pmid stringlengths 8 8 | pmcid stringlengths 8 11 ⌀ | source stringclasses 2
values | rank int64 1 9.78k | sections unknown | tokens int64 3 46.7k |
|---|---|---|---|---|---|
39921902 | PMC11842047 | pmc | 5,526 | {
"abstract": "Abstract Summary Constraint-based metabolic models offer a scalable framework to investigate biological systems using optimality principles. Construction and simulation of detailed models that utilize multiple kinds of constraint systems pose a significant coding overhead, complicating implementation of new types of analyses. We present an improved version of the constraint-based metabolic modeling package COBREXA, which utilizes a hierarchical model construction framework that decouples the implemented analysis algorithms into independent, yet re-combinable, building blocks. By removing the need to re-implement modeling components, assembly of complex metabolic models is simplified, which we demonstrate on use-cases of resource-balanced models, and enzyme-constrained flux balance models of interacting bacterial communities. Notably, these models show improved predictive capabilities in both monoculture and community settings. In perspective, the re-usable model-building components in COBREXA 2 provide a sustainable way to handle increasingly complex models in constraint-based modeling. Availability and Implementation COBREXA 2 is available from https://github.com/COBREXA/COBREXA.jl , and from Julia package repositories. COBREXA 2 works on all major operating systems and computer architectures. Documentation is available at https://cobrexa.github.io/COBREXA.jl/ .",
"discussion": "4 Discussion Resource allocation constraints are increasingly used in metabolic modeling, due to their ability to mechanistically describe complex metabolic phenomena. We showcased that even a simplified RBA model recapitulates the experimental findings of E. coli metabolism better compared to an enzyme-constrained model: building on theoretical developments, we included both membrane and cytosolic protein capacity constraints to model the onset of overflow metabolism ( Zhuang et al. 2011 , De Groot et al. 2020 ). Interestingly, both the ec-FBA and sRBA models asserted higher sensitivity of the onset to membrane capacity limitations ( Supplementary Section S2.1 ). In the community setting, our results suggest that enzyme capacity plays an important role in determining steady-state community abundances. However, simulations of a 4-member community of E. coli mutants still showed a substantial compositional variability of near-optimal growth rates ( Supplementary Fig. S5 ), which is contradicted by experimental results ( Mee et al. 2014 ), suggesting existence of additional significant driving forces behind the community dynamics. We expect that substantially more complex models will be required to provide sufficient predictive power for such communities. We demonstrated that COBREXA 2 and the constraint trees framework enable rapid implementation of advanced modeling approaches, like community-scale resource-constrained models. In the future, we hope that the model construction approach introduced by COBREXA 2 will aid further refinement of such models with new kinds of constraint systems, enabling further improvements of the model predictive abilities and thus deeper investigation of the biological mechanisms at the root of complex metabolic phenomena."
} | 794 |
39921902 | PMC11842047 | pmc | 5,526 | {
"abstract": "Abstract Summary Constraint-based metabolic models offer a scalable framework to investigate biological systems using optimality principles. Construction and simulation of detailed models that utilize multiple kinds of constraint systems pose a significant coding overhead, complicating implementation of new types of analyses. We present an improved version of the constraint-based metabolic modeling package COBREXA, which utilizes a hierarchical model construction framework that decouples the implemented analysis algorithms into independent, yet re-combinable, building blocks. By removing the need to re-implement modeling components, assembly of complex metabolic models is simplified, which we demonstrate on use-cases of resource-balanced models, and enzyme-constrained flux balance models of interacting bacterial communities. Notably, these models show improved predictive capabilities in both monoculture and community settings. In perspective, the re-usable model-building components in COBREXA 2 provide a sustainable way to handle increasingly complex models in constraint-based modeling. Availability and Implementation COBREXA 2 is available from https://github.com/COBREXA/COBREXA.jl , and from Julia package repositories. COBREXA 2 works on all major operating systems and computer architectures. Documentation is available at https://cobrexa.github.io/COBREXA.jl/ .",
"discussion": "4 Discussion Resource allocation constraints are increasingly used in metabolic modeling, due to their ability to mechanistically describe complex metabolic phenomena. We showcased that even a simplified RBA model recapitulates the experimental findings of E. coli metabolism better compared to an enzyme-constrained model: building on theoretical developments, we included both membrane and cytosolic protein capacity constraints to model the onset of overflow metabolism ( Zhuang et al. 2011 , De Groot et al. 2020 ). Interestingly, both the ec-FBA and sRBA models asserted higher sensitivity of the onset to membrane capacity limitations ( Supplementary Section S2.1 ). In the community setting, our results suggest that enzyme capacity plays an important role in determining steady-state community abundances. However, simulations of a 4-member community of E. coli mutants still showed a substantial compositional variability of near-optimal growth rates ( Supplementary Fig. S5 ), which is contradicted by experimental results ( Mee et al. 2014 ), suggesting existence of additional significant driving forces behind the community dynamics. We expect that substantially more complex models will be required to provide sufficient predictive power for such communities. We demonstrated that COBREXA 2 and the constraint trees framework enable rapid implementation of advanced modeling approaches, like community-scale resource-constrained models. In the future, we hope that the model construction approach introduced by COBREXA 2 will aid further refinement of such models with new kinds of constraint systems, enabling further improvements of the model predictive abilities and thus deeper investigation of the biological mechanisms at the root of complex metabolic phenomena."
} | 794 |
37936059 | PMC10629017 | pmc | 5,527 | {
"abstract": "Background Heavy metal contamination has been a severe worldwide environmental issue. For industrial pollutions, heavy metals rarely exist as singular entities. Hence, researches have increasingly focused on the detrimental effect of mixed heavy metal pollution. Genome analysis of Lampropedia strains predicted a repertoire of heavy metal resistance genes. However, we are still lack of experimental evidence regarding to heavy metal resistance of Lampropedia , and their potential in mixed heavy metal removal remain elusive. Results In this study, a Lampropedia aestuarii strain GYF-1 was isolated from soil samples near steel factory. Heavy metal tolerance assay indicated L. aestuarii GYF-1 possessed minimal inhibition values of 2 mM, 10 mM, 6 mM, 4 mM, 6 mM, 0.8 mM, and 4 mM for CdCl 2 , K 2 CrO 4 , CuCl 2 , NiCl 2 , Pb(CH 3 COO) 2 , ZnSO 4 , and FeCl 2 , respectively. The biosorption assay demonstrated its potential in soil remediation from mixed heavy metal pollution. Next the draft genome of L. aestuarii GYF-1 was obtained and annotated, which revealed strain GYF-1 are abundant in heavy metal resistance genes. Further evaluations on differential gene expressions suggested adaptive mechanisms including increased lipopolysaccharides level and enhanced biofilm formation. Conclusion In this study, we demonstrated a newly isolated L. aestuarii GYF-1 exhibited mixed heavy metal resistance, which proven its capability of being a potential candidate strain for industrial biosorption application. Further genome analysis and differential gene expression assay suggest enhanced LPS and biofilm formation contributed to the adaptation of mixed heavy metals. Supplementary Information The online version contains supplementary material available at 10.1186/s12866-023-03093-4.",
"conclusion": "Conclusions In this study, we isolated a mixed heavy metal resistant L. aestuarii strain from soil samples near steel factory. Heavy metal tolerance assay indicated L. aestuarii GYF-1 possessed MIC value of 2 mM, 10 mM, 6 mM, 4 mM, 6 mM, 0.8 mM, and 4 mM for CdCl 2 , K 2 CrO 4 , CuCl 2 , NiCl 2 , Pb(CH 3 COO) 2 , ZnSO 4 , and FeCl 2 , respectively. The biosorption assay demonstrated its capacity in bioremediation of soil polluted by mixed heavy metals. Genome analysis revealed abundance of heavy metal resistance genes in the genome of L. aestuarii GYF-1. Further evaluation on differential gene expressions under stress condition suggest enhanced LPS and biofilm formation contributed to the adaptation of mixed heavy metals. This study demonstrated L. aestuarii GYF-1 can be selected as a potential candidate strain for biosorption application.",
"introduction": "Introduction Heavy metal contamination has been a severe worldwide environmental issue due to their toxicity, accumulative, and nonbiodegradable properties. The primary cause of heavy metal pollution has emerged from industrial activities, such as mining, electroplating, paints and pigments, batteries, tanning and textile, steel industries. Use of pesticides, insecticides, fertilizers in agricultural fields have been secondary source of heavy metal contamination [ 1 ]. Recently, research focuses on metal pollution have trends to shift from single metals to mixed metals. For industrial pollutions, heavy metals rarely exist as singular entities. Hence, researches have increasingly focused on the detrimental effect of mixed heavy metal pollution to ecosystem and health of living system. For example, heavy metal mixture induces global iron starvation resulted in decreased activity of biological nitrate removal [ 2 ]. Exposure to mixed heavy metals is negatively associated with renal function via oxidative stress disorder [ 3 ]. Accumulation of mixed heavy metals is highly related to the occurrence of cancer [ 4 ]. Thus, effective approaches must be taken for remediation of mixed heavy metals contamination. Microorganisms have developed various mechanisms for adaptation of mixed heavy metal stress, resulting in an eco-friendly and cost-effective strategy called biosorption. The detoxification of mixed heavy metals includes biosorption via production of extracellular polymeric substances (EPS), efflux of toxic metals by active transporters, intracellular sequestration, surface precipitation, metal reduction. To date, many bacterial groups are considered as potential bioagents for mixed heavy metal removal, such as Bacillus sp. , Pseudomonas sp. , Alcaligenes sp. , Rhizopus sp. , Sphingomonas sp. , Azospira sp. and Cupriavidus sp. , etc [ 5 , 6 ]. Therefore, discovery of novel bacteria contributes to mixed heavy metal removal is of great importance, which would benefit not only to better understanding of the bacterial adaptation mechanisms, but also to development of new biosorption strategies. Lampropedia spp . is a Gram-negative, Neisser-positive, non-spore forming coccus. L. hyalina was firstly isolated from polluted water sample by Schroeter in 1886 [ 7 ]. Since then, the other three L. hyalina strains, ATCC 11,041, ATCC 43,383, DSM 15,336 were identified from dairy farm yard [ 8 ], rumen [ 9 ], and activated sludge [ 10 ], respectively. L. hyalina DSM 15,536 was identified as phosphorus removal bacteria because its capability of synthesizing polyphosphate and polyhydroxyalkanoates accumulating bacteria [ 10 ]. This suggest Lampropedia hyalina might also be resistant to heavy metal stress via polyphosphates-mediated detoxification. Later, researches isolated three other species of Lampropedia , including Lampropedia aestuarii YIM MLB12 from a sediment sample of the Maliao River estuary [ 11 ], Lampropedia puyangensis 2-bin from cankered bark tissue of Populus × euramericana [ 12 ], and Lampropedia cohaerens CT6 from arsenic rich microbial mats of a Himalayan hot water spring [ 13 ]. To note, the draft genome of L. cohaerens CT6 also indicated a repertoire of heavy metal resistance genes against arsenic, copper, cobalt, zinc, magnesium, and cadmium [ 13 ]. However, to our knowledge, we are currently lack of experimental evidence regarding to heavy metal resistance of Lampropedia , and their potential in mixed heavy metal removal remain elusive. The present study aimed to evaluate the heavy metal resistance and application value of Lampropedia aestuarii GYF-1 isolated from soil samples near steel industry. Results demonstrated resistances of L. aestuarii GYF-1 to Cd 2+ , Cr 6+ , Cu 2+ , Ni 2+ , Pb 2+ , Zn 2+ and Fe 2+ as well as its capability of removing mixed heavy metal contamination. Further integrated genomic analysis and relative gene expression results elucidate adaptive mechanisms for reducing mixed heavy metal stress.",
"discussion": "Discussions The present study isolated one heavy metal-resistant L. aestuarii strain from soil samples near steel factory and demonstrated its potentiality in biosorption of mixed heavy metals. Currently, the Lampropedia genus is represented by four species, L. hyalina , L. puyangensis , L. cohaerens , and L. aestuarii , however, their heavy metal resistance and potential in application remain elusive. Here, we experimentally demonstrated L. aestuarii GYF-1 is resistance to multiple heavy metal stress with increased LPS level and biofilm formation. L. aestuarii GYF-1 was isolated from soil samples containing heavy metals, which suggest strain GYF-1 have developed adaptation mechanism to mixed heavy metal stress. The annotations revealed a repertoire of metal resistance genes in L. aestuarii GYF-1 genome (Table S4 ), suggest they might be responsible for efflux of the heavy metals. Bacterial chromium resistance usually via chromium reductase encoded by chromium reductase and efflux system, which involves chromium transporter encoded by chrA and a chrA positive regulator chrB [ 21 ]. The predictions of Cr-resistance genes indicated one gene GYF_02978 encodes chromate transporter gene chrA in Scaffold 5, however, no chrB gene was identified, suggest the ChrA functions alone in strain GYF-1. Two cupin-like domain transcription activator genes chrR were found in Scaffold 1 and Scaffold 9. For copper resistance, yfiH (GYF_00340) encodes multicopper polyphenol oxidoreductase laccase was found in Scaffold 1. The copZ is a known encodes cytoplasmic copper chaperone, which was predicted in Scaffold 13 (GYF_00913). A MerR-family transcription factor cueR was predicted in Scaffold 8 (GYF_03485). In addition, GYF_01534 was predicted as a potential copper(I)-binding protein although its function was not annotated in Lampropedia spp. Analysis of nickel transporter predicted many genes encode nikABCDE transporter in GYF-1 genome. This could explain previous finding that strain GYF-1 have high nickel tolerance but with no adsorption in nickel stress, because the main mechanism of nickel resistance used by GYF-1 is via efflux mediated by transporters. A whole nikABCDE gene cluster (GYF_01655–01659) was annotated in Scaffold 2, and partial nik gene clusters were predicted in Scaffold 3, 8, 11, 16, 20, suggest the nik transporter complex in GYF-1 might be different with those classic systems reported in model bacteria. Besides those single heavy metal specific transporter, many multiple heavy metal transporters were also annotated. The czcABC gene cluster (GYF_02595–02597) was annotated in Scaffold 4, which encode heavy metal efflux pump of cobalt, zinc, and cadmium. Two P1B-type ATPase genes zntA (GYF_03486 and GYF_03587) which confer resistance specifically to Pb 2+ , Zn 2+ , and Cd 2+ were found in Scaffold 8. The ABC-type Mn 2+ /Zn 2+ transport system znu was predicted in Scaffold 5, with two znuB (GYF_02916 and GYF_02917) and one znuC (GYF_02918), suggest the difference of znuABC complex in L. aestuarii GYF-1 from other studied species. The presence of multiple efflux transporters could also explain L aestuarii GYF-1 exhibited high tolerance to individual heavy metal stress as indicated by MIC values (Table 2 ). To note, the mechanism that bacterial responses to individual heavy metal or mixed stress might be different. Therefore, the role of these transporters in heavy metal stress needs to be further validated. Polyphosphates (polyP), as polyanions, are involved in detoxification of heavy metals [ 22 ]. Annotation revealed two polyphosphate kinase encoding genes, ppk1 (GYF_02745) and ppk2 (GYF_02770), are involved in polyphosphate metabolic process. The expression of ppk1 and ppk2 was assessed to evaluate if inorganic polyP contributes to intracellular heavy metal sequestration. The expression of ppk1 was 2.18-fold higher in 1 × SHMM than in control medium, and no significant up-regulation was found with regards to ppk2 expression (Fig. S9 ). This indicated the polyP might contribute to mitigation of mixed heavy metal stress, but might not play the major role for biosorption. One explanation is that the intracellular compartment depends on the concentration of heavy metals inside the cells. In addition, genome annotation also predicted the presence of lipoic acid biosynthesis pathway [ 23 ] and superoxide dismutase (SOD) [ 24 ], which are important to alleviate reactive oxygen stress generated by heavy metals. Biosorption is defined as adsorption of substances by using passive physiochemical pathways, such as electrostatic forces and ion/proton displacement, while bioaccumulation is active metabolic event in which heavy metals are taken up into the cell [ 5 ]. In this study, our current finding suggested detoxification of mixed heavy metal stress mainly through biosorption. For bioaccumulation, genome annotations indicated the presence of two proteins containing heavy metal binding motif (Table S1 ). GYF_02298 was annotated as a thiol-disulfide isomerase containing CXXC motif which is responsible for multi heavy metal binding [ 25 ]. GYF_03532 was predicted as a heavy metal sensor kinase containing a sensor histidine kinase domain [ 26 ]. However, if these two proteins actively contribute to bioaccumulation of heavy metal remain elusive. For biosorption application, we applied two-round adsorption strategy into removal of mixed heavy metals pollution. L. aestuarii GYF-1 showed biosorption efficiency of greater than 90% in removing Cd 2+ , Cr 6+ , and Pb 2+ , however, relatively low efficiency in adsorbing Cu 2+ and Zn 2+ and no affinity to Ni 2+ , individually. The biosorption efficiency is impacted by various factors, such as pH, temperature, initial concentration of the heavy metals, cell density, treatment time, etc [ 5 ]. Thus, optimization of adsorption condition might be necessary to improve the biosorption efficiency. On the other hand, mixed heavy metal biosorption assay proved the concept that L. aestuarii GYF-1 could be used as biosorbent for bioremediation of soil samples in a lab-scale. Several bioprocess factors also should be taken into considerations when scale-up to industrial level, for example type of bioreactor, pH, and temperature control, mixing and agitation, feeding strategy (in batches or in continuous mode) [ 27 , 28 ]."
} | 3,280 |
38475460 | PMC10934396 | pmc | 5,528 | {
"abstract": "The application of biostimulants has been proven to be an advantageous tool and an appropriate form of management towards the effective use of natural resources, food security, and the beneficial effects on plant growth and yield. Plant-growth-promoting rhizobacteria (PGPR) are microbes connected with plant roots that can increase plant growth by different methods such as producing plant hormones and molecules to improve plant growth or providing increased mineral nutrition. They can colonize all ecological niches of roots to all stages of crop development, and they can affect plant growth and development directly by modulating plant hormone levels and enhancing nutrient acquisition such as of potassium, phosphorus, nitrogen, and essential minerals, or indirectly via reducing the inhibitory impacts of different pathogens in the forms of biocontrol parameters. Many plant-associated species such as Pseudomonas , Acinetobacter , Streptomyces , Serratia , Arthrobacter , and Rhodococcus can increase plant growth by improving plant disease resistance, synthesizing growth-stimulating plant hormones, and suppressing pathogenic microorganisms. The application of biostimulants is both an environmentally friendly practice and a promising method that can enhance the sustainability of horticultural and agricultural production systems as well as promote the quantity and quality of foods. They can also reduce the global dependence on hazardous agricultural chemicals. Science Direct, Google Scholar, Springer Link, CAB Direct, Scopus, Springer Link, Taylor and Francis, Web of Science, and Wiley Online Library were checked, and the search was conducted on all manuscript sections in accordance with the terms Acinetobacter , Arthrobacter , Enterobacter , Ochrobactrum , Pseudomonas , Rhodococcus , Serratia , Streptomyces , Biostimulants, Plant growth promoting rhizobactera, and Stenotrophomonas . The aim of this manuscript is to survey the effects of plant-growth-promoting rhizobacteria by presenting case studies and successful paradigms in various agricultural and horticultural crops.",
"conclusion": "11. Conclusions In the rhizosphere, bacteria produce phytohormones such as cytokinins, ethylene, abscisic acid, gibberellins, and auxin. PGPR can be categorized into free-living rhizobacteria, which live outside plant cells, and symbiotic bacteria, that can be found inside plants and exchange metabolites with them directly. The PGPR exert their impacts through helping to counteract pathogen attack, facilitating food intake, and regulating plant hormone levels. Bacteria of the genus Arthrobacter can play an important role in plant growth via direct action mechanisms such as increased iron uptake via iron-chelating siderophores, production of phytohormones, production of volatile components, and solubilization of inorganic phosphates that can influence plant signaling pathways and metabolism. Different biotic stress parameters can influence the growth and reproduction of PGPR in plants. Pseudomonas strains are important in the control of different diseases triggered by fungal phytopathogens, such as Fusarium solani causing okra root rot, damping off disease caused by Pythium spp., foliage blight disease caused by Phytophthora nicotianae , and Rhizoctonia solani associated with Rhizoctonia root-rot. Pseudomonas stutzeri , Pseudomonas fluorescens , and Pseudomonas aeruginosa secrete chitinase and are known as good biocontrol agents, while Pseudomonas putida produces different mycolytic enzymes, such as amylase, chitinase, protease, lipase, and cellulase. Streptomyces spp. has been found to significantly enhance the intensity of mycorrhizal root colonization in agricultural crops, whereas root colonization by Streptomyces in legume crops promotes root nodulation frequency, and Acinetobacter can help in plant growth by producing gibberellin, IAA, antibiotics, and siderophore, with an important role in the solubilization of zinc and phosphate. Many strains of Ochrobactrum sp. also have a significant role in toxicity mitigation to different crops, especially legume plants. Rhodococcus species is a soil-borne organism that is widespread in the environment, especially in soil, having an important function in plant growth as well as a wonderful ability to metabolize harmful environmental pollutants. Different Serratia and Stenotrophomonas spp. species have plant-growth-promoting abilities that can influence crops by the production of phytohormone indole-3-acetic acid, which has an important role in plant growth promotion. In conclusion, the application of plant-growth-promoting rhizobacteria-based biostimulant bacteria is both an environmentally friendly practice but also a promising methodology that can noticeably improve the use efficiency of natural resources. The main PGPR effects are regulating plant hormone levels, increasing counteraction to pathogen attack, and improving food intake, yet the application of PGPR in the agricultural industry represents a small fraction within the agricultural industry worldwide, which is mainly because of the inconsistent characteristics of the inoculated PGPR that can influence agricultural production. Important parameters that can influence the successful utilization of PGPR are environmental parameters, the interaction capability with indigenous microflora in soil, the compatibility with the crop on which it is inoculated, and their survival in soil. More studies are needed to determine the important effects of PGPR as important parameters to ensure the productivity and the stability of agricultural systems in sustainable agriculture.",
"introduction": "1. Introduction Biostimulants can be used to complement the application of chemical inputs, including the utilization of beneficial rhizosphere microbiome like advantageous fungi and plant-growth-promoting rhizobacteria [ 1 , 2 , 3 ]. The major biostimulants effects on crops include improving the visual quality of final products, stimulating the immune systems of plants, inducing the biosynthesis of plant defensive biomolecules, removing heavy metals from contaminated soil, improving crop performance, reducing leaching, improving root development, improving seed germination, inducing tolerance to abiotic and biotic stressors, accelerating crop establishment, and promoting nutrient uptake and nutrient use efficiency [ 4 , 5 ]. Biostimulants are components that increase plant growth but do not qualify as essential plant nutrients, but biofertilizers are live microbes whose primary impact is to increase plant growth. Moreover, as biofertilizers are live microbes whose primary influence is to increase crop growth, biopesticides are live organisms whose primary effect is to directly control and manage crop diseases and pests. It is important to consider the point that the main difference between biostimulants and biofertilizers is that biofertilizers contain many nutrients but biostimulants do not have the plant nutrients. Ochrobactrum species are Gram-negative, non-enteric, non-fermenting bacteria that are closely associated with the genus Brucella , which are found in wide range of environments including in animals, plants, soil, aircraft, and water [ 6 , 7 ]. Sipahutar and Vangnai [ 8 ] observed that Ochrobactrum sp. MC22 can improve the yield of soybean and mung bean with significant function for rhizoremediation in a crop area with triclocarban contamination. Acinetobacter is a Gram-negative bacterium found in nature, especially in the rhizosphere of many plants, playing an important function as a plant-growth-promoting bacterium and being known to produce gibberellin, siderophore, IAA, biosurfactants/bioemulsifiers, and antibiotics, as well as solubilize zinc, potassium, and phosphate. Bacteria of the Enterobacter species belong to the ESKAPE ( Enterobacter spp., Pseudomonas aeruginosa , Acinetobacter baumannii , Klebsiella pneumoniae , Staphylococcus aureus , and Enterococcus faecium ) group of pathogens [ 9 , 10 , 11 ]. It has been reported that microbial biostimulants can synthesize IAA, which can promote root branching and plant growth of the plants under both abiotic and biotic stresses [ 4 , 5 ]. Microbial biostimulants can alleviate salt stress as they are associated with a high level of IAA and can improve them in the re-establishing of favorable water potential gradients under water shortage conditions as well as increasing film hydration around the roots [ 1 , 2 , 3 , 4 , 5 ]. Both PGPR and plants have ACC-deaminase, which has the ability to reduce the concentration of ethylene in the root zone and roots, with PGPR-derived ACC-deaminases being able to decrease ethylene-induced inhibition by reducing root zone ethylene [ 12 , 13 , 14 ]. Enterobacter species notably increased below- and aboveground responses in rice plants [ 12 ]. Arthrobacter species are obligate aerobes and Gram-positive chemoorganotrophs that are often found among soil bacteria [ 13 ], which is a major aerobic bacterium under the class of Actinobacteria and the family Micrococcaceae [ 14 ]. The most well-known and important in situ bioremediation of them is Cr(VI) reduction abilities [ 15 ]. The biomass of Arthrobacter protophormiae was used to detach Cd(II) from an aqueous solution [ 16 ]. Arthrobacter echigonensis MN1405 helped Phytolacca acinosa Roxb. in obtaining high remediation effectiveness of Mn removal and accumulation in the Mn contamination area [ 17 ]. Pseudomonas are known as plant-associated and soil-dwelling species because of their biological activity of controlling plant diseases, both indirectly through inducing plant defense resistance responses or directly via producing antagonistic metabolites, especially by their large class of secondary metabolites, which are known as cyclic lipopeptides [ 18 ]. Different species of the Pseudomonas genus are well known as exhibiting plant growth promotion traits, such as indole acetic acid (IAA) biosynthesis, phosphate solubilization, and stress alleviation enzyme production, which are important characteristics for the development of effectual plant biostimulants [ 19 ]. Rhodococcus spp. are a group of non-model Gram-positive bacteria that have various catabolic activities with high adaptive capabilities, making them unique because of different applications in lignocellulosic biomass conversion, environmental bioremediation, and whole-cell biocatalysis [ 20 ]. This genus includes a small number of opportunistic pathogens and species, making them appropriate from the point of view of safety [ 21 ]. Serratia is a Gram-negative, rod-shaped bacterium that is an important ubiquitous member of the Enterobacteriaceae family [ 22 ]. Serratia proteamaculans suppressed Rhizoctonia solani in vivo and in vitro, increased plant growth parameters, and stimulated tomato defense machinery [ 23 ]. Kang et al. [ 24 ] reported that Serratia nematodiphila PEJ1011 can regulate the endogenous ABA levels in pepper plants while reducing the endogenous salicylic acid and jasmonic acid contents. Streptomyces spp. is a filamentous and Gram-positive prokaryote, being the main clade of the phylum Actinobacteria , belonging to the family of Streptomycetaceae [ 25 ], which are ubiquitous in marine sediments and soils; moreover, they are usually found inside plant roots and in the rhizosphere [ 26 ]. It is the most well-known genus of Actinobacteria , being a Gram-positive mycelial sporulating bacteria with a great ability to be resident in dry and saline soils or live within spontaneous plants of drylands, with the ability of producing IAA and siderophores, as well as solubilizing mineral phosphates [ 27 ]. It can be dispersed by arthropods, insects, and other microbes, and it contains unique metabolites that can influence insect behavior and bacterial growth [ 28 ]. The complexity of soil environments and the interactions of Streptomyces with other organisms is the main reason for the production of secondary metabolites [ 29 ]. They have shown great potential for protection against fungal disease, plant growth promotion, and colonization ability in cereals [ 30 ]. Kaari et al. [ 30 ] concluded that Streptomyces sp. UT4A49 can be considered as a promising biocontrol factor for tomato bacterial disease control, that of Ralstonia solanacearum . Elango et al. [ 31 ] reported that Streptomyces lydicus and Streptomyces griseus, together with Trichoderma harzianum and Bacillus subtilis, are recommended for the control of red root rot disease of tea plants. Some of the most important Streptomyces with biocontrol activities are Streptomyces griseus , Streptomyces kasugaensis , Streptomyces J-1, Streptomyces sp., Streptomyces sanglieri , Streptomyces griseorubens E44G, Streptomyces rochei ACTA1551, and Streptomyces felleus YJ1, and notable Streptomyces with plant-growth-promoting activities are Streptomyces anulatus S37, Streptomyces sp., Streptomyces matansis BG5, Streptomyces sp. RSF17, Streptomyces vinaceus CRF2, Streptomyces pulcher CRF17, Streptomyces PRIO41, Streptomyces mutabilis , and Streptomyces fumangs gn-2 [ 32 , 33 ]. Stenotrophomonas is one of the most important aerobic plant-growth-promoting bacterium with around 18 well-characterized species, having shown a high capability in phosphate solubilization, nitrogen fixation, siderophore production, the production of plant growth regulators, and antagonism against pathogenic microorganisms [ 34 ]. Its species also have important functions in the bioremediation process by helping in biofortification and degrading xenobiotic compounds, as well as having significant roles in the improvement of crop plant health [ 35 ]. The action mechanisms of microbial biostimulants in plants are not well understood yet; however, PGPR can upregulate the expression of genes related to cell growth and cellulose biosynthesis; promote shoot length; increase water use efficiency, photosynthesis, and water retention in drought conditions; increase gas exchange; promote plant biomass; decrease the levels of lipid peroxidation; increase organic acid, protein, soluble sugar, and biomass production; improve chlorophyll, ABA levels, and compatible solutes; and enhance the activity of antioxidant enzymes as well as reducing membrane permeability in salt stress conditions [ 3 , 4 , 5 , 6 , 7 , 8 , 9 , 10 ]. The goal of this review article is to survey the impacts of plant-growth-promoting rhizobacteria by presenting case studies and successful paradigms in different agricultural and horticultural crops. The title of this article was selected as PGPR has an important role in plant growth through direct action mechanisms, and understanding the roles and impacts of different types of PGPR is important in terms of achieving more sustainable agricultural goals. In this article, we also tried to study the major mechanisms of action of different types of biostimulant products with an emphasis on the application and integration of microbial-based biostimulant products in horticultural and agricultural crop production. This research examines the scientific literature on biostimulants from 1991 to December 2023 by conducting a bibliometric analysis of the literature published on the Web of Science database, including more than one thousand articles. The information provided was obtained from randomized control experiments, review articles, and analytical observations and studies that have been gathered from various literature sources such as PubMed, Science Direct, Scopus, and Google Scholar. The keywords used were the Latin and common names of different agricultural and horticultural species, as well as microbial biostimulants, such as “ Ochrobactrum ”, “ Acinetobacter ”, “ Arthrobacter ”, “ Enterobacter ”, “ Pseudomonas ”, “ Rhodococcus ”, “ Serratia ”, “ Streptomyces ”, “Biostimulants”, “Plant growth promoting rhizobactera”, and “ Stenotrophomonas ”."
} | 4,009 |
35641308 | PMC9544493 | pmc | 5,530 | {
"abstract": "Summary Fine root endophyte mycorrhizal fungi in the Endogonales (Mucoromycota arbuscular mycorrhizal fungi, M‐AMF) are now recognized as at least as important globally as Glomeromycota AMF (G‐AMF), yet little is known about the environmental factors which influence M‐AMF diversity and colonization, partly because they typically only co‐colonize plants with G‐AMF. Wild populations of Lycopodiella inundata predominantly form mycorrhizas with M‐AMF and therefore allow focussed study of M‐AMF environmental drivers. Using microscopic examination and DNA sequencing we measured M‐AMF colonization and diversity over three consecutive seasons and modelled interactions between these response variables and environmental data. Significant relationships were found between M‐AMF colonization and soil S, P, C:N ratio, electrical conductivity, and the previously overlooked micronutrient Mn. Estimated N deposition was negatively related to M‐AMF colonization. Thirty‐nine Endogonales Operational Taxonomic Units (OTUs) were identified in L . inundata roots, a greater diversity than previously recognized in this plant. Endogonales OTU richness correlated negatively with soil C:N while community composition was mostly influenced by soil P. This study provides first evidence that M‐AMF have distinct ecological preferences in response to edaphic variables also related to air pollution. Future studies require site‐level atmospheric pollution monitoring to guide critical load policy for mycorrhizal fungi in heathlands and grasslands.",
"introduction": "Introduction Mycorrhizal fungi enable plants to obtain up to 80% of their nutritional resources, mostly nitrogen (N) and phosphorus (P) otherwise bound in soil, in exchange for photosynthates (Smith and Read, 2008 ). Factors affecting plant nutrient availability are key drivers of ecosystem processes within heathlands, which are nutritionally poor (Read et al ., 2004 ) and rapidly declining in their British stronghold and across Europe (Diaz et al ., 2006 ). Mycorrhizal species richness within a habitat is directly linked to plant species richness and adaptation to the local environment (Johnson et al ., 2005 , 2010 ) and vice versa. Some host plants demonstrate mycorrhizal fungal specificity or preference (van der Heijden et al ., 2015 ). This is the case of the clubmoss Lycopodiella inundata , a locally rare perennial lycophyte which favours wet heathland habitats. It establishes a mutualism with multiple closely related taxa within the Endogonales (Mucoromycota) arbuscular mycorrhizal fungi (M‐AMF) clade. Many M‐AMF taxa can be harboured within a single host root and thus far, L . inundata has been found to associate primarily with M‐AMF (Hoysted et al ., 2019 ) while other lycopods are colonized predominantly by Glomeromycota AMF (G‐AMF) (Benucci et al ., 2020 ; Rimington et al ., 2020 ). Other plant lineages, however, form endomycorrhizal associations with both G‐AMF and M‐AMF simultaneously (Field et al ., 2016 ; Rimington et al ., 2020 ). In these cases, it is difficult to distinguish microscopically the two endomycorrhizal AM fungal groups within roots, their specific functional roles or responses to environmental variables. Hyphal diameter and vesicle size help distinguish between the two fungal groups. Mucoromycota‐AMF form often‐branching thin hyphae <2 μm (typically 0.5–1.5 μm in diameter), with small (5–15 μm in length) intercalary and terminal vesicles (Hoysted et al ., 2019 ; Kowal et al ., 2020a ). In contrast, G‐AMF hyphae are coarse, with a larger hyphal diameter >3 μm and longer (20–30 μm) vesicles (Orchard et al ., 2017a ; Hoysted et al ., 2019 ). Both G‐AMF and M‐AMF form arbuscules in tracheophytes but due to the finer hyphal diameter of M‐AMF, previous studies have referred to these taxa as fine root endophytes or FRE (Orchard et al ., 2017a ; Hoysted et al ., 2019 ). Thus, M‐AMF‐dominated L . inundata represents an ideal system to study associations between vascular plants and this ubiquitous, but long neglected, mycorrhizal fungal clade. Moreover, it also allows investigations on how this endomycorrhizal symbiosis responds to changes in environmental factors such as atmospheric CO 2 concentration (Hoysted et al ., 2019 ), atmospheric pollution and edaphic variables. Here we focus on how the latter variables affect M‐AMF host plant root colonization, abundance and diversity which remain understudied despite the widespread distribution of M‐AMF across the land plant phylogeny, including food crops (Orchard et al ., 2017a ; Hoysted et al ., 2018 ; Sinanaj et al ., 2020 ). Considerable M‐AMF diversity has been discovered recently and delimited into 36 species in Endogonaceae and Densosporaceae (Endogonales, Mucoromycota) (Rimington et al ., 2018 , 2019 , 2020 ), but most taxonomic levels remain formally undescribed (Bonfante and Venice, 2020 ) and it is still unknown whether all M‐AMF are Endogonales. The placement of M‐AMF and G‐AMF as subphyla has been the subject of recent taxonomic discussion (Spatafora et al ., 2016 ; Orchard et al ., 2017b ), furthered by Endogonales systematics examined by Desirò et al . ( 2017 ). However, as G‐AMF remain a distinct clade from Endogonales (M‐AMF), we maintain the higher taxonomic order of phyla herein (Tedersoo et al ., 2018 ). Molecular, structural and functional differentiation of M‐AMF from G‐AMF (Field et al ., 2016 ; Field et al ., 2019 ) have been starting points to unravel their distinct and complementary ecological roles within shared hosts and habitats (Albornoz et al ., 2020 ). However, despite recent findings on FRE (M‐AMF) prevalence and colonization phenology in L . inundata (Kowal et al ., 2020a ) and the functional role of M‐AMF in L . inundata N uptake (Hoysted et al ., 2019 ) as well as their widespread occurrence in grasses and other vascular plants (Orchard et al ., 2017a ; Albornoz et al ., 2020 ), little is known about the environmental drivers of M‐AMF community composition and colonization. Understanding the ecological dynamics of M‐AMF is important to maintain diverse mycorrhizal communities supporting host plant and habitat resilience, which in the case of this rare lycopod with declining European populations, are critical. Plants associate with different mycorrhizal fungi which facilitate nutrient mobilization and uptake of soil N and P (Read and Perez‐Moreno, 2003 ; de la Fuente Cantó et al ., 2020 ), the two major growth‐limiting minerals required by autotrophic plants. For example, ericoid mycorrhizal fungi (ErM), forming the main endomycorrhizal type on N‐limited heathlands, access organic N for their Ericaceae host plants (Smith and Read, 2008 ; Leopold, 2016 ). In P‐limited habitats, G‐AMF are responsible for up to 100% of P uptake in some plant species (Smith et al ., 2003 ) and are functionally crucial. But recent findings point to fundamentally different nutritional functions between M‐AMF and G‐AMF; G‐AMF are more efficient than M‐AMF in P uptake and transfer to liverwort hosts, independent of N availability. Conversely, host plants colonized only by M‐AMF receive substantial transfer of N, including organic N, from their fungal partners alongside P (Field et al ., 2019 ). Air pollution resulting in excess nutrients in the environment is one of the major threats to biodiversity (CBD, 2019 ; IPBES, 2019 ). At the ecosystem level this results in changes in plant and fungal species composition, loss and/or shifts in plant and fungal species diversity, and nutritional imbalances in plants (SAEFL, 2003 ; Field et al ., 2014 ; Suz et al ., 2014 ; van der Linde et al ., 2018 ). The effects of pollutant critical loads (below which significant adverse effects to the ecosystem do not occur), and critical levels (above which harmful effects may occur) (CLRTAP, 2004 ), may be detected even once a site is no longer in exceedance, as ecosystem recovery might take time (Suz et al ., 2021 ). Long‐term models suggest that acidic heathland habitats, which predominantly harbour ErM, are highly susceptible as their recovery can be prolonged well after pollution has ceased (Payne et al ., 2013 , 2017 ; Stevens, 2016 ). Specifically, N deposition pollution affects soil characteristics which are important in shaping ErM diversity in heathlands (van Geel et al ., 2020 ). Excess soil N or changes in pH can also influence plant nutrient availability, increasing prevalence of N‐tolerant plant species, and ultimately altering species composition through shading or competition (Stevens et al ., 2018 ). The link between woodland‐dominant ectomycorrhizal fungi (EcM) and changes in plant nutrient status through N deposition to ecosystems is also well established (van der Linde et al ., 2018 ; Suz et al ., 2021 ), and similar effects have been reported for N additions to G‐AMF (Corkidi et al ., 2002 ; Johnson, 2010 ; Liu et al ., 2012 ; Jiang et al ., 2018 ; Ceulemans et al ., 2019 ). In this study we investigated abiotic soil and environmental interactions which influence M‐AMF plant root colonization, richness and community composition in heathlands across environmental gradients while generating diversity data for these groups of fungi. We hypothesized that N deposition will be one of the main factors influencing these fungi. We also studied whether modelled air pollution could be related to soil covariates known to affect G‐AMF (Johnson, 2010 ). Finally, we investigated the presence of G‐AMF in L . inundata roots and explored their potential contribution to host plant nutrition.",
"discussion": "Discussion Lycopodiella inundata consistently hosts M‐AMF Phylogenies and OTU assemblages using near‐complete 18S DNA sequences indicate that L . inundata hosts at least 39 M‐AMF OTUs distributed within seven taxonomic clades across 13 Endogonales clades. This represents a significant advance in our knowledge of M‐AMF diversity among early‐diverging vascular plants. Earlier phylogenies including M‐AMF detected in roots of L . inundata were either poorly resolved within Densosporaceae (Rimington et al ., 2015 ) or based on partial (400–700 bp) 18S rDNA sequences (Hoysted et al ., 2019 ). Furthermore, our OTU accumulation curves indicate that there may be yet more diversity to uncover (Fig. S2 a). The presence of three main M‐AMF OTUs across most sites highlights the ubiquitous presence of these fungal lineages and their consistent association with L . inundata . We only detected G‐AMF in 7.5% of samples across the 12 sites. These T1 molecular findings agree with the root colonization observed across sites, except for the Netherlands, where G‐AMF structures were observed microscopically but not detected in the DNA analyses. Environmental factors affect M‐AMF colonization, community richness and composition Our analyses show that atmospheric pollution may be affecting M‐AMF colonization levels and OTU richness, at least indirectly. This is indicated by the negative relationship between percentage of roots colonized and total N deposition. Also, our model indicates that soil C:N is a key variable affecting M‐AMF colonization negatively. This is similar to other studies focused on G‐AMF (Johnson et al ., 2010 ; Tedersoo and Bahram, 2019 ). In fact, soil C:N has been shown to be affected negatively by N deposition across Europe (Mulder et al ., 2015 ), in both grasslands and moorlands (Evans et al ., 2006 ; Volk et al ., 2016 ). We found indirect links between air pollution variables and M‐AMF OTU richness. Soil C:N was negatively correlated with M‐AMF OTU richness, raising the possibility that higher N atmospheric pollution could indirectly lead to a decrease in M‐AMF symbionts and changes in diversity and composition of AMF functional groups as previously observed in soil carbon richness gradients (Johnson et al ., 2013 ). In forest systems, atmospheric N negatively affects diversity and composition of mycorrhizal fungi (Lilleskov et al ., 2019 ). However, M‐AMF's specialization for providing N from organic sources to host plants, at least in microcosm experiments with dual G‐AMF/M‐AMF host liverworts (Field et al ., 2019 ), may be critical for host plant resilience when changes in levels of inorganic N from air pollution results in an imbalance of N resources. Nonetheless, given that little is still known with respect to C:N dynamics in G‐AMF (Corrêa et al ., 2015 ) and M‐AMF, field manipulation studies are needed to improve predictions of these feedback cycles. Thus far, the direct impact of N deposition on M‐AMF diversity and the role of certain M‐AMF OTUs in mining organic N in heathlands remain to be tested. The unique M‐AMF OTUs identified from Scottish root samples, where atmospheric N is lowest, may provide a clue where to begin investigations. The strong relationships we found between several soil variables and the extent of M‐AMF colonization are consistent with similar research on Trifolium subterraneum colonized by M‐AMF (Albornoz et al ., 2020 ). We also found distinct soil environmental niches between M‐AMF and G‐AMF presence regarding soil pH. Our BD model (site level) shows a negative relationship between pH and M‐AMF colonization, but a positive relationship with G‐AMF. This agrees with Tedersoo et al . ( 2020 ) where they found contrasting ecological preferences between M‐AMF and G‐AMF in extensive soil environmental DNA sampling across habitats in Estonia and North Latvia, with M‐AMF preferring acidic soils. In that study, pH had the strongest effect on the diversity of fungi. This finding is also consistent with Albornoz et al . ( 2022 ) which found M‐AMF preference for acidic soils across a wide sampling of agricultural sites in Australia. Still, more studies have focused on G‐AMF ecological requirements, without differentiating M‐AMF. For instance, in semi‐natural plant–soil feedback systems, soil pH is the principal driver affecting G‐AMF community composition (Dumbrell et al ., 2010 ). Similarly, presence of keystone G‐AMF taxa in agroecosystems is best explained by soil pH, P levels, bulk density and salinity (Liu et al ., 2014 ; Banerjee et al ., 2019 ), but these studies ignored M‐AMF. Our NMDS analysis shows a similar influence of soil P in M‐AMF composition across sites. Both models indicated that M‐AMF colonization is related to soil S (negatively) and Mn (positively) at both plot and site levels and M‐AMF density per root is also negatively related to both S and Mg. Soil sulfate is linked directly to the atmospheric concentrations of sulfur dioxide (Feinberg et al ., 2021 ). Reductions in SO 2 levels over the last decades in both the United Kingdom and the Netherlands may be relevant to the observed pH effect on colonization in this study. In a related heathland manipulation study, Tibbett et al . ( 2019 ) found that elemental S additions were the primary factor affecting soil pH and a negative G‐AMF colonization response. Our temporal data suggest significant negative correlations between SO \n x \n deposition and M‐AMF colonization in spring, followed by negative correlations between soil S and M‐AMF colonization in autumn. Thus, a reduced SO \n x \n deposition followed by limited availability of soil S could be affecting root colonization by M‐AMF and heathland recovery in general, but further work is needed to assess this M‐AMF specific response. It is possible that in acidic soils, such as in heathlands, high levels of these micronutrients are required because of their poor solubility (Millaleo et al ., 2010 ). The precise role of micronutrients such as Mn and Mg has been less tested than other soil variables and nutrients but our study indicates they are important indirect factors likely affecting soil pH, at least in heathlands. Presence of Glomeromycota in Lycopodiella inundata roots Four of the eight G‐AMF DNA sequences were from roots also colonized by M‐AMF, confirming that some plants of L . inundata are co‐colonized by both groups of fungi. This finding is in keeping with the association between G‐AMF and other Lycopodiopsida (Rimington et al ., 2015 ) and dual colonization by M‐AMF and G‐AMF across different plant lineages as seen with mutualisms in liverworts (Field et al ., 2016 ; Rimington et al ., 2020 ), grasses (Hoysted et al ., 2019 ) and angiosperms (Orchard et al ., 2017a ). Nonetheless, the strong preference for M‐AMF by L . inundata is certainly consistent. We found a positive relationship between soil P and the presence of G‐AMF in some L . inundata roots. Experimental microcosms using dual G‐AMF and M‐AMF host plants show that G‐AMF may be more efficient than M‐AMF in supporting plant P acquisition (Field et al ., 2019 ; Hoysted et al ., 2019 ). While not directly comparable, this could help explain the rare presence of G‐AMF. Furthermore, we observed that the rare G‐AMF structures within L . inundata were more likely to be present in roots also colonized by M‐AMF rather than occurring on their own. This suggests that L . inundata largely relies on M‐AMF for its P requirements as previously assumed (Hoysted et al ., 2019 ), but may also recruit G‐AMF symbionts under certain opportunistic conditions, such as when N deposition is high and/or M‐AMF OTU richness is lower. This differs from previous studies (Orchard et al ., 2017a ) which suggested that M‐AMF enhance host plant P uptake rather than providing primary access to P. Further field studies focusing on the functional role of these fungi are needed. Our model also showed a positive relationship between NO \n x \n and G‐AMF. However, despite data suggesting some Glomus (G‐AMF) species might be N‐tolerant, responses to N deposition by G‐AMF can be variable (Treseder et al ., 2007 , 2018 ) and given their rarity in L . inundata roots, extensive root DNA sequencing across pollution gradients would be required to confirm this relationship. Limitations of local and national modelling interactions and grid resolutions The models may be underestimating or masking the relationship strength between M‐AMF colonization and N deposition due to limitations relating to EMEP model resolution and collinearity among the N covariables (Methods S4 and S5 ). Work carried out as part of the UK Joint Nature Conservation Committee‐led project Nitrogen Futures has demonstrated the variability within grids when modelled at different resolutions. Mean NH 3 concentration and N deposition across all locations in a grid was higher with a resolution of 1 × 1 km 2 compared to 2 × 2 m 2 . Conversely, the maximum for any location in the grid was higher at the lower resolution (Thomas et al ., 2020 ). Future directions The season‐specific correlations observed provide some evidence that relationships with covariates may not be consistent over time due to variation in climate factors and changes during the growing season. Therefore, future work should investigate potential interactions between time of year and abiotic drivers. Leaf C content significantly correlated with M‐AMF colonization, suggesting there may be a link between root colonization and plant tissue C content, as previously shown (Zhu et al ., 2014 ; Mathur et al ., 2018 ), and this may provide a simple non‐invasive tool to infer relative host C allocation to these fungi. We observed that one site in southern England with low M‐AMF colonization (Aldershot), which is also the L . inundata population under greatest decline, lacked the dominant OTU1, present however in all other study sites. Roots in neighbouring sites where OTU1 was present had much higher colonization despite having similar N deposition values. This could suggest that OTU1 may provide their host plants with coping mechanisms to N deposition stress. It is also possible that the widely dispersed OTU1 is present but not yet detected in Aldershot by our sampling effort or it may be unable to compete with the vegetation changes occurring near this population. Transfer experiments of plants hosting OTU1 from nearby thriving populations in southern England and monitoring whether this facilitates population stability and growth over time would allow testing this hypothesis. We expected to find G‐AMF more commonly and opportunistically colonizing L . inundata where G‐AMF plants (e.g. Molinia caerulea ) were more dominant and less where ErM, NM and/or EcM plants were more frequent. However, given the low variability in vegetation composition across our sites, we were unable to test vegetation as a categorical predictor of M‐AMF and G‐AMF colonization of L . inundata roots. Molecular analyses coupled with experimental microcosms – testing donor and target plants – would help disentangle this putative association. This study found that several soil characteristics influence M‐AMF colonization and richness. Furthermore, our analyses also indicate that atmospheric pollution may indirectly interact with these same soil characteristics, and therefore indirectly influence M‐AMF colonization and community composition. However, higher‐resolution air pollution monitoring is needed, at the field experiment scale, to couple air pollution monitoring data with M‐AMF resilience and diversity measures. Without such an investment to test and set air pollution critical load and levels specifically for mycorrhizal fungi in vulnerable habitats such as heathlands, we may be overlooking irreversible ecosystem changes occurring belowground."
} | 5,402 |
36821718 | null | s2 | 5,531 | {
"abstract": "Chemically labile ester linkages can be introduced into lignin by incorporation of monolignol conjugates, which are synthesized in planta by acyltransferases that use a coenzyme A (CoA) thioester donor and a nucleophilic monolignol alcohol acceptor. The presence of these esters facilitates processing and aids in the valorization of renewable biomass feedstocks. However, the effectiveness of this strategy is potentially limited by the low steady-state levels of aromatic acid thioester donors in plants. As part of an effort to overcome this, aromatic acid CoA ligases involved in microbial aromatic degradation were identified and screened against a broad panel of substituted cinnamic and benzoic acids involved in plant lignification. Functional fingerprinting of this ligase library identified four robust, highly active enzymes capable of facile, rapid, and high-yield synthesis of aromatic acid CoA thioesters under mild aqueous reaction conditions mimicking in planta activity."
} | 246 |
25229207 | PMC4199106 | pmc | 5,532 | {
"abstract": "DNA nanostructures constitute attractive devices for logic computing and nanomechanics. An emerging interest is to integrate these two fields and devise intelligent DNA nanorobots. Here we report a reversible logic circuit built on the programmable assembly of a double-stranded (ds) DNA [3]pseudocatenane that serves as a rigid scaffold to position two separate branched-out head-motifs, a bimolecular i-motif and a G-quadruplex. The G-quadruplex only forms when preceded by the assembly of the i-motif. The formation of the latter, in turn, requires acidic pH and unhindered mobility of the head-motif containing dsDNA nanorings with respect to the central ring to which they are interlocked, triggered by release oligodeoxynucleotides. We employ these features to convert the structural changes into Boolean operations with fluorescence labelling. The nanostructure behaves as a reversible logic circuit consisting of tandem YES and AND gates. Such reversible logic circuits integrated into functional nanodevices may guide future intelligent DNA nanorobots to manipulate cascade reactions in biological systems.",
"discussion": "Discussion We have demonstrated the programmable assembly of [3]pseudocatenane with two branched-out heads functionalized with i-motif and G-quadruplex DNAs. The mobility of interlocked circles triggered by the ROs, and the proton-induced cyclizations of their head-motifs are verified by gel electrophoresis, AFM and fluorescence experiments. The whole process can be reset by the addition of cROs and an increase of the pH to 8, to remove the ROs and to disrupt the i-motif structure, respectively. The mobility and cyclization steps can easily be monitored by the RG/BHQ1 and Cy3/Cy5 system, respectively, that behave as two tandem logic gates (YES and AND). We show that the interlocking allows for the direct coupling between the i-motif assembly and the formation of the G-quadruplex structure. By this means, a novel reversible fluorescent logic circuit is built and then employed to control a repeatedly operating DNAzyme nanoswitch on the basis of difference in robustness between the designed bimolecular i-motif and the G-quadruplex. An important feature of the here described logic device is its robust reversibility. To our knowledge, so far only a single example of a reversible DNA-based logic gate has been described by Turberfield and colleagues 21 . They described a reversible AND gate comprising a DNA hairpin motif that can balance between ON and OFF conformations depending on toehold exchange of complementary input ODNs. The ON-state was reported by an increase in fluorescence by an opened molecular beacon. As pointed out by Turberfield 21 , reversibility of DNA logic circuits is important, (i) to better respond to changes in inputs, (ii) to reduce, or even avoid, error accumulation, and (iii) to prevent hysteresis resulting from sequential inputs. We introduce a completely different system that consists of tandem YES and AND gates, and show how to employ a reversible DNA logic circuit to control a complex DNA nanostructure by switching DNAzyme activity reversibly and repeatedly. These systems can be envisioned to be combinable with other reversible DNA logic circuits to achieve further cascading of logic gates. The two heterogeneous inputs, H + and toehold ROs, could potentially be coupled to more complex future functional DNA nanostructures. To achieve a possible future coupling of our circuit with other DNA computing devices in follow-up applications, an option would be to operate our system via photoswitched ROs 28 and light-controlled pH change 48 49 that has successfully been applied to non-reversible DNA logic arrays 50 . Importantly, no additional ROs, cROs and proton are required for our reversible DNA circuit, which should further facilitate the use of this system as a component in larger DNA computing circuits. Our study demonstrates how to harness unique and precisely controllable structural features of interlocked DNA nanoarchitectures for functionalization, and provides a further important step towards integrating DNA computing with functional nanodevices. Such reversible logic circuits integrated into functional nanodevices may serve as a guide for intelligent DNA nanorobots to manipulate cascade reactions in biological systems 14 51 52 ."
} | 1,081 |
23055455 | null | s2 | 5,534 | {
"abstract": "One on each side: gold nanoparticles (AuNPs) and semiconducting quantum dots (QDs) are integrated on a single DNA origami scaffold. Streptavidin-functionalized QDs bind to biotin anchors on one side of the DNA origami, while DNA-coated AuNPs bind through DNA hybridization to single-stranded DNA on the other side of the scaffold. This approach offers a new path toward the organization of complex systems consisting of disparate materials."
} | 110 |
35771939 | PMC9271185 | pmc | 5,535 | {
"abstract": "Significance Biofilms are an important mode of bacterial growth where surface-attached cells enjoy survival advantages provided by the extracellular matrix composed of secreted, diffusible biopolymers. An important question is how diffusible matrix production can be stable under the evolutionary pressure of exploitation. Here, we show that some matrix proteins in Vibrio cholerae biofilms can indeed be exploited by nonproducers, but the exploitation happens within a quantifiable spatial range around the producer cell clusters. An ecological model considering the length scales of exploitation and cell group structure reveals the stable conditions for diffusible matrix production that are consistent with those in natural environments. Our study provides concepts and tools for studying public goods sharing in spatially structured bacterial biofilms.",
"discussion": "Discussion Cell–cell attachment and cell–surface adhesion are often critical to the evolutionary advantage of biofilm formation. The extracellular matrix plays an important role in controlling biofilm organization and adhesion ( 37 – 39 , 45 , 49 – 54 ), and understanding the evolutionary dynamics of its production remains an active area of work in this research domain. The biofilm matrix often contains components that vary in their diffusion properties and potential for exploitation by cells that do not produce them. In this study, we explored the population dynamics of diffusible matrix protein production in V. cholerae biofilms. We carried out competition assays under static and flow conditions, both of which are frequently encountered by V. cholerae in native habitats and in hosts ( 47 , 48 , 55 ). We found that the matrix components RbmC and Bap1, as diffusible public goods, can be exploited and confer resistance to physical disturbance to clusters of nonproducing cells within a finite spatial range around producer cell clusters. This exploitation range depends on the diffusion–advection condition of the environment, where a continuous flow effectively shrinks the exploitation range and consequently suppresses cheater exploitation. In their marine and freshwater reservoirs, V. cholerae cells often attach to floating food particles and experience shear flows ( 47 , 48 ). During infection, V. cholerae biofilms are primarily localized on the tips of intestinal villi and therefore experience peristaltic flow in the gut ( 55 ). The flow speed and the associated shear stress used in this study fall within the range measured in these environments ( 56 ). The population densities of V. cholerae cells in natural ecosystems have been measured as ranging from 10 2 to 10 5 cells per liter ( 57 ), corresponding to an average distance between cells on the order of millimeters in 3D space. Thus, our results suggest that the combination of flow and low cell density drives the competition dynamics in favor of RbmC and Bap1 producers and leads to the evolutionary stability of these diffusible matrix proteins in V. cholerae . It is interesting to contrast our results with those for the other major matrix protein in V. cholerae biofilms, RbmA, which (together with VPS) is responsible for high-density cell–cell packing within biofilm cell clusters. RbmA is secreted and shared in a limited fashion within cell lineage groups producing it, conferring protection from exploitation with little dependence on the distance between clusters of producing versus nonproducing cells ( 4 ). We envision that the different sharing dynamics of RbmA, RbmC, and Bap1 are constrained by their function as components of the biofilm matrix. RbmA, which holds mother-daughter cell lineages together, must be sequestered and stay in close proximity to producing cells to perform this function. RbmC and Bap1, on the other hand, must travel a distance away from producing cells to confer adhesion between cell groups and underlying surfaces, making them inherently sharable and exploitable; the exact biochemical and biophysical mechanisms underlying this difference await future research. As a consequence, different matrix components—even those produced by cells of one species within the same biofilm—can have different environmental constraints and population structures as conditions for their evolutionary stability. Our results have a straightforward correspondence with social evolution theory. Hamilton’s rule provides the canonical condition under which cooperation is favored by selection: The fitness benefit of receiving cooperative help, weighted by the relatedness coefficient that quantifies the correlation between recipient genotype and cooperative actor genotype ( 58 – 60 ), must exceed the cost of the cooperative behavior ( 18 ). Putting this principle into the context of our experiments, the cost of RbmC and Bap1 production is set by their regulation (here, nearly constitutive) and nutrient supply conditions, and the benefit is set by the extent to which shear stress is applied to biofilms ( SI Appendix , Fig. S8 ). The relatedness coefficient—taking note again that this effectively measures the extent to which RbmC/Bap1-producing cells benefit each other relative to the total population composition ( 28 , 58 , 60 – 62 )—is controlled by the range over which RbmC and Bap1 diffuse and the spatial distribution of producer and cheater cell clusters ( 26 , 28 , 29 , 58 , 62 , 63 ). In many reported cases, the clustering of clonal cell lineages within mixed biofilm communities can be sufficient to stabilize cooperation against cheating ( 25 – 29 ). Here, however, the spatial range of RbmC/Bap1 sharing was larger than the size of clonal cell clusters due to the leakage of these proteins from groups of producing cells, and the associated public goods dilemma could not be resolved by clonal group clustering alone. Instead, the solution is rooted in the combination of biophysical mechanisms of protein retention and diffusion, community spatial structure, and environmental perturbations, all captured in the spatial exploitation model ( Fig. 3 ). The key parameter in our model, the exploitation radius R , can additionally depend on many physiological factors including the production rate of the matrix, duration of biofilm growth, aggregation of adhesion proteins, uptake of adhesion protein by cheater matrix, and biofilm dispersal ( 64 ). Competition for binding to the exopolysaccharides and for surface adsorption from other biomacromolecules in the environment may also reduce R ( 65 ). In addition, although we have assumed R to be constant in our spatial model of P protection , R does appear to vary and drop to a much smaller value when the inoculation number density goes above 10 −1 cells/100 µm 2 ( SI Appendix , Fig. S9 ). This can be understood as follows: At a high total inoculation number density, the size of each producer cluster is smaller due to nutrient limitation, and the concentration of secreted proteins around each producer cluster is consequently reduced. Additionally, the elevated incorporation of adhesion proteins into the surrounding cheater clusters further constrains the exploitation range. In the limit of a confluent biofilm layer, the system reduces to the well-studied case of a densely packed bacterial colony ( 21 , 23 ). Another approximation in our P protection model is that the density of producer clusters is taken as the inoculation number density of producer cells, σ 0 , p . In practice, we observed a modest increase in the number of producer clusters at the end of the experiment due to cell dispersal and recolonization ( SI Appendix , Fig. S10 ). To account for these details, a multiscale model that considers these additional factors should be implemented, which we leave for future studies. Nevertheless, our simplified spatial model already provides a good quantitative description of the underlying dynamics of cooperative matrix protein production, especially in the low seeding density regime most relevant to the natural conditions of V. cholerae . While the concept of exploitation range has been discussed in relation to public goods production in bacterial communities ( 26 – 29 , 66 – 69 ), direct quantitation of the exploitation radius and a close comparison between model and experiment, as we achieved here, represent important steps forward. Our study thus provides both a conceptual guideline and technical tool set for future studies on how to quantify public goods sharing in relation to population structure in the biofilm context. Finally, previous work has shown that in the cases of extracellular enzymes ( 27 , 70 , 71 ), siderophores ( 23 , 72 ), and autoinducers ( 73 – 75 ) as diffusive public goods, clonal segregation and efficient consumption of public goods (or their enzymatic products and complexes) through high local cell density and a high uptake rate can minimize the potential for exploitation in the context of continuous films ( 21 , 26 , 27 ). In the context of spatially discrete clonal clusters, we expect the evolutionary stability of these public goods to be determined by the interplay between the exploitation range and the spatial structure as shown here."
} | 2,302 |
34208358 | PMC8231175 | pmc | 5,537 | {
"abstract": "Cell-free synthetic biology is a maturing field that aims to assemble biomolecular reactions outside cells for compelling applications in drug discovery, metabolic engineering, biomanufacturing, diagnostics, and education. Cell-free systems have several key features. They circumvent mechanisms that have evolved to facilitate species survival, bypass limitations on molecular transport across the cell wall, enable high-yielding and rapid synthesis of proteins without creating recombinant cells, and provide high tolerance towards toxic substrates or products. Here, we analyze ~750 published patents and ~2000 peer-reviewed manuscripts in the field of cell-free systems. Three hallmarks emerged. First, we found that both patent filings and manuscript publications per year are significantly increasing (five-fold and 1.5-fold over the last decade, respectively). Second, we observed that the innovation landscape has changed. Patent applications were dominated by Japan in the early 2000s before shifting to China and the USA in recent years. Finally, we discovered an increasing prevalence of biotechnology companies using cell-free systems. Our analysis has broad implications on the future development of cell-free synthetic biology for commercial and industrial applications.",
"introduction": "1. Introduction In 1961, Heinrich Matthaei and Marshall Nirenberg used Escherichia coli S30 extract to decipher the genetic code, which later earned them the 1968 Nobel prize in Physiology or Medicine [ 1 ]. Since then, cell-free protein synthesis (CFPS) has enabled the engineering and study of cellular processes without using living cells [ 2 , 3 , 4 , 5 , 6 , 7 , 8 ]. To achieve CFPS, cellular machinery for transcription, translation, and energy regeneration is extracted from cells, then assembled with reaction components in a range of reaction vessel types, including well-plates [ 9 ], droplets [ 10 ], or liposomes [ 11 ] ( Figure 1 ). CFPS offers several advantages over other methods. First, the open nature of the reaction allows the user to directly influence the biochemical systems of interest. As a result, new components can be added or synthesized, and these can be maintained at precise concentrations. This can include, for example, non-canonical amino acids for incorporation into proteins that expand the chemistry of life. Second, the chemical environment can be controlled, actively monitored, and rapidly sampled. Third, processes that take days or weeks to design, prepare, and execute in cells can be done more rapidly in a cell-free system, since no time-consuming cloning steps are needed and reactions are easily automated [ 5 , 8 , 12 ]. Despite the advantages, CFPS still has challenges to widespread commercial adoption including costs, enzyme stability, reaction duration, inefficiency of some post-translational modifications, and unfamiliarity with the technology [ 5 ]. Recent work has attempted to overcome these challenges. For instance, E. coli strains were modified to enhance protein synthesis yield and reduce production costs [ 13 ]. Furthermore, protein synthesis machinery was enhanced to synthesize proteins with glycosylation and non-canonical amino-acids [ 14 , 15 , 16 , 17 ]. In addition, the preparation process of CFPS was simplified and streamlined to minimize cost and batch variability [ 18 , 19 ]. These recent innovations have boosted the commercial interest in CFPS. Several companies now sell CFPS kits, including Life Technologies, Promega, Arbor Biosciences, Thermo Scientific, WEPRO, Qiagen, Lenio Bio, New England Bio Labs, and GeneFrontier, among others [ 20 ]. There are also increasing interests in using CFPS to produce commercial therapeutic proteins, synthesize glycoproteins, prototype cellular pathways, enable portable molecular diagnostics, and synthesize virus-like particles as vaccines [ 8 , 17 , 21 , 22 , 23 ] ( Figure 1 ). Recent papers have reviewed the history of CFPS and its myriad of applications [ 4 , 5 , 24 ]. Based on the recent and rapid development in CFPS, what can we expect for the future of cell-free synthetic biology? Can we learn lessons from the field’s history to guide future development? For instance, which areas of innovation usher in new CFPS-based applications? Answers to these questions may help us prioritize the next investments in cell-free systems both in academia and industry. Toward these answers, patents are often used as the indicators for future commercial applications or investment activity. Patent analysis has been invaluable in clarifying the trend and magnitude of technology development [ 25 ]. However, patent and publication trends are unable to fully capture the exact state of research in a given field as the underlying economic, scientific, and societal factors coalesce into the observed output of papers and patents. Here, we cataloged patent publications and peer-reviewed manuscripts relating to CFPS from public databases. We then curated these publications by subject matter, applicants/authors, and countries. Finally, we analyzed the type of innovations that spurred new waves of development in cell-free systems.",
"discussion": "4. Discussion Here, we analyzed the emergence of the cell-free synthetic biology field based on publication and patent frequency. The picture that emerges is that system development and applications are coming of age, with an increasing rate of research and potential. There are a few caveats in analyzing the progress and impact of a field using the patent and publication data. This study focuses on the total count of published patents and peer-reviewed manuscripts and their separation into various groups based on relevant metadata, such as their topic and authors. It does not investigate the long-term impact of these documents. Few highly influential advancements could impact the field far more than numerous others. We did not investigate the revenue generated by each patent, but our data suggest that various stakeholders were willing to invest the resources and effort necessary to file patents on the technology. It is also important to note that the observed changes in publications or patent filings are influenced by a variety of economic and societal factors that the current data set is unable to distinguish between, making the establishment of causal connections challenging. Despite these caveats, we can make several conclusions. The number of patents and publications have been steadily increasing in cell-free synthetic biology over the last decade ( Figure 2 ). Between 2015 and 2020, the most rapid growth in both patents and publications was related to improving CFPS systems ( Figure 3 ). The constant growth in publications relates to applications of CFPS, which exceeded the other categories in recent years ( Figure 3 ). In addition, there is growing commercial interest in CFPS technology by companies, as evident by numerous new start-ups in this space ( Figure 4 a). China has shown the most rapid growth in both publications and patents over the past few years, while Japan has shown diminishing numbers ( Figure 5 ). Based on these conclusions, we envision a few developments that could enhance investment and research interest in CFPS technology. Recent work has investigated the robust quality control of CFPS [ 50 ]. The focus on quality control will be critical to delivering cell-free products for practical use, especially in applications such as on-demand therapeutic production and point-of-care detection [ 21 ]. Furthermore, production yields of green fluorescent protein (GFP) are often used to benchmark the quality of CFPS, but GFP production does not reflect actual use cases. For instance, our recent work has benchmarked a new CFPS system using a range of proteins with different sizes, including a peptide nanocage, GFP, and Cas9 [ 51 ]. The creation of robust and multifunctional CFPS systems will also allow production of a broad range of proteins, varying in protein size, post-translational modification, and folding. This includes peptides and proteins with non-canonical amino acids [ 52 , 53 , 54 , 55 ]. In addition, the drastic reduction of the cost of CFPS systems and the ability to scale reactions will allow for the incorporation of these systems into routine lab workflows. Finally, the ability to readily store, distribute, and activate freeze-dried cell-free systems by simply adding water has opened new opportunities for on-demand biomanufacturing and point-of-care diagnostics [ 8 , 17 ]. Since CFPS is still less known and used in society, further development of educational kits [ 43 , 56 , 57 ] and courses will increase exposure and awareness of the technology to encourage the long-term expansion of the field."
} | 2,181 |
35247636 | null | s2 | 5,538 | {
"abstract": "Across diverse research and application areas, dynamic functionality-such as programmable changes in biochemical property, in mechanical property, or in microscopic or macroscopic architecture-is an increasingly common biomaterials design criterion, joining long-studied criteria such as cytocompatibility and biocompatibility, drug release kinetics, and controlled degradability or long-term stability in vivo. Despite tremendous effort, achieving dynamic functionality while simultaneously maintaining other desired design criteria remains a significant challenge. Reversible dynamic functionality, rather than one-time or one-way dynamic functionality, is of particular interest but has proven especially challenging. Such reversible functionality could enable studies that address the current gap between the dynamic nature of in vivo biological and biomechanical processes, such as cell traction, cell-extracellular matrix (ECM) interactions, and cell-mediated ECM remodeling, and the static nature of the substrates and ECM constructs used to study the processes. This review assesses dynamic materials that have traditionally been used to control cell activity and static biomaterial constructs, experimental and computational techniques, with features that may inform continued advances in reversible dynamic materials. Taken together, this review presents a perspective on combining the reversibility of smart materials and the in-depth dynamic cell behavior probed by static polymers to design smart bi-directional ECM platforms that can reversibly and repeatedly communicate with cells."
} | 399 |
35056492 | PMC8779262 | pmc | 5,540 | {
"abstract": "Based on the importance and sensitivity of microbial communities to changes in the forest ecosystem, soil microorganisms can be used to indicate the health of the forest system. The metagenome sequencing was used to analyze the changes of microbial communities between natural and plantation Castanea henryi forests for understanding the effect of forest types on soil microbial communities. Our result showed the soil microbial diversity and richness were higher in the natural forests than in the plantation. Proteobacteria , Actinobacteria , and Acidobacteria are the dominant categories in the C . henryi rhizosphere, and Proteobacteria and Actinobacteria were significantly enriched in the natural forest while Acidobacteria was significantly enriched in the plantation. Meanwhile, the functional gene diversity and the abundance of functions in the natural forest were higher than that of the plantation. Furthermore, we found that the microbial network in the natural forests had more complex than in the plantation. We also emphasized the low-abundance taxa may play an important role in the network structure. These results clearly showed that microbial communities, in response to different forest types, provide valuable information to manipulate microbiomes to improve soil conditions of plantation.",
"conclusion": "5. Conclusions In summary, we found that the microbial diversity and richness of natural forest were significantly higher than that of plantation forest. Acidobacteria , Proteobacteria , and Actinobacteria are the three dominant phyla, and the abundance was affected by different forest types and cultivated varieties. Proteobacteria and Actinobacteria are enriched in natural forests, while Acidobacteria is massively enriched in the plantation. Meanwhile, we found that a more complex network structure and more positive interactions exist in natural forest than plantation. The keystone taxa are mostly rare and moderate abundant taxa. Our results provide a scientific basis for the sustainable management of the plantation. Furthermore, the follow-up work should further study the variation law and influencing factors of microbial community under seasonal and interannual differences for realizing more reasonable and effective agricultural management measures in the future.",
"introduction": "1. Introduction Soil microorganisms are an important part of the soil ecosystem [ 1 , 2 , 3 ]. The diversity of soil microbial composition and function play a critical role in maintaining soil productivity and stability, such as nutrient cycling [ 4 , 5 ], and pollutant degradation [ 6 , 7 ]. Meanwhile, rhizosphere microbiota can largely determine plant metabolism and physiological activities, including driving nutrient acquisition [ 8 , 9 ], regulating plant growth [ 10 ], and protecting hosts from biological and abiotic stresses [ 11 , 12 , 13 ]. Soil microbial community structure was reported to be sensitive to changes in soil environment. Due to the complexity of forest soil ecosystem, vegetation types [ 14 , 15 ], fertilization [ 16 , 17 ], irrigation [ 18 , 19 ], and land-use type [ 20 ] have a great impact on the number and species of soil microorganisms. For instance, the conversion of natural forest to poplar forest plantation reduced the organic matter content and humidity values and lowered the diversity of soil microbial communities [ 21 ]. After the natural forest was transformed into rubber plantation, the content of soil organic matter and total nitrogen decreased and the abundance of Actinomycetes , arbuscular mycorrhizal fungi, fungi, bacteria and protozoa decreased significantly [ 22 ]. Meanwhile, the tree species could influence soil microbial community [ 23 , 24 ]. Studies had shown that fungal diversity increased with the increase in tree diversity [ 25 , 26 ]. In addition, vegetation types can also affect the richness and activity of soil microorganisms by plant litter. Cong et al. [ 14 ] considered that significant differences in soil microbial communities among three vegetation types (a coniferous forest, a mixed broadleaf forest, and a deciduous broadleaf forest) were probably explained by decomposing the plant litter with different chemical structures. In view of the sensitivity of soil microbial community structure and function to environmental changes, the characteristics of soil microbial community structure can be used as one of the important indicators of soil quality changes [ 27 , 28 ]. Castanea henryi , an important economic species, has a long history of artificial cultivation and is mainly distributed in Zhejiang, Jiangxi, Hunan, and Fujian, among which Fujian has the largest planting area and yield in the country [ 29 , 30 ]. With the continuous expansion of artificial C. henryi planting regions, the area of the remaining natural forest has dropped sharply, which was half of that of plantation forest [ 31 ]. The natural forest has complex community structure and plant species, while the plantation is characterized by intensive monoculture [ 32 ]. The amount of litter returned in plantation is significantly lower than that in natural forest [ 33 , 34 , 35 ]. In addition, long-term fertilization and spraying pesticides in plantation have led to soil acidification, consolidation, soil erosion, and organic pollution [ 36 , 37 ], which would destroy the dynamic balance of soil microbial community [ 38 , 39 ]. According to the statistics, in 2016, the area of soil erosion of C. henryi forest in Songxi County, Fujian Province, was 665.8 hectares, accounting for 90% of the area of C. henryi forest [ 40 ]. A large number of researchers showed that soil erosion and soil degradation can seriously break the dynamic balance of soil microbial community, and the imbalance of soil microbial community will further aggravate soil degradation [ 41 , 42 , 43 ]. Therefore, the study on the structure and functional diversity of soil microbial community is of great significance to understand the soil quality of artificial C. henryi forest. Meanwhile, it could provide guidance in preventing soil fertility decline and scientific managing of plantation C. henryi forest. At present, the research of C. henryi plantation mainly focused on the development and utilization of germplasm resources, genetic diversity, high-yield cultivation techniques, and occurrence and control of main diseases and pests [ 29 , 44 , 45 ]. However, the research on soil microbial community of C. henryi plantation is less investigated. In this study, metagenomic sequencing technology with microbial molecular ecological networks were combined to (1) characterize the change of composition and function of rhizosphere microbial community assembly at different forest types and cultivated varieties, (2) analyze the soil fertility degradation from the perspective of microorganisms, and (3) elucidate the effect of forest types on microbial molecular ecological networks. Here, we hypothesize that the monoculture and unreasonable management of the plantation may reduce the species and functional diversity of the microbial community and exacerbate competition among microorganisms to nutrients. The results of this research were expected to provide a scientific basis for the sustainable management of the plantation.",
"discussion": "4. Discussion The long-term continuous cropping of plantation leads to the decline of forest productivity [ 32 ] and soil degradation [ 54 ], and the loss of soil microbial diversity is considered as the main threat to the balance of ecosystems [ 55 ]. Different forest planting methods can affect soil microbial communities and the change of microbial communities may affect the function of soil ecosystem [ 3 , 15 ]. In this research, we compared the differences of microbial diversity among rhizosphere soil of C. henryi forest cultivated in natural and plantation forests. Our findings clearly showed that the natural forests have higher diversity and richness at species composition and functional genes and the network relationship is more complex. The soil organic matter of natural forest can be satisfied by the natural decomposition of abundant litter [ 56 ]. Plantations are often characterized by intensive single cultivation [ 32 ] and the return of litter is significantly lower than that of natural forest, whether at a young age or after maturity [ 56 ]. For instance, Yang et al. found that the content of soluble organic nitrogen in 0–10 cm soil decreases by 27.7% when the natural forest of Castanopsis kawakamii converts into three kinds of artificial forests [ 35 ]. Allen et al. found that the soil fertility, microbial biomass, and NH 4+ conversion rate was gradually decreased when the forests were transformed into the two kinds of rubber plantations [ 57 ]. The main nutrient source of plantation soil is fertilization. Short-term and reasonable fertilization can effectively improve soil conditions, while long-term and unreasonable fertilization will imbalance soil nutrient elements and cause soil acidification [ 37 , 58 , 59 ]. According to statistics, in some major areas of China, over-application of chemical fertilizers has caused serious soil acidification and reduced soil pH value by 0.13 to 2.20 [ 59 ]. However, pH is one of the key factors that drives the microbial community construction and it can well predict the microbial diversity and richness, whether on a continental scale [ 60 ] or on different altitude gradients [ 61 ] and even on different plant compartments [ 62 ]. Meanwhile, due to the monoculture in C. henryi plantation and the frequent occurrence of serious diseases and pests, C. henryi plantation mainly depends on spraying pesticides for control. Low concentration and short-term application of chemical pesticides can well control diseases and pests, but the long-term excessive application will cause pesticide residues and environmental pollution [ 38 , 63 ]. This unreasonable operation caused soil hardening, nutrient imbalance, and soil erosion, thus reducing the diversity and abundance of microbial communities. Plant root exudates are also a key factor affecting the accumulation of soil microbial species [ 64 ]. Previous studies have shown that the plants with different genotypes can secrete different types and amounts of root exudates [ 65 , 66 ]. For instance, there are significant differences in specific compounds in the root exudates of the three soybean varieties, with the variety Nice-Mecha exuding about two times more organic acids than variety Bara, and the variety Svapa exuding more sugars than variety Bara [ 67 ]. Hence, we speculated that the differences among different cultivated varieties were very likely to be explained by the difference in root exudates. Our study showed Proteobacteria , Acidobacteria , and Actinobacteria are the dominant flora in the rhizosphere soil of two C. henryi forest types, accounting for >72.9% of all the phyla. The abundance of Proteobacteria and Actinobacteria in rhizosphere soil of natural forest was higher than that of plantation, while the abundance of Acidobacteria in plantation was higher than that of natural forest. By analyzing the microbial communities of litter and soil of forests, Urbanova et al. found that Proteobacteria and Actinobacteria were major phyla, accounting for more than 50 and 13% both in litter and soil [ 26 ], suggesting that they are the main populations for the decomposition and transformation of forest lit-ter. Actinobacteria can decompose the refractory organic matter in the high carbon content soil [ 68 ], affecting the input of soil nutrients and the formation of soil aggregate structure. Meanwhile, Actinobacteria can participate in the decomposition of aromatic compounds and other complex compounds [ 69 , 70 ]. In addition, they are able to produce a wide spectrum of secondary metabolites, such as antibiotics and organic acid [ 71 ], to resist pathogenic microorganisms and harmful environmental microorganisms [ 71 , 72 ]. Proteobacteria is also an important flora for decomposing soil organic matter, fixing nitrogen and carbon, and dissolving phosphorus [ 26 , 73 , 74 , 75 ]. It has been reported that Proteobacteria can also degrade aromatic compounds to reduce the organic pollution [ 38 ]. Our research found that the carbon metabolism pathway, the biosynthesis of antibiotics pathway, and the biosynthesis of secondary metabolites pathway were enriched in the natural forest. Therefore, natural forests may increase the carbon metabolism pathways of microbial communities by enriching Proteobacteria and Actinobacteria to enhance the decomposition of understory litter and release nutrients to improve soil quality. At the same time, Proteobacteria and Actinobacteria also enhance the biosynthesis of secondary metabolites and antibiotics pathway to improve the ability of the host to resist the invasion of pathogens. Acidobacteria mostly exists in acidic environments, and the abundance decreased with pH value [ 76 , 77 ]. Meanwhile, it is reported that some Acidobacteria members can tolerate the environment of nutrient deficiency. Therefore, it can be used as an evaluation index of soil fertility [ 78 ]. Compared with the natural forest, the C. henryi plantation has lower pH and soil nutrients, as well as lower soil moisture (SM). Unreasonable fertilization and pesticide spraying in plantations leads to deterioration of soil quality, such as soil acidification and compaction and nutrient imbalance. Hence, the characteristic of a single planting structure and unreasonable management measures may cause the decline of plantation soil fertility. The degradation of soil may be the main reason for the accumulation of Acidobacteria in the plantation. From our network analyses, we found that higher nodes, links, and average degree in the natural forests network indicate that the network of the natural forest is more complex than that in the plantation [ 79 ]. The interaction patterns in the natural forests network were identified as predominantly positive, meaning that microbial interactions tended to favor symbiosis rather than competition [ 80 , 81 ]. As indicated in previous studies, the positive correlations are dominant among rhizosphere bacteria which are usually rich in nutrients [ 17 , 79 ]. Li et al. found that the percentage of positive links of soil microbial communities in Panicum virgatum increased from 72.0 to 87.9% under three N fertilizer levels (0, 56, 196 kg ha −1 ) [ 17 ]. Hence, rich nutrients may alleviate the competition of microorganisms. However, we found more negative correlations existed in plantations. Studies on the social network structure of animals have shown that the availability of resources is one of the key drivers of network structures [ 82 , 83 ]. Similarly, finite nutrients can invoke a relentless war among diverse microorganisms [ 84 ]. We conjecture that poor plantation soil triggered microorganism competition. In our study, the top 10 nodes which have high connections (that is, degree) were selected as keystone taxa in each network and the removal of keystone taxa will lead to drastic changes in the composition and function of the microbiome [ 85 , 86 ]. We observed that the keystone taxa, whether in natural forest or plantation network, are mostly rare and moderate abundant taxa, consistent with Xiong et al. [ 52 ] and Xue et al. [ 87 ]. However, although the dominant taxa with high relative abundance are very important and can affect ecosystem functioning or a specific process, the taxa with low abundant taxa should not be neglected because they act as an important or potentially important role in maintaining microbial networks [ 88 ]."
} | 3,942 |
38076825 | PMC10705423 | pmc | 5,541 | {
"abstract": "To adjust to sudden shifts in conditions, microbes possess regulated genetic mechanisms that sense environmental challenges and induce the appropriate responses. The initial evolution of microbes in new environments is thought to be driven by regulatory mutations, but it is not clear how this evolution is affected by how quickly conditions change (i.e. dynamics). Here, we perform experimental evolution on continuous cultures of tetracycline resistant E. coli in different dynamical regimens of drug administration. We find that cultures evolved under gradually increasing drug concentrations acquire fine-tuning mutations adapting an alternative efflux pump to tetracycline. However, cultures that are instead periodically exposed to large drug doses evolve transposon insertions resulting in loss of regulation of the main mechanism of tetracycline resistance. A mathematical model shows that sudden drug exposures overwhelm regulated responses, which cannot induce resistance fast enough. These results help explain the frequent loss of regulation of resistance in clinical pathogens.",
"introduction": "Introduction To enable the colonization of ecological niches where conditions change frequently 1 , 2 , microbes are equipped with inducible mechanisms that sense changes in their surroundings and initiate the appropriate cellular programs 3 . In some cases, responses are not time sensitive and mostly tune gene expression to optimal levels, such as when cells respond to shifts in nutrient conditions 4 , 5 . However, when cells are exposed to antibiotics or other harmful compounds, cell survival depends on the quick deployment of its defenses, while gene expression is still possible 6 , 7 . Therefore, even when microbes carry mechanisms to deal with hostile environments, sudden and frequent changes in conditions still threaten cells that are too slow to respond. Dynamic environments, where conditions shift rapidly, pose fundamentally different selective pressures on the evolution of antibiotic responses 8 – 10 . Antibiotic resistance mechanisms originate in the soil, where antibiotic concentrations are typically low 11 , within a selection window that inhibits the growth of sensitive strains while enriching resistant subpopulations carrying beneficial mutations 12 , 13 . In contrast, antibiotics are used in the clinic in high doses, with the intent of wiping out entire microbial populations 14 , 15 . Such extreme environments pose strict bottlenecks, resulting in selective sweeps of any surviving mutants 16 , 17 , however unfit 18 , 19 . However, while much attention has been devoted to the evolution of antibiotic resistance under steady drug concentrations 20 – 22 , we still lack an understanding of how evolution proceeds in dynamic environments, where drug concentrations are high and change quickly. Recent studies suggest that the short-term evolution of microbes in such challenging environments relies heavily on mutations affecting regulatory pathways 23 – 25 . To study the evolution of antibiotic responses, we focus on tetracycline resistance in Escherichia coli , which is mediated by two inducible efflux mechanisms capable of transporting the drug out of the cell - the tet and acr operons 26 , 27 ( Fig. 1A ). While the acr operon transports a wide variety of toxic compounds and is part of the E. coli core genome 28 , 29 , the tet operon is a tetracycline specific module of the E. coli pan-genome that provides the bulk of resistance in strains where it is present 30 , 31 . Both mechanisms are regulated by repressors of the same family – TetR and AcrR, respectively 32 , 33 – that can bind tetracycline and lose affinity for DNA, releasing expression of their respective efflux pumps, tetA and acrAB . Since the acr operon is not optimized for tetracycline efflux, several mutations in acrB have been reported to increase tetracycline resistance 34 , 35 . Additionally, active tetracycline transport via TetA involves an ion exchange that has been shown to disrupt the membrane potential, thereby posing a trade-off between resistance and toxicity in tetA expression 36 . Ultimately, tetracycline resistance depends on the interaction between these two inducible mechanisms, which differ in costs/benefits and drug specificities. Resistance can be increased by acquiring mutations in either one. Here, we use a system of automated continuous cultures to experimentally evolve tetracycline resistant E. coli populations under different regimens of drug administration. We compare evolution under a steady drug environment, where the drug concentration changes gradually, with a dynamic environment where the population is periodically subjected to sudden exposures to high drug concentrations.",
"discussion": "Discussion Performing experimental evolution using carefully controlled continuous cultures, we studied the role of the dynamics of drug delivery in the evolution of antibiotic responses. We found that steady environments where drug concentrations change only gradually led to the refinement of the AcrB efflux pump through point mutations that optimize it for tetracycline efflux. Meanwhile, dynamic environments with sudden exposures to large drug concentrations did not result in such refinement of protein function, but instead led to the abolishment of regulation of the tet operon, the main mechanism of tetracycline resistance. Such mutation was unexpected in a changing environment, since loss of regulation had been previously understood to be a long-term outcome of evolution in constant environments 55 . These experiments show that regulation is not only important for the parsimonious use of cellular programs that are only needed occasionally, but also needs to guarantee a timely activation of cell defenses in hostile environments, while gene expression is still possible. Regardless of environment, evolution proceeded first through regulatory mutations 56 – 58 , with a high prevalence of transposon insertions. Previous work shows that exposure to low levels of tetracycline results in high levels of mobile element activity via the sensing of cellular stress 59 – 61 . Evolution under high drug levels requires quick adaptation, and transposon insertions provide an accessible mechanism for genome remodeling. Transposon insertions also offer the advantage of being reversible 62 . The overwhelming presence of transposon insertions rewiring the regulation of resistance genes, as well as mutations in genes regulating and operating transposable elements, indicates an important role of this mechanism in the short-term evolution of microbes in challenging environments, which is often missed in whole-genome sequencing analysis. Sudden drug exposures are likely to play a role in the evolution of pathogens in clinical environments. Most antibiotic treatments are well designed to reach and maintain high drug concentrations, but short courses such as intravenous or inhalation result in sharp increases of drug bioavailability at the site of infections 63 . Indeed, loss-of-function mutations in transcription repressors of resistance genes are often found in clinical isolates 64 , 65 . Our results suggest that these mutations can serve both the purposes of bypassing the induction of drug responses and of de-repressing resistance mechanisms that are not sufficiently activated by the drug. Constitutive expression of resistance would be even more beneficial for resistance mechanisms that are induced only by downstream effects of the drug and respond much more slowly than the tet operon 66 . This work raises interesting questions about how the dynamics of the environment shapes evolution, such as whether dynamic regimens with lower drug concentrations might ease evolutionary bottlenecks and ultimately select for faster responses, or whether longer periods in the absence of drug might reduce the benefits of constitutive expression of resistance. Dynamic environments could also promote the coexistence of multiple mutants or species that are optimized for different drug levels, shaping the composition and evolution of complex microbial communities 67 ."
} | 2,037 |
27128992 | PMC5148194 | pmc | 5,543 | {
"abstract": "Arbuscular mycorrhizal fungi (AMF) occur in the roots of most plants and are an ecologically important component of the soil microbiome. Richness of AMF taxa is a strong driver of plant diversity and productivity, thus providing a rationale for characterizing AMF diversity in natural ecosystems. Consequently, a large number of molecular studies on AMF community composition are currently underway. Most published studies, at best, only address species or genera-level resolution. However, several experimental studies indicate that variation in plant performance is large among plants colonised by different individuals of one AMF species. Thus, there is a potential disparity between how molecular community ecologists are currently describing AMF diversity and the level of AMF diversity that may actually be ecologically relevant. We propose a strategy to find many polymorphic loci that can define within-species genetic variability within AMF, or at any level of resolution desired within the Glomermycota. We propose that allele diversity at the intraspecific level could then be measured for target AMF groups, or at other levels of resolution, in environmental DNA samples. Combining the use of such markers with experimental studies on AMF diversity would help to elucidate the most important level(s) of AMF diversity in plant communities. Our goal is to encourage ecologists who are trying to explain how mycorrhizal fungal communities are structured to take an approach that could also yield meaningful information that is relevant to the diversity, functioning and productivity of ecosystems.",
"conclusion": "Conclusions Although molecular AMF community ecology has yielded exciting new information about distribution of AMF taxa, we emphasize that researchers interested in the ecological role of AMF diversity should consider results of ecological experiments to help define the relevant level of AMF diversity to study. A second step should be to develop molecular-based approaches that will enable the measurement of such diversity. If not, ecologists have the possibility of generating very large amounts of sequence data with limited ecological relevance. If important levels of AMF variation are below the species level then ecologists need to embrace new approaches, despite the enormous challenges. The approaches that we suggest are certainly not caveat-free, and could probably not easily be undertaken by one research group alone. However, they have the potential to help uncover the role of variation in AMF on the ecology and functioning of communities and ecosystems. For this reason, we think that a concerted effort by several research groups, and a true interchange between experimental ecologists and researchers developing molecular tools, is required to effectively elucidate the role of AMF diversity in natural ecosystems.",
"introduction": "Introduction Arbuscular mycorrhizal fungi (phylum Glomeromycota) are, without question, an ecologically important component of soils, forming symbioses with ~200 000 plant species in all the major biomes ( Davison et al., 2015 ; van der Heijden et al. , 2015 ). Arbuscular mycorrhizal fungi improve plant P acquisition and also other soil nutrients ( Harrison, 1999 ; Hodge et al. , 2001 ). Arbuscular mycorrhizal fungi exert a strong influence on plant community structure, plant diversity and ecosystem function ( van der Heijden et al. , 1998a , 1998b ). Despite their ecological importance, surprisingly little is known about basic AMF ecology, their biogeography and the factors governing the structure of AMF communities. As with most soil microbiota, molecular techniques to measure diversity have allowed ecologists to study AMF communities in a way that was not possible a few years ago. Most of the studies on AMF diversity evaluate variation in DNA sequences derived from amplicons of specific targeted genes or regions of the genome; sometimes referred to as a meta-genomic approach. Despite this worthwhile pursuit, in this Perspectives article, we argue that in order to understand the role of AMF in shaping plant communities, molecular data need to be aligned with relevant levels of AMF diversity. Here we explain why approaches might currently not be aligned and we propose some possible solutions that may be more ecologically relevant for understanding how the diversity of AMF influences plant ecology. In a first step, researchers using molecular approaches to document AMF diversity should consider findings from experimental studies that identify which are the ecologically relevant levels of AMF diversity that influence plant ecology. In a second step, these scientists then need to develop molecular approaches to allow the measurement of those ecologically relevant levels of AMF diversity. In a third step, researchers need to then measure ecologically relevant levels of AMF diversity in nature or in experiments that will help to pinpoint the role of AMF diversity in plant ecology and ecosystem functioning. In order to understand the problem, we first need to consider the ecological rationale for measuring AMF diversity in communities and also understand the way in which molecular techniques are currently used to measure such diversity."
} | 1,305 |
36770256 | PMC9919079 | pmc | 5,544 | {
"abstract": "A three-terminal synaptic transistor enables more accurate controllability over the conductance compared with traditional two-terminal synaptic devices for the synaptic devices in hardware-oriented neuromorphic systems. In this work, we fabricated IGZO-based three-terminal devices comprising HfAlO x and CeO x layers to demonstrate the synaptic operations. The chemical compositions and thicknesses of the devices were verified by transmission electron microscopy and energy dispersive spectroscopy in cooperation. The excitatory post-synaptic current (EPSC), paired-pulse facilitation (PPF), short-term potentiation (STP), and short-term depression (STD) of the synaptic devices were realized for the short-term memory behaviors. The IGZO-based three-terminal synaptic transistor could thus be controlled appropriately by the amplitude, width, and interval time of the pulses for implementing the neuromorphic systems.",
"conclusion": "4. Conclusions In conclusion, a novel three-terminal synaptic transistor have been fabricated and evaluated for application in the neuromorphic systems. The device without the HfAlO x layer performs a high working current, which causes high power consumption. For the devices with thick CeO x (~70 nm), the variation was worse than that with thin CeO x (~30 nm). In addition, the extended channel width (200 μm to 400 μm) and length (100 μm to 400 μm) are analyzed considering the identical area to optimize the device. Experimentally, the CeO x 30 nm device (D30) with HfAlO x layer is appropriate as an efficient synaptic device when the channel width and length are 200 and 100 μm, respectively. In particular, the fluctuation of V th was 0 to 0.1 V when the gate voltage is applied up to 5 V. The EPCS, PPF, STP, and STD have been performed to mimic the biological synapse. The larger the amplitude (7 V), the higher the pulse number (100), or the longer the pulse width (1000 μs), the higher the current. Additionally, as the current increases, the time to return to the steady state is increased, which has been proven through the PPF test. Since D30 has short-term memory characteristics, it causes potentiation or depression under different interval times between pulses. By controlling the pulse amplitude, conductance is updated linearly, making it simple to predict. Moreover, the property is a factor that allows a high recognition rate in the artificial neural network towards the online learning.",
"introduction": "1. Introduction The capability to store and process data is a crucial feature for handling large quantities of data without loss in the current era of big data [ 1 ]. To fulfill such demands, computing systems should be equipped with high-performance transistors. The high-performance transistors are closely related to their scalability and numbers. Assuming that the total number of transistors that can be integrated within a limited area under Moore’s law doubles every 18 to 24 months, the present capacity for transistors is tens of millions [ 2 , 3 , 4 ]. The performances and circuit functionalities can be improved by increasing the scalability and the number of transistors; however, it is difficult to increase them due to the integration limit and the heat generated during operation [ 5 ]. The von Neumann architecture is the most widely used scheme in computing systems and has an additional process of transmitting data to the memory after processor operation [ 6 , 7 , 8 , 9 ]. This could cause bottlenecks in data transmission [ 10 , 11 ]. Therefore, new architectures are being developed and evaluated in recent times instead of the serial von Neumann architecture. Parallel structures of neuromorphic systems are currently emerging to handle large amount of data effectively [ 12 , 13 ]. These neuromorphic systems mimic the neurons and synapses which are connected in parallel in the biological neural networks [ 14 ]. The interconnected neurons process large quantities of data more efficiently than transistors because the neurons operate simultaneously [ 15 , 16 ]. If human information processing capabilities could be carried out with computing systems, energy could be consumed more efficiently for data processing [ 17 , 18 , 19 , 20 ]. The working of the neurons and synapses in a neuromorphic system should be duplicated concerning those of biological systems for efficient implementation. The synapses update the weights by reacting to the signals from other neurons [ 21 ]. This behavior has a large similarity with that of a memory device, but unlike existing memory systems, the results are not simply classified into “0” or “1” state [ 22 , 23 ]. The hardware shows several states, rather than the two on/off states. Next-generation memories such as phase-change memory (PRAM) [ 24 ], ferroelectric random-access memory (FRAM) [ 25 ], and resistive-switching random-access memory (RRAM) can be used in the neuromorphic systems [ 26 , 27 ]. The PRAM stores data using a material that changes into an amorphous or crystalline state by temperature. The amorphous condition is a high-resistance state (HRS), whereas the crystalline condition is a low-resistance state (LRS), which allows data to be written to and erased from a device. The HRS and LRS allow capability of realizing multiple states via varying the operation schemes. The FRAM undergoes state change by the polarization of a ferroelectric material. The RRAM stores data by the resistive-change switching of an insulator using an external electric field, where the switching characteristics are determined by the insulator and electrodes. If a reset voltage is applied, the device is switched to the HRS; conversely, if a set voltage is applied, the device is moved to the LRS. Energy efficiency for switching and the compatibility to CMOS integration are, hence, essential for high-density memory and neuromorphic systems [ 28 , 29 , 30 , 31 , 32 ]. Compared with PRAM and FRAM, the metal–insulator–metal (MIM) structure of the RRAM enables flexible fabrication and fewer material restrictions [ 33 ]. Owing to these advantages, numerous studies have been reported on RRAMs [ 34 , 35 , 36 , 37 , 38 ]. Depending on switching layer and electrode materials, the RRAM shows various switching characteristics, including unipolar [ 39 ], bipolar [ 35 ], threshold [ 40 ], long-term [ 41 ], and short-term [ 38 ], which facilitate the construction of more hardware-oriented artificial neural networks (ANNs). The three-terminal synaptic transistor that is similar to the RRAM structure was developed to enhance the integration and to allow learning and signal transmission simultaneously for advanced ANNs [ 42 , 43 , 44 , 45 ]. The three-terminal synaptic transistor is derived from the MOSFET structure. The interface charge traps in the channel layer are controlled by the gate voltage to achieve multistate operation capability. The two-terminal device does not learn simultaneously while receiving signals. As a limitation of the terminal, there is no distinction between the presynapse and postsynapse [ 46 ]; hence, at least three terminals (gate: presynapse, drain: postsynapse, and source: ground) are required for separating the terminals. When a signal affects the channel by the gate (presynapse), the channel conductance is altered; this activates the device to update and read data by a drain (postsynapse), which imitates the biological system. It is, therefore, possible to perform an inference operation in a learning operation concurrently so that there is less loss than in a two-terminal device [ 47 , 48 ]. An electrical synaptic device conducts similarly to a synapse in the biological neural network. The biological synaptic weight is the conductance from the perspective of the synaptic device. The stimulus (signal) transmitted from the presynapse (top electrode) is accumulated over time, and the increased conductance (fired weight) indicates the excitatory postsynaptic current (EPSC) that performs another signal transmission [ 49 ]. The EPSC is further triggered as an applied electrical stimulus to the gate. In another case, paired-pulse facilitation (PPF) is generated with different interval times between two pulses [ 50 ]. This means that the conductance increases by repeated pulses before recovering to the initial state, such as the process that a spike is observed before a neuron fires and returns to the steady state. Further, if the conductance is enhanced continuously by multiple pulses, it is considered as potentiation, and the opposite as depression [ 51 ]. Through the process of conductance modulation, the device realizes learning and memory operations similar to the flexible interconnects in the neural network. In the present work, a three-terminal device was fabricated with the W gate/drain/source, HfAlO x and CeO x layers as gate oxides, and the IGZO channel. These materials are not only compatible with CMOS circuits but are also widely used in RRAMs. CeO 2 is a naturally synthesized material with a single Ce 4+ ion and two O 2− ions, but cerium oxide deposited by sputtering has an amorphous formation [ 52 ]. The sputtering deposited CeO 2−x has numerous oxygen vacancies (V o ) within the material layer. Many V o are moved by the electrical force to switch the state of the channel [ 53 , 54 , 55 ]. However, an excess of vacancies can eventually cause a leakage current from the gate to the source. This is not appropriate for neuromorphic systems requiring low power consumption. Therefore, a high-κ material is introduced to prevent unnecessary energy consumption [ 56 ]. The notable high-κ materials include HfO 2 (gate dielectric) and Al 2 O 3 (blocking oxide dielectric) [ 57 ]. HfO 2 has a very high dielectric constant but weak endurance against switching. Al 2 O 3 tends to have high endurance despite a relatively low permittivity. HfAlO x formation, in which HfO 2 and Al 2 O 3 are stacked alternately to utilize their respective advantages, shows high endurance and permittivity [ 58 , 59 ]. Accordingly, the movements of many V o are suppressed by the HfAlO x layer to enable a low-power three-terminal synaptic transistor. IGZO is widely used in the thin-film transistors (TFT) as a channel material with a very low off-current that is also related to the leakage current [ 60 , 61 , 62 ]. Therefore, a three-terminal device was attempted in this study as a low-power neuromorphic device by adopting IGZO as the channel layer.",
"discussion": "3. Results and Discussion The image of each layer of the nanodevice, from the source to the bottom electrodes, is shown in Figure 1 b. The TEM image tends to have a higher/lower contrast as the element becomes heavier/lighter [ 63 ]. The W layer, source, and bottom gate exhibit high contrasts because it is the heaviest element among the components (W, Ce, Hf, Al, O, In, Zn, Ga) of the device. The order of element weights is listed as follows: W, Hf, Ce, In, Ga, Zn, Al, and O. It is confirmed from the image that the contrast differs depending on the atomic number. Line scan ( Figure 1 c) and elemental mapping ( Figure 1 d) were performed by an EDS and TEM. The elements are represented by the same colors in both figures. The line scan shows the normalized atomic percent. The IGZO, which is the channel material, has about a 1:1:1 ratio of In, Ga, and Zn, considering the ratio of the sputtering target (In:Ga:Zn = 1:1:1). In the case of HfAlO x , a similar atomic ratio is observed as the proportion of Hf and Al in the ALD system. A non-stoichiometric state CeO x (1.3 < x < 2) film was deposited by a sputtering system. The initial state of the device has a very high resistance, which prevents rapid current flow. Hence, the device must be subjected to a soft breakdown by adjusting the compliance current as a forming process [ 64 ]. Similarly, the three-terminal synaptic transistor also has many defects including interface traps in the CeO x /HfAlO x /IGZO layer. These defects need to be induced that the electrons are trapped to control the conductance of the IGZO channel. The negative gate voltage releases electrons from the IGZO-HfAlO x interface, whereas the positive gate voltage attracts electrons from the IGZO channel. In other words, the negative (positive) bias makes the channel more n-type (p-type) and increases (decreases) the barrier height between the source/drain and IGZO channel. Figure 2 shows the I-V curves of the gate voltage versus drain current. The drain current increases as the conductance of the IGZO channel increases. In Figure 2 b, the gray line indicates the I-V curve when applying a voltage bias from −7 V to 7 V to the gate terminal for the forming process. As shown in Figure S1b , the hysteresis loops from around 3 V are different with and without the forming process. The accumulated charges are a factor that controls the synaptic characteristics of the three-terminal device [ 65 ]. The drain voltage of 0.1 V is used for reading when applying voltages of −5 V to 5 V (green lines), 6 V (orange lines), and 7 V (brown lines) with width/length (μm) of 200/100 to the ~30 nm CeO x device (D30), as shown in Figure 2 b. The applied negative bias (−5 V) causes the device to return to its initial state. As the gate voltage increases in the positive direction, the drain current also increases, but when the gate voltage decreases in the negative direction by the DC sweep, the drain current decreases. The difference in current in the forward and backward directions also depends on the maximum gate voltage. The values of drain current ratio above gate voltage = 4 V are as follows: V g = 5 V , I d , a f t e r / I d , b e f o r e : 7.41 V g = 6 V , I d , a f t e r / I d , b e f o r e : 36.3 V g = 7 V , I d , a f t e r / I d , b e f o r e : 49.5 \nwhere V g is the gate voltage, I d , a f t e r is the drain current after sweeping, and I d , b e f o r e is the drain current before sweeping. As the voltage increases, the drain current increases. This tendency also appears in Figure 2 a,c which are the results applied to the same process as Figure 2 b including the forming switching which is not marked. Here, the gray lines depict the cell-to-cell variations. I d,after /I d,before values in Figure 2 a ( Figure 2 c) are 4.96 (3.65), 67.5 (76.7), and 147 (450) when V g values are 5, 6, and 7 V, respectively. The ~70-nm CeO x device (D70) in Figure S2 , which has the same fabrication process but different deposition times of the CeO x layer, was also measured for cycle-to-cycle and cell-to-cell (CTC) changes. Figure S2b −d shows the I-V curves during cycles and Figure S2e −g shows the I-V curves of CTC. D70 typically demonstrates that the curve shifts as charge accumulates from cycle to cycle. Even in CTC, the grain lines are curves of other cells, but they are not consistent. The D70 has the worst cycle-to-cycle and CTC variations compared to D30. The results indicate that the thick layer, which produced more V o , is accompanied by fluctuations [ 66 , 67 ]. The ~30-nm CeO x device without the HfAlO x layer was also fabricated for obtaining a switching material with higher dielectric constant. As the result, the drain current went significantly higher ( Figure S3 ) than those of D30 and D70 although it was before the forming process. The sum of the current is very large if the devices used in the neuromorphic system are connected in parallel, so high-κ materials might play an essential role for low-power operation. The threshold voltage (V th ) is the turn-on point of the three-terminal device. The V th values of D30 and D70 with 200/100 μm width/length are plotted in Figure 3 . The ranges of V th in D30 are 4.1–4.2 V (V g : 5 V), 4–4.2 V (V g :6 V), and 3.8–4 V (V g :7 V). Similarly, the ranges in D70 are 3.2–3.4 V, 3.1–3.4 V, and 2.6–3 V. The V th fluctuations are relatively larger in D70 than in D30 owing to the larger V o . The D30 (W/L = 200/100 μm) is appropriate for a device with a small fluctuation to mimic synapses. It is an ANN system that is learned by transmitting signals derived from the stimulus to the postsynapse. For learning, the synapse is activated by updating the weights. From the device perspective, the gate acts as the presynapse (the terminal into which a pre-synaptic neuron signal comes), the drain acts as the postsynapse (the terminal through which a post-synaptic neuron signal goes out), and the channel layer acts as the weight reactant. Accordingly, it is proposed that the current of the three-terminal synaptic transistor should be changed by an electrical signal. This current is commonly referred to as EPSC. First, as shown in Figure 4 a−c, an amplitude of 1–7 V (N = 1, with = 100 μs), a pulse number of N = 1–50 (amplitude = 5 V, width = 100 μs), and a pulse width of 10–1000 μs (N = 1, amplitude = 5 V) are applied to the gate terminal. Further, a bias of 0.1 V is continuously applied to the drain terminal to read the device state. The drain current increases as the amplitude, pulse number, and pulse width increase. After the pulse is removed at the gate, the triggered current gradually decreases over time, which proves the D30 device has short-term characteristics. It is also identified that the recovery time varies with the magnitude of the current generated by the pulse. Table 1 shows the decay times of the drain currents immediately after removing the pulse. In Table 1 , I 0 is the triggered drain current directly after applying the pulse, and t decay indicates the time it takes for I 0 to reduce by 100% or more. In Figure 4 d, the EPSC is obtained by applying a pulse width (interval time) of 500 ms (1 ms) and amplitude of 8 V. The resulting current tends to recover to a steady state due to the first pulse, but it accumulates as the next pulse is input. The behaviors of brain show the property of remembering for a long time if the frequency of stimulation is high, while memory duration becomes short if the frequency is low [ 68 ]. The number of frequencies can be modulated by controlling the interval time between pulses. The PPF, which measures dissimilarity between only two pulses with time, is shown in Figure 4 e. The interval time is set from 1 ms to 2000 ms, and the PPF is defined as Equation (1).\n (1) P P F % = 100 × I 2 / I 1 In the D30 with short-term, as the interval time is longer, the difference between the first and second current is small. On the contrary, if the time is short, it has a large change of approximately 107%. Utilizing the property, short-term potentiation (STP) and short-term depression (STD) can be realized. Short-term synaptic plasticity means that synaptic efficiency changes when the period of presynapse is mirrored: STP is the increase of synaptic strength under repeated stimuli, and STD is the decrease under repeated stimuli [ 69 ]. The STP is conducted by input signals that are repeatedly applied at 5 V and 300 ms with 20 ns interval time in Figure 5 b. Due to short-term characteristics, the conductance accumulated in the channel recovers over time. If the same signal is applied before returning to the steady state in time without inducing the potentiation, the reduction speed is delayed. STD takes place when the same signal with intervals of 1 ms as STP is applied to the devices. The recovering memory is delayed by controlling only the interval time of the pulse, which is the same as STP. It is possible to show the depression due to the slowdown time. The stable state is achieved when 0 V (blue) is applied to the gate terminal, and the activated state is observed when the pulse with 5 V for 300 ms (red) is applied. For 10,000 cycles, it displays extremely strong endurance and has an on/off ratio of roughly 50%. By modifying the design of the pulse, the conductance of the channel can be set as predictive. High recognition can be obtained in artificial neural simulations since the high linearity in weight update [ 70 ]. The results, as shown in Figure 5 e, are obtained by applying the increasing/decreasing pulse design. The linearly-changing input scheme in potentiation or depression shows better linearity in weight change than identical-input scheme. Consequently, the D30 can control channel conductance by optimizing the pulse scheme. This is equal to the results in D70 ( Figure S4a,b )."
} | 5,082 |
37292929 | PMC10245885 | pmc | 5,545 | {
"abstract": "While artificial machine learning systems achieve superhuman performance in specific tasks such as language processing, image and video recognition, they do so use extremely large datasets and huge amounts of power. On the other hand, the brain remains superior in several cognitively challenging tasks while operating with the energy of a small lightbulb. We use a biologically constrained spiking neural network model to explore how the neural tissue achieves such high efficiency and assess its learning capacity on discrimination tasks. We found that synaptic turnover, a form of structural plasticity, which is the ability of the brain to form and eliminate synapses continuously, increases both the speed and the performance of our network on all tasks tested. Moreover, it allows accurate learning using a smaller number of examples. Importantly, these improvements are most significant under conditions of resource scarcity, such as when the number of trainable parameters is halved and when the task difficulty is increased. Our findings provide new insights into the mechanisms that underlie efficient learning in the brain and can inspire the development of more efficient and flexible machine learning algorithms.",
"introduction": "Introduction The brain is an incredibly efficient machine as it can process and store vast amounts of information with exceptional speed and accuracy while utilizing the energy of a small light bulb 1 , 2 . Beyond its energetic efficiency, and unlike most artificial learning systems, the biological brain can also learn continuously over a lifetime 3 , 4 . It constantly integrates new information with existing knowledge without catastrophically forgetting what was previously learned 5 , forming a complex web of concepts and ideas. This efficient learning, in terms of both energy and performance, is believed to be facilitated by a process known as neural plasticity 6 – 11 . By forming, eliminating, and fine-tuning connections between neurons 12 , neural plasticity allows the brain to adjust to newly acquired information and affect cognitive functions such as perception 13 , 14 , motor control 15 – 17 , and decision-making 18 – 20 . Neural plasticity can be divided into two broad and overlapping categories, functional and structural plasticity 21 . Functional plasticity typically refers to changes in the strength of individual synapses in response to neural activity 22 , mediated by alterations in the release probability of neurotransmitters 23 , and the number of receptors on the postsynaptic neuron 24 . Functional plasticity is thought to underlie many forms of learning and memory, including long-term potentiation (LTP) and long-term depression (LTD) 25 . Structural plasticity, on the other hand, typically refers to changes in the overall structure of the nervous system 26 , including the growth of new dendrites, filopodia, and axons 27 , the formation of new and the elimination of existing synapses 28 as well as changes in the size/shape of spines 29 . Synaptic turnover, a form of structural plasticity, refers to the dynamic process of continuous formation and dismantling of synapses 30 – 32 and is strongly associated with learning in multiple areas of the brain 21 , 33 – 38 . Studies have shown that synaptic turnover is vital during development, whereby the formation and elimination of synapses are critical for establishing functional neural circuits 39 – 41 . However, synaptic turnover is not restricted to developmental periods 36 , 42 . Several studies have shown that it continues throughout the lifespan 43 – 45 , providing a mechanism to adjust the wiring diagram of neuronal circuits in addition to the strength of synapses. Similarly, synaptic pruning is important for stabilizing and consolidating memories by removing synapses that are no longer needed 46 – 48 . While numerous experimental and modeling studies have investigated the role of neural plasticity in learning and memory, little is known about the role of synaptic turnover, especially when it operates in conjunction with other biological features such as active dendrites and other plasticity forms. Synaptic turnover is especially interesting because of its potential to expand the processing capabilities of sparsely connected networks like those often seen in biological brains. This is achieved by the dynamic formation and dismantling of task-specific microcircuits 49 . Moreover, the ability to rewire through synaptic turnover can dramatically increase the storage capacity in simulated biological networks 50 – 52 and enhance the ability to transmit information efficiently across neurons 53 . Experiments in transgenic animals with higher synaptic turnover rates showed improvements in learning a fear conditioning task 35 and enhanced motor learning ability 54 . Moreover, the higher level of spine turnover is correlated with greater capacity for subsequent song imitation in birds 55 , thus verifying its importance in behaving animals. Computational modeling suggested that the mechanism via which learning was improved resided at the dendritic level, whereby increased turnover led to synapse clustering in active dendrites and sparser memory encoding 56 and may protect memories from subsequent modifications 57 . To investigate whether synaptic turnover underlies the learning efficiency of biological circuits, we expanded a biologically constrained spiking neural network (SNN) model 56 and assessed its learning capacity on various discrimination tasks. We found that the presence of synaptic turnover resulted in higher performance accuracy and faster learning across all scenarios tested. Moreover, these improvements were highest under conditions of resource scarcity, namely when the number of trainable parameters was halved and when the task difficulty was increased. Improvements were due to a more accurate representation of the different image categories by the network weights through their flexible and more efficient utilization. Our results highlight the important role of structural plasticity in optimizing learning in biological circuits and open new avenues for exploring the applicability of such biological plasticity rules in machine learning applications.",
"discussion": "Discussion We investigated the impact of synaptic turnover, a type of structural plasticity, on the learning and classification performance of a bio-realistic SNN model with Ca 2+ -dependent learning rules and few compartment neurons. The SNN model was based on prior work 56 , 58 and was extended to include a teaching signal, different interneuronal to pyramidal cell connectivity, and to receive image-based input signals. The model was applied to custom-based and MNIST binary classification tasks to understand how synaptic turnover may facilitate learning. While we focused on synaptic turnover, the proposed SNN model provides a useful framework for studying the role of various biological features in learning and classification. These include different learning rules, types of inhibition, wiring structure, compartmentalization, etc. We found that the presence of synaptic turnover resulted in higher performance accuracy and faster learning across all classification tasks tested. It also allowed learning with a much smaller number of training examples to achieve the same performance as a network without synaptic turnover. Importantly, these learning improvements were highest under conditions of resource scarcity, namely when the number of trainable parameters was halved, and the task difficulty was increased. When trainable parameters were doubled, the contribution of synaptic turnover in performance accuracy diminished while speed and training sample benefits persisted. These improvements were due to a more accurate representation of the different image categories by the network weights through their flexible and more efficient utilization. Flexibility is enhanced because synapses that do not convey important information (due to random initialization) can turnover until they find an appropriate presynaptic partner, in which case they can effectively contribute to the task at hand. This process can be further facilitated by restricting turnover so that new synapses are formed within the sub-population in which they were eliminated, thus ensuring an even distribution of resources to both class-specific populations. Recent advancements in neural networks incorporating bio-inspired learning rules and dendritic mechanisms have generated enthusiasm in the scientific community as they promise solutions to challenging problems like continual learning and credit assignment in the absence of error backpropagation 79 – 81 . However, the ways in which biological plasticity rules, anatomical and biophysical properties of the neurons cooperate to drive efficient learning remain explored. Towards this goal, we used a SNN model that incorporates various forms of functional and structural plasticity in multicompartmental neuronal models with active dendrites. We selected these because of their established impact on learning and memory processes 82 , 83 and their prevalence across regions and species 84 . The proposed model can thus serve as a basis for a more extensive exploration of the integrative effects of these mechanisms in biological and machine learning. Our study highlights the potential of using bio-inspired learning rules for machine learning applications. While currently, such training rules cannot match the performance of traditional algorithms for practical applications, this approach holds promise for future generations of SNN models. Our bio-inspired, supervised learning method requires a very small amount of labeled data, contrary to classical methods 85 , while achieving very high-performance accuracy. As such, it opens the way for the construction of complex SNN architectures that incorporate biological features like recurrent connections, cross-linked associations, reinforcement, attention, and other types of layers. Numerous studies have explored the development of local learning rules for SNNs 86 – 92 (for recent reviews see 93 , 94 ), as well as the hardware implementation of neuromorphic computing systems with analog weights and spiking architectures 95 , 96 (for a review see 97 ). Therefore, the development of SNN systems like the one proposed here, which are readily implementable on neuromorphic hardware devices, will facilitate the creation of real-time, energy-efficient information processing systems for various applications (e.g., communication, household appliances, industrial production, robotics) 98 . It is crucial to acknowledge the limitations of our modeling approach. First, our model has only been evaluated on binary classification tasks. While increasing the size of our network in terms of neural units and synapses would allow us to tackle more complex tasks, such expansion is beyond the scope of this study, whose goal is to demonstrate the efficiency gains of synaptic turnover. Second, proving a teaching signal during training is not in line with single-area models, as teaching signals in the brain arise from already established networks, as demonstrated in various regions 99 , 100 . Ideally, these networks should be formed through unsupervised learning algorithms. A more biologically plausible alternative could be a self-supervised learning framework, where teaching signals arise from the similarity and dissimilarity of inputs rather than class labels directly 101 . Despite the limitations of our teaching signal approach, our study is still valuable as it allows us to evaluate the contribution of the synaptic turnover mechanisms to efficient learning. The demonstration that topologically constrained synaptic turnover can improve performance and facilitate learning has the potential to inspire the integration of such mechanisms into advanced architectures in the AI field. The most important take-home message of this work is that synaptic turnover improves the efficiency of learning in our SNN model. This is important because brains evolved under evolutionary pressure and have limited resources (neurons, synapses, energy). Thus, coming up with mechanisms that allow accurate learning with the use of limited resources is critical for survival. Given the advent of technology, resource limitations may not apply to current machine learning systems. However, the fact that state-of-the-art ML systems, such as GPT-4, require the energy of a small city to be trained, charts a non-sustainable future for such AI models. The findings of this study can inspire future developments to adopt biological features to improve the efficiency and sustainability of ML systems."
} | 3,183 |
25815781 | PMC4500129 | pmc | 5,547 | {
"abstract": "The increasing ease of producing nucleic acids and proteins to specification offers potential for design and fabrication of artificial synthetic “organisms” with a myriad of possible capabilities. The prospects for these synthetic organisms are significant, with potential applications in diverse fields including synthesis of pharmaceuticals, sources of renewable fuel and environmental cleanup. Until now, artificial cell technology has been largely restricted to the modification and metabolic engineering of living unicellular organisms. This review discusses emerging possibilities for developing synthetic protocell “machines” assembled entirely from individual biological components. We describe a host of recent technological advances that could potentially be harnessed in design and construction of synthetic protocells, some of which have already been utilized toward these ends. More elaborate designs include options for building self-assembling machines by incorporating cellular transport and assembly machinery. We also discuss production in miniature, using microfluidic production lines. While there are still many unknowns in the design, engineering and optimization of protocells, current technologies are now tantalizingly close to the capabilities required to build the first prototype protocells with potential real-world applications.",
"conclusion": "6. Conclusions Building a complex self-assembling nanomachine from a collection of diverse biological components would be a significant scientific achievement, but one that has not yet been realized. Although many of the design principles have been established and there is precedent for some of the important componentry to be successfully reconstituted in vitro , attaining sufficient knowledge to effectively utilize the components in construction of a viable protocell represents a substantial undertaking. The task may be significantly greater than has yet been accomplished in characterizing the proteins in a native environment. Nevertheless, the potential benefits make the effort more than worthwhile. The simplicity, non-autonomous nature and finite lifespan of non-replicative protocells relative to living organisms holds many potential applications for synthetic biology and molecular medicine. Part of a good design remit is to choose a minimal system that will function as required. Protocells with well-defined synthetic applications, such as fuel or pharmaceutical synthesis, could be designed as stripped-down models, incorporating essential biosynthetic and bioenergetic pathways, but eliminating a multitude of other complex processes required for “life”. Another means of simplification is inherent in the separation between the protocell and its construction process. Not everything has to be encapsulated. Offloading some key functions, for example through component production in a microfluidic production line and self assembly via engineered scaffolds, would result in protocells of significantly reduced complexity that will still be capable of carrying out a designated function. Many designs reported to date are tantalizingly close to the complexity required for an elementary functioning protocell. The use of technologies that have been evaluated and implemented in vitro to a prototype stage, in combination with a range of components that have been functionally reconstituted, also in vitro , will serve as the basis of useful protocells where the finite lifetime is determined only by the longevity of the components. There is significant ground to be covered before functional protocells become a reality. As well as basic construction issues, design will focus on maintaining compatibility—whether by separation of processes or tailoring of reagents. Steady progress in the area has still not yet advanced to the point where a fully functional self-replicating cell can be built. Many would consider the major stalling points to be capabilities for self-replication and division. Whilst impressive steps have been made towards this goal, it is likely to be quite some time before a true autonomous artificial cell can be realized, whereas the design and construction of simpler machines built for a vast array of important synthetic or bio-delivery purposes is almost within grasp.",
"introduction": "1. Introduction Advances in recombinant molecular biology in the 1990s heralded new approaches enabling exploration of biological systems in unprecedented detail. The ability to clone, modify and express a vast array of cellular proteins was extended yet further by the capacity to reconstitute recombinant macromolecules into simple artificial systems. The resulting explosion of knowledge has yielded a far deeper understanding of fundamental processes underscoring cellular biology. Science is fast approaching the point where such information can be harnessed and used in the design and generation of “protocells”, rudimentary synthetic organisms inspired by biology. Carefully selected components of metabolic pathways can, in theory, be utilized to produce purpose-built protocells with highly specific functions. The potential applications are significant and varied; for example in environmental cleanup, biofuel production, and medical and pharmaceutical devices for drug production and drug delivery. At present, designer organisms are based on living cells that have undergone modifications to selected metabolic pathways or have had new pathways incorporated. This type of tailoring for desirable characteristics can be thought of as a “top down” approach to design. Amongst the most impressive examples to date are the generation of usable fuels by modification of both yeast and bacteria, including direct synthesis of alkanes from free fatty acids by introduction of the key enzymes fatty acid reductase and aldehyde decarbonylase [ 1 ]. Others include the production of alcohols with examples of the extensive metabolic modification to produce isobutanol [ 2 , 3 , 4 , 5 , 6 ], butanol [ 2 , 7 ], propanol [ 2 ], isopropanol [ 8 ] and 3-methyl-1-butanol [ 9 ] as fuel components. Impressively the production of some of these fuel molecules has even been linked to CO 2 fixation in photosynthetic organisms [ 10 , 11 ]. There are also instances of pharmaceutical precursors being successfully produced in microorganisms by introduction of biosynthetic pathways, lowering the potential cost of drug manufacture. These include precursors of the antimalarial artemisinin produced in E. coli [ 12 ] or a modified yeast strain [ 13 , 14 ], and of the chemotherapeutic taxol in E. coli [ 15 ]. The simplicity of modifying existing organisms in this fashion is not without appeal. “Housekeeping” metabolic processes, including the ability to generate energy currency, are already present within host organisms and their endogenous biosynthetic pathways can be harnessed to produce precursors for synthetic chemistry. Despite these obvious benefits there are some inherent limitations. For instance, altering an organism’s metabolism decreases its overall fitness, as a result of the build up of inhibitory precursors [ 16 , 17 , 18 ] or from the inherent toxicity of the preferred end products to the host cell. For example yeast cells engineered to produce artemisinic acid, a chemical precursor to artemisinin, are characterized by a marked increase in indicators of cellular stress responses [ 19 ]. In such instances “protocells” could offer a superior alternative. Designed instead from the bottom upwards, they comprise elementary systems containing only a minimal complement of components required to execute their desired function(s). The significant design potential of protocells has advantages over top down approaches that could prove particularly important in the design of medical devices, for example, where biocompatibility could be engineered from first principles. Theoretically, the degree of control over protocell design far exceeds that which can be obtained by re-directing living cells, which still holds many unknowns. While optimization can be accomplished to some extent through metabolic engineering of existing organisms, e.g., by evolution and selection [ 20 ] host organism modification for product production is not trivial (discussed in [ 21 ]). By cherry-picking key metabolic pathways or enzyme cascades and omitting non-essential components, protocell performance can be optimized and many of the problems associated with modifying existing organisms can be avoided. In simpler protocells, protective cellular pathways can be eliminated and resistance to product toxicity engineered—for instance by exclusion of target molecules. The formation of unwanted byproducts can be minimized by avoidance of competing pathways. In the absence of the cellular components responsible for degrading a product of interest, yields can be improved and efficiency maximized. This level of control comes at a cost. To date the development of true protocells has been hindered by the significant technical challenges involved, largely due to the difficulty in combining individual components of diverse cellular origins into a self-maintaining and functionally competent system. The high inherent complexity of living cells, with multiple interconnected pathways and complex regulatory processes, is a consequence of their evolution over millennia by a process of natural selection. A complete genome equips them to replicate, service internal energy needs and exist independently. The very simplicity that advantages protocells could also be their downfall, in that in vitro assembly requires that all essential components be provided. At a minimum, this includes a collection of enzymatic components required for energy generation and product synthesis. Despite these hurdles, many of the technologies required to build protocells are now accessible. Recent studies have combined promising design with incorporation of relatively complex protein machinery into liposomal compartments—the membrane-bound capsules that isolate the working components from external factors just as living cells compartmentalize themselves with lipid membranes. This article provides an overview of protocell technology, moving from what is currently known to suggesting how novel application of existing methodologies could be utilized as groundwork to construct complex artificial systems. The first section discusses building simple prototype protocell machinery from individual components. The second is a future perspective on self-assembling protocells—which represent a significant leap in complexity. One might envisage protocells with a limited “genome” of nucleic acids and the machinery required for maintenance, repair and auto-production of specific proteins—all controlled by systems that sense and respond to external triggers. A third section presents possibilities for automating protocell production. The review finishes with a selection of points for contemplation and consideration. While there are boundless possibilities, the designs considered here follow a classical template of a lipid membrane-bound compartment in which proteins and/or nucleic acids are incorporated. The challenges facing the field of synthetic biology are substantial but not insurmountable and the potential of this exciting and emergent frontier area remains to be realized."
} | 2,841 |
27487207 | PMC4990128 | pmc | 5,548 | {
"abstract": "Summary Bacteria of the SAR11 clade constitute up to one half of all microbial cells in the oxygen-rich surface ocean. DNA sequences from SAR11 are also abundant in oxygen minimum zones (OMZs) where oxygen falls below detection and anaerobic microbes play important roles in converting bioavailable nitrogen to N 2 gas. Evidence for anaerobic metabolism in SAR11 has not yet been observed, and the question of how these bacteria contribute to OMZ biogeochemical cycling is unanswered. Here, we identify the metabolic basis for SAR11 activity in anoxic ocean waters. Genomic analysis of single cells from the world’s largest OMZ revealed diverse and previously uncharacterized SAR11 lineages that peak in abundance at anoxic depths, but are largely undetectable in oxygen-rich ocean regions. OMZ SAR11 contain adaptations to low oxygen, including genes for respiratory nitrate reductases (Nar). SAR11 nar genes were experimentally verified to encode proteins catalyzing the nitrite-producing first step of denitrification and constituted ~40% of all OMZ nar transcripts, with transcription peaking in the zone of maximum nitrate reduction rates. These results redefine the ecological niche of Earth’s most abundant organismal group and suggest an important contribution of SAR11 to nitrite production in OMZs, and thus to pathways of ocean nitrogen loss.",
"conclusion": "Conclusions Collectively, our findings identify diverse and abundant SAR11 lineages whose genome content and environmental distribution reflect adaptation to an anoxic niche, unlike all other SAR11 bacteria characterized to date. The experimentally verified NO 3 - reductase activity in the Gamma-type SAR11 nar variant, along with the high expression levels of divergent SAR11 nar genes in the functionally anoxic core of the OMZ, suggest that persistence in this niche is linked to NO 3 - respiration, consistent with the fundamental importance of this process in OMZs. Nitrate respiration in OMZs constitutes the primary mode for organic carbon mineralization and the main production route of NO 2 - , a critical substrate for the major nitrogen loss processes of anammox and denitrification. The presence and activity of nar operons in SAR11, as well as the high abundance of nar -associated SAR11 clades in the OMZ, implicate these versatile organisms as major contributors to the initiation of OMZ nitrogen loss. Together, these findings redefine the ecological niche of one of the planet's most dominant group of organisms, providing a set of genomic references to establish SAR11 as a model for studies of nitrogen and carbon cycling in OMZs.",
"introduction": "Introduction Alphaproteobacteria of the SAR11 clade form one of the most ecologically dominant organism groups on the planet, representing up to half of the total microbial community in the oxygen-rich surface ocean 1 – 5 . All characterized SAR11 isolates, including the globally ubiquitous Pelagibacter genus, are aerobic heterotrophs adapted for scavenging dissolved organic carbon and nutrients under the oligotrophic conditions of the open ocean 6 – 9 . Gene-based surveys have also revealed diverse SAR11 lineages at high abundance in the deep waters of the meso- and bathypelagic realms 10 – 13 . However, the functional properties that distinguish SAR11 living in distinct ocean regions remain unclear. All known SAR11 genomes are small (typically less than 1.5 Mbp), with genomic streamlining as a potential adaptation to the nutrient limiting conditions of the open ocean. 11 It has been hypothesized that adaptations in SAR11 do not involve large variations in gene content 6 , 8 , suggesting that SAR11’s contribution to ocean biogeochemistry is primarily through its role in aerobic oxidation of organic carbon. Although genetic or biochemical evidence of anaerobic metabolism has not been reported for SAR11, high abundances of SAR11-related genes have been detected under anoxic conditions in marine oxygen minimum zones (OMZs). Permanent OMZs extend over ~8% of the oceanic surface area (O 2 < 20 µM) 14 , with the largest and most intense OMZs in upwelling regions of the Eastern Pacific. In the cores of these regions microbial respiration of high surface primary production combines with low ventilation to deplete oxygen (O 2 ) from mid-water depths, resulting in O 2 concentrations below detection (~10 nM) over a major portion (~100-700 m) of the water column 15 . In the absence of O 2 , respiratory nitrate (NO 3 - ) reduction to nitrite (NO 2 - ) becomes the dominant process for organic matter oxidation 16 , with respiratory NO 3 - reductases (Nar) being among the most abundant and highly expressed enzymes in OMZs 17 – 19 . NO 3 - respiration results in a substantial accumulation of NO 2 - in OMZs, often to micromolar concentrations 20 . This NO 2 - pool is actively cycled through NO 2 - -consuming microbial metabolisms, notably the anaerobic processes of denitrification and anaerobic ammonium oxidation (anammox) 21 , 22 , which together in OMZs account for 30-50% of the loss of bioavailable nitrogen from the ocean as either gaseous dinitrogen (N 2 ) or nitrous oxide (N 2 O) 21 , 22 . Surprisingly, SAR11 bacteria are often the most abundant organisms in the NO 2 - -enriched N-loss zone of OMZs where O 2 is undetectable, representing ~20% (range: 10-40%) of all 16S rRNA genes and protein-coding metagenome sequences in the 0.2 to 1.6 µm biomass fraction 18 , 19 , 23 , 24 . Such high abundances imply that SAR11 make up a substantial fraction of the OMZ community and raise the question of SAR11’s role in OMZ biogeochemistry. Here, we analyzed single amplified genomes (SAG) to identify the metabolic basis for SAR11's dominance in anoxic OMZs. We focused on SAR11 SAGs obtained from the Eastern Tropical North Pacific (ETNP) OMZ off Mexico, the world’s largest OMZ accounting for 41% of global OMZ surface area 14 ( Fig. 1a ). Oxygen concentration ([O 2 ]) at this site declined from ~200 μM at the surface to ~400 nM at the bottom of the oxycline (30-85 m) and was typically at or below the detection limit (~10 nM) from ~90 m to 700 m. At the time of sample collection, NO 3 - reduction rates increased with depth into the OMZ, peaking at ~9.5 nM N d -1 at 300 m 19 , paralleling an increase in the abundance of sequences encoding Nar-type NO 3 - reductases in coupled metagenomes and metatranscriptomes ( Fig. 1c ). In contrast, aerobic NO 2 - oxidation peaked at 100 m (260 nM N d -1 ) where trace O 2 was available and NO 2 - was abundant, before declining 20-fold with depth into the OMZ ( Fig. 1c ). However, NO 2 - oxidation rates are likely overestimated due to slight O 2 contamination in incubations 20 . These data highlight a transition to anoxia within the ETNP OMZ 15 , 19 , with in situ [O 2 ] at least an order of magnitude lower than the inhibitory threshold for NO 3 - reduction, denitrification, and anammox 25 , 26 , consistent with micromolar accumulations of NO 2 - from NO 3 - reduction in this zone."
} | 1,751 |
32581040 | PMC8143649 | pmc | 5,549 | {
"abstract": "Proteomics and metaproteomics are important tools for studying the spatiotemporal heterogeneous ecosystem in our gut. We review strategies and their applications to gut ecology studies, such as building a dynamical model of the MLI.",
"conclusion": "CONCLUSION We are still at the early stage of exploring the ecosystem principles that maintain the homeostasis of our gut microbial community and host-microbiome relationship. Recent development of proteomics and metaproteomics technologies can provide promising contribution to studying spatiotemporal host-microbiome interaction at the MLI. Important approaches facilitating such studies include isolation strategies for different MLI components, enrichment methods to obtain designated array of proteins, probing for specific pathways, isotopic labeling for tracking nutrient flow, and the use of in vitro MLI models. Therefore proteomics and metaproteomics, based on properly selected protocols, can provide information on functional diversity, matter and energy flow, and site-specific insights that are suitable for mathematical modeling of the MLI ecosystem."
} | 274 |
35040946 | PMC9119000 | pmc | 5,552 | {
"abstract": "Abstract The goal of cost-effective production of fuels and chemicals from biomass has been a substantial driver of the development of the field of metabolic engineering. The resulting design principles and procedures provide a guide for the development of cost-effective methods for degradation, and possibly even valorization, of plastic wastes. Here, we highlight these parallels, using the creative work of Lonnie O'Neal (Neal) Ingram in enabling production of fuels and chemicals from lignocellulosic biomass, with a focus on ethanol production as an exemplar process.",
"introduction": "Introduction The invention of vulcanized rubber in 1893 opened the doors for the development of synthetic polymers (Barker, 1940 ). The versatility in polymer types and blends allows diverse applications ranging from household goods to medical equipment and single-use supplies (Andrady & Neal, 2009 ). Currently, most plastic polymers are derived from petrochemicals in processes designed to produce stable, durable materials. This stability to various abiotic and biotic processes results in the accumulation of synthetic plastic polymers in the environment, including microplastics (Sharma & Chatterjee, 2017 ). The global production of plastic continues to increase (Elhacham et al., 2020 ) and much of this plastic waste ends up in landfills and water bodies (Law et al., 2020 ). Various studies have shown that approximately 80 wt% of the debris collected from the ocean floor is plastic (Selvam et al., 2021 ). Thus, there is a need for the development of processes that reuse, recycle or repurpose these materials (Lau et al., 2020 ). Here, we focus on biologically mediated repurposing of these materials (Fig. 1 ). Ongoing efforts to improve chemical and thermal-based processes are described elsewhere (Liu et al., 2021 ; Monsigny et al., 2018 ; Padhan & Sreeram, 2019 ; Qureshi et al., 2020 ; Rahimi & Garcia, 2017 ; Vollmer et al., 2020 ). It has also been demonstrated that plastic monomers traditionally derived from petroleum can be produced by engineered microbes (Karp et al., 2017 ) and plants (Hillmyer, 2017 ; Rasutis et al., 2015 ). Fig. 1. Overview of possible means of plastic degradation and valorization. The environmental accumulation of carbon-rich plastics due to an insufficient biological sink is reminiscent of the carbonaceous period in which lignin production drastically outpaced its biodegradation, and much of this lignin still exists in the form of coal and shale oil deposits (Robinson, 1990 ). This leads to the intriguing premise that petroleum-derived plastics are ultimately derived from lignin and this novel biological functionality that arose 300 million years ago is still wreaking havoc, but this question is beyond the scope of this review. Persistence of lignocellulosic biomass in the environment is no longer a problem and it is actually an appealing source of carbon and energy for the microbial production of fuels and chemicals. The continued progress in the engineering of microbes for utilization and valorization of biomass makes them a tantalizing candidate for addressing the plastic waste problem. The challenges of developing microbial cell factories that can degrade, or even valorize plastic waste in an economically viable process parallels the challenge of developing microbial cell factories for the valorization of lignocellulosic biomass (Fig. 2 ). The topic of microbial degradation, and possibly valorization, of plastic waste has been repeatedly reviewed elsewhere, as described below. Here, we highlight the parallels between plastic utilization and biomass utilization, with a focus on Lonnie O. Ingram's extensive body of work related to biomass valorization. Fig. 2. Summary of the overall process design procedure.",
"discussion": "Discussion Here we have attempted to highlight the similarities between the development of processes for the microbial valorization of lignocellulosic biomass and possible processes for the microbial valorization of plastic waste. This review has focused on the work of Lonnie O'Neal (Neal) Ingram and his colleagues at the University of Florida, but the field of lignocellulosic biomass utilization includes many excellent researchers who have made unique and valuable contributions not described here. This document has mainly described Ingram's work with the production of ethanol from lignocellulosic biomass, but his research group worked with many other biochemical products not described here. Finally, this review has not described Ingram's dedication to commercialization and technology transfer. Neal is often quoted as saying “The development of technology for the cost-effective conversion of modern, renewable biomass into a clean-burning automotive fuel has the potential to free the United States and other nations from oil-dependence and to allow a redistribution of wealth based on productivity and ingenuity rather than natural resources.” The assertion of this work is that the development of technology for the cost-effective degradation and valorization of plastic waste has the potential to free the global community from the scourge of plastic waste and that the productivity and ingenuity of our scientists and engineers are sufficient to accomplish this goal for the benefit of all."
} | 1,321 |
31147580 | PMC6542837 | pmc | 5,555 | {
"abstract": "A biofilm has a unique structure composed of microorganisms, extracellular polymeric substances (EPSs), etc., and it is layered on a substrate in water. In material science, it is important to detect the biofilm formed on a surface to prevent biofouling. EPSs, the major component of the biofilm, mainly consist of polysaccharides, proteins, nucleic acids, and lipids. Because these biomolecules have a variety of hydrophilicities or hydrophobicities, the substrate covered with the biofilm shows different wettability from the initial state. To detect the biofilm formation, this study employed a liquid-squeezing-based wettability assessment method with a simple wettability index: the liquid-squeezed diameter of a smaller value indicates higher wettability. The method is based on the liquid-squeezing behaviour of a liquid that covers sample surfaces when an air-jet is applied. To form the biofilm, polystyrene surfaces were immersed and incubated in a water-circulated bioreactor that had collected microorganisms in ambient air. After the 14-d incubation, good formation of the biofilm on the surfaces was confirmed by staining with crystal violet. Although the contact angles of captive bubbles on the surfaces with the biofilm were unmeasurable, the liquid-squeezing method could distinguish between hydrophilic and hydrophobic initial surfaces with and without biofilm formation using the diameter of the liquid-squeezed area. The surface wettability is expected to be a promising property for in-situ detection of biofilm formation on a macroscopic scale.",
"introduction": "Introduction Commonly, a biofilm is defined as an aggregate of microorganisms in which cells that are frequently embedded within a self-produced matrix of extracellular polymeric substances (EPSs) adhere to each other and/or to a surface. A biofilm is a system that can be adapted internally to environmental conditions by its inhabitants. The self-produced matrix of EPSs, which is also referred to as slime, is a polymeric conglomeration generally composed of extracellular biopolymers in various structural forms 1 . The biofilm formation is a phased process: (1) substances dissolved in water are physically adsorbed onto a surface, and this is referred to as a conditioning film; (2) next, microorganisms adhere onto the conditioning film as a scaffold; (3) the microorganisms secrete EPSs embedding the microorganisms themselves, and this results in a microorganism colony and biofilm; and (4) some of the microorganisms are detached from the biofilm and dispersed into the surrounding environment 2 , 3 . In the biofilm, microorganisms adapt to their surrounding area and easily grow; therefore, the biofilm affects the surrounding environmental and sanitary conditions 4 , 5 . Furthermore, because the biofilm has electrochemical properties, a substrate on which the biofilm has formed (referred to as a biofilm-formed substrate) is more easily corroded as compared with a biofilm-free substrate 6 , 7 . Detection of the biofilm is normally confirmed by the presence of EPSs, in particular, staining of EPSs is a common approach 8 – 10 . This study focused on the EPSs that consist of various types of substances; that is, polysaccharides, proteins, nucleic acids, and lipids, which show a variety of hydrophilicities or hydrophobicities 11 . Through biofilm formation, the surface covered with the biofilm is assumed to be hydrophilic; as a result, the wettability assessment becomes a method for detecting biofilm formation in-situ . Wettability is a surface property that shows affinity of the surface to a contact liquid 12 . This property can be visualized by the behaviour of a liquid droplet on the surface; liquid droplets spread on a high wettability surface and are repelled on a low wettability one. The contact angle is a standard index to quantify the surface wettability in various scientific fields 12 – 16 . As an in-liquid measurement, the captive bubble method has been used for determining the contact angle of a bubble attaching on a cell surface in liquid 12 , 17 – 19 . However, the captive bubble method is unsuitable for some high wettability cases and it is difficult to use in an upright position where the sample surface faces upwards. Recently, a method for assessing surface wettability based on the behaviour of a liquid squeezed by an air-jet has been proposed 20 , 21 . In this method, the diameter of the liquid squeezed area during air-jet application can be used as an index of surface wettability with a high correlation to contact angle 21 . Furthermore, this index is measurable even in situ . Through the wettability assessment by liquid squeezing, this study demonstrated that the surface wettability is a useful property for in-situ detection of biofilm formation on a macroscopic scale.",
"discussion": "Discussion Because of the increasing stain densities on the dishes after the 7-d incubation as compared with those at the initial state, the biofilm was speculated to be well formed on both PS and VGP-PS dishes. Since small objects having rod- or sphere-like shapes and other larger structures were observed by confocal scanning laser microscopy for the 14-d incubated dishes, the steps in the biofilm formation in this bioreactor were inferred to include not only physical adsorption of substances, resulting in the formation of a conditioning film, but also, the growth of microorganisms and the accumulation of EPSs. However, the dispersion stage was unable to be determined in this experiment, although the biofilm detachment occurred on the PS dish at 28 d. In previous studies 22 , 23 , incubation with the bioreactor successfully yielded well-formed biofilms on various types of material surfaces including glass, metals, and plastics other than PS. Therefore, this incubation system was judged as applicable to the biofilm formation on PS and VGP-PS surfaces. Various types of surface modifications affecting hydrophobic and hydrophilic properties have been investigated for reduction of biofilm formation 24 – 28 , and in the present study both hydrophilic VGP-PS and hydrophobic bare PS surfaces had no remarkable effect for preventing biofilm formation. The cause of biofilm formation on both PS and VGP-PS dishes was speculated to be the enhancement of physical absorption such as by a hydrophobic effect in the PS case and hydrogen bond formation to a hydroxy group in the VGP-PS case 29 . Hydrophobic PS surfaces generally cause the hydrophobic effect which enhances the physical absorption of hydrophobic sites of molecules in water onto hydrophobic surfaces by repelling the surrounding water 30 . On the other hand, because the VGP process introduces hydroxy groups onto a surface, a hydrogen bond occurs between the hydroxy groups on the VGP surface and some types or parts of molecules in solution. Therefore, the hydrogen bond on the VGP-PS surface was speculated as assisting the physical absorption of molecules onto that surface 31 . The 14-d incubations in both PS and VGP-PS cases produced the biofilm that was the same as the biofilm formed following the 21-d incubation and relatively stabler in comparison with the early stage biofilm for the 7-d incubation or the biofilm for the longer 28-d incubation. Those PS and VGP-PS surfaces at 14 d were more hydrophilic than the original surfaces, even in the case of the originally hydrophilic VGP-PS, the surface with the biofilm was more hydrophilic than the original one. Because the contact angle of the captive bubble on the PS dish surface was smaller than that on the VGP-PS one, the PS and VGP-PS dishes at the initial state had relatively hydrophobic and hydrophilic properties, respectively. This result agreed with the general understanding of the VGP process, which can change a surface into a more hydrophilic by the addition of the hydroxy group onto the surface. In a previous study regarding liquid-squeezing-based wettability assessment 21 , a clear relationship between the liquid-squeezed diameter and contact angle of the water droplet onto bare and atmospheric pressure nitrogen gas plasma-treated polystyrene surfaces was confirmed. The larger liquid-squeezed diameter during air-jet application and the residual squeezed area on the PS dish indicated this dish had a better hydrophobic property compared with the VGP-PS dish, and the results agreed with the results from the captive bubble method. Although the contact angle of the captive bubble could not be measured due to no attachment of the air bubble for biofilm-formed dishes, the liquid-squeezing approach worked on these dishes. Both dishes after the biofilm formation had smaller mean values of liquid-squeezed diameter during air-jet application and the liquids were fully recovered over the dishes, resulting in no residual squeezed area. Therefore, liquid-squeezing-based wettability assessment was an appropriate method for detecting biofilm formation via its wettability change. In situ Raman spectroscopy 32 and FTIR spectroscopy 33 have revealed the existence of lipids, which are fundamental hydrophobic substances, in biofilms. However, the lipids were speculated to be embedded in a micelle structure in the EPSs or microorganisms because of interaction by surrounding water, therefore, the wettability of surfaces on both PS and VGP-PS dishes expressed hydrophilicity. Especially, because of the non-zero values in the case of biofilm-formed dishes, instead of zero in the residual squeezed diameter, the mean value of liquid-squeezed diameter during air-jet application was a more useful index for quantitative evaluation. To date, various types of in-situ approaches for biofilm characterization have been proposed and practically used (Table 1 ). The present wettability approach has the two important features: (1) the capability for in-situ assessment of an intact hydrous biofilm and (2) the capability for macroscopic representation of physicochemical surface information via the wettability index (liquid-squeezed diameter). Unfortunately, though local detailed information such as a microscopic surface profile cannot be obtained directly, the wettability approach is useful for the simple assessment of surface change in the incubation of anti-biofilm material development with complementary use of a microscopic approach. Table 1 In-situ approaches 11 for biofilm characterization with additional references. Approach Features Reference Optical microscopy Local observation of planar microstructure in an intact biofilm in case of no staining. \n 38 ,\n 39 \n Confocal laser scanning microscopy (CLSM) Volumeric microstructure of biofilm can be observed in situ from laser intensity without staining. \n 40 –\n 44 \n Environmental scanning electron microscopy Observation of local microstructure in near-atmospherical vacuum environment with higher resolution than optical diffraction limit. \n 45 ,\n 46 \n Raman spectroscopy Chemical structures in biofilm can be assessed in situ . \n 22 ,\n 39 ,\n 47 \n Fourier transform infraredspectroscopy (FTIR) The chemical functional groups in biofilm can be analysed macroscopically in situ in case of attenued total reflection technique. \n 33 ,\n 34 ,\n 47 \n Quartz crystal microbalance (QCM) Time-series biofilm formation can be indirectly detected in situ via frequency changes. \n 48 \n Wettability In-situ macroscopic representation of physicochemical information of intact hydrous biofilm. This study"
} | 2,859 |
26152584 | PMC4495169 | pmc | 5,556 | {
"abstract": "ABSTRACT Multicellular biofilm formation and surface motility are bacterial behaviors considered mutually exclusive. However, the basic decision to move over or stay attached to a surface is poorly understood. Here, we discover that in Bacillus subtilis , the key root biofilm-controlling transcription factor Spo0A~P i (phosphorylated Spo0A) governs the flagellum-independent mechanism of social sliding motility. A Spo0A-deficient strain was totally unable to slide and colonize plant roots, evidencing the important role that sliding might play in natural settings. Microarray experiments plus subsequent genetic characterization showed that the machineries of sliding and biofilm formation share the same main components (i.e., surfactin, the hydrophobin BslA, exopolysaccharide, and de novo -formed fatty acids). Sliding proficiency was transduced by the Spo0A-phosphorelay histidine kinases KinB and KinC. We discovered that potassium, a previously known inhibitor of KinC-dependent biofilm formation, is the specific sliding-activating signal through a thus-far-unnoticed cytosolic domain of KinB, which resembles the selectivity filter sequence of potassium channels. The differential expression of the Spo0A~P i reporter abrB gene and the different levels of the constitutively active form of Spo0A, Sad67, in Δspo0A cells grown in optimized media that simultaneously stimulate motile and sessile behaviors uncover the spatiotemporal response of KinB and KinC to potassium and the gradual increase in Spo0A~P i that orchestrates the sequential activation of sliding, followed by sessile biofilm formation and finally sporulation in the same population. Overall, these results provide insights into how multicellular behaviors formerly believed to be antagonistic are coordinately activated in benefit of the bacterium and its interaction with the host.",
"conclusion": "Conclusions. In toto , we demonstrated how the model organism B. subtilis can cope with the basic, although fundamental, decision that it must take when it is attached and committed to a surface: to move or remain in place. We present a novel mechanistic model, containing a Spo0A command that explains the coordinated expression of the different behaviors of B. subtilis . In this model, the spatiotemporal regulation of KinB and KinC by potassium determines the Spo0A~P i amount, which in turn orchestrates the onset and sequential progression of sliding motility, biofilm formation, and finally sporulation/fruiting body formation ( Fig. 8 ; see also Fig. S7 in the supplemental material). Recent publications demonstrate the necessity of biofilm formation for the efficient colonization of the root surface and biocontrol properties of B. subtilis ( 33 , 34 ). The common signal (potassium) and the common regulatory network under Spo0A control (the phosphorelay) ( 25 ) for sliding (this work), biofilm formation (this work and references 10 and 32 to 34 ), and sporulation-fruiting body formation ( 54 ) and our in vitro experiments ( Fig. 2G ) suggest that sliding might also contribute to the ability of the plant growth-promoting and biocontrol bacterium B. subtilis to reach the root surface and efficiently colonize the rhizosphere. Once again, B. subtilis offers an example of simplicity in how distinct prokaryotic social behaviors previously believed to be antagonistic and independent from each other, i.e., surface motility, biofilm formation, and sporulation, might work together to benefit the bacterium and the host.",
"introduction": "INTRODUCTION How do bacteria move from one location to another in natural niches? Most bacteria are able to swim in aquatic environments powered by rotating flagella, whereas a range of different mechanisms have evolved that facilitate surface spreading ( 1 – 4 ). While swimming is considered to be an individual behavior, cells are able to migrate together and cooperatively during surface translocation ( 1 , 3 , 4 ). Surface movement can depend on the presence of flagella (i.e., swarming), the extension and retraction of type IV pili (i.e., twitching motility), the involvement of focal adhesion complexes (i.e., gliding), or “passive” surface translocation (i.e., sliding). Although the mechanisms of swarming, twitching, and gliding motilities have been extensively studied in most bacteria with appendages, the information about the mechanism of sliding, its regulation, and its importance is sparse. Since its original definition, more than 4 decades ago, the concept of sliding as a passive surface translocation driven by expansive forces in the growing colony has not varied much ( 2 , 3 ). However, sliding motility represents a heavily exploited mechanism that different pathogens of global importance (i.e., Bacillus anthracis , Salmonella enterica , Staphylococcus aureus , Legionella pneumophila , and mycobacteria) use for spreading ( 3 , 5 , 6 ). Bacillus subtilis is a Gram-positive endospore-forming bacterium that has been extensively studied due to its diverse differentiation processes ( 7 – 11 ). Different B. subtilis strains swarm on semisolid agar plates ( 4 , 12 ) or form architecturally complex biofilms with vein-like structures and apical tips (fruiting bodies) that project the formed spores into the air ( 7 , 13 – 15 ). In addition, wild and undomesticated B. subtilis isolates ( 7 , 16 ) have beneficial growth-promoting effects on plants and animals ( 17 , 18 ) as well as probiotic effects in humans ( 19 – 21 ). If biofilm formation and an active surface motility are antagonistic but important attributes of a bacterium, how then is the collective decision to move over or stay attached to a surface taken and controlled? In this work, we use the model organism B. subtilis to investigate the genetic mechanism and regulatory network of sliding for surface colonization and their relationship with another prominent cooperative surface behavior, i.e., biofilm formation.",
"discussion": "RESULTS AND DISCUSSION The master regulator of sporulation and multicellular biofilm formation, Spo0A, controls sliding motility in B. subtilis . If bacteria use sliding in a cooperative manner to move across surfaces without the necessity for flagella or any other appendages, how does it take place and what are the regulatory networks that induce and control it? To solve this puzzle, we used two wild (undomesticated) B. subtilis strains of different genetic lineages, the Marburg-related strain NCIB3610, able to swarm and slide ( 7 , 12 , 22 ), and the human-probiotic natto-related strain RG4365 ( 16 , 21 ), which only slides (see Fig. S1 in the supplemental material). It is known that the global transcription factor SinR is essential for the swimming and swarming motilities in B. subtilis ( 23 ). While, as expected, the inactivation of sinR in the NCIB3610 strain yields a completely defective swarming phenotype ( 24 ), the inactivation of sinR in the RG4365 B. subtilis natto strain yields an almost unaffected sliding-proficient phenotype ( Fig. 1A, B, and G ). If SinR is not required to slide, is there any other transcription factor that contributes to the regulatory network of sliding? In B. subtilis , other multicellular and developmental programs (biofilm formation, fruiting body formation, and sporulation) are governed by the master transcription factor Spo0A of the phosphorelay signaling pathway ( 7 , 25 ). Taking into account the dispensability of SinR activity for sliding proficiency, we wondered if this was also the case for Spo0A in sliding. Although swarming of an NCIB3610-isogenic spo0A- deletion strain was not affected (see below), a spo0A -deletion RG4365-derived strain was completely unable to slide on the agar surface ( Fig. 1C and G ). This result suggested that the master regulator of sporulation and biofilm formation, the protein Spo0A, would also be a key regulator of social sliding. The inactivation of spoIIAC ( sigF ), the distal gene of the tricistronic spoIIA operon, coding for the first compartment-specific sporulation sigma transcription factor (σ F ), did not affect the sliding phenotype of wild-type RG4365 cells ( Fig. 1D and G ) and suggested the independency of sliding from the sporulation program in B. subtilis . It is known that the absence of Spo0A activity results in an increase of the activities of the transcription factors SinR and AbrB ( 26 ) that could be responsible for the absence of sliding ability in the spo0A natto strain. If this were the case, the regulatory proteins SinR and/or AbrB could be an inhibitor of sliding motility. As shown in Fig. 1E to G , the inactivation of abrB but not sinR was able to restore (although only partially) the sliding ability of Spo0A-deficient cells and suggested that AbrB was negatively controlling the expression of at least one gene whose product was necessary for sliding proficiency (see below). To confirm the essential role of Spo0A for the proficiency of social sliding in B. subtilis , we constructed an RG4365-derived strain that harbored, in addition to a deletion of the wild-type copy of spo0A , an isopropyl-β- d -thiogalactopyranoside (IPTG)-inducible form of Spo0A (Sad67) that is active in the absence of phosphorylation (phosphorelay independent) ( 27 – 29 ). As shown in Fig. 1H and I , the supplementation with IPTG restored the sliding ability of the spo0A -deletion but Sad67-carrying strain and confirmed the essential role of Spo0A for social sliding proficiency in B. subtilis . FIG 1 Revealing the genetic regulation of sliding motility in B. subtilis . (A to G) Sliding phenotype (A to F) and kinetic characterization (G) of different B. subtilis natto strains (see Table S1 in the supplemental material) affected in the expression of key regulators of gene expression. B. subtilis cells were cultured and inoculated on LB-0.7% agar plates as indicated in Materials and Methods. The arrows in panel G indicate the developmental times when the photographs shown in panels A to F were taken. Strain references for the symbols in panel G correspond to the reference colors shown in panels A to F. The horizontal black line at 8.5 cm shows the maximal size of motility related to the size of the agar plate used. Each value is the average from three replicates. (H and I) Active Spo0A (Sad67) triggers sliding motility in the B. subtilis natto strain. In the absence of IPTG supplementation, Spo0A-deficient but Sad67-positive cells are not motile on soft agar plates (H), but in the presence of IPTG, these cells recover full sliding proficiency (I). Solid IPTG, one or two grains, was poured on top of the solidified LB-0.7% agar (at the points indicated in panel I) in order to allow the dissolution of IPTG in the medium and the formation of a continuous gradient of the inducer. How widespread is sliding motility and how conserved is the role of Spo0A in different B. subtilis isolates? Although flagellum production is essential for swarming motility in the Marburg-related strain NCIB3610 ( 12 ), it has been reported that NCIB3610-derived hag strains (unable to make flagella) are able to slide on solid surfaces after longer periods of incubation (24 h or more) ( 30 ). Therefore, we wanted to know if the essential role of Spo0A for sliding proficiency in the B. subtilis natto strain is also manifest in Marburg-derived cells. To this end, we analyzed the surface translocation ability of different NCIB3610-derived strains under two experimental conditions: incubation on soft Luria-Bertani (LB) medium as shown in previous work and the conditions previously used by other groups, i.e., soft minimal salts glycerol glutamate (MSgg) agar medium, to investigate the sliding ability of Marburg cells ( 31 ). As shown in Fig. 2 , the inactivation of spo0A in NCIB3610 cells did not affect their motility behavior (compare Fig. 2A and B ). As expected, the interruption of flagellin synthesis in the NCIB3610 hag -derived strain RG4384 (see Table S1 in the supplemental material) blocked the surface translocation in both soft media as monitored at a developmental time of 20 h ( Fig. 2C ). However, after longer incubation, the NCIB3610 hag cells moved on the agar surface by a sliding mechanism as previously reported ( Fig. 2D ) ( 31 ). Remarkably, the inactivation of spo0A in the Marburg-derived hag strain ( Δhag Δspo0A double mutant strain RG4385 [see Table S1 ]) completely abolished the ability of these flagellum-less and Spo0A-deficient NCIB3610-derived cells to translocate on the agar surface and confirmed the key role of Spo0A as the master regulator of social sliding motility in B. subtilis ( Fig. 2E and F ). FIG 2 The key regulator of multicellular behavior, Spo0A, controls sliding motility in B. subtilis . (A and B) Spo0A activity is fully dispensable for the swarming proficiency of NCIB3610 cells. (C and D) Inactivation of flagellar synthesis ( hag mutation) impairs swarming motility in NCIB3610 cells (C), but after a longer incubation (24 h or more), sliding proficiency is turned on in Hag-deficient cells (D). (E) Spo0A activity is essential for surface translocation ability of Hag-deficient cells. Photos shown in panels A to C and panels D and E correspond to the sliding migration of the indicated strains after 15 h and 40 h of incubation, respectively. (F) Kinetics of swarming and sliding motilities in Spo0A- and Hag-positive or -deficient NCIB3610 cells. (G) Important role of Spo0A and sliding proficiency for plant root colonization. As indicated in the supplemental material, as soon as sanitized wheat seeds germinated on LB-diluted agar plates, 3.0 µl of stationary-phase cultures of wild-type and spo0A mutant cells was inoculated at the points indicated by the white dotted circles in the top panel. After 24 h of incubation, the wild-type (WT) and spo0A cells formed rounded colonies of similar size and appearance that were confined to the point of inoculation (data not shown). After 3 days of bacterial inoculation, wild-type cells, but not the spo0A cells, efficiently slid on the agar surface (the boundaries of the sliding disc are denoted by the orange dashed circle in the top panel). After 5 days (bottom panel), the wild-type cells were able to colonize the root rhizosphere (blue dashed circle) while the spo0A cells remained immobilized. Representative images of several independent experiments are shown. What might be the importance of bacterial sliding in nature? B. subtilis is a beneficial bacterium that improves plant and animal growth ( 17 , 18 ) as well as possessing advantageous probiotic properties in humans ( 19 , 21 ). One desired attribute of a host-colonizing bacterium is the ability to spread over and colonize a particular niche (i.e., the rhizosphere) and establish a long-lasting community (i.e., a biofilm) associated with the host ( 19 , 32 ). Notably, as shown in Fig. 2G , Spo0A plays a key role in the ability of the plant growth-promoting rhizobacterium B. subtilis ( 33 , 34 ) to activate social sliding and colonize the root rhizosphere. Microarray analysis of B. subtilis cells under sliding conditions. Our initial analysis uncovered the novel roles of the transcription factors AbrB and Spo0A as a repressor and an activator of sliding, respectively. What other genes are important for sliding proficiency in B. subtilis and what is the role of AbrB and Spo0A in their expression? To answer these questions, we performed microarray experiments under different environmental conditions and in various genetic backgrounds. On the one hand, we compared the global gene expression of the spo0A mutant RG4370 with that of the wild-type natto strain RG4365 on LB plates with a 0.7% agar concentration, where, as was shown, the wild-type and the spo0A mutant strains were proficient and impaired in sliding, respectively (see Table S2A in the supplemental material). In a second type of experiment, we examined wild-type cells under sliding-restrictive conditions using LB plates with a 1.5% agar concentration and compared their transcriptome with the pattern of gene expression under sliding-permissive conditions on 0.7% agar plates (see Table S2B ). These microarray analyses showed that 310 and 295 genes were significantly ( P value, <10 −4 ) up- or downregulated in the spo0A mutant strain compared to the wild-type strain, respectively, while 72 and 100 genes were found to be up- and downregulated, respectively, at an increased agar concentration (see Table S2 ). Interestingly, most of the genes belonging to the σ D regulon, which is related to flagellum motility and chemotaxis, were activated in the spo0A mutant strain under sliding-permissive conditions. However, the B. subtilis natto strain lacks flagella under this and all tested genetic backgrounds (see Fig. S1E to G in the supplemental material and data not shown). It has been suggested that Marburg-related wild-type cells (i.e., NCIB3610) lack flagellum production for translocation on solid surfaces depending on extracellular surfactin and potassium ion ( 22 ). More recently, it was shown that synthesis of the exopolysaccharide (EPS) of the extracellular matrix is genetically coupled to the inhibition of flagellum-mediated motility ( 23 ), and as we show below, EPS expression is increased under sliding-permissive conditions. While mutation in spo0A resulted in differential expression of various genes under sliding-permissive conditions on LB medium, we did not find any sporulation-related gene in the wild-type strain to be differentially expressed under this experimental condition (sliding turned on [see Table S2A in the supplemental material]). One simple explanation for this observation is that under the condition used (i.e., rich LB medium and sliding-permissive conditions), sporulation is not activated in the wild-type strain and, therefore, mutation in spo0A has no effect on these genes in the wild-type strain during sliding. In contrast, we observed elevated expression of sporulation σ G -dependent genes in wild-type cells grown under non-sliding-permissive conditions (higher agar concentration [see Table S2B in the supplemental material]). This induction of sporulation genes in samples from 1.5% agar plates is probably due to the fact that under this sliding-restrictive condition, B. subtilis cannot spread, nutrients around cells become limited, and sporulation is started similarly to the conditions during formation of complex colony biofilms ( 11 , 35 , 36 ). Which other genes are expressed during active sliding? We found that in wild-type cells under sliding-permissive conditions (see Table S2B in the supplemental material) and in comparison to the spo0A mutant strain (see Table S2A ), genes related to biofilm matrix production ( sipW , tasA , and eps in the case of the spo0A mutant), biofilm surface layer ( bslA ), fatty acid synthesis ( fab ), and surfactin synthesis ( srfAC in the case of the spo0A mutant) were upregulated. Sliding but not swarming depends on the bslA and eps genes. The microarray experiments presented above showed that genes related to biofilm formation and biofilm surface layer are upregulated under the conditions when sliding is feasible and suggest that this gene repertoire could include novel and necessary components of the sliding machinery in B. subtilis ( Fig. 3A ). Therefore, mutations in bslA , epsG , or tasA genes were introduced into the wild-type B. subtilis natto strain RG4365. Mutations in bslA or epsG abolished sliding of the B. subtilis natto strain ( Fig. 3B ). On the other hand, mutation of the tasA gene did not alter the sliding properties of the B. subtilis natto strain. Are EPS production and BslA synthesis required for sliding proficiency? We examined the effect of the mutations of bslA and epsG in the B. subtilis NCIB3610 wild-type strain (proficient in swarming and sliding) and its hag derivative (proficient only in sliding). While swarming of bslA and epsG mutants in the NCIB3610 background was not altered ( Fig. 3C ), sliding properties of hag bslA and hag epsG double mutant strains were decreased similarly to the bslA and epsG single mutants of the B. subtilis natto strain ( Fig. 3D ). These experiments show that both the BslA protein and the EPS, which are essential components of the biofilm matrix in B. subtilis , are indispensable for sliding. The microarray analysis showed that abrB was downregulated under sliding-permissive conditions (see Table S2A in the supplemental material), which is in agreement with our experimental results that showed the partial restoration of sliding proficiency of the spo0A mutant strain when abrB was also deleted ( Fig. 1E ). Accordingly, AbrB is a repressor of bslA ( 37 , 38 ), a gene that here was shown as required for full sliding proficiency ( Fig. 3B ). FIG 3 Sliding and swarming ability of mutant B. subtilis strains. (A) The cartoon highlights the B. subtilis genes with possible novel roles in sliding motility as suggested by the microarray experiments performed (see text for details). (B) Sliding properties of B. subtilis natto strain derivatives; from top to bottom, wild-type, tasA , epsG , and bslA strains. (C) Swarming of B. subtilis NCIB3610 and its derivatives comparable to those in panel B. (D) Sliding of B. subtilis NCIB3610 hag strain and double mutant hag tasA , hag epsG , and hag bslA strains. (E) Transcomplementation of sliding-deficient strains. Neither bslA nor eps mutant cells are able to slide separately, but when they are poured together, sliding proficiency is restored. In line with the experiments on the B. subtilis natto strain, swarming and sliding were not decreased after a mutation was introduced into the tasA gene of NCIB3610 or into hag strains, respectively ( Fig. 3C and D ), suggesting that the other essential component of the biofilm matrix, TasA, has no role in the ability of B. subtilis to slide. It is interesting that based on the microarray analysis, several genes encoding antimicrobial metabolites (i.e., bacilysin, bacillibactin, and plipastatin) were increased under sliding-permissive conditions (see Table S2 in the supplemental material). In this scenario, the known antimicrobial activity of TasA ( 39 ) might suggest a role of this protein in protection of sliding cells against predators instead of a crucial role of this protein in motility. Understanding the role of the cellular components of the sliding machinery in B. subtilis. The array experiments showed that the operon related to surfactin production ( srf ) was induced under sliding-permissive conditions (see Table S2 in the supplemental material), a result that is in agreement with the essential role of this surfactant in sliding proficiency (see Fig. S2A ). Surfactin is a secreted lipopeptide molecule that in addition to its function in cell-cell communication, as a paracrine signal during multicellularity ( 10 , 40 ), has been proposed to allow the spreading of multicellular colonies through the production of surfactant waves that decrease the surface tension of the surrounding space ( 41 ). In addition, the EPS overproduced during active sliding is also a component of the extracellular matrix that has been identified as a major force driving biofilm growth, due to the osmotic stress generated by its secretion in the extracellular space ( 42 ). Therefore, it is feasible that during sliding the secreted surfactin and EPS, by producing waves of surfactant and gradients of osmotic pressure in the intercellular space of the motile community, respectively, constitute two major forces driving the cooperative sliding of the cells sitting on the surface. In addition, the overexpression of the KinA inhibitor SivA (differential expression of sivA is indicated in Table S2A in the supplemental material) ( Fig. 3A ) ( 43 ) would ensure that sporulation is not triggered during active sliding and suggests that KinA is not the histidine kinase involved in the activation of Spo0A for sliding proficiency (see below). What would be the role of BslA, the other major molecule overproduced during active sliding? It has been shown that BslA is a hydrophobin-like protein secreted to the extracellular space, where it forms surface layers at both the agar-cell and air-cell interfaces around the biofilm ( 44 , 45 ). Because of its physiochemical properties, BslA behaves as an elastic and highly hydrophobic layer coating the biofilm ( 44 – 46 ). Here, we also confirm that BslA is a major contributor to the water repellence of sliding cells (see Fig. S2B and C in the supplemental material). Moreover, the deficiency in BslA synthesis allows the aqueous solution to pass through the cells immediately (see Fig. S2C ). These results suggest that this hydrophobin-like protein could play a role during active sliding as a protector of sliding cells against surface wetting, as was proposed previously for biofilms ( 45 , 47 ). In this scenario, we hypothesize that the sliding-deficient phenotype of eps and bslA cells could be circumvented when the two types of cells are present together under sliding-permissive conditions. Supporting this hypothesis, a previous study on B. subtilis biofilm formation also suggests that these components can be shared among strains producing one but not the other component ( 48 ). As shown in Fig. 3E , mixing epsG and bslA derivatives of the B. subtilis natto strain restores the sliding ability. Interestingly, during the time that our work was under review, van Gestel et al. showed that sliding of B. subtilis 3610 depends on the division of labor between matrix (EPS) and surfactin producer subpopulations ( 49 ). Analyzing a specific set of mutants, they could also show the complete deficiency in sliding of srfA and eps mutants. In contrast to our observation of the dispensability of TasA activity for sliding in B. subtilis natto and Hag-deficient NCIB3610 cells poured on LB soft medium, sliding of wild-type 3610 on MSggN agar medium was only partly impaired in colony expansion ( 49 ). This partially different observation of the TasA requirement for sliding may be due to the different media and growth conditions used for the experiments. While the genes involved in fatty acid (FA) synthesis (i.e., fabF , fabHBA , fabG , etc.) were overexpressed in wild-type cells under sliding-permissive conditions, those genes involved in FA degradation (i.e., fadR , fadA , fadE , etc.) were downregulated at the same time (see Table S2 in the supplemental material). Is an active lipid synthesis, and therefore active membrane formation and remodeling, necessary to slide? In order to confirm the in silico results and test the formulated hypothesis, we proceeded to specifically block de novo FA synthesis in B. subtilis cells grown under swarming- and sliding-supportive conditions. To this end, we treated B. subtilis cells with the antibiotic cerulenin, which is a specific inhibitor of the FabF condensing enzyme ( 14 ), at sub-MICs (below 2 µg ⋅ ml −1 ), which do not affect the vegetative growth of the NCIB3610 and RG4365 strains ( 50 ) (see Fig. S3A and B ). Our results show that sub-MICs of cerulenin produce a dose-dependent impediment of sliding motility as well as swarming in B. subtilis ( Fig. 4A ; also see Fig. S3C and D ). These results confirm the microarray data and suggest that an active de novo FA synthesis constitutes an overlooked requirement for surface (sliding and swarming) motility. FIG 4 De novo branched fatty acid synthesis is required for swarming and sliding proficiencies in B. subtilis . (A) Dose-dependent inhibitory effect of sub-MICs of cerulenin on swarming and sliding proficiencies of NCIB3610-related wild-type and bslA and hag mutant strains, respectively. Sliding and swarming experiments were performed as indicated in the legend to Fig. 1 but with the inclusion of the indicated cerulenin concentration in the soft agar plates. (B) Exogenous branched FAs but not linear FAs (palmitic [nC 16:0 ] and oleic [nC 18:1 ] acids) restore the sliding proficiency of the B. subtilis natto strain in the absence of de novo FA synthesis. Surprisingly, the supplementation with exogenous FAs (nC 16:0 and nC 18:1 , palmitic and oleic acids, respectively) of LB soft agar plates containing cerulenin (2 µg ⋅ ml −1 ) did not bypass the inhibition of surface motility in B. subtilis but allowed the resumption of the planktonic growth of a similar cerulenin-treated culture incubated under shaking conditions (see Fig. S4A in the supplemental material). Why did the addition of nC 16:0 and nC 18:1 FAs not suppress the negative effect of cerulenin on sliding but allow the resumption of planktonic growth? To answer this question, we analyzed the FA profile in samples of wild-type cells grown under sliding-permissive conditions and liquid shaking culture (see Fig. S4B ). B. subtilis , unlike Escherichia coli , synthesizes linear and branched (iso- and anteiso-) saturated fatty acids at 37°C to maintain an adequate membrane fluidity. We found that under active sliding there is a predominance of the synthesis of saturated FAs with lower melting points (anteiso-C 15:0 and anteiso-C 17:0 ) and a decrease in the synthesis of the FAs with higher melting points (linear nC 15:0 and nC 16:0 ). Overall, during active sliding, the percentage of linear FAs drops from 25.0 to 4.0% while the content of anteiso-FAs rises from 26.0 to 54.0%. Simultaneously, the global content of iso-FAs remains around 50.0%, independently of the growth conditions (see Fig. S4B ). The notable increase in the synthesis of low-melting-point anteiso-FAs and the simultaneous decrease in the synthesis of linear FAs would allow the synthesis of membrane lipids with lower melting points and therefore the biogenesis of membranes with higher fluidity. We hypothesize that the synthesis of cellular membranes with a higher fluidity might facilitate the group translocation of B. subtilis cells on solid surfaces in the absence of the propelling force of the flagella (see Fig. S1B ). In this scenario, the increase in membrane fluidity that might be required to slide could not be reached with the supply of linear FAs to cerulenin-treated cultures under sliding-permissive conditions. To test this idea, we supplemented B. subtilis cells incubated under sliding-permissive conditions in the presence of cerulenin with branched FAs. As predicted by the hypothesis, the sliding proficiency was fully restored when branched FAs were added as a supplement to the cells with an interrupted de novo FA synthesis ( Fig. 4B ). The phosphorelay signaling system coordinates multiple multicellular behaviors in B. subtilis. Until now, two multicellular behaviors of B. subtilis have been known to be under the control of the phosphorelay signaling system: fruiting body formation (including spore formation) and biofilm development ( 7 , 11 , 51 ). As demonstrated in this work, sliding motility is a type of cooperative behavior under the novel control of Spo0A phosphorylated by inorganic phosphate (Spo0A~P i ) and therefore of the phosphorelay. Which phosphorelay histidine kinase governs sliding motility? To solve this question, we constructed RG4365-isogenic phosphorelay- kin mutant strains to analyze their sliding behavior. As shown in Fig. 5A , the wild-type and the kinA , kinD , and kinE single mutant strains show similar and proficient patterns of sliding motility. On the other hand, mainly the kinB and, to a much lesser extent, the kinC single mutant strains display a partial impairment in sliding. To confirm that KinB and KinC are sensor kinases involved in the control of sliding, we constructed a kinB kinC double mutant strain and compared its sliding phenotype with those of the spo0A mutant and the other phosphorelay-defective control strain, the spo0F single mutant, which are completely impaired in surface translocation. As shown in Fig. 5B , the kinB kinC double mutant strain displayed a complete impairment in sliding proficiency that was equivalent in magnitude to the sliding deficiency of the spo0A and spo0F mutant strains. In Fig. 5C , the sliding kinetics of the different phosphorelay mutant strains confirm that KinB and KinC are the two sensor kinases that govern sliding motility in B. subtilis and that KinB activity is more significant than KinC activity for the proficiency in that behavior. As expected, the transcomplementation (into the nonessential amyE locus) of the kinB and kinC mutant strains with a wild-type copy of kinB and kinC , respectively, restored full sliding proficiency (see Fig. S5 in the supplemental material). FIG 5 The phosphorelay sensor kinases KinB and KinC govern social sliding motility in B. subtilis . (A) Sliding phenotype of single kin mutant B. subtilis natto strains (see Table S1 in the supplemental material) after 40 h of incubation on soft LB agar plates at 37°C. (B) Complete sliding-deficient phenotype of spo0A , spo0F , and double kinB kinC mutant strains of the B. subtilis natto strain under conditions of incubation similar to those indicated for panel A. (C) Kinetics of sliding motility of different phosphorelay mutants over time. Note that the line with pink squares is common to the spo0A , spo0F , and kinB kinC mutant strains. (D) Sporulation, biofilm formation (at atmospheric oxygen level), and sliding motility are Spo0A-dependent developmental programs that B. subtilis preferentially regulates by duos of phosphorelay sensor kinases. Interestingly, the sliding-controlling KinC kinase (this work) has been proposed (along with KinD) to govern the onset of biofilm formation ( 10 , 33 – 35 , 52 ). Further, KinA and KinB have been suggested to alter biofilm development on certain media and at reduced oxygen levels ( 53 ). We confirm (data not shown) that in the RG4365 natto strain, as well as in the NCIB3610 strain ( 33 , 34 ), both sensor kinases, KinC and KinD, govern the onset of biofilm development and extracellular matrix production in response to plant-derived polysaccharides that constitute one of the signals able to induce both kinases ( 33 ). In toto , B. subtilis employs duos of phosphorelay histidine kinases to control different multicellular behaviors. The histidine kinase duos KinA/KinB, KinC/KinD, and KinB/KinC govern the onset of sporulation and fruiting body formation ( 7 , 8 , 28 , 54 , 55 ), biofilm development under atmospheric oxygen pressure ( 10 , 33 – 35 , 51 , 52 ), and social sliding (this work), respectively ( Fig. 5D ). The sliding signal. What is the nature of the signal, acting on KinB and/or KinC, which is responsible for triggering sliding motility in B. subtilis ? To answer this fundamental question, we had two premises. First, the sliding-inducing signal, responsible for the autophosphorylation of KinB and/or KinC, should not be strong enough to trigger KinB~P i -/KinC~P i -dependent activation (phosphorylation) of Spo0A to the high levels of the regulator (Spo0A~P i ) needed for spore formation, because we did not observe induction of sporulation genes under conditions of active sliding (see Table S2 in the supplemental material). Second, the same signal that activates the autophosphorylation of KinB and/or KinC sensor kinases to make Spo0A~P i and induce sliding motility would prevent initial KinB~P i -dependent and/or KinC~P i -dependent biofilm formation. A recent report suggested that KinB is controlled by the respiratory apparatus via its second transmembrane segment ( 53 ). It is proposed that under conditions of reduced electron transport, KinB becomes active (formation of KinB~P i ) via a redox switch involving its second transmembrane segment with one or more cytochromes to induce biofilm formation and sporulation ( 53 ). We envision that under active sliding, in rich soft medium, the physiological conditions of the sliding cells would be different from the conditions of sessile cells forming a biofilm. Under conditions of biofilm formation, a crowded population of cells exists encased in the biofilm matrix with nutrients that become rapidly exhausted ( 11 ). Furthermore, if a reduced electron transport triggers KinB~P i -dependent biofilm formation and sporulation, surface translocation (sliding) would not be activated at the same time since the two are antagonist responses ( 53 ). Therefore, we consider it unlikely that the status of the respiratory apparatus, sensed by the second transmembrane domain of KinB, could be the physiological condition triggering sliding. Interestingly, it has been proved that intracellular potassium represents a negative signal for KinC ( 10 , 52 ). Potassium is a major intracellular ion that impairs KinC activation through interaction with the cytoplasmic PAS-PAC sensor domain of the kinase ( 10 ). The intracellular potassium concentration decreases as B. subtilis cells reach the late logarithmic phase when newly synthesized surfactin, through its membrane pore formation activity, and the putative potassium channel YugO secrete the ion to the outside of the cell ( 10 , 52 ). In this model, the surfactin/YugO-mediated intracellular drop in potassium concentration activates KinC. Curiously, in contrast to the round colonies formed on LB and LBY (LB medium supplemented with 4.0% yeast extract; see also reference 14 ) agar plates by the wild-type RG4365 strain and its isogenic kinC derivative, the kinB mutant strain forms colonies and biofilms with a tendril-shaped morphology that are very similar to the morphology of wild-type B. subtilis colonies grown on CM (casein digest-mannitol medium) plates, a solid medium with potassium deficiency ( 22 ) ( Fig. 6A and B ). Basically, low (micromolar) and high (millimolar) levels of potassium ions favor tendril-like and rounded colony formation, respectively ( 22 ). Due to the similar colony phenotypes of the kinB mutant and the wild-type strain grown on solid medium with low levels of potassium, we were motivated to investigate if potassium is involved in the regulation of KinB. Remarkably, as shown in Fig. 6C , we discovered a dose-dependent positive effect of potassium ions on sliding motility of the wild-type and kinC strains, which are proficient in kinB expression. This sliding stimulation was observed at potassium levels between 50 mM and 100 mM (data not shown) (with an optimal sliding-stimulatory concentration of 75 mM) that are comparable to the potassium concentrations that inhibited KinC from triggering biofilm formation ( 10 , 52 ). In contrast, there was no effect of potassium supplementation on the sliding proficiency of the kinB mutant ( Fig. 6C ). These results strongly suggest that potassium represents a positive signal for KinB activation. A closer examination of the colony and biofilm phenotypes of the kinB mutant strain ( Fig. 6A and B ) seems to indicate that KinB might inhibit KinC from stimulating biofilm formation, and we are currently investigating this phenomenon. FIG 6 Potassium is the physiological signal that regulates sliding motility in B. subtilis . (A and B) Tendril-like morphology of RG4365-isogenic kinB colonies and biofilms (complex colonies) formed on LB (A) or LBY (B) medium. (C) Potassium stimulates sliding of KinB-positive (wild-type and kinC mutant strains) but not KinB-deficient ( kinB strain) cells. (D) Potassium does not represent a signal for sporulation proficiency. Sporulation proficiencies of wild-type (wt) ( kinA + \n kinB + ) and kinA mutant ( kinA kinB + ) strains in the presence and absence of added potassium ions (75 mM) are shown. Viable cells and spores were determined after 30 h of growth in SM as previously described ( 27 ). Results presented in panels C and D are representative of three experiments performed separately. It is known that the sensor histidine kinase KinB is, in addition to KinA, the main sporulation kinase of B. subtilis ( 54 ). Although a kinB mutant strain is proficient in sporulation (Spo + phenotype), a double kinA kinB mutant strain is almost unable to sporulate (Spo0 phenotype) ( 54 ). Therefore, we were interested to investigate if the positive effect of potassium on KinB-dependent sliding proficiency was also valid for spore formation. As shown in Fig. 6D , potassium did not stimulate sporulation in either of the two KinB-proficient strains, i.e., wild-type (KinA + KinB + ) and kinA mutant (KinA − KinB + ), that were analyzed. Consequently, potassium constitutes a specific signal for sliding motility that precisely fulfills the two hypothesized premises to be ineffective in triggering sporulation (first premise, Fig. 6D ) but, at the same time, strong enough to trigger sliding motility (second premise, Fig. 6C ) ( 10 , 52 ). In addition, the dual role of potassium ions (present at high intracellular levels at early times of growth) as activators and inhibitors of KinB and KinC activities, respectively, points to KinB as the phosphorelay kinase responsible for the start of the cooperative sliding movement (see below). Does potassium represent a direct or an indirect signal to activate KinB? Searching for conserved domains and sequence motifs present in KinB that might be involved in the potassium response, we discovered a disregarded sequence (SLKTNGTG) residing on the ATP-binding region of KinB that is absent in the sequences of the other four phosphorelay sensor kinases ( 56 ) ( Fig. 7A ). This sequence possesses a significant homology to the highly conserved K + -filter (selectivity) sequence of the pore loop domain (P-domain) of potassium channels (T/S-x-x-T-x-G-x-G consensus sequence) ( 57 , 58 ). Despite the many protein motifs and domains present in different types of potassium channels ( 58 ), the KinB K + -selectivity-like sequence (here called K* for simplicity) is the only common element related to K + channels. While active KinB is a dimer ( 59 ), typical potassium channels are tetramers made up of predominantly identical subunits clustered to form the ion permeation pathway across the membrane ( 60 – 62 ). In addition to the absence of the pore motifs that surround the selectivity filter ( Fig. 7B ), KinB also lacks the different domains that have been described in different potassium channels ( 58 , 60 , 61 ). Furthermore, the K* resides in the cytosolic region of the kinase, while in all known (eukaryotic and prokaryotic) potassium channels the K + -filter resides in transmembrane domains. While these topological features exclude KinB as a potassium channel, they open the possibility that the kinase might sense the intracellular concentration of the ion throughout its cytosolic K*. Therefore, we tested if the K* plays a role or not in sliding motility. To this end, we constructed two types of kinB mutant strains harboring specific mutations in the K* (see Fig. S6 in the supplemental material). In one case, three out of the four conserved amino acids of the consensus K* were replaced by alanines (the fourth conserved amino acid, G, of the consensus sequence was not altered as it overlaps with a predicted ATPase motif of the kinase [ Fig. 7A ]) to give rise to the mutant KinB K*→A (see Fig. S6 ). In the second constructed kinB mutant strain, 7 out of the 8 amino acids of the K* were deleted (mutant strain KinB ΔK* [see Fig. S6 ]). One important consideration for both mutant strains before analysis of their roles in sliding is that they must be functional (i.e., promote spore formation). In this sense, the sliding-promoting activity of KinB should be separable from its biofilm/sporulation-promoting activities, a scenario that would explain why KinB-dependent sliding proficiency and KinB-dependent biofilm/spore formation are not simultaneously activated (see below). FIG 7 KinB harbors a cytosolic selectivity filter motif responsive to potassium ions that specifically allows sliding proficiency. (A) Amino acid sequence of KinB. The six continuous underlines indicate the six transmembrane domains of KinB. The histidine highlighted in green corresponds to the residue of autophosphorylation; the blue amino acid triplets represent the top and bottom sites of the ATP-binding domain of the kinase. The yellow box highlights the cytosolic sequence in KinB with homology to the potassium selectivity filter sequence present in potassium channels. (B to D) Sporulation and sliding proficiencies are separable KinB functions. KinB mutant strains affected in the integrity of the selective filter sequence were able to restore full sporulation proficiency of a Spo0 kinA kinB double mutant strain (A − B − in panel B) but did not restore KinB-dependent sliding activity in that A − B − background (B), either in kinB (B − ) (C)- or in kinB kinC (B − C − ) (D)-deficient mutant strains. B k + →A and B Δk + indicate KinB proteins with Ala-exchanged and Ala-deletion K + -filter domains, respectively. (E) Mutation of the potassium selectivity sequence in KinB abolished the ability of B. subtilis to slide in response to potassium addition. Sliding and sporulation proficiencies were measured as indicated in Materials and Methods. Results presented in panels B to E are representative of four experiments performed separately after 40 h of incubation. Remarkably, while the complementation of a kinA kinB double mutant strain (originally Spo0 and deficient in sliding) with a wild-type copy of kinB restored sporulation and sliding proficiencies, both types of alterations in KinB (KinB K*→A and KinB ΔK* ) were able to complement full sporulation proficiency but did not restore the KinB-driven sliding proficiency in that genetic ( kinA kinB ) background ( Fig. 7B ). Simultaneously, both constructed KinB mutants failed to restore sliding proficiency of kinB ( Fig. 7C ) and kinB kinC ( Fig. 7D ) mutant strains and were also insensitive to the stimulation of sliding after potassium supplementation ( Fig. 7E ). These results confirm that the K + -selectivity-like sequence present in KinB (K*) plays an essential role in potassium sensing, leading to sliding proficiency that is located on a different part of the KinB protein than the electron transport-sensing transmembrane segment 2 responsible for triggering biofilm and spore formation ( 54 ). Potassium constitutes the signal for the fine-tuned interconnection of social sliding and biofilm development. Potassium represents the most abundant ion in the cytoplasm (~200 mM in E. coli versus 7 mM content in LB medium) ( 58 ). Unlike most other intracellular cations, the high intracellular concentrations of potassium do not interfere significantly with vegetative growth. However, apart from the stimulatory effect of potassium on KinB (this work), it is known that potassium is a strong inhibitor of KinC activity ( 10 , 52 ), but enigmatically, the activity of both histidine kinases (KinB and KinC) is required for full sliding proficiency ( Fig. 5A and B ). The simplest explanation for this apparent paradox is that KinB and KinC should work at different times of sliding development accompanying the drop in the intracellular concentration of potassium that happens during the transition from the log phase to the early stationary phase. We favor a scenario (i.e., in LB soft agar plates) in which, at the onset of sliding and at the edge of an active sliding colony (where the youngest cells would be present), the intracellular potassium concentration would be high because cells have plenty of nutrients and are in log phase. During this time, KinB (which is the first phosphorelay kinase to be expressed and therefore is present early on in the cell) ( 63 ) should be active (due to the potassium stimulus) in driving the synthesis of the sliding machinery while KinC activity would remain low because of its reduced expression at early times of growth and because of the presence of high levels of intracellular potassium ( 10 , 52 ) ( Fig. 8A , left image). As soon as sliding cells approach the late log phase on LB soft agar plates (probably at the inner, older portion of the sliding disc), surfactin and the potassium exporter YugO are expressed on the cellular membrane ( 10 , 52 ). Therefore, the intracellular potassium concentration decreases, and KinB activity is downregulated and replaced by active KinC to maintain the appropriate levels of Spo0A~P i required for the continued expression of genes needed for full sliding proficiency (i.e., bslA , srf , eps , and fab ) and biofilm formation (see below) at the inner part of the sliding disc ( Fig. 8A , right image). Supporting the view of this spatiotemporal regulation of the kinases, when wild-type (KinA- and KinB-positive) B. subtilis cells are loaded on an optimized soft agar medium that in addition to sliding motility also allows biofilm formation (i.e., soft LBY agar) ( 14 ), it is possible to observe the formation of structures typical of a biofilm in the inner part of the sliding disc, while the outer borders remain flat ( Fig. 8B , left image). This result strongly suggests that (under sliding-permissive conditions) KinC and KinB are active at the interior and at the edge of the sliding disc, respectively ( Fig. 8A , right image). In agreement with this result, when kinC (KinB-positive and KinC-deficient) cells were incubated under similar conditions, the sliding disc, as expected, was smaller than the sliding disc formed by wild-type cells, and more importantly, no structure that resembles a biofilm was formed ( Fig. 8B , middle image). Concordantly, a kinB strain (KinC-positive and KinB-deficient) formed typical biofilm wrinkle-like structures at the inside and outside regions of the slowly sliding cells ( Fig. 8B , bottom right images). Overall, the former results ( Fig. 6 to 8 ) suggest that potassium ions and KinB~P i act earlier than KinC~P i to activate the onset of sliding and confirm, once the potassium concentrations start to be different in distinct regions of the sliding disc, the spatiotemporal regulation of KinB and KinC ( Fig. 8A ). In addition, the KinC~P i -dependent biofilm formation observed at the interior (older) part of the sliding disc after incubation in a medium that allows biofilm and sliding (LBY soft agar plates) (left and middle images in Fig. 8B ) suggests that the potassium-mediated activation of KinB to make Spo0A~P i and start sliding motility ( Fig. 6C and 7 ) is not only insufficient to induce KinB~P i -dependent spore formation ( Fig. 6D ) but also insufficient to trigger the onset of KinB~P i -dependent biofilm formation (middle image in Fig. 8B ). FIG 8 Spatiotemporal regulation of the sliding-inducing kinases. (A) At the onset of sliding, as soon as cells were poured on LBY-0.7% agar plates, KinB was the first-acting kinase while KinC remained inactive. This differential activity of KinB and KinC is due to the intracellular potassium input that activates and inhibits each kinase, respectively (left panel). As progression of sliding continues, there is a drop in the intracellular potassium concentration in the cells at the inner part of the sliding disc. Under this physiological condition, KinB and KinC become inactive and active inside the sliding disc, respectively, while KinB remains active at the newest part (border) of the sliding community. (B) Wild-type B. subtilis RG4365 and its isogenic derivatives mutated in kinC or kinB were loaded on soft agar plates of LBY medium and incubated for 20 h at 37°C. Under these conditions of simultaneous stimulation of sliding and biofilm formation proficiencies, the formation of a structured biofilm in the inner part of the wild-type sliding disc is observed. In contrast, in the sliding disc of the kinC strain no biofilm structure is formed, suggesting that the biofilm observed in the inner part of the sliding disc of wild-type cells is a product of the KinC activity. In both cases (wild-type and kinC strains), the borders of the sliding discs are flat and unstructured. In the case of the kinB cells (right panel), KinC activity drives the formation of typical wrinkled structures, representative of a mature biofilm, inside and at the borders of the colony. (C) Two models for the temporal progression of multicellularity in B. subtilis as described in the text. (D) β-Galactosidase production from P abrB -lacZ in B. subtilis cells grown on petri dishes filled with media that favor the expression of different social behaviors under the control of Spo0A~P i : sliding motility (LB-0.7% agar; green line), biofilm formation (LBY-1.5% agar; red line), sporulation (SM-1.5% agar; black line), or none (LB-1.5% agar; blue line) (see text for details). Cells were taken from the petri dishes at the times indicated in the figure and assayed as described in the supplemental material. The results shown are representative of three independent experiments made in duplicate. M.U., Miller units. (E to H) Five microliters of an overnight culture of the natto strain RG4382 (Δ spo0A ::Ery / P spac -spo0A- sad67 Cat) was inoculated on the middle of solidified LBY-0.7% agar medium prepared with different concentrations of IPTG as shown in the figure. After 20 h of incubation at 37°C, photographs were taken and the sporulation frequency (after elution of the cells from the petri dishes) was determined as described in Materials and Methods. The results are representative of seven independent experiments performed in triplicate. How can sliding and the sessile lifestyle of biofilm formation, which are social behaviors believed to be antagonistic to each other, be positively controlled by the same regulatory pathway (Spo0A~P i ), and why are the biofilm formation and sporulation pathways not activated upon increased KinB autophosphorylation during promotion of sliding motility? Fujita and Losick reported that a gradual increase in the levels and activity of Spo0A results in the expression of different sets of genes in B. subtilis ( 28 ). For instance, high levels of Spo0A~P i stimulate sporulation (fruiting body formation) and low levels of Spo0A~P i stimulate biofilm formation ( 64 ). Therefore, we hypothesize that sliding and biofilm formation developments require, as suggested by Chai et al. ( 64 ) for the cases of sporulation and biofilm formation, different levels of Spo0A~P i to become active. Until now, it was clearly demonstrated that the sliding-permissive conditions used in this work (i.e., soft LB agar) and the potassium input that activates KinB (which at the same time inhibits KinC activity) ( 10 ) are adequate to support sliding (Fig. 1 and 6C) but insufficient to trigger sporulation ( Fig. 6D ) and biofilm formation ( Fig. 8B , middle image, KinB + KinC − strain). Therefore, sliding motility in B. subtilis seems to be activated before biofilm formation and sporulation. But how is sliding and not the other developmental pathways triggered when KinB become active? In other words, which behavior needs lower levels of Spo0A~P i to be triggered? We hypothesize two alternative progressions. In one scenario ( Fig. 8C , model I), the surface-committed cells first slide and later on form the biofilm structures at the center. In the other situation ( Fig. 8C , model II), the surface-attached cells first produce a sessile biofilm and cells at the edge of the biofilm engage in sliding later on. In both scenarios, the social behavior that is triggered first (sliding or biofilm formation) is the one that requires the smaller amount of Spo0A~P i ( 28 , 64 ). One approach to monitor the in vivo levels of Spo0A~P i is to measure the expression of abrB , the most sensitive reporter of the Spo0A~P i levels present in the cell ( 27 , 28 ). Spo0A~P i is a strong repressor of abrB , and very low levels of Spo0A~P i (insufficient to trigger biofilm and sporulation) are sufficient to downregulate abrB ( 27 , 28 ). Therefore, to obtain more insight into the levels of Spo0A~P i present during sliding and biofilm, we measured the levels of β-galactosidase (β-Gal) activity driven by the expression of an abrB-lacZ transcriptional fusion under conditions that favor sporulation (growth on sporulation medium [SM]-1.5% agar plates), sliding (growth on LB-0.7% agar plates), biofilm (growth on LBY-1.5% agar plates), or none of the abovementioned behaviors (growth on LB-1.5% agar plates). At different times, cells were removed from the petri dishes and the abrB -driven β-galactosidase activity was measured ( Fig. 8D ). As expected, the abrB expression was the lowest and highest (indicating the highest and smallest amounts of Spo0A~P i , respectively) when the cells were grown on SM- and LB-1.5% agar plates, respectively ( Fig. 8D ). Interestingly, the levels of abrB expression under conditions of active sliding (growth on LB-0.7% agar plates) were significantly higher than the AbrB levels observed under conditions of active biofilm formation (growth on LBY-1.5% agar plates). These results ( Fig. 8D ) suggest that the levels of Spo0A~P i required to trigger sliding motility are lower than the levels of Spo0A~P i needed to trigger biofilm formation, and therefore, the former behavior (sliding motility) would be triggered before biofilm formation when B. subtilis is attached and committed to a sessile differentiation (model I in Fig. 8C ). To confirm this interpretation, the Δ spo0A sad67 strain (see Table S1 in the supplemental material) ( 27 – 29 ), where the synthesized active Spo0A (Sad67) level depends on the supplemental IPTG, was cultivated on LBY-0.7% agar plates supplemented with different amounts of IPTG. In this experiment, we hypothesized that the behavior (biofilm formation, sliding, or fruiting body formation-sporulation) expressed at the lowest IPTG concentration would reflect the multicellular B. subtilis response that requires the smallest amount of active Spo0A (Sad67) to be produced. As expected, in the absence of IPTG addition ( Fig. 8E ), sad67 is not expressed and B. subtilis is unable to display (because Spo0A activity is completely absent in Δ spo0A cells) any of its different multicellular behaviors. As soon as the LBY-0.7% agar medium is supplemented with a small amount of IPTG (0.01 µM), B. subtilis cells start to slide ( Fig. 8F ). When the IPTG concentration is increased to 10 µM (and therefore more active Spo0A is produced in the surface-committed cells), B. subtilis triggers complex colony biofilm formation ( Fig. 8G ). At the largest amount of supplemental IPTG (1,000 µM), the growth of B. subtilis is restricted (because high levels of active Spo0A inhibit vegetative division) ( 27 – 29 ) and B. subtilis directly induces the formation of fruiting bodies filled with spores ( Fig. 8H ). Overall, these results strongly suggest that the increase in the levels of active Spo0A (Spo0A~P i ) ( 28 , 64 ) allows the expression of the different behaviors of B. subtilis following a temporal sequence of social sliding motility, multicellular biofilm formation, and finally fruiting body formation (sporulation) (see Fig. S7 ). Conclusions. In toto , we demonstrated how the model organism B. subtilis can cope with the basic, although fundamental, decision that it must take when it is attached and committed to a surface: to move or remain in place. We present a novel mechanistic model, containing a Spo0A command that explains the coordinated expression of the different behaviors of B. subtilis . In this model, the spatiotemporal regulation of KinB and KinC by potassium determines the Spo0A~P i amount, which in turn orchestrates the onset and sequential progression of sliding motility, biofilm formation, and finally sporulation/fruiting body formation ( Fig. 8 ; see also Fig. S7 in the supplemental material). Recent publications demonstrate the necessity of biofilm formation for the efficient colonization of the root surface and biocontrol properties of B. subtilis ( 33 , 34 ). The common signal (potassium) and the common regulatory network under Spo0A control (the phosphorelay) ( 25 ) for sliding (this work), biofilm formation (this work and references 10 and 32 to 34 ), and sporulation-fruiting body formation ( 54 ) and our in vitro experiments ( Fig. 2G ) suggest that sliding might also contribute to the ability of the plant growth-promoting and biocontrol bacterium B. subtilis to reach the root surface and efficiently colonize the rhizosphere. Once again, B. subtilis offers an example of simplicity in how distinct prokaryotic social behaviors previously believed to be antagonistic and independent from each other, i.e., surface motility, biofilm formation, and sporulation, might work together to benefit the bacterium and the host."
} | 15,293 |
31344298 | PMC7027739 | pmc | 5,558 | {
"abstract": "Abstract Species‐rich plant communities have been shown to be more productive and to exhibit increased long‐term soil organic carbon (SOC) storage. Soil microorganisms are central to the conversion of plant organic matter into SOC, yet the relationship between plant diversity, soil microbial growth, turnover as well as carbon use efficiency (CUE) and SOC accumulation is unknown. As heterotrophic soil microbes are primarily carbon limited, it is important to understand how they respond to increased plant‐derived carbon inputs at higher plant species richness (PSR). We used the long‐term grassland biodiversity experiment in Jena, Germany, to examine how microbial physiology responds to changes in plant diversity and how this affects SOC content. The Jena Experiment considers different numbers of species (1–60), functional groups (1–4) as well as functional identity (small herbs, tall herbs, grasses, and legumes). We found that PSR accelerated microbial growth and turnover and increased microbial biomass and necromass. PSR also accelerated microbial respiration, but this effect was less strong than for microbial growth. In contrast, PSR did not affect microbial CUE or biomass‐specific respiration. Structural equation models revealed that PSR had direct positive effects on root biomass, and thereby on microbial growth and microbial biomass carbon. Finally, PSR increased SOC content via its positive influence on microbial biomass carbon. We suggest that PSR favors faster rates of microbial growth and turnover, likely due to greater plant productivity, resulting in higher amounts of microbial biomass and necromass that translate into the observed increase in SOC. We thus identify the microbial mechanism linking species‐rich plant communities to a carbon cycle process of importance to Earth's climate system.",
"introduction": "1 INTRODUCTION Biodiversity loss through anthropogenic changes in the global environment is threatening ecosystem functions and services. Grassland ecosystems are predicted to experience most biodiversity losses as a consequence of land‐use change, such as the conversion of grasslands into croplands (Sala et al., 2000 ) and recent studies revealed concomitant negative impacts on soil carbon cycling (Chen & Chen, 2019 ; Tang et al., 2019 ). Terrestrial ecosystems store most organic carbon in soils where it has the potential to become stable soil carbon and thus can be sequestered for longer time periods. Globally, terrestrial carbon storage is dominated by forests (39% of the total terrestrial organic carbon stored in forest soils and vegetation), but grasslands also contribute substantially (34% of the total terrestrial carbon) as they cover a large part of the world's landmass, with ~53 × 10 6 km 2 grassland area versus ~29 × 10 6 km 2 forest area (White, Murray, & Rohweder, 2000 ). Soil organic carbon (SOC) represents the largest carbon reservoir in global grasslands, with up to 98% carbon stored belowground (Hungate et al., 1997 ). As such, understanding the mechanisms that sustain grassland SOC storage is of utmost importance for estimating the potential of grasslands to reduce atmospheric carbon dioxide (CO 2 ) concentrations and mitigate feedbacks from the biosphere to the climate system. Plant diversity is increasingly recognized to be central to grassland SOC storage, with observations from biodiversity experiments demonstrating clear links between plant diversity, primary productivity, and ecosystem carbon cycling (Cong et al., 2014 ; De Deyn et al., 2011 ; Fornara & Tilman, 2008 ; Lange et al., 2015 ; Naeem, Thompson, Lawler, Lawton, & Woodfin, 1994 ). Hereafter, we use plant diversity as a term to describe both plant species number and functional composition, and specify when referring specifically to plant species richness (PSR), functional group richness, or functional group identity. Higher aboveground plant productivity as a consequence of increased plant diversity is usually also accompanied by increased belowground plant biomass production, although the latter may occur only after a delay (Cong et al., 2014 ; Fornara & Tilman, 2008 ; Ravenek et al., 2014 ). However, while there is evidence that increasing plant diversity translates into greater aboveground primary productivity (Roscher et al., 2005 ; Spehn et al., 2005 ; Tilman, Wedin, & Knops, 1996 ), few studies have investigated the mechanisms linking plant diversity and plant productivity with SOC dynamics. This is partly due to the paucity of long‐term biodiversity experiments that allow for exploration of typically slow changes in SOC storage. Indeed, we are aware of only four of such grassland biodiversity experiments globally. Studies from these experiments have consistently shown positive effects of plant diversity on SOC storage, and have largely ascribed this to increased rhizosphere carbon inputs (Cong et al., 2014 ; De Deyn et al., 2011 ; Fornara & Tilman, 2008 ; Lange et al., 2015 ; Steinbeiss, Beßler, et al., 2008 ). Yet, how this mechanism is linked to microbial carbon processing has rarely been empirically tested, limiting our ability to implement microbial carbon dynamics in climate‐carbon models and dynamic global vegetation models (Crowther et al., 2016 ). The build‐up of organic carbon ultimately depends on the balance between carbon inputs and outputs from the system, which is determined by plant biomass production, and SOC formation and decomposition, and is therefore, to a high degree, governed by the activity of soil microbes. Most plant‐derived carbon is taken up by soil microbes and used to either generate energy (and thus CO 2 ) or generate biomass. After death, microbial necromass becomes part of the nonliving soil organic matter pool (Miltner, Bombach, Schmidt‐Brucken, & Kastner, 2012 ). Estimates of the proportion of microbially derived carbon transformed into nonliving SOC range from 40% (Kindler, Miltner, Richnow, & Kastner, 2006 ) to 80% (Liang & Balser, 2011 ), but the role of necromass carbon for SOC build‐up is not well tested in the context of changing PSR. It is, therefore, important to distinguish between microbial catabolic and anabolic pathways in order to disentangle their specific contributions to SOC accumulation. One way to synthesize microbial physiology is the widely used metric of microbial carbon use efficiency (CUE), which describes the efficiency by which microbes convert organic carbon into growth (Manzoni, Taylor, Richter, Porporato, & Agren, 2012 ; Sinsabaugh, Manzoni, Moorhead, & Richter, 2013 ). When incorporated into microbial biomass, carbon has the potential to become part of the soil organic matter pool and can reside in soils for longer time periods. Accordingly, a high microbial CUE favors SOC storage, although other physiological characteristics of the soil microbial community like microbial growth and turnover may equally promote SOC accumulation. Moreover, microbial CUE was shown to scale positively with microbial growth (Zheng et al., 2019 ) and to be maximized at highest growth rates (Manzoni et al., 2017 ). Nevertheless, despite the general importance of these microbial processes to SOC accumulation, their relationships with plant diversity are almost entirely unknown. In this study, we explicitly addressed the question of how soil microbial physiology responds to increasing plant diversity. PSR, functional group richness, and functional composition have all been shown to promote aboveground and belowground plant productivity in the Jena Experiment (Marquard et al., 2009 ; Ravenek et al., 2014 ). However, increases in root biomass were more strongly determined by PSR than by functional group richness (Ravenek et al., 2014 ) and further led to greater rhizosphere carbon inputs in high‐diversity plant communities (Chen et al., 2017 ; Lange et al., 2015 ). We here focus on how microbial activity impacts the transformation of detrital organic material to unravel the causal physiological pathways through which soil microbes promote SOC accumulation (Figure 1 ). Specifically, we used the long‐term biodiversity experiment in Jena (Roscher et al., 2004 ) to measure gross rates of microbial community growth, turnover, and CUE in grassland plots differing in plant diversity. Plant diversity was considered in three metrics: PSR (1, 2, 4, 8, 16, and 60 plant species); plant functional group richness (one, two, three, and four plant functional groups, composed of grasses, legumes, small herbs, and tall herbs); and plant functional group identity (the presence/absence of a certain plant functional group). As depicted in Figure 1 , we hypothesized (a) that microbial growth and turnover rates would increase with increasing PSR, resulting in higher amounts of microbial biomass and necromass that in turn lead to SOC accumulation; and (b) that higher PSR would increase microbial growth more than respiration, correspondingly promoting microbial CUE and leading to increased SOC storage. Figure 1 Conceptual model depicting the hypothetical relationships between plant species richness and microbial physiology that are expected to promote soil organic carbon (SOC) build‐up in species‐rich plant communities (pool sizes within, microbial processes without text frames; mic, microbial; CUE, carbon use efficiency)",
"discussion": "4 DISCUSSION Species‐rich grasslands are fundamental for many ecosystem processes and services and are important for increasing the carbon storage of terrestrial ecosystems (Hungate et al., 2017 ). Higher SOC storage is believed to be either due to greater plant inputs and/or due to lower losses of organic carbon at high levels of plant diversity, the latter of which reflects a higher efficiency of soil microbial carbon cycling. We found here that increasing PSR promoted microbial biomass both directly and indirectly through higher plant carbon inputs (as indicated by higher root carbon mass per area) and faster microbial growth. This increase in microbial biomass was, in turn, mechanistically coupled with the build‐up of SOC, as shown by piecewise SEM. Moreover, microbial turnover rates increased with increasing PSR, which most likely triggered increases in microbial necromass and thus contributed to the higher SOC content found in species‐rich plant communities. Although these connections could not be demonstrated in a single common SEM, fungal necromass significantly determined SOC in a reduced structure of the SEM (Figure S3 b). In contrast, changes in microbial respiration or CUE were not causally linked to the increase in SOC content with increasing PSR. The established positive relationship between plant diversity and productivity is commonly coupled with increased aboveground living and dead plant biomass, as well as with higher belowground biomass production and root exudation (El Moujahid et al., 2017 ; Fornara & Tilman, 2008 ; Ravenek et al., 2014 ; Roscher et al., 2005 ). Root biomass and root‐associated products, such as belowground litter and root exudates, are the main forms of carbon input into soils and represent important carbon sources for soil microbes. Greater root inputs into soils, however, can trigger decreases (Steinbeiss, Temperton, et al., 2008 ) or increases in SOC storage (Xu, Liu, & Sayer, 2013 ), depending on the responses of microbial carbon metabolism and the extent of rhizosphere priming effects. Our findings of higher microbial biomass and activity in response to increasing PSR concomitant with increased belowground carbon input as evidenced by a higher root biomass carbon is in line with earlier studies from the Jena Experiment (Eisenhauer et al., 2010 ; Lange et al., 2015 ; Strecker et al., 2015 ). However, when splitting overall ‘microbial activity’ into anabolic and catabolic processes, we observed a more pronounced increase in growth (twofold) than in respiration (1.5‐fold), indicating a relatively greater anabolic capacity of soil microbial communities at high PSR levels. We suggest that this caused soil microbial biomass to increase, which explains the higher growth rates observed per unit of soil mass. Interestingly, biomass‐specific respiration rates and microbial CUE did not respond to changes in PSR, while biomass‐specific growth rates increased significantly with increasing PSR. This is important because biomass‐specific rates represent the microbial physiology independently of microbial biomass. The latter is equivalent to microbial turnover at steady state conditions (i.e., when microbial biomass remains constant in the short term, as expected in a 24 hr measurement period, and as shown by Zheng et al. ( 2019 )), explaining why both microbial growth and microbial biomass turnover rates accelerated under increasing PSR. In the long term (at decadal scales), accelerated microbial growth and faster microbial turnover rates will promote microbial necromass formation. This accelerated production and turnover of microbial biomass is expected to promote SOC storage via ongoing iterative cycles of microbial proliferation, growth, and death, ultimately leading to incorporation of higher amounts of microbial‐derived carbon in the SOC pool of more diverse plant communities. Thus, microbial growth increased through higher plant carbon inputs has the potential to fuel the soil organic matter reservoir with microbially derived carbon due to both accelerated biomass and necromass formation (Liang, Cheng, Wixon, & Balser, 2011 ). We further stress the importance of measuring microbial respiration and growth simultaneously, when assessing microbial contributions to SOC accumulation, as both processes affect SOC dynamics in different ways (Data S1 ). The observed increase in microbial necromass carbon with PSR was mainly driven by increases in the formation of fungal necromass, since bacterial necromass did not respond to manipulations in PSR. As such, the fungal to bacterial necromass ratio also increased with increasing PSR. Fungal‐derived necromass was shown to significantly contribute to soil organic matter accumulation that was also strongly promoted by efficient microbial biomass production (Kallenbach, Frey, & Grandy, 2016 ; Li et al., 2015 ). While a previous study from the same experiment reported no PSR but plant functional group richness effects on fungal to bacterial biomass ratios based on phospholipid fatty acid analysis (Lange et al., 2014 ), we found that the corresponding necromass ratio was more strongly driven by PSR (Table 2 ) than by functional group richness (Table S1 ). More diverse plant mixtures most likely support soil microbial communities with a larger amount and higher diversity of resources, which is supported by previous work at this experiment demonstrating a higher diversity of organic compounds of low molecular weight, such as organic acids, at higher levels of PSR (El Moujahid et al., 2017 ). This suggests that the diversity of more complex compounds, such as lignins, proteins, and condensed tannins, may also be higher with increasing PSR. Both the quality and quantity of substrates are known to affect bacterial and fungal growth, with fungal growth being more promoted by complex carbon substrates and increased loading rates of available substrate compared to bacterial growth (Rousk & Baath, 2011 ). This could have translated into the higher fungal to bacterial necromass ratios observed here. Alternatively, it cannot be ruled out that the recycling of bacterial necromass by the active microbial community is faster than that of the fungal necromass at higher PSR, for example, because bacterial remains are thought to be richer in nutrients (Sterner & Elser, 2002 ) or because fungal necromass decomposition is retarded by melanin impregnation (Fernandez, Langley, Chapman, McCormack, & Koide, 2016 ). Increased labile carbon inputs can trigger the activation of dormant microbes (Blagodatskaya & Kuzyakov, 2013 ) by alleviation of their carbon limitation (Demoling, Figueroa, & Baath, 2007 ). We found little evidence, however, that soil microbes were released from carbon limitation through increased plant carbon inputs at higher PSR, since we observed no response in microbial CUE and in biomass‐specific respiration. Microbial CUE was shown to decrease and biomass‐specific respiration to increase under conditions of increasing carbon availability (Manzoni et al., 2012 ; Spohn & Chodak, 2015 ). However, biomass‐specific respiration, determined as the ratio of soil basal respiration to soil microbial biomass, does not provide any information about how much of the carbon taken up by microbes is used for microbial growth and thereby is incorporated into microbial biomass. Therefore, although biomass‐specific respiration and CUE both refer to microbial utilization of carbon, biomass‐specific respiration should not be used as a proxy for microbial CUE, which is defined by the ratio of growth over carbon uptake. Nonetheless biomass‐specific respiration is a valuable indicator as a relative measure of the degree of substrate limitation of the soil microbial community (Wardle & Ghani, 1995 ). No response of microbial biomass‐specific respiration and CUE implies that although more plant‐derived carbon will have entered the soil in more diverse plant communities, the soil microbial community most likely did not change in nutritional limitations but remained carbon limited or carbon to nutrient colimited, even at high PSR levels. This is in concordance with observations from a large suite of soils differing in land use, soil organic matter content, nutrient status, soil pH, and spanning a wide range of soil carbon to nitrogen ratios, which have shown that soil microbial growth, determined by radiotracer incorporation approaches, is most commonly limited by a lack of carbon or energy (Alden, Demoling, & Baath, 2001 ; Demoling et al., 2007 ; Kamble & Baath, 2014 ). We did not find support for our expectation that microbial CUE would change with PSR. Changes in microbial CUE, therefore, cannot explain the increase in SOC accumulation with PSR. Increasing resource carbon to nutrient ratios for soil microbial communities have been shown to decrease microbial CUE (Manzoni et al., 2012 ). In the Jena Experiment, not only the quantity but also the quality of plant biomass responded to changes in plant diversity as carbon to nitrogen ratios increased significantly with PSR. This is thought to be a consequence of altered nutrient allocation and carbon fixation patterns of aboveground vegetation (Abbas et al., 2013 ; Vogel, Eisenhauer, Weigelt, & Scherer‐Lorenzen, 2013 ), and because of shifts in the identity and proportional composition of plant functional groups, especially in the case of root stoichiometry (Chen et al., 2017 ). Plant detrital material can be expected to have even wider carbon to nutrient ratios compared to living plant tissues, due to remobilization of nutrients prior to litter production. These changes in carbon to nitrogen ratios of plant biomass and detritus most likely translated into unfavorable substrate stoichiometries for soil microbial communities, as also reflected in the increasing root and soil carbon to nitrogen ratios observed here with increasing PSR (Tables 1 and 2 for root stoichiometry only). Constant microbial CUE, therefore, also suggests that microbial communities here operate below their threshold element ratio and therefore experience persistent carbon limitation (Mooshammer, Wanek, Zechmeister‐Boltenstern, & Richter, 2014 ). When compared to PSR, functional group richness was of less importance to microbially driven SOC build‐up. Specifically, while functional group richness also promoted microbial biomass increases that translated into the build‐up of SOC, this effect was less pronounced and was neither mediated through root carbon input nor through microbial growth. This is important because previous findings have shown that both PSR and functional group richness increase aboveground community biomass (Marquard et al., 2009 ), but we demonstrate here that only PSR effects extend belowground. Despite this, we found clear effects of the presence versus absence of different functional groups, and particularly of legumes, on the soil microbial system. Specifically, we found that legumes decreased root biomass, microbial growth, microbial biomass, and turnover rates. While we did not observe a legume‐induced reduction in biomass‐specific respiration, as previously reported from the same experiment (Strecker et al., 2015 ), our findings add support to the notion that legumes have a negative impact on soil microbial processes. This is coupled with the fact that legumes have been shown to decrease root biomass (Ravenek et al., 2014 ). As a consequence, our findings suggest that legumes, as the only functional group here with negative effects on the soil microbial system, are responsible for the reductions in SOC content observed in this experiment, and act via their inhibitory influence on community‐level root biomass, and thereby on microbial biomass and activity. We posit that this legume effect arises due to the ability of legumes to fix nitrogen through symbioses with nitrogen‐fixing bacteria, causing increased soil nitrogen availability and leading to a reduced need to allocate photosynthetic carbon to root biomass at the community level. By comparison, grasses are known to invest relatively extensively in root biomass, which may be responsible for our observations that grasses supported microbially driven SOC build‐up (Figure S4 b). Tall herbs did not affect the soil microbial system, whereas small herbs significantly increased microbial growth and biomass and thus led to increases in SOC (Figure S4 c). This positive effect of small herbs was rather unexpected and needs further clarification. In conclusion, species‐rich plant communities, most likely through greater plant organic matter inputs, promoted the growth of soil microbial communities more strongly than their respiratory activity, triggering increases in microbial biomass. At the same time, microbial biomass turnover rates increased, thereby promoting microbial necromass formation. We show that these mechanisms together led to SOC accumulation. Clearly changes to the soil system are themselves a driver of change in the plant community, and thus the changes we observed to some extent reflect the coupling between shifts in plant communities and the soil system. This is the first evidence of causal links between microbial physiology, microbial biomass, and necromass build‐up and SOC storage in the context of plant biodiversity."
} | 5,672 |
35013592 | null | s2 | 5,559 | {
"abstract": "Cyanobacteria of the genus Trichodesmium provide about 80 Tg of fixed nitrogen to the surface ocean per year and contribute to marine biogeochemistry, including the sequestration of carbon dioxide. Trichodesmium fixes nitrogen in the daylight, despite the incompatibility of the nitrogenase enzyme with oxygen produced during photosynthesis. While the mechanisms protecting nitrogenase remain unclear, all proposed strategies require considerable resource investment. Here we identify a crucial benefit of daytime nitrogen fixation in Trichodesmium spp. that may counteract these costs. We analysed diel proteomes of cultured and field populations of Trichodesmium in comparison with the marine diazotroph Crocosphaera watsonii WH8501, which fixes nitrogen at night. Trichodesmium's proteome is extraordinarily dynamic and demonstrates simultaneous photosynthesis and nitrogen fixation, resulting in balanced particulate organic carbon and particulate organic nitrogen production. Unlike Crocosphaera, which produces large quantities of glycogen as an energy store for nitrogenase, proteomic evidence is consistent with the idea that Trichodesmium reduces the need to produce glycogen by supplying energy directly to nitrogenase via soluble ferredoxin charged by the photosynthesis protein PsaC. This minimizes ballast associated with glycogen, reducing cell density and decreasing sinking velocity, thus supporting Trichodesmium's niche as a buoyant, high-light-adapted colony forming cyanobacterium. To occupy its niche of simultaneous nitrogen fixation and photosynthesis, Trichodesmium appears to be a conspicuous consumer of iron, and has therefore developed unique iron-acquisition strategies, including the use of iron-rich dust. Particle capture by buoyant Trichodesmium colonies may increase the residence time and degradation of mineral iron in the euphotic zone. These findings describe how cellular biochemistry defines and reinforces the ecological and biogeochemical function of these keystone marine diazotrophs."
} | 506 |
34552573 | PMC8450586 | pmc | 5,560 | {
"abstract": "In the plant rhizosphere and endosphere, some fungal and bacterial species regularly co-exist, however, our knowledge about their co-existence patterns is quite limited, especially during invasion by bacterial wilt pathogens. In this study, the fungal communities from soil to endophytic compartments were surveyed during an outbreak of tobacco wilt disease caused by Ralstonia solanacearum . It was found that the stem endophytic fungal community was significantly altered by pathogen invasion in terms of community diversity, structure, and composition. The associations among fungal species in the rhizosphere and endosphere infected by R. solanacearum showed more complex network structures than those of healthy plants. By integrating the bacterial dataset, associations between fungi and bacteria were inferred by Inter-Domain Ecological Network (IDEN) approach. It also revealed that infected samples, including both the rhizosphere and endosphere, had more complex interdomain networks than the corresponding healthy samples. Additionally, the bacterial wilt pathogenic Ralstonia members were identified as the keystone genus within the IDENs of both root and stem endophytic compartments. Ralstonia members was negatively correlated with the fungal genera Phoma, Gibberella , and Alternaria in infected roots, as well as Phoma, Gibberella , and Diaporthe in infected stems. This suggested that those endophytic fungi may play an important role in resisting the invasion of R. solanacearum.",
"conclusion": "Conclusion The bacteria wilt pathogen Ralstonia members and the infected root endophytic fungal genera Phoma, Gibberella, Alternaria, Haematonectria, Cryptococcus, Podospora, Spodiobolus, Malassezia, Aleuria, Dioszgia , and Davidiealla , as well as the infected stem endophytic fungal genera Phoma, Gibberella, Didymella , and Diaporthe , were negatively correlated, and these fungi may be potential biocontrol resources in dealing with tobacco bacterial wilt disease. At present, there are few reports on the exploration and application of tobacco endophytic fungal resources. This study will provide potential ideas and theoretical support for enriching the study of tobacco endophytic fungal resources and controlling tobacco bacterial wilt disease. Further experiment with species isolation and verification is needed to confirm these findings.",
"introduction": "Introduction Ralstonia solanacearum , the causative agent of soil-borne bacterial wilt disease in plants, is often found in agricultural land used for tobacco cultivation. Once the pathogen invades the plant root system, it rapidly spreads to the stem, causing an internal system imbalance of the entire tobacco plant, accelerating senescence and death ( Li et al., 2017 ; Wei et al., 2018 ). As important members of the plant microecosystem ( Nilsson et al., 2019 ), fungi play a vital role in promoting the material cycle of agro-ecosystem and affecting plant growth and health ( Lilleskov et al., 2011 ; Zhang et al., 2021 ). Therefore, dynamic changes in the structure and composition of the plant soil fungal community can indicate the alterations of soil micro-ecological environment ( Shi et al., 2020 ). Plant endophytic fungi exist within the host plant and interact closely with other microorganisms to promote plant growth, resist the invasion of plant pathogens, and improve the disease resistance of host plant through their own metabolites or induction of the host’s metabolites ( Arnold, 2007 ; Mousa et al., 2016 ; Bastias et al., 2017 ). However, our knowledge about how the soil and endophytic fungal communities change under the invasion of R. solanacearum is quite limited. With the surge of research on microbial communities in various ecological environments, the interdomain relationships between different types of microbial communities has attracted great attention ( Kapitan et al., 2019 ; Wang et al., 2020 ). The association of plant bacterial and fungal communities is critical to overall microbial community structure and plant health ( Hassani et al., 2018 ), and intrigues numerous botanists and microbiologists. Plant microbial communities live and colonize several zones including the bulk soil, rhizosphere soil, phyllosphere, and endosphere ( Bulgarelli et al., 2013 ; Kumar et al., 2017 ). They play a vital role in the acquisition of nutrients by plants, mutual defense, and co-evolution ( Martin et al., 2017 ; Fitzpatrick et al., 2018 ). Understanding the relationship between related microorganisms in natural ecosystems may help us better explore and learn about the assembly, diversity, and stability of plant-related communities ( Haq et al., 2014 ; van der Putten, 2017 ). Recently, researchers have attempted to identify plant-related microbial communities from bacteria to fungi, as well as their associations, by means of microbiome annotation database or molecular-based experimental methods, broadening our basic knowledge of these types of microbial communities ( Jackrel et al., 2017 ; Bamisile et al., 2018 ; Levy et al., 2018 ). Endophytic fungi are often considered to be beneficial to their host plants ( Qian et al., 2019 ), because they may play various ecological roles such as promoting plant growth, enhancing the absorption of nutrient, resistance against various plant pathogens, as well as tolerance against various biotic and abiotic stresses ( Waqas et al., 2012 ; Jia et al., 2016 ; Terhonen et al., 2016 ; Ripa et al., 2019 ). Current studies on tobacco wilt, a bacterial disease, mainly focus on the diversity and structural composition of the tobacco soil microbial community and its correlation with soil physicochemical properties, as well as the associations between the pathogen and bacterial species ( Jiang et al., 2017 ; Li et al., 2017 ). However, the associations between the closely related bacterial and fungal communities in various zones of the plant-soil microecosystem under the invasion of R. solanacearum and the roles of endophytic fungi remains unclear. The purposes of the current study are to: (i) Illuminate the characteristics of fungal communities in the bulk and rhizosphere soils and the root and stem endophytic compartments of healthy tobacco plants and those infected by R. solanacearum ; (ii) Reveal the associations among species of fungal communities from various zones of the plant-soil microecosystem via molecular ecological network analysis; (iii) Explore the associations between fungi and bacteria through interdomain ecological network (IDEN) analysis; (iv) Study the associations between pathogenic Ralstonia members and fungi through sub-network analysis, and to explore fungal biocontrol resources that may antagonize the bacterial wilt pathogen. From the obtained results, we will provide a new strategy and theoretical support for enriching the study of tobacco endophytic fungal resources and for exploring the antagonistic fungal resources targeting R. solanacearum.",
"discussion": "Discussion The occurrence of plant diseases is closely related to the microbial diversity in the soil and endophytic compartments ( Shi et al., 2019 ; Ulloa-Munoz et al., 2020 ). Our previous studies have shown that the diversity of endophytic bacterial communities in the roots and stems of plants with bacterial wilt infected was significantly higher than that of healthy samples ( Hu Q.L. et al., 2020 ). In this study, there was no significant difference in the diversity of the root endophytic fungal community between infected and healthy samples. Endophytic fungal community diversity of the infected stem samples was significantly higher than that of the healthy samples. This may be explained by that the bacterial wilt pathogen invasion success within root endophyte could result in the destruction of plant’s defense system. The bacterial wilt pathogen and fungal members interacted closely in the roots, inducing more fungal communities to migrate into the stem, and therefore resulted in an increased diversity of the stem endophytic fungal community compared to the healthy samples ( Tan and Zou, 2001 ; Thebault and Fontaine, 2010 ; Kefi et al., 2012 ). Moreover, high diversity and close associations among various microorganism are beneficial to the stability of microbial communities, thereby boosting the microbial community’s resistance to pathogen invasion ( McCann, 2000 ; Wehner et al., 2010 ; van Elsas et al., 2012 ; Mallon et al., 2015 ). The increased diversity of fungal communities in stem endophytes may be a middle-late-stage immune response of the plant to the bacterial wilt pathogen invasion. The invasion of bacterial wilt pathogen may cause changes in the fungal community composition of the various zones of the tobacco microecosystem. From the perspective of relative abundance, fungal composition displayed significant changes at the genus level between infected and healthy samples in the bulk soil, rhizosphere soil, and root and stem endophytic compartments. In the infected bulk soil and rhizosphere soil, the relative abundances of Rhodotorula, Ceratobasidium, Cyberlindnera, Podospora, Conocybe, Monoblepharis, Paraconiothyrium , and Phoma were significantly decreased, whereas the relative abundances of Gibberella , Cryptococcus, Mucor, Nectria, Debaryomyces , and Haematonectria were significantly increased, compared to the corresponding healthy samples. Such changes in composition might be because the invasion of bacterial wilt disease made pathogenic Ralstonia members the dominant species in soil and thus altered the composition of the soil fungal community. In the infected endophytic samples, the genera that significantly declined were Cryptococcus, Didymella, Mortierella, Paraphoma, Davidiealla, Phoma , and Mucor . Many of the secondary metabolites produced by these endophytic fungi have been reported to have inhibiting or antibacterial abilities ( Melo et al., 2014 ; Xia et al., 2015 ; Li et al., 2018 ). These results indicated that the beneficial endophytes were either actively repelled by the host immune system or defeated by the more dominant migrating microbial community ( Lundberg et al., 2012 ; Velásquez et al., 2017 ). The relative abundances of Haematonectria, Gibberella, Ceratobasidium, Nectria, Bionectria , and Didymella were significantly enhanced in the infected endophytic samples, indicating that they may benefit during the pathogen invasion process. It is possible they are opportunists that took advantage of the potential niche opened by pathogen invasion and entered the plant endophytic compartment ( Lundberg et al., 2012 ). The compositional changes of these fungal communities may be caused by changes in root exudates or complex changes in the plant immune system during pathogen invasion ( Martinoia and Baetz, 2014 ; dos Santos et al., 2020 ), and this promoted the differential recruitment and/or differential rejection of microorganisms to resist the invasion of bacterial wilt pathogen in plant roots and stems ( Kwak et al., 2018 ). Microbe-microbe associations are essential for the function of microecosystems in soil and endophytic compartments ( Barberán et al., 2012 ). Molecular ecological network (MEN) analysis has been increasingly employed to explore potential microbial associations in various ecosystems ( de Menezes et al., 2015 ). However, there are few reports on microbial associations in the fungal community of plants under invasion by bacterial wilt pathogen. In this study, we applied network analysis to quantify and visualize the associations among microorganisms of the fungal community under the invasion of R. solanacearum . The results showed that the fungal networks of infected samples had higher complexity and more links than the healthy samples in the rhizosphere soil and root and stem endophytic compartments. Furthermore, the corresponding topological structures demonstrated significant differences as well. Together this indicated that the invasion of bacterial wilt pathogen changed the composition of the soil fungal community and further strengthened the associations among species in the fungal community. The highly connected and modularized fungal community associations were conducive to regulating the stability of the community ( Eisenhauer et al., 2013 ; Downing et al., 2014 ; Tardy et al., 2014 ), thereby controlling the propagation and colonization of the pathogen. Hence, it is necessary to study the associations of microorganisms in soil and endophytic fungal communities for more effective prevention and control of diseases. Soil is one of the main habitats for bacteria and fungi ( Effmert et al., 2012 ). Endophytic microorganisms, including fungi and bacteria, live in the intercellular or intracellular spaces of plant tissues. The associations between fungi and bacteria are part of the communication network maintaining the balance of this microhabitat ( Bamisile et al., 2018 ). We adopted IDENs to analyze the association network between fungi and bacteria in each zone of the plant-soil microecosystem, and found more complex and tighter fungal-bacterial associations in the infected samples than the corresponding healthy samples for all tested zones. In addition, the IDEN of the infected root endophytic compartment presented the most network links and the highest number of links per microorganism, suggesting a closer bacterial-fungal associations in this network. Interestingly, the Zi-Pi plots demonstrated that the pathogenic Ralstonia members were the keystone genus in the root and stem endophytic bacterial-fungal association networks. The reason for these results may be that with the invasion of the bacterial wilt pathogen, more soil fungi and bacteria developed a mutually beneficial relationship and entered the plant root endophytic community together, resulting in more complex associations among microorganisms ( Hu Q.L. et al., 2020 ). It may also be that the competition of nutrient resources or niche space caused more diversified associations between fungi and bacterial microorganisms ( Ghoul and Mitri, 2016 ). A third possibility is that because the microecological balance was broken by the pathogen invasion, leading to more intense antagonistic relationships between fungal, other bacterial members, and the pathogen ( García-Bayona and Comstock, 2018 ; Hu J. et al., 2020 ). These phenomena were more prominent in infected root and stem endophytic compartments. To further clarify which fungi interacted closely with the pathogenic Ralstonia members in the endophytic roots and stems, we built sub-networks centered on the pathogen, Ralstonia members, and included its associations with fungi. The results showed that the root endophytic fungal genera Phoma, Gibberella, Alternaria, Haematonectria, Cryptococcus, Podospora, Spodiobolus, Malassezia, Aleuria, Dioszgia , and Davidiealla , and the stem endophytic fungal genera Phoma, Gibberella, Diaporthe , and Didymella were all negatively correlated with Ralstonia members. Plant-associated endophytic fungi are rich sources of novel bioactive and structurally diverse secondary metabolites and other natural products, which were generally considered to protect their host plants by blocking or inhibiting the appropriate pathogenic microorganisms ( Rustamova et al., 2020 ). According to previous research, the active compound named as barceloneic acid C isolated and purified from the secondary metabolites of the endophytic fungus Phoma sp. JS752, isolated from Phragmites communis Trinius, demonstrated an antibacterial activity against pathogenic gram-positive bacteria Listeria monocytogenes and Staphylococcus pseuditermedius , and gram-negative bacteria such as Escherichia coli and Salmonella typhimurium ( Xia et al., 2015 ). The purified secondary metabolites [(3S)-3,6,7-trihydroxy-α-tetralone, Cercosporamide, β-Sitosterol and trichodermin] of Phoma sp. ZJWCF006, which was screened and isolated from the Arisaema erubescens endophytes, showed remarkable antibacterial activity against four plant fungal pathogens ( Fusarium oxysporium, Rhizoctonia solani, Colletotrichum gloeosporioides, and Magnaporthe oryzae ) and two plant bacterial pathogens ( Xanthomonas campestris and Xanthomonas oryzae ) ( Wang et al., 2012 ). The secondary metabolity compound ergosterol peroxide from the endophytic fungus Gibberella moniliformis JS1055, isolated from a halophyte Vitex rotundifolia ( Kim et al., 2018 ), exhibited moderate inhibitory activity against bacteria Staphylococcus aureus and Escherichia coli ( Zhu et al., 2017 ). The endophytic fungus Alternaria alternata AE1, isolated from Azadirachta indica A. Juss, could produce highly effective bioactive metabolites that showed a strong inhibitory effect on pathogenic bacteria Listeria monocytogenes and Escherichia coli ( Chatterjee et al., 2019 ). The antibacterial activities of compounds phomosines A and C produced by endophytic fungus Diaporthe sp. F2934, isolated from the tropical plant Aegle marmelos , showed an antibacterial activity against a variety of gram-negative and gram-positive bacteria, and its inhibitory zone diameter (IZD) against Staphylococcus aureus was 20% larger than the standard antibiotic vancomycin ( Sousa et al., 2016 ). It can be seen that these fungi revealed by our study and their secondary metabolites have been reported with antibacterial ability or activity against some bacteria, and they may have potential resistance to bacterial wilt pathogen invasion."
} | 4,387 |
37982076 | null | s2 | 5,561 | {
"abstract": "Interbacterial antagonism can significantly impact microbiome assembly and stability and can potentially be exploited to modulate microbes and microbial communities in diverse environments, ranging from natural habitats to industrial bioreactors. Here we highlight key mechanisms of interspecies antagonism that rely on direct cell-to-cell contact or diffusion of secreted biomolecules, and discuss recent advances to provide altered function and specificities for microbiome engineering. We further outline the use of ecological design principles based on antagonistic interactions for bottom-up assembly of synthetic microbial communities. Manipulating microbial communities through these negative interactions will be critical for understanding complex microbiome processes and properties and developing new applications of microbiome engineering."
} | 212 |
38733019 | PMC11086169 | pmc | 5,562 | {
"abstract": "The burgeoning interest in intelligent transportation systems (ITS) and the widespread adoption of in-vehicle amenities like infotainment have spurred a heightened fascination with vehicular ad-hoc networks (VANETs). Multi-hop routing protocols are pivotal in actualizing these in-vehicle services, such as infotainment, wirelessly. This study presents a novel protocol called multiple junction-based traffic-aware routing (MJTAR) for VANET vehicles operating in urban environments. MJTAR represents an advancement over the improved greedy traffic-aware routing (GyTAR) protocol. MJTAR introduces a distributed mechanism capable of recognizing vehicle traffic and computing curve metric distances based on two-hop junctions. Additionally, it employs a technique to dynamically select the most optimal multiple junctions between source and destination using the ant colony optimization (ACO) algorithm. We implemented the proposed protocol using the network simulator 3 (NS-3) and simulation of urban mobility (SUMO) simulators and conducted performance evaluations by comparing it with GSR and GyTAR. Our evaluation demonstrates that the proposed protocol surpasses GSR and GyTAR by over 20% in terms of packet delivery ratio, with the end-to-end delay reduced to less than 1.3 s on average.",
"conclusion": "5. Conclusions This study introduces the MJTAR protocol tailored for urban vehicular ad-hoc networks. MJTAR operates as a geographic routing protocol, leveraging 2-hop junction-based vehicle traffic density and curve metric distance to the destination. The protocol innovates by proposing an E-IFTIS mechanism that estimates vehicle traffic conditions up to 2-hop junctions without dependence on fixed infrastructure. Additionally, it introduces the OMSS mechanism, employing an ant colony algorithm. Notably, MJTAR surpasses GSR and GyTAR in network performance, offering a probability-based algorithm to explore multiple junction paths with low latency and a high probability of packet delivery success. In future investigations, we aim to explore mechanisms for predicting vehicle traffic density using multi-junction-based machine-learning techniques. Furthermore, we intend to investigate forwarding node selection techniques that account for link quality to mitigate packet loss on roads between junctions. Finally, we plan to analyze the probability of packet collisions that may occur when multiple sources transmit packets to a destination in network simulations and study collision avoidance methods.",
"introduction": "1. Introduction Intelligent transportation systems (ITSs) are garnering increasing attention globally due to the proliferation of vehicles and the emergence of various issues such as traffic congestion, air pollution, and traffic accidents. ITS integrates wireless and advanced IT technologies into mobile vehicles to enhance traffic safety and driver convenience. IEEE has introduced a new wireless communication standard called wireless access in vehicular environment (WAVE), based on IEEE 802.11p, to deliver these services. Consequently, research on vehicular ad-hoc networks (VANETs) is becoming increasingly active [ 1 , 2 ]. VANETs represent a next-generation network technology that facilitates vehicle-to-vehicle (V2V) and vehicle-to-infrastructure (V2I) communications and is a subclass concept of traditional mobile ad-hoc networks (MANETs). MANETs can establish independent ad-hoc networks among mobile nodes to construct autonomous networks. However, due to the mobile nodes’ limited characteristics, such as low bandwidth, low power, and limited resources, MANETs have limitations in realizing various services. In contrast, VANETs can support a broad spectrum of applications as vehicles are equipped with onboard units (OBUs) capable of robust processing without power limitations. These applications include safety measures like vehicle collision avoidance, emergency message dissemination, and traffic accident notification; driver convenience features like alternative route guidance, parking lot location, and gas station payment; and entertainment offerings such as games, movies, and music [ 3 , 4 ]. In VANETs, multi-hop routing protocols are crucial for supporting application services via wireless communication between vehicles. While various dynamic routing protocols for mobile networks have been proposed in existing MANET research [ 5 , 6 , 7 , 8 ], these protocols are unsuitable for VANETs. Designing a routing protocol for vehicular networks poses several challenges. Vehicles move rapidly, leading to frequent topology changes and network link breaks. Moreover, real-time variations in link quality occur based on vehicular traffic density. For instance, low traffic density results in higher packet delay due to fewer vehicles being available to relay packets, whereas high traffic density leads to lower packet delay. Urban environments can exacerbate these challenges by causing packet loss due to obstacles like tall skyscrapers interfering with radio signals. Therefore, vehicular routing protocols must address these demanding conditions diligently. Greedy perimeter stateless routing (GPSR) stands out as a geographically based routing protocol deemed highly suitable for VANETs. It leverages GPS to identify nearby neighbors and employs greedy forwarding to direct packets to the node closest to the destination. GPSR is recognized for its speed, adaptability, and scalability, facilitated by its maintenance of a one-hop neighbor table. However, GPSR is primarily designed for highway scenarios, so it encounters frequent communication drops and high packet latency in urban scenarios [ 9 ]. Map-based geographic source routing (GSR) represents a junction-based geographic routing protocol that combines geographic-based (GPSR) and topology-based (DSR) protocols [ 10 ]. This method employs the vehicle’s digital map to establish a fixed sequence of junctions for routing packets to their destination. However, the junction sequence prioritizes the shortest distance to the destination, neglecting the traffic density through which the packets traverse. Conversely, the enhanced greedy traffic-aware routing protocol (GyTAR) integrates traffic awareness into the existing GSR protocol, introducing a junction-based geographic routing protocol [ 11 ]. GyTAR dynamically selects one junction at a time as traffic conditions evolve. The criterion for junction selection is based on the junction’s highest density of vehicle traffic and shortest distance to the destination among neighboring junctions. The GyTAR routing protocol has exhibited superior network performance compared to GSR. With the rapid proliferation of vehicles in recent years, traffic monitoring systems have become an indispensable component of ITS. These systems can monitor sensor-equipped traffic vehicles for identification, speed, and traffic congestion in real-time, offering various applications. The smart traffic monitoring system (STMS) [ 12 ] is a traffic surveillance framework designed for monitoring traffic congestion and managing traffic lights. It operates as a fog node, gathering real-time data from geographically dispersed sensors and transmitting them to the cloud for storage and processing. Additionally, it can be adapted to detect traffic incidents that require immediate assistance amidst congested traffic. However, the STMS faces challenges due to the need to transmit the collected data to a base station, which can lead to bandwidth constraints, substantially impacting latency-sensitive applications. The infrastructure-free traffic information system (IFTIS) [ 13 ] offers a distributed mechanism for vehicles to collaborate in collecting traffic information on a road segment without relying on roadside units (RSUs). In this approach, the group leader of each cell in a road segment gathers traffic data and forwards them to the next junction using a greedy strategy. This technique offers substantial cost advantages as it eliminates the need for fixed infrastructure such as base stations. Moreover, it benefits from improved connectivity with an increase in the number of vehicles. Recent research in vehicular networks has attempted to utilize artificial intelligence to address network latency and energy efficiency issues. In [ 14 ], they proposed an algorithm to improve QoS in vehicular networks by jointly scheduling deep neural network (DNN) inference tasks at the microarchitectural and network levels. This algorithm is a technique for making two-level scheduling decisions that utilizes deep reinforcement learning to respond to dynamic environments. The technique aims to minimize the total weighted sum of response time and energy consumption for all jobs under the constraints of response time, energy consumption, and storage capacity. In [ 15 ], an energy-efficiency secure offloading (EESO) technique based on asynchronous advantage actor–critic (A3C) was proposed in non-orthogonal multiple access (NOMA) offloading scenarios using physical layer security (PLS) techniques. This technique applies asynchronous deep reinforcement learning for highly dynamic automotive edge computing to reduce the energy consumed by the system and protect confidential information from eavesdropping. The centralized routing scheme with mobility prediction (CRS-MP), proposed in [ 16 ], introduces a centralized routing protocol for end-to-end unicast communication. By leveraging artificial neural networks (ANNs) to learn vehicle mobility patterns, the protocol offers a routing technique to improve packet delivery probability and minimize delay. Notably, the scheme does not necessitate continuous monitoring of vehicle mobility. Instead, it dynamically selects the routing path based on the probability of vehicle mobility, either by the software-defined networking (SDN) controller or the RSU/base station (BS). Our study proposes multiple-junction-based traffic-aware routing (MJTAR), demonstrating lower latency and higher packet delivery success rates than existing traffic-aware geographic routing methods such as GyTAR [ 11 ]. MJTAR incorporates two fundamental mechanisms. First, it introduces an enhanced infrastructure-free traffic information system (E-IFTIS) mechanism to identify vehicle traffic from multiple junctions. E-IFTIS extends the capabilities of existing IFTIS [ 13 ] to collect vehicle traffic data for a road segment from up to two junctions. Second, MJTAR offers an optimal multiple junction selection scheme (OMSS) mechanism to determine the optimal packet path based on multiple junctions. OMSS utilizes an ant colony optimization algorithm [ 17 ] to compute the connectivity probability for each road segment between junctions. Subsequently, it dynamically selects multiple junctions (two-hop-based junctions) based on the connectivity probability of the road segments. The research contributions of the MJTAR routing protocol are outlined as follows: We propose an optimal multiple junction selection scheme (OMSS) algorithm that utilizes the ACO algorithm to select the optimal multiple junctions. This algorithm employs a stochastic formula to explore the optimal multiple junctions by mimicking the behavior of biological ants. We present a distributed mechanism for estimating vehicle traffic density based on multiple junctions in a purely ad-hoc environment, eliminating the need for fixed infrastructure such as roadside units (RSUs). The remainder of the paper is structured as follows: Section 2 briefly introduces the related work in this research area. Section 3 outlines the necessity of this work and introduces the MJTAR mechanism proposed in this study. Subsequently, Section 4 presents the performance evaluation based on extensive simulations using NS-3 [ 18 ] and SUMO [ 19 ] simulators. Finally, Section 5 concludes the thesis."
} | 2,958 |
36399572 | PMC9674295 | pmc | 5,563 | {
"abstract": "Micro-objects stick tenaciously to each other—a well-known show-stopper in microtechnology and in handling micro-objects. Inspired by the trigger plant, we explore a mechanical metastructure for overcoming adhesion involving a snap-action mechanism. We analyze the nonlinear mechanical response of curved beam architectures clamped by a tunable spring, incorporating mono- and bistable states. As a result, reversible miniaturized snap-through devices are successfully realized by micron-scale direct printing, and successful pick-and-place handling of a micro-object is demonstrated. The technique is applicable to universal scenarios, including dry and wet environment, or smooth and rough counter surfaces. With an unprecedented switching ratio (between high and low adhesion) exceeding 10 4 , this concept proposes an efficient paradigm for handling and placing superlight objects.",
"introduction": "INTRODUCTION Nature has evolved enviable materials architectures with programmable response to ensure survival ( 1 ). A prominent example is the so-called snap-through elastic instability, which allows, for example, the Venus flytrap to catch prey by closing its leaves ( 2 ), the hummingbird to elastically snap the beak to catch insects ( 3 ), or the Australian triggerplant Stylidium to immobilize and pollinate insects when touched ( Fig. 1A ) ( 4 ). Inspired by such powerful actuation mechanisms, applications include inflatable soft jumpers ( 5 ), bistable valves for autonomous control ( 6 ), bifurcation-based embodied logic ( 7 ), wave transition with bidirectionality ( 8 ), stable memory with reprograming ability ( 9 ), multistable inflatable origami ( 10 ), and devices to absorb energy under dynamic impact ( 11 ) or to tailor the mechanical behavior ( 12 ). To trigger the snap-through, stimuli such as swelling ( 7 ), hydraulic or pneumatic control ( 5 , 6 , 13 ), dielectric ( 14 , 15 ), and temperature ( 16 , 17 ) have been reported. Here, we present a bioinspired design that derives heuristically from the phenomenon that the triggerplant Stylidium uses for pollination by snap-through action and for solving a crucial problem in microrobotic handling: the release and placement of superlight objects that, for dimensional reasons, tend to stick tenaciously to other surfaces. Fig. 1. Concept of the bioinspired two-way switchable adhesive based on a curved beam with tunable boundary condition. ( A ) Release of elastic energy for dynamic pollination of the triggerplant ( Stylidium ). ( B ) Schematic demonstrating the switching principle derived. At small compressive preloads, the object is picked up and then moved to a new position. For release, higher loading induces the snap-through bistability. Upon release, the system relaxes back to its original shape. Insets represent the force analyses that include the weight of the object, F g and the adhesive force, F ad . ( C ) Schematic of the curved beam model with a fully (left) and partially (right) clamped end. The horizontal displacement of the right end is constraint by a spring with stiffness, k . When the applied normal displacement, d , is exerted at the center of the beam, the length of the beam changes from L 0 to L . ( D ) The two limiting boundary conditions lead to either monostable ( k = 0) or bistable ( k = ∞) response. ( E ) Normalized force-displacement and ( F ) elastic strain energy-displacement curves of a curved beam with the shape parameters L 0 = 24 mm, h / L 0 = 0.2, h / t = 4.8, and b = 6 mm. The normalized stiffness of the spring k ¯ = k L 0 3 / E I was 200 (yellow), 700 (green), and 20,000 (purple). Solid lines represent the analytical solutions, and the dot-dashed lines represent the results of the finite element analysis (FEA). ( G ) Design map for monostable (yellow and green) and bistable (purple) response in terms of the normalized stiffness k L 0 3 / E I ∈ [ 1 0 2 , 1 0 5 ] and the shape parameter h / L 0 ∈ [0.03,0.33]. Blue circles correspond to the individual curves shown in (D) and (E). A prerequisite for the occurrence of a snap-through instability is the existence of bi- or metastable states. Two common designs that have been well studied are the curved beam ( 18 – 23 ) and the curved shell ( 5 , 24 , 25 ). By using beam and shell theories, the mechanical responses including force-displacement and pressure-volume relationships have frequently been investigated. However, an open question is the impact of varying boundary conditions on the snap-through bistability and the formulation of explicit conditions for the transition from mono- to bistable states. The application of such a transition between two different mechanical responses could constitute a switching strategy, e.g., for innovative adhesive grippers that are actuated by robotic systems ( 26 ). Switchable adhesives are programmable systems exhibiting an “on” state—for adhering to the target object—and an “off” state—for the release on demand ( 27 ). Existing concepts rely on electric and electromagnetic ( 28 – 31 ), chemical ( 32 – 34 ), thermal ( 35 , 36 ), and mechanical inputs ( 37 – 42 ). Most of them, however, are based on specific responsive materials or function only in combination with subsidiary devices, both of which enhance the complexity of the adhesive system and are difficult to integrate into microsystems. An exception is the purely mechanical release that can be integrated into the trajectory of a robotic system without further equipment. Examples include compression-induced detachment via buckling ( 37 – 42 ) and the shear actuation of unidirectional adhesives ( 43 – 47 ). Intriguingly, we learn that the fast movement of the column of the triggerplant is evoked by external mechanical stimuli, i.e., the touching by the nectar-gathering insects ( Fig. 1A ). The pollen is then released from the column, which inspires us to release the micro-objects in a similar way without requiring additional means like heat, chemical, electric/magnetic-field generators, etc. In this work, we use a metastructure concept to create a snap-through instability for gripping and releasing microscopic objects. The approach is based on the tuning of boundary conditions for the snap-through instability, which allows a reversible transition from mono- to bistability. The metastructure design consists of three parts: a rigid frame, a laminated spring, and a curved beam (U-shaped) unit ( Fig. 1B ). During the pick-and-place process, as shown in the schematic in Fig. 1B , the rigid frame tunes the boundary condition of the curved beam unit from “compliant” to “stiff” in conjunction with the compression load. Thus, because of the triggering of the snap-through of the curved beam, the contact area between the curved beam and the target object separates, resulting in the releasing action. Thereafter, the compressed laminated spring pushes the curved beam out of the rigid frame for cyclic operation of the metastructure. First, we theoretically analyze a curved beam model with tunable boundary conditions ( Fig. 1C ), i.e., a fully and a partially clamped end, where the latter is confined by a spring with a given stiffness. We systematically investigate and optimize the evolution of the force-displacement response of the curved beam and validate our theoretical model with numerical simulations and experiments. Then, we design a two-way switchable adhesive as a general platform for reversibly switching between mono- and bistability. Last, a two-photon direct writing is used to downscale the design for handling superlight objects with smooth and rough surfaces and under versatile circumstances including dry and underwater conditions.",
"discussion": "DISCUSSION An important figure of merit for evaluating switchable adhesive mechanism is the switching ratio, i.e., the adhesion force ratio between the on and the off states. Most demonstrated values so far typically ranged from 2 to 10 for mechanically triggered detachments ( 27 , 39 ). For calculating the switching ratio of our design, the gravitation of the lightest object of 6 μg in our demonstrations can be considered as the smallest adhesion due to its successful release. Last, a high switching ratio of 10 4 is derived (see section S2 in the Supplementary Text). Such a high value exceeds those for most switchable adhesives reported in the literature ( 50 , 51 ) and, especially, is much higher than those associated with devices with mechanical triggers (typically ~2 to 10) ( 27 , 39 ). In the present work, we introduce a novel strategy for a release mechanism based on a reversible snap-through mechanism inspired by biological examples. A curved beam model was proposed as a two-way switchable device. The mechanics was examined in analytical detail and validated by numerical simulations and experiments. The following conclusions can be drawn: 1) The proposed beam model led to the successful realization of a bistable mechanism. The response of the curved beam could be programmed by the boundary conditions. If one end of the curved beam was constrained by a spring, then the deformation was (i) monostable for low stiffness and (ii) bistable for high stiffness. 2) A novel design map predicts the relationship between the beam parameters and the switching behavior. It allows for rational design of the desired functionality. In an improved design, with variable thickness along the beam, the instability was triggered at a smaller displacement, and higher buckling modes were suppressed. This can result in more precise and reliable handling performance. 3) Miniaturization was demonstrated by high-resolution direct printing from two resins of different moduli using two-photon lithography. Handling and release of micro-objects under dry and wet conditions on smooth and rough surfaces were successfully demonstrated. After the object was released, the elastic restoring force returned the system to its initial state. Pick-and-place of a 6-μg silicon chip (the lightest among all tested samples) was demonstrated, with an unprecedented switching ratio of more than 10 4 , which is three orders of magnitude larger than by a typical buckling-induced release using micropillars ( 39 , 40 ). The proposed snap-through concept has the potential to create an efficient paradigm for handling and placing superlight objects such as, e.g., microlight-emitting diodes, microlenses, or components for microelectrical-mechanical systems."
} | 2,613 |
38380057 | PMC10877092 | pmc | 5,564 | {
"abstract": "Abstract The effective utilization of cellulose and hemicellulose, the main components of plant biomass, is a key technical obstacle that needs to be overcome for the economic viability of lignocellulosic biorefineries. Here, we firstly demonstrated that the thermophilic cellulolytic fungus Myceliophthora thermophila can simultaneously utilize cellulose and hemicellulose, as evidenced by the independent uptake and intracellular metabolism of cellodextrin and xylodextrin. When plant biomass serviced as carbon source, we detected the cellodextrin and xylodextrin both in cells and in the culture medium, as well as high enzyme activities related to extracellular oligosaccharide formation and intracellular oligosaccharide hydrolysis. Sugar consumption assay revealed that in contrast to inhibitory effect of glucose on xylose and cellodextrin/xylodextrin consumption in mixed-carbon media, cellodextrin and xylodextrin were synchronously utilized in this fungus. Transcriptomic analysis also indicated simultaneous induction of the genes involved in cellodextrin and xylodextrin metabolic pathway, suggesting carbon catabolite repression (CCR) is triggered by extracellular glucose and can be eliminated by the intracellular hydrolysis and metabolism of oligosaccharides. The xylodextrin transporter MtCDT-2 was observed to preferentially transport xylobiose and tolerate high cellobiose concentrations, which helps to bypass the inhibition of xylobiose uptake. Furthermore, the expression of cellulase and hemicellulase genes was independently induced by their corresponding inducers, which enabled this strain to synchronously utilize cellulose and hemicellulose. Taken together, the data presented herein will further elucidate the degradation of plant biomass by fungi, with implications for the development of consolidated bioprocessing-based lignocellulosic biorefinery.",
"introduction": "Introduction Plant biomass is the most abundant renewable resource on earth and considered as a potential feedstock for producing biofuels and biochemicals. Lignocellulosic biomass is primarily composed of cellulose, hemicellulose, and lignin, combining to form a complex rigid structure that is difficult to degrade ( 1 ). The efficient and complete utilization of all components of plant biomass is critical to the economic viability of lignocellulosic biorefineries. Cellulose is a homopolymer composed exclusively of hexose sugars. It consists of linear chains of d -glucopyranose units, each linked by β-1,4-glycosidic bonds. These chains are further stabilized by both intramolecular and intermolecular hydrogen bonds ( 2 ). Hemicellulose, which is interwoven with cellulose fibers, is a diverse biopolymer. Its composition includes xylose, arabinose, mannose, galactose, among other sugars, contributing to its heterogeneity ( 3 ). Xylan, a predominant form of hemicellulose in plant cell walls, features a linear backbone of β-1,4-linked xylosyl residues. This structure is occasionally modified by the addition of side groups, such as arabinose and glucuronic acid, which can influence its properties and interactions ( 4 ). In traditional biorefineries, fermentable sugars are obtained from biomass via a pretreatment and saccharification, with glucose and xylose being the most abundant monosaccharides in the hydrolysate ( 5 ). The asynchronous and inefficient utilization of monosaccharides due to carbon catabolite repression (CCR) in the microorganisms used for fermentation is a major factor preventing the cost-effective bioconversion of plant biomass ( 6–8 ). Furthermore, some microorganisms traditionally used for fermentations (e.g. Saccharomyces cerevisiae ) lack pathways for the catabolism of xylose and arabinose, which are the main building blocks of hemicellulose. To overcome these problems, rational pathway engineering ( 9 , 10 ), evolutionary engineering ( 11 , 12 ), and global regulatory system rewiring ( 13 , 14 ) have been applied for increasing the efficiency of xylose assimilation. Moreover, the engineering of sugar transporter ( 15 , 16 ) and adjustments to the metabolic flux of glucose and xylose ( 17 , 18 ) have been performed to enable the simultaneous utilization of glucose and xylose. Nevertheless, the catabolic rate of xylose is still significantly slower than that of glucose. In addition, the cofermentation of cellobiose and xylose has been proposed as an alternative strategy, as replacing glucose with cellobiose eliminates the inhibitory effects of glucose on xylose uptake by S. cerevisiae ( 19 ). Compared with the process of separate hydrolysis and fermentation with abundant monosaccharides as intermediate products, consolidated bioprocessing (CBP) that integrates enzyme production, saccharification, and fermentation in one step by one or more microorganisms offers a viable alternative for the efficient degradation and utilization of cellulose and hemicellulose ( 20 ). Several filamentous fungi have evolved sophisticated enzymatic machineries to directly deconstruct biomass into soluble fermentable sugars ( 21–23 ), making them potentially useful for CBP. The synthesis and secretion of these lignocellulolytic enzymes are tightly regulated in different nutrient environments, in which cellulase and hemicellulase can be induced by their respective substrates (i.e. cellulose and hemicellulose, respectively) to initiate the carbon utilization ( 24–26 ). Unlike the single carbon source, lignocellulosic biomass has a complex composition that requires filamentous fungi to precisely regulate carbon metabolism-related processes during biomass degradation. However, it remains unclear how filamentous fungi coordinate the metabolism of different lignocellulose components, especially cellulose and hemicellulose, in the face of a complex lignocellulose. Earlier research confirmed that CCR is an evolved trait in filamentous fungi ( 8 ), and that it affects the production of cellulolytic enzymes as well as the uptake and catabolism of other carbon sources in the presence of glucose ( 27–29 ). The strategies employed by filamentous fungi to mitigate CCR and maintain effective lignocellulosic degradation need to be elucidated. \n Myceliophthora thermophila , which can efficiently degrade biomass ( 30 ), has presented a prospect in the development of cell factories to produce fuels and chemicals directly from lignocellulose by CBP ( 31 ). In the previous study, we found that M. thermophila could efficiently produce malic acid directly from corncobs, with no detectable accumulation of pentose and hemicellulose during the fermentation process. This is inconsistent with the belief that pentose metabolism is inhibited by glucose during biomass hydrolysis, suggestive of the synchronous utilization of cellulose and hemicellulose in M. thermophila . The current investigation was conducted to further explore this interesting phenomenon. We revealed that in this thermophilic fungus, CCR due to extracellular glucose can be largely avoided by using cellodextrin and xylodextrin derived from the degradation of cellulose and hemicellulose via the independent induction of cellulolytic and hemicellulolytic enzymes. Cellodextrin and xylodextrin are defined as β-1,4- d -glucosyl- and xylosyl-oligosaccharides, respectively, with a polymerization degree between 2 and 7 ( 32 , 33 ). The xylodextrin transporter MtCDT-2, which can tolerate high cellobiose concentrations, facilitates the synchronous utilization of cellodextrin and xylodextrin. These findings will be beneficial to clarifying the metabolic landscape of the filamentous fungi used for degrading lignocellulose as well as for improving the selection of chassis cells for CBP and engineering various cell factories to produce natural compounds.",
"discussion": "Discussion Filamentous fungi possess the capacity of deconstructing lignocellulose by the synergistic action of large amounts of hydrolytic enzymes. Some soluble substrates released have been identified to induce the expression of lignocellulosic enzymes, but their functions differ among fungi. Cellobiose and its derivative can induce cellulase gene expression in N. crassa , T. reesei , and Aspergillus species ( 26 , 27 , 43 ), but not in Phanerochaete chrysosporium ( 44 ). Sophorose is a cellulase inducer in T. reesei ( 45 , 46 ), but it is nonfunctional in N. crassa and Aspergillus niger ( 26 , 47 ). Our transcriptome analysis indicated cellobiose can induce the expression of cellulase genes in M. thermophila (Fig. 5 A). In N. crassa , the inductive effect of cellobiose is masked by its degradation to glucose by endogenous BGLs ( 26 ), while in M. thermophila , cellobiose can directly induce the expression of majority cellulase genes (Fig. 5 A). In addition, hemicellulase gene expression is also highly induced by xylobiose and xylose (Fig. 5 A and E), but xylobiose has a stronger inductive effect (Fig. 5 E and Dataset S8 ). The deletion of bxl1 , bxl2 , or Mtcdt-2 in M. thermophila cultured on xylan can significantly increase the extracellular xylanolytic activity (Fig. 3 G), implying xylobiose or its derivatives may function as the inducer of hemicellulases, like cellobiose inducing the expression of cellulase genes. In this study, M. thermophila synchronously utilized cellobiose and xylobiose (i.e. main products of lignocellulose decomposition) in the G2X2 medium (Fig. 1 E and F), suggesting that cellobiose and xylobiose have two relatively independent metabolic pathways involving oligosaccharide transport, intracellular hydrolysis, and even the catabolism of their constitutive monosaccharides. Consistent with this suggestion, the xylodextrin transporter gene Mtcdt-2 , intracellular βBXL genes bxl1 and bxl2 , and the xylose catabolic pathway genes were consistently highly expressed in M. thermophila grown in xylobiose and G2X2 media (Fig. 4 B and D). In N. crassa , deletion of Nccdt-2 resulted in significant growth defects on cellulose and hemicellulose ( 37 ). However, deletion of Mtcdt-2 mainly affected the growth of mycelium on hemicellulose (xylan), indicating NcCDT-2 and MtCDT-2 differ regarding their functions related to lignocellulose utilization. Sugar transport assay using S. cerevisiae demonstrated the preference of MtCDT-2 to xylobiose, partly contributing to the simultaneous utilization of xylobiose and cellobiose. The interconnections among catabolic pathways of the components from plant biomass have been investigated ( 39 ). Mannodextrins can inhibit cellulase production, thereby inhibiting the growth of N. crassa on cellulose. The crosstalk of perception pathways and competitive inhibition between cellodextrins and mannodextrins have been observed both during uptake by cellodextrin transporters and within cells. This inhibition phenomenon was also detected in M. thermophila and T. ressei . However, the expression of cellulase and hemicellulase genes is independently induced in mixed G2X2 medium (Fig. 5 A and B), indicating that the intracellular coexistence of xylodextrin/cellodextrin and their degradants (xylose and glucose) does not trigger the crosstalk and CCR effect. Additionally, in this study, the presence of glucose inhibited the utilization of xylose, xylobiose, and cellobiose (Figs. 1 B and S2 ). Accordingly, the CCR effect in M. thermophila was activated following the perception of extracellular glucose, unlike in S. cerevisiae , where glucose signals that activate the main glucose repression pathway are derived from the plasma membrane receptor and the intracellular intermediates of glucose catabolism ( 8 , 48 ). In an earlier study on glucose and galactose utilization, the glucose concentration regulated galactose metabolism-related gene expression, which was almost completely suppressed by the treatment with 10 g/L glucose ( 49 ). In the present study, trace amounts of extracellular monosaccharides were detected during xylobiose and cellobiose metabolism (Fig. S7 ), suggesting that M. thermophila may maintain a relatively low extracellular glucose level, thereby eliminating the CCR effect during lignocellulose utilization. This may also explain why the intracellular xylobiose and cellobiose metabolism of M. thermophila is greater than the extracellular hydrolysis of monosaccharides. Additionally, metabolizing oligosaccharides avoids the competition for glucose with glucose-utilizing soil microbes and may be more energetically favorable because oligosaccharide transport does not consume excessive amounts of energy and enzymes do not need to be secreted ( 50 , 51 ). Intracellular oligosaccharide metabolism strategies present the advantage of M. thermophila in development of CBP technique, which can be borrowed for fermentation strain design. For example, a previous study involving S. cerevisiae confirmed the inhibition of xylose transport by glucose can be bypassed by using xylose and cellobiose as substrates ( 19 ). In terms of industrial production, the conversion of plant biomass to soluble oligosaccharides catalyzed by a hydrolytic enzyme cocktail avoids the inhibitory effects of glucose on hydrolases. Moreover, the simultaneous utilization of substrates during fermentation is ideal. Furthermore, industrial fungal strains can be rationally designed according to the cellobiose/xylobiose metabolic activities in M. thermophila to facilitate the simultaneous use of multiple carbon sources. In conclusion, the independent oligosaccharide catabolic pathway in the thermophilic fungus M. thermophila may be exploited for the simultaneous utilization of cellulose and hemicellulose in lignocellulosic biorefineries."
} | 3,423 |
36074002 | PMC9709963 | pmc | 5,566 | {
"abstract": "Abstract The sunburst anemone Anthopleura sola is an abundant species inhabiting the intertidal zone of coastal California. Historically, this species has extended from Baja California, Mexico to as far north as Monterey Bay, CA. However, recently the geographic range of this species has expanded to Bodega Bay, CA, possibly as far north as Salt Point, CA. This species also forms symbiotic partnerships with the dinoflagellate Breviolum muscatinei , a member of the family Symbiodiniaceae. These partnerships are analogous to those formed between tropical corals and dinoflagellate symbionts, making A. sola an excellent model system to explore how hosts will (co)evolve with novel symbiont populations they encounter as they expand northward. This assembly will serve as the foundation for identifying the population genomic patterns associated with range expansions, and will facilitate future work investigating how hosts and their symbiont partners will evolve to interact with one another as geographic ranges shift due to climate change.",
"introduction": "Introduction The sunburst anemone, Anthopleura sola is a large, solitary anemone inhabiting the intertidal zone of the Pacific coast from Baja California, Mexico to central California ( Fig. 1 ). Within the past half century, the geographic range of this species has expanded northward ( Denny and Gaines 2007 ), likely ending between Bodega Bay, CA and Salt Point (Mendocino County), CA (BHC, pers. obs.). Northward expansions of species historically relegated to more equatorial latitudes along the California coast have been documented during temporary periods of increased temperatures near geographic range edges ( Sanford et al. 2019 ), but have nonetheless prompted researchers to begin assessing how these populations will evolve to match novel geographic locations they encounter during longer-term range expansions. A particularly important feature of the A. sola expansion is that they are likely encountering novel symbiont populations that historically have only interacted with 2 other symbiotic species that are members of the genus Anthopleura , A. xanthogrammica and A. elegantissma whose geographic ranges extend to Alaska. Previous work has shown that these symbionts are shared between these 3 species—with the exception of the southernmost populations of A. sola and A. elegantissima , where symbionts are partitioned by host species ( Cornwell and Hernández 2021 ). As A. sola continues to move northward, interactions between newly arriving hosts and naive symbiont populations will become more common, which will allow researchers to identify patterns of molecular and physiological coevolution in both partners as geographic ranges shift with climate change. Because this symbiosis is analogous to the partnership between tropical corals and dinoflagellates, characterizing how novel symbiotic partnerships evolve along the California coast will have global implications. Fig. 1. \n Anthopleura sola polyp in sandy habitat (image credit: B. Cornwell). Well-assembled genomes are an important tool for identifying genomic patterns associated with range expansions and coevolution with symbiont partners. A. sola exhibits little population structure across its geographic range with no evidence for historical population bottlenecks, which likely contributes to the highest average level of heterozygosity of the 3 symbiotic species of Anthopleura on the Pacific coast of North America ( π = 0.0095; Cornwell and Hernández 2021 ). A draft assembly of A. sola has already been generated without long reads, resulting in a contig N50 of 5,224 bp and scaffold N50 of 16,096 with a total estimated genome size of 434 Mb ( Cornwell 2020 ). Here, we present a new assembly for A. sola which substantially improves on previously published versions, and creates a new resource for marine scientists studying how marine populations will evolve as their geographic ranges shift with warming conditions.",
"discussion": "Discussion \n A. sola is one of several species whose geographic range is shifting poleward as ocean temperatures warm, making this assembly integral to future studies aimed at detecting the genomic basis of local adaptation, identifying gene flow across the former, current, and future geographic range of A. sola , and understanding the genetic basis of partner compatibility with the endosymbiont Breviolum muscatinei (an analogous partnership to their tropical coral cousins). Recent work suggests that interactions between host and symbiont genomes play a role in structuring some genetic loci between the 2 partners ( Cornwell and Hernández 2021 ), and this improved assembly will allow for a much higher level resolution—both in terms of marker density and location in the genome—of those genetic interactions. Within the genus, long-read assemblies using Oxford Nanopore chemistry yielded an assembly of 243 Mb in 5,359 contigs for the sister species A. elegantissima ( Dimond et al. 2021 ), which largely agrees with the estimated size of this assembly—ca. 289Mb. Both of these assemblies are much smaller than the previous estimate for A. sola of 434 Mb ( Cornwell 2020 ), suggesting that assemblies using only short reads may not properly resolve repetitive regions or could erroneously divide contigs or scaffolds that represent the same location in the genome. One reason for this might be the high levels of heterozygosity that are characteristic not just of Anthopleura spp., but of many marine invertebrates, which highlights the value long reads in constructing genome assemblies for highly heterozygous organisms that inhabit ocean environments. Recent assemblies for other cnidarian species ( Acropora ) recover 14 chromosome-level scaffolds with a larger estimated genome size of 450 to 475 Mb ( Fuller et al. 2020 ; L ópez-Nandam et al. 2021). This assembly is contained within 270 scaffolds, but no clear threshold in the scaffold sizes of this assembly suggest a chromosome number (the amount each scaffold adds to the overall assembly size appears to reach an inflection point at N = 28 but this is far from definitive)."
} | 1,535 |
34390514 | PMC9291133 | pmc | 5,568 | {
"abstract": "Abstract Bacteria often cooperate by secreting molecules that can be shared as public goods between cells. Because the production of public goods is subject to cheating by mutants that exploit the good without contributing to it, there has been great interest in elucidating the evolutionary forces that maintain cooperation. However, little is known about how bacterial cooperation evolves under conditions where cheating is unlikely to be of importance. Here we use experimental evolution to follow changes in the production of a model public good, the iron‐scavenging siderophore pyoverdine, of the bacterium Pseudomonas aeruginosa . After 1200 generations of evolution in nine different environments, we observed that cheaters only reached high frequency in liquid medium with low iron availability. Conversely, when adding iron to reduce the cost of producing pyoverdine, we observed selection for pyoverdine hyperproducers. Similarly, hyperproducers also spread in populations evolved in highly viscous media, where relatedness between interacting individuals is increased. Whole‐genome sequencing of evolved clones revealed that hyperproduction is associated with mutations involving genes encoding quorum‐sensing communication systems, while cheater clones had mutations in the iron‐starvation sigma factor or in pyoverdine biosynthesis genes. Our findings demonstrate that bacterial social traits can evolve rapidly in divergent directions, with particularly strong selection for increased levels of cooperation occurring in environments where individual dispersal is reduced, as predicted by social evolution theory. Moreover, we establish a regulatory link between pyoverdine production and quorum‐sensing, showing that increased cooperation with respect to one trait (pyoverdine) can be associated with the loss (quorum‐sensing) of another social trait.",
"introduction": "1 INTRODUCTION Microbes are social organisms. Over the last three decades, a wealth of research has uncovered that microbial communities are shaped by complex networks of competitive and cooperative interactions (Figueiredo & Kramer, 2020 ; Ghoul & Mitri, 2016 ; Granato et al., 2019 ; Little et al., 2008 ; West et al., 2007a ). Competitive interactions may involve the secretion of toxins against competitors, hunting or competitive exclusion (Granato et al., 2019 ; Hibbing et al., 2010 ; Pérez et al., 2016 ). Examples of cooperative behaviours include mutualistic cross‐feeding, communication via signalling molecules, and sharing the benefits of secreted molecules such as proteases, siderophores and biofilm components (D’Souza et al., 2018 ; Dragoš et al., 2018 ; Kramer et al., 2020 ; Robinson et al., 2020 ; Wilder et al., 2011 ). Microbial cooperation often underlies important biological processes, including the establishment of infections (Ackermann et al., 2008 ; Granato et al., 2018 ), nutrient fixation in the rhizosphere (Denison et al., 2003 ) and mutualistic interactions with hosts (Verma & Miyashiro, 2013 ). Microbial cooperation has attracted the attention of evolutionary biologists not only because of its variety in form and function, but also because it incurs a cost for the actor while benefitting other individuals (West et al., 2007b ). Cooperation could thus select for “cheating” variants that do not cooperate themselves, but benefit from cooperative acts performed by others (Ghoul, Griffin et al., 2014 ). How then can cooperative traits be maintained on evolutionary timescales? This question has spurred an enormous body of research focusing on how microbes cope with the threat of cheating (Özkaya et al., 2017 ; Smith & Schuster, 2019 ; Strassmann & Queller, 2011 ; Travisano & Velicer, 2004 ; Wechsler et al., 2019 ). While interesting in its own right, the focus on cheating might have diverted attention from other factors that could also influence the evolution of cooperative traits (see Zhang & Rainey, 2013 for a critique). While we currently know much about the evolution of cheating resistance mechanisms and environmental factors that maintain cooperation in the face of cheating, we know little about cooperative trait evolution in environments where cheating plays no role. For example, can higher levels of cooperation be selected for, or can cooperative traits be lost for reasons other than cheating, for instance through disuse (Velicer et al., 1998 ; Zhang & Rainey, 2013 )? Here, we tackle these questions by focusing on the evolution of pyoverdine production in the bacterium Pseudomonas aeruginosa , one of the most widely studied social traits in microbes (Buckling et al., 2007 ; Dumas & Kümmerli, 2012 ; Griffin et al., 2004 ; Harrison, 2013 ; Harrison et al., 2017 ; O’Brien et al., 2017 ; Ross‐Gillespie et al., 2015 ). Upon sensing iron scarcity, cells secrete pyoverdine, a siderophore that chelates environmental or host‐bound ferric iron. Iron‐loaded pyoverdine is then imported via a specific receptor, which is followed by iron reduction, release from the siderophore and subsequent recycling of pyoverdine (Kramer et al., 2020 ; Schalk et al., 2020 ). Pyoverdine production is a cooperative trait, because pyoverdines are costly for the individual cell to produce, but once loaded with iron they become available to other cells in the local neighbourhood (Buckling et al., 2007 ; Harrison, 2013 ). It is well established that pyoverdine, as a so‐called “public good,” can select for cheating in severely iron‐limited and well‐mixed environments, where cheats can freely access secreted pyoverdines (Griffin et al., 2004 ; Kümmerli et al., 2015 ). It is less clear, however, how pyoverdine production would evolve in other environments where, for example, iron is less stringently limited, thereby reducing the advantage of cheating, and/or where environmental viscosity would limit cell mixing and thus cheating opportunities. Several scenarios can be envisaged. First, in iron‐rich environments pyoverdine production might be selectively lost because it is not required (Zhang & Rainey, 2013 ). Second, in viscous environments increased levels of pyoverdine production and thus cooperation could be favoured because viscosity reduces cell dispersal and siderophore diffusion (Julou et al., 2013 ; Kümmerli et al., 2009 ), which ensures that pyoverdine‐mediated social interactions occur more often between genetically related individuals (Dobay et al., 2014 ). Third, iron availability and environmental viscosity might interact and jointly affect the cost and benefit of pyoverdine production, and thereby select for an altered production level that matches the optimal cost‐to‐benefit ratio of the environment in which that bacteria evolved. Finally, the social trait might diversify without the production levels being affected. Indeed, a great diversity of pyoverdine variants exists among Pseudomonas strains (Butaite et al., 2017 ), but each strain produces only one type of pyoverdine, together with its cognate receptor. Mathematical models suggest that pyoverdine and receptor structure diversification could help individuals escape cheating (Lee et al., 2012 ) or gain an edge over competitors in the race for iron, even under iron‐rich conditions (Niehus et al., 2017 ). To test for these alternative evolutionary outcomes, we experimentally evolved replicated populations of the laboratory strain P . aeruginosa PAO1 for 200 days (~1200 generations) in nine different environments. We manipulated environmental conditions along two gradients, iron availability and media viscosity, with each gradient entailing three levels in a 3 × 3 full‐factorial design. We hypothesized that evolution under increased iron availabilities should result in a reduction of pyoverdine production, because less pyoverdine is required when iron is abundant. In contrast, evolution under increased environmental viscosities should result in higher pyoverdine production, due to increased relatedness between interacting individuals. After experimental evolution, we quantified changes in pyoverdine production and investigated the extent to which pyoverdine remained shareable among cells. We then assessed how phenotypic changes affect the fitness of evolved populations and clones. Finally, we sequenced the genomes of 119 evolved clones to map evolved phenotypes to genotypic changes. We expected that reduced pyoverdine production should be associated with mutations in the pyoverdine locus (synthesis and regulation), while increased pyoverdine production should be associated with mutations in loci responsible for gene expression regulation.",
"discussion": "4 DISCUSSION Because bacterial cooperative traits affect virulence, ecosystem functioning and microbiome assembly, enormous interest has focused on their evolution (Davis & Isberg, 2019 ; Ebrahimi et al., 2019 ; Magnúsdóttir et al., 2015 ). Particular attention has been paid to cheating (Harrison et al., 2017 ; Strassmann & Queller, 2011 ; Tarnita, 2017 )—the loss of cooperation through mutants that exploit the cooperative efforts of others. In contrast, cooperative trait evolution remains poorly examined in environments where cheating plays no role (but see Velicer et al., 1998 ). Here we explored how different environments shape the evolutionary trajectory of a model bacterial cooperative trait, the production of the iron‐scavenging siderophore pyoverdine by the opportunistic pathogen Pseudomonas aeruginosa . We experimentally evolved this species for ~1200 generations in nine different environments that varied in iron availability and environmental viscosity. We found that selection for reduced pyoverdine production due to cheating occurred only in one of the nine environments (low iron – low viscosity), where relatedness among individuals is low and the benefit of exploitation is high. In stark contrast, pyoverdine production either did not change or even increased over time in four environments each (Figure 2 ). Selection for increased pyoverdine production occurred predominantly in environments with either high iron availability, where the social trait is expressed at a relatively low level and thus incurs reduced costs, or with high viscosity, where interacting individuals are more closely related. Whole‐genome sequencing revealed that point mutations in genes of the pyoverdine locus, particularly the iron‐starvation sigma factor pvdS , are significantly associated with decreased pyoverdine production. Conversely, SNPs and large‐scale deletions in QS genes are associated with pyoverdine hyperproduction. Together, our results suggest that bacterial social traits can undergo rapid evolutionary change and allow bacteria to adapt to the prevailing environmental and social conditions. The evolution of pyoverdine nonproducers or reduced‐producers only consistently occurred in one environment (low iron, low viscosity). Because pyoverdine is most needed under low iron availability, our findings indicate that non‐ and reduced‐producers have a selective advantage due to cheating (Figure 2 ; Figure S1 ). This pattern is consistent with previous studies, where nonproducing cheats spread because they used pyoverdine produced by others (Ghoul, West et al., 2014 ; Harrison et al., 2017 ; Kümmerli et al., 2015 ). The mutational patterns discovered are also consistent with previous work showing that pyoverdine non‐ or reduced‐production predominantly arose by mutations in pvdS and its promotor region in nonhypermutator clones (Granato & Kümmerli, 2017 ; Kümmerli et al., 2015 ; O’Brien et al., 2019 ; Tostado‐Islas et al., 202 1). The sigma factor pvdS regulates the expression of the entire pyoverdine biosynthesis machinery (Cunliffe et al., 1995 ; Ringel & Brüser, 2018 ). Mutations in it are thus often associated with (i) reduced production of this public good, and (ii) major cost saving when the entire pyoverdine locus is shut down. In contrast to nonhypermutators, we found that mutations in pyoverdine biosynthesis genes were overrepresented in hypermutators. We hypothesize that mutations in these genes that reduce pyoverdine production costs are rare, and thus only surface in clones with elevated mutation rates. Our data further suggest that pyoverdine producers can adapt to cheats by reducing the compatibility of their pyoverdines (Figure 5 ). This phenomenon predominantly surfaced in the low‐iron/low‐viscosity environment, where nonproducer prevalence was highest. Specifically, we observed that pyoverdines of evolved producers stimulated the growth of an engineered nonproducer less than the pyoverdine of the ancestral wildtype. However, our genomic analysis revealed that this change in compatibility is not due to modification of the pyoverdine backbone structure, a commonly proposed (yet unproven) adaptative response to cheating (Inglis et al., 2016 ; Lee et al., 2012 ; Smith et al., 2005 ; Stilwell et al., 2018 ). Instead, we observed that reduced pyoverdine‐mediated growth benefits for nonproducers were associated with mutations in the qsrO‐vqsM‐PA2228 operon and its upstream region (Figure 6f ). The exact function of this operon is unknown, but deletions in it can lead to a complete shutdown of all three QS systems of P . aeruginosa (Köhler et al., 2014 ; Liang et al., 2014 ). Given that vqsM is a putative transcriptional regulator (Huang et al., 2019 ), we believe that the change in pyoverdine compatibility is not a direct consequence of QS silencing, but rather a pleiotropic effect of an as yet to be described regulatory link of this operon and pyoverdine. We speculate that this operon is involved in post‐synthetic pyoverdine modifications, affecting the compatibility of this molecule among different members of the community. The main finding of our experiments is the evolution of pyoverdine hyperproduction in multiple environments. It highlights that increased levels of cooperation can evolve across a range of conditions. In line with our hypothesis, we found that pyoverdine production increased in high‐viscosity environments regardless of iron availability. High viscosity reduces individual dispersal and molecule diffusion (Kümmerli et al., 2009 ). Both effects increase the relatedness among interacting individuals, as clonemates are more likely to stay together and social interactions take place across a smaller physical range (i.e., in smaller groups) (Julou et al., 2013 ; Weigert & Kümmerli, 2017 ). Limited dispersal and small group size are known to favour cooperation (Dobay et al., 2014 ; El Mouden & Gardner, 2008 ; Hamilton, 1964 ; Queller, 1994 ). While previous work showed that high relatedness can maintain cooperation when cooperators and genetically engineered cheaters are mixed in short‐term experiments (Bastiaans et al., 2016 ; Gilbert et al., 2007 ; Griffin et al., 2004 ; Kümmerli et al., 2009 ), we here demonstrate that it can actually favour increased levels of cooperation during more long‐term experiments. We further observed that hyperproducers also emerged and spread in low‐viscosity environments with intermediate and high levels of iron. This finding opposes our original hypothesis that increased iron availability should select for reduced pyoverdine production, because fewer molecules are necessary to scavenge iron when it is abundant. An alternative explanation for our findings is that increased pyoverdine production could be sustainable in these environments, because the cost‐to‐benefit ratio of this trait is altered. Costs might be relatively low because the pyoverdine production level is still considerably lower compared to what strains produce in the low‐iron/low‐viscosity environment (Figure 1 ). Benefits might be relatively high because pyoverdine is still required to scavenge iron bound to the bipyridyl chelator. Overall, our results suggest that increased resource availability may favour the evolution of higher cooperative efforts because cooperation becomes cheaper while its benefits persist (Brockhurst et al., 2008 ). We found that mutations in QS regulatory genes ( lasR , rhlR and pqsR ) and full deletions of the Las‐regulon were associated with an increase in pyoverdine production, suggesting a link between pyoverdine and QS regulons. We propose two non‐mutually exclusive reasons to explain why these mutations were favoured in our experiment. First, mutations in QS loci may be advantageous per se, and pyoverdine hyperproduction may simply be a by‐product of these mutations. In our growth medium, QS‐controlled traits such as proteases or biosurfactants are not needed, so abolishing QS could save substantial metabolic costs (Özkaya et al., 2018 ). In support of this by‐product hypothesis, we found that QS‐mutants occurred in all environments (Figure 6d ), suggesting that it is generally favourable to lose QS. Second, mutations in QS genes may be advantageous because they cause pyoverdine hyperproduction. Our results also support this hypothesis, because pyoverdine hyperproducers did not spread in the low‐iron/low‐viscosity environment, where cheaters dominated. Instead, they reached high frequencies in high‐viscosity environments, where cooperation is predicted to be favourable. Thus, mutations in the QS‐regulon may accomplish two goals at once: silencing of a superfluous regulon, and the possibility to increase pyoverdine cooperation. The evolutionary trajectories and associated genetic changes we describe are remarkably similar to patterns of P . aeruginosa evolution during chronic infections in humans (Marvig et al., 2015 ; Winstanley et al., 2016 ). For example, longitudinal studies on cystic fibrosis patients with P . aeruginosa lung infections revealed that social traits are often under selection (Andersen et al., 2015 ; Jiricny et al., 2014 ). For instance, both pyoverdine hyper‐ and nonproducers evolve in human patients in sequential steps (Andersen et al., 2018 ). Furthermore, the widespread accumulation of QS mutants that we observed in the laboratory has parallels in human lungs (Bjarnsholt et al., 2010 ; Wilder et al., 2009 ) and in animal infection models (Granato et al., 2018 ; Jansen et al., 2015 ). Moreover, hypermutators and isolates that over‐express efflux pumps are also commonly observed in clinical isolates (Mena et al., 2008 ; Rees et al., 2019 ; Sobel et al., 2005 ), where they seem to contribute to antibiotic resistance. Consistently, we found hypermutators in almost all our experimental treatments, and the widespread spreading of efflux pump mutants, probably in response to bipyridyl toxicity. Together, these parallels suggest that in vitro studies, despite their obvious limitations, may help understand general evolutionary trajectories taken by pathogens. In conclusion, we found that P . aeruginosa can quickly evolve alternative social phenotypes to match prevailing abiotic (iron availability) and biotic (relatedness) conditions. Cheating, which was the focus of many previous studies, seems to be only favoured in one of the environments studied here. Instead, our data suggest that P . aeruginosa adapts its pyoverdine production profile to match environmental requirements, often by up‐regulating pyoverdine production, but never by losing the trait altogether. We further show that social traits should not be studied in isolation, as they are connected through an intricate regulatory network (Balasubramanian et al., 2013 ). Selection for changes in one trait, such as the loss of QS, affect other traits, such as increased pyoverdine production. An integrative approach that considers the multifaceted social profiles of bacteria is needed to fully understand the evolution of sociality in these microbes."
} | 4,945 |
21475736 | PMC3058493 | pmc | 5,570 | {
"abstract": "Foragers can improve search efficiency, and ultimately fitness, by using social information: cues and signals produced by other animals that indicate food location or quality. Social information use has been well studied in predator–prey systems, but its functioning within a trophic level remains poorly understood. Eavesdropping, use of signals by unintended recipients, is of particular interest because eavesdroppers may exert selective pressure on signaling systems. We provide the most complete study to date of eavesdropping between two competing social insect species by determining the glandular source and composition of a recruitment pheromone, and by examining reciprocal heterospecific responses to this signal. We tested eavesdropping between Trigona hyalinata and Trigona spinipes , two stingless bee species that compete for floral resources, exhibit a clear dominance hierarchy and recruit nestmates to high-quality food sources via pheromone trails. Gas chromatography–mass spectrometry of T. hyalinata recruitment pheromone revealed six carboxylic esters, the most common of which is octyl octanoate, the major component of T. spinipes recruitment pheromone. We demonstrate heterospecific detection of recruitment pheromones, which can influence heterospecific and conspecific scout orientation. Unexpectedly, the dominant T. hyalinata avoided T. spinipes pheromone in preference tests, while the subordinate T. spinipes showed neither attraction to nor avoidance of T. hyalinata pheromone. We suggest that stingless bees may seek to avoid conflict through their eavesdropping behavior, incorporating expected costs associated with a choice into the decision-making process. Electronic supplementary material The online version of this article (doi:10.1007/s00265-010-1080-3) contains supplementary material, which is available to authorized users.",
"introduction": "Introduction Animals at multiple trophic levels actively search for patchily distributed food such as mobile prey, flowering or fruiting trees, or carrion. Such consumers can improve search efficiency, and ultimately fitness, by using information provided by the food itself or other organisms in the vicinity (Giraldeau 1997 ; Dornhaus and Chittka 2004 ). Use of social information ( sensu Danchin et al. 2004 ) by foragers appears to be widespread. Information can come from signals (features or behaviors that have evolved to alter the behavior of the receiver in a specific way) or cues, which did not evolve because of such effects (Maynard Smith and Harper 2003 ). When signals provide such information, unintended receivers that use it are exhibiting “interceptive eavesdropping.” Because signals evolve through selection for information flow, they are vulnerable to selective pressures exerted by eavesdroppers (Peake 2005 ). Evolutionary and ecological effects of eavesdropping may be particularly strong and diverse in the context of foraging because resultant increases in food discovery efficiency cascade through food webs (Kean et al. 2003 ). Many examples show predators and prey benefiting from social information to locate prey and avoid predators, respectively (e.g., sources in Stowe et al. 1995 ; Peake 2005 ; Seppänen et al. 2007 ; Valone 2007 ). However, social information can also improve search efficiency within a trophic level. In this latter context, heterospecific eavesdropping (on signals) and “spying” (using social information provided by cues; Wisenden and Stacey 2005 ) can affect community structure. Such strategies can (1) increase the frequency of interaction among competitors (Seppänen et al. 2007 ) or (2) drive the formation and maintenance of foraging groups (Goodale et al. 2010 ) that provide benefits (e.g., protection) that overcome costs of food sharing (Stevens and Gilby 2004 ). Despite the ecological implications of eavesdropping, little is known about how dominant and subordinate species competing for food use social information. To date, only a handful of studies have investigated interceptive eavesdropping on food location or quality signals by heterospecifics. Exploitation of heterospecific food location cues has also received some attention, primarily with social insects. Experiments suggest that subordinate species can avoid competitors (e.g., Pimm et al. 1985 ; Fletcher 2008 ; Evans et al. 2009 ; Slaa and Hughes 2009 ) or depleted resources (e.g., Nakashima et al. 2002 ; Yokoi et al. 2007 ) by using heterospecific visitation signals and cues. This latter phenomenon may be quite sophisticated; bumble bees can learn to be attracted or repelled by social information depending on their past experience with the food source, use it to determine when flowers replenish their nectar, and increase rejection of visited flowers when the visitor was an aggressive species. Avoidance of depleted resources through detection of chemical or visual cues likely saves bees time (Goulson 2009 ), thereby increasing foraging efficiency. Dominant species, on the other hand, are sometimes attracted to food cues and signals of heterospecific competitors (reviewed in Slaa and Hughes 2009 ; Goodale et al. 2010 ). This diversity of responses suggests the rules governing eavesdropping and spying within a trophic level are more complex than in predator–prey situations and require more sophisticated decision making (Coolen et al. 2003 ). To understand these rules, and the effects of social information on communities and signaling systems, we must further investigate within-trophic-level eavesdropping and spying across a range of species. Highly social insects are an excellent system for studying eavesdropping and spying by competitors. Species often compete for food with sympatric relatives (Hubbell and Johnson 1977 ; Hölldobler and Wilson 1990 ; Eltz et al. 2002 ) and exhibit clear dominant–subordinate relationships (e.g., Fellers 1987 ; Lichtenberg et al. 2010 ). Social insects combine excellent associative learning (Dukas 2008 ) with powerful olfactory detection (Greenfield 2002 ) for successful foraging. Using social information may thus provide fitness benefits through improved search efficiency. Both ants and several genera of eusocial stingless bees (Apidae, Meliponini) recruit nestmates by depositing attractive pheromones at a high-quality food source or as a trail between food and nest (Hölldobler and Wilson 1990 ; Nieh 2004 ). Because recruitment pheromones are signals present in the public domain, they are susceptible to eavesdropping. Stingless bee trails may be particularly at risk because the trails are much less heavily guarded than are ant foraging trails (EML, personal observation). Where ant eavesdropping has been observed, it is typically between parabiotic “garden ant” species that share nests in epiphytes and thus are highly likely to encounter each others’ trails (Slaa and Hughes 2009 ). Research shows that both intra- (Boogert et al. 2006 ; Jarau 2009 ) and interspecific (Nieh et al. 2004a ) eavesdropping occurs in stingless bees. Coupled with patterns of response to the visual presence of heterospecifics on flowers (Slaa et al. 2003 ) and anecdotal observations (Kerr et al. 1963 ; Johnson and Hubbell 1975 ; Johnson 1983 ), these studies suggest that stingless bees actively avoid food sources of decreased resource quality or to which they will have limited access. In particular, stingless bees appear to avoid resources occupied by dominant species, thereby steering clear of conflict (the dominance motivation hypothesis). Dominant species may also benefit by following subordinates’ pheromone trails, using this social information to discover high-quality food sources that they can take over. Because stingless bees serve as major pollinators of tropical plants (Endress 1994 ), eavesdropping interactions between sympatric colonies may significantly affect bees’ foraging patterns and, ultimately, plant gene flow. Here, we tested olfactory eavesdropping between two trail-laying stingless bee species that have a clear dominance relationship. Trigona hyalinata and Trigona spinipes overlap in distribution (Camargo and Pedro 2007 ), exhibit similar floral utilization (Lichtenberg et al. 2010 ; unpublished data), and likely compete for resources. Both species use odor trails to recruit large numbers of nestmates to rich resources such as mass-flowering trees and sucrose feeders (Nieh et al. 2003 ; Nieh et al. 2004b ). Trigona hyalinata foragers easily displace T. spinipes from food sources (Lichtenberg et al. 2010 ), both by arriving at a food source en masse and by attacking individual T. spinipes foragers (see Supplemental movies). In T. spinipes and all other trail-laying stingless bee species studied to date, recruitment pheromones come from the cephalic labial glands (Jarau 2009 ). An eavesdropper must both detect the target pheromone and distinguish it from its own. To show that eavesdropping is possible between these two species, we determined the chemical composition and attractiveness to nestmates of T. hyalinata labial gland secretions. The pheromone of T. spinipes is already known to have one main component: octyl octanoate (Schorkopf et al. 2007 ). We then tested eavesdropping between these species with preference tests. Under the dominance motivation hypothesis, we predicted that the subordinate T. spinipes (Lichtenberg et al. 2010 ) would avoid T. hyalinata recruitment pheromone, while the dominant T. hyalinata would be attracted to T. spinipes pheromone.",
"discussion": "Discussion Recent evidence suggests that social information use by foragers is widespread, but our understanding of how animals use such information remains limited. Most examples of interceptive eavesdropping and “spying” occur between trophic levels or emphasize copying behavior (reviewed in Dall et al. 2005 ; Peake 2005 ; Seppänen et al. 2007 ; Valone 2007 ; Goodale et al. 2010 ). Our results provide the most complete example to date of interceptive eavesdropping by competing social insects: we determined the composition and probable glandular source of the chemical signal, and examined reciprocal heterospecific responses to this signal in preference tests with multiple colonies. We show heterospecific avoidance eavesdropping by a stingless bee: T. hyalinata avoids the recruitment pheromone of T. spinipes . Among pollinating social bees, this within-trophic-level social information can help foragers avoid unprofitable resources or conflict (e.g., Stout et al. 1998 ; Nieh et al. 2004a ). Chemical analysis of T. hyalinata and T. spinipes LGE demonstrates that the pheromones should be (1) detectable by both species because both contain octyl octanoate in relatively high concentrations and (2) differentiable because T. spinipes recruitment pheromone consists of one major component while T. hyalinata has six (Fig. 2 ). One of these, hexyl octanoate, is reported for the first time as a component of stingless bee recruitment pheromones. Unlike the pattern reported for other social insects (Slaa and Hughes 2009 ), eavesdropping responses did not depend on relative dominance of the eavesdropping and signaling species. In our study, the dominant species, T. hyalinata , avoided the recruitment pheromone of the subordinate species. Trigona spinipes showed no attraction to or avoidance of the dominant species’ pheromone. Under the dominance motivation hypothesis, if eavesdropping decisions were based solely on relative dominance, we would have seen attraction by T. hyalinata foragers and avoidance by T. spinipes foragers to heterospecific recruitment pheromone. Recruitment pheromones Four trail-laying stingless bee species (in three genera), in addition to our study species, are attracted to labial gland secretions (Jarau 2009 ; Stangler et al. 2009 ). These results, taken with the identical chemistry of T. spinipes LGE and odor marks (Schorkopf et al. 2007 ), indicate that stingless bees’ recruitment pheromones are secreted by the labial glands. Our finding that T. hyalinata foragers are strongly attracted to nestmate LGE, and chemical similarity with congeners’ LGEs, strongly suggests that recruitment pheromone comes from the labial glands in this species as well. \n Trigona hyalinata recruitment pheromone composition is consistent with recruitment pheromones of congeners (Jarau 2009 ) and other odor-marking stingless bees (Stangler et al. 2009 ), which contain carboxylic and terpene esters. Octyl hexanoate, octyl octanoate, hexyl decanoate, octyl decanoate, and decyl hexanoate are shared with other species (Jarau 2009 ; Jarau et al. 2010 ), while hexyl octanoate is reported for the first time as components of stingless bee recruitment pheromones. Behavior of other Trigona species suggests that foragers require the entire blend of chemicals to exhibit natural trail-following behavior (Jarau 2009 ). Interestingly, the esters found in T. hyalinata recruitment pheromone are also found in other glandular extracts thought to have an attractive function. These include the Dufour’s gland in Andrena (Fernandes et al. 1981 ; Hefetz 1987 ), Dufourea (Wheeler et al. 1985 ), and Svastra (Duffield et al. 1984 ) bee species, and mandibular and preputial glands that likely produce sex pheromones in several Myrmecocystus ant species (Lloyd et al. 1989 ) and the Brandt’s vole (Zhang et al. 2007 ), respectively. In intraspecific preference tests, T. spinipes showed weaker preference (65%) than did T. hyalinata (80%). Trigona spinipes preferences were also weaker than in other preference experiments (73%, Nieh et al. 2004b ; 90%, Schorkopf et al. 2007 ), while T. hyalinata shows greater consistency across studies (81%, Nieh et al. 2003 ). Three major differences in life history traits between these species could be related to this, although the last two seem less likely. First, T. spinipes appear to be highly generalist in their floral visitation, visiting 51% of 562 plant species for which we have collated stingless bee resource use data. Trigona hyalinata , however, were found on only 5% of the plants and seem to specialize on dense floral patches: trees and shrubs (unpublished data). The relationship between floral preference patterns and reliance on social information is not clear and bears further investigation. Second, at the species level T. hyalinata are more dominant and aggressive than T. spinipes (Lichtenberg et al. 2010 ). Third, T. hyalinata colonies have almost three times the number of workers as T. spinipes colonies have (D.W. Roubik, personal communication; Wille and Michener 1973 ). This is likely related to the dominant status of T. hyalinata ; species with larger colonies tend to be more dominant (Lichtenberg et al. 2010 ). It is unlikely that species’ dominance affects degree of reliance on information provided by nestmates, since subordinate Melipona species show strong attraction to nestmates’ chemical “footprints” (e.g., 91%; Nieh 1998 ). Most previous research on stingless bee recruitment pheromones utilized artificial trails rather than presenting odors at a food source (the feeder), as we did. This raises the possibility that pheromones deposited at the food source differ from those used in odor trails. However, no evidence indicates that stingless bees have two separate recruitment pheromones. Our protocol used the same pheromone-producing gland as artificial trail studies. In one case, chemically analyzed odor marks were collected from the feeder (Schorkopf et al. 2007 ). In addition, T . spinipes and T. hyalinata create polarized odor trails, depositing the majority of their pheromone within 1 m of the feeder (Nieh et al. 2004b ). Because the trail is an extension of odor marks at the food source, it is reasonable to assume that pheromone deposited on the feeder is the same as trail pheromone, and that the bees obtain the same meaning from our experimental setup and odor trails. The pheromone concentration that we used, 0.1 bee equivalents, elicited a natural response in a congener, Trigona recursa (Jarau et al. 2004 ), and in our experiments. Eavesdropping responses of T. hyalinata and T. spinipes were the same whether the treatment was fresh pheromone that had elicited strong natural recruitment or LGE ( T. hyalinata , 29% vs. 34% of bees choosing the feeder with pheromone; T. spinipes , 52% vs. 51%). Each species also showed a highly significant ( p < 0.0001) preference for nestmate recruitment pheromone at 0.1 bee equivalents (Fig. 3 ; T. hyalinata , 80% choosing the feeder with pheromone; T. spinipes , 65%). The strength of these preferences was similar to those shown by each species in preference tests that employed odor trails (Nieh et al. 2003 ; Nieh et al. 2004b ). Greater or lesser amounts of LGE may elicit different eavesdropping responses than those reported here. Predatory ants eavesdropping on fig volatiles exhibit such a dose-dependent response, showing greater attraction to larger quantities of figs (Ranganathan and Borges 2009 ). However, our results show that 0.1 bee equivalents are sufficient to elicit both conspecific and heterospecific responses that are the same as responses to natural-deposited pheromone at approximately the same concentration. Interspecific eavesdropping The limited amount of natural habitat near the laboratory and Trigona preferences for nesting high in trees limited the number of colonies that we were able to work with. Despite this, we feel that our results reflect species-typical behaviors. Results were highly consistent across replicate colonies (Table 4 ), and each colony showed the same response to pheromone from two different heterospecific colonies (A3). This consistency across colonies and months also indicates that a species’ eavesdropping behavior does not vary much, if at all, with the current food needs of the colony. \n Trigona spinipes foragers clearly could detect the presence of heterospecific pheromone. Preference for nestmate pheromone decreased significantly, albeit slightly (Fig. 5 ; going from 65% to 59% choosing the T. spinipes LGE; p = 0.002), when bees chose between conspecific and heterospecific pheromones. Thus, T. spinipes does recognize T. hyalinata pheromone as different. Despite this detection ability, T. spinipes foragers exhibited a behavioral lack of choosiness in eavesdropping tests, showing no preference between feeders with no odor and T. hyalinata pheromone (Fig. 4 ). A similar failure to use social information has been found for three-spined sticklebacks, although the related nine-spined sticklebacks use similar social information (Coolen et al. 2003 ). Trigona spinipes are attracted to footprint cues of the subordinate Melipona rufiventris at certain locations (Nieh et al. 2004a ), suggesting they have a species- and context-specific response to social information, and heterospecific signals and cues do not always alter their movements. Bumble bees also facultatively use social information, exhibiting visual local enhancement only when approaching unfamiliar flower types (Kawaguchi et al. 2007 ). Alternately, T. spinipes may ignore T. hyalinata pheromones when they cannot also see foragers on the marked food source. Apis mellifera ignore olfactory information when sufficient visual information is available (Giurfa et al. 1994 ). Contrary to our expectation, T. hyalinata foragers showed strong avoidance of the subordinate species’ recruitment pheromone (Fig. 4 ). This result was surprising, given previous patterns of social information use by social insects (Slaa and Hughes 2009 ) and the highly dominant behavior exhibited by T. hyalinata (Lichtenberg et al. 2010 ). It is unlikely that our results reflect avoidance of all non-nestmate recruitment pheromones by T. hyalinata foragers. Individual T. hyalinata foragers will fly to and attack other species at food sources (Lichtenberg et al. 2010 ). Given that dominant stingless bee species such as T. hyalinata appear to be relatively poor at discovering new food sources (the dominance-discovery trade-off, Fellers 1987 ; Nagamitsu and Inoue 1997 ), avoiding all non-nestmate odors would severely limit food intake by T. hyalinata colonies. Our findings are consistent with a hypothesis of conflict avoidance through eavesdropping decisions. Attacking to gain control of an occupied resource can inflict mortality losses even for highly dominant species (Johnson and Hubbell 1974 ; Nieh et al. 2005 ). The recruitment pheromone we presented to eavesdropping T. hyalinata in trials with fresh odor marks and LGE corresponds to the presence of numerous subordinate foragers. In our fresh odor mark trials, we collected marks once a “pulse” (Nieh et al. 2004b ) of at least 30 bees arrived at the feeder. Stingless bee trails must be actively maintained by bees interrupting their food collection, and begin to fade out after approximately 20 min without such maintenance. Thus, recruitment pheromones provide current information on both resource availability and abundance of bees already present at the resource. Under such conditions, attack may be costly for a T. hyalinata colony because it may require the participation of hundreds of bees, which could otherwise be recruited to non-contested food sources. Trigona hyalinata foragers’ decisions to not choose a resource at which social information predicts high numbers of subordinate heterospecifics may be similar to the failure of dominant Trigona silvestriana to drive away large numbers of subordinate bees (Johnson and Hubbell 1974 ). Indeed, while T. hyalinata can easily displace a group of foraging T. spinipes , they do not attempt to do so every time they encounter the subordinate species (personal observation). This behavior merits further study. One possible explanation is that social insect eavesdropping decisions include expected costs associated with each choice; research that we are currently conducting investigates this. Different lines of evidence suggest that eavesdropping on signals and spying on cues affect the movements of social bees. First, we have observed T. hyalinata depositing odor marks on flowers. Second, feeders are discovered more quickly by other stingless bee species when they bear recruitment pheromone or a large quantity of footprints (Johnson 1983 ). Finally, interspecific interactions increase between-plant movement of honey bees (Greenleaf and Kremen 2006 ) and bumble bees (Kawaguchi et al. 2007 ), and may do the same for eavesdropping stingless bees. Our results indicate that social information use by competitors is governed by complex rules. Potentially large ecological and evolutionary impacts make this an important area for future investigation."
} | 5,735 |
36772225 | PMC9919609 | pmc | 5,571 | {
"abstract": "Tiny machine learning (TinyML) has become an emerging field according to the rapid growth in the area of the internet of things (IoT). However, most deep learning algorithms are too complex, require a lot of memory to store data, and consume an enormous amount of energy for calculation/data movement; therefore, the algorithms are not suitable for IoT devices such as various sensors and imaging systems. Furthermore, typical hardware accelerators cannot be embedded in these resource-constrained edge devices, and they are difficult to drive real-time inference processing as well. To perform the real-time processing on these battery-operated devices, deep learning models should be compact and hardware-optimized, and hardware accelerator designs also have to be lightweight and consume extremely low energy. Therefore, we present an optimized network model through model simplification and compression for the hardware to be implemented, and propose a hardware architecture for a lightweight and energy-efficient deep learning accelerator. The experimental results demonstrate that our optimized model successfully performs object detection, and the proposed hardware design achieves 1.25× and 4.27× smaller logic and BRAM size, respectively, and its energy consumption is approximately 10.37× lower than previous similar works with 43.95 fps as a real-time process under an operating frequency of 100 MHz on a Xilinx ZC702 FPGA.",
"conclusion": "6. Conclusions With the development of deep learning technology and rapid growth in the area of IoT, deploying deep learning models on IoT devices has become an emerging field with TinyML. However, most deep learning algorithms are too complex to be executed on these resource-constrained edge devices. The algorithms are not suitable for battery-operated IoT devices owing to their high computation and energy consumption requirements. Therefore, a lightweight deep learning model and its well-optimized hardware through HW/SW co-optimization are required; hence, we proposed an optimized model for the hardware to be implemented and a lightweight and energy-efficient hardware architecture in this paper. We presented an optimized model based on a backbone network by employing model simplification and compression. Model reduction techniques, such as involving maxpool operations in convolution, alleviates the computation complexity and latency, and hardware-oriented parameter simplification enables software and hardware to be co-optimized. Additionally, data quantization, as a model compression technique, was performed to reduce the storage space requirement for the parameters. Experiment results showed that the optimized model successfully performed object detection and was subjected to both qualitative and quantitative evaluations. Furthermore, a lightweight and energy-efficient hardware architecture was proposed with a 3D tensor-like PE structure, generation of input feature map matrix, and a memory-efficient on-chip memory management strategy. The 3D tensor-like PE structure deals with several input feature map matrices at the same time, leading to low latency and eventually low energy consumption. In addition, logic size can be reduced owing to parallel processing between layers and channels through combination of the 3D tensor-like PE structure and input feature map matrix generation. Besides, the on-chip memory management strategy enables incremental access to an external memory without frequent and irregular data read/write operations, leading to the use of a few and small on-chip memories, and also low power consumption owing to the simple access to the memories. As a result, the proposed hardware design achieves 1.25× and 4.27× smaller logic and BRAM sizes, respectively, and consumes approximately 10.37× less energy than those of previous similar works with 43.95 fps as a real-time process under an operating frequency of 100 MHz on a Xilinx ZC702 FPGA. It indicates that the proposed hardware is capable of being embedded in the resource-constrained edge devices and can be applied to the battery-operated IoT devices.",
"introduction": "1. Introduction Deep learning has been popular because of the availability of computing power and the development of big data [ 1 ], and various reviews and discussions on deep learning have been extensively conducted in recent years [ 2 , 3 , 4 , 5 , 6 ]. It has been widely applied in many fields such as image recognition [ 7 ], object detection [ 8 ], autonomous driving [ 9 , 10 ], and robotics [ 11 ]. Moreover, deep learning networks have been shown to be successful for these fields [ 12 ], and nowadays it has become important even in the field of IoT with the rapid development of IoT devices and network infrastructure [ 13 ]. Accordingly, deep learning operation in real time with low energy on these resource-constrained edge devices has emerged as essential work in the era of IoT [ 14 ]. However, deep learning models are generally too complex, and they also require considerable amounts of data and their computation [ 15 ]. Model complexity in deep learning is a fundamental issue in terms of model framework, model size, optimization process, and data complexity. Most deep learning models have a complex model framework, such as a convolutional neural network (CNN), and their model size is so huge owing to numerous parameters, layers, and filters. In addition, the configuration, such as layer width and filter size, also affects model size. As a result, running a deep learning model requires so much memory to store those numerous parameters and a tremendous amount of intermediate data. Furthermore, high energy consumption is inevitable for computing their calculation and moving so much data from/to memory. Therefore, it is very challenging to implement the hardware to accelerate these deep learning models on the battery-operated IoT devices. Thus, TinyML is an important and emerging area for operating machine learning applications on small embedded IoT devices, and hence it has been actively researched recently [ 16 , 17 , 18 , 19 , 20 ]. It aims at designing and developing algorithms and hardware capable of performing inferences on resource-constrained devices at extremely low energy. Accordingly, it takes into account the characteristics of hardware to be operated and tries to optimize the model for the hardware and reduce computational load and memory demand by deploying approximation and compression, like pruning. Moreover, the hardware design has to be implemented as lightweight to be embedded in low-cost resource-constrained devices and energy-efficient so that the deep learning model works smoothly on the battery-operated devices, and low-latency so as to run in real time on IoT edge devices, considering the fast sensory data streams. In this paper, we present the optimized network model for hardware to be implemented. The proposed optimized model is based on SqueezeNet [ 21 ], which is a mobile-oriented network. We perform model reduction and parameter simplification on the backbone model network through model simplification, and integer quantization is adopted for activation and parameters through model compression. Furthermore, a lightweight and energy-efficient hardware architecture is proposed, and an implemented design is able to perform parallel processing between layers and channels deploying a 3D tensor-like processing element (PE) structure. It results in low latency and reduction in energy consumption. Besides, a small amount of on-chip memory is required owing to the proposed on-chip memory management strategy, which makes the design lightweight and low-power. As a result, the experimental results demonstrate that our optimized model successfully performs object detection, and the proposed hardware design achieves 1.25× and 4.27× smaller logic and BRAM size, respectively, and its energy consumption is approximately 10.68× lower than previous related works with 43.95 fps as a real-time process under an operating frequency of 100 MHz on a Xilinx ZC702 FPGA. The rest of this paper is structured as follows: Section 2 provides some background of lightweight deep learning techniques with our backbone model and its related works. Section 3 presents the proposed model optimization through model simplification and compression for hardware to be implemented. In Section 4 , the proposed hardware architecture is presented for the hardware design to be lightweight and energy-efficient. Experimental results for the performance of the proposed model and hardware architecture are shown in Section 5 . Lastly, Section 6 discusses the conclusion."
} | 2,157 |
39203560 | PMC11356974 | pmc | 5,572 | {
"abstract": "(1) Background: Plant diversity has long been assumed to predict soil microbial diversity. The mutualistic symbiosis between forest trees and ectomycorrhizal (EM) fungi favors strong correlations of EM fungal diversity with host density in terrestrial ecosystems. Nevertheless, in contrast with host tree effects, neighboring plant effects are less well studied. (2) Methods: In the study presented herein, we examined the α-diversity, community composition, and co-occurrence patterns of EM fungi in Quercus acutissima across different forest types (pure forests, mixed forests with Pinus tabuliformis , and mixed forests with other broadleaved species) to ascertain how the EM fungi of focal trees are related to their neighboring plants and to identify the underlying mechanisms that contribute to this relationship. (3) Results: The EM fungal community exhibited an overall modest but positive correlation with neighboring plant richness, with the associations being more pronounced in mixed forests. This neighboring effect was mediated by altered abiotic (i.e., SOC, TN, LC, and LP) and biotic (i.e., bacterial community) factors in rhizosphere soil. Further analysis revealed that Tomentella_badia , Tomentella_galzinii , and Sebacina_incrustans exhibited the most significant correlations with plant and EM fungal diversity. These keystone taxa featured low relative abundance and clear habitat preferences and shared similar physiological traits that promote nutrient uptake through contact, short-distance and medium-distance smooth contact-based exploration types, thereby enhancing the potential correlations between EM fungi and the neighboring plant community. (4) Conclusions: Our findings contribute to the comprehension of the effect of neighboring plants on the EM fungal community of focal trees of different forest communities and the biodiversity sensitivity to environmental change.",
"conclusion": "5. Conclusions Our study corroborates significant neighboring plant diversity effects on the EM fungal diversity of focal Q. acutissima trees. In general, compared to the strong correlation between directly interacting host plant and EM fungal diversity that has been widely documented, indirect effects of neighboring trees contribute to moderate correlations to EM fungi. EM fungal interactions inferred from network topology ( R 2 = 0.27–0.58, p < 0.05) showed the strongest correlations with neighboring plant diversity, followed by α-diversity index ( R 2 = 0.20–0.51, p < 0.05), but the correlation with β diversity was weak and insignificant ( R 2 = 0.04, p = 0.11). This result indicated insignificant neighboring plant effects on EM fungal community compositional dissimilarity. Mixture forests could enhance the strength of the positive correlation. These results were consistent with the observed increasing EM fungal diversity in mixture forests with Pinus tabuliformis , of which the α-diversity index increased by 15.93–53.58%, network complexity-summarized topological parameters increased by 67.72%, while β diversity decreased by 1.2% compared with pure forests. By establishing linkages between environmental heterogeneity and plant-microbe relationships, our analysis revealed the significant roles of high assortativity but lower average path distance of the cross-kingdom networks, indicating that neighboring plants increased EM fungal diversity by increasing the complexity of the coexisting bacterial community in rhizosphere soil. Among the abiotic factors investigated, the slopes of the diversity relationship regression were found to be positively associated with total nutrients (TN and LP) but negatively associated with total C (SOC and LC). These results suggest that the neighboring plant–EM correlation may be more susceptible to environmental fluctuations in systems with high nutrient levels. Overall, our study reveals moderate neighboring plant diversity effects on the EM fungal diversity of Q. acutissima in natural secondary forests on the Loess Plateau, pointing to the possibility that this indirect effect of plant diversity may be attributed primarily to altered bacterial communities and nutrients in the rhizosphere soil.",
"introduction": "1. Introduction The relationship between plants and soil microorganisms provides insights into the ecological drivers of community structure and function. Ectomycorrhizal (EM) fungi constitute a substantial component of the plant–soil microbiome, colonizing up to 95% of the root tips of woody plants [ 1 ]. EM symbiotic mutualists benefit the host plant by enhancing nutrient and water uptake from the soil, and the hosts provide fungi with photosynthetic products [ 2 ]. Considering the biological functions of the symbiotic process, the correlations between the diversity of EM fungal species and the diversity of EM plant species have been extensively investigated [ 3 , 4 ]. However, the authors of these studies typically collected soil samples that contained mixed roots from co-occurring EM plant species, which potentially overestimates the relationships between EM diversity and host diversity by including the overall positive effect of plant species [ 5 ]. In addition, an increasing number of studies have demonstrated that taxonomically and functionally diverse EM fungal communities are associated with the root systems of individual trees [ 6 , 7 ]. Therefore, investigation of the role of individual host species on EM fungal diversity is necessary beyond EM host plant diversity. Plant species grow in aggregated clumps; a neighborhood of other potential host plants and non-host plants may also affect EM fungi in focal plants. Despite the obvious links, the role of neighboring plant communities in affecting EM fungal communities remains enigmatic. Analogous to the host effects, the diversity of neighboring plants has also been assumed to be positively related to EM fungal diversity, but no direct or negative relationships have also been reported [ 8 , 9 ]. A field investigation into the EM fungal communities of willows (Salix) in an alpine glacier forefield further indicated that the diversity of a plant community only influences the occurrence of a limited number of particular fungi [ 10 ]. The neighboring tree effect on focal EM fungal communities warrants improving our understanding of plant mixture impacts on soil processes and competition among microorganisms belowground. The neighboring effects may be attributed to alterations in abiotic and biotic properties in rhizosphere soil. A number of abiotic factors influence the EM fungal community, with the impact of soil nutrient heterogeneity being particularly noteworthy. These nutrients may be present in the soil solution as soluble inorganic or organic compounds or may be liberated from decomposing organic matter [ 11 ]. Leaf litter exhibits significant variation between forest types, and the quality and quantity of leaf litter affects soil microorganisms in both direct and indirect ways. Directly, leaf litter serves as a substrate for soil microorganisms. At the outset of the decomposition process, water-soluble compounds present in the leaf litter are leached into the soil, thereby becoming available for microorganisms [ 12 ]. Indirectly, EM fungi are capable of mobilizing nutrients from soil microsites enriched with organic or mineral residues through the secretion of hydrolytic and oxidative enzymes [ 13 ]. The alteration in nutrient cycling as a result of litter quality and quantity consequently impacts the population sizes of root-associated microbes by modifying plant–EM interactions [ 14 ]. Furthermore, the exchange of nutrients constitutes a critical mechanism through which EM fungi and plants exert reciprocal selective forces on one another. The efficacy of both partners as mutualists may vary. This asymmetry in nutrient exchange has the potential to destabilize EM fungal interactions, as selection favors individuals that provide reduced benefits and incur lower costs [ 13 ]. Nevertheless, despite the extensive research on the impact of soil and litter nutrients on the diversity of EM fungi, the underlying mechanisms by which soil nutrient heterogeneity regulates relationships between EM fungal diversity and plant diversity remain unclear. Alternatively, microbial species in natural ecosystems do not exist in isolation as individual populations; rather, they interact with one another to serve a variety of ecosystem functions [ 15 ]. Therefore, our understanding of microbial communities should extend beyond the individual/species-level characteristics of species richness and abundance to encompass the interspecific characteristics of the intricate microbial communities. EM fungi inhabit small pore spaces, where their diminutive size allows for close interactions with rhizosphere soil bacteria, thereby potentially influencing both the composition of the communities and their function [ 16 ]. Indeed, evidence suggests that the establishment of EM fungi on tree roots may be influenced by the bacterial communities present in the rhizosphere through the production of cell wall hydrolytic enzymes, phytohormones, amino acids, and/or vitamins [ 17 ]. Some of these biochemical activities may exert a direct influence on the germination and growth rate of EM fungal structures, while others may impact root development and susceptibility to infection [ 17 ]. Furthermore, it was discovered that interspecies interactions influence the habitat affinities or shared physiologies of microbial community members [ 18 ]. These studies have yielded significant insights, indicating that biotic interactions play a pivotal role in shaping microbial communities. Nevertheless, our understanding of the relative importance of environmental heterogeneity and cross-kingdom co-occurrence patterns in the diversity of forest EM fungi and of how bacterial communities regulate the plant–EM fungal diversity relationship remains limited. This represents a key knowledge gap to elucidate the linkages between aboveground and belowground biodiversity, particularly in ecologically fragile areas. The present study aims to determine how the diversity patterns (abundance, α-diversity index, and compositional dissimilarities) and co-occurrence patterns of EM fungi symbiosed with Quercus acutissima are related to the diversity of neighboring plants in three different forest types, i.e., pure forests of Q. acutissima (PF), mixed forests of Q. acutissima and Pinus tabuliformis (QPF), and mixed forests of Q. acutissima and other broadleaved species (QBF). The potential mechanisms of neighboring plant diversity effects on EM fungal communities were also determined by incorporating the physiochemical properties (i.e., soil pH, moisture, nutrients, and litter nutrients) and biotic factors (i.e., soil bacterial and cross-kingdom species associations) of rhizosphere soil into the linear mixed-effects model of plant–EM fungal regressions. We sought to address the following questions: (1) how do the EM fungal diversity and co-occurrence patterns of Q. acutissima vary across different forest types? Considering the higher resource heterogeneity observed in mixed forests, it is reasonable to hypothesize that EM fungal diversity will be higher in mixed forests than in pure forests. (2) What is the relationship between the EM fungal diversity of focal trees and the diversity of neighboring plants across different forest types? Considering the functional role of soil microbes as the primary decomposers of plant-derived substrates, it is reasonable to hypothesize a positive relationship between them. (3) What environmental factors are responsible for neighboring plant diversity effects on EM fungal communities? The hypothesis is that bacterial community exerts a greater influence on neighboring plant–EM fungal relationships than soil abiotic factors.",
"discussion": "4. Discussion 4.1. Effects of Forest Types on EM Fungal Diversity Two antithetic processes, turnover resulting from species replacement and nestedness resulting from species loss, have been demonstrated to drive the responses of microbial community composition to environmental changes [ 26 ]. The present study examined turnover and nestedness patterns in the species β diversity of EM fungal communities symbioses with Q. acutissima in different forest types and found that the dissimilarity pattern observed in the EM fungal community was predominantly shaped by species turnover ( Table S1 ) caused by spatial constraints or environmental sorting. Similarly, Wu et al. [ 27 ] reported significant species turnover patterns for bacteria in lake ecosystems, whereas those of diatoms and chironomids exhibited significant nestedness. The disparate results may be attributed to the distinctive attributes of the investigated taxonomic groups, such as dispersal ability, trophic position, and body size [ 27 , 28 ]. The high spatial turnover of EM fungi also indicates the potential for evolutionary adaptation to environmental conditions across forest types [ 29 ]. It is noteworthy that although the contribution of species nestedness in our study was comparable to that found in other regions of the world (usually <20%), the relative importance of the nestedness component was higher in the QPF (17.04%) and QBF (21.89%) than in the PF (13.73%). This may be partially explained by the presence of EM fungal specialists in mixtures of coniferous and other broadleaved tree species, whose loss might contribute to a significant loss of community functions along the environmental heterogeneity. In light of these findings, it can be posited that the loss or gain of species across forest types should also be considered as a significant factor influencing the dynamics of the EM fungal community. The occurrence patterns of EM fungal communities are also influenced by the forest types. In comparison to pure forests, networks in pine mixture forests exhibited markedly elevated values of betweenness centrality and edge and node numbers in subnetworks and a significantly enhanced multi-complexity index ( p < 0.05; Figure 1 d), which is consistent with hypothesis 1. This finding aligns with prior research indicating that soil fungal networks in mixed forests exhibit elevated topological properties [ 30 ]. These alterations in topological parameters, such as more connections and increased centrality, could facilitate the accelerated transmission of external perturbative effects among networked species, thereby enhancing the efficiency of the system [ 31 ]. Moreover, the co-occurrence network constitutes a valuable methodology for identifying potential ecological interactions among diverse microbial taxa within the soil across various forest types [ 32 ]. The potential EM fungal networks from the PF or QBF exhibited a higher incidence of negative correlations compared to those from the QPF ( Figure 1 d), indicating increased inter-specific competition within pure forests. Such inter-specific competition could be attributed to the restricted C resources of pure forests, which are typified by sparse vegetation and low primary productivity due to water deficiency. Considering that EM fungal symbiosis is an energy-intensive process, the reduced soil C availability in pure forests may have exerted more significant impacts on potential EM fungal competition than the mixed forest soils. This inference was corroborated by the observation of lower soil water content and nutrient availability in pure forests ( Figure S1 ). Alternatively, the mixed forest provided adequate substrates and diverse habitats for the colonization of EM fungi and facilitated niche differentiation, thereby resulting in a weakening of microbial interactions within the rhizosphere soil [ 30 , 33 ]. 4.2. Effects of Neighboring Plant Community on EM Fungal Diversity Our study provides evidence for a moderate overall correlation between neighboring plant diversity and the EM fungal community of focal Q. acutissima trees with respect to α- diversity, β- diversity, and network topological characteristics ( Figure 2 and Figures S2–S4 ), which is consistent with hypothesis 2. The observed positive relationships provide empirical support for functional interdependence between plants and soil microorganisms at a global scale. An increase in plant diversity can directly affect the diversity patterns of EM fungal communities by altering the amount of photosynthetically derived C allocated to their symbiotic partners [ 34 ]. Furthermore, the diverse physiochemical attributes of plant detritus entering the belowground ecosystem could result in the formation of discrete niches for specific decomposers [ 35 ]. A plant community with higher diversity may be more productive and provide more resource input, thus serving as more bioavailable substrates (e.g., non-recalcitrant C) and leading to an increase in microbial metabolic activity and potential interactions related to resource utilization and ecological niche differentiation [ 31 ]. However, it was noteworthy that compared to the strong correlation between directly interacting host plant and EM fungal diversity, the indirect effects of the neighboring plant community, which encompasses both neighboring host trees, shrubs, and herbs, may be the most plausible explanation for the observed moderate correlation in our study. The presence of other hosts and non-host plants may hamper the development of typical EM fungal communities due to priority effects, allelopathy, and competition among fungi [ 5 , 36 ], leading to a more gradual increase in EM fungal diversity with neighboring plant richness. When forest types were considered, the EM fungal community showed disparate relationships with neighboring plant diversity ( Figure 2 ). The associations between plant richness and EM fungal diversity were more pronounced in the QPF, whereas in pure forests, plant diversity exhibited a decoupling from Simpson diversity and betweenness centrality. These findings suggest significant roles of tree species composition in the EM fungal symbionts of hosts, thereby influencing the concurrence of both plant and microbial communities. Prior research has indicated that the strength and direction of effects of tree combination on EM fungal communities are associated with plant phylogenetic relationships [ 37 ]. Trees with more phylogenetically distant neighbors tend to have higher fungal richness and community dissimilarity compared to trees surrounded by closely related host species [ 9 ]. For example, the EM community in Picea abies is more sensitive to the presence of Betula pendula neighbors than P. sylvestris neighbors [ 4 ]. Furthermore, the relatively less stressed environments in the QPF may be of particular significance in this context. Forest stands with an admixture of broad-leaved or coniferous tree species contribute to an enhanced variety of root exudates and organic materials, and an abundance of microclimates at floor level and in the soil of forest systems fosters more divergent microhabitats and suitable conditions for a greater variety of EM fungal species in the QPF [ 38 , 39 ]. It is increasingly recognized that keystone microbial communities exert significant influences on ecosystem engineering, influencing the assembly and functioning of microbial communities [ 40 ]. The Zi-Pi analysis identified eleven keystone EM fungal taxa (nine connectors and two module hubs), mainly belonging to the genera Tomentella , Sebacina , Inocybe , and Hysterangium ( Figure 5 ). When examining the relationship between keystone taxa, EM fungal community, and neighboring plant diversity, ASV552 ( Tomentella_badia ), ASV609 ( Tomentella_galzinii ), and ASV447 ( Sebacina_incrustans ) exhibited the most significant correlations. One possible explanation for this result may relate to the physiological traits of these keystone taxa, especially their exploration types, which are commonly assumed to denote the spatial foraging patterns and resource-related niches of extraradical mycelia [ 41 ]. Although not all species in a genus were recognized by exploration type in the available records, the genera Tomentella and Sebacina are generally contact, short-distance, and medium-distance smooth types [ 42 ] that have been proposed to maximize the area of hydrophilic hyphae that extend into the soil and promote the rapid uptake of mobile nutrients [ 43 ], thereby enhancing the potential correlations with neighboring plant and the EM fungal community. Alternatively, these keystone species are distinguished by their low relative abundance (0.011–0.638%) and distinctive habitat preferences. The results indicate the potential for strong correlations at finer taxonomic resolutions or for specific microbial functional groups, of which the low-abundance microorganisms may be among the key groups that play crucial roles in diversity maintenance [ 44 ]. The disproportionate influence of rare taxa is inconsistent with the expected high functional redundancy within microbial communities [ 44 ]. The significance of rare taxa may be attributed, at least in part, to the conditional rarity of these microbial communities, which comprise approximately 1.5–28% of all microbes. The conditional rarity indicates that the impact of these microorganisms is subject to external environmental influences and is not a permanent state [ 45 ]. Rather, they may serve as repositories of genetic resources that could be mobilized under appropriate conditions [ 46 ]. This is reflected in the distinct habitat preferences of these rare taxa. Further research focusing on specific EM fungal functional groups, with clearly identified host plants within the plant community, could significantly enhance the precision of estimates pertaining to plant–microbial relationships. 4.3. Environmental Variables Regulate Diversity Relationships between EM Fungi and the Neighboring Plant Community While it is plausible that soil nutrient heterogeneity facilitates greater diversity of both plant and soil microbial communities, the identified correlations between neighboring plant-EM fungal regression and soil properties in our study do not fully substantiate this perspective. Soil total N was found to be positively correlated with the plant–EM fungal Shannon diversity and Simpson diversity relationship. Similarly, a stronger EM fungal community compositional relationship was observed to be positively correlated with LP ( Table 1 ). However, soil total C was negatively related to the plant–EM fungal Shannon diversity and Simpson diversity relationship, whereas litter total C is the major factor negatively associated with the plant–EM fungal community nestedness dissimilarity correlations. The contrasting correlations of plant–EM fungi relationships to substrate C and nutrient availability may be associated with symbiotic processes between plants and EM fungi [ 47 ]. Plants typically allocate 20–40% of their C to EM fungi, which in turn forage for soil nutrients that are not easily accessible to plants [ 13 ]. This means that host plants depend on their fungal partners for resource acquisition through a reciprocal exchange of photosynthesis products [ 2 ]. The secondary forests under investigation on the Loess Plateau are organic matter-limited ecosystems. The retention and release of other nutrients, such as N and P, by organic matter depends on microbial function and activity [ 47 ]. It seems probable that plants growing in these low-organic matter environments formed stronger relationships with microbes through direct symbiosis or nutrient recycling-related feedback [ 37 ]. EM fungal colonization can produce nutrient-acquiring enzymes that could probably mine carbon and N from poorly decomposable litters for their own growth and could utilize pine and oak litter as a phosphorous source [ 48 ]. This increased availability of nutrients results in an increase in the growth and diversity of plants and soil microorganisms and potential interactions among them. Although the influence of biotic interaction on biodiversity is a well-documented topic of research, there is comparatively less investigation into the potential implications for the plant–microbial relationship. Our results revealed that plant–EM fungal correlations were significantly related to high assortativity but lower average path distance of the cross-kingdom networks, which indicated a more pronounced influence of coexisting bacterial communities in rhizosphere soil. Further analysis reveals that the biotic factors have more direct and significant effects on EM fungal communities than abiotic factors in natural secondary forests at the southern Loess Plateau ( Figure 4 ), thereby supporting the third hypothesis. Diverse communities are predicted to be more productive than species-poor assemblages due to their enhanced efficiency in capturing limited resources [ 49 ]. This has been attributed to niche partitioning allowing for the coexistence of different species by utilizing distinct resources. Furthermore, coexisting species may engage in positive interactions, leading to enhanced community performance [ 18 ]. Prior research has indicated that a diverse range of bacteria reside in the surrounding EM symbiosis [ 50 ]. These bacteria have been shown to facilitate the establishment of plant–EM fungal symbioses through the stimulation of mycelial extension, increased root–EM fungal colonization, and the mitigation of adverse conditions that impede mycelial spread, referred to as mycorrhiza helper bacteria [ 17 ]. These bacterial communities differ from those present in bulk soil and exhibit a greater capacity to solubilize inorganic nutrients compared to non-mycorrhizosphere bacteria [ 17 ]. Therefore, the rhizosphere provides a niche for specific bacteria, which, in turn, play an important role in the symbiotic system of trees. It is now evident that EM fungal diversity significantly impacts a wide array of ecosystem processes, encompassing biogeochemical cycles and eco-evolutionary dynamics within above- and below-ground communities, particularly in the context of global change. Linking plant diversity to EM fungal diversity and soil ecosystem function in field studies is therefore far more challenging, as plant communities are influenced by more than just soil abiotic conditions [ 17 , 47 ]. Our findings provide empirical evidence that rhizosphere bacterial communities significantly and predictably enhance the strength of plant–EM fungi relationships in natural secondary forest ecosystems. It is crucial to acknowledge, however, that association studies do not necessarily attribute the causal effects of plants on the soil community, given the myriad of mechanisms through which the EM fungal community can interact with the plant community [ 1 ]. More elaborate indoor experiments under controlled conditions should be conducted to evaluate the relative contributions of species, functional, and genetic diversity in driving these processes, as well as the influence of extrinsic factors in modulating biodiversity–function relationships in future work."
} | 6,796 |
20886303 | null | s2 | 5,573 | {
"abstract": "Immunodominance refers to the phenomenon in which simultaneous T cell responses against multiple target epitopes organize themselves into distinct and reproducible hierarchies. In many cases, eliminating the response to the most dominant epitope allows responses to subdominant epitopes to expand more fully. The mechanism that drives immunodominance is still not well understood, although various hypotheses have been proposed. One of the more prevalent views is that immunodominance is driven by passive T cell competition for space on antigen presenting cells (APCs) or for access to specific MHC:epitope complexes on the surface of APCs. However, several experimental studies suggest that passive competition alone may not fully explain the robustness of immunodominance under physiological conditions or varying proportions of antigen-specific precursor T cells and APCs. These studies propose that a mechanism of active suppression among T cells gives rise to immunodominance.In this work, we present the novel hypothesis that mutual suppression of simultaneous T cell responses results from the appearance of adaptive regulatory T cells (iTregs) during the course of the overall T cell expansion. We extend the mathematical model of T cell expansion proposed in Kim et al. (Bull. Math. Biol. 2009, doi: 10.1007/s11538-009-9463-1 ) to consider multiple, concurrent T cell responses. The model is formulated as a system of independent feedback loops, in which antigen-specific T cell population produces a nonspecific feedback response. Our simulations show that the fastest response to expand gives rise to a de novo generated population of iTregs that induces a premature contraction in slower or weaker T cell responses, leading to a hierarchical expansion as observed in immunodominance. Furthermore, in some cases, removing the dominant T cell response allows previously subdominant responses to develop more fully."
} | 481 |
27148239 | PMC4839258 | pmc | 5,574 | {
"abstract": "As habitats change due to global and local pressures, population resilience, and adaptive processes depend not only on their gene pools but also on their associated bacteria communities. The hologenome can play a determinant role in adaptive evolution of higher organisms that rely on their bacterial associates for vital processes. In this study, we focus on the associated bacteria of the two most invasive seaweeds in southwest Iberia (coastal mainland) and nearby offshore Atlantic islands, Asparagopsis taxiformis and Asparagopsis armata . Bacterial communities were characterized using 16S rRNA barcoding through 454 next generation sequencing and exploratory shotgun metagenomics to provide functional insights and a backbone for future functional studies. The bacterial community composition was clearly different between the two species A. taxiformis and A. armata and between continental and island habitats. The latter was mainly due to higher abundances of Acidimicrobiales, Sphingomonadales, Xanthomonadales, Myxococcales, and Alteromonadales on the continent. Metabolic assignments for these groups contained a higher number of reads in functions related to oxidative stress and resistance to toxic compounds, more precisely heavy metals. These results are in agreement with their usual association with hydrocarbon degradation and heavy-metals detoxification. In contrast, A. taxiformis from islands contained more bacteria related to oligotrophic environments which might putatively play a role in mineralization of dissolved organic matter. The higher number of functional assignments found in the metagenomes of A. taxiformis collected from Cape Verde Islands suggest a higher contribution of bacteria to compensate nutrient limitation in oligotrophic environments. Our results show that Asparagopsis -associated bacterial communities have host-specificity and are modulated by environmental conditions. Whether this environmental effect reflects the host's selective requirements or the locally available bacteria remains to be addressed. However, the known functional capacities of these bacterial communities indicate their potential for eco-physiological functions that could be valuable for the host fitness.",
"introduction": "Introduction Stress-tolerance and adaptation in disturbed environments are mainly studied at the scale of populations of single species. Population genetic diversity is a key factor for adaptation to changing environments (e.g., Massa et al., 2013 ). The local gene pool influences a population's capacity to persist and to expand beyond the native range (for non-indigenous species—NIS) or as habitat changes (e.g., for edge populations; Pauls et al., 2013 ). Besides, most introduced/edge populations are only a subset of the entire species gene pool, often displaying very limited genetic diversity (Bridle and Vines, 2007 ; Dlugosch and Parker, 2008 ). Driven by the rapid advance of culture-independent technologies, there is increasing evidence that the entire holobiome, involving distinct types of symbiosis between microorganisms and host eukaryotes, and the genetic richness of those microbial communities can play a determinant role both in adaptation and evolution of higher organisms (Zilber-Rosenberg and Rosenberg, 2008 ; Tonon et al., 2011 ; Dittami et al., 2014 ). In some cases, like the human gut microbiome, called the forgotten organ (O'Hara and Shanahan, 2006 ), the success of the host is so dependent on the associated microorganisms that the microbial genome functions as an extension of the genome of the host (Mandrioli and Manicardi, 2013 ). The hologenome theory of evolution considers that the holobiont and its hologenome (the sum of host and associated microbiota genome) act in consortium, as a unit of selection in evolution. Genetic variation in holobionts can arise from changes in either the host or the symbiotic microbiota genomes (Rosenberg et al., 2010 ). The diverse microbial symbiont community can aid the holobiont in surviving, multiplying and acquiring time necessary for the host genome to evolve and keep up with rapid and drastic environmental changes (Rosenberg et al., 2010 ). Some bacterial strains associated to invasive seaweeds and absent in the native range, are suggested to play a role in stress tolerance (Aires et al., 2013 , 2015 ), as found in terrestrial plants. Microbial communities establish stable associations with eukaryotic hosts, influencing host fitness and mutually fulfilling several crucial functions (Thompson et al., 2015 ). Bacterial assemblages associated with distinct marine eukaryotes include groups involved in important metabolic processes such as nitrification, nitrogen fixation (Chisholm et al., 1996 ), sulfate reduction (Crump and Koch, 2008 ), photosynthesis (Barott et al., 2011 ), plant growth enhancement (Orole and Adejumo, 2011 ), morphogenesis induction (Nakanishi and Nishijima, 1996 ), or chemical defense (Lee et al., 2009 ; Burke et al., 2011 ). It is uncertain how bacterial assemblages organize across eukaryote hosts, what drives their organization and how stable those communities are. In the specific case of seaweeds, associated bacterial communities can be species-specific or even variety-specific (e.g., Aires et al., 2013 , 2015 ), but they can also change seasonally (Lachnit et al., 2011 ), or result from a competitive lottery model where bacteria with similar metabolic abilities (functions) will be stochastically recruited (Burke et al., 2011 ). The interactive effects of both host and environment have only recently been discussed (Campbell et al., 2015 ; Marzinelli et al., 2015 ). Studies on corals and seaweeds suggested that bacterial assemblages determined by host specificity can be disrupted by environmental pressures (e.g., anthropogenic pollution; Marzinelli et al., 2015 ) and the host will tend to adapt to local conditions by selecting a more advantageous “hologenetic” background from the available bacterial guild (Kelly et al., 2014 ). However, species-specificity disruption can make the host more prone to diseases (Morrow et al., 2012 ). Among the species included in the lists of the “worst invasive alien species threatening biodiversity in Europe” (EEA, 2007 ) and the Mediterranean Sea are those in the red seaweed genus Asparagopsis Montagne (Bonnemaisoniales, Rhodophyta) (Streftaris and Zenetos, 2006 ; Andreakis et al., 2007 ) that has been spreading rapidly across European waters (Andreakis et al., 2009 ). Although, with contrasting geographical distributions, together, A. armata and A. taxiformis are present along all continents and in all oceans across the world, partially due to multiple introduction events (Andreakis et al., 2004 ; Sherwood, 2008 ). A. taxiformis is considered cosmopolitan in subtropical and tropical communities worldwide (Abbott, 1999 ) and, so far, five cryptic lineages have been described for this species, with distinct geographic distributions (Dijoux et al., 2014 ). Due to its complexity and cryptic nature, its natural vs. invasive distribution is still under discussion for some parts of the globe (Dijoux et al., 2014 ). The more temperate species A. armata consists of two cryptic lineages naturally distributed along western and southern Australia and New Zealand and non-indigenous in the Northeast Atlantic and Mediterranean coasts (Andreakis et al., 2007 ; Dijoux et al., 2014 ). In this study, we characterize bacterial communities associated to both Asparagopsis species in contrasting environments: mainland (coastal) and islands (offshore, where only A. taxiformis is present) across the southern Northeast Atlantic, where both species are described as invasive. With a cosmopolitan distribution and, consequent, high adaptive potential (despite the presumably low genetic variability inherent to invasive species), these species represent a good model to study holobiont adaptation to different environments and anthropogenic influences. Considering the whole genetic pool (host + associated bacteria) and expecting a more prompt response from the bacterial partners when compared to that of the host (Rosenberg et al., 2010 ) we believe that the integration of this genetic component will provide innovative insights and the right tools to anticipate the spread of invasive species. Bacterial communities are characterized through next generation barcoding of the V5–V8 region of 16S rRNA gene, in combination with exploratory functional assessments using shotgun metagenomics. Based on previous studies in other marine organisms (e.g., Burke et al., 2011 ; Aires et al., 2015 ), we make the following predictions: (1) host-associated bacterial communities will cluster differentially according to host species if the host species plays an important role in the association; (2) part of the host-associated bacterial guild will be habitat dependent, if they are necessary for coping with the habitat requirements; (3) if the host has “habitat-related” functional requirements fulfilled by the microbiome, then these should be mirrored in the functional profiles of the associated bacteria; (4) if associated bacterial communities (after subtraction of the environmentally available community) are assembled solely according to a lottery model, then none of the above hypotheses should be confirmed. In this dynamic perspective of the holobiont, both host and environment would be decisive for the dynamics of the holobiont structure and the “habitat-dependent” part might offer, to the host, some resilience to disturbed habitats.",
"discussion": "Discussion This study is the first report of species and environmental specificity of bacterial communities of a cosmopolitan/invasive species in contrasting habitats (Mainland vs. offshore Islands), using metabarcoding and metagenomics. Our results showed that the genus Asparagopsis carries bacterial communities that are well differentiated between the closely related sister species A. taxiformis and A. armata within the same habitat. However, the species-specific community composition of A. taxiformis showed a striking differentiation associated with contrasting environments (Mainland vs. offshore Islands). These results are novel and unexpected because other studies of invasive seaweed species have shown that hosts tend to maintain their specific bacterial community in invaded vs. native ranges thousands of Km apart, showing a strong consistent species-specific pattern (Aires et al., 2015 ). Yet, geographical versus environmental specificity are not contradictory or exclusive; as in contrast with Aires et al. ( 2013 ), here we compared contrasting different environments (coastal vs. offshore, polluted/ anthropogenically disturbed vs. pristine). Likewise, in the green alga Bryopsis , differences in bacterial communities were largely due to environment (13%) and host phylogeny (10%), while geography only explained 2% (Hollants et al., 2013 ). Our functional profiles are in agreement with Burke et al. ( 2011 ) who found that there is functional genetic profile equivalence even when bacterial community composition is different. We show that the two different species (sharing the same environmental pressures) are closer in the dendrogram than the conspecific groups. In addition, in our study, bacterial community taxonomy also shows some overlapping in addition to the functional profiles. However, for a direct comparison with these authors' study, a more detailed and replicated approach would be needed. The shared OTUs that could be considered closest to a core bacterial community were highly unevenly represented across the three groups and mostly mirrored the individual plots for groups' unique OTUs. The most widely shared bacteria, with high abundances for all groups, were members of the order Sphingobacteriales (see Table S1 and Figure S2 ) and might be a structural part of the Asparagopsis microbiome. In contrast, several other bacteria strongly contributed to the observed differentiation (Figure 3 , Figure S2 ) showing high correlation to the different groups. Even though functional assignments based on 16S rDNA barcoding have to be done cautiously, some putative functions can be related to bacteria taxonomic classification, which we based on similar studies. Bacterial community composition (associated with different organisms) has been shown to change when facing stress. The orders Acidimicrobiales, Sphingomonadales, Xanthomonadales, and Myxococcales increased in the rhizosphere of halophyte plants ( Halimione portulacoides and Sarcocornia perennis ssp. perennis ), in estuarine salt marshes in Portugal, when the concentration of hydrocarbons in sediment increased along with the presence of homologous genes related to OH degradation (Oliveira et al., 2014 ). Acidimicrobiales and Sphingomonadales were, in our case, part of mainland groups, but absent from the islands groups (Figure 3 ), whereas Xanthomonadales and Myxococcales were more dominant in the A. taxiformis mainland group. Specifically, 100% of the Sphingomonadales order were assigned to the Erythrobacter genus commonly detected in petroleum-contaminated soil, groundwater, and coastal seawater (Alonso-Gutiérrez et al., 2009 ). Besides, the genus Roseobacter (only highly correlated with A. taxiformis Mainland samples) was found to be closely related to oil-metabolizing functions in polluted environments (McKew et al., 2007 ). The Alteromonodales order was only found as distinguishing factor in the two coastal groups (Figure 3 ). This is in accordance with previous studies where bacterial communities associated with the polychaete Ophelina sp. react to the increase of heavy metals by an increase of abundance of bacteria from the Alteromonadales order (Neave et al., 2012 ). Likewise, Alteromonodales have also been found related to sites affected by urbanization and eutrophication (Marcial Gomes et al., 2008 ; Zeng et al., 2010 ) and some of it members are metal-resistant and capable of binding Cu 2+ and Zn 2+ cations thereby reducing their toxicity (Vincent et al., 1994 ). The described functions of the groups here found are in agreement with our metagenomic data which showed a higher proportion of metabolic functions associated to stress response and resistance to toxic compounds, more specifically, most hits on resistance or tolerance to heavy metals (as copper, cobalt, arsenic, and zinc) in Asparagopsis from the mainland sites. Also, in seaweeds, the toxic effect of heavy metals, and other environmental stresses, appears to be related to production of reactive oxygen species (ROS), which impose oxidative stress on the cells (Dring, 2006 ) and results in unbalanced cellular redox status. Algae respond to heavy metals by induction of several antioxidants, including diverse enzymes such as glutaredoxins and the synthesis of low molecular weight compounds such as glutathione (Pinto et al., 2003 ; Mellado et al., 2012 ). This was reflected in our metagenomic data which showed functional assignments to the glutathione pathway, only found in mainland samples (where heavy metals and other environmental stresses are expected to be higher, which leads to ROS) which is used to scavenge the ROS produced (Rijstenbil et al., 2000 ; Ratkevicius et al., 2003 ). Only samples collected in mainland showed hits on alkaloids. As an extra mechanism of defense, alkaloids are produced to enable plant/seaweed protection against pathogens and herbivores (War et al., 2012 ) and as detoxifiers in polluted soils by terrestrial plants (Khodjaniyazov, 2012 ). Genes of secondary metabolites, other than alkaloids, were also more abundant in our samples presumably more-exposed to pollution. It has been suggested that macroalgae depleted from their own chemical defense are able to rely on the secondary metabolites produced by their associated bacteria (Egan et al., 2000 ). Since the number of metagenome reads assigned to eukaryotes was very low, it is likely that these functions inferred were provided by the bacterial community. All these results suggest complex interactions between macro and microorganisms. It is likely that that part of the seaweed microbiome might be more under influence of the environment rather than the host. However, there is also a degree of specific association as shown when invasive seaweed species switch environments yet maintaining their own specific microbiome (Aires et al., 2015 ). Regardless of the lack of specific data on environmental variables of our sampling sites, the mainland coastal area, where our samples were taken, is under constant anthropogenic influence with fishing/commercial and recreational maritime activities. Also, due to the proximity of an industrial area (including a hydroelectric plant, refineries, and petrochemical industries) and a fishing port, the levels of pollutants are known to be high (Anonymous, 2008 ). The discriminative presence of the referred orders in the coastal groups, together with the higher metabolic assignments on protection mechanisms, might well be related with the host necessity of “gathering specific symbionts” with “specific functional capacities” for environmental remediation in order to survive and persist. Nevertheless, some of these bacteria may be just looking up for themselves and their necessity of coping with the adversities. The hypothesis that the seaweed would not take advantage of the described mechanisms, cannot be discarded and should also be considered. Bacterial communities associated to A. taxiformis from offshore Islands were distinguished by the presence of the family Hyphomonadaceae of the Caulobacterales order (in the combination of the results shown in Figure 3 and Figure S2 ). Hyphomonadaceae members are widely distributed in marine environments (Anast and Smit, 1988 ) and especially common in oligotrophic waters (Alain et al., 2008 ). They are believed to play an important role in the mineralization of dissolved organic matter (Abraham et al., 1999 ), an important trait in oligotrophic conditions (Biddanda et al., 2001 ). These studies are in agreement with our findings for the A. taxiformis Islands group sampled from oligotrophic/non-eutrophic waters. The Island group show a distinct abundance of metabolic functions assigned to osmotic stress but there is insufficient evidence to suggest that the salinity, in these very dry islands, would be significantly different from the other locations. This apparent association of the bacterial community with osmotic stress might be simply due to the lack of other stresses in these relative pristine conditions. Dittami et al. ( 2016 ) showed that the capacity of Ectocarpus cultures to grow in diluted standard seawater medium is correlated with the presence of representatives of the Sphingomonadales order. This order is the best represented in all the replicates throughout all groups not allowing us to draw any direct link between these findings. Overall, and besides the lack of replicates for metagenomics, our bacterial characterization through 16S barcoding is in agreement with metagenomic results. It is apparent that the environmental conditions (polluted, under anthropogenic influence coastal sites vs. offshore more pristine island waters) may shape not only the bacterial community composition, but consequently mirror the compositional differences in differentiated functional profiles. Our results suggested that the microbiome composition may primarily be influenced by the host traits (narrowing down from phylogenetic to individual host differences) and then by environmental conditions in which functional capability should be more important than bacterial taxonomic composition. A recent study (Hester et al., 2016 ) described that bacterial communities associated with seaweeds can be divided in two groups; the stable symbionts which are found specifically associated with a particular taxonomic group, and the sporadic symbionts which can be either the product of stochastic events or the response to environmental pressures (as supported by studies as Kelly et al., 2014 ). This suggests that changes in the symbiont members can lead to holobionts better adapted for particular conditions. These results are in line with our findings. Moreover, we hypothesize that community assembly through competitive lottery (as shown by Burke et al., 2011 ) may not be a simple gamble. From our results the bacterial communities found in both disturbed environments are not only known for the same set of functional capacities but are also phylogenetically related. So, this lottery may not only be restricted to the bacterial availability within the guild (where the ocean might be the source), but also to their ability to perform functions required under specific environmental demand. This might explain the larger number of OTUs (Figure S1 ) and metabolic similarities shared between the distinct Asparagopsis species from the same “disturbed” coastal environment when compared to the two conspecific, but in opposite environments, groups. Also, the Islands group showed reduced bacterial community richness when compared to the “disturbed environments” groups (Table S4 ). In agreement with other studies (Kelly et al., 2014 ; Marzinelli et al., 2015 ), we hypothesize that the host species might be the first factor shaping their bacterial community assembly. Yet, host-specific communities will likely adapt (as other traits) when facing environmental stresses by shifting to alternative communities (some specific bacteria and/or advantageous metabolic genes) that might act as in situ bioremediators to the host's advantage. As the environment changes, there is reassembling and the acquisition of new members. However, in contrast to our descriptive study, experimental studies are required to evaluate the mechanisms behind these processes and assembly: Do hosts really select advantageous bacteria or is their abundance directly linked to their availability in the environment? Are those bacterial genes providing function to their macroalgal host by assuring defense against toxicity or are they just covering bacterial needs? The supportive metagenomic data in this study showed the importance of combining tools, 16S barcoding and metagenomics as well as metatranscriptomics in the future, to unveil the important factors shaping host-associated bacterial communities. Although, more detailed sampling with the inclusion of different replicates in the metagenomic data, as well as experimental and manipulative studies (concerning bacterial communities), are needed to determine the whole adaptive potential of seaweed species under climate change and anthropogenic influence. Holobionts should be understood as a dynamic unit (Hester et al., 2016 ), not a static group of genomes that evolve together as suggested by the Hologenome theory of evolution (Rosenberg et al., 2007 )."
} | 5,763 |
36514531 | PMC9731602 | pmc | 5,576 | {
"abstract": "Abstract Due to their versatility and the high biomass yield produced, cultivation of phototrophic organisms is an increasingly important field. In general, open ponds are chosen to do it because of economic reasons; however, this strategy has several drawbacks such as poor control of culture conditions and a considerable risk of contamination. On the other hand, photobioreactors are an attractive choice to perform cultivation of phototrophic organisms, many times in a large scale and an efficient way. Furthermore, photobioreactors are being increasingly used in bioprocesses to obtain valuable chemical products. In this review, we briefly describe different photobioreactor set‐ups, including some of the recent designs, and their characteristics. Additionally, we discuss the current challenges and advantages that each different type of photobioreactor presents, their applicability in biocatalysis and some modern modeling tools that can be applied to further enhance a certain process.",
"conclusion": "5 CONCLUDING REMARKS In a world in which the consumption of energy is unceasingly rising, and where most of this energy comes from non‐renewable sources, fuels and other products sustainably obtained must play a global major role towards a greener future. In this context, algae present themselves as a very attractive tool to fight climate change, and achieve the goals proposed in the Paris Agreement. There are still many challenges to face when using PBRs (Table 2 ), especially regarding industrial scale production of biomass, while also optimizing energy consumption and light harvesting, but we firmly believe that these setbacks are as challenging as they are rewarding. Combining the great variety of PBR types (Table 1 ) together with technologies such as light filters, different types of illumination and/or novel aeration or mixing methods is likely to result in new efficient strategies for large scale cultivation of biomass and production of useful substrates.",
"introduction": "1 INTRODUCTION There are plenty of different species of phototrophic microorganisms, ubiquitous all around the globe, with estimations ranging from a moderate number of 30,000 to the astonishing count of 1 million [ 1 ]. These organisms are indeed so numerous, that it is calculated that they are responsible for almost 50% of the total photosynthesis of our planet [ 2 ]. Amongst them stand the microalgae and the Gram‐negative prokaryotic cyanobacteria. They stand out as mayor players for commercial applications, already being widely discussed for biofuel production [ 3 ], or as natural sources of valuable products such as pigments [ 4 ], bioactive compounds or bulk chemicals. Furthermore, they present exciting results for their application in the food and feed industry [ 5 , 6 ]. An excellent example of this is the production of cyanobacteria of the genus Spirulina , frequently used for animal and human consumption because of their outstanding nutritional value [ 7 , 8 ]. On the other hand, taking into account the alarming increase in global CO 2 levels, which results in severe climate change effects, and considering the high sequestration capability that algae present, which is estimated in 1.83 kg of CO 2 per kg of dry biomass [ 9 ], it was proposed that they could be part of the way towards a more sustainable economy [ 10 ]. Yet other environmental applications of algae in which they showed great promise is the treatment of affluents [ 11 , 12 ], or as sustainable biofertilizers, which reduces the use of polluting synthetic fertilizers [ 11 , 12 , 13 , 14 , 15 , 16 ]. All these applications have generated a growing interest in these organisms that is also reflected in the steady growth in the number of papers regarding this subject, as seen on Figure 1 . FIGURE 1 Number of microalgae publications published each year in the last 25 years, in which it is possible to see the growing interest in microalgae. Analysis made using “Web of Science” database searching “microalgae” as topic on the 15‐06‐2021 PRACTICAL APPLICATION Photobioreactors are an attractive choice to perform microalgae cultivation and biocatalytic processes in an efficient way, sometimes even in a large‐scale. Although photobioreactors have been used for years, they are still being studied and improved using different shapes and illumination techniques in order to optimize the process. In this review, we aim to gather the latest information about different photobioreactor set‐ups, including some of the most recent designs, their main characteristics, the main parameters affecting the photobioprocess, and the advantages and disadvantages of each type of photobioreactor in order to provide a clear idea of them, and acknowledge their present situation and future challenges. Furthermore, we discuss the application of photobioreactors in biocatalysis and the application of different mathematical modeling tools that can be used to improve the efficiency of a given process. Since these microorganisms are photosynthetic, their growth is conditioned to the presence of a source of light. Nevertheless, this is not the sole factor to consider, since also pH, CO 2 , nitrogen availability, salinity, temperature, oxygen removal and medium mixing all play a major role, and must be tuned for each specific case, given that the growth rate and biomass maximum concentration of algae tend to vary between species [ 17 ]. One of the most common ways of growing phototrophic microorganisms is using open systems, such as raceway ponds. Although these are economic in terms of construction, operation and maintenance, the control on growth conditions tends to be poor, which makes them unsuitable for the production of fine chemicals, and pharmaceutical or food ingredients [ 9 ]. Additionally, there is a non‐negligible risk of contamination by parasites and predators of microalgae, such as rotifers or ciliates, which can destroy microalgal biomass in days, causing severe losses to productivity [ 17 ]. In contrast, closed photobioreactors (PBRs) offer an excellent control on culture conditions with minimal risk of contamination, but with a higher initial investment, and operational and maintenance costs [ 18 ]. The use of phototrophic organisms in biocatalysis has been a discipline in constant expansion in the last years, with several important fine chemicals being produced in this way both in industrial and lab scale [ 19 ], typically using either flat panel or tubular PBRs. Nevertheless, other types of PBRs have been the subject of many studies, with interesting advances regarding their design and application. We encourage the curious reader to take a look at some of the excellent works that have been published on the topic of PBRs, such as the ones from Carvalho et al. [ 20 ], Dasgupta et al. [ 21 ], Zitelli et al. [ 22 ], Chang et al. [ 23 ], and Płaczek et al. [ 24 ]. These authors present different perspectives on the topic and provide a great amount of information. Nevertheless, it should be considered that many years have passed since their publication, and therefore an update was needed. More recently, Johnson et al. [ 25 ], Sero et al. [ 26 ], and Legrand et al. [ 27 ] published very thorough reviews in which they briefly describe different types of PBRs, with an special emphasis in cultivation strategies in the first one, advances in biophotonics in the second one, and modelling in the third. Different from the abovementioned literature, in this review we significatively expand the information provided by these previous studies regarding the specifications of PBRs, while also providing examples of state‐of‐the‐art developments in the field. Furthermore, we also discuss the applicability of PBRs in biocatalysis, and provide a brief overview of mathematical modeling tools that might be applied in each process. With the objective of making this manuscript as practical as possible for the reader, the technical characteristics of PBRs and a comparison between different designs are presented in a tabulated form (Tables 1 and 2 , respectively). TABLE 1 Technical specifications of different PBRs PBR type Temperature control Mixing Gas exchange Stirred tank Heat exchanger/reactor jacket Stirrer/impeller Gas injection Conventional Tubular Vertical column Independent unit Airlift/bubbles Gas exchange at head space Horizontal tubular Water spraying, shading Circulation with pumps Injection into feed. Dedicated degassing unit Flat panel Heat exchange coils Airlift/bubbles Gas exchange at head space Bag Independent unit Circulation with pump Injection in a separate unit Unconventional Pyramid Independent unit Airlift/bubbles Gas injected through sparger Nature inspired Thermostatic bath and shell and tube heat exchanger Pump/small bubbles from turbulent flow Gas injected through sparger Hybrid Membrane Independent unit Circulation with pump/bubble mixing Gas injection Open‐close Heat exchanger Recirculation with pumps Gas injection in one of the units John Wiley & Sons, Ltd. TABLE 2 Comparison between different PBRs PBR type Use in photo‐biotrans formations Advantages Limitations Conventional Conventional Stirred tank Yes [ 44 ] Easier to control. Possibility of running axenic cultivation Low surface to volume ratio, low efficiency of light absorption and low productivity [ 102 ] Tubular Vertical column Yes [ 55 ] Good biomass growth, high efficiency of photosynthesis, high potential of scalability. Cheap and easy to maintain. Low energy use and suitability for algae immobilization Small area of light exposition that is additionally reduced with the increase of column diameter. Low surface to volume area. Possibility of biofilm formation on reactors walls Horizontal tubular Yes [ 17 ] Large surface to volume ratio. Allow high biomass density High energy consumption, photo bleaching, Scale up challenging due to the loss of the surface to volume ratio Flat panel Yes [ 57 ] Large surface of exposition to light. High surface to volume ratio. High productivity of biomass. Ability to maintain uniform access to light across entire volume of cultivation. Small concentration of dissolved oxygen Increase of production scale requires the use of numerous modules [ 51 ]. Difficulties in control of cultivation temperature. Risk of fouling. Potential hydrodynamic stress in some algae species.Photoinhibition Bag No Good adaptability, simplicity, and cost‐effectiveness Need for periodic bag replacement.Considerable amount of waste when using disposable plastic bags.Prone to photoinhibition and leakage Unconventional Pyramid No Small area of land needed. Completely automated, Possible to have internal and external illumination. Fully automated [ 26 , 78 ] Further studies are required Nature‐inspired No Extremely high surface area to volume ratio. Good performance in multiphasic flow Further studies are required Hybrid Membrane Yes [ 88 ] Easy operational maintenance and low operating temperature High pressure drops, fouling formation and difficult to optimize in full scale plant Open‐close Yes [ 86 ] High surface area/volume ratio. Maximizes solar harvest Need to improve the design's biomass and oil yields John Wiley & Sons, Ltd."
} | 2,802 |
25935554 | PMC4425913 | pmc | 5,580 | {
"abstract": "Background Lactic acid is the building block of poly-lactic acid (PLA), a biopolymer that could be set to replace petroleum-based plastics. To make lactic acid production cost-effective, the production process should be carried out at low pH, in low-nutrient media, and with a low-cost carbon source. Yeasts have been engineered to produce high levels of lactic acid at low pH from glucose but not from carbohydrate polymers (e.g. cellulose, hemicellulose, starch). Aspergilli are versatile microbial cell factories able to naturally produce large amounts of organic acids at low pH and to metabolize cheap abundant carbon sources such as plant biomass. However, they have never been used for lactic acid production. Results To investigate the feasibility of lactic acid production with Aspergillus , the NAD-dependent lactate dehydrogenase (LDH) responsible for lactic acid production by Rhizopus oryzae was produced in Aspergillus brasiliensis BRFM103. Among transformants, the best lactic acid producer, A. brasiliensis BRFM1877, integrated 6 ldhA gene copies, and intracellular LDH activity was 9.2 × 10 −2 U/mg. At a final pH of 1.6, lactic acid titer reached 13.1 g/L (conversion yield: 26%, w/w) at 138 h in glucose-ammonium medium. This extreme pH drop was subsequently prevented by switching nitrogen source from ammonium sulfate to Na-nitrate, leading to a final pH of 3 and a lactic acid titer of 17.7 g/L (conversion yield: 47%, w/w) at 90 h of culture. Final titer was further improved to 32.2 g/L of lactic acid (conversion yield: 44%, w/w) by adding 20 g/L glucose to the culture medium at 96 h. This strain was ultimately able to produce lactic acid from xylose, arabinose, starch and xylan. Conclusion We obtained the first Aspergillus strains able to produce large amounts of lactic acid by inserting recombinant ldhA genes from R. oryzae into a wild-type A. brasiliensis strain. pH regulation failed to significantly increase lactic acid production, but switching nitrogen source and changing culture feed enabled a 1.8-fold increase in conversion yields. The strain produced lactic acid from plant biomass. Our findings make A. brasiliensis a strong contender microorganism for low-pH acid production from various complex substrates, especially hemicellulose.",
"conclusion": "Conclusion The overexpression of a heterologous ldhA gene from R. oryzae made it possible to produce significant amounts of L-LA by A. brasiliensis recombinant strains. The best L-LA producer strain, A. brasiliensis BRFM1877, integrated 6 ldhA gene copies in its genome and demonstrated an intracellular LDH activity of 9.2 x 10 −2 U/mg. Switching the nitrogen source from ammonium sulfate to Na-nitrate prevented a strong drop in pH and increased the L-LA conversion yields. With nitrate, at pH 3, L-LA conversion yield reached 44% (w/w) with low quantities of by-product. L-LA productivity was higher at near-neutral pH than at pH 3, but the lower pH prevents the production of gluconic acid resulting in similar L-LA conversion yields with or without pH neutralization. The recombinant strain was able to produce L-LA from xylose and associated arabinose, the main pentose monomers found in plant biomass. It was also able to produce L-LA directly from starch and birchwood xylan without addition of polysaccharide depolymerizing enzyme cocktails. Further studies including process and metabolic engineering are expected to increase both titer and yield. These findings raise new prospects for using Aspergillus species as new hosts for L-LA production from plant biomass.",
"discussion": "Results and Discussion Selection of transformants in ammonium medium without pH regulation After co-transformation of A. brasiliensis BRFM103 with the lactate dehydrogenase A gene ( ldhA ) and the hygromycin resistance gene ( hph+ ), isolated transformants were screened for L-LA production in MM1 liquid medium without pH regulation. pH of the medium dropped from 5 to 2.5 in 40 h then slowly decreased to a final pH of 1.6 at 138 h culture. Eight transformants produced L-LA and were selected for further analysis. The supernatants of these clones were analyzed by chromatography to determine residual glucose, L-LA and by-product concentrations. At 138 h incubation, all the glucose was consumed. L-LA production at 138 h incubation was low (0.03 to 0.6 g/L) for three of the L-LA-producing transformants (BRFM1879, BRFM1873, and BRFM1881). Five recombinant strains (BRFM1872, BRFM1874, BRFM1875, BRFM1877 and BRFM1880) demonstrated high L-LA production (11.7 to 13.9 g/L) from 50 g/L of glucose at 138 h of culture. L-LA conversion yields ranged from 0.1% (w/w) for BRFM1879 to 27% (w/w) for BRFM1880 (Figure 1 ). Recombinant strains produced low ethanol conversion yields ranging from 1 to 14% (w/w). For the transformants exhibiting L-LA conversion yields higher than 23% (w/w), ethanol production was nearly abolished with final conversion yields below 1% (w/w), suggesting that the heterologous ldhA efficiently compete with native pyruvate decarboxylase (PDC) for pyruvate utilization in high L-LA-producing strains. Similar results were obtained in yeasts with native ethanol production such as Candida sonorensis, where the natural production of around 20 g/L of ethanol was nearly abolished by ldh insertion in neutralized culture conditions [ 13 ]. In S. cerevisiae , which naturally produces larger amounts of ethanol, additional modifications of the ethanol pathway are necessary to abolish ethanol production [ 34 , 35 ]. L-LA production had no severe impact on growth of the recombinant strains. However, the biomass conversion yields of high L-LA-producing strains, at 22 to 24% (w/w), were lower than the biomass conversion yields of low L-LA-producing strains, at 28 to 31% (w/w) (Figure 1 ). Figure 1 Conversion yields of glucose to biomass, lactic acid, and ethanol by the L-LA-producing transformants of A. brasiliensis . Conversion yields of lactic acid at 138 h of culture (orange square), ethanol at 138 h of culture (blue square), and dry biomass at 144 h of culture (gray square), are expressed in g/100 g of glucose consumed. Transformant strains are sorted on the basis of their L-LA conversion yields. Gene copy number and intracellular LDH activity in recombinant strains ldhA gene copy number and intracellular NADH dehydrogenase activities were determined for the recombinant A. brasiliensis-ldhA strains. Fresh cultures of A. brasiliensis-ldhA transformants in MM2 liquid medium without pH regulation were used for these experiments. The wild-type BRFM103 strain was studied as a control. The genome of LA-producing strains contained between 3 and 16 ldhA gene copies (Figure 2 A). Intracellular LDH activities increased with ldhA gene copy number up to 6 copies, corresponding to a maximum LDH activity of 9.1x10 −2 U/mg (Figure 2 A). Above 6 ldhA gene copies, LDH activity decreased gradually to a minimal value of 1.4x10 −2 U/mg. The reduced LDH activity could be a consequence of ldhA mRNA degradation due to gene silencing occurring in Aspergillus [ 36 ]. Furthermore, some transformants with equal gene copy numbers showed variable LDH activity levels, probably as a result of ectopic integration of the ldhA gene, which is a feature of the transformation system [ 37 ]. Indeed, expression level can be different in different regions of the genome. Figure 2 Relation between intracellular LDH activity and gene copy number (A) or lactic acid conversion yields (B) . The gene copy number was determined with fresh 24 h cultures. Activities are expressed in U per mg of total proteins, where 1 U is the amount of μmole NADH, H + reduced by intracellular extract per min. For intracellular LDH activity, measurements were carried out in triplicate at 72 h of culture; error bars show the standard deviation. Conversion yields were measured during the screening assays at 138 h incubation. Increasing LDH activity had a positive impact on LA yields up to 6.7x10 −2 U/mg (Figure 2 B). Above this threshold, there was no further observable increase in yield. This suggests that the LDH activity is not a limiting step in lactic acid production above 6.7x10 −2 U/mg, and so another parameter, such as medium composition or LA toxicity, might be limiting LA production. Since high LDH activity is likely to empower LDH in the competition for pyruvate utilization, A. niger BRFM1877, showing the highest intracellular LDH activity, was chosen for further investigation. Impact of pH on L-LA production by A. brasiliensis BRFM1877 Over the course of liquid culture of BRFM1877 in MM1, pH of the medium dropped from 5 to 2.4 in 40 h then slowly decreased to a final value of 1.6 at 138 h of culture. This intense acidification illustrates the acid tolerance of Aspergilli. For citric acid production, a pH of 2 increases production yield [ 38 ]. However, at low external pH, more energy is needed for the export of lactate, since undissociated L-LA from the broth can freely reenter the cell. This is suggested to be the main reason for weak acid stress in yeasts [ 11 , 13 ]. In order to evaluate the impact of pH acidification during L-LA production by BRFM1877, the strain was cultured in MM2 liquid medium (50 g/L glucose initial) containing Na-nitrate as nitrogen source instead of ammonium sulfate. In this experiment, extra glucose (20 g/L) was added at 96 h growth to increase final titers. With Na-nitrate as nitrogen source, pH dropped from 5 to 3.2 in 48 hours then remained unchanged during further culture. The L-LA titers and conversion yields reached mean values of 17.7 g/L and 32.2 g/L, 47% (w/w) and 44% (w/w), at 90 h (before extra glucose addition) and 260 h of culture, respectively (Figure 3 A). These titers and yields are higher than those obtained at pH below 2 (13.1 g/L, 26%, w/w), highlighting that ammonium sulfate as a nitrogen source restricted L-LA production, most probably via strong pH acidification. Na-nitrate was therefore used as nitrogen source to further investigate the impact of pH on L-LA production by A. brasiliensis BRFM1877. Figure 3 Impact of pH on organic acid production by A. brasiliensis BRFM1877 in glucose–nitrate-containing medium. Glucose (blue square), gluconic acid (black triangle), ethanol (blue circle), and lactic acid (orange circle) concentrations in cultures carried out in MM2 liquid medium started with 50 g/L of glucose then added with 20 g/L of glucose at 96 h of culture, without (A) and with (B) addition of 80 g/L CaCO 3 after 24 h of culture. The evolution of pH in both cultures, with (black triangle) and without (violet square) CaCO3 addition, is also presented (C) . Experiments were carried out in triplicate. Error bars show standard deviations. Cultures at near-neutral pH were performed in MM2 liquid medium with CaCO 3 addition and extra glucose addition at 96 h. CaCO 3 was added after spore germination (24 h) at a final concentration of 80 g/L. Immediately after CaCO 3 addition, pH increased from 4.0 to 6.5 and stabilized around 7.5 within a few hours and throughout subsequent fungal growth. L-LA productivity was strongly improved from 0.13 g/(L.h) to 0.32 g/(L.h) by pH neutralization (Figure 3 ). Furthermore, the L-LA titers reached mean values of 20.8 g/L and 35.6 g/L at 96 h (before extra glucose addition) and 120 h of culture, respectively (Figure 3 B). The conversion yields from 70 g/L of glucose were therefore 51% (w/w) and 46% (w/w) with and without CaCO 3 addition, respectively. This result shows that adding CaCO 3 has a small effect on L-LA titers and conversion yields of A. brasiliensis BRFM1877 in the conditions tested. Moreover, cultures at pH above the pKa of lactic acid (3.8) led to the production of calcium lactate and complicate the recovery of free lactic acid [ 2 ]. Therefore, concentration of free lactic acid will be higher in culture at pH 3 than culture at pH 7.5 (with CaCO 3 addition). The low effect of pH neutralization on L-LA titers and conversion yields could be explained by the significant increase of gluconic acid. The titer of gluconic acid obtained in neutralized conditions was 22.9 g/L whereas only 6.2 g/L was produced at pH 3. Glucose oxidase, the enzyme responsible for gluconic acid production, is known to be less efficient at a pH below 3 [ 38 ]. Therefore, the slight positive effect of pH neutralization on lactic acid production might be attenuated by the negative effect of gluconic acid production. One additional advantage of culturing A. brasiliensis BRFM1877 at pH 3 is that it reduces the production of gluconic acid without genetic modification of the gluconic acid metabolic pathway. Studies on recombinant yeast strains expressing heterologous ldh genes have reported various effects of pH on L-LA production [ 7 , 13 , 39 , 40 ], the most prevalent being the effect of host genetic background and origin of the heterologous LDHA on low pH L-LA production [ 13 , 41 ]. Looking at the recombinant yeasts expressing the same ldh gene from R. oryzae , A. brasiliensis BRFM1877 compares favorably for low-pH L-LA production. For instance, the production yield of S. cerevisiae cultivated at pH 3.5 was 32% (29.2 g/L of L-LA from 92 g/L of glucose) [ 40 ] whereas C. sonorensis cultivated without pH regulation produced 12% of L-LA (6 g/L from 50 g/L of glucose) [ 13 ]. Here, L-LA yielded 44% (32.2 g/L of L-LA from 70 g/L of glucose) in nitrate minimum medium at pH 3, demonstrating that A. brasiliensis BRFM1877 shows promising performances for L-LA production at low pH. Production of lactic acid from various substrates by A. brasiliensis BRFM1877 In agreement with the versatile ability of Aspergilli to break down the lignocellulose building blocks, we evaluated the ability of A. brasiliensis BRFM1877 to assimilate the main pentose monomers from hemicelluloses, i.e. D-xylose and associated L-arabinose, compared to the main hexose, D-glucose, in MM2 liquid medium without pH neutralization. As expected from the literature [ 15 ], the best L-LA production yield (34%, w/w) was obtained from D-glucose (Figure 4 A). D-xylose was also an efficient carbon source, at 24% (w/w) L-LA production yield (Figure 4 B). Using L-arabinose, production yield was 18% w/w (Figure 4 C). A similar decrease in product formation when pentoses are used instead of glucose is reported for ethanol production by yeasts. [ 42 ]. A redox imbalance due to the need for NADPH in the reduction step of xylose degradation is proposed to be responsible for this decreased production [ 42 ]. In Aspergillus, D-xylose and L-arabinose degradation also use NADPH. Moreover, L-arabinose has a greater need for NADPH than D-xylose [ 43 ]. Similar redox imbalance could explain the lower L-LA yield observed with L-arabinose compared to D-xylose. The maximum concentration of L-LA (18.5 g/L) was obtained at 120 h culture from D-glucose (Figure 4 A). Production time was extended by using pentoses. The strain produced maximal titers of 12.3 g/L L-LA at 216 h from D-xylose (Figure 4 B) and 9.0 g/L at 264 h from L-arabinose (Figure 4 C). The various production timescales could be explained by the lag phase due to adaptation of the strain transferred from Sabouraud broth (glucose) to media containing pentoses as carbon sources. Alternatively, fungal growth could be slowed down by its xylose and arabinose uptake rates [ 17 ]. Figure 4 Lactic acid production and sugar consumption from monosaccharides and plant biomass polysaccharides. Lactic acid (orange square), D-glucose (blue square), D-xylose (black square), L-arabinose (green square) concentrations were determined when A. brasiliensis was grown on monosaccharides (A, B, C) with D-glucose (A) , D-xylose (B) or L-arabinose (C) . A. brasiliensis was also grown on polysaccharides (D) to produce lactic acid from starch (blue circle) or xylan (black circle). For each, initial substrate concentration was of 50 g/l and 50 g/kg for monosaccharides and polysaccharides, respectively. Experiments were carried out in triplicate. Error bars show standard deviations. Since A. brasiliensis BRFM1877 was able to produce L-LA from glucose and xylose, we investigated its ability to directly convert glucose and xylose-based polymers, i.e. potato starch and birchwood xylan. L-LA titers reached 11.7 g/L at 120 h and 8.6 g/L at 168 h of cultivation from starch and xylan, respectively (Figure 4 D). Production yields were 23% (w/w) from starch and 17% (w/w) from xylan. The loss of yield on these polymers compared to the production yields of their corresponding monomers (glucose 34%, w/w; xylose 24%, w/w) could be attributed to the extra energy needed for the production of starch or xylan-degrading enzymes. Alternatively, L-LA production could be impacted by an incomplete degradation of the polymers. Interestingly, there was no significant lag in L-LA production times on polymers compared with their corresponding monomers. Indeed, the maximum L-LA titers were obtained at 120 h from glucose- and starch-containing media. Surprisingly, our results even suggest that xylan was consumed faster than xylose. Note that we used the same preculture to inoculate all the flasks in this experiment. This result suggests that degradation of the polymer was not a limiting step for L-LA production from starch and xylan. The major bottleneck for cost-efficient metabolite production from polymers is the cost of enzymatic pretreatment of raw materials [ 23 ]. Here, the production of L-LA directly from starch and xylan, without addition of enzyme cocktail, emerges as a novel and promising biotechnological method for L-LA production."
} | 4,436 |
22900014 | PMC3416842 | pmc | 5,582 | {
"abstract": "We study the properties of the dynamical phase transition occurring in neural network models in which a competition between associative memory and sequential pattern recognition exists. This competition occurs through a weighted mixture of the symmetric and asymmetric parts of the synaptic matrix. Through a generating functional formalism, we determine the structure of the parameter space at non-zero temperature and near saturation (i.e., when the number of stored patterns scales with the size of the network), identifying the regions of high and weak pattern correlations, the spin-glass solutions, and the order-disorder transition between these regions. This analysis reveals that, when associative memory is dominant, smooth transitions appear between high correlated regions and spurious states. In contrast when sequential pattern recognition is stronger than associative memory, the transitions are always discontinuous. Additionally, when the symmetric and asymmetric parts of the synaptic matrix are defined in terms of the same set of patterns, there is a discontinuous transition between associative memory and sequential pattern recognition. In contrast, when the symmetric and asymmetric parts of the synaptic matrix are defined in terms of independent sets of patterns, the network is able to perform both associative memory and sequential pattern recognition for a wide range of parameter values.",
"introduction": "Introduction Neural networks were originally developed to model the behavior of the brain. However, due to the great complexity of the brain's neural circuitry and of the synaptic interactions, it was necessary to propose simplified models, such as the McCulloch-Pitts model [1] and the Hopfield model [2] which, although simple, still capture some important characteristics of the neuronal dynamics. One important question in this field is how and when a neural network is able to memorize a given set of patterns. Two main different mechanisms to store information in a neural network have been identified, to wit, the Associative Memory (AM) on the one hand, and the Sequential Pattern Recognition (SPR) on the other hand. The neural network performs AM when its dynamical attractors are fixed points, each corresponding to one of the patterns that we want to store in the network. This type of dynamic behavior is characterized by a symmetric interaction matrix that contains the connection strength between the neurons. Some examples of this dynamics are the Hopfield model and the Little model [2] , [3] , [4] , [5] . Contrary to the above, in SPR the network memorizes a fixed set of patterns which are retrieved in certain order in time. From a dynamical point of view, this corresponds to a cyclic attractor consisting of the sequence of patterns stored in the network in a given order. A necessary condition for SPR to occur is that the matrix of neuron-neuron interactions has to be asymmetric. A well known example of this type of dynamics is the asymmetric Hopfield model [2] , [6] . Several dynamical phases have been identified in both the AM and SPR models. Broadly speaking, these phases characterize how well the network can recognize its set of patterns, and the type of memory, i.e., whether AM or SPR. Both AM and SPR have been widely studied separately. Nonetheless, there is evidence showing that in real neural networks the synaptic connections are neither fully symmetric nor fully asymmetric [7] , [8] . Rather, they can be considered as a mixture of these two cases, generating an interaction network with a complex topology. Additionally, there is evidence that the brain is capable to perform both AM and SPR [8] , [9] . For instance, recalling the color of a simple object would be an example of AM, whereas recalling the digits in a phone number in the proper order would constitute an example of SPR. Since these two types of pattern retreival coexist in the brain, several authors have introduced modifications to the Hopfield model in order to obtain both types of pattern retrieval within the same network [10] , [11] , [12] . One approach to this problem was proposed by Coolen and Sherrington in Ref. [10] : They introduced a model in which the interaction matrix has two parts, one symmetric and one asymmetric. These two parts are weighted by a mixture parameter \n , in such a way that the interaction matrix , (also called the synaptic matrix), can be written as (1) where and are symmetric and asymmetric matrices, respectively. For only the symmetric part is present (the classical Hopfield Model) and therefore the network performs AM, while for only the asymmetric part survives (the asymmetric Hopfield Model) and the network performs SPR. For intermediate values of , there is a competition between the symmetric and asymmetric parts of the synaptic matrix. One of the main questions in this model, which we will refer to as the Coolen-Sherrington model , (or the CS model for short), is how the network dynamics transit from AM to SPR as varies from 1 to 0. The point is that for all the patterns are stored as independent attracting fixed points, whereas for the patterns are stored as part of a single large cyclic attractor. Is this transition from AM to SPR continuous or discontinuous? Can some of the patterns be stored in a cyclic attractor whereas some other patterns are stored as fixed point attractors? As we will see, the answer to these questions depends on the definition of the symmetric and asymmetric parts of the synaptic matrix. In the original CS model the symmetric part and the asymmetric part are correlated because they are defined in terms of the same set of patterns (see the definition of the model in the next section). Additionally, Coolen and Sherrington studied this model for the particular case where the number of patterns stored in the network is smaller that the number of neurons ( ). In terms of the load parameter , the above condition corresponds to in the thermodynamic limit (this regime is termed non saturated .) Within this regime, Coolen and Sherrington found that for parallel updating and for , the network dynamics exhibit only fixed point attractors, i.e. the network performs AM. However, when is decreased, a first order phase transition appears: Below a certain critical value that depends on the temperature , the dynamical trajectories end up either in cyclic attractors (the networks exhibits SPR), or in stable mixed states that consist of combinations of the desired patterns. However, no coexistence of AM and SPR was found. Afterwards, in Ref. [13] the authors studied the CS model using correlated patterns. They found that it is possible to have SPR (the stable cycle limit is still present) when the correlation between the patterns is small. In the case of AM the network goes to a fixed point attractor but this attractor does not coincide with any of the desired patterns. Metz and Theumann [14] , [15] presented a full study of the stability of the patterns in a multi-layered neural network with competition between AM and SPR, finding the phase space regions where the network performs AM, SPR and the region for the spin-glass solutions (SGS), but no coexistence between AM and SPR was found either. By “coexistence” of AM and SPR we mean that some of the patters are stored as fixed point attractors while other patterns are stored in larger cyclic attractors. When such a coexistence does not exist, then all patterns are stored either as independent fixed point attractors or as a single large cyclic attractor. Recently, the same authors in [16] studied a model similar to the CS model by means of the generating functional technique. They present a study of the stationary states and the different regions on the phase space where either fixed points or cyclic attractors are attained. It is important to note that the lack of coexistence of AM and SPR in the original CS model is not a trivial result. The synaptic weights are a weighted mixture of symmetric and asymmetric matrices. Therefore, especially for intermediate values of the mixture parameter , it could have happened that some of the patterns were recognized as fixed-point attractors whereas some other patterns were recognized in sequential order. But this was not the case: either all the patterns are fixed point attractors or all of them form a huge cyclic attractor (remember that we are working in the saturated regime where the number of stored patters is a finite fraction of the number of neurons: , which becomes infinite in the thermodynamic limit ). This all-or-none behavior was not expected for the original CS model and deserved a careful study carried out by several authors. Using the generating functional formalism developed in Refs. [17] , [18] , [19] , [20] – [21] we investigate how the network transits from pure AM to pure SPR for the CS model, first in the case in which the symmetric and asymmetric parts of the synaptic matrix are correlated (defined in terms of the same set of patterns) and then when they are independent (defined each in terms of different sets of patterns). Both cases are studied for systems in which the number of patterns is a finite fraction of the total number of neurons , namely, in the saturated regime for which even in the thermodynamic limit . We compute the phase space over the parameters , and the temperature , finding the regions where the networks performs AM and/or SPR, as well as the spin-glass region, for different values of the mixing parameter . As might have been expected, we find that when and are correlated, the network either performs AM or SPR, but it is incapable to perform both for the same set of parameter values. In contrast, when and are independent of each other, AM and SPR coexist within a large region of the parameter space. We present a complete explicit characterization of the different phases, as well as the transition between AM and SPR when these behaviors coexist. In the next section we present the two versions of the CS model we study in this work. In Sec. we compute the dynamical equations that determine the temporal evolution of the network using the generating functional formalism developed in Refs. [17] , [18] , [19] , [20] . In Sec. we present the results for the original CS model and determine the structure of the parameter space identifying the regions of highly correlated, weakly correlated and spin-glass solutions. We do this for the AM and SPR dynamics and show that these two types of pattern retrieval do not coexist. In Sec. we present analogous results but for the modified version of the CS model, and we show that in this case AM and SPR dynamics do coexist. Finally, in Sec. we summarize our results.",
"discussion": "Discussion We have obtained the complete phase diagram and analyzed the transitions from AM to SPR in a neural network model proposed by Coolen and Sherrington in which both types of pattern retrieval compete, as well as in a simple modification of the model in which both types of pattern retrieval may coexist. In these systems, the AM and SPR dynamics are encoded in the symmetric and asymmetric parts of the synaptic matrix, respectively, and the contribution of each of these parts is weighted by a parameter in such a way that when only the symmetric part survives, whereas when only the asymmetric part is present. In the original Coolen and Sherrington model, the same set of patterns are used to define the symmetric and asymmetric parts of the synaptic matrix. Using the standard functional generating formalism, we obtained the phase diagram of the system which shows that in the original Coolen-Sherrington model AM and SPR dynamics cannot coexist. This is simply due to the fact that a given pattern cannot be a fixed point and part of a larger cyclic attractor at the same time. Therefore, the original CS model can retrieve patterns only for the limiting cases or . For intermediate values of the network is “frustrated” and can perform neither AM nor SPR. To prevent the system from falling into a “frustrated” state as mentioned above, we modified the CS model by using two independent sets of patterns in order to define separately the symmetric and asymmetric parts of the synaptic matrix. In doing so we allow the possibility for the network to have fixed points belonging to one set of patterns, and simultaneously cyclic attractors constructed with the patterns that belong to the other set. Our goal was to determine how the network transits from AM to SPR as varies from 1 to 0 in this new case where the two sets of patters were independent. As expected, in this case the AM and SPR dynamics coexist for a wide range of values of . However, some other aspects of the model can be analyzed. For instance, quasi periodic states are known to occur in the original CS model (with only one set of patterns) and it would be interesting to determine to what extent these quasi periodic states exist in the modified CS model (with two independent sets of patterns). Also, it is possible to have an intermediate situation in which the two sets and share some of the patterns. In this case the two sets would not be fully independent and the transition from AM to SPR dynamics could be more complicated. Finally, the generating functional approach that we used to determine the structure of the phase space works very well when the patterns are uncorrelated and the network is fully connected. It would also be interesting to extend the analysis to networks with more realistic topologies, such as the small-world and scale-free topologies, in order to determine how the network topology affects the dynamics."
} | 3,442 |
27386515 | PMC4928996 | pmc | 5,583 | {
"abstract": "We identify the first quantitative trait loci for antioxidant capacity in corals, providing possible new avenues for management and restoration approaches.",
"introduction": "INTRODUCTION The impacts of global climate change, ocean acidification, and other anthropogenic disturbances are growing concerns for coral reef ecosystems ( 1 – 3 ). Coral bleaching [that is, the loss of obligate dinoflagellate photosymbionts, Symbiodinium spp., and/or their photopigments from coral tissues ( 4 )] is a common stress response in corals; it results from a variety of factors, including high and low temperatures ( 5 ), high irradiance levels, low salinity, sedimentation, pollution, and herbicides ( 6 ). Extended periods of higher-than-usual summer temperatures have caused mass coral bleaching events and have led to considerable coral mortality worldwide. The capacity of corals to resist thermal bleaching is reduced when they are also exposed to high nutrient levels associated with terrestrial runoff ( 7 ). Furthermore, terrestrial runoff on its own is also known to have a negative impact on coral health and can cause coral bleaching ( 8 ). Similar to coral reefs in other regions, the Great Barrier Reef (GBR) is experiencing environmental challenges due to land modification in adjacent coastal areas, which leads to discharges of sediments, chemicals, and nutrients into nearshore waters ( 9 ). Intraspecific variation in bleaching tolerance thresholds is common both among and within coral populations ( 10 ). This can be attributed to acclimatization and adaptation to local environments by the coral host, as well as endosymbiotic dinoflagellates and other microbial symbiont communities ( 11 , 12 ). Unveiling factors underpinning intraspecific variation in coral stress tolerance will not only enhance our understanding of biological functioning but also provide information that is relevant to coral reef management and restoration approaches. For instance, the identification of coral stress tolerance genes will allow relatively tolerant wild colonies to be selected for fragmentation, reseeding, and restoration ( 13 ) based on a simple genotyping assay. Alternatively, such colonies could be used for selective breeding to rear offspring with enhanced environmental stress tolerance ( 14 ) or for assisted translocation ( 15 – 17 ). Divergent selection pressures exerted by contemporary environmental gradients provide researchers an excellent opportunity to examine the genetic basis of intraspecific phenotypic variation in adaptive traits. The GBR extends over 14° latitude, providing an extensive north-to-south temperature gradient, and also spans a cross-shelf (west to east) water quality gradient associated with proximity to agriculture on adjacent coastlines and terrestrial runoff during the wet season through several large river systems ( 9 ). Here, we use gene-by-environment and genotype-by-phenotype association analyses to identify quantitative trait loci (QTLs) for antioxidant capacity or environmental stress tolerance in the genome of the coral animal Acropora millepora .",
"discussion": "DISCUSSION The three independent data sets obtained in our study (summarized in Table 1 ) provide compelling evidence that C29226S281 and C70S236 are true QTLs; the presence of G and T alleles at these loci, respectively, is indicative of a relatively higher antioxidant capacity and tolerance to temperature stress and/or low water quality, especially in the homozygous state. Whereas the 3 to 6% higher CoQH 2 levels in colonies of the GG and TT genotypes for C29226S281 and C70S236, respectively, might seem insubstantial as compared to other genotypes in both 27° and 32°C experimental treatments ( Fig. 2B ), this difference is considerable and biologically relevant given that severe coral bleaching causes an approximately 10% drop in CoQH2% ( 23 ). Both genetic markers are located in genes that play a role in the ubiquitination process, which is known to be involved in the coral thermal stress response ( 20 , 27 ). Frontloading of stress tolerance genes (such as heat shock proteins) and antioxidant enzymes, as demonstrated in our laboratory experiment, has also been observed in a transcriptomic study of experimentally heat-stressed colonies of the congener Acropora hyacinthus in American Samoa ( 27 ). Table 1 Summary of results. Main results from the three independent data sets obtained in this study. Data set Locus C70S236 Locus C29226S281 Gene-by-environment correlation (poor water quality, high SST range, and low mean SST) Higher frequency of T allele Higher frequency of G allele 2006 nonbleached versus bleached corals (temperature stress) 12% higher frequency of T allele in nonbleached colonies No difference 2009 nonbleached versus bleached corals (salinity and turbidity stress) No difference 28% higher frequency of G allele in nonbleached colonies Experimental heat stress Higher antioxidant capacity (CoQH 2 ) in TT genotypes: 35.2% explained by genotype Higher antioxidant capacity (CoQH 2 ) in GG genotypes: 14.6% explained by genotype Resistance to photosynthetic damage ( F v / F m ) in TT genotypes: 10.5% explained by genotype A number of invertebrates, including oysters ( 28 ) and corals ( 29 , 30 ), are known to recruit the same pathways in response to a range of environmental stressors as well as pathogen infection. This response involves a defense against ROS ( 2 ) (for example, antioxidant and chaperone proteins), apoptosis, cytoskeleton reorganization, and the innate immune response ( 27 , 31 ). High temperature and poor water quality commonly lead to oxidative stress and subsequent bleaching due to a reduction in Symbiodinium photosynthetic efficiency caused by increased Symbiodinium density ( 32 , 33 ) and by photoinhibition of photosystem II ( 4 ). High levels of ROS may trigger the coral host innate immune response, leading to high levels of the reactive nitrogen species, nitric oxide, a common immune pathway in animals ( 34 ). The two markers identified in our study are QTLs for antioxidant capacity, and their role in the coral stress response to both high and low temperatures as well as poor water quality is in line with our current understanding of common invertebrate responses to environment- and pathogen-induced stress. Nevertheless, our results suggest that some level of response specificity of the markers may exist. The laboratory heat stress experiment demonstrated that the effect size of C70S236 on CoQH 2 and F v / F m (antioxidant capacity and photochemical efficiency) was larger than that of C29226S281. Therefore, C70S236 may have a greater influence on temperature-related stress tolerance. This is consistent with the gene-by-environment association analysis that showed a significant correlation between allele frequencies at C70S236 and range in SST ( Fig. 1 ). Furthermore, the natural bleaching data demonstrated that the bleaching tolerance of colonies collected in the summer that was characterized by high temperature (2006) was significantly associated with allele frequencies at the C70S236 locus, whereas C29226S281 showed allelic associations with bleaching conditions caused by poor water quality ( Fig. 2A and fig. S3). This also explains the smaller (but still significant) genotypic effects on tolerance to heat stress at the C29226S281 locus in the laboratory experiment. We predict that these two SNP loci will represent markers for tolerance to a range of other environmental stressors that are known to lead to cellular oxidative stress. Environmental variables related to temperature and water quality have cumulative effects on coral fitness ( 3 , 35 , 36 ). However, mean SST and NO 3 concentration at the sampling sites in this study have an inverse relationship, suggesting that the populations investigated are exposed to either high mean SSTs or high NO 3 , rather than to both simultaneously. In contrast, range in SST is positively correlated with NO 3 concentration across sampling locations, indicating that locations with poor water quality (that is, high nutrient concentrations) are characterized by wider temperature fluctuations ( Fig. 1 ). Four populations (Halftide Rocks, Magnetic Island, Humpy Island, and Halfway Island Reefs) that showed allelic differentiation from the remainder of the populations at the two loci are characterized by both high temperature fluctuation and poor water quality (as represented by high chlorophyll and NO 3 concentrations) (fig. S2). Furthermore, a multiple regression model that included NO 3 concentration and range in SST showed a significant fit to the allele frequency data for both loci (C29226S281 and C70S236). This increase in the goodness of fit of the model, despite a nonsignificant and small effect of range in SST when analyzed independently (table S2), suggests that there may be synergistic effects between selection driven by NO 3 concentration and range in SST. Finding different allele frequencies in thermally variable environments is consistent with previous experimental evidence demonstrating that corals from fluctuating habitats are more tolerant to temperature stress ( 37 ). Whereas phenotypic traits are not necessarily under the control of the QTL itself [that is, they can instead be regulated by genes in the same linkage group ( 38 )], the two SNP markers identified in our study are representative of the hard-wired genetic components underpinning phenotypic variation and can be used in a wide range of applications relevant to coral reef management and restoration. Extrinsic factors, such as temperature and water quality, are critical drivers of the degradation of coral populations. Spatial mapping of the C29226S281 and C70S236 genotypes can provide high-resolution data, which are used to predict the environmental stress susceptibility of corals and allow the identification of resilient and susceptible populations and individuals. Additional QTLs can be developed to provide more confidence in such initiatives. In combination with the information on connectivity among populations, this will enable the identification of key targets for conservation, such as highly resilient populations that have the ability to seed surrounding reefs. Furthermore, active human interference through assisted migration and selective breeding may be necessary to facilitate the survival of coral reefs in the future ( 14 ). Individuals with double homozygous genotypes at the two genetic markers, such as those identified in our study, can be targeted to select stress-tolerant brood stock for translocation and selective breeding to restore highly damaged reefs. A great advantage of the use of genetic markers for identifying resilient corals, rather than relying on historical environmental averages, is the greater resolution associated with it (that is, to the colony rather than reef level). Detection of genetically determined phenotypic variants within the population minimizes issues often encountered in transplant studies, such as acclimatization-associated trade-offs and environmentally regulated heritable factors ( 11 ). In conclusion, our study has revealed two QTLs for antioxidant capacity and environmental stress tolerance in reef-building corals. The identification of stress-tolerant genotypes will facilitate exploration of new management and restoration options for the world’s rapidly degrading coral reefs."
} | 2,858 |
34594312 | PMC8478078 | pmc | 5,586 | {
"abstract": "In natural and agricultural ecosystems, survival and growth of plants depend substantially on residing microbes in the endosphere and rhizosphere. Although numerous studies have reported the presence of plant-growth promoting bacteria and fungi in below-ground biomes, it remains a major challenge to understand how sets of microbial species positively or negatively affect plants’ performance. By conducting a series of single- and dual-inoculation experiments of 13 plant-associated fungi targeting a Brassicaceae plant species ( Brassica rapa var. perviridis ), we here systematically evaluated how microbial effects on plants depend on presence/absence of co-occurring microbes. The comparison of single- and dual-inoculation experiments showed that combinations of the fungal isolates with the highest plant-growth promoting effects in single inoculations did not have highly positive impacts on plant performance traits (e.g., shoot dry weight). In contrast, pairs of fungi with small/moderate contributions to plant growth in single-inoculation contexts showed the greatest effects on plants among the 78 fungal pairs examined. These results on the offset and synergistic effects of pairs of microbes suggest that inoculation experiments of single microbial species/isolates can result in the overestimation or underestimation of microbial functions in multi-species contexts. Because keeping single-microbe systems under outdoor conditions is impractical, designing sets of microbes that can maximize performance of crop plants is an important step for the use of microbial functions in sustainable agriculture.",
"introduction": "Introduction Plants in natural and agricultural ecosystems are associated with diverse taxonomic groups of microbes, forming both positive and negative interactions with the microbiomes ( Lundberg et al., 2012 ; Peay et al., 2016 ; Busby et al., 2017 ; Toju et al., 2018b ). In particular, bacteria and fungi found within and around root systems have been reported as key determinants of plants’ survival and growth ( Hiruma et al., 2016 , 2018 ; Castrillo et al., 2017 ; Trivedi et al., 2020 ). A number of rhizosphere bacteria, for example, are known to stimulate plants’ growth by producing phytohormones ( Lugtenberg and Kamilova, 2009 ; Bhattacharyya and Jha, 2012 ; Finkel et al., 2020 ). Mycorrhizal fungi are ancient symbionts of land plants ( Remy et al., 1994 ; Taylor et al., 1995 ), providing soil phosphorus and/or nitrogen to their hosts ( Richardson et al., 2009 ; Tedersoo et al., 2010 ; Jansa et al., 2019 ). Moreover, a growing number of studies have shown that diverse clades of endophytic and soil fungi support host plants by provisioning inorganic/organic forms of nutrients ( Usuki and Narisawa, 2007 ; Newsham, 2011 ; Hiruma et al., 2016 ), activating plant immune systems ( van Wees et al., 2008 ; Pieterse et al., 2014 ), and suppressing populations of pathogens/pests in the rhizosphere ( Narisawa et al., 2004 ; Khastini et al., 2012 ; Gu et al., 2020 ). Thus, developing scientific bases for maximizing the benefits from those plant-associated microbiomes is an essential step for fostering sustainable crop production and restoring forest/grassland ecosystems ( Bulgarelli et al., 2013 ; Carlström et al., 2019 ; Wagg et al., 2019 ; Saad et al., 2020 ). One of the major challenges in utilizing plant-associated microbial functions is to design sets of microbial species/isolates ( Vorholt et al., 2017 ; Paredes et al., 2018 ; Toju et al., 2018a ; Wei et al., 2019 ). While a single microbial species or isolate can have specific functions in promoting plant growth, broader ranges of positive effects on plants are potentially obtained by introducing multiple microbial species/isolates ( Wang et al., 2011 ; Ważny et al., 2018 ; He et al., 2020 ). For example, a fungal species degrading organic nitrogen ( Newsham, 2011 ) and another fungus suppressing soil pathogens ( Vinale et al., 2008 ) may collectively provide plants with a broader spectrum of physiological functions than each of them alone, potentially having additive or synergistic effects on the growth of their hosts. Meanwhile, sets of microbes trying to colonize the plant endosphere or rhizosphere may compete for resources/space ( Kennedy et al., 2009 ; Werner and Kiers, 2015 ; Toju et al., 2016 ) or inhibit each other’s growth ( Helfrich et al., 2018 ), making their impacts on host plants more negative than that observed under single-inoculation conditions (i.e., offset effects) ( Nelson et al., 2018 ). Given that multiple microbial species inevitably interact with a single plant in agroecosystems ( Toju et al., 2018a ), knowledge of those synergistic and offset effects in plant-associated microbiomes is crucial for optimizing microbial functions in crop production. A starting point for designing sets of microbes is to use the information of single-inoculation assays, in each of which a single microbial species/isolate is introduced to a target plant species/variety ( Ahmad et al., 2008 ; Harbort et al., 2020 ). Through this initial assay, respective species/isolates are scored in terms of their functions (e.g., plant-growth promotion effects) under single-inoculation conditions ( Nara, 2006 ; Dai et al., 2008 ; Taurian et al., 2010 ; Tsolakidou et al., 2019 ). The next step is to consider how these single-inoculation scores can be used for designing sets of microbes that potentially promote plant growth in synergistic ways. As the number of combinations inflates with that of constituent species/isolates [e.g., {N × (N – 1)}/2 combinations in two-species systems], prioritizing candidate species/isolate combinations based on single-inoculation results is an important step ( Paredes et al., 2018 ; Toju et al., 2018a , 2020 ). The simplest way of exploring best sets of microbes is to combine microbes with highest single-inoculation scores. This strategy of combining microbes in highest ranks is promising if synergistic (or additive) effects are common in plant-associated microbiomes. In contrast, if offset effects of multiple microbes on plant performance are ubiquitous, alternative strategies for exploring species/isolate combinations are required to maximize benefits from plant-associated microbiomes. Thus, knowledge of the prevalence and intensity of such synergistic and offset effects is essential in synthetic microbiome studies. Nonetheless, although there have been studies reporting synergistic/offset effects of multiple plant-associated microbes ( Han and Lee, 2006 ; Wang et al., 2011 ; Ważny et al., 2018 ; He et al., 2020 ), experimental studies systematically evaluating the commonness of those effects are scarce. In this study, we tested the hypothesis that synergistic effects on plant growth are common in below-ground fungal biomes in a series of single- and dual-inoculation experiments. By using 13 plant-associated fungal species belonging to various taxonomic groups, we first evaluated their basic effects on plant growth in a single-inoculation experiments with a Brassicaceae species ( Brassica rapa var. perviridis ). We also performed dual-inoculation experiments for all the 78 possible combinations of the fungal species and then evaluated the performance of the combinations in light of single-inoculation results. The data then provided a platform for testing whether plant-growth promoting effects exceeding those of all the single-inoculation conditions are attainable under dual-inoculation conditions. Overall, this study provides a basis for understanding to what extent plant-growth promotion effects of microbiomes can be expected from the information of single-species inoculations, illuminating the potential importance of “non-additivity” in multi-microbe contexts.",
"discussion": "Discussion By using taxonomically diverse plant-associated fungi, we here evaluated plant-growth promoting effects of pairs of fungal isolates in light of those observed in single-isolate inoculation experiments. The 13 fungal isolates differed greatly in their independent effects on Brassica plants ( Figures 2 , 3 ), providing an ideal opportunity for examining how the ranking of plant-growth promoting effects in single-inoculation contexts were related to that in multi-species (dual-inoculation) contexts ( Figures 4 – 6 ). Such information of synergistic and offset effects in the presence of multiple microbial species is indispensable for understanding to what extent we can predict functions of microbial communities (microbiomes) from the datasets of single-species/isolate screening. A series of single- and dual-inoculation experiments indicated that greater performance of plants were potentially obtained in multi-species than in single-species contexts ( Figure 2 ). This result, itself, is consistent with previous reports of enhanced plant growth by specific pairs of bacteria/fungi ( Han and Lee, 2006 ; Wang et al., 2011 ; Ważny et al., 2018 ; He et al., 2020 ). Meanwhile, our experiments on 78 combinations of fungi systematically suggested that pairs of microbes, each of which had greatly positive impacts on plant growth in single inoculations, could show minor effects on plants under multi-species conditions. For example, the strategy of combining the two species/isolates with highest performance in the single inoculation experiments (i.e., V. simplex and Alternaria sp. KYOCER00001239) did not result in high plant-growth promoting effects ( Figure 3 ): rather, offset effects were observed in those “top × top” pairs ( Figures 4 – 6 ). Thus, biological functions at the community (microbiome) level may be rarely maximized by the “bottom-up” exploration of sets of microbes based solely on single-inoculation experiments. Our experiments also suggested that pairs of microbes with subordinate performance in single inoculation assays could show largest growth-promoting effects on plants ( Figure 2 ). This result suggests that single-species/isolate screening does not always provide sufficient information for predicting microbial performance at the multi-species level ( Toju et al., 2018a ). Interestingly, the fungal pairs with highest synergistic effects in our experiment involved fungi in the genera Fusarium and Curvularia ( Figure 4A ), which were often described as plant pathogenic taxa ( Michielse and Rep, 2009 ; Ma et al., 2013 ; Manamgoda et al., 2015 ). Basically, physiological effects on plants vary remarkably among species/isolates within taxa as evidenced by the presence of Fusarium and Curvularia species enhancing plant health and growth ( Olivain et al., 2006 ; Nahalkova et al., 2008 ; Priyadharsini and Muthukumar, 2017 ). In fact, the Fusarium and Curvularia isolates examined in our study had positive effects on Brassica plants even in the single-inoculation assays ( Figure 2 ). Moreover, the results of the dual inoculation experiments suggested that some fungi in these predominantly plant-pathogenic genera can have even greater positive effects on plants in combination with specific other fungi ( Figures 2 , 3 ). Our results on synergistic effects in multi-species contexts further illuminate the potential use of diverse endosphere/rhizosphere microbes whose biological functions have been underestimated in conventional screening of single inoculations. The fact that microbial functions critically depend on combinations of microbial species/isolates highlight the importance of “bird’s-eye” views of designing microbiomes. Given that microbial functions at the community (multi-species) levels are not the simple sums/averages of functions in single-species contexts ( Figures 2 , 7 ), research strategies taking into account not only each microbe’s roles but also the nature of microbe–microbe interactions will provide platforms for optimization of microbiome functions ( Agler et al., 2016 ; Toju et al., 2016 ; Banerjee et al., 2018 ). In this respect, interdisciplinary approaches integrating the observational, genomic, and metagenomic information of microbial functions ( Bulgarelli et al., 2015 ; Levy et al., 2018 ; Ichihashi et al., 2020 ) with community ecological analyses of species interaction networks ( Agler et al., 2016 ; van der Heijden and Hartmann, 2016 ; Toju et al., 2017 ) will help us explore highly functional and stable microbial sets among numerous candidate combinations of species ( Paredes et al., 2018 ; Saad et al., 2020 ; Toju et al., 2020 ). In other words, information of microbial functions in single-species contexts is utilized by being combined with insights into dynamics and processes within microbiomes. While the experiments conducted in this study provided a unique opportunity for systematically evaluating synergistic/offset effects of microbes on plants, the obtained datasets should be interpreted with caution given the following limitations. First, physiological mechanisms by which the examined fungi affected plant growth were unexplored in the current study. Although detailed physiological and/or molecular biological investigations have been done for some of the fungal species used in this study [e.g., C. tofieldiae ( Hiruma et al., 2016 ), V. simplex ( Guo et al., 2018 ), and C. chaetospira ( Harsonowati et al., 2020 )], metabolites and genes involved in the plant–fungus interactions are unknown for the remaining species. For more mechanistic understanding of interactions involving plants and multiple microbial species, we need to perform transcriptomic analyses targeting plants’ responses to each microbe as well as those comparing plants’ gene expression patterns between single- and multiple-symbiont conditions. Comparative transcriptomic analyses across experiments with different environmental conditions (e.g., soil nutrient concentrations) will provide essential insights into microbial functions as well. Second, the inoculation test based on single plant species precluded us from understanding how general synergistic/offset effects existed in plant–fungal biome interactions. Although some of the fungal taxa used in this study have been reported to interact with multiple families of plants ( Hermosa et al., 2012 ; Toju et al., 2018b ), impacts of endophytic/soil fungi on plants can vary depending on plant taxa and environmental conditions ( Kiers et al., 2011 ; Pineda et al., 2013 ; Rudgers et al., 2020 ). Therefore, to gain more robust insights into synergistic/offset effects in interactions of plants and multiple microbial species/isolates, the reproducibility of the patterns observed in this study should be examined in inoculation experiments targeting diverse other plant species. Third, it is important to acknowledge that the complexity of the microbial sets examined in this study is minimal (i.e., two fungal species): different types of phenomena may be observed in combinations of three or more bacterial/fungal species ( Durán et al., 2018 ; Paredes et al., 2018 ; Carlström et al., 2019 ; Wei et al., 2019 ). Moreover, it remains to be examined how we can increase microbial functions (e.g., host plant growth rates) by increasing the number of microbial species/isolates. The presence of microbial pairs outperforming single-microbe systems ( Figure 2 ) leads to the working hypothesis that compatible sets of three or more microbial species yield greater functions than simpler communities by playing complementary roles. Meanwhile, it is expected that benefits of microbiomes do not increase linearly with increasing number of microbial species (i.e., saturating curves of benefits against increasing number of microbes) ( van der Heijden et al., 1998 ), at least in terms of specific functions such as provisioning of soil phosphorus or blocking of soil pathogens. We here showed that screening based on inoculations of single microbial species/isolates can result in the underestimation of the microbes that potentially have large plant-growth promoting effects in combinations with specific other microbes. Given that plants are inevitably associated with hundreds or more of microbial species in agricultural and natural ecosystems ( Lundberg et al., 2012 ; Schlaeppi and Bulgarelli, 2015 ; van der Heijden and Hartmann, 2016 ; Thoms et al., 2021 ), such nonlinearity found in microbe–microbe associations deserve future intensive research. It may be important, for instance, to examine how antagonistic relationships between salicylic-acid- and jasmonic-acid-related plant physiological responses ( Niki et al., 1998 ), which are activated by different types of bacteria/fungi ( Robert-Seilaniantz et al., 2011 ) [but see Betsuyaku et al. (2018) ], can result in such nonlinear effects of multiple microbes on plant performance. Interdisciplinary studies on relationships between microbiome compositions and their ecosystem-level functions are awaited toward the maximization of microbial functions for sustainable agriculture and ecosystem restoration."
} | 4,248 |
21278815 | null | s2 | 5,587 | {
"abstract": "Amphiphlic block copolymers consisting of hydrophilic, poly(acrylic acid) randomly decorated with acrylate groups and hydrophobic, rubbery poly(n-butyl acrylate) self-assembled into well-defined micelles with an average diameter of ~21 nm. Radical polymerization of acrylamide in the presence of the crosslinkable micelles gave rise to hybrid, elastomeric hydrogels whose mechancial properties can be readily tuned by varying the BCM concentration."
} | 112 |
24625633 | PMC3953072 | pmc | 5,588 | {
"abstract": "In recent years, there has been an increased interest in the research and development of sustainable alternatives to fossil fuels. Using photosynthetic microorganisms to produce such alternatives is advantageous, since they can achieve direct conversion of carbon dioxide from the atmosphere into the desired product, using sunlight as the energy source. Squalene is a naturally occurring 30-carbon isoprenoid, which has commercial use in cosmetics and in vaccines. If it could be produced sustainably on a large scale, it could also be used instead of petroleum as a raw material for fuels and as feedstock for the chemical industry. The unicellular cyanobacterium Synechocystis PCC 6803 possesses a gene, slr2089 , predicted to encode squalene hopene cyclase (Shc), an enzyme converting squalene into hopene, the substrate for forming hopanoids. Through inactivation of slr2089 ( shc ), we explored the possibility to produce squalene using cyanobacteria. The inactivation led to accumulation of squalene, to a level over 70 times higher than in wild type cells, reaching 0.67 mg OD 750 \n −1 L −1 . We did not observe any significant growth deficiency in the Δ shc strain compared to the wild type Synechocystis , even at high light conditions, suggesting that the observed squalene accumulation was not detrimental to growth, and that formation of hopene by Shc is not crucial for growth under normal conditions, nor for high-light stress tolerance. Effects of different light intensities and growth stages on squalene accumulation in the Δ shc strain were investigated. We also identified a gene, sll0513 , as a putative squalene synthase in Synechocystis , and verified its function by inactivation. In this work, we show that it is possible to use the cyanobacterium Synechocystis to generate squalene, a hydrocarbon of commercial interest and a potential biofuel. We also report the first identification of a squalene hopene cyclase, and the second identification of squalene synthase, in cyanobacteria.",
"conclusion": "Conclusions We have shown that it is possible to use the cyanobacterium Synechocystis to generate squalene, a hydrocarbon of commercial interest and a potential raw material for biofuels. We also demonstrate the first identification and functional verification of an active squalene hopene cyclase in cyanobacteria. Inactivating the shc gene in Synechocystis led to accumulation of squalene to a level about 70 times higher than in the wild type. No growth impairments were detected in the engineered strain. Furthermore, we identified a gene putatively encoding squalene synthase, and could show that inactivation of this gene abolished squalene synthesis in the cells. The isoprenoid biosynthesis is a source of many useful compounds with wide application in biotechnology, medicine, chemistry, as food additives, and potentially as fuels. For large scale generation of useful isoprenoid compounds, it would be preferable to use photoautotrophic microorganisms such as cyanobacteria as hosts, since they can grow photosynthetically using solar energy, water and carbon dioxide from the atmosphere to generate the desired product. The squalene producing strain of Synechocystis generated in this study serves as an example of such a production system. Improvements of the system using metabolic engineering techniques are possible and will be addressed in future work, as well as further investigation of the native metabolism that leads to squalene production.",
"introduction": "Introduction Isoprenoids, also called terpenoids, are a large family of compounds including carotenoids, tocopherol, phytol, sterols and hormones. In most prokaryotes, in algae, and in plant plastids, isoprenoids can be produced via the methyl-erythritol-4-phosphate (MEP) pathway ( [1] , see Fig. 1 ). This pathway was first characterized in Escherichia coli , and uses pyruvate and glyceraldehyde 3-phosphate as substrates [2] to form, in several steps, the products isopentenyl diphosphate (IPP) and dimethylallyl diphosphate (DMAPP) (see Fig. 1 ). IPP and DMAPP are the building blocks for all other isoprenoids, some of which have useful commercial applications in nutrition, medicine, chemistry, and potentially as biofuels. 10.1371/journal.pone.0090270.g001 Figure 1 Squalene biosynthesis pathway through the MEP-pathway in Synechocystis . Abbreviations used: G3P = glyceraldehyde 3-phosphate; DXP = deoxyxylulose 5-phosphate; MEP = methylerythritol 4-phosphate; IPP = isopentenyl diphosphate; DMAPP = dimethylallyl diphosphate; GPP = geranyl diphosphate; FPP = farnesyl diphosphate; PSPP = presqualene diphosphate. For enzymes: Ipi = isopentenyl diphosphate delta isomerase; CrtE = geranylgeranyl pyrophosphate synthase; Sqs = squalene synthase; Shc = squalene hopene cyclase. Based on data from [3] , [5] , [6] , [7] , [8] . The sequenced genome of the unicellular cyanobacterium Synechocystis sp. PCC 6803 (from here on referred to as Synechocystis ) [3] contains all the genes needed to encode the enzymes involved in the MEP pathway in E. coli \n [4] (see Fig. 1 ). Only a few studies have investigated this pathway in cyanobacteria [5] , [6] , [7] , [8] , despite the fact that it is the origin of many interesting and potentially useful compounds. Recently, it was shown that Synechocystis can be used for production of isoprene, a small (C 5 H 8 ) volatile hydrocarbon [9] , and for photosynthetic generation of β -phellandrene (C 10 H 16 ), an essential oil [10] , as a side reaction to the isoprenoid biosynthesis. We are now interested in investigating if it is possible to use cyanobacteria for generation of longer-chain isoprenoid hydrocarbons. Using cyanobacteria for direct production of a biofuel is advantageous, since they can grow photosynthetically on minimal media, fixing carbon dioxide from air and using sunlight as an energy source to generate the product. The isoprenoid squalene is a 30-carbon pure hydrocarbon, the formation of which is catalyzed by the enzyme squalene synthase. Squalene synthase performs a two-step reaction, where two molecules of farnesyl-diphosphate (FPP) are first combined to form presqualene diphosphate (PSPP), which is subsequently converted into squalene, in a NADPH-dependent step ( [11] , Fig. 1 ). The mechanism of this reaction has been thoroughly investigated, primarily in eukaryotes [12] , [13] . Today, commercial uses of squalene include as an ingredient in cosmetic products and in vaccines, as an additive in some adjuvant formulations, but if it could be produced sustainably and in large quantities, it could also be used as a raw material for biofuels and as feedstock for the chemical industry. In a wide range of bacteria, squalene is used as the substrate for formation of hopene, a complex pentacyclic hydrocarbon which is further modified to form hopanoids [14] , [15] . The enzyme catalyzing the formation of hopene from squalene, squalene hopene cyclase (Shc) has been characterized in a number of organisms [16] , [17] , [18] , [19] , [20] , and the structure of Shc from Alicyclobacillus acidocaldarius has been determined [21] . Presence of hopanoids in the outer membrane and in the thylakoid membranes have been observed in the cyanobacterium Synechocystis PCC 7614 [22] , however, to our knowledge, no investigation has yet been carried out regarding production of squalene, or its use in the cell by the action of squalene hopene cyclase, in cyanobacteria. In this study, we have generated a squalene-producing strain of the cyanobacterium Synechocystis . This was achieved by inactivating the gene slr2089 , putatively encoding the enzyme squalene hopene cyclase (Shc). Inactivation of this single gene leads to accumulation of squalene in the cell. In addition, we identified the gene encoding squalene synthase in Synechocystis .",
"discussion": "Results and Discussion Genes in Synechocystis Putatively Involved in Synthesis and Use of Squalene In the genome of Synechocystis \n [3] , it is possible to identify all the genes encoding the MEP pathway in E. coli , as has previously been noted [4] . Since the aim of this study was to investigate production of squalene in Synechocystis , we looked further for genes that could be involved in synthesis and utilization of squalene in this strain. One gene, slr2089 , can be identified as likely encoding squalene hopene cyclase (Shc) in Synechocystis . The putative Shc amino acid sequence is homologous to other Shcs in the databases, with identity/similarity of 43%/58% to the structurally known Shc from A. acidocaldarius (PDB: 2SQC_A), and contains known conserved motifs such as the catalytic aspartate identified in A. acidocaldarius \n [21] , a DXDD motif in the active site cavity important for the activity of the enzyme [23] , and repeated QW-motifs [24] . It exhibits the highest similarities to other putative Shcs in cyanobacteria (>60% identity, >70% similarity). However, shc does not appear to be universally present in cyanobacteria. Based on cyanobacterial genome sequences available in the Cyanobase ( http://genome.kazusa.or.jp/cyanobase/ ) and JGI ( http://img.jgi.doe.gov/cgi-bin/w/main.cgi ) databases, shc is present in about 45% of the sequenced strains (data not shown). This is in agreement with data for other organisms, where estimates of the distribution of hopanoid biosynthesis range from 4% of microorganisms in the oceans [25] to 50% of a set of cultured strains [26] . It is clear that the presence of shc and hopanoid biosynthesis is not universal and may be an unusual trait in the global microbiome. A blast search for squalene synthase in the Synechocystis genome resulted in identification of the gene sll0513 , annotated as encoding a hypothetical protein. The amino acid sequence of the sll0513 gene product (GenBank accession no BAA10820) shows similarities with squalene synthases in other organisms ( e.g. 52%/72% identity/similarity to farnesyl-diphosphate farnesyltransferase from Bacillus megaterium WSH-002, 31%/42% identity/similarity to Sqs from Botryococcus braunii (GenBank accession no AAF20201.1), 26%/42% identity/similarity to Sqs from Saccharomyces cerevisiae (ERG9, GenBank accession no AAA34597.1)), with the highest similarities to other cyanobacterial sequences (>60% identity, >80% similarity to a number of putative cyanobacterial squalene synthases). In the cyanobacterium Thermosynechococcus elongatus BP-1, squalene synthase, encoded by sqs, has been experimentally verified [27] . However, there are substantial differences between sll0513 and sqs in T. elongatus ; sll0513 encodes a 277 aa protein, whereas Sqs in T. elongatus is 359 aa, and their mutual identity/similarity is 29.5%/41.1%. sll0513 is also similar to phytoene synthases, however, in Synechocystis there is another gene, crtB , which has been shown to encode phytoene synthase [28] . Furthermore, the deduced amino acid sequence of sll0513 contains previously identified conserved domains common to squalene synthases [11] , [12] , [29] , including a putative NADPH binding site not present in phytoene synthase [30] (data not shown). The substrate for the squalene synthase, farnesyl diphosphate, is formed through linking of one molecule of IPP and one molecule of DMAPP to form geranyl-diphosphate, followed by addition of another molecule of IPP. In the Synechocystis genome, there is one gene, crtE ( slr0739 ), annotated as geranylgeranyl diphosphate synthase, which is likely to encode the enzyme that performs these steps. \n Figure 1 summarizes the proposed pathway for squalene synthesis and utilization in Synechocystis . Inactivation of shc in Synechocystis and Detection of an shc Transcript The gene slr2089 in the Synechocystis genome [3] , putatively encoding Shc, was inactivated by replacing a 606 bp region of the gene with a neomycin resistance cassette ( Fig. 2A ). The inactivated gene construct was transferred to Synechocystis via natural transformation to generate a Δ shc strain. Transformants were isolated by selection with appropriate antibiotics, and replacement of the wild type copy of the gene with the inactivated version was confirmed by PCR. Expected PCR fragments were amplified from the successful Δ shc inactivation strains ( Fig. 2B ). Furthermore, RNA was isolated from the wild type and Δ shc strains and used for detection of shc transcript in RT-PCR experiments. Transcripts could be detected in both wild type and Δ shc cells; however, amplification of transcripts from the deleted region resulted in a product only from the wild type strain ( Fig. 2C, left panel s). This shows that the gene is actively transcribed under standard photoautotrophic growth conditions in the wild type Synechocystis , and that while transcription of the gene is still active in the Δ shc strain, there is no intact transcript present. Amplification of 23S cDNA was used as a positive control ( Fig. 2C, right panel s). 10.1371/journal.pone.0090270.g002 Figure 2 Knock-out strategy and screening of mutant. (A) Schematic overview of the knock-out strategy for shc . A neomycin ( npt ) cassette is inserted into shc by homologous recombination thereby deleting a part of shc . (B) PCR screening of genomic DNA for checking segregation of the mutant compared to wild type. Primers used for screening were a-d where a = shc_usR_as; b = shc_middle_F; c = shc_middle_R; d = shc_dsF_as. (C) RT-PCR of Δ shc and wild type cDNA using reverse transcriptase (top) and without (bottom) as negative controls. Primers amplifying 23S were used as a control of equal loading of RNA. Sequencing of genomic DNA from the Δ shc strain was done to further verify the inactivation, and the results reaffirmed that the antibiotic cassette was positioned inside the shc gene (data not shown). Extraction and Detection of Squalene in the Δ shc and Wild Type Strains After inactivation of shc , we hypothesized that squalene may be accumulating in the cells. To investigate this possibility, a method for extraction and detection of squalene from Synechocystis was developed, based on the method for total lipid extraction by Bligh and Dyer [31] (see Materials and Methods for details). Total lipids were extracted from cultured cells using methanol and chloroform, the resulting lipids were dissolved in heptane, and squalene content was determined using HPLC ( Fig. 3A ), by comparison to a commercially available squalene standard. To further verify the identity of the squalene peak observed by HPLC of cyanobacterial lipid extracts, the peak was eluted and analyzed by GC-MS ( Fig. 3A ). In both wild type and Δ shc , it was found that the eluted peak exhibited the correct fragmentation as compared to a squalene standard ( Fig. 3B ). However, only minimal amounts of squalene could be detected in the wild type, confirming our results from HPLC. Synechocystis cells with shc inactivated grown to stationary phase had a squalene content of 0.67±0.102 mg OD 750 \n −1 L −1 while the wild type contained 0.0093±0.0031 mg OD 750 \n −1 L −1 ( Fig. 3C ). Thus, squalene accumulated in the Δ shc strain to a level more than 70 times the level in the wild type. This result, together with the RT-PCR results showing active transcription of slr2089 , suggests that slr2089 does indeed encode a functional squalene hopene cyclase, and also that if there are other enzymes in Synechocystis that may use squalene as a substrate, they do not consume all squalene produced under the conditions tested. 10.1371/journal.pone.0090270.g003 Figure 3 Analysis of squalene accumulation. Squalene was extracted from wild type and Δ shc cultures and the extracts analyzed by HPLC using a squalene standard for identification and quantification of the squalene peak. The identity of the squalene peak was confirmed using GC-MS. (A) HPLC chromatograms (left panels) and GC-MS chromatograms (right panels) for the detection of squalene. A squalene standard (top) was compared with cell extracts from wt (middle) and Δ shc mutant (bottom). (B) Fragmentation of squalene standard detected using GC-MS (C) Comparison of squalene content between wild type, Δ shc , Δ shc :pPMQ shc complemented strain, and Δ sqs cells. n.d. = no squalene detected. Complementation of the Δ shc Strain To confirm that the observed squalene accumulation in the Δ shc cells is due to the deletion in slr2089 , we performed a complementation of the deletion in the Δ shc background. For this purpose, slr2089 and an approximately 1200 bp region immediately upstream of the gene were cloned in a self-replicating vector and used to transform the Δ shc strain. In the resulting Δ shc :pPMQ shc strain, squalene accumulation was 0.0529±0.0031 mg OD 750 \n −1 L −1 , and thus it was strongly reduced compared to the level in the Δ shc cells ( Fig. 3C ), showing that the introduced shc -gene did complement the inactivation in Δ shc . However, the level of squalene was not as low as in the wild type (see above and Fig. 3C ). This may be due to insufficient expression from the plasmid construct. Inactivation of sll0513 ( sqs ) As described above, we identified one gene, sll0513 , in the genome sequence of Synechocystis, putatively encoding squalene synthase. Since this gene is not very similar to the only cyanobacterial squalene synthase characterized so far, sqs in T. elongatus \n [27] , we decided to investigate its function by making a deletion of this gene. We found that in the lipid extracts from the sll0513 deletion strain, Δ sqs, no squalene peak could be detected by HPLC ( Fig. 3C ). Wild type cells did contain a low level of squalene (see above), probably present as an intermediate metabolite. The complete absence of any squalene peak in the Δ sqs cell extracts therefore indicates that sll0513 really does encode squalene synthase, essential for squalene formation, in Synechocystis . The results presented above show that Synechocystis certainly exhibits a squalene synthase activity, and this together with the conserved sequence features present in sll0513, the lack of squalene production in the Δ sqs strains, and the lack of any other obvious candidate squalene synthase genes in the Synechocystis genome, present a strong indication that sll0513 does indeed encode squalene synthase in Synechocystis , despite the observed differences between the deduced amino acid sequence of sll0513 and the squalene synthase of T. elongatus BP-1 ( tll1096 ) [27] . Thus, we suggest that squalene, and hopene, formation in Synechocystis takes place according to the pathway presented in Fig. 1 , and that sll0513 is sqs in Synechocystis . Growth Characteristics of Δ shc Compared to Wild Type Synechocystis \n In order to assess the growth characteristics of the Δ shc strain, wild type Synechocystis and Δ shc were grown in parallel cultures under photoautotrophic growth conditions. In other organisms, it has been found that inactivating shc and thereby the production of hopanoids led to membrane damage [32] , [33] . We therefore hypothesized that a lack of hopanoids might affect membrane systems, potentially including the thylakoid membranes, of Δ shc , and lead to photosynthetic growth defects. To determine the effect of Δ shc on the growth at different light intensities, mutant and wild type cultures were inoculated from cells grown at normal light and then grown at low light (LL, 5 µmol photons m −2 s −2 ), normal light (NL, 50 µmol photons m −2 s −2 ) and high light (HL, 500 µmol photons m −2 s −2 ) ( Fig. 4 ). To quantify growth, OD 750 was measured every 24 hours up to 192 hours. There was a marked difference in growth between different light intensities where LL had a slower initial growth but in the end reached the same OD. The difference between wild type and Δ shc , however, was not significant, suggesting that a loss of Shc has no impact on normal photoautotrophic growth, nor on high light induced stress tolerance under the conditions tested. Thus, if membranes in Synechocystis are affected by inactivating shc , potentially resulting in a lack of hopanoids, the effect is not so severe as to impact growth under any of the different light conditions tested. Furthermore, the accumulation of squalene was not detrimental to cell growth. 10.1371/journal.pone.0090270.g004 Figure 4 Growth curve of Synechocystis wild type and Δ shc strain under different light conditions. Wild type = solid lines; Δ shc = dashed lines; LL = low light (squares, 5 µmol photons m −2 s −2 ); NL = normal light (triangles, 50 µmol photons m −2 s −2 ); HL = high light (diamonds, 500 µmol photons m −2 s −2 ). While inactivation of shc did not lead to sensitivity to high light stress, it is possible that other stress conditions may reveal a Δ shc phenotype, and we plan to address this question in future studies. It is clear from the literature that hopanoids may have different roles and be of varying importance for the growth of different organisms [32] , [33] , [34] . Our work in this study forms the foundation for further investigation of the role of hopanoids in cyanobacteria, which will be interesting given that they are oxygenic photosynthetic autotrophs and have a lifestyle quite different from other microorganisms where such studies have been performed so far. Accumulation of Squalene Depending on Light, Growth Phase In order to investigate whether squalene production is connected to light conditions or a specific growth phase in Synechocystis , we collected samples of the Δ shc strain from cells in different growth phases and under two different light conditions, low light (LL, 5 µmol photons s −1 m −2 ) and normal light (NL, 50 µmol photons s −1 m −2 ), and examined the squalene content of the samples (see Fig. 5 ). The cultures were harvested in the exponential phase (40 h), late exponential phase (88 h) and in the stationary phase (280 h). Squalene content in the LL cultures at 40 hours decreased from the level in the seed culture (0 h, grown at NL prior to the start of the experiment), but then levels increased at similar rates as for the NL cultures as they grew ( Fig. 5A ). At each time point, cells grown under NL had higher squalene content, measured per OD 750 and volume culture, than did LL-grown cells. This may be an effect of the lower growth rate at LL, as squalene seems to accumulate in the cells during growth. LL cultures at 280 hours reached similar OD as normal light cultures at 88 hours and also similar squalene accumulation, suggesting a correlation of squalene production and cell density. Thus, faster growing cells, leading to higher cell density, would have a higher squalene content at any given time point in a batch experiment. It should also be noted that the cultures used in the experiment had been pre-cultivated under NL and was split and moved to LL and NL conditions at the start of the experiment. The drop in squalene content in the LL grown cells after 40 h in LL presumably reflects this change in growth conditions and shift to a slower rate of growth at LL. If this change occurs because of a lower squalene production compared to growth rate, effectively diluting out the content with every cell division, or through some other mechanism is yet unclear. 10.1371/journal.pone.0090270.g005 Figure 5 Accumulation of squalene at different growth stages and light intensities. (A) Squalene content of Δ shc cells grown at different light intensities and extracted at different growth stages. The mutant strain was grown under normal light, and at 0 hours low light (LL, 5 µmol photons m −2 s −2 ) and normal light (NL, 50 µmol photons m −2 s −2 ) cultures were inoculated. Squalene was then extracted at different time points to investigate the correlation between squalene production and growth phase. (B) Growth curve of Δ shc (dashed lines) from the same experiment, and of wild type Synechocystis (solid lines), using low light (squares) or normal light (triangles)."
} | 6,032 |
35298339 | PMC8944584 | pmc | 5,591 | {
"abstract": "Significance Horizontal gene transfer (HGT)—the transfer of DNA between lineages—is responsible for a large proportion of the genetic variation that contributes to evolution in microbial populations. While HGT can bring beneficial genetic innovation, the transfer of DNA from other species or strains can also have deleterious effects. In this study, we evolve populations of the bacteria Helicobacter pylori and use DNA sequencing to identify over 40,000 genetic variants transferred by HGT. We measure the cost of many of these and find that both strongly beneficial mutations and deleterious mutations are genetic variants transferred by natural transformation. Importantly, we also show how recombination that separates linked beneficial and deleterious mutations resolves the cost of HGT.",
"conclusion": "Conclusions In natural populations of microbes, DNA can spread horizontally by a range of mechanisms. For example, plasmid conjugation and bacteriophage selfishly promote their own spread and can prevail despite significant costs ( 62 ). In contrast, for other mediums of HGT—nanotubes, vesicles, and natural transformation—the integration of DNA sequences by HGT is random with respect to its fitness effect in a given environment. Here, we study HGT by natural transformation, which provides an opportunity to separate the evolutionary forces that shape the dynamics of HGT variants from the confounding effects of infectious mechanisms of HGT. Our results confirm two important roles of recombination in populations evolving with HGT. First, HGT from diverged populations can introduce a vast amount of genetic variation into the evolving population, even without selection for those variants. The scale of genetic variation in HGT compared with non-HGT treatment populations ( Fig. 3 and SI Appendix , Fig. S3 ) shows how the genome content of a local donor population could significantly influence the evolution of a naturally competent recipient population. Second, our data show that Fisher–Muller mechanisms function in recombining populations of microbes and can ameliorate the costs of the horizontal import of deleterious genetic variants in adapting populations. Although H. pylori has an exceptionally high recombination rate ( 63 ), the observation of gene-specific sweeps in natural microbiomes ( 21 , 64 , 65 ) suggests that the two impacts of recombination observed in this study—the introduction and reshuffling of novel genetic variation—are likely to be common in natural microbial populations. Future experimental and theoretical studies need to take horizontal gene flow and recombination into account when considering adaptation in natural and clinical microbial communities."
} | 674 |
38604750 | PMC11062418 | pmc | 5,592 | {
"abstract": "Abstract Major progress in developing Saccharomyces cerevisiae strains that utilize the pentose sugar xylose has been achieved. However, the high inhibitor content of lignocellulose hydrolysates still hinders efficient xylose fermentation, which remains a major obstacle for commercially viable second-generation bioethanol production. Further improvement of xylose utilization in inhibitor-rich lignocellulose hydrolysates remains highly challenging. In this work, we have developed a robust industrial S. cerevisiae strain able to efficiently ferment xylose in concentrated undetoxified lignocellulose hydrolysates. This was accomplished with novel multistep evolutionary engineering. First, a tetraploid strain was generated and evolved in xylose-enriched pretreated spruce biomass. The best evolved strain was sporulated to obtain a genetically diverse diploid population. The diploid strains were then screened in industrially relevant conditions. The best performing strain, MDS130, showed superior fermentation performance in three different lignocellulose hydrolysates. In concentrated corncob hydrolysate, with initial cell density of 1 g DW/l, at 35°C, MDS130 completely coconsumed glucose and xylose, producing ± 7% v/v ethanol with a yield of 91% of the maximum theoretical value and an overall productivity of 1.22 g/l/h. MDS130 has been developed from previous industrial yeast strains without applying external mutagenesis, minimizing the risk of negative side-effects on other commercially important properties and maximizing its potential for industrial application.",
"introduction": "Introduction The bioethanol industry has contributed significantly to reduction of greenhouse gases, increasing energy security, and also supported economic development in many countries. Ethanol produced from biomass accounts for the largest fermentation product by volume (Meyer Hans-Peter 2016 ). It is among the most promising green fuel alternatives to fossil fuels in the short term. Currently, ethanol is produced mainly from feedstocks such as corn and sugar cane, known as first-generation (1 G) bioethanol. 1 G bioethanol production has greatly optimized fermentation processes and has grown to an economically highly competitive industry. 1 G bioethanol is currently the only transportation fuel alternative that can be produced in an economically viable manner (Mizik 2021 ). However, feedstock competition with the food supply has drawn considerable attention, especially because of the impact on developing countries. Additionally, the increased demand for 1 G feedstock might have contributed to deforestation when forest land was converted into agricultural land. This is claimed to have resulted, at least in some cases, in low or even no net greenhouse gas savings by the 1 G bioethanol production process (Jeswani et al. 2020 ). On the other hand, second generation (2 G) ethanol production utilizes lignocellulosic feedstocks such as agricultural and forest residues, as well as the organic fraction of municipal solid waste. Ethanol produced from such waste biomass is considered to be a near carbon-neutral or even carbon-negative renewable fuel alternative. However, until now, economically feasible industrial-scale production of bioethanol from waste biomass has remained a major challenge, with only limited industries advancing to commercial scale and their economic viability remaining unclear (Devi et al. 2022 , Raj et al. 2022 ). The main hurdle for economically sustainable 2 G ethanol production lies in the recalcitrant nature of lignocellulosic biomass. Complete hydrolysis of lignocellulosic biomass into fermentable sugars in an economically viable manner has not been achieved because of its complex structure and composition (Vasic et al. 2021 ). Lignocellulose materials are composed of polymers of cellulose, hemicellulose, and lignin. These polymers provide rigidity to the biomass, are structurally complex and are resistant to chemical, physical, and enzymatic hydrolysis. The most effective hydrolysis processes to date implement harsh chemical and physical pretreatment conditions before enzymatic hydrolysis. Such harsh pretreatment steps release inhibitory chemicals such as weak acids, furan, and phenolic compounds (Palmqvist Eva 2000 , Soares et al. 2017 , Seidel et al. 2019 ). These chemicals significantly reduce the efficiency of fermentation by inhibiting the metabolism of the fermenting microorganisms. In addition, the industry is confronted with the difficulty of finding robust microorganisms that are able to completely convert all the hydrolyzed sugars derived from the cellulose and hemicellulose into bioethanol under industrial conditions. Since the biomass is a major contributor to the production cost, a high overall ethanol yield, utilizing all available sugars from the biomass, is an essential condition for the industrial process. The sugars include hexose sugars such as glucose and mannose, and pentose sugars, mainly xylose and arabinose. On top of that, minimum water usage is important to reduce operational costs. As a result, there is a need to use biomass hydrolysates with a high solids load. High solid content, however, implies high concentrations of inhibitors, of which the toxicity is further enhanced by the ethanol produced (Koppram et al. 2014 ). As a consequence, an organism that can ferment both hexose and pentose sugars and displays high inhibitor tolerance is a prerequisite for economically viable industrial production of bioethanol from lignocellulosic biomass. \n Saccharomyces cerevisiae is the dominant microorganism in 1 G bioethanol production because it produces the highest ethanol titer from hexose sugars at nearly stochiometric yield and is tolerant to ethanol and a wide range of inhibitors. The major bottleneck for the use of S. cerevisiae in cellulosic ethanol production is its inability to utilize pentose sugars. However, in recent years, recombinant S. cerevisiae strains that can efficiently ferment xylose have been developed by metabolic and evolutionary engineering (Jeffries 2006 , Hahn-Hagerdal et al. 2007 , Brat et al. 2009 , Demeke et al. 2013a ). Yet, the efficiency of hexose and especially pentose fermentation by the engineered strains is greatly hampered by the inhibitors present in lignocellulosic hydrolysates. Though glucose utilization is reasonably maintained in the presence of moderate to high concentrations of inhibitors, xylose fermentation is much more sensitive and severely reduced or even eliminated already at low inhibitor concentrations (Bellissimi et al. 2009 , Vanmarcke et al. 2021 ). This forced the industry to use low density lignocellulose hydrolysates to achieve good ethanol yields and productivity. However, this results in the use of excessive amounts of water, and also results in a low final ethanol titer. As a consequence, the downstream processing becomes less cost effective. Therefore, yeast strains able to ferment all sugars in hydrolysates at high solids loads and inhibitor concentrations are necessary for economically viable cellulosic ethanol production. Evolutionary engineering has been applied widely for improvement of industrial properties in S. cerevisiae . In evolutionary engineering, a cultivation condition with a selective pressure to industrially relevant traits is applied and spontaneous or induced mutant strains with a selective advantage under the condition used are selected (Mans et al. 2018 ). To enhance the mutation frequency, chemical or physical mutagens can be applied, followed by selection for the specific trait of interest (Smith et al. 2014 , Inokuma et al. 2017 ). However, the large number of mutations generated by such mutagens frequently results in side effects that compromise other important industrial traits, such as the proliferation rate under conditions used for industrial yeast production and the general robustness under varying, stressful industrial conditions (Cakar et al. 2012 , Demeke et al. 2013a ). In this work, we have developed a robust industrial S. cerevisiae strain with efficient xylose-fermentation capacity in nondetoxified lignocellulose hydrolysates at high solids loading. For that purpose, we developed and employed a novel evolutionary engineering strategy that can be applied for industrial strain development with selectable traits such as efficient substrate utilization and high inhibitor tolerance. The resulting strain, MDS130, demonstrated complete conversion of both glucose and xylose to ethanol with a high yield and overall productivity, and reaching ethanol titers >7% (v/v) in concentrated undetoxified cellulosic hydrolysates.",
"discussion": "Discussion Impressive progress has been achieved in the past decades for the development of superior S. cerevisiae strains with very good xylose-fermentation capacity. However, strains that show rapid and complete xylose utilization in synthetic medium were found to have much slower and incomplete xylose utilization in lignocellulose hydrolysates (Demeke et al. 2013a , Li et al. 2016 , Costa et al. 2017 ). Though glucose utilization is not severely affected in lignocellulose hydrolysate fermentations, xylose utilization is hindered even at lower inhibitor concentrations. This could be due to the fact that xylose is not a natural substrate for S. cerevisiae , with its metabolism as a result not being well-integrated into the regulatory network of cellular metabolism, such as the response to stressful conditions (Deparis et al. 2017 ). The slow and incomplete xylose utilization observed when the strains are used in fermentations of concentrated undetoxified lignocellulose hydrolysates, which are notoriously rich in fermentation inhibitors, has forced the industry to introduce a costly washing step or to use more diluted hydrolysate to reduce the inhibitor concentration. However, cost competitive lignocellulose based ethanol production requires the use of concentrated hydrolysates to achieve a minimum ethanol concentration of at least 4% v/v (Palmqvist Eva 2000 , Koppram et al. 2014 ). In this work, we have developed using a novel evolutionary engineering method an industrial yeast strain able to efficiently utilize xylose in nondetoxified concentrated lignocellulose hydrolysates. We have employed an evolutionary engineering approach that combined the superior industrial properties of a robust 1 G strain, Ethanol Red, and a 2 G strain GSE16-T18. We used a tetraploid hybrid, rather than a regular diploid hybrid strain generated from two haploid derivative strains, in the adaptive evolution cycles for two main reasons. First, to obtain a diploid hybrid strain, one must isolate haploid segregants from each diploid parent strain with the same industrial trait as the diploid parent. Isolation of haploid segregants carrying all the superior properties of its diploid parent strain is cumbersome and sometimes impossible (Kim et al. 2017 ). Instead, by using a tetraploid hybrid, we maintain all the genetic factors of both parent strains that determine important industrial traits. In our case, the main industrial traits are the rapid xylose-fermentation capacity and good inhibitor tolerance of the 2 G strain GSE16-T18, and the very good propagation capacity, high temperature tolerance and absent flocculation of the 1 G strain Ethanol Red. Second, considering a limited rate of random mutations for a certain genome size per generation, we anticipated that tetraploid strains that carry double the genome of diploid strains generate more mutations per generation than diploid strains during adaptive evolution (Selmecki et al. 2015 ). This would facilitate the generation of beneficial mutations that could be selected with the applied selective pressure. Indeed, all four cultures containing diploid strains dropped out due to the loss of fermentation capacity in the first five cycles of the adaptive evolution step. One of the two tetraploid hybrids maintained higher fermentation capacity, probably due to the rapid generation of mutants able to tolerate the high level of inhibitors. However, the better performance by the tetraploid hybrid strains could also be due to their inherent genetic fitness since they carry all genetic elements coming from each diploid parent. Acid pretreated spruce material is among the most inhibitory lignocellulose material. It contains high concentrations of acetic acid, HMF and furfural (Demeke et al. 2013b , Vanmarcke et al. 2021 ). We used a spruce hydrolysate with a high concentration of inhibitors for the adaptive evolution steps. Improvement of inhibitor tolerance to this high concentration of inhibitors should in principle also result in better performance in other relatively less inhibitory hydrolysates. As expected, the best isolate MDS130 showed significant improvement in spruce hydrolysate as well as in two other hydrolysates originating from sugar cane bagasse and corn cob. Adaptive evolution in inhibitor-rich hydrolysate has been used previously to improve fermentation performance of 2 G yeast strains. Nevertheless, enhanced fermentation is often associated with improved glucose utilization, while xylose utilization usually shows much less or even no improvement at all (Zhu et al. 2015 , Liu 2017 et al. 2017 ). This is probably because most hydrolysate media contain much more glucose than xylose, and S. cerevisiae prefers glucose over xylose (Zaldivar et al. 2002 ). The yeast’s innate preference for glucose would therefore hamper xylose metabolism and its adaptation in the presence of inhibitors. For that reason, improvement of xylose utilization in inhibitor-rich hydrolysates was found to be difficult (Liu et al. 2017 ). To address this problem, we have first used xylose-enriched hydrolysate and second, hydrolysate in which all glucose was converted with a regular yeast strain into ethanol. In the first hydrolysate, the yeast had to improve its xylose-utilization capacity in the inhibitor-rich hydrolysate to proliferate faster. The same was true in the second hydrolysate, which however, had the additional challenge of the presence of high ethanol, which exacerbates the toxicity of the lignocellulose-derived inhibitors. Encouraging results have been obtained with rational metabolic engineering for improvement of inhibitor tolerance in 2 G yeast strains. However, similar to the results with evolutionary engineering approaches, the inhibitor tolerance was often only improved for glucose utilization while xylose-fermentation capacity was sometimes even reduced (Wallace-Salinas et al. 2014 , Jayakody et al. 2018 , Brandt et al. 2021 ). Targeted genetic modifications to improve inhibitor tolerance possibly only work well for glucose utilization because the genetic factors were originally also identified for fermentation in glucose medium. In that sense, evolutionary engineering for improved xylose utilization and inhibitor tolerance in the absence of glucose, and especially in the presence of high ethanol, is an interesting alternative. In our study, we performed the evolutionary adaptation in glucose-depleted and xylose-enriched hydrolysate to facilitate development of inhibitor tolerance of xylose fermentation by overcoming the effect of glucose repression on xylose utilization. In the first-round adaptation, there was still a small amount of glucose in the medium to help initiate growth. Once the strains were adapted to the presence of the inhibitors, the second-round adaptive evolution was subsequently done using the best isolate from the first round in glucose-depleted medium in the presence of only xylose as a carbon source and also high ethanol. As a result of the two consecutive approaches, significant improvement in the xylose-fermentation rate was observed over the course of the adaptive evolution cycles, indicating improvement of xylose-fermentation tolerance to higher concentrations of inhibitors. This was clearly demonstrated by the significant improvement of the xylose-utilization rate in the presence of inhibitors by one of the evolved single cell isolates, MD4, compared to GSE16-T18, without much difference in the glucose utilization rate (Fig. 6 ). Unraveling the genetic modifications responsible for the simultaneous improvement of xylose fermentation and inhibitor tolerance could reveal targets for further improvement by targeted metabolic engineering, though that is beyond the scope of this work. Simultaneous improvement of xylose fermentation and inhibitor tolerance has been achieved previously in lignocellulose hydrolysates (Smith et al. 2014 ) and also specifically to weak acid and high temperature (Inokuma et al. 2017 ). While those studies used chemical or physical mutagenesis to create population diversity, no mutagenesis step was used in our study to avoid negative effects by background mutations. Reduced growth rate, for instance, was manifested in previous studies (Demeke et al. 2013a , Smith et al. 2014 ). Instead, adaptive evolution of tetraploid hybrid strains followed by sporulation, generated a very diverse population of diploid strains carrying different combinations of genetic elements from the two parent strains as well as new combinations of mutations generated during the adaptive evolution. The evolved tetraploid strain MD4 itself showed significant improvement in xylose fermentation in corn cob and sugarcane bagasse hydrolysates (Fig. 6 ). However, xylose fermentation by MD4 was not complete, even after 90 h of fermentation. Since generation of heterozygous mutations is more common during evolutionary adaptation than generation of homozygous mutations (Sellis et al. 2016 ), we envisaged that the effect of recessive mutations masked by the wild type allele could be manifested when the wild type allele was lost during subsequent chromosome segregation. This could further improve the xylose-fermentation performance and inhibitor tolerance upon selection for these traits. Additionally, genetically diverse diploid populations of strains could be generated by sporulation, and with an efficient selection step, further improved superior strains could be obtained. Indeed, after sporulation and screening of more than 200 segregants in inhibitor-rich hydrolysate, several diploid strains showed better fermentation performance in lignocellulose hydrolysates (Figs 8 and 9 ). The best diploid isolate, MDS130, showed a stable superior fermentation profile in three different lignocellulose hydrolysates. MDS130 showed very good xylose utilization in nondetoxified, concentrated lignocellulose hydrolysates. The very high ethanol yield of 91% in such very inhibitory medium appears to be unprecedented. The ethanol yield by the control strain GSE16-T18 was 81%, which is comparable to that of most engineered 2 G strains in previous reports (Jansen et al. 2017 ). Additionally, the almost complete xylose utilization by MDS130 within 48 h resulting in an ethanol titer of 5.6% w/v (about 7% v/v) in such highly challenging lignocellulose hydrolysates is unprecedented and might already meet industrial requirements. Nonetheless, compared to glucose utilization, xylose utilization is still much slower resulting in a prolonged overall fermentation time. Hence, further improvement of the xylose utilization rate in the presence of inhibitors would further reduce the overall fermentation time to make industrial 2 G ethanol production more economically feasible. The fermentation temperature employed in this study was 35°C, which is higher than the regular 30°C–32°C fermentation temperatures for S. cerevisiae . Though 35°C is optimum for the first generation strain Ethanol Red, fermentation of lignocellulosic material is sensitive to higher temperatures (Zhu et al. 2015 ). We used 35°C during the evolutionary adaptation, which resulted in superior performance by MDS130 at this temperature. We chose 35°C for two main reasons. First, higher temperature is more relevant for the simultaneous saccharification and fermentation (SSF) process (Mutturi and Lidén 2013 ). SSF is an interesting process option for bioethanol production that uses enzymatic hydrolysis together with fermentation. It is preferred for higher product yield and requires less investment. Though SSF is not yet fully practical, partial saccharification could be possible during fermentation even in a separate hydrolysis and fermentation process, when the enzymes used at the saccharification step are not inactivated. In that respect, fermentation at higher temperature will have an advantage because it is closer to the optimum saccharification temperature. On top of that, fermentation at higher temperature reduces the cost of cooling fermentation tanks (Prado et al. 2020 ). In conclusion, we have developed a robust industrial strain MDS130, that can completely utilize xylose in undetoxified lignocellulose hydrolysate with high ethanol yield and titer. To achieve this, we devised a new evolutionary engineering method that allows to combine and select the superior industrial properties from two diploid industrial strains into a single diploid strain, and also resulted in more rapid evolution without external mutagenesis step. The method can be applied for development of industrial strains with selectable traits without introducing chemical or physical mutagenesis steps, which are known to frequently introduce undesired genetic changes. MDS130 is a promising strain for direct industrial application of cellulosic ethanol production."
} | 5,394 |
26613343 | PMC5029182 | pmc | 5,594 | {
"abstract": "Dense microbial groups such as bacterial biofilms commonly contain a diversity of cell types that define their functioning. However, we have a limited understanding of what maintains, or purges, this diversity. Theory suggests that resource levels are key to understanding diversity and the spatial arrangement of genotypes in microbial groups, but we need empirical tests. Here we use theory and experiments to study the effects of nutrient level on spatio-genetic structuring and diversity in bacterial colonies. Well-fed colonies maintain larger well-mixed areas, but they also expand more rapidly compared with poorly-fed ones. Given enough space to expand, therefore, well-fed colonies lose diversity and separate in space over a similar timescale to poorly fed ones. In sum, as long as there is some degree of nutrient limitation, we observe the emergence of structured communities. We conclude that resource-driven structuring is central to understanding both pattern and process in diverse microbial communities.",
"introduction": "Introduction A key determinant of whether a microbial cell survives and divides is the identity of surrounding cells. These neighboring cells determine the signaling molecules it will perceive, whether it will be infected by plasmids or viruses, whether it will be attacked by toxins, and, most fundamentally, its access to resources. On an evolutionary timescale, the pressures exerted by neighbors can lead to the competitive exclusion of a genotype or to new evolutionary adaptations ( Kerr et al. , 2002 ; Habets et al. , 2006 ; Kim et al. , 2014 ). These adaptations include both cooperative and competitive phenotypes that shape the productivity and functioning of the group, such as the secretion of polymers that allow cells to get better access to nutrients ( Kim et al. , 2014 ), or the evolution of lethal antibiotics ( Kreft, 2004 ; Nadell et al. , 2009 ; Mitri and Foster, 2013 ; Koch et al. , 2014 ). In order to understand the population dynamics of microbial groups, it is therefore necessary to understand how and why genotypes and phenotypes organize in space ( Shapiro, 1995 ; Johnson and Boerlijst, 2002 ; Korolev et al. , 2014 ; Murray et al. , 2014 ; Van Gestel et al. , 2014 ). Theory and experiments have revealed that, in the absence of mutation and selection, initially diverse and well-mixed microbial populations will commonly lose their diversity when growing in dense groups such as bacterial colonies, leading to large patches of single genotypes ( Ben-Jacob et al. , 1994 ; Golding et al. , 1999 ; Habets et al. , 2006 ; Hallatschek et al. , 2007 ; Xavier and Foster, 2007 ; Hallatschek and Nelson, 2010 ; Korolev et al. , 2012 ). Theory suggests that nutrient limitation is a general factor that underlies this process ( Nadell et al. , 2010 ). Specifically, nutrient limitation ensures that only cells at the edge can grow, which leads to bottlenecks and genetic drift that drives a loss of diversity ( Hallatschek et al. , 2007 ; Nadell et al. , 2010 ). However, we lack empirical tests that demonstrate this key link between nutrient levels and the genetic or phenotypic organization of microbial groups. Here we use a combination of simulation modeling and mixed-genotype Pseudomonas aeruginosa bacterial colonies to evaluate the role of nutrient levels in the spatial structuring of microbial groups. We find that, as predicted by theory, colonies with abundant resources maintain large unstructured regions. High resource supply means that a large sub-population of cells will be dividing rapidly, allowing many different lineages to be maintained as the group expands. However, while high-nutrient groups remain well mixed for many more cell divisions than low-nutrient groups, they also expand more rapidly. Given sufficient space to expand, therefore, high- and low-nutrient groups experience a comparable loss of diversity over time. Our work suggests that the effects of resource limitation will be central to understanding the organization and evolution of many microbial communities.",
"discussion": "Discussion The spatial structure of genotypes within microbial communities is considered central to their form and function. Strong spatial structuring is generally associated with increased genetic drift and weakened natural selection ( Hallatschek and Nelson, 2008 ; Hallatschek and Nelson, 2010 ; Korolev et al. , 2011 ). Nevertheless, the emergence of structure can result in strong natural selection for cooperative traits that benefit the cells of the same genotype in the surroundings and promote productivity and general resilience ( Nadell et al. , 2010 ; Mitri et al. , 2011 ; Mitri and Foster, 2013 ; Kim et al. , 2014 ). In addition, spatial structure can have important ecological impacts, which includes making communities more stable by reducing the strength of interactions between species ( Coyte et al. , 2015 ; Kim et al. , 2008 ). Understanding what drives spatial structuring in microbial groups, therefore, is a fundamental challenge for microbiology. Our work suggests that nutrient limitation is a key mechanism underlying the emergence of spatial structure in cell groups. Unless nutrients are saturating through a cell group, there will be a limited number of cells dividing at the growing edge, which promotes the stochastic loss of cell lineages. The number of cells dividing at the growing edge increases with nutrient levels and determines the expansion velocity of the colony. Importantly, increases in this velocity are strongly associated with the maintenance of diversity. The power of this model to explain diversity loss can be summarized by plotting the rate of diversity loss in space as a function of colony expansion velocity ( Figure 7 ). However, our work also emphasizes that any gradation in nutrient levels has the potential to lead to diversity loss, as long as there is sufficient space for expansion. Although high-nutrient conditions allow cells to maintain large well-mixed areas, because the groups are growing rapidly, they too will lose diversity on similar timescales as groups under much lower-nutrient levels. Because of these differences in growth rate, then, colonies will end up with similar levels of genetic diversity independently of nutrient abundance, assuming that they are not limited by space. Therefore, in the absence of other mechanisms that strongly disrupt spatial structure—such as the production of surfactants ( Xavier et al. , 2011 ) or surface motility ( Figure 1b )—nutrient gradients always have potential to generate strong spatial separation. Consistent with this conclusion, evidence suggests that the demixing of diverse populations into clonal groups is common in microbes, including P. aeruginosa ( Korolev et al. , 2011 ), E. coli ( Korolev et al. , 2010 ; Korolev et al. , 2011 ), Saccharomyces cerevisiae ( Hallatschek and Nelson, 2010 ; Korolev et al. , 2010 ; Müller et al. , 2014 ), Dictyostelium discoideum ( Buttery et al. , 2012 ) and Bacillus subtilis ( Ben-Jacob et al. , 1994 ; Golding et al. , 1999 ), and under different growth conditions in the laboratory (on agar plates ( Shapiro, 1995 ; Golding et al. , 1999 ; Hallatschek et al. , 2007 ; Freese et al. , 2014 ) and in flow cells ( Nielsen et al. , 2000 ; Pamp and Tolker-Nielsen, 2007 ; Momeni et al. , 2013a )). An exception can occur when the different strains and species depend on each other to grow, such as in cross-feeding interactions, where strains are immune to the demixing processes we describe here ( Momeni et al. , 2013a ; Müller et al. , 2014 ). In this case, however, the positively interacting species can effectively be viewed as a single ecological unit that will, in turn, segregate with respect to other competing strains or mixtures of strains ( Mitri et al. , 2011 ; Momeni et al. , 2013b ). It is important to note that the large-colony and high-nutrient conditions often studied in the laboratory setting are not reflective of many natural conditions where densities and growth rates can be much lower. Our study does, however, consider relatively low-nutrient conditions, and we observe that spatial structure can emerge on very small spatial scales (tens of microns, Figure 1d ). Our observation of fine-scale structuring contrasts with the emphasis on sequencing studies that sample over large scales and detect extremely high species diversity ( Gans et al. , 2005 ; Roesch et al. , 2007 ). However, there is a growing body of work that applies fluorescent in situ hybridization to natural communities, which also highlight a significant potential for fine-scale spatial structuring of genotypes ( Stacy et al. , in press ), such as in burn wounds ( Fazli et al. , 2009 ; Malic et al. , 2009 ), in the gut microbiota ( Dejea et al. , 2014 ; Millet et al. , 2014 ; Engel et al. , 2015 ; Earle et al. , 2015 ), in bladder infections ( Kikuchi et al. , 2009 ) or on leaf surfaces ( Monier and Lindow, 2005 ). There is also variability in the degree of structure, however, and significant genotypic mixing is also seen in some contexts, such as human dental plaque ( Palmer et al. , 2003 ; Zijnge et al. , 2010 ) or infections of the middle ear ( Weimer et al. , 2010 ). The observation of natural spatial structure does not, of course, demonstrate that this structure was driven purely by nutrient-limited growth. Spatial separation can result from other factors, including sparse seeding of a population or a patchy environment of microniches ( Kikuchi et al. , 2009 ; Kubo et al. , 2011 ). Although our experiments were conducted with microbes, we hypothesize that our findings will be common to any initially diverse population containing multiple heritable cell types with limited access to resources. Resource gradients are expected to be common in all cellular systems, occurring not only owing to nutrient limitations but also low-nutrient diffusivity, high cell growth rates or low yields of converting nutrients to biomass ( Nadell et al. , 2010 ). The logic also applies to any diffusible substrate that is necessary for growth, such as signaling molecules. Outside of microbial communities, our conclusions are likely to be particularly relevant to the organization and evolution of cancerous tumors. Viewed through the lens of evolutionary ecology, tumors are populations of cells that acquire mutations and diversify as they grow and expand in space, followed by migrations to different parts of the body (metastases; Korolev et al. , 2014 ; Gundem et al. , 2015 ). These parallels to other evolving populations may help predict tumor evolution, leading to novel therapies ( González-García et al. , 2002 ; Gatenby et al. , 2013 ; Korolev et al. , 2014 ). Indeed, preliminary data using computer simulations together with ‘radiogenomics'—where the spatio-genetic structures resulting from these processes are revealed by linking radiology to the genetic characterization of cancer cells ( Gatenby et al. , 2013 )—suggest that competition between neighboring cells is likely to lead to the spatial separation of genotypes, as seen in our colonies ( González-García et al. , 2002 ; Carmona-Fontaine et al. , 2013 ; Gatenby et al. , 2013 ). Similarly, cell competition over resources has been suggested to be important during multicellular development. Experiments with neutral markers to distinguish lineages of developing cells in the Drosophila melanogaster wing have shown spatial separation of cells reminiscent of that observed in microbial colonies, which is driven by gradients in signaling molecules rather than nutrient gradients ( Johnston, 2009 ). In sum, our analyses show that resource gradients can have a pivotal role in purging diversity and shaping the spatial structure of microbial groups. It is clear that these effects will act alongside a number of other processes that affect spatial structure, ecology and evolution. Nevertheless, our findings predict that the process of spatial structuring and the concomitant loss in diversity will be common due to the commonness of nutrient gradients ( Stewart et al. , 2008 ). Although genomics often documents extremely high species diversities, the near-universality of resource limitation suggests that local diversity may be much lower than these estimates. A gradual reduction in diversity as populations grow may prove fundamental to the structure and function of microbial communities and other cellular groups."
} | 3,138 |
37333853 | PMC10272358 | pmc | 5,595 | {
"abstract": "Introduction \n Burkholderia thailandensis is a study model for Burkholderia pseudomallei , a highly virulent pathogen, known to be the causative agent of melioidosis and a potential bioterrorism agent. These two bacteria use an (acyl-homoserine lactone) AHL-mediated quorum sensing (QS) system to regulate different behaviors including biofilm formation, secondary metabolite productions, and motility. Methods Using an enzyme-based quorum quenching (QQ) strategy, with the lactonase Sso Pox having the best activity on B. thailandensis AHLs, we evaluated the importance of QS in B. thailandensis by combining proteomic and phenotypic analyses. Results We demonstrated that QS disruption largely affects overall bacterial behavior including motility, proteolytic activity, and antimicrobial molecule production. We further showed that QQ treatment drastically decreases B. thailandensis bactericidal activity against two bacteria ( Chromobacterium violaceum and Staphylococcus aureus ), while a spectacular increase in antifungal activity was observed against fungi and yeast ( Aspergillus niger , Fusarium graminearum and Saccharomyces cerevisiae ). Discussion This study provides evidence that QS is of prime interest when it comes to understanding the virulence of Burkholderia species and developing alternative treatments.",
"conclusion": "5 Conclusion Using a combination of phenotypic and molecular approaches we demonstrate that lactonase-mediated quorum quenching strongly alters the behavior and the virulence of B. thailandensis , a model strain for studying the highly virulent pathogen B. pseudomallei . Biofilm formation, motility, proteolytic activity were strongly impacted upon enzymatic treatment. Interestingly we show that antimicrobial activities of B. thailandensis against other microorganisms are highly mediated by the lactonase, bactericidal effect towards both Gram positive and negative strains being significantly decreased while fungicidal activity being tremendously enhanced against both yeast and fungi. This work, combining fine and exhaustive microbiological characterizations together with proteomic analyses, will help to understand the role of QS in bacterial virulence and microbial interactions and will contribute to promote the use of quorum quenchers for controlling pathogens and modulate cooperation and competition in complex microbial populations.",
"introduction": "1 Introduction Most bacteria use a molecular communication system, referred to as quorum sensing (QS). It relies on the production and detection of small molecules known as autoinducers. QS enables the adaptation of bacterial behaviors in a cell density-dependent manner and allows bacteria to act cooperatively and induce different physiological functions ( Mion et al., 2019 ). Among the mechanisms activated by QS, bioluminescence was the first described in Vibrio sp. ( Tobias et al., 2020 ). Other mechanisms were later described, such as biofilm formation, virulence factor production, and antibiotic production ( Zhao et al., 2020 ). Many autoinducers have been identified, including oligopeptides which are mainly used by Gram-positive bacteria, acyl-homoserine lactones (AHLs) by Gram-negative bacteria, and the diester AI-2 which is used by both ( Rémy et al., 2018 ; Tobias et al., 2020 ). A wide panel of AHLs has been reported. These molecules share a lactone ring but their lateral acyl chain varies in term of modifications and/or length ( Billot et al., 2020 ). Depending on the bacterial species, one or several specific AHLs can be produced and recognized by the cells to trigger virulence and biofilm formation . Biofilm formation allows bacteria to settle and limits the effectiveness of antimicrobial treatments ( Rémy et al., 2018 ; Bowler et al., 2020 ). To counteract AHL-mediated bacterial virulence, strategies to disrupt QS, referred to as quorum quenching (QQ), have been developed. Various approaches have been reported, such as the use of natural or synthetic QS inhibitors ( Rémy et al., 2018 ), autoinducer sequesters ( Grandclément et al., 2016 ; Rémy et al., 2018 ) and QQ enzymes (QQEs), which can inactivate autoinducers ( Rémy et al., 2018 ). One major advantage of QQ is that this strategy interferes with QS to prevent noxious behaviors without killing bacteria, thus minimizing the appearance of resistance mechanisms ( Rémy et al., 2018 ). QQEs are particularly promising in light of their capacity to act catalytically and extracellularly for efficient QS disruption ( Billot et al., 2020 ). Among the QQEs, the archaeal enzyme Sso Pox ( Hiblot et al., 2013 ) has been particularly investigated over the past decade considering its lactonase activity against a large panel of AHLs ( Hiblot et al., 2013 ) and its tremendous resistance to temperature, process and long-term storage ( Del Vecchio et al., 2009 ; Rémy et al., 2016 ). Enzyme engineering approaches made it possible to generate several Sso Pox variants with enhanced properties ( Billot et al., 2022 ). The QQ potential of these variants has been demonstrated against different bacterial strains, including human pathogens (such as Pseudomonas aeruginosa ) and environmental strains (such as Chromobacterium violaceum ). Significant phenotypic alterations upon Sso Pox treatment have been reported, including biofilm disruption ( Rémy et al., 2020 ; Mion et al., 2021 ), reduced virulence in in vitro ( Rémy et al., 2020 ; Mion et al., 2021 ) and in vivo models ( Hraiech et al., 2014 ), as well as modulation of antimicrobial production ( Mion et al., 2021 ). Here, the QQ effect of Sso Pox against Burkholderia thailandensis was evaluated. Besides some rare cases of human infections ( Chang et al., 2017 ; Gee et al., 2018 ), B. thailandensis is of special interest as it constitutes a relevant study model for Burkholderia pseudomallei due to their high genome relatedness ( Majerczyk et al., 2014b ). B. pseudomallei is the causative agent of melioidosis ( Klaus et al., 2018 ) and a potential bioterrorism agent ( Oliveira et al., 2020 ). B. thailandensis constitutes a safe surrogate for studying and tackling B. pseudomallei virulence. Three QS systems were identified in B. thailandensis , each system being composed of an AHL synthase and a receptor to produce and sense a specific AHL signal. These three systems are shared with B. pseudomallei ( Majerczyk et al., 2014b ), QS-1 (BtaI1-BtaR1) QS-2 (BtaI2-BtaR2) and QS-3 (BtaI3-BtaR3) related to the production of C 8 -HSL, 3-OH-C 10 -HSL and 3-OH-C 8 -HSL respectively ( Majerczyk et al., 2014b ). In addition, two orphan receptors were also identified ( Majerczyk et al., 2014a ) (BtaR4 or MalR ( Gupta et al., 2017 ) and BtaR5). QS regulates various phenotypes in B. thailandensis such as biofilm formation ( Tseng et al., 2016 ), motility ( Ulrich et al., 2004 ) and antimicrobial molecules including bactobolin ( Klaus, 2020 ), 4-hydroxy-3-methyl-2-alkenylquinoline ( Majerczyk et al., 2014a ; Majerczyk et al., 2014b ) and malleilactone ( Majerczyk et al., 2014a ). In this study, the effect of Sso Pox V82I, an improved monovariant of Sso Pox with enhanced lactonase activity ( Billot et al., 2022 ), was selected and evaluated on B. thailandensis . The impact of Sso Pox V82I lactonase was assessed through a proteomic approach complemented by phenotypic analysis. The results obtained show that Sso Pox V82I treatment significantly disturbs B. thailandensis proteome and QS-related phenotypes, including the modulation of antimicrobial compound production.",
"discussion": "4 Discussion \n B. thailandensis E264 is a soil saprophyte bacteria and is considered to be a non-pathogenic homolog of B. pseudomallei . Both bacteria contain three QS-systems which regulate the expression of numerous genes as well as the production of secondary metabolites. In this study, the effect of Sso Pox V82I lactonase was evaluated on B. thailandensis through proteomic and phenotypic analysis. AHLs used by B. thailandensis were described as C 8 -HSL, 3-OH-C 8 -HSL and 3-OH-C 10 -HSL by LC/LC/MS analysis ( Duerkop et al., 2009 ). Here, in vitro degradation assays towards these lactones were used to evaluate the performance of four previously described Sso Pox variants, and Sso Pox V82I was chosen. The QQ ability of this variant was then confirmed on B. thailandensis and a significant reduction in biofilm formation and cell aggregation, two QS-related phenotypes of B. thailandensis , was achieved ( Chandler et al., 2009 ; Majerczyk et al., 2014a ; Tseng et al., 2016 ). To get a broader picture of global changes involved by QS and QQ, proteomic analyses were performed. Proteomic analyses provided evidence that QQ by Sso Pox V82I treatment generates a huge shift in B. thailandensis with 18.3% of the proteome being significantly impacted. Several metabolic pathways were impacted, including primary metabolites such as sugar and amino acid metabolisms. AHLs synthesis is directly related to methionine metabolism, therefore, in addition to AHLs signal disruption, QQ might also generate an improper balance in the primary metabolism, impacting many metabolic pathways. Proteomic analysis revealed that the treatment with Sso Pox V82I results in activation of numerous pathways: 72% of impacted proteins were upregulated, suggesting that QS represses several key functions in B. thailandensis . In contrast, previous studies, using other Sso Pox variants on P. aeruginosa and C. violaceum did not induce such an upregulation upon lactonase treatment, suggesting that QS equally upregulated or downregulated proteins on these strains ( Rémy et al., 2020 ; Mion et al., 2021 ). QS regulation is known to be complex in B. thailandensis , and could repress or activate AHL-responsive genes depending on the growth stage ( Majerczyk et al., 2014a ). To confirm proteomic changes at the phenotypic level, different bioassays were performed which showed that motility and proteolytic activities were increased after Sso Pox V82I treatment and that antimicrobial compound production and secretion were modulated. Indeed, competition assays using cell-free supernatants showed that the production of an antifungal molecule was induced by QQ while the production of an antibacterial molecule was repressed. Motility was already known to be negatively regulated by QS in B. thailandensis ( Ulrich et al., 2004 ; Chandler et al., 2009 ; Passos da Silva et al., 2017 ), and the increased motility observed after QQ treatment confirmed these results. Regarding proteolytic activity, previous studies did not show any effect on protease production in a QS-mutant strain, suggesting that proteolytic activity was not related to QS ( Chandler et al., 2009 ). Under our conditions and using a QQE, protease activity was increased, as observed on milk agar plates, suggesting that proteolytic activity is negatively regulated by QS in B. thailandensis . Biosynthesis of some antimicrobial molecules and secondary metabolites, in B. thailandensis , was also previously described to be QS-controlled ( Majerczyk et al., 2014a ). In our conditions, QQ treatment by Sso Pox V82I led to a decrease in bactericidal effect against C. violaceum and S. aureus . Several proteins involved in antibiotic molecule synthesis (including malleilactone, terphenyl and bactobolin) were identified as being downregulated by Sso Pox in proteomic analyses. Malleilactone is a cytotoxin that contributes to virulence in B. pseudomallei and B. thailandensis ( Klaus et al., 2018 ), and its bactericidal effect against Gram positive bacteria was described (including in S. aureus and B. subtilis) ( Biggins et al., 2012 ), and bactericidal activity against B. subtilis was reported for terphenyl ( Biggins et al., 2011 ). Proteins related to bacteriocin biosynthesis ( Russell et al., 2014 ; Majerczyk et al., 2016 ; Mao et al., 2017 ) were identified and downregulated by lactonase treatment in proteomic analysis. In addition, proteins involved in bactobolin production were decreased after Sso Pox V82I treatment. Bactobolin is a well described antibiotic compound against different bacteria such as S. aureus ( Klaus, 2020 ) and C. violaceum ( Chandler et al., 2012a ). Synthesis of bactobolin is controlled by QS-2 (3-OH-C 10 -HSL, BtaI2-BtaR2) ( Seyedsayamdost et al., 2010 ; Majerczyk et al., 2014a ) and has been described as the sole active antibiotic compound remaining in cell-free supernatants after 0.22 µm filtration ( Klaus et al., 2020 ). Furthermore, LC-MS/MS analysis revealed a reduced response in bactobolin in treated samples. It is, therefore, probable that the bactericidal effect observed without Sso Pox V82I treatment is due to bactobolin production. Interestingly, while decreasing antibacterial effect, QQ treatment led to a strong increase in antifungal activity against A. niger, F. graminearum and S. cerevisiae. A similar increase in antifungal activity was also reported in Burkholderia ambifaria , a member of the Burkholderia cepacia complex , when hmqG or hmqA , two genes belonging to the hmqABCDEFG operon responsible for HMAQ production and under QS regulation ( Coulon et al., 2019 ), were mutated ( Vial et al., 2008 ). This operon was also described in B. thailandensis and B. pseudomallei ( Vial et al., 2008 ). In addition to an increase in antifungal activity, hmqG or hmqA mutations led to an intensification in proteolytic activity ( Vial et al., 2008 ). Here, the HmqG protein (WP_009893822.1) was identified in the proteomic analysis as being downregulated by Sso Pox V82I treatment, confirming previous results obtained in B. ambifaria. However, when extractions were performed and analyzed in LC-MS, HMAQ molecules and derivatives were identified, but a high variability between samples was observed and did not allow to draw conclusion about the impact of QQ on these molecules. Interestingly, these extractions were tested and, surprisingly, the extracted supernatants had lost their antifungal effect, suggesting a potential alteration of the bioactive compounds during the extraction step. Previous studies have demonstrated the conservation of QS regulons and QS-regulated functions in B. pseudomallei and B. thailandensis ( Majerczyk et al., 2014b ) and within Burkholderia cepacia complex ( Wopperer et al., 2006 ). In addition, previous studies investigated QS disruption using QS mutants ( Ulrich et al., 2004 ; Chandler et al., 2009 ; Tseng et al., 2016 ) and QQ lactonase ( Ulrich, 2004 ; Wopperer et al., 2006 ) in Burkholderia species have demonstrated that QS is involved in carbon metabolism and the regulation of phenotypes such as biofilm formation, motility and proteolytic activity. Here, the global proteomic approach offers an overview of the implication of QS disruption in B. thailandensis and could be extended to other Burkholderia species. Proteomic and phenotypic analyses outline the importance of QS disruption in antimicrobial production modulation and its impact on the interbacterial or interkingdom competition model. Recent studies on several microbiota such as human gut microbiota ( Coquant et al., 2021 ) and human oral cavity microbiota ( Muras et al., 2020 ) have detected the presence of AHLs. In addition, their production could be related to inflammatory diseases in gut and dental plaque formation. In environmental isolates, AHLs have been detected in coral and their presence is related to coral bleaching ( Zhou et al., 2020 ). A recent study using an Sso Pox variant demonstrated biofilm disruption in a complex sample from soil and revealed that Sso Pox impacted not only bacteria with an AHLs QS system but also bacteria not known to sense or produce AHLs ( Schwab et al., 2019 ). Using a lactonase enzyme, another QQ study confirmed the role of QS on oral microbiota and the impact of a QQE on oral biofilm formation ( Parga et al., 2023 ). These studies highlight the impact of QQ in a heterogeneous population through the modulation of antimicrobial compounds such as antibiotics, fungicides, and secretion systems. These molecules can induce dysbiosis and led to pathology genesis. Taken together, these studies open the way for further investigations of the impact of QQ on microbiota samples, to gain a better understanding of these complex microenvironments."
} | 4,102 |
31080272 | null | s2 | 5,600 | {
"abstract": "Fabricating powerful neuromorphic chips the size of a thumb requires miniaturizing their basic units: synapses and neurons. The challenge for neurons is to scale them down to submicrometer diameters while maintaining the properties that allow for reliable information processing: high signal to noise ratio, endurance, stability, reproducibility. In this work, we show that compact spin-torque nano-oscillators can naturally implement such neurons, and quantify their ability to realize an actual cognitive task. In particular, we show that they can naturally implement reservoir computing with high performance and detail the recipes for this capability."
} | 163 |
29018433 | PMC5622926 | pmc | 5,601 | {
"abstract": "Effects of biodiversity on productivity are more likely to be expressed when there is greater potential for niche complementarity. In soil, chemically complex pools of nutrient resources should provide more opportunities for niche complementarity than chemically simple pools. Ectomycorrhizal (ECM) fungal genotypes can exhibit substantial variation in nutrient acquisition traits and are key components of soil biodiversity. Here, we tested the hypothesis that increasing the chemical complexity and forms of soil nutrients would enhance the effects of intraspecific ECM diversity on host plant and fungal productivity. In pure culture, we found substantial variation in growth of strains of the ECM fungus Laccaria bicolor on a range of inorganic and organic forms of nutrients. Subsequent experiments examined the effects of intraspecific identity and richness using Scots pine ( Pinus sylvestris ) seedlings colonized with different strains of L. bicolor growing on substrates supplemented with either inorganic or organic forms of nitrogen and phosphorus. Intraspecific identity effects on plant productivity were only found under the inorganic nutrient amendment, whereas intraspecific identity affected fungal productivity to a similar extent under both nutrient treatments. Overall, there were no significant effects of intraspecific richness on plant and fungal productivity. Our findings suggest soil nutrient composition does not interact strongly with ECM intraspecific richness, at least under experimental conditions where mineral nutrients were not limiting. Under these conditions, intraspecific identity of ECM fungi becomes more important than richness in modulating plant and fungal performance.",
"conclusion": "Conclusion Intraspecific variation in L. bicolor nutrient resource utilization did not interact with substrate nutrient complexity when the fungi were grown in symbiosis with Scots pine. As similar levels of intraspecific variation in nutrient utilization have been observed with other ECM species (e.g., Amanita muscaria , Sawyer et al., 2003 ; Hebeloma cylindrosporum , Guidot et al., 2005 ), our findings suggest that ECM intraspecific richness may be important for plant and fungal productivity under nutrient limiting conditions, relative to resource heterogeneity under higher resource conditions. Our data suggests intraspecific identity of ECM fungi becomes more important than richness in modulating plant and fungal performance under nutrient rich conditions.",
"introduction": "Introduction Research on biodiversity-ecosystem function relationships has tended to focus on species diversity ( Loreau et al., 2001 ; Tilman et al., 2014 ) with much less emphasis on within-species diversity ( Hughes et al., 2008 ). Increasingly, however, intraspecific diversity has been shown to have wide-ranging effects on numerous ecosystem processes in a myriad of systems ( Hughes et al., 2008 ; Johnson et al., 2012 ). Yet, little is known about the nature of the biodiversity-ecosystem function relationship with respect to forest ecosystems and intraspecific diversity of soil microorganisms. A key group of forest soil microorganisms are the ectomycorrhizal (ECM) fungi, which compose a significant proportion of the soil microbial biomass ( Högberg and Högberg, 2002 ), and colonize the majority of the fine root systems of trees in boreal and temperate biomes ( Smith and Read, 2008 ). Ectomycorrhizal fungi have important functional roles in forest ecosystems, particularly for biogeochemical cycling and nutrient acquisition to trees ( Smith and Read, 2008 ). Molecular analyses of ECM fungal communities demonstrate considerable variation both within and between species, at a range of spatial scales ( Johnson et al., 2012 ). Indeed, the roots of an individual Populus tremula tree were found to support between 182 and 207 species of ECM fungi, and 23 ITS genotypes of Cenococcum geophilum ( Bahram et al., 2010 ). Recent findings indicate that intraspecific diversity of ECM fungi can have significant impacts on ecosystem processes and multifunctionality ( Wilkinson et al., 2010 , 2012 ; Hazard et al., 2017 ). For example, intraspecific diversity of Paxillus obscurosporus was found to affect fungal productivity and respiration in pure culture ( Wilkinson et al., 2010 ). A more challenging set of experiments found that intraspecific diversity of Laccaria bicolor affected productivity of host plants, fungal productivity and soil nutrient retention, and that the effect of richness on these metrics were stronger at the intra- versus interspecific level of diversity ( Hazard et al., 2017 ). These studies suggest that intraspecific identity and richness of ECM fungi can play a direct role in influencing forest ecosystem functioning, and complementarity, derived from niche differentiation and facilitation, is likely the explanatory mechanism ( Johnson et al., 2012 ). This is supported by the finding that L. bicolor populations overyield when in mixture compared to what would be predicted from their performance in monoculture ( Hazard et al., 2017 ). While these studies highlight the need to consider the role of intraspecific diversity in regulating ecosystem functioning, a major gap in understanding concerns the conditions in which such effects are likely to occur. Positive, negative or neutral effects of ECM intraspecific diversity on ecosystem functioning metrics, such as plant productivity, are likely context dependent, as seen in studies manipulating interspecific richness. For example, the biomass of birch plants, originally associated with eight ECM species, was doubled in comparison to single ECM associations but only when soil fertility was low ( Jonsson et al., 2001 ). Thus, there is a need to conduct experiments under various abiotic environmental conditions to increase predictability power of the impact ECM diversity could have on forest ecosystem functioning. The productivity of boreal and temperate forests is contingent on nitrogen (N) availability: these systems are typically low in available N ( LeBauer and Treseder, 2008 ; Inselsbacher and Näsholm, 2012 ). Characteristically, ECM plants dominate these forests, where their associated ECM fungi possess the ability to scavenge for essential nutrients from the soil, in particular N and phosphorus (P), and transfer a portion of these nutrients to their tree hosts in exchange for reduced carbon ( Smith and Read, 2008 ). Boreal and temperate forest soils contain both inorganic and organic sources of N and P, although organic forms are found in much greater quantities ( Inselsbacher and Näsholm, 2012 ; Bol et al., 2016 ). ECM fungi have evolved to utilize both inorganic forms and low molecular weight organic forms of N such as amino acids, which can be taken-up intact (e.g., Näsholm et al., 1998 ), and ECM fungi can also grow on organic forms of P such as phytic acid ( Antibus et al., 1992 ). Differences in nutrient complexities in soil may be an important mechanism by which niche separation and substrate use preferences occur that would reduce competition between individuals. Resource heterogeneity in soils may also be a key selective force leading to differences in functional traits of ECM fungi, given that resource heterogeneity is an important factor that supports the biodiversity of grassland and animal communities ( Kerr and Packer, 1997 ; Harpole and Tilman, 2007 ). Evidence from culture studies has shown that species of ECM fungi differ in their utilization of N sources ( Abuzinadah and Read, 1988 ; Finlay et al., 1992 ; Anderson et al., 1999 ), and even between individuals within a species ( Finlay et al., 1992 ; Anderson et al., 1999 ; Sawyer et al., 2003 ; Guidot et al., 2005 ). It is likely that strains of ECM fungi also differ in the ability to utilize organic P forms, as evident by variation in the production of organic P degrading enzymes ( Ho, 1989 ; Meyselle et al., 1991 ; Antibus et al., 1992 ). For example, strains of L. bicolor have been shown to vary in their solubilization of inorganic phosphorus sources in culture ( Nguyen et al., 1992 ; de la Bastide et al., 1995 ). It is therefore plausible that due to differences in N and P use efficiencies between strains within a species that ECM intraspecific diversity could affect components of the N cycle in forest ecosystems. Also, as host plant growth can be contingent on the ECM strain ( Hedh et al., 2009 ), intraspecific diversity could have a key role in tree productivity. Indeed, in Hazard et al. (2017) , L. bicolor intraspecific diversity was found to have considerable effects on the productivity of both fungi and Scots pine hosts in microcosms. Their experimental design consisted of a gradient of intraspecific richness, with each strain in monoculture and in mixture to enable testing for both intraspecific identity and richness effects, and transgressive overyielding ( Loreau, 1998 ; Loreau and Hector, 2001 ). Here we use the same L. bicolor strains and experimental design, but with the growth medium enriched with either inorganic or organic forms of N and P nutrient resources. An a priori pure culture experiment showed the ability of the L. bicolor strains to differentially utilize forms of inorganic and organic N, and showed greater variation in utilization for organic versus inorganic N, and suggesting the potential for nutrient partitioning. Here, we test for direct effects of L. bicolor intraspecific diversity on Scots pine and fungal productivity, and determine whether these effects are driven by interactions among fungal strains and between strains and nutrient complexity. We predict that the effects of L. bicolor intraspecific identity and richness on Scots pine plant and fungal productivity will be greatest when the chemical composition and variety of nutrient resources is complex, due to the greater likelihood of niche partitioning.",
"discussion": "Discussion In this study, we tested the effect of L. bicolor intraspecific identity and richness on the productivity of Scots pine and associated ECM fungal productivity in microcosms supplied with either an inorganic or organic composition of N and P nutrient resources. We predicted that identity and richness effects would be greatest when the nutrient resources in the substrate are chemically complex and diverse, due to the greater likelihood of niche partitioning ( Turner, 2008 ). This prediction was supported by our observation of significant variation in growth of L. bicolor on a range of nutrient forms. When plants were grown in symbiosis, however, the experimental findings did not support our prediction. Intraspecific identity effects on plant productivity were only found when the Scots pine seedlings were grown in the inorganic nutrient substrate. Despite that intraspecific identity effects on fungal productivity were found under both the inorganic and organic nutrient substrate, the effects were not greater in response to the organic nutrient treatment. Furthermore, no effects of L. bicolor intraspecific richness on Scots pine plant and fungal productivity were found under either nutrient treatment. The lack of richness and identity effects on productivity under the organic nutrient treatment reflects similar utilization of the nutrient resources rather than nutrient specialization by the L. bicolor strains. Our current understanding of intraspecific ECM variation in N source use is largely based on pure culture studies ( Finlay et al., 1992 ; Anderson et al., 1999 ; Sawyer et al., 2003 ; Guidot et al., 2005 ), which can only serve as an estimate of the strains fundamental niche ( Bogar and Peay, 2017 ). We demonstrated intraspecific variation in organic N source utilization by L. bicolor strains in pure culture, but these findings were not consistent with the effects on plant and fungal productivity when in symbiosis. For example, there were significant differences in mycelial biomass between the strains LbC and LbD when grown on media with glutamic acid and phenylalanine as the sole source of N, but not when grown on ammonium nitrate after 28 days. Furthermore, in comparison to ammonium nitrate, mycelial biomass was greater on glutamic acid and glycine. However, contrary to these findings, when in monoculture, Scots pine inoculated with monocultures of LbC and LbD did not differ in ECM-RRL in response to manipulation of nutrient resources, and when in mixture, additive effects were found under the inorganic nutrient substrate but not for the organic nutrient substrate. These contrasting findings could reflect the realized niches of the strains in regards to soil nutrient resources. Further examination of organic N source utilization and transfer to host plants, and competition for nutrient resources between strains of ECM species within soil would help to define realized niches. It is, however, possible that free-living microbes could have invaded the initially sterile microcosm substrate and caused mineralization of the organic N and P forms and thereby releasing inorganic nutrient forms and reducing the potential for niche partitioning by the L. bicolor strains. Alternatively, nutrient complexity effects on intraspecific richness and identity for plant and fungal productivity may only occur if plants are nutrient starved. For example, it has been shown that when soil fertility is low the biomass of birch plants, originally associated with eight ECM species, doubled in comparison to single ECM associations ( Jonsson et al., 2001 ). In our study, the mean shoot N and P concentrations (37 ± 9 mg N g -1 ; 9 ± 2 mg P g -1 ) of Scots pine relative to the optimal content for Scots pine needles (13–18 mg N g -1 ; 1.6–2.2 mg P g -1 ; Moilanen et al., 2010 ), suggest the plants were not nutrient limited. Soil nutrient availability is likely a strong influencing factor on complementarity and competition intensity among strains for nutrient resources ( Jonsson et al., 2001 ). Furthermore, the lack of intraspecific richness effects on plant and fungal productivity could be related to the number of available nutrient resources rather than the chemical form of the nutrients. In this study, the organic nutrient treatment was comprised of four N sources, whereas the inorganic nutrient treatment was comprised of two. Based on resource competition theory, the addition of a limiting resource (e.g., N from the addition of glutamic acid) making that resource (i.e., N) no longer limited could reduce the competitive trade-off of the strains ( Tilman, 1982 ). Competitive trade-offs have recently been reported for ECM species at the root system scale ( Moeller and Peay, 2016 ), and presumably competitive trade-offs could also occur between individuals within an ECM species. In this scenario, strain co-existence would be predicted to decrease with increasing resource number ( Harpole et al., 2016 ). To test this hypothesis, ECM strain abundance data would be required, and an experimental design that disentangles nutrient complexity from resource number effects on productivity. Although identity and richness effects of plant and fungal productivity were not greater in the organic nutrient treatment, the results do suggest that identity and richness effects on plant and fungal productivity are context dependant. In this study, L. bicolor identity effects on plant productivity were only found under the inorganic nutrient substrate, and although identity effects on fungal productivity were found for both nutrient complexities, different trends in productivity emerged between the nutrient treatments. For example, under the inorganic nutrient substrate, Scots pine ECM-RRL was greater in the L. bicolor mixtures than the monocultures, and transgressive overyielding occurred, suggesting a degree of complementarity. In the organic nutrient substrate, Scots pine ECM-RRL varied within and amongst the L. bicolor mixtures and monocultures. This finding possibly reflects variation in nutrient source utilization by the L. bicolor strains, but also that competition for a nutrient resource could be occurring in some of the mixtures. No significant richness effects on plant and fungal productivity were found, which contrasts with past work ( Hazard et al., 2017 ). Hazard et al. (2017) used the same L. bicolor strains, similar experimental design but with a different, single growth medium and lower concentration of the inorganic nutrient solution, and found that intraspecific richness increased Scots pine root productivity and ECM-RRL. In the present study, significant transgressive overyielding of ECM-RRL occurred in the mixtures that contained two L. bicolor strains, and there was a weak trend, of increasing ECM-RRL with increasing intraspecific L. bicolor richness under the inorganic nutrient treatment. Conversely, L. bicolor identity affected plant and fungal productivity, an effect also seen by Hazard et al. (2017) , but in the present study, the identity effects were stronger. Shoot height, shoot biomass and root length of Scots pine were significantly greater under the organic nutrient treatment for some of the L. bicolor strains, mostly for the monocultures, whereas shoot P content of Scots pine was greater under the inorganic nutrient treatment. The difference in the strength of the richness and identity relationships with plant and fungal productivity found between these studies might reflect the different growth media, i.e., peat-vermiculite versus sand-vermiculite, and the difference in nutrient availability. Peat-vermiculite may support more physical niches and diverse forms of nutrients, due to the inherit differences in water holding capacity, aeration, bulk density and cation exchange capacity ( Sahin et al., 2002 ). Also, a lower nutrient rich substrate may enhance resource complementarity ( Jonsson et al., 2001 ) between the L. bicolor strains. Collectively, these findings suggest that substrate physical and chemical conditions interact with ECM fungal richness and identity to affect plant and fungal productivity metrics. Also, these findings suggest that the substrate properties are driving the development of certain functional traits by interacting with particular L. bicolor strains. Further studies conducted with other ECM species and strains within a species, and under more natural nutrient conditions and across a range of forest soil types would be valuable for identifying patterns for ECM diversity effects on plant and fungal productivity."
} | 4,640 |
39182035 | PMC11344385 | pmc | 5,602 | {
"abstract": "Background At lower concentrations copper (Cu), zinc (Zn) and nickel (Ni) are trace metals essential for some bacterial enzymes. At higher concentrations they might alter and inhibit microbial functioning in a bioreactor treating wastewater. We investigated the effect of incremental concentrations of Cu, Zn and Ni on the bacterial community structure and their metabolic functions by shotgun metagenomics. Metal concentrations reported in previous studies to inhibit bacterial metabolism were investigated. Results At 31.5 μM Cu, 112.4 μM Ni and 122.3 μM Zn, the most abundant bacteria were Achromobacter and Agrobacterium . When the metal concentration increased 2 or fivefold their abundance decreased and members of Delftia , Stenotrophomonas and Sphingomonas dominated. Although the heterotrophic metabolic functions based on the gene profile was not affected when the metal concentration increased, changes in the sulfur biogeochemical cycle were detected. Despite the large variations in the bacterial community structure when concentrations of Cu, Zn and Ni increased in the bioreactor, functional changes in carbon metabolism were small. Conclusions Community richness and diversity replacement indexes decreased significantly with increased metal concentration. Delftia antagonized Pseudomonas and members of Xanthomonadaceae . The relative abundance of most bacterial genes remained unchanged despite a five-fold increase in the metal concentration, but that of some EPS genes required for exopolysaccharide synthesis, and those related to the reduction of nitrite to nitrous oxide decreased which may alter the bioreactor functioning. Supplementary Information The online version contains supplementary material available at 10.1186/s12866-024-03437-8.",
"conclusion": "Conclusion Low concentrations of copper, zinc and nickel (31.5 μM copper, 112.4 μM nickel, 122.3 μM zinc) altered the bacterial community structure in a bioreactor seeded with wastewater that changed even more with higher concentrations of these metals. Achromobacter and Agrobacterium were enriched at low concentration of the metals, while Delftia, Stenotrophomonas and Sphingomonas at higher concentrations. Delftia was antagonistic towards Pseudomonas and other Xanthomonadaceae . Assimilatory reduction of nitrate, and assimilation of methylmercaptopropionate to dimethylsulfide was negatively affected by the metal concentration, which may impair the removal of nitrate and some organo-sulfur compounds in metal contaminated water. Functional metagenomics produced by shotgun sequencing help to improve our understanding on what mechanisms and population dynamics (e.g., antagonisms) occur in the microbial community.",
"discussion": "Discussion Bacterial community diversity and structure In this study, the presence of the metals reduced the microbial diversity in the bioreactor as has been reported before [ 24 ]. In this study, the richness and diversity replacement decreased as the metal concentration increased, but not for Hill numbers at q = 1 and q = 2. In other words, dominant genera were replaced by others when the metal concentration increased. As such, the presence of the genera or species defined the dominance or evenness. Achromobacter and Agrobacterium were the most dominant genera in the HM-1 sample, while Delftia , Stenotrophomonas and Sphingobacterium in the HM-2 and HM-5 samples. This might be explained by the increased metal concentration, or the differential impairment of metabolic functions by the metals. Members of these genera have strategies to resist metals [ 7 , 23 , 27 ]. Bacteria of the genera [ 22 ] , Delftia [ 2 ] , Stenotrophomonas [ 28 ] and Agrobacterium [ 38 ] have been found to adhere to surfaces via external polymeric substances [ 15 ]. This might lead to the formation of sediment microparticles that might immobilize metals reducing their bioavailability and toxicity [ 34 ]. This phenomenon has been observed for other genera, such as Pseudomonas, Accumulibacter, Competibacter, Mesorhizobium, Xanthomonas and Pseudoxanthomonas [ 15 , 22 , 23 , 34 , 36 ]. Some of these genera, such as Pseudomonas (relative abundance 0.3%), Mesorhizobium (0.06%) , Xanthomonas (0.3%) and Pseudoxanthomonas (0.2%) were found in the metal amended bioreactor, but their relative abundance was low. The relative abundances of Pseudomonas, Xanthomonas and an unknown Xanthomonadaceae were correlated and the partial correlation between them decreased (r 2 ) when the metal concentration was deleted from the correlation coefficients, suggesting that the apparent correlations between the genera was due mainly to the metals. Metabolic functionality It was hypothesized that changes in bacterial richness and diversity would also result in changes in the metabolic functions in the bioreactor samples, but that was mostly not the case. The relative abundance of most metabolic functions did not change despite the increase in metal concentration, i.e. the bacterial metabolic functionality was more stable than the bacterial community structure despite the variations in metal concentration. Heterotrophic metabolism was the main function, and carbohydrate and amino acid metabolism the most abundant. This may suggest that despite the increases in metal concentration, organic material decomposition was not affected. This agrees with our previous report using putative/predictive functions from 16S metagenomics [ 1 ]. The reduction of nitrite to nitric oxide ( nirK, nirS ) and from nitrite to nitrous oxide ( nifK/nirS, norBC ) and DMSP demethylation, however, all decreased significantly with increased metal concentrations. In all these metabolic functions, oxido-reduction biochemical reactions are involved. The narB, nasA, nifK, nirS, nasA, and norB genes involved in the N cycle encode for proteins binding iron, while narB and nifK may bind molybdenum ions, whereas norB is from the cytochrome family. The decrease in their relative abundance might reflect the lethality of these metals to bacteria that possess those genes. The DMSP demethylation occurs thanks to proteins encoded by the dmd gene [ 16 ]. Proteins encoded by the dmdD gene bind magnesium ions (Uniprot) and products of dmdBCD are still being investigated. Dey [ 16 ] reported that dmdBCD products are involved in carbon and sulfur fixation by plankton and that a decrease in their protein expression might result in a higher production of dimethylsulfide (DMS), an air contaminant in acid rain. The relative abundance of the metabolic functions showed only small changes. This would suggest that most of the functions are either required for the microbial community, e.g. autotrophy, or they were present and became detectable as the substrate became available, such as heterotrophy, in both bioreactors. In other words, the forces driving the occurrence of specific metabolic functions may include randomness, necessity and/or substrate availability. Apart from the different metal concentrations, conditions were similar in both bioreactors. It can thus be assumed that, despite biodiversity, the range of metabolic functions were similar as evidenced by our results. Contrario sensu, different conditions might be driving forces for changes in metabolic functions. For instance, in this study, the metals in the HM bioreactor might lead an increase in metal resistance mechanisms. Among them, the transformation and transport of metals, and EPS production are the most well-described mechanisms. The formers, transformation (redox) and transport mechanisms have been described in Lysinibacillus, Agrobacterium and some genera of the Xanthomonadaceae, such as Stenotrophomonas and Xanthomonas [ 28 – 30 , 32 , 38 ]. The relative abundance of Agrobacterium and Achromobacter decreased in the HM-2 and HM-5 samples, suggesting that the minimum inhibitory concentration of the metals was between that found in the HM-1 and HM-2 samples, which might also affect the metabolic functions within the bacterial community. One of the main metabolic-related mechanisms is the production of EPS from carbohydrates and proteins. The main pathways for EPS production comprise curdlan, xanthan, colanic and alginate synthesis, which involve crd, gum, wca and alg operons [ 34 ]. Succinoglycan ( exo operon), the vibrio-like polysaccharide ( cysE and vps operons), the enterobacterial antigen ( wec operon) and Poly-N-acetyl-glucosamine ( ica and pga operons) are other, but less frequent genes encoding for EPS. The crd, gum, wca and alg operons have been detected in Agrobacterium, Xanthomonas, Enterobacteriaceae and Pseudomonas [ 34 ]. The crd operon has been reported in Agrobacterium [ 34 ], but it does not appear yet in the RefSeq gene database or KEGG. In this study, the relative abundance of the wecA gene increased with increased metal concentration. This gene has been reported in members of the Enterobacteriaceae family, such as Escherichia , Yersinia and Klebsiella [ 9 ]. The relative abundance of these genera was low in the bioreactor (≤ 0.04%) and no correlation was found between them (or the Enterobacteriaceae family) and the wecA gene. As such, further research into the prevalence of the wecA gene in other bacterial groups might be needed as the EPS production genes may also help other genera to adhere to surfaces and survive. Although EPS production by Delftia has not yet been reported, its ability to autoaggregate and coaggregate in biofilms with other bacteria has been reported [ 2 ], which might explain its ability to colonize surfaces [ 15 ]. This coaggregation seems to be sugar-specific with Citrobacter and Enterobacter , but not with Pseudomonas , whose interaction appears to be protein–protein mediated [ 2 ]. In this study, Pseudomonas seemed to antagonize Delftia within the biofilm. It has been reported that a strain of Delftia was able to inhibit Pseudomonas growth, motility and biofilm production [ 31 ]. In this study , Delftia was also antagonistic against Xanthomonas and an unknown Xanthomonadaceae . This antagonism prevailed even after eliminating the metals from the correlation analysis, suggesting that it was independent of the presence of the metals. Also, Achromobacter, Delftia, Pseudomonas and Stenotrophomonas have been found to produce biofilms and adhere to them [ 15 ]. A strain of Achromobacter has been reported to denitrify and be able to immobilize 10 mg L −1 (89 μM approximately) cadmium ions (Cd) in wastewater [ 22 ]. Although this immobilization of Cd has been hypothesized to occur via precipitation with carbonate, its precipitation is larger with a composite sediment containing EPS. Achromobacter and Stenotrophomonas strains carrying copLAB genes that have been isolated from an agricultural field are able to resist 0.8 mM Cu. This ability of Achromobacter together with Agrobacterium to produce EPS [ 34 ] might explain their higher abundance in the HM-2 sample. As mentioned above, the EPS may be a scaffold for Delftia to grow better when the metal concentration increased. This might have helped Delftia to adhere to the biofilm and inhibit the growth and adhesion of other bacteria, as has been described previously [ 9 , 31 ]. This might explain why the relative abundance of Delftia was highest in the HM-5 sample by antagonizing Pseudomonas and other Xanthomonadaceae , and also why the abundance of Achromobacter decreased. These observations agreed with a previous study of [ 1 ] based on 16S metagenomics, where an increase in the relative abundance of Stenotrophomonas and Sphingomonas at high metal concentrations (HM-2 and HM-5) and the increase of Agrobacterium at low metal concentrations (HM-1) were reported. Although 16S metagenomics gave a good insight in the behavior of the bacterial community, shotgun metagenomics allow a deeper knowledge in microbiota diversity and gene/function prevalence, such as the EPS production genes."
} | 3,018 |
31179193 | PMC6545100 | pmc | 5,603 | {
"abstract": "Background Cultivars of bahiagrass ( Paspalum notatum Flüggé) are widely used for pasture in the Southeastern USA. Soil microbial communities are unexplored in bahiagrass and they may be cultivar-dependent, as previously proven for other grass species. Understanding the influence of cultivar selection on soil microbial communities is crucial as microbiome taxa have repeatedly been shown to be directly linked to plant performance. Objectives This study aimed to determine whether different bahiagrass cultivars interactively influence soil bacterial and fungal communities. Methods Six bahiagrass cultivars (‘Argentine’, ‘Pensacola’, ‘Sand Mountain’, ‘Tifton 9’, ‘TifQuik’, and ‘UF-Riata’) were grown in a randomized complete block design with four replicate plots of 4.6 × 1.8 m per cultivar in a Rhodic Kandiudults soil in Northwest Florida, USA. Three soil subsamples per replicate plot were randomly collected. Soil DNA was extracted and bacterial 16S ribosomal RNA and fungal ribosomal internal transcribed spacer 1 genes were amplified and sequenced with one Illumina Miseq Nano. Results The soil bacterial and fungal community across bahiagrass cultivars showed similarities with communities recovered from other grassland ecosystems. Few differences in community composition and diversity of soil bacteria among cultivars were detected; none were detected for soil fungi. The relative abundance of sequences assigned to nitrite-oxidizing Nitrospira was greater under ‘Sand Mountain’ than ‘UF-Riata’. Indicator species analysis revealed that several bacterial and fungal indicators associated with either a single cultivar or a combination of cultivars are likely to be plant pathogens or antagonists. Conclusions Our results suggest a low impact of plant cultivar choice on the soil bacterial community composition, whereas the soil fungal community was unaffected. Shifts in the relative abundance of Nitrospira members in response to cultivar choice may have implications for soil N dynamics. The cultivars associated with presumptive plant pathogens or antagonists indicates that the ability of bahiagrass to control plant pathogens may be cultivar-dependent, however, physiological studies on plant-microbe interactions are required to confirm this presumption. We therefore suggest that future studies should explore the potential of different bahiagrass cultivars on plant pathogen control, particularly in sod-based crop rotation.",
"conclusion": "Conclusions We detected a few differences in community composition and diversity of soil bacteria among bahiagrass cultivars, suggesting a moderate impact of cultivar choice on the soil bacterial community. Further, cultivar choice affected the relative abundance of sequences assigned to members of the nitrite-oxidizing bacterial genus Nitrospira with possible implications for soil-N dynamics. In contrast, soil fungal composition and diversity was not altered by the different cultivars. Several bacterial and fungal indicator species assigned to either a single cultivar or a combination of cultivars were presumptive plant pathogens or antagonists. In view of this, we suggest future work that explores the potential of bahiagrass cultivars to control plant pathogens.",
"introduction": "Introduction Bahiagrass ( Paspalum notatum Flüggé), native to South America ( Burton, 1967 ), is a widespread, warm-season perennial, commonly used as pasture in the Southeastern USA. Following its introduction into many countries worldwide, the sod-forming grass is also common in Australia and Japan ( Hirata, 2000 ; Wilson, 1987 ) and has become naturalized in the USA. It was first introduced into the USA in 1913 ( Scott, 1920 ) and is extensively cultivated on more than 1.5 million hectares in southeast USA, making it the most common and widely used perennial grass across southern states ( Newman, Vendramini & Blount, 2011 ). Bahiagrass grows well in sandy, low fertile soils, requires low inputs, and it exhibits tolerance towards short-term drought and flooding events as well as continuous cattle stocking ( Gates, Quarin & Pedreira, 2004 ; Newman, Vendramini & Blount, 2011 ). Low winter temperatures and aridity limit its geographic distribution ( Gates, Quarin & Pedreira, 2004 ). ‘Pensacola’, ‘Tifton 9’, ‘TifQuik’, and ‘UF-Riata’ are among the most popular cultivars in the Southeastern USA. They exhibit differences in growth habit, cold tolerance, seasonal and total yield, seed production and grazing tolerance ( Newman, Vendramini & Blount, 2011 ). Cultivars also can differ in their resistance to diseases ( Hancock et al., 2010 ; Trenholm, Cisar & Unruh, 2011 ). Further, cultivar-specific nutrient use efficiencies may reduce nitrate leaching and fertilizer input costs ( Wiesler & Horst, 1993 ; Liu, Hull & Duff, 1997 ; Baligar, Fageria & He, 2001 ). Therefore, cultivar choice is an important factor for the maintenance of soil health. It is well established that plant community composition and diversity influences the belowground microbial community and vice-versa ( Berg, 2009 ; Berg & Smalla, 2009 ; Van der Heijden et al., 1998 ; Kourtev, Ehrenfeld & Häggblom, 2003 ; Kowalchuk et al., 2002 ; Lange et al., 2015 ; Reynolds et al., 2003 ; Wardle et al., 2004 ; Zak et al., 2003 ). Beneficial plant-microbe interactions, such as mycorrhizal symbiosis or root colonization of plant growth-promoting rhizobacteria (PGPR) are known to enhance host plant growth ( Artursson, Finlay & Jansson, 2006 ; Lugtenberg & Kamilova, 2009 ), pathogen resistance ( Azcón-Aguilar & Barea, 1997 ; Harrier & Watson, 2004 ; Van Loon, Bakker & Pieterse, 1998 ; Maherali & Klironomos, 2007 ), and abiotic stress tolerance ( Evelin, Kapoor & Giri, 2009 ; Vurukonda et al., 2016 ; Wu, Zou & Xia, 2006 ; Yang, Kloepper & Ryu, 2009 ). Whereby belowground, mycorrhiza symbionts depend on organic carbon supply via host roots ( Smith & Read, 2008 ) and PGPR can be attracted via root exudates ( Badri & Jorge, 2009 ; Somers, Vanderleyden & Srinivasan, 2004 ), creating a complex plant-microbe-soil feedback system ( Miki et al., 2010 ). Emerging evidence shows that plant cultivars can be one of the factors affecting the composition of the rhizosphere microbiome ( Briones et al., 2002 ; Dalmastri et al., 1999 ; Diab El Arab, Vilich & Sikora, 2001 ; Germida & Siciliano, 2001 ; Schweitzer et al., 2008 ). Different grass species have been shown to be capable of altering soil microbial communities, mainly due to differences in nutrient acquisition strategies and rhizodeposits ( Bardgett et al., 1999 ; Grayston et al., 1998 ; Vandenkoornhuyse et al., 2003 ). A few studies reported that rhizosphere bacterial populations vary across different grass cultivars ( Miller, Henken & Veen, 1989 ; Rodrigues et al., 2016 ), whereas the potential effect of different grass cultivars on the composition of fungal communities remains widely unexplored. Identifying alterations of the soil microbiome by cultivar choice is of importance as specific microorganisms can have specific lifestyles, including mutualism, parasitism or involvement in diverse saprotrophic activities. These processes are directly linked to the fitness of the host plants and soil fertility. Alterations of belowground microbial communities can have significant impact on plant performance. In several managed grassland ecosystems, Proteobacteria, Acidobacteria, Actinobacteria, and Bacteroidetes have been found to be the most abundant soil bacterial phyla ( Cao et al., 2017 ; Kaiser et al., 2016 ; Nacke et al., 2011 ; Rodrigues et al., 2016 ; Zhou et al., 2003 ). Members of these phyla contribute to essential soil functions, such as biological nitrogen fixation (BNF) ( Baldani et al., 1997 ). Further, beneficial rhizobacteria can stimulate plant growth via the production of plant hormones, suppress soil-borne plant pathogens, supply nutrients to plants and improve soil structure ( Berg, 2009 ; Hayat et al., 2010 ; Van der Heijden, Bardgett & Van Straalen, 2008 ; Weller et al., 2002 ). Hence, PGPR such as Arthrobacter , Azotobacter , Burkholderia , and Pseudomonas species have been used to enhance agricultural production for decades ( Bhattacharyya & Jha, 2012 ; Vessey, 2003 ). Besides bacteria, symbiotic associations with mycorrhizal fungi can improve plant resistance to pathogens ( Selosse, Baudoin, & Vandenkoornhuyse, 2004 ; Wehner et al., 2010 ) as well as improve plant nutrition, particularly by enhancing plant phosphorus (P) acquisition ( Li et al., 2006 ; Smith, Mette Grønlund & Andrew Smith, 2011 ; Smith, Smith & Jakobsen, 2003 ). Many arbuscular mycorrhizal (AM) fungi communities under grass have been shown to be dominated by the families Glomeraceae, Gigasporaceae and Acaulosporaceae ( Hiiesalu et al., 2014 ; Oehl et al., 2005 ; Xu et al., 2017 ). In some grassland soils, the genus Glomus was identified as the most abundant AM fungi ( Gai et al., 2009 ; Wang et al., 2003 ). Glomus is the largest genus of AM fungi described ( Schwarzott, Walker & Schüssler, 2001 ). In association with peanut ( Arachis hypogaea ) and lettuce ( Lactuca sativa ) plants, Glomus spp. were demonstrated to promote plant growth, P and micronutrient uptake ( Krishna & Bagyaraj, 1984 ) and increased drought tolerance ( Ruiz-Lozano, Azcon & Gomez, 1995 ). Using next generation amplicon sequencing, the aim of this study was to determine whether different bahiagrass cultivars interactively influence the belowground microbial community composition and diversity. To achieve this aim, we recovered bacterial 16S ribosomal RNA (16S rRNA) and fungal ribosomal internal transcribed spacer (ITS) 1 gene sequences from soil samples of six different bahiagrass cultivars grown in a randomized complete-block design. We hypothesized that bahiagrass cultivar choice affects the microbial community composition and diversity of both, soil bacteria and fungi. Given the significant role of soil microorganisms in soil nutrient cycling and plant nutrition, our research outcomes can provide insight into bahiagrass-associated soil bacterial and fungal communities, as well as the plant-microbe-soil feedback system among grass cultivars and better our understanding of the grassland ecosystem.",
"discussion": "Discussion Soil bacterial communities across bahiagrass cultivars The soil bacterial communities across managed bahiagrass cultivars exhibited parallels to the communities of diverse grassland ecosystems at phylum and class level. For example, the top three dominant soil bacterial phyla across all bahiagrass plots (Proteobacteria, Acidobacteria, and Actinobacteria) as well as the dominance of the Alpha-, Delta-, and Gammaproteobacteria were also reported for managed grassland soils ( Cao et al., 2017 ; Nacke et al., 2011 ; Rodrigues et al., 2016 ; Zhou et al., 2003 ). Further, the greater relative abundance of the phyla Acidobacteria and Actinobacteria agrees with other studies investigating bacterial communities in grassland soils ( Kaiser et al., 2016 ; Nacke et al., 2011 ; Rodrigues et al., 2016 ; Will et al., 2010 ). The most abundant bacterial genus that was taxonomically assigned across cultivars of bahiagrass, Candidatus Udaeobacter , is ubiquitous in soils and frequently recovered using 16S rRNA gene sequencing approaches. Recently, Brewer et al. (2016) reported that an affiliate of this genus, Candidatus Udaeobacter copiosus , can account for almost one third of the soil bacterial taxa in grasslands. Further, Candidatus Udaeobacter copiosus has shown dominance in soil samples even across geographic distance ( Brewer et al., 2016 ). Despite its great relative abundance in soils worldwide, the ecology and physiology of members of the genus Candidatus Udaeobacter largely remain unknown. Our second most abundant soil bacterial genus ( Bradyrhizobium ) that matched our sequences was previously found as one of the most prominent genera in other grassland ecosystems ( Brewer et al., 2016 ; McCaig, Glover & Prosser, 1999 ; Thomson et al., 2010 ). Many Bradyrhizobium species have the ability to denitrify ( Bedmar, Robles & Delgado, 2005 ; Fernández et al., 2008 ; Kaneko et al., 2002 ; Mesa, Göttfert & Bedmar, 2001 ) and are proposed to play a key role in denitrification ( Jones et al., 2016 ). Moreover, several Bradyrhizobium affiliates are capable of fixing atmospheric N 2 and are considered to contribute significantly to BNF in soils ( Zahran, 1999 ). The abundance of Bradyrhizobium , however, cannot serve as an indicator of their N 2 fixation rates as shown in a recent study on native switchgrass ( Panicum virgatum ) ( Bahulikar et al., 2014 ). Thus, although a genetic potential for denitrification and BNF is given by our second most dominant soil bacterial genus, its contribution to N cycling in soil of bahiagrass remains unclear and requires further investigations on functional level. In line with the other dominant genera that we taxonomically assigned, the genus Candidatus Solibacter , our third most abundant genus, has been reported as one of the top genera recovered from grassland soils ( Kaiser et al., 2016 ). Even for the most frequently investigated affiliate of the genus, Candidatus Solibacter usitatus , detailed ecological and physiological information is still lacking ( Dedysh et al., 2017 ; Ward et al., 2009 ). Soil fungal communities across bahiagrass cultivars In line with previous results from grassland ecosystems ( Barnard, Osborne & Firestone, 2013 ; Chen et al., 2017 :201; Porras-Alfaro & Bayman, 2011 ; Tedersoo et al., 2014 ; Yang et al., 2017 ), sequences assigned to Ascomycota numerically dominated over all other fungal phyla across cultivars. The dominant fungal classes in our bahiagrass plots (Sordariomycetes, Glomeromycetes, Dothideomycetes, and Agaricomycetes) were similar to those found in Californian grassland soils ( Barnard, Osborne & Firestone, 2013 ). Many species of our most abundant taxonomically assigned fungal genus across cultivars, Penicillium , have been identified as plant growth-promoting fungi for several plants including grasses ( Khan et al., 2008 ; Wakelin et al., 2004 ; Whitelaw, Harden & Bender, 1997 ). A well reported mechanism of plant growth promotion by Penicillium spp. is their ability to solubilize P for plant nutrition in soil ( Asea, Kucey & Stewart, 1988 ; Kucey, 1987 ; Wakelin et al., 2004 ). We found that the potentially phytopathogenic genus Fusarium was assigned as the second most abundant genus across all plots. Fusarium spp. are cosmopolitans that are present in all types of ecosystems ( Summerell et al., 2010 ) and were reported to be one of the most abundant soil fungal taxa in some grassland ecosystems ( Khidir et al., 2010 ; Orgiazzi et al., 2012 ; Warcup, 1951 ; Yang et al., 2017 ). Fusarium diseases are, except under rare conditions, considered as not serious for bahiagrass under field conditions ( Singh, 2009 ). Besides Penicillium and Fusarium , sequences assigned to the genus Mortierella were dominant and have also been shown to be highly abundant in grassland soils ( Warcup, 1951 ; Yang et al., 2017 ). Members of the genus Mortierella are a diverse, ubiquitous and abundant group of filamentous fungi in soils that exhibit a saprophytic lifestyle ( Uehling et al., 2017 ; Wagner et al., 2013 ). Additionally, some species were recently described as root endophytes ( Bonito et al., 2016 ; Johnson et al., 2019 ). There is evidence that several Mortierella species can promote the growth of certain plant species whereby for some species, similar to Penicillium , one of the identified mechanisms for plant growth promotion is their ability to solubilize P for plant uptake ( Osorio & Habte, 2001 ; Osorio & Habte, 2013 ; Osorio & Habte, 2014 ; Sharma et al., 2013 ; Zhang et al., 2011 ). Soil bacteria and fungi under different bahiagrass cultivars Numerous studies have shown that plant cultivars or varieties can affect the composition of the associated soil rhizosphere bacterial and fungal communities ( Bell et al., 2014 ; Briones et al., 2002 ; Dalmastri et al., 1999 ; Diab El Arab, Vilich & Sikora, 2001 ; Germida & Siciliano, 2001 ; Jie, Liu & Cai, 2013 ; Schweitzer et al., 2008 ). Different grass cultivars can exhibit dissimilar nutrient requirements ( Ashworth et al., 2017 ; Oliveira et al., 2017 ) as well as root exudate quantities and qualities ( Christiansen-Weniger, Groneman & Van Veen, 1992 ; Guo, McCulley & McNear, 2015 ), which are likely to affect populations of root-associated microorganisms. The bahiagrass cultivars differed in productivity, stand establishment and growth rate, and temperature sensitivity ( Chambliss & Sollenberger, 1991 ; Newman, Vendramini & Blount, 2011 ). Thus, considering the holistic approach of plant-microbe-soil as a feedback system ( Miki et al., 2010 ), it is likely that different bahiagrass cultivars affect the rhizosphere microbiome and alter plant-microbe-soil traits. In our study, differences in microbial community composition in response to cultivar choice were only detected for bacterial communities. It should be noted that our soil samples were a mixture of rhizosphere and bulk soil, which may have contributed to the low number of detected differences in the composition and diversity of soil microbial communities among cultivars. Soil microbial community functional diversification is thought to be crucial for soil microbiome stability and resilience ( Griffiths & Philippot, 2013 ; Shade et al., 2012 ). Therefore, the comparingly low bacterial alpha diversity (Simpson’s index) and evenness (Simpson’s evenness) in TifQuik soil ( Fig. 2C ), may signal a decreased potential of the soil bacterial community to counter perturbations. Differences in community composition among cultivars were limited to bacterial communities among Argentine and Sand Mountain and Argentine and TifQuik. Cultivar choice further affected relative abundance of the cosmopolitan genus Nitrospira ( Fig. 3 ). Nitrospira affiliates are present in a wide range of habitats, including deep sea sediments ( Nunoura et al., 2015 ), cold deserts ( Gupta et al., 2015 ), and tropical sponges ( Sharp et al., 2007 ). Traditionally, members of Nitrospira are described as nitrite-oxidizing bacteria, performing the second oxidation-step in nitrification. Recently, however, Daims et al. (2015) reported complete nitrification by a member of the genus Nitrospira , which completely changes our understanding of ammonia-oxidizing and nitrite-oxidizing bacteria. Apart from their place in the nitrification pathway, the increased relative abundance of Nitrospira under Sand Mountain compared to UF-Riata ( Fig. 3 ) may indicate a greater potential for nitrite oxidation activity in soil of Sand Mountain. In 2014 and 2015, Dubeux et al. (2017) determined the bahiagrass yield and crude protein content of all six bahiagrass cultivars at our experimental site. The yield of Sand Mountain was among the greatest of all six cultivars and out-yielded Argentine. Although no statistically significant differences in crude protein content were detected among cultivars, it is worth mentioning that Sand Mountain showed the greatest mean crude protein content ( Dubeux et al., 2017 ). Wedin & Tilman (1990) reported a close relationship between soil-N cycling and the choice of perennial grass species. Several studies showed that certain grass species can suppress nitrification ( Ishikawa et al., 2003 ; Lata et al., 2004 ; O’Sullivan et al., 2016 ; Subbarao et al., 2009 ). In contrast, Hawkes et al. (2005) demonstrated that invasive grass species can increase nitrification rates and the abundance of ammonia-oxidizing bacteria in soil of Californian grassland. Studies that explored the role of grass root exudates on nitrification mainly focused on nitrification inhibition as a strategy for reduced nitrate leaching from soil. Numerous studies reported nitrification inhibitors in root exudates of grasses ( Subbarao et al., 2006 ; Subbarao et al., 2009 ; Sun et al., 2016 ; Zakir et al., 2008 ). The composition of grass root-exudates has been shown to be affect by both cultivar and fungal endophytes ( Guo, McCulley & McNear, 2015 ). It remains unclear whether certain bahiagrass cultivars affect nitrification rates. However, we speculate that some cultivars may promote or less supress nitrifying soil microorganisms to increase N availability, particularly in the absence of N fertilization like at our experimental site. Sand Mountain further harboured two indicator species, one OTU anchored in the genus Pajaroellobacter and the other in the genus Bauldia ( Table S3 ). The genus Pajaroellobacter is not well characterized, except for Pajaroellobacter abortibovis , the etiologic agent of epizootic bovine abortion in cattle, which is a vector transmitted disease by the tick Ornithodoros coriaceus ( Brooks et al., 2016 ; King et al., 2005 ). Likewise, the genus Bauldia is largely unexplored. Sequences assigned to the genus Haliangium was found characteristic for the cultivars Pensacola and Tifton 9 ( Table S3 ). Haliangium spp. have been recovered from soil samples before, even with great geographic distance among samples ( Ding et al., 2014 ; Fulthorpe et al., 2008 ). Some members of Haliangium have the capability to produce the antifungal metabolite haliangicin which can supress the growth of a broad range of fungi ( Fudou, Iizuka & Yamanaka, 2001 ; Kundim et al., 2003 ). There is no application of Haliangium in plant protection yet, however, the potential of myxobacteria to produce unique secondary metabolites has been recognized ( Reichenbach & Höfle, 1993 ; Wenzel & Müller, 2009 ). For all cultivars but Argentine, an OTU of the abundant bacterial family Nitrosomonadaceae was assigned as an indicator species ( Table S1 ). They are characterized as lithoautotrophic of ammonia-oxidizing bacteria and harbour the well-characterized genera Nitrosomonas and Nitrosospira . In view if this result and the relative abundances of Nitrospira , we suggest that the dynamics of soil-N cycling under different bahiagrass cultivars should be further investigated. Half of the cultivars (Pensacola, Sand Mountain, and Tifton 9) harboured a sequence assigned to a member of the Ceratobasidiaceae as an indicator species ( Table S4 ). Genera of this fungal family include economically relevant phytopathogens like Rhizoctonia , which cause, for example, ‘brown patch’ disease on turfgrasses ( Oniki et al., 1986 ). In rotation systems, bahiagrass has shown to reduce Rhizoctonia population densities in soil and associated diseases on peanuts ( Johnson et al., 1999 ), and vegetables (cucumber ( Cucumis sativus ‘Comet’) and snap bean ( Phaseolus vulgaris ‘Strike’)) ( Sumner et al., 1999 ). The two tested bahiagrass cultivars in the above-mentioned studies on peanuts and vegetables were Pensacola and Tifton 9, respectively. Since Pensacola, Sand Mountain, and Tifton 9 were characterized by an OTU assigned to a member of the Ceratobasidiaceae, our bahiagrass cultivars may differ in their ability to suppress Rhizoctonia population in soils. Therefore, it may be valuable to screen bahiagrass cultivars for disease suppression when used in sod-based crop rotations (i.e., 1 to 8 years of peanuts or vegetables rotated with 2 to 10 years of bahiagrass). An OTU assigned to the widespread family Orbiliaceae was identified as an indicator species of the cultivars Sand Mountain, TifQuik, Tifton 9, and UF-Riata ( Table S4 ). Several members of this family are carnivorous fungi which trap nematodes in soils ( Pfister, 1997 ; Rubner, 1996 ). The underlying mechanisms of biocontrol of nematodes by microorganisms are well described ( Li et al., 2015 ). Rotations of bahiagrass with peanuts, soybean ( Glycine max ), or vegetables have shown the potential to increase nematode control ( Rodriguez-Kabana et al., 1988 ; Rodriguez-Kabana et al., 1989 ; Sumner et al., 1999 ). However, there is a lack of studies comparing the performance of different bahiagrass cultivars on nematode control. Based on our molecular results, we speculate that bahiagrass cultivar screening may improve nematode biocontrol."
} | 6,080 |
39893160 | PMC11787353 | pmc | 5,604 | {
"abstract": "Artificial neural networks have advanced due to scaling dimensions, but conventional computing struggles with inefficiencies due to memory bottlenecks. In-memory computing architectures using memristor devices offer promise but face challenges due to hardware non-idealities. This work proposes layer ensemble averaging—a hardware-oriented fault tolerance scheme for improving inference performance of non-ideal memristive neural networks programmed with pre-trained solutions. Simulations on an image classification task and hardware experiments on a continual learning problem with a custom 20,000-device prototyping platform show significant performance gains, outperforming prior methods at similar redundancy levels and overheads. For the image classification task with 20% stuck-at faults, accuracy improves from 40% to 89.6% (within 5% of baseline), and for the continual learning problem, accuracy improves from 55% to 71% (within 1% of baseline). The proposed scheme is broadly applicable to accelerators based on a variety of different non-volatile device technologies.",
"introduction": "Introduction The increasing demand for large-scale neural network models has prompted extensive research on approaches to optimize model efficiency and accelerate computations. Quantized neural networks, which utilize reduced-precision representations for model parameters and activations, have emerged as a promising avenue for achieving significant computational gains without compromising performance. Recent works advance this approach to the extreme demonstrating effective 1-bit 1 and 1.58-bit 2 (ternary) quantization of the parameter space with minimal performance loss. In addition to precision reduction, memory-based hardware accelerators are emerging as another frontier to enhance neural network efficiency. In particular, in-memory computation on a physical chip of memory devices called memristors offers a synergistic solution that can complement efficiency gains achieved by network quantization 3 . Memristors are two-terminal non-volatile memory devices that exhibit unique programmable resistive switching behavior. Their intrinsic characteristics enable co-location of computation and memory, mimicking aspects of synaptic functionality in biological systems 4 , 5 . By arranging memristors over a two-dimensional array such that the devices are placed at intersection points, an architecture commonly referred to as a crossbar 6 – 8 , underlying device physics can be exploited to implement parallelized vector-matrix multiplication—a critical operation in artificial neural networks—in the analog domain. Since memristor crossbars are programmable and non-volatile 9 , 10 , they can be utilized to build dedicated hardware accelerators for deep neural networks. Diverse technologies including resistive random-access memory (ReRAM) and phase change memory are being considered as promising crossbar candidates to implement the multiply and accumulate operations representing the standard synaptic weights model used in most neural networks 11 – 19 . Hardware accelerators based on these technologies can overcome von Neumann architecture limitations as they minimize data movement and energy consumption, both of which are key bottlenecks for large-scale neural network workloads. For these reasons, memristive neural network accelerators have the potential to transform capabilities of machine learning systems and thereby usher in a neuromorphic era of artificial intelligence computing at the edge. A comprehensive exploration of the interplay between quantized neural networks, dedicated hardware accelerators, and memristive technologies becomes imperative for advancing this capability with the goal to unlock unprecedented efficiency gains in real-world deep learning applications. The study of memristive neural network accelerators faces several challenges. A purely experimental approach is unfeasible since commercial tape-outs have long timelines and significant design and fabrication costs 4 , 20 , 21 . Before hardware prototyping can be practically motivated, results from hardware-aware 22 – 24 simulations—simulations that take hardware characteristics into account—are needed to reliably predict the performance of these systems. The critical issue is that these devices can exhibit complex non-idealities such as cycle-to-cycle variability (inconsistency in performance from one switching cycle to the next), device-to-device variability (variation in behavior between individual devices), and even tuning failure (where the resistance state remains stuck at a certain level). Operating arrays of these devices introduces system-level non-idealities as well, such as noise and precision/resolution limitations arising from analog-to-digital converters (ADCs), digital-to-analog converters (DACs), and trans-impedance amplifiers. These non-idealities manifest as deviations in the outputs of the underlying vector-matrix multiplications, rendering memristive neural networks incapable of achieving software-equivalent accuracies without some mechanism for fault tolerance in place. There have been numerous advancements in fault tolerance schemes for bridging this performance gap in inference accuracy of non-ideal memristive networks compared to their software counterparts 6 , 25 – 30 , with more recent works focusing on challenging continual learning settings 31 , 32 . Existing literature on the subject can be classified into two broad categories. The first focuses on circuit-level and device-level optimizations 6 , 25 such as alterations to the crossbar configuration and circuitry or the device material stack, and the second on network-level algorithmic optimizations 26 – 30 , 33 such as advanced programming or weight-to-device mapping and encoding schemes. Algorithmic investigations can be further divided into two sub-categories. The first utilizes information about hardware defects and attempts to train defect or hardware-aware solutions for the memristive hardware, while the second attempts to transfer a pure-software solution on the memristive hardware intelligently. Both approaches share the goal of maximizing the performance of the memristive network. The first achieves this by training redundancy into the solution, while the second achieves this by averaging out induced currents from multiple mapped instances of one or more solutions on the crossbar. We offer a detailed comparison of prominent works from literature that propose fault tolerance solutions for memristive neural networks in Supplementary Table 1 . Most works are simulation-oriented and are not validated with system-level neural network results from a hardware prototype. They make assumptions on non-idealities that do not fully represent a practical memristor crossbar system. For example, a common assumption about stuck-at faults is that stuck-low and stuck-high devices take on conductances \\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}$${G}_{{OFF}}$$\\end{document} G O F F 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}$${G}_{{ON}}$$\\end{document} G O N (commonly formulated in terms of resistances as high resistance state and low resistance state respectively), while all operable devices can be tuned within this range to a relatively high precision (typically \\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}$$\\ge 6$$\\end{document} ≥ 6 bits). Additionally, some studies overlook the fact that defect maps can evolve over time due to repeated switching cycles, or due to the application of damaging voltages which can vary from one device to another. With the prototyping system and memristive devices used in this work (presented later in the Methods: Experimental Setup section), we found that devices with stuck-at faults have conductance values that are significantly beyond the conductance range of non-stuck, operable devices. Although this is a common occurrence in array characterization studies 34 , 35 , few existing works on fault tolerance schemes account for this, as indicated by the stuck-at state formulation column in Supplementary Table 1 . We also found that tuning an operable device to the conductance of a stuck device tends to permanently damage the device due to irreversible dielectric breakdown, limiting device endurance 9 , 36 . This leads to an outdated defect map and, if unaddressed, a considerable reduction in the overall usability of the chip for any computational workload. This means that one must exercise caution when tuning operable devices, ensuring that the limited dynamic range 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}$$[{G}_{{OFF}},\\,{G}_{{ON}}]$$\\end{document} [ G O F F , G O N ] is strictly respected across the array. Under these limitations, traditional fault tolerance schemes based on redundant rows or columns of memristors lose their effectiveness as a single stuck device in the positive conductance matrix can no longer be compensated by a single operable device in the negative conductance matrix (or vice versa). This problem is further exacerbated by system-imposed limitations on device tunability. For example, even though we can tune individual devices to multi-bit precision (presented later in the Results: Device and Array Characterization section), operating contiguous blocks in parallel and realizing parallel vector-matrix multiplication for inference imposes new constraints on conductance states due to limited precision and noise from system components such as ADCs, DACs, transimpedance amplifiers, and other transient sources. To compensate, we artificially limit target conductance states for our devices to a lower precision of 1-bit. From a device perspective, this regime is relevant beyond memristors, as certain device technologies—such as magnetic tunnel junctions 37 —only support two conductance states. Just as a unified prototyping system is essential for the development of emerging memory devices 38 , a device-agnostic fault tolerance algorithm that can fairly compare the neural network performance of different technologies is equally critical. These practical considerations form the foundation of our work on layer ensemble averaging. We propose layer ensemble averaging (summarized in Fig. 1 , see Methods: Layer Ensemble Averaging for more details), a hardware-correction fault tolerance scheme designed for improving the performance of memristive neural networks irrespective of the underlying device technology. It leverages the concept of redundant devices and eliminates the need for high device and system tunability, making it a versatile, assumption-free, and universally applicable solution for fault tolerance in these networks. It also does not require re-training and is thus particularly suitable in the context of a fixed prototyping system designed for evaluating device technologies or parameters against one another. It comprises of block-by-block weight mapping and encoding algorithms that operate together to directly correct vector-matrix multiplication outputs in hardware at the level of individual neural network layers. Contrary to existing related literature 26 , 30 where neural network outputs are obtained by polling outputs of an ensemble of neural networks, not necessarily mapping the same solution, here ensemble outputs are polled at the level of each layer by mapping the same solution multiple times. This key difference makes our approach suitable for cases where the availability of software solutions is limited (perhaps due to high training costs or long training times) and one wishes to deploy a pre-trained model directly to memristive hardware for acceleration with minimal performance loss. It also leads to a straightforward way to investigate hardware outputs relative to software outputs of the vector-matrix multiplication operation for any given layer. Consequently, the proposed approach has the added benefit of being useful for domains beyond deep learning that require accurate multiply-and-accumulate or vector-matrix multiplication operations such as digital signal processing, image and video processing, scientific computing, and financial modeling. Fig. 1 Layer ensemble averaging. Demonstration of the proposed layer ensemble averaging technique for mapping an example weight matrix \\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}$${{\\bf{W}}}_{{\\bf{ideal}}}$$\\end{document} W ideal to a crossbar of non-ideal memristive devices. The first step is to determine appropriate locations on the crossbar for mapping (see Supplementary Algorithm 1 ), followed by an encoding step (see Table 1 and Supplementary Algorithm 2 ) that converts the weight matrix to device conductance matrices \\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}$${{\\bf{G}}}_{{\\bf{pos}}}$$\\end{document} G pos 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}$${{\\bf{G}}}_{{\\bf{neg}}}$$\\end{document} G neg . Each encoded conductance matrix is written to contiguous blocks of devices on the non-ideal chip according to respective mappings that contain additional information about row-wise defects based on the summed conductance variation metric from Eq. ( 1 ). This is utilized during inference to suppress currents from highly defective rows from participating in the averaging process (see Eq. ( 3 )). Green and orange coloring represents devices in low and high conductance states respectively, with varying shades indicating device-to-device variability, while black and white coloring represents devices that are stuck/faulty or inactive respectively. For simplicity, only the layer ensemble mapping for \\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}$${{\\bf{G}}}_{{\\bf{pos}}}$$\\end{document} G pos is shown. Redundancy parameters shown: \\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}$$\\alpha=3$$\\end{document} α = 3 , indicating that the weight matrix (represented by conductance matrices \\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}$${{\\bf{G}}}_{{\\bf{pos}}}$$\\end{document} G pos ) is mapped on to the crossbar three times 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}$${{\\bf{G}}}_{{\\bf{pos}}}^{\\left({\\bf{i}}\\right)}$$\\end{document} G pos i representing 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}$$i$$\\end{document} i -th mapping 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}$${{\\bf{G}}}_{{\\bf{pos}}}$$\\end{document} G pos (similarly for \\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}$${{\\bf{G}}}_{{\\bf{neg}}}$$\\end{document} G neg ), 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}$$\\beta=2$$\\end{document} β = 2 , indicating that from this layer ensemble of size 3, currents from 2 active rows (based on the SCV metric from Eq. ( 1 )) will be averaged for each output. In this work, we demonstrate that pre-trained and quantized neural network solutions written to a non-ideal memristive crossbar can attain near-software inference performance using the proposed layer ensemble averaging fault tolerance scheme. We present simulations informed by experiments where we compare the performance of layer ensemble averaging with hardware correction methods from literature (bold entries in Supplementary Table 1 ) for classifying the MNIST handwritten digit dataset 39 . Realistic device and system non-idealities are modeled in these simulations. We offer comparisons with two existing state-of-the-art hardware-oriented fault tolerance schemes Mapping Algorithm with inner fault tolerant ability (MAO) 27 , 33 and the Committee Machines (CM) algorithm 26 . We do not offer comparisons with software-oriented schemes such as those based on error correcting codes and architectures 40 – 43 , because they are orthogonal solutions that can be paired with any device redundancy-based fault tolerance scheme. Moreover, we do not offer comparisons with schemes involving re-training, because one of our goals is to have a fault tolerance scheme that is independent of the underlying technology, and solutions involving re-training inherently utilize device-specific properties such as tunability. Additionally, we provide an end-to-end system demonstration where we validate the overall effectiveness of layer ensemble averaging as a fault tolerance approach on a continual learning problem based on the Yin-Yang dataset 44 using an array of ReRAM devices with our custom-built mixed-signal hardware prototyping platform called Daffodil 38 (summarized in Fig. 2 , see Methods: Experimental Setup for more details). This platform consists of an integrated chip with 20,000 memristive devices, a mixed-signal printed circuit board (PCB), a Zynq-based (certain commercial processes and software are identified in this article to foster understanding. Such identification does not imply recommendation or endorsement by the National Institute of Standards and Technology, nor does it imply that the processes and software identified are necessarily the best available for the purpose) field programmable gate array (FPGA) development board, and an accompanying software suite guided by principles of hardware-software co-design for hardware experiments as well as accurate system simulations. Fig. 2 Overview of the experimental setup. a Image of the complete working setup. The FPGA development board interfaces with the Daffodil printed circuit board housing the 20,000 device ReRAM chip (outlined in white) via a FPGA mezzanine card (FMC) connector. The development board also connects to a host system over a serial terminal through which high-level applications can be run. b A zoomed-in view of the outlined mixed-signal Daffodil board showing major hardware components present on-board. ADCs, DACs, and the FMC connector are located under the board.",
"discussion": "Discussion In summary, we proposed and demonstrated layer ensemble averaging (LEA)—a hardware-correction fault tolerance scheme that can improve the inference performance of memristive neural networks up to some software baseline. The scheme does not require any re-training after deployment to crossbars of emerging memory devices and does not make any assumptions on multi-bit device tunability, making it a versatile, generally applicable, device-agnostic fault tolerance scheme. We demonstrated that layer ensemble averaging outperforms both the Mapping Algorithm with Inner-Fault tolerance (MAO) and Committee Machines (CM)—two existing state-of-the-art hardware-based fault tolerance methods—by simulating a 2-layer perceptron network for classifying the MNIST handwritten digit dataset (see Fig. 4 ). Additionally, we experimentally validated the fault tolerance capability of layer ensemble averaging by mapping a 3-layer perceptron network to tackle a multi-task classification problem (see Fig. 5 ) on our end-to-end mixed-signal prototyping platform featuring an in-house fabricated chip with 20,000 ReRAM devices. MAO, CM, and LEA are all fault tolerance schemes that rely on device-level redundancy. The first algorithm, MAO, operates by mapping network parameters to devices such that each parameter is modeled by the summed conductance of multiple devices. For some software weight matrix \\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}$${{\\bf{W}}}_{{\\bf{ideal}}}$$\\end{document} W ideal , this leads to the expression \\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}$${{\\bf{W}}}_{{\\bf{ideal}}}\\propto {\\sum }_{i=1}^{\\alpha }{{\\bf{G}}}_{{\\bf{pos}}}^{\\left({\\bf{i}}\\right)}-{\\sum }_{i=1}^{\\alpha }{{\\bf{G}}}_{{\\bf{neg}}}^{\\left({\\bf{i}}\\right)}$$\\end{document} W ideal ∝ ∑ i = 1 α G pos i − ∑ i = 1 α G neg i , where \\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}$$\\alpha$$\\end{document} α expresses the number of redundant crossbars 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}$${{\\bf{G}}}_{{\\bf{pos}}}^{({\\bf{i}})}$$\\end{document} G pos ( i ) expresses 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}$$i$$\\end{document} i -th \\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}$${{\\bf{G}}}_{{\\bf{pos}}}$$\\end{document} G pos conductance matrix (similarly for \\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}$${{\\bf{G}}}_{{\\bf{neg}}}$$\\end{document} G neg ). The impact of devices with stuck-at faults can be compensated by tuning operable devices participating in the mapping of a given parameter directly. This scheme works well if stuck-at fault conductance ranges are within the conductance ranges of operable devices, but we find that it becomes ineffective when stuck-at fault ranges are beyond the range of operable devices (which themselves have low tunability, as summarized in Supplementary Table 1 ). The second algorithm, CM, operates by mapping \\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}$$\\alpha$$\\end{document} α unique solutions to the crossbar, each of which is mapped 1 time, in contrast to MAO where a single solution is mapped \\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}$$\\alpha$$\\end{document} α times. When the final prediction must be made, output currents are averaged from all networks in the committee. Mathematically, this leads to the representation \\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}$${{\\mathbf{W}}}_{{\\mathbf{ideal}},{\\mathbf{final}}}\\propto \\frac{1}{{\\rm{\\alpha }}}\\left({\\sum }_{i=1}^{\\alpha }{{\\mathbf{G}}}_{{\\mathbf{pos}}}^{\\left({\\mathbf{i}}\\right)}-{\\sum }_{i=1}^{\\alpha }{{\\mathbf{G}}}_{{\\mathbf{neg}}}^{\\left({\\mathbf{i}}\\right)}\\right)$$\\end{document} W ideal , final ∝ 1 α ∑ i = 1 α G pos i − ∑ i = 1 α G neg i for the final layer, 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}$${{\\mathbf{W}}}_{{\\mathbf{ideal}}}^{\\left({\\mathbf{i}}\\right)}\\propto {{\\mathbf{G}}}_{{\\mathbf{pos}}}^{\\left({\\mathbf{i}}\\right)}-{{\\mathbf{G}}}_{{\\mathbf{neg}}}^{\\left({\\mathbf{i}}\\right)}$$\\end{document} W ideal i ∝ G pos i − G neg i for 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}$$i$$\\end{document} i -th mapping of preceding layers. This scheme works well when the trained solutions and devices have a high number of states, but we find that the performance degrades when the tunability is limited and stuck-at faults are high. Finally, our algorithm LEA operates by mapping the same solution \\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}$$\\alpha$$\\end{document} α times to memristive crossbars, similar to MAO. Unlike CM, the averaging happens at each layer instead of just the final layer. The mathematical representation is thus \\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}$${{\\mathbf{W}}}_{{\\mathbf{ideal}}}\\propto \\frac{1}{{\\rm{\\beta }}}\\left({\\sum }_{i=1}^{\\beta }{{\\mathbf{G}}}_{{\\mathbf{pos}}}^{\\left({\\mathbf{i}}\\right)}-{\\sum }_{i=1}^{\\beta }{{\\mathbf{G}}}_{{\\mathbf{neg}}}^{\\left({\\mathbf{i}}\\right)}\\right)$$\\end{document} W ideal ∝ 1 β ∑ i = 1 β G pos i − ∑ i = 1 β G neg i for each layer, where \\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}$$\\beta \\le \\alpha$$\\end{document} β ≤ α (see Methods: Layer Ensemble Averaging for a detailed discussion). Although this averaging incurs additional overhead compared to MAO and CM, we showed that our proposed algorithm LEA leads to a consistently higher degree of fault tolerance and is thus a suitable scheme for memristive neural network implementations where practical non-idealities are present. The overheads in LEA are still comparable to MAO and CM. At similar levels of redundancy, all schemes require the same number of devices. For a mixed-signal system such as our Daffodil prototype that we utilize for experimental validation in this work, there is no additional hardware overhead. This is because ADCs and DACs are reconfigurable, and inference operations such as current summation and averaging can be done in software on the hard processor before being sent to subsequent analog layers. In such an implementation, the only overhead is in terms of time complexity. For an implementation where all operations are implemented in hardware, layer ensemble averaging would have a constant time overhead because of the scaling operation for averaging that must happen at each layer after current summation, compared to no averaging in MAO and averaging only at the final layer in CM. The scaling unit in hardware can be time-shared across various layers, meaning that the hardware overhead of layer ensemble averaging would still be the same as CM. We argue that this would be an acceptable overhead given the performance gains that are possible with layer ensemble averaging, as reported in our simulation results in Fig. 4 as well as experimental results in Fig. 5 . Furthermore, the redundancy 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}$$\\alpha$$\\end{document} α 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}$$\\beta$$\\end{document} β can be used to implement a trade-off between the performance of the memristive neural network against the total number of devices that participate in mapping each weight matrix. Provided enough devices are available for the mapping to succeed, this trade-off can bring the performance of the memristive network on-par with its software counterpart. Although the exact level of redundancy required to attain near-software inference performance depends on the network architecture and task under investigation and layer-wise sensitivities to faults, layer ensemble averaging can be employed as a general hardware-correction technique with even a limited redundancy level to correct vector-matrix multiplication outputs and improve network performance. Additionally, it can be paired with software-correction techniques for additional fault tolerance if needed, since software-correction would be an orthogonal solution. By design, layer ensemble averaging assumes that hardware-redundancy is possible. However, pairing layer ensemble averaging with a software-correction scheme can assist in the case where the number of devices required to map parameters for a certain layer redundantly exceed the physical number of available devices. In an extreme situation where no block-wise redundancy can be implemented due to hardware limitations, one could relax the mapping phase to allow for row-wise duplication, as discussed in the Methods: Layer Ensemble Averaging section. In conjunction with a software-correction scheme, another workaround for added robustness can be to average currents from the same mapping multiple times per layer. This can be accomplished with minimal changes to the system controller. Our results provide an experimental proof-of-concept demonstration of the possibility of performing accurate vector-matrix multiplications in non-ideal memristive crossbars. We showed that layer ensemble averaging is a versatile training-free fault tolerance scheme that can improve the performance of memristive neural networks irrespective of the underlying device technology. We presented several variants suitable for different situations. If optimizing network performance is the priority, greedy mapping is the better candidate. If device defects are known to be uniformly distributed or if fast algorithmic run time is a requirement, then random mapping is a better candidate as a single sampling iteration would produce mappings that have a similar statistical distribution of defects compared to the exhaustive greedy search. Similarly, simple encoding is generally the better candidate since it has a faster algorithmic run time compared to the reduced mapping error encoding, but the latter can be useful algorithm if the application would benefit from slightly better mapping errors. A multi-objective optimization framework for producing layer ensemble averaging mappings at redundancy levels proportional to layer-wise sensitivities could be devised in the future. While this work focused exclusively on utilizing layer ensemble averaging for fully connected neural network inference with ternary weights, the methodology is broadly applicable to a variety of architectures (such as convolutional layers, recurrent layers, transformer layers, etc.) and domains requiring accurate multiply-and-accumulate and vector-matrix multiplication operations. Dedicated memristive hardware accelerators for such cases could be envisioned in the future owing to their power consumption and energy benefits compared to traditional computing systems, and the proposed layer ensemble averaging approach could help mitigate their defects accordingly."
} | 8,837 |
31406214 | PMC6690927 | pmc | 5,605 | {
"abstract": "Benthic foraminifera are known to play an important role in marine carbon and nitrogen cycles. Here, we report an enrichment of sulphur cycle -associated bacteria inside intertidal benthic foraminifera ( Ammonia sp. (T6), Haynesina sp. (S16) and Elphidium sp. (S5)), using a metabarcoding approach targeting the 16S rRNA and aprA -genes. The most abundant intracellular bacterial groups included the genus Sulfurovum and the order Desulfobacterales. The bacterial 16S OTUs are likely to originate from the sediment bacterial communities, as the taxa found inside the foraminifera were also present in the sediment. The fact that 16S rRNA and aprA –gene derived intracellular bacterial OTUs were species-specific and significantly different from the ambient sediment community implies that bacterivory is an unlikely scenario, as benthic foraminifera are known to digest bacteria only randomly. Furthermore, these foraminiferal species are known to prefer other food sources than bacteria. The detection of sulphur-cycle related bacterial genes in this study suggests a putative role for these bacteria in the metabolism of the foraminiferal host. Future investigation into environmental conditions under which transcription of S-cycle genes are activated would enable assessment of their role and the potential foraminiferal/endobiont contribution to the sulphur-cycle.",
"conclusion": "Conclusion To date, sulphur-cycle related putative endobionts in benthic foraminifera have been largely overlooked and understudied, however, as our data shows, the genetic potential for both sulphur oxidation and sulphate reduction is abundant in the studied foraminiferal species from two different locations. Furthermore, these SOB and SRB are phylogenetically closely related to known symbiotic bacteria of other marine eukaryotes. We therefore hypothesize, that these putative endobionts, which foraminifera may derive from the ambient sediment, could be linked to foraminiferal carbon / nutrient acquisition, allowing the foraminifera to inhabit the periodically anoxic and sulphidic intertidal sediments. Future studies targeting the activity of the putative endobionts are needed to confirm their functions and roles in foraminiferal ecology.",
"introduction": "Introduction Benthic foraminifera are unicellular eukaryotes widespread across marine environments. Due to their high abundance and predominance in the benthic ecosystem, they play an important role in the sedimentary carbon cycle by participating in phytodetritus processing and organic matter uptake 1 , 2 . Living at and also deeper within the sediment implies that these foraminifera sometimes live under oxygen depleted conditions and potentially rely on alternative biogeochemical pathways 3 . Benthic foraminifera are known to actively take part in the nitrogen cycle, as several species have the ability to take up and store nitrate intracellularly 4 – 7 and to perform complete denitrification 4 , 8 . A recent genome analysis of two Globobulimina species suggests the existence of a novel denitrification pathway encoded by foraminifera’s own genome 8 . In certain areas, benthic foraminifera may even be responsible for the majority of the benthic denitrification process, which highlights their global importance in the nitrogen cycle 9 , 10 . In addition to carbon and nitrogen cycling, recent evidence of foraminiferal sulphur uptake in labelling experiments 11 suggests that foraminifera may also play a role in the sedimentary sulphur cycle. Benthic foraminifera are known to harbour a range of potential bacterial endobionts, including putative denitrifying bacteria and sulphur-oxidizing bacteria 12 – 14 . Recently, methanotrophs were also found to be associated with benthic foraminifera 15 . As such, the function of the putative endobiont community may be diverse, ranging from metabolic strategies to the ability to inhabit otherwise hostile environments, such as dysoxic, sulphidic sediments 13 , 16 . Endosymbiotic relationships are also common in other marine eukaryotes, offering them potential evolutionary benefits, as they help the host to adapt to unstable conditions and survive in unfavourable environments 17 . For example, ciliates are known to harbour a variety of endobionts linked to carbon, nitrogen and sulphur cycles, which are crucial for the survival of the host species 16 , 18 . In ciliates, the endosymbiotic relationships are known to have developed independently and species-specifically, and they persisted on long geological time scales 19 . The origin of the benthic foraminiferal endobionts is currently not well understood. It has been suggested that they may be transferred from generation to generation 14 . Alternatively, they may be drawn from the ambient sediment, similar to planktonic foraminifera, which are suggested to have evolved their endosymbioses via interactions with water column bacteria 20 . However, so far very little is known of the interactions between foraminifera and the surrounding sediment bacterial community. Sedimentary bacterial communities may play a role in foraminiferal diet. However, bacterial uptake by benthic foraminifera has been shown to be unselective, implying that bacterivory occurs mainly in association with potential deposit-feeding behaviour 21 . Feeding strategies, as well as organic matter turnover rates, appear to be species-specific 21 – 23 . For example, in situ experiments with 15 N labelled bacteria have shown that benthic foraminifer Ammonia tepida prefers algae in its diet over bacteria 24 , suggesting that bacteria are not its primary food source. In addition to bacterivory, foraminifera are known to have a variety of other feeding strategies, such as herbivory, carnivory and even direct dissolved organic carbon uptake 21 , 23 , 25 , 26 . Mixotrophy is also an important trophic strategy for some shallow-water benthic foraminifera with the ability to harbour photosymbionts or kleptoplasts 27 – 29 . The photosymbiont associations can be diverse and flexible, and it has even been suggested that some species are able to shuffle their photosymbionts to increase adaptability 30 . Kleptoplast-driven photosynthesis and associated inorganic carbon fixation is shown to be an important carbon sequestration mechanism for Haynesina germanica 31 . Additionally, kleptoplasts may serve as an energy reservoir under unfavourable conditions 32 . Despite the significant contribution of these ubiquitous and abundant organisms to both C and N cycling e.g. 1 , 2 , 4 , 8 – 10 very little is known of the potential interactions between the sediment bacteria and benthic foraminifera. Previously, endobiont studies have been mainly based on transmission electron microscope (TEM) observations 33 and lacked direct comparisons to the sediment microbial community. Recently, 16S rDNA metabarcoding has provided insights into intracellular bacterial communities of planktonic foraminifera, allowing the identification of putative species-specific endobionts 20 , 34 . Here, we use a metabarcoding approach to target the 16S rRNA gene and focus on 3 benthic species, Ammonia sp. (T6), Elphidium sp. (S5), and Haynesina sp. (S16) 23 collected from intertidal localities in the Dutch Wadden Sea. We compare the 16S rDNA metabarcoding –derived intracellular bacterial operational taxonomic units (OTUs) to those of the ambient sediment to determine which bacterial groups are enriched within foraminifera, and link the findings to sediment porewater chemistry and sediment bacteria distribution. We examine potential species-specific intracellular bacterial 16S OTUs in foraminifera, as well as, the effect of sediment depth and sampling location. Furthermore, the sulphur cycle-related aprA (dissimilatory APS reductase) functional gene is quantified and sequenced, to explore the potential for intracellular bacteria-driven sulphur oxidation/sulphate reduction in benthic foraminifera, and to study the phylogenetic relationships of the associated bacteria.",
"discussion": "Discussion The 16S rDNA metabarcoding revealed a wide diversity of both SRB and SOB enriched among the intracellular bacterial 16S OTUs of foraminifera compared to the surrounding sediment bacterial community. Although, some intraspecific variation was observed (Fig. 2 ), statistically intracellular bacterial 16S OTUs as well as aprA OTUs of foraminifera were species-specific. Overall, alpha diversity was lower in foraminifera compared to sediments, which could potentially imply a selective uptake, although differences in the amount of sample material (0.25 g sediment vs . a single foraminiferal cell) may also be the driver of the lower diversity obtained. The genetic potential for sulphur oxidation and sulphate reduction was further identified by targeting and quantifying the aprA gene, which was found to be abundant across different foraminiferal species. In contrast, intracellular N-cycle associated bacteria were not successfully targeted in foraminifera, implying that they play a trivial role in our specimens compared to S-cycle associated bacteria. We were able to amplify a rather long DNA fragment (approximately 550 bp) of intracellular bacterial DNA from foraminifera, implying that the bacterial DNA was not degraded by digestion, which would typically limit the length of the fragments obtained 40 . As such, the intact nature of the extracted DNA implies that the bacteria inside the foraminifera may be alive and putatively endobiotic. Intact and dividing bacteria have also been previously observed inside intertidal benthic foraminifera Ammonia sp. (T6), suggesting that putative bacterial endobionts could exist at least in this species 41 . To verify the activity and function of the putative endobionts, RNA and/or FISH analysis are recommended. The presence of bacterial OTUs inside the foraminifera solely due to bacterivory cannot be completely excluded but seems very unlikely. Although similar bacterial taxa were present in the foraminifera and the surrounding sediment, these occurred in contrasting relative abundancies. Foraminifera digest bacteria randomly 21 while deposit feeding, which would likely result in intracellular bacterial composition that would more closely mirror that of the ambient sediment. Random deposit feeding would also not be expected to result in species-specific bacterial 16S OTUs observed here. In addition, previous work has shown that bacterivory plays only a minor role in fulfilling the carbon requirements of benthic foraminifera 21 , 24 , 42 . Instead, the elphidiid specimens ( Elphidium sp. S5, Haynesina sp. S16), are kleptoplastic 29 and thus are likely to have a dietary preference for diatoms 23 , whereas Ammonia sp. T6 has been suggested to exhibit also carnivorous behaviour 23 , 43 . The relatively low yield of chloroplasts in elphidiids in this study, compared to the study of Chronopoulou et al . (2019) where universal 18S primers were used to target eukaryotes on the same specimens, may be related to the limited ability of 16S primers to target these organelles. Thus, to better target intracellular algal signal we recommend the use of universal eukaryotic primers. As an alternative to being solely a food source, the sediment bacterial community could also provide endobionts to benthic foraminifera in a similar way that the endobionts of pelagic foraminifera have been linked to the surrounding water column 20 . Previous studies have suggested that SOB could potentially have an endobiotic relationship with foraminifera 14 . In addition, sulphur ( 34 S) incorporation under dysoxia was observed for Ammonia sp. (T6), implying an ability to potentially synthesize sulpholipids through a sulphate activation pathway 11 . In other marine eukaryotes, endo- and ectobionts associated with the sulphur cycle are widespread in marine environments and occur in several phyla 17 . The host can benefit from them in various ways. For example, SOB symbionts can fix carbon autotrophically while deriving energy from sulphur-oxidation, and provide the host with organic carbon sources 17 . Sulphur-oxidizing symbioses have been discovered in marine sponges 39 , nematodes 44 , ciliates 18 , 45 , 46 and oligochaete worms 36 , 47 . In turn, SRB symbionts can produce sulphide by oxidizing either organic compounds e.g. acetate or inorganic compounds e.g. hydrogen. Some eukaryotes, such as the oligochaete worm Olavius algarvensis , can even harbour both SOB and SRB, forming an endosymbiotic sulphur cycle, potentially helping the host cope with sediments with variable sulphide concentrations 36 , 47 . In our data, several aprA OTUs in the SRB I branch cluster together with the O. algarvensis Delta 1 symbiont. Overall, the sulphur cycle –related OTUs were almost equally distributed between SRB and SOB branches, in which they grouped with known symbiotic bacteria of other marine eukaryotes, suggesting that some of the aprA OTUs could be putative endosymbionts for the foraminifera. These closely related symbionts included, for example, ectosymbionts of the oligochaete worm Tubificoides benedii isolated from the Wadden Sea coastal sediments 38 . In the SRB I branch, the foraminiferal intracellular bacterial OTUs contained a large abundance of sulphate-reducing Deltaproteobacteria, belonging to bacterial families Desulfobulbaceae and Desulfobacteraceae . In ciliates, these same bacterial taxa have been identified as endobionts growing autotrophically and potentially providing the host with amino acids 18 . Similarly, kleptoplast-bearing foraminifera are able to receive photosynthates, such as amino acids, directly from their symbionts 48 . In addition, utilisation of dissolved amino acids has been observed in benthic foraminifera 25 , although the exact mechanism for this is poorly understood. We suggest that sulphur-cycle related endobionts could potentially benefit foraminifera by providing carbon or other vital compounds, such amino acids, to the host. Dynamic environments with changing redox conditions, such as intertidal mudflats, have been estimated to be a potential hotspot for sulphur-cycle related symbiotic associations 38 . They are in general characterized by a very shallow oxygen penetration depth 49 and variable redox stress due to non-steady state porewater geochemistry associated with tidal activity and bioturbation. In this study, at de Cocksdorp, the sediment became oxygen depleted after 0.2 mm and free H 2 S was detected below 1 cm sediment depth (Fig. 1 ). Despite the challenging conditions, foraminifera can thrive in these environments due to their ability to survive long periods of anoxia 5 , 50 , 51 and tolerate sulphidic conditions 52 . In the Wadden Sea, all three species are commonly encountered and contribute significantly towards benthic biomass 1 , 22 , 42 . Under anoxia, the metabolic rate of foraminifera decreases 53 and their cytoplasm gets thinner 41 . However, despite reduced metabolism, foraminifera must still sustain their vital functions and have been shown to continue to grow and calcify 54 . In other eukaryotes, such as ciliates, endobionts are hypothesized to be crucial for survival in anoxic/dysoxic environments 18 , 45 . In a similar way, endobionts could help benthic foraminifera to adapt to changing environmental conditions. The relatively diverse composition of intracellular bacterial 16S OTUs may also provide the foraminifera an advantage in responding to environmental stress, as they could potentially utilize the most appropriate endobiont community, in a similar fashion to photosymbiont-bearing foraminifera that have been suggested to potentially shuffle their internal symbiont pool in response to changes in environmental conditions 30 , 55 . In addition to sulphur-cycle bacteria, the intracellular bacterial 16S OTUs observed in this study included chloroplasts, of which the most abundant one was similar to kleptoplastic endobionts previously isolated from Virgulinella fragilis 14 . It has been suggested that harbouring chloroplasts along with SOB symbionts may have the advantage of reducing the harmful effects of H 2 S 12 , 14 , which could benefit kleptoplastic species such as Elphidium sp. and Haynesina sp. The genetic and metabolic diversity of putative sulphur cycle-associated endobionts might help foraminifera to colonize unstable, dynamic environments, where oxygen is limited but sulphate and sulphide is abundant."
} | 4,120 |
35651318 | PMC9795978 | pmc | 5,606 | {
"abstract": "Abstract The extensive use of petrochemicals has produced serious environmental pollution problems; fortunately, bioremediation is considered an efficient way to fight against pollution. In line with Synthetic Biology is that robust microbial chassis with an expanded ability to remove environmental pollutants are desirable. Pseudomonas putida KT2440 is a robust lab microbe that has preserved the ability to survive in the environment and is the natural host for the self‐transmissible TOL plasmid, which allows metabolism of toluene and xylenes to central metabolism. We show that the P . putida KT2440 (pWW0) acquired the ability to use octane as the sole C‐source after acquisition of an almost 62‐kb ICE from a microbial community that harbours an incomplete set of octane metabolism genes. The ICE bears genes for an alkane monooxygenase, a PQQ‐dependent alcohol dehydrogenase and aldehyde dehydrogenase but lacks the electron donor enzymes required for the monooxygenase to operate. Host rubredoxin and rubredoxin reductase allow metabolism of octane to octanol. Proteomic assays and mutants unable to grow on octane or octanoic acid revealed that metabolism of octane is mediated by redundant host and ICE enzymes. Octane is oxidized to octanol, octanal and octanoic acid, the latter is subsequently acylated and oxidized to yield acetyl‐CoA that is assimilated via the glyoxylate shunt; in fact, a knockout mutant in the aceA gene, encoding isocitrate lyase was unable to grow on octane or octanoic acid.",
"conclusion": "CONCLUSIONS \n Pseudomonas putida (pWW0) acquired an ICE with oct genes from a microbial community, that allowed oxidation of octane to octanoic acid, the clone was named EM2‐4 and it was characterized. The acquired oct cluster was atypical in the sense that it lacked the required electron transfer proteins, which were ‘parasitized’ from the host. Once octanoic acid is made, it is acylated and degraded via β‐oxidation to produce high levels of acetyl‐CoA, which is metabolized through the glyoxylate shunt (Figure 3 ). In agreement with this proposal, the set of enzymes involved in β‐oxidation and the glyoxylate shunt was induced, and a mutant in aceA failed to grow on octane or octanoic acid. Our proteomic analysis revealed a redundancy of acyl‐CoA and β‐oxidation enzymes induced in octane growing cells, which agrees with the fact that no mutants unable to grow on octane in this set of enzymes were found. We have previously generated a collection of mutants in KT2440 and found mutants in several of these enzymes. All of them grew on butanol, supporting the referenced redundancy of acylating enzymes related above. Accordingly, with operation of the glyoxylate shunt, isocitrate lyase—a key enzyme of the pathway—was one of the most abundant proteins when octane was used as the sole C. In a previous study, other mutations in the glyoxylate shunt were found, i.e. in malate synthase (GcbB), that limited the catabolism of butanol. This mutant failed to use octanoic acid (not shown). Therefore, the set of induced proteins and mutants support that medium‐chain alkanes are metabolized and β‐oxidized to acetyl‐CoA, which is channelled to central metabolism via the glyoxylate shunt, while the Krebs cycle playing a minor role. Meanwhile, respiratory chains adapt to optimize electron flow and growth using octane under aerobic conditions.",
"introduction": "INTRODUCTION Over the last century, the chemical industry has been based on the exploitation of petrochemicals as fuels for combustion and as raw materials for the synthesis of a wide range of compounds. Many of these chemicals reach the environment, and while some are degraded others remain in soils or water either because of their recalcitrance or their rate of deposit is still far superior to that of removal (Danso et al., 2019 ; Mori & Kanaly, 2021 ; Nagata et al., 2019 ). Pollution related to petrochemical use has led to global warming and the search is now on to replace polluting chemicals with new green alternatives (Calero & Nikel, 2019 ; García‐Franco et al., 2021 ; Godoy et al., 2021 ; Ramos & Duque, 2019 ); although the new (bio)chemistry for production of chemicals is still developing (Brooks & Alper, 2021 ; Liu et al., 2020 ; Reed & Alper, 2018 ). Bioremediation is an efficient way to fight against pollution and to warrant environmental performance, biodegradation platforms that are able to remove multiple pollutants are of interest. In the selection of such microorganisms, it must be noted that some pollutants are intrinsically toxic (Beites & Mendes, 2015 ; Ramos et al., 2015 ; Whyte et al., 1997 ). Hence, the selection of biodegradation platforms should be based not only on the basic metabolic potential of the microorganisms but also on their endurance and stress responses. Current legislation limits the use of recombinant microbes in the environment and for this reason microbes that have naturally acquired catabolic genes for removal of pollutants should be considered as a preferential way to build new pathways to expand their metabolic potential (Ramos & Timmis, 1987 ; Reineke & Knackmuss, 1988 ). The soil bacterium Pseudomonas putida survives well in soils and is highly tolerant to hydrocarbons and xenobiotic compounds. Thus, it has the potential to serve as a robust platform for the degradation of chemicals. Pangenome analysis of the species has revealed a series of core genes that provide it with a versatile central metabolism, and which enable it to metabolize natural and xenobiotic compounds. In addition, strains of this species are endowed with a series of efflux pumps that extrude toxic chemicals from the cytoplasm, the periplasm and the outer membrane to the surrounding medium to provide a defensive mechanism against toxic chemicals (reviewed by Udaondo et al., 2012 ; Ramos et al., 2015 ; Udaondo et al., 2016 ). Furthermore, their metabolic potential allows bacteria of this species to adapt to rapidly changing conditions (e.g. oxidative stress, temperature challenges and sudden osmotic perturbations) (Cuenca et al., 2016a ; Cuenca et al., 2016b ; Matilla et al., 2007 ; Molina‐Santiago et al., 2014 ; Nikel & de Lorenzo, 2013 ; Nikel & de Lorenzo, 2014 ; Nogales et al., 2020 ). As such, the core functions define the P . putida species as a robust chassis for environmental survival and operation (Belda et al., 2016 ; Molina et al., 2000 ; Nogales et al., 2020 ; Udaondo et al., 2016 ). Analysis of P . putida accessory genes also revealed a number of catabolic properties shared by two or more strains but not all, for example, chromosomally encoded pathways for degradation of aromatic hydrocarbons—the tod pathway (Zylstra & Gibson, 1989 ) or pathways linked to self‐transmissible plasmids—the TOL plasmid for degradation of toluene via the toluene monooxygenase pathway (Worsey & Williams, 1975 ). \n Pseudomonas putida KT2440 strain is a derivative of P . putida mt‐2, which was isolated as a degrader of 3‐methylbenzoic acid from a soil sample in Japan in 1960 (Nakazawa, 2002 ). This metabolic property was linked to the presence of the catabolic self‐transmissible TOL plasmid pWW0, which encodes a set of enzymes that enable P . putida to grow on several aromatic hydrocarbons such as toluene, m ‐xylene and p ‐xylene (Bagdasarian et al., 1981 ; Worsey & Williams, 1975 ). The strain has also been subjected to mutagenesis and subsequent selection to isolate clones able to degrade recalcitrant chemicals such as p ‐ethylbenzoate (Ramos et al., 1987 ). Pseudomonas putida KT2440 strain is a well‐characterized member of this species and has become a model laboratory microbe for bioremediation. It has also retained its ability to survive and thrive in edaphic and aquatic environments despite a long history in the laboratory (Molina et al., 2000 ). The strain was included as part of a bacterial consortium used to remove petroleum‐based hydrocarbons from soils (Pizarro‐Tobias et al., 2015 ). In this niche it is resilient; field tests revealed that the strain could survive for months in the roots of plants (Molina et al., 2000 ; Niqui‐Arroyo et al., 2013 ; Ronchel & Ramos, 2001 ). In silico analysis of the 6,181,873 bp long genome sequence of P . putida KT2440 (Belda et al., 2016 ; Nelson et al., 2002 ) revealed the lack of any virulence factor in the genome, which was the basis of the early designation of the strain as a generally safe host for recombinant DNA constructs and use in the environment (Kampers et al., 2019 ; Nikel & de Lorenzo, 2018 ; Poblete‐Castro et al., 2013 ; Poblete‐Castro et al., 2020 ; Timmis, 2002 ). The genome of KT2440 encodes a broad array of enzymes that can remove oxygen‐free radicals, i.e. superoxide dismutase, peroxidases and other enzymes capable of dealing with oxidative agents. In a previous study, Niqui‐Arroyo et al. ( 2013 ) described a consortium of microorganisms that grew well with a wide range of linear (i.e. octane, decane, dodecane) and branched hydrocarbons present in wastewaters generated by cleaning airport runways. Analysis of the 16S rRNA amplicons of this consortium revealed a wide range of microbes belonging to genus Acinetobacter , Brevibacterium , Sphingobium , Cupridiavirus , Yersinia , Alteromonas and others. Unfortunately, certain limitations prevent the direct use of the consortium as a bioremediation agent. Namely, some of the microbes have potentially pathogenic characters and certain bacteria could not be recovered and cultured on solid agar plates. To overcome these limitations, we hypothesized that it would be possible to rescue the catabolic potential of the bacteria in hydrocarbon‐enriched samples (i.e. the ability to degrade linear hydrocarbons) by incorporating the corresponding genes via natural gene transfer into the KT2440 (pWW0) chassis (Nogales et al., 2020 ). This was carried out to enhance the ability of KT2440 to catabolize and degrade pollutants.",
"discussion": "RESULTS AND DISCUSSION \n Achieving octane degradation in Pseudomonas putida \n KT2440 ( pWW0 ) \n Because some of the octane‐degrading microorganisms in the consortium did not grow on solid medium (Niqui‐Arroyo et al., 2013 ) we chose to carry out liquid culture matings to transfer genes to KT2440 and enable the strain to grow on linear hydrocarbons. We set up liquid mating by mixing 1 ml of KT2440 (pWW0) grown with toluene as the sole C‐source and 1 ml of the consortium supplied by Bio‐Iliberis R&D. The mixed cultures were kept at 30°C with soft agitation. To counter select the KT2440 (pWW0) capable of degrading octane, we spread 0.1 ml of the mating culture on M9 minimal medium plates with octane as the sole C‐source and 30 μg ml −1 of rifampicin—an antibiotic that KT2440 is resistant to but that none of the consortium bacteria is resistant to. The rate of appearance of Rif R Oct + colonies on these plates was 1 out 10 9 per recipient cell. To test if the colonies on the plates were derived from KT2440 (pWW0), we performed a series of quick metabolic tests which consisted of checking that the cells used three carbon sources: (i) toluene, which is assimilated through the upper and meta cleavage pathway of the TOL plasmid present in the original recipient cells (Worsey & Williams, 1975 ); (ii) citrate, a tricarboxylic acid that P . putida is able to transport into the cytoplasm and is used as a sole carbon and energy source (Herrero et al., 1990 ); and (iii) the newly acquired property of degrading octane. In all cases all strains were positive. To further assure that KT2440 was the host, we used the 27F (5′‐AGAGTTTGATCCTGGCTCAG‐3′) and 1492R (5′‐GGCTCGAGCGGCCGCCCGGG‐3′) primers and amplified the 16S rRNA gene. The sequences we obtained matched >99% to those of P . putida KT2440, which led us to believe that KT2440 (pWW0) had acquired the genes involved in the degradation of octane. Because matings were kept for 48 h, we cannot discard siblings and for this reason a random colony, which we named EM2‐4, was chosen for further characterization. The EM2‐4 strain was able to grow exponentially on octane with duplication times of 105–120 min and yields of about 0.36–0.4 CDW g −1 octane. Surprisingly the strain did not use any other medium‐chain or branched hydrocarbons. It is known that P . putida cells exposed to hydrocarbons fortify their membranes by isomerization of both C16 and C18 cis fatty acids to the trans forms. In agreement with this observation is that the cis : trans ratio of EM2‐4 changed from about 5.9–7.8 in glucose grown cells to 1.04–1.14 in cells growing on octane (Table S1 ). Similar results have been reported before for several Pseudomonas strains (Chen et al., 1995 ). As such, the acquisition of the ability to grow on octane at a high rate does not preclude one of the toxic chemical stress responses of the P . putida KT2440 strain. \n Sequence analysis of the genome of P . putida EM2‐4 \n We sequenced the genome of strain EM2‐4 using Illumina and a genome coverage of 200×. The 6.2 Mb genome was comprised of 78 contigs with sizes ranging between 43,286 and 549,539 bp. We found a contig of 66,272 bp that contained a 61,974 bp insert with flanking chromosomal DNA, which allowed us to identify that the set of genes was on the chromosome and not on the TOL plasmid. In fact, the genetic element gained was located between PP_4491 ( phhB , which encodes a pterin‐4‐alpha‐carbinolamine dehydratase involved in the folate biosynthetic pathway) and PP_4492 ( yhhS , which encodes a carbohydrate efflux transporter) (Figure 1 ). Sequence analysis supported that the acquired DNA with a G + C of 64.86% and encoding 58 ORFs corresponded to an integrative conjugative element (ICE) confirmed by using ICEberg v2.0 database ( https://db-mml.sjtu.edu.cn/CEfinder.html ) (Johnson & Grossman, 2015 ; Liu et al., 2019 ). The online tool from ICEberg database, ICEfinder detects recombination and conjugation modules in the query DNA sequence using Hidden Markov Models and then looks for the oriT region, then it performs a pattern‐based colocalization of gene groups in the element (Li et al., 2018 ). Further analysis comparing the ICE acquired by P . putida against other elements in the ICEberg dataset revealed that several ICE modules of the ICE genes are present on chromosomes or plasmids of multiple microorganisms from different taxa. It appears that the ICE present in KT2440 gained successive sets of genes that most likely originated from Burkholderia , Delftia , Aeromonas , Ralstonia , Shewanella , Paraburkholderia and others (see Table 1 ). The set of ORFs present on the ICE was distributed into three defined regions (Table 1 ; Figure 1 ). The first one is flanked by an integrase (ORF1 of the ICE) and a series of IS 5 , IS 6 , IS 21 and IS 478 sequences (ORFs 33–39 in Table 1 ). This region was populated by genes that encoded proteins of unknown function, and an efflux pump (ORF23‐25). Downstream, a second region contains seven genes putatively involved in the degradation of alkanes, namely: ORF41, which encodes an alkane 1‐monooxygenase that could oxidize octane into octanol; ORF42, a quinone‐dependent alkan‐1‐ol dehydrogenase that transforms octanol to the corresponding aldehyde; ORF44, an aldehyde dehydrogenase that transforms octanol to octanoic acid; and ORF43, a putative acyl‐CoA ligase that permits the entry to the acylated fatty acid to the β‐oxidation cycle (Table 1 ), and two transcriptional regulators that belong to the LysR and AraC families. In addition, we found 13 proteins (from ORF46 to ORF58) which encode a series of transfer functions of the P‐type conjugative system (i.e. TrbCBJKLFGI). FIGURE 1 Schematic representation of the ICE acquired for octane degradation by M2‐4. Arrows indicate direction of transcription. The ORFs are shown: in green are gene clusters related to DNA transfer and stabilization; in orange an efflux pump of the RND family; in red genes related to octane degradation TABLE 1 ORFs present in the ICE and identification of the genetic clusters BlastN best hit NCBI nr/nt bacteria database ORF Amino acids Function Organism ORF Query cover Percent intent \n E ‐value 0 119 Pterin‐4‐alpha‐carbinolamine dehydratase \n Pseudomonas putida \n PP_491 1 401 Integrase \n Burkholderia cenocepacia \n A3203_15290 99 96.33 0.0 2 355 Nuclease domain‐containing protein \n Delftia lacustris \n I6G47_25115 100 93.02 0.0 3 170 DNA repair protein RadC \n Pseudomonas aeruginosa \n EIP87_28665 100 95.69 0.0 4 117 Mlr6156 protein \n Pseudomonas putida \n E6B08_23585 100 97.72 4.00E‐169 5 93 Hypothetical protein \n Pseudomonas putida \n E6B08_23580 100 95.34 6.00E‐121 6 84 Hypothetical protein \n Aeromonas sp. C2U47_04775 100 98.81 4.00E‐122 7 277 UPF0380 proteins YafZ and homologues \n Burkholderia cenocepacia \n A3203_15330 100 96.99 0.0 8 688 Putative plasmid stabilization protein, ParB \n Aeromonas caviae \n WP3S18E02_07940 100 92.89 0.0 9 105 y4eB gene homologue \n Pseudomonas aeruginosa \n K0E62_28125 100 98.1 2.00E + 152 10 111 Transcriptional regulator, Xre family \n Pseudomonas aeruginosa \n MCN99_21695 98 96.34 1.00E‐149 11 117 Putative lipoprotein \n Burkholderia cenocepacia \n A3203_15355 100 96.58 0.0 12 60 Hypothetical protein \n Burkholderia cenocepacia \n CP015036 (3492922…3493101) 100 99.44 4.00E‐71 13 257 Hypothetical protein \n Burkholderia cenocepacia \n A3203_15360 100 94.81 0.0 14 95 Helix–turn–helix domain‐containing protein \n Delftia lacustris \n I6G47_25210 100 96.14 1.00E‐127 15 286 Plasmid replication initiator protein \n Ralstonia solanacearum \n RSc2606 100 95.45 0.0 16 73 Hypothetical protein \n Enterobacter sp. CU081_17605 100 97.65 3.00E‐119 17 213 ParA‐like protein \n Pseudomonas aeruginosa \n CWI20_06980 100 94.21 0.0 18 91 ParB‐like protein \n Thioalkalivibrio sulfidiphilus \n Tgr7_1880 100 93.41 3.00E‐109 19 186 Hypothetical protein \n Pseudomonas aeruginosa \n EIP87_28580 100 94.99 0.0 20 200 S26 family signal peptidase \n Sphingomonas sp. EIK56_27525 93 92.50 0.0 21 43 VirD2 components relaxase \n Variovorax paradoxus \n CP091716 (1641334…1641462) 100 90.77 1.00E‐39 22 665 VirD2 components relaxase \n Burkholderia cenocepacia \n A3203_15395 100 97.46 0.0 23 296 Efflux RND transporter periplasmic adaptor subunit, mdtA \n Shewanella chilikensis \n GII14_05840 100 95.5 0.0 24 1108 Efflux RND transporter permease subunit, mdtB \n Thauera sp. Tmz1t_2073 100 93.8 0.0 25 495 Efflux transporter outer membrane subunit, outer membrane factor (OMF) lipoprotein \n Thauera sp. Tmz1t_2072 100 93.36 0.0 26 583 Ubiquinone biosynthesis protein UbiB \n Diaphorobacter sp. I3K84_20315 93 93.8 0.0 27 325 Patatin‐like phospholipase family protein \n Variovorax paradoxus \n L3V85_07690 100 99.79 0.00E + 00 28 254 Octaprenyl diphosphate synthase \n Hydrogenophaga pseudoflava \n HPF_12290 100 96.33 0.0 29 898 Processive diacylglycerol beta‐glucosyltransferase \n Hydrogenophaga pseudoflava \n HPF_12295 100 96.51 0.0 30 234 CerR family C‐terminal domain‐containing protein \n Hydrogenophaga pseudoflava \n HPF_12300 100 99.72 0.0 31 203 Hypothetical protein \n Hydrogenophaga pseudoflava \n HPF_12305 100 99.51 0.0 32 236 Hypothetical protein \n Hydrogenophaga pseudoflava \n HPF_12310 100 98.59 0.0 33 113 Transposase \n Variovorax paradoxus \n L3V85_07650 100 100.00 8.00E‐176 34 107 IS66 family insertion sequence element accessory protein TnpB \n Variovorax paradoxus \n L3V85_07645 100 100.00 3.00E‐149 35 511 IS66 family transposase \n Variovorax paradoxus \n L3V85_07640 100 100.00 0.0 36 418 IS5 family transposase \n Acidovorax sp. BSY15_3119 100 90.91 0.0 37 515 IS21 family transposase \n Acidovorax sp. Ajs_3594 100 93.72 0.0 38 246 IS21‐like element helper ATPase IstB \n Comamonas thiooxydans \n LCH15_25240 100 94.04 0.0 39 94 IS1478 transposase \n Acidovorax sp. BSY15_166 91 92.66 9.00E‐100 40 371 AraC family transcriptional regulator \n Variovorax paradoxus \n L3V85_07635 100 99.91 0.0 41 754 Alkane‐1 monooxygenase \n Variovorax paradoxus \n L3V85_07630 100 99.96 0.0 42 559 Alkan‐1‐ol dehydrogenase, PQQ‐dependent \n Variovorax paradoxus \n L3V85_07625 100 100.00 0.0 43 542 3‐Methylmercaptopropionyl‐CoA ligase \n Variovorax paradoxus \n L3V85_07620 100 99.94 0.0 44 580 Aldehyde dehydrogenase \n Variovorax paradoxus \n L3V85_07615 100 99.94 0.0 45 315 Transcriptional regulator, LysR family \n Variovorax paradoxus \n L3V85_07610 100 100.00 0.0 46 93 Entry exclusion lipoprotein TrbK \n Variovorax paradoxus \n L3V85_07605 100 100.00 1.00E‐142 47 678 Conjugal transfer protein TraG \n Variovorax paradoxus \n L3V85_07600 100 100.00 0.0 48 159 CopG family transcriptional regulator \n Variovorax paradoxus \n L3V85_07595 100 100.00 0.0 49 355 P‐type conjugative transfer ATPase TrbB \n Variovorax paradoxus \n L3V85_07590 100 100.00 0.0 50 130 Conjugative transfer protein TrbC, TrbC/VirB2 \n Variovorax paradoxus \n L3V85_07585 100 100.00 0.00E + 00 51 91 Family type IV secretion system protein, TrbB/VirB3 \n Variovorax paradoxus \n L3V85_07580 100 100.00 0.00E + 00 52 825 Conjugative transfer protein TrbE \n Pseudomonas aeruginosa \n PERCYII40_2689 100 91.53 0.0 53 252 Conjugative transfer protein TrbJ \n Comamonas sp. FOZ74_04005 100 93.83 0.0 54 95 Conjugative transfer protein TrbK \n Pseudomonas aeruginosa \n FOY97_18900 100 96.84 6.00E‐131 55 460 Conjugative transfer protein TrbL \n Pseudomonas aeruginosa \n FIU24_22485 100 95.88 0.0 56 235 Conjugative transfer protein TrbF \n Pseudomonas aeruginosa \n MBI8613852.1 100 98.72 3.00E‐168 57 333 Conjugative transfer protein TrbG \n Pseudomonas aeruginosa \n PERCYII40_2686 100 97.85 0.0 58 423 Conjugative transfer protein TrbI \n Pseudomonas aeruginosa \n WP_023127271.1 100 98.79 0.0 59 82 Hypothetical protein \n Proteobacteria \n WP_100441186 100 100.00 7.00E‐51 60 401 Putative membrane protein \n Pseudomonas putida \n PP_4492 ORFs in ICE element integrated in EM2‐4. Shaded in different grey tones are the three different sets of genes on the ICE. Alkane monooxygenases of Pseudomonas GPo1 that hydroxylate medium‐chain hydrocarbons usually comprise three components: (i) the AlkB monooxygenase (51% identity with ORF41 in the ICE acquired by EM2‐4), (ii) a soluble NADH‐rubredoxin reductase and (iii) the soluble transfer protein rubredoxin (Alonso & Roujeinikova, 2012 ; Chen et al., 1995 ). The alk cluster in EM2‐4 is atypical in the sense that genes encoding electron donors of the monooxygenase, such as the rubredoxin oxidoreductase or rubredoxin, were not present. Therefore, we assumed that host proteins would participate in ALK reaction electron transfers (Li et al., 2019 ; Rojo, 2009 ). More concretely, the genome of P . putida KT2440 contains a gene (PP_5315 or rubA ) that encodes a rubredoxin‐I and two genes ( alkT or PP_5314 and PP_5371) encoding two rubredoxin‐NAD(+) reductases; enzymes responsible of the reduction of oxidized rubredoxins. Therefore, the acquisition of the octane degradation capability by P . putida KT2440 appears to be the result of a new pathway made of acquired environmental genes and host genes, which work coordinately. We tested if the ICE element was stable and if it can be transferred to other strains. We grew EM2‐4 cells for 100 generations in LB and then cells were spread on M9 with toluene as sole C‐source, M9 with octane as the sole C‐source, or M9 with citrate as the sole C‐source. All of the cells were able to grow in each condition, suggesting stable inheritance of the ICE element. To test if the ICE can be transferred horizontally to other Pseudomonas strains, we mated EM2‐4 with P . putida PSC303 and PSC2078, two Km R derivatives of KT2440 marked with a mini‐Tn 5 , that grow on citrate as a C‐source but not on octane (del Castillo et al., 2008 ). Donor and recipient cells were harvested in the mid logarithmic phase or in the late stationary phase and the mattings were done on solid LB medium plates, M9 minimal medium plates with glucose as a C‐source and in liquid M9 medium for up to 24 h. No Km R Oct + clones were recovered indicating that the rate of transfer of the ICE element under these conditions is below 10 −10 per recipient. \nProteomic analysis reveals how octane is channelled to central metabolism in this strain\n Given that EM2‐4 grew exponentially with octane and that the ICE was genetically stable, we endeavoured to further characterize the phenotype. We carried out proteomic analysis to define the route through which the cells metabolize the linear hydrocarbon. For this, three replicates were performed in two different growth conditions: M9 with glucose (control), and M9 with octane (0.5%, vol./vol.) as sole carbon source (octane metabolism). Doubling times in the exponential phase with glucose (97 ± 3 min) and octane (110 ± 5 min) were similar, and cells were harvested by centrifugation at the mid‐log phase (14,000 g 5 min) when cultures had reached a turbidity of about 0.8 at OD 660 . Cell pellets were frozen at −80°C until used and lysed as previously described (Molina et al., 2019 ). The proteins were digested with trypsin after reduction of disulfur bridges with 100 mM Tris‐carboxyethyl‐phosphine, then alkylated with 200 mM chloroacetamide (see Supplementary Experimental procedures ) and finally TMT‐6plex labelled (Altelaar et al., 2013 ). Then, peptides were identified by high‐throughput‐tandem mass spectrometry and quantified as described in the Supplementary Experimental procedures . We compared the control condition (glucose) with cells grown in octane as sole carbon source in terms of protein enrichment using the T‐fold method of PROTEOBIOTICS, as recommended by the Proteomic Facility of CNB. A total of 2677 different quantification peptide groups were confidently assigned (Table S2 ; data are hosted on figshare and available via https://figshare.com/s/d5bc13bcb41bdaf951fc and DOI: 10.6084/m9.figshare.17061629). Their relative quantities were estimated for each condition based on their respective spectral counts and normalized spectral abundance and the whole‐cell proteomes were compared on the basis of their detection in the three replicates. Data are reported in Table S2 . Figure 2 shows a general overview of the functional categories of the whole cell proteome, weighted by the Normalized Spectral Abundance Factor (Zybailov et al., 2006 ) of the identified proteins in all conditions tested. Proteins involved in central metabolism comprised nearly 67% of total proteins in terms of quantities of the whole cell proteomes. Proteins involved in translation and transcription, cell envelope biogenesis and cell motility and secretion represented each about 10% of total proteins (see Figure 2 ). These ratios are similar to those reported previously for Pseudomonas with other C‐sources (Cuenca et al., 2016a ). This global view of P . putida EM2‐4 protein content indicates no specific bias in the proteomic strategy and points to central metabolism as key octane‐related functional categories for analysis. Using a log 2‐fold change threshold equal or above 1.5 and a stringent statistical level of confidence ( q ‐value <0.05), a list of 89 proteins was identified as statistically significant more abundant in the octane grown than in the glucose grown (Table S3 ), while 37 proteins were less abundant in the whole‐cell proteome in cells growing with octane (Table S4 ). FIGURE 2 Proteomic analysis. Functional categories of proteins displaying loss or gain in cells grown in glucose or octane. Relative quantities of proteins (NSFA) detected in whole‐cell proteome. Proteins (67%) belonged to cell metabolism while 10%, 95% and 11% of proteins were related to DNA transcription/translation, cell envelope and functional organelles \nInsights into octane metabolism based on proteomic analysis\n Octane needs to enter into EM2‐4 cells for oxidation and we identified an uncharacterized porin PP_2662 that was 25‐fold more abundant in octane grown than in glucose grown cells and two porins, OprD and OpdD, involved in amino‐acid uptake that were induced fivefold. It is likely that octane enters into the periplasmic space through these porins (Table S3 , Doncheva et al., 2019 ). Octane is then oxidized to octanol and subsequently to octanal at the membrane level, and eventually to octanoic acid. Regarding oxidation of octane by EM2‐4, we wish to emphasize that host proteins may support the initial oxidation steps up to octanal and of the latter to octanoic as we found that a PQQ‐dependent alcohol dehydrogenase (PP_2679) and an aldehyde dehydrogenase (PP_2680) were induced. This agrees with the notion that octane is being oxidized to octanoic acid. Among the proteins with the higher fold change were enzymes related to acylation of medium‐chain fatty acids, the β‐oxidation cycle and the glyoxylate shunt (Table S3 and Figure 3 ). We identified a set of enzymes involved in the potential acylation of octanoic acid and its subsequent β‐oxidation. The corresponding proteins were induced between 10‐ and 86‐fold, with the highest induction found for two acyl‐CoA dehydrogenase (PP_4948, PP_0370) (see Table S3 ) which are involved in each cycle of β‐oxidation and yield a trans ‐double bond between C2 and C3, which is the substrate of β‐keto‐thiolases (PP_2137 and PP_3754) that are also induced (7‐ to 18.4‐ fold) (Table S3 ). In addition, acyl‐CoA synthetases PP_2351 and PP_4487 were induced 4.7‐ and 8.7‐fold, respectively. Figure 4 shows a STRING‐derived interactome. This figure shows that, among the proteins highly induced in the presence of octane, there were a set of membrane proteins (PP_2662, PP_2663 PP_2667, PP_2669, PP_2674, PP_2675, PP_2678, PP_2679 and PP_2680). FIGURE 3 Schematic representation of octane metabolism in the EM2‐4. Octane likely enters the cells via porins and is oxidized to octanoic acid at the membrane level. Upon acylation it enters the β‐oxidation cycle which yields acetyl‐CoA that is metabolized via the glyoxylate shunt with excess acetyl‐CoA channelled to PHA for storage. Red arrows indicate the cycles under operation with octane and the green arrows indicate the two steps missing in the Krebs cycle in octane‐growing cells FIGURE 4 STRING‐derived interactome of a set of highly abundant proteins in cells growing in octane. The scheme shows that the glyoxylate shunt AceA and GcbB proteins, the set of enzymes related to acylation of medium‐chain fatty acids and the β‐oxidation cycle The second highest fold change (40‐fold) was isocitrate lyase (PP_4116) (Table S3 ), a protein that is involved in central metabolism through its role in the glyoxylate shunt, where glyoxylate is subsequently converted to malate and, as expected, malate synthase (PP_0356, GcbB) was induced (7.1‐fold) (Figure 3 , Table S3 ). Concomitantly, isocitrate dehydrogenase and oxo‐glutarate dehydrogenase, two key enzymes of the Krebs cycle were strongly repressed, which supports blockage of the TCA cycle and full operation of the glyoxylate shunt (Table S4 ). To further confirm the role of the glyoxylate shunt in octane assimilation we generated an ace A mutant by homologous recombination using a Km R cassette that interrupted the ace A gene (see Supplementary Experimental procedures). The Km R \n ace A mutant failed to grow on octane and butanol as expected. We also found induced 4.5‐ to 7‐fold enzymes that are involved in polyhydroxyalkanoate (PHA) synthesis (PP_5007 and PP_5008), suggesting that excess carbon is stored in the form of PHA. This was confirmed by transmission electron microscopy of cells growing on octane as the sole C source (data not shown). PP_4203, an electron transfer flavoprotein‐ubiquinone oxidoreductase, a cytochrome (PP_2675) and cbb3 ‐2 terminal cytochrome oxidase (PP_4256, PP_4255, PP_4258 and PP_4257) were also induced. The induction of this set of respiratory proteins is concomitant with repression of another ubiquinol oxidase involved in electron transfer (PP_4651) and several cytochromes (PP_0813, PP_0105, PP_4193, PP_4251 and PP_4250). This suggests that octane utilization leads to optimization of electron flow by synthesis and removal of a series of electron transfer protein from the respiratory chain. Octane exerts stress in the cells and a number of defence strategies were activated in response to this linear hydrocarbon, for example, the chaperon IbpA was induced 12‐fold (Table S3 ), as well as an efflux pump (PP_2019), and an alkyl‐hydro peroxidase induced over fivefold. The histone‐like HU protein was also induced fivefold (Table S3 ) suggesting a potential role in maintenance of internal chromosome structure. We did not identify an increase in the alkane monooxygenase component or the adjacently encoded alcohol dehydrogenase, which suggests that the oct gene cluster may be expressed constitutively. This is in agreement with the absence of specific alkS/alkT regulatory genes, although other regulators were found adjacent to these catabolic genes in the island. Pathway evolution is thought to follow this route: for hydrocarbon degradation it has been proposed that cells first acquire the catabolic capacity and later the regulatory genes (Pérez‐Pantoja et al., 2021 ). This may also be the case with the acquired oct genes. In the same vein, we did not see changes in the level of the cis‐trans isomerase, in spite of the above experimental data showing a cis to trans isomerization of unsaturated fatty acids (Table S1 ) (Junker & Ramos, 1999 ). This agrees with previous observations that cti isomerase in several strains of Pseudomonas is expressed constitutively (Junker & Ramos, 1999 ) and became functional once exposed to different toxic compounds. As such, responses to toxic hydrocarbons through phospholipid modification are an inherent and constitutive function in Pseudomonas , regardless of the biodegradative potential of the strain. \n Simultaneous degradation of aromatic and lineal hydrocarbons by P . putida \n EM2 ‐4 \n Toluene, xylenes and their corresponding alcohols [benzyl alcohol/3‐methylbenzyl alcohol (3MBA)] and acids [benzoate/3‐methylbenzoate (3MBz)] are degraded by enzymes that are encoded by the TOL plasmid (Worsey & Williams, 1975 ). However, octane is degraded by a hybrid OCT pathway located on the host chromosome. We devised a way to test whether, in EM2‐4, octane can be degraded simultaneously with one of three molecules: (i) glucose, which is metabolized through the Entner–Doudoroff pathway, (ii) the TOL upper pathway substrate 3MBA, or (iii) the TOL lower pathway substrate 3MBz. EM2‐4 cells were grown on glucose in the absence and in the presence of hydrocarbons and when cells reached a turbidity of about 1 cells were harvested and suspended at a turbidity OD 660 of 10 to prepare resting cells. Degradation assays were set up with 50 ppm octane and 3 mM glucose, 1 mM 3MBA or 1 mM 3MBz. We then measured the C‐source concentrations over time. In accordance with the constitutive expression of the OCT pathway we found that regardless of an additional C‐source, octane was removed at a rate of 4.1 ± 0.4 to 15.6 ± 2.0 mg L −1 h −1 (Table 2 ). In fact, octane degradation rate was highest with glucose—an observation that could be explained by the fact that energy generation from glucose could favour cell metabolism. We also found that TOL pathway substrates were consumed at a rate of 12–24 mg L −1 h −1 in the absence and in the presence of octane. Therefore, the EM2‐4 strain has the ability to degrade several pollutants simultaneously. TABLE 2 Rates of OCT and TOL pathways substrate consumption by EM2‐4 resting cells. Substrate consumption rate at 2 h (mg L −1 h −1 ) \n n ‐octane Glucose 3MBA or 3MBz \n n ‐octane 4.1 n.a. n.a. \n n ‐octane + glucose 15.6 544.4 n.a. \n n ‐octane + 3MBA 4.2 n.a. 24.2 \n n ‐octane + 3MBz 5.0 n.a. 12.5 The consumption rate is r = ( c substrate at t \n 0 – c substrate at t \n 1 )/( t \n 1 – t \n 0 ). Octane or aromatic compounds substrate consumption rate ( r ) was determined for the initial 2 h of the assay, while the glucose consumption rate was determined during the initial 60 min. The displayed values are the mean of at least three independent assays. Standard deviations are between 10% and 20% of the provided values."
} | 9,080 |
39921326 | PMC11947988 | pmc | 5,607 | {
"abstract": "Abstract Malonic acid (MA) is a high‐value‐added chemical with significant applications in the polymers, pharmaceutical, and food industries. Microbial production of MA presents enzyme inefficiencies, competitive metabolic pathways, and dispersive carbon flux, which collectively limit its biosynthesis. Here, the non‐conventional oleaginous yeast Yarrowia lipolytica is genetically engineered to enhance MA production. Initially, the malonyl‐CoA pathway, comprising a malonyl‐CoA hydrolase from Saccharomyces cerevisiae , is confirmed as the most efficient for MA production in Y. lipolytica . To further enhance MA production, two novel malonyl‐CoA hydrolases exhibiting higher activity than the hydrolase from S. cerevisiae , are identified from Y. lipolytica and Fusarium oxysporum , respectively. The introduction of the malonyl‐CoA hydrolase from F. oxysporum increases the MA titer to 6.3 g L −1 . Subsequently, advanced metabolic engineering strategies are performed to ensure a sufficient flux of the precursors acetyl‐CoA and malonyl‐CoA for MA production, resulting in a production of 13.8 g L −1 MA in shaking‐flasks. Finally, by employing the fermentation conditions and feeding strategies, a maximum concentration of 63.6 g L −1 of MA is achieved at 156 h with a productivity of 0.41 g L −1 h −1 in fed‐batch fermentation. This study provides a new way for engineering Y. lipolytica to enhance MA production at high titer.",
"introduction": "1 Introduction Malonic acid (MA) is an organic dicarboxylic acid with a wide range of applications in food, pharmaceuticals, manufacturing, and chemical industries. As a platform chemical, MA serves as a precursor for numerous flavors, fragrances, and pharmaceuticals, including cinnamic acid, 3,4,5‐trimethoxycinnamic acid, and γ‐nonanolactone. [ \n \n 1 \n \n ] Additionally, MA can also be used in manufacturing industries, particularly in electronics, [ \n \n 2 \n \n ] specialty solvents, and polymer cross‐linking. [ \n \n 3 \n \n ] For these applications, MA has been listed as one of the top 30 chemicals that can be produced from biomass by the United States Department of Energy. [ \n \n 4 \n \n ] Currently, MA is predominantly produced industrially by chemical synthesis, specifically through the hydrolysis of diethyl malonate and cyanoacetic acid. [ \n \n 5 \n \n ] However, the hydrolysis of diethyl malonate is prone to reversible reactions, MA undergoes decarboxylation, decomposing into acetic acid, water and carbon dioxide by heat under high temperature conditions, which results in low product yield. Conversely, the hydrolysis of cyanoacetic acid is a complex process that generates impurities, thereby reducing the purity of MA. Therefore, there is an urgent need for the development of a sustainable, efficient and environmentally friendly method for the production of MA. Biological production of MA has been demonstrated in the microorganisms of Escherichia Coli , Myceliophthora thermophila , and Saccharomyces cerevisiae . [ \n \n 6 \n \n ] At present, MA can be synthesized via three intermediates: β‐alanine, [ \n \n 6a \n \n ] oxaloacetate, [ \n \n 6c \n \n ] and malonyl‐CoA. [ \n \n 6b \n \n ] The β‐alanine pathway was constructed to produce MA from β‐alanine in E. coli by introducing the β‐alanine pyruvate transaminase encoded by pa0132 gene from Pseudomonas aeruginosa and overexpressing the E. coli succinate semialdehyde dehydrogenase encoded by yneI gene, resulting in 3.60 g L −1 by fed‐batch fermentation. [ \n \n 6a \n \n ] Then, a novel MA synthetic pathway was designed and constructed in M. thermophila using oxaloacetate as a precursor. This pathway involved the conversion of oxaloacetate to malonate‐semialdehyde via oxaloacetate dehydrogenase (Mdc), followed by the reduction of malonate‐semialdehyde to MA by the dehydrogenase YneI, which produced only 42.5 mg L −1 MA. [ \n \n 6c \n \n ] Given the inherent robustness of the budding yeast S. cerevisiae , such as its resistance to acidic conditions and lake of phage contamination, the β‐alanine pathway was first ported from E. coli to S. cerevisiae to facilitate the production of MA in our previous study. [ \n \n 7 \n \n ] However, the maximum titer of MA achieved was only 91.5 mg L −1 , which was significantly lower than that observed in E. coli . It has been reported that the native gene encoding 3‐hydroxyisobutyryl‐CoA hydrolase in S. cerevisiae could be mutated to exhibit malonyl‐CoA hydrolase activity and catalyze the conversion of malonyl‐CoA to MA. [ \n \n 6b \n \n ] To enhance the production of MA, we next constructed the malonyl‐CoA pathway by targeting the mitochondrial 3‐hydroxyisobutyryl‐CoA hydrolase gene EHD3 from S. cerevisiae to the cytoplasm and mutating its active sites to obtain malonyl‐CoA hydrolase activity. [ \n \n 8 \n \n ] This genetic modification, combined with the efforts of improving the precursor supply of malonyl‐CoA and optimizing the fermentation conditions, the MA titer was increased to 1.6 g L −1 after the fed‐batch fermentation. [ \n \n 8 \n \n ] However, this MA titer remains significantly lower than that achieved via the β‐alanine pathway E. coli in the previous study, [ \n \n 6a \n \n ] indicating that the activity of malonyl‐CoA hydrolase might be not high enough for MA production. In addition, using S. cerevisiae as a production host usually challenged with ethanol accumulation, [ \n \n 9 \n \n ] leading to the scattered carbon flux and limiting MA production. Therefore, employing a more efficient malonyl‐CoA hydrolase and an attractive host with abundant malonyl‐CoA flux could potentially enhance MA production to a higher level. \n Yarrowia lipolytica is a “generally regarded as safe” (GRAS) yeast, [ \n \n 10 \n \n ] which has several advantages, including the ability to metabolize various carbon sources, excellent acid tolerance, elevated levels of acetyl‐CoA and malonyl‐CoA, a broad pH tolerance range, the capacity to achieve high cell densities and being unaffected by glucose repression. [ \n \n 11 \n \n ] Recently, the advancements in genetic tools have highlighted the significant potential of Y. lipolytica to produce various malonyl‐CoA‐derived products. [ \n \n 12 \n \n ] \n Y. lipolytica has been effectively engineered and optimized as a microbial cell factory for the production of lipids, [ \n \n 13 \n \n ] natural product [ \n \n 14 \n \n ] and terpenoids, [ \n \n 15 \n \n ] such as squalene, [ \n \n 16 \n \n ] farnesene, [ \n \n 17 \n \n ] germacrene A, [ \n \n 18 \n \n ] and carotenoids. [ \n \n 19 \n \n ] This is largely attributed to the robust activity of its tricarboxylic acid (TCA) cycle and acetyl‐CoA metabolism. As a prototypical oleaginous yeast, Y. lipolytica exhibits exceptional lipid accumulation capabilities, achieving a lipid titer of 72.7 g L −1 and an oil content of 81.4% in bioreactor settings at an industrial scale. [ \n \n 20 \n \n ] Furthermore, the biosynthetic pathway for polydatin has been successfully established in Y. lipolytica , yielding 6.88 g L −1 , which represents the highest reported level of polydatin production. [ \n \n 21 \n \n ] Additionally, the synthesis of another malonyl‐CoA derivative, 3‐hydroxypropionic acid (3‐HP), has been efficiently accomplished, with yields reaching 1.128 g L −1 in shake flask fermentation and 16.23 g L −1 in fed‐batch fermentation using the recombinant strain Po1f‐NC‐14. [ \n \n 22 \n \n ] These achievements are closely linked to the high supply of acetyl‐CoA and malonyl‐CoA precursors of Y. lipolytica . [ \n \n 23 \n \n ] \n In this study, we aim to develop a more efficient Y. lipolytica cell factory for MA production. The efficiencies of three MA synthesis pathways of malonyl‐CoA pathway, malonyl‐CoA and malonate‐semialdehyde pathway, oxaloacetate and malonate‐semialdehyde pathway are evaluated. The malonyl‐CoA pathway is identified as the most effective pathway for MA production. Subsequently, two novel malonyl‐CoA hydrolases exhibiting higher activities are identified from Y. lipolytica and Fusarium oxysporum , respectively, and utilized for MA production. To enhance the precursor supply of acetyl‐CoA and malonyl‐CoA, their competing pathways are further inhibited and the key enzymes are overexpressed. Ultimately, through the combined efforts of fermentation optimization and fed‐batch fermentation, an MA titer of 63.6 g L −1 is achieved. This represents the highest titer reported to date and constitutes a significant breakthrough in the biosynthesis of MA.",
"discussion": "3 Discussion With the ability to grow on various carbon sources including waste cooking oil, [ \n \n 35 \n \n ] excellent acid tolerance, and being unaffected by glucose repression, [ \n \n 36 \n \n ] \n Y. lipolytica has been recognized as an ideal strain for biosynthesis of high value‐added products based on its abundant acetyl‐CoA and malonyl‐CoA precursors. [ \n \n 37 \n \n ] In this study, an effective malonyl‐CoA pathway was constructed to produce MA in Y. lipolytica by introducing the malonyl‐CoA hydrolase from S. cerevisiae and F. oxysporum . Enhancing the fluxes of acetyl‐CoA and malonyl‐CoA enabled the engineered strain to produce a high level of MA both in the shake flasks and the scaled‐up fermentation. Compared with other MA‐producing method, this engineered strain could synthesize MA with higher titer and productivity. Microbial production of MA has been successfully performed in E. coli , M. thermophila and S. cerevisiae via three intermediates of β‐alanine, [ \n \n 6a \n \n ] malonate‐semialdehyde, [ \n \n 6c \n \n ] and malonyl‐CoA, [ \n \n 8 \n \n ] respectively. The β‐alanine pathway requires the involvement of more than six enzymes to convert glucose to MA. This process is further complicated by competition for carbon sources from other metabolic pathways and the low activity of semialdehyde dehydrogenase (YneI), which results in the accumulation of the byproduct of β‐alanine. These factors collectively contribute to a low titer of MA. [ \n \n 6a \n \n ] Our previous study has demonstrated that the β‐alanine pathway is less efficient than the malonyl‐CoA pathway in S. cerevisiae for MA production. [ \n \n 7 \n , \n 8 \n \n ] However, the cytoplasmic levels of malonyl‐CoA S. cerevisiae might insufficient for optimal MA production. Consequently, Y. lipolytica might serve as a more suitable host for MA production via malonyl‐CoA pathway. Currently, no studies have reported the MA production by the two malonate‐semialdehyde pathways in yeast, the Mdc pathway and the McrC pathway. The Mdc pathway is shorter than the β‐alanine pathway, however, the poor acid tolerance of M. thermophila limits its utility as a host for synthesizing MA through MDC pathway. [ \n \n 6c \n \n ] Conversely, the Mdc pathway may perform better in Y. lipolytica due to its excellent acid tolerance. In addition, the McrC pathway, another malonate‐semialdehyde pathway, has significant advantages in the production of 3‐HP. [ \n \n 22 \n , \n 24 \n \n ] To develop an effective synthesis pathway for MA, we evaluated the efficiencies of the malonyl‐CoA pathway and two malonate‐semialdehyde pathways of Mdc pathway, and McrC pathway for MA production in Y. lipolytica . The malonyl‐CoA pathway was constructed by introducing malonyl‐CoA hydrolase derived from the mutated 3‐hydroxyisobutyryl‐CoA hydrolase ScEhd3** for producing MA in S. cerevisiae in our previously study, the highest titer of MA was only 13.6 mg L −1 after shake flask fermentation before engineering the MA synthetic pathway. [ \n \n 8 \n \n ] By expressing the malonyl‐CoA hydrolase from ScEhd3** in the Y. lipolytica Po1f strain, the engineered YMA‐1 strain produced 1.0 g L −1 MA (Figure 1b ), which was a 73.5‐fold increase compared to that produced by the same malonyl‐CoA pathway in S. cerevisiae . This titer was also higher than those produced by the two malonate‐semialdehyde pathways, namely the Mdc pathway and McrC pathway. Future studies aimed to increase MA production via malonate‐semialdehyde pathway should focus on identifying more efficient enzymes, specifically the oxaloacetate decarboxylase, malonyl‐CoA reductase and malonate‐semialdehyde dehydrogenase. These findings indicated that the enrichment of MA synthesis precursors, especially acetyl‐CoA and malonyl‐CoA in Y. lipolytica , making this non‐traditional yeast an ideal host for MA production via the malonyl‐CoA pathway. At present, only the mutation of 3‐hydroxyisobutyryl‐CoA hydrolase (ScEhd3**) in S. cerevisiae has been reported to have the malonyl‐CoA hydrolase activity. In this study, the malonyl‐CoA pathway was constructed by overexpressing the malonyl‐CoA hydrolase ScEhd3** in Y. lipolytica , resulting in a production yield of 3.0 g L −1 following shake flask fermentation (Figure 1e ). However, the conversion rate was still very low, given the utilization of 50 g L −1 glucose. These findings indicated that the activity of malonyl‐CoA hydrolysis is crucial for efficient MA production. To identify more efficient malonyl‐CoA hydrolases for the conversion of malonyl‐CoA to MA, a preliminary screening for malonyl‐CoA hydrolases from fungal origin was performed using a phylogenetic tree based on the amino acids of 3‐hydroxyisobutyryl‐CoA hydrolase ScEhd3 and its conserved amino acids for malonyl‐CoA hydrolase (F121I and E124S) available in NCBI and UniPort databases. Two new malonyl‐CoA hydrolases from Y. lipolytica and F. oxysporum with higher malonyl‐CoA hydrolase activities than ScEhd3** were identified. The MA titer was increased to 6.3 g L −1 by overexpressing FoEhd3** (Figure 2g ). To increase the production of MA, further study might be needed to explore more efficient malonyl‐CoA hydrolases from other origins including plants, animals as well as bacteria. In addition to malonyl‐CoA hydrolase activity, the malonyl‐CoA supply is another limiting step for MA production, the intracellular levels of which is tightly regulated by the accumulation of its precursor of acetyl‐CoA. In Y. lipolytica , more acetyl‐CoA and malonyl‐CoA are used for lipid synthesis, as wildtype strains can accumulate lipids up to 70% of dry biomass. [ \n \n 38 \n \n ] We employ a ‘restrain–pull’ strategy effectively switching the carbon metabolic flux from lipid to acetyl‐CoA by deleting two diacylglycerol acyltransferases of Dga1 and Dga2 to decrease the consumption of acetyl‐CoA, overexpressing the enzymes involved in β‐oxidation pathway and acyl‐CoA synthase Faa1 to increase the acetyl‐CoA supply. The titer of MA was significantly increased to 10.2 g L −1 through reducing the synthesis of lipids (Figure 3c,d ). The regulatory mechanism of the acetyl‐CoA carboxylase (Acc1) in Y. lipolytica has not been well studied yet, although it has been reported that the activity of ScAcc1 could be tightly regulated by Snf1 kinase by direct phosphorylation. [ \n \n 28 \n \n ] Here, by comparing the amino acid sequence of ScAcc1 and YlAcc1, we found that YlAcc1 also shared the predicted phosphorylation sites (S667 and S1178) of Snf1 kinase. The MA production was increased 34.8% by overexpressing the mutated YlAcc1 S667A‐S1178A (Figure 3e ), indicating that similar regulation mechanism might be exited for YlAcc1 as the ScAcc1 by protein kinase of Snf1. In fed‐batch fermentation, despite the improved MA level using glycerol as a carbon source, the highest biomass of YMA‐23 reached only ≈ 150 (OD 600 ), which was just 68.2% of that using glucose as a carbon source (Figure 5a,b ). The addition of glycerol resulted in the elevation of the acetyl‐CoA pools, facilitating the flux of carbon metabolism to MA thereby competing for some carbon sources, which may have resulted in the low biomass. [ \n \n 30b \n \n ] Owing to the Crabtree effect, S. cerevisiae produces a large amount of ethanol in the presence of oxygen and excess glucose, leading to a loss of carbon for the biosynthesis of non‐ethanol chemicals. [ \n \n 39 \n \n ] In contrast, Y. lipolytica is not inhibited by high sugar concentrations because of its Crabtree‐negative characteristic, and its rate of sugar consumption is much higher than that of other microorganisms. [ \n \n 35a \n \n ] Therefore, the initial concentrations of the carbon sources were usually set to a relatively high level. [ \n \n 40 \n \n ] Currently, microbial cell factories utilizing Y. lipolytica as host cells exhibit generally low product conversion rates. A significant proportion of carbon sources is diverted toward lipid synthesis and other pathways, resulting in wasted carbon sources, [ \n \n 41 \n \n ] which limits the industrial application of Y. lipolytica as a host strain. To maximize the utilization of different carbon sources, enhancing the activity of key enzymes within the synthetic pathway emerges as the most effective strategy. Given that lipids constitute the primary by‐products of Y. lipolytica , promoting lipid degradation and facilitating their conversion into desired products is also a viable approach. Furthermore, coupling the TCA cycle with the crucial intermediate acetyl coenzyme A via the citric acid cycle can further enhance carbon source utilization. Organelle engineering holds the potential to localize the production of target compounds, thereby minimizing substrate loss and improving the carbon conversion rate. [ \n \n 16 \n , \n 42 \n \n ] Additionally, transporter engineering can address the limiting factors such as product inhibition, thereby enhancing overall process efficiency. [ \n \n 43 \n \n ] \n In conclusion, an efficient MA‐producing strain was constructed by the selection of different MA synthetic pathways and identification of the effective malonyl‐CoA hydrolases in this study. To improve the supply of acetyl‐CoA and malonyl‐CoA, the strain was metabolically engineered to improve carbon metabolic flux from lipid to acetyl‐CoA by blocking the flow of TAG, enhancing the β‐oxidation pathway as well as eliminating the inhibition of Acc1 via mutation of its two possible phosphorylation sites. Finally, the titer of MA was increased to 63.6 g L −1 by using glycerol as carbon source and supplementing MnCl 2 in the culture medium in fed‐batch fermentation. This work will facilitate the realization of the industrial production of MA and provide a reference for the production of other malonyl‐CoA derivatives in the Y. lipolytica ."
} | 4,576 |
38057633 | null | s2 | 5,610 | {
"abstract": "Herein, peptide nucleic acids (PNAs) are employed in the design of a participatory duplex PNA-peptide crosslinking agent. Biophysical and mechanical studies show that crosslinkers present during peptide assembly leading to hydrogelation participate in the formation of fibrils while simultaneously installing crosslinks into the higher-order network that constitutes the peptide gel. The addition of 2 mol % crosslinker into the assembling system results in a ~100 % increase in mechanical stiffness without affecting the rate of peptide assembly or the local morphology of fibrils within the gel network. Stiffness enhancement is realized by only affecting change in the elastic component of the viscoelastic gel. A synthesis of the PNA-peptide duplex crosslinkers is provided that allows facile variation in peptide composition and addresses the notorious hydrophobic content of PNAs. This crosslinking system represents a new tool for modulating the mechanical properties of peptide-based hydrogels."
} | 250 |
33522527 | null | s2 | 5,613 | {
"abstract": "The extracellular matrix (ECM) is a water-swollen, tissue-specific material environment in which biophysiochemical signals are organized and influence cell behaviors. Electrospun nanofibrous substrates have been pursued as platforms for tissue engineering and cell studies that recapitulate features of the native ECM, in particular its fibrous nature. In recent years, progress in the design of electrospun hydrogel systems has demonstrated that molecular design also enables unique studies of cellular behaviors. In comparison to the use of hydrophobic polymeric materials, electrospinning hydrophilic materials that crosslink to form hydrogels offer the potential to achieve the water-swollen, nanofibrous characteristics of endogenous ECM. Although electrospun hydrogels require an additional crosslinking step to stabilize the fibers (allowing fibers to swell with water instead of dissolving) in comparison to their hydrophobic counterparts, researchers have made significant advances in leveraging hydrogel chemistries to incorporate biochemical and dynamic functionalities within the fibers. Consequently, dynamic biophysical and biochemical properties can be engineered into hydrophilic nanofibers that would be difficult to engineer in hydrophobic systems without strategic and sometimes intensive post-processing techniques. This Review describes common methodologies to control biophysical and biochemical properties of both electrospun hydrophobic and hydrogel nanofibers, with an emphasis on highlighting recent progress using hydrogel nanofibers with engineered dynamic complexities to develop culture systems for the study of biological function, dysfunction, development, and regeneration."
} | 426 |
39408575 | PMC11476194 | pmc | 5,615 | {
"abstract": "The efficient production of biobased organic acids is crucial to move to a more sustainable and eco-friendly economy, where muconic acid is gaining interest as a versatile platform chemical to produce industrial building blocks, including adipic acid and terephthalic acid. In this study, a Saccharomyces cerevisiae platform strain able to convert glucose and xylose into cis , cis -muconic acid was further engineered to eliminate C2 dependency, improve muconic acid tolerance, enhance production and growth performance, and substantially reduce the side production of the intermediate protocatechuic acid. This was achieved by reintroducing the PDC5 gene and overexpression of QDR3 genes. The improved strain was integrated in low-pH fed-batch fermentations at bioreactor scale with integrated in situ product recovery. By adding a biocompatible organic phase consisting of CYTOP 503 and canola oil to the process, a continuous extraction of muconic acid was achieved, resulting in significant alleviation of product inhibition. Through this, the muconic acid titer and peak productivity were improved by 300% and 185%, respectively, reaching 9.3 g/L and 0.100 g/L/h in the in situ product recovery process as compared to 3.1 g/L and 0.054 g/L/h in the control process without ISPR.",
"introduction": "1. Introduction To transition the current fossil fuel-based economy to a more resilient biobased economy focused on sustainability, alternative production processes for everyday chemicals are required. Microbial production of platform molecules used to synthesize these chemicals offers a valuable alternative in this respect. The biobased production of 2,4-hexadienedioic acid, also referred to as muconic acid (MA), has attained attention over the last decade. This unsaturated dicarboxylic acid with two reactive dicarboxylic groups can be used to produce a diverse set of polymers, for example, polyesters [ 1 ]. In addition, MA can be converted into other platform chemicals of great significance. For example, MA can be hydrogenated to yield adipic acid, one of the monomers in nylon-6,6 [ 2 ]. Moreover, MA can be converted into levulinic acid, another platform chemical with a projected market size of 2400 tons yearly by 2025 [ 3 ]. Biobased synthesis of MA, more specifically the cis , cis -MA isomer (ccMA), has been successfully achieved through the metabolic engineering of various microorganisms. The synthesis is either based on the bioconversion of aromatic (lignin-derived) feedstocks or on de novo synthesis from carbohydrates. The former has resulted in yields of up to 85 g/L [ 4 ], where the main challenge remains the complexity of lignin as a feedstock, which requires extensive upstream purification [ 5 , 6 ]. The highest reported titer from de novo production to date is 88.2 g/L, accomplished with Corynebacterium glutamicum using glucose as the carbon source and maintaining the pH at 7.0 [ 7 ]. Yeast microbial cell factories are an appealing alternative due to, amongst other things, their high tolerance towards organic acids at low pH, resistance towards toxic inhibitors, and the possibility of them to utilize carbon sources from second-generation biomass [ 8 , 9 , 10 ]. Although research on non-conventional yeasts has gained momentum [ 11 ], most research to establish a eukaryotic production host has focused on Saccharomyces cerevisiae . However, the production of ccMA in S. cerevisiae is constrained by the significant inhibitory effect of MA on yeast viability, resulting in maximum titers, yields, and productivities in a fed-batch 2 L bioreactor of 22.5 g/L, 77 mg/g glucose, and 0.191 g/L/h, respectively [ 12 ]. While these results were achieved in buffered conditions at pH 6.0, the toxicity at low pH is significantly increased, as ccMA exists primarily in its protonated form, allowing it to diffuse into the cells. In unbuffered conditions, concentrations as low as 5 g/L ccMA caused a significant reduction of 43% in the maximum specific growth rate of the laboratory reference strain S. cerevisiae CEN.PK 113-7D [ 13 ]. Despite this, low-pH fermentation remains attractive, as medium neutralization requires high amounts of acids and bases, typically resulting in significant salt waste, and can therefore compromise economic feasibility [ 14 , 15 ]. In that respect, continuous extraction of organic acids during fermentation in in situ product recovery (ISPR) processes has emerged as an attractive technique to reduce product inhibition and increase fermentation performance. Here, reactive extraction using organic phases containing amines, phosphorous compounds, and ionic liquids has been described to efficiently separate organic acids from fermentation broth. This, moreover, leads to a partial purification and concentration of the target product in the organic phase and can hence reduce downstream purification efforts [ 16 ]. This approach has been previously evaluated for MA production using S. cerevisiae , where adding polypropylene 400 to shake flask fermentations extracted MA but did not further improve strain performance [ 17 ]. In another study, a biocompatible reactive extraction mixture was developed containing CYTOP 503 in canola oil and was successfully applied to fermentation processes at the shake flask level with S. cerevisiae , resulting in a growth and MA titer improvement of 44% and 18%, respectively. A proof of the strategy at 10 L bioreactor scale led to a maximum MA titer of 4.33 g/L at pH 4.0 [ 18 ]. The present study uses a previously engineered MA-producing S. cerevisiae strain, originally constructed by Nicolaï and coworkers [ 17 ], as a basis for further engineering. The modifications present in that original strain are shown in Figure 1 . This strain can convert glucose and xylose to ccMA as it holds a heterologous pathway for MA biosynthesis from 3-dehydroshikimate (DHS). DHS dehydratase from Podospora anserina , protocatechuic acid (PCA) decarboxylase (PCAD) from Klebsiella pneumoniae , and oxygen-consuming catechol 1,2-dioxygenase (CDO) from Candida albicans were expressed, while ethanol production was eliminated by deletion of PDC1 , PDC5 , and PDC6 . In addition, feedback inhibition for the entry into the shikimate pathway was eliminated by the integration of a feedback-resistant DAHP synthase. Moreover, the overexpression of PAD1 in this strain enhanced the production of prenylated flavine mononucleotide (prFMN), a cofactor of PCAD. To alleviate the growth on glucose of this pyruvate-decarboxylase negative (Pdc − ) strain, an internal deletion of MTH1 , MTH1 ΔM41-T78 , was introduced [ 19 , 20 , 21 ]. This strain, referred to as S. cerevisiae TN22, served as the starting point for the current work, where engineering efforts were focused on the reduction of the C2 dependency and the improvement of ccMA tolerance ( Figure 1 ). The improved strain was evaluated in a fed-batch fermentation at bioreactor scale with ISPR, aiming to further reduce product inhibition and enhance the process performance at low-pH to contribute to a more sustainable production of MA and, by extension, other biobased platform chemicals.\n\n2.1. Elimination of C2 Dependency for ccMA Production by Reintroduction of PDC5 Fermentation trials with the previously constructed MA-producing host strain, S. cerevisiae TN22, demonstrated its dependence on an endogenously added C2 carbon source, e.g., ethanol, to support ccMA production from glucose or xylose ( Figure 2 A). When supplemented with 1% ethanol, fermentation in yeast extract-peptone-dextrose (YP4%D) and yeast extract-peptone-xylose (YP4%X) yielded titers of 2.09 g/L and 2.17 g/L ccMA, respectively. In contrast, only 0.158 and 0.372 g/L ccMA were produced in YP4%D and YP4%X without ethanol. In addition, growth in media without ethanol was limited. To address its C2 dependence, S. cerevisiae TN22 was subjected to adaptive laboratory evolution (ALE). Ten populations were cultured in media where ethanol concentrations were gradually lowered. For five of these populations, prior to evolution, additional genetic diversity was introduced by performing EMS mutagenesis. Over a period of three months, a total of 15 media transfers were performed, alternating glucose and xylose as carbon sources to avoid undesired trade-offs. At regular intervals, cells were plated. A total of 129 evolved clones were tested for their ability to produce ccMA in shake flasks in YP4%D, without the addition of ethanol ( Figure 2 B). Upon fermentation completion, 19 evolved clones yielded a ccMA titer higher than 0.5 g/L. The top-performing clone produced 1.57 g/L ccMA from 4% glucose, corresponding to a yield of 39.9 mg/g glucose and approaching the yield of the unevolved strains on YPD supplemented with ethanol (41.8 mg/g glucose and ethanol). However, for the top-performing mutants, xylose conversion into ccMA in YPD was compromised, delivering titers below 0.350 g/L ccMA when YP4%X was fermented. Instead, as an alternative approach to address the ethanol dependency of S. cerevisiae TN22, PDC5 was reintroduced. Four constitutive promoters of variable strengths were used. Fermentation profiles are shown in Figure 2 C,D. PDC5 expression under the TEF1 and RPL18B promoters showed restoration of ccMA production in YP4%D without the addition of ethanol. During the first phase of the fermentation, ethanol was produced (24 h), which was subsequently completely consumed to produce ccMA. Moreover, growth of the modified strains was significantly improved compared to S. cerevisiae TN22. Strain S. cerevisiae TN22 RPL18Bp_PDC5 was chosen as a starting strain for further engineering efforts.",
"discussion": "3. Discussion In this work, the previously constructed MA-producing host strain S. cerevisiae TN22 [ 17 ] was further improved. To abolish ethanol formation as a byproduct, which diverts carbon away from MA, pyruvate decarboxylase-encoding genes were deleted in this strain. Therefore, like other Pdc − strains, it is unable to grow in excess glucose, a defect likely caused by repression of the respiratory metabolism and a lack of cytosolic C2 supply [ 20 ]. Even though Nicolaï and coworkers engineered a mutation known to alleviate this defect, MTH1 ΔM41-T78 [ 17 ], strain S. cerevisiae TN22 only showed limited growth with barely any ccMA production in medium without a C2 source (0.158 g/L in YP4%D). Since the requirement of ethanol addition constrains the industrial applicability of this MA production host, we attempted to address this defect by subjecting TN22 to ALE. Over the course of 15 rounds, several populations were transferred in media with ethanol concentrations that were gradually lowered. Fast-growing clones were selected for evaluation and yielded multiple promising clones, with the best isolate delivering a titer of 1.57 g/L ccMA from 4% glucose. Unfortunately, these promising isolates failed to substantially improve MA production in a xylose-containing medium without the addition of a C2 source. Two conclusions can be drawn from these observations. First, mutations that alleviate C2 dependency for MA production can be carbon source-dependent. Second, adaptive mutations that overcome the C2 auxotrophy of Pdc-negative strains for the utilization of xylose seem to be harder to acquire than mutations that allow the utilization of glucose. Since the ability to utilize xylose as a carbon source is vital for any industrial application of an MA-producing host strain, an alternative strategy was explored. The ability to produce ethanol was partially restored by the reintroduction of PDC5 . A set of four promoters of variable strengths was used to express PDC5 , and expression under an RPL18B promoter was sufficient to support MA production when glucose or xylose served as carbon sources. Since no further improvement in MA yield, productivity, or titer could be observed when the stronger TEF1 promoter was applied, S. cerevisiae TN22 RPL18Bp_PDC5 was chosen for further engineering efforts. We subsequently increased tolerance to MA by enhancing its efflux. To this end, the multidrug transporter encoded by QDR3 , known to mediate MA export, thereby conferring MA tolerance [ 22 ], was overexpressed. With increasing promoter strengths applied, MA titers of the engineered strains increased, with expression under a TEF1 promoter resulting in the highest MA titers. Conversely, accumulation of the pathway intermediate PCA decreased with stronger promoter strengths. This reduction in byproduct formation is crucial for process economics, as it minimizes carbon diversion to unwanted products and simplifies product purification. Given the beneficial nature of this modification, it was engineered in strain S. cerevisiae TN22 RPL18Bp _ PDC5 , where similar improvements were observed. Given that the overexpression constructs with the highest promoter strengths yielded the largest improvements, at this moment it remains unclear whether further increasing QDR3 expression, i.e., by integrating multiple copies of the overexpression constructs, may further improve the production host. The improved S. cerevisiae TN22 RPL18Bp_PDC5 ; TEF1p_QDR3 strain was subsequently implemented in fed-batch fermentations in 2 L bioreactors, reaching an end titer of 3.1 g/L with a peak productivity of 53.9 mg/L/h, which was in line with the shake flask experiments. However, product inhibition was distinct, and at a titer of 2.2 g/L of MA, production and growth stagnated and productivity had sharply decreased. At this timepoint, the pH of the medium had dropped to 4.0 from the initial value of 5.5. This observation of product inhibition of MA is in accordance with other studies ( Table 2 ) showing that the highest performances were reached by neutral-pH fermentation, while low-pH fermentation has been limited by product inhibition. Toxicity towards S. cerevisiae was observed in fermentations without pH control at titers as low as 2.9 g/L [ 17 ] and 5 g/L [ 13 ]. Also, in other yeast species such as P. occidentalis , growth was completely stopped after a drop of pH to 2.0 and could only be restored partly at pH 3.5-4.0. Titers of MA in that case were reduced to 7.2 g/L compared to 38.8 g/L at a neutral pH [ 8 ]. The reason for the pronounced inhibition is directly linked to the dissociation behavior of MA being protonated at pH levels below pH 4.0 (pKa 1 : 2.9, pKa 2 :4.0) [ 23 ], enabling the diffusion into the cells. To address the product inhibition, reactive extraction emerged as an attractive strategy to extract organic acids from low-pH fermentation broth, as commonly applied extractants in the literature show enhanced performance for protonated organic acids [ 36 , 37 ]. One constraint is reported to be the toxicity of the extractants to the microorganisms, but mixing extractants with biocompatible diluents can reduce toxicity and yield mixtures that can directly be added to the fermentation broth [ 38 ]. In our previous study, such a mixture containing the phosphorous based extractant CYTOP 503 (12.5 v%) and canola oil was developed and proved to be non-toxic for S. cerevisiae while retaining high extraction capacity for MA [ 18 ]. These findings are in line with the here attained data, and no negative interference with the new yeast strain was found. On the contrary, all performance parameters were significantly increased when ISPR was applied, including MA production, productivity, growth, and yield. To that end, an overall maximum titer of 9.3g/L was obtained in the ISPR fermentation. Nevertheless, at the end of the fermentation, product inhibition was found recurrent, and while the results prove the effectiveness of ISPR, they also reveal the limitations of the ISPR process in a batch mode. In this respect, various options to industrially apply ISPR based on reactive extraction are described, where a continuous separation with a constant inflow of fresh organic phase and a constant outflow of loaded organic phase allows further process intensification. This can be achieved internally in the bioreactor with specialized equipment or externally by circulating the broth or fermentation supernatant over an external extraction unit [ 39 , 40 , 41 ]. Here, resource recycling is crucial, and future research should focus on efficient back-extraction techniques for the recovery of the organic acid and to allow the recycling of the reactive extraction mixture. In parallel, strain engineering efforts should be focused on increasing MA tolerance at low pH to enhance the overall fermentation performance to ultimately yield an economic process. ALE in S. cerevisiae presents an attractive approach to enhancing tolerance to ccMA, as it has been successfully applied to improve tolerance to other compounds such as acetic acid and lactic acid tolerance [ 42 , 43 ]. Notably, increases in the copy number of QDR3 were previously observed during laboratory evolution experiments on adipic acid and glutaric acid, two other dicarboxylic acids [ 22 ]. In conclusion, in this study, an MA producing strain was successfully engineered with enhanced MA tolerance, production, and growth, while substantially reducing the side production of the intermediate PCA. Since organic acid production is limited by product inhibition at low-pH, ISPR was successfully applied to fermentation processes using the improved strain. By adding an organic reactive extraction mixture to the process, MA was continuously extracted, resulting in significant improvements in MA production and growth. This can serve as a basis for a sustainable, biobased production of MA and, by extension, other organic acids at low pH. Future research should investigate specialized industrial equipment for continuous operation, as well solvent recycling and reuse."
} | 4,464 |
20393580 | PMC2872612 | pmc | 5,616 | {
"abstract": "Sustainable biofuel alternatives to fossil fuel energy are hampered by recalcitrance and toxicity of biomass substrates to microbial biocatalysts. To address this issue, we present a culture-independent functional metagenomic platform for mining Nature's vast enzymatic reservoir and show its relevance to biomass conversion. We performed functional selections on 4.7 Gb of metagenomic fosmid libraries and show that genetic elements conferring tolerance toward seven important biomass inhibitors can be identified. We select two metagenomic fosmids that improve the growth of Escherichia coli by 5.7- and 6.9-fold in the presence of inhibitory concentrations of syringaldehyde and 2-furoic acid, respectively, and identify the individual genes responsible for these tolerance phenotypes. Finally, we combine the individual genes to create a three-gene construct that confers tolerance to mixtures of these important biomass inhibitors. This platform presents a route for expanding the repertoire of genetic elements available to synthetic biology and provides a starting point for efforts to engineer robust strains for biofuel generation.",
"introduction": "Introduction Global environmental problems related to the combustion of fossil fuels and increasing concerns about their supply underscore the importance of developing renewable fuel alternatives with a reduced environmental footprint. The application of synthetic biology ( Baker et al , 2006 ; Ro et al , 2006 ) to engineer biocatalysts that produce biofuels from diverse lignocellulosic materials including waste and low agricultural intensity biomass holds promise to deliver one such sustainable alternative ( Farrell et al , 2006 ; Tilman et al , 2006 ; Fargione et al , 2008 ; Searchinger et al , 2008 ). However, bioconversion of lignocellulose to biofuels is currently limited by biomass recalcitrance ( Himmel et al , 2007 ) and toxicity of non-fermentable compounds in the original substrate and formed as byproducts of biomass pretreatment ( Klinke et al , 2004 ). Although the identity and inhibitory concentrations of these compounds have been characterized, their mechanisms of toxicity are poorly understood, and genes conferring tolerance to most of these compounds have not been identified. A synthetic biology approach to design efficient biocatalysts for biofuel generation requires a diverse inventory of functional genetic machinery allowing usage of or conferring tolerance toward these compounds. As plant biomass is constantly recycled in the environment ( Kirk and Farrell, 1987 ), a reservoir of enzymatic machinery must exist in the soil microbiome that allows for the tolerance and complete processing of its constituent chemicals. However, the majority of this microbial processing machinery has remained inaccessible to synthetic biology and metabolic engineering, as a majority of microbes in the soil are recalcitrant to culturing ( Torsvik et al , 1998 ). We show the utility of culture-independent metagenomic functional selections for discovery of novel functional genes from the soil microbiome, enabling expansion of the synthetic biology toolbox for lignocellulosic biomass conversion and tolerance. The key steps of this method involve extraction of metagenomic DNA from arbitrary environmental sources ( Rondon et al , 2000 ), transformation of environmental metagenomic libraries into the microbial host of interest, and selection of functional genetic elements conferring the desired phenotype compatible with the chosen host ( Figure 1 ). This method is well suited for biomass catalysis, as the functional genes that allow the host to use recalcitrant substrates or tolerate toxic chemicals can be directly selected from arbitrary metagenomic libraries.",
"discussion": "Results and discussion Engineered Escherichia coli strains have been shown to harbor many advantages as biocatalysts for biofuel production, including the ability to ferment a majority of plant-derived monosaccharides, no requirements for complex growth factors, and earlier industrial use, but still suffer from lower tolerance to biomass inhibitors when compared with other candidate biofuel producing organisms ( Dien et al , 2003 ). Therefore, as a proof of concept, we applied metagenomic functional selections to select a number of functional genetic elements from diverse soil microbiomes that confer resistance in E. coli to different classes of biomass inhibitors. We extracted metagenomic DNA from four different soil microbiomes ( Table I ), with an optimized protocol to carefully purify high molecular weight DNA (Materials and methods; Supplementary methods ). We chose to create large-insert (40–50 kb) libraries to allow for the potential discovery of phenotypes requiring multiple genes. Four metagenomic libraries of sizes ranging from 0.2 to 2.5 Gb were created in a single-copy fosmid vector, and transferred into an E. coli host using phage transduction ( Table I ; Materials and methods). The concentrations of seven important biomass chemicals that inhibit the growth of the wild-type E. coli host were determined using growth assays on Luria Broth (LB) agar media with sparse concentration range screening ( Supplementary Table I ) based on previously published results ( Zaldivar and Ingram, 1999 ; Zaldivar et al , 1999 , 2000 ). These chemicals span the three major classes of biomass inhibitors (organic acids, alcohols, and aldehydes), which accumulate during the pretreatment of biomass and agricultural and municipal waste, resulting in the inhibition of microbial biofuel fermentation ( Klinke et al , 2004 ). Selection of the four metagenomic libraries on solid media containing each of the seven biomass compounds at their determined inhibitory concentrations yielded metagenomic library clones that survived in 15 of the 28 combinations ( Table I ). Tolerant clones were identified against all seven inhibitors. Methylcatechol was the only compound for which tolerant clones were identified from all four soil libraries. Hydroquinone and furfural tolerant clones were identified in one library. Interestingly, the medium-sized (1 Gb) dairy farm soil library yielded clones tolerant to the most inhibitors (six out of seven), whereas the pH 5.5 bog soil library with the largest size (2.5 Gb) yielded clones tolerant to two of the seven inhibitors. These phenotypic results show the utility of metagenomic functional selections for transferring lignocellulosic inhibitor tolerance phenotypes to E. coli . As these phenotypes are encoded on large DNA inserts, it is important to identify which and how many specific genetic elements within these operon-sized stretches are responsible for these phenotypes. As a proof of concept, we chose to focus on a subset of clones for more in-depth analysis of growth kinetics in the presence of inhibitors, sequencing of the full-length inserts, and sub-cloning of functionally annotated genes. Inhibitory chemicals are derived from two major sources during biomass pretreatment—depolymerization of complex lignin polymers and degradation of biomass sugars ( Klinke et al , 2004 ). Accordingly, we chose clones conferring tolerance to one lignin monomer, syringaldehyde, and one biomass sugar degradation product, 2-furoic acid, for further phenotypic and genetic analysis. For each inhibitor, 20 metagenomic clones with improved tolerance on solid selection media ( Supplementary Table I ) were retested in liquid culture at three concentrations spanning the range of previously reported inhibitory concentrations (Materials and methods), and for all clones improved phenotype was confirmed ( Supplementary Figure 1 ). From this set, one clone for syringaldehyde (mgSyrAld) and one clone for 2-furoic acid (mgFurAc), with the greatest difference in cell growth when compared with control at 1.55 g/l syringaldehyde and 0.8 g/l 2-furoic acid, respectively, were selected for analysis of growth kinetics (Materials and methods). Metagenomic inserts from these clones were extracted and retransformed into wild-type E. coli to confirm that the improved phenotype was due to the presence of the metagenomic insert (Materials and methods). The phenotypic improvements were 5.7-fold for syringaldehyde and 6.9-fold for 2-furoic acid, expressed as fold improvements in cell growth at an inhibitor concentration that results in a 90% reduction of wild-type E. coli cell growth ( Figure 2A and B ). The mgSyrAld and mgFurAc metagenomic inserts were sequenced at three-fold coverage, assembled, and annotated ( Figure 2C and D ) (Materials and methods; Supplementary Tables II and III ). Regions of the metagenomic sequences with the highest detectable homology to the NCBI non-redundant nucleotide database using BLAST ( Altschul et al , 1990 ) are 7% of mgFurAc with 79% identity to a region of the Pelobacter propionicus DSM 2379 genome and 1% of mgSyrAld with 73% identity to a region of the Burkholderia ambifaria AMMD chromosome 2, indicating that the selected metagenomic sequences are largely novel. Based solely on the sequence and computational annotation of the inserts, it is difficult to predict which genes are responsible for the improved tolerance especially as the mechanism of toxicity is poorly characterized for these compounds. Therefore, we performed a loss-of-function study with mgSyrAld and mgFurAc using transposon mutagenesis to identify the functional genetic elements contributing to the selected phenotypes ( Figure 2C and D ) (Materials and methods). The 192 transposon-inserted clones per inhibitor created for sequencing of the mgSyrAld and mgFurAc fosmids were individually subjected to kinetic growth survival assays in the presence of 1.4 g/l syringaldehyde and 0.8 g/l 2-furoic acid, respectively. For mgSyrAld, 3 of the 192 unique transposon insertions resulted in a knockdown of the improved syringaldehyde tolerance, all mapping to either the promoter or the N-terminal coding region of a 348 amino acid gene product annotated to be a UDP-glucose-4-epimerase ( Figure 2D ) (Materials and methods). For mgFurAc, 7 of the 192 unique transposon insertions resulted in a knockdown of the improved 2-furoic acid tolerance, with three hits mapping to the coding region of a 342 amino acid gene product annotated to be a RecA protein, and four hits mapping to a 111 amino acid gene product with predicted membrane-spanning domains but of unknown function ( Figure 2C ) (Materials and methods; Supplementary information ). Interestingly, these two genes are more than 10 kb apart in the mgFurAc metagenomic fragment, and none of the annotated genes between these locations appear to contribute to the selected phenotype based on the transposon mutagenesis results. Although the mechanism of inhibition in E. coli by syringaldehyde and 2-furoic acid is unknown, the gene hits identified in our selection may provide starting points for discovery of the underlying modes of inhibition and rescue ( Supplementary information ). To verify that the three genes implicated in the loss-of-function studies were necessary and sufficient for the syringaldehyde and 2-furoic acid tolerance phenotypes, we PCR amplified each gene from the corresponding selected metagenomic fosmid (mgUdpE from mgSyrAld, and mgRecA and mgOrfX from mgFurAc), sub-cloned them into an expression vector, transformed these into wild-type E. coli , and repeated the tolerance assays (Materials and methods; Supplementary information ). In all three cases, the individual genes exhibited improved tolerance to the inhibitors when compared with wild-type E. coli ( Figure 3 ). The improved syringaldehyde tolerance because of the mgUdpE gene was very similar to mgSyrAld ( Figure 3B ). In contrast, both mgRecA and mgOrfX were individually unable to completely recapitulate the level of mgFurAc tolerance ( Figure 3A ), which might be expected if both genes are required for the full observed phenotype of mgFurAc. To test this hypothesis, we created a bicistronic construct (mgRecA_mgOrfX) by sub-cloning the mgOrfX gene and its upstream ribosome-binding site (RBS) between the mgRecA gene and the transcription terminator site. The bicistronic mgRecA_mgOrfX construct exhibited improved 2-furoic acid tolerance when compared with both individual genes, and more closely resembled the growth behavior of mgFurAc ( Figure 3A ). This shows the utility of metagenomic selections using large-insert metagenomic libraries to identify multiple genetic elements, which can together contribute to improved phenotypes despite not being immediately co-localized in gene sequence. For instance, a metagenomic library smaller than 10 kb would have been unable to capture the improved phenotype derived from the action of both genes. One of the goals of synthetic biology is to improve microbial phenotypes by integrating multiple functional genetic elements from arbitrary genetic or engineered sources. To test whether the selected genes that confer tolerance to E. coli when exposed to the biomass inhibitors syringaldehyde and 2-furoic acid individually could also confer tolerance to a mixture of these inhibitors, we created a tri-cistronic construct (mgRecA_mgOrfX_mgUdpE) by sub-cloning the mgUdpE gene and its upstream RBS just downstream of the bicistronic mgRecA_mgOrfX construct. The tri-cistronic construct exhibited improved growth phenotypes in the presence of mixtures of syringaldehyde and 2-furoic acid ( Figure 3C ; Supplementary Figure S4 ) showing that genes selected using this metagenomic selection platform against individual inhibitors can be combined to create constructs that confer tolerance toward inhibitor mixtures. Strain optimization has previously been achieved through adaptive evolution ( Yomano et al , 1998 ; Herring et al , 2006 ; Liu, 2006 ) where rare beneficial genomic mutations can be selected without earlier knowledge about the mode of inhibition. Adaptive evolution is ideally suited for optimization of the genomic inventory of functions in a given strain, but the timescale for evolving entirely new functions is generally prohibitive, as they typically require a substantial number of specific mutations. One of the strengths of metagenomic functional selections for strain improvement is that its success in discovering functional genetic elements is similarly independent of earlier knowledge regarding the mode of inhibition of selected chemicals, while sampling a large reservoir of novel genes from the metagenome. The genetic machinery that we can evolve, select, and engineer into microbial biocatalysts for tolerating or degrading biomass inhibitory chemicals can be considered mechanistically analogous to genetic machinery encoding microbial antibiotic resistance. Both biomass-derived inhibitors and most antibiotic chemicals are produced naturally in the environment ( Walsh, 2003 ), and microbial communities have likely evolved similar biochemistries to tolerate and process these xenobiotics. Mechanisms of antibiotic resistance include (a) adaptive genomic mutation that obscure the cellular target of the drug without compromising the native cellular function, (b) acquired mechanisms to degrade or expel the drug, often gained through horizontal transfer through plasmid exchange, and (c) adjusting expression in other pathways to increase target production or bypass it with redundant functions ( Davies, 1994 ; Walsh, 2003 ). Metagenomic functional selections for biomass tolerance and conversion are likely to uncover genetic machinery that parallels the second and third mechanisms. Indeed, we hypothesize that the metagenomically selected proteins homologous to RecA and UDP-glucose-4-epimerase encoding resistance to 2-furoic acid and syringaldehyde, respectively, may function to complement or rescue the activity of their putatively compromised native cellular counterparts ( Supplementary information ). In future applications of this methodology, mechanisms to enzymatically metabolize biomass inhibitors may have enhanced utility because they would increase the nutrient yield and net carbon flux, though these mechanisms would be undesirable when the intended tolerance phenotype is against the biofuel product (e.g. ethanol). The screening of metagenomic clone libraries from diverse environmental sources has previously yielded numerous biomolecules including novel proteases, amylases, cellulases, and antibiotics ( Brady and Clardy, 2000 ; Rondon et al , 2000 ; Daniel, 2005 ; Warnecke et al , 2007 ), and the yields of these methods appear primarily limited by the number of clones that can feasibly be screened ( Daniel, 2005 ). In comparison, a library-wide selection scheme allows for exhaustive interrogation of the enzymatic reservoir encoded within metagenomic libraries that can be made using current techniques (⩽5 × 10 10 bp) ( Riesenfeld et al , 2004 ; Daniel, 2005 ). Functional selections have been designed for hundreds of anabolic, catabolic, and resistance phenotypes in E. coli ( Neidhardt et al , 1996 ; Sommer et al , 2009 ), and opportunities exist for the design of selections for other biotechnologically relevant phenotypes including controlled flocculation, surface adhesion, natural competency, and cell–cell communication ( Williamson et al , 2005 ). We have shown that metagenomic functional selections can successfully discover functional genetic elements encoding chemical tolerance relevant to biomass conversion. The same platform can be applied to select for microbial usage and production of specific biomass chemicals. The repertoire of biomass substrates that can be used by a microbial biocatalyst has been expanded by transfer of specific genetic machinery for substrate metabolism from other microbes with these properties ( Jin et al , 2005 ). Similarly, substrate usage machinery may be selected from a metagenomic clone library by providing the substrate as the sole source of a required nutrient (e.g. carbon, nitrogen), only allowing clones expressing functional genes enabling substrate usage to grow selectively. Functional selections for chemical production can be achieved in a biocatalyst metagenomic clone library that contains a biochemical circuit that links the presence of the desired product to a selectable resistance or usage phenotype. For instance, a circuit can be designed where a transcription factor responsive to the stoichiometric presence or absence of the product controls the expression of an antibiotic resistance gene ( Desai and Gallivan, 2004 ). A distinguishing feature of synthetic biology is the emphasis on integrating arbitrary genetic elements to generate robust and predictable biocatalysts to solve multiple biological, chemical, and engineering problems including fuel generation, environmental remediation, and pharmaceutical production ( Baker et al , 2006 ). Our work shows that metagenomic functional selections enable the direct discovery of novel genetic elements from Nature's enzymatic catalogue, providing a route for expanding the synthetic biology tool box."
} | 4,773 |
36345483 | PMC9636870 | pmc | 5,617 | {
"abstract": "Coral reefs are declining worldwide primarily because of bleaching and subsequent mortality resulting from thermal stress. Currently, extensive efforts to engage in more holistic research and restoration endeavors have considerably expanded the techniques applied to examine coral samples. Despite such advances, coral bleaching and restoration studies are often conducted within a specific disciplinary focus, where specimens are collected, preserved, and archived in ways that are not always conducive to further downstream analyses by specialists in other disciplines. This approach may prevent the full utilization of unexpended specimens, leading to siloed research, duplicative efforts, unnecessary loss of additional corals to research endeavors, and overall increased costs. A recent US National Science Foundation-sponsored workshop set out to consolidate our collective knowledge across the disciplines of Omics, Physiology, and Microscopy and Imaging regarding the methods used for coral sample collection, preservation, and archiving. Here, we highlight knowledge gaps and propose some simple steps for collecting, preserving, and archiving coral-bleaching specimens that can increase the impact of individual coral bleaching and restoration studies, as well as foster additional analyses and future discoveries through collaboration. Rapid freezing of samples in liquid nitrogen or placing at −80 °C to −20 °C is optimal for most Omics and Physiology studies with a few exceptions; however, freezing samples removes the potential for many Microscopy and Imaging-based analyses due to the alteration of tissue integrity during freezing. For Microscopy and Imaging, samples are best stored in aldehydes. The use of sterile gloves and receptacles during collection supports the downstream analysis of host-associated bacterial and viral communities which are particularly germane to disease and restoration efforts. Across all disciplines, the use of aseptic techniques during collection, preservation, and archiving maximizes the research potential of coral specimens and allows for the greatest number of possible downstream analyses.",
"introduction": "Introduction Coral reefs provide sustenance, goods, and services for coastal communities worldwide and likely harbor more than one third of all marine species ( Fisher et al., 2015 ). However, corals and reef frameworks are increasingly being degraded due to anthropogenic disturbances ( Intergovernmental Panel on Climate Change, 2022 ). Climate change has severely affected coral reef health on a global scale, primarily through increased sea surface temperatures leading to devastating coral bleaching events. The increased frequency and intensity of these events reduces the capacity for reef recovery and restoration efforts ( Heron et al., 2016 ; van Hooidonk et al., 2016 ; Sully et al., 2019 ), and successive bleaching events have decreased live coral cover by up to 60% in some localities ( Miller et al., 2009 ; Raymundo et al., 2019 ; Dalton et al., 2020 ). Up to one third of all reef-building corals species may be at risk of extinction from the combined effects of bleaching and local stressors such as nutrient pollution, overfishing, and habitat destruction ( Pandolfi et al., 2003 ; Carpenter et al., 2008 ; Plaisance et al., 2011 ; Hughes et al., 2017 , 2018 , 2019 ). Given the increased frequency and severity of bleaching events, scientists and restoration practitioners need to study coral bleaching and disease more. One way to achieve greater efficiency is through the implementation of a common framework recently developed for coral bleaching experiments ( Grottoli et al., 2021 ). Another is by reducing the number of duplicative efforts more broadly and maximizing the number of analyses that can be performed on sampled specimens through greater collaboration. Identifying common methodological pipelines in collecting, preserving, and archiving Between 2014 and 2021 over 20,000 coral specimens and samples were collected for bleaching studies ( McLachlan et al., 2021 ), many of which are suitable for additional analyses that could address new questions concerning various aspects of bleaching. The technology and methods commonly used in coral biology research have quickly progressed in recent decades ( Cziesielski, Schmidt-Roach & Aranda, 2019 ; Grottoli et al., 2021 ). The combination of traditional and modern genomic insights, physiological metrics, and microscopy and imaging analytics have together given scientists an ever-expanding toolkit to interrogate the mechanisms and results of coral bleaching and restoration efforts at the subcellular, cellular, tissue, and organismal levels. Integration of these approaches thus allows individual specimens to be used for multiple downstream applications and expands the potential utility of every coral sample collected. Despite this, scientists and practitioners tend to sample, preserve, and archive specimens in a manner specific to their own specialized applications or aims, and on average only conduct one or two downstream analyses per study ( McLachlan et al., 2021 ). However, it is unclear how many or how often archived samples are utilized. Yet, limits exist on how many tools individual researchers can manage, conduct, and financially support. Trained in increasingly complicated fields of study, it is impractical for any one scientist, or even a team of scientists, to have the breadth of knowledge, skills, and resources to conduct the full range of possible Omics, Physiology, and Microscopy and Imaging analyses on any given set of specimens. However, with effective documentation during sampling ( Grottoli et al., 2021 ), coupled with strategic preserving and archiving decisions, specimens could be available to additional research teams, who could increase the number of analyses ultimately conducted on a given set of samples, contributing to a better understanding of bleaching mechanisms with less sampling and experimental damage to reefs. As the numbers and expertise of scientific investigators expand, so do the tools, methods, and perspectives at their disposal. We brought together investigators from around the world to further synthesize research methods in order to identify low-cost and practical ways to share specimens, reduce duplicative efforts, and increase the end-use potential of samples generated in coral bleaching research and restoration programs ( Fig. 1 ). We identified and consolidated working pipelines that could (1) expand the number of potential analyses on currently archived samples, and (2) assist in future project planning to maximize the number of potential downstream analyses while minimizing any extra work, time, or funds required. While no single methodological pipeline can be all-inclusive, several critical steps in these methodological pipelines were found to optimize the potential utility of each coral specimen within the constraints of a given study design ( Fig. 2 ). 10.7717/peerj.14176/fig-1 Figure 1 Flow chart of conceptual design for workshop on methods of collecting preserving and archiving coral bleaching specimen. 10.7717/peerj.14176/fig-2 Figure 2 Methodological pipeline used during the preservation and archiving of coral specimens for research and restoration purposes for various downstream analyses. The orange columns on the left-side of the figure ( i.e ., columns B and C) indicate the methods used during the collection and archiving of coral fragments, categorized by the chemical preservatives and fixatives used and the method of temperature storage ( i.e ., room temperature (RT), frozen (F), refrigerated (RF), or oven dried (D)). For example, a specimen collected using the row 7 pipeline is immediately frozen using liquid nitrogen, and then subsequently stored in a conventional freezer ( e.g ., −80 °C) whereas a specimen collected using the row 20 pipeline is first stored in ethanol and then placed in a refrigerator at 4 °C. The remaining columns ( i.e ., columns D–AM) describe whether a specimen collected using a given pipeline is suitable for a variety of downstream measurements such as DNA analyses or chlorophyll quantification. Downstream analyses are categorized into three disciplines: (1) Omics, (2) Physiology, and (3) Microscopy and Imaging ( i.e ., row 2). These columns are further subdivided based on the specific type of coral material being used ( i.e ., coral host tissue (CH), algal symbionts (AS), microbiome (M), skeleton (SK) or gametes (G)). Five levels of appropriateness are herein described: Optimal (O), Acceptable (A), Undesirable (U), not acceptable (N) and unknown (?). These designations are based upon publish methodological data as well as the consensus scientific opinions of 30 coral scientists who attended the Coral Bleaching Research Coordination Network meeting in June 2020. The percentage of downstream analyses which were afforded an optimal or acceptable appropriate designation is shown in column AN. The total number of potential pipelines that are acceptable or optimal for a given downstream analysis are shown in row 59. The total number of potential pipelines for which the suitability is unknown is shown in row 60. Consolidating methods for broadening participation Cheap and unifying methods can serve to increase participation and inclusion in coral bleaching and restoration research, particularly for those with minimal funding. Clear, simple guidelines for specimen and sample collection, manipulation, and preservation can also make it easier for experts working on parallel questions in non-coral systems to bring their hypotheses and approaches to bear on the coral bleaching and restoration fields. Adapting and expanding sampling, preserving, and archiving of specimens in ways that allow for additional downstream analyses can generate research opportunities for early career scientists and students, providing a mechanism for additional collaboration and more entry points into the field of coral research, as well as creating new opportunities for collaborations and networking between researchers with distinct yet complementary areas of inquiry, thereby fostering advances and new ideas within the field. These efforts support the inclusion of researchers in the field who may not currently conduct marine fieldwork due to lack of access to resources ( e.g ., funding, SCUBA gear, boat access, laboratory equipment), training ( e.g ., scientific dive certifications), and/or physical or logistical capability. A separate challenge in promoting diversity and inclusion in the broader field of coral research is to connect researchers that have samples with other scientists and managers (including undergraduate trainees and volunteers) from diverse disciplines and backgrounds that can run additional analyses. A database of samples and researchers (and their research interests/skill sets) could be useful in identifying and jump-starting fruitful collaborations and sample sharing. Numerous community-based resources can also provide data storage options to both facilitate data archiving and reuse, including those specific to coral research, restoration and biodiversity ( e.g ., GEOME ( Deck et al., 2017 ; Riginos et al., 2020 )). Going forward, implementation of specific collection, preservation, and archiving pipelines developed herein could further maximize and foster more collaboration among diverse community members and stakeholders. Consolidating methods for restoration specimens Coral restoration and rehabilitation programs aim to assist in the recovery of reef ecosystems through passive and active means, and for the ultimate goal of creating a reef that can independently continue to develop without further intervention ( Boström-Einarsson et al., 2020 ). Recent efforts to explore the success and failure of some restoration programs have revealed a lack of coordinated efforts among restoration practitioners, scientists, and managers. Further, some restoration programs remain unlinked to scientific endeavors that could track natural biological, chemical, and oceanographic phenomena that provide mechanistic context for why some coral propagation and outplanting efforts result in success while others do not. Collaborative work to engage in scientific inquiries before, during, and after restoration efforts, along with standardized practices, could accelerate and advance restoration programs. For example, genetic, physiological, and microbiome sampling of specimens from restoration corals that are successfully outplanted have revealed key aspects of why some genotypes and species are more resistant or resilient to local and global stressors ( Baums, 2008 ; Lohr & Patterson, 2017 ; Morikawa & Palumbi, 2019 ; Klinges et al., 2020 ; van Woesik et al., 2021 ; Voolstra et al., 2021 ). Thus, the consolidated methods presented herein can be used to bridge the gaps between the restoration and research communities more readily and completely. General considerations for collecting, preserving, and archiving coral bleaching specimens The central aim of our workshop was to identify simple and low-cost methods within the three broad categories of Omics, Physiology, and Microscopy and Imaging analyses that could increase the impact of every coral bleaching study in an effort to best understand scientific principles and increase restoration and conservation success. In the process, we uncovered several key issues that all researchers and managers can consider regardless of individual subfields, including: (1) specimen and sample provenance and metadata, (2) sample collection considerations, and (3) sample handling and storage considerations . It is also important to consider how collection, preservation, and storage methods may shift the accuracy or precision of downstream analyses. For a more elaborate discussion of specific methods see the Supplemental Materials . Specimen/Sample provenance and metadata Museums, research aquariums, and private collections have standard protocols for documenting the history, or origin, of individual specimen ( Smithsonian Institution, 2006 ; National Science & Technology Council, Interagency Working Group on Scientific Collections, 2009 ; National Academies of Sciences, Engineering, & Medicine, 2020 ). Researchers and practitioners can optimize the use of their data and samples by rigorously cataloguing, and formally documenting as many experimental ( e.g ., temperature ramp rate, light level, flow), biological ( e.g ., coral color, morphotype, taxonomy, provenance), and environmental ( e.g ., depth, nutrient concentrations, reef type) variables as possible ( Grottoli et al., 2021 ) because these data provide needed context for each collection. We refer to these descriptive, contextual data as metadata. Representative samples can also be properly ‘vouchered’ with a museum for long-term preservation, and such specimen can have important applications for a wide range of future work from these biological collections. First, if such samples include both tissue and the taxonomically informative skeleton, they can provide a taxonomic reference in the event that cryptic species are discovered, or to assign identity of the samples with future changes to taxonomy. Such vouchered samples also provide invaluable reference barcodes for databases that are becoming increasingly important as environmental DNA (eDNA) approaches become commonplace. Likewise, techniques change through time and questions that would have been impossible to address from such samples a couple of decades ago have become common place today with the advent of high-throughput and single-cell sequencing. Finally, even in cases where there is no obvious need to preserve the samples, the value of having historical samples has been showcased repeatedly in the field of epidemiology and emerging zoonotic disease research, where natural history collections have been integrated with host-pathogen research to resolve pathways of transmission ( Thompson et al., 2021 ). The questions that will be answered by historical samples may yet be unknown, but it is only possible to address them if the samples are collected, vouchered and properly maintained. Sample provenance also includes the documentation of how and where samples and their resulting data and metadata are physically and digitally stored. Growing recognition of the value of historical data and appreciation for FAIR (findability, accessibility, interoperability, and reusability) data standards ( Wilkinson et al., 2016 ) is inspiring the efforts to archive sample data and metadata in ways that facilitate reuse and ensure archived data is available to future researchers ( Zerbino et al., 2018 ; Davis et al., 2019 ; Percie du Sert et al., 2020 ). For example, the Genomic Observatories Metadatabase ( Deck et al., 2017 ), stores metadata archives permanently linked to -omics resources stored at the National Center of Biotechnology Information’s (NCBI, Bethesda, MD, USA) and the National Science Foundation’s Data Management Office (BCO-DMO) serve as repositories where samples and associated metadata are linked to the researchers who produced those studies and can be contacted about collaboration or specimen sharing. Last, in regard to data provenance, many funding agencies have specific data management and dissemination requirements ( e.g ., BCO-DMO at the National Science Foundation, GenBank at the National Center for Biotechnology Information, Environmental Data Service at the Natural Environment Research Council). However, relevant details concerning these samples and their province legacy data are often overlooked by researchers. For example, a recent sampling of the Sequence Read Archive (SRA) of GenBank found that only ~14% of all archived specimens associated with a sequencing project included both collection year and site as basic metadata that would be required for the reuse of archived genomic data in future studies ( Toczydlowski et al., 2021 ). As the culture of global research and reef conservation and restoration have moved toward more open and collaborative models, there is growing pressure from funding bodies, journals, management agencies, and researchers alike to provide these data in open-access formats ( Sibbett, Rieseberg & Narum, 2020 ), and develop community-wide cyberinfrastructure that facilitates the discovery and reuse of material samples ( e.g ., iSamples; Davies et al., 2021 ). Such consolidated efforts stand to benefit the advancement and accessibility of the field of coral bleaching research and restoration science and effort as a whole. Specimen/Sample collection considerations There is a myriad of possible techniques for collecting, processing, and archiving most coral specimens (for more details see Supplemental Materials ). However, unique differences among coral taxa including individual colonies, their morphotypes, tissue thicknesses, skeletal density, and variation in life history demand special consideration as these variables may affect the biology and chemistry of collected coral samples and could dictate the applicability of many downstream procedures. Additionally, colony and specimen/sample size as well as species-specific variation can affect how corals respond to and recover from stress ( Brandt, 2009 ; Thomas & Palumbi, 2017 ; Álvarez-Noriega et al., 2018 ; Levas et al., 2018 ). The quantity of available sample material can also affect what downstream techniques are possible. Improving the precision of measurements of colony and specimen size is an active area of research ( Table S1 ) with the advent of new technological developments such as 3D laser scanning and photogrammetry ( House et al., 2018 ; Vivian et al., 2019 ; Zawada, Dornelas & Madin, 2019 ). Information about the original size of the parent colony or outplant specimen can provide helpful information for interpreting resulting data because size has been shown to be an important bleaching predictor ( Álvarez-Noriega et al., 2018 ). Collection permits may also restrict the number of samples that can be collected, which can affect the types of analytical methods that are possible downstream and how much excess material may or may not be available for archiving and future research. Further, some agencies restrict the use of specimen for explicitly defined goals within a given project or on a specific permit and thus may be unavailable for alternative end goals. Researchers should be aware of such agency-based limitations and enquire with providers on any downstream use restrictions. Last, developmental stage can have significant impacts on which methods are suitable and practical for any methodological pipeline. For example, the amount of material required for some analyses may be prohibitive when working with coral larvae or gametes, but easily performed on adult tissues. Thus, the types of research questions that can be addressed will vary depending on the life stage of the specimen and dictate the types of downstream analyses and collaborations that are most productive. Temperature and sample storage considerations When collecting and preserving coral bleaching and restoration specimens for short and long-term use, documenting a sample’s temperature history is critical (see Box 1 on freezing and cryopreservation). In general, altered temperatures can cause rapid state changes in live specimen physiology, microbiology, and geochemistry. Many subcellular and cellular processes can change within minutes to hours when corals undergo shifts in ambient temperature ( Hillyer et al., 2017a ), and swift sample processing is important to capture those responses. Once samples are preserved, temperature can further influence the integrity of each sample for some types of analyses. For example, cells lyse if samples are too cold, thus making them unsuitable for imaging of intact cells. Each scientific discipline (Omics, Physiology, and Microscopy and Imaging) has guidelines for optimal preservation temperatures suitable to ensure the integrity for their analytic process (see Fig. 2 ). The duration of storage for these specimens can also dictate ideal archiving temperature conditions. If samples are intended to be stored for tens of years ( e.g ., in coral gamete biobanks), cryopreservation and downstream restoration, rapid-freezing in liquid nitrogen, and storing at −80 °C are the safest holding temperatures. If the tissues or cells can tolerate freeze-drying, and the final packaging is vacuum-sealed, then such specimens can be maintained for many years at room temperature. 10.7717/peerj.14176/table-1 Box 1 Freeze it and forget it? Freezing material is at the heart of maintaining robust tissue archives. But what are the consequences of some of these freezing processes in terms of tissue quality over time? Before deciding how to store samples, both the sensitivity of the downstream analysis ( e.g ., RNA stability) and how long that process needs to be viable should be considered. The cryopreservation field is rapidly evolving, especially for human samples. For example, standard practice for understanding tumor physiology was to fix in formalin, embed in paraffin, and store at room temperature. However, delicate RNA can degrade over time under these conditions but remains robust if stored at −80 °C ( Baena-Del Valle et al., 2017 ). Thus, coral RNA and enzyme specimens may best be stored at −80 °C, potentially remaining stable for up to 10 years at these low temperatures and making them suitable for additional downstream analyses. For corals, storing at −80 °C allows for the highest number of downstream analyses ( Fig. 2 ). However, longer-term stable storage (>tens of years) at liquid nitrogen temperatures (−196 °C) is preferable ( Ortega-Pinazo et al., 2019 ; Kelly et al., 2019 ), though highly impractical for many researchers due to the cost and equipment needs associated with ultra-cold storage. In contrast, many laboratory analyses can be reliably performed on specimens stored at −20 °C ( Fig. 2 ) for 2 to 5 years. \n Frozen But Alive: Cryopreservation Holds Material Safely for Many Years \n Cryobiology is the study of cells and tissues at cold temperatures. The central principle in cryopreservation is to avoid the formation of lethal intracellular ice. Generally, cryopreservation uses permeating cryoprotectants or solutes, such as dimethyl sulfoxide (DMSO), methanol or propylene glycol, and non-permeating solutes, such as sugars ( e.g ., glycerol), to allow the permeating cryoprotectants to enter cells and block ice crystal formation, and to permit the non-permeating solutes to dehydrate and remove intracellular water to reduce and avoid ice formation. Once cells and tissues are safely cryopreserved and held at liquid nitrogen temperatures, most biological processes are reduced. Theoretically, if cells are maintained at liquid nitrogen temperatures, they can survive for thousands of years with minimal damage. Thus, cryopreservation of living coral tissue and maintenance in liquid nitrogen ( e.g ., cryobanks) provides access to a multitude of scientific and restoration uses because the tissues are frozen, but also alive. Once the cryoprotectants are warmed and the cells are rehydrated, they are alive, and any number of analyses can be done post-thawing. However, cryopreserved cultured cells are equally robust at either −196 °C or −80 °C using a number of metrics over 8 years ( Miyamoto et al., 2018 ). Even in properly cryopreserved samples, tissue degradation can occur if samples are removed from a freezer to subsample and then refrozen or exposed to heat transients by opening and closing of a freezer door. Thus, avoiding any changes in freezer temperatures is ideal. To date, cryopreservation processes have been used to preserve coral sperm from over 48 species worldwide ( Hagedorn et al., 2012 ). This international collaboration has used frozen sperm to subsequently fertilize coral eggs and create new coral larvae ( Hagedorn et al., 2017 ). Moreover, frozen sperm has also been used to demonstrate the feasibility of assisted gene flow in the critically threatened coral Acropora palmata ( Hagedorn et al., 2021 ). Frozen coral material is now archived in biorepositories around the world, and some of the material for the assisted gene flow experiments was stored for up to 10 years before successful use in fertilization experiments. However, coral bleaching and restoration research is often conducted in locations where adequate freezing agents and materials ( e.g ., liquid nitrogen, dry shippers, or even ice) may not always be available or reliable. Although not all methods require temperature stabilization, many do ( Fig. 2 ). Therefore, if possible, all researchers should record above and below water (1) the transport holding temperature, (2) any altered temperatures during transport, and (3) the duration of transport. For example, if live or dead specimens were removed from an offshore reef, transported to shore, and placed in new containment, the method and duration of transport as well as the temperature of any onshore activities ( e.g ., freezer storage, water temperature manipulation) should be documented and included in the sample metadata. Specimen handling and sterility considerations There is increased interest in how the coral holobiont microbiome ( i.e ., Symbiodiniaceae, bacteria, viruses, and other microscopic eukaryotes) responds to, and may be involved in, preventing or exacerbating coral bleaching and/or increasing or reducing restoration success. Many ecological and physiological bleaching studies can be easily paired with Symbiodiniaceae analyses ( e.g ., cell densities, gene sequencing) through shared samples, but the potential for coral bacterial and/or viral analyses is severely compromised when sterile collection tools ( e.g ., gloves, bone-cutters) and sterile receptacles (bags or tubes) are not used. The use of aseptic handling techniques during coral collection and processing is a relatively small and inexpensive change in the methodological pipeline that can enable additional downstream microbiome analyses ( Fig. 3 ). For example, a suitable aseptic technique in the field may be as simple as wearing nitrile gloves when handling corals and using sterile receptacles, such as Whirl-Pak sample bags, and minimizing cross-contamination by replacing with clean equipment or sterilizing in between handling different specimens. For microbiome work, additional sampling of the environment ( e.g ., water and sediments) can provide information about sources of potential contamination. Importantly, while aseptic techniques are ideal for many downstream applications, it is impractical if not impossible to maintain underwater and in some handling situations. 10.7717/peerj.14176/fig-3 Figure 3 Pictogram outlining and quantifying some of the most commonly used sampling and preservation pipelines for exploring questions about coral bleaching and restoration. Shown in the decision tree are (1) the choice of different steps in coral specimen handling (aseptic on the left and non-aseptic on the right) (2) the range of initial short term storage techniques (ST) which include different types of temperature storage or chemical preservation and the range of preservation methods (3) and a selection of long-term (LT) storage options. Using summary data from Fig. 2 , we calculated the total numbers of possible downstream analyses (values in colored boxes, ranging from fewer techniques (blue) to higher (bright pink)) that can be conducted for the three primary disciplines discussed in this review (omics, physiology, and microscopy techniques) when different combinations of aseptic or septic technics along with different ST and LT options are chosen. The use of aseptic techniques (depicted on the left-hand side) in sample collection, preservation, storage, and archiving increases the number of possible downstream analyses shown at the bottom, relative to specimens handled using non-aseptic tools and receptacles (right-hand side), particularly in the Omics category. Similarly, freezing of samples at any point in the pipeline may limit the number of Microscopy and Imaging analyses that can be applied, though higher temperature storage points (>4 °C) and sample state changes limit the utility of specimens for many Omics and Physiological analyses as detailed in Fig. 2 . Freezing of samples in liquid nitrogen (LN 2 ) or between –80 °C and –20 °C while using aseptic technique allows for the greatest number of the common downstream analyses depicted, represented by the bright pink cells. While many other applications and methods are possible, those depicted here are a small representation of common methods and not meant to be a fully inclusive list. Specific conditions for many analyses are more thoroughly outlined in the table of Fig. 2 . Caveats and considerations for methodologies, accuracy, and usability In each discipline there may be recommended and, in some cases, well benchmarked standard operating protocols for each individual method discussed below and in the Supplemental. However, many of the methodological pipelines discussed below which are suitable for some aspects of coral-bleaching and restoration research but have not yet been fully evaluated in terms of their accuracy and precision. Therefore, deviations from standard procedures for a given discipline could potentially result in data that are inaccurate, uninterpretable, or unusable. It is important to consider the potential caveats when using any non-standard procedure in field and laboratory work. Yet, as research techniques improve and additional methods and protocols are confirmed as having high precision and accuracy, more of the potential pipelines discussed below may be employed with confidence in any given discipline. For example, using chemically fixed ( e.g ., in formaldehyde) samples for genomic-based analysis has been considered non-standard in the past, but new work shows that these preserved specimens can be reliably used to gain insight into various aspects of coral biology, retrospectively ( Greene et al., 2020 ). Identification of consolidated methodological pipelines for general use in coral bleaching and restoration studies A previous literature review identified many methodologies in coral-bleaching studies ( McLachlan et al., 2021 ), broadly categorized into three disciplinary areas: Omics ( e.g ., genomics, epigenomics, transcriptomics, metagenomics, amplicon analysis, proteomics, and metabolomics), Physiology ( e.g ., chlorophyll, lipid/protein/carbohydrate concentrations, biomass, tissue and skeletal stable isotopes), and Microscopy & Imaging-based analyses ( e.g ., Symbiodiniaceae density measures, electron microscopy, histology, Raman spectroscopy). Using 36 defined analytical assessments ( Fig. 2 columns), we quantitatively determine which methodological pipelines can maximize the number of downstream procedures across these three disciplinary areas. We assigned several broad categorical terms to determine whether a step in the pipeline was ‘ O ptimal,’ ‘ A cceptable,’ ‘ U ndesirable,’ or ‘ N ot Acceptable.’ Steps within pipelines marked ‘undesirable’ indicate that there may be research to show the method is not ideal, or that it is illogical to pursue a particular pipeline based on past evidence. Thus, caution should be taken when evaluating these incomplete pipelines. Further, in many cases it was unclear if limitations existed for a particular downstream method or pipeline due to a lack of existing references, and thus we also designated many cells in the matrix as ‘unknown’ ( Fig. 2 ). These ‘unknowns’ are likely to have resulted from insufficient testing or knowledge in a particular area as opposed to the method being truly unacceptable; testing these approaches may present fruitful areas for future research. We then summed the number of ‘optimal’ and ‘acceptable’ cells to determine which pipelines best served a given set of downstream methodologies. In evaluating the various methodological approaches used in specimen collection, preservation, and archiving, we were able to identify several pipelines that maximize the number of downstream analyses that are possible ( Fig. 2 green cells; Fig. 3 fuchsia pink cells). Freezing or fixation methods dictate most methodological pipelines Instantaneous freezing or ‘rapid freezing’ in liquid nitrogen upon initial collection followed by ultra-cold storage ( e.g ., −80 °C) maximizes the number of possible downstream analyses (supports ~64% of 36 methods; Fig. 2 blue and green cells row 7). Analyses that could concurrently or sequentially be conducted after specimen rapid-freezing and cold storage fell primarily within the Omics and Physiology disciplines, while rapid freezing is inappropriate for most tissue Microscopy and Imaging because it alters tissue integrity (see Box 1 ). Freezing post-collection using −80 °C and more conventional −40 °C or −20 °C freezers were also suitable for several procedures within the Omics- and Physiology-based methods (supports ~83% of 18 Omics- and Physiology analytics), except for some RNA-based analytics, which always require immediate rapid freezing or preservation ( e.g ., in RNAlater, TRIzol). Within the Microscopy and Imaging discipline, preserving in paraformaldehyde followed by refrigeration allowed for the greatest number of downstream analytics (70% of all Microscopy and Imaging techniques, row 48), including some Omics methods. However, few if any of the Physiological methods could be conducted on samples stored in these aldehydes. Methodological considerations are needed to determine the suitability of some collecting, preserving and archiving sample pipelines A few analyses, including metabolomics, quantification and identification of mycosporine-like amino acids (MAAs), soluble lipid analysis, and histology techniques stood out as highly restrictive in their requirements for initial and secondary storage methods. Such methodological limitations could be due to stringency in storage requirements or, as suggested by the large number of unknowns in Fig. 2 , due to insufficient testing of potential alternative methods. Thus, we summed the number of ‘unknown cells’ to determine which methods had the most uncertainty in terms of how samples could be collected, preserved, and archived. Numerous methods had many ‘unknowns’ ( Fig. 2 , row 60) limiting our ability to find suitable additional pipelines to access outside of their standard procedures. For example, biomass quantification and Raman spectroscopy each respectively had 51% and 83% unknowns for the different possible methodological pipelines we tracked. Below we discuss considerations specific to each major discipline: Physiology, Omics, and Microscopy and Imaging, given these methodological differences. Furthermore, we add more details about standard operating procedures for each of the major downstream analyses within Fig. 2 and throughout the Supplemental Material . While not an exhaustive list, we aimed to give researchers enough information to consider how to collect, preserve, and archive their specimens for many potential applications. We also recognize that methods are continuously evolving with the advent of new technologies. It is likely that newer, better methods will eventually become available and, thus, future researchers should take steps to confirm that additional procedures have not become available following the publication of this work."
} | 9,417 |
23869111 | null | s2 | 5,618 | {
"abstract": "We synthesized and characterized a new family of di-block copolymers based on the amino acid sequences of "
} | 26 |
31246995 | PMC6597161 | pmc | 5,622 | {
"abstract": "Integration of trees in agroforestry systems can increase the system sustainability compared to monocultures. The resulting increase in system complexity is likely to affect soil-N cycling by altering soil microbial community structure and functions. Our study aimed to assess the abundance of genes encoding enzymes involved in soil-N cycling in paired monoculture and agroforestry cropland in a Phaeozem soil, and paired open grassland and agroforestry grassland in Histosol and Anthrosol soils. The soil fungi-to-bacteria ratio was greater in the tree row than in the crop or grass rows of the monoculture cropland and open grassland in all soil types, possibly due to increased input of tree residues and the absence of tillage in the Phaeozem (cropland) soil. In the Phaeozem (cropland) soil, gene abundances of amoA indicated a niche differentiation between archaeal and bacterial ammonia oxidizers that distinctly separated the influence of the tree row from the crop row and monoculture system. Abundances of nitrate ( napA and narG ), nitrite ( nirK and nirS ) and nitrous oxide reductase genes ( nosZ clade I) were largely influenced by soil type rather than management system. The soil types’ effects were associated with their differences in soil organic C, total N and pH. Our findings show that in temperate regions, conversion of monoculture cropland and open grassland to agroforestry systems can alter the abundance of soil bacteria and fungi and soil-N-cycling genes, particularly genes involved in ammonium oxidation.",
"conclusion": "Conclusion The trees in our agroforestry systems increased the fungi-to-bacteria ratio and altered the abundance of AOA and AOB amoA genes (particularly in the cropland on a Phaeozem soil), suggesting a niche differentiation. These may be due to the long-term absence of fertilization and cultivation and tree litter input within the tree row of agroforestry. Abundances of genes encoding for nitrate ( napA and narG ) and nitrite reductase ( nirK and nirS ) as well as nitrous oxide reductase gene nosZ clade I were less affected by the management system than by soil type. Overall, our results show that temperate agroforestry can alter the abundance of soil bacteria and fungi and soil-N-cycling genes compared to monoculture and open grassland. It should be noted that this study relies on a single measurement period, which does not allow temporal extrapolation of our findings. Future studies should thus focus on the temporal dynamics of the genes involved in soil-N cycling and its controlling factors in order to gain deeper understanding of the services provided by trees in agroforestry systems.",
"introduction": "Introduction Modern alley cropping systems are innovative agroforestry systems, where rows of fast growing trees are planted alternately with rows of annual crops [ 1 ]. Spatial arrangement of the tree and crop components allows ecological interactions between them, which can increase the overall efficiency of the resource use if the trees and crops are not competing for the same resources [ 2 , 3 ]. For example, due to deeper root growth, most tree species can access water and nutrients from deeper soil layers than the associated crops [ 3 , 4 ]. Deep-rooting trees are thus capable of providing a ‘safety-net’ for leached nutrients from the crop rhizosphere by deep-root nutrient uptake [ 5 – 8 ]. In contrast to monoculture agriculture, agroforestry can be more sustainable by conserving and improving soil quality through increasing soil organic C and nutrient availability [ 9 , 10 ] and its potential for C sequestration [ 10 ]. Increases in organic C stocks are mainly due to litter input and root decay of the tree rows [ 11 ]. Inclusion of N 2 -fixing trees further enhances physico-chemical as well as biological properties of soil and thus contributes to an improvement of soil fertility [ 12 – 14 ]. Most trees in temperate alley cropping systems are fast-growing trees like Salix and Populus species, suitable for biomass production. Soil microbial community structure and function are shaped by resource availability, which in turn is controlled, among other factors, by the quantity and quality of plant litter input as well as root exudation and decay [ 15 , 16 ]. Thus, it is likely that trees integrated in agroforestry systems affect soil microbial communities on both structural and functional levels. For example, a recent study in three poplar-based temperate agroforestry systems found greater soil fungal C-to-bacterial C ratios in the tree rows than the crop rows [ 17 ], probably due to increased input of lignin- and suberin-rich tree residues. Soil enzymatic activities and substrate utilization patterns of soil microbial communities also indicate that microbial communities in the tree and crop rows of agroforestry systems are functionally different [ 11 , 18 , 19 ]. Furthermore, other studies have reported beneficial effects of agroforestry on the abundance and diversity of the soil microbial community [ 12 , 20 , 21 ]. A recent study investigated for the first time the soil bacterial community structure in temperate agroforestry systems using next generation sequencing [ 22 ]. They found that the integration of trees increases the abundance and species richness of soil bacteria, which can have several causes. Plant diversity and productivity affect soil microbial processes such as N mineralization [ 23 ] and the nitrification-denitrification pathway [ 24 ]. For example, increased crop diversity by intercropping of maize ( Zea mays ) with peanut ( Arachis hypogaea ) resulted in an increase in nifH , ammonia-oxidizing archaea (AOA) and ammonia-oxidizing bacteria (AOB) amoA and narG abundance, whereas nirK , nirS and nosZ clade I remained unaffected [ 25 ]. Although soil microbial responses to agricultural management are highly complex [ 26 ] and not uniform among microbial groups [ 27 ], there is substantial evidence that agricultural intensification is one of the main drivers of observed losses in soil microbial diversity and connectivity [ 28 – 30 ]. The decreased complexity of intensive agricultural systems contributes to the loss of soil microbial biodiversity and is expected to negatively affect soil processes like nutrient cycling and retention [ 31 , 32 ]. The increase in system complexity associated with agroforestry is likely to alter soil microbial community structure and function, and thus soil-N-cycling. Several studies have shown that temperate agroforestry systems affect N cycling and enzymatic activities in soils; however, associated changes in the abundance of microbial communities involved in soil-N cycling as well as soil fungi and bacteria remain to be elucidated. The main aim of this study was to test whether the abundance of genes encoding enzymes involved in soil-N cycling differ under temperate monoculture and agroforestry cropland and open grassland and agroforestry grassland. A further aim was to assess differences in the abundance of bacterial and fungal populations by quantifying bacterial 16S rRNA and fungal 18S rRNA genes. We hypothesized that the tree row in the agroforestry systems increases the above- and belowground litter input and nutrient cycling, which in turn i) changes the abundance of functional genes involved in nitrification and denitrification, and ii) increases relative fungal abundance. In contrast to prior studies analyzing substrate utilization patterns and enzymatic activities of soils under laboratory conditions, this is, to our knowledge, the first field-based study investigating differences in the abundance of genes involved in soil-N cycling between temperate monoculture and agroforestry cropland and open grassland and agroforestry grassland.",
"discussion": "Discussion The increase in fungi-to-bacteria ratio in the tree row in all soil types ( Fig 1 ) suggests that the tree component of agroforestry increased the relative fungal abundance. This finding concurred with the findings of Beuschel et al . [ 17 ] who reported larger fungal C-to-bacterial C ratios in the tree rows than the crop rows of poplar-based temperate agroforestry systems. As tree litter is more recalcitrant than crop or grass residues [ 54 ], increased tree litter likely accounts for the higher relative abundance of fungi in the agroforestry tree row. Moreover, tillage practiced in cropland reduces the growth of fungal hyphae and disrupts mycorrhizal networks [ 55 – 57 ], which may have contributed to the reduction of fungal abundance in the arable land ( Fig 1 ). Other factors, such as changes in microclimate and theabsence of fertilization within the tree row, may also have contributed to the observed pattern. The ratio of fungal 18S-to-bacterial 16S rRNA gene abundances does not equal to a biomass ratio, as the copy number of rRNA genes is taxa-specific [ 58 ]. Such ratio, however, allows to detect shifts in the composition of microbial communities across soils [ 59 ]. The larger AOA than AOB amoA gene abundances ( Fig 2 ) suggests that archaea were the predominant prokaryotes involved in ammonia oxidation at our study sites. Significant contribution of archaea to nitrification has been observed in previous studies (e.g. [ 60 – 63 ]). In contrast, recent studies demonstrated that although AOA amoA genes may numerically dominate over AOB amoA , enzymes encoded by AOB amoA genes control nitrification [ 64 – 67 ]. In the cropland on Phaeozem soil, the opposing trends of AOA and AOB amoA gene abundance with increasing distance from the tree row in the agroforestry and the large AOB in the monoculture system suggest a niche differentiation of AOA and AOB, e.g. due to their preference for an ammonium source [ 68 , 69 ]. We attributed this pattern of AOA and AOB niche differentiation to the influence of trees and the application of N fertilizer both in the agroforestry crop row and in the monoculture system. AOA abundance was shown to increase by mineralized ammonium derived from soil organic matter or low ammonium concentrations in soils [ 70 , 71 ], whereas AOB is favoured by ammonium applied as mineral fertilizer [ 71 , 72 ]. Thus, long-term absence of mineral N fertilization and cultivation as well as tree litter input within the tree row of the agroforestry system may have contributed to the opposite trends of AOA and AOB amoA gene abundance. The influence of other soil factors on the abundance of AOA and AOB populations under field conditions is still not extensively investigated, and it is likely that a combination of soil factors also influence AOA and AOB abundance as well as niche partitioning. Apart from the ammonium source, the positive correlations of AOA and AOB amoA gene abundance with WFPS and plant-available P, K and Mg ( Fig 4 ) across sites, although with weaker correlations for AOB ( S4 Table ), indicated that soil moisture and macronutrients may also influence the abundance of AOA and AOB amoA genes. Szukics et al . [ 73 ] claimed that the population adaptability of AOA is greater than of AOB, which would enable AOA to adapt rapidly to changing environmental conditions. We found that AOA amoA gene abundance was more responsive to soil moisture than AOB, based on the correlation coefficients ( S4 Table ). Plant-available P as well as K and Mg are rarely measured when quantifying genes involved in soil-N cycling. Previous studies showed that AOA are positively influenced by NaHCO 3 - and Bray-extractable available P, whereas the response of AOB to available P was opposite [ 74 – 76 ]. Our findings that both AOA and AOB amoA gene abundances increased with increasing plant-available P, exchangeable K and Mg indicated a link between ammonium cycling and these rock-derived nutrients, which was not driven by the general bacterial and fungal genes since their abundance were not correlated with these nutrients ( Fig 4 , S4 Table ). There are only a limited number of studies to which we can relate the relative abundance of nitrite oxidoreductase nxrB gene to amoA genes. A possible explanation for the higher ratio of amoA (AOA+AOB) to nxrB gene abundances in the cropland soil (Phaeozem) than those in the grassland soils (Histosol and Anthrosol) may be the selectivity of our nxrB primer pair (NxrB-1F/R) for Nitrobacter strains [ 39 , 77 ]. The functional group of nitrite-oxidizing bacteria (NOB) is composed of six genera, namely Nitrobacter , Nitrotoga , Nitrococcus , Nitrospina , Nitrospira and Nitrolancetus [ 78 ]. Attard et al . [ 79 ] demonstrated that changes in nitrite oxidation are contributed by the shifts between Nitrobacter -like and Nitrospira -like NOB rather than within Nitrobacter -like populations. Thus, greater ratios between amoA and nxrB gene abundances in the Phaeozem soil (cropland) may originate from non- Nitrobacter dominated NOB populations, which cannot be assessed by our primer pair. The increased napA gene abundances in the tree row of the Phaeozem soil (cropland) ( Fig 3A ) may indicate a more effective microbial removal of NO 3 - in the tree row than the monoculture system. Root exudation of easily available C by trees is likely to facilitate microbial NO 3 - removal through denitrification, dissimilatory NO 3 - reduction and immobilization [ 80 – 82 ]. Apart from NO 3 - uptake by tree roots, microbial NO 3 - removal promoted by tree root exudates may additionally contribute to reduced NO 3 - leaching in agroforestry systems. Overall, genes involved in denitrification were rather affected by soil type than management system ( Fig 3A , Fig 3B , S4 Fig ). The stark differences in napA , narG , nirK and nosZ clade I gene abundances among soil types across management systems ( Fig 3A , Fig 3B , S4A Fig , S4C Fig ), i.e. Histosol (grassland) and Anthrosol (grassland) > Phaeozem (cropland), were mainly attributed to higher soil organic C and total N as well as lower soil pH of the grassland than the cropland ( Fig 4 , S1 Table ). Similar correlations were found for nirS gene, although with lower correlation coefficients ( S4 Table ), which agree with previous findings [ 83 , 84 ]. Studies reporting the relationships between the abundance of denitrifying genes and soil organic C and total N are contradictory, which may be attributed to different soil types and soil management practices [ 83 , 85 – 88 ]. Nevertheless, consistent with our findings, positive correlations between the abundances of napA , narG , nirK , nirS and nosZ clade I genes and soil organic C and total N have been observed in different ecosystems [ 85 , 86 , 88 , 89 ]. In addition, there are conflicting findings on the relationship between denitrification genes and soil pH, which relate to the range of pH covered in the studies. For example, increasing soil pH in acidic spruce forests exhibits a positive effect on nirK and a negative effect on nirS gene abundance [ 83 ], whereas the opposite was reported for grassland soils with a pH range from 6.4–7.1 [ 87 ]. Other studies even suggest that the abundance of both nir genes increases with increasing soil pH [ 84 , 90 ]. As the number of studies focusing on denitrifying genes and soil pH is limited and some studies cover rather narrow pH ranges, the effect of soil pH on the abundance of denitrification genes under field conditions is still not well understood. Recent phylogenetic studies identified a previously overlooked clade of microorganisms harbouring nosZ [ 45 , 91 ]. This clade ( nosZ clade II) has attracted great attention as its abundance and phylogenetic diversity has shown to be critical to the reduction of N 2 O to N 2 in soils [ 92 – 94 ]. The nosZ clade II-type microorganisms cover a broad range of bacterial and archaeal phyla, whereas clade I-type microorganisms have been shown to consist exclusively of members of α-, β- and γ-proteobacteria [ 45 ]. The higher nosZ clade II-to- nosZ clade I ratio in the Phaeozem soil (cropland), signifying higher potential for N 2 O consumption, compared to the Histosol and Anthrosol soils (grassland) ( Fig 3D ) may suggest that this cropland soil may have a greater potential to consume atmospheric N 2 O to N 2 relative to its potential for complete NO 3 — to-N 2 reduction within the soil."
} | 4,075 |
32580337 | PMC7357121 | pmc | 5,623 | {
"abstract": "Bacteria use quorum sensing signaling for cell-to-cell communication, which is also important for their interactions with plant hosts. Quorum sensing via N -acyl-homoserine lactones (AHLs) is important for successful symbioses between legumes and nitrogen-fixing rhizobia. Previous studies have shown that plant hosts can recognize and respond to AHLs. Here, we tested whether the response of the model legume Medicago truncatula to AHLs from its symbiont and other bacteria could be modulated by the abundance and composition of plant-associated microbial communities. Temporary antibiotic treatment of the seeds removed the majority of bacterial taxa associated with M. truncatula roots and significantly altered the effect of AHLs on nodule numbers, but lateral root density, biomass, and root length responses were much less affected. The AHL 3-oxo-C 14 -HSL (homoserine lactone) specifically increased nodule numbers but only after the treatment of seeds with antibiotics. This increase was associated with increased expression of the early nodulation genes RIP1 and ENOD11 at 24 h after infection. A 454 pyrosequencing analysis of the plant-associated bacteria showed that antibiotic treatment had the biggest effect on bacterial community composition. However, we also found distinct effects of 3-oxo-C 14 -HSL on the abundance of specific bacterial taxa. Our results revealed a complex interaction between plants and their associated microbiome that could modify plant responses to AHLs.",
"conclusion": "5. Conclusions Our results showed that plant phenotypic responses to bacterial AHLs could be modified by the abundance and composition of plant-associated bacterial communities. In addition, the modification of the AHL responses was specific for the AHL-tested and the phenotype studied. This indicated that studies examining the responses of plants to AHLs need to carefully consider the presence of plant-associated microbes that may directly or indirectly alter plant responses to AHLs. Interestingly, exposure of plants to AHLs also altered the relative abundance of specific bacterial taxa associated with the plant. Future studies could be aimed at determining the molecular mechanisms by which AHLs, plants, and associated microbes interact, for example, through molecular mimicry, quorum quenching, or indirect changes to plant metabolism and defense responses.",
"introduction": "1. Introduction Plant-associated bacteria can have beneficial effects on plant performance, for example, by altering plant growth and development or by inducing tolerance to diverse biotic and abiotic stresses [ 1 , 2 , 3 ]. Bacterial and fungal species can colonize all parts of the plant, including surfaces and the apoplast of roots, shoots, and seeds. Some bacteria also infect plants intracellularly, most prominently nitrogen-fixing bacteria that invade the inside of root hairs and cortical cells, although they remain outside the plant plasma membrane [ 4 ]. Plants can be colonized by various bacteria during different developmental stages and can also inherit some of their bacterial consortia via seeds in a process called vertical transmission. The seed-borne bacterial community structure is able to disperse systemically throughout the plant, colonizing roots and nodules [ 5 , 6 ]. Many plant-associated bacteria use cell-to-cell signaling via quorum sensing signals to coordinate behaviors that support successful colonization or invasion of plant hosts [ 7 , 8 ]. Gram-negative bacteria typically use acyl-homoserine lactone (AHL) signals, which control behaviors important for plant-microbe interactions, such as biofilm formation, bacterial motility, expression of pathogenicity genes, plasmid transfer, production of antibiotics, expression of genes required for successful symbiotic nitrogen fixation, and many others, in a density-dependent manner [ 9 ]. A possible advantage for controlling these behaviors in a density-dependent manner is that bacterial populations can build up to effective numbers before triggering plant defense responses [ 10 , 11 , 12 ]. Plants can detect the presence of AHLs and respond in a number of specific ways, even to very low (pM) concentrations of AHLs [ 13 ]. It has been shown that exposure of plants to AHLs can elicit a wide range of molecular and physiological responses in plants that are relevant for the outcome of plant–microbe interactions. One of the responses of plants to AHLs includes the production of quorum sensing mimic compounds, which can interfere with bacterial quorum sensing [ 13 , 14 , 15 ]. Other responses include changes in defense and developmental pathways in plants. AHLs can interfere with plant growth and development by altering plant hormone and signaling pathways, e.g., [ 16 , 17 , 18 , 19 , 20 , 21 ]. In addition, AHLs mediate plant defense responses and can often confer enhanced tolerance to pathogens, e.g., [ 22 , 23 , 24 , 25 , 26 , 27 ]. In legumes, which are colonized by nitrogen-fixing rhizobia, exposure to AHLs can influence the outcome of the symbiosis. During the legume-rhizobia symbiosis, rhizobia are initially present in the soil and are attracted to flavonoid exudates of host roots [ 28 ]. Flavonoids activate the synthesis of rhizobial nodulation (Nod) factors, which are perceived by the plant host and trigger a signaling cascade necessary for the invasion of rhizobia into the root and the initiation of nodule development [ 29 , 30 , 31 ]. In addition to Nod factors, exopolysaccharides are necessary for the successful invasion of rhizobia into root hairs and later inside nodules [ 32 ]. AHLs synthesized by rhizobia can influence the outcome of symbiosis by regulating bacterial attachment to root surfaces, host invasion via exopolysaccharide synthesis, plasmid transfer, movement, and expression of nitrogen-fixation genes [ 33 , 34 , 35 , 36 , 37 , 38 , 39 , 40 ]. On the plant side, the perception of rhizobial AHLs by legume hosts is also likely to be important for fine-tuning the symbiosis. Exposure of legumes to AHLs improved nodulation in the model legume Medicago truncatula , but this was observed after exposure to the symbiont-produced AHL, 3-oxo-C 14 -HSL (3-oxo-C 14 -homoserine lactone) in one study [ 41 ], while a smaller increase in nodule numbers was observed in a different study, which also showed that specific AHLs, including 3-oxo-C 14 -HSL, accelerated the rate of nodulation in M. truncatula [ 42 ]. The mechanism of the nodulation response remains unknown. Because legume tissues are concurrently colonized by rhizobia, as well as a multitude of other microbes [ 43 , 44 ], it is possible that quorum sensing signals from these different species cause a range of distinct and overlapping responses in the host plant. It is also likely that plant hosts perceive signals from multiple bacteria at once, and this could influence the overall responses to their specific symbionts. Here, we tested whether the presence of naturally-occurring bacteria associated with the plant host would alter the response of M. truncatula to AHLs and how this influenced the plant symbiosis with its specific symbiont, Sinorhizobium meliloti . We decided to test this by comparing responses to AHLs in plants from surface-sterilized seeds in comparison with seeds receiving an additional antibiotic treatment that was able to remove the majority of the bacterial taxa associated with the plant. This was done to make sure that all the bacteria that were present were naturally associated with the plant and well able to colonize the plant. The seeds used in our study originated from field-grown M. truncatula plants that we observed to harbor a variety of plant-associated bacteria.",
"discussion": "3. Discussion The aim of our study was to explore the possibility that the presence and composition of plant-associated bacteria could affect plant responses to AHLs. Many previous studies have demonstrated that exposure of plants to bacterial AHLs affects diverse phenotypes, including plant growth, development, and immune responses [ 13 , 16 , 19 , 20 , 22 , 23 , 25 , 26 , 27 , 48 ]. These studies typically grow plants under semi-sterile conditions with surface-sterilized seeds to minimize the influence of AHLs from contaminating bacteria. We showed here that 1) even roots growing from surface-sterilized seeds of field-grown M. truncatula that did not show any obvious signs of contamination under plate-grown conditions contained multiple bacterial species, and 2) that the presence of these plant-associated bacteria could modulate the response of the plant to AHLs. We further showed that the interaction between the AHL effect and the bacterial community was dependent on the AHL structure and the phenotype investigated. In our experiments, we used an antibiotic treatment that was applied for 6 h immediately after surface-sterilization of the seeds with sodium hypochlorite. We did not find any systematic effects of the antibiotic treatment on plant phenotypes, and the antibiotic was removed by extensive washing; therefore, we believed that it is unlikely that the influence of the antibiotic was a direct effect on the plant, but that it acted by removing the majority of bacterial taxa present in the plants. Because the seeds we used were field-grown, it is likely that the origin of some of the bacteria is located either on or below the seed coat, but some of the bacteria might also be endophytic. We did not further investigate the location or origin of these bacteria, but simply concluded for the purpose of this study that they were associated with the plant. The bacteria that we found associated with M. truncatula seedlings belonged to genera that have been found associated with plants in other studies. The most abundant genera were Erwinia , a genus with many known plant pathogens [ 49 ], Enterobacter, some of which are plant growth-promoting endophytes [ 50 , 51 ], and Pantoea species, which include seed endophytes and plant growth-promoting species as well as pathogens [ 52 , 53 ]. Other genera represented were species of Rhizobium , Pseudomonas, and Serratia . Similarly, identification of Medicago sativa seed endophytes (from surface-sterilized seeds) also found members of Enterobacter , Pantoea , non-nodulating rhizobia, and Pseudomonas species, among others [ 54 ]. A recent microbiome study of M. truncatula found various species of Pseudomonas and Rhizobium species associated with M. truncatula roots, although these were from soil-grown plants [ 55 ]. While in that microbiome study, Proteobacteria represented more than 50% of root-associated taxa, we found Proteobacteria almost exclusively in our study, suggesting that the seed treatment could have selectively removed several taxa or that soil exposure increased members of the other taxa. Future experiments could be aimed at a more detailed metagenomics analysis of the bacterial species inhabiting the field-grown M. truncatula seeds. Treatment with the antibiotics reduced total numbers and diversity of bacteria drastically and shifted the relative abundance of bacteria strongly towards Pseudomonas species, with lower numbers of Rhizobium species, although none of these were known symbionts of M. truncatula . The predominance of Pseudomonas spp. in antibiotic-treated roots is likely a result of Pseudomonas sp . being able to develop antibiotic resistance to a broad range of antibiotics [ 56 ]. While we only used one broad-spectrum antibiotic in this study, it is likely that other methods of manipulating the M. truncatula microbiome could have yielded different results. The shift of bacterial community composition and abundance associated with M. truncatula clearly affected the responses of the roots to AHLs. This could occur via several mechanisms, including (1) background production of AHLs by the plant-associated bacteria that are perceived by the plant at the same time as the AHL used as the experimental treatment, (2) destruction of the AHL used as the treatment by plant-associated bacteria, or (3) modulation of plant responses by plant-associated bacteria independent of their AHL production. These three mechanisms are further discussed below. The predominant genera— Erwinia , Enterobacter , Pantoea , Pseudomonas, and Rhizobium— are all known to synthesize AHLs, although of different structures. Some of the AHLs synthesized by these species overlap the AHLs used here, for example, S. meliloti produces C 8 -HSL, 3-oxo-C 8 -HSL, C 14 -HSL, 3-oxo-C 14 -HSL, and C 18 -HSL [ 14 , 46 ], while Pantoea spp. [ 57 ] and Erwinia spp. [ 58 ] synthesize 3-oxo-C 8 -HSL and Pseudomonas spp. [ 7 ] and Enterobacter spp. [ 59 ] synthesize C 4 -HSL. However, the exact AHL structures of the specific species of most of the identified bacteria in our study are not known. Thus, the perception of external AHLs by the plant, if colonized by these bacteria, is likely a mixed response to the combination of AHLs the plant perceives. For example, 3-oxo-C 14 -HSL appears to enhance nodulation in the absence of substantial numbers of plant-associated bacteria and thus likely in the absence of substantial concentrations of other AHLs. However, other AHLs may have negative effects on nodule numbers. Thus, if the plant perceives a mixture of these AHLs, the net outcome might be unchanged numbers of nodules. This would depend, of course, on the location where these AHLs are made and perceived, the mechanism by which different AHLs are perceived, and the mechanisms by which each of them acts. Currently, no plant receptors that directly bind different AHLs have been identified. Another important consideration is whether the bacteria identified to colonize M. truncatula roots would actually occur at densities at which effective concentrations of AHLs are produced. We estimated the number of culturable bacteria recovered from the non-antibiotic-treated roots to be around 120,000 per root, approximately 4 cm long. This does not take into account that there would be unculturable bacteria and that different members of those microbial populations would likely to be dense in some locations and absent from other parts of the root. Estimations of quorum sizes for bacteria on dry leaf surfaces, where diffusion is limited, is small, with populations as low as 10 cells able to communicate via AHLs, while this number increases in wetter environments with higher diffusion rates [ 60 ]. The reporter used to detect AHLs start to respond to 3-oxo-C 6 -HSL at 100 nM [ 60 ]. Similar results have been reported by microscopic analysis of quorum sensing calling distances on root surfaces of sand-grown tomato plants, uncovering that AHL-medicated quorum sensing occurs even at low bacterial densities and can be exerted over distances far beyond the length of a bacterial cell [ 61 ]. This has been done using an AHL reporter sensitive to concentrations above 20 nM of 3-oxo-C 12 -HSL. These studies suggest that even small clumps of cells on a plant surface could contain locally significant concentrations of AHLs. Whether these are sufficient to trigger plant responses would be interesting to investigate. Secondly, some of the identified plant-associated bacteria are able to destroy AHLs, often from other species, and this quorum quenching mechanism is likely a widespread mechanism for interfering with quorum sensing across bacterial species [ 62 , 63 ]. Quorum quenching activities have been characterized in rhizobia [ 63 ] and Pseudomonas [ 64 ] species, but could also be present in some of the other species identified here. Thus, the presence of plant-associated bacteria could alter the active concentrations of any applied AHLs, lessening or altering their effect on the plant. Thirdly, the presence of plant-associated bacteria could potentially induce a number of biochemical and physiological responses in the plant that indirectly interfere with the response to an applied AHL. Plant-associated bacteria are able to synthesize hormones, effectors, and signals that could interfere with plant responses to AHLs, in particular, those that alter development and defense [ 65 ]. We observed significant effects of the antibiotic treatment alone in some of the AHL treatments, for example, on plant biomass ( Figure 4 ). This might reflect the production of plant hormones by the plant-associated bacteria, although this would have to be investigated in detail by further experiments. If plant hormone responses or plant immune responses were altered by the ‘background’ presence of plant-associated bacteria, the AHLs that cause changes to the same pathways might lead to different overall responses by the plant. Some of the plant-associated microbes that were reduced by the antibiotic treatment included known plant pathogens, including Erwinia , Pantoea, and Serratia spp., although their direct effects on M. truncatula were not studied in detail. The focus of our study was mainly on nodulation, as rhizobia are well documented to require AHL signaling for effective symbiosis [ 35 ]. We confirmed that the application of 3-oxo-C 14 -HSL at 1 μM concentration significantly increased nodule numbers, similar to what previous studies found [ 41 , 42 ], but only in antibiotic-treated roots. While roots in our previous study [ 41 ] were treated with the same antibiotic, a separate study, which found a significant but relatively smaller increase in nodule number than shown here, did not treat seeds with antibiotics, and the source of seeds and growing conditions were different [ 42 ]. It would be interesting to compare the effects of AHLs on different batches of seeds of the same species originating from different sources with different inherent seed-associated bacteria side by side in the same experiment. None of the other AHLs had positive effects on nodule numbers in this study, but some had effects in another study [ 42 ]. This highlights the difficulty of comparison between studies, with some of the variability likely due to the presence of seed-borne bacteria and further variability likely due to growth conditions used. It is also likely that different strains of rhizobia could exert different effects. The S. meliloti 1021 strain used here lacks one of the AHL receptors, ExpR, which affects some of its AHL-dependent behaviors [ 66 ]. It would, therefore, be interesting to test the nodulation responses in strains encoding a functional copy of ExpR . It would, furthermore, be interesting to examine the effects of AHLs on nitrogen fixation, which we did not do here. At the time of harvesting at three weeks after inoculation, biomass changes accompanying the higher nodule numbers in 3-oxo-C 14 -HSL-treated roots were not evident, suggesting that any potential increases in nitrogen fixation had not occurred or not led to enhanced biomass (yet), but this would have to be quantified in future studies. We also showed that the effect of 3-oxo-C 14 -HSL was manifested in the increased expression of early nodulation genes within the first 24 h after inoculation with rhizobia. This suggested that this AHL, which is produced by its symbiont S. meliloti [ 14 ], was specifically perceived by the plant host and enabled different responses to rhizobial Nod factors. The early nodulation genes— ENOD11 , ERN1 , NIN1, and RIP1— which showed the highest expression in 3-oxo-C 14 -HSL-treated roots in the presence of the antibiotics, when most nodules were observed, are inducible by Nod factors in M. truncatula root hairs within 24 h and are known markers for successful Nod factor signaling [ 47 ]. However, only RIP1 and ENOD11 expression levels showed a similar significant interaction between AHL and antibiotic treatment, as was seen for nodule numbers ( Figure 6 ). RIP1 is a Rhizobium-induced peroxidase necessary for the formation of reactive oxygen species that are required for the infection of rhizobia [ 67 ], while ENOD11 is a repetitive proline-rich cell-wall protein that is an early marker for successful infection [ 68 ]. This result suggested that 3-oxo-C 14 -HSL acted at the earliest stages of nodulation. In the future, it will be interesting to investigate the mechanism of 3-oxo-C 14 -HSL perception and action during nodulation in M. truncatula in more detail, for example, through quantifying infection events and more comprehensive analysis of gene expression changes, including defense genes that could regulate early infection events. Interestingly, the same AHL also affects immune responses in non-legumes [ 26 , 27 ]. An additional interesting aspect to examine would be to test whether 3-oxo-C 14 -HSL may affect Ca 2+ -signaling in the root hair since Ca 2+ spiking is one of the earliest responses to Nod factors from rhizobia [ 69 ] and the short-chain AHL—C 4 -HSL—has been shown to trigger changes in Ca 2+ signaling in Arabidopsis [ 70 ]. Apart from the specific nodulation responses to AHLs, there were negligible effects on root length, lateral root numbers, and plant biomass. Other studies have found similar results at 1 μM concentrations of AHLs, while other concentrations can have strong effects on root elongation [ 16 , 18 , 22 , 41 , 42 , 71 ]. This underlines the importance of interpreting AHL responses in the context of their concentrations. Overall, our finding that some AHL responses were dependent on the presence of plant-associated bacteria clearly cautions studies on plant responses to AHLs because it is unlikely that results obtained under sterile or semi-sterile laboratory conditions can be translated to plants growing in the field, where plants are exposed to many different AHLs at once, as well as a myriad of other signals from plant-associated bacteria. It might also explain why different studies do not always find similar results or why different plant species show different responses to the same AHLs [ 72 ]. Finally, an unexpected result from our study was that the application of 3-oxo-C 14 -HSL to M. truncatula altered the composition of the plant-associated bacterial community. This was evident both in the presence and absence of antibiotics. While this effect was not large, several bacterial species were significantly affected in abundance by AHL treatment. This suggested a further complication of studies looking at AHL responses in plants (or other eukaryotes), as the effect of an applied AHL could include indirect effects of bacteria associated with the host. We currently do not understand the mechanism by which AHLs could shift bacterial community composition associated with the plant host. AHLs might alter the exudation of secondary metabolites that could affect bacterial community composition. For example, exposure of M. truncatula to AHLs has been shown to alter the expression of the flavonoid synthesis pathway [ 13 ] and the accumulation of flavonoid metabolites [ 41 ], and flavonoids can alter microbiome composition [ 73 ]. In addition, exposure of M. truncatula to AHLs has altered the exudation of metabolites that can interfere with quorum sensing between bacteria [ 13 ]. Thus, altered quorum sensing in the plant-associated microbiome could alter its community composition through alteration of bacterial movement, selective biofilm formation, competition for nutrients, or plasmid transfer, which are regulated by quorum sensing in rhizobia [ 35 ]. Altogether, a complex picture emerges that warrants further investigation between plant hosts, their associated microbiome, and the role of bacterial quorum sensing signals in shaping both plant phenotypes and the microbial community composition. Our study could be extended in the future in many directions. First, while the choice of antibiotics and the treatments applied were likely to influence the outcome of our study, our study demonstrated that, in principle, a shift in the plant-associated microbiome did make a difference in the outcome of plant AHL responses. Further studies could be aimed at varying the way the microbiome is altered. For example, rather than removing existing taxa, the addition of specific species of bacteria to the plant could be trialed. Additional experiments should also be aimed at gaining a comprehensive view of the actual AHL structures and concentrations present in and around roots over time and in different locations. AHL concentrations are likely to be very dynamic in time and often variable, depending on the location different bacteria are present in or around the root. In addition, the AHLs were mixed into the growth media that the plants were growing on and would likely have started breaking down during the experiment [ 74 ], although the pH of the medium was kept acidic. It is hard to predict what the actual active concentrations of AHLs were at different times and whether breakdown products might have altered the responses over time. Our study did not capture that complexity, and this could be explored in the future. The fact that the much shorter-term exposure to AHLs caused changes in nodulin gene expression in a similar pattern to the increased nodule numbers three weeks later suggested that the effect of 3-oxo-C 14 -HSL was quite specific."
} | 6,309 |
36516226 | PMC9797087 | pmc | 5,624 | {
"abstract": "Feedforward network models performing classification tasks rely on highly convergent output units that collect the information passed on by preceding layers. Although convergent output-unit like neurons may exist in some biological neural circuits, notably the cerebellar cortex, neocortical circuits do not exhibit any obvious candidates for this role; instead they are highly recurrent. We investigate whether a sparsely connected recurrent neural network (RNN) can perform classification in a distributed manner without ever bringing all of the relevant information to a single convergence site. Our model is based on a sparse RNN that performs classification dynamically. Specifically, the interconnections of the RNN are trained to resonantly amplify the magnitude of responses to some external inputs but not others. The amplified and non-amplified responses then form the basis for binary classification. Furthermore, the network acts as an evidence accumulator and maintains its decision even after the input is turned off. Despite highly sparse connectivity, learned recurrent connections allow input information to flow to every neuron of the RNN, providing the basis for distributed computation. In this arrangement, the minimum number of synapses per neuron required to reach maximum memory capacity scales only logarithmically with network size. The model is robust to various types of noise, works with different activation and loss functions and with both backpropagation- and Hebbian-based learning rules. The RNN can also be constructed with a split excitation-inhibition architecture with little reduction in performance.",
"introduction": "Introduction Binary classification is a basic task that involves dividing stimuli into two groups. Machine-learning solutions to this task typically use single- or multi-layer perceptrons [ 1 ] in which, almost invariably (but see [ 2 ]), the output that delivers the network’s decision comes from a unit that collects information from all of the units in the previous layer. Collecting all of the evidence in one place (i.e. in one unit) is an essential element in the design of these networks. In humans and other mammals, tasks like this are likely performed by neocortical circuits that have a recurrent rather than feedforward architecture, and where there are no obvious highly convergent ‘output’ neurons. Instead, all of the principal neurons are sparsely connected to a roughly equal degree. This raises the question of whether a network can classify effectively if the information needed for the classification remains dispersed across the network rather than being concentrated at a single site. Here we explore how and how well recurrent networks with sparse connections and no convergent output unit can perform binary classification. We study sparse RNNs that reach decisions dynamically. Despite their sparse connectivity, these networks are able to compute distributively by propagating information across their units. To add biological realism, we also constrain our sparse RNN to have a split excitation-inhibition architecture. The model maintains high performance despite this constraint. To investigate capacity and accuracy, networks were trained by back-propagation through time (BPTT). With extensive training, these models can categorize up to two input patterns per plastic synapse, matching the proven limit of the perceptron [ 3 ]. The model is robust to different types of noise, training methods and activation functions. The number of recurrent connections per neuron needed to reach high performance scales only logarithmically with network size. To investigate biologically plausible learning, we also constructed networks using both one-shot and iterative Hebbian plasticity. Although performance is significantly reduced compared to BPTT, capacity is still proportional to the number of plastic synapses in the RNN.",
"discussion": "Discussion We presented a biologically plausible neural network architecture that solves the binary classification task with high capacity. This architecture is based on a sparse RNN that solves the task dynamically. Our main purpose was to show that categorization can be achieved in a truly distributed way without the convergence of information onto any single locus. Sparse RNNs can categorize roughly two patterns per plastic synapse, matching classic perceptron performance [ 3 ]. These networks are robust to various types of noise and across training methods and activation functions. Our approach supports separate populations of excitatory and inhibitory neurons, and the resulting E/I networks perform well. The performance of sparse RNNs for categorization, scaling of the number of patterns proportional to the number of synapses, is in keeping with results from other network studies. In sparse Hopfield-type models, the number of stored bits scales with the number of synapses [ 8 ]. In Hopfield-style models of recognition memory, the number of patterns than can be identified as familiar or novel scales with the number of synapses [ 9 ], a result that also holds for feedforward networks [ 10 ] and for various plasticity rules [ 10 , 11 ]. Categorization by sparse networks has been considered previously by Kushnir and Fusi [ 2 ], who used a committee-machine-like readout on a recurrent network with a fixed number of recurrent connections per unit. Fixed here means a number of connections per neuron that was independent of the number of neurons ( N ), as opposed to the case we studied with sparse connections proportional to N per neuron (or log N in Minimum sparseness). In addition, in their study [ 2 ], recurrent connections were not learned. Nevertheless, Kushnir and Fusi showed that the recurrent connections play a crucial role in maintaining classification accuracy with sparse readouts. They proved that their model can classify a number of inputs proportional to the number of plastic synapses, as reported by other studies [ 3 , 12 , 13 ] and in ours. Because of the fixed number of connections per neuron, in the regime of large numbers of neurons, their RNN eventually becomes disconnected. However, the largest connected subnetwork scales linearly with N as long as f > 1 N [ 14 , 15 ], a reasonable assumption which requires that, on average, every neuron have at least 1 synapse and which was used in many network dilution studies [ 2 , 16 , 17 ]. Though the Kushnir and Fusi results hold in light of the largest connected component, some neurons in the RNN cannot contribute to the computation. These disconnected neurons consume resources but do not help solving the task. Our results, particularly from Minimum sparseness, suggest that it suffices to have ∼ log N connections per neuron to have all neurons contribute distributively to solve the task. This is only a small price to pay, compared to the case of fixed number of connections per neuron, for having the RNN be fully efficient, since log(10 11 ) < 26. Our results suggest that it is possible to generate decisions in a dynamic and distributed manner in RNNs. This is almost certainly closer than perceptron models to how the bulk of decisions made by biological networks are computed. We suggest the following model: motor or premotor circuits are held in a state of readiness during a go/no-go task, but are not activated until the decision is made. Relevant information is conveyed to the neurons in this circuit, much like the patterns are conveyed to the RNNs we studied. If these inputs are appropriate for a no-go decision, the motor/premotor circuit may be perturbed, but it does not make a transition to a fully activated state. This is analogous to the blue curves in Fig 1C . If, on the other hand, the evidence favors action, the motor/premotor circuit reacts more strongly, analogous to the red curves in Fig 1C , and the motor action is initiated."
} | 1,967 |
34949814 | null | s2 | 5,625 | {
"abstract": "How and when the microbiome modulates host adaptation remains an evolutionary puzzle, despite evidence that the extended genetic repertoire of the microbiome can shape host phenotypes and fitness. One complicating factor is that the microbiome is often transmitted imperfectly across host generations, leading to questions about the degree to which the microbiome contributes to host adaptation. Here, using an evolutionary model, we demonstrate that decreasing vertical transmission fidelity can increase microbiome variation, and thus phenotypic variation, across hosts. When the most beneficial microbial genotypes change unpredictably from one generation to the next (for example, in variable environments), hosts can maximize fitness by increasing the microbiome variation among offspring, as this improves the chance of there being an offspring with the right microbial combination for the next generation. Imperfect vertical transmission can therefore be adaptive in varying environments. We characterize how selection on vertical transmission is shaped by environmental conditions, microbiome changes during host development and the contribution of other factors to trait variation. We illustrate how environmentally dependent microbial effects can favour intermediate transmission and set our results in the context of examples from natural systems. We also suggest research avenues to empirically test our predictions. Our model provides a basis to understand the evolutionary pathways that potentially led to the wide diversity of microbe transmission patterns found in nature."
} | 397 |
37847029 | PMC10715075 | pmc | 5,627 | {
"abstract": "ABSTRACT Understanding the changes in bacterial community structure in different microenvironments of Camellia oleifera is essential to better explore the benign interaction between beneficial microorganisms and plants. Using Camellia oleifera trees, a Chinese wooden oil plant as a model ecosystem, we characterized the archaeal and bacterial microbiome across five different tissue-level niches using 16S rRNA gene analyses. Our research indicates that the diversity of Camellia oleifera endophytic bacterial communities is highly dependent on the plant compartment. The species replacement process (69.90%) is the dominant factor in the differences in bacterial community structure. The dominant bacteria phyla ( Proteobacteria , Acidobacteria , Actinobacteria , Bacteroidetes , Firmicutes , Chloroflexi , and Verrucomicrobia ) of Camellia oleifera show a significant plant compartment (roots, stems, leaves, fruits) enrichment effects. A variety of bacteria ( Hymenobacter , Allorhizobium-Neorhizobium-Pararhizobium-Rhizobium , Mesorhizobium , Bradyrhizobium , Bacillus , Ochrobactrum , Pantoea , Pseudomonas , etc.) with nitrogen-fixed potentials are enriched in Camellia oleifera tissue. In addition, the hub bacterial groups of Camellia oleifera are Nitrospira , Haemophilus , Staphylococcus , Ruminiclostridium , and Ochrobactrum . They are widespread colonization in various tissues with a low relative abundance and may play an important role in the nitrogen cycle, host life promotion, and plant defense. This study provides a holistic understanding of the endosphere bacterial community structure, which is one of the most complete ecological niche-level analyses of Camellia oleifera . These results provide a scientific theoretical basis for an in-depth discussion of plant-endosphere microbial interaction and better exploration of benign interaction of beneficial microorganisms and plants. IMPORTANCE Microorganisms inhabited various tissues of plants and play a key role in promoting plant growth, nutritional absorption, and resistance. Our research indicates that the diversity of Camellia oleifera endophytic bacterial communities is highly dependent on the plant compartment. Proteobacteria , Acidobacteria , Actinobacteria , Bacteroidetes , Firmicutes , Chloroflexi , and Verrucomicrobia are dominant bacteria phyla. The tissues of Camellia oleifera contain various bacteria with nitrogen fixation potential, host life promotion, and plant defense. This study provides a scientific theoretical basis for an in-depth discussion of plant-endosphere microbial interaction and better exploration of benign interaction of beneficial microorganisms and plants.",
"conclusion": "Conclusion In general, in this study, highly diversified and structured niche-specific groups were observed in different sample types of Camellia oleifera . The diversity of endophytic bacterial communities in Camellia oleifera is highly dependent on plant compartments, and each compartment represents a unique niche of the bacterial community. Our study not only confirms the niche differentiation of the microbes at the soil-root interface but also demonstrates the fine-tuning and adaptation of the endophytic microbiota in the stem, leaf, and fruit compartments. In addition, we have identified the hub bacterial microbes of Camellia oleifera . This study provides a relevant model for the systematic study of the changes in microbial community in the organizational level niche of Camellia oleifera plants. These results fill the knowledge gap of the endophytic bacterial community of Camellia oleifera and provide a theoretical basis for the subsequent exploration of microbial functions and research on bio-fertilizers.",
"introduction": "INTRODUCTION \n Camellia oleifera is a unique oil plant with high nutritional value in China and one of the world’s four major wooden oil plants. Its uses have high economic flexibility, such as the production of Camellia seed oil, cosmetics, chemicals, brachycal, gums, activated carbon, and the cultivation of edible fungi ( 1 ). In addition, Camellia oleifera are basically grown in mountains and hill regions that are not suitable for grain production, thereby avoiding competing for land resources with grain. However, the long-term predatory and extensive management have led to a significant decrease in soil nutrients and in a low-yield state. Plants are colonized by complicated multi-kingdom microbial communities (bacteria, fungi, native creatures, etc.) ( 2 , 3 ). Each part of the plant is a unique ecosystem of microorganisms. Compared with other plant tissues (including roots, stems, leaves, flowers, and seeds), it has a unique microbial assembly ( 4 \n – \n 7 ). The results of Xiong et al. ( 8 ) pointed out that some members of Burkholderiaceae , Microbacteriaceae , Streptomycetaceae , and Rhizobiaceae were enriched on the surface of phylloplane and rhizoplane at the maize seedling stage. Lei et al. ( 9 ) show that the bacterial community structure between different parts of Macleaya cordata has a significant change, among which Sphingomonas and Methylobacterium dominate the fruits and leaves, respectively. In addition, a significant plant compartment effect was observed in the microbiome of tomato, willow, poplar, agave, and cactus ( 10 \n – \n 13 ). In addition, only adaptive or non-picky bacterial populations can survive or flourish within plant tissues due to filtration and selection, which leads to a low degree of microbial diversity ( 14 ). These microorganisms inhabiting various parts of plants can play a key role in promoting plant growth, nutrient absorption, and resistance to biotic or abiotic stresses (diseases, insect pests, high temperature, saline-alkali soil or drought, etc.) ( 15 \n – \n 17 ). One of the strategies developed by non-legume plants to increase the supply of nitrogen is to form a nutritional alliance with endophyte nitrogen fixation bacteria. So far, a large number of nitrogen-fixed nutrient bacteria ( Azospirrillum brasiliense , Gluconacetobacter diazotrophicus , Herbaspirillum seropedicae , Azoarcus, etc.) have been identified as epibiotic or endophyte bacteria, combined with cereal and grass ( 18 ). Recently, Deynze et al. ( 19 ) reported that they identified landrace maize that could benefit from the atmospheric nitrogen fixed by the related endophytic nitrogen-fixing bacteria ( Azospirillum brasilense , Herbaspirillum seropodicae , and Burkholderia unamae ). The aerial root mucus produced by this special maize is proved to be the environmental niche of the nif gene pool, which can provide up to 85% of assimilated nitrogen. Under drought stress, Bacillus sp . (12D6) and Enterobacter sp . (16i) will rapidly colonize in the rhizosphere of maize seedlings, stimulate the secretion of auxins and gibberellins, significantly increase the root length, root surface area, and number of root tips of maize seedlings, to obtain more water and alleviate drought stress ( 20 ). Therefore, analyzing the characteristics of bacterial communities in different ecological niches of healthy hosts may help to improve soil quality, crop growth, and stress resistance, thus reducing the dependence on fertilizers in production activities. It is of great importance for promoting the sustainable development of Camellia oleifera production and understanding the contribution of Camellia oleifera to ecosystem services. Most of the previous research is concentrated in the ecological position of the bacterial community of the soil-root interface ( 21 \n – \n 23 ). Contrary to the knowledge of bacterial microbial community differentiation related to the rhizosphere ( 24 \n – \n 26 ), there are few reports on the structural composition of bacterial communities in different tissues of Camellia oleifera , especially the relationship between underground and aboveground communities. Here, we employed 16S rRNA sequencing to evaluate the niche differentiation of bacterial communities related to bulk soil and roots, stems, leaves, and fruits of Camellia oleifera . The analysis of niche differentiation (bulk soil, root, stem, leaf, and fruit) of endophytic bacteria community in Camellia oleifera can provide a scientific basis for further exploring the mechanism of plant endophytic microbial interaction and tapping the biological potential of benign interaction between plant growth and development and beneficial microorganisms.",
"discussion": "DISCUSSION The diversity of endophytic bacteria community in Camellia oleifera is highly dependent on compartment The results show that the richness and diversity of the Camellia oleifera bacterial community gradually decreased from the bulk soil to the endophytic compartment ( Fig. 1 ; Table S1). This result is consistent with the general view of endophyte colonization. This is because the rhizosphere soil-root interface acts as a selective barrier, and only a limited number of bacteria can adapt to the endogenous lifestyle and dominate the endogenous combination, thus forming a unique, highly rich, and diverse microbial community ( 37 ). Some studies have found that niche differentiation, especially between soil and plant tissue, can lead to changes in bacterial community structure ( 38 , 39 ). In this study, there are significant differences in the bacterial community structure in different niches of Camellia oleifera , especially between bulk soil and plant tissues ( Fig. 2 ). The species replacement process (69.90% contribution rate) is the dominant factor causing this difference ( Fig. 3 ; Table 1 ). This result is consistent with the view that each plant compartment is a unique niche of microbial entities and has a unique microbial combination compared with other plant tissues (including roots, stems, leaves, flowers, and seeds) ( 8 , 40 , 41 ). Each compartment represents the unique niche of Camellia oleifera bacteria The plant endophytic environment is considered to be a restricted niche. A variety of biological factors (infiltration pathway, plant genotype, strain type, etc.) and abiotic factors (ultraviolet radiation, temperature, dryness, etc.) limit the colonization of endophytic bacteria ( 7 , 42 , 43 ). In this study, the number of OTUs in fruit, leaves, root, and stem was reduced by an average of 33.33% compared to the bulk soil, indicating that it is the main site of microbial colonization and activity in the soil, which harbors a rich and diverse bacterial population as compared to other plant ecological niches ( 44 ). In addition, 1,269, 833, 860, and 1,169 differential OTUs were obtained from fruit, leaves, root, and stem, respectively, compared to bulk soil samples ( Fig. 5 ). Contrary to stem (up 365 vs down 804), fruit showed obvious enrichment effect (up 795 vs down 474). This result is consistent with the conclusion that the diversity of endophytic bacterial communities in Camellia oleifera is highly dependent on compartment, indicating that only adaptive and non-selective bacterial populations can survive and/or proliferate in the tissues of Camellia oleifera ( 45 ). Finally, we found that most OTUs are not shared, especially those found in bulk samples ( Fig. 5 ). These findings are consistent with studies showing plant niche differentiation ( 11 , 46 ). Of these, four OTUs (such as Ruminococcaceae , Chitinophagaceae , and Diplorickettsiaceae ) are unique to the root tissue and one ( Methylopila ) to the leaf tissue. In all, 205 OTUs are unique to the soil of Camellia oleifera forest, with the majority of them belonging to functional groups that participate in nutrient transformation ( Candidatus_Xiphinematobacte , HSB_OF53-F07 , and FCPS473 ), promote plant growth ( Mucilaginibacter ), and decompose organic matter ( Candidatus_Udaeobacter , Chthoniobacter , and Pedosphaera ) ( 47 \n – \n 52 ). At the species level, Proteobacteria , Actinobacteria , and Bacteroidetes are the common dominant bacteria of Camellia oleifera , watermelon, Arabidopsis, rice, Antarctic vascular plants, Dendrobium, and other plants ( 2 , 45 , 53 \n – \n 55 ). This indicates that the composition of endophytic bacterial communities in plants may be similar at the phyla level. It has been found that Proteobacteria plays a dominant role in endoderm ( 56 , 57 ), leaf ( 45 ), and stem ( 11 ). In this study, Proteobacteria , Bacteroidetes, and Firmicutes were significantly enriched in the fruit, leaf, root, and stem of Camellia oleifera (Table S2), and Proteobacteria was the dominant phylum ( Fig. 6 ; Table S2). Sphingomons are a common bacteria associated with each other in different plant tissues ( 58 ), which is significantly enriched in Camellia oleifera fruits (21.59%). Sphingomons is not only an important regulator of Arabidopsis thaliana leaf microbiota ( 59 ) but also the most characteristic microorganism in rice seed disease resistance phenotype. It plays the role of “extending the immune system” in the “disease triangle,” and can be passed from generation to generation in the microbiome of healthy plant seeds ( 60 ). However, the significance of sphingomons expression in Camellia oleifera fruits needs further study. Massilia can colonize in plant tissues such as Alopecurus aequalis Sobol and ryegrass, and degrade polycyclic aromatic hydrocarbons (PAH) compounds ( 61 ). In this study, Massilia was significantly enriched in Camellia oleifera fruits (9.12%) and stems (3.57%), indicating that Camellia oleifera may be able to eliminate environmental pollution by degrading PAH through Massilia , providing a new perspective for plants to control PAH absorption through endophytic bacteria and reflect the ecological function of Camellia oleifera forest. In addition, we also found a variety of bacteria with nitrogen fixation potential in Camellia oleifera fruits ( Hymenobacter , Allorhizobium-Neorhizobium-Pararhizobium-Rhizobium ), leaves ( Bacillus , Mesorhizobium , Ochrobacterium , Pantooa , Pseudomonas , Stenotrophomona , etc.), roots ( Bradyrhizobium ), and stems ( Devosia ) (Fig. S3). The above results indicate that the specific bacteria in the endophytic bacteria of Camellia oleifera may play an important role in eliminating environmental pollution and obtaining nutrition. Potential ecological functions of hub microbes in Camellia oleifera \n Hub microbes can be important nodes in the community, rich taxonomic groups in the network structure of the microbial community, or microbes significantly related to ecological functions ( 62 ). In our study, five cluster modules were identified in the Camellia oleifera bacterial co-occurrence network ( Fig. 8 ). According to the analysis of Cluster 1 with the highest score, the hub bacteria of Camellia oleifera are Nitrospira (0.07%), Haemophilus (0.05%), Staphylococcus (0.17%), Ruminiclostridium (0.04%), and Ochrobactrum (0.14%) ( Table 2 ). Nitrospira i s one of the most widely distributed and diverse nitrites oxidizing bacteria, and also a key nitrifying bacteria in natural ecosystems ( 63 , 64 ). Ochrobacterium and Staphylococcus can promote the growth of host plants by generating indole-3-acetic acid (IAA) or cooperating with silicate ( 65 , 66 ). Ruminiclostridium , as a cellulose-degrading bacterium, its cellulose-degrading products can provide a carbon source for the growth of other microorganisms on the one hand ( 67 ), and may play an indirect role in activating plant defense on the other hand ( 68 ). Haemophilus is usually related to human pathogens, but it has been found that the genus inhabits plants ( 69 ). The hub microbes identified by us are relatively low in abundance, but they are widely colonized in various tissues (fruit, leaf, stem, root, and bulk soil) of Camellia oleifera , and may play an important role in nitrogen cycling, host growth promotion, and plant defense. Conclusion In general, in this study, highly diversified and structured niche-specific groups were observed in different sample types of Camellia oleifera . The diversity of endophytic bacterial communities in Camellia oleifera is highly dependent on plant compartments, and each compartment represents a unique niche of the bacterial community. Our study not only confirms the niche differentiation of the microbes at the soil-root interface but also demonstrates the fine-tuning and adaptation of the endophytic microbiota in the stem, leaf, and fruit compartments. In addition, we have identified the hub bacterial microbes of Camellia oleifera . This study provides a relevant model for the systematic study of the changes in microbial community in the organizational level niche of Camellia oleifera plants. These results fill the knowledge gap of the endophytic bacterial community of Camellia oleifera and provide a theoretical basis for the subsequent exploration of microbial functions and research on bio-fertilizers."
} | 4,246 |
37799163 | null | s2 | 5,628 | {
"abstract": "DNA origami is a promising technology for its reproducibility, flexibility, scalability and biocompatibility. Among the several potential applications, DNA origami has been proposed as a tool for drug delivery and as a contrast agent, since a conformational change upon specific target interaction may be used to release a drug or produce a physical signal, respectively. However, its conformation should be robust with respect to the properties of the medium in which either the recognition or the read-out take place, such as pressure, viscosity and any other unspecific interaction other than the desired target recognition. Here we report on the read-out robustness of a tetragonal DNA-origami/gold-nanoparticle hybrid structure able to change its configuration, which is transduced in a change of its plasmonic properties, upon interaction with a specific DNA target. We investigated its response when analyzed in three different media: aqueous solution, solid support and viscous gel. We show that, once a conformational variation is produced, it remains unaffected by the subsequent physical interactions with the environment."
} | 283 |
36944647 | PMC10030830 | pmc | 5,629 | {
"abstract": "Neurobiological systems continually interact with the surrounding environment to refine their behaviour toward the best possible reward. Achieving such learning by experience is one of the main challenges of artificial intelligence, but currently it is hindered by the lack of hardware capable of plastic adaptation. Here, we propose a bio-inspired recurrent neural network, mastered by a digital system on chip with resistive-switching synaptic arrays of memory devices, which exploits homeostatic Hebbian learning for improved efficiency. All the results are discussed experimentally and theoretically, proposing a conceptual framework for benchmarking the main outcomes in terms of accuracy and resilience. To test the proposed architecture for reinforcement learning tasks, we study the autonomous exploration of continually evolving environments and verify the results for the Mars rover navigation. We also show that, compared to conventional deep learning techniques, our in-memory hardware has the potential to achieve a significant boost in speed and power-saving.",
"introduction": "Introduction In the last decades, artificial intelligence (AI) has drawn inspiration from the biological world, where humans and animals interact with one another and the surrounding environment to improve the efficiency of routine tasks 1 . This continuous and mutual interplay enables a constant boost of the abilities, the knowledge, and the complexity of the organisms, which become increasingly resilient to the daily life 2 . Currently, achieving efficient adaptation to the continually evolving situations of life is a major objective of the AI community, whose principal aim is to build machines able to infer concepts and to make decisions 3 . The experience-based knowledge, where agents evolve by trial-and-error episodes throughout their entire life, is an interdisciplinary subject of biology, computer science and neuroscience known as “reinforcement learning” 4 . During the last decades there have been several studies to contextualize the framework of reinforcement learning. For instance, the Markov Decision Process introduces a numerical framework under the hypothesis that the state probability and the reinforcement learning operations are known and accessible 5 , 6 . Such decision-making procedure introduces a probability function \\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}$$P({{{{{\\rm{s}}}}}},\\, {a},\\,{s}^{{\\prime} })$$\\end{document} P ( s , a , s ′ ) which weights the value \\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}$$V\\left(s\\right)$$\\end{document} V s of a certain position “ \\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}$$s$$\\end{document} s ” for moving toward another state “ \\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}$${s}^{{\\prime} }$$\\end{document} s ′ ”. In equation: 1 \\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}$$V\\left(s\\right)={{{\\max }}}_{a}(R\\left(s,\\, a\\right)+\\alpha \\mathop{\\sum}\\limits_{s^{\\prime} }P(s,\\, a,\\, s^{\\prime} )V({s}^{{\\prime} })),$$\\end{document} V s = max a ( R s , a + α ∑ s ′ P ( s , a , s ′ ) V ( s ′ ) ) , The solution of the Markov process is a policy method which defines, if the model of the environment is known, the most convenient action to take at every available state 6 . However, in biology, organisms do not often have a model of the environment a priori, and they have to handle their own policies relying on the current occurrences by direct interaction with the surroundings. In this context, the Q-learning theory is a model-free algorithm used to assess the quality of an action in a particular state 7 . In formula: 2 \\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}$$Q\\left(s\\right)=R\\left(s,\\, a\\right)+\\alpha \\mathop{\\sum}\\limits_{s^{\\prime} }P(s,\\, a,\\, s^{\\prime} ){{{\\max }}}_{a^{\\prime} }Q({s}^{{\\prime} },{a}^{{\\prime}} ),$$\\end{document} Q s = R s , a + α ∑ s ′ P ( s , a , s ′ ) max a ′ Q ( s ′ , a ′ ) , where \\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}$${\\max }_{a^{\\prime} }Q({s}^{{\\prime} },\\, a^{\\prime} )$$\\end{document} max a ′ Q ( s ′ , a ′ ) is the maximum of all the possible \\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}$$V\\left(s\\right)$$\\end{document} V s . Consequently, a quality of a certain position \\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}$$Q\\left(s\\right)$$\\end{document} Q s is dependent on the quality of the nearest states \\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}$${s}^{{\\prime} }$$\\end{document} s ′ . The Q-values can also map the value of each position with respect to the environmental modulations in time, thus defining the so called “temporal difference (TD)” framework 8 : 3 \\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}$${{TD}}_{t}\\left(s,\\, a\\right)=\\beta ({Q}_{t}\\left(s,\\, a\\right)-{Q}_{t-1}\\left(s,\\, a\\right)),$$\\end{document} T D t s , a = β ( Q t s , a − Q t − 1 s , a ) , where \\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}$$\\beta$$\\end{document} β is the inverse of the learning rate of the current \\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}$$Q$$\\end{document} Q value with respect to the previous \\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}$$Q$$\\end{document} Q value ( \\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}$${Q}_{t-1}\\left(a,\\, s\\right)$$\\end{document} Q t − 1 a , s ). These models can map the behaviour of the agent developing a decision-based policy by exploiting the interaction with the environment and taking a decision whose effect, in turn, constitutes part of the experience of the agent 9 – 11 . All these intuitions have been demonstrated in “Dyna”, where learning methods were used for managing the planning results and for developing a cause-consequence model of the agent’s actions 12 . To study the spatial learning and memory, several experiments were carried out in the field of behavioural neuroscience, such as the water maze exploration 13 , 14 . In particular, the Morris Maze navigation has been investigated by neuroscientists to study the effect of cognitive diseases related to the spatial learning 15 . Such studies also modelled the physiological basis of reward-based behaviours using Hebbian learning and spiking neurons 16 . In this context, it has been observed that when a penalty/reward event occurs, humans and animals release in brain dopamine, a pleasure-related neurotransmitter which become the reinforcement variable for the elaboration of the experience 13 . All these findings have been sources of inspiration for building intelligent hardware computing elements. In particular, in the last years, recurrent synaptic connections have been addressed as key elements for reproducing reward-based decision-making demonstrators 17 using both CMOS-based platforms 18 and non-volatile memories 19 . CMOS technology is the most mature approach for the AI hardware design, highlighted by the results achieved by deep learning with AlphaGo 20 , 21 . However, the first hardware setup of AlphaGo required 1920 central processing units (CPUs) and 280 graphics processing units (GPUs), with a peak power of half a megawatt 22 . Such power requirement is far from what is observed in brain-computation for mainly two reasons: (i) the slow and energy-hungry training procedures of deep learning techniques, for instance the “backpropagation” 21 ; (ii) the communication delay between the processing units and the dynamic random-access memory (DRAM), also known as “Von Neumann bottleneck”, while biological computation happens in-situ, i.e. in the same place where the information is stored 23 . For this reason, memristors, such as resistive switching devices (RRAMs) and phase change memories (PCMs), appear interesting for emulating the stochastic neuro-plausible computing, thanks to the reduced area, 3D stacking capability in the backend-of-the-line, increased parallelism, and analogue storage 24 – 27 . A key advantage of networks based on these emerging devices is the fast computation exploited by vector-matrix multiplications which can intrinsically perform in-situ multiplication and summation via Ohm’s and Kirchhoff’s laws 28 , 29 . Memristor arrays have shown enhancements in speed and energy for both in-memory supervised learning 30 – 32 and unsupervised learning 33 – 36 . Furthermore, they are the best candidates for neurocomputing, boosting algorithms such as the spike-timing dependent plasticity (STDP) 35 , 37 and the homeostatic mechanisms to stabilize the divergent growth of the weights under pure Hebbian learning 38 , 39 . Such features offer key abilities for the implementation of resilient bio-inspired systems but, generally, are not as accurate as standard deep learning approaches, which, on the other hand, lack plasticity. These dichotomies of artificial neural networks with respect to the biological word was summed up since the early years of investigation in AI with the sentence “stability-plasticity dilemma” 40 . In this work, we propose a neuromorphic hardware based on Silicon Oxide (SiO x ) RRAM devices able to join state-of-the-art accuracy and bio-inspired plasticity for autonomous and resilient navigation at low-power. The network relies on bio-inspired algorithms, such as STDP and plastic homeostasis, to adjust the parameters along a temporal sequence, as in recurrent neural networks (RNNs) 41 . The RRAM devices are used for both Hebbian learning processes (integration, fire, potentiation/depression of the synapses) and to map the recurrent internal state of each neuron. In particular, the multilevel capability of the devices is used to modulate the neuronal threshold, acting as homeostatic boundary of the firing activities 42 . To test the resilience of the hardware, a two-dimensional dynamic maze showing environmental changes in time is experimentally configured in a field-programmable-gate-array (FPGA), thus mimicking biology 16 and deep-learning software-based approaches 43 – 45 . The bio-inspired hardware described in this work is also tested for complex cases such as the Mars rover navigation, thus investigating the properties of the system in terms of scalability and reconfigurability. The network starts from stochastic trials, it progressively maps the configuration of the environment, it becomes a master of the problem trial after trial, and it finally finds the optimum path towards the objective. Furthermore, we benchmark our work with respect to deep learning techniques, finally demonstrating that our solution overcomes the standard approaches used for autonomous navigation. In this context, we also present a theoretical framework which highlights the main benefits of the RRAM-based in-situ computation such as the high efficiency, resilience, low power consumption and accuracy. In the Supplementary Discussion of this manuscript, we also provide a further appendix on the numerical modelling of the bio-inspired approach to reinforcement learning and a more technical insight about the experimental setup.",
"discussion": "Discussion Deep learning techniques using standard Von Neumann processors enable accurate autonomous navigation but require great power consumption and long time for making training algorithms effective, Fig. 4f, g ( https://pypi.org/project/pyqlearning/ ). In particular, the environmental information is often sparse, noisy and delayed, while training procedures are supervised and require direct association between inputs and targets during the backpropagation. Hence, complex models of convolutional neural networks are needed to numerically find the best combination of parameters for the deep reinforcement computation 62 , ( https://pypi.org/project/pyqlearning/ ). Thus, the standard approaches to reinforcement learning enable free-policy learning by reinforcement, but this is paid in terms of lower accuracy with respect to the same environmental configuration, Fig. 4f , and in higher cost of the resources when a specific performance is targeted, Fig. 4g . Furthermore, standard processors require data transmission back and forth the DRAM (Von Neumann bottleneck) while in-memory computing assures a local processing of the information where it is stored, Fig. 1a 22 , 23 , 63 . Note also that deep Q-learning techniques suffer from unstable learning under some conditions of bias overestimation which requires a mutual training of a multi-layer-perceptron (MLP) network and a correct setting of the learning rate 64 . This could affect the effectiveness of the training algorithm when the system must map the environment autonomously, requiring a network of several layers with the Adam optimizer applied for stochastic optimization 65 . All these features assure high accuracy in, at least, 1000 episodes for each trial. Contrarily, the bio-inspired learning procedure relies on training-free in-situ hardware computation. This approach improves a lot the time efficiency, Fig. 4b , and the energy consumption, Fig. 4c , while keeping high the accuracy, Fig. 4f . Furthermore, the STDP does not require dedicated methods for stochastic optimization, and it assures an optimum behaviour also when the configuration of the space to explore is not constant 36 . Related to this context, in the supplementary appendix “Comparison of the resilient properties between bio-inspired and deep learning approaches”, we report a theoretical study over the adaptation capabilities of the neuromorphic solution with respect to the standard Python-based approach ( https://pypi.org/project/pyqlearning/ ). The Markov decision process, Q-learning, TD(λ) and deep learning are not the only topics to which the scientific community refers to for modelling and designing reinforcement learning algorithms. For instance, the multi-bandit problem is often taken as benchmark. The multi-armed bandit problem deals with an agent that attempts to make decisions as a consequence of previous experiences but, at the same time, it needs to acquire new knowledge for the next decision-making events. To cope with this framework, several works have proposed the use of RNNs for enhancing the re-use of past information 66 and for building “meta-learners”, i.e., systems trained on a distribution of similar tasks featuring a generalization capability when novel goals are targeted 67 , 68 . However, even considering these meta-approaches, several CNN-based training algorithms are anyway necessary to provide the system with an optimum policy map for the required navigation task, thus falling again in the power and time bottleneck. In conclusion, we proposed an event-based hardware based on RRAM devices capable of self-adaptation to get efficient neurocomputing in reinforcement learning tasks. We studied the experimental behaviour of the network highlighting the resilient capability of the autonomous navigation under various environmental difficulties, such as obstacles and dynamic modifications of the maze. We also proposed a study of the hardware reconfigurability of the system under the Mars rover navigation test. Finally, we introduced a theoretical framework for bio-inspired reinforcement learning highlighting the main outcomes of RRAM-based computation with respect to the state-of-the-art. This work highlights the relevance of bio-inspired approaches for artificial intelligence and underlines the computational benefits of non-volatile memories for autonomous hardware systems."
} | 4,654 |
39056601 | PMC11275114 | pmc | 5,630 | {
"abstract": "Microbial alkane degradation pathways provide biological routes for converting these hydrocarbons into higher-value products. We recently reported the functional expression of a methyl-alkylsuccinate synthase (Mas) system in Escherichia coli , allowing for the heterologous anaerobic activation of short-chain alkanes. However, the enzymatic activation of methane via natural or engineered alkylsuccinate synthases has yet to be reported. To address this, we employed high-throughput screening to engineer the itaconate (IA)-responsive regulatory protein ItcR (WT-ItcR) from Yersinia pseudotuberculosis to instead respond to methylsuccinate (MS, the product of methane addition to fumarate), resulting in genetically encoded biosensors for MS. Here, we describe ItcR variants that, when regulating fluorescent protein expression in E. coli , show increased sensitivity, improved overall response, and enhanced specificity toward exogenously added MS relative to the wild-type repressor. Structural modeling and analysis of the ItcR ligand binding pocket provide insights into the altered molecular recognition. In addition to serving as biosensors for screening alkylsuccinate synthases capable of methane activation, MS-responsive ItcR variants also establish a framework for the directed evolution of other molecular reporters, targeting longer-chain alkylsuccinate products or other succinate derivatives.",
"conclusion": "4. Conclusions The efficient conversion of methane to liquid fuels and other value-added chemicals remains the “holy grail” of catalysis; the conversion of short-chain alkanes is similarly challenging. Enzymatic/microbial approaches hold promise in overcoming the many catalytic hurdles and inefficiencies associated with existing chemical process technologies [ 29 ]. Whereas oxygen-dependent biological processes suffer from inherently large carbon and energy losses, anaerobic methane/alkane activation and conversion may significantly improve efficiency [ 1 ]. Toward the directed evolution of enzyme-based methane activation via fumarate addition, here, we report the first-ever genetically encoded biosensors of MS based on variants of the ItcR repressor. The differential responsivity of WT-ItcR to IA (high) and MS (negligible) compared to the Var7 sensor’s response—high to MS and reduced to IA—highlights the ability of subtle structural changes resulting from relatively few amino acid substitutions to significantly impact ligand binding and allosteric regulation. In addition to serving as a tool that enables high-throughput screening of mutant gene libraries for novel MS biosynthesis, an ItcR-based MS sensor serves as a “parent” in the further directed evolution of biosensors for alkylsuccinates resulting from the addition of fumarate to ethane and other short-chain alkanes.",
"introduction": "1. Introduction Methane and short-chain alkanes are abundant feedstocks in the chemical and energy industries. The controlled activation and conversion of these hydrocarbons into value-added products remains a major technical hurdle [ 1 , 2 ]. Microbial alkane degradation pathways provide biological routes to activate and metabolically convert small gaseous alkanes into higher-value products [ 3 , 4 , 5 ], offering a potential solution to the large-scale, capital-intensive challenges associated with existing gas-to-liquid technologies (e.g., Fischer-Tropsch) [ 1 ]. Anaerobic bio-activation of n -alkanes occurs predominantly through their addition to the double bond of fumarate via the activity of alkylsuccinate synthase enzymes [ 6 ]. The alkylsuccinate products are further degraded through rearrangement, decarboxylation, and β-oxidation, coupled with fumarate regeneration [ 7 , 8 ]. A key challenge to taking advantage of these pathways lies in the difficulty of functionally expressing alkylsuccinate synthases, along with their partner “activating” enzymes (AEs), in a host organism suitable for metabolic engineering and bioprocessing. We recently demonstrated the first-ever functional expression of methyl-alkylsuccinate synthase (Mas) from the Azoarcus sp. strain “HxN1”, along with its partner AE, in a recombinant host ( Escherichia coli ) [ 4 ]. Consistent with the Mas activity reported for HxN1 [ 6 ], we observed the activation of C3-C6 linear alkanes. Organisms capable of anaerobic degradation of ethane and propane have been identified [ 5 , 8 , 9 ], but there are no indications of alkylsuccinate synthase-based methane activation in nature. Having established the functional expression of Mas, we now seek to improve the biosynthesis of alkylsuccinates as well as engineer Mas for (enhanced) activity toward shorter alkanes and, potentially, methane. Given the many variables and possible bottlenecks in this transformation, it is desirable to employ a high-throughput screening of many different gene libraries to identify mutations that confer enhanced product formation. High-throughput screening is also a powerful approach for identifying enzyme variants having improved substrate specificity, or even activity toward non-native substrates. In this context, a common limitation in high-throughput mutational analysis is the lack of a sensitive and compound-specific screening method. For the case of alkylsuccinates produced within a mixture of many dicarboxylic acids, we know of no such assay. The use of natural or engineered bacterial transcription factors as endogenous biosensors that report on the production of target small molecules is now an established, powerful approach to screening for novel or improved biosynthesis in whole cells [ 10 , 11 , 12 ]. Here, we describe the design of genetically encoded biosensors that specifically report on the presence of methylsuccinate (MS, the product of methane addition to fumarate) via the expression of a fluorescent reporter protein. These biosensors were constructed by altering the ligand specificity of the ItcR repressor. ItcR is a LysR-type transcriptional regulator located in the opposite direction of an operon encoding genes responsible for the catabolism of itaconate (IA, also called methylenesuccinate(2-)) in Yersinia pseudotuberculosis [ 13 ]. Gene repression by ItcR is relieved when the repressor binds itaconate [ 13 ]. Itaconate differs from MS by just one C-C bond ( Figure 1 ), yet ItcR shows a 100-fold lower induced gene expression response to MS at 5 mM inducer [ 13 ]. We tested whether wild-type ItcR (WT-ItcR), when expressed in E. coli , could be induced by other alkylsuccinates of interest: (1-ethyl)succinate (ES), (1-methylethyl)succinate (MES), and (1-methylpentyl)succinate (MPS). In all cases, the response was lower than that observed for MS ( Table S2 ). Given the chemical similarity of itaconate to MS, we reasoned that ItcR could be engineered to instead respond to MS without great difficulty. Indeed, a single round of random mutagenesis yielded an ItcR variant that, when expressed in E. coli , shows a >9-fold improved response to 1 mM MS (added to the culture broth). Subsequent rounds of mutagenesis resulted in a collection of other variants having further enhanced sensitivity and specificity toward MS. The structural modeling and analysis of the ItcR ligand binding pocket provides mechanistic insights into the altered molecular recognition. In addition to serving as a tool to screen for the Mas-variant-catalyzed, oxygen-independent activation of methane, MS-responsive ItcR variants also serve as starting points for further directed evolution of inducer specificity toward other alkylsuccinates (products of fumarate addition to short-chain alkanes) and succinate derivatives.",
"discussion": "3. Results and Discussion 3.1. Directed Evolution of MS-Responsive ItcR Variants The evolution progression of MS-sensitive variants originating from WT-ItcR is depicted in Figure 2 . We initially used random mutagenesis across the entire ItcR ligand-binding domain (LBD) coding region. Screening was performed using FACS, with the induced expression of mCherry red fluorescent protein (RFP) [ 15 ] serving as the reporter. To remove “leaky” variants that cause high RFP fluorescence in the absence of the inducer, random mutagenesis libraries were first sorted to isolate a low-background population in the absence of MS. This population was next subjected to “positive” screening in the presence of 1 mM MS. First-round screening resulted in variant #1 (“Var1”) (E127V, V144A, and K269R), showing 9.3 ± 0.9-fold improvement in transcriptional response to 1 mM MS. We next checked the contribution of each individual amino acid substitution found in Var1 toward this improved response by constructing variants #2, #3, and #4 (“Var2”, etc.). Var2 (E127V and V144A) showed the same response as Var1, while Var3 (V144A and K269R) and Var4 (E127V and K269R) showed a 5.2 ± 0.8-fold and 1.5 ± 0.2-fold lower induced gene expression response to 1 mM MS (relative to Var1), respectively. This result led us to generate a site-saturation mutagenesis (SSM) library targeting AA positions 127 and 144, using Var1 as the parent. Here, “NDT” codons were used, resulting in a library consisting of 144 possible variants, out of which 400 isolated clones were screened. No variant showed an improved response relative to Var1. A second round of random mutagenesis (with Var1 as parent) and FACS sorting (using 1 mM MS) was next conducted. The resulting variants, Var5 (E127V, V144A, K269R, and E173K), Var6 (E127V, V144A, K269R, and S280T), and Var7 (E127V, V144A, K269R, and L195I), showed 15.1 ± 2-, 15.2 ± 0.9-, and 17.3 ± 2-fold improvement in the RFP expression response to 1 mM MS (relative to WT-ItcR), respectively. From these newly identified amino acid substitutions, we next constructed variants Var8 (E127V, V144A, K269R, S280T, and E173K) and Var9 (E127V, V144A, K269R, L195I, and E173K). The E173K substitution was added to Var6 and Var7 because Var5 showed the highest response to a relatively low concentration of MS (100 µM). Although Var8 and Var9 showed lower-fold induced RFP at the high concentration (1 mM), Var8 showed a 1.2 ± 0.2-fold improved response to 100 µM MS compared to Var5. This result, in turn, guided the generation of another SSM library, this time targeting positions 173 and 280, with Var8 as the parent. The library was screened in the presence of 1mM MS, yielding Var10 (E127V, V144A, K269R, and E173V) showing 1.6 ± 0.3-fold induced RFP expression in 1 mM MS (as compared to Var8), but no improvement with 100 µM MS. 3.2. Characterization of MS Biosensors Dose–response curves were fitted to a modified Hill equation [ 28 ], as shown below: (1) y = a + ( b − a ) / ( 1 + k x n ) Here, y is the response (i.e., the normalized fluorescence of cells, resulting from the expression of RFP) at x concentration of the inducer (note that this is the concentration of the exogenous inducer—that which was added to the culture broth). Parameters a , b , and k describe the response at zero concentration, the maximum response, and the concentration corresponding to 50% of the maximum response (may be considered an “apparent K d ” for the response to the exogenous inducer), respectively. The Hill coefficient ( n ) describes the cooperativity of the biosensor. The calculated values of these parameters for each designed biosensor are reported in Table 1 . The corresponding plotted data and fitted equations are provided in Figure S6 of the Supplementary Materials . One representative dose–response curve (that for Var7) is provided in Figure 3 . WT-ItcR, and presumably its variants, operates as a dimer with two possible ligand binding sites. We determined a Hill coefficient of ~1.7 for WT-ItcR with IA ( Table 1 ). Most variants similarly have Hill coefficients in this range (~1.5–1.9) with MS, suggesting a conserved cooperative binding mechanism. The values for Var8 and Var9 are lower but still indicative of positive, partial cooperativity. Such variation in n may reflect alterations in the binding site and/or protein conformation that affect ligand binding cooperativity. The sensitivity of each variant sensor system, quantified by ‘ k ’ in Equation (1), is plotted against their respective background RFP expression levels (parameter ‘ a ’, representing “leaky” expression) in Figure 4 . All ItcR variants show enhanced sensitivity to MS. Var8 and Var9 display sensitivities to MS comparable to that of WT-ItcR to IA (~0.44 mM), though notably with much higher leaky expression. All other variants show background levels comparable to that of WT-ItcR. Table 2 lists the fold induced RFP expression value for each variant at 0.1 mM, 1 mM, and 5 mM MS. Var7 shows a ~25-fold induced expression response to 5 mM MS, which represents a 10-fold improvement compared to WT-ItcR at the same concentration. The inducer specificity of the MS-responsive ItcR variants was assessed by comparing the RFP expression response to MS with that to IA, as well as other, potentially “competing” ligands when screening for MS biosynthesis (i.e., fumarate and succinate). As shown in Table 3 , all biosensors show high specificity toward MS over both fumarate and succinate (≥12). Specificity over IA is also significantly improved compared to WT-ItcR, with Var6 having ~77-fold increased MS/IA specificity. 3.3. Binding-Site Modeling and Analysis: Variant 7 vs. WT-ItcR The structures of the WT-ItcR and Var7 LBDs were modeled using AlphaFold [ 23 ]; this was performed prior to the publication of the solved ItcR X-ray crystal structure [ 22 ]. The predicted structure exhibited remarkable similarity to the experimentally solved structure, with an RMSD of 0.64 Å. The predicted LBD structure of Var7 (using AlphaFold) deviated from the WT-ItcR LBD structure by an RMSD of 0.66 Å, suggesting the amino acid substitutions in Var7 did not likely confer substantial structural changes to the LBD. IA and MS (both R- and S-isomers) were next docked into the LBDs of WT-ItcR and Var7 (refer to Methods Section). The resulting Glide docking parameters are provided in Table S4 . Detailed docking poses for WT-ItcR with IA, WT-ItcR with MS, Var7 with IA, and Var7 with MS (R- and S-isomers), including relevant interaction distances, are provided in Figure S8 . Figure 5 depicts docking pose overlays for WT-ItcR with IA vs. MS (S-isomer) and WT-ItcR with IA vs. Var7 with MS. For ease of visualization, only the most relevant binding pocket residues are included in these overlays. In comparing the docking pose of IA to that of MS in WT-ItcR ( Figure 5 a), there is a clear difference in ligand orientation. Perhaps most notably, whereas the methylene group of IA lies 3.9 Å above the F196 ring (forming a critical hydrophobic interaction [ 22 ]), the methyl group of MS is instead pointing away from F196. The Glide docking scores ( Table S4 ) indicate stronger binding for IA in WT-ItcR (−5.32 kcal/mol) as compared to MS (−4.01 kcal/mol). It is important to note that while MS is a much poorer inducer of WT-ItcR as compared to IA, MS still acts as a binding pocket ligand. Figure 5 b shows a clear shift in the position of MS in Var7 relative to that in WT-ItcR. For reference, the C-atom RMSD between IA and MS docked into WT-ItcR is 2.97 Å, while that between IA in WT-ItcR and MS in Var7 is 1.62 Å. Significantly, we now see the MS methyl group positioned 4.2 Å above F196 in the Var7-MS complex, similar to the methylene of IA in the WT structure. The L195I substitution of Var7 likely helps to reposition F196 to support this interaction. L195I may further alleviate steric hindrance by the slightly bulkier MS ligand. It is also noteworthy that the T98 hydroxyl group in Var7 now lies 2.8 Å from O3- of MS (the same distance from O2- of IA in the WT-ItcR complex), as compared to 4.1 Å between these atoms for the case of MS docked into WT-ItcR. H-bonding interactions with R148 and S100 are similarly more conserved between the Var7-MS and WT-ItcR-IA complexes as compared with those for MS docked into WT-ItcR."
} | 3,997 |
36923508 | PMC10009093 | pmc | 5,631 | {
"abstract": "Highlights • Acidithiobacillus ferrooxidans was engineered to overexpress quorum sensing genes. • Engineered cells exhibited enhanced mineral attachment. • Engineered cells led to higher bioleaching efficiencies of covellite under low iron conditions. • The afeI gene under control of the tac promoter was superior to overexpression of the endogenous operon.",
"conclusion": "4 Conclusions The effect of overexpression of QS production machinery in A. ferrooxidans was examined on the bioleaching of the copper sulfide, covellite. The engineered cells not only improved the cell attachment to the mineral by increasing EPS production and biofilm formation, but also enhanced the bioleaching efficiency by accelerating iron oxidation of attached cells. Overall, the improvement of the process performance by engineered cells was more significant when the overexpression of quorum sensing gene was regulated under tac promoter than the native promoter.",
"introduction": "1 Introduction Bioleaching is a promising metal recovery technology which can reduce environmental impacts and improve process economics as compared to conventional pyrometallurgy. During bioleaching, microorganisms solubilize metals either via non-contact or contact mechanisms [1] . As many natural ores are sulfidic deposits [2] , cell attachment to the solid surface is thought to be an essential step during the initiation of bioleaching. The direct interaction between the microorganisms and surface involves biofilm formation, which is regulated by quorum sensing (QS) in many bacteria [3] . QS is an intra- or interspecies chemical communication system that is used to regulate the expression of specific genes in response to variations in bacterial populations. This can lead to changes in the synthesis of extracellular polymeric substances (EPS) and encourage cell aggregation [3] , [4] , [5] . Characterization and potential modulation of the QS machinery of extremophiles is particularly important, considering that population regulation is a fundamental strategy used to tolerate harsh growth environments [6] . Acidithiobacillus ferrooxidans is a representative acidophilic chemolithotrophic bioleaching microorganism [7] , which obtains energy from oxidation of iron and/or reduced sulfur species. A Lux I/R-like type AI-1 QS system (AfeI/R or qs-I) has been identified and characterized in A. ferrooxidans \n [4] . The addition of a QS signaling molecule, acyl-homoserine lactone (AHL), improved the cell attachment of A. ferrooxidans to solid materials [8] , [9] , [10] , [11] , suggesting an important role for QS in the regulation of cell-surface interactions. Gao, Liu, Fu, Gu, Lin, Liu, Pang, Lin and Chen [12] observed that overexpression of the endogenous qs-I operon improved EPS production and sulfur oxidation activities of A. ferrooxidans during the bioleaching of reduced sulfur compounds (sulfur and pyrite). The authors further characterized a substrate-dependent modulation of the QS machinery in A. ferrooxidans , where stimulation was observed in the presence of sulfur (0.8% w/v). However, substantial inhibition was observed when iron (10 g/L) was present in the growth medium [4] . These observations raise questions as to whether the manipulation of QS systems can be exploited to enhance the bioleaching of copper sulfide minerals, since iron cycling is an important driving force for mineral oxidation. In addition, it is not clear how QS will be influenced by the presence of both sulfur and iron in the bioleaching milieu. The effect of the overexpression of the QS system on bioleaching of copper minerals such as covellite has not yet been reported. Here we engineered A. ferrooxidans for the recombinant overexpression of the endogenous qs-I operon or just the afeI gene (encoding AHL synthase) under control of the tac promoter. The tac promoter is a strong constitutive promoter for gene expression in A. ferrooxidans , and this allows for the effect of continuous transcription of the afeI gene under the control of tac promoter to be compared with the endogenously regulated gene expression by the native promoter. We explored the impact of this modification on the bioleaching of covellite, for which the bioleaching efficiency is generally lower than for other copper sulfides, such as chalcocite. The genetic modifications not only led to enhanced cell adhesion to covellite, but also improved the bioleaching efficiency of the mineral under low iron conditions. In addition, the expression of just the afeI gene under control of the constitutive tac promoter further enhanced the bioleaching kinetics as compared to the overexpression of the native operon.",
"discussion": "3 Results and discussion 3.1 Construction of engineered cells Two A. ferrooxidans strains were created by transformation with plasmids encoding enzymes implicated in QS (Table S2). The first contained the endogenous qs-I operon (AF QS ) which expresses afeI, afeR and a hypothetical protein under control of their endogenous promoters. The second contained just the afeI gene encoding the AHL synthase (AF QS-tac ), under control of the tac promoter. A strain expressing GFP (AF GFP ) was used as a control in addition to the wild type cells (AF), to explore the effect of plasmid transformation on cell performance. During the initial growth in F2S medium, AF QS and AF QS-tac cells showed 19–23-fold and 34–42-fold higher transcriptional gene expression of afeI than the two control strains, respectively (Table S3). This confirms the effective overexpression of afeI gene in both engineered cell lines, and four strains were used for bioleaching. 3.2 Effect on cell attachment To examine the effect of QS gene overexpression in A. ferrooxidans on cell attachment to minerals, the total EPS concentration and biofilm formation were analyzed, at the end of the bioleaching experiments. AF and AF GFP strains showed comparable EPS concentrations and biofilm formation ( Fig. 1 ), indicating that the expression of an unrelated gene (i.e., GFP) had little impact on cell attachment. On the other hand, both QS cell lines exhibited higher total EPS concentrations (1.3–1.5 times higher than AF and AF GFP cells). It was observed that both the loosely bound (LB-EPS) and tightly bound (TB-EPS) EPS fractions increased under the conditions with AF QS and AF QS-tac . In the same manner, a 2.1–2.5 fold increase in biofilm formation was observed in the conditions with engineered cells compared to the controls, which is consistent with the fact that cell-bound EPS fractions are important constituents in biofilm matrices [20] . Interestingly, AF QS-tac further enhanced both EPS and biofilm formation compared to AF QS , indicating that the effect of QS overexpression was more evident when afeI was expressed alone with the strong tac promoter. Improved cell attachment by engineered cells was also observed visually, where more cells were located at the covellite surface, as compared to AF and AF GFP cells ( Fig. 2 ). The higher carbon content in the conditions with AF QS and AF QS-tac also indicates that more cells adhered to the covellite surface (Fig. S2A). Notably, porous biofilm configurations were observed only in the conditions with engineered cells (Fig. S2B). A similar morphology was previously observed when adjacent cells bridge to each other, creating a regular biofilm structure [21] . These observations indicate that QS gene overexpression in A. ferrooxidans promotes EPS production and biofilm formation, thus facilitating cell adherence to the covellite surface. Fig. 1 The concentration of total EPS with (A) soluble, loosely-bound (LB-EPS), and tightly-bound (TB-EPS) fractions and (B) biofilm formation after the covellite bioleaching under the different conditions. Error bars indicate standard deviations, and the asterisks indicate statistical significance. AF, wild type cells; AF GFP , engineered cells with GFP; AF QS , engineered cells with overexpression of quorum sensing; AF QS-tac , engineered cells with overexpression of quorum sensing under tac promoter. Fig 1 Fig. 2 FISH Detection of A. ferrooxidans cells on the covellite residues after bioleaching with (A) the wild type A. ferrooxidans (AF), (B) the wild type cells with GFP (AF GFP ), (C) the engineered cells overexpressing the quorum sensing operon (AF QS ), or (D) or afeI under control of the tac promoter (AF QS-tac ). The bars represent 10 mm. Fig 2 3.3 Effects on bioleaching efficiency Copper bioleaching efficiency of covellite (CuS) by the four A. ferrooxidans strains was monitored. Iron was initially withheld from the experiments, and bioleaching was not observed. The addition of iron on Day 5 facilitated copper solubilization as cells were able to generate Fe 3+ as a primary oxidant for covellite leaching [22] . In particular, AF QS and AF QS-tac rapidly oxidized Fe 2+ after iron supplementation, compared to the controls, while AF QS-tac showed a higher iron oxidation rate than AF QS (Fig. S3). The rapid iron oxidation by the engineered cells coincided with the higher bioleaching efficiencies ( Fig. 3 ), as compared to the controls. Accordingly, the cultures with AF QS and AF QS-tac achieved final bioleaching efficiencies of 43% and 51%, respectively, which are 1.4–1.7 times higher than AF and AF GFP . After bioleaching, the mineral residues included jarosite, pyrite, and covellite, regardless of the different experimental conditions (Fig. S4), which agrees with previous observation [ 17 , 23 ]. Fig. 3 Copper bioleaching from covellite by different cell lines, through the 26-days of the bioleaching experiments. Ferrous iron was added on Day 5 of the bioleaching experiments to achieve a 10 mM final concentration. Error bar presents standard deviations. AF (black), wild type cells; AF GFP (gray), engineered cells with GFP; AF QS (blue), engineered cells overexpressing the QS operon; AF QS-tac (yellow), cells overexpressing afeI under the tac promoter. Asterisks indicate statistical significance of the engineered cells compared to either of AF or AF GFP (*), and between engineered cells (**) at each time point. Fig 3 The QS overexpression in A. ferrooxidans improved both cell attachment and bioleaching efficiency of covellite, which is consistent with the pivotal role of cell-surface interactions on bioleaching of solid minerals [24] , [25] , [26] . Interestingly, both engineered cells improved iron oxidation during bioleaching. This result is in contrast to previous observations made when iron was used as an energy source, as QS overexpression led to suppressed iron respiration and decreased EPS and biofilm formation [ 3 , 4 ]. A minimal amount of iron is necessary to enable bioleaching, and here we found that a low iron concentration (10 mM Fe 2+ used in this study) was likely below a threshold necessary to be impacted by the overexpression of the QS genes. It is also possible that the presence of sulfur (provided by covellite) could compensate for the inhibitory effect of iron on QS regulation, considering the substrate-dependent effects that have been reported with the QS machinery [ 4 , 12 ]. Even though the iron oxidation rate was slower in the control cultures with AF QS and AF QS-tac as compared to the engineered strains, the final concentration of Fe 3+ at the end of the experiments was comparable among all four strains (Figure S3). However, the engineered strains exhibited enhanced bioleaching efficiencies as compared to the two control strains ( Fig. 3 ). This is consistent with a mechanism where localized oxidation by mineral-attached cells is more beneficial for the bioleaching of covellite, as compared to the non-contact iron-mediated oxidation performed by planktonic cells. Thus, the improved bioleaching efficiencies ( Fig. 3 ) can likely be attributed to the enhanced iron oxidation rates and which may be more effective when the cells are concentrated in the biofilms attached to the cells [17] . 3.4 Transcriptional gene expression of afeI During the bioleaching, the transcriptional gene expression of afeI of the four cell lines was increased (6.8–53-fold) compared with that during the cell growth (Tables 1 and S3). This suggests that the cells controlled the gene expression preferably when they are exposed the mineral environment. In addition, AF QS and AF QS-tac showed up to 3.9- and 25-fold higher transcriptional gene expression of afeI during the bioleaching, compared to AF and AF GFP ( Table 1 ).Given that the type AI-I QS system characterized in A. ferrooxidans includes AHL synthase (AfeI) [4] , the higher afeI concentration by the engineered cells demonstrates the improved AHL-mediated gene expression by QS overexpression, during the bioleaching of covellite. This also confirms no repression of QS system by the low amount of iron (10 mM Fe) present during the bioleaching. Since the expression and regulation of QS is a strategy for enabling bacterial communication in response to cell populations [27] , the higher transcriptional expression of afeI by the engineered cells is consistent with the enhanced biofilm formation by these strains ( Fig. 1 ). Moreover, the higher afeI gene concentration of AF QS-tac compared to AF QS indicates that afeI overexpression under tac promoter drives stronger expression than native promoter in qs-I operon. Table 1 Concentrations of the afeI gene transcripts in the different experimental cell lines at the end of the covellite bioleaching experiments. Data were measured in triplicate and error bars represent standard deviations. Table 1 Cell strain afeI gene concentration (copies/g DNA) a AF (7.3 ± 0.8) × 10 8 †, ‡, # AF GFP (1.0 ± 0.1) × 10 9 †, ‡, # AF Qs (3.0 ± 2.2) × 10 9 *, **, ‡ AF Qs-tac (1.9 ± 0.3) × 10 10 *, **, ‡, # a Values showing statistically significant differences ( p < 0.05) among different types of cells (*, compared to AF; **, compared to AF GFP ; †, compared to AF QS ; ‡, compared to AF QS-tac ) and between the same types of cells with different growth conditions (Table S3) (#). Taken together, QS overexpression in A. ferrooxidans enhanced cell attachment to the mineral and thereby improved the bioleaching efficiency of covellite, which was particularly effective when afeI expression was controlled under the tac promoter. These results suggest an interesting strategy to improve bioleaching efficiency from covellite and potentially other copper sulfide minerals, by overexpression of QS machinery in bioleaching microbes under low iron conditions . In addition to primary ore sources, this approach can also be applied to economically advantageous secondary sources, such as electronic wastes, where the cell attachment can also be a critical for effective bioleaching."
} | 3,711 |
39215820 | PMC11365853 | pmc | 5,632 | {
"abstract": "Coral reefs rely heavily on reef fish for their health, yet overfishing has resulted in their decline, leading to an increase in fast-growing algae and changes in reef ecosystems, a phenomenon described as the phase-shift. A clearer understanding of the intricate interplay between herbivorous, their food, and their gut microbiomes could enhance reef health. This study examines the gut microbiome and isotopic markers (δ 13 C and δ 15 N) of four key nominally herbivorous reef fish species ( Acanthurus chirurgus , Kyphosus sp., Scarus trispinosus , and Sparisoma axillare ) in the Southwestern Atlantic’s Abrolhos Reef systems. Approximately 16.8 million 16S rRNA sequences were produced for the four fish species, with an average of 317,047 ± 57,007 per species. Bacteria such as Proteobacteria, Firmicutes, and Cyanobacteria were prevalent in their microbiomes. These fish show unique microbiomes that result from co-diversification, diet, and restricted movement. Coral-associated bacteria ( Endozoicomonas , Rhizobia , and Ruegeria ) were found in abundance in the gut contents of the parrotfish species Sc. trispinosus and Sp. axillare . These parrotfishes could aid coral health by disseminating such beneficial bacteria across the reef. Meanwhile, Kyphosus sp. predominantly had Pirellulaceae and Rhodobacteraceae. Four fish species had a diet composed of turf components (filamentous Cyanobacteria) and brown algae ( Dictyopteris ). They also had similar isotopic niches, suggesting they shared food sources. A significant difference was observed between the isotopic signature of fish muscular gut tissue and gut contents, pointing to the role that host genetics and gut microbes play in differentiating fish tissues. Supplementary Information The online version contains supplementary material available at 10.1007/s00248-024-02423-x.",
"conclusion": "Conclusions This study highlights key features of the herbivorous fish gut microbiome in Abrolhos. Firstly, the study reveals the potential role of reef fish, specifically Scarus trispinosus and Sparisoma axillare parrotfish, in spreading mutually beneficial and potentially harmful bacteria. However, due to the drastic reduction in these fish populations as a result of overfishing, their role in disseminating these bacteria on the Abrolhos reefs is likely diminished. Given the rising local and global changes threatening the Abrolhos reefs, the management of fisheries must be reviewed to mitigate their negative impact. Secondly, the study identifies a partial overlap in isotopic niches among four fish species, suggesting similarities in feeding. This similarity also implies that microbes might play a role in isotopic partitioning within the fish tissues. Though these fish feed from the same resource pool, each species may depend on its unique microbiome for isotopic differentiation. This suggests that gut microbiomes could be driving forces in differentiating fish tissue isotopic signatures. Bacteria associated with fermentation and nutrient absorption within the microbiome could play pivotal roles in host health and metabolism. This study underscores the importance of considering not only the fish feeding but also the influence of their microbiomes on reef ecology and health. These findings pave the way for future research and more effective conservation strategies.",
"introduction": "Introduction Herbivorous fish play a crucial role in maintaining the health of coral reefs [ 5 , 30 , 77 ]. Known as the primary regulators of algae cover, these fish can eradicate over 90% of daily algae and turf production in shallow coral reefs [ 8 ]. Overfishing, however, has led to a decline in herbivores, promoting the growth of fast-spreading organisms like algae that dominate the sea floor, a phenomenon referred to as the phase-shift process [ 47 ]. Anthropogenic stress, i.e., human-induced nutrient pollution, threatens coral reefs by causing endogenous microbial shifts in wild reef fish [ 19 ]. Coral reef herbivory is complex, involving an array of primary consumers whose feeding habits, diets, morphology, physiology, and ecological functions greatly vary [ 17 , 48 ]. Evolutionary developments like gizzard-like stomachs and bristle-like teeth have enabled herbivorous fish to feed on specific producers of coral reefs [ 63 , 77 ]. Diet strongly influences the gut microbiome of surgeonfishes (family: Acanthuridae) [ 55 ]. For instance, while some Acanthurus fish, as A. chirurgus , feed mainly on detritus, macroalgae, turf, and invertebrates, and are considered sediment suckers [ 28 , 77 ], some parrotfish, as Scarus trispinosus and Sparisoma axillare , consume corals, crustose coralline algae (CCAs), turf, epilithic, and endolithic Cyanobacteria, and microalgae [ 14 , 32 , 34 ]. Macroalgae is not their primary food source but is consumed incidentally in parrotfishes [ 14 ]. Both parrotfish and Acanthurus species significantly influence the dynamics of detritus and sediment [ 77 ]. Kyphosus feed mainly on macroalgae [ 9 ] with gut and mouth structures well adapted for this diet. Although they also consume invertebrates such as calanoid copepods, likely opportunistically, their feeding behavior varies with the age and size of the fish [ 72 ]. Parrotfish and some grazing acanthurids as A. chirurgus use a mechanical grinding process in their pharyngeal mill or gizzard-like stomach, respectively, relying mainly on swift gut throughput and protein, lipid, and soluble carbohydrate digestion [ 14 ]. In contrast, kyphosids have a longer gut that may rely on refractory carbohydrate fermentation [ 56 ]. Acanthurus species have acidic stomachs, ideal for digesting tough algal material. Algae-derived carbohydrate fermentation is a key energy source, producing short-chain fatty acid (SCFA) [ 45 ]. SCFA production can also occur in the final part of the intestine (hindgut chamber or caecal pouch) in Kyphosus vaigienses [ 67 ]. Past studies in Lizard Island, Australia, show that K. vaigiensis predominantly consumes phaeophytes, while K. cinerascens eats a substantial amount of rhodophytes [ 13 ]. The bacteroidota genera Alistipes and Rikenella are found abundantly in the furthest part of the lumen section, where SCFA levels are at their peak. These bacteria play a crucial role in breaking down seaweed into compounds beneficial for the fish [ 76 ]. In addition, metagenomic analyses of K. sydneyanus hindgut contents have revealed the degradation pathways for vital algae dietary substrates like mannitol, alginate, laminarin, fucoidan and galactan, agar, and carrageenan [ 76 ]. Moreover, a noticeable uptake of fermentation products such as acetate in K. sydneyanus hindgut was observed [ 56 ]. These variations in fermentation by-products may explain the isotopic differentiation seen in fish tissue. The gut of fish is known to house many microbial cells (1.71 × 109 g −1 dry wt feces; [ 73 ]). A recent study of the Acanthurus gut microbiome discovered the presence of Lachnospiraceae, a microbe thought to facilitate the digestion and absorption of carbohydrate-dense brown macroalgae [ 71 ]. The gut microbiome of seagrass-specializing parrotfish ( Calotomus spinidens ) in Fiji’s reef areas and surgeonfish ( Acanthurus nigricauda , Ctenochaetus striatus ) in the Florida Keys was predominantly Proteobacteria [ 52 ]. In contrast, the gut of Paracirrhites bicolor , P. arcatus , P. xanthus , and P. nisus fish observed in the Line Islands of the Republic of Kiribati was teeming with Firmicutes [ 37 ]. It is important to note that the Paracirrhites species consists of piscivores/invertivores [ 84 ], which have very different diets compared to the herbivorous studied here. Notably, Acanthurus triostegus from Australia’s Great Barrier Reef had a high quantity of Firmicutes , Epulopiscium [ 60 ]. This microbe also held dominance in the gut of seven herbivorous coral reef fish varieties in the South China Sea [ 38 ]. Regrettably, data detailing the microbiome composition, fermentation triggers, and possible symbionts of herbivores inhabiting Southwestern Atlantic reefs remains scarce. Fish contribute to reef fertilization through a steady supply of microbe-rich feces [ 40 ]. An example is the parrotfish cyanobacteria diet that enriches the coral ecosystem. Besides that, parrotfishes have the potential to disperse Symbiodinium, the photosynthetic microalgae indispensable for stony coral survival [ 10 , 42 , 58 ]. Corallivorous fish, such as parrotfish, may affect the health of coral by sharing microbial symbionts by biting coral and distributing their waste across reefs [ 66 ]. Thus, parrotfish corallivory could be a beneficial process for disseminating helpful bacteria over reefs. This activity has been linked to the transmission of mutualistic bacteria [ 41 ]. Nevertheless, parrotfish may also contribute to reef bioerosion and sediment production and movement [ 51 , 62 ]. Herbivorous fish, including Scarus and Acanthurus , consume cyanobacterial turf and feces of planktivorous fish [ 65 ]. Coprophagy, or feces-eating, apparently provides significant nutrient and energy intake [ 65 ]. Additionally, climate change could trigger the growth of potentially disease-causing bacteria in reef fish [ 25 ]. However, the composition of the feces of Southwestern Atlantic coral reef fish remains largely unknown. Metagenomics (eDNA) analysis of gut contents infers that herbivorous reef fish may have a varied diet and diversified feeding behavior [ 57 ]. In one recent study conducted in an upwelling region, Acanthurus chirurgus and Sparisoma axillare shared some similarities in diets comprised of red calcareous articulated algae, red corticated algae, detritus, and diatoms [ 9 ]. These fish also consumed other common food items like Cyanobacteria, red and green filamentous algae [ 9 ]. The δ 15 N signatures of their food sources, potentially including macroalgae, suspended particulate matter, and sediment (ranges from 6 to 7‰), were consistent with those found earlier in macroalgae (6.1‰), suspended particulate matter (6.2 − 6.5‰), and sediment organic matter (6.4 − 7.1‰). However, despite having similarities in diets and gut contents, A. chirurgus (δ 13 C: − 18.7 ± 0.3‰; δ 15 N: 12.3 ± 0.2‰) and Sp. axillare (δ 13 C: − 16.0 ± 0.1‰; δ 15 N: 10.8 ± 0.2‰) displayed different isotopic signatures, possibly due to differences in their genetics and/or microbiomes, such as through microbial fermentation [ 9 ]. Different microbiomes may have different fermentation pathways and therefore different fermentation products. The four species belong to different genera and have distinct intestinal features, leading to unique feeding biologies. Nevertheless, they all share the common trait of being nominally herbivorous. These two herbivorous fish species ( A. chirurgus and Sp. axillare ) feed on many types of coral reef substrates, suggesting similar microbial compositions. However, additional research on fish microbiomes is required to validate this hypothesis. In the Abrolhos Reef systems of the Southwestern Atlantic Ocean (Brazil), there are four nominally herbivorous fish: Scarus trispinosus , Sp. axillare , Acanthurus chirurgus , and Kyphosus sp. [ 31 – 33 , 35 ]. Among these, the greenbeak parrotfish ( Sc. trispinosus ) is the largest, while the gray parrotfish ( Sp. axillare ) is the smallest. Individuals of Sc. trispinosus larger than 30 cm are excavators [ 51 ], biting corals, turf, and CCAs, probably searching for epilithic and endolithic Cyanobacteria, and microalgae (see [ 14 ]), while individuals smaller than 30 cm are scrapers of the same type of substrates [ 28 , 32 , 34 , 51 ]. In contrast, individuals of Sp. axillare are browsers, biting mainly turf and macroalgae [ 9 , 28 , 32 , 34 ]. Over the past decade, the abundance of these four reef fish species has declined in Abrolhos, coinciding with noticeable phase shifts [ 7 , 31 ]. Following the sharp decline of large carnivorous reef fishes, parrotfishes were progressively targeted by commercial fisheries in Brazil, resulting in a global population decline of 50% for S. trispinosus [ 36 ]. A significant development related to this is the increase in turf algae, which covers over 60% of the benthic cover in some areas of the Abrolhos Bank [ 7 , 81 ]. Along with the coral reef phase shift, there has been a significant increase in the proportion of fermentative bacteria ( Rikenella , Akkermansia , Desulfovibrio , Brachyspira ) in the gut of reef fishes particularly in the Siganidae family present in the Indian Ocean [ 12 ]. Consequently, the changes in the abundance of herbivorous fish also impact the coral reef benthic communities [ 61 , 68 ]. However, more information is needed regarding the composition of the microbiome and potential symbionts of the herbivorous fish in Abrolhos. This study sought to investigate the microbiome composition of the four nominally herbivorous reef fish of the Abrolhos Bank. We examined microbial diversity (using 16S rRNA Illumina sequencing) from various sections of the fish’s gut. We also analyzed the isotopic composition of carbon (δ 13 C) and nitrogen (δ 15 N) in the fish tissue and gut content, which helped to identify potential food sources. The gut anatomy of Sc. trispinosus and Sp. axillare is homogeneous, naturally with anatomy composed of three major sections (stomach absent). On the other hand, Khyphosus sp. and A. chirurgus have five distinct sections in their gut. However, one aspect that remains unclear is how the microbiome could vary throughout their gut.",
"discussion": "Results and Discussion Microbiome Diversity and Temporal Dynamics in Herbivorous Fish Species from the Abrolhos Coral Reefs We generated a total of 16.80 million 16S rRNA sequences across four fish species, resulting in an average of 317,047 ± 57,007 sequences per library ( n = 53 libraries) (Supplementary Table 1 ). Proteobacteria, Firmicutes , and Cyanobacteria made up about 90% of the microbiomes of A. chirurgus and Sp. axillare (Fig. 3 A). However, each fish species displayed a unique microbiome. Specifically, A. chirurgus and Sp. axillare demonstrated higher quantities of Ruminococcaceae. On the other hand, Sc. trispinosus exhibited an increased abundance of Alteromonadaceae and Fusobacteriaceae, while Kyphosus sp. showed greater amounts of Pirellulaceae and Rhodobacteraceae (Fig. 3 B). Apart from this, we also observed temporal changes in the gut microbe compositions of the fish taxa between february 2016 and october 2017 (Fig. 4 ). These annual variations in microbiome profiles within the same species may be attributed to changes in diet due to seasonal food availability. For example, Kyphosus vaigiensis generally consumes Sargassum in the summer and switches to Dictyota , Plocamium , and Gelidium during winter [ 9 ]. Fig. 4 Major microbial families in Acanthurus chirurgus and Kyphosus sp. ( A ), with five divisions of the intestinal tract ( A ), and bellow, Sparisoma axillare with Scarus trispinosus ( B ), having three natural divisions of tract ( B ), in 2016 and in 2017 Microbiomes Reveal a Shared Pool of Food Items Among Herbivorous Fish All four studied fish species exhibited identifiable sequences of Dictyopteris undulata in their digestive systems, although A. chirurgus presented the highest amount. The finding of this brown alga’s sequences in the fishes’ systems implies active herbivory. Also, Cyanobacteria ranging between 1 and 2% were mainly composed of Cyanobium PCC-6307 in Kyphosus and Acanthurus species (Fig. 3 C). Other varieties such as Trichodesmium IMS101, Schizothrix LEGE 07164, Cylindrospermopsis CRJ1, Synechococcus PCC-7336, Phormidium MBIC10003, Synechococcus CC9902, Cyanothece sp. WH 8902, and Arthrospira PCC-7345 were also detected. These Cyanobacteria might be elements of turfs [ 81 ]. The higher presence of Cyanobacteria in A. chirurgus and Kyphosus sp. could signal pronounced herbivory over turf. Concurrently, families like Ruminococcaceae were more prevalent in A. chirurgus and Sp. axillare , while Vibrionaceae, Alteromonadaceae, and Fusobacterium were more present in Sc. trispinosus . CFUs measured on marine agar revealed greater counts in A. chirurgus and Kyphosus sp. compared to Sc. trispinosus and Sp. axillare . This could underpin that A. chirurgus and Kyphosus sp. might depend on microbial fermentation. Both Khyphosus sp. and A. chirurgus exhibit five distinct gut portions, while A. chirurgus has also a gizzard-like stomach. This higher anatomical complexity offers further backing for the fermentation hypothesis [ 17 ]. Contribution of the Gut Microbiome to Fish Health The midgut section of Kyphosus sp. showed the highest richness and diversity (Chao1 mean = 161.550 and Shannon mean = 4.101), followed by the midgut (Chao1 mean = 110.833 and Shannon = 3.928) and the gizzard-like stomach (Chao1 mean = 106.000 and Shannon = 3.888) of A. chirurgus. Finally, in the Simpson index, which evaluates dominant taxa of the microbiome, the midgut of Kyphosus sp. is the most dominant (Simpson mean = 0.963), followed by the end midgut of A. chirurgus (Simpson mean = 0.961) and the foregut of Kyphosus sp. (Simpson mean = 0.952) (Fig. 7 C). Despite observed fluctuations in the abundance of gut microbiome components across different fish specimens and years, it is evident that each fish species has a unique microbiome profile (Fig. 5 ). To further understand the relative significance of different influencing factors on microbial evolution, we used a null model based on βNTI and RC-Bray metrics to study the roles of both stochastic processes (such as dispersal limitation, homogenizing dispersal, and ecological drift) and deterministic processes (like homogeneous and heterogeneous selection) [ 75 ]. The findings indicate that these forces are primarily driven by stochastic processes (Fig. 6 ). Specifically, among the four fish species, dispersal limitation exerted the most influence (52.94%) on microbial evolution. This suggests that each fish species co-diversifies with specific microbiomes. The most prevalent microbiome components are potentially stable symbionts, which have co-diversified to offer reciprocal benefits to the host and the microbiome. Fig. 5 Non-metric Multidimensional Scaling (NMDS) with Bray–Curtis similarity index of gut content microbiomes. Approximately 440 family types contribute to form the dimensions of the graph with the Bray–Curtis index. The two specimens (gathered per year) of each fish, collected in 2016 and 2017, demonstrate that the Acanthurus chirurgus microbial group is closer to Sparisoma axillare , while the other two species have different patterns Fig. 6 Evolutionary driving forces in fish microbiomes The gut microbiome of a host significantly influences the immune system by stimulating the development of gut cells/tissues, the gut circulatory system, and the immune system itself [ 2 , 74 ]. Gut microbes can generate a broad range of secondary metabolites, such as hormone precursors like tryptophan/serotonin, vitamins like B12, SCFAs like butyric, lactic, and propionic acids, 4-indolecarbaldehyde and L-3-phenyllactic acid. All of these contribute significantly to the host’s metabolism. Certain bacterial groups, such as Prevotella and Bifidobacterium , may influence nutrient production through fermentation, absorption, and body weight gain [ 79 ]. The Proteobacteria Vibrionaceae , Rhodobacteraceae , and Fusobacteria are noted to be abundant in fish guts and are potentially associated with fermentative processes [ 21 , 82 , 83 ]. Clostridia might also play a role in the fermentative digestion process in species like Kyphosus sp., Sc. trispinosus , and A. chirurgus [ 16 , 29 , 67 ]. Rhodobacteria may help in the assimilation of bile and cholesterol [ 1 , 69 ], while Fusobacteria produces butyrate [ 3 ], a short-chain fatty acid often associated with carbohydrate fermentation, including that found in coral mucus [ 78 , 80 ]. Butyrate, common in the intestines of herbivorous and omnivorous fish [ 15 , 17 ], may inhibit the growth of potential fish pathogens [ 59 ]. Scarus trispinosus is known to prey on corals, including Montastraea cavernosa and Mussismilia [ 34 , 35 ]. Notably, these corals have an abundant source of mucus composed of sulfated glycoproteins, sugars like fucose, and mucin proteins. This study demonstrates that Sc. trispinosus and Sp. axillare may have a beneficial effect on coral reef health by dispersing potential mutualistic bacteria across the reef (Fig. 7 A). However, these fish might also spread potentially harmful bacteria (Fig. 7 B). Both Kyphosus sp. and A. chirurgus showed a reduced potential for spreading bacteria due to the lower occurrence of either mutualists or pathogens in their hindgut (Fig. 7 ). Corallivore feces are likely to contain a high volume of mutualistic bacteria (such as Endozoicomonas , Ruegeria , and Rhyzobia ) and a lower volume of potential coral pathogens (vibrios) in comparison to the feces of algae grazers and detritivorous fish [ 24 , 40 ]. The presence of ammonia-oxidizing Pirellulaceae and a low volume of Vibrionaceae in the gut of Kyphosus sp. suggests a complex microbial community, potentially influenced by the species' diet and habitat. Although these bacterial groups have been previously found in Abrolhos corals [ 27 , 39 ], it is important to note that Kyphosus species are essentially herbivorous, primarily feeding on different types of algae. This indicates that the intestinal microbiome of these fish may also reflect the environmental bacteria present in their surroundings. Fig. 7 Abundance of potentially mutualistic ( A ) and pathogenic ( B ) bacteria for corals across the fish gut. Box plot of alpha diversity metrics (Chao1, Shannon, and Simpson) ( C ) for Acanthurus chirurgus , Kyphosus sp., Scarus trispinosus , and Sparisoma axillare at the prokaryotic genus level. F, foregut; MF, mid foregut; M, midgut; EM end midgut; H, hindgut. Kyphosus sp.: F = foregut/stomach, MF = anterior intestine, M = midgut, EM = end midgut, H = hindgut. A. chirurgus : F = foregut/anterior stomach, MF = gizzard-like stomach, M = anterior intestine, EM = midgut, H = hindgut. Coral mutualistic bacteria ( Ruegeria , Endozoycomonas , and Rhyzobia ). Coral pathogenic bacteria ( Vibrio , Photobacterium , Alteromonas ) Corallivore feces, therefore, might contribute to coral health by supplying mutualistic microbes. Indeed, fish feces have been identified as a hotspot for Symbiodinium [ 42 ]. In corals, the growth of Rhodobacteraceae, Verrucomicrobiaceae, Flavobacteriaceae, Vibrionaceae, Fusobacteriaceae, Campylobacteraceae, and Cohaesibacteraceae is triggered by fish feces [ 23 ]. Fish that prey on corals showed an increase in potentially beneficial bacteria (such as Oceanospirillum and Ruegeria ) and a decrease in opportunistic bacteria (such as Flammeovirgaceae, Rhodobacteraceae, Rhodospiralles, Glaciecola ) [ 24 ]. The microbiome of the facultative coral-feeding butterflyfish ( Chaetodon capistratus ) varies across Caribbean reefs [ 18 ]. In degraded reefs (with low coral cover), microbiomes were more heterogeneous, enriched with fermentative bacteria and vibrios, and had fewer potential coral mutualists like Endozoicomonas [ 18 ]. Analysis of the foregut, midgut, and hindgut facilitated the assessment of potential variations in the microbiome composition throughout the entire fish gut (Fig. 7 ). Three microbial groups were differentiated within the Abrolhos fish gut microbiomes: (i) stable microbes (e.g., Alteromonas , Vibrio , and Photobacterium in Kyphosus ), (ii) microbes with increased presence (e.g., Ruegeria , Endozoycomonas , and Rhyzobia in Scarus and Sparissoma ), and (iii) microbes with decreased presence (e.g., Ruegeria , Endozoycomonas , and Rhyzobia in Kyphosus sp.) (Fig. 7 ). The observed variation in certain microbiome members suggests that the fish gut could act as a fermenter, encouraging the proliferation of specific microbial types (such as Rhodobacteraceae ) and/or eliminating other bacteria throughout the gut. To further investigate the influence of food sources on the fish microbiome, isotopic analysis was conducted. The Isotopic Signature of the Gut Contents Supports Similarities in Feeding Among the Four Herbivorous Reef Fish We obtained 44 isotopic signatures in total, which include gut content (20), fish muscular gut tissue (20), and food sources – namely, one rhodolith, one turf, and two corals (Table 1 ). The food sources composition was (δ13C and δ15N, ‰): Montastraea cavernosa (− 32.9 and 1.4), Mussismilia braziliensis (− 25.1 and 1.2), turf (− 19.1 and 1.8), rhodolith (− 12.8 and 5.0). Rhodolith exhibited the highest δ 13 C and δ 15 N values, while the Mo. cavernosa and Mu. braziliensis corals recorded the lowest. The δ 13 C values of fish gut content lay between − 23.1 and − 12.3‰. Observably, the isotopic signatures of Kyphosus sp. and Sp. axillare were different. Likewise, Sparisoma ’s δ 13 C values and total tissue carbon levels differed significantly from those of Scarus and Acanthurus . We noted a significant variation in the gut content δ 13 C of A. chirurgus and Kyphosus sp. (ANOVA, F = 4.54, p -value = 0.017). The δ 13 C range for fish tissue was − 20.9 to − 10.2‰, with Kyphosus sp. showing a substantial difference in its tissue content δ 13 C compared to other fish (ANOVA, F = 15.74, p -value < 0.001).\n Table 1 Taxa, tissue, number of samples ( n ), mean ± standard deviation, and enrichment factors (Ɛ) of δ 13 C and δ 15 N of Sparisoma axillare , Scarus trispinosus , Acanthurus chirurgus , and Kyphosus sp. at Abrolhos Bank, Southwestern Atlantic. Significant interspecific variations within a given tissue and intraspecific variations among different tissues are denoted by uppercase and lowercase letters, respectively ( p -value < 0.05). * p -values refer to the differences between gut content and tissue for each fish species (δ 13 C/δ 15 N) Tissue n δ 13 C (‰) δ 15 N (‰) Ɛ δ 13 C (‰) Ɛ δ 15 N (‰) p -value* Kyphosus sp. Gut content 8 -14.4 ± 0.9 Aa 3.0 ± 0.8 Aa 0.07/0.000 Tissue 8 -12.4 ± 0.8 Ab 5.9 ± 0.8 Ab 1.8 2.9 Acanthurus chirurgus Gut content 3 -19.8 ± 3.2 Ba 2.7 ± 0.3 Aa 0.03/0.005 Tissue 3 -18.5 ± 3.7 Ba 4.3 ± 0.5 Bb 1.0 1.6 Scarus trispinosus Gut content 5 -17.9 ± 4.1 Aa 2.7 ± 0.5 Aa 0.01/0.000 Tissue 5 -16.5 ± 1.1 BCa 4.0 ± 0.3 Bb 1.1 1.3 Sparisoma axillare Gut content 4 -15.1 ± 1.4 Aa 2.9 ± 0.7 Aa 0.14/0.007 Tissue 4 -14.4 ± 1.8 Ca 4.1 ± 0.2 Bb 0.5 1.2 The gut content δ 15 N varied between 1.7 and 4.0‰ and showed no significant difference among the four herbivorous fish (ANOVA, F = 0.26, p -value = 0.856). On the other hand, the fish tissue content δ 15 N ranged from 3.8 to 6.9‰. Kyphosus sp. possessed significantly higher δ 15 N compared to the other three fish species (ANOVA, F = 14.27, p -value < 0.001) (Table 1 ). It can be noted that Kyphosus’ s habits distances itself from the other fish. Previous studies demonstrate seasonal shifts in diets are relevant. For instance, in summer, fish species are more generalist in their feeding characteristics. [ 20 ]. Remarkably, Kyphosus sp. was the only species displaying a significant difference in δ 13 C between gut and tissue contents (ANOVA, F = 19.36, p -value = 0.001). Comparable δ 13 C values between gut and tissue contents were noted in A. chirurgus (ANOVA, F = 0.22, p -value = 0.665), Sc. Trispinosus (ANOVA, F = 0.57, p -value = 0.470), and Sp. axillare (ANOVA, F = 0.38, p -value = 0.558). There was a higher δ 15 N in fish tissue relative to gut content for all fish species (Table 1 ; Kyphosus sp. – ANOVA, F = 3.76, p -value < 0.001; A. chirurgus – ANOVA, F = 2.49, p -value = 0.011; Sc. Trispinosus – ANOVA, F = 1.99, p -value = 0.001; Sp. axillare – ANOVA, F = 1.98, p -value = 0.015). The tissue and gut contents varied significantly in terms of δ 13 C and δ 15 N for all four species (Table 1 ). The highest discriminant factors were found in Kyphosus sp. (1.8 and 2.9‰, respectively), whereas Sp. axillare exhibited the lowest (0.5 and 1.2‰, respectively). Some items in the diet of herbivorous fish, such as turf, brown algae, rodoliths, and corals, has been detailed (Fig. 8 ). As derived from the SIBER analysis, there was considerable overlap (> 64%) in gut contents between Sc. Trispinosus and A. chirurgus . This, along with similar CR, suggests that their food source is mainly turf (Fig. 8 , Tables 2 and 3 ). Of note, Sc. Trispinosus demonstrated the most diverse and uneven diet among herbivorous fish, as evidenced by the highest values for CR, TA, SEAc, MNND, and SDNND in gut content (Table 3 ). Fig. 8 The standard ellipse area corrected for small samples (SEAc) of gut muscular tissue and gut content (microbiome fraction), both specified on Supplementary Table 2 , for the four herbivorous fishes Sparisoma axillare , Scarus trispinosus , Acanthurus chirurgus , and Kyphosus sp. for each specimen ( n ) at Abrolhos Bank, Southwestern Atlantic Table 2 Overlapping niche (SEAc, %) among reef fishes. Values are gut contents/tissue Kyphosus sp. A. chirurgus S. trispinosus S. axillare Gut contents Kyphosus sp. -/- -/- 45.9/- 60.4/- Acanthurus chirurgus -/- -/- 79.5/- -/- Scarus trispinosus 13.6/- 64.5/- -/- 27.3/22.4 Sparisoma axillare 40.4/- -/- 61.4/25.9 -/- Table 3 Isotopic niche metrics. Significant interspecific variations within a given tissue and intraspecific variations among different tissues are denoted by uppercase and lowercase letters, respectively ( p -value < 0.05). Abbreviations definitions (NR, CR, TA, SEAc, CD, MNND, and SDNND) in the text. Values are gut contents/gut muscle tissue NR (‰) CR (‰) TA (‰) SEAc (‰ 2 ) CD (‰) MNND (‰) SDNND (‰) Kyphosus sp. 2.2/2.3 2.4/2.2 2.8/3.0 2.5/2.3 1.0 Aa /1.0 Aa 0.5 Aa /0.5 Aa 0.4 Aa /0.2 Ba Acanthurus chirurgus 0.6/0.8 6.1/6.9 1.8/0.6 6.9/2.3 2.2 Aa /2.6 Ba 3.0 Ba /3.0 Ba 0.2 Aa /1.6 Ab Scarus trispinosus 1.1/0.9 9.9/2.6 5.5/1.2 8.5/1.4 2.7 Aa /0.8 Aa 2.0 ABa /0.7 Aa 2.8 Ba /0.2 Bb Sparisoma axillare 1.2/0.3 3.3/3.7 2.2/0.6 3.7/1.2 1.2 Aa /1.3 ABa 1.5 ABa /1.0 Aa 0.3 Aa /0.7 Aa In contrast, Kyphosus sp. and Sp. axillare showed isotopic niche overlaps between 40 to 60%, indicating mutual reliance on a food source associated to turf and rodolith, backed by high NR values. As uncovered from the fish tissue’s isotopic profiles, A. chirurgus and Kyphosus sp. displayed distinct ellipses in their isotopic niches, indicating no overlap with other species. This implies unique assimilation patterns. On the other hand, Sp. axillare and Sc. Trispinosus exhibited overlap between 22 and 26%. Traces of M. braziliensis coral, turf, and rhodoliths consumption were found in A. chirurgus . In contrast, Kyphosus sp. appeared to favor resources associated to rodolith, demonstrated by the lower CD values and the lowest MNND and SDNND values. A recent study established that the isotopic signatures of A. chirurgus (δ 13 C: − 18.7 ± 0.3‰; δ 15 N: 12.3 ± 0.2‰) and Sp. axillare (δ 13 C: − 16.0 ± 0.1‰; δ 15 N: 10.8 ± 0.2‰) varied in an upwelling subtropical region [ 9 ]. However, this study found both fish species to have similar nitrogen isotopic signatures, A. chirurgus had a mean ± standard deviation (stdev) of δ 13 C − 18.5 ± 3.7‰ and δ 15 N 4.3 ± 0.5‰, while S. axillare showed δ 13 C − 14.4 ± 1.8‰ and δ 15 N 4.1 ± 0.2‰. The decreased δ 15 N in A. chirurgus tissue from Abrolhos tropical reefs compared to the upwelling region could potentially be due to higher nitrogen levels in upwelling food items. The gut content’s δ 13 C and δ 15 N signatures were similar among the four species in Abrolhos. Even though these herbivorous fish consume some similar food sources in Abrolhos, differences in their microbiomes, host genetic backgrounds, and metabolic rates may contribute to the diversity in nutrient assimilation, resulting in varying tissue isotopic signatures."
} | 8,075 |
39149420 | PMC11324577 | pmc | 5,633 | {
"abstract": "Bacterial biofilms are organized heterogeneous assemblages of microbial cells encased within a self-produced matrix of exopolysaccharides, extracellular DNA and proteins. Over the last decade, more and more biofilm-associated proteins have been discovered and investigated. Furthermore, omics techniques such as transcriptomes, proteomes also play important roles in identifying new biofilm-associated genes or proteins. However, those important data have been uploaded separately to various databases, which creates obstacles for biofilm researchers to have a comprehensive access to these data. In this work, we constructed BBSdb, a state-of-the-art open resource of bacterial biofilm-associated protein. It includes 48 different bacteria species, 105 transcriptome datasets, 21 proteome datasets, 1205 experimental samples, 57,823 differentially expressed genes (DEGs), 13,605 differentially expressed proteins (DEPs), 1,930 ‘Top 5% differentially expressed genes’, 444 ‘Threshold-based DEGs’ and a predictor for prediction of biofilm-associated protein. In addition, 1,781 biofilm-associated proteins, including annotation and sequences, were extracted from 942 articles and public databases via text-mining analysis. We used E. coli as an example to represent how to explore potential biofilm-associated proteins in bacteria. We believe that this study will be of broad interest to researchers in field of bacteria, especially biofilms, which are involved in bacterial growth, pathogenicity, and drug resistance. Availability and implementation: The BBSdb is freely available at http://124.222.145.44/#!/ .",
"introduction": "1 Introduction Bacterial biofilms are adhesion structure formed by single or multiple bacteria and their metabolites. In clinical practice, biofilms can greatly improve the ability of pathogenic bacteria to resist antibiotics, thus increasing the risk of infection ( Jamal et al., 2018 ; Schwarzer et al., 2020 ). Biofilm-associated proteins are defined as a type of protein molecules closely related to bacterial biofilm formation, which include constitutive proteins located downstream of the biofilm regulatory network and upstream transcriptional regulators. The understanding and discovery of biofilm-associated genes and proteins can help us to better understand the molecular mechanisms of bacterial biofilm formation. Over the last decade, more and more biofilm-associated genes and proteins have been discovered and investigated with the development of omics techniques including transcriptomes and proteomes ( Lasaro et al., 2009 ; Hay and Zhu, 2015 ; Wang et al., 2017 ; Jia et al., 2022 ). During the development and formation of biofilm, the transcription profile of bacteria changes, and some genes with obviously variable expression levels, which are proved by previous experiments, often play an important role in biofilm formation. Therefore, it is important for researchers to obtain these data and analyze gene and protein expression profile in the background of biofilms. However, those important transcriptome and proteome data have been uploaded separately to various databases, which make biofilm researchers pain to have a comprehensive access to these data. Although several resources provide biofilm data, such as Quorumpeps ( Wynendaele et al., 2013 ) for QS-derived signaling peptides, BiofOmics ( Lourenço et al., 2012 ) for biofilm experimental information, BaAMPs ( Di Luca et al., 2015 ), aBiofilm ( Rajput et al., 2018 ), dpABB ( Sharma et al., 2016 ) for antibiofilm Agents, BSD8 ( Urbance et al., 2020 ) for structural information. There is an urgent need to combine multi-omics data for the prediction and analysis of biofilm-associated proteins. Here, we developed BBSdb, an online database focusing on experimentally validated biofilm-associated proteins. In addition, BBSdb provided a predictor for prediction of biofilm-associated protein, in which users could upload their interested protein sequence to predict candidate biofilm-associated proteins and browse corresponding entries of DEG. BBSdb can serve as a useful resource to make researchers pain-free to obtain transcriptomes, proteomes in biofilm research, query information of experimentally validated biofilm-associated proteins, and utilize developed predictor for protein prediction.",
"discussion": "4 Discussion In this work, we provided a comprehensive database for biofilm research, which could be an infrastructure for the biofilm research community. In addition, we developed a predictor assisting researchers to explore potential biofilm-associated proteins. However, several questions remain to be addressed. For example, the BBSdb database only collected bacterial proteomes through \n Supplementary Materials \n in literature without collecting, processing, and analyzing the original bacterial proteome data in public databases such as PRIDE ( Perez-Riverol et al., 2022 ). Regarding the existence of multiple post-transcriptional regulation, proteomes will provide useful information for researchers, therefore more proteomes should also be integrated in the BBSdb. Biofilm-associated proteins are participating in the biofilm formation process of their respective bacteria referring to previous research ( Lasa and Penadés, 2006 ). In this study, we defined biofilm-associated protein as those that have been reported in the literature and validated experimentally as components of the biofilm structure or involved in the regulation of biofilm development. In practical applications, we need to take into account the differences between different biofilm model systems and many other factors, such as the duration of the experiment and the specific strain, in order to obtain condition-dependent biofilm-associated proteins ( Edel et al., 2019 ; Flemming et al., 2023 ). These proteins can effectively explain the diversity of mechanisms in bacterial biofilm development. In addition, homologous genes and proteins of different bacteria species can perform different functions, it is necessary to develop a bacteria-specific predictive model for better guiding practice in the next release. In the future, other questions also remain to be addressed, including the long-term maintenance and update of BBSdb and improvement of predictive ability upon big data accumulation."
} | 1,571 |
35757877 | PMC9234814 | pmc | 5,634 | {
"abstract": "Global warming, habitat loss and overexploitation of limited resources are leading to alarming biodiversity declines. Ecosystems are complex adaptive systems that display multiple alternative states and can shift from one to another in abrupt ways. Some of these tipping points have been identified and predicted by mathematical and computational models. Moreover, multiple scales are involved and potential mitigation or intervention scenarios are tied to particular levels of complexity, from cells to human–environment coupled systems. In dealing with a biosphere where humans are part of a complex, endangered ecological network, novel theoretical and engineering approaches need to be considered. At the centre of most research efforts is biodiversity, which is essential to maintain community resilience and ecosystem services. What can be done to mitigate, counterbalance or prevent tipping points? Using a 30-year window, we explore recent approaches to sense, preserve and restore ecosystem resilience as well as a number of proposed interventions (from afforestation to bioengineering) directed to mitigate or reverse ecosystem collapse. The year 2050 is taken as a representative future horizon that combines a time scale where deep ecological changes will occur and proposed solutions might be effective. This article is part of the theme issue ‘Ecological complexity and the biosphere: the next 30 years’.",
"introduction": "1 . Introduction Over the last decades, a general consensus among scientists from very diverse disciplines has been emerging about the future of our planet and our society (Intergovernmental Panel on Climate Change) and provides a grim picture of how global warming will affect the biosphere in multiple ways and across scales [ 1 ]. Regional, continental and planetary-scale changes are taking place at an accelerated pace. Greenhouse gases are the most obvious example of such a trend, with CO 2 in particular displaying a fast increase that has no equivalent over the past 500 Myr. This rise is a consequence of industrialization and the parallel population growth, particularly in urban areas ( figure 1 a ). By 2050, 70% of humankind will live in cities. Despite the deceleration of this process (largely due to reduced fertility rates and changes in women’s status), the predicted expansion gives a staggering 9.7 billion people. The ultimate reason for this explosive growth has to be found in the mathematics of population dynamics. The historical record of modelling in climate science and conservation studies starts long ago. In many cases, predicted outcomes were tied to theory (either mathematical or computational) that would help quantify future scenarios of change, decay and recovery [ 4 , 5 ]. A common goal (and a nontrivial problem) in all these approaches is prediction.\n Figure 1 . Ecological complexity challenges for 2050. With the rise of global temperatures, population growth and the resulting pressure on resources and habitats, biodiversity will face major threats. One crucial role of science is to develop reliable predictions of future trends. Here, four examples are chosen (left) along with current forecasts (central column, estimated 2050 states indicated with a red circle) and examples of the complex systems approaches used (right). ( a ) Urban centres (image of Central Park, New York, by Ajay Suresh, Creative Commons) are rapidly expanding as massive migrations occur towards cities. Human population growth (centre) is slowly decelerating, but two extra billion humans will be added to the current numbers, reaching 9.7 billion by 2050. The current trend is a consequence of the nonlinearities associated with hyperbolic dynamics, which predicts a singularity at a given finite time t c (right). ( b ) Rainforests (left image by Gleilson Miranda, Creative Commons) are experiencing rapid loss and fragmentation of their habitats, with predicted critical points (centre plot, grey bar, see [ 2 ]) to be reached in a few decades. These critical points correspond to percolation thresholds (right panel). ( c ) Drylands (image courtesy of David Huber) are expanding and will grow from the current 40% to more than 50% in just three decades. Models of drylands involving vegetation cover as a key variable predict sharp transitions between alternative states, connected through three different shifts [ 3 ]. Here two of them are indicated. ( d ) Marine ecosystems, and coral reefs (left image by Toby Hudson, Creative Commmons) in particular, are being affected by warming ocean temperatures, eutrophication, pathogens and overfishing. Reef cover is rapidly shrinking and might experience massive decays in the next decades. Here, the previous and predicted time series of coral reef cover in Hawaii is shown (centre, data from https://19january2017snapshot.epa.gov/cira/climate-action-benefits-coral-reefs_.html ). Multiple alternative states have been identified (right) with different sources of stress causing jumps from one state to another. Historical examples of long-term prediction include the famous 1972s The limits to growth report that was intended to present the first long-term simulation of economic and population growth [ 6 ]. It involved a simplified description of human systems and their interactions with a world with finite resources. The model efforts, led by Donella Meadows, incorporated several key variables known to grow with time, including human population, food production, industrialization, pollution and consumption of non-renewable natural resources. The methodology was inspired by the work of Jay Forrester [ 7 ]. He was a pioneer of so-called Systems Science, a field that takes a complexity view of the world where interactions among many components are treated as simplified, deterministic dynamical systems. The report was cautious about the assumptions and its potential implications: ‘The model we have constructed is, like every other model, imperfect, oversimplified and unfinished’ [ 6 ]. One of the key predictions made by the report is described as follows (pp. 23–24): If the present growth trends in world population, industrialization, pollution, food production and resource depletion continue unchanged, the limits to growth on this planet will be reached sometime within the next one hundred years. The most probable result will be a rather sudden and uncontrollable decline in both population and industrial capacity. Despite all the unknowns, the crucial outcome of the report was clear. Business as usual in a planet with limited resources and a rapidly (exponentially) expanding human population can only end up in unsustainable growth and collapse. A second message from the report sounds familiar nowadays: ‘the trends depicted above could be modified provided that sustainable growth is introduced, in such as way that rational use of resources allows the maintainance of stability while the basic material needs of each person on earth are satisfied and each person has an equal opportunity to realize his individual human potential’ [ 6 , p. 24]. An obvious limitation of this kind of study is the requirement of model simplifications, such as ignoring geography or different sources of fluctuations, along with the inevitable limitations associated with parameter estimation. Most importantly, the use of a small number of variables seems inappropriate when trying to represent the complexity of the real world. The goal was to examine the interactions between the five variables within a two-century window (1900–2100). It thus includes past information that was used to calibrate some of the required parameters. In this way, Meadows’ model became the first integrated global model and inspired a great deal of studies since [ 8 ]. Nowadays, any realistic assessment of the future of the planet requires consideration of the explicit role played by climate. As global warming and an intensive exploitation of planet resources keep rapidly increasing, the analysis of past climates and modelling efforts suggest that future changes can unfold in potentially catastrophic ways [ 9 ]. As far-from-equilibrium, dissipative structures, ecological systems exhibit nonlinear dynamical properties that pervade their stability but are also responsible for their fragility under stress. They are in fact complex adaptive systems (CAS) [ 10 ]. Crucial features of CAS include spatial and temporal heterogeneity, diversity and nonlinearity [ 3 ]. It is in this context that integrative approaches to climate and the biosphere are of fundamental relevance. Wide weather fluctuations, alarming biodiversity declines and social unrest are already here. Future potential tipping points have been identified, while most predicted climate change scenarios seem confirmed and consistent with worst-case outcomes. What can be done to reverse, counterbalance or prevent tipping points? Many different proposals have been suggested based on sustainable growth, restoration strategies and increased clean energy use. But the time scale for effective measures is rapidly shrinking. Confronted with a planet decline where humans are part of a complex, endangered ecological network, novel approaches need to be taken. All these approaches include unsolved, multiscale problems and will need to be applied in a social context dominated by cities, political instability and rising inequality. A complex systems perspective including all key aspects of the problem is required, pointing to an agenda of well-defined alternatives. What are the challenges ahead for the next decades? Using 2050 as a potential time horizon, here we summarize some of the key issues associated with the future of the biosphere under a complex systems perspective. Much has taken place since the publication of The \n limits to growth and the use of models is nowadays widespread. How can humans be included as part of modelling efforts? What kind of information is required to feed these models? What can be safely predicted? Answering these and other fundamental questions was the goal of a workshop hosted by the Santa Fe Institute in 2021. The meeting convened a group of researchers from diverse fields, from theoretical and conservation ecology to synthetic biology. This Theme Issue summarizes several key concepts associated with the nonlinear, complex nature of our biosphere and how these nonlinearities affect future trends. But the year 2050 needs to be seen also as a window to plan for interventions: what can be done from conservation, restoration and engineering?",
"discussion": "5 . Discussion A general consensus among scientists from very diverse disciplines has been emerging about the future of our planet and our society. As global warming and an intensive exploitation of planet resources keeps rapidly increasing, WS indicate that potentially catastrophic transitions will unfold within this century. Wide weather fluctuations, alarming biodiversity declines and social unrest are already here. Predicted climate change scenarios seem confirmed and consistent with worst-case outcomes. What can be done to reverse, counterbalance or prevent tipping points? Many different solutions have been suggested based on sustainable growth, restoration strategies and increased clean energy use. But the time scale for effective measures is rapidly shrinking. Confronted with a planet decline where humans are part of a complex, endangered ecological network, novel approaches need to be taken. All these approaches include unsolved, multiscale problems and will need to be applied in a social context dominated by cities, political instability and rising inequality. A complex systems perspective including all key faces of the problem is required, pointing to an agenda of well-defined alternatives. Changing ecosystems, either following bottom-up (synthetic biology) or top-down (afforestation, geoengineering) approximations needs to be carefully considered, and different strategies are compatible. What is the optimal way of bringing together biodiversity and human interests? As pointed out by Howard Odum, any ecological engineering approach needs to join human design and environmental self-design so that they are mutually symbiotic [ 83 ]. To make this a reality, preserving and fostering biodiversity is a necessary condition. How can we know for sure the state of the biosphere by 2050? Can our ambition of an accurate prediction be fulfilled? As shown by the success of climate science, predictions are not only possible but essential to define strategic mitigation and adaptation roadmaps. The diverse range of proposals discussed here span a range of views that needs to be used as a source of alternative, but complementary solutions. We cannot yet know if 2050 will be characterized by the success of large-scale protection or instead will be (as pointed to in [ 87 ]) dominated by novel ecosystems. As pointed out by the physicist Denis Gabor, predicting the future might be difficult, but we can also think out of the box. Our species has been a too successful ecosystem engineer, transforming a planet where ecosystems are nowadays being dismantled. We face an uncertain future with limited resources exploited by a fast-growing human population and where biodiversity needs to be protected. Biodiversity is central in providing society with the required goods and services to sustain itself [ 132 ]. Action is needed to preserve it while ensuring the well-being of humans. Any future solution will necessarily involve considering the whole range of strategies described by the different contributions of this theme issue. Science will also need citizen awareness of the problems involved and a proper governance. As we write this paper, humanity is moving out from a 2-year pandemic event that is a reminder of the global nature of the Anthropocene [ 133 ]. Dealing with COVID-19 required an enormous collective scientific action that ended up with effective vaccines in a very short time window. But it has also revealed our weaknesses. The reality of climate change and its consequences are upon us and global decisions will be needed again. The complex, nonlinear nature of our biosphere makes it difficult to design simple solutions. New ideas and integrative strategies involving multiple scales will be needed as we keep pushing our understanding of this unique planet and reconsider our future place as a node of the living web."
} | 3,610 |
33536879 | PMC7848953 | pmc | 5,635 | {
"abstract": "Animals rely on internal motivational states to make decisions. The role of motivational salience in decision making is in early stages of mathematical understanding. Here, we propose a reinforcement learning framework that relies on neural networks to learn optimal ongoing behavior for dynamically changing motivation values. First, we show that neural networks implementing Q-learning with motivational salience can navigate in environment with dynamic rewards without adjustments in synaptic strengths when the needs of an agent shift. In this setting, our networks may display elements of addictive behaviors. Second, we use a similar framework in hierarchical manager-agent system to implement a reinforcement learning algorithm with motivation that both infers motivational states and behaves. Finally, we show that, when trained in the Pavlovian conditioning setting, the responses of the neurons in our model resemble previously published neuronal recordings in the ventral pallidum, a basal ganglia structure involved in motivated behaviors. We conclude that motivation allows Q-learning networks to quickly adapt their behavior to conditions when expected reward is modulated by agent’s dynamic needs. Our approach addresses the algorithmic rationale of motivation and makes a step toward better interpretability of behavioral data via inference of motivational dynamics in the brain.",
"introduction": "Introduction Motivational salience is a cognitive process that motivates, or propels, an individual’s behavior toward or away from a particular object, event, or outcome ( Zhang et al., 2009 ). Such process describes an a priori defined “wanting” of an outcome. It regulates behaviors toward particular goals, adjusts the amounts of time and energy that an individual is willing to expend in pursuit of each desired outcome, and sets the acceptable levels of related risk ( Zhang et al., 2009 ; Berridge, 2012 ). Motivational salience, or, as we will call it here for brevity, motivation , describes animals’ a priori desire or aversion to receive a particular outcome, which should be contrasted with liking or disliking of an outcome that is experienced a posteriori . Mathematically, motivation can be viewed as a subjective modulation of the expected value of reward, determined before the reward is received. Behavior-based models of motivation emerged as a part of the broader effort to understand reward-guided behaviors in humans and other animals [reviewed by Miller (2008) ]. Motivational levels in these models were described as the subjects’ drives toward certain outcomes (e.g., appetite and thirst). To estimate the relative dynamics of different drives (not observable in experiment) psychologists offered human or animal subjects to approach/avoid different combinations of stimuli and titrated responses based on the valency/strength of these inputs ( Sears and Hovland, 1941 ; Miller et al., 1943 ). Such “conflict” experiments resulted in detailed models of motivational dynamics. Motivational drives can therefore be viewed as temporarily varying representations of motivational salience. Neuronal correlates of motivation-related variables were discovered in the ventral pallidum (VP). VP is a part of the basal ganglia that receives the inputs from a number of mesocorticolimbic areas ( Kelley et al., 1982 ; Reep and Winans, 1982 ; Fuller et al., 1987 ; Grove, 1988 ; Martinez-Murillo et al., 1988 ; Heimer et al., 1991 ; Maslowski-Cobuzzi and Napier, 1994 ; Maurice et al., 1997 ; Berridge, 2012 ). As the major output of the ventral basal ganglia ( Saper and Loewy, 1980 ; Heimer et al., 1987 ; Mogenson and Yang, 1991 ; Leung and Balleine, 2013 ), it sends substantial projections to the lateral habenula (LHb), dorsal and medial raphe nuclei (DR/MR), ventral tegmental area (VTA), substantia nigra pars compacta and pars reticulata (SNc and SNr), and mediodorsal thalamus ( Haber and Knutson, 2010 ; Richard et al., 2016 ). Thus, the VP is a hub linking areas involved in reward processing with motor output regions, and is anatomically poised to mediate motivated behaviors. Indeed, lesions in the VP induce aphagia and adipsia, the lack of motivation to eat and drink, respectively ( Morgane, 1961 ; Stellar et al., 1979 ; Humphries and Prescott, 2010 ), and anhedonia, an inability to feel pleasure ( Berridge, 1996 ). An intact VP is also necessary for drug seeking behaviors ( McFarland and Kalivas, 2001 ; Harvey et al., 2002 ; Miller et al., 2006 ; Vijayaraghavan et al., 2008 ) and for active avoidance and aversive learning ( Ishihara et al., 1991 ; Page et al., 1991 ; Root et al., 2013 ). Human brain imaging studies indicate that the VP activities correlate with motivational vigor ( Pessiglione et al., 2007 ; Root, 2013 ; Singh-Bains et al., 2016 ). In vivo single unit recording studies in rodents and monkeys indicate that the VP neuron firing correlates with motivational salience ( Berridge and Schulkin, 1989 ; Tindell et al., 2004 ; Smith and Berridge, 2007 ; Tachibana and Hikosaka, 2012 ; Jiang et al., 2015 ; Richard et al., 2016 ). In the experiments in which sodium starvation was introduced in rats, the responses of the VP neurons to the conditioned stimulus (CS) associated with normally aversive sodium stimulus have changed to match those to the CS associated with normally attractive sucrose ( Berridge, 2012 ). These observations suggest that the VP is critically involved in “positive motivation,” including the “liking” (the pleasurable impact of reward consumption) and the “wanting” (the attractiveness of a stimuli or incentive salience) aspects of behaviors ( Berridge and Schulkin, 1989 ). It may also be involved in “negative motivation”, the drive to avoid aversive stimuli. In this study, we investigated the circuit mechanism of representation of motivational information in VP networks. Computational models for motivated behaviors are best represented by reinforcement learning (RL) models. RL is the area of machine learning and artificial intelligence that deals with the strategies that rational agents can employ while navigating in an environment to maximize future rewards ( Sutton and Barto, 1998 ; Zhang et al., 2009 ). As such, RL models are successful in predicting and explaining adaptive choice behaviors in both human and animals, and have been successful in predicting the causal changes in neuronal responses ( Schultz et al., 1997 ; Schultz, 1998 ; Dayan and Abbott, 2001 ; Lee et al., 2012 ). The underlying RL theory has been widely adapted as a framework for both interpreting experimental data and designing new experiments ( Schultz, 2007 ; Lee et al., 2012 ). Specifically, it is successful in explaining how the brain adjusts the estimates of future rewards and updates these expectations based on experience ( Schultz et al., 1997 ; Schultz, 1998 ; Dayan and Abbott, 2001 ; Lee et al., 2012 ). Motivation has been approached in RL from multiple angles. In the research on intrinsic motivation the agents were additionally rewarded for exercising “curiosity” to try new strategies useful for prospective goals ( Chentanez et al., 2005 ; Singh et al., 2010 ; Kulkarni et al., 2016 ). In multi-objective RL (MORL), motivations affected the available actions favoring particular behaviors to prioritize certain objectives [reviewed in Liu et al. (2014) ]. In both intrinsic motivation and MORL, the concepts of motivation were introduced to achieve certain computational flexibility with no focus on building a plausible model of the human/animal decision-making. RL models are mostly concerned with the learning aspect of behavior. However, fluctuations in physiological states can profoundly affect behavior. Recent suggestions include using time-varying multiobjective reward functions in biological context ( Koulakov, 2018 ; Palm and Schwenker, 2019 ). Modeling such factors is thus an important goal in computational neuroscience and is in the early stages of mathematical description ( Berridge, 2012 ; Berridge and Robinson, 2016 ). In this study, we develop a computational network model of motivational salience in the context of RL. Since RL relies on future rewards to generate behavior, and these rewards are modulated by motivational states, complete understanding of complex behavioral choices is impossible without incorporating motivation. We compare the results of our model to the previously published mouse data obtained in the classical conditioning paradigm ( Stephenson-Jones et al., 2020 ), in which recordings from the VP neurons are available. We show that our motivated RL model both learns to correctly predict motivation-dependent rewards/punishment and generates neural responses consistent with the responses of the VP neurons. In particular, we show that, similarly to real neurons, RL neural networks contain two oppositely-tuned populations of neurons responsive to rewards and punishment. In the model, these two populations form a recurrent network that helps maintain motivation-dependent variables when inputs are missing. Our RL-based model is both consistent with previously published experimental data and suggests a hypothesis for the structure of connectivity in the VP networks. We show that networks with motivation can adapt their behavior to changes in reward functions without relearning network weights and can do so without prespecified goals. We demonstrate how our network model can form the basis of the hierarchical RL system. Overall, we argue that neural networks implementing motivational salience in the brain may enable compact representation of dynamic behaviors accommodating to the shifts in the needs of agents.",
"discussion": "Discussion Motivational salience, which we call here, for brevity, motivation, has been defined previously as the need-based modulation of reward magnitude ( Zhang et al., 2009 ; Berridge, 2012 ). Here we proposed an RL approach to the neural networks that can be trained to include motivation into the calculation of action. We considered a diverse set of example networks that can solve different problems using a similar architecture. In each task, we aimed to use a simple model capable of successfully learning the task. This approach both minimized the training time for each model, and constrained the models to generalize their behaviors across the inputs instead of memorizing the input-output pairs. The networks received both current motivation and state variables as inputs and were trained to compute the magnitude of cumulative motivation-dependent future rewards (Q-function). The action was then selected as a maximum over the Q-function. The network weights were updated using TD rule via the conventional backpropagation algorithm. We found that the networks can learn correct behaviors in this setting, including behaviors that reflect relatively complex scenarios of future motivation changes. Thus, our model, in the transport network example, is capable of solving an NP-complete task without relearning the connection weights. Our approach is based on the previous model by Zhang et al. (2009) , with a few critical differences. First, in the aforementioned work the state of the agent (reflecting the cue/conditioned stimulus, CS) was the only input of a value function; we considered the value function to be explicitly a function of the state and motivation. This way, our models were able to learn the relation between the state, motivation, and their joint incentive value. Second, to interpolate between the multidimensional inputs of the value function, we used deep neural networks. Deep RL models are capable of learning the generalized rules in their weights from the first principles. In our case, the models were capable of generalizing the relation between the rewards, motivations, and their incentive values (i.e., the product of the reward and motivation) and were also able to extrapolate their policies to the novel motivation schedules via developing new courses of action (e.g., the “delayed migration” policy). Third, in our models, the motivations were multidimensional and dynamic, forcing the agents to learn the dynamics of motivations to develop the optimal behaviors. Overall, our work combines the Berridge’s model of motivation with deep RL and previous models of motivational drives to provide an interpretable framework for studying motivated behaviors and their algorithmic rationale in real-world agents and settings. Although motivation seemingly can be viewed as a part of the agent’s state, there are multiple reasons to consider them separately. First, motivation is generally a slowly changing variable. Thus, an animal’s appetite does not change substantially during a few seconds of a single behavioral trial. At the same time, the animal’s actions may lead to immediate changes of its position. Second, the research in neuroscience suggests that motivation and state may be represented and computed separately in mammalian brain. Whereas motivation is usually attributed to the regions of the reward system, such as the ventral pallidum (VP) ( Berridge and Schulkin, 1989 ; Berridge, 2012 ), the state is likely to be computed elsewhere, e.g., in the hippocampus ( Eichenbaum et al., 1999 ) or cortex. Such distinction in the brain may be based on an algorithmic rationale that facilitates computations and is yet to be understood. Finally, in hierarchical RL (HRL) implementation, motivation is provided by a higher-level network, while information about the state is generated externally. For these reasons, in this work we consider an agent’s state s → t and its motivation μ → separately. Although the Q-function with motivation (2) is similar to the Q-function in goal-conditioned RL ( Schaul et al., 2015 ; Andrychowicz et al., 2017 ), the underlying learning dynamics of these two models are different. Motivated behavior simultaneously pursues multiple distributed sources of dynamic rewards. The Q-function therefore accounts for the internal motivation dynamics. This way, an agent with motivation chooses what reward to pursue – making it different from RL with subgoals ( Sutton et al., 1999 ). As we show in this work, simple motivational schedules give rise to large varieties of behaviors. A reduction in numbers of handcrafted features suggests that motivation could provide a step toward more general methods of computation – a goal identified recently by Richard Sutton ( Sutton, 2019 ). Our model of motivation is consistent with the large body of existing motivational models and behavioral observations. In a recent work, Keramati and Gutkin (2014) show that homeostatic RL explains prominent motivation-related behavioral phenomena including anticipatory responding ( Mansfield and Cunningham, 1980 ), dose-dependent reinforcement ( Hodos, 1961 ), potentiating effect of deprivation ( Hodos, 1961 ), inhibitory effect of irrelevant drives ( Dickinson and Balleine, 2002 ), etc. Although homeostatic RL defines the rewards as the gradients of the cost function with a fixed point, the theoretical predictions generalize to the models with linear, or approximately linear, multiplicative motivation. We therefore expect the behaviors of our models to be consistent with the large body of experimental data described above ( Hodos, 1961 ; Mansfield and Cunningham, 1980 ; Dickinson and Balleine, 2002 ). Biologically-grounded choices of motivation dynamics enable our model to reproduce realistic behaviors, including those related to drug addiction. Here we show that a simplistic model, where motivation toward “smoking” grows large compared to motivations toward the other rewards, qualitatively accounts for the binging behavior. Our model suggests that the smoking frequency can be explained with the temporal discounting parameter γ defining the relative impact of the rewards near and far in the future. Our framework, offering a way to derive behaviors from the first principles, can be combined with the classical results regressing the craving rates to a variety of environmental cues (e.g., McKennell, 1970 ) to build finetuned models of addicted behaviors. For example, motivational dynamics may change over time. Addictive drugs can become less rewarding (‘liked’) after repeated experience despite increases in the motivational salience and/or craving for drugs – which can be accounted for in the model with an additional layer in motivation hierarchy (akin to the Transport network task). The discrepancy between low ‘liking’ of the drug and the high rate of cravings can be formalized as the difference between motivational salience and motivational vigor, the generalized willingness to expend energy toward a reward. Motivational vigor can be incorporated into the model by the means of the actor-critic formalism ( Sutton and Barto, 1998 ) which computes log likelihoods for every action. Including these and other parameters to our motivation framework may help building detailed models of addictive behaviors in future work. We trained recurrent neural networks (RNNs) to estimate future motivation-dependent subjective values of the reward in the Pavlovian conditioning task. In contrast to purely feedforward networks, the RNNs allow learning the temporal sequences of events such as the associations between reward-predicting cues (conditioned stimuli, CS) and following rewards (unconditioned stimuli, US). The ability to learn the temporal US-CS associations makes RNNs a rational choice for the models of animal behavior and neuronal activity in Pavlovian conditioning tasks ( Sutton and Barto, 1987 ). Since the structure of network for computing motivation-dependent reward expectations is not fully understood, modeling this circuit as an RNN seems to be a simple and plausible first step, similarly to the models of persistent activity and working memory. It is not clear at the moment whether RNN obtained here is fully contained in VP or is represented by some other part of the reward circuit. Our mathematical model does not specify where the recurrent connectivity facilitating the persistent activity is formed; such structure could occur in PFC where neurons are known to maintain cue-reward associations ( Gottfried et al., 2003 ; Bray and O’Doherty, 2007 ), in VP, or in some other brain region. Previously published findings suggest that VP may not contain connectivity that is strong enough to maintain persistent activity ( Stephenson-Jones et al., 2020 ). Our study may motivate the search for the recurrent circuit that can maintain cue-reward associations. We found that neurons in the RNNs trained to recognize motivation can be clustered into two oppositely tuned populations: positive and negative motivation neurons. These populations display increased firing in reward and punishment trials, respectively. Similar two groups of neurons are found in the previously published data ( Stephenson-Jones et al., 2020 ) on neural responses in the mouse’s VP: a basal ganglia region implicated in motivation-dependent estimates of reward ( Richard et al., 2016 ). Thus, our neural networks develop response patterns comparable to experimentally observed in the brain. We found that the recurrent network structure in this Pavlovian conditioning case is compatible with the conventional models of working memory. The general idea is that the information about an upcoming reward – once supplied by a cue – is maintained in the network due to the positive recurrent feedback. This feedback can be produced by disinhibition between two oppositely tuned populations of neurons, namely positive and negative motivation sensitive cells. Thus, the presence of two subpopulations of neurons may be a consequence of the functional requirements on the network to maintain persistent variables within a trial. This function is reflected in both neural responses and architecture. Motivation offers a framework compatible with other methods in machine learning, such as R-learning, goal-conditioned RL, and hierarchical RL (HRL). R-learning is an average-reward reinforcement learning model ( Schwartz, 1993 ; Sinakevitch et al., 2018 ). Specifically, the cumulative sum of future rewards is computed with respect to the average level of reward. The average reward level is computed across multiple trials, which makes it similar to motivation. In goal-conditioned RL – the closest counterpart to RL with motivation – the Q-function depends on three parameters: Q ( s → t , a t , g ) , where g is the current static goal. In the motivation framework, multiple dynamic goals are present at the same time and it is up to an agent to decide which one to pursue – based on the future motivation dynamic learned by the network. HRL methods include the options framework ( Sutton and Barto, 1987 ; Sutton and Barto, 1998 ), RL with subgoals ( Sutton et al., 1999 ), feudal RL ( Dayan and Hinton, 2000 ; Bacon and Precup, 2018 ), and others. In HRL, complex tasks are solved by breaking them into smaller, more manageable pieces. In both the case of motivated agents and HRL, the reward function is manipulated by an external process, such as a higher level manager ( Sutton et al., 1999 ). HRL approaches have several advantages compared to traditional RL, such as transfer of knowledge from already learned tasks and the ability to faster learn solutions to complex tasks. Although HRL methods are computationally efficient and generate behaviors separated into multiple levels of organization – which resemble animals’ behavior – a mapping of HRL methods to brain networks is missing. Here, we suggest that motivation offers a way for HRL algorithms to be implemented in the brain. In case of motivation, the goal of the agent is not explicitly specified and may shift in course of behavior if motivational variables change their values. Moreover, multiple goals may simultaneously be presented to an agent, whose aim is to select the one that yields the highest subjective reward. We present an example of how HRL can be implemented in motivation setting for the case of transport network. Overall, we suggest that motivation-based networks can generate complex ongoing behaviors that can rapidly adapt to dynamic changes in an organism’s demands without changes in synaptic strengths. Thus, neural networks with motivation can both encompass more complex behaviors than networks with a fixed reward function and be mapped onto neuronal circuits that control rewarded behaviors. Since animal performance in realistic conditions depends on the states of satiety, wakefulness, etc., our approach should help build more realistic computational models that include these variables."
} | 5,683 |
39414881 | PMC11484895 | pmc | 5,636 | {
"abstract": "Understanding how coral reefs respond to disturbances is fundamental to assessing their resistance and resilience, particularly in the context of climate change. Due to the escalating frequency and intensity of coral bleaching events, it is essential to evaluate spatio-temporal responses of coral reef communities to disentangle the mechanisms underlying ecological changes. Here, we used benthic data collected from 59 reefs in the Red Sea over five years (2014–2019), a period that encompasses the 2015/2016 mass bleaching event. Reefs were located within three different geographic regions with different environmental settings: north (Duba; Al Wajh), central (Jeddah; Thuwal), and south (Al Lith; Farasan Banks; Farasan Islands). Coral community responses were region-specific, with communities in the south being more promptly affected than those in the northern and central regions, with hard and soft coral cover dropping drastically in several reefs from around > 40% to < 5% two years after bleaching. Coral bleaching effects were particularly evident in the decrease of cover in branching corals. Overall, we documented a shift towards a dominance of macroalgae, turf algae, and crustose coralline algae (CCA). Using remote sensing data, we analyzed sea surface temperature (SST) regimes at the study sites to infer potential drivers of changes in benthic composition. Both SST and Degree Heating Weeks (DHW) only partially aligned with the responses of benthic communities, highlighting the need for more accurate predictors of coral bleaching in the Red Sea. In times of intense coastal development along Saudi Arabia’s Red Sea coast, our study provides crucial baseline information on developments in coral reef community composition, as well as to guide decision-making, namely restoration efforts. Supplementary Information The online version contains supplementary material available at 10.1038/s41598-024-74956-7.",
"conclusion": "Conclusions This study investigated the response of benthic communities along a latitudinal gradient on the Saudi coast of the Red Sea. We observed a significant regional variation in the effects of the mass bleaching event in 2015–2016. The southern region was the most affected, showing a major decline in hard coral cover, a major loss of branching corals, and a shift in the benthic composition towards an algal (macro- and turf) community, accompanied by a loss in coral diversity. In contrast, the northern and central reefs registered a small decline in coral cover (1–5%), and in some cases coral cover remained stable or even slightly increased.",
"introduction": "Introduction Coral reefs are currently facing multiple stressors arising from a combination of local and global threats that have contributed to coral cover loss 1 – 3 . Increasing sea surface temperature (SST) has resulted in more frequent and intense climate-induced bleaching events, impacting coral reefs globally 4 – 7 . Thermal anomalies and coral bleaching have increased in frequency, from once every 25 to 30 years at the beginning of the 80’s, to once every 5.9 years by 2016 6 , affecting even regions considered coral refuges 8 – 11 . Regardless of the projected future climate scenarios, coral reefs are expected to be degraded, resulting in changes in species composition 12 , including a general loss of biodiversity and ecosystem services 13 , 14 . Indeed, by 2050, it is predicted that coral cover will decrease globally by 5–14%, while by the end of the century more than 40% of coral may be lost from the world’s oceans 15 . Coral bleaching occurs when, under sustained stress, corals expel their algal symbionts from their transparent polyp tissue, thereby exposing the white, calcareous coral skeleton 16 . In a zooxanthellate coral, these symbionts may provide over 90% of the energetic requirements for the host 17 . If the duration is long enough, bleaching can cause coral death 18 and consequently reef degradation 16 , 19 , thus directly affecting reef functioning 20 , 21 . Reef degradation often manifests as a shift towards dominance by macroalgae or other non-calcifying organisms (e.g. soft corals, sponges) which are able to outcompete calcifiers (e.g. hard corals) after a disturbance 3 , 22 , 23 . This shift in community structure and the potential loss in coral reef-associated biodiversity has a direct effect on reef functioning, affecting diverse ecological processes such as biogeochemical fluxes, primary production, herbivory, and predator-prey interactions 24 – 26 . Ultimately, these ecological changes will be translated into negative economic effects 27 . The assessment of long-term changes to coral reef benthic community composition as a response to stressors associated with climate change requires the investigation across large temporal scales, supported wherever possible by well-established and standardized monitoring programs 28 . Nevertheless, in many tropical countries long-term monitoring programs are scanty and rarely standardized; not only are temporal data scarce or missing, but often baseline data themselves is absent, making evaluation of the ecological effects of coral bleaching challenging 29 . Such lack of data hampers the ability to make projections based on future scenarios. This is particularly true coral reef communities in the Red Sea, which remain largely understudied compared to other major reef systems (e.g. the Great Barrier Reef) 30 . Red Sea coral reef community responses (and recovery trajectories) to bleaching are not always documented as national monitoring programs have been limited in scope and duration. Monitoring by local stakeholders often faces the same obstacles (but see PERSGA 31 ). Over the last decade, the Red Sea’s coral reefs have seen several heat-related coral bleaching events 32 – 34 , and more recently, seasonal bleaching due to cold stress 35 . Though heat-related bleaching events in the Red Sea are beginning to garner attention, Genevier et al. 36 reported that thermal anomalies in the Red Sea have increased in frequency and intensity since 1998, mirroring the trend seen worldwide 37 , 38 . In the aforementioned study, the authors determined historical marine heatwave events in the region by analyzing 35 years of SST data, identifying areas at risk, and highlighting potentially undetected bleaching events. According to the study, marine heatwaves encompassed more extensive areas and occurred more frequently than previously reported, even in the non-El Niño years, suggesting that coral bleaching may have been underestimated in the past. The Red Sea is characterized by well-known north-south gradients of salinity, SST, and primary productivity 39 , making the basin an ideal natural laboratory to study coral reef dynamics. Communities in the three regions of the Red Sea (northern, central, and southern) are exposed to different environmental conditions and, therefore, may respond to SST anomalies in contrasting ways due to their different environmental history. Based on a network of reefs that have been monitored since 2014 along the Saudi Arabian Red Sea coast (revisited nearly every two years) using standardized photo-transects, we investigated spatial and temporal changes in the community composition of coral reef benthic organisms over a six-year period. This period (2014–2019) was comprised of surveys conducted before and after the third and longest global-scale coral bleaching event caused by the El Niño-Southern Oscillation (ENSO) in 2015/2016 19 , 40 . This study provides insights into the dynamics of benthic community structure across regional and temporal scales in the hypersaline and naturally warm Red Sea, and lays the foundation to track future trajectories of recovery or changes in these coral reefs. The relationship between patterns in Degree Heating Weeks (DHW) and SST data was also investigated (following findings by DeCarlo et al. 41 ), as were the reported regional differences in bleaching prevalence. We hypothesize that coral communities of the northern Red Sea will be the most resistant to marine heat waves due to their higher thermal tolerance; indeed, this region is often considered a coral refuge that could potentially help to repopulate other regions 42 , 43 .",
"discussion": "Discussion This study documented latitudinal and temporal changes in the composition and structure of coral reef benthic communities over a 5-year period encompassing the 2015–2016 mass bleaching event along the Saudi Arabian coast of the Red Sea. The results show that Red Sea coral reef benthic communities responded differently to a mass bleaching event dependent on the environmental context they are set in. This study also highlights the value of in situ biological data regards to assessing the effects and recovery trajectory of bleaching events. Indeed, we show here a strong mismatch between the predictions of bleaching using DHW metrics and observations, reinforcing the synergistic effects between SST and other variables, like nutrients, triggering bleaching as previously reported by DeCarlo et al. 41 in the Red Sea. Interestingly, responses of coral genera appear to be region-specific; although some genera seem to generally be more sensitive (e.g., Pavona , Montipora , and Acropora branching), or more resistant (e.g., Porites massive), this does not hold true across all sites or regions. The nearshore reefs of the most affected region (SRS) did not show clear signs of recovery four years after the mass bleaching event. The variability in coral bleaching across different sites, where some corals are affected while others remain unscathed, can be attributed to multiple factors. This diversity in bleaching patterns suggests that the phenomenon is influenced by a complex interplay of elements, probably including water temperature, water quality, light intensity, local adaptation of coral organisms and acclimatization, individual thermal history, and the type of symbiotic zooxanthellae present 63 , 64 . The increased CCA cover reported in this study may signify an encouraging possibility for coral reef regeneration, as CCA plays a pivotal role in fortifying the reef substrate through the secretion of calcium carbonate, and is linked to the settlement and attachment of coral larvae 65 , 66 . This calcified deposit provides an anchor for coral development and acts as a protective shield, mitigating the effects of physical erosion 67 . Even so, any potential for positive outcomes are counterbalanced by the influence of human-induced pressures. Factors such as the overfishing of reef herbivores and nutrient enhancement can impede these beneficial effects, thus hindering coral reef recovery 68 . Region-specific responses of Red Sea benthic communities to the 2015/2016 mass bleaching event Investigating the patterns of variability in coral reef benthic communities through space and time, we found a region-specific response that aligned with the prevailing SST trends and the known productivity patterns in the region 45 . Specifically, benthic communities from the southern Red Sea were characterized by high coral cover (25.9%±19.4 in period 1), a low cover of fleshy macroalgae (3.1%± 6.0), and a substantial cover of soft corals (8.0%±8.4) before the bleaching event, but were severely affected by the 2015/2016 mass bleaching event. This result based on in situ photo transects supports the findings by DeCarlo et al. 41 The authors detected a higher intensity of the 2015/2016 bleaching event in the southern Red Sea compared to previous bleaching events affecting the Red Sea region (i.e., 1982, 1995, and 2002), hence a region-specific signature of these events. Although the SRS region is known for its comparatively high primary productivity due to Gulf of Aden Intermediate Water (GAIW) intrusion 69 , 70 and monsoon-driven upwelling 71 , 72 , corals are still very sensitive to the synergistic effects of both temperature and nutrient stressors when above certain thresholds. Indeed, it seems that the combination of thermal anomalies and an increase in nutrients from an early upwelling caused by the effects of El Niño may determine the intensity of bleaching events in the region 41 , 72 . In other words, corals in the Red Sea can resist an increase in SST if not accompanied by an increase in nutrients, as observed in 1998 or 2010 in the southern Red Sea 41 . Another study reported that environmental pressures like heat stress and higher levels of dissolved organic carbon (DOC) could potentially accelerate the breakdown of the coral-algae symbiotic by boosting the rates of nitrogen fixation linked to corals 73 . In this study, a relatively high turf cover was also recorded in period 1 (16.6%), and it increased after the bleaching event (24.1%); this trend was strongest in the SRS which increased from 16.5% in period 1 to 28.4% in period 2). Although there are several factors related to the decline and loss of coral species, including coral disease, invasive species, and climate change, PERSGA 31 reports that major threats in this region are primarily linked to human pressures such as overfishing, coastal development, population expansion, and industrial pollution. Thus, the observed increase in turf algae cover may be a part of the broader pattern of ecological shifts driven by these anthropogenic factors and not an outcome of the bleaching event. It is important to note that these factors can contribute to decrease the tolerance of coral species to additional stressors like those related to increased SST usually associated with bleaching events. The most remarkable, yet expected, response to the 2015/2016 bleaching event was a significant decrease in coral cover. Particularly for hard corals (both branching and massive forms demonstrated clear reductions, showing a 20-fold drop in cover (from 25.8 to 1.2%). A similar, though less drastic, decrease in soft coral cover was also detected (from 8 to 0.3%). Such sharp decreases in coral cover have been reported in response to bleaching events worldwide 6 , 74 , 75 , followed by either an increase in turf or macroalgae (an increase from 15.3 to 25.9%, and 3.1–34.8%, respectively for the present study) 76 , 77 . Even though Red Sea coral reef communities have widely been considered resilient, namely due to their high recruitment rates 78 , 79 , more frequent bleaching events in the region may hinder their potential for recovery, as observed elsewhere 16 , 80 . This can be particularly true for nearshore reefs that are more affected by the bleaching events due to, among other stressors, more drastic variation in SST than offshore reefs 81 . Indeed, considering that four years after bleaching the percent cover of hard corals was about 12% of the pre-bleaching levels ( pre-bleaching: 29.8%±18.8%; four years post-bleaching: 3.6%±3.2%), signs of solid recovery were not detected. Corals in the southern Red Sea may face a serious threat as mass bleaching events are expected to occur on average every 5.9 years globally 6 . Nevertheless, though slow, signs of recovery were noticed. Average hard coral cover almost doubled in two years (period 2: 1.2 ± 0.8%; period 3: 3.6%±3.2%). It is important to note that such signs of recovery ere not homogenous, as clarified by the high standard deviation. Taxa susceptibility to bleaching: toward a prevalence of massive forms with a loss of rapid growing branching corals During a heat stress event, coral species will respond differently to changes in environmental conditions. In some cases, high plasticity may be observed, and response patterns may not be consistent across reef habitats or across a shelf gradient 82 . In general, branching corals are considered more prone to bleaching and subsequent mortality 83 – 86 . A similar pattern has been detected in previous studies in the Red Sea 32 , 34 and is supported by our findings. A previous study on the Saudi Arabian Red Sea coast showed that the coral genera more sensitive to changes in environmental conditions were indeed mainly genera with branching growth forms, such as Acropora , Stylophora , Seriatopora , and some species of Porites 87 . In our study, Porites , Pocillopora , Pavona , and tabular and branching Acropora were the most abundant corals in the south in the first sampling period contributing to ~ 12% total cover. After the bleaching event, these genera specifically (and all branching corals in general) were severely affected and almost disappeared, confirming the growth form’s sensitivity to bleaching. In our study, the massive Porites category was the main driver of the community composition. Consistent levels (and a general dominance) of coverage highlights its relative tolerance to global warming effects. This is supported by a previous study in the Great Barrier Reef, which found that acroporids bleach faster and present a higher whole-colony mortality rate, while species such as Platygyra daedalea and Porites lobata , both massive species, were characterized by a later onset of bleaching and partial mortality 88 . Changes in benthic community composition does not necessarily correspond with heat stress metrics It has previously been shown that the occurrence of widespread coral bleaching and heat stress in the northern Red Sea have so far been disconnected from each other 43 . The present study also showed a discrepancy between heat stress and observed changes in benthic reef communities. Although showing constantly high heat stress each summer during the study period, reefs in the northern Red Sea experienced changes in benthic cover on a much lower scale compared to reefs in the central and southern regions (with the exceptions of some reefs at Al Wajh). These observations are in line with the proposal of the northern Red Sea as a thermal refuge for scleractinian corals 42 , 43 . However, this region is not without its own unique set of environmental stressors. Intense coastal development in the region associated with two major development projects (NEOM, and the Red Sea Global 89 , 90 ) which can affect turbidity, nutrient availability, and local hydrodynamics, all of which have been shown to influence the response of corals to bleaching 91 – 93 . Temperature variability patterns experienced by corals can, in some cases, be more relevant to the susceptibility of corals than the maximum temperature populations experience over a certain period of time 92 – 95 . Significant coral cover decline in some of the CRS and SRS locations observed here may indicate that corals in these areas are already close to their thermal tolerance thresholds, while this may not be the case yet for corals in the NRS 43 . However, it remains unclear for how long reefs in the NRS will be able to withstand widespread bleaching and coral cover loss in the future given the already high rates of ocean warming in the north 96 . The DHW peaks particularly in the southern locations combined with the early summer upwelling / GAIW-intrusion most likely played a major part in the bleaching pattern observed in the southern Red Sea, as reported by DeCarlo et al. 41 , 72 . Upwelling in the southern Red Sea in June to August 2015 was the strongest recorded since the 1980s but ceased earlier than usual, leading to relatively low SST during the upwelling but a longer than usual subsequent warming period 41 , 72 . This strong upwelling of GAIW also led to comparatively high nutrient levels in otherwise oligotrophic waters where (sub-) tropical coral reefs thrive, another major stressor for corals 41 . The combination of a long and intense warming period and a strong upwelling event leading to high ambient nutrient levels are most probably what caused bleaching in southern-central and southern Saudi Arabian reefs in 2015 and potentially the subsequent coral cover decline observed at these sites."
} | 4,963 |
33885483 | null | s2 | 5,637 | {
"abstract": "Through targeted binding to the cell membrane, structural DNA nanotechnology has the potential to guide and affix biomolecules such as drugs, growth factors and nanobiosensors to the surfaces of cells. In this study, we investigated the targeted binding efficiency of three distinct DNA origami shapes to cultured endothelial cells via cholesterol anchors. Our results showed that the labeling efficiency is highly dependent on the shape of the origami as well as the number and the location of the binding overhangs. With a uniform surface spacing of binding overhangs, 3D isotropic nanospheres and 1D anisotropic nanorods labeled cells effectively, and the isotropic nanosphere labeling fit well with an independent binding model. Face-decoration and edge-decoration of the anisotropic nanotile were performed to investigate the effects of binding overhang location on cell labeling, and only the edge-decorated nanotiles were successful at labeling cells. Edge proximity studies demonstrated that the labeling efficiency can be modulated in both nanotiles and nanorods by moving the binding overhangs towards the edges and vertices, respectively. Furthermore, we demonstrated that while double-stranded DNA (dsDNA) bridge tethers can rescue the labeling efficiency of the face-decorated rectangular plate, this effect is also dependent on the proximity of bridge tethers to the edges or vertices of the nanostructures. A final comparison of all three nanoshapes revealed that the end-labeled nanorod and the nanosphere achieved the highest absolute labeling intensities, but the highest signal-to-noise ratio, calculated as the ratio of overall labeling to initiator-free background labeling, was achieved by the end-labeled nanorod, with the edge-labeled nanotile coming in second place slightly ahead of the nanosphere. The findings from this study can help us further understand the factors that affect membrane attachment using cholesterol anchors, thus providing guidelines for the rational design of future functional DNA nanostructures."
} | 511 |
35630345 | PMC9147336 | pmc | 5,638 | {
"abstract": "Rhizospheric plant–microbe interactions have dynamic importance in sustainable agriculture systems that have a reduced reliance on agrochemicals. Rhizosphere signaling focuses on the interactions between plants and the surrounding symbiotic microorganisms that facilitate the development of rhizobiome diversity, which is beneficial for plant productivity. Plant–microbe communication comprises intricate systems that modulate local and systemic defense mechanisms to mitigate environmental stresses. This review deciphers insights into how the exudation of plant secondary metabolites can shape the functions and diversity of the root microbiome. It also elaborates on how rhizosphere interactions influence plant growth, regulate plant immunity against phytopathogens, and prime the plant for protection against biotic and abiotic stresses, along with some recent well-reported examples. A holistic understanding of these interactions can help in the development of tailored microbial inoculants for enhanced plant growth and targeted disease suppression.",
"conclusion": "4. Conclusions and Future Perspectives The interaction intricate chemical signaling between plants and their associated microbiome deciphers the communication dynamics that are underway within the rhizosphere and their outcomes in terms of plant productivity and development. Over the last decade, quorum sensing has been studied in the gene expression of both beneficial and harmful interactions between inter- and intramicrobial species and in the interkingdom signaling between plants and PGPR, i.e., nitrogen-fixing bacteria and rhizobia. Most of the plant metabolites that have been reported to date have been characterized by the Arabidopsis thaliana plant model. The recent literature suggests that other plants need to be investigated for the analysis of more plant metabolites and chemical cues. Expanding the research on rhizosphere communities will aid in the discovery of new chemical cues and their potential to enhance plant productivity in terms of symbiosis, resistance against environmental stresses, and immunity or defense mechanisms toward phytopathogens. Based on experimental evidence, there is no doubt that plants create an environment that recruits a specific rhizomicrobiome and shape microbial associations that are beneficial for their growth. Therefore, the escalating pressure of the demand for high crop production has opened up new avenues for research on rhizosphere signaling to reduce dependency on synthetic processes and agrochemicals. The alterations in the metabolic pathways of plants, microbes, rhizodeposits, and signaling molecules could be effective in the development of a rhizomicrobiome that is beneficial for effective plant productivity and resistance toward biotic and abiotic stresses, which could ultimately lead to more sustainable agriculture. Moreover, future research can be expanded by using novel “multi-omics techniques”, which encompass genomics, proteomics, transcriptomics, phenomics, metagenomics, metabolomics, metatranscriptomics, metaproteomics, and metagenomics and could reveal multi-layered information about new chemical cues, cellular mechanisms, signaling mechanisms, and plant genes. Furthermore, future research on rhizosphere engineering is essential to dissect the targeted outcomes of the beneficial plant–microbe communications that could broaden the application of sustainable agriculture.",
"introduction": "1. Introduction The rhizosphere is a most captivating environment, which harbors a variety of microorganisms that are deeply involved in plant–microbe communication. This high-density niche allows plants to interact with associated microorganisms through chemical signals that are produced in response to specific stimuli, which in turn activate many regulatory mechanisms [ 1 ]. Rhizospheric regions possess higher bacterial activity than non-rhizospheric regions. The composition of a microbiome is determined by different biotic and abiotic factors, i.e., the climate, type of soil, and chemical signals that are produced by the plant and its associated microbes [ 2 , 3 ]. Plants and microbes have diverse interactions that involve close interactions, either positive or negative, including mutualism (symbiosis), parasitism, and commensalism [ 4 ]. Positive interactions include those with microorganisms, e.g., rhizobia, plant growth-promoting rhizobacteria (PGPR), and mycorrhiza, which result in beneficial outcomes, such as growth promotion, nutrient accessibility, and protection against abiotic and biotic environmental stresses [ 5 , 6 , 7 ]. On the other hand, plant interactions with microbial pathogens result in negative outcomes, i.e., plant diseases [ 8 , 9 ]. The interaction between both partners depends on specialized signaling molecules or chemical signals that are also significant in cooperative, as well as competitive, microbial behavior [ 10 , 11 ]. The chemical cues or secondary metabolites act as mediators in plant–microbe and microbe–microbe communication and also trigger plant responses [ 12 ]. In the past decade, progress has been made in the understanding of the types of chemical signals that are responsible for controlling the activities of plants and associated microbes [ 13 , 14 , 15 ]. The most studied signaling compounds are N-Acyl homoserine lactones (N-AHLs), which are produced by a variety of bacterial taxa and regulate quorum sensing and pathogenicity within a bacterial population [ 16 , 17 , 18 ]. Likewise, plant roots secrete a variety of metabolites as exudates, including photosynthetically derived carbon compounds, e.g., organic acids, vitamins, flavonoids, polysaccharides, amino acids, and sugars. These root exudates create an enriched environment for the rhizomicrobiome to interact and increase diversity based on the composition of the exudates [ 15 , 19 , 20 ]. Moreover, plants release secondary metabolites against pathogens and insects that act as defensive signals [ 21 , 22 ]. Plants use adaptive strategies to enhance the defensive capacity of their innate resistance to biotic and abiotic factors by interacting with beneficial microbes [ 23 , 24 ]. To unravel the process of microbial interaction with plants, an understanding of the types of chemical communication between all members is necessary. Thus, the known microbial community and their interactions could help in the optimal use of beneficial microbes for better plant growth. Some of the literature on different signaling molecules that participate in the development of interactions within the rhizosphere for enhanced plant growth has been reviewed [ 15 , 25 , 26 , 27 , 28 ]. However, much still needs to be explored in terms of the significance of these interactions as far as microbial or chemical diversity and the understanding of signaling molecules are concerned. This review aims to enlighten our understanding of rhizosphere signaling in plant–microbe communication, both cooperative and competitive, and its significance in plant productivity and the development of sustainable agriculture systems. Nonetheless, the signaling compounds have a prodigious potential to escalate plant functions when they are understood in depth."
} | 1,799 |
12914651 | PMC193635 | pmc | 5,640 | {
"abstract": "With 16 complete archaeal genomes sequenced to date, comparative genomics has revealed a conserved core of 313 genes that are represented in all sequenced archaeal genomes, plus a variable 'shell' that is prone to lineage-specific gene loss and horizontal gene exchange."
} | 67 |
26222031 | PMC4539577 | pmc | 5,642 | {
"abstract": "Ammonia- and nitrite-oxidizers are collectively responsible for the aerobic oxidation of ammonia via nitrite to nitrate and play essential roles for the global biogeochemical nitrogen cycle. The physiology of these nitrifying microbes has been intensively studied since the first experiments of Sergei Winogradsky more than a century ago. Urea and ammonia are the only recognized energy sources that promote the aerobic growth of ammonia-oxidizing bacteria and archaea. Here we report the aerobic growth of a pure culture of the ammonia-oxidizing thaumarchaeote Nitrososphaera gargensis 1 on cyanate as the sole source of energy and reductant, the first organism known to do so. Cyanate, which is a potentially important source of reduced nitrogen in aquatic and terrestrial ecosystems 2 , is converted to ammonium and CO 2 by this archaeon using a cyanase that is induced upon addition of this compound. Within the cyanase gene family, this cyanase is a member of a distinct clade that also contains cyanases of nitrite-oxidizing bacteria of the genus Nitrospira. We demonstrate by co-culture experiments that these nitrite-oxidizers supply ammonia-oxidizers lacking cyanase with ammonium from cyanate, which is fully nitrified by this consortium through reciprocal feeding. Screening of a comprehensive set of more than 3,000 publically available metagenomes from environmental samples revealed that cyanase-encoding genes clustering with the cyanases of these nitrifiers are widespread in the environment. Our results demonstrate an unexpected metabolic versatility of nitrifying microbes and suggest a previously unrecognized importance of cyanate for N-cycling in the environment."
} | 422 |
39496612 | PMC11535045 | pmc | 5,643 | {
"abstract": "Materials scientists have taken a learn-from-nature approach to study the structure-property relationships of natural materials. Here we introduce a nature-inspired composite architecture showing a hierarchical assembly of granular-like building blocks with specific topological textures. The structural complexity of the resulting architecture is advanced by applying the concept of grain orientation internally to each building block to induce a tailored crack resistance. Hexagonal grain-shaped building blocks are filled with parallel-oriented filament bundles, and these function as stiff-blocks with high anisotropy due to the embedded fiber reinforcements. Process-induced interfacial voids, which provide preferential crack paths, are strategically integrated with cracks to improve fracture toughness at the macroscopic scale. This study discusses the structural effects of the local/global orientations, stacking sequences, feature sizes, and gradient assemblies of granular blocks on crack tolerance behavior. Alternating stacking sequences induce cracks propagating in the arrestor direction, which boost the fracture energy up to 2.4 times higher than the same layup stacking sequence. Gradient arrangements of feature sizes from coarse to fine or fine to coarse result in the coexistence of stiffness and toughness. Our approach to applying crystallographic concepts to complex composite architectures inspires for original models of fracture mechanics.",
"introduction": "Introduction Nature is a plentiful source of inspiration for overcoming the current challenges facing the science of artificial materials. As the mechanical limitations of existing engineering materials have become apparent, material scientists have turned their attention to natural biological systems to find original models for toughening mechanisms 1 – 3 . For example, biomineralized materials, such as the brick-and-mortar structure of abalone shell nacres 4 , the gradient structure of bamboo stem vascular bundles 5 , and the bouligand structure of crustacean exoskeletons 6 , have provided abundant structural motifs to inspire the design of damage-tolerant architectures. These structures usually originate from simple elementary building blocks, the hierarchical assembly of which exhibits a synergistic mechanical performance beyond the capabilities of the individual building blocks 7 , 8 . The recent emergence of a bottom-up 3D printing process, which corresponds to the architectural concept of natural evolution, has accelerated the expansion of research on nature-inspired structures by providing design freedom in geometric constructions 9 – 13 . For example, the 3D printing of specific topological arrangements using two different materials, stiffer platelets and softer matrix, resulted in tough composites inspired by mineralized natural materials 12 . The granular microstructure of crystalline materials in metallurgy has also served as an inspiration motif for 3D printing. By mimicking elements of crystal microstructures such as crystal phases, grain boundaries and precipitates, a range of robust and damage-tolerant architectures has been developed 13 . By expanding our interest to composite materials, we can find further applications 14 , 15 . Among many studies, much attention has been focused on the process-induced shear alignment and hierarchical distribution of fibrous or 2D platelet reinforcements 16 – 20 . Induced magnetic fields and nozzle rotations have also made significant contributions to the tunable alignments of reinforcements in concentric, stacked, and spiral patterns 17 – 20 . In terms of fracture mechanics, recent reports demonstrated that fine planar alignment of discontinuous reinforcement, when induced layer by layer, improves fracture toughness accompanied by increased crack resistance and fracture surface 21 , 22 . Indeed, a model of nature-inspired fracture mechanics designed by combining bouligand and nacrous staggered structures showed a hybrid-toughening mechanism of crack twisting and crack bridging along the planar fiber orientations gradually rotated by a certain pitch angle 21 . Moreover, the anisotropic arrangement and heterogeneous distribution of fibers over the micro and macro scales control the crack propagation mode and assign enhanced fracture toughness far beyond monolithic anisotropic structures. The monolithic anisotropic structures have a simple in-plane arrangement, which allows cracks to propagate straight along the lowest toughness direction. In contrast, the anisotropy and heterogeneity of fiber orientation force crack tips to repeatedly deflect and arrest, which increases local crack resistance and causes additional fracture energy dissipation throughout the system 23 , 24 . On the negative side, the spontaneous formation of void defects that are inherent in 3D printing has been identified as a critical issue 25 , 26 . However, recent studies have shown that controlled void formation in biomimetic composites may not be detrimental in terms of nature-inspired engineering 27 , 28 . The strategic incorporation of voids in printed composites can help realize engineering solutions that imbue effects such as crack blunting and crack deflection 6 , 27 . Accordingly, there has been much discussion about the hierarchical organization of subfilament bundles prepared by FDM printing, and it has been clarified that the geometric arrangement and interfacial bonding strength of subfilament bundles can have a significant impact on fracture toughness 29 . The macroscopic crack planes along weak interfaces, accompanied by multiple crack deflections and branchings, reduced the fracture driving force and induced progressive damage. Since large amounts of filaments and interfaces participated in plastic energy dissipation, the total amount of fracture energy that can be consumed throughout the system increased correspondingly. Rather, contrary to expectations, excessive interfacial bonding strength inhibited crack blunting and led to fatal damage. Interestingly, the strategically integrated void-crack provides detectable warning and a wide margin before final failure, thereby guiding a safe failure mechanism with a high residual load carrying capacity. This synergistic void-crack interaction has been well addressed in simple layup monolithic composites 6 , 27 , 29 , 30 , but a meaningful attempt in complex hierarchical architectures has so far not been undertaken, and furthermore, experimental evidence has not been presented to verify the basic assumptions. Our efforts on these issues would occupy this blank space in two aspects: the presentation of nature-inspired complex hierarchical structures and the scientific understanding of their failure mechanisms. We report a concept of nature-inspired architecture, which is hierarchically assembled from grain-like building blocks with specific topological textures. Granular domains have been defined by sequential local printing, filling each sub-region with aligned filament bundles of anisotropic properties. The individual granular domains, which serve as fundamental building blocks, were strategically positioned in a planar layer with varying sizes, orientations, and organizations, and then hierarchically layered in various stacking sequences. The reinforcing fibers embedded in the elementary filaments contributed to the local anisotropy from granular orientation 15 , 31 . The macroscopic topological texture was specified by the textural contrast of the interfaces between the subfilament bundles in the building block assembly. Our study starts with the three-dimensional visualization of a series of nature-inspired architectures, assembled by local/global orientations, stacking sequences, feature sizes, and gradient assemblies of granular blocks, respectively, and then is extended to the observation of the corresponding fracture behaviors. The main objective is to investigate the resulting characteristic granular cracking behaviors, which are inherently induced by the synergistic crack-void interactions and the obstacle effects of aligned fibers. It is well known that crack growth along grain phases or grain boundaries in crystallography determines the embrittlement and damage tolerance of materials, which should be properly compared to our hierarchical artifacts 32 . The current work focuses specifically on the spatiotemporal capture of three-dimensional fracture geometry along process-induced multiscale arrangements of discontinuous reinforcing fibers and weak void interfaces, aiming to clarify the resulting anisotropic fracture behavior by presenting experimental evidence. In fact, previously presented surface observations have limited usefulness due to the loss of spatial information about the interior of the bulk material, which leads to incomplete understanding or incorrect information about the failure mechanisms of 3D structures 33 , 34 . Here the crack geometry was correctly identified in the micro-macroscopic range, and consequently, the understanding of the microstructural effects of hierarchical orientation and void-crack interactions on macroscale crack propagation was improved with definite scientific validation.",
"discussion": "Discussion The present work has provided a natural inspiration motif for the 3D printing of artificial granular architectures and demonstrated composites’ fracture behaviors correlating with the poly-granular cracking behavior mechanisms found in natural mineralized materials. The process-induced microstructural features, i.e., a high degree of material alignment and weak interfacial void formation, were used to systematically organize hierarchical granular architectures of specific geometric topology in the filamentary phase. Large-field tomographic observation has contributed to the three-dimensional visualization of various structural configurations of granular building blocks (local/global orientation, stacking sequence, feature size, and gradient structure) and provided impressive internal microstructures and crack morphologies for identifying the segmented constituents and understanding the resulting toughness mechanisms. The significance of this work is that the structural complexity of the resulting architecture was advanced by the systematic organization of granular building blocks consisting of composite filament bundles (compared to the general case using pure materials in the existing literature), and consequently, the resulting hierarchical composite architecture contributed to the exceptionally high crack resistance. Our initial attention focused on the characteristic cracking behaviors, i.e., inter- and intra-granular cracking behaviors, inherently induced by synergistic crack-void interactions and the obstacle effects of aligned fibers. Crack growth along grain phases or grain boundaries tailored the crack driving force with progressive crack patterns. Eventually, the alternating stacking sequence led cracks propagating in the divider direction to cracks propagating in the arrester direction, which caused complex multiple cracking modes, including shear crack opening, resulting in a dramatic increase in fracture toughness. To the best of the author’s knowledge, this is the first thorough observation of the three-dimensional cracking behavior of 3D-printed nature-inspired composite architecture. This challenge means that further studies are needed to understand the mechanical behaviors of more diverse and complex architectures as a function of materials, compositions, and configurations, and such work can provide plenty of research space in the field of structural engineering. This approach using the inherent characteristics of 3D printing will help to break through technical limitations from a nature-inspired engineering perspective and will broaden the scope of 3D printing into a more extensive range of research and engineering applications. In future work, we will investigate the reinforcement effects of embedded fibers, including fiber length, fiber content, length-to-diameter ratio, and interfacial bonding strength between fibers and matrix, which are important physical/geometrical parameters but have been overlooked here due to paper limitations. To clarify these effects, we will explore more suitable structures and develop efficient analytical procedures while maintaining the intention of this study to gain scientific inspiration from the robust and dynamic structural systems of natural materials."
} | 3,121 |
21845185 | PMC3145257 | pmc | 5,646 | {
"abstract": "Horizontal gene transfer (HGT) plays an important role in the evolution of life on the Earth. This view is supported by numerous occasions of HGT that are recorded in the genomes of all three domains of living organisms. HGT-mediated rapid evolution is especially noticeable among the Bacteria, which demonstrate formidable adaptability in the face of recent environmental changes imposed by human activities, such as the use of antibiotics, industrial contamination, and intensive agriculture. At the heart of the HGT-driven bacterial evolution and adaptation are highly sophisticated natural genetic engineering tools in the form of a variety of mobile genetic elements (MGEs). The main aim of this review is to give a brief account of the occurrence and diversity of MGEs in natural ecosystems and of the environmental factors that may affect MGE-mediated HGT.",
"conclusion": "Concluding Remarks The high rate of horizontal gene exchange in natural ecosystems is evident from both retrospective and prospective types of studies. The microbial world around us can be seen as a giant microbiome, with the continuous flow of genes between its different compartments. This flow is sustained by a variety of sophisticated natural genetic engineering tools, MGEs, which have been selected during the evolution as providing the means for re-shuffling the available genetic material and picking the best responses possible to cope with the continuously changing environmental challenges. The recent relatively short history of the “antibiotic era” (Aminov, 2010 ) demonstrates the ultimate success of this strategy and urges us to rethink our own when interacting with the microbial world. Continuous discoveries of novel MGEs and mechanisms of HGT, together with the findings of unexpectedly high HGT rates in natural ecosystems, indicate that we are still far from the understanding of the true extent of HGT in nature. The contribution to the better understanding may be envisaged as the combination of retrospective and prospective approaches. On the dry lab side, the history of past HGT events, which is recorded in the wealth of genomic/genomic information, can be more vigorously interrogated on the basis of our knowledge about MGEs and with the help of bioinformatics tools that are able to detect the events consistent with HGT. On the wet lab side, it is the development of in situ technologies that are more sensitive, less intrusive and applicable to the field studies. The microcosm experiments should model real environmental situations, working with native microbiota, with a lesser dependence on model organisms. These developments may help to elaborate better strategies to deal with the pressing needs such as the emergence of novel infections and opportunistic pathogens as well as antibiotic resistance genes."
} | 702 |
37626072 | PMC10457335 | pmc | 5,647 | {
"abstract": "Upscaling the utilization of polymer wastes together with the valorization of undesirable waste rice straw (RS) will minimize the environmental impact of waste disposal by traditional tools. This present work assesses the utilization of polyethylene terephthalate wastes in enhancing the production of polyester-(high density polyethylene) HDPE from Rice straw polyol composites. In this respect, the polyester from rice straw polyol in hybrid with glycolysis polyethylene terephthalate wastes (Gly-WPET) was assessed in comparison with that resulted from RS-polyol, using FTIR, non-isothermal analysis, and mechanical tests. The data showed the positive role of Gly-WPET in hybrid with RS-polyol in production polyester with high thermal stability and mechanical properties. It provided an increase in activation energy of degradation, elongation, Young's Modulus, and modulus of toughness from 184.5 to 1201 kJ/mole, from 4.7 to 9.8%, from 47.5 to 66.5 MPa, and from ~ 4.0 to 23 J/m 3 , respectively. This behavior was reflected in the properties of HDPE-polyester polyol (PEPO) composites, especially in improving elongation (from 55.4 to 72%). These promising data persuade us to recommend the influential role of Gly-WPET in using PEPO from liquefied RS as a plasticizer.",
"conclusion": "Conclusions This work deals with promoting the utilization of bottle plastic wastes and undesirable rice straw in production of green products and avoiding the environmental risk from their disposing by burning. A potential route for using WPET to enhance the performance of polyester resulting from RS-polyol was evaluated. The production variables (amount of curing agent and replacing a percentage of RS-polyol by Gly-WPET) were optimized by estimating the thermal stability and mechanical properties of ester products. It was found that polyester from hybrid liquefied RS and Gly-WPET resulted in increase the thermal stability (E a increased from 184.5 to 1201 kJ/mol), elongation (from 4.7 to 9.8%), young’s modulus (from 47.5 to 66.5 MPa), and toughness (~ 4.0 to 23 J/m 3 ) than PEPO produced from RS-polyol. Additionally, it promotes the application of the PEPO as a plasticizer and in producing high-performance HDPE-polyester composites.",
"introduction": "Introduction Minimizing the solid wastes (e.g., plastics and agricultural) is essential for preserving the environment from traditional tools used to dispose of these wastes, which affects human health. Hence, the disposal of agricultural solid waste is one of the most pressing environmental issues 1 . Polyethylene terephthalate waste (WPET) is an example of the most important commercially consumed plastics in our daily life. On the other hand, rice straw (RS) is regarded as undesirable agricultural waste. These wastes are in massive amounts, and most are not fully utilized. Therefore, this waste management with the upcycling concept has become an important social issue. There are mainly two ways for PET recycling, which can be done mechanically or chemically. Mechanical recycling is easy to employ but presents some limitations as the properties of the final product decrease from the second cycle, while chemical recycling offers versatile procedures. The potential for chemical recycling is that it enables an integrated recycling process, where PET is depolymerized to its original constitution allowing the synthesis of new high value-added products 2 , 3 . Upgrading the utilization of undesirable agro-fibers as precursors for the production of carbon materials (activated carbon and carbon nanotubes), functional paper, artificial wood, hydrogels for agricultural purposes, as well as nanoparticles for controlling the release of fertilizer as well as bioactive and optical compounds 4 – 12 , is a beneficial approach, also for the production of low-cost products. Agricultural wastes in polymer composites have received interest in academic and industrial sectors. Many types of natural fibers have been used to reinforce the polymer composites 13 – 16 . The thermoplastics polymers used as matrices in the production of composites with natural fibers are high density (HDPE) and low density (LDPE) polyethylene, chlorinated polyethylene (CPE), polypropylene (PP), normal polystyrene (PS), and polyvinylchloride (PVC). Unfortunately, the relatively poor compatibility between the composite components noticed in the scanning electron microscopy (SEM) study 17 is a drawback. Different methods are performed to improve the compatibility by modifying the fiber matrix via adding coupling agents or accepting the fiber’s hydrophobic properties by grafting with hydrophobic monomers or esterification 15 , 16 , 18 , 19 . Moreover, extracting hemicellulose from fibers enhances its performance for improving the produced composite's water resistance and tensile strength of the produced composite 20 . Polymer–polymer composites are also reported in literature. Blending polyolefins with polyethylene terephthalate (PET) has involved significant study activity. Guerrero et al. 21 observed that the blending of PET increased the tensile strength property of HDPE; however, the reverse trend was noticed in the case of the elongation property. Studying the composite from recycled HDPE and PET in the presence and absence of compatibilizer, using a co-rotating twin screw extruder, proved to be effective in progressively increasing the interactions between two phases and enhancing the phase dispersion of the blends 22 . The mechanical properties of HDPE-based composites were also improved by adding sawdust, polyethylene- g -maleic anhydride, and commercial alumina additives. These additives also improved thermal and flammability properties 23 . Continuing our work on the valorization of lignocellulosic-byproducts in the production of valuable products, such as, carbon nanostructures, functional wood and paper as well as hydrogels for agricultural purposes and controlled release systems 4 – 7 , 9 , 23 , the current work was focused on enhancing the RS-based polyol in preparation of polyester polyols (PEPO) and further applied in HDPE-composites. The synergistic effect of glycoside polyethylene terephthalate waste (Gly-WPET) on RS-Polyol was evidenced from increasing the thermal stability and mechanical properties (tensile strength, elongation, young's modulus, and toughness) of PEPO. Additionally, it promotes the application of PEPO as a plasticizer and in the production of high-performance HDPE-PEPO composites.",
"discussion": "Results and discussion Characterization of RS-polyol and evidence of the polyesters formation Liquefaction of rice straw with ethylene glycol using sulfuric acid as a catalyst resulted in polyol with an average yield of 58.7%, while the solid residue was 41.34%. This liquefaction yield approach was found by Liang et al. 29 when they examined the liquefaction of crop residues in various solvents (ethylene carbonate, ethylene glycol, and polyethylene glycol). They found the liquefaction yield of crop residues using ethylene glycol was 60. Acid value and hydroxyl value were also estimated for liquefied rice straw, where acid value refers to the mass of potassium hydroxide (KOH) in mg required to neutralize one gram of examined compound. The acid and hydroxyl values for the prepared liquefied rice straw were 8.2 and 82.7 mg KOH/g, respectively. Different analyses were carried out for evidence of the formation of polyesters from RS-polyol and Gly-WPET, namely FTIR, 1 H NMR and molecular weight (MW) by GPC. The FT-IR spectra of RS-polyol (Fig. 1 ) showed a broad band at 3309 cm −1 indicating the presence of OH groups and the band at about 1050 cm –1 assigned to C–O–C asymmetry stretching. Observation of functional groups, such as –OH and C–O, implying that polyols were successfully prepared by liquefying rice straw under the studied conditions. The decrease in intensity of OH band on esterification with the sharp band at 1727 cm −1 , which matched the carbonyl group, evidenced the modification of polyol to polyester. Concerning the spectra of Gly-WPET and its polyester. Figure 1 FT-IR of polyol rice straw and polyester polyol. Figure 2 showed the following bands in the case of Gly-WPET: at 3437 cm −1 indicates the alcoholic group –OH; at 2985 cm −1 and 2864 cm −1 corresponds to the asymmetric and symmetric deformation of C–H stretching vibration, respectively. An intense band of the stretch carbonyl ester C=O is observed at 1713 cm −1 , together with bands at 1267 cm −1 and 1099 cm −1 of asymmetric and symmetric vibration of C–O ester. Similar to the spectrum of polyester from RS-polyol the reduction in the intensity of OH groups (3520 cm −1 ) and increasing in the intensity of peak at 1713 cm −1 which related to carbonyl ester group was observed in spectrum of polyester from Gly-WPET. Figure 2 FT-IR of Gly-WPET and its polyester. The formation of polyesters from RS-polyol and Gly-WPET was also evidenced by 1 H-NMR analysis. The chart of RS-polyol based polyester (Fig. 3 a) showed the signal at 5.0 ppm for hydroxyl protons (CH 2 –OH), as well as signal at 3.6 ppm which is assigned to the methylene proton attached to the primary alcohol group (CH 2 –OH). The signal at 4.085 ppm referred to the CH 2 –O (C=O) proton, while the signal at 2.315 ppm referred to CH 2 –O (C=O). Due to protons CH 2 –CH 2 –O–(C=O), the signal at 1.62 ppm appears 30 . In the case of 1 H NMR of polyester from the glycolysis product of WPET, Fig. 3 b showed a signal due to the proton of aromatic CH 2 of terephthalic acid at 8 ppm. The signal at 5.0 ppm is assigned to the hydroxyl protons (CH 2 -OH). Also, signals between 3.7 and 3.9 ppm are assigned to the methylene proton attached to the primary alcohol group, CH 2 –OH, while signals between 4.2 and 4.7 ppm and at 2.3 ppm are referred to the CH 2 –O (C=O), and CH 2 –(C=O) protons. respectively. Finally, the signal due to protons CH 2 –CH 2 –O–(C=O) appears at 1.64 ppm 16 . Figure 3 1 H-NMR of prepared polyesters ( a ) from RS-polyol, and ( b ) from Gly-WPET. The estimated number-average molecular weight (Mn), weight-average molecular weight (Mw), and polydispersity (PD = Mw/Mn) of polyesters from RS-polyol and its hybrid with glycolysis product of WPET (Gly-WPET) were recorded in Table 2 . The total Mn and Mw ranged from 2.47 × 10 5 to 3.14 × 10 5 and from 2.74 × 10 5 to 3.87 × 10 5 , respectively, during the esterification of liquefied rice straw and its hybrid with glycolysis product of WPET. Interestingly, using liquefied RS-WPET improved the monodisperse of the resulting polyester, where the PD (M w /M n ) was decreased from 1.232 to 1.089–1.109. Table 2 Molecular weight of polyesters from RS-polyol and its hybrid with glycolyzed WPET. PEPO sample GPC parameter M n M w PD PEPO from RS-polyol 3.14 × 10 5 3.87 × 10 5 1.232 PEPO from 90% RS-polyol + 10% Gly. WPET) 4.06 × 10 5 4.42 × 10 5 1.089 PEPO from 50% RS-polyol + 50% glycolysis WPET 2.47 × 10 5 2.74 × 10 5 1.109 Optimizing RS-polyol/Gl-WPET hybrid-based polyesters In this study, two effects were carried out to optimize the role of glycolysis WPET for preparing PEPO from RS-polyol, via changing the amount of 1,6-diisocyanatohexane as curing agents and the substitution ratio of polyol by glycolysis of WPET. Effect of curing agent In this study different ratios of1,6-diisocyanatohexane (10, 20, 30 and 50%) were added to cure the polyester from RS-polyol to specify the optimum ratio, which performed polyester with high mechanical properties. Figure 4 a–c illustrates the changes in mechanical properties, including tensile strength, elongation at break, young’s modulus, and modulus of toughness for prepared polyester versus the isocyanate ratio. The tensile strength increased gradually with increasing the isocyanate ratio from 10 to 50%, where the tensile strength and young's modulus increased from 0.22 to 1.28 MPa and from 3.77 to 47.47 MPa, respectively. Figure 4 Mechanical properties of polyester polyol from rice straw versus amount isocyanate curing agent ( a ) Tensile strength, ( b ) elongation at break and ( c ) young’s modulus. In contrast, the elongation at break was decreased with increasing the isocyanate ratio. The improvement in tensile strength and young's modulus may be ascribed to increased crosslinking density with increasing isocyanate ratios. The positive effect of curing agents on mechanical properties agrees with that reported by Lee et al. 25 . Effect of substituting RS-polyol by Gly-WPET From the preceding data, 50% curing agent was provided from polyol with relatively high strength properties. This percentage was used for further study to evaluate the role of substituting RS-polyol by Gly-WPET on the properties of produced polyester. The obtained data are illustrated in Fig. 5 , which shows the positive effect of replacing a part of RS-polyol by Gly-WPET. The tensile strength of polyester increased gradually with the substitution ratio of glycolysis WPET (increased from 1.28 MPa to 2.92, 3.19, and 3.7 for substitution 10%, 20%, and 30%, respectively. Further increase in the substitution ratio to 50% also improved the tensile strength of polyester (2.45 MPa) but less than 30% (Fig. 5 a). Figure 5 ( A ) Mechanical properties of polyester polyols from substituting the RS-polyol by Gly-WPET ( B ) toughness for stress–strain curve. The elongation at break and Young's Modulus (Fig. 5 a) increased with increasing the Gly-WPET till 50% (from 4.7 to 9.8% and from 47.47 to 66.48 MPa, respectively). As can be seen that the glycolysis of WPET provided improvement in modulus of toughness which was estimated from area of stress–strain curves (Fig. 5 b), and the maximum improvement at 30% substitution [23.03 (J/m 3 )/control 3.98 J/m 3 ]. The influence of Gly-WPET on polyester from RS-polyol was also clear from FTIR spectra and TGA analysis. Figure 6 showed that the FT-IR spectra of all polyesters of hybrid liquefied samples included the same bands, but the only difference in increasing the intensity of bands at 1249 and 750 cm −1 with increasing the substitution ratio from 10 to 50%, which indicated the formation of a polymer network between polyester and curing agent and creation of the ether linkage. Moreover, the terephthalate group (OOCC 6 H 4 COO) was included in the product. The disappearance of peak N=C=O at 2250 cm −1 confirming that diisocyanate of curing agents completely reacted with polyester polyol 24 . Figure 6 FTIR of polyester from RS-polyol and its hybrid with 10% (PEPO 10), 20% (PEPO 20), 30% (PEPO 30) and 50% (PEPO 50) Gly-WPET. Regarding the effect of substituting RS-polyol by Gly-WPET on the resulting polyester's thermal stability, Fig. 7 and Table 3 illustrated that the PEPO from liquefied RS was decomposed in two stages. First at lower than 100 °C, which was attributed to water dehydration, while the second at 250–500 °C regarded as the main degradation stage and related to the degradation of urethane linkage and bonds of network formation between polyester and the curing agent. The peak maximum of this main degradation peak appeared at 347.2 °C with an onset temperature of 301.0 °C. The resulting polyester from substituting the RS-polyol with Gly-WPET the decomposition proceeded in two and three stages. These additional stages are attributed to the dissociation of the interaction between RS-polyol, Gly-WPET, and curing agent with the formation of different strength bonds, in addition to the decomposition of the polymer chain of WPET, especially at a relatively high substitution ratio, which usually started to degrade at 375–400 °C according to the literature 31 , 32 . As can be noticed (Table 3 ), the percentage of residual weight increased with the substitution of PET from 37.7 to 45.5%. This is back to the fact that thermal pyrolysis of PET left much solid residue that did not decompose at examined temperature (600 °C). This is attributed to the interlinking reaction between the decomposed products of PET and with the formation of stabilized products 33 . Figure 7 TGA and DTG curves of polyesters from RS-polyol and its hybrid with Gly-WPET. Table 3 TGA and DTGA measurements of the main degradation stage of polyesters from RS-polyol and its hybrid with different ratios od Gly-WPET. Sample code Decomposition step Temp. range (°C) DTG peak (°C) Onset temp (°C) Weight loss% Ash % at 1000 °C n order R 2 S e E a kJ/mol PEPO-0 2nd 271.5–505.9 347.2 301.0 50.990 37.77 2.000 0.966 0.287 184.516 ∑ E a = 184.5 PEPO -10 2nd 264.9–408.9 331.2 282.4 37.110 37.03 2.000 0.965 0.206 181.604 3rd 405.4–483.9 440.8 12.360 2.000 0.953 0.219 424.532 ∑ E a = 606.1 PEPO -20 2nd 271.3–364.7 333.4 293.0 22.617 45.56 2.000 0.965 0.227 273.677 3rd 365.7–415.7 386.2 14.308 2.000 0.936 0.212 464.153 4th 416.7–498.2 443.5 13.880 2.000 0.964 0.209 458.512 ∑E a = 1196.3 PEPO -30 2nd 272.1–351.5 324.2 295.6 18.743 38.81 2.000 0.949 0.216 265.340 3rd 354.4–415.5 384.5 20.445 2.000 0.941 0.235 394.969 4th 413.7–480.5 444.7 13.037 2.000 0.953 0.220 500.428 ∑E a = 1160.7 PEPO -50 2nd 289.9–347.5 322.3 303.9 19.704 41.47 1.500 0.993 0.064 301.328 3rd 347.5–409.0 377.9 16.021 2.000 0.937 0.221 381.067 4th 409.0–480.1 448.4 18.839 2.000 0.954 0.228 518.728 ∑E a = 1201.1 The kinetic parameters (temperature range, DTG peak temperature, weight loss, correlation coefficient (R 2 ), order \"n\", standard error estimate (S e ), and activation energies (E a ) of the main degradation in all samples are gathered in Table 3 . In all cases, substituting RS-polyol with liquefied WPET provided lower onset temperature and DTG peaks than the stage of polyester from pure RS-polyol. At the same time, their activation energies increased gradually with increasing the substitution percentage. The activation energies of PEs from hybrid liquefied samples ranged from 606 to 1201 kJ/mol, while E a of polyerster from pure RS-polyol was 184.5 kJ/mol). This indicated that liquefied WPET enhanced the bond formation during the esterification of RS-polyol, which provided thermal stabilization in the final product, which requires relatively higher E a to degrade. Assessment of HDPE-PEPO composites Blending polyolefins and polyesters have attracted considerable research because they are among the most consumed plastics. Therefore, HDPE was blended with the prepared polyesters. Blending these two materials could offer a very attractive balance of mechanical properties. Figure 8 showed that incorporating the polyesters, especially 5%, from RS polyol and Gly-WPET, individually or in hybrid (50:50), positively affected the mechanical properties of HDPE. The improvement in elongation was more significant than the tensile strength, whereas the tensile strength improved from 21.1 MPa to 22.0, and 21.9 MPa, respectively. Above this ratio (10%), the tensile strength decreased. At the same time, the elongation at break increased from 24.8% for pure HDPE to be 55.4%, 64.6 and 72.0% for HDPE/5% polyester of RS-polyol, HDPE/5% hybrid polyesters, and HDPE/5% polyester WPET. As can be noticed, blending HDPE with 10% PEPO from hybrid RS-polyol with Gly-WPET, also improved the elongation as in the case of 5% WPET-PEPO. This promising behavior of 10% PE rather than 5% was also noticed in the case of toughness, where the area of stress–strain curve increased from 246.4 to 597.1 and 700.6 T (this is the wrong unit, should be J/cm 3 , please double check) for 5% and 10% PE blends. Based on the definition of plasticizer and the obtained data, we recommend the utilization of these PEPOs as a plasticizer. Figure 8 Mechanical properties of high-density polyethylene-polyester polyol (HDPE-PEPO) composites. ( a ) Tensile strength, ( b ) elongation, ( c ) stress–strain curve and ( d ) Toughness. The morphology of HDPE and HDPE-PEPO blends with different types and content from prepared polyesters is shown in Fig. 9 A,B. Figure 9 a clarifies a good dispersion of silica fillers on the HDPE matrix surface. Figure 9 b–d illustrates the micrographs of HDPE–PEPO blends incorporating 5% PEPO from RS-polyol and Gly-WPET individually or in hybrid. These micrographs clarify the adhesion of polyester and silica fillers to the HDPE matrix was good at low content. The good adhesion results in enhancing the mechanical properties, which agrees with the other conclusion from our experimental results. In comparison, Fig. 9 e–g showed that the adhesion of PEPO to the HDPE matrix was poor at high loading (10%). The weak adhesion results in damage or alteration of the mechanical properties. The SEM micrograph of higher PEPO loading showed some aggregates indicating inadequate additive dispersion within the HDPE matrix. These aggregates have become more present and more significant in size by increasing the ratio of polyester to 10%, especially in the case of WPET-100. This observation emphasized te trend of mechanical properties illustrated in Fig. 8 . Figure 9 ( A ) SEM morphology of high-density polyethylene (HDPE) and HDPE blended with different types and 5% polyesters. ( B ) SEM morphology of high-density polyethylene (HDPE) and HDPE blended with different types and 10% polyesters."
} | 5,292 |
31908923 | PMC6940699 | pmc | 5,649 | {
"abstract": "Cyanobacterial biofuels have the potential to reduce the cost and climate impacts of biofuel production because primary carbon fixation and conversion to fuel are completed together in the cultivation of the cyanobacteria. Cyanobacterial biofuels, therefore, do not rely on costly organic carbon feedstocks that heterotrophs require, which reduces competition for agricultural resources such as arable land and freshwater. However, the published product titer achieved for most molecules of interest using cyanobacteria lag behind what has been achieved using yeast and Escherichia coli ( E. coli) cultures. In Synechocystis sp. PCC 6803 ( S . 6803), we attempted to increase the product titer of the sesquiterpene, bisabolene, which may be converted to bisabolane, a possible diesel replacement. We tested 19 strains of genetically modified S . 6803 with five different codon usage sequences of the bisabolene synthase from the grand fir tree ( Abies grandis ). At least three ribosome binding sites (most designed using the RBS Calculator) were tested for each codon usage sequence. We also tested strains with and without the farnesyl pyrophosphate synthase gene from E. coli . Bisabolene titers after five days of growth in continuous light ranged from un-detected to 7.8 mg/L. Bisabolene synthase abundance was measured and found to be well correlated with titer. Select strains were also tested in 12:12 light:dark cycles, where similar titers were reached after the same amount of light exposure time. One engineered strain was also tested in photobioreactors exposed to a simulated outdoor light pattern with maximum light intensity of 1600 μmol photons m −2 s −1 . Here, the bisabolene titer reached 22.2 mg/L after 36 days of growth. Dramatic improvements in our ability to control gene expression in cyanobacteria such as S . 6803, and the co-utilization of additional metabolic engineering methods, are needed in order for these titers to improve to the levels reported for engineered E. coli .",
"discussion": "4 Discussion Following other work to generate bisabolene from microbial cultures ( Davies et al., 2014 ; Peralta-Yahya et al., 2011 ), we sought first to show that bisabolene could be produced in genetically modified S. 6803. The bisabolene synthase from A. grandis was selected as it had shown the highest activity in E. coli among four codon-optimized genes from species of trees ( Peralta-Yahya et al., 2011 ). We anticipated that S. 6803 may have a small concentration of the precursor metabolite, farnesyl pyrophosphate (FPP), because it appears to lack a dedicated farnesyl pyrophosphate synthase. Instead, the production of FPP in S. 6803 likely depends on geranylgeranyl pyrophosphate synthase sometimes releasing FPP before another isopentenyl pyrophosphate reacts with FPP in the active site. Therefore, we co-expressed FPP synthase from E. coli , codon optimized by GenScript for expression in S. 6803, in an operon structure with bisabolene synthase. Transcription of this biscistronic mRNA was initiated by the tic promoter. This promoter has relatively strong expression in S. 6803 ( Albers et al., 2015 ), and the presence of two lac operators maintains low expression until induced by IPTG. This was expected to be useful in the event that the expression of the two genes was toxic to S. 6803. We did not, however, find there to be significant impacts on growth from the expression of bisabolene synthase. The initial strain produced 1.6 ± 0.2 mg bisabolene/L after 5 days or 0.31 ± 0.04 mg/L/day, a significantly higher titer than previously reported 0.6 mg/L bisabolene titer in cyanobacteria after 4 days ( Davies et al., 2014 ). Building on an initial proof of concept that bisabolene synthase from A. grandis could be expressed and functional in S. 6803, we sought to increase the expression of this gene and increase the bisabolene titer achieved. Lacking a high-throughput screen for bisabolene production, we constructed a set of S. 6803 strains with varied codon usage (five variants) and varied RBS sequence (three or four RBS sequences for each codon usage variant). The simple measure of CAI correlated well with bisabolene titer when compared using the same RBS. All three commercially designed gene sequences could be used to generate functional bisabolene synthase. When RBS sequences were varied, the range of bisabolene titers achieved using each of these three codon optimizations were significantly overlapping. Kudla et al. (2009) measured the fluorescence of 154 different coding sequences of gfp expressed in E. coli and found that the stability of mRNA secondary structure near the RBS accounted for about half of all variation in the expression level, while the CAI was a poor predictor of fluorescence ( Kudla et al., 2009 ). This suggests that it may also be possible to improve heterologous gene expression without changing much of the sequence, and thereby avoid the requirement to synthesize the full gene sequences which may or may not be expressed at a higher level than the wild-type sequence. The RBS sequences tested in this work show that the sequence of the 5′ untranslated region clearly impacts gene expression, but at this time we can’t make accurate predictions about the expression level of a gene in S . 6803 based only on this sequence. The RBS Calculator poorly predicted the relative expression of bisabolene synthase in S. 6803. Similar results were found by Wang et al. (2018) when expressing ethylene forming enzyme in S. 6803. This tool was originally validated using fluorescent protein expression in E. coli and, later, also in Pseudomonas fluorescens , Salmonella typhimurium LT2, and Corynebacterium glutamicum ( Farasat et al., 2014 ; Salis et al., 2009 ). Although we do expect translation to be a similar process across species of prokaryotes, the only concession the RBS Calculator makes to differences between organisms appears to be the 16s rRNA sequence which interacts with the RBS in the mRNA. S. 6803 and E. coli have nearly identical sequences in the last 14 nucleotides at the 3′ end of their 16S rRNA and, therefore, The RBS Calculator uses the same anti-Shine-Dalgarno sequence for S. as it does for E. coli (only the final nucleotide of the 16s rRNA differs between these organisms). The translation initiation rates predicted using the RBS Calculator for the sequences used in study are generally the same or similar between the two organisms. The ability to accurately predict relative gene expression rates would facilitate the development of more complex genetic circuits in S. 6803 for applications in industrial biotechnology. Genetic studies provide some evidence that translation may be different between E. coli and S. 6803. For example, a smaller proportion of genes in S. 6803 than in E. coli appear to be initiated by Shine-Dalgarno (SD) sequences (26% for S. 6803 versus 57% for E. coli ) ( Ma et al., 2002 ). It is not clear whether SD sequences should be expected to increase translation initiation rates in S. 6803 as much as they do in E. coli . SD-antiSD hybridization is thought to reduce the impact of mRNA secondary structure on translation initiation ( de Smit and van Duin, 1994 ). However, cyanobacterial mRNA that lack SD sequences are generally predicted to have weaker secondary structure adjacent to the start codon (on either side) than the mRNAs of γ-proteobacteria that lack a SD sequence ( Scharff et al., 2011 ). Optimal spacing between the start codon and SD sequence can also influence translation initiation rates. RBS Calculator penalizes deviation in spacing from five nucleotides between the SD and start codon ( Salis et al., 2009 ). Ma et al. found the most frequent spacing for between SD and the start codon to be five for E. coli , and seven for S. 6803 ( Ma et al., 2002 ). We were also surprised that cultivation of the best performing strain in photobioreactors supplemented with CO 2 did not significantly increase the product titer beyond what was measured from shake flask cultivation. We expected these cultures to reach higher optical densities in the CO 2 -rich environment and also to have higher precursor availability for bisabolene synthase. However, both the maximum optical density and the titer after five days growth were similar to what was measured from shake flask cultures. It is likely that, in both cases, a nutrient other than carbon becomes limiting and prevents further growth and bisabolene production. For example, Clark et al. (2018) increased the concentration of nitrate by more than 10-fold, the concentration of phosphate by nearly 80-fold and the concentration of iron by more than 35-fold in Media A to reach higher cell densities of Synechococcus sp. PCC 7002 ( Clark et al., 2018 ). In both shake flask and PBR cultures subjected to light:dark cycles, bisabolene titer continued to increase after the cultures reached stationary phase. The bisabolene titer in PBR cultures increased significantly after they reached stationary phase and the specific titer increased from 0.26 ± 0.03 mg/L/OD after 14 days of growth to 2.6 ± 0.19 mg/L/OD after 36 days of growth. This result suggests that the methylerythritol 4-phosphate pathway may remain active during stationary phase, and continue to make IPP, DMAPP, and FPP available for bisabolene synthesis. Carotenoid production is also reliant on the MEP pathway and may increase during the stationary phase as the cells experience low light in the high-density culture. Carotenoids also have photoprotective properties, and their increased production may also be necessary for the periods of time that the cell are exposed to the high light intensity of the surface of the culture. These results are similar to the findings of others. In terms of the product titer, Davies et al. (2014) expressed the same DNA2.0 codon optimized bisabolene synthase sequence in Synechococcus sp. PCC 7002 using a cpcBA promoter from S. 6803. Their tests resulted in a bisabolene titer of 0.6 mg/L after 96 h of growth in continuous light (6 μg/L/hr). Wichmann et al. (2018) engineered strains of the algae Chlamydomonas reinhardtii to produce titers of 3.9 mg/L of bisabolene in phototrophic conditions and 11.0 mg/L in mixotrophic conditions after seven days of growth. Wang et al. (2018) tested a small set of RBS sequences driving expression of ethylene forming enzyme in S. 6803, achieving a 2.5-fold increase above the base strain. We utilized a solvent layer of dodecane to collect produced bisabolene, following the work of others ( Davies et al., 2014 ; Peralta-Yahya et al., 2011 ). The dodecane layer formed an emulsion after a few days of growth in both the shake flasks and the photobioreactors. Occasionally, samples of this emulsion required brief centrifugation in order to collect a completely organic sample. We did not find any growth inhibition due to the addition of dodecane in shake flasks and bisabolene was not detected following extraction ( Bligh and Dyer, 1959 ) of cells harvested from cultures grown with dodecane (data not shown). Molecular dynamics models have supported the hypothesis that bisabolene may diffuse through the cell membrane into aqueous media, or, more rapidly, into dodecane in contact with the cell ( Vermaas et al., 2018 ). Conversely, it has also been reported, in studies using Dunaliella salina , that a stagnant dodecane layer did not result in cell death while dodecane sparged through a culture did result in cell death ( Kleinegris et al., 2011 ). The collection of excreted, hydrophobic, and volatile products from cyanobacterial or microalgal cultures which are grown outside to utilize sunlight is an area that requires further research into both the mechanism by which they are excreted (as this may limit the production rate), as well as into cost effective means of capturing the excreted product. Our findings reinforce that it continues to be necessary to test combinations of genetic components in S. 6803 to obtain a desired outcome because present tools for predicting the effects of the components are inadequate. Combinations of RBSs and codon optimizations resulted in 3.3-fold increase in the concentration of bisabolene synthase over our original strain. However, even the highest expressing strain, bisabolene production likely remains limited by the expression of this enzyme. In this strain, 2.0–10xB, bisabolene synthase was estimated at just 0.4% of the total protein. Continued progress in metabolic engineering of cyanobacteria would benefit from improved understanding of translation initiation mechanisms in cyanobacteria and how they may differ from our understanding of translation initiation mechanisms in E. coli . Further improvements to genetic component design rules are needed to reduce the impacts of context dependence on the function of the components."
} | 3,241 |
39836436 | PMC11750059 | pmc | 5,650 | {
"abstract": "ABSTRACT With many species interacting in nature, determining which interactions describe community dynamics is nontrivial. By applying a computational modeling approach to an extensive field survey, we assessed the importance of interactions from plants (both inter‐ and intra‐specific), pollinators and insect herbivores on plant performance (i.e., viable seed production). We compared the inclusion of interaction effects as aggregate guild‐level terms versus terms specific to taxonomic groups. We found that a continuum from positive to negative interactions, containing mostly guild‐level effects and a few strong taxonomic‐specific effects, was sufficient to describe plant performance. While interactions with herbivores and intraspecific plants varied from weakly negative to weakly positive, heterospecific plants mainly promoted competition and pollinators facilitated plants. The consistency of these empirical findings over 3 years suggests that including the guild‐level effects and a few taxonomic‐specific groups rather than all pairwise and high‐order interactions, can be sufficient for accurately describing species variation in plant performance across natural communities.",
"introduction": "1 Introduction A central aim in ecology is to understand the maintenance of species diversity (Levin 1970 ; Hobbs and Mooney 1985 ). Niche‐based explanations for the mechanisms underlying species coexistence rely on the demonstrated importance of biotic interactions for species performance (e.g., growth, fecundity). These effects have been widely studied by combining phenomenological models and experimental manipulations using pairs of competing species within the same trophic guild (Levine and HilleRisLambers 2009 ; Godoy and Levine 2014 ). Yet, ecologists are only recently quantifying these pairwise interaction strengths for whole communities or considering a wider set of interactions within and across trophic guilds simultaneously (but see García‐Callejas et al. 2023 ; Chang et al. 2023 ; Bimler et al. 2024 ). On the contrary, research on food webs and plant‐pollinator systems usually examines the structure and complexity of the interaction network without integrating within guild interactions (Godoy et al. 2018 ; Vitali et al. 2023 ). As such, it remains untested how the structure of biotic interactions and the relative importance of within‐ versus across‐guild interactions drive species performance and thus the maintenance of local diversity (Pilosof et al. 2017 ). Bimler and Mayfield phenomenological models of population growth that evaluate coexistence based on an individual performance framework (hereafter called individual performance models) provide context‐dependent results. Indeed, they assume that the most expected (not necessarily the most likely in a statistical sense) nature of an interaction between a species pair is the only possible one (Simha et al. 2022 ). For instance, first, they assume that plants always compete for resources (Craine and Dybzinski 2013 ; Lanuza, Bartomeus and Godoy, 2018 ; Johnson and Hastings 2022 ). Second, based on their mutualistic behavior, pollinators always have positive impacts on plant performance (Vázquez et al. 2015 ; Aizen 2021 ). Finally, herbivores act as natural enemies and thus always have negative impacts (Barber et al. 2012 ; Aguirrebengoa et al. 2023 ). The signs and strengths of interactions between two types of species (or the same type of species) can and do, however, produce different effects as circumstances change. For example, some plants facilitate each other more strongly than they compete for resources (Bimler et al. 2018 ); some pollinators have negative effects on plants (Magrach et al. 2017 ); and some herbivores have positive effects on plant performance by promoting further growth (Génin et al. 2021 ). Such counterexamples are not unusual and their presence in nature begs the question of whether or not we might want to take a more holistic perspective when studying coexistence and the diversity maintenance of real communities—notably one that allows for a continuum of negative and positive species interactions regardless of trophic guild and type of interaction (Koffel, Daufresne, and Klausmeier 2021 ; Gómez, Iriondo, and Torres 2023 ; Bimler et al. 2023 ; Allen‐Perkins et al. 2023 ). In addition to simplifying interactions to a single nature (sign), most individual performance models also assume that interactions are all pairwise and direct (Mayfield and Stouffer 2017 ). Multispecies interactive effects have, however, been shown to be common and important in many natural systems (Bimler and Mayfield 2023 ). It remains unclear if ignoring multispecies interactions in individual performance models has minimal effects on model performance or if this omission removes important biological realism. A common way to incorporate multispecies interactions is to allow for higher‐order interactions (HOIs), which occur when an interaction between two species is modified by the presence of a third (Li et al. 2021 ). Awareness of the importance of HOIs has increased with growing interest in applying theories designed for pairwise interactions to multispecies natural systems (Levine et al. 2017 ; Mayfield and Stouffer 2017 ; Bimler and Mayfield 2023 ; Buche, Bartomeus, and Godoy 2024 ). The few empirical studies that have assessed the importance of HOIs in natural systems have occurred for plant–plant interactions (Mayfield and Stouffer 2017 ; Lai et al. 2022 ; Li et al. 2021 ), microbe interactions (Ishizawa et al. 2024 ), arthropod interactions (Barbosa, Fernandes, and Morris 2023 ) and for one type of interaction between two contrasted guilds—plants and pollinators (Buche, Bartomeus, and Godoy 2024 ). These studies all align with theoretical expectations that HOIs are important factors in predicting species' performances (Bairey, Kelsic, and Kishony 2016 ; Kleinhesselink et al. 2022 ; Gibbs, Levin, and Levine 2022 ). Despite the literature increasingly supporting the importance of HOIs, most individual performance models omit HOIs, allowing only for direct pairwise interactions (Li et al. 2021 ). Attempting to include additional complexity in individual performance models yields several well‐understood challenges. These models are prone to over‐fitting under the classic assumption that each interacting species offers unique insights into the dynamics of a species' performance (Bimler et al. 2023 ). However, from the macroevolution literature, we know that closely related species are often more similar than distantly related species (Cavender‐Bares et al. 2009 ) (e.g., compare forbs vs. legumes) and that most natural communities exhibit functional redundancy, the phenomenon in which many species have the same ecological role in a given community (Laliberte et al. 2010 ; Chang et al. 2023 ). These frameworks suggest that groups of species with similar functional characteristics might interact similarly (Barbier et al. 2018 ) and can be potentially lumped to simplify models; however, how to best group species remains an open question. Traditionally, species have been grouped based on taxonomy or functional groups (Martyn et al. 2020 ; Uriarte et al. 2004 ; Straub, William, and Snyder 2006 ). A more powerful approximation is to use sparse matrix modeling approaches, which allow us to identify what interaction strengths and at which grouping level parameters are and are not important for parameterizing individual performance models, allowing us to focus only on the unique species interactions that actively affect the performance of a focal species (Hastie, Tibshirani, and Wainwright 2015 ; Weiss‐Lehman et al. 2022 ). Here, we examine the degree of complexity in the nature and structure of biotic interactions necessary to explain the performance of plant species (i.e., seed production, Figure 1a ). Specifically, we test whether the details of within and cross‐species interactions are required to explain plant performance in a highly diverse community. We address this question by comparing effects aggregated across species in each trophic guild (‘guild‐level terms’) with details unique to specific functional or taxonomic groups (e.g., family grouping). We coupled data on species abundances and plant performance collected across 3 years in an annual plant community in southern Europe with a Bayesian sparse matrix modeling approach. This approach explores whether the inclusion of four interaction types and two sources of complexity improves the description of individual performance for four focal annual plant species. The four types of direct interactions considered mirror the trophic guilds present in the system: intraspecific plant interactions, interspecific plant–plant, plant–pollinator and plant–herbivore interactions. The two sources of complexity are the inclusion of higher‐order interactions and the variation in sign and strength of net interactions. Examining patterns in interactions' nature, we answered the following questions: (i) Are all pairwise plant–plant, plant–pollinator and plant–herbivore interactions and their potential three‐way HOIs necessary for describing observational patterns of plant performance? (ii) How do within‐ and cross‐species interactions vary in signs and strengths (from negative to positive and strong to weak)? (iii) Is there consistency in which interactions increase model performance and their nature (sign and strength) across years? FIGURE 1 Study system and methodological illustration: Depiction of the four annual plant species studied: \n Chamaemelum fuscatum \n (CHFU), Leontodon maroccanus (LEMA), \n Hordeum marinum \n (HOMA) and \n Centaurium tenuiflorum \n (CETE), from left to right. We collected data on the number of seeds produced, plant neighbors, herbivores and pollinators for each focal species (panel a). These observations were used to fit an individual performance model to estimate species interactions. The effects could be positive (green) or negative (yellow) based on whether an individual promotes or harms the performance of the focal annual plant. Each trophic level (panel b for herbivores, panel c for interspecific plants and panel d for pollinators) has a guild‐level effect that aggregates across interactions and taxonomic‐specific deviations from the guild‐level effect (red distribution; represented by colored species). For example, a ‘grass‐specific effect’ could diverge from the plant‐level effect. Credit to Nerea Montesperez for the illustration and Biorender. Extended version is shown in Figure S12 .",
"discussion": "4 Discussion By combining multiple years of extensive field data collection in a Mediterranean grassland with a new method for estimating the nature and strength of species interactions (Weiss‐Lehman et al. 2022 ), our study quantifies the importance of four common types of ecological interaction while allowing for higher‐order interactions on plant performance. Our results are based on an adaptive and computational robust modeling approach, which can be applied in highly diverse systems without over‐fitting the model. They show, at least for our system, that cross‐trophic interaction types, while overall weak, vary from positive to negative signs. Only a few taxonomic‐specific direct interactions and cross‐trophic HOIs were needed to characterize plant reproductive success within our study system. These patterns were robust to the different groups we examined, regardless of grouping at the species, family, or functional group. Our result provides a welcome starting point for other researchers when deciding how to simplify potential interactions to include in diverse, multitrophic models—especially if applying non‐sparse statistical frameworks, where models would fail to converge if including all potential species interactions. Detailed exploration of guild‐level direct terms points to a more nuanced story. Our findings suggest that though most species within trophic guilds have redundant effects, one individual term at the species, family or functional group level often emerged as diverging from the relevant guild‐level term. Such taxonomic‐specific terms often coincided with dominant species. Further, while there was strong evidence that the nature of guild‐level direct interactions varied extensively between positive and negative (Figure 3 ), most net, abundance‐weighted effects were very weak (Figure 4 ), except for interspecific plant–plant interactions, which tended to be strongly negative (i.e., competitive). Our study clearly shows that most single species within a trophic level do not have distinct effects on plant performance, but some species do, and these distinct effects are strong and important. The redundancy in statistical estimates of species interaction strengths for most species within trophic guilds was surprising, given the diversity of the functional and evolutionary details in this highly speciose community. The weak importance of taxonomic‐specific effects highlights that the density of neighbors matters more than their identity. For instance, to empirically estimate pollinators' effects on plant performance, the number of total visits is critical; who is doing the visiting might not be so much (Vázquez and Simberloff 2002 ). This is because the log‐normal distribution of species abundance commonly found in nature (McGill et al. 2007 ; Cadotte and Tucker 2017 ) decreases the detection of taxonomic‐specific effects over the common effect of a trophic guild (Lewis et al. 2023 ). Additionally, the redundant effect of species within trophic guilds on ecological function is commonly found in conservation ecology (Walker 1992 ; Biggs et al. 2020 ). Like the functional redundancy principle, most species might have a generalizable effect on ecological patterns, with a few key species having disproportionate importance. Identifying the key taxonomic groups that exhibit a divergent interaction effect from their trophic guild has important ramifications for ecological theory and conservation strategy (Walker 1992 ). For instance, the “beetle” group and the Nitidulidae family in particular, were found to have a specific positive effect on Leontodon maroccanus (LEMA) in 2020. Within the Nitidulidae family, the genus Brassicogethes , also known as pollen beetle, is the most abundant in our system (Figure S9 ). These beetles are generalist pollen and nectar feeders (Seimandi Corda et al. 2018 ), with adults and juveniles moving around flower heads (Wäckers, Romeis, and van Rijn 2006 ). Yet, they are specialized on the Brassicaceae plant family for oviposition (Seimandi Corda et al. 2018 ). The negative impact of flower beetles may be mainly restricted to their host plant, as suggested by Seimandi Corda et al. ( 2018 ), and act as pollinators in other instances, especially facilitating pollen transport within a single flower in self‐compatible plants such as LEMA (Hurtado, Godoy, and Bartomeus 2023 ). Additionally, we found that grasses, when grouped at the functional level (‘Grass’), family level (Poaceae) or species level ( \n Hordeum marinum \n , HOMA), had an effect that diverged from other heterospecific plants on LEMA. While plant interactions are dominantly negative, the specific effect of grass individuals reinforced such competition. LEMA and HOMA are the most abundant in this system, which could explain their strong competitive effect; we speculate this could occur in other grassland ecosystems as the Poaceae and Asteraceae families have distinct ecological strategies leading them to be efficient invaders in many systems around the world (Huang et al. 2024 ). Understanding if these families can coexist with the rest of the plant community despite their strong competitive abilities or are slowly excluding others from the system is critical to predicting the accurate state of the community and potentially managing it (Aoyama et al. 2022 ). Similarly, understanding the importance of higher‐order interactions in the speciose multitrophic systems needs further attention. We found little evidence of the importance of HOIs despite the one occurrence in 2020, for LEMA. The effect of the ‘bee’ grouping on LEMA reinforced intraspecific competition yet had a marginal realized effect on its performance (< 1%). Negative effects of pollinators on intraspecific interaction can occur through dilution effect (Benadi and Pauw 2018 ) and/or stronger competition for the attention of potentially scarce pollinators (Lázaro, Lundgren, and Totland 2009 ). Despite the high sensitivity of detecting HOIs with our approach, based on simulated data (95.6%, Appendix S4.12 ), few HOIs were detected in our system. The lack of detectable HOIs is perhaps surprising as there is evidence of their importance in other systems (e.g., Mayfield and Stouffer 2017 ; Bimler and Mayfield 2023 ). A notable difference between our studies and others is the use of aggregate groups of species and the inclusion of multiple trophic levels. HOIS may be more detectable with only species‐level groupings. Given the different approach we used to other studies, it is important to note that our study does not conflict with other empirical studies that have found significant HOIs (Buche, Bartomeus, and Godoy 2024 ), or theoretical within‐guild models (Mayfield and Stouffer 2017 ; Barbosa, Fernandes, and Morris 2023 ; Lai et al. 2024 ). Certainly, our results suggest the need for further investigations of the importance of HOIs in complex natural communities. Previous studies investigating species interactions have restrained them to a priori directions in their effects (Gómez, Iriondo, and Torres 2023 ; Bimler et al. 2023 ). Allowing for interactions to vary along a continuum of positive‐to‐negative effects revealed that the effect of pollinators on plants was primarily positive while heterospecific plant‐on‐plant interactions were mainly competitive, as expected (Ollerton et al. 2011 ; Rodger, Bennett, and Razanajatovo 2021 ; Adler et al. 2018 ; Yang et al. 2022 ); yet, surprisingly, intraspecific plants and herbivore interactions with plants were highly variable, including some strong positive effects. Given that we used seed set as our proxy for performance, a positive effect from herbivory may have resulted from the allocation of more energy to seed production due to the stress induced by leaf damage (Bartomeus, Gagic, and Bommarco 2015 ), or increased growth in compensation for the removal of aging (or young) tissues. Similarly, the positive effects of some intraspecific plants counter most theoretical expectations. Still, such positive effects are commonly observed in nature, especially in populations persisting at low densities (aka Allee effect) (Heyes et al. 2020 ; Bowler et al. 2022 ) or in the presence of favorable micro‐environments (Bimler et al. 2018 ). While positive interactions among individuals of the same species can lead to uncontrolled population growth (Hart 2023 ), this positive‐feedback loop may be limited by negative effects from individuals of different species (Sheley and James 2014 ), the presence of higher trophic levels (Cervantes‐Loreto et al. 2021 ), or temporal variations in the direction of interactions within the same species (Zou, Yan, and Rudolf 2024 ), as evidenced in this study. Except for heterospecific plants, the overall net strength of species interactions across trophic levels was weak, suggesting an emerging neutrality in our system. This finding is aligned with classic ecological theory, which posits that the feasibility and stability of ecological systems are promoted by weak species interactions (May 1972 ; Yang et al. 2023 ). The effect of the heterospecific plants was, however, strong in some instances—showing a potentially strong competition for resources in a system with strong annual climatic variation. While the 3 years considered have relatively similar precipitation regimes (Figure S3 ), explicitly accounting for interannual precipitation variation might elucidate additional mechanisms of such competition (Bimler et al. 2018 ; Hallett et al. 2019 ). Indeed, the tendency of our system towards neutrality might indicate that fluctuation‐dependent mechanisms, such as the spatial and temporal storage effect (Tan et al. 2017 ) or relative non‐linearity (Hallett et al. 2019 ), could play an important role in driving coexistence in our system. Without further study, however, our results cannot be used to determine which, if either, of these mechanisms is involved in maintaining the diversity of this system, but targeted experiments to test for these mechanisms are a high priority for future studies. Overall, our findings provide critical empirical evidence on the nature and strength of species interactions in a highly‐speciose ecosystem. Our system presents, on average, guild‐level weak effects rather than being involved in complex sets of pairwise and higher‐order interactions. As these effects range from positive to negative interactions, they should not be predefined with one specific direction but allowed to vary along a continuum. This does not mean we should study complex systems by assuming a random structure of biotic interactions. Instead, we should identify the redundant effects within trophic guilds and the specific interactions that deviate from this redundancy. This can be particularly important for future theoretical work on diversity and conservation strategies for managing strong competitors. Lessons from this study advance our understanding of the structure of biotic interactions under high‐dimensional natural systems."
} | 5,435 |
39095421 | PMC11297175 | pmc | 5,651 | {
"abstract": "Most triblock copolymer-based physical hydrogels form three-dimensional networks through micellar packing, and formation of polymer loops represents a topological defect that diminishes hydrogel elasticity. This effect can be mitigated by maximizing the fraction of elastically effective bridges in the hydrogel network. Herein, we report hydrogels constructed by complexing oppositely charged multiblock copolymers designed with a sequence pattern that maximizes the entropic and enthalpic penalty of micellization. These copolymers self-assemble into branched and bridge-rich network units (netmers), instead of forming sparsely interlinked micelles. We find that the storage modulus of the netmer-based hydrogel is 11.5 times higher than that of the micelle-based hydrogel. Complementary coarse grained molecular dynamics simulations reveal that in the netmer-based hydrogels, the numbers of charge-complexed nodes and mechanically reinforcing bridges increase substantially relative to micelle-based hydrogels.",
"introduction": "Introduction Network frameworks in organisms are created through hierarchical self-assembly and defect-free interlinking of biopolymers such as actin, collagen, and mucin 1 – 5 . The primary function of a network is to give the organism its shape and mechanical resistance to deformation. In addition, biopolymers self-assemble into intricate network structures that exhibit differentiated functionality. For example, F-actin, a microfilament constituting the intracellular cytoskeleton, is a complex and dynamic network that supports biological processes such as cell migration and contractile dynamics 3 . Collagen, the most abundant microfiber in the extracellular matrix, gives a wide range of stiffness and elasticity, from soft skin to hard bone tissues 4 . Mucins assemble into branched networks and form mucus hydrogels, which are responsible for molecular transport and a physical barrier preventing infection 5 . Chemical and topological structures of polymers together dictate the properties of networks 6 , 7 . Polymer network topology indicates the extent to which the strands of polymers are connected. Given that the elasticity of polymer networks has topological origins, there is a growing interest in understanding and controlling topology from a molecular perspective in order to design networks with specific mechanical properties. Inspired by biological networks, attempts have been made to mimic the network structure in organisms using synthetic polymers 8 – 12 . However, these efforts have not been fully successful due to the emergence of topological defects in the network (e.g., inhomogeneity in node/strand density, dangling/unreacted strands and/or entanglements, and loop formation) 13 , 14 . Primary loops, the simplest cyclic topologies, form when the two ends of a strand reside in the same node. Since the loops do not impose direct physical crosslinks between nodes, they only weakly contribute to the elasticity of the network through the rare loop entanglement of interlocked polymers from independent nodes 15 . A common approach to designing hydrogel network forming block copolymers is to use an ABA triblock construct wherein the middle and end blocks are hydrophilic and charged blocks, respectively 16 – 19 . Mixtures of oppositely charged ABA triblock copolymers assemble through electrostatic interactions and can produce distinct micellar structures in which the coacervate nodes of the micelles are surrounded by loops formed by the hydrophilic middle blocks. At concentrations higher than the critical gelation concentration (CGC), some loops of the micelles can transform into bridges that connect the coacervate nodes, leading to the formation of a sparsely linked network (Fig. 1a ). Therefore, the key components affecting the mechanical properties of ABA triblock hydrogels are the density of bridges and nodes in the network rather than that of loops 15 , 20 . Previous observations have shown that the density of bridges increases with increasing relative length of the middle block and concentration of ABA triblock copolymers 21 , 22 . However, despite numerous efforts, there appears to be a limit to maximizing the fraction of bridges based on the loop-rich micelle structure, with the largest reported bridge fraction being only 63% in a hydrogel 22 . In addition, even when the loop formation was suppressed with the ABC triblock terpolymer through independent crosslinking of the A and C blocks, the mechanical properties of the hydrogel were not significantly improved 23 , 24 . These observations point to the need for a design strategy for strong hydrogels that focuses less on micellar loops and more on forming bridges. Hence, we aim here to design block polyelectrolytes that directly assemble to form a network of highly interlinked nodes with minimal formation of loops. In this design, it is the bridging process that imparts strong mechanical properties to the resulting network (Fig. 1b ). We start with a conventional mixture of triblock polyelectrolytes (P(APTC 150 - b -DMA 750 - b -APTC 150 ) and P(AMPS 150 - b -DMA 750 - b -AMPS 150 )), which will serve as a point of comparison for designed pentablock (P(DMA 250 - b -APTC 150 - b -DMA 250 - b -APTC 150 - b -DMA 250 ) and P(DMA 250 - b -AMPS 150 - b -DMA 250 - b -AMPS 150 - b -DMA 250 )) and nonablock polyelectrolytes (P(DMA 250 - b -APTC 150 - b -DMA 250 - b -APTC 150 - b -DMA 250 - b -APTC 150 - b -DMA 250 - b -APTC 150 - b -DMA 250 ) and P(DMA 250 - b -AMPS 150 - b -DMA 250 - b -AMPS 150 - b -DMA 250 - b -AMPS 150 - b -DMA 250 - b -AMPS 150 - b -DMA 250 )). Constructs of triblock and pentablock polyelectrolytes with identical molecular weight and charge density per chain were synthesized to compare equivalent polymer counts at the same weight percentage, as shown in Supplementary Fig. 1a , c (note that the length of every neutral block in pentablock polyelectrolytes is one-third of the middle neutral block of triblock polyelectrolytes). We aimed to increase the thermodynamic penalty of micellization in pentablock polyelectrolytes to create another self-assembly 25 . The design strategy was to increase the bending energy cost of loop formation by introducing short middle neutral blocks and disrupting the associativity of charged termini (that are responsible for loop closure) by replacing them with neutral blocks 26 . Consequently, the self-assembled pentablock polyelectrolytes exhibit a structural distinction, in which a branched network unit (netmer) with several densely interlinked coacervate nodes was formed instead of barely linked micelles. At concentrations above the CGC, the triblock polyelectrolytes formed a sparsely linked network of micelles (Fig. 1c ). Meanwhile, the pentablock polyelectrolytes formed a densely linked network by hierarchically stacking the netmers (Fig. 1c ). To further increase gelation efficiency, we designed the longer nonablock polyelectrolytes (composed of the same block species and block molecular weights present in the pentablock polyelectrolytes); the gelation efficiency was maximized due to the formation of larger netmers (Supplementary Fig. 1d , c ). The nonablock polyelectrolytes with longer chain lengths can promote the formation of a larger number of linked coacervate nodes per netmer relative to the pentablock polyelectrolytes. Due to the structural advantages of the more highly connected networks of the pentablock and nonablock polyelectrolytes, the storage modulus of the resulting hydrogels was substantially greater than that of the conventional triblock polyelectrolyte hydrogels. Fig. 1 Schematic representation of the network structures and gelation mechanisms. Two types of hydrogel networks. a In the traditional hydrogel network formed through the self-assembly of ABA block copolymers, the nodes are surrounded by loops. b In the proposed densely linked network, the coacervate nodes are closely linked by a large number of bridges. c Schematic of the hierarchical assembly mechanism from the unimer to the hydrogel network in three different polyelectrolytes. As the polymer concentration increased, the triblock polyelectrolytes formed micelles, leading to the formation of a tri-PEC (polyelectrolyte complex) hydrogel network with a sparsely linked structure. Pentablock polyelectrolytes and nonablock polyelectrolytes preferentially formed netmers with branched structures and produced the densely linked penta- and nona-PEC hydrogel networks, respectively.",
"discussion": "Results and discussion To achieve a nearly complete monomer conversion, all multi-block polyelectrolytes (Supplementary Fig. 1 ) were synthesized from acrylamide derivatives with high propagation rate coefficients ( k p ) through the one-pot aqueous reversible addition-fragmentation chain transfer (RAFT) polymerization 27 – 29 . A well-controlled sequence was successfully obtained for each block with a desired monomer conversion, degree of polymerization, and polydispersity index (PDI), as confirmed by 1 H nuclear magnetic resonance (NMR) and gel permeation chromatography (GPC) (Supplementary Figs. 2 – 4 ; details of the synthesis procedure are provided in the Methods section). In this study, four types of multiblock polyelectrolytes were compared as follows (Supplementary Fig. 1 ). First, triblock and pentablock polyelectrolytes with the same overall length of the polymer were compared (Supplementary Fig. 1a , c ). At this time, the length of the middle block of the triblock polyelectrolytes was three times longer than that of the pentablock polyelectrolytes, such that the triblock polyelectrolytes were referred to as tri(long middle, LM)block polyelectrolytes. In addition, we designed triblock polyelectrolytes with the same length as the middle block of the pentablock polyelectrolytes (this is referred to as tri(short middle, SM)block polyelectrolytes, (P(APTC 150 - b -DMA 250 - b -APTC 150 ) and P(AMPS 150 - b -DMA 250 - b -AMPS 150 ))) and determined the effect of the middle block length on the mechanical properties of polyelectrolyte complex hydrogels (Supplementary Fig. 1b ). Finally, we designed nonablock polyelectrolytes composed of the same blocks present in the pentablock polyelectrolytes but with increased blocks, and compared it with other multiblock polyelectrolytes (Supplementary Fig. 1d ). Self-assembled structures at concentrations below the CGC To determine the tri(LM)block, tri(SM)block, pentablock, and nonablock polyelectrolytes self-assembled complex structures (referred to as tri(LM)-, tri(SM)-, penta-, and nona-PEC (polyelectrolyte complex), respectively), cryo-TEM, SANS, and multiangle DLS experiments were performed for 0.2 wt% PEC solutions. First, 0.2 wt% PECs were visualized by cryo-TEM observations (Fig. 2a–d and Supplementary Fig. 5a–d ). All PECs formed spheroid coacervate nodes, penta- and nona-PEC had a small node diameter (23.85 ± 3.66 nm and 21.64 ± 3.75 nm, respectively), but tri(LM)- and tri(SM)-PEC had a large coacervate node diameter (31.14 ± 5.51 nm and 33.84 ± 6.39 nm, respectively) (Fig. 2e–h ). Noteworthy, although the ionic blocks of all polyelectrolytes have the same length, the tri(LM)- and tri(SM)-PEC node diameter is larger than that of the penta- and nona-PEC. It is considered that the sequence of each multiblock polyelectrolyte acts as an important factor in forming a node and determining its size. This trend is consistent with theoretical expectations, considering the steric constraints and potential conformations of the ionic blocks within the node structure (Fig. 2 l, m ) 26 . In particular, tri(LM)- and tri(SM)-PEC, which feature ionic blocks terminated within the polymer chain, promote the formation of larger nodes by accommodating a conformation with a radius corresponding to the ionic block length. In contrast, penta- and nona-PEC, which have an ionic block sandwiched between two neutral blocks, supposedly result in a more compact node structure due to the steric hindrance imparted by the neutral blocks. The structural analysis of the node of PECs was also confirmed through small-angle neutron scattering (SANS) experiments in 0.2 wt% PEC solutions, with node diameters of 27.4 ± 9.0 nm for tri(LM)-PEC, 31.8 ± 5.4 nm for tri(SM)-PEC, 22.8 ± 5.2 nm for penta-PEC and 18.2 ± 8.4 nm for nona-PEC (Fig. 2 i, j ). The coacervate node diameter of each PEC calculated by SANS coincided with trends in the cryo-TEM results. After that, we calculated the aggregation number of the polyelectrolytes to form the coacervate node (aggregation number, 270 for tri(LM)-PEC, 421 for tri(SM)-PEC, 155 for penta-PEC, 79 for nona-PEC) (Fig. 2k ). Requiring fewer polymers to form a node means that a higher node density can be achieved under the same polymer concentration conditions. Therefore, penta- and nona-PEC are expected to have significantly higher node density than tri(LM)- and tri(SM)-PEC. Fig. 2 Morphology analysis in PEC. Cryo-TEM images of the 0.2 wt% tri(LM)- ( a ), tri(SM)- ( b ), penta- ( c ), and nona-PEC ( d ). Histogram of core diameter distribution for tri(LM)- ( e ), tri(SM)- ( f ), penta- ( g ), and nona-PEC ( h ) calculated from cryo-TEM images. The total counts are 200, and the red line is a normal distribution. Core diameters are presented as mean values ± standard deviation. i SANS profiles of tri(LM)- (green open circles), tri(SM)- (olive green open circles), penta- (black open circles), and nona-PEC (blue open circles) with 0.2 wt% polymer concentration. All SANS profiles were fitted by the best model of the form factor of the core-shell sphere (red solid line). j Node diameter of four different PECs. The open rectangle represents cryo-TEM results, and the closed rectangle represents SANS results. Node diameters are presented as mean values ± standard deviation. k Calculated aggregation number of four different PECs. Schematic of the expected ionic block conformations in the coacervate node of tri(LM)- and tri(SM)-PEC ( l ) and penta- and nona-PEC ( m ). The gray circle represents the coacervate node, the solid line represents the neutral block, and the dotted line represents the ionic block. To rationalize and visualize the self-assembly of PECs, we performed coarse-grained molecular dynamics (CGMD) simulations of tri(LM)block, tri(SM)block, pentablock and nonablock polyelectrolytes in which the dense phase of the polymers is in coexistence with the dilute phase (Fig. 3a–c , Supplementary Fig. 6a and Supplementary Movies 1 – 4 ). This CG model of polyelectrolytes (see coarse-grained molecular dynamics simulation section in Supplementary Information) was developed to qualitatively assess—rather than quantitatively—the behavior of hydrogels. We employed the CG polymer model in implicit solvent to effectively explore the behavior of PECs on a computationally accessible macroscopic scale while preserving essential physical properties such as charge patterns and molecular weight ratio of polyelectrolytes. We carefully selected parameters such as the strength of electrostatic interactions, the dielectric constant, and the effective temperature to reflect the ambient conditions. As is evident from the representative CGMD snapshots (Fig. 3a–c and Supplementary Fig. 6a ), tri(LM)block and tri(SM)block polyelectrolytes mainly self-assemble into loop rich micelles. Most of the neutral blocks in tri(LM)-PEC form loops; however, on average 1-2 bridges can form between a pair of coacervate nodes (Fig. 3 d, e ), which can impose weak connectivity between micelles resulting in low node density (Fig. 3f ). In contrast, loop formation is mostly suppressed in penta-PEC for two reasons: (1) the larger bending energy cost of a shorter middle block closure and (2) the neutral ends can hamper micellization through hindering interactions with the charged blocks. Thus, in the absence of loop formation, pentablock polyelectrolytes can assemble into netmeric units in which the coacervate nodes are tightly linked through several bridges, resulting in high node density (Fig. 3b and 3d–f ). Interestingly, although the length of the neutral block of nonablock polyelectrolytes is as short as that of pentablock and tri(SM)block polyelectrolytes, the larger molecular weight of nonablock polyelectrolytes promotes a richer connection between the coacervate nodes resulting in the highest node density (Fig. 3c–f ). In addition to observing differences in the self-assembled form of PECs, we found a significant difference in the node population. In particular, the coacervate nodes of tri(LM)- and tri(SM)-PEC are larger and less numerous than penta- and nona-PEC, see the top right panels of Fig. 3a–c , top left magnified image of Supplementary Fig. 6a (the coacervate nodes are visualized with different colors, for the sake of clarity, the neutral blocks are omitted). Similar to the node size calculated by cryo-TEM and SANS described above, the simulation results also showed that the node size of tri(LM)-PEC was significantly larger than that of penta- and nona-PEC (Supplementary Fig. 7 ). Fig. 3 CG-MD simulation and size analysis in PEC. Coarse-grained molecular dynamics (CG-MD) simulations of self-assembled tri(LM)- ( a ), penta- ( b ), and nona-PEC ( c ) in coexistence with a dilute phase; the yellow, blue, and red monomers of polymers represent the neutral, negatively charged and positively charged species, respectively. Tri(LM)-PEC showed distinct micellar structures, and penta- and nona-PEC featured largely interlinked coacervate nodes. The coacervate nodes of the PEC in a pure dense phase are shown in the top right of panels, for the sake of clarity, the neutral blocks of the PEC are not shown in the coacervate nodes visualization. d The interlinks between a pair of nodes in three PECs. e The number of loops per node in three PECs. f The node density of three PECs in the dense phase. g Corresponding q 2 dependence of the mean decay rate (gamma, Γ ) for three PECs from multiangle DLS experiments. h Comparison of hydrodynamic diameter for three PECs. Schematic of the expected PEC cluster structure for penta- ( i ), nona- ( j ), and tri(LM)-PEC ( k ). CG-MD simulation data and hydrodynamic diameter are presented as mean values ± standard deviation. Next, a multiangle DLS experiment was conducted to compare the size of the PEC network. The multiangle DLS results revealed mean decay rate (gamma, \\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}$$\\bar{{\\Gamma }}$$\\end{document} Γ ¯ ) was proportional 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}$${q}^{2}$$\\end{document} q 2 in all 0.2 wt% PECs (Fig. 3g and Supplementary Fig. 6b ). Therefore, the mutual diffusion coefficient ( \\documentclass[12pt]{minimal}\n\t\t\t\t\\usepackage{amsmath}\n\t\t\t\t\\usepackage{wasysym} \n\t\t\t\t\\usepackage{amsfonts} \n\t\t\t\t\\usepackage{amssymb} \n\t\t\t\t\\usepackage{amsbsy}\n\t\t\t\t\\usepackage{mathrsfs}\n\t\t\t\t\\usepackage{upgreek}\n\t\t\t\t\\setlength{\\oddsidemargin}{-69pt}\n\t\t\t\t\\begin{document}$${D}_{m}$$\\end{document} D m ) of PEC was determined by linear regression 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}$$\\bar{{\\Gamma }}={D}_{m}{q}^{2}$$\\end{document} Γ ¯ = D m q 2 ( D tri(LM)-PEC = 47,181 nm 2 /s, D penta-PEC = 154,460 nm 2 /s, D nona-PEC = 85,874 nm 2 /s and D tri(SM)-PEC = 808,810 nm 2 /s). Afterward, the calculated diffusion coefficient was substituted into the Stokes-Einstein equation ( R h = k B T/6πη 0 D m , k B is the Boltzmann constant, T is the absolute temperature and η 0 is the solvent viscosity, R h is the hydrodynamic radius) to confirm the hydrodynamic diameter of the PEC network (Fig. 3h and Supplementary Fig. 6c ). In the case of tri(LM)-PEC, since the neutral block was made the longest despite a few bridges, connectivity between PEC was secured, creating the largest PEC network (6.96 ± 1.38 µm, Fig. 3 h, k ). When confirming the tri(LM)-PEC network through CG-MD simulations, a network of sparse and large micelles connected by a few bridges was visible (Supplementary Movie 5 ). On the other hand, unlike tri(LM)-PEC formed a micelle-based PEC network, penta-PEC formed netmers and its size was 2.13 ± 0.16 μm, which was significantly smaller than tri(LM)-PEC (Fig. 3 h, i ). This means that although penta-PEC is rich in bridges, it has poor connectivity due to short middle blocks and mainly consists of small netmers (Supplementary Movie 6 ). However, nona-PEC formed larger netmers (3.82 ± 0.21 μm, Fig. 3 h, j ) than penta-PEC because it had four ionic blocks within one polymer to ensure connectivity. Accordingly, the nona-PEC network showed a tight, densely connected netmer conformation in CG-MD simulations (Supplementary Movie 7 ). Meanwhile, in the case of tri(SM)-PEC, the smallest PEC network was shown due to the small number of bridges and short middle block (0.40 ± 0.05 μm, Supplementary Fig. 6c , d ). These results show differences in PEC network morphology depending on the polymer structure, which may affect gelation efficiency and mechanical properties at high polymer concentrations 6 , 7 . PEC hydrogels at concentrations above the CGC After analyzing the diluted solutions of tri(LM)-, tri(SM)-, penta-, and nona-PEC, we systematically investigated the mechanical properties of the PEC hydrogels. First, the CGC was confirmed by measuring the change in the zero-shear viscosity according to the concentration (Supplementary Fig. 8 ) 30 . In general, the higher the connection efficiency of the network, the lower the CGC. As expected, nona-PEC (0.8 wt%) showed the lowest CGC than other type of PEC. Interestingly, tri(LM)-PEC (1.0 wt%) showed lower CGC than penta-PEC (1.5 wt%). Penta-PEC requires a relatively higher polymer concentration to construct the network netmers; note that tri(LM)-PEC can benefit from the longer middle block to create interlinks at low polymer concentrations. In the case of tri(SM)-PEC, which had a small number of bridges with a short middle block, it showed the highest CGC (3.0 wt%) due to shortcomings in network connectivity. Next, rheological examinations were conducted to determine the viscoelastic properties of the studied hydrogels such as storage modulus ( G ´) and loss modulus ( G ´´), which represent the ability to store energy elastically and dissipate energy through heat, respectively. To investigate the effect of polymer concentration on the mechanical properties of the PEC hydrogels, strain and frequency sweep experiments were conducted for all four types of hydrogels at polymer concentrations ranging from 3.0 to 10.0 wt% (Fig. 4a–d , g, h and Supplementary Figs. 9 , 10 (Note that the tri(SM)-PEC experiments were conducted starting from a concentration of 4.4 wt%, measured stably)). Fig. 4 Rheological properties of the PEC hydrogels. Strain sweep was performed in a strain range from 1 to 1000% at a frequency of 3 rad s −1 . Storage ( G ′, closed rectangles) and loss ( G ″, open rectangles) moduli of the tri(LM)- (green), tri(SM)- (olive green), penta- (black), and nona-PEC (blue) hydrogels with 3.0 ( a ), 4.4 ( b ), 7.0 ( c ) and 10.0 wt% ( d ) polymer concentrations. Angular frequency dependencies of the complex viscosity of the hydrogels from the non-equilibrium CGMD simulations ( e ) and experiments ( f ). Rheological properties of the tri(LM)-, tri(SM)-, penta-, and nona-PEC hydrogels with various polymer concentrations. G´ obtained at a 1% strain and frequency of 3 rad s −1 and the dotted line represents ideal G´ calculated through the phantom network theory of rubber elasticity ( g ) and crossover strains of the G´ and G´´ determined from the strain amplitude sweep measurement ( h ). i Free-standing and stretchable performance of nona-PEC hydrogels with and without blue dye in the form of cubes. 20 g weighted cube-shaped 10.0 wt% tri(LM)- ( j ), penta- ( k ) and nona-PEC ( l ) hydrogels. m PEC hydrogels compared to synthetic and natural hydrogels. Gray closed labels represent synthetic block copolymer hydrogels, and gray open labels represent natural hydrogels, including chitosan, gelatin, lignin, and peptide. The results revealed that G ´ was substantially larger than G ´´ at low strain values, and their intersection was observed at higher strains due to the yielding of the gel. Furthermore, tan δ (defined as G ´´/ G ´ in the linear elastic regime) produced similar values between 0.15 and 0.48 at all polymer concentrations, which indicates that all PEC hydrogels behave more like elastic materials rather than viscous liquids (Supplementary Fig. 9 ). The G ´ values and crossover strains according to the polymer concentration of all hydrogels are summarized in Table 1 , Fig. 4g, h . For tri(SM)-PEC hydrogels, as can be expected from previous results, it showed the lowest mechanical properties at all polymer concentrations. Interestingly, G ´ of tri(LM)-PEC hydrogels was higher than that of penta-PEC hydrogels in the concentration range from 3.0 to 4.4 wt% (Fig. 4 a, b ), but the trend was reversed from 7.0 wt% (Fig. 4 c, d ). Thus, at low polymer concentrations, the penta-PEC netmers are less likely to form a hydrogel network; as shown in the CGC of penta-PEC was higher than tri(LM)-PEC. This suggests that at low polymer concentrations, where the gelation force is insufficient for penta-PEC hydrogel formation, longer nonablock polyelectrolytes can utilize their greater molecular weight to enhance spatial correlation among the nodes. This characteristic facilitates gelation through the formation of large netmers, where multiple coacervate nodes are interconnected. These nodes are linked by electrostatic interactions and bridged by neutral blocks, enabling the formation of several coacervate nodes per nonablock polyelectrolyte. Thus, the mechanical properties of nona-PEC hydrogel were pronounced at low polymer concentrations. For example, at 3.0 wt%, the G ´ of the nona-PEC hydrogel (476 Pa, 118 µM) was 3.6 times higher than that of the tri(LM)-PEC hydrogel (133 Pa, 214 µM) (Table 1 ). Table 1 Summary of the mechanical properties of the hydrogels Hydrogel Weight Concentration (wt%) Molar Concentration (µM) G´ at a strain of 1% a (Pa) Crossover strain (%) a Tri(LM)-PEC hydrogel 3 214 133 147 4.4 314 578 243 7 499 959 131 10 714 1924 211 Tri(SM)-PEC hydrogel 4.4 486 23 6 7 774 335 72 10 1105 1138 77 Penta-PEC hydrogel 3 214 26 30 4.4 314 416 366 7 499 1699 304 10 714 221 433 Nona-PEC hydrogel 3 118 476 319 4.4 173 1557 413 7 275 4065 306 10 392 22,173 614 a Determined by strain amplitude sweep at a fixed frequency of 3 rad s -1 . At high polymer concentrations, all the netmer clusters were well-connected, and thus, netmer bridges acted as elastic springs 15 , 20 . Indeed, at concentrations above 7.0 wt%, G ´ of penta-PEC hydrogel was larger than G ´ of tri(LM)-PEC hydrogel (Fig. 4 c, d ). In particular, at 10.0 wt%, the penta-PEC hydrogel showed G ´ that was about 4 times greater than that of the tri(LM)-PEC hydrogel. As seen in previous results, penta-PEC had structural advantages over tri(LM)-PEC, such as abundant bridges and high coacervate node density, which benefited the mechanical properties according to rubber elasticity theory 6 , 7 . However, these structural advantages were expressed in mechanical properties only if network connectivity was guaranteed 15 . Therefore, the structural advantages of the penta-PEC hydrogel began to emerge in the high polymer concentrations (in this case, from 7.0 wt%), where the polymer concentration complemented network connectivity. Note that nona-PEC hydrogel had guaranteed network connectivity and structural advantages even at low polymer concentrations so that at 10.0 wt%, the G ´ of the nona-PEC hydrogel (22,173 Pa, 392 µM) was 11.5 times higher than that of the tri(LM)-PEC hydrogel (1924 Pa, 714 µM) (Table 1 ). Next, the experimentally determined modulus of the hydrogel was compared with the predicted modulus for the 100% bridging network model. For this purpose, based on the phantom network theory of rubber elasticity, assuming that all polymers constituting the hydrogel exist as bridges, G’ of the hydrogel was calculated (see Fig. 4g and phantom network theory of rubber elasticity section in Supplementary Information). The experimental G’ of tri(LM)- and tri(SM)-PEC hydrogels was similar to or lower than the theoretical G’ at all polymer concentrations, which was considered to be a negative deviation caused by the loops present in the actual hydrogel. Interestingly, starting from 7.0 wt% polymer concentration, the penta-PEC hydrogel exhibited an experimental G’ much higher than the theoretical G’ . It can be expected that the abundance of entanglement between bridges existing between nodes caused positive deviation 31 – 33 . In particular, the nona-PEC hydrogel shows an experimental G’ greater than the theoretical G’ starting at 4.4 wt%, and the difference from the theoretical value becomes larger as the concentration increases. This is because the longest nonablock polyelectrolytes may have more dominant entanglement between adjacent bridges during the self-assembly process. To computationally assess the mechanical properties of the hydrogels, we performed non-equilibrium (NE) CGMD simulations of polyelectrolytes dense phase under an oscillatory shear deformation (see the Supplementary Fig. 11 and NE-CGMD simulations section in Supplementary Information) and characterized the angular frequency dependency of the complex viscosity (Fig. 4e ). The complex viscosities of the hydrogels from NE-CGMD simulations can predict the trend of the experimentally characterized complex viscosities (Fig. 4f ). To gain further insight into the structure of the PEC network, the SAXS profile of the hydrogel was investigated (Supplementary Fig. 12a–d ). From the SAXS results, as the polymer concentration was increased from 3.0 to 10.0 wt%, the domain spacing of the three hydrogels decreased; the domain spacing was calculated from the primary scattering peak (defined as d = 2 π / q ) (Supplementary Fig. 12e ). Over the entire concentration range, the penta-PEC hydrogels possessed a much shorter domain spacing, than that of the tri(LM)-PEC hydrogels due to the shorter neutral blocks between the ionic blocks (Supplementary Fig. 1a , c ). Since the tri(SM)block polyelectrolytes had the same neutral block length as the pentablock polyelectrolytes, it showed a domain spacing that was smaller than that of the tri(LM)-PEC hydrogel and similar to that of the penta-PEC hydrogel (Supplementary Fig. 1b ). In the case of tri(SM)-PEC hydrogel, domain spacing increased at 10.0 wt% polymer concentration, which means that as the polymer concentration increased, the network structure changed due to the rearrangement of the polymer chains 17 . Since tri(SM)-PEC hydrogel had a low molecular weight, the molar concentration at 10.0 wt% was higher than that of other PEC hydrogels, which seemed to accelerate the rearrangement of the polymer chains. Since nonablock polyelectrolytes had high polydispersity, the diversity of chain lengths and sizes caused the scattering vector of nona-PEC hydrogels to spread, resulting in broader peaks in the SAXS profile. Interestingly, at high polymer concentrations, nona-PEC hydrogel displays similar domain spacing to penta-PEC hydrogel due to the same length of neutral blocks (Supplementary Fig. 1c , d ). Afterward, diffraction peaks were confirmed at a polymer concentration of 10.0 wt% for all hydrogels (diffraction peaks, q 1 , √ 3 q 1 and √ 12 q 1 for tri(LM)-PEC hydrogel, q 1 and √ 7 q 1 for penta-PEC hydrogel, q 1 for nona-PEC hydrogel and q 1 , √ 4 q 1 and √ 12 q 1 for tri(SM)-PEC hydrogel). All PEC hydrogels were sufficiently dilute that they only exhibited a disordered structure, rather than ordered mesophases. Future research at higher polymer concentrations will likely be necessary to confirm the ordered structure. To further explore the network structure dependency of the PEC hydrogel on polymer concentration, two-dimensional correlation spectroscopy (2DCOS) was used by one-dimensional (1D) Raman spectra. By comparing the synchronous and asynchronous 2DCOS spectra, it was possible to confirm the sequential order of various chemical constituents and the transformation of the network structure according to the polymer concentration 34 – 37 . In the synchronous 2D Raman correlation spectrum, red peaks mean two correlated peaks change in the same direction (increase or decrease together depending on increasing concentration), whereas blue peaks mean two correlated peaks change in the opposite direction. According to Noda’s rule 35 , in the asynchronous 2D Raman correlation spectrum, if two correlated peaks have the same color at the same position as the synchronous spectrum, the peak on the x -axis changes before the peak on the y -axis according to the change in concentration. If the colors are different, the peak on the y -axis changes before the peak on the x-axis according to the change in concentration. Generally, in the 2DCOS analysis, bonds with a fast sequence are either a bond constituting the main chain of a polymer that undergoes a significant conformational change 34 , 38 , 39 or non-covalent interactions (electrostatic interactions or hydrogen bond) that significantly increase or decrease 40 – 43 . The representative Raman bands of our PEC hydrogels were observed at wavenumbers of 2940 (C–H), 1160 (C–N + ), 1050 (SO 3 − ), 930 cm −1 (C–C) (Supplementary Figs. 13 – 16 ). Upon comparison of the 2DCOS results of the tri(LM)-PEC hydrogels, it was found that the sequential order of the bands upon increasing the polymer concentration was as follows: C–N + < SO 3 − < C–H < C–C (Supplementary Fig. 13 and Supplementary Table 13 ). Furthermore, for tri(SM)-PEC hydrogels, the sequential order of the bands upon increasing the polymer concentration was as follows: C–N + < SO 3 − < C–C < C–H (Supplementary Fig. 14 and Supplementary Table 13 ). In other words, the hydrocarbon bonds that make up the main chain changed more rapidly than those composing the charged blocks, indicating considerable structural changes in the polymer chain. This result is consistent with the conformational change of the polymer chain in the tri(LM)- and tri(SM)-PEC hydrogels, where the micellar loops are transformed into expanded bridges as the polymer concentration increases (Fig. 1c ) 22 . On the other hand, for penta- and nona-PEC hydrogels, the sequential order of the bands upon increasing the polymer concentration was as follows: C–H < C–C < SO 3 − < C–N + (Supplementary Figs. 15 , 16 and Supplementary Table 13 ). In other words, the bonds that make up the charged blocks changed faster than those in the main chain. Therefore, as the polymer concentration increased, coacervation of the charged blocks occurred predominantly, connecting the netmers hierarchically (Fig. 1c ). Finally, as a simple functional test, we investigated the self-healing properties of the physical hydrogels. As shown in Supplementary Fig. 17a–d , after the strain-induced failure, all hydrogels immediately and completely restored both their G ´ and G ´´, indicating the rapid recovery of the entire network. In particular, the self-healing hydrogels by attaching different nona-PEC hydrogels with and without blue dye in the form of cubes were free-standing and showed good stretchable performance (25% strain) (Fig. 4i ). Other PEC hydrogels were also capable of self-healing, but due to their low mechanical properties, they broke easily upon deformation, unlike the nona-PEC hydrogel. In addition, when a weight of 20 g was placed on tri(LM)-, penta-, and nona-PEC cube-shaped 10.0 wt% hydrogels (having similar sizes and heights), there were distinct mechanical responses; tri(LM)-PEC hydrogel (Fig. 4j ) failed to withstand the weight and shattered, whereas penta- and nona-PEC hydrogels sustained the stress without failure (Fig. 4 k, l ). Afterward, we compared the G’ and crossover strain of our hydrogels with those of other synthetic and natural hydrogels with physical bonds (Fig. 4m ) 17 , 23 , 24 , 44 – 48 . Our hydrogel exhibited higher G’ and crossover strain than other hydrogels, suggesting that a simple method of controlling the sequence of the multiblock copolymer can improve the quality of the physical hydrogel by maximizing the fraction of elastically effective bridges in the network. We anticipate that our methodology holds significant promise for creating multiblock copolymer-based physical hydrogels with reduced structural defects. Herein, we demonstrated a simple strategy to produce rigid, water-rich hydrogels through the self-assembly of oppositely charged multiblock polyelectrolytes. Loop formation was inevitable when conventional ABA triblock polyelectrolytes self-assembled into core coacervate micelles, creating topological defects in subsequent network formation. Our strategy was to synthesize pentablock and nonablock polyelectrolytes by adjusting the sequence of linear polymers. As revealed by the CGMD simulations, the pentablock and nonablock polyelectrolytes inhibited loop formation; thereby, these polymers were directly self-assembled into highly branched netmers with several bridges formed between the coacervate nodes. As a result, the mechanical properties of the netmer-based hydrogels were greater than those of the conventional micelle-based hydrogels. Our CGMD simulations and experimental investigation show that linear polymers with appropriate charge sequence modulation can form highly interlinked and mechanically robust networks. Given that, in addition to electrostatic interactions, various physical interactions such as hydrophobic interactions, hydrogen bonding, and metal-ligand coordination can be incorporated into netmer compositions, we expect further development of robust and physical hydrogels in the future."
} | 9,678 |
29090283 | PMC5651552 | pmc | 5,654 | {
"abstract": "The microbially derived polyhydroxyalkanoates biopolymers could impact the global climate scenario by replacing the conventional non-degradable, petrochemical-based polymer. The biogenesis, characterization and properties of PHAs by Bacillus species using renewable substrates have been elaborated by many for their wide applications. On the other hand Bacillus species are advantageous over other bacteria due to their abundance even in extreme ecological conditions, higher growth rates even on cheap substrates, higher PHAs production ability, and the ease of extracting the PHAs. Bacillus species possess hydrolytic enzymes that can be exploited for economical PHAs production. This review summarizes the recent trends in both non-growth and growth associated PHAs production by Bacillus species which may provide direction leading to future research towards this growing quest for biodegradable plastics, one more critical step ahead towards sustainable development.",
"conclusion": "7 Conclusion In view of the recent advances in biopolymer research, primarily the PHAs have significant impact as a potential substitute of petro-chemical based plastics. The major challenge for the economical production of biopolymers (PHAs) depends on the selection of potential microbes by polyphasic approach and a cost-effective production approach. This suggests selection of suitable Bacillus species capable of efficient consumption and bioconversion of inexpensive substrates into a broad range of PHAs with diverse properties and applications. Among the various explored waste material, activated sludge seems to be the most promising for the Bacillus species. Combining the batch and fed-batch fermentations for enhanced productivity compared to the other methods available in the public domain can be another process intervention. Considering the controllable nature of chemostat, fed-batch fermentation seems to great potential to enhance productivities. All such efforts at the laboratory scale will need to be validated at pilot-scale for future industrial production and wide application of this biopolymer to tap the application potential of such bacterial species in general, and the genus Bacillus in particular.",
"introduction": "1 Introduction In developing countries several activities are transforming local problems into international issues in this global village. Plastics with favourable mechanical integrity and excellent durability have been one of the fall-outs of the rapid progress in material science technology. Having its utility in diverse sectors, plastics have became an essential part of the modern life. In the global commodity petrochemical based plastic production has grown two hundred fold from 1.5 million tons in 1950 to 299 million tons with an annual growth rate of 9% in 2013 [1] , [2] . These are typical petroleum-based, non-biodegradable polymers gather or aggregate around our ecosystem which is a far cry from few years back ecosystem [2] . Degradation of such solid wastes is a global concern. Even though it is difficult to completely ban the use of plastics due to their versatile utilities, it is possible to replace or reduce their use with alternative biodegradable polymers with similar properties. Among the entire bio-based and bio-degradable polymer, polyhydroxyalkanoates (PHAs) are well-known. These are bio-based and biodegradable without waste and also recycled to CO 2 and water. The endocellular PHAs are biosynthesized hydroxy-fatty-acids stored as lipid inclusions when carbon source is in abundance and nutrients like nitrogen, phosphorus, oxygen or sulphur are limited. These are secondary metabolites produced by various microbes in response to environmental stress. Such microorganisms can be located in diverse ecological niches like costal water body sediments, marine region, rhizospheric soil, water sediments and sludge [3] . These environments are often brimming over with organic nutrients and poor in other nutrients to support active growth and meet the metabolic requirements of the starving PHAs accumulating microbial population [4] . Extensive research provides a clear vision on several PHAs producers, that these microbes synthesize PHAs inclusions in the late log phase of growth cycle. Then, in later stage of their life cycle they use it as a carbonosoms [5] , [6] . Through metabolic activities, PHA is normally depolymerized to D-hydroxy-butyrate on demand, and then metabolized to acetoacetate and acetoacetyl-CoA [7] to provide energy to the cell. Though these carbonosoms accumulation has been investigated in various bacterial isolates, Bacillus species are extensively studied in PHAs world since the exploration of poly- β -hydroxybutyrate (PHB) in the cytosol of Bacillus megaterium by the French Lemoigne, in 1926 [8] . Some Bacillus species have been reported to produce as much as 90% (w/w) PHAs of dry cells during nutrients imbalance [9] . Bacillus species becoming model organisms in industry and academic world attributed primarily to its genetic stability [10] . Apart from higher growth rate compared to other bacteria, the use of Bacillus species to produce PHAs is advantageous over others due to the absence of lipopolysaccharides external layer in them which makes PHAs extraction much simple [11] . Bacillus species are also capable of producing PHAs copolymers utilizing the relatively simple, inexpensive and structurally unrelated carbon sources. Moreover, the isolates possess the ability to secrete number of hydrolytic enzymes that can be exploited for cost affordable PHAs production by utilizing, for instance, agro-industrial and other waste materials [12] . The major drawback of Bacillus species in PHAs production is their sporulating nature. Practically the fact of sporulation and deposition of PHAs granules provoked due to stress factors [13] . To overcome the predicament research on pilot scale PHB productions by B. cereus in the media that depresses sporulation, under acidic pH [14] and potassium deficiency [15] conditions. These pores over strategies not only inhibit spore formation in Bacillus but also can enhance the PHAs productivity. Several studies of PHAs are dealing with mostly on upstream and downstream process, its applications [16] , [17] and with genetic modifications or mutations to increase the yield [9] , [18] . Now these expertises become an impediment, being economically nonfeasible to market. This review summarizes these recent trends in PHAs production by Bacillus species as an effort to provide direction and leads to future research and development towards the growing quest for biodegradable plastics, one more critical step ahead towards an eco-sustainable development."
} | 1,686 |
27381069 | PMC5052336 | pmc | 5,655 | {
"abstract": "Spore counts, species composition and richness of arbuscular mycorrhizal fungi, and soil glomalin contents were evaluated in a soil contaminated with Zn, Cu, Cd and Pb after rehabilitation by partial replacement of the contaminated soil with non-contaminated soil, and by Eucalyptus camaldulensis planting with and without Brachiaria decumbens sowing. These rehabilitation procedures were compared with soils from contaminated non-rehabilitated area and non-contaminated adjacent soils. Arbuscular mycorrhizal fungi communities attributes were assessed by direct field sampling, trap culture technique, and by glomalin contents estimate. Arbuscular mycorrhizal fungi was markedly favored by rehabilitation, and a total of 15 arbuscular mycorrhizal fungi morphotypes were detected in the studied area. Species from the Glomus and Acaulospora genera were the most common mycorrhizal fungi. Number of spores was increased by as much as 300-fold, and species richness almost doubled in areas rehabilitated by planting Eucalyptus in rows and sowing B. decumbens in inter-rows. Contents of heavy metals in the soil were negatively correlated with both species richness and glomalin contents. Introduction of B. decumbens together with Eucalyptus causes enrichment of arbuscular mycorrhizal fungi species and a more balanced community of arbuscular mycorrhizal fungi spores in contaminated soil.",
"conclusion": "Conclusion The concentration of heavy metals in the soil of the areas under rehabilitation affected the diversity and the composition of AMF communities. The rehabilitation system 2, which included planting of Eucalyptus in rows and Brachiaria in the inter-rows, presented the most favorable conditions for the development and occurrence of AMF, with potential application in rehabilitation program of contaminated areas. Although at a lesser degree, positive effects were also observed in system 1. It was evident that Glomus sp. is adapted to the stress caused by different concentrations of heavy metals, since it was detected in all the sites contaminated with Zn, Cu, Cd, and Pb. Diversity of AMF and glomalin contents can be considered good indicators of rehabilitation of soils contaminated with Zn, Cu, Pb and Cd.",
"introduction": "Introduction Arbuscular mycorrhizal fungi (AMF – Phylum Glomeromycota) are important components of the soil microbiota that contribute to the diversification and stability of natural ecosystems. 1 AMF are obligate mutualistic symbionts that colonize the roots of more than 80% of plant families, and which establish an association known as arbuscular mycorrhiza. 2 Several studies have reported on the abundance and occurrence of AMF in contaminated soils, 3 , 4 , 5 , 6 and have highlighted the high resistance of AMF isolates to several types of stress, including water stress, soil acidification, disaggregation, and absence or scarcity of vegetation cover. Mining activities produce wastes which may contain heavy metals as contaminants. These residues are generally deposited on the ground, and often occupy large extensions. In order to prevent chemical elements contained in these residues to be exposed to leaching processes, or to the action of the winds, rehabilitation programs are usually employed in these areas through the establishment of vegetation cover. However, plants often have very limited development in contaminated areas. Due to their beneficial effects on plant growth under stressed conditions, AMF have been used to enhance rehabilitation of contaminated soils through phytoremediation. 7 Besides, mycorrhizal plants and their associated microbiota can regulate absorption, transformation and removal of soil pollutants. 8 The contribution of AMF to plant growth in soils contaminated with heavy metals is related to several factors, including diversity, abundance, persistence, and efficiency of AMF populations, which might vary between different locations, and are related to the environmental variables and to the presence of vegetation. 9 The insoluble glycoprotein glomalin, which is abundantly produced by the hyphae of some AMF, is known for accumulating in the soil where it can retain large amounts of heavy metals. 10 It has been suggested that soil glomalin content is related to changes associated with land use and rehabilitation of degraded soils. 11 These changes in vegetation cover may directly influence AMF communities and the quantity and quality of compounds produced by them, such as glomalin. 4 Therefore, abundance and richness of AMF or glomalin content may be useful indicators of rehabilitation. Studies showed that these attributes of AMF are inversely related to concentration of heavy metals in the soil. 3 Therefore, improved knowledge on the dynamics of these fungi in contaminated rehabilitated soils is highly relevant, considering that AMF may contribute to revegetation and phytostabilization. 12 The aim of this study was to evaluate the occurrence and AMF community composition in soils from areas contaminated with zinc (Zn), cadmium (Cd), copper (Cu), and lead (Pb), and subject to rehabilitation.",
"discussion": "Discussion Replacement of the contaminated soil with tree plantation in rows with or without B. decumbens sowing in the inter-rows allowed successful rehabilitation of the soil. These rehabilitation programs promote reduction of total and soluble contaminant metals (Zn, Cu, Pb, and Cd) in the soil. Compared to the concentrations found at the beginning of the rehabilitation process, these metals were reduced by 90%, and allowed the establishment and the growth of the introduced plants. Decrease in metal contents varied according to the element, and were higher in rows than in inter-rows due to the amount of soil replaced and to the growth of new seedlings. However, Zn and Cd concentrations exceeded a bit the limits set by COPAM for the agricultural use of soil (450 and 38 mg kg −1 of dry soil of Zn and Cd, respectively). 15 In turn, NR site presented the highest concentration of heavy metals. Thus, it demonstrated that rehabilitation by soil excavation and replacement with non-contaminated soil was effective to create an adequate environment for plant initial growth. As shown in the Fig. 1 , plants were able to develop well in the contaminated area, and consequently showed continuous growth which promoted AMF multiplication. In fact, it was found a total a 15 different AMF morphotypes in the area, and most of these species had been previously reported in studies of AMF occurrence and ecology in these sites before rehabilitation. 3 However, differently from these authors, in rehabilitated area, it was found Acaulospora longula and Rhizophagus diaphanus , which indicates that rehabilitation allowed sporulation of other species unable to sporulate in the non-rehabilitated soils. Disregarding the control sites (P and C), 15 AMF species were found in the present study. This number was similar to, or higher than the ones reported by other studies carried out in heavy metals contaminated soils. Study carried in India found only four AMF species in a soil contaminated with heavy metals (Cu, Zn, Cr, Cd and Pb) collected from a tannery treatment plant. 23 In other studies carried in bauxite mining areas in Brazil, and in sewage sludge deposition areas that were contaminated with Zn, Cd, Cu, Ni and Pb in Germany, researchers found six AMF species. 4 , 24 Fifteen species were detected in soils contaminated with Zn, Cu, Pb, Ni, and Cd in India. 25 The present results showed that revegetation with Eucalyptus and Brachiaria associated with replacement of contaminated soil by uncontaminated soil may contributed to a continuous revegetation process of the area. It was found predominance of species from the Glomus and Acaulospora genera, which confirms that they are better adapted to stressful conditions of soil contaminated with heavy metals. 26 In fact, Glomus sp. was a generalist species since it was the only one found in all the sites. The species P. occultum was also considered a generalist, since it was detected by the trap culture technique in most plots. As pointed out by other authors, 27 , 28 , 29 \n Acaulospora and Glomus genera are well adapted to excess of heavy metals, and few reports reveal the occurrence of P. occultum in contaminated soils, as observed in the present study. However, Melloni, Siqueira and Moreira 4 reported the occurrence of P. occultum in mining soil under rehabilitation. Rhizophagus intraradices recorded in this study had only been previously recorded in soils contaminated with Zn, Cu, Cd and Pb in bauxite mining areas in the north of Brazil and in gypsum mining areas. 3 , 28 , 30 , 31 Among the AMF species detected in direct field samples, Acaulospora sp. and C. pelucida were not found in the trap culture pots. In turn, the species A. foveata , A. colombiana , and Glomus sp. (sporocarp) were exclusively detected by this method. Except for A. colombiana , which was found only in the trap culture from Cerrado, it is suggested that the change in the AMF community structure may be related to alterations in the stress levels. In this study, stress levels were higher in the field (higher heavy metal concentrations), when compared to trap cultures (lower heavy metal concentrations promoted by diluting the original soil in the preparation of traps), which is supported by Rydlová and Vosátka. 32 These authors emphasized that AMF may lose their adaptation and stress tolerance when sub-cultured without initial stress. These results highlight the importance of detecting AMF diversity by multiplying them in trap culture pots, in addition to the survey of spores directly collected from the field, in order to promote access to a larger number of species in studies on diversity and to afford a more complete picture of the biology, ecology, and diversity of AMF communities. 33 Nevertheless, several authors suggest that the trap culture method might select for species that easily sporulate, and might omit the ones that colonize roots, although they do not sporulate. 34 After the evaluation of the direct field samples, it was found greater AMF species richness in EcB-IR, whereas for the trap cultures, the greatest species richness was detected in Ec-R. Nevertheless, both results show AMF as an indicator of soil rehabilitation, since the AMF species richness found in the rehabilitated sites was higher, when compared to NR site, where the concentrations of heavy metals were higher. The presence of only one Glomus sp. in Cerrado field samples may be related to the absence of, or to the low sporulation of other AMF species at sampling time. However, by using trapping, it was possible to recover two other species ( A. colombiana and P. occultum ). It should also be considered that sampling occurred in the rainy season, and at this condition, low amount of spores is usually found. 35 , 36 On the other hand, AMF total diversity found in Cerrado was lower, when compared to rehabilitation systems and pasture, possibly because Cerrado is more stable than the other environments regarding the variation in soil characteristics. These conditions could act as selection pressure on fungal communities with reduced sporulation, or produce spores with low capacity to resist adverse conditions. 37 , 38 This is consistent with findings of low spore density in climax ecosystems. 39 Thus, although several studies have demonstrated high diversity of AMF in Cerrado, edaphic conditions and vegetation cover might also promote changes in spore density. 4 In a study on soils from a bauxite mining area under rehabilitation, it was found that mining affected spore density, which increased after the beginning of rehabilitation. 4 The lower spore density in the C and P, might be attributed to the natural stability of these ecosystems, especially due to the constant presence of hosts, and to the absence of abrupt variations in soil fertility. These characteristics might ensure the survival of fungal species with low natural sporulation ability, or which produce spores with poor resistance to adverse conditions. 4 , 40 Similar effects might also be found in areas subjected to agricultural cultivation, such as pasture, which are influenced by soil handling techniques, such as tilling and fertilization, extensive monoculture, and the use of agricultural toxins. According to some authors these handling techniques might reduce the incidence of some AMF and sporulation. 40 The number of spores counted in the direct field samples was lower, when compared to the count in the trap culture pots in all the investigated sites. It has been reported that the number of spores in the direct field samples is usually low, and they are parasitized. 41 Besides that, usually all of the subcellular structures needed for accurate identification of the species are not intact. Therefore, AMF multiplication system in host plants under controlled conditions is used. In the present study, the use of the trap culture technique allowed the observation that AMF species from sites highly contaminated with heavy metals, such as NR, presented high sporulation capacity under less contaminated conditions, as well as in the presence of a mycotrophic host plant, B. decumbens , which was used as a bait plant in the trap culture pots. In addition, the greater sporulation found in EcB-IR might be associated with the presence of B. decumbens in the inter-row areas, which probably favored sporulation. Studies that evaluated the effect of B. decumbens in revegetation of degraded soils, showed 400% increase in soil spores, which confirmed that this grass species is effective in multiplying AMF. 42 Soil restoration system 2 ( Table 1 ) provided better balance between AMF populations, since no predominance of any AMF species was observed ( Fig. 3 ). This result suggests that the soil cultivated with Eucalyptus and Brachiaria in the inter-rows led to the best soil conditions, where microhabitats protected better the soil against sudden disruptions, and showed less competition for nutrients and niches among microorganisms, ensuring low sporulation. 28 In the present study, BRSP was related to soil rehabilitation, since NR, where the soil presented lower vegetation cover and higher concentrations of heavy metals, presented lower EE-BRSP and BRSP contents, as suggested for degraded areas in general. 11 In conditions with high concentrations of heavy metals, AMF usually presents protective mechanisms, such as greater production and renewal of external mycelium, which contributes to glomalin synthesis. 43 , 44 However, it is worth noting that regardless of the rehabilitation program (systems 1 and 2), EE-BRSP and BRSP values found in the contaminated areas were higher than, or similar to those found by other authors in areas with other type of contaminants. Studies showed values of EE-BRSP that varied from 0.4 to 0.07 mg g soil −1 , together with an increase in the level of contamination with Cr. 45 The high levels of glomalin in Cerrado soil were probably represented by old stocks of this glycoprotein, since its stability and persistence in preserved environments can reach 42 years in tropical soils. 46 However, none of the treatments reached BRSP values similar to the 60 mg g soil −1 found in forest soils of temperate areas. 46"
} | 3,856 |
37333661 | PMC10272403 | pmc | 5,656 | {
"abstract": "Microbial communities can be considered complex adaptive systems. Understanding how these systems arise from different components and how the dynamics of microbial interactions allow for species coexistence are fundamental questions in ecology. To address these questions, we built a three-species synthetic community, called BARS (Bacillota A + S + R). Each species in this community exhibits one of three ecological roles: Antagonistic, Sensitive, or Resistant, assigned in the context of a sediment community. We show that the BARS community reproduces features of complex communities and exhibits higher-order interaction (HOI) dynamics. In paired interactions, the majority of the S species ( Sutcliffiella horikoshii 20a) population dies within 5 min when paired with the A species ( Bacillus pumilus 145). However, an emergent property appears upon adding the third interactor, as antagonism of species A over S is not observed in the presence of the R species ( Bacillus cereus 111). For the paired interaction, within the first 5 min, the surviving population of the S species acquires tolerance to species A, and species A ceases antagonism. This qualitative change reflects endogenous dynamics leading to the expression for tolerance to an antagonistic substance. The stability reached in the triple interaction exhibits a nonlinear response, highly sensitive to the density of the R species. In summary, our HOI model allows the study of the assembly dynamics of a three-species community and evaluating the immediate outcome within a 30 min frame. The BARS has features of a complex system where the paired interactions do not predict the community dynamics. The model is amenable to mechanistic dissection and to modeling how the parts integrate to achieve collective properties.",
"conclusion": "5. Conclusion The BARS three-species synthetic community serves as a valuable higher-order model, providing clues to important ecological questions. For instance, it helps to understand how complex systems arise from different components and the role of microbial interaction dynamics in promoting species coexistence. At higher-order levels, the bars model demonstrates the emergence of coexistence between sensitive and antagonistic species, where antagonistic competition is no longer observed. Of great significance is that the BARS model allows for evaluating the immediate response of bacteria in antagonistic competition. This contrasts to most studies, which evaluate community assembly over hours or days, leaving the initial encounter between community members as a black box; we miss the game’s climax and are left only with a picture of the winners. The BARS synthetic community model simplifies the study of dynamic interactions but maintains emergent features of ecological communities.",
"introduction": "1. Introduction Microbial communities form the basis of all biological systems and are dynamic and complex. The complexity of microbial communities results, among other things, from the great diversity of interacting species. The stability in highly diverse communities is proposed to be the result of systems of more than two competitors forming a network of competitive relationships, the structure of which should influence the dynamics of the system as a whole ( Bissett et al., 2013 ; Konopka et al., 2015 ). The stability of communities is considered an emergent property. Extensive effort has been dedicated to understanding the processes that contribute to community assembly (Reviewed in Nemergut et al., 2013 ). Synthetic ecology offers an approach to studying communities, reducing the number of variables, particularly the number of interacting species, the environment, and the spatial setting. One critical aspect, though often underappreciated, is that synthetic ecology allows controlling the timing of the experiments. Competitive interactions are a defining characteristic of microbial communities, and competition can be indirect, through competition for nutrients (exploitative competition) or direct, when one individual harms another (interference competition; Hibbing et al., 2010 ; Cornforth and Foster, 2013 ). Individuals possess unique genetic repertoires that allow them to respond rapidly to environmental changes and perceived threats from other community members. For instance, they can synthesize siderophores to compete for nutrients, such as iron ( Wandersman and Delepelaire, 2004 ). For direct competition, bacteria can produce antibiotics, toxins, and surfactants ( Yim et al., 2006 ; Cornforth and Foster, 2013 ; Niehus et al., 2021 ). Antimicrobial compounds mediate interactions between different species, related strains of the same species, or genetically identical individuals in a population ( Kerr et al., 2002 ; Abrudan et al., 2015 ; Stubbendieck and Straight, 2015 ; Lyons and Kolter, 2017 ), and some have evolved strategies for delivering antibiotics and toxins in a cell-to-cell contact fashion, such as type VI secretion (T6S) and contact-dependent growth inhibition (CDI) ( Garcia, 2018 ). Understanding the diverse adaptation strategies of microbial communities to compete and survive is essential for comprehending the functioning and evolution of microbial ecosystems. A common approach to studying interactions among microbes is through the analysis of pairwise interactions. This strategy was, after all, the source of discovery of natural products encoded by bacteria and fungi that have been the sources of many antibiotics in current clinical use. Pairwise interactions are commonly studied to understand microbe interactions, but modeling multispecies communities from paired interactions has limitations ( Carrara et al., 2015 ). Friedman et al. (2017) suggest that species that coexist in pairs will survive, and those excluded by any surviving species will go extinct in a multispecies setting. However, in higher-order interactions, the interactions between a group of species depend on the presence and interactions of other species in the community. The presence of additional species can modulate interactions, making predictions from paired interactions complicated ( Billick and Case, 1994 ). Simplified microbial models have been widely employed to study community coexistence and assembly. One example is the non-transitive rock-paper-scissors model developed by Kerr et al. (2002) using three isogenic E. coli strains (toxin-producer, toxin-sensitive, and toxin-resistant strain). This model demonstrated that coexistence can occur in spatially structured environments, but not in mixed environments where microbes are continuously dispersed and interact directly and indirectly through diffusible molecules, leading to unstable coexisting communities ( Kerr et al., 2002 ). Kerr’s model uses a single species and isogenic laboratory strains with a single trait altered, which is useful for studying specific mechanisms. Another synthetic community is the “THOR” model by Lozano et al. (2019) , which explores paired interactions with different species that have an antagonistic effect on colony expansion and biofilm formation. However, Lozano’s model evaluated the antagonistic interaction only after 10 h and up to 2 days, leaving a gap in information regarding the early dynamics of the interaction. Despite being composed of only three organisms, the models discussed above exhibit emergent properties that are not observed in paired interactions as a result of what is described as Higher Order Interactions (HOIs). In such interactions, the effect of one competitor on another depends on the presence of a third species. These interactions arise with increased complexity in bacterial communities, and the presence of at least three organisms is required for HOIs to emerge. The term “higher-order interactions” has been defined in various ways in the ecological literature ( Billick and Case, 1994 ). One use of the term HOI implies that indirect interactions arise, for instance, when the impact of one species on another requires or is modified by the presence of a third species ( Wootton, 1994 ). HOIs have to be considered an intrinsic property of complex microbial communities. In describing collective behaviors, emergence refers to how collective properties arise from the properties of parts, how behavior at a larger scale arises from the detailed structure, and relationships at a finer scale. Emergent behaviors have been observed from large ensembles of elementary agents such as ant colonies and bird flocks. In ecology, emergence seeks to explain how biodiversity is maintained. The main question being to what extent the interaction between any two species is influenced by other species in the system. Understanding higher-order interactions (HOIs) is crucial for comprehending the dynamics of communities and what allows species coexistence and makes them robust to the interactions between them and the environment ( Grilli et al., 2017 ). In a higher-order interaction, one competitor modifies the competition between the other two ( Levine et al., 2017 ). Although HOIs have been observed in bacterial communities, they were not explicitly described in some cases ( Gallardo-Navarro and Santillán, 2016 ; Lozano et al., 2019 ; Piccardi et al., 2019 ). Nevertheless, there is a growing body of research on HOI dynamics in microbial communities ( Mickalide and Kuehn, 2019 ; Sanchez-Gorostiaga et al., 2019 ). In microbial communities, response to competition can be affected by various factors, such as growth conditions, life-history and functional traits, and interactions. However, communities are frequently disturbed by various factors, such as dispersion and invasion, which introduce new neighbors and require that their members be ready to respond to potential threats from competing organisms. The time of a response is an essential factor in competition, and a delay in responding, to antagonism, for instance, could lead to the death of a population. Nevertheless, immediate responses in competitive interactions have been the subject of limited research, as most laboratory studies on microbial interactions are conducted over periods of hours ( Kerr et al., 2002 ; Kirkup and Riley, 2004 ; Vaz Jauri and Kinkel, 2014 ; Lozano et al., 2019 ), and in other cases, after days of interactions ( Gallardo-Navarro and Santillán, 2016 ; Mickalide and Kuehn, 2019 ; Piccardi et al., 2019 ). Considering that the generation time of most aerobic, free-living bacteria is 20 to 60 min (albeit under laboratory conditions), the data from these experiments represent a late picture of an interaction that began within minutes of the bacteria’ first encounter. In this context, there is a need for models that allow us to study the immediate response to competition in bacterial communities. In previous work, after studying a transitive interaction network of 78 bacterial species, a repeated pattern of Resistant (R), Antagonist (A), and Sensitive (S) species was observed that reflected a history of interactions in communities ( Pérez-Gutiérrez et al., 2013 ). In the referred work, competitive interactions were evaluated in pairs, by recording halos of growth inhibition formed around a drop of a given species culture on a “mat” formed by another species extended on semi-solid medium. The result of the evaluation of 6,084 interactions, was a transitive network that resembled a food web. The interactions revealed three different ecological properties: “Antagonists,” “Sensitive” (sensitivity to the antagonists), as well as some non-connected “Resistant” species. A simulation with a cell automata model revealed that despite the antagonism no extinction of sensitive species occurred in a “structured environment” ( Zapién-Campos et al., 2015 ). Based on the above information where three interaction properties enable a complex natural microbial community, the hypothesis in this work is that a synthetic microbial community can be created using three members, each representing a different type of interaction (A, R, S). The simplicity of this community will allow the description of the assembly process and the exploration of the mechanisms underlying the paired and higher-order interactions. We anticipate that this higher-order community will exhibit emergent properties that cannot be observed in paired interactions. Specifically, we expect that the community will allow for the coexistence of sensitive and antagonistic strains, and that the sensitive strains will respond rapidly response to antagonistic interactions in order to survive. Additionally, we anticipate that the role of the resistant strain will only become apparent in the interaction of all three strains. In order to fully understand the dynamics of these interactions, we propose that evaluations must be conducted in minutes, even before the interacting bacteria have divided. This is a new approach that has not been attempted in previous works and will allow us to capture the early stages of community assembly. Through this work, we hope to shed light on the mechanisms that underlie the interactions between the different strains in the community, and how they contribute to the community dynamics. To answer these questions, we constructed and evaluated a three-species synthetic community. Consisting of three Bacillota species each with a different characteristic: antagonistic (A), resistant to antagonism (R), and sensitive to antagonism (S). Each species had a distinct colony morphology that allowed simple quantification of the viability of each species in the interactions. The BARS community allowed us to observe emergent properties within minutes of the interaction, making it a unique model for exploring the immediate response to antagonism. The emergent property observed in BARS is the coexistence of the three different species, despite the antagonism observed in a paired interaction. BARS dynamics reproduce those of ecological communities: The effect of a strong competitor on the activity of its interactor, habitat modification, and changes in the abundance of the sensitive species. The BARS synthetic community model, with a simple experimental approach, has properties of a complex system with HOI dynamics.",
"discussion": "4. Discussion In this work, we present a three-species synthetic community that we named BARS, consisting of three Bacillota species characterized as the antagonist (A), resistant to antagonism (R), and sensitive to antagonism (S). The community exhibits endogenous dynamics that provide clues as to how complex systems arise from the interaction of a few components allowing species to coexist in antagonistic interaction. The emergent property observed in BARS is the coexistence of the three different species, despite the antagonism observed in a paired interaction. Microbial communities exhibit complex adaptive system behavior, and the BARS model, exemplifies an HOI community. It displays emergent states and suggest non-linear dynamics. The BARS model achieved a stable, coordinated system immediately after starting a triple interaction assay. The system’s emergent stability was achieved by changing the “rules” of antagonism to a non-antagonistic mode, resulting in endogenous and higher-order properties that were not predictable from either the examination of individual members or the paired interactions. The essential feature of the BARS model is its rapid response to interaction, occurring within 30 min. In this model, during a paired interaction, the sensitive species population survived antagonism and became tolerant and the A species cease to antagonize. In the triple interaction the R species promoted the survival of the sensitive species. The BARS three-species model, summarized in Figure 5 , captures the properties of complex communities with highly interactive organisms that engage in metabolic coupling, promote endogenous dynamics, and form self-organizing structures. The results suggest mechanisms of negative interference competition mediated by antibiotics and antibiotic resistance genes. We show that the survival of the sensitive strain was due to the induction of tolerant cells during interaction, and that antagonism was caused by diffusible antimicrobial compounds. We also observed that the detoxification of the medium and degradation of antimicrobial substances, and that the antagonism and protection offered by the resistant strain were density-dependent. Figure 5 The three-species BARS synthetic community: a model of a complex system with emergent properties. The BARS synthetic community serves as a model of a complex system with emergent properties. (A) Each species exhibits known properties from paired interactions, but their behavior in a triple community cannot be predicted. (B) As a higher-order interaction system, it adopts features of complex communities, with the resistant species acting as a mediator (blue shield symbol). During paired interactions, a strong antagonism occurs in the first 5 min, but no further antagonism is observed in the next 20 min (the curve decreases and then remains stable). Notably, after 30 min, a qualitative change is observed, as the A species no longer antagonizes S cells, and S cells become tolerant to A. The top line in the figure shows that a stable three-species community emerges and antagonism is not observed, suggesting that an immediate dynamic takes place. The three BARS species were selected based on specific criteria, including their distinct ecological “roles” assigned in a previous work, where paired interactions in a 78 × 78 species matrix were evaluated ( Pérez-Gutiérrez et al., 2013 ). The species are part of a collection from five sampled communities for which antagonism interaction networks within each community have been described. Within each of the five networks, less antagonism was observed than in the combined network. This suggests that there may have been a history of community interactions, in which species that were able to tolerate one another were enriched within each community. As such, it is possible that these bacteria have acquired particular antibiotic production genes and that an “arms race” has been playing out, by which these bacteria have evolved multiple forms of antibiotic resistance. The genetic repertoire for exerting antagonism, inducing resistance, and detoxification capacity may have been acquired and fine-tuned through their evolution in their sediment communities. Our own research has shown intraspecies variation in Bacillus strains co-occurring in the sediment communities of the Churince system in Cuatrocienegas, Mexico. We have shown that these strains exhibit high levels of intraspecific phenotypic and genotypic diversity ( Gómez-Lunar et al., 2016 ; Rodríguez-Torres et al., 2017 ), indicating a microevolutionary process that optimizes their ability to respond to environmental changes and microbial interaction challenges. Communities experience invasion continuously, or their members get shuffled around to find themselves with new neighbors. They have evolved adaptations to respond immediately to biotic and abiotic challenges in complex communities, making them experts in integrating fine-grained information and collectively finding solutions that we observe as emergent phenomena. Their capacity to respond immediately was probably shaped by the history of past environmental states. As Flack (2017) notes, biological systems make hypothesis about the present and future environments they or their offspring will encounter based on the history of past environmental states they or their ancestors have experienced. The importance of BARS and other models in synthetic ecology is that they allow us to address questions regarding the assembly of communities, a process where biotic and abiotic filters define the diversity and abundance of species. Given the complexity of this endeavor, having models that reduce the number of individuals and environmental variables is essential. However, equally important and missing in most microbial models, is being able to address the immediate consequence of the invasion of a community by new individuals. Different statistical and information theory methods can be used to detect causal relationships. Herrera Paredes et al. (2018) investigated the predictive performance of various statistical models. They demonstrated that it is possible to infer causality between microbiome composition and host phenotypes in complex systems. We speculate that within the complex systems, the BARS model behavior may represent a case of “explainable” weak emergence. Weak emergence describes the way in which complex systems exhibit new properties or behaviors that are not present in their individual parts but can be explained in terms of the interactions between the parts. To test this hypothesis, experiments could be conducted to determine whether the emergent features in the triple interaction are simply an extension of the mechanisms already observed in the paired, lower-level interaction. This investigation would require an examination of the molecular mechanisms of antagonism and resistance. Various strategies for antagonism have been reported, including the secretion of enzymes or metabolites into the extracellular space or through outer membrane vesicles (OMVs; Manning and Kuehn, 2011 ; Kulkarni et al., 2015 ; Liao et al., 2015 ; Lekmeechai et al., 2018 ; Roszkowiak et al., 2019 ). Toxin production is a common response to competitive interactions, but factors such as species frequency, nutrient level, relatedness, and the cost of toxin production all influence bacterial decision-making ( Niehus et al., 2021 ). Another property of the dynamics of the BARS community is the induced tolerance of a fraction of the sensitive species population to antagonism. Our results suggest that under the conditions evaluated, the observed antagonism occurred during the interaction and there were no preexisting tolerant cells, but rather the cells that survived are those that induced resistance to antagonism. This resembles antibiotic resistance induction that has been extensively studied in bacteria, with some examples specifically in Bacillus ( Chancey et al., 2012 ). The mechanism of defense can include enzymes that degrade antibiotics or toxins. The detoxification of the environment by one of the interactors in a community has been described. The idea of a third interactor as a mediator of antagonism was reported by Kelsic et al. (2015) . They proposed that the antibiotic-resistant species in a community have the function of degrading the toxins and antibiotics in the environment, allowing a sensitive species to cohabit in the community with the antibiotic producer species. In studies on antibiotic resistance, detoxifier capacities have been attributed to species able to express enzymes that can degrade antibiotics, such as β-lactamases. In a gram-negative model, the presence of β-lactamases was reported as a modulator of an antagonistic interaction by the cleavage and deactivation of antibiotics ( Medaney et al., 2015 ; Frost et al., 2018 ). We speculate that the R species can neutralize or detoxify the environment from the antagonist substance produced by the antagonist species. Microbial models often fail to capture the immediate consequences of new individuals invading a community, a crucial aspect akin to missing challenges and conflicts in a story plot and only finding out how the story concludes. In the BARS experimental model, by assessing the dynamics at 5 min, the strong antagonism and immediate responses between the paired interactions of A and S were observed, as well as the seemingly instantaneous emergent properties in the triple interaction of A, R, and S, that allowed the three species to coexist. Understanding these early stages of community formation is critical for predicting long-term dynamics and the development of stable, functional communities. The rapid molecular responses in bacteria can explain these rapid responses. B. subtilis serves as an example of a rapid defensive response against antibiotics. Cao et al. (2002) showed that in Bacillus subtilis , Extracellular Sigma Factors (ECF) coordinate the defensive response within 3 to 10 min, specifically in response to antibiotics that inhibit cell wall synthesis. ECF sigma factors help maintain cell envelope integrity by activating the expression of genes that inactivate or detoxify antibiotics or alter cell surface properties. Therefore, a quick response by the bacterial population through their defense mechanisms is crucial for survival. This rapid response in BARS suggests that the response is regulated and has important implications for understanding natural communities. Niehus et al. (2021) modeled the competition of species and the effect of a regulated versus a constitutive production of a toxin. They found significant benefits of regulation of timing as this minimizes cost and maximizes the effect on an opponent. They suggest that this is similar to classical predictions from the game theory of animal combat and that the regulation of combat in bacteria is about timing an attack and turning off defenses once the attack is over ( Niehus et al., 2021 ). To address timing and gene expression in synthetic communities, it is exciting the significant advancements in microfluidic analysis. These devices offer single-cell analysis, cell–cell interactions and growth dynamics with high spatio-temporal resolution, allowing for the control of environmental conditions (physical, biological, and chemical stimuli) in a high-throughput manner (reviewed in Burmeister and Grünberger, 2020 ). HOI can be observed in studies of synthetic communities of microorganisms reported in recent years ( Abrudan et al., 2015 ; Gallardo-Navarro and Santillán, 2016 ; Lozano et al., 2019 ; Mickalide and Kuehn, 2019 ; Piccardi et al., 2019 ), some of them showing some properties similar to the BARS community dynamics. Gallardo-Navarro and Santillán (2016) described a model in which a resistant species ( Staphylococcus sp.) seems to protect a sensitive one ( Bacillus. aquimaris ) from the antagonism of a B. pumilus species. Another example comes from the work of Lozano et al. (2019) , in a model named THOR (the hitchhikers of the rhizosphere), with a similar three-species community; interestingly, in this model, the resistant species is also B. cereus that protects a species of Flavobacterium johnsoniae from growth inhibition by Pseudomonas koreensis. Given the emergent properties that we and others have observed, maybe we have to carefully evaluate more data before applying a “rule” such as the one proposed by Friedman for the prediction that species that all coexist with each other in pairs will survive, whereas species that are excluded by any of the surviving species will go extinct ( Friedman et al., 2017 ). Since complex network systems with non-linear dynamics are capable of self-organization, characterized by feedback, stability, and hierarchy, and such systems display new arising or emergent features, we do not think that paired interactions can predict the stability or survival of a species. One of the key features of complex systems is that they often exhibit non-linear dynamics, making them more susceptible to sudden transitions or changes in behavior. In BARS, the ratio of resistance cells relative to those of the antagonist and sensitive species defines whether the antagonistic interaction is subdued or increased. In the report from Medaney et al. (2015) , they observed an increase in the cell density of the interactors as an emergent property since the population of the sensitive species was positively correlated with the cell density of the resistant one. It was intriguing that protection from the R species occurred when its proportion relative to the antagonist was small (10:1:0.25) but when the cell density of the resistant species was lowered (10:1:0.05), the sensitive species seemed to be more affected by the antagonist. The density experiment with the resistant species surprised us, as a slight decrease in the ratio of R seemed to send the system in the opposite direction. Since the BARS model dynamics are very sensitive to the initial species cell ratio, suggesting non-linear dynamics, experiments at different densities and frequencies of species A, S, and R are needed for the mathematical modeling of the interaction. Understanding the rules governing microbial community dynamics over time is still limited. However, mathematical modeling has emerged as a valuable tool in this field. For example, Coyte et al. (2015) employed a mathematical approach to model microbial communities. They found that high species diversity is more likely to be stable when competitive rather than cooperative interactions dominate the system. The principles gathered from studying such a model will allow to describe the interaction between components of the system, to build models that test predictions and increase our understanding of microbial community dynamics. Although they are intended to predict community dynamics, scaling models of synthetic ecology with culturable microorganisms can be challenging. In the study of human gut communities, some research groups are working on building increasingly large synthetic communities (currently up to 100 species) ( van Leeuwen et al., 2023 ), and mathematical modeling is being used to predict the metabolic behavior of gut communities. Metagenomic and transcriptomic approaches are now commonly used to study microbial communities at large scales and over time ( Taş et al., 2021 ). For instance, using metagenome and metatranscriptome analysis, Chuckran et al. (2021) demonstrated that soil microbial communities respond to carbon inputs within hours by altering gene expression. Despite the difficulties in replicating complex natural ecosystems in laboratory settings, numerous metagenomic studies have been conducted to investigate gut communities, water treatment plants, and industrial bioreactors, among others. Mathematical models are employed to better understand how community composition and emergent functions translate to natural systems ( van den Berg et al., 2022 ). Given the simplicity of the BARS community, transcriptomics analysis will allow identifying genes involved in the immediate response to community interactions. The results will aid in the identification of substances involved in tolerance and antagonism, and may reveal if there is a connection between the dynamics of the paired and the triple interaction."
} | 7,678 |
33731836 | PMC8397785 | pmc | 5,657 | {
"abstract": "Microbial populations often experience fluctuations in nutrient complexity in their natural environment such as between high molecular weight polysaccharides and simple monosaccharides. However, it is unclear if cells can adopt growth behaviors that allow individuals to optimally respond to differences in nutrient complexity. Here, we directly control nutrient complexity and use quantitative single-cell analysis to study the growth dynamics of individuals within populations of the aquatic bacterium Caulobacter crescentus . We show that cells form clonal microcolonies when growing on the polysaccharide xylan, which is abundant in nature and degraded using extracellular cell-linked enzymes; and disperse to solitary growth modes when the corresponding monosaccharide xylose becomes available or nutrients are exhausted. We find that the cellular density required to achieve maximal growth rates is four-fold higher on xylan than on xylose, indicating that aggregating is advantageous on polysaccharides. When collectives on xylan are transitioned to xylose, cells start dispersing, indicating that colony formation is no longer beneficial and solitary behaviors might serve to reduce intercellular competition. Our study demonstrates that cells can dynamically tune their behaviors when nutrient complexity fluctuates, elucidates the quantitative advantages of distinct growth behaviors for individual cells and indicates why collective growth modes are prevalent in microbial populations.",
"conclusion": "Conclusions Our work uses a well-studied and ecologically relevant model system to directly control nutrient complexity in order to understand the effect on transitions between aggregative and solitary growth behaviors. Our results have important and direct implications for how the spatial associations and dynamics of individual cells influence the ecological and evolutionary properties of microbial populations in natural ecosystems. For example, cells live and divide in close proximity of each other within biofilms, which not only represent an important growth mode for bacteria in nature [ 45 ], but are also of relevance in industrial and medicinal applications. Finally, our findings elucidate the importance of a potentially ubiquitous strategy in bacterial populations, whereby individuals can transition between distinct behavioral states in response to fluctuating environments.",
"introduction": "Introduction Bacteria in natural environments exhibit distinct behavioral states such as living in close spatial proximities of each other within surface-attached biofilms or in solitary planktonic states [ 1 – 4 ]. Transitions between aggregation and planktonic behaviors are common and are governed by diverse molecular cues [ 4 ]. However, the functionality of distinct behavioral modes, which in all likelihood represent adaptations to different environmental conditions, remains understudied and unclear. This is because it is challenging to identify the benefits that collective or solitary growth modes provide to individual cells and analyze when it is advantageous for cells to switch between growth modes. One explanation could be that the benefit of different growth modes depends strongly on the nutrient environment, since nutrients fundamentally influence cellular pathways that drive growth. Therefore, we asked if the complexity of the growth substrate determines whether cells engage in solitary or aggregative behaviors. Aggregative growth, where individuals self-organize in close spatial association, can be beneficial when cells release compounds that modify the extracellular environment. Prominent examples of such compounds include iron-chelators [ 5 ] and enzymes that degrade complex polysaccharides like chitin [ 6 ] and disaccharides like sucrose [ 7 ]. These compounds generate diffusible resources – for example, simple sugars or metals in a bioavailable form – that are transiently accessible for all nearby cells. Cells in close spatial associations can thus benefit from diffusible resources emerging from the activity of other individuals [ 1 , 5 , 8 ]. In contrast, solitary behavior, where cells are planktonic and often motile, reduces local competition and allows dispersal to new habitats [ 2 , 4 ]. This suggests that growth substrates can be key modulators of behavior in bacteria. An important question thus arises: How does the complexity of the nutrient affect dynamic behavioral transitions in bacteria? Nutrient environments that bacteria encounter in natural ecosystems are dominated by complex polysaccharides [ 9 – 12 ]. These generally have large molecular sizes and are often particulate. Thus, cells must degrade these polymeric substrates using extracellular enzymes into simpler mono- and multi-meric forms, which can then be taken up by cells and catabolized [ 13 ]. In well-mixed environments, extracellular degradation products generated by the enzymatic machinery of an individual cell can be lost due to diffusion [ 5 ]. As a result, polymeric growth substrates are expected to reduce the productivity of microbial populations relative to monomeric substrates. Therefore, we reasoned, in accordance with previous work on yeast [ 7 ], that growth on complex carbohydrates should favor growth of cells as collectives. This is because group behavior will lead to an increase in the per capita payoff due to the reduction in diffusional loss and higher benefit from the degradative activities of neighboring cells. Our goal here was to obtain a quantitative estimate of the advantages of aggregation at the level of individual cells and to understand when it is beneficial for cells to switch from aggregative to planktonic growth. We used Caulobacter crescentus as a model system to study growth behaviors of cells on xylan, a naturally abundant polysaccharide. Xylan is a major component of plant biomass (up to 30%) and thus is a common recalcitrant compound in natural ecosystems [ 14 ]. C. crescentus is ubiquitous in both aquatic and terrestrial environments, and has the biochemical repertoire to metabolize xylan in addition to several other complex polysaccharides [ 12 , 15 – 17 ]. Asymmetrical cell division in C. crescentus gives rise to two different cell types, a sessile stalked cell and a flagellated swarmer cell. The swarmer cell is motile and differentiates into a sessile stalked cell before initiating cell division [ 18 ]. The differentiation between motile swarmer cells and sessile stalked cells is controlled by nutrient signals [ 19 ] and thereby, C. crescentus represents a good model system to study behavioral responses to nutrient complexity in bacteria. We used well-mixed batch cultures to study the growth dynamics of C. crescentus populations and combined this with a quantitative analysis of behaviors at the single-cell level in the presence of either the polysaccharide xylan or its constituent monosaccharide xylose (Fig. 1a ). In order to track and quantify the growth and behavior of individual cells, we used a combination of microfluidics, time-lapse microscopy and automated image analysis. These measurements were performed in microfluidic devices containing micron-scale growth chambers (Supplementary Fig. 1 ), where bacterial cells could grow and move, and nutrients could freely diffuse, albeit at a reduced rate compared to well-mixed conditions [ 20 ]. The environment that cells experienced within chambers could be controlled and altered by changing the nutrient source from polymers to monomers and vice-versa, making it possible to measure behavioral transitions at the level of individual cells. Fig. 1 The polymer xylan limits the growth of C. crescentus compared to the monomer xylose. a Caulobacter crescentus CB15 cells were grown in the same concentration (%weight/volume) of the polymer 0.05% xylan or its constituent monomer 0.05% xylose and b the growth dynamics of populations (optical density at 600 nm) were measured. c Maximum growth rate and d maximum optical density observed over the course of a growth cycle. Compared to populations on xylan, populations grown on xylose achieve higher growth rates (h −1 ) (independent samples t -test, P = 0.0019, R 2 (eta 2 ) = 0.82, n populations = 4 for each treatment) and greater maximum optical density (independent samples t -test, P = 0.0001, R 2 (eta 2 ) = 0.94, n populations = 4 for each treatment). Squares, horizontal lines and whiskers indicate the individual measurements for each biological replicate population ( n populations = 4), the mean and the 95% confidence interval (CI), respectively, on xylan (yellow) and xylose (blue). Asterisks indicate significant differences.",
"discussion": "Discussion In natural ecosystems, cell aggregates or clumps are frequently observed when cells colonize and degrade particulate organic matter like chitin and alginate using chitinases and alginate lyases, respectively [ 10 , 31 – 33 ], or degrade simpler polysaccharides like sucrose through the action of invertases [ 7 ]. Our study shows that such aggregative behavior can be a beneficial trait for bacteria growing on complex polysaccharides. An increase in local cell density results in an increase in the growth rate of cells. Since, the enzymatic activity is localized on individual cells, group formation likely allows cells to benefit from the activity of their immediate neighbors [ 20 ] through the exchange of breakdown products over small spatial distances and can potentially increase the rate of degradation of polymers [ 30 , 33 – 35 ]. In addition, aggregation could be a strategy to exclude cells of other species or strains so that only kin benefit from breakdown products [ 30 , 34 ]. That dispersal from aggregates and biofilms is triggered in response to changes in nutrient availability is well known [ 2 , 4 ]. Our results showing departure from aggregates not only align well with previous findings but also provide novel insight into the role of nutrient complexity in driving behavioral transitions. It is known that bacterial cells can integrate information on nutrient availability and cell density to time their departure from biofilms [ 2 ]. Therefore, when the benefit of collectively degrading a resource no longer exists, cells are able to respond and engage in dispersal to solitary growth modes, which can serve multiple purposes. First, when simpler nutrients like monomers are present, cells can potentially reduce competition amongst individuals through reduced local densities [ 4 , 17 , 32 ]. The finding that an increased final cell density on the monomer xylose can reduce the final growth rate supports the existence of competition at high density. Second, cells can depart from aggregates to find and colonize new nutrient patches [ 4 , 36 ]. The presence of monomers or absence of polymer in the environment could serve as a signal for cells that polymeric resources on particulate organic matter have reduced. In addition, cells can use chemotaxis systems in order to respond to such nutrient gradients in nature. It is known that C. crescentus possess multiple chemotaxis clusters which differentially regulate swimming behaviors towards nutrients like xylose or attachment to surfaces [ 37 , 38 ]. Finally, since the formation of aggregates presumably requires the investment of cellular resources into attachment to surfaces or to other cells, dispersal into solitary modes might represent a cost-saving strategy for bacteria when aggregation no longer provides a benefit. Understanding how cells perceive environmental cues about nutrient complexity or changes in nutrient concentration and respond with appropriate behaviors will shed light on the regulatory pathways that govern such dynamic transitions. The molecular mechanisms that drive collective growth behaviors and environment-induced transitions in bacteria have been widely studied, albeit in relatively non-natural environments [ 39 , 40 ]. Our work uses a well-studied genetic model system to study growth on natural substrates and thus provides a basis to extend mechanistic studies in the context of the natural ecology of bacteria. It is known that cellular appendages like pili and holdfast adhesins allow C. crescentus cells to colonize surfaces [ 41 , 42 ] and thus are likely involved in colony formation on xylan. In contrast, flagella allow cells to swim. Since, both flagellated swarmer and holdfast bearing stalked cells are present within colonies, it is likely that the motile swarmer cells drive the bulk of dispersal events. A systematic test of the role of such cellular structures on natural growth substrates will enable the unraveling of their ecological functions. Interestingly, it is known that the nutritional quality of the environment influences the activity of key regulators that modulate cell differentiation and adhesion in C. crescentus [ 19 ]. The loss of genes involved in in the biosynthesis of the adhesive holdfast can reduce growth and thus impose a fitness disadvantage in lake water [ 43 ], an observation in line with our findings with the holdfast deficient strain. In contrast, the disruption of surface attachment confers a fitness advantage when growing on simple sugars [ 43 ]. These findings imply that the ability to attach and form aggregates can be advantageous in natural ecosystems, where polymeric carbon substrates represent the dominant growth substrate [ 44 ]. We suggest that future work should focus on addressing the importance of processes that are involved in signaling within the cell and the regulation of these behaviors in mediating distinct phenotypic behaviors that help bacterial cells grow on ecologically relevant substrates."
} | 3,426 |
35500122 | PMC9171643 | pmc | 5,658 | {
"abstract": "Significance Corals exhibit highly variable responses to marine heat waves as well as to local biological and ecological circumstances that moderate them across reef seascapes. This variability makes identifying refugia—reefs possessing conditions that increase coral resilience—nearly impossible with traditional surveys. We developed and applied an airborne coral mortality mapping approach to identify reef refugia in a major marine heat wave across the Hawaiian Islands. A combination of human and environmental factors, including reduced coastal development and lower sedimentation levels, advantaged the majority of refugia over neighboring reefs. High-resolution monitoring of coral mortality reveals a reef geography of both resilience and vulnerability to climate change.",
"discussion": "Discussion Our study highlights the unique role that coral mortality mapping can play in support of reef conservation and management as ocean climate continues to change. Whether in a laboratory or field setting, understanding which corals bleach is important for assessments of heat and light stress response. However, coral morality mapping provides an avenue to determine net rates of coral loss over large areas, which reveal ecosystem-level resilience. Maps of coral mortality following marine heat waves and other disturbances provide input to the growing field of coral reef restoration, which must grapple with numerous issues that limit reef recovery efforts ( 20 , 21 ). Given the pronounced limitations of carrying out coral restoration at large scales, more tactical approaches are urgently needed to identify areas of both heightened coral mortality and survivorship. Coral mortality mapping can focus subsequent effort on moderately impacted reefs that might benefit from direct interventions while avoiding areas of severe decline that may prove fruitless in restoration efforts. Still other reefs showing resilience can be identified and designated for protection. These types of emergent options could help manage enormous areas, such as the Great Barrier Reef ( 22 ). While we found reefs with elevated coral resilience to the 2019 marine heat wave, pronounced within-refugium heterogeneity of coral mortality was also observed. This may be an indicator of genetic variability in thermal tolerance within and between coral species ( 23 ). Such complex spatial patterns of coral performance may, therefore, provide options for the selection of thermally tolerant corals for study. Subsequent selective propagation, based on within-refugium coral performance in heat waves, can help managers and conservationists to enhance reef resiliency by strengthening the pool of genotypes used for restoration. Our airborne results reveal a highly complex geography of coral vulnerability to thermal events at individual reef to archipelago scales. While high temperature and light stress remain broad forces of coral mortality, the scale dependence of coral loss and survivorship is far more complex than can be accounted for by these drivers alone. In the 2019 Hawaiʻi marine heat wave, a range of coral mortality levels was observed, and on some reefs in areas of less coastal development and lower land to sea sedimentation, corals performed better than in other areas lacking these land-based impacts. Such results point to the need for management to directly address land–reef issues in a changing climate. Despite this understanding, we still lack nearshore maps of subsurface groundwater discharge, currents, and internal waves, all of which may greatly affect coral outcomes in a warming climate ( 24 , 25 ). Because preexisting live coral cover was a strong predictor of coral resilience in the 2019 marine heat wave, it may be the case that some of these unmapped factors are supporting coral survival. In fact, mapping coral persistence may provide a way to identify where these factors are biologically relevant. Coral resilience results from a combination of genetic thermal tolerance ( 26 ) and mediating factors, such as human stressors and local hydrodynamics ( 18 ). Yet, these interacting factors remain extremely difficult to constrain via traditional studies. Repeat coral mortality mapping could be integrated with laboratory and field-based approaches to better understand options and to drive efforts at scales of greater ecological efficacy. While the remote sensing with imaging spectroscopy used here is currently only available via a few airborne platforms, space-based systems are in rapid development ( 27 ), which will make coral mortality mapping routine in a few years. The discovery and monitoring of coral reef refugia offer an important pathway for meaningful conservation interventions that protect more corals in our changing climate."
} | 1,190 |
35479988 | PMC9037638 | pmc | 5,660 | {
"abstract": "Furfural is a promising renewable platform molecule derived from hemi-cellulose, which can be further converted to fossil fuel alternatives and valuable chemicals due to its highly functionalized molecular structure. This mini-review summarizes the recent progress in the chemo-catalytic and/or bio-catalytic conversion of furfural into high-value-added chemicals, including furfurylamine, C 6 carboxylic acid, i.e. , furandicarboxylic acid, furfural alcohol, aromatics, levulinic acid, maleic acid, succinic acid, furoic acid, and cyclopentanone, particularly the advances in the catalytic valorization of furfural into useful chemicals in the last few years. The possible reaction mechanisms for the conversion of furfural into bio-chemicals are summarized and discussed. The future prospective and challenges in the utilization of furfural through chemo- and bio-catalysis are also put forward for the further design and optimization of catalytic processes for the conversion of furfural.",
"conclusion": "3. Conclusions and outlook This review presented a comprehensive overview on the latest advancements in the catalytic valorization of furfural into value-added chemicals through chemo- and bio-catalysis approaches, focusing on the development of catalysts and the reaction mechanisms of furfural conversion. The wide utilization of biomass-derived furfural opens a new avenue to reduce the dependence on oil and fossil resources and provides a sustainable development for mankind. Despite the great progress achieved in the efficient utilization of furfural to produce useful chemicals including furandicarboxylic acid, aromatics, and cyclopentanone, and the fact that some bottlenecks have been addressed and some emerging strategies have been exploited, there are still critical challenges that hinder the utilization of furfural to produce value-added chemicals on a large scale. (1) The bio-catalytic conversion of furfural provides a green protocol to valuable chemicals. Compared to chemo-catalysis, bio-catalysis has some advantages including high product selectivity and environmentally friendly nature, but it has some limitations, such as low productivity and low feedstock concentration, leading to extensive energy consumption for product separation and purification in the downstream processes. Thus, it is urgent to develop some novel processes by integrating bio-catalysis and chemo-catalysis to make full use of their advantages and overcome their drawbacks in the transformation of biomass into useful chemicals. (2) Significant challenges remain in designing and developing industrially feasible reaction pathways through heterogeneous catalysts and/or bio-catalysts for the conversion of furfural into useful chemicals including diacids, diols, aromatics, and cyclopentanone. To date, most research on the utilization of furfural is still in its infancy, such as the production of furandicarboxylic acid, aromatics, succinic acid, and cyclopentanone. Thus, great efforts need to be further undertaken to meet the commercial-scale application requirements. Recently, artificial intelligence (AI)-driven synthesis has been shown to be a very strong tool in organic synthesis, materials design, etc. This novel technique will remarkably facilitate the screening of excellent catalysts and the development of efficient and economic pathways for the valorization of furfural to value-added chemicals. (3) In general, the catalytic conversion of furfural to bio-chemicals requires cascade reactions, especially for hydrogenation and hydrodeoxygenation, which are usually conducted under harsh reaction conditions at high temperatures and pressures. These harsh reaction conditions result in the degradation of furfural to produce other by-products or humins, which will increase the production cost of the desired products. In addition, the hydrothermal reaction conditions easily lead to leaching of the active components and even deactivation of solid catalysts. Thus, it is of great significance to develop efficient and robust heterogeneous catalysts for the conversion of furfural. Recently, the photo-, electro-, and photoelectron-catalytic conversion of biomass into bio-fuels and bio-chemicals have attracted increasing attentions due to their mild reaction conditions, high selectivity, and environmentally friendly nature. The development of photoelectrochemical processes can minimize the energy input, while affording high conversion efficiency and selectivity. It can be expected that some innovative reaction pathways and novel valuable bio-chemicals will be yielded through these new techniques. (4) It is necessary to have an in-depth understanding of the reaction mechanism for further improving the design and catalytic properties of solid catalysts. Besides the identification of the key reaction intermediates, the surface/interface reaction kinetics also need thorough studies. However, it is a significant challenge to in situ dynamically identify and follow the evolution of reaction intermediates, especially under the harsh hydrothermal reaction conditions. Thus, in situ and even operando characterization techniques and theoretical calculation may act as useful tools for this goal, which can propel our understanding of the structure–activity relationship of catalysts and promote the development of high-efficient catalysts for the valorization of furfural.",
"introduction": "1. Introduction With the depletion of fuel reserves and increasing demands for fuel, great efforts have been devoted to developing novel sustainable and renewable energies and resources. 1–5 Biomass is the only renewable organic carbon source in nature, which is environmentally friendly, has abundant reserves and inexpensive, providing unique advantages for the production of fuels and industrially important chemicals. 5–7 Furfural is a versatile platform molecule, and thus has received considerable attention in recent years. The presence of an aldehyde group in furfural makes it possible to further synthesize higher value chemicals and polymer monomers, including furandicarboxylic acid, furfurylamine, furfural alcohol, cyclopentanone, and levulinic acid and other molecules through oxidation, hydrogenation, and hydrolysis ( Fig. 1 ). 6 Hence, the high-value utilization of biomass-derived furfural is important for sustainable development in the future. Fig. 1 Furfural as a platform compound for diverse reactions. Furfural is produced from agricultural resources containing xylose or xylan, such as wood wastes and corn cob. 8 The main producers of furfural are China (∼70% total production capacity), The Dominican Republic (Central Romana Corporation, 32 kton per year), and South Africa (20 kton per year). These three countries account for approximately 90% of the global furfural capacity (280 kton per year). 9,10 In the past, most studies have been carried out to reveal the reaction mechanism for the conversion carbohydrates into furfural at the molecular level. As shown in Fig. 2 , initially, the formation of furfural from hemicellulose-containing biomass includes the acid-catalyzed depolymerization of the hemicellulose component to form xylose. Subsequently, the isomerization of xylose to xylulose occurs followed by a dehydration reaction to yield furfural. 11 To date, various reaction systems such as biphasic reaction systems and ionic liquids have been developed to boost the yield of furfural. Fig. 2 Conversion pathway of hemicellulose component into furfural. Some excellent reviews about the conversion of hemicellulose to furfural and the valorization of furfural into bio-fuels have been published. 5,6,12–14 A systematic review of the conversion of xylan into furfural is beyond the scope of this manuscript. However, a comprehensive summary of latest advances in the valorization of furfural to value-added chemicals, especially the progress in the last five years, has not been reported to date. This mini-review summarizes the recent advances in the conversion of furfural into valuable chemicals through chemo- and bio-catalysis. This mini-review mainly focuses on the catalytic strategies for the valorization of furfural into useful chemicals including selective oxidation reactions, hydrogenation, hydrogenolysis, reductive amination, and aromatization. The valuable products, include maleic anhydrides, maleic acid, succinic acid, furoic acid, furfurylamine, furandicarboxylic acid, furfural alcohol, aromatics, levulinic acid, and cyclopentanone, especially for advancements in the catalytic conversion of furfural to useful chemicals in the last few years. The possible reaction mechanisms for the conversion of furfural into bio-chemicals are summarized and discussed. In addition, future prospects are also provided to highlight the challenges and opportunities for the valorization of furfural."
} | 2,213 |
25309305 | null | s2 | 5,663 | {
"abstract": "Superhydrophobic, porous, 3D materials composed of poly( ε -caprolactone) (PCL) and the hydrophobic polymer dopant poly(glycerol monostearate- "
} | 35 |
23029297 | PMC3454355 | pmc | 5,667 | {
"abstract": "A water drop on a superhydrophobic surface that is pinned by wire loops can be reproducibly cut without formation of satellite droplets. Drops placed on low-density polyethylene surfaces and Teflon-coated glass slides were cut with superhydrophobic knives of low-density polyethylene and treated copper or zinc sheets, respectively. Distortion of drop shape by the superhydrophobic knife enables a clean break. The driving force for droplet formation arises from the lower surface free energy for two separate drops, and it is modeled as a 2-D system. An estimate of the free energy change serves to guide when droplets will form based on the variation of drop volume, loop spacing and knife depth. Combining the cutting process with an electrofocusing driving force could enable a reproducible biomolecular separation without troubling satellite drop formation.",
"conclusion": "Conclusions Using a variety of techniques to fabricate superhydrophobic surfaces and knives, a drop of water pinned to wire loops could be cut in a gentle fashion so that no satellite drops are formed. This technique complements a previous publication [4] describing how isoelectric focusing can be conducted on an aqueous drop resting on a superhydrophobic surface. Drop cutting is a further step that completes the separation of proteins that preferentially migrate to an electrode during focusing. A 3-D model is presented and experimental data compared to a 2-D thermodynamic model which can guide the separation distance of the wire loop electrodes needed for cutting as a function of the initial drop volume. Surfaces that repel other types of liquids, such as superoleophobic surfaces, could also cut organic or ionic liquids via the method described.",
"introduction": "Introduction In the past few years there has been a spectacular growth in the number of scientific articles describing the manufacture of water repellent surfaces, also known as superhydrophobic surfaces, for a wide range of consumer and industrial uses [1] – [3] . Similarly, droplet microfluidics, which is generally described as the creation and manipulation of defined droplets in an insoluble continuous phase, is a popular topic in biotechnology, analytical instrumentation, high-throughput screening, and other instrumentation development. A particular focus is using the unique surfaces to enable the manipulation of individual drops so that a complex mixture can be rapidly and inexpensively resolved into individual components. In general, a major challenge in biomolecular separations is to separate a large number of key proteins from biological fluids for a variety of clinical and biotechnological applications. Multiprotein separation is vital for the detection of important proteins that provide valuable information on gene expression and can serve as early signals of a disease state. Currently, technological solutions are limited to using specific labels ( e.g. , ligands or antibodies) or an array of instruments with accompanying sample preparation steps, which usually require expert handling. Finding a rapid, efficient and simple means of separating components in a small sample, such as a drop, without using channels, stationary phase, gels or other transfer media is a two-pronged problem. One needs, firstly, a suitable means of generating conditions within the drop for separating molecules and, secondly, a means of collecting one or more components separated from the rest. Previously, the generation of a pH gradient suitable to create isoelectric focusing in a drop sitting on a superhydrophobic surface was demonstrated [4] . Although a micropipette system or some other suitable, additional instrument could remove the isolated protein, our group became interested in developing a strategy to divide the drop without generating undesired mixing effects or satellite drops. Simply stretching the drop in order to divide it generates a meniscus shape in the liquid leading to a thinning of the bridge followed by a catastrophic rupture often resulting in small satellite drops as can be seen in liquid jets [5] – [9] . Instead, a means to cut the drop at a particular location with a superhydrophobic knife would be more helpful in achieving the desired goal of separating molecular components within a single drop. Satellite droplets are often formed when the Weber number ( We = ρU 2 L/γ ) of a system is greater than 1. As a comparison, Weber numbers for our system are on the order of 1×10 −5 . This is mostly due to the very slow speed of the knife descending upon the stretched droplet. Bormashenko & Bormashenko [10] have recently demonstrated that for an unpinned droplet, the speed of a superhydrophobic knife is important when cutting a coated liquid marble or a liquid drop on a superhydrophobic surface. However, our parallel and independent research effort is focused on drop cutting when the system is at equilibrium and while the drop is pinned by two wire loops that can function as electrodes. The physics of this drop cutting method is rather simple compared to other droplet formation and manipulation strategies (flowing streams–Rayleigh-Tomotika analysis [11] – [13] and electrowetting–fully assessed in 2003 [14] ). It depends entirely on the difference in the surface free energy between a distorted drop shape and two separate drops pinned to the wire loops. Our method does not rely on standard flow models or the viscosity of the two fluids of the liquid/air interface. Furthermore, our droplets are not being stretched enough such that they can begin to behave like an unstable fluid filament. In this strategy, a force applied to the surface using a slow-moving superhydrophobic knife–after components are spatially separated using isoelectric focusing or some other separation method–assists the drop-cutting step and leads to the creation of two drops that can be further processed or collected and analyzed. 10.1371/journal.pone.0045893.g001 Figure 1 Schematic contrasting splitting a drop by pulling each end with the method of cutting with a superhydrophobic knife. Both drops lie on superhydrophobic surfaces. A simple rectangle can estimate the shape and contour of the drop ( h top for the top contour of the stretched drop, and h o for the bottom contour of the stretched drop). The split drop results in two equally sized spheres, both of radius r d (equal to the radius of the wire loops used to pin and stretch the original droplet).",
"discussion": "Results and Discussion Understanding the closely related area of liquid jet breakage through experiments and theory development continues to be a relatively active research topic, but there is much less work involving splitting a single drop. Liquid jet breakup research does show that unless vibrations are carefully controlled [6] , [8] , satellite drops are formed. Stretching a drop ( Figure 1 – top) cannot be controlled unless an additional force, such as through the use of controlled vibrations, is superimposed. High throughput flowing droplet formation microfluidic research also shows that careful control of focusing conditions is important in order to control the neck between droplets and hence to avoid satellite drops [17] , [18] . Satellite formation in microfluidics is similar to what occurs in free liquid jets, and it can be generally attributed to the action of droplets coming apart while liquid is held in a thin bridge between the two larger droplets. 10.1371/journal.pone.0045893.t001 Table 1 Upper Limit Predictions and Results. Volume of Drop V s (µL) Measured Separation Distance h o (mm) Upper Limit (mm) Measured h top value (mm) Did Drop Split in Next Frame? 50 8.3 12.9 11.0 No 50 8.3 12.9 12.9 \n Yes \n 60 9.3 14.2 14.2 No 60 9.2 14.2 14.2 No 60 9.2 14.2 14.3 No 60 9.3 14.2 14.3 \n Yes \n 60 10.7 15.0 12.8 No 60 10.7 15.0 12.9 \n Yes \n 70 13.0 17.1 14.5 No 70 13.1 17.1 14.6 \n Yes \n Examples of upper limit predictions as compared to ImageJ measured values and observed drop cutting for a zinc coated superhydrophobic knife and a Teflon superhydrophobic surface (based on the 2-D model). Out of 15 videos capturing drop splitting, there were 74 photos analyzed. In 17 of them (23%), the measured values for h top exceed the calculated upper limit, however all of those were still within 7% of the calculated upper limit. As shown in Figure 1 (a), some information can be gleaned from droplet formation via liquid jet as it generates a concave meniscus that can create satellite drops due to the elastic response of the drop during breakage, which is characterized by oscillations [6] , [8] . In distinct contrast of spontaneously generating a concave meniscus, our technique involves pressing down on the liquid cylinder with a superhydrophobic knife to form two convex menisci, as shown in Figure 1 (b). This form of cutting does not create satellite drops since the surface tension driving force “folds” the liquid inwards on each daughter drop respectively [16] . When slowly cutting a drop pinned on each side with a superhydrophobic knife, the slicing or cutting action gradually eliminates the bridge while allowing time for the liquid to be folded into one droplet or the other. For a given separation distance between the wire loops, the superhydrophobic knife can be slowly lowered until it touches the superhydrophobic surface and then slowly raised back above the drop(s) surface, indicating a reversible system until breakage–in contrast to a liquid jet. A droplet can either split, if shape distortion is sufficiently achieved by the blade, or the drop can resume its original shape upon blade removal. \n Figure 2 (a) shows a typical still image from digital video for a drop that could not be cut, while Figure 2 (b) shows a drop at the point of being cut in two. Still sequences from videos were collected to document the distortion of the upper surface by the superhydrophobic knife in order to compare the profile to an appropriately simple2-D thermodynamic model prediction of the range of conditions that would favor cutting of the drop. The wire loop separation favorable to drop splitting and the contour lengths of the droplets are of particular interest. The following analysis creates a framework to guide drop cutting and provides a range of values needed to achieve cutting. From the analysis of Young and the mathematics of Laplace, the change of the shape of a drop of water is determined by balancing pressure and surface tension in order to achieve a fluid static condition. For our analysis, we only examine two different shapes – a sphere and a cylinder. Assuming a perfectly superhydrophobic surface and that a drop elongated by connecting to two wire hoops creates a cylinder, the energy for both static states at constant volume can be described. Because no material is lost in the drop splitting process, the volume, V , remains constant throughout, but can be expressed for clarity by: V \n s = 4/3 πr s \n 3 for the original sphere, V d = 1/2 V \n s = 2/3 π d \n 3 for the daughter droplets, (assuming equal splitting, subscript d for daughter droplets, of which there are two) and V c = πr c \n 2 \n h for the cylindrically stretched droplet, (subscript c for cylinder) where h is the length of a stretched droplet. See Figure 1 and the discussion below. Surface area varies for the different configurations and gives rise to significant differences in the Gibbs Free Energy, which provides a framework or waypoints to understand the cutting process. Given the surface area of the original sphere, A s = 4 πr s \n 2 , the Gibbs Free Energy to form the spherical interface is then, Δ G = 4 πr s \n 2 \n γ . Similarly, the surface area for a cylinder, A c = 2 πr c h + 2 πr c \n 2 , results in Δ G = (2 πr c h + 2 πr c \n 2 ) γ . For the process of stretching a droplet in preparation for splitting, the droplet shape changes from a sphere to a cylinder, which is a positive free energy change, Δ G = 2 πγ ( r c h + r c \n 2 – r s \n 2 ). Setting the volumes equal to each other, solving for r s and substituting gives us the following relation: (1) \n For the process of cutting the cylinder into two equal size spheres, the energy change is calculated by comparing the energy of a cylinder with two spheres equal to the original volume defined above: (2) \n This free energy change can be positive or negative depending upon the ratio of the cylinder radius to its height ( r c / h ). Assuming that the cylinder starts at a relatively small value of h with respect to r c , by stretching the cylinder so that h is 4–5 times the radius, the free energy change for this process is negative meaning that the drop would minimize its free energy by forming two separate drops. This relatively simple two stage model fully and simply describes the 3-D energetics of the system. To provide a quantitative assessment for the imaging data of the drop cutting, a reduced model is needed since full 3-D imaging is not desirable or especially feasible. The important outcome is to design a rubric to identify conditions favoring drop cutting that can be deduced from the 2-D free energy analysis and can be effectively compared to experiments by imaging of a side view of the cutting process. In this model, the energy change considers only the circumference of the cylinder and spheres that are at the liquid/air interface, and assumes that the surface tension force acts across the length of the interfacial line. Here, the liquid/air cylinder circumference is considered to be only the top ( h top ) and bottom ( h o ) lengths (see Figure 1 ). For this 2-D analysis, the liquid/air circumference prior to splitting is a rectangle, and if the drop splits, the liquid/air circumference of the two drops is two half-circles. The ends of the rectangle and other half of the circles pinned by the wire loops during the splitting process are not considered in this model since these surfaces undergo no changes in interfacial energy. The surface length of h top is the only length assumed to be increasing due to the action of the superhydrophobic knife. The 2-D model yields the following free energy equation assuming that the drop splits evenly: (3) \n Thus, for a given length of a cylindrically shaped drop, the creation of a higher interfacial length ( h top ) can lead to a negative free energy change. As the drop is stretched more due to a wider spacing between the wire loops, it is comparatively easier to slice the drop with the superhydrophobic knife since the increase in the needed interface length as a percentage of the original length is decreased. The lower limit to achieve drop splitting is found by setting equation (3 ) equal to zero and solving for h top in terms of V . The minimum value of h top in order to achieve drop cutting is: (4) \n However, this underestimates the value of h top needed when the drop volume is larger. The upper limit in principle is the maximum distortion the knife can impose upon the droplet. To calculate the upper limit, an arc length must be calculated for the top contour of the drop. For the upper limit estimate, the following general equation is used to approximate the drop profile upon cutting near the knife: (5) where A corresponds to the height of the droplet’s upper surface above the Teflon slide, which can be obtained from the diameter of the hoops, d ; B corresponds to normalized time assuming the knife is moving at a constant velocity, and is equal to the value 1 for the purposes of calculating the upper limit; and C is the ratio h o / V s . Figure 3 shows four typical still images from the videos with a curve drawn from equation (5 ) superimposed to demonstrate how this equation reasonably describes the droplet profile during splitting. The upper limit of h top is found from the droplet being distorted to the maximum by the superhydrophobic knife. Taking the arc length of the drop profile at the point when the knife has descended as far as it can before touching the Teflon slide surface, and substituting in the meaningful variables for A , B , and C as described above gives: (6) where: \n (7) \n Table 1 compares calculated upper limits for various droplet scenarios with the measured values of h top using the ImageJ software and shows whether the deformed drop was about to split in the next frame of the video. For a range of separation distances of 3–13 mm and drop sizes of 15–70 µL, the measured h top values when drop cutting occurs more closely follow the upper limit prediction. Our observations also show that measured h top values which do not well exceed the lower limit are insufficient to generate drop splitting. Larger droplets can achieve splitting before reaching the upper limit due to their instability. We recommend that a separation distance between wire loops or other pinning surfaces of 4–5 times the original drop radius is a good rule of thumb to cut drops in this fashion, which is similar to the geometric rule of thumb as noted for laminar jets [8] , [9] . Conclusions Using a variety of techniques to fabricate superhydrophobic surfaces and knives, a drop of water pinned to wire loops could be cut in a gentle fashion so that no satellite drops are formed. This technique complements a previous publication [4] describing how isoelectric focusing can be conducted on an aqueous drop resting on a superhydrophobic surface. Drop cutting is a further step that completes the separation of proteins that preferentially migrate to an electrode during focusing. A 3-D model is presented and experimental data compared to a 2-D thermodynamic model which can guide the separation distance of the wire loop electrodes needed for cutting as a function of the initial drop volume. Surfaces that repel other types of liquids, such as superoleophobic surfaces, could also cut organic or ionic liquids via the method described."
} | 4,477 |
27801909 | PMC5322295 | pmc | 5,669 | {
"abstract": "The cellulolytic protist Trichonympha agilis in the termite gut permanently hosts two symbiotic bacteria, ‘ Candidatus Endomicrobium trichonymphae' and ‘ Candidatus Desulfovibrio trichonymphae'. The former is an intracellular symbiont, and the latter is almost intracellular but still connected to the outside via a small pore. The complete genome of ‘ Ca. Endomicrobium trichonymphae' has previously been reported, and we here present the complete genome of ‘ Ca. Desulfovibrio trichonymphae'. The genome is small (1 410 056 bp), has many pseudogenes, and retains biosynthetic pathways for various amino acids and cofactors, which are partially complementary to those of ‘ Ca . Endomicrobium trichonymphae'. An amino acid permease gene has apparently been transferred between the ancestors of these two symbionts; a lateral gene transfer has affected their metabolic capacity. Notably, ‘ Ca. Desulfovibrio trichonymphae' retains the complex system to oxidize hydrogen by sulfate and/or fumarate, while genes for utilizing other substrates common in desulfovibrios are pseudogenized or missing. Thus, ‘ Ca. Desulfovibrio trichonymphae' is specialized to consume hydrogen that may otherwise inhibit fermentation processes in both T. agilis and ‘ Ca. Endomicrobium trichonymphae'. The small pore may be necessary to take up sulfate. This study depicts a genome-based model of a multipartite symbiotic system within a cellulolytic protist cell in the termite gut.",
"introduction": "Introduction Termites require symbioses with gut microbes, in order to digest dead plant matter and obtain nitrogenous compounds ( Brune, 2014 ; Hongoh, 2011 ). Members of phylogenetically basal (‘lower') termite taxa harbour in their guts a dense community of protists, bacteria and archaea. The protists generally establish a symbiotic relationship with multiple species of prokaryotes, which reside in their cytoplasm, nucleoplasm or attach onto the cell surface ( Brune, 2014 ; Sato et al. , 2014 ). Although metagenome, metatranscriptome and metabolome analyses of the microbiota in the gut of lower termites have been performed ( Tartar et al. , 2009 ; Do et al. , 2014 ; Tokuda et al. , 2014 ), the functions of individual microbial species and their interrelationships mostly remain unclear. In particular, the multipartite symbiotic system comprising cellulolytic protists and their multiple prokaryotic endo- and/or ectosymbionts has not been characterized in detail. Trichonympha agilis , a cellulolytic parabasalid protist that is present in the gut of the termite Reticulitermes speratus , hosts two bacterial symbionts, ‘ Candidatus Endomicrobium trichonymphae' phylotype Rs-D17 (class Endomicrobia ; Stingl et al. , 2005 ; Ohkuma et al. , 2007 ) and ‘ Candidatus Desulfovibrio trichonymphae' phylotype Rs-N31 (class Deltaproteobacteria ; Sato et al. , 2009 ). The cellular association between T. agilis and these two bacteria species is permanent: ca . 4,000 and 1,800 cells of ‘ Ca . Endomicrobium trichonymphae' and ‘ Ca . Desulfovibrio trichonymphae', respectively, always inhabit the T. agilis cell in specific subcellular locations, as shown in Figures 1a and b ( Sato et al. , 2009 ). These bacteria account for ca . 4% and 2% of the total prokaryotic cells in the R. speratus gut, respectively ( Sato et al. , 2009 ). The complete genome sequence of the uncultured, intracellular symbiont ‘ Ca . Endomicrobium trichonymphae' (1.15 Mb, including plasmids) was previously obtained using a whole genome amplification (WGA) technique ( Hongoh et al. , 2008a ). Its small genome showed the potential to synthesize various amino acids and cofactors and to ferment monosaccharides to acetate, lactate, ethanol, CO 2 and H 2 ( Hongoh et al. , 2008a ). ‘ Ca. Desulfovibrio trichonymphae' is uncultured and was previously considered to be an intracellular symbiont ( Sato et al. , 2009 ). However, transmission electron microscopy (TEM) of Trichonympha globulosa in the gut of the termite Incisitermes marginipennis revealed that its Desulfovibrio symbionts are localized in deep invaginations of the host ( Trichonympha ) plasma membrane that are open to the exterior of the host cell ( Strassert et al. , 2012 ). Our re-examination showed that ‘ Ca. Desulfovibrio trichonymphae' phylotype Rs-N31 cells were almost completely buried in the host cytoplasm but still connected to the outside through a small pore ( Figures 1c and d ). Analyses using the reverse transcription polymerase chain reaction (RT-PCR) showed that ‘ Ca. Desulfovibrio trichonymphae' phylotype Rs-N31 transcribed the dsrAB and apsA genes that are responsible for sulfate reduction and hynA for hydrogen oxidation ( Sato et al. , 2009 ). No other information on the functions of ‘ Ca. Desulfovibrio trichonymphae' has been available hitherto. In this study, we attempted to acquire the complete genome sequence of ‘ Ca . Desulfovibrio trichonymphae' phylotype Rs-N31, in order to predict its functions and roles in the symbioses with T. agilis and ‘ Ca . Endomicrobium trichonymphae' phylotype Rs-D17. Our results provide a genome-based model of a tripartite symbiotic system within a cellulolytic protist cell in the termite gut.",
"discussion": "Discussion The present study revealed that ‘ Ca . Desulfovibrio trichonymphae' retains the complex system required to oxidize H 2 by sulfate and/or fumarate in spite of its reduced genome size. In contrast, the bacterium has lost the ability to utilize other electron donors common in desulfovibrios, such as lactate, formate, and ethanol. The corruption of the repressor gene rex suggested that ‘ Ca . Desulfovibrio trichonymphae' constitutively expresses genes for sulfate respiration. Since the protist host and the co-inhabiting ‘ Ca . Endomicrobium trichonymphae' most probably generate H 2 during the fermentation of sugars ( Yamin, 1980 ; Odelson and Breznak, 1985 ; Hongoh et al. , 2008a ; Zheng et al. , 2016 ), we suggest that interspecies H 2 transfer is one of the driving forces for the evolution of this tripartite symbiosis. Consistently, ‘ Ca . Desulfovibrio trichonymphae' cells are localized immediately adjacent to hydrogenosome-like compartments ( Figure 1 ; Sato et al. , 2009 ), organelles that produce H 2 ( Müller et al. 2012 ). Although the H 2 partial pressure in the protist-inhabiting portion (paunch) of agarose-embedded termite guts is extremely high (for example, 15–30 kPa in Reticulitermes santonensis ), the H 2 -emission rate of living termites is 30 to 50 fold lower ( Ebert and Brune, 1997 ; Pester and Brune, 2007 ). In addition, the H 2 partial pressure steeply decreases toward the peripheral gut region ( Ebert and Brune, 1997 ; Pester and Brune, 2007 ). These indicate that H 2 is rapidly removed by H 2 -oxidizers, especially in the gut of living termites, and imply that competition for H 2 can occur among those microbes. Indeed, exogenously-supplied H 2 greatly enhances the methanogenic activity of the termites Zootermopsis angusticolis and Reticulitermes flavipes , both of which harbour methanogens that produce CH 4 from H 2 and CO 2 in their gut ( Messer and Lee, 1989 ; Ebert and Brune, 1997 ). Therefore, it is tempting to infer that an ancestor of ‘ Ca . Desulfovibrio trichonymphae' might have colonized the surface of a Trichonympha cell for an abundant H 2 supply, and subsequently evolved into a vertically transmitted, almost intracellular symbiont. The small pore may be necessary for sulfate uptake from the outside of the host protist cell. The colonization on the Trichonympha cells should also have eliminated the cost and need for its own motility that is otherwise required to keep the bacterium at nutritionally optimal sites in the gut and to prevent washout from the gut. The removal of H 2 by ‘ Ca . Desulfovibrio trichonymphae', in turn, likely benefits the protist host and the co-inhabiting ‘ Ca . Endomicrobium trichonymphae' by decreasing the inhibitory effect of H 2 against their fermentation processes. The fact that all the T. agilis cells in the R. speratus gut harbour ‘ Ca . Desulfovibrio trichonymphae' strongly suggests that they have a mutualistic relationship. On the other hand, certain Trichonympha species do not possess Desulfovibrio symbionts ( Strassert et al. , 2012 ). Other endo- and/or ectosymbionts might substitute the role of ‘ Ca . Desulfovibrio trichonymphae'. Otherwise, since H 2 diffuses rapidly in the gut environment, a portion of the gut protist community might not need to house H 2 -oxidizers if there are enough H 2 -oxidizing activities in total in the gut. Thus, the need for cellular association between a protist and H 2 -oxidizers might also depend on the total abundance of H 2 -oxidizers in the gut. Trichonympha -associated Desulfovibrio phylotypes are not monophyletic ( Supplementary Figure S8 ), implying that independent acquisitions of Desulfovibrio symbionts by Trichonympha protists have occurred ( Sato et al. , 2009 ; Strassert et al. , 2012 ; Ikeda-Ohtsubo et al. , 2016 ). Trichonympha collaris in the gut of Zootermopsis nevadensis harbours rod-shaped Desulfovibrio ectosymbionts, which are laterally attached to the host cell surface ( Ikeda-Ohtsubo et al. , 2016 ). The Desulfovibrio ectosymbionts of Trichonympha globulosa are held by invaginations much deeper than those of T. collaris ( Ikeda-Ohtsubo et al. , 2016 ; Strassert et al. , 2012 ). These might be in intermediate stages in the evolution to the nearly intracellular symbiont like ‘ Ca . Desulfovibrio trichonymphae' phylotype Rs-N31. Since the concentration of sulfate in the termite gut is not high (for example 0.00–0.01 mM in R. speratus and 0.09–0.33 mM in H. sjostedti ) ( Sato et al. , 2009 ), the supply of malate from the host cytoplasm for fumarate respiration may be another driving force for the symbioses between these Trichonympha and Desulfovibrio species. Interestingly, T. collaris harbours a third symbiont, ‘ Candidatus Adiutrix intracellularis', in addition to an Endomicrobium endosymbiont and the Desulfovibrio ectosymbiont ( Ikeda-Ohtsubo et al. , 2016 ). ‘ Ca . Adiutrix intracellularis' is an intracellular symbiont, which belongs to the ‘Rs-K70 group', a deeply branching clade in the class Deltaproteobacteria . A draft genome analysis of this bacterium suggested that it produces acetate from H 2 and CO 2 ( Ikeda-Ohtsubo et al. , 2016 ). Transcription of the gene coding for a key enzyme, hydrogenase-linked formate dehydrogenase, was detected in situ ( Rosenthal et al. , 2013 ). Interrelationships among these three bacterial symbionts are unknown. Hydrogen-oxidizing activity has been experimentally demonstrated in bacterial symbionts of other termite-gut protists. Eucomonympha sp. in the gut of H. sjostedti harbours an intracellular rod-shaped bacterium, ‘ Candidatus Treponema intracellularis' (order Spirochaetales ), which also produces acetate from H 2 and CO 2 ( Ohkuma et al. , 2015 ). ‘ Candidatus Azobacteroides pseudotrichonymphae' (order Bacteroidales ), an intracellular symbiont of the protist Pseudotrichonympha grassii in the gut of the termite Coptotermes formosanus , exhibited a strong H 2 -uptake activity ( Inoue et al. , 2007 ; Hongoh et al. , 2008b ). The ability to synthesize amino acids and cofactors is, in part, complementary between ‘ Ca . Desulfovibrio trichonymphae' and ‘ Ca . Endomicrobium trichonymphae' ( Supplementary Tables S4 and S5 ). Given that the termite host feeds on nitrogen-poor wood materials, biosynthesis and provision of essential nitrogenous compounds are critically important for the termite, and probably also for the protists ( Odelson and Breznak, 1985 ). In addition, provision of cobalamin from ‘ Ca . Desulfovibrio trichonymphae' to ‘ Ca . Endomicrobium trichonymphae' should be important because the latter has a cobalamin-dependent methionine synthase and a cobalamin-dependent ribonucleotide reductase ( Hongoh et al. , 2008a ). The synthesized methionine, in turn, can be provided through the host cytoplasm to ‘ Ca . Desulfovibrio trichonymphae', which possesses the transporter MetINQ. Whereas ‘ Ca . Desulfovibrio trichonymphae' probably takes up amino acids, including methionine, also from the gut lumen through the small pore, provision of nitrogen sources from the host cytoplasm without competition seems to be beneficial to the bacterium. Thus, the complementary supply of nitrogenous compounds may be another driving force for the evolution of this symbiosis. It remains unknown how such compounds are released to the host cytoplasm from these symbiotic bacteria; the protist host may digest the symbionts like exogenously supplied bacterial cells ( Odelson and Breznak, 1985 ). Because the gene for ammonium transporter is pseudogenized in ‘ Ca . Endomicrobium trichonymphae' ( Hongoh et al. , 2008a ) and absent in ‘ Ca . Desulfovibrio trichonymphae', amino acids should be their primary nitrogen sources. Interestingly, the aromatic amino acid transporter aroP gene has obviously been laterally transferred between the ancestors of ‘ Ca . Desulfovibrio trichonymphae' and the endosymbiotic endomicrobia. Genes required for the fumarate respiration of ‘ Ca . Desulfovibrio trichonymphae' may have also been laterally acquired from other bacterial lineages ( Supplementary Table S7 ). It has been suggested that ‘ Ca . Adiutrix intracellularis' laterally acquired genes required for its reductive acetogenesis ( Ikeda-Ohtsubo et al. , 2016 ). Thus, LGT plays important roles in the symbiotic system in the termite gut. In conclusion, our analyses of the complete genomes of the two symbionts ‘ Ca . Desulfovibrio trichonymphae' and ‘ Ca . Endomicrobium trichonymphae' unveiled an elaborate mutualism within the protist cell. A schematic tripartite symbiosis is depicted in Figure 5 . The gut protists phagocytose wood particles, hydrolyze the cellulose and hemicellulose to monosaccharides, and ferment the monosaccharides to acetate, CO 2 and H 2 ( Odelson and Breznak, 1985 ; Yamin, 1980 ). A portion of the monosaccharides are imported by ‘ Ca . Endomicrobium trichonymphae' and fermented to acetate, CO 2 , ethanol, and H 2 ( Hongoh et al. , 2008a ). The produced acetate is the main carbon and energy source of the termite host ( Brune, 2014 ). ‘ Ca . Desulfovibrio trichonymphae' takes up the generated H 2 as the energy source and acetate and CO 2 as the carbon sources. Malate, produced during the fermentation process of the protist, enables ‘ Ca . Desulfovibrio trichonymphae' to oxidize H 2 even when sulfate is unavailable. The removal of H 2 should promote the fermentation processes of the protist and ‘ Ca . Endomicrobium trichonymphae'. The protist provides these bacteria with the metabolites and the habitat, and in turn, obtains various amino acids and cofactors. Future studies of the functions of the uncultivable protist host are needed to fill the picture describing this complex symbiotic system."
} | 3,772 |
36133945 | PMC9417631 | pmc | 5,670 | {
"abstract": "Cellulose, the most abundant natural polymer, has good biocompatibility, biodegradability, and non-toxicity, which make it and its derivatives promising candidates for the fabrication of multifunctional materials, while maintaining sustainability and environmental friendliness. The combination of electrospinning technology and cellulose (and its derivatives) provides a feasible approach to produce nanostructured porous materials with promising functionalities, flexibility, renewability and biodegradability. At the same time, it enables value-added applications of cellulose and its derivatives that are derived from nature or even biomass waste. This review summarizes and discusses the latest progress in cellulose-based electrospun nanofibers, including their construction methods and conditions, various available raw materials, and applications in multiple areas (water treatment, biomaterials, sensors, electro-conductive materials, active packaging, and so on), which are followed by the conclusion and prospects associated with future opportunities and challenges in this active research area.",
"conclusion": "4. Conclusion and future prospects Recent progress in cellulose-based electrospun nanofiber production and applications is summarized. In one aspect, research work focused on exploring new cellulose solvents and cellulose derivatives that are suitable for electrospinning and improving electrospinning apparatus for better fiber quality; in another, various applications in water treatment, biomaterials, sensors, electro-conductive materials, active packaging, and so on have been proposed to make full use of these renewable, biodegradable and nontoxic nanofibrous networks with a large surface area and interconnected porous structure. Different forms of cellulose (wood pulp, nanocellulose, etc. ) and derivatives (CA, hydroxypropyl methylcellulose, carboxymethyl cellulose, azido-cellulose, aldehyde cellulose, ethyl cellulose, etc. ) have been employed as raw materials for electrospinning. However, the optimized electrospinning conditions of all these materials have relatively low polymer concentrations in spinning solutions and low flow rates during spinning, resulting in low productivity. Therefore, future research on new solvents and cellulose derivatives is still required to address these issues and improve productivity. It is worth noting that the electrospinning technology improved largely in the past 15 years; for example, advanced electrospinning apparatuses allow the production of nanofabrics with multiple structures (aligned, multi-layered, composite, etc. ). However, current research work mainly focuses on the formulations and electrospinning conditions of nanofibers, and not many novel structures have been reported. Hence, the rational design of electrospun nanofibers with better performance is expected in future research, which should be based on the understanding of the structure–property relationship and linked tightly with the targeted application.",
"introduction": "1. Introduction Electrospinning is a facile and efficient method for preparing fibers with nano-scaled diameters, which have a large specific surface area, good interconnectivity, and structural stability. Meanwhile owing to the random arrangement of nanofibers prepared by electrospinning, a large number of isotropic pores evenly distribute among fibers. 1,2 These characteristics make electrospun nanofibers useful in a wide range of applications, including filtration & adsorption materials, “smart” materials, catalytic materials, and so on. 3–7 However, the existing nanofibers fabricated by electrospinning are usually based on non-degradable polymers, which may cause environmental issues after disposal. As the most abundant biodegradable polymer on the earth, cellulose possesses many fascinating properties, such as biocompatibility, environmental friendliness, and inexhaustible renewability. 8 Taking advantage of the reactive, numerous, and regularly arranged hydroxyl groups along cellulose chains, a wide variety of cellulose derivatives have been developed by simple reactions between functional substituents and active hydroxyls. 9 The derivatives possess interesting properties inherited from both cellulose and modified functional groups, including good solubility and processability, excellent flexibility, high mechanical strength, and so on. 10,11 Hence, the combination of electrospinning technology and cellulose (and its derivatives) provides a feasible approach to produce nanostructured porous materials with desirable properties. At the same time, it enables value-added applications of cellulose and its derivatives that are derived from nature or even biomass waste. Overviewing the research work published in last 3 years (2018–2021), various electrospun nanofibers have been successfully prepared from cellulose and cellulose derivatives in different solvent systems. 12–17 These nanofibers demonstrated promising applications in the fields of filtration, tissue engineering and biomedical engineering. 18–22 According to SciFinder, 17 literature reviews with the keywords “cellulose” and “electrospin” have been published during 2018–2021. It is certain that this is an active research area. However, all these reviews focus on either a single compound ( e.g. , cellulose acetate and nanocellulose)/single solvent system ( e.g. , ionic liquids) or a single application ( e.g. , supercapacitors, adsorbents, and wound healing materials). Therefore, this review will provide the researchers with a holistic view of recently reported cellulose-based electrospun nanofibers, including their raw materials (cellulose, cellulose derivatives, and nanocellulose), construction methods (electrospinning solvents and conditions), and applications (water treatment, biomaterials, sensors, electro-conductive materials, active packaging, etc. )."
} | 1,467 |
28333583 | PMC5467749 | pmc | 5,672 | {
"abstract": "To efficiently learn from feedback, cortical networks need to update\nsynaptic weights on multiple levels of cortical hierarchy. An effective and\nwell-known algorithm for computing such changes in synaptic weights is the error\nbackpropagation algorithm. However, in this algorithm, the change in synaptic\nweights is a complex function of weights and activities of neurons not directly\nconnected with the synapse being modified, whereas the changes in biological\nsynapses are determined only by the activity of presynaptic and postsynaptic\nneurons. Several models have been proposed that approximate the backpropagation\nalgorithm with local synaptic plasticity, but these models require complex\nexternal control over the network or relatively complex plasticity rules. Here\nwe show that a network developed in the predictive coding framework can\nefficiently perform supervised learning fully autonomously, employing only\nsimple local Hebbian plasticity. Furthermore, for certain parameters, the weight\nchange in the predictive coding model converges to that of the backpropagation\nalgorithm. This suggests that it is possible for cortical networks with simple\nHebbian synaptic plasticity to implement efficient learning algorithms in which\nsynapses in areas on multiple levels of hierarchy are modified to minimize the\nerror on the output.",
"introduction": "1 Introduction Efficiently learning from feedback often requires changes in synaptic weights\nin many cortical areas. For example, when a child learns sounds associated with\nletters, after receiving feedback from a parent, the synaptic weights need to be\nmodified not only in auditory areas but also in associative and visual areas. An\neffective algorithm for supervised learning of desired associations between inputs\nand outputs in networks with hierarchical organization is the error backpropagation\nalgorithm ( Rumelhart, Hinton, & Williams,\n1986 ). Artificial neural networks (ANNs) employing backpropagation have\nbeen used extensively in machine learning ( LeCun et\nal., 1989 ; Chauvin & Rumelhart,\n1995 ; Bogacz, Markowska-Kaczmar, &\nKozik, 1999 ) and have become particularly popular recently, with the\nnewer deep networks having some spectacular results, now able to equal and\noutperform humans in many tasks ( Krizhevsky,\nSutskever, & Hinton, 2012 ; Hinton\net al., 2012 ). Furthermore, models employing the backpropagation\nalgorithm have been successfully used to describe learning in the real brain during\nvarious cognitive tasks ( Seidenberg &\nMcClelland, 1989 ; McClelland, McNaughton,\n& O’Reilly, 1995 ; Plaut,\nMcClelland, Seidenberg, & Patterson, 1996 ). However, it has not been known if natural neural networks could employ an\nalgorithm analogous to the backpropagation used in ANNs. In ANNs, the change in each\nsynaptic weight during learning is calculated by a computer as a complex, global\nfunction of activities and weights of many neurons (often not connected with the\nsynapse being modified). In the brain, however, the network must perform its\nlearning algorithm locally, on its own without external influence, and the change in\neach synaptic weight must depend on just the activity of presynaptic and\npostsynaptic neurons. This led to a common view of the biological implausibility of\nthis algorithm ( Crick, 1989 )—for\nexample: “Despite the apparent simplicity and elegance of the\nback-propagation learning rule, it seems quite implausible that something like\nequations […] are computed in the cortex” ( O’Reilly & Munakata, 2000 , p. 162). Several researchers aimed at developing biologically plausible algorithms for\nsupervised learning in multilayer neural networks. However, the biological\nplausibility was understood in different ways by different researchers. Thus, to\nhelp evaluate the existing models, we define the criteria we wish a learning model\nto satisfy, and we consider the existing models within these criteria: Local computation. A neuron performs computation\nonly on the basis of the inputs it receives from other neurons weighted\nby the strengths of its synaptic connections. Local plasticity. The amount of synaptic weight\nmodification is dependent on only the activity of the two neurons the\nsynapse connects (and possibly a neuromodulator). Minimal external control. The neurons perform\nthe computation autonomously with as little external control routing\ninformation in different ways at different times as possible. Plausible architecture. The connectivity\npatterns in the model should be consistent with basic constraints of\nconnectivity in neocortex. \n The models proposed for supervised learning in biological multilayer neural\nnetworks can be divided in two classes. Models in the first class assume that\nneurons ( Barto & Jordan, 1987 ; Mazzoni, Andersen, & Jordan, 1991 ; Williams, 1992 ) or synapses ( Unnikrishnan & Venugopal, 1994 ; Seung, 2003 ) behave stochastically and receive\na global signal describing the error on the output (e.g., via a neuromodulator). If\nthe error is reduced, the weights are modified to make the produced activity more\nlikely. Many of these models satisfy the above criteria, but they do not directly\napproximate the backpropagation algorithm, and it has been pointed out that under\ncertain conditions, their learning is slow and scales poorly with network size\n( Werfel, Xiew, & Seung, 2005 ). The\nmodels in the second class explicitly approximate the backpropagation algorithm\n( O’Reilly, 1998 ; Lillicrap, Cownden, Tweed, & Akerman,\n2016 ; Balduzzi, Vanchinathan, &\nBuhmann, 2014 ; Bengio, 2014 ; Bengio, Lee, Bornschein, & Lin, 2015 ;\n Scellier & Bengio, 2016 ), and we\nwill compare them in detail in section 4 . Here we show how the backpropagation algorithm can be closely approximated\nin a model that uses a simple local Hebbian plasticity rule. The model we propose is\ninspired by the predictive coding framework ( Rao\n& Ballard, 1999 ; Friston,\n2003 , 2005 ). This framework is\nrelated to the autoencoder framework ( Ackley, Hinton,\n& Sejnowski, 1985 ; Hinton &\nMcClelland, 1988 ; Dayan, Hinton, Neal,\n& Zemel, 1995 ) in which the GeneRec model ( O’Reilly, 1998 ) and another approximation of\nbackpropagation ( Bengio, 2014 ; Bengio et al., 2015 ) were developed. In both\nframeworks, the networks include feedforward and feedback connections between nodes\non different levels of hierarchy and learn to predict activity on lower levels from\nthe representation on the higher levels. The predictive coding framework describes a\nnetwork architecture in which such learning has a particularly simple neural\nimplementation. The distinguishing feature of the predictive coding models is that\nthey include additional nodes encoding the difference between the activity on a\ngiven level and that predicted by the higher level, and that these prediction errors\nare propagated through the network ( Rao &\nBallard, 1999 ; Friston, 2005 ).\nPatterns of neural activity similar to such prediction errors have been observed\nduring perceptual decision tasks ( Summerfield et\nal., 2006 ; Summerfield, Trittschuh,\nMonti, Mesulam, & Egner, 2008 ). In this letter, we show that when\nthe predictive coding model is used for supervised learning, the prediction error\nnodes have activity very similar to the error terms in the backpropagation\nalgorithm. Therefore, the weight changes required by the backpropagation algorithm\ncan be closely approximated with simple Hebbian plasticity of connections in the\npredictive coding networks. In the next section, we review backpropagation in ANNs. Then we describe a\nnetwork inspired by the predictive coding model in which the weight update rules\napproximate those of conventional backpropagation. We point out that for certain\narchitectures and parameters, learning in the proposed model converges to the\nbackpropagation algorithm. We compare the performance of the proposed model and the\nANN. Furthermore, we characterize the performance of the predictive coding model in\nsupervised learning for other architectures and parameters and highlight that it\nallows learning bidirectional associations between inputs and outputs. Finally, we\ndiscuss the relationship of this model to previous work.",
"discussion": "4 Discussion In this letter, we have proposed how the predictive coding models can be\nused for supervised learning. We showed that they perform the same computation as\nANNs in the prediction mode, and weight modification in the learning mode has a\nsimilar form as for the backpropagation algorithm. Furthermore, in the limit of\nparameters describing the noise in the layer where output training samples are\nprovided, the learning rule in the predictive coding model converges to that for the\nbackpropagation algorithm. 4.1 Biological Plausibility of the Predictive Coding Model In this section we discuss various aspects of the predictive coding\nmodel that require consideration or future work to demonstrate the biological\nplausibility of the model. In the first model we presented (see section 2.2 ) and in the simulations of handwritten digit\nrecognition, the inputs and outputs corresponded to layers different from the\ntraditional predictive coding model ( Rao\n& Ballard, 1999 ), where the sensory inputs are presented to\nlayer l = 0 while the higher layers extract underlying\nfeatures. However, supervised learning in a biological context would often\ninvolve presenting the stimuli to be associated (e.g., image of a letter, and a\nsound) to sensory neurons in different modalities and thus would involve the\nnetwork from “input modality” via the higher associative cortex to\nthe “output modality.” We focused in this letter on analyzing a\npart of this network from the higher associative cortex to the output modality,\nand thus we presented s out to nodes at layer\n l = 0. We did this only for this case because it is easy to\nshow analytically the relationship between predictive coding and ANNs.\nNevertheless, we would expect the predictive coding network to also perform\nsupervised learning when s in is presented to layer\n0, while s out to layer\n l max , because the model minimizes the errors\nbetween predictions of adjacent levels so it learns the relationships between\nthe variables on adjacent levels. It would be an interesting direction for\nfuture work to compare the performance of the predictive coding networks with\ninput and outputs presented to different layers. In section 3.3 , we briefly\nconsidered a more realistic architecture involving both modalities represented\non the lowest-level layers. Such an architecture would allow for a combination\nof supervised and unsupervised learning. If one no longer has a flat prior on\nthe hidden node but a gaussian prior (so as to specify a generative model), then\neach arm could be trained separately in an unsupervised manner, while the whole\nnetwork could also be trained together. Consider now that the input to one of\nthe arms is an image and the input at the other arm is the classification. It\nwould be interesting to investigate if the image arm could be pretrained\nseparately in an unsupervised manner alone and if this would speed up learning\nof the classification. We now consider the model in the context of the plausibility criteria\nstated in section 1 . The first two criteria\nof local computation and plasticity are naturally satisfied in a linear version\nof the model (with f ( x ) =\n x ), and we discussed possible neural implementation of\nnonlinearities in the model (see Figure 3 ).\nIn that implementation, some of the neurons have a linear activation curve (like\nthe value node x 1 ( 2 ) in Figure\n3 ) and others are nonlinear (like the node f ( x 1 ( 2 ) ) ), which is consistent with the variability of\nthe firing-input relationship (or f-I curve) observed in biological neurons\n( Bogacz, Moraud, Abdi, Magill, &\nBaufreton, 2016 ). The third criterion of minimal external control is also satisfied by\nthe model, as it performs computations autonomously given input and outputs. The\nmodel can also autonomously “recognize” when the weights should be\nupdated, because this should happen once the nodes converged to an equilibrium\nand have stable activity. This simple rule would result in weight update in the\nlearning mode, but no weight change in the prediction mode, because then the\nprediction error nodes have activity equal to 0, so the weight change (see equation 2.19 ) is also 0.\nNevertheless, without a global control signal, each synapse could detect only if\nthe two neurons it connects have converged. It will be important to investigate\nif such a local decision of convergence is sufficient for good learning. The fourth criterion of plausible architecture is more challenging for\nthe predictive coding model. First, the model includes special one-to-one\nconnections between variable nodes ( x i ( l ) ) and the corresponding prediction error nodes\n ( ε i ( l ) ) , while there is no evidence for such special\npairing of neurons in the cortex. It would be interesting to investigate if the\npredictive coding model would still work if these one-to-one connections were\nreplaced by distributed ones. Second, the mathematical formulation of the\npredictive coding model requires symmetric weights in the recurrent network,\nwhile there is no evidence for such a strong symmetry in cortex. However, our\npreliminary simulations suggest that symmetric weights are not necessary for\ngood performance of predictive coding network (as we will discuss in a\nforthcoming paper). Third, the error nodes can be either positive or negative,\nwhile biological neurons cannot have negative activity. Since the error neurons\nare linear neurons and we know that rectified linear neurons exist in biology\n( Bogacz et al., 2016 ), a possible way\nwe can approximate a purely linear neuron in the model with a biological\nrectified linear neuron is if we associate zero activity in the model with the\nbaseline firing rate of a biological neuron. Nevertheless, such an approximation\nwould require the neurons to have a high average firing rate, so that they\nrarely produce a firing rate close to 0, and thus rarely become nonlinear.\nAlthough the interneurons in the cortex often have higher average firing rates,\nthe pyramidal neurons typically do not ( Mizuseki\n& Buzsáki, 2013 ). It will be important to map the nodes\nin the model on specific populations in the cortex and test if the model can\nperform efficient computation with realistic assumptions about the mean firing\nrates of biological neurons. Nevertheless, predictive coding is an appealing framework for modeling\ncortical networks, as it naturally describes a hierarchical organization\nconsistent with those of cortical areas ( Friston, 2003 ). Furthermore, responses of some cortical neurons\nresemble those of prediction error nodes, as they show a decrease in response to\nrepeated stimuli ( Brown & Aggleton,\n2001 ; Miller & Desimone,\n1993 ) and an increase in activity to unlikely stimuli ( Bell, Summerfield, Morin, Malecek, &\nUngerleider, 2016 ). Additionally, neurons recently reported in the\nprimary visual cortex respond to a mismatch between actual and predicted visual\ninput ( Fiser et al., 2016 ; Zmarz & Keller, 2016 ). 4.2 Does the Brain Implement Backprop? This letter shows that a predictive coding network converges to\nbackpropagation in a certain limit of parameters. However, it is important to\nadd that this convergence is more of a theoretical result, as it occurs in a\nlimit where the activity of error nodes becomes close to 0. Thus, it is unclear\nif real neurons encoding information in spikes could reliably encode the\nprediction error. Nevertheless, the conditions under which the predictive coding\nmodel converges to the backpropagation algorithm are theoretically useful, as\nthey provide alternate probabilistic interpretations of the backpropagation\nalgorithm. This allows a comparison of the assumptions made by the\nbackpropagation algorithm with the probabilistic structure of learning tasks and\nquestions whether setting the parameters of the predictive coding models to\nthose approximating backpropagation is the most suitable choice for solving\nreal-world problems that animals face. First, the predictive coding model corresponding to backpropagation\nassumes that output samples are generated from a probabilistic model with\nmultiple layers of random variables, but most of the noise is added only at the\nlevel of output samples (i.e., Σ i ( 0 ) > > Σ i ( l > 0 ) ). By contrast, probabilistic models\ncorresponding to most of real-world data sets have variability entering on\nmultiple levels. For example, if we consider classification of images of\nletters, the variability is present in both high-level features like length or\nangle of individual strokes and low-level features like the colors of\npixels. Second, the predictive coding model corresponding to backpropagation\nassumes a layered structure of the probabilistic model. By contrast,\nprobabilistic models corresponding to many problems may have other structures.\nFor example, in the task from section 1 of\na child learning the sounds of the letters, the noise or variability is present\nin both the visual and auditory stimuli. Thus, this task could be described by a\nprobabilistic model including a higher-level variable corresponding to a letter,\nwhich determines both the mean visual input perceived by a child and the sound\nmade by the parent. Thus, the predictive coding networks with parameters that do\nnot implement the backpropagation algorithm exactly may be more suited for\nsolving the learning tasks that animals and humans face. In summary, the analysis suggests that it is unlikely that brain\nnetworks implement the backpropagation algorithm exactly. Instead, it seems more\nprobable that cortical networks perform computations similar to those of a\npredictive coding network without any variance parameters dominating any others.\nThese networks would be able to learn relationships between modalities in both\ndirections and flexibly learn probabilistic models well describing observed\nstimuli and the associations between them. 4.3 Previous Work on Approximation of the Backpropagation Algorithm As we mentioned in section 1 ,\nother models have been developed describing how the backpropagation algorithm\ncould be approximated in a biological neural network. We now review these\nmodels, relate them to the four criteria stated in section 1 , and compare them with the predictive coding model. O’Reilly (1998) \nconsidered a modified ANN that also includes feedback weights between layers\nthat are equal to feedforward weights. In this modified ANN, the output of\nhidden nodes in the equilibrium is given by (4.1) o i ( l ) = f ( ∑ j = 1 n ( l + 1 ) w i , j ( l + 1 ) o j ( l + 1 ) + ∑ j = 1 n ( l − 1 ) w j , i ( l ) o j ( l − 1 ) ) , and the output of the output nodes satisfies in\nequilibrium the same condition as for the standard ANN (an equation similar to\nthe one above but including just the first summation). It has been demonstrated\nthat the weight change minimizing the error of this network can be well\napproximated by the following update ( O’Reilly, 1998 ): (4.2) Δ w i , j ( l ) ∼ o i ( l − 1 ) , t r a i n o j ( l ) , t r a i n − o i ( l − 1 ) , p r e d o i ( l ) , p r e d . This is the contrastive Hebbian learning weight update rule ( Ackley et al., 1985 ). In equation 4.2 ,\n o j ( l ) , p r e d denotes the output of the nodes in the\nprediction phase, when the input nodes are set to o j ( l max ) = s j i n and all the other nodes are updated as\ndescribed above, while o j ( l ) , t r a i n denotes the output in the training phase when,\nin addition, the output nodes are set to y j ( 0 ) = s j o u t and the hidden nodes satisfy equation 4.1 . Thus, according to\nthe plasticity rule, each synapse needs to be updated twice—once after\nthe network settles to equilibrium during prediction and once after the network\nsettles following the presentation of the desired output sample. Each of these\ntwo updates relies just on local plasticity, but they have the opposite sign.\nThus, the synapses on all levels of hierarchy need “to be aware”\nof the presence of s out on the output and use\nHebbian or anti-Hebbian plasticity accordingly. Although it has been proposed\nhow such plasticity could be implemented ( O’Reilly, 1998 ), it is not known if cortical synapses can\nperform such form of plasticity. In the above GeneRec model, the error terms δ \nare not explicitly represented in neural activity, and instead the weight change\nbased on errors is decomposed into a difference of two weight modifications: one\nbased on target value and one based on predicted value. By contrast, the\npredictive coding model includes additional nodes explicitly representing error\nand, thanks to them, has a simpler plasticity rule involving just a single\nHebbian modification. A potential advantage of such a single modification is\nrobustness to uncertainty about the presence of s out \nbecause no mistaken weight updates can be made when\n s out is not present. Bengio and colleagues ( Bengio,\n2014 ; Bengio et al., 2015 )\nconsidered how the backpropagation algorithm can be approximated in a\nhierarchical network of autoencoders that learn to predict their own inputs. The\ngeneral frameworks of autoencoders and predictive coding are closely related, as\nboth of the networks, which include feedforward and feedback connections, learn\nto predict activity on lower levels from the representation on the higher\nlevels. This work ( Bengio, 2014 ; Bengio et al., 2015 ) includes many\ninteresting results, such as improvement of learning due to the addition of\nnoise to the system. However, it was not described how it is mapped on a network\nof simple nodes performing local computation. There is a discussion of a\npossible plasticity rule at the end of Bengio\n(2014) that has a similar form as equation 4.2 of the GeneRec model. Bengio and colleagues ( Scellier\n& Bengio, 2016 ; Bengio &\nFischer, 2015 ) introduce another interesting approximation to\nimplement backpropagation in biological neural networks. It has some\nsimilarities to the model presented here in that it minimizes an energy\nfunction. However, like contrastive Hebbian learning, it operates in two phases,\na positive and a negative phase, where weights are updated from information\nobtained from each phase. The weights are changed following a differential\nequation update starting at the end of the negative phase and until convergence\nof the positive phase. Learning must be inhibited during the negative phase,\nwhich would require a global signal. This model also achieves good results on\nthe MNIST data set. Lillicrap et al. (2016) focused\non addressing the requirement of the backpropagation algorithm that the error\nterms need to be transmitted backward through exactly the same weights that are\nused to transmit information feedforward. Remarkably, they have shown that even\nif random weights are used to transmit the errors backward, the model can still\nlearn efficiently. Their model requires external control over nodes to route\ninformation differentially during training and testing. Furthermore, we note\nthat the requirement of symmetric weights between the layers can be enforced by\nusing symmetric learning rules like those proposed in GeneRec and predictive\ncoding models. Equally, we will show in a future paper that the symmetric\nrequirement is not actually necessary in the predictive coding model. Balduzzi et al. (2014) showed\nthat efficient learning may be achieved by a network that receives a global\nerror signal and in which synaptic weight modification depends jointly on the\nerror and the terms describing the influence of each neuron of final error.\nHowever, it is not specified in this work how these influence terms could be\ncomputed in a way satisfying the criteria stated in section 1 . Finally, it is worth pointing out that previous papers have shown that\ncertain models perform similar computations as ANNs or that they approximate the\nbackpropagation algorithm, while in this letter, we show, for the first time,\nthat a biologically plausible algorithm may actually converge to\nbackpropagation. Although this convergence in the limit is more of a theoretical\nresult, it provides a mean to clarify the computational relationship between the\nproposed model and backpropagation, as described above. 4.4 Relationship to Experimental Data We hope that the proposed extension of the predictive coding framework\nto supervised learning will make it easier to test this framework\nexperimentally. The model predicts that in a supervised learning task, like\nlearning sounds associated with shapes, the activity after feedback,\nproportional to the error made by a participant, should be seen not only in\nauditory areas but also visual and associative areas. In such experiments, the\nmodel can be used to estimate prediction errors, and one could analyze precisely\nwhich cortical regions or layers have activity correlated with model variables.\nInspection of the neural activity could in turn refine the predictive coding\nmodels, so they better reflect information processing in cortical circuits. The proposed predictive coding models are still quite abstract, and it\nis important to investigate if different linear or nonlinear nodes can be mapped\non particular anatomically defined neurons within a cortical microcircuit ( Bastos et al., 2012 ). Iterative refinements\nof such mapping on the basis of experimental data (such as f-I curves of these\nneurons, their connectivity, and activity during learning tasks) may help\nunderstand how supervised and unsupervised learning is implemented in the\ncortex. Predictive coding has been proposed as a general framework for\ndescribing computations in the neocortex ( Friston, 2010 ). It has been shown in the past how networks in the\npredictive coding framework can perform unsupervised learning, attentional\nmodulations, and action selection ( Rao &\nBallard, 1999 ; Feldman &\nFriston, 2010 ; Friston, Daunizeau,\nKilner, & Kiebel, 2010 ). Here we add to this list supervised\nlearning, and associative memory (as the networks presented here are able to\nassociate patterns of neural activity with each other). It is remarkable that\nthe same basic network structure can perform this variety of the computational\ntasks, also performed by the neocortex. Furthermore, this network structure can\nbe optimized for different tasks by modifying proportions of synapses among\ndifferent neurons. For example, the networks considered here for supervised\nlearning did not include connections encoding covariance of random variables,\nwhich are useful for certain unsupervised learning tasks ( Bogacz, 2017 ). These properties of the predictive coding\nnetworks parallel the organization of the neocortex, where the same cortical\nstructure is present in all cortical areas, differing only in proportions and\nproperties of neurons and synapses in different layers."
} | 6,735 |
36792670 | PMC9932165 | pmc | 5,673 | {
"abstract": "The process of natural silk production in the spider major ampullate (Ma) gland endows dragline silk with extraordinary mechanical properties and the potential for biomimetic applications. However, the precise genetic roles of the Ma gland during this process remain unknown. Here, we performed a systematic molecular atlas of dragline silk production through a high-quality genome assembly for the golden orb-weaving spider Trichonephila clavata and a multiomics approach to defining the Ma gland tri-sectional architecture: Tail, Sac, and Duct. We uncovered a hierarchical biosynthesis of spidroins, organic acids, lipids, and chitin in the sectionalized Ma gland dedicated to fine silk constitution. The ordered secretion of spidroins was achieved by the synergetic regulation of epigenetic and ceRNA signatures for genomic group-distributed spidroin genes. Single-cellular and spatial RNA profiling identified ten cell types with partitioned functional division determining the tri-sectional organization of the Ma gland. Convergence analysis and genetic manipulation further validated that this tri-sectional architecture of the silk gland was analogous across Arthropoda and inextricably linked with silk formation. Collectively, our study provides multidimensional data that significantly expand the knowledge of spider dragline silk generation and ultimately benefit innovation in spider-inspired fibers.",
"introduction": "Introduction Spiders (Order Araneae) are abundant generalist arthropod predators including more than fifty thousand extant species 1 , 2 . All spiders produce silks, which are natural high-performance proteinaceous fibers that are crucial for spider survival and reproduction 3 , 4 . Silk production is a fascinating spider-salient trait of particular economic interest, mainly due to the exceptional properties of these fibers, including high tensile strength and toughness, low density, self-powered rotational actuation, and biocompatibility 5 – 7 . Numerous researchers have tried to emulate natural spidroin (the main proteins of spider silk) production and spinning processes for the biomimetic generation of artificial materials with spider silk-like properties 8 – 11 ; however, much of our current understanding of spider silk formation is based on physical and material studies that have provided only a partial picture of its nature 12 , 13 . Thus, the elucidation of the molecular biological mechanisms involved in the natural silk production system will be valuable for gaining an in-depth understanding of spider silk 14 – 17 . Although orb-web spiders have multiple silk-producing glands, the major ampullate (Ma) gland is often used as a model system in silk production research due to its relatively large size and, especially, the impressive properties of its product, dragline silk 18 . Accordingly, most attempts at innovating dragline silk-inspired fibers have generally involved the silk proteins and microenvironment produced by the Ma gland 11 , 12 , 19 . The Ma gland can be divided into three macroscopic segments, the Tail, the Sac, and the Duct, which are characterized by gradients of pH values, ion concentrations, and shear forces 20 – 22 . Liquid silk protein is synthesized and stored at a very high concentration in the Tail and Sac and transformed into insoluble fiber via the Duct 23 , 24 . In this context, recombinant major ampullate spidroins (MaSps) have been constructed to achieve specific physical properties of silk, including strength, extensibility, and stickiness 25 – 27 . Spider silk-constituting elements (SpiCEs), which are nonspidroin proteins, have been utilized to increase the tensile strength in the case of composite silk films 28 , 29 . In addition, a microfluidic device designed to closely simulate natural ionic and pH conditions allowed fibers to be directly pulled from the outlet and then reeled in air, as in natural spinning 30 – 32 . It is becoming apparent that the detailed mechanisms underlying dragline silk production, including the cellular architecture and molecular function of the Ma gland as well as the biocomposition and formation of dragline silk, are fundamental to advanced fiber innovation 11 , 33 . To shed light on these mechanisms, we herein present a high-quality chromosome-scale reference genome for the golden orb-web spider Trichonephila clavata , which exhibits a colorful body and constructs a large and impressive orb web (Fig. 1a ). By multiomic analysis of the Ma gland and dragline silk, we traced the origins of dragline silk components from the Ma gland segments (Tail, Sac, and Duct), elucidated the epigenetic and post-transcriptomic regulatory features of spidroin genes, and built a single-cell spatial architecture of the tri-sectional Ma gland, which provides the first detailed cytological definitions of the spider Ma gland based exclusively on segment-specific, cell-type-specific, space-specific, and dragline silk-related gene expression classification. These results allowed us to generate a comprehensive molecular atlas of natural silk production within the tri-sectional Ma gland, thereby elucidating the generation mechanism of dragline silk. The data were further extended to reveal the convergent evolution of silk production in the silkworm Bombyx mori , a model species from another arthropod group, and exhibited the shared molecular characteristics of the spider and silkworm silk glands. Our multiomics datasets are accessible in SpiderDB ( https://spider.bioinfotoolkits.net ) and will be valuable for future explorations of the evolutionary origins of silk production strategies and the creation of biomimetic spider silks. Fig. 1 Chromosomal-scale genome assembly and full spidroin gene set of T. clavata . a Photograph of T. clavata showing an adult female and an adult male on the golden orb-web (above) and the female and male karyotypes (below). SCS, sex chromosome system. b Circular diagram depicting the genomic landscape of the 13 pseudochromosomes (Chr1 – 13 on an Mb scale). c Twenty-eight T. clavata spidroin genes anchored on chromosomes. d Spidroin gene groups of another orb-web spider, T. antipodiana . The published genomic data of T. antipodiana 35 was analyzed to identify the location information of spidroin genes. e Spidroin gene catalog of six orb-web spider species. f Expression clustering of silk glands (major and minor ampullate (Ma and Mi), flagelliform- (Fl), tubuliform- (Tu), aggregate- (Ag), and aciniform & pyriform (Ac & Py) glands) and venom glands. The pink line shows the closest relationship between the Ma and Mi glands. g Morphology of T. clavata silk glands. Similar results were obtained in three independent experiments and summarized in Source data. h Expression patterns of 28 spidroin genes in different types of silk glands. Source data are provided as a Source Data file.",
"discussion": "Discussion Herein, we report the first chromosome-scale reference genome assembly of T. clavata . This high-quality genome combined with multiomics analyses enabled the elucidation of silk biosynthesis processes in the tri-sectional Ma gland. Our findings led us to hypothesize that the Tail performs the primary function in MaSps secretion, while the Sac, acting as a storage site, secretes a wide range of proteins (MaSps and nonspidroin proteins), and the Duct plays a limited role in protein secretion but a crucial role in protein structural transition 20 , 21 , 24 . We propose a molecular model comprehensively describing the mechanism underlying dragline silk biosynthesis in the T. clavata Ma gland (Fig. 6 ): (i) The Tail, consisting of two ubiquitous cell types, is the major site of the secretion of organic acids and 13 dragline silk proteins constituting the inner layer of dragline silk. The presence of organic acids can affect protein solubility in water by altering conformational states 55 . In the Tail, these 13 dragline silk genes are highly activated, among which the transcripts encoding MaSp -Group 1 proteins are the dominant components. Relatively high CA in this segment, along with mCG methylation, might contribute to this abundant gene expression. (ii) The Sac, consisting of two ubiquitous and four unique cell types, is the major site of the secretion of lipids and 28 dragline silk proteins constituting the middle layer of dragline silk. The lipids act as a coat that wraps around the dragline silk to regulate its water content 55 . These 28 dragline silk genes are also highly activated in the Sac, among which MaSp -Group1, MaSp -Group2, MaSp-like , SpiCE-DS1 , SpiCE-DS2 , SpiCE-DS4 , and SpiCE-DS13 are the dominant components. Relatively high CA and mCG methylation might also contribute to the high gene expression observed in this segment. (iii) The Duct, consisting of one ubiquitous and four unique cell types, is the major site of the secretion of chitin, cuticular proteins, ions (Ca 2+ and H + ), and 13 dragline silk proteins. Chitin and cuticular proteins form the cuticular intima, where shear forces are generated 21 , 56 . The ions are responsible for a multidimensional state of flux, including ion exchange, pH gradient formation, and dehydration 11 , 20 , and the 13 dragline silk proteins constitute the outer layer of dragline silk. The transcription of dragline silk genes is less active in the Duct than in the Tail and Sac, and only 0.76% of the transcripts come from these genes. Relatively low CA in this segment, along with mCG methylation, might contribute to this decrease in expression. Overall, our results not only define the proteomic and metabolic components of dragline silk but also trace their origins from the tri-sectional Ma gland. The numbers of cell types in the Tail, Sac, and Duct are positively correlated with the diversity of their functions. Fig. 6 Molecular basis of tri-sectional dragline silk generation in the T. clavata Ma gland. Orange solid circles represent the genes with an FPKM > 10,000, and orange hollow circles represent the genes with an FPKM < 10,000, CA chromatin accessibility, Me methylation. Adobe Illustrator 2020 was used to create the image. Our results provide genomic clues for the hierarchically ordered biosynthesis of spidroins. We documented that the MaSp1a–c & MaSp2e , MaSp2a–d , and MiSp-a–e genes are distributed in three distinct groups. In addition, we demonstrated that the MaSps within each of these groups exhibited concerted SC and ST expression profiles in the tri-sectional Ma gland. We also identified the group-specific common TF motifs at the epigenetic level and constructed group-specific lncRNA-miRNA-mRNA networks at the ceRNA level. Such results revealed novel structural, expressional, and regulatory characteristics of spidroin genes that have not been reported in other spider genomes. Spidroins are thought to exert important influences on the mechanical properties of dragline silk 16 , 57 , 58 . As the group distributions of spidroin genes have already been identified in the high-quality spider genomes of T. antipodiana 35 and Latrodectus elegans 59 with cataloged spidroin proteins, our data fill an important information gap regarding the arrangement of spidroin genes on whole chromosomes and provide a general entry point for the mechanical differentiation of silk and the further study of spider genome evolution. In addition to the above genomic findings, our results provide cellular clues for the tri-sectional organization and functional division of the Ma gland. While previous studies have shown that there are two or three cell types that secrete spidroins in the Ma gland of orb-web spiders 21 , 60 , the number of cell types in the Ma gland has been a subject of debate. The divergent findings concerning these cell types have suggested that they may vary between species but probably also reflect technological difficulties in the capture of intact cells after sample preparation for cellular and morphological studies 20 , 21 , 61 . Our study provides well-verified scRNA-seq and ST data indicating the existence of ten cell types in the whole Ma gland, thus contributing to answering this contentious question. The fascinating single-cell spatial architecture of the gland further indicates the differentiation trajectory of Ma gland cells during cell state transitions, suggesting that the Sac and Duct are derived from the early cell type present in the Tail via cell differentiation. Our genomic, proteomic, and metabolomic analyses further reveal details about the generation of silk in silk-producing glands. Physical and material studies have shown that liquid silk dope is transformed into insoluble fiber through the combined effects of pH and ion gradients as well as extensional and shear forces as it migrates through the silk gland 3 , 62 . Our work provides biological evidence of this role of the silk gland and further demonstrates high convergence of several molecular functions in the spider Ma gland Duct and the silkworm ASG, including calcium ion binding, chitin binding, signal transduction, and V-type H + -transporting ATPase expression. Such overlaps were silk gland-specific and not identified in other tissue types (hemocyte and ovary) (Supplementary Fig. 30 ). Furthermore, we identified convergent silk components in spider and silkworm silks, including two proteins (mucin-19 and GDH) and two major metabolites (choline and DL-malic acid). These analogous aspects indicated critical processes and components essential to silk formation, which contribute to understanding the evolution of silk spinning and can be used as references for the optimization of Duct-based artificial spinning devices and the design of silk protein solutions in artificial spinning. In conclusion, our current work comprehensively illustrates the biological basis of dragline silk formation in an orb-weaving spider. To our knowledge, this study is the first to reveal the biological mechanism of silk spinning in spiders in such detail. We believe that the chromosome-scale reference genome of the golden orb-web spider and the molecular atlas of the tri-sectional Ma gland presented here will facilitate the understanding of the spider silk-spinning process in spiders and further serve as a powerful platform for evolutionary studies of silk-spinning organisms. More importantly, the significant molecular characteristics revealed by our results and the generated datasets should ultimately pave the way for producing optimized synthetic silks through genetic and biomimetic manipulations."
} | 3,643 |
31737024 | PMC6831723 | pmc | 5,674 | {
"abstract": "The plant economics spectrum proposes that ecological traits are functionally coordinated and adapt along environmental gradients. However, empirical evidence is mixed about whether aboveground and root traits are consistently linked and which environmental factors drive functional responses. Here we measure the strength of relationships between aboveground and root traits, and examine whether community-weighted mean trait values are adapted along gradients of light and soil fertility, based on the seedling censuses of 57 species in a subtropical forest. We found that aboveground traits were good predictors of root traits; specific leaf area, dry matter, nitrogen and phosphorus content were strongly correlated with root tissue density and specific root length. Traits showed patterns of adaptation along the gradients of soil fertility and light; species with fast resource-acquisitive strategies were more strongly associated with high soil phosphorus, potassium, openness, and with low nitrogen, organic matter conditions. This demonstrates the potential to estimate belowground traits from known aboveground traits in seedling communities, and suggests that soil fertility is one of the main factors driving functional responses. Our results extend our understanding of how ecological strategies shape potential responses of plant communities to environmental change.",
"conclusion": "Conclusion The plant economics spectrum helps explain ecological strategies, community assembly, and ecosystem functioning ( Reich, 2014 ), but empirical evidence is still limited to evaluate how broadly applicable these patterns are across diverse systems. Our study provides evidence for a plant economics spectrum in a subtropical seedling community, presents the potential to extrapolate from readily measured aboveground traits to more inaccessible belowground plant functions, and suggests that limitation by soil fertility is the main factor driving the functional response in this forest. We found that both variation in organ morphology and biomass allocation play important roles in responding to local environments. These results provide insights into how plant ecological strategies shape community assembly and responses of plant communities to environmental change. The understanding of plant strategies will continue to improve by combining large scale studies with more species and measured traits (e.g. tree height and production) in the future.",
"introduction": "Introduction Functional traits can be powerful indicators of the ecological strategies of plant species ( Wright et al., 2004 ; Freschet et al., 2010 ; de la Riva et al., 2016 ), and trait-based approaches have provided important insights toward understand what drives the structure of biological communities ( Messier et al., 2017 ; Umana et al., 2017 ). One significant achievement in trait-based ecology is the identification of an economics spectrum of plant traits that helps conceptually organize trade-offs between resource acquisition and conservation ( Wright et al., 2004 ; Chave et al., 2009 ; Weemstra et al., 2016 ). The plant traits associated with the leaf economics spectrum (LES) and wood economics spectrum (WES) tend to be coordinated; there is often a one-dimensional trade-off gradient from resource conservative to resource acquisitive strategies ( Wright et al., 2004 ; Chave et al., 2009 ; Pivovaroff et al., 2014 ). However, in part because root traits are more difficult to collect than leaf and wood traits, the root economics spectrum (RES) is much less well understood than its aboveground counterparts ( Weemstra et al., 2016 ; Bergmann et al., 2017 ). Plant economics spectrum (PES) theory attempts to integrate leaf, stem, and root traits to explain plant ecological strategies ( Reich, 2014 ; Diaz et al., 2016 ). It predicts that (1) the leaf, stem, and root traits are correlated with each other and coordinated across different organs, and (2) functional traits should have adaptive significance along resource gradients. This leads to two inferences: (1) because all organs should be consistently resource-acquisitive or resource-conservative for all resources, readily measured aboveground traits should be useful predictors of root traits that are more difficult to measure; and (2) if traits from different organs are coordinated, then they should response to environmental gradients consistently (e.g., fast-growing, resource-acquisitive species should be more abundant in resource-rich environments, and slow-growing, resource-conservative species should be more abundant in resource-poor environments) ( Reich, 2014 ; Kramer-Walter et al., 2016 ). The hypothesis of trait coordination has general support. For example, plant traits regulating photosynthetic and hydraulic capacity are coordinated, and those traits are also related to anatomical, leaf, and stem traits ( Santiago et al., 2004 ; Aasamaa et al., 2005 ; Taylor and Eamus, 2008 ). Traits are significantly correlated across leaves, stems, and roots in a subarctic flora ( Freschet et al., 2010 ), in Mediterranean forests and shrublands ( de la Riva et al., 2016 ), and in a Mediterranean rangeland ( Perez-Ramos et al., 2012 ). However, some studies did not support consistent coordination between aboveground and belowground traits. For instance, Fortunel et al. (2012) found that root traits were closely aligned with stem traits but not with leaf traits across 758 Neotropical tree species ( Fortunel et al., 2012 ), and Kramer-Walter et al. (2016) reported that although many root, stem, and leaf traits of tree seedlings were coordinated, specific root length was not ( Kramer-Walter et al., 2016 ). Recent studies suggest that because roots face a more complex environment than aboveground organs ( Weemstra et al., 2016 ), root traits are more multidimensional than those of leaves, so a single acquisition-conservation axis cannot adequately capture the variety of belowground functions and tradeoffs that drive differences in plant performance across species ( Kramer-Walter et al., 2016 ; Weemstra et al., 2016 ; Bergmann et al., 2017 ; Shen et al., 2019 ). In addition, because the relationships between leaf and root traits can shift depending on soil properties ( Craine et al., 2005 ), water availability ( Fort et al., 2013 ), temperature ( Geng et al., 2014 ), and other environmental conditions, it is unclear whether aboveground traits can be used reliably to predict belowground plant functions. If the plant economics spectrum is a robust feature of plant communities, traits from different organs should be coordinated within species, and should show adaptive patterns along gradients of associated resources. We would expect, for example, that plants in canopy gaps and areas of fertile soil should have a fast, resource-acquisitive strategy; that is, species should have high specific leaf area (SLA), specific root length (SRL), leaf and root nitrogen (LN, RN), and have traits associated with low tissue density (e.g., low leaf dry matter content, root tissue density) (see Figure 1 for details) ( Poorter et al., 2008 ; Martinez-Vilalta et al., 2010 ; Wright et al., 2010 ; Mommer et al., 2011 ; Hajek et al., 2013 ; Weemstra et al., 2016 ). In contrast, slow-growing species with a resource-conservative strategy (with traits contrasting to those above) should be more abundant in environments with low light and limiting soil nutrients ( Reich, 2014 ; Kramer-Walter et al., 2016 ). However, empirical studies found that these associations were inconsistent. For example, Freschet et al. (2010) reported that resource acquisitive species tended to grow in the soils with high N and litter water content ( Freschet et al., 2010 ); High leaf and litter quality (e.g., high N) were associated with high N availability in soil, while high root quality did not show the same relationships to soil properties ( Orwin et al., 2010 ). Kramer-Walter et al. (2016) found that root traits other than tissue density were not coordinated with variation in aboveground traits along a soil fertility gradient ( Kramer-Walter et al., 2016 ). There is a critical need for further studies that simultaneously consider belowground and aboveground traits along environmental gradients to test the prediction of the plant economics spectrum for coordinated variation in whole-plant traits. Figure 1 Illustration of the expected relationships among soil fertility, canopy openness, aboveground traits, and root traits based on the plant economic spectrum. Solid arrows represent positive effects or correlations, dashed arrows represent negative effects or correlations. Grey arrows indicate the effects of environmental conditions on traits, black arrows indicate the relationships between aboveground and root traits. See Table 1 for trait abbreviations. Table 1 The definition, abbreviation, and unit for 16 functional traits in this study. Trait Abbreviation Definition unit Aboveground trait Leaf area LA Leaf surface area cm 2 Leaf area ratio LAR The ratio of total leaf area to biomass of an individual cm 2 g -1 Leaf carbon content LC Leaf carbon content per mass % Leaf dry matter content LDMC The ratio of leaf dry mass to leaf fresh mass g g -1 Leaf nitrogen content LN Leaf nitrogen content per mass mg g -1 Leaf phosphorus content LP Leaf phosphorus content per mass mg g -1 Specific leaf area SLA Dividing leaf area by dried mass cm 2 g -1 Specific stem length SSL The ratio of stem length to stem dry weight cm g -1 Leaf thickness T Leaf thickness at the widest part of each leaf cm Root trait Fine-root diameter DIAM Average fine root diameter mm Root branching intensity RBI Number of root tips per fine-root length tips cm -1 Root nitrogen content RN Root nitrogen content per mass mg g -1 Root phosphorus content RP Root phosphorus content per mass mg g -1 Root tissue density RTD The ratio of root dry mass to its volume g cm -3 Specific root area SRA The root surface area per dry mass cm 2 g -1 Specific root length SRL The root length divided by its dry mass cm g -1 To address this need, we combined information on woody plant censuses of 17,148 seedlings of 57 common species across 10 years with seedling functional traits of roots, stems, and leaves, to examine the relationships between aboveground and belowground traits, and to examine whether these traits show patterns of adaptation along environmental gradients of light (canopy openness) and soil fertility in a typical subtropical forest in Southern China. Specifically, we addressed the following two questions: 1) How well do aboveground traits of 57 subtropical tree seedling species serve as indicators for corresponding root traits? 2) Do traits have adaptive associations along the gradients of soil fertility and light? Specifically, are seedling species with fast resource-acquisitive traits associated with environments of high light and soil fertility, and resource-conservative seedling species associated with resource-poor local environments?",
"discussion": "Discussion This study demonstrated that aboveground and belowground traits of plants tended to be coordinated, aboveground traits were good predictors of root traits, and traits from different organs showed corresponding patterns of adaptation along the gradients of soil fertility and light. Species that presented resource-acquisition traits were associated with environments with high soil P, soil K, and canopy openness plus low N and organic matter; plant species with traits adapted for resource conservation were instead associated with environments with the inverses of those environmental characteristics ( Figure 5 ). Light environment and soil fertility were the focus of our study because they are considered the most limiting resources in tropical and subtropical forests ( Uriarte et al., 2005 ; Wright et al., 2011 ; Pasquini and Santiago, 2012 ; Santiago et al., 2012 ; Record et al., 2016 ; Han et al., 2017 ). However, other environmental conditions also have the potential to drive functional responses of plants. Aboveground Traits Are Good Predictors of Root Traits One of the main predictions from the plant economics spectrum is that traits from different organs should be coordinated ( Reich, 2014 ); this suggests that root traits should be predictable from known aboveground traits. We found, for 57 species of subtropical tree seedlings, that aboveground and root traits were highly correlated ( Table 2 ), and that the signs of the correlations were congruent across organs (see expected relationships in Figure 1 ), highlighting the existence of a plant economics spectrum for this subtropical seedling community. Previous studies have found similar patterns in other systems: N, P, and dry matter content of roots, leaves, and stems were positively correlated ( Freschet et al., 2010 ; Geng et al., 2014 ), plants with low leaf or root N content tended to also have high leaf and root tissue density ( Craine et al., 2005 ; Perez-Ramos et al., 2012 ), SRL was positively correlated with SLA ( Freschet et al., 2015 ), and RTD and LDMC were positively correlated ( Bergmann et al., 2017 ). However, traits of root diameter (DIAM) and branching intensity (RBI) seem to be independent of the whole-plant economics spectrum, highlighting the multidimensionality of root traits ( Kramer-Walter et al., 2016 ; Weemstra et al., 2016 ). This points to the need to incorporate the complexity of the soil environment, root form and function, and multiple belowground resource uptake strategies (e.g., mycorrhizal pathways) in future studies ( Weemstra et al., 2016 ). We identified several aboveground traits that are most useful as predictors of corresponding root traits and evaluated the strength of the associations ( Table 3 ). Predictive power reached up to 22–40% for PES root traits (excluding DIAM and RBI) only using 1–3 aboveground traits, confirming a moderately useful predictive potential of aboveground traits when direct measurement of root traits is unavailable. Overall, leaf area ratio (LAR) and leaf nitrogen (LN) were the best predictors of PES root traits in this subtropical forest. However, the relationships among traits in different organs may vary across environmental gradients. Craine et al. (2005) found that relationships between leaf and root traits were stronger at regional than global scales, and that environmental factors such as soil properties might affect their relationships ( Craine et al., 2005 ). Others noted that the SLA-SRL relationship could shift from negative to positive with increasing temperature ( Geng et al., 2014 ); LN area /RN length increased and SLA/SRL and LN mass /RN mass decreased from semi-arid to arid environments ( Liu et al., 2010 ). These results demonstrate that uncovering the trait relationships among different plant organs should be linked to environmental gradients, especially for large-scale studies. Limitation by Soil Fertility Is the Main Factor Driving the Functional Response Another fundamental prediction of plant economics spectrum theory is that resource-associated traits from different organs should show adaptive congruence along resource gradients ( Reich, 2014 ). Species with fast resource acquisitive strategy should be more abundant in fertile soil, especially for the key soil nutrients (e.g., N, P, and K concentration) ( Lu et al., 2010 ; Wright et al., 2011 ; Pasquini and Santiago, 2012 ; Santiago et al., 2012 ; Record et al., 2016 ). In our study, we used a 10-year census of 1158 seedling plots to examine this hypothesis. Traits tended to vary predictably along the soil PC1 axis, indicating that limitation by soil fertility was the main ecological factor driving the functional response ( Perez-Ramos et al., 2012 ). Our results showed that nutrient-acquisitive species were more abundant at the low end of the soil PC1 axis ( Figures 3 and 5 ), where soils had high phosphorus and potassium but low nitrogen and organic matter ( Figure 2 ). This contrasts with results from previous studies, where many traits did not show adaptive patterns along resource gradients. For example, fast-growing, nutrient-acquisitive species were unexpectedly more abundant in soils with high N or P ( Freschet et al., 2010 ; Orwin et al., 2010 ); low plant tissue density (e.g., leaf dry matter content or root tissue density) corresponded to soils with high N, but SRA increased with increasing soil C:N ratio ( Perez-Ramos et al., 2012 ). In addition, soil nutrient (N, P, K) availability increased LAR, SLA, but decreased SRL ( Freschet et al., 2015 ); increasing soil P and decreasing C:N ratio were associated with decreasing RTD, RBI and SRL but increasing RN and DIAM ( Kramer-Walter et al., 2016 ). We suggest that multidimensional traits may lead to these mixed results, because some traits fall outside of the plant economics spectrum, implying that these traits are constrained by environmental drivers that are not necessarily related to resource uptake ( Weemstra et al., 2016 ), resulting in weak or inconsistent relationships with environments. However, although plant traits show clear patterns of adaptive matching along the resource gradient represented by the soil PC1 axis, our results do not fully support the expectations of the plant economics spectrum. We note signs of resource conservation in resource-poor environments and resource acquisition in resource-rich environments ( Figure 1 ), a clear trade-off between high nitrogen and organic matter on one end and phosphorus and potassium at the other ( Figure 2 ), and resource-acquisitive species are more abundant in the soil high in P and K but low in N and OM ( Figures 3 and S4 ). Because limiting resources should be the main ecological factors that drive the functional response of plants ( Perez-Ramos et al., 2012 ), our results suggest that P and K may be the most limiting soil nutrients in this subtropical forest, so that species adapted for resource acquisition were more successful in the environments where those nutrients were abundant. Previous studies came to similar conclusions that P was the most limited element in plant growth in old, weathered soils of tropical forests ( Walker and Syers, 1976 ; Condit et al., 2013 ; Liu et al., 2018 ) and that K addition increased photosynthesis rates and height growth of tropical seedlings ( Pasquini and Santiago, 2012 ; Santiago et al., 2012 ). Functional Traits Response to Canopy Openness Besides soil nutrients, seedling response to light could be another important axis that differentiates niches; previous results showed that seedlings are limited by light in the shaded understory of tropical and subtropical forests ( Uriarte et al., 2005 ; Han et al., 2017 ) and variation in ecophysiological traits of trees are expected to help capture the variation in light ( Reich et al., 1998 ; Poorter and Rose, 2005 ; Sanchez-Gomez et al., 2006 ; Coble et al., 2017 ). In our study, however, functional traits showed much weaker response to canopy openness than to soil fertility ( Figure 5 ). Only 5 out of 16 traits showed a significant relationship with canopy openness ( Figure 4 ). These traits did show patterns of adaptation along a light gradient, where species with traits adapted for fast resource acquisition (e.g., high LAR, SSL, SRA, and low RTD) were more abundant in the seedling plots with greater canopy openness, consistent with the PES expectation ( Figure 1 ). There are at least two factors that could lead to the relatively weak response of traits to canopy openness. First, light in most of the seedling plots was consistently very low ( Table S1 ), with >90% of seedling plots <5% canopy openness. There were very few canopy gaps associated with plots in this study, resulting in low spatial heterogeneity of light availability and corresponding difficult to measure trait-based responses among plants ( Record et al., 2016 ). Second, we did not measure temporal heterogeneity of light, since we only measured canopy openness in 2017; unidentified temporal variation in canopy openness could also contribute to the weak relationship between plant traits and the light environment. Biomass Allocation Plays an Important Role in Responding to Environments Both leaf area ratio (LAR) and specific stem length (SSL) are traits involve in biomass allocation, determining how much leaf area or stem length is produced per unit of plant mass ( Poorter et al., 2012 ). Both these traits were associated with gradients of soil fertility and canopy openness ( Figures 3 and 4 ), indicating that plants adapt to different environments not only through organ morphology but through variation in biomass allocation ( Freschet et al., 2015 ). For example, shifts in biomass allocation were important in plant responses to soil nutrients ( Freschet et al., 2015 ). However, some relationships between biomass allocation, traits, and environments in our study were not consistent with previous results and expectations ( Figure 1 ). For example, Freschet et al. (2015) found that increased soil nutrient supply drives an increase in SLA, leaf mass fraction (LMF, leaf dry mass/total plant dry mass), and LAR, but a decrease in SRL, root mass fraction (RMF, root dry mass/total plant dry mass), and root length ratio (RLR, root length/total plant dry mass) ( Freschet et al., 2015 ). While increased light supply drives a decrease in SLA, LMF, and LAR ( Reich et al., 1998 ), it increases RMF. In our study, SLA, SRL, LAR, and SSL all increased with increasing soil P and K and with decreasing N and OM; increasing light availability also increased LAR and SSL. We suggest that these relationships may derive from the specific environments of tropical and subtropical seedlings; where light and soil nutrients were extremely limited in the understory in these regions ( Uriarte et al., 2005 ; Lu et al., 2010 ; Wright et al., 2011 ; Pasquini and Santiago, 2012 ; Santiago et al., 2012 ; Record et al., 2016 ; Han et al., 2017 ), all variation in organ morphology and biomass allocation are coordinated to best use the limited resources. For example, in response to increased availability of P, K, and light, seedling species with a fast resource acquisition strategy not only increased SLA and SRL to enhance the resource acquisition but also increased biomass allocation to leaf area (LAR) and height (SSL) to achieve a relative high resource use efficiency and growth rate."
} | 5,624 |
28649958 | null | s2 | 5,675 | {
"abstract": "Advances in synthetic biology allow us to engineer bacterial collectives with pre-specified characteristics. However, the behavior of these collectives is difficult to understand, as cellular growth and division as well as extra-cellular fluid flow lead to complex, changing arrangements of cells within the population. To rationally engineer and control the behavior of cell collectives we need theoretical and computational tools to understand their emergent spatiotemporal dynamics. Here, we present an agent-based model that allows growing cells to detect and respond to mechanical interactions. Crucially, our model couples the dynamics of cell growth to the cell's environment: Mechanical constraints can affect cellular growth rate and a cell may alter its behavior in response to these constraints. This coupling links the mechanical forces that influence cell growth and emergent behaviors in cell assemblies. We illustrate our approach by showing how mechanical interactions can impact the dynamics of bacterial collectives growing in microfluidic traps."
} | 266 |
32292655 | PMC7144587 | pmc | 5,677 | {
"abstract": "Rhizospheric and endophytic fungi are key factors which influence plant fitness and soil fertility. Atractylodes macrocephala is one of the best-known perennial herbs used in traditional Chinese medicine. Continuous cropping has been shown to have a negative effect on its growth and renders it more susceptible to microbial pathogen attacks. In this study, we investigated the effects of continuous cropping on the endophytic and rhizospheric fungi associated with A . macrocephala using culture-independent Illumina MiSeq. Continuous cropping was found to decrease fungal diversity inside plant roots, stems, leaves and tubers. Additionally, we found that the structure and diversity of rhizospheric and endophytic fungal communities were altered by root-rot disease. Fusarium was overrepresented among root-rot rhizospheric and endophytic fungi, indicating that it has a major negative impact on plant health during A . macrocephala monocropping. Canonical correspondence analysis of the control and diseased samples revealed that pH, hydrolysis N, electrical conductivity and Hg content were well-correlated with fungal community composition during continuous cropping. Taken together, these results highlight the ecological significance of fungal communities in maintaining plant fitness and will guide the development strategies to attenuate the negative impacts of A . macrocephala continuous cropping.",
"conclusion": "Conclusions Overall, we found that continuous cropping and severe root-rot disease could both significantly affect the structure and diversity of A. macrocephala endophytic and soil fungal communities. Rhizospheric soil pH and organic matter content decreased with increasing continuous cropping time. Moreover, the abundance and diversity of fungal communities decreased, while the prevalence of severe root-rot disease increased with prolonged continuous cropping. Of all root-rot rhizospheric and endogenous fungal species, Fusarium was most significantly enriched upon continuous cropping. Further, soil pH, hydrolysis N, electrical conductivity, and Hg were most strongly correlated with fungal community composition. Simultaneously examining both endophytic fungal populations (from surface-sterilized plant tissues) and the rhizospheric samples (including soils from the roots) allowed us to gain a deeper understanding of how continuous cropping alters fungal populations. Our results suggest that changes in fungal diversity can be used to predict disease outbreaks in A. macrocephala continuous cropping systems. The findings of this research can also guide the development of management strategies to improve A. macrocephala production.",
"introduction": "Introduction Plants and rhizospheric soil are colonized by fungal communities that can impact their fitness by influencing nutrient acquisition, causing soil-borne diseases and affecting the activity of plant pathogens ( Tan et al., 2017a ). Many previous studies have shown that studying plant-microorganism interactions can lead to improvements in several agronomic processes, including crop rotation and tillage ( Somenahally et al., 2018 ), pesticide application ( Regar et al., 2019 ), irrigation ( Dang et al., 2019 ), fertilizer application ( Nguyen et al., 2018 ) and continuous cropping ( Ali et al., 2019 ; Xiong et al., 2015 ). Continuous cropping is widely adopted in Chinese agricultural production and is defined as cultivating the same or similar crop species for a long period of time ( Shipton, 1977 ). Studies have shown that continuous cropping leads to compromised growth, yield loss, disease susceptibility and quality deterioration ( Hontoria et al., 2019 ). The detrimental effects of continuous cropping have been demonstrated in a variety of crop species, including Atractylodes macrocephala ( Zheng et al., 2018 ). A. macrocephala is a perennial herb that has been cultivated for over 700 years in temperate and subtropical regions. The tuber of A. macrocephala is commonly referred to as “ Baizhu ”, and has been used to treat cancer, osteoporosis, gastrointestinal dysfunction, obesity and fetal irritability in traditional Chinese medicine in East Asia ( Zhu et al., 2018b ). It has recently been shown that continuous cropping of A. macrocephala can lead to reduced yield and quality ( Zheng et al., 2018 ), possibly due to alterations in soil enzymatic activities, allelochemical substance enrichment, soil microbial community changes or soil-borne pathogen accumulation ( Xiong et al., 2015 ). Change in soil microbiota communities has been singled out as one of the major causes of yield loss during A. macrocephala continuous cropping ( Chen et al., 2014 ; Shi, 2018 ). Root-rot diseases, which are often associated with A. macrocephala continuous cropping, also play a major role in loss of yield and quality during A. macrocephala continuous cropping ( Zheng et al., 2018 ). Most of these diseases are associated with fungal pathogens such as Fusarium oxysporum , Rhizoctonia solani , and Ceratobasidium sp. ( Liu, 2012 ; You et al., 2013 ; Zhang, 2015 ). Rhizospheric fungi play key roles in organic matter decomposition, nutrient cycling and soil fertility maintenance ( Miao et al., 2016 ; Zhou et al., 2017 ). Some rhizospheric fungi have also been associated with pathogen growth inhibition, leading to their application as biocontrol agents ( Venneman et al., 2019 ). Additionally, the overall level of endophytic fungi has been shown to be a good indicator of host plant health ( Zhu et al., 2018a ). Recent studies have proposed that continuous cropping results in imbalances in endophytic and rhizospheric soil fungal community diversity and structure, and rapid accumulation of fungal pathogens ( Liu et al., 2019 ; Qin et al., 2017 ). Thus, a healthy and stable endophytic and rhizospheric fungal community may be essential for maintaining long-term continuous cropping and stable crop yields. Both the mechanism by which continuous cropping affects rhizospheric soil fungal communities and how these changes influence soil productivity are still largely unknown. To date, there has been no culture-independent studies of rhizospheric soil and endophytic fungal community diversity during monocropping of A. macrocephala ( Xiong et al., 2015 ; Zhou & Wu, 2012 ). In this study, we used molecular characterization to examine changes in rhizospheric fungal communities, rotation soils, roots, leaves, stems, and tubers in A. macrocephala grown in continuous cropping fields. We aimed to investigate the link between soil physiochemical properties and the structure, composition and diversity of fungal communities. Additionally, we examined the underlying mechanism by which continuous cropping of A. macrocephala influenced endophytic and rhizospheric fungal communities.",
"discussion": "Discussion Sustainable A. macrocephala production management calls for a deep understanding of how continuous cropping alters the structure and diversity of fungal communities. To gain more insight in this, we assessed the effects of continuous cropping on fungal communities in rhizospheric soil, rotation soil and endophytes of A. macrocephala . We found that the diversity, structure and composition of rhizospheric soil and endophyte fungal communities were greatly affected by continuous cropping. Rhizospheric soil fungal alpha-diversity indices, including Chao1 and Shannon indices, decreased after continuous cropping practices. In addition, fallow soils showed relatively higher overall fungal activity, whereas fungal diversity was similar between continuously cropped soil and the S_CK, suggesting that fungal diversity gradually recovers to unplanted control soil with time of fallowing. Severe root-rot disease has long been known as a major problem in continuous cropping ( Tan et al., 2017a ). Consistently, we detected a relatively higher fungal community diversity in healthy tissues and rhizospheric soils compared to diseased samples. Therefore, the decrease of soil and endophytic fungal diversity may play a role in disease development during A. macrocephala continuous cropping. Rhizospheric soil fungal communities were also strongly influenced by continuous cropping of A. macrocephala . Zygomycota, Ascomycota, and Basidiomycota were the dominant fungal phyla in both continuous cropping and fallow samples. At the genus level, the relative abundance of Talaromyces and Fusarium increased significantly with cropping time. Talaromyces and Fusarium genera fungi contain several potential pathogens, such as Fusarium oxysporum and Talaromyces helices ( Wu et al., 2016 ). Thus, increases in these genera may negatively contribute to A. macrocephala continuous cropping, eventually leading to increased disease pressure. Moreover, the fungal community profile of fallow soil became more similar to the control soil profile as the number of years of fallowing increased, indicating that recovery of a healthy soil profile is possible. Endophytic fungi usually inhabit different plant tissues without harming their hosts. However, we observed reduced abundance of endophytic fungal OTUs in diseased samples, signifying reduced endophytic fungal diversity. Ascomycota, Basidiomycota, and Zygomycota were the top three fungal phyla detected in both healthy and diseased A. macrocephala plants, which was similar to the profile of soil fungi. This pattern fits with earlier publications, which have shown that endophytic fungi are primarily derived from soil fungi that enter the plants via roots, tubers, leaves and stems ( Dai et al., 2010 ; Tan et al., 2017b ). At the genus level, Fusarium and Alternaria abundance significantly increased in the root-rot diseased A. macrocephala samples. This may be due to increases in other root-rot disease associated species, such as F. oxysporum , F. solani , and Alternaria gansuense . Increases in these pathogenic fungi likely caused a decrease in endophytic fungi due to limited availability of space and nutrients ( Zeng, 2016 ; Zhang et al., 2015 ). These findings warrant future studies on the changes in microbial communities, as they can be used as indicators of overall soil and plant health during continuous cropping. A deeper understanding of soil properties during A. macrocephala continuous cropping systems is key to improving soil productivity. Long-term monoculture of A. macrocephala has been reported to reduce organic matter content and soil pH, due to the return of organic material and the application of fertilizer to the soil ( Geng et al., 2015 ). Increased soil hydrolysis N, available P, and available K contents often result from fertilizer application ( Shi, 2018 ). The optimum pH range for A. macrocephala growth is 5.1 to 6.6 ( Zhu et al., 2018b ), and the pH decline seen during continuous cropping may therefore increase the disease susceptibility of A. macrocephala . In agricultural ecosystems, soil microbial communities have a major impact on soil organic matter accumulation and nutrient cycling, which are often used as indicators of soil quality ( Ashworth et al., 2017 ). In our study, soil pH, hydrolysis N, electrical conductivity, and Hg content were most strongly correlated with fungal community structure during continuous cropping of A. macrocephala . Soil pH may directly alter fungal community composition by inhibiting fungal survival and growth, as seen in earlier publications indicating fungal taxa were unable to grow under a certain soil pH ( Zhang et al., 2016 ). In addition, soil communities can affect soil N dynamics, while hydrolysis N has a major impact on the composition of fungal community under continuous cropping. The change in fungal composition may be a result of both fertilization and the observed increase in hydrolysis N cycling ( Thompson & Kao-Kniffin, 2019 ). Agricultural management practices, such as high fertilizer usage and restricted irrigation can immensely influence electrical conductivity variation ( Adviento-Borbe et al., 2006 ; Kim et al., 2016 ). Electrical conductivity is associated with soil salinity and our results suggest it may be an important predictor of fungal community compositions in continuous cropping soils of A. macrocephala . Moreover, Acremonium and Penicillifer show a negative correlation with heavy metal contents in the soil, especially Hg and Cr."
} | 3,097 |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.