber and the gel are combined face-to-face. The observation results show that there is no phase separation between the fiber and the matrix material. The interface structure is stable and shall not be damaged. (2) The self-healing process was determined including healing agent diffusion and crack healing. Firstly, the artificial skin samples were kept at 25 °C to observe the diffusion behavior of the healing agent into the gel, which was photographed at six-day intervals. As time progressed, the healing agent spread to both sides, with the fibers as the axis. Within 18 days, the healing agent had spread throughout the gel, and the color of the gel gradually deepened over the subsequent period. There was no shrinkage in the gel size and no surface cracking during this period, indicating that self-healing of the gel by the fibers did occur. (3) The XRD and FT–IR results indicated that the self-healing agent could enter the matrix material through fiber damage or release and that it chemically reacted with the matrix material, thereby changing the chemical structure of the damaged matrix. (4) Considering the tensile strength ratio before and after healing, the self-healing efficiency of the artificial skin composite was measured by a tensile fracture test. In order to simplify the complexity of the experiment, a single optical fiber was embedded in the matrix parallel to the tensile direction. It was found that the temperature greatly affected the self-healing efficiency.",
"introduction": "1. Introduction As one of the most important organs of the human body, skin not only has the basic function of protecting the body, secreting oils, and excreting metabolites, but it is also an important system for human perception, interaction, and communication with the outside world [ 1 , 2 ]. Based on the functions of human skin, various artificial skins with corresponding functions have been developed. These bionics have extensive applications in body state detection, artificial limbs and robots, and human–computer interaction [ 3 , 4 , 5 ]. With the development of humanoid robots, more and more research has been carried out on artificial intelligence robot skin. With the gradual application of the artificial skin of robots in different fields, many problems have been revealed in the process, among which the breakage of artificial skin during use has become the primary problem that hinders the large-scale application of artificial skin. In practice, the broken area of artificial skin is small compared to the overall area, and replacing large areas of artificial skin can be costly financially and time-consuming [ 6 , 7 , 8 ]. It is well known that flexibility, tensile properties, and mechanical durability affect the service life of artificial skin [ 9 ]. Despite continuous refinement and improvement in the abovementioned directions, the problem of breakage repair in the use of artificial skin has never been well solved and has largely hindered its practical application [ 10 , 11 , 12 ]. Therefore, the ability of rapid repair after in-use damage is therefore key to the practical application of artificial skin. Self-healing technology has the advantages of low cost, long aging, easy preparation, and multiple materials selection. Nowadays, self-healing artificial skin has a broad application potential in the field of artificial limbs and artificial intelligent robots [ 13 , 14 ]. To prolong the service life of artificial skin, reversible-bond self-healing, microcapsule self-healing, and microvascular self-healing have been gradually explored [ 15 , 16 , 17 , 18 ]. Research shows that most reversibly bonded artificial skins, while solving the problem of self-healing, take a long time to repair after breakage and require a relatively demanding repair environment; these shortcomings limit the practical application of this class of artificial skin [ 19 , 20 ]. The appearance of microcapsule technology has enhanced the healing of microcracks in artificial skin and has enhanced the anti-aging effect of artificial skin. This technology enables it to achieve both anti-aging and self-healing. Nowadays, it is relatively mature, which has been proven in the fields of road and composite self-healing [ 21 , 22 , 23 ]. Reasonable design of the wall material can also make the core material continuously release to achieve the purpose of self-nutrition [ 24 ]. Low core content of microcapsules is less capable of repair in the face of larger damage to the artificial skin. Moreover, empty microcapsules tend to form stress concentrations after the core material is released, resulting in a reduction in the mechanical properties of the substrate [ 25 ]. Therefore, the direction of research to prolong the life of artificial skin should take account of self-healing properties and various mechanical properties [ 26 ]. In the face of the problems of reversible adhesive artificial skin in artificial intelligence robots prone to aging and microencapsulated artificial skin prone to stress concentrations due to a small amount of microcapsules, hollow fiber technology can be used not only as a reinforcement material to improve the mechanical properties of artificial skin, but also as a release healing agent to enable the artificial skin to self-heal after damage. This technology has been used successfully in the self-healing of asphalt, membranes, polymers, and other materials [ 27 , 28 , 29 ]. In previous work, self-healing artificial skin using fibers has been successfully designed with improved mechanical properties, which helps to promote the widespread application of artificial skin especially in artificial intelligence robot field [ 29 , 30 , 31 ]. A propylene glycol–polyvinyl alcohol–gelatin network gel is used as the matrix and polyvinylidene fluoride hollow fibers containing a healing agent are used as the reinforcement, with the fibers bonded to the gel matrix by being pre-laid into a mold. In previous studies, we have investigated the mechanical properties of artificial skin and the effect of the pH and the content of the restorative solution on the effectiveness of self-healing [ 32 ]. Based on the above research background and research basis, the microstructure and self-healing capability of artificial skin composites using biomimetic fibers containing a healing agent were investigated from the perspective of physics and chemistry. Microstructure analysis analyzed the fiber surface, distribution, and interface of this composite. The entire process of self-healing was observed and determined including fiber breakage and the agent diffusion steps. The self-healing agent release was detected based on the chemical reaction. Last, the temperature effect on self-healing efficiency was also investigated.",
"discussion": "3. Results and Discussion 3.1. Microstructure of Fibers Figure 1 a–c shows the SEM cross-section morphology of hollow PVDF fibers from different angles of observation. From the figures, it can be roughly distinguished that the diameter of the fiber is 1 mm and the thickness of the fiber is 150 μm. The internal and external surfaces of the fibers are smooth without defects. The fiber cross-section presents a regular circle. It can also be seen from the shape of the cross-section that the fiber thickness is uniform without large deviation ( Figure 1 a). There are many small cavities in the fiber wall material with an average diameter of 1 μm, which is evenly distributed and has a uniform pore size ( Figure 1 b). The pore size is similar to the reported result [ 30 ]. Previous studies have also shown that these small cavities can allow the liquid inside the fiber to slowly release into the matrix material around the fiber [ 30 , 31 ]. The fiber has a certain mechanical strength and can be deformed after compression. From the shape of fiber compression ( Figure 1 c), it can be preliminarily determined that it can withstand a certain amount of pressure without damage. 3.2. Interface Morphology between the Gel Matrix and the Fibers The interface problem of composite materials is a factor that cannot be neglected in materials research. The close bonding of the interface of each component in a composite material is the key to the overall performance of the composite material and the various materials that make it up. Therefore, for new composite materials, the interface stability between the gel and fibers of artificial skin is worth investigating. When the artificial skin is damaged with interface separation, the gel and fibers will not break simultaneously. However, the mechanical properties of the composite sample will be reduced along with the following two conditions. (A) The fibers break, and the gel does not break. The healing agent inside the fiber flows out and reacts with the gel, causing the healing agent to be wasted. (B) The gel is damaged, and the fibers are not broken. This situation is even worse than situation A and can leave the gel wound without timely healing and cause further wound enlargement during subsequent use of the artificial skin. To address this vital issue, SEM was used to analyze the interface binding stability between the artificial skin fibers and gel. Interface damage is often very complex. In this study, the interface stability under steady-state conditions, stress damage, and material leakage are mainly considered. Figure 2 shows the interfacial bonding of the empty fiber and the gel after fracturing under different conditions. For better observation, we have post-colored these images, where the fiber part is colored yellow and the gel part is colored green. Figure 2 a,b shows the SEM cross-section morphologies of an artificial skin sample observation from different angles. The cross-section is cut by freezing with a sharp slicer, and its surface is flat and smooth. The place pointed by the red arrow is the interface between the fiber and the gel. There is no apparent separation between the two interfaces, and the fracture surface is relatively neat and basically in the same plane. Therefore, it can be concluded that no two-phase interface separation between the fiber and the gel occurs in the face of tensile and shear forces exerted by sharp objects. Secondly, Figure 2 c,d shows the artificial skin after it was stretched and fractured at the aperture. It can be noted that the fracture surfaces of the fiber and the gel are not at the same level, which is due to the different modulus of the two. Even so, the gel remains attached to the fibers, and it can be seen that even when the artificial skin is stretched with forces parallel to the interface, the gel remains firmly attached to the fibers without phase separation. Thirdly, Figure 2 e,f shows the fracture of the artificial skin after fracture due to bending. It can be seen clearly that the fiber cross-section is damaged by repeated bending without phase separation between the fiber and the gel ( Figure 2 e). The gel remains intact. The artificial skin does not break after bending, but the gel appeared to fall off ( Figure 2 f). After careful observation, it can be clearly confirmed that the prominent part in the picture is the fiber, but the fiber surface is still wrapped by fiber. Therefore, it can be considered that the interface between the fiber and the gel is quite stable even after the fiber is bent. Figure 3 shows SEM images of the artificial skin after a period of self-healing and treatment breakage, in which the fibers are stained yellow, the gel is dyed green, and the healing agent inside the fibers is stained red. In Figure 3 a–c, the healing agent has penetrated the micro-porous structure. The red arrow indicates the direction of liquid penetration. The micro-porous structure on the fiber wall in the figure was filled with the healing agent, and there was no phase separation between the gel and the fiber, indicating that the penetration of the healing agent did not affect the adhesion of the gel to the fiber. Figure 3 d–f shows the whole process of liquid penetration into the outer matrix of the fiber through micro-pores. It can even be found that the healing agent infiltrates the gel through the micro-porous structure and fuses with the matrix gel. The reacted gel penetrates deeper into the fiber along the micro-porous structure, further strengthening the interfacial bond. 3.3. Determination of the Self-Healing Process There are numerous ways to identify self-healing, and direct observation is one of the most visual methods. The healing agent within the fibers in the artificial skin sample in Figure 4 was stained with methyl violet. The samples were left at 25 °C to observe the healing behavior of the healing agent on the gel, which was photographed at six-day intervals. As time progressed, the healing agent spread to both sides with the fibers as the axis ( Figure 4 a–d). By day 18, the healing agent had spread throughout the gel, and the color of the gel gradually deepened over the subsequent period ( Figure 4 e–h). There was no shrinkage in the gel size and no surface cracking during this period, indicating that self-healing of the gel by the fibers did occur. Figure 5 a shows the FT–IR curves of the exact location of the artificial skin on day 0, day 6, day 12, and day 18 of the self-healing observation. These spectra have similar trends and peak positions; there are only minor differences in the peak position, intensity, and shape. The 3100–3700 cm −1 range has distinct peaks in all four spectra. Still, the peak intensity decreases sequentially with increasing self-healing time, which is characteristic of the O-H stretching vibration in the hydroxyl group, which reacts with boric acid to form borate ester bonds as the reflection proceeds. As the self-healing reaction proceeded, several distinct peaks of borates in the curve became more apparent, such as 1340 cm −1 and 1420 cm −1 (asymmetric stretching vibration peaks of the B-O-C complex). The characteristic peak of the spectrum at 1083 cm −1 was related to the vibrational stretching of the C-OH bond in the artificial skin and the matrix, and the C-OH content decreased with the self-healing reaction. The 1083 cm −1 characteristic peak also falls. Figure 5 b shows the XRD curves of the same location of the artificial skin on day 0, day 6, day 12, and day 18 of the self-healing observation. The curves show two distinct diffraction peaks at 2θ = 19.4° and 2θ = 22.0°, which correspond to the orthogonal lattice structure of the semi-crystalline PVA. With the self-healing of artificial skin, the diffraction peak at 2θ = 19.4° shifted and weakened. The diffraction peak at 2θ = 22.0° disappeared, indicating that the hydroxyl group of PVA reacted strongly with the borate, and the borate ester bond formed between them destroyed the structure of the crystalline region of PVA. 3.4. Measurement of the Mechanical Strength of the Composites at Different Temperatures Temperature is a significant factor limiting the use of artificial skin. Temperature affects the mechanical properties of the artificial skin during regular use and its ability to heal itself. This experiment simulates the healing behavior of artificial skin at different temperatures including low temperature (5 °C), room temperature (25 °C), and high temperature (45 °C). Figure 6 shows a set of comparative experiments on the maximum stretching of artificial skin at 5 °C. Figure 6 (a 1 ) is the artificial skin without any treatment after destruction (named as A 1 ), Figure 6 (a 2 ) is the artificial skin with self-healing after collapse (named as A 2 ), and Figure 6 (a 3 ) is the artificial skin gel without any treatment (named as A 3 ). It is evident that the maximum stretch of A 2 is much larger than that of A 1 , and the stretch length of A 2 is about 60% of that of A 3 . Figure 6 b shows the stress–strain curves of the three groups of samples. Regarding strain, A 2 (320 KPa) had much higher stress than A 3 (127 KPa) during the stretching process. The lower temperature affected the performance of the artificial skin, making the performance of A 3 lower, while the low temperature weakened the large elasticity of A 2 while giving it greater strength. In terms of strain, the maximum strain of A 2 was 1.2, the maximum strain of A 3 was 1.96, and the repair efficiency of the artificial skin at 5 °C was 61.2%. Figure 6 c shows the comparative experiments of the maximum tensile degree of the artificial skin at an ambient temperature of 25 °C. The samples were treated the same way as the experiments at 5 °C named as C 1 , C 2 , and C 3 as shown in Figure 6 (c 1 –c 3 ). Due to the increase in temperature, the difference between C 3 (120 KPa) and C 2 (130 KPa) in stress is insignificant. In terms of strain, the maximum strain was 1.3 for C 2 and 1.75 for C 3 . The self-healing efficiency of the artificial skin at 25 °C was 74.3%, which shows that the increase in temperature enhanced the elasticity of the repaired artificial skin. However, it reduced the strength of the healed skin. The ambient temperature of the comparison experiment of the maximum stretching degree of the artificial skin in Figure 6 e is 45 °C, and the sample processing and experimental methods are the same as the previous two groups of experiments ( Figure 6 (e 1 –e 3 ) are named E 1 , E 2 and E 3 ). Figure 6 f shows the stress–strain curves of the three groups of samples. The stress of E 2 (185 KPa) is slightly higher than that of E 3 (175 KPa), the strain of E 2 (1.65) is almost the same as that of E 3 (1.7 KPa), and the repair efficiency reaches 97.0%. At higher temperatures, the self-healing reaction is more rapid and complete, so that the mechanical behavior of the repaired artificial skin can be restored to the previous level. According to the performance of the artificial skin at different temperatures, the A 3 , B 3 , and C 3 versions of artificial skin at various temperatures did not differ much, which is because the artificial skin gel matrix was prepared with the addition of propanetriol, a natural antifreeze agent, which ensured that the artificial skin maintained a more stable performance at different temperatures. The performance of temperature has a significant impact on the self-healing ability of the artificial skin. The temperature affects the self-healing reaction of the healing agent and the gel. Higher temperatures bring higher self-healing efficiency; at the same time, the temperature also affects the gel after the reaction. Higher temperatures result in a higher reaction degree, which indirectly improves efficiency."
} | 5,353 |
37425631 | PMC10323715 | pmc | 2,971 | {
"abstract": "Microbial fuel cells (MFCs) are widely acknowledged to be a promising eco-friendly abatement technology of pollutants, and are capable of generating electricity. However, the poor mass transfer and reaction rate in MFCs significantly decrease their treatment capacity for contaminants, especially hydrophobic substances. The present work developed a novel MFC integrated with an airlift (ALR) reactor using a polypyrrole modified anode to promote the bioaccessibility of gaseous o -xylene and attachment of microorganisms. The results indicated that the established ALR-MFC system showed excellent elimination capability, with removal efficiency exceeding 84% even at high o -xylene concentration (1600 mg m −3 ). The maximum output voltage of 0.549 V and power density of 13.16 mW m −2 obtained by the Monod-type model were approximately twice and sixfold higher than that of a conventional MFC, respectively. According to the microbial community analysis, the superior performances of the ALR-MFC in terms of o -xylene removal and power generation were mainly ascribed to the enrichment of degrader ( i.e. Shinella ) and electrochemical active bacteria ( i.e. Proteiniphilum ). Moreover, the electricity generation of the ALR-MFC did not decrease at a high O 2 concentration, as O 2 was conducive to o -xylene degradation and electron release. The supplication of an external carbon source such as sodium acetate (NaAc) was conducive to increasing output voltage and coulombic efficiency. The electrochemical analysis revealed that released electrons can be transmitted with the action of NADH dehydrogenase to OmcZ, OmcS, and OmcA outer membrane proteins via a direct or indirect pathway, and ended up transferring to the anode directly.",
"conclusion": "4 Conclusions This work established a MFC coupling system with AL reactor for efficient pollutant elimination and power generation. Both the o -xylene removal efficiency and the output voltage of the system were enhanced significantly, resulted from the existence of abundant bacteria with capability of degrading o -xylene and producing electricity in biofilm. The impacts of operation conditions such as oxygen content and external carbon source on reactor performance were also evaluated. A decrease of 8.2% in o -xylene removal efficiency appeared as the oxygen content varied from 21% to 10%, demonstrating the enhancement effect of oxygen for o -xylene degradation. Besides, except for the coulombic efficiency improvement, the output voltage achieved as high as 0.606 V with NaAc addition, 10% higher than the reported average voltage (0.55 V). The comparison of CV curves measured before and after running process indicated that electrons can be transferred from biofilm to anode directly via inner membrane protein such as NADH dehydrogenase as well as outer membranes proteins such as OmcZ, OmcS, and OmcA.",
"introduction": "1 Introduction Microbial fuel cells (MFCs) are an emerging green energy technology that can generate electricity directly during the efficient degradation of contaminants using microbes instead of expensive materials as catalysts. 1 The anode is the fundamental component in MFCs, and its characteristics play a critical role in electron transfer and the redox reaction, and have been the primary reason underlying the low efficiency in various MFC prototypes. 2 The common carbon-based materials, including carbon paper, 3 carbon cloth, 4 carbon felt, 5 graphite felt, 6 graphite brush, 7 and graphite sheet, 8 are widely applied in MFCs owing to their stability and affordability. However, their inherently hydrophobic property is inconducive for microbial adhesion, contributing to poor electron transfer capacity. 2 Many attempts have been made with respect to modification in enhancing MFCs performance and viability. Polypyrrole (PPy) is an attractive polymer used for electrode modification due to its excellent biocompatibility and conductivity. PPy deposition improves the surface roughness and electrical conductivity of the anode, and thus the maximum power density of the MFC was 2.3 times that using unmodified carbon felt. 9 Besides, substrate mass transfer is another important consideration in the design of reactors, which dominates MFCs performance. 10 Integrated systems are established to improve the insufficient gas–liquid mass transfer of contaminants, especially hydrophobic volatile organic compounds (VOCs). MFC with biological flowing reactor achieved the removal efficiency of ethyl acetate as high as 93.8%. 11 The airlift reactor (ALR) has been extensively utilized in wastewater treatment, air purification, petroleum desulfurization, and other fields due to its exceptional mass and heat transfer rate, 12 ALR was proved to enhance fluid turbulence and accelerate mass transfer rates, obtaining 95.4% o -xylene removal efficiency. 13 The removal efficiency and stability of the airlift packing bioreactor was significantly higher than the ALR in removing dichloromethane and toluene, which means that the set of obstacles is profitable in VOCs removal. 14 Moreover, the addition of external carbon source has been confirmed as another effective strategy to enhance system performance. The maximum power density of MFC can be increased from 9.1 to 28.3 W m −3 using glucose as an additional substrate. 15 Oxygen content is a vital contributing parameter for operation capacity of coupled system due to the sensitivity of microorganisms to oxygen, 16 which has not been investigated deeply yet. In this work, we aimed to develop an ALR-MFC system to overcome the mass transfer limitation in hydrophobic o -xylene removal, and the anode modified with PPy was employed to supply more microbial attachment sites and accelerate electron transfer. The system performances including pollutant elimination and electricity generation under different o -xylene concentrations were monitored. Also, the effects of oxygen content and external carbon source on reactor operation were evaluated. More detailed, the enhancement mechanism of integrated construction for o -xylene degradation was explored based on analysing microbial structure and electrochemical characteristic of the biofilm.",
"discussion": "3 Results and discussion 3.1 \n O -xylene removal performance of ALR-MFC The removal efficiency of ALR-MFC maintained at 92% at o -xylene concentration lower than 800 mg m −3 and then gradually reduced to 84%, while the mineralization efficiency displayed the sustained declining trend as o -xylene concentration increased ( Fig. 2 ). A more significant decrease in mineralization efficiency compared to that in removal efficiency at all tested o -xylene concentrations was contributed to the recalcitrance of o -xylene. 20 Normally, the degradation of o -xylene will be terminated before the step of ring-opening. The removal efficiency and elimination capacity of ALR-MFC were 84% and 57.6 g m −3 h −1 at o -xylene load of 67.7 g m −3 h −1 , which were 21.7% and 91.9% greater than that of the traditional biological filter obtained at lower pollutant load, respectively. 21 Moreover, the o -xylene elimination behaviour of the reactor was also superior, compared to two-chambered MFC whose removal efficiency and elimination capacity were 78% and 4.3 g m −3 h −1 at 5 g m −3 h −1 o -xylene load, 1 respectively, possibly ascribed to the occurrence of stronger gas–liquid mass transfer in our established system. Fig. 2 The effect of o -xylene concentration on removal and mineralization efficiencies of ALR-MFC reactor. 3.2 Power generation performance of ALR-MFC As shown in Fig. 3 , the Monod-type model was used to describe the relationship between the power generation performance and the o -xylene concentration. The power generation capacity of ALR-MFC displayed a continuous improvement with inlet gas concentration rising from 200 to 1600 mg m −3 . It should be noted that the output voltage increased slowly when the inlet o -xylene concentration over than 1000 mg m −3 . The results of the model fitting ( R 2 = 0.997) indicated that the maximum generated voltage was 0.549 V, more than two times higher than that of ordinary MFC reactor. 1 Larger bioanode surface was one of the primary contributors to the greater output voltage. The maximum power density of ALR-MFC achieved as high as 13.16 mW m −2 , approximately sixfold better than that of the reported data (2.3 mW m −2 ). 22 The significant advantages on output voltage and elimination capacity of developed ALR-MFC system demonstrated that the reactor has great potential for efficient conversion from chemical energy of o -xylene into electrical energy. Fig. 3 Monod kinetics modelling of output voltage and power density with respect to o -xylene concentration. The coulombic efficiency of ALR-MFC in Fig. 4 revealed that the largest coulombic efficiency of 5.37% can be attained at 200 mg m −3 o -xylene, and then decreased gradually under higher pollution load condition. Although the coulombic efficiency was over twice greater than the MFC feeding with toluene (2.6%), 23 which was poorer than that of MFCs supplied by acetate (coulombic efficiency = 31%), 24 and glucose (coulombic efficiency = 37%). 25 The results emphasized the importance of carbon source types to reactor performance, and the relevant explorations have been conducted in the later section. Fig. 4 Effect of o -xylene concentration on coulombic efficiency of ALR-MFC. 3.3 Microbial community in biofilm of ALR-MFC The succession of microbial community structure before and after operation was shown in Fig. 5 . The origin microbial community was mainly consisted of Acidovorax (38.05%), Brevundimonas (28.6%) and Dysgonomonas mossii (7.9%), accounting for 74.55% of the total microbes. The power generation relied on the genus, including Acidovorax ( Table 1 ) and Dysgonomonas mossii . The dedicated o -xylene degraders such as Brevundimonas and Comamonas also had ability to degrade toluene and p -nitrophenol. 26,27 For long-term operation, the relative abundances of Proteiniphilum , Shinella , and Dokdonella increased markedly. Both Shinella and Dokdonella can degrade toluene and o -xylene compounds simultaneously. 28,29 Moreover, previous report has illustrated that Proteiniphilum played an essential role in efficient electricity generation. 30 Alicycliphilus was in capacity acquiring and consuming electrons, 31,32 lowering the power density. 22 Other bacteria, such as Leifsonia and Mycobacterium played synergistic roles in o -xylene degradation in the bioanode. Fig. 5 Microbial community in ALR-MFC reactor in start-up stage and after 150 days operation. Main genus of anode biofilm in start-up stage and after 150 days operation a Genus Proportion (%) Main function Electron utilization Ref. Before operation After operation \n Acidovorax \n 38.05 <0.3 A & B Electron consumption \n 33 \n \n Proteiniphilum \n <0.3 25.37 A Electron consumption \n 30 \n \n Brevundimonas \n 28.60 <0.3 B — \n 26 \n \n Mycobacterium \n <0.3 6.73 B — \n 34 \n \n Shinella \n <0.3 12.96 B — \n 29 \n \n Dysgonomonas mossii \n 7.9 <0.3 A Electron consumption \n 35 \n \n Dokdonella \n <0.3 8.87 A & B Direct electron transfer \n 28 and 36 \n Comamonas \n 6.72 <0.3 A & B Direct electron transfer \n 27 and 37 \n Alicycliphilus \n 3.12 <0.3 A Electron consumption \n 38 and 39 \n Leifsonia \n <0.3 3.45 B — \n 40 \n \n Dysgonomonas \n 1.79 <0.3 A & B Electron consumption \n 41 \n a A represents the main function of power generation, and B represents the main function of xylene degradation. 3.4 Effect of operating conditions on ALR-MFC performance 3.4.1 Inlet oxygen concentrations \n Fig. 6 shows that the oxygen content of o -xylene gas has a notable influence on system performance. In Fig. 6a , a decrease of 8.2% in removal efficiency appeared as the oxygen content varied from 21% to 10%, likely due to the activity inhabitation of aerobic predominant degrader such as Shinella under low oxygen concentration. 42 Differently, an increase of 9.7% in mineralization efficiency emerged with oxygen content decreasing, which was because of the activity promotion for some catabolic enzymes with high oxygen affinities. 43 The output voltages acquired at different oxygen concentrations indicated that generated voltage was not inhibited by a high oxygen concentration ( Fig. 6b ). Even though the existence of oxygen would induce the competition of electrons with the anode which was supposed to limit output voltage, oxygen could also promote the ring-opening and complete degradation of o -xylene to make up the voltage loss caused by oxygen in the mixed gas. Fig. 6 Effect of oxygen concentration on (a) o -xylene removal and mineralization efficiencies, and (b) output voltage of ALR-MFC reactor. 3.4.2 External carbon source NaAc with the concentration of 0.5 g L −1 was added into electrolyte to determine the function of external carbon during o -xylene degradation. As presented in Fig. 7a , both the removal efficiency and output voltage dropped after the addition of NaAc, which was caused by the higher bioavailability of NaAc and succession of microbial community. Notably, the output voltage raised to 0.606 V after cultivation for 16 d, exceeding the average voltage reported previously (0.55 V). 44 The improvement in output voltage and decline in removal efficiency suggested that NaAc can enhance the electricity generation performance but not the o -xylene elimination capacity of ALR-MFC. Additionally, the coulombic efficiency of system with NaAc shown in Fig. 7b was obviously higher than that of the system without NaAc shown in Fig. 4 . The coulombic efficiency was more than 6.9% at all measured o -xylene concentrations (150–1200 mg m −3 ), and the largest coulombic efficiency could reach 47.2%, verifying the enhancement effect of external carbon source on chemical energy to electricity energy. Fig. 7 (a) The removal efficiency and output voltage, and (b) coulombic efficiency of ALR-MFC reactor with 0.5 g per L NaAc addition. 3.5 Electrochemical characteristic CV tests were carried out to identify the redox-active components that may be involved in exocellular electron transfer (EET) between biofilm and anode. 45 The CV analyses were measured in the following two conditions, including after replacement of fresh medium and after 48 h of reactor operation. It can be seen from Fig. 8a that there was no significant difference between two CV curves, manifesting that microbial secretions did not participate in the EET process. Fig. 8 CV curves of (a) bioanode measured before and after 48 h operation, and (b) anode biofilm. The CV curve of anode biofilm was shown in Fig. 8b . The reduction peak at −0.360 to −0.040 V is probably pertained to OmcZ (−0.420 to −0.06 V), OmcS (−0.360 to −0.04 V) and OcmA (−0.350 to −0.08 V). 46 One characteristic oxidation peak at −0.420 to −0.300 V may be corresponded to NAD + /NADH (−0.477 V). 46 Based on the detected oxidation peaks and its potential corresponding proteins, the electron pathway involved in the o -xylene powered ALR-MFC was proposed in Fig. 9 . Firstly, o -xylene is oxidized to CO 2 , H + and electrons by anodic attached microorganisms. The released electrons and H + could be accepted by intracellular dehydrogenase to produce NADH, 47 which is then transferred H + to periplasm and simultaneously transfer electrons to cytochrome. The electrons were then transmitted to outer membrane proteins such as OmcZ, OmcS, and OmcA and finally taken up by anode directly or indirectly. 48 Fig. 9 Proposed electron transfer pathway between biofilm and anode."
} | 3,913 |
18093303 | PMC2242787 | pmc | 2,972 | {
"abstract": "Background According to the classical model of Macevicz and Oster, annual eusocial insects should show a clear dichotomous \"bang-bang\" strategy of resource allocation; colony fitness is maximised when a period of pure colony growth (exclusive production of workers) is followed by a single reproductive period characterised by the exclusive production of sexuals. However, in several species graded investment strategies with a simultaneous production of workers and sexuals have been observed. Such deviations from the \"bang-bang\" strategy are usually interpreted as an adaptive (bet-hedging) response to environmental fluctuations such as variation in season length or food availability. To generate predictions about the optimal investment pattern of insect colonies in fluctuating environments, we slightly modified Macevicz and Oster's classical model of annual colony dynamics and used a dynamic programming approach nested into a recurrence procedure for the solution of the stochastic optimal control problem. Results 1) The optimal switching time between pure colony growth and the exclusive production of sexuals decreases with increasing environmental variance. 2) Yet, for reasonable levels of environmental fluctuations no deviation from the typical bang-bang strategy is predicted. 3) Model calculations for the halictid bee Lasioglossum malachurum reveal that bet-hedging is not likely to be the reason for the graded allocation into sexuals versus workers observed in this species. 4) When environmental variance reaches a critical level our model predicts an abrupt change from dichotomous behaviour to graded allocation strategies, but the transition between colony growth and production of sexuals is not necessarily monotonic. Both, the critical level of environmental variance as well as the characteristic pattern of resource allocation strongly depend on the type of function used to describe environmental fluctuations. Conclusion Up to now bet-hedging as an evolutionary response to variation in season length has been the main argument to explain field observations of graded resource allocation in annual eusocial insect species. However, our model shows that the effect of moderate fluctuations of environmental conditions does not select for deviation from the classical bang-bang strategy and that the evolution of graded allocation strategies can be triggered only by extreme fluctuations. Detailed quantitative observations on resource allocation in eusocial insects are needed to analyse the relevance of alternative explanations, e.g. logistic colony growth or reproductive conflict between queen and workers, for the evolution of graded allocation strategies.",
"conclusion": "Conclusion Up to now bet-hedging as an evolutionary response to variation in season length has been the main argument to explain field observations of graded resource allocation in annual eusocial insects. However, our analysis shows that the effect of moderate environmental fluctuations does not select for deviation from the classical bang-bang strategy and that the evolution of graded allocation strategies can be triggered only by extreme fluctuations. Thus, the widespread belief that graded control in social insects is most probably a type of spreading of risk was premature. Both, additional behavioural mechanisms at the colony level and gene-centred or individual-centred approaches can provide promising alternative explanations. Detailed quantitative field or laboratory observations on resource allocation in eusocial insects are required to analyse the relevance of alternative explanations, e.g. logistic colony growth or reproductive conflict between queen and workers, for the evolution of graded allocation strategies.",
"discussion": "Discussion Our analysis of the optimal resource allocation pattern in eusocial insect colonies clearly demonstrates that moderate fluctuations of environmental conditions (length of foraging season) will not necessarily foster the evolution of bet-hedging allocation strategies. This deviation from the rather intuitive predictions of Oster and Wilson is readily explained by the inherent buffering capacity of the bang-bang allocation strategy; finite worker productivity and mortality rates determine an extended reproductive phase at the end of the season, when sexuals are produced exclusively. Thus, even if season length would be rather short, colonies could nonetheless expect certain fitness (successful production of sexuals) as long as the season ends after the onset of the reproductive phase. Consequently, rather high fluctuations in environmental conditions are needed to promote the evolution of graded allocation strategies with the simultaneous production of workers and sexuals. The results of our model are rather robust against variation in model parameters (worker mortality, worker survival and mean season length). Parameter modifications within a plausible range did not change results markedly. It seems plausible to assume that increasing mean season length might reduce the effect of environmental variance, as identical environmental variance decreases relatively when mean season length increases. This is not the case. Increasing mean season length will just prolong the period of complete worker production, but not influence strategy transition. Even more, very short mean season length could result in strategies which start with the production of sexuals right from the beginning [ 33 , 46 - 48 ]. As long as a season ends after the onset of the reproductive phase the pure bang-bang strategy is buffered against complete reproductive failure. The switch from the bang-bang strategy to a graded strategy thus strongly depends on the length of the reproductive phase. Worker efficiency and survival are the main determinants of the switching time in the deterministic case without environmental fluctuations and the optimal duration of sexual production decreases with increasing worker survival ( q ) and increasing worker efficiency ( c ) [ 8 ]. Populations in ideal conditions with high worker survival and high worker efficiency will thus switch to graded allocation strategies for much smaller variability of environmental conditions than populations that live under harsh environmental conditions. Yet, at least for Lasioglossum malachurum the broad transition zone between pure colony growth and reproduction cannot be explained as an adaptive response to fluctuating environmental conditions. Even for normally distributed season length the observed transition zone is more than twice as long as predicted based on realistic estimates for the coefficients of variation for environmental fluctuations. Model results for uniformly distributed season length indicate that with more realistic distribution functions and observed variability of season length graded control is rather improbable. According to our model, a transition phase between worker and sexual production is not necessarily characterised by a smooth continuous (sigmoid) increase in the production of sexuals as has been predicted by Oster and Wilson [ 11 ]. The specific form of this transition strongly depends on the frequency distribution of season length. For uniformly distributed season length with a very steep left flank, the transition zone is characterised by a peak in the amount of resources invested into reproduction of sexuals at the beginning of the transition phase. On the other hand, for the normal distribution, with its very smooth left flank the transition is characterised by a monotonic increase in the amount of resources invested into reproduction. We have chosen the rectangular and the normal distribution because they are both commonly used types of frequency distributions representing the opposite ends of a continuum of distributions with increasingly steep flanks [ 30 , 33 ]. However, the normal distribution is an unbounded distribution while the minimum as well as the maximum length of a season is obviously bounded. Thus, assuming a normal distribution of season length may easily produce artefacts in the context of bet hedging. This can be seen, when the variance of season length is greatly increased. As we have to limit season length to a minimum of one day an increase in the variance of the normal distribution necessarily leads to steeper left flanks of this distribution. When we do this, the peak in resource allocation observed in the case of the uniform distribution emerges again. An increase of the lower boundary of the season length (in Fig. 3 and Fig. 5 we assume a minimum seasons length of one day, L min = 1) will also reestablish the humped shape for predicted resource allocation. A lower boundary of L min = 27 (about half of the mean season length) which cuts only 1% of the normal distribution in Fig. 4c will result in a prominent peak in the strategy curve. Immediately after the onset of the reproductive phase nearly 30% of the resources are invested into the production of sexuals. This rather high investment decreases subsequently to values around 15% before it rises again to end with a pure sexual phase (Fig. 6 ). In general this hump becomes more pronounced when either the length of the season becomes more variable or when the left flank of the density distribution of the season length becomes steeper (Fig. 2 ). Figure 6 Optimal investment strategy (a), when 1% of the frequency distribution of season length is cut at the left side of the distribution (b). There is a prominent peak in sexual production just before minimum season length. Model parameters: worker productivity rate c = 0.15, survival rate q = 0.95, mean season length μ = 50, standard deviation σ = 8.7. We have shown that variation in season length is not likely to be the reason for graded control in halictids. We have given two arguments. In general, environmental variation can already be buffered by simple bang-bang strategies. Parameter and effect estimation for the sample species L. malachurum in fact lead to the conclusion that the environmental variation this species is exposed to is too low to necessitate graded control as an evolutionary answer. Instead, we believe that alternative mechanisms can be responsible for the evolution of graded control in this species. For the sake of simplicity (and to accord with the simple model by Macevicz and Oster) we have assumed constant productivity and mortality rates during the season. However, it is reasonable that worker (per capita) productivity declines, when colony size increases. That this in fact occurs has been shown by Michener for several halictid species [ 49 ]. Although different theoretical approaches tried to analyse the effect of decreasing worker productivity within the framework of optimal dynamic resource allocation, the most straight forward analysis has never been performed: replacing the linear dependence of resource allocation on worker number (see eqns. 1 and 2) by a logistic relationship. Surprisingly, Macevicz and Oster [ 8 ] used such a relationship only for the worker equation (see eqn. 1 here) to estimate model parameters from field data, but not for the dynamics of the sexuals, thus inevitably favouring the early production of sexuals. Their remark, that saturation of productivity can promote graded control only for very restrictive parameter combinations seems premature, and further theoretical effort on this topic should be promising. Beekman et al. [ 9 ] investigated the effect of limited egg laying rate in bumble bees. This is equivalent to saturation of colony productivity, but unfortunately they allowed only for dichotomous strategy switches in their model, i.e. did not foresee the evolution of graded control. In this case the evolution of early switching is always accompanied by a waste of time. As the egg laying rate of L. malachurum queens is limited, too [ 35 ], this may be a reason for the graded allocation strategy observed in this species. First model calculations (unpublished) support this hypothesis. Thus, a thorough analysis should incorporate a broader set of strategic options to predict the influence of rate limitations on resource allocation strategies. However, it has to be kept in mind that workers of L. malachurum are not sterile. Thus, the queen's limited egg laying rate does not necessarily imply a saturation of egg production, as workers might provide additional eggs [ 35 ]. The original model of Macevicz and Oster [ 8 ] as well as our approach consider the colony (homologous to a single individual) as the unit of selection. As long as we assume that only a single trait is variable (and independent from others) and that all individuals of the colony do not have any further behavioural options, this perspective is also valid at the individual and the genetic level. However, since Macevicz and Oster there has been a wealth of theoretical and empirical analyses on individual worker fitness and the balance between gene, individual and colony level selection (see [ 50 ] for review). Selfishness, e.g. a worker disappearing into hibernation to become a queen the following year or leaving the colony for independent nest founding, may also account for 'graded control', even though this may be sub-optimal at the level of the colony. For most eusocial halictids, workers probably have a variety of reproductive options ( see [ 51 ] for cases of 'worker-sized' queens in L. malachurum that enter hibernation to found colonies the following year). The same may be true for social polistine wasps, where the distinction between gyne and worker toward the end of the colony cycle (and even earlier in the colony cycle) is difficult [ 52 ]. The timing of the production of males is another aspect that has benefited from a gene-centred analysis [ 53 ]."
} | 3,455 |
29988471 | PMC6025826 | pmc | 2,976 | {
"abstract": "Background Microalgae biomass is regarded as a potential feedstock for bioenergy purposes through anaerobic digestion (AD). Even though AD is a well-proven technology, the use of new feedstocks requires in-depth studies. A lot of research has been conducted assessing methane yield without paying attention to the anaerobic microbiome and their activities. For such a goal, the present investigation was designed to link methane yield to those two later sludge characteristics. In this sense, different anaerobic sources were tested, namely adapted to microalgae biomass and adapted to sewage sludge. Results Despite the registered differences for the anaerobic microbiome analysis and specific methane activities towards model substrates, sludge adapted to digest sewage sludge did not affect the methane yield of Chlorella sorokiniana and Scenedesmus sp. Opposite to that, sludge samples adapted to digest microalgae exhibited a concomitant increase in methane yield together with increasing digestion temperatures. More specifically, the values attained were 63.4 ± 1.5, 79.2 ± 3.1 and 108.2 ± 1.9 mL CH 4 g COD in −1 for psychrophilic, mesophilic and thermophilic digestions, respectively. While psycro- and mesophilic digestion supported similar yields (most probably linked to their anaerobic microbiome resemblance), the values attained for thermophilic digestion evidenced the usefulness of having a highly specific microbiome. The relative abundance of Firmicutes, particularly Clostridia , and Proteobacteria together with an important abundance of hydrogenotrophic methanogens was highlighted in this inoculum. Conclusion Overall, this study showed that working with tailored anaerobic microbiome could help avoiding pretreatments devoted to methane yield enhancement.",
"conclusion": "Conclusion This research demonstrated that despite the differences related to their anaerobic microbiome and SMA towards model substrates, the anaerobic digestion of microalgae biomass was not influenced by different inocula sources adapted to digest sewage sludge. Opposite to that, the sludge samples adapted to digest microalgae biomass exhibited better performances. Mesophilic sludge adapted to the digestion of microalgae consortium mainly composed by Scenedesmus showed greater methane yields than adapted to digest sewage sludge. The same could not be concluded with other microalgae biomass ( Chlorella ). Thus, it could be concluded that the anaerobic microbiome was tailored to degrade mainly Scenedesmus . With regard to the adapted inocula, psychrophilic digestion displayed lower methane productivity while methane yield was comparable to mesophilic digestion with adapted sludge. Most remarkably was the methane yield achieved by the thermophilic adapted sludge. Even though, a high microbial diversity might play a positive role in maintaining the stability of the system, the anaerobic microbiome of thermophilic digester presented a low diversity but highly efficient for the anaerobic digestion of Scenedesmus sp. The relative abundance of Firmicutes, particularly Clostridia , and Proteobacteria together with an important abundance of hydrogenotrophic methanogens was highlighted in this inoculum. Linking process engineering to microbial community in AD reactors could bring new insights to pay the way out to a better digester performance and avoid pretreatments by working with a highly specific anaerobic microbiome.",
"discussion": "Results and discussion Sludge samples and microalgae chemical characteristics All anaerobic sludge samples exhibited a standard VS/TS ratio ranging 0.66 ± 0.05, VSS/TSS = 0.80 ± 0.04 and VSS/VS = 0.57 ± 0.07. These values were in good agreement with other sludge samples used for BMP tests [ 20 , 21 ]. In this manner, despite the different feedstock used during their previous activity, namely sewage sludge (mixture of primary and secondary sludge) (S1, S2 and S3) or microalgae biomass (S4, S5 and S6), all of them were within the usual values. With regard to the microalgae biomass employed as substrate, both biomass showed a prevalent protein composition (Table 2 ). This feature is quite normal of microalgae grown without any stressful condition [ 20 ]. In the case of Scenedesmus sp., proteins were followed by carbohydrates (approx. 34%) and a minor proportion of lipids (11%). Opposite to that, C. sorokiniana displayed a higher lipid content (24%) than the carbohydrate fraction (10%). It should be noted that this strain has been reported to have a particular tendency to accumulate lipids [ 22 , 23 ]. Table 2 Macromolecular distribution of the microalgae biomass employed as feedstock for anaerobic digestion Chemical parameter Scenedesmus sp. \n C. sorokiniana \n TS g L −1 26.7 ± 0.7 37.5 ± 0.2 VS g L −1 23.0 ± 1.2 34.1 ± 0.0 Carbohydrates (%) 33.8 ± 2.4 10.4 ± 0.7 Proteins (%) 41.0 ± 0.7 56.3 ± 0.7 Lipids (%) 11.5 ± 3.4 24.1 ± 0.7 Ash (%) 13.8 ± 2.5 9.2 ± 0.3 \n Sludge samples: anaerobic microbiome Pyrosequencing was performed to characterize the anaerobic microbiome of all sludge samples. With regard to Shannon’s index of diversity at genera level, all the sludge samples were in the range of 1.36–1.75, exception made for S1 and S6, which exhibited 3.07 and 0.32, respectively. This indicated that the diversity of S1 was considerably higher than the rest of sludge samples while S6 displayed a highly specific consortium in which the biodiversity was really reduced. At phylum level, bacterial distribution was mainly represented by Proteobacteria. This phylum ranged 46–51% of the bacteria retrieved in the adapted sludge samples while in S2, the prevalence of Proteobacteria was outcompeted by Actinobacteria (27%) and S1 and S3 exhibited 35 and 42% of the bacterial population (Fig. 1 a). Proteobacteria are frequently reported to be present at high proportion in anaerobic sludge [ 24 ]. Groups of Proteobacteria as Rhizobiales, Rhodobacterales, Sphingomonadales and Burkholderiaels – Comamonadaceae found in this work have been also reported in studies related to microalgae-based wastewater treatment [ 25 ]. Reactors S3–6 were adapted to digest microalgae–bacteria consortia harvested from a photosynthetic-based wastewater treatment. This inocula adaptation can explain the prevalent presence of this group of bacteria in these inocula. It has been shown that Proteobacteria population increases from 13 to 50% when changing the feeding from sewage sludge to raw Chlorella [ 26 ]. This fact was also observed in the present investigation, in which the sludge samples adapted to digest microalgae (S4–S6) exhibited slightly higher population of this phylum. This feature could be explained by the above-mentioned prevalence of proteins in microalgae biomass. Fig. 1 Taxonomic profiles at phylum ( a ) and genera ( b ) level for the bacterial community of the different inoculum sources used in the BMPs \n The second main phyla retrieved from the sludge samples was Firmicutes. Non-adapted sludge (S1–S3) showed a relative abundance of approximately 20% of Firmicutes (Fig. 1 a). This bacterial community is also normally present in anaerobic sludge digesting sewage sludge. Interestingly, adapted sludge samples (S4–S6) decreased their abundance of Firmicutes in psychrophilic and mesophilic sludge to around 10% while that of the thermophilic displayed 28% relative abundance. The sequences were mainly affiliated to the order Clostridiales . More specifically, the relative abundance of Clostridia in S2–S5 were in the narrow range of 62–66% of the Firmicutes, while S1 and S6 showed markedly higher percentages (86 and 96%, Fig. 1 b). It is important to note that Clostridia are responsible of conducting macromolecules hydrolysis [ 27 ] and syntrophic acetate oxidation coupled with hydrogenotrophic methanogenesis [ 28 ]. Clostridiales is an order of obligate anaerobic bacteria with chemoorganotrophic fermentative metabolisms. Members of this group of microorganisms are frequent in soils, sediments, rumen and intestinal tracts of animals and insects. In addition, Clostridiaceae family members have in common their saccharolytic fermentative metabolism in environmental substrate decomposition [ 29 ]. Bacteroidetes were present at really low relative abundance (2–6% for all sludge samples). Bacteroidetes are anaerobic microorganisms which are included in animal intestinal microflora [ 30 ] due to its importance in cellulose and protein degradation [ 31 ]. When digesting raw and thermally pretreated Chlorella , Bacteroidetes abundancy increased to 20% with regard to 12% registered in the inoculum used as seed of the reactor (adapted to digest sewage sludge, [ 26 ]). Likewise, Bacteroidetes also found the dominant phylum in the digestion of Scenedesmus obliquus at mesophilic range [ 32 ]. Nevertheless, in the present study, their abundance remained low regardless of the inoculum source. Chloroflexi phylum are aerobic bacteria commonly found in activated sludge systems [ 33 ], and therefore, this phylum usually ends up in the anaerobic digesters of wastewater treatment plants. Due to their filamentous morphology, their presence is normally associated to bulking phenomena. In this manner, no trend could be withdrawn in sludge samples collected from wastewater treatment and their relative abundance varied from 6% for S3 to 27% for S1. A striking feature was the fact that the thermophilic adapted sludge showed 1% of this phylum (Fig. 1 a). This value was really low when compared to the rest of sludge samples. Nevertheless, this feature was in agreement with Greses et al. [ 34 ] who also reported extremely low relative abundance (1.1%) in a thermophilic CSTR fed with S cenedesmus . Authors attributed this fact to the low-ammonium tolerance of this phyla and the operational temperature. It should be also highlighted that S6 displayed 16% Thermotogae relative abundance (classified as other in Fig. 1 a). The presence of this phylum was negligible in the rest of the sludge samples. Thermotogae have been described to release hydrolytic enzymes catalyzing the degradation of polysaccharides into acetate, carbon dioxide and hydrogen [ 35 ] and also to play a key role in interspecies hydrogen transfer [ 36 ]. Thermotogae species are obligate anaerobes and hyperthermophiles bacteria found in large range extremophile environments, including mammalian, ruminant and termite digestive tracts. All those ecophysiology characteristics explain this abundance in S6 sludge [ 37 ]. The archaeal domain accounted for 7–8% of the identified population in the case of sludge adapted to digest sewage sludge (S1–S3), while this value decreased to 3–4% for all the sludge samples adapted to digest microalgae (S4–S6). In this manner, it was clear that microalgae digestion affected archaeal relative abundance. With regard to the phyla determined in the different sludge samples, the abundance of the strict acetotrophic Methanosaeta was prevailing in S1–S5, ranging from 67 to 82% of the archaeal population (Fig. 2 ). Opposite to that, methanogenesis in the thermophilic sludge S6 was mainly conducted by Methanothermobacter and Methanosarcina (59 and 40% of the archaeal population retrieved in the sample, respectively). Methanothermobacter is a hydrogenotrophic methanogen [ 38 ] while Methanosarcina is more versatile and can metabolize both hydrogen and acetate as energy source [ 39 ]. Indeed, acetoclastic methanogens are commonly outcompeted by hydrogenotrophic methanogens in thermophilic digesters [ 40 ]. This fact is related to the lower stability of thermophilic digestion conditions as a result of acetate accumulation and acidification. In this manner, according to the bacterial and archaeal population it can be assumed that this was the case as well in S6. Hydrogenotrophic methanogens were also represented by the Methanomicrobiales community identified in S1–S5, however, their relative abundance was much lower than in the thermophilic digester (10–30% vs. 60%). Fig. 2 Taxonomic profiles at phylum level for the archaeal community of the different inoculum sources used in the BMPs \n There are no other studies published that focused in the comparison of the microbial population inocula adapted to digest microalgae and sewage sludge at different temperatures ranges. As an example, many other articles have characterized the microbial composition of inocula adapted to digest microalgae and activated sludge only at one temperature range [ 24 , 25 , 29 , 34 ]. The novelty of this article is the comparison of different microbiome composition adapted to three different operational temperature ranges and feedstocks (microalgae and raw sewage). Overall, the anaerobic microbiome analysis of all sludge samples used in the present study evidenced different bacterial and archaeal population not only in terms of relative abundances but also in the phyla and genera identified. In this manner, it could be concluded that the different feedstocks fed in the digesters (sewage sludge and microalgae biomass) as well as the operational temperature had an impact on the inoculum sources. Specific methanogenic activity of sludge samples toward model substrates: BSA and cellulose According to Angelidaki et al. [ 41 ] specific methanogenic activity (SMA) was measured using different model substrates representing proteins (BSA) and carbohydrates (cellulose). Those two substrates were selected as model for protein and carbohydrate degradation since those are the fractions which present lower biodegradability in the microalgae biomass [ 16 , 20 ]. As it can be seen in Fig. 3 , SMA was higher and faster for the BSA model substrate than with cellulose. The sludge S1 exhibited considerably higher SMA than the adapted sludge samples. More specifically, in the first day of digestion, S1 displayed threefold SMA (0.157 g COD consumed/g VS day) than S4, S5 and S6 (0.055 g COD consumed/g VS day, Fig. 3 a). This fact was probably mediated by the higher microbial diversity encountered in this sludge in comparison to the adapted ones (Section “ Sludge samples: anaerobic microbiome ”). A higher diversity implies more options that the appropriate degrading microorganisms is within the microbial consortium, which ultimately implies more degrading pathways that are strictly required for anaerobic digestion. Even though digesters with low diversity might operate under stable conditions, higher microbial diversity ensures a higher resistance resilience and functional redundancy, which ultimately results in good AD performances [ 42 ]. Fig. 3 Specific methanogenic activities (SMA) towards model substrates: a BSA and b cellulose Since hydrolysis is the first step on anaerobic digestion, it can be assumed that these data reflected a higher proteolytic activity of S1 with regard to the other tested sludge samples. Proteolytic bacteria occur in high numbers in anaerobic digesters receiving raw sewage sludge [ 24 ]. This would support the fact that S1-SMA was the highest since the other sludge samples were adapted to digest mainly microalgae biomass. More interestingly, S1 was the sludge with lower Proteobacteria relative abundance (35% out of the total bacterial number vs . 46–51% for the adapted sludge, Fig. 1 a). In this manner, it can be highlighted that despite the lower relative abundance, the enzymes secreted by the proteolytic bacteria of S1 were more suitable and/or active for BSA hydrolysis. The SMA using BSA as a substrate was diminished along digestion time. After approximately 7 days of digestion, the SMA attained for BSA was minimal (Fig. 3 a) while the one corresponding to cellulose as model substrate slowly increased (Fig. 3 b). Thus, it could be concluded that all sludge samples had a higher proteolytic activity than cellulolytic toward the model substrates. One or other prevalent hydrolytic activity strongly depend on the selected inoculum. In principle, protein hydrolysis is slower than the hydrolysis of carbohydrates [ 43 ]. Nevertheless, hydrolysis constants and, therefore, methane production is highly dependent on several factors such as microbial community, selected model substrate and digestion temperature. In this manner, a wide range of hydrolytic constant might be found in the literature [ 44 ]. As observed for S1 and the adapted sludge samples (S4–S6), the SMA obtained when using cellulose model substrate was also clearly different (Fig. 3 b). The only common behavior for all the sludge samples was the lack of methane production during the 2–3 first days of digestion. Similarly to what it was observed in the previous case, S1 exhibited three-fold higher SMA than the adapted sludge samples after 5 days of digestion (0.07 g COD consumed/g VS day for S1 vs ± 0.022 g COD consumed/g VS day for S4–S6). Once again, the higher biodiversity determined in S1 mediated higher SMA for cellulose than the rest of the sludge samples. While in the case of the proteolytic activity, SMA steadily decreased along digestion, in the case of cellulolytic activity, the sludge samples responded differently. Almost barely any activity was observed after 7 days of digestion for S1, while the adapted sludge samples maintained an SMA of 0.021–0.024 along 19 days of digestion regardless of the digestion temperature. This fact resulted in a similar methane yield but lower methane productivity of the adapted sludge samples. Microalgae digestion using different anaerobic sludge inocula Based on the above-mentioned analysis, it was obvious that differences existed among the inocula used herein. Incubation with different inocula at mesophilic range resulted in similar BMP values for Scenedesmus sp. biomass, since no significant differences were observed between the different non-adapted mesophilic inocula (Fig. 4 a). S1–S3 supported methane yields of 63.1 ± 3.1 mL CH 4 g COD in −1 . Slightly higher values (79.2 ± 3.1 mL CH 4 g COD in −1 ) were determined for the sludge adapted to digest microalgae biomass (S4). With regard to the sludge samples adapted to digest microalgae biomass at different temperature, methane yields increased concomitantly with digestion temperature. More specifically, the values attained were 63.4 ± 1.5, 79.2 ± 3.1 and 108.2 ± 1.9 mL CH 4 g COD in −1 for psychrophilic, mesophilic and thermophilic digestions, respectively (Fig. 4 b). Despite of the differences registered in terms of the anaerobic microbiome and the metabolic activities towards model substrates, no big differences were evidenced among the mesophilic sludge samples. This feature was opposite to what it is described in the literature when testing different inocula sources for the digestion of different biomass. For instance, Gu et al. [ 45 ] evaluated different inocula sources on the anaerobic degradation of rice straw and their results showed that digested manures were more active than anaerobic sludge. Similarly, Cordoba et al. [ 11 ] also proved that the selection of inocula sources could affect the anaerobic digestion of liquid swine wastewater. Nevertheless, in some cases, the effect of inocula source is only detectable depending on the targeted substrate [ 46 ]. In the present study, the only remarkable differences were observed for the sludge samples adapted to digest microalgae biomass at meso- and thermophilic range. In principle, thermophilic digesters are usually operated as close as possible to 50 °C. This temperature, being close to the optimum for enzymatic activity, frequently results in faster reaction rates compared to mesophilic digestion, leading to shorter retention times. Therefore, advantages of thermophilic digestion involve faster hydrolysis and acidogenesis while being more sensitive to ammonia toxicity. Within thermophilic digestion of microalgae in BMP mode, Capson-Tojo et al. [ 47 ] digested lipid-extracted Nannochloropsis gaditana at mesophilic and thermophilic range. They concluded that mesophilic digestion supported higher anaerobic biodegradability than thermophilic, however, they also observed that this later digestion temperature supported higher COD solubilization. This was attributed to the fact that the used anaerobic sludge was indeed mesophilic for both assays, and thus, the digestion run at thermophilic range was too short to get the anaerobic microorganisms adapted to the thermophilic temperature. As a matter of fact, the investigation presented herein proved that when the inoculum was adapted to digest microalgae at thermophilic conditions, methane yield was the highest of all trials. Overall, adapted sludge samples to the digestion of microalgae have shown to be beneficial for the biodegradation of Scenedesmus sp. Digestion conducted at psychrophilic range supported similar methane yields than the sludge samples adapted to digest sewage sludge (Fig. 4 ). Therefore, the same yield could be achieved at lower energy cost for maintaining the digester temperature. Methane yields of psycro- and mesophilic digestions were quiet similar most likely due to their anaerobic microbiome resemblance (bacterial and archaeal population). The most remarkable difference in methane yields was achieved by the thermophilic consortium, which provided the highest value. Methane yield at thermophilic range was 1.36-fold higher than mesophilic range. The benefits of using thermophilic digestion over mesophilic digestion seems to be specie specific [ 5 ]. According to Zamalloa et al. [ 48 ], the digestion of S. obliquus in thermophilic range increased the methane yield by 24% when compared to the digestion in mesophilic range operating continuous digesters. To the best of the authors’ knowledge, no thermophilic digestion in batch mode has been published for Scenedesmus and thus, no comparison could be made. Nevertheless, pretreatments devoted to enhance methane yields of this microalgae strain in BMPs reported similar enhancement values [ 49 ]. Thus, this study highlighted the potential of working with highly specific consortia to increase methane yield. Given the differences in the anaerobic microbiomes of the different sludge samples, it can be concluded that the microorganism’s consortium developed in the adapted sludge was linked to the higher methane yields achieved. As hypothesized by Greses et al. [ 34 ] the results obtained herein seemed to confirm that the high relative abundance of Firmicutes in the thermophilic sludge (Fig. 2 a) compared to the rest of the sludge samples could be the explanation for the higher methane yields achieved with this inocula. Fig. 4 Cumulative methane production achieved by mesophilic sludge samples ( a ) and adapted to microalgae sludge samples ( b ) when digesting Scenedesmus sp. biomass \n According to De Vrieze et al. [ 46 ] the substrate employed could be of paramount importance when dealing with the effect of different inocula sources. Since Scenedesmus sp. is most probably the hardest microalgae to digest [ 15 ], a similar approach was conducted with some easier digestible microalgae. The results attained for the digestion of Chlorella sorokiniana upon the use of selected inocula was shown in Fig. 5 . The main difference was the methane productivity while the final yields were not affected by the inoculum. After 15 days of digestion, methane yields ranged 105–114 mL CH 4 g COD in −1 for the three tested sludge. In this context, the beneficial effect observed on the anaerobic digestion of Scenedesmus sp. by the inoculum adapted to digest microalgae (S4) was not evident in the digestion of C. sorokiniana . Nevertheless, it cannot be neglected that the productivity was higher for S4 than for the rest of the sludge samples. In this manner, S4 achieved maximum methane yield after 10 days of digestion while S1 required 18 days, despite of the highest specific methanogenic activity registered for this sludge (Fig. 3 ). When compared to C. sorokiniana , the better results obtained for Scenedesmus sp. were related to the fact that the microalgae biomass digested by S4–S5 and S6 was mainly composed by Scenedesmus , Dictyosphaerium , Coelastrum , Micractinium and Chlorella . Most probably, the anaerobic microbiome developed in the anaerobic inocula of the adapted sludge samples was particularly suitable for the digestion of Scenedesmus and thus, the positive effect was more evident in that biomass. Fig. 5 Cumulative methane production achieved by mesophilic sludge samples when digesting Chlorella sorokiniana biomass"
} | 6,137 |
16519800 | PMC1420336 | pmc | 2,977 | {
"abstract": "Background Biochemically detailed stoichiometric matrices have now been reconstructed for various bacteria, yeast, and for the human cardiac mitochondrion based on genomic and proteomic data. These networks have been manually curated based on legacy data and elementally and charge balanced. Comparative analysis of these well curated networks is now possible. Pairs of metabolites often appear together in several network reactions, linking them topologically. This co-occurrence of pairs of metabolites in metabolic reactions is termed herein \"metabolite coupling.\" These metabolite pairs can be directly computed from the stoichiometric matrix, S . Metabolite coupling is derived from the matrix ŜŜ T , whose off-diagonal elements indicate the number of reactions in which any two metabolites participate together, where Ŝ is the binary form of S . Results Metabolite coupling in the studied networks was found to be dominated by a relatively small group of highly interacting pairs of metabolites. As would be expected, metabolites with high individual metabolite connectivity also tended to be those with the highest metabolite coupling, as the most connected metabolites couple more often. For metabolite pairs that are not highly coupled, we show that the number of reactions a pair of metabolites shares across a metabolic network closely approximates a line on a log-log scale. We also show that the preferential coupling of two metabolites with each other is spread across the spectrum of metabolites and is not unique to the most connected metabolites. We provide a measure for determining which metabolite pairs couple more often than would be expected based on their individual connectivity in the network and show that these metabolites often derive their principal biological functions from existing in pairs. Thus, analysis of metabolite coupling provides information beyond that which is found from studying the individual connectivity of individual metabolites. Conclusion The coupling of metabolites is an important topological property of metabolic networks. By computing coupling quantitatively for the first time in genome-scale metabolic networks, we provide insight into the basic structure of these networks.",
"conclusion": "Conclusion This study presents the first calculations of metabolite coupling for curated, genome-scale metabolic networks. As expected, metabolite coupling is dominated by a few highly connected pairs (H + /H 2 O, ATP/ADP, etc.). However, it was also shown that the coupling of metabolites with lower individual connectivity demonstrated a striking amount of regularity, with the number of reactions shared by pairs of metabolites in a given metabolic network closely approximating a line on a log-log plot. We also probed the network to discover which metabolite pairings occurred much more or less than would be expected from their individual connectivities alone. We found that the highly connected metabolites contributed a disproportionate percentage of the enriched coupling interactions in that highly connected metabolites are more connected to each other than would be expected from their individual connectivities alone. These preferentially coupled pairs of metabolites highlight chemical relationships between molecules. In this study, it is interesting that all of the top 15 most coupled metabolites have at least one other metabolite with which they preferentially couple in E. coli , highlighting the importance of metabolite coupling to the emergence of highly-connected compounds in metabolic networks.",
"discussion": "Results and discussion Basic network properties We characterized six metabolic networks based on topological features, with representatives from each of the three primary domains of life. The metabolic networks considered represent three bacteria ( E. coli [ 2 ], Helicobacter pylori [ 19 ], Staphylococcus aureus [ 20 ]), one member of the archea ( Methanosarcina barkeri [ 21 ]), one unicellular eukaryote ( S. cerevisiae [ 22 ]), and one human cellular organelle (cardiac mitochondrion [ 3 ]). These are, to the best of our knowledge, the only genome-scale metabolic networks that are manually curated as well as elementally and charge balanced available to date. We computed the symmetric metabolite coupling matrix M for each network by multiplying the binary form of the stoichiometric matrix S by its transpose (see Methods for details). Each diagonal element of M (m ii ) represents the number of reactions in which a particular metabolite appears and each off-diagonal element (m ij , i ≠ j) represents the number of reactions in which two particular metabolites appear together. Figure 1 shows a graphical representation of the first 15 rows and columns of M for all six networks. Both the box size and the histogram height are logarithmically scaled and normalized by the number of unique metabolite pairs that occur in each network. This scaling and normalization allows for the visualization of pairs of metabolites that participate in anywhere from zero to several hundred reactions together in any given network. The color of each box and bar represents one particular network. Any matrix constructed by the multiplication of another matrix with its transpose is symmetric, and thus the points above the diagonal of M are identical to those below; thus, each histogram below the diagonal corresponds to one set of colored boxes above the diagonal. It is notable that, of the 15 most connected metabolites, only three pairs never couple in the six networks analyzed here: ADP/NADH, ADP/NADPH, and NH 4 /CoA. The first two metabolite pairs that never couple are perhaps surprising, because it is well known that cells use electron carriers to phosphorylate ADP. However, because this process occurs in a series of reactions instead of a single reaction, these metabolites are not considered to be coupled. The computation of second-degree coupling (where two metabolites are coupled if they couple with at least one common third metabolite) would find this relationship, but would also introduce significant difficulties because many metabolite pairs would be second-degree coupled solely because they are individually coupled with the proton or water. Basic properties of each network and its corresponding M are shown in table 1 . The metabolic network of E. coli has 621 compounds and thus its M is a square symmetric matrix with 621 2 elements. Each compound in the network must participate in at least one reaction, so each diagonal element of M must be greater than zero. However, there is no general topological reason that an arbitrary pair of metabolites must participate in any reaction together, so zeroes are feasible for off-diagonal elements. In fact, the density of M for E. coli is 2.2%, meaning that approximately 98% of the (((621 2 )/2)-621) unique off-diagonal elements are zero. Figure 2 is a binary representation of M for E. coli , constructed by keeping all zeros in place (white spaces) and replacing all non-zeros with 1 (black points), that clearly demonstrates this sparsity. Smaller networks tend to have a denser M . Metabolite coupling For each of the six networks, we determined the number of reactions in which each possible pair of metabolites is coupled. We defined two metabolites as coupled in a given reaction if they both occur in that reaction, in contrast to the recently presented concept of metabolite concentration coupling analysis which studies metabolites whose concentrations are linked by network stoichiometry [ 23 ]. The distribution of metabolite coupling is shown for E. coli (figure 3 ) and S. cerevisiae (figure 4 ). Most pairs of metabolites are never coupled and cannot be shown on a log-log plot. Many pairs of metabolites occur together in only one reaction and are illustrated by the leftmost data point of each figure. The pairs of coupled metabolites that occur in many reactions together are the rightmost points in each plot and represent pairs of metabolites that are either small and ubiquitous (e.g. H + , PP i ) or traditionally referred to as cofactors (e.g. NAD + , NADH). In the six networks considered, the two most common metabolite pairs are always H + /H 2 O and ATP/ADP. The pair ATP/ADP is ranked first in H. pylori and S. aureus but second in the remaining four networks. The identity and order of coupled metabolite pairs diverges after these two pairs, as illustrated by the plots for the two model microorganisms, E. coli and S. cerevisiae . The number of reactions in which each possible combination of two metabolites participates approximates a line on a log-log scale in all networks studied, with average slopes between -2 and -3 (figure 5 ). The slope of the best-fit line (linear on a log-log scale) was determined for each network from the left-most 10 points and is indicated in the figure legend, along with the corresponding r 2 value. While the dominance of certain cofactor pairs in metabolic networks was known, the strong fit to a line for the less connected pairs demonstrates significant regularity in the use of pairs of metabolites in metabolic networks. The slope of this line provides a quantitative measure to describe the organization of metabolite coupling in the network and can be used to compare coupling properties of different networks of various sizes. Because the coupling relationships fit a line so well, it is possible in principle to determine the relative dominance of certain pairs of metabolites across networks. A smaller (negative) slope generally implies that there are either fewer metabolite pairs that only appear once or twice or that there are more pairs of metabolites that participate in an intermediate number of reactions (for example, 5 to 10). A smaller slope suggests that relatively more metabolite pairs influence more than one reaction and link reactions together. Metabolite connectivity For each reconstructed metabolic network, we calculated the number of reactions in which an individual metabolite occurs, m ii , as a measure of its individual connectivity in each network (figure 6 ). The first reconstructed genome-scale network for Haemophilus influenzae metabolism [ 24 ] suggested that this property had a power-law distribution and fit a line on a log-log scale. Similar calculations made for many networks [ 7 ] using a somewhat different network formalism described earlier yielded similar results. The fit to a line in figure 6 begins after the number of occurrences is two or greater, since a metabolite must participate in at least two reactions to be involved in reactions that can be active at steady state. This result is different from the graph-based approaches that plot in-degree and out-degree separately [ 7 ], essentially converting all metabolites that participate in two reactions into two occurrences of one reaction each. Furthermore, the slope of each line for single metabolite connectivities (figure 6 ) is different than the slope of the corresponding line for metabolite coupling (figure 5 ). When the ratio of the slope for metabolite coupling to the slope for single metabolite participation is computed, the results range from 0.95 (mitochondrion) to 1.54 ( S.aureus ) with a mean of 1.11 and standard deviation of 0.10. Furthermore, when the slopes are rank ordered from greatest to least, it is observed that the networks are ordered differently when considering single metabolites than when considering metabolite coupling. Thus, the widely differing slope ratios and differential ordering indicates that the distribution of metabolite coupling (off-diagonal elements in M ) is not simply a direct result of the distribution of metabolite connectivity (diagonal elements in M ). Relation between metabolite coupling and metabolite connectivity Although the slopes of best-fit lines vary considerably for single metabolites and metabolite pairs, there is an important relationship between metabolite coupling and individual metabolite connectivity because a particular metabolite cannot be highly coupled with other metabolites if it does not appear in many reactions. Thus, each off-diagonal element of M (m ij ) is subject to both of the following criteria: m ij ≤ m ii m ij ≤ m jj The metabolite coupling and connectivity is shown graphically for E. coli in figure 2 , with the rows and columns ordered by overall connectivity. The most connected metabolites couple with the greatest number of other metabolites, as shown by the large number of black points toward both the top and the left side of the figure. The number of black points in each row (not counting the one black point on the diagonal) is the number of unique compounds with which a given compound couples. For example, in E. coli, hydrogen couples with 479 metabolites and thus there are 479 black points in row 1 (not counting the diagonal). The matrix is symmetric, so there are also 479 black points in column 1. Figure 7 further illustrates that the most connected metabolites couple with considerably more different metabolites than less-connected metabolites. The sum of each row of figure 2 , added to the sum of all previous rows and normalized gives a data point in figure 7 . For all networks, the first point, representing hydrogen, is between 0.05 and 0.06 in figure 7 , signifying that hydrogen accounts for between 5% and 6% of all coupling interactions. The second point, representing the percentage of coupling interactions involving either hydrogen or water, is close to 0.10 (10%) in all networks. The number of metabolites that couple with of the top 15 most connected metabolites is listed in table 2 , the second column of which is the numerical result of summing across a row in figure 2 . It is not surprising that metabolites like the proton and ATP, which each occur in many reactions, couple with many different metabolites. In figure 7 , the nearly constant slope in the middle region of the curve (between approximately metabolite 100 and 500) for all networks except the mitochondrion is interesting, however. While it takes fewer than the most-connected 100 metabolites to account for 50% of the coupling, it takes between 300 and 500 metabolites to account for 90% of the coupling in any given network, excluding the mitochondrion. Furthermore, each of the 300 to 500 metabolites in that region between 50% and 90% participates roughly equally in the cumulative coupling interactions; that is, each of these metabolites couples with a roughly equal number of other metabolites. This is likely due to the simple fact that most of these metabolites each participate in two reactions. The slight decrease in slope for the final 50 to 100 metabolites suggests that the least connected metabolites each account for slightly less coupling, which is unsurprising since most of these metabolites occur in only one reaction, and are dead-ends in their respective networks. The mitochondrion is a special case that does not conform well to the previous generalizations. An examination of the mitochondrial metabolic network shows that, relative to its size, the mitochondrion contains a disproportionately high number of metabolites that are highly coupled in other networks. Whereas the 15 most connected metabolites represent only 2.4% of the 621 metabolites present in E. coli , they account for over 10% of the mitochondrion's 145 metabolites. We present the results for the mitochondrion in the interest of completeness, but caution that it may be difficult to extract meaning from a direct comparison to other, more comprehensive, metabolic networks. While normalization of the results to some measure of network size would bring the data in figure 7 into closer agreement, it would not change the simple fact that the mitochondrial network is fundamentally more limited in scope and over-represents highly connected metabolites relative to the genome-scale matrices. This high connectivity and coupling within the mitochondrion suggests that the individual metabolic reactions therein are more interrelated and may affect each other to a greater extent, on average, than those within a genome-scale bacterial metabolic network. Preferentially coupled metabolite pairs There are some pairs of metabolites that occur in reactions together at a higher rate than would be expected given their individual connectivities. In order to locate these pairs, we used a Monte Carlo approach to uniformly sample the possible values of each off-diagonal element of the metabolite coupling matrix given that the diagonal (the frequency of a given metabolite's participation in the network) remained unchanged. Based on the two diagonal elements corresponding to each off-diagonal element (m ii and m jj correspond to m ij ), each off-diagonal element is assigned a random, feasible value. Complete details are in the methods section. After doing this many times for each pair of metabolites in each network, it is possible to determine which metabolite pairs in each network are coupled more than expected for a random network with the same individual connectivity. The pairs that never couple as often in 10,000 randomizations as they do in the real network are listed for E. coli in table 3 . The spread of preferentially coupled metabolite pairs (those that couple more often in the real network than in 99% of the randomizations) is graphically represented for E. coli in figure 8 . The very existence of preferentially coupled metabolites demonstrates the non-random organization of metabolite coupling. This non-random coupling can be a function of fundamental chemical principles or biological needs, such as the general need for both ATP and ADP to transfer a phosphate moiety. Furthermore, a number of less connected metabolites preferentially couple, demonstrating that this effect is not limited to the highly-connected metabolites. Although many of the preferentially coupled pairs are clustered toward the left side of figure 8 and correspond to the most connected metabolites, the remainder are spread throughout much of the figure, and as a result, the network as a whole. Thus, participating in many reactions is not necessary to detect non-random preferential coupling. This suggests that metabolites with low connectivity overall in the network are still tightly connected to certain other metabolites, often through shared chemical structural properties. Examination of the list presented in table 3 supports this assertion. The proton couples preferentially with several metabolites that gain or lose a proton during the course of balanced biochemical transformations. The adenine and phosphate containing metabolites preferentially couple in a number of cases. Sugars preferentially couple with other sugars, and fatty acids preferentially couple with other fatty acids. Preferentially uncoupled metabolite pairs The same Monte Carlo randomization used to locate overly coupled pairs of metabolites was also used to find metabolites that occur together less often than expected. This procedure located relatively high connectivity metabolites (those that individually participate in many reactions) that do not couple as often as they would in a network with random coupling; results are shown for E. coli in table 4 . For example, this procedure indicates that ATP and NADH couple less than expected based on the number of reactions in which they participate together in every genome-scale network. This preferential uncoupling between ATP and NADH allows the network more flexibility in controlling energy and redox needs. Although these two metabolites never occur together in any reaction in the mitochondrion, having even 95% confidence that this is a non-random result would require that they participate in fewer than zero reactions together and thus this observation was not statistically significant. This demonstrates a limitation of the Monte Carlo approach to finding under-coupled metabolite pairs – it cannot locate with high confidence metabolites that are relatively scarcely connected individually."
} | 5,039 |
25369026 | PMC4219724 | pmc | 2,980 | {
"abstract": "We present a unified approach to describing certain types of collective decision making in swarm robotics that bridges from a microscopic individual-based description to aggregate properties. Our approach encompasses robot swarm experiments, microscopic and probabilistic macroscopic-discrete simulations as well as an analytic mathematical model. Following up on previous work, we identify the symmetry parameter, a measure of the progress of the swarm towards a decision, as a fundamental integrated swarm property and formulate its time evolution as a continuous-time Markov process. Contrary to previous work, which justified this approach only empirically and a posteriori , we justify it from first principles and derive hard limits on the parameter regime in which it is applicable.",
"introduction": "Introduction Swarm robotics refers to the concept of using a number of autonomous and often very simple robots to collaboratively accomplish a task. For some scenarios, this can provide an attractive alternative to deploying a single, complex robot. Typically, three main advantages of a swarm-based approach are cited, which can also be viewed as design goals [2] – [4] . (i) Robustness : swarm performance does not critically depend on individuals and degrades gracefully when individuals malfunction. (ii) Scalability : swarm behaviour can scale well for a wide range of problem sizes (and swarm sizes). (iii) Flexibility : swarms are assumed to adapt their behaviour flexibly to changing environmental conditions. When designing swarm-control mechanisms, researchers and engineers are faced with the challenge to develop a set of rules at the individual (microscopic) level such that a desired behaviour at the group (macroscopic) level is achieved [4] . This is a very difficult task since there is no general systematic way to devise individual behaviours that reliably achieve a desired group behaviour. Thus design choices can usually only be tested in experiments or simulations. Performing swarm experiments is expensive and requires considerable effort and time comitment [3] , [4] . Simulations, on the other hand, are efficient and fast but cannot achieve the same degree of realistic behaviour as physical experiments. Any approach that allows us to derive predictions of a swarm's behaviour analytically would thus be of significant advantage. In an attempt to address this, the present paper presents an approach that brings together physical experiment, simulation and analytic predictions. Generally, swarm simulations are categorized depending on their level of abstraction [3] . (i) Microscopic simulations model the behaviour of individual robots and the interaction between robots. (ii) Macroscopic models describe the number of robots in the different behavioural states [5] , [6] . It is useful to distinguish between macroscopic-discrete and macroscopic-continuous methods [7] . Discrete approaches model the count of robots in each one of a finite set of states, while continuous approaches model the (real-valued) fraction of the whole population in each of the states. Macro-discrete models are amenable to a master equation approach and thus typically treated stochastically [6] , [8] , [9] , while macro-continuous approaches typically result from an averaging procedure [8] and are hence deterministic. A further possibility is to have an infinite number of states or continuous state variables. This can be treated deterministically or stochastically [1] . Ideally, a unified approach to modelling a robot swarm will derive parameters for a microscopic description from experiments, derive macroscopic equations from the microscopic model and perform microscopic simulations to validate the macroscopic description [4] . Several factors seem to make such an approach challenging. First, microscopic simulations, while more accessible than physical experiments, generally require substantial computational resources if they involve a large number of robots. Second, physical experiments are expensive, time-consuming and can usually only be conducted under sanitized laboratory conditions [4] . Third, deriving macroscopic descriptions from probabilistic microscopic ones is usually hard, in particular if spatial aspects need to be taken into account [8] , [9] . The present paper explores the feasibility of such a unified approach using the example of a typical collective-decision making problem [1] . We show that, despite the above challenges, such an approach is feasible provided the process under investigation meets specific requirements. We derive our modelling method from first principles based on chemical kinetics. This enables us to analyze the requirements for its applicability and its limitations systematically and in detail. It also allows us to have confidence in the approach beyond a purely empirical justification if these requirements are met. This article builds upon Hamann et al. \n [1] . It extends this work in several regards. Firstly, we present a consistent multi-scale approach that spans microscopic, macroscopic-discrete and macroscopic descriptions. Secondly, we make a first attempt at deriving the aggregate continuous-time Markov model from first principles. This was justified only a posterori in Hamann et al. \n [1] . We do so by transforming the microscopic equations for the individual agent motion into an aggregate macroscopic-discrete reaction system. This transformation is firmly grounded on established techniques from chemical kinetics. This approach vastly improves the confidence we may have in a continuous-time Markov description provided the system under consideration is within well-defined limits. Thirdly, using a mathematical toy model, we show a way to analytically derive the characteristic parameters of the stochastic differential equation for the aggregate model from the macroscopic-discrete description. Several other authors in recent years have applied methods from chemical reaction networks to analyze the dynamics of robot swarms [10] – [15] . Like these earlier works our approach is rooted in the theory of reaction kinetics, but our main concerns differ from these works subtly but importantly. Firstly, we are interested in quantifying transient behaviour as much as steady-state behaviour, for example the time to reach a particular state. Previous work has often only been concerned with steady state behaviour [10] , [13] . Secondly, we are specifically interested in how group behaviour emerges from local interactions and communication between swarm members. Previous work has often explicitly excluded local communication and interaction from the methodology [11] , [12] , [15] . Thirdly, we are interested in systematically abstracting, based on first principles, a spatially extended scenario into a non-spatial macroscopic model without excluding spatial inhomogenities that may emerge from local interactions. Previous work has not explicitly made this connection. [10] explicitly incorporate spatial inhomogenities but restrict these to the ones known a-priori, such as aggregation at a set of predefined boundaries. The paper is structured as follows. We start by introducing the density estimation task and our implementation on a microscopic level using the multi-agent framework flame ( http://www.flame.ac.uk/ ) in Section “Microscopic approach''. Here we also reproduce some of the results from [1] . In Section “Kilobot experiments'' we present results from a physical implementation of the density classification task using a swarm of K ilobots \n [16] . We then ask the question if the microscopic (agent-based) description can be translated into a spatially-homogeneous macroscopic-discrete formalism (Section “The macroscopic-discrete approach''). This section contains our derivation of the macroscopic-discrete Master equation and numerical validations using the software package inchman \n [17] , [18] . Finally, we analytically solve the macroscopic-discrete Master equation for a simplified model and obtain expressions for the coefficients of the associated Fokker-Planck equation (Section “Constructing a time coarse-grained Markov process for the symmetry parameter''). In this way we can justify the aggregate continous-time Markov model and close the loop.",
"discussion": "Discussion In this article we have presented a consistent multiscale approach to modelling a typical decision making scenario in swarm robotics. We perform microscopic simulations of the swarm (Section “Microscopic approach'') and conduct physical experiments using a swarm of K ilobots (Section “Kilobot experiments'') to validate our simulations. Following up on previous work [1] , we identify a symmetry parameter as the fundamental collective swarm variable and tentatively suggest a continuous-time Markov process to describe its evolution. We derive various macroscopic swarm properties, such as decision time and splitting probability, from the time evolution of the symmetry parameter and compare these results to the data obtained from the simulations and the experiments. Extending upon previous work, we identify approximate spatial homogeneity as a key requirement for this type of modelling and investigate the conditions under which it is valid to assume spatial homogeneity. For this regime we derive a macroscopic-probabilistic model using techniques from chemical kinetics (Section “The macroscopic-discrete approach''). We simulate the macroscopic-discrete model and assess its agreement with the previous results. Finally, we detail an approach to deriving the defining properties of the continuous-time Markov process for the symmetry parameter analytically by solving the macroscopic-probabilistic model (Section “Constructing a time coarse-grained Markov process for the symmetry parameter''). Generally, the decision process is approximated well by the continuous-time Markov process. This is a surprising result given the very restrictive assumptions which are required to obtain the aggregate description. We investigate the validity of the multiscale approach for different regimes of the microscopic swarm properties. We identify the limiting factors and derive hard quantitative limits for the applicability of the approach. Firstly, depending on the magnitude of the avoidance radius, the assumption of spatial homogeneity breaks down. The mean free path length is the characteristic quantity for spatial homogeneity. If the mean free path length is much smaller than the container dimensions robots will start to form clusters. One case of this is where the avoidance radius is large. Clustering affects the behaviour of the symmetry parameter and, in general, predictions made based on this collective property will be less accurate. Future work will investigate further if spatial structures can be incorporated into a collective description. Secondly, even if spatial homogeneity is satisfied, the lumped process describing the evolution of the symmetry parameter will, in general, not be memoryless and hence not Markov. The symmetry parameter will generally only be well-approximated if the process is ‘almost Markov’, i.e. the memory influences the macroscopic behaviour only marginally. Our analytic toy model suggest that the Markov property will be satisfied better if the decision depth is larger. Future work will extend this analytic model to accomodate a larger decision depth and eventually the full macroscopic-probabilistic model. We believe that this extension, while cumbersome, should not pose any in-principle difficulties. The analytic treatment of our case study has allowed us to identify the properties under which the proposed empirical parameter estimate of the macroscopic model yields a good approximation of the system's behaviour. For other scenarios these properties would have to be established on a case-by-case basis. How this can be done analytically in other scenarios is beyond the scope of the proposed methodology. In how far the proposed methodology can be generalized thus requires further investigation. Our description hinges on the assumption of spatial homogeneity. Some probabilistic models use homogeneous compartments and incorporate spatial separation by introducing delay times for travelling between compartments [7] . Stochastic delay-differential equations are hard to formulate and even harder to solve and hence we do not expect our approach to work in this case. On the other hand, several studies that assume spatial homogeneity can successfully address problems of collaborative manipulation [8] and task allocation [9] . We believe that our approach can be fruitfully applied to these and related scenarios. Demonstrating the applicability of our approach to these and related scenarios in future work will hopefully allow us to gain increased confidence that this simpler empirical macroscopic method approximates a broad spectrum of application scenarios reliably."
} | 3,245 |
39148012 | PMC11325573 | pmc | 2,981 | {
"abstract": "Background The symbiosis among plants, rhizobia, and arbuscular mycorrhizal fungi (AMF) is one of the most well-known symbiotic relationships in nature. However, it is still unclear how bilateral/tripartite symbiosis works under resource-limited conditions and the diverse genetic backgrounds of the host. Results Using a full factorial design, we manipulated mungbean accessions/subspecies, rhizobia, and AMF to test their effects on each other. Rhizobia functions as a typical facilitator by increasing plant nitrogen content, plant weight, chlorophyll content, and AMF colonization. In contrast, AMF resulted in a tradeoff in plants (reducing biomass for phosphorus acquisition) and behaved as a competitor in reducing rhizobia fitness (nodule weight). Plant genotype did not have a significant effect on AMF fitness, but different mungbean accessions had distinct rhizobia affinities. In contrast to previous studies, the positive relationship between plant and rhizobia fitness was attenuated in the presence of AMF, with wild mungbean being more responsive to the beneficial effect of rhizobia and attenuation by AMF. Conclusions We showed that this complex tripartite relationship does not unconditionally benefit all parties. Moreover, rhizobia species and host genetic background affect the symbiotic relationship significantly. This study provides a new opportunity to re-evaluate the relationships between legume plants and their symbiotic partners. Supplementary Information The online version contains supplementary material available at 10.1186/s12870-024-05488-5.",
"conclusion": "Conclusions Our study revealed that tripartite symbiosis does not unconditionally benefit all parties under all conditions, and the complex relationship does not conform to one universal outcome. Instead, different parts of the tripartite relationship could have distinct results (Fig. 7 ). This provides an interesting starting point to further investigate their complex interactions and the role of each party in the tripartite symbiosis. In short, our study expands the knowledge of the effects of the tripartite symbiosis of V. radiata at the subspecies/accession level, and the investigation of symbiosis between two rhizobia and mungbean accessions provides useful information on the combination with the highest efficiency.",
"discussion": "Discussion Tripartite symbiosis among rhizobia, AMF, and mungbean The benefit of symbionts can vary depending on the environment and timing. Therefore, to reduce the use of chemical fertilizers, it is essential to understand the interactions within the tripartite symbiosis involving AMF, rhizobia, and the host. Additionally, different cultivars may exhibit diverse responses influenced by their genetic factors to symbiotic microbes, leading to distinct preferences for symbiosis. Under adverse conditions, how plants interact with these symbiotic microbes could be one of the most important issues for legume agriculture. To date, most recent studies have focused on the relationships of a single plant accession with few rhizobia/AMF species, and only a few large-scale studies have investigated multiple within-species accessions and multiple symbiotic microbes. Therefore, a study integrating multiple accessions and microbes is needed because the host’s genetic background could be diverse and affect symbiotic outcomes. In this study, we selected five accessions from each of the wild ( V. radiata ssp. radiata ) and cultivated ( V. radiata ssp. sublobata ) mungbean subspecies, representing worldwide genetic variation in this species (Table S1). We aimed to investigate the outcomes of bilateral/tripartite interactions involving mungbean subspecies, AMF, and rhizobia. To minimize the potential influence of various P and N supplies, we performed the entire experiment under the same environment with fixed low P (25 μM) and low N (1 mM) supplies. Low P and low N were used because previous studies showed that sufficient P or N supplies might decrease the symbiotic efficiency between the microbe and the host [ 32 , 42 ], which could introduce uncertainty in the evaluation of the nutrient uptake ability of the microbes. The results indicate that under the experimental conditions, Bj was a more beneficial partner of mungbean than Ri , although the benefit of Bj depends on the specific mungbean accession. Under the tripartite interaction, the nodule number and FW of Bj decreased when mungbean was inoculated with Ri , whereas the colonization rate of Ri increased when mungbean was co-inoculated with Bj . Overall, during the timelapse of the experiment, the presence of Bj increased the FW, RCC, and N in the host, but decreased P. In contrast, Ri negatively affected FW and RCC, but increased P (Fig. 7 ). Fig. 7 Summary of the tripartite symbiosis among mungbean, rhizobia and AMF. Mungbean genetic background affected rhizobia but not AMF colonization. Rhizobia generally had positive effects on the two other parties except for decreasing phosphorus content in plants. AMF generally had negative effects on the two other parties except for increasing phosphorus content in plants Among the hypotheses on plant–microbe interactions, the latest hypothesis stands that short-term competition in the biosphere tends to evolve into a long-term cooperative system, increasing biodiversity and decreasing ecological conflicts, called “cooperative equilibrium [ 43 ].” When the environment changes, especially when the N:P:C ratio is altered, mutualism might become parasitism in plant–microbe interactions [ 3 , 11 , 44 ] until a new equilibrium is formed. The previous studies showed that under certain degrees of available N and P or under natural field conditions, AMF is a beneficial partner to mungbean in tripartite symbiosis [ 9 , 27 – 30 ]. However, under N- and P-limited conditions, our mungbean-rhizobia-AMF system did not conform to the previous results. The cooperative equilibrium was broken under this condition. Although the existence of Bj benefited the mungbean, Ri reduced the biomass of the host while increasing the host’s P content. The symbiosis between Ri and mungbean shifted from mutualism to parasitism, showing the nutrient tradeoff symptoms between these two organisms. Therefore, under N- and P-limited conditions, Bj could be a stable partner of mungbean with fewer ecological conflicts, whereas the mutualism between Ri and mungbean largely depended on the existing nutrients in the environment. Further study is necessary to address the complexity of outcomes in this tripartite system. Wild and cultivated mungbean subspecies exhibit notable distinctions in their phenotypic and genetic characteristics, which may lead to varying responses to symbiotic interactions. Wild mungbean generally appears to be more plastic, exhibiting stronger negative and positive impacts from AMF and rhizobia, respectively. For example, AMF treatment reduces the positive effects between plants and the rhizobia fresh weight, and the effect appears to be much stronger in the wild than in cultivated mungbean (Fig. 3 ). Although the lack of rhizobia decreased the leaf chlorophyll content over time, the effect also appeared to be stronger in the wild group than in the cultivar group (Fig. 4 A). In summary, the findings suggest that wild mungbean varieties may exhibit greater sensitivity to microbes, while cultivated mungbean subspecies could display reduced responsiveness to these microorganisms. Moreover, cultivated mungbean demonstrated enhanced nutrient uptake efficiency, as evidenced by their significantly greater biomass and higher total N and P contents (Fig. 3 & Fig. S3). Previous studies have shown that dual inoculation of AMF and rhizobia could promote legume growth in Medicago truncatula [ 8 ], V. radiata [ 9 ], Glycine max , Phaseolus mungo , Cicer arietinum , Lens culinaris and Phaseolu mungo [ 12 ]. Here, we found that the positive Bj effect and negative Ri effect are not simply additive: strong interaction effects exist for many plant traits. Regardless of whether we modeled the effect of rhizobia in terms of qualitative treatment effect or quantitative nodule weight, we observed similar trends in which Ri had detrimental effects on plant weight, but rhizobia ameliorated these negative effects (Fig. 3 & Fig. S3). Afkhami et al . [ 8 ] modeled the relationship between plant fitness and rhizobia fitness (nodule number), calling this positive relationship the “fitness alignment” between symbionts. The authors showed that AMF could increase the responsiveness of M. truncatula to the beneficial effects of rhizobia, while our study showed that Ri , when it had an overall negative effect, further decreased the responsiveness of V. radiata to the benefits of Bj . In addition, while Afkhami et al . [ 8 ] focused on pod number, we showed that different plant parts have distinct effects. When focusing on those plants inoculated with Bj and modeling nodule weight as Bj fitness, Ri had a negative impact on the relationship between plants and Bj only in plant shoots but not roots, and the effect appeared to be specific to the wild subspecies (Fig. 3 & Table S7). In this experiment, Ri showed a characteristic position in the mutualist-parasite spectrum, where it increased the P content but reduced the biomass of both the plants and Bj . In contrast to previous studies, here, we show that the effect is not universal but could differ among genetic backgrounds and plant parts. Noticeably, we only used Ri as an AMF in this study, and it remains unclear whether the outcome will be similar to that of other AMFs. Although Ri has been reported to increase the biomass of mungbean [ 27 ], it showed opposite effects in the conditions we used. A previous study also reveals that the mixture of AMF inoculum (a mixture of five species of AMF, including Ri ) has no significant effects on the mungbean, while the dual inoculation of rhizobia and AMF increases the yield [ 28 ]. It is unclear whether the different AMFs have different mutualist-parasite spectrums to the mungbean, or whether the results we observed could apply to the field soil conditions (we used pots with sterile sands in our system). In addition, AMF has been reported to benefit plant hosts with different growth periods [ 45 ]. The syngeneic effects of dual inoculation of AMF and rhizobia were observed after 11 wpi in mungbean in the previous works [ 29 , 30 ]. In our experiments, we only recorded traits as long as 6 wpi, and it is unknown if Ri ’s effects would become positive at the prolonged time points. Further studies are needed to clarify these issues. Fitness of rhizobia species toward various mungbean accessions In our study, we found that Bj had various effects on different mungbean accessions, but how does Bj affect other rhizobia species? Among the dominant genera of rhizobia, Sinorhizobium and Bradyrhizobium are classified into a fast-growing rhizobia and a slow-growing rhizobia, respectively [ 16 ]. In previous studies, Bj has been shown to be more dominant than Sf when two rhizobia species co-inoculate the same host plant [ 17 – 21 , 35 ]. On the other hand, compared with Bj , Sf could form more nodules on soybeans at the same bacterial concentration [ 35 ]. Our studies showed that Bj formed more nodules than Sf (Table S9). Additionally, compared with Sf , Bj promoted more RCC in mungbean and was less specific to host genotypes, indicating that mungbean might have a more efficient symbiosis with slow-growing Bj than with fast-growing Sf . Our data showed that many accessions had strong symbiosis with Bj but had no or very low affinity for Sf (Fig. 6 B). Thus, these accessions could be candidates for studying the mechanism of symbiosis between the host and rhizobia."
} | 2,968 |
29468690 | PMC5887887 | pmc | 2,982 | {
"abstract": "Summary \n There is consensus that plant species richness enhances plant productivity within natural grasslands, but the underlying drivers remain debated. Recently, differential accumulation of soil‐borne fungal pathogens across the plant diversity gradient has been proposed as a cause of this pattern. However, the below‐ground environment has generally been treated as a ‘black box’ in biodiversity experiments, leaving these fungi unidentified. Using next generation sequencing and pathogenicity assays, we analysed the community composition of root‐associated fungi from a biodiversity experiment to examine if evidence exists for host specificity and negative density dependence in the interplay between soil‐borne fungi, plant diversity and productivity. Plant species were colonised by distinct (pathogenic) fungal communities and isolated fungal species showed negative, species‐specific effects on plant growth. Moreover, 57% of the pathogenic fungal operational taxonomic units (OTUs) recorded in plant monocultures were not detected in eight plant species plots, suggesting a loss of pathogenic OTUs with plant diversity. Our work provides strong evidence for host specificity and negative density‐dependent effects of root‐associated fungi on plant species in grasslands. Our work substantiates the hypothesis that fungal root pathogens are an important driver of biodiversity‐ecosystem functioning relationships.",
"introduction": "Introduction The diversity of plant species within natural grasslands is known to enhance plant productivity (i.e. biomass), both above and below ground (Tilman et al ., 2006 ; Cardinale et al ., 2007 ; Mueller et al ., 2013 ; Ravenek et al ., 2014 ). The resultant positive relationship between biodiversity and productivity is often explained by resource partitioning; as individual plant species differ in resource use and acquisition, the more species grown together in a community, and the more resources the plant community can obtain (Parrish & Bazzaz, 1976 ; Berendse, 1982 ; Dimitrakopoulos & Schmid, 2004 ; Oram et al ., 2018 ). More recently, an alternative hypothesis was suggested, where pathogenic soil biota play a key role in regulating the relationship between plant diversity and productivity (Maron et al ., 2011 ; Schnitzer et al ., 2011 ). The pathogen hypothesis is built upon two assumptions about plant–pathogen interactions within plant communities. The first is that plant species accumulate species‐specific pathogens, referred to as host specificity. These species‐specific pathogens can reduce the performance of their host species, but have little impact on other plant species. The second assumption is negative density dependence, which suggests that the accumulation of plant species‐specific pathogens and the negative impact on host performance declines with decreasing relative abundances of their host plants. Hence, along a gradient of plant species richness, pathogen pressure, potentially limiting biomass production, is high at low species richness (i.e. in monocultures where the relative abundance of the host is 100%) and expected to decrease in plant species‐rich mixtures, where the relative abundance of host plants will be lower (Kulmatiski et al ., 2012 ; Bever et al ., 2015 ). This ‘pathogen hypothesis’ is analogous to crop yield reductions from pests in agriculture, which typically increase with repeated cultivation of the same crop on the same field, and can be reduced by rotations with different crops (Bullock, 1992 ). So far, the effect of pathogenic soil biota on the plant diversity–productivity relationship has mainly been investigated using a ‘black box’ approach, in which the growth response of plants on soil with an intact (but unknown) soil microbial community is compared to that on soils with their associated soil microbial community eliminated via sterilisation or application of fungicide. For example, Maron et al . ( 2011 ) showed that the significant positive relationship between plant species richness and biomass observed on an intact soil was absent on fungicide‐treated soil. Similarly, Schnitzer et al . ( 2011 ) demonstrated that soil sterilisation enhanced plant biomass production at low species richness, leading to a reduction of the positive relationship between plant species richness and biomass on sterilised soil as compared to field soil. Importantly, Schnitzer et al . ( 2011 ) included an additional treatment, in which sterilised soil was re‐inoculated with a soil wash from field soil containing bacteria and saprotrophic and pathogenic fungi, but not mycorrhizal fungi. This re‐inoculation ‘restored’ the positive relationship between plant species richness and biomass to a level similar to that observed on nonsterilised field soil. Together, these two studies suggest that pathogenic soil biota, and soil‐borne fungi in particular, play an important role in regulating the plant diversity–productivity relationship. However, treatments such as soil sterilisation or fungicide application do not resolve the identity of the soil biota involved and, as such, provide only indirect evidence for how they interact with plant species richness. Here, we determined the taxonomic and functional diversity of soil‐borne fungal communities in roots from a 10‐yr‐old biodiversity experiment (van Ruijven & Berendse, 2009 ) using next generation sequencing approaches. We used these data to assess host specificity and negative density dependence of the root‐associated fungal community. It is important to note that these two factors may have contrasting effects on the relationship between plant and fungal species richness. If host specificity is strong and different plant species accumulate different fungal species, mixtures of plant species can be expected to harbour a mix of the fungal communities associated with each of the component plant species, potentially leading to an increase in fungal species richness with increasing plant species richness (Rottstock et al ., 2014 ; Dassen et al ., 2017 ). On the other hand, if negative density dependence leads to a reduction in the accumulation of host‐specific fungal species with increasing plant species richness (i.e. decreasing host density), this may lead to a dilution of host‐specific pathogens in mixtures, resulting in a less positive, neutral or even negative relationship between plant species richness and fungal richness. We focused on the soil‐borne fungal communities associated with the plant roots, as these were hypothesised to be the main pathogenic factors structuring natural plant communities (Gilbert, 2002 ; Alexander, 2010 ; Fisher et al ., 2012 ), but are still a hitherto largely unexplored component of the root microbiome (but see Hannula et al ., 2017 ). Using detailed analyses of the taxonomic and functional composition of the soil‐borne community in experimental grassland plots, our study aims to reveal the main players and their rules of play in promoting biodiversity effects in plant communities. We test the following hypotheses: (1) different plant species harbour different root‐associated fungal (both total and pathogenic) communities (i.e. host specificity); (2) the community composition of (pathogenic) root‐associated fungi from plants grown in species‐rich mixtures is determined by the relative abundance of the host plants (i.e. negative density dependence); and (3) the diversity of root‐associated fungal communities increases with plant species richness. These three hypotheses were investigated by assessing the root‐associated fungal communities from roots of plant monocultures and two‐, four‐ and eight‐plant species mixtures from the Wageningen biodiversity experiment. We combined full characterisation of the fungal communities using next generation sequencing and related it to specific measurement of below‐ground plant abundance on the same samples using quantitative PCR (qPCR) to quantify species‐specific root biomasses on two soil depths. The spatial aspect was explicitly taken into account because plant abundance changes along this spatial gradient, potentially affecting the composition of the fungal community (Jumpponen et al ., 2010 ). We also isolated and identified fungi from symptomatic sections of roots of two selected plant species and performed bioassays to assess species‐specific pathogenicity.",
"discussion": "Discussion Our findings constitute the first steps in opening the ‘black box’ of root‐associated fungal pathogens regulating the positive relationship between plant biodiversity and productivity. By quantifying fungal communities from roots across a gradient of plant species richness, we provide an empirical test of the two assumptions underlying the role of pathogens in biodiversity–productivity relationships. Our results provide strong support for host specificity, because plant species differed in pathogenic fungal community composition in monocultures. This result would predict that when multiple plant species are grown in mixtures their unique species‐specific root‐associated fungal communities would also exist together, and thus fungal species richness would be positively correlated with plant species richness. However, our results demonstrate that when plant species are grown together it does not lead to an increase in the diversity of root‐associated fungal communities. Indeed, 57% of the pathogenic fungal OTUs recorded in plant monocultures were not detected at high plant diversity, consistent with fungal pathogen dilution occurring with increased plant species richness. Host specificity of soil‐borne fungi in monocultures The roots of the plant species growing in monocultures harboured significantly different fungal communities, indicating host specificity. Host specificity was evident at the level of the total fungal community as well as within the main phyla, such as Ascomycota and Basidiomycota. Consistent with the wide host ranges of many arbuscular mycorrhizal fungal species, host specificity was not apparent for the Glomeromycota (Smith & Read, 2008 ; Dumbrell et al ., 2010 ), although future work should confirm this finding with primers that are more specific for this phylum. Within the context of our study it is more important to investigate host specificity at the functional level of fungal guilds (i.e. root pathogens and endophytes). Recently, fungi have begun to be classified and analysed within functional groups based on phylogeny and the literature, and although this field is in its infancy it can provide useful indications of the activities of fungal communities (Nguyen et al ., 2016 ; Hannula et al ., 2017 ). We performed this classification as thoroughly as possible by carefully checking the literature (Domsch et al ., 2007 ; Farr & Rossman, 2014 ). Fortunately, for our functional groups of interest –pathogens and endophytes – there is a large amount of information available due to their economic importance in agriculture, making our classifications more robust. Our study demonstrates host specificity of pathogenic fungal communities in natural grassland species. Inoculation of two plant species with isolates of two of these fungi revealed host‐specific pathogenic effects. These bioassays provide proof of principle that these below‐ground organisms from biodiversity experiments can actually act as pathogens and have differential effects on different plant species. However, such single‐fungus single‐plant bioassays are still relatively simple (see also Sarmiento et al ., 2017 ). The next step would be to determine the effects of multiple fungal interactions on plant growth (e.g. Klironomos, 2002 ; Hersh et al ., 2012 ), as fungal co‐infection is the rule rather than the exception (Oyarzun et al ., 1993 ; Peay et al ., 2013 ). Combining complex culture‐based approaches and next generation sequencing will reveal the most influential fungal pathogenic actors and their modes of play in biodiversity–ecosystem functioning. Work by Gilbert & Webb ( 2007 ) has shown that the chance of fungal infection is inversely related to phylogenetic distance between the host species. Consistent with this finding, we found a significant effect of the plant functional group (i.e. grass vs forbs) on the overall fungal community composition and also on the root pathogens and endophytes. The latter may explain findings from plant–soil feedback studies, showing that plants experience the strongest growth‐reducing effects on soil conditioned by species from the same plant functional group (Petermann et al ., 2008 ; Hendriks et al ., 2013 ). Patterns of host specificity were apparent at both depths, but fungal communities differed between the two soil layers. This finding suggests that niche differentiation over depth occurs among fungi (Jumpponen et al ., 2010 ; Unterseher et al ., 2011 ; Mujic et al ., 2016 ). Future work should reveal whether this depth differentiation is associated with decreased root biomass (Mueller et al ., 2013 ; Ravenek et al ., 2014 ), altered root exudation profiles (Kawasaki et al ., 2016 ) or differences in abiotic conditions (e.g. organic matter content; Cong et al ., 2014 ) across depths. Host abundance effects on soil‐borne fungi in plant mixtures In plant mixtures, 58% of the variation in fungal community composition was explained by the variation in root biomass across different plant species, suggesting the host specificity observed in monocultures may also be important in regulating fungal communities in more diverse plant communities. However, there were only a few direct links between fungal community composition and the density (species‐specific root biomass) of specific hosts. Only the density of the grass A. odoratum had a significant effect on the composition of the total fungal community, as well as on the root pathogenic community. If host density does have an overarching effect on fungal community composition, then additional significant relationships between host densities and fungal community composition would be expected. The lack of relationships may have three causes: First, in experimental systems, like those used in this study, the soil physico‐chemical environment was initially similar across plots. Over time, however, abiotic differences may have emerged that are driven by plant identity and diversity effects (Cong et al ., 2014 ; Lange et al ., 2015 ; Chen et al ., 2017 ). These abiotic differences may also be drivers of the fungal community (Tedersoo et al ., 2014 ; Thomson et al ., 2015 ), rather than the plant species alone. Second, and perhaps more importantly, the fungal community may not only be determined by the density of the host at the plot level, but also by the identity of the closest neighbouring plants (i.e. individuals belonging to other species). Disease transmission is probably determined by the ‘contagiousness’ of neighbours (Otten et al ., 2005 ). For example, some soil‐borne fungal pathogens can persist on a nonhost species without causing disease symptoms. A host plant with such ‘asymptomatic’ neighbours is more likely to become infected than a plant surrounded by a true nonhost, which is not colonised (Malcolm et al ., 2013 ). Closely related neighbours are more ‘contagious’ than others (Gilbert & Webb, 2007 ). As a consequence, the chance of fungal infection increases with decreasing phylogenetic distance between plant species (Parker et al ., 2015 ). However, the set‐up of most biodiversity experiments is such that host density and neighbour identity are simultaneously varied and do not allow the disentanglement of these two components. Third, host density effects on pathogen abundance may only be straightforward when the pathogen has a narrow host range, and not in communities – which include generalist pathogens – as a whole. This is illustrated by the fact that we observed a significant negative relationship between the number of sequence reads of the specialist fungus P. chrysanthemicola and root biomass of its host L. vulgare in mixed root samples. However, we did not find a similar relationship for the grass specialist G. incrustans , a fungus known to have a broader host range than P. chrysanthemicola . The decline in sequence reads of the specialist pathogen ( P. chrysanthemicola ) with decreasing density of its host ( L. vulgare ) is in line with negative density dependence, although quantitative statements based on the number of DNA sequences from next generation sequencing approaches should be treated with caution. So far, negative density dependence has mainly been demonstrated for tree seedlings in tropical (Bell et al ., 2006 ; Mangan et al ., 2010 ; Liu et al ., 2015 ) and temperate (Johnson et al ., 2012 ) forests. In most cases, this is based on seedling survival as a function of either distance from, or density of, adult trees. Empirical data of the abundance of the pathogens involved are scarce, and our study constitutes the first step towards the functional characterisation of the soil‐borne community along a diversity gradient. Fungal diversity along the plant species richness gradient We found no evidence that fungal species richness increased with plant species richness in our study. This contradicts the findings from two other studies. First, Dassen et al . ( 2017 ) found a positive relationship between fungal diversity and plant species richness, but investigated bulk soil rather than roots. The authors attributed the positive relationship to increased diversity of litter in the soil and other abiotic soil conditions, rather than plant species richness. Second, Rottstock et al . ( 2014 ) showed that the number of leaf pathogenic fungal groups increased with plant species richness. However, they reported the cumulative number of pathogen groups found across all plant species, which suggests that the results may be partly due to increased sampling efforts with plant species richness. In our study, equal sequencing effort was employed per sample. For the vast majority of our samples, OTU–rarefaction curves were approaching asymptotes and thus it is unlikely we differentially missed fungal species due to an artefact. The combination of our finding that fungal diversity does not increase with plant species richness, and that each plant species harboured a distinct fungal community in monoculture, suggests that part of the fungal species present in plant monocultures disappear in plant mixtures. Many of these ‘lost’ fungi were found to be pathogens, with 57% of the pathogenic fungal taxa recorded in plant monocultures not detected in highly diverse plant communities. The reduction of pathogenic OTUs with plant species richness suggests that fungal root pathogens have trouble locating their hosts in diverse plant communities. Future perspective Our study is the first attempt to identify the fungi that play a role in plant biodiversity–productivity relationships. We show that soil‐borne fungal pathogens may be reduced at high plant species richness, which may explain at least in part why plant productivity increases with increasing plant species richness, as previously hypothesised (Maron et al ., 2011 ; Schnitzer et al ., 2011 ; Hendriks et al ., 2013 ; Bever et al ., 2015 ). Our work has enhanced the field of research into biodiversity–productivity relationships by providing further evidence of the importance of below‐ground pathogens by validating two underlying assumptions: host specificity and negative density dependence. Moreover, we postulate that the identity of neighbouring plants may be a factor that is particularly important in soil‐borne fungal community assembly. The effect of neighbours on the loss of fungal diversity has been referred to as the root camouflage hypothesis (Gilbert et al ., 1994 ), which poses that fungal root pathogens ‘get lost’ in the ‘jungle’ of roots in complex plant communities. Root systems in species‐rich grassland communities are indeed tightly intermingled (Kesanakurti et al ., 2011 ; Ravenek et al ., 2014 ; Frank et al ., 2015 ). One potential mechanism behind root camouflage could be that increased root biomass of different plant species leads to higher concentrations of root‐secreted antifungal chemicals (Bednarek & Osbourn, 2009 ), which can reduce the build‐up of pathogen communities. Another potential explanation may be that the exudates and volatiles of other plant species and their root microbiomes may interfere with the signalling cues needed for spore germination and/or directional hyphal growth of fungi toward the host root. Future work will have to reveal which mechanism(s) operate below ground. To disentangle the role of host density from the role of neighbours it would be necessary to separately analyse roots (and fungal communities) of different plant species in plant mixtures. Understanding the rules of the ‘hide‐and‐seek’ game played between plants and pathogens below ground is particularly important for biodiversity conservation and the maintenance of ecosystem functioning, but may also help to improve the design of mixed cropping systems (Li et al ., 2014 ; Brooker et al ., 2015 ). Our work suggests that increasing diversity of (phylogenetically different) plants may be an important prerequisite to reducing disease pressure in both natural and agricultural ecosystems."
} | 5,363 |
37636045 | PMC10448109 | pmc | 2,983 | {
"abstract": "Summary Electric syntrophy between fatty acid oxidizers and methanogens through direct interspecies electron transfer (DIET) is essential for balancing acidogenesis and methanogenesis in anaerobic digestion. Promoting DIET using electrically conductive additives proved effective in enhancing methanogenesis; however, its possibility to affect other microbial redox reactions in methanogenic systems has been little studied. This study provides the first confirmation of the electro-syntrophic coupling of sulfide oxidation to S 0 with CO 2 -reducing methanogenesis in sulfur-rich methanogenic cultures supplemented with conductive magnetite (100–700-nm particle size). The H 2 S content in biogas, initially exceeding 5000 ppmv, decreased to below 1 ppmv along with an accumulation of extracellular S 0 (60–70 mg/L; initially <1 mg/L) at a magnetite dose of 20 mM Fe, while there were no significant changes in methane yield. A comprehensive polyphasic approach demonstrated that the S 0 formation occurs through electro-syntrophic oxidation of sulfide coupled with CO 2 -reducing methanogenesis, involving Methanothrix as the dominant methanogen.",
"conclusion": "Conclusions A novel electric syntrophy coupling anaerobic sulfide oxidation to S 0 with electrotrophic CO 2 reduction to methane was confirmed by a polyphasic approach including physicochemical, electrochemical, microscopic, and molecular analyses. Magnetite addition induced microbial oxidation of sulfide to S 0 during the AD of sulfur-rich waste mixtures, significantly decreasing the H 2 S production in biogas. The enrichment of diverse electroactive microorganisms and the increase in ETS activity under magnetite-supplemented conditions suggested the involvement of DIET. Experiments on semi-continuous cultures with methanogenic inhibitors (BES and fluoroacetate) and 13 C-bicarbonate (DNA-SIP) revealed that the anaerobic oxidation of sulfide to S 0 was coupled by DIET with electrotrophic methanogenesis, especially by Methanothrix . This study is the first to confirm a novel electric syntrophy coupling the oxidation of sulfide to S 0 with the electrotrophic reduction of CO 2 to methane in anaerobic environments. Although more research is needed to identify the exoelectrogenic ASOB in syntrophy with electrotrophic methanogens and to better understand their interactions, this newly discovered syntropy likely contributes to the cycling of sulfur in both natural and engineered anaerobic environments.",
"introduction": "Introduction Converting organic waste into methane-rich biogas through anaerobic digestion (AD) is a key process for sustainable waste management and is receiving increasing attention as a viable source of renewable energy. One of the most exciting recent findings in the field of AD is the electric syntrophy between exoelectrogenic bacteria and electrotrophic methanogens via direct interspecies electron transfer (DIET). 1 , 2 , 3 DIET is considered energetically and kinetically more favorable than conventionally known indirect interspecies electron transfer (IIET), which is mediated by electron-rich metabolites, such as hydrogen or formate, 4 although they both play a crucial role in the methanogenic degradation of fermentation products, such as volatile fatty acids (VFAs) and ethanol. 5 An imbalance between the production and consumption of fermentation products leads to digester souring (i.e., a sudden pH drop with an accumulation of acids) and even failure. Therefore, efficient syntrophic methanogenesis via interspecies electron transfer is essential for the stabilization of and energy recovery from organic matter through AD. DIET does not require complex enzymatic steps for producing and utilizing H 2 or formate that is necessary for IIET, and is not inhibited by increased hydrogen partial pressure. 4 , 5 Accordingly, the possibility has been raised that the methanogenic degradation of fermentation products may be improved by promoting DIET. 5 In support of this hypothesis, Kato et al. 2 first reported in 2012 the acceleration of methanogenesis by the addition of (semi)conductive iron oxides in rice paddy soil enrichments. The authors suggested that (semi)conductive iron oxides could promote the electric syntrophy between Geobacter and Methanosarcina by serving as conduits for DIET. Subsequently, many studies have demonstrated the effectiveness of adding conductive materials (mostly carbon- or iron-based) in promoting DIET and thus accelerating methanogenesis in different AD processes. 5 , 6 , 7 DIET via abiotic conductive materials bypasses the need for cell-to-cell electrical contacts via biological components, such as outer membrane cytochromes and conductive pili (e-pili), necessary for biological DIET, 8 which makes adding conductive materials a simple and efficient approach to promote DIET. These findings presented new possibilities for improving the methanogenic performance and stability of anaerobic digesters. Our understanding of the role and mechanism of DIET in methanogenic environments has advanced significantly over the last decade, and engineering DIET using conductive additives is now considered a promising strategy to boost AD, 7 although performance improvement may not be due solely to the promotion of DIET. 9 Studies on the effects of conductive additives in mixed-culture AD processes have thus far focused almost exclusively on the enhancement of methane production and the microbial reactions involved in electro-syntrophic methanogenesis. However, given that electroactive microorganisms are commonly present in different anaerobic environments, including anaerobic digesters, 10 , 11 conductive additives could also affect, either directly (i.e., promoting electric syntrophy) or indirectly (i.e., altering electron flow), microbial redox reactions other than those involved in methane production in AD processes. 12 Therefore, it is likely that the effect of conductive additives on electron flow in methanogenic microbial communities will be complex, particularly when electron sinks other than methanogenesis are abundant. Although changes in electron flow directly influence the performance of anaerobic digesters, these changes and their implications have been studied very little to date. In a recent study, we investigated the effect of adding magnetite (Fe 3 O 4 ), a commonly used conductive material for promoting DIET, during continuous AD of a sulfur-rich waste mixture and observed no apparent enhancement of methane production. 12 Instead, we found that sulfide produced by dissimilatory sulfate reduction was biologically oxidized to elemental sulfur (S 0 ) and accumulated extracellularly in the presence of magnetite, resulting in a significant reduction in the H 2 S content in biogas. This novel finding offers a new perspective for in situ H 2 S removal and S 0 recovery in digesters, which are of particular interest in the management of AD plants. H 2 S inhibits methanogenesis and causes odor and corrosion problems, and therefore the control of H 2 S formation and emission is important for stable AD. 13 , 14 Furthermore, the H 2 S content in biogas (usually from several hundred to several thousand ppmv) must be reduced by a costly gas-cleaning process to meet the standards for different uses of biogas. 15 Based on molecular, electrochemical, and thermodynamic analyses, we proposed a novel electric syntrophy between anaerobic microbial sulfide oxidation to S 0 and methanogenic reduction of CO 2 via magnetite-mediated DIET. Given that naturally occurring conductive minerals, such as metal ores, are abundant in nature, the proposed electric syntrophy may be a new type of sulfur metabolism in anaerobic environments, playing a significant role in the global sulfur cycle. Although strong circumstantial evidence was provided, direct evidence for the proposed electric syntrophy was not obtained in our previous study, which focused primarily on the formation of extracellular S 0 . Recently, Jiao et al. 16 also proposed an electric syntrophy between sulfide-oxidizing bacteria and methanogens through pyrite (FeS 2 )-mediated DIET in batch AD of sewage sludge based on the metagenome-assembled genomes recovered from the digesters. Although the genetic potential of sulfide-oxidizing bacteria to participate in DIET was suggested, direct experimental evidence for the proposed DIET-based syntrophy was not provided. Therefore, the present study aimed to substantiate the electro-syntrophic coupling of sulfide oxidation to electrotrophic methanogenesis in the presence of magnetite during the AD of a waste mixture of sulfur-rich macroalgal biomass and cheese whey ( Figure 1 ). The accumulation of extracellular S 0 from the microbial oxidation of sulfide was reproduced in duplicate lab-scale digesters, and the proposed electric syntrophy was verified by a series of semi-continuous cultures with the digester slurry. This is the first study to show an electro-syntrophic association between anaerobic sulfide oxidation to S 0 and electrotrophic methanogenesis from CO 2 . The findings of this study not only advance our understanding of electro-syntrophic microbial metabolisms in methanogenic environments but also help to improve strategies for engineering DIET in AD processes for enhanced energy recovery and process stability. Figure 1 Overview of the experimental flow and key research questions DIET, direct interspecies electron transfer; Mag, magnetite; BES, 2-bromoethanesulfonate (to inhibit all methanogenic pathways); FAc, fluoroacetate (to selectively inhibit aceticlastic methanogenesis).",
"discussion": "Results and discussion H 2 S removal and S 0 formation Duplicate anaerobic continuously stirred tank reactors (RM1 and RM2) showed very similar reaction profiles with stable organic removal and methane production across the experimental phases with increasing doses (Phases M0, M8 and M20 at 0, 8 and 20 mM Fe, respectively) of magnetite (100–700-nm particle size) ( Table 1 ; Figure 2 ). In line with our previous observation, 12 magnetite addition had no significant effect on methane production but substantially reduced H 2 S production in both reactors. The methane yield remained consistent regardless of the magnetite dose at 0.24–0.25 L/g chemical oxygen demand (COD) fed, corresponding to around 0.37 L/g COD removed, which is close to the theoretical yield of 0.35 L/g COD removed. This indicates that methanogenesis was active and the primary electron sink in the reactors throughout the experiment. The H 2 S content in biogas was over 5,200 ppmv in Phase M0, which is within the typical range of 0.1–2% (v/v) depending on substrate composition but too high for use in engines or vehicles (<1–1000 ppmv). 17 With the addition of magnetite, the H 2 S content fell drastically to around 500 ppmv in Phase M8 and further below 1 ppmv in Phase M20. Meanwhile, the residual concentration of total dissolved sulfide (TDS) remained around 60 mg S/L across the experimental phases ( Table 1 ), which is lower than the inhibitory concentration to methanogens (100–800 mg/L). 17 These results suggest that the reduced H 2 S production in the presence of magnetite was not due to the suppression of dissimilatory sulfate reduction but rather to the removal of produced sulfide. It should be noted that, in Phases M8 and M20, there was an accumulation of extracellular S 0 (61.9–69.9 mg/L) without detectable formation of FeS (<0.1 mg Fe/L), indicating that H 2 S was not removed by FeS precipitation but rather through the oxidation of sulfide to S 0 under magnetite-added conditions. Supporting this observation, the results of cyclic voltammetry (CV) of Phase M20 biomass in RM1, cultivated for 24 h using acetate and sulfate as the sole carbon and sulfur sources, showed a pair of anodic and cathodic peaks (+0.37 and −0.02 V vs. Ag/AgCl) corresponding to the redox reactions between (poly)sulfide and S 0 ( Figure S1 ). 18 In contrast, no peaks relevant to these redox reactions were detected in the cyclic voltammogram of the Phase M0 biomass, further evidencing the essential role of magnetite in mediating these reactions. The experimental results summarized above, which agree with our previous study, 12 reconfirm the oxidation of sulfide to extracellular S 0 in the presence of magnetite under methanogenic conditions. Table 1 Steady-state process performance data for each experimental phase Process parameter Unit Phase M0 Phase M8 Phase M20 RM1 RM2 RM1 RM2 RM1 RM2 COD removal % 66.3 (11.4) a 64.1 (22.7) 62.6 (10.5) 63.2 (9.2) 66.0 (25.7) 65.9 (21.1) Residual VFAs mg COD/L 10.9 (1.4) 9.12 (0.7) ND b , ∗ ND ∗ ND ∗ ND ∗ CH 4 production rate mL/L·d 68.9 (7.0) 68.3 (4.2) 68.1 (8.9) 67.5 (7.8) 64.8 (2.6) 64.0 (3.1) CH 4 yield L/g COD fed 0.25 (0.01) 0.25 (0.01) 0.25 (0.02) 0.25 (0.01) 0.24 (0.0) 0.24 (0.01) CH 4 yield L/g COD removed 0.38 (0.1) 0.38 (0.0) 0.38 (0.1) 0.37 (0.1) 0.37 (0.0) 0.37 (0.0) H 2 S production rate mL/L·d 0.043 (0.01) 0.046 (0.0) 0.005 (0.0) ∗ 0.004 (0.0) ∗ <0.001 ∗ <0.001 ∗ H 2 S content in biogas ppmv 5,283 (683) 5,217 (161) 553 (49) ∗ 484 (42) ∗ <1 ∗,# <1 ∗,# Total dissolved sulfide mg S/L 58.1 (4.4) 57.5 (3.2) 58.9 (6.2) 61.6 (1.2) 58.6 (1.1) 55.4 (4.3) Extracellular S 0 mg/L <1 <1 69.2 (1.3) ∗ 69.9 (0.5) ∗ 63.3 (2.0) ∗ 61.9 (8.0) ∗ Residual magnetite mM Fe ND ND 8.5 (0.1) 8.2 (0.0) 19.7 (0.3) 23.4 (0.5) Symbols indicate statistically significant differences (p < 0.05) compared to Phase M0 (∗) and to Phase M8 (#). a Standard deviation represents in parenthesis. b Not detected. Figure 2 Methane and H 2 S production profiles of duplicate reactors RM1 and RM2 The reactors were operated at increasing magnetite concentrations (Phases M0, M8, and M20 at 0, 8, and 20 mM Fe, respectively). The average extracellular S 0 concentrations of the reactors for each experimental phase under steady-state conditions are presented. Error bars represent standard deviations of duplicate measurements. Microbial sulfide oxidation to S 0 Anaerobic batch experiments using the same inoculum sludge as for RM1 and RM2 were performed with sodium acetate (5 g COD/L) to determine whether the oxidation of sulfide to S 0 is biological or chemical. The experiments were carried out with or without the addition of magnetite, under biotic and abiotic (autoclaved sludge or no inoculation) conditions. During 28 days of cultivation, a noticeable formation of extracellular S 0 was observed only in the biotic cultures dosed with magnetite (20 mM Fe), and the S 0 accumulation was significantly lower when sodium sulfide was added as the sulfur source (11.2 mg S/L) than when sodium sulfate was added (67.5 mg S/L). The difference in S 0 accumulation could be attributed to the volatilization loss of H 2 S in the former, especially during the initial period of incubation, when the sulfide concentration is high. The calculated H 2 S to HS − ratio at equilibrium (pKa = 6.76 at 37°C, 19 ) was approximately 8% in the cultures (initial pH of 7.87), and therefore a considerable amount of sulfide in the medium must have been released as H 2 S into the gas phase during the incubation of the cultures with sodium sulfide. Meanwhile, no detectable amount of S 0 (<1 mg/L) formed in any abiotic culture regardless of the presence or absence of magnetite. These results indicate that the magnetite-aided sulfide oxidation to S 0 was microbially mediated rather than through chemical reactions between sulfide and magnetite, 20 which agrees with our previous observation. 12 Accordingly, the residual magnetite concentration remained unchanged from the dosed levels throughout the experiment in the duplicate reactors ( Table 1 ), and the X-ray diffraction (XRD) profiles of the reactor samples taken on Day 525 (Phase M20) matched that of pure magnetite ( Figure S2 ). These results show that the added magnetite remained intact with no significant dissolution or chemical transformation during the reactor operation. The microscopic examination of the reactor samples revealed that magnetite particles and microorganisms agglomerated to form flock-like aggregates, in comparison with the bare carbon tape where the specimen was loaded ( Figures 3 A and S3 ). Several previous studies on DIET-promoted AD using submicron magnetite particles have reported the formation of conductive cell-magnetite aggregates, 12 , 21 , 22 which shortens the distance between syntrophic partners involved in various microbial redox reactions in methanogenic systems and facilitates their electron exchange. 5 , 23 Additionally, magnetite can compensate for the lack of e-pili and c -type cytochromes in extracellular electron exchange 24 and can facilitate DIET even between microorganisms lacking these biological connectors. 3 , 25 , 26 Accordingly, the electron transport system (ETS) activity increased more than 3-fold, from approximately 4 to over 12 μg/mL⋅min, in both RM1 and RM2 with the addition of magnetite ( Figure 3 B). Comparable levels of ETS activity increase have been reported in other studied using conductive additives to promote DIET in AD. 27 , 28 ETS activity is an indicator of microbial respiration and redox reactions, which has been shown to increase under DIET-promoting conditions with magnetite addition. 29 , 30 , 31 , 32 Taken together, magnetite did not act as an electron shuttle but likely provided cell-to-cell electrical connections for syntrophic sulfide oxidation to S 0 in the duplicate reactors. Fluorescence microscopy with a sulfane sulfur–specific probe, SSP4, confirmed the extracellular deposition of S 0 in proximity to cell-magnetite aggregates ( Figure 3 C), supporting the conversion of sulfide to S 0 via extracellular electron transfer. Figure 3 Scanning electron microscopy images of RM1 and RM2 sludge samples collected in Phase M20 (A) ETS activity of reactor sludge at each experimental phase (B), deposition of extracellular S 0 in RM1 sludge (Phase M8) visualized by fluorescence microscopy with a sulfane sulfur-specific probe SSP4 (C) (A and C) Magnetite particles, cells, and S 0 are indicated by white, red, and green arrows, respectively. (B) Asterisks indicate statistically significant differences (p < 0.05) in ETS activity compared to Phase M0. Error bars present standard deviations of triplicate measurements. Microbial responses to magnetite addition High-throughput sequencing (HTS) was performed to investigate the effects of magnetite addition on the microbial communities in the experimental reactors. A total of 104,133 reads were obtained from rRNA (i.e., cDNA) libraries, and they were clustered into 398 operational taxonomic units (OTUs; 381 bacterial and 17 archaeal). The taxonomic affiliations of major OTUs (>2% of the total bacterial or archaeal reads in at least one sample) are presented in Figure 4 . The microbial community structures of the reactors changed greatly with the addition of magnetite, as seen in the bacterial and archaeal cluster dendrograms generated from the distribution of OTUs, where the community structures in Phase M0 were clustered remotely from those in Phases M8 and M20 ( Figure 5 ). Twenty-three phyla were identified from the bacterial sequences, with Bacteroidetes , Proteobacteria , and Spirochaetes being the dominant phyla, and 28 sequences (28 OTUs) were unclassified at the phylum level. Changes in bacterial community structure after adding magnetite were apparent at the phylum level, with increases (e.g., Bacteroidetes , Spirochaetes , Fibrobacteres , and Planctomycetes ) and decreases (e.g., Proteobacteria , Cloacimonetes , and Hydrogendentes ) in the relative abundance of different phyla. Notably, the abundance of the putative electroactive genus Paludibacter 33 , 34 (OTUs B1 and B3) increased dramatically from less than 10% to nearly 50% between Phases M0 and M8 ( Figure 4 ). Although present in low numbers, electroactive Geobacter and Sphaerochaeta genera, 35 which were not detected in Phase M0, also exhibited a significant increase in relative abundance under magnetite-added conditions during Phases M8 and M20 ( Geobacter , 0.3–0.5%; Sphaerochaeta , 0.5–4.2%) ( Figure S4 A). Interestingly, magnetite addition also significantly affected the community structure of sulfate-reducing bacteria (SRB). The relative abundance of Desulfomicrobium , the major SRB genus comprising more than 20% of the total bacteria in Phase M0, decreased sharply to below 2% in the following phases, whereas that of the genus Desulfobulbus remained at similar or higher levels in the presence of magnetite ( Figures 4 and S4 B). These different responses can be explained by the fact that Desulfobulbus species possess the omcF gene encoding a c -type outer membrane cytochrome for extracellular electron transfer and are probably DIET-active in anaerobic sulfate-containing environments, 36 but Desulfomicrobium species have not been shown to grow syntrophically. 37 The above results suggest that magnetite enriched diverse electroactive bacteria and promoted electro-syntrophic associations via DIET in the reactor microbial communities. Figure 4 Taxonomic affiliation and relative abundance (%) of major OTUs (>2% in at least one archaeal or bacterial library) Cells with relative abundance values are colored in a heatmap-like fashion: green for archaeal and red for bacterial sequences. OTU, operational taxonomic unit. a The lowest rank assigned by UCLUST against the RDP database (p, phylum; c, class; f, family; g, genus). b Not detected at all (zero read). Please note that ‘0.0’ means non-zero read, but in very low relative abundance (<0.1%). Figure 5 Taxonomic distribution of retrieved sequences and cluster dendrograms constructed from the distribution of OTUs in the bacterial (A) and archaeal (B) 16S rRNA gene libraries of reactor samples Each sample is labeled with the corresponding reactor name and experimental phase. Sequences with a relative abundance lower than 1% in all samples were classified as ‘Others’. Bootstrap values higher than 70% (1,000 replicates) are shown at the nodes in the dendrograms. It is noteworthy that known anaerobic sulfide-oxidizing bacteria (ASOB), such as purple sulfur bacteria, purple non-sulfur bacteria, green sulfur bacteria and nitrate-dependent sulfide-oxidizing bacteria, 38 , 39 were not detected in any of the reactor samples, although a considerable amount of extracellular S 0 was generated in Phases M8 and M20. The absence of these ASOB is understandable given that the experimental reactors were run anaerobically in the dark with a negligible amount of nitrate (0.0–0.3 mg NO 3 − -N/L). It is therefore likely that the sulfide oxidation to S 0 in the presence of magnetite was mediated by unknown ASOB, presumably such as the populations belonging to the phyla Fibrobacteres and Planctomycetes . The relative abundance of these phyla increased significantly with the addition of magnetite ( Figure S4 C), as observed in our previous study. 12 \n Fibrobacteres shares a unique common ancestor and is closely related with the phylum Chlorobi , which contains green sulfur bacteria (i.e., Chlorobiaceae ) that phototrophically oxidize sulfide and deposit extracellular S 0 globules. 40 , 41 These bacteria have flavocytochrome c -sulfide dehydrogenate catalyzing H 2 S-dependent cytochrome c reduction, which enables the sulfide oxidation to S 0 outside the cells. 42 Given that outer membrane c -type cytochromes play an important role in extracellular electron transfer in exoelectrogens 24 and many neighboring families of Chlorobiaceae are chemoheterotrophic, 43 , 44 it could be possible that unknown Fibrobacteres bacteria chemotrophically contributed to the oxidation of sulfide to S 0 through DIET-based syntrophy. 12 In support of this suggestion, the growth of green sulfur bacteria in light-limited anaerobic environments, such as deep seawater and hydrothermal vents, has been observed. 41 , 45 , 46 \n Planctomycetes species have membrane-bound polysulfide reductase oxidizing sulfide to S 0 , which is common in known ASOB, such as Wolinella and Chlorobium species. 47 , 48 \n Planctomycetes cells have unique electron-dense crateriform structures scattered over the cell surface 43 and a comparative genomic study found a putative extracellular electron transfer gene cluster (PCC4) in Planctomycetes species. 49 It is therefore possible that unknown Planctomycetes bacteria participated in the electric syntrophy for the oxidation of sulfide to S 0 . The methanogenic community was dominated by the family Methanotrichaceae (>86% of the total archaea in all 16S rRNA libraries) across the experimental phases ( Figure 5 B). Of note is the drastic dominance shift between two major Methanothrix OTUs (A1 and A2) with the addition of magnetite ( Figure 4 ). OTU A1 was the most abundance methanogenic population, accounting for more than 90% of the total archaeal sequences in Phase M0; however, its relative abundance dropped to 2.1–8.4% in Phases M8 and M20. Conversely, OTU A2, which was not detected in the absence of magnetite, became the most abundant methanogenic population in the presence of magnetite. Meanwhile, correlation analysis based on the relative abundance of OTUs found that OTU A2 correlated positively with the magnetite dose and negatively with the H 2 S content in biogas (p < 0.05), whereas OTU A1 showed the opposite pattern ( Figure S5 ). Given the recent finding that Methanothrix , previously thought to be strictly aceticlastic, can electrotrophically reduce CO 2 to methane. 50 These results suggest that the addition of magnetite stimulated OTU A2, which is possibly more electroactive and less tolerant to sulfate-reducing conditions, over OTU A1. Electro-syntrophic methanogenesis through magnetite-mediated DIET appears to have developed as a major route of methane production in Phases M8 and M20. Furthermore, OTU A2 also had significant positive correlations with the extracellular S 0 concentration and the Planctomycetaceae -related OTU B18 (p < 0.05). This supports the possibility noted above that unknown Planctomycetes bacteria, particularly OTU B18, were involved in the syntrophic oxidation of sulfide to S 0 , which was likely coupled to electrotrophic methanogenesis through DIET. Meanwhile, the relative abundance of the hydrogenotrophic order Methanomicrobiales represented by OTU A4 increased notably with the addition of magnetite ( Figures 4 and 5 B), and it had a significant positive correlation with the magnetite dose (p < 0.05) ( Figure S5 ). An enrichment of Methanomicrobiales in the presence of conductive material has been reported in several studies 51 and a recent study revealed that Methanospirillum hungatei belonging to this order has electrically conductive extracellular filaments, although their function has yet to be elucidated. 52 Therefore, although the DIET ability of Methanomicrobiales species has not been verified in defined co-cultures, 53 magnetite may have stimulated the electro-syntrophic growth of Methanomicrobiales . It may also be possible that the enrichment of syntrophic VFA-oxidizing Syntrophaceae populations (OTUs B7, B16 and B17), possibly capable of both IIET and DIET, 54 , 55 facilitated the growth of hydrogenotrophic methanogens in Phases M8 and M20. Overall, it appears that magnetite addition created a DIET-promoting environment and enriched different electroactive microorganisms in the experimental reactors. Electro-syntrophic coupling of sulfide oxidation and electrotrophic methanogenesis Calculations based on the formal reduction potential (E 0 ′, vs. standard hydrogen electrode at pH 7°C and 25°C) indicated that, among many reduction reactions commonly occurring in AD processes, the reductions of CO 2 to CH 4 and of SO 4 2− to H 2 S can be coupled with the oxidation of sulfide to S 0 to form a spontaneous redox process (Reactions 1–4). 56 , 57 S 0 + H + + 2e – → HS − E 0 ′ = −0.270 V (Reaction 1) S 0 + 2H + + 2e – → H 2 S E 0 ′ = −0.274 V (Reaction 2) SO 4 2− + 10H + + 8e – → H 2 S + 4H 2 O E 0 ′ = −0.220 V (Reaction 3) CO 2 + 8H + + 8e – → CH 4 + 2H 2 O E 0 ′ = −0.244 V (Reaction 4) However, as in our previous study, 12 only the electrotrophic reduction of CO 2 to CH 4 (Reaction 4) was able to make the overall reaction thermodynamically favorable (E 0 ′ cell = 0.132–0.346 V in Phases M8 and M20) when taking into account the measured concentrations of reactants and products in the reactors. This result suggests that electrotrophic methanogenesis was likely the electron-accepting reaction coupled with the anaerobic oxidation of sulfide to S 0 through DIET. To verify the formation of the proposed electric syntrophy, semi-continuous cultures (140-mL working volume and 20-day hydraulic retention time [HRT]) inoculated with the sludge taken from RM1 in Phase M20 were incubated with or without the continuous addition of magnetite (20 mM Fe) ( Figure 1 ). During 28 days of incubation, the cultures with (M) or without (C) magnetite addition showed comparable sulfate-reducing activities (i.e., comparable TDS concentrations) and methane productivities ( Table 2 ). Meanwhile, in C cultures, the extracellular S 0 concentration decreased over time (4.8 mg/L on Day 28), along with an accumulation of H 2 S (3,577 ppmv in biogas), as the magnetite derived from the RM1 sludge was washed out with the effluent ( Figure 6 ). However, in M cultures, the extracellular S 0 concentration increased with cultivation and reached approximately 60 mg/L, with the H 2 S content of the biogas remaining low at 69–126 ppmv. These results agree with the observations in RM1 and RM2 and further confirm the essential role of magnetite in the anaerobic oxidation of sulfide to S 0 . Notably, MB cultures supplemented with magnetite and 2-bromoethanesulfonate (BES; 50 mM), a potent inhibitor of all methanogenic pathways, 58 which were tested in parallel with C and M cultures, showed a significantly faster decrease in the extracellular S 0 concentration (complete removal in 21 days of incubation) than C cultures. Virtually no methane was produced in MB cultures, with the H 2 S content in biogas remaining comparable to that in C cultures. Therefore, it was demonstrated that active methanogenesis is essential for the anaerobic oxidation of sulfide in the presence of magnetite. For a deeper understanding, M cultures were supplemented with fluoroacetate (20 mM), which selectively inhibits aceticlastic methanogenesis, 58 and renamed MF cultures. Interestingly, the extracellular S 0 concentration in MF cultures did not decrease over time and remained at levels comparable to those in M cultures during 28 days of semi-continuous cultivation. MF cultures still produced methane, although the methane content and production rate were lower compared to C and M cultures, while maintaining a low H 2 S content in biogas (50–136 ppmv). In addition, sulfate reduction was active in all tested cultures with comparable TDS concentrations, regardless of the presence or absence of magnetite or methanogenesis inhibitors ( Table 2 ). Overall, the above results confirm that CO 2 -reducing methanogenesis, not aceticlastic methanogenesis, is the main reduction reaction driving the syntrophic oxidation of sulfide to S 0 in the presence of magnetite. Table 2 Production of methane, H 2 S, and total dissolved sulfide in semi-continuous bottle cultures with or without magnetite and methanogenic inhibitors Culture Magnetite (20 mM Fe) Inhibitor a CH 4 production rate (mL/d) CH 4 content in biogas (%) H 2 S production rate (mL/d) H 2 S content in biogas (ppmv) Total dissolved sulfide (mg/L) C – b – 5.1–7.0 10.8–39.6 0.03–0.06 2,196–3,577 61.0–70.3 M + c – 4.6–7.3 10.5–40.8 0.001–0.003 ∗,# 69–126 ∗,# 59.6–71.0 MB + BES 0.03–0.06 ∗,†,‡ 0.3–0.4 ∗,†,‡ 0.02–0.03 †,‡ 1,292–3,489 †,‡ 59.5–66.3 MF + Fluoroacetate 3.8–6.5 11.9–14.6 0.001–0.003 ∗,# 50–136 ∗,# 46.1–62.0 Symbols indicate statistically significant differences (p < 0.05) compared to C cultures (∗), MB cultures (#), M cultures (†), and MF cultures (‡). a BES to inhibit all methanogenic pathways and fluoroacetate to selectively inhibit the aceticlastic pathway. b Not added. c Added. Figure 6 Temporal changes in extracellular S 0 concentrations in semi-continuous bottle cultures with or without magnetite and methanogenic inhibitors: M cultures supplemented with magnetite (20 mM Fe), MB cultures supplemented with magnetite and BES (50 mM), and C cultures with neither magnetite nor BES (control) After 28 days of cultivation, M cultures were supplemented with fluoroacetate (1 mM) and further incubated for 28 days (referred to as MF cultures). Error bars are standard deviations of quadruplicate cultures. Symbols indicate statistically significant differences ( p < 0.05) compared to C cultures (∗), MB cultures (#), M cultures (†), and MF cultures (‡). The 16S rRNA-targeted HTS results of the biomass samples taken from the tested cultures at the end of the experiment revealed that Methanothrix and Methanobacterium dominated the methanogenic communities in C and M cultures (>91% of the total archaeal reads) ( Figure 7 ). As expected, there were virtually no active methanogens in MB cultures (≤3 archaeal reads out of >13,024 total prokaryotic reads), while several hundreds and thousands of archaeal reads were identified in all other cultures. Complete inhibition of methanogenesis disturbs the interspecies electron transfer required for the anaerobic degradation of VFAs and suppresses the activity of syntrophic VFA oxidizers. 5 Under the imbalanced conditions in MB cultures (≤3,850 mg VFAs as COD/L), fiber-degrading Anaerosporobacter occurred as the dominant fermentative bacteria producing VFAs ( Figure S6 ). 59 The methanogenic communities in MF cultures were dominated by hydrogenotrophic methanogens, especially Methanobacterium , which is reasonable because acetate is converted to methane by syntrophic acetate oxidation coupled with hydrogenotrophic methanogenesis when aceticlastic methanogenesis is inhibited. 60 , 61 Correspondingly, the relative abundance of (putative) syntrophic acetate-oxidizing bacteria belonging to the class Clostridia 62 , 63 , 64 , 65 increased in MF cultures ( Figure S6 ). Recently, a Methanobacterium strain capable of growing via DIET with Geobacter metallireducens was reported, although all other strains tested so far were not electroactive. 66 This finding suggests the possibility that Methanobacterium thrived in electric syntrophy with ASOB or other electroactive bacteria. However, given that its relative abundance was much lower in the cultures without inhibitors, the predominance of Methanobacterium in MF cultures was likely due to the inhibition of aceticlastic methanogens rather than the stimulation of its electro-syntrophic growth. An interesting observation is that a notable amount of Methanothrix remained in MF cultures, which contrasts with virtually no detection of methanogens in MB cultures ( Figure 7 ). Given that electrotrophic methanogenesis is the only non-aceticlastic pathway known so far for Methanothrix , 50 this result strongly suggests the electrotrophic growth of Methanothrix in MF cultures. Figure 7 Taxonomic distribution of retrieved archaeal sequences at the genus level in the 16S rRNA gene libraries for semi-continuous cultures with or without magnetite and methanogenic inhibitors: M cultures supplemented with magnetite (20 mM Fe), MB cultures supplemented with magnetite and BES (50 mM), MF cultures supplemented with magnetite and fluoroacetate (1 mM), and C cultures supplemented with none of them (control) Samples for analysis were taken from each of the quadruplicate bottles at the end of the incubation. The average numbers of prokaryotic and archaeal sequences (average ±standard deviation) from quadruplicates are indicated on top of the bars (prokaryotes/archaea). Sequences with a relative abundance lower than 1% in all samples were classified as ‘Others’. MB cultures were excluded from analysis since virtually no archaeal sequences were detected (≤3 out of ≥13,024 prokaryotic sequences). Noteworthy is that the relative abundance of Methanothrix was significantly lower in MF cultures than in C or M cultures, although MF cultures maintained comparable extracellular S 0 levels to M cultures ( Figures 6 and 7 ). This result suggests that electro-syntrophic methanogenesis using sulfide as an electron donor could support substantially less growth of Methanothrix in MF cultures compared to its growth in C and M cultures, in which aceticlastic methanogenesis was not inhibited. The limited contribution of electro-syntrophic methanogenesis to the growth of Methanothrix , despite the efficient oxidation of sulfide to S 0 , under magnetite-added conditions can be attributed to the production of very marginal amounts of H 2 S (i.e., limited electron donor availability) compared to methane in the biogas (<0.07%, v/v) ( Table 1 ). Based on the assumption that all electrons generated during sulfide oxidation to S 0 were used for CO 2 -reducing methanogenesis (Reactions 2 and 4), the resulting increase in methane production rate can only be estimated to be approximately 3 mL per day, drawing upon the results obtained from Phases M8 and M20 under steady-state conditions. 12 This finding corresponds to the fact that no significant increase in methane production was observed as a consequence of promoted electro-syntrophic methanogenesis in the experimental reactors, although extracellular S 0 accumulated with the addition of magnetite ( Figure 2 ; Table 1 ). Electrotrophic Methanothrix as electron acceptor for sulfide oxidation To verify the contribution of Methanothrix to the electro-syntrophic oxidation of sulfide to S 0 , a set of DNA stable-isotope probing (SIP) experiments was performed. The quadruplicates of C cultures were divided into two groups of duplicate cultures amended with 13 C-bicarbonate, one with magnetite addition (SM) and the other without (SC), and they were cultivated semi-continuously in the same manner as C cultures ( Figure 1 ). In agreement with the observations in C and M cultures, a significant formation of extracellular S 0 was observed only in SM cultures. After three volume turnovers to allow sufficient time for metabolizing 13 C-bicarbonate, the 16S rRNA gene concentrations of Methanothrix and archaea were analyzed by digital polymeric chain reaction (dPCR) for each culture. Methanothrix accounted for the majority of 13 C-labeled archaea, with even higher Methanothrix -to-archaea ratios than those in the 12 C-DNA libraries, in SM cultures, whereas the presence of 13 C-labeled Methanothrix (and other methanogens) was negligible in SC cultures ( Table 3 ). This result confirms that magnetite promoted the electrotrophic growth of Methanothrix given that it is the only route through which Methanothrix can utilize CO 2 for methanogenesis. 3 Although acetate generated by homoacetogenesis from H 2 and CO 2 might be aceticlastically utilized by Methanothrix , 67 the large difference in the enrichment of 13 C-labeled Methanothrix between SC and SM cultures shows that its effect is marginal. Table 3 Concentrations of archaeal and Methanothrix 16S rRNA genes in the 13 C- and 12 C-DNA fractions from DNA-stable isotope probing experiments with 13 C-bicarbonate in semi-continuous bottle cultures with or without magnetite Fraction C6ulture Magnetite a 16S rRNA gene concentration (copies/g DNA) Methanothrix /archaea ratio (%) Archaea Methanothrix 13 C SC1 – ND b ND n.a. c SC2 – ND ND n.a. SM1 + 8.8 × 10 12∗ 6.0 × 10 12∗ 68.5% ∗ SM2 + 1.4 × 10 13∗ 1.2 × 10 13∗ 83.4% ∗ 12 C SC1 – 1.4 × 10 13 8.9 × 10 12 61.8% SC2 – 1.5 × 10 13 9.1 × 10 12 59.0% SM1 + 7.4 × 10 12 3.8 × 10 12 51.1% SM2 + 1.0 × 10 13 6.4 × 10 12 64.6% Asterisks indicate statistically significant differences (p < 0.05) compared to SC cultures. a +, added (20 mM Fe); –, not added. b Not detected. DNA concentration was below the detection limit (<0.010 μg/mL). c Not applicable. The results of the series of experiments discussed above collectively demonstrate that the anaerobic oxidation of sulfide to S 0 can be driven by electrotrophic methanogenesis, especially by Methanothrix , via mineral-mediated DIET in sulfur-rich methanogenic environments. This study is the first to demonstrate the occurrence of this type of electric syntrophy, uncovering a new route for anaerobic sulfur metabolism. Sulfur is an essential element in all living organisms and is widely distributed in the Earth’s crust (approximately 0.05% of the lithosphere’s weight). 68 Given the abundance of conductive minerals, such as magnetite and other metal ores, in nature, it is likely that electro-syntrophic sulfide oxidation plays a significant role in the global sulfur cycle and affects the fate of other elements, particularly in anaerobic environments – for example, carbon in methanogenesis. Furthermore, this newly found electric syntrophy presents an interesting possibility for in situ H 2 S control and S 0 recovery in anaerobic digesters, leading to improved process stability and economy. 12 While substantial further research is required for practical application, the magnetic separation of cell-magnetite aggregates from the effluent may offer a feasible strategy for S 0 recovery and microbial retention. 21 As numerous electro-syntrophic associations exist and respond differently to environmental conditions in anaerobic digesters, understanding the interactions among electroactive microorganisms is important in engineering DIET to enhance digester performance and stability. Conclusions A novel electric syntrophy coupling anaerobic sulfide oxidation to S 0 with electrotrophic CO 2 reduction to methane was confirmed by a polyphasic approach including physicochemical, electrochemical, microscopic, and molecular analyses. Magnetite addition induced microbial oxidation of sulfide to S 0 during the AD of sulfur-rich waste mixtures, significantly decreasing the H 2 S production in biogas. The enrichment of diverse electroactive microorganisms and the increase in ETS activity under magnetite-supplemented conditions suggested the involvement of DIET. Experiments on semi-continuous cultures with methanogenic inhibitors (BES and fluoroacetate) and 13 C-bicarbonate (DNA-SIP) revealed that the anaerobic oxidation of sulfide to S 0 was coupled by DIET with electrotrophic methanogenesis, especially by Methanothrix . This study is the first to confirm a novel electric syntrophy coupling the oxidation of sulfide to S 0 with the electrotrophic reduction of CO 2 to methane in anaerobic environments. Although more research is needed to identify the exoelectrogenic ASOB in syntrophy with electrotrophic methanogens and to better understand their interactions, this newly discovered syntropy likely contributes to the cycling of sulfur in both natural and engineered anaerobic environments. Limitations of the study While we demonstrated in this study the occurrence of a novel electric syntrophy, coupling anaerobic sulfide oxidation to S 0 with electrotrophic CO 2 reduction to CH 4 , using a polyphasic approach, we were unable to directly observe the flow of electrons from the former to the latter reactions due to technological limitations. Although Methanothrix was identified as the major electrotrophic methanogen involved in this syntrophy, the possibility of other microorganisms directly accepting electrons from the sulfide oxidation cannot be ruled out due to the complex nature of methanogenic microbial communities. We did not identify the exoelectrogenic ASOB in syntrophy with Methanothrix , despite proposing Fibrobacteres -and Planctomycetes -related bacteria as potential candidates. These outstanding questions merit further research for a deeper understanding of the electro-syntrophic interactions in anaerobic microbial communities."
} | 11,054 |
31447806 | PMC6691176 | pmc | 2,984 | {
"abstract": "The most common quorum sensing (QS) system in Gram-negative bacteria consists of signaling molecules called N- acyl-homoserine lactones (AHLs), which are synthesized by an enzyme AHL synthase (LuxI) and detected by a transcriptional regulator (LuxR) that are usually located in close proximity. However, many recent studies have also evidenced the presence of LuxR solos that are LuxR-related proteins in Proteobacteria that are devoid of a cognate LuxI AHL synthase. Pandoraea species are opportunistic pathogens frequently isolated from sputum specimens of cystic fibrosis (CF) patients. We have previously shown that P. pnomenusa strains possess QS activity. In this study, we examined the presence of QS activity in all type strains of Pandoraea species and acquired their complete genome sequences for holistic bioinformatics analyses of QS-related genes. Only four out of nine type strains ( P. pnomenusa , P. sputorum , P. oxalativorans, and P. vervacti ) showed QS activity, and C8-HSL was the only AHL detected. A total of 10 canonical luxI s with adjacent luxR s were predicted by bioinformatics from the complete genomes of aforementioned species and publicly available Pandoraea genomes. No orphan luxI was identified in any of the genomes. However, genes for two LuxR solos (LuxR2 and LuxR3 solos) were identified in all Pandoraea genomes (except two draft genomes with one LuxR solo gene), and P. thiooxydans was the only species that harbored no QS-related activity and genes. Except the canonical LuxR genes, LuxIs and LuxR solos of Pandoraea species were distantly related to the other well-characterized QS genes based on phylogenetic clustering. LuxR2 and LuxR3 solos might represent two novel evolutionary branches of LuxR system as they were found exclusively only in the genus. As a few luxR solos were located in close proximity with prophage sequence regions in the genomes, we thus postulated that these luxR solos could be transmitted into genus Pandoraea by transduction process mediated by bacteriophage. The bioinformatics approach developed in this study forms the basis for further characterization of closely related species. Overall, our findings improve the current understanding of QS in Pandoraea species, which is a potential pharmacological target in battling Pandoraea infections in CF patients.",
"conclusion": "Conclusions Multiple species of the genus Pandoraea were frequently isolated from sputum samples of CF patients from all over the world, and Pandoraea species are identified as emerging pulmonary pathogen associated with CF. While some species were obtained from the environments, clinically isolated species such as P. pnomenusa has also been recovered from soils in the environment. This suggests the ubiquitous nature of this group of bacteria, and they are thus identified as opportunistic pathogens. The recent report on the QS activity in P. pnomenusa rapidly caught the attention of the scientific community as QS systems have been linked to the regulation of virulence factors, antibiotic resistance, and various traits that are dangerous to patients. Although this study revealed that only four type strains of nine species of genus Pandoraea possess AHL-based QS activity, we also reported the presence of two highly conserved luxR solos in most of their genomes. Our analyses had revealed that these LuxR solos belonged to different clusters of novel evolutionary branches in QS systems. We hypothesize that these LuxR solos in Pandoraea could potentially be responsive to AHLs or different signals produced by neighboring species and coordinate regulation of gene expression, thus playing important roles in the infection process and persistence of these pathogens in cystic fibrosis patients. In the process, we developed an in silico systematic bioinformatics prediction workflow, which is useful for LuxI and LuxR genes identification of other species. To summarize, this study lays the foundation for future study on QS systems of Pandoraea as a potential antimicrobial target in the treatment of Pandoraea infections.",
"introduction": "Introduction Cystic fibrosis (CF) results from mutations in the cystic fibrosis transmembrane conductance regulator (CFTR) gene that functions in modulating chloride ion transport across epithelial cells ( Trapnell et al., 1991 ; Pier et al., 1996 ). As a consequence to this gene abnormality, majority of CF patients suffer from secretion of thick and viscous mucus in their respiratory tracts. These copious respiratory secretions become the breeding ground for microorganisms, which lead to both chronic and transient pulmonary infections, inflammation, obstruction of airways, and ultimately life-threatening pulmonary dysfunction ( Gibson et al., 2003 ). Staphylococcus aureus , Pseudomonas aeruginosa , Burkholderia cepacia, and a spectrum of other Gram-negative bacteria are frequently associated with bacterial lung infections in CF patients ( Gibson et al., 2003 ; LiPuma, 2010 ). However, recent reports revealed unprecedented infections by a number of bacteria and Pandoraea species are among the novel bacteria associated with pulmonary infections in CF patients ( Atkinson et al., 2006 ; Davies and Rubin, 2007 ). The genus Pandoraea was proposed to accommodate a group of isolates cultured from sputum specimens of CF patients that were initially misidentified as B. cepacia and genus Ralstonia . In the process of taxonomical characterization, some members of genus Burkholderia are reclassified into Pandoraea based on genotypic characteristics as well ( Coenye et al., 2000 , 2001 ). Members of genus Pandoraea are commonly recovered from sputum specimens of patients with cystic fibrosis, but some species were isolated from various environmental sources too. These bacteria have been considered as emerging multi-drug resistant pathogens in the context of cystic fibrosis ( Davies and Rubin, 2007 ), but our understanding about the epidemics of Pandoraea species remains scarce. Bacterial cells are able to interact with one another via production and release of diffusible signaling molecules into their living environment. Detection of such molecules enables bacteria to coordinate gene expression in response to both high and low cell population densities. The process is termed as quorum sensing (QS) or bacterial cell-to-cell communication ( Williams, 2007 ). The canonical LuxI/R QS system is one of the most studied QS systems in bacteria. In this system, the responsible signaling molecules are N -acyl homoserine lactones (AHLs) that are produced by an AHL synthase, LuxI activates a cognate transcriptional regulator, LuxR if the concentration of AHLs achieves a threshold. Upon activation, LuxR binds to the promoters or regulators of targeted genes in response to the cell density and causes coordinated gene expression in the bacterial population ( Fuqua and Greenberg, 2002 ; Williams, 2007 ). Via this system, bacteria regulate a variety of activities including biofilm formation, production of extracellular enzymes, regulation of virulence genes, and so on. With the advancement in DNA sequencing technologies, novel subgroups of luxI and luxR homologs have been identified in numerous bacterial species. While it led to reports that most typical luxI / R QS systems have both genes involved located almost adjacent to each other, additional luxR homologs that do not pair with a cognate luxI are frequently found. These unpaired luxR homologs that are termed as luxR solos possess modular homologies to the canonical LuxR with an AHL-binding domain at their N-terminus and a DNA-binding helix-turn-helix (HTH) domain at the C-terminus ( Subramoni and Venturi, 2009 ). Bioinformatics prediction of QS genes in proteobacterial genomes had revealed the presence of numerous additional orphan luxR homologs with no luxI homologs in close proximity ( Fuqua, 2006 ). In addition to their widespread distributions in proteobacteria, some of these LuxR solos are phylogenetically related and several surveys provided evidence on clustering of LuxR solos into different functionally relevant groups ( Brameyer et al., 2014 ; Gan et al., 2015 ; Subramoni et al., 2015 ). It is believed that the presence of additional LuxR solos increases the range of gene regulatory activities by responding to self-produced AHLs or eavesdropping on exogenous AHLs and even other signaling molecules produced by other species ( Hudaiberdiev et al., 2015 ; Subramoni et al., 2015 ). Interestingly, some LuxR solos harbored by non-QS bacteria are responding to non-AHL signaling molecules such as OryR of Xanthomonas oryzae pv. oryzae and XccR of Xanthomonas campestris pv. campestris that are capable of interacting with plant signaling molecules and play essential roles in their pathogenicity ( Zhang et al., 2007 ; Ferluga and Venturi, 2009 ). Members of genus Pandoraea have been reported with QS activity and are able to communicate via the production of AHLs ( Ee et al., 2014 ). In this study, we investigated (1) if AHL-mediated QS is a common activity employed by all type strains of Pandoraea species and (2) the distribution of QS genes in Pandoraea genus. Our work was initiated by obtaining type strains of all species of the genus from culture collection to characterize their QS activity before we sequenced their complete genomes to provide molecular data on distributions and phylogenetic relationships of QS genes in the study species. A systematic bioinformatics prediction workflow was developed for the identification of LuxI and LuxR of Pandoraea species. Our findings indicated that AHL synthases of genus Pandoraea represent a novel evolutionary branch of QS system. We also identified the presence of two conserved LuxR solos in most members of Pandoraea genus, which prompted us to further discuss the acquisition mechanisms and possible roles of these LuxR solos in this study.",
"discussion": "Results and Discussion Detection and Characterization of Quorum Sensing Activity in Pandoraea Genus Following our previous discovery on QS activity in P. pnomenusa RB-38 and RB-44 ( Han-Jen et al., 2013 ; Ee et al., 2014 ), we hypothesized that QS could be a common activity employed by all members in Pandoraea genus. To prove our hypothesis, all nine type strains of Pandoraea were first screened for their QS activity using CVO26 biosensor ( McClean et al., 1997 ), and other possible AHLs were also extracted from spent culture prior to characterization of AHL signaling molecules using MRM MS analysis. Of nine strains, only four ( P. pnomenusa, P. sputorum, P. oxalativorans, and P. vervacti ) activated the CVO26 biosensor, and C8-HSL was the only AHL detected in MRM MS analysis ( Supplementary Figure S1 ). Interestingly, not all clinically isolated strains were detected positive for AHL production as P. apista DSM 16535 T and P. pulmonicola DSM 16583 T have no QS activity detected. Complete Genome Sequencing of Nine Type Strains of Pandoraea Species As QS is not a common activity employed by all Pandoraea species, we questioned (1) if other non-AHL-producing Pandoraea species ( P. apista, P. pulmonicola, P. norimbergensis, P. faecigallinarum, and P. thiooxydans ) are actually possessing mutated luxI and/or luxR , incapable of producing or detecting AHL ( Sandoz et al., 2007 ) and (2) if they harbor luxR solo in their genomes? To prove these hypotheses, we sequenced the complete genomes of all nine type strains of Pandoraea using SMRT sequencing technology to facilitate the identification of the QS genes in the genomes. HGAP assembler was employed to assemble all genomes to complete closure, and circularization was performed to provide a high confidence genomic size of each Pandoraea strains. Plasmids were distinguished from the chromosomal DNA, and designation code was provided for each strain ( Table 1 ). Besides, all Pandoraea genomes available in GenBank were also retrieved for investigation ( Supplementary Table S3 ). Table 1 Designation code, sequencing information and general features of nine circularized genomes of Pandoraea type species. Strains Code Assembler (coverage) Chromosome Plasmid Accession no. Genome size, bp G + C content Gene (protein) Accession no. Genome size, bp G + C content Protein (gene) P. pnomenusa DSM 16536 T Ppn DSM HGAP3 (244.6×) CP009553.2 5,389,285 64.90 4,811 (4,586) – – – – P. pulmonicola DSM 16583 T Ppu DSM HGAP3 (110.5×) CP010310.1 5,867,621 64.30 5,022 (4,820) – – – – P. sputorum DSM 21091 T Psp DSM HGAP3 (215.16×) CP010431.1 5,751,958 62.80 5,037 (4,859) – – – – P. apista DSM 16535 T Pap DSM HGAP3 (243.8×) CP013481.1 5,507,928 62.63 5,042 (4,864) CP013482.1 77,293 62.63 5,042 (4,864) P. norimbergensis DSM 11628 T Pno DSM HGAP3 (165.0×) CP013480.1 6,167,399 63.10 5,417 (5,195) – – – – P. faecigallinarum DSM 23572 T Pfa DSM HGAP3 (75×) CP011807.1 5,261,138 63.70 4,619 (4,412) CP011808.1 CP011809.1 402,292 124,395 61.00 59.30 388 (334) 130 (103) P. oxalativorans DSM 23570 T Pox DSM HGAP3 (215.2×) CP011253.2 5,639,839 63.10 4,983 (4,711) CP011518.1 CP011519.1 CP011520.1 CP011521.1 640,227 135,985 85,789 46,278 63.80 60.60 59.80 59.20 516 (452) 136 (120) 94 (70) 52 (40) P. thiooxydans DSM 25325 T Pth DSM HGAP3 (165.54×) CP011568.1 4,464,186 63.20 4,117 (3984) – – – – P. vervacti DSM 23571 T Pve DSM HGAP2 (185×) CP010897.1 5,656,222 63.50 4,912 (4697) CP010898.1 105,231 62.00 102 (95) Genome sizes of species in Pandoraea genus range between 4.5 and 6.2 Mb with genomes of P. thiooxydans DSM 25325 T and P. norimbergensis DSM 11628 T representing the smallest and largest, respectively ( Table 1 ). The G + C content of these genomes varies from 62.63 to 64.9%. The presence of plasmids was identified in four of nine type strains. Notably, P. apista DSM 16535 T is the only one harboring plasmid of five clinically isolated strains. The other three type strains harboring plasmid are P. faecigallinarum DSM 23572 T , P. oxalativorans DSM 23570 T , and P. vervacti DSM 23571 T that were isolated from oxalate-enriched cultures from different environments ( Table 1 ; Sahin et al., 2011 ). ANI analysis was subsequently performed to investigate the genetic and evolutionary distances of all Pandoraea species ( Supplementary Table S4 ). Generally, 95% of ANI value is the accepted cut-off threshold for species-species delineation ( Richter and Rosselló-Móra, 2009 ). The Pandoraea genomes in this study formed several clusters on a phylogenomic tree constructed using neighbor-joining algorithm ( Figure 2 ). In Cluster 1 that included several P. pnomenusa and P. pulmonicola DSM 16583 T , we observed that the clinically isolated strains P. pnomenusa DSM 16536 T were distinguished from the other P. pnomenusa that were obtained from the environments by slightly further distances. Figure 2 Phylogenomic analysis depicting the genetic and evolutionary distances of all Pandoraea species. In general, Pandoraea genus was separated into five distinct clusters: Cluster 1 ( P. pnomenusa , P. pulmonicola ), Cluster 2 ( P. sputorum , P. oxalativorans , P. faecigallinarum , and P. vervacti ), Cluster 3 ( P. apista ), Cluster 4 ( P. norimbergensis ), and Cluster 5 ( P. thiooxydans ). Bar, 0.2 substitutions per nucleotide position. By contrast, P. sputorum DSM 21091 T was the only clinically isolated species that was clustered together with P. oxalativorans DSM 23570 T , P. faecigallinarum DSM 23572 T , and P. vervacti DSM 23571 T that were isolated from various environments despite being obtained from different origins. The strains P. norimbergensis DSM 11628 T and P. thiooxydans were found to be most distantly related to all other Pandoraea species (<85 and <79% ANI values, respectively; Supplementary Table S4 ) forming outgroups in the phylogenomic analysis of Pandoraea genus ( Figure 2 ). Results from the ANI analysis also suggested reclassification of several Pandoraea strains with uncertain taxonomic status. We deduce from the analysis that Pandoraea sp. E26 could be reclassified as P. pnomenusa E26 as it shares 99% ANI value with all P. pnomenusa . On the other hand, P. pnomenusa strains 6,399 and 7,641 that demonstrated ANI value <90% against all other P. pnomenusa suggested that reclassification might be necessary. This is supported by the observations that both the strains exhibited the highest ANI values with P. apista species (>88% ANI values; Supplementary Table S4 ) and were placed out of Cluster 1 (consisted of all other P. pnomenusa and P. pulmonicola DSM 16583 T ) in phylogenomic tree ( Figure 2 ). Unfortunately, the taxonomy of Pandoraea strains SD6-2 and B-6 remained questionable as both of them were having low ANI values (<88 and <86%, respectively; Supplementary Table S4 ) with all other Pandoraea in this study. Identification of luxI and luxR Homologs in Genomes of Genus Pandoraea \n To provide high confidence in authenticity of all LuxI and LuxR identified in this study, we created a stringent and effective systematic bioinformatics prediction of LuxI and LuxR as presented in Figure 1 . A total of 10 luxI s were identified in genomes of P. pnomenusa, P. sputorum, P. oxalativorans, and P. vervacti ( Supplementary Table S5 ). A typical authentic LuxI contains three signature InterPro domains (IPR016181, IPR001690, and IPR018311) and 10 signature conserved residues. Although all 10 LuxIs of Pandoraea species identified were found to contain only domains IPR016181 and IPR001690 ( Figure 3 ), our previous gene cloning data of PpnI RB38 confirmed that LuxI of Pandoraea species could function properly despite the absence of domain IPR018311 ( Lim et al., 2015a ). Multiple alignment analysis of LuxI also revealed a consistent profile of signature conserved residues in all LuxIs of Pandoraea species, thus concordantly supported the evidence that these are authentic functional LuxI for the production of C8-HSL ( Supplementary Table S6 ). No orphan luxI was identified in any of the Pandoraea genomes in this study. Figure 3 Signature InterPro domain of LuxI (top) and LuxR (bottom). LuxI of Pandoraea species contain two signature domains (IPR016181 and IPR001690), while LuxR of Pandoraea species contains four signature domains (IPR005143, IPR011991, IPR016032, and IPR000792). Intriguingly, besides the expected canonical luxR , two additional luxR solos (named luxR 2 solo and luxR 3 solo) were identified in most Pandoraea genomes ( Supplementary Table S5 ). P. thiooxydans DSM 25325 T is the only exception and does not harbor any canonical luxI / R1 and luxR solo in its genome ( Supplementary Table S5 ). P. apista TF81 and Pandoraea sp. E26 were also found to harbor only luxR2 solo. However, we hypothesized that luxR 3 solo could be missing in the gap of their draft genomes. All canonical LuxR and LuxR solos identified in this study contained all the four signature InterPro domains (IPR005143, IPR011991, IPR016032, and IPR000792), and all nine signature conserved residues (six key amino acids in autoinducer-binding domain and three key amino acids in DNA-binding domain) found in typical LuxR ( Supplementary Table S7 ; Subramoni et al., 2015 ). To date, there have been reports on LuxR solos responding to non-AHL signals. For examples, the PluR of Photorhabdus luminescens senses α-pyrone ( Brachmann et al., 2013 ), while PauR of P. asymbiotica detects dialkylresorcinols and cyclohexanediones ( Brameyer et al., 2015 ) signaling molecules instead of AHLs. These non-AHL-binding LuxRs, however, harbor substitutions in the conserved amino acid motif of autoinducer-binding domain compared to that in AHL sensors ( Brameyer and Heermann, 2015 ). The autoinducer-binding domain of all LuxR solos in Pandoraea species contained the six conserved amino acids (W57, Y61, D70, P71, W85, and G113) with respect to TraR ( Supplementary Table S7 ) and thus reflected a conserved motif for AHL-binding LuxR proteins. From our analysis, we noticed that majority of the annotation pipelines often annotated luxR 2 and luxR 3 solo genes as hypothetical proteins making it a challenge in their identification process. Hence, we employed an in silico systematic bioinformatics prediction of these genes to aid in future identification of luxR 2 and luxR 3 solos. For nomenclature purpose, the gene products of canonical luxI / R identified in Pandoraea genomes were given designation with the first alphabet of the genus followed by the first two alphabet of species, for instances, PpnI, LuxI of P. pnomenusa, and PspR, LuxR of P. sputorum ( Supplementary Table S5 ). Additionally, to differentiate between canonical LuxR and LuxR solos, canonical LuxR were given designation as LuxR1 (e.g., PpnI/R1 and PspI/R1), while LuxR solos were given designation as LuxR2 and LuxR3 solos. Gene designations, accession numbers, amino acid length, GC content, and genetic orientation of all canonical LuxI/R1 and LuxR solos of Pandoraea species are presented in Supplementary Table S5 . No QS gene was found on plasmid. LuxI of Pandoraea Species Represents a Novel Evolutionary Branch of Quorum Sensing System To determine the relatedness of LuxI in Pandoraea genus with the other well-characterized LuxI, phylogenetic and pairwise identity matric analyses were conducted. Figure 4 shows the recent phylogenetic tree of LuxI from several groups that are most closely related to the LuxI in Pandoraea genus, together with pairwise identity matrix analysis. The analyses revealed that LuxI of Pandoraea species were highly conserved in Pandoraea genus forming a distinct cluster separated from the LuxI of Burkholderia species, SolI of Ralstonia solanacearum , and RhlI and LasI of Pseudomonas aeruginosa , representing a novel evolutionary branch of QS system ( Figure 4 ). Ralstonia and Burkholderia are closely related genera to Pandoraea , and they shared highly similar phenotypic profiles that often resulted in the misidentification of Pandoraea species ( Coenye et al., 2001 ; Henry et al., 2001 ). As Pandoraea were also predominantly recovered from CF patients, QS genes of P. aeruginosa (model organism for QS and CF patients) were also included in the analysis ( Barr et al., 2015 ). Figure 4 LuxI phylogenetic tree and pairwise identity matrix analyses of Pandoraea species and closely related species. LuxI of Pandoraea species form a distinct cluster against LuxI of Burkholderia species and P. aeruginosa representing an evolutionary distinct branch of QS system. Bootstrap values (expressed as percentages of 1,000 replications) greater than 50%. Bar, 0.2 substitutions per amino acid position. While analysis on the amino acid pairwise identity revealed that similarity of PpnI of different P. pnomenusa strains can be as high as 90%, pairwise identity shared between the LuxI of different Pandoraea species varied from about 50–90%, even though they were all placed in the same cluster in phylogenetic tree of LuxI ( Figure 4 ). When compared to the LuxI of other genera, the LuxI of Pandoraea species are only 41–55% in pairwise identity with the well-characterized CepI of Burkholderia cenocepacia that catalyzes primarily the synthesis of C8-HSL and a minority of C6-HSL ( Lewenza et al., 1999 ; Lewenza and Sokol, 2001 ); 48–55% in pairwise identity with SolI, which catalyzes the synthesis of C6-HSL and C10-HSL ( Flavier et al., 1997 ); 26–41% in pairwise identity with CciI, which catalyzes primarily the synthesis of C6-HSL and a minority of C8-HSL ( Malott et al., 2005 ); and lastly, about 26–33% pairwise identity with RhlI and LasI, which catalyzes primarily the synthesis of C4-HSL and 3-oxo-C12-HSL ( Latifi et al., 1996 ). Besides, we also identified the 20 bp lux box located in the upstream region of the luxI of Pandoraea species ( Supplementary Figure S2 ). The lux box is a 20 bp palindromic sequence located upstream in the promoter region of luxI, which is required for the binding of AHL-activated LuxR ( Lewenza et al., 1999 ). All luxI of Pandoraea species shared consensus in 14 of 20 lux box sequence ( Supplementary Figure S2 ). Canonical LuxR1 and LuxR Solos in Pandoraea Genus Phylogenetic and pairwise identity matric analyses performed on all identified canonical LuxR1 of genus Pandoraea demonstrated close clustering with CepR from genus Burkholderia and SolR from genus Ralstonia with 96% bootstrap value ( Figure 5 ). Similar to CepI of genus Burkholderia , the LuxI of Pandoraea produce C8-HSL, but SolI from Ralstonia produces two short-chain AHL signals (C6-HSL and C10-HSL). As all identified luxR 1 of Pandoraea are located adjacent to luxI , it is believed that the primary function of LuxR1 is for the detection of C8-HSL produced by its canonical LuxI. A comparison on the amino acid sequences of multiple LuxR groups revealed the highest pairwise identity among canonical LuxR1 of Pandoraea (71–100%) but lower identity to all the LuxR of other groups (<41%), including the LuxR2 and LuxR3 solos in Pandoraea . Figure 5 LuxR phylogenetic tree and pairwise identity matrix analyses of Pandoraea species and closely related species. Canonical LuxR1 clustered closely with CepR and SolR. LuxR2 and LuxR3 solos formed a distinct cluster with CciR and LasR as the outgroup, respectively, representing evolutionary distinct branches of LuxR. Bootstrap values (expressed as percentages of 1,000 replications) greater than 50%. Bar, 0.2 substitutions per amino acid position. Intriguingly, LuxR2 and LuxR3 solos formed two separated clusters on phylogenetic tree with the canonical LuxR identified in Pandoraea genus. Both the LuxR solos were distinctive from each other and had CciR of Burkholderia cenocepacia and LasR of P. aeruginosa as outgroups of the clusters, respectively ( Figure 5 ). The LuxR2 solos are highly conserved in the genus Pandoraea showing >85% in amino acid pairwise identity among different species, as compared to the canonical LuxR1 (>71% in pairwise identity) and LuxR3 solos (>56% in pairwise identity) ( Figure 5 ). From the phylogenetic analysis, it might imply that these LuxR solos in Pandoraea represent two novel evolutionary branches of LuxR in QS system. This is supported by a comprehensive search in various databases, which did not return significant matches with any other species and thus indicated that LuxR2 and LuxR3 solos were found exclusively only in Pandoraea species. The widespread distribution of LuxR solos in almost every Pandoraea species (except P. thiooxydans DSM 25325 T ) indicated that they could be playing potential roles in survival and persistence of these species. For Pandoraea species that possess QS activity ( P. pnomenusa, P. sputorum, P. oxalativorans, and P. vervacti ), additional LuxR solos could function in detecting endogenous AHL signals produced by the AHL synthase to increase the regulatory targets of the complete canonical LuxI/R QS system. Notably, QS positive P. pnomenusa and P. sputorum, which are clinically isolated, might possess LuxR solos for their survival and persistence in respiratory tracts of CF patients as well as regulation of virulence factors. Similar phenomenon was observed in QscR solo of Pseudomonas aeruginosa, which is a LuxR solo that responds to endogenous 3-oxo-C12-HSL produced by LasI to control the timing of AHL production in the species for regulating expression of virulence factors. A study on qscR mutant demonstrated that it is hypervirulent in killing its host indicating that QscR solo is important for efficient regulation of QS-mediated virulence factors ( Chugani et al., 2001 ). In addition, the LuxR solos in Pandoraea species could be essential for detecting exogenous AHLs produced by neighboring species, especially for Pandoraea species that do not own a LuxI/R AHL system. In this study, AHL production was not observed in P. apista DSM 16535 T and P. pulmonicola DSM 21091 T that were isolated from sputa of CF patients. The presence of LuxR solos in these strains could be responsible for eavesdropping by detecting exogenous AHL molecules produced by P. aeruginosa that is chronically colonizing the respiratory tracts of CF patients. It is also noteworthy that many Gram-negative bacteria with QS activity such as Burkholderia and Ralstonia are common pathogens causing lung infections in CF patients. In fact, there are bacteria that possess LuxR solos even though they do not harbor any type of AHL synthase such as Escherichia coli and Salmonella enterica serovar Typhimurium. These bacteria carry a LuxR homolog, and SdiA was reported able to detect and respond to AHL signaling molecules produced by other bacterial species to activate their gene expression ( Ahmer, 2004 ). Comparative Gene Mapping of All Quorum Sensing Genes and Putative Acquisition Mechanism of LuxR Solos in Type Strains of Pandoraea \n Since this is the first documentation of luxR 2 and luxR 3 solos in Pandoraea genus, we are determined to investigate the acquisition mechanism of these genes in Pandoraea genus. Hence, we performed comparative gene mapping to study the degree of conservation of all QS genes. All QS genes of Pandoraea were found to be highly conserved at syntenic genomic locations ( Figure 6 ): all canonical luxI / R 1 were found to be convergently inverted (only luxI / R 1 of P. sputorum and P. norimbergensis overlapped with each other) and located upstream of an alcohol dehydrogenase and an ABC transporter ATP-binding protein ( Figure 6A ); luxR 2 solos were consistently located between a LysR transcriptional regulators and a RND transporter ( Figure 6B ); and luxR 3 solos were always found located downstream of cytochrome c oxidase subunits I and II and a membrane protein ( Figure 6C ). As there are hypothetical proteins located in the immediate upstream of luxR 2 and luxR 3, we questioned if these hypothetical proteins could be the canonical luxI that have mutated and lost its function or domain. However, after a comprehensive domain prediction was performed on these hypothetical proteins, there was no residue of luxI in these hypothetical proteins. Figure 6 Comparative gene mapping of all QS genes in type strains of Pandoraea species. All QS genes were highly conserved at syntenic genomic location. (A) Canonical luxI and luxR 1 were found be convergently inverted and located upstream of alcohol dehydrogenase and ABC transporter ATP-binding protein. (B) \n luxR 2 solos located between LysR transcriptional regulators and RND transporter. (C) \n luxR 3 solo located downstream of cytochrome c oxidase subunits I and II and a membrane protein. Subsequently, we also performed an extensive search for the presence of any QS genes in genomic island, prophages, and mobile genetic element regions to determine the possibility of horizontal gene transfer event. No QS gene was found on any genomic island, and no residue of transposase was found in close proximity of all QS genes. Although no QS gene was found within any intact prophage region, there are, however, few luxR solos that were found in close proximity with incomplete and intact prophage sequences, such as 53,482 bp between PpuR2 (63.0% GC content; 544,114–544,881 bp) with an incomplete prophage region 1 (63.7% GC content; 597,363–605,870 bp); 62,330 bp between PoxR2 (60.0% GC content; 478,160–478,927 bp) with an incomplete prophage region 1 (62.91% GC content; 533,510–541,257 bp); and 6,231 bp between PoxR3 (65.2% GC content; 2,404,930–2,405,844 bp) with an intact prophage region 8 (63.2% GC content; 2,391,675–2,398,699 bp). These observations suggested that luxR 2 and luxR 3 solos could be transmitted into Pandoraea genus by transduction event mediated by prophage. However, parts of these prophage sequences might be lost during evolution. Various QS-related genes had been reported in the genomes of bacteriophages including homologs of accessory gene regulator ( agr ) in the genome of Clostridium difficile phage phiCDHM1 ( Hargreaves et al., 2014 ) and regulatory protein LuxR in the Azospirillum brasilense Cd bacteriophage’s genome ( Boyer et al., 2008 ). QS activity in Pandoraea species has been related to the regulation of virulence factors, biofilm formation, extracellular enzymes production, antibiotic resistance, and various other lethal traits. Although not all Pandoraea species exhibit QS activity, findings in this study revealed that almost every Pandoraea species (except P. thiooxydans DSM 25325 T ) possess LuxR solos genes in their genomes. The repertoire of LuxR solos in the genus increases the range of gene regulatory activities and is anticipated to play roles in QS-dependent regulation of phenotypic functions, which should be investigated further. The data presented are also useful in future application including quorum quenching (QQ) study that attempts to disrupt the bacterial cell-to-cell communication of Pandoraea species through QS ( See-Too et al., 2018 ). QQ has been suggested as alternative antibacterial strategy to antibiotics, which might lead to emergence of multi-drug resistant bacteria ( Tang and Zhang, 2014 ). Last but not least, we hope that findings from this study contribute to further research to elucidate the downstream roles of QS genes in Pandoraea species, including their LuxR solos."
} | 8,417 |
39209831 | PMC11362324 | pmc | 2,985 | {
"abstract": "Carbon capture and utilization (CCU) covers an array of technologies for valorizing carbon dioxide (CO 2 ). To date, most mature CCU technology conducted with capture agents operates against the CO 2 gradient to desorb CO 2 from capture agents, exhibiting high energy penalties and thermal degradation due to the requirement for thermal swings. This Perspective presents a concept of Bio-Integrated Carbon Capture and Utilization (BICCU), which utilizes methanogens for integrated release and conversion of CO 2 captured with capture agents. BICCU hereby substitutes the energy-intensive desorption with microbial conversion of captured CO 2 by the methanogenic CO 2 -reduction pathway, utilizing green hydrogen to generate non-fossil methane.",
"introduction": "Introduction The global emissions of anthropogenic greenhouse gases have risen immensely during the past decades, with energy consumption from fossil fuel combustion being the main driver for around two-thirds of the total greenhouse gas emissions 1 . In 2022, the annual global greenhouse gas emissions from energy-related fossil fuel combustion and industrial processes reached 41.3 Gt CO2-eq , where carbon dioxide (CO 2 ) emissions accounted for 89% of the total emissions (36.8 Gt CO2 ) 2 . However, the United Nations Climate Change Conference (COP28) declared the ‘beginning of the end’ of the fossil fuel era by the end of 2023 3 . The adoption of a renewable-based energy infrastructure has accelerated in the last decade with renewable electricity generation constituting 7858 TWh in 2021 4 . The transformation of the energy system towards heavy electrification has been envisioned as a significant factor in achieving carbon neutrality by 2050 5 . However, the hard-to-abate sector, including heavy transport and industrial processes, relies heavily on chemical energy carriers with a high volumetric energy density and remains difficult to electrify and decarbonize 6 . Here, Power to X (PtX) provides a potential solution that enables the conversion of electricity to chemical energy carriers, and combining it with Carbon Capture and Utilization (CCU) represents a critical technology to supply carbon-based energy carriers (e-fuels) for reaching the 2050 net zero emission target. Various post-combustion carbon capture technologies are currently available, with chemical scrubbing being the most mature technology relying on solvent-based absorption and desorption conducted on the scale of megaton CO 2 captured annually 7 , 8 . Conventional CCU with amine scrubbing consists of multiple steps, including CO 2 capture by solvent absorption, CO 2 release by heat desorption, CO 2 dehydration and compression for storage and transport, and long-term storage or utilization in various downstream chemical and biological processes. Despite the many point sources of flue gas and a high potential for capturing CO 2 , conventional carbon capture with chemical scrubbing is attenuated by high energy penalties due to the dilute CO 2 concentrations in flue gas, constituting up to 30% of the typical power plant output 9 . In conventional CO 2 desorption from capture agents, the energy equivalent to the heat of absorption is added in a reboiler unit at the desorption column to induce the desorption and release of CO 2 . The typical thermal reboiler duty for the CO 2 desorption column is 3.5–4.0 GJ t −1 CO2 10 – 12 . Substantial research is therefore directed toward reducing the reboiler duty by modifying the capture agents and using blends that comprise capture agents with additives. Activators enhance the absorption and desorption kinetics and chemicals minimize the oxidative and thermal degradative behaviors 13 . The reboiler duty can furthermore be reduced by optimizing the desorption process parameters, such as adjusting operating pressures in the desorption unit and regulating reboiler temperatures 14 . Hence, multiple pilot-scale studies have demonstrated routes for reducing the reboiler duty of the desorption unit to a range of 1.9–3.6 GJ t −1 CO2 15 – 17 . In the future fossil-free society based on a circular economy, the captured CO 2 is considered a valuable commodity that must be utilized to produce chemicals and fuels, which currently are fossil-based. However, carbon utilization preceded by carbon capture is limited by additional expenses for storing and transporting CO 2 , which requires dehydration to avoid corrosion and compression (~0.4 GJ t −1 CO2 ) 18 . Thus, a group of new concepts based on integrated carbon capture and utilization (ICCU) is currently being developed, where captured CO 2 is directly synthesized into valuable compounds simultaneously with the desorption from the capture agents. Hereby, the energy expenses for CO 2 purification by desorption, transportation, dehydration, and compression are eliminated, improving the competitiveness of the capture process. The current ICCU processes include thermocatalysis 19 , electrochemical catalysis 20 , photoelectrocatalysis 21 , and biofixation with microalgae and cyanobacteria 22 . Nevertheless, several drawbacks have challenged the ICCU technologies. Thermocatalytic ICCU processes entail high temperatures and pressure to convert the CO 2 . Although isothermal operation can mitigate the energy penalty from the desorption, substantial efforts are required to identify a potentially stable and efficient CO 2 sorbent capable of synergistically matching the reaction with the catalyst 23 . Furthermore, electrochemical catalysis offers the prospects of being conducted at milder temperatures, but this approach necessitates the utilization of expensive catalysts due to the chemical inertness of CO 2 24 . Accordingly, additional gas conditioning of the raw flue gas will be essential to avoid poisoning the catalysts. In contrast to chemical catalysts, biocatalysts such as methanogens can handle various contaminants such as hydrogen sulfide (H 2 S) and sulfur dioxide (SO 2 ) 25 , which otherwise would deactivate the chemical catalysts used in both ICCU and CCU processes 26 . Methanogens belong to the domain Archaea and can be divided into three different physiological categories based on their substrate: hydrogenotrophic methanogens, methylotrophic methanogens, and acetoclastic methanogens 27 . The former is a key component in biological methanation, a robust Power to X (PtX) technology, where CO 2 is reduced to methane (CH 4 ) using renewable H 2 from water electrolysis. Species important for the biomethanation process include Methanobacterium, Methanobrevibacter , and Methanoculleus , which all have been reported to be enriched during biogas upgrading, whereas the relative abundance of acetoclastic methanogens (species within the Methanosarcinaceae family) is usually experienced to decrease during biomethanation 28 , 29 . A common denominator for these microorganisms is their anaerobic trait, which renders them inhibited by oxygen (O 2 ). Biomethanation has currently exclusively been utilized with feedstocks of biogas CO 2 25 , 30 , 31 and syngas CO 2 32 , 33 and has not been suited for direct conversion of flue gas CO 2 due to the flue gas composition containing O 2 , which is detrimental to the obligate anaerobic methanogens. Furthermore, the CO 2 is diluted with N 2 , which creates a dilute CH 4 stream of <20% without any potential downstream application. The development of a technology that enables the biological utilization of flue gas CO 2 would thus increase the potential of CO 2 feedstock substantially, as biogenic CO 2 from biogas only constituted 0.024 Gt CO2 by 2020 in Europe 34 , which is a mere fraction (~1%) of the 2.5 Gt CO2 emitted in Europe 35 . Targeting CH 4 creates a versatile energy vector that indirectly electrifies the hard-to-abate sector while utilizing the already established natural gas grid infrastructures in the TWh magnitude 6 . However, the CH 4 must comply with the standards of the distribution and storage grids, which are defined by the specific pipeline networks in the range 70–98% CH 4 in the EU 36 . This Perspective unfolds a concept for ICCU with simultaneous desorption and conversion of CO 2 to CH 4 based entirely on a microbiological driving force that enables biomethanation to be applied to dilute flue gases. Using methanogens for ICCU will ultimately alleviate the energy penalties and thermal degradation of the capture agent related to conventional carbon capture while reducing the number of process units required for CCU. This CO 2 transformation route of bio-integrated carbon capture and utilization (BICCU) is currently demonstrated at a low Technology Readiness Level (TRL), but the concept, challenges, and prospects are presented here."
} | 2,185 |
21166901 | null | s2 | 2,986 | {
"abstract": "Bacterial populations frequently act as a collective by secreting a wide range of compounds necessary for cell-cell communication, host colonization and virulence. How such behaviours avoid exploitation by spontaneous 'cheater' mutants that use but do not contribute to secretions remains unclear. We investigate this question using Pseudomonas aeruginosa swarming, a collective surface motility requiring massive secretions of rhamnolipid biosurfactants. We first show that swarming is immune to the evolution of rhlA(-) 'cheaters'. We then demonstrate that P. aeruginosa resists cheating through metabolic prudence: wild-type cells secrete biosurfactants only when the cost of their production and impact on individual fitness is low, therefore preventing non-secreting strains from gaining an evolutionary advantage. Metabolic prudence works because the carbon-rich biosurfactants are only produced when growth is limited by another growth limiting nutrient, the nitrogen source. By genetically manipulating a strain to produce the biosurfactants constitutively we show that swarming becomes cheatable: a non-producing strain rapidly outcompetes and replaces this obligate cooperator. We argue that metabolic prudence, which may first evolve as a direct response to cheating or simply to optimize growth, can explain the maintenance of massive secretions in many bacteria. More generally, prudent regulation is a mechanism to stabilize cooperation."
} | 362 |
25054325 | PMC4108414 | pmc | 2,987 | {
"abstract": "Habitat complexity strongly affects the structure and dynamics of ecological communities, with increased complexity often leading to greater species diversity and abundance. However, habitat complexity changes as communities develop, and some species alter their environment to themselves provide habitat for other species. Most experimental studies manipulate basal substrate complexity, and while the importance of complexity likely changes during community development, few studies have examined the temporal dynamics of this variable. We used two experiments to quantify the importance of basal substrate complexity to sessile marine invertebrate community development through space and time. First, we compared effects of substrate complexity at 70 sites across ten estuaries. Sites differed in recruitment and community development rates, and after three months provided spatial variation in community development stage. Second, we tested for effects of substrate complexity at multiple times at a single site. In both experiments, complexity affected marine sessile invertebrate community composition in the early stages of community development when resource availability was high. Effects of complexity diminished through time as the amount of available space (the primary limiting resource) declined. Our work suggests the presence of a bare-space threshold, at which structural complexity of the basal substrate is overwhelmed by secondary biotic complexity. This threshold will be met at different times depending on local recruitment and growth rates and is likely to vary with productivity gradients.",
"introduction": "Introduction Habitat complexity, or the physical structure of an environment, influences community composition in a number of ways. Complex habitats can promote species coexistence by providing a wide range of niches, thereby reducing niche overlap and increasing diversity [1] , [2] . Classic work by MacArthur & MacArthur (1961) found a correlation between bird species diversity and foliage height diversity, rather than plant species composition [3] , and similar relationships between species diversity and habitat complexity have since been observed in terrestrial [4] – [6] , freshwater [7] – [9] , and marine systems [10] , [11] . Habitat complexity can also be important in mediating predation, since cryptic habitats provide refuge for smaller organisms that would otherwise be vulnerable [12] – [15] . In many systems habitat complexity varies through space and time. Available structure can change seasonally [6] , in response to disturbance [16] , and as a result of interactions between species and their environment. Individual organisms can both reduce and add to habitat complexity: resource utilization decreases the amount of available substrate, but some species can themselves provide habitat for others. In marine communities, habitat-forming organisms such as barnacles and algae provide substrate for other organisms to settle and grow, and can become the main source of structure once basal substrate becomes rare [17] , [18] . Effects of basal substrate complexity may therefore change over time, as the complexity of the substrate is buffered by habitat complexity provided by resident species. Many studies have observed strong effects of basal substrate complexity on community structure, but few have examined how this changes over the course of community development. Southwood et al (1979) showed that the relative influence of habitat complexity changed over the course of succession in a birch woodland, and structural complexity became more important to species diversity in the later stages of succession [19] . A similar study by Brose (2003) in temporary wetlands suggested that structural complexity was independent of successional stage, and that the quantity of structural complexity determined community richness and diversity [20] . Work in fouling assemblages has suggested a declining importance of complexity effects with time, but has not explicitly compared stages of community development [21] , [22] . In sessile invertebrate communities, population dynamics and community composition are dependent on space availability, as larvae require space to settle and grow [23] . Space is abundant in the early stages of community development, allowing active larval choice of settlement substrate [24] – [28] . Larvae may preferentially recruit to structural features of the environment, which can influence larval survival [29] . Fine-scale structural complexity changes the hydrodynamics, physical cues, and refuge quality of substrates, which may in turn alter larval settlement [25] , [30] . Hydrodynamics may encourage settlement in grooves or crevices, as larvae become trapped in eddies that form on the leeward side of structural features [25] . Larvae may also choose to settle in the comparatively protected substrate of these features as a means of refuge from predators [31] – [33] . In hard-substrate environments, larvae frequently settle preferentially in structural features such as grooves, pits, or crevices [9] , [32] – [34] . However, as the amount of available bare space declines during community development, the importance of basal substrate complexity to sessile invertebrate communities might also diminish and the role of biotic complexity may become more important. Here we used two experiments to investigate how the role of habitat complexity changes throughout community development. We hypothesized that structural complexity would become less predictive of species abundance and diversity over time, as basal substrate is sequestered. Following previous studies, we represented habitat complexity by cutting varying numbers of grooves into the surface of a flat, hard substrate [32] , [34] , [35] . We considered variation in the stage of community development in two ways, through space and time. First, we compared communities across multiple estuaries, in which communities differed naturally in their assembly-rate and species composition. By comparing communities across estuaries at a single point in time, we captured communities at different stages of community development. Second, we observed recruitment density over the course of community development at multiple times. In each of these approaches we examined the relationship between basal substrate complexity and community composition. Together, these experiments offer insights into the generality of the relationships between habitat complexity and diversity in the context of community development for marine sessile invertebrate communities.",
"discussion": "Discussion Basal habitat complexity was important to marine sessile invertebrate community composition, but only in the early stages of community development when resource availability was high. The rate of space sequestration differed between and within estuaries and thus the effects of basal habitat complexity varied through space. Effects were strongest in Wagonga Inlet where recruitment and growth were low and there was the most available space. Diversity indices and species abundances increased with fine-scale complexity in this estuary, but effects were weaker elsewhere. Basal complexity effects also appeared to diminish with time. Complexity effects were present after one month when there was a significant amount of bare space, but not after three months when space availability was low. Both the spatial and temporal studies suggest that bare space, or perhaps the absence of biotic habitat complexity, may mediate the role of basal complexity in community development of sessile marine invertebrate communities. However, these results may also apply more generally to other environments, such as bare soil and freshwater rocks, in which bare substrate is actively colonized [39] , [40] . In our study, the mere presence of basal complexity in space-rich environments increased species diversity and abundance. Increasing complexity generally did not affect community patterns beyond the changes observed at the lowest level of complexity, indicating a threshold beyond which the effect of complexity remains constant. There is clearly a minimum level at which species begin to respond to complexity, which has been documented in fish and invertebrates as function of predation refuge [12] , [41] . Species response to complexity is likely a matter of scale, and Kelaher (2003) suggests that there is an upper threshold at which the addition of more structural components leads to a decline in species diversity and abundance [42] . Our results suggest that our complexity treatments provided a level of complexity intermediate to these two potential thresholds. The role of complexity is likely also dependent on whether species exhibit mobile or sessile life histories [34] . For example, the importance of complexity as a predation refuge may be more important to sessile species in the vulnerable early stages of development, while mobile species may continue to find refuge in habitat complexity over the course of their lifetimes. However, regardless of mobility, the importance of complexity is likely dependent on the scale of complexity relative to species size. Our work suggests the existence of a bare-space threshold at which the presence of structural complexity of the basal substrate becomes irrelevant to community development. Effects of complexity diminished rapidly when the availability of bare space fell below 30 to 50% in both the spatial and temporal experiments, although additional time points may be necessary in future studies to more clearly define the boundaries of a temporal threshold. When bare space is available, settling organisms can take advantage of fine-scale structural features of the basal substrate. However, as bare space becomes increasingly limited over the course of community development, or as a result of other factors such as high larval supply or settlement rate, the subtleties of fine-scale structure may be overwhelmed by biotic structure from recruitment. The presence of primary recruits alters both larval settlement rates and post-settlement mortality by controlling available space [17] , [18] and creating new micro-habitats for subsequent recruits [43] . Further, chemical cues may induce larvae to settle on resident adults rather than other structures [44] , [45] . There may be a certain recruit density at which resident species facilitate future settlement to a greater extent than the presence of basal structural features. In fact, in our spatial complexity experiment, we even observed declines in species richness and abundance at sites with high biotic cover and minimal bare space – a relationship that may depend on the identity of primary recruits. For example, initial settlement of complex basal substrate by a dominant or competitive species could lead to declines in the richness and abundance of future settlers via competitive exclusion. It is difficult to determine the role of species identity in our observed community patterns. In our spatial study, large differences in community composition among estuaries make it hard to isolate the role of species identity in development from complexity effects. However, broadly speaking, in communities with distinct complexity effects (e.g. Wagonga Inlet in the spatial study or after one month in the temporal study), barnacle and bryozoan taxonomic groups were significant drivers of complexity patterns, despite the fact that these species have different life history strategies (i.e. solitary vs. colonial). Colonial organisms are expected to dominate solitary species over the course of community development, and may be superior competitors in space-limited environments as a result of asexual, indeterminate growth and fouling ability [46] . Further work in this system should address the importance of species identity, specifically with regard to coloniality, in mediating fine-scale basal complexity effects. However, the fact that we did not see differences based on species identity in our study strengthens the suggestion that an alternative mechanism, such as bare space availability, may be more important in this system. In our study, early differences in recruitment patterns, as driven by the presence of structural complexity, did not translate to long-term differences in the mature communities. This result has implications for understanding the importance of priority effects in marine benthic communities [47] . Sutherland and Karlson (1977) showed that after the initial developmental period in a sessile invertebrate community, subsequent changes to the community were dependent on the identity of both the resident adults and the newly settling larvae, and thus on the order of recruitment by distinct species. In long-term populations, adult mortality led to the release of approximately 20–60% of bare space annually, and the colonization of this space by new species dramatically changed community composition over time [48] . The fact that our communities became increasingly similar to each other over the course of community development suggests that priority effects become less pertinent with time, barring space-freeing processes such as adult mortality, disturbance, or predation that allow for variable larval settlement. Accordingly, more recent work has shown that interactions between resident adults and new recruits that affect juvenile persistence are strongest within hours of larval settlement [49] . The relationship between complexity effects and bare space availability suggests that the presence of basal habitat structure may be more important in communities that are recruitment limited. Recruitment limitation is the idea that population size and species densities may be limited by larval supply [50] , [51] . “Supply-side” processes such as larval supply, settlement rate, and post-settlement mortality are closely tied to the presence of bare space. Roughgarden, Iwasa, and Baxter (1985) explore the role of bare space and “supply-side” factors in their classic model for the demography and population dynamics of an open population with space-limited recruitment. They suggest that in the presence of a low settlement rate, a steady state is reached where free space is present and the relative spatial abundance of species is determined by variation in settlement and mortality rates [52] . Thus, in recruitment-limited environments, the presence of ample bare space may support variable community responses to available structural features. Previous work in marine and terrestrial systems have shown largely positive effects of complexity on species diversity measures, but few studies have defined the conditions in which this is important. This study suggests that fine-scale habitat complexity increases marine sessile invertebrate diversity measures and species abundance, but only in the early stages of community development. There may be a bare-space threshold at which structural complexity becomes overwhelmed by recruitment and community development loses sensitivity to structural complexity. However, this threshold will be met at different times depending on local recruitment and growth rates and is therefore likely to vary with gradients of productivity."
} | 3,829 |
24255625 | null | s2 | 2,988 | {
"abstract": "We consider a general swarm model of self-propelling agents interacting through a pairwise potential in the presence of noise and communication time delay. Previous work has shown that a communication time delay in the swarm induces a pattern bifurcation that depends on the size of the coupling amplitude. We extend these results by completely unfolding the bifurcation structure of the mean field approximation. Our analysis reveals a direct correspondence between the different dynamical behaviors found in different regions of the coupling-time delay plane with the different classes of simulated coherent swarm patterns. We derive the spatiotemporal scales of the swarm structures, as well as demonstrate how the complicated interplay of coupling strength, time delay, noise intensity, and choice of initial conditions can affect the swarm. In particular, our studies show that for sufficiently large values of the coupling strength and/or the time delay, there is a noise intensity threshold that forces a transition of the swarm from a misaligned state into an aligned state. We show that this alignment transition exhibits hysteresis when the noise intensity is taken to be time dependent."
} | 299 |
34847792 | PMC8633776 | pmc | 2,989 | {
"abstract": "For billions of years, photosynthetic microbes have evolved under the variable exposure to sunlight in diverse ecosystems and microhabitats all over our planet. Their abilities to dynamically respond to alterations of the luminous intensity, including phototaxis, surface association and diurnal cell cycles, are pivotal for their survival. If these strategies fail in the absence of light, the microbes can still sustain essential metabolic functionalities and motility by switching their energy production from photosynthesis to oxygen respiration. For suspensions of motile C. reinhardtii cells above a critical density, we demonstrate that this switch reversibly controls collective microbial aggregation. Aerobic respiration dominates over photosynthesis in conditions of low light, which causes the microbial motility to sensitively depend on the local availability of oxygen. For dense microbial populations in self-generated oxygen gradients, microfluidic experiments and continuum theory based on a reaction–diffusion mechanism show that oxygen-regulated motility enables the collective emergence of highly localized regions of high and low cell densities.",
"introduction": "1 . Introduction Photosynthesis is a fundamental process of life that converts light into chemical energy [ 1 ]. As a metabolic process, it is ubiquitous in highly diverse groups of species, ranging from higher level plants to prokaryotic cyanobacteria [ 2 ] and eukaryotic microalgae [ 3 ]. These unicellular photosynthetic microbes inhabit almost any ecosystem on our planet and are dispersed to heterogeneous microhabitats. Their regulation, acclimation and adaptation to light are key to the survival and propagation in their environment [ 4 ]. Light is perceived by photoreceptors [ 5 ], which control a multitude of essential biological functionalities including circadian rhythm [ 6 – 8 ], sexual reproduction [ 9 ], directed motion along a light gradient (phototaxis) [ 10 – 13 ], and adhesion to surfaces [ 14 ]. The photosynthetic machinery itself is regulated by the light harvesting complex of chlorophyll a/b, representing another pathway to respond to light cues [ 15 ]. In unfavourable light conditions, photosynthetic microbes can still survive and produce energy through aerobic respiration, allowing them to produce ATP by the consumption of oxygen [ 16 ]. This consumption however can result in self-generated dark anoxia [ 17 , 18 ], where the cell is deprived of both light and oxygen. Light directly controls the activities of both metabolic pathways, photosynthesis and aerobic respiration, and thus has profound implications on the life of microbial populations. We show that, by inhibiting photosynthetic activity, a confined suspension of Chlamydomonas reinhardtii cells can form large-scale aggregations. Collective aggregation is controlled via the microbial motility, which sensitively depends on the availability of light and oxygen. The appearance as well as the dynamics of aggregation are governed by the light intensity and wavelength as control parameters, providing a direct link between the activity of the photosynthetic machinery, microbial motility and large-scale self-organization.",
"discussion": "6 . Discussion We unravel the interplay of microbial motility and microbial responses to light conditions and local oxygen concentration fields, which enables isolation of the feedback mechanism for the localized aggregation of an ensemble of motile photoactive microbes: under low light conditions, C. reinhardtii cells switch their energy production metabolism from photosynthesis to aerobic respiration. If the cell density is sufficiently high and oxygen is not replenished externally at a sufficiently high rate, oxygen is locally deprived resulting in a decrease in microbial swimming velocity. As a result of a generic coupling between cell density and swimming velocity, given through the power law ρ ∝ v −2 , microbial aggregation occurs in regions of reduced motility. An inverse density–velocity proportionality has been shown to be able to cause spontaneous aggregation in active matter systems through a positive-feedback mechanism called motility-induced phase separation (MIPS) [ 33 ]. However, the strength of the power-law dependence observed in our system is not sufficient for an aggregation to spontaneously form through a MIPS-like phenomenon, but rather appears to be a generic trait of the aggregation mechanism at work. Even though the deprivation of oxygen decreases the swimming velocity of the motile microbes, there is no effect on their tumbling time, which only depends on the cell density. This phenomenon of light-regulated microbial aggregation is a direct consequence of the ability of planktonic photosynthetic microbes to reversibly switch their energy production from photosynthesis to oxygen respiration. Such aggregates of motile microbes form in the complete absence of external light gradients (phototaxis) [ 11 , 13 , 19 ], nutrient sources (chemotaxis) [ 34 , 35 ], photokinesis [ 20 ], aerotaxis (see electronic supplementary material, figure S8) and quorum sensing and thus represent manifestations of another remarkable collective behaviour of motile living cells."
} | 1,301 |
25360746 | PMC4216011 | pmc | 2,990 | {
"abstract": "The light dependency of respiratory activity of two scleractinian corals was examined using O 2 microsensors and CO 2 exchange measurements. Light respiration increased strongly but asymptotically with elevated irradiance in both species. Light respiration in Pocillopora damicornis was higher than in Pavona decussata under low irradiance, indicating species-specific differences in light-dependent metabolic processes. Overall, the coral P. decussata exhibited higher CO 2 uptake rates than P. damicornis over the experimental irradiance range. P. decussata also harboured twice as many algal symbionts and higher total protein biomass compared to P. damicornis , possibly resulting in self-shading of the symbionts and/or changes in host tissue specific light distribution. Differences in light respiration and CO 2 availability could be due to host-specific characteristics that modulate the symbiont microenvironment, its photosynthesis, and hence the overall performance of the coral holobiont.",
"conclusion": "Conclusions Light-saturated respiration rates (R light O2 micro ) were similar in both corals and multiple times higher than steady-state dark respiration rates (R dark O2 micro ). This is interpreted as the activity of light-driven metabolic pathways that increase with increasing irradiance. The light respiration rates show, that differential CO 2 uptake rates of the two species examined could indicate that carbon availability influences the metabolic processes of the holobiont. Although both coral hosts are known to harbour the same Symbiodinium subclade C1 [42] , it seems that they experience different host-specific microenvironmental conditions (see Figure 3 ).",
"introduction": "Introduction The success of scleractinian corals in oligotrophic tropical waters is based on the endosymbiosis between the coral host and single-celled microalgae, i.e., dinoflagellates in the genus Symbiodinium that reside within the host's endodermal cells. The algal symbionts translocate up to 95% of their photosynthetically fixed carbon (C) to the coral host under optimal conditions [1] , whilst the algal symbionts receive nutrients and shelter from the host [2] , [3] . There is considerable genotypic variation within the Symbiodinium genus [4] that can modulate the stress resilience of the holobiont [5] . The dark reactions of photosynthesis fix CO 2 into organic carbon using the enzyme Ribulose-1,5-bisphosphate-carboxylase/oxygenase (RuBisCO). Symbiodinium contains a prokaryotic-type II RuBisCO, which has a low affinity for CO 2 \n [6] – [9] . High concentrations of CO 2 are therefore necessary to promote carbon assimilation and to meet the hosts' energetic demand for symbiont-derived photosynthates [10] – [12] . Holobiont respiration may present an additional internal CO 2 source contributing to the complex carbon exchange and transfer system within corals. Chlororespiration, involving plastoquinone (PQ) oxidation with O 2 and a terminal oxidase (PTOX) [13] can be active within the chloroplasts of Symbiodinium . Furthermore, calcification occurring in the calicodermis of the coral [14] and host mitochondrial respiration can further contribute to the internal CO 2 supply in the holobiont [15] , [16] . Coral host respiration is just one source of inorganic carbon for symbiont photosynthesis [17] – [19] ; external inorganic carbon sources such as seawater are also utilised. However, the supply of inorganic carbon via passive diffusion from the surrounding seawater and host tissue is restricted by several factors: 1) the generally low CO 2 content of seawater, 2) the presence of a diffusive boundary layer, and 3) the presence of multiple membranes of the host tissue surrounding the endodermal Symbiodinium cells, which need to be traversed. Both, coral host and symbionts employ a range of carbon concentrating mechanisms (CCMs) [20] – [24] to enhance the carbon supply from the external medium and thus increase CO 2 availability to the Symbiodinium chloroplasts [25] as well as for calcification purposes [26] . The rate of photosynthesis by the symbionts and therefore their carbon demand is closely correlated with photon irradiance [27] , and may become carbon limited under high irradiance [28] . As the delivery of carbon to the algal symbionts is controlled by the activity of CCMs (of coral host as well as algal symbionts), as well as host respiration [19] , the host metabolism can thus have a strong impact on symbiont photosynthesis, e.g., by supplying sufficient inorganic carbon under high irradiance. While demands on the host-supplied carbon shift with irradiance, e.g., due to extra demand in light-enhanced calcification [29] , there are only few experimental investigations of such responses in the literature [26] , [30] . We investigated if respiratory-dependent processes in the coral would follow a typical asymptotic rise with increasing irradiance, as it is known for photosynthetic processes. Photosynthesis and calcification require carbon as substrate [31] , [32] ; photosynthesis is directly dependent on light and coral calcification is known to be light-enhanced [33] , [34] . Indeed, there is a close interplay of internal utilization of metabolically derived carbon for both processes. Carbonic anhydrase enzymes catalyse the reaction CO 2 +H 2 O ↔ HCO 3 \n − +H + , and therefore generate substrate for the calcification reaction (CO 2 +H 2 O+Ca ++ ↔ CaCO 3 +2H + ), as well as for photosynthesis: CO 2 +H 2 O ↔ CH 2 O+O 2 \n [35] , [36] . The exchange of respiratory gases (O 2 and CO 2 ) in photosynthetic symbioses is difficult to study in the light because respiratory O 2 uptake is masked by the O 2 production from photosynthesis. At low irradiance, where symbiont photosynthesis is lower than respiratory activity in the coral, i.e., below the irradiance compensation point net O 2 uptake and CO 2 release can be measured [37] . To measure these gas exchange patterns in corals is challenging, as several discrete ‘compartments’ of respiration operate in parallel and in close proximity, and therefore there is a close coupling between autotrophic and heterotrophic processes [38] . Enhanced post-illumination dark respiration (EPIR), which is the respiratory activity measured just after transition from light to darkness, has been used to support assumptions about light-driven respiratory processes in corals [16] , [34] . However, in the absence of light there is no production of reducing agents due to the absence of photosynthetic light reactions, so that EPIR likely underestimates light respiration. To quantify respiration in the light, O 2 microsensors can be used to quantify gross photosynthesis rates (GP O2 micro ) in corals independent of respiration [14] , [39] , [40] . In conjunction with flux calculations of the net photosynthetic rate (Pnet O2 micro ) from measured steady-state O 2 concentration profiles, microsensor measurements allow for the determination of respiration rates in the light [41] . In this study, we present the first direct measurements of light respiration in corals as a function of irradiance. We combine O 2 microsensor measurements with detailed CO 2 exchange measurements to assess the relationship between CO 2 exchange and symbiont gross photosynthesis rates in two scleractinian corals, Pocillopora damicornis (Linnaeus, 1758) and Pavona decussata (Dana, 1846), that are known to harbour the same Symbiodinium subclade (C1) [42] . The light dependency of external carbon uptake and respiratory activity was also examined, to see if respiratory processes followed an asymptotic rise with irradiance similar to photosynthetic processes.",
"discussion": "Discussion This is the first study reporting an integrated approach measuring coral light respiration and gross photosynthesis with O 2 microsensors and CO 2 gas exchange techniques across a range of irradiance. The two main finding of this study are that i) light-saturated (at 210 µmol photons m −2 s −1 ) respiration rates (R light O2 micro ) were multiple times higher than steady-state dark respiration rates (R dark O2 micro ) (11 times for P. decussata and 25 times for P. damicornis , and ii) P. damicornis and P. decussata differ in their photophysiological function despite likely harbouring the same symbiont subclade C1 [42] (see Fig. 3 for a conceptual diagram of the main findings). 10.1371/journal.pone.0110814.g003 Figure 3 Conceptual model of light and carbon availability, in the two hard coral species, Pocillopora damicornis and Pavona decussata in moderate light (∼100 µmol photons m −2 s −1 ). The schematic diagram of a coral shows the coral tissue containing algal symbionts (green circles), which lies above the calicoblastic layer. Photosynthetic active radiation (PAR) (rainbow arrow) penetrates the coral tissue. In P. decussata a higher density of symbionts reduced light availability compared to P. damicornis . Dissolved inorganic carbon (grey arrows; quantity is relative to arrow thickness) can originate from internal sources such as the calicoblastic layer or from the external environment, where P. decussata draws stronger on the external carbon uptake. Light respiration (R) (strength indicated through size), was greater in P. damicornis than in P. decussata . Sufficient supply of CO 2 to the algal symbionts is of paramount importance for the functioning of a coral symbiosis [18] , [52] , [53] , where an increased supply enhances photosynthesis [31] . Gross photosynthesis rates (GP O2 micro ) were similar for both coral species across the applied irradiance levels. However, gross CO 2 uptake rates, as well as algal symbiont density were generally higher in P. decussata ( Fig. 1 B). These results raise the question as to why a coral with twice as many symbionts and greater CO 2 uptake ( P. decussata ) did not show a greater photosynthetic productivity. The coral P. decussata had a much greater protein biomass than the coral P. damicornis and the algal symbionts would have been more densely packed within the coral tissue. Self-shading of the algal symbionts [54] , as well as species-specific differences in light propagation within the host tissue [46] , [55] could explain our findings for P. decussata . A model of how canopy-understory development can influence the photosynthesis-irradiance (P-I) relationship has previously been introduced [56] . Here we could expand that model to introduce the light respiratory activity as well as carbon uptake in relation to how canopy-understory influences the P-I relationship in the two corals examined here (see Fig. 3 ). Light respiration in P. damicornis reached its maximum at a lower irradiance than in P. decussata and exceeded dark respiration ( Fig. 2 ). A higher proportion of GP O2 micro was therefore contributed by light respiration in P. damicornis than in P. decussata . Our results suggest therefore that species-specific light-driven respiratory processes are active within the two coral species. Light-driven respiration is often coupled to calcification in the calicodermis [14] , [29] , [33] , [36] , [57] and it seems possible that the calcification process accounts for a large fraction of the light respiration. For calcification to take place, O 2 and photosynthate are necessary so that the coral host can liberate adenosine-triphospate (ATP) for the calcifying process [58] , [59] . The hyperbolic increase in light respiration for both species, up to the maximum measured photon irradiance (1100 µmol photons m −2 s −1 ; Fig. 2 ) suggests that host respiration is closely coupled to release of photosynthates from zooxanthellae. However, recent attempts to investigate calcification and light respiration rates in corals, using an indirect measuring technique, found that light respiration increased the most in zooxanthellae as opposed to the coral host [60] . Given these results, it seems more likely that metabolic activity supporting calcification, e.g., Symbiodinium 's photosynthetic reaction and carbon fixation, are responsible for most of the increase in light respiration. Calcification itself is a positive feedback mechanism for Symbiodinium photosynthesis, as CO 2 is being produced during skeleton accretion [29] . Both species showed steady and light-independent gross CO 2 uptake rates at >78 µmol photons m −2 s −1 , where calcification could then fuel the photosynthetic activity through internal carbon release. However, the recently proposed ‘proton flux hypothesis’ [36] , where the shedding of protons generated during the calcification process is proposed to result in a lag of CO 2 uptake could also explain our results. Whether light respiration is simply controlled by the availability and source of carbon substrates or other metabolic controls remains to be investigated. In both corals, P. damicornis and P. decussata , light-saturated respiration rates (R light O2 micro ) at 210 µmol photons m −2 s −1 were similar. Light stimulated respiration in P. damicornis increased to a greater degree than that in P. decussata (25 versus 11 times). Light-saturated respiration rates in both species reached an asymptotic value of 5 nmol cm −2 s −1 at photon irradiances >210 µmol photons m −2 s −1 ( Fig. 2 ). The strong increase of respiration rates during the light as compared to steady-state dark respiration rates are most likely due to the low-light acclimation of the experimental corals (40 µmol photons m −2 s −1 ). Dark respiration rates are generally dependent upon pre-experimental incubation irradiances [61] , [62] . Under low light adaptation steady-state dark respiration rates are low but once exposed to light, the metabolic activity increases and so do light respiration rates and other oxygen uptake processes. The magnitude of this increase is independent on the pre-experimental incubation irradiance [62] . Photoacclimation is a process of morphological (here in terms of coral host) and physiological adjustments of a phototrophic organism towards growth irradiances. Pigmentation (coral host pigmentation [63] and light harvesting pigments such as accessory pigments and chlorophyll [64] ), as well as photochemical quenching capacity (xanthophyll pool [65] , [66] ) can be increased and decreased in abundance and concentrations. During high light exposure these adjustments help acclimatization in the phototroph only to some extend, and as a result, high light stress results in the accumulation of reactive oxygen species [67] , the stimulation of alternative electron transport systems [68] , [69] , often consuming oxygen, and of photorepair mechanisms [70] , [71] . The cost of all these processes results in low net photosynthesis [62] , due to increased respiration and other oxygen uptake [39] , [72] . The light source in the experiments of this study excluded the naturally occurring ultraviolet radiation, which corals experience in the field and which is a major cause of photodamage [73] , [74] . Translating our findings to corals in the field, the increase of oxygen uptake rates on going from dark to light (or from low to high light) might therefore not be as great as found in this study; however, once photorepair processes are entrained the actual oxygen uptake rates might be just as high or even higher. Pronounced stimulation of respiration in light has been reported for the coral species Galaxea fascicularis , where light respiration was ∼12 times higher than dark respiration under an irradiance of 140 µmol photons m −2 s −1 \n [14] . Kühl et al. [39] observed values of light respiration to be ∼6 times higher than during dark respiration in Favia sp. under an irradiance of 350 µmol photons m −2 s −1 . Here light respiration accounted for 77% of the gross photosynthetic O 2 production. The differing increase of respiration rates from dark to light between the reporting studies and our results are probably due to species differences and differential pre-experimental and experimental irradiances. In our study light respiration accounted for 88% of gross photosynthetic O 2 production in P. decussata and 97% of gross photosynthetic O 2 production in P. damicornis at 210 µmol photons m −2 s −1 . Maximum gross photosynthetic O 2 production were on average ∼0.53 nmol O 2 cm −2 s −1 for both coral species ( Fig. 1 ) and were of a similar magnitude to other microsensor measurements of gross photosynthesis rates in corals [75] . Light dependent increase in O 2 consumption through respiratory processes has been discussed previously [68] . Tchernov et al. [68] concluded that ongoing activity of the MAP cycle could be accounted for by the increased O 2 uptake with increasing photon irradiance. Indeed, various light-driven O 2 consuming processes, such as photorespiration [76] , [77] and the MAP cycle [68] , [78] , [79] could also be involved in the high level of light respiration observed here. However, the activity of the MAP cycle does not result in net O 2 concentration changes [78] ; it therefore cannot be measured in O 2 exchange measurements with microsensors [80] . Hence, we conclude that the only other process to explain the light respiration results apart from light-stimulated mitochondrial O 2 uptake is photorespiration, involving oxygenase activity of RuBisCO [81] . However, further investigations are needed to verify and describe these processes."
} | 4,388 |
30797225 | PMC6462302 | pmc | 2,992 | {
"abstract": "Background: Magnetotactic bacteria are a heterogeneous group of Gram-negative prokaryote cells that produce linear chains of magnetic particles called magnetosomes, intracellular organelles composed of magnetic iron particles. Many important applications have been defined for magnetic nanoparticles in biotechnology, such as cell separation applications, as well as acting as carriers of enzymes, antibodies, or anti-cancer drugs. Since the bacterial growth is difficult and the yield of magnetosome production is low, the application of magnetosome has not been developed on a commercial scale. Methods: Magnetospirillum gryphiswaldense strain MSR-1 was used in a modified current culture medium supplemented by different concentrations of oxygen, iron, carbon, and nitrogen, to increase the yield of magnetosomes. Results: Our improved MSR-1 culture medium increased magnetosome yield, magnetosome number per bacterial cell, magnetic response, and bacterial cell growth yield significantly. The yield of magnetosome increased approximately four times. The optimized culture medium containing 25 mM of Na-pyruvate, 40 mM of NaNO3, 200 µM of ferrous sulfate, and 5-10 ppm of dissolved oxygen (DO) resulted in 186.67 mg of magnetosome per liter of culture medium. The iron uptake increased significantly, and the magnetic response of the bacteria to the magnetic field was higher than threefold as compared to the previously reported procedures. Conclusion: This technique not only decreases the cultivation time but also reduces the production cost. In this modified method, the iron and DO are the major factors affecting the production of magnetosome by M. gryphiswaldense strain MSR-1. However, refining this technique will enable a further yield of magnetosome and cell density.",
"introduction": "INTROUDUCTION One of the most common Gram-negative prokaryotic cells with heterogeneous characteristics is magnetotactic bacteria being able to produce linear chains of bacterial magnetic particles (BacMPs) called magnetosomes[ 1 ]. BacMPs are intracellular organelles composed of magnetic iron particles surrounded individually by a phospholipid bilayer. The size of the magnetosome particles often varies within the species, ranging from 35 to 120 nm[ 2 ]. Magnetosomes mainly composed of magnetite (Fe 3 O 4 ) or greigite (Fe 3 S 4 ) are assembled as one or more chain(s) depending on different elements and typically located close to the cytoplasmic membrane. Meanwhile, the number of magnetosomes in Magnetospirillum gryphiswaldense MSR-1 often differs with regard to the environmental conditions[ 3 , 4 ]. Magnetic bacteria have high biomineralization ability and are able to adjust themselves to new environmental conditions such as sever deprivation of metals[ 5 ]. It has been suggested that prokaryotes can be classified into biologically induced and biologically controlled mineralization based on their ability in synthesis of minerals[ 6 ]. Microbial reduction of metals and formation of magnetosomes within microorganisms in a marine environment were first discovered by Blakemore in 1975[ 7 ]. The biological production of the biomineralized magnetosomes is strictly controlled at the gene level, and the magnetosomes are normally formed in different sizes and shapes in magnetosome membrane[ 6 ]. Magnetosome synthesis has recently been proposed as a model for the formation of prokaryotic organelles and biomineralization[ 8 , 9 ]. Although the details of the mechanism for the synthesis of magnetosomes are not exactly clear, studies have shown that the formation of magnets is a cellular process that depends on several stages, including the separation of the internal membrane of the cell, the transfer of ions, the crystallization of magnetite within these vesicles, and the formation and arrangement of adult crystals as a linear structure of the cellular skeleton[ 1 , 10 ]. Unlike the chemical synthesis of other nanocrystals, magnetosomes are synthesized via unique features, including a perfect crystallographic appearance, a narrow and single magnetic domain in nanosize range with a permanent magnetization, and the formation of a biocompatible lipid bilayer around each mineral particle[ 2 , 11 - 13 ], which bring about an exceptional importance in biotechnological applications of magnetic nanoparticles such as nuclear magnetic resonance, cell separation assays as drug carriers, and destruction of tumor cells by hyperthermia[ 14 - 18 ]. Since 1991, several applications including carriers for enzymes[ 19 ], nucleic acids[ 8 , 20 ], and antibodies[ 19 ] as well as anticancer drugs[ 8 , 9 , 21 ] have been reported for bacterial magnetosomes. However, because of the difficulty in growing magnetotactic bacteria and the low-yield production of magnetosomes, these applications have not been extended to commercial scale[ 10 , 13 ]. Various kinds of culture media such as the optimized flask medium (OFM), large-scale medium, magnetic spirillum growth medium, and optimized growth medium have been developed to fulfill the requirements of magnetotactic bacteria[ 11 ]. Adjustment of oxygen, temperature, and redox potential have been demonstrated to be remarkably effective in magnetosome production and magnetotactic bacterial yield in fed-batch flasks and bioreactors[ 11 , 12 , 22 ]. Moreover, most of the magnetotactic bacteria strains have been found to consume oxygen, ferric quinate, and nitrate as electron acceptors and use succinate or lactate, acetate, and nitrate as electron donors[ 11 ]. Culture medium optimization was very effective for high-yield cultivation of magnetosome, as it has previously been reported that at higher dissolved oxygen (DO) level, cell growth would be greater, but for higher yield of magnetosome, low DO concentrations are the prerequisite[ 23 , 24 ]. Thus, to resolve this situation, it is necessary to enhance DO to an optimum level by stirring the medium to increase the magnetotactic bacteria growth and allow the microbe to lower DO by the respiration process to an optimum level. Some of the important obstacles in the mass culture of magnetic bacteria are adjusting the oxygen level and optimizing the culture medium. The main aim of this study was to investigate the effect of various concentrations of medium compounds, including oxygen, iron, carbon, and nitrogen to improve the growth of M. gryphiswaldense MSR-1 and to increase the yield of magnetosome.",
"discussion": "DISCUSSION In this report, we attempted to increase the production of magnetosome in MSR-1 strain of M. gryphiswaldense by optimizing the supply of oxygen, iron, carbon, and nitrogen. The results showed that the oxygen has a critical role in the synthesis of magnetosomes. At DO level above 5-10 ppm, the iron uptake and magnetosome production reduced, but the bacterial growth was normal. However, at lower DO (<5-10 ppm), the rate of iron uptake, Cmag, and magnetosome production increased that is likely associated with the slow growth of bacteria. These observations are in agreement with those reported by Yang et al .[ 13 ]. It has also been reported that magnetosome yield could be significantly increased under microaerobic conditions; however, the exact role of oxygen in magnetosome biomineralization is still unclear[ 11 ]. Most reports have claimed that oxygen is necessary to maintain the required redox potential for magnetosome growth[ 12 , 28 ]. M. gryphiswaldense MSR-1 is facultative anaerobic bacteria with aerobic condition preference for the cell growth[ 13 ]. During magnetotactic bacteria growth the concentration of O 2 had a specific effect on the synthesis of magnetosomes. It increases the magnetosome production at concentrations lower than 5-10 ppm, whereas at higher concentrations, it decreases the magnetosomes[ 4 , 29 , 30 ]. Therefore, controlling DO level in the culture medium or altering aerobic/anaerobic conditions is favorable for magnetosome production[ 4 , 25 , 31 - 33 ]. In addition to oxygen levels, our study showed that the iron source is another important factor for more efficient uptake of this vital ion and the magnetosome production. Ferrous sulfate was better source of iron for M. gryphiswaldense than ferric quinate and citrate, perhaps because the reduced form of ferrous is more soluble compared to the oxidized analogue. This result is in line with the findings of Yang et al .[ 24 ]. We showed that the excess concentrations of iron in the culture medium can be toxic for M. gryphiswaldense MSR-1. A slight increase in extracellular iron concentration elevated the iron uptake, magnetosome production, Cmag, and cell growth, but the ferrous sulfate concentration above 150 μM led to reduction in the cell growth and enhancement of the iron uptake rate, Cmag, and magnetosome production. Iron concentrations ≥300 μM could damage the bacteria seriously and significantly reduce the rate of iron uptake, Cmag, magnetosome production, and the growth rate ( Fig. 3 ). In magnetotactic bacteria, iron not only acts as a protein cofactor but also accelerates the biomineralization process in the cells[ 6 , 14 , 29 ]. It has been demonstrated by Faivre et al .[ 29 ] that M. gryphiswaldense utilizes soluble ferrous ions and ferric (ferritin form) synchronously for the magnetosome production. Maximum magnetosome production was observed at 200 μM of iron, which was in agreement with the reported results[ 34 - 36 ]. Using TEM analysis, we proved that in the optimized culture medium, both the quantity and size of magnetosome increased ( Fig. 7 ). This observation was also noted by Liu et al .[ 12 ]. Furthermore, the increased number of double chains in bacteria indicates the optimal medium is an appropriate condition for the growth of M. gryphiswaldense ( Fig. 7 ). Our results confirmed the superiority of pyruvate and lactate for magnetosome production over acetate and succinate as reported previously[ 12 , 28 ]. This preference may indicate that pyruvate and lactate have better redox potential in comparison with succinate and acetate as the electron donor for ATP synthesis[ 37 ]. We showed that 25 mM sodium pyruvate was the optimal concentration for magnetosome production by Magnetospirillum sp ., as reported also by Liu et al .[ 12 ] and Zhang et al .[ 22 ]. They concluded that such behavior can be related to the bacterial susceptibility to increased osmotic potential[ 22 ]. For magnetosome synthesis, the role of nitrogen is very important to form proteins for the assembly of magnetosome and to transport and incorporate iron in the magnetotactic bacteria. Our data showed that in the magnetotactic bacteria, the use of NaNO 3 , as a nitrogen source, leads to more magnetosome production versus using NH 4 Cl and (NH 4 ) 2 SO 4 . This result contradicts the results of Liu et al .[ 12 ]. Nitrate is a strong oxygen acceptor having redox potential, which supports our observations ( Fig. 6 ). The culture medium containing the low concentration of nutrients (in particular carbon and nitrogen) is a key limiting factor that affects magnetosome production and cell density of all magnetotactic bacteria. Our results showed that in magnetosome synthesis, the rate of iron uptake and Cmag enhanced at lower concentrations of NaNO 3 (40 mM), as compared to the higher concentrations (60 mM), which support previous results[ 38 ]. Proteins promote the nucleation of iron crystals, leading to magnetosome production in magnetotactic bacteria after the vesicles assembling[ 39 ]. It has recently been illustrated that the magnetosome-associated membrane proteins play a significant role in magnetosome crystals growth inside the magnetotactic bacteria[ 39 , 40 ]. Our result showed that within 50-h culture, a maximum amount of magnetosome was produced (about 186.87 mg L −1 ), and 5.76 g L −1 bacterial cells were obtained, whereas the cell growth rate was 2.768 g L −1 /day. The magnetic nanoparticles are synthesized in the presence of low oxygen and high level of iron concentrations under the influence of intracellular microbial reduction as trivalent iron compounds[ 4 , 41 ]. The numbers and shapes of the magnetosomes vary depending on the conditions of the culture medium, in particular, the concentration of available soluble iron, DO, nutrients, temperature, pH, and degradation power. On average, between 10 and 30 magnetosomes are synthesized in each bacterial cell. For the synthesis of magnetosomes, the bacterial cells need to uptake a large amount of iron from its surroundings and put in the bio-mineralization process[ 11 , 38 ]. Also, these crystals are required to be in well-defined numbers, shapes, and crystal size, in order to play an efficient magnetic field sensor role. The composition of additives in culture medium can have significant effects on the size and other magnetic properties of these nanoparticles[ 28 ]. In summary, we report an improved MSR-1 culture medium condition for increased magnetosome yield, magnetosome number per cell, magnetic resonance, and cell yield in a shorter time and the reduced cost. Our method allows achieving the mass production of magnetosomes by MSR-1 in a fermentor scale by choosing the suitable sources and concentrations of culture medium composition. The most important advantages of this protocol are: (1) the concentrations of oxygen, iron, carbon, and nitrogen sources in the medium can be auto-controlled at a constant level by pH-stat feeding, leading to ease of manipulation and elimination of the possibility of nutrient exhaustion during the culture process and (2) easy up-scale process for industrial production without need for genetic manipulation. The DO and iron are the major factors affecting the magnetosome production in the bacterial culture. Further refinements of this protocol to overcome other shortages of the magnetosome production of magnetotactic bacteria are ongoing in our laboratory."
} | 3,476 |
30258172 | PMC6331573 | pmc | 2,994 | {
"abstract": "Delignification, or lignin-modification, facilitates the decomposition of lignocellulose in woody plant biomass. The extant diversity of lignin-degrading bacteria and fungi is underestimated by culture-dependent methods, limiting our understanding of the functional and ecological traits of decomposers populations. Here, we describe the use of stable isotope probing (SIP) coupled with amplicon and shotgun metagenomics to identify and characterize the functional attributes of lignin, cellulose and hemicellulose-degrading fungi and bacteria in coniferous forest soils from across North America. We tested the extent to which catabolic genes partitioned among different decomposer taxa; the relative roles of bacteria and fungi, and whether taxa or catabolic genes correlated with variation in lignocellulolytic activity, measured as the total assimilation of 13 C-label into DNA and phospholipid fatty acids. We found high overall bacterial degradation of our model lignin substrate, particularly by gram-negative bacteria (Comamonadaceae and Caulobacteraceae), while fungi were more prominent in cellulose-degradation. Very few taxa incorporated 13 C-label from more than one lignocellulosic polymer, suggesting specialization among decomposers. Collectively, members of Caulobacteraceae could degrade all three lignocellulosic polymers, providing new evidence for their importance in lignocellulose degradation. Variation in lignin-degrading activity was better explained by microbial community properties, such as catabolic gene content and community structure, than cellulose-degrading activity. SIP significantly improved shotgun metagenome assembly resulting in the recovery of several high-quality draft metagenome-assembled genomes and over 7500 contigs containing unique clusters of carbohydrate-active genes. These results improve understanding of which organisms, conditions and corresponding functional genes contribute to lignocellulose decomposition.",
"introduction": "Introduction The incomplete decomposition of woody biomass in coniferous forests is an important global carbon sink, with approximately one third of a gigaton of carbon accruing on an annual basis [ 1 ]. Lignocellulose decomposition is influenced by the structure and function of microbial communities [ 2 , 3 ] and terrestrial carbon cycling models increasingly parameterize these properties [ 4 , 5 ]. However, efforts are constrained by our rudimentary knowledge of the composition and ecology of decomposer populations, stemming from limitations of culture-dependent methods and the complexity of soil communities. The best characterized decomposers inhabit forest soil litter, where conditions favour rapid lignocellulose degradation [ 6 , 7 ], yet the decay of lignin-rich plant polymers in soil occurs in a continuum governed by conditions and substrate accessibility, resulting in diversified niches for decomposers [ 8 – 10 ]. To resolve the catabolic and ecological traits of decomposers, we must utilize culture-independent methods, like stable isotope probing (SIP), that better reflect in situ conditions. The nature of microbial lignin-degradation is poorly described beyond the canonical breakdown of lignin in woody biomass by specialized, aerobic, litter-inhabiting wood-rot fungi. Yet, these fungi are largely absent in deeper mineral soil where conditions favour decomposition by bacteria [ 11 ], who represent the most likely lignin-degraders when oxygen is limited [ 12 – 14 ]. Several soil bacteria can degrade model lignin compounds in pure culture, suggesting a role in the catabolism of low-molecular weight, partially-degraded forms of lignin [ 15 – 21 ]. Knowledge of bacterial lignin-degradation in environmental contexts is limited, with all studies of soil degraders originating from tropical forests which describe active populations of predominantly Alpha and Gammaproteobacteria [ 13 , 22 ]. The existing evidence for bacterial lignin-degradation demonstrates the need to characterize both bacterial and fungal populations and contrast their roles in different soil environments to better understand in situ the processes that govern the decomposition of lignocellulose. Delignification rapidly increases the rate of lignocellulose decomposition and may have evolved primarily as a means of accessing more readily degradable plant carbohydrates, exemplified by the strategies of white and brown-rot Agaricomycota [ 23 – 25 ]. The capability to co-degrade lignin and other lignocellulosic polymers has not been thoroughly explored in bacteria. There is some evidence for the co-metabolism of lignocellulosic polymers by various Streptomyces spp . [ 26 ], while a comparative genomics study revealed that cellulose-degrading bacteria also possess higher numbers of hemicellulases [ 27 ]. The extent to which catabolic traits are conserved within specific taxa versus communities will influence models of microbial decomposition. In the simplest case, the abundance of highly adapted, multi-substrate degrading taxa may predict rates of decomposition, which is supported by a small number of studies [ 2 , 28 ]. However, forces of genomic streamlining in bacteria lead to functional diversification among closely related species, particularly in extra-cellular processes that produce common goods [ 29 ], evident in the species-level conservation of bacterial endoglucanases [ 30 ]. A multi-substrate SIP experiment provides the means to identify whether forest soil decomposers can assimilate 13 C from various lignocellulosic polymers and, with shotgun metagenomic sequencing, determine whether their genomes encode the necessary suite of catabolic genes. To address the above knowledge gaps, we utilized SIP microcosm-based experiments to investigate the composition and degradative potential of hemicellulose, cellulose and lignin-degrading populations from organic and mineral layer soils in coniferous forests across North America. The identity of degraders and their genomic content were assessed using amplicon and shotgun sequencing of DNA enriched in 13 C from labeled substrates. Lignocellulolytic activity was quantified according to the amount of 13 C assimilated into total DNA and phospholipid fatty acids (PLFA). The in situ abundances of lignocellulolytic populations were determined on the basis of previously reported pyrotag libraries from the same field samples [ 31 , 32 ]. This is currently the most comprehensive cultivation-independent study of lignocellulolytic populations in forest soils and yields new insights to the taxa responsible and the importance of certain catabolic gene families.",
"discussion": "Discussion We failed to find support for our initial hypothesis that individual species would degrade multiple components of lignocellulose. Instead, we identified a diverse array of taxa assimilating 13 C from three major lignocellulosic polymers, demonstrating the degree of specialization in decomposer populations. While approximately one third of taxa enriched in 13 C-DNA pools possessed members that could collectively degrade more than one substrate, and, though a handful of MAGs encoded a putative suite of catabolic enzymes, none of the individual bacterial OTUs (at a 99% similarity threshold) were definitively enriched from all three substrates. Even closely related taxa, like Caulobacter and Asticcacaulis , exhibited differences in their levels of enrichment on cellulose and lignin and in CAZy gene content. In contrast, the multi-substrate degradative capacity of fungi was evident in the increased relative abundances of genes encoding both delignifying (peroxidases) and cellulolytic enzymes in 13 C-DNA pools. The recovery of a Myceliophthora MAG from 13 C-metagenomes was consistent with its known capability for complete lignocellulose decomposition [ 69 , 70 ]. Therefore, our findings suggest that complementation among functional guilds may be necessary for decomposition of lignocellulose by bacteria and not necessarily fungi. One caveat is the possibility that our method was biased towards identifying single substrate utilizing bacteria. The activity of multi-substrate users may have been masked by populations that grew faster on individual substrates, particularly the relatively labile hemicellulose, as was observed in other SIP experiments [ 71 ]. At the very least, our study demonstrates the decomposition of lignocellulosic polymers can commonly occur via a division of labour among specialized taxa. The substantial assimilation of carbon from our model lignin substrate by bacteria supports our hypothesis that bacteria contribute significantly to degradation of native forms of lignin in situ. Bacterial activity was particularly evident in deeper mineral layers of forest soil, suggesting that lignin decomposition can occur throughout the soil column. Interestingly, the most novel lignin-degraders from mineral soils belong to uncultured clades of Caulobacteraceae, Acidobacteria, Solirubrobacterales, Elusimicrobia, Nevskiales, and Cystobacteraceae. The low activity of fungal lignin degraders in our microcosms was unlikely due to unmet nutritional or environmental conditions, since metabolically competent taxa like Myceliophthora [ 69 , 70 ] assimilated carbon from cellulose under the same conditions. One possible explanation is that fungi were involved in lignin-degradation but did not metabolize degradation products to become sufficiently labeled, perhaps because of bacterial cross-feeding, evident in the inhibitory effect of fungicide on lignin incorporation by bacteria. Indeed, the results from the fungicide-treatments indicated significant interactions occurred between bacteria and fungi in lignin degradation. However, fungi were not essential for bacterial lignin degradation, given the persistent and substantial incorporation of 13 C-lignin by bacteria in fungicide-treated soils. This study provides strong evidence for the capacity of bacteria to degrade lignin, highlighting the need for further characterization of their activity and ecology in soil. The substantial number of taxa identified by SIP with previously reported lignocellulolytic activity was testament to the success of previous culturing-based efforts to characterize soil decomposers and a validation of our SIP approach. Approximately 72% (14/19) of hemicellulolytic taxa and 74% (31/42) of cellulolytic taxa were previously reported to degrade the corresponding substrate in vitro. Far fewer lignolytic taxa (~28%, 8/29) had previously reported degradative activity, likely due to less frequent and formalized testing of lignin degradation. However, there was considerable agreement in the bacterial genera we identified and those identified in other culture-independent studies of lignin-degraders in forest soil, such as Caulobacter , Sphingomonas , Sphingobacterium , Nocardia , Telmatospirillum and Azospirillum [ 13 , 22 , 72 ]. A lower proportion of lignin-degrading taxa associated with mineral layer soil (5/16) had previously reported activity compared to those associated with the organic layer (8/12), suggesting the former may be more difficult to culture or were less frequently targeted for study. Many of the novel cellulolytic groups we identified belong to cultivation-resistant phyla, Planctomycetes, Verrucomicrobia, Chloroflexi and Armatimonadetes (formerly OP10), which commonly predominate soil communities [ 73 , 74 ]. Each of these phyla possesses at least one isolate capable of degrading cellulose [ 75 – 78 ] and have been designated cellulolytic in other SIP-cellulose experiments [ 28 , 71 , 79 , 80 ]. Notably, all contigs that contained clusters of ten or more CAZymes (27 contigs), from 13 C-cellulose metagenomes, were classified to taxa from the aforementioned groups and contained both CBMs (26/27) and endoglucanases (20/27). These findings supported our assertion that the richness of described lignocellulolytic taxa is underestimated in soils, which limits our understanding of the ecology of decomposition and may afford new types of biocatalysts for processing lignocellulosic biomass. The widespread capacity among members of Caulobacteraceae to degrade the three components of lignocellulose was unexpected, yet consistent with reports of their enrichment on decaying wood [ 81 , 82 ] and during early stages of litter decomposition [ 83 ]. Caulobacter were first isolated from cellulose-amended lake-water [ 84 ], but are primarily known as oligotrophic, aquatic organisms [ 85 ]. Their role in degradation of plant carbohydrates was first postulated based on the analysis of the C. crescentus genome [ 86 ] and subsequently demonstrated by growth on cellulose, xylose and vanillate [ 87 – 90 ]. Although we provide the first evidence for the role of Caulobacteraceae in degrading all three polymers of lignocellulose, several surveys of forest soils report enrichment of Caulobacter or Asticcacaulis in samples amended with cellulose [ 80 , 91 , 71 ] or lignin [ 13 , 22 ]. The role of Caulobacteraceae in decomposition in forest soils may prove significant, given their relatively high abundance (0.5–2.5% of total libraries) and their capacity to adhere to insoluble polymers, like lignocellulose. The taxonomy and catabolic capacity of lignocellulose degraders were moderately associated with variation in the total lignocellulolytic activity. The influence of compositional differences between soil layers was, in certain cases, overshadowed by the strong negative correlation between organic matter and 13 C enrichment. Yet, in regression modeling, CAZy gene content and community composition had equivalent or greater explanatory power than other predictors. The abundances of several prominent lignocellulolytic taxa were significantly correlated with higher 13 C assimilation, demonstrating that certain taxa play more significant roles, at least under conditions we tested. Strongly oxidative bacterial enzymes commonly studied in relation to lignin-degradation, such as DyP-type peroxidases and laccases [ 92 – 94 ], were less predictive of lignolytic activity than aryl alcohol oxidases, suggesting a greater role of the latter class of enzymes in bacterial lignin-degradation. Overall, fewer endoglucanase genes were predictive of 13 C-assimilation compared to AA gene families. This difference suggests either the existence of a greater diversity of endoglucanases, minimizing the explanatory power of any single gene, or a narrower activity range of AA families with a correspondingly higher relevance to lignolytic activity. This result agrees with the finding that the composition of decomposer communities had increased explanatory power for the rate of litter decomposition during later stages when greater proportions of lignin remained (Cleveland et al. 2014). The comprehensive study of all three major lignocellulosic polymers enabled us to examine the co-occurrence of lignocellulolytic traits, addressing several knowledge gaps in this area. The results indicated unexpected specialization among bacterial populations for degradation of individual lignocellulosic polymers and revealed several novel lignocellulolytic taxa, highlighting current limits to our knowledge of decomposition. This research supports the view that bacterial decomposition of oligomeric lignin is a ubiquitous soil process, with the potential to occur in deeper soil layers following early stages of litter decomposition. As hypothesized, variation in community composition was found to constrain lignocellulolytic activity across various forest types and soil layers in North America. The relationship between these communities and process rates should receive continuing study to refine our understanding of soil carbon stabilization and terrestrial carbon cycling models. Furthermore, the large number of degradative gene clusters from uncultivated lignocellulolytic taxa represent a trove of potentially novel enzymes for biotechnological applications."
} | 4,002 |
34168314 | PMC8630044 | pmc | 2,995 | {
"abstract": "Elevated seawater temperatures have contributed to the rise of coral disease mediated by bacterial pathogens, such as the globally distributed Vibrio coralliilyticus , which utilizes coral mucus as a chemical cue to locate stressed corals. However, the physiological events in the pathogens that follow their entry into the coral host environment remain unknown. Here, we present simultaneous measurements of the behavioral and transcriptional responses of V. coralliilyticus BAA-450 incubated in coral mucus. Video microscopy revealed a strong and rapid chemokinetic behavioral response by the pathogen, characterized by a two-fold increase in average swimming speed within 6 min of coral mucus exposure. RNA sequencing showed that this bacterial behavior was accompanied by an equally rapid differential expression of 53% of the genes in the V. coralliilyticus genome. Specifically, transcript abundance 10 min after mucus exposure showed upregulation of genes involved in quorum sensing, biofilm formation, and nutrient metabolism, and downregulation of flagella synthesis and chemotaxis genes. After 60 min, we observed upregulation of genes associated with virulence, including zinc metalloproteases responsible for causing coral tissue damage and algal symbiont photoinactivation, and secretion systems that may export toxins. Together, our results suggest that V. coralliilyticus employs a suite of behavioral and transcriptional responses to rapidly shift into a distinct infection mode within minutes of exposure to the coral microenvironment.",
"introduction": "Introduction Coral reefs are declining worldwide due to rising sea surface temperatures and increasing prevalence of coral disease outbreaks [ 1 – 3 ]. Elevated sea surface temperatures cause physiological stress in corals [ 4 ] and provide distinct advantages for some coral pathogens [ 5 ]. One well-studied coral pathogen, Vibrio coralliilyticus BAA-450, displays tightly regulated temperature-dependent virulence against its coral host, Pocillopora damicornis . While this V. coralliilyticus strain is avirulent at temperatures below 24 °C, it is capable of attacking the coral symbiotic dinoflagellates [ 6 , 7 ] and lysing coral tissue [ 8 ] at temperatures above 27 °C. V. coralliilyticus displays two distinct behavioral adaptations enabling targeted infection of corals that are physiologically stressed and therefore more vulnerable to pathogenic invasion. First, the bacterial pathogen uses chemotaxis to target chemical signatures present in the mucus of stressed corals [ 9 ]. Second, V. coralliilyticus displays chemokinesis, which is the ability to change swimming speed in response to a change in chemical concentration, to potentially enable faster environmental exploration in the presence of its coral host mucus [ 9 , 10 ]. While both chemotaxis and chemokinesis are behaviors associated with motility and chemical sensing, chemotaxis specifically refers to the ability to detect and follow chemical gradients (without necessarily any change in swimming speed), whereas chemokinesis refers to the ability to change swimming speed in response to an overall change in concentration in the environment (without any reference to whether cells follow gradients). In V. coralliilyticus , these two behaviors combine to enable efficient and rapid targeting of stressed corals. However, chemokinesis, in contrast to chemotaxis, has remained more rarely studied in bacteria and is almost entirely undescribed in the context of marine disease [ 11 – 13 ]. Coral mucus—in addition to triggering increased motility and chemotaxis which are behaviors necessary for infection by V. coralliilyticus [ 14 , 15 ]—also represents the critical interface where pathogen activities can dictate the outcome of an infection. Corals secrete up to half of the carbon assimilated by their algal symbionts as mucus [ 16 , 17 ] and its production represents a sizable energetic investment that is important for nutrient cycling across the entire reef system [ 18 – 21 ]. In addition, mucus provides corals with protection against desiccation and is an ancient and evolutionarily conserved first line of defense against pathogens [ 22 ]. During infection studies, corals have been observed to actively expel ingested pathogens by spewing out bacteria-laden mucus from the mouths of polyps [ 23 , 24 ]. However, entry into host mucus may also signal to the bacterial pathogen that contact with a potential host is imminent. Thus, elucidating the behavioral and transcriptional responses of V. coralliilyticus in the context of its coral host environment, and in particular coral mucus, is important in elucidating the mechanisms underpinning coral infection. Here, we present experiments in which we simultaneously used video microscopy and RNA sequencing to measure the behavioral and transcriptional responses of V. coralliilyticus upon a sudden exposure to mucus from its coral host. To study chemokinesis independently from chemotaxis, we conducted our experiments in the absence of chemical gradients. This represents the first investigation to couple behavioral and transcriptomic analyses to decipher the mechanisms promoting coral infection. We show that behavioral and transcriptional responses occur concomitantly over a surprisingly rapid timescale of only minutes, highlighting the agility of the pathogens in seizing what are likely to be limited windows of opportunity [ 25 ] to target and ultimately infect their host.",
"discussion": "Discussion We have reported a rapid behavioral and transcriptional response of V. coralliilyticus to coral mucus exposure, which led to a two-fold increase in swimming speed and significant differential expression of 53% of the genes in the genome within 10 min. Our findings identify coral mucus as a potential chemical signal that induces pathogens to prepare for host colonization and infection. These responses are in line with the behavioral and physiological versatility characteristic of marine copiotrophic bacteria, which are often adapted to boom and bust lifestyles [ 64 , 65 ]. Yet the extent and the rapidity of the responses observed here suggest that temporally precise orchestration of behavior and gene expression is important for coral host colonization by V. coralliilyticus . Chemokinesis in response to exposure to coral mucus is potentially a strategy for V. coralliilyticus to seize a limited window of opportunity to reach the coral surface. By increasing swimming speed, bacteria also enhance their chemotactic velocity, leading to a decrease in the time required to follow a chemical gradient to its source. This was previously shown for V. coralliilyticus using microfluidic gradient experiments [ 9 , 10 ] and appears to be a more general feature of Vibrios, having also been observed in V. alginolyticus chemotaxing toward amino acids [ 13 ]. While swimming fast is expensive in the typically dilute ocean environment [ 66 ], energy is no longer limiting once nutrient-rich mucus is available. Instead, what is limiting is the window of time during which bacteria can exploit that mucus signal to reach the host. Not only can ambient water currents transport bacteria past the coral surface, but intense vortical flows produced by the corals themselves through cilia on their surfaces—moving at speeds much greater than bacterial swimming speeds—can result in rapid alternation of transport toward and away from the coral surface [ 25 ]. In this hydrodynamic environment, the colonization of a host by a bacterial pathogen is a challenging behavioral feat, where the opportunity to home in and attach to the coral surface may only last minutes or even less. The rapid behavioral response we reported here is consistent with this dynamic environment. In particular, the strong chemokinesis—where bacteria doubled their speed—is consistent with the need to reduce the time required to migrate to the coral surface once the detection of mucus indicates the presence of a coral. Furthermore, we observed that chemokinesis in response to coral mucus was almost entirely absent at a temperature at which V. coralliilyticus is avirulent (18.7 °C), which is consistent with the temperature dependence of chemokinesis observed in our previous study [ 10 ]. Thus, we propose that chemokinesis is a virulence trait that is important for successful host colonization by bacterial pathogens in the dynamic host surface environment. Entry into coral mucus represents a dramatic change in nutrient exposure for V. coralliilyticus compared to the oligotrophic reef waters. Accordingly, V. coralliilyticus rapidly upregulated metabolic pathways of nutrients that are present in coral mucus, which may fuel the energetically expensive chemokinesis trait, as well as protein production (ribosome and tRNA biosynthesis) and cell growth ( ftsZ and rpoD ) genes that may enable rapid proliferation and confer a competitive advantage to the pathogens as they invade the coral host microbiome [ 67 , 68 ]. Chemokinesis upon homogeneous addition of nutrients has been observed in other bacteria including Rhodobacter sphaeroides [ 69 ], Escherichia coli [ 70 ] and Azospirillum brasilense [ 71 ], and it has been speculated that this swimming speed enhancement is mediated by increasing the proton motive force that is responsible for flagellar rotation [ 12 , 71 ]. In line with this, V. coralliilyticus exposed to coral mucus increased the expression, on a similar timescale as the chemokinesis behavior, of genes encoding the Na + -NQR enzyme, suggesting that regulation of periplasmic sodium levels may help control swimming speed. Thus, we hypothesize that the metabolism of mucus substrate stimulates Na + -NQR activity, which in turn enables sustained chemokinesis. Additional experimental work is required to test this hypothesis. Despite the ~2× increase in swimming speed observed through video microscopy, flagellar genes were downregulated at the early RNA-seq time point (10 min). The swimming phenotype may thus persist using the existing polar flagellum, while downregulation of flagellar genes may be a strategy to prevent further replenishment of the flagellar apparatus during the transition to a non-motile phase, evidenced by the concurrent upregulation of biofilm genes. This observation has a parallel in the removal and downregulation of flagella observed in pathogens within the human mucosa, where it is speculated to be a strategy to escape immunological detection by the host, since flagella are strong inducers of pro-inflammatory signaling [ 72 ]. While corals possess innate and adaptive-like immunity [ 73 ], whether a similar dynamic occurs on the coral surface is currently not known. Following only 10 min of exposure to coral mucus, the master regulator of Vibrio virulence, ToxR, and its associated protein ToxS, were upregulated. ToxR is known to be essential for coral infection by V. coralliilyticus [ 74 , 75 ], and in other Vibrio pathogens the ToxR regulatory system coordinates the transcription of colonization, motility, and virulence genes in response to environmental conditions [ 76 , 77 ]. These downstream effects of ToxR were indeed observed in our RNA-seq results. The temporal modulation of quorum-sensing autoinducer molecule (AI-1, AI-2, CAI-1) producers as seen in our RNA-seq data may be a strategy to coordinate metabolic and lifestyle transitions at the population level, as has been observed in Vibrio harveyi [ 78 ]. One such lifestyle transition may be biofilm formation, which is tightly regulated by quorum sensing in Vibrio pathogens [ 47 , 79 ]. Indeed, we observed the upregulation of biofilm-related vps and rbm gene clusters in coral mucus at 10 min. Furthermore, we observed the upregulation of important Vibrio toxins, VcpB zinc metalloprotease, and VchA and VchB hemolysins, in coral mucus at both 10- and 60-min time points. Similarly, several secretion systems were upregulated, including the Sec-dependent and type 2 secretion systems, which are together responsible for extracellular secretion of a broad range of proteins, including toxins and degradative enzymes involved in the pathogenesis of many Gram-negative bacteria [ 80 – 82 ]. Type 6 secretion systems have been observed to be responsible for the injection of toxic effector proteins into bacterial cells in antagonistic interactions [ 83 – 85 ]. Taken together, the upregulation of toxR and toxS , as well as their downstream gene expression effects, suggest that coral mucus serves as an environmental signal for V. coralliilyticus to activate host colonization and virulence gene expression programs. The VcpB zinc metalloprotease is a key virulence factor of V. coralliilyticus that causes photoinactivation of coral endosymbionts and coral tissue lesions [ 7 ], and its rank as one of the most highly and significantly upregulated genes in our RNA-seq dataset suggests that the bacterium rapidly responded to coral mucus as a cue to initiate its virulence program. However, the second zinc metalloprotease that has been implicated in V. coralliilyticus infections of corals, VcpA (EEX33179) [ 8 ], was downregulated in our experiment. The two zinc metalloproteases (VcpA and VcpB) may thus play redundant roles in V. coralliilyticus infections and may be important in different environmental contexts (Supplementary Discussion). Our results underscore the rapidity of behavioral and transcriptional changes that occur in a coral pathogen upon entry into the host environment (Fig. 5 ). These changes in swimming and gene expression patterns paint a clear sequence of events immediately preceding infection—although further validation with direct phenotypic evidence is required. Upon exposure to coral mucus, the coral pathogen V. coralliilyticus (known to chemotax towards coral mucus [ 9 ]) increases swimming speed by up to two-fold within minutes, a response that, in the natural environment, would lead to faster chemotaxis and a halving of the time required for the pathogen to track the coral surface from which the mucus signal originates. This capacity to rapidly chemotax into the coral surface microenvironment is important because of the short window of opportunity available to pathogens in the hydrodynamic environment surrounding corals. Simultaneously, transcriptional changes indicate that mucus exposure immediately prompts V. coralliilyticus to increase nutrient metabolism and prepare for host colonization and damage. The downregulation of motility genes, puzzling at first in view of the strong chemokinetic response, is in fact consistent with the upregulation of quorum sensing and biofilm formation genes, together suggesting a “final dash” to the coral surface enabled by enhanced swimming speed, followed by a rapid transition to a non-motile, coral surface-associated lifestyle. The upregulation of metabolism, growth, and antibiotic resistance genes suggests that the pathogen takes advantage of mucus as an energy source, and prepares to colonize the coral surface and compete with commensal bacteria. The upregulation of host damage genes and secretion systems responsible for toxin export suggests preparation for the infection process itself. Precise temporal control of pathogenesis is a hallmark of Vibrio pathogens [ 86 , 87 ], which are capable of rapidly modulating their lifestyle between free-swimming and biofilm phases in response to their environment, in particular temperature changes [ 34 , 86 , 88 ]. The frequency of acute temperature-rise in reef waters is increasing [ 89 ], giving additional opportunities for temperature-dependent bacterial pathogens, such as V. coralliilyticus , to infect corals [ 90 , 91 ]. In this context, understanding the mechanisms underlying the earliest stages of bacterial infections is critical in anticipating future disease outbreaks and curbing coral mortality to protect the ecosystems that they support. Fig. 5 Putative infection timeline of V. coralliilyticus . Our results suggest that exposure to coral mucus triggers a suite of behavioral and transcriptional responses in V. coralliilyticus leading up to infection. Within 2 min of coral mucus exposure, V. coralliilyticus induces strong chemokinesis, which, coupled with chemotaxis [ 9 , 10 ], allows the pathogens to reach the coral surface faster. Also early upon coral mucus exposure, upregulation of genes for metabolism of mucus components, biofilm formation, quorum sensing, and antibiotic resistance, and downregulation of flagella- and chemotaxis-related genes, enable host colonization and competition with commensal bacteria. Toxin genes (zinc metalloproteases and hemolysins; yellow stars) and secretion system genes are upregulated in coral mucus, which may lead to host tissue and symbiont damage. Solid arrows indicate bacterial responses for which we have direct observational evidence; dotted arrows indicate hypothesized phenomena based on our RNA-seq data. Figure adapted from Garren et al., 2014 [ 9 ]."
} | 4,270 |
35851269 | PMC9530717 | pmc | 2,997 | {
"abstract": "Methane metabolism in wetlands involves diverse groups of bacteria and archaea, which are responsible for the biological decomposition of organic matter under certain anoxic conditions. Recent advances in environmental omics revealed the phylogenetic diversity of novel microbial lineages, which have not been previously placed in the traditional tree of life. The present study aimed to verify the key players in methane production, either well-known archaeal members or recently identified lineages, in peat soils collected from wetland areas in Japan. Based on an analysis of microbial communities using 16S rRNA gene sequencing and the molecular cloning of the functional gene, mcrA , a marker gene for methanogenesis, methanogenic archaea belonging to Methanomicrobiales , Methanosarcinales , Methanobacteriales , and Methanomassiliicoccales were detected in anoxic peat soils, suggesting the potential of CH 4 production in this natural wetland. “ Candidatus Bathyarchaeia”, archaea with vast metabolic capabilities that is widespread in anoxic environments, was abundant in subsurface peat soils (up to 96% of the archaeal community) based on microbial gene quantification by qPCR. These results emphasize the importance of discovering archaea members outside of traditional methanogenic lineages that may have significant functions in the wetland biogeochemical cycle.",
"conclusion": "Conclusions In the present study, cultivation-independent molecular analyses based on the 16S rRNA gene amplicon and functional mcrA gene were used to evaluate key microbial groups and their potential activities in metabolic methane production. Members of well-known methanogenic archaea were detected, which corresponded with the detection of the mcrA gene in anoxic subsurface peats. Members of Ca. Bathyarchaeia, the uncultivated archaea that are considered to play a role in biogeochemical cycles, were abundant in the present study. The results obtained prompt the further development of culturing innovations (culture-based experiments) and complete genomic characterization, which will be useful for providing comprehensive metabolic insights into Ca. Bathyarchaeia.",
"discussion": "Results and Discussion Surface water chemistry Water collected at the sampling site was used to assess the ORP, pH, and ion content. Based on the pH value, peat soil was mildly acidic (pH 6.45). Furthermore, according to data from samples collected during the rainy season (a period between June and July with a high average precipitation of 250–300 mm, according to the Japan Meteorological Agency) and the value obtained from previous field measurements (pH 4.78) on 28 November 2019, soil pH may have been slightly higher than expected due to the dilution effect of rainfall. The surface water redox potential was 204 mV, suggesting oxidizing conditions at the soil surface. The ion composition analysis revealed that sulfate (307 μM), nitrate (200 μM), and Ca 2+ (238 μM) were major ions found in surface water. Other minor cations and anions detected in surface water are listed in Table 1 . Calcium and 3 other base cations (Na + , Mg 2+ , and K + ) are important components that generally form the majority of cation groups found in peat surface water ( Bourbonniere, 2009 ). The level of calcium ions in the present study was slightly higher than the maximum concentrations in peat bogs in Canada and northern USA (170 μM), northern and central Europe (125 μM), and in subtropical peatlands in central China (73 μM). The concentrations of other base cations in our analysis were in the range measured in surface water from northern hemisphere bogs ( Bourbonniere, 2009 ). Variations in major cations may be dependent on mineral rock fragments and the ion exchange capacity of Sphagnum plants ( Verry, 1975 ; Sjörs and Gunnarsson, 2002 ). The presence of nitrate in surface peat water may be attributed to microbial nitrification under an elevated surface water temperature and higher pH ( Freeman et al. , 1993 ; Whitfield et al. , 2010 ). Nitrate concentrations in surface water from Bogatsuru were higher than those reported from northern and central Europe peat bogs (maximum concentration of 39 μM) ( Bourbonniere, 2009 ). However, the contribution of and variations in nitrate contents in the present study warrant further study to establish whether biologically relevant or anthropogenic disturbances occurred. It is important to note that this research has been interpreted only from the surface water chemistry profile (at a time point); therefore, an analysis of the physicochemical characteristics of subsurface waters will provide insights into more environmental features than the present limited information. Microbial community in the surface layer (BO10) The microbial community composition of peat soils was analyzed using the 16S rRNA gene amplicon and next-generation sequencing platform. 16S rRNA gene sequencing lacks the representation of actual microbial abundance in samples due to the limitation of PCR amplification and sequencing ( e.g. , the method of DNA extraction and purification, primer selection, and errors from sequencing technology) ( Schloss et al. , 2011 ). In the present study, a total of 31,794 microbial sequences were obtained from peat soil samples (8,237 reads from BO10, 10,702 reads from BO45, and 12,855 reads from BO90) with an average length of 464 bp. The taxonomic classification and relative abundance of microorganisms are summarized in Fig. 1 . At the domain level, bacterial sequences were dominant at all depths. Taxonomic classification at the phylum level revealed that members of the phyla, Proteobacteria , Acidobacteriota , Planctomycetota , and Cyanobacteria , were dominant in surface peat soil ( Fig. 1 B). The water analysis revealed that a higher nitrate content (200 μM) was detected than the analytical range of surface peat waters (0.3–39 μM) in northern peatlands ( Bourbonniere, 2009 ), indicating the availability of nitrogen-transforming reactions in the Bogatsuru habitat. Microbial nitrogen-transformation pathways ( e.g. , nitrogen fixation, nitrification, and denitrification) involve diverse groups of microorganisms. Based on the microbial community profile and taxonomic classification, several groups of bacteria associated with nitrogen fixation and transformation were identified in this study. According to taxonomic characterization at lower levels, the bacterial sequences of the orders Rhizobiales and Planctomycetales were detected at the highest proportion in surface peat soil (BO10). These bacterial groups have been reported to play a functional role in nitrogen cycling. Rhizobiales ( Bradyrhizobium spp.) are nitrogen-fixing bacteria that generally live symbiotically with plant legumes ( Kuypers et al. , 2018 ). Some members of Planctomycetota oxidize ammonium anaerobically using nitrite as an electron acceptor ( Fuerst, 2005 ; Fuerst and Sagulenko, 2011 ). Cyanobacteria , which were only detected in BO10, have also been shown to assimilate nitrogen for growth through nitrogenase catalysis ( Berman-Frank et al. , 2003 ; Bothe et al. , 2010 ). Microbial community in middle and deep layers (BO45 and BO90) Based on the results of the soil gas analysis, CH 4 was only detected in BO45 (0.27±0.14 mM). By referring to 16S rRNA gene amplicon sequencing, archaeal sequences were detected in BO45 and BO90 ( Fig. 1 A). The community compositions of peat soil at depths of 45 and 90 cm were slightly different from those of surface samples, mainly Anaerolineales , Nitrospirales , Syntrophobacterales , and the candidate order GIF9, which were detected at a higher proportion (>500 sequence read counts) than in surface soil. Bacteria from the genus Nitrospira (belonging to the phylum Nitrospirota ) were recently discovered to undergo complete ammonia oxidation (comammox) ( Daims et al. , 2015 ; van Kessel et al. , 2015 ). Furthermore, diverse clades of comammox Nitrospira have been detected in the sediment along estuarine tidal flat wetlands ( Sun et al. , 2020 ). Nitrospira members were dominant in the low dissolved oxygen reactor of a wastewater treatment system ( Roots et al. , 2019 ) and were proven to oxidize formate using nitrate as an electron acceptor under anoxic conditions ( Koch et al. , 2015 ). Other major bacterial sequences detected in subsurface peats were affiliated to the phylum Chloroflexota ( Fig. 1 B). Members of Chloroflexota have been identified in sediments and are suggested to be involved in the subsurface carbon cycle ( Blazejak and Schippers, 2010 ; Kadnikov et al. , 2012 ). The metabolic lifestyles of Chloroflexota in sediments retrieved from genomic analyses include sugar and amino acid degradation, acetate utilization, and nitrate respiration and nitrification ( Hug et al. , 2013 ). The sequences of well-known methanogenic archaea in the orders Methanomicrobiales and Methanosarcinales were detected at a depth of 45 cm. Thermoplasmatales were also identified in subsurface peats (BO45 and BO90). In addition, Ca. Bathyarchaeia sequences were more abundant in BO45 and BO90 than other archaeal sequences. Other archaeal groups, including Iainarchaeota , Hadarchaeia , Nitrososphaeria , and Nanoarchaeota , were also detected in subsurface peat soils. Microbial gene abundance The distribution of microbial gene numbers along the vertical soil depth ( Table 2 ) was quantified by qPCR using specific primer sets. Prokaryotic 16S rRNA gene numbers ranged between 2.56×10 8 and 8.73×10 8 genes g –1 peat. The archaeal 16S rRNA gene number was lower than those of the prokaryotes at all soil depths, ranging between 2.81×10 6 and 2.58×10 7 genes g –1 peat. The abundance of archaeal genes was the highest in the middle depth layer. Furthermore, the ratio of archaeal 16S rRNA genes to prokaryotic 16S rRNA genes ranged between 0.4% and 2.9%, suggesting the low abundance of archaea at all depths. The abundance of the mcrA gene was interpreted based on a qPCR data analysis and gel-electrophoresis confirmation ( Fig. S2 ), and mcrA genes were only detected in subsurface soils (45 and 90 cm) with the highest copy number of 3.91×10 6 genes g –1 peat at a depth of 45 cm. mcrA gene numbers were higher than those previously observed at subsurface peats (ranging from 10 4 –10 5 genes g –1 peat) in Japanese wetlands ( Akiyama et al. , 2011 ). If we assume that archaea and methanogens carry one copy of the 16S rRNA and mcrA genes, respectively ( Kembel et al. , 2012 ; Louca et al. , 2018 ), MCR-containing archaea in the present study may have accounted for approximately 15% of the archaeal sequences at a depth of 45 cm. Furthermore, the high copy number of the mcrA gene corresponded with the detection of CH 4 at a depth of 45 cm from the soil gas analysis, suggesting the production potential of CH 4 from methanogenic archaea. Ca. Bathyarchaeia 16S rRNA genes were detected at depths of 45 and 90 cm using the modified primers, with copy numbers of 4.45×10 6 and 4.59×10 6 genes g –1 peat, respectively. If we assume that the copy number of the 16S rRNA gene of Ca. Bathyarchaeia is equal to 1, the ratios of Ca. Bathyarchaeia to archaea in BO45 and BO90 were 18 and 97%, respectively, indicating the distribution of Ca. Bathyarchaeia in the archaeal community in the Bogatsuru wetland. Microbial gene abundance is summarized in Table 2 . Phylogenetic composition of the mcrA gene The phylogenetic diversity of mcrA was assessed using molecular cloning. The taxonomic classification of mcrA sequences is shown in Fig. 2 A, while the phylogenetic tree is shown in Fig. 2 B. A total of 23 and 22 clones were obtained in the mcrA clone library of BO45 and BO90 samples, respectively. Based on the results obtained, most of the mcrA nucleotide sequences in the BO45 library were phylogenetically classified into Methanomicrobiales , which accounted for approximately 78% of all mcrA clone sequences. Methanobateriaceae \n mcrA was dominant in BO90, comprising approximately 59% of all sequences. The methanogenic lineage in Methanosaetaceae was detected as a minority group at depths of 45 and 90 cm. This result corresponded with previous finding showing the dominance of Methanomicrobiales in wetlands in Hokkaido, followed by a small proportion of Methanosaetaceae ( Narihiro et al. , 2011 ). Methanomassiliicoccales accounted for 9% of all mcrA clones in the deepest peat soil (BO90). Phylogenetic composition and metabolic potential of Ca. Bathyarchaeia Ca. Bathyarchaeia sequences obtained from the 16S rRNA gene amplicon analysis were aligned and affiliated with the phylogenetic tree of archaea ( Fig. 3 ). Based on the phylogenetic analysis, Ca. Bathyarchaeia detected in the present study belonged to various subgroups (Subgroup-5a, 5b, 5bb, 7, 9, 13, 17, and 18) ( Zhou et al. , 2018 ), indicating the diversity of this archaeal lineage in the terrestrial wetland ecosystem. Ca. Bathyarchaeia sequences have previously been detected in more than half of the archaeal populations in various peatlands ( Rooney-Varga et al. , 2007 ; Hawkins et al. , 2014 ; Xiang et al. , 2017 ). Nevertheless, their ecological functions in peatland ecosystems have yet to be confirmed. Based on physiological and genomic characterizations, members of Ca. Bathyarchaeia possess diverse trophic and metabolic properties, including methanogenesis, and have been reported to utilize proteins, aromatic compounds, plant-derived carbohydrates, and lignin ( Lloyd et al. , 2013 ; Meng et al. , 2014 ; Lazar et al. , 2016 ; Yu et al. , 2018 ). However, the Ca. Bathyarchaeia sequences retrieved herein deviated from the recognized methane-metabolizing groups BA1 (subgroup-3) and BA2 (subgroup-8), which have been proposed to encode methyl coenzyme M reductase ( Evans et al. , 2015 ). Notably, qPCR quantification showed that the proportion of Ca. Bathyarchaeia was high in archaea, which positively encourages the need for further studies. Future research that focuses on the characterization of metabolic capability, particularly the confirmation of MCR-containing Ca. Bathyarchaeia, is warranted. Methanogenic potential and biogeochemical interaction in wetland soils The mcrA gene phylotype revealed that members of Methanomicrobiales , which are well-recognized hydrogenotrophic methanogens that generally reduce CO 2 to methane with H 2 and/or formate as the electron donor, were mainly detected in the present study. They have been found in diverse anaerobic natural habitats, such as freshwater and marine sediments, rice paddies, animal digestive tracts, and wetlands ( Jabłoński et al. , 2015 ). Known Methanomicrobiales representatives that have been successfully isolated from peat include Methanosphaerula palustris E1-9c ( Cadillo-Quiroz et al. , 2008 ; Cadillo-Quiroz et al. , 2009 ) and Methanoregula boonei 6A8 ( Bräuer et al. , 2006 ). The minor groups of methanogenic archaea present in peat soils were Methanobacteriaceae and Methanoseataceae . Methanobacteriaceae also perform CO 2 reduction coupled with H 2 oxidation for methanogenesis. However, Methanoseataceae are considered to be acetate utilizers, with acetate cleaved to form methane and carbon dioxide as terminal products through acetoclastic methanogenesis. Culture representatives of Methanobacteriales and Methanosarcinales are Methanobacterium paludism SWAN1 ( Cadillo-Quiroz et al. , 2014 ) and Methanothrix thermoacetophila PT ( Kamagata et al. , 1992 ), respectively. The order Methanomassiliicoccales , which was detected in the deepest soil in the present study, is an obligate methylotroph that produces methane from the reduction of methanol, methyl sulfide, and methylated amines ( Paul et al. , 2012 ; Lang et al. , 2015 ). The representative of this order was initially isolated from human feces and called Methanomassiliicoccus luminyensis B10 ( Dridi et al. , 2012 ). Thermoplasmatales , which are members of the class Thermoplasmata , were also detected based on 16S rRNA gene amplicon sequencing. The detection of methanogenic archaea and a functional gene for methanogenesis in the present study suggested the potential of methane production using peat soils, either via hydrogenotrophic or acetoclastic methanogenesis. In contrast, net methane emission from marine sediments and terrestrial environments may be neutralized by anaerobic methanotrophic archaea (ANME) via the anaerobic oxidation of methane prior to its escape into the atmosphere ( Knittel and Boetius, 2009 ). Based on 16S rRNA gene reads, no sequences of ANME were detected in the present study. Another group of microbes may also utilize methane aerobically, namely, aerobic methanotrophic bacteria ( Dedysh and Knief, 2018 ). These organisms use methane monooxygenase to convert methane to methanol. Sequences of aerobic methanotrophic bacteria of the phylum Verrucomicrobiota , Methylacidiphilum , were detected at subsurface peats (BO45 and BO90). Genomic analyses have shown that representative strains of Methylacidiphilum possess monooxygenase, similar to methanotrophs in the phylum Proteobacteria , which demonstrates a capability for aerobic methane oxidation ( Dunfield et al. , 2007 ; Op den Camp et al. , 2009 ). In anaerobic environments, methanogenic archaea compete with sulfate-reducing bacteria for available common substrates ( Muyzer and Stams, 2008 ). Therefore, the presence of sulfate in such an environment is a key factor in trophic competition. In the present study, sulfate was detected based on the geochemistry of surface water ( Table 1 ). The results of 16S rRNA gene amplicon sequencing revealed that the sulfate-reducing bacteria, Desulfobacca , affiliated with Desulfobacterota , were dominant in anoxic subsurface peats (BO45 and BO90), as depicted by the constitution of a relative high ratio of Desulfobacterota sequences in Fig. S1 . Desulfobacca accounted for approximately 46 and 49% of all Desulfobacterota sequence reads in BO45 and BO90, respectively. Bacterial isolates belonging to the genus Desulfobacca have been isolated from granular sludge ( Göker et al. , 2011 ). Furthermore, a physiological analysis revealed that they utilized acetate as the sole carbon source and sulfate as an electron acceptor. Competition between sulfate reducers and acetoclastic methanogens for acetate utilization may occur in Bogatsuru wetlands because Desulfobacca and Methanoseataceae were detected in the present study. Combined environmental omics approaches have been extensively proven as an advantageous strategy for identifying unknown microbial diversity, mainly in the context of examining key players in biogeochemical processes. Since methanogenic archaea in peatlands are difficult to culture due to, for example, their specific optimal growth requirements, potential syntrophic bacterial partners, environmental conditions, and generation time ( Wolfe, 2011 ; Khelaifia et al. , 2013 ; Narihiro and Kamagata, 2013 ; Carson et al. , 2019 ), the recovery of genomic data may reveal taxonomic profiles and imply their functional properties. Further studies need to focus on the cultivation and isolation of uncultured methanogens and microbial syntrophs, which will contribute to our understanding of and reveal important information on microbial physiology and their functions that may provide feedback regarding the global methane and carbon cycle."
} | 4,911 |
25621168 | null | s2 | 2,998 | {
"abstract": "Interaction with water causes shrinkage and significant changes in the structure of spider dragline silks, which has been referred to as supercontraction in the literature. Preferred orientation or alignment of protein chains with respect to the fiber axis is extensively changed during this supercontraction process. Synchrotron x-ray micro-fiber diffraction experiments have been performed on "
} | 98 |
21980274 | PMC3182867 | pmc | 2,999 | {
"abstract": "Collective navigation and swarming have been studied in animal groups, such as fish schools, bird flocks, bacteria, and slime molds. Computer modeling has shown that collective behavior of simple agents can result from simple interactions between the agents, which include short range repulsion, intermediate range alignment, and long range attraction. Here we study collective navigation of bacteria-inspired smart agents in complex terrains, with adaptive interactions that depend on performance. More specifically, each agent adjusts its interactions with the other agents according to its local environment – by decreasing the peers' influence while navigating in a beneficial direction, and increasing it otherwise. We show that inclusion of such performance dependent adaptable interactions significantly improves the collective swarming performance, leading to highly efficient navigation, especially in complex terrains. Notably, to afford such adaptable interactions, each modeled agent requires only simple computational capabilities with short-term memory, which can easily be implemented in simple swarming robots.",
"introduction": "Introduction Many organisms exhibit complex group behavior [1] – [10] , including collective navigation observed in the flight of birds [11] , trail organization in ants [12] , and swarming of locust [13] , fish [14] and bacteria [15] , among others. The aggregation results in highly complex collective behavior, with new functionality and computational ability. Simple interaction models, which describe how each agent acts according to the result of a ‘computation’ it performs on the locations of the other agents, have been used to demonstrate and study the fundamental building blocks of complex group behavior [16] – [24] . In computational models, swarming behavior can arise from simple rules, and in particular demonstrate qualitive (and sometimes quantitative) features of collective behavior observed in nature: Vicsek et al. [16] introduced the ‘self-propelling particles’ (SPP) model, in which the motion of each individual is determined by the mean orientation of its local neighborhood with some noise induced perturbation. The SPP model can exhibit random or coherent motion of group clusters depending on the particle density and on the noise of each individual; high density and low noise results in a coherent group motion. Later derivatives of the SPP model included individual preferential movement directions, collision avoidance, and attraction [17] – [23] . Couzin et al. [17] , [18] studied a model in which the direction of motion of each individual is determined by a set of rules: repulsion (from neighbors who are too close), attraction (to prevent fragmentation), alignment (of velocity directions and speed), and reaction to the environment. A swarm using these interaction rules can come to a collective ‘decision’ about its direction of movement without leadership and a small fraction of individuals ‘in agreement’ are needed for such a cohesive decision to be made. Recently, Torney et al. [25] presented a model in which the individual agents adapt their interactions according to local conditions. A special feature of this model is that it leads to the emergence of collective navigation although each agent does not possess individual navigation capabilities. Even bacteria show remarkably sophisticated collective behaviors. Some bacteria strains can form large colonies with intricate complex architectures, which allows them to expand efficiently by taking advantage of the available resources [26] – [29] . They construct intricate multicellular structures utilized for protection and cooperation of cells [30] – [33] . In addition, bacteria display complicated movement dynamics, in which cells organize into vortices, form traffic lanes, or move collectively in a common direction [34] – [36] . Bacteria swarming behavior in colonies was explained by considering attractive and repulsive forces between colony parts [10] , [28] , [37] , [38] , communication capabilites [39] – [43] , physical interactions between cells, and forces from the environment [44] . Bacteria navigate using chemotaxis, i.e., moving according to gradients in the chemical concentration [45] – [49] . Bacteria are too small to detect the chemical gradients across their body receptors, and thus detect the concentration as they swim, and delay their tumble if the concentration increases. As a result, they make longer excursions towards areas of higher concentration. Each bacterium may only acquire local and partial cues from the environment, but as a group bacteria can navigate through challenging environments. In such cases, the optimal local direction may be completely independent of the global environment. In addition, microorganisms are especially sensitive to noise, due to stochastic variations in their internal mechanisms, sensory system, and the external environment. Information pooling was shown to improve decision making in animal groups [1] , [50] – [52] , such as the accuracy of navigating birds. In addition, it has been shown that schooling can improve the collective ability of groups of chemotactic organisms, such as bacteria, to climb gradients [53] . Interaction between individuals such as repulsion, alignment, and attraction, may exist in bacteria due to the associations between single cells by mechanical and chemical means. Mechanical interactions can result in collision or adhesion of cells. Chemical interactions, by secretion and detection of various diffusible chemicals, can result in repulsion or attraction. In high densities, interactions between elongated cells cause alignment of cell bodies and velocities. Here, motivated by bacteria swarming, we study the collective behavior of agents with self-navigation capabilities (particularly, a tractable variant of chemotaxis) and performance dependent adaptable interactions. Specifically, when the change of chemical concentration is positive, an agent is more likely to continue in its previous direction, thus, decreasing the influence of the other agents, and vice versa. This implies that the interaction network among agents is plastic – similar in spirit to the approach in machine learning and neuroscience [54] – [56] and the recent work by Torney et al. [25] . The current approach enables quanitative comparison between the efficiency of collective navigation in the case of static and adaptable interactions. We found that the adaptable interactions become more important for more complex terrains. We also found that collective navigation of agents with adaptable interactions is more robust to the initial conditions, to the internal noise in the system, and to the values of the interactions.",
"discussion": "Discussion We introduced a collective behavior model of a group of interacting agents, in which each group member senses the environment and adaptively weighs its own evidence and the behavior of its neighbors to navigate in a complex environment. We expanded a model that originated from the self propelling particles model and has been used to describe swarming in many complex systems [17] – [23] , [25] , [53] , [57] . We investigated the navigation capabilities of the swarm in a complex terrain and showed that independent agents create fragmented groups while each agent performs an independent biased random walk towards the target. Interacting agents were far better in finding the target than independent agents, and also demonstrated emergent collective swarming, but were affected strongly by global noise due to positive feedback. Previously, Torney et al. [25] showed that adaptable interactions can lead to the emergence of collective navigation in swarms composed of agents that do not posses navigation capabilities as individuals. Here, we studied collective navigation of agents which do possess navigation capabilities as individuals while focusing on the advantage of performance dependent interactions. When we added a learning mechanism to the network of agent-agent interactions, these swarms had a higher probability of finding the source, and significantly faster. Moreover, performance-dependent adaptive interactions improved the efficiency of the collective navigation beyond that of agents with static interactions, even for an optimal set of static interaction parameters. The adaptable interactions enabled agents to adjust the weight they gave to their neighbors according to local conditions. We used a hard limit weighting, in which agents either followed their neighbors or balanced equally between them and their individual direction, and this was enough to significantly improve the navigation efficiency of swarms in a complex terrain. We note that we did not add memory beyond the measurement of the change in concentration, which already exists in the navigation of an independent agent, or additional computational capabilities to the agents. Using the immediate environment as a teacher the weights of each agent in the network changes dynamically. This gave a form of noise reduction, where the influence of erroneous agents on the system was reduced, and the power of sub-groups changed and resulted in a dynamically shifting leading cluster of agents that comprises of only the most successful performers. We found that the adaptable interactions model is more robust to internal noise and to diversity in the agents' control mechanism parameters. In particular, the model was robust to the radii of interactions. The system of agents with adaptable interactions changed dynamically according to each agent's success and as a result, the system as a whole transforms into a robust yet “plastic” network. Models of swarm intelligence and their analysis have the potential to export ideas and algorithms from nature into novel computational tools, including distributed algorithms for optimization and other complex problems in addition to mechanisms for robotic systems [58] - [61] . The model we studied here can be viewed as a distributed network of sensors, with the capability of having local effects on each other. The problem at hand is a function optimization task, where the function samples contain local and global errors. Each sensor can only sample the function at one position and the next sampling position is in the local proximity of the previous one. We investigated how the local effects or interactions between the sensors affect the function optimization time of the network under different conditions. We found that adaptable interactions benefit the system as a whole in a complex navigation task making it faster to find the target under more diverse conditions than before. In the current study, our swarms constituted of identical individuals with equal measurement capabilities. Natural extension would be investigating the effect of variability, for example in the interaction ranges and noise distributions of agents, on the swarm's collective navigation performance. It is known that many biological mechanisms benefit from variability in the system in the presence of noise [62] - [64] . We expect agent variability to be advantageous for navigation in the case of both spatial and temporal noise. The combination of sensor diversity and adaptable interactions can constitute a solution to navigation in the presence of spatial and temporal noise such as in the case of a time-changing terrain. Bacteria have developed various communication capabilities such as direct and indirect cell-cell physical and chemical interactions, chemical signaling, such as quorum sensing, and chemotaxis signaling [30] , [32] , [36] , [40] , [43] , [65] – [67] . Thus, the communication mechanisms necessary to sustain adaptable interactions already exist in bacteria; in fact, the interaction capabilities found in some strains of social bacteria are far more sophisticated and have yet to be understood [68] . Adaptable interactions, similar to what we have suggested, may be found in other groups of simple organisms such as fish. Moreover, we suggest that performance dependent adaptable interactions exist in more complex networks, such as social networks."
} | 3,052 |
23095941 | PMC3586100 | pmc | 3,000 | {
"abstract": "There is a particularly high interest to derive carotenoids such as β-carotene and lutein from higher plants and algae for the global market. It is well known that β-carotene can be overproduced in the green microalga Dunaliella salina in response to stressful light conditions. However, little is known about the effects of light quality on carotenoid metabolism, e.g., narrow spectrum red light. In this study, we present UPLC-UV-MS data from D. salina consistent with the pathway proposed for carotenoid metabolism in the green microalga Chlamydomonas reinhardtii . We have studied the effect of red light-emitting diode (LED) lighting on growth rate and biomass yield and identified the optimal photon flux for D. salina growth. We found that the major carotenoids changed in parallel to the chlorophyll b content and that red light photon stress alone at high level was not capable of upregulating carotenoid accumulation presumably due to serious photodamage. We have found that combining red LED (75 %) with blue LED (25 %) allowed growth at a higher total photon flux. Additional blue light instead of red light led to increased β-carotene and lutein accumulation, and the application of long-term iterative stress (adaptive laboratory evolution) yielded strains of D. salina with increased accumulation of carotenoids under combined blue and red light. Electronic supplementary material The online version of this article (doi:10.1007/s00253-012-4502-5) contains supplementary material, which is available to authorized users.",
"introduction": "Introduction Microalgae have great potential in many aspects of the conversion of conventional petrol-based manufacturing to bio-based manufacturing: in production of biofuels as well as bio-factories producing valuable pharmaceuticals, food additives, and cosmetics (Cordero et al. 2011 ; Lamers et al. 2008 ; Takaichi 2011 ; Vilchez et al. 2011 ; Wijffels and Barbosa 2010 ). Carotenoids are extremely important for human and animal nutrition, and they are distributed broadly in both phototrophic and non-phototrophic organisms (Takaichi 2011 ). However, humans and animals cannot synthesize necessary carotenoids and must obtain them from their diets (Takaichi 2011 ). Carotenoids can be divided into two groups based on their chemical structure: the carotenes such as β-carotene and the xanthophylls such as lutein. Among important carotenoids for humans, β-carotene is a major source of vitamin A which is necessary for functions of the retina and has an effect on many tissue types (Amengual et al. 2011 ; von Lintig et al. 2005 ) through its action as a regulator of gene expression. In addition, β-carotene helps protect the skin against photoaging by its antioxidant activity (Darvin et al. 2011 ). Lutein and zeaxanthin are also of particular interest for their role in reducing the development and progression of age-related macular degeneration (Carpentier et al. 2009 ; Fernandez-Sevilla et al. 2010 ). In higher plants as well as in green algae, the antenna pigment molecules (Telfer 2002 ) bound to light harvesting (or antenna) complexes in the thylakoid membrane help to harvest light and transfer energy to the reaction center of photosystems, e.g., PS II. The antenna pigments usually consist of carotenoids, chlorophyll b , and chlorophyll a (Jahns and Holzwarth 2012 ; Telfer 2002 ). Chlorophyll a is very different from chlorophyll b in functions since it acts uniquely as primary electron donor in the reaction center of photosystems, though it also helps to transfer energy in the antenna complex (Jahns and Holzwarth 2012 ). Carotenoids such as β-carotene and lutein (Jahns and Holzwarth 2012 ; Telfer 2002 ) play a central role in PS II, harvesting blue light and transferring energy to photosystem reaction centers and protecting the photosynthetic apparatus against photo-oxidative damage by deactivating reactive oxygen species (ROS) and reducing the ROS formation under excess light. To study carotenoid metabolism in the green microalga Dunaliella salina , it will be informative and important to profile all related antenna pigments. The unicellular green microalga D. salina has been useful in studying carotenoid metabolism as it is able to accumulate large amounts of carotenoids (Ye et al. 2008 ). To date, some researchers have addressed the effects of different abiotic environment conditions (Gómez and González 2005 ; Coesel et al. 2008 ; Lamers et al. 2010 ; Ramakrishna et al. 2011 ) on the accumulation of carotenoids in D. salina , and it is widely accepted that light intensity is a key stimulus for β-carotene overproduction in D. salina (Lamers et al. 2010 ). With regard to the regulation of genes involved in carotenoid biosynthesis in D. salina , it has also been suggested that Lcy-β steady-state transcript levels were upregulated in response to all stress conditions tested, e.g., salt, light, and nutrient depletion (Ramos et al. 2008 ). However, little is known about the light regulation underlying carotenoid metabolism, and it remains unclear whether accumulation of β-carotene and major carotenoids is related to light quality, although an increase in β-carotene accumulation has been observed in Dunaliella cultivated under white light combined with UV-A, compared with white light alone (Salguero et al. 2005 ; Lamers et al. 2008 ; Mogedas et al. 2009 ). The progressive light-emitting diode (LED) technology that is currently emerging has a high conversion efficiency from electricity to light while providing narrow spectra of wavelengths, and the application of LED in photobioreactors (PBRs) marks a great advance over existing indoor agricultural lighting (Yeh and Chung 2009 ). In addition, LED illumination induced light stress on Dunaliella cells at lower incident photon fluxes, e.g., 170 and 255 μE/m 2 /s, while as high as 1000 μE/m 2 /s photon flux is usually provided for driving Dunaliella cells to overproduce β-carotene by conventional lights such as fluorescent lamp and high-pressure sodium lamp (Lamers et al. 2010 ). In this study, the effects of nearly monochromatic light (20 nm bandwidth at half peak height), e.g., red light with a narrow spectrum as well as combined blue and red light, on D. salina were evaluated, with regard to both growth rate and the accumulation and composition of major carotenoids. Adaptive laboratory evolution (ALE) has been widely utilized as a tool for developing new biological and phenotypic functions and exploring strain improvement in synthetic biology for bacteria (Palsson 2011 ). Specifically, ALE has been utilized to evolve strains to better adapt to defined conditions, e.g., a certain carbon source, energy source, or to cope with environmental stress. However, ALE is still a novel solution for improving strain performance in microalgal biotechnology (Fu et al. 2012 ). With the aid of redesigned LED-based PBRs combining blue LED with red LED, we have set out to use ALE to develop D. salina strains with increased yields of carotenoids.",
"discussion": "Discussion Global climate change has called for immediate reduction of CO 2 emission and development of sustainable manufacturing. This study has provided data relating to such an issue, the photosynthesis-based production of valuable compounds such as β-carotene and lutein using microalgae. The use of well-designed LED lighting for D. salina illustrates the potential for enhancing sustainable production of carotenoid products such as β-carotene and lutein efficiently by microalgal biotechnology (Lamers et al. 2008 ; Ribeiro et al. 2011 ). Major results of this study are summarized in Fig. 5 . Firstly, the metabolic profile of major pigments was determined by UPLC-UV-MS and found to be consistent with a model of carotenoid metabolism proposed for green algae (Chang et al. 2011 ). Then, we studied the effect of red LED lighting on the average growth rate and photon based biomass yield. The photon flux of 128 μE/m 2 /s red light (660 nm) was determined optimal for efficient growth of D. salina . The major antenna pigments in D. salina were analyzed under different red LED lighting conditions, and it was inferred from the results that the stress due to supra-optimal photon flux of narrow bandwidth red light did not yield higher carotenoid levels in D. salina presumably due to significant photodamage. The major carotenoids increased relative to chlorophyll b at high light intensities. Finally, using redesigned lighting combining red LED (75 %) arrays with blue LED (25 %) arrays, we were able to utilize adaptive laboratory evolution (ALE) to develop strains yielding higher levels of carotenoids. Fig. 5 Schematic pathways to developing D. salina with increased yields of carotenoids. BP biomass productivity, CC carotenoids content \n In plant cells, mechanisms regulating carotenoid biosynthesis and accumulation are complex (Lu and Li 2008 ). It has been suggested that light plays a key role in the biosynthesis of carotenoids through light signal sensing and downstream regulation (Lamers et al. 2008 ). However, it is difficult to study light effects on Dunaliella cells in depth without high-quality lighting. With well-designed LED lighting, it has become possible to investigate effects of nearly monochromatic light, e.g., red light and blue light with narrow spectra, on both Dunaliella cell growth and carotenoid metabolism. Our results demonstrate that Dunaliella cells (UTEX LB #200) are sensitive to a red LED photon flux of 170 μE/m 2 /s and higher applied in this study and fail to acclimate to such environments. Further analysis of the ratio of zeaxanthin to the VAZ pool (Fig. S3 in the ESM) is indicative of stress responses to high photon flux supporting the observation that with a basic flux of 128 μE/m 2 /s red LED light, additional red LED light is more stressful than additional blue LED light. The light signal transduction of blue light may be different from that of red light since plants usually have different photoreceptors/domains, e.g., blue light-regulated and red light-regulated, although these photoreceptors have both overlapping and distinct functions (Chory 2010 ). It has also been found that blue LED light enhanced growth of the green microalga Haematococcus pluvialis in the early exponential phase but caused the suppression of growth later in batch culture, while accumulation of astaxanthin was significantly enhanced (Katsuda et al. 2004 ). In addition, it has been reported that blue light stimulates carotenoid synthesis in non-photosynthetic bacteria such as Myxococcus xanthus (Ruiz-Vázquez et al. 1993 ). A possible mechanism may be that blue light signal transduction in D. salina involving major carotenoids (Chory 2010 ; Jahns and Holzwarth 2012 ) is separate from red light at a high level of total light intensity. It has been shown that adaptation to environmental factors varies along native clines, and it has been suggested that changes in photoreceptor family members are important determinants in adaptation to the natural variation of light sensitivity (Chory 2010 ). The stable difference in the adapted strain (HI 001) could be either a consequence of accumulative mutations or due to selection of variants already found in the original culture (UTEX LB #200). Further study, e.g., by reference genome sequencing, needs to be performed to decipher the nature of the differences in the Dunaliella strains (both UTEX LB #200 and HI 001) once the genome sequences of D. salina strains UTEX 1644 and CCAP 19/18 are published by the US DOE Joint Genome Institute ( http://genome.jgi.doe.gov/genome-projects/ ). Biosynthesis of carotenoids is complex and coordinated with the biogenesis of chlorophylls and proteins of the photosynthetic apparatus (Bohne and Linden 2002 ) as well as electron transport (Cardol et al. 2011 ). The content of the major carotenoids appears to be regulated in concert with the chlorophyll b content in D. salina cells (Figs. 3c and 4 ). It is possible that carotenoid metabolism is regulated along with chlorophyll b through the geranyl geranyl diphosphate pathway (KEGG database) as summarized in Fig. S4 in the ESM. Although supra-optimal irradiation with red light did not increase carotenoids but seriously inhibited growth, we found that adding excess blue light, and applying ALE on the contrary, led to increased β-carotene and lutein yields. Our results show that well-designed ALE is an effective way to increase sustained productivity in contrast to established methods where carotenoid accumulation in D. salina is usually achieved with low biomass productivity and cell density. We have shown that an efficient culture system with increased light energy efficiency and economy of operation can be developed using innovation in lighting technology in combination with genetically based methods such as ALE for strain development. In conclusion, light quality is critical for both D. salina growth and carotenoid accumulation. ALE combined with redesigned LED lighting has allowed a substantial increase in growth yield per photon flux and in the level of sustainable production of β-carotene and lutein (Fig. 4 ). These results are also a demonstration of the technical feasibility of LED-based PBRs for direct conversion of CO 2 to valuable chemicals."
} | 3,352 |
37206316 | PMC10189395 | pmc | 3,001 | {
"abstract": "Increasing energy demands and environmental pollution concerns press for sustainable and environmentally friendly technologies. Soil microbial fuel cell (SMFC) technology has great potential for carbon-neutral bioenergy generation and self-powered electrochemical bioremediation. In this study, an in-depth assessment on the effect of several carbon-based cathode materials on the electrochemical performance of SMFCs is provided for the first time. An innovative carbon nanofibers electrode doped with Fe (CNFFe) is used as cathode material in membrane-less SMFCs, and the performance of the resulting device is compared with SMFCs implementing either Pt-doped carbon cloth (PtC), carbon cloth, or graphite felt (GF) as the cathode. Electrochemical analyses are integrated with microbial analyses to assess the impact on both electrogenesis and microbial composition of the anodic and cathodic biofilm. The results show that CNFFe and PtC generate very stable performances, with a peak power density (with respect to the cathode geometric area) of 25.5 and 30.4 mW m −2 , respectively. The best electrochemical performance was obtained with GF, with a peak power density of 87.3 mW m −2 . Taxonomic profiling of the microbial communities revealed differences between anodic and cathodic communities. The anodes were predominantly enriched with Geobacter and Pseudomonas species, while cathodic communities were dominated by hydrogen-producing and hydrogenotrophic bacteria, indicating H 2 cycling as a possible electron transfer mechanism. The presence of nitrate-reducing bacteria, combined with the results of cyclic voltammograms, suggests microbial nitrate reduction occurred on GF cathodes. The results of this study can contribute to the development of effective SMFC design strategies for field implementation.",
"conclusion": "4 Conclusions Understanding the influence of electrode material on the microbial communities involved in the electrogenic processes (and consequently, on the electrochemical performance) is essential to accelerate the translation of SMFCs into practical implementations. This study provides for the first time an in-depth investigation on the effect of different cathode materials on SMFC performance, combining electrochemical analyses with advanced microbial profiling. The use of an innovative CNF–Fe cathode led to an electrochemical performance similar to Pt-doped carbon cloth, thus providing an excellent low-cost alternative. Nonetheless, graphite felt showed better performance than all the other electrodes tested. Not only it exhibited higher voltages, catalytic activity, and lower resistance values, but it also favoured the enrichment of Hydrogenophaga, which potentially led to the improvement of cathodic activity by increasing electron uptake via H 2 evolution and/or facilitating nitrate reduction. On the other hand, graphite felt showed much lower reproducibility in results and higher mass transport losses, which might be related to the lack of an oxygen reduction reaction catalyst. Microbial taxonomic profiling of the cathode biofilms revealed taxa related to oxygen reduction or involved in utilizing alternative electron acceptors other than oxygen, such as nitrate. An uncultured Gammaproteobacterium was the most prevalent taxon on all cathode biofilms tested, indicating its importance in drawing electrons from the cathode. Overall, the results generated with this study can inspire future research on low-cost and high-performing SMFCs for practical applications in energy harvesting and bioremediation.",
"introduction": "1 Introduction Renewable energy is the only answer to balance increasing global energy demands with the net-zero emissions target by 2050. It is unlikely that one single approach could solve such a great challenge and help minimise our dependence on fossil fuels. Therefore, the most promising strategy is to explore the synergistic contribution of several green energy systems. Amongst the different options, microbial electrochemical technologies hold great potential, given their cost-effectiveness and unique ability to degrade organic matter in biological waste with the concomitant production of electrical energy. In this context, soil microbial fuel cell (SMFC) is a particularly attractive carbon-neutral and affordable technology, able to exploit the use of the endogenous microorganisms present in the soil to convert the organic matter in the soil into useful electricity [ 1 ]. The estimated number of bacterial species per gram of soil varies between 2000 and 8.3 million [ 2 , 3 ], and the abundance of organic matter, despite the variation between different soil types, is approximately 100 mg g −1 [ 4 , 5 ]. These features make the soil a very sustainable source of energy. SMFC technology also offers great opportunities as a self-powered and sustainable in situ bioremediation strategy for soils contaminated with metals, hydrocarbons, and pesticides [ 6 ]. Contrary to traditional (i.e., liquid-based) microbial fuel cells, SMFC technology is still in its infancy, with studies mainly focused on assessing its potential for soil bioremediation [ 7 , 8 ]. Nonetheless, SMFCs are characterised by simple, low-cost, and low-maintenance designs, which facilitate translating the technology from the lab to the field [ 9 ]. In microbial fuel cells, the electrode material plays an important role in performance. Current generation in such systems is highly dependent on the cathode's reduction kinetics, and in the case of air-cathode designs, the oxygen reduction reaction (ORR) is one of the major limiting factors in electricity generation due to its high activation energy [ 10 , 11 ]. This high energy is traditionally overcome by using platinum as a catalyst, which is expensive and suffers from biofouling effects [ 12 , 13 ]. Many high-performing transition metal-based catalysts (Fe, Co, Ni, etc.) have been developed to provide a low-cost, durable alternative [ [14] , [15] , [16] , [17] ]. Composite electrode materials, consisting of transition metals supported by a carbon matrix and conductive polymers, have also been proposed [ 11 , 17 ]. To develop cathode materials with high catalytic activity, activated carbon has been functionalised with a metal–organic framework (MOF) [ 18 ], while the performance of air cathodes in microbial fuel cells has been enhanced by functionalising activated carbon with reduced graphene oxide [ 19 ]. Nanostructure-based carbon materials, such as carbon nanotubes, nanofibers, and graphene, are gaining increasing interest due to their large surface area and functionality [ [20] , [21] , [22] ] and, therefore, approaches like doping metal particles on carbon nanostructure are very attractive [ 23 , 24 ]. Carbon nanofibers (CNFs) have been explored for the functionalization of electrodes in microbial fuel cells since they provide very high electrical conductivity, high chemical stability, high porosity, and high specific surface area [ [23] , [24] , [25] , [26] , [27] , [28] ]. Electrospinning of CNFs has proven to be more effective in terms of catalytic activity than physical or chemical modification due to the larger surface area achieved [ 29 , 30 ]. Nonetheless, so far, CNFs have been only tested for liquid-fed microbial fuel cells, and, despite their great potential, only very few studies have exploited their benefits [ [31] , [32] , [33] , [34] , [35] ]. Doping CNFs with Ni nanoparticles can lead to high electrical conductivity due to spontaneous nitrogen defects upon treatment at 900 °C; high power density was achieved using these electrodes as the cathode in liquid microbial fuel cells [ 28 ]. Table S1 provides an overview of the cathode material used for SMFCs. Typically, in SMFCs, unmodified carbon-based materials, such as carbon felt, carbon cloth, and granular activated carbon, are used as a cathode [ [36] , [37] , [38] , [39] , [40] ]. Nonetheless, carbon-based and stainless steel-based electrodes functionalised with Pt (0.1–0.2 mg cm −2 ) [ 41 , 42 ] or alternative catalysts, such as Ni, in combination with platinum, have also been reported [ [43] , [44] , [45] ]. The power densities obtained with different configurations of SMFCs, such as single chamber [ 36 , 44 , 45 ], column type [ 41 ], tubular [ 46 ], or U type [ 42 ], are within the range of 0.85–77.5 mW m −2 [ 36 , 39 , 40 , 42 ], except for few cases where a polymer electrolyte membrane was used in dual chamber microbial fuel cell, in which a value as high as 7.5 mW cm −2 was obtained [ 38 ]. It is expected that the SMFC performance improves with the use of an ORR catalyst at the cathode, however, the power output generated by SMFCs with a Pt-coated cathode (0.1–0.2 mg cm −2 ), was reported to be within the range of 8.8–39 mW m −2 [ 41 , 43 , 46 ], while composite electrodes, fabricated by modifying a stainless-steel matrix with conductive carbon black, can lead to power densites over 250 mW m −2 with no need for an expensive catalyst [ 44 , 45 ]. Although a proper comparison among the different systems is difficult given the great difference in the design parameters used for each study, it is clear that there is great scope to develop robust and efficient cathodes to improve the performance of SMFCs. In this study, we investigate and compare for the first time the performance of four different cathodes in membrane-free air-cathode SMFCs: carbon cloth (SMFC-CC), Pt-doped carbon cloth (SMFC-PtC), graphite felt (SMFC-GF), and Fe-doped CNFs (SMFC-CNFFe). Electrochemical tests over prolonged operating times are complemented by microbial taxonomic analyses to elucidate the influence that the electrode material has on the cathodic and anodic biofilm formation and, consequently, on the overall electrochemical performance, thus guiding the development of highly performing SMFC designs.",
"discussion": "3 Results and discussion 3.1 Electrochemical results The output voltage generated over time by the four types of SMFCs was compared. Fig. 2 a shows the output voltage generated by the SMFCs. The biofilm enrichment and voltage stabilization time may greatly vary in microbial fuel cells, depending on the experimental conditions and design, and it can extend up to several weeks [ 54 ]. Incorporating a corrosion-resistive current collector, such as Ti mesh, with carbon electrodes is a promising strategy to improve electrode conductivity [ 55 ]. In our work, the use of Ti mesh improved the overall performance without directly contributing to the current generation, as demonstrated by a control study in which the cathode consisted of Ti mesh only ( Fig. S4 ). As shown, the use of carbon cloth as the cathode led to the poorest performance; the output voltage generated by SMFC-CC reached a steady state voltage of 0.086 ± 0.03 V, reached after 10 days of operation. The performances of SMFC-CNFFe and SMFC-PtC were very similar, with steady output voltages of 0.241 ± 0.038 and 0.244 ± 0.033 V, respectively, which remained stable over approximately 40 days, suggesting a consistent electrocatalytic activity. A different trend in the evolution of the output voltage over time was observed in SMFC-GF. In this case, the output voltage initially stabilised at 0.25 ± 0.08 V; however, a second exponential increase in voltage was observed, reaching a second steady value of 0.36 ± 0.06 V after about 25 days of operation. This trend has usually been observed by us with graphite felt cathodes, and an in-depth investigation on the related causes has been discussed in Ref. [ 56 ]. Despite the higher output voltage obtained, the performance of SMFC-GF was not stable, and large fluctuations were observed, especially during and after the second exponential increase. As shown in Fig. 2 b, the anode performance was not affected by the cathode material, thus confirming that the difference in the overall performance of the several SMFC designs tested largely depended on the cathode, which is consequently the limiting electrode in the system [ 49 ]. Fig. 2 b shows that after about 10 days of operation, the anode potential vs Ag/AgCl stabilises at −0.365 ± 0.024 V (for SMFC-CNFFe), −0.355 ± 0.043 V (for SMFC-PtC), −0.408 ± 0.014 V (for SMFC-CC), and −0.287 ± 0.11 V (for SMFC-GF). The cathode potentials ( Fig. 2 c) of CNF–Fe and PtC were similar (−0.121 ± 0.031 and −0.123 ± 0.02 V, respectively), while the cathode potentials for carboncloth was −0.321 ± 0.014 V, suggesting a limited cathodic activity. In the case of SMFC-GF, the average cathode potential reflected the overall fuel cell output voltage trend and, accordingly, is characterised by two stages. A first steady potential value of −0.177 ± 0.045 V was reached after 13 days, remaining stable until day 26. Afterwards, the cathode potential increased to 0.077 ± 0.08 V on day 28. The superior performance of graphite felt as cathode over other materials has been previously demonstrated and attributed to the very high specific surface area of the material that favours biofilm growth [ 57 ]. In particular, graphite felt has been shown to facilitate the proliferation of oxygen-reducing microbial communities at the cathode [ 57 ]. On the other hand, the graphite-felt cathode led to the most unstable performance of the SMFC, as shown by the large variability in the measurements. Earlier studies with biocathodes also reported a slow increment in the cell voltage and subsequent slow stabilization, attributed to a slow bacterial growth onto the cathode surface [ [58] , [59] , [60] ]. Fig. 2 SMFC performance. a , Output voltage generated by the four SMFCs. b , Anode potential over time. c , Cathode potential over time. Data represent the average of three replicates, and error bars represent the standard error of the mean. Fig. 2 For all four types of SMFC, after about 50 days, an increase in both the anode and the cathode potential was observed, which led to a decrease in the output voltage generated. This decline could be attributed to a local decrease in the organic matter available to the biofilm on the electrode surface. This decrease was more pronounced in the SMFC-GF. The polarization and power curves in tests performed after 27 days of operation ( Fig. 3 ) confirm the superior performance of the SMFC-GF, with a peak power nearly three times higher than the peak power generated by SMFC-PtC and 3.4 times the power generated by SMFC-CNFFe. In particular, SMFC-GF generated a peak power of 0.43 ± 0.009 mW, corresponding to a power density of 87.3 mW m −2 compared to 0.15 ± 0.025 mW (power density 30.4 mW m −2 ) generated by SMFC-PtC, and 0.127 ± 0.026 mW (power density 25.5 mW m −2 ) generated by SMFC-CNFFe ( Fig. 3 b). The lowest power was generated by SMFC-CC (0.021 mW, with a power density 4.285 mW m −2 ), approximately 95% lower than SMFC-GF, and 86% and 83% lower than the SMFC-PtC and SMFC-CNFFe respectively. The power densities obtained in our study are comparable or, in some cases, higher than those previously reported for SMFCs, which range from 22.9 [ 40 ] to 73.5 mW m −2 [ 44 ]. Our study reports a power density of 87.3 mW m −2 , which is, to the best of our knowledge, the highest reported for a single chamber, individual SMFC operated without the addition of any external substrate. The power density value for the CNFFe electrode is 25.5 mW m −2 , comparable to values obtained with other catalyst-coated electrodes, including Pt, ranging between 0.85 [ 42 ] and 39 mW m −2 [ 46 ]. Fig. 3 Polarization tests after 27 days of operation. a , Polarization curves; b , Power curves. Values are the average of triplicates, and error bars represent the standard error of the mean. Fig. 3 To date, no study provides a direct comparison between the use of PtC and CNF–Fe at the cathode of an SMFC. Nevertheless, previous studies confirmed the benefits of using modified CNF at the cathode of microbial fuel cells with surface power densities of up to 60 mW m −2 [ 30 ] and volumetric power densities of up to 14.4 W m −3 [ 23 ]. Electrospun N–CNF reported a power density of around 125 mW m −2 in a liquid anolyte single-chamber microbial fuel cell [ 28 ]. These studies, however, refer to liquid microbial fuel cells (usually operated with synthetic media), with the electrodes physically separated by an ion exchange membrane, which makes a direct comparison impracticable. The much lower electrical conductivity of the soil used in this study (710 μS cm −1 compared to 20 mS cm −1 for a phosphate buffer-based electrolyte) is a reason for the lower output power generated by SMFC-CNFFe. The EDS analysis ( Fig. S5 ) suggests that the Fe content of the CNF–Fe cathode did not decrease much with over 50 days of continuous operation: 4.74 wt% initially compared to 3.49 wt% at the end of the experiment. Although the accuracy of this analysis is low, and other components from soil were detected, this result still indicates the long-term stability of the CNF–Fe electrode. Future studies may be directed towards understanding the stability of the electrodes under different operating conditions. SEM images of the four cathodes, before and after the operation, show the presence of a biofilm in each case ( Fig. 4 ). Due to the finer structure of CNF and the large specific surface area of the nanostructure used, greater microbial colonization onto the CNF–Fe electrode surface can be observed. Fig. 4 Morphological analysis by SEM of the electrodes before and at the end of the experiment. The figure shows SEM of cathode electrodes before and after the experiment: a , carbon cloth; b , graphite felt; c , Pt-doped carbon cloth; d , CNF–Fe. Acceleration voltage: 7 kV. Magnification: 5000X. Fig. 4 The I–V curve for SMFC-GF indicates a reduced activation loss in the lower current region but an increased concentration loss in the higher current (low resistance) region ( Fig. 3 a). Aerobic metabolism at the cathode decreases the activation energy at higher resistances by enhancing the final reduction reaction [ 61 ]; however, the concentration loss in the lower resistances suggests that the electrons discharged at the anode are not reduced sufficiently fast at the cathode [ 62 ]. It has also been previously reported that microbial fuel cells with carbon cloth cathodes suffer from high activation losses and ohmic losses due to poor microbial activity and the absence of catalysts [ 63 ]. The polarization curves for CNF–Fe and PtC suggest that there is almost a linear drop in the current, mainly attributed to ohmic losses, which is lower than in the case of non-doped electrodes; however, no major concentration loss can be detected. The ohmic loss can be attributed to the restricted ionic conductivity between anodic and cathodic electrolytes due to the low conductivity of the soil [ 64 ]. Although the graphite felt cathode is able to reduce the activation losses due to improved microbial activity and efficient cathode reduction reactions, it suffers from concentration losses that could be alleviated by, for instance, surface functionalization. EIS tests were performed on the SMFCs at the beginning and the end of their operation by configuring the cathode as the working electrode. As shown in Fig. 5 , the ohmic resistance (R Ω ) significantly differs for the four SMFC designs. At the start of the experiment, R Ω values for all the electrodes (solid symbols in Fig. 5 ) varied between 295 and 409 Ω, with the lowest value (295 Ω) observed for SMFC-PtC and the highest (409 Ω) value observed for SMFC-CC. Values of different circuit elements are reported in Table 1 . Since the SMFCs were operated with the same electrolyte (soil) and moisture content, the observed differences can be attributed primarily to the different electrode materials used for the cathodes. At the end of the experiment, R Ω values significantly decreased for all the SMFC systems tested ( Table 1 ). This reduction indicates the improved conductivity of the electrodes and soil, likely caused by the formation of an electroactive biofilm on the surface of both the anode and cathode [ 65 ]. The equivalent circuit model shows that R2 (R ct ), associated with the diameter of the first semicircle, varied between different electrodes at day 0 and was comparatively higher for the catalyst-coated electrodes. We assume that the initial high charge transfer resistance for PtC in this study may be due to operational fluctuations and poor connectivity. After 56 days of operation, the decrease in charge transfer resistance decreased in the following order: PtC (14.3 Ω) < CNF–Fe (20.6 Ω) < Graphite felt (27.7 Ω) < Carbon cloth (33.3 Ω), suggesting that the catalyst-coated electrodes established better connectivity and exhibited better electrocatalytic activity since the biofilm formation decreases the charge transfer resistance [ [65] , [66] , [67] ]. Semicircles in the high-frequency region were followed by near linear segments in the low-frequency region due to diffusion resistance, signifying the diffusion process of oxygen in the electrode-electrolyte interface (R3) [ 57 ]. After 56 days of operation, the total impedance decreased for all the electrodes ( Fig. 4 ). Moreover, after this time, the effect of diffusion resistance, which is defined by the linear portion in the low-frequency region [ 30 ], was no longer observed for SMFC-CNF-Fe and SMFC-PtC (and reduced for SMFC-GF), and instead, closed depressed semicircles were observed for these SMFCs. This result suggests better oxygen diffusion at the electrode-soil interface. Overall, the observed decrease in the total resistance suggests a better charge transfer efficiency and ORR rates due to the formation of an electroactive biofilm onto the surface of the cathode. The graphite-felt cathode led to the lowest total resistance, in agreement with earlier findings [ 30 ]. Fig. 5 Nyquist plots of the SMFCs at the start (day 0) and the end (day 54) of the experiment. Solid symbols correspond to data at the start of the experiment, and empty symbols to data at the end of the experiment. Solid lines show the fitting of the equivalent circuit model (ECM) to the data. The inset figure shows an expanded graph for the high-frequency region of day 0. Fig. 5 Table 1 Values of different circuit elements after fitting the ECM to the Nyquist plots. Table 1 Cathode material R1 (Ω) R2 (Ω) R3 (Ω) Day 0 Day 54 Day 0 Day 54 Day 0 Day 54 Carbon cloth 409 111.7 45 33.3 - - Graphite felt 347.6 92.16 77 27.7 - - CNF–Fe 310.1 223.8 89.5 20.6 - 792 Pt-doped carbon cloth 295.6 87.3 196 14.3 - 198 The cyclic voltammetry studies reveal interesting findings about the electrochemical behaviour of the four cathodes tested ( Fig. 6 ). For all four SMFCs, the redox currents obtained at day 0 are very low ( Fig. 6 b). SMFC-GF showed a reduction current of −3 × 10 −5 A and an oxidation current lower than 1 × 10 −5 A, while SMFC-CNFFe was characterised by a reduction current of approximately −6 × 10 −5 A and an oxidation current of about 1.3 × 10 −5 A. The capacitive behaviour of the carbon cloth electrode is indicative of the lack of redox reactions mediated by microbial species and is in line with previously reported results [ 68 ]. Fig. 6 Cyclic voltammograms of the four cathodes. a , Comparing results at the start (day 0) and the end (day 54) of the experiment; b , Zoom in on day 0. Fig. 6 CV tests at the end of the experiment (day 54) revealed higher peak current values ( Fig. 6 a), which increased by a factor of 10 2 . This result confirms that the formation of a cathodic biofilm enhanced the charge transfer kinetics. Previous studies also showed improved catalytic current values after the formation of an electroactive biofilm in biocathodes [ 57 ]. SMFC-CC showed the poorest performance, with a reduction current of around −1 × 10 −3 A and an oxidation current of around 0, followed by SMFC-CNFFe (approximately 1.5 × 10 −3 to −3 × 10 −3 A). On the other hand, SMFC-PtC and SMFC-GF showed the best performance, with an observed oxidation current between 2 × 10 −3 and 4 × 10 −3 A and a reduction current between −7 × 10 −3 and 8 × 10 −3 A. These results suggest that, similarly to PtC, the electroactive cathodic biofilm that develops onto the graphite felt plays an important role in ORR reactions and electron transfer processes. Moreover, for the graphite felt electrodes, a pair of quasi-reversible redox peaks at around +0.41 and −0.1 V were observed, suggesting a redox reaction on the electrode. The higher surface area of graphite felt enables a better biofilm attachment, resulting in higher ORR activity. The appearance of a pair of distinct redox peaks at around +0.2 and −0.1 V, with CV tests carried out in both air and nitrogen-saturated catholyte in microbial fuel cells with a graphite felt cathode, have been previously associated with the presence of a cathodic biofilm [ 57 ]. 3.2 Taxonomic analysis of microbial communities The cathode material may affect the composition of the microbial biofilms colonizing the surface of both the anode and cathode. To evaluate this effect, taxonomic analysis was performed by amplicon sequencing of the 16S rRNA genes of the bacteria in the different communities. Samples were taken from the cathode and anode of each type of SMFC and the soil in the vicinity of the cathode. Since SMFC-GF outperformed the SMFCs with other cathode materials, a duplicate SMFC-GF cathode was also sequenced to evaluate this observation. Analysis of the amplicon data revealed the microbial taxa present at each sampling location ( Fig. 7 ). For each SMFC type tested in the anodic biofilm, the ASVs with the highest relative abundance were members of the family Geobacteraceae (range 5.7–36.2%), followed by the genera Pseudomonas (6.2–26.1%), Geobacter (5.8–11.0%), and Bacteroides (2.1–23.8%); the family Xanthobacteraceae (5.5–8.8%); the genera Udaeobacter (3.1–5.8%) and Faecalibacterium (0.0–9.8%); WCHB1-31 from the phylum Bacteriodota (1.6–5.5%), and Citrifermentans (1.5–4.8%); and an unidentified Gammaproteobacteria (1.6–4.0%). Fig. 7 Heatmap showing the relative abundances of amplicon sequence variants (ASV) detected in each sample, grouped by sample type (anodes, cathodes, and soils in the vicinity of the cathode). Soil T0 refers to the soil used for inoculation of the SMFCs. ASVs with relative abundance >2% present in at least one sample are shown at the lowest taxonomic rank assigned for each ASV. Fig. 7 The cathode biofilms were dominated by the ASV identified as the unknown Gammaproteobacterium also observed on the anodes (10.9–21.6%), followed by members of the genus Methylobacterium / Methylorubrum (1.5–16.9%); the family Xanthobacteraceae (5.2–11.5%); the genera Hydrogenophaga (2.8–11.2%), Methyloversatilis (2.1–12.0%), Hyphomicrobium (1.9–11.7%), Bacteroides (0.9–14.7%), Nitrospira (1.9–8.2%), and Methylibium (1.6–8.9%); the family Comamonadaceae (1.9–6.7%); the genera Pseudomonas (2.4–4.1%), Faecalibacterium (0.1–10.8%), and OLB12 from the family Microscillaceae (1.2–3.2%), the families Geobacteraceae (1.7–4.6%) and Methyloligellaceae (1.4–4.0%), and the genera Mycobacterium (1.1–3.9%) and Lacunisphaera (0.4–8.1%). The most prevalent taxa in the soil used for inoculation of the SMFCs included the family Muribaculaceae (16.9%), the genera Udaeobacter (15.8%), Bacteroides (10.8%), the families Bacillaceae (9.7%) and Xanthobacteraceae (8.8%), the genera Candidatus Xiphinematobacter (7.5%), Castellaniella (6.4%), and Chlorobium (4.7%). Interestingly, after the experiment, the soil near each cathode showed a similar microbial composition to the soil at day 0. The most abundant ASVs in the soil near the cathodes included the families Xanthobacteraceae (13.0–23.4%) and Bacillaceae (8.1–11.5%), the genera Candidatus Udaeobacter (8.0–15.4%), Candidatus Xiphinematobacter (5.4–7.6%), and Mycobacterium (4.5–10.8%); the families Rhodocyclaceae (1.7–8.4%), Methyloligellecaea (3.8–5.9%), Muribaculaceae (0.0–15.7%), and Erwiniaceae (0.0–17.8%), and the genera Bacteroides (2.8–6.2%). Canonical Correspondence Analysis (CCA) was carried out to determine which taxa clustered with anodes, cathodes, and soil samples ( Fig. 8 ). The microbial composition of the samples was found to be unique for each sampling location, suggesting selective enrichment on the anodes and cathodes throughout the experiment. The family Geobacteraceae and the genera Geobacter and Pseudomonas highly correlated with anodes. Both Geobacter and Pseudomonas are well-known electroactive species; Geobacter sulfurreducens is a model organism for direct electron transfer from cell to anode [ 69 ], while Pseudomonas spp. are known to produce redox compounds that may mediate the transfer of electrons to the anode by other members of the community [ 70 , 71 ]. In addition to Geobacter and Pseudomonas, a lower enrichment of Citrifermentans was also observed across the anode samples. This genus belongs to the family of Geobacteraceae and has been suggested to partake in exoelectrogenic processes [ 6 , 72 ]. Other electroactive taxa were found to be associated with the anodes; however, their presence was not ubiquitous to all SMFCs. These include Aeromonas in the case of SMFC-PtC and Faecalibacterium in the case of SMFC-GF and SMFC-PtC. Both genera have shown to be electroactive [ 73 , 74 ]. In addition, the genus Bacteroides was also found to be enriched in the anodic biofilm of SMFC-GF. Bacteroides have been previously detected in anodic biofilms, suggesting they may exhibit an electrochemical function [ 75 , 76 ]. Fig. 8 Canonical Correspondence Analysis (CCA) plot showing the association of taxa with anode, cathode, and soil samples. The red circle indicates time “0” (initial) soil sample. Fig. 8 While the relative abundances of species were consistent across the anodic biofilms, the several cathodes presented a different community profile with more variation, possibly as a result of the difference in the cathode material ( Fig. 7 ). On the other hand, the soil near the cathode was associated with a unique set of taxa , which clustered with the initial soil used for inoculum, and not with the cathode electrodes ( Fig. 8 ), suggesting that cathodic enrichment was limited to the electrode's surface. To determine the effect of cathode material on microbial composition, CCA was carried out for the cathode samples only ( Fig. 9 ). Taxa found enriched on the cathodic biofilms included facultative denitrifiers, electroactive species, as well as hydrogen-producing and hydrogenotrophic bacteria. Fig. 9 Canonical Correspondence Analysis (CCA) plot showing the association of taxa with different types of cathode materials used. CC: Carbon cloth; CNF–Fe: Carbon nanofibre doped with iron; PtC: Carbon cloth doped with platinum; GF: Graphite felt (two replicates: GF1 and GF2). Fig. 9 The facultative denitrifiers Methylobacterium/Methylorubrum , Hyphomicrobium, Methyloversatilis , Hydrogenophaga , and Comamonadaceae are able to reduce both nitrate and oxygen as the final electron acceptor, and all have been previously detected in cathodic biofilms in liquid microbial fuel cells [ [77] , [78] , [79] , [80] , [81] , [82] , [83] , [84] , [85] ]. Electroactive species were also detected on the cathodes, such as Faecalibacterium and Geobacter , albeit to a lesser extent than on the anodes. Several hydrogenotrophic taxa were observed on the cathodes, capable of using H 2 as an energy source. However, their abundance varied depending on the cathode material. Methylobacterium was associated with CNF–Fe and CC cathodes, Hydrogenophaga with GF cathodes, and Methyloversatilis with both CNF–Fe and GF cathodes. Hydrogenophaga harbour membrane-bound hydrogenases capable of catalysing both H 2 evolution and oxidation [ 86 ], while Methylobacterium was reported to produce formate from CO 2 , utilizing electrons from a cathode [ 87 ]. Similarly, the presence of Methyloversatilis was reported to positively correlate with the availability of H 2 [ 88 ], as well as with microbial corrosion of carbon steel [ 89 ], suggesting that this genus is also able to utilise H 2 to generate reducing power. Considering the taxonomic composition of the cathodes, we hypothesise that H 2 evolution and oxidation are important mechanisms of electron transfer from cathodes to the biofilm via membrane-bound uptake hydrogenases. It has also been suggested, in the case of liquid-microbial fuel cells, that extracellular hydrogenases and formate dehydrogenases are involved in electron transfer via rapid metabolite cycling [ 90 ]. Thus, cathodic environments would favour the enrichment of taxa harbouring such enzymes. The presence of hydrogenotrophs in the cathode biofilms may be associated with improved cathodic activity and overall SMFC performance since the rate of H 2 consumption at the cathode may determine the rate of H 2 evolution from the cathode [ 91 ]. The only taxon strongly associated with both replicates of the best-performing GF cathodes was the genus Hydrogenophaga . The greater redox activity observed in the case of SMFC-GF ( Fig. 5 ) may indicate an improved capacity of the bacterial communities to accept electrons directly from the cathode to catalyze the oxygen reduction reaction [ 92 ], which could be attributed to the greater enrichment of Hydrogenophaga on these cathodes [ 93 ]. Other taxa were found enriched on either but not both SMFC-GF replicates, including Lacunisphaera (SMFC-GF1) and Faecalibacterium , Bacteroides , and Parabacteroides (SMFC-GF2) . Of these genera, Bacteroides have been reported to produce hydrogen [ 94 , 95 ], whereas the ability to reduce nitrate was reported for both Lacunisphaera and Bacteroides [ 96 , 97 ]. Since the redox potential of the NO 3 − /NO 2 − couple is +421 mV, likely, the peaks at 419 ± 20 mV observed for SMFC-GF in the CV tests (shown in Fig. 5 ) are due to the microbial reduction of nitrate to nitrite under aerobic conditions [ 98 ]. This result suggests that the enrichment of denitrifiers in the SMFC-GF cathode biofilms, specifically Hydrogenophaga , Lacunisphaera , and Bacteroides, may have facilitated nitrate reduction reaction. Despite the variability of taxa across different cathodes, an unidentified Gammaproteobacterium was the most prevalent ASV in all cathodic samples. Although the presence of this ASV was low in the inoculum (0.4%), it was enriched not only in the cathodic biofilms of all SMFCs (10.9–21.6%), but also in the anodic biofilms, albeit to a lesser extent (1.6–4.0%). The nucleotide sequence of this ASV was searched against the NCBI database using the BLAST tool, which returned a hit with 100% homology: the sequence of this uncultured bacterium was previously detected in paddy soil microbial fuel cells [ 99 ]. In addition, unclassified Gammaproteobacteria were previously reported to dominate biofilms of high-performing oxygen-reducing cathodes [ 85 ]. Due to its ubiquity in electrode biofilms in this experiment, it can be hypothesised that this uncultured Gammaproteobacterium is involved in electrochemical processes, particularly in the cathodic biofilm."
} | 8,826 |
39124221 | PMC11314055 | pmc | 3,003 | {
"abstract": "The global ecosystem relies on the metabolism of photosynthetic organisms, featuring the ability to harness light as an energy source. The most successful type of photosynthesis utilizes a virtually inexhaustible electron pool from water, but the driver of this oxidation, sunlight, varies on time and intensity scales of several orders of magnitude. Such rapid and steep changes in energy availability are potentially devastating for biological systems. To enable a safe and efficient light-harnessing process, photosynthetic organisms tune their light capturing, the redox connections between core complexes and auxiliary electron mediators, ion passages across the membrane, and functional coupling of energy transducing organelles. Here, microalgal species are the most diverse group, featuring both unique environmental adjustment strategies and ubiquitous protective mechanisms. In this review, we explore a selection of regulatory processes of the microalgal photosynthetic apparatus supporting smooth electron flow in variable environments.",
"conclusion": "7. Concluding Remarks The photosynthetic apparatus is a sophisticated and intertwined machinery that maintains efficient energy conversion rates under varying environmental conditions. In this review, we covered the basic blueprint of how photosynthetic electron transfer generates the pmf and how fine-tuning the latter is pivotal for survival in an everchanging surrounding. We highlighted several special adaptations of oxygenic photosynthesis in microalgal systems. Besides already exploited feats such as sourcing lipid-rich biomass, the microalgal group in its yet to be fully explored diversity holds promising photoprotective traits that may be beneficial for photosynthesis in the field. Assembled data from different niches should therefore hold a key constituent for future studies, which could pave the road for bioengineering a more resistant, adaptable, and efficient system.",
"introduction": "1. Introduction 1.1. The Diversity of Microalgal Oxygenic Photosynthesis Photosynthesis is a relatively ancient development of life on earth that uses light to capture CO 2 via Ribulose-1,5-bisphosphate carboxylase/oxygenase (RuBisCO) activity. This review will focus on the most successful type of photosynthesis which is oxygenic and uses water as an electron donor for CO 2 fixation in the Calvin Benson Bassham (CBB) cycle. It was initially developed in proto-cyanobacterial organisms which are dated as far as 2.3 Ga [ 1 ], with the earliest fossil findings pointing to 1.9 Ga [ 2 ]. As of now, there is an agreement that all oxygenic photosynthesizers originate from a single lineage of organisms, which possessed both type I and II photosynthetic reaction centers (aka PSI and PSII) [ 3 , 4 , 5 ]. These proto-cyanobacteria later evolved to the current day cyanobacteria to eventually engage in an endosymbiosis event (or events, see [ 6 , 7 , 8 ]), giving rise to the O 2 -producing plastids occurring in the lineage of Archaeplastida which includes green and red algae as well as Glaucophytes, but also in other domains such as Stramenopiles (e.g., diatoms) and Alveolata (e.g., dinoflagellates and Chromerida) [ 9 ]. As a very diverse group, microalgae can be found virtually everywhere, conquering both aquatic and terrestrial habitats including soil, aeroterrestrial and epiphytic habitats by developing unique adaptations [ 10 ]. The combination of a relatively short life cycle paired with a large eukaryotic genome, around 20–150 Mb (excluding exceptions [ 11 ]), might have helped microalgae to succeed in a competitive environment. Unsurprisingly, species of the same genus, such as Chlorella , were found in completely different environments—from Antarctic oceans [ 12 , 13 ] to Mediterranean deserts [ 14 ]—and in many cases feature distinct gene expression patterns in response to their habitat and the associated stress types [ 15 ], yet holding little genomic variation. On the other hand, the phenotypic expression of different algal lineages of similar habitats exhibits such converged traits that previous classification attempts led to a grand mix-up of genetic lineages [ 16 ]. In this review, we will shed light on the regulation of electron transfer processes that generate a transmembrane electrochemical proton gradient, also referred to as proton motive force ( pmf ). Our focus will be on how the pmf is fine-tuned for sustained photosynthetic productivity and how environmental adaptations altered these regulatory processes in different microalgae. However, functional microalgal photosynthesis research is entangled with research on other phototrophs since conserved fundamental processes are involved, such as energy stabilization upon water splitting in the oxygen-evolving complex (OEC). To fully cover how the pmf is regulated, this review will also lean on extrapolated knowledge derived from other photosynthetic domains. We will provide an overview of unique aspects of photosynthesis regulation in a selection of microalgal examples, acknowledging that covering the entirety of microalgal diversity will be beyond the scope of this review. To illustrate the heterogeneity of the term ‘microalgae’, we included a simplified phylogenetic tree ( Figure 1 , based on recent studies [ 9 , 17 , 18 ]), presenting the most relevant model organisms of eukaryotic microbial phototrophs. 1.2. The Oxygenic Photosynthetic Apparatus In all oxygenic photosynthesizers, the pmf across the thylakoid membrane is constituted of two components: chemical (osmotic H + gradient, ΔpH) and electric (membrane potential, ΔΨ). The electrons which are released during water oxidation (at the OEC of PSII) reduce a plastoquinone (PQ) molecule, situated in the acceptor side of PSII (Q B ), converting it to plastoquinol (PQH 2 ). As PQH 2 diffuses within the membrane, it can reduce the cytochrome b 6 f complex (Cyt b 6 f ) and by doing so, increase the capacity of pmf generation [ 19 , 20 ]. The electron transfer between Cyt b 6 f and PSI is then mediated by either plastocyanin (Pc) or cytochrome c 6 (Cyt c 6 ). This variation originates from the altered metal cofactors and their environmental abundance, which in some cases determines the expression levels of Pc (containing a copper atom) and Cyt c 6 (containing an iron–heme cofactor) [ 21 ]. Some lineages, such as red algae, have lost the genes encoding Pc, while other lineages such as Charophytes and the derivative lineage of land plants almost exclusively rely on it. These lineages were thought to have completely lost the genes encoding Cytc 6 , although recent studies discovered Cytc 6 orthologs that are still poorly characterized (e.g., Cytc 6A and Cytc 6B ) [ 22 ]. Adequately, these adaptations also triggered alterations of the interacting residues, situated on the PSAF loop of PSI, in both the green lineage during the transition to land [ 23 ] and across other photosynthetic lineages [ 24 , 25 , 26 ]. Following Pc/Cytc 6 diffusion towards and reduction of photo-oxidized PSI, the energy stored within its excitation is channeled to the three [4Fe–4S] centers (F X , F A , F B ). PSI then most prominently reduces ferredoxin (FDX), which is a small soluble electron carrier [ 27 , 28 ] that mediates a plethora of redox reactions, such as NADPH production via FDX:NADP + oxidoreductase (FNR) [ 29 , 30 ]. The photo-reduced [2Fe-2S] cluster of FDX feeds into diverse redox carrier pools, such as thioredoxins and thioredoxin-like proteins [ 31 ]. Broadly, these processes are fine-tuned by an intricate regulatory network, aiming to maintain a proper pmf which allows bioenergetic membranes to engage in chemiosmosis via ATP synthase (F O F 1 ) [ 32 ]. In Figure 2 , we present a schematic illustration of the photosynthetic apparatus, based on green microalgal physiology. The boxes highlight the sections covered in this review, comprising a selection of the latest findings in the field."
} | 1,986 |
39763919 | PMC11702635 | pmc | 3,005 | {
"abstract": "Microbes of nearly every species can form biofilms, communities of cells bound together by a self-produced matrix. It is not understood how variation at the cellular level impacts putatively beneficial, colony-level behaviors, such as cell-to-cell signaling. Here we investigate this problem with an agent-based computational model of metabolically driven electrochemical signaling in Bacillus subtilis biofilms. In this process, glutamate-starved interior cells release potassium, triggering a depolarizing wave that spreads to exterior cells and limits their glutamate uptake. More nutrients diffuse to the interior, temporarily reducing glutamate stress and leading to oscillations. In our model, each cell has a membrane potential coupled to metabolism. As a simulated biofilm grows, collective membrane potential oscillations arise spontaneously as cells deplete nutrients and trigger potassium release, reproducing experimental observations. We further validate our model by comparing spatial signaling patterns and cellular signaling rates with those observed experimentally. By oscillating external glutamate and potassium, we find that biofilms synchronize to external potassium more strongly than to glutamate, providing a potential mechanism for previously observed biofilm synchronization. By tracking cellular glutamate concentrations, we find that oscillations evenly distribute nutrients in space: non-oscillating biofilms have an external layer of well-fed cells surrounding a starved core, whereas oscillating biofilms exhibit a relatively uniform distribution of glutamate. Our work shows the potential of agent-based models to connect cellular properties to collective phenomena and facilitates studies of how inheritance of cellular traits can affect the evolution of group behaviors.",
"introduction": "Introduction Bacterial biofilms are large communities of cells that exist in nearly every environment [ 10 ]. They are bound together by an extracellular matrix that provides both stability and protection [ 3 , 8 , 1 ]. Biofilms exhibit a variety of emergent behaviors that give biofilm-dwelling microbes advantages unavailable to planktonic cells [ 16 , 31 , 42 ]. For example, cells within biofilms differentiate into heterogeneous phenotypes [ 22 , 45 , 46 , 21 ], divide labor [ 28 , 36 ], and coordinate behavior via chemical signals [ 13 , 33 , 49 , 23 ]. These group phenomena have led researchers to assert that biofilms represent a transition between single-celled and multicellular life [ 38 , 7 ]. A striking multicellular behavior is the presence of cell-to-cell electrochemical signals that influence metabolism in Bacillus subtilis biofilms [ 25 , 34 ]. As a biofilm expands, fewer nutrients penetrate to the center; most are consumed by exterior cells [ 43 , 51 ]. The paucity of nutrients in the interior raises a problem: if interior cells are starved, the integrity of the biofilm is at risk [ 25 ]. In vitro B. subtilis biofilms exhibit a behavior that seems to allow them to navigate this challenge. When interior cells become starved, they release potassium, depolarizing nearby cells and hampering their ability to absorb glutamate. In turn, nearby cells become distressed, release potassium, and hyperpolarize, eventually leading to a wave of potassium release. This wave propagates to the biofilm exterior [ 34 ]. It has been hypothesized that glutamate consumption among cells in the exterior slows down enough that glutamate can diffuse to the center [ 25 ]. Once interior cells have enough glutamate, they cease releasing potassium, allowing exterior cells to repolarize and resume consumption, eventually leading to stress and another wave of potassium release. These repeated waves of potassium release have been referred to as a form of microbial “signaling” [ 34 , 29 , 11 ]. Potassium signaling has been proposed to allocate nutrients efficiently at the colony level [ 25 , 24 ], and it is heterogeneous at the cellular level. Some cells participate in signaling and hyperpolarize during signaling waves, whereas others do not [ 20 ]. It is unknown how cellular variation in signaling behavior affects biofilm-level properties, such as distributions of nutrients. In order to answer this question, we need models that can connect cell-level properties, such as signaling state and inheritance of signaling behavior, to colony-level phenomena. Several computational models of B. subtilis signaling behavior have been introduced to explore hypotheses about the causes and effects of signaling. Zhai and colleagues (2019) proposed an agent-based model to explain observations they had made about signaling. Their in vitro experiments revealed that a roughly constant proportion of cells signal each oscillation, that the same cells tend to release potassium in repeated signaling waves, and that signaling behavior is weakly heritable—that is, daughter cells of signaling cells are more likely than average to participate in signaling waves. They modeled signaling as a percolation process in which a cell only signals during a depolarization wave if it both has a binary trait that predisposes it to signaling and is adjacent to another signaling cell in the biofilm. Using an agent-based model in which agents represent individual cells allowed them to test whether signaling in this manner would transmit a signal across the biofilm consistently. However, their model focused on small sub-regions of the biofilm to match the limitations of their experimental system—a roughly 35×230 rectangle of cells at the edge of the biofilm, where the colony is close to two-dimensional. Their model also did not include nutrient diffusion or uptake, preventing its use for studying how individual cell behaviors affect the distribution of nutrients or growth of the biofilm. Other models of B. subtilis depolarization waves are based on systems of differential equations. Martinez-Corral et al. (2018) produced a model of a one-dimensional slice of the biofilm, extending from the center to an edge. Ford et al. (2021) extended this to two dimensions, simulating a complete biofilm. Both models aimed to capture signaling and nutrient patterns at the scale of an entire biofilm. These models explicitly simulate nutrient diffusion and metabolism and have signaling operate through mechanisms that depend on internal glutamate concentration, providing powerful and accurate recreation of biofilm-wide signaling dynamics. However, modeling these complex interactions at a larger scale using differential equations comes at the cost of resolution. These models describe phenomena on the scale of the biofilm but do not distinguish individual cells. Their advantages are thus opposite those of the agent-based model of Zhai and colleagues, but neither can describe the effects of individual cell behaviors on broad patterns of nutrient distribution or signaling. The model we propose strikes a compromise between the flexibility and resolution of the agent-based approach of Zhai and colleagues (2019) and the scalability of ODE models. Our approach is agent-based, but the agent-based elements are overlaid on a simplified version of the ODE model developed by Martinez-Corral et al. (2019). Via this hybrid strategy, our model retains some of the benefits of both previous approaches. Our model enables simulation at the scale of an entire flow-cell biofilm [ 15 , 35 ], comprising approximately 51,000 individual cells, each with unique potassium, glutamate, membrane potential, and signaling dynamics. We validate our model by comparing the behaviors of simulated biofilms with those observed in experiments, including signaling patterns at local and colony-wide scales, response to various stressors, and growth patterns. We show that many of the distinctive features of B. subtilis signaling, including waves of depolarization and the fraction, identity, and descent of cells that participate in signaling, can emerge naturally from our model. We then demonstrate the application of our model by exploring open questions regarding synchronization of oscillations among neighboring biofilms [ 24 ] and the effect of signaling on glutamate distribution.",
"discussion": "Discussion We introduced a computational model of metabolic signaling in B. subtilis biofilms. Previous models of this behavior have either been small in scope, only able to examine local behaviors of cells and omitting nutrients, or large in scope but unable to study heterogeneity in cell-level behavior [ 50 , 11 ]. We have developed a model that bridges this gap, allowing the examination of the effect of cell-level behaviors on broader signaling patterns and the concentration of nutrients across the biofilm. We were able to replicate both individual-cell and biofilm-scale observations from previous work and new experiments, including oscillation and growth patterns, signaling in interior and exterior cells, and synchronization between neighboring biofilms. We also found support for the hypothesis that signaling results in a more even distribution of glutamate, which may extend the lifespan of a biofilm during periods of stress. Previous models of B. subtilis signaling have adopted various assumptions about the effects of signaling on individual cells and the biofilm. On one hand, the models of Prindle and colleagues (2015), Martinez-Corral and colleagues (2018, 2019), and Ford and colleagues (2021) encoded assumptions that imply that signaling will increase glutamate uptake for the signaling cell both by directly increasing the cell’s ability to absorb glutamate, and suppressing glutamate absorption for neighboring cells. On the other hand, the models in Larkin et al. (2018) and Zhai et al. (2019) prioritized the observation that hyperpolarized cells experience slower growth [ 24 ], although more recent work has suggested that the slow growth of hyperpolarized cells may be an artifact of ThT staining itself inhibiting growth [ 14 ]. Larkin and colleagues hypothesized signaling to be costly to the individual cell but beneficial to the biofilm as a whole, as it promotes a more even distribution of glutamate. Further, they noticed that the fraction of cells that signal in a given wave was close to the minimum number of cells necessary for the signaling wave to propagate across the exterior of the biofilm as predicted by percolation theory [ 41 ] (where signalers are randomly distributed among non-signalers and a signal is propagated by direct contact between two signaling cells). They interpreted this observation as being consistent with the idea that signaling cells act altruistically, sacrificing their own growth to promote the integrity of the biofilm. In our model, we adopt assumptions similar to those of Prindle et al. (2015) and Martinez-Corral et al. (2019) that lead to signaling typically increasing the glutamate uptake of the signaling cell. At the same time, we replicate the heterogeneity in signaling behavior, the fraction of signaling cells, and the individual-level consistency of signaling across waves emphasized by Larkin et al. (2018) and Zhai et al. (2019). Thus, the individual-cell-level signaling patterns observed by the latter studies—and particularly a fraction of signaling cells near the percolation-theory threshold for signal transmission—can be attained without an explicit trade-off between individual-level growth and group-level glutamate distribution. However, like Larkin et al. (2018) and Zhai et al. (2019), our results are consistent with the idea that cell-level heterogeneity is important. In our model, a particular amount of variation in propensity to signal is necessary to achieve synchronized oscillations. In the presence of such variation, the cells with the highest propensity to signal hyperpolarize first. Once enough cells participate, a wave of signaling occurs, relieving glutamate stress and suppressing further signaling. Under this hypothesis, the participating fraction of cells may be near the level predicted by percolation theory because once that level is reached, stress is relieved and further signaling is not required. The observation that a requisite level of variation in signaling propensity is necessary to produce coordinated waves of signaling in our model raises further questions. What could be the source of variation in signaling propensity, and how could this variation be maintained? In vitro biofilms observed to participate in signaling are typically clonal, so variation in signaling behavior is unlikely to be genetic in well-studied cases. Yet signaling behavior is observed to be heritable, in the sense that daughter cells are more likely to participate in signaling waves if their mother cell signals. One speculative possibility is that the regulatory network controlling potassium channel expression [ 27 ] results in multi-generational epigenetic inheritance of signaling [ 45 , 32 ]. Whatever the source of the variation, on the basis of current observations, if the apparent individual-level cost of signaling is in fact an artifact of ThT staining [ 14 ], cells with a proclivity to signal might be expected to increase in frequency within the biofilm, taking up more glutamate than their neighbors, dividing more quickly, and potentially transmitting (non-genetically) their elevated propensity to signal to their offspring. Depending on how propensity to signal is realized and transmitted, such a process could lead to a decline of variation in propensity to signal, or at least to a decline of heritable variation, if continued long enough and if there are no forces generating new heritable mutation (analogous to mutation). (Our model contains such a force, as random deviations from a parent cell’s signaling threshold are partially inherited by offspring.) In our model, if too many cells signal, oscillations cease to be coordinated, and the distribution of internal glutamate—while much more even than in the complete absence of signaling—leaves some cells at the interior of the biofilm starved of glutamate. Thus, our model raises a possibility that is almost the reverse of the one raised by Larkin et al. (2018) and Zhai et al. (2019)—if signaling improves glutamate uptake for the signaling cell and reduces glutamate uptake for its neighbors, we might think of the cells that do not signal, rather than the ones that do, as acting altruistically, giving up their access to glutamate so that interior cells are not starved. There remain other possibilities—there may in fact be a cost of signaling to the individual, the increase in glutamate uptake from signaling may be dependent on the signaling state of a cell’s neighbors, or any of a number of others. In our current implementation, reproduction is not dependent on internal glutamate, so we do not explore such questions, but they are important for future theoretical and experimental work. Another area of future study involves extending our model to predict how other processes are altered by emergent electrochemical signaling. For example, the expression of some genes has been proposed to be regulated by ion-responsive kinases [ 12 ]. By coupling cellular potassium flux to gene expression in our model, we could predict patterns of gene expression heterogeneity that would arise due to signaling. In addition, other cell phenotypes are regulated by nutrient conditions, notably matrix production and sporulation [ 26 ]. By modeling the response of genetic circuits that control the differentiation into these phenotypes [ 5 ], we could predict how the altered distribution of nutrients in signaling biofilms in turn alters the distribution of matrix producers and spores [ 46 , 40 , 6 ]. Our model may prove valuable to understanding the feedback between cellular phenomena and emergent nutrient conditions within biofilms, a topic of recent interest [ 17 ]. Overall, our work shows that combining agent-based and diffusion-based models can account for the emergence of community-level properties from interactions of individual cells. Doing so allows us to study the effect of signaling behavior on the biofilm as a whole, and on individual cells, taking into account heterogeneity among cells. That so many of the collective and cell-level signatures of B. subtilis biofilm signaling can be observed in a simple model hints at a relatively simple set of principles governing in vitro signaling behavior."
} | 4,111 |
35007419 | PMC8796234 | pmc | 3,006 | {
"abstract": "Furfural\nand 5-hydroxymethyl\nfurfural (HMF) are abundantly available\nbiomass-derived renewable chemical feedstocks, and their oxidation\nto furoic acid and furan-2,5-dicarboxylic acid (FDCA), respectively,\nis a research area with huge prospective applications in food, cosmetics,\noptics, and renewable polymer industries. Water-based oxidation of\nfurfural/HMF is a lucrative approach for simultaneous generation of\nH 2 and furoic acid/FDCA. However, this process is currently\nlimited to (photo)electrochemical methods that can be challenging\nto control, improve, and scale up. Herein, we report well-defined\nruthenium pincer catalysts for direct homogeneous oxidation of furfural/HMF\nto furoic acid/FDCA, using alkaline water as the formal oxidant while\nproducing pure H 2 as the reaction byproduct. Mechanistic\nstudies indicate that the ruthenium complex not only catalyzes the\naqueous oxidation but also actively suppresses background decomposition\nby facilitating initial Tishchenko coupling of substrates, which is\ncrucial for reaction selectivity. With further improvement, this process\ncan be used in scaled-up facilities for a simultaneous renewable building\nblock and fuel production.",
"conclusion": "Conclusions We report here molecular catalysts for the direct\ncatalytic oxidation\nof furfural and HMF to furoic acid and FDCA, respectively, using alkaline\nwater as the formal oxidant. The oxidation is associated with the\ngeneration of pure H 2 gas with no detectable CO contamination\n(detection limit: 15 ppm), suitable for direct utilization in a proton-exchange\nmembrane fuel cell. When the ruthenium acridine PNP complex 6 was used as the catalyst, furoic acid/FDCA was obtained\nwith high yield (>95%) by aqueous oxidation of furfural/HMF. Mechanistic\nstudies revealed an initial hydride transfer from the substrate to\nthe catalyst ligand backbone under the conditions, generating the\ndearomatized complex 9 , which subsequently catalyzed\nthe oxidation. Complex 9 also catalyzes the Tishchenko\ncoupling of substrates, which is essential for background decomposition\nsuppression. Overall, the Ru-acridine PNP-based system is unique,\nwith its atypical reactivity, in catalyzing the selective furfural/HMF\noxidative reactions with complete inhibition of substrate decomposition,\nresulting in high furoic acid/FDCA and H 2 yields. We believe\nthat this report will initiate further investigations toward the homogeneous\ncatalytic oxidation of furfural/HMF, largely overlooked until now,\nespecially with water as the formal oxidant, given the dire importance\nof transitioning toward renewable material and fuel synthesis in the\ncontext of modern sustainability. With sufficient improvements in\nthe conditions, such a homogeneous process could be ideal for large-scale\nFDCA (and furoic acid) and H 2 synthesis from the HMF (and\nfurfural)–water mixture in an industrial setup, compared to\nthe equivalent (photo)electrochemical processes.",
"introduction": "Introduction Owing to the negative\nconsequences of fossil fuel use, intensive\nresearch is ongoing, focusing on transitioning toward a renewable\nframework for fuel and materials production. 1 − 5 Furfural and 5-hydroxymethyl furfural (HMF) are chemical\nfeedstocks produced by hydrolysis of biomass waste. 6 − 8 Because of their\nrenewable nature, the synthesis of commodity chemicals from furfural\nand HMF has garnered increasing attention. 9 − 12 Among many products obtainable\nfrom furfural and HMF, their oxidation products, furoic acid and furan\ndicarboxylic acid (FDCA), respectively, hold particular interest ( Figure 1 A). Furoic acid has\nmany applications including plastic plasticizer, food preservative,\npharmaceutical intermediate, and FDCA precursor, and has potential\napplications in optics technology because of its unique crystal properties,\nwith large-scale synthesis plants operated by multiple companies. 13 − 16 Similarly, FDCA is a promising renewable alternative to terephthalic\nacid for polymers synthesis. 17 FDCA-based\nrenewable biopolymers often show improved mechanical, thermal, and\ngas transport properties compared to their terephthalic acid based\ncounterparts found in the market ( Figure 1 B), 18 , 19 with the United States\nDepartment of Energy identifying FDCA as 1 of the 12 priority chemicals\nfor the establishment of a green chemical industry in the future. 20 , 21 Several companies started pilot plants for FDCA synthesis from HMF\nover the past decade because of its growing market in the polymer\nindustry; however, the markedly different approaches undertaken reflect\nthe lack of an economically optimized process for the desired synthesis. 22 Figure 1 Different aspects of furfural and HMF oxidation to furoic\nacid\nand FDCA. (A) Chemical equations with previous approaches. (B) Bioplastics\n(PEF and PPF) derived from FDCA. (C) HMF oxidation by water to FDCA\nwith H 2 evolution with previous and new approach. (D) Challenges\nfaced in this study and its circumvention. (E) Relevant ongoing catalytic\naqueous oxidation reactions during the process. The most explored selective oxidative routes to access furoic acid\nand FDCA from furfural and HMF use heterogeneous catalysts, such as\nsupported PbPt/C, Au, Ag 2 O/CuO, and AuPd/Mg(OH) 2 , with excess oxidants (mainly high-pressure oxygen or air). 23 − 26 The process produces water as the side product for HMF to FDCA conversion,\nwhich although environmentally benign, does not hold any economic\nvalue. Alternative oxidation methods such as electrochemical 27 − 29 and bio 30 -enzymatic 31 − 34 oxidation of furfural and HMF\nto furoic acid and FDCA have also been explored. Recent reports have\nelegantly coupled H 2 production from water with biomass\noxidation to (photo)electrochemically produce H 2 and furoic\nacid/FDCA from water and furfural/HMF mixture ( Figure 1 C). 35 − 40 These photoelectrochemical systems, however, require advanced specialized\nmaterials and can be challenging to rationally improve. Besides, their\nlarge-scale implementation can be difficult because of the need for\nsophisticated infrastructures and low working concentrations. 41 In contrast to the explored heterogeneous,\nbiological, enzymatic,\nand (photo)electrochemical processes, catalytic homogeneous systems\nfor furfural and HMF oxidation are extremely limited. The use of well-defined\nhomogeneous complexes for furfural/HMF oxidation is challenging because\nof the facile substrate decomposition pathways at high temperatures\nin alkaline/aerobic conditions leading to the formation of polymeric\nproducts ( Figure 1 D). 42 , 43 Goldberg and co-workers have reported complexes that are active\nin catalyzing the aqueous reforming of other aldehydes to acids, but\ndisplay minimal activities when furfural/HMF is used as a substrate. 44 , 45 Interestingly, Nakajima and co-workers have recently reported an\nN-heterocyclic carbene organocatalyst for furfural to furoic acid\nconversion in the presence of 1,8-diazabicyclo[5.4.0]undec-7-ene (DBU)\nbase, which use O 2 as the oxidant, but can only partially\noxidize HMF to 5-hydroxymethylfurancarboxylic acid intermediate. 46 Herein, we report the catalytic homogeneous\noxidation of furfural\nand HMF to furoic acid and FDCA, respectively, using alkaline water\nas the formal oxidant. The reaction is catalyzed by well-defined ruthenium\ncomplexes with acridine-based PNP pincer ligands 47 , 48 and generates pure H 2 gas as the reaction byproduct.\nMechanistic studies indicate that the Ru complexes not only catalyze\nthe substrate oxidation to acid but also induce rapid substrate disproportionation\nin the initial hour, which is crucial in preventing substrate decomposition.\nFurther reaction involves the dehydrogenative oxidation of the generated\nalcohol by water ( Figure 1 E). The scalability of the process is demonstrated by carrying\nout a gram-scale reaction. Notably, the system can theoretically produce\nup to a substantial 3.48 wt % H 2 when HMF is used as a\nsubstrate and LiOH as the base if a neat system is developed, generating\nboth renewable fuel and material precursors in one simple homogeneous\nprocess.",
"discussion": "Results and Discussion Furfural Oxidation to Furoic Acid Catalyst\nScreening Our investigation started by exploring\nthe aqueous furfural oxidation to furoic acid in the presence of the\nRu-PNN bipyridyl complex 1 ( Figure 2 A), which is reported by us to catalyze aqueous\ndehydrogenative oxidation of alcohols to carboxylic acid salts by\nalkaline water ( Table S1 , Supporting Information). 49 However, attempts toward aqueous oxidation of\nfurfural at 135 °C by complex 1 (1 mol %) in 1,4-dioxane/alkaline\nwater resulted in complete decomposition of furfural with no generation\nof H 2 or furoic acid. A control reaction revealed that\nfurfural is prone to degradation at elevated temperature under the\nreaction conditions, even without any catalyst. We subsequently screened\nseveral ruthenium-based pincer complexes developed in our group for\nthe desired dehydrogenative oxidation reaction. However, all efforts\ninvolving complexes 1–4 resulted in substrate\ndecomposition with no significant gas generation ( Figure 2 A). In the case of the Ru-PNNBPy Ph complex ( 5 ), the furfural decomposition rate\nslowed down, obtaining the acid and alcohol as the reaction products,\nhowever, with only a small amount of H 2 generated (8%)\n( Table S1 , entry 7). Remarkably, the acridine\nPNP complex 6 catalyzed the reaction with high H 2 (80%) yield and furoic acid (87%) yield ( Table S1 , entry 8). GC analysis of the generated gas mixture\nshowed only H 2 gas with no CO contamination, suitable for\nits use in a proton-exchange membrane (PEM) fuel cell without further\npurification. Optimization of the catalyst amount showed that 1 mol\n% catalyst loading is ideal for both high H 2 and furoic\nacid yields, with lower catalyst loadings being detrimental, especially\nfor H 2 yields ( Figure 2 A). A stoichiometric base was necessary for the reaction,\nand under the catalytic base, decreased yields were observed. Furoic\nacid and H 2 were obtained in >95% yield by increasing\nthe\nreaction time from 36 to 48 h. Figure 2 Catalytic oxidation of furfural and HMF\nwith water as the formal\noxidant. (A–D) Furfural to furoic acid: (A) Catalyst and condition\nscreening. (B) Effect of base. (C) Catalyst ligand substitution effect.\n(D) Temperature effect on furfural to furoic acid conversion. (E–G)\nHMF to FDCA: (E) Reaction equation. (F) Different possible parallel\npathways. (G) Condition optimization to obtain FDCA and H 2 . Reactions in (A)–(D) were conducted using 1 mmol of furfural,\n1 mol % catalyst, and 1.2 equiv of the base in 1,4-dioxane (2 mL)/water\n(1 mL) unless otherwise specified. Reactions in (B) and (D) are with 6 (1 mol %) as the catalyst, and reactions in (C) used NaOH\n(1.2 equiv) as the base with 48 h reaction time. Reactions in (G)\nused 0.5 mmol of HMF in 1,4-dioxane (2 mL)/water (1 mL), with other\nreaction conditions as specified. Yields calculated by 1 H NMR (mesitylene standard) and gas buret (H 2 ). Yields\ncorrespond to furoic acid or FDCA salts before acidification with\nisolated yields in parentheses. Effect of the Base, Catalyst Substitution, and Temperature Next, the effect of different bases was explored ( Figure 2 B; Table S2 ). Strong bases such as NaOH, KOH, and LiOH were effective\nfor high furoic acid and H 2 yield, with relatively weaker\nbases such as K 3 PO 4 and Na 2 CO 3 being similarly effective (>90% furoic acid yield). Decreasing\nthe base strength further as with NaHCO 3 resulted in decreased\nfuroic acid yield (52%), with the rest of the product being the alcohol.\nAmong amine-based organic bases, DBU was moderately effective for\nacid generation (78%), whereas with dimethyl aminopyridine (DMAP),\nalmost no acid formed, with 85% furfural being unreacted. We also\nexplored the effect of different ligand substituents of the catalyst\nstructure on product yield ( Figure 2 C). Complexes Ru-Acr( i Pr) ( 6 ) and Ru-Acr(Cy) ( 7 ) displayed similar catalytic activities\nunder the conditions for furfural oxidation to furoic acid. In contrast,\nthe Ru-Acr(Ph) complex 8 , with electron-withdrawing Ph\nsubstitutions onto phosphorus donor atoms, was slightly less active\nin catalyzing the reaction. Optimization studies regarding reaction\ntemperature revealed that 135 °C is required for reaction completion\nin 48 h and decreasing the temperature to 125 or 115 °C resulted\nin lower yields ( Figure 2 D). HMF Oxidation to FDCA The direct\noxidation of HMF to\nFDCA via this homogeneous dehydrogenative aqueous oxidation method\nwas subsequently explored ( Figure 2 E). As mentioned earlier, direct FDCA synthesis from\nHMF by water is a process with a great prospect in the industrial\nproduction of biobased renewable polymers and fuels, for which current\nmethods are limited. HMF oxidation to FDCA can occur via two different\nroutes— one via the generation of 5-hydroxymethyl-2-furancarboxylic\nacid (HMFCA) from the initial oxidation of the aldehyde group to acid\nor via initial oxidation of the alcohol group in HMF forming diformylfuran\n(DFF), whose subsequent oxidation generates FDCA ( Figure 2 F). With use of our method,\nheating HMF at 135 °C in the presence of complex 6 (4 mol %), NaOH (2.2 equiv) in 1,4-dioxane/water (2:1 mL/mL), FDCA\nformation in 15% yield was observed after 18 h ( Figure 2 G, entry 1). The primary reaction product\nwas the HMFCA intermediate (75%), signifying that oxidation of the\naldehyde group is easier than that of the alcohol group under the\nconditions ( Figure S12 ). These two products,\nalong with the disproportionation product bis(hydroxymethyl)furan\n(BHMF, 10%), accounted for all the HMF conversion (99%), indicating\nthe absence of any polymeric side pathways. A higher yield of FDCA\n(70%) was obtained by using a more alkaline solution (4 equiv of base)\nand a longer reaction time (60 h) (entry 2). Similar to furfural oxidation,\ncomplex 8 was less active for HMF to FDCA oxidation,\ntoo (entry 3). FDCA yield increased to 95% when the reaction temperature\nwas increased to 150 °C, using 6 as the catalyst\n(entry 4). Under optimized conditions, FDCA in high yield (95%) was\nobtained (H 2 yield: 93%) with 2 mol % of complex 6 and 2.2 equiv of NaOH, at 160 °C after 68 h of reaction\n(entry 5). LiOH was similarly active as NaOH as a base in facilitating\nFDCA formation under the reaction conditions (entry 6). Thus, it is\nshown that complex 6 can catalyze the direct and selective\nHMF oxidation to FDCA in the presence of alkaline water with high\nyields while also generating quantitative pure H 2 gas.\nIn the absence of catalyst, decomposition of HMF into unidentifiable\nproducts was observed (entry 7). Mechanistic Investigation We subsequently explored\nthe reactivity complex 6 with aldehyde, base, and water\nto understand the reaction mechanism ( Figure 3 ). Initial experiments were carried out with\nbenzaldehyde as the furfural surrogate, which is more stable at higher\nreaction temperatures and easier to follow. Complex 6 does not react with benzaldehyde (5 equiv) under neutral conditions\nin a THF/water solvent mixture (0.5:0.1 mL/mL), even when heated at\na high temperature of 130 °C for 0.5 h ( Figure 3 a). On the other hand, when NaOH (5 equiv)\nwas added, and the solution was subsequently heated at 130 °C\nfor 10 min inside a J. Young NMR tube, generation of two new complexes\nwere observed in the 31 P{ 1 H} NMR spectrum ( Figure S31 ) with their characteristics 31 P chemical shifts at 74.1 and 87.5 ppm, respectively, at a 0.8:1.0\nratio (parent complex 31 P signal chemical shift is at 69.1\nppm). In the 1 H NMR spectrum, surprisingly, the 9H acridine\naromatic protons from both complexes were missing, which appear around\n8–9 ppm as a singlet, with new sets of peaks around 3.5–4\nppm, which were assigned to CH 2 protons at the 9 position\nof the acridine ring ( Figure S33 ). On the\nbasis of further NMR analysis and our previously reported observations\nwith the Ru-Acr system, these two complexes were identified as the\ndearomatized 9H acridine complex ( 9 ) and Ru-Acr9H phenyl-carboxylate\ncomplex ( 9a ) ( Figure 3 a). 47 Thus, under the reaction\nconditions, hydride transfer from the substrate initially takes place\nto the 9CH position of the catalyst’s acridine backbone, leading\nto the formation of dearomatized complexes, 50 , 51 which further catalyze the reaction. Figure 3 Mechanistic and control\nexperiments. (a) Reactivity of complex 6 with benzaldehyde\nin the presence or absence of NaOH. (b)\nComplex 9 catalyzed furfural oxidation to furoic acid.\n(c) Reactivity of complex 9 with furfural and water.\n(d) Reactivity of 9 with furoic acid to generate the\nfuroate complex 9b . Additional reactivity of 9b in the presence of base NaOH and water. No reaction took place without\nthe base. (e) Catalytic activity of complex 9b in catalyzing\nfurfural to furoic acid in the presence of NaOH. Accordingly, complex 9 catalyzed the dehydrogenative\noxidation of furfural under basic conditions with similar activity\nas the aromatized complex 6 (furoic acid yield 98%, Figure 3 b). Subsequent mechanistic\nexperiments focused on the reactivity of complex 9 with\ndifferent substrates. In the absence of base, complex 9 reacts with water and furfural at 130 °C to generate the furoate\ncomplex 9b after 10 min ( Figure 3 c, Figure S35 ). 9b can also be accessed alternatively by mixing 9 with furoic acid in THF at room temperature ( Figure 3 d, Figure S36 ). 52 9b was found to be stable under\nneutral conditions in THF in the presence of water (10 equiv), even\nwhen subjected to high temperature (130 °C) ( Figure S37 ). On the other hand, when NaOH (3 equiv) was added\nto the solution, the formation of a new complex was observed at RT\nin the 31 P NMR, along with the generation of sodium furoate\n(observed in 1 H NMR) ( Figure 3 d). The new complex slowly decomposed at\nRT, which was facilitated at elevated temperature ( Figure S38 ) and is tentatively assigned the structure of high-energy\nhydroxide intermediate 9c . The furoate complex 9b was observed to catalyze the aqueous oxidation of furfural\nin the presence of an external base; however, its catalytic activity\nsubsided when no base was present in the system ( Figure 3 e). Thus, complex 9b seemingly acts as a deactivating species under the reaction conditions,\nand the addition of a stoichiometric base is required to remove the\nchelating furoate ligand to generate the product while at the same\ntime opening relevant coordination sites for catalytic turnover. On the basis of these observations, a mechanism cycle as shown\nin Figure 4 is proposed.\nInitial hydride transfer from the substrate to the catalyst ligand\nbackbone generates the dearomatized complex 9 (step i).\nComplex 9 , in the presence of water, generates the hydroxide\ncomplex 9c with H 2 evolution (step ii). 50 In the presence of furfural, the hydroxo complex\ngenerates the hemiacetal complex 9d via attack of the\nhydroxide ligand onto furfural (step iii), akin to the mechanism proposed\nat heterogeneous metal water interfaces. 53 − 55 Further β-hydride\nelimination generates the furoic acid complex (step iv), H 2 liberation leading to the furoate complex 9b (step\nv). In the absence of an external base, the catalytic cycle halts\nat this stage; however, when a base is present, the furoate ligand\ndetaches as product furoate salt while regenerating the hydroxide\ncomplex (step vi). It should be noted here that an alternate mechanism\ninvolving the initial free acetal formation, followed by acetal dehydrogenation\nand beta hydride elimination cannot be entirely ruled out, based on\nour mechanistic observations ( SI, section 7.5 ). Figure 4 Mechanistic cycle. A plausible scheme for furfural to furoic acid\nformation using water catalyzed by 6 involving the generation\nof dearomatized complexes. Elemental steps: (i) initial dearomatization\nof the catalyst, (ii) initial dehydrogenation of water by 9 generating hydroxide complex, (iii) hydroxide attack on the aldehyde,\n(iv) beta hydride elimination, (v) H 2 evolution, and (vi)\nproduct elimination and substitution. Simultaneous to the aldehyde dehydrogenative coupling with water,\nleading to generation of the furoate salt and H 2 , disproportionation\nof furfural also takes place to generate furoate and furfuryl alcohol\nas the reaction products. These two processes result in the quick\nconsumption of the initial aldehyde during the reaction. Total consumption\nof the aldehyde was observed after the initial 15 min of reaction\nalong with 60% of furoic acid and 20% H 2 yield (40% of\nfurfuryl alcohol side product; Figure 5 A). The subsequent reaction completion involves the\nconversion of the generated furfuryl alcohol to furoic acid. The observed\nH 2 evolution time profile suggests that the aldehyde dehydrogenative\ncoupling reaction with water is quick, with alcohol dehydrogenation\nbeing comparatively slower ( Figure S16 ).\nSimilar to furfural, HMF oxidation is also surmised to proceed involving\na combination of direct dehydrogenative oxidation to FDCA and disproportionation–oxidation\npathway involving BHMF. Accordingly, when BHMF was tried as a substrate\ninstead of 5-HMF, FDCA in high yields (81%) was isolated (similar\nreaction conditions as in Figure 2 G, entry 4) (see Supporting Information ). The reaction pathways ongoing during the reactions of furfural\nand HMF oxidation are detailed in Figure S45 . Figure 5 Active decomposition suppression by Ru complexes. (a) Product distribution\nfrom furfural oxidation after 15 min with and without the presence\nof catalyst. Furfural (0.5 mmol), water (1 mL), 6 (1\nmol %), NaOH (1.2 equiv), and 1,4-dioxane (2 mL). (b) Catalytic formation\nof furfuryl furoate by Tishchenko coupling of furfural catalyzed by 9 . Furfural (0.5 mmol), 9 (1 mol %), and 1,4-dioxane\n(2 mL). (c) Alternative [Ru] and base-mediated pathway for furfural\ndisproportionation under the reaction conditions as compared to Cannizzaro\ndisproportionation. Active Decomposition Inhibition\nby [Ru] The ruthenium\ncomplex takes an active part in the substrate disproportionation process.\nAs mentioned, under standard reactions with complex 6 , furfural is entirely consumed in 15 min via combined dehydrogenative\naqueous oxidation (to acid and H 2 ) and disproportionation\n(to acid and alcohol) to produce furoate and furfuryl alcohol ( Figure 5 A), selectively.\nInterestingly, when the reaction was conducted without catalyst, only\n55% furfural conversion was observed after 15 min, with only 20% accounting\nfor acid and alcohol products by Cannizzaro disproportionation, the\nrest being unidentified decomposition products ( Figure 5 A). These results suggest a parallel disproportionation\npathway for furfural, in addition to the Cannizzaro mechanism, in\nthe presence of Ru catalyst. Accordingly, when furfural was heated\nat 135 °C in 1,4-dioxane in the presence of complex 9 (1 mol %) for 3 h, formation of furfuryl furoate in 65% yield was\nobserved, signifying that complex 9 can catalyze the\nTishchenko coupling of furfural ( Figure 5 B). 51 , 56 The resulting ester,\nunder the reaction conditions of Table S1 is hydrolyzed to furoate salt and furfuryl alcohol ( Figure 5 C). This alternative disproportionation\npathway via Tishchenko coupling followed by base-mediated hydrolysis\nleads to quick consumption of all furfural/HMF before the onset of\ndecomposition and is crucial for the observed high oxidation selectivity\nwith catalyst 6 or 9 . A tentative mechanism\nof the Tishchenko reaction catalyzed by complex 9 is\nshown in Supporting Information (Figure S40) . Catalyst Recycling and Scale-Up Focusing on the practicality\nof the system for large-scale implementation, we also explored the\npossibility of catalyst recycling after the reaction. The catalyst\nwas recovered from the postreaction solution by evaporation of the\nsolvents and extracting the catalyst with benzene (detailed procedure\nin Supporting Information ). Following this\nprotocol, the catalyst was recycled for three cycles and its catalytic\nactivities for furfural oxidation were retained after the third cycle.\n(97%, 91%, and 83% sodium furoate yield, respectively, in the first,\nsecond, and third cycle), demonstrating the viability of catalyst\nrecycling. We also conducted a gram-scale experiment with 15 mmol\nof furfural (1.44 g) to check the scalability of the process. After\n68 h of reaction at 150 °C, 1.3 g of furoic acid (77% yield)\nwas isolated from the reaction, along with 316 mL of H 2 collected (87% yield), demonstrating the scalability of the process."
} | 6,170 |
21227922 | PMC3082886 | pmc | 3,007 | {
"abstract": "Caldicellulosiruptor bescii DSM 6725 utilizes various polysaccharides and grows efficiently on untreated high-lignin grasses and hardwood at an optimum temperature of ∼80°C. It is a promising anaerobic bacterium for studying high-temperature biomass conversion. Its genome contains 2666 protein-coding sequences organized into 1209 operons. Expression of 2196 genes (83%) was confirmed experimentally. At least 322 genes appear to have been obtained by lateral gene transfer (LGT). Putative functions were assigned to 364 conserved/hypothetical protein (C/HP) genes. The genome contains 171 and 88 genes related to carbohydrate transport and utilization, respectively. Growth on cellulose led to the up-regulation of 32 carbohydrate-active (CAZy), 61 sugar transport, 25 transcription factor and 234 C/HP genes. Some C/HPs were overproduced on cellulose or xylan, suggesting their involvement in polysaccharide conversion. A unique feature of the genome is enrichment with genes encoding multi-modular, multi-functional CAZy proteins organized into one large cluster, the products of which are proposed to act synergistically on different components of plant cell walls and to aid the ability of C. bescii to convert plant biomass. The high duplication of CAZy domains coupled with the ability to acquire foreign genes by LGT may have allowed the bacterium to rapidly adapt to changing plant biomass-rich environments.",
"conclusion": "CONCLUSIONS Caldicellulosiruptor bescii is the most thermophilic anaerobic bacterium capable of utilizing cellulose as well as multiple polysaccharides and unprocessed plant biomass. From an analysis of its genome, coupled with transcriptomic and proteomic data, we suggest that not one particular feature but a combination of properties that act in synergy enables the bacterium to degrade various polysaccharides and plant biomass:\n Enrichment in multi-modular, multi-functional CAZy proteins each containing two catalytic modules specific to different components of plant cell walls combined with multiple CBMs. Presence of three PLs of different families absent from the genome of C. saccharolyticus. Concentration of all multi-modular, multifunctional CAZy genes including three PLs and all CBM3s in one large functional gene cluster. Multiplication of CAZy modules within the large CAZy gene cluster increasing ‘dosage’ of particular CAZy modules, in particular, three cellobiohydrolase GH48s in combination with CBM3s. Absence of modules with motifs of dockerin or cohesin domains, which mediate the assembly of the cellulosome. This confers C. bescii with more flexibility to produce combinations of ‘free’ enzymes to degrade a variety of insoluble polysaccharides. Binding of C. bescii to xylan and switchgrass is mediated by proteins containing conserved non-CAZy modules known to bind polysaccharides or cell wall components, and proteins with CBM and membrane-binding modules. This binding is dynamic, in contrast to the irreversible binding of the cellulosome to cellulose. Hypothetical/conserved hypothetical genes located inside or in the vicinity of CAZy or sugar transport genes/operons are related to plant biomass conversion. Hypothetical/conserved hypothetical genes that are highly regulated during growth on polysaccharides or related substrates are likely involved in plant biomass conversion. \nCurrently there is an increased interest in members of the Caldicellulosiruptor genus that display the ability to degrade multiple polysaccharides as well as plant biomass. Like the prototypical cellulose-degrader, C. thermocellum , these bacteria have a high potential for use in efficient two-step biomass-sugar–biofuel conversion processes. The data presented here are a valuable source of information that can be utilized for further characterization of the Caldicellulosiruptor species that will lead to a deeper understanding of the mechanisms of the non-cellulosomal plant biomass conversion process.",
"introduction": "INTRODUCTION Lignocellulosic plant biomass is the most abundant renewable alternative to petroleum as a source of fuel ( 1 ). It consists mainly of cellulose and hemicellulose in combination with up to 20% lignin. Biological conversion of this chemically and physically complex material, represents a major challenge ( 2 , 3 ). Expensive thermal and chemical pretreatments are needed to decrease its recalcitrance and expose the polysaccharides to carbohydrate-active enzymes (CAZy) and carbohydrate-binding modules (CBMs) that help destroy the plant cell walls ( 4 , 5 ). Despite intensive studies, many aspects of microbial and enzymatic biomass-to-biofuel conversion are still not understood. Thermophilic anaerobic bacteria hold great promise as they display higher bioconversion rates, minimize the risk of contamination, facilitate product recovery and synthesize highly thermostable enzymes ( 6 , 7 ). However, only a relatively small number of anaerobic thermophiles are able to convert crystalline cellulose into soluble fermentable sugars, and only a few of them are able to metabolize simultaneously the hexose and pentose sugars that are produced from cellulose and hemicellulose, respectively ( 1 , 8 ). One of the best studied of the cellulolytic microbes is Clostridium thermocellum , which grows optimally at 60°C ( 9 ). It produces ethanol and is being used for the consolidated bioprocessing of plant biomass ( 6 , 8–10 ). Its cellulolytic system is a large multi-protein complex called the cellulosome, the enzymatic components of which act synergistically to degrade crystalline cellulose ( 9–11 ). The recent availability of genetic systems in C. thermocellum and a related thermophile ( 12 ) provides a much needed tool to investigate the mechanisms of cellulose degradation. Several members of the genus Caldicellulosiruptor are able to degrade cellulose at even higher temperatures (up to 90°C) and they also utilize pentose sugars ( 13–18 ). The genomes of C. saccharolyticus DSM 8903 ( 19 ) and C. bescii DSM 6725 have been sequenced ( 20 ) and some CAZy enzymes have been purified and characterized from both species ( 21 , 22 ). Representatives of this genus have potential utility in biomass-to-sugars conversion processes but more comprehensive studies are needed to understand the degradative mechanisms involved. Caldicellulosiruptor bescii grows at temperatures up to 90°C and is the most thermophilic bacterium capable of growth on crystalline cellulose ( 16 ). It also utilizes xylan, pectin and starch and is also able to grow efficiently on untreated plant biomass with high lignin content ( 14 , 16 ). The bacterium is capable of using cellulose and xylan simultaneously. Its ability to grow on the hardwood poplar is of particular interest as this hardwood can be genetically manipulated to potentially decrease recalcitrance ( 23 ). For example, one transgenic poplar line overexpressing xyloglucanase is less recalcitrant to cellulolytic enzymes ( 24 ). In the present article we analyze the genome of C. bescii with a particular focus on genes encoding enzymes involved in plant biomass conversion. We also present transcriptomic and proteomic data and compare its genome with those of other anaerobic thermophiles, including its close relative C. saccharolyticus . This analysis will contribute to our understanding of plant biomass conversion at extreme temperatures and will provide a genetic basis for the plant biomass-degrading properties displayed by this remarkable organism.",
"discussion": "RESULTS AND DISCUSSION General features and comparative genomics The C. bescii genome contains a 2 919 718 bp circular chromosome with 35.2% GC content and is slightly smaller than the size of the average bacterial genome (3.3 Mb; www.ncbi.nlm.nih.gov/genomes/lproks.cgi ). It contains two native circular plasmids termed AX710673 pBAL (8294 bp, 38.5% GC) and AX710687 pBAS2 (3653 bp, 42.9% GC), both of which were isolated and sequenced previously ( 35 ). The sequence of AX710687 pBAS2 reported here is identical but that of AX710673 pBAL has 8 deletions, 11 insertions and 7 mismatches ( Supplementary Figure S2 ). AX710673 and AX710687 encode eight and four open reading frames (ORFs), respectively. AX710687 encodes exclusively uncharacterized proteins, while AX710673 encodes two putative regulators and two proteins involved in nucleic acid metabolism. The chromosome is predicted to contain 2654 protein coding genes. Their arrangement on the two strands suggests that it has two equal replicores with a positive correlation between the direction of transcription and replication ( Figure 1 ). The 16 S RNA sequences confirmed that what was formerly termed Anaerocellum thermophilum is a member of the phylum Firmicutes , class Clostridia , order Clostridiales and that it should be classified in the genus Caldicellulosiruptor ( 15 ). C. hydrothermalis and C. kronotskiensis are the closest known relatives of C. bescii and C. saccharolyticus DSM 8903 is the closest relative with a sequenced genome ( 15 , 19 ).\n Figure 1. Diagram of C. bescii chromosome. From outside to inside, the circles show (i) COG categories (two circles), (ii) mean centered GC content of the genome, (iii) genes (two circles) with functions related to CAZy (green), sugar transporters (red) and cell-adhesion (blue), (iv) GC skew plot(orange-purple circle) and (v) RNA genes (ribosomal in red, tRNA in blue and others in aquamarine). The GenomeViz was used to construct the circular chromosome wheel ( http://www.uniklinikum-giessen.de/genome/index.html ). We compared the general features of the C. bescii genome to those of five anaerobic thermophiles containing significant numbers of CAZy-related genes potentially involved in plant biomass degradation ( Table 1 ): C. saccharolyticus DSM 8903, C. thermocellum ATCC 27405, Thermotoga maritima MSB8, Thermoanaerobacter pseudethanolicus ATCC 33223 and T. tengcongensis MB4. The genome size of C. bescii is similar to that of C. saccharolyticus (2.97 Mb) ( 19 ). Both utilize crystalline cellulose and xylan and are very closely related with over 2300 C. bescii genes having as their top Blast hit in the C. saccharolyticus sequence. The C. bescii genome is smaller than that of the cellulolytic but not xylanolytic C . thermocellum ATCC 27405 ( T opt 60°C, 3.8 Mb) ( 36 ), and larger than the genome of the xylanolytic but not cellulolytic T. maritima MSB8 ( T opt 80°C, 1.86 Mb) ( 37 ). By 16 S rRNA analysis, C. bescii is closely related to the Thermoanaerobacter genus. Its genome is most similar in size to that of T. tengcongensis MB4 ( T opt 75°C, 2.7 Mb) ( 38 ) and slightly smaller than that of T. pseudethanolicus ATCC 33223 (formerly T. ethanolicus strain 39E, T opt 65°C, 2.4 Mb). Both C. bescii and C. saccharolyticus grow on polysaccharides such as starch but do not utilize cellulose or xylan.\n Table 1. General features and comparative genomics of C. bescii DSM 6725 General features Caldicellulosiruptor bescii DSM 6725 Caldicellulosiruptor saccharolyticus DSM 8903 Thermoanaerobacter pseudethanolicus ATCC 33223 Thermoanaerobacter tengcongensis MB4 Clostridium thermocellum ATCC 27405 Thermotoga maritima MSB8 Length of chromosome (Mbp) 2.9 3.0 2.4 2.7 3.8 1.9 G + C content (%) 35.2 35.3 34.5 37.6 39.0 46.2 Coding density (%) 85.4 86.1 86.7 86.8 83.5 86.8 Total no. of predicted protein coding genes 2662 2679 2243 2588 3191 1858 Average length of protein-coding genes (bp) 942 957 915 905 1008 905 Total no. of predicted tRNA 47 46.0 56 55 56.0 46.0 Total no. of rRNA genes (no. of operons) 9 ( 3 ) 9 ( 3 ) 16 ( 4 ) 12 ( 5 ) 12 ( 4 ) 3 ( 3 ) Secreted Proteome (SignalP prediction) 394 362 244 257 404 207 Membrane Proteins (TMHMM prediction) 344 348 282 401 481 239 Percentage secreted proteome (SignalP prediction) 14.80 13.51 10.88 9.93 12.66 11.14 Percentage membrane proteins (TMHMM prediction) 12.92 12.99 12.57 15.49 15.07 12.86 IS elements Full copies 34 91 42 69 100 3 Partial copies 132 130 77 104 56 22 Ismax-full-copy a eISCsa4 eISCsa4 eISTps4/eISTps5 eISTps1 ISCth3 eISTma2 MaxCopy no. b 12 33 2 17 18 2 Growth on cellulose and xylan c Cellulose, xylan Cellulose, xylan Does not grow Does not grow Cellulose Xylan, CMC d a ISmax is the IS element with the largest full copy number. b MaxCopy no. is the largest full copy number. c See text for references. d CMC: carboxymethyl cellulose. Using the COG approach to predict gene function ( 39 ), we analyzed the genomes of 41 anaerobic thermophiles ( Supplementary Tables S2 and S3 ). The genome of C. bescii is significantly enriched in genes encoding proteins involved with cell motility and secretion (COG group N) and cell division (group D), in agreement with the SignalP result suggesting that C. bescii has a large number of secreted proteins. Interestingly, the cellulolytic C. thermocellum ATCC 27405 has a significantly higher than average number of genes responsible for DNA replication, recombination or repair, which may have facilitated the development of a genetic system for this organism ( 28 ). On the other hand, the C. thermocellum genome appears to have fewer genes involved in intracellular trafficking and in defense mechanisms. The genome of C. bescii has a lower number of uncharacterized genes (groups R and S) and a higher percentage of genes not assigned to COG categories. Of the 2666 proteins encoded by the C. bescii genome, 394 (14.8%) and 344 (12.9%) are predicted to have signal peptides and transmembrane helices, respectively. Using a previously-developed program ( 32 ), it was found that the 2666 genes in C. bescii are predicted to be organized into 1209 transcriptional units, 577 of which are multi gene. The 259 genes with functions related to carbohydrate metabolism and sugar transport were predicted to form 180 transcriptional units, 111 of which are multi-gene and 69 are single-gene operons ( Supplementary Table S4 ). Furthermore, C. bescii (and its close relative C. saccharolyticus ) is predicted to contain 14 [12] sigma factors, 8 [7] anti-sigma modulators, 97 [60] putative transcription regulators and 18 [25] histidine kinases. Among the 18 putative histidine kinases, 11 are predicted to be membrane-associated. The high number of putative regulators suggests that C. bescii is highly responsive to changing environmental conditions and nutrient availability. This is in line with a recent report showing that the composition of C. thermocellum cellulosome varies with the growth substrate ( 40 ). Genome dynamics Insertion sequences (IS) have been found to be actively involved in the genomic recombination and horizontal gene transfer events in prokaryotic genomes ( 41 ). The coding region of an IS is flanked by fixed-length non-coding terminal regions, which are essential in mediating transposition and genomic recombination ( 31 , 42–44 ). Unfortunately, in many cases genome annotations include only the potential coding sequences carried by the elements and ignore their terminal regions. The statistics of IS elements in six genomes of anaerobic thermophiles are summarized in Table 1 (see also Supplementary Table S5 ). Thermotoga maritima harbors the smallest number of IS elements, with only 3 full copies and 22 partial copies. Caldicellulosiruptor bescii also harbors much fewer full IS copies [34] than the other four bacteria, especially compared to its closest relative C. saccharolyticus [91]. Full copies of IS elements are typically the results of recent proliferation. In light of these data, the genome of C. bescii is probably more stable than the other genomes, except for that of T. maritima. The presence of IS elements suggests that all of the organisms listed in Table 1 likely have a history of horizontal gene transfer events ( Supplementary Table S5 ). Accordingly, the C. bescii genome has multiple sequences that are much more closely related to those in other genomes than they are to those in C. saccharolyticus , suggesting that these regions are the results of such events. As shown in Supplementary Figure S3 and Table S6 , these include three of the thermophilic organisms listed in Table 1 , C. thermocellum (23 genes), T. tengcongensis (21 genes) and T. pseudethanolicus (18 genes), as well as Petrotoga mobilis (11 genes), Thermoanaerobacter sp . X514 (11 genes) and Dictyoglomus thermophilum (14 genes). In addition, eleven C. bescii genes show the highest similarity to those in C. phytofermentans a mesophilic anaerobe that, like C. bescii , is both celluloytic and xylanolytic. These ‘horizontally transferred’ genes in C. bescii are predicted to encode ABC transporters [25], carbohydrate-active enzymes (CAZy) [17], mobile-element related [18], signal transducers and DNA binding (all containing a helix-turn-helix motif: 15), and genes encoding domains of unknown function like conserved domain UPF0236 [7], KWG leptospira repeat [6] and a radical SAM domain [6], many of which may be involved in various catabolic and anabolic pathways ( 45 ). Distribution of CAZy genes within genomes of anaerobic thermophiles The distribution of CAZy genes ( http://www.cazy.org ) related to plant biomass degradation within the genomes of the 41 anaerobic thermophiles is shown in Supplementary Tables S2 and S7 . Glycoside transferases were not considered in this group as they are mainly involved in the biosynthesis of polysacchardies. Among these thermophiles, the 16 genomes of the archaea encode very few CAZy proteins. They do not contain polysaccharide lyases (PLs) and 13 of the 16 genomes do not encode CBMs, which are critical for degradation of insoluble polysaccharides. Six of the archaeal species grow on starch, although three of them are not predicted to contain genes that encode CBMs. Three of the genomes encode CBMs, glycosyde hydrolases (GHs) and carbohydrate esterases (CEs). Two of them, Pyrococcus furiosus and Thermococcus kodakaraensis , do not grow on cellulose or xylan, but do grow on starch, while the other, Thermofilum pendens , does not grow on any polysaccharide that has been examined although its genome encodes several GHs, CBMs and CEs. The presence of two GH13s, the recombinant forms of which are amylolytic enzymes, suggests that this organism can grow on starch or cyclodextrins ( 46 ). In contrast to the anaerobic thermophilic archaea, all of the genomes of the 25 anaerobic thermophilic bacteria encode CBMs, GHs and CEs ( Supplementary Tables S2 and S7 ). However, in many cases growth of these organisms on components of plant biomass has not been reported. Starch is the most common polysaccharide to be used by this group and 14 of them, including C. bescii , have this ability and all contain α-amylase-type enzymes (GH13). PLs are identified in eight of these bacteria, and five of them have been shown to grow on pectin, including C. bescii , C. saccharolyticus , C. thermocellum , T. lettingae and T. maritima . The genome of C. saccharolyticus does not contain PLs but it does encode two GH28s that are putatively involved in hydrolysis of pectin backbone. Based on the numbers of CBMs and GHs that they contain, these anaerobic thermophilic bacteria ( Supplementary Table S2 ) can be classified into three groups wherein (i) both CBMs and GHs are low (7 genomes); (ii) CBMs are low but GHs are high (15 genomes) and (iii) both CBMs and GHs are high (3 genomes). The latter category includes C. bescii , C. saccharolyticus and C. thermocellum . All representatives of the Caldicellulosiruptor genus grow on cellulose, xylan, pectin and starch ( 13 , 16 ). C. thermocellum does not grow on xylan as it cannot consume xylose, however, it depolymerizes xylan into xylose, xylobiose and xylooligosaccharides ( 9 ). Consequently, there is a clear correlation between the number of representatives of CAZy genes in a genome and the plant biomass-degrading abilities of a microorganism. Comparison of CAZomes of C. bescii and C. saccharolyticus The modular architecture of the 88 CAZy genes in C. bescii is shown in Supplementary Table S8 . There is a comparable number of such genes in C. saccharolyticus [94]. Other common characteristics include (i) a similar module arrangement for CAZy-related proteins that do not contain CBMs, (ii) all proteins containing CBMs are predicted to be extracellular based on the presence of signal peptides (extracellular CAZy proteins in C. bescii are shown in Table 3 ); and (iii) the major CBM3s present in the enzymes from both organisms are derived from subfamilies 3a and 3b. CBM3a/3b bind tightly to crystalline cellulose and thus enhance the access of cellulases to their substrate relative to other cellulose-directed CBMs ( 47 , 48 ). In this respect, these two Caldicellulosiruptor species are similar to cellulolytic clostridia that produce cellulosomes, where CBM3 plays a pivotal role in substrate targeting of their respective cellulase complexes ( 10 , 11 ). The clostridial enzymes generally contain additional CBMs that direct the cellulose-tethered complex to specific regions of the cell wall, consistent with the activity of the enzyme containing these additional targeting modules ( 49 , 50 ). In contrast, C. bescii and C. saccharolyticus contain fewer of these additional, non-crystalline cellulose-binding CBM families ( Supplementary Tables S2 and S8 ). The most significant of these are five CBM22s, and one CBM36 that likely targets xylan ( Supplementary Table S8 ). Within this context it should be noted that CBM22s bind tightly to isolated xylan chains but not to hemicellulose within the plant cell wall ( 51 ). Thus, CBM22-containing enzymes likely target xylans that have been released from the plant cell wall. It appears, therefore, that CBM3s work in both of these bacteria as the primary mechanism for the attachment of enzymes to plant polysaccharides. Furthermore, the majority of CBM3-containing enzymes contain multiple CBM3s. These are likely to confer extremely tight binding to cellulose to offset the dissociation promoted by elevated temperatures. Indeed, it has been suggested that there is a general correlation between the growth temperature at which an organism and the frequency of finding enzymes with multiple CBM copies ( 52 ).\n Table 2. Distribution of COG within some genomes of anaerobic thermophiles Computed frequency a P -value b Average Function COG Cbes Csac Teth TTE Cthe Tmar Cbes Csac Teth TTE Cthe Tmar A 0.00 0.00 0.00 0.00 0.00 0.00 0.240 0.240 0.240 0.240 0.240 0.240 0.056 RNA processing B 0.00 0.11 0.00 0.10 0.05 0.07 0.167 0.390 0.167 0.388 0.747 0.304 0.149 Chromatin structure C 6.08 5.98 6.67 6.91 6.02 8.16 0.069 0.063 0.892 0.871 0.065 0.726 9.581 Energy D 2.75 1.96 2.60 2.25 1.80 1.54 0.018 0.768 0.967 0.108 0.679 0.499 1.537 Cell division E 10.18 9.69 11.98 12.41 8.42 13.54 0.202 0.141 0.480 0.395 0.044 0.207 11.876 Amino acids F 3.52 3.12 3.78 3.30 3.10 3.70 0.373 0.181 0.478 0.742 0.172 0.524 3.739 Nucleotides G 12.29 12.18 10.32 9.22 7.40 12.28 0.099 0.895 0.750 0.629 0.406 0.099 8.169 Carbohydrates H 5.63 4.87 5.66 3.40 4.58 3.98 0.394 0.245 0.600 0.068 0.198 0.881 6.126 Coenzymes I 1.79 2.01 2.24 3.04 2.22 2.30 0.168 0.222 0.711 0.436 0.282 0.308 2.860 Lipids J 8.90 7.99 9.03 8.17 7.77 9.42 0.311 0.220 0.675 0.763 0.201 0.631 10.513 Translation K 8.45 8.31 7.73 8.43 8.75 5.93 0.158 0.813 0.663 0.161 0.107 0.152 7.207 Transcription L 8.13 13.29 10.44 9.11 14.67 6.28 0.485 0.992 0.134 0.689 0.001 0.207 8.047 DNA M 5.89 5.88 5.72 5.81 7.59 5.16 0.378 0.619 0.584 0.605 0.096 0.454 5.359 Cell membrane N 8.51 3.76 3.07 3.72 4.35 4.12 0.016 0.598 0.489 0.592 0.314 0.348 3.136 Cell motility and secretion O 3.84 3.18 3.78 4.19 4.07 3.77 0.181 0.036 0.159 0.671 0.726 0.157 4.521 Posttranslational modification P 4.87 4.39 5.78 6.34 4.77 8.37 0.077 0.038 0.774 0.635 0.067 0.126 6.809 Inorganic ions Q 3.59 1.32 1.24 1.78 1.48 1.47 0.075 0.202 0.179 0.646 0.750 0.245 2.153 Secondary metabolites T 5.57 6.88 5.13 6.86 8.28 5.09 0.383 0.785 0.553 0.783 0.094 0.547 4.779 Signal transduction U 0.00 2.12 3.01 2.20 2.31 2.51 0.033 0.695 0.067 0.273 0.767 0.171 1.658 Transport V 0.00 2.81 1.83 2.72 2.27 2.30 0.035 0.878 0.449 0.859 0.276 0.736 1.707 Defense mechanism Z 0.00 0.16 0.00 0.00 0.09 0.00 0.310 1.000 0.310 0.310 0.022 0.310 0.018 Cytoskeleton R 9.30 11.00 10.79 11.63 10.00 13.02 General prediction only S 5.20 6.61 7.94 6.88 5.77 17.00 Function unknown Not in COG 32.11 18.48 13.64 14.61 22.23 9.85 Not assigned a Frequency was computed as percentage of genes assigned to each COG group among all genes with COG assignment. When a gene was assigned to multiple COG groups, it would be counted multiple times. b The P -value was calculated based on the assumption that the distribution of the frequence in each COG group follows a normal distribution. Cbes, Caldicellulosiruptor bescii DSM 6725; Csac, Caldicellulosiruptor saccharolyticus DSM 8903; Teth, T. pseudethanolicus ATCC 33223; TTE, T. tengcongensis MB4; Cthe, C. thermocellum ATCC 27405; Tmar, Thermotoga maritima MSB8. \n Table 3. Primary extracellular proteins of C. bescii involved in utilization of insoluble components of plant biomass Gene CAZy module architecture CAZy module activity ( www.cazy.org ) Transcriptomics Proteomics CBM Catalycic (main activities) Cell. Signif. ExtP Membr. Cbes_0089 GH11-CBM36 Xylan Xylanase Up Yes Cbes_0182 GH43-CBM22 a -GH43-CBM6 b Xylan a,b , amorphous cellulose b Xylanase, β-xylosidase, arabinanase Up Yes Cbes_0183 CBM22-CBM22-GH10 Xylan Endo-1,4-, endo-1,3-β-xylanase Up Yes Cbes_0458 GH1 β-glucosidase, β-galactosidase, β-mannosidase, β-glucuronidase Up No C Cbes_0594 GH5-CBM28-SLH-SLH-SLH Amorphpus cellulose, cellooligosaccharides Mannanase, cellulase, lichenase, xylanase Up Yes C Cbes_0609 CBM41 a -CBM48 b -GH13-CBM20 c Starch a,c , glycogen b , cyclodextrines c Starch Up Yes C C Cbes_0610 CBM20 Starch, cyclodextins Up Yes CX Cbes_0618 CBM22-CBM22-GH10 Xylan Endo-1,4-, endo-1,3-β-xylanase Up Yes X Cbes_1439 GH23 Peptidoglycan lyase Down Yes Cbes_1462 CE4 Acetyl xylan esterase Up Yes Cbes_1829 CE4 Acetyl xylan esterase Up No Cbes_1853 PL11-CBM3 Cellulose Rhamnogalacturonan lyase Down No CX Cbes_1854 CBMX-PL3 Pectate lyase Down Y/N CX X Cbes_1855 CBMX-PL9 Pectate lyase, exopolygalacturonate lyase Down Yes CX Cbes_1857 GH10 a -CBM3-CBM3-GH48 b Cellulose endo-1,4-, endo-1,3-β-xylanase a , cellobiohydrolase b Up Yes CX X Cbes_1859 GH5 a -CBM3-CBM3-GH44 b Cellulose Mannanase a ; Xyloglucanase, endoglucanase b Up Yes CX X Cbes_1860 GH74 a -CBM3-CBM3-GH48 b Cellulose Xyloglucanase, endoglucanase a ; cellobiohydrolase b Up Yes CX X Cbes_1865 GH9 a -CBM3-CBM3-CBM3-GH5 b Cellulose Endoglucanase a ; mannanase b Up No CX Cbes_1866 GH5 a -CBM3-CBM3-CBM3-GH5 b Cellulose Mannanase a , cellulase b Down Y/N CX Cbes_1867 GH9 a -CBM3-CBM3-CBM3-GH48 b Cellulose Endoglucanase a , cellobiohydrolase b Up Yes CX CX Cbes_2593 GH13 Starch Up Yes Primary extracellular proteins are CAZy proteins where each contains a signal peptide and, in most cases, a CBM. The superscripts on the CBM and GH domains (a, b or c) indicate the corresponding CAZy module activity. Transcriptomics and proteomics show regulation of gene transcription on cellulose versus glucose, and protein identification using LC-MS/MS. N-terminal GH5 modules in Cbes_1859, Cbes_1865 and Cbes_1866 are identical to N-terminal GH5 module of Csac_1077, and C-terminal module in Cbes_1866 is identical to the C-terminal module in Csac_1077 which has been experimentally shown to display mannanase and cellulase/lichenase activities, respectively ( 40 ). TMD, transmembrane domain; Membr., membrane protein fraction; C, cellulose; X, xylan; CBMX, an unknown module possibly pectin binding. Notably, the CBM3s in the genomes of both C. bescii and C. saccharolyticus are concentrated only in one gene cluster and this encodes mainly CAZy proteins (Cbes_1853-_1867 and Csac_1076-_1085: see Figure 2 and Supplementary Figure S4 ). However, there is a significant difference in the arrangement of these gene clusters. In C. bescii this cluster is enriched in CBM3s, which are present as double or triple modules within one gene product, in comparison to the cluster of C. saccharolyticus (16 versus 10). Specifically, C. bescii contains three genes encoding PLs of different families that are absent from C. saccharolyticus . Moreover, of all thermophilic anaerobes, only C. bescii has PLs of three different families ( Supplementary Table S2 ). The C. bescii cluster also contains three GH48s versus one in C. saccharolyticus . The GH48s are key enzymes in crystalline cellulose hydrolysis and are uniquely arranged in C. bescii. There is no other known example of three modules of this type in combination with a second catalytic module of different CAZy activity ( Figure 2 A and B). Interestingly, a deletion mutant of C. thermocellum lacking two GH48s was able to completely hydrolyze crystalline cellulose, albeit at a slower rate than the wild-type ( 28 ). This cluster in C. bescii also has three GH5 mannanases (versus one gene in C. saccharolyticus ) and six genes encoding multifunctional CAZy proteins (versus three genes in C. saccharolyticus ), each containing two catalytic modules of different hydrolytic activity separated by double or triple CBM3s.\n Figure 2. Comparison of the two relative gene clusters involved in biomass conversion in C. bescii DSM 6725 ( A ) and C. saccharolyticus DSM 8903 ( B ). Abbreviations: GH5, GH9, GH10, GH44, GH48 and GH74, glycoside hydrolases of families 5, 9, 10, 44, 48 and 74, respectively; CBM3, carbohydrate-binding module of family 3 where ‘b’ and ‘c’ are of subgroups within CBM3; GT39, glycosyl transferase of family 39; PL3, PL9 and PL11, polysaccharide lyases of families 3, 9 and 11; X, module of unknown function with homology to pfam CBM_4_9, Signal peptides, linkers and fragments of unknown function are shown in violet, blue and grey colors, respectively. CelA, CelB and ManA (encoding by Csac_1076, _1077 and _1078) are enzymes with experimentally demonstrated activities. Consequently, this CAZy-enriched gene cluster in C. bescii uniquely contains CBM3s that potentially mediate the binding of 13 catalytic modules to the insoluble substrate, while in C. saccharolyticus there are only eight catalytic modules attached to CBM3s. The C. bescii gene cluster also contains a GH74 module, which is a putative xyloglucanase ( Table 3 ). This enzyme has an important role in biomass degradation as it hydrolyzes xyloglucan networks ( 53 ). In C. bescii the GH74 enzyme is part of a multi-modular protein with two CBM3s and GH48 (Cbes_1860), the combination of which is predicted to display synergism by binding the two catalytic modules to xyloglucan, hydrolyzing xyloglucan and releasing and hydrolyzing cellulose. In C. saccharolyticus the corresponding gene is truncated to GH74-CBM3 ( Figure 2 A and B). In Cbes_1867, GH48 is combined with GH9 via triplet of CBM3s. The combination of GH9 (endoglucanase) and GH48 (exoglucanase) assumes a synergy in hydrolysis of amorphous and crystalline parts of cellulose. Similarly, Cbes_1857 has the modular structure GH10-CBM3-CBM3-GH48 where GH10 is a xylanase, and the catalytic modules can act in a concert on mixed type xylan/cellulose substrates. In general, the C. bescii gene cluster encodes a powerful set of CAZy enzymes active against major components of plant cell walls (cellulose, xylan, xylogluvan, pectin and mannan). In contrast, the analogous cluster in C. saccharolyticus is significantly truncated and lacks genes encoding some important biomass-related activities. All but of one the CBM3s in the C. bescii gene cluster has >99% identity, the three GH48s are 100% identical, and there are also three GH5s with high degree of sequence identity. Such gene duplication in the main CAZy-containing cluster in C. bescii suggests that both diversity of CAZys, and the ‘dosage’ of individual CAZy are important for this bacterium to adapt to new growth substrates, including various polysaccharides and related materials derived from plant biomass ( 54 ). Our analysis of the CAZy-related genes in C. bescii revealed that the NCBI annotation of several of these sequences is incorrect and/or incomplete. For example, Cbes_1853 is annotated as cellulose 1,4-β-cellobiosidase, Cbes_1857, _1860 and _1867 are annotated as glycoside hydrolases family 48 and Cbes_1865 is annotated as a glycoside hydrolase family 9. Based on comparisons with other sequences in the CAZy database, we propose that Cbes_1853 is a rhamnogalacturonan lyase; Cbes_1857, _1860, _1865 and _1867 are bifunctional enzymes containing GH10/GH48, GH74/GH48, GH9/GH5 and GH9/GH48, respectively. These changes are listed in Supplementary Table S8 . Our new annotations suggest that these genes contain multiple domains, and such combination of multiple domains could be the key to biomass degradation. Sugar transport A total of 257 genes in the C. bescii genome are predicted to encode transporters including 171 involved in sugar transport ( Supplementary Table S9 ). Cellular transport systems can be classified into seven main classes ( http://www.chem.qmul.ac.uk/iubmb/mtp/ ). Although the total number of transporter genes is similar in the genomes of the two Caldicellulosiruptor species, C. saccharolyticus contains 18 more genes of family 3.A.1 that transport organic and inorganic molecules of various sizes, while C. bescii has 11 more genes of family 2.A.1 that transport molecules of small sizes including lactose ( Supplementary Table S10 ). ABC transporters in bacterial genomes are composed of an inner membrane component (IMC) and an ATPase component. In the C. bescii genome ( Supplementary Table S9 ) in most cases the IMCs are paired and encoded by one operon suggesting that the ABC transporter system is tetrameric (two IMCs and two ATPases). The ATPases are typically not linked and are located remotely from the IMCs, suggesting that one ATPase serves multiple IMCs ( 55 , 56 ). Multiple solute-binding proteins (SBPs) were also identified in both genomes. They are generally located close to IMCs, but often are predicted in separate operons. Many SBPs belong to functional category COG1653 that includes putative proteins transporting various oligosaccharides and simple sugars. In many cases sugar transport systems are found in the same operon or in the vicinity of genes encoding CAZy related proteins. This observation suggests that these transporters are involved in the transport of sugars released by the corresponding enzymes encoded by these CAZy related operons or genes. In particular, the gene cluster Cbes_0050-_0063 contains ABC transporters and four glycosyl transferases of families GT2 and GT4 that transfer mannosyl, rhamnosyl, N-acetyl-glucosaminyl, β-galactosaminyl and galactosyl, glucosyl, mannosyl or xylosyl groups, respectively. It seems likely that ABC transporter elements located in the same operon are involved in transport of related sugars. The neighboring operons, Cbes_0174-_0181 and Cbes_0182-_0187 encode elements of ABC transporters and glycoside hydrolases GH43, GH39 and GH10, which encode xylanase, xylosidase and arabinofuranosidases, respectively. Transporter operon Cbes_1107-_1112 is located close to genes Cbes_1103 (GH51 with putative activities endoglucanase or arabinofuranosidase) and Cbes_1104 (GH4 displays activities of α-glucosidase, α-galactosidase and α-glucuronidase) ( Supplementary Table S9 ). These observations imply that genes encoding sugar transport and sugar metabolism are typically closely associated. Comparison of metabolic pathways In comparing the pathways present in C. bescii and C. sacharolyticus assigned by the KEGG database, we found that both genomes are similar in term of the number of genes present in assigned pathways, as shown in Supplementary Table S11 . However, there is one pathway present in C. bescii only. Its genome includes four genes essential for the biosynthesis of deoxythymidine-diphosphate rhamnose (dTDP- l -rhamnose) from glucose-1-phosphate, which is produced from cellobiose by cellobiose phosphorylase ( Supplementary Table S11 and Figure S5 ) . This is of particular interest as the activated sugar donor, glucose-1-phosphate, could be an energy source or could participate in the glycosylation of extracellular proteins and flagella biosynthesis ( 57 , 58 ), particularly since the genome of C. bescii is enriched in genes related to secretion and motility. In some bacteria, arabinogalactan is attached to peptidoglycan via a rhamnose-N-acetylglucosamine disaccharide linker unit ( 59 ) so it is not clear whether this pathway in C. bescii is essential for conversion of components of plant biomass. There is also a difference between the two Caldicellulosiruptor species in alanine metabolism. In particular, C. bescii and C. sacharolyticus contains eight and one copy, respectively, of homologs of alanine racemase (EC. 5.1.1.1), which reversibly converts l -alanine to d -alanine. However, they both contain only a single copy of d -alanine- d -alanine ligase (EC. 6.3.2.4), which converts d -alanine to d -alanyl- d -alanine, an enzyme involved in peptidoglycan metabolism in Gram-positive bacteria. The consequences of this are not clear at present. Caldicellulosiruptor bescii CAZy and sugar transport genes with closest homologs in genomes other than C. saccharolyticus Seventeen CAZy genes in the genome of C. bescii do not have their closest relatives in C. saccharolyticus (based on Blast analysis; see Supplementary Table S12 ). These genes were probably acquired from thermophilic [12] and mesophilic [5] microorganisms. Fourteen of these microorganisms degrade polysaccharides and three of them produce ethanol, but they also include three methanogenic archaea, which are not known to degrade polysaccharides. Ten of the 17 C. bescii genes are organized into three clusters that contain multiple CAZy-related proteins: Cbes_0052-_0061 and Cbes_0154-_0157 are all composed of GTs transferred from mesophiles, Cbes_1853-_1855 encodes PLs enzymes acquired from a thermophile and two mesophiles and Cbes_1853-_1855 was incorporated into a region containing multiple GH and CBM-containing genes. The latter gene cluster is discussed further below. There are also 25 genes related to ABC transporters that do not have their closest relatives in C. saccharolyticus ( Supplementary Table S13 ). It is assumed that these were acquired by lateral gene transfer but in this case only from bacteria. The closest relatives of the 25 genes are found in 17 bacteria, many of which are capable of metabolizing polysaccharides with some generating ethanol as an end product. The ABC transporter genes appear to have been acquired predominantly [12] from mesophiles. Interestingly, 17 of the 25 genes are organized into 5 gene clusters, 3 of which are adjacent to 3 of the CAZy-gene clusters discussed above. In particular, cluster Cbes_2371-_2376 has four of its top Blast hits in C. phytofermentans , an organism that is capable of producing high concentrations of ethanol during cellulose fermentation. The same cluster encodes 3 ABC transporter genes, a GH43, a histidine kinase and a response regulator, suggesting that this six-gene cluster was horizontally transferred more or less intact from a Clostridium species, and may play a significant role in biomass degradation. Similarly, the Cbes_2076-_2094 cluster contains two ABC transporters (six genes), a GH2 and two integrase-related genes. A large number of genes in this cluster have their top Blast hits in two species, B. subtilis and D. thermophilum , indicating that this region could be a hot spot for DNA integration or genome rearrangement in C. bescii . These data suggest that the exchange of genetic information has had a significant impact on the metabolic capabilities of C. bescii , and that this exchange has occurred between very different microorganisms, including (i) archaea and bacteria, (ii) aerobes and anaerobes, (iii) Gram-positive and Gram-negative bacteria and (iv) (hyper)thermophiles and mesophiles (and even psychrophiles). Moreover, these observations provide conclusive evidence for the divergent evolution of what appear to be two very closely-related species, C. bescii and C . saccharolyticus . Genes encoding proteins potentially involved in cell–carbohydrate adhesion In some cellulolytic microorganisms such as C. thermocellum , the strong interactions between the cells and the insoluble polysaccharide substrate are mediated by the cellulosome. The genome analyses of C. bescii shows that it does not produce a cellulosome complex as no dockerin- and cohesin-like domains of either types I or II were identified. In addition, genes encoding extracellular CAZy enzymes did not contain similar domains of unknown function that might encode new types of dockerins. However, microscopy studies show that C. bescii cells directly attach to xylan and switchgrass ( Figure 3 ). The attachment is dynamic as many cells are also planktonic, enabling cell densities to be used as a measure of cell growth ( 14 , 15 ). Although the mechanism is not known, analysis of the genome of C. bescii reveals many genes that are predicted to encode modules that could be involved in such cell–substrate interactions ( Table 4 ). They include surface-layer homology (SLH) domains which are known to mediate the binding of proteins to cell surfaces ( 60 ), fibronectin type 3-like (Fn3) domains containing binding sites for the cell surface ( http://pfam.sanger.ac.uk ), and lysine motif (LysM) domains found in a variety of enzymes involved in bacterial cell wall degradation that may have a general peptidoglycan-binding function ( 61 ). In addition, C. bescii contains Fn3-like domains that have sequence similarity to so-called ‘X’ domains, which have shown to bind carbohydrates ( 62 ).\n Figure 3. Scanning electron microscopy (SEM) images of C. bescii cells attached to xylan from oat spelts ( A ) and to switchgrass ( B ). The bars indicate (A) 1 μm and (B) 2 μm, respectively. \n Table 4. Caldicellulosiruptor bescii genes encoding proteins with putative cell adhesion, protein–protein interaction or carbohydrate-binding function Gene name Protein (AAs) SP Annotation Domain structure Transcriptomics Proteomics FilterPaper Significant ExtP Membrane WC Cbes_0012 3027 Y Q466C0 Putative uncharacterized protein SLH-SLH-SLH - Fn3 -VWA- RHS Up Yes CXn CXn Cbes_0077 1710 Y A4J714 S-layer domain protein SLH-SLH-SLH -Transglut_core Up Yes CXn C Cbes_0594 755 Y Q59154 Endoglucanase GH5- CBM28 - SLH-SLH-SLH Up Yes C Cbes_0608 547 Y A3DET8 Cellulose 1,4-beta-cellobiosidase X- SLH-SLH-SLH Up Yes Cbes_0438 1157 Y A4XG20 S-layer domain protein SLH-SLH -X Up No Xn Yes Cbes_1839 575 Y A4XM24 S-layer domain protein SLH-SLH -X Up Yes Cbes_1943 277 Y A4XI88 S-layer domain protein SLH-SLH Yes No Yes Cbes_2295 1074 Y A4XM87 S-layer domain protein SLH-SLH -X Yes Yes Xn C Cbes_2341 484 Y A4XH32 S-layer domain protein X- SLH-SLH Y/N No Cbes_1573 1055 Y A4XH96 Putative uncharacterized protein SLH-SLH - SpoVT_AbrB Y/N No Cbes_2303 1018 Y A4XM93 S-layer domain protein SLH -X Y/N No CXn CXn Cbes_2342 1010 Y A4XH31 Putative uncharacterized protein SLH -X Y/N No Yes Cbes_1944 1201 Y A4XI87 Fibronectin, type III domain protein X- Fn3 -X Y/N No Yes Cbes_1945 1265 Y A4XI87 Fibronectin, type III domain protein X- Fn3 -X Y/N No Cbes_0190 582 N A4XM45 Peptidase M23B X- LysM -G5-peptidase M3 UP Yes Cbes_0508 203 Y A4XIM2 Allergen V5/Tpx-1 family protein LysM -SH3_3 Y/N No Yes Cbes-0560 507 Y A4XHM2 Peptidoglycan-binding LysM LysM-LysM Yes No Yes Cbes_1391 109 Y A4XKU6 Peptidoglycan-binding LysM LysM Y/N No Yes Cbes_2402 511 N A4XGE4 Peptidoglycan-binding LysM X- LysM Y/N No Yes Cbes_0174 951 Y Extracellular solute-binding protein family 1 CBM_X-SBP1 Up No Xn Xn Cbes_0181 595 Y Extracellular solute-binding protein family 1 SBP1 - CBM_X Up Yes Xn CXn SP, Signal Peptide; SLH, surface layer homology domain; SBP1, solute-binding protein of family 1; X, domain not present in PFAM; CBM_X, Pfam annotation of PF06204; RHS, multiple tandem 22-residue repeats each containing strongly conserved dipeptide YD; WC, cell-extract; C, cells grown on cellulose; Xn, cells grown on xylan. Specifically, Cbes_0594 has an SLH domain combined with GH5 and CBM28. Binding of C. stercorarium xylanase to the cell wall via its SLH domains has been demonstrated ( 63 ). Cbes_0174 and Cbes_0181 contain bacterial solute-binding domains (SBPs), which are typically attached to an outer membrane and are components of sugar transport systems ( 64 ). Cbes_0174 has an N-terminal and Cbes_0181 has C-terminal modules with BLAST hits to CBM6 and pfam CBM_4_9, respectively (designated here as CBM_X, Table 4 ). All three proteins are candidates for binding to both cell (by SLH, SBP) and polysaccharides (by CBM28, CBM_X). There are also many modules of unknown biological function listed in Table 4 (modules designated as ‘X’, LysM, Fn3, RHS, etc.) that contain signal peptides and could potentially be presented on the cell surface of C. bescii that may display novel catalytic and/or binding functions. Hypothetical genes and their location According to the NCBI annotation, the C. bescii genome contains a total of 826 ORFs of unknown function that are annotated as encoding either hypothetical (723 HP) or conserved hypothetical proteins (103 CHP: Supplementary Table S14 ). We have now assigned a putative function to 46 of them via the KEGG [2] and COG [44] databases, and using the CAZy database another previously annotated CHP is annotated as a GT4 (Cbes_1572). In order to obtain some insight into the likely function of some of other C/HPs, we utilized the fact that genes transcribed in the same operon or gene cluster are often functionally related ( 65 , 66 ). A gene cluster is defined here as set of genes encoded on the same DNA strand with intergenic distances between adjacent genes of <300 bp ( 66 ). We found that 17 C/HPs are in the same operon with or located adjacent to CAZy genes, therefore, they are predicted to be functionally related to carbohydrate metabolism and potentially plant biomass conversion. As an example, Supplementary Figure S6 shows genes encoding a CHP (Cbes_0178) associated with genes encoding sugar transporters, suggesting that this CBP is likely involved in the same function. Consequently, using the KEGG and COG annotations, operon and gene cluster prediction analyses, putative functions can be assigned to a total of 295 HPs (41%, 428 remain unassigned) and 44 CHPs (43%, 59 remain unassigned: Supplementary Table S15 ). Insights into gene function from proteomic and transcriptomic analyses A total of 1429 (54%) of 2666 predicted protein-coding sequences (PCSs) were confirmed by proteomic analyses ( Supplementary Table S16 ) and 1790 (67.1%) PCSs were confirmed by transcriptomic data (Supplementary Tables S17 and S18). Therefore, a total of 2196 (83%) of the annotated PCSs were confirmed, including 46.6% by both methods, 18.5% by proteomics and 34.9% by transcriptomics. Among 88 genes annotated as CAZy-related genes, 59 (67.0%) are confirmed experimentally, including 28.8% by both methods. Among 826 PCSs that were annotated as encoding C/HPs, 613 (74%) were expressed on different substrates according to the transcriptomic and proteomic results. These data also allowed us to correct putative transcription unit (TU or operon) boundaries for 18 gene pairs (or 5% of the gene pairs with proteomic data: Supplementary Table S19 ). This leads to the splitting and merging of 20 TUs into 30 TUs. The 257 genes predicted to encode sugar transporters and CAZys are organized into 180 TUs ( Supplementary Table S4 ). Of the 171 transporters predicted to be sugar-related and 88 CAZy genes, expression at the RNA or protein level was shown for 136 (79%; Supplementary Table S9 ) and 84 ( Supplementary Table S8 ), respectively, have been detected. When C. bescii was grown on crystalline cellulose (filter paper) versus glucose, a total of 1203 genes had a significant change in expression level, as shown in Supplementary Table S17 . These included 64 CAZys (32 down- and 32 up-regulated: Supplementary Table S8 ), 90 transporters (29 down- and 61 up-regulated: Supplementary Table S9 ) and 358 C/HPs (124 down- and 234 up-regulated: Supplementary Table S14 ). Among the 21 primary CAZy genes (encoding proteins with signal peptides and CBMs) ( Supplementary Table S8 ), 16 were up-regulated on cellulose and 14 proteins were identified on both cellulose and xylan ( Table 3 ). Of 21 genes putatively related to cell–substrate adhesion ( Table 4 ), 12 genes were up-regulated on cellulose and 8 proteins were identified on cellulose and xylan. More detailed analyses were conducted with operons/gene clusters ( Supplementary Table S20 , see also Tables S21 and S22 ). Among the gene clusters whose expression is up-regulated during growth on cellulose, there are six of potential interests. These include (i) Cbes_1856-Cbes_1864 encoding the majority of CAZy multi-modular multifunctional enzymes discussed above, as well as two HPs; (ii) Cbes-2371-Cbes_2375 encoding GH43 and two membrane components of ABC transporters, a gene cluster that is missing in C. saccharolyticus ; (iii) Cbes_2413-Cbes_2421, Cbes_2494-Cbes_2500 and Cbes_0261-Cbes_0265, all of which encode HPs; and (iv) Cbes_2591-Cbes_2595 encoding an α-amylase, a DNA repair protein and three HPs. The up-regulation of these gene clusters on cellulose suggests that they are involved in plant cell wall conversion. Five clusters, including genes encoding proteins of different metabolic pathways, were down-regulated on cellulose indicating the plasticity of transcription regulation upon changing growth conditions. Upon switching from glucose to cellulose, four genes of the Cbes_1853-Cbes_1864 cluster are down-regulated while five other genes of the same cluster are up-regulated. This differential regulation validates our operon prediction that this cluster contains multiple transcription units. Among 42 predicted TFs with significant changes in gene expression, 17 and 25 were down- and up-regulated, respectively, when cells were grown on crystalline cellulose versus glucose ( Supplementary Table S23 ). It was previously suggested ( 67 ) that the level of expression of a TF is proportional to the number of operons that it regulates. This observation was used to predict the number of operons regulated by the TFs. Of 13 TFs with >4-fold changes, 7 and 6 were down- and up-regulated, respectively. In particular, TF Cbes_1856, a component of the major CAZy gene cluster, Cbes_1856-Cbes_1864, is up-regulated 3.4-fold suggesting that it is involved in plant biomass conversion. Cbes_2264 is up-regulated 17-fold. This TF is part of an operon encoding sugar transporters (Cbes-_2265-Cbes_2266), which are also up-regulated. These data suggest that these transporters utilize soluble oligomeric products of cellulose hydrolysis rather than glucose. In contrast, TF Cbes_2033 is down-regulated >8-fold, and it is located upstream of gene cluster Cbes_2029-Cbes_2031, which contains a predicted sugar transporter that is up-regulated 4-fold. This cluster is presumably not involved in cellulose metabolism. TF Cbes_1901 is down-regulated >9-fold, although the adjacent HP gene is up-regulated <2-fold, supporting the prediction that this TF regulates multiple operons ( Supplementary Table S23 ). It is also evident that the production of some CAZys, sugar transporters and C/HPs are sugar-specific ( Supplementary Tables S8 and S16 ). Two of them were found only when cells were grown on xylan: Cbes_0618 (CBM22-CBM22-GH10) and Cbes_0152 (CE7). This is in accord with the CAZy annotation, as the CBM22 domain binds xylan, GH10 is an endo-xylanase and CE7 is an acetyl–xylan esterase. These proteins are assumed to play a pivotal role in hemicellulose degradation. Other proteins were detected only upon growth on cellulose and cellobiose, but not on xylan. They include Cbes_0097 (GH30) and Cbes_0458 (GH1), which are potential β-glucosidases related to cellulose degradation, Cbes_0468 (GH36, potential α-galactosidase) and Cbes_0609 (CBM41-CBM48-GH13-CBM20 with CBMs binding to α-linked polysaccharides and α-amylase). Production of the latter two proteins during growth on cellulose suggests that the stereospecificity of the sugar linkage is not important for the regulation of the respective genes. Cbes_0459 and Cbes_0460 are putative cellobiose/cellodextrin phosphorylases (GH94) detected in much higher amounts than β-glucosidases. This is consistent with the energetics of cellulose degradation as cellobiose/cellodextrin phosphorylases provide an advantage for anaerobic cellulolytic microorganisms. They convert cellobiose/cellodextrins into glucose and glucose-1-phosphate without utilizing valuable ATP, which can be conserved for energy-consuming reactions. In contrast, β-glucosidase hydrolyses cellobiose into two glucose molecules, which must be phosphorylated with ATP before they can be utilized. Hence, in general, interpretation of the microarray and proteomic data is consistent with the CAZy database classification. Five and three sugar transporters were identified only after growth on xylan or on cellulose/cellobiose, respectively, consistent with the specificity of these proteins for certain oligosaccharides. Furthermore, 8 and 16 C/HPs were detected on xylan and cellulose/cellobiose, respectively. Two of the xylan-specific proteins (Cbes_2729 and _2368), and 2 cellulose/cellobiose specific proteins (Cbes_2630 and _1288) were detected at relatively high levels. These data suggest that these previously uncharacterized proteins play important roles in hemicellulose/cellulose metabolism, even though they have no recognizable CAZy domains."
} | 13,541 |
34112928 | PMC8192556 | pmc | 3,008 | {
"abstract": "Bio-photovoltaic devices (BPVs) harness photosynthetic organisms to produce bioelectricity in an eco-friendly way. However, their low energy efficiency is still a challenge. A comprehension of metabolic constraints can result in finding strategies for efficiency enhancement. This study presents a systemic approach based on metabolic modeling to design a regulatory defined medium, reducing the intracellular constraints in bioelectricity generation of Synechocystis sp. PCC6803 through the cellular metabolism alteration. The approach identified key reactions that played a critical role in improving electricity generation in Synechocystis sp. PCC6803 by comparing multiple optimal solutions of minimal and maximal NADH generation using two criteria. Regulatory compounds, which controlled the enzyme activity of the key reactions, were obtained from the BRENDA database. The selected compounds were subsequently added to the culture media, and their effect on bioelectricity generation was experimentally assessed. The power density curves for different culture media showed the BPV fed by Synechocystis sp. PCC6803 suspension in BG-11 supplemented with NH 4 Cl achieved the maximum power density of 148.27 mW m −2 . This produced power density was more than 40.5-fold of what was obtained for the BPV fed with cyanobacterial suspension in BG-11. The effect of the activators on BPV performance was also evaluated by comparing their overpotential, maximum produced power density, and biofilm morphology under different conditions. These findings demonstrated the crucial role of cellular metabolism in improving bioelectricity generation in BPVs.",
"introduction": "Introduction Discovering the inherent capability of microorganisms to generate bioelectricity has opened a new era of renewable energy production. Regarding the emergence of microbial fuel cells (MFCs) as a system to extract electrons from the complex organic substrate during the anaerobic biodegradation process, the critical role of particular microorganisms as biocatalysts in wastewater treatment and energy generation has become crystal clear 1 . Although the MFCs are highly in academic interests 2 , the advantages of bio-photovoltaic devices (BPVs), which can operate without organic sources and carbon dioxide emission, reveal more benefits in comparison with MFCs encountering difficulties of supplying organic feed and carbon dioxide management 3 . In BPVs, the oxygenic photosynthetic organism is being used as a bio-anode catalyst, deriving electrons from water photolysis under the light emission 4 . Despite the high potential of BPVs as a promising power generator, the main bottleneck restricting the commercial way is the intense competition of intracellular pathways for energy resources and intrinsic metabolic losses resulting in low power outputs generation 5 . Therefore, numerous researches were undertaken to understand the intracellular electron pathways of photosynthetic organisms. Among unicellular photosynthetic organisms, Synechocystis sp. PCC 6803 (henceforth referred to as Synechocystis ), the well-known model organism, received a great deal of attention to investigate the electron transfer pathways 6 . In this regard, investigation of inhibitors' effects on the photosynthetic metabolic pathways 7 , creation of mutant strains to redirect electron flux 8 , and eliminating photosystem II 9 , has been demonstrated in previously published studies to understand electrons generation mechanism in Synechocystis . Bombelli et al. investigated the use of classic inhibitors such as DCMU, DBMIB, and methyl viologen (MV), to identify the electron transfer chain suggesting that photo-electrons are originated from photolysis of water and end with the reducing form of PS I 7 . Bradley et al. redirected electron flux from competing electron sinks by creating mutant strains of Synechocystis lacking respiratory terminal oxidase complexes and using ferricyanide as an electron acceptor. When terminal oxidase complexes were inactivated, 10% of the respiratory electron flux was redirected to ferricyanide in the dark condition 8 . Cereda et al. studied the effect of photosystem II (PS II) on electron generation through both inhibitor addition and gene deletion. It was also proved that the primary photocurrent generation requires the activation of PS II during water photolysis 9 . Moreover, their case study on the Synechocystis revealed the incapability of transferring electrons to the electrode owing to numerous metabolic pathways facing fierce competition with the electrode for the electrons. Thus, extracellular photo-electrons are attributed to a small proportion of the total organism electron flux 9 . This matter has been pointed out in Glazier's work that cellular metabolism used only 3% of electrons produced by photosynthetic light reactions 10 , and thus the metabolism has considerable potential for electricity generation. However, switching the cellular metabolism to the generation of NADH is required. Despite the careful investigations on the photosynthetic electron pathways, the importance of a system-oriented strategy applying metabolic modeling to uncover other metabolic pathways that may contribute to electrons production has been neglected. Mao and Verwoerd applied a metabolic network using flux balance analysis (FBA) to explore the inherent capacity of electron generation for Synechocystis ; however, the lack of experimental works in their studies is evident 11 . Enhancing intracellular electron carriers such as NADH and NAD + can be considered a positive approach to enhancing energy generation in bioelectrochemical systems. In this regard, Han et al. regenerated NADH by genetically engineered Clostridium ljungdahlii , and increased the maximum power density of the MFC more than twofold 12 . Moreover, It was demonstrated that the electricity generation of Synechocystis depended on NADH production, which is profoundly affected by various metabolic pathways 11 . Therefore, concerning the vital role of metabolic modeling to identify intracellular constraints and switch the cellular metabolism to the generation of NADH, a regulatory defined medium can be developed to shift the cell metabolism to the object of interest 13 . In this study, a system-oriented strategy based on metabolic modeling is used to design a regulatory defined medium, reducing the intracellular constraints in the bioelectricity generation of Synechocystis via the cellular metabolism alteration to improve the NADH generation. Thus, a comprehensive investigation is achieved to identify the key reactions that play a critical role in enhancing electricity production in Synechocystis via a linear algorithm for finding multiple optimal solutions (LAMOS) 14 . Consequently, the predicted reactions are regulated by adding enzyme regulators to the culture medium. These compounds regulate the enzyme activity of the target reactions and are found in the BRENDA database 15 . Finally, regulators' effect on BPV performance is evaluated by comparing polarization and power density curves under various conditions.",
"discussion": "Discussion The proposed systemic approach based on metabolic modeling identified bottleneck genes in metabolism for enhancing bioelectricity generation. To this end, LAMOS and BRENDA were applied for the initial screening of existing compounds for switching metabolism to improve NADH production, and therefore increase power density generation (Fig. 7 ). It is important to note that proposed system-oriented strategy is a screening that seeks to make suggestions from a large number of available components which the effect of their addition to the medium can be evaluated. This strategy suggested the components affect the enzyme of key bottleneck genes, and thus, increasing the chance of improving bioelectricity production. In this regard, LAMOS identified the reactions that had the highest potential in improving bioelectricity production, listed in Tables S1 and S2 . Then, for each reaction, the potential regulators in the BRENDA database were searched. Besides, in addition to the experimentally reported effect of the regulator on the bottleneck genes, the regulator may be effective on some other enzymes; therefore, assessing the effect on the final product, which is the production of electrons, maybe the best solution. This method has also been implemented successfully in previous studies 13 , 37 leading to improved production. Figure 7 Graphical abstract. After an in-depth screening of all components in the BRENDA database, phosphotransacetylase, and glutamate dehydrogenase were selected to be experimentally investigated due to the following reasons. Firstly, since it is undeniable that regulators can affect other enzymes' activity as well 13 ; thus, the adverse effect of regulators on other metabolic pathways in improving power generation was investigated. In other words, as it was shown in Table S6 , ammonium chloride, which is an activator of phosphotransacetylase, did not have a substantial adverse effect on other critical metabolic pathways. However, potassium chloride, an activator of glutamate dehydrogenase, possessed an adverse effect on some critical metabolic pathways, including the inhibitory effect on oxidative phosphorylation pathways (Table S6 ). It was concluded that using the regulators that had an adverse effect on crucial pathways could decrease the bioelectricity production, albeit its positive effect on the targeted reaction. It is worth mentioning that the detailed effects of additives on cellular metabolism cannot be accurately evaluated, but with the aid of BRENDA, an attempt has been made to provide an estimate of the regulator’s effect on cellular metabolism, and components screening has been carried out based on this. Furthermore, for some enzymes, such as glycerol-3-phosphatedehydrogenase, no regulators were specified in the BRENDA database. Most recently, Lai et al. 24 also developed a recipe for BG11 (nBG11), compatible with the BPV operation. The nBG11 medium was modified by removing organic pH buffers (HEPES/TES), optimizing MnCl 2 concentration, and chelating trace metal mix solution (TMM) with EDTA. Optimizing culture medium using 1 mM ferricyanide as the mediator led to the long‐term photo‐current generation of about 20 µA from Synechocystis during repeated light–dark cycles for over 12 days. Similarly, in our study, redesigning the culture medium, particularly the supplementation of NH 4 Cl, generated a longer stationary phase (Fig. 4 A), which shows a more stable current density. Moreover, Bombelli et al. 38 were also supplemented BG-11 with NaCl to study bioelectricity generation of Synechocystis ; therefore, the closest minerals to NaCl were obtained and studied. Our results demonstrate that adding activators could provoke the Synechocystis as a strain engaged in noticeable intrinsic metabolic losses to reduce its activation overpotential significantly and accelerate bioelectricity generation. The overexpression of the phosphotransacetylase by NH 4 Cl exceedingly facilitates electrons transfer and stimulates cell electrogenesis metabolic pathways, which brings about lower activation losses. Although KCl led to an increase in glutamate dehydrogenase activity, the experimental results showed that it could not be helpful in electrogenesis metabolism as much as NH 4 Cl. The investigation of overshoot occurrence in the BPV also demonstrated the culture medium used NH 4 Cl as an activator produced a better performance in compensating electron depletions. Furthermore, the addition of NH 4 Cl had a small effect on the biofilm morphology. Interestingly, our experimental studies showed that the addition of 40 mM NH 4 Cl resulted in a 49% decrease in optical density (OD750) in the strain culture, which is consistent with the previous study, representing a photodamage 39 . On the other hand, it was reported that ammonium chloride saves up to 30% of the reducing equivalents produced by the photosynthetic light reactions in cyanobacteria 40 . More specifically, nitrate needs to be reduced to ammonium via two sequential reactions catalyzed by nitrate reductase and nitrite reductase to be assimilated by cyanobacteria. The reductions of nitrate to nitrite to ammonium are Fd-dependent, consuming two and eight electrons (Table S7 ), respectively 41 . Thus, although ammonium chloride was reported to cause photodamage and reducing cell density, its role in reducing intracellular constraints, saving reducing equivalents, and facilitating electrons transfer is undeniable. Additionally, this study focused on reducing intracellular constraints to enhance bioelectricity generation. However, some researches were conducted to minimize extracellular constraints to improve the BPV performance. Wenzel et al. 42 enhanced current generation from photosynthetic biofilms of Nostoc punctiforme and Synechocystis by using porous translucent electrodes in which a twofold increase in current generation was observed. Bombelli et al. 38 harnessed a high power density of 105 mW m −2 for Synechocystis by miniaturizing geometries using microfluidic BPV. In microfluidic devices, due to the reduction of the distance between the electrodes, ohmic losses are reduced 33 , and achieving the maximum power density is more feasible. Thus, diminishing intracellular constraints by the method presented in this study together with optimizing the BPV to reduce extracellular constraints, would be a big step towards the industrialization of this technology. Future research should also aim to improve the electrogenesis metabolic pathways by implementing systems metabolic engineering approach 43 . Zhu et al. 44 developed a BPV with a constrained electron flow based on a d -lactate mediated consortium consisting of engineered Synechococcus elongatus UTEX 2973 and Shewanella oneidensis MR-1. The cyanobacteria harvested solar energy to produce d -lactate, which is subsequently used by the exoelectrogenic microorganism, releases electrons from it, and transfers them to the anode for bioelectricity generation. A power density of 150 mW·m −2 in a temporal separation setup was produced by implementing this method. Thus, unlike the addition of mediators such as potassium ferricyanide, which is unsuitable for large-scale 7 , it is more interesting to implement the systemic approach to introduce novel microbial interactions capable of sustainable bioelectricity production. Besides, instead of supplementing the classic inhibitors such as DCMU and DBMIB, which were used to investigate the source of electrons in the photosynthetic organism, the addition of regulators identified by the systemic approach would bring about more benefits by improving electrical power generation. Finally, it is worth to mention that the produced power density of Synechocystis is still relatively lower than exoelectrogenic bacteria such as Shewanella or Geobacter . However, it would be interesting to implement this method to enhancing exoelectrogenic bacteria bioelectricity generation."
} | 3,788 |
35531503 | PMC9072019 | pmc | 3,009 | {
"abstract": "While studies have shown that anaerobic co-digestion of chicken manure (CM) and corn stover (CS) is an efficient method to treat these agricultural wastes, the microbial ecology of these systems and optimal parameters for the digestion process are yet to be determined. In this study, the effects of different initial substrate concentrations and CS : CM mixture ratios on co-digestion and microbial community structure were evaluated. Results demonstrated that both the highest cumulative methane yields and methane production rates were obtained from reactors with a CS : CM ratio of 1 : 1 during hemi-solid-state anaerobic digestion (HSS-AD). Cumulative methane yields and methane production rates were 24.8% and 42% lower in solid-state anaerobic digestion (SS-AD) reactors using the same CS : CM ratios. Analysis of microbial community structures revealed that cellulolytic bacteria and a diversity of syntrophic microorganisms capable of direct interspecies electron transfer (DIET) and hydrogen interspecies transfer (HIT) were enriched in the best-performing reactors. Methanosarcina species also dominated during HSS-AD, and their presence was positively correlated with methane production in the reactors.",
"conclusion": "Conclusions The HSS-AD reactor provided with feedstock at a 1 : 1 CS : CM ratio performed better than the other three operational parameters (HSS-AD3:1, SS-AD1:1, SS-AD3:1). These optimal parameters promoted methane production while suppressing the accumulation of VFAs and FAN, both compounds known to inhibit the biomethanogenic process. Analysis of bacterial and archaeal communities demonstrated conditions associated with the highest performing reactors enriched for bacteria that could form syntrophic partnerships with methanogens ( i.e. Methanosarcina ). Results from this study should be taken into consideration in future attempts to optimize full-scale digester performance.",
"introduction": "Introduction Agricultural wastes, such as animal manure and crop residue (for example, corn stover), are a major concern to the farming industry and are known to be major contributors to soil pollution and greenhouse gas production. 1 Each year, 400 million tons of chicken manure (CM) and 300 million tons of corn stover (CS) are produced in China alone. 2 Anaerobic digestion (AD) is a process that can convert these wastes into biogas as a renewable energy source, and any residual solids can be turned into nutrient-rich fertilizers. 3 However, mono-digestion of these feedstocks is not desirable because CS contains a high proportion of recalcitrant lignocellulose, 4 and CM has a low C/N ratio, 5 both factors that lead to poor AD performance. To minimize these inhibitory factors and improve biogas yields, co-digestion of both of these substrates is recommended. 6 While studies have already shown that CS : CM co-digestion strategies improve reactor performance, 7 use of large-scale anaerobic digesters has not been frequently implemented due to low energy production and poor economic viability. This limited application is largely because various operating conditions and the effects they have on microbial ecology have not yet been adequately studied. The type of AD applied often depends on the percentage of total solids (TS) in the waste to be treated. Wet state anaerobic digestion (WS-AD) is used for waste with TS content <10%, hemi-solid-state anaerobic digestion (HSS-AD) is used for waste with TS content between 10–15%, and solid-state anaerobic digestion (SS-AD) is used for waste with TS content >15%. While WS-AD is generally used for treatment of animal manure, HSS-AD and SS-AD are more effective during treatment of complex materials such as lignocellulosic crop residue 8 and are likely to be more compatible with co-digestion feedstocks. Other advantages of HSS-AD and SS-AD include higher volumetric methane output, enhanced nutrient balance, sufficient buffer capacity, smaller reactor requirements, and lower energy input and water addition. 9 Another major factor that must be taken into consideration when optimizing AD is the microbial ecology of the system. Successful AD relies on a diverse microbial consortium working together to convert complex organic matter into methane using a characteristic four-step process: hydrolysis, acidogenesis, acetogenesis, and methanogenesis. 10 Hydrolysis and acidogenesis lead to rapid production of fatty acids and alcohols, and the robust microbes involved in these reactions can survive a wide range of environmental conditions. Further conversion of fatty acids and alcohols relies on interspecies electron transfer (IET) between syntrophic bacteria and methanogens. IET via hydrogen/formate interspecies electron transfer (HFIT) has been recognized as the conventional route of biogas formation for several decades. 11–13 A more recently discovered form of IET is direct interspecies electron transfer (DIET), in which syntrophic partners forge biological extracellular protein connections to exchange electrons. 14–18 IET often occurs relatively slowly and is extremely sensitive to operating conditions. 10,11 Therefore, understanding how to promote these syntrophic interactions, especially in anaerobic digesters with high organic loads, is of great interest to the scientific community. 15 Although many studies have shown that supplementation of conductive materials such as granular activated carbon, 19 biochar, 20 and carbon cloth 21 can enhance IET during AD, 22–25 studies to determine how to enhance syntrophic metabolism by simply adjusting operational parameters are lacking. Evidence does suggest that altering initial substrate concentrations (ISC) based on total solid content affects IET. For example, extremely high solid content during AD may cause volatile fatty acids (VFAs), such as propionate or butyrate, to accumulate. 26,27 Large quantities of VFAs are toxic to methanogens and will lead to high hydrogen partial pressures which inhibit hydrogen interspecies transfer (HIT). 28 In addition, data regarding bacterial and methanogenic communities enriched during co-digestion of CS and CM is needed as it will help provide insight into mechanisms of IET that might be occurring in the reactors. This information can then be used to enhance the co-digestion process. For example, studies have shown that co-digestion of CM with microalgae stimulates growth of Methanothrix and Methanosarcina species, both organisms known to participate in DIET. 29 The discovery of optimal operational conditions for treatment of agricultural wastes with high TS content, and more importantly, the elucidation of microbial interactions key to this process, are essential for future large-scale application. Therefore, reactors were assembled with a variety of ISC (based on total solid content) and CS : CM mixture ratios to identify which operational parameters yielded the best results. Bacterial and archaeal communities were also characterized and relationships between community structure, methane production, ISC, and mixture ratios were identified.",
"discussion": "Results and discussion Anaerobic digestion performance using varying operating parameters In order to determine which operating parameters promoted the best anaerobic digestion of CS and CM, four different reactor conditions were tested in triplicate; HSS-AD with a CS : CM ratio of 3 : 1 (HSS3:1), HSS-AD with a CS : CM ratio of 1 : 1 (HSS1:1), SS-AD with a CS : CM ratio of 3 : 1 (SS3:1), and SS-AD with a CS : CM ratio of 1 : 1 (SS1:1). Methane concentrations were monitored during the entire operation period (40 days) ( Fig. 1 ). Cumulative methane production in the HSS-AD reactors was significantly ( p < 0.01) higher than the SS-AD reactors ( Fig. 1A ). The HSS1:1 reactors produced the highest amount of cumulative methane (223.7 ± 16.0 mL per g VS added ), with yields about 33.0% higher than the SS1:1 reactors. Cumulative methane formed by the HSS1:1 reactors was about 15% higher than yields from a similar reactor in a previous study, which may have been due to differences in inoculum. 7 The experiments also showed that reactors with 1 : 1 CS : CM ratios yielded methane production rates that were 35.7 ± 7.7% higher than reactors provided with a 3 : 1 feedstock ratio. These differences can likely be attributed to higher lignocellulose content in the 3 : 1 feedstock, making it significantly harder to degrade. Fig. 1 (A) Cumulative methane yields from anaerobic co-digestion and (B) daily methane production in digesters containing different mixture ratios and ISCs. Error bars represent standard deviations from triplicate samples. In addition to higher overall methane production, the HSS-AD reactors started producing methane early in the digestion process, with the highest daily methane yields being generated around day 7 ( Fig. 1B ). Conversely, the SS-AD reactors did not show strong start-up performance, and overall daily methane yields never exceeded 10.0 mL per g VS added . Both daily methane yields and methane content during the start-up period were lowest in the SS3:1 reactors ( Fig. 1B and ESI Fig. S1 † ), and total cumulative methane was only 202.6 ± 13.5 mL per g VS added in these reactors ( Fig. 1A ). While the 3 : 1 feedstock clearly hindered the start-up period in the SS-AD reactors, it did not appear to impact the start-up period in the HSS-AD reactors ( Fig. 1A ). These results suggest that hydrolysis/acidification was not a limiting factor in HSS-AD reactors. Impact of CS and CM feedstock mixture ratios on volatile fatty acids, nitrogen-containing compounds, and volatile solids VFAs, FAN, and TAN were also measured throughout the 40 day co-digestion period in order to assess the potential impact of CS and CM feedstock mixture ratios on microbial metabolism ( Fig. 2 ). VFAs accumulated rapidly during the early stages of treatment in all four reactor conditions ( Fig. 2A ), demonstrating that hydrolytic and acidogenic microorganisms were active. 36 However, VFA formation was higher in reactors provided with 1 : 1 CS : CM feedstock ratios, indicating that the higher lignocellulose content in the 3 : 1 reactors may have limited hydrolysis. 9,37 In addition, accumulation of VFAs throughout the co-digestion process was significantly greater in SS-AD reactors ( Fig. 2A ). These results suggest that microorganisms that could cooperatively metabolize VFAs to methane were enriched early in the digestion process in the HSS-AD reactors. Fig. 2 Change in (A) VFAs, (B) FAN, (C) and TAN during anaerobic co-digestion with different mixture ratios and ISCs. Error bars represent standard deviations from triplicate samples. Two large methane peaks were produced in the HSS-AD reactors over the course of the experiment ( Fig. 1B ). The first large peak occurred between days 5–10, which was also when VFA concentrations were highest ( Fig. 2A ), and the second peak occurred between days 19–21 for HSS1:1 reactors and days 27–29 for HSS3:1 reactors when VFAs dropped below 1107 mg L −1 and 2350 mg L −1 , respectively. These peaks may have corresponded with shifts in the dominant methanogenic communities over time. Early in the experiment, methanogens were likely to be using hydrogen generated by fermentative bacteria. These hydrogenotrophic methanogens initially kept H 2 concentrations below the thermodynamic threshold, 38–40 however, in the organically rich HSS-AD reactors, H 2 production eventually outpaced H 2 consumption and H 2 producing pathways are not thermodynamically favorable at high H 2 concentrations. 41 Therefore, fermentative microorganisms tend to shift their metabolism from H 2 producing pathways to VFA ( i.e. propionate, butyrate, acetate) producing pathways when H 2 concentrations are high. 41–44 Homoacetogenesis is also favored over hydrogenotrophic methanogenesis at elevated hydrogen partial pressures. 45 Acetoclastic methanogens can then directly utilize acetate as a substrate for methanogenesis, 46 and VFAs like propionate and butyrate can serve as electron donors for syntrophic partnerships between bacteria and methanogens. 47–49 Evidence supporting the theory that hydrogenotrophic methanogens were dominant when the first peak was formed comes from the finding that VFAs were most abundant during this period as methanogens were not yet able to convert these compounds to methane. During the second peak, on the other hand, VFA concentrations were low because acetoclastic methanogens and methanogens that could participate in DIET became more active. Further investigations into microbial community structure in the HSS-AD reactors during the first and second methane peaks are warranted. Free ammonia nitrogen concentrations were highest in SS-AD reactors with CS : CM ratios of 1 : 1 ( Fig. 2B ), with concentrations remaining well above 1000–2000 mg L −1 throughout the digestion process. The highest TAN was also observed using this operating parameter ( Fig. 2C ). High FAN and TAN negatively impact methanogenesis and are known to be major inhibitors of anaerobic digestion. 50,51 In fact, FAN concentrations >1110 mg L −1 have been shown to completely inhibit anaerobic systems. 52 This explains why the SS1:1 reactors produced the lowest cumulative methane yields ( Fig. 1A ). FAN remained ≤1000 mg L −1 in the other three reactor conditions, indicating that ammonia concentrations did not adversely affect metabolic activity in these digesters ( Fig. 2B ). Removal of VS and methane conversion rates also varied significantly between the different reactor conditions ( Fig. 3 ). Significantly higher proportions ( p < 0.01) of VS were removed in HSS1:1 (60.1 ± 2.4%) and HSS3:1 (59.0 ± 1.4%) reactors than in SS1:1 (51.4 ± 0.8%) and SS3:1 (56.2 ± 2.0%) reactors. Higher methane conversion rates were also observed in HSS1:1 (372.1 ± 27.7 mL methane per g VS removed ) and HSS3:1 (370.6 ± 19.3 mL methane per g VS removed ) reactors than in SS1:1 (327.3 ± 6.1 mL methane per g VS removed ) and SS3:1 (360.2 ± 24.9 mL methane per g VS removed ) reactors ( Fig. 3 ). Overall, HSS-AD reactors provided with a CS : CM ratio of 1 : 1 performed best with the highest cumulative methane yields and methane production rates. Fig. 3 Removal of volatile solids (VS) and methane conversion in anaerobic co-digestion using different mixture ratios and ISCs. Error bars represent standard deviations from triplicate samples. Microbial community characterization High-throughput sequencing of bacterial and archaeal 16S rRNA gene fragments was done to determine the impact that reactor operational parameters (ISC and CS : CM mixture ratios) can have on microbial community structure in reactors treating high-solid agricultural waste ( Fig. 4 ). Fig. 4 Relative distribution of (A) bacterial and (B) archaeal 16S rRNA gene sequences at the genus level. Sequences that accounted for less than 1.0% of the population were classified as “Others”. Bacterial community analysis At the genus taxonomic level, there were several key differences between bacterial communities associated with well-performing HSS-AD reactors and lower methane-yielding SS-AD reactors ( Fig. 4A ). Both of the HSS-AD communities were abundant in species from the genera Ercella , Clostridium , Pelobacter , and Herbivorax , which accounted for 65.5% ± 0.8% of the population. These same four species accounted for only 16.8% in SS1:1 reactors, which were the poorest performing reactors ( Fig. 1 ). \n Ercella and Clostridium were the most abundant genera in both of the well-performing HSS-AD reactors and SS3:1 ( Fig. 4A ). Ercella and Clostridium are members of the Clostridiales order which plays many roles in the anaerobic digestion process, including high-rates of cellulose hydrolysis, protein catabolism, and acidogenesis leading to production of short-chain fatty acids, CO 2 , and H 2 . 53 Although both of these genera are primarily known for their fermentative metabolisms, some Clostridium species are capable of extracellular electron transfer to such electron acceptors as Fe( iii ), sulfur, and current-harvesting electrodes 54–58 and Ercella can transfer electrons to insoluble sulfur compounds. 58 The presence of Clostridium and Syntrophomonas species has been previously reported to be an indicator of AD stability. 59,60 This is consistent with the high abundance of Clostridium and Syntrophomonas ( p < 0.01) in HSS-AD communities, and the positive correlation between their abundance and cumulative methane production ( p < 0.01, R 2 = 0.78) ( Fig. 1A , and 4A ). This correlation between methane production and abundance of these bacteria can be explained by the fact that both of these genera have been shown to form syntrophic partnerships with methanogens. 61 In addition to their cooperative growth with methanogens via HFIT, 61 it is also possible that these species are capable of DIET to a methanogenic partner. There are two hallmarks of a bacterium's capacity for DIET, the first is the ability to transfer electrons across the cell membrane to an extracellular electron acceptor ( i.e. another microorganism), 62,63 and the second is the presence of Geobacter -like type IV pilin proteins. Geobacter pili contain aromatic amino acids that are located at key positions within the protein chain 64 and account for ≥9% of the amino acid residues with small gaps between them. 65,66 Although the role of pili in extracellular electron transfer is under debate, pili from all of the bacteria known to be capable of DIET thus far have these characteristics. 65,66 In fact, recent identification of a Geobacter -like pilin protein in Syntrophus acetitrophicus led to the discovery that this species, once thought to only be capable of HFIT, 67,68 also has the capacity for DIET. 66 A number of genomes from Clostridiales species known to participate in HFIT, 61 including Syntrophomonas , also have genes coding for Geobacter -like pili, 66 suggesting that these species might also be capable of DIET. Support for this comes from the finding that Syntrophomonas species were likely to be participating in DIET in anaerobic digesters supplemented with ferroferric oxide. 69 In addition, both Ercella and Clostridium species are able to transfer electrons outside of the cell suggesting that it is possible that these genera were not only providing electrons to methanogens through fermentative by-products, but may have also been participating in DIET. \n Pelobacter species were also significantly enriched in HSS-AD reactors (9.1% of total community in HSS3:1 and 5.6% in HSS1:1). However, sequences from this genus were barely detected in SS-AD reactors (0.02% of total community in SS3:1 and 0.5% in SS1:1). Pelobacter has been shown to grow syntrophically with methanogens via HFIT, 70–72 but several species from this genus also have traits that are characteristic of bacteria with the capacity for DIET. For example, Pelobacter species are capable of electron transfer to Fe( iii ) and S0, 73–75 and have Geobacter -like type IV pili. 76 Species from the genus Sporanaerobacter were also significant members of all four digesters, but had the highest abundance in SS1:1 (14.8%). Sporanaerobacter metabolize proteins and carbohydrates with reduction of elemental sulfur, an extracellular electron acceptor. 77 In a previous study, Sporanaerobacter species were enriched in anaerobic digesters supplemented with conductive carbon cloth, and were likely participating in DIET with Methanosarcina species. 77,78 \n Herbivorax and Cellulosibacter , both genera with cellulolytic metabolisms, 79,80 were enriched in HSS-AD reactors ( Fig. 4 ). Their abundance was significantly lower in SS-AD reactors suggesting that their scarcity could be linked to impairment of lignocellulose degradation and lower methane production rates. Overall, results from the bacterial community analysis suggest that a variety of syntrophic bacteria involved in both HFIT and/or DIET were significantly ( p < 0.05) enriched in HSS-AD reactors, signifying that syntrophic–methanogenic associations were favored during HSS-AD. Archaeal community analysis The predominant methanogenic genera found in all samples included Methanosarcina , Methanoculleus , Methanothrix , and Methanosphaerula , however, the relative abundance of each genus varied between conditions ( Fig. 4B ). Methanosarcina was dominant in HSS3:1 (86.6% of the overall community), HSS1:1 (95.3%), and SS3:1 (85.5%). However, Methanoculleus (50.2%) were more abundant than Methanosarcina (42.4%) in SS1:1. The relative abundance of Methanosarcina species was positively correlated with cumulative methane production ( P < 0.01, R 2 = 0.84) ( Fig. 1A , and 4B ). These results are consistent with previous studies that have shown that Methanosarcina species promote stable digestion, particularly in reactors treating complex organic waste. 81,82 \n Methanosarcina species are mixotrophic methanogens that are able to metabolize acetate, hydrogen, and C1 compounds, 83 and along with Methanothrix (formerly Methanosaeta ), are some of the few methanogens shown to be capable of DIET in culture studies. 71 Methanothrix are obligate acetoclastic methanogens, and accounted for <4.2% of the overall archaeal community within all digesters ( Fig. 4B ). While high acetate concentrations in the reactors should have favored growth of both Methanosarcina and Methanothrix , 84–86 Methanothrix are sensitive to high concentrations of VFAs. 87,88 Methanosarcina , on the other hand, are typically able to survive exposure to environmental stressors, such as high VFA or FAN concentrations, because they can aggregate to form irregular clumps that protect them from harsh environmental conditions. 88,89 Strictly hydrogenotrophic Methanoculleus species predominated SS1:1 reactor archaeal communities (50.2%) ( Fig. 4B ). Methanoculleus abundance was positively correlated with concentrations of FAN and TAN ( P < 0.01, R 2 = 0.78) ( Fig. 2B, C , and 4B ), and negatively correlated with cumulative methane production ( P < 0.01, R 2 = −0.84) ( Fig. 1A , and 4B ). This is consistent with reports that Methanoculleus tend to dominate reactors with high concentrations of VFAs and ammonia ( Fig. 2 ). 60,90 Sequences from other hydrogenotrophic methanogenic genera, including Methanospaerula and Candidatus-Methanoplasma , were also found in all of the digesters ( Fig. 4B ). However, their abundance was low, indicating that these hydrogenotrophic methanogens played a lesser role. It is clear that most of the methanogenic activity detected in efficiently operating reactors (HSS3:1, HSS1:1, SS3:1) could be attributed to Methanosarcina . Not only is this genus able to survive harsh conditions found in anaerobic digesters with high organic loads, it is also capable of syntrophic growth with syntrophic bacteria through both HFIT and DIET 72,91 and can use a variety of fermentative products as substrates for methanogenesis (acetate, H 2 /CO 2 , formate, C1 compounds). 83 Correlation between microbial community composition and reactor performance Redundancy analysis (RDA) was conducted to elucidate relationships between microbial community structure, operational conditions (ISC and mixture ratios), and reactor performance (cumulative methane production) ( Fig. 5 and 6 ). RDA showed that both ISC and mixture ratio had a significant impact on bacterial community composition ( Fig. 5 ). The relative abundance of bacterial sequences from Herbivorax , Cellulosibacter , Ercella , Syntrophomonas , and Pelobacter positively corresponded with methane yields ( Fig. 5 ). As previously discussed, these genera are associated with degradation of cellulose and syntrophic metabolism of alcohols or short-chain fatty acids, both of which enhance reactor performance by preventing the accumulation of inhibitory intermediates. On the other hand, these genera were negatively correlated with ISC and thus primarily clustered with the HSS-AD samples. This indicates that a decrease in ISC promotes the enrichment of these beneficial genera which aid in the production of methane ( Fig. 1 ). Abundance of Ercella , Herbivorax , and Clostridium species were partially associated with an increase of CS in feedstocks, which makes sense because these are cellulolytic bacteria that would be enriched when lignocellulose content is high. Fig. 5 Redundancy analysis triplot ordination diagram using the bacterial community at the genus level (bacterial genera fitting greater than 90% were displayed). Arrows represent reactor performance, operational conditions, or bacterial genera. Vector length indicates the strength of correlation with a given ordination axis. The cosine of the angle between two arrows indicates the correlated degree. Solid stars represent HSS-AD samples, and diamonds represent SS-AD samples (3 : 1 for CS : CM = 3 : 1, 1 : 1 for CS : CM = 1 : 1). The distance between sample symbols approximates their dissimilarity. Fig. 6 Redundancy analysis triplot ordination diagram using the archaeal community at the genus level (archaeal genera fitting greater than 90% were displayed). The presence of such fermentative genera as Clostridium , Gallicola , and Garciella , were also significantly influenced by operational parameters and their presence appeared to promote methanogenesis ( Fig. 5 ). In contrast, the relative abundance of sequences from Caldicoprobacter and Tissierella were positively correlated with ISC and negatively correlated with mixture ratio ( Fig. 5 ), suggesting that these bacteria are likely favored in digesters with high ammonia concentrations (SS1:1). This is consistent with previous reports that showed that Caldicoprobacter and Tissierella were dominant in anaerobic digesters with excess ammonia or alkaline conditions. 92,93 ISC and mixture ratio also directly influenced archaeal community structure ( Fig. 6 ). For example, Methanoculleus and Methanosphaerula showed a significant ( P < 0.01) positive correlation with ISC, and were clustered closely with the SS1:1 sample, which were parameters that yielded the lowest methane productions. The high prevalence of obligate hydrogenotrophic methanogens in SS1:1 reactors suggests that hydrogen was the main electron carrier between bacteria and methanogens. 84 In this syntrophic scenario, the hydrogen partial pressure must typically be kept lower than 10 −5 atm by hydrogen consumers, and the value of Gibbs free energy discharge of the entire methanogenic process is thereafter negative, making syntrophic methanogenesis thermodynamically feasible. 94 However, in SS1:1 reactors, excess FAN likely inhibited hydrogenotrophic methanogenesis, and diffusion limitations may have also resulted in local hydrogen accumulation, all of which may cause a significant increase of hydrogen partial pressure in these reactors, further inhibiting the syntrophic metabolism. 51,84,95 \n Methanosarcina showed a significant ( P < 0.01) positive correlation with cumulative methane production, and was negatively correlated with ISC ( Fig. 6 ). The metabolic plasticity and morphological characteristics of Methanosarcina are likely to give them a competitive advantage in these systems. They can utilize DIET for methane production when conditions ( i.e. high ammonia concentrations) are inhibitory for hydrogenotrophic methanogenesis. 95"
} | 6,881 |
24683461 | PMC3967904 | pmc | 3,010 | {
"abstract": "Controlled experiments show that arbuscular mycorrhizal fungi (AMF) can increase competitiveness of exotic plants, potentially increasing invasion success. We surveyed AMF abundance and community composition in Centaurea stoebe and Potentilla recta invasions in the western USA to assess whether patterns were consistent with mycorrhizal-mediated invasions. We asked whether (1) AMF abundance and community composition differ between native and exotic forbs, (2) associations between native plants and AMF shift with invading exotic plants, and (3) AMF abundance and/or community composition differ in areas where exotic plants are highly invasive and in areas where they are not. We collected soil and roots from invaded and native forb communities along invasion gradients and in regions with different invasion densities. We used AMF root colonization as a measure of AMF abundance and characterized AMF communities in roots using 454-sequencing of the LSU-rDNA region. All plants were highly colonized (>60%), but exotic forbs tended to be more colonized than natives ( P < 0.001). We identified 30 AMF operational taxonomic units (OTUs) across sites, and community composition was best predicted by abiotic factors (soil texture, pH). Two OTUs in the genera Glomus and Rhizophagus dominated in most communities, and their dominance increased with invasion density ( r = 0.57, P = 0.010), while overall OTU richness decreased with invasion density ( r = −0.61, P = 0.006). Samples along P. recta invasion gradients revealed small and reciprocal shifts in AMF communities with >45% fungal OTUs shared between neighboring native and P. recta plants. Overall, we observed significant, but modest, differences in AMF colonization and communities between co-occurring exotic and native forbs and among exotic forbs across regions that differ in invasion pressure. While experimental manipulations are required to assess functional consequences, the observed patterns are not consistent with those expected from strong mycorrhizal-mediated invasions.",
"introduction": "Introduction Non-native plant invasions can reduce plant diversity and alter ecosystem processes (Levine et al. 2003 ). Despite substantial research efforts (Richardson and Pyšek 2008 ), we still lack a predictive understanding of how some exotic plants establish and outcompete native plants. The two most popular hypotheses state that exotic plants become invasive due to an escape from natural enemies in their native range (enemy release hypothesis; Keane and Crawley 2002 ) or fail to become invasive because they encounter new enemies in their exotic range (biotic resistance hypothesis; Elton 1958 ). Mutualists may be equally important for driving plant invasions. For example, invasions may fail due to an absence of specific mutualists in the exotic range (Richardson et al. 2000 ) or succeed because the invader either brings its mutualists or is able to associate with native or cosmopolitan mutualists (Dickie et al. 2010 ). It has been suggested that some invasions are successful due to encounters with new, and better, mutualists (Reinhart and Callaway 2006 ) although solid experimental data for this are lacking. Invasive plants may also disrupt native mutualisms and decrease the competitiveness of the community being invaded (Stinson et al. 2006 ). Overall, a better understanding of how plant invasions affect – and are affected by – mutualists is critical for improving our understanding of how some exotic plants become invasive. Here, we focus on arbuscular mycorrhizal fungi (AMF) in the phylum Glomeromycota that form a symbiosis, arbuscular mycorrhiza (AM), with the majority of land plants in which the fungi receive photosynthate in exchange for phosphorus, nitrogen, and other putative services (Smith and Read 2010 ). Due to their ubiquity and location in the root–soil interface, AMF have been referred to as keystone mutualists (O'Neill et al. 1991 ) with the potential to influence ecosystem processes such as productivity, carbon and nutrient cycling, water use, and soil structure (Rillig 2004 ). In general, perennial, coarse-rooted plants appear most dependent on AM (Wilson and Hartnett 1998 ), but responses are contingent on the particular plant–fungus combination (Klironomos 2003 ) and environmental conditions (Johnson et al. 1997 ; Johnson and Graham 2012 ). Evaluating the likelihood that AMF mediate plant invasions includes assessments of potential differences in the mycorrhizal dependency of the invading exotic and native plants, as well as of invasion-driven changes in AMF abundance and community composition (Pringle et al. 2009 ). Previous work has shown that AMF have the potential to influence plant invasive success. For example, the successful invaders Euphorbia esula (spurge) and Centaurea jacea (knapweed) have a high AM dependency (Klironomos 2002 ) and Centaurea stoebe becomes more competitive toward native plants when grown with AMF (Marler et al. 1999 ). Both invaders also increase the overall abundance of AMF and harbor different AMF communities than adjacent native grasslands (Lekberg et al. 2013 ). Another mycotrophic invader, Solidago canadensis , alters AMF communities in ways that promote its own growth more than a competing native plant (Zhang et al. 2010 ). In contrast, Allaria petiolata (garlic mustard) and Bromus tectorum (cheatgrass) are nonmycorrhizal or have a low mycorrhizal dependency (Stinson et al. 2006 ; Busby et al. 2011 ). These two invaders reduce the overall abundance of AMF and the root colonization of mycotrophic native plants (Stinson et al. 2006 ; Lekberg et al. 2013 ), which could promote further invasions by reducing native plant competitiveness (Vogelsang and Bever 2009 ). Because most experiments involving AMF and plant invasions have been conducted under controlled greenhouse conditions (e.g., Marler et al. 1999 ; Zhang et al. 2010 ), it is unclear whether responses are realized in the field in the presence of multiple interacting factors (Smith and Read 2010 ; Johnson and Graham 2012 ; but see Callaway et al. 2004 ). Observational studies incorporate all the complexity of the real world and are capable of revealing the end result of long-term ecological processes in ways that short-term greenhouse studies cannot (Diamond 1983 ). For example, survey data revealed a likely co-invasion by Pinus contorta and its ectomycorrhizal fungi in New Zealand as well as P. contorta's greater reliance on cosmopolitan mutualists relative to native plants (Dickie et al. 2010 ). Taking a similar approach, we assessed the potential for AM to influence the invasion success of C. stoebe and Potentilla recta by asking whether (1) the exotic forbs harbor a different AMF abundance and/or community composition than native forbs, (2) mycorrhizal associations in natives are altered in the presence of P. recta (i.e., assessing the ability of P. recta to interrupt native mycorrhizas), and (3) AMF abundance and/or community composition differs between areas where the two forbs are highly invasive and areas where they are not. The observed patterns of AM were examined for consistency with patterns that might be expected in a mycorrhizal-mediated invasion. For example, increases in colonization or changes in composition of the AMF community in more densely invaded plots (presumably toward a stronger association with, or a more beneficial suite of symbionts as in Zhang et al. 2010 ) would be consistent with mycorrhizal-mediated invasion. We want to stress, however, that unaltered AMF abundance and community compositions does not disprove mycorrhizal-mediated invasion, as AM function can vary with abiotic parameters and host plant identity. Further, we did not manipulate our factor of interest (level of invasion by P. recta and C. stoebe ) and controlled experiments are therefore required to assess causation as well as any functional consequences of differences observed (Gotelli and Ellison 2013 ).",
"discussion": "Discussion Using survey data, we assessed the potential of AMF to influence invasion success. We specifically asked (1) whether native and exotic forbs at the same sites harbor different AM colonization and AMF communities, (2) whether the composition of AMF communities in native forbs is different in the presence of exotic forbs, and (3) whether exotic forbs harbor different AM colonization and AMF communities across regions (and simultaneously across a gradient of invasion density). The nature of the data sampled also allowed us to assess whether the AMF community composition correlated more strongly with the soil environment, plant communities, or spatial distance among sites. Previous comparisons between native and exotic invasive plants have focused on specific native species with selected invaders and found significant shifts in AMF communities associated with native plants when co-occurring with invaders (Mummey et al. 2005 ; Hawkes et al. 2006 ; Stinson et al. 2006 ; Busby et al. 2011 ). However, generalizing those results to apply broadly to invasive species may be problematic because comparisons included different plant functional groups, which is arguably the most important factor in predicting mycorrhizal effects (Hoeksema et al. 2010 ; Lekberg et al. 2013 ). To minimize confounding factors, we focused our study on non-N 2 -fixing, mycorrhizal forbs across multiple, independent invasions to isolate the effect of plant provenance from factors of mycorrhizal status and plant functional group. Do AMF abundance and community composition differ between native and exotic forbs? All forbs, exotic and native, were highly colonized with rates ranging from 61% to 96%. High rates of colonization may imply a high degree of mycorrhizal dependency (Hetrick et al. 1996 ; Wilson and Hartnett 1998 ), even in the invasive plants C. stoebe and P. recta , which conflicts with previous suggestions that many invasive plants have a low or evolved mycorrhizal dependency (Seifert et al. 2009 ; Pringle et al. 2009 ; Vogelsang and Bever 2009 ; but see Klironomos 2002 and Shah et al. 2008 ). Indeed, several studies have shown that C. stoebe depend on AMF to be competitive (Marler et al. 1999 ; Callaway et al. 2004 ). The greater relative abundance of arbuscules in native compared with exotic forbs (Table 1 ) could indicate increased nutrient exchange between plant and fungus (Smith and Read 2010 ). More arbuscules in native forbs may also reflect some degree of preferential resource allocation between plants and fungi, because arbuscules (and the analogous hyphal coils) are the primary symbiotic interface across which nutrients are transferred (Smith et al. 2001 ; Smith and Smith 2011 ), and more arbuscules have been observed in hosts that deliver more carbon (Lekberg et al. 2010 ). Hausmann and Hawkes ( 2009 ) observed a similar increase in arbuscules in a greenhouse experiment of native and exotic grasses and concluded that AMF prefer native hosts. However, our study is based on a one-time sampling and differences may also be attributed to seasonal growth patterns of plants. A more focused study on costs and benefits across seasons is required to determine whether arbuscule abundance is greater in native plants year round, and to assess possible consequences for mycorrhizal function. Although richness of AMF taxa in roots is greater than once believed (Öpik et al. 2006 ), plant roots are typically dominated by just a few taxa that can account for 40% of all recovered sequences (Dumbrell et al. 2010 ). We observed the same pattern here; all but two of the seventeen plant communities sampled (both native and invasive) were dominated by two OTUs (OTU 1 and 16, closely related to Glomus microaggregatum and Rhizaphagus irregularis, respectively), which accounted for 54% of all sequences recovered in this study. By considering the invasion gradient, we can infer how invading P. recta interacts with existing AM. In the pre-invasion (native community), OTU 1 and 16 made up 44% of the recovered sequences, but this increased to 62% of recovered sequences in the “post-invasion” (exotic monoculture community). These data suggest that this exotic forb forms symbioses with pre-existing dominant AMF rather than promoting rare AMF taxa. Whether or not these OTUs contribute to invasiveness or respond to an increased abundance of a preferred host is unknown and will require manipulating OTU abundances. It is interesting to note, however, that both Rhizophagus irregularis and Glomus microaggregatum are considered generalists in their host and environmental preferences, which is consistent with suggestions that exotic plants may associate with generalist, cosmopolitan mutualists to a greater extent than native plants (Dickie et al. 2010 ; Moora et al. 2011 ; Öpik and Moora 2012 ). The greater dominance of OTU 1 and 16 in P. recta relative to native forbs was coupled, not unexpectedly, with a reduced AMF richness. Lower richness might be expected because we sampled only a single plant species within invasions but included multiple species in native plant communities. At least some host preference exists in the AM symbiosis (e.g., Vandenkoornhuyse et al. 2003 ), and plant species richness correlates with AMF species richness under at least some conditions (van der Heijden et al. 1998 ; Landis et al. 2004 ). However, we did not see a reduced AMF richness in the C. stoebe invasions, which corroborates earlier findings (Lekberg et al. 2013 ) and shows that richness can be high in monodominant stands of mycotrophic invasive forbs. P. recta and C. stoebe affected AMF richness differently, yet they had similar biomass production (Gallagher 2011 ) and AM colonization – two factors that may correlate with belowground C-allocation and in turn overall fungal richness (Waldrop et al. 2006 ). These apparently inconsistent results indicate that more work is required to determine the mechanisms by which host plants affect AMF richness. Do native plants have altered AMF associations in the presence of exotic plants? Altering the AMF of native neighbors may be one mechanism by which exotic forbs increase competitiveness (Zhang et al. 2010 ). C. stoebe can alter the mycorrhizae of adjacent Pseudoroegneria spicata (bluebunch wheatgrass plants; Mummey et al. 2005 ), and exotic annual grasses can decrease AMF richness in co-occurring native perennial grasses (Hawkes et al. 2006 ). In sampling across the three P. recta invasion gradients, we observed only a small shift in AMF communities. Some rare taxa were lost during invasion, but OTU 1 and 16 remained dominant (Table 1 ). In two of our three sites (MPG1 and MPG2), the AMF communities found in the roots of P. recta and the native forbs growing in the transition area were both intermediate to those found in the “pre-invasion” (native) community and the “postinvasion” monoculture. Apparently, neighboring plants of the same functional group can influence each other's AMF communities. However, whether these shifts are significant enough to affect overall mycorrhizal function is debatable and will require more detailed in situ functional studies. Neighboring P. recta and native forbs shared both dominant and rare OTUs (Fig. 3 ). The redundancy in AMF taxa and the close proximity of these plants (<10 cm) make it highly likely that these plants were connected via common mycorrhizal networks (CMNs). Nutrients such as nitrogen (He et al. 2009 ) and phosphorus (Wilson et al. 2006 ) can be transferred through CMNs, and the terms of carbon/nutrient trade between fungi and host plant have important implications for plant competitive interactions. Careful in vitro studies have found convincing evidence for bidirectional control of the carbon/nutrient trade between plants and fungi (Lekberg et al. 2010 ; Hammer et al. 2011 ; Kiers et al. 2011 ; Fellbaum et al. 2012 ). If exchange is truly “tit for tat,” then CMNs may create a positive feedback where larger hosts contribute more carbon and receive more benefit than their smaller neighbors (Weremijewicz and Janos 2013 ; but see Walder et al. 2012 ). This alone could help explain the success of some fast growing exotic species. Further, some plant-AMF species combinations are more beneficial to host plants than others (Klironomos 2002 , Kiers et al. 2011 ; Walder et al. 2012 ), and exotic competitiveness could increase if exotic plants are capable of shifting AMF communities toward more cooperative AMF taxa, although we did not find evidence of this. Finally, there may be temporal differences in exchange, which have not yet been explored, but are critical to understanding the implications of CMNs for invasion success across relevant time periods. Do AMF abundance and/or community composition differ in areas where exotic plants are highly invasive and areas where they are not? Arbuscular mycorrhizal fungi abundance and community composition of two exotic forbs were compared between regions where these exotics are highly and moderately invasive, similar to the biogeographical approach taken by previous investigators to identify drivers of invasions (e.g., Callaway et al. 2004 ; Lankau 2010 ). We hypothesized that if AMF influence plant invasions, then AMF abundance and/or the community composition would differ between areas where the forbs are highly invasive (Montana) and moderately invasive areas (western Washington). We found that AM colonization tended to be higher in Montana plants, but all exotic forbs were highly colonized (>60%), and we found no strong shifts in AMF community composition. Thus, while some of these changes were statistically significant, it is unclear whether they are functionally important. Overall drivers of AMF community assembly The ordination of our data from all sites suggests that soil abiotic factors were more important than host plant identity and site for determining AMF community composition when plants from the same functional group were compared. This agrees with previous work and reinforces the importance of pH and texture (Fierer and Jackson 2006 ; Lekberg et al. 2007 ) as important environmental habitat filters for microbial community assembly. Of the 30 OTUs detected, each plant community only hosted between 6 and 14 different OTUs and most were dominated by two OTUs that appear to have global distributions. Genetic dominance does not necessarily translate into relative abundance in roots however. Structure formation differs between AMF genera (Dodd et al. 2000 ); and a higher density of vesicles, which contain thousands of copies of nuclei (Bécard and Pfeffer 1993 ), can result in a higher number of sequences per unit fungal biomass. Both G. micoraggregatum and R. irregularis typically produce large numbers of vesicles, which may at least partially explain their dominance throughout this data set. Overall vesicle formation was consistent across all plant communities, however, so shifts in dominance from P. recta to native forbs are likely to reflect shifts in relative abundance of taxa in roots. The overall richness found here is lower than was found in a similar study by Lekberg et al. ( 2013 ), but we used substantially fewer sequences, and thus, our focus here was on dominant OTUs rather than an exhaustive characterization of all taxa present. It is also interesting to note that site-specific taxa tended to be rare, which agrees with Öpik et al. ( 2006 ). Summary Utilizing survey data that can reveal the end result of long-term ecological processes, we observed greater mycorrhizal colonization in exotic than in native forbs, and in regions where exotics are highly invasive. This could imply that AMF directly benefit exotic plants and may contribute to invasions, although natives were only slightly less colonized and fungal abundance alone seems an unlikely explanation for invasive success. Samples collected along invasion gradients suggested that P. recta joins existing mycorrhizal networks, simultaneously increasing dominance of the two most abundant OTUs and decreasing richness of rare taxa. If rare and specialist taxa are important for native forbs, this decrease in richness may indirectly benefit P. recta by decreasing native forb competitiveness. But again, the shifts were small and may not be biologically significant. Also, there were only slight differences in AMF abundance and community composition in areas where P. recta and C. stoebe are in high abundance compared with areas where they are not, which appears inconsistent with a strong, mycorrhizal-mediated invasion. Across our study, community assemblages were best predicted by abiotic parameters (soil pH and texture), not plant community type, adding to the growing body of evidence that some taxa are widespread generalists and environmental conditions are important filters. While controlled experiments are required to verify the functional consequences of the patterns observed here, we found only weak support for mycorrhizal-mediated invasions of C. stoebe and P. recta in the intermountain west."
} | 5,309 |
36932455 | PMC10024408 | pmc | 3,011 | {
"abstract": "The relationships between biodiversity-ecosystem functioning (BEF) for microbial communities are poorly understood despite the important roles of microbes acting in natural ecosystems. Dilution-to-extinction (DTE), a method to manipulate microbial diversity, helps to fill the knowledge gap of microbial BEF relationships and has recently become more popular with the development of high-throughput sequencing techniques. However, the pattern of community assembly processes in DTE experiments is less explored and blocks our further understanding of BEF relationships in DTE studies. Here, a microcosm study and a meta-analysis of DTE studies were carried out to explore the dominant community assembly processes and their potential effect on exploring BEF relationships. While stochastic processes were dominant at low dilution levels due to the high number of rare species, the deterministic processes became stronger at a higher dilution level because the microbial copiotrophs were selected during the regrowth phase and rare species were lost. From the view of microbial functional performances, specialized functions, commonly carried by rare species, are more likely to be impaired in DTE experiments while the broad functions seem to be less impacted due to the good performance of copiotrophs. Our study indicated that shifts in the prokaryotic community and its assembly processes induced by dilutions result in more complex BEF relationships in DTE experiments. Specialized microbial functions could be better used for defining BEF. Our findings may be helpful for future studies to design, explore, and interpret microbial BEF relationships using DTE. Supplementary Information The online version contains supplementary material available at 10.1186/s40793-023-00478-w.",
"conclusion": "Conclusion The dilution-to-extinction experiments involve complex microbial ecological processes. We found that deterministic processes become important with increasing dilution levels because of the selection of copiotrophs and the loss of rare species. The structural shift from oligotroph dominance to copiotroph dominance caused by dilution-to-extinction can change functional performance and lead to more complex BEF relationships in DTE studies, especially for broad functions. Microbial specialized functions could be better used for quantifying BEF relationships in DTE experiments as well as in field BEF observations. In addition, the selection of copiotrophs may cause a higher mean rrn copy number and make qPCR, if not corrected, ineffective in representing the true biomass. Our findings are helpful for future studies exploring microbial BEF relationships.",
"introduction": "Introduction Microbes are key components of biodiversity and play important roles in ecosystem functioning [ 1 , 2 ]. For a long time, the ecosystem functions of the microbial community were thought as highly redundant [ 3 , 4 ] because the diversity of the microbial community is tremendous [ 5 ], and the functional genes are highly redundant [ 6 ]. Recently, a few studies have suggested that the loss of biodiversity in microbial communities also impairs the ecosystem functioning in different ecosystems [ 7 – 9 ]. In fact, the number of biodiversity-ecosystem functioning (BEF) studies in microbial communities is much fewer than the number of BEF studies in macroscopic communities, which does not match the important roles of microbes in different ecosystems [ 10 , 11 ]. In recent years, the development of high-throughput sequencing has enabled the quantification of microbial diversity and thus facilitated the exploration of microbial BEF [ 12 ]. Dilution-to-extinction (DTE) has mostly been used to manipulate microbial diversity to study microbial BEF relationships in recent years [ 13 , 14 ]. During DTE, the high dilution level could reduce the abundance of species and then remove rare species to obtain lower diversity [ 15 ]. DTE has become an important method to study microbial BEF relationships [ 10 , 12 , 16 , 17 ] and provides evidences that rare species play vital roles in ecosystem functioning [ 15 , 18 , 19 ]. An important advancement of ecology in the last twenty years is the understanding of how stochastic processes contribute to assembling communities [ 20 , 21 ]. Now, it is well recognized that stochastic and deterministic processes shape the community together, but their relative importance in community assembly may vary [ 21 , 22 ]. It is also interesting to know how community assembly processes determine functional performance [ 22 ]. For example, many researchers believe that different community assembly processes will change BEF relationships [ 23 – 25 ]. A model study and an experimental study based on microbial communities showed that the dominance of stochastic processes would impair ecosystem functioning generating negative BEF relationships, and deterministic processes could result in positive BEF relationships [ 24 , 26 ]. Thus, understanding how microbial communities are assembled during DTE experiments is very important for in-depth analysis of the microbial BEF relationships. In ecology, there is a fundamental life strategy-based spectrum running from r -strategists, which achieve their instinct growth rate ( r max ) when resources are sufficient, to K -strategists, which maintain their population size near the carrying capacity ( K ) when resources are limited [ 27 ]. In microbial ecology, there is a framework similar to this spectrum. It is the copiotroph–oligotroph spectrum, where copiotrophs are thought to be fast-growing while oligotrophs are thought to grow slowly and efficiently [ 28 ]. The ribosomal RNA operon ( rrn ) copy number in the microbial genome is a candidate index for distinguishing copiotrophs and oligotrophs because of its good prediction of maximum growth rates [ 29 – 32 ]. In a primary succession of microbial community, copiotrophs, those with high rrn copy numbers, are dominant in early succession and later replaced by oligotrophs, those with low rrn copy numbers [ 33 – 35 ]. The abundance-weighted mean rrn copy number at the community level consequently reduced with succession of microbial community [ 33 , 35 ]. Therefore, the application of rrn copy number fitted the understanding of copiotroph-oligotroph spectrum and could help reveal processes behind community dynamics. There are some suggestions that a higher dilution level might result in a higher ratio of copiotrophs [ 10 , 17 ], because the available nutrient level is relatively high compared to the low microbial abundance in diluted communities. However, this possibility has rarely been studied. This possibility should not be neglected as copiotrophs and oligotrophs have contrasting functional traits and performance [ 28 ], which may influence BEF relationships in DTE studies. On the one hand, the selection of copiotrophs could contribute to the dominance of deterministic processes at high dilution levels. On the other hand, the community assembly of rare species is driven mainly by stochastic processes [ 36 , 37 ]; The stochastic processes could be weakened as the loss of rare species is an important process occurring at higher dilutions in DTE experiments [ 15 ]. In this study, we conducted a microcosm study using DTE and further verified our results using a meta-analysis of DTE studies. We aimed to determine how the selection of copiotrophs and loss of rare species in DTE contributed to the community assembly and how the shift in microbial community assembly processes would influence the BEF relationships.",
"discussion": "Discussion DTE, as an important method to manipulate microbial diversity, is widely used to explore BEF relationships in different ecosystems and significantly promotes our understanding of the importance of microbial diversity [ 10 , 17 , 24 ]. However, most studies focused on taxonomic diversity changes caused by DTE. Here, we observed the selection of copiotrophs and the reduction of rare species as well as enhanced deterministic processes in microbial community assembly towards higher dilutions using microcosm study and meta-analysis. These change in community structure may result in more complex BEF relationships in microbial communities when broad microbial functions are considered. We also found more neutral and less positive BEF relationships in broad functions than specialized functions in DTE studies. Selection of copiotrophs is responsible for stronger deterministic processes at a higher dilution level Deterministic processes have been demonstrated to become stronger at a higher dilution level because of reduced microbial diversity [ 19 , 24 ]. However, the communities with low diversity are not necessarily dominated by deterministic processes, as some studies found stronger stochastic processes than deterministic processes in communities with low diversity rather communities with high diversity [ 48 , 54 ]. We highlight the contributions of loss of rare species and selection of copiotrophs to community assembly. Community assembly of rare species is more commonly driven by stochastic processes than that of abundant species [ 55 ], which are also observed in this study (Figure S4 ; Figure S6 ). Loss of rare species weakens the stochastic processes, while selection of copiotrophs strengthens deterministic processes. At a higher dilution level, bacteria may spend a longer time in regrowth, i.e., obtaining biomass/abundance similar to undiluted communities [ 10 ]. During regrowth, copiotrophs could outcompete oligotrophs like what happening in the early stage of primary succession [ 33 ]. Similarly, Abreu et al. found that the copiotrophs are more likely to outcompete the oligotroph at higher dilution rates using 2- to 5-species coculture experiments [ 56 ]. This could be owing to the fast growth rates which could favor copiotrophs to quickly occupy the empty niche caused by disturbance [ 33 , 57 ] or dilution [ 56 ]. Therefore, OTUs with high rrn copy numbers are more likely to persist in highly diluted microbial communities. Betaproteobacteria , which are thought to be copiotrophs in both freshwater [ 58 ] and soil ecosystems [ 28 , 59 ] were found to increase with dilution level in both microcosm and the meta-analysis studies. The strong deterministic processes in communities with low diversity at high dilution level could be negative to ecosystems if the limited species selected by deterministic processes are not the ones carrying out important ecosystem functioning [ 23 ]. Loss of rare species causes the loss of specialized functions in DTE In real scenarios, not all species face the same danger of extinction [ 60 ]. Species with low abundance in natural ecosystems are more likely to be lost due to different stressors, habitat fragmentation and drift [ 61 ]. DTE is thought to remove rare species, meeting the need for rare species loss [ 61 , 62 ] and make DTE a popular method [ 10 , 17 ]. Rare species making up the majority in natural communities play an essential role in ecosystem functioning [ 63 ]. Many specialized functional genes are carried by rare species in microbial communities such as the sulphate reduction or phenanthrene degradation [ 47 , 64 , 65 ]. Thus, the low “redundancy” of specialized functional genes makes these functional performances more vulnerable to diversity loss [ 66 ]. For example, the abilities of chitin and cellulose degradation [ 67 ], xenobiotic carbon degradation [ 68 , 69 ], N 2 O reduction [ 70 ], sulfate reduction [ 71 ], and Fe(III) reduction [ 72 ] are easily lost within a few steps of 10-fold dilution. It explains our observation that microbial specialized functions were impaired by biodiversity loss in most DTE studies (Fig. 4 ). Furthermore, the microbial specialized functions are better used for defining BEF relationships in DTE experiments as well as in field BEF observations. Different ratios of copiotrophs might change BEF relationships Biodiversity not only includes the number of existing taxa, but also includes functional and phylogenetic information [ 10 , 73 ]. Functional diversity is well believed to be a better predictor for ecosystem functioning than simple taxonomic richness [ 73 , 74 ]. For microorganisms, some functional traits at the genome level, such as the rrn copy number studied in this study, are effective in predicting species performance [ 75 ]. When those traits are applied at the community level, they are potential to give better predictions for the differences in microbial functional performance [ 76 ]. The biodiversity effect can be divided into the complementarity effect and the sampling effect (also called as selection effect) [ 77 ]. The complementarity effect means that different species could enhance ecosystem functions through niche separation or positive interactions [ 77 ]. The sampling effect means that the dominant species may have a strong effect on ecosystem functions and that the productive species are more likely to be present in diverse communities [ 77 ]. If the dominant species strongly favor a certain ecosystem function, the functional performance could be high even in a low-diversity community [ 78 ]. In the DTE experiments, the higher ratio of copiotrophs in low diversity communities could result in higher functional performance through strong sampling effects. Compared to oligotrophs, copiotrophs adapt better to resource-rich conditions with faster growth rates and quick response to substrate addition [ 28 ]. For example, copiotrophs could utilize carbon substrates more widely and quickly than oligotrophs [ 28 , 79 ]. This could explain why we found neutral BEF relationships in the diversity of carbon substrates (Fig. 3 a, b). At the microbial community’s level, highly diluted communities may have a higher relative abundance of broad function-related genes than less diluted communities in DTE experiments [ 19 ]. Thus, when considering broad functions, the sampling effect could outweigh the complementarity effect and lead to a neutral and even negative BEF relationship. Implications for DTE The rrn copy number is well known to vary from 1 to 15 in bacterial genomes and 1 to 4 in archaeal genomes [ 80 ]. The species with a high rrn copy number could result in a high abundance in sequencing, although with a low cell number. To resolve this case, a pipeline to remove this bias has been built up [ 45 ]. We observed a significant change in the rrn copy number after dilution, which means that the rrn copy number should be used to correct the abundance data and obtain the true bacterial cell number in later DTE studies. Similarly, a correction is also needed for the result of quantitative polymerase chain reaction (qPCR). In dilution-to-extinction studies, the most widely used method to monitor the regrowth of microbial biomass is qPCR [ 9 , 24 , 81 ]. Most of the studies used it to represent the bacterial biomass directly except the study by Domeignoz-Horta et al. [ 81 ]. This method may not make a significant change in the natural community, where the ratio of copiotrophs is quite low [ 82 , 83 ] and their mean rrn copy numbers are close to others. As we proved, the mean rrn copy number increased with dilution level and, thus, the direct use of qPCR may be inaccurate to represent biomass recovery. It is important to use mean rrn copy number to correct the result of qPCR or using other methods, such as cell density using a flow cytometer [ 10 ] and microbial carbon [ 84 ] to quantify the microbial biomass."
} | 3,892 |
31275251 | PMC6593306 | pmc | 3,012 | {
"abstract": "Challenges to the reclamation of pyritic mine tailings arise from in situ acid generation that severely constrains the growth of natural revegetation. While acid mine drainage (AMD) microbial communities are well-studied under highly acidic conditions, fewer studies document the dynamics of microbial communities that generate acid from pyritic material under less acidic conditions that can allow establishment and support of plant growth. This research characterizes the taxonomic composition dynamics of microbial communities present during a 6-year compost-assisted phytostabilization field study in extremely acidic pyritic mine tailings. A complementary microcosm experiment was performed to identify successional community populations that enable the acidification process across a pH gradient. Taxonomic profiles of the microbial populations in both the field study and microcosms reveal shifts in microbial communities that play pivotal roles in facilitating acidification during the transition between moderately and highly acidic conditions. The potential co-occurrence of organoheterotrophic and lithoautotrophic energy metabolisms during acid generation suggests the importance of both groups in facilitating acidification. Taken together, this research suggests that key microbial populations associated with pH transitions could be used as bioindicators for either sustained future plant growth or for acid generation conditions that inhibit further plant growth.",
"conclusion": "Conclusions This research demonstrates that microbial activity plays a significant role in acid generation at moderately acidic pH levels where robust plants are still actively growing. The development of microbiomes in mine tailings undergoing phytostabilization at moderately acidic pH conditions appear to be characterized by competing dynamics between acidophilic organoheterotrophic and lithoautotrophic activity that develop acidic conditions and a diverse microbiome that supports plant growth. Taxonomic transitions between these two communities may serve as potential bioindicators of future conditions that either promote or inhibit plant growth. However, taxonomic transitions need to be further validated using physiologically-based studies to tease apart the functional activity between these two competing communities. Efforts to understand the transition between these acidifying and plant-supporting communities would include a more accurate understanding of diagnostic thresholds between a soil microbiome that supports continued plant growth and a microbiome that facilitates the generation of acidity. Combining insights of below ground phytostabilized soil microbiome dynamics with current knowledge of above ground plant growth may better assure sustained long-term vegetation establishment for mine tailing reclamation.",
"introduction": "Introduction Metal mining has left a lasting legacy of environmental degradation across the globe where a primary concern is the long-term generation of acid mine drainage (AMD) from residual mine wastes that contain large quantities of metal sulfide ores, such as pyrite ( Schippers et al., 2000 ; Lu and Wang, 2012 ). Acidic mine tailings and AMD microbial communities have been well-characterized as being dominated by acidophilic microbial communities that catalyze iron and sulfur oxidation, thereby accelerating the dissolution of pyrite and the continual generation of acidity ( Baker and Banfield, 2003 ; Schippers et al., 2010 ; Chen et al., 2016 ). However, the microbiology of moderately acidic mine tailings, such as those undergoing revegetation, has been less well-characterized. The kinetics of abiotic reactions involved in pyrite oxidation at circumneutral pH under aerobic conditions initially suggests that biotic activity plays a minimal role in acid generation before environmental conditions become acidic ( (Bill) Evangelou and Zhang, 1995 ). However, microenvironments between soil aggregates create niches of varying oxygen availability and pH, where microbial activity may play a significant role in acid generation. Furthermore, experimental results suggest that biological activity can play an intricate role in increasing the rate of pyrite oxidation at circumneutral and moderately acidic pH conditions ( Singer and Stumm, 1970 ; Moses and Herman, 1991 ; Ranjard and Richaume, 2001 ; Meruane and Vargas, 2003 ; Balci et al., 2007 ; Druschel et al., 2008 ; Luther et al., 2011 ). Compost-assisted phytostabilization is the active establishment of vegetation in mine tailings after the addition of a compost soil amendment to create conditions conducive to plant growth, but challenges remain to prove its long-term efficacy ( Mendez and Maier, 2008 ; Conesa et al., 2012 ). Sustained vegetative growth requires the establishment of a functioning soil substrate and the incorporation of soil amendments into pyritic mine wastes is an attempt to initiate the rapid development of these functional characteristics from undeveloped parent material ( Zanuzzi et al., 2009 ; Li and Huang, 2015 ). Studies of mine tailings undergoing both natural and anthropogenic revegetation have attempted to capture the transition of both physical and biological properties from raw mine tailings to reclaimed soil substrates capable of supporting plant growth ( Ye et al., 2000 ; Li et al., 2015 , 2016a ). Among lithotrophic populations in unamended mine tailings, there appears to be a transition of dominant populations across a pH gradient ( Chen et al., 2013 ). This transition may be similar to what is experienced during phytostabilization when acidity in mine tailings is often neutralized by a soil amendment to promote vegetation growth. Chen et al. (2014) found that during the moderate pH stage (3–5) of pyrite oxidation, genera with an average relative abundance above 5% included Tumebacillus , Alicyclobacillus , Bacillus , Acidithiobacillus , and Leptospirillum . Dominant genera in the acidic stage (pH < 3) were Ferroplasma , Sulfobacillus , and Leptospirillum . Intriguingly, several of the genera that were found to naturally colonize pyritic material were capable of both organoheterotrophic as well as lithoautotrophic metabolism. Plant growth itself has shown potential to suppress microbial populations thought to be involved in iron and sulfur oxidation and thus prevent acid generation. A study recently quantified Acidithiobacillus and Leptospirillum , common iron and sulfur oxidizing bacteria found in AMD, using quantitative PCR and determined that vegetative growth during mine tailing revegetation lowered the abundance of these key iron and sulfur oxidizing bacteria ( Li et al., 2016b ). Valentín-Vargas et al. (2014) and Valentin-Vargas et al. (2018) showed that plant growth in a revegetation mesocosm experiment acted to delay acidification and was associated with a robust presence of plant-growth-promoting heterotrophs and suppression of lithoautotrophic iron and sulfur oxidizers. Across these studies, it appears that plants have the ability to suppress certain populations of lithotrophic microbes and thereby decrease the rate of acid generation making conditions more conducive to plant growth. The study presented here contrasts with these previous studies by taxonomically evaluating the microbial dynamics initiated when a plant-growth-supporting microbial community from a compost inoculum is incorporated into a mine tailings environment dominated by acidophiles. The goal is to evaluate a compost-assisted phytostabilization strategy for highly acidic pyritic mine tailings by characterizing the microbial community succession during a 6-year field study following a single application of compost. There are two objectives addressed by this research. The first is to characterize the changes in the bacterial and archaeal microbiome over a 6-year period following a single application of compost to highly acidic pyritic mine tailings in the field. The second is to use a controlled microcosm enrichment study to identify the subset of microbial populations that enable the development and maintenance of acidic conditions when reduced iron and sulfur are present. The combined analysis of microbiome development during a 6-year field study and in a controlled microcosm experiment has elucidated bioindicators that mark transitions between plant-growth-supporting and highly acidic environmental conditions. This research will improve the understanding of bacterial and archaeal microbiome development during phytostabilization and identifies potential bioindicators of the acidification process to improve management of mine tailings reclamation efforts.",
"discussion": "Discussion This study explored microbiome dynamics in a compost-assisted phytostabilization field study for 6 years following a single application of compost at the beginning of the field trial. Results show a temporal shift over the 6 years from a diverse community capable of supporting plant growth to a community dominated by AMD-associated acidophiles. Results from the complementary microcosm experiments were used to bridge the gap between community taxonomy and function in the field study by empirically associating microbial communities with the functional ability to facilitate the generation or maintenance of acidic conditions. However, evaluation of individual population contributions to acid generation activity becomes nuanced due to poor resolution in assigning definitive taxonomy through amplicon sequencing and the fact that some relevant taxa are associated with both organoheterotrophic and lithoautotrophic metabolisms. A survey of literature characterizing cultured isolates from families that were abundant in the microcosm experiment ( Table 1 ) suggests that both organoheterotrophic and lithoautotrophic energy metabolisms are present in microbial communities that facilitate acid generation. While lithotrophic activity that catalyzes the oxidation of reduced iron and sulfur is most strongly associated with AMD microbial communities, carbon mineralization through organoheterotrophic activity indirectly promotes environmental conditions conducive to lithotrophic activity and this activity is recognized to play an important role in AMD microbial communities ( Baker and Banfield, 2003 ). Key bacteria and archaea, known to be related to acid generation, are discussed in the following sections in the context of populations identified in this study. Bacterial Populations Facilitating Acidification-Putative Lithoautotrophs Well-characterized obligate lithoautotrophs that directly facilitate the formation of acid in AMD environments were present in both the field and microcosm communities. One example is Acidithiobacillus ferroxidans , an aerobic chemolithoautotroph capable of both iron and sulfur oxidation ( Ohmura et al., 2002 ; Valdés et al., 2008 ). This species has been shown to increase the rate of ferrous iron oxidation below pH 4.5 relative to abiotic conditions, reaching a maximum rate at pH 3 ( Meruane and Vargas, 2003 ). In the present study, all reads corresponding to the family Acidithiobacillaceae were associated with 13 OTUs that were classified to the Acidithiobacillus genus level. A second example is Leptospirillum , a well-characterized acidophilic chemolithoautotroph that is highly sensitive to organic carbon and only capable of ferrous iron oxidation ( Hallmann et al., 1992 ; Sand et al., 1992 ). Recent research has characterized the species Leptospirillum ferrodiazotrophum and its ability to fix nitrogen as an important component of AMD microbial communities ( Parro and Moreno-Paz, 2003 ; Tyson et al., 2005 ). All reads corresponding to the family Leptospirillaceae were associated with seven OTUs that were classified as Leptospirillum . Both of these acidophilic families (Acidithiobacillaceae and Leptospirillaceae) were found at all times throughout the field study (Groups 1–5) and increased in abundance over time. However, as noted earlier, the relative abundance of these families alone was not a good predictor of pH in the field trial. Bacterial Populations Facilitating Acidification-Putative Organoheterotrophs Abundant OTUs from this study that have been linked to organoheterotrophic activity and acidification included four families, the Acetobacteraceae , Alicyclobacillaceae , Xanthomonadaceae , and Sulfobacillaceae . For the Acetobacteraceae, 66% of reads from the microcosm experiments and 51% of the reads from the field study corresponded to OTUs classified to the genus Acidiphilium . Acidiphilium is well-characterized as an organoheterotroph in acidic AMD microbial communities ( Wichlacz et al., 1986 ; Chen et al., 2016 ). Its ability to mineralize organic carbon may play a role in removing products inhibitory to lithotrophic microbial populations ( Marchand and Silverstein, 2002 ; Liu et al., 2011 ). In the family Alicyclobacillaceae , a reported isolate, Alicyclobacillus H1B4, has been shown to be an obligate heterotroph that also has the capacity to oxidize iron (but not sulfur) ( Joe et al., 2007 ). A second reported isolate, A. ferrooxydans, is also heterotrophic, but can oxidize both ferrous iron and reduced sulfur compounds ( Jiang et al., 2008 ). In the present study, a total of 62 and 58% of the Alicyclobacillaceae reads in the microcosm experiments and field trial respectively, corresponded to Alicyclobacillus OTUs. A subset of the Alicyclobacillus reads (78% for the microcosm experiment) and (26% for the field trial) corresponded further to A. ferrooxydans OTUs. Similarly, genera within the Xanthomonadaceae (found in the buffered moderately acidic FeO and FeSO treatments) have been found in AMD. One species in this family, Dyella thiooxydans (not found in this study), has been shown to oxidize sulfur while utilizing several carbon substrates ( Lu et al., 2010 ; Anandham et al., 2011 ; González-Toril et al., 2011 ; Chen et al., 2014 ). A second species that has been found associated with acidic environments is Metallibacterium scheffleri strain DKE6, a facultatively anaerobic iron reducer ( Ziegler et al., 2013 ). In this study, a majority of OTUs classified within the Xanthomonadaceae family were not classified to a specific genus. Finally, a species in the phylum Firmicutes , Sulfobacillus acidophilus, has been shown to oxidize both iron and sulfur with the capacity to switch between autotrophic and heterotrophic growth ( Norris et al., 1996 ; Karavaiko et al., 2006 ; Watling et al., 2008 ). In this study, 13 Firmicutes OTUs were identified as Sulfobacillus and these OTUs were primarily found under more acidic conditions; in the highly acidic microcosms and in Groups 4 and 5 of the field study. Archaeal Populations Facilitating Acidification The archaea related to acidification in the microcosm and field study samples were comprised of Ferroplasma and Thermogymnomonas , both of which are associated with highly acidic AMD conditions ( Reysenbach and Brileya, 2014 ). Comparing the potential functional niche of these two genera, Ferroplasma acidiphilum has been described as an obligate lithoautotroph ( Golyshina et al., 2000 ). A more recent study has described F. acidarmanus as capable of coupling organic carbon oxidation with ferric iron as an alternate electron acceptor under anaerobic conditions ( Dopson et al., 2004 , 2007 ). Another study characterized two strains within the Ferroplasma that did not grow without a minimal presence of yeast extract and that were also capable of tetrathionate oxidation ( Okibe et al., 2003 ). In contrast, the species Thermogymnomonas acidicola has been described as an aerobic chemoheterotroph ( Itoh et al., 2007 ). There was a distinct shift in the distribution of Ferroplasmaceae and Thermogymnomonas in the field study. The former was in higher abundance in Group 3 representing higher pH (3–4), while Thermogymnomonas was in higher abundance in Groups 4 and 5 at lower pH (2–3) in the field study ( Figure 1 ). This is intriguing since neither genus was present to any great extent in the moderately acidic microcosms but both were present in the highly acidic microcosm (pH 2–3) with Ferroplasma dominating. In the highly acidic microcosm, Ferroplasma represented 97% of the Picrophilaceae reads and Thermogymnomonas represented only 3% of reads. There are similar contrasts in the literature about Ferroplasma . It has previously been reported to be abundant in highly acidic AMD microbial communities ( Chen et al., 2014 ). However, Korehi et al. (2014) found that Ferroplasma was present during early stages (pH > 5) of natural pyrite oxidation. These results suggest that the behavior of these two archaea may be driven by a combination of pH and other biogeochemical characteristics that has yet to be fully elucidated. Microbiome Dynamics in the Field Study Broadly, results from the IKMHSS field study show an increase in relative abundance of organisms capable of lithotrophic energy metabolisms as the site acidified. The microcosm experiments demonstrate further that it is a co-establishment of potential organoheterotrophic and lithoautotrophic populations that facilitates acidification. The results confirm previous reports of organoheterotrophic organisms in the microbial communities found in highly acidic AMD and early stage moderately acidic pyrite oxidation ( Hao et al., 2010 ; Korehi et al., 2014 ). These organoheterotrophic and mixotrophic populations may fill a similar niche to that of Acetobacteraceae in mineralizing organic compounds that inhibit lithoautotrophic activity ( Marchand and Silverstein, 2002 ; Liu et al., 2011 ). Unique to this study are the insights it provides into how the native acidophilic microbial community found in highly acidic mine tailings environment becomes stressed by the addition of an organic amendment (compost) and how this community can eventually recover. Clearly, recovery of the acidophilic community following compost addition requires the participation of organoheterotrophs to mineralize organic carbon to recreate oligotrophic microenvironments that facilitate iron and sulfur oxidation. Such chemoheterotrophic activity may also be important in directly catalyzing the oxidation of pyrite. Hao et al. (2009) showed that co-cultures of chemoheterotrophic and chemolithotrophic bacteria were capable of compromising a phospholipid passivation layer prepared on the surface of pyrite crystals, thereby increasing the rate of pyrite dissolution over monocultures of a lithotrophic bacteria. We conclude that microorganisms that facilitate acid generation in moderately acidic conditions are key to determining the final outcome in a mine tailings system undergoing compost-assisted phytostabilization. Shifts in the microbial community over the course of this compost-assisted phytostabilization study can be compared to other studies of the progression in microbial communities associated with pyrite oxidation across a pH gradient. Chen et al. (2014) studied natural pyrite oxidation under greenhouse conditions across circumneutral pH to acidic conditions and described a similar set of microbial populations related to Alicyclobacillaceae , Xanthomonadaceae and Acetobacteraceae in more moderately acidic conditions. A second study examined several different mine tailings that exhibited a range of pH ( Chen et al., 2013 ). A 16 s rRNA pyrosequencing analysis of two of the mine tailings, one with higher levels of organic carbon and a pH of 6.4 and the other with lower organic carbon and a pH of 2.4, revealed the following. The site with circumneutral pH conditions had high relative abundance of Thiobacillus, Legionella, Gemmatimonas, and Sphingomonas, while the acidic site had high relative abundance of Ferroplasma , Acidithiobacillus , Leptospirillum , Sulfobacillus and Thermogymnomonas . These results suggest the co-occurrence of Ferroplasma and Thermogymnomonas in highly acidic conditions and not the distinct shift between the two populations seen in the IKHMSS field study. While the present study focused on understanding constituent populations of the microbiome that facilitate acidification during phytostabilization, microbial families that support plant growth are equally important to the success of plant establishment. The below ground biomass of plants and the ability to exude organic carbon into the amended soil stimulate organoheterotrophy and suppress lithoautotrophic activity. This process is important to success of above ground biomass production and the presence of plant-growth associated microbial populations actively helps support plant growth, especially in stressed environments ( Glick, 2010 ; Ma et al., 2011 ). In the taxonomic analysis of the field study microbiome, 8 of the 10 families that exhibited a large decrease in relative abundance under acidifying conditions were either Proteobacteria or Actinobacteria , taxonomic units strongly associated with natural soils and improved plant growth ( Glick, 1995 ; Spain et al., 2009 ). One of these families, Microbacteriaceae , has related isolates that have been shown to increase the dry weight and length of roots in Rapeseed ( Brassica napus ) grown in lead and copper contaminated soils ( Sheng et al., 2008 ; He et al., 2010 ). These results suggest that a complete understanding of how best to facilitate phytostabilization will require improved knowledge of how these potential plant-growth-promoting taxonomic families facilitate sustained vegetation growth with a parallel understanding of microbes facilitating acidification. After all, it is the dynamics between these two communities that will decide the success of a phytostabilization attempt. Therefore, using shifts in energy metabolisms as an indicator for sustained plant growth in reclaimed mine tailings is dependent on the taxa accomplishing this shift in metabolism. Studies of functional activity by transcriptional activity or functional assays would improve the understanding of differences in active energy metabolisms, as well as other functional processes, between plant growth supporting and acid generating microbial communities. Importantly, functional activity studies would be able to elucidate the contribution of mixotrophic populations in microbiome processes and determine how they actively contribute to processes such as iron and sulfur oxidation or carbon cycling across environmental conditions."
} | 5,638 |
20232871 | null | s2 | 3,013 | {
"abstract": "Postsynthetic covalent modification of metal-organic frameworks (MOFs) with long alkyl substituents is shown to protect these materials against moisture. These MOFs, which normally display hydrophilic properties, are readily converted into hydrophobic or superhydrophobic materials. Overcoming the inherent sensitivity of MOFs to water represents a major step forward in their widespread use in technology applications."
} | 104 |
35639693 | PMC9191642 | pmc | 3,014 | {
"abstract": "Significance Primary endosymbiosis allowed the evolution of complex life on Earth. In this process, a prokaryote was engulfed and retained in the cytoplasm of another microbe, where it developed into a new organelle (mitochondria and plastids). During organelle evolution, genes from the endosymbiont are transferred to the host nuclear genome, where they must become active despite differences in the genetic nature of the “partner” organisms. Here, we show that in the amoeba Paulinella micropora , which harbors a nascent photosynthetic organelle, the “copy-paste” mechanism of retrotransposition allowed domestication of endosymbiont-derived genes in the host nuclear genome. This duplication mechanism is widespread in eukaryotes and may be a major facilitator for host–endosymbiont integration and the evolution of organelles.",
"discussion": "Discussion Establishment of a novel organelle is critically dependent on the early steps of host–endosymbiont integration, yet we have little knowledge about the mechanisms that govern this process. Analyzing chromatophore and nuclear genes in P. micropora , we provide key insights into the paths and mechanisms by which endosymbiont/chromatophore genes that have been transferred to the host nuclear genome are rearranged and expressed to enable the evolution of an organelle. Here, we posit that retrotransposition is a major facilitator of these processes, contributing to the expansion, reorganization, and acquisition of light regulation by the endosymbiont-derived genes in the host. Phylogenetic analyses revealed that the HLI gene family in Paulinella has polyphyletic origins. In the common ancestor of photosynthetic Paulinella spp., multiple hli genes were likely transferred from the endosymbiont genome to the nuclear genome of the host ( Fig. 4 A ). Most HLI genes in KR01 are retrocopies and are arranged in pairs positioned in a head-to-head orientation. The HLI genes in each head-to-head retrogene pair, despite being homologs, never belong to the same clade (i.e., they are not recent duplicates of each other), which raises the question of how the clusters were formed. This head-to-head gene configuration is uncommon for HLI genes in cyanobacteria and cyanophage genomes, where they are often found in tandem head-to-tail arrangements ( 24 , 25 ). Therefore, it seems unlikely that they were transferred to the host in this arrangement; rather, they were transferred separately and later rearranged as head-to-head pairs in the host nuclear genome. In other eukaryotic genomes, retrogenes are sometimes found in a head-to-head orientation, possibly because inherently bidirectional promoters that are transcriptionally active facilitate the insertion of transposable elements flanking these active promoters ( 26 , 27 ). Thus, it is plausible that the HLI gene pairs in Paulinella were generated through their retrotransposition into the same genomic position in a divergent orientation, either simultaneously or stepwise (i.e., the insertion of one gene might have facilitated the insertion of the other). Alternatively, the conserved repeats (homopyrimidine tracts) that flank these gene pairs may have promoted recombination events that led to this genomic arrangement ( 28 ). Regardless of the mechanism underlying the generation of the gene pairs, the finding that the conserved retrotransposon domains located contiguous to the retrocopied genes are of eukaryotic origin suggests that the organization and fixation of these genes were driven by the host. Furthermore, because this HLI arrangement is found in both P. micropora and P. chromatophora , it most likely evolved during the early stages of primary endosymbiosis, predating the divergence of these species (∼60 Mya) and generated at least three different HLI pairs that contain genes from four different HLI clades (clades 1 to 4) ( Fig. 4 A ). In addition, we found that most of these retrogenes are differentially regulated in response to HL, whereas the nonretrogenes are mostly not HL responsive, despite being transcriptionally active. This observation leads us to hypothesize that the HLI genes acquired HL-responsive regulatory elements through retrotransposition. We also found that these HLI retrogene pairs were later expanded through DNA-mediated duplications, independently in the genomes of these two Paulinella species. This demonstrates a convergent evolutionary path in which the HL-regulated retrocopies are preferentially retained. Fig. 4. Hypothetical model for the evolution of the HLI gene family in Paulinella. ( A ) At the early stages of primary endosymbiosis, before the divergence of P. chromatophora and P. micropora species, different HLI genes were transferred from the cyanobacterial endosymbiont/chromatophore into the host nuclear genome via independent EGT events. These genes were duplicated and retrotransposed, forming at least three kinds of head-to-head gene pairs with different HLI gene compositions. After P. chromatophora and P. micropora diverged, the HLI gene pairs were highly duplicated, independently in the two species (via a DNA-mediated mechanism), resulting in expansion of this gene family. Genes were also lost after transfer to the nuclear genome as a consequence of pseudogenization because they were not effectively expressed and/or not functionally effective. ( B ) The ancestral EGT-derived HLI genes were most likely associated with their original promoter and not regulated by stress in the host’s nuclear genome. Different HLI paired retrocopies may have associated with or acquired a bidirectional promoter that is inducible by stress, leading to an early-persistent expression pattern through retrotransposition. Extensive duplications of the HLI pairs and degeneration of this ancestral promoter could have resulted in the evolution of two new expression patterns that are complementary, partitioning the early-persistent pattern into the more specialized early and late expression patterns. In addition to the HLI retrogenes, we identified active retrocopies of the EGT-derived PSAE and PSAI genes, supporting our hypothesis that retrotransposition has facilitated adaptation of the endosymbiont-derived genes in the amoeba genome. Gene duplications facilitate adaptation by providing new genetic material for mutation, drift, and selection to act upon ( 29 ). Moreover, the “copy-paste” mechanism of retrotransposition provides additional opportunities for the acquisition of host gene promoters with adaptive regulatory features ( 30 ). Thus, we hypothesize that retrotransposition plays a key role in “rewiring” the expression of prokaryotic genes transferred to eukaryotic genomes by replacing their original promoters, which may not be suitable for expression in the eukaryotic host, with those of host origin, which may confer appropriate regulatory features. Indeed, we found that most of the HL-induced HLI genes were generated through retrotransposition, with most of the nonretrocopies being transcriptionally active but not HL regulated. Furthermore, the HL-regulated copies of PSAE3 and PSAI are retrocopies. Consistent with and broadening this hypothesis are the findings that genes transferred from bacteria to worm and insect eukaryotic genomes are embedded in a region of DNA enriched in transposon/retrotransposon elements that may mediate gene duplication events in the host ( 4 , 31 , 32 ). Moreover, retrotransposition played a role in the expansion of genes involved in acclimation to stress, which included bacteria-derived genes, in the cold-adapted green microalga Chlamydomonas antarctica ( 33 ). Overall, our data strongly suggest that retrotransposition of Paulinella genes can generate multiple gene copies that enable the rewiring of genes and their regulatory features to sustain host–endosymbiont interactions. This process provides the foundation for the evolution of additional molecular changes that favor a more efficient and resilient primary endosymbiotic association. Because of its potential to generate deleterious effects, retrotransposition is suppressed under normal conditions. However, when organisms face extreme stress, this mechanism can become active and induce extensive genetic rearrangements that can lead to more “adaptive” functions and/or the evolution of novel functions that enhance cell survival ( 23 ). For instance, large-scale retrotransposition events appear to be triggered by dramatic changes in climate/conditions, resulting in expanded gene families that facilitate stress acclimation. This process is also associated with photosynthesis and the establishment of the symbiosis between cnidarians and their Symbiodiniaceae endosymbionts ( 18 , 34 ). Similarly, retrotransposition has been associated with the adaptation of C. antarctica to cold temperatures ( 33 ). Our results show that the HLI , PSAE , and PSAI retrogene copies acquired the ability to respond to HL-induced stress in KR01 during the period in which the endosymbiont was transitioning from a free-living organism to a nascent organelle inside the Paulinella host. Because the photosynthetic endosymbiont in Paulinella produces ROS even under relatively low-light conditions ( 7 ) and retrotransposons are induced by oxidative stress in plants, fungi, and mammals ( 35 – 38 ), we hypothesize that during the early stages of the integration of host–endosymbiont metabolisms, the discord between the metabolisms of two organisms elicited high and sustained levels of ROS production that induced retrotransposon-dependent gene duplications. These events could have elevated Paulinella resilience by increasing gene copy number and facilitating altered regulation of endosymbiont-derived genes critical for acclimation of the cyanobacterial ancestor of the chromatophore to oxidative stress. The finding that HLI gene pairs generated through retrotransposition in the common Paulinella ancestor were later expanded independently in each of the two Paulinella species suggests that the retrocopies conferred an advantage to the host and were subject to selection. For the evolution of the photosynthetic Paulinella spp., an increase in HLI gene dosage could lessen the potential damaging impact caused by the absorption of excessive excitation energy/light stress. An increase in copy number could have also allowed mutations in these genes that enhance or optimize transcriptional regulation in the host genome to become fixed, tuning HLI levels to the physiological conditions. Here, we identified three HLI expression patterns in response to HL stress: early, late, and early-persistent. Because all clades of HLI retrocopies show the early-persistent pattern, this expression pattern was likely ancestral to the other two that may have evolved later. Interestingly, the combined expression features of the two putative newly evolved patterns are complementary to the putative ancestral pattern (i.e., early + late = early-persistent). The duplication, degeneration, and complementation model for the preservation of gene duplicates predicts that mutations in regulatory elements can increase the probability of gene preservation, usually by creating more specific gene functions and patterns of expression by partitioning ancestral functions rather than evolving new ones ( 39 ). Partitioning of gene expression has been shown in humans and yeast, where it leads to tissue- or subcellular localization–specific expression ( 40 – 42 ). Here, we hypothesize that the two newly evolved expression patterns in KR01 resulted from modification of the ancestral promoter to allow the “subfunctionalization” of different HLI genes—not in space, but in time—allowing the organism to more effectively cope with different features of light stress (e.g., photosystem damage during initial light stress versus ROS accumulation after prolonged HL exposure) ( Fig. 4 B ). These optimization steps could reduce the oxidative stress generated by the nascent organelle and accelerate host–endosymbiont integration. By uncovering the genetic mechanisms necessary for domestication of endosymbiont-derived genes in the host nuclear genome of Paulinella spp., we provide insights into primary endosymbiosis. The optimization of transcriptional control of these genes by the host via extensive retrotransposition (potentially promoted by light stress) has allowed the loss of the ancestral gene copies in the endosymbiont genome. This led to a genetic dependency on the host that was likely key for stabilizing the integration of the partner organisms and enabling the evolution of a new photosynthetic organelle. Finally, even though this study reveals a major role for retrotransposition in facilitating the evolution of a photosynthetic organelle in Paulinella spp., this duplication mechanism is ubiquitous in many eukaryotes. Thus, the role of retrotransposition in gene domestication might extend to other eukaryotic genomes, with horizontally transferred genes that become established and functionally active to evolve adaptive or novel functions in their recipient organisms ( Fig. 5 ). Fig. 5. Hypothetical role of retrotransposition in the domestication of prokaryotic genes in eukaryotic genomes based on the Paulinella model. Genome fragments containing genes are transferred from the endosymbiont into the host nuclear genome ( A ). The recently transferred endosymbiont genes of prokaryotic origin are likely not optimally expressed in the eukaryotic genome ( B ). Retrotransposition events cause duplication and relocation of the genes in different locations in the nuclear genome ( C ) that allow the acquisition of new promoters ( D ). These new promoters, in addition to optimization over time, improve transcription of the EGT-derived gene retrocopies. Additional DNA-based duplications allow further optimization of transcription and subfunctionalization of some of the retrocopies, while the suboptimally regulated copies become pseudogenized ( E ). The optimized transcriptional control of the endosymbiont-derived genes in the host nucleus relaxes selective pressure on the endosymbiont copies, resulting in their pseudogenization and loss over time. Consequently, endosymbiont genes become under host control in the nucleus while the endosymbiont genome is reduced. mRNA, messenger RNA."
} | 3,604 |
22792198 | PMC3391224 | pmc | 3,015 | {
"abstract": "Biofilms play an important role as a settlement cue for invertebrate larvae and significantly contribute to the nutrient turnover in aquatic ecosystems. Nevertheless, little is known about how biofilm community structure generally responds to environmental changes. This study aimed to identify patterns of bacterial dynamics in coral reef biofilms in response to associated macrofouling community structure, microhabitat (exposed vs. sheltered), seasonality, and eutrophication. Settlement tiles were deployed at four reefs along a cross-shelf eutrophication gradient and were exchanged every 4 months over 20 months. The fouling community composition on the tiles was recorded and the bacterial community structure was assessed with the community fingerprinting technique Automated Ribosomal Intergenic Spacer Analysis (ARISA). Bacterial operational taxonomic unit (OTU) number was higher on exposed tiles, where the fouling community was homogenous and algae-dominated, than in sheltered habitats, which were occupied by a variety of filter feeders. Furthermore, OTU number was also highest in eutrophied near-shore reefs, while seasonal variations in community structure were most pronounced in the oligotrophic mid-shelf reef. In contrast, the macrofouling community structure did not change significantly with seasons. Changes in bacterial community patterns were mostly affected by microhabitat, seasonal and anthropogenically derived changes in nutrient availability, and to a lesser extent by changes in the macrofouling community structure. Path analysis revealed a complex interplay of various environmental and biological factors explaining the spatial and temporal variations in bacterial biofilm communities under natural conditions.",
"introduction": "Introduction Microbial biofilms play an important role in aquatic systems by providing a conditioned surface for larval settlement and metamorphosis of sessile organisms [1] and by contributing to nutrient turnover and productivity [2] – [3] . Biofilms generally have a high microbial diversity, which is maintained by exogenous [4] as well as endogenous [3] mechanisms. Exogenous drivers may consist of either a top-down control by predation or viral lysis of bacteria, which limits the dominance of certain species in the community and would allow for the co-existence of different species within the same niche, or a bottom-up control may consist of the wide variety of energy sources and substrates available in an ecosystem, which offer a large variety of niches for bacteria [4] . Endogenous mechanisms include interactions among microbial species, with dynamic exchanges of metabolites, which thereby further contribute to the formation of various ecological niches [3] . Although diversity is generally high, biofilm community structure can greatly vary with changes in environmental conditions [5] – [6] , such as nutrient availability, temperature, salinity and light, which can moreover fluctuate over space and time [2] , [7] – [9] . Nutrient availability was found to be one of the major factors affecting biofilm diversity and composition (reviewed by [2] ), and to vary with seasons [7] , [10] or with human impact, i.e. eutrophication [9] , [11] – [12] . Higher nutrients generally cause a shift from autotrophic to heterotrophic and to sulphur reducing bacteria as a response to decreased light availability and increased load in organic material [9] , [11] , [13] , while the overall biofilm diversity has been found to either remain unaffected [8] , [14] or to increase [12] , [15] . Those changes in the biofilm community may further affect the behaviour and success of larval settlement of sessile macroorganisms [16] – [19] . Conversely, macroorganisms reacting to environmental changes may modify their chemical composition and consequently affect their associated bacterial community [20] – [21] . Negri et al. [22] demonstrated that the algae encrusted substrate affected the overlying bacteria, which further influenced the settlement of coral larvae in laboratory experiments [22] . These studies demonstrate well the interactions between biofilm composition and macrofouling, where initial biofilm formation influences the settlement of macroorganisms and the macroorganisms may in turn affect biofilm formation (see also [23] – [24] ). However, these studies were all species-specific and so far did not consider community level approaches so as to assess the interplay between microbial and macrobial communities (assembly of fleshy and calcareous algae, bryozoans, ascidicans, barnacles, spirorbid worms, etc.). Furthermore, biofilm studies on hard substratum in coral reefs are scarce. To the best of our knowledge, our study is the first attempt to date to simultaneously address the interactions of microbial biofilms and macrofouling communities in coral reefs in response to spatio-temporal changes in the environment. Understanding these dynamics is of increasing importance in the context of anthropogenic water quality changes, which may strongly affect the interplay of the two communities and consequently the stability of the coral reef ecosystem. Here, the diversity and dynamics of colonizing bacterial communities were investigated on tiles, which were deployed so as to create sheltered and exposed microhabitats in several coral reefs of the Spermonde Archipelago, Indonesia, over 20 months ( Fig. 1 ). Spermonde is characterized by an eutrophication gradient between nutrient-rich coastal and oligotrophic offshore waters and by seasonal changes in nutrient input and turbidity mainly due to variations in rain fall. The macrofouling community that settled on the tiles was included in the analyses to examine its relationship with biofilm diversity and dynamics. The following main questions were addressed: (1) How much do eutrophication, geographic location and seasonality affect bacterial diversity and community structure? (2) How important is the presence and composition of the macrofouling community for microbial dynamics in the context of microhabitats, eutrophication and seasonality? (3) What effects do microhabitats have on the bacterial community structure? By disentangling the respective effects of space, season, and environmental parameters by using a combination of multivariate statistical tools, a community ecology approach was applied to examine the principles underlying the assembly and shifts of the microbial community and its reciprocal interactions with the macrofouling communities in their natural environment. 10.1371/journal.pone.0039951.g001 Figure 1 Map of the Spermonde Archipelago, SW Sulawesi (Indonesia) with the study sites indicated by stars: Near-shore Lae Lae (LAE), near mid-shelf Samalona (SAM), mid-shelf Bonebatang (BBA) and off-shore Lanyukan (LNK).",
"discussion": "Discussion The aim of this study was to assess bacterial diversity patterns of biofilms in coral reefs, their dynamics in relation to changes in the associated macrofouling community and their response to seasonal and spatial variations in environmental conditions. Changes in bacterial community structure associated with hard or sediment surfaces in higher latitude coral reefs have already been related to spatial differences in response to eutrophication [13] , light availability (depth gradient) [19] , wave energy [25] and seasonality [26] , while the effect of associated macrofouling communities has not been addressed so far. Additionally, our study provides first data on the dynamics of bacterial communities in biofilms of equatorial coral reefs in the global centre of marine biodiversity [27] . Here, the investigated reefs were exposed to seasonal fluctuations mostly determined by the amount of nutrient supply through land run-off (rain fall) and to a lesser extent by temperature or light availability [28] . The results of our study revealed significant environmental effects, reflected in marked differences in bacterial community structure, which could be related to microhabitat, eutrophication, season, and to changes in the fouling community. Effect of Microhabitat (Sheltered vs. Exposed) The microhabitats created by the tiles provided a sheltered/shaded and exposed/high light environment. They were found to be associated with the greatest differences in OTU number and significant changes in community structure of colonizing bacteria as compared to seasonal or regional influences. While the sheltered habitat was characterized by low light and a heterogeneous macrofouling community structure that would have provided a variety of surfaces for bacterial colonization, the exposed habitat was characterized by high light and a rather homogeneous algal cover. Interestingly, bacterial OTU number was higher on the exposed tiles, suggesting that bacterial communities may depend more on energy availability than on the offered substrate diversity. Organic material permanently settles on the upper tiles and is produced by the algae themselves, which may thus provide a large variety and quantity of nutrients for bacterial growth [29] and diversification [2] , [15] . Additional reasons for a higher diversity on the exposed tiles may be a higher disturbance (e.g. grazing by fishes on the upper tiles), which continuously varies UV exposure (changes in UV may change bacterial communities; [30] ) and thus creates space for new species to establish [6] . Although the high densities of cryptic filter-feeders are also known to harness significant amounts of organic material on both, artificial [31] and natural reef surfaces [32] , thus potentially enriching this microhabitat, the reduced photosynthesis and lack of sedimentation may have reduced the overall supply of organic material to the microbial communities. In addition, the potentially antibacterial activity of the diverse secondary metabolites produced by macrofouling organisms (e.g. [20] , [33] – [34] ) may have contributed to the lower OTU numbers on the lower side of the tiles. Effect of Eutrophication The eutrophication gradient was clearly evidenced in the macrofouling community structure by a shift from filamentous to crustose algae, which was previously described as an indicator for eutrophication [35] – [36] . Also the benthic community structure revealed strong signs of eutrophication and pollution in the most near-shore reef by a shift from hard corals to soft heterotrophic filter-feeding organisms [37] . However, the eutrophication gradient was less evidenced by the measured water parameters, although POC and Chl a have been found to be useful indicators in this area [38] , most likely due to the localized and discontinuous nature of the water sampling campaign in this study. Along the decreasing eutrophication gradient, bacterial communities were characterized by decreasing OTU numbers ( Fig. 2D ), but by increasing bacterial community heterogeneity (i.e. beta diversity; Fig. 4 ). Higher local bacterial diversity might be obtained in accordance to the energy-diversity relationships (e.g. [39] – [40] ), which would favour high diversity under higher nutrient availability in near-shore reefs. This explanation would be further supported by our observations at the microhabitat level, where the bacterial community on the upper tiles experienced a stronger effect of eutrophication, since they harbour dense filamentous algae carpets that may serve as effective traps for sediment, particulate organic matter and associated pollutants in near-shore reefs. The more similar community structure obtained in the eutrophied areas would be supported by previous observations of wider range of energy sources available in human impacted areas, which allow for more similar species to co-exist and for the selection of bacterial species that are generally more robust against the higher load of pollutants and toxins in eutrophied environments [15] . At the less impacted sites, the opposite processes would then be true, with less impact on local richness but more on the variation in community structure. From a technical viewpoint, it should also be noted that ARISA fingerprinting may not necessarily provide an accurate depiction of bacterial species richness, because e.g. the discrimination between bacterial types is based on ITS sequence length and ribosomal operon copy number may vary across microbial genomes [41] , which would overall result in overestimating OTU number. Shifts in community structure based on ARISA community fingerprinting are yet useful to infer changes in diversity across samples and are generally in good agreement with patterns obtained by sequencing-based approaches of ribosomal genes (e.g. [26] , [42] – [43] ). Changes in water quality can be crucial for some bacterial species and hence may limit their distribution. This seemed to be the case for 29 OTUs, which change in relative abundance and distribution was purely due to water parameters, regardless of variation in geographic location or seasons ( Fig. 5 ). Such behaviour has been previously related to that of specialist species, while species purely affected by location and not by environmental conditions or seasonal fluctuations were considered generalist [44] . Here we followed this classification but, in addition, propose to also classify as generalists OTUs that were found at least once at each site regardless of season, tile side, or environmental conditions (n = 267 OTUs, from which 61 were already considered in the OTU numbers reported in Fig. 5 ). Due to limitations in detecting rarer OTUs, it should be mentioned that OTUs only found at a site may in fact have broader distribution range in the environment. Effect of Seasonality Seasonality was clearly reflected in the variation of water parameters, bacterial community structure, but only weakly in changes of the fouling community. The OTU composition in mid-shelf reefs (SAM and BBA) revealed a pronounced seasonality, while the near- and off-shore reefs (LAE and LNK) featured a lower seasonality. This suggests a lower dependency of the bacterial community on substrate availability (i.e. fouling community), but rather a dependency on nutritional load. To explain those observations, we may thus propose the following scenario: At near-shore LAE, chronically high nutrient supply selects for specific bacterial community structure throughout the year. At near mid-shelf SAM, OTU numbers were high throughout the seasons and community structure was variable. This may be explained by a nutrient supply, which is comparatively high but still variable in its composition throughout the year, entailing variations in community structure. The oligotrophic mid-shelf BBA was associated with the largest differences in OTU number and community structure. Low OTU numbers were evidenced during the most nutrient-depleted period that occurs at the end of the transition period from wet to dry season, while tiles sampled in the dry season experienced occasional impacts of land run-off (rainfall) at the end of the season, which increased nutrient load levels and OTU number. At off-shore LNK, other oceanic factors (e.g. currents, upwelling) weakened the effects of monsoon related variability in nutrient supply (e.g. land run-off) that normally dominate on the shelf. Although each factor was found to significantly affect changes in community structure when all parameters were included in the analyses, the overall low amount of explained community variation (variation partitioning) found in our study may be typical of classical community ecology studies (e.g. [45] – [46] ). This is also consistent with the idea that environmental variables significantly explaining variation in community structure should be taken as putative ecological scenarios rather than as true reconstruction of the sources of variation under complex natural conditions. Relationships between Bacterial and Macrofouling Communities Numerous studies have demonstrated the effects of biofilms on larval settlement by revealing large variations in species-specific responses and sensitivities (reviewed by [1] ). A strong interplay between biofilm compositions and macrofouling has been suggested, with reciprocal effects between the biofilm community and settling macroorganisms [22] – [24] . Such reciprocal interactions between the two communities were also detected in our study by the path analysis approach, further justifying the use of complementary multivariate statistical approaches. However, the dependence was rather of weak magnitude due to differential responses of the two communities to external changes. Indeed, eutrophication clearly affected both macrofouling and bacterial communities, while seasonal fluctuations in environmental parameters only affected bacterial community structure, but hardly the macrofouling community structure. Although we cannot exclude that the characterization of the two communities at different levels of taxonomic resolution might have masked some patterns of community covariation, the following reasons may most likely explain the low dependence: First, microorganisms may be expected to have much shorter generation times and larger population sizes than macroorganisms, which lead to much higher dynamics in bacterial than in macrofouling communities in response to environmental changes. Second, after 4 months of tile deployment, bacterial populations would have established their own microenvironments [2] – [3] , which might have sheltered them from subsequent changes in the macrofouling community. The same may be valid for the macrofouling community, which also reaches an advanced state of succession after 4 months [47] , allowing a certain independence from surface conditioning. Third, the bioactivity of secondary metabolites of the macrofouling community changes with seasons and this may affect microbial communities [48] , while the macrofouling community structure may itself remain mostly unchanged. In conclusion, this study identified bacterial patterns in biofilms of equatorial coral reefs and the likely factors that significantly affect them, be they variations in microhabitat, eutrophication level, seasonality or co-occurring macrofouling community. The complex interplay of all those factors was disentangled and resulted in new hypotheses: Not only nutrient availability and its seasonal fluctuations, but also the specific locations and degree of exposure, may be key parameters that shape bacterial community structure, diversity and ultimately functions. Because all factors had significant, yet modest contributions, other yet-unknown factors may also be at play in the study area and would need to be identified in the future (e.g. succession, secondary metabolites, other environmental parameters), as well as the impact of inter- and intra-species interactions within each of the investigated communities."
} | 4,730 |
23070318 | null | s2 | 3,016 | {
"abstract": "Polymicrobial interactions are widespread in nature and play a major role in maintaining human health and ecosystems. Whenever one organism uses metabolites produced by another organism as energy or nutrient sources, it is called cross-feeding. The ecological outcomes of cross-feeding interactions are poorly understood and potentially diverse: mutualism, competition, exploitation, or commensalism. A major reason for this uncertainty is the lack of theoretical approaches linking microbial metabolism to microbial ecology. To address this issue, we explore the dynamics of a one-way interspecific cross-feeding interaction in which food can be traded for a service (detoxification). Our results show that diverse ecological interactions (competition, mutualism, exploitation) can emerge from this simple cross-feeding interaction and can be predicted by the metabolic, demographic, and environmental parameters that govern the balance of the costs and benefits of association. In particular, our model predicts stronger mutualism for intermediate by-product toxicity because the resource-service exchange is constrained to the service being neither too vital (high toxicity impairs resource provision) nor dispensable (low toxicity reduces need for service). These results support the idea that bridging microbial ecology and metabolism is a critical step toward a better understanding of the factors governing the emergence and dynamics of polymicrobial interactions."
} | 367 |
31466527 | PMC6716841 | pmc | 3,017 | {
"abstract": "Background Poly(( R )-3-hydroxybutyrate- co -( R )-3-hydroxyhexanoate) [P(3HB- co -3HHx)] is a bacterial polyester with high biodegradability, even in marine environments. Ralstonia eutropha has been engineered for the biosynthesis of P(3HB- co -3HHx) from vegetable oils, but its production from structurally unrelated carbon sources remains unsatisfactory. Results Ralstonia eutropha strains capable of synthesizing P(3HB- co -3HHx) from not only fructose but also glucose and glycerol were constructed by integrating previously established engineering strategies. Further modifications were made at the acetoacetyl-CoA reduction step determining flux distribution responsible for the copolymer composition. When the major acetoacetyl-CoA reductase (PhaB1) was replaced by a low-activity paralog (PhaB2) or enzymes for reverse β-oxidation, copolyesters with high 3HHx composition were efficiently synthesized from glucose, possibly due to enhanced formation of butyryl-CoA from acetoacetyl-CoA via ( S )-3HB-CoA. P(3HB- co -3HHx) composed of 7.0 mol% and 12.1 mol% 3HHx fractions, adequate for practical applications, were produced at cellular contents of 71.4 wt% and 75.3 wt%, respectively. The replacement by low-affinity mutants of PhaB1 had little impact on the PHA biosynthesis on glucose, but slightly affected those on fructose, suggesting altered metabolic regulation depending on the sugar-transport machinery. PhaB1 mostly acted in the conversion of acetoacetyl-CoA when the cells were grown on glycerol, as copolyester biosynthesis was severely impaired by the lack of phaB1 . Conclusions The present results indicate the importance of flux distribution at the acetoacetyl-CoA node in R. eutropha for the biosynthesis of the PHA copolyesters with regulated composition from structurally unrelated compounds.",
"conclusion": "Conclusions We herein engineered Ralstonia eutropha for the biosynthesis of P(3HB- co -3HHx) from structurally unrelated glucose and glycerol by conferring glucose assimilation ability, enhancing glycerol assimilation ability, and installing an artificial pathway for biosynthesis of the copolyester. Further modifications at the acetoacetyl-CoA reduction step demonstrated the importance of flux distribution at the acetoacetyl-CoA node in R. eutropha for the biosynthesis of the PHA copolyesters with regulated composition from the structurally unrelated compounds. The moderate weakening of the ( R )-specific reduction of acetoacetyl-CoA or enhancement of reverse β-oxidation allowed efficient biosynthesis of the copolyesters from glucose with high 3HHx composition, possibly due to enhanced formation of butyryl-CoA, a precursor of the C 6 -intermediates, from acetoacetyl-CoA via ( S )-3HB-CoA. This study can provide important information for the engineering of R. eutropha for the production of PHAs as well as other acetyl-CoA-derived compounds.",
"introduction": "Introduction of low affinity mutants of PhaB1 and the effects on PHA biosynthesis from glucose We had attempted protein engineering of PhaB1 based on the crystal structure [ 37 ], and obtained mutants with low affinity toward acetoacetyl-CoA, as described in Additional Information. The kinetic parameters of PhaB1 and the mutants using N -His 6 -tagged recombinant proteins are shown in Table 1 . We confirmed very high affinity of PhaB1 to acetoacetyl-CoA with a K m value of 2 μM, as previously reported [ 37 , 38 ], and observed substrate inhibition at acetoacetyl-CoA concentrations higher than 12 μM. N142V and Y185F mutants of PhaB1 (designated as PhaB1 NV and PhaB1 YF ) showed much larger K m values toward acetoacetyl-CoA, whereas V max values were not affected by the N142V mutation and retained at 78% by the Y185F mutation, when compared with those of the parent wild-type enzyme. The double mutations of N142V and Y185F markedly decreased catalytic efficiency, as K m and V max of the double mutant PhaB1 NVYF to acetoacetyl-CoA increased 50-fold and decreased by one-third of those of PhaB1, respectively. Table 1 Kinetic parameters of PhaB paralogs and PhaB1 mutants from R. eutropha toward acetoacetyl-CoA Enzyme Mutation(s) K m [μM] V max [U mg −1 ] k cat [s −1 ] k cat / K m [μM −1 s −1 ] PhaB1 1.99 ± 0.23 162 ± 6 71.2 ± 2.7 35.8 ± 4.4 PhaB1 NV N142V 58.5 ± 18.1 175 ± 31 76.7 ± 2.7 1.31 ± 0.47 PhaB1 YF Y185F 86.2 ± 20.6 127 ± 17 55.8 ± 7.6 0.65 ± 0.18 PhaB1 NVYF N142V/Y185F 109 ± 38 53.9 ± 8.3 23.7 ± 3.7 0.22 ± 0.08 PhaB2 2.48 ± 0.63 10.5 ± 0.6 4.90 ± 0.28 2.00 ± 0.52 PhaB3 1.27 ± 0.53 88.9 ± 8.7 40.8 ± 4.0 32.7 ± 14.0 PhaB1 NV (low-affinity) and PhaB1 NVYF (low-affinity and low-activity) were here applied with the aim of achieving moderate weakening of the ( R )-specific reduction step in the P(3HB- co -3HHx) biosynthesis pathway. The two mutant genes were individually introduced into the strain NSDG-GG-∆B1 downstream of phaA in the chromosome of NSDG-GG-∆B1 (original phaB1 locus) (Fig. 1 ), and the resulting strains NSDG-GG-B1 NV and NSDG-GG-B1 NVYF were used as hosts for pBPP-ccr Me J4a-emd or pBPP-ccr Me J Ac -emd. When these strains were cultivated on glucose, however, unexpectedly the copolyester biosynthesis properties (entries 4, 5, and 11 in Fig. 2 and Additional file 1 : Table S1) were not greatly changed when compared with those of the phaB1 -deleted strains (entries 3 and 9), except for NSDG-GG-B1 NV /pBPP-ccr Me J Ac -emd (entry 10) that accumulated slightly more P(3HB- co -3HHx) than the phaB1 -deleted strain.\n\nIntroduction of low-activity paralog PhaB2 and the effects on PHA biosynthesis from glucose Although a previous study revealed the roles of the three PhaB paralogs in P(3HB) biosynthesis by R. eutropha [ 14 ], the catalytic properties of PhaB2 and PhaB3 had yet to be determined. Table 1 also shows the results of kinetic analysis of PhaB2 and PhaB3 using the N -terminal His 6 -tagged recombinant proteins. Both PhaB2 and PhaB3 showed very high affinity to acetoacetyl-CoA, with K m values of 2.5 μM and 1.3 μM, respectively, and substrate inhibition as well as PhaB1, while the V max values of PhaB2 and PhaB3 were one order of magnitude lower and about half, respectively, when compared with that of PhaB1. As PhaB2 was supposed to be applicable as a low-activity reductase for the moderate weakening of the ( R )-specific reduction of acetoacetyl-CoA, the strain NSDG-GG-B2 was constructed by inserting phaB2 downstream of phaA in the chromosome of NSDG-GG-∆B1 (Fig. 1 ). The strain transformed with pBPP-ccr Me J4a-emd or pBPP-ccr Me J Ac -emd accumulated P(3HB- co -5.9–7.0 mol% 3HHx) with 71.4–72.3 wt% cellular content on glucose (entries 6 and 12 in Fig. 2 ), which demonstrated that the insertion of phaB2 into the pha operon increased PHA production when compared with the phaB1 -deleted strains. The amounts of the 3HB and 3HHx units incorporated into the polyester were intermediate between those by NSDG-GG ( phaB1 + ) and NADG-GG-∆B1 ( phaB1 – ) strains (Additional file 1 : Table S1). This was consistent with the altered flux distribution from acetoacetyl-CoA to ( R )-3HB-CoA and ( S )-3HB-CoA by PhaB2. No significant difference in the PHA production was observed between the strains harboring PhaJ Ac and PhaJ4a.",
"discussion": "Discussion The microbial production of P(3HB- co -3HHx) has usually utilized vegetable oils or fatty acids as carbon sources because the provision of ( R )-3HHx-CoA monomer can be simply achieved by ( R )-specific hydration of 2-enoyl-CoA intermediate in β-oxidation catalyzed by R -hydratase (PhaJ) [ 12 ]. The industrial production of P(3HB- co -3HHx) from palm oil by recombinant R. eutropha has been demonstrated by Kaneka Corp., Japan, since 2011. In addition to vegetable oils, the use of other inexpensive biomass feedstocks, such as sugars and glycerol, is expected to be another promising way of achieving low-cost production and consequent wide applications. From this perspective, we previously engineered R. eutropha for the expansion of utilizable carbon sources [ 28 , 32 ] and for the generation and polymerization of ( R )-3HHx-CoA from fructose through an artificial pathway [ 23 ]. These were here integrated into R. eutropha , which led to the construction of strains capable of synthesizing P(3HB- co -3HHx) from not only fructose but also glucose and glycerol. However, the yield and 3HHx composition of the copolyesters were insufficient, so further investigation focused on improving the strains to achieve efficient production of the copolyesters with a higher 3HHx fraction. NADPH-dependent acetoacetyl-CoA reductase (PhaB) is an ( R )-3HB-CoA-providing enzyme in most P(3HB)-producing microorganisms, and R. eutropha possesses three paralogs of PhaB (PhaB1, PhaB2, PhaB). The roles of the multiple PhaBs in R. eutropha have been investigated and discussed based on the biosynthesis of P(3HB) homopolymer [ 14 ]. PhaB1 from R. eutropha has also been frequently applied in PHA biosynthesis by engineered E. coli strains. Nevertheless, the effects of modifications in the acetoacetyl-CoA reduction step on the biosynthesis of PHA copolymers have not been well considered. In the case of P(3HB- co -3HHx) biosynthesis from fructose by the previously engineered R. eutropha [ 23 ], deletion of phaB1 was an important modification for the incorporation of 3HHx unit into the polyester fraction, although this accompanied the reduction of PHA production. It was supposed that ( R )-3HB-CoA provision in the phaB1 -lacking strain, supported by the minor paralog PhaB3 (associated with only 2%–5% of total NADPH-dependent acetoacetyl-CoA reductase activity in cell extracts of R. eutropha [ 14 , 33 ]), was significantly weakened when compared with the parent phaB1 + -strains. This tradeoff between production and composition of the copolyesters depending on the presence or absence of PhaB1 coincided with the formation of butyryl-CoA mainly via ( S )-3HB-CoA, and suggested the importance of flux distribution at the acetoacetyl-CoA node for the copolyester biosynthesis through the artificial pathway. We assumed that moderate weakening of the ( R )-specific reduction of acetoacetyl-CoA would establish metabolic flux distribution from acetoacetyl-CoA to ( R )- and ( S )-3HB-CoAs suitable for P(3HB- co -3HHx) synthesis. The R. eutropha strains were thus modified by introducing low-affinity mutants (PhaB1 NV and PhaB1 NVYF ) and a low-activity paralog (PhaB2) of PhaB1 for moderate weakening of the ( R )-specific reduction. On glucose, the use of PhaB2 instead of PhaB1 resulted in production of the copolyester with cellular content and 3HHx composition in-between those by the phaB1 + and ∆ phaB1 strains, whereas the two kinds of low-affinity mutant of PhaB1 did not significantly affect PHA biosynthesis. These results demonstrated that the flux distribution from acetoacetyl-CoA toward C 4 - and C 6 -monomers can be regulated by specific activity levels of the reductase when the enzyme retains high affinity to acetoacetyl-CoA. As previous metabolomic analysis of R. eutropha showed a very low intracellular concentration of acetoacetyl-CoA [ 40 ], high substrate affinity of the reductase was required to change the metabolic fluxes. Another idea to overcome the tradeoff was the enhancement of reverse β-oxidation forming 2-enoyl-CoAs from 3-oxoacyl-CoAs via ( S )-3HA-CoAs. The introduction of the second copies of had and crt2 into the pha operon in the ∆ phaB1 strain increased the 3HHx composition of the polyester fraction without a serious negative impact on PHA production. The combination of the enhanced reverse β-oxidation and weakened ( R )-specific reduction possibly increased butyryl-CoA formation from acetoacetyl-CoA and the following elongation to 3-oxohexanoyl-CoA, as well as the next reverse cycle to ( R )-3HHx-CoA (Fig. 5 ). As seen on vegetable oils [ 18 ], the copolyester biosynthesis was also affected by the substrate specificity of R -hydratase. The use of short-chain-length-specific PhaJ Ac tended to increase the C 4 units and decreas the C 6 units when compared with medium-chain-length-specific PhaJ4a. This was because crotonyl-CoA was the second node for redistribution of the flux to the C 4 - and C 6 -monomers. When PhaJ Ac was functional, crotonyl-CoA was partially intercepted to form ( R )-3HB-CoA and thus the decrease in the C 4 units caused by the lack of PhaB1 was compensated to some extent (Fig. 5 ). Wang et al. reported on the biosynthesis of P(3HB- co -3HHx) from glucose by recombinant E. coli strains harboring trans -2-enoyl-CoA reductase from Treponema denticola in BktB-dependent condensation pathway (14.2 wt%, 4.0 mol% 3HHx) or reverse β-oxidation pathway using FadBA from E. coli (12.4 wt%, 10.2 mol% 3HHx) [ 41 ]. In this study, the practically useful P(3HB- co -12.1 mol% 3HHx) could be produced with cellular content of 75.3 wt% from glucose by the engineered R. eutropha strain NSDG-GG-HC/pBPP-ccr Me J Ac -emd. Fig. 5 Proposed pathway for P(3HB- co -3HHx) biosynthesis from glucose by R. eutropha NSDG-GG-HC/pBPP-ccr Me J Ac -emd. PhaA and BktB, β-ketothiolases; PhaB1 and PhaB3, NADPH-acetoacetyl-CoA reductases; Had, NAD-( S )-3HB-CoA dehydrogenase; Crt2, crotonase; PhaC NSDG , N149S/D171G mutant of PHA synthase from A. caviae ; PhaJ Ac , short-chain-length-( R )-enoyl-CoA hydratase from A. caviae ; Ccr Me , crotonyl-CoA carboxylase/reductase from Methylorubrum extorquens ; Emd Mm , ethylmalonyl-CoA decarboxylase from Mus musculus \n The modifications of the acetoacetyl-CoA reduction step had similar effects on P(3HB- co -3HHx) biosynthesis from fructose to those from glucose (Fig. 3 and Additional file 1 : Table S2), although differences were observed when the PhaB1 mutants were introduced. On fructose, the recombinant strains having PhaB1 NV or PhaB1 NVYF accumulated more PHA with a larger 3HB fraction than the ∆ phaB1 strain, which was not seen on glucose. This result strongly suggested actual functions of the PhaB1 mutants in ( R )-3HB-CoA formation on fructose, despite the low affinity to acetoacetyl-CoA. This might be due to altered metabolic regulation depending on the sugar uptake machinery. In the R. eutropha H16-derived strains, fructose is incorporated and 6-phosphorylated by ATP-binding cassette transporter (FrcACB) [ 42 ] and fructokinase, respectively, while the uptake and 6-phosphorylation of glucose is mediated by mutated GlcNAc-specific PTS (NagFE) [ 28 ]. It is well known that glucose-specific PTS is associated with catabolite repression in E. coli . Likewise, the glucose-transportation by the mutant of NagFE in the R. eutropha strains may affect gene expression and consequent changes in metabolic regulation, such as a further decrease in intracellular pool of acetoacetyl-CoA, toward which catalytic functions of the low-affinity PhaB1 mutants were limited. The engineered strain NSDG-GG could grow and accumulate P(3HB) well on glycerol, whereas the deletion of phaB1 severely impaired PHA biosynthesis from glycerol, unlike from sugars. It has been demonstrated that PhaB1 and PhaB3 contributed to P(3HB) biosynthesis on fructose, while only PhaB1 had a role on palm oil [ 14 ]. This was probably due to weak expression of phaB3 on palm oil as shown by microarray analysis [ 35 ]. The present results strongly suggested that phaB3 expression was tightly repressed on not only vegetable oils but also another non-sugar substrate, glycerol. We further observed that introduction of the genes of PhaB1 mutants, PhaB2, or Had-Crt2 into the ∆ phaB1 strain only faintly restored the PHA biosynthesis. The highly efficient and highly expressed PhaB1 was essential to convert acetoacetyl-CoA to ( R )-3HB-CoA under the conditions on glycerol. A further engineering strategy to enhance the C 6 unit-formation pathway even with the functions of PhaB1 should be investigated to achieve the production of P(3HB- co -3HHx) from glycerol. Such strategy is also expected to be useful for more efficient production of PHA copolyesters and other compounds by phaB1 + -strains from various structurally unrelated carbon sources."
} | 4,045 |
33925903 | PMC8123457 | pmc | 3,018 | {
"abstract": "Arctic bacteria employ various mechanisms to survive harsh conditions, one of which is to accumulate carbon and energy inside the cell in the form of polyhydroxyalkanoate (PHA). Whole-genome sequencing of a new Arctic soil bacterium Pseudomonas sp. B14-6 revealed two PHA-production-related gene clusters containing four PHA synthase genes ( phaC ). Pseudomonas sp. B14-6 produced poly(6% 3-hydroxybutyrate- co -94% 3-hydroxyalkanoate) from various carbon sources, containing short-chain-length PHA (scl-PHA) and medium-chain-length PHA (mcl-PHA) composed of various monomers analyzed by GC-MS, such as 3-hydroxybutyrate, 3-hydroxyhexanoate, 3-hydroxyoctanoate, 3-hydroxydecanoate, 3-hydroxydodecenoic acid, 3-hydroxydodecanoic acid, and 3-hydroxytetradecanoic acid. By optimizing the PHA production media, we achieved 34.6% PHA content using 5% fructose, and 23.7% PHA content using 5% fructose syrup. Differential scanning calorimetry of the scl- co -mcl PHA determined a glass transition temperature (T g ) of 15.3 °C, melting temperature of 112.8 °C, crystallization temperature of 86.8 °C, and 3.82% crystallinity. In addition, gel permeation chromatography revealed a number average molecular weight of 3.6 × 10 4 , weight average molecular weight of 9.1 × 10 4 , and polydispersity index value of 2.5. Overall, the novel Pseudomonas sp. B14-6 produced a polymer with high medium-chain-length content, low T g , and low crystallinity, indicating its potential use in medical applications.",
"conclusion": "4. Conclusions PHA has attracted attention as an alternative material to replace petroleum plastics. The composition of monomers, especially medium-chain-length monomers such as 3HO and 3HD, influences the properties of the polymers, decreasing the rigidity and high crystallinity of PHB and increasing the toughness of the polymer. Pseudomonas species have been studied as possible mcl-PHA producers, including many mesophile Pseudomonas strains. We previously isolated the novel Arctic strain Pseudomonas sp. B14-6, which shared a similar gene structure with Pseudomonas sp. 61-3 and contained two different gene clusters for scl-PHB and mcl-PHA. Although the two species had similar gene clusters and alignments, the new strain produced a PHA with a different monomer ratio than previously reported PHAs. In particular, the strain produced 94% medium-chain-length monomer content, compared with less than 60% for Pseudomonas sp. 61-3. As a result, the scl- co -mcl PHA from Pseudomonas sp. B14-6 exhibited totally different T m , T g , and T c values from other reported PHAs. Furthermore, Pseudomonas sp. B14-6 was able to utilize a wide range of carbon sources, including glucose, fructose, galactose, glycerol, N-acetylglucosamine, and sucrose. The carbon source adaptability suggests the industrial potential of the strain, considering that the well-studied P. putida cannot utilize substrates such as sucrose to produce PHA without metabolic engineering. Finally, we optimized carbon and nitrogen sources, cultivation time, temperature, and inoculum size, and evaluated fructose syrup as a cost-effective feedstock for PHA production. In conclusion, Pseudomonas sp. B14-6 could be used for commercial PHA production with an economical starting material, resulting in a unique polymer that contains a very high proportion of mcl-PHA.",
"introduction": "1. Introduction Some microbes have developed various tactics to survive the harsh Arctic environment [ 1 , 2 , 3 ]. One of these survival mechanisms is to adapt to temperature and nutrition fluctuations by accumulating energy and carbon sources within the cell as polyhydroxyalkanoate (PHA) granules [ 4 , 5 ], which also function as chaperone-like molecules to protect internal cellular systems [ 6 , 7 , 8 ]. Although several microbes thrive at low temperatures and accumulate PHAs, the metabolism of Arctic bacteria is generally quite slow, weakening their potential as PHA producers [ 7 ]. PHAs are produced by bacteria using various renewable feedstocks, and are easily degraded under biological conditions [ 9 ]. Properties such as the monomer composition, distribution, and molecular weight of these biocompatible polyesters are controlled by different substrates or production hosts [ 10 , 11 , 12 ]. Depending on the desired property, production can be manipulated to yield different short-chain-length monomers such as butyrate and valerate, and medium-chain-length monomers such as hexanoate, octanoate, and chains of up to 14 carbon atoms in length [ 13 , 14 , 15 , 16 ]. Short-chain-length PHAs (scl-PHAs) such as polyhydroxybutyrate (PHB) have high rigidity but are brittle, while medium-chain-length PHAs (mcl-PHAs) are flexible and create a sticky surface, which provides them with thermoelastomeric properties suitable for biomedical applications such as in skin adhesives and drug delivery systems [ 13 , 17 , 18 ]. However, the production of mcl-PHAs has been relatively challenging due to difficulties in controlling the metabolic flux of the desired monomers [ 11 , 19 ]. Pseudomonas spp. are known for their ability to survive in wide-ranging temperatures and habitats, suggesting industrial potential thanks to their hydrolase activity and capacity for PHA and exopolysaccharide production [ 20 ]. Pseudomonas have been investigated as major PHA producers, usually producing mcl-PHA and rarely producing scl-PHA [ 21 , 22 ]. As the PHA polymer structure is affected by the substrates utilized during cultivation, the selection of the carbon source is important [ 23 ]. PHA production by Pseudomonas spp. using carbohydrate as a substrate usually leads to a low mcl-PHA yield, with a heterogeneous monomeric composition [ 24 ]. Selection of the proper substrate and optimization are needed in order to increase mcl-PHA yield and variety. Although Pseudomonas putida is the most studied PHA-producing strain, species including P. putida and its variants have difficulty using sucrose directly [ 25 ]. Because sugarcane-based feedstocks and waste fructose syrup are cheap and abundant sources for PHA production, several efforts have been undertaken to metabolically engineer strains to fully uptake sucrose, pretreat feedstock to degrade sucrose to glucose and fructose, or discover strains that are able to utilize sucrose to produce PHA [ 25 , 26 , 27 ]. In the current study, we attempted to produce a unique PHA with a heterogenous monomer composition from the novel Arctic microbe Pseudomonas sp. B14-6. Media composition and culture conditions such as temperature, inoculum size, and cultivation time were optimized to achieve high PHA production and content. Additionally, we cultured Pseudomonas sp. B14-6 under optimal conditions with fructose, a single monosaccharide, and fructose syrup, a cost-effective complex carbon source. Finally, we analyzed the physical properties of the produced PHA to hypothesize possible applications of the strain and the PHA.",
"discussion": "3. Results and Discussion 3.1. Screening of PHA-Related Genes in Pseudomonas sp. B14-6 We previously reported that whole-genome sequencing of Pseudomonas sp. B14-6 revealed a genome 6,776,772 base pairs in length. The strain showed dramatic membrane changes with temperature, and important players of membrane modification were identified [ 28 ]. Using the genome sequencing results, we identified two PHA-synthesis-related gene clusters and compared them with PHA-related genes from Ralstonia eutropha , Pseudomonas, and other PHA-producing species by BLAST ( Figure 1 ). The two gene clusters had the orders phaR-phaP1-phaC1-araC-phaB-phaA-phaC2 and phaC4-phaZ-phaC3-phaD-phaP2-phaP3 , which shared high similarity with Pseudomonas sp. 61-3, although the annotation might differ [ 35 ]. Amongst these two PHA synthesis-related gene clusters, we identified four PHA synthase genes, named phaC1 – C4 , from the open reading frame order assigned by whole-genome sequencing. As PHA synthases can be classified as types I to IV [ 36 ], we attempted to classify each phaC by the BLAST search results. One cluster had acetyl-CoA thiolase ( phaA) , acetoacetyl-CoA reductase ( phaB ), and two PHB synthases ( phaC1 and phacC2 ), sharing high similarity with the PHA synthesis cluster of Ralstonia eutropha (a PHA synthase type-I that produces scl-PHA) ( Figure 1 a). The order of the PHA synthesis gene in R. eutropha is phaC1-phaA-phaB1 ; however, the order in Pseudomonas sp. B14-6 was phaB-phaA-phaC2 , in the reverse order. We found that phaC1 was a PHA synthetase, not a PHA synthase, with a much longer sequence than other phaC genes; it contained a PHA synthase type-IV-related domain, although its function was unknown. Next to phaC1 was the phaR gene, known for the subunit required to produce PHA with phaC in the PHA synthase type-IV system. However, it was difficult to determine its specific function because phaC had a longer sequence than the PHA synthase type-IV of Bacillus megaterium, and was similar to that of Pseudomonas sp. 61-3 [ 36 ]. In addition, the two PhaC genes in R. eutropha are located very far apart, with one cluster having the order phaC1-phaA-phaB1-phaR-bktB (β-ketothiolase) and the other having the order phasin2-phaC2-phaB2 . The gene cluster of Pseudomonas sp. B14-6 was thus unique, with both phaC genes at neighbor locations. The other PHA gene cluster also had two phaC genes ( phaC3 and phaC4) and a PHA depolymerase ( phaZ ) resembling a PHA synthase type-II that appears in typical PHA-producing Pseudomonas spp. Compared with the PHA gene cluster of Pseudomonas spp., the length of the phaC3 gene was the same as that of Pseudomonas putida , and the phaD was slightly longer than that of P. putida and Pseudomonas aeruginosa , but shorter than the long phaD gene of Pseudomonas oleovorans ( Figure 1 b). As both Pseudomonas sp. 61-3 and Pseudomonas sp. B14-6 have similar gene clusters and two different types of PHA synthesis gene clusters for scl-PHAs ( phaC2) and mcl-PHAs ( phaC3 and phaC4 ), it was expected that Pseudomonas sp. B14-6 would produce similar PHAs as Pseudomonas sp. 61-3 or Pseudomonas sp. MPC6. 3.2. Analysis of PHA from Pseudomonas sp. B14-6 According to the gene search results, we focused on two points: (1) Pseudomonas sp. B14-6 could produce PHA, and (2) PHA produced by the strain might contain monomeric units, both medium and short chain length. As the growth temperature range of Pseudomonas sp. B14-6 is between 4 °C and 30 °C, we first measured growth and PHA accumulation at different temperatures in standard PHA production media. Several studies have reported that low temperatures could stimulate higher PHA content in single cells, but relatively low biomass. However, when Pseudomonas sp. B14-6 was cultured at 15 °C, 25 °C, and 30 °C, dry cell weight (DCW) was highest at 25 °C and PHA content was highest at 30 °C ( Figure 2 a). Although Pseudomonas sp. B14-6 was isolated from Arctic soil and could grow at low temperatures, a moderate temperature of 30 °C was optimal for PHA production. The produced PHA was analyzed by GC-MS, revealing seven different monomeric units containing 3HB as a short-chain monomeric unit; 3HHx, 3HO, 3HD, 3HdD, and 3HtD as medium-chain monomeric units; and 3-hydroxy-5- cis -dodecenoic acid (HdDe) as an unsaturated monomeric unit ( Figure 2 b). The ratio of monomers in poly(6.0% 3HB- co -8.5% 3HHx- co -39.8% 3HO- co -33.6% 3HD- co -5.6% 3HdDe- co -6.5% 3HdD- co -0.1% 3HtD) differed from that in Pseudomonas sp. 61-3 poly(44% 3HB- co -5% 3HHx- co -21% 3HO- co -25% 3HD- co -3% 3HdD- co -3% 3HdDe) and Pseudomonas sp. MPC6 poly(89.5% 3HB- co -1.8% 3HHx- co -3.3% 3HO- co -4.4% 3HD- co -1.1% 3HdD) [ 37 , 38 ]. Pseudomonas sp. B14-6 produced a very high proportion of medium-chain-length monomers (96% of PHA content), compared with 56% and 89.5% from Pseudomonas sp. 61-3 and Pseudomonas sp. MPC6, respectively [ 39 , 40 ]. It was expected that the physical properties of the PHA produced by Pseudomonas sp. B14-6 would differ from those of previously reported PHAs. 3.3. Optimization of mcl-PHA Production To determine the optimal nutrients for mcl-PHA production, we screened various carbon sources, including glucose, fructose, xylose, N-acetyl glucosamine, galactose, lactose, and sucrose, and compared monomeric composition from the carbon source ( Table 1 ). Among them, we focused on fructose as the main carbon source ( Figure 3 a). Additionally, we screened yeast extract, malt extract, peptone, and tryptone as complex nitrogen sources, as well as ammonium sulfate and ammonium chloride as defined nitrogen sources. Among these, yeast extract yielded the highest PHA content ( Figure 3 b). Further optimization was conducted using fructose as the carbon source and yeast extract as the nitrogen source. We determined that 5% fructose yielded the highest PHA content of 36% ( Figure 3 c), while optimal biomass and PHA content were achieved with 0.1% yeast extract ( Figure 3 d). Another parameter that affects PHA production is the amount of microbes inoculated in the media. Thus, we evaluated PHA production using inoculum at 0.5% to 4% of the final volume, determining 0.5% as the optimal inoculum size ( Figure 3 e). PHA production is also affected by cultivation time, in which gene expression and polymer building is not completed during a short cultivation, and accumulated PHA is degraded as an energy source due to exhaustion of the initial carbon source during a long cultivation. Therefore, cultivation time for PHA production was evaluated for 168 h, revealing that culture for 120 h yielded optimal biomass and PHA content, but volumetric productivity was the highest at 48 h ( Figure 3 f). The composition of PHA was changed by different temperatures (data not shown). As a result, the optimal conditions for PHA production with Pseudomonas sp. B14-6 were set as 5% fructose, 0.1% yeast extract, 0.5% inoculum size, and 120 h culture time at 30 °C. PHA production in optimal conditions was higher than other psychrophilic Pseudomonas spp. ( Table 2 ). 3.4. Monitoring Time-Dependent PHA Production and Fructose Syrup Application The time-dependent production of PHA by Pseudomonas sp. B14-6 was investigated under optimized conditions, measuring DCW, amount of PHA produced, and carbon source consumption over 120 h. Almost no PHA was produced during the first 24 h, but both PHA content and DCW increased from 48 h until 72 h. DCWs at 72 h and 120 h were similar and volumetric productivities were slightly decreased, but the highest PHA content was obtained at 96 h ( Figure 4 a). The consumption of fructose increased after 48 h, and the pH began decreasing from approximately 7 to 6.69 at 120 h ( Figure 4 b). We then evaluated the use of fructose syrup composed of 20% glucose, 36% fructose, and 44% sucrose as an economical carbon source. The production of PHA with media that contained 5% fructose syrup achieved 23.7% PHA content ( Figure 4 c). Use of fructose syrup induced a different pH pattern, optimal culture time, and lower PHA content than using fructose ( Figure 4 d). No catabolite repression occurred during cultivation, as all sugars were consumed together. 3.5. Physical Properties of Produced PHA The thermal properties of poly(5.0% 3HB- co -7.8% 3HHx- co -25.6% 3HO- co -40.0% 3HD- co -9.9% 3HdDe- co -11.0% 3HdD- co -0.7% 3HtD) were investigated by DSC, including the glass transition temperature (T g ), melting temperature (T m ), crystallization temperature (T c ), and melting enthalpy (ΔH m ). For the scl- co -mcl PHA, T g = 15.3 °C, T c = 86.8 °C with an enthalpy = 38.6 J/g, and T m = 112.8 °C with ΔH m = 41.2 J/g were obtained ( Figure 5 a). The thermal characteristics of our scl- co -mcl PHA differed both in T m and T c from those of the previously reported mcl-PHA from Pseudomonas sp. PAMC28620 composed of poly(25.5% 3HO- co -52.1% 3HD- co - 5.7% 3HdD- co -16.7% 3HtD) with T m = 172.8 °C, T g = 3.99 °C, and T c = 54.61 °C [ 30 ]. The thermal properties also differed from those of the mcl-PHA from Pseudomonas sp. MPC6 composed of poly(89.5% 3HB- co -1.8% 3HHx- co -3.3% 3HO- co -4.4% 3HD- co -1.1% 3HdD) with T m = 163.5 °C, T g = 2.3 °C, and T c = 46.0 °C [ 37 ]. The scl- co -mcl PHA from Pseudomonas sp. B14-6 had lower T m values and higher T g and T c values due to the low 3HB content and high 3HO and 3HD content, which lowered the melting point. As an important polymer parameter, the degree of crystallinity (X c ) was calculated from the enthalpy, revealing 3.82% crystallinity. This X c value was extremely low compared to that reported for mcl-PHAs from Pseudomonas sp. PAMC28620 (X c = 43.7%) [ 30 ], Bacillus thermoamylovorans PHA005 (X c = 43.0%) [ 44 ], and R. eutropha (X c = 91%) [ 45 ]. As highly crystalline polymers have limited application in industrial and medical fields [ 46 ], the scl- co -mcl PHA from Pseudomonas sp. B14-6 may have potential in applications that require a sticky and relatively low-temperature modeling polymer, as well as biodegradable characteristics. The molecular weight of the obtained PHA was characterized by GPC, demonstrating a retention time peak start at 13.02 min, peak maximum at 16.62 min, and peak end at 20.32 min ( Figure 5 b). The scl- co -mcl PHA had average values, with a number average molecular weight (M n ) of 3.6 × 10 4 , weight average molecular weight (M w ) of 9.1 × 10 4 , Z-average (M z ) of 1.9 × 10 5 , and viscosity average molar mass (M v ) of 8.0 × 10 4 . The polydispersity index value of scl- co -mcl PHA was 2.5, which differed from other values of scl-PHA or mcl-PHA, due to the mixed monomeric unit composition ( Table 3 ). The high polydispersity index value was due to the various monomer unit composition and breakdown of the polymer in the sample preparation steps. The physical properties of the scl- co -mcl PHA produced by Pseudomonas sp. B14-6 differed from PHAs produced by other species. Therefore, we hypothesize that the scl- co -mcl PHA has potential applications in biomedical or similar industrial fields."
} | 4,547 |
37108429 | PMC10138535 | pmc | 3,019 | {
"abstract": "Bacterial adaptation is regulated at the population level with the involvement of intercellular communication (quorum sensing). When the population density is insufficient for adaptation under starvation, bacteria can adjust it to a quorum level through cell divisions at the expense of endogenous resources. This phenomenon has been described for the phytopathogenic bacterium Pectobacterium atrosepticum ( Pba ), and it is called, in our study, adaptive proliferation. An important attribute of adaptive proliferation is its timely termination, which is necessary to prevent the waste of endogenous resources when the required level of population density is achieved. However, metabolites that provide the termination of adaptive proliferation remained unidentified. We tested the hypothesis of whether quorum sensing-related autoinducers prime the termination of adaptive proliferation and assessed whether adaptive proliferation is a common phenomenon in the bacterial world. We showed that both known Pba quorum sensing-related autoinducers act synergistically and mutually compensatory to provide the timely termination of adaptive proliferation and formation of cross-protection. We also demonstrated that adaptive proliferation is implemented by bacteria of many genera and that bacteria with similar quorum sensing-related autoinducers have similar signaling backgrounds that prime the termination of adaptive proliferation, enabling the collaborative regulation of this adaptive program in multispecies communities.",
"conclusion": "5. Conclusions Both known quorum sensing mediators of Pba (AHL and AI-2) are components of the signaling background that “notify” Pba that it is time to terminate adaptive proliferation and proceed to the persistence-related stage of adaptation. AI-2 seems to have a greater contribution to the termination of adaptive proliferation than AHL. However, neither AI-2 nor AHL are strictly required to terminate adaptive proliferation, and these two signaling molecules can compensate for each other’s actions (at least partially) during the termination of adaptive proliferation, possibly along with some other as-yet unidentified signaling molecules. In turn, both AHL and AI-2 are required for the formation of cross-resistance in the course of adaptive proliferation and for retaining full virulence following adaptive proliferation; herewith, the AHL contributes more to cross-resistance than AI-2. Adaptive proliferation is a phenomenon common to a wide range of bacterial species that use different molecules as quorum sensing signals. The signaling background that primes the termination of adaptive proliferation in one AHL-producing species can be perceived by and terminate adaptive proliferation in (at least partially) other AHL-producing species, but not in AHL-non-producing species. This indicates that adaptive proliferation can be regulated at the interspecies level in complex bacterial communities. Bacterial cells “prepare” themselves for perceiving the signaling background that primes the termination of adaptive proliferation; the sensitivity to this signaling background is absent (or low) in exponentially growing cells, but it gradually develops during the course of adaptive proliferation.",
"introduction": "1. Introduction Adaptive programs in bacteria are regulated at the population level, involving intercellular signaling [ 1 , 2 , 3 , 4 , 5 ]. Without intercellular communication, bacteria cannot properly adapt following exposure to a stressor. However, by the example of the phytopathogenic bacterium Pectobacterium atrosepticum ( Pba ), it has been shown that, if the cell concentration is below the one that permits intercellular communication, bacteria are able to increase the population density up to a quorum level, even in the absence of an exogenous growth substrate [ 6 ]. Such a phenomenon of an increase in population density under starvation conditions takes place only at a low initial cell titer of 10 1 –10 5 CFU/mL until it reaches the value of ~10 6 CFU/mL. In our study, we refer to this phenomenon as adaptive proliferation. Following adaptive proliferation, Pba cells retain high virulence and develop cross-protection, i.e., become resistant to a number of stressors [ 7 ]. Importantly, in the absence of exogenous organic carbon, irrespective of whether the initial cell titer is 10 1 or 10 5 CFU/mL, the adaptive proliferation yields a population density of ~10 6 CFU/mL. This means that the adaptive proliferation is terminated not after a particular number of cell divisions, but after the population density reaches a particular level that presumably provides intercellular communication. This is supported by the facts that (1) the termination of the adaptive proliferation is coupled with the accumulation of autoinducers of one of the quorum sensing systems, acyl homoserine lactones (AHL), as well as with the upregulation of AHL-synthase gene, and (2) adaptive proliferation is terminated prematurely (at a level below 10 6 CFU/mL) in the presence of metabolites accumulated in the starving cultures that had already passed through the process of adaptive proliferation [ 6 ]. Apparently, adaptive proliferation is carried out at the expense of endogenous cell resources, which is evident from cell morphology: cells after adaptive proliferation have reduced sizes and a dramatically reduced volume of cytoplasm [ 6 ]. Therefore, to prevent the waste of resources, it is important for bacterial cells to timely terminate adaptive proliferation when it is no longer needed, i.e., when the population density reaches a level enabling the intercellular communication necessary for an effective adaptation. It is obvious that cells “feel” this level via extracellular signaling molecules that remain unidentified to date. Quorum sensing autoinducers are the most explicit candidate signal molecules to indicate that population density is sufficient for adaptation and that it is reasonable to terminate adaptive proliferation. In Pba , two quorum sensing systems have been described to date: (1) mediated by 6-oxo- and 8-oxo-AHL—the products of the expI gene and (2) mediated by autoinducer of the second type, AI-2, the product of the luxS gene [ 8 ]. AHL, which is used as quorum sensing signal, not only by Pectobacterium species, but also by species of some other genera, are widely shown to control virulence, motility, biofilm formation, and oxidative stress resistance in Pectobacterium species [ 9 , 10 , 11 ]. In turn, the role of AI-2, which is a “universal” inter-species signal produced by most, if not all, bacteria, in Pectobacterium physiological responses, is much less understood [ 11 , 12 , 13 ]. The aim of our study was to determine whether AHL and AI-2 are involved in the termination of adaptive proliferation in Pba . In addition, we sought to check whether, other than Pba bacterial species, both AHL-producing and AHL-non-producing were able to implement adaptive proliferation, and if so, whether the signaling background that primes the termination of adaptive proliferation is genus-specific or universal for different bacterial genera, enabling the collaborative regulation of this adaptive program in multispecies communities.",
"discussion": "3. Discussion In the present study, we aimed to understand for what reasons bacterial cell division under starvation at low initial population densities (adaptive proliferation) is terminated when the population density reaches a value of 10 6 CFU/mL. We put forward the hypothesis that it is the accumulation of the quorum-related autoinducers that primes the termination of adaptive proliferation. We first examined whether exogenous AHL would repress the adaptive proliferation. We showed that exogenous 6-oxo-AHL caused premature termination of adaptive proliferation, whereas 8-oxo-AHL enhanced the effect of 6-oxo-AHL. However, this effect of AHL was manifested only at concentrations higher than those accumulated in Pba cultures. Then, we monitored adaptive proliferation in two quorum-related Pba mutants, one of which with knocked out gene encoding AHL-synthase (Δ expI ) was AHL-deficient, while the other one with knocked out gene encoding AI-2-synthase (Δ luxS ) was AI-2-deficient. Unexpectedly, in both mutant strains, the adaptive proliferation proceeded similarly to the wild type strain, including that it was terminated at a population density of 10 6 CFU/mL. However, in mutant strains, in contrast to the wild type strain, we observed a rather sharp decrease in CFU titers following adaptive proliferation. We proposed that both AHL- and AI-2-mediated quorum sensing systems were required to gain stress resistance following adaptive proliferation. Indeed, following adaptive proliferation, the level of cross-protection and the level of virulence were reduced in both mutants compared with the wild type. Based on the monitoring of adaptive proliferation in the Δ expI and Δ luxS mutants, AHL or AI-2 were required for obtaining the increased stress resistance, but they were unlikely to be involved in priming the termination of adaptive proliferation. Then, we assumed that if these two autoinducers do not indeed participate in the target physiological process, then the supernatants of cultures of mutant strains that passed through adaptive proliferation would cause the premature termination of adaptive proliferation, similar to the supernatants of the wild type. However, compared to the supernatants of the wild type cultures, the supernatants from mutant strains had a reduced ability to prime the premature termination of adaptive proliferation, especially the supernatants of the Δ luxS mutant. This fact clearly shows that both AHL and AI-2 (especially the latter) are involved in priming the termination of adaptive proliferation. However, at the same time, it remained unclear why the termination of adaptive proliferation in AHL- and AI-2-deficient strains occurred in a manner similar to the wild type strain. We proposed that the loss of one of the two studied autoinducers (AHL or AI-2) can be compensated for by the enhanced activity of the other or by additional yet unidentified intercellular signaling molecules, which would lead to the termination of adaptive proliferation in AHL- and AI-2-deficient strains. To check this, we determined the relative levels of AHL and AI-2 in the wild type and mutant strains following adaptive proliferation. Indeed, in the supernatants of starving cultures of the Δ expI mutant, we observed an increased level of AI-2 compared to the wild type, while in those of the Δ luxS mutant, an increased level of AHL accumulated. So, if a strain is unable to produce one quorum-related autoinducer during adaptive proliferation, it produces a greater amount of the other one, which presumably contributes to the termination of adaptive proliferation, even in the absence of one of the regulators of this process. This is in accordance with our results showing that only increased concentrations of exogenous AHL could trigger the premature termination of adaptive proliferation in Pba . Thus, the termination of adaptive proliferation is triggered by the combinatory and mutually compensatory actions of two quorum sensing systems and maybe other additional metabolites involved in intercellular communication that remain to be identified. Herewith, the interaction of different regulatory systems during the termination of adaptive proliferation is likely to be organized in such a way as to ensure the successful realization of this physiological process, even in the case of a failure of one of the regulatory systems involved in its regulation. Quorum sensing autoinducer signal integration expressed in the cooperative action of two quorum sensing systems in terms of the activation of various phenotypes has been previously reported for Pseudomonas aeruginosa and the bioluminescent bacterium Vibrio harveyi [ 14 , 15 , 16 ]. In V. harveyi , different quorum-related phenotypes were manifested at different ratios of AI-2 and AHL concentrations. Herewith, the manifestation of some phenotypes required the presence of both autoinducers, while other phenotypes could be induced by only one of the autoinducers, and the second one only enhanced the effect of the first autoinducer [ 16 ]. Our results show that AI-2 and AHL seem to enhance each other’s actions in terms of priming the termination of adaptive proliferation, but neither of these two autoinducers is strictly required for such priming. The results of the conducted experiments also indicate that the specific extracellular signaling background is not the sole element required for the termination of adaptive proliferation. Additionally, during adaptive proliferation, the sensitivity (competence) of cells to this extracellular signaling background must be increased. We were able to cause the premature termination of adaptive proliferation (at CFU levels below 10 6 CFU/mL) using various treatments, but we were never able to completely repress adaptive proliferation, even when cells were transferred to supernatants of wild type cultures that had passed through adaptive proliferation, despite the fact that these supernatants contained the entire extracellular signaling background for the termination of adaptive proliferation. In our opinion, de novo inoculated cells are unable to perceive this signaling background and, due to this, proceed to the division. However, after several rounds of division, cells acquire the ability to recognize these signals, and adaptive proliferation is terminated prematurely. This scenario implies that bacterial cell sensitivity to quorum sensing signals is not constant and can change depending on the physiological status. Indeed, it has been previously shown that bacteria ( V. harveyi ) can adjust the sensitivity to quorum sensing signals [ 17 , 18 ]. In turn, such an adjustment of the sensitivity to autoinducers is likely to play an important role in the termination of adaptive proliferation. In our study, we also found that the phenomenon of adaptive proliferation is not unique to Pba and assessed whether the signaling background that primes the termination of adaptive proliferation is genus-specific or common in different genera. All six analyzed bacterial species were able to implement adaptive proliferation during starvation at a low population density. The termination of adaptive proliferation in different species occurred at 1–6 × 10 6 CFU/mL. The supernatants of starving Pba cultures could cause premature termination of adaptive proliferation in other AHL-producing species, but not in AHL-non-producing species. This fact additionally supports the role of AHL in the termination of adaptive proliferation in AHL-producing species. However, within AHL-producing species, the composition of the signaling background that primes the termination of adaptive proliferation seems to have some genus-specific features, since supernatants of Pba had a greater inhibitory effect on the adaptive proliferation of P. ananatis than of D. solani . The reason for this may be that the Dickeya species, in addition to producing AHL, synthesize another autoinducer, vfm (virulence factor modulating), as the main quorum sensing signal [ 19 , 20 ]. It has been previously shown that bacteria within complex communities perceive not only self-produced quorum-related molecules, but also those that are produced by other species [ 21 , 22 , 23 ]. This is quite expected for AI-2, which is synthesized by most bacterial species. Herewith, AHL also widely participated in interspecies communication [ 24 , 25 , 26 , 27 ]. Due to this, the bacteria of a particular species can sense not only their own population density, but also the density of the whole bacterial community (or its significant part) and exploit autoinducers produced by other members of the community. For example, Pectobacterium wassabiae (a phytopathogenic bacterium closely related to Pba ) can sense AHL synthesized by other bacteria in the multispecies community of potato tubers, and an integral AHL pool produced by the whole community determines the level of virulence of this species [ 26 ]. Our results show that the composition of an integral pool of quorum-related molecules determined by the composition of the bacterial community can influence not only virulence, but also adaptive reactions in bacteria. Herewith, bacteria of different species can use quorum-related molecules of each other to initiate the adaptive program. In particular, adaptive proliferation is likely to be regulated at an interspecies level in multispecies communities. Thus, our study shows that adaptive proliferation is a widespread phenomenon in the bacterial world. The existence of this phenomenon indicates that bacterial cells contain a significant amount of resources and energy that can be spent to provide multiple cell divisions in the absence of a growth substrate in order to successfully adapt to unfavorable conditions. The consumption of these resources is tightly regulated via intercellular communication to avoid their waste when the quorum level of population density required for adaptation is achieved. Our study provides the first information about the components of the extracellular signaling background that prime the termination of adaptive proliferation. These components are AHL and AI-2, which act synergistically and mutually compensatory (possibly with other as-yet unidentified metabolites) to provide timely termination of adaptive proliferation."
} | 4,403 |
25926899 | PMC4406998 | pmc | 3,021 | {
"abstract": "Rhizosphere, the interface between soil and plant roots, is a chemically complex environment which supports the development and growth of diverse microbial communities. The composition of the rhizosphere microbiome is dynamic and controlled by multiple biotic and abiotic factors that include environmental parameters, physiochemical properties of the soil, biological activities of the plants and chemical signals from the plants and bacteria which inhabit the soil adherent to root-system. Recent advancement in molecular and microbiological techniques has unravelled the interactions among rhizosphere residents at different levels. In this review, we elaborate on various factors that determine plant-microbe and microbe-microbe interactions in the rhizosphere, with an emphasis on the impact of host genotype and developmental stages which together play pivotal role in shaping the nature and diversity of root exudations. We also discuss about the coherent functional groups of microorganisms that colonize rhizosphere and enhance plant growth and development by several direct and indirect mechanisms. Insights into the underlying structural principles of indigenous microbial population and the key determinants governing rhizosphere ecology will provide directions for developing techniques for profitable applicability of beneficial microorganisms in sustainable agriculture and nature restoration.",
"conclusion": "CONCLUDING REMARKS This minireview aims to consolidate our current understanding of chemistry and biology of rhizosphere with regard to both host plant and the residing microbial community with an emphasis to appraise the role of root exudates in shaping rhizobiota. It also highlights the beneficial interactions between plants and different PGPM and BCA. This information, collectively, may serve as foundation for further research towards development of novel methods in sustainable agricultural practice. Despite the historical commitment and attention to the soil microbiome, its study has not kept pace with the recent surge of research activity as that of human microbiome. Plant microbiome is often termed as the second genome of plant and it is to be remembered that plant-microbe interaction is a “dialogue” rather than a unilateral relationship. Further investigation unraveling the ‘specific’ interactions in the rhizosphere will benefit the fundamental understanding of plant biology and provide stability of food production.",
"introduction": "INTRODUCTION Rhizosphere, the soil adjacent to plant-roots, is a unique niche for microbial colonization. The term, rhizosphere, was first coined by Lorenz Hiltner [ 1 ]. It is a complex chemical matrix replete with diverse microbial species. Rhizosphere microbial community is recruited from the surrounding soil which acts as a microbial seed bank, while the plants determine which members of this bulk soil reservoir of microorganisms will flourish and thrive in the rhizosphere [ 2 ]. The present review is organized mainly into three sections. In the first section, we define rhizosphere effect and discuss the role of different factors such as plant genotype, plant age and environmental interferences that affect rhizodeposition processes. In the second section, we elaborate how the chemistry at the root-soil interface influences the microbial community with a highlight on plant-microbe and microbe-microbe interactions. Finally, few studies have been cited which exploit the mechanistic and ecological knowledge of rhizosphere into the programs linked to bio-based agriculture and economy."
} | 886 |
39266507 | PMC11393082 | pmc | 3,025 | {
"abstract": "Resistive memory devices feature drastic conductance change and fast switching dynamics. Particularly, nonvolatile bipolar switching events (set and reset) can be regarded as a unique nonlinear activation function characteristic of a hysteretic loop. Upon simultaneous activation of multiple rows in a crosspoint array, state change of one device may contribute to the conditional switching of others, suggesting an interactive network existing in the circuit. Here, we prove that a passive resistive switching circuit is essentially an attractor network, where the binary memory devices are artificial neurons while the pairwise voltage differences define an anti-symmetric weight matrix. An energy function is successfully constructed for this network, showing that every switching in the circuit would decrease the energy. Due to the nonvolatile hysteretic function, the energy change for bit flip in this network is thresholded, which is different from the classic Hopfield network. It allows more stable states stored in the circuit, thus representing a highly compact and efficient solution for associative memory. Network dynamics (towards stable states) and their modulations by external voltages have been demonstrated in experiment by 3-neuron and 4-neuron circuits.",
"introduction": "Introduction Resistive memory devices possess rich dynamics, including electrical, thermal, and structural effects 1 – 5 . By exploiting the switching dynamics or simply the programmable resistance attribute, resistive memory enables highly efficient computations, such as stateful logic 6 , temporal information processing 7 , and analog matrix computing 8 . Thanks to the nonvolatile switching and the crosspoint architecture, the resistive memory-based computations are promising to alleviate the von Neumann bottleneck and to provide massive parallelism 9 . For instance, crosspoint resistive memory arrays have been frequently employed to accelerate the matrix-vector multiplication (MVM) in many algorithms, among which neural networks have drawn the most attention 10 , 11 . The attractor network is a kind of recurrent neural network (RNN) model working with interactive feedback, whereas the Hopfield network is a typical representative 12 . It has a strong connection with biological neural circuits and memory mechanisms and finds applications in associative memory and combinational optimization 13 – 15 . In the classic Hopfield network, a set of McCulloch-Pitts (MP) neurons are interconnected through a symmetric weight matrix. During the computing process, the neuronal states are updated (usually asynchronously) by performing a sequence of MVM operations and pointwise nonlinear functions 12 . Recently, there has been a series of works on the hardware implementation of the Hopfield network with crosspoint resistive memory arrays to accelerate the MVM operations 16 – 19 . In these works, the memory devices are used as static, programmable analog resistors to emulate highly simplified synapses. On the other hand, the neurons are usually simulated by traditional amplifier circuits, which, however, are considered inefficient in terms of area, time, and energy costs. In addition, as only the MVM operation is performed in this architecture, the algorithmic iterations are discretely conducted, which causes extra latency. In particular, the discrete iteration of analog computing requires analog-digital conversion interfaces, which incur a stubborn bottleneck that limits the computing efficiency. Observe that a resistive memory device is not merely a programmable resistor, rather it has its own dynamics, which appear as thresholded, recyclable and nonvolatile switching behaviors 20 . The relationship between a column of resistive memory devices is described by the Kirchhoff’s current law (KCL), which contributes naturally a dot product operation. As a result, it is possible to sketch out an attractor network model, considering that: (1) resistive memory devices show a large ratio (typically 10−100) and fast speed (nanoseconds) of conductance change, thus the resistive switching (set and reset) can be viewed as a nonlinear activation 21 , (2) when applying voltages simultaneously to a column of resistive memory devices, switching of one device is conditional on the linear combination of other devices’ states described by KCL in the circuit, (3) the switching of one device changes the overall distribution of potentials in the circuit, which plays feedback to trigger further switching events, updating the state vector of the devices. This approach is an emergent result of a simple circuit consisting of passive resistive switching devices, which is in sharp contrast to the hardware mapping of the Hopfield network with resistive devices and active amplifiers. As such, the resistive memory circuit could be an intrinsic attractor network where the device states (upon application of a set of voltages) evolve from an initial vector and eventually stabilize at some states after a sequence of state changes, which are attractors of the circuit. In this network, the weight parameters and, hence the state stability are determined and modulated by the applied voltages. Given the high compactness of resistive memory devices and the absence of active components, this approach should deliver high efficiency of attractor networks for relevant applications. In addition, because of the unique neuronal activation, the resistive memory network is prone to store more attractors than the Hopfield network, which is very beneficial to improving the storage capacity in associative memory.",
"discussion": "Discussion The emergent attractor network works with the physical computation of the dot product by KCL, where the two operands are conductance and voltage, as well as the hysteretic set/reset transitions. While the applied voltages may be provided by precise digital-to-analog converters and device parameters \\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}$${G}_{{LCS}}$$\\end{document} G L C S , \\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}$${G}_{{HCS}}$$\\end{document} G H C S , \\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}$${V}_{{set}}$$\\end{document} V s e t 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}$${V}_{{re}{set}}$$\\end{document} V r e s e t are inherently stochastic. To investigate the impact of variations of these parameters on the network characteristics (including the number of stable states and transition paths), we have performed sufficient device measurements to collect their statistics. By characterizing 14000 cycles of direct current (DC) I - V sweeps, 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}$${V}_{{set}}$$\\end{document} V s e t 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}$${V}_{{re}{set}}$$\\end{document} V r e s e t of each cycle are extracted, and their statistics are shown in SI Appendix , Supplementary Fig. S8 . Note that the test parameters in the DC mode have been explored so that the mean values 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}$${V}_{{set}}$$\\end{document} V s e t 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}$${V}_{{re}{set}}$$\\end{document} V r e s e t are close to the results in the pulse mode. Both \\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}$${V}_{{set}}$$\\end{document} V s e t 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}$${V}_{{re}{set}}$$\\end{document} V r e s e t show normal distributions with their own mean values and standard deviations, namely \\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}$${V}_{{set}}=$$\\end{document} V s e t = \n \\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}$$N$$\\end{document} N (0.63, 0.14) V 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}$${V}_{{re}{set}}=$$\\end{document} V r e s e t = \n \\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}$$N$$\\end{document} N (− 1.12, 0.21) V. As shown in the SI Appendix , Supplementary Fig. S9 , the statistics 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}$${G}_{{LCS}}$$\\end{document} G L C S 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}$${G}_{{HCS}}$$\\end{document} G H C S also follow normal distributions, subject 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}$${G}_{{LCS}}=$$\\end{document} G L C S = \n \\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}$$N$$\\end{document} N (65.5, 18.6) μS 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}$${G}_{{HCS}}=$$\\end{document} G H C S = \n \\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}$$N$$\\end{document} N (1082.9, 54.7) μS. Based on these results, we have carried out comprehensive evaluations of their impacts on the 4-neuron network applied with voltages \\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}$${{{{\\boldsymbol{v}}}}}_{4}$$\\end{document} v 4 . In Eq. 4 , the network energy is related to both threshold voltages ( \\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}$${V}_{{set}}$$\\end{document} V s e t 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}$${V}_{{re}{set}}$$\\end{document} V r e s e t ) 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}$${V}_{{BL}}$$\\end{document} V B L , the latter of which, in turn, is determined by the conductance states of neuron devices. Given that \\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}$${V}_{{set}}$$\\end{document} V s e t , \\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}$${V}_{{re}{set}}$$\\end{document} V r e s e t , \\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}$${G}_{{LCS}}$$\\end{document} G L C S 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}$${G}_{{HCS}}$$\\end{document} G H C S follow normal distributions, the network energy becomes rather a distribution function instead of a constant value. For the 4-neuron network applied with \\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}$${{{{\\boldsymbol{v}}}}}_{4}$$\\end{document} v 4 , the energies calculated for each state vector are summarized in the SI Appendix , Supplementary Fig. S10 . Due to the variations of energies, all the transitions between state vectors become probabilistic, and the limitation on the transitions between certain states may be broken ( SI Appendix , Supplementary Fig. S10 ). Fig. 4b shows the probabilities of the final states for a given initial state after considering the energy distributions, where the vertical/horizontal axis dictates the initial/final state, and the corresponding experimental case is marked with a red box. Such simulations have been extended to larger networks, including 6-neuron and 8-neuron networks, and the results of the energy function and state transition probability ( SI Appendix , Supplementary Figs. S11 , S12 ) show similar trends as the 4-neuron network. In such realistic cases, the emergent attractor network turns out to be similar to the Boltzmann machine concept 33 , which is a probabilistic version of the classic Hopfield network. In particular, the switching with probabilistic thresholds ( \\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}$${V}_{{set}}$$\\end{document} V s e t 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}$${V}_{{re}{set}}$$\\end{document} V r e s e t ) depicts very closely the behavior of the Boltzmann machine. Specifically, the transitions of unstable states become more diverse with a distribution of probability. In the deterministic situation, state ‘0100’ will not transit to ‘1100’ due to the requirement of energy change threshold (even though state ‘1100’ has a lower energy). In the probabilistic situation, however, there is a probability of 36% that this transition may happen. Some deterministic attractor states also may cascade to a lower-energy state, such as the transitions from ‘0001’ or ‘0010’ to ‘0011’ with a probability of 10%, and those from ‘1001’ or ‘1010’ to ‘1011’ with a probability of 9%. Only the two attractors with the lowest energy (‘0011’ and ‘1011’) are deterministically stable. Such behaviors may be utilized to explore the application of resistive memory networks to probabilistic computing tasks. Because of the disturbance of probabilistic switching that breaks out of the constraint on state transition, the number of attractors is also affected. Supplementary Figs. S13 and S14 in the SI Appendix show the impact of each factor including \\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}$${V}_{{set}}$$\\end{document} V s e t , \\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}$${V}_{{re}{set}}$$\\end{document} V r e s e t , \\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}$${G}_{{HCS}}$$\\end{document} G H C S 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}$${G}_{{LCS}}$$\\end{document} G L C S analyzed one by one, for the given voltages \\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}$${{{{\\boldsymbol{v}}}}}_{4}$$\\end{document} v 4 . The first observation is that the impact of threshold voltage on the results is substantially greater than that of conductance. In particular, 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}$${V}_{{set}}$$\\end{document} V s e t distribution has the highest impact, which should be ascribed to the fact that, in this specific case, states with more bits of ‘0’ are prone to be unstable and thus experience more set transitions. Meanwhile, the impact 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}$${G}_{{LCS}}$$\\end{document} G L C S distribution is almost ignorable, due to its small conductance value involved in the computation. By including distributions of all parameters in the simulation, the resulting number of attractors also presents a normal distribution ( SI Appendix , Supplementary Fig. S15 ). We note the load resistor with conductance \\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}$${G}_{0}$$\\end{document} G 0 may also have an effect on the network behaviors ( SI Appendix , Supplementary Fig. S16 ), although it should be much more controllable than the stochastic resistive memory devices. In the model, the LCS and HCS of resistive memory have been taken as logical 0 and 1 to participate in the deduction of equations, which should be an idealized situation. In practice, the conductance of LCS is absolutely not 0, and a finite ratio \\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}$$\\varDelta$$\\end{document} Δ = \\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}$${G}_{{LCS}}$$\\end{document} G L C S / \\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}$${G}_{{HCS}}$$\\end{document} G H C S of should be considered. We have also investigated the impact of this factor on the network behaviors, through both numerical simulation ( SI Appendix , Supplementary Fig. S17 ) and theoretical analysis ( SI Appendix , Supplementary Text S3 ). The results show that the ratio \\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}$$\\varDelta$$\\end{document} Δ does not cause a substantial change to the evolvement of the number of stable states with increasing voltages, demonstrating that the model’s moderate sensitivity to it. In addition, we have also considered the situation when the network is scaled up. It is observed from Supplementary Fig. S18 in the SI Appendix that as the network scales up, the change of the curve becomes even more subtle, undoubtedly supporting the scalability of the model with non-ideal conductance. Regarding the power consumption, as only one RRAM device switches each time (sequentially) in the circuit, the switching current should not be accumulated. It turns out to be like a chain of automatic memory writing/erase operations and should not cause line destruction and drive circuitry overhead. In between two switching events, the accumulated current on the bit line is given by \\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}$${I}_{{BL}}=\\frac{{\\sum}_{i}{V}_{i}{g}_{i}}{{g}_{0}{\\sum}_{i}{g}_{i}}$$\\end{document} I B L = ∑ i V i g i g 0 ∑ i g i , which is a multiply-and-average (MAV) operation. The sum of conductances in the denominator term should relax the accumulated currents. This is in contrast to the common dot product operation, where more high-conductance devices would contribute to a higher accumulated current, but here the current would be averaged out by the conductance sum. The relaxed current on the bit line should also help alleviate the IR drop issue caused by the line resistance, for which we have carried out circuit simulations of the 8-neuron network. The results show that even when all the devices are in state ‘1’ (HCS), the voltage drop caused by line resistance is merely in the range of mV, which is negligible compared to the threshold voltages ( \\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}$${V}_{{set}}$$\\end{document} V s e t 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}$${V}_{{reset}}$$\\end{document} V r e s e t ) and should have a very limited impact on the circuit dynamics. In addition, although the endurance of RRAM devices is limited with respect to conventional volatile memories 34 , i.e., SRAM and DRAM, the recalling of attractor states through a sequence of switching events is essentially a kind of inference process in the attractor network. Different from the conventional memory application where high endurance is essential, the requirement on the endurance performance of RRAM devices should be much weaker for inference, which is also true for other neural networks for the same application. In this work, we theoretically and experimentally demonstrate that a circuit composed of a column of resistive switching devices is inherently an attractor network. A nonvolatile resistive memory device is an artificial neuron that provides a hysteretic nonlinear activation function. The recursion in the network is enabled by the interaction between the devices, which is formulated as an anti-symmetric weight matrix dictated by the applied voltages. Such an approach facilitates associative memory applications with less hardware cost, sharply opposite to the conventional implementation of the Hopfield network where resistive devices are used as synapses and voltages represent neuronal input/output. We also elaborate an energy function for this network and the convergence analysis therein, by referring to the classic Hopfield network. The unique activation function and synaptic weights endow the resistive memory network with a high storage capacity, which can be tuned in the exponential range. By considering device variations, the network turns into a probabilistic model, which is intriguing and promising for relevant applications. The concept should be extended to other device species, such as unipolar or threshold-switching devices, thus developing or inspiring new attractor network models and hardware solutions."
} | 7,178 |
32687537 | PMC7392344 | pmc | 3,026 | {
"abstract": "With antibiotic resistance rates on the rise, it is critical to understand how microbial species interactions influence the evolution of resistance. In obligate mutualisms, the survival of any one species (regardless of its intrinsic resistance) is contingent on the resistance of its cross-feeding partners. This sets the community antibiotic sensitivity at that of the ‘weakest link’ species. In this study, we tested the hypothesis that weakest link dynamics in an obligate cross-feeding relationship would limit the extent and mechanisms of antibiotic resistance evolution. We experimentally evolved an obligate co-culture and monoculture controls along gradients of two different antibiotics. We measured the rate at which each treatment increased antibiotic resistance, and sequenced terminal populations to question whether mutations differed between mono- and co-cultures. In both rifampicin and ampicillin treatments, we observed that resistance evolved more slowly in obligate co-cultures of E . coli and S . enterica than in monocultures. While we observed similar mechanisms of resistance arising under rifampicin selection, under ampicillin selection different resistance mechanisms arose in co-cultures and monocultures. In particular, mutations in an essential cell division protein, ftsI , arose in S . enterica only in co-culture. A simple mathematical model demonstrated that reliance on a partner is sufficient to slow the rate of adaptation, and can change the distribution of adaptive mutations that are acquired. Our results demonstrate that cooperative metabolic interactions can be an important modulator of resistance evolution in microbial communities.",
"introduction": "Introduction The ability of pathogens to rapidly evolve antibiotic resistance is a pressing global challenge. While resistance frequently evolves in complex microbial communities, relatively little is known about how species interactions influence the evolution of antibiotic resistance [ 1 – 3 ]. Most of the studies that do incorporate multiple species focus more on the role of horizontal transfer of antibiotic resistance between species, and less on the de novo evolution of resistance within genomes [ 3 – 7 ]. Additionally, antibiotic resistance studies in multi-species systems typically involve unknown interactions between species, with some exceptions in modelling [ 8 – 11 ]. The role of specific interspecies interactions in the evolution of antibiotic resistance in microbial communities, therefore, remains largely unexplored. Positive interactions are common in bacterial communities [ 12 ]. One important type of positive interaction is cross-feeding, wherein two species exchange essential metabolites [ 13 ]. The resilience of metabolically interdependent microbial systems to environmental disturbances is a growing field of study, with research being conducted into how these systems resist invasion [ 14 , 15 ] and respond to abiotic environmental changes [ 8 , 16 – 18 ]. We have previously shown that obligate cross-feeding influences the effect of antibiotics over short time scales (i.e. within a single growth curve): the least resistant member of an obligate cross-feeding community constrains the ability of more resistant community members to grow at high antibiotic concentrations, producing “weakest-link” dynamics [ 18 ]. In this case, antibiotic resistance in monoculture (genetic mechanisms conferring an ability to grow at higher antibiotic concentrations) did not predict the sensitivity in co-culture (a phenotypic trait describing ability to grow at high antibiotic concentrations), as the sensitivity was limited by the dependence on the least resistant species. This idea has also been demonstrated by others through modelling approaches [ 8 ]. We hypothesize that the weakest link pattern described above should also hold over evolutionary timescales; that is, at any given point during the evolution of resistance in a metabolically interdependent community, one ‘weakest-link’ species should set the antibiotic sensitivity of the whole community. The obligate cross-feeding interactions would then require that the weakest link species evolve resistance in order for the whole community to rise in the concentration of antibiotic at which growth is possible. We therefore hypothesize that metabolically interdependent communities will be slower to adapt to rising antibiotic levels than their single-species counterparts. Importantly, in this study, we differentiate between sensitivity (the ability of a species or community to grow at a certain antibiotic concentration) and resistance (a genetic change or mechanism that modulates tolerance) [ 19 ]. This distinction is important as we have previously shown that obligate mutualistic interactions can decouple sensitivity and resistance when a highly resistant species is dependent on a less resistant one [ 20 ]. We also hypothesize that weakest link dynamics should affect the mechanisms of resistance evolution. Cross-feeding should change the benefit of some mechanisms; for example, the evolution of shared resistance through antibiotic-degrading enzymes [ 3 , 21 , 22 ] or induction of resistance in a partner species [ 23 , 24 ] could be uniquely selected for in co-culture. Cross-feeding may also make some resistance mechanisms more costly. For instance, altering cell wall permeability may be particularly maladaptive in cross-feeding communities, where exchanged nutrients will typically be at low concentrations [ 25 – 27 ]. Finally, a reduction in the rate of adaptation may drive different mechanisms of resistance to evolve. Varying the rate at which antibiotics are increased has been shown to lead to different evolutionary trajectories [ 28 ]. More rapid changes in antibiotic concentration tend to select for mutations with larger effects that are more costly [ 28 ]. These big effect mutations can trap populations on sub-optimal fitness peaks [ 28 ]. We therefore hypothesize that we will see different mechanisms of resistance evolve in co-culture vs. monoculture, though the exact nature of these differences is unclear. We investigated whether obligate cross-feeding altered the rate and mechanism of antibiotic resistance evolution. We used a previously engineered two-species system wherein a methionine-auxotrophic E . coli consumes lactose and excretes carbon by-products that S . enterica uses as a carbon and energy source, and S . enterica overproduces the methionine required by E . coli . We evolved six replicate populations of each species growing in monoculture (providing E . coli with lactose and methionine, and S . enterica with glucose) and in obligately cross-feeding co-culture (providing both species with lactose only, hereafter referred to as ‘co-culture’) along identical antibiotic gradients of rifampicin or ampicillin. At each transfer, the MIC (minimum inhibitory concentration) was assessed. We also constructed a mathematical model of resistance and used it to assess the generality of our findings, specifically with respect to changes in population size or physiology between monocultures and co-cultures. Our results show that growth in an obligate co-culture slows the rate of adaptation, and sometimes leads to different mechanisms of resistance.",
"discussion": "Discussion Our results demonstrate that obligate cross-feeding can slow the rate of antibiotic resistance evolution, and in some contexts change the mechanisms through which resistance evolves. Relative to monoculture, obligate cross-feeding slowed the rate of adaptation of both E . coli and S . enterica in the face of two different drugs. In rifampicin, genes that acquired resistance mutations in co-culture were a subset of the genes involved in adaptation in monoculture. In ampicillin, in contrast, we observed different resistance mechanisms arising in monoculture and in co-culture. Additionally, we used a model to demonstrate that mutualism may generally slow the rate at which populations evolve resistance. Critically, the model demonstrates that interdependency is sufficient to alter rates and mechanisms of evolution, independent of differences in population size or physiology. Overall, our experimental and modeling results showed that monocultures evolved more resistance than co-cultures, irrespective of species and antibiotic identity, suggesting that slower resistance evolution in cross-feeding co-cultures is a general phenomenon. These results demonstrate the importance of taking ecological context into account when studying the evolution of antibiotic resistance, and our work adds to a growing body of literature on how species interactions shape evolutionary rates in microbial communities [ 36 , 37 ]. We have shown that mutualistic interactions between species can result in slower evolution of antibiotic resistance. This reduced rate of adaptation to antibiotics was highly robust appearing for both bacteria under both drug treatments. However, it is important to note that populations in co-culture experienced additional differences which could have influenced rates of adaptation. For example, in co-culture S . enterica likely consumes a mix of carbon sources while in monoculture it was provided only glucose. Carbon source can alter the evolution of resistance [ 38 ]. S . enterica also had a smaller population in monoculture and previous work has suggested that changes in population size are an important mechanism through which species interactions may influence evolutionary rate [ 36 , 37 ]. However, we do not think that the differences in rate were driven solely by these effects in our experiments for two reasons. First, there was no significant difference in E . coli carbon source or population size in monoculture and co-culture (p = 0.43, S2 Fig ), yet E . coli evolved resistance more slowly in co-culture. While the S . enterica carbon was more variable and its population was smaller in co-culture, the fact that both species adapted more slowly in co-culture argues against carbon source or population size being the sole drivers of the difference in evolutionary rate. Additionally, our mathematical model demonstrates that interdependency is sufficient to drive differences in rates of adaptation to abiotic stressors. In the model resources are abstracted and population sizes of each species were set to be equal in monoculture and co-culture, thereby demonstrating that differences in rate should be expected as a result of the species interactions independent of other effects. Obligate cross-feeding also changed the mechanisms through which resistance evolved, in some cases. While in the rifampicin experiments our results were consistent with similar adaptive trajectories in monoculture and coculture, in ampicillin we saw several cases of genes repeatedly acquiring adaptive mutations in one treatment but not in the other. This was most notable for the ftsI mutations in S . enterica . In this case the difference in evolutionary trajectories between treatments was driven by environmentally-dependant costs. Mutations in ftsI repeatedly arose in co-culture replicates, but were never observed in S . enterica monocultures. While the observed ftsI mutations were viable in co-culture, they were lethal in monoculture. This is particularly interesting as ftsI activity is implicated in interactions between S . enterica and the human immune system; knocking out ftsI increases survival of S . enterica inside macrophage cells [ 35 ]. In previous work, acidic conditions were found to drive survival of ftsI knockouts [ 35 ]. We find a similar phenotypic response to acid in our mutants, though we do not believe that pH is the mechanism allowing growth of the mutants in our co-culture as we do not observe any change in the pH of our media over the course of co-culture growth. One other factor that may contribute to the different mechanisms of resistance evolution in S . enterica is carbon source. In co-culture S . enterica consumes acetate and other carbon sources while in monoculture they were provided glucose. However, ftsI mutants did not grow on acetate minimal medium so we do not believe that acetate is sufficient to drive the observed differences in mechanism. In future work we will further investigate the mechanisms through which our bacterial interactions alter the evolution of a gene involved in interactions with the human immune system. Our current results demonstrate that changing the costs of resistance mutations is at least one way through which co-culture can alter how bacteria evolve resistance to antibiotics. Obligate cross-feeding also changed the mechanisms through which resistance evolved, in some cases. While in the rifampicin experiments our results were consistent with similar adaptive trajectories in monoculture and coculture, in ampicillin we saw several cases of genes repeatedly acquiring adaptive mutations in one treatment but not in the other. This was most notable for the ftsI mutations in S . enterica . In this case the difference in evolutionary trajectories between treatments was driven by environmentally-dependant costs. Mutations in ftsI repeatedly arose in co-culture replicates, but were never observed in S . enterica monocultures. While the observed ftsI mutations were viable in co-culture, they were lethal in monoculture. This is particularly interesting as ftsI activity is implicated in interactions between S . enterica and the human immune system; knocking out ftsI increases survival of S . enterica inside macrophage cells [ 35 ]. Our results indicate that co-culture can alter evolution by altering the cost of some resistance mechanisms, and that this change in cost can lead to changes in genes involved in interactions with the human immune system. It is less clear why growth in co-culture altered the cost of mutations in ftsI . One possibility is that co-culture provided access to different carbon sources. S . enterica consumes acetate and other carbon sources in co-culture and glucose in monoculture. However, ftsI mutants did not grow on acetate minimal medium so we do not believe that acetate is sufficient to drive the observed differences in mechanism. In previous work, acidic conditions were found to drive survival of ftsI knockouts [ 35 ]. We find a similar phenotypic response to acid in our mutants, though we do not believe that pH is the mechanism allowing growth of the mutants in our co-culture either as we do not observe any change in the pH of our media over the course of co-culture growth. Though we do not suspect pH is the driver in our system a growing body of literature implicates pH as a modulator of microbial community dynamics including relative growth rates [ 39 , 40 ], interspecies interactions [ 41 ], and antibiotic resistance [ 42 – 44 ]. It is likely that a similar environmental modification alters the cost of ftsI mutations in our system, however future work will be needed to uncover the physiological mechanism. Our computational model suggests an additional reason mutualism may broadly alter the mechanisms through which organisms evolve resistance to abiotic stress. Our model abstracts the specific genetic mechanisms of our system, and instead focuses on how mutualism alters selection on mutations with different effect sizes. In agreement with previous work [ 45 , 46 ], when large-effect resistance mutations were infrequent, evolution in monoculture simulations was driven by the mutations that caused the biggest MIC impact, despite their rarity. In contrast, in co-culture, resistance evolution occurred primarily through mutations of smaller effect. This outcome makes sense as there is little benefit to evolving resistance beyond that of obligate partners. If small effect mutations are more common and provide the same benefit as big effect mutations, then small effect mutations are more likely to be observed in evolutionary trajectories. We do not observe evidence for this effect in our current experiments likely due to the limits of our sample sizes. As noted above in our experiments, different mechanisms of resistance appear to be driven by changes in environment-dependent costs associated with specific mutations. However, the model demonstrates a broadly applicable mechanism that can lead populations involved in obligate mutualisms to evolve along different evolutionary trajectories than independent populations. Though it was not considered here, cross-protection is also likely to alter evolution of resistance in microbial communities. Cross-protective resistance commonly arises through antibiotic degradation [ 21 , 22 , 47 – 52 ], though other mechanisms such as chemical signalling [ 23 , 24 , 53 ] and spatial coordination [ 54 ] have also been implicated. Neither our E . coli nor S . enterica have β-lactamases or other enzymes that degrade antibiotics, so we did not expect or observe this mechanism in our system. However, in cases where cross-protection is mutationally accessible, we would expect that MIC in co-culture could increase more quickly than in monocultures, as a single resistance mutation would protect all species in the community. There are likely many mechanisms through which bacterial interactions can alter evolution of resistance. We hope that the subset that we have demonstrated here will provide a starting point for further exploration. The observation that mutualism changed the rate and mechanisms of adaptation to abiotic stress has several significant implications. First, the work may inform attempts to combat the crisis of antibiotic resistance. For example, treatment with broad-spectrum antibiotics that inhibit both a pathogen and its obligate cross-feeding partners could slow the rate at which pathogens acquire resistance. More broadly the work provides data on the ongoing debate over how mutualisms will respond to our rapidly changing climate [ 55 , 56 ]. Our results suggest that organisms involved in obligate mutualisms, such as plant-pollinator interactions, may be constrained in their ability to adapt to abiotic stress. Microbial communities provide tractable systems for developing and testing predictive frameworks for the co-evolution of mutualisms. Overall, our work emphasizes the need to take ecological context (and particularly partner resistance) into account when studying the rates and mechanisms through which populations adapt to abiotic challenges."
} | 4,637 |
36936293 | PMC10018510 | pmc | 3,027 | {
"abstract": "Furanic polymers, currently mainly represented by polyethylene\n2,5-furandicarboxylate (PEF), also known as polyethylene furanoate,\nhave a fantastic potential to replace fossil-based polymers: for example,\npolyethylene terephthalate (PET). While 2,5-furandicarboxylic acid\n(FDCA), a precursor of PEF, and its derived polymers have been studied\nextensively, 2,5-bis(hydroxymethyl)furan (BHMF) has received relatively\nlittle attention so far. Similarly to FDCA, BHMF is a biobased platform\nchemical derived from renewable sources such as sugars. This review\nhighlights different polymerization techniques for BHMF-based polyesters\nand addresses BHMF’s relative instability during the synthesis\nof BHMF-derived polymers, including polycarbonates and polyurethanes.\nFurthermore, the degradability of furanic polyesters is discussed\nand BHMF’s toxicity is briefly elaborated.",
"conclusion": "Conclusion This article highlights the synthesis of\nBHMF-based polymers and\nthe degradation of furanic polyesters. The synthesis of BHMF-based\npolymers has been demonstrated to be possible via enzymatic polymerization\nand solution polymerization, but only low molecular weights have been\nachieved to date. The thermal stability issues of BHMF in polymer\nproduction was reported and addressed in the synthesis of polyurethanes,\npolycarbonates, and epoxy monomers. Recent studies have shown that\nBHMF-based polymers showed good thermal stability. A wide range of\nthermal properties of these BHMF-based polymers can be tuned by altering\nthe number of methylene units in the dicarboxylic acid comonomer.\nFurthermore, the furanic moieties in the polymeric chains can be used\nfor thermoreversible cross-linking by Diels–Alder chemistry\nusing a bismaleimide. Considering the diverse properties possessed\nby BHMF-based polymers, understanding their property–application\nrelationship becomes increasingly important. Several studies\nreported on the degradability and compostability\nof furanic polyesters, but no real biodegradability tests have yet\nbeen performed. However, promising results are shown for the enzymatic\ndegradation of these polymers. Furthermore, the degradation rate of\nfuranic polyesters is dependent on multiple properties, including\nmolecular weight, crystallinity, thermal properties, polymeric structure,\nparticle size, and hydrophobicity. It is hypothesized that BHMF-based\npolyesters degrade slowly, but this will be highly dependent on the\nconditions and polymeric properties. In summary, biobased BHMF-derived\npolymers show interesting properties\nand have the potential to replace fossil-based polymers. Understanding\ntheir production route and properties will be critically important\nto reach the full potential of BHMF as a renewable building block\nand will in turn undoubtedly lead us to technological advancements\nin the discipline of green polymer materials. The next step is the\nproduction of BHMF-based polymers in higher amounts to evaluate their\nmacroscopic properties, such as mechanical and gas barrier properties,\nin order to consider in which field these polymers might be applicable.\nIn addition, the development of an industrial or straightforward synthesis\nroute is key to bring BHMF-based polymers to the next stage.",
"introduction": "Introduction Conventional polyethylene terephthalate\n(PET) is indispensable\nin today’s society, since it is ubiquitously used in bottles,\npackaging, fibers, and more applications due to its favorable properties:\nhigh strength to weight ratio, light transmittance, and low permeability. 1 − 3 Although PET is generally seen as nonbiodegradable, a great deal\nof research has been performed on enzymatic degradation or catalytic\ndepolymerization of PET. 1 , 4 , 5 PET is industrially produced from fossil-based ethylene and para -xylene, which are converted to ethylene glycol and\nterephthalic acid or dimethyl terephthalate, which act as the monomers. 2 Although the production of 100% Bio-PET has been\nachieved, a multitude of disadvantages are present, including high\nfeedstock costs, limited resources, and low yields. 2 Biobased polyethylene furanoate (PEF) may be a greener\nalternative\nto PET due to its similarities in chemical structure and enhanced\nproperties, such as a higher T g , lower T m , higher yield stress, and improved gas barrier\nproperties, i.e. lower water, CO 2 , and O 2 permeability. 6 − 10 However, PEF has typically a lower elongation at break and a reduced\ncrystallinity and is still far from cost competitive with PET. 8 , 11 The Dutch company Avantium is one of the main drivers commercializing\nPEF on a large scale by a novel catalytic process for the production\nof furanics. 7 , 9 PEF can be synthesized from the\nmonomers ethylene glycol (EG) and 2,5-furandicarboxylic acid (FDCA).\nEG can be obtained from bioethanol, and FDCA is mostly produced from\nsugars through a multistep process. 7 − 9 5-Hydroxymethylfurfural\n(HMF) is typically derived from sugars\nand is considered a renewable platform chemical due to its versatile\nreactivity and role as a precursor in many possible applications,\ne.g., the production of FDCA. 11 − 13 Another less highlighted monomer\nderived from HMF is 2,5-bis(hydroxymethyl)furan (BHMF). BHMF can be\neasily obtained from HMF in very high yields, up to 99%, via an enzymatic\nor chemical reduction reaction. 11 , 14 − 16 Synthetic routes toward BHMF have been discussed extensively in\nthe literature, including the recent and comprehensive review by Aricò. 15 − 18 Although BHMF is mainly studied as a monomer, it can also act as\na building block for plasticizers, biodiesel additives, flame retardants,\nand surfactants. 19 − 24 Both FDCA and BHMF have the advantage of a relatively higher thermal\nand chemical stability compared to HMF, which is susceptible to degradation\nto levulinic acid, humins, and other degradation products. 13 , 15 , 25 , 26 An example of a potential synthetic production pathway for BHMF\nfrom d -fructose is illustrated in Figure 1 , derived from the work of Cao et al. and\nThananatthanachon and Rauchfuss. 14 , 26 The latter\nconverted d -fructose to HMF in DMSO with formic acid as a\ncatalyst, yielding 99% HMF. On the other hand, Cao et al. used a low-cost\nreusable Cu/SiO 2 catalyst for the HMF reduction to BHMF\nin a 97% yield. More than 30 synthetic production paths have been\nelaborated in the review by Aricò, for example from HMF or\ndirectly from fructose. 15 A simple, cheap,\nand effective route directly from a stable compound is preferred over\nthe use of unstable HMF as the starting material. Furthermore, the\nisolation, scale-up, and purification of the BHMF reaction mixtures\nneed special attention. 15 Figure 1 Two step synthesis route\nfor the production of BHMF from D-fructose,\ndata obtained from refs. ( 14 and 26 ) The extensively studied BHMF synthesis routes are\npromising for\nlarge-scale production; thus, the valorization of BHMF to biobased\npolymers is worth highlighting. Although BHMF is difunctional, it\nusually requires a second difunctional comonomer to produce a polymer.\nIn the case of a polyester, a diacid or a dimethyl ester is necessary\nin order to allow polycondensation. Succinic acid and adipic acid\nare suitable comonomers and are (potentially) obtained from biobased\nresources, like carbohydrates, using micro-organisms like fungi and\nbacteria. 27 , 28 This highlights that it is feasible to produce\nfully biobased BHMF-based polyesters. BHMF is structurally very\nsimilar to FDCA, since both monomers\ncontain an aromatic furan ring and have short side groups at positions\n2 and 5. A difference between BHMF- and FDCA-based polyesters is the\nposition of the ester group, as illustrated in Figure 2 . FDCA and BHMF possess significant differences\nin their melting points, which are 342 and 74–77 °C, respectively. 29 , 30 Typically, FDCA appears as a white crystalline powder, while BHMF\nis more yellowish. However, this difference in color needs to be carefully\ntaken into account, as this yellow color could also be caused by minor\nimpurities, since BHMF has also been reported to appear as a white\nsolid. 31 While FDCA itself is acidic, when\nBHMF is exposed to an acidic environment (e.g., the Brønsted\nacid H 2 SO 4 ) it degrades to humins. 32 Another physiochemical difference between these\ntwo furan compounds is their thermal stability. FDCA can withstand\nhigh temperatures, for example during bulk polymerization at a temperature\nhigher than 200 °C. 33 In contrast,\nBHMF has limited thermal stability and its degradation is already\ninduced at 120–130 °C. 34 , 35 These differences\nmight be explained by the difference in reactivity and electronegativity\nof the carboxylic acid groups and the hydroxy groups close to the\nfuran ring. The carboxylic acid groups of FDCA are already in the\nhighest oxidation state, while the hydroxyl groups of BHMF are still\nprone to oxidation. HMF contains, similar to BHMF, a methoxy group\nattached to the furan ring. It is known that this side of HMF is able\nto react with water to open the ring structure and form 2,5-dioxo-6-hydroxyhexanal\nvia a multistep process, which turns subsequently into humins. 36 , 37 A similar reaction might take place during thermal degradation of\nBHMF in the presence of moisture. Detailed degradation pathways of\nboth FDCA and BHMF are an interesting topic to explore. Other comparisons\nbetween them are their toxicity properties, since it is expected that\nthe different functional groups will influence toxicity as well, which\nis briefly discussed later in this review. Figure 2 Structural representation\nof poly(2,5-furandimethylene succinate)\n(PFS) and poly(ethylene 2,5-furandicarboxylate) (PEF). Nevertheless, based on the chemical similarities\nit is expected\nthat BHMF-based polyesters will have properties similar to those of\ntheir FDCA-based counterparts. Consequently, polymers derived from\nrenewable BHMF are very interesting to study as sustainable alternatives\nfor fossil-based polymers. The current status of literature\ncovering BHMF-based polymers is\nlimited compared to that for other furanic polyesters. This indicates\nthat the promising precursor BHMF is still in the shadow of other\nbiobased building blocks and polymers. Four recent reviews by Lalanne\net al., Zhang et al., Kashparova et al., and Chernyshev et al. focused\nmainly on the synthesis of furanics and FDCA polymers and only shortly\nelaborated on a few examples of BHMF-based polymers. 11 , 38 − 40 To provide a comprehensive overview of the field,\nthis review highlights and summarizes the major peer-reviewed articles\non BHMF-based polymers, thereby discussing the synthesis routes and\ntheir potential. The topic is broadened by including the (bio)degradability\nof furanic polyesters."
} | 2,678 |
29062965 | PMC5637227 | pmc | 3,028 | {
"abstract": "Systems metabolic engineering is a multidisciplinary area that integrates systems biology, synthetic biology and evolutionary engineering. It is an efficient approach for strain improvement and process optimization, and has been successfully applied in the microbial production of various chemicals including amino acids. In this review, systems metabolic engineering strategies including pathway-focused approaches, systems biology-based approaches, evolutionary approaches and their applications in two major amino acid producing microorganisms: Corynebacterium glutamicum and Escherichia coli, are summarized.",
"introduction": "1 Introduction Systems metabolic engineering is an emerging discipline that combines the concepts of systems biology, synthetic biology and evolutionary engineering [1] , as defined by Lee [2] , it involves the application of omics data and the utilization of omics data for synthetic biology and evolutionary engineering for strain breeding and process improvement. With the development of high-throughput technologies, computational methods and simulation approaches, systems biology has become much more mature and applicable, and has already manifested its giant potential in providing genome-wide information and clues for synthetic biology and evolutionary engineering. The various combinations of systems biology, synthetic biology and evolutionary biology have been successfully applied for metabolic engineering of industrial strains [3] , [4] , [5] . Undoubtedly, systems metabolic engineering could dig out the maximum potentials of microbial cell factories. The microbial production of amino acids is a large area where systems metabolic engineering strategies have been successfully applied, mainly in two important producing microorganisms: Corynebacterium glutamicum and Escherichia coli . Since the first discovery of the l -glutamate producing strain C. glutamicum in 1957, strain breeding has become a fierce competing spot of leading amino acid manufacturing enterprises with the expanding market demand for amino acids. l -Glutamate is the major bulk amino acid, which covers nearly two thirds of the amino acid market. The market demand of l -lysine ranks just next to l -glutamate, with an current annual production of over 2200000 tons [6] . Various strain breeding approaches have been developed, whilst genetically defined metabolic strategies have gradually taken the place of the conventional random mutagenesis-selection method and become the mainstream. While local metabolic engineering of microorganisms that focuses on the engineering of one or a few specific genes or metabolic pathways generally has the limitation of being not able to take the whole metabolic process into consideration. Systems metabolic engineering tries to overcome this limitation by combined approaches to obtain rationally designed strains. In this review, systems metabolic engineering strategies and applications for amino acid producing strain improvement are summarized, mainly focusing on the two major industrial production microorganisms: C. glutamicum and E. coli ."
} | 778 |
34960845 | PMC8708435 | pmc | 3,029 | {
"abstract": "An initial drop shape can alter the bouncing dynamics and significantly decrease the residence time on superhydrophobic surfaces. Elliptical footprint drops show asymmetric dynamics owing to a pronounced flow driven by the initial drop shape. However, the fundamental understanding of the effect of viscosity on the asymmetric dynamics has yet to be investigated, although viscous liquid drop impact on textured surfaces is of scientific and industrial importance. Here, the current study focuses on the impact of elliptical footprint drops with various liquid properties (density, surface tension, and viscosity), drop sizes, and impact velocities to study the bouncing dynamics and residence time on non-wettable ridged surfaces numerically by using a volume-of-fluid method. The underlying mechanism behind the variation in residence time is interpreted by analyzing the shape evolution, and the results are discussed in terms of the spreading, retraction, and bouncing. This study provides an insight on possible outcomes of viscous drops impinging on non-wettable surfaces and will help to design the desired spraying devices and macro-textured surfaces under different impact conditions, such as icephobic surfaces for freezing rain or viscous liquids.",
"conclusion": "4. Conclusions Numerical simulations were carried out for the bouncing behavior and residence time of elliptical footprint drops on the ridged surface using the VOF method. For low viscosities (Cases 1–4), e 0 and e + drops on macro-ridge patterns could cause at least 25% and 40% reductions in residence time compared with e 0 drops on flat surfaces, respectively. For moderate viscosity, the residence time of Case 5 only slightly changed with impact velocity. For high viscosities (Cases 6–7), drops increased the residence time constantly so that incorporating a macroscopic texture would no longer promote a reduction in residence time. The trend agreed with the fact that the low P regime for impacts with P < 1 enabled the mass redistributions of drops, which were important for a fall reduction in residence time, such as the formation of the butterfly shape. The asymmetric bouncing dynamics observed in the current work can be regarded as an initial value problem with liquid properties and drops’ ellipticity, compared with the conventional impact dynamics based on circular symmetry. This study provides insight on possible outcomes of impinging viscous drops on the ridged surfaces, and will help to develop the spraying system and superamphiphobic or SHSs for many applications, such as an anti-icing strategy during freezing rain or oil [ 10 , 19 ]. In addition, the ellipticity of drop shape is a controllable factor, which can help to adjust drop repellency from the textured surfaces in industrial applications, such as a dropwise condensation for enhancement of heat exchange performances [ 33 ]. Bouncing dynamics on non-wettable macro-textured surfaces can be extended to applications, such as environmental oil–water separation, waterproofing, and biomedical interactions. For example, the impact of core-shell or Janus drops on the surfaces can be used for a liquid separation method using the asymmetric bouncing dynamics induced by different liquid properties, such as density, viscosity, and interfacial tension [ 34 ]. In addition, bouncing dynamics on non-wettable macro-textured surfaces can also offer practical implications for designing water-repellent fibers or fabrics, such as hydrophobic fibers [ 35 ] and hair array [ 36 , 37 ].",
"introduction": "1. Introduction Drop impacting onto solid surfaces is a natural phenomenon [ 1 , 2 ] and essential for many engineering applications, such as spraying cooling [ 3 ], forensic application [ 4 ], pesticide deposition [ 5 ], inkjet printing [ 6 ], and impact erosion [ 7 ]. Previous studies on the liquid drops impacting on solid surfaces with various wettabilities assisted us to design surfaces for self-cleaning [ 8 , 9 ], anti-icing [ 10 ], and increasing the efficiency of heat exchange [ 11 ] in solar photovoltaics, condensers, and steam turbines, etc. For the past two decades, fluid repellency from the hydrophobic and superhydrophobic surfaces (SHSs) has been an active field for understating the fundamentals and developing diverse applications [ 8 , 12 , 13 ]. SHSs show outstanding anti-wetting properties characterized by the higher water contact angles (>150°) and very small contact angle hysteresis (<5°). The dynamics of the solid−liquid interactions can yield the theoretical Rayleigh limit given by 2.22 ( ρD 3 /8 σ ) 1/2 , independent of the impact velocity, where ρ , D , and σ are the density of liquid, equilibrium diameter, and interfacial tension, respectively [ 12 , 13 ]. The theoretical limit has been accepted as the shortest residence time of symmetrical drop bouncing from SHSs [ 12 ]. The impact dynamics can be determined by the following dimensionless numbers [ 1 , 2 ]: Weber number, We = ρDU 2 / σ , Reynolds number, Re = ρDU / μ , Ohnesorge number, Oh = μ /( ρDσ ) 1/2 , and capillary number, Ca = μU / σ , where U is the impact velocity and μ is the viscosity of liquid. The group of dimensionless numbers allows us to comprehend the relative magnitudes of the inertial and viscous forces and the surface tension. In addition, a dimensionless number, called Impact number ( P = We / Re 4/5 ), was used for a comprehensive estimation on whether capillary or viscosity governed the drop dynamics [ 13 ]. If Impact number was greater than unity, the viscosity effect was dominant in the hydrodynamics. Water is a fluid of which viscosity can be ignored ( μ ~1 mPa s), so it has been widely used in the field of drop impact and wetting. However, effects of viscosity can become visible and significantly change the residence time at low temperatures because the viscosity of water drop increases. Unfortunately, not much attention has been paid to viscous impacts, and most of the studies have been devoted to inviscid impacts, except for the following. Mao et al. [ 14 ] studied viscous liquid drop ( μ ~1–100 mPa s) impact on surfaces with various wettabilities. They predicted the maximum spreading diameter and tendency of drop rebounding as a function of viscosity and static contact angle. Bartolo et al. [ 15 ] investigated the retraction behavior of viscous liquid drop ( μ ~1–205 mPa s) and presented inertial-capillary and viscous-capillary regimes based on Ohnesorge number. Lin et al. [ 16 ] conducted the systematical investigation on viscous liquid drop ( μ ~1–398 mPa s) impact on solid surfaces with various wettabilities, from hydrophilic to SHSs. They focused on studying maximum spreading diameter and spreading time by introducing a modified inertial-capillary time scale and Weber number. Yeong et al. [ 17 ] demonstrated viscous liquid drop ( μ ~1–8 mPa s) impact on inclined SHSs and found that, as the viscosity of the fluid increased, the receding angle of the surfaces reduced significantly, thereby altering a drop’s rebound characteristics. Abolghasemibizaki et al. [ 18 ] investigated liquid drops with various viscosities ( μ ~8–100 mPa s) and impact velocities. They reported that the drop dynamics was related to residence time on non-wettable flat and textured surfaces, and the retraction velocity could be scaled as both inertial-capillary velocity (~( σ / ρD ) 1/2 ) and viscous-capillary velocity (~ σ / μ ). Raiyan et al. [ 19 ] studied the effect of viscosity ( μ ~1–23 mPa s) on bouncing dynamics with and without a macro-ridge by investigating the conditions for observing drop splitting with various viscosities. Recent studies reported that the residence time can be altered below the inertial-capillary time scale ( τ 0 = ( ρD 3 /8 σ ) 1/2 ) by macroscopic surface structures [ 20 , 21 , 22 ] and modifying the initial drop shapes [ 23 , 24 , 25 ] to challenge the limit of the time scale in symmetric bouncing. Bird et al. [ 20 ] demonstrated the residence time reduction by using a single macroscopic ridge, which induced a higher retraction velocity on the ridge and the subsequent redistribution of the drop into butterfly shapes. Afterward, Gauthier et al. [ 21 ] introduced repellent wires to SHSs to investigate the drop dynamics for several drop diameters, wire sizes, and impact velocities. They found a steplike decrease in residence time at intermediate- and high-impact velocities, suggesting the residence time relation, t 0 / n 1/2 , where t 0 is the residence time of the drop on a surface without the macro-texture, and n is the number of lobes of drops. Patterson et al.’s [ 22 ] experimental results showed that the number of intersecting spokes had an effect on Leidenfrost drops’ residence time, and the previous relation of t 0 / n 1/2 did not follow the experimental results when the drop was split to n > 2. The control of drop mobility might be effective in the situation where the target solid must be adjustable in such aforementioned studies. However, if the target solid is uncontrollable, the target liquid or initial shape of the drop might be one of the candidates to modify the impact dynamics. The author’s previous study confirmed that the initial drop shape can alter the bouncing behavior and significantly decrease the residence time if ellipsoidal drops collided on flat surfaces [ 23 ]. Recently, the author’s previous studies proposed the asymmetry of bouncing behavior of spheroidal and ellipsoidal water drops on SHSs to demonstrate a collaboration between the initial drop shapes and ridged surfaces [ 24 , 25 ]. The studies reported the feasibility of shortening the residence time by ~50% using a volume-of-fluid (VOF) method, compared with symmetric bouncing of spherical drops on flat surfaces. For different geometric relationships between the drops and ridge, the non-spherical shapes induced different dynamics in the directions parallel and perpendicular to the macro-texture during the whole of the impact. However, previous explanations for residence time reduction are valid only for water drops and do not hold for viscous liquids. Furthermore, it is necessary to understand how shape distortions of viscous liquid drops affect the liquid repellency from surfaces using macro-texture in practical terms. In this study, it is hypothesized that viscosity and initial drop shape might alter the bouncing behavior considerably and play an important role in designing the desired impinging system and surface modification for many practical applications. The current work focuses on studying the impact of ellipsoidal drops with various liquid properties ( μ ~1–100 mPa s) and drop sizes by predicting the bouncing behavior and residence time ( t c ) on ridged surfaces numerically, using the VOF method [ 26 ]. A rectangular shape of the ridge is chosen as a representative of a ridge because it is simple to be manufactured and widely used. The impact velocity is a very essential factor, responsible for viscous liquid’s repellency from surfaces. The numerical simulation provides a proof-of-concept for a reduction in the residence time compared with symmetric bouncing of spherical drops. The underlying mechanism behind the variation in t c is investigated by analyzing the drop dynamics. In addition, the results are discussed in terms of the spreading, retraction, and bouncing.",
"discussion": "3. Results and Discussion Bouncing behavior of drops on the ridge surface can depend on the geometric configuration between the drops and ridge, as shown in the illustration of Figure 2 b. After drops spread on the ridged surface, they are split into two parts by the ridge, which induces the formation of the inner rim to retract in the outward direction (away from the ridge). Finally, drops behave differently in the x and z directions during the retraction and then bounce off from the surfaces. e + drops can offer an efficient way for decreasing the residence time ( t c ) noticeably compared with spherical and e − drops. For example, e + and e − drops on the ridged surfaces decreased the residence time by approximately 55% and 38% below e 0 drops on flat SHSs for pure water liquid at We = 47, respectively [ 25 ]. The findings were explained in terms of an initial mass distribution and a pronounced flow driven by the distribution during the spreading. The initial shape of the e + drop intrinsically induces the pronounced flow outward in the z direction, which can evolve itself into widespread liquid along the ridgeline, as shown in the illustration of Figure 2 c. In addition, after the split, the two parts start to retract outward and inward along the x direction, thereby leading to high aspect ratios of the liquids aligned on the z direction before the bouncing. In other words, a fast bouncing of the e + drop can originate from a unidirectional retraction that induces the mass and momentum transfer from the x to the y directions, while the role of the z direction is negligible during the retraction. In contrast, the initial shape of the e − drop intrinsically drives the pronounced flow outward in the x direction, which can evolve itself into widespread liquid along the direction perpendicular to the ridgeline, as shown in the illustration of Figure 2 c. Thus, the e − drops are considerably elongated in the x direction, and two fragments continue to move outward in the direction after the mother drop is split. For the asymmetric dynamics of the e − drops, the roles of x and z directions in the mass and momentum transfer cannot be ignored. Accordingly, the shape evolution and residence time of the e − drops show a striking contrast with those of e + drops because the unidirectional retraction can be crucial for rapid bouncing. Table 1 shows various liquid properties used in the current study, which correspond to ethanol or glycerin aqueous solutions in a certain weight percentage. Impact number ( P ) was used for a comprehensive estimation on whether capillary or viscosity governs the drop dynamics. In the current study, assuming that the impact velocity is constant, P increases exponentially with Case number (from Case 1 to Case 7). For example, the hydrodynamics of Case 1 ( P = 0.02–0.1 and Oh = 0.002) and Case 4 ( P = 0.1–0.4 and Oh = 0.008) can be determined by the inertia and capillary forces. For the highest viscosity, the hydrodynamics of Case 7 ( P = 1.3–4.8 and Oh = 0.25) can be governed by the viscosity force. Shape evolutions of ellipsoidal drops with liquid properties and different drop sizes were investigated. First, snapshots of ethanol aqueous solutions (Cases 2 and 4) at a fixed impact velocity are shown in Figure 3 a–f. Each last snapshot for e 0 , e + , and e − drops was captured at the moment of bouncing of drops from the surface and ridge. The drops in Case 4 spread more widely than those in Case 2, as shown in the Figure 3 a–f at 3 ms, because Case 4 has lower surface tension, lower contact angle, and higher viscosity than Case 2, according to the liquid properties. Solid, long dashed, and short dashed lines represent the temporal variation of the half widths, x 1 , x 2 , and z 1 , respectively, as depicted in the inset of Figure 3 h. Temporal evolutions of the half widths for the four Cases indicate that the outer rims ( x 1 ) retract inward further to approach the inner rims ( x 2 ) as P decreases (from Case 4 to Case 1), as shown in Figure 3 g–i. Evidently, low P can play an important role in shortening the residence time in a capillary-dominant regime. High viscosities had a significant effect on altering the bounce dynamics and residence time. Figure 4 shows evolutions in shape and dimensionless half width for Cases 5 and 7 of water/glycerin mixtures, which reveal that a high P leads to small deformations in the spreading and retraction processes, compared with the drops in the capillary-dominant regime. Separated drops retract slowly and then bounce off near the ridge, as shown in Case 5 of Figure 4 a–c. In addition, directly after drops are bouncing, contracted shapes along the z direction are found in e 0 and e + drops, whereas vertically elongated shapes are found in e − drops. Meanwhile, drops are not split by a ridge and then evolve their shapes into spheroids directly after bouncing, as shown in Case 7 of Figure 4 d–f. For high Oh (Cases 5–7), inertial and capillary forces only slightly affect the shape evolution, and viscous force is relatively dominant, which is different from the bouncing dynamics of drops observed at a low Oh (Cases 1–4). Snapshots of the drops with the diameters ( D ) of 1.3, 2.0, and 3.0 mm for Case 1 were obtained at the fixed We , as shown in Figure 5 a–f. The shape evolutions of the two Cases exhibit distinct features of bouncing dynamics, such as a formation of liquid alignment on the z and x directions for e + and e − drops, respectively. This phenomenon leads to the significant reduction in t c of e + drops because the newly formed inner and outer rims retract to the x direction, thereby inducing upward motions of the drops. Figure 5 g–i indicate that the outer rims ( x 1 ) retract to the x direction toward the inner rims ( x 2 ) for e + drops, whereas the outer and inner rims move further away from a ridge until drops are detached from the surface for e − drops. The residence time was predicted as a function of impact velocity for e 0 , e + , and e − drops, as shown in Figure 6 a–c, respectively. e 0 and e + drops show a constant decline in t c at low impact velocity, whereas e − drops show no significant changes in t c at low impact velocity. In addition, e 0 and e + drops exhibit a substantial fall in t c for Cases 1–4 and no significant change at impact velocity above the thresholds, whereas e − drops never decrease the residence time for any cases, although the impact velocity increases. It is found that e 0 , e + , and e − drops on macro-ridge patterns cause 40–54%, 25–47%, and 0–35% reductions in residence time compared with e 0 drops on flat surfaces at U = 1.3 m/s, respectively, which are obtained from Cases 1–4. For a moderate viscosity, the residence time of Case 5 only slightly changes with the impact velocity, whereas those of Cases 6 and 7 for high viscosities increase constantly, as shown in Figure 6 a–c. The trend agrees with the fact that low P regime enables the mass redistributions of drops, which are important for reduced t c , such as the formation of a butterfly shape, whereas high P regime cannot form the shape properly. In other words, the viscous dissipation can retard the retraction process and suppress drop splitting, thereby leading the residence time to increase as the impact velocity increases. Figure 6 d shows the residence time normalized by the initial-capillary time scale as a function of Impact number for Cases 1, 4, 6, and 7. Cases 1 and 4 exhibit almost similar changes in t c / τ 0 with respect to P , which presents a striking contrast with Cases 6 and 7 showing a monotonic increase in t c / τ 0 . It is confirmed that Cases 1–4 under the capillary-dominant regime ( P < 1) have a similar scenario of variation in t c / τ 0 . In the same manner, Cases 6 and 7 under the viscosity-dominant regime ( P > 1) have a similar scenario of variation in t c / τ 0 . Hence, it is concluded that Impact number can govern whether a macro-ridge can lead to a steplike reduction in the residence time or not. Moreover, when the normalized residence time is plotted with the capillary number, it is found that there is a prerequisite of Ca < 0.4 for a fall decrease in the residence time to occur, as shown in Figure 7 . The regime of small Ca indicates that the surface tension attains dominance over the viscosity. To investigate the underlying mechanism behind the variation in residence time, t c * was defined as the residence time of ellipsoidal drops normalized by that obtained from the drop impact on a flat surface for the same e . In addition, t c * can be characterized in terms of several durations, t 1 *– t 4 *; that is, t 1 is the duration for spreading along the ridge, t 2 is the duration for spreading on the substrate, t 3 is the duration for retraction on the substrate, and t 4 is the duration for ascending the ridge, as depicted in the inset of Figure 8 a. To examine the effects of drop size and liquid properties on reduction in the residence time, t c (symbol) and t c * (vertical stack) were investigated as a function of P , as shown in Figure 8 . The t c linearly decreases with an increase of P owing to a decrease in drop diameter, D , as shown in Figure 8 a. The e + drops exhibit the minimal t c among the drops in Case 1. The variation in t c * shows that a large D (low P ) has a significant effect on the reduction in the residence time of e 0 drops, whereas a small D (high P ) has a significant effect on the reduction in the residence time of e − drops, as shown in the cases of D = 3.0 and 1.3 mm of Figure 8 a, respectively. The case of D = 1.3 mm presents a low deviation of t c * between e 0 , e + , and e − drops. Figure 8 b indicates that t c slightly increases with Impact number at P < 1 (capillary-dominant regime), whereas t c greatly increases with Impact number at P > 1 (viscosity-dominant regime). In the capillary-dominant regime, e + and e − drops exhibit the minimal and maximal t c under Cases 1–4, respectively. Moreover, a low P has a substantial influence on the reduction in t c * of the e 0 and e − drops, as shown in Cases 1–4 of Figure 8 b. Meanwhile, for Cases 5–7, the minimal t c appears at e − drops. In the viscosity-dominant regime, t c increases significantly with P because t 4 , the duration for climbing up the ridge, mainly contributes to an increase in t c , as shown in Cases 6 and 7."
} | 5,502 |
37635215 | PMC10463938 | pmc | 3,030 | {
"abstract": "Background Poly-β-hydroxybutyrate (PHB), produced by a variety of microbial organisms, is a good substitute for petrochemically derived plastics due to its excellent properties such as biocompatibility and biodegradability. The high cost of PHB production is a huge barrier for application and popularization of such bioplastics. Thus, the reduction of the cost is of great interest. Using low-cost substrates for PHB production is an efficient and feasible means to reduce manufacturing costs, and the construction of microbial cell factories is also a potential way to reduce the cost. Results In this study, an engineered Sphingomonas sanxanigenens strain to produce PHB by blocking the biosynthetic pathway of exopolysaccharide was constructed, and the resulting strain was named NXdE. NXdE could produce 9.24 ± 0.11 g/L PHB with a content of 84.0% cell dry weight (CDW) using glucose as a sole carbon source, which was significantly increased by 76.3% compared with the original strain NX02. Subsequently, the PHB yield of NXdE under the co-substrate with different proportions of glucose and xylose was also investigated, and results showed that the addition of xylose would reduce the PHB production. Hence, the Dahms pathway, which directly converted D-xylose into pyruvate in four sequential enzymatic steps, was enhanced by overexpressing the genes xylB , xylC , and kdpgA encoding xylose dehydrogenase, gluconolactonase, and aldolase in different combinations. The final strain NX02 (Δ ssB , pBT xylBxylCkdpgA ) (named NXdE II) could successfully co-utilize glucose and xylose from corn straw total hydrolysate (CSTH) to produce 21.49 ± 0.67 g/L PHB with a content of 91.2% CDW, representing a 4.10-fold increase compared to the original strain NX02. Conclusion The engineered strain NXdE II could co-utilize glucose and xylose from corn straw hydrolysate, and had a significant increase not only in cell growth but also in PHB yield and content. This work provided a new host strain and strategy for utilization of lignocellulosic biomass such as corn straw to produce intracellular products like PHB.",
"conclusion": "Conclusions Through eliminating the competitive pathway of PHB production and enhancing the Dahms pathway, an engineered S . sanxanigenens NXdE II was obtained that could utilize xylose alone or mixed sugar (glucose and xylose) better. The final strain produced 21.49 ± 0.67 g/L PHB from corn straw total hydrolysate (CSTH) with a content of 91.2% CDW and the PHB yield was 0.30 ± 0.01 g/g sugar. This was a small advance in reducing manufacturing costs by utilizing lignocellulosic biomass to produce PHB. In the past decades, many research groups have studied PHB production with readily available alternative and inexpensive carbon sources such as lignocellulosic biomass [ 31 ]. Among them, some have achieved good results and the most commonly used strains include C . necator (named R . eutropha earlier), Pseudomonas spp., Bacillus spp., E . coli , and so on. Annamalai and Sivakumar study the PHB production using wheat bran hydrolysate as the carbon source with C . necator NCIMB 11,599 and produced cell dry weight, PHB and yield of 24.5 g/L, 62.5%, and 0.319 g/g sugar respectively [ 47 ]. Recently, Lee et al. constructed a coculture system of C . necator NCIMB 11,599 and Bacillus sp. SM01 can successfully utilize xylose of lignocellulosic biomass to produce PHB. This co-culture system can not only increase PHB production, but also overcome the limitation of sugar consumption [ 33 ]. However, the co-culture system does not solve the problem of co-utilizing glucose and xylose derived from lignocellulosic biomass fundamentally. About one year later, Lee et al. selected a novel Loktanella sp. SM43, which showed high utilization of both glucose and xylose, from a marine environment. Loktanella sp. SM43 could produce PHB using various lignocellulosic hydrolysates as feedstock and PHB production reached the highest at 3.66 ± 0.01 g/L when pine tree hydrolysates were used [ 31 ]. Therefore, the engineered strain NXdE II was excellent in co-utilizing glucose and xylose to produce considerable PHB without a co-culture system. The performance of NXdE II in producing PHB from CSTH also has an advantage over some others without genetic engineering in similar research. In this work, we constructed an engineered strain to produce considerable PHB from lignocellulose waste. It is not only saving the raw material cost but also producing an excellent substitute that is “environment friendly” for petrochemically derived plastics. Furthermore, the engineered strain provided a new platform to produce other valuable chemicals from pyruvate.",
"discussion": "Results and discussion Blocking the production of exopolysaccharide for improving PHB production The precursor of PHB synthesis is acetyl-CoA, which is catalyzed by pyruvate dehydrogenase complex. Pyruvate was formed mainly from the Embden-Meyerhof-Parnas (EMP) pathway, pentose phosphate (PP) pathway, and Entner-Doudoroff (ED) pathway [ 37 ]. However, in strain NX02, the branch pathway towards glucose-1-phosphate, which would be ultimately converted to precursors of exopolysaccharide synthesis existing at the glucose-6-phosphate node, caused the carbon loss towards PHB synthesis. Thus, the relationship between PHB synthesis and exopolysaccharide synthesis is competition in strain NX02 (Fig. 1 ). \n Fig. 1 Overview of vital carbon metabolism pathways (glucose metabolism, xylose metabolism, exopolysaccharide synthesis, and PHB synthesis) in NX02. Arrows indicate the pathway steps encoded by the adjacent labeled gene products. Black arrows indicate the native pathways. Green arrows indicate the overexpressed pathways by addition of copies of native enzymes. Dotted black arrows indicate pathways abolished by deletion of the first glycosyltransferase ( ssB ). The white oval granules represent the product PHB. Orf 00050 , orf 02310 , orf 02740 , orf 06335 , orf 09175 , orf 10285 , orf 13370 , orf 14780 , and orf 14815 represent genes encoding isoenzyme of β-ketothiolase; Orf 02315 , orf 02745 , orf 02775 , orf 02830 , orf 09170 , orf 09185 , and orf 26300 represent genes encoding isoenzyme of acetoacetyl-CoA reductase; Orf 00055 , orf 02300 , orf 02730 , orf 09265 , orf 09550 , orf 20850 , orf 22525 , and orf 28645 represent genes encoding isoenzyme of PHB synthase. Abbreviations: EMP, Embden-Meyerhof-Parnas; ED, Entner-Doudoroff; PP, pentose phosphate; TCA, tricarboxylic acid; phbA , encoding β-ketothiolase; phbB , encoding NADPH-dependent acetoacetyl-CoA reductase; phbC , encoding PHB synthase; xylB , encoding xylose dehydrogenase; xylC , encoding xylonolactonase; kdpgA , encoding 2-keto-3-deoxy D-xylonate aldolase \n To improve PHB production, the metabolic pathway of exopolysaccharide synthesis was interrupted to accumulate more acetyl-CoA. Glucose-1-phosphate was involved in the formation of cell walls, and blocking the pathway from glucose-6-phosphate to glucose-1-phosphate might affect the growth and division of bacterium. Thus, gene ssB encoding the first glycosyltransferase responsible for polysaccharide biosynthesis was selected to be deleted by homologous recombination, resulting in strain NXdE (Fig. 1 ). To test the effect of deleting gene ssB on PHB production, the shake flask fermentations of strain NX02 and NXdE were performed and the cell dry weight and PHB production were detected. As shown in Fig. 2 A, NX02 produced 6.08 ± 0.23 g/L PHB with a content of 63.7% CDW, and NXdE produced 9.24 ± 0.11 g/L PHB with a content of 84.0% CDW. Compared with NX02, the PHB production of NXdE was significantly increased by 76.3%. The results indicated that PHB production had a great improvement by blocking its competitive exopolysaccharide synthesis pathway. To gain insight into the scale-up possibility of PHB production by strain NXdE, batch cultivation was conducted in a 5 L autoclavable fermenter and the results are shown in Fig. 2 B. The strain NXdE obtained 15.32 ± 0.19 g/L cell dry weight and produced 11.45 ± 0.43 g/L PHB at 72 h, which were further increased by 23.9% compared with flask fermentation. According to our previous study, NX02 could produce 14.88 ± 0.83 g/L exopolysaccharide and 6.08 ± 0.23 g/L PHB in 5 L autoclavable fermenters under the same condition [ 37 ]. Both NX02 and NXdE had no glucose left at the end of the fermentation. Notably, the strain NXdE had an excellent sugar conversion rate compared with the original strain NX02 in terms of producing PHB. The high PHB production in NXdE may be due to its multi-copied key genes related to PHB biosynthesis. There are eight copies of phbC , seven copies of phbB , and nine copies of phbA in the genome of S. sanxanigenens NXdE (shown in Fig. 1 ) [ 38 ]. In recent years, some researchers isolated a novel Sphingomonas sp. from argan soil that can produce PHB. They studied the PHB production from argan seeds waste by this strain and the putative PHB produced was 1.92 g/L [ 39 ]. However, the PHB yield of S . sanxanigenens NXdE was 11.45 g/L, which was the highest reported yet among the Sphingomonas genus [ 37 ]. \n Fig. 2 Comparison of fermentation using glucose as a sole carbon source by strain NX02 and NXdE. ( A ) Comparison of PHB yield and cell dry weight of shake flask fermentation in strain NX02 and NXdE; ( B ) Performance of batch fermentation in a 5 L fermenter by engineered strain NXdE. “**” indicates the significant differences (p-value<0.05) of cell dry weight between NX02 and NXdE, and “***” indicates the significant differences (p-value<0.01) of PHB yield between NX02 and NXdE \n Utilization of different proportions of glucose and xylose for PHB production The strain NX02 can naturally utilize glucose and xylose simultaneously to produce bio-products [ 38 ], and its derivative strain NXdE also has this property. Based on this feature, we tested the possibility of strain NXdE co-utilizing glucose and xylose to produce PHB. The flask fermentation of NXdE under different proportions of glucose and xylose (the total amount of sugar mixture was 40 g/L) was performed. Results showed that fermentations using glucose and xylose as co-substrates were not as good as that of using glucose alone, and with the increase of xylose ratio, the cell dry weight and PHB production showed a basically decreasing trend. E.g., the PHB production was decreased to 5.00 ± 0.03 g/L from 9.20 ± 0.29 g/L and the cell dry weight was decreased to 7.48 ± 0.27 g/L from 11.83 ± 0.62 g/L, which were decreased by 45.7% and 36.8%, respectively (Fig. 3 ). It was obvious that PHB yield was reduced with the decrease of cell dry weight, and this phenomenon was common for intracellular products [ 40 ]. Moreover, the PHB content showed a slight downward trend with the increase of xylose ratio in fermentations using a mixed carbon source (Fig. 3 ). We speculated that the decrease of PHB production with the increase of xylose ratio was mainly due to the decrease of bacterium quantity caused by lower energy and precursor supply from xylose metabolism. Theoretically, 1 mol glucose can be converted to 2 mol pyruvate through the glycolysis pathway, and pyruvate can be used in the synthesis of PHB and enter the TCA pathway to provide a mass of energy. One mol xylose will form 1 mol pyruvate through the XI pathway and Dahms pathway while part of xylose will directly come into the TCA pathway through the Weimberg pathway. Thus, 1 mol xylose could only be converted into less than 1 mol pyruvate, which caused the decrease of energy and precursor supply from xylose metabolism (Fig. 1 ). Moreover, the deficiency of precursors was also a potential reason considering the PHB content that from 76.1 to 71.1% showed a decreasing trend with the increase of xylose ratio (Fig. 3 ). \n Fig. 3 The shake flask fermentations of strain NXdE under different proportions of glucose and xylose \n Enhancing Dahms pathway for improvement of PHB yield from xylose Lignocellulose such as straw contains high levels of the suitable fermentable sugar D-xylose, the second-most abundant carbohydrate in nature after glucose [ 41 ]. To decrease the cost of production and implement the high conversion of lignocellulose into biodegradable and environmental-friendly polymer PHB, it is necessary to enhance the utilization of xylose by the engineered strain NXdE. Firstly, we tested the PHB performance with xylose as a sole carbon source between engineered strain NXdE and parent strain NX02. The results showed that the PHB production of NXdE was also increased from 2.41 ± 0.09 g/L to 3.41 ± 0.09 g/L and the cell dry weight was increased from 4.93 ± 0.31 g/L to 7.93 ± 0.45 g/L under the condition of xylose fermentation (Fig. 4 A). Compared with NX02, the PHB production and CDW were increased by 41.5% and 60.9%, respectively. As shown in Fig. 1 , pyruvate is a key metabolite linking the TCA, glycolysis pathway, and xylose metabolism, and is also consumed through PHB synthesis. Therefore, it is necessary to choose a faster metabolic pathway from D-xylose to pyruvate for more efficient PHB synthesis. In this situation, the Dahms pathway was enhanced, which directly converted D-xylose into pyruvate in four sequential enzymatic steps: D-xylose is firstly oxidized by D-xylose dehydrogenase (encoded by gene xylB ) to D-xylonolactone. Subsequently, D-xylonolactonase (encoded by gene xylC ) catalyzed D-xylonolactone to D-xylonate, which then is converted to an intermediate, 2-keto-3-deoxyd-xylonic acid, by xylonate dehydratase (encoded by gene xylD ). Lastly, 2-keto-3-deoxyd-xylonate is oxidized into pyruvate and glycolaldehyde by 2-keto-3-deoxy D-xylonate aldolase (encoded by gene kdpgA ) [ 42 ]. It has been reported that introduction of the Caulobacter crescent genes xylB and xylC allowed the establishment of the Dahms pathway in E . coli . However, the engineered strain had poor growth because the carbon flux of xylose was flown mainly to the Dahms pathway, which led to reduced ATP availability and poor glycolysis [ 43 , 44 ]. Based on the research above, the native genes xylB and xylC of strain NX02 were overexpressed (Fig. 1 ). The resulting strain was NXdE I, and the shake flask fermentation results showed that the CDW of NXdE I (8.93 ± 0.06 g/L) had a 12.6% increase but the PHB production (3.45 ± 0.48 g/L) had no significant change compared with the control strain NXdE (3.41 ± 0.09 g/L PHB and 7.93 ± 0.45 g/L CDW), which were quite different from the report. This might be caused by the existence of the Weimberg pathway which can enter the TCA cycle and supplement the energy for cell growth. Next, the native gene kdpgA encoding 2-keto-3-deoxy D-xylonate aldolase, the key enzyme that could catalyze 2-keto-3-deoxy-xylonate to pyruvate and glycolaldehyde of the Dahms pathway, was overexpressed to increase the amount of precursor pyruvate for improved PHB production (Fig. 1 ). Other three isoenzymes of aldolase from E . coli MG1655, which were encoded by yagE , yjhH , and garL severally were also overexpressed to explore the most appropriate isozyme [ 45 , 46 ]. The resulting strain named NXdE II, NXdE III, NXdE IV, and NXdE V, respectively. Additionally, the enzyme xylonate dehydratase of the Dahms pathway as a shared pathway of the Weimberg pathway was not overexpressed in case excessive carbon flux will flow into the Weimberg pathway followed by the TCA cycle leading to a metabolic overflow. The shake flask fermentations of four engineered strains (NXdE II, NXdE III, NXdE IV, and NXdE V) were performed with xylose as a sole carbon source to explore the ability of PHB production. After overexpressing the three genes xylB - xylC - kdpgA , both PHB production and cell dry weight of strain NXdE II were increased to 7.65 ± 0.31 g/L and 12.07 ± 0.25 g/L drastically, which were increased by 124.3% and 52.2% respectively compared with NXdE I (Fig. 4 A). On the other hand, the increased cell dry weight and PHB production were probably due to the increased pyruvate provided from Dahms pathway. In theory, 1 mol xylose can be converted into 1 mol pyruvate through both Dahms pathway and XI pathway, but the routes are different. In the case of Dahms pathway, the xylose is firstly dehydrogenated to xylonolactone and release a NADPH. Then, the xylonolactone is catalyzed into pyruvate by three steps without energy consumption [ 36 ]. In the case of XI pathway, the xylose is firstly catalyzed to xylulose by xylose isomerase. The resulting xylulose is subsequently phosphorylated to xylulose-5-phosphate and this step will consume ATP as energy. After that, the xylulose-5-phosphate will enter into the PP pathway. In summary, compared with XI pathway which will consume lots of energy and have a long series of reactions, the Dahms pathway takes less energy and is a faster route for synthesis of pyruvate. The more pyruvate also improved the energy productivity of TCA pathway (Fig. 1 ), which was better for cell growth. However, the results of three isoenzymes were not good enough compared with the native aldolase of NX02 (Fig. 4 A). NXdE III, NXdE IV, and NXdE V produced 4.12 ± 0.06, 4.18 ± 0.13, and 5.39 ± 1.03 g/L PHB, respectively. In addition, the PHB/CDW value of NXdE II was about 63.33% and it was the highest among all seven strains (Fig. 4 B). \n Fig. 4 The shake flask fermentations using xylose as a sole carbon source by parent strain NX02 and six engineered strains. ( A ) The PHB yield and CDW of seven strains using xylose as a sole carbon source; ( B ) The PHB content of seven strains using xylose as a sole carbon source \n The PHB content can also be observed simply using an optical microscope (Fig. 5 ). After crystal violet staining, the PHB granules were hard to be stained and it was white while other parts of the cells would be purple. It is obvious that strain NX02 could coproduce PHB and exopolysaccharide (fibrous) while strain NXdE II only produced PHB granules and the PHB granules are larger than those in NX02. The relative gene transcription levels of strain NXdE II were compared with parent strain NX02 through RT-PCR analysis (Fig. 6 ), which indicated that the gene ssB has been knocked down and all the genes related to enhancing the Dahms pathway have been successfully overexpressed. The relative transcription levels of gene xylB , xylC , and kdpgA were increased by 6.77-fold, 3.27-fold, and 4.86-fold respectively. Therefore, NXdE II was chosen for further research. \n Fig. 5 The microscopic observation of strain NX02 ( A ) and NXdE ( B ) at the end of fermentation using xylose as a sole carbon source \n \n Fig. 6 Relative transcriptional levels of genes ssB , xylB , xylC , and kdpgA in strain NX02 and NXdE II cultivated to exponential phase \n Batch and fed-batch fermentations of strain NXdE II using glucose and xylose to produce PHB The batch fermentations with glucose or xylose as carbon sources were performed to investigate the growth and fermentation of the engineered strain NXdE II. The results of batch fermentations are shown in Fig. 7 . When glucose was used as a carbon source, the strain reached the peak of growth at 48 h and produced 6.77 ± 0.61 g/L PHB with a content of 53.0% CDW (Fig. 7 A). The highest yield of PHB in NXdE II was 0.29 ± 0.02 g/g glucose at 48 h. It is obvious that PHB production in glucose by engineered strain enhancing the Dahms pathway was decreased compared with the original strain (Figs. 2 B and 7 A). This phenomenon might be due to that the strategy of enhancing the Dahms pathway caused an imbalance of glucose and xylose metabolism and the cell growth and PHB synthesis was disturbed. The fermentation results using xylose as a carbon source showed that the strain reached the peak of growth at 60 h and produced 12.50 ± 0.18 g/L PHB with a content of 79.9% CDW (Fig. 7 B). The highest yield of PHB in NXdE II was 0.31 ± 0.01 g/g xylose at 60 h. The PHB production of NXdE II through batch fermentation using xylose as a sole carbon source was increased by 63.4% compared with flask fermentation. This result verified that the engineered strain NXdE II could grow and ferment well in xylose. \n Fig. 7 Batch fermentation performance of strain NXdE II using glucose ( A ) or xylose ( B ) as a sole carbon source in a 5 L fermenter \n Based on the constructed engineered strain NXdE II, corn straw total hydrolysate (CSTH) was selected to characterize the PHB production potential [ 34 ]. The detailed method of preparation of corn straw total hydrolysate preparation of CSTH was indicated in our previous work [ 34 ] and herein a brief introduction was given in Materials and Methods part. The CSTH obtained containing 52.6 g/L glucose, 15.1 g/L xylose, and 1.8 g/L arabinose. The arabinose was not considered in this article due to its low content. The fermentation results using 40 g/L CSTH as a carbon source showed that the strain reached the peak of growth at 48 h and produced 11.07 ± 0.30 g/L PHB with a content of 73.9% CDW (Fig. 8 A). The yield of PHB in batch fermentation using CSTH of NXdE II was 0.37 ± 0.01 g/g sugar, which was 1.28-fold higher than batch fermentation using glucose and 1.19-fold higher than batch fermentation using xylose. It can be seen from the results above that the engineered strain enhancing the Dahms pathway has achieved significant improvement in the mixed sugar fermentation of glucose and xylose. On one hand, the fermentation period of strain NXdE II was 24 h shorter than strain NXdE. The shorter production period means lower costs of manufacturing production. On the other hand, the PHB yield of strain NXdE II using CSTH (11.07 ± 0.30 g/L) reached the level of strain NXdE using glucose as a sole carbon source (11.45 ± 0.43 g/L) statistically. Moreover, it can be seen from the fermentation curve (Fig. 8 A) that the proportion of glucose and xylose in the CSTH could be consumed properly by the strain at the same time. This phenomenon proved that the metabolic balance of strain NXdE II was more appropriate for the co-utilization of glucose and xylose in CSTH. In addition, strain NXdE II produced 21.49 ± 0.67 g/L PHB with a content of 91.2% CDW through fed-batch fermentation using CSTH as a carbon source and feeding CSTH (Fig. 8 B). The yield of PHB in fed-batch fermentation using CSTH of NXdE II was 0.30 ± 0.01 g/g sugar. In summary, the resulting strain NXdE II produced 11.07 ± 0.30 g/L PHB with a yield of 0.37 ± 0.01 g/g sugar in the batch fermentation and 21.49 ± 0.67 g/L PHB with a yield of 0.30 ± 0.01 g/g sugar in the fed-batch fermentation. Therefore, corn straw total hydrolysate will be a better alternative substrate to produce PHB in NXdE II, which is a more promising PHB producing strain due to its capability of co-utilization of glucose and xylose efficiently. \n Fig. 8 Batch fermentation performance of strain NXdE II using corn straw total hydrolysate as a carbon source in 5 L fermenters. ( A ) batch fermentation; ( B ) fed-batch fermentation feeding corn straw total hydrolysate"
} | 5,789 |
36211070 | PMC9535648 | pmc | 3,032 | {
"abstract": "In spite of the enormous\npotential of cyanobacteria as a renewable\nenergy source, elevated UV exposure is a major impediment to their\ncommercial viability and productivity. Fremyella diplosiphon is a widely explored cyanobacterium with great biofuel capacity\ndue to its high lipid content. To enhance UV stress tolerance in this\nspecies, we overexpressed the photoreactivation gene ( phr\nA ) that encodes for photolyase DNA repair enzyme in the wild\ntype F. diplosiphon (B481-WT) by genetic\ntransformation. Our efforts resulted in a transformant (B481-ViAnSa)\nwith a 3808-fold increase in the phr A mRNA transcript\nlevel and enhanced growth under UV-B stress. Additionally, DNA strand\nbreaks in the transformant were significantly lower after 12 and 16\nh of UV radiation, with significantly higher dsDNA recovery in B481-ViAnSa\n(98.1%) compared to that in B481-WT (81.5%) at 48 h post irradiation.\nPhotosystem II recovery time in the transformant was significantly\nreduced (48 h) compared to that in the wild type (72 h). Evaluation\nof high-value fatty acid methyl esters (FAMEs) revealed methyl palmitate,\nthe methyl ester of hexadecenoic acid (C16:0), to be the most dominant\ncomponent, accounting for 53.43% of the identified FAMEs in the transformant.\nResults of the study offer a promising approach to enhance UV tolerance\nin cyanobacteria, thus paving the way to large-scale open or closed\npond cultivation for commercial biofuel production.",
"introduction": "1 Introduction The negative impact of\nfossil fuels on the environment and human\nhealth has sparked enormous interest in the development of biofuels\nas a renewable energy source. While cyanobacteria are an ideal third-generation\nfeedstock for a variety of fuels including biodiesel, ethanol, and\nbiogas, these photosynthetic organisms face an immense threat due\nto global climatic changes. 1 , 2 In recent years, a decrease\nin the stratospheric ozone layer due to excessive release of air pollutants\nsuch as chlorofluorocarbons, organobromides, and reactive nitrogen\nspecies has resulted in increased solar UV-B (280–320 nm) reaching\nthe Earth’s surface. 3 Several physiological\nand biochemical processes such as motility, photo-orientation, and\nCO 2 uptake in cyanobacteria are impaired by UV radiation.\nIn addition, it is known to adversely impact biomolecules in these\norganisms, with nucleic acids being the primary targets. 4 Studies by Rastogi et al. 5 and Castenholz\nand Garcia-Pichel 6 have reported that cyanobacterial\ngenomic function and fidelity are adversely affected by UV-B, as the\nDNA molecules directly absorb UV-B radiation inducing DNA strand breaks.\nA variety of mutagenic and cytotoxic DNA lesions including cyclobutane-pyrimidine\ndimers (CPDs), 6-4 photoproducts (6-4PPs), and their Dewar valence\nisomers are induced, disrupting genomic integrity. In addition, cyanobacterial\nUV-B-induced DNA degradation due to thymine dimerization has been\nconfirmed by Anabaena , Nostoc , and Scytonema sp. 7 Additionally,\nUV-B-induced DNA lesions (CPDs and 6-4PPs) can also cause primary\nand secondary breaks since they are associated with transcription\nand replication blockages, leading to the collapse of replication\nforks in CPD-containing DNA. 8 To\ncombat the negative effects of radiation stress, cyanobacteria\nemploy a variety of direct and indirect defense strategies that enable\ntolerance to fluctuating UV levels. The first line of defense employed\nby most cyanobacterial species is the avoidance by migration through\nself-shading or mat formation. 7 Cyanobacteria\nsuch as Anabaena sp. Nostoc commune and Scytonema sp. have the capacity to produce\nUV-absorbing compounds mycosporine-like amino acids and scytonemin\nas a response to UV radiation. 9 Although\ncyanobacteria use these defense mechanisms to combat UV stress, these\nrepair systems can be rapidly overwhelmed by sustained UV radiation. 10 However, some species employ photoreactivation,\na process in which photolyase is activated by the blue wavelength\nof solar light, to reverse and modify nitrogenous bases to their normal\nstate followed by thymine dimer formation caused by UV radiation. 11 Fremyella diplosiphon is a well-studied\ncyanobacterial species that has great potential as a third-generation\nbiofuel agent due to its fatty acid methyl esters. 12 In addition to growth in varying light intensities by modifications\nof its light-harvesting complexes, the organism is extremely amenable\nto genetic transformation. 13 Efforts to\nenhance value-added traits such as halotolerance and cellular lipid\ncontent in this species have enabled unique environmental applications. 14 , 15 A report by Vass et al. 16 has indicated\nthat the Phr A gene plays a role in the DNA repair\nmechanism of Synechocystis sp. PCC 6803 and mutant\ncells lacking the gene were highly susceptible to UV-B damage. To\nthe best of our knowledge, there are no reports to augment UV-B tolerance\nin this organism. The development of a UV-B-tolerant strain would\nbe invaluable to maximize its potential for biofuel production in\nscale-up systems. The objective of the present study was to overexpress\nthe photoreactivation ( phr A ) gene in F. diplosiphon B481-WT to enhance UV-B tolerance.\nSuccessful transformation was confirmed using RT-qPCR and fluorometric\nanalysis of DNA unwinding assays and photosynthetic efficiency, which\nis known to be adversely impacted by UV-B, evaluated as a measure\nof photosystem II functionality. Additionally, the fatty acid methyl\nester (FAME) profile of the transformant was compared to the wild\ntype to determine its biofuel efficacy.",
"discussion": "2 Results\nand Discussion 2.1 Identification, Cloning,\nand Expression of\nPhotolyase phr A Gene in F. diplosiphon In this study, F. diplosiphon UV-B tolerance was enhanced by taking advantage of the gene expression\nsystem of a plasmid vector containing the photolyase gene. Gel electrophoresis\nof the double-digested vector construct revealed bands at the expected\nsizes of ∼1500 and ∼3000 bp for Phr A gene and pGEM-7Zf plasmid, respectively ( Figure S1 ). The high similarity of 97.82% to the phr A gene from B481-WT indicated homology to the photolyase gene in Nostoc sp. ( Figure 1 ). Quantification of the phr A gene transcript\nlevels in the transformant revealed a 3808-fold increase ( p < 0.05) compared to that of the wild-type strain ( Figure 2 ). The phr\nA -overexpressing F. diplosiphon strain was designated as B481-ViAnSa, and the gene sequence was\ndeposited in NCBI Genbank with the accession number MW357071 . DNA\nphotolyase homologues have been reported to be the major factors for\nUV resistance in the cyanobacterium Synechocystis sp. PCC 6803. 16 As observed in our study,\nenhanced gene transcript levels indicated enhanced UV-B tolerance\nvia phr A gene overexpression. The photolyase-deficient Synechocystis sp. mutants incurred a significant amount\nof 70% damaged DNA compared to 30% in the wild type following UV-B\nradiation and were incapable of repairing UV-B-induced DNA damages.\nIn addition, upregulation of the phr A gene by 5.19-\nand 9.98-fold after 30 and 60 min of rewetting dried biofilms in the\ndesert cyanobacterium Chroococcidiopsis exposed to\nMars-like UV flux has been reported. 17 Figure 1 Basic\nlocal alignment search tool analysis of F.\ndiplosiphon B481-WT photolyase gene on the National\nCenter for Biotechnology Information (NCBI) showing a 97.82% similarity. Figure 2 Quantification of photolyase transcript levels in F. diplosiphon wild type (B481-WT) and transformant\n(B481-ViAnSa). Error bars indicate Δ C t values at a 95% confidence interval across four replicates. Given that cyanobacteria use solar energy for essential\nenergy-dependent\nprocesses, harmful UV-B radiation affects several physiological and\nbiochemical processes such as photosynthesis, growth, survival, cell\ndifferentiation, genome integrity, and total lipid profiles. 7 Therefore, we evaluated the efficacy of the transformant\nunder simulated UV-B conditions at an intensity of 3.0 W m –2 at the surface of the cell culture. Our results revealed significantly\nhigh UV-B tolerance in the transformant radiated for 20–160\nmin. While a significant reduction ( p < 0.05)\nin the growth of B481-WT was observed even at 20 min of UV-B exposure,\nthe transformant (B481-ViAnSa) showed no significant reduction of\ngrowth at an exposure time of 40–160 min ( Table S1 ). Interestingly, we observed a significantly rapid\ngrowth recovery of the B481-ViAnSa strain compared to that of B481-WT\n( Figure 3 ). Furthermore,\nirradiation of B481-ViAnSa for 20 min significantly ( p < 0.05) increased the growth rate over a 14 day period compared\nto the nonirradiated transformant, indicating exceptional growth performance\nunder UV stress. These results correlate to the report of Sinha et\nal., 18 where 30 min of UV-B irradiation\nimpaired 50% cell growth in Anabaena sp., Nostoc sp., Nostoc carmium , N. commune , and Scytonema sp., and Anabaena sp. with no recuperation even\nafter 120 min. We observed a reduction in the growth of B481-WT and\nB481-ViAnSa strains in a dose-dependent manner, which is in accordance\nto a previous report in which UV-B inhibited growth in Nostoc muscorum , Pediastrum boryanum , and Aphanothece sp. 19 It is known that thymine–thymine dimer photoproducts, which\nare biologically the most relevant UV-B-induced lesions, account for\n∼75–80% of all UV-B-induced damage, resulting in cyanobacterial\ngrowth reduction. 11 , 16 In addition, an increase in the\nfrequency of thymine dimers was reported in Anabaena sp., Nostoc sp., and Scytonema sp., and Anabaena variabilis PCC\n7937 when exposed to UV-B. 5 , 7 Figure 3 Growth of F. diplosiphon wild-type\n(B481-WT) and transformant (B481-ViAnSa) strains irradiated under\nUV-B (3.0 W m –2 ) for 0–160 min and grown\nin BG11/HEPES media. Growth comparisons of strains at UV-B treatments\nfor 0, 20, 40, 80, and 160 min exposure are shown in panels (A–E).\nDifferent letters above the final time point indicate significance\namong treatment means ( p < 0.05). 2.2 Comparison of DNA Strand Breakages in the\nTransformant and Wild-Type F. diplosiphon Strains FADU assay, an accurate and powerful method for\nthe quantitative analysis of DNA damage, was used to measure DNA strand\nbreaks in the transformant engineered with the phr A (B481-ViAnSa). In prior studies, this technique has accurately measured\nUV-B-induced strand breakage in the cyanobacterium A. variabilis PCC 7937. 5 Quantification of dsDNA damage detected by fluorescence analysis\nof the fluorochrome-bound DNA revealed maximal fluorescence at 450\nnm, while it was lower in ssDNA ( Figure 4 ). The difference between the upper and lower\nfluorescence limits of ds- and ss-samples provided a more reliable\nanalysis of DNA strand breaks in UVt samples since the amount of DNA\ndamage in treated cells is expressed by the difference in fluorescence\nintensities. Using this assay, we detected significantly higher ( p < 0.05) DNA damage in both B481-WT and B481-ViAnSa\nstrains exposed to UV-B at 12 and 16 h ( Figure 5 ) compared to the untreated control (sample-ds).\nThese results are corroborated by previous studies in which DNA lesions\nand strand breaks were reported in the cyanobacteria A. variabilis PCC 7937 and Synechocystis sp., PCC 6803 exposed to UV-B. 5 , 16 While our results indicated\na significant reduction ( p < 0.05) in the dsDNA\nof both strains exposed to UV-B, the transformant exhibited significantly\nless dsDNA breakages compared to wild type. In addition, a significantly\nhigh ( p < 0.05) dsDNA of 60.3, 70.2, and 98.1%\nwere observed in B481-ViAnSa compared to that of B481-WT (50.4, 55.6,\nand 81.5%) at 0, 24, and 48 h post UV-B irradiation. Interestingly,\nwe noted significantly higher ( p < 0.05) dsDNA\nrecovery in B481-ViAnSa (98.1%) compared to that in B481-WT (81.5%)\nafter 48 h ( Figure 5 B). Figure 4 Fluorescence excitation of F. diplosiphon DNA-bound Hoechst 33258. Emission data (emission peak 450 nm) were\nobtained using the maximum wavelength of the excitation peak at 343\nnm. The double-stranded (ds) DNA was not subjected to alkaline unwinding,\nwhile the single-stranded (ss) DNA was subjected to complete alkaline\nunwinding. Figure 5 Percentage double-stranded (ds) DNA in F. diplosiphon wild-type (B481-WT) and transformant\n(B481-ViAnSa) strains after\nexposure to UV-B radiation (3.0 W m –2 (∼8.0\nμmol m –2 s –1 )) for 12 h\n(A) and 16 h (B). Different letters above the standard error bars\nindicate significance between percentages ( p <\n0.05). Based on these results, we hypothesize\nthat higher photolyase activity\ncould have resulted in more efficient DNA repair in the transformant.\nIn addition, this strain exhibited a significantly higher ( p < 0.05) percentage of dsDNA at 16 h of UV-B radiation.\nComparison of gene transcription in the transformant and DNA damage\nshowed an inverse correlation. While the phr A gene\noverexpression in the transformant was significantly high ( p < 0.05) compared to that in the wild type, DNA damage\nas indicated by FADU assay was low. These results indicate that the\noverexpression of the photolyase gene could have reduced thymine dimers\ncaused by UV-B. Our findings are consistent with a report by Chen\net al. 20 where UV-B radiation of Anabaena sp. and Microcystis viridis significantly\ndecreased ( p < 0.05) the percentage of dsDNA due\nto ROS-induced damage, elucidating a correlation between oxidative\nstress and DNA damage. It is known that ROS generated under UV radiation\nstress damages cyanobacterial DNA by reacting with sugars, purines,\nand pyrimidines. 8 Further, ROS-induced\ndamage can indirectly activate Ca 2+ -dependent endonucleases\nin response to increasing intracellular free Ca 2+ and inhibiting\nenzymes involved in DNA replication. Consequently, DNA strand breakage\nis common in cells subjected to oxidative stress linked to UV-B-induced\ndouble-strand breaks. 2.3 Evaluation of Photosystem\nII Activity and\nChlorophyll a Content in the Transformant B481-ViAnSa The ratio of variable and maximum fluorescence ( F v / F m ) of the dark-adapted\nchlorophyll a fluorescence parameter was used to\nmeasure the photochemical efficiency of photosystem II reaction centers.\nComparison of photosystem II activity and chlorophyll a content between the wild-type and transformant strains did not reveal\nsignificant differences ( Figure 6 ). However, we noticed a significant difference ( p < 0.05) in the photosystem II (PSII) recovery rate\nof the UV-treated transformant compared to that of the wild type.\nWhile the transformant PSII recovered in 48 h following UV-B radiation\nat an intensity of 3.0 W m –2 for 1 h, the wild-type\nstrain took 72 h, indicating enhanced photolyase gene activity in\nthe transformant contributing to UV stress tolerance. Our results\nwere consistent with the findings of Vass et al. 16 where the phr A -deficient mutant Synechocystis sp., PCC 6803, lacked the capacity to restore\nPSII activity following UV-B irradiation. The study reported a 70%\nloss of PSII activity in the mutant Synechocystis sp., PCC 6803, and only 30% in the wild type. UV-B radiation is\nknown to affect cyanobacterial photosynthetic performances by causing\ndisassembly of the phycobilisome complexes and photobleaching of critical\nsolar harvesting pigments, which include chlorophyll a , carotenoids, and phycobiliproteins. 21 In Synechocystis sp., UV-B was reported to interfere\nwith cyanobacterial solar energy harvesting phycobilisomes leading\nto their disintegration and potential cell death. 22 Figure 6 Evaluation of photosystem II activity (A) and chlorophyll a (B) content in F. diplosiphon B481-WT and B481-ViAnSa strains after 12 h UV-B radiation. Different\nletters above the error bars indicate significant differences ( p < 0.05). 2.4 Fatty\nAcid Methyl Ester Composition in UV-B-Irradiated\nand Nonirradiated F. diplosiphon Strains,\nB481-ViAnSa and B481-WT Previously, researchers have reported F. diplosiphon to possess valuable biodiesel qualities,\nwhich can maximize biofuel production. 19 , 23 Hence, we\ncompared the high-value saturated and unsaturated FAMEs in the transformant\nto the wild-type strain. Our results showed methyl palmitate, the\nmethyl ester of hexadecenoic acid (C16:0), to be the most dominant\nFAME component, accounting for 53.43 and 51.69% in B481-ViAnSa and\nB481-WT, respectively. Methyl octadecenoate (C18:0), the second abundant\nFAME, accounted for 30.12 and 33.02% in B481-ViAnSa and B481-WT, respectively.\nThis was followed by methyl octadecenoate (C18:1), which accounted\nfor 22% in B481-ViAnSa and 23.02% in B481-WT. Additionally, we detected\nmethyl tetradecanoate (C14:1), methyl hexadecanoate (C16:1), and methyl\noctadecadienoate (C18:2) in both strains ( Table 1 ). The FAMEs identified in the present study\ncorroborate with previous studies in F. diplosiphon treated with gold and iron nanoparticles. 15 , 23 Interestingly, UV-B radiation significantly reduced ( p < 0.05) the percentage of all FAME components, including methyl\npalmitate, which was reduced by 20.51 and 19.25% in B481-ViAnSa and\nB481-WT, respectively ( Table 1 ). This observation has also been reported by Kumar et al., 24 in which UV-B radiation resulted in a decline\nin the FAME content of the microalgae Chlorella sorokiniana . Due to the exposure of cultures to simulated UV-B radiation for\n4 continuous h, a reduction of FAMEs is expected. However, we observed\nsignificantly higher ( p < 0.05) amounts of saturated\nFAMEs in both strains irradiated with UV-B when compared to those\nof the untreated control. Our results are consistent with a previous\nreport study in Lyngbya purpurem , where\nthe saturated FAMEs and lipid saturation index were significantly\ngreater ( p < 0.05) in UV-B-exposed cultures compared\nto those in UVA or PAR. 25 Table 1 Quantitative Composition of Fatty\nAcid Methyl Esters in Transesterified Lipids of Nonirradiated and\nUV-B-Radiated F. diplosiphon Wild-Type\n(WT) and Transformant (B481-ViAnSa) Strains nonirradiated\n(%) UV-B-irradiated\n(%) FAME type B481-ViAnSa B481-WT B481-ViAnSa B481-WT methyl palmitate (C16:0) 53.43 51.69 42.47 41.74 methyl octadecanoate (C18:0) 30.12 33.02 23.33 24.18 methyl octadecanoate (C18:1) 22 23.02 15.99 17.05 methyl tetradecanoate (C14:1) 5.19 5.07 1.18 2.07 methyl hexadecanoate (C16:1) 4.76 4.59 0.87 0.81 methyl octadecadienoate (C18:2) 3.01 3.13 0.91 0.78 In summary, our results indicate that overexpression\nof the phr A gene enhanced F. diplosiphon UV stress tolerance, with enhanced PSII reversal rate and no negative\nimpact on lipids. Considering future projections of increased UV-B\nradiation reaching the Earth’s surface due to environmental\npollution and depletion of the ozone layer, 25 this study has paved the way for cultivating F. diplosiphon in large-scale outdoor systems. Future studies will aim toward screening\nof diverse protective sunscreen compounds in the UV-tolerant transformant,\nwhich lead to the production of environment-friendly sunscreen and\nmoisturizers."
} | 4,783 |
38858621 | PMC11163716 | pmc | 3,035 | {
"abstract": "Background Quorum sensing (QS) is a cell density-based intercellular communication system that controls virulence gene expression and biofilm formation. In Pseudomonas aeruginosa ( P. aeruginosa ), the LasR system sits at the top of the QS hierarchy and coordinates the expression of a series of important traits. However, the role of lasR in phage infection remains unclear. This study aims to investigate the role of lasR QS in phage infection. Methods The P. aeruginosa phage was isolated from sewage, and its biological characteristics and whole genome were analyzed. The adsorption receptor was identified via a phage adsorption assay. Following lasR gene knockout, the adsorption rate and bactericidal activity of phage were analyzed. Finally, real-time quantitative polymerase chain reaction (RT-qPCR) was conducted to explore how lasR promoting phage infection. Results The lytic phage vB_Pae_PLY was isolated and lipopolysaccharide (LPS) was identified as its adsorption receptor. The adsorption rate and bactericidal activity of vB_Pae_PLY were reduced after lasR knockout. RT-qPCR results showed that the expression of galU , a key gene involved in LPS synthesis, was down-regulated, and several genes related to type IV pili (T4P) were also down-regulated in the lasR mutant PaΔ lasR . Conclusions The study showed that QS lasR may promote phage vB_Pae_PLY infection by involving in the synthesis of LPS and T4P. This study provides an example of QS in promoting phage infection and deepens the understanding of phage-bacteria interactions. Supplementary Information The online version contains supplementary material available at 10.1186/s12866-024-03349-7.",
"conclusion": "Conclusions In conclusion, this study demonstrates that lasR may promote phage infection by positively regulating the biosynthesis of LPS and T4P. Disrupting lasR expression leads to decreased phage sensitivity. Given the complex and multifaceted role of QS in host-phage interactions, the future research is needed to uncover the various mechanisms by which QS participates in phage infection.",
"discussion": "Discussion Studies have shown that in some Gram-negative bacteria like Vibrio and Escherichia coli ( E. coli ), QS regulated phage resistance by down-regulating phage receptors on the cell surface. M Mozammel et al. indicated that QS may mediate phage resistance by reducing or eliminating O antigen expression in Vibrio cholerae [ 30 ]. Down-regulation of the outer membrane protein K (OmpK) receptor mediated by QS in Vibrio anguillarum reduced phage adsorption [ 17 ]. In E. coli , the QS system reduced phage receptors [ 16 , 31 ]. Additionally, the study by Xuan et a l. revealed that N-acyl homoserine lactones (AHLs) protected Shewanella baltica from phage infection by decreasing LPS-mediated phage adsorption [ 32 ]. The above studies all hypothesized that QS negatively regulates the expression of receptors required for phage infection. However, while the phenomena and mechanisms of QS-controlled phage resistance have been observed in these bacteria, it has not been clear whether QS has similar effects in other bacteria. Interestingly, the present study found that the knockout of lasR in P. aeruginosa PAO1 decreased the phage sensitivity and adsorption (Fig. 7 ), indicating that lasR positively regulated phage infection. This is consistent with previous observations in P. aeruginosa [ 18 , 33 ]. This study demonstrated that lasR promotes phage infection in P. aeruginosa . Phage adsorption to the bacterial surface is the first and most critical step in phage infection [ 34 , 35 ]. Molecules on the bacterial surface, such as LPS, pili, peptidoglycan components, and outer membrane proteins, can all be attachment sites for phage tails [ 35 – 38 ]. LPS is related to virulence in various Gram-negative bacteria, and is also the receptor of many phages [ 39 – 42 ]. Some LPS-specific phages against P. aeruginosa , and phage-resistant strains generated by mutations in LPS biosynthetic genes have been described in the literature [ 43 – 47 ]. T4P plays a central role in the expression of many phenotypes including motility, multicellular behavior, sensitivity to phages, natural genetic transformation, and adhesion [ 48 , 49 ]. Studies show that many P. aeruginosa phages utilize pili as the primary receptor to infect cells, and mutations or inhibition of T4P synthesis may reduce the adsorption [ 18 , 40 , 41 , 46 , 47 ]. Downregulation of pilA , pilB and pilQ genes reduced T4P-mediated phage adsorption to protect P. aeruginosa from phage infection [ 41 , 50 ]. RT-qPCR analysis in the present study revealed that galU and T4P-related genes ( pilA , pilB , pilC and pilD , pilQ and pilV , pilW and pilY ) were significantly decreased in the PaΔ lasR (Fig. 8 ). LPS is composed of O antigen, core oligosaccharide, and lipid A. galU encodes UDP-glucose pyrophosphorylase (GalU), and was found to be involved in the synthesis of the core region of P. aeruginosa LPS [ 43 , 51 ]. The current study demonstrates that the phage vB_Pae_PLY utilizes surface LPS as the receptor for adsorption (Fig. 6 ). Thus, the reduced phage sensitivity in PaΔ lasR can be largely attributed to the impaired LPS synthesis. In addition, pili also play an important role in phage infection; therefore, downregulation of T4P may also be another reason for reduced phage sensitivity. LPS and T4P are important virulence factors for the opportunistic pathogen P. aeruginosa . On the other hand, serving as phage receptors, LPS and T4P facilitate phage infection with their synthesis being positively regulated by QS. Therefore, lasR QS inhibition will affect the synthesis of bacterial LPS and T4P, resulting in reduced phage adsorption with LPS and T4P as receptors, and weaken the bactericidal effect of phages to a certain extent."
} | 1,471 |
36196006 | PMC10099755 | pmc | 3,036 | {
"abstract": "Summary \n Ester‐linked p ‐coumarate ( p CA) is a hallmark feature of the secondary cell walls in commelinid monocot plants. It has been shown that p CA groups arise during lignin polymerisation from the participation of monolignol conjugates assembled by p ‐coumaroyl‐CoA:monolignol transferase (PMT) enzymes, members of the BAHD superfamily of acyltransferases. Herein, we report that a eudicot species, kenaf ( Hibiscus cannabinus ), naturally contains p ‐coumaroylated lignin in the core tissues of the stems but not in the bast fibres. Moreover, we identified a novel acyltransferase, Hc PMT, that shares <30% amino acid identity with known monocot PMT sequences. Recombinant Hc PMT showed a preference in enzyme assays for p ‐coumaroyl‐CoA and benzoyl‐CoA as acyl donor substrates and sinapyl alcohol as an acyl acceptor. Heterologous expression of Hc PMT in hybrid poplar trees led to the incorporation of p CA in lignin, but no improvement in the saccharification potential of the wood. This work illustrates the value in mining diverse plant taxa for new monolignol acyltransferases. Furthermore, the occurrence of p CA outside monocot lineages may represent another example of convergent evolution in lignin structure. This discovery expands textbook views on cell wall biochemistry and provides a new molecular tool for engineering the lignin of biomass feedstock plants.",
"introduction": "Introduction Lignin is a phenolic polymer found predominantly in the secondary cell walls of vascular plants. The deposition of lignin alters the physicochemical properties of cellulosic cell walls, thereby contributing to mechanical strength, water conduction and plant defence (Barros et al ., 2015 ). At the same time, lignin presents a formidable challenge for the industrial use of plant biomass due to its considerable chemical recalcitrance (Yoo et al ., 2020 ). An improved understanding of lignin biosynthesis and evolution could be harnessed to enhance feedstock species, thereby economising the sustainable use of plant biomass for the production of biochemicals, biofuels, biomaterials and an array of other bioproducts. Lignin is assembled primarily from three monolignol precursors – p ‐coumaryl, coniferyl and sinapyl alcohols – which, respectively, form p ‐hydroxyphenyl (H), guaiacyl (G) and syringyl (S) lignin units once incorporated into a polymer (Boerjan et al ., 2003 ). The biosynthesis of monolignols occurs in the cytosol via the phenylpropanoid pathway. Following the export of monolignols from the cell, lignin polymerisation proceeds by the coupling of monolignol radicals generated by peroxidase and laccase enzymes. As this assembly process is stochastic and dependent solely upon the supply and chemical propensity of monomers for radical coupling, the composition and structure of lignin varies widely across different cell types and plant lineages (Ralph et al ., 2004 ; Novo‐Uzal et al ., 2012 ). The diversification of plants has given rise to a host of other lignin constituents in addition to the three canonical monomers. For example, the lignin of commelinid monocots features ester‐linked p ‐coumarate pendent groups ( p CA; Smith, 1955a ; Higuchi et al ., 1967 ). Similarly, p ‐hydroxybenzoate moieties ( p HB) decorate the lignins of poplar and willow (Smith, 1955b ; Goacher et al ., 2021 ). The lignin found in the bast fibres of kenaf is naturally acetylated (Ralph, 1996 ). And quite remarkably, the fronds of Canary Island date palms have lignin with ester‐linked p CA, p HB, benzoate (BA), ferulate (FA), vanillate (VA) and acetate (Ac) groups (Karlen et al ., 2017 ). Although the biological role of these lignin acylations remains largely unknown, there is considerable interest in harnessing this biochemistry to create designer lignins (Mottiar et al ., 2016 ). Ester‐linked pendent groups arise from the incorporation of γ‐acylated monolignol conjugates that participate in radical coupling reactions (Lu & Ralph, 2002 , 2008 ; Fig. S1A ). Members of the BAHD superfamily of acyltransferases catalyse the assembly of these conjugates before their export and polymerisation (R. D. Hatfield et al ., 2008 ); Withers et al ., 2012 ; Marita et al ., 2014 ; Petrik et al ., 2014 ; Smith et al ., 2022 ). Moreover, due to the inherent plasticity of lignification, acyltransferases have been successfully deployed for lignin engineering. For example, heterologous expression of p ‐coumaroyl‐CoA:monolignol transferase (PMT) genes from rice, Brachypodium and maize led to the incorporation of novel p CA groups into the lignin of transgenic plants, including in Arabidopsis, poplar and alfalfa (Smith et al ., 2015 ; Marita et al ., 2016 ; Sibout et al ., 2016 ). Conjugates of p CA are predominantly assembled in monocots with sinapyl alcohol (Grabber et al ., 1996 ), suggesting that PMT enzymes may have a bias in monolignol substrate preference. It has been proposed that one role of acylated monolignols may be to enhance the incorporation of S‐lignin units during lignification through radical transfer mechanisms (Takahama et al ., 1996 ). Although p CA is a good substrate for peroxidase and laccase enzymes and it can become integrated into synthetic dehydrogenative polymers formed in vitro , p CA occurs primarily as free phenolic pendent groups in lignin (Ralph et al ., 1994 ). This is because phenoxy radicals of p CA preferentially undergo radical transfer rather than coupling reactions during lignification (R. Hatfield et al ., 2008 ); Gani et al ., 2019 ), meaning that only the monolignol moieties of conjugates participate in polymerisation. By contrast, monolignol–FA conjugates engage in radical coupling efficiently at both ends, and may thereby become integrated into lignin polymer backbones (Wilkerson et al ., 2014 ). In this way, the incorporation of backbone‐integrated ferulate esters introduces chemically labile linkages that readily improves the efficiency of lignin depolymerisation (Kim et al ., 2017 ). On the contrary, the impact of pendent p CA groups on biomass recalcitrance remains less clear. Cell‐wall‐bound p CA groups have traditionally been considered a unique feature in the cell walls of monocots (Harris & Hartley, 1980 ; Karlen et al ., 2018 ). However, herein we report that the lignin of kenaf ( Hibiscus cannabinus ), a eudicot species, naturally contains ester‐linked p CA pendent groups in the core tissues of stems. This was unequivocally shown using NMR and derivatisation followed by reductive cleavage (DFRC) methods. We further characterised a BAHD acyltransferase from kenaf that showed activity as a PMT in assays of recombinant enzyme, and when expressed heterologously in transgenic poplar trees. These observations support a growing consensus that lignin acylation, including the incorporation of p CA groups, is not phylogenetically limited to a few select clades but rather occurs more broadly in diverse plant taxa.",
"discussion": "Results and Discussion Kenaf lignin is naturally p ‐coumaroylated The bast fibres of kenaf contain an S‐rich lignin that is extensively acetylated (Ralph, 1996 ). It has been shown that acetyl groups predominantly occur on the γ‐OH of S‐lignin units and that they derive from the incorporation of preacylated monolignols during lignification (Lu & Ralph, 2002 ). In contrast to the bast fibres, kenaf stem core tissues inherently possess a lignin that is less acetylated and has lower levels of S units (Seca et al ., 1998 ). Furthermore, alkaline hydrolysis indicated that mature kenaf stems may have cell‐wall‐bound p CA and FA (Geronikaki et al ., 1979 ; Morrison III et al ., 1999 ). However, pyrolysis GC–MS of kenaf uncovered only trace amounts of hydroxycinnamic acids (Gutiérrez et al ., 2004 ). To examine this further, we analysed kenaf bast and core tissues using solution‐state two‐dimensional 1 H– 13 C heteronuclear single‐quantum coherence (HSQC) NMR. Spectra collected from enzyme‐lignin preparations showed well‐resolved aromatic signatures corresponding to H‐, G‐, and S‐lignin in both bast and core tissues, with bast fibres having a higher proportion of S units (Fig. 1a ). This analysis also revealed the presence of p CA and FA groups in the core tissues, but not in bast fibres. DFRC reactions, which cleave β‐aryl ether bonds in lignin while leaving ester‐linkages intact, were then performed next (Fig. 1b ). Diagnostic dihydro‐ p ‐coumarate (DH p CA) products derived from p CA conjugates with S‐lignin (S–DH p CA) and G‐lignin units (G–DH p CA) were detected in kenaf stem core tissues, but not in bast fibres. Products corresponding to monolignol–FA conjugates were also observed, as has been reported previously (Karlen et al ., 2016 ). Notably, p CA was associated with both G and S units, whereas FA was exclusively conjugated with G units. We also used a modified version of DFRC to look for acetylated lignin units; however, these groups were only detected in the bast, as has been reported previously (Lu & Ralph, 2002 ). Fig. 1 Evidence that kenaf core lignin naturally contains p CA groups. (a) 2D‐NMR 1 H– 13 C HSQC spectra for enzyme‐lignin preparations of kenaf core and bast tissues, as labelled, showing the aromatics region. The colour coding and peak annotations for p ‐hydroxyphenyl (H, mauve), guaiacyl (G, blue) and syringyl units (S and S′, dark and light purple), as well as p ‐coumarate ( p CA, orange) and ferulate groups (FA, pink), are elaborated with the structures shown. Also shown, in grey, are peaks corresponding to phenylalanine (Phe), tyrosine (Tyr), cinnamyl alcohol endgroups (I), cinnamaldehyde endgroups (J) and spirodienone units (D). (b) Multiple reaction monitoring traces from GC–MS analysis of derivatisation followed by reductive cleavage reaction products corresponding to kenaf core and bast tissues, as labelled. The spectrum for kenaf bast is delayed by 15 s. Peaks corresponding to cis‐ and trans‐ isomers of monolignols (H, mauve; G, blue; S, dark purple) and monolignols conjugated with p ‐coumarate (coniferyl 7,8‐dihydro‐ p ‐coumarate, G–DH p CA and sinapyl 7,8‐dihydro‐ p ‐coumarate, S–DH p CA, yellow) and ferulate (coniferyl 7,8‐dihydro‐ferulate, G–DHFA, pink) are labelled and colour‐coded with reference to the structures shown. The observation that p CA groups occur exclusively on the lignin of core tissues in kenaf stems whereas acetyl groups prevail in bast fibres illustrates that kenaf lignin acylation may be highly controlled in a spatiotemporal manner. This is certainly not the only example of such fine‐scale patterning of lignin composition and structure. For example, the outer tissues of Canary Island date palm fronds contain lignin with p CA and p HB groups, whereas the lignin found in the inner tissues contains p HB moieties as well as fewer p CA and BA groups (Karlen et al ., 2017 ). At an even finer scale, p HB groups in poplar stems occur in the cell walls of xylem fibres but are largely absent from adjacent vessels (Goacher et al ., 2021 ). Whether such variability in acylation substantially impacts the physicochemical properties of cell walls, and thereby plays any important biological role remains unclear (Mottiar & Mansfield, 2022 ). Differences in lignin acylation may be driven by spatially restricted expression of the corresponding monolignol acyltransferases; however, the supply of requisite substrates could also play an important role. Our analysis proves the presence of p CA in the core tissues of kenaf stems and provides a more complete view of kenaf lignin structures (Fig. S1B,C ). Recently, p CA was observed (but not remarked upon) in mulberry ( Morus alba ), another eudicot species but from the family Moraceae (Yamamoto et al ., 2020 ). These two examples illustrate that p ‐coumaroylation is not restricted to monocot lineages, as has long been the prevailing view. In an extensive survey of dicots, p CA was released from the cell walls of diverse plant taxa including members of the Bignoniaceae, Caryophyllaceae, Gesneriaceae, Meliaceae, Rubiaceae and Styracaceae families (Hartley & Harris, 1981 ). There have even been reports of small amounts of p CA in Populus spp., particularly in the bark tissues (Pearl & Beyer, 1960 ; Sun et al ., 2001 ). However, it remains to be seen how widely p CA occurs within the eudicots and whether p CA represents a substantial fraction of the total lignin in any of these taxa. Whether or not p ‐coumaroylation, or any lignin acylation for that matter, has an important biological function remains an open question. However, it is becoming clear that acylated lignin is far more common than had been previously recognised (del Río et al ., 2022 ). To date, five PMT enzymes have been identified and characterised, and all occur in commelinid monocots: rice ( Os PMT; Withers et al ., 2012 ), Brachypodium ( Bd PMT; Petrik et al ., 2014 ), maize ( Zm PMT; Marita et al ., 2014 ), sorghum and switchgrass ( Sb PMT and Pv PMT; Smith et al ., 2022 ). In order to shed light on the occurrence of p ‐coumaroylation within the eudicots and potentially develop a new tool for lignin engineering, we set out to characterise a novel BAHD acyltransferase from kenaf, Hc PMT. Recombinant \n Hc PMT produces monolignol– pCA conjugates \n Hc PMT protein was expressed in E. coli and purified by immobilised metal affinity chromatography (Fig. S3 ). The recombinant enzyme was first tested in a series of pairwise assays with individual acyl donor substrates ( p CA‐CoA, BA‐CoA, p HB‐CoA, FA‐CoA and Ac‐CoA) and monolignol acyl acceptors ( p ‐coumaryl, coniferyl and sinapyl alcohols). When provided with p CA‐CoA and any of the three monolignols, recombinant Hc PMT produced monolignol– p CA conjugates (Fig. S4A–C ). Hc PMT also reacted with BA‐CoA to produce coniferyl–BA and sinapyl–BA, but not p ‐coumaryl–BA (Fig. S4D–F ). Similarly, coniferyl– p HB and sinapyl– p HB were produced with p HB‐CoA (Fig. S4G–I ). However, feruloyl‐CoA yielded only sinapyl–FA as a product (Fig. S4J–L ). As FA is linked only to G units in kenaf lignin, Hc PMT is likely not responsible for feruloylation, consistent with a previous report that a distinct feruloyl‐CoA:monolignol transferase (FMT) occurs in kenaf (Wilkerson et al ., 2018 ). And finally, Hc PMT was able to produce sinapyl–Ac and coniferyl–Ac when provided with Ac‐CoA (Fig. S4M–O ). To determine the preferred acyl donor, Hc PMT was subjected to a series of substrate competition assays with all five acyl‐CoA substrates along with each monolignol provided separately. In the presence of the five acyl donors and p ‐coumaryl alcohol, recombinant Hc PMT produced only p ‐coumaryl– p CA (Fig. 2a ). With coniferyl alcohol as the acyl acceptor, coniferyl– p CA was produced along with a lesser amount of coniferyl–BA (Fig. 2b ). Similarly, with sinapyl alcohol and all five acyl donors provided, sinapyl– p CA and sinapyl–BA were produced (Fig. 2c ). The identities of the conjugation products were confirmed by mass spectrometry and with reference to authentic standards. Thus, among the five acyl donor substrates tested, recombinant Hc PMT was most active with p CA‐CoA and BA‐CoA. Since acetyl‐CoA was not a preferred substrate, it is very unlikely that Hc PMT is a contributor to lignin acetylation in kenaf. In summary, the analysis of recombinant Hc PMT supports its designation as a p ‐coumaroyl‐CoA:monolignol transferase enzyme having a preference for sinapyl alcohol. Fig. 2 Competitive enzyme activity assays testing Hc PMT substrate preferences. (a) In the presence of p ‐coumaryl alcohol and five CoA thioesters ( p CA‐CoA, BA‐CoA, p HBA‐CoA, FA‐CoA and Ac‐CoA), Hc PMT produces only p ‐coumaryl– p CA. (b) In the presence of coniferyl alcohol and all five CoA thioesters, Hc PMT produces coniferyl– p CA and coniferyl–BA. (c) In the presence of sinapyl alcohol and all five CoA thioesters, Hc PMT produces sinapyl– p CA and sinapyl–BA. Chromatographic traces show the absorbance at 273 nm. The insets show mass spectra for monolignol– p CA conjugates in blue and monolignol–BA conjugates in orange with m / z values labelled. All reactions were performed at room temperature for 60 min with 1 μg purified recombinant enzyme, 1 mM dithiothreitol, 1 mM of each CoA thioester and 1 mM of each monolignol in 50 mM phosphate buffer (pH 6). Transgenic poplar trees expressing \n Hc PMT have altered lignin Although in vitro assays are invaluable for assessing substrate preference and catalytic potential, they do not necessarily reflect the true enzyme activity in vivo as enzymes that are apparently bifunctional may be catalytically limited by the availability of substrates. We therefore generated transgenic hybrid poplar trees to test the effects of Hc PMT in planta . Five transgenic poplar lines representing independent stable transformation events were selected for in‐depth analysis and were grown under controlled conditions in a glasshouse (Fig. 3a ). After 16 wk of cultivation, some of the transgenic lines appeared to be stunted compared with wild‐type (WT) control trees. Lines 4 and 5 were significantly shorter, and line 5 had a smaller average stem diameter (Fig. 3b ). These lines also showed the highest levels of transgene expression, as measured by RT‐qPCR (Fig. S5 ). In addition, line 5 exhibited extensive sylleptic branching, which can be a sign of metabolic stress in hybrid poplars. Fig. 3 Transgenic hybrid poplar trees expressing Hc PMT. (a) Representative photographs of each transgenic line and the wild‐type (WT) control. (b) Heights and stem diameters measured at the time of harvest are plotted on the left and right axes, respectively, in blue and orange. Values marked with an asterisk are significantly different from the WT control (one‐way ANOVA with Dunnett's test, n = 5 for each line using technical triplicates, P ‐value < 0.05). Error bars correspond to SD. To ascertain whether the expression of Hc PMT affected the biosynthesis and assembly of cell wall components, the Klason lignin content was measured using samples of extractive‐free powdered wood from each line. Compared with WT control trees, there were no significant differences in total lignin content, nor in the proportions of acid‐insoluble or acid‐soluble lignin (Fig. 4a ). Similarly, no consistent differences were noted in the composition of structural polysaccharides (Table S3 ). There were, however, significant differences in lignin composition with lines 1 and 5 having a greater proportion of S units compared with WT control trees (Fig. 4a ). Microscopy of stem cross sections stained with phloroglucinol‐HCl and toluidine blue revealed that there were no obvious changes in lignin distribution, nor in the proportion or dimensions of vessels (Fig. S6 ). Next, the molecular weights of enzyme‐lignin preparations were assessed using gel permeation chromatography. Compared with WT control trees, lignin from lines 1, 3, 4 and 5 had greater weight‐averaged molecular weight, M \n W , as well as an increased dispersity index, Đ \n M (Fig. 4b ). Fig. 4 Lignin analysis of transgenic poplar expressing Hc PMT. (a) The Klason lignin content is plotted on the left axis with the acid‐insoluble and acid‐soluble fractions shown in dark and light blue, respectively. The ratio of syringyl (S) : guaiacyl (G) lignin monomers is plotted on the right axis in orange. (b) The molecular weight of enzyme‐lignin samples is plotted on the left axis on a weight basis ( M \n W ) in dark green and on a number basis ( M \n N ) in light green. The dispersity ( Đ \n M = M \n W / M \n N ) is plotted on the right axis in magenta. Values marked with an asterisk are significantly different from the wild‐type (WT) control (one‐way ANOVA with Dunnett's test, n = 5 for each line using triplicates with P ‐value < 0.05 in a and n = 3 for each line with P ‐value < 0.1 in b). Error bars correspond to SD. Mild alkaline hydrolysis reactions were then performed in attempts to selectively release ester‐linked cell‐wall‐bound components (Figs 5a , S7 ). Substantial quantities of p CA were released from extractive‐free wood powder for all five transgenic lines, whereas WT control trees yielded only trace amounts. Poplar lignin is naturally p ‐hydroxybenzoylated and slight reductions in p HB groups were observed as a consequence of p CA incorporation. Interestingly, it was observed that the biosynthesis and the incorporation of p CA groups did not come at the expense of p HB when Os PMT was expressed in poplar (Smith et al ., 2015 ). Although recombinant Hc PMT also showed some activity in vitro on BA‐CoA, and it is known that BA is compatible with lignification in poplar (Kim et al ., 2020 ), only trace amounts of BA groups were detected (Fig. S7B inset). There was also no change in cell‐wall‐bound Ac groups, further indicating that Hc PMT does not produce acetylated monolignols in planta . Fig. 5 Analysis of lignin acylations in transgenic poplar expressing Hc PMT. (a) The amounts of ester‐linked p ‐coumarate ( p CA) and p ‐hydroxybenzoate ( p HB) released by alkaline hydrolysis are plotted on the left axis in orange and blue, respectively. Cell‐wall‐bound acetyl is plotted on the right axis in green. (b) The amounts of the peracetates of sinapyl dihydro‐ p ‐coumarate (S–DH p CA) and coniferyl dihydro‐ p ‐coumarate (G–DH p CA) released by DFRC are plotted on the left axis in light orange and dark orange, respectively. Sinapyl p ‐hydroxybenzoate (S– p HB) is plotted on the right axis in blue. Values marked with an asterisk are significantly different from the wild‐type (WT) control (one‐way ANOVA with Dunnett's test, n = 5 in (a) and n = 3 in (b) for each line using technical triplicates, P ‐value < 0.05). Error bars correspond to SD. (c) Partial 2D‐NMR 1 H– 13 C HSQC spectra for enzyme‐lignin samples of line 4 and the WT control, as labelled, showing the aromatic region. The peak annotations for p ‐hydroxyphenyl (H, light purple), guaiacyl (G, blue) and syringyl units (S and S′, dark and light purple), as well as p ‐coumarate (orange) and p ‐hydroxybenzoate pendent groups (yellow) are elaborated with the structures shown at the side. Proportions for integrated peak volumes are provided as the mean ± SE for three biological replicates. Derivatisation followed by reductive cleavage reactions were performed next to release intact ester‐linked monolignol conjugates (Fig. 5b ). This work revealed that p CA predominantly occurred as conjugates with S‐lignin units (S–DH p CA) since only minor quantities of G conjugates (G–DH p CA) were detected. The activity of Hc PMT in transgenic poplar was largely consistent with the native activity in kenaf core tissues, considering that poplar is an S‐lignin‐rich species. Endogenous p HB was conjugated only to S units in poplar, as has been observed previously (Mottiar et al ., 2022 ). As a final confirmation of p CA incorporation, enzyme‐lignin preparations from line 4 and the WT control were examined by NMR (Fig. 5c ). HSQC spectra corroborated the thioacidolysis and alkaline hydrolysis results and showed that the incorporation of p CA in poplar slightly altered the S : G lignin monomer ratio and resulted in only a small reduction in p HB groups. Again, no BA groups were observed, indicating that Hc PMT was unable to generate monolignol–BA conjugates in poplar likely because the preferred p CA‐CoA substrate was readily available. Volume integrations indicated that p CA groups represented c . 6.6% of the total lignin in transgenic poplar trees expressing Hc PMT, much greater than the endogenous levels observed in kenaf core tissues. Inspection of the aliphatics region of the spectra showed that the incorporation of additional γ‐acylated monolignol conjugates also led to a small increase in the proportion of β‐aryl ether linkages at the expense of phenylcoumaran and resinol structures (Fig. S8 ). Such changes in lignin composition, molecular weight and structure could potentially impact cell wall recalcitrance, thereby affecting the efficiency of industrial biomass utilisation. Industrial relevance of \n p CA \n The occurrence of p CA in monocot cell walls enhances lignin solubility due to the abundance of free phenolic groups (Lapierre et al ., 1989 ). Similarly, lignin from transgenic Arabidopsis and poplar with novel p CA groups resulting from the expression of Bd PMT was significantly more soluble in alkali (Sibout et al ., 2016 ; Lapierre et al ., 2021 ). However, it has been reported that lignin interacts with hemicelluloses in plant cell walls through electrostatic and other non‐covalent forces (Kang et al ., 2019 ). A greater abundance of p CA pendent groups decorating lignin polymers could conceivably promote such interactions, thereby restricting access to structural polysaccharides. As such, the question of whether p ‐coumaroylation enhances or impedes cell wall digestibility remains largely unresolved. To evaluate whether the expression of Hc PMT in poplar affected biomass recalcitrance, powdered wood was subjected to various pretreatments (dilute alkali, dilute acid and no pretreatment) followed by 72 h of enzymatic hydrolysis (Fig. S9 ). Regardless of pretreatment regime, no improvement in saccharification potential was observed among any of the transgenic lines in terms of either glucose or xylose release. In fact, several lines showed slight reductions in digestibility compared with the WT control. This contrasts with the increased sugar release observed for transgenic Arabidopsis expressing Bd PMT (Sibout et al ., 2016 ). However, no change in cell wall digestibility was reported for alfalfa lines with elevated p CA levels resulting from the expression of Zm PMT (Marita et al ., 2016 ). In a study of inbred maize lines with varying contents of hydroxycinnamates, p CA was negatively correlated with cell wall degradability (Zhang et al ., 2011 ). A possible explanation for these contrasting observations is that the lignin content was reduced in the transgenic Bd PMT Arabidopsis, but unchanged in Hc PMT‐expressing poplar and in these other examples. The observed increases in the average molecular weight of lignin in Hc PMT transgenic poplar could offer an alternative explanation for reduced digestibility, as lignin polymer size and structure may play an important role in chemical recalcitrance (Yoo et al ., 2020 ). Indeed, recalcitrance is a complex parameter influenced by myriad attributes of the biomass, the severity of the pretreatment regime and the efficacy of enzymatic hydrolysis (McCann & Carpita, 2015 ). Whether the introduction of ester‐linked pendent groups impacts biomass recalcitrance remains an open question. In contrast to our findings, a recent report on the expression of Bd PMT in transgenic poplar found a significant increase in saccharification due to lignin p ‐coumaroylation (Lapierre et al ., 2021 ). Similarly, we also recently reported that transgenic poplar with elevated levels of p ‐hydroxybenzoate groups showed reduced recalcitrance (Mottiar et al ., 2022 ). One possible explanation could lie in the timing of lignin acylation. In the current study, we used a constitutive promoter that may have produced acylated monolignols well in advance of lignification. By contrast, those other studies used cell‐wall‐specific promoters, and this could have caused spatiotemporal differences with acylated monomers being incorporated at different stages of lignification and perhaps even in different locations (i.e. in the middle lamella vs in the cell wall). To resolve this question, it would be interesting to analyse poplar biomass from both constructs simultaneously. Although lignin p ‐coumaroylation may or may not improve biomass digestibility, p CA could be a lucrative engineering target in its own right. As p CA groups are ester‐bound to the lignin backbone, treatment with mild alkali can be used to efficiently release these pendent groups. In some applications involving fermentation, such free phenolics could be cytotoxic and detrimental to growth and conversion rates. However, these ‘clip‐offs’ would be commercially valuable if they could be economically separated from the residual biomass (Karlen et al ., 2020 ; Timokhin et al ., 2020 ). Liberated p CA could be used directly as there are well‐established applications in foods, pharmaceuticals and cosmetics (Boz, 2015 ; Pei et al ., 2016 ; Boo, 2019 ). Alternatively, p CA could be converted into platform chemicals and then upgraded into various high‐value bioproducts and biomaterials. For example, work is ongoing to develop enhanced strains of Pseudomonas and Rhodococcus capable of efficient catabolism of p CA to produce organic acids that can be readily transformed into an array of valuable chemicals (Otani et al ., 2014 ; Becker et al ., 2015 ; Mohamed et al ., 2020 ). In this way, the discovery and development of new monolignol acyltransferases may provide valuable molecular tools for lignin engineering and the pursuit of designer lignins (Mottiar et al ., 2016 ). \n \n Hc PMT differs substantially from monocot PMT sequences A comparison of PMT amino acid sequences reveals noteworthy differences and similarities (Fig. 6a ). Although Os PMT, Bd PMT and Zm PMT all share 60% sequence identity, Hc PMT is < 30% identical. When amino acid similarities are considered, Hc PMT is at most 43% similar to the other PMTs. All four sequences bear the universally conserved HXXXD and DXGWG motifs that are hallmarks of BAHD acyltransferases (St‐Pierre & De Luca, 2000 ). The conserved histidine in the former sequence is known to deprotonate the hydroxyl group of acyl acceptors, whereas the latter motif is apparently structurally important (Ma et al ., 2005 ). All the PMT sequences also have conserved motifs characteristic of clade Va, namely YPFAGR, QVTXXXCGG and GXYGN (Tuominen et al ., 2011 ). Fig. 6 Analysis of PMT sequences. (a) An alignment of the Hc PMT amino acid sequence with those of the previously characterised PMTs from rice ( Os PMT), Brachypodium ( Bd PMT) and maize ( Zm PMT). Fully conserved residues are shown in bold, and residues conserved only among the commelinid monocot sequences are shown in blue. The universal motifs characteristic of BAHD enzymes are highlighted in purple, clade‐specific motifs are highlighted in orange, and additional residues discussed in the text are highlighted in green (labelled as Note 1 and Note 2). (b) A phylogenetic tree prepared by maximum likelihood analysis showing the relationship of Hc PMT with putative homologues identified in Populus trichocarpa (Potri), Arabidopsis thaliana (AT), Vitis vinifera (GSVIV), Medicago truncatula (Medtr) and Solanum lycopersicum (Solyc). Sequences for Os PMT, Bd PMT and Zm PMT were included as an outgroup. The values at each node represent statistical confidence based on 1000 bootstrap iterations. The scale bar corresponds to 0.2 substitutions per site. Sequences shown in green are discussed in the text. Although no crystal structures are currently available for any PMT enzymes, comparisons with other BAHD acyltransferases may be insightful. For example, structure‐functional analysis of the hydroxycinnamoyl‐CoA : shikimate hydroxycinnamoyltransferases (HCT) enzymes from Coffea canephora and Sorghum bicolor revealed that glycine and tryptophan residues guard the entrance to the substrate binding pocket (Lallemand et al ., 2012 ; Walker et al ., 2013 ). Although these residues are conserved in the monocot PMTs, alanine and leucine occur instead in Hc PMT, perhaps resulting in a larger active site opening (see Note 1 in Fig. 6a ). Similarly, comparison of the Hc PMT sequence with the recently reported structure of FMT from Angelica sinensis (Liu et al ., 2022 ) points to a series of residues that comprise the acyl acceptor pocket. Interestingly, none of these are conserved between Hc PMT and the other PMT sequences (see Note 2 in Fig. 6a ). However, further interpretations, particularly with regard to the binding of the acyl donor, will undoubtedly require a PMT‐specific structural model. \n p ‐Coumaroylation in kenaf may be the product of convergent evolution Putative homologues of Hc PMT were identified in the genomes of five eudicot species by amino acid homology and then used in the assembly of a phylogenetic tree (Fig. 6b ). As has been observed previously with BAHD acyltransferases (Tuominen et al ., 2011 ), these sequences predominantly clustered by taxon, pointing to gene duplication and neofunctionalisation events that occurred after the common ancestors diverged. This pattern of taxon‐specific expansion is typical of genes involved in secondary metabolism (Ober, 2005 ) and likely reflects the versatility in the active site of BAHD enzymes, which are known to employ a wide range of acyl donors and acceptors. Included among the putative homologues of Hc PMT are several previously‐characterised BAHD enzymes that exhibit diverse activities. For example, AT3G03480 encodes an acetyl‐CoA:( Z )‐3‐hexen‐1‐ol acetyltransferase involved in the biosynthesis of leaf volatiles released upon wounding (D'Auria et al ., 2007 ), and AT5G17540 is involved in the metabolism and homeostasis of brassinosteroids in Arabidopsis (W. Zhu et al ., 2013 ). In grape, GSVIVT01022762001 encodes an anthraniloyl‐CoA : methanol transferase that impacts flavour and aroma (Wang & De Luca, 2005 ). The alcohol acetyltransferases encoded by Solyc08g005770 and GSVIVT01024259001 in tomato and grape, respectively, accept a range of alcohol substrates (Goulet et al ., 2015 ; Maoz et al ., 2018 ). And finally, Potri.013G074500 and Potri.019G043600 encode BAHD enzymes that act as benzoyl‐CoA : benzyl/salicyl alcohol O ‐benzoyltransferases in poplar (Chedgy et al ., 2015 ). Clearly, sequence homology is a poor predictor of activity in BAHD enzymes, as has been noted previously (Luo et al ., 2007 ). The observation that some of the closest homologues of Hc PMT in eudicots have activity on various other acyl donors and acceptors suggests that Hc PMT most likely evolved its enzymatic activity independently of the monocot PMTs. Taxon‐specific clustering in the phylogenetic analysis further indicates that Hc PMT and the related eudicot BAHD sequences arose following the divergence of their common ancestors ( i.e . long after the division between monocots and eudicots). Interestingly, convergent evolution has been documented for BAHD enzymes previously (Luo et al ., 2007 ; Peng et al ., 2016 ). At present, it remains unclear how broadly p CA occurs within the eudicots beyond kenaf (order: Malvales) and mulberry (order: Rosales; Yamamoto et al ., 2020 ). However, it is also possible that PMT enzymes evolved repeatedly among the eudicots. Only further investigations on lignin acylations in diverse lineages will resolve this question. Of course, it is also possible that p ‐coumaroylation, or lignin acylation more generally, could be an ancestral trait that was subsequently lost from many extant genera. However, given that a majority of plant taxa evidently lack lignin acylations altogether, homoplasy appears to offer a more plausible explanation. Independent evolution of plant cell wall components is hardly atypical. For example, mixed‐linkage glucans evolved separately in horsetails ( Equisetum spp.) and in the Poaceae monocots (Fry et al ., 2008 ; Sørensen et al ., 2008 ). Similarly, the biosynthesis of sinapyl alcohol, the monomer that yields S‐lignin units, developed independently in angiosperms, the spikemoss Selaginella moellendorffii and apparently in various basal gymnosperms (Weng et al ., 2008 ; Espiñeira et al ., 2011 ). It has also been postulated that the ability to acylate monolignols with ferulate has arisen at least twice since nonhomologous FMT sequences occur in the eudicot A. sinensis and among the commelinid monocots (Wilkerson et al ., 2014 ; Karlen et al ., 2016 ). Indeed, the same logic may hold for other lignin acylations. For example, p ‐hydroxybenzoylation of lignin has been detected in the disparate eudicot families Salicaceae and Araliaceae and in the monocot families Arecaceae and Posidoniaceae (Smith, 1955b ; Pearl et al ., 1959 ; Hibino et al ., 1994 ; Rencoret et al ., 2020 ). The discovery of p CA in a eudicot species elicits new questions on the evolution of lignin. An equally important pursuit for future studies will be to investigate whether lignin acylations have any important biological role(s) in cell walls that could be adaptive for plants. In the meantime, the availability of Hc PMT provides a valuable new tool to modify lignin pendent groups and enable further progress in lignin engineering."
} | 9,262 |
36792979 | PMC9930361 | pmc | 3,039 | {
"abstract": "Background Artisanal and small-scale gold mining activities are producing contamination with heavy metals and metalloids (HMM) into soils and water worldwide. The HMM are considered as one of the major abiotic stresses due to their long-term persistence in soil. In this context, arbuscular mycorrhizal fungi (AMF) confer resistance to a variety of abiotic plant stressors including HMM. However, little is known regarding the diversity and composition of AMF communities in heavy metal polluted sites in Ecuador. Methods In order to investigate the AMF diversity, root samples and associated soil of six plant species were collected from two sites polluted by heavy metals, located in Zamora-Chinchipe province, Ecuador. The AMF 18S nrDNA genetic region was analyzed and sequenced, and fungal OTUs were defined based on 99% sequence similarity. Results were contrasted with AMF communities from a natural forest and from reforestation sites located in the same province and with available sequences in GenBank. Results The main pollutants in soils were Pb, Zn, Hg, Cd and Cu with concentrations exceeding the soil reference value for agricultural use. Molecular phylogeny and OTU delimitation showed 19 OTUs, the family Glomeraceae was the most OTU-rich followed by Archaeosporaceae, Acaulosporaceae, Ambisporaceae and Paraglomeraceae. Most of the OTUs (11 of 19) have been found at other locations worldwide, 14 OTUs were proven from nearby non-contaminated sites in Zamora-Chinchipe. Conclusion Our study showed that there are no specialized OTUs at the studied HMM polluted sites, but rather generalists adapted to a wide variety of habitats. Their potential role in phytoremediation approaches remains to be investigated.",
"conclusion": "Conclusion We investigate the AMF diversity and associated soil of six plant species growing at two sites polluted by heavy metals. Overall results showed that there are no specialized OTUs at the studied HMM polluted sites, but rather generalists adapted to a wide variety of habitats.",
"discussion": "Results and discussion Physico-chemical analyses of soil samples and presence of AMF colonization The analyses of 9 soil samples from Chinapintza and 21 soil samples from La Pangui showed low pH values, low nutrient content and high heavy metal concentration (Tables 1 , 2 ). The lowest pH value and the highest concentration of heavy metals were detected at Chinapintza site (Table 2 ). The main heavy metals found at both sites were Pb, Zn, Hg, Cd and Cu, all with concentrations exceeding the soil reference value for agricultural use [ 10 ]; Table 2 ). There was a wide variability in the concentrations of the heavy metals from the different samples, which was also observed by Chamba et al. [ 11 ]. In our study, we consider the same species as [ 11 , 12 ], Axonopus compressus , Erato polymnioides and Miconia zamorensis , but in addition, Medinilla sp., Colacasia sp. and Cyathea sp. Table 1 Physico-chemical parameters analysed from soil samples collected from Chinapintza and La Pangui sites of Zamora-Chinchipe, Ecuador Name of site pH Levels of major elements in ppm Levels of trace elements in m.eq / 100 ml Level of Na in meq / 100 g soil NH4 P S Cl K Ca Mg Fe Mn B Chinapintza * 3.39 ± 0.56 47.75 ± 29.25 17.34 ± 7.66 116.00 ± 80.84 89.80 ± 16.20 0.21 ± 0.13 0.68 ± 0.28 0.53 ± 0.45 1473.80 ± 926.2 150.64 ± 44.36 0.66 ± 0.14 0.22 ± 0.02 La Pangui ** 4.25 ± 2.21 75.44 ± 39.66 22.54 ± 10.46 168.01 ± 33.99 83.80 ± 17.20 0.19 ± 0.14 2.98 ± 1.02 0.68 ± 0.23 661.20 ± 406.80 67.13 ± 40.14 0.82 ± 0.12 0.28 ± 0.10 * Mean value of 9 data ** Mean value of 21 data Table 2 Levels of heavy metals on soils recorded from Chinapintza and La Pangui sites of Zamora-Chinchipe, Ecuador Name of site Levels of heavy metals in ppm Al Cd Cr Cu Pb Zn Hg Au Chinapintza * 9426.83 ± 2211.24 4.25 ± 9.73 17.43 ± 3.51 131.84 ± 117.89 1501.25 ± 915.31 886.16 ± 595.30 26.26 ± 13.03 15.99 ± 7.13 La Pangui ** 7922.37 ± 3556.19 3.13 ± 2.19 13.02 ± 8.50 94.83 ± 60.44 560.53 ± 283.53 460.42 ± 390.55 17.33 ± 12.88 9.58 ± 6.39 Soil quality reference *** 1.4 – 22.0 64.0 – 87.0 63.0 – 91.0 70.0 – 600.0 200.0 – 360.0 6.6 – 50.0 * Mean value of 4 data ** Mean value of 6 data *** Canadian Council of Ministers of the Environment (2007). The lower level is a reference value for agricultural and higher level is for industrial use Despite the adverse soil conditions, examined root samples in both sites were moderate to highly colonized by AMF (40–80%) and showed the usual characteristics of AMF such as arbuscules, coils, extra and intracellular hyphae and vesicles (data not shown). This finding is consistent with Chamba et al. [ 11 ] in the same area of Chinapintza that showed mycorrhizal colonization of up to 70% in E. polymnioides , M. zamorensis and A. compressus (70 ± 7, 60 ± 5 and 50 ± 5%) respectively. Long et al. [ 35 ] earlier reported moderate to high degree of mycorrhizal colonization in five plant species growing in acidic soils severely polluted with Zn, Pb, Cu, and Cd. The extent of AMF colonization can be interpreted as positive correlation to plant dependence on symbiosis [ 55 ], even more under extreme soil conditions. The lack of data on essential soil factors, such as measurements of cation exchange capacity (CEC), phosphorus, and organic matter, is certainly a limitation of this study. Alguacil et al. [ 1 ] observed an increase in the percentage of colonization by AMF and a decrease in the concentration of heavy metals in native plants growing in polluted soil with organic amendments, indicating an increase in the resistance to heavy metals stress. In other cases, immobilization of metals such as Pb and Cd in roots and stems increases plant tolerance to heavy metals in presence of AMF colonization [ 24 , 52 ]. Molecular phylogeny and OTU delimitation Successful PCR amplification was obtained from 18 plant samples in total, 7 samples from Chinapintza and 11 samples from La Pangui. After cloning 78 sequences of AMF were obtained, 64 sequences were grouped in 19 OTUs (Fig. 1 a, b and Table 3 ) and 14 sequences were singletons. The family Glomeraceae was the most diverse family displaying 52% (33) of all sequences and 53% (10) of all OTUs (OTU 1 to 10, Fig. 1 a). The Archaeosporaceae were represented with 15 sequences and 5 OTUs (OTU 13 to 17), Acaulosporaceae with 12 sequences and 2 OTUs (OTU 11 and OTU 12), Ambisporaceae with 3 sequences and 1 OTU (OTU 18) and Paraglomeraceae with 1 sequence and 1 OTU (OTU 19) (see Fig. 1 b). The most frequent OTU 1 Glomeraceae occurred in 72% of the samples (Table 3 ). The OTU 11 Acaulospora species is also frequent, occurring in 56% of the samples and a further Glomeraceae OTU 6 present in 44% of the samples. Many of the other OTUs were proven in smaller numbers, 8 only once. Fig. 1 Phylogram inferred from a Maximum Likelihood (ML) analysis of partial 18S nrDNA sequences of Glomeromycotina associated with six plant species growing at heavy metal polluted soils in Chinapintza and La Pangui sites including sequences from the databases NCBI and MaarjAM with high similarity. The values that support the nodes correspond to Maximum Likelihood bootstrap. Only bootstrap values greater than 50% are shown. The tree was outgroup rooted with Endogone pisiformis X58724. OTUs are defined on a 99% similarity threshold. Sequences from this study are indicated by the species name, followed by individual number, clone number and location (Ch = Chinapintza and LP = La Pangui). Sequence provenances: BRA = Brazil, CAN = Canada, CHN = China, DEU = Germany, ECU = Ecuador, GBR = United Kingdom, IND = India, IRL = Ireland, ITA = Italy, JPN = Japan, LSO = Lesotho, MYS = Malaysia, NZL = New Zealand, SPN = Spain, USA = United States, VEN = Venezuela, VNM = Vietnam, ZAF = South Africa. Phylogenetic tree was divided into Fig. 1a and 1b Table 3 Frequency of AMF OTU at each plant individual recorded from samples collected from Chinapintza and La Pangui sites of Zamora-Chinchipe, Ecuador Plan species OTU * 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 Axonopus compressus 1 1 1 1 3 Cyathea sp 1 1 1 1 1 4 Cyathea sp 2 1 1 2 Colacasia sp 1 1 1 1 3 Colacasia sp 2 1 1 1 1 4 Erato polymnioides 1 1 1 1 1 1 5 Erato polymnioides 2 1 1 Erato polymnioides 3 1 1 1 3 Erato polymnioides 4 1 1 1 1 1 5 Medinilla sp 1 1 1 1 1 4 Medinilla sp 2 1 1 2 Medinilla sp 3 1 1 2 Medinilla sp 4 1 1 1 1 1 5 Medinilla sp 5 1 1 1 1 1 5 Miconia sp 1 1 1 1 3 Miconia sp 2 1 1 1 3 Miconia sp 3 1 1 2 1 5 Miconia sp 4 1 1 1 1 1 5 Total 13 4 1 1 1 8 1 2 1 1 10 2 5 5 1 2 2 3 1 64 La Pangui 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 15 Chinapintza 1 1 1 1 1 1 1 1 1 1 1 11 * OTU 1 to 10 correspond to Glomeraceae, OTU 11 and OTU 12 to Acaulosporaceae, OTU 13 to 17 Archaeosporaceae, OTU 18 Ambisporaceae and OTU 19 Paraglomeraceae Most of the OTUs (14 of 19) were previously found in other locations in Ecuador, mostly in Zamora-Chinchipe province [ 20 , 22 ], whereas 9 OTUs were previously found in Ecuador and several locations worldwide. For three OTUs (OTUs 12, 13 and 17) there were no proofs from other sites (Fig. 1 b and Table 4 ).\n Table 4 List of sequences and previous recorded distribution corresponding at each OTU recorded from samples collected from Chinapintza and La Pangui sites of Zamora-Chinchipe, Ecuador OTU Freq Plant species Location Family of Glomeromycotina Previous recorded distribution LP Ch 1 13 Colacasia sp. 1_8 1 Glomeraceae Ecuador 1,2 and several locations worldwide (MYS 3 , JPN 4 , ZAF 3 , GBR 3 ) Medinilla sp. 1_7 1 Miconia zamorensis 3_8 1 Miconia zamorensis 2_8 1 Cyathea sp. 2_5 1 Medinilla sp. 3_4 1 Cyathea sp. 1_2 1 Medinilla sp. 2_5 1 Medinilla sp. 5_3 1 Erato polymnioides 3_4 1 Erato polymnioides 4_1 1 Erato polymnioides 1_3 1 Miconia zamorensis 4_2 1 2 4 Erato polymnioides 1_9 1 Glomeraceae Ecuador 1,2 and USA 5 . Correspond to VTX00074 at MaarjAM database Medinilla sp. 4_3 1 Cyathea sp. 1_4 1 Erato polymnioides 4_6 1 3 1 Erato polymnioides 3_2 1 Glomeraceae Ecuador and GBR 3 4 1 Axonopus compressus 1_2 1 Glomeraceae Ecuador 5 1 Axonopus compressus 1_5 1 Glomeraceae Several locations worldwide (CHN 6 , USA 7 , DEU, JPN) Correspond to VTX00309 at MaarjAM database 6 8 Miconia zamorensis 4_4 1 Glomeraceae Ecuador 1 , VEN 8 and BRA 3 . Correspond to VTX00268 at MaarjAM database Erato polymnioides 1_1 1 Erato polymnioides 2_1 1 Medinilla sp. 1_1 1 Erato polymnioides 4_9 1 Medinilla sp. 4_5 1 Medinilla sp. 3_9 1 Medinilla sp. 2_3 1 7 1 Medinilla sp. 5_9 1 Glomeraceae Ecuador 1 . Correspond to VTX00292 at MaarjAM database 8 2 Miconia zamorensis 3_5 1 Glomeraceae Ecuador. Correspond to VTX00292 at MaarjAM database Miconia zamorensis 4_1 1 9 1 Colacasia sp. 2_5 1 Glomeraceae Ecuador 1 and several locations worldwide (SPN 9 , CHN 10 , IND 3 ). Correspond to VTX00166 at MaarjAM database 10 1 Miconia zamorensis 2_1_X 1 Glomeraceae Brazil. Correspond to VTX00418 at MaarjAM database 11 10 Medinilla sp. 4_4 1 Acaulosporaceae Ecuador 1 and several locations worldwide (VEN, JPN, CHN 3 ) Medinilla sp. 5_4 1 Colacasia sp. 2_9 1 Miconia zamorensis 1_3 1 Miconia zamorensis 2_5 1 Colacasia sp. 1_11 1 Medinilla sp. 1_3 1 Miconia zamorensis 3_6 1 Miconia zamorensis 3_9 1 Miconia zamorensis 4_5 1 12 2 Colacasia sp. 2_8 1 Acaulosporaceae Miconia zamorensis 1_6 1 13 5 Cyathea sp. 1_5 1 Archaeosporaceae Medinilla sp. 5_5 1 Medinilla sp. 4_8 1 Cyathea sp. 2_1 1 Erato polymnioides 4_4 1 14 5 Erato polymnioides 1_4 1 Archaeosporaceae Ecuador 1 and several locations worldwide (GBR 3,10 , BRA 3 , CHN 3 , NZL 11 ). Correspond to VTX00005 at MaarjAM database Medinilla sp. 1_2 1 Medinilla sp. 5_10 1 Miconia zamorensis 3_7 1 Colacasia sp. 1_1 1 15 1 Erato polymnioides 1_7 1 Archaeosporaceae Ecuador 1,2 16 2 Medinilla sp. 4_1 1 Archaeosporaceae Ecuador 1 Cyathea sp. 1_1 1 17 2 Erato polymnioides 3_11 1 Archaeosporaceae Erato polymnioides 4_10 1 18 3 Miconia zamorensis 1_4 1 Ambisporaceae Ecuador 1,2 and NZL 12 Miconia zamorensis 4_3 1 Colacasia sp. 2_1 1 19 1 Axonopus compressus 1_6 1 Paraglomeraceae Ecuador and several locations worldwide (JPN, DEU 13 , USA). Correspond to CVTX00238 at MaarjAM database [ 20 ], 1 [ 21 ], 2 [ 42 ], 3 [ 59 ], 4 [ 49 ], 5 [ 35 ], 6 [ 31 ], 7 [ 3 ], 8 [ 2 ], 9 [ 13 ], 10 [ 9 ], 11 [ 16 ], 12 [ 48 ] 13 Sequences are indicated by the species name, followed by individual number and clone number. Freq Frequency. Locations correspond to Ch Chinapintza and LP La Pangui. Sequence provenances as give in Fig. 1 ab: BRA Brazil, CHN China, DEU Germany, ECU Ecuador, GBR United Kingdom, IND India, JPN Japan, MYS Malaysia, NZL New Zealand, SPN Spain, USA United States, VEN Venezuela, ZAF South Africa At La Pangui site 15 OTUs were detected, while at Chinapintza site 11 OTUs (Table 3 ). Seven OTUs were present at both sites, including all frequent ones. 3 to 8 OTUs per plant species and 1 to 5 OTUs per plant individual are present (Table 3 and Table 4 ). All plant species harbor Glomeraceae-OTUs and members of Archaeosporales with exception of Axonopus compressus , most of them also Acaulosporaceae (Table 3 ). The molecular analysis of AMF showed a species-rich community with 19 OTUs belonging to five different families: Glomeraceae, Acaulosporaceae, Archaeosporaceae, Ambisporaceae and Paraglomeraceae. Diversity of AMF fungi in heavy metal polluted mining areas The highest concentration of heavy metals was detected at Chinapintza, with average values of Pb, Zn and Cu approximately double the value detected at La Pangui (Table 2 ). Previous studies have shown that an increase of HMM concentration decreases AMF richness [ 19 , 60 – 62 ]. However, the observed difference in the number of OTUs between La Pangui and Chinapintza cannot attributed to differences in the concentration of heavy metals as the number of samples in both sites are not equivalent due to plant rarity. Within Glomeraceae, OTU 1 was the most frequent followed by OTU 6, both present in the nearby sites of Zamora-Chinchipe, but also present elsewhere in the world. Our results are similar to those of previous studies showing the dominance of Glomeraceae in soils contaminated with heavy metals [ 8 , 19 , 28 , 35 , 44 , 60 ]. In contrast, the most abundant AMF in several heavy metal contaminated soils Rhizophagus intraradices and Funneliformis mosseae [ 8 , 19 , 28 , 60 ] were not found in our contaminated sites. Further studies are needed to determine whether host plant identity or site characteristics, such as climate or soil, have a significant effect on the composition of AMF communities. In contrast to our results, in a recent study Faggioli et al. [ 14 ] found a rich AMF community dominated by members of Paraglomeraceae followed by Glomeraceae, in Pb-contaminated soils using an Illumina approach. In our study, together with the fact that 14 OTUs were previously found in other locations in Ecuador, it can be concluded that there are no specialists in heavy metal polluted sites, but generalists adapted to disturbed sites. However, the overall composition of the AMF-community of these heavy metal contaminated sites is similar to many other AMF-communities at family level: Glomeraceae dominates in terms of OTUs and frequencies, followed by Archaeosporaceae and Acaulosporaceae. Zamora-Chinchipe, a place where artisanal and small-scale gold mining is a deeply rooted activity, provoking contamination for several decades, was object of study to investigate the diversity and composition of AMF communities associated with the roots of six plant species sampled from heavy metal polluted soils located in Chinapintza and La Pangui sites. Although several members of Glomeromycota are consider as cosmopolitan species [ 51 ], its local distribution is affected by several factors [ 23 ]. Some AM fungal taxa have only been reported in the highly contaminated areas, which could represent ecotypes adapted to this extreme environment [ 61 , 62 ]. For phytoremediation of metal-contaminated soils, the use of indigenous fungi is recommended considering that they are adapted to particular abiotic and stressful conditions [ 41 ]. The potential role in phytoremediation approaches of the dominant fungi detected in our study remains to be investigated."
} | 4,097 |
33338066 | PMC7748156 | pmc | 3,040 | {
"abstract": "Collective behavior, and swarm formation in particular, has been studied from several perspectives within a large variety of fields, ranging from biology to physics. In this work, we apply Projective Simulation to model each individual as an artificial learning agent that interacts with its neighbors and surroundings in order to make decisions and learn from them. Within a reinforcement learning framework, we discuss one-dimensional learning scenarios where agents need to get to food resources to be rewarded. We observe how different types of collective motion emerge depending on the distance the agents need to travel to reach the resources. For instance, strongly aligned swarms emerge when the food source is placed far away from the region where agents are situated initially. In addition, we study the properties of the individual trajectories that occur within the different types of emergent collective dynamics. Agents trained to find distant resources exhibit individual trajectories that are in most cases best fit by composite correlated random walks with features that resemble Lévy walks. This composite motion emerges from the collective behavior developed under the specific foraging selection pressures. On the other hand, agents trained to reach nearby resources predominantly exhibit Brownian trajectories.",
"introduction": "1 Introduction Collective behavior is a common but intriguing phenomenon in nature. Species as diverse as locusts, and some families of fish or birds exhibit different types of collective motion in very different environments and situations. Although the general properties of swarms, schools and flocks have been widely studied (see e.g. [ 1 ] for a review), the emergence of global, coordinated motion from the individual actions is still a subject of study. Different approaches, ranging from statistical physics to agent-based models, have led to new insights and descriptions of the phenomenon. Statistical physics models are very successful at describing macroscopic properties such as phase transitions and metastable states [ 2 – 4 ], but in order to apply the powerful tools of statistical mechanics, these models normally simplify the individuals to particles that interact according to certain rules dictated by the physical model adopted, as for instance the Ising-type interaction of the spins in a lattice. A different type of models are the so-called self-propelled particle (SPP) models [ 5 – 8 ], which enable higher complexity in descriptions at the individual level but still allow one to employ the tools of statistical physics. They describe individuals as particles that move with a constant velocity and interact with other individuals via fixed sets of rules that are externally imposed. In SPP models, the description of the interactions is not restricted to physically accepted first principles, but can include ad hoc rules based on specific experimental observations. In this work, we follow a different approach and model the individuals as artificial learning agents. In particular, we apply Projective Simulation (PS) [ 9 ], which is a model of agency that can incorporate learning processes via a reinforcement learning mechanism. The individuals are thus described as PS agents that interact with their surroundings, make decisions accordingly and learn from them based on rewards provided by the environment. This framework allows for a more detailed, realistic description in terms of the perceptual apparatus of the agent. One of the main differences with respect to previous models is that the interaction rules between agents are not imposed or fixed in advance, but they emerge as the result of learning in a given task environment. This type of agent-based models that employ artificial intelligence to model behavior are gaining popularity in the last few years. Artificial neural networks (ANN) have been used, for instance, in the context of navigation behaviors [ 10 , 11 ] and reinforcement learning (RL) algorithms have been applied to model collective behavior in different scenarios, such as pedestrian movement [ 12 ] or flocking [ 13 , 14 ]. In contrast to other learning models such as neural networks, PS provides a transparent, explicit structure that can be analyzed and interpreted. This feature is particularly useful in modeling collective behavior, since we can study the individual decision making processes, what the agents learn and why they learn it. This way, we can directly address the questions of how and why particular individual interactions arise that in turn lead to collective behaviors. Initial work by Ried et al. [ 15 ], where the authors use PS to model the density-dependent swarm behavior of locusts, laid the foundations of the present work. Since the interaction rules are developed by the agents themselves, the challenge is to design the environment and learning task that will give rise to the individual and, consequently, collective behavior. In previous works, the agents are directly rewarded for aligning themselves with the surrounding agents [ 15 ] or for not losing neighbours [ 14 ]. Instead of rewarding a specific behavior, in this work we set a survival task that the agents need to fulfill in order to get the reward, and then analyze the emergent behavioral dynamics. As a starting hypothesis, we consider the need to forage as an evolutionary pressure and design a learning task that consists in finding a remote food source. Due to this particular survival task, our work relates to the investigation of foraging theories and optimal searching behavior. There is a vast number of studies devoted to the analysis of foraging strategies in different types of environments e.g., [ 16 – 19 ]. In the particular case of environments with sparsely distributed resources (e.g. patchy landscapes), there are two main candidates for the optimal search model: Lévy walks [ 20 – 22 ] and composite correlated random walks (CCRW) [ 23 , 24 ]. The former are described by a single distribution of step lengths that is characterized by a power-law p ( ℓ ) ∼ ℓ − μ with exponent 1 < μ ≤ 3, whereas the latter consider that the movement is composed of two different modes, characterized by two exponential distributions with different decay rates. Although the mathematical models behind them are fundamentally different, they have some common features that make the movement patterns hard to distinguish [ 24 – 28 ]. In broad terms, both models can produce trajectories that are a combination of short steps (with large turning angles in 2D), which are useful for exploring the patch area, and long, straight steps, which are efficient to travel the inter-patch distances. Even though both models have theoretical [ 22 , 23 ] and experimental (e.g. [ 29 , 30 ]) support, it is not yet clear if animal foraging patterns can be described and explained by such models or if they are too complex to admit such simplifications. Furthermore, regarding the Lévy walks, there is an ongoing debate on the question whether they emerge under certain animal foraging strategies. Currently there exist two main hypotheses, referred to as the evolutionary and the emergentist. The evolutionary hypothesis (also called Lévy flight foraging (LFF) hypothesis) states that certain species have evolved according to natural selection to develop an optimal foraging strategy consisting of Lévy walk movement patterns (see e.g. [ 31 ] and references therein). On the other side, the emergentist hypothesis argues that the LFF hypothesis is not sufficient to account for the complexity of animal behavior since it does not explain certain anomalies observed experimentally (see [ 32 ] and references therein). It argues that Lévy walks can emerge spontaneously as a consequence of the features of the environment, which lead to certain responses from the foraging organism. Thus, these responses are not part of an evolved strategy developed over the course of generations, but can arise from innate behaviors and lead to Lévy patterns spontaneously when the animal is confronted with certain environmental conditions. Due to the fact that our learning task is directly related to foraging strategies, we link the present work to the aforementioned studies by analyzing the individual trajectories the agents produce as a consequence of the behavior developed in the different learning contexts. The paper is organized as follows: an introduction to Projective Simulation and a detailed description of the model and the learning setup are given in Sec. 2. In Sec. 3, we present different learning tasks and analyze the resulting learned behaviors. In Sec. 4, we study the emergent group dynamics and individual trajectories within the framework of search models to determine if they can be described as Lévy walks or composite correlated random walks. Finally, we summarize the results and conclude in Sec. 5.",
"discussion": "5 Discussion We have studied the collective behavior of artificial learning agents, more precisely PS agents, that arises as they attempt to survive in foraging environments. More specifically, we design different foraging scenarios in one-dimensional worlds in which the resources are either near or far from the region where agents are initialized. This ansatz differs from existing work in that PS agents allow for a complex, realistic description of the sensory (percepts) and motor (actions) abilities of each individual. In particular, agents can distinguish how other agents within visual range are oriented and if the density of agents is high or low in the front and at the back of their visual area. Based on this information, agents can decide whether to continue moving in their current direction or to turn around and move in the opposite direction. Crucially, there are no fixed interaction rules, which is the main difference that sets our work apart from previous approaches, like the self-propelled particle (SPP) models or other models from statistical physics. Instead, the interactions emerge as a result of the learning process agents perform within a framework of reinforcement learning. The rewards given as part of this learning process play a role analogous to environmental pressures in nature, by enhancing the behaviors that led the agent to be rewarded. Therefore, by varying the task and reward scheme and studying the resulting behaviors, our approach allows us to test different causal explanations for specific observed behaviors, in the sense of environmental pressures proposed to have led to these behaviors. In this work, we have considered scenarios where the food is situated inside or far from the region where agents are initialized and we have observed that the initially identical agents develop very different individual responses —leading to different collective dynamics— depending on the distance they need to cover to reach the reward (food source). Agents learn to form strongly aligned swarms to get to distant food sources, whereas they learn to form cohesive (but weakly aligned) swarms when the distance to the food source is short. Since we model each individual as an artificial learning agent, we are able not only to study the collective properties that arise from the given tasks, but also to analyze the individual responses that agents learn and that, in turn, lead to the swarm formation. Thus, we observe for instance that the tendency to align with the neighbors in the d F = 21 case increases with the density of neighbors surrounding the agent. In the case of a training with d F = 4, we observe that the individuals tend to move to the region with higher number of neighbors, which leads to high cohesion at the collective level. We note that the task faced by our artificial agents, of reaching a food source, is closely related to the behaviors studied in the context of foraging theory. For this reason, we compare the individual trajectories that result from the learning process to the principal theoretical models in that field. We show that most of the individual trajectories resulting from the training with distant resources —which leads to strongly aligned swarms— are best fitted by composite correlated random walks consisting of two modes, one intensive and one extensive, whose mean step lengths are λ ^ I - 1 ≃ 2 . 7 and λ ^ E - 1 ≃ 75 , respectively. A smaller fraction of these trajectories is best fitted by power-law distributions with exponents μ ^ ≅ 1 . 6 , that is, Lévy walks. The exponent of the power-law distribution we obtain is close to 2, which is the optimal Lévy walk for maximizing the rate of target encounters in environments with sparsely distributed, renewable resources [ 22 , 31 , 68 ]. Moreover, our results are in agreement with the study of Reynolds [ 67 ] that shows that animals can approximate Lévy walks by adopting a composite correlated random walk. In contrast, agents that were trained to find nearby resources and follow the dynamics of cohesive swarms present normal-diffusive, Brownian-like trajectories that do not exhibit two movement modes but just one. One crucial point of this analysis is that our simulated agents move in a multi-agent context and their movement patterns are therefore determined by the swarm dynamics they have developed through the learning process. In particular, we provide a new perspective and additional insight on the studies mentioned above regarding Lévy walks and CCRW, since the individual trajectories that are best fit by these two models arise from a collective motion with very specific features such as strong alignment and decaying cohesion. This, together with the fact that the individual responses emerge as a result of the learning process, provides an example of how trajectories with features that resemble Lévy walks can emerge from individual mechanisms that are not generated by a Lévy walk process. In this sense, our work provides an unusual example to consider within the emergentist versus evolutionary debate on Lévy walks (see e.g. [ 31 , 32 ]). In particular, our work supports the former view point insofar as the agents do not have built-in interaction rules that come from a certain mathematical model such as that of the Lévy walk, but nevertheless exhibit trajectories that resemble Lévy walks as a result of the swarm dynamics that emerged from certain foraging environmental pressures. To conclude, we have applied a model of artificial agency (PS) to different foraging scenarios within the framework of collective motion. We have shown that, without any prior hard-wired interaction rules, the same agents develop different individual responses and collective interactions, depending on the distance they need to travel to reach a food source. Agents form strongly aligned swarms to stabilize their trajectories and reach distant resources, whereas they form cohesive, unaligned swarms when the resources are near. In addition, we have shown that Lévy-like trajectories can be obtained from individual responses that do not have a simple theoretical model as the underlying process, but instead are generated and arise from the interplay of a fine-grained set of learned individual responses and the swarm behavior that emerges from them at a collective level. This work provides a new framework for the study of collective behavior, which supports more detailed and realistic representations of individuals’ sensory and motor abilities and different types of environmental pressures. It would be interesting to apply this approach to the more complex collective behaviors that arise in two- and three-dimensional environments. Furthermore, the PS model allows for a variety of new scenarios to explore in the context of behavioral biology, since different reward schemes can easily be implemented and studied."
} | 3,954 |
40071206 | PMC11895702 | pmc | 3,042 | {
"abstract": "Microbial Electrochemical Technology (MET) offers a promising avenue for CO 2 utilization by leveraging the ability of chemolithotrophic microorganisms to use inorganic carbon in biosynthetic processes. By harnessing the power of electroactive bacteria, METs can facilitate the conversion of inorganic carbon into organic compounds. Therefore, this work combines biosurfactant production at the anode and PHB production at the cathode of Microbial Fuel Cells (MFCs), while testing the efficiency of Microbial Electrosynthesis Cells (MECs), and traditional culture in liquid media. This study employed a consortium of Pseudomonas aeruginosa PA1430/CO1 and Shewanella oneidensis MR-1, to provide reducing equivalents to Cupriavidus necator DSM428 for CO 2 fixation and polyhydroxybutyrate (PHB) production. Glycerol was used as a carbon source by the anode consortium to investigate biosurfactant production. Additionally, Adaptive Laboratory Evolution (ALE) was employed to enhance the efficiency of this process to develop biofilms capable of synthesizing PHB from CO 2 in MFCs under a controlled gas atmosphere (10% CO 2 , 10% O 2 , 2% H 2 , 78% N 2 ). Observed results showed a higher direct CO 2 removal from the gas mix in MECs (73%) than in MFCs (65%) compared to control cultures. Anionic (18.8 mg/L) and non-ionic (14.6 mg/L) surfactants were primarily present at the anodes of MFCs. Confocal microscope analysis revealed that the accumulation of PHBs in C. necator was significantly higher in MFCs (73% of cell volume) rather than in MECs (23%) and control cultures (40%). Further analyses on metabolites in the different systems are ongoing. Our data gave evidence that the anode consortium was able to provide enough electrons to sustain the chemolithotrophic growth of C. necator and the biosynthesis of PHBs at the cathode of MFCs, in a mechanism suggestive of the direct interspecies electron transfer (DIET), naturally occurring in natural environment.",
"conclusion": "Conclusion It is possible to produce PHBs and surfactants from, respectively, inorganic carbon and glycerol in MFCs inoculated with preformed, mature biofilms of C. necator at the cathode and Pseudomonas aeruginosa/Shewanella oneidensis consortium at the anode in a self-sustaining overall process, where a mechanism similar to direct electron transfer process established among the electrogenic consortium (at the anode) and electrotrophic microorganisms (at the cathode). Such a mechanism seems to establish when well mature biofilms, already accustomed to the nutritional conditions established in MFCs are present at the electrodes, providing also electrochemical stability to the system. The use of MFCs without preformed bioelectrodes as well as inoculated with strains not acclimatized to the conditions established in MFCs resulted in unstable systems. MECs were more efficient in removing CO 2 , from the gas mix although the amount of PHBs produced was lower than in MFCs. This result seems to indicate that, in MECs, the energy provided to the systems allowed the production of metabolites whose biosynthesis requires a higher amount of energy. Proper analyses of catholyte are needed to characterize the pool of metabolites produced by C. necator when grown at −955 mV cathode potential. A dedicated measurement system will be used to improve the monitoring configuration of the prototypes as future improvement as well. The preliminary results from confocal microscopy confirmed the trend obtained by the fluorescence measurements, but a quantitative analysis of PHBs produced in each system and the amount of energy spent by each system for their biosynthesis is needed, as well as the identity of the produced biosurfactants. In any case, our research showed that combining biosurfactants and PHB biosynthesis in MFCs is possible. The metabolic analyses on C. necator showed a prolonged metabolic activity as well as an increased ability to use molecules containing an aromatic ring such as p -hydroxybenzoic acid (PHBA), polysorbates (Tween 40) and aromatic amines, besides of aminoacids and polymers, as observed in a similar experimental activity in C. saccharoperbutylacetonicum NT-1. Further investigations, even at genomic level, are needed in order to evaluate the outcomes of the ALE used and the meaning of the metabolic changes observed in C. necator , the increased production of PHBs at the cathode and investigate the metabolic profiles of P. aeruginosa and S. oneidensis to further develop MFCs for CO 2 capture and glycerol fermentation. Overall, this study presents a novel approach to simultaneously achieve CO 2 capture, PHB production, and biosurfactant synthesis in MFCs, advancing the field of MET. The results demonstrate the enhanced metabolic activity of C. necator , with an increased ability to utilize aromatic compounds such as p-hydroxybenzoic acid and polysorbates. These findings offer a promising pathway for sustainable bioproduction. Further investigations, including genomic and metabolic profiling, are necessary to optimize MFCs for CO 2 capture and glycerol fermentation in industrial applications.",
"introduction": "Introduction The need for new green processes for the biosynthesis of value-added compounds has driven scientific research toward electrosynthesis and other bioprocesses. Among them, the bioelectrochemical system (BES) is a hybrid technology that combines microbiology and electrochemistry, providing a sustainable and efficient approach to address the pressing issues of climate change and environmental pollution ( Ramírez-Vargas et al., 2018 ; Nastro et al., 2023 ). BES offers a promising approach to transforming CO 2 into value-added multi-carbon compounds. BES can reduce CO 2 electrochemically by harnessing the power of electrotrophic microorganisms providing the required energy input to drive CO 2 fixation in the biocathode ( Cabrera et al., 2021 ; Naderi et al., 2023 ; Temirbekova et al., 2023 ; Wu et al., 2023 ). This process, often called Microbial Electrosynthesis (MES), enables the biosynthesis of several organic compounds, (such as ethanol, acetate, formate, propionate, and isopropanol) ( Kronenberg et al., 2017 ; Alvarez Chavez et al., 2022 ) by applying an external electric potential (<2.5 volts). The energy provided is the primary mechanism to reduce the thermodynamic stability of CO 2 , via pathways like the Wood-Ljungdahl pathway in Clostridium spp. ( Yang et al., 2022 ). In recent years, MES has developed as a promising technique for the biosynthesis of polyhydroxyalkanoates (PHAs), which are polyesters classified as follows by their carbon side chain length: short-chain (3–5 carbons), medium-chain (6–14 carbons), and long-chain (14–25 carbons). Their general formula is [—O—CRH—CH 2 —CO—]n, where R denotes the side chain (e.g., methyl group for 3-hydroxybutyric acid (HB) or ethyl group for 3-hydroxyvaleric acid (HV)). Poly(3-hydroxybutyrate) (P(3HB)) is a prominent example of a short-chain PHA ( Acharjee et al., 2022 ; Fu et al., 2023 ). The interest in PHAs in general, and PHBs in particular, lies in their potential industrial applications as a replacement for fossil fuel-based materials ( Ranaivoarisoa et al., 2019 ). These bio-based polyesters are widely used in bioplastics, chemical additives, medicine, agriculture, wastewater treatment, and cosmetics. Besides CO 2 , PHAs can also be produced from carbon-rich waste substrates ( Javaid et al., 2020 ; Shin et al., 2021 ). From a biochemical perspective, PHAs are energy-storage molecules synthesized by various microorganisms under stress conditions (limited essential nutrients) with excess carbon sources ( Saravanan et al., 2022 ). The PAHs are a convincing alternative to synthetic polymers, demonstrating adequate mechanical properties similar to polypropylene and being thermoformable into diverse bio-based products, which directly reduce the detrimental impact of humans on the environment. The Cupriavidus necator H16 is a well-established microorganism that produces PHBs at an industrial scale ( Al Rowaihi et al., 2018 ; de Melo et al., 2023 ). Nevertheless, the high production costs, 4–10 times greater than those of fossil-fuel-derived polymers ( Saravanan et al., 2022 ; de Melo et al., 2023 ), hinder its widespread adoption. The 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, and BESs have the potential to fulfill both requirements and contribute to taking PHAs on the market in the future ( Alvarez Chavez et al., 2022 ). Since the pioneering work of Srikanth et al. (2012) demonstrated PHA biosynthesis at the cathode of Microbial Electrosynthesis Cells (MECs) using a consortium of microorganisms, research on electrochemically-driven PHA production has gained traction. Recent studies ( Alkotaini et al., 2018 ; Lai and Lan, 2021 ; Shin et al., 2021 ; Pillot et al., 2022 ) various strategies regarding biosynthesis of PHAs in BES, including syngas utilization by Cupriavidus necator ( Langsdorf et al., 2024 ), highlighted the potential of BESs for CO 2 capture and PHAs production. Parallelly, BESs demonstrate promise in producing platform chemicals like biosurfactants at the anode ( Pasternak et al., 2023 ; de Rosset et al., 2024 ). Biosurfactants are amphiphilic surface-active molecules due to their hydrophobic and hydrophilic moieties and they find applications in oil processing, the pharmaceutical sector (moisturizers, creams, and medicines), food (as emulsifiers), medical (antimicrobial agents), agricultural (fertilizers), and civil (waste and sewage treatment) industries. ( Nikolova and Gutierrez, 2021 ; Sarubbo et al., 2022 ; Vigil et al., 2024 ). Among the substrates that can be used to produce biosurfactants, glycerol, a significant byproduct of biodiesel production, is a desirable substrate ( Monteiro et al., 2018 ; Liu et al., 2022 ). Its utilization in MES for biosurfactant production (e.g., rhamnolipids from Pseudomonas aeruginosa ( Bezerra et al., 2019 ; Zhao et al., 2021 )) offers economic and environmental benefits by adding value to a waste product. Glycerol, a byproduct that can constitute up to 10% of total biodiesel production, necessitates well-organized management to lessen the overall cost of the biodiesel production process. This byproduct has several uses in different industries, such as chemical, textile, pharmaceutical, and food sectors. Additionally, it can be utilized to produce value-added compounds such as ethanol, hydrogen, propanoic acid, butanol, citric acid, and polyunsaturated fatty acids, primarily through microbial metabolism ( Chilakamarry et al., 2021 ), and, as previously said, biosurfactants. The ability of Pseudomonas aeruginosa to produce rhamnolipids, a class of glycolipids, as metabolites of glycerol degradation in both aerobic and anaerobic conditions and the possible use of crude glycerol as a substrate for rhamnolipids production is very attractive from an industrial perspective ( Bezerra et al., 2019 ; Zhao et al., 2021 ). Furthermore, recent studies have identified biosurfactant production, specifically glycolipids, in environmental isolates of the genus Shewanella . Nevertheless, the metabolic pathway for glycolipids biosynthesis in Shewanella remains to be clarified ( Joe et al., 2019 ; Ram et al., 2019 ; Hassanshahian et al., 2020 ). Several approaches, aiming at emulating the natural environment and gaining access to the untapped natural resources emerging from crosstalk between partners, have been widely employed in biomanufacturing to produce pharmaceuticals, nutraceuticals, food, and drinks on a large scale and play a prominent role in the bioremediation and bioenergy sectors ( Kapoore et al., 2022 ). Co-culturing experiments involving the direct mixing of microorganisms have been proven to enhance microbial functions and accomplish tasks that are difficult to achieve with monocultures. Instead, exopolysaccharides (EPS) immobilize cells in biofilms by providing mechanical stability and keeping them close. As a result, in biofilm, an inter-connected cohesive three-dimensional polymer network is realized, where crosstalk between cells forms synergistic micro-consortia ( Kapoore et al., 2022 ), in our case, a dual-species consortium. Starting from this assumption, we cocultured, since the very beginning of our experiment, P. aeruginosa and S. oneidensis firstly in liquid media and then in biofilms, to possibly enhance their interactions with potential advantages in terms of electron transfer to the anode and glycerol fermentation. The integration of CO₂ capture at the cathode with surfactant biosynthesis at the anode in BESs leads to several advantages, including high operational efficiency, reduced production costs, lower energy consumption, and mitigation of environmental impacts ( Jourdin et al., 2018 ; Gong et al., 2020 ; Ayol et al., 2021 ; Li and An, 2022 ). Hence, this integration approach provides a sustainable solution to climate change and pollution challenges ( Hao et al., 2022 ). Electrosynthesis in MECs involves applying an electrochemical potential to control microbial metabolism or sustain abiotic electrochemical processes ( Dattatraya Saratale et al., 2022 ). Alternatively, MFCs can generate reducing equivalents via electrogenic bacteria at the anode to support chemolithotrophic growth at the cathode. These systems utilize CO₂ with H₂ (or electrons and protons) to produce metabolites, providing a stable dynamic equilibrium amid the anode and cathode compartments ( Nastro et al., 2023 ). Unlike what happens in MECs, MFCs work properly if the sum of the Gibbs Energy of the bioelectrochemical processes occurring at the anode and the cathode is negative [ΔG tot = (ΔG anode + ΔG cathode) <0], leading to a self-sustaining system ( Abdulwahhab et al., 2023 ; Nastro et al., 2023 ). Direct Interspecies Electron Transfer (DIET) is a key mechanism enhancing electron flow among microbial populations. This process, mediated by conductive materials (e.g., biochar), conductive pili, cytochromes, or molecular shuttles, enables the syntrophic growth of exoelectrogenic and electrotrophic bacteria, as seen in Geobacter metallireducens and Methanosarcina barkeri ( Holmes et al., 2018 ; Chen et al., 2022 ; Valentin et al., 2023 ). In C. necator , conductive pili may enhance DIET efficiency by increasing cell-electrode contact distances and enabling more cells to participate in electron transfer. Their expression can significantly improve MFC performance by optimizing metabolic activity and power generation, paving the way for advanced bioelectronic materials and integration of such systems into bioelectronics applications. Starting from this information, we set up MFCs using a consortium of highly electroactive bacteria to verify whether such bacteria, once cocultured for several hundreds of generations, might be able to satisfy the demand of electrons of C. necator DM428 and, therefore, sustain its chemolithotrophic growth. C. necator utilizes the Calvin-Benson-Bassham Cycle for the assimilative reduction of CO 2 , facilitating the biosynthesis of glyceraldehyde-3-phosphate. In this process, H₂ molecules serve as electron donors for the O 2 -tolerant [NiFe]-hydrogenases, while O₂ acts as the terminal electron acceptor ( Teramoto et al., 2022 ). We hypothesized that in MFCs, in the absence of an external source of electrons, a mechanism like DIET might be established between the anode consortium and C. necator , thus fostering the syntrophic growth of bacteria at the anode and cathode, with concurrent biosynthesis of biosurfactants (at the anode) and CO 2 capture at the cathode, with concurrent synthesis of potential added-value compounds. To test this, in this study, we aimed to investigate PHB production in C. necator DSM428 cells by applying different conditions: the presence of an external potential of −955 mV ( v s Ag/AgCl reference electrode in MECs), of the electrons provided by P. aeruginosa/S. oneidensis consortium in MFCs, without electrogenesis in microbiological cultures in liquid medium. We drastically reduced the start-up period of BESs by using already mature biofilms at both anodes and cathodes, and microorganisms progressively grown in oligotrophic conditions. Once we set up the first BESs and control cultures, we repeatedly used strains isolated from anolytes and catholytes of BESs at the end of a cycle to start a new experimental session according to an Adaptive Laboratory Evolution (ALE) approach. ALE is a technique for the selection of strains with better phenotypes by long-term culture under a specific selection pressure or growth environment ( Wang et al., 2023 ). Adaptive laboratory evolution (ALE) is the process based on which the principles of natural evolution are implemented in the laboratory to a specific population under controlled conditions. In ALE, natural selection is directed toward a selected environment with the desired conditions for microbial populations to obtain better fitness to improve one or more metabolic abilities of a given microorganism. Further, ALE is used as a tool for a deeper understanding of the genetic and/or metabolic pathways of evolution and, also, to improve the biosynthesis of products of (high) added value, such as ethanol, butanol, and lipids ( Mavrommati et al., 2022 ). Following this approach, we also investigated the metabolic profile and intracellular PHB content in C. necator when used for several thousands of generations in MFCs. The obtained profile was compared with the original strain purchased at DSMZ, which had never grown in chemolithotrophic conditions. To our knowledge, this is the first time such an experimental scheme has been applied to C. necator ."
} | 4,489 |
35432228 | PMC9010870 | pmc | 3,043 | {
"abstract": "Methane, a potent greenhouse gas produced in freshwater ecosystems, can be used by methane-oxidizing bacteria (MOB) and can therefore subsidize the pelagic food web with energy and carbon. Consortia of MOB and photoautotrophs have been described in aquatic ecosystems and MOB can benefit from photoautotrophs which produce oxygen, thereby enhancing CH 4 oxidation. Methane oxidation can account for accumulation of inorganic carbon (i.e., CO 2 ) and the release of exometabolites that may both be important factors influencing the structure of phytoplankton communities. The consortium of MOB and phototroph has been mainly studied for methane-removing biotechnologies, but there is still little information on the role of these interactions in freshwater ecosystems especially in the context of cyanobacterial growth and bloom development. We hypothesized that MOB could be an alternative C source to support cyanobacterial growth in freshwater systems. We detected low δ 13 C values in cyanobacterial blooms (the lowest detected value −59.97‰ for Planktothrix rubescens ) what could be the result of the use of methane-derived carbon by cyanobacteria and/or MOB attached to their cells. We further proved the presence of metabolically active MOB on cyanobacterial filaments using the fluorescein isothiocyanate (FITC) based activity assay. The PCR results also proved the presence of the pmoA gene in several non-axenic cultures of cyanobacteria. Finally, experiments comprising the co-culture of the cyanobacterium Aphanizomenon gracile with the methanotroph Methylosinus sporium proved that cyanobacterial growth was significantly improved in the presence of MOB, presumably through utilizing CO 2 released by MOB. On the other hand, 13 C-CH 4 labeled incubations showed the uptake and assimilation of MOB-derived metabolites by the cyanobacterium. We also observed a higher growth of MOB in the presence of cyanobacteria under a higher irradiance regime, then when grown alone, underpinning the bidirectional influence with as of yet unknown environmental consequences.",
"introduction": "Introduction Freshwater ecosystems are estimated to be among the largest natural sources of atmospheric methane ( Kirschke et al., 2013 ; Saunois et al., 2020 ; Rosentreter et al., 2021 ), a potent greenhouse gas of which atmospheric concentrations may increase due to feedback mechanisms as the result of global warming ( Marotta et al., 2014 ). However, diffusive lake methane fluxes are mostly (30–99%, Bastviken et al., 2008 ) mitigated by aerobic methane-oxidizing bacteria (MOB), mostly belonging to Alpha- and Gammaproteobacteria ( Bodelier et al., 2019 ) using methane for energy generation and cellular carbon. In this way, biogenic CH 4 can subsidize the pelagic food web as an alternative energy and carbon source ( Bastviken et al., 2003 ; Deines et al., 2007 ; Jones and Grey, 2011 ; Agasild et al., 2014 ) by predation of MOB by protozoa and metazoa, who transfer it into the pelagic food web ( Bastviken et al., 2003 ; Kankaala et al., 2007 ; Schilder et al., 2017 ). Aerobic methanotrophs can also thrive under oxygen-deficient conditions, which are common in eutrophic and hypertrophic stratified lakes ( Yang et al., 2019 ; Mayr et al., 2020 ). This can be explained by their versatility in using other electron acceptors, e.g., nitrite/nitrate, humic acids, or ferric ions ( Kits et al., 2015 ; Oswald et al., 2017 ; Naqvi et al., 2018 ; van Grinsven et al., 2020 , 2021 ). However, the near complete consumption of methane was observed in anoxic layers of a stratified lake when algae could perform photosynthesis and supply MOB with oxygen ( Milucka et al., 2015 ; Oswald et al., 2015 ). A similar phenomenon was described in the leaves and stems of aquatic plants that supported the methanotrophic activity ( Yoshida et al., 2014 ). Such cohabitation of MOB and photosynthetically active organisms has already been investigated for potential use in effective methane-removing biotechnologies, in which algae provide MOB with oxygen and MOB produce CO 2 in return ( van der Ha et al., 2011 ; Badr et al., 2019 ). However, this exchange of substrates between MOB and photosynthetic phytoplankton has mostly been demonstrated in reactor/biotechnological settings. An in situ example of this interkingdom methane-derived carbon exchange has been described in peat bogs, where methanotrophs were present in hyaline cells and on stems and leaves of Sphagnum and rapidly oxidized methane to CO 2 , which constituted a significant source of carbon (10–15%) for the peat moss ( Raghoebarsing et al., 2005 ). Organic compounds released by methanotrophs such as methanol, formate, acetate, and other metabolites can potentially support a broad range of microbes ( Chistoserdova and Kalyuzhnaya, 2018 ). In this respect, cyanobacteria can grow mixotrophically ( Schmetterer and Flores, 1988 ; Stebegg et al., 2019 ), while genes related to methane oxidation have been detected in the “cyanosphere” of two blooming cyanobacteria ( Pascault et al., 2021 ), indicating the possibility of metabolite exchange between these bacterial guilds. Carbon dioxide is an important factor influencing the structure of phytoplankton communities ( Shapiro, 1997 ). Primary production increases with trophy ( Peters, 1986 ), and thus eutrophic and hypertrophic waters may be undersaturated with CO 2 due to high phytoplankton productivity ( Finlay et al., 2009 ; Lazzarino et al., 2009 ). Dense blooms often deplete dissolved CO 2 below the atmospheric equilibrium ( Talling, 1976 ; Maberly, 2008 ; Balmer and Downing, 2011 ). However, in deep, eutrophic and stratified lakes, the meta- and hypolimnion can be rich in CO 2 and HCO 3 as the result of biological decomposition and chemical reactions ( Heaney et al., 1986 ). The oxidation of methane may account for a high proportion of excess inorganic carbon accumulation in the hypolimnion of stratified lakes ( Houser et al., 2003 ), turning methane-derived carbon into a more relevant carbon source than photosynthetically produced carbon under more eutrophic conditions ( Schilder et al., 2017 ). However, a quantitative prediction of the feedback between phytoplankton growth, methane oxidation, and CO 2 drawdown in aquatic ecosystems has garnered little attention. This is surprising since CO 2 may perform a crucial role in the competition among phytoplankton species, including harmful cyanobacteria that threaten the water quality of many eutrophic and hypertrophic lakes and cause severe ecological and economic damage worldwide. Cyanobacteria with high-flux bicarbonate uptake systems can benefit from elevated CO 2 levels ( Ji et al., 2017 ). Although the light conditions in the meta- and hypolimnion can be poor due to high primary production, cyanobacteria are adapted to them. For example, Aphanizomenon has a competitive advantage in light-limited conditions due to its affinity for light ( De Nobel et al., 1998 ). Low light intensity and high nutrient concentrations, in combination with CO 2 provided by methane oxidation in the hypo- and metalimnion, may provide cyanobacteria with an advantage over other phytoplankton, e.g., green algae ( Shapiro, 1997 ). Moreover, some cyanobacteria can regulate their vertical distribution (buoyancy) and move up and down the water column to seek optimal light and nutrient conditions ( Walsby and Booker, 1980 ; Walsby et al., 2004 ; Carey et al., 2012 ). In this way, cyanobacteria can gain extra benefits by positioning themselves closer to the chemocline, where in stratified lakes, the majority of methane is oxidized by MOB ( Schubert et al., 2010 ). Biogenic methane has the lowest isotopic carbon ratio compared to other natural sources because it is extremely depleted in 13 C ( Whiticar, 1999 ). Organisms assimilating this 13 C-depleted carbon directly (e.g., MOB) or indirectly (e.g., phyto- or zooplankton) will have more negative δ 13 C values of their cellular carbon compared to other components of the food chain. Cyanobacteria exhibit variable δ 13 C values, which are quite often below −30.6‰ ( Vuorio et al., 2006 ; Agasild et al., 2019 ). These low δ 13 C values are much lower than isotopic signatures usually found in inorganic carbon sources ( Rinta et al., 2015 ) that can be utilized during photosynthesis. Previously reported CH 4 oxidation linked to photosynthetic activity and overlapping niches motivated us to explore the possibility of associations occurring between MOB and cyanobacteria. We suspected that there are associations between cyanobacteria and MOB, which would result in the low observed δ 13 C values, and the latter could be an alternative carbon source supporting cyanobacteria growth. To test this hypothesis, we analyzed filamentous cyanobacteria collected in the field, for the presence of metabolically active MOB dwelling in the cyanosphere or attached to cyanobacteria. We also tested several cyanobacterial laboratory strains for the presence of particulate methane monooxygenase gene pmoA . We also hypothesized that carbon derived from CH 4 oxidation by MOB subsidizes photoautotrophs, such as cyanobacteria in stratified lakes. We tested this hypothesis under laboratory conditions, first testing the possibility of cyanobacteria growth in the presence of methane and methanotrophic bacteria without external input of CO 2 , and second, testing the 13 C transfer from CH 4 to cyanobacteria. Since we had no access to axenic cultures of cyanobacteria and there are known examples of methanogens and MOB associations with phytoplankton ( Grossart et al., 2011 ; Mulhollem et al., 2016 ; Samad et al., 2020 ; Li et al., 2021 ), we decided to test the possibility of carbon transfer via MOB metabolites utilized by other organisms that later release CO 2 or organic substances. We also tested the influence of light conditions on the overall performance of the consortium. There are known examples from aquatic ecosystems showing that light intensities as low as 4.1 μmol photons m –2 s –1 can significantly decrease methane oxidation ( Murase and Sugimoto, 2005 ), and laboratory experiments showed 90% growth inhibition for Methylosinus and enriched cultures of natural methanotroph communities ( Dumestre et al., 1999 ). We expected that stronger light conditions would inhibit the growth of light-sensitive MOB and cyanobacteria and reduce the overall performance of both organisms (e.g., CH 4 oxidation and cyanobacteria production), while low light intensity would enhance the overall performance of both organisms. Here, in this study, we used a model system of the filamentous cyanobacterium Aphanizomenon gracile (strain SAG 31.79), common to eutrophic lakes, and the methane oxidizing bacterium Methylosinus sporium (NIOO collection, strain L17-3), isolated from a freshwater lake.",
"discussion": "Discussion Cyanobacteria—Methane-Oxidizing Bacteria Associations Meta- and hypolimnion in eutrophic lakes can be rich in CH 4 , produced mainly by methanogenic Archaea in the sediments among other CH 4 lacustrine sources ( Günthel et al., 2019 ). The archaeal methanogenic pathway discriminates against 13 C ( Whiticar, 1999 ), thus organisms assimilating this 13 C-depleted carbon directly (MOB) or indirectly (phytoplankton) should have more negative δ 13 C values of their cellular carbon compared to other components of the food chain. The oxidation of methane may account for a high proportion of excess inorganic carbon accumulation in the hypolimnion of stratified lakes ( Houser et al., 2003 ). Additionally, MOB produce organic metabolites that can potentially support a broad range of microbes ( Chistoserdova and Kalyuzhnaya, 2018 ). Cyanobacteria are frequently found in the oxic-anoxic zone where the majority of CH 4 is oxidized to CO 2 and with their high-flux bicarbonate uptake systems they can benefit from elevated CO 2 levels ( Ji et al., 2017 ) or even metabolites released by MOB. We hypothesized that there are direct associations between cyanobacteria with MOB and/or that cyanobacteria utilize CO 2 and metabolites from oxidized CH 4 which would result in the low observed δ 13 C values. Using different yet supplementary approaches, we showed that MOB can be physically attached to and stimulate the growth of cyanobacteria via metabolic links. Thus, we tested phytoplankton collected in the field using FTCP labeling coupled with FACS, and we were able to distinguish active MOB cells that were either single (free-living) or in association with algae and cyanobacteria ( Figure 2 ). Field reports have described the presence of MOB attached to phytoplankton using different techniques. For instance, alpha-proteobacterial MOB were detected using fluorescence in situ hybridization (FISH) in an algal cell ( Milucka et al., 2015 ) or mRNA reads related to methane metabolism in the cyanosphere of bloom-forming cyanobacteria ( Pascault et al., 2021 ). With respect to the technique applied to the study of the active MOB, it should be noted that FTCP could also label active nitrifiers, as this fluorophore can react with both ammonia monooxygenase and methane monooxygenase ( McTavish et al., 1993 ). On the other hand, we were able to amplify the pmoA gene in several cyanobacterial isolates cultured under laboratory conditions. We could even detect pmoA sequences associated with non-axenic cyanobacterial cultures that have been cultivated under laboratory conditions for a long time. This suggests that methanotrophs survive and even propagate in cultures where no methane is added. However, cyanobacteria can produce small amounts of methane under oxic conditions, which is very likely connected to their photosynthetic activity ( Bižić et al., 2020 ). Even though the amount of methane produced may be low, both Methylocystis spp. ( Knief and Dunfield, 2005 ; Cai et al., 2016 ) and Methylocapsa spp. ( Tveit et al., 2019 ) can thrive well under these conditions. Hence, the Methylocystis -type sequences we obtained from cyanobacterial isolates ( Supplementary Table 4 ) match this idea of oligotrophic strains being able to survive in the cyanosphere. Additionally, other methanotrophs may be able to survive in the cyanosphere under natural conditions, where large amounts of methane may be formed during cyanobacterial blooms ( Li et al., 2021 ). Another option for supporting MOB growth and activity may be hydrogen gas that is generated during nitrogen fixation by cyanobacteria ( Lopes Pinto et al., 2002 ; Dutta et al., 2005 ). Some methanotrophs, such as Methylocystis possess hydrogenases, allowing them to generate energy from H 2 oxidation ( Hakobyan et al., 2020 ). Initially, we conducted isotopic analyses of field samples containing phytoplankton and its selected members. Most of the δ 13 C values were in agreement with previous literature (e.g., Vuorio et al., 2006 ), but some taxa expressed lower isotopic ratios than expected, i.e., P. rubescens and Dinobryon sp. ( Figure 1 ). Their low δ 13 C values could not originate only from DIC or CO 2 , as they were more positive (minimum values noted in Budzisławskie Lake were −5.8 and −16.4‰, respectively). High isotopic values for inorganic carbon sources may be the result of acetotrophic methanogenesis, where δ 13 C values of CO 2 can be 30‰ higher than those in acetate used as a substrate in methane production ( Conrad, 2005 ; Steinmann et al., 2008 ). The CO 2 originating from acetotrophic methanogenesis does not exclude methane-derived inorganic carbon sources for phytoplankton, as observed in our experiments, but adds another source into the pool of inorganic carbon, and as a result, the mixture of different sources of CO 2 has higher δ 13 C values. However, the differences in δ 13 C values of Dinobryon could be explained by the fact that they are mixotrophs and can feed on microorganisms, including MOB. For example, phototrophic flagellates can graze up to 79% of the total bacteria consumed ( Sanders et al., 1989 ), and Dinobryon can selectively graze on Archaea ( Ballen-Segura et al., 2017 ), which discriminate against 13 C during methanogenesis. Cyanobacteria are also known for mixotrophy, but they can only incorporate dissolved organic carbon sources ( Stebegg et al., 2019 ). This led us to think that there was another carbon source influencing the δ 13 C values of P. rubescens and possibly other phytoplankton groups. The qPCR revealed presence of high copy numbers of pmoA for Type II methanotrophs in Licheńskie and Łódzko-Dymaczewskie lakes ( Supplementary Figure 2 ). These bacteria are capable of oxidizing dissolved CH 4 , and can add to the pool of CO 2 available for cyanobacteria or release organic metabolites available for other microbes or mixotrophs. Our data do not allow for assigning the types of MOB attached to cells or designating the dissolved CO 2 as the main source of carbon responsible for low δ 13 C values detected in the field samples. Hence, future work should determine which MOB species are active in the cyanosphere, what is their contribution to CH 4 oxidation and carbon transfer to cyanobacteria, and which metabolic pathways are active in the cyanosphere dwelling MOB as compared to their free-living counterparts. It is also still not known what type of molecules are interchanged between MOB and cyanobacteria and how this interchange is regulated by environmental factors such as light or temperature. Further quantitative studies are required to assess the importance of different paths of carbon and energy flow, e.g., the uptake of metabolites by cyanobacteria from MOB. Cyanobacteria—Methane-Oxidizing Bacteria Interactions To further investigate whether MOB can promote the growth of cyanobacteria, we conducted two laboratory experiments in which filamentous A. gracile grew in the presence of CH 4 and M. sporium without external input of CO 2 , followed by 13 C transfer from labeled CH 4 to cyanobacteria. The results demonstrated that the methane-oxidizing bacterium Methylosinus subsidized Aphanizomenon with carbon. Moreover, there are two possible pathways of carbon transfer from MOB to cyanobacteria. One option is via direct transfer of CO 2 produced during methane respiration by MOB, as shown in the first experiment. In these incubations where the cyanobacterium grew alone, the carbon dioxide concentrations decreased significantly, and the methane concentration was not affected ( Figure 3 ). However, the cyanobacterium had the highest yield in the presence of Methylosinus , and there was a significant methane decrease with no CO 2 limitation ( Figure 3 ). The second possible carbon pathway is the direct uptake of other released metabolites coming from the methanotrophs ( Strong et al., 2015 ; Gilman et al., 2017 ) by the cyanobacterium or the possibility of carbon transfer via MOB metabolites utilized by other organisms attached to the non-axenic Aphanizomenon , which later release CO 2 or organic substances. When exploring these pathways in the second experiment, we found that the Aphanizomenon monocultures incubated with only labeled methane were enriched in 13 C derived from methane at roughly half of the amount in the mixed culture. This can be explained by the fact that MOB are present in the cyanobacterial cultures, as we demonstrated by pmoA PCR, including the strain SAG 31.71 used in our experiments. Additionally, the highest 13 C incorporation was found in Aphanizomenon alone and without CH 4 when it was supplied with spent filtered medium from Methylosinus containing no 13 CO 2 but only labeled soluble exudates of the methanotrophs ( Figure 6 ). The most plausible explanation is the presence of different bacteria feeding on labeled metabolites from Methylosinus , and producing labeled CO 2 and/or Aphanizomenon gracile was able to assimilate organic carbon released by MOB. We have no data on the capability of A. gracile to assimilate organic carbon, but there are examples of it in other cyanobacteria, such as Nostoc sp., Anabaena sp., and Synechococcus sp. ( Rippka et al., 1979 ; Schmetterer and Flores, 1988 ; Stuart et al., 2016 ; Stebegg et al., 2019 ). The influence of MOB cannot be excluded, but it was rather limited since in the second stage of the experiment with metabolites there was no CH 4 in the atmosphere and thus no substrate for growth. Although cyanobacteria are capable of producing methane, e.g., Nodularia spumigena , and a phosphonate-degrading gene cluster was found in 28 sequenced cyanobacterial strains isolated from the Baltic Sea, no such genes were found in the Aphanizomenon genus ( Teikari et al., 2018 ). Light Influence on Growth of Aphanizomenon , Methylosinus and Methane Consumption Both organisms had higher yield when grown together, but this coexistence depended on light conditions. Contrary to our expectations, the consortium grown in high light intensity conditions had the highest CH 4 consumption and CO 2 and biomass production. Methylosinus alone grew better at low light than at high light intensity ( Figure 4 ), and the whole consortium had a significantly lower yield at low light ( Figures 4 , 5 ). Moreover, the yield of Aphanizomenon in consortium under high light conditions was higher than that under low light intensity, and the yield of Methylosinus in consortium was similar to that in the treatment where Methylosinus grew alone under low light intensity conditions. Surprisingly, methane consumption in consortium treatment was similar at both light levels when compared to Methylosinus alone ( Figure 3C ). This suggests that the shading effect (i.e., self-shading by increasing biomass of organisms in a culture) had a role in it, and allowed for a better growth of Aphanizomenon and more efficient methane consumption by Methylosinus in high light intensity. The high light intensity condition in our experiment reached 105 μmol s –1 m –2 PAR, which is far above the known inhibiting light levels ( Murase and Sugimoto, 2005 ), and it did not suppress CH 4 consumption in the consortium nor in the Methylosinus alone treatment. However, when the Methylosinus was cultured alone the number of DNA copies was significantly lower under high light intensity when compared to low light intensity, which may suggest a change in the apparent cell-specific activity, i.e., increased metabolic activity but decreased abundance probably due to higher costs of growth in strong light stress. This response is different to known examples from aquatic ecosystems showing that light intensities as low as 4.1 μmol photons m –2 s –1 can significantly decrease methane oxidation ( Murase and Sugimoto, 2005 ), and laboratory experiments showed 90% inhibition for Methylosinus and enriched cultures of natural methanotroph communities ( Dumestre et al., 1999 ). Another example of mutual support is the exchange of “oxygen for methane”—when both organisms are present, methane is oxidized by the methanotroph, which produces excess CO 2 , thereby avoiding carbon limitation of Aphanizomenon , and the full consortium had the highest yield as a result. Previous work ( van der Ha et al., 2011 ; Milucka et al., 2015 ; Oswald et al., 2015 ) showed that methanotrophs can benefit from oxygen produced during photosynthesis, and Raghoebarsing et al. (2005) showed that methane-derived CO 2 provides 10–15% of carbon for photosynthesis in Sphagnum . However, our experiment started in a normal atmosphere enriched in CH 4 , with a surplus of oxygen. Thus, we cannot be certain whether Aphanizomenon supported or performed an important role in supplementing Methylosinus with oxygen. Synthesis Species composition of phytoplankton depends on the magnitude of change in CO 2 ( Low-Décarie et al., 2011 ), and different requirements of various phytoplankton taxa for carbon influence interspecies competition. Dense blooms often deplete the dissolved CO 2 in surface waters ( Finlay et al., 2009 ; Lazzarino et al., 2009 ), however thermal stratification entraps methane and predicted climate warming will intensify this process. The methane is oxidized by MOB and may account for a high proportion of excess inorganic carbon accumulation in the hypolimnion of stratified lakes ( Houser et al., 2003 ). Additionally, MOB produce organic metabolites that can potentially support a broad range of microbes ( Chistoserdova and Kalyuzhnaya, 2018 ). The majority of CH 4 is oxidized in chemocline, where cyanobacteria are frequently found. Thus, in eutrophic and hypertrophic lakes low light intensity and high nutrient concentrations, in combination with CO 2 provided by methane oxidation in hypo- and metalimnion may provide buoyant cyanobacteria with an advantage over other phytoplankters ( Shapiro, 1997 ). Our study showed that indeed cyanobacterial growth can be significantly stimulated by interaction with MOB in several ways, i.e., via direct supplement with CO 2 and via metabolites. It is not clear which path is more relevant and further investigations are required, especially on the possibility of metabolite uptake by cyanobacteria. It is also not known whether cyanobacteria intake metabolites, e.g., from free-living bacteria, directly from bacteria living in cyanosphere or if they use CO 2 released by other microbial organisms utilizing metabolites from MOB. Whichever path is true we can say that MOB support the growth of cyanobacteria. The influence of methane-derived carbon dioxide and MOB on photosynthetic organisms, such as cyanobacteria, has been overlooked in aquatic ecology, but methane-derived carbon may perform an important role in the development of harmful cyanobacteria that threaten the water quality of many lakes. Only recently has the influence of phytoplankton biomass on greenhouse gas production been brought to attention ( Bartosiewicz et al., 2021 ), and climate warming together with eutrophication strongly enhances the production of these gases. Our findings put the relative roles of MOB in the aquatic food web in a different perspective especially in the case of climate-induced increases in carbon and nutrient loading. MOB prevent the release of the diffusive CH 4 fluxes from the water column ( Bastviken et al., 2008 ), thus preventing release of a potent greenhouse gas to the atmosphere. Also, methane oxidation performs an important role in an aquatic microbial loop as predation on MOB by protozoa and zooplankton transfers energy and carbon to higher trophic levels ( Bastviken et al., 2003 ; Kankaala et al., 2007 ; Schilder et al., 2017 ). Cyanobacteria often form intense blooms due to increasing eutrophication ( Glibert et al., 2014 ) and global warming ( Visser et al., 2016 ). Even though the larger the bloom the more CO 2 can be sequestered, it should be noted that it eventually leads to hypoxia and in consequence the increase in CH 4 production. As a consequence a positive feedback loop may enhance both cyanobacteria proliferation and methane production ( Bartosiewicz et al., 2021; Bižić , 2021 ). In our study we have pointed to the overlooked alternative path that may have important consequences for lake ecosystem functioning."
} | 6,830 |
35522739 | PMC9075801 | pmc | 3,044 | {
"abstract": "There is an increasing trend of combining living cells with inorganic semiconductors to construct semi-artificial photosynthesis systems. Creating a robust and benign bio-abiotic interface is key to the success of such solar-to-chemical conversions but often faces a variety of challenges, including biocompatibility and the susceptibility of cell membrane to high-energy damage arising from direct interfacial contact. Here, we report living mineralized biofilms as an ultrastable and biocompatible bio-abiotic interface to implement single enzyme to whole-cell photocatalytic applications. These photocatalyst-mineralized biofilms exhibited efficient photoelectrical responses and were further exploited for diverse photocatalytic reaction systems including a whole-cell photocatalytic CO 2 reduction system enabled by the same biofilm-producing strain. Segregated from the extracellularly mineralized semiconductors, the bacteria remained alive even after 5 cycles of photocatalytic NADH regeneration reactions, and the biofilms could be easily regenerated. Our work thus demonstrates the construction of biocompatible interfaces using biofilm matrices and establishes proof of concept for future sustainable photocatalytic applications.",
"introduction": "INTRODUCTION Living biological material systems, in contrast with their nonliving counterparts, have an inimitable combination of features including environmental adaption, self-propagation, responsiveness to environmental stimuli, and self-repair ability ( 1 – 4 ). Inorganic systems, particularly inorganic nanomaterials such as semiconducting nanoparticles (NPs), have their own unique properties such as high surface-to-volume ratios, broad-range light-harvesting ability, and extremely efficient photoelectrochemical properties ( 5 – 7 ). Thus, the idea of combining biological and inorganic nanomaterial systems has attracted much attention, with researchers in multiple areas exploring potentially vast application opportunities in areas including catalysis ( 8 ), electronics ( 9 ), precision drug delivery ( 10 ), diagnostics ( 11 ), and disease treatment ( 12 ). In particular, the combination of living microbes with inorganic semiconductors has recently become a research hotspot in the nascent materials research field known as “semi-artificial photosynthesis” ( 13 , 14 ). In semi-artificial photosynthesis systems, the design of biocompatible interfaces to hybridize microbial cells and inorganic semiconductors is viewed as the key to achieve high-efficiency solar-driven fuel (such as H 2 or liquid alcohol) ( 15 , 16 ) or chemical production (such as complex metabolite shikimic acid) ( 17 ). Several strategies have been previously reported for the integration of microbial cells with semiconducting NPs for photoinduced biocatalytic applications ( 13 ). Through CdS NP precipitation on the cell surface of the bacterium Moorella thermoacetica , photoinduced CO 2 fixation to acetic acid was achieved using photogenerated electrons as the reducing power ( 18 ). During the reaction procedure, the oxidative holes produced by CdS upon light irradiation caused cell damage and even complete cell rupture at higher light intensities. An alternative method, the “nanobiohybrid organism” was developed on the basis of the intracellular uptake of quantum dots (including CdS, CdSe, and InP QDs) into bacteria (including Azotobacter vinelandii and Cupriavidus necator ) for photoinduced conversion of CO 2 , N 2 , and H 2 O into renewable biochemicals (such as butanediol and poly-β-hydroxybutyric acid) ( 19 ). Despite having impressive catalytic activities, these systems unfortunately lacked a sustainable bio-abiotic interface for continuous solar-to-chemical conversion, as intracellular QDs could also impair the cell viability originating from the toxicity of the ligands (such as cysteamine or 3-mercaptopropionic acid) modified on the surface of QDs ( 19 ) or the insertion of NPs into the cell membrane ( 20 ). Thus, despite important progress, creation of ideal interfaces between cells and nanomaterials, which had both interfacial stability and minimum cell damage, was still an unresolved challenge due to functional differences between biological and inorganic nanosystems ( 21 , 22 ). Biofilms are natural consortia of microbial cells embedded in their secreted extracellular matrix, composed of amyloid fibers exhibiting superior resilience to external environmental stresses ( 23 , 24 ). Here, inspired by the vast application potential of bacterial biofilms as engineered living materials ( 2 ) and biocatalysts ( 24 ), we developed an in situ mineralization strategy for the construction of photocatalyst-mineralized biofilms and harvested such mineralized biofilms as living bio-abiotic interfaces to implement diverse photocatalytic applications. Specifically, we first fused A7 peptides, containing two cysteines having strong affinity toward transition metal ions, to the C terminus of the Escherichia coli curli major subunit CsgA protein to construct functional CsgA A7 nanofibers, which site-specifically promoted the in situ extracellular mineralization of CdS NPs. Then, we used functional Tc Receiver /CsgA A7 biofilms to generate photocatalyst-mineralized biofilms through in situ mineralization of CdS NPs on E. coli biofilm nanofibers, thereby retaining the attractive photocatalytic potential of the hybrid catalyst while alleviating impairment through segregation of CdS NPs from bacterial cells ( Fig. 1 ). The CdS NPs within photocatalyst-mineralized biofilms generated electrons upon light irradiation, and the electrons were used and transferred to the redox centers of catalytic enzymes via electron mediators [such as methyl viologen (MV) or nicotinamide adenine dinucleotide (NAD)]. Last, we successfully achieved single enzyme to whole-cell photocatalytic applications including photocatalytic trimethylpyruvic acid (TMP) reduction using purified leucine dehydrogenase (LDH) enzyme ( Fig. 1A ) and photocatalytic CO 2 reduction using a single whole cell coexpressing extracellular CsgA A7 fibers and intracellular formate dehydrogenase (FDH) ( Fig. 1B ). In particular, the cells remained alive even after 5 cycles of photocatalytic NADH (reduced form of NAD + ) regeneration reactions. The proof-of-concept demonstration of cyclable photocatalytic applications suggested that the biofilms can serve as excellent platforms for the construction of biocompatible bio-abiotic interfaces for sustainable production of chemicals and stocks. Our constructed photocatalyst-mineralized biofilms exhibited alleviated injuries through segregation and higher catalytic efficiency through more active site exposure. The design expands the frontier of semi-artificial photosynthesis and has the potential to drive further research that can solve future energy and environmental issues. Fig. 1. Schematic of living photocatalyst-mineralized biofilms explored for single enzyme to whole-cell photocatalysis. The A7 peptides within the biofilms are functional peptides specifically displayed on the CsgA A7 nanofibers that can be mineralized via in situ formation of CdS NPs generating photocatalyst-mineralized biofilms. The mineralized biofilms harbor the CdS NPs that can produce electrons upon light irradiation; the electrons are then harnessed and transferred to the catalytic centers of redox enzymes through electron mediators (MV or NAD) for various photocatalytic applications including reduction of TMP into l -tert-leucine coupling with purified LDH ( A ) and CO 2 reduction coupling with a single cell coexpressing CsgA A7 fibers and FDH ( B ).",
"discussion": "DISCUSSION The design of biocompatible interfaces between microbes and semiconductors holds great promise for diverse applications, especially for sustainable production of chemicals through technologies from the emerging area of semi-artificial photosynthesis ( 14 ). Seeking to alleviate the possible damage that can occur at the cell-inorganic interfaces in photocatalytic systems ( 18 , 20 ), we here produced living bio-abiotic interfaces where E. coli biofilms engineered with CsgA A7 curli nanofibers as scaffolds for mineralization with CdS NPs were used to construct bio-abiotic interfaces that spatially segregate the mineralized CdS NP component from the bacterial cells. After mineralization, these NPs retained their useful photoelectric properties and functioned as efficient semiconductors. We showed that our photocatalyst-mineralized biofilms can achieve photocatalyzed transfer of electrons to diverse electron mediator compounds. We then exploited these photoinduced electrons for diverse photocatalytic applications, including redox-based synthesis of l -tert-leucine with an exogenously applied enzyme, and reduction of CO 2 to generate formic acid using a single engineered bacterial strain producing both the CsgA A7 nanofibers and the FDH enzyme. Previous attempts for the design of bio-abiotic interfaces include anchoring QDs on E. coli biofilms for photocatalytic H 2 production ( 29 ) and growing Sporomusa ovata biofilms on the silicon nanowire array electrode for acetate production coupling with a photovoltaic device ( 34 ). The preparation of QDs and silicon nanowire arrays necessarily involves the use of toxic reagents and harsh reaction conditions, while the coupling of the biofilm chassis with biomediated synthesis of nanomaterials, introduced in the current work, represents an important advantage over previous approaches owing to the simple, mild, and environmentally friendly reaction conditions ( 35 , 36 ). Furthermore, our initial demonstration clearly illustrates the utility of this approach for enabling seamless integration of inorganic and biological materials while protecting delicate cells from damage by high–energy state semiconductor for photocatalytic applications. Given that diverse types of amyloid nanofibers are widespread in nature, for example, in the Gram-positive Bacillus subtilis ( 37 ) or the facultative Shewanella oneidensis famous for its electron conductive capacity ( 38 ), the design principles demonstrated in the present study could be extended to other biofilm systems. In particular, the capacity to engineer the genomes of the various host strains should support the transfer of additional metabolic modules for applications in additional areas of semi-artificial photosynthesis or other pathways that require extensive local generation of reductants (e.g., NADH and NAD phosphate)."
} | 2,634 |
35418216 | PMC8830174 | pmc | 3,048 | {
"abstract": "Background This study examines the destiny of macromolecules in different full-scale biogas processes. From previous studies it is clear that the residual organic matter in outgoing digestates can have significant biogas potential, but the factors dictating the size and composition of this residual fraction and how they correlate with the residual methane potential (RMP) are not fully understood. The aim of this study was to generate additional knowledge of the composition of residual digestate fractions and to understand how they correlate with various operational and chemical parameters. The organic composition of both the substrates and digestates from nine biogas plants operating on food waste, sewage sludge, or agricultural waste was characterized and the residual organic fractions were linked to substrate type, trace metal content, ammonia concentration, operational parameters, RMP, and enzyme activity. Results Carbohydrates represented the largest fraction of the total VS (32–68%) in most substrates. However, in the digestates protein was instead the most abundant residual macromolecule in almost all plants (3–21 g/kg). The degradation efficiency of proteins generally lower (28–79%) compared to carbohydrates (67–94%) and fats (86–91%). High residual protein content was coupled to recalcitrant protein fractions and microbial biomass, either from the substrate or formed in the degradation process. Co-digesting sewage sludge with fat increased the protein degradation efficiency with 18%, possibly through a priming mechanism where addition of easily degradable substrates also triggers the degradation of more complex fractions. In this study, high residual methane production (> 140 L CH 4 /kg VS) was firstly coupled to operation at unstable process conditions caused mainly by ammonia inhibition (0.74 mg NH 3 -N/kg) and/or trace element deficiency and, secondly, to short hydraulic retention time (HRT) (55 days) relative to the slow digestion of agricultural waste and manure. Conclusions Operation at unstable conditions was one reason for the high residual macromolecule content and high RMP. The outgoing protein content was relatively high in all digesters and improving the degradation of proteins represents one important way to increase the VS reduction and methane production in biogas plants. Post-treatment or post-digestion of digestates, targeting microbial biomass or recalcitrant protein fractions, is a potential way to achieve increased protein degradation. Supplementary Information The online version contains supplementary material available at 10.1186/s13068-022-02103-3.",
"conclusion": "Conclusions This study focused on residual methane potential and the destiny of macromolecules in different full-scale biogas processes, from substrate to digestate. Our results showed that: Protein was the most abundant macromolecule in the digestates from plants operating on FW and sewage sludge (3–21 g/kg), while free sugars and fat were efficiently degraded High residual protein content was partly coupled to recalcitrant protein, but also to the formation of microbial biomass during substrate degradation Unstable digestion processes (i.e. high total concentration of volatile fatty acids and low volatile solids [VS] reduction) due to ammonia inhibition (> 0.7 mg NH 3 -N/kg), partly caused by digestion at elevated temperatures and/or deficiencies of trace elements (mainly cobalt), led to the accumulation of VFAs and to high RMP in the digestates Co-digestion of sewage sludge with fat increased protein degradation efficiency with 18%, possibly through the mechanism referred to as priming Furthermore, theoretical gas yields (TMP red ) were calculated based on the contents of macromolecules in the substrates and digestates. This parameter gave a more accurate assessment of the overall biogas efficiency compared with VS degradation, as it took the gas potential of the different types of VS into account. TMP red was approximately 10%-units higher compared to VS reduction since it considered that fat (which were almost completely degraded) holds a higher gas potential than for example carbohydrates. Together with the RMP, which is related to the degradability of the remaining VS, different strategies to access the gas potential of the residual fractions could be formulated: To improve biogas yields of protein-rich digestates, post-treatment prior to post-digestion is desirable, as the microbial biomass formed during AD would be targeted as well. Post-treatment should focus on methods directed at disrupting microbial biomass and recalcitrant protein structures For improved digestion where TMP is high, for example, with high residual content of carbohydrates, extending HRTs by, for example, post-digestion would be suitable to reduce RMP and thus limit GHG emissions during digestate storage and application Use of trace element supplementation may be necessary to obtain a stable process, especially when operating at higher temperatures (i.e. 55 °C)",
"introduction": "Introduction Anaerobic digestion (AD) is a well-established technology for both generating renewable biogas and valorizing organic waste fractions [ 1 ]. Biogas consists primarily of methane and carbon dioxide, with methane being a versatile energy carrier that can replace fossil fuel in vehicles, heat and power production, and industrial processes [ 1 ]. In addition to biogas, AD produces a nutrient-rich residue (i.e. digestate) that can be used as fertilizer in agriculture, thereby recycling nutrients between urban and rural areas [ 2 ]. Due to the multifunctionality of the process and its high-value outputs, AD can be seen as central to achieving a circular bioeconomy. Consequently, the number of AD plants in Europe has increased in recent years, and waste and residues from agriculture and industry as well as municipal organic waste and sewage sludge are now increasingly treated in biogas plants [ 3 ]. To reach environmental goals and to optimize economic output, AD plants should be operated at high efficiency, i.e. at a high biogas yield per reactor volume and time combined with a high degree of degradation. The degradation efficiency of the biogas process is important for both biogas production and nutrient levels in the digestate and is also decisive for the digestate’s residual methane production [ 4 ]. A well-digested material with low residual methane potential (RMP) will decrease greenhouse gas (GHG) emissions associated with subsequent digestate storage. The efficiency of a biogas process depends on several factors, often interlinked, including the composition and pre-treatment of the ingoing substrate, operational parameters such as organic load, hydraulic retention time, mixing and digester fluid behaviour and temperature, as well as digester technology [ 5 ]. In addition, an active and well-synchronized microbial community is needed [ 6 ]. AD processes proceed through several degradation steps performed by metabolically linked microbial groups, typically operating in a synchronized manner [ 5 , 7 ]. In the first step, hydrolysis, extracellular enzymes (e.g., lipases, proteases, and cellulases) attack and break up fat, proteins, and carbohydrates. The rate of this step is strongly dependent on the substrate accessibility, which in turn depends on the chemical composition [ 8 ]. In the next steps, the hydrolysis products (i.e. oligo- and monomers) are converted to fatty acids (via acidogenesis) and alcohols, followed by further oxidation of the acids mainly to acetic acid and hydrogen/carbon dioxide (via acetogenesis). In the final step, acetate and H 2 /CO 2 are converted to methane by acetotrophic and hydrogenotrophic methanogens, respectively. The composition and activity of the prevailing microbial community depend on both the external operational parameters and internal environmental conditions of the digester (e.g., pH, ammonia levels, and volatile fatty acid concentration [ 6 ]). In addition, the levels and availability of various ions and trace metals are crucial for microbial activity (mainly at the enzyme level) and thus for substrate degradation efficiency [ 9 ]. Moreover, the proportion of acetotrophic versus hydrogenotrophic methanogenesis depends strongly on the prevailing environmental conditions. At a high ammonia level or a thermophilic temperature, methane formation proceeds mainly via syntrophic acetate oxidation (SAO) coupled to hydrogenotrophic methanogenesis [ 10 ]. The AD of substrates and the subsequent methane production can be improved by various operational strategies such as: (a) pre-treatments to break up complex macromolecular structures for increased substrate accessibility and microbial degradation (reviewed by Atelge et al. [ 11 ] and Mirmohamadsadeghi et al.[ 12 ]); b) co-digestion approaches or the use of process additives, often trace metals, to improve the nutrient balance of the AD process (reviewed by Mata-Alvarez et al. [ 13 ]); c) adjusting the organic loading rate (OLR) and hydraulic retention time (HRT) or implementing post-digestion, i.e. implementing a second step of digestion to ensure sufficient degradation time; or d) using thermophilic operational conditions for enhanced degradation rates [ 14 ]. The optimal strategy to use will vary depending on the substrate mix. Based on RMP values determined in different studies, it is clear that the residual organic matter in digestates varies and can have significant biogas potential [ 4 , 15 , 16 ]. However, knowledge of the factors dictating the size and composition of this residual fraction and of how they correlate with RMP is currently lacking. The aim of this study was therefore to generate additional knowledge of the compositions of the residual fractions in digestates and to understand how they correlate to various operational and chemical parameters, in order to formulate strategies for enhanced methane production. This was undertaken by characterizing the composition of both the substrates and digestates from nine full-scale biogas plants, including both the main digesters and post-digesters, when present. The residual organic fractions were then linked to parameters such as substrate type, pH, levels of trace metals and ammonia, viscosity, operational parameters (e.g., OLR, HRT, and temperature), RMP, and enzyme activity (i.e. cellulase, protease, and lipase). The selected plants operated with different categories of substrates: four plants co-digested primarily food waste (FW), two plants operated on plant-based agricultural wastes (AW), one plant operated on AW together with manure (AWM), and two plants were wastewater treatment plants (WWTPs) using a mix of primary and waste activated sludge.",
"discussion": "Discussion What remains and why? The chemical compositions of the organic residues in digestate have been evaluated and discussed in a number of publications, mainly focusing on the stability of the digestate relative to its use as soil amendment/biofertilizer [ 17 – 19 ]. These studies suggest that carbohydrate, fat, and protein structures are all reduced during the AD process, leaving mainly more stable aliphatic and, to some extent, aromatic structures originating from fibre in the digestate. However, little is known about the correlations between operational parameters and the composition of residual organic matter in digestate, or about the correlation between digestate composition and residual methane production. Cluster analysis showed that the substrate category determined the clustering of the biogas plants in terms of chemical and operational parameters, although FW2 was an exception, clustering with AW plants rather than other FW plants (Fig. 1 ). Clustering according to substrate category was expected as the composition and chemistry of the feedstock should affect both operational management and the character of the resulting digestate, and similar categorization has been observed previously for microbial gene expression [ 20 , 21 ]. These results are also supported by a previous study conducting a non-target analysis of the dissolved organic matter composition in full-scale biogas reactors, showing that digestates originating from the AD of sewage sludge had characteristics distinctly different from those of digestates from co-digestion processes digesting different mixtures of organic wastes [ 22 ]. In that study, the operational temperature of the co-digestion reactors was also reported as a likely important parameter regarding differences in the residual dissolved organic matter characteristics. The study also showed that proteins were enriched in the digestates relative to carbohydrates [ 22 ]. Below follows a detailed discussion of the fate of different macromolecules relative to operational parameters and substrate category. Protein In line with the findings of Shakeri Yekta et al. [ 23 ], protein was the most abundant macromolecule in all investigated digestates, except that from AW1. Moreover, the in-depth mass balance study of three of the plants (i.e. FW1, FW-TD, and WWTP2) showed that the protein fraction of VS increased from substrate to digestate, while the fat and carbohydrate fractions were degraded to a larger extent. The degradation efficiencies of the biogas plants digesting AW displayed large variation, with protein degradation efficiencies ranging from 27–44% in AWM to about 55% in AW2 (Fig. 3 ). The protein degradation in AW1, determined to be 100%, however, is likely incorrect. This plant (AW1) received a larger proportion of easily degradable carbohydrates than did the other investigated plants (Table 2 ), resulting in high overall VS reduction, and since the estimation of Y prot (protein from microbial growth) is based on VS reduction, the calculation of residual protein might be underestimated (see Sect. 4.5). In contrast, AWM displayed poor hydrolysis of the protein, which could partly be connected to the type of substrate, which mainly comprised manure (67%; Table 1 ). In a previous study of the AD of pig manure, similarly low values for protein degradation efficiency were obtained (average 40% [ 24 ]). In line with this, a study investigating the AD of manure showed that the residual gas production from digestates mainly derived from the cellulose and hemicellulose fraction and that the protein fraction remained undigested [ 25 ]. Low protein degradation was also observed in the main digesters of the FW1 and FW2 processes (21% and 32%, respectively, calculated from values in Fig. 2 ), possibly caused by the slaughterhouse waste used as substrate in these plants, which likely contained recalcitrant protein fractions such as collagens which is known to be difficult to hydrolyse [ 26 ]. Consequently, proteins appear to have the highest potential for increased biogas production in digestates from different types of substrate categories. Higher protein hydrolysis, i.e. degradation efficiency would also improve the digestate quality as it would result in high levels of plant-available ammonium (NH 4 + ) as well as potentially decreased emissions of GHG (i.e. CH 4 ) from digestate storage [ 25 ]. The low degradation and higher VS fraction of protein could partly be explained by the fact that the outgoing proteins to some extent originated from recalcitrant microbial biomass (Y prot ) produced during the anaerobic degradation of VS. It is well known that substrates dominated by bacterial biomass, such as secondary sludge from WWTP, are rich in proteins and quite recalcitrant and therefore difficult to hydrolyse [ 27 , 28 ]. Similarly, the WWTPs in this study, which received secondary sludge as substrate (ca. 10–30% of ingoing material), displayed even lower protein degradation than did the other investigated biogas plants, as well as low VS reduction of 45–50%. Comparing the protein degradation efficiencies of the sewage sludge digesters (i.e. WWTPs) interestingly revealed that WWTP1 and WWTP2-B, both of which received grease separator sludge in addition to mixed sludge, had similar protein degradation efficiencies of 59% and 62%, respectively (Fig. 3 ). WWTP2-A, which received only mixed sludge, had a relatively lower protein degradation efficiency of 44% even though the amount of primary sludge going into this digester was higher at the time of sampling (27% vs. 17 and 12% in WWTP2-A and -B, respectively, based on TS). This indicates that the fat additions could have positive effects, promoting the more efficient degradation of the WWTP substrate. This could be due to a synergistic co-digestion effect observed for other substrates, perhaps generated by increased activity of the microbial community through improved environmental prerequisites for the active microorganisms [ 29 , 30 ]. Another explanation could be that adding easily degradable substrate increased the growth and/or activity of the microorganisms in this environment. This phenomenon has previously been observed in several studies of the microbial activity and degradation of organic matter in soil, and has in that field been referred to as the “priming effect” [ 31 ]. Hydrolysis of protein releases ammonium and ammonia, which can be problematic as ammonia-induced process disturbances in continuous digesters have been observed over a wide range of ammonia concentrations, i.e. 0.2–1.5 g/L NH 3 -N (reviewed by Capson-Tojo et al. [ 32 ]). The disturbances are typically more pronounced at higher operational temperatures, as this causes the level of free ammonia to increase. Ammonia inhibition could be one reason for the VFA accumulation and low VS reduction observed in FW-TD (> 8 g/L VFAs, > 0.7 g/kg NH 3 -N), and previous studies have shown a correlation between these two parameters (reviewed by Capson-Tojo et al. [ 32 ]). This high VFA accumulation was probably one reason why FW-TD did not cluster with any of the other plants in the clustering analysis, although the thermophilic conditions and higher TS content of the digestate than in the other plants likely also contributed. It has been observed that ammonia inhibition can cause the accumulation of acetate and, in more severe inhibition cases, propionate [ 10 , 33 ], the latter being the major acid produced in FW-TD. In this study, NH 4 + -N production correlated positively to the accumulation of acetate in the different reactors ( p < 0.05). Another plant with high VFA concentration was FW2, which displayed the lowest VS reduction of all studied plants. Of special interest here was the difference in performance between FW1 and FW2. Both processes treated similar substrates (although FW2 with a slightly shorter HRT) and had similar NH 3 -N levels, but still FW2 showed indications of instability with higher levels of VFAs as well as lower VS reduction compared with FW1. The metal analysis revealed that the level of the trace metal cobalt (Co) was low in FW2 (0.2 mg/kg), which is close to what has previously been determined to be critical for FW digestion [ 33 ] and lower than that in the other FW plants (0.7–0.9 mg/kg). Similarly, the Ni and Se contents were relatively low in FW2 (0.3 mg/kg and 0.04 mg/kg, respectively). The importance of Co, Ni, and Se for methanogenesis and SAO is well established [ 34 ], and the low levels of these elements in FW2 suggest a lack of trace elements, resulting in elevated VFA levels in the FW2 main digester. The level of Co in FW-TD was also low (0.1 mg/kg), which might be one factor contributing to the observed high levels of VFAs in this digester, particularly considering that requirements for trace elements (including Co) appear to be even higher in thermophilic conditions [ 35 ] and at high NH 4 -N concentrations [ 33 ]. The fact that the Co and Ni contents affect the overall digestion efficiency was also supported by the positive correlation of Co and Ni to VS reduction ( p < 0.05, r > 0.7) found in this study. The measured protease activity did not correlate to the protein degradation efficiency, which is likely explained by the build-up of protein-rich microbial biomass (Y prot ) masking the protein degradation of the substrate. This assumption is supported by the correlation between protease activity and the concentration of nitrogen, in both NH 4 -N and Kjeldahl-N forms, released during amino acid degradation, which represents an indirect measure of protein degradation. In summary, considering the protein concentration across different plants and from substrate to post-digester in this study, post-digestion results in substantially increased protein hydrolysis and hence degradation. Processes not applying post-digestion, for example, in the WWTP and AWM plants, had low protein degradation compared with the other plants. This could be because, first, both fat and free sugars are quickly degraded, while protein degradation is relatively slow, particularly under acidifying conditions [ 36 ], and, second, the microbial biomass produced during substrate degradation is rich in protein s and could be considered a post-digester substrate. However, accessing the organic material of the microbial biomass is challenging, and requires treatment before post-digestion. Carbohydrates The AW plants displayed high VS reduction (78 and 77% for AW1 and AW2, respectively), which could partly be explained by the relatively long HRTs applied by this plant category. AWM, which used a relatively shorter HRT, displayed a lower VS reduction of 62%. Long HRTs are generally necessary when treating AW rich in recalcitrant lignocellulosic materials which are slowly hydrolysed [ 25 , 37 , 38 ]. In a survey of 21 full-scale plants operating on AW, only two plants used HRTs under 60 days, and the average digestion time was about 100 days [ 15 ]. In a German survey of biogas plants operating primarily on plant-based AW (i.e. 55–100% crop residues), the HRTs were 46–191 days ( n = 24, mean = 98 days; [ 38 ]). However, a study of Danish biogas plants operating mainly (> 70%) on manure demonstrated that the HRTs were typically shorter than in digesters running on plant-based substrates only, with HRTs of 18–62 days ( n = 11, mean = 32 days; Supplementary Material in Hamelin et al. [ 37 ]). Even though lignocellulose-rich materials are known to be difficult to hydrolyse and hence degrade during AD, the VS reduction in the plant-based digesters was similar to that in FW1 and FW3, showing that high VS reduction can be obtained with long HRTs. However, the benefits of prolonging the HRT for enhanced degradation should be weighed against the drawbacks of inefficient biogas production per time unit, which reduces plant profitability. In all plants, VS reduction correlated to the degradation of sugars (e.g., galactose, glucose, mannose, and arabinose) as well as to the cellulase activity per g VS (Additional file 4 ). An exception to this was FW2, which displayed very low degradation of sugars (5%) and was among the lowest in cellulase activity, i.e. just 6% of the highest activity (Fig. 7 ). The sugar content of the substrate in this plant was much lower than that in the other FW plants, for example, 4 g/kg versus 40–46 g/kg in FW1 and FW3, whereas the level of VFA was higher at 160 mmol/L versus 50–60 mmol/L in FW1 and FW3 (Additional file 1 ). Combined, this indicates that pre-hydrolysis of the sugars in the substrate occurred before entering the digester. This conclusion is also supported by the VFA profile of the substrate mixture of FW2, as a high ratio of acetate and butyrate combined with a low pH (5.4) is characteristic of dark fermentation (i.e. H 2 production [ 39 ]). The difference in carbohydrate degradation between the biogas plants (except FW2) was likely related to their origin and composition. AW1 and AW2 primarily digested lignocellulosic-rich materials (e.g., crop silage, corn silage, cereals, and grain chaff), the carbohydrate degradability of which is limited due to their recalcitrant hemicellulose–lignin structures (reviewed by Carrere et al. [ 40 ]). In addition, a large fraction of the substrate in AWM was manure, which also contains recalcitrant fibre fractions remaining after feed digestion as well as bedding materials [ 24 ], which could explain the low degradation efficiency of carbohydrates in this plant (71%) versus the others (91–92%). Degradation of free or complexed xylose did not correlate with VS reduction, but as this sugar monomer is one of the main components of hemicellulose [ 41 ], this result is likely related to the recalcitrance of the lignocellulosic substrates discussed above. The silage at AW1 was rich in xylose (185 g/kg), but despite the low degradation efficiency of xylose at this plant, the VS reduction in the main digester was relatively high at 67%, likely due to the easily degradable sugars in the starch slurry. The decrease in lignin-like structures observed in FW1, WWTP2, and FW-TD (23–63%) was unexpected, as native lignin is considered highly recalcitrant under anaerobic conditions. One explanation could be overestimation of the ADL fraction in the lignin analysis, as a high content of free lipids (> 10% of TS) could interfere with the analysis [ 42 ]. The residual amounts of cellulose and hemicellulose detected in the detailed investigation of the three biogas plants differed somewhat, as shown in Fig. 5 . The fact that the cellulose content was higher than the hemicellulose content in the substrate of FW1, while the opposite was seen in the FW-TD substrate, was probably partly due to the 2% (volumetric) addition of garden waste to FW-TD (Table 2 ), and partly because this plant used brown paper bags to collect the FW (FW1 used plastic bags for collecting the organic FW). When comparing the FW digesters, the degradation efficiencies of hemicellulose were similar at 70–76%, while the cellulose degradation was much lower at 55% in FW-TD versus 82% in FW1. The reason for the lower degradation of cellulose could be related to the high ammonia content in the FW-TD digester resulting from the high temperature and high ammonium-nitrogen concentration (Table 1 ). In previous studies, a high ammonia concentration of 0.3–0.41 g/L was found to reduce cellulose degradation efficiency [ 43 , 44 ]. Fat In general, fat degradation was efficient in all plants for which data were available. The degradation of fat can be challenging at high loads and lead to problems with foaming, floatation, and inhibition by long-chain fatty acids (LCFAs; [ 45 , 46 ]). In this study, FW1 had the highest substrate fat content at 32% of total VS, yet with low concentrations of fat in the digestate. The efficient fat degradation could relate to a high abundance of syntrophic β-oxidizing bacteria and Cloacimonadota, which have been correlated to efficient degradation of lipids and long-chain fatty acids, respectively [ 47 , 48 ]. In addition, FW digesters might have a microbial community better adapted to maintaining the low partial pressure of hydrogen required for efficient lipid oxidation [ 49 ], as high ammonia concentrations result in more abundant or active hydrogenotrophic methanogens that via syntrophic acetate-oxidation produce methane through hydrogen consumption. From a physical/mechanical perspective, the often higher TS resulting from co-digestion with FW compared with, for example, sewage sludge digestion, provides more surface area for the lipids to adhere to, and could thus lead to better mixing with the fats [ 50 ] and improved lipid accessibility for the microorganisms. Furthermore, co-digestion plants typically have mandatory pasteurization of their substrate mixtures (as in FW1–FW3 in this study), in which heating, melting, and dissolution of the fat fractions result in better mixing with the remaining substrate and thereby improved degradation. The fat degradation efficiency of the FW2 main digester was notably low at 11%, but this might, as discussed in \" Protein \", be related to the low levels of trace elements and the combination of high levels of ammonia and VFAs that strongly indicates inhibition of the methanogens. Synergetic co-inhibition caused by LCFAs and high ammonia has previously been observed by Tian et al. [ 51 ], who argued high ammonia led to the accumulation of hydrogen and acetate, in turn rendering the β-oxidation pathway thermodynamically unfavourable, concomitantly with an accumulation of LCFAs leading to inhibition. Lipase activity correlated positively to residual fat content in the main digesters (Fig. 8 ; Additional file 4 ), and this was particularly evident in FW2, which displayed the highest lipase activity and had a fat content over three times higher than those of the other FW processes. Since the raw fat analysis includes both lipids and the LCFAs that are soluble in petrol (used during the extraction), it cannot be fully determined that there was an accumulation of LCFA, but, based on the enzyme activity and the poor performance of the digester, this was likely the case. Unfortunately, high lipase activity in a system in which LCFA degradation is hampered increases the risk of inhibition from LCFAs, as observed previously [ 52 ]. In addition, correlation analysis revealed a negative correlation between Fe concentration and lipase activity ( r = –0.7; Additional file 4 ), but whether Fe has a direct effect on lipase activity or whether the correlation is because of an indirect relationship is unclear and will require further investigation. Viscosity, EPS, and SMP Low VS reduction results in higher VS content in the digestate, and in the present study high viscosity was clearly correlated to high VS content ( r = 0.74 at a shear rate of 100/s; Additional file 4 ). The fact that higher viscosity was observed in AW1 and AW2 despite having similar or lower VS contents compared with FW-TD indicates that the composition rather than concentration of the VS is important when comparing digesters operating on different substrates. High viscosity can, apart from increasing the power demand [ 53 ], also negatively affect the mixing efficiency of digesters (reviewed by Lindmark et al. [ 54 ]), causing the formation of dead zones, sedimentation, and floating layers and ultimately leading to reduced degradation efficiency of the biogas process ([ 55 ]). In addition to VS content, a positive correlation between viscosity and the presence of EPS/EPS p and of cations (i.e. K, Mg, and Mn) could be identified in the biogas plants. Both cations and EPS have previously been found to increase viscosity, as the cations bridge and strengthen the polymer network, thus affecting the viscosity accordingly [ 56 – 58 ]. In line with this, there was a positive correlation of viscosity to arabinose (likely in the form of plant polysaccharides) at both investigated shear rates, with the highest content of arabinose being seen in AW1, followed by AW2 and AWM (Additional file 1 ). Lastly, the viscosity in the post-digesters was lower than in the main digesters (Fig. 8 ), possibly because the carbohydrate fraction of the EPS (EPSc) decreased due to the degradation of these molecules in the first digestion step, but not enough post-digesters were sampled to confirm this statistically. RMP, TMP, and TMP red The residual methane potential (RMP) is a good quantitative measure of how much of the remaining organic material in the digestate could actually contribute to increased biogas production, although it only shows the methane production that could be obtained without any further treatment [ 59 ]. Previous studies of RMP from various digestates report values of 20–240 mL CH 4 /g VS, in extreme cases corresponding to as much as 50% of the total biogas production, depending on the type of substrate and operation [ 15 , 16 , 60 ]. Several studies show a positive correlation between RMP and OLR and a negative correlation between RMP and HRT [ 4 , 16 ]. In this study, RMP correlated positively to the OLR of the digester ( p < 0.05; Additional file 4 ), while there was no clear negative correlation between RMP and HRT. The plants with both among the longest HRT (AW2) and shortest HRT (WWTP1A-B) had low RMPs, and the plant with the longest HRT (AW1) had among the highest RMPs. This was likely explained by the characteristics of the different substrates used; in contrast, Ruile et al. [ 16 ] included only agricultural digesters in their study. Instead, in this study, RMP (measured in L CH 4 /kg) was positively correlated to VS content. Comparing the specific RMPs (L CH 4 /kg VS), AW2 and the WWTPs displayed low potentials, which indicates that the quality of the outgoing VS from these plants was lower than that of the other plants in terms of potential biogas production. It also suggests that efforts to increase the gas production from these plants only by increasing the HRT would not likely be worthwhile. In addition, the low protein degradation efficiency in these plants (48–62%), together with the relatively high residual protein content in the digestates, suggests the presence of recalcitrant protein structures, and hence that post-treatment targeting proteins would be more efficient than only increasing the HRT in accessing this potential gas production. This conclusion is supported by the fact that only 6–13% of the theoretical methane potential (TMP) was obtained during the RMP tests in the WWTPs, corresponding to 0.2–1.2 L of additional CH 4 per kg substrate. In all digestates, except that in AW1, protein was the largest contributor to residual TMP, emphasizing the importance of targeting this macromolecule for increased biogas production. WWTP1 had relatively high RMP and low VS reduction compared with WWTP2-A and -B, which were also treating grease separator sludge (55% vs. 63 and 61% in WWTP2-A and -B), further supporting the conclusion in Sect. 3.1.1 that protein degradation was more efficient when fat was added as a co-substrate. Among the plants with higher specific RMP, FW2 and FW-TD both had high VFA contents and low VS degradation. FW-TD also contained a high level of crude fat relative to FW1 (4.1 vs. 1.5 g), which is also correlated to high gas potential and likely contributed to the high RMP in this case. However, the reason for this is partly related to process instabilities, meaning that the RMPs would likely be reduced if process operation was adjusted (i.e. adding trace elements or lowering the digestion temperature in the thermophilic digester). AWM had the highest RMP (170 mL CH 4 /g VS) and this suggests that a longer HRT could be a suitable measure to enhance the methane production. In FW1, FW2, FW-TD, and AWM (Fig. 6 and corresponding values for FW2 and AWM, not shown), only 40–46% of the TMP was obtained during the RMP tests (Fig. 4 B), even though they were run for about 100 days. In AWM and FW-TD, this resulted in 8.6 and 9.1 NL CH 4 /kg substrate, respectively, indicating that increasing the digestion time or adding a post-digestion step should be considered, although for plants with high VFAs (i.e. FW-TD and FW2), improving the performance of the main digestion to prevent VFA and LCFA accumulation should be the primary focus. Post-digestion could be a relatively easy measure to extract more methane from existing substrate and, in light of methane as a potent GHG, it would also be a way to reduce methane emissions during digestate storage before land application. Such emissions can be substantial, as demonstrated in a study finding that digestate emissions at a manure and FW co-digestion plant without post-digestion amounted to 12% of the plant’s yearly methane production [ 60 ]. Similarly, a survey of manure-based digesters showed that an estimated increase in methane production of about 20% could be achieved by prolonging the HRT from 30 to 60 days [ 37 ]. The reduction in theoretical methane potential (TMP red ) measures how much of the ingoing TMP is harvested in the biogas process, i.e. a high TMP red indicates an efficient process. Interestingly, the TMP red was higher than the VS reduction in several of the processes included in the mass balance, likely because fat, proteins, and carbohydrates have different gas yields [ 5 ]. For example, the TMP red of FW1 was 88% compared with a VS reduction of 77%, which is related to the fact that almost all the fat, which carries the highest methane production potential per g, was degraded, while some proteins and carbohydrates remained (harbouring lower methane production potential per g). From a methane production perspective, this somewhat lowers the incentive to further treat and digest the digestate in FW1, since the additional methane it would produce (a maximum of 12% of the substrate methane potential) might not compensate for the added costs. In AWM, on the other hand, TMP red was lower than the VS reduction (55% vs. 62%), suggesting that a significant amount of gas could yet be extracted. In FW-TD, the TMP red reached only 74%, indicating that the process disturbance observed and/or lack of post-digestion limited the degree of degradation and that up to 26% more gas could theoretically be produced from the substrate. In summary, when high residual organic content and high RMP are consequences of process instability and VFA accumulation, the first action to take would be to pinpoint the reason for the instability (e.g., lack of trace elements or ammonia inhibition) and adjust the process operation accordingly. If the digestate still has a high RMP, another action could be to prolong the digestion time of the main digester or implement post-digestion. Lastly, as clearly demonstrated in this study, many digestates covering different AD processes seem to contain recalcitrant protein fractions, either from the substrate or in microbial biomass. To better access the methane contained in this material, targeted treatment before post-digestion could be promising. In addition, it is logical to suggest post-treatment of the residual fractions rather than pre-treatment of the substrates to increase the efficiency of biogas processes, as this would avoid spending energy or chemicals to also treat the already easily accessible fractions of the substrate (e.g., fats and sugars)."
} | 9,557 |
32681001 | PMC7367848 | pmc | 3,050 | {
"abstract": "Recurrently connected networks of spiking neurons underlie the astounding information processing capabilities of the brain. Yet in spite of extensive research, how they can learn through synaptic plasticity to carry out complex network computations remains unclear. We argue that two pieces of this puzzle were provided by experimental data from neuroscience. A mathematical result tells us how these pieces need to be combined to enable biologically plausible online network learning through gradient descent, in particular deep reinforcement learning. This learning method–called e-prop–approaches the performance of backpropagation through time (BPTT), the best-known method for training recurrent neural networks in machine learning. In addition, it suggests a method for powerful on-chip learning in energy-efficient spike-based hardware for artificial intelligence.",
"introduction": "Introduction Networks of neurons in the brain differ in at least two essential aspects from deep neural networks in machine learning: they are recurrently connected, forming a giant number of loops, and they communicate via asynchronously emitted stereotypical electrical pulses, called spikes, rather than bits or numbers that are produced in a synchronized manner by each layer of a deep feedforward network. We consider the arguably most prominent model for spiking neurons in the brain: leaky integrate-and-fire (LIF) neurons, where spikes that arrive from other neurons through synaptic connections are multiplied with the corresponding synaptic weight, and are linearly integrated by a leaky membrane potential. The neuron fires—i.e., emits a spike—when the membrane potential reaches a firing threshold. But it is an open problem how recurrent networks of spiking neurons (RSNNs) can learn, i.e., how their synaptic weights can be modified by local rules for synaptic plasticity so that the computational performance of the network improves. In deep learning, this problem is solved for feedforward networks through gradient descent for a loss function E that measures imperfections of current network performance 1 . Gradients of E are propagated backwards through all layers of the feedforward network to each synapse through a process called backpropagation. Recurrently connected networks can compute more efficiently because each neuron can participate several times in a network computation, and they are able to solve tasks that require integration of information over time or a non-trivial timing of network outputs according to task demands. Therefore, the impact of a synaptic weight on the loss function (see Fig. 1 a) is more indirect, and learning through gradient descent becomes substantially more difficult. This problem is aggravated if there are slowly changing hidden variables in the neuron model, as in neurons with spike-frequency adaptation (SFA). Neurons with SFA are quite common in the neocortex 2 , and it turns out that their inclusion in the RSNN significantly increases the computational power of the network 3 . In fact, RSNNs trained through gradient descent acquire then similar computing capabilities as networks of LSTM (long short-term memory) units, the state of the art for recurrent neural networks in machine learning. Because of this functional relation to LSTM networks these RSNN models are referred to as LSNNs 3 . In machine learning, one trains recurrent neural networks by unrolling the network into a virtual feedforward network 1 , see Fig. 1 b, and applying the backpropagation algorithm to that (Fig. 1 c). This method is called backpropagation through time (BPTT), as it requires propagation of gradients backwards in time. Fig. 1 Schemes for BPTT and e-prop. a RSNN with network inputs x , neuron spikes z , hidden neuron states h , and output targets y * , for each time step t of the RSNN computation. Output neurons y provide a low-pass filter of a weighted sum of network spikes z . b BPTT computes gradients in the unrolled version of the network. It has a new copy of the neurons of the RSNN for each time step t . A synaptic connection from neuron i to neuron j of the RSNN is replaced by an array of feedforward connections, one for each time step t , which goes from the copy of neuron i in the layer for time step t to a copy of neuron j in the layer for time step t + 1. All synapses in this array have the same weight: the weight of this synaptic connection in the RSNN. c Loss gradients of BPTT are propagated backwards in time and retrograde across synapses in an offline manner, long after the forward computation has passed a layer. d Online learning dynamics of e-prop. Feedforward computation of eligibility traces is indicated in blue. These are combined with online learning signals according to Eq. ( 1 ). With a careful choice of the pseudo derivative for handling the discontinuous dynamics of spiking neurons, one can apply BPTT also to RSNNs, and RSNNs were able to learn in this way to solve really demanding computational tasks 3 , 4 . But the dilemma is that BPTT requires storing the intermediate states of all neurons during a network computation, and merging these in a subsequent offline process with gradients that are computed backwards in time (see Fig. 1 c). This makes it very unlikely that BPTT is used by the brain 5 . The previous lack of powerful online learning methods for RSNNs also affected the use of neuromorphic computing hardware, which aims at a drastic reduction in the energy consumption of AI implementations. A substantial fraction of this neuromorphic hardware, such as SpiNNaker 6 or Intel’s Loihi chip 7 , implements RSNNs. Although it does not matter here whether the learning algorithm is biologically plausible, the excessive storage and offline processing demands of BPTT make this option unappealing. Hence, there also exists a learning dilemma for RSNNs in neuromorphic hardware. We are not aware of previous work on online gradient descent learning methods for RSNNs, neither for supervised learning nor for reinforcement learning (RL). There exists, however, preceding work on online approximations of gradient descent for non-spiking neural networks based on 8 , which we review in the Discussion Section. Two streams of experimental data from neuroscience provide clues about the organization of online network learning in the brain: First, neurons in the brain maintain traces of preceding activity on the molecular level, for example, in the form of calcium ions or activated CaMKII enzymes 9 . In particular, they maintain a fading memory of events where the presynaptic neuron fired before the postsynaptic neuron, which is known to induce synaptic plasticity if followed by a top–down learning signal 10 – 12 . Such traces are often referred to as eligibility traces. Second, in the brain, there exists an abundance of top–down signals such as dopamine, acetylcholine, and neural firing 13 related to the error-related negativity, that inform local populations of neurons about behavioral results. Furthermore, dopamine signals 14 , 15 have been found to be specific for different target populations of neurons, rather than being global. We refer in our learning model to such top–down signals as learning signals. A re-analysis of the mathematical basis of gradient descent learning in recurrent neural networks tells us how local eligibility traces and top–down learning signals should be optimally combined—without requiring backprogation of signals through time. The resulting learning method e-prop is illustrated in Fig. 1d . It learns slower than BPTT, but tends to approximate the performance of BPTT, thereby providing a first solution to the learning dilemma for RSNNs. Furthermore, e-prop also works for RSNNs with more complex neuron models, such as LSNNs. This new learning paradigm elucidates how the brain could learn to recognize phonemes in spoken language, solve temporal credit assignment problems, and acquire new behaviors just from rewards.",
"discussion": "Discussion We propose that in order to understand the computational function and neural coding of neural networks in the brain, one needs to understand the organization of the plasticity mechanisms that install and maintain these. So far, BPTT was the only candidate for that, as no other learning method provided sufficiently powerful computational function to RSNN models. But as BPTT is not viewed to be biologically realistic 5 , it does not help us to understand learning in the brain. E-prop offers a solution to this dilemma, as it does not require biologically unrealistic mechanisms, but still enables RSNNs to learn difficult computational tasks, in fact almost as well as BPTT. Furthermore, it enables RSNNs to solve these tasks in an energy-efficient sparse firing regime, rather than resorting to rate coding. E-prop relies on two types of signals that are abundantly available in the brain, but whose precise role for learning have not yet been understood: eligibility traces and learning signals. As e-prop is based on a transparent mathematical principle (see Eq. ( 3 )), it provides a normative model for both types of signals, as well as for synaptic plasticity rules. Interestingly, the resulting learning model suggests that a characteristic aspect of many biological neurons—the presence of slowly changing hidden variables—provides a possible solution to the problem how a RSNN can learn without error signals that propagate backwards in time: slowly changing hidden variables of neurons cause eligibility traces that propagate forward over longer time spans, and are therefore able to coincide with later arising instantaneous error signals (see Fig. 3 b). The theory of e-prop makes a concrete experimentally testable prediction: that the time constant of the eligibility trace for a synapse is correlated with the time constant for the history-dependence of the firing activity of the postsynaptic neuron. It also suggests that the experimentally found diverse time constants of the firing activity of populations of neurons in different brain areas 30 are correlated with their capability to handle corresponding ranges of delays in temporal credit assignment for learning. Finally, e-prop theory provides a hypothesis for the functional role of the experimentally found diversity of dopamine signals to different populations of neurons 14 . Whereas previous theories of reward-based learning required that the same learning signal is sent to all neurons, the basic Eq. ( 1 ) for e-prop postulates that ideal top–down learning signals to a population of neurons depend on its impact on the network performance (loss function), and should therefore be target-specific (see Fig. 2 c and Supplementary Note 6 ). In fact, the learning-to-learn result for e-prop in ref. 31 suggests that prior knowledge about the possible range of learning tasks for a brain area could optimize top–down learning signals even further on an evolutionary time scale, thereby enabling for example learning from few or even a single trial. Previous methods for training RSNNs did not aim at approximating BPTT. Instead, some of them were relying on control theory to train a chaotic reservoir of spiking neurons 32 – 34 . Others used the FORCE algorithm 35 , 36 or variants of it 35 , 37 – 39 . However, the FORCE algorithm was not argued to be biologically realistic, as the plasticity rule for each synaptic weight requires knowledge of the current values of all other synaptic weights. The generic task considered in ref. 35 was to learn with supervision how to generate patterns. We show in Supplementary Figs. 1 and 7 that RSNNs can learn such tasks also with a biologically plausible learning method e-prop. Several methods for approximating stochastic gradient descent in feedforward networks of spiking neurons have been proposed, see e.g., refs. 40 – 44 . These employ—like e-prop—a pseudo-gradient to overcome the non-differentiability of a spiking neuron, as proposed previously in refs. 45 , 46 . References 40 , 42 , 43 arrive at a synaptic plasticity rule for feedforward networks that consists—like e-prop—of the product of a learning signal and a derivative (eligibility trace) that describes the dependence of a spike of a neuron j on the weight of an afferent synapse W j i . But in a recurrent network, the spike output of j depends on W j i also indirectly, via loops in the network that allow that a spike of neuron j contributes to the firing of other neurons, which in turn affect the firing of the presynaptic neuron i . Hence, the corresponding eligibility trace can no longer be locally computed if one transfers these methods for feedforward networks to recurrently connected networks. Therefore, ref. 40 suggests the need to investigate extensions of their approach to RSNNs. Previous work on the design of online gradient descent learning algorithms for non-spiking RNNs was based on real-time recurrent learning (RTRL) 8 . RTRL itself has rarely been used as its computational complexity per time step 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}$${\\mathcal{O}}({n}^{4})$$\\end{document} O ( n 4 ) , if n is the number of neurons. But interesting approximations to RTRL have subsequently been proposed (see ref. 47 for a review): some stochastic approximations 48 , which are \\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}$${\\mathcal{O}}({n}^{3})$$\\end{document} O ( n 3 ) or only applicable for small networks 49 , and also recently two deterministic \\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}$${\\mathcal{O}}({n}^{2})$$\\end{document} O ( n 2 ) approximations 50 , 51 . The latter were in fact written at the same time as the first publication of e-prop 31 . A structural difference between this paper and 50 is that their approach requires that learning signals are transmitted between the neurons in the RNN, with separately learnt weights. 51 derived for rate based neurons a learning rule similar to random e-prop. But this work did not address other forms of learning than supervised regression, such as RL, nor learning in networks of spiking neurons, or in more powerful types of RNNs with slow hidden variables such as LSTM networks or LSNNs. E-prop also has complexity \\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}$${\\mathcal{O}}({n}^{2})$$\\end{document} O ( n 2 ) , in fact \\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}$${\\mathcal{O}}(S)$$\\end{document} O ( S ) if S is the number of synaptic connections. This bound is optimal—except for the constant factor—since this is the asymptotic complexity of just simulating the RNN. The key point of e-prop is that the general form ( 13 ) of its eligibility trace collects all contributions to the loss gradient that can be locally computed in a feedforward manner. This general form enables applications to spiking neurons with slowly varying hidden variables, such as neurons with firing rate adaptation, which are essential ingredients of RSNNs to reach the computational power of LSTM networks 3 . We believe that this approach can be extended in future work with a suitable choice of pseudo derivatives to a wide range of biologically more realistic neuron models. It also enables the combination of these rigorously derived eligibility traces with—semantically identical but algorithmically very different—eligibility traces from RL for reward-based e-prop (Eq. ( 5 )), thereby bringing the power of deep RL to RSNNs. As a result, we were able to show in Figs. 2 – 5 that RSNNs can learn with the biologically plausible rules for synaptic plasticity that arise from the e-prop theory to solve tasks such as phoneme recognition, integrating evidence over time and waiting for the right moment to act, and winning Atari games. These are tasks that are fundamental for modern learning-based AI, but have so far not been solved with RSNNs. Hence, e-prop provides a new perspective of the major open question how intelligent behavior can be learnt and controlled by neural networks of the brain. Apart from obvious consequences of e-prop for research in neuroscience and cognitive science, e-prop also provides an interesting new tool for approaches in machine learning where BPTT is replaced by approximations in order to improve computational efficiency. We have already shown in Supplementary Fig. 4 that e-prop provides a powerful online learning method for LSTM networks. Furthermore, the combination of eligibility traces from e-prop with synthetic gradients from ref. 52 even improves performance of LSTM networks for difficult machine learning problems such as the copy-repeat task and the Penn Treebank word prediction task 31 . Other future extensions of e-prop could explore a combination with attention-based models in order to cover multiple timescales. Finally, e-prop suggests a promising new approach for realizing powerful on-chip learning of RSNNs on neuromorphic chips. Whereas, BPTT is not within reach of current neuromorphic hardware, an implementation of e-prop appears to offer no serious hurdle. Our results show that an implementation of e-prop will provide a qualitative jump in on-chip learning capabilities of neuromorphic hardware."
} | 4,577 |
30796234 | PMC6384942 | pmc | 3,051 | {
"abstract": "Resistive switching (RS) was demonstrated in four-terminal planar memristive devices fabricated on reduced TiO 2 (TiO 2−x ) single crystal substrates. In the device, a pair of diagonally opposing electrode terminals is used to modify the distribution of oxygen vacancies in the region between another pair of diagonally opposing electrode terminals. This allowed microscopic visual observations of the oxygen vacancy distribution based on electrocoloring. The visual contrast observed in the TiO 2−x reflects the oxygen vacancy concentration in the electrically active zone of the device, which can be modified by application of various external voltages to the electrodes. The current that flows in the device is significantly dependent on the modified oxygen vacancy distribution and the resultant resistance is switchable when the polarization of the applied external voltage is reversed. The crystallographic orientation of the TiO 2−x substrate has a strong influence on the reversible RS phenomenon. Mechanisms behind the voltage-driven resistance change are elaborated with the aid of microscopic analysis for both crystalline and electronic structures in the electrically active zone of the device. Suppression of the formation of irreversible conductive structures comprised of accumulated oxygen vacancies is a key to establishing reversible RS in the device.",
"conclusion": "Summary and Conclusions We have demonstrated RS on the basis of carrier conductivity modulation caused by modification of the oxygen vacancy distribution in four-terminal planar devices fabricated on rutile TiO 2−x single-crystal substrates. The electrocoloring phenomenon effectively facilitates visual observation of the oxygen vacancy distribution and local transitions driven by an external voltage. Thus, a one-to-one correspondence between the electrically controlled LRS/HRS and the morphologically controlled oxygen vacancy distribution in the device is unambiguous. STEM analysis revealed that the electrically active zone in the (100) device is highly reduced and the atomic arrangements are strongly perturbed. However, the region with LRS in the (001) device was only moderately reduced and no significant changes in the crystal structure were observed. The reversibility of the RS phenomenon is strongly dependent on the crystallographic orientation of the TiO 2−x crystal; the superior characteristics of the (001) device with respect to the (100) device are attributed to suppression of the formation of conductive filaments in the RS zone. As also suggested in the previous report 12 , the feasibility of resistance control by modification of the oxygen vacancy distribution is expected to lead to the implementation of multilevel resistance in memristive devices. This functionality is particularly required for realizing continuous weight of synaptic connection in devices for neuromorphic computing. Such functional design will be investigated in the near future. In addition, the unique characteristics of complementary RS in the four-terminal memristive device provide an opportunity to not only develop gate-tuning memristive devices but also to pave the way for new types of memory architectures.",
"introduction": "Introduction Resistive switching (RS) in memristors has attracted a great deal of attention for application to nonvolatile memory devices and neuromorphic computing 1 – 8 . In one class of memristors, it is typically composed of a metal/transition-metal-oxide/metal structure, in which the transition metal oxide can function as a conductive or resistive layer, depending on the applied voltage or the current injected into the device. The RS properties of memristors are required to be well controlled and the devices must exhibit reliable reversibility. To date, the mechanism for the RS phenomenon has been explained based on the generation and redistribution of oxygen vacancies in the oxide caused by application of an external voltage 2 , 9 – 21 . As-fabricated memristors generally have high resistance and often require an electroforming process where a conductive filament is preformed to lower the resistance locally in the oxide. Oxygen vacancies play a key role in modulating the conductive filaments, and their distribution is considered to determine the high or low resistance states (HRS or LRS) of the device. Non-filamentary, homogeneous RS has also been demonstrated based on the bulk ion conductivity of an oxygen-deficient amorphous main-group oxide thin film 22 . The RS behavior has been well explained by considering coupled ion drift and diffusion motion, and the oxygen concentration profile. Reduced titanium dioxide (TiO 2−x ) is a typical RS material for use in memristors 1 , 2 , 10 . The filamentary mechanism is arguably the most commonly accepted explanation for the RS phenomenon in TiO 2−x 2 , 23 , 24 , which is closely related to the local phase transformation that creates oxygen-deficient compounds, such as rutile TiO 2 crystals with polytype shear plane defects or Magnéli phases in TiO 2 2 , 25 . However, with regard to memristors driven by the filamentary RS mechanism, it is frequently argued that the formation of such conductive filaments is rather stochastic, has less controllability, and is more or less accompanied by other physical changes in the device structure, such as blow-off or massive redistribution of electrodes 24 – 28 , thereby posing a serious problem in the practical use of such devices. Therefore, non-filamentary-type RS operation based on modulation of the bulk carrier conduction is essential for the development of memristors with high stability, reliability, and controllability. In this study, we demonstrate precise control of the oxygen vacancy distribution and related bulk carrier conduction properties in non-filamentary-type memristive devices with a four-terminal structure fabricated on TiO 2−x single crystal substrates. In this device, a pair of diagonally opposing electrodes is used to modify the oxygen vacancy distribution in the electrically active zone between another pair of diagonally opposing electrodes. Although the size of this device is on the order of several tens of micrometers, which is not as small as conventional two-terminal memristors, our strategy is to embrace this size and explore modification of the oxygen vacancy distribution in the electrically active zone of the device. Microscopic visual observations of the oxygen vacancy distribution were conducted based on the electrocoloring effect in TiO 2−x 29 , in which regions with higher concentrations of Ti 3+ appear dark blue, indicating the presence of oxygen vacancies 30 . This allows the motion of oxygen vacancies under a DC electric field to be traced during device operation. The correlation between the modulation of the oxygen vacancy distribution in the electrically active zone and the RS properties (HRS and LRS) was investigated.",
"discussion": "Results and Discussion Figure 1 shows reflection-mode optical micrographs of our four-terminal memristive devices fabricated on (100) and (001) substrates. The terminals are labeled T1 to T4, with the T2 and T4 terminals being used to modify the oxygen vacancy distribution in the region between the T1 and T3 terminals. Figure 1 Optical micrographs showing four-terminal planar device structures fabricated on reduced rutile TiO 2−x ( a ) (100) and ( b ) (001) surfaces. The basic protocol for electrical measurements using the device consisted of write and non-disturbing read voltage application steps. First, the current flow between T1 (T2) and T3 (T4), I 1–3 ( I 2–4 ), was measured at an applied voltage of 1 V to probe the initial state. A constant write voltage V C , was then applied to both T2 and T4 for a certain duration T C , under the condition that both T1 and T3 were grounded. Subsequently, I 1–3 ( I 2–4 ) was again measured at an applied voltage of 1 V to probe the state of the device. This protocol was repeated while the values of V C and T C were varied. Figure 2 shows typical sequences of voltage application and current measurements for the device. As a proof-of-concept demonstration to evaluate V C -driven control of the oxygen vacancy distribution, the sequence shown in Fig. 2(a) was employed. On the other hand, the sequence shown in Fig. 2(b) was used to evaluate the repeatability of RS, where the V C value was varied to generate a ramp-up and ramp-down waveform. Figure 2 Typical sequences for measurement of RS properties of devices. ( a ) Diagram illustrating the controllability of the oxygen vacancy distribution in the electrically active zone of the devices and ( b ) the reversibility of the RS properties in the devices. We first examined the interplay between the resistance change and the modulation of the oxygen vacancy distribution in the (100) device. Figure 3(a,b) respectively show typical results for the variations of I 1–3 and I 2–4 as a function of the cumulative time T C for application of the write voltage V C according to the sequence of Fig. 2(a) . The initial values of I 1–3 and I 2–4 were almost the same, as expected; however, after application of V C = 8 V for 500 s, I 1–3 increased while I 2–4 decreased. This indicates that the resistance in the region between T1 (T2) and T3 (T4) was changed due to the application of V C . Both the current values then remained constant until 950 s, even though the polarity of V C was changed and the value was negatively and stepwise increased from −1 V to −8 V, as shown in Fig. 2(a) . A significant decrease (increase) of I 1–3 ( I 2–4 ) occurred after application of an additional −8 V for 50 s (at cumulative time of 1000 s), and the resultant I 1–3 ( I 2–4 ) value was slightly lower (higher) than the initial value. Subsequent high and low (low and high) values of I 1–3 ( I 2–4 ) were obtained after the application of V C at +8 and −8 V, respectively, which indicated a reversible change of the resistance in the region between diagonally opposing electrodes in the device. Figure 3 Variations of ( a ) I 1–3 and ( b ) I 2–4 at 1 V as function of cumulative time T C for application of voltage V C . (Inset) Labels a–f show measurement stage of I 1–3 and I 2–4 at cumulative times of 0, 500, 950, 1000, 1100, and 1150 s, respectively. To obtain insight into the correlation between this alternate current variation and the oxygen vacancy distribution, the internal structure of the device was inspected using optical microscopy. In reduced TiO 2 , introducing oxygen vacancies produces trivalent titanium ions (Ti 3+ ) that act as color centers. Therefore, change in the configuration of electro-coloring regions should reflect the change in the distribution (local concentration) of Ti 3+ ions and oxygen vacancies, which is considered to be caused by drift motion of positively charged oxygen vacancies by an applied electric field. Figure 4(a–f) depict a series of representative transmission-mode optical micrographs of the device subjected to V C for cumulative times of 0, 500, 950, 1000, 1100, and 1150 s, respectively, the observation stages of which are marked by red solid circles in Fig. 3 . At the initial stage, the TiO 2−x area was faintly colored light blue, as shown in Fig. 4(a) , after which an area colored dark blue appeared to bridge T1 and T3 while the other areas became colorless after application of V C = +8 V for 500 s (Fig. 4(b) ). This type of coloration in the TiO 2−x crystal is attributed to the electrocoloring phenomenon and the configuration of the electrocolored areas is modified by the application of an external voltage to the device. Therefore, when a positive voltage of V C is applied to T2 and T4 under the condition where T1 and T3 are grounded, positively-charged oxygen vacancies are likely repelled from T2 and T4, and then accumulate around T1 and T3 due to drift motion induced by the electric field. The region colored dark blue contains a higher concentration of oxygen vacancies and is likely more conductive, whereas the colorless region has a lower concentration of oxygen vacancies and is less conductive. Thus, the configuration with T1 and T3 connected to each other through a zone with a high concentration of oxygen vacancies, i.e., high conductivity, accounts for the observed increase in I 1–3 . In contrast, T2 and T4 are observed to be surrounded by colorless areas; therefore, the resistance between them is determined by resistive areas with less conductivity, which results in a decrease in I 2–4 compared with the initial value. Figure 4 Representative optical microscope images showing internal structure of (100) device corresponding to measurement stage of I 1–3 and I 2–4 at cumulative times of T C = ( a ) 0, ( b ) 500, ( c ) 950, ( d ) 1000, ( e ) 1100, and ( f ) 1150 s. This type of bridge structure was then maintained until a time of 950 s (Fig. 4(c) ), which is consistent with the lack of significant changes in I 1–3 and I 2–4 shown in Fig. 3 . In contrast, as shown in Fig. 4(d) , application of an additional −8 V for 50 s (cumulative time of 1000 s) made this structure completely inverted, where a colorless area bridged T1 and T3, while the other areas were colored dark blue. These features are also consistent with the observed I 1–3 decrease; T1 and T3 are connected through a resistive region with a low oxygen vacancy concentration. On the other hand, the observed increase in I 2–4 is likely due to a substantial increase in the areas where the T2 and T4 Pt electrodes are observed to be surrounded by a dark blue conductive region. These changes at the cumulative time of 1000 s look rather abrupt and discontinuous to be brought by the redistribution of oxygen vacancies under the constant voltage application. This intriguing behavior of abrupt switching is difficult to understand solely by considering field driven drift motion of the vacancies. One possible explanation for this is increased sample temperature by Joule heating 2 , 26 , which may strongly affect (promote) the migration and redistribution of vacancies after some incubation time. After subsequent applications of V C at + 8 V and −8 V, as shown in Fig. 4(e,f) , the contrast was inverted repeatedly, displaying features of dark blue and faint blue T1–T3 bridges, which correspond to the measured high and low I 1–3 values, respectively. It should be noted that the colored structures become complicated and the degree of contrast inversion is reduced. The remaining faint blue contrast reflects the slightly higher I 1–3 value measured at 1150 s compared with that at 1000 s, as shown in Fig. 3(a) . Next, to examine the RS reversibility in the four-terminal devices, electrical measurements were implemented by applying the sequence shown in Fig. 2(b) . Figure 5(a,b) ) shows representative variations of resistance between T1 (T2) and T3 (T4), R 1–3 ( R 2–4 ) as a function of the applied V C in the device fabricated on the TiO 2−x (100) substrate. Here, the maximum and minimum V C values were set to +8 V and −8 V, respectively, with a T C of 100 s and a cycle where positive and negative V C were applied was repeated three times. Note that once the R 1–3 value is decreased, the R 2–4 value is increased, and vice versa when the polarity of V C is inverted, which indicates that the variations dependent on V C are complementary to each other. Specifically, in the first cycle with an increasing V C value, R 1–3 ( R 2–4 ) is gradually decreased (increased) to the value at the positive maximum V C , so that the LRS (HRS) is achieved. The LRS (HRS) is stable while decreasing V C but a sharp increase (decrease) in R 1–3 ( R 2–4 ), a so-called RESET (SET) operation of the device, occurs at −7 V, so that the HRS (LRS) is achieved. The HRS (LRS) is then stable again while increasing V C . In the second and third cycles, the initial values of R 1–3 and R 2–4 start from the same values as those measured last in the previous cycle, and the trends of the R 1–3 and R 2–4 variations are almost similar to the first cycle. However, the maximum values of R 1–3 and R 2–4 in each cycle are observed to decrease with the measurement cycle, and this trend is much more obvious for R 1–3 than for R 2–4 . Figure 5 ( a,b ) Variations of resistance between T1 (T2) and T3 (T4) as function of applied voltage V C in device fabricated on TiO 2−x (100) substrate. ( a ) R 1–3 and ( b ) R 2–4 . ( c , d ) Optical micrographs showing internal structure of (100) substrate device surface. Images were obtained at V C = ( c ) +8 V and ( d ) −8 V during the second cycle measurement. Solid and dotted arrows indicate positions of dark lines elongated along [001] and <011> directions, respectively. Microscopic observation of the device was performed to clarify the origin of this anomalous resistance change during the cyclic application of V C . At the stages of V C = +8 V and −8 V in the first cycle, the characteristic oxygen vacancy distribution (data not shown) was confirmed to be quite similar to those in Fig. 4(b,d) , respectively. In contrast, the internal structures of the device subjected to the second and third cycles did have distinct features. Figures 5(c,d) show transmission-mode optical micrographs of the device at the stages of V C = +8 V and −8 V in the second cycle, respectively. Comparison with the first cycle case indicates the presence of additional dark line contrast elongating mainly along the [001] direction, as well as along <011> directions in the zone bridging between T1 and T3 (see arrows in Fig. 5 ). In particular, these dark lines still remain in the T1-T3 bridge, as shown in Fig. 5(d) , although the dark blue behind them becomes faint due to the application of negative V C . Scanning transmission electron microscopy (STEM) analysis demonstrated that these dark lines are closely related to shear plane defects in the TiO 2−x matrix, which result from a local accumulation of excess oxygen vacancies. Figure 6 shows cross-sectional high-angle annular dark field (HAADF)-STEM images acquired from the electrically active zones in the four-terminal devices. As shown in Fig. 6(a) , high-contrast bright planar structures (PS) were observed in cross-section extending along 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}$$\\bar{1}00$$\\end{document} 1 ¯ 00 ] depth direction, as well as the [010] direction in the vicinity of the surface of the (100) plane device, both of which correspond to the dark lines shown in Fig. 5 . The selected area electron diffraction (SAED) pattern for the PS shows diffuse streaks along the <100> and <010> directions, as well as spots that correspond to single-crystal rutile TiO 2−x , which indicates the existence of stacking disorders along these directions (Fig. 6(a) inset). This was verified from the high resolution HAADF-STEM image shown in Fig. 6(c) , which shows the atomic arrangements locally disturbed in the form of shear plane defects 31 in the single-crystal matrix. The Ti-L edge electron energy loss spectrum (EELS) acquired at the PS in the (100) device (PS100) is shown in Fig. 6(e) . The spectrum for the PS shows a clear difference from that for the as-reduced TiO 2−x (AR001), especially in the shape and position of the L 3 peaks. In addition, both the onset energy of the Ti-L 2,3 edges ( E on ) and the energy splitting between the e g and t 2g orbital peaks in the L 3 edges (Δ E ) for the PS show smaller values than those for AR (Fig. 6(f) ), which suggests that the TiO 2−x phases in the PS turn into a further reduced state. Figure 6 ( a – d ) Cross-sectional HAADF-STEM images and SAED patterns (insets) for electrically active zones in TiO 2−x four-terminal devices. ( a , c ) (100) device and ( b , d ) (001) device. Red arrows in ( c ) show the stacking disordered regions. ( e ) Ti L 2,3 edge EELS spectra of electrically active zones in the different devices. ( f ) Energy onset E on , and e g -t 2g peak splitting Δ E , of Ti L 3 edge. AR001, PS100, and DB001 denote as-reduced TiO 2−x , PS in (100) device, and dark-blue regions in (001) device, respectively. These changes in the atomic and electronic structures possibly occur as a result of oxygen vacancy accumulation caused by application of an external voltage to the device. In TiO 2 -based memristors, shear plane defects or the Magnéli phases are well known to act as highly conductive current paths, i.e., a conductive filament 24 , 25 . Furthermore, recent studies using density functional theory have predicted that the ordered arrangement of oxygen vacancies along the [001] direction is thermodynamically favorable and forms a conductive filament in rutile TiO 2 32 , 33 . Thus the observed PS possibly reflects current leakage paths that are additionally formed in the T1–T3 bridge during application of the positive V C . Such leakage paths are not decomposed, even after application of the negative V C , so that the resistance of the T1–T3 bridge gradually decreases as the V C cycle is repeated. This phenomenon eventually tends to deteriorate the controllability of the RS properties. Detailed analytical results of the crystallographic and electronic structures in the electrically active zone of these TiO 2−x memristive devices are also reported elsewhere 34 . The (100) substrate case shows the crude reversibility of RS against cyclic V C application, which is mainly attributed to the irreversible structural change in the electrically active zone. In contrast, we have confirmed that much higher RS reversibility can be obtained for the device fabricated on the (001) substrate. Typical results for the variation of R 1–3 and R 2–4 as a function of the applied V C in the (001) substrate device are shown in Fig. 7(a,b) , respectively. Although some peculiar changes in both resistances were observed with an increase in V C at the initial stage of the first cycle, the variation of R 1–3 ( R 2–4 ) exhibits an orderly clockwise (counterclockwise) rotation as the V C cycle is repeated. Furthermore, the threshold voltages for SET and RESET operations can be unambiguously determined — ±4 V in this case — and the maximum and minimum values of the resistances did not change significantly, even after the third cycle. These characteristics explicitly demonstrate the superiority of the (001) substrate device to the (100) device in terms of the stability of the RS properties. Figure 7 ( a , b ) Variations of resistance between T1 (T2) and T3 (T4) as function of applied voltage V C , in device fabricated on TiO 2−x (001) substrate. ( a ) R 1–3 and ( b ) R 2–4 . ( c , d ) Optical micrographs showing internal structure 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}$$(001)$$\\end{document} ( 001 ) device surface. Images were obtained at V C = ( c ) +6 V and ( d ) −6 V during the first cycle measurement. Inspection of the oxygen vacancy distribution was also conducted and the results are shown in Fig. 7(c,d) , where the inner structures are displayed for the device at the stages subjected to application of V C at +6 V and −6 V in the first cycle, respectively. The interfaces between the dark blue and colorless areas are much broader than those evident for in the (100) device shown in Fig. 4 . Figure 7(c) also shows that the dark blue region is constricted at the center between T2 and T4 by a diluted blue color. Comparison of this feature and those shown in Fig. 4 clearly indicates that the drift and diffusion behavior of oxygen vacancies is significantly dependent on the crystal orientation of TiO 2−x and the direction of the electric field applied to the device. The oxygen vacancy distribution shown in Fig. 7 accounts for the observed current variations, i.e., the formation of the dark-blue-colored region bridging T1 and T3 at V C = +6 V corresponds to the lowest R 1–3 (highest R 2–4 ) measured, and this configuration is flipped to bridge between T2 and T4 at V C = –6 V where the highest R 1–3 (lowest R 2–4 ) was measured. This type of pattern flipping was confirmed to be repeated in the second and third cycles. The irreversible structures, such as the dark line contrasts as observed in the (100) device, are not formed in the electrically active zone in this case. The absence of distinct structural changes in the crystal was also confirmed by STEM observations. The low magnification STEM image in Fig. 6(b) shows no significant features in the dark-blue-colored region of the electrically active zone of the (001) device (DB001). The SAED pattern and the high-resolution STEM image confirmed a single-crystal rutile structure, as shown in Fig. 6(b,d) . These observations are in clear contrast to the case for the (100) devices where strong perturbation in the crystal structure was observed. However, the Ti-L edge EELS spectrum acquired for the DB001 device did indicate some changes compared with those for as-reduced TiO 2−x : the onset energy E on and the L 3 peak splitting Δ E have lower values than those for the AR001 device (Fig. 6(f) ), which indicates that the region is locally more reduced due to voltage application. It was thus deduced that this change in the electronic structure without a distinct change in the crystalline structure is the origin of the reversible RS characteristics in the (001) device. Finally, a demonstration of the repeatability of the RS phenomenon, i.e., an endurance test for the memristive device fabricated on the (001) substrate is presented in Fig. 8 . LRS and HRS can be switched more than 150 times in a stable manner, which ensures the stable operation of memristive devices based on carrier conductivity modulation controlled by modification of the oxygen vacancy distribution in the TiO 2−x crystal substrate. Concerning the reproducibility of the data, we confirmed that all the results presented here are reproducible in several times experiments. While at the same time, we observed some scattering and deviations in detailed behavior of individual devices, which may be attributed to local inhomogeneity of oxygen vacancy concentration caused by the initial crystal reduction process. One possible remedy for this is more precise control of the crystal (non-)stoichiometry based on thin film processes, and work along these lines are currently underway. Figure 8 Endurance test result for (001) substrate device up to 180 cycles. V C of ±6 V with T C of 5 s was applied to switch the device between LRS and HRS."
} | 6,769 |
29868643 | PMC5983913 | pmc | 3,052 | {
"abstract": "AI-2–mediated quorum sensing of E. coli is directly connected to sugar metabolism via the phosphocarrier HPr protein of PTS.",
"introduction": "INTRODUCTION Bacteria, despite being unicellular organisms, are capable of coordinating population-level behavior through a process termed quorum sensing (QS). In this process, bacteria secrete chemical signaling molecules called autoinducers (AIs), which accumulate as cell density increases. Once the AI level reaches a threshold, signaling a “quorum” of cells, the AI signals are transported intracellularly, where they activate gene expression and enable coordinated phenotypic responses in the population. The importance of QS in biofilm formation and maintenance ( 1 ), bacterial persistence ( 2 ), and pathogenicity ( 3 ) has appeared in many review articles. Further, the interplay between the signaling components that comprise QS systems has been the subject of many studies in metabolic engineering and synthetic biology, wherein genetic networks have been developed to enable “programmed” design and control of metabolic pathways and bacterial phenotype ( 4 , 5 ). For example, researchers have exploited QS circuitry for the design and implementation of “smart” bacteria that target and destroy cancers and pathogens ( 6 , 7 ). Some of these systems depend on the “orthogonality” of the signaling system with the metabolic activity of the host organism. For instance, Saeidi et al . ( 8 ) engineered bacteria that sense acyl-homoserine lactone QS signals produced by Pseudomonas aeruginosa and subsequently release toxins to eradicate the pathogen. Other systems rely on the interdependence of QS activity and host metabolism. For example, an autoinduction system was constructed by Tsao et al . ( 9 ), in which Escherichia coli secrete AI-2 [dihydroxy-pentane-dione (DPD)] and, at the appropriate time for gene expression, self-induce expression of a recombinant protein by amplifying expression from the native lsr (LuxS-regulated) promoter using T7 polymerase. Further mechanistic understanding of how the cell regulates QS processes, for example, based on the availability of substrates like glucose, will further enable researchers to exploit these QS systems for the design of synthetic biology systems with new capabilities. There is evidence that the AI-2–mediated QS system is partially regulated by substrate availability and cell metabolism. LuxS synthesizes AI-2 as a by-product of the activated methyl cycle ( 3 ), after which AI-2 accumulates extracellularly. AI-2 is imported by LsrACDB ( 10 ) and phosphorylated by the kinase LsrK, sequestering it within the cell ( 11 ). Phosphorylated AI-2 relieves LsrR-mediated repression of the lsr operon ( 12 ), allowing transcription of the lsr genes and acceleration of AI-2 uptake. It has recently been reported that another key bacterial process, chemotaxis, is linked to AI-2 signaling via LsrB and enhances not only self-aggregation of E. coli but also its coaggregation with Enterococcus faecalis ( 13 , 14 ). Several studies suggest that the bidirectional lsr operon, in addition to being regulated by LsrK and LsrR, is also subject to carbon catabolite repression (CCR). For instance, activation of the lsr promoter does not occur in the presence of glucose ( 10 ) or glycerol ( 15 ) and requires the global regulators cyclic adenosine monophosphate (cAMP) and cAMP receptor protein (CRP) ( 10 ). Binding sites for cAMP-CRP exist between the lsrR and lsrACDB promoters ( 12 ), and binding of cAMP-CRP likely modulates the promoter activity ( 16 ). Here, we propose a new mechanism linking cell metabolism to the AI-2 QS system. Specifically, we have discovered that the activity of LsrK is regulated by the phosphoenolpyruvate (PEP)–dependent sugar phosphotransferase system (PTS) protein HPr. PTS is important for sugar uptake and regulation of carbohydrate metabolism. It comprises three units—EI, HPr, and the EII protein complex—that sequentially transfer a phosphoryl group from PEP to the transported carbohydrate. The active transport of PTS sugars affects the phosphorylation state of each of the PTS components. Although EI and HPr are general PTS proteins, EII is specific to the carbohydrate being transported, and one of the most commonly studied is a subunit involved in glucose transport, EIIA Glc . The phosphorylation state of EIIA Glc regulates the activity of adenylate cyclase, which synthesizes cAMP, a global regulator within the cell ( 17 ). As discussed above, the cAMP-CRP complex regulates transcription from the lsr promoter. There is evidence to suggest that the phosphorylation state of the other general PTS proteins also regulates AI-2 QS activity. Pereira et al . ( 18 ) demonstrated that phosphorylated EI is required for the initial uptake of AI-2, although the mechanism behind this has not been elucidated. Our research demonstrates that LsrK is tightly bound to the PTS protein HPr, and that HPr directly influences LsrK activity. LsrK was previously demonstrated to have a range of substrate specificities ( 19 ), perhaps illustrating this enzyme’s involvement in additional functions besides QS. Moreover, we demonstrate that uptake of PTS carbohydrates has direct involvement in signaling and show for the first time that PTS regulates AI-2 QS not only through the global regulator cAMP but also directly through specific interactions with LsrK. This finding suggests that bacteria have evolved sophisticated mechanisms for incorporating information about substrate availability and cell metabolism into QS processes. This discovery is of fundamental importance, as phenomena such as pathogenicity, bacterial persistence, and biofilm formation have previously been shown to be influenced by nutrient availability and QS. Here, we suggest that these processes may be more closely linked than previously thought.",
"discussion": "DISCUSSION The presence of de novo regulation of LsrK mediated by HPr was inductively revealed from structural studies of the LsrK/HPr complex. That is, we discovered that the yield of purified LsrK using the previous complicated expression methodology ( 21 ) seemed to be critically dependent on the in vivo quantity of the HPr protein. We showed that the structure of the His LsrK/HPr complex is likely to be an inactive form, as the absence of DPD in the His LsrK/HPr crystals grown with both ATPγS and DPD is, in part, explained by the following: (i) Crystallization of the His LsrK protein already contained the HPr protein, and (ii) HPr is more tightly bound to ATP:LsrK ( K i 1 , 0.08 ± 0.01 μM) than to ATP:LsrK:DPD ( K i 2 , 0.27 ± 0.03 μM). Although the exact mechanism by which HPr inhibits LsrK activity is still unknown, the absence of DPD in the His LsrK/HPr complex structure and the 1D STD NMR data of DPD in the presence of the His LsrK/HPr complex also suggest that a structural change of domain I induced by HPr binding is not adequate for recruiting the DPD molecule. The in vivo ratio of HPr and p-HPr is closely related to the rate of sugar uptake, and we note that about 25-fold lower inhibition of LsrK activity by p-HPr, compared to unphosphorylated HPr, likely enables the QS process to be dependent on the E. coli sugar metabolism. Although a strict steric hindrance during protein-protein interactions generally results in an all-or-none response, the less stringent steric hindrance imported by the phosphorylated H15 during the interaction between LsrK and HPr likely enables fine-tuning of the AI-2–mediated QS according to both in vivo concentration and the ratio of HPr and p-HPr. Different K m values of LsrK proteins from E. coli and Salmonella typhimurium LT2 The previously reported K m values of hXK ( 31 ) for d -xylulose and of ecGK ( 29 ) for glycerol are 24 ± 3 and 16 ± 5 μM, respectively, and the k cat of hXK is 35 ± 5 s −1 . The kinetic parameters of the Salmonella typhimurium ( S. typhimurium ) LT2 LsrK protein for DPD ( K m , 1.0 ± 0.2 mM; k cat , 7.6 ± 0.6 s −1 at 0.5 mM ATP) ( 22 ) are different from our kinetic values of the E. coli LsrK ( K m , 58 ± 11 μM; k cat , 17 ± 1 s −1 at 0.75 mM ATP). Perhaps, the higher K m values noted for the S. typhimurium LT2 LsrK might be a result of the assembly of an LsrK/HPr complex, because (i) the K m values of GK and XK for their substrate sugars are similar to those of the E. coli LsrK for DPD, (ii) the apparent effect of HPr binding mimics the increased K m value of LsrK activity (−1/ K m is the x intercept in the Lineweaver-Burk plot) ( Fig. 3A ), and (iii) the HPr-binding sequence of S. typhimurium LT2 is almost identical to that of E. coli ( Fig. 2 ). On the other hand, the different K m values of LsrK between E. coli and S. typhimurium likely suggest that a critical concentration of AI-2 for the initiation of the AI-2–mediated QS signaling may vary with different bacterial species. Resolution to these two possibilities requires further studies. QS is integrated to sugar metabolism via HPr of PTS PTS is a key bacterial system that monitors the available sugars in the environment, regulating inherent metabolic processes. The phosphorylation status of the PTS components plays an important role via protein-protein interactions. The phosphorylation state of EIIA Glc is a key determinant for the CCR, also known as glucose effect in E. coli ( 17 ). In the CCR of Gram-positive bacteria, phosphorylation at S46 of HPr plays a central role ( 33 ). As HPr-S46 is located in the binding surface for LsrK, we suggest that inhibition of LsrK activity by HPr carrying the phosphorylated S46 could be abolished ( Fig. 1D ), perhaps indicating differences in QS regulation among Gram negatives and Gram positives. Although the dephosphorylated EIIA Glc directly inhibits many metabolic enzymes (adenylate cyclase, GK, lactose permease, and maltose permease) through the direct protein-protein interaction ( 17 , 34 ), there have been several reports for the interaction between HPr and non-PTS proteins in E. coli : (i) Both HPr and p-HPr bind glycogen phosphorylase (GP), but only HPr binding appreciably increases GP activity ( 35 ); (ii) HPr, but not p-HPr, directly antagonizes the activity of Rsd, a negative regulator of the primary sigma factor (σ 70 ), which enables the σ 70 -RNA polymerase holoenzyme to occupy the promoter regions of various housekeeping genes ( 34 ). The involvement of HPr, not EIIA Glc , in the regulation of GP and Rsd (anti-sigma factor) seems to be more reasonable, because these two functions are not sugar-specific but, instead, are global regulatory functions of E. coli . The in vivo concentration of HPr is also higher than that of EIIA Glc in E. coli ( 36 ). Because the AI-2–mediated QS signaling accompanies the changes of bacterial global regulation, it seems reasonable that the upstream PTS component, HPr, would be a better regulator than EIIA Glc . A model for LsrK/HPr interplay between QS and PTS At low cell density, secreted AIs diffuse away and otherwise do not accumulate. Rather, as cells grow and continuously secrete AIs, the accumulative production of AIs at high density initiates their detection and consecutive responses. Individual cell behaviors, in turn, become properties of the collective. This transition, called QS, enables persistence in biofilms and many other phenotypes. Abstractions of the signal transduction mechanisms have led to the design and implementation of many application-specific functions. The synchronous responses of bacteria at high density therefore provide great benefits for concerted QS signaling. QS is generally thought to be under control of transcriptional regulation, exemplified by bacteria-specific QS phenomena such as bioluminescence, biofilm formation, and virulence factor secretion. The inhibition of LsrK by HPr discovered here provides another regulatory mechanism of QS. It is particularly noteworthy that regulation of enzyme activities typically occurs at much faster characteristic times than regulation of activities involving gene expression. This will have ramifications on the design of new synthetic biology constructs. We depict a simplified view of the regulatory structure of LsrK and HPr and their involvement in PTS and QS ( Fig. 7 ). That is, the phosphorylation status of HPr enables the QS of bacteria to be directly linked to their metabolic state; the previous report showing the involvement of ptsI in the import of AI-2 and the transcription of lsr operon ( 18 ) can now be explained by the direct interaction of HPr and p-HPr with the LsrK protein. The de novo regulation of LsrK activity by HPr and p-HPr that uses a protein-protein interaction and, at the same time, enables self-identification of a bacterium’s metabolic state could provide additional ways to synchronize the responses of bacteria to AI-2 and, more importantly, to differentiate the QS of certain bacteria in mixed microbial flora according to their metabolic states. Fig. 7 Summarized model for the role of HPr between two distinct bacterial processes: QS and PTS. QS signaling should be strongly related to a bacterial metabolic status, because QS accompanies a transition from planktonic to sessile growth. General transcriptional regulation is under control of σ 70 and σ S during the exponential growth and stationary phase, respectively. In the PTS sugar transport system, HPr delivers the activated phosphate group from EI to EIIA and then the phosphate is eventually transferred to imported glucose (PTS sugars) via the transporter (EIIB and EIIC). The native HPr is dominant during the exponential growth, but p-HPr is accumulated in the stationary phase largely due to the absence of sugar import. The K i 1 (0.08 μM) and K i 2 (0.27 μM) values of the HPr protein were very low, and thus, LsrK inhibition by HPr is likely to be released during the late stationary phase when the accessible HPr is extremely limited. In addition, at these later times, AI-2 is taken up by the Lsr transporter and more LsrK is expressed. In the current scenario, LsrK activity is modulated directly via glucose metabolism and by transcriptional regulation via AI-2 QS. The former is presumably far quicker than the latter, indicating that QS activity and sugar metabolism are tightly linked, and hence, population-scale behavior can be influenced rapidly by rapid changes in nutrient availability within a particular niche."
} | 3,617 |
33093542 | PMC7581756 | pmc | 3,053 | {
"abstract": "Trophic downgrading in coastal waters has occurred globally during recent decades. On temperate rocky reefs, this has resulted in widespread kelp deforestation and the formation of sea urchin barrens. We hypothesize that the intact kelp forest communities are more spatially variable than the downgraded urchin barren communities, and that these differences are greatest at small spatial scales where the influence of competitive and trophic interactions is strongest. To address this, benthic community surveys were done in kelp forests and urchin barrens at nine islands spanning 1230 km of the Aleutian Archipelago where the loss of predatory sea otters has resulted in the trophic downgrading of the region’s kelp forests. We found more species and greater total spatial variation in community composition within the kelp forests than in the urchin barrens. Further, the kelp forest communities were most variable at small spatial scales (within each forest) and least variable at large spatial scales (among forests on different islands), while the urchin barren communities followed the opposite pattern. This trend was consistent for different trophic guilds (primary producers, grazers, filter feeders, predators). Together, this suggests that Aleutian kelp forests create variable habitats within their boundaries, but that the communities within these forests are generally similar across the archipelago. In contrast, urchin barrens exhibit relatively low variability within their boundaries, but these communities vary substantially among different barrens across the archipelago. We propose this represents a shift from small-scale biological control to large-scale oceanographic control of these communities.",
"introduction": "Introduction Trophic downgrading occurs when apex predators have been extirpated over large geographic regions, which can lead to important consequences for ecosystem functioning due to both direct and indirect cascading effects 1 . This has been observed globally across a variety of terrestrial and marine ecosystems 2 – 7 . Often, trophic downgrading triggers increases in herbivore populations, thereby changing overall community structure 6 , 8 , 9 and altering patterns of ecosystem productivity 10 – 12 . Trophic downgrading can be especially important if it ultimately affects ecosystem engineers that provide habitat, which modifies the physical environment, regulates primary production and energy flow, and generally supports high biodiversity. For example, the extirpation of gray wolves from Yellowstone National Park, USA in the early 1900s resulted in reduced predation on elk and increased herbivory on forest-forming trees 13 . This ultimately led to changes in the morphology and hydrology of the region’s river systems and its riparian plant communities 14 , 15 . Similarly, the loss of sea otters from the nearshore habitats of the Aleutian Archipelago during the 1980s and 1990s resulted in reduced predation on herbivorous sea urchins and a subsequent overgrazing of the regions kelp forests 2 . This led to reduced biodiversity 16 , altered food web dynamics 17 , and reduced benthic ecosystem productivity 12 in the coastal environment. Further, when a community has been downgraded, its resilience (i.e., recovery and stability after a disturbance) can decrease in comparison to intact communities 3 . This has been an important consideration in the design of marine protected areas (MPAs) and terrestrial parks and reserves, which typically protect apex predators to maintain diversity and normal ecosystem functioning 7 , 18 – 20 . Spatial and temporal variability in community structure are important components of ecological systems, and understanding how variability changes in space and time can infer a wide range of ecological processes 21 – 27 . For example, resistance to biological invasions is strongly correlated with variability in environmental conditions 28 and community structure 29 , with less variable communities being more resistant to invasion than highly variable communities. However, the loss of foundation species can increase susceptibility to biological invasions 30 , 31 . Moreover, patterns of variability can themselves change at different temporal and spatial scales coincident with environmental and demographic forcing factors 21 , 25 , 26 , 32 , 33 . Indeed, Levin 21 noted that the problem of pattern and scale is the central problem in ecology and that it is important to find ways to quantify patterns of variability in space and time, understand how patterns change with scale, and understand the causes and consequences of pattern. This can be especially important in ecosystems where different forcing factors affect communities across a range of spatial scales 33 – 35 . Kelp forests are benthic, biogenic habitats that include highly productive primary producers rivaling those of cultivated agricultural fields and tropical rainforests in productivity 36 – 39 . This productivity and the associated formation of complex, three-dimensional biogenic habitat enhances local biodiversity and secondary productivity 16 , 40 , 41 . However, kelp forests in many areas of the world have been trophically downgraded as their apex predators have been removed 42 , 43 . For example, in Tasmania, the loss of predatory lobsters has led to increases in sea urchin abundance and an increased risk of catastrophic shifts to widespread sea urchin barrens 44 . Such deforestation of kelp forests due to sea urchin grazing is becoming more common in mid-latitudes 45 – 47 (see also citations in Steneck et al. 9 ). This generally results in a loss of biodiversity and associated ecosystem services, as kelp forests generally support more species 16 , 48 (this study) and exhibit greater primary productivity and habitat complexity 12 , 49 , 50 than sea urchin barrens. Consequently, we expected there would be more combinations of species that could spatially differentiate the kelp forest communities than the sea urchin barren communities. This would be especially important at small spatial scales where the influence of biological interactions (e.g., competition, grazing, and predation) in spatial structuring of communities can be strongest and potentially obscure large-scale variability due to climate or oceanographic variability 51 – 55 . The shallow subtidal regions of the Aleutian Archipelago (Fig. 1 ) are a well-studied simple system under top-down control 56 . Here, sea otters that were previously hunted to near extinction by the fur trade began to recover in the early 1900s 57 . This recovery continued through the 1980s when the removal of sea otters by killer whales became apparent by the 1990s 56 , 57 . Following the removal of otters, the abundance and biomass of their primary prey, herbivorous sea urchins, dramatically increased 2 . These hyper-abundant sea urchins quickly overgrazed the macroalgal communities, causing widespread deforestation of the region’s kelp forests, and increases in the prevalence and extent of urchin barrens. These urchin barrens are largely devoid of all fleshy macroalgae but instead are dominated by sea urchins and coralline red algae 12 , 16 . They can be of varying ages depending on the timing of sea otter recovery and subsequent sea otter removal, but once formed they are stable over periods of at least several years 58 . Here, we use this trophically-mediated deforestation to examine spatial patterns of community variability in kelp forests and urchin barrens throughout the Aleutian Archipelago. We ask if non-downgraded communities with more species (i.e., intact kelp forests) exhibit greater spatial variability compared to downgraded communities with fewer species (i.e., urchin barrens) 16 . We then compare how these patterns of variability are distributed across different spatial scales, from meters to hundreds of kilometers. We do this for the whole communities and for different trophic guilds (i.e., primary producers, grazers, filter feeders, and predators). Based on previous observations we have made during numerous visits to the archipelago, we hypothesized that (1) the community state with higher biodiversity (intact kelp forests) will have greater overall spatial variability in community structure than the community state with lower biodiversity (downgraded urchin barrens), (2) patterns of variability will be greatest at small spatial scales for the kelp forests but greatest at large spatial scales for the urchin barrens, and (3) these patterns will be consistent among different trophic guilds. Figure 1 Map of study area showing the nine islands sampled across the Aleutian Archipelago (inset shows portion of archipelago where islands are located). Coordinates in decimal degrees for approximate sampling locations are: Attu: 52.92°, 173.20°; Nizki/Alaid: 52.74°, 174.00°; Kiska: 51.97°, 177.58°; Amchitka: 51.41°, 179.28°; Tanaga: 51.81°, − 177.94°; Adak: 51.87°, − 176.66°; Atka: 52.10°, − 174.69°; Yunaska: 52.66°, − 170.74°; and Chuginadak: 52.84°, − 169.75°. Image from Metzger et al. 2019.",
"discussion": "Discussion Trophically downgraded ecosystems are becoming more common worldwide, which is leading to conservation concerns due to the associated losses of biodiversity, productivity, and community resilience 1 , 9 . Downgrading often results in the formation of alternate stable states 4 , 6 , 8 in which herbivores become hyper-abundant and autotrophs become rare. When this affects the distribution and abundance of ecosystem engineers, such as forest-forming trees and kelps that create habitat and supply energy for their communities, ecosystem function can be altered 12 and biodiversity reduced 16 . Such ecosystem changes, however, can vary among different spatial scales due to a variety of forcing factors 21 , 24 , 32 , 33 . Here, we demonstrate that the intact kelp forests in the Aleutian Archipelago have more species than the trophically downgraded urchin barrens, and that this is consistent using both biomass and density data. Further, the intact kelp forest communities were most variable at the smallest spatial scale examined (among quadrats within each forest) and least variable at the largest spatial scale (among islands). This suggests that kelp forests create a variable habitat within their boundaries with high potential for species interactions (e.g., competition, grazing, predation), which are important drivers of small-scale variability 54 , 55 and which can mask the influence of large-scale climate or oceanographic factors 51 – 53 . These habitat characteristics are then repeated over larger spatial scales (i.e., different kelp forests typically contain similar species assemblages and biological interactions), which reduces larger-scale variability. Following trophic downgrading that resulted in reduced biodiversity as the kelp forests transitioned to urchin barrens 16 , the number of potential species interactions also declined. Although the timing of these changes varied among islands, the actual changes were consistent across the archipelago, which led to a transformation of the landscape 61 . This also led to changes in spatial variability in community structure and how it was distributed among the different spatial scales, but this depended on the type of data used. Specifically, when species abundances were estimated based on their densities, there was an overall reduction in total spatial variability in community structure at each spatial scale, and the barrens remained most variable at the smallest spatial scale and least variable at the larger spatial scales. In contrast, when species abundances were estimated based on their biomasses, variability in community structure again decreased at the smallest spatial scale, but it increased substantially at the larger spatial scales. This resulted in the urchin barren communities being least variable at the smallest spatial scale and most variable at the largest spatial scale. It also resulted in the urchin barrens exhibiting more total spatial variability in community structure than the kelp forests. These differences between the two data types appeared largely due to how each estimated the abundance of filter feeders and grazers (discussed below). Regardless, we believe this increase in large-scale variability reflected variation in oceanographic, topographic and hydrodynamic conditions, which can mask the importance of small-scale biological influences 33 , 62 – 64 . Environmental conditions can vary among the islands that make up the Aleutian Archipelago. For example, island size and coastline morphology, differences in the dimensions and depths of the many oceanic passes that separate them, their associated shelf bathymetries, and their water mass properties can also influence the environment 65 . Currents influencing water masses in this region are complex. In the central and western Aleutians, the Alaska Stream flows westward along the Aleutian shelf break, providing Bering Sea in flow from Samalga Pass to Near Strait 66 , 67 . These waters are relatively cool, saline, and nutrient rich 66 . In contrast, the Aleutian North Slope Current flows along the Bering side of the islands from Amchitka Pass eastward 68 , 69 . These waters become warmer, fresher, and more nutrient poor to the east due to the influence of the Alaska Coastal Current 67 . An example of the complexities of the Aleutian environment can be seen at Samalga Pass, a well-described biogeographic break for some cold-water corals, zooplankton, demersal fishes, seabirds, and deep marine faunal communities 65 , especially for benthic invertebrate and macroalgal assemblages on coastal rocky reefs 70 . We propose that these large-scale oceanographic influences play a larger role in structuring the trophically downgraded urchin barrens, while prior to downgrading, the Aleutian kelp forest were presumably locally controlled by biological interactions. When the different taxa within the two habitats were organized into their respective trophic guilds, we found similar results to those observed for the whole communities, with the exception of filter feeders. In particular, when abundances were estimated using density data, each of the trophic guilds was more variable in the kelp forest than the urchin barrens. Further, primary producer, grazer and predator assemblages were each most variable at the smallest spatial scale and least variable at the largest spatial scale within the kelp forests, but they followed the opposite pattern in the urchin barrens. In contrast, filter feeder assemblages were most variable at the largest spatial scale in kelp forests but at the smallest spatial scales in urchin barrens. However, given that many of the benthic filter feeders in this study (i.e., sponges and tunicates) exhibit indeterminate growth, a wide range in body sizes, and encrusting morphologies, we believe that estimates of variability based on density are insufficient to properly describe their abundances or spatial variation. Consequently, when abundances were estimated based biomass data, variability in filter feeder assemblages followed the same patterns observed for the whole communities and the for other feeding guilds; they were each most variable at the smallest scale and least variable at the largest scale in the kelp forests, but followed the opposite pattern in the urchin barrens. We believe that the differences in filter feeder scaling patterns between the two data types are likely not a sampling artifact but rather because biomass is the better measure of filter-feeder abundance and variation. Indeed, other studies have found discrepancies in community structure and distribution patterns for sponges when examining both density and biomass 71 , 72 . Likewise, the grazers were dominated by gastropods, whose heavy calcium carbonate shells may have dominated their biomass estimates, and thus were likely better described by their densities. In other words, the two sampling methods assess different taxa in the communities and this is important to how these communities are organized spatially. Consequently, we propose that each trophic guild followed the same scaling patterns as did the whole communities, and thus were likely affected by the same forcing factors. Specifically, they were each most variable at the smallest spatial scale within the kelp forests, which again is likely due to enhanced biological interactions 51 , 53 , 64 . As the kelp forests transitioned to urchin barrens and species diversity was reduced, the effects of these interactions became less important at larger spatial scales and the importance of physical factors and oceanographic control increased 35 , 62 , 73 – 75 . Our findings are in agreement with previous studies in coastal marine systems that have found variability in the structure of biological communities to exist at multiple spatial scales 21 , 26 , 33 – 35 . These patterns are often due to different forcing factors that operate across a range of scales 21 , 32 , 33 , 54 . For example, variation in biological interactions can drive variation at small scale (< 10 m) patterns, while habitat heterogeneity can drive local scale (< 10 km) differences in community structure 33 , 76 , 77 . At mesoscales (10–100 km) or regional scales (> 100 s km), coastal marine communities may differ across oceanographic boundaries where multiple physical parameters change simultaneously 35 , 78 , 79 . This can have important consequences to our understanding of the relative importance of the factors that structure these communities, as small-scale biological interactions can mask the influence of large-scale oceanographic and climate factors 51 – 53 , while large-scale factors can mask the influence of small-scale biological interactions 33 , 62 – 64 . Here, we show that variation in Aleutian kelp forest communities generally follows a scaling pattern, which is similar to patterns observed for giant kelp, Macrocystis pyrifera , populations along the coasts of California, USA and Baja California, MEX 33 , 35 . There, as with our current study, the greatest amount of variability was observed at spatial scales that encompassed 10 s of meters (i.e., among samples within each kelp forest) compared to scales that encompassed 100 s to 1000 s of kilometers (i.e., among geographic regions). However, that scaling pattern was altered by large-scale changes in ocean temperature and wave intensity during the 1997–1998 ENSO event, which affected the kelp populations at large (100 s–1000 s km) spatial scales and masked patterns of small-scale variability 27 . Specifically, this resulted in a reduction in small-scale variability and a shift towards large-scale variability 33 . Given the Aleutian nearshore ecosystem is a top-down driven system where, in general, environmental variables play a small role in determining community structure and the presence of alternate stable states 56 , we suggest the effects of trophic downgrading in the Aleutian kelp forests were similar to the ENSO effects in the California and Mexico kelp populations in that they reduced overall spatial variability in community structure and shifted this variability from small scales (i.e., within forests) to large scales (among geographic regions), and propose that this occurred due to a shift in the relative importance of biological interactions towards oceanographic forcing. Ecological theory predicts that global warming will increase the importance of foundation species in maintaining ecosystem function because they can ameliorate environmental stress 79 – 81 . In the Aleutian Archipelago, we have demonstrated that the trophic downgrading that has greatly eliminated foundational kelp species has also reduced variability in overall community structure and in the structure of trophic guilds. While functional redundancies among predators and herbivores can make diverse systems more stable 7 , 9 , simple food webs, such as those found in the Aleutian Archipelago, do not have the functional redundancies at higher trophic levels needed to maintain stability. Trophic rewilding (i.e., the restoration of apex predators) has been suggested for systems that have been downgraded 82 and may be one way that Aleutian kelp forests and their spatial variability can be restored. However, this may also be problematic for the Aleutian ecosystem given that killer whales still may be hunting sea otters there. Further, the establishment of a successful sea urchin fishery (i.e. human predation) to reduce sea urchin numbers may also not be feasible because the urchins in these barrens are generally characterized as having small gonads of poor quality due to a lack of macroalgal food 83 . But, as ocean temperatures in this region continue to rise with climate change 84 , the possibility of urchin disease may increase 85 , 86 , which could overtake predation as the primary source of urchin mortality 87 and result in substantial declines in urchin populations 88 – 90 . This can ultimately lead to a restoration of the ecosystem as seen in other areas of the world 80 , 88 , 91 , 92 . Regardless, we believe some mechanism of widespread urchin mortality may be necessary to return normal variability in ecosystem structure."
} | 5,324 |
34155335 | PMC8630232 | pmc | 3,054 | {
"abstract": "Permanently cold marine sediments are heavily influenced by increased input of iron as a result of accelerated glacial melt, weathering, and erosion. The impact of such environmental changes on microbial communities in coastal sediments is poorly understood. We investigated geochemical parameters that shape microbial community compositions in anoxic surface sediments of four geochemically differing sites (Annenkov Trough, Church Trough, Cumberland Bay, Drygalski Trough) around South Georgia, Southern Ocean. Sulfate reduction prevails in Church Trough and iron reduction at the other sites, correlating with differing local microbial communities. Within the order Desulfuromonadales , the family Sva1033, not previously recognized for being capable of dissimilatory iron reduction, was detected at rather high relative abundances (up to 5%) while other members of Desulfuromonadales were less abundant (<0.6%). We propose that Sva1033 is capable of performing dissimilatory iron reduction in sediment incubations based on RNA stable isotope probing. Sulfate reducers, who maintain a high relative abundance of up to 30% of bacterial 16S rRNA genes at the iron reduction sites, were also active during iron reduction in the incubations. Thus, concurrent sulfate reduction is possibly masked by cryptic sulfur cycling, i.e., reoxidation or precipitation of produced sulfide at a small or undetectable pool size. Our results show the importance of iron and sulfate reduction, indicated by ferrous iron and sulfide, as processes that shape microbial communities and provide evidence for one of Sva1033’s metabolic capabilities in permanently cold marine sediments.",
"conclusion": "Conclusion This study has shown how the microbial communities in sub-Antarctic South Georgia surface sediments are shaped by the dominant TEAP; sulfate reduction in Church Trough and iron reduction in Cumberland Bay, Drygalski Trough, and Annenkov Trough. We provide evidence for microbial iron reduction as one of the metabolic capabilities of the family Sva1033 using RNA-SIP with Cumberland Bay surface sediments. Coincidentally, in all iron reduction sites, Sva1033 was the dominant member of Desulfuromonadales found in situ, while other known marine iron reducers were scarce. We also identified iron-reducing capabilities of Sva1033 members in similar surface sediments from Potter Cove in the Antarctic Peninsula. Therefore, this clade might be very important for iron reduction in permanently cold marine sediments given the input of iron from enhanced glacial erosion, weathering, and glacial melt as a result of global warming. Furthermore, our data show high relative abundance of persistent sulfate reducers and suggest their activity in the iron reduction zone of marine sediments potentially participating in cryptic sulfur cycling, with the produced sulfide precipitating as metal sulfide mineral or being reoxidized.",
"introduction": "Introduction Organic matter degradation is the main source of electron donors and carbon for microbial metabolism in marine sediments [ 1 , 2 ]. The estimated 5.39 × 10 29 microbial cells in marine sediments [ 3 ] form a microbial food chain. Below the oxic zone, the anaerobic portion of this food chain starts with specialists, which perform hydrolytic and fermentative processes [ 4 ], and ends with anaerobically respiring microorganisms, which oxidize fermentation products with available terminal electron acceptors such as nitrate, Mn(IV), Fe(III), sulfate, and CO 2 . Because nitrate and Mn(IV) are rapidly depleted in the uppermost centimeters of most coastal and upper slope surface sediments [ 2 , 5 ], sulfate and Fe(III) are the most abundant terminal electron acceptors utilized by microorganisms for mineralization of fermentation products in these depositional environments [ 6 – 8 ]. Iron enters the ocean from various sources including terrigenous origins via weathering and erosion and subsequent transport by rivers and windblown dust; hydrothermal vents [ 9 ]; melting sea ice and icebergs [ 10 ]; and glacial associated erosion, weathering, and meltwater [ 11 – 14 ]. Due to global warming, glacial associated input of iron is predicted to increase in the future, resulting in enhanced amounts of iron reaching coastal sediments and adjacent ocean areas especially in higher latitudes [ 13 , 15 , 16 ]. Sulfate, which is generally present in high concentrations in the water column (~28 mM), is supplied to the sediment by downward diffusion, accelerated by bio-irrigation and other advective processes [ 7 , 8 , 17 ]. In addition, it is the final product of reoxidation of sulfide [ 18 , 19 ], which itself is produced by sulfate reduction [ 20 ], potentially resulting in a cryptic sulfur cycle [ 18 , 19 , 21 ]. Iron reduction is constrained by the reactivity and lower availability of ferric iron compared to sulfate [ 5 , 22 ]. Therefore, while iron reduction is favored in certain marine settings [ 23 , 24 ], organic matter oxidation by sulfate reduction is often more important than iron reduction in marine sediments [ 7 , 8 ], a competition shown to be also regulated by the availability and reactivity of organic matter and ferric iron [ 22 , 25 ]. Geochemical and biogeochemical factors have been previously shown to be key parameters shaping the microbial communities in marine sediments [ 1 , 26 , 27 ]. Besides the availability of terminal electron acceptors [ 1 ], i.e., Fe(III) and sulfate, these factors include quantity, composition, and reactivity of organic matter [ 26 , 28 , 29 ], sediment geochemistry [ 27 , 30 , 31 ], salinity [ 32 ], temperature [ 33 ], ocean currents [ 34 ], primary productivity in the overlying water column [ 35 ], and sedimentation rate [ 24 ]. The permanently cold surface sediments around the island of South Georgia in the South Atlantic Ocean, which we have investigated in this study, are influenced by high organic matter content around the shelf areas (0.65 wt% Cumberland Bay [ 36 ]), and high iron content within or close to the fjords (ref. [ 37 ], Cumberland Bay: total Fe solid phase 47 g/kg [ 38 ]; 0.7 wt% ferrihydrite and lepidocrocite [ 39 ]). In addition, the studied sediments were found to be characterized by widespread active methane seepage within the fjords and on the shelf, mostly associated with cross-shelf glacial troughs [ 36 , 40 , 41 ]. So far, we studied the microbial communities inhabiting deeper sediments (down to 10 m below seafloor) at three sites around South Georgia [ 42 ] and in the present study, a detailed analysis of the surface sediments at very fine scales is provided. The sediments of the second study site Potter Cove, a small fjord located at the southwest of King George Island/Isla 25 de Mayo (South Shetland Islands) on the northern tip of the West Antarctic Peninsula, are characterized by a high input of iron from glacial meltwater and bedrock erosion [ 13 , 14 , 43 ]. Especially, sediments close to the glacier termination show a deeper ferruginous zone compared with sediments not directly influenced by glaciers [ 14 ] similarly to Cumberland Bay, South Georgia [ 41 ]. We hypothesize that differing geochemical characteristics in the surface sediments (top 20–30 cm) at various sites around South Georgia shape the local microbial communities. To test this hypothesis, four sites, located on the outer shelf (Annenkov Trough, Church Trough) or within or close to one of the fjords (Cumberland Bay, Drygalski Trough), were selected around the island of South Georgia. These sites were characterized by either high ferrous iron or hydrogen sulfide concentrations. The microbial communities of these sediments were investigated by 16S rRNA gene sequencing, quantitative PCR and RNA stable isotope probing (SIP) incubations. Correlation and multivariate regression analyses were performed to identify which geochemical parameters primarily shape the microbial community composition. The active iron-reducing microbial community of South Georgia sediments was compared to those in geochemically similar sediments of Potter Cove (Antarctic Peninsula) using RNA-SIP experiments.",
"discussion": "Discussion Permanently cold coastal sediments from sub-Antarctic and Antarctic regions are subject to increased input of iron and other terrigenous compounds as a consequence of intensified weathering, erosion, and glacial melt due to observed global warming [ 10 – 15 , 89 , 90 ]. The impact of these altered element and material flux on the microbial communities in such sediments is currently understudied. Likewise, the bacterial communities present in surface sediments around the sub-Antarctic island South Georgia were not previously studied in detail. This study investigated the impact of environmental change on microbial communities in permanently cold (sub-)Antarctic sediments. Our findings show how geochemical characteristics such as the predominant electron accepting process and quality of organic matter potentially shape sediment communities in various sites around South Georgia. Importantly, in the iron reduction sites, we obtained evidence for dissimilatory iron reduction as one of Sva1033 clade’s ecological roles in permanently cold sediments using RNA-SIP. Finally, indications for concurrent sulfate reduction were obtained, despite the dominance of iron reduction in incubation experiments. Geochemical footprints shape microbial community composition Selective survival of taxa buried below the upper 10 cm bioturbation zone has been identified as the significant process relevant for microbial community assembly in marine sediments [ 91 – 94 ]. Using the geochemical parameters as environmental factors for selection of the microbial community composition in the dbRDA (Figs. 4 and S5 ), various trends were observed. For example, depth-wise variation in community composition across all sites was strongly explained by the ammonium concentrations (Fig. 4 ), whose presence—along with DIC—is an indicator for organic matter degradation [ 95 ]. This was reflected in the core microbial community; the taxa Flavobacteriales , Rhodobacterales , Cellvibrionales , and Verrucomicrobiales , known for degradation of labile organic matter such as proteins, amino acids, polysaccharides, and simple sugars [ 96 – 99 ], were more abundant in the surface and decreased with depth across all sites (Fig. 3A ). A similar trend was previously observed for some of these taxa in sediments of the Antarctic shelf [ 100 ]. In contrast, known “persister” microorganisms [ 91 – 93 ] such as Anaerolineae , Phycisphaerae , and the Atribacteria clade JS1 [ 101 – 104 ], showed a consistent increase in relative abundance along increasing depth across all sites (Fig. 3A ). Differing supply of fresh organic matter on the outer shelf sites (Church Trough and Annenkov Trough) compared to sites located closer to the island was a possible explanation for likely higher microbial activity at these sites, as corroborating data from the geochemical profiles and gene copy numbers of microorganisms in the sediments (Figs. 2 , 3 B, 5 , and S4B ) indicated. This idea was supported by known large phytoplankton blooms and high primary production on the outer shelves around South Georgia [ 37 , 40 , 105 ]. Beyond ammonium shaping the communities along the sediment depth gradient, the dbRDA similarly showed a distinct selection of microbial communities in the study sites based on ferrous iron and sulfide concentrations. Thus, the likely dominant TEAP, i.e., iron and sulfate reduction, served as a factor for identifying the sites as either a group of iron reduction sites (Annenkov Trough, Cumberland Bay, Drygalski Trough) or sulfate reduction site (Church Trough; Fig. 4 ). A strong dependency of microbial community composition on TEAP was previously demonstrated in deeper sediments (down to 10 m below seafloor) from South Georgia [ 42 ] and from the Baffin Bay in the Arctic [ 27 ], in which iron and sulfate or iron and manganese reduction dominated, respectively. Since ferrous iron and sulfide as products of microbial iron and sulfate reduction, respectively, were recognized by dbRDA as the environmental factors in our sediments to shape local communities, we hypothesize that the microorganisms contributing to these processes are important members of the microbial community. Thus identification of potential sulfate and iron reducers in the sediments will reveal which microorganisms are likely involved in the terminal respiratory processes. Amongst the sulfate reducers (Fig. 5 ), Desulfobacteraceae , Desulfobulbaceae , and Desulfarculaceae were most dominant, even down to the deeper layers of Cumberland Bay and Church Trough sediments [ 42 ]. The order Desulfuromonadales harbors many species with the metabolic capability to perform dissimilatory iron reduction [ 74 , 75 , 77 ], but also sulfur reduction [ 106 – 110 ] and even a few microorganisms are capable of sulfate reduction [ 111 ]. Members of Desulfuromonadales are typically found in ferruginous sediments (e.g., [ 45 , 112 , 113 ]). This order was the most abundant potential dissimilatory iron reducing clade in this study. Here, the main representative identified was the family Sva1033 (Figs. 3 A, 7A , and S6 ), which was recently suggested to be capable of iron reduction in a Terrestrial Mud Volcano site [ 82 ] and Arctic sediments [ 112 ]. Based on the calculated phylogenetic tree (Fig. S7 ), this family is closely related to the clade “Desulfuromonas 2” (as assigned by Silva release 138 [ 64 ]). Until now, there are no cultivated members of this clade and its metabolic capabilities are yet to be confirmed. The significant correlation between depth-wise Fe 2+ concentrations with relative abundance of Sva1033 in Annenkov Trough and Drygalski Trough (Fig. S6 ) strengthens the hypothesis that Sva1033 is involved in microbial iron reduction in surface sediments of South Georgia. The Sva1033 clade is capable of dissimilatory iron reduction As the Sva1033 clade was first identified in Arctic sediments [ 114 ], we tested the hypothesis that this clade is ecologically adapted to perform iron reduction as one of its metabolic capabilities in permanently cold sediments. This was done by setting up RNA-SIP incubations using acetate as labeled substrate with Cumberland Bay and Antarctic Potter Cove sediments, especially as both Potter Cove and Cumberland Bay are characterized by a broad ferruginous zone [ 14 ]. Due to thermodynamic constraints in dissimilatory utilization, acetate has been frequently used successfully for specifically tracing anaerobically respiring microorganisms such as iron reducers (e.g., [ 78 , 86 ]). Within Cumberland Bay sediments, iron reduction is very likely the dominant TEAP occurring in all SIP incubations as indicated by the increasing Fe 2+ concentrations in the treatments, including controls (Fig. S8A ). The slurry likely retained endogenous iron oxides and organic matter from the original sediment. In surface sediments from Cumberland Bay (same sampling site but previous expedition), total Fe content of the solid phase of 46.8 g/kg was reported [ 38 ], of which ferrihydrite and lepidocrocite contributed 0.65–0.7 wt% Fe [ 39 ]. Therefore, iron reduction could be stimulated without the amendment of additional electron acceptors or donors (see unamended control treatment, Fig. S8A ). Given the similarity in the microbial community composition and proportion of enriched taxa in the heavy fractions (Fig. 6B ), we conclude that dissimilatory iron reduction is the most likely dominant process conducted by the labeled taxa enriched in the heavy fractions of all treatments, i.e., members of Desulfuromonadales with Sva1033 as the most abundant taxon. This conclusion was supported by the observed similar geochemistry in the treatments (Fig. S8 ). The possibility that other processes such as sulfate and sulfur reduction, which could occur in these incubations, stimulated the enrichment of Desulfuromonadales is not supported by the formation of Fe 2+ in the incubation experiment over time (Fig. S8A ). The likelihood that Sva1033 performs iron reduction in situ as in the RNA-SIP incubations is supported by the close relation of OTU sequences from the in situ sediment and the SIP experiment (Fig. S7 ). Likewise, in Potter Cove sediment incubations, Sva1033 was also identified as the most dominant organism taking up the acetate label (40% relative abundance in ultra-heavy fraction, Figs. 7C and S9 ). Our study thus provides evidence for the capability for microbial iron reduction in the uncultured Sva1033 clade from permanently cold (sub-)Antarctic sediments with acetate as electron donor and carbon substrate (Fig. 7 ). Other taxa enriched in the heavy fractions of the SIP experiments included Desulfuromonas , Geopsychrobacter , and Geothermobacter , species with known iron reducing capabilities [ 75 , 80 , 115 , 116 ]. Activity of sulfate reducers in the iron reduction sites Sulfate reducers are metabolically flexible. Their primary metabolic capabilities are essential for the global sulfur cycle as utilizers of the most oxidized form of sulfur [ 117 ] and they are capable of syntrophic growth, e.g., with methanogens [ 118 , 119 ]. In addition, sulfate reducers are capable of growth with TEAPs such as nitrate [ 84 , 120 ] and Fe(III) under sulfate limitation [ 83 , 121 ]. In similar permanently cold marine sediments from the Arctic, sulfate reducers were relatively less abundant compared to iron reducers when iron reduction predominated [ 27 , 112 ], which is in contrast to this study (Fig. 5 ). Despite the high abundance of sulfate reducers in the sediments of the iron-rich sites (Figs. 3 and 5 ), evidence for sulfate reduction was not directly obtained from the pore water profiles (Fig. 2 ). Hence, an open question emerges regarding the metabolism that keeps sulfate reducers persistent across these sites such that their relative abundances outnumber the iron reducers who likely perform the clearly more dominant TEAP in situ (Fig. 2 ). We hypothesize for our study that sulfate reduction, masked by the reoxidation of the produced sulfide back to sulfate, fuels the persistence of sulfate reducers [ 19 , 21 ] in the iron reduction sites. The results from SIP incubations showed that sulfate reducers were present and active in all treatments. However, they were present in higher abundance in the light compared to the heavy fractions (Fig. 6B ). This can be explained by their high abundance in the starting sediment material (42%, Fig. S10 ). Sulfate reducers responded to the addition of electron acceptor as evidenced by their increased relative abundance in the heavy fractions of the sulfate-amended treatment compared to the other treatments (25% vs. 7–18% ultra-heavy labeled fraction). In general, the lower enrichment of sulfate reducers in the heavy fractions compared to the potential iron reducers is likely because (I) iron reducers were more efficient in the uptake of electron donors such as the provided acetate [ 122 ]; (II) iron reduction was the dominant biogeochemical process observed (as discussed above, Fig. 6B and S8A ); or/and (III) sulfate reducers were thriving on different, sediment endogenous electron donors [ 123 ]. Nevertheless, the detection of dsrA transcripts in the SIP fractions (Fig. 6C ) supported the hypothesis of co-occurring sulfate reduction in the treatments. Following the observations of the SIP experiment, we suggest that minor concurrent sulfate reduction, in the background of the dominant TEAP i.e. iron reduction, likely fuels the persistence of sulfate reducers in situ in the iron reduction sites Annenkov Trough, Cumberland Bay, and Drygalski Trough (Figs. 3 A and 5 ). The non-detection of sulfide in the incubations could be explained by precipitation with Fe 2+ forming mackinawite (FeS) and/or pyrite (FeS 2 ) [ 124 , 125 ], or reoxidation microbially or abiotically by reactive Fe(III) oxides [ 21 ]. Recently, this type of cryptic sulfur cycling, indicated by high sulfide oxidation rates in surface sediments, was shown in multiple studies [ 21 , 126 , 127 ]. In addition, concurrent sulfate and iron reduction in the same zone was reported multiple times [ 17 , 20 , 77 , 124 ], but detailed information about the associated microbial community is lacking. These studies [ 21 , 126 – 129 ] assign the majority of ferrous iron production to the abiotic process of iron reduction by sulfide oxidation. Although this process likely also occurs in the sediments investigated in this study, sulfide concentrations below detection limit (Fig. 2 ) and the high activity of mainly iron reducing microorganisms (Fig. 6B ) indicate that these abiotic processes provide a minor contribution to observed high Fe 2+ concentrations (Figs. 2 and S8A ). A limitation of our study is the unexpected lack of dissolved Fe 2+ over time in the acetate, lepidocrocite and molybdate treatment from the SIP incubation (Fig. S8A ). While this result suggested that iron reduction did not occur in this treatment, the enrichment of Desulfuromonadales members in similar proportion as in the other treatments shows that iron reduction certainly occurred (Fig. 6B ). Besides, inhibition of iron reduction by molybdate has not been shown previously. In comparable studies, molybdate concentrations of 10 mM [ 83 ] or even 20 mM [ 124 ] did not inhibit microbial iron reduction. Instead, in our study, the produced dissolved Fe 2+ reacted abiotically with the added molybdate, preventing the detection of Fe 2+ (see supplementary material, Figs. S11 and S12 , for details)."
} | 5,438 |
35252822 | PMC8894265 | pmc | 3,055 | {
"abstract": "Summary Natural biological materials provide a rich source of inspiration for building high-performance materials with extensive applications. By mimicking their chemical compositions and hierarchical architectures, the past decades have witnessed the rapid development of bioinspired materials. As a very promising biosourced raw material, silk is drawing increasing attention due to excellent mechanical properties, favorable versatility, and good biocompatibility. In this review, we provide an overview of the recent progress in silk-based bioinspired structural and functional materials. We first give a brief introduction of silk, covering its sources, features, extraction, and forms. We then summarize the preparation and application of silk-based materials mimicking four typical biological materials including bone, nacre, skin, and polar bear hair. Finally, we discuss the current challenges and future prospects of this field.",
"conclusion": "Conclusion and outlooks The excellent mechanical properties, favorable versatility, and good biocompatibility render silk a very promising raw material for constructing high-performance biomimetic materials. In this review, silk-based structural and functional materials inspired by some typical natural biological materials including bone, nacre, skin, and polar bear hairs, are thoroughly displayed and discussed. It should be noted that the bionic systems are certainly not limited to the above-mentioned ones, as silk fiber itself is also a classic example to be extensively imitated ( Ling et al., 2017b ). Given that several good review articles had systematically summarized the development of bioinspired spinning and regenerated silk fibers ( Koeppel and Holland, 2017 ; Liu et al., 2019b ; Shang et al., 2019 ), we herein do not mainly focus on this subject. But in respect to expanding raw material sources, those artificial silk fibers prepared from waste silk, or those with mechanical performance superior to native silk, will make a considerable contribution. Despite the progress achieved in silk-based bioinspired materials during the past decades, there are still challenges for their structural design and bulk production. Indeed, it is hard to completely replicate the full-scale hierarchical architectures of natural biological materials at ambient conditions. To this end, we try to propose a viewpoint combining reductionism and integratism. Specifically, reductionism requires us to perfectly analyze the composition and structure of biological materials from macroscale to nanoscale. For example, Reznikov et al.’s 3D observation and characterization resulted in an expansion of the previously known hierarchical structure of bone to at least 12 levels ( Reznikov et al., 2018 ), with particularly elucidating the organization and relationship between bone’s principal components—mineral and collagen, which will definitely help to understand the delicate design of bone in every single detail. Integratism, generally speaking, reminds us to use identical or similar components to build biomimetic materials, with emphasizing on the integration of the multiple components or multiscale structures. For example, Mao et al. fabricated a millimeter-thick synthetic nacre that highly resembles both the chemical composition and the hierarchical structure of natural nacre ( Mao et al., 2016 ). This was achieved by an “assembly-and-mineralization” approach inspired by the natural process in mollusks, where silk fibroin formed the organic layers between the aragonite layers and played a vital role for adhesion. In this regard, will it be a better alternative using catechol-groups-modified silk molecules ( Burke et al., 2016 ) with improved adhesive capability? The answer is unknown. Whereas, in terms of bone-like composites, genetically engineered silk protein fused with HAP-binding domain VTKHLNQISQSY (VTK) had been confirmed to promote biomineralization and thus further enhanced osteoinductivity ( Dinjaski et al., 2017 ). In addition, the inherent arginine–glycine–aspartate (RGD) motif existed in some wild silk could induce cell attachment and proliferation in bone tissue engineering ( Behera et al., 2017 ; Sahu et al., 2015 ). Hence, taking full advantage of silk by directional selection or optimization of its features may provide a solid material foundation for the future silk-based bioinspired functional composites. With challenges remaining for building high-performance bioinspired materials in practical form and in bulk, a systematic methodology for dealing with the structural complexity of a truly bioinspired structure whose dimensions span from the nanoscale to the macroscale is highly demanded for scientists and engineers, while combining two or more fabrication technologies may work out as a solution ( Wegst et al., 2015 ). Certainly, the development of silk-based bioinspired structural and functional materials will also follow this route. Hopefully, there will be a breakthrough in the near future.",
"introduction": "Introduction Many biological materials, such as bone, nacre, bamboo, spider silk, polar bear hair, and skin, possess remarkable mechanical properties, as well as some unique functions, generally originating from their multiscale structures ( Eder et al., 2018 ; Meyers et al., 2013 ; Wegst et al., 2015 ). Learning from these natural examples will stimulate the development of high-performance structural and functional materials, due to their great demands in various important fields, including aircraft technology, automobile manufacturing, electronic sensing, tissue engineering, and personal thermal management ( Bhattacharjee et al., 2017 ; Chortos and Bao, 2014 ; Cui et al., 2018 ; Guan et al., 2020 ; Li et al., 2020a ; Mao et al., 2021 ; Pan et al., 2021 ). During the past decades, significant progress has been achieved in bioinspired materials. A series of methodologies, from component selection, biomimetic structure design, fabrication procedure, to application demonstration, have been intensively explored. For example, mimicking natural nacre’s elegant “brick-and-mortar” architectures, bulk synthetic nacre could be successfully constructed by infiltrating epoxy resin into a predesigned nanomaterial framework, which exhibited excellent strength and toughness comparable or even superior to natural nacre ( Bai et al., 2016 ; Du et al., 2019a ; Han et al., 2019 ; Zhao et al., 2020 ). It is also worth noting that high-performance biological materials typically consist of components like proteins, polysaccharides, or minerals, which are biodegradable and sustainable ( Mohanty et al., 2018 ; Wegst et al., 2015 ). This has become increasingly important as the petroleum-based polymers are difficult to biodegrade, leading to rapidly growing pollution problems worldwide ( Geyer et al., 2017 ; Jambeck et al., 2015 ). Therefore, in part or in whole, using biopolymers as building blocks to construct bioinspired materials emerge as an attracting alternative ( Li et al., 2021 ). To this extent, many biopolymers including collagen, silk, cellulose, chitin, chitosan, and alginate have been investigated to develop composites with biomimetic structures. Among them, silk stands out for its excellent mechanical properties, favorable versatility, and good biocompatibility ( Holland et al., 2019 ; Melke et al., 2016 ; Omenetto and Kaplan, 2010 ; Shao and Vollrath, 2002 ; Tao et al., 2012 ; Xu et al., 2019 ), making it a promising raw material for building multiple biomimetic systems. Recently, silk-based bioinspired structural and functional materials have drawn increasing attention. In this work, we overview the progress of bioinspired materials based on silk. We first give a brief introduction of silk, covering its sources, features, extraction, and forms. Then, we summarize the preparation and application of silk-based materials mimicking bone, nacre, skin, and polar bear hair. Finally, we discuss the current challenges and future prospects of the development of silk-based bioinspired materials. This review provides a better understanding to silk’s roles in biomimetic composites, which will further inspire scientists and engineers to explore its potential applications."
} | 2,055 |
38757653 | PMC11251569 | pmc | 3,056 | {
"abstract": "Abstract Bioinspired synaptic devices have shown great potential in artificial intelligence and neuromorphic electronics. Low energy consumption, multi‐modal sensing and recording, and multifunctional integration are critical aspects limiting their applications. Recently, a new synaptic device architecture, the ion‐gating vertical transistor (IGVT), has been successfully realized and timely applied to perform brain‐like perception, such as artificial vision, touch, taste, and hearing. In this short time, IGVTs have already achieved faster data processing speeds and more promising memory capabilities than many conventional neuromorphic devices, even while operating at lower voltages and consuming less power. This work focuses on the cutting‐edge progress of IGVT technology, from outstanding fabrication strategies to the design and realization of low‐voltage multi‐sensing IGVTs for artificial‐synapse applications. The fundamental concepts of artificial synaptic IGVTs, such as signal processing, transduction, plasticity, and multi‐stimulus perception are discussed comprehensively. The contribution draws special attention to the development and optimization of multi‐modal flexible sensor technologies and presents a roadmap for future high‐end theoretical and experimental advancements in neuromorphic research that are mostly achievable by the synaptic IGVTs.",
"introduction": "1 Introduction Understanding and replicating the intricate functions of the human brain using electronics is not an easy feat. Over a century ago, Nobel Laureate neuroanatomist Santiago Ramón y Cajal discovered that the human central nervous system is composed of an intricate web of connections between countless neurons. [ \n \n 1 \n \n ] The brain contains ≈100 trillion synapses, which act as bridges between one hundred billion neurons. [ \n \n 2 \n \n ] These biological neural networks process information through synaptic events. Nerve impulses in the form of action potentials are generated by presynaptic neurons in response to received input or stimulus information. These impulses are then transmitted to postsynaptic neurons via the synapses. Generally, the neuron membrane devises a resting potential (from −60 to −70 mV) that is either excited or inhibited by incoming stimuli. [ \n \n 3 \n \n ] Excitatory or inhibitory stimuli can create positive or negative voltage, whereas the combination of inputs beyond a certain threshold generates an action potential that controls the influx of calcium cations into the cytoplasm. This event releases neurotransmitters into the synaptic cleft. On the postsynaptic neuron, such transmitters connect to receptors, causing the ligand‐gated channel to allow sodium ions to enter the postsynaptic membrane. This transmits to the subsequent neuron the input stimuli. [ \n \n 4 \n , \n 5 \n \n ] The neurotransmitters' chemical interaction with the postsynaptic receptors changes their shape and induces succeeding reactions for ion gating. The abrupt charge disparity creates the potential that initiates another action stimulus in the postsynaptic terminal for information transmission. [ \n \n 6 \n \n ] Enzymatic processes deactivate the neurotransmitters linked to the receptors, causing their departure toward the presynaptic region for the subsequent transmission. [ \n \n 7 \n \n ] The synaptic weight between two neighboring neurons determines the information transmission efficiency, i.e., the memory effect. The stimuli‐responsive change in synaptic weight is known as plasticity, [ \n \n 8 \n \n ] which is dependent on the amplitude and repetition of action potential spikes. In recent years, scientists have made significant progress in developing bio‐inspired synaptic systems and artificial multisensory neurons for memory and perception ( Figure \n \n 1 a.1 ). [ \n \n 9 \n , \n 10 \n , \n 11 \n , \n 12 \n \n ] These neuromorphic devices are capable of imitating basic and advanced neural functions, such as pain perception, pattern recognition, [ \n \n 13 \n , \n 14 \n \n ] and light‐, sound‐ and pressure sensing, [ \n \n 15 \n , \n 16 \n , \n 17 \n \n ] besides emulating the five primary human sensory systems (Figure 1a.2 ) via multisensory neural networks (Figure 1a.3 ). Among the various neuromorphic device technologies, which have led to the exponentially increasing number of publications since the 2000s, transistors have played an important role especially after 2008 (Figure 1b ). Particularly, the ion‐gating transistor (IGT) has gained significant attention due to its unique features arising from its ionic‐ and/or electronic current modulation, such as high transconductance, large gate capacitance, low operating frequency, and ultrahigh carrier concentration. [ \n \n 18 \n , \n 19 \n , \n 20 \n \n ] These properties and the IGT's structural characteristics make this device suitable for artificial synapse applications. [ \n \n 21 \n , \n 22 \n , \n 23 \n \n ] Despite the challenges in device fabrication due to the transistor's 3‐terminal integration, the expressive current modulation enabled by ion‐gating is crucial for the simultaneous reading and writing operations required by memory. [ \n \n 24 \n , \n 25 \n \n ] Upon the application of voltage stimuli, the IGT's gate ( G ) can emulate the presynaptic function, whereas the postsynaptic response can be acquired at the source ( S ) and drain ( D ) electrodes. Multi‐gated stimulations on an IGT can be processed through this single device, providing ample room for information collection from various sources, resulting in spatiotemporal recognition apparently impossible in two‐terminal devices. [ \n \n 26 \n , \n 27 \n , \n 28 \n \n ] Additionally, IGTs allow the integration of inorganic‐ and/or organic materials into multifunctional artificial synapses, providing versatility, mechanical flexibility, biocompatibility, and low‐cost manufacturing (e.g., printing and coating), with the ability to modify their properties via supramolecular engineering and cross‐linking. [ \n \n 29 \n , \n 30 \n , \n 31 \n \n ] \n Figure 1 Human‐brain‐inspired multisensory functions and neuromorphic research. a) Schematic illustrations of a.1) the five primary sensory systems in the human body, a.2) multisensory functions processed by the human brain, and a.3) their emulation by artificial neural networks. Adapted with permission. [ \n \n 12 \n \n ] Copyright 2021, Tan et al. Published by Springer Nature. b) The increasing number of publications regarding neuromorphic applications (viz. logarithmic scale) over the past years to October 2023. Data collected from Web of Science on November 2023. Examination fluctuations (upper error bars) arise from possible keyword combinations, article categorization, and search refinement. Recent research on IGTs has unveiled a game‐changer neuromorphic device − the ion‐gating vertical transistor (IGVT). [ \n \n 32 \n , \n 33 \n \n ] Accordingly, Figure 1b also showcases the number of publications focused on neuromorphic applications of vertical transistors, putting forward that research on this new device technology is just getting started. Additionally, Figure 1b indicates that vertical transistors have the potential to become one of the next exponentially rising technologies in the field of neuromorphic science. Among the vertical transistors in Figure 1b , the IGVTs are in pride of place given their distinguished mixed ionic‐electronic current modulation. Furthermore, compared with conventionally planar‐configuration devices, IGVTs display higher bias‐stress constancy, lower working voltage, and more efficient power management. [ \n \n 34 \n , \n 35 \n \n ] This is primarily due to the IGVT channel length ( L ) being determined by the S ‐to‐ D vertical separation, naturally, the thickness of a thin film − a feature that allows for downscaling the ionic‐electronic transport pathway to the tiny sub‐10 nm range. [ \n \n 36 \n , \n 37 \n , \n 38 \n \n ] Secondly, whereas in conventional IGTs the lateral charge transport is more vulnerable to defects at the semiconductor/electrolyte interface, in IGVTs, the vertical charge transport significantly eliminates this side effect. [ \n \n 34 \n \n ] IGVTs have indeed more outstanding mechanical stability compared to planar IGTs since the vertical charge transport enhances the semiconductor channel resilience to fissures and displacements triggered by deformation. [ \n \n 39 \n , \n 40 \n \n ] Such unique and promising features have been capturing the attention of artificial synapse research, which turned its focus recently to IGVTs. In just a few years, such devices have already demonstrated low voltage operation, low energy consumption, high recognition accuracy, and promising mechanical flexibility, multiplex integrability, and multisensory capability. [ \n \n 40 \n , \n 41 \n , \n 42 \n , \n 43 \n , \n 44 \n , \n 45 \n \n ] \n Our review explores the exciting potential of using vertical transistor technology to enhance ion‐gating properties in neuromorphic applications. Although there have been previous reviews that cover resembling topics related to active materials, [ \n \n 46 \n , \n 47 \n , \n 48 \n \n ] device architectures and design principles, [ \n \n 23 \n , \n 34 \n , \n 43 \n \n ] ion‐gating properties, [ \n \n 22 \n , \n 49 \n , \n 50 \n \n ] and neuromorphic applications, [ \n \n 35 \n , \n 44 \n \n ] the inherent importance and novelty displayed by the use of IGVTs to boost the capabilities of state‐of‐the‐art neuromorphic applications make this review no sooner vital. In due course, here we offer an up‐to‐date overview of the latest advancements in IGVT technology, with a focus on cutting‐edge demonstrations that can revolutionize the next generation of brain‐inspired artificial synapses. We begin with an examination of the IGVT's architecture and ion‐gating principle (Section 2 ), highlighting the unique features that make this device ideal for high‐end neuromorphic applications. Subsequently, the topical achievements in IGVT fabrication reliability, monolithic integration processes, and electrical characteristics are discussed in Section 3 , providing the standards for IGVT‐based artificial‐synapse development. Then, Section 4 brings up insightful demonstrations of neuromorphic IGVTs in high performance, high energy‐efficiency, flexible, multisensory, and multi‐modal applications. As a roadmap for advanced neuromorphic technologies, Section 5 points out the current open challenges in the field and offers our view on the most useful theoretical and experimental tools that are currently available for researchers and specialists to tackle each standing situation. Finally, Section 6 provides a critical outlook on the future of synaptic IGVTs."
} | 2,660 |
29593682 | PMC5858530 | pmc | 3,057 | {
"abstract": "Arbuscular mycorrhizal and ectomycorrhizal symbioses are among the most important drivers of terrestrial ecosystem dynamics. Historically, the two types of symbioses have been investigated separately because arbuscular mycorrhizal and ectomycorrhizal plant species are considered to host discrete sets of fungal symbionts (i.e., arbuscular mycorrhizal and ectomycorrhizal fungi, respectively). Nonetheless, recent studies based on high-throughput DNA sequencing technologies have suggested that diverse non-mycorrhizal fungi (e.g., endophytic fungi) with broad host ranges play roles in relationships between arbuscular mycorrhizal and ectomycorrhizal plant species in forest ecosystems. By analyzing an Illumina sequencing dataset of root-associated fungi in a temperate forest in Japan, we statistically examined whether co-occurring arbuscular mycorrhizal ( Chamaecyparis obtusa ) and ectomycorrhizal ( Pinus densiflora ) plant species could share non-mycorrhizal fungal communities. Among the 919 fungal operational taxonomic units (OTUs) detected, OTUs in various taxonomic lineages were statistically designated as “generalists,” which associated commonly with both coniferous species. The list of the generalists included fungi in the genera Meliniomyces, Oidiodendron, Cladophialophora, Rhizodermea, Penicillium , and Mortierella . Meanwhile, our statistical analysis also detected fungi preferentially associated with Chamaecyparis (e.g., Pezicula ) or Pinus (e.g., Neolecta ). Overall, this study provides a basis for future studies on how arbuscular mycorrhizal and ectomycorrhizal plant species interactively drive community- or ecosystem-scale processes. The physiological functions of the fungi highlighted in our host-preference analysis deserve intensive investigations for understanding their roles in plant endosphere and rhizosphere.",
"introduction": "Introduction In terrestrial ecosystems, most plant species form intimate interactions with mycorrhizal fungi, which play essential roles in the growth and survival of their hosts (van der Heijden et al., 2008 ; Bever et al., 2010 ; Peay et al., 2016 ). Those fungi, for example, supply soil nitrogen and phosphorous to associated plants, thereby enhancing hosts' physiological states (Smith and Read, 2008 ). They are also known to reduce deleterious effects of pathogens on host plants (Marx, 1972 ; Azcón-Aguilar and Barea, 1997 ; Borowicz, 2001 ). Moreover, mycorrhizal fungi can contribute to physiological homeostasis of plants by increasing hosts' resistance to abiotic stress (Grover et al., 2011 ). Therefore, understanding and managing below-ground integrations between plants and their mycorrhizal fungal symbionts are major challenges not only in basic ecology but also in forestry and agronomy. Among the several categories of mycorrhizal fungi, arbuscular mycorrhizal, and ectomycorrhizal fungi are major groups of below-ground fungal communities in temperate forests (Smith and Read, 2008 ; Peay et al., 2016 ). Arbuscular mycorrhizal fungi (the phylum Glomeromycota) first appeared early in the history of land plants (Remy et al., 1994 ) and hence they associate with plant species in diverse plant taxa (Schüβler et al., 2001 ). They are obligate mutualistic symbionts and hence rely entirely on carbon supply from host plants (Smith and Read, 2008 ). While they are abundant in root systems of herbaceous plants (Hiiesalu et al., 2014 ), they are hosted also by diverse tree species (Liu et al., 2015 ). Ectomycorrhizal fungi, which consist mainly of the phyla Ascomycota and Basidiomycota, appeared in the era of seed plant diversification (Hibbett and Matheny, 2009 ). In contrast to arbuscular mycorrhizal fungi, some of them may obtain carbon not only from plants but also from soil by decomposing dead organic matter (Talbot et al., 2008 ) (but see Lindahl and Tunlid, 2015 ). Ectomycorrhizal fungi play important roles in forest community dynamics because they promote the dominance of the specific plant families (e.g., Pinaceae, Fagaceae, Betulaceae, and Dipterocarpaceae; Tedersoo et al., 2010 ; Tedersoo and Smith, 2013 ) through “positive plant–soil feedbacks” (Booth, 2004 ; McGuire, 2007 ; Bennett et al., 2017 ). Due to the difference in their major host taxa, arbuscular mycorrhizal and ectomycorrhizal fungi have been considered to form distinct sets of symbioses with their arbuscular mycorrhizal plant and ectomycorrhizal plant hosts (Smith and Read, 2008 ), potentially driving discrete community ecological dynamics. As a consequence, arbuscular mycorrhizal and ectomycorrhizal symbioses have been investigated separately in most mycological studies. Nonetheless, recent studies integrating high-throughput DNA sequencing and host–symbiont network analyses have shown that diverse non-mycorrhizal fungi with broad host ranges are associated with roots of both arbuscular mycorrhizal and ectomycorrhizal plants within terrestrial ecosystems (Toju et al., 2014a , 2015 ). Furthermore, mycorrhizal, endophytic, and other types of root-associated fungi have been reported to co-occur within/around a tiny segment of plant roots (Read and Haselwandter, 1981 ; Mandyam and Jumpponen, 2005 ; Nakamura et al., 2017 ), potentially interacting with each other positively or negatively (Toju et al., 2016b ) (cf. Kennedy et al., 2009 ; Werner and Kiers, 2015 ). Interestingly, an increasing number of studies have shown that non-mycorrhizal fungi (e.g., endophytic fungi) can supply host plants with phosphorous, potentially playing physiological roles similar to those of mycorrhizal fungi (Jumpponen, 2001 ; Narisawa et al., 2002 ; Newsham, 2011 ; Hiruma et al., 2016 ; Almario et al., 2017 ). Thus, host plant ranges of those non-mycorrhizal fungi are of particular interest because they will provide a basis for uncovering potential sharing of soil nutrients between arbuscular mycorrhizal and ectomycorrhizal plants and its consequences on the community- or ecosystem-level dynamics (Kadowaki et al., in review). However, while an increasing number of studies have evaluated host preferences (or generality) of diverse functional groups of root-associated fungi including possible endophytes (Huang et al., 2008 ; Kernaghan and Patriquin, 2011 ; Botnen et al., 2014 ; Sato et al., 2015 ), most studies have investigated either arbuscular mycorrhizal or ectomycorrhizal plant species but not both. Consequently, we still have limited knowledge of how co-occurring arbuscular mycorrhizal and ectomycorrhizal plant species can interact with each other indirectly through below-ground webs of symbioses involving not only mycorrhizal but also diverse non-mycorrhizal fungi. In this study, we statistically examined host preferences of not only mycorrhizal but also root-endophytic fungi in a mixed forest of arbuscular mycorrhizal and ectomycorrhizal coniferous trees in Japan. We sampled roots of Chamaecyparis obtusa (arbuscular mycorrhizal) and Pinus densiflora (ectomycorrhizal) and then revealed community compositions of the fungi associated with the two plant species based on Illumina sequencing. The dataset allowed us to classify those fungi in terms of their host preferences, highlighting endophytic fungi preferentially found from either Chamaecyparis or Pinus , and those commonly associated with both plant species. Overall, this study provides a basis for future studies examining how diverse functional groups of below-ground fungi mediate interactions between arbuscular mycorrhizal and ectomycorrhizal plant species in terrestrial ecosystems.",
"discussion": "Discussion Our data provided a novel opportunity to compare mycorrhizal and non-mycorrhizal fungal communities between arbuscular mycorrhizal ( Chamaecyparis ) and ectomycorrhizal ( Pinus ) plants in a temperate forest. One of the recent conceptual advances in mycology is that plant species in the wild interact not only with mycorrhizal fungi but also with diverse taxonomic/functional groups of endosphere and rhizosphere fungi (Mandyam and Jumpponen, 2005 ; Newsham, 2011 ). Those recent findings challenge the classic view that plant species differing in mycorrhizal type form discrete sets of below-ground plant–fungus interactions. Hereafter, we discuss fungi potentially mediating arbuscular mycorrhizal and ectomycorrhizal symbioses as well as those that preferentially interact with either mycorrhizal type of plant hosts. Many of the fungi found commonly from both plant species belonged to major orders in Ascomycota, namely, Helotiales, Chaetothyriales, and Eurotiales (Figure 2B ; Table 1 ). Among them, Meliniomyces (Helotiales) (Hambleton and Sigler, 2005 ) showed surprisingly high infection rates, appearing in 70 and 86% of Chamaecyparis and Pinus root samples, respectively. Although some species in Meliniomyces–Rhizoscyphus complex have been confirmed to be ectomycorrhizal in pure culture synthetic trials, the most abundant OTU detected in this study was allied to M. variabilis , which has been inferred as saprotrophic, endophytic, or ericoid mycorrhizal (Vrålstad et al., 2002a , b ; Hambleton and Sigler, 2005 ; Tedersoo et al., 2009 ). Interestingly, M. variabilis obtained from a Norway spruce ( Picea abies ) microhabitats lacking ericaceous plants formed ericoid mycorrhizae with European blueberry ( Vaccinium myrtillus ) under experimental conditions, promoting the growth of the host (Vohník et al., 2013 ). Another Helotiales fungus frequently detected from both Chamaecyparis and Pinus roots belonged to the genus Rhizodermea . A fungus in the genus has been reported to enhance heavy-metal stress tolerance of host plants (Yamaji et al., 2016 ). Our analysis also detected Cladophialophora (Chaetothyriales), a lineage of so-called “dark septate endophytes” (Jumpponen and Trappe, 1998 ; Jumpponen, 2001 ). A species in the genus ( C. chaetospira ) has been known to enhance growth and pathogen resistance of host plants (Morita et al., 2003 ; Usuki and Narisawa, 2007 ). Penicillium (Eurotiales) fungi are also reported frequently from roots of diverse plant taxa, although they are generally considered as saprotrophic soil fungi (Watanabe, 2010 ) or postharvest pathogens of fruits (Agrios, 2005 ). However, given the repeated reports of Penicillium fungi from seemingly benign roots of diverse plant species (Cao et al., 2002 ; Toju et al., 2016b ), some of them may play positive roles. Indeed, some Penicillium species associated with wheat are known to solubilize phosphorous in rhizosphere or endosphere (Wakelin et al., 2004 ). Penicillium species are also known to produce a series of antibiotics, which potentially inhibits growth of plant pathogens (Yang et al., 2008 ). We also detected Mortierella and Oidiodendron fungi as common symbionts of Chamaecyparis and Pinus roots (Table 1 ). Fungi in the genus Mortierella are often isolated from soil and root systems in various types of habitats (Watanabe, 2010 ). Although they are generally regarded as saprotrophs, some of them potentially promote plant growth by suppressing root-knot nematodes or phytopathogens such as Rhizoctonia and Cercospora (Eroshin and Dedyukhina, 2002 ; AL-Shammari et al., 2013 ). Fungi in the genus Oidiodendron (anamorph of Myxotrichum ) are also reported from diverse soil environments, while the genus include ericoid mycorrhizal fungi, O. maius and O. griseum (Couture et al., 1983 ; Douglas et al., 1989 ; Rice and Currah, 2005 ; Vohník et al., 2005 ). Oidiodendron fungi were also reported from roots of non-ericaceous plants such as Betula, Picea , and Abies trees in a boreal forest (Kernaghan and Patriquin, 2011 ). While there were ectomycorrhizal fungi frequently detected from both Chamaecyparis and Pinus roots ( Rhizopogon ), no arbuscular mycorrhizal fungi were designated as “host generalists” in our study (Table 1 ). This pattern is of particular interest in light of previous studies reporting asymmetry in host–symbiont associations between ectomycorrhizal and arbuscular mycorrhizal symbioses (Plattner and Hall, 1995 ; Dickie et al., 2001 ). For example, colonization of ectomycorrhizal fungi might be deleterious to non-ectomycorrhizal plants as reported in a herbaceous plant species, whose roots suffered from severe necrosis after infection of the truffle fungus, Tuber melanosporum (Plattner and Hall, 1995 ) (see also Booth, 2004 ). Thus, in the forest studied in this study, the presence of Pinus and its ectomycorrhizal fungi may have negative impacts on Chamaecyparis , although possibilities that those ectomycorrhizal fungi play positive or neutral roles in Chamaecyparis root systems deserve further investigations. Arbuscular mycorrhizal fungi have been also reported to interact with non-typical host plant species. For example, an oak species ( Quercus rubra ) is known to host not only ectomycorrhizal but also arbuscular mycorrhizal fungi in the vicinity of arbuscular mycorrhizal plants (Dickie et al., 2001 ). The nearly complete absence of arbuscular mycorrhizal fungi in Pinus roots in our study (Figure 2A ) highlights context dependency in such host–symbiont associations that span conventional categories of mycorrhizal symbioses. The statistical analysis conducted in this study also allowed us to explore fungal species preferentially associated with either Chamaecyparis or Pinus . As expected, many arbuscular mycorrhizal fungi were found almost exclusively from Chamaecyparis . Meanwhile, a Helotiales fungus in the genus Pezicula (anamorph, Cryptosporiopsis ; Verkley, 1999 ) (Chen et al., 2016 ) showed a strong preference for Chamaecyparis . Given that fungi in the genus produce secondary metabolites (e.g., mullein and echinocandin) that inhibit growth of plant pathogens (Noble et al., 1991 ; Schulz et al., 1995 ; Wang et al., 2014 ), Chamaecyparis hosts may be benefited by the presence of the endophytic fungi. Among the fungi preferentially associated with Pinus , an ascomycete fungus in the genus Neolecta (Neolectales) displayed the strongest host preference. Although their functions remain unknown, Neolecta fungi are known to associate with plant roots (Redhead, 1979 ; Landvik et al., 2003 ): an observation of co-occurrence of a Neolecta fungus and an ectomycorrhizal fungus in root tips (Redhead, 1979 ) is intriguing in postulating their functions. Although these results on potential host preferences are of particular ecological interest, it should be acknowledged that this study did not take into account possible spatial heterogeneity of edaphic factors (e.g., soil pH and C/N ratios) within the study site: there were too many sampling positions to perform detailed chemical analyses. In the dataset, we observed spatial autocorrelations in the occurrences of Chamaecyparis / Pinus root samples (Supplementary Figure 4 ) and fungal community structure (Figure 4 ). To evaluate relative contributions of host preference and spatial environmental heterogeneity, more sophisticated statistical analyses (e.g., latent variable model analyses; Warton et al., 2015 ) needs to be tried in future studies. Our screening of plant-associated below-ground fungi with narrow/broad host ranges provides crucial implications for the understanding of dynamic linkage between plant and below-ground fungal communities (Klironomos, 2002 ; Bever et al., 2010 ; van der Putten et al., 2013 ). Previous studies on arbuscular mycorrhizal plants have shown “negative plant–soil feedbacks”, in which increases of host-specific soil microbes result in decline of the host plant populations (Bever, 2002 ; Kardol et al., 2007 ; Mangan et al., 2010 ). In contrast, positive feedbacks leading to monodominance have been suspected for interactions between ectomycorrhizal plants and their ectomycorrhizal fungi (Booth, 2004 ; McGuire, 2007 ; Bennett et al., 2017 ). While most of those previous studies focused on plant–soil feedbacks operating in interactions involving a single plant species and their mycorrhizal (and pathogenic) fungi, arbuscular mycorrhizal, and ectomycorrhizal fungi often coexist within a forest (Dickie et al., 2001 ; Toju et al., 2014a ), potentially driving feedbacks across different mycorrhizal types (Kadowaki et al., in review). In this respect, the observed asymmetry in infection patterns of arbuscular mycorrhizal and ectomycorrhizal fungi (Figure 2A ) helps us postulate possible directionality in such across-mycorrhizal-type dynamics of plant and below-ground fungal communities. Also intriguingly, this study identified a number of endophytic fungi associated with both arbuscular mycorrhizal and ectomycorrhizal plants and those specific to either mycorrhizal type of plant species (Tables 1 , 2 ; Figure 5 ). Given the prevalence of endophytic fungi and their functional effects on host plant growth (Jumpponen and Trappe, 1998 ; Jumpponen, 2001 ; Newsham, 2011 ), understanding of plant–soil feedbacks would be never complete without taking into account the entire association networks involving not only mycorrhizal but also non-mycorrhizal fungi. Among endophytic fungal taxa potentially playing pivotal roles in such plant–soil feedbacks, the ascomycete order Helotiales (Tedersoo et al., 2009 ; Almario et al., 2017 ; Nakamura et al., 2017 ) is of particular interest because they included not only OTUs specific to either arbuscular mycorrhizal or ectomycorrhizal plant species but also generalist OTUs associated with both categories of host plants. Experimental studies testing the roles of host-specific and generalist endophytic fungi are awaited to build frameworks for describing and forecasting forest community dynamics."
} | 4,450 |
33089655 | PMC7888465 | pmc | 3,059 | {
"abstract": "Mass‐production of clean (i.e. in vitro produced), safe and robust inoculum of arbuscular mycorrhizal fungi (AMF) at affordable costs remains a critical challenge. Engineering plants for enhanced delivery of lipids to AMF could represent an innovative avenue to produce a novel generation of high‐quality and cost‐effective bio‐fortified AMF inoculants for application in agro‐ecosystems.",
"conclusion": "Concluding remarks We envision that future efforts to extend the number of AMF species grown in vitro and to achieve mass‐production of these organisms could focus on lipid engineering of the symbiotic interface. Success in this attempt will inevitably necessitate intense methodological research including combinatorial lipid metabolic engineering, selection of mycorrhized TAG‐accumulating host, lipid flux analysis and spores domestication. This new concept will offer versatile and multi‐beneficial options: (i) increase TAG‐based carbon sources in the AMF, with vesicles, intra‐ and extra‐radical spores accumulating more lipids for a higher germination and root‐colonization potential (bio‐fortification = best quality), ii) stimulate the asexual reproduction machinery to produce more spores in Petri plates and bioreactors (biomass production = high quantity), decreasing costs of in vitro spore production systems (cost‐efficiency = industry profitable). Ultimately, it is expected that research will transform this concept into a novel generation of high‐quality and cost‐effective bio‐inoculant for large‐scale application in agro‐ecosystems.",
"introduction": "Introduction Arbuscular mycorrhizal fungi (AMF) are ubiquitous soil microorganisms forming a mutualistic symbiosis with an estimate of 72% of terrestrial plant taxa. They are key players in agro‐ecosystems, improving plant nutrition (Smith and Read, 2008 ) and increasing their resistance/tolerance to biotic and abiotic stresses (Pozo et al ., 2013 ; Plouznikoff et al ., 2016 ). Unfortunately, overuse of chemical fertilizers and pesticides and intensive agricultural practices have recklessly decreased mycorrhizal diversity in modern agricultural systems (Sosa‐Hernández et al ., 2019 ), a situation which is exacerbated under climate changes believed to impart dramatic effects on plant–soil–AMF interactions by altering the community structure, abundance, composition and functional activity of plant‐associated microbial taxa (Sergaki et al ., 2018 ). Because of that, applying agricultural practices favouring/restoring AMF diversity or introducing selected mass‐produced AMF inoculants into soils have become key challenges of the next big revolution in the development of sustainable agricultural systems."
} | 667 |
25313520 | PMC4196933 | pmc | 3,060 | {
"abstract": "To investigate the distribution and dynamics of microbial community in anaerobic digestion at agitated and non-agitated condition, 454 pyrosequencing of 16s rRNA was conducted. It revealed the distinct community compositions between the two digesters and their progressive shifting over time. Methanogens and syntrophic bacteria were found much less abundant in the agitated digester, which was mainly attributed to the presence of bacterial genera Acetanaerobacterium and Ruminococcus with relatively high abundance. The characterization of the microbial community corroborated the digestion performance affected at the agitated condition, where lower methane yield and delayed methane production rate were observed. This was further verified by the accumulation of propionic acid in the agitated digester.",
"conclusion": "Conclusion The study was conducted to investigate the effect of agitating on anaerobic digestion, focusing on the microbial community dynamics that was revealed to be heavily influenced by agitation. While methanogens and syntrophic bacteria were identified at low relative abundance, species related to Acetanaerobacterium , Ruminococcus and Ruminococcaceae that were known to produce H2 through sugar fermentation were abundant in the agitated digester. The affected digestion performance under agitation may be inferred that the presence of minor amount of H 2 inhibited the syntrophic interaction and VOAs degradation.",
"introduction": "Introduction The divergent effect of agitation on anaerobic digestion has been reported by some studies, while most of which investigated the conventional physiochemical properties [1] , [2] , few evaluated the link between digestion performance and microbial activities, focusing specifically on a small group of bacteria and archaea [3] . However, the ecophysiological role of a great variety of microbes that participate in anaerobic digestion has yet to be fully understood. Classically, environmental microbial communities are analyzed by construction of 16S rRNA clone libraries and the subsequent sequencing of individual clones. The approach, termed Sanger sequencing, has been applied in Tian et al (2013) to compare the microbial community structures for anaerobic digestion at agitated and non-agitated condition, led to the identification of some major microbial phylotypes in bacteria or archaea domain. Due to the fact that a few numbers of clones can be affordably sequenced, Sanger method has its limitation in revealing the whole complexity of microbial communities and is unlikely to adequately represent the genetic diversity. New development of high-throughput next-generation sequencing technologies (NGS), such as 454 barcoded pyrosequencing, not only eliminates the laborious step of preparing clone libraries, but also makes large scale environmental sequencing cost effective and keeps the bias small [4] . In Liang et al., 79% of the genetic variations detected by NGS 454 pyrosequencing were not detected by Sanger cloning-sequencing, especially the low-frequency variations (<20% of amplified population) which can be of major significance [5] . However, phylogenetic assignments based on 454 pyrosequencing could be less precise due to shorter read lengths (200 bp) compared to relatively larger sequence lengths resulting from classical 16S rRNA cloning-sequencing (400 bp and above). In the context of anaerobic digestion, 454 pyrosequencing has been utilized in characterizing biogas-producing communities, revealing a number of new bacteria involved [6] , [7] , but as far as we are aware studying the effect of agitation on entire microbial communities using the technique has rarely been reported. There were limitations in the previous study [8] with respect to determining the complex communities at different taxonomical hierarchies, as well as identification of dynamics of key microbial populations. One of the main objectives of the present study was to address the gap by characterizing the live microbial consortia through pyrotag sequencing, and to relate the community structures to digestion performance and parameters under agitated and non-agitated conditions, such as CH 4 production, soluble Chemical Oxygen Demand (sCOD) and Volatile Organic Acids (VOAs) profiles.",
"discussion": "Discussion Digestion performance comparison The non-agitated digester 1 achieved higher CH 4 yield and CH 4 production rate than the agitated digester 2. In trial 2, digester 1 showed higher sCOD degradation rate and CH 4 production rate than in trial 2 due to that the digester was inoculated with liquor recovered from the previous trial, which had been adapted to tailings decomposition. Digester 2 was also inoculated with adapted liquor from trial 1, but neither CH 4 production nor substrate decomposition showed accelerated rate. This suggested that the agitation affected digester 2 performance adversely. The VOA profiles revealed an initial accumulation of acetic acid in both digesters, but it rapidly disappeared in digester 1 while it persisted for 10 to 12 days before degradation in digester 2. Conversion of propionic acid appeared problematic in digester 2 as high accumulation was observed, especially for trial 2. In anaerobic digestion, acetogenic bacteria ferment propionate to acetate which is then utilized by acetotrophic methanogens to produce CH 4 . The high accumulation of propionic acid in digester 2 probably suggested the inhibition of propionate degradation. The accumulation was further increased in trial 2, indicating digestion inhibition could be exacerbated if the digester liquor was exposed to agitating and used for inoculation continuously. It is well accepted that accumulation of propionic acid indicates an anaerobic process instability [11] , but it can also be considered as the cause of the process failure as its accumulation has been reported inhibitory for methanogens activity [12] . Difference in microbial community composition Digester 1 and digester 2 clearly varied in their bacterial and archaeal community compositions as the PCA ordination indicated. The operating conditions of the two digesters only differed in agitation status, suggesting agitation could have a strong influence driving microbial community of anaerobic digestion. Though both digesters were dominated by phylotype Thermotogales , Petrimonas and an unclassified bacterium, the separation between the digesters was likely linked to the different distribution of some less dominant but important organisms. Methanogens and Anaerobaculum related species were mostly characterized in digester 1, whereas Acetanaerobacterium , Ruminococcus and Ruminococcaceae related species were prevalent in digester 2. The PCA plot formed two main clusters: digester 1 samples from day 19 to day 30, and digester 2 samples from day 22 to day 32, which were retrieved from trial 2. On the contrary, samples from trial 1 were highly variable. This is an indication of the progress that microbial populations shifted substantially from the inoculum community in response to the onset of operation and progressively adapted to tailings decomposition as the operation continued in trial 2. The tighter clustering of digester 2 (day 22 to day 32) suggested a lower microbial diversity as verified by the Chao1 richness estimation (results not shown) in comparison with digester 1. Interestingly, the community evolving between digester 1 and 2 seemed to follow a similar path from day 0 to day 5 but subsequently developed in separate ways to disparate compositions. This could imply the effect of agitation was not instantaneous but rather cumulative. Microbial community composition and digestion performance 16s rDNA sequences reads were compared to the entries of RDP database, and assigned to phylogeneic groups. However, a large portion of sequences were classified to an unclassified phylotype, suggesting the complexity of microbial communities in anaerobic digestion is yet to be characterized. Among identified phylotypes, Anaerobaculum has been known to degrade peptide and a limited number of carbohydrates [13] – [15] . Sugihara et al studied the propionate-degrading ability of a microbial consortium exposed to periodic propionate pulses in sequencing fed batch reactor and reported an Anaerobaculum -related species being dominant [16] . Phylotype Anaerobaculum seemed to play a role in propionate degradation, even though it has not been recognized as a syntrophic propionate utilizing bacterium. It can be postulated that the low abundance of Anaerobaculum in digester 2 probably resulted in accumulation of propionic acid. Archaeal (mostly methanogens) versus bacterial community ratio generally agreed with the reported methanogen proportions ranging from 0.1% to 15% of the total microbial population [17] , [18] . Identified methanogens were closely related to Methanoculleus and Methanosarcina species. Members of Methanoculleus are hydrogenotrophic methanogens [19] , while Methanosarcina species are mostly acetoclasic but also able to use H 2 \n [20] . In digester 1, sequences originating from phylotype Methanosarcina were generally more abundant than from Methanoculleus. The relative abundance of Methanosarcina was seen to follow a dynamic that coincided with the digestion performance. The marked increase from day 3 to day 5 in trial 1 and day 1 to day 4 in trial 2 corresponded to the high CH 4 production rate and the significant reduction of acetic acid during a similar time frame. Methanosarcina spp. has been reported to have higher growth rates and tolerance to pH changes and could potentially lead to stable methenogensis in anaerobic digestion [21] . It should be noted there was an almost complete lack of methanogens in digester 2 throughout trial 2. This does not necessarily imply they are absent, but indicated a fairly low abundance of these organism compared with those in digester 1. It was, however, correlated to the low CH 4 production and the accumulation of propionic acid in digester 2 particular for trial 2. Phylotypes Acetanaerobacterium , Ruminococcus and Ruminococcaceae were found with high relative abundance in digester 2. Those species were closely related and all belonged to family Ruminococcaceae , which are known to degrade cellulose and produce hydrogen (H 2 ) as one of the fermentation products [22] – [24] . Application of Ruminococcus species has been widely used for H 2 production from a variety of feedstock [25] , [26] . Interestingly, some species of Ruminococcus were reported to produce propionate other than ethanol as fermentation products, which could also lead to propionic acid accumulation [27] . Possible functions of selected organisms In anaerobic digestion, hydrogen could be generated through fermentation of intermediate products as sugars or VOAs [28] . However, hydrogen (H 2 ) production is energetically unfavorable due to proton being a poor electron acceptor. The development of syntrophic communities allows H 2 production to become energetically favorable and sustain degradation of organic compounds and production of CH 4 . Due to that syntrophic metabolism, methanogenic activity has to be suppressed in order to produce free H 2 or it would have been readily converted to CH 4 in anaerobic digestion. pH control is desirable for H 2 production because methanogenic activity drops sharply in an acidic environment [29] .However, production of free H 2 at low partial pressure has been reported in anaerobic processes operated at neural or near neutral pH [30] , [31] , which is similar to the operational condition of digester 2 (pH 7.2±0.2). It was speculated that Ruminococcus related bacteria produced free H 2 in digester 2 even though it was typically considered energetically unfavorable. Some studies seemed to support this speculation. Rychlik and May investigated the effect of Methanobrevibacter smithii on growth rate, organic acid production and specific ATP activity of Ruminococcus albus in the co-ulture [32] . The result indicated no increase in the growth rate, acetate or ATP production, suggesting Ruminococcus albus did not receive energetic advantage from co-culturing with the methanogen and the syntrophic metabolism was not preferred. Zhou et al investigated the effect of methanogenic inhibitors on methaneogens and three rumen bacteria Fibrobacter succinogenes, Ruminococcus albus and Ruminococcus flavefaciens \n [33] . While the anti-methanogen compounds effectively reduced the population of methanogens, the inhibiting effect was insignificant or none on the bacterial population, suggesting the syntrophic relation was weak or did not exist. It appeared that, unlike typical hydrogen producing bacteria, phylotype Ruminococcus may not always rely on the syntrophic relationship with methanogens to grow and produced free H 2 in digester 2. Its concentration was expected to be very low that was not analyzed in biogas composition, but may have exerted enough inhibition on methanogensis as discussed below. Phylotypes Desulfotomaculum , Pelotomaculum and Syntrophomonas were detected in both digester 1 and 2 at low proportion (see Table S2 ). Species in these genera are well known as obligate syntrophic bacteria that play crucial role in the degradation of short chain fatty acid such as propionate and butyrate [34] , [35] . The lower CH 4 yield and the low abundance of methanogens in digester 2 may be attributed to the extreme sensitivity of obligate syntrophic bacteria to H 2 . Even at low partial pressure, H 2 can inhibit syntrophic metabolism and thereby limit the substrate supply to methanogens. This inhibition seemed to be cumulative as the methanogen abundance decreased with time. Inhibition of syntrophic interactions also result in accumulations of VOAs, as suggested by the high accumulation of propionic acid observed for digester 2, particularly in trial 2. Menes and Muxi [14] reported H 2 inhibition on glucose utilization by Anaerobaculum mobile , which could explain the low abundance of phylotype Anaerobaculum phylotype in digester 2."
} | 3,540 |
34361434 | PMC8348448 | pmc | 3,061 | {
"abstract": "Macromolecular assembly into complex morphologies and architectural shapes is an area of fundamental research and technological innovation. In this work, we investigate the self-assembly process of recombinantly produced protein inspired by spider silk (spidroin). To elucidate the first steps of the assembly process, we examined highly concentrated and viscous pendant droplets of this protein in air. We show how the protein self-assembles and crystallizes at the water–air interface into a relatively thick and highly elastic skin. Using time-resolved in situ synchrotron x-ray scattering measurements during the drying process, we showed that the skin evolved to contain a high β-sheet amount over time. We also found that β-sheet formation strongly depended on protein concentration and relative humidity. These had a strong influence not only on the amount, but also on the ordering of these structures during the β-sheet formation process. We also showed how the skin around pendant droplets can serve as a reservoir for attaining liquid–liquid phase separation and coacervation from the dilute protein solution. Essentially, this study shows a new assembly route which could be optimized for the synthesis of new materials from a dilute protein solution and determine the properties of the final products.",
"conclusion": "4. Conclusions Natural and synthetic spider silk materials are both mechanically processed to achieve their mechanical properties, usually by a spinning process. Earlier studies showed that the three-dimensional poly(alanine) crystals with the β-sheet structure are formed only during drawing steps [ 38 ]. Future studies may cover both the early formation stages of secondary structures, as well as their transformation into higher-order structures due to additional processing. Another phenomenon in natural spider silk, which strongly depends on the basic formation mechanisms of secondary structures, is the so-called super contraction, where unrestrained spider dragline silk contracts to about 50% of its original length when wetted or exposed to highly humid environments [ 46 ]. Our results also show that a humidity of 80% leads to a significant reduction in β-sheets. Specifically, our results provide insights into the formation processes of secondary structures from solutions containing proteins inspired by spidroins. This work thereby highlights the importance of the water–air interface during the self-assembly of tough silk fiber in the natural spinning process, and clarifies why synthetic processes, i.e., wet spinning of engineered or reconstituent silk proteins, do not recapitulate the properties of native fibers. On a more general level, these results build the basis for further investigations of such formation processes to optimize the production of new materials with tailored functionalities from solutions of engineered macromolecules toward sustainable production of next-generation high-performance material with much broader applications.",
"introduction": "1. Introduction Fabricating materials from biological macromolecules includes utilizing unique molecular interactions to formulate condensed multiscale superstructures. The formation of protein-based materials is closely related to multi-scale supramolecular self-assembly processes and conformational conversion by developing secondary structural motifs, which is not yet fully understood [ 1 , 2 ]. One fascinating example of such ultrastructural material is spider silk, which exhibits exceptional mechanical properties in comparison to any natural or human-made materials. It is a unique material, with high stiffness, strength, and extensibility, and considerable overall toughness [ 3 , 4 ]. Despite extensive study in recent years, and the fact that spider silk and the spinning process have served as a source of inspiration for the design of next-generation high-performance materials, little is known about the intermediate process steps from dilute spidroin solution to the final dried silk fiber, nor the conformational conversions taking place during this process at ambient conditions [ 5 , 6 , 7 , 8 , 9 , 10 ]. Riekel and Vollrath investigated dragline spider silk strand extracted from living spiders in an in situ X-ray experiment by combining small-angle X-ray scattering (SAXS) and wide-angle X-ray scattering (WAXS) measurements [ 11 ]. They showed that the fibers are composed of crystalline and amorphous domains, stabilized by a polymeric network. β-sheet structures were present even shortly after exiting the spider, demonstrating that the secondary structures form either directly at the exit or already inside the body [ 12 ]. There is a general understanding that the remarkable mechanical properties of spider silk are heavily influenced by β-sheets motifs [ 6 , 13 , 14 , 15 , 16 ]. Although this has been mainly investigated for dried silk solutions or fibers, little is known about the formation processes that occur in the protein solutions while drying, or the development of secondary structural features. Several synthetic pathways exist to produce recombinant silk, and they provide an excellent basis for in situ investigations of the formation process of secondary structures [ 17 ]. In this work, we explore the development of conformational structures during evaporation-induced self-assembly of pendant droplets of recombinantly produced spider silk-like proteins ( Figure 1 A,B). Their tri-block architecture is based on our earlier reports consisting of two folded terminal domains and a highly repetitive amphiphilic mid-block spidroin sequence ( Figure 1 A) [ 18 , 19 , 20 , 21 ]. We anticipated the geometry of the pendant droplet interface to be the most suitable approach for understanding the self-assembly and conformational evolution of dense silk solution with concentrations of spidroins in the same range as in the glands before fiber formation ( Figure 1 B). This is otherwise problematic to probe using any other means, such as planar surfaces using very dilute protein solutions, as described previously [ 22 , 23 ]. As a starting point, under controlled relative humidity at room temperature, we formed a pendant droplet in the air from a never-dried 2% w / v protein in pure water for one hour. To our surprise, we noted that the same protein could undergo two very distinct self-assemblies. First, part of the protein is self-assembled and crystallized at the water–air interface into an elastic skin ( Figure 1 ). Second, we observed that the skin acts as a barrier, enabling the protein within to gradually be concentrated and undergo liquid–liquid phase separation (LLPS) and thereby forming liquid-like coacervate (LLC) droplets ( Figure 1 ). We extensively characterized the physio-chemical nature of the LLPS for this protein previously [ 18 , 19 , 20 , 21 ]. Therefore, in this work we focused on a detailed characterization of self-assembly, leading to the formation of the skin with the associated evolution of secondary structures.",
"discussion": "3. Results and Discussions 3.1. Evolution of Skin at the Water–Air Interface To better understand the assembly process of the protein at the interface, we vitrified the entire droplets at various time points in liquid ethane–propane (50%:50%) mixture, followed by fracturing the droplets and probing the evolution of the skin using high-resolution SEM imaging to have direct observation of the assembly. Figure 2 illustrates droplets corresponding to 2 min, 20 min, 30 min, and 50 min into the incubation in air, respectively. The formation of a free-standing coherent skin, covering the entire surface of the pendant drop over time, is clearly evident. The thickness of the skin increased as more protein was absorbed at the interface of the drop. The thickness of the skin ranged from 10 nm to about 25 µm during incubation from 2 to 50 min. Notably, the skin appeared to become increasingly dense as it matured. 3.2. Structural Evolution of Skin To investigate the structural features of the skin at the molecular scale, we followed in situ self-assembly of the protein at the water–air interface using synchrotron WAXS measurements. Figure 3 A shows representative time-resolved 1D WAXS curves extracted from 2D diffraction patterns corresponding to Figure 2 . The measurement was expanded up to 120 min. Scans were performed every minute at 500 µm underneath the blunt needle tip at the center point, keeping the position fixed throughout the measurement. At the onset of the measurement, a strong signal corresponding to the so-called water peak could be seen with the D-spacing ranging from 3 to 3.5 Å [ 36 , 37 ]. Over time, scattering intensity declined due to the gradual evaporation of the water molecules. This gradually resulted in the emergence of the two visible peaks towards the last scans. A sharp peak between 4 and 5 Å, with the maximum peak intensity at around 4.5 Å, is known to correspond to β-sheet interstrand spacing, as described previously, by combining experimental and computational simulation [ 19 , 38 , 39 , 40 ], A broader peak appeared between 8.5 and 11 Å, with its maximum intensity at 10.6 Å. This spacing is commonly assigned to β-sheet interspacing. As the content of β-sheet structures was rising during the drying process, water loss and drying can be interpreted as the key triggering mechanism for crystallization [ 18 , 38 , 39 , 40 ]. These observations are in agreement with complementary RAMAN spectroscopy measurements to determine the conformationally sensitive amide I signal which has been dominantly assigned to β-sheet structures with an emerging peak at a wavenumber of 1657 cm −1 ( Figure 3 B) [ 41 , 42 ]. Additionally, performing polarized microscopy of the skin at various time points placed between two crossed polarizers at 45° with respect to their axis demonstrated an increase in the birefringence of the samples over time, indicative of conformational conversion and evolution of β-sheet motifs ( Figure 3 C). 3.3. Mechanical Properties of the Skin As the skins matured over time into the free-standing, highly elastic 2D membrane in both the hydrated and dehydrated state, we set to determine their mechanical properties. To minimize possible experimental errors during tensile measurements due to inherent variation in the thickness of the skin throughout the surface of a single droplet, and also possible variation between different droplets, we rolled the 2D skins into fiber-shaped materials ( Figure 4 A). This enabled us to measure the cross-sectional area for each specimen accurately by SEM. By doing so, we were able to make fiber-shaped materials with an average diameter of 500 ± 30 μm. Figure 4 A,B shows the results of stress–strain tests carried out for the corresponding samples at 50% RH. As described, the skins were highly elastic in the hydrated state, however their ultimate strength values were substantially low, and therefore we were not able to detect this using our experimental setup. Hence, we exclude any discussion on the hydrated specimens. However, for the dried samples at 50% RH, the general shape of the stress–strain curves showed ductile fracture features [ 43 ]. Stress–strain curves of the skins showed four distinct regimes: (1) a linear elastic deformation region from 0 to 1.12% strain and 0 to 25 MPa stress, (2) a yield point (1.24), (3) a short hardening region expanding from 1.12 to 2.6% strain and 25 to 33.2 Mpa stress, followed by (4) decreasing plastic deformation region until catastrophic failure. The materials showed an ultimate strength of 33.2 ± 4.6 MPa and an ultimate strain of 6.89%. High-resolution fractography after the tensile test revealed the origin of the ductile behavior of the materials. Besides the internal molecular rearrangement and sliding of the protein backbone, we noted the formation of multiscale tears throughout the material, which is indicative of inelastic deformation and a decrease in the tensile properties before undergoing catastrophic failure. 3.4. Absorption Kinetics of the Protein at the Water–Air Interface To better understand the absorption kinetics of the protein at the water–air interface, we performed pendant drop tensiometry for the protein solutions at the various concentrations ( Figure 5 ). In addition to what we indicated as a saturated silk solution (2% w / v ), we also tested very dilute (0.2% w / v ) and highly concentrated (15% w / v ) solutions. We also included MQ water as a control. Data obtained from the DST measurements revealed two distinct trends illustrating the absorption of the protein at the water–air interface, very similar to what is known for surface-active proteins [ 44 , 45 ]. First, with the increase in protein concentration, we noted a substantial decrease in surface tension. Second, surface tension decreased substantially as the droplets were incubated for a longer period. These were only pronounced for 2% w / v and 15% w / v . DST data points for the 0.2% w / v sample remained nearly constant at around 72.5 mN m −1 , indicating that there is no apparent absorption of the proteins at the interface. Data for 0.2% w / v was nearly identical to the MQ water. Both 2% w / v and 15% w / v showed a similar trend in the evolution of the interfacial tension. This includes a steep initial decrease in the surface tension, followed by a very gradual reduction for the rest of the measurements. However, we also observed a few noticeable differences. For instance, surface tension for the 2% w / v decreased from 71 to 58 mN m −1 , whereas this was ranging from 69 to 47 mN m −1 for the 15% w / v protein solution. Additionally, the initial plunge was expanded for around 10 min for the 2% w / v , whereas this was about 20 min for 15% w / v sample. 3.5. Time-Dependent Conformational Conversion of the Proteins at the Water–Air Interface To approach the structural characteristic and conformational conversion more systematically we probed in situ X-ray scattering experiments on drying droplets ( Figure S1, Supplementary Materials ) of 0.2%, 2%, and 15% w / v at various humidities (20, 40, and 80% RH). We primarily examined the most prominent peak at 4.5 Å to evaluate the formation of secondary structures. The area and the position of this peak were probed as a function of time (drying of the droplets) and the circular chord (corresponding to the distance the X-ray beam has to travel through the droplet). Measurements were performed at various positions of the droplets, starting from the edge and moving to the center ( Figure 6 A–F and Figure S2 ). Results demonstrated that the formation of β-sheet structures was directly correlated with the increase in the protein concentration and inversely correlated with the increase in humidity. We noted that the amount of β-sheets, indicated by the peak area ( Figure 6 A–F and Figure S2 ), was lowest for 80% RH and protein concentrations of 0.2% w / v , reaching almost the range of a threshold error. Samples with 2% and 15% w / v at 20 and 40% RH showed a clear appearance of β-sheet structures over time. Both 2% and 15% w / v samples showed relatively similar development of β-sheets over time, with the differences that samples with 15% w/v reach a higher content of β-sheets in comparison to 2% w / v in most humidities. However, it is notable that the 15% w / v sample showed a lower slope ( Figure 6 A–C) of the peak area, increasing overtime at the onset of measurements. Seidel et al. showed that a peak position at around 4.4 Å is correlated to silk with only isolated or partially stacked β-sheet structures, while materials (usually after drawing) with larger microcrystals exhibit a peak for lateral distance at 4.6 Å. Therefore, not only larger peak areas, but also higher peak positions (larger D-spacing) can be assumed to correlate with a higher degree of ordering of β-sheets [ 38 ]. In general, D-spacing rises overtime for all samples ( Figure 6 D–F). It is also important to notice that, in almost all the combinations, a slight delay before β-sheet formation was observed. This was more pronounced for low protein concentrations and high humidities. Contrary to this, the combination of high protein concentrations and low humidities showed the least delay, leading to the assumption that the water content and the drying of the droplet is of great importance for the entire process. Furthermore, looking at the β-sheet formation over the circular chord (corresponding to the distance the X-ray beam has to travel through the droplet), for most samples, we noted that the assembly is higher at the edge of the droplet, but we found no relation between the path through the droplet and the peak position, and therefore ordering, of the β-sheets."
} | 4,190 |
39738176 | PMC11686099 | pmc | 3,062 | {
"abstract": "Nonlinearity is a crucial characteristic for implementing hardware security primitives or neuromorphic computing systems. The main feature of all memristive devices is this nonlinear behavior observed in their current-voltage characteristics. To comprehend the nonlinear behavior, we have to understand the coexistence of resistive, capacitive, and inertia (virtual inductive) effects in these devices. These effects originate from corresponding physical and chemical processes in memristive devices. A physics-inspired compact model is employed to model and simulate interface-type RRAMs such as Au/BiFeO \\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}$$_{3}$$\\end{document} /Pt/Ti, Au/Nb \\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}$$_{\\textrm{x}}$$\\end{document} O \\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}$$_{\\textrm{y}}$$\\end{document} /Al \\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}$$_{2}$$\\end{document} O \\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}$$_{3}$$\\end{document} /Nb, while accounting for the modeling of capacitive and inertia effects. The simulated current-voltage characteristics align well with experimental data and accurately capture the non-zero crossing hysteresis generated by capacitive and inductive effects. This study examines the response of two devices to increasing frequencies, revealing a shift in their nonlinear behavior characterized by a reduced hysteresis range Fourier series analysis utilizing a sinusoidal input voltage of varying amplitudes and frequencies indicates harmonics or frequency components that considerably influence the functioning of RRAMs. Moreover, we propose and demonstrate the use of the frequency spectra as one of the fingerprints for memristive devices.",
"conclusion": "Conclusion This paper explores the nonlinear behavior of BFO and DBMD devices using Fourier analysis, which is crucial for applications such as neuromorphic computing and hardware security. A physics-based compact model based on the cloud-in-a-cell scheme is used to investigate the simultaneous presence of resistive, capacitive, and inductive effects. The capacitance arises from the change in depletion layer width at the Schottky contacts, tunnel barrier, and within the oxide due to the movement of ions or vacancies. Inductive effects occur due to the inertia of the movable defects. The I – V curves show that capacitive effects have a significant impact on the resistive switching of the BFO device, resulting in non-zero crossing hysteresis. In contrast, DBMD’s behavior is less affected due to its lower capacitance. The frequency-dependent I – V curves of both devices exhibit decreasing hysteresis at higher frequencies, approaching linear resistor-like behavior. However, the amplitude spectrum reveals increased nonlinearity at low frequencies and uniform convergence of harmonics. Meanwhile, for high frequencies, the dominance of odd harmonics indicates reduced nonlinear behavior observed from the patternless or nearly no change in internal state attractors due to reduced time for the particles to react. The corresponding phase plots show spiral patterns at low frequencies with randomly varying angles indicating the mixing of different harmonics, while for high frequencies, the phase changes only as multiples of 90 degrees. Based on the observations from this study, we propose using frequency spectra as a fingerprint in addition to conventional I – V curves to better understand nonlinear dynamics.",
"introduction": "Introduction While CMOS devices have been the backbone of electronics for decades, memristive devices have emerged as a promising alternative with distinctive characteristics. Memristive devices exhibit intrinsic nonlinear behavior, as evidenced by the pinched hysteresis in their current-voltage characteristics ( I – V curves), a well-established fingerprint defining their unique attributes 1 . These devices retain information about the amount of charge that has passed through them, presented in the history-dependent resistance function 2 – 4 . As the charge accumulates, the resistance (conductance) of the memristor can alter, and this modification is retained until the charge is reset. While traditional circuit components such as resistors, capacitors, and inductors can exhibit nonlinearity under certain conditions, this is typically less pronounced and more predictable compared to that observed in memristive devices. This is due to the fact that the behaviors of capacitors and inductors are governed by derivatives and integrals, which introduce inherent nonlinearity. Memristive devices, in contrast, inherently exhibit a significantly greater degree of nonlinearity due to their state-dependent resistance and memory properties. This intrinsic nonlinearity is not merely a minor deviation from linear behavior; rather, it is a fundamental characteristic due to various stochastic processes that give rise to distinctive and intricate responses. This pronounced nonlinearity enables memristive devices to exhibit unique and complex behaviors, such as state-dependent resistance and memory properties, making them particularly well-suited for applications in physical unclonable functions (PUFs), true random number generators (TRNGs), and neuromorphic computing. The enhanced nonlinearity in memristive devices significantly amplifies their sensitivity to initial conditions, promoting diverse state transitions and increasing their unpredictability. These characteristics are crucial for the high-performance requirements of PUFs, TRNGs, and neuromorphic computing, where robustness, security, uniqueness, and complex adaptive behaviors are of paramount importance. The functioning of these devices relies on their internal state, which evolves over time and leads to intriguing and intricate nonlinear characteristics 5 . The physical and chemical mechanisms that contribute to resistive switching, such as redox reactions, drift-diffusion of mobile defects, and filament formation 6 change the internal state of the device and thus the resistance of memristive devices. In nanoscale electronic devices, nonlinear behavior can arise from various unknown sources beyond those mentioned here. While memristive devices have prominent variable resistance, capacitive and virtual inductive effects can also contribute significantly to nonlinearity. To be more specific, the inductive effects are not real and are correlated and referred to as inertia effects, as described by Yarragolla et al. 7 , 8 . It can be manifested that resistive, capacitive, and inertia memory coexist in such devices 9 . A detailed explanation is provided in subsequent sections. The nonlinear behavior serves as the most important feature that is crucial for implementing neuromorphic dynamic systems such as memristive reservoirs, memristive oscillatory neural networks, and memristive chaotic systems 1 , 10 . Furthermore, the nonlinear behavior can be strategically utilized as a valuable entropy source for implementing hardware security applications such as physical unclonable functions (PUFs) or true random number generators (TRNGs) 11 , 12 . By taking advantage of this behavior, the specific resistive states, transition patterns, harmonics, and resonant frequencies of memristive devices can be used as intrinsic and device-specific features, which are difficult to replicate or predict. In practical terms, the presence of such harmonics from a frequency spectrum has been empirically confirmed, as demonstrated in experiments on a \\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}$$\\textrm{Ge}_{2}\\hbox {Se}_{3}$$\\end{document} -based self-directed channel memristor conducted by Frank et al. 13 . This evidence highlights the real-world occurrence of phenomena predicted by the simulations in this paper and underscores the potential utility of these characteristics in creating cryptographic keys, seeds, or true random numbers with high entropy. Moreover, although nonlinear behavior may be an important aspect, it presents several obstacles to overcome, such as achieving reproducibility and reliability in the face of unavoidable variability and ensuring stability and predictability in dynamic synaptic-like plasticity. These challenges require adequate investigation before pursuing practical and effective implementation of memristive devices. Therefore, comprehending the origin and examining the nonlinearity of memristive devices is imperative for their efficient utilization in neuromorphic computing and hardware security. Memristive devices often exhibit a pinched hysteresis in their I – V curves, with a zero crossing indicating zero current for zero input voltage. However, some devices, particularly RRAMs, display non-zero crossing hysteresis, suggesting the presence of stored charge. Several studies attribute this phenomenon to resistive, capacitive, and/or inertia effects 9 , 14 , 15 . So, a more appropriate approach would be to consider the change in resistive switching as a shift in memristive device impedance rather than just memristance. This modification reflects the interplay of different effects, which introduces a frequency-dependent impedance, resulting in nonlinearities and intricate responses to sinusoidal inputs. Additional evidence supporting the existence of these effects is shown in Nyquist plots of their impedances, which display double semicircles 16 – 18 . One semicircle can be attributed to the presence of an RC circuit, while a negative curve could be interpreted as an inductive loop. It is important to note that this inductive behavior does not correspond to conventional electromagnetic inductance. Instead, we propose that it arises from inertial effects, which contribute to the formation of the second curve. The appearance of different semicircles in a Nyquist plot indicates the presence of multiple time constants or frequency-dependent behavior in a system. Moreover, as mentioned above, the presence of harmonics in the frequency spectrum of a \\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}$$\\textrm{Ge}_{2}\\hbox {Se}_{3}$$\\end{document} -based device, also provides empirical support for the presence of various RLC-like effects in real devices. Therefore, the coexistence of resistive, capacitive, and inertia effects in memristive devices (as shown in Fig. 1 ) can result in diverse behaviors at distinct frequencies. Therefore, utilizing the frequency-dependent characteristics of memristive devices instead of the commonly used I – V curves can offer a different perspective on understanding the switching in these devices. Fig. 1 Illustration of the concert of co-existence of resistive, capacitive, and inductive effects in interface-type memristive devices. From the previous discussion, it is critical to consider both particle (ion/vacancy) transport and capacitive effects when modeling memristive devices. These two components contribute significantly to the dynamic alterations in memristance. Capacitive effects, encompassing charge storage and redistribution within the device, are particularly imperative and play a pivotal role alongside ion transport mechanisms. The capacitive components of the device react as the charge accumulates or depletes, impacting the charge distribution and, subsequently, the resistance state. The dual contribution of ion transport and capacitive effects enhances the nonlinear properties observed in memristive devices. Understanding the capacitive contributions is critical to comprehending the complex dynamics that result in non-volatility, hysteresis, and memory-dependent changes in resistance. A deeper understanding of the nonlinear behavior of memristive devices can be achieved through comprehensive modeling of ion transport and capacitive effects. Inertia effects are not modeled independently in this work but are included through particle transport. Subsequent sections provide details on how inertia effects are integrated into the overall model for a comprehensive interpretation. The paper examines in detail the nonlinear behavior of specifically interface-type memristive devices. Apart from their nonlinear behavior, these devices offer significant advantages for neuromorphic computing systems and hardware security applications, such as analog-type switching without current compliance, robust endurance, minimal temperature dependence, and reduced variability, making them highly suitable for such advanced applications 12 , 19 – 21 . In contrast, filamentary-type devices, which are more commonly studied, typically exhibit digital switching, less robust endurance, temperature-dependent performance, and require current compliance 22 . To address the knowledge gap and facilitate a deeper understanding of the distinct behaviors exhibited by interface-type memristive devices a cloud-in-a-cell (CIC) scheme-based model for the double barrier memristive device (DBMD) 23 and the bismuth ferrite oxide memristive device (BFO) 7 , 20 is used in this work. Unlike the state-of-the-art compact models 24 – 26 , this method more precisely incorporates the stochastic ion or vacancy transport like the multidimensional computational models 27 – 30 , it is fast and accurate like the state-of-the-art compact models. The model was validated through a series of experimental tests, including assessments of variability, temperature, stress amplitude, retention, and plasticity. The calculated results were in close agreement with the experimental data, demonstrating the accuracy and reliability of the CIC method for modeling the behavior of BFO. This study aims to develop a precise CIC model for interface-type devices, considering capacitive effects, as the first step in understanding their nonlinear behavior. Therefore, to study the nonlinear behavior, the CIC models proposed by Yarragolla et al. 20 , 23 , 31 are further modified to incorporate the capacitive effects. The following section will provide a detailed explanation of the modeling of capacitive effects in interface-type memristive devices. Furthermore, this paper serves as a proof of concept for using the frequency spectra as a fingerprint to identify and characterize memristive devices, providing an alternative tool to the I – V curves for obtaining further insights into nonlinear dynamics. A parametric study is conducted to investigate the variations in resistive switching of the device resulting from various inputs and device parameters coupled with Fourier analysis of the frequency spectra, uncovering harmonics and sidebands. It is important to note that in this study, we focus on understanding the nonlinear behavior of individual memristive devices. This detailed analysis serves as a fundamental step before evaluating the collective behavior of multiple devices typically used in PUFs, TRNGs, or neural networks. Fig. 2 The modified equivalent circuits of ( a ) double barrier memristive device 23 and ( b ) bismuth ferrite oxide memristive device 20 with parallel capacitors across different layers. SC: Schottky contact, TB: tunnel barrier, and SE: solid-state electrolyte.",
"discussion": "Results and discussion The I – V curves shown in Fig. 3 illustrate the nonlinear behavior of BFO and DBMD devices resulting from the coexistence of resistive, capacitive, and inertia effects. To obtain the I – V curves of BFO device shown in Fig. 3 a, an input voltage of 8.5 V was utilized, in conjunction with the parameters used by Yarragolla et al. 20 (given in Supplementary Table S1 ). The figure compares the I – V curves of BFO with and without including capacitive effects. Although the curves with and without capacitive effects exhibit a similar nonlinear change in current with the input voltage, the curve with included capacitive effects highlights a noticeable non-zero crossing hysteresis. This curve depicts the precise analog change in current with the crossing point observed at approximately −2.9 V in the experimental I – V curves. This confirms the presence of charges in interfacial-type memristive devices and how they result in capacitive effects. In contrast to the BFO device, the I – V curves of the DBMD device, as shown in Fig. 3 e, exhibit a notable absence of non-zero crossing hysteresis. This difference can be attributed to the much higher capacitive reactance in the DBMD device as seen in Fig. 3 g, which is at least two orders of magnitude higher than that calculated for the BFO device as in Fig. 3 c. The simulated I – V curves for the DBMD obtained for parameters given in Supplementary Table S2 , both with and without the inclusion of capacitive effects, agree remarkably well with the experimental I – V curves. The reduced capacitance in the pF range contributes to the subtle variations observed in the curves with and without capacitive effects, indicating a minimal effect on the current. Notably, at lower input voltage frequencies for both BFO and DBMD, the inclusion or exclusion of capacitive effects may not result in significant changes in current. However, as the input voltage frequency increases, the capacitive effects become more pronounced, potentially affecting the overall device behavior (detailed later). To more effectively study the resistive switching mechanisms in these devices, we examined the impedance changes over time, as shown in Fig. 3 b and f. We opted to plot impedance to account for resistive, capacitive, and inertial effects, as opposed to the conventional approach of plotting resistance. As the voltage increases, vacancies in the BFO device and ions in DBMD migrate towards the Pt and Au electrodes, respectively, transitioning the device from a high impedance state (HIS) to a low impedance state (LIS) and vice versa. Fig. 4 The I – V of ( a ) BFO and ( c ) DBMD devices for sinusoidal voltage of constant frequency and varying maximum voltage amplitude and their corresponding change in internal state shown in ( b ) and ( d ). We analyze the particle transport by plotting the average mobility and drift velocity of vacancies or ions over time, as shown in Fig. 3 d,h. During the part of the cycle with increasing positive voltage, the particle velocity increases until the activation energy is relatively low, while the mobility remains moderate. For a decreasing voltage from 8.5 V to 0 V, the activation energy is more near the Pt electrode, so the drift velocity and the mobility are reduced. Conversely, for negative voltage, the mobility of vacancies increases by two orders of magnitude, as observed by Du et al. 32 . However, in DBMD, the mobility of ions does not increase significantly, and the velocity remains relatively stable. The activation energy for oxygen vacancy drift in BFO ranges from approximately 0.55 eV at the Au electrode to 0.76 eV near the Pt electrode due to the diffusion of Ti \\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}$$^{4+}$$\\end{document} ions into BiFeO \\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}$$_{3}$$\\end{document} layer 20 . Meanwhile, for oxygen ion drift in DBMD, it remains nearly constant at 0.76 eV across 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}$$\\rm{Nb_{\\rm x}O_{\\rm y}}$$\\end{document} layer 23 . As a result, the velocity of ions in DBMD is significantly lower than that of vacancies in BFO, leading to lower mobility in DBMD compared to BFO. The nonlinearity in memristive devices undergoes significant changes with variations in frequency; therefore, understanding the frequency-dependent behavior of such devices is crucial. For this, different I – V curves of both devices are plotted in Figs. 4 and 6 for a sinusoidal input voltage of different amplitudes and frequencies. The following observations can be drawn from the three plots: (a) For a constant frequency and an increasing maximum device voltage \\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}$$(V_{\\textrm{Device, max}})$$\\end{document} , the hysteresis lobe area increases for BFO shown in Fig. 4 a and DBMD in Fig. 4 c 23 , 47 . This is because the device’s internal state, measured based on the position of oxygen vacancies in BFO (Fig. 4 b) and oxygen ions in DBMD Fig. 4 d, undergoes greater changes at higher voltages. As the voltage increases (decreases), the charged particles experience a greater (lesser) force, resulting in their drift toward the respective interfaces. This leads to an increase (decrease) in current and the difference between the low and high resistance states. However, this argument holds true only for the lower frequency range, a common feature observed in most memristive devices. (b) For a constant \\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}$$(V_{\\textrm{Device, max}})$$\\end{document} , as the frequency increases, the hysteresis loop area decreases as depicted in Figs. 5 a and 6 a 23 , 47 . This phenomenon occurs because the vacancies lack time to jump to a different lattice position and become confined in the same potential well. At lower frequencies, the slower response of the device allows more time for resistive switching mechanisms to manifest, resulting in a larger hysteresis area. As the frequency exceeds 10 Hz for BFO and 1 Hz for DBMD, the I – V curves begin to exhibit behavior similar to that of a normal resistor. Moreover, the change in hysteresis area with frequency in memristive devices can be attributed to the dynamic interplay of capacitive and resistive effects. As the frequency increases, the capacitive reactance \\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}$$(X_{\\textrm{C}}=1/2\\pi fC)$$\\end{document} decreases, the inductive reactance \\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}$$(X_{\\textrm{L}}=2\\pi fL)$$\\end{document} increases, and vice versa. The impedance equation \\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}$$Z=R+j(X_{L}-X_{C})$$\\end{document} illustrates that capacitive and inductive reactances balance at certain frequencies, approaching a more resistive behavior. The behavior of memristive circuits changes from nonlinear to more or less linear resistor-like behavior as the operational frequency varies due to the varying contribution of resistive, capacitive, and inertia effects. Fig. 5 ( a ) The simulated I – V curves of BFO obtained for a sinusoidal input voltage of 8.5 V with various frequencies, Frequency spectra of BFO corresponding to the I – V curves depicted in Figure (a) showing ( b ) amplitude spectra and ( c ) phase spectra, ( d ) the attractors depict the internal state change of the device for different frequencies, plotted for five consecutive input voltage cycles. The legend for figure (a) applies to figures (b) and (c). Points in Figures (a) and (b) that are not immediately visible are superimposed. Fig. 6 ( a ) The simulated I – V curves of DBMD obtained for a sinusoidal input voltage of 3 V with various frequencies, Frequency spectra of DBMD corresponding to the I – V curves depicted in Figure (a) showing ( b ) amplitude spectra and ( c ) phase spectra, and ( d ) the attractors depict the internal state change of the device for different frequencies, plotted for five consecutive input voltage cycles. The legend for figure (a) applies to figures (b) and (c). Points in Figures (a) and (b) that are not immediately visible are superimposed. (c) The presence of different effects in these devices has a substantial impact on the zero-crossing current in memristive devices when the frequency of the applied voltage changes. Figure 5 a shows that the crossing points move further from the origin as the frequency increases 9 , and this effect becomes more pronounced with increasing voltage. As frequency increases, the capacitive current also increases due to the reduced capacitive reactance. This observation was previously made by Yarragolla et al. 7 . The battery effect in a circuit becomes more pronounced with increasing frequency. The battery effect, commonly linked to the existence of capacitive components, is characterized by the storage and release of charges even when there is no external voltage source. The increased capacitance with frequency is accompanied by a decrease in capacitive reactance, which improves the capacity to store charges. The observed increased capacitance leads to a more pronounced battery-like effect, where the capacitive component functions as a virtual battery, storing charges and impacting the circuit’s overall behavior. As a result, this is associated with a shift of the crossing point further away from the origin. Moreover, the effects of inertia also contribute to the shift in the crossing points. This has been explained in detail by Yarragolla et al. 7 . For a frequency of 10 Hz, the change in current is almost linear, and the non-zero crossing disappears. To evaluate the presence of harmonics, we present the frequency spectra of the BFO and DBMD devices in Figs. 5 b and 6 b, respectively, corresponding to the I – V curves shown in Figs. 5 a and 6 a. Both devices exhibit a consistent pattern in their amplitude spectra. At low frequencies, we observe a gradual reduction in spikes, indicating a decrease in the dominance of harmonic components in the signal. This phenomenon can be explained by the fact that the impedance obtained from resistive, capacitive, and inertial components exhibits less variation or better balance at lower frequencies. In other words, a uniform mixing of both even and odd harmonics is present at low frequencies leading to a uniform convergence of the amplitude spectrum. It is important to note that no resonances were observed in BFO and DBMD. However, it is possible that resonances may be present in other devices. Further investigation is required to confirm this possibility. Moreover, the linear I – V curves observed for frequencies >1 Hz for BFO and >0.05 Hz for DBMD suggest a predominance of resistive behavior due to the dominance of odd harmonics in the amplitude spectra. However, the irregular converging harmonics in the frequency spectrum could be possibly originating from capacitiveor inertia effects. At higher frequencies, the capacitive effects within devices become more pronounced, resulting in a capacitor-resistor-capacitor structure. In BFO, the nearly symmetric I – V curves result in diminished or nullified even-order terms in the signal’s Taylor series expansion. As a result, odd-order harmonics dominate the spectrum, despite the presence of nonlinearities. However, for DBMD, the high-frequency I – V curves exhibit slightly more symmetry compared to the low-frequency curves. This leads to non-uniform convergence of harmonics, with irregular dominance of even or odd harmonics as frequency increases. For higher frequencies of more than 10 Hz, the prevalence of odd harmonics becomes more apparent. This suggests a clear trend towards the dominance of odd harmonics. Therefore, the devices can be characterized as displaying linear resistor-like behavior at high frequencies and pronounced inherent nonlinear characteristics at low frequencies. Further investigation through experiments is needed to clarify the specific mechanisms contributing to these observed behaviors and their implications for device performance. Notably, although not for the devices described in this paper, a similar spectrum with such non-uniform converging harmonics has been observed for a Ge \\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}$$_2$$\\end{document} Se \\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}$$_3$$\\end{document} device experimentally 13 . This observation can now be proven to indicate that the occurrence of such non-uniform convergence of harmonics is not a measurement error but rather results from the various RLC effects that change with the frequency. The phase plot of frequency spectra of BFO and DBMD are shown in Figs. 5 c and 6 c, respectively. At low frequencies, due to a balance between multiple frequencies, components such as resistive, capacitive, and inertia effects contribute to phase shifts ranging between \\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}$$0^{\\circ }$$\\end{document} 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}$$360^{\\circ }$$\\end{document} . The wide range of phase angles, and not just the multiples of 90, suggests the mixing of even and odd harmonics at low frequencies, as observed in the amplitude spectrum. A more or less uniform and continuous spiral pattern indicates that the system exhibits nonlinear behavior with respect to phase shifts. Moreover, at higher frequencies, the phase shifts are limited to multiples 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}$$\\pm \\,90^{\\circ }$$\\end{document} , which could be due to the reduced capacitive reactance and increased inertia effects or inductive reactance. This results in phase shifts of even (odd) harmonics at \\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}$$\\pm \\,180^{\\circ }$$\\end{document} and odd (even) harmonics at \\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}$$\\pm \\,90^{\\circ }$$\\end{document} degrees for BFO (DBMD). From these observations, we believe that the intricate combination of device-specific characteristics, nonlinear effects, impedance dynamics, and material factors contributes to the distinct phase shift patterns observed in the memristive devices that are different from one another. For more insights, frequency spectrum plots at different \\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}$$V_{\\textrm{Device, max}}$$\\end{document} are shown in Fig. S4 and Fig. S5 of the supplementary material. To investigate the complex dynamics of memristive devices at various frequencies further, we conducted a detailed analysis of the phase-space attractor plots for both the BFO and DBMD devices. As shown in Fig. 5 d for the BFO device and Fig. 6 d for the DBMD device, we observed unique trajectories corresponding to five consecutive cycles of sinusoidal voltages. At lower frequencies, the phase-space trajectory quickly transitions from the initial conditions into a relatively stable orbit. However, as the frequency increases, we observe a sequence of trajectories typically associated with changes in initial conditions. Memristive devices exhibit a shift in internal state with each cycle due to the reduced response time of the particles at higher frequencies, which prevents them from returning to their original positions. This leads to a change in initial conditions for every cycle, and therefore we observe a horizontal shift in the attractors. At even higher frequencies, the device tends to exhibit linear I – V curves, with the attractors displaying an absence of any noticeable pattern. Furthermore, by subjecting memristive devices to a known input signal and analyzing the resulting output signal, one can calculate the total harmonic distortion (THD) as a measure of the distortion introduced by harmonics. Specific THD signatures based on the inherent nonlinear characteristics of different memristive devices or materials can provide a unique identifier. Variations in THD values can indicate differences in the structural or material properties of the memristive device, which contribute to its unique fingerprint. Using THD as a unique identifier for memristive devices enhances identification by introducing greater specificity and complexity to complement other characterization methods, such as I – V curves and frequency spectrum analysis. Fig. 7 Total harmonic distortion in ( a ) BFO and ( b ) DBMD memristive device. At lower frequencies, the interactions between resistive, capacitive, and inertia components are more, resulting in gradually reducing harmonics and higher. The higher THD is observed for both BFO and DBMD devices, as seen in Fig. 7 . This is because, at lower frequencies, the reactive components (capacitive and inductive) have more time to affect the signal, leading to phase shifts and distortion. In other words, as mentioned above, there is an equal contribution from both even and odd harmonics. On the other hand, higher frequencies exhibit reduced nonlinear effects, and the interactions between different elements tend to be less complex compared to lower frequencies. The reactive components have less time to affect the signal, resulting in a nearly linear resistor-like response with lower THD. The preceding discussion indicates that the frequency spectra of memristive devices can function as a fingerprint, similar to the well-established I – V curves that are critical for device identification and characterization. This observation enhances our comprehension of the nonlinear behavior of these devices. Engineers can use the frequency-dependent behavior to improve device performance, gain insights into device behavior, and implement neuromorphic functionalities such as pattern recognition, memory, and learning. Fourier analysis decomposes the current response into frequency components, which can identify specific patterns within the signal, as shown in Figs. 5 b,c and 6 b,c. These features can be used for machine learning algorithms to classify different patterns. The frequency response of memristive devices can also be applied in the hardware security domain to detect tampering. Any discrepancies in the response would indicate that the device has been tampered with. Additionally, the frequency response of a memristive device can be used to generate random numbers. This is possible because the frequency response of a memristive device is inherently nonlinear, as shown in Figs. 5 and 6 . By applying a periodic input signal to the device and observing the output signal, which is the device’s response, one can derive a sequence of random numbers. The stochastic and nonlinear nature of the device collectively ensures that the generated sequence of random numbers is unpredictable and statistically random. A detailed analysis of the statistical properties, e.g., true randomness, remains for future work."
} | 9,739 |
36837052 | PMC9968140 | pmc | 3,063 | {
"abstract": "Current methods for the protection of metal surfaces utilize harsh chemical processes, such as organic paint or electro-plating, which are not environment-friendly and require extensive waste treatments. In this study, a two-step approach consisting of electrochemical assisted deposition (EAD) of an aqueous silane solution and a dip coating of a low surface energy silane for obtaining a superhydrophobic self-cleaning surface for the enhanced protection of copper substrate is presented. A porous and hierarchical micro-nanostructured silica basecoat (sol-gel) was first formed by EAD of a methyltriethoxysilane (MTES) precursor solution on a copper substrate. Then, a superhydrophobic top-coat (E-MTES/PFOTS) was prepared with 1H,1H,2H,2H-Perfluorooctyltriethoxysilane (PFOTS) for low surface energy. The superhydrophobic coating exhibited anti-stain properties against milk, cola, and oil, with contact angles of 151°, 151.5°, and 129°, respectively. The EAD deposition potential and duration were effective in controlling the microscopic morphology, surface roughness, and coating thickness. The E-MTES/PFOTS coatings exhibited chemical stability against acids, bases, and abrasion resistance by sandpaper. The proposed 2-layer coating system exhibited strong chemical bonding at the two interfaces and provided a brush-like surface morphology with long-lasting superhydrophobicity. The developed method would provide an environment-friendly and expedient process for uniform protective coatings on complex surfaces.",
"conclusion": "4. Conclusions In this work, a two-step process for obtaining superhydrophobic coatings was investigated for enhancing the protection of copper substrates. The base coat was prepared by the EAD process using MTES as the precursor of the sol-gel solution. It is shown that the surface topography, thickness, and surface roughness of the E-MTES films can be adjusted by controlling the deposition time and potential. The hierarchical surface topography of the EAD silica base coat ensured the penetration of the low surface energy PFOTS top-coat to be chemically bonded with the base coat and provided superhydrophobic properties. The E-MTES/PFOTS coatings exhibited excellent self-cleaning properties with WCAs of 153°, 151°, 151.5°, and 129° for milk, cola, and vegetable oil, respectively. The superhydrophobic E-MTES/PFOTS coating showed high chemical stability in corrosive solutions (acids and bases) and high abrasion resistance against sandpaper. The developed 2-layer coating system demonstrated two distinctive benefits including: (1) the EAD process for a complaint uniform base-coat on the complex substrate with high adhesion strength; (2) a simple dip coating of PFOTS for the superhydrophobic top-coat with chemical bonding to the base-coat. The reported method would provide an environment-friendly and widely applicable approach for fabricating sol-gel silica-based superhydrophobic surfaces for enhancing the surface protection of metal substrates.",
"introduction": "1. Introduction Metal surfaces are prone to oxidation and corrosion in ambient conditions and require surface protection. Currently, surface coatings such as organic paint and metallic coatings (e.g., nickel and chrome) by electroplating processes [ 1 ] are applied to provide protection. However, the electroplating process uses harsh chemicals such as strong acids and bases, which are environmentally unfriendly. The process creates some hazardous waste products, which can severely damage the environment if they are not properly disposed of [ 2 ]. Therefore, alternative environmentally friendly processes with enhanced surface protection functionality are in high demand with urgency. Sol-gel coatings are made from hydrolyzed solutions of organometallic precursors, typically silica-based materials, which are chemically stable without harmful substances, and provide environmentally friendly surface protection. However, under harsh environments, sol-gel thin-film coating alone may be insufficient for surface protection. Inspired by the self-cleaning effect of lotus leaves [ 3 ], self-cleaning coatings are expected to repel liquids (including water, acids, bases, and solvents) from contacting the metal surfaces and, therefore, further enhance surface protection function. Superhydrophobic surfaces with water contact angles >150° have been shown to exhibit the self-cleaning effect and are studied for a wide range of applications [ 4 ]. Various strategies for constructing superhydrophobic surfaces on metal substrates have been investigated in the past decades [ 5 , 6 ]. Techniques for surface modification include chemical etching [ 7 ], laser patterning methods [ 8 , 9 ], wet etching [ 10 ], Chemical Vapor Deposition (CVD) [ 11 ], micro-arc oxidation (MAO) [ 12 ], and surface coating of intrinsically low energy materials, such as PDMS, fluoropolymers, and fluorinated silanes [ 13 , 14 , 15 , 16 ]. These methods either employ expensive machines or use harsh chemicals to change the metal surface conditions. These complex or environmentally unfriendly processes have hindered further widespread applications of protective self-cleaning coatings for metals, especially for 3D complex surfaces. Superhydrophobic coatings are usually prepared by a two-step process, where a surface with sufficient roughness is firstly constructed by various physical and chemical methods [ 17 , 18 , 19 , 20 ] and then further modified with low surface energy materials. Sol-gel methods [ 21 , 22 ] are increasingly adopted for the preparation of superhydrophobic surfaces due to the wide range selection of precursors, the relative simplicity of the wet chemical process, and the use of inexpensive and environmentally friendly raw materials. Conventionally, silica-based sol-gel coatings are applied by dip-coating, spray-coating, roller-coating, or spin-coating [ 23 , 24 ] processes. The driving force for the generation of thin films by these self-assembly-like methods relies heavily on the degree of the hydrolytic condensation reaction of the sol-gel monomer [ 25 ]. For this reason, it is difficult to deposit thicker silica films uniformly, especially on 3D complex industrial parts [ 26 , 27 ]. The inherent mechanical properties and weak adhesive strength of inorganic sol-gel coatings on metal substrates have restricted their industrial applications as a pre-treatment layer for coating systems. EAD is a recently developed method for the preparation of coatings on conductive materials based on the sol-gel principle [ 28 ] and allows for the preparation of thicker and rougher sol-gel films. EAD is based on the principle of in-situ catalysis of sol-gel chemistry (hydrolysis and condensation reactions) by the application of negative potential. The immediate condensation of the solute on the electrode surface can be accelerated by the reduction of oxygen or some specific ions supporting the electrolyte or by hydrolysis, which leads to the increase of pH value near the electrode surface. The chemical reaction formula in this process is expressed as Equations (1)–(3) [ 29 , 30 ].\n 2H 2 O + 2e − = 2OH − + H 2 (1) \n O 2 + 2H 2 O + 4e − = 4OH − (2) \n NO 3 + H 2 O + 2e − = NO 2− + 2OH − (3) A schematic diagram of the described reaction process is shown in Figure 1 a. It is worth noting that during the above “electrodeposition” process, the silane component does not lose or gain electrons at the electrode surface and the pH value of the bulk solution itself does not change throughout the process, and no new substances are introduced [ 31 , 32 , 33 ]. Compared to conventional self-assembled films, these electrodeposited films (e-films) have several unique advantages: (i) The base catalysis near the cathode surface provides an additional driving force for film formation, resulting in thicker and rougher films [ 34 , 35 ]; (ii) Gelation is separated from solvent evaporation and occurs during electrodeposition, producing films with greater porosity and better intra-film cohesion; (iii) The derived OH- ions also catalyze the chemical bonding process between the sol-gel film and the substrate [ 36 ]. EAD has the significant advantage of depositing coatings on complex non-planar geometries and controlling the thickness and composition of the nanocomposite and ensuring the surface structures are ohmically connected to the metal substrate. Mandler et al. [ 37 , 38 ] demonstrated the electrochemical co-deposition of sol-gel monomer with copper ions or gold nanoparticles leading to the formation of thin nanocomposite films. Their results showed that the ratio of sol-gel monomer to copper ions/gold nanoparticles strongly influenced the morphology and grain size of the films. This method demonstrated the general applicability of EAD in templating nano-sized objects in thin films. Electrodeposited silica-based films are often used in research on porous electrode materials, electrochemical sensors, biodegradable polymer films, and the preparation of superhydrophobic films [ 39 , 40 , 41 ]. Compared with conventional electrochemical plating processes, EAD requires lower energy input and no need for waste treatment. Therefore, EAD offers a low-cost and environmentally friendly process. It is suited for producing relatively thicker and rougher films uniformly on complex conductive surfaces and thus provides a new and convenient method for preparing superhydrophobic surfaces. The construction of superhydrophobic surfaces can be realized by the one-step method or two-step method. The former is achieved by one-step deposition of inorganic and organic composite film from a mixed silane precursor solution with various fillers [ 42 ]. The high content of organosilane components in the one-step superhydrophobic films tends to cause weaker mechanical properties than pure inorganic silica films. The latter is obtained by the EAD of silica films followed by low surface energy modification with long alkyl chain organosilanes or fluorosilanes [ 43 ]. Wang et al. [ 44 ] prepared superhydrophobic membranes with a water contact angle of 153.5 ± 3° in a two-electrode cell using a synthetic 3-undecyl-4-amino-5-mercapto-1,2,4-triazole (UAMT) in a one-step electrolytic method using copper foil as the anode and platinum wire as the cathode. Wu et al. [ 45 ] reported the deposition of silica films at a typical constant potential of −1.3 V on 306 L stainless steel (SS) or indium tin oxide (ITO) glass electrodes, followed by hydrophobic modification with dodecyltrimethoxysilane (DTMS) to obtain excellent water repellency. Further, Hu et al. [ 46 ] mixed dodecyltriethoxysilane (DTMS) and tetraethoxysilane (TEOS) sol-gel precursors and prepared SiO 2 /DTMS hybrid sol-gel superhydrophobic films on mild steel substrates by a one-step electrodeposition method. A two-step method to create superhydrophobic surfaces on mild steel was reported by Zhang et al. [ 43 ]. First, highly porous and graded nano/microstructured silica films were prepared by EAD, and then were further treated with long alkyl chain dodecyltrimethoxysilanes. The superhydrophobicity provided by the rigid SiO 2 matrix showed high mechanical durability against abrasion and good repair ability for thermal damage treatments. Despite the above-mentioned methods, the combination of EAD and self-assembly of the sol-gel top layer to obtain superhydrophobicity on the copper substrate has not been systematically investigated. In this work, a two-step method is studied for the preparation of superhydrophobic coatings on a copper substrate. The first step is to generate a rough micro/nanostructures sol-gel coating on Cu substrates from MTES precursor solutions by EAD. The effects of key EAD parameters including deposition potential and duration on the surface morphology and thickness of the films are investigated. The second step is to modify the sol-gel coatings by fluorination with PFOTS to obtain superhydrophobic surfaces with low surface energy. The coating stability, durability, and self-cleaning ability are compared to those prepared by the one-step method. It is aimed to provide an environment-friendly and expedient process for uniform protective coatings.",
"discussion": "3. Results and Discussion 3.1. Process Control of the EAD of E-MTES Coatings During the electro-deposition process, the time and current were recorded for each coating sample. Figure 2 a shows the correlation between the deposition current and the different deposition time under a typical deposition voltage of −1.2 V vs. Ag/AgCl. Due to the double layer charging process at the electrode and solution interface, the potential-step curve [ 48 ] was used in this setup. The current increased sharply during the first few seconds of recording and then gradually reached a quasi-steady state value. It was also noticed that the current response curve flattened when the deposition time reached 600 s. This indicated the completion of the film growth on the substrate. The current-time relationship of the samples at different voltages was studied using 600 s as a threshold value. As shown in Figure 2 b, with the potential absolute value increased from −1.0 V to −1.4 V, the current curve shifted to more negative values. The higher current densities and the higher applied potentials led to a larger amount of hydroxide produced and, thus, the higher pH values at the electrode surface and increased the rate of film formation. Therefore, the film thickness can be adjusted by varying the potential of the working electrode, which controls the film formation rate. 3.2. Surface Microstructure and Coating Thickness by EAD The deposition process of sol-gel silica films depends on various parameters, such as deposition time and applied potential. Figure 3 shows the FE-SEM images of the top-view of the E-MTES coating on Cu substrates, showing that film deposition took place consecutively when prolonging the duration at the applied potential of −1.2 V vs. Ag/AgCl. The FE-SEM results showed that the surface morphology of the film changed with increasing deposition time. At a shorter deposition duration ( Figure 3 a,b), the surface was relatively flat, and there were still visible polishing marks on the copper surface. This was due to the incomplete growth of the film, which had not yet covered the polished marks on the Cu substrate. As the deposition time increased, the surface became rougher. At a deposition time of 200 s, the film consisted of a tightly packed arrangement of micro- and nano-sized silica spheres ( Figure 3 c,d). As the deposition time was increased to 400 s, the E-MTES films continued to grow, with flocculent polymers gradually covering the surface ( Figure 3 e,f), resulting in thicker films. This was attributed to more OH − aggregation in the vicinity of the cathode, which facilitated the condensation process. When the time was increased to 600 s, the E-MTES films showed a layered morphology consisting of micro-structures and the films appeared highly porous ( Figure 3 g,h), leading to a further increase in the surface roughness. The film-growing principle of the EAD process consists of a few stages, from nucleation, growing into an island, further growing into continuous film, and then becoming porous due to the dissolution of the already formed film under prolonged immersion in the chemical solution (600 s). The rough and porous morphology facilitated the subsequent fluoro-silane surface modification to construct superhydrophobic surfaces. Cross-sectional SEM images and measurements of the coating thickness values at different deposition times: (a) 100 s, (b) 200 s, (c) 400 s, and (d) 600 s are shown in Figure S1 . Table S1 lists the coating thickness values with varying deposition durations. The thickness of the coating increased exponentially with time. Similarly, the cross-sectional SEM images and measurements of coating thickness values at different deposition potentials: (a) −1.0 V, (b) −1.1 V, (c) −1.2 V, and (d) −1.3 V are shown in Figure S2 . The effect of deposition potential on the coating surface morphology was studied using 600 s as a time reference for all the samples ( Figure 4 ). Typically, organosilane precursors underwent hydrolytic condensation to produce silica nuclei and growing “islands” that gradually covered the substrate and formed the coating after curing. In this study, an initial “triangular” micro-structure was observed by EAD ( Figure 4 a,b) at −1.0 V, and the EDS elemental analysis ( Figure S3 ) showed that the structure was mainly composed of Si and O elements. With increased deposition potential values, the micro-structure became finer ( Figure 4 c,d) and more inter-connected, as shown in Figure 4 e–h. As the deposition potential became more negative, more OH − ions accumulated at the cathode and at the same time accelerated the condensation process, resulting in a rougher and more porous E-MTES coating ( Figure 4 f,h). The coating thickness values were measured on the cross-sectional SEM images of the samples as shown in Figures S1 and S2 . The thickness values were plotted against electrochemical deposition potential and deposition time as shown in Figure 5 and summarized in Table S2 . It is seen from the curves and the values that with more negative deposition potential, the film thickness increased exponentially. Similarly, the film thickness increased with increasing deposition time. 3.3. Chemical Analyses of the Coated Surfaces The coating surface with a deposition potential of −1.2 V vs. Ag/AgCl for 600 s was further analyzed. The FE-SEM image and EDS spectra and elemental mapping are presented in Figures S4 and S5 . The E-MTES film contained a high amount of C, O, and Si, indicating that the coating had been successfully deposited onto the copper surface. The atomic weight ratios of C, O, and Si elements at different deposition times are shown in Figure 4 a–d. The results indicate that the Si content of the coating increased with increasing deposition time, which was consistent with the results observed in FE-SEM ( Figure 3 ). Figure 6 shows the surface profile analyses and chemical analyses of the superhydrophobic coating. It can be observed from Figure 6 a that the surface became relatively smoother after fluorination modification, and the pores were “sealed.” Figure 6 b shows the 3D morphology of the superhydrophobic coating, corresponding to Sa = 8.35 μm and Sq = 10.67 μm. The FT-IR spectra of the E-MTES and E-MTES/PFOTS coatings are shown in Figure 6 c. The band at 2970 cm −1 is due to C–H stretching in CH 3 . The band at 1270 cm −1 corresponds to CH 3 symmetric bending in Si-CH 3 [ 40 ]. The main absorption peak at 1025 cm −1 corresponds to the asymmetric stretching vibration of Si-O-Si, which is the result of the silanol condensation reaction [ 49 ]. The results confirmed the formation of silica film on the copper substrate. With the increase of negative potential, the cathodic OH − concentration increased, which greatly promoted the formation of well-grown cross-linked network chains with Si-O-Si bonds. The band at 1118 cm −1 is ascribed to Si-O-C. The band at 779 cm −1 shows that the condensation of Si-OH groups was almost completed and only a few Si-O-CH 2 CH 3 groups remained [ 50 ]. The weak bands at 1228 and 1190 cm −1 observed on the superhydrophobic surface E-MTES/PFOTS coating were assigned to C-F stretching vibration. These results confirmed that PFOTS molecules had been successfully assembled on the porous E-MTES surface by the dip-coating method. To further confirm the formation of the PFOTS film, a high-resolution XPS spectrum of the PFOTS hydrophobic coating is shown in Figure 6 d. The binding energy at 528.28 eV and 99.81 eV are O1s and Si2p, respectively. The strong energy peak at 683.88 eV represents elemental fluorine, which confirms the successful grafting of the PFOTS molecule. The fluorine modification is an important factor to achieve the desired superhydrophobicity. The elemental F energy peak was also observed in the two EDS spectra shown in Figure 6 a, which were measured at two points on the superhydrophobic surface. The F atomic contents of these two points were about 10%, which further confirmed the result obtained by XPS. Figure 7 illustrates the chemical structure of each coating layer and the condensation reactions between hydroxyl groups of hydrolyzed silanes, which led to the formation of Si-O-Si bonds upon drying and condensation. This confirms the chemical bonding between layers and the brush-like fluoro-chains, which provided intrinsically low surface energy status. The coatings were subjected to a crosshatch adhesion test and achieved the grade “5B” (“5B” is the best grade of the adhesion test) according to ASTM D3359 [ 47 ]. 3.4. Surface 3D Profile and Roughness It is well known that micro/nanostructures play a crucial role in the wettability of the coating surface. The surface roughness of the coating was assessed according to the deposition time and deposition potential. Figure 8 shows the 3D morphologies and 2D profiles of the sample surfaces after different deposition times. It is seen that when deposition time is short, the surface morphology is not uniform with the deposition occurring at some preferred area or points on the surface. With increasing deposition time to 600 s, the layer grew and gradually covered the whole surface, leading to a more regular and uniform surface structure. The increase in surface roughness was in good agreement with the FE-SEM images shown in Figure 3 g,h, where the porous film was formed. Figure 9 shows the 3D morphologies and 2D profiles of the sample surfaces at different deposition potentials. It is seen that deposition potential was more effective in controlling the surface morphology. With deposition potential becoming more negative at −1.2 V and −1.3 V, the surfaces became rougher with more regular structures as shown in the 2D profiles of the surfaces in Figure 8 d and Figure 9 c,d, in which the peaks and valleys in the 2D profiles were more evenly distributed. Correspondingly, Figure 10 shows the average roughness (Sa) and the root means square roughness (Sq) of the surfaces. It is clear that as the deposition time increased and the deposition potential was more negative, the surface gradually became rougher, and the surface roughness increased almost exponentially. As shown in Figure 10 , after increasing the deposition time from 100 s, 200 s, 400 s, to 600 s, Sa increased from 0.203 ± 0.0288 μm, 0.438 ± 0.079 μm, 2.17 ± 0.271 μm to 8.29 ± 0.829 μm, respectively, and Sq increased from 0.308 ± 0.056 μm, 0.833 ± 0.221 μm, 3.50 ± 0.246 μm to 10.56 ± 0.985 μm, respectively. Similarly, when depositing potential changed from −1.0 V, −1.1 V, and −1.2 V to −1.3 V, Sa increased from 0.306 ± 0.044 μm, 0.431 ± 0.017 μm, 5.01 ± 0.334 μm to 8.51 ± 0.244 μm, and Sq increased from 0.519 ± 0.114 μm, 0.636 ± 0.077 μm, 7.55 ± 0.392 μm, to 10.75 ± 0.185 μm, respectively. These results indicate that the EAD process parameters (e.g., time and potential) played a significant role in improving the coating surface morphology and facilitating the further formation of superhydrophobic surfaces. The regular distributed peaks and valleys on surface morphology acted as the anchoring force between the top-coat and base-coat and thus ensured strong adhesion between the two layers. By combining the strong chemical bonding between the silica basecoat and the fluorinated silane top-coat with the physical interlocking force, the 2-layer coating system is expected to provide long-lasting superhydrophobicity as well as mechanical durability. 3.5. Water Contact Angle (WCA) Measurements and Stability Tests Contact angle measurement is the main method to evaluate the surface wettability of a solid material. The contact angles of the electrochemically deposited silicon films were measured by the Sessile drop method using a contact angle instrument [ 51 ]. Figure 11 a shows the WCAs of the coatings made by EAD at different deposition parameters, with a comparison to the dip-coated surfaces at the same immersion times. It is seen that both coating methods showed increasing trends of WCAs with increasing deposition time, while EAD-coated surfaces exhibited higher WCAs (>100° after 200 s). Figure 11 b shows the WCAs with varying EAD deposition potential. It shows a maximum WCA of 104° at −1.2 V. Figure 11 c shows WCAs of bare copper before and after fluorination modification. Compared to the bare copper with a WCA of 57°, the fluorination-modified surface showed hydrophobicity (WCA of 126°). Figure 11 d compares the WCAs of E-MTES coating (prepared at −1.2 V vs. Ag/AgCl) before and after fluorination modification. It shows that the EAD-coated surface had a WCA of 105°, while after fluorination modification, the WCA increased to 153°. This is attributed to the fact that the silane precursor MTES has a terminal group of -CH 3 , which is unable to produce long-chain (C-H) bonds. It is also difficult to obtain a superhydrophobic surface by relying only on the surface morphology of the E-MTES coating. The fluorination modification provided a long –CF chain at the outer surface of the EAD-coated rough surface, which enabled superhydrophobicity. The chemical stability of the superhydrophobic surface was evaluated by measuring the WCAs with water of different pH values ranging from 1 to 13, and pH adjustment was done using HCl and NaOH. As shown in Figure 11 e, the WCAs of the E-MTES/PFOTS coating remained above 150° with a sliding angle of less than 10° at all times and the pH variation did not affect the WCA. The results demonstrated that the E-MTES/PFOTS coating had great resistance to both acidic and basic solutions. The superhydrophobicity of the E-MTES/PFOTS coating confirmed the mechanism for producing superhydrophobic surfaces by combining surface morphology with low surface energy coating materials. When a liquid comes into contact with a rough superhydrophobic surface, it cannot fully penetrate into the surface structure but forms a heterogeneous wetting state with a composite interface combining solid, liquid, and air. The Cassie-Baxter model can be used to analyze the heterogeneous wetting state, and the equation is described as Equation (4) [ 52 ]: (4) cos θ = γ S f S cos θ 1 + ( 1 − f S ) cos θ 2 \nwhere f S describes the area fraction of the liquid−air interface occluded by the surface texture, and γ S describes the non-dimensional surface roughness factor of the superhydrophobic surface. θ is the real contact angle of the superhydrophobic surface, θ 1 is the equilibrium contact angle on solid, and θ 2 is the equilibrium contact angle on air. Since the value of the equilibrium contact angle on air is 180°, thus Equation (4) could be written as Equation (5) [ 53 ]: (5) cos θ = γ S f S cos θ 1 + f S − 1 Therefore, increasing the portion of the liquid-air interface and reducing the portion of the solid-liquid interface on a superhydrophobic surface can enhance its water repellency. The E-MTES coating’s highly porous micro-/nano-structure traps more air into the surface structure, thus increasing the fraction of the liquid-air interface. Therefore, after modifying PFOTS, the sample could reach a higher WCA. 3.6. Self-Cleaning Test and Durability Evaluation To demonstrate the self-cleaning property of the superhydrophobic surfaces, black ink was used to stain the surface. The contact angle was measured at least three times on each sample to ensure accuracy. Samples were tested for self-cleaning at a tilting angle of 8°. Figure 12 a shows the self-cleaning test results of the superhydrophobic surface with water-based black ink. It can be seen that when the ink was dropped onto the surface from the top, the droplet rolled downwards to the bottom, leaving no trace on the surface, which indicated that the surface was stain-repellent. Another self-cleaning test was conducted on the superhydrophobic surface by staining with carbon black powder as shown in Figure 12 b. The carbon black powder was randomly spread on the surface, and then water droplets were applied from the top of the sample. It is seen that the carbon black powder was easily removed, and no black residue was observed on the E-MTES/PFOTS surface. Therefore, the superhydrophobic E-MTES/PFOTS coating with self-cleaning capability would be suitable for practical applications to prevent contamination on copper surfaces. Figure 12 c–f are schematic illustrations of the self-cleaning behavior. Figure 12 c,d shows the movement of a water droplet on bare copper, where the droplet almost mixed with the contaminant due to the hydrophilic nature of the surface and the wettability of the substrate (CA < 90°). When the water droplet rolled on the superhydrophobic surface ( Figure 12 e,f), it rolled off together with the contaminant due to the water repellency of the superhydrophobic surface, leaving a clean surface after the experiment. Figure 13 a shows the process of the anti-sticking test of the superhydrophobic surface. The testing process was performed by squeezing the end of a fixed micro syringe allowing a water drop (size about 3 μm) to form at the tip and moving up and down the platform below to collect the drop. It was observed that even with forced contact of the water drop with the surface, the water drop was easily separated from the superhydrophobic surface without leaving any visible residues. During the contact, the water drop deformed and deflected to the side of the needle, indicating the non-sticking property of the surface. Figure 13 b shows the testing process of dust removal by water droplets. The dust was absorbed by the water droplet and well removed from the surface, confirming the anti-sticking property of the surface. In addition to water, the coating was shown to exhibit excellent resistance to wetting by various liquids, including milk, cola, and vegetable oil ( Figure 13 c) with contact angles of 151°, 151.5°, and 129°, respectively. This indicates excellent self-cleaning properties of the coated surfaces, which would have potential applications for daily uses. Wear and abrasion resistance pose a serious challenge for many practical applications of superhydrophobic surfaces under natural environmental conditions. The superhydrophobic coating was abraded using 1000-grit sandpaper (contact area 150 mm 2 ) moving 20 cm distance per cycle under a load of 50 g. The abrasion resistance of the superhydrophobic E-MTES/PFOTS coating was evaluated on the basis of the measured wettability after 10 cycles of abrasion by sandpaper. As shown in Figure 13 d, the WCA of the E-MTES/PFOTS coating surface did not change noticeably after 10 cycles and remained superhydrophobic (CA = 151°) with a sliding angle of about 8°. The good abrasion resistance of the 2-layer coating is due to the strong chemical bond between the basecoat and top-coat, as well as the complaint basecoat with a chemical bond on the copper substrate (as shown in Figure 7 ). The strength of these chemical bonds is much higher than any physical bonds, which demonstrates the intended design mechanism of the 2-layer coating system in this study. 3.7. Mechanisms of Chemical Bonding and Durability of Superhydrophobicity The excellent abrasion resistance and durability of the superhydrophobicity of the developed 2-layer coating system are believed due to the strong chemical bonding and physical interlocking between the two layers, as well as the chemical bonding between the E-MTES layer and the substrate induced by the EAD process. Figure 14 illustrates the mechanism of the 2-layer coating structural design. As shown in the chemical reaction formula in Equations (1)–(3), the EAD process can accelerate the reduction of oxygen or increase the pH value near the electrode surface and induce more –OH groups by hydrolysis, which cause immediate condensation of the solute on the electrode surface and formation of the Si-O-Si network of the base-coat. Due to the concurrent deposition and reduction processes, a porous and rough surface is formed. This rough surface would provide local gaps for the penetration of the PFOTS top-coat. Meanwhile, the hydrolyzed silane groups in the PFOTS material would form a condensation reaction with the hydroxyl groups in the base coat. Therefore, strong bonding is established at the interfaces of the top coat and the base coat. The fluoro-side groups in the PFOTS chemical structure have high stability and tend to protrude on the surface providing a brush-like structure and long-lasting superhydrophobic properties. Some hydrophilic nano-silica particles were added into the PFOTES solution during hydrolysis, in which the nano-particle surfaces were grafted with the PFOTES’s fluoro-chains to further enhance the favorable surface morphology for superhydrophobicity. 3.8. Comparison with Other Research Work in Literature Table 2 provides a comparison of the roughness, WCA, and reported properties of this work with other published results in literature. It is seen that most of the superhydrophobic coatings consist of silanes and fluoro-containing substances. Compared to the published results, this work has the advantage of using mild chemicals applied for copper substrate with a thicker and more durable coating."
} | 8,347 |
38046629 | PMC10690873 | pmc | 3,064 | {
"abstract": "Recently, there has been a significant increase in academic and industrial interest in self-healing polymers (SHPs) due to their remarkable ability to regenerate scratched surfaces and materials of astronomical significance. Scientists have been inspired by the magical repairing mechanism of the living world. They transformed the fiction of self-healing into reality by designing engrossing polymeric materials that could self-repair mechanical abrasions repeatedly. As a result, the durability of the materials is remarkably improved. Thus, the idea of studying SHPs passively upholds economic and environmental sustainability. However, the critical areas of self-healing (including healing efficiency, healing mechanism, and thermo-mechanical property changes during healing) are under continuous scientific improvisation. This review highlights recent notable advances of SHPs for application in regenerating scratched surfaces with various distinctive underlying mechanisms. The primary focus of the work is aimed at discussing the impact of SHPs on scratch-healing technology. Beyond that, insights regarding scratch testing, methods of investigating polymer surfaces, wound depths, the addition of healing fillers, and the environmental conditions maintained during the healing process are reviewed thoroughly. Finally, broader future perspectives on the challenges and prospects of SHPs in healing surface scratches are discussed.",
"introduction": "1. Introduction Self-healing is a ubiquitous surviving prowess of the living world to repair physical damage routinely. The intrinsic repairing ability in nature has inspired scientists to develop self-healing polymers (SHPs) with recent application in vital areas, including electronic gadgets, 1 sensors, 2 e-skin, 3 adhesives, 4 supercapacitors, 5 wearable devices, 6 automobiles, and steel coatings. 7,8 The self-repairing properties in synthetic materials have improved the final design's lifetime, strength, and safety. This can substantially decrease per capita purchases, industrial waste, and environmental pollution. 9 Thus, SHPs can exquisitely reverse climate change. 10 In the 1970s, as interfacial macromolecular interpenetration began to be studied as one of the repair mechanisms for damage occurring in thermoplastic resins, the concept of self-healing emerged as a research topic for polymeric materials. 11 However, it is not an exaggeration to say that the systematic research on full-scale self-healing began with the study of the microcapsule polymer complex by White's group in 2001. 12 This is the first extrinsic self-healing polymer composite, and this concept was expanded to microvascular network in 2007 by the same group. 13 In 2002, Chen et al. presented a polymer with a cross-linked structure that can be re-mendable by heat using the Diels–Alder/retro-Diels–Alder (DA/rDA) reaction, and which is the first intrinsic self-healing polymer based on reversible covalent bonding. 14 In the same year, research on shape memory effects and self-healing began. 15 In 2008, a self-healing polymeric material based on a supramolecular assembly relying on hydrogen bonds was presented by Leibler et al. , pioneering as the first self-healing rubber using non-covalent bonds. 16 In 2011, the Leibler group first reported the concept of ‘vitrimer’, which can maintain a constant crosslinking density of polymers that can be reformed by heating based on transesterification reaction of epoxy self-healing polymers. 17 Since 2011, studies on endogenous self-healing mechanisms for various dynamic covalent and non-covalent bonds have increased intensively and extensively. It has been well-known that polymers with self-healing mechanisms are broadly classified into two types: extrinsic and intrinsic. 18 Depending on the specific mechanism, self-healing polymers can be differentiated into several different types. However, most self-healing polymers involve two important steps in the wound-healing process. 19 As illustrated in Fig. 1 , when polymers are damaged by external impact ( e.g. , scratches, cracks, cuts, etc. ), the wound should be sutured through physical flow or diffusion of the polymer or by the elastic shape memory effect towards the injured area. 11 In the case of intrinsic self-healing, one of these mechanisms is required, whereas in the case of extrinsic self-healing, the release of substances such as monomers that can be polymerized autonomously included in microcapsules or microvascular network structures. 13 In the next step, substantial self-healing occurs through molecular-level recombination of the damaged polymer owing to physical or chemical bonding. Intrinsic self-healing includes both physical and chemical rebonding. Physical bonds include hydrogen bonding, ionic bonding, metal–ligand complexes, etc. Chemical bonds include Diels–Alder, 20 disulfides, 21 imines, 22 hindered urea, 23 thiourethane, 24 etc. Extrinsic self-healing, on the other hand, is exemplified by polymer formation through methods such as ROMP that complete the self-healing process. 25 Recently, research on hybrid self-healing materials has been actively conducted, which can operate through various mechanisms triggered by external stimuli such as light, pH, specific chemicals, and magnetic fields, as well as simple external stimuli such as temperature. 26 Fig. 1 Schematic illustration of the self-healing repairing mechanisms and types of SHPs when the scratch is induced on the surface of a film or coating. In reference to the surface scratches on display panels and automobiles, the SHPs are critically important to recover physical abrasions quickly. 27 Additionally, since damage or large cracks that can affect the lifespan of all products always start from scratches or small cracks, the problem of formation and recovery of scratches on polymer surfaces is an important issue. Therefore, this review focuses on the formation of scratch wounds and their self-healing and deals with the most recent developments of SHPs with application in self-repairing surface scratches by describing various underlying mechanisms. Moreover, the advancements made in designing such polymers to perform under hostile environmental changes such as extreme temperature, radiation, electric field, corrosion, and acidic and alkaline conditions have been analysed. The review will give a detailed account of the factors influencing polymer scratch and recovery. Besides, scratch testing (such as sharp/blunt tip scratching or abrasive scratching), methods of investing polymer surfaces, wound depths, the addition of healing fillers, and conditions maintained during the healing process are reviewed thoroughly. Further, various progressive surface examinations (including sliding friction, micro to nano-scratch, nanoindentation, and evaluation of penetration depth, residual depth, residual scratch width, contact scratch width, and elastic recovery rates) are also included. Finally, a broader future perspective on the challenges and prospects of SHPs in healing surface scratches is presented."
} | 1,774 |
20070870 | null | s2 | 3,066 | {
"abstract": "It is widely recognized that biofuel production from lignocellulosic materials is limited by inadequate technology to efficiently and economically release fermentable sugars from the complex multi-polymeric raw materials. Therefore, endoglucanases, exoglucanase, pectate lyases, cutinase, swollenin, xylanase, acetyl xylan esterase, beta glucosidase and lipase genes from bacteria or fungi were expressed in Escherichia coli or tobacco chloroplasts. A PCR-based method was used to clone genes without introns from Trichoderma reesei genomic DNA. Homoplasmic transplastomic lines showed normal phenotype and were fertile. Based on observed expression levels, up to 49, 64 and 10, 751 million units of pectate lyases or endoglucanase can be produced annually, per acre of tobacco. Plant production cost of endoglucanase is 3100-fold, and pectate lyase is 1057 or 1480-fold lower than the same recombinant enzymes sold commercially, produced via fermentation. Chloroplast-derived enzymes had higher temperature stability and wider pH optima than enzymes expressed in E. coli. Plant crude-extracts showed higher enzyme activity than E. coli with increasing protein concentration, demonstrating their direct utility without purification. Addition of E. coli extracts to the chloroplast-derived enzymes significantly decreased their activity. Chloroplast-derived crude-extract enzyme cocktails yielded more (up to 3625%) glucose from filter paper, pine wood or citrus peel than commercial cocktails. Furthermore, pectate lyase transplastomic plants showed enhanced resistance to Erwina soft rot. This is the first report of using plant-derived enzyme cocktails for production of fermentable sugars from lignocellulosic biomass. Limitations of higher cost and lower production capacity of fermentation systems are addressed by chloroplast-derived enzyme cocktails."
} | 464 |
35423546 | PMC8695656 | pmc | 3,067 | {
"abstract": "With the help of dopamine, we constructed a hydroxyl-rich secondary reaction platform on a surface formed by interwoven nylon 56 and cotton fibres. Octadecyl mercaptan and vinyl trimethoxysilane (VTMS) are used for the click coupling preparation of superhydrophobic reagents, which are grafted onto polydopamine aggregates and successfully used to prepare superhydrophobic nylon 56/cotton-interwoven fabric. The static contact angle was 161° and the sliding angle was 8°. Note that the prepared superhydrophobic fabric can withstand corrosive liquids, water washing, ultraviolet radiation and mechanical abrasion, it has excellent superhydrophobic stability, and self-cleaning and oil–water-separation functionalities. This simple, fast and environmentally friendly method can be applied to other substrates and shows tremendous potential for expanding the field of superhydrophobic applications.",
"conclusion": "4. Conclusion We oxidised and polymerised dopamine to form a polydopamine coating, built a secondary reaction platform rich in hydroxyl on the surface of cotton and nylon 56 fibres and grafted a superhydrophobic reagent onto the polydopamine aggregate via the click coupling of octadecyl mercaptan and vinyl trimethoxysilane; this not only improved the fabric's surface roughness but also reduced its surface energy. Thus, we successfully manufactured superhydrophobic textiles. The CA and SA of the superhydrophobic textiles are 161° and 8°, respectively. These textiles can withstand acidic and alkaline environments, organic solvents, washing with water, UV irradiation and mechanical wear. Moreover, they have excellent superhydrophobic stability and functionality for self-cleaning and oil–water separation. Our method is simple, rapid, environmentally friendly and universally applicable and can be applied to endow other substrates with high hydrophobicity. Thus, our technique shows good prospects for superhydrophobic applications.",
"introduction": "1. Introduction Usually, superhydrophobic surfaces are defined as those with static water contact angles of >150° and sliding angles of <10°. 1–4 Due to the superhydrophobicity, the water droplet can easily move down a tilted surface and it does not stick to the surface and also, it bounces when it is dropped on the surface from a height. Such surfaces exhibit unique wettability characteristics and have attracted considerable attention because of their potential applications, which include corrosion protection, anti-icing, self-cleaning, 5 oil–water separation, 6 antifouling, 7 and microfluid transportation. 8 There are two key points to consider when preparing superhydrophobic surfaces, namely, the construction of the micro–nano roughness and the use of materials with low surface energies. 9–11 Scientists have developed multiple preparation methods, including nanoparticle loading, 12 layer-by-layer assembly, 13 and chemical vapour deposition. 14 Using these methods, research institutions successfully created superhydrophobic surfaces on different substrates to meet various complex and changeable requirements. 15–17 Unfortunately, certain superhydrophobic surfaces remain fragile and unstable at present and are easily affected by the external environment, resulting in a loss of their superhydrophobic properties. Moreover, there are certain problems with the manufacturing process, including its length, its high cost, and environmental pollution. Therefore, in this study, we provide a simple, fast, environmentally friendly, and pollution-free method for preparing durable superhydrophobic surfaces. Cotton fibre, the world's most abundant plant fiber, is inexpensive and provides excellent comfort, reproducibility and biodegradability. Its primary component is cellulose. Because of the rich hydroxyl on the surface of the cellulose substrate, it absorbs water, and there it becomes wet and its durability reduces. Therefore, superhydrophobic properties must be present on the cotton fibre surface. 18 Polyamide 56, commonly known as nylon 56, is amongst the most popular synthetic fibres; it has a wide range of applications because of its excellent durability, abrasion resistance, and corrosion resistance. 19 By developing a micro–nano rough structure on nylon 56 fiber to give it superhydrophobic properties, it should be possible to broaden its field of applications. Inspired by the excellent adhesion of marine mussels, a growing number of researchers are focusing on dopamine. As a surface modifier, dopamine plays an important role in multiple research fields, including biology, 20 chemistry 21 and medicine. 22 For example, Chen et al. prepared a nickel/graphene/polydopamine-composite coating using dopamine-assisted electrodeposition technology. The durability and corrosion resistance of this composite coating were significantly improved, and it had hydrophobic properties. 23 Based on the principle of self-polymerisation of dopamine and fluorescent polymer, Shi et al. prepared a yellow-green fluorescent–starch-based phosphor with high stability using an environmentally friendly method. This phosphor exhibited strong water dispersibility and high biocompatibility. 24 The method for generating polydopamine through the oxidative polymerisation of dopamine is simple and easy to implement; polydopamine can be formed by self-polymerisation only under an aerobic weak base and can act with strong adhesion to almost any matrix. The surface of polymerized polydopamine coating is rich in hydroxyl, amino and other groups. In addition to modifying membrane materials 25 and nanomaterials, 26 it can be used to biomimic the surfaces of fibres and fabrics. By improving the surface properties of these materials, they can be made to provide a secondary reaction platform for subsequent experiments and see use in preparing functional textiles. 27 Therefore, a series of superhydrophobic surfaces can be prepared with dopamine assistance. Sharpless, the winner of the 2001 Nobel Prize in Chemistry, proposed the concept of click chemistry. 28 Thiol–ene click chemistry offers many advantages, including a high product yield, easy separation and purification, regiospecificity, stereospecificity, and mild reactions. Furthermore, it has a wide range of applications; 29–31 e.g. , Bao et al. prepared photochromic cotton fabrics with good durability based upon the principle of thiol–ene click chemistry. These fabrics have good resistance to ultraviolet rays and excellent fatigue resistance. 32 Ning et al. adopted the thiol–ene click chemistry method to prepare a scratch-resistant coating with ultra-high adhesion on a styrene–butadiene rubber–elastomer matrix, thus making it superhydrophobic. 33 These results indicate that thiol–ene click chemistry can be applied to the study of surface functional polymers and provide a new concept for preparing superhydrophobic surfaces. In this work, we develop a simple, fast, environmentally friendly, and versatile method for preparing superhydrophobic nylon-56/cotton-interwoven fabrics ( Fig. 1 ). With dopamine used to build a hydroxyl-rich secondary-reaction platform on the surface of nylon 56 and cotton fibres, octadecyl merthiol and vinyl trimethoxysilane were employed for click coupling to prepare superhydrophobic reagents and to graft them onto the polydopamine aggregate to form superhydrophobic textiles. In other existing superhydrophobic literature reports, some methods use fluorine-containing hydrophobizing agents as raw materials, However, despite their considerable hydrophobic effect, they contribute to environmental pollution. Dopamine is a non-toxic, environmentally friendly, and adaptable substance. Most organic materials can be modified by dopamine on its surface, so it can form a secondary reaction platform on the surface of nylon 56/cotton interwoven fabric. Some existing methods for preparing superhydrophobic surfaces take a long time. Using click chemistry to graft low surface energy substances, based on its advantages of fast reaction rate and modularity, we can prepare superhydrophobic fabrics in a short time. The superhydrophobic fabric prepared by this method possesses good anti-ultraviolet (UV) ageing, abrasion resistance, acid and alkali resistance, acetone solvent resistance, and oil–water separation performance. Fig. 1 Formation mechanism of superhydrophobic textiles.",
"discussion": "3. Results and discussions 3.1 Brief formation mechanism of superhydrophobic nylon-56/cotton-interwoven fabric Dopamine is first oxidized to benzoquinone and then undergoes intramolecular cyclization, generates 5,6-dihydroxyindole through intramolecular rearrangement reactions and finally forms a polydopamine coating with irreversible high-strength covalent bonds on the substrate. The formed polydopamine coating is rich in hydroxyl functional groups. Vinyltrimethoxysilane is hydrolysed to form silanol, which undergoes a condensation reaction with the hydroxyl groups in polydopamine to form hydrogen bonds. Octadecyl mercaptan is then assisted by the photoinitiator click reaction with vinyl trimethoxy silane to form a micro–nanoscale rough structure on the substrate, thereby preparing the super hydrophobic textile. 3.2 Surface morphology and composition of the superhydrophobic nylon-56/cotton-interwoven fabric \n Fig. 2 shows scanning electron images of the original nylon-56/cotton-interwoven fabric, the dopamine-modified nylon-56/cotton-interwoven fabric and the superhydrophobic nylon-56/cotton-interwoven fabric. From Fig. 2a and b , we can see that the original nylon-56/cotton fibres show relatively smooth surfaces with a clean appearance and no other substances attached. Fig. 2c and d show that after the dopamine-modification treatment, the surfaces of these fibres are covered with polydopamine aggregates, thus building a secondary reaction platform rich in hydroxyl groups. Fig. 2e and f illustrate that the surface morphologies of the nylon 56 and cotton fibres have significantly changed and that a large number of granular block polymers have appeared; this is attributed to silane is hydrolysed and condenses with the hydroxyl group in polydopamine to form a hydrogen bond, after which it forms a polymer through click reactions with thiol under UV light; these nano-scale particles increase the roughness of the fibre surface, thus lending the fabric superhydrophobic properties. Fig. 2 SEM images of (a) original nylon-56 fibre; (b) original cotton fibre; (c) dopamine-modified nylon-56 fibre; (d) dopamine-modified cotton fibre; (e) superhydrophobic nylon-56 fibre; (f) superhydrophobic cotton fibre. \n Fig. 3 and 4 show the EDS energy spectra of the nylon-56 and cotton fibres before and after finishing, respectively. Fig. 3b shows that the dopamine-modified nylon-56 fibre comprises 65.555% C, 19.075% O, and 15.370% N. This shows a change in the content ratios of C, O, and N compared to Fig. 3a , indicating that polydopamine aggregates are attached to the surfaces of nylon 56 fibres. Two new elements (S and Si) appear, revealing that a click reaction occurred and that the hydrophobic reagent was successfully grafted on the nylon 56 fibre surface. Fig. 3c indicates that S accounts for 1.2% of the content on the superhydrophobic nylon 56 fibre while Si accounts for 2.350%. Fig. 3f shows a nitrogen content of 8.047% in dopamine-modified cotton fibre. Compared with Fig. 3e , the appearance of N shows that polydopamine aggregates are successfully arranged on the surface of the cotton fibre; Fig. 3f indicates that S accounts for 3.192% of the superhydrophobic cotton fibre while Si accounts for 4.197%. Fig. 4g and h shows the distribution of S and Si on the surface of the modified fabric; the elements were evenly distributed. Fig. 3 Element composition content of (a) original nylon-56 fibre, (b) dopamine-modified nylon-56 fibre and (c) superhydrophobic nylon-56 fibre, (d) original cotton fibre, (e) dopamine-modified cotton fibre, (f) superhydrophobic cotton fibre. Fig. 4 EDS spectra of (a) original nylon-56 fibre, (b) dopamine-modified nylon-56 fibre and (c) superhydrophobic nylon-56 fibre, (d) original cotton fibre, (e) dopamine-modified cotton fibre, (f) superhydrophobic cotton fibre. (g) Element mapping of superhydrophobic nylon-56 fibre, (h) Element mapping of superhydrophobic cotton fibre. \n Fig. 5 shows the infrared spectra (ATR) of the original nylon-56/cotton-interwoven fabric, the dopamine-modified nylon-56/cotton-interwoven fabric and the superhydrophobic nylon-56/cotton-interwoven fabric. The vibration peak at 1420 cm −1 is attributed to the pulling of the aromatic ring. The extensional vibrations show that dopamine successfully forms a polydopamine coating on the surface of the interwoven fabric. The vibrational peaks appearing at 2818 cm −1 and 2850 cm −1 are attributed to the symmetrical and asymmetrical vibrations of the –CH 2 group in the long alkyl chain. Si–C-bending vibration peaks appear at 1260 cm −1 . The vibration peak at 798 cm −1 is attributed to the tensile vibration of Si–O–Si. The above characteristic peaks indicate that the thiol–ene click–chemistry reaction has successfully occurred on the polydopamine coating. Fig. 5 Infrared spectra of original nylon-56/cotton-interwoven fabric, dopamine-modified nylon-56/cotton-interwoven fabric and superhydrophobic nylon-56/cotton-interwoven fabric. To validate the hydrophobisation process, an XPS test is performed. The XPS measurement spectrum presented in Fig. 6a shows that the nitrogen characteristic-peak area (at 398.4 eV) of the dopamine-modified nylon 56 fibre has changed compared with the original nylon 56 fibre. Fig. 6b shows a comparison with the original cotton fibre. The characteristic peak of nitrogen (at 398.4 eV) appears on the dopamine-modified cotton fibre, thereby confirming that dopamine successfully forms a polydopamine coating on the surfaces of nylon 56 and cotton fibres. Moreover, the area of the characteristic peak (at 284.8 eV) of carbon dramatically increases for the superhydrophobic textile, which can be attributed to the effect of the long-chain alkyl of octadecyl thiol. For superhydrophobic fibres, we can clearly observe four new peaks in Fig. 6a and b ; these occur at 150.5 eV (Si 2s), 100.1 eV (Si 2p), 229.0 eV (S 2s) and 164.0 eV (S 2p). This shows that the superhydrophobic reagent prepared by click coupling with octadecyl mercaptan and vinyl trimethoxysilane is successfully grafted on the polydopamine aggregates. Fig. 6 (a) XPS spectra of the original nylon-56 fibre, dopamine-modified nylon-56 fibre and superhydrophobic nylon-56 fibre; (b) XPS spectra of raw cotton fibre, dopamine-modified cotton fibre and superhydrophobic cotton fibre. 3.3 The effect of UV-irradiation time and reactant ratio on the superhydrophobic finishing of nylon-56/cotton-interwoven fabric \n Fig. 7a displays the changes in the CA and SA of the superhydrophobic fabric under different UV irradiation times. For an illumination time of 30 min, the CA reaches 156.38° and the SA is <10°. The change in CA increases with illumination time and finally stabilises, which may be attributed to the completion of click reaction. The results show that the best illumination time is 30 min. Fig. 7 (a) Changes in CA and SA with reaction time. (b) Changes in CA and SA with the ratio of vinyltrimethoxysilane to octadecyl mercaptan. Vinyl trimethoxy silane and octadecyl mercaptan are used to generate polymers through the thiol–ene click reaction in which mercaptan and vinyl functional groups are almost equally consumed. Here, vinyl trimethoxy silane is hydrophobic and octadecyl mercaptan contains hydrophobic groups; hence, the ratio of vinyltrimethoxysilane to octadecyl mercaptan is crucial for determining surface wettability. Fig. 7b shows that CA and SA change with the molar-mass ratio; when the molar-mass ratio of vinyltrimethoxysilane and octadecyl mercaptan is 1 : 2, CA reaches 161°, SA reaches 8° and the nylon-56/cotton-interwoven fabric becomes superhydrophobic. The degree of hysteresis of the contact angle represents how easy it is for the liquid to detach from the solid surface. The greater the difference between the advancing contact angle and the receding contact angle, the less likely it is for the liquid to move on the surface of the fabric. As shown in Fig. 8 , the contact angle hysteresis of the samples is almost all less than 10°, indicating that the droplets are easy to fall off the surface of the superhydrophobic fabric. Fig. 8 The data of the sample's advancing contact angle, receding contact angle and contact angle hysteresis. 3.4 Mechanical stability of superhydrophobic nylon-56/cotton-interwoven fabric The mechanical stability of superhydrophobic fabrics should be considered for real-life applications because this largely determines their durability. Thus, we conducted a simple sandpaper-abrasion test. As shown in Fig. 9a , the fabric is placed on 1000-CW sandpaper, pressed with a weight of 100 g, and pulled 15 cm along the ruler. The CA of water was recorded after every five wear cycles to show the change in the fabric surface's wettability ( Fig. 9b ). The CA of the fabric remained >150° over 25 abrasion cycles; this may be attributed to the excellent adhesion of polydopamine. The nano aggregates on the fibre did not significantly fall off, resulting in stable and excellent hydrophobic properties. The mechanical stability of superhydrophobic fabrics also includes tensile stability, which is mainly characterized by testing mechanical strength. The test results are shown in the Fig. 10 . Compared with the original fabric, the breaking strength and breaking elongation of the superhydrophobic fabric in the warp direction have not changed significantly, while the breaking strength and breaking elongation in the weft direction have been improved. The experimental finishing method not only caused no damage to the fabric, but also improved its mechanical properties. In summary, it shows that the superhydrophobic fabric has good mechanical stability, thereby improving the potential for superhydrophobic textile applications. Fig. 9 (a) Picture of sandpaper-abrasion test. (b) CA and SA changes of the superhydrophobic nylon-56/cotton-interwoven fabric after 25 abrasion cycles. Fig. 10 Tensile test of original fabric and superhydrophobic fabric. 3.5 Chemical stability and durability of superhydrophobic nylon-56/cotton-interwoven fabric In the course of daily life, fabrics will be exposed to harsh and complex environments such as strong acids or bases, organic solvents, washing and ultraviolet radiation; therefore, it is necessary to improve the chemical stability and durability of superhydrophobic fabrics. To test the effects of ultraviolet radiation, our fabric sample is exposed to UV light. The fabric is situated 15 cm away from the light source. Each light cycle lasts 4 h, after which the CA and SA of the fabric are tested ( Fig. 11a ). From the figure, it can be seen that, after four photoperiods, the CA of the fabric gradually decreases with UV irradiation time, although it remains above 150°; the SA, moreover, increases slightly. Good hydrophobic properties are therefore observed, indicating that superhydrophobic fabric exhibits UV resistance. To test the effect of organic solvents on such fabrics, fabric samples are immersed in acetone every 4 h for a certain time period and then rinsed with deionised water and dried at 70 °C. As shown in Fig. 11b , after the fabric is soaked in acetone for 24 h, the CA remains above 155°, indicating that the fabric has good durability against acetone; this may be attributed to the chemical polymer grafted onto the polydopamine coating has good resistance to corrosion by organic solvents. To test the influence of the acid–base environment on the superhydrophobic fabric, the fabric samples are immersed in solutions of different pH (pH = 1, 3, 5, 7, 9, 11 and 13) for 48 h. Fig. 11c shows the CA; although there is a slight change, it still remains at >150°. The results show that the acid–base environment has a certain effect upon the superhydrophobic properties of the fabric, but that the fabric still maintains a high hydrophobicity. Therefore, superhydrophobic fabric has a certain resistance to acidic and alkaline environments. This may be because the air trapped upon the fabric's surface can inhibit acid or alkali corrosion. To test the effect of washing on superhydrophobic fabrics, washing durability is tested with reference to the standard washing-machine procedure outlined in the AATCC Test Method 61-2003 Test No. 1A. The washing times are set to 0, 45, 90, 135, 180, 225, and 270 min. As shown in Fig. 11d , the CA and SA of the fabric change with the washing time; however, the former remains above 155° while the latter remains below 10°. This may be attributed to the super adsorption of polydopamine, which makes it difficult for low-surface-energy substances to fall off of the surface of the substrate. Therefore, superhydrophobic fabric exhibits excellent UV resistance, solvent resistance, acid and alkali resistance, and water-washing resistance, all of which is conducive to industrial production. Fig. 11 The changes in the CA and SA of superhydrophobic nylon-56/cotton-interwoven fabric under the following conditions: (a) UV irradiation over various times; (b) soaking in acetone over various times; (c) soaking in various-pH solutions over 48 hours; (d) washing over various times. 3.6 Self-cleaning performance of the superhydrophobic nylon-56/cotton-interwoven fabric In this work, a reactive blue dye was used as a contaminant for self-cleaning tests; as shown in Fig. 12a and b , when water droplets fall, we can clearly see the difference between the original fabric and the superhydrophobic fabric. The surface of the original fabric is completely wetted and contaminated with dye; however, on the superhydrophobic fabric, this dye is removed by water droplets to ensure the surface is clean. To demonstrate the water-repellent characteristics of the superhydrophobic fabric, it, in addition to the original fabric, was completely immersed in water. The original fabric was reported to sink in water after being released while the superhydrophobic fabric floated ( Fig. 13a ). In practical applications, fabrics will come into contact with liquids commonly used in daily life; here, we use salt water, coffee, milk, dyed water, cola and tea. As shown in Fig. 13b and c , the original fabric was wetted and contaminated by these liquids; however, the surface of the superhydrophobic fabric showed spherical droplets and no contamination occurred. Fig. 14a and b show that when the droplets contact the original fabric, the fabric is completely penetrated. However, the droplets on the superhydrophobic fabric are in a spherical form. The results indicate that the permeability of the fabric is significantly reduced. This may be the hydrophobic agent grafted on the fiber surface achieves super hydrophobic effect. These results indicate that the superhydrophobic fabric has excellent self-cleaning and antifouling properties that are suitable for use in daily life. Fig. 12 Self-cleaning test of nylon-56/cotton-interwoven fabric: (a) original fabric, (b) superhydrophobic fabric. Fig. 13 (a) Water soaking of the original nylon-56/cotton-interwoven fabric and the superhydrophobic nylon-56/cotton-interwoven fabric; (b) the state of different droplets on the original nylon-56/cotton-interwoven fabric; (c) the state of different droplets on the superhydrophobic nylon-56/cotton-interwoven fabric. Fig. 14 (a) Penetration effect of original fabric, (b) penetration effect of superhydrophobic fabric. 3.7 Oil–water-separation performance of superhydrophobic nylon-56/cotton-interwoven fabric As industrialisation progresses, a large amount of oily wastewater will be generated during manufacturing processes. A superhydrophobic fabric can be used for oil–water separation, which is beneficial for environmental protection. Fig. 15a and b display methylene chloride and carbon tetrachloride stained with Oil Red O. The results illustrate that the superhydrophobic fabric can adsorb dichloromethane and carbon tetrachloride. Fig. 15c presents a schematic of an oil–water-separation test using a superhydrophobic fabric. Dichloromethane is dyed red with Oil Red O and water is dyed blue with methyl blue. The oil/water mixture (with both oil and water having volumes of 100 mL) is poured into an oil–water separation device. The oil quickly penetrates the fabric while the water is left on top; it can be seen that the oil–water separation performance of the superhydrophobic fabric is relatively excellent. Regardless of the density of the oil, the superhydrophobic fabric can selectively absorb oil. And based on the excellent durability, abrasion resistance and corrosion resistance of nylon 56, the super-hydrophobic fabric can be used in harsh environments, broadening the application in the field of oil–water separation. Fig. 15 (a and b) The selective adsorption of superhydrophobic nylon-56/cotton-interwoven fabric to methylene chloride (dyed with Oil Red O) and carbon tetrachloride in water (dyed with Oil Red O); (c) oil–water-separation test of superhydrophobic nylon-56/cotton-interwoven fabric."
} | 6,342 |
35665560 | PMC9543302 | pmc | 3,068 | {
"abstract": "Abstract Mammals rely on the metabolic functions of their gut microbiota to meet their energetic needs and digest potentially toxic components in their diet. The gut microbiome plastically responds to shifts in host diet and may buffer variation in energy and nutrient availability. However, it is unclear how seasonal differences in the gut microbiome influence microbial metabolism and nutrients available to hosts. In this study, we examine seasonal variation in the gut metabolome of black howler monkeys ( Alouatta pigra ) to determine whether those variations are associated with differences in gut microbiome composition and nutrient intake, and if plasticity in the gut microbiome buffers shortfalls in energy or nutrient intake. We integrated data on the metabolome of 81 faecal samples from 16 individuals collected across three distinct seasons with gut microbiome, nutrient intake and plant metabolite consumption data from the same period. Faecal metabolite profiles differed significantly between seasons and were strongly associated with changes in plant metabolite consumption. However, microbial community composition and faecal metabolite composition were not strongly associated. Additionally, the connectivity and stability of faecal metabolome networks varied seasonally, with network connectivity being highest during the dry, fruit‐dominated season when black howler monkey diets were calorically and nutritionally constrained. Network stability was highest during the dry, leaf‐dominated season when most nutrients were being consumed at intermediate rates. Our results suggest that the gut microbiome buffers seasonal variation in dietary intake, and that the buffering effect is most limited when host diet becomes calorically or nutritionally restricted.",
"introduction": "1 INTRODUCTION Mammals that consume diets rich in cellulose and other plant fibres depend on gut microbes to extract short chain fatty acids (SCFAs) and additional energy sources from food items (Flint et al., 2008 ; Flint & Bayer, 2008 ; Lambert, 1998 ; Mackie, 2002 ). In mammals relying heavily on plant foods, the gut microbiome plays an essential role in breaking down phenolics and other plant toxins (Dearing & Kohl, 2017 ; Greene et al., 2020 ; Kohl et al., 2014 , 2016 ). Additionally, dietary shifts, environmental changes in food availability, rainfall and temperature, and habitat differences, such as forest structure and anthropogenic changes, appear to more strongly influence the gut microbiome composition of animals that ingest a high proportion of indigestible plant fibre than animals that ingest a diet composed principally of fruit, flowers and/or invertebrates (Frankel et al., 2019 ; Greene et al., 2019 ). Moreover, there is evidence of functional and compositional convergence in the gut microbiome of animals with high‐fibre diets, both within and across taxonomic groups (Amato et al., 2019 ; Hale et al., 2018 ; Ley et al., 2008 ; Muegge et al., 2011 ). Gut microbial composition can rapidly change in response to dietary variation (e.g., high fat to low fat consumption) (David et al., 2014 ; Turnbaugh et al., 2009 ). Several studies have identified seasonal shifts in gut microbial composition, including in wild mice ( Apodemus sylvaticus ) (Maurice et al., 2015 ), wild Bale monkeys ( Chlorocebus djamdjamensis ) (Trosvik, Rueness, et al., 2018 ), captive and wild giant pandas ( Ailuropoda melanoleuca ) (Wu et al., 2017 ; Xue et al., 2015 ), wild flying squirrels ( Pteromys volans orii ) (Liu et al., 2019 ), wild sage grouse ( Centrocercus urophasianus ) (Drovetski et al., 2019 ), wild Verreaux's sifakas ( Propithecus verreauxi ) (Springer et al., 2017 ), wild tench ( Tinca tinca ) (Dulski et al., 2020 ) and wild geladas ( Theropithecus gelada ) (Baniel et al., 2021 ; Trosvik, Muinck, et al., 2018 ). Seasonal shifts in the gut microbiome of these species are linked to changes in the specific foods consumed, food availability or macronutrient composition of the diet (Kartzinel et al., 2019 ; Orkin et al., 2019 ; Ren et al., 2017 ). For example, in African great apes ( Gorilla gorilla gorilla and Pan troglodytes troglodytes ), the gut microbiome contains more fibre‐degrading taxa and higher cellulose degradation functional potential when individuals are consuming a leaf‐heavy diet (Hicks et al., 2018 ). In Chinese alligators ( Alligator sinensis ) and hibernating ground squirrels ( Ictidomys tridecemlineatus ), mucin‐degrading gut microbial taxa increase in abundance during periods of fasting, when less energy is available from dietary sources (Carey et al., 2012 ; Tang et al., 2019 ). Gut microbial community stability and resilience improve host health by reducing long periods of gut microbial dysbiosis (Allaway et al., 2020 ; Sommer et al., 2017 ). However, the ability of the microbiome to plastically respond to short‐ and long‐term changes in environmental conditions could also be important for host health and fitness. This plastic response in gut microbiome composition and function influences host phenotype, resulting in variable host physiology and behaviour in different environmental circumstances (Davidson et al., 2018 ; Moeller & Sanders, 2020 ; Stearns, 1989 ; West‐Eberhard, 2003 ). Evidence is beginning to emerge that diet‐related gut microbial changes may buffer energy and nutrient availability. In humans, we see marked, rapid shifts in gut microbiome composition and SCFA production immediately after an ultramarathon (Grosicki et al., 2019 ) or in response to increases of other forms of exercise (Estaki et al., 2016 ; Keohane et al., 2019 ), that are consistent with an increase in gut microbial efficiency beneficial for host health. Gut microbial community composition shifts also have been observed in pregnant and lactating monkeys, resulting in a more metabolically efficient gut microbiome (Mallott et al., 2020 ; Mallott & Amato, 2018 ). Similarly, seasonal shifts in gut microbiome composition and SCFA production in black howler monkeys ( Alouatta pigra ) appear to compensate for decreases in energy intake, buffering against nutrient and energy shortfalls (Amato et al., 2015 ). The extent to which changes in microbiome composition result in shifts in potential host‐relevant functions and microbial metabolism (beyond changes in SCFA production) remains unclear. The microbial functions actively being expressed vary more between individuals than either gene family or metabolic pathway presence in the microbiome, and variability in microbial gene expression appears to play a larger role in influencing host phenotypic plasticity than do changes in the taxonomic composition or functional potential of the gut microbiome (Barroso‐Batista et al., 2020 ; Heintz‐Buschart & Wilmes, 2018 ; Tanca et al., 2017 ). Thus, studies of gut microbiome composition and function using marker gene or metagenomic sequencing may miss important variation in actively expressed microbial functions. Metabolomics, identifying the small molecules produced during metabolism, allows us to examine how microbial metabolism responds to changes in nutrient and energy intake in the host (Bäckhed & Crawford, 2010 ; Ursell et al., 2014 ). Several studies have shown that microbially associated metabolites are strongly linked to host health and metabolism and respond to dietary change in much the same way as microbiome composition (Filippis et al., 2016 ; Maier et al., 2017 ; Mchardy et al., 2013 ; Sharon et al., 2014 ). Metabolomics offers greater insight into the fine‐scale plasticity in microbial metabolism that acts as a buffering agent against nutrient and energy shortfalls. Investigating how the metabolome responds to changing diet and/or energetic needs in wild systems helps us understand the role of the gut microbiome in buffering energetic or nutrient shortfalls in animals; however, few studies have used metabolomics to examine gut microbial functional shifts in wild animals (Garber et al., 2019 ; Gomez et al., 2015 , 2016 ). To determine how host–microbe cometabolic processes dynamically respond to seasonal changes in diet and nutrient intake, we examined the metabolome of a black howler monkey population experiencing seasonal changes in feeding behaviour, nutrient intake, gut microbiome composition and SCFA production, as reported in our previous research (Amato et al., 2014 , 2015 , 2017 ). This population experiences three distinct seasons—Wet, Fruit‐Dominated (WFD), Dry, Leaf‐Dominated (DLD), and Dry, Fruit‐Dominated (DFD)—each characterized by distinct energy and nutrient intake profiles (Table 1 ). The intake of protein, energy, lipids, neutral detergent fibre and nonstructural carbohydrates was highest in the WFD season and lower in the DLD and DFD seasons. Therefore, we found this population optimal for addressing three research questions: (i) Does the metabolome of black howler monkeys vary seasonally? (ii) If so, are seasonal changes in the metabolome associated with changes in gut microbial community composition and linked to changes in nutrient and energy intake? (iii) Do seasonal changes in the metabolome suggest that the gut microbiome is buffering shortfalls in energy or nutrient intake? If buffering is occurring, then when the consumption of a specific macronutrient declines, we expect to see an increase in metabolite concentrations related to those specific macronutrient metabolic pathways. TABLE 1 Energy and nutrient intake profiles for each of the three seasons experienced by the study population of black howler monkeys. Energy consumption is expressed as energy per metabolic body weight (MBW) and nutrient intake values are expressed as grams per metabolic body weight Season Energy (kcal/MBW) Protein (g/MBW) Lipid (g/MBW) Nonstructural carbohydrate (g/MBW) Neutral detergent fibre (g/MBW) Wet, Fruit‐Dominated (WFD) High (177.4–182.9) Intermediate to high (9.4–11.9) High (3.2–4.2) Intermediate to high (29.0–29.7) High (42.3–50.3) Dry, Leaf‐Dominated (DLD) Intermediate (105.5–172.4) Intermediate to high (7.7–10.0) Low to intermediate (1.2–3.2) Low (12.4–16.2) Intermediate (27.2–39.9) Dry, Fruit‐Dominated (DFD) Low (106.6–114.5) Low (4.7–5.3) Low to intermediate (1.8–2.1) Intermediate to high (16.3–17.0) Low (21.9–24.7)",
"discussion": "4 DISCUSSION In this study, we examined how seasonal variation in macronutrient intake influences the gut metabolome in wild black howler monkeys. We found that the gut metabolome varied seasonally in our study population, similar to previous studies (Amato et al., 2015 ). We found partial support for the hypothesis that changes in the gut microbiome and corresponding changes in the metabolome buffer seasonal energy and/or nutrient shortfalls. Additionally, we found strong associations between faecal metabolites and ingested plant metabolites. However, we did not find significant associations between the metabolome and microbiome composition. The faecal metabolome of black howler monkeys is dominated by lipids, amino acids, carbohydrates and organic acids, similar to other species of nonhuman primates (Garber et al., 2019 ; Gomez et al., 2015 , 2016 ; Ni et al., 2021 ). Some of the long‐chain fatty acids prominent in the black howler monkey metabolome—palmitic acid, stearic acid and vaccenic acid—are associated with high‐fat diets in mice (Daniel et al., 2014 ) and fruit‐dominated diets in lowland gorillas (Gomez et al., 2016 ). Palmitic acid and stearic acid are the most common saturated fatty acids in nature while vaccenic acid is one of the most common unsaturated fatty acids (Sommerfeld, 1983 ). For primates that principally consume plant foods, such as black howler monkeys, the majority of lipid intake comes from fruits and their seeds (Norconk et al., 2009 ). Given that this population of black howlers has not been observed to consume large amounts of seeds and commonly voids undigested seeds in their faeces, the fatty acid profiles of their faecal metabolomes are likely to be strongly influenced by a yearly diet composed of 57.3% fruits (per cent dry weight) (Amato & Garber, 2014 ). We found significant seasonal differences in faecal metabolome profiles, both in composition and in pathway enrichment. Several metabolic pathways were differentially enriched between seasons, with lower concentrations of metabolites present during the WFD season when compared to both the DLD and DFD seasons. Additionally, metabolite network structure varied across seasons. The metabolite network was highly connected but less stable in the DFD season, when black howler monkey diets are calorically and nutritionally more constrained. In contrast, during the WFD season when black howler monkeys are consuming more energy‐ and nutrient‐rich diets, the metabolite network was diffuse. The most stable metabolite network occurred during the DLD, when most nutrients were being consumed at intermediate rates. In addition, edge betweenness in the microbial–metabolite interaction networks was highest in the DLD season, indicating that consistent, strong connections between individual bacteria and metabolic products may be contributing to metabolite network stability in this season. These patterns suggest that metabolic cross‐feeding may be more necessary for the gut microbial community when nutrients from the host diet are less readily available in the gut. Because cross‐feeding can increase the metabolic efficiency and/or ecological stability of the microbial community (Coyte et al., 2015 ; Coyte & Rakoff‐Nahoum, 2019 ; D'Souza et al., 2018 ; Evans et al., 2020 ; Goldford et al., 2018 ; Gudelj et al., 2016 ; Liu & Sumpter, 2017 ; Smith et al., 2019 ), our data provide preliminary evidence of improved nutritional buffering by the gut microbiome during the DLD. During the DFD season, when metabolite network connectivity was high but stability was low, metabolic cross‐feeding may be taking place, but the cross‐feeding relationships are not as consistant or stable over time. Microbial responses to variation in the intake of specific nutrients over time also provide evidence of nutritional buffering by the gut microbiome. For example, given that leaves tend to be high in fibre and low in host‐metabolizable energy (Norconk et al., 2009 ), we expected to see enrichment of pathways related to structural carbohydrate and lipid metabolism during the DLD, as microbes degrade fibre to produce SCFAs. While the pathway enrichment analysis did not show an increase in carbohydrate and lipid metabolism, we found that metabolites related to lipid metabolism and SCFA metabolism became more important in metabolite networks during the DLD season. In addition, the metabolism of essential vitamins (B3) and amino acids were enriched in the DLD season, potentially compensating for a diet poor in specific nutrients and aiding in the digestion of protein‐rich leaves. Similarly, although fruits have more host‐metabolizable energy than leaves, they are lower in protein compared to leaves, and overall caloric intake during the DFD season was reduced by 37.4%–40.0% compared with the WFD season and 0%–33.6% compared with the DFD season (Table 1 ). Therefore, we expected microbial buffering to result in an enrichment of amino acid synthesis pathways and lipid metabolism pathways. We did see an enrichment of amino acids pathways during the DFD and lipid metabolites having a more central role in the metabolite network (Figure 3 ). In contrast, we observed a de‐enrichment of many pathways during the WFD season when the black howler monkey diet was least nutritionally constrained. These results confirm earlier work in this population (Amato et al., 2014 , 2015 ), as well as other studies in mammals (Gomez et al., 2015 ; Koren et al., 2012 ; Mallott & Amato, 2018 ; Springer et al., 2017 ; Sun et al., 2016 ; Wu et al., 2017 ), that indicate the gut microbiome acts as a potential buffer limiting energy and nutrient shortfalls due to seasonal changes in diet or changes in host nutrient requirements. The gut microbiome also provides nutritional benefits to hosts by processing plant secondary metabolites that otherwise act as toxins or digestive inhibitors. This relationship has been documented in desert woodrats (Dearing & Kohl, 2017 ; Kohl et al., 2014 , 2016 ) and may be important to highly folivorous primates such as black howler monkeys, whose yearly diet contains 33.1% young and mature leaves (% dry weight) (Amato & Garber, 2014 ). Interestingly, our plant‐faecal metabolite interaction network from all seasons combined showed faecal target metabolites associated with essential nutrients, fatty acid metabolism, and the metabolism of chemical defensive compounds clustered with two groups of plant source metabolites: simple carbohydrates and fatty acids. Specifically, the plant secondary metabolites included dehydroabietic acid in the upper largest cluster, a diterpenoid for chemical defence commonly found in tree resin (Helfenstein et al., 2017 ), and epicatechin, a flavan‐3‐ol and major component of condensed tannin (Ferreira et al., 1999 ; Khanbabaee & Ree, 2001 ), in the right smaller cluster. This suggests that individuals ingested toxic plant secondary metabolites while consuming sugar and fatty acid metabolites used for fatty acid/glucose‐related biosyntheses. This relationship could indicate a potential trade‐off in that foods they consume with the highest protein and caloric value also contain high amounts of tannins and other difficult to digest plant secondary metabolites. While some mammals might avoid food items high in particular plant secondary metabolites, evidence suggests that other mammals readily consume foods high in secondary metabolites if those foods are also high in energy, protein or water (Felton et al., 2009 ; Lambert & Rothman, 2015 ; Remis et al., 2001 ; Simpson & Raubenheimer, 2001 ; Villalba & Provenza, 2005 ). The presence of tannin‐ or toxin‐degrading bacteria in the gut microbiome could facilitate this behaviour by allowing animals to tolerate higher concentrations of plant secondary metabolites in their diet. We also identified an aromatic thiol (organosulphur compound), positioned in the larger cluster of sugar/fatty acid metabolites along with dehydroabietic acid, that is either derived from plant leaves (Gonulalan et al., 2019 ), from soil (Shen et al., 2020 ), or is an intermediate byproduct of another aromatic hydrocarbon (such as butylbenzene) that was degraded by sulphur‐reducing gut bacteria (such as Desulfovibrio , which degrades hydrocarbons with a sulphate redox reaction) (Lyles et al., 2014 ; Widdel et al., 2006 , 2010 ) after uptake in the diet. Either way, the presence of this aromatic thiol might be another indicator of how the black howler monkey gut microbiome facilitates a higher tolerance for increased amounts of plant secondary metabolites during dietary shifts. Mechanistic experiments testing the capacity of the howler monkey gut microbiome to degrade these potential plant toxins will be necessary to verify this relationship. Although our data suggest that the gut microbiome buffers wild howler monkey hosts against nutritional challenges, they also indicate that these microbial services are likely to have limits. Our metabolite networks indicated increased cross‐feeding and community stability in response to nutritional constraints during the DLD compared to the WFD. However, during the DFD season, as black howler monkey diets became even more calorically and nutritionally constrained, the stability of the metabolite networks began to decline. This trend suggests there may be a threshold past which specific dietary changes alter the underlying structure of the microbiome in a way that compromises the nutritional services the microbiome provides to the host. For example, reductions in the intake of specific macronutrients consumed in large amounts such as fibre, or reductions that persist for extended time periods could lead to the loss of key microbial taxa (Sonnenburg et al., 2016 ). Studies of black howler monkeys in anthropogenically altered environments report reduced microbial diversity as well as reduced relative abundances of SCFA‐producing microbial taxa (Amato et al., 2013 ). These losses in microbial taxa are correlated with losses of plant species, and presumably particular macronutrients, from the howler diet, and are likely to inhibit nutritional buffering by the microbiome. Identifying the dietary thresholds past which the gut microbiome loses its ability to buffer hosts from nutritional shortfalls in black howler monkeys as well as other wild animals will provide important insight into host ecology and conservation as well as microbial community dynamics. Data describing howler monkey physiology are critical for verifying the magnitude of impact of potentially beneficial microbiome functions on hosts in variable environments. While some changes in microbial metabolism could benefit the host by increasing the availability and subsequent absorption of nutrients or key vitamins lacking in the diet, they might also be harmful to the host if scarce nutrients are diverted into networks of gut microbial metabolism. Alternatively, the impacts of changes in microbial metabolism on the host might be negligible. These potential outcomes probably occur along a gradient and are context‐dependent. Evaluating host nutrient and energy balances using noninvasive markers as well as performing controlled laboratory experiments to assess microbial metabolism and interactions with host dietary substrates in real time will provide crucial insight. Regardless of the extent to which the observed shifts in microbial metabolism buffer host nutrition, our data suggest they are strongly driven by host diet. Although previous research has shown that nutrient limitations in the large intestine resulting from dietary changes influence microbial community composition (Reese et al., 2018 ), wild animals systems such as this one make it difficult to test the extent to which host‐driven changes in the intestinal environment result in preferential recruitment of specific microbial taxa or genes. However, it is widely accepted that changes in host diet alter the nutritional environment in the gut such that microbes will differentially regulate their metabolic pathways (Fontaine & Kohl, 2020 ). This alters both competitive and mutualistic interactions between microbes, as suggested by the shifting networks of microbial taxa and metabolites that we observed across seasons. Furthermore, one of the strongest relationships we detected was between the faecal metabolome and the metabolite content of the plants consumed by black howler monkeys. For example, during the DLD season, we saw an increase in the centrality and importance of tannins and other plant secondary metabolites in the faecal metabolite networks that is probably related to the higher concentrations of these compounds in leaves compared with fruit. During the DLD, the consumption of mature leaves increased to 64% (per cent dry weight) compared with 37% during the WFD season and 43% during the DFD season (Amato et al., 2015 ; Amato & Garber, 2014 ). While mature leaves probably contribute more tannins to howler monkey diets compared with young leaves, rates of mature leaf consumption were found to be relatively low and constant in this population (1%–8% dry weight) (Amato & Garber, 2014 ). We also found a high number of significant direct associations between plant and faecal metabolites. Previous studies demonstrated similarly strong relationships between primate dietary intake and gut microbiome composition and function using DNA‐based approaches (Amato et al., 2015 ; Mallott et al., 2018 ; Orkin et al., 2019 ). Fewer have identified these relationships with faecal metabolites (Garber et al., 2019 ; Gomez et al., 2015 , 2016 ). In conclusion, we found strong relationships both between seasonal changes in diet and gut microbiome function and between the consumption of specific plant metabolites and faecal metabolite profiles. These patterns suggest that the gut microbiome might be buffering howler monkeys against seasonal variations in nutrient intake. However, we also identified evidence of a potential threshold in dietary intake past which the ability of the gut microbiome to buffer howler monkeys could be diminished. Moving forward, combining detailed studies of nutrient consumption and data on gut microbial community composition and function with biomarkers of host energy status and physiology will help to clarify the magnitude of these potential benefits and limitations and more precisely identify relevant mechanisms of interaction between both hosts and microbes as well as among different microbes."
} | 6,222 |
24285981 | null | s2 | 3,069 | {
"abstract": "The mussel byssal cuticle employs DOPA-Fe"
} | 10 |
26949525 | PMC4758376 | pmc | 3,073 | {
"abstract": "This review is a short synopsis of some of the latest breakthroughs in the areas of lignocellulosic conversion to fuels and utilization of oils for biodiesel. Although four lignocellulosic ethanol factories have opened in the USA and hundreds of biodiesel installations are active worldwide, technological improvements are being discovered that will rapidly evolve the biofuels industry into a new paradigm. These discoveries involve the feedstocks as well as the technologies to process them.",
"conclusion": "Conclusions Plants as sources of biofuels have many advantages, particularly a neutral carbon balance. Although much rhetoric has surfaced to discourage the growth of biofuel crops because they utilize farmland that should be dedicated to food crops, in reality the productivity and yield of fuels from dedicated energy crops appears to be on a steep, upward trajectory. Moreover, the technology to produce ethanol and biodiesel from plant biomass has progressed at a phenomenal rate, generating confidence that the industry will be profitable and sustainable. Because biomass varies widely in chemical composition and structure, and processes applicable to each biomass type vary as well, smaller biorefineries may be the standard of this industry rather than requiring large biorefineries to reach economic feasibility\n 31 ,\n 39 . This conclusion is based on the cited studies that look at specific processes applied to specific biomass types, such as woody biomass and GVL, which would enable a biorefinery to produce fuels efficiently from a specific biomass type. Although not discussed here, a contributing factor to the profitability of the industry will be the manufacture of co-products from the biomass: carbon fibers, fillers, resins, or polymer blends from lignin\n 22 ,\n 40 ,\n 41 or pre-extracted plastics\n 42 or enzymes\n 43 prior to deconstructing the biomass.",
"introduction": "Introduction An important mitigation strategy for the impact of fossil fuels on the environment is to use biofuels from renewable sources for transportation. Biofuels from plants represent the most abundant source of renewable fuels, offering the manufacture of ethanol and butanol (as gasoline additives) and long-chain hydrocarbons (for diesel additives or as jet fuels) from starch, cellulose, hemicellulose, and oils. The source of the energy captured by plants is the sun, which will be a constant source of energy for the next few billion years. The carbon released from the burning of biofuels is continually cycled rather than being released from ancient fixed carbon sources, as is the case for fossil petroleum and natural gas. The problem is that the cost of production of fuels from lignocellulose and plant oils is high and this nascent industry cannot compete with oil prices. \n Current progress : For the past two decades, ethanol has been produced primarily from cornstarch and cane sugar. Fourteen billion gallons of ethanol were produced in the USA from cornstarch in 2014 (\n Figure 1 ). Also shown in\n Figure 1 is that corn-based ethanol production has plateaued (\n http://www.ethanolrfa.org/wp-content/uploads/2015/09/23d732bf7dea55d299_3wm6b6wwl.pdf ). Approximately 40% of the current USA corn crop is used to produce ethanol and is not likely to expand anymore, because the remainder of the crop is being used for animal feed and human food. Ethanol is produced from cane sugar in Brazil at a level of 7.3 billion gallons in 2014 (\n http://sugarcane.org/sugarcane-products/ethanol ). Together, Brazil and the USA produce more than 90% of the world’s supply of ethanol. Figure 1. Ethanol production volumes from cornstarch over the last 9 years in the USA. Production has increased from approximately 4 billion gallons (15.1 billion liters) in 2005 to over 14 billion gallons (53 billion liters) in 2014. Biodiesel is a renewable fuel that has received considerable attention recently because it is also non-polluting. It is carbon neutral because the carbon present in vehicle exhaust was recently fixed from atmospheric carbon\n 1 . Biodiesel can be manufactured from numerous oils and fats including virgin vegetable oils, such as canola, soybean, and camelina, from waste cooking oils, or from animal fats, such as tallow or lard. The global biodiesel industry has grown considerably over the last several years\n 2 , although since 2008 a dip occurred based on world economic recession. Europe has produced the greatest volume of biodiesel over the years, followed by the USA. Worldwide production in 2012 comprised 6 billion gallons (22.5 billion liters) (\n http://www.uabio.org/img/files/docs/140526-wba-gbs-2014.pdf ). Production in the USA in 2012 was approximately 0.89 billion gallons but increased to over 1.25 billion gallons in 2014 (Energy Information Administration as shown in\n Figure 2 ). Figure 2. Biodiesel production volumes from all source oils over the last 9 years in the USA. Production has increased from 112 million gallons in 2005 to over 1.3 billion gallons in 2014. A large dip in production was seen from 2008 to 2010 during the economic recession. \n The Renewable Fuel Standard II (RFS II) : RFS II is the motivation for increasing the production of renewable fuels from green plants (\n http://www.ethanolrfa.org/policy/regulations/renewable-fuel-standard/ ). This standard was set in 2005 and revised in 2007 to mandate quantities of renewable fuels to be incorporated into the transportation industry in the USA. The goal for 2022 is set at 36 billion gallons of renewable fuels, with 16 billion gallons required to be from lignocellulosic feedstocks, and 1 billion gallons per year of biodiesel. Additionally, 58% of the fuels produced by 2022 should be “advanced biofuels”, e.g. non-starch ethanol or other types of fuels such as long-chain hydrocarbons or butanol that achieve a 50% reduction in greenhouse gas emissions. The feedstocks for these fuels are lignocellulose and oils. However, intense research is necessary to make this cost-competitive. Breakthroughs are being made in feedstock structure, ease of processing, efficiency of conversion, co-product manufacture, and sustainability. As these discoveries come together, they can be incorporated into new industrial applications."
} | 1,554 |
39663647 | PMC11635138 | pmc | 3,076 | {
"abstract": "ABSTRACT Under accelerating threats from climate‐change impacts, marine protected areas (MPAs) have been proposed as climate‐adaptation tools to enhance the resilience of marine ecosystems. Yet, debate persists as to whether and how MPAs may promote resilience to climate shocks. Here, we use 38 years of satellite‐derived kelp cover to empirically test whether a network of 58 temperate coastal MPAs in Central and Southern California enhances the resistance of kelp forest ecosystems to, and their recovery from, the unprecedented 2014–2016 marine heatwave regime that occurred in the region. We also leverage a 22‐year time series of subtidal community surveys to mechanistically understand whether trophic cascades explain emergent patterns in kelp forest resilience within MPAs. We find that fully protected MPAs significantly enhance kelp forests' resistance to and recovery from marine heatwaves in Southern California, but not in Central California. Differences in regional responses to the heatwaves are partly explained by three‐level trophic interactions comprising kelp, urchins, and predators of urchins. Urchin densities in Southern California MPAs are lower within fully protected MPAs during and after the heatwave, while the abundances of their main predators—lobster and sheephead—are higher. In Central California, a region without lobster or sheephead, there is no significant difference in urchin or kelp densities within MPAs as the current urchin predator, the sea otter, is protected statewide. Our analyses show that fully protected MPAs can be effective climate‐adaptation tools, but their ability to enhance resilience to extreme climate events depends upon region‐specific environmental and trophic interactions. As nations progress to protect 30% of the oceans by 2030, scientists and managers should consider whether protection will increase resilience to climate‐change impacts given their local ecological contexts, and what additional measures may be needed.",
"introduction": "1 Introduction Marine protected areas (MPAs) are an essential conservation tool whose coverage has globally expanded in the past decades (Duarte et al. 2020 ; Lubchenco and Grorud‐Colvert 2015 ). Their importance is reflected in recent international policies aiming to protect 30% of coastal and open oceans, as specified within Target 3 of the Kunming‐Montreal Global Biodiversity Framework (Convention of Biological Diversity 2022 ). Following mounting evidence of increasing impacts of climate change on marine ecosystems (Schoeman, Bolin, and Cooley 2023 ), the new conservation framework includes climate mitigation and adaptation in Target 8 (Convention of Biological Diversity 2022 ). The assumption underlying this framework is that protected areas may enhance climate adaptation and ecosystem resilience. While some empirical evidence supporting this expectation exists for individual MPAs and species (Jacquemont et al. 2022 ), clear empirical evidence at regional scales and for whole ecosystems is lacking. There is strong consensus that well‐managed and fully protected (i.e., no‐take) MPAs promote biodiversity and habitat conservation (Gill et al. 2017 ; Lester et al. 2009 ; Sala and Giakoumi 2018 ), but the extent to which MPAs confer ecological resilience to climate change impacts remains poorly understood. One prominent manifestation of anthropogenic climate change is the increase in the frequency and intensity of extreme climate shocks, in particular marine heatwaves (MHWs) (Oliver et al. 2018 ). MHWs have caused mass mortality of sessile or low‐mobility species (Garrabou et al. 2022 ; Szuwalski et al. 2023 ), losses of habitat‐forming species such as corals and kelp, and regime shifts, among other impacts (Arafeh‐Dalmau et al. 2019 ; McPherson et al. 2021 ; Smale et al. 2019 ; Wernberg 2021 ). For example, MHWs in Australia and in the northeast Pacific Ocean have caused extensive losses of kelp over large areas and a shift into alternative stable ecosystem states dominated by less‐productive algae or by sea urchin “barrens,” that have resulted in large‐scale economic losses (Rogers‐Bennett and Catton 2019 ; Wernberg 2021 ). Given that MHWs are becoming more frequent and longer (Oliver et al. 2018 ), it is a research priority to understand whether and how MPAs might increase resilience to their impacts. Whether MPAs provide resilience to ecosystems experiencing climate shocks is debated and challenging to study (White et al. 2025 ). The operational definition for resilience used here is resistance to and recovery from disturbance (Connell and Sousa 1983 ), although resilience is a multifaceted concept (O'Leary et al. 2017 ). MPAs are designed to provide protection from local anthropogenic disturbance, primarily from extractive activities. They cannot directly mitigate the broad‐scale impacts of climate shocks (Filbee‐Dexter et al. 2024 ; Tittensor et al. 2019 ). However, by reducing extractive activities such as fishing, MPAs may allow the recovery of key species for ecosystem functioning, which in turn can promote resilience to climate shocks (Benedetti‐Cecchi et al. 2024 ; Jacquemont et al. 2022 ; Roberts et al. 2017 ; Sala and Giakoumi 2018 ; Schindler, Armstrong, and Reed 2015 ). The empirical evidence surrounding this argument is still emerging and mixed. Some studies have found no evidence that MPAs confer resilience to climate impacts (Bruno, Côté, and Toth 2018 ; Freedman et al. 2020 ; Smith et al. 2023 ). On the other hand, other studies have shown increased resilience to climate change in MPAs: for instance, in Baja California, Mexico, juvenile recruitment and adult abundance of pink and green abalone recovered faster within MPAs following a mass mortality of benthic invertebrates due to climate‐driven hypoxia and warming (Micheli et al. 2012 ; Smith et al. 2022 ). In California, USA, species diversity recovered 75% faster from a series of MHWs within MPAs compared to adjacent unprotected areas (Ziegler et al. 2023 ). Additionally, a recent global analysis found that well‐enforced MPAs can buffer the impacts of MHWs on reef fish by promoting the stability of fish at the community and metacommunity levels (Benedetti‐Cecchi et al. 2024 ). Ultimately, a clear understanding of the conditions under which MPAs can provide climate resilience for whole ecosystems, including habitat‐forming species and their associated communities, remains limited, due to the challenge of detecting resilience within MPAs. One key challenge with detecting resilience emerges from the scarcity of long‐term, sufficiently replicated and spatially extensive data needed to characterize the state of the marine systems within and outside MPAs, before, during, and after climate extremes occur. MPAs must also be sufficiently large and be in place for a sufficient duration for any benefits of protection to emerge (Claudet et al. 2008 ). With a general paucity of studies with the necessary experimental design and statistical power, it is challenging to characterize the natural temporal variability and the inherent spatial heterogeneities of marine environments to achieve consensus on whether and under what circumstances MPAs might increase resilience to climate change impacts. Here we overcome these challenges by utilizing long‐term datasets to evaluate whether MPAs can promote kelp forest resilience to an unprecedented series of MHWs in California. During 2014–2016, the California coast was subjected to one of the largest and longest MHW regimes ever documented on Earth, with consistent mean temperature anomalies of 1°C–4°C (Cavole et al. 2016 ; Di Lorenzo and Mantua 2016 ; Frölicher and Laufkötter 2018 ), providing a unique opportunity to investigate the dynamics of MPAs and ecosystem resilience. The combination of the 2014 warm‐water anomaly and the 2015–2016 El Niño Southern Oscillation led to extremely warm waters (Cavole et al. 2016 ; Frölicher, Fischer, and Gruber 2018 ) that caused species range shifts (Favoretto, Sánchez, and Aburto‐Oropeza 2022 ; Sanford et al. 2019 ; Smith et al. 2023 ), a widespread loss of kelp forests from Northern California to Baja California Sur, Mexico (Bell et al. 2023 ), and an outbreak of sea urchins that eroded kelp forest resilience (Rogers‐Bennett and Catton 2019 ). Additionally, California has a network of MPAs that cover 16% of state waters (Saarman and Carr 2013 ), decades of satellite‐derived estimates of kelp cover (Bell et al. 2023 ), and underwater surveys of kelp forest communities (Malone et al. 2022 ). With the rich ecological monitoring data that exist in this ecosystem, we can evaluate for the first time the resilience of kelp forest ecosystems to MHWs within MPAs at a regional scale and the underlying mechanisms that facilitate this resilience. Trophic cascades are one of the proposed mechanisms by which MPAs can provide climate resilience. It has been hypothesized that, by protecting key predators of sea urchins (a voracious grazer of kelp) MPAs may indirectly control sea urchin abundance, thus increasing both kelp resistance to, and recovery from, MHWs (Ripple et al. 2016 ). Outside MPAs, where fishers target urchin predators, there are fewer predators and more urchins (Eisaguirre et al. 2020 ). When a disturbance leads to severe kelp loss, urchins can shift their behavior from hiding in protective cracks and eating drift kelp to being exposed, eating any remaining kelp and preventing kelp establishment (Harrold and Reed 1985 ; Kriegisch et al. 2019 ). Overharvesting and depletion of urchin predators can then lead to a high abundance of urchins that overgraze kelp forests (Cowen 1983 ). If MPAs protect and foster greater abundances of urchin predators (which otherwise would be commonly fished), then protected kelp forests may be more likely to recover and even resist change in the face of a disturbance, compared to unprotected kelp forests. In this study we investigated the recovery of the giant kelp ( \n Macrocystis pyrifera \n ) and bull kelp ( Nereocystis luetkeana ), henceforth “kelp,” following the 2014–2016 MHWs in Central and Southern California. The main objectives were to determine (1) whether kelp forests within a network of MPAs were more resilient to the 2014–2016 MHWs compared to unprotected kelp forests, (2) whether resilience of kelp forests differed between regions, and (3) whether there is evidence that trophic cascades are a mechanism underlying resilience to climate shocks. To address these questions, we assessed changes in kelp area during and after the 2014–2016 MHW using satellite‐derived estimates of kelp area spanning 1984–2021 and analyzed 22 years of subtidal monitoring datasets to investigate possible evidence for trophic cascades. We tested the following hypotheses: (i) kelp resistance and recovery are higher within fully protected and partially protected MPAs compared to unprotected areas in both Central and Southern California during and after the MHWs; (ii) urchin abundances are lower within MPAs compared to unprotected areas during and after the MHWs, enabling the recovery of kelp forests; and (iii) abundances of giant kelp, urchins, and urchin predators are inversely related, as predicted by the trophic cascade hypothesis.",
"discussion": "4 Discussion This study provides empirical evidence that fully protected MPAs can promote the resilience of kelp forests to climate impacts specifically when natural predators of urchins are protected, resulting in reduced grazing on kelp. Full protection improved both kelp resistance to and recovery from extreme MHWs, an effect evident from both satellite‐based canopy estimates and underwater surveys, but this effect varied by region. In Central California, where the main urchin predators were extirpated by a disease outbreak (i.e., sunflower sea stars) or are protected statewide and therefore not directly influenced by MPA status (i.e., sea otters), kelp decreased and sea urchins increased dramatically during and after the MHW, across both protected and unprotected sites. In contrast, in Southern California, MPAs had significantly greater abundances of urchin predators and fewer urchins within both partially and fully protected MPAs during and after the 2014–2016 MHWs. These results lend support to the role of trophic cascades as a mechanism for ecological resilience, and fully protected MPAs as an effective climate‐adaptation tool. Our findings provide evidence that trophic cascades may be a mechanistic path through which MPAs provide climate resilience to kelp forest ecosystems; however, these benefits are context‐dependent and vary regionally. Multiple studies have shown that fully protected MPAs increase the biomass and abundance of the predators of urchins (Caselle et al. 2015 ; Hamilton and Caselle 2015 ; Lenihan et al. 2022 ), which exert top‐down control on urchin populations, thereby supporting stability and resilience of kelp populations (Ling et al. 2009 ; Peleg, Blain, and Shears 2023 ). Here, we show that this mechanism also applies under climate impacts because we observed that there were fewer sea urchins, less loss of kelp, and greater recovery of kelp populations inside fully protected MPAs during and after the 2014–2016 MHW in Southern California. Corroborating this interpretation of our results, we found that urchin abundances were negatively correlated with those of spiny lobster and California sheephead and that giant kelp densities were positively correlated with spiny lobster abundances. One potential reason that there is not a similar positive correlation between densities of giant kelp and California sheephead is the large recruitment event of sheephead during 2015 and 2016 (Figure S16 ) while kelp suffered losses. This indirect relationship was previously documented before the MHWs in Hamilton and Caselle et al. ( 2015 ), consistent with the trophic cascade hypothesis. Overall, these results suggest that the recovery from overfishing of urchin predators within MPAs is likely controlling urchin populations and potential behavior, thus preventing overgrazing and allowing kelp to recover faster from disturbances than in unprotected areas. In Central California, we found no measurable effect of protection status on kelp resistance and recovery, likely because spatial protection does not confer additional benefits to the main mesopredators of urchins in the region—sea otters and sunflower sea stars—whose dynamics are largely independent of fishing effort and, consequently, protection status. Sea otters are federally protected and have not been actively hunted for over a century, thus benefiting from protection throughout their range. Further, sea urchin abundance started to increase exponentially both inside and outside MPAs following the mass mortality of sunflower stars due to the outbreak of sea star wasting disease in 2013–2015, which led this sea urchin predator to near extinction (Harvell et al. 2019 ; Montecino‐Latorre et al. 2016 ; Rogers‐Bennett and Catton 2019 ). We assume that the level of protection has no influence on recovery of sunflower stars, as this species is not fished and has yet to recover. These results support the notion that following major MHWs, multiple predators are needed to prevent urchin outbreaks and maintain kelp abundance through time in Central California (Selgrath et al. 2024 ). Also, these observations illuminate how non‐spatial policies, such as species‐specific interventions (i.e., the federal protection conferred over sea otters, and possibly the proposed active restoration of depleted seastar populations) may be needed to promote ecosystem resilience. Interestingly, during the MHWs we see opposing population trends of urchins within Central and Southern California. In Southern California, there was a consistent decline in urchins across all categories of protection during 2014–2016, although densities were still high with an average of 3.89 m 2 in 2016. This may be explained by the large increase in the number of California sheephead during these years or that purple sea urchins are already living closer to their upper thermal limit in Southern California (Hammond and Hofmann 2010 ), thus a temperature anomaly of a few degrees may have caused mortality in Southern California but not Central California. However, hundreds of kilometers further south in Baja California, Mexico, purple sea urchins have increased their abundance after the MHWs (Arafeh‐Dalmau et al. 2020 ), indicating that the conditions are suitable. Most importantly, after the MHWs there were fewer urchins in fully protected MPAs in Southern California, providing evidence that trophic cascades play a role in lowering urchin abundances. Our results expand on other studies in the region, emphasizing evidence for trophic cascades—preserved by MPAs—as the mechanism separating healthy kelp forests from urchin barrens. For example, trophic cascades were found to enhance macroalgal abundances in MPAs in the northern Channel Islands a year after the MHWs (Eisaguirre et al. 2020 ). However, another study in the Channel Islands found contrasting evidence: there was an increase of urchins within MPAs, in part due to the release of red urchins from fishing pressure within MPAs, which outweighed any effect of trophic cascades (Malakhoff and Miller 2021 ), though the authors of this study did not consider the response of urchins to the MHWs. In comparison, we found fewer urchins within MPAs, but only during and after the MHWs. Notably, when we took into consideration the year of establishment of the MPAs, we found that protection led to fewer urchins in Southern California through time (Figure S8 ). Therefore, by expanding the spatial and temporal scale of analysis, our results reconcile previously contrasting conclusions. Our work is also subject to some limitations. First, kelp canopy area is an estimate from satellite imagery, which may add some sources of error (Alix‐Garcia and Millimet 2022 ). However, ongoing methodological improvements have addressed most detection gaps (see Bell et al. 2020 for more detail). For the subtidal data, while we have size structure information for California sheephead that allows us to evaluate biomass, such data are not available for spiny lobsters as it is difficult to measure their size in the field. Also, accurate estimates of lobster density are difficult to obtain from this type of general long‐term monitoring of the benthos, as larger and more‐tailored surveys, which investigate all crevices, are needed to obtain accurate counts. However, the method used to estimate lobster densities in this study is consistent for all sites and years; therefore, potential accuracy limitations are less likely to influence the overall findings of this study. Moreover, we did not include in our analyses other smaller species, such as crabs, which may benefit from MPAs and influence urchin populations by feeding on juveniles (Clemente et al. 2013 ). We excluded these species because of the current limited understanding of their role as urchin predators. Finally, we were not able to explore evidence for trophic cascades within Central California as there is no population data for otters at the same scale and resolution of the MLPA data. Besides trophic interactions, there are additional potential reasons why spatial protection in Central California was not associated with increased climate resilience for kelp forests. First, this region was less impacted by the MHWs themselves. Notably, on average kelp area remained 1.5–3 times higher in Central than Southern California during and after the MHW (Figure 2 ), although the underwater densities are similar (Figure 4 ). It is no surprise that level of protection had no effect in Central California because, regardless of protection, giant and bull kelp forests were not as impacted from the MHWs, even though they experienced a steady decline after the MHWs. In addition, large areas of Central California are less accessible to people and therefore are less impacted by human activities, including fishing (Free et al. 2023 ), than in Southern California, and because density of the remaining urchin predators (federally protected sea otters) is largely uncorrelated with protection status. Our results are in general agreement with previous studies that also found limited contribution of MPAs to climate resilience for kelp forest communities in Central California (Smith et al. 2023 ). These findings suggest that it is a priority to assess the benefits of MPAs for providing climate resilience in regions that are more impacted by climate change and human activities. Our study casts new light on differences in climate resilience between two regions in California and, most importantly, highlights the importance of the local ecological context in determining whether MPAs can be expected to buffer climate extremes. Our findings have important implications for evaluating the benefits that MPAs can confer in terms of mitigating the impacts of climate change, and also for informing approaches to climate‐smart management and establishment of new MPAs (Arafeh‐Dalmau et al. 2023 ) as nations make progress toward protecting 30% of the oceans by 2030 while adapting to climate change (Convention of Biological Diversity 2022 ). Understanding which mechanisms confer climate resilience at different levels of biological organization (species, population, and ecosystem), and at local to regional scale, is crucial to inform realistic expectations of the climate benefits that MPAs or other management options may provide. There is a need for deeper understanding of the mechanisms that drive ecosystem resilience to understand where placing MPAs may increase climate resilience. Furthermore, such understanding requires continued investment in long‐term monitoring and standardized metrics to define and measure ecological resilience to evaluate the conditions under which MPAs confer resilience to climate impacts. The most important implication of our findings is that protection of top predators confers benefits that propagate through the ecosystem, boosting resilience to and recovery from acute impacts of climate change. While this goal often underpins the establishment of MPAs, its effectiveness in providing climate resilience is seldom supported by empirical evidence. Additional research is required to assess the generality of our findings, but they provide a strong motivation to carefully manage fishing pressure in the coastal zone as climate extremes become more frequent and intense (Oliver et al. 2018 ; Schoeman, Bolin, and Cooley 2023 ). MPAs offer many benefits from preventing continued destruction of habitats (including blue carbon ecosystems such as seagrass and mangroves), increasing food security, and increasing resilience to climate shocks and environmental variability, ultimately increasing overall ecosystem resilience (Aburto‐Oropeza et al. 2011 ; Jacquemont et al. 2022 ; Miteva, Murray, and Pattanayak 2015 ; Selig and Bruno 2010 ). However, MPAs are not a panacea to the ongoing and projected impacts of climate change. In particular, our results of context‐dependent roles of MPAs in conferring climate resilience highlight the urgency to carefully consider what and where additional measures are needed, such as the protection of wide‐ranging top predators to the active restoration of habitat and critical species interactions. Crucially, the root causes of climate change and global biodiversity loss must be urgently addressed before the efficacy of our adaptation tools is lost (Mills et al. 2023 )."
} | 5,930 |
29387787 | PMC5786439 | pmc | 3,078 | {
"abstract": "Clustered ferromagnetic Josephson junctions form ultralow energy synaptic elements.",
"introduction": "INTRODUCTION Recently, neuromorphic computing has been demonstrated in a number of hardware platforms, including modified complementary metal-oxide semiconductor (CMOS) architectures ( 1 ), static random-access memory–type synapses with CMOS neurons ( 2 ), and memristive synapses with CMOS neurons ( 3 ). These implementations have demonstrated significant improvements in efficiency compared to software neural networks run on CMOS platforms using a standard von Neumann architecture, and yet, all of these implementations are orders of magnitude less efficient than the human brain ( 2 ). The synapse is widely believed to be integral for both learning and memory ( 4 ). With approximately 10 15 synapses in the human brain, it is a critical component of neural circuitry. Hence, finding a simple, low-energy, artificial synapse is an important step in making a neuromorphic computer that can approach the level of complexity of the human brain. The potential gains from neuromorphic computing have led to intense research in devices that can mimic the functionality of a synapse ( 5 – 12 ). Here, we demonstrate a new physical implementation of an artificial synapse that is orders of magnitude more efficient than that of a human brain, with a spiking energy that is sub-attojoule per synaptic event. These new artificial synapses are compatible with single flux quantum (SFQ) Josephson junction (JJ) circuits that can provide the underlying technology platform needed to scale large neuromorphic systems ( 13 – 16 ). On the basis of the Josephson effect ( 17 ), a JJ can produce fast, low-voltage spikes of a few picoseconds in duration when the current through the junction exceeds its critical current I c ( 13 ). These SFQ spikes have a time-integrated voltage amplitude given by the flux quantum Φ 0 = 2.068 × 10 −15 V·s. With its intrinsic spiking behavior, neuromorphic SFQ circuitry has been proposed and demonstrated at a basic level ( 18 – 23 ). A missing component of these neuromorphic systems has been a compact low-energy plastic synapse analog. Here, we demonstrate a synaptic element based on a dynamically reconfigurable JJ synapse capable of non-Hebbian learning ( 4 , 24 ). We also demonstrate the basic energy and size scaling behavior required to realize a low-power, complex neuromorphic system based on these artificial synapses combined with SFQ neurons.",
"discussion": "DISCUSSION A new type of dynamically tunable JJ has been demonstrated with potential application as a synthetic synaptic element for superconducting neuromorphic computing. In these devices, the JJ critical current, in a zero applied magnetic field, can be varied in an analog manner. Although digital implementations of JJs with magnetic barriers have been extensively demonstrated ( 26 , 35 , 42 , 43 ), to our knowledge, reproducible zero-field analog control has not previously been shown. The device behaves naturally as an artificial synapse that can interact with low-energy electrical pulses. There are potentially other applications of these novel JJs. For example, the low energy involved in changing the order of the magnetic clusters makes this barrier an excellent candidate for cryogenic memory applications, possibly as the magnetic free-layer material in a pseudospin-valve magnetic JJ. In addition, because of their natural spiking behavior, these devices could be used as integrate and fire neurons. The device demonstrated here requires an external magnetic field to order the magnetic nanoclusters in the JJ synapse. One physical implementation would be to lithographically pattern magnetic field control lines above the JJ synapses. This type of implementation should work up to considerable scale, as evidenced by field-switched magnetic random-access memory products currently available with more than 1.6 × 10 7 bits per chip ( 44 ). However, implementations without additional field control lines would allow an even better scaling. It has previously been demonstrated that spin transfer torques can be used to switch a magnetically soft layer in a two-state JJ with a magnetic pseudospin-valve barrier ( 26 , 35 ). Similar spin-polarizing layers should also allow for the elimination of a field control line in future JJ synapses. In addition, if the energy required to reorder the magnetic clusters is on the order of the SFQ pulse energy, then the JJ synapses will be capable of Hebbian learning ( 24 ). The scaling of pulse energy with device size appears to be a promising path toward realizing magnetic cluster reordering energy that will be on the order of the SFQ pulse energy. One of the significant advantages of this system is its ability to leverage existing digital SFQ JJ logic infrastructure. Digital SFQ circuits have been demonstrated with more than 10 5 JJs ( 45 ). In addition, because of their superconducting nature, SFQ circuits can have so-called ballistic communications, both on-chip and chip-to-chip, with data transmission rates demonstrated up to 60 gigabits/s ( 46 – 48 ). The main advantage is that the transmission of SFQ pulses is not limited by the typical capacitive charging, which semiconductor electronics typically face. In addition, there is very little energy cost in transmitting the pulses along superconducting wires/transmission lines ( 30 ), and zero static dissipation SFQ circuits have recently been demonstrated ( 49 ). This results in the main energy dissipation mechanism as the pulse generation itself. For example, in our basic circuit element in Fig. 3 , the energy dissipated for a presynaptic spike, followed by the synaptic weighting and output spike (if the synaptic weight is above threshold), would be ~3 I b Φ 0 ~ 1 aJ. The main additional cost is cooling to 4 K, which, for large-scale systems, is ~1000, bringing the energy required for this operation up to ~1 fJ. There would be an additional cost for supplying the bias currents, but these currents could be shared within each layer because they are not the tunable element. The high-speed communications could be combined with one of the many neuromorphic interconnect strategies that have been developed in CMOS, such as matrix-vector multipliers combined with thresholding ( 50 ), tree routing ( 51 ), or self-timing ( 52 ). Leveraging of fabrication tools and techniques that have been developed for CMOS and modified for both commercial magnetoresistive random-access memory and large-scale digital SFQ logic should greatly accelerated the progress of neuromorphic circuits using JJ neurons and JJ synapses. However, we should note that, although the demonstrated device and the circuit modeling are quite promising, considerable work remains to demonstrate a large neuromorphic architecture with complex layout and routing using this new technology."
} | 1,717 |
35914169 | PMC9371685 | pmc | 3,079 | {
"abstract": "Significance Wetlands are the major natural source of methane, an important greenhouse gas. The sulfur and methane cycles in wetlands are linked—e.g., a strong sulfur cycle can inhibit methanogenesis. Although there has historically been a clear distinction drawn between methane and sulfur oxidizers, here, we isolated a methanotroph that also performed respiratory oxidization of sulfur compounds. We experimentally demonstrated that thiotrophy and methanotrophy are metabolically compatible, and both metabolisms could be expressed simultaneously in a single microorganism. These findings suggest that mixotrophic methane/sulfur-oxidizing bacteria are a previously overlooked component of environmental methane and sulfur cycles. This creates a framework for a better understanding of these redox cycles in natural and engineered wetlands.",
"conclusion": "Conclusions A methanotrophic bacterium showing respiratory oxidation of sulfur compounds was discovered, which greatly expands the current concept of facultative methanotrophy. This study revealed that thiotrophy and methanotrophy are metabolically compatible, which blurs the long-observed distinctness of methanotrophic and thiotrophic microorganisms. It also highlights the difficulties in inferring methanotrophy and thiotrophy based on phylogeny. Further studies on the regulation, kinetics, and evolution of thiotrophic and methanotrophic metabolisms in this strain are required. Due to possible combination with other metabolisms, unexpected metabolic versatility of methanotrophs could remain to be discovered in various sediment environments. Together, our findings set a framework for better understanding methane- and sulfur-cycle interaction in the oxic–anoxic interface zone of natural and engineered ecosystems. Isolated Methylovirgula Species. Strain HY1 is a novel species of the genus Methylovirgula of the family Beijerinckiaceae and is capable of both methane and sulfur oxidation for growth. We propose the following candidate status: Taxonomy. (i) Etymology. The taxonomy for Methylovirgula thiovorans sp. nov. is as follows: thi.o.vo’rans. Gr. neut. n. theion, sulfur; L. pres. part. vorans, eating; N.L. part. adj. thiovorans, sulfur-eating. (ii) Habitat. An acidic wetland on top of Mount (Mt.) Daeam located in South Korea. (iii) Diagnosis. Cells are straight rods, with a diameter of 0.2 to 0.3 µm and a length of 0.4 to 1.7 µm ( SI Appendix , Fig. S19 ). Grows between 15 and 30 °C with an optimum at 27 °C and between pH 4.0 to 6.0 with an optimum at pH 4.5. The strain grows via oxidizing methane, various multicarbon organic compounds, and reduced sulfur compounds ( Table 1 ). The G + C content of the type strain is 60.1 mol%. The 16S rRNA gene sequence similarity is 98.7% with M. ligni and ranges from 95.8 to 97.3% with closely related genera: Methylocapsa , Methyloferula , Beijerinckia , and Methylocella .",
"discussion": "Results and Discussion Isolation of Methanotrophic Strain HY1. Wetland samples were incubated at pH 4.0 in a low-salt mineral (LSM) medium under methanotrophic conditions. From a methane-oxidizing enrichment culture transferred biweekly for about 6 mo, colonies of methanotrophs were retrieved by using a floating filter technique. A methanotrophic isolate designated HY1 showed 98.7% 16S ribosomal RNA (rRNA) gene-sequence identity to Methylovirgula ligni within the family Beijerinckiaceae in the order Hyphomicrobiales ( SI Appendix , Fig. S1 ). Culture purity was verified via Illumina sequencing of a 16S rRNA gene amplicon, as well as via full genome sequencing and assembly. The average nucleotide identity value for the genome of strain HY1 was 73.6% with that of M. ligni and ranged from 72.2 to 73.6% with selected members of the Beijerinckiaceae, indicating that the isolate represents a novel species of the genus Methylovirgula . Growth was observed at moderate temperature in an acidic pH range between pH 4 and 6 and an optimum at pH 4.5 on methanol ( SI Appendix , Fig. S2 ), similar to other methanotrophic members of the Beijerinckiaceae ( SI Appendix , Table S1 ). Thiotrophic Growth. Previously isolated strains of Methylovirgula are methylotrophs that cannot oxidize CH 4 as an energy source ( 27 ), although they are closely related phylogenetically to the methanotrophic genera Methylocella and Methyloferula within the Beijerinckiaceae ( 28 ) ( Fig. 1 and SI Appendix , Fig. S1 ). Strain HY1 was observed in batch culture to grow on methane as well as various C1–C4 alcohols, organic acids, and short-chain alkanes ( Table 1 ). This substrate range is similar to facultative methanotrophs of the genus Methylocella ( Table 1 ) ( 11 , 12 , 29 , 30 ), which show 95.8 to 96.3% 16S rRNA gene-sequence similarity with strain HY1. Table 1. \nSubstrate utilization by strain HY1\n Substrate Concentration Growth Strain HY1 M. silvestris BL2 M. ligni BW863 Organic C1 Methane 2.5 to 20% ++ ++ – Methanol 20 mM ++ ++ ++ Formate 5 mM + + – Formate 20 mM – ++ – C2 Ethane 2.5 to 20% ++ ++ – Ethanol 5 mM, 20 mM + + + Acetate 5 mM + ++ – Acetate 20 mM – – – Oxalate 5 mM ++ – – Oxalate 20 mM + – – C3 Propane 2.5% + ++ – Propane 20% – ++ – 1-Propanol 5 mM + – – 1-Propanol 20 mM – – – 2-Propanol 5 mM + ++ – 2-Propanol 20 mM + ++ – 1,2-Propanediol 5 mM, 20 mM ++ ++ – Acetone 5 mM ++ ++ – Acetone 20 mM + + – Acetol 5 mM ++ ++ – Acetol 20 mM + ++ – Pyruvate 5 mM – ++ + Pyruvate 20 mM + ++ ++ C4 Butane 2.5% + – – Butane 20% – – – 1-Butanol 5 mM ++ – – 1-Butanol 20 mM – – – 2-Butanol 5 mM ++ – – 1-Butanol 20 mM – – – Butanal 5 mM ++ – – Butanal 20 mM + – – 2-Butanone 5 mM ++ – – 2-Butanone 20 mM + – – Succinate 5 mM + ++ – Succinate 20 mM ++ ++ – Malate 5 mM + + + Malate 20 mM ++ ++ ++ Inorganic Sulfur * Thiosulfate 1 mM + – – Thiosulfate 15 mM + – – Tetrathionate 2 mM + – – Elemental sulfur (S°) 3 g/L + – – Others H 2 20% – – – CO 20% – – – Each substrate tested was supplied as a sole energy source to batch cultures. Growth is indicated as follows: ++, growth to OD 600 > 0.25; +, growth to OD 600 ≥ 0.04 to 0.25; –, growth to OD 600 < 0.04. All the substrates were tested with 5% CO 2 . M. silvestris BL2 and M. ligni BW863 were retested in parallel to strain HY1, and the results largely conformed to previous reports ( 11 , 12 , 29 , 30 ). Accumulation of sulfate was also used as an indicator of biological sulfur oxidation. * Due to rapid autoxidation of sulfide in our condition ( 126 , 127 ), growth on sulfide could not be tested. Instead, a microrespirometry experiment was performed to show the oxidation of sulfide by strain HY1 by using high-density cells ( Table 2 ). Surprisingly, strain HY1 also grew on the inorganic sulfur compounds thiosulfate, tetrathionate, and elemental sulfur (S°), a capability never before observed in any methanotroph ( Table 1 ). Because growth of a single organism on methane and reduced sulfur was an unprecedented observation, these processes were carefully verified by time-course analysis of batch cultures using analytical methods described in SI Appendix, Analytical Methods . Consumption of methane or thiosulfate as sole growth substrates and production of sulfate from thiosulfate were observed concurrent with population growth ( Fig. 2 ). A near-stoichiometric conversion of thiosulfate to sulfate at 1:1.98 was observed when low thiosulfate concentrations (<5 mM) were added ( Eq. 1 ) ( Fig. 2 B ). [1] S 2 O 3 2 − + 2 O 2 + H 2 O → 2 SO 4 2 − + 2H + . Fig. 2. Growth and time course of methane and thiosulfate oxidation by batch cultures of strain HY1. ( A ) Time courses of methane oxidation and the concomitant growth of strain HY1. Due to oxygen depletion, methane was not completely oxidized. ( B ) Time courses of thiosulfate oxidation, sulfate production, and the concomitant growth of strain HY1. The stoichiometry of thiosulfate consumption to sulfate production was nearly equal to the predicted 1:2. Specific analytical assays for methane, sulfate, and thiosulfate in batch cultures are described in SI Appendix , Analytical Methods . Error bars represent ±1 SD of three biological replicates. V/v, vol/vol. Abiotic decomposition of thiosulfate was observed only at pH < 3.0 ( SI Appendix , Fig. S3 ). In this study, strain HY1 was cultivated in a buffered medium at pH 4.5 to 5.0, and abiotic decomposition of thiosulfate was therefore assumed to be negligible. An increase in 16S rRNA gene copies during the oxidation of methane or thiosulfate indicated that both substrates supported population growth ( Fig. 2 ). Biomass production was also demonstrated via quantification of cellular protein in the cultures and measurement of molar growth yield Y x/m (g dry cell weight⋅mol −1 substrate) ( Fig. 3 ). Furthermore, methane and thiosulfate were oxidized concomitantly when both substrates were provided simultaneously ( SI Appendix , Fig. S4 ). Accordingly, the biomass produced when methane and thiosulfate were simultaneously used was almost equal to the sum of the biomass produced when a similar amount of methane and thiosulfate was utilized individually ( Fig. 3 A ). The biomass molar growth yields ( Y x/m ) were comparable between methane- and thiosulfate-grown cells ( Fig. 3 B ), as expected from the similar standard free-energy changes for their oxidations (introduction). Fig. 3. Biomass production in strain HY1 grown on methane and thiosulfate. ( A ) The growth yield was calculated as milligrams of cellular protein produced per culture volume (mg⋅protein⋅L −1 ) after the complete oxidation of the substrate(s). Strain HY1 was grown in 100 mL of LSM medium at pH 5.0 with methane (15%, vol/vol), thiosulfate (4 mM), and methane+thiosulfate (15%, vol/vol; 4 mM), respectively, in 160-mL serum vials. For the complete oxidation of substrates, 60% (vol/vol) oxygen was supplied. ( B ) The molar growth yield ( Y x/m ) was calculated as gram of dry cell weight per mol of substrate consumed (g dry cell weight⋅mol −1 ⋅substrate). Error bars represent ±1 SD of three biological replicates. Genomic Properties. The observed capacity of strain HY1 to grow on both methane and reduced sulfur compounds prompted us to investigate the genomic basis for these processes. The final assembled genome contained two circular contigs: a circular chromosome and a 278-kb circular megaplasmid. The overall genomic features of strain HY1 compared with other methanotrophs in the Beijerinckiaceae are presented in SI Appendix , Table S1 . The most surprising finding from the genomic analysis was the identification of a comprehensive genetic repertoire encoding the utilization of reduced sulfur compounds as electron donors ( Fig. 1 and Datasets S1 and S2 ). The assembly verified the presence of genes encoding methane oxidation and sulfur oxidation within a single organism’s genome. Genes encoding for methane and short-chain alkane oxidation. The key genes encoding CH 4 oxidation predicted from the genomic analysis are presented in Dataset S2 , and the predicted pathways for methane oxidation are presented in Fig. 4 . While a gene cluster for sMMO, a soluble diiron monooxygenase family enzyme ( 31 ), is present, genes for particulate methane monooxygenase (pMMO) are absent. The genes encoding sMMO, mmoXYBZDC , are closely related phylogenetically to those of other methanotrophs possessing only sMMO, such as Methylocella , Methyloferula , and Methyloceanibacter ( 32 , 33 ), and the gene arrangements are highly conserved among these methanotrophs ( SI Appendix , Fig. S5 ). Fig. 4. Proposed central carbon and energy metabolism in strain HY1 and differential protein abundance between methane-grown and thiosulfate-grown cells. The color scale indicates whether proteins have higher abundance in methane-grown (red) or thiosulfate-grown (blue) cells. The intensity of the color in each protein indicates the relative fold change difference (log 2 FC). Methane oxidation: Methane is oxidized to methanol by the soluble methane monooxygenase, sMMO (MHY1_02902–2908). The produced methanol is oxidized to formaldehyde via the lanthanide-dependent MDH, XoxF (MHY1_02202 was the most abundant MDH). Formaldehyde oxidation to formate then proceeds via the tetrahydromethanopterin (H 4 MPT) pathway, and C1 incorporation into the serine cycle is mediated by the tetrahydrofolate (H 4 F) carbon-assimilation pathway. Sulfur oxidation: In the periplasm, two thiosulfate molecules are oxidized to tetrathionate by thiosulfate dehydrogenase, DoxDA (MHY1_01298), and then TetH (MHY1_02468) hydrolyzes tetrathionate to sulfate and disulfane monosulfonic acid, which most probably decomposes spontaneously to thiosulfate and sulfur. Sulfane sulfur derived from thiosulfate and sulfide via SoxYZAB (MHY1_00063–66) and Sqr (MHY1_02376), respectively, is transported into the cytoplasm via PmpAB (MHY1_00234–235 and MHY1_01361–1362; only MHY1_01361–1362 are indicated here), then transferred to the DsrEFH (MHY1_00081–83) and DsrC (MHY1_00084) via the sulfur-transporting complex [rhodanese (MHY1_01281)-TusA (MHY1_00072)-DsrE2A (MHY1_00073)]. The persulfurated DsrC is oxidized to DsrC and sulfite by DsrAB sulfite reductase (MHY1_00079–80), thereby releasing electrons to the iron–sulfur flavoprotein, DsrL (MHY1_00087). Sulfite is probably transported to the periplasm by a TauE-like exporter (MHY1_01299). The sulfite:cytochrome c oxidoreductase, SorAB (MHY1_p00095–0096), encoded in the megaplasmid, might be involved in sulfite oxidation. It is speculated that the TauD/DsrQ protein (MHY1_00078) catalyzes the release of sulfite during the breakdown of sulfonates. Another key genomic trait of strain HY1 is the absence of genes encoding Ca 2+ -dependent methanol dehydrogenase (MDH) or PQQ-dependent alcohol dehydrogenase. Instead, strain HY1 contains four genes encoding lanthanide-dependent MDH (XoxF-type MDH) affiliated with the XoxF3 and XoxF5 clades ( 34 ) ( SI Appendix , Table S2 ). Consistent with this finding, strain HY1 could not grow on methane or methanol without the addition of lanthanides to the medium ( SI Appendix , Fig. S6 ). Aside from the verrucomicrobial methanotrophs ( 35 – 37 ), strains PC1 and PC4 of Methylocella tundrae have been reported to contain only XoxF-type MDH and are dependent on lanthanides for their growth ( 30 ). The tetrahydromethanopterin (H 4 MPT) pathway for formaldehyde oxidation to formate and the tetrahydrofolate (H 4 F) pathway mediating C1 transfer to the serine cycle, which are widely conserved in Beijerinckiaceae methanotrophs, are both encoded in strain HY1 ( Dataset S2 ). Strain HY1 also utilized butane (2.5%, volume [vol]/vol) and its possible catabolic intermediates (i.e., l-butanol and isobutanol) ( Table 1 ), which is unique among known methanotrophs ( 29 ). Methylocella silvestris BL2 is equipped with genes encoding two different soluble diiron monooxygenase family enzymes—i.e., sMMO and propane monooxygenase—and can utilize C 1 –C 3 alkanes for growth ( 12 ). In contrast, the genome of strain HY1 encodes a single diiron monooxygenase, sMMO. Nevertheless, it cannot only grow on C 1 –C 3 alkanes, but also on butane (C 4 ). Notably, the dedicated propane monooxygenase from M. silvestris BL2 requires high propane concentrations (20%, vol/vol) ( Table 1 ). Thus, the absence of a related propane monooxygenase may explain why strain HY1 utilizes propane (and butane) only when provided at low concentrations (2.5%, vol/vol). We hypothesize that the single sMMO in strain HY1 has a broad substrate range, including butane, and initiates oxidation of various short-chain alkanes to both primary and secondary alcohols, as reported for other soluble methane monooxygenases ( 38 – 40 ). The growth of strain HY1 on various alcohols is likely attributable to the presence of a repertoire of various XoxF-type MDHs with potential broad substrate specificity ( SI Appendix , Table S2 ) ( 37 , 41 – 43 ). The enzymes required to oxidize the ketones derived from secondary alcohols remain elusive ( 44 – 47 ). The complete set of genes encoding for the tricarboxylic acid cycle in strain HY1 ( Dataset S2 ) is a common trait of alphaproteobacterial methanotrophs ( 48 ), but is also consistent with strain HY1’s utilization of multicarbon compounds. Genes encoding for oxidation of reduced sulfur compounds. A repertoire of genes encoding enzymes involved in sulfur oxidation ( soxYZAB , dsrABEFHCMKLJOPN , sqr , sorAB , tetH , and doxAD ) suggested that strain HY1 is capable of using various reduced sulfur compounds for growth ( Fig. 1 and Dataset S2 ). Predicted pathways of sulfur oxidation are summarized in Fig. 4 and described in detail below. In strain HY1, the sox , dsr , and other genes form a single gene cluster ( SI Appendix , Fig. S7 ). The combination of a periplasmic, truncated Sox system, which lacks SoxCD, with a cytoplasmic rDsr system in strain HY1 is widespread in bacterial sulfur oxidizers and occurs in at least four class-level lineages, including Alphaproteobacteria, Betaproteobacteria, Gammaproteobacteria, and Chlorobia ( Fig. 1 ). Many organisms containing these genes, such as green and purple anoxygenic phototrophic sulfur bacteria, form sulfur deposits as a characteristic intermediate en route to the end-product sulfate ( 49 – 51 ). The oxidation of reduced sulfur compounds—e.g., thiosulfate, sulfide, or polysulfides—to sulfate is always initiated in the periplasm, and we can confidently state that the SoxAYZAB proteins in strain HY1 constitute a periplasmic multienzyme system ( Fig. 4 ). SoxYZ serves as a carrier protein to which reduced sulfur compounds remain attached during the oxidation process. SoxA(X) catalyzes the oxidative fusion of the sulfur substrate (e.g., thiosulfate) to a conserved cysteine of SoxY. A gene encoding SoxX was not found in the genome of strain HY1, as has been reported for the sulfur oxidizers Halomonas halophila and Beggiatoa sp. PS ( 52 ). In classical heterodimeric SoxAX proteins, the c -type cytochrome SoxX serves as the site of electron storage and transfer to an electron-transfer partner cytochrome c during turnover of the enzyme, while SoxA harbors the catalytically active site ( 52 ). It is therefore conceivable for strain HY1 that SoxA alone is active and that it transfers electrons directly to a separate c -type cytochrome acceptor encoded elsewhere in the genome. Similar observations have been made for thiosulfate dehydrogenase, where the encoding gene tsdA is accompanied by the gene tsdB for the electron-accepting cytochrome c in many, but not in all, TsdA-containing organisms ( 53 ). Once thiosulfate is bound to SoxYZ, the sulfone group (−SO 3 − ) is hydrolytically released by SoxB. The SoxB of strain HY1 is related phylogenetically to other alphaproteobacterial SoxB proteins ( SI Appendix , Fig. S8 ) and shows the closest affiliation to the enzymes from known sulfide- and/or thiosulfate-oxidizing organisms of the order Hyphomicrobiales, including aerobic members of the genus Magnetospirillum ( 54 ). Other alphaproteobacterial methanotrophs that contain soxB are Methylosinus trichosporium OB3b ( 55 ) and Methylosinus sp. 3S-1 ( 56 ), but additional components of the Sox machinery (SoxA and SoxX) are not encoded in these organisms ( Dataset S1 ). Thus, it is unlikely that they can oxidize thiosulfate. A complete Sox system containing a SoxCD has been reported in genomes of the gammaproteobacterial methanotroph genera Methylotuvimicrobium , Methylicorpusculum , Methylobacter , and Methylosarcina , but their potential for sulfur oxidation has not been experimentally proven ( Dataset S1 ). Methylicorpusculum oleiharenae XLMV4 could not be grown on thiosulfate compounds, despite possessing a complete Sox system ( 57 ). The lack of a potential redox-active motif (CxxxC) in SoxD of methanotrophs ( SI Appendix , Fig. S9 ) implies that it might be nonfunctional or has a different metabolic function ( 58 , 59 ). In that case, they would need an auxiliary system, such as the rDsr or Hdr (heterodisulfide reductase)-like system, for sulfur oxidation to sulfite. Strain HY1 encodes rDsr, but these enzymes are not encoded in any other methanotroph genomes presently available ( Dataset S1 ). As the genome of HY1 does not encode sulfane dehydrogenase SoxCD, sulfane sulfur atoms (−S − ) bound to SoxYZ cannot be further oxidized in the periplasm. The described attachment of thiosulfate and release of sulfate likely repeats, and chains of sulfur atoms are formed that probably lead to the intermediary formation of sulfur deposits. Two other enzyme systems can also contribute to sulfur formation. Strain HY1 encodes enzymes of the so-called S 4 intermediate (S 4 I) pathway that act on thiosulfate in the periplasm ( Fig. 4 ): Two thiosulfate molecules are first oxidized to tetrathionate by thiosulfate dehydrogenase (DoxDA), and then tetrathionate hydrolase (TetH) hydrolyzes tetrathionate to sulfate and disulfane monosulfonic acid, which probably decomposes spontaneously to thiosulfate and sulfur ( 60 , 61 ). In addition, gene MHY1_02376 encodes a type I Sqr, a periplasmically oriented monotopic membrane protein catalyzing the oxidation of sulfide and probably releasing hydrophilic polysulfides that can also contribute to sulfur deposition ( Fig. 4 ) ( 62 , 63 ). In the next step, sulfur formed either via the Sox or S 4 I pathways or resulting from the action of Sqr is transferred into the cytoplasm ( Fig. 4 ). How this is achieved in general by sulfur oxidizers is unclear. A YeeE/YedE protein resembling the thiosulfate transporter from Escherichia coli and Spirochaeta thermophila ( 64 ) is possibly involved in the process in Hyphomicrobium denitrificans. However, this organism does not contain Dsr proteins ( 65 ). Strain HY1 does not encode a full-length YeeE/YedE homolog; instead, two sets of pmpA and pmpB homologous genes are present (MHY1_00234–235 and MHY1_01361–1362). PmpAB conform to the first third of the YeeE/YedE domain and have also been discussed as potential transport components for sulfur-containing compounds ( 66 ). Notably, a gene encoding a periplasmic DsrE-like sulfurtransferase (MHY1_00068) resides in the sox operon of strain HY1, and a membrane-bound DsrE-like sulfurtransferase (MHY1_00073) is encoded in the immediate vicinity, between the sox and dsr clusters. We consider the possibility that the first may be involved in sulfur transfer to the latter, which may then mediate transport of the sulfur into the cytoplasm. Inside the cytoplasm, sulfur never occurs in free form, but is handled by further sulfurtransferases, such as TusA (MHY1_00072) and DsrEFH (MHY1_00081–83), and treated by the concerted action of dsr -encoded enzymes ( Fig. 4 ) ( 49 , 50 , 67 ). Fully in line with this concept, the key enzyme of the pathway, siroheme-containing dissimilatory sulfite reductase (DsrAB) from strain HY1 falls into the rDsr group (i.e., oxidative DsrAB used for sulfur oxidation) ( SI Appendix , Figs. S7 and S10 ). It is related to rDsrAB sequences from other Alphaproteobacteria (most notably organisms from the order Hyphomicrobiales). Members of the order Hyphomicrobiales, containing closely related rDsrAB—i.e., Rhodobium orientis and Rhodomicrobium vannielii —have a documented capacity for autotrophic growth on sulfide or thiosulfate ( 68 , 69 ). A further closely related rDsrAB sequence is that of a metagenome-assembled genome of a canonical sulfur-oxidizing phototroph (purple nonsulfur bacterium) of the genus Rhodomicrobium , recovered from a permafrost thaw wetland, which was also found to contain sMMO and PMO gene clusters ( 26 ). The protein DsrC (MHY1_00084) is characterized by a conserved carboxyl-terminal motif (Cys B –X 10 –Cys A ) and plays a central role in the rDsr pathway, as it acts as a sulfur carrier that is loaded by DsrEFH ( 70 ). The membrane-bound DsrMKJOP complex is assumed to oxidize persulfurated DsrC, thus generating DsrC trisulfide, which then serves as the substrate for rDsrAB. The electrons released by the oxidation of the sulfur bridged between DsrC-Cys A and Cys B to sulfite are probably transferred to NAD + . This reaction is catalyzed by the NADH:acceptor oxidoreductase DsrL (MHY1_00087) ( 49 , 71 , 72 ). Notably, two additional DsrC-like proteins are encoded in strain HY1 (MHY1_00075 and MHY1_00097). Both lack Cys A and probably function as regulatory sulfur-related proteins (RpsA) ( 73 ). In strain HY1, sulfite formed by the rDsr pathway does not appear to be further processed in the cytoplasm, as genes encoding adenosine-5′-phosphosulfate reductase (AprBA) and ATP sulfurylase (Sat) ( 49 ) or the cytoplasmically oriented membrane-bound sulfite-oxidizing enzyme SoeABC ( 74 ) are all absent. Instead, sulfite is probably transported to the periplasm by a TauE-like exporter (MHY1_01299) ( 75 ) and oxidized to sulfate by periplasmic sulfite:cytochrome c oxidoreductase (SorAB; MHY1_p00095–96 residing on the megaplasmid in strain HY1). Within the family Beijerinckiaceae, genetic equipment of sulfur oxidation closely resembling that of strain HY1 is found in Rhodoblastus acidophilus ( Fig. 1 and SI Appendix , Figs. S7 and S10 ), but, surprisingly, the growth of this bacterium was not supported by sulfur compounds in laboratory tests ( 76 ). A putative cytoplasmic αβγδ-heterotetrameric, bidirectional hydrogenase is also encoded by MHY1_02662–2665. This resembles Pyrococcus furiosus sulfhydrogenase that catalyzes H 2 production and H 2 oxidation coupled with the reduction of elemental sulfur and polysulfide to sulfide ( 77 , 78 ). The importance of this enzyme for sulfur metabolism is unclear. Genes Encoding Autotrophy. In order for strain HY1 to grow by oxidizing reduced sulfur as electron donors, a system for autotrophic CO 2 fixation is required since the serine cycle can be used only in conjunction with methane or methanol oxidation. Indeed, strain HY1 encodes genes required for a complete Calvin–Benson–Bassham (CBB) cycle, as observed in other methylotrophs and methanotrophs of the family Beijerinckiaceae ( Fig. 1 and Dataset S2 ). The RubisCO large subunit (CbbL) of strain HY1 shares 94.5 to 96.1% amino acid similarities with the form I enzyme from strains of Methylocella and Methylocapsa . The CBB cycle is widespread in proteobacterial autotrophic sulfur oxidizers ( 51 ) and is also used for chemolithotrophic and methanotrophic growth by verrucomicrobial methanotrophs ( 20 , 79 ). A methylotroph closely related to HY1, Beijerinckia mobilis , also grows autotrophically on methanol by using the CBB cycle ( 80 ), further suggesting the possible role of this cycle in carbon fixation during the growth of strain HY1 on sulfur compounds. In contrast, none of the gammaproteobacterial methanotrophs with the complete Sox system contained cbbL and cbbS genes for the CBB cycle ( Fig. 1 ), which may explain the inability of gammaproteobacterial methanotrophs to grow chemolithoautotrophically. As expected, the growth of strain HY1 at pH 4.5 with methane, thiosulfate, or both as substrates was strictly dependent on CO 2 supplementation in the headspace (10%, vol/vol), possibly due to the requirement for CO 2 fixation via the serine or CBB cycle ( SI Appendix , Fig. S11 ). Similarly, the growth of verrucomicrobial methanotrophs requires supplementation of the medium with CO 2 ( 19 , 20 ). This indicates that a high mixing ratio of CO 2 is critical for autotrophs growing at acidic pH below the p K a of bicarbonate (pH 6.1), at which bicarbonate is converted to CO 2 , which has limited water solubility. Induction of Methane and Sulfur Oxidation. To determine whether methane and sulfur metabolisms are constitutive or inducible in strain HY1, oxygen-consumption rates were measured in microrespirometry experiments by using whole cells grown on methane, thiosulfate, or methane+thiosulfate ( Table 2 ). Methane-grown cells consumed oxygen in the presence of methane, but not in the presence of reduced sulfur compounds. Similarly, thiosulfate-grown cells could not consume oxygen in the presence of methane. Thus, neither methanotrophy nor oxidative sulfur metabolism is constitutive. Thiosulfate-grown cells actively consumed oxygen in the presence of tetrathionate, indicating that the S 4 I pathway is involved in thiosulfate oxidation. Both methane- and thiosulfate-grown cells consumed oxygen with methanol, ethanol, and 1-propanol at similar rates, indicating constitutive expression of alcohol-utilization enzymes ( Table 2 ). Table 2. Substrate-specific oxygen-consumption rate by strain HY1 Substrate Oxygen uptake rate (µmol⋅mg⋅protein −1 ⋅h −1 ) Methane-grown cells Thiosulfate-grown cells Methane+thiosulfate-grown cells Tetrathionate nd 9.36 ± 0.53 1.53 ± 0.07 Thiosulfate nd 7.97 ± 0.41 1.40 ± 0.05 Sulfur nd 2.22 ± 0.11 0.28 ± 0.01 Sulfite 0.51 ± 0.02 * 6.34 ± 0.38 1.01 ± 0.05 Sulfide nd 6.53 ± 0.27 1.21 ± 0.04 Methane 3.48 ± 0.18 nd 2.20 ± 0.11 Methanol 7.64 ± 0.39 4.95 ± 0.21 6.54 ± 0.33 Ethanol 5.45 ± 0.24 4.59 ± 0.26 5.10 ± 0.32 1-Propanol 5.15 ± 0.29 4.11 ± 0.19 4.75 ± 0.22 2-Propanol nd nd nd Methane+thiosulfate nt nt 2.52 ± 0.14 Data are expressed as means ±1 SD ( n = 5). Values (µmol⋅mg⋅protein −1 ⋅h −1 ) are calculated based on the amount of substrate consumed by resting cell suspension at OD 600 = 0.3. The substrates’ concentrations tested were as follows: 500 μM methane; 20 μM alcohol, sulfite, and sulfide; 4 mM thiosulfate and tetrathionate; and 100 mg of sulfur in a 1-mL vial. nd, not detected; nt, not tested. * Reaction stopped in the presence of 200 μM sulfite. Oxygen-consumption rates of thiosulfate-grown cells on reduced sulfur compounds (except for elemental sulfur) were 1.8 to 2.7 times higher than those of methane-grown cells on methane ( Table 2 ), coinciding with the higher μ max on thiosulfate than on methane ( Fig. 2 ). Although the affinity of methane-grown cells for methane [ K m(app) = 148 ± 20 µM] was 12 times higher than that of thiosulfate-grown cells for thiosulfate [ K m(app) = 1,859 ± 76 µM] ( SI Appendix , Fig. S12 ), the affinity of thiosulfate-grown cells to sulfide [ K m(app) = 2.7 ± 0.2 µM] was much higher. Sulfide is a common and preferred sulfur species by all sulfur-oxidizing lithotrophs ( 71 , 81 ). The affinity of the methane-grown cells to methane is similar to values estimated in other methanotrophs containing sMMO, but much lower than the affinities estimated in methanotrophs containing pMMO ( 12 , 82 , 83 ). Since methane concentrations in Yongneup ( Materials and Methods ) are less than the K m(app) , the low-affinity sMMO-methanotroph strain HY1 may benefit greatly from utilizing other energy sources, such as substrates containing carbon–carbon bonds and reduced sulfur compounds ( Table 1 ). In the thiosulfate+methane batch cultures, methane and thiosulfate were oxidized concomitantly for growth ( SI Appendix , Fig. S4 ). As expected, both methane and thiosulfate contributed to net O 2 consumption in microrespirometry experiments using cells grown mixotrophically on methane and thiosulfate ( Table 2 ). These results support that methane- and thiosulfate-oxidation pathways functioned concurrently during the oxidation of both methane and thiosulfate ( SI Appendix , Fig. S13 A ). Accordingly, genes of the serine and CBB cycles for CO 2 fixation were expressed during simultaneous oxidation of methane and thiosulfate ( SI Appendix , Fig. S13 B ). In the microrespirometry experiments, substrate-specific oxygen-consumption rates for each substrate in the mixotrophically grown cells were lower than those in cells grown on either substrate alone. Consistent with this, the mixotrophic growth rate ( μ max = 0.023 ± 0.001 h −1 ) ( SI Appendix , Fig. S4 ) was also lower than the growth rate on thiosulfate ( μ max = 0.033 ± 0.004 h −1 ), although similar to the growth rate on methane ( μ max = 0.022 ± 0.006 h −1 ) ( Fig. 2 ). The facultative methanotroph Methylocystis sp. H2s was also shown to grow more slowly in the presence of methane and acetate than when methane was provided as the only available substrate ( 13 ). As demonstrated in other microorganisms ( 13 , 84 , 85 ), our results indicate that simultaneous utilization of methane and sulfur compounds by strain HY1 may not result in higher growth rates than when a single substrate is utilized, although they do benefit from a greater total substrate pool ( Fig. 3 ). Carbon Monoxide and Hydrogen Metabolism. Strain HY1 possesses coxLMS , encoding carbon monoxide dehydrogenase (CODH) ( Dataset S2 ), which is rare in methanotrophs and found only in the genomes of Methyloferula stellata and M. gorgona ( 16 , 17 , 86 ). The CODH belongs to the form I type, which can be used for the respiratory oxidation of CO ( 87 ) ( SI Appendix , Fig. S14 ). Although strain HY1 could not grow with CO as the sole energy source, CO (<5% CO, vol/vol, in headspace) was concomitantly consumed in the presence of other electron donors, such as methane or thiosulfate ( SI Appendix , Fig. S15 ). Similarly, M . gorgona was found to oxidize CO only in the presence of methane ( 18 ). The inability to grow on CO as a sole substrate is unexpected since strain HY1 contains a complete CBB cycle to support CO 2 fixation. The inability of M . silvestris BL2 T to grow on methane in the presence of CO suggests a function of CODH for detoxification of CO ( SI Appendix , Fig. S16 ). Other potential advantages of CO oxidation in strain HY1, such as increasing starvation survival ( 86 ), are yet unclear. Although many methanotrophs encode hydrogenases ( 20 ), strain HY1 encodes a Group 1e Isp-type hydrogenase, which has not been found in any other methanotroph ( SI Appendix , Fig. S17 ). The genes encoding large, small, and membrane-anchored subunits of hydrogenase are located adjacently on the megaplasmid and are closely related to genes in other sulfur-oxidizing bacteria (e.g., 79.1% and 71.2% similarity with large and small subunits of hydrogenase in Acidithiobacillus sulfuriphilus , respectively) ( Dataset S2 ). Proteobacterial methanotrophs can consume H 2 , although, to date, this process has only been reported as providing reductants to supplement methanotrophic growth ( 88 – 90 ). Strain HY1 could not consume H 2 under any conditions, but a trace amount of hydrogen was accumulated during methane oxidation ( SI Appendix , Fig. S15 ). Since this type of hydrogenase is reversible ( 91 , 92 ), excess reducing power generated from electron donors could be diverted to proton reduction to produce H 2 , as observed in other methanotrophs ( 93 – 95 ) and thiotrophs ( 96 , 97 ). Proteomic Analyses. In order to provide evidence supporting the predicted pathways of reduced sulfur oxidation and methane oxidation in strain HY1 ( Fig. 4 ), proteomes from cells grown on methane and thiosulfate were compared (see the expression of key and all proteins in Datasets S2 and S3 , respectively). The proteome of ethanol-grown cells was analyzed for comparison. The abundances of the structural (MmoX, MmoY, and MmoZ) and chaperone (MmoG) proteins for sMMO were not significantly different between methane- and thiosulfate-grown cells, falling in the range of 0.5 to 4.1% of the total proteins. Abundances of these proteins were higher than those in ethanol-grown cells ( Dataset S2 ). The high abundance of Mmo proteins in thiosulfate-grown cells was unexpected since these cells did not consume oxygen in the microrespirometry experiments when provided with methane as the sole substrate ( Table 2 ). Expression of some genes encoding MMO in the absence of methane was observed previously ( 19 , 98 ). The presence of these proteins in all of the conditions tested might be an adaptive regulatory trait for rapid induction of methane oxidation in response to dynamic fluxes of methane in wetland environments. The reductase (MmoC), regulatory proteins (MmoD and MmoB), and transcription factor (MmoR), which are all necessary for sMMO activity, were more than threefold increased ( P < 0.05) in the methane-grown cells compared with thiosulfate- and ethanol-grown cells, potentially explaining why methane oxidation was observed only in methane-grown cells ( Table 2 ). MDH XoxF5 (MHY1_02202; 3.9 to 7.0% of proteins) and cytochrome c 553i , XoxG5 (MHY1_00490; 0.12 to 0.27% of proteins) were constitutively expressed on all growth substrates. Proteins involved in H 4 MPT-mediated formaldehyde oxidation to formate, aldehyde oxidation, and propionate metabolism were also constitutively expressed ( Dataset S2 ). These results are consistent with the observed oxygen consumption in the presence of various primary alcohols by cells grown on all the three substrates ( Table 2 ) and might be associated with preferential oxidation of methanol over methane (or thiosulfate), as observed in other methanotrophs possessing only sMMO ( 11 , 17 , 99 ). Among chemotaxis receptor proteins, an aerotaxis receptor (Aer) (MHY1_p00042) was more than 2.6-fold ( P < 0.05) more abundant in methane-grown cells than in other conditions. The increased abundance of flagellin and structural proteins of flagella formation in the methane-grown cells compared to thiosulfate- and ethanol-grown cells indicates that motility is important during growth on methane to allow cells to locate optimal conditions within steep methane and O 2 gradients. The abundances of proteins predicted to be involved in sulfur oxidation in strain HY1, such as the Sox and rDsr systems, DoxAD/TetH, SorAB, and sulfhydrogenase, were all greatly increased in thiosulfate-grown cells compared with methane- and ethanol-grown cells ( Fig. 4 and Dataset S2 ). For example, the abundance of key Sox (e.g., SoxYZ) and rDsr (e.g., DsrEFH) proteins was >31-fold ( P < 0.05) higher in thiosulfate-grown than in methane- or ethanol-grown cells. Selective occurrence of DoxAD/TetH in thiosulfate-grown cells indicates the operation and relevance of the S 4 I pathway as expected from the microrespirometry experiments ( Table 2 ). The abundances of two types of cytochromes, a homolog of cytochrome c 2 (MHY1_02786) and the cytochrome c 552 (CycM; MHY1_01465), increased in thiosulfate-grown cells compared with methane-grown cells (more than threefold, P < 0.05), indicating that they may transport electrons from sulfur oxidation to the terminal oxidase. Notably, the subunits of high-affinity cbb 3 -type cytochrome c oxidase were >2.7-fold more abundant ( P < 0.05) in thiosulfate-grown cells than in cells grown on other substrates ( Dataset S2 ), although cells were grown under oxygen-replete conditions (>15%, vol/vol). Sulfur oxidation in strain HY1 might be facilitated in microaerobic oxic–anoxic transition zones. The expression of key enzymes of the CBB cycle (CbbL and CbbS) significantly increased (>4.8-fold, P < 0.05) in thiosulfate-grown cells compared to methane-grown or ethanol-grown cells ( Dataset S2 ). This result supports the coupled induction of sulfur oxidation and the CBB cycle during chemolithoautotrophic growth of strain HY1. An N-type ATPase-encoding operon was found in addition to the other F-type ATPase-encoding operons in strain HY1 ( Dataset S2 ). Biochemical evidence indicates that the c-subunits of Burkholderia pseudomallei N-type ATPase predominantly bind H + and are involved in pumping protons ( 100 ). Thr65, Met66, and Tyr69 in the C-terminal helix, which are likely to contribute to the hydrogen-bonding network around the proton-binding site, were conserved in the c-subunit sequence of strain HY1 N-type ATPase ( SI Appendix , Fig. S18 ). In addition, we speculate that the N-type ATPase of strain HY1 may function as an ATP-driven proton pump for H + homeostasis to survive in acidic stress caused by sulfur oxidation, as observed previously ( 100 ). In support of this notion, the N-type ATPase operon is selectively expressed in the cells grown on thiosulfate compared to methane ( Dataset S2 ). Ecological Relevance. Wetlands are periodically or permanently water-saturated soil environments with a water table at or close to the soil surface. Consequently, steep gradients in soil-redox conditions are developed by a complex pattern of biogeochemical cycling of elements. While the environmental activity of strain HY1 remains unknown, its ability for complete sulfur oxidation could be critical for replenishing the sulfate pool and sustaining sulfate reduction in wetlands. When methane production is repressed by the presence of an active sulfur-redox cycle, thiotrophic capacity could greatly benefit methanotrophs that harbor only sMMO, an enzyme with a low methane affinity ( SI Appendix , Fig. S12 ), by allowing them to switch between methane and sulfur oxidation, depending on which substrate is more available ( 12 , 83 ). Due to the changes in groundwater tables, wetlands frequently shift in the dominance of sulfate reduction over methanogenesis and vice versa ( 101 , 102 ). As shown in Table 2 , activities of mixotrophically grown cells of strain HY1 to sulfur species and methane were lower than those of thiosulfate- or methane-grown cells of strain HY1. Concordantly, the growth rate of strain HY1 simultaneously utilizing methane and thiosulfate was not higher than those utilizing thiosulfate or methane as a single substrate ( Fig. 2 and SI Appendix , Fig. S4 ). This growth pattern is well-described in microorganisms simultaneously utilizing multiple substrates ( 13 , 84 ). In this context, metabolically flexible thiotrophic methanotrophs may outcompete strict methanotrophs or thiotrophs only in highly fluctuating environments, where they have access to a greater pool of net substrates."
} | 10,490 |
36975833 | PMC10100796 | pmc | 3,080 | {
"abstract": "ABSTRACT Anaerobic oxidation of methane (AOM) coupled with reduction of metal oxides is supposed to be a globally important bioprocess in marine sediments. However, the responsible microorganisms and their contributions to methane budget are not clear in deep sea cold seep sediments. Here, we combined geochemistry, muti-omics, and numerical modeling to study metal-dependent AOM in methanic cold seep sediments in the northern continental slope of the South China Sea. Geochemical data based on methane concentrations, carbon stable isotope, solid-phase sediment analysis, and pore water measurements indicate the occurrence of anaerobic methane oxidation coupled to metal oxides reduction in the methanic zone. The 16S rRNA gene and transcript amplicons, along with metagenomic and metatranscriptomic data suggest that diverse anaerobic methanotrophic archaea (ANME) groups actively mediated methane oxidation in the methanic zone either independently or in syntrophy with, e.g., ETH-SRB1, as potential metal reducers. Modeling results suggest that the estimated rates of methane consumption via Fe-AOM and Mn-AOM were both 0.3 μmol cm −2 year −1 , which account for ~3% of total CH 4 removal in sediments. Overall, our results highlight metal-driven anaerobic oxidation of methane as an important methane sink in methanic cold seep sediments. IMPORTANCE Anaerobic oxidation of methane (AOM) coupled with reduction of metal oxides is supposed to be a globally important bioprocess in marine sediments. However, the responsible microorganisms and their contributions to methane budget are not clear in deep sea cold seep sediments. Our findings provide a comprehensive view of metal-dependent AOM in the methanic cold seep sediments and uncovered the potential mechanisms for involved microorganisms. High amounts of buried reactive Fe(III)/Mn(IV) minerals could be an important available electron acceptors for AOM. It is estimated that metal-AOM at least contributes 3% of total methane consumption from methanic sediments to the seep. Therefore, this research paper advances our understanding of the role of metal reduction to the global carbon cycle, especially the methane sink.",
"conclusion": "Conclusions. Methane oxidation occuring in the methanic zone driven by sedimentary microbial communities is an important mechanism that controls natural emissions of methane from the gas hydrate-bearing area. It happens mainly due to the presence of alternative electron acceptors other than sulfate to react with methane. Abundant Fe/Mn-(oxyhydr)oxides preserved in the shelf sediments might be migrated into the study region due to the rapid increase of anomalous subsidence toward the deep water areas in the Qiongdongnan basins ( Fig. 5 ). Therefore, high amounts of buried reactive Fe(III)/Mn(IV) minerals seem to be important available electron acceptors for AOM in the methanic zone of the Haima methane seep, accompanied by the generation of highly alkaline, extremely δ 13 C DIC -depleted and Fe(II)/Mn(II)-enriched pore waters, abundant Fe-Mn carbonates, along with authigenic magnetite by microbial iron/manganese reduction. In methanic sediments, abundant active ANME groups (ANME-1 and ANME-2c) and potential dissimilatory iron reducers (e.g., ETH-SRB1) are potentially involved in metal-AOM in situ . Mechanistically, the apparent ability of ANME-2c to oxidize methane via the release of single electrons in this study should also be able to respire solid electron acceptors directly via extracellular metal reduction, which would confirm the presence of previously reported methane oxidation coupled to insoluble Fe(III)/Mn(IV) reduction. It is estimated that metal-AOM at least contributes 3% CH 4 removal from methanic sediments to the seep. Overall, metal-AOM could significantly impact the biogeochemical cycles in consuming CH 4 in modern marine seep sediments. FIG 5 Simplified scenario for how buried reactive Fe(III)/Mn(IV) minerals offer electron acceptors for AOM in methanic sediments under the high seepage flux of methane. At high seepage activities, methane gas and fluids move along migration pathways from deep sediments to the seabed in cold seeps with gas hydrate reservoirs. When high amounts of buried reactive Fe/Mn(oxyhydr)oxides from the slope sediments are exposed to methanic environments, diverse ANME groups actively mediated methane oxidation coupled to insoluble Fe(III)/Mn(IV) reduction either independently or in syntrophy with metal reducers. It results in authigenic Fe/Mn mineral (carbonate Fe/Mn and magnetite) precipitation and extremely δ 13 C DIC -depleted and Fe(II)/Mn(II)-enriched pore waters in the methanic zone. EET, extracellular electron transfer.",
"introduction": "INTRODUCTION Marine cold seeps are not only an indicator of gas hydrate reservoirs but also an important methane source to the oceans, which has a significant impact on the global carbon cycle and climate change ( 1 ). It is estimated that 0.02 gigaton (Gt) methane is consumed annually in the sediment with an additional 0.02 Gt methane releasing annually into the overlying ocean by seafloor cold seeps ( 2 ). Thus, the production and consumption of methane are key components of the carbon cycle for cold seeps ( 3 ). Anaerobic oxidation of methane (AOM) driven by microbial communities plays a key role in decreasing methane emissions to the atmosphere ( 4 ). Anaerobic methanotrophic archaea (ANME) mediate this process through the coupling of methane oxidation to the reduction of nitrite, nitrate, manganese/iron oxides, and sulfate ( 5 – 8 ). Sulfate-driven AOM (S-AOM) by assemblages of ANME and sulfate-reducing bacteria (SRB) ( 6 , 9 ) is regarded as the major process for methane sink within cold seep sediments, reaching the highest activities within the sulfate-methane transition zone (SMTZ) ( 10 ). Early studies show that methane oxidation possibly coupled with metal oxidation can still occur at a considerable rate in the methanic zone below the SMTZ when sulfate has been completely depleted or at very low levels ( 11 ). A range of efforts have been undertaken to demonstrate the occurrence of metal oxides-driven AOM (metal-AOM) (including high-valence iron/manganese oxides) ( 5 , 12 , 13 ). This microbial process is mediated by ANME through the reverse methanogenesis pathway, typically in syntrophy with dissimilatory iron/manganese-reducing bacteria ( 5 , 14 ). Investigations of enrichment cultures have also revealed that ANME-2a, ANME-2c, and ANME-2d can perform AOM coupled to the extracellular dissimilatory reduction of iron and manganese oxides independently using, e.g., a unique set of multiheme cytochromes (MHCs) ( 12 , 13 , 15 – 17 ). The activity rates of Fe-AOM are efficiently estimated by incubation experiments or geochemical modeling but rarely for Mn-AOM ( Table 1 ) ( 18 – 26 ). Microbial culture experiments from the Eel River Basin seep have found that manganese oxides can drive AOM as electron acceptors more efficiently than ferrihydrite ( 5 ). However, microorganisms involved in the coupling between AOM and metal reduction in marine environments are still largely unknown, especially for manganese reduction ( 27 ). The contribution of Mn-AOM for methane removal in in situ marine environments is still not fully identified, either. Without in-depth understanding of the role of metal-AOM in the biogeochemical cycle, the contribution of metal reduction to the global carbon cycle, especially the methane sink, is likely to be undervalued ( 28 ). TABLE 1 Summary of the estimated rates of S-AOM, Fe-AOM, and Mn-AOM in sediments from various freshwater and marine environments Ecosystem Environment In situ concn (μmol/L) Model-derived rates (μmol CH 4 cm −3 yr −1 ) Depth-integrated rates (μmol CH 4 cm −2 yr −1 ) The fraction in total CH 4 oxidation Method a Reference/source Fe 2+ Mn 2+ S-AOM Fe-AOM Mn-AOM S-AOM Fe-AOM Mn-AOM S-AOM (%) Fe-AOM (%) Mn-AOM (%) Marine Haima cold seep 148 2,289 0.66 0.02 N.A. b 20.05 0.31 0.32 97 1.5 1.5 Modeling This study Hikurangi margin 184 N.A. 0.48 0.0005 N.A. 3 0.4 N.A. 88 12 N.A. Modeling \n 22 \n Black Sea 800 23 0.07 1.46E−05 N.A. 5.9 0.04 N.A. 99 0.70 N.A. Modeling \n 20 \n Baltic Sea 600 N.A. 0.27 0.0011 N.A. 8.8 2.5 N.A. 78 22 N.A. Modeling \n 21 \n Bothnian Sea 1,830 N.A. N.A. N.A. N.A. 78.7 8 N.A. 90 9 N.A. Modeling \n 23 \n 6.94E−04 1.32 N.A. 58.38 1.63 N.A. 97 3 N.A. 13 CH 4 /Modeling \n 19 \n Jiaolong cold seep 27 N.A. 328.57 13.87 N.A. N.A. N.A. N.A. N.A. N.A. N.A. 14 CH 4 \n 26 \n North Sea 380 40 2.04 0.03 N.A. N.A. N.A. N.A. 98 2 N.A. 14 CH 4 \n 18 \n Eel River Basin seep N.A. N.A. 52 6 14 N.A. N.A. N.A. N.A. N.A. N.A. 13 CH 4 \n 5 \n Freshwater Lake Kinneret 70 N.A. N.A. 1.26 N.A. N.A. N.A. N.A. N.A. N.A. N.A. 13 CH 4 \n 25 \n Dover Bluff salt marsh 30 80 2.41 1.42 0.876 N.A. N.A. N.A. N.A. N.A. N.A. 14 CH 4 \n 24 \n Hammersmith Creek River 500 400 5.66 4.5 1.314 N.A. N.A. N.A. N.A. N.A. N.A. 14 CH 4 a Modeling, geochemical modeling estimates; 13 CH 4 , 13 CH 4 incubations; 14 CH 4 , 14 CH 4 incubations. b N.A., not available. The Haima cold seep was firstly discovered as an active cold seep in the Qiongdongnan basin on the northwest slope of the South China Sea by the dives of a remotely operated vehicle named Haima in 2015 (see Fig. S1a in the supplemental material) ( 29 ). A large number of findings have since been emerging about the biogeochemistry of cold seep carbonates, benthos, and sediments in the Haima cold seep ( 30 – 33 ). Massive amounts of terrigenous metal oxides are supplied into the continental slope of the South China Sea from rivers ( 34 ). Consequently, iron/manganese-containing minerals are part of the major components in the sediments of this region with high-flux methane seeps, rendering it a natural laboratory to investigate the role of metal oxides in the methane cycle. In this study, combining geochemical and microbial analyses of the Haima cold seep sediments, we aimed to (i) reveal the occurrence of metal-AOM in the methanic zone, (ii) identify microorganisms involved in the metal-AOM and their key mechanisms, and (iii) estimate the contribution of removal of methane by Fe/Mn-AOM. Our findings provide insights into the coupling mechanism between iron/manganese reduction and AOM as well as the role of metal-AOM in the biogeochemical cycle.",
"discussion": "RESULTS AND DISCUSSION Geochemical data indicate anaerobic methane oxidation in the methanic zone. Due to sampling difficulties, a 4.3-meter-long piston core at a water depth of 1,375 m, named core HM-S11, was retrieved from the Haima cold seep in the South China Sea, where a gas chimney and bubble plumes were observed, indicative of ongoing seepage activities (Fig. S1b to d). Concentrations of methane (CH 4 ) were likely undervalued due to the sample artifacts of degassing, even though the methane range of the core clearly differed. CH 4 was the dominant seeping hydrocarbon gas (0.13 to 919.57 μM), along with ethane (C 2 H 6 ) being detected (0.07 to 3.96 μM) below 130 centimeters below the seafloor (cmbsf) (see Table S1 in the supplemental material). Sulfate (SO 4 2− ) concentrations experienced a quasilinear decrease ( R 2 = 0.98) at the interval of 130 to 230 cmbsf and were almost depleted below ( Fig. 1a ; see also Table S1). Correspondingly, CH 4 concentrations increased rapidly from 15.74 μM at 170 cmbsf to 781 μM at the 210 cmbsf ( Fig. 1a ; Table S1). Therefore, based on methane and sulfate profiles, our sediment core samples were categorized into three biogeochemical zones ( 35 , 36 ) as follows: sulfate reduction zone (surface 130 cmbsf), sulfate - methane transition zone (~130 to 230 cmbsf), and methanic zone (230 to 430 cmbsf). FIG 1 Geochemical profiles of the sediment core HM-S11 in the Haima seep. (a) Profiles of methane (CH 4 ) and sulfate (SO 4 2− ) contents in porewater. (b) Concentrations of dissolved inorganic carbon (DIC) and stable carbon isotope ratios δ 13 C DIC in porewater. (c) Concentrations of dissolved Fe 2+ and Mn 2+ in porewater. (d) Contents of Fe 2 O 3 T and MnO 2 T in sediments. (e to h) Sequential extraction of iron minerals in sediments. Fe carb , carbonate-associated Fe; Fe ox1 , amorphous iron (oxyhydr)oxides; Fe mag , magnetite Fe; Fe py , pyrite Fe. Zones A, B, and C are suggested as the sulfate reduction zone, the sulfate-methane transition zone, and the methanic zone. In the methanic zone, methane concentrations fluctuated between 53 μM and 920 μM, with a notable decrease from 781 μM at 210 cmbsf to 53 μM at 330 cmbsf ( Fig. 1a ; Table S1). The stable carbon isotope ratios (δ 13 C) of CH 4 became heavier from −68.77‰ to −64.33‰ when CH 4 decreased (Table S1), which is consistent with the preferential use of lighter isotopic values by microbes leading to residual inorganic carbon enriched in δ 13 C-CH 4 ( 25 ). In accordance with this, the dissolved inorganic carbon (DIC) values increased from 250 cmbsf (10.38 mM) to 330 cmbsf (17.27 mM) ( Fig. 1b ), implying the formation of HCO 3 − in the methanic zone ( 37 ), which can be also evidenced by the increase in total alkalinity ( 38 ) from 16.01 mM to 28.94 mM (Table S1). The measured δ 13 C DIC values were maintained at lower than −42.30‰ ( Fig. 1b ) in the methanic zone and were much more 13 C-depleted than that of typical marine organic matters (approximately −20‰) in this sea area ( 39 ). As microbes preferentially use the lighter carbon isotopes (δ 12 C), AOM usually results in 13 C-depleted DIC and slightly heavier 13 C values of the residual CH 4 ( 25 ). In this cold seep site, the observed high concentrations of DIC, extreme 13 C-depletion δ 13 C DIC , and heavier δ 13 C values of residual CH 4 indicate that the DIC increase was probably caused by microbial methane oxidation rather than microbial degradation of other organic matter. Additionally, low concentration profiles of phosphate (PO 4 3− ) and ammonium (NH 4 + ), lower than 41.65 μM and 56.72 μM, respectively (Table S1), also support that organic matter degradation was not the main reason for increased DIC concentrations in these sediment samples ( 40 ). Diverse ANME actively mediated methane oxidation in the methanic zone. Anaerobic methanotrophs are assigned to three distinct clades (ANME-1 with subgroups a and b; ANME-2 with subgroups a, b, c, and d; and ANME-3) within the phylum “Halobacteriota” ( 41 ). To identify potential ANME clades in the methanic zone, we performed DNA and RNA sequencing of sediment samples from amplicons and reconstructed 16S rRNA gene sequences from metagenomes by the phyloFlash pipeline ( 42 ) (see Fig. S2 and Table S2 in the supplemental material). Detailed results of microbial community composition at phylum level were described in the supplemental material. Taxonomy classifications of archaeal 16S rRNA gene amplicons indicate that ANME accounted for 69 to 87% of the whole archaeal community in the methanic zone ( Fig. 2a ; see also Table S3 in the supplemental material). These ANME populations are phylogenetically diverse, including ANME-1a (up to 80% at 330 cmbsf), ANME-1b (up to 24% at 430 cmbsf), ANME-2c (up to 16% at 290 cmbsf), and ANME-3 (up to 68% at 370 cmbsf). Relative abundances of 16S rRNA transcripts suggest that ANME-1a (up to 54% at 330 cmbsf), ANME-3 (up to 84% at 370 cmbsf), ANME-2c (up to 63% at 290 cmbsf), and ANME-1b (up to 19% at 410 cmbsf) were the dominant and active players for AOM occurring in the methanic zone ( Fig. 2a ; Table S3). Taxonomic metagenome profiles show a similar distribution profile as 16S rRNA gene amplicon sequencing, such as ANME-1a (up to 71% at 330 cmbsf), ANME-1b (up to 40% at 430 cmbsf), and ANME-2c (up to 5% at 290 cmbsf), but with less relative abundance of ANME-3 (up to 18% at 370 cmbsf) ( Fig. 2a ; Table S3). FIG 2 Gene- and genome-resolved view of the dominant bacteria and archaea. (a) Distribution of genus-level ANMEs and SRBs at different sediment depths in the sediment core HM-S11. (Top) Relative abundance based on 16S rRNA gene amplicons. (Middle) Relative abundance based on 16S rRNA transcript amplicons. (Bottom) Reconstruction of full-length 16S rRNA genes from the metagenomes. ANMEs are shown on the left, and SRBs are shown on the right. (b) Maximum-likelihood phylogenetic tree of nine reconstructed metagenome-assembled genomes with mcrA sequences. Metagenomic assembly and binning yielded 17 metagenome-assembled genomes (MAGs) taxonomically affiliated with methane-metabolizing lineages that either produce or consume methane (see Table S4 in the supplemental material). Among them, nine harbor sequences encoding the catalytic subunit of methyl-coenzyme M reductases (McrA) ( 43 ) involved in methyl reduction during methane oxidation ( Fig. 2b ; see also Table S5 in the supplemental material). They belong to clades of ANME-1 ( n = 3) and ANME-2 ( n = 4). Additionally, S11_12_1 and S11_6_17 belonging to Methanosarcinaceae were predicted to have the capability to perform the methanogenesis pathway. MAGs for ANME-3 lineages were not recovered despite its high relative abundance based on 16S rRNA genes and transcripts (Table S3). Based on read mapping (Table S4), species represented by ANME-1 (i.e., “ Candidatus Methanophagales” [ 44 ]) S11_3_46 (0.4 to 7.7% of the whole microbial community) and S11_6_2 (0.4 to 7.3%) were observed to be the most abundant in the methanic zone, followed by S11_10_20 (0.1 to 1.2%) and S11_8_35 (0.1 to 0.6%) from ANME-2c (“ Candidatus Methanogaster” [ 41 , 43 ]). Metatranscriptomic analyses (see Table S6 in the supplemental material) showed that three ANME-1 genomes (Co_S11_862, S11_3_46, and S11_6_2) highly expressed mcrA genes in the methanic zone (up to 4,382 transcript per million [TPM] at 410 cmbsf). The transcripts of mcrB genes from ANME-1 (Co_S11_862) and ANME 2c (S11_12_8) genomes also had high expression levels (up to 2,378 TPM at 410 cmbsf). These results further suggest that ANME populations were actively responsible for the observed anaerobic methane oxidation in the methanic zone. Methane oxidation is coupled to metal oxide reduction in the methanic zone. In the methanic zone, dissolved ferrous iron (Fe 2+ ) and manganese (Mn 2+ ) concentrations in pore water were found to reach up to 148 μM at 370 cmbsf and as high as 2,289 μM at 340 cmbsf, respectively ( Fig. 1c ; Table S1). The Spearman correlation ( Fig. 3a ) results further show that Fe 2+ and Mn 2+ concentrations have a strong positive covariance with CH 4 (ρ = 0.874 and 0.699, respectively) and DIC (ρ = 0.891 and 0.818, respectively); Fe 2+ has a strong negative relationship with δ 13 C DIC (ρ = −0.655; P < 0.05). These data indicate that the high amounts of dissolved Fe 2+ and Mn 2+ are associated with the fluctuation of methane concentrations in the methanic zone ( 15 , 45 ). Correspondingly, the solid-phase sediment analysis revealed richer Fe 2 O 3 T (3.17 to 3.74%) and MnO 2 T (0.06%) in the methanic zone than those of the SMTZ (Fe 2 O 3 T , 2.31 to 3.60%; MnO 2 T , 0.04%) ( Fig. 1d ; see Table S7 in the supplemental material). The reactive iron minerals, including carbonate-associated iron (Fe carb ) (up to 1.16%), amorphous iron (oxyhydr)oxides (Fe ox1 ) (up to 1.19%), and magnetite iron (Fe mag ) (up to 0.61%) were detected with the higher contents in the methanic zone ( Fig. 1e to g ; see Table S8 in the supplemental material). Therefore, these data implied sufficient supplies of reactive Fe-oxides and the occurrence of Fe authigenic minerals (carbonate Fe/Mn and magnetite) as the products of iron reduction ( 46 , 47 ). Similar to that of iron, total manganese (MnO 2 T ) is also elevated from 0.04% in the SMTZ to 0.06% in the methanic zone ( Fig. 1d ; Table S7). Given the elevated MnO 2 T and extremely high dissolved Mn 2+ (up to 2,289 μM), the contribution of manganese reduction to AOM cannot be ignored in this seep. Overall, porewater and solid-phase profiles support metal-driven methane oxidation in methanic sediments. FIG 3 Spearman correlations of sediments in the core HM-S11. (a) Spearman correlation coefficients between depth-wise distribution of geochemical parameters. (b) Correlation between geochemical parameters and abundances of MAGs belong to ANMEs and metal reduction bacteria. The stars symbolize P values of correlation. ***, P < 0.001; **, P < 0.01; *, P < 0.05. Potential microorganisms involved in dissimilatory metal reduction. For in-depth understanding of Fe(III)/Mn(IV)-dependent AOM in Haima cold seep, it is critical to identify the indigenous microorganisms responsible for this process. Members in different ANME clades are suggested to mediate metal-driven AOM by extracellular electron transfer (EET) to Mn(IV)/Fe(III) (oxyhydr)oxides or metal-reducing partners. In iron-reducer Geobacter sulfurreducens , the process of EET is carried out via MHCs during metal reduction ( 27 , 48 ). For ANME-2d from freshwater Fe-AOM enrichment, a set of MHCs for extracellular dissimilatory Fe(III) reduction were highly expressed ( 12 , 49 , 50 ). Here, all analyzed ANME genomes were found to contain the genes encoding several c-type and periplasmic cytochromes (see Table S9 in the supplemental material). Among all ANME genomes, three MAGs, S11_2_24, S11_5_37 and S11_6_25, belonging to ANME-2c, also encode S-layer-associated multiheme c-type cytochromes, implying a role of ANME-2c archaea with an S-layer protein in conducting electron derived from reverse methanogenesis shuttling from the archaeal membrane to the outside of the cell ( 50 ). S11_12_8 and S11_6_25 affiliated with the family ANME-2c also encode outer membrane cytochrome Z ( omcZ ) gene (see Table S10 in the supplemental material), which plays an important role in Fe(III) reduction ( 51 ). Furthermore, S11_12_8 not only actively expressed at zone C with the maximum of MAG’s abundance and TPM values with mcrB gene but also had a significantly positive relation with CH 4 (ρ = 0.790; P < 0.01), Fe 2+ (ρ = 0.734; P < 0.01), and Mn 2+ (ρ = 0.601; P < 0.05) ( Fig. 3b ). To identify potential dissimilatory metal reducers in the methanic zone, we performed sequencing of bacterial 16S rRNA gene and transcript amplicons, and the 16S rRNA gene from metagenome (Fig. S2 and Table S2). The typical partner SRB of ANME-1, i.e., family Desulfobacteraceae clustering into the SEEP-SRB1 (seep-endemic sulfate-reducing bacteria) clade ( 52 ), were much more abundant in the methanic zone where sulfate was depleted (7 to 26% in DNA libraries, 9 to 42% in RNA libraries, and 7 to 35% in metagenomic libraries) than other zones ( Fig. 2a ; Table S3). Additionally, 2 to 4% in DNA libraries, 2 to 12% in RNA libraries, and 1 to 4% in metagenomic libraries of bacterial sequences in the sulfate zone and SMTZ were identified as Desulfatiglans (family Desulfarculaceae ), which is another common SRB associated with ANMEs in methane seep environments ( 38 ). Besides, the SEEP-SRB2 clade was detected with 2 to 4% in DNA libraries, <0.7% in metagenomic libraries, but 22 to 33% in RNA sequence libraries in the SMTZ ( Fig. 2a ), implying the metabolic activity of SEEP-SRB2 involved in Sulfate-AOM ( 52 ). Therefore, according to 16S rRNA gene and transcript amplicons, and 16S rRNA metagenomes in sediment samples, members of the SEEP-SRB1 clade were the dominant and active bacteria in the methanic zone, potentially involved in dissimilatory metal reduction in situ . Based on the metabolic pathways with metal reduction (Table S6), gene encoding metal (iron/manganese) reduction enzymes, such as decaheme c-type cytochrome ( mtrC ), were present in S11_6_22 and Co_S11_566 affiliated with ETH-SRB1 (ethane-dependent sulfate-reducing bacteria) from the order Desulfobacterales , which were identified as the marine SEEP-SRB1 group of Desulfosarcina -affiliated sulfate-reducing Deltaproteobacteria ( 53 ). The two MAGs (S11_6_22 and Co_S11_566) have the higher abundance in the methanic zone (mean, 0.20% and 0.12%) than the SMTZ (mean, 0.18% and 0.07%) (Table S4). Besides, Spearman’s correlation results ( Fig. 3b ) show that Co_S11_566 closely related with concentrations of Fe 2+ (ρ = 0. 699) and Mn 2+ (ρ = 0. 650). Consistent with the results of microbial communities based on 16S rRNA gene and 16S rRNA transcripts, metagenomic and metatranscriptomic evidence indicated ETH-SRB1 (identified as the marine SEEP-SRB1 group) probably act as the role of metal reducing bacteria in the methanic zone. We also found the presence of hypothetical proteins attributed to porins, cytochrome c binding motif sites (CxxCH), and Geobacter -related gene markers ( omc ) for iron reduction in Co_S11_933 (Table S10), belonging to Zixibacteria , which was reported with pathways of either oxidation or reduction of ferric/ferrous iron and arsenate/arenite and nitrate/nitrite ( 54 ). Co_S11_933 also displayed a higher abundance in the methanic zone (0.03 to 0.06%) than in other zones (Table S4). Contribution of metal-AOM to methane consumption. Geochemical observations and microbiological analyses support that Fe and Mn oxides reduction is coupled to methane oxidation in the methanic zone. We then used reactive transport numerical modeling to predict their contributions to methane consumption. Sensitivity tests of the model results suggest that the modeled profiles are insensitive to the changes of sedimentation rates (see Fig. S3 in the supplemental material). This is because the major porewater profiles in the methane seeps are controlled by methane supply and AOM rather than particulate organic carbon degradation. Constrained by the measured porewater data and Fe leaching experiments ( Fig. 1 ; Tables S1, S7, and S8), the results of the reaction-transport modeling predict the model-derived rates for Fe-AOM of up ~0.02 μmol CH 4 cm −3 year −1 in the methanic zone ( Fig. 4 and Table 1 ). Our estimated Fe-AOM rate is lower than those derived from stimulated microbial communities in laboratory incubations with the sufficient supply of substrates (CH 4 and Fe oxides) ( Table 1 ) ( 5 , 24 , 25 ). Despite that, it is more than 20 times as big as the estimated potential Fe-AOM rates by kinetic modeling from in situ marine methanic sediments with a much higher Fe 2+ concentration (approximately 180 to 800 μM) ( 20 – 23 ) ( Table 1 ) because of much higher total AOM rates related to intense methane bubbling in the Haima seep. FIG 4 Modeled reaction rate profiles of S-AOM (green) (a) and Fe-AOM (blue) (b) according to reaction-transport modeling. As Mn speciation data were not available, we used the diffusive Mn 2+ flux calculated based on the quasilinear concentration gradient at the depth interval of ~250 cm and 350 cm to represent the rate of Mn-AOM. Based on the porewater profiles, our estimated diffusive flux for Mn 2+ is 1.276 μmol cm −2 year −1 . Thus, taking into account that only one molecule of CH 4 is needed to reduce four molecules of MnO 2 , CH 4 removal by Mn-AOM is estimated to be 0.319 μmol CH 4 cm −2 year −1 [ 19 ]. This is a minimum estimate as the potential Mn 2+ consumption by authigenic minerals is not taken into account. Depth-integrated rates of Fe-AOM and Mn-AOM are both 0.3 μmol cm −2 year −1 in the methanic zone, which are considerably lower than the S-AOM rate (~20 μmol cm −2 year −1 ) and account for ~1.5% of total CH 4 removal by microbial metabolism, respectively ( Table 1 ). The high S-AOM rate is caused by methane bubble dissolution, while its upward-ascending and enhanced sulfate supply from seawater is due to bubble irrigation. The estimated depth-integrated rate of Fe-AOM in the Haima seep falls within the range of those reported in different environments globally ( 20 – 23 ). These data from the Haima cold seep provide the first in situ evidence for quantitatively significant manganese-dependent AOM in marine sediments. Given an apparent elevated sedimentary manganese content in the methanic zone (from 0.04% to 0.06%) and high concentration of dissolved Mn 2+ (up to 2,289 μM), the contribution for Mn-AOM consumed by authigenic minerals could have been underestimated. Conclusions. Methane oxidation occuring in the methanic zone driven by sedimentary microbial communities is an important mechanism that controls natural emissions of methane from the gas hydrate-bearing area. It happens mainly due to the presence of alternative electron acceptors other than sulfate to react with methane. Abundant Fe/Mn-(oxyhydr)oxides preserved in the shelf sediments might be migrated into the study region due to the rapid increase of anomalous subsidence toward the deep water areas in the Qiongdongnan basins ( Fig. 5 ). Therefore, high amounts of buried reactive Fe(III)/Mn(IV) minerals seem to be important available electron acceptors for AOM in the methanic zone of the Haima methane seep, accompanied by the generation of highly alkaline, extremely δ 13 C DIC -depleted and Fe(II)/Mn(II)-enriched pore waters, abundant Fe-Mn carbonates, along with authigenic magnetite by microbial iron/manganese reduction. In methanic sediments, abundant active ANME groups (ANME-1 and ANME-2c) and potential dissimilatory iron reducers (e.g., ETH-SRB1) are potentially involved in metal-AOM in situ . Mechanistically, the apparent ability of ANME-2c to oxidize methane via the release of single electrons in this study should also be able to respire solid electron acceptors directly via extracellular metal reduction, which would confirm the presence of previously reported methane oxidation coupled to insoluble Fe(III)/Mn(IV) reduction. It is estimated that metal-AOM at least contributes 3% CH 4 removal from methanic sediments to the seep. Overall, metal-AOM could significantly impact the biogeochemical cycles in consuming CH 4 in modern marine seep sediments. FIG 5 Simplified scenario for how buried reactive Fe(III)/Mn(IV) minerals offer electron acceptors for AOM in methanic sediments under the high seepage flux of methane. At high seepage activities, methane gas and fluids move along migration pathways from deep sediments to the seabed in cold seeps with gas hydrate reservoirs. When high amounts of buried reactive Fe/Mn(oxyhydr)oxides from the slope sediments are exposed to methanic environments, diverse ANME groups actively mediated methane oxidation coupled to insoluble Fe(III)/Mn(IV) reduction either independently or in syntrophy with metal reducers. It results in authigenic Fe/Mn mineral (carbonate Fe/Mn and magnetite) precipitation and extremely δ 13 C DIC -depleted and Fe(II)/Mn(II)-enriched pore waters in the methanic zone. EET, extracellular electron transfer."
} | 7,677 |
27199908 | PMC4850331 | pmc | 3,081 | {
"abstract": "Marine methane seep habitats represent an important control on the global flux of methane. Nucleotide-based meta-omics studies outline community-wide metabolic potential, but expression patterns of environmentally relevant proteins are poorly characterized. Proteomic stable isotope probing (proteomic SIP) provides additional information by characterizing phylogenetically specific, functionally relevant activity in mixed microbial communities, offering enhanced detection through system-wide product integration. Here we applied proteomic SIP to 15 NH 4 + and CH 4 amended seep sediment microcosms in an attempt to track protein synthesis of slow-growing, low-energy microbial systems. Across all samples, 3495 unique proteins were identified, 11% of which were 15 N-labeled. Consistent with the dominant anaerobic oxidation of methane (AOM) activity commonly observed in anoxic seep sediments, proteins associated with sulfate reduction and reverse methanogenesis—including the ANME-2 associated methylenetetrahydromethanopterin reductase (Mer)—were all observed to be actively synthesized ( 15 N-enriched). Conversely, proteins affiliated with putative aerobic sulfur-oxidizing epsilon- and gammaproteobacteria showed a marked decrease over time in our anoxic sediment incubations. The abundance and phylogenetic range of 15 N-enriched methyl-coenzyme M reductase (Mcr) orthologs, many of which exhibited novel post-translational modifications, suggests that seep sediments provide niches for multiple organisms performing analogous metabolisms. In addition, 26 proteins of unknown function were consistently detected and actively expressed under conditions supporting AOM, suggesting that they play important roles in methane seep ecosystems. Stable isotope probing in environmental proteomics experiments provides a mechanism to determine protein durability and evaluate lineage-specific responses in complex microbial communities placed under environmentally relevant conditions. Our work here demonstrates the active synthesis of a metabolically specific minority of enzymes, revealing the surprising longevity of most proteins over the course of an extended incubation experiment in an established, slow-growing, methane-impacted environmental system.",
"conclusion": "Conclusion This study represents one of the most comprehensive SIP proteomics investigations of a complex environmental milieu carried out to date, revealing protein synthesis in slow-growing methane seep sediment communities with a high degree of functional and phylogenetic detail. The detection of multiple functionally relevant orthologs provides a broad sense of in situ active metabolic pathways, while the incorporation of stable isotope probing methods reveals the subset of proteins actively produced under laboratory-based methanotrophic conditions. Components of all of the proteins involved in the reverse methanogenesis pathway were identified, and Mcr was shown to account for 10.4% of all detected peptides. The prevalence of 15 N enriched orthologs involved in reverse methanogenesis and sulfate reducing pathways bolsters our understanding that AOM is the dominant biogeochemical process in seep-simulating incubation conditions, indicating that metabolic activity need not scale with community composition as determined by 16S rRNA gene relative abundances. MS-based PTM identification is well-suited to reveal a wider functional diversity than that encoded by nucleic acid sequences alone, a potentially common reality with significant metabolic repercussions. We were able to detect several enriched, pervasive proteins that lack functional annotations, highlighting the utility of proteomic SIP as a discovery platform for ecologically important proteins that should be prioritized for subsequent biochemical characterization. Finally, proteomic SIP can inform more focused proteomic investigations to quantify enzymes of pathways of interest; such values could then be integrated into flux balance models of carbon, sulfur, and nitrogen cycles to better constrain the dynamics and rates of reactions at both the consortium and ecosystem scales. Similar analyses of methanogenic habitats or cultures would provide additional information on the differences between methanogenic and methanotrophic pathways, potentially revealing activity-based controls on methane-linked metabolism. Tracking the metabolic activity of energy-limited microbial communities has long challenged analytical techniques that address specific microscale anabolic reactions, system-wide metabolic potential, or broad biogeochemical transformations. Proteomic SIP enables the functionally-, phylogenetically-, and temporally-constrained understanding of protein synthesis, opening access to a previously untapped array of distributed low-growth microorganisms active in selected catabolic pathways. Developing a more nuanced appreciation of metabolite flow and interspecies interaction in this way is crucial for the improved understanding and management of microbial ecosystems and their global impacts.",
"introduction": "Introduction Marine methane seep sediments harbor complex microbial communities that play significant roles in the global carbon and sulfur biogeochemical cycles (Jørgensen and Kasten, 2006 ; Reeburgh, 2007 ). Studies of these features have predominately focused on biodiversity (e.g., Bidle et al., 1999 ; Knittel et al., 2005 ; Pernthaler et al., 2008 ), the energetic and biochemical basis of syntrophic partnerships (Alperin and Hoehler, 2009 ; Stams and Plugge, 2009 ), and ecosystem-wide contributions to global methane processing (Reeburgh, 2007 ). Such studies revealed that one of the dominant metabolisms at seep sites—the anaerobic oxidation of methane (AOM)—is enacted by consortia of anaerobic methanotrophs (ANME) most closely related to methanogens and sulfate-reducing deltaproteobacteria. Phylogenetically-linked measures of anabolic activity and growth of AOM consortia using stable isotope probing have identified intricately coupled microbial metabolisms (Krüger et al., 2008a ; Dekas et al., 2009 ; Orphan et al., 2009 ), while metagenomics and metatranscriptomics studies have pointed to metabolic potential and associated biochemical pathways (Hallam et al., 2004 ; Pernthaler et al., 2008 ; Meyerdierks et al., 2010 ; Stokke et al., 2012 ; Wang et al., 2014 ). Proteomic stable isotope probing (SIP) couples these experimental objectives by offering a functionally- and phylogenetically-constrained enzymatic profile of constituent organisms as well as a temporal reporter of protein synthesis and metabolic response to distinct conditions (Pan et al., 2011 ; Seifert et al., 2012 ; Justice et al., 2014 ; Mohr et al., 2014 ). In environments such as anoxic methane seep sediment and an array of subsurface habitats, where energy can become limiting, microbes frequently exhibit extremely slow growth rates and are particularly recalcitrant to activity-based analyses. Radiotracers are sensitive probes to rates on the order of 10 −19 –10 −16 mol cell −1 day −1 (Parkes et al., 1990 ), but the range of discoverable metabolic reactions is severely constrained. Fluorescence in situ hybridization (FISH) coupled with nanoscale secondary ion mass spectrometry (nanoSIMS; e.g., Wagner, 2009 ) can detect assimilation rates as low as 10 −17 mol cell −1 day −1 (Morono et al., 2014 ) and visualize phylogenetically constrained spatial associations, though experimental throughput is low and only assimilatory processes can be queried. Whole-cell bioorthogonal non-canonical amino acid tagging (BONCAT) coupled with FISH can be used to fluorescently visualize microorganisms active in protein synthesis (Hatzenpichler et al., 2014 ), but identification of specific newly synthesized proteins has only been successfully applied in a few cases (Babin et al., 2016 ) and requires further development for complex environmental systems. Proteomic SIP represents an important entry in the analysis of metabolic activity in low-energy microbial systems, due to its spatially broad, yet functionally- and phylogenetically-specific search space. The procedure is able to identify particular metabolic pathways or enzyme-mediated responses (Bozinovski et al., 2012 ; Justice et al., 2014 ) that can be integrated across constituents of a particular lineage, offering an opportunity to access a segment of the low-activity biosphere that might go undetected by other methods due to low levels of anabolism by individual organisms. Although challenges remain—particularly surrounding protein extraction, peptide quantification, database collation, and the interpretation of non-detections—proteomic SIP offers a promising method for assessing the in vivo activity and catalytic function of microbial communities. Culture-independent studies of energy-limited methane seep settings have included meta-omics investigations focused largely on the pathway responsible for AOM. These investigations have revealed that AOM likely utilizes the same enzymes as methanogenesis, operating in the reverse direction (Hallam et al., 2004 ; Meyerdierks et al., 2010 ). ANME-1 draft genomes and fosmids, however, lack the mer gene (Meyerdierks et al., 2010 ; Stokke et al., 2012 ), prompting the proposition of a reverse-methanogenesis bypass (Meyerdierks et al., 2010 ). ANME-2 lineages, including ANME-2a (Wang et al., 2014 ), and nitrate-reducing ANME-2d (Haroon et al., 2013 ) genomes, as well as a magnetic enrichment of ANME-2c consortia (Pernthaler et al., 2008 ), contained the mer gene. Corresponding gene expression profiles revealed ANME-1 methylenetetrahydromethanopterin dehydrogenase ( mtd ), heterodisulfide reductase subunits A and B ( hdrAB ), and methyl-coenzyme M reductase subunit A ( mcrA ) transcripts (Meyerdierks et al., 2010 ). ANME-2a and ANME-2d transcriptomic datasets exhibited highly expressed methyl-tetrahydromethanopterin coenzyme M methyltransferase ( mtr ), mcr, mer , and methenyltetrahydromethanopterin cyclohydrolase ( mch ) and substantially lower levels of mtd and several formylmethanofuran dehydrogenase ( fmd ) subunits (Haroon et al., 2013 ; Wang et al., 2014 ). Proteomic SIP in environmental samples enables the identification of isotopically enriched proteins synthesized by a microbial assemblage after incubation with an isotopically labeled substrate (e.g., 15 N-ammonium). The application of proteomic SIP to slow growing methane seep microbial communities offers an opportunity to examine production patterns of reverse methanogenesis enzymes central to ANME-facilitated methane-oxidation, as well as to identify the proteins expressed by syntrophic bacterial partners and other organisms within the seep ecosystem. This approach can highlight metabolic shifts resulting from incubation conditions, reveal the fraction of proteins that are newly synthesized, and potentially facilitate the construction of a framework for interspecies metabolic relationships (Jehmlich et al., 2008 ; Hawley et al., 2014 ). The greatest insights from environmental metaproteomic studies to date have been associated with time-resolved analyses of low-complexity microbial systems (i.e., acid mine drainage microbial biofilms; Denef et al., 2010 ; Pan et al., 2011 ), while similar efforts with sponge symbionts (Liu et al., 2012 ), sinking marine particulates (Moore et al., 2012 ), and the human gut (Verberkmoes et al., 2009 ; Xiong et al., 2015 ) have clarified interspecies relationships and dominant biogeochemical functionalities. There have been comparatively few studies applying proteomics to deep-sea methane seep ecosystems; these have focused on individual enzymes (Krüger et al., 2003 ), explored specific biochemical questions (Glass et al., 2014 ), or achieved relatively low coverage of the full ecosystem (Stokke et al., 2012 ). Here we use proteomics and proteomic SIP to examine multiple methane seep sediment incubations containing diverse ANME and bacterial lineages and exhibiting a range of methanotrophic activity levels.",
"discussion": "Results and discussion Activity and community composition in seep sediment microcosms Samples of methane seep sediment from Hydrate Ridge (hereafter referred to as #3731 and #5133) were each allocated into two anaerobic microcosm incubations overpressured with methane and amended with either 14 NH 4 + (unlabeled treatment) or 15 NH 4 + (labeled treatment; Figure 1 ). Cell aggregate abundance and sulfide concentrations, both of which correlate positively with sulfate-based AOM activity (Iversen and Jorgensen, 1985 ), were monitored throughout the incubation period. Both #3731 and #5133 were collected from actively seeping areas, yet the #3731 incubation, recovered from the lower 6–12 cm sediment horizon, demonstrated indications of lower metabolic activity (Table 2 ) and 15 N incorporation (see below). Potential reasons for this discrepancy include differing in situ methane concentrations despite general markers of activity, localized heterogeneity of methanotrophic constituents (Table 2 ), and sediment horizon depth-related variability in AOM activity. Table 2 Data on incubation activity, as measured by sulfide production and cell abundance . Days After Incubation Set-Up #3731 #5133 #3731 \n 14 N #3731 \n 15 N #5133 \n 14 N #5133 \n 15 N Aggregate Counts (per mL) Sulfide Concentration (mM) Aggregate Counts (per mL) Sulfide Concentration (mM) Aggregate Counts (per mL) Sulfide Concentration (mM) Aggregate Counts (per mL) Sulfide Concentration (mM) 0 3.4 × 10 6 0.9 3.4 × 10 6 0.9 2.4 × 10 7 2.4 × 10 7 6 1.2 1 17 1.4 2 20 3.3 3.8 65 2.2 1.6 12.8 13 160 3.4 × 10 7 14.3 4.1 × 10 7 17.9 326 2.7 × 10 6 4.2 7.9 × 10 5 3 Microbial communities were analyzed using 16S rDNA iTAG sequencing to compare the changes in diversity between the two time points of #3731 and between sediment #3731 and #5133 (Supplemental Data Sheet 3a). Across all three samples, the most abundant representatives were Deltaproteobacteria from the uncultivated SEEP-SRB1 (accounting for 7.2% of sequences on average, ± 1.3% SD) and SEEP-SRB2 (5.1 ± 1.7% SD) clades (Schreiber et al., 2010 ), Sulfurovum (6.7 ± 2.8% SD), and ANME-1a (6 ± 2.6% SD) and ANME-1b (5 ± 1.3% SD). ANME-2 subclades accounted for an average of 3.2% of sequences (±0.56% SD); however, the primer set used has been shown to preferentially detect ANME-1 over ANME-2 (Case et al., 2015 ), and independent FISH-based microscopy indicates a prevalence of ANME-2 consortia in the initial inoculum of both samples (data not shown). In all three samples, ANME sequences represented a substantial proportion of the archaeal relative abundance (67.8% on average, ± 6.3% SD). AOM-related enzymes of unspecified lineage putatively come from ANME organisms given that methanogenic lineages comprised < 0.1% of the archaeal community across all samples at the order level, as determined by 16S rRNA gene tag sequencing (Supplemental Data Sheet 3b). Protein identifications The metaproteomic dataset produced here, using optimized extraction techniques and Orbitrap mass spectrometers, represents the deepest proteomic measurement of methane seep sediments to date and rivals most other environmental proteomics studies. Analyses at the 1% FDR level yielded 5664 unique proteins across all samples under the 1TP, 1UP condition, and 3495 unique proteins under the 2TP, 1UP condition (see Table 3 for quantification details). Compared with previous environmental studies (Supplemental Presentation 1; Table S1), the large number of proteins and selective incorporation of stable isotope label demonstrated here indicates that increased microbial diversity, low overall activity levels, and challenging physicochemical conditions are not insurmountable challenges in SIP-resolved metaproteomic studies. Nonetheless, the quantitative recovery, detection, and characterization of all proteins is not technically feasible because of protein adsorption to silicate minerals (Ding and Henrichs, 2002 ), complexation (Nguyen and Harvey, 2001 ) as well as the difficulty of solubilizing membrane-bound constituents (Tan et al., 2008 ). Table 3 The number of proteins identified in this study based on distinct identification criteria . Sample Protein Identifications 2TP 1UP 1TP 1UP Hydrate Ridge Seep Sediment #3731 14 N T 17d 1562 2648 Hydrate Ridge Seep Sediment #3731 15 N T 17d 1695 2741 Hydrate Ridge Seep Sediment #3731 14 N T 326d 847 1477 Hydrate Ridge Seep Sediment #3731 15 N T 326d 820 1484 Hydrate Ridge Seep Sediment #5133 14 N T 160d 1850 3027 Hydrate Ridge Seep Sediment #5133 15 N T 160d 1179 2260 Unique Proteins from All Samples 3495 5664 2TP 1UP = two total peptides, one unique peptide required for positive identification; 1TP 1UP = one total peptide, one unique peptide required for positive identification . A core set of 456 (1TP, 1UP) and 283 (2TP, 1UP) proteins were detected in all six samples; of these, the five most common categories were “proteins of unknown function” with best matches to hypothetical genes (comprising 31.1% of protein identifications averaged between the search conditions), adenylyl sulfate reductase subunit A (AprA, 7.4%), McrB (5.7%), McrG (4.9%), and ATPase subunit A (4.6%). Phylogenetic assignments of the cumulative metaproteome demonstrated a higher relative abundance of Bacteria-classified proteins (56.9% 1TP, 1UP; 58.7% 2TP, 1UP) than those assigned to Archaea (37.2% 1TP, 1UP; 35.2% 2TP, 1UP). 15 N enrichment of proteins expressed in AOM sediment incubations To provide additional context for 15 N incorporation-based activity levels, the isotopic composition of bulk protein extract and intact ANME-SRB consortia was measured by nanoscale Secondary Ion Mass Spectrometry (nanoSIMS). NanoSIMS analysis of #3731 sediment bulk protein extracts from the 15 NH 4 + amended incubation revealed a slow-growing community with a low but measurable enrichment at the early (0.45 atom % enrichment for 15 N T 17 d ) and late (0.53 atom % at 15 N T 326 d ) time points compared to the 14 NH 4 + treated controls (0.38 atom % at both time points; Table 4 ). In silico identification of enriched peptides from sample #3731 also exhibited a small but detectable isotopic enrichment of the protein pool during the 326-day incubation. The 15 N atom % enrichment of the entire #3731 15 N peptide pool based on Sipros analysis was 0.4% at T 17 d , increasing to 0.8% at T 326 d (while corresponding values for 14 N control incubations were 0.5 and 0.4%, respectively). Table 4 15 N enrichment values of proteome-derived peptides and bulk protein extract . Sample ID Incubation Time (d) Peptide LC MS/MS Analysis (SE) Bulk Protein Extract nanoSIMS Analysis 15 N Atom % Mean % of Enriched Peptides (>5% 15 N) 15 N Atom % Mean (SE) #3731 14 N T 17d 17 0.5 (0.03) 0.44 (0.03) 0.38 (0) #3731 15 N T 17d 17 0.4 (0.01) 0.52 (0.07) 0.45 (5x10-5) #3731 14 N T 326d 326 0.4 (0.09) 0.62 (0.07) 0.38 (0) #3731 15 N T 326d 326 0.8 (0.04) 1.48 (0.04) 0.53 (5x10-5) #5133 14 N T 160d 160 0.5 (< 0.01) 0.87 (0.03) NA #5133 15 N T 160d 160 9.5 (18.34) 18.06 (0.68) NA Peptides recovered from sample #5133 15 N after 160 days of incubation with 15 NH 4 + demonstrated significantly more uptake of the 15 N label, with Sipros-calculated isotopic enrichment of 9.5 atom % 15 N, while peptides from the corresponding 14 NH 4 + control exhibited a mean enrichment of 0.5%. NanoSIMS analysis of ANME-SRB consortia from the same sample showed substantially higher 15 N incorporation levels, reaching 23.8% after 64 days (Supplemental Presentation 1; Figure S1). The broader peptide-based enrichment value may provide a more representative constraint on community-wide anabolic rates because ANME-SRB consortia constitute a particularly active subset of the community (Knittel and Boetius, 2009 ). Peptides that uniquely mapped to ANME archaea had a mean 15 N value of 13.8%, while those attributed to the dominant deltaproteobacterial sulfate-reducing group ( Desulfobacterales ) had a mean enrichment value of 10.1%. Proteins previously linked to the reverse methanogenesis pathway had a mean enrichment of 18.3%, whereas proteins associated with dissimilatory sulfate reduction exhibited a mean enrichment of 9.3%. These values suggest that ANME lineages responded to the methane incubation conditions with heightened biosynthetic activity compared to SRB, and that a higher proportion of reverse methanogenesis enzymes were newly synthesized compared to other identified ANME proteins. The observation of enhanced 15 N incorporation in the nanoSIMS analysis of individual intact AOM consortia relative to the calculated 15 N values from bulk ANME and SRB-associated proteins recovered from the same incubation is not fully understood. Potential explanations include the incomplete detection of ANME and deltaproteobacterial proteins, enrichment discrepancies among organisms or different cellular nitrogen sources, or variation in the turnover of proteins and other cellular nitrogen stores (Reitzer, 2003 ). Membrane proteins are believed to comprise 20–30% of prokaryotic proteomes based on genomic analysis (Stevens and Arkin, 2000 ), yet solubilizing membranes and accessing these proteins is a technical challenge (Trötschel and Poetsch, 2015 ); preferential incorporation of 15 N into membrane proteins, such as receptors, transporters, or oxidoreductases may be partially responsible for the 15 N unaccounted for in our SIP metaproteomic analysis. Peptide enrichment values exhibited a marked bimodal distribution, with an unenriched peak and a 15 N enriched peak at 49% for the #5133 15 N T 160d sample and 64% for the #3731 15 N T 326 d sample (Figure 2 ). The presence of two distinct peptide pools in the data presented here implies that most of the detected proteins (made up of peptides with 15 N values below 5%) were likely synthesized before the addition of 15 NH 4 + and were not fully degraded, while the remainder was newly synthesized. The lack of peptides with intermediate enrichment values argues against multiple protein precursor pools (Justice et al., 2014 ) or substantial heterotrophic turnover of the system's primary producers during the course of the 160-day and 326-day incubations. Figure 2 15 N enrichment distributions for all peptides identified in (A) #5133 15 N T 160 d and (B) #3731 15 N T 326 d incubations . In both cases, the precursor pool of NH 4 + was 496 μM and 1 mM of 15 NH 4 + was added at the beginning of the incubation period. Functional and phylogenetic distributions of enriched proteins Of the proteins identified from the #5133 15 N sample, 32% ( n = 377) were isotopically enriched using the conservative (2TP, 1UP) criterion. To determine which functional traits or phylogenetic affiliations may be disproportionately represented in either the enriched or unenriched protein pools, annotated proteins were partitioned by enzyme type or phylogenetic assignment of the corresponding ORF. The relative abundance of these categories was calculated for both the enriched and unenriched proteins, and the variance contributed by each bin was determined (calculated by the square of the difference in relative abundances). Eighteen functional protein categories accounted for 79.8% of the variance between enriched and unenriched proteins, with each contributing at least 1% of the variance (Figure 3A ). The largest contributor, whose distribution produced 24.8% of the observed variance, was RNA polymerase (RNAP), which was more prominent in the unenriched protein pool. By examining the phylogenetic assignments of RNAP orthologs, the enzyme serves as a representative signal of organism-wide regulation. The detection of remnant, unenriched RNAP proteins reveals their resistance to degradation over the course of the 160-day experiment and offers the possibility that all transcription needs were met by pre-existing, long-lived copies. The relative lack of newly produced copies could indicate that the incubation conditions were deleterious to active transcription and, perhaps, survival of some microorganisms; this scenario is consistent with the detection of few enriched proteins from similar phylogenetic groupings, suggesting a broader shutdown of transcriptional machinery. In this context, the lineages responding most negatively to the anaerobic incubation conditions include Methylococcales, Sulfurovum, Chromatiales, Alteromonadales , and Pseudomonas (Figure 3A ), indicating a shift away from putative oxygen- or nitrate-based metabolisms common within the oxidizing conditions of upper seep sediment layers. Figure 3 (A) Protein functional groups accounting for at least 1% of variance between the enriched and unenriched #5133 15 N T 160d protein pools. Positive values reveal protein types that are more abundant in the enriched fraction; negative values indicate those more prevalent in the unenriched fraction. Inset table shows the phylogenetic affiliations of unenriched RNAP and enriched GroEL proteins. (B) Phylogenetic associations accounting for at least 1% of variance between the enriched and unenriched #5133 15 N T 160d protein pools. Positive values reveal phylogenetic assignments whose protein products are more abundant in the enriched fraction; negative values indicate those more prevalent in the unenriched fraction. Phylogenetic assignments were made at the family level; higher level assignments are provided if no family-level specificity was available, and genus-level assignments are provided if no other genera were observed in the same family. Archaeal Gzfosmids are ANME-affiliated sequences reported by Hallam et al. ( 2004 ). Of the predominantly enriched protein types in #5133, chaperonin GroEL accounts for the most variance (10.9%) between the enriched and unenriched pools, followed by F-type ATPase subunits B (8.4%), and A (6.1%). GroEL is frequently detected in environmental metaproteomic studies (Benndorf et al., 2007 ; Sowell et al., 2009 ; Verberkmoes et al., 2009 ) and has been demonstrated to prevent amorphous protein aggregations (Hartl et al., 2011 ), facilitate proper folding of a wide variety of enzymes (Kerner et al., 2005 ), and mediate stress response (Kim et al., 2013 ); changing pressure or energetic regimes in our incubations may have prompted such a response. F-type ATPase and pyrophosphatase PpaC (1.4%) are both involved in ATP generation, concordant with previous observations of heightened energy production under conditions of physiological stress (Alexandre et al., 2001 ) such as those potentially experienced by slow-growing methanotrophic consortia (Valentine, 2007 ). Several high-variance proteins are involved in reverse methanogenesis and sulfate reduction pathways. AprA (5.1%), McrB (3.0%), McrG (2.8%), sulfite reductase subunit B (DsrB; 2.4%), McrA (2.1%), MtrA (1.7%), AprB (1.4%), and sulfate adenylyltransferase (Sat, 1.3%) were all more prevalent in the 15 N enriched fraction. This observation is consistent with active sulfate-coupled AOM as demonstrated by sulfide production during the incubation as well as quantitative increases in and growth of ANME-SRB consortia (Table 2 , Supplemental Presentation 1, Figure S1). Of the high-variance proteins that were detected at higher levels within the unenriched fraction, methyl-accepting chemotaxis proteins (accounting for 1.6% of the variance) may be used in biofilm-linked cell signaling or orientation toward favorable environmental conditions (Morgan et al., 2006 ). Most of the observed chemotaxis proteins exhibit greatest homology with Gammaproteobacteria , and their near-absence among 15 N-enriched proteins suggests decreased need for mobility-linked sensing and/or less communication among these constituents. Glutamine synthetase (3.0%) catalyzes the formation of glutamine from glutamate and ammonia, and its minimal presence among enriched proteins could indicate the accessibility of free glutamine (and likely other amino acids) from biomass degraded during the incubation, or a heightened reaction rate—facilitated by increased NH 4 + concentrations—that obviated new enzyme production. Predominantly unenriched proteins may reveal functions of lower importance under incubation conditions, while reflecting remnant capabilities that highlight the importance of proteins' ability to outlive upstream markers of activity (e.g., transcripts (Moran et al., 2013 ). Fifteen phylogenetic bins each contributed at least 1% of the variance between the relative abundances of enriched and unenriched proteins recovered from #5133 15 N, representing 92.5% of the total variance (Figure 3B ). Among the more notable differences was the prevalence of enriched proteins linked to methane-metabolizing lineages (accounting for 32.3% of the variance) and Deltaproteobacteria (27.7%), particularly orders with known sulfate reducers (18.3%) expressing proteins from carbon fixation and sulfate reduction pathways, GroEL, and ribosomal proteins. Proteins attributed to Epsilonproteobacteria (20.8%) and especially Sulfurovum RNAP, Apr, and ATPase (15.3%) were markedly less prevalent in the enriched pool. The dearth of enriched epsilonproteobacterial protein products is consistent with these organisms' proposed susceptibility to changing conditions (Alain et al., 2004 ; Sievert et al., 2007 ; Toner et al., 2012 ) as well the potentially unfavorable conditions for supporting active oxygen- or nitrate-coupled sulfide oxidation. The high relative abundance of Sulfurovum in the #5133 tag sequencing data (Supplemental Data Sheet 3a) highlights the utility of SIP proteomics: the abundance and persistence of 16S rRNA genes and proteins may not always be reflective of biosynthetically active participants in an environment. Finally, of the proteins that remained unannotated (i.e., proteins of unknown function using the 2TP, 1UP criterion), 26 were enriched in the #5133 15 N T 160d sample and were present in all six proteomes, suggesting they are important in seep sediment habitats and are actively synthesized under conditions supporting AOM. Approximately half of the unknown proteins were attributed to uncultured Archaea while most of the remainder was associated with sulfate-reducing Deltaproteobacteria (Table 5 ). Up to 40% of expressed bacterial proteins currently lack functional classification, and there is a growing need for further categorization and characterization of these proteins (Galperin and Koonin, 2004 ). These 26 consistently detected, uncharacterized proteins represent notable targets for functional biochemical study (Guengerich and Cheng, 2011 ; Goodacre et al., 2014 ). Table 5 UniProt results for the 26 proteins that were detected in all six proteomes, enriched in #5133 15 N T 160 d , and were unannotated via KAAS and UniProt . UniProt Annotation Phylogenetic Assignment UniProt ID % Identity E- value 3datasets_contig_19395_3 Putative uncharacterized protein Uncultured archaeon D1J9F4 ** 51.7 2 E-64 5datasets_contig_137335_2 Putative uncharacterized protein Uncultured archaeon D1J9F4 ** 90.5 1 E-45 5datasets_contig_214909_1 Putative uncharacterized protein Uncultured archaeon D1J9F4 ** 79.1 2 E-47 5datasets_contig_354507_3 Putative uncharacterized protein Uncultured archaeon D1J9F4 ** 88.6 1 E-38 5datasets_contig_426791_2 Putative uncharacterized protein Uncultured archaeon D1J9F4 ** 61.2 2 E-64 VO1_contig_3180_1 Putative uncharacterized protein Uncultured archaeon D1JGR9 ** 78.5 4 E-42 3datasets_contig_52223_1 Hypothetical secreted protein * Uncultured archaeon D1JHB5 ** 42.9 2 E-61 3datasets_contig_128337_2 Hypothetical secreted protein * Uncultured archaeon D1JHB5 ** 38.1 3 E-52 5datasets_contig_120634_2 Hypothetical secreted protein * Uncultured archaeon D1JHB5 ** 46.6 3 E-55 3730_contig_6723_4 Putative uncharacterized protein Uncultured archaeon D1JHC5 ** 77 5 E-35 5datasets_contig_264150_1 Putative uncharacterized protein Uncultured archaeon Q2Y4L6 ** 57.5 3 E-62 3730_contig_4042_9 Uncharacterized protein Uncultured archaeon Q6MZD7 ** 52.2 4 E-44 3datasets_contig_80833_1 Uncharacterized protein Uncultured archaeon Q6MZD7 ** 54.1 4 E-42 VO1_contig_16363_1 Uncharacterized protein Uncultured archaeon Q6MZD7 ** 52.6 2 E-41 5datasets_contig_14875_4 Uncharacterized protein Desulfococcus oleovorans A8ZRQ4 36.1 5 E-67 3730_contig_10399_1 Uncharacterized protein Desulfonatronospira thiodismutans D6SNE1 62.3 6 E-42 3730_contig_109916_2 Uncharacterized protein Desulfonatronospira thiodismutans D6SNE1 63.8 4 E-44 5datasets_contig_13906_2 Putative uncharacterized protein Uncultured Desulfobacterium sp. E1YFV9 62.6 0 E+00 5datasets_contig_110039_2 Putative uncharacterized protein Desulfobulbus propionicus E8RH53 62.7 5 E-49 3730_contig_26432_2 Hypothetical protein Desulfatirhabdium butyrativorans UPI0003FB5DCE 45.7 5 E-101 5datasets_contig_120508_2 Hypothetical protein Desulfobulbus japonicus UPI0003FCAF71 54.9 2 E-32 5datasets_contig_26951_1 Hypothetical protein Desulfosarcina sp. BuS5 UPI0004837EE9 55.8 0 E+00 5datasets_contig_175666_3 Hypothetical protein Thermodesulfatator atlanticus UPI0003B3E199 40.4 2 E-39 5datasets_contig_93287_2 Putative uncharacterized protein Moorea producens F4XS28 47.4 7 E-131 5datasets_contig_80150_1 Hypothetical protein Longispora albida UPI000381E695 43.8 5 E-105 5datasets_contig_244181_2 Uncharacterized protein * Uncultured bacterium K1YAD4 51.4 1 E-114 * Contains an InterPro-predicted transmembrante domain . ** Best-match ORFs derived from Meyerdierks et al. ( 2010 ) . Metabolic pathways Methane metabolism The anaerobic oxidation of methane (AOM) is a dominant metabolism at Hydrate Ridge methane seeps. Many distinct orthologs of all enzymes involved in the reverse methanogenesis pathway were recovered from incubated seep sediment (Figure 4 ; phylogenetic affiliations provided in Supplemental Data Sheet 4a), signifying unprecedented completeness and diversity of pathway protein detection. Methyl-coenzyme M reductase (Mcr) is one of the most abundant enzymes in ANME, accounting for 7% of total extractable proteins from the ANME-1 rich Black Sea mats (Krüger et al., 2003 ). Across all six of our proteomic data sets, 10.4% of peptides were attributable to Mcr; 34.6% of 15 N enriched peptides in sample #5133 were associated with this critical enzyme. Considerable diversity of detected Mcr proteins was observed, with approximately 20 orthologs of each Mcr subunit identified across all six samples, representing 47% of the mcr protein sequences within our seep metagenome database. The #5133 microcosm thus appeared to stably maintain a genetically diverse population of active ANME archaea over the course of the extended incubation. Of the expressed alpha, beta, and gamma subunits of MCR, one-third was affiliated with ANME-1 (31% of detected orthologs, see Supplemental Data Sheet 4a), with smaller contribution of ANME-2c (19%), Methanosarcinaceae (14%), and ANME-2a derived (12%) orthologs. The three non-catalytic Mcr components (McrA2, McrC, McrD) were also identified, but these putative (methanogenic) activation domains (Prakash et al., 2014 ) were notably less prevalent in the metaproteome. This result may indicate McrA2, McrC, and McrD re-use as activating components, or their decreased relevance under methanotrophic conditions, potentially reflective of a methane-oxidizing enzymatic mechanism that is distinct from a strict reversal of methane formation. Figure 4 Metaproteomic data for enzymes involved in the reverse methanogenesis pathway . Filled boxes indicate an enzyme constituent's presence in the associated sample's proteome. Due to space constraints, Hdr detections are not shown, but relevant information is provided as Protein orthologsdetected(enrichedorthologs)∕orthologsinmetagenomicdatabase. Several previous ANME-targeted meta-omics efforts have sought to reconstruct the reverse methanogenesis pathway (Table 6 ; e.g., Hallam et al., 2004 ; Meyerdierks et al., 2010 ; Stokke et al., 2012 ; Haroon et al., 2013 ; Wang et al., 2014 ), but the Mer protein in particular has proven challenging to identify (Hallam et al., 2004 ; Meyerdierks et al., 2010 ; Stokke et al., 2012 ). In our proteomic experiments, seven Mer orthologs were identified (six of which were 15 N-enriched), exhibiting best homology to ANME-2a, M. burtonii , and other Methanosarcinales lineages. Our detection of 15 N-enriched Mer orthologs under incubation conditions favoring AOM is consistent with a reverse methanogenesis metabolism. It is possible that ANME-1 methane oxidation proceeds via the mer bypass (Meyerdierks et al., 2010 ), whereby activated methane is converted first to methyl-containing intermediates and then to formaldehyde, which is processed to methylene-H 4 MPT by a formaldehyde-activating enzyme (Fae) and hexulose-6-phosphate (Hps). The viability of the methanogenic version of this pathway has been demonstrated with an M. barkeri mer deletion mutant (Welander and Metcalf, 2008 ). An ANME-1 mer copy has not yet been identified, and 12 orthologs of the Fae/Hps fusion protein were detected across all samples (1TP, 1UP). Table 6 Central reverse methanogenesis enzymes identified in metagenomic, metatranscriptomic, and metaproteomic analyses . Green, identified; orange, not addressed; red, not identified. ANME proteins newly detected in seep environments during this study are highlighted in bold text. (McrABG subunits associated with ANME-2c derived fosmids were also detected; no other ANME-2c genes related to reverse methanogenesis were in the metagenomic database.) Building upon the 356 proteins detected from seep sediments in the Nyegga area (Norwegian Sea; Stokke et al., 2012 ), additional subunits of enzymes putatively involved in the reverse methanogenesis pathway were newly detected in our study, including McrC and McrD (Table 6 ). Of the components not found in the proteome, some of those present in the co-located metagenomes (MtrBCDEF, HdrDE) are membrane-bound proteins (Thauer et al., 2008 ) and thus are expected to be more difficult to quantitatively recover (Trötschel and Poetsch, 2015 ). Given the relatively poor genomic coverage of the ANME-2 lineages, combined with the observed dominance of ANME sequences relative to methanogens in our iTAG diversity surveys (Supplemental Data Sheet 3b), it is likely that most of the reverse methanogenesis pathway orthologs of poor homology derive from ANME representatives. Mcra post translational modifications Mcr is a critical enzyme in the reverse methanogenesis pathway, activating methane with a disulfide bond and initiating the multi-step anaerobic oxidation of methane (Krüger et al., 2003 ; Shima and Thauer, 2005 ). To better understand the range of variability found between McrA orthologs and potentially distinguish between methanogenic and methanotrophic forms of the enzyme, we conducted a computational search for selected post-translational modifications (PTMs; Camacho et al., 2009 ). MS-based PTM analysis is a developing field, and it is difficult to draw functional conclusions from detected modifications; nonetheless, the approach has been used to corroborate modifications (Selmer et al., 2000 ) and as a discovery tool supplemented by subsequent crystallographic confirmation (Moellering and Cravatt, 2013 ). Spectra from sample #5133 14 N T 160d were used due to its enhanced metabolic activity relative to #3731 (Table 2 ) and unlabeled peptides, which ensured the accuracy of PTM mass windows. PTMs were observed among 10 of the 18 McrA orthologs, 13 of 19 McrB orthologs, and 7 of 17 McrG orthologs; none were identified on five McrA2, one McrC, and three McrD orthologs. Twenty-eight distinct modification events of McrA were observed, including mono-, di-, and tri-methylations, hydroxylation, beta-methylthiolation, and acetylation (Figure 5 ). PTMs were detected with high frequency (occurring an average of once every 84 amino acids among ANME Mcr subunits), but only 11% were affiliated with multiple orthologs. This observation, in addition to their positioning away from active site residues, suggests a dominance of PTMs responsive to environmental variation (Olsen et al., 2010 ) that may be involved in enzyme regulation (Olsen and Mann, 2013 ). Figure 5 A phylogenetic and PTM-based analysis of the McrA subunit . The left panel shows a Muscle-aligned phylogenetic tree of all McrA sequences in the metagenomic database, as well as additional cultured organisms to account for all seven methanogenic orders and other ANME representatives. The metaproteomic detection panel shows the presence/absence bars from Figure 4 , positioned next to their corresponding ORF label. The active site/PTM analysis panel provides alignments generated by Jalview for selected amino acid positions of note that together account for all PTMs observed in this and previous studies, as well as residues implicated in the enzyme's active site. Amino acids detected in proteomics analyses are highlighted in gray; colored residues exhibited PTMs specified in the figure key. PTMs observed in previous studies (Elias and Gygi, 2007 ; Sharma et al., 2012 ; Wang et al., 2013 ) are portrayed by color-coding of amino acid position numbers at the top of the alignment; boxed numbers correspond with active site residues described in Sharma et al. ( 2012 ). ORF coverage was calculated by dividing the number of amino acids in an ORF's detected peptides by that ORF's full length. All PTM analysis only includes data from #5133 14 N T 160d . The frequent occurrence and differential expression of PTMs among McrA orthologs is a discovery that suggests modifications are common and are phylogenetically and/or environmentally heterogeneous. Functional traits can be substantially affected by the expression of such modifications (Moellering and Cravatt, 2013 ) and they represent potentially fruitful targets for future work aimed at distinguishing enzymatic modulation or differentiating methanogenic from methanotrophic Mcr in vivo . Sulfur metabolism The dissimilatory sulfate reduction pathway was well-represented in the metaproteomes from all six seep sediment incubation samples (Figure 6 ). In total, 17 Sat, 37 AprA, 13 AprB, 11 DsrA, and 18 DsrB proteins were recovered, accounting for 28% of the potential metagenomic search space among these enzyme subunits. Previous protein-based examinations of sulfate reduction in methane seep environments have reported a complete sulfate reduction pathway within the ANME-1 dominated “reefs” in the Black Sea (Basen et al., 2011 ) and have localized AprB (Wrede et al., 2012 ), Sat, and Dsr enzymes (Milucka et al., 2013 ) to ANME-affiliated sulfate SRB via antibody detection. Sulfate reduction pathway proteins were also recovered from Nyegga seep sediments, with the exception of DsrB (Stokke et al., 2012 ). Figure 6 Metaproteomic data for enzymes involved in the sulfate reduction pathway . For key, see Figure 4 . Phylogenetic affiliations of sulfate reduction pathway orthologs (Sat, AprA, AprB, DsrA, and DsrB) identified across all six proteomic experiments are shown in Supplemental Data Sheet 4b,c. Of those classified to class level, deltaproteobacterial enzymes (84%) dominated, with Desulfobacterales as the most common family (39% of all class-level partitioned orthologs). Based on the 16S rDNA community data, the majority of the deltaproteobacteria were members of the SEEP-SRB1 and SEEP-SRB2 clades within the Desulfobacterales (Supplemental Data Sheet 3a), an observation supported by FISH visualization (Trembath-Reichert et al., 2016 ) and the frequent SEEP-SRB association with ANME (Orphan et al., 2002 ; Knittel et al., 2005 ). Sulfate-coupled oxidation of non-methane hydrocarbons can also play a substantial role in sulfur metabolic processing at hydrocarbon seeps (Bose et al., 2013 ). This process is typically attributed to the DSS clade, with additional contributions from non-DSS Desulfobacteraceae, Desulfobulbaceae, Syntrophobacteraceae , and Desulfurellales (Widdel et al., 2010 ; Kleindienst et al., 2014 ). It is difficult to attribute the presence of these organisms exclusively to higher-order alkane oxidation, but Sat enzymes linked to Syntrophobacteraceae were detected, and all of the above lineages were represented in the 16S rRNA gene tag data (Supplemental Data Sheet 3a). 90 (1TP, 1UP) proteins—89 of which did not correspond with previously annotated proteins—demonstrated closest homology to the Desulfosarcina strain BuS5, the only pure culture SRB known to degrade short chain alkanes (Kniemeyer et al., 2007 ). Seven thiosulfate reductase orthologs attributed to Delta - and Epsilonproteobacteria were detected (2TP, 1UP), suggesting that thiosulfate reduction is a relatively common metabolic capability within seep sediments. This observation is consistent with previous evidence of thiosulfate reduction from ANME-1-rich seep sediment incubations from Eckernförde Bay (Jagersma et al., 2012 ) and thiosulfate disproportionation in ANME-1- and ANME-2-dominant samples (Nauhaus et al., 2005 ). In both cases, the process was seemingly disconnected from AOM, and we observe no evidence of thiosulfate reductase synthesized during our incubations. The connection between AOM-linked sulfate reduction and thiosulfate metabolism remains uncertain, but may impact sulfur cycling in important ways. Enzymes facilitating sulfur oxidizing pathways were present in the co-localized metagenomes, indicating that the sediments possess a widespread potential for sulfur compound oxidation. Tag sequencing data revealed a substantial proportion of the Sulfurovum genus, accounting for between 4 and 9.6% of total 16S rRNA gene sequences (Supplemental Data Sheet 3a). Sulfurovum species are frequently invoked as sulfur-oxidizing constituents in anoxic zones (Sylvan et al., 2012 ; Schunck et al., 2013 ; Urich et al., 2014 ); their high 16S rRNA gene relative abundance, combined with the low recovery of associated sulfur-oxidizing proteins, suggests that DNA signatures of this lineage are retained over the experimental timescale despite the absence of substantial protein-based biosynthesis."
} | 11,609 |
25918393 | PMC4434766 | pmc | 3,082 | {
"abstract": "Significance Microorganisms are key players in emissions of the greenhouse gas (GHG) methane from anoxic carbon-rich peat soils of the Arctic permafrost region. Although available data and modeling suggest a significant temperature-induced increase of GHG emissions from these regions by the end of this century, the controls of and interactions within the underlying microbial networks are largely unknown. This temperature-gradient study of an Arctic peat soil using integrated omics techniques reveals critical temperatures at which microbial adaptations cause changes in metabolic bottlenecks of anaerobic carbon-degradation pathways. In particular taxonomic shifts within functional guilds at different levels of the carbon degradation cascade enable a fast adaptation of the microbial system resulting in high methane emissions at all temperatures.",
"conclusion": "Conclusions By the end of this century, summer temperatures in the Arctic possibly will increase by 1–6 °C, with substantial spatial and temporal variations. Our results showed that within a short period (30 ± 10 d), reflecting the Arctic active summer season, the peat microbiota quickly adapted to large temperature shifts. CH 4 production below 10 °C was higher than in lower latitude ecosystems and increased rapidly even with small increases in temperature. Temperature changes were associated with changes in the microbial community and metabolic network and affected pathways of SOC decomposition (including CH 4 production), metabolic bottlenecks, and the microbial loop. The threshold temperature of 7 °C—a temperature that might be reached in many high-latitude anoxic peat systems periodically within the next century—defined the switch between two different system balances. At temperatures below 7 °C, syntrophic oxidation of propionate was the rate-limiting step for CH 4 production, depending on the efficiency of acetotrophic methanogens. Above 7 °C, hydrolysis of polysaccharides was the rate-limiting step. Temperatures higher than 7 ° C thus might relieve limitations inherent to terminal mineralization, including syntrophic fermentation and methanogenesis. Particularly striking was the plasticity of the microbial community involved in terminal mineralization, because cascade effects of temperature altered substrate availability in terminal steps. This work has pointed out the remarkable ability of anaerobic Arctic peat microbiota to adapt to a wide range of temperatures within a short time, reflecting their response to fluctuating temperatures during the short Arctic summer. The consequence of increased temperatures is a large increase in CH 4 and CO 2 production, which potentially is of crucial importance to the global carbon cycle in a warming world.",
"discussion": "Discussion Low-Temperature Adaptation. We investigated the temperature response of microorganisms responsible for anaerobic degradation of SOC and CH 4 production in Arctic peat soil. The relatively high CH 4 production rate at 4 °C (25% of the rate at 25 °C) implies that the Arctic peat microbiota investigated here is well adapted to low temperature. At lower latitudes, the CH 4 production shows a stronger response to increasing temperature: In temperate peat soils, the rate at 4 °C is <1% of the rate at 25 °C ( 40 ); in temperate lake sediments, the rate is 10% of that at 25 °C ( 41 ); and in sub-Arctic peat soils, the rate is 17% of that at 27 °C ( 22 ). Also methane fluxes in wetland, rice paddy, and aquatic ecosystems located within a broad range of latitudes have a higher temperature sensitivity; the average CH 4 emissions increases by 57-fold from 1 °C to 30 °C ( 42 ). The results show that this methane-producing Arctic ecosystem has a lower temperature sensitivity than temperate/sub-Arctic ecosystems. The proportion of hydrogenotrophic methanogenesis in the Arctic peat soil was high (35%), in contrast to lake sediments ( 41 , 43 – 45 ) and rice soil ( 46 , 47 ), where only a minor proportion or no CH 4 originates from hydrogenotrophic methanogenesis at low temperature. This finding is in accordance with studies of sub-Arctic ( 22 ) and boreal peat soil ( 21 ), where the proportion of hydrogenotrophic methanogenesis is 30 and 80%, respectively. In lake and tundra ecosystems at low temperatures, homoacetogens are important H 2 consumers, and acetate is the major precursor for methanogenesis ( 17 , 44 , 45 , 48 ). However, homoacetogens were not important in this Arctic peat soil, as shown by the low relative abundance of fhs transcripts and the high and constant proportion of hydrogenotrophic methanogenesis at all temperatures. Temperature-Dependent, Rate-Limiting Steps for CH 4 Production. The step in anaerobic SOC decomposition with the lowest rate defines the rate-limiting step for CH 4 production. In rice soils, polysaccharide hydrolysis is the rate-limiting step for CH 4 formation ( 16 , 18 ). In our study, the accumulation of propionate and acetate below 7 °C indicated that terminal processes rather than upstream polysaccharide hydrolysis were rate limiting for CH 4 production. Terminal processes also might be rate limiting in situ, considering that 7 °C is within the range of current Arctic summer soil temperatures, and high concentrations of fermentation intermediates have been measured in summer ( 20 ). In our study, propionate was syntrophically fermented at low temperature by propionate oxidizers within the Firmicutes in association with formate- and H 2 -using methanogens within the Methanobacteriales, thus keeping the formate and H 2 concentrations in the microcosms low. However, the high and variable concentrations of acetate and propionate at low temperature, together with the prevalence of low-affinity acetate-using Methanosarcinaceae over high-affinity Methanosaetaceae indicated that acetotrophic methanogenesis and propionate oxidation were not syntrophically associated below 7 °C. However, the two processes were still connected, as shown by the nearly perfect negative correlation between acetate and propionate concentrations. These findings suggest that below 7 °C the relationship is cyclic, with opposite cycles of activity of propionate oxidizers and acetotrophic methanogens, increasing the Gibbs free-energy change of the reactions of both groups. Because of the high propionate and low acetate concentrations in several microcosms below 7 °C, propionate oxidation was exergonic, with a Gibbs free-energy change around the minimum found to sustain life in propionate-oxidizing bacteria (−6 to 12 kJ/mol) ( 49 ). Above 7 °C, the propionate pool was depleted and was associated with a shift in the functional guild responsible for propionate oxidation, with Firmicutes replaced by Bacteroidetes and Deltaproteobacteria. Calculations showed that between 7 °C and 12 °C, propionate oxidation to acetate, H 2 , and formate was endergonic, whereas the alternative pathway of propionate oxidation to acetate and H 2 was exergonic. This finding suggests that propionate oxidizers might alter their pathways to enable energy conservation depending on the conditions, as discussed previously ( 19 ). The metatranscriptomic indications that at higher temperatures Bacteroidetes oxidize propionate to acetate and H 2 but not to formate support this interpretation. Also, the concentration of acetate decreased with increasing temperature, in line with the increased abundance and activity of high-affinity acetotrophic methanogenic Methanosaetaceae ( 50 ), suggesting that more efficient acetate utilization was associated with more efficient propionate oxidation. A similar association was seen in a triculture of a high-affinity acetotrophic methanogen, a propionate oxidizer, and a formate-using methanogen ( 51 ), where low acetate and formate concentrations increased the efficiency of propionate oxidation over that of dicultures without the acetotrophic methanogen. The same relationship between acetate concentrations and propionate oxidation has been observed in wastewater treatment at low temperatures ( 52 ) and in anaerobic degradation in taiga pond sediment ( 17 ). The removal of the terminal bottleneck above 7 °C resulted in hydrolysis of polysaccharides becoming the rate-limiting step for CH 4 production. Microbial Loop Alters Methanogenesis Pathways. The most pronounced temperature-related taxonomic shift was a tenfold increase in SSU rRNA of the predatory protist phylum Cercozoa ( 53 ). This shift and the overall low abundance of protists indicated a low grazing pressure at in situ temperatures and substantially higher grazing pressure at higher temperatures. The difference in grazing pressures is supported further by increased activity and substrate turnover not resulting in a larger microbial biomass. The increase in Cercozoa correlated with the increased abundance of transcripts for methanogenesis from mono- and dimethylamines. Sources of methylamines include glycine-betaine, sarcosine, and glycine ( 54 , 55 ). Glycine comprises up to 60% of some proteins ( 54 ). Several plant cell-wall structural proteins contain long stretches of glycine ( 56 ), and Gram-positive bacterial cell walls contain pentaglycine bridges that make up a substantial fraction of peptidoglycan ( 57 ). The increase in the relative abundance of transcripts of proteins involved in glycine betaine, sarcosine, and glycine degradation suggested that these compounds are among the primary sources of methylamines in Arctic soils. Glycine from undigested cell walls or amino acids could be released by Cercozoa during grazing on bacteria. Indeed, nitrogen ( 58 ), sometimes as amino acids and peptides ( 59 ), is released during protist grazing on bacteria. Furthermore, the compatible solute glycine-betaine, used as an intracellular osmo- and thermoprotectant in psychrophiles, could have been released during the shift from low to high temperature. However, the relative abundance of transcripts for proteins involved in glycine-betaine reduction was low. Taken together, this circumstantial evidence illustrates how higher-level trophic interactions of the soil microbial loop can indirectly impact the substrate availability for an important guild of GHG-producing microorganisms. The storage compound for excess carbon and energy, PHB, can provide organisms with a crucial advantage under fluctuating carbon and/or limiting nutrient availability ( 60 , 61 ). The high abundance of transcripts for PHB metabolism, mainly assigned to Actinobacteria, suggested that PHB is an important storage compound at low temperature. The decrease with increasing temperature suggested that the advantage of intracellular carbon storage in this taxon is restricted to low temperature. Both methanogenesis from methylamines ( 25 ) and fermentation of glycine ( 54 ) release ammonium (NH 4 + ). Thus, the higher availability of nitrogen might explain the indicated decrease in PHB metabolism. However, other explanations are possible also. Taxonomic and Metabolic Shifts Enable Rapid Thermal Adaptation of Microbiota. A remarkable effect of temperature change was the taxonomic switches within functional guilds and the functional switches within taxa. A taxonomic switch within the functional guild for syntrophic propionate oxidation occurred with increasing temperature, where Firmicutes were replaced by Bacteroidetes and to some extent by Deltaproteobacteria. Other taxonomic switches occurred with increasing temperature: Within the guild for methanogenesis from H 2 /CO 2 and formate, Methanomicrobiales replaced Methanobacteriales; within the guild for acetotrophic methanogenesis, Methanosaetaceae replaced Methanosarcinaceae; Methanosarcinaceae in turn changed their metabolism from acetotrophic methanogenesis to methylotrophic methanogenesis, thus exhibiting remarkable metabolic flexibility during thermal adaptation. Such taxon and metabolic shifts correspond well to the temperature dependence of microbial community metabolism in aquatic ecosystems modeled by Hall et al. ( 62 ). These authors propose that organisms adapt physiologically to be competitive for substrates within specific temperature ranges, and that this adaptation results in changes in their relative contribution to community metabolism. We suggest that similar mechanisms are triggered in the Arctic peat ecosystem studied, indicating that part of the complexity found in microbial communities is attributable to their flexibility across environmental gradients such as temperature. The switch from acetotrophic to methylotrophic methanogenesis of Methanosarcinaceae likely is a direct result of increased substrate availability, possibly from a more pronounced microbial loop, but also might have been triggered by more competitive Methanosaetaceae having a higher affinity to acetate. All observed effects are in the terminal steps of anaerobic decomposition, i.e, syntrophic oxidation of organic acids and methanogenesis, indicating that the temperature effect, which is primarily thermokinetic in upstream metabolism, becomes systemic (via changes in pathways and taxa) in downstream and terminal metabolism. The results agree with those of previous studies showing that thermal adaptation is the switch from cold-adapted taxa to warm-adapted taxa ( 63 ). However, in the tightly connected anaerobic SOC degradation network of this Arctic peat soil, additional mechanisms resulting in altered substrate concentrations are as important as thermal adaptation in shaping the community response to temperature increases. Missing Sinks and Sources. Mass balance calculations indicated that there were missing sinks and sources of carbon and reducing equivalents. The putative sink of acetate can be explained by either assimilation ( 64 , 65 ) or adsorption to solids ( 66 ), whereas the unknown redox kinetics of both solid and dissolved humic matter ( 67 ) might have been a sink for reducing equivalents (H 2 and formate). Among the putative sources of CO 2 are primary fermentation of hexose or pentose sugars, anaerobic respiration, and/or CH 4 oxidation. Because sulfate, nitrate, and nitrite were not detected, alternative electron acceptors must be considered. The CO 2 proportion of mineralized carbon was below 50%, indicating that the carbon source deviated from C 6 H 12 O 6 , as is consistent with the indications for methanogenesis from methylamines originating from the degradation of proteins or fats, both of which would yield a CO 2 proportion of mineralized carbon below 50%. However, the increase in CO 2 proportion with temperature suggested that other processes, such as CH 4 oxidation, might have affected the proportions of terminal products. High numbers of transcripts and SSU rRNA were assigned to CH 4 -oxidizing bacteria within the Methylococcales. The assignment of the majority of the transcripts to M . tundripaludum ( 68 , 69 ) indicated that this CH 4 -oxidizing bacterium might be important not only in oxic soil layers ( 20 , 70 ) but also under anoxic conditions. However, the data generated in this study are not suitable to test this hypothesis."
} | 3,781 |
39125189 | PMC11314048 | pmc | 3,083 | {
"abstract": "This research aims to explore how functionally active structures affect the physical, mechanical, thermal, and fire-resistant properties of elastomeric compositions using ethylene–propylene–diene rubber as a base. The inclusion of aluminosilicate microspheres, microfibers, and a phosphorus–boron–nitrogen–organic modifier in these structures creates a synergistic effect, enhancing the material’s heat-insulating properties by strengthening coke and carbonization processes. This results in a 12–19% increase in heating time for unheated sample surfaces and a 6–17% increase in residual coke compared to existing analogs. Microspheres help counteract the negative impact of microfibers on composition density and thermal conductivity, while the phosphorus–boron–containing modifier allows for controlling the formation of the coke layer.",
"conclusion": "5. Conclusions The findings suggest that incorporating pre-assembled functional–active systems into elastomeric fire-resistant materials improves their interaction with the elastomeric matrix, leading to more effective distribution. This targeted delivery of the modifier to the interfacial layer enhances coking processes, specifically at the interface. By introducing both microspheres and microfibers during coking initiation, structures are formed where microfibers encircle microspheres, strengthening the coke layer and enhancing its erosion resistance.",
"introduction": "1. Introduction The development of heat-resistant materials using elastomers that can endure high-temperature conditions for short durations is a crucial objective for safeguarding structures across diverse industries, including aviation, rocketry, and the oil-and-gas sector [ 1 , 2 ]. These materials find application in coating rocket engine combustion chambers and nozzles, as well as in gas generator casings and other related equipment. Researchers Bhuvaneswari, Walter, and Ahmed have been conducting investigations into the development of fire-resistant polymer materials [ 3 , 4 , 5 ]. Fire-heat composite materials (FHPMs) are subject to a wide array of demanding criteria, often presenting conflicting requirements: improving one parameter may lead to the degradation of other properties. Scientists are confronted with a complex, multifaceted challenge in finding an optimal balance that enables the creation of the most efficient material. The effectiveness of fire-retardant materials hinges on the physicochemical transformations of components, their thermal decomposition, and alterations in the material’s chemical structure [ 6 , 7 ]. Processes involving structuring and the formation of a protective layer with low thermal conductivity are activated under the influence of high temperatures [ 8 , 9 , 10 , 11 ]. In most compositions designed for ablative protection against high temperatures, inorganic additives play a pivotal role by aiding in the formation of carbon coke during material thermal decomposition. To enhance coke formation processes, components containing elements such as halogens, phosphorus, nitrogen, boron, metals, or combinations thereof are introduced [ 12 , 13 , 14 , 15 ]. Recently, researchers have increasingly used mixtures of different flame retardants or substances in which molecules simultaneously contain elements such as phosphorus, boron, and nitrogen to achieve a higher effectiveness in reducing the flammability of materials. The mechanism of action of the coating containing the mentioned compounds is based on the fact that, when exposed to flames, as well as during the polymer material’s destruction and oxidation, phosphorus–boron-containing compounds form polyphosphoric and boron-containing acids. These acids distribute as a film on the material’s surface and hinder the ingress of oxygen necessary to sustain the combustion process. Moreover, a porous glassy coating of polyphosphoric acid, which has low thermal conductivity, develops on the surface of the polymer, thereby decreasing the level of heat that enters the polymer’s interior. Phosphorus–boron-containing compounds promote reactions of cyclization, condensation, and carbonization of the decomposition products during combustion, leading to the formation of a “coke cap”. During the pyrolysis of polymers containing phosphorus compounds, phosphoric acid and its anhydrides are formed, which catalyze dehydration, dehydrogenation, and contribute to carbonization. Phosphorus–organic compounds primarily act in the condensed phase, altering the direction of the decomposition processes and increasing the coke residue, while reducing the amount of gaseous combustible products [ 6 ]. Some types of phosphorus-containing flame retardants decompose to form gaseous compounds, wherein phosphorus facilitates the formation of sooty conglomerates, reducing combustion completeness. The introduction of phosphorus-containing fragments into polymer compositions not only reduces their flammability but often enhances adhesion, corrosion resistance, and other beneficial properties. However, the range of highly effective phosphorus-based flame retardants produced by the global industry is insufficient and can be expanded through the synthesis of new synergistic phosphorus–nitrogen–boron–halogen-containing compounds. Furthermore, the application of fire-protective coatings on composite and elastomeric materials, considering aspects such as adhesion, thermal stability, and the effects of static and dynamic mechanical loads, remains incompletely studied. In challenging operational environments, fire- and heat-resistant materials face not only high temperatures and pressures but also rapid gas flow, resulting in the erosion of the surface layer of ablative thermal protection materials (FHPM), material thinning, and a subsequent decrease in effectiveness. Addressing this issue involves incorporating microfibrous fillers such as kaolin, basalt, carbon fibers, and similar materials to create a reinforcing framework, enhancing the material’s ability to withstand erosion. The introduction of these fillers into polymers leads to diverse interactions at the polymer–filler interface, impacting the composite material’s mechanical properties, physicochemical stability, and thermal resistance [ 16 , 17 , 18 ]. However, the use of these fillers can result in technological problems, such as fiber agglomeration, reduced material homogeneity, and increased density and thermal conductivity. To address these issues, various treatments can be applied to improve the uniform distribution of microfibers and enhance interaction with the polymer. The phosphorus–boron–nitrogen-containing modifier (PBN) has been shown to be an effective system for improving physical and mechanical properties, as well as fire-retardant and thermophysical characteristics of the composite. In References [ 19 , 20 ], chemical processes in flames upon the introduction of phosphorus-containing additives; transformation mechanisms of additives; and mechanisms of the influence of these compounds on the combustion rate, structure, limits of spread of hydrogen, and hydrocarbon flames are considered. Flame inhibition by phosphorus-containing additives was explained by the authors of [ 20 , 21 ] by an increase in the rate of recombination of hydrogen proton and hydroxide anion and reactions with phosphorus oxides and oxyacids. The described processes are characteristic of the gas phase. The action in the condensed phase consists of the fact that, upon the decomposition of the flame retardant, residues of phosphoric acid are formed that act as a dehydrating agent, promoting the formation of carbonized structures. At the same time, aerosols may also form, contributing to the deactivation of radicals through a wall effect. It has been noted in works [ 22 , 23 ] that, in this phase, phosphorus compounds and their decomposition products cause dehydration of the polymeric structure, cyclization, crosslinking, aromatization, and graphitization; in other words, they act as crosslinking agents. Polyphosphoric acids are formed on the surface of the coating. As mentioned in [ 24 ], these compounds can act in both phases. For example, when phosphorus-containing flame retardants are added to polystyrene, the rate of its decomposition in the condensed phase increases (rather than decreases, as one might expect), and triphenylphosphine and sulfur exhibit a synergistic effect in inhibiting the combustion of polystyrene. It should be noted that microencapsulated phosphorus can be used as a modifying agent, and when introduced into epoxy polymers, an increase in their fire resistance is observed, while the physical–mechanical and dielectric properties remain practically unchanged [ 17 ]. The application of a product based on melamine, aminotrimethylenephosphonic acid, and FeCl 3 ·6H 2 O (2d-CFA) as an effective modifier for polymer compositions is well-known [ 25 ]. The “labyrinth” effect of 2d-CFA nanosheets is beneficial for hindering mass and heat transfer during the early pyrolysis of PLA. In Reference [ 26 ], a novel bio-based intumescent flame retardant (PP-Fe) with a nanosheet structure was successfully fabricated using simple self-assembly technology. The resulting material exhibited high mechanical properties and flame retardancy. Modified nanotubes based on aluminosilicates have been applied [ 27 ] to enhance the fire safety and thermostability of polylactic acid polymers.",
"discussion": "3. Discussion The introduced functional–active components, forming relatively large microstructures, lead to a slight decrease in strength indicators, but their values still remain above the normative values ( Table 2 ). More importantly, as the content of microspheres increases, the density of the compositions decreases and tends to align with that of the control sample. The reduction of density is significant in creating FHPMs for aviation and rocket technology [ 28 ]. Samples that contain a ratio of 5 microspheres to 10 microfibers with functional–active structures demonstrate the most effective fire-resistant characteristics. As heat moves through the fire-resistant material, various adaptive processes take place: the upper layers of the material experience degradation of the polymer matrix, while the added functional–active components aid in creating a denser porous coke layer strengthened with microfibers (see Figure 2 ). The addition of phosphorus–boron–nitrogen–organic modifiers on the surface of microspheres in the deeper layers of the material triggers the formation of coke, which leads to a slower heating rate. If we assume that the coke residue is formed solely from the mineral part of the recipe, the introduction of 5–10 wt. To evaluate the performance retention of fire-resistant composite materials containing a combination of modified aluminosilicate microspheres and microfibers at elevated temperatures (75–150 °C), elastic and tensile properties were determined using a Shimadzu AG-X Plus tensile testing machine equipped with a thermal chamber. As shown in Table 3 , the introduction of a combination of microfibers and microspheres in a ratio of 10:5 provides the highest retention of material properties. The manufacturing technology of fire-resistant materials involves a multilayer structure, and strong bonding between the layers is achieved by using an appropriate adhesion promoter. The presence of nitrogen in the synthesized compound suggests that it may exhibit adhesive activity and interact with segments of the film-forming polymer macromolecules. The results of comparative analysis show that the introduction of a modifying additive increases the adhesive strength. The highest adhesive strength is achieved when bonding rubber compounds based on ethylene–propylene rubber with adhesive 88SA, with the addition of 5–10 wt. ( Figure 3 ). Further increasing the content of the modifier does not significantly affect the strength characteristics, which may be associated with a weakening of the diffusive nature of the interaction between the adhesive and the substrate. The synthesized modifier in elastomeric material can also act as a co-agent of adhesion, contributing to the strengthening of bond strength in the adhesive–elastomer system. Its presence improves diffusive processes during bonding. The effectiveness of the investigated additives is confirmed by differential thermal analysis (DTA) and thermogravimetric (TGA) analyses ( Figure 4 ). The amount of energy expended on structuring processes, material coking, modifier decomposition, and its chemical changes due to heat can be assessed by examining the area beneath the endothermic peak on the DTA curve. The introduction of functional–active structures leads to an increase in coke residue by 4–30% and an increase in the area under the endothermic peak by 24%. In Figure 5 , the tested samples after erosion-resistance testing under high-speed heat flow conditions are presented. The control sample ( Figure 5 b) is characterized by a significant mass loss and a short ignition time. The presence of functional–active structures intensifies coke-formation processes, while microfiber contributes to the creation of a strong coke with low thermal conductivity ( Figure 5 e). Based on the experimental findings, we were able to formulate a hypothesis regarding the mechanism through which materials containing functional–active structures provide fire protection ( Figure 6 ). When exposed to a high-temperature flow, the following zones can be distinguished within the cross-section of an FRCM [ 29 ]: 1. Zone of intense thermal destruction of the coke layer in direct contact with the gas flow (temperature ranging from 2500 to 4000 °C): In this zone, thermal destruction of foamed coke occurs, leading to its mineralization, volatilization of inorganic compounds, and carbon itself (above 3700 °C). The gas pyrolysis zone is also affected by erosion caused by the gas flow. 2. Zone of coking and foamed coke formation: In this zone, the polymer undergoing pyrolysis undergoes coking, resulting in the formation of a porous coke structure. The size of the pores in the plastic zone increases, starting from the polymer binder pyrolysis zone at temperatures above 300 °C, or for ethylene–propylene rubbers, above 400 °C. When the coke loses its plasticity at temperatures above 500–600 °C, some micro-cracking and delamination of the foamed coke occur (stratification in microzones). Gas pyrolysis products flow out through the formed pores into the high-temperature zone. 3. Zone of pyrolysis of the polymer binder at temperatures ranging from 300 to 400 °C: In this zone, an approximately 1 mm thick thermal decomposition of the polymer (pyrolysis) takes place. Chemical bonds in high-molecular-weight macromolecules of rubber and other polymers and resins introduced into the FHPM are broken, resulting in the formation of low-molecular-weight products (pyrolysis gases) and a sharp loss of material mass. As viscosity sharply decreases in this zone, pore formation and the development of a porous structure begin. Comparatively small pores are formed here, and they increase in size with temperature until the material retains certain plastic properties (pore size exceeding 0.3 mm). The thickness of the sample sharply increases during pore formation. Deformation processes occurring in the samples intensify pore formation, alter pore shape, and create through-channels. The formation of coke occurs in already-foamed rubber, and its porous structure is largely determined by the process of pore formation in rubber. Additionally, fillers, plasticizers, and modifiers influence the coking process by catalyzing carbonization (acting as centers for crystallization, leading to dehydrogenation reactions, and modifying the ratio of volatile pyrolysis fractions). 4. Pre-pyrolysis layer: Processes of thermodecomposition of weak chemical bonds begin at temperatures above 523 K, and at temperatures around 673 K, pore-formation processes start in partially destructured materials. Simultaneously, materials based on butadiene–nitrile rubbers and other diene polymers undergo structuring and cyclization processes, resulting in material shrinkage (523–573 K), while materials based on chlorinated polyethylene transition to a thermoplastic state within this temperature interval. During this temperature interval, the growth of pores begins, originating from nucleation bubbles on filler particles and microdroplets of plasticizer. Nucleation pores have sizes smaller than 1 μm. The growth of bubbles occurs due to the diffusion of gaseous decomposition products of rubber and dissolved or adsorbed water within them. 5. Zone of material heating from the pore formation zone to the wall. During the initial stages of heating (up to 373–423 K), thermal expansion of the material occurs, along with melting processes of certain crystalline components and the formation of nucleation pores under the influence of volatile components. Depending on the heating conditions, the nature of the rubber, and the coating thickness, local thermal explosions can occur in FHPM, leading to irregular destruction of the porous rubber layer and coke. The impact of the gas flow, thermal deformations in the carbonizing FHPM consisting of layers with different moduli, leads to the cracking of the coke layer and the formation of through-cracks and -channels through which pyrolysis gases escape (the “blowout effect”). The coking process strongly depends on the type of material; in structuring materials, a dense network of the polymer matrix is formed, promoting the creation of dense fine–porous coke. In destructuring materials made of thermoplastic, coarse–porous rubber, loose, weak coke is formed that is easily destroyed by the gas flow. At the same time, the heat-protection function is not only related to coke formation but also to the thermal costs of rubber thermodegradation. In this regard, materials based on EPDM have a significant advantage over materials based on butadiene–nitrile rubbers, where thermal transformations are accompanied by significant effects of structuring and cyclization. The protective effect is also associated with a number of other physicochemical transformations in the sample, such as melting, sublimation, boiling of components, chemical reactions of modifiers, significant influence on processes of pore formation, thermodegradation, pyrolysis, and coking of fillers, plasticizers, vulcanizing agents, and other components of FHPM. Under the influence of high-temperature heat flow, the introduced functional–active components transform into a reinforced coke layer with increased resistance to erosion and reduced thermal conductivity. In this case, microspheres act as coke-formation centers. The process is initiated by the PBN modifier layer on the microspheres’ surface, which captures and holds the microfibers essential for enhancing the strength of coke when subjected to material erosion conditions. Under high-temperature exposure in the coking zone, melting and destruction of microspheres occur, but even on the surface of fragments, microfibers are retained and continue to perform their function of reinforcing coke."
} | 4,797 |
29720635 | null | s2 | 3,084 | {
"abstract": "The formation of condensed matter typically involves a trade-off between structural order and flexibility. As the extent and directionality of interactions between atomic or molecular components increase, materials generally become more ordered but less compliant, and vice versa. Nevertheless, high levels of structural order and flexibility are not necessarily mutually exclusive; there are many biological (such as microtubules"
} | 107 |
24357897 | null | s2 | 3,085 | {
"abstract": "Pneumatically actuated, non-elastomeric membrane valves fabricated from polymerized polyethylene glycol diacrylate (poly-PEGDA) have been characterized for temporal response, valve closure, and long-term durability. A ~100 ms valve opening time and a ~20 ms closure time offer valve operation as fast as 8 Hz with potential for further improvement. Comparison of circular and rectangular valve geometries indicates that the surface area for membrane interaction in the valve region is important for valve performance. After initial fabrication, the fluid pressure required to open a closed circular valve is ~50 kPa higher than the control pressure holding the valve closed. However, after ~1000 actuations to reconfigure polymer chains and increase elasticity in the membrane, the fluid pressure required to open a valve becomes the same as the control pressure holding the valve closed. After these initial conditioning actuations, poly-PEGDA valves show considerable robustness with no change in effective operation after 115,000 actuations. Such valves constructed from non-adsorptive poly-PEGDA could also find use as pumps, for application in small volume assays interfaced with biosensors or impedance detection, for example."
} | 308 |
33392172 | PMC7773924 | pmc | 3,086 | {
"abstract": "The textile and clothing industry is the first manufacture sector in Tunisia in terms of employment and number of enterprises. It generates large volumes of textile dyeing wastewater (TDWW) containing high concentrations of saline, alkaline, and recalcitrant pollutants that could fuel tenacious and resilient electrochemically active microorganisms in bioanodes of bioelectrochemical systems. In this study, a designed hybrid bacterial halothermotolerant bioanode incorporating indigenous and exogenous bacteria from both hypersaline sediment of Chott El Djerid (HSCE) and TDWW is proposed for simultaneous treatment of real TDWW and anodic current generation under high salinity. For the proposed halothermotolerant bioanodes, electrical current production, chemical oxygen demand (COD) removal efficiency, and bacterial community dynamics were monitored. All the experiments of halothermotolerant bioanode formation have been conducted on 6 cm 2 carbon felt electrodes polarized at −0.1 V/SCE and inoculated with 80% of TDWW and 20% of HSCE for 17 days at 45°C. A reproducible current production of about 12.5 ± 0.2 A/m 2 and a total of 91 ± 3% of COD removal efficiency were experimentally validated. Metagenomic analysis demonstrated significant differences in bacterial diversity mainly at species level between anodic biofilms incorporating allochthonous and autochthonous bacteria and anodic biofilm containing only autochthonous bacteria as a control. Therefore, we concluded that these results provide for the first time a new noteworthy alternative for achieving treatment and recover energy, in the form of a high electric current, from real saline TDWW.",
"conclusion": "Conclusion This work is the first to demonstrate the potential to develop a novel halothermotolerant bioanode incorporating allochthonous and autochthonous bacteria from both hypersaline sediment and TDWW. A core bacterial community composed of three autochthonous strains, Achromobacter sp., C. acidovorans, D. gadei , and four allochthonous strains, P. aquaticus, G. sulfurreducens, G. metallireducens , and M. sediminum , ensures the effectiveness of the bioanode by producing high current density (12.5 A/m 2 ) and a total of 91% of COD removal efficiency. These findings, achieved under both hypersaline (165 g/L) and thermophilic conditions (45°), could lead to possible applications of BES technology for treatment and energy recovering from high-temperature and high-saline wastewaters.",
"introduction": "Introduction The textile and clothing industry has become one of the most important sectors of activity. Despite the use of high-tech equipment and modern technologies, it remains among the highest water consuming industries. As reported by the United States Environmental Protection Agency (USEPA), the production of 9,072 kg of finished textile per day requires about 36,000 L of water only for wet processing (Ghaly et al., 2014 ; Berkessa et al., 2020 ). Moreover, almost the water consumed generates large volumes of textile dyeing wastewater (TDWW). The TDWW contains more than 1.5 g/L of waste dye as the uptake of these dyes by textile fabrics is very poor, high salinity of about 5–6% NaCl and 5% of Na 2 SO 4 and other chemicals, such as various acids, alkalis, sulfur, naphthol, surfactant-dispersing agents, formaldehyde-based dye fixing agents, hydrocarbon-based softeners, and heavy metals (Verma et al., 2012 ; Lin et al., 2015 ; Pazdzior et al., 2018 ). Most of these chemicals and products of their degradation, i.e., metabolites, are recalcitrant in nature and severely affect both aquatic and terrestrial life (Ben Mansour et al., 2012 ; Kant, 2012 ; Chandrakant et al., 2016 ). It is therefore well-established that these hazardous pollutants should be removed from TDWW by appropriate and effective methods prior to their release into the environment. Different TDWW treatment methods and techniques have been implemented and evaluated over the past two decades. These methods involve (i) physical methods (coagulation–flocculation, adsorption, and filtration techniques), (ii) oxidation methods categorized as advanced oxidation processes (cavitation, photocatalytic oxidation, Fenton chemistry) and chemical oxidation using oxidizing agents (O 3 and H 2 O 2 ), and (iii) bioremediation methods (fungi, algae, bacteria) (Gosavi and Sharma, 2014 ; Yeap et al., 2014 ; Chandrakant et al., 2016 ; Chouchane et al., 2018 ; Jiang et al., 2018 ). For several reasons, such as eco-friendly, cost competitive, less sludge production, and giving non-hazardous metabolites or full mineralization (Hayat et al., 2015 ), the biological methods are qualified as the most sustainable method for wastewater treatment. However, biological methods cannot guarantee the achievement of required results of the TDWW, as some of the dye molecules or other chemical components are hazardous and/or recalcitrant to microorganism-driven degradation. The most appropriate method should be cost-effective, valuable in the degradation of resistant compounds, and produce safe and good quality effluent. There is therefore a great environmental and economical need to develop a processing technology that addresses these severe challenges. Tenacious, resistant, and exoelectrogenic microorganisms used in bioelectrochemical systems (BES) can perform the dual function of degrading pollutants and recovering in the form of electrical energy, the energy resulting from the oxidation of these pollutants. Indeed, previous studies reported that exoelectrogenic populations have favorably demonstrated an added value for the treatment of recalcitrant pollutants, such as azo dyes, petroleum hydrocarbon, and heavy metals (Adelaja et al., 2015 ; Choudhury et al., 2017 ; Monzon et al., 2017 ; Grattieri and Minteer, 2018 ; Vijay et al., 2018 ; Askri et al., 2019 ; Elabed et al., 2019a , b ). As already reported by Xie et al. ( 2011 ), organic and inorganic compounds in TDWW contain almost five times more energy than that consumed to treat it. Thus, it is hypothetically possible to extract electrical energy from the TDWW by applying well-adapted electrochemically active microorganisms capable of degrading recalcitrant pollutants, transmuting dangerous metabolites, and capturing an electronic flow that can be converted into electrical energy under high salinity. In this context, an earlier study by Askri et al. ( 2019 ) demonstrated the enrichment of efficient exoelectrogenic microorganisms from hypersaline sediment of Chott El Djerid (HSCE) able to produce a current density in the range of 7 A/m 2 under combined high temperature and hypersaline conditions (temperature 45°C, salinity 165 g/L) using lactate as carbon source, i.e., anodic fuel. Halothermophilic microorganisms are thus suitable candidates for the treatment of the high saline wastewaters generated, for example, in the textile dyeing (2–10 g/L), seafood processing (8–20 g/L), tannery (40–80 g/L), and petroleum industries (few g/L to 300 g/L) (Shehab et al., 2017 ; Cherif et al., 2018 ; Askri et al., 2019 ). The aim of this work was therefore to demonstrate, for the first time, proof of the feasibility of designating an efficient microbial halothermotolerant bioanodes from both hypersaline sediment (HSCE) and saline TDWW microbium able to treat textile wastewater and generating an electrical current collected via an electrode. At this stage of progress, the exploitation of the anodic current flow generated is not investigated at all. It is only quantified in terms of bioelectrochemical kinetics, from the acquisition of the J = f(E) curves. It is thus not at all a question here of developing a microbial fuel cell (MFC), a microbial electrolysis cell (MEC), or any other BES, judged too fluctuating and random to focus attention on the precisely targeted phenomenon, i.e., the formation of a bioanode without limitation caused by a limiting step of a larger and more complete BES process. On the contrary, bioanode formation was studied here in three-electrode electrochemical bioreactor installations to ensure well-controlled electroanalysis conditions, as explained in the review by Rimboud et al. ( 2014 ), who point out the fundamental basic principles and advantages of the three-electrode arrangement compared with MFC or other BES devices. Once the potential for electrical current generation and TDWW treatment of the halothermotolerant bioanodes had been proven via electrochemical and analytical tools, contributions of autochthonous microorganisms from TDWW and allochthonous from HSCE were investigated through comparative metagenomic analyses of biofilms, HSCE, and TDWW.",
"discussion": "Results and Discussion Bioanode Growth and Electrochemical Characterization Four microbial bioanodes were independently grown by chronoamperometry with a VSP multichannel potentiostat (Biologic SAS), setting the potential value of the WEs at −0.1 V/SCE. This electrode potential has particularly shown its relevance to efficiently lead to the formation of particularly robust and efficient bioanodes from wastewater (Blanchet et al., 2015 ) and also from sediments (Erable et al., 2017 ). The four bioanodes grown at this polarization potential are referred to as TDWW, TDWWS1, TDWWS2, and TDWWS3 all through the figures and text. The constant electric polarization was maintained for 17 days, which is a reasonable amount of time to obtain matured bioanode on carbon based electrodes colonized by halo-exoelectrogenic bacteria (Erable et al., 2009 ; Rousseau et al., 2014 ; González-Muñoz et al., 2018 ; Askri et al., 2019 ). Current densities vs. time plots for −0.1 V/SCE polarization potential are shown in Figure 1 . The start of current production was observed almost after 2 days of polarization for all three bioanodes co-inoculated with HSCE, in comparison with the TDWW bioanode, where current production was only recorded from day 5 with very low densities throughout the experiment. This suggests that the electroactive community developed faster and more efficiently in reactors inoculated with HSCE. Remarkably, the maximal current performance reached during the experiments was almost the same for the three bioanodes TDWWS1, TDWWS2, and TDWWS3. A reproducible current production of about 12.5 ± 0.2 A/m 2 was obtained. However, maximum current density peaks were reached after different periods of polarization for 4, 6, and 10 days for TDWWS3, TDWWS1, and TDWWS2, respectively. Successive batches of addition of textile wastewater usually make it possible to harmonize the current densities over the long term of several replicates of experiments. Here, the reproducibility of the TDWWS replicates observed from the first batch is to be underlined and certainly comes from the inexistence of competitive reactions due to the inhibition of non-electroactive microorganisms by the high salinity, the toxicity of the pollutants, and the possible presence of oxygen traces. Figure 1 Evolution of the current density (A/m 2 ) vs. time (days) for experiments on a carbon felt electrode of 6 cm 2 projected surface area polarized at −0.1 V/SCE in a reactor containing textile dyeing wastewater (TDWW) and reactors containing 80% of TDWW and 20% of saline sediments (TDWWS1, TDWWS2, TDWWS3). Previous works have shown that the performance of bioanode systems is basically inversely proportional to the complexity of the substrate as in the case of TDWW (Pant et al., 2010 ; Pandey et al., 2016 ; Heidrich et al., 2018 ). However, it is worth noting that in this study, the oxidation currents obtained are almost doubled compared with those obtained with the same inoculum (HSCE) using lactate (5 g/L) as substrate (Askri et al., 2019 ), and the initial COD for both sets of experiments was 975 and 1,432 mg/l, respectively. The diversity of the TDWWS biofilm composed jointly of autochthonous and allochthonous bacteria may be critical to its performance. Overall, wastewater used to feed BES, mainly MECs and MFCs, was either domestic wastewater or industrial wastewater from a wide variety of sources (breweries, dairies, refineries). The highest current densities obtained by BES fed with real industrial effluents were 10.7 and 10.3 A/m 2 where biorefinery (Pannell et al., 2016 ) and brewery wastewaters (Yu et al., 2015 ) have been used, respectively. However, for BES fed with domestic wastewater, the highest current densities obtained were 3.8 (Ullery and Logan, 2015 ) and 3.5 A/m 2 (Blanchet et al., 2016 ). The average current densities calculated from 48 research papers are 2.6 A/m 2 for industrial wastewater and 0.8 A/m 2 for domestic wastewater. Two main reasons could explain in part this significant difference in average current densities: (i) industrial wastewater is generally more conductive (7 mS/cm) than domestic wastewater (1.5 mS/cm) (Yen et al., 2016 ) and (ii) the total organic matter concentration in industrial wastewater is between 5,000 and 12,000 mg/L (Rajeshwari et al., 2000 ), whereas that in domestic wastewater is between 320 and 740 mg/L (Almeida et al., 1999 ). The obtained results demonstrated that the TDWW, although rather recalcitrant to biological treatment, is found as the most suitable effluent to generate electric current using hypersaline sediment as a source of tenacious and exoelectrogen biocatalysts. CV was used as a tool to confirm the presence of electroactive biofilm on the electrode surface. Results from CV tests for the four different bioanodes are shown in Figure 2 . The initial CVs of the TDWW (negative control) and TDWWS1, S2, and S3 on the porous carbon felt electrode clearly showed the absence of an electroactive biofilm due to the fate shape of the voltammogram from 0.0 to +0.3 V. Interestingly from 0.0 to −0.6, a sharp decrease of the current was observed very likely due to the electrochemical reduction of some compounds in the medium. Another CV cycles were performed when the maximum values of the current density were reached for all reactors (turnover CVs). For the TDWW, no noticeable change from its initial state is visible, confirming the total absence of development of an electroactive biofilm on the corresponding carbon felt electrode. On the other hand, for the other samples, TDWWS1, S2, and S3, a remarkable difference in the general shape of the turnover CVs is identified compared with their initial states and with the TDWW. Figure 2 Cyclic voltammetry performed in different conditions: (A) CV of porous carbon felt electrode immersed in TDWW; (B–D) CV of porous carbon felt electrode immersed in TDWW and HSCE sediment obtained at two different times: CV I (–): at the beginning of the experiment; CV M (–) at maximum current density. All three TDWWs turnover CVs have a zero current potential very close to −0.4 V/SCE. Below this potential, a reduction current is detected that increases to the lower potential limit of −0.6 V/SCE. This reduction current is a less pronounced residual of the same reduction phenomenon(s) observed on all electrodes at t = 0. Concerning the visible oxidation current for potentials higher than −0.4 V/SCE, it is constantly increasing from −0.4 to −0.1 V or even 0.0 V/SCE. Beyond these potential values, i.e., for potentials more positive than −0.1 V or 0.0 V/SCE, the maximum speed of the exchange current is reached, and a current plateau is therefore observable between −0.1 V and +0.3 V/SCE. The current density even tends to decrease over this anode potential range in the case of Figures 2B,D because the scanning speed is certainly too fast to ensure a steady state of bacterial metabolic phenomena limiting the production of the electron flow. In sum, the CV results clearly indicate the establishment of a biofilm with electrocatalytic properties on the carbon felt electrodes, as well as a successful enrichment of this electroactive biofilm with bacteria capable of using electron acceptor electrodes (Harnisch and Freguia, 2012 ; Rimboud et al., 2014 ). The hysteresis identified between the forward and return curves of the CVS ( Figures 2B–D ) is classic of bioanodes formed in saline or hypersaline environments (Erable and Bergel, 2009 ; Rousseau et al., 2014 ), where capacitive phenomena related to the high ionic charge of the electrolyte and especially the capacity of the electrode material and electroactive biofilm couple coexist. COD Measurement Table 1 indicates that reactors inoculated with HSCE showed satisfactory performance in terms of both current production and COD removal that were about 12.5 ± 0.2 A/m 2 and 91 ± 3%, respectively. Interestingly, these reactors demonstrated COD removal proportional to the production of current generation. However, COD removal in the reactor not inoculated with HSCE was much lower, i.e., 42.3%, than those in the inoculated reactors, and the production of the current also remains very low. For bioremediation of wastewater containing recalcitrant pollutants, these findings revealed that reactors inoculated with HSCE were found to be comparatively higher in COD removal efficiency than other textile wastewater treatments, such as (i) anaerobic internal circulation reactor (COD removal = 87%), (ii) Fenton's process with and without pH adjustment where COD removal was 89 and 33%, respectively (Hayat et al., 2015 ), and (iii) the combination of homogenization decantation and membrane treatments (COD removal = 66%) (Buscio et al., 2015 ). Table 1 COD removal efficiency of TDWW from different reactors. Reactors Contents Temperature ( ° C) Duration (day) Maximal current density (A/m 2 ) COD removal rate (%) 1 20% HSCE + 80% TDWW 45 17 12.5 91.0 2 20% HSCE + 80% TDWW 45 17 12.3 93.6 3 20% HSCE + 80% TDWW 45 17 12.7 88.3 4 (negative control) 100% TDWW 45 17 0.3 42.3 Microscopy Analysis of Biofilms Morphology Colonization of the carbon felt electrodes was evaluated at the end of the experiments, mainly on the outer surface of the felts but also within the porosity of the felt. Indeed, previous work has shown the difficulty of electroactive biofilms to colonize the internal surfaces of felt electrode structures (Chong et al., 2019 ), especially when real effluents either viscous, highly charged with suspended solids, or very highly loaded with COD are used (Blanchet et al., 2016 ). Carbon felt WEs possibly covered with HSCE and/or TDWW microorganisms were removed from the reactors and imaged by epifluorescence microscopy ( Figure 3 ). Samples of carbon felt WEs were TDWW that is a bioanode with very low current production (300 mA/m 2 ) and TDWWS1, TDWWS2, and TDWWS3 that produced similar current densities of 12.5 ± 0.2 A/m 2 . Figure 3 Evaluation of microbial proliferation on the outer surface of carbon felt electrodes by epifluorescence microscopy imaging. Carbon felt electrode surface has been stained with acridine orange to localize living and dead microbial cells. TDWW was the electrode sample with the lowest surface colonization rate. Only a few fibers on the surface of the carbon felt electrode were partially colonized by scattered, non-continuous, sparse groups of bacterial colonies with low bacterial cell density. Since this bioanode generated almost no current, it is not surprising to observe a low rate of colonization. The few isolated clusters of bacterial colonies are probably mostly similar to suspended biological substances, kinds of microbial flocs participating in the non-bioelectrochemical oxidation of the COD of textile effluents. In comparison, the electrode surfaces of TDWWS1, TDWWS2, and TDWWS3 were much more thickly colonized. In spite of relatively similar bioelectrochemical behaviors, but with a time lag, the microscopic patterns of the biofilms show significant differences. TDWWS1 was colonized by a continuous biofilm settled between the surface fibers. The production of exopolymeric substances was also much more prominent for this biofilm. TDWWS2 showed a non-homogeneous colonization of the carbon fibers. The appearance of the biofilm was like the control not inoculated with sediment from HSCE, but the density and number of microbial clusters were much higher. Finally, TDWWS3 showed an intermediate colonization profile between TDWWS1 and TDWWS2, with a thin continuous biofilm that closely covered the surface fibers of the carbon felt. This type of electroactive biofilm enveloping the carbon fibers is generally a feature of biofilms in hypersaline environments (Rousseau et al., 2014 ). However, we have also recently revealed that the physical structure of electroactive biofilms formed from sediments from HSCE, i.e., enveloping the fibers or distributed between the fibers, was not systematically a sign of improved or, on the contrary, decreased current production (Askri et al., 2019 ). Microscopic inspection of the internal porosity of the carbon felt electrodes revealed a very low degree of internal porosity colonization for the electrodes TDWWS1, TDWWS2, and TDWWS3 ( Figure 4 ). Even so, no trace of biofilm could be seen in the core of the TDWW electrode belonging to the electrochemical reactor not inoculated with HSCE. The poor accessibility to the porosity of the carbon felt is usually due to the clogging phenomena of the external pores of the electrode exerted by particles in suspension present in the wastewater (Blanchet et al., 2016 ). Insofar as the external faces of the carbon felts do not present a colonization obstructing the pores, i.e., the inter-fiber spaces, this hypothesis is not at all conceivable here. The lack of agitation of the reaction medium, the hydrophobicity of the carbon felt, and the low growth rate of microorganisms in the textile wastewater are certainly individual or combined tracks that can partly justify the low internal colonization rate of the 3D felt electrodes. Figure 4 Observation of the internal microbial colonization of the carbon felt electrodes. A 1 cm thick cross-section of the electrodes was prepared, and observations were then performed in the center of the porosity of the electrodes after staining the dead and living microbial cells with acridine orange. Molecular Analysis of the Bacterial Communities The composition and abundance distribution of each sample (rTDWW and HSCE) and biofilms (TDWW, TDWWS1, TDWWS2, and TDWWS3) at the taxonomic levels of phylum and species are shown in Figures 5A,B . Figure 5 (A) Bacterial distribution at phylum level of samples from raw textile dyeing wastewater (rTDWW), hypersaline sediment of Chott El Djerid (HSCE), and biofilms from a reactor containing textile dyeing wastewater (TDWW) and reactors containing 80% of TDWW and 20% of saline sediments (TDWWS1, TDWWS2, and TDWWS3). (B) Bacterial distribution at species level of samples from rTDWW, HSCE, and biofilms from a reactor containing TDWW and reactors containing 80% of TDWW and 20% of saline sediments (TDWWS1, TDWWS2, and TDWWS3). At phylum level ( Figure 5A ), raw textile dyeing wastewater (rTDWW) bacterial community, considered as the autochtone community, was mainly characterized by Bacteroidetes (48.5%), Actinobacteria (14.5%), Proteobacteria (12.5%), Planctomycetes (9.5%), and Chloroflexi (8%). Previous studies have demonstrated that the bacterial phyla Acidobacteria, Planctomycetes, and Chloroflexi were more abundant in samples from textile wastewater (Meerbergen et al., 2016 ). However, in hypersaline sediment sample of Chott el Djerid (HSCE), source of the allochthonous community, the most abundant phylum was Proteobacteria (86%) approximately seven times higher than that in rTDWW (12.5%). In addition to Proteobacteria, the HSCE sample was hosted by other low-grade phyla, such as Firmicutes (6.8%), Cyanobacteria (5%), and Bacteroidetes (2%). Earlier studies using different molecular methods (Ben Abdallah et al., 2016 , 2018 ) have demonstrated that the bacterial community in HSCE was dominated by Proteobacteria, followed by Firmicutes, Bacteroidetes, Cyanobacteria, and Actinobacteria. Based on the species assignment results ( Figure 5B ), Dysgonomonas gadei and Filimana aquilariae were the two dominant species in rTDWW, wherein their relative abundances were 21 and 19%, respectively. From the detected phyla Actinobacteria, Micrococcus sp. and Rhodococcus sp. were the most representative species at relative abundances of 12.6 and 11.5%, respectively. However, the Proteobacteria phylum shows a bacterial profile with the enrichment of Pseudomonas sp. (12.5%), Comamonas acidovorans (12%), Lysobacter sp. (11%), Aeromonas hydrophila (10.2%), Burkholderia cepacia (9%), and Achromobacter sp. (8.5%). By contrast, sample coming from HSCE was represented by various species of Proteobacteria with a predominance of Psychrobacter aquaticus (14.5%), Marinobacter sediminum (13.5%), Psychrobacter alimentarius (12.5%), Geobacter sulfurreducens (9.5%), Marinobacter hydrocarbonoclasticus (9%), Geobacter metallireducens (9%), and Halomonas spp. (7.5%). Figure 5B indicates that the relative abundances of bacteria at phylum level in TDWWS1, TDWWS2, and TDWWS3 biofilms are quite similar, whereas those in TDWW biofilm show some differences. TDWWS1, TDWWS2, and TDWWS3 biofilms demonstrated the presence of the same phyla and largely at the same pattern of abundance Proteobacteria > Bacteroidetes = Firmicutes > Actinobacteria > Chloroflexi = Planctomycetes > Thermotogae. As example, TDWWS1 biofilm showed Proteobacteria (30.5%) > Bacteroidetes (17.43%) = Firmicutes (17.5%) > Actinobacteria (5.8%) > Chloroflexi (1.5%) = Planctomycetes (1.6%) > Thermotogae (0.35%). It is worth noting that the bacterial phyla identified in the three biofilms show heterogeneous profiles composed of phyla found in rTDWW and in HSCE. In this case, TDWWS1, TDWWS2, and TDWWS3 biofilms harbored autochthonous phyla from rTDWW represented in particular by Bacteroidetes, Actinobacteria, Chloroflexi, and Planctomycetes and allochthonous phyla from HSCE, such as Proteobacteria, Firmicutes, and Thermotogae. The heterogeneous bacterial profile of the different biofilms occurs more clearly at species level. P. aquaticus (8.06–16.02%), P. alimentarius (3.04–6.5%), G. sulfurreducens (8.2–11.8%), G. metallireducens (2.5–5.4%), M. hydrocarbonoclasticus (6.3–7.4%), and M. sediminum (2.3–11.7%) as allochthonous species (and not found in rTDWW sample) were the most abundant species in TDWWS1, TDWWS2, and TDWWS3 biofilms. Furthermore, B. cepacia (2.4–7.6%), Achromobacter sp. (10.1–16.4%), C. acidovorans (8.4–14.2), D. gadei (2.6–15.0%), Rhodococcus sp. (3.2–9.6%), and Micrococcus sp. (5.8–15.7%) as autochthonous species (and not found in HSCE) were also abundant in TDWWS1, TDWWS2, and TDWWS3 biofilms. By comparing the bacterial community of these three biofilms to that hosted TDWW, considered as a control, we found that TDWW harbored less bacteria belonging to the Proteobacteria (9.47%) and Bacteroidetes (3.22%) phyla. However, it hosted more bacteria from the Planctomycetes (4.5%), Chloroflexi (3.6%), and Spirochaeta (1.5%) phyla. At species level, the most abundant bacteria in TDWW biofilm were D. gadei (19.04%), Pseudomonas sp. (16.66%), Achromobacter sp. (15.47%), F. aquilariae (14.28%), and Lysobacter sp. (8.33%). This biofilm (TDWW) composed only of autochthonous bacteria was not electrochemically effective, as shown in Figure 1 . Only few hundred milliamperes were obtained as current generation (300 mA). These strains were also noticed in previous studies on microbial communities in textile wastewater (Meerbergen et al., 2016 ). Strikingly, high current production was obtained (12.5 ± 0.2 A/m 2 ) with biofilms (TDWWS1, TDWWS2, and TDWWS3) incorporated by both bacteria from HSCE and TDWW samples. A core bacterial community that was shared by the three biofilms could be highly involved in the current production. This core was composed of three autochthonous strains, Achromobacter sp., C. acidovorans, D. gadei , and four allochthonous strains, P. aquaticus, G. sulfurreducens, G. metallireducens , and M. sediminum ."
} | 7,030 |
29311325 | PMC5789925 | pmc | 3,087 | {
"abstract": "Significance Complex systems in many fields, because of their intrinsic nonlinear dynamics, can exhibit a tipping point (point of no return) at which a total collapse of the system occurs. In ecosystems, environmental deterioration can lead to evolution toward a tipping point. To predict tipping point is an outstanding and extremely challenging problem. Using complex bipartite mutualistic networks, we articulate a dimension reduction strategy and establish its general applicability to predicting tipping points using a large number of empirical networks. Not only can our reduced model serve as a paradigm for understanding the tipping point dynamics in real world ecosystems for safeguarding pollinators, the principle can also be extended to other disciplines to address critical issues, such as resilience and sustainability.",
"discussion": "Discussion Complex dynamical systems exhibiting a tipping point are widespread, and it is of interest to understand the dynamical mechanism of the tipping point and to develop predictive tools. To accomplish these goals, a viable solution is dimension reduction. We focus in this paper on bipartite mutualistic networks, not only as a concrete example to show the use of dimension reduction but also, because of the fundamental values of safeguarding pollinators to human survivability ( 28 ). In a mutualistic network system, a tipping point typically exists. As the environment continues to deteriorate, the system can drift toward the tipping point, where the catastrophic phenomenon of pollinator collapse will occur. The backbone that supports the functioning of such a network is mutualistic interactions between the pollinators and plants. To understand the role of the interactions with respect to the emergence of a tipping point, both species of the bipartite network must be retained in a reduced model. That is, the minimum dimension of the reduced system should be two [a 1D reduced model ( 18 ) is inadequate to describe mutualistic interactions]. With this in mind, we carry out a dimension reduction process by resorting to different types of averaging methods for species abundances. In particular, given an empirical mutualistic network, we carry out averaging processes to arrive at a 2D model with two collective dynamical variables: one for the pollinators and another for the plants. The average can be either unweighted or weighted. We show that our 2D reduced model captures the essential features of all 59 available real world mutualistic networks, not only in terms of the average abundances but more importantly, in terms of the occurrence of the tipping point, even in the presence of stochastic disturbances. We also find that, because of the lack of sufficient randomness in real mutualistic networks, a weighted average (e.g., based on degrees or eigenvectors) is necessary for the reduced model to exhibit a tipping point at the same critical parameter value as with the original network. Our 2D model can thus serve as a generic paradigm for understanding the tipping point dynamics in real world mutualistic networks. For example, the 2D model can be exploited to investigate a variety of nonlinear dynamical phenomena in mutualistically interacting networked systems, such as bifurcations ( 29 ), basin structures ( 30 , 31 ), crises ( 32 ), and transient chaos ( 33 – 38 ), which would otherwise be infeasible with the original systems because of their high dimensionality."
} | 860 |
31362385 | PMC6789575 | pmc | 3,089 | {
"abstract": "One approach to understanding how life-like properties emerge involves building synthetic cellular systems that mimic certain dynamical features of living cells such as bacteria. Here, we developed a model of a reaction network in a cellular system inspired by the ability of bacteria to form a biofilm in response to increasing cell density. Our aim was to determine the role of chemical feedback in the dynamics. The feedback was applied through the enzymatic rate dependence on pH, as pH is an important parameter that controls the rates of processes in cells. We found that a switch in pH can be used to drive base-catalyzed gelation or precipitation of a substance in the external solution. A critical density of cells was required for gelation that was essentially independent of the pH-driven feedback. However, the cell pH reached a higher maximum as a result of the appearance of pH oscillations with feedback. Thus, we conclude that while feedback may not play a vital role in some density-dependent behavior in cellular systems, it nevertheless can be exploited to activate internally regulated cell processes at low cell densities.",
"conclusion": "5. Conclusions The design of reaction networks that mimic the feedback-driven behavior of biological systems remains a formidable challenge. There has been increasing interest in obtaining quorum sensing type behavior in synthetic systems [ 5 , 12 , 33 , 34 , 35 , 36 , 37 ]. The underlying processes are vastly different however all examples exploit seemingly similar principles of an autocatalytic signaling process and communication between entities via a common surround. Here, taking inspiration from bacteria, we have proposed a model to determine the role of feedback in certain density-dependent behavior. The model involved an enzyme catalyzed reaction in cells which influenced the pH and the dynamics of an external pH-regulated process: the base-catalyzed formation of a gel or precipitate. We found that a sharp switch in pH was obtained in the multi-cellular system above a critical number of cells and gelation/precipitation did not require the presence feedback. However, we only observed pH oscillations with feedback and the oscillations resulted in a higher pH maximum in the cells. We conclude that feedback may be used to activate internal cellular processes at low substrate concentrations and low cell densities, even when it plays no role in the external dynamics. Although the chemistry of the underlying feedback processes was not necessarily relevant to quorum sensing, similar dynamical responses may arise in cells. This model system was also based on processes that might be used for synthetic quorum sensing analogues in biotechnology applications [ 38 , 39 ].",
"introduction": "1. Introduction Bacteria are arguably one of the most prolific forms of life on earth, capable of surviving the harshest of environmental conditions. The processes that allow bacteria to maintain homeostasis, adapt to changing environmental conditions, and switch between vastly different states in response to external stimuli are often driven by internal feedback mechanisms. Feedback, when a process is regulated by its output, plays a vital role in the functioning of living organisms [ 1 ]. Feedback is also utilized to drive a sharp switch from single cell to multi-cellular behavior in a phenomenon known as quorum sensing [ 2 ]. Bacteria communicate by release and amplification of signaling species resulting in population-level responses such as the formation of biofilm above a threshold number or density [ 3 ]. Biofilms consist of a sticky matrix of polysaccharides and other biomolecules that protect the bacteria against the external environment. The development of antimicrobial resistance or tolerance typically involves the formation of a biofilm driven by a quorum sensing mechanism. The design of biologically-relevant reaction networks that can reproduce dynamical cell behaviors may provide an insight into the role of feedback in such phenomena [ 4 , 5 ]. Progress in systems and synthetic biology has resulted in oscillations and quorum sensing driven by autocatalytic networks in genetically modified organisms [ 6 ]. The systems approach to chemistry is also undergoing rapid development [ 7 ] but although feedback based on DNA [ 8 ], peptides [ 9 ] and other biomolecules [ 10 , 11 ] has been implemented, autocatalytic communication mechanisms have been mainly investigated in regard to inorganic redox processes [ 12 , 13 ]. There are relatively few robust examples of chemical feedback in which cell-to-cell communication can be explored. Taking inspiration from bacteria, we developed a model to explore the role of feedback in a cellular system when the feedback is coupled to other process, such as gelation or precipitation of a substance in the external solution. The model consisted of an enzyme-catalyzed reaction network based on urease, a virulence factor produced by bacteria [ 14 ] and associated with biomineralization, the precipitation of inorganic salts by living organisms [ 15 ]. It has been demonstrated that the urease reaction can display feedback as a result of the enzyme rate dependence on pH [ 16 , 17 ]. The underlying chemistry, although biologically relevant, is not implicated in quorum sensing in bacteria. However, feedback through pH is of interest as pH plays an important role in the rates of many processes and oscillations in pH have been observed in living systems as well as implicated in the origin of life [ 18 , 19 ]. Here, we used the model to explore the influence of increasing density of cells on the dynamics and determined the importance of feedback on the resultant behavior. Our aim was to build a simple model that may shed light on the role of feedback in natural processes but also to aid in the design of synthetic cells that may be exploited in biotechnological applications. There have been great advances in the development of synthetic cells based on, for example, gene networks or enzymatic reactions in liposomes [ 20 , 21 ], but few examples exploiting feedback. Hence our model is based on an experimentally-realizable reaction network that might be implemented in a synthetic cell, rather than a realistic mimic of biological processes in bacteria.",
"discussion": "4. Discussion Feedback plays a vital role in the functioning of living systems [ 1 ]. One of the important applications of feedback in bacteria is the release and detection of a small diffusible molecule, the autoinducer, in a phenomenon known as quorum sensing [ 24 ]. Quorum sensing is used by bacteria to activate population-level responses such as the formation of biofilm and a transition from single cell to multi-cellular behavior [ 3 ]. In biofilms, cells are bound together in a matrix of polysaccharides and other biomolecules that protect them from the external environment. Here, inspired by biofilm formation, we have developed a reaction network in a cellular system coupling feedback with other chemical processes such as gelation in the external solution. We chose to examine feedback through pH because pH is an important parameter that controls the rates of many processes. Our system is also based on experimentally-realizable processes that could be implemented in a synthetic cell. In our simple model, the enzyme reaction took place in a cell coupled to an external solution containing substrate, S, and acid. Feedback arose through the characteristic bell-shaped rate-pH curve of the enzyme-catalyzed reaction. The reaction produced a weak base and the rate of the enzyme reaction depended on the pH, with a maximum at pH 7. In a cell with no external supply of species, we obtained a switch from low pH to high pH after a time lag accompanied by the sudden appearance of product, N, and rapid consumption of substrate. This is an example of pH-driven feedback with the basic product of the enzyme reaction acting as autocatalyst. If the feedback was removed by making the enzyme rate independent of pH then we observed no rate acceleration in the loss of substrate. Nevertheless, the basic product, N, suddenly appeared after a time lag and there was a switch to high pH. In this acid-base reaction system, the initial acid reacted with the base and kept the pH low until sufficient base was produced to overcome the acid. This is similar to the substrate-depletive chemical clock reactions driven by inhibitor removal in the absence of feedback i.e., acid acted as an inhibitor suppressing formation of OH − [ 25 ]. When the mass transport of species from an external reservoir of constant concentrations was included, we obtained oscillations in pH. The differential transport of acid and substrate was essential for the observation of oscillations, i.e., the transport rate constant k H > k S . This is in agreement with earlier work on pH oscillations in some enzyme-catalyzed reactions [ 26 , 27 , 28 ]. Oscillations typically require positive feedback coupled with delayed negative feedback. The net rate of the transport of acid into the cell increased with increasing difference between the inner and outer pH, i.e., as the pH increased the rate of transport of acid into the cell increased. Therefore, the acid provided the necessary delayed negative feedback that terminated the autocatalytic, base-producing process. There are few experimental examples of transport-driven instabilities in chemical systems, where the differential rates of transport of species cause oscillations in space or time [ 29 ]. The concept was first explored in Turing’s pioneering modelling work on the basis of morphogenesis, in which spatial patterns were observed in a ring of coupled cells when the inhibitor transport was greater than that of the substrate in an autocatalytic process [ 30 ]. Differential transport may arise naturally in acid-base systems as acid, in the form H 3 O+, diffuses typically much quicker than other small ions in solution: the diffusion coefficient of acid can be up to nine times greater than small ions. In the model presented here, the chemical species were produced in a cell and crossed the cell wall with rates that were related to the permeability coefficients and radius of the cell. Permeability coefficients vary widely and depend on the nature of the membrane components, membrane thickness as well as the species involved [ 31 ]. In general, charged species do not cross cell membranes with appreciable rates. However, in a synthetic cell system constructed with phospholipids, fast acid transport could be facilitated by the incorporation of ion channels in the cell membrane [ 32 ]. Living cells are filled with buffering agents, so it is important to understand the influence that buffers might have on pH-driven feedback. We found that oscillations were, not surprisingly, completely suppressed for values of the buffer concentration greater than the substrate concentration. However, for lower concentrations the buffer primarily reduced the maximum of the oscillations. Thus, buffers could be used to control the amplitude of oscillations with a pH maximum less than 7. We mapped the behavior of the single cell using bifurcation diagrams of pH as a function of substrate concentration in the reservoir, [S r ]. The transition from a low pH state to oscillations was obtained with feedback whereas only a single steady state pH was observed when no feedback was present. Nevertheless, we obtained a sharp switch in pH as [S r ] was increased in both cases. Thus the system may be considered ultrasensitive to small changes in substrate, even in the absence of feedback [ 1 ]. The acid acting as an inhibitor played an important role in generating this response. In the model with multiple cells, we demonstrated that the increase in pH could be used to drive a base-catalyzed gelation or precipitation process in the external solution. This process was inspired by the quorum sensing of bacteria that leads to the formation of biofilm containing polysaccharides or the precipitation of inorganic minerals such as calcium carbonate, which occur in natural systems as the pH of the solution increases. We found that above a threshold density of cells, d = 0.3, the pH of the external solution increased, and gelation/precipitation occurred. A value of d = 0.01 corresponds to a number density of 2.7 million cells per ml of solution, assuming cells of radius 1 µm; comparable with bacterial cell cultures. The critical density of cells for a switch to high pH and gel formation was not influenced by the presence of feedback, rather the inhibition by acid played the major role in this switch-like behavior. We note that the behavior here was obtained for relatively low transfer rates in which a decoupling between the internal and external system was observed: there was a difference in the timescales of the pH dynamics in the cell and the pH dynamics in the outer solution. No large amplitude oscillations in pH in the external environment were obtained for any values of the parameters explored. The transfer coefficients ranged from 0.0014–0.008 s −1 ; with a cell of the order 1 μm, this gives permeability coefficients from 10 −9 –10 −8 m s −1 , on the low side for transport of neutral molecules in synthetic vesicle systems. It may be that for larger transport coefficients the feedback plays a greater role in the switch for gelation. However, we demonstrated that the feedback played an important role in the internal cell dynamics. The emergence of oscillations above a critical density of cells resulted in a much higher cell pH than without feedback. Thus, we conclude that the presence of feedback could be used to drive an internal switch that activates other pH-regulated processes within the cell at low substrate or cell densities."
} | 3,450 |
38398981 | PMC10892042 | pmc | 3,090 | {
"abstract": "Memristor devices have diverse physical models depending on their structure. In addition, the physical properties of memristors are described using complex differential equations. Therefore, it is necessary to integrate the various models of memristor into an unified physics-based model. In this paper, we propose a physics-informed neural network (PINN)-based compact memristor model. PINNs can solve complex differential equations intuitively and with ease. This methodology is used to conduct memristor physical analysis. The weight and bias extracted from the PINN are implemented in a Verilog-A circuit simulator to predict memristor device characteristics. The accuracy of the proposed model is verified using two memristor devices. The results show that PINNs can be used to extensively integrate memristor device models.",
"conclusion": "5. Conclusions To date, there has been much research on memristors, but no generic model has been proposed. In this study, we propose a methodology for training memristor state ODEs using the unsupervised learning of PINNs. This methodology has a simpler structure and better prediction performance than conventional neural networks. This can solve traditional memristor state ODEs that are complex to analyze numerically, which allows the integration of many existing memristor device models and enables a compact model for circuit simulator. Two models are presented to validate the memristor compact modeling methodology. The resistive states of the two models are analyzed using PINN learning. The trained data can be used to predict the physical properties of memristors with high accuracy. This proves that the memristor compact modeling methodology using PINN is reasonable. In addition to presenting a PINN-based compact model under ideal conditions, the inclusion of variability in the PINN memristor model is essential. Future research will include a PINN methodology for modeling non-ideal device characteristics such as device-to-device variability and cycle-to-cycle variability.",
"introduction": "1. Introduction The memristor was first proposed by L. Chua in the context of traditional passive devices such as resistors (R), inductors (L), and capacitors (C). This is described as the interaction between electric charge and magnetic flux [ 1 ]. The conventional von Neumann architecture faces challenges such as high-power consumption and low computing speed. Memristors can serve as integral components in neuromorphic computing to overcome these problems [ 2 ]. Recently, memristors have been the focus of significant research, particularly in their applications in various fields such as the study of non-volatile CMOS memristors [ 3 ]. Furthermore, there is active discussion on research utilizing Hopfield neural networks (HNNs) and heterogeneous discrete neural networks (HDNNs) that consider synaptic behavior [ 4 , 5 ]. A memristor is a resistive and non-volatile memory that changes state when a voltage is applied [ 6 ]. Essentially, a memristor device has a two-terminal metal/insulator/metal (MIM) sandwich structure [ 7 ]. In addition, memristors have various switching mechanisms depending on the material and structure such as filament formation/rupture [ 8 ], transition by amorphous/crystalline phase change [ 9 ], reversal of ferroelectric polarization direction [ 10 ], current-induced magnetization [ 11 ], etc. The physical analysis of memristors with such nonlinearity and complexity is very difficult and time-consuming [ 12 ]. The memristor I–V characteristics are described with a hysteresis curve depending on the on/off state, which, in mathematical terms, has the form of a differential equation [ 13 ]. Memristor devices have diverse and complex models depending on their structure, and there are various differential equations to solve them. In order to develop a comprehensive memristor model, researchers have tried building the framework of memristor models such as window function modification [ 14 ], hybrid memristor memory [ 15 ], capacitive connections of memristors [ 16 ], and tantalum oxide memristor models [ 17 ]. It is, therefore, not surprising that we need to unify various types of memristor models [ 18 , 19 , 20 ] for device and circuit simulators. Solving differential equations based on traditional numerical methods, such as structural modeling of memristors and the memristor model of Messaris et al., requires significant time and effort [ 21 , 22 , 23 , 24 , 25 , 26 ]. To replace traditional numerical methods, the long short-term memory (LSTM) neural network methodology, a highly accurate model, is used to describe the hysteresis phenomenon of memristors. However, it has the disadvantage of having to consider past time and is more complex than the traditional multi-layer perceptron (MLP) model [ 27 ]. Moreover, the study on memristor-based neural networks utilizing recurrent neural network (RNN) [ 28 ] demonstrates excellent performance in data prediction. Nonetheless, the structure of the neural network is highly complex. This study proposes a methodology to solve the difficult and complex differential equations of a memristor with a physics-informed neural network (PINN). PINN is an artificial neural network (ANN) for numerically solving differential equations and has a more intuitive and simpler structure than the traditional LSTM model. It can also obtain solutions to differential equations faster and more accurately than traditional models [ 29 , 30 , 31 ]. PINN can be used to predict the I–V characteristics of a memristor and extract the weight and bias of the predicted function. The neural network incorporating extracted weights and biases is implemented with Verilog-A for circuit validation. For more accurate verification, we use two memristor models. The equation for the conductance of a memristor device consists of differential equations. This implies that various differential equations exist for different structural models. PINN provides a direct and easy solution of structure-dependent physics-based differential equations. Consequently, we validate that the PINN methodology is applicable for integrating these diverse models, which has not been explored in previous memristor research. In Verilog-A, the application of the extracted weights and biases demonstrates the potential of the circuit simulator. As a result, this methodology allows for the compact modeling of various memristor devices. Additionally, it can be applied to the study of various devices that require physical analysis and integrated modeling. In this study, compact modeling has the advantage that each of the different model equations can be easily analyzed using one unified modeling technique. Accordingly, it reduces the time and cost required to construct the physical modeling of new devices. In this work, Section 2 introduces memristor behavior and resistive state characteristics depending on voltage and time. Furthermore, we describe the model normalization procedure in PINN training. Section 3 details the PINN method, including the loss function and configuration of the neural networks. Section 4 presents the simulation process using the PINN methodology. In this section, we also illustrate the results of circuit simulator implementation in Verilog-A. Finally, Section 5 concludes the paper.",
"discussion": "4. Results and Discussion A PINN compact model is implemented using the two models described in Section 2 . The training model of PINN is constructed using the python package SciANN [ 38 ]. The neural network is composed of two layers with 10 neurons and 20 neurons for each of the two models. The use of a simple neural network structure implies a significant reduction in computer execution time. Furthermore, the Verilog-A simulation takes a relatively short time as it operates simply with the weights and biases extracted from the neural network. The simulation process unfolds as follows: 1. Train the two models outlined in Section 2 through Python-based PINN learning. 2. Extract weight and bias values from the trained results. 3. Generate symbols in Verilog-A using the extracted weight and bias values. 4. Implement the circuit using the generated symbols. 5. Verify whether the physical characteristics obtained through traditional numerical analysis methods and PINN learning are identical. For GMMS model, the parameters required for training the GMMS model are provided in Table 1 [ 21 ]. The most difficult part of training the memristor state through PINN is insufficient training in regions where the time axis is close to 0 and the voltage axis is at inflection points. To increase the training accuracy on the time axis, we further divided the time input from 2000 intervals to 5000 intervals from 0 to 0.1. We also reduced the training rate from 0.1 to 0.001, which helped improve accuracy. Since this is unsupervised learning, the presented initial conditions must also be correct. When Equation (1) is trained by PINN, this can produce a memristor state as demonstrated in Figure 2 . Figure 2 a is the state described by the numerical solution data of Equation (1), and Figure 2 b is the state produced from PINN training data. In the GMMS model, the memristor state approaches 1 as time and voltage increase. Figure 3 a,b show the error between Figure 2 a,b, which are numerical solution data and PINN training data, respectively. The calculated maximum error is 0.018, and the MSE is 5 × 10 − 6 , giving a training accuracy of 98.5%. The loss function is a measure of how well the neural network is performing during training. The loss function is calculated via Equations (14), (15) and (17). The loss function in Figure 4 shows a significant reduction in loss in the early epochs and is optimized by the optimization algorithm. Finally, a loss of 6 × 10 − 7 is obtained, indicating that learning is being performed very well. The weights and biases are extracted through the PINN learning method. These values are used to predict the I–V characteristic curve by substituting them into Equations (18)–(20). To calculate the current and voltage, the conductance of the memristor was calculated first, as shown in Figure 5 . The input of simulation is a sin waveform with an amplitude of 0.2 for 0.1 s at different frequencies 10 Hz, 100 Hz, and 1 kHz. Figure 6 shows the I–V characteristic curves obtained from Verilog-A circuit simulation [ 23 , 39 ]. The dotted line represents the numerical solution data and the solid line represents the PINN training data. Figure 6 a–c show the results when the input signals have a frequency of 10 Hz, 100 Hz and 1 kHz, respectively. Accurate I–V characteristic curves were predicted for input frequencies under various conditions. The current–voltage characteristics of the GMMS model can be influenced by variations in R O N and R O F F , as shown in Table 1 . If R O N and R O F F decrease, the set current and reset current increase, respectively. The simulation for the memristor model of Messaris et al. is performed. Table 2 presents the necessary parameters for training the memristor model of Messaris et al. [ 23 ]. The memristor model of Messaris et al. uses PINN to train the positive and negative regions, respectively. When the differential equations for resistance in the memristor model of Messaris et al. are solved using PINN, there is an issue with values becoming excessively large. This disrupts the proper progress of training in the neural networks. To address this issue, we proceed with training by applying the normalization method mentioned in Section 2 . This method transforms resistance into an equation related to the resistive state. The resistive state in the positive region is illustrated in Figure 7 . Figure 7 a,b represent numerical solution data and PINN training data, respectively. In the positive region, the error between the numerical solution data in Figure 7 a and the PINN training data in Figure 7 b is depicted in Figure 8 . There are many inflection points, but the maximum error is very small at 0.001. The resistive state in the negative region is depicted in Figure 9 . Figure 9 a represents the numerical solution data and Figure 9 b represents the PINN training data. Figure 10 shows the error between the numerical solution data and PINN training data in the negative region. The maximum error rate is 0.006. In contrast to the positive region, the negative region exhibits more inflection points. The PINN training loss for the memristor model of Messaris et al. is depicted in Figure 11 . Figure 11 a,b represent the learning loss in the positive region and the negative region, respectively. The memristor model of Messaris et al. was set up with 1000 epochs and a learning rate of 0.01. The loss for the positive region and the negative region converges to the level of approximately 10 − 6 and 10 − 4 , respectively. The simulation uses the same input as the GMMS model, consisting of a 10 Hz, 100 Hz, and 1 kHz sin waveform. The conductance can be calculated using the current equation of Equation (9) for the memristor model of Messaris et al., as illustrated in Figure 12 . Other models [ 19 ] show the same magnitude of conductance in the positive and negative regions for a sin waveform input. However, the magnitude of the positive and negative regions is not equal due to the four fitting parameters given in Equation (9) of the memristor model of Messaris et al. The I–V characteristic curve for the memristor model of Messaris et al. is depicted in Figure 13 . The difference between the cases at 10 Hz and 100 Hz frequencies is not observed clearly. However, it is possible to analyze the differences by enlarging the I–V characteristic curve, as shown in the inset pictures of Figure 13 . At 10 Hz, one hysteresis curve is generated, while the number of hysteresis curves at 100 Hz and 1 kHz is ten and one hundred, respectively. To verify accurate current predictions at high frequencies, the square pulses with a frequency of 1 MHz (pulse width 1 μ s ) are applied to the GMMS and Messaris models, as shown in Figure 14 a,b. It can be observed that accurate current predictions are achieved even with high frequency square pulses. Stochastic non-ideal properties of memristors, such as noise [ 40 ] and variability [ 41 ], can be modeled in the form of differential equations [ 42 , 43 ]. The methodology in this paper can facilitate the stochastic modeling of memristors."
} | 3,619 |
39010147 | PMC11251334 | pmc | 3,092 | {
"abstract": "Background Amidst the escalating carbon dioxide levels resulting from fossil fuel consumption, there is a pressing need for sustainable, bio-based alternatives to underpin future global economies. Single-carbon feedstocks, derived from CO 2 , represent promising substrates for biotechnological applications. Especially, methanol is gaining prominence for bio-production of commodity chemicals. Results In this study, we show the potential of Komagataella phaffii as a production platform for itaconic acid using methanol as the carbon source. Successful integration of heterologous genes from Aspergillus terreus ( cadA , mttA and mfsA ) alongside fine-tuning of the mfsA gene expression, led to promising initial itaconic acid titers of 28 g·L −1 after 5 days of fed-batch cultivation. Through the combined efforts of process optimization and strain engineering strategies, we further boosted the itaconic acid production reaching titers of 55 g·L −1 after less than 5 days of methanol feed, while increasing the product yield on methanol from 0.06 g·g −1 to 0.24 g·g −1 . Conclusion Our results highlight the potential of K. phaffii as a methanol-based platform organism for sustainable biochemical production. Supplementary Information The online version contains supplementary material available at 10.1186/s13068-024-02541-1.",
"conclusion": "Conclusion Overall, the success of our IA production process using K. phaffii as a platform is attributed to the combined efforts of metabolic engineering and process engineering, emphasizing the importance of both aspects in developing efficient and sustainable production processes. Although our strain cannot yet compete with industrially relevant yields of 0.58 g·g glucose −1 produced by A. terreus [ 33 ] in terms of carbon conversion, one must consider the origin of the carbon, as the use of glucose in the current industrial process is unfavorable for a sustainable development of bio-based production. Our process establishes for the first time a high-yield MeOH-based production system with a high IA production rate of 0.45 g·L −1 ·h −1 . This high production rate in combination with the advantages of MeOH as carbon source, the additional safety of K. phaffii as a non-pathogenic, well-established production host and the minimal by-product formation streamlining downstream purification highlight the relevance of our process. With additional metabolic and process engineering advancements, K. phaffii is poised to emerge as a significant industrial production host for IA utilizing MeOH.",
"introduction": "Introduction Due to the continuous exploitation of fossil fuels, the atmospheric content of carbon dioxide is rapidly increasing [ 1 ], and it becomes clear that alternative, sustainable, bio-based products and feedstocks must form the foundation of the global economy in the future. Efforts are underway to pioneer novel technologies and production platforms for products traditionally produced from gas, coal or oil [ 2 , 3 ]. Many of these advancements use microbial systems to enable conversion of organic materials into biofuels as well as chemicals with a wide range of complexity [ 4 – 8 ]. The replacement of fossil resources with renewable first-generation feedstocks, however, raises social and ethical concerns due to the high demand for alternative food and feed sources driven by the growing world population [ 9 , 10 ]. Naturally, this has prompted greater technological investment in alternative substrates, aiming for a neutral carbon footprint and sidestepping ethical implications. Sustainable resources of interest include various waste products such as crude glycerol [ 11 ], lignocellulosic biomass [ 12 ], food waste [ 13 ] and more unconventionally single-carbon (C1) substrates derived from CO 2 [ 14 ]. Valorising CO 2 and its derived C1 substrates is a growing area of interest, being developed in various biotechnological [ 4 , 15 – 17 ] and chemical fields [ 18 ]. Methanol (MeOH), a low-cost renewable single-carbon feedstock, is gaining attention for bio-production of commodity chemicals. With the additional interest in MeOH, production capacity has increased to 174 million metric tons by 2022 and is steadily increasing [ 19 ]. Traditionally sourced from syngas, MeOH can now also be generated from methane and CO 2 , offering a means of sequestering greenhouse gasses to support a sustainable bio-economy [ 20 ]. Methylotrophic organisms, possessing the inherent capability to utilize MeOH as their exclusive carbon and energy source, stand out as promising candidates as platform organisms, with the yeast Komagataella phaffii being one of them [ 20 – 22 ]. Although K. phaffii is widely recognized in the industry for recombinant protein production [ 23 ], recent findings indicate its significant potential in generating platform chemicals [ 24 – 27 ]. One platform chemical gaining an increasing worldwide interest is itaconic acid (IA), an unsaturated dicarboxylic acid, which holds considerable promise as a biochemical building block, as its versatility lies in its ability to serve as a monomer for various products such as resins, plastics, paints, and synthetic fibers [ 28 , 29 ]. Traditionally, microbial production of IA started in 1960 by fermenting Aspergillus terreus in sugar-containing media [ 30 ]. Other microorganisms such as Ustilago sp . , Candida sp. and Rhodotorula sp. have since shown the ability to produce IA; however, A. terreus has remained the predominant industrial production host, achieving titers of up to 150 g·L −1 during a cultivation for 9.7 days [ 13 , 31 – 33 ]. The natural biosynthetic pathway for IA follows a pathway analogous to citric acid formation, involving the tricarboxylic acid cycle (TCA). In A. terreus , IA is produced from cis -aconitate by the key enzyme aconitate decarboxylase, CadA, residing in the cytosol [ 34 – 36 ]. Further two transporters are involved in shuttling either the precursor cis -aconitate from the mitochondria to the cytosol, i.e., the mitochondrial cis -aconitate transporter (MttA) or the product itself to the extracellular space, i.e., the major facilitator superfamily transporter (MfsA) [ 36 , 37 ]. The commercial production of IA by A. terreus has improved since its establishment [ 31 , 38 , 39 ], but it mainly relies on monosaccharides that could also be utilized as food or feed. Alternative, more sustainable substrates for production with A. terreus or alternative hosts have long been under investigation, with MeOH being one of the least explored [ 30 , 38 , 40 , 41 ]. Production using the C1-substrate could potentially lower manufacturing costs, contribute to a circular economy and improve the overall production sustainability [ 15 , 16 , 20 ]. Hence, K. phaffii as a natural methylotroph and established production host in biotechnology is a promising candidate for producing itaconic acid from MeOH. In this study, we explored the potential of K. phaffii as a production organism for IA using MeOH as the primary carbon source. For that, the heterologous genes cadA , mttA as well as mfsA from A. terreus were successfully engineered into K. phaffii . The engineered pathway is illustrated in Fig. 1 . Upon small-scale preliminary experiments in 24-deep-well plates and shake flasks, fed-batch cultivations were carried out in lab-scale reactors, where optimal results were achieved by combining strain engineering with process engineering. Fig. 1 The engineered metabolic pathway for itaconic acid production from methanol in K. phaffii . CadA: \n cis -aconitate decarboxylase, MttA : mitochondrial cis- aconitate transporter, MfsA: Major Facilitator Superfamily transporter of itaconic acid",
"discussion": "Discussion Traditionally, itaconic acid (IA) production has primarily been associated with filamentous fungi such as Aspergillus spp. or Ustilago maydis using glucose as a carbon source [ 39 , 58 , 59 ]. However, the search for alternative hosts has been ongoing, driven by the need for more sustainable production methods and the exploration of diverse substrates. In this study, we show that K. phaffii can produce IA from methanol (MeOH) reaching titers between 50 and 55 g·L −1 . This was achieved by a combinatorial approach focusing on strain engineering and process optimization. In several natural non-producing organisms, overexpression of cadA has enabled IA production from different carbon substrates, including C1 feed stocks [ 58 ]. Just to mention a few, cadA expression enabled Methylorubrum extorquens AM1 IA production from MeOH reaching titers of 31.6 mg·L −1 [ 60 ], and autotrophic IA production from CO 2 was enabled in strains such as Synechocystis sp. PCC6803 and a synthetic autotrophic K. phaffii strain, yielding 14.5 mg·L −1 and 530 mg·L −1 , respectively [ 49 , 61 ]. Also in our hands, expression of cadA enabled IA production in the range of 0.23 g·L −1 on MeOH in K. phaffii. The production capacity was, however, boosted with the coexpression of the MttA transporter, reaching titers of 1.8 g·L −1 . Expression of mttA creates possibly a pull effect of cis -aconitate from the mitochondria to the cytosol, thereby providing more precursor molecules for conversion by CadA to IA [ 62 , 63 ]. Further, we demonstrated that by incorporating an A. terreus gene encoding a plasma membrane transporter, mfsA , under the control of a strong promoter, IA titers of 3.0 g·L −1 were reached in shake flask cultivations. The pivotal role of both transporters has previously been underscored by Wierckx et al. [ 64 ], a finding corroborated by our observations. Furthermore, our results suggested that utilizing a strong promoter for mfsA proves advantageous for IA production. Subsequent toxicity screenings revealed that MfsA expression also improved growth and viability of K. phaffii at high IA concentrations. This phenomenon is likely attributed to the efficient export of IA by the cell, thereby mitigating weak organic acid-induced stress within the cellular environment. Upscaling the process to bioreactor cultivations was successful, reaching IA titers of 28.2 g·L −1 with the strain expressing a single copy of all three heterologous genes cadA , mttA and mfsA . However, excessive biomass accumulation was a significant side effect, hence process optimization was required. The combination of nitrogen limitation and an increase in process temperature from 25 to 30 °C resulted in a twofold increase in the MeOH yield for the cadA + mttA + mfsA pGAP strain. In addition, it led to a 13% reduction in biomass accumulation, favoring the conversion of MeOH to IA over biomass production. A substantial increase in IA production was anticipated, considering the prior evaluation of CadA activity [ 35 ]. Here, in vitro studies revealed a positive correlation between enzyme activity and temperature, peaking at 42 °C in A. terreus. Further process optimization of the multicopy strains, exploring temperatures ranging from 28 to 34 °C, revealed that the engineered strain exhibited optimal performance within the temperature range of 30–32 °C. Notably, cellular metabolism was adversely impacted by 34 °C leading to decreased viability over time and less efficient MeOH utilization compared to a temperature of 30 °C. This decline in cell viability is particularly undesirable for downstream processing of the fermentation broth, as reduced viability results in increased cell lysis and thus presence of unwanted cell debris [ 65 ]. The absence of metabolic by-products during all fed-batch cultivations was confirmed with HPLC analysis, which would further simplify downstream isolation of IA. Next to this, we could show that increasing the number of all heterologous genes markedly improved the potential of K. phaffii as an IA production host, as it increased IA productivity (0.45 g·L −1 ·h −1 ) simultaneously with an efficient conversion of MeOH into IA (0.26 g·g −1 ). In several organisms [ 54 , 55 ], the integration of multiple copies of cadA has proven beneficial for production. The instability of CadA was previously described [ 31 , 66 ], though whether the necessity for additional cadA expression is linked to this instability remains unanswered at this point. However, our findings suggest that the integration of additional cadA copies had a minimal impact on IA production in K. phaffii , unless it coincides with the multicopy integration of mttA and mfsA . The heightened expression of these two transport proteins enhanced the pull effect. Notably, strains harboring a high mttA GCN showed severe growth impairments in comparison to the remaining multicopy strains. Similar results were obtained by Baumschabl et al. [ 49 ], indicating that a strong promoter regulation for mttA expression severely impacts growth in the autotrophic K. phaffi i strain. A high MttA activity might efficiently drain the TCA cycle from its intermediate and result in accumulation of IA in the cytosol, thus possibly inflicting acid-induced stress within the cell. The additional expression of mfsA in the cadA + mttA + mfsA pGAP and MC_cadA + mttA + mfsA strains possibly compensates for this phenotype as IA-toxification becomes less likely due to the increased export."
} | 3,328 |
30429482 | PMC6235870 | pmc | 3,093 | {
"abstract": "Web spiders synthesize silk fibers of unique strength and extensibility through the controlled self-assembly of protein building blocks, so-called spidroins. The spidroin C-terminal domain is highly conserved and connects two polypeptide chains through formation of an all-helical, intertwined dimer. Here we use contact-induced fluorescence self-quenching and resonance energy transfer in combination with far-UV circular dichroism spectroscopy as three orthogonal structural probes to dissect the mechanism of folding and dimerization of a spidroin C-terminal domain from the major ampullate gland of the nursery web spider Euprosthenops australis . We show that helices forming the dimer core assemble very rapidly and fold on association. Subsequently, peripheral helices fold and dock slowly onto the preformed core. Lability of outer helices facilitates formation of a highly expanded, partially folded dimer. The high end-to-end distance of chain termini in the partially folded dimer suggests an extensibility module that contributes to elasticity of spider silk.",
"introduction": "Introduction Web spiders use up to seven specialized glands to synthesize silk fibers of outstanding mechanical properties tailored for various tasks including prey capture, reproduction, and shelter 1 . Spider silk is characterized by a unique combination of strength and extensibility, which results in toughness that supersedes that of man-made threads 2 – 4 . At the beginning of the process the constituting protein building blocks, so-called spidroins, are stored in soluble form and at high concentration in the ampulla of the spinning gland, which is located in the spider’s abdomen. On demand, silk is formed by the controlled self-association of spidroins within a tapering spinning duct. Chemical and mechanical stimuli within the duct induce phase and structural transitions of spidroins that transform into silk 1 , 3 , 4 . Spidroins from the major ampullate (Ma) gland form dragline silk, which is the toughest fiber and used by spiders as a lifeline or to build the web frame. Ma silk is a current focus of biomimetic material sciences 3 , 4 . The central segment of Ma spidroins consists of repetitive peptide motifs of simple amino acid composition involving alanine-, glycine-, and proline-rich stretches, which are unstructured under storage conditions and form mainly β-sheet structure in fibers 4 . The N- and C-terminal domains (NTD and CTD) flank the repetitive core domain and are highly conserved across glands and species. Conservation underscores their important functional roles in silk. A recent genomic study shows that sequences of CTDs are slightly more diverse compared to those of NTDs 5 . In particular, sequences from the pyriform and aggregate glands differ from those of other glands, which may reflect the fact that pyriform and aggregate silk has adhesive rather than fiber-forming function 5 . Both NTD and CTD are regularly folded five-helix bundles that provide solubility and connectivity of spidroins 6 – 8 . The NTD forms an antiparallel dimer, a process that is induced by a change of pH and salt composition along the spinning duct. A comprehensive body of biophysical work shows that pH-triggered dimerization of NTDs involves changes of protein surface electrostatics and conformation 3 , 6 , 8 – 14 . The CTD, on the other hand, forms a constitutive, parallel dimer already in the ampulla of the gland and connects two spidroin chains permanently. Although NTD and CTD share the same secondary structure they are very different on the level of tertiary and quaternary structure, highlighting their different functional roles. In contrast to the NTD, the fold of the CTD is intertwined: the C-terminal ends of each subunit undergo a domain swap and dock onto the core helix of the opposing subunit 7 . The topology of the fold resembles a molecular clamp. Parallel orientation of CTD subunits aligns the appended, repetitive segments, prestructuring and priming them for shear-induced phase and structural transitions within the duct 7 , 15 . Ma spidroin CTDs contain a single, conserved cysteine that forms a disulfide connecting two core helices of the dimer interface covalently. Structural studies show that the clamp-like fold is highly conserved across glands and species 16 – 18 despite the lack of a disulfide in CTDs of non-Ma spidroins 19 , 20 . Conservation of structure underscores a common functional role. Here, we investigate folding/unfolding transitions associated with self-assembly of the CTD of spidroin 1 from the Ma gland (MaSp1) of the nursery web spider Euprosthenops australis . We use site-directed mutagenesis and chemical modification to establish extrinsic fluorescence reporter systems based on resonance energy transfer (FRET) and contact-induced self-quenching tailored to probe specific conformational coordinates. We apply steady-state spectroscopy and kinetic experiments to probe structure, dynamics and energetics of folding and dimerization yielding mechanistic insights. We identify a three-state mechanism of self-assembly that involves a partially folded, dimeric intermediate. In the first step, the unfolded chains of two monomers associate to form the helical core of the dimerization interface at remarkably fast rate constant. In the second step, the peripheral helices fold slowly onto the preformed core. Lability of peripheral helices facilitates formation of a highly expanded structure that may contribute to elasticity of spider silk.",
"discussion": "Discussion Spidroins undergo phase and structural transitions within a spinning duct, which are induced by mechanical and chemical stimuli. The process involves a change of pH from 7.0 to 5.7 and has different effects on spidroin terminal domains 17 . In general, pH values above or below a protein’s isoelectric point (pI) induce repulsive electrostatic forces between correspondent surface charges, which counteract protein–protein association. Most proteins have a pI below 7. As the solution pH approaches the pI, proteins tend to associate because net surface charges are neutralized and repulsive forces vanish. Web spiders seem to make use of this phenomenon to assemble spidroins along a pH gradient within their duct. At very low pH, however, many proteins tend to undergo acid-induced denaturation, which can lead to aggregation 30 . It may therefore not surprise that CTDs are destabilized at pH values below 7 (refs. 15 , 17 and 31 ,). The CTD from the minor ampullate gland of Araneus ventricosus is reported to even form aggregates and β-sheet fibrils at low-pH values between 5.5 and 5.0 17 . MaSp1 CTDs of Latrodectus hesperus and Nephila clavipes form unstructured aggregates at pH 5 and below 15 , 31 . But in vitro aggregation of CTDs commonly takes hours or even longer and, in some cases, is induced only at highly acidic nonphysiological conditions. Web spiders synthesize silk fibers within a fraction of a second 3 . The relevance of observed CTD aggregation for silk formation by the animal is therefore unclear 15 . However, unfolding of the CTD during silking may induce transitions of functional relevance like the formation of β-sheet secondary structure or phase transition. In agreement with previous studies 15 , 17 we found that the MaSp1 CTD from E. australis is destabilized at low pH, but remained fully folded on the physiological time scale (Fig. 1e, f ). Our results showed that self-assembly of the CTD follows a three-state mechanism. This was evident from equilibrium and kinetic experiments where we observed two discrete structural transitions and determined the underlying rate constants and energetics. We deduced a mechanism of formation of the molecular-clamp topology of the native CTD dimer from initially unfolded monomers (Fig. 5a ). In the first step, the unfolded monomers associate to form the core helices of the dimer interface, i.e., the central scaffold built by helices 4. The rate constant of this process was remarkably fast ( k ass = 5.2 ± 0.4 × 10 7 M −1 s −1 ). By comparison, the basal rate constant of protein–protein association is ~1 × 10 5 M −1 s −1 (refs. 32 – 34 ), i.e., ~500 times slower than the value observed here. Theory assumes that two structures containing complementary association interfaces must orient in solution before they can form a productive complex in a lock-and-key fashion 32 , 33 . The entropic constraint of molecular orientation slows association to the basal value 32 . Intriguingly, the CTD dimer interface folds upon association. This was evident from the loss of helical secondary structure observed in the second transition of chemical denaturation data (Fig. 1d ). Since folding and association are cooperative events, the entropic constraint of molecular orientation may drop out and explain fast dimerization. Alternatively, steering forces caused by long-range electrostatics may accelerate association beyond the basal value 33 , 34 . Ultrafast association is also observed for the MaSp1 NTD, which by contrast to the CTD proceeds from folded monomers 12 , 35 . Fig. 5 Reaction scheme of CTD self-assembly and structural model of the intermediate. a Reaction scheme, rate constants and structural model of two-step folding and dimerization of the CTD. The two subunits constituting the dimer are colored blue and cyan. Unstructured coils and helices are depicted in cartoon representation. C-terminal helix 5 in N 2 , which swaps domain and threads through the chain of the opposing subunit, is highlighted red. b Structural model of the dimeric intermediate I 2 . Chain dimensions of unfolded helices 1–3 inferred from FRET spectroscopy and polymer theory, and of the dimer core formed by helices 4 and 5, are indicated. The disulfide that covalently connects the N-terminal ends of helices 4 is highlighted in red stick representation The second step was folding and docking of the peripheral N-terminal helices onto the dimer scaffold. This process was on the surprisingly slow seconds time scale ( k f = 0.78 ± 0.07 s −1 ). Isolated helices are known to typically fold within a microsecond 36 and their docking can occur within a millisecond 37 . How can the dramatic slowing of helix folding and docking be rationalized? In the native CTD helix 5 undergoes a domain swap and threads through the chain of the opposing subunit where it clamps onto helix 4. The structure formed by helices 4 and 5 is the docking interface for N-terminal helices 1–3. The intertwined fold resembles a knotted protein. In knotted proteins, the knotting event limits the time constant of folding to the slow seconds-to-minutes scale 38 – 40 . Two-step folding of the CTD observed here resembles that of a dimeric domain from Helicobacter pylori , which also contains a molecular clamp and was engineered to knot 41 . Folding and docking of helices 1–3 in the CTD are thus likely to be rate-limited by domain swap, which involves threading and clamping of C-terminal helices 4 and 5. The rate constant of CTD dissociation, i.e., of transition I 2 to D ( k diss = 8 ± 1 × 10 −3 s −1 ), matched the rate constant of unfolding of N-terminal helices 1–3, i.e., of transition N 2 to I 2 ( k u = 9 ± 2 × 10 −3 s −1 ). This suggests that undocking/unfolding of peripheral helices limits the rate constant of dimer dissociation. This is reasonable because the peripheral helices have to first dismantle from the core before the chains can untangle to form unfolded monomers. Why has the three-state mechanism of CTD self-assembly gone unnoticed in previous biophysical studies? Previously, unfolding of CTDs was described by a conventional two-state model 7 , 15 , 17 , 18 , 31 . But deviations from two-state behavior are reported: Andersson et al. 17 find that unfolding of a CTD from the minor ampullate gland of A. ventricosus deviates from two-state. Hagn et al. 7 report indications for early unfolding of helix 1 of a CTD from A. diadematus from NMR spectroscopy. Multistate mechanisms are difficult to detect if energetics of transitions are not well separated. In such cases, three-state mechanisms may appear two-state. But a third state reflecting a bimolecular event can be detected from protein concentration-dependent experiments 21 , 41 , 42 . We simulated theoretical denaturation curves using the three-state model applied to our experimental data. In this model we artificially reduced the free energy of dimerization (Δ G I2-D ). As Δ G I2-D drops from the here experimentally determined 14 kcal mol −1 over 12.5 to 11.5 kcal mol −1 , the two discrete unfolding transitions of 10 µM CTD progressively overlay and appear two-state (Fig. 6 ). Fig. 6 Theoretical denaturation curves of the CTD of varying free energy of dimerization. Simulated denaturation curves of 10 µM CTD using the model for a three-state equilibrium involving a dimeric intermediate. The curves contain the thermodynamic parameters derived from the fit to experimental data of wild-type CTD (free energy of dimerization: Δ G I2-D = 14 kcal mol −1 , black line). Simulated denaturation curves using the same model but lowering Δ G I2-D to 12.4 and to 11.5 kcal mol −1 are shown as blue and red line, respectively Elasticity of spider silk is attributed to repetitive, glycine/proline-rich sequence areas that form helical structures in silk 43 , 44 . Infrared spectroscopy finds evidence for helix-to-coil transitions in spider silk upon expansion induced by mechanical strain 45 . But the CTD also plays a role: deletion of CTDs in engineered spidroins impairs extensibility of synthetic fibers 46 . The underlying molecular mechanisms remained elusive. Here we found that peripheral helices of the CTD are very labile (Δ G N2–I2 = 2.6 ± 0.5 kcal mol −1 ) compared to the dimer core (Δ G I2-D = 14 ± 1 kcal mol −1 ). Imposing stress on a spider silk fiber will give rise to a stretching force on the N-terminal ends of a CTD dimer where the central, repetitive spidroin domains append. This stretching force may induce early unfolding of labile CTD helices 1–3. We estimated expansion of the CTD upon unfolding of its N-terminal helices. The loss of helical content associated with the transition N 2 to I 2 determined by CD spectroscopy was 58 ± 3%. This value was in good agreement with the fractional secondary structure of 54% formed by helices 1–3, as seen in the structure 7 . The Tanford β value 47 of the folding of I 2 to N 2 , revealed by chevron analysis, was low ( β T = m f /( m f + m u ) = 0.46). This showed that the transition state of folding of helices 1–3 was expanded and denatured-like, supporting an expanded N-terminus in I 2 . FRET spectroscopy showed that the N-termini in I 2 exhibited an end-to-end distance of >10 nm (Fig. 2 ). The average end-to-end distance 〈 r 2 〉 of an unfolded polypeptide chain can be estimated using polymer theory 48 . Helices 1–3 comprise a 64-residue polypeptide segment. An unfolded 64-residue segment has an end-to-end distance of ~9 nm (〈 r 2 〉 = C ∞ nl 2 ) 48 . Here, C ∞ is the characteristic ratio, which reflects chain stiffness and takes a value of ~9 for long, generic polypeptides 49 , n is the number of peptide bonds, and l is the distance between two neighboring backbone amide bonds in a polypeptide ( l = 0.38 nm). From polymer theory we can thus estimate that the two unfolded chain segments in I 2 form an extended structure of ~21 nm end-to-end distance (Fig. 5b ). This corresponds to a ~sevenfold expansion of the CTD in state I 2 compared to in state N 2 . A linear and fully stretched 64-residue polypeptide, which may be formed at high tensile stress just before the fiber breaks, has a length of nl = 24 nm. This corresponds to a 17-fold expansion of maximally stretched I 2 compared to N 2 . In Ma silk, further expansion is blocked by the covalent disulfide between helices 4 (Fig. 5b ). In conclusion, we identified a two-step mechanism of self-assembly of the CTD. Fast folding and dimerization of central helix 4 is followed by slow domain swap and threading of the C-terminal helix forming the scaffold for docking of the N-terminal helices. The CTD self-assembles in the ampulla of the gland where the residence time of spidroins is sufficiently long for slow folding. On the other hand, slow kinetics of helix docking gives rise to their lability, which could facilitate partial unfolding in solid fibers induced by mechanical stress. The CTD may thus act as extensibility module that, in addition to helical central spidroin segments, contributes to elasticity of spider silk."
} | 4,155 |
22550959 | PMC3527305 | pmc | 3,094 | {
"abstract": "Background Poly(4-hydroxybutyrate) [poly(4HB)] is a strong thermoplastic biomaterial with remarkable mechanical properties, biocompatibility and biodegradability. However, it is generally synthesized when 4-hydroxybutyrate (4HB) structurally related substrates such as γ-butyrolactone, 4-hydroxybutyrate or 1,4-butanediol (1,4-BD) are provided as precursor which are much more expensive than glucose. At present, high production cost is a big obstacle for large scale production of poly(4HB). Results Recombinant Escherichia coli strain was constructed to achieve hyperproduction of poly(4-hydroxybutyrate) [poly(4HB)] using glucose as a sole carbon source. An engineering pathway was established in E. coli containing genes encoding succinate degradation of Clostridium kluyveri and PHB synthase of Ralstonia eutropha. Native succinate semialdehyde dehydrogenase genes sad and gabD in E. coli were both inactivated to enhance the carbon flux to poly(4HB) biosynthesis. Four PHA binding proteins (PhaP or phasins) including PhaP1, PhaP2, PhaP3 and PhaP4 from R. eutropha were heterologously expressed in the recombinant E. coli, respectively, leading to different levels of improvement in poly(4HB) production. Among them PhaP1 exhibited the highest capability for enhanced polymer synthesis. The recombinant E. coli produced 5.5 g L -1 cell dry weight containing 35.4% poly(4HB) using glucose as a sole carbon source in a 48 h shake flask growth. In a 6-L fermentor study, 11.5 g L -1 cell dry weight containing 68.2% poly(4HB) was obtained after 52 h of cultivation. This was the highest poly(4HB) yield using glucose as a sole carbon source reported so far. Poly(4HB) was structurally confirmed by gas chromatographic (GC) as well as 1 H and 13 C NMR studies. Conclusions Significant level of poly(4HB) biosynthesis from glucose can be achieved in sad and gabD genes deficient strain of E. coli JM109 harboring an engineering pathway encoding succinate degradation genes and PHB synthase gene, together with expression of four PHA binding proteins PhaP or phasins, respectively. Over 68% poly(4HB) was produced in a fed-batch fermentation process, demonstrating the feasibility for enhanced poly(4HB) production using the recombinant strain for future cost effective commercial development.",
"conclusion": "Conclusion In summary, Escherichia coli strain JM109 harboring an engineering pathway encoding succinate degradation genes of Clostridium kluyveri and PHB synthase gene of Ralstonia eutropha together with its native succinate semialdehyde dehydrogenase genes sad and gabD inactivated, was able to achieve significant level of poly(4HB) biosynthesis from glucose. Additional expression of four PHA binding proteins PhaP or phasins in the recombinant strain, respectively, led to a further improvement of poly(4HB) accumulation. PhaP1 was found most useful among the four PhaPs used. Over 68 wt% poly(4HB) was produced in a fed-batch fermentation process, demonstrating the feasibility for enhanced poly(4HB) production using the recombinant strain for future cost effective commercial development.",
"discussion": "Discussion As a strong pliable thermoplastic material with good flexibility, poly(4HB) has been approved by FDA as a suture material ( http://www.tepha.com ). Biomedical applications are usually not sensitive to high cost. However, a reduction on poly(4HB) production cost should allow for more application exploitation. High production cost for poly(4HB) comes from expensive 4HB precursors including 4-hydroxybutyric acid, γ-butyrolacton or 1,4-butanediol [ 41 , 42 ], and from very low yield of poly(4HB) by recombinant bacteria. Therefore, simple and low cost substrates as well as a highly productive strain can help reduce poly(4HB) production cost. The anaerobic succinate degradation pathway employed in this study conferred on the recombinant E. coli the ability to utilize glucose as a sole carbon source for poly(4HB) production. The additional expression of PHA granule associate protein PhaP provided a further enhancement on poly(4HB) yield, allowing for further fermentor exploitation. While in wild E. coli strain, succinate semialdehyde can be degraded to succinate by SSA dehydrogenase (SSADH) encoded by sad and gabD [ 30 ] , leading to a decreased metabolic flux to 4HB production (Figure 1 ). To channel more flux to 4HB, the native SSADH genes of E. coli were inactivated in the poly(4HB) producing recombinant. Shake flask studies of E. coli JM109 and its SSADH deficient mutant JM109SG harboring pMCSH5 and pKSSE5.3 showed that inactivation of SSADH genes significantly improved poly(4HB) synthesis compared with the wild strain JM109 which had no poly(4HB) production at all (Table 1 ). On the other hand, the highest 4HB molar fraction in P3HB4HB synthesized from glucose in E. coli was 11% reported so far [ 28 ]. Our result indicated that the recombinant enzymes in this pathway were active enough to provide sufficient 4HB from glucose for polymerization. Expression of all four PhaPs (phasin) cloned from R. eutropha provided additional improvement on poly(4HB) accumulation in the order of PhaP1 > PhaP3 > PhaP2 > PhaP4 (Table 1 ). The differences of their different influences are not clear yet but probably due to the different roles of PhaP played on PHA granules formation. PhaP1 was the major phasin with the highest expression amount in R. eutropha while PhaP2, PhaP3 and PhaP4 were small in quantity [ 39 , 40 ], indicating its dominating function for PHA granule formation, and PhaP3 was expressed at a significantly high level in PhaP1 deficient strains, other PhaPs were in much lower levels. Our results therefore suggested that the poly(4HB) yields were positively related to the expression levels of PhaP. The recombinant E. coli JM109SG (pKSSEP1, pMCSH5) grown to 12 g L -1 CDW under a well-controlled fermentor run in a fed-batch process accumulated over 68% poly(4HB) using glucose as the only carbon source over a 52 h period (Figure 3 ). This is by far the highest yield for poly(4HB). In its exponential growth period of 8–24 h after innoculation, poly(4HB) content increased most rapidly and reached a relatively stable level when cells entered the stationary phase. As in the exponential phase, TCA cycle is most active, supplying the most succinyl-CoA for the poly(4HB) synthesis, leading to a rapid poly(4HB) accumulation rate. A continuous fermentation process that maintains the cells in their exponential growth phase may further improve poly(4HB) accumulation level."
} | 1,643 |
16524470 | PMC1434774 | pmc | 3,095 | {
"abstract": "Background Recently there has been a lot of interest in identifying modules at the level of genetic and metabolic networks of organisms, as well as in identifying single genes and reactions that are essential for the organism. A goal of computational and systems biology is to go beyond identification towards an explanation of specific modules and essential genes and reactions in terms of specific structural or evolutionary constraints. Results In the metabolic networks of Escherichia coli, Saccharomyces cerevisiae and Staphylococcus aureus , we identified metabolites with a low degree of connectivity, particularly those that are produced and/or consumed in just a single reaction. Using flux balance analysis (FBA) we also determined reactions essential for growth in these metabolic networks. We find that most reactions identified as essential in these networks turn out to be those involving the production or consumption of low degree metabolites. Applying graph theoretic methods to these metabolic networks, we identified connected clusters of these low degree metabolites. The genes involved in several operons in E. coli are correctly predicted as those of enzymes catalyzing the reactions of these clusters. Furthermore, we find that larger sized clusters are over-represented in the real network and are analogous to a 'network motif. Using FBA for the above mentioned three organisms we independently identified clusters of reactions whose fluxes are perfectly correlated. We find that the composition of the latter 'functional clusters' is also largely explained in terms of clusters of low degree metabolites in each of these organisms. Conclusion Our findings mean that most metabolic reactions that are essential can be tagged by one or more low degree metabolites. Those reactions are essential because they are the only ways of producing or consuming their respective tagged metabolites. Furthermore, reactions whose fluxes are strongly correlated can be thought of as 'glued together' by these low degree metabolites. The methods developed here could be used in predicting essential reactions and metabolic modules in other organisms from the list of metabolic reactions.",
"discussion": "Discussion and Conclusion In this paper we have observed that the lowest degree metabolites are implicated in two distinct properties of the metabolic networks, one, the existence of essential metabolic reactions (and lethal single metabolic gene knockouts), and two, existence of functional clusters in the metabolic networks (and associated regulatory modules). To some extent the identification of UP/UC metabolites depends on the way the metabolic network is curated. For example, the networks we have used leave out certain non-enzymatic reactions such as protonation-deprotonation reactions. Since their inclusion would render some of the presently UP(UC) metabolites non-UP(UC), our definition of UP(UC) could be criticized as being somewhat arbitrary. In this context it is worth noting that for the networks as they stand, our definition of UP(UC) allows us to establish a connection between distinct properties of the network (e.g., between essentiality, a functional property and the UP/UC character, a topological property), and that our main findings hold for metabolic networks of three distinct organisms. This suggests that UP/UC reactions as defined by us do capture a certain pattern. In our view the important point is not that other definitions of the network would obscure the pattern, but rather, that there do exist systematic definitions of the network in which a pattern is visible. In metabolic networks the very existence of essential reactions is an indicator of the fragility of the system: Even though the network has many reaction nodes, the removal of a single essential reaction node destroys the functionality of the network completely by blocking the flow of an essential intermediate. Isozymes are a way of dealing with this fragility. However, not all essential reactions have isozymes [ 39 ]; this means that evolution has tolerated this fragility. Our finding that essential reactions are tagged by low degree metabolites may provide some insight into why this is the case. Metabolites that participate in very few reactions perhaps do so in part because some feature of their chemical structure prohibits ready association with other molecules, i.e., their low degree is a consequence of constraints coming from chemistry. Then evolution tolerates the reactions that produce or consume such metabolites as essential because chemistry leaves it no choice. Alternatively, it could be that this fragility happens to be a byproduct of some other desirable structural property that contributes to robustness or evolvability, such as modularity. We have drawn attention to the fact that low degree metabolites also play a role in functional clustering of reactions in the metabolic network. We have further provided evidence that the UP-UC clusters at the metabolic level correspond, with a high probability, to sets of genes forming modules at the regulatory level in E. coli . This raises the question: if low degree metabolites contribute to modularity, could it be that the evolutionary advantages of that have outweighed the disadvantage of the above mentioned fragility caused by the same low degree metabolites? Is it the case that evolution has preferred 'chemically constrained' low degree metabolites in spite of the fragility they cause because they contribute to modularity? A goal in biology is to understand highly evolved biological organization in terms of simpler and more inevitable structures [ 40 ]. Here we have presented evidence that certain genetic regulatory modules, in particular certain operons, mirror the low degree structure of the metabolites whose production and consumption they regulate. This could be an example of how the origin of certain regulatory structure can be traced to simple chemical constraints."
} | 1,494 |
19909344 | null | s2 | 3,097 | {
"abstract": "The PhoBR regulatory system is required for the induction of multiple genes under conditions of phosphate limitation. Here, we examine the role of PhoB in biofilm formation and environmental stress response in Vibrio cholerae of the El Tor biotype. Deletion of phoB or hapR enhanced biofilm formation in a phosphate-limited medium. Planktonic and redispersed biofilm cells of the DeltaphoB mutant did not differ from wild type for the expression of HapR, suggesting that PhoB negatively affects biofilm formation through an HapR-independent pathway. The DeltaphoB mutant exhibited elevated expression of exopolysaccharide genes vpsA and vpsL compared with the wild type. Deletion of hapR enhanced the expression of the positive regulator vpsT, but had no effect on the expression of vpsR. In contrast, deletion of phoB enhanced the expression of the positive regulator vpsR, but had no effect on the expression of hapR and vpsT. The DeltaphoB mutant was more sensitive to hydrogen peroxide compared with the wild type and with an isogenic DeltarpoS mutant. Conversely, the DeltaphoB mutant was more resistant to acidic conditions and high osmolarity compared with the wild type and with an isogenic DeltarpoS mutant. Taken together, our data suggest that phosphate limitation induces V. cholerae to adopt a free-swimming life style in which PhoB modulates environmental stress response in a manner that differs from the general stress response regulator RpoS."
} | 364 |
39814778 | PMC11735844 | pmc | 3,098 | {
"abstract": "Trait-based approaches are revolutionizing our understanding of high-diversity ecosystems by providing insights into the principles underlying key ecological processes, such as community assembly, species distribution, resilience, and the relationship between biodiversity and ecosystem functioning. In 2016, the Coral Trait Database advanced coral reef science by centralizing trait information for stony corals (i.e., Subphylum Anthozoa, Class Hexacorallia, Order Scleractinia). However, the absence of trait data for soft corals, gorgonians, and sea pens (i.e., Class Octocorallia) limits our understanding of ecosystems where these organisms are significant members and play pivotal roles. To address this gap, we introduce the Octocoral Trait Database, a global, open-source database of curated trait data for octocorals. This database houses species- and individual-level data, complemented by contextual information that provides a relevant framework for analyses. The inaugural dataset, OctocoralTraits v2.2, contains over 97,500 global trait observations across 98 traits and over 3,500 species. The database aims to evolve into a steadily growing, community-led resource that advances future marine science, with a particular emphasis on coral reef research."
} | 317 |
35664929 | PMC9149024 | pmc | 3,099 | {
"abstract": "Electroactive microorganisms (EAMs) play important roles in biogeochemical redox processes and have been of great interest in the fields of energy recovery, waste treatment, and environmental remediation. However, the currently identified EAMs are difficult to be widely used in complex and diverse environments, due to the existence of poor electron transfer capability, weak environmental adaptability, and difficulty with engineering modifications, etc. Therefore, rapid and efficient screening of high performance EAMs from environments is an effective strategy to facilitate applications of microbial fuel cells (MFCs). In this study, to achieve efficient degradation of methyl orange (MO) by MFC and electricity harvest, a more efficient exoelectrogen Shewanella carassii- D 5 that belongs to Shewanella spp . was first isolated from activated sludge by WO 3 nanocluster probe technique. Physiological properties experiments confirmed that S. carassii- D 5 is a Gram-negative strain with rounded colonies and smooth, slightly reddish surface, which could survive in media containing lactate at 30 °C. Moreover, we found that S. carassii- D 5 exhibited remarkable MO degradation ability, which could degrade 66% of MO within 72 h, 1.7 times higher than that of Shewanella oneidensis MR-1. Electrochemical measurements showed that MFCs inoculated with S. carassii- D 5 could generate a maximum power density of 704.6 mW/m 2 , which was 5.6 times higher than that of S. oneidensis MR-1. Further investigation of the extracellular electron transfer (EET) mechanism found that S. carassii -D 5 strain had high level of c -type cytochromes and strong biofilm formation ability compared with S. oneidensis MR-1, thus facilitating direct EET. Therefore, to enhance indirect electron transfer and MO degradation capacity, a synthetic gene cluster ribADEHC encoding riboflavin synthesis pathway from Bacillus subtilis was heterologously expressed in S. carassii- D 5 , increasing riboflavin yield from 1.9 to 9.0 mg/g DCW with 1286.3 mW/m 2 power density output in lactate fed-MFCs. Furthermore, results showed that the high EET rate endowed a faster degradation efficient of MO from 66% to 86% with a maximum power density of 192.3 mW/m 2 , which was 1.3 and 1.6 times higher than that of S. carassii- D 5 , respectively. Our research suggests that screening and engineering high-efficient EAMs from sludge is a feasible strategy in treating organic pollutants.",
"conclusion": "4 Conclusions In summary, a Shewanella carassii- D 5 with strong bioelectricity generation capacity was isolated from activated sludge by the electrochromism of WO 3 nanoclusters, which could reach a maximum power density of 704.6 mW/m 2 , ∼5.6 times higher than that of the model electroactive strain S. oneidensis MR-1 (125.0 mW/m 2 ). Subsequently, two EET mechanisms of S. carassii -D 5 were further identified, based on the analysis of electrophysiological indicators including the amount of c -type cytochromes, adhesion biomass, and riboflavin. Notably, these results demonstrated that S. carassii- D 5 had a higher density of cytochromes and stronger biofilm formation ability than S. oneidensis MR-1, thus facilitating direct EET. Moreover, to enhance the indirect extracellular electron transfer of S. carassii- D 5 , riboflavin synthesis was further improved from 1.9 to 9.0 mg/g DCW by overexpression of the ribADEHC gene cluster from B. subtili . The maximum power density of S. carassii- D 5 -C 5 reached 1286.3 mW/m 2 , which was 1.8 times higher than that of S. carassii -D 5 (704.6 mW/m 2 ). For the purpose of promoting the degradation of azo dyes and bioelectricity generation based on the above study, the MFC with S. carassii- D 5 -C 5 as the anode biocatalyst and MO as the terminal electron acceptor was subsequently constructed. The MO degradation rate of S. carassii- D 5 -C 5 increased from 66% to 86%, while the maximum power density improved by 1.6 times from 117.9 to 192.3 mW/m 2 , respectively. In conclusion, the S. carassii -D 5 strain isolated from sludge shows superior organic matter degradation and bioelectricity generation ability, which provides a new possibility for MFCs in pollutant degradation and electricity recovery.",
"introduction": "1 Introduction Environmental pollution and energy crisis have become serious challenges for global sustainable development. Domestic, industrial, and farming wastewaters are essential categories of environmental pollutants that contain enormous amount of chemical energy stored in organics [ 1 ]. Nevertheless, these energies are often largely lost or dissipated in the actual treatment process. Recovering potential energy meanwhile treating wastewater by cost-effective technologies can contribute to relieving energy shortage and environmental pollution issues [ 2 ]. Microbial fuel cells (MFCs) have attracted widespread attention for their dual function of energy recovery and wastewater treatment by conversion of chemical energy from partial carbohydrates and organic acids in wastewater into electrical energy by biocatalysts [ 1 , [3] , [4] , [5] , [6] ]. Notably, the physiological characteristics and electron transfer ability of electroactive microorganisms (EAMs) as biocatalysts in MFCs directly affect the catalysis performance [ 7 ]. Generally, EAMs have been isolated and obtained mainly from Geobacter , Shewanella , Pseudomonas , and Rhodoferax sp [ 7 ]. Among these strains, the extracellular electron transfer (EET) mechanism and synthetic biology modification strategies based on S. oneidensis MR-1 and G. sulfurreducens are extensively studied [ 8 , 9 ]. Both strains have multiple EET mechanisms, including direct EET mediated by c -type cytochromes [ [10] , [11] , [12] ] and conductive nanowires [ [13] , [14] , [15] , [16] ], and indirect electron transfer mediated by electron mediators [ [17] , [18] , [19] ]. However, the low electron transfer capacity of these EAMs limits the potential applications in wastewater treatment and energy recovery [ 20 , 21 ]. Therefore, to overcome this issue, numerous studies have been conducted by researchers from broadening substrate utilization, improving electron generation, and enhancing extracellular electron transfer [ [22] , [23] , [24] ]. Although these synthetic biology strategies could facilitate electron production and efficient transfer at a certain level, a higher threshold of extracellular electron transfer capacity is still not achieved due to their physiological properties and the different of EET mechanisms. Thus, there is an urgent need to isolate and screen novel EAMs with superior physiological characteristics and electron generation capacity. Activated sludge contains a diverse range of microorganisms, of which EAMs are one of the most important categories, actively involved in the degradation and resourceful conversion of wastewater pollutants [ 25 , 26 ]. In order to improve wastewater treatment and resource utilization, researchers have developed various methods for enrichment and screening of EAMs using activated sludge as an inoculum source in recent years, including MFC electrode enrichment, WO 3 nanoprobe coloration [ [27] , [28] , [29] , [30] ], peroxidase coloration [ [31] , [32] , [33] ], and dye reduction [ [34] , [35] , [36] ]. However, most of the microorganisms isolated and selected from activated sludge mainly were exoelectrogens with low electron transfer efficiency, which cannot realize efficient degradation and power recovery of organic wastewater containing azo dyes [ 37 , 38 ] and domestic wastewater [ 5 , 39 ]. Here, to realize efficient degradation of methyl orange (MO) meanwhile harvest clean electricity, a highly electroactive strain S. carassii -D 5 was isolated from activated sludge. The physiological analysis confirmed that S. carassii- D 5 was a Gram-negative strain which could maintain survival and power production using lactate as carbon source and electron donor. We further measured the electrochemical characterization of S. carassii -D 5, which showed that the maximum power density was 704.6 mW/m 2 . The extracellular electron transfer mechanism of S. carassii -D 5 was probed by using electrochemical and spectroscopic methods, demonstrating that S. carassii -D 5 mostly conducted extracellular electron transfer via a direct electron transfer channel mediated by cytochromes and electroactive biofilms. In addition, to further enhance its extracellular electron transfer and MO degradation capacity, a riboflavin synthesis gene cluster ribADEHC from Bacillus subtilis was heterologously expressed in S. carassii -D 5 , increasing riboflavin yield from 1.9 to 9.0 mg/g DCW with 1286.3 mW/m 2 power density output in lactate fed-MFCs. Furthermore, the strengthen extracellular electron transfer capacity resulted in an increase in the first order rate constants ( k ) of the MO degradation level from 0.0175 h −1 to 0.0316 h −1 with a maximum power density of 192.3 mW/m 2 in MFC.",
"discussion": "3 Results and discussion 3.1 Bacterial enrichment, isolation, and identification In this study, for the purpose of enriching and isolating highly electroactive microorganisms, the MFCs were constructed using activated sludge from wastewater treatment plants as anolyte as well as microbial community in the sludge as anode microorganisms. The voltage output performance of activated sludge-MFCs was investigated ( Fig. S1 ). It was obvious that the activated sludge-MFC could maintain a maximum output of 242.3 ± 12 mV over three generation cycles, which indicated that a stable microbial community was formed at the electrode to release electrons. After 250 h of enrichment, anode carbon cloth containing electroactive bacteria was resuspended in PBS buffer, and then incubated with 10 −9 CFU/mL concentration on LB agar plates by combining WO 3 nanomaterials as probes for rapidly identifying EAMs [ 28 ]. Based on WO 3 electrochromic, we obtained an electroactive strain with strong electrochromic ability ( Fig. S2 ). To further identify the properties of the isolated electroactive strain, the 16S rRNA fragment was amplified by PCR with universal primers 27F/1492R. Based on the results of the 16S rRNA gene sequencing ( Seq. S1 ), we successfully constructed a phylogenetic tree of the strain using the neighbor-joining method and identified that the strain belonged to the genus Shewanella carassii , with 99.79% sequence similarity to Shewanella carassii strain LZ2016-166 (GenBank accession number MF164483.1 ) in Fig. 1 . We thus named it as Shewanella carassii- D 5 . The information on the other isolated strains was provided in Table S1 . Fig. 1 Phylogenetic tree based on the results of a neighbor-joining analysis of 16S rRNA sequences for the strain Shewanella carassii- D 5 and various members of Shewanella spp. Numbers at nodes indicate bootstrap values > 50% (expressed as percentages of 1000 replications). Fig. 1 3.2 Biological characterization of the Shewanella carassii- D 5 strain To characterize the morphological properties of Shewanella carassii- D 5 strain, we first observed its cellular morphology on LB agar plate using a smart camera. As shown in Fig. 2 A, the colonies were round or elliptical with neat edges and opaque pink-orange color. The Gram stain results indicated that S. carassii- D 5 was Gram-negative ( Fig. S3 ) with a short rod-like morphology, similar to S. oneidensis MR-1 [ 49 ]. Scanning electron microscopy (SEM) and transmission electron microscope (TEM) were further used to observe cell ultrastructure ( Fig. 2 B–C). It was found that S. carassii- D 5 was rod-shaped strain with a size of 2.0–3.0 μm and a rough cell surface containing extracellular polymeric substances (EPS), which may contribute to biofilms formation to promote energy conversion and protect them against adverse environmental influences [ 50 ]. Fig. 2 Characterization of biological properties of S. carassii- D 5 . (A) Colony morphology photo of the bacteria inoculated on LB agar plate. (B) SEM characterization of the isolated strain. (C) TEM characterization of the bacteria. (D) OD 600 of inoculated with the bacteria adding acetate, lactate, glycerol, fructose, sucrose, xylose, galactose, and glucose as substrates at concentration ranging from 5 to 20 mM. (E) Optimization of optimum growth temperature of S. carassii -D 5 with 20 mM lactate as the sole carbon source. (F) Comparison of MO degradation ability between S. carassii- D 5 and S. oneidensis MR-1. Three biological replicates were performed. (∗∗: p < 0.01; ∗: p < 0.05). Fig. 2 In addition, to determine the optimal carbon source for the growth of S. carassii- D 5 , the cell growth was monitored by adding different concentration gradients (5–20 mM) acetate, lactate, glycerol, fructose, sucrose, xylose, galactose, and glucose to the M9 medium as sole carbon sources, respectively. As shown in Fig. 2 D, the results indicated that S. carassii- D 5 enabled metabolizing lactate and acetate but cannot utilize other carbon sources. In particular, cell growth was enhanced with increasing of lactate concentration, which is the same as the profiles of Shewanella . To investigate the optimum culture temperature of S. carassii- D 5 , we further determined its growth activity at different temperatures using 20 mM lactate as the carbon source. As revealed in Fig. 2 E, the optimum culture temperature of S. carassii- D 5 was 30 °C, which is consistent with S. oneidensis MR-1. However, the growth of S. carassii- D 5 strain is relatively weaker than S. oneidensis MR-1 ( Fig. S4 ). The UV–Vis spectra of MO solution in cathode of MFCs inculcated with the wide type S. carassii -D 5 under illumination at different reaction times was seen in Fig. S5 that the S. carassii -D 5 achieved the degradation of the MO solution. The MO removal rate of S. carassii- D 5 was shown in Fig. 2 F, which showed a more rapid MO removal process compared to S. oneidensis MR-1. Especially, within 72 h, S. carassii -D 5 could degrade about 66% of MO, which was about 1.7 times higher than that of S. oneidensis MR-1 (40%), indicating that the superior MO degradation ability was mainly caused by the high EET of S. carassii -D 5 rather than the cell growth. 3.3 Electrochemical characterization of S. carassii- D 5 The power density and the polarization curve are important indicators to assess the bioelectricity generation performance of MFCs. Here, to evaluate the specific power production capacity of S. carassii -D 5 , we systematically measured the above indicators in MFCs. As shown in Fig. 3 A, the power density output curves shown that S. carassii- D 5 strain obtained a maximum power density of 704.6 mW/m 2 , which was increased by 5.6 times compared with that of the S. oneidensis MR-1 (125.0 mW/m 2 ). The polarization curve reflected the relationship between the potential and the current when the external resistance was reduced. The dropping slope in the polarization curve of S. carassii- D 5 was smaller ( R int = 1670.0 Ω), indicating the smaller internal resistance of MFC comparing with S. oneidensis MR-1 ( R int = 5317.0 Ω). The corresponding output voltage was shown in Fig. S6 . Moreover, the redox reaction kinetics at cell-electrode interfaces was further conducted by CV [ 19 , 51 ]. In Fig. 3 B, S. carassii- D 5 and S. oneidensis MR-1 exhibited an obvious riboflavin redox peak originating from ∼ −0.4 V (vs. Ag/AgCl). Notably, the redox peak of S. carassii- D 5 was slightly lower than that of S. oneidensis MR-1. In the meantime, we also found a redox peak on the CV curve starting from ∼ −0.3 V, which corresponded to OM c -Cyts, thus speculated that outer membrane cytochromes main played a major role in extracellular electron transfer in S. carassii- D 5 strain. Fig. 3 Electrochemical characterization of S. carassii- D 5. (A) LSV and the polarization curves and (B) CV curves of S. carassii -D 5 (OD 600 = 1.0) inoculated in MFCs with lactate as electron donor. (C) Cytochrome level measurement in the LB fermentation broth at OD 600 = 1.0. (D) The attached biofilm and colony count on anode carbon cloth in MFCs with lactate as electron donor. (E) Water contact angle observation and the affinity of the cells for n-hexadecane. (F) Riboflavin production in the LB fermentation broth at OD 600 = 1.0. Three biological replicates were performed. (∗∗: p < 0.01; ∗: p < 0.05; ns: no significance; p > 0.05). Fig. 3 Numerous studies demonstrated that the mechanism of EET between EAMs and electrodes is conducted mainly by direct transfer mediated by c -type cytochromes or conductive nanowires and indirect transfer mediated by electron mediators [ 9 ]. Therefore, to reveal the EET mechanism of S. carassii- D 5 , we further measured the cytochromes, adhesion biomass and riboflavin in MFC which were regarded as playing important roles in EET process of Shewanella . A full wavelength scan results showed that S. carassii- D 5 had a higher density of cytochromes expression than that of S. oneidensis MR-1, suggesting that S. carassii- D 5 could depend on abundant cytochromes directly involving in extracellular electron transfer ( Fig. 3 C). The DCW of the anode chamber and biofilm loading on anodic carbon cloth was determined. As shown in Fig. 3 D, S. carassii -D 5 had a relatively high biomass (131.0 ± 9.6 μg/cm 2 ) on the anode surface, ∼2.5-fold higher than S. oneidensis MR-1 (52.3 ± 14.9 μg/cm 2 ). Meanwhile, the cell count results also showed that there were massive S. carassii -D 5 cells attached on the anode carbon cloth (58.2 ± 12.0 × 10 7 /cm 2 ), which was consistent with the SEM results ( Fig. S7 ). However, there was no significant difference in the total biomass in the anode chamber, which indicated a significant advantage in biofilm formation of S. carassii- D 5 over S. oneidensis MR-1. In addition, it is well know that bacterial attachment to a surface is the initial step in biofilm formation and hydrophilicity might be the dominant factor affecting bacterial adhesion, and strains with relatively strong hydrophobicity are more likely to aggregate to form biofilms [ 52 ]. To elucidate the mechanism of high loading of the anodic biofilm, we further determined the contact angle (θ W), as well illustrated in Fig. 3 E, the contact angle of S. carassii -D 5 was 16.15°, while that of S. oneidensis MR-1 was 13.23°. Normal hexadecane-grown cells showed strong capacity in adhering to the hydrocarbon-water interface [ 53 ]. Here, it was further shown that the hydrocarbon adhesion rate of S. carassii -D 5 was 66.3%, which was ∼1.5 times higher of S. oneidensis MR-1 (42.6%). In brief, the cell surface hydrophobicity of S. carassii -D 5 was higher than that of S. oneidensis MR-1, the abundant EPS of cells which was strengthened contact interaction between cells and electrode surface, thus enhancing the cell coverage on electrode. Besides, we further determined the amount of riboflavin synthesized in S. carassii -D 5 ( Fig. 3 F). It was found that S. carassii- D 5 could produce a riboflavin level of 1.9 mg/g DCW, similar to S. oneidensis MR-1, which further supported the previous speculation that direct EET mechanism played an essential role. 3.4 Enhanced indirect electron transfer by expressing riboflavin synthesis pathway in S. carassii- D 5 Recent studies have found that enhancing the extracellular electron transfer rate not only enhances power output but also promotes MO degradation [ 19 , 54 , 55 ]. Therefore, to further boost extracellular electron transfer and MO downgrade, an available inducible promoter P tac and the riboflavin biosynthetic gene cluster ribADEHC from B. subtili were assembled into S. carassii- D 5 to construct recombinant strain S. carassii -D 5 -C 5 ( Fig. 4 A). Subsequently, the inducible plasmid expression level was further optimized by measuring cell density. As shown in Fig. 4 B, the optimum concentrations of IPTG and kana obtained after optimization were 0.2 mM and 20 mg/L, respectively. Fig. 4 C showed that overexpressing the riboflavin synthesis pathway in S. carassii -D 5 exhibited a superior growth under aerobic conditions, indicating that riboflavin synthesis promoted cellular metabolic activity. As shown in Fig. 4 D, the riboflavin yield secreted by the engineered strain was ∼9.0 mg/g DCW, which was 4.7-fold higher than that of the control strain (1.9 mg/g DCW). To evaluate the electricity generation capacity of using lactate, S. carassii- D 5 -C 5 was further inoculated to lactate-fed MFC. The output voltage results showed that the electron output capacity of S. carassii- D 5 -C 5 was significantly higher than that of S. carassii- D 5 ( Fig. S8 ) . In Fig. 4 E, the maximum power density of recombinant strain S. carassii -D 5 -C 5 was 1286.3 mW/m 2 , ∼1.8 times than that of S. carassii- D 5 (704.6 mW/m 2 ). The slope of the polarization curve of S. carassii- D 5 -C 5 was smaller than that of S. carassii- D 5 , which indicated that the MFC inoculated with S. carassii -D 5 -C 5 strain ( R int = 803.0 Ω) had a lower internal resistance compared to S. carassii- D 5 ( R int = 1670.0 Ω). Moreover, the redox reaction kinetics at cell-electrode interfaces was conducted by CV in Fig. 4 F. Compared to S. carassii -D 5 , it is obviously that the S. carassii -D 5 -C 5 exhibited a higher riboflavin redox peak originating from ∼ −0.4 V (vs. Ag/AgCl), which indicated that the riboflavin-mediated EET was strengthened. Meanwhile, a higher OM c -Cyts redox peak on the CV curve starting from ∼ −0.3 V was found, which indicated more riboflavin can bound with cytochromes to form semi-quinones, accelerating the extracellular electron transfer by \"single-electron redox reaction\" [ 56 ]. In conclusion, all these results suggested that the presence of a relatively higher level of riboflavin synthesized by integrating riboflavin synthesis module could increase the bioelectricity generation and extracellular electron transfer capacity. Fig. 4 Enhanced production of riboflavin by engineering S. carassii -D 5 via synthetic biology approach. (A) Schematic plasmid map of P tac promoter combined with RibA, RibD, RibE, RibH, and RibC genes expressing vectors. The plasmid was assembled into S. carassii- D 5 to construct recombinant strain S. carassii -D 5 -C 5 for the enhanced riboflavin metabolism. (B) The appropriate concentrations of IPTG and kana for growth metabolism of engineered strain. (C) Growth curves of the S. carassii -D 5 -C 5 and S. carassii -D 5 in LB fermentation broth with final optimal concentration of 0.2 mM IPTG and 20 mg/L kana at 30 °C for 32 h. (D) Concentration of riboflavin produced by recombinant strain S. carassii -D 5 -C 5 and the control strain S. carassii -D 5 in the broth. (E) LSV and the polarization curves and (F) CV curves of recombinant strain S. carassii -D 5 -C 5 and the control strain S. carassii -D 5 (OD 600 = 1.0) inoculated in MFCs with lactate as electron donor. Three biological replicates were performed. (∗∗: p < 0.01; ∗: p < 0.05). Fig. 4 3.5 MO degradation with simultaneous power generation by engineered S. carassii The MO removal capacity of MFC makes it a sustainable technology in wastewater treatment. In this study, to synergistically and simultaneously achieve MO degradation and energy harvesting, we constructed a MFC using S. carassii -D 5 -C 5 as biocatalyst and MO as cathode electrolyte in Fig. 5 A. Using MO solution as electron acceptors in cathode (pH = 3, initial MO concentration was 50 mg/L), the MO solution was reduced as the azo-bond broken, becoming clear and transparent gradually. The digital multimeter was used to record the output voltage of the MFCs, as illustrated in Fig. 5 B, the MFC inoculated with S. carassii -D 5 -C 5 only took 3.5 h to quickly increase its bioelectricity generation from the initial 50.0 mV to the maximum voltage of 168.6 mV, while the MFC inoculated with S. carassii- D 5 increased from the initial 20.0 mV to 103.0 mV used ∼7.5 h at the same condition accompanied with MO degradation. The MFCs inoculated with the engineered strain had smaller internal resistance, suggesting that enhanced riboflavin biosynthesis further strengthened the EET capacity of S. carassii- D 5 . The power density and polarization curves were calculated and shown in Fig. 5 C, which showed that S. carassii -D 5 -C 5 strain obtained a maximum power density of 192.3 mW/m 2 , ∼1.6 times higher than that of S. carassii -D 5 . Notably, the slope of the polarization curve of S. carassii- D 5 -C 5 was smaller than that of S. carassii- D 5 , suggesting that the synthesis and secretion of riboflavin reduced the resistance of extracellular electron transfer and thus enhanced the EET [ 57 , 58 ]. Fig. 5 MO reduction by the recombinant strain S. carassii -D 5 -C 5 and S. carassii -D 5 . (A) Structural design of MFC for simultaneous MO decolorization and bioelectricity generation. (B) Voltage output obtained from the MFCs operated for one discharge cycle using MO solution as electron acceptor in cathode (pH = 3, initial MO concentration was 50 mg/L). (C) LSV and the polarization curves of the MFCs. (D) The UV–Vis spectra of MO solution in cathode of MFCs inculcated with the recombinant strain S. carassii -D 5 -C 5 under illumination at different reaction times. (E) Characterization of MO degradability in cathode. (F) Anaerobic reduction and reduction kinetic curves of MO. Three biological replicates were performed. Fig. 5 Additionally, the biodegrade of MO in MFCs was also investigated. Fig. 5 D illustrated that the UV–Vis spectra of MO solution in the cathode of MFCs inoculated with the strain S. carassii- D 5 -C 5 at different reaction times, the gradual decrease in the absorbance at 465 nm revealed that MO was reduced gradually and suggested that the azo-bond of MO was depleted. The reduction rate of MO by S. carassii- D 5 -C 5 was further identified by ultraviolet and visible spectrophotometer. As shown in Fig. 5 E, S. carassii- D 5 -C 5 exhibited more efficient degradation of MO. Within approximately 72 h, 86% of MO was reduced by strain S. carassii- D 5 -C 5 , which was 1.3 times higher than that of S. carassii- D 5 (66%). To further investigated MO reduction kinetics in MFCs, the pseudo-first-order kinetics equation was used to describe the MO removal kinetics (Eq. (3) ). Fig. 5 F revealed that the calculated kinetics constant of S. carassii- D 5 -C 5 was 0.0316 h −1 , which was 1.8 times higher than that of S. carassii- D 5 (0.0175 h −1 ). These results suggested that improving the synthesis of riboflavin not only enhanced EET but also promoted MO degradation of S. carassii -D 5 ."
} | 6,734 |
35400206 | PMC9040811 | pmc | 3,100 | {
"abstract": "ABSTRACT Quorum sensing (QS) is a unique mechanism for microorganisms to coordinate their activities through intercellular communication, including four main types of autoinducer-1 (AI-1, namely, N -acyl homoserine lactone [AHL]), AI-2, AI-3, and diffusible signaling factor [DSF]) based on signaling molecules. Quorum quenching (QQ) enzymes can disrupt the QS phenomenon by inactivating signaling molecules. QS is proposed to regulate biofilm formation in extremely acidic environments, but the QS/QQ-related genomic features in most acidophilic bacteria are still largely unknown. Here, genome annotation of 83 acidophiles from the genera Acidithiobacillus , Leptospirillum , Sulfobacillus , and Acidiphilium altogether revealed the existence of AI-1, AI-3, DSF, and AhlD (AHL degradation enzyme). The conservative investigation indicated that some QS/QQ-related proteins harbored key residues or motifs, which were necessary for their activities. Phylogenetic analysis showed that LuxI/R (AI-1 synthase/receptor), QseE/F (two-component system of AI-3), and RpfC/G (two-component system of DSF) exhibited similar evolutionary patterns within each pair. Meanwhile, proteins clustered approximately according to the species taxonomy. The widespread Acidithiobacillus strains, especially A. ferrooxidans , processed AI-1, AI-3, and DSF systems as well as the AhlD enzyme, which were favorable for their mutual information exchange and collective regulation of gene expression. Some members of the Sulfobacillus and Acidiphilium without AHL production capacity contained the AhlD enzyme, which may evolve for niche competition, while DSF in Leptospirillum and Acidithiobacillus could potentially combine with the cyclic diguanylate (c-di-GMP) pathway for self-defense and niche protection. This work will shed light on our understanding of the extent of communication networks and adaptive evolution among acidophiles via QS/QQ coping with environmental changes. IMPORTANCE Understanding cell-cell communication QS is highly relevant for comprehending the regulatory and adaptive mechanisms among acidophiles in extremely acidic ecosystems. Previous studies focused on the existence and functionality of a single QS system in several acidophilic strains. Four representative genera were selected to decipher the distribution and role of QS and QQ integrated with the conservative and evolutionary analysis of related proteins. It was implicated that intra- or intersignaling circuits may work effectively based on different QS types to modulate biofilm formation and energy metabolism among acidophilic microbes. Some individuals could synthesize QQ enzymes for specific QS molecular inactivation to inhibit undesirable acidophile species. This study expanded our knowledge of the fundamental cognition and biological roles underlying the dynamical communication interactions among the coevolving acidophiles and provided a novel perspective for revealing their environmental adaptability.",
"introduction": "INTRODUCTION Acidophiles, a category of primary extremophiles with unique features, are widely distributed in acid mine/rock drainage (AMD/ARD) environments ( 1 ). In these natural ecosystems, acidophilic microorganisms take active parts in the elemental cycle of sulfur and iron globally by oxidizing reduced inorganic sulfur compounds (RISCs) to sulfate and transforming ferrous and ferric ions ( 2 , 3 ). It has been revealed that biofilm formation by attached bacterial cells is correlated with extracellular polymeric substance (EPS) production ( 4 ). As the prevalent and predominant genus thriving in AMD, the application of Acidithiobacillus in bioleaching has been studied extensively, as well as its adaptive evolution to extreme environments ( 5 , 6 ). Acidithiobacillus , together with other acidophilic genera such as Leptospirillum , Sulfobacillus and Acidiphilium , exists in sulfide-bearing mineral environments between 20 and 40°C with pH lower than 3 ( 7 , 8 ). Their attachment on ore surface and subsequent biofilm formation have been deciphered in great detail ( 9 ). Quorum sensing (QS) is a sophisticated cell-to-cell communication process that enables bacteria to sense environmental changes (especially cell densities) and then orchestrate behaviors collectively, such as bioluminescence, motility, virulence factor production, and biofilm formation ( 10 , 11 ). QS relies on the production, release, accumulation, and detection of extracellular signal molecules, called autoinducers (AIs). With the increasement of bacterial population density, AIs accumulate in the outer environment. Bacteria monitor the change of AI concentration, namely, the shift in cell amounts, and then jointly alter the expression levels of specific genes once the threshold has been reached ( 12 ). A novel conception called quorum quenching (QQ) enzymes, which inactivate QS molecules, has also emerged. QQ is termed a QS interference method, which is deemed to have evolved either by QS owners to eliminate excess signals or by other competitive organisms to attenuate their QS communication pathway ( 13 ). Specifically, the autoinducer-1 (AI-1) system, composed of the canonical LuxI/LuxR pair, is one of the most well-studies QS systems in bacteria and is prevalent in Gram-negative (G – ) bacteria such as Proteobacteria ( 14 ). LuxI catalyzes reactions between the homoserine lactone moiety contributor S -adenosylmethionine (SAM) and the acyl carrier protein (ACP) and principally synthesizes 3-oxo-hexanoyl homoserine lactone (OHHL), which is an N -acyl homoserine lactone (AHL) with the 3-oxo group. Then LuxR protein recognizes and binds to AHLs and consequently activates the expression of various QS-dependent genes ( 15 ). Chemical degradation of AHL compounds is a featured instance of QQ enzymes. AHL-lactonases and AHL-acylases have been described in several bacteria and developed as promising tools to block unnecessary gene expression and pathogenic phenotypes in medicine, aquaculture, and other fields ( 13 ). As a major type employed by both Gram-positive (G + ) and G − bacteria, the autoinducer-2 (AI-2) pathway serves in intra-species as well as inter-species communication modes ( 16 ). AI-2 is generated by the LuxS enzyme through a series of reactions and is recognized by three specific receptors, LuxP, LsrB, and RbsB ( 17 ). The autoinducer-3 (AI-3) is an amination product of aromatic compounds, as a less common interkingdom QS system, which is mainly found and elucidated in the enteric bacterium enterohemorrhagic Escherichia coli (EHEC). The QS E. coli regulators B and C (QseBC) in conjunction with E and F (QseEF) are key components of the AI-3/epinephrine (Epi)/norepinephrine (NE) signaling circuits. After binding with AI-3/Epi/NE, QseC auto-phosphorylates, and then it mediates the phosphorylation of QseB, thereby coordinating the expression of flagellar, motility, and virulence genes ( 18 ). QseEF share a similar mechanism but have a narrower distribution and play essential roles in the regulation of attaching and effacing (AE) lesion formation ( 18 , 19 ). The diffusible signaling factor (DSF) represents a novel kind of QS system, which is exemplified in Xanthomonas campestris pv. campestris . DSF family signals synthesized by RpfF are transduced by the sensor protein RpfC to its receptor RpfG, to regulate the expression of DSF-controlled genes based on the signaling cascade encompassing RpfB, cyclic diguanylate (c-di-GMP), and Clp in X. campestris pv. campestris ( 20 ). Bioinformatic prediction and experimental validation have determined the existence of QS signaling communication within acidophiles. For example, it is well known that Acidithiobacillus ferrooxidans possesses a functional AHL-type QS. The system has evolved unique regulatory patterns specific to the energy substrates, which expands our understanding of AI-1 to adjust the gene transcription of A. ferrooxidans for cell growth and population development in Fe/S-enriched extremely acidic environments ( 21 , 22 ). It is known that several A. ferrooxidans and Acidithiobacillus thiooxidans strains synthesize AHLs, which could certainly exert functions within the same species or might make A. ferrooxidans / A. thiooxidans communication happen. Although two Leptospirillum ferrooxidans strains, DSMZ 2391 and DSMZ 2705, could not produce AHL, they can sense external AHLs secreted by other microbes located in its ecological niche by expressing a SdiA-like protein, just like the case of Escherichia and Salmonella ( 23 ). After addition of exogenous DSF, the attached cells of Acidithiobacillus caldus , Leptospirillum ferriphilum , and Sulfobacillus thermosulfidooxidans on mineral surfaces decline obviously, so the Leptospirillum spp. are proved to have the ability to produce the DSF family compounds ( 24 ). It will be fascinating to examine the effects of DSF in mixed cultures involving Acidithiobacillus , Leptospirillum , and heterotrophic Acidiphilium spp. in the future ( 24 ). Hence, QS are powerful mediators of intra- to inter-species communication circuitry, and some acidophilic members maintain complex interactions with others by this approach. With the increasing development of sequencing techniques, more and more genomic data of acidophiles are readily available. In this study, we reported the bioinformatic survey of QS or QQ in the four above-mentioned acidophilic genera and focused on the AI-1 system being inspected frequently. With regard to the qualified QS/QQ-related proteins, the sequence alignment was conducted to check key residues or motifs for authentication validity. Then phylogenetic analysis was carried out to determine evolutionary relationships. Finally, we deciphered the system distribution properties and assessed our current understanding of QS in acidophiles, aiming to offer genomic evidence for its potential role and function in extreme AMD ecosystems.",
"discussion": "DISCUSSION QS is an intercellular communication mechanism which enables bacteria to feel surrounding cues and coordinately adjust their density and behavior. It has been reported that some acidophilic bacteria have utilized this novel strategy to coordinate EPS synthesis, energy metabolism, or biofilm formation to control bioleaching activity and environmental adaptation ( 24 , 39 , 40 ). Here, a bioinformatic survey has been conducted to predict the existence of a QS system in some acidophiles at the genomic level. Then, the conservation, probable function, and evolutionary aspects of QS-related proteins were explored. A schematic distribution of QS/QQ systems among acidophiles was concluded ( Fig. 7 ). Acidithiobacillus strains encoded three types of QS system and one QQ type of AHL degradation enzyme. Some individuals of Sulfobacillus and Acidiphilium have evolved an enzyme for AHL inactivation, potentially playing crucial roles against other competitors. An effective DSF-type signaling system was notably specific to Leptospirillum in particular. FIG 7 Overview of proposed QS system and QQ enzyme distribution in acidophiles. (1) AI-1 QS system; (2) AHL degradation enzyme of QQ; (3) QseE/F of AI-3 QS system; (4) DSF-QS system. AHL molecules cleaved by acidophiles with(out) AHL production capacity for effective niche exploitation. The AI-1 system is one major QS type which has been characterized widely. Meanwhile, the related LuxI/LuxR and their homologues are focused on and examined extensively via in silico analyses ( 41 , 42 ). Using the domain-based strategy and key residue comparison, we found that an adjacent LuxI/R homolog pair was carried by most A. ferrooxidans and A. thiooxidans strains, similar to the distribution of the AfeI/AfeR system based on sequence BLAST analysis ( 22 ). The AfeI/R (LuxI/R like) system has been defined and verified in A. ferrooxidans , which could intelligently drive energy metabolism, cell growth, and EPS secretion in Fe 2+ /S 0 -enriched medium to benefit themselves ( 22 ). Nine different chemical AHL molecules with diverse C-3 substitutions (hydroxyl and oxo) and only even numbers (between 8 and 16) of carbons in the acyl chain (3-hydroxy-C 8 -AHL, 3-hydroxy-C 10 -AHL, C 12 -AHL, 3-oxo-C 12 -AHL, 3-hydroxy-C 12 -AHL, C 14 -AHL, 3-oxo-C 14 -AHL, 3-hydroxy-C 14 -AHL, and 3-hydroxy-C 16 -AHL) are detected from A. ferrooxidans ATCC 23270 cultures ( 21 ). Plentiful 3-hydroxy-C 14 -AHL is traced in Fe 2+ - and S 0 -enriched media, which could stimulate regulation of EPS synthesis and cell growth in S 0 -enriched medium but not work in Fe 2+ -enriched medium ( 22 ). In addition, for the two kinds of 3-oxo-AHL compounds with 12 or 14 carbons in the large acyl chains, they are produced only by sulfur- and thiosulfate-grown cells, and the function effects are still to be characterized ( 21 ). It is common that AHL synthases are able to produce more than one kind of AHL. The SinI of Sinorhizobium meliloti synthesizes five different forms of AHLs, ranging from C 12 -AHL to C 18 -AHL, including some oxo-AHLs and a monounsaturated AHL ( 43 ). Longer-chain AHLs seem to be more insulated from chemical degradation and utilized by microorganisms in harsher environments, suggesting that cross talk might emerge between A. ferrooxidans and other bacterial species inhabiting the bioleaching ecological niche ( 44 ). According to the evolutionary relatedness, LuxI and LuxR homolog proteins have similar clustering characteristics, indicating that they may be coevolving and cofunctioning ( 45 ). The tree clades clustered mainly in accordance to the species taxonomy, and the AI-1 system of A. ferrooxidans and A. thiooxidans exhibited a clear division. In contrast, the system showcased a relatively higher degree of evolutionary relatedness within the same species. Moreover, the LuxI homolog solo of Acidiphilium sp. 37-67-22 was evolutionarily distant and shared similar crystal structures and vital substitution of a mutant RhlI, suggesting its authenticity and functionality in AHL with 3-oxo group synthesis. Though RhlI is reported to mainly produce N -butanoyl (BHL) and N -hexanoyl (HHL) homoserine lactones, which are AHLs without 3-oxo groups ( 31 ). Chemical communication between Acidiphilium strains and Acidithrix strains has effectively enhanced growth and Fe(II) oxidation rates. The mediated QS molecules still need in-depth inspection, then ( 46 ). The LuxR homolog solo in L. ferrooxidans C2-3 has also been confirmed, and the same critical residues have been identified in a bioinformatic survey, clustering with LuxR solos of Methylacidiphilum fumariolicum , Methylacidiphilum infernorum , and Nitrospira defluvii in the phylogenetic tree ( 47 ). The phenomenon of biofilm formation triggered by QS has been observed through a transcriptome technique ( 48 ); we thus postulated that its LuxR solo could sense AI-1 produced by other bacteria such as A. ferrooxidans , which always live with them in extremely acidic waters, and then modulate relevant gene expression. As a dominant organism in AMD environments, A. ferrooxidans has all kinds of the above-mentioned communication tools, which may provide strong assistance for its desirable adaptation. Since AHL-dependent signaling strategies are widespread and attractive among bacteria, a signal interference method that disrupted the QS system has come into view over the past decade. The AHL-degradation enzyme AhlD in Arthrobacter sp. has been discussed and predicted to exist in some other bacteria ( 32 ). The AhlD distribution spanned various species of Acidithiobacillus , Sulfobacillus , and Acidiphilium and harbored a conserved motif, HXHXDH≈H≈D, necessary for enzymatic activity. Interestingly, except for A. ferrooxidans BY-3 and Acidiphilium sp. 37-67-22, the others did not possess any AHL-QS system-linked proteins. Some bacteria could cleave their own AHL signal, such as Agrobacterium and Pseudomonas ( 13 ). Owing to extensive QQ activities within the phyla Proteobacteria , Bacteroidetes , Actinobacteria , and Firmicutes , it is likely that many acidophiles may fight against others with AHL-emitting capacity or balance the amount of AHL produced by themselves for efficient resource and niche utilization by this approach ( 13 ). DSF combined with c-di-GMP to jointly regulate biofilm formation among acidophiles. The DSF-based QS system represents an intriguing type of cell-to-cell communication mechanism in diverse G – bacteria. Multiomics and genetic methods have unveiled a complete DSF system possessed by L. ferriphilum DSM 14647 T ( 49 ) and a similar rpf gene cluster with a complete RpfC homologue contained in L. ferrooxidans C2-3 ( 50 ). As well as in Leptospirillum species, we also found the DSF type QS carried by several A. ferrivorans and A. thiooxidans strains, A. ferrooxidans BY0502, A. ferridurans IO-2C, and A. albertensis DSM 14366. By sequence alignments, signature residues or motifs were screened in RpfF essential for DSF biosynthesis, and RpfB engaged in the turnover of the DSF family signals. The signaling sensor RpfC and transduction RpfG were distributed in terms of taxonomic lineages and displayed similar evolutionary patterns in the phylogenetic tree. Several previously described studies have discussed the vital biological effects of DSF family compounds in Leptospirillum . After adding DSF, the amounts of L. ferriphilum and S. thermosulfidooxidans adhering on minerals decrease, leading to biofilm dispersal and preventing the formation of a passivation layer, which is essential for the bioleaching performance ( 51 ). Moreover, when DSF production by L. ferriphilum microcolonies is prior to the addition of S. thermosulfidooxidans , DSF molecules specially suppress Fe 2+ oxidation of exogenous S. thermosulfidooxidans cells, thereby making the energy resource available specifically to the DSF releaser ( 52 ). Here, the Leptospirillum sp. DSF system may be an efficient niche protection strategy and could resist against other unfavored biomining populations. QS and c-di-GMP signaling system are considered to be the primary methods modulating biofilm formation and EPS production in G – acidophiles ( 53 ). The c-di-GMP metabolism elements have been predicted and compared in A. thiooxidans , A. ferrivorans , A. ferrooxidans , A. caldus , and A. albertensis ( 54 ). Meanwhile, its functions in adjusting motility and adherence have been confirmed in several Acidithiobacillus and Leptospirillum species ( 24 , 55 – 57 ). It has been acknowledged that the activated RpfG has phosphodiesterase ability and could degrade c-di-GMP, the congenital ligand of the transcription factor Clp. Ultimately, derepressed Clp alters the expression level of abundant genes, such as those coding virulence factors ( 58 ). There is reason to expect that the DSF integrated with c-di-GMP pathways plays pivotal roles in population communication and adaptation, just like the connection between AHL-mediated QS and c-di-GMP during the process of colonization and dissolution of minerals ( 4 , 59 ). Concluding remarks. In summary, our analysis provided a picture of the distribution, phylogeny, and putative functions of QS/QQ-related proteins among acidophiles belonging to four genera. The presence and authenticity of QS systems and QQ enzymes were emphasized, which played a critical role in establishing communication circuits or disturbing signal propagation for valuable niche exploitation. Intra- or interspecies contact could occur via “dialects” in the acidophilic microbe world, opening new perspectives for the regulatory networks of gene expression and adaptive evolution. More experiments are needed to investigate the ecological functions of QS/QQ in microbial communities among acidophiles."
} | 4,980 |
26075362 | null | s2 | 3,101 | {
"abstract": "It is widely believed that the archaeal ancestor was hyperthermophilic, but during archaeal evolution, several lineages - including haloarchaea and their sister methanogens, the Thaumarchaeota, and the uncultured Marine Group II and Marine Group III Euryarchaeota (MGII/III) - independently adapted to lower temperatures. Recent phylogenomic studies suggest that the ancestors of these lineages were recipients of massive horizontal gene transfer from bacteria. Many of the acquired genes, which are often involved in metabolism and cell envelope biogenesis, were convergently acquired by distant mesophilic archaea. In this Opinion article, we explore the intriguing hypothesis that the import of these bacterial genes was crucial for the adaptation of archaea to mesophilic lifestyles."
} | 196 |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.