From antibiotic to chalkophore: the biology and evolution of SF2768 in Streptomyces.

This study aimed to identify some important virulence factors, including pap, fim, sfa, aer and hly genes, typical of uropathogenic Escherichia coli (UPEC) in isolates collected from diarrheic and urinary tract infected patients in Khartoum State by multiplex polymerase chain reaction (PCR) assay. A total of 100 clinical specimens (50 urine and 50 diarrhea) were collected. Samples were cultured and identified by conventional method. Most study population were females 57/100 (57%); 42 suffering from urinary tract infections (UTIs) and 15 from diarrhea, while males were 43/100 (43%); 8 suffering from UTIs and 35 from diarrhea. Among enrolled subjects, 83 were positive for one or more uropathogenic E. coli virulent genes, while 17 isolates were negative for all genes. The results of multiplex PCR revealed that thirty two (n=32) diarrheal samples and fourteen (n=14) urine samples were aer positive. Thirty three (n=33) urine samples and eight (n=8) diarrheal samples appeared as fim positive. The genes pap and hly were found in 24 and 14 urine samples, respectively and in 9 and 3 diarrheal samples, respectively, while sfa gene was detected only in 15 urine specimens. The study concluded that fim gene was highly prevalent among UTI patients while aer gene was highly prevalent among diarrhea patients. Key words: Uropathogenic Escherichia coli, fimH, aer, pap, sfa, hly, Sudan.

This study aimed to identify some important virulence factors, including pap, fim, sfa, aer and hly genes, typical of uropathogenic Escherichia coli (UPEC) in isolates collected from diarrheic and urinary tract infected patients in Khartoum State by multiplex polymerase chain reaction (PCR) assay. A total of 100 clinical specimens (50 urine and 50 diarrhea) were collected. Samples were cultured and identified by conventional method. Most study population were females 57/100 (57%); 42 suffering from urinary tract infections (UTIs) and 15 from diarrhea, while males were 43/100 (43%); 8 suffering from UTIs and 35 from diarrhea. Among enrolled subjects, 83 were positive for one or more uropathogenic E. coli virulent genes, while 17 isolates were negative for all genes. The results of multiplex PCR revealed that thirty two (n=32) diarrheal samples and fourteen (n=14) urine samples were aer positive. Thirty three (n=33) urine samples and eight (n=8) diarrheal samples appeared as fim positive. The genes pap and hly were found in 24 and 14 urine samples, respectively and in 9 and 3 diarrheal samples, respectively, while sfa gene was detected only in 15 urine specimens. The study concluded that fim gene was highly prevalent among UTI patients while aer gene was highly prevalent among diarrhea patients.   Key words: Uropathogenic Escherichia coli, fimH, aer, pap, sfa, hly, Sudan.

Background: The present study assessed the occurrence and biofilm forming potential of ESBL-producing Escherichia coli isolates recovered from raw salad vegetables collected from 32 retail vegetable shops of Mirzapur district, Uttar Pradesh, India (n=224). Methods: Standard bacteriological culture methods using cefotaxime-supplemented EMB agar were used for initial isolation of cefotaxime-resistant E. coli which were subjected to phenotypic detection of ESBL production by Kirby-Bauer disc diffusion test. Subsequently, all the phenotypically positive ESBL-producing E. coli isolates were subjected to determine their biofilm-producing ability by Congo red agar (CRA) assay and microtiter plate assay (MPA) followed by genotypic confirmation of biofilm production by PCR assay targeted at fim-H, Sfa and papC genes. Additionally, the isolates were genotypically explored for a potential presence of Shiga-like toxin genes (stx1 and stx2). Result: Altogether, 276 cefotaxime-resistant E. coli isolates were recovered, of which, 99.27% (274/276) were observed to be phenotypically positive for ESBL production. The brown, dry and rough (BDAR) and red, dry and rough (RDAR) colony morphotypes were observed by 70.80% (194/274) and 29.19% (80/274) of the isolates, respectively by CRA method revealing their potential to form biofilms. Biofilm formation was evident in 94.16% (258/274) of the isolates by MPA method, of which 79.45% (205/258), 17.05% (44/258) and 3.48% (9/258) were considered as weak, moderate and strong biofilm producers, respectively. The fimH gene was found to be the predominant genetic determinant of biofilm formation which was detected in 77% (211/274) of the tested isolates followed by Sfa gene (11.31%) and papC gene (4.01%). A significant difference (p less than 0.05) was evident between CRA and MPA methods as well as phenotypic and genotypic methods for detection of biofilm formation among the recovered isolates. In addition, significant differences (p less than 0.05) were also observed upon vegetable-wise comparison of ESBL-producing E. coli isolates and the presence of biofilm producing genes. The stx1 and stx2 genes were found to be present in one isolate recovered from a carrot sample and two isolates from coriander samples, respectively. . The findings of this study highlight the widespread occurrence of ESBL-producing E. coli from raw salad vegetables with significant biofilm-forming potential posing a public health risk. The detection of Shiga-like toxin genes in some isolates further underscores the need for stringent surveillance and improved hygiene practices to mitigate contamination risk in fresh produce.

The Gut Microbiome in Health and Disease

Extreme environments, such as Antarctica, select microbial communities that display a range of evolutionary strategies to survive and thrive under harsh environmental conditions. These include a diversity of specialized metabolites, which have the potential to be a source for new natural product discovery. Efforts using (meta)genome mining approaches to identify and understand biosynthetic gene clusters in Antarctica are still scarce, and the extent of their diversity and distribution patterns in the environment have yet to be discovered. Herein, we investigated the biosynthetic gene diversity of the biofilm microbial community of Whalers Bay, Deception Island, in the Antarctic Peninsula and revealed its distribution patterns along spatial and temporal gradients by applying metagenome mining approaches and multivariable analysis. The results showed that the Whalers Bay microbial community harbors a great diversity of biosynthetic gene clusters distributed into seven classes, with terpene being the most abundant. The phyla Proteobacteria and Bacteroidota were the most abundant in the microbial community and contributed significantly to the biosynthetic gene abundances in Whalers Bay. Furthermore, the results highlighted a significant correlation between the distribution of biosynthetic genes and taxonomic diversity, emphasizing the intricate interplay between microbial taxonomy and their potential for specialized metabolite production.IMPORTANCEThis research on antarctic microbial biosynthetic diversity in Whalers Bay, Deception Island, unveils the hidden potential of extreme environments for natural product discovery. By employing metagenomic techniques, the research highlights the extensive diversity of biosynthetic gene clusters and identifies key microbial phyla, Proteobacteria and Bacteroidota, as significant contributors. The correlation between taxonomic diversity and biosynthetic gene distribution underscores the intricate interplay governing specialized metabolite production. These findings are crucial for understanding microbial adaptation in extreme environments and hold significant implications for bioprospecting initiatives. The study opens avenues for discovering novel bioactive compounds with potential applications in medicine and industry, emphasizing the importance of preserving and exploring these polyextreme ecosystems to advance biotechnological and pharmaceutical research.

Slow sand filters (SSFs) are widely applied to treat potable water; the removal of contaminants (e.g., particles, organic matter, and microorganism) occurs primarily in the top layer. However, the development of the microbial community and its metabolic function is still poorly understood. In the present study, we analyzed the microbial quantity and community of the influents sampled from the effluent of the last step (rapid sand filtration) and of the top layers of SSFs (Schmutzdecke, 0–2 cm, 4–6 cm, 8–10 cm) sampled near terminal head loss when the Schmutzdecke (SCM) was most developed in two full-scale drinking water treatment plants (DWTPs). The two DWTPs use the same artificially recharged groundwater source. The biomass in the filter, quantified by flow cytometric intact cell counts (ICC) and adenosine triphosphate (ATP), decreased rapidly along the depth till 8–10 cm (>1 log TCC; >75% ATP); the decrease was most pronounced from the SCM to the surface sand layer (0–2 cm), after which the biomass stabilized quickly at lower depths (2–10 cm). Remarkably, beta diversity showed that SSFs layers of the same depth in two DWTPs with distinctive filter age and plant location clustered together, which indicated their insignificant effects in shaping microbial communities in SSFs. The alpha diversity indices followed the trend of the biomass, suggesting more active and diverse communities in SCM layer. PICRUSt-based function prediction revealed significant over-representation of metabolism and degradation of complex organic matters (e.g., butanoate, propanoate, xenobiotic, D-Alanine, chloroalkene, and bisphenol) in SCM layer, the functional importance of which was confirmed by the co-occurrence patterns of the dominant taxa and metabolic functions. Using an island biogeography model, we found that microbial communities in SSFs were strongly assembled by selection (68 OTUs, 50.0% sequences), rather than by simple accumulation of the microbial communities in the influents (120 OTUs, 44.8% sequences). Our findings enhance the understanding of microbial community assembly and of metabolic function in the top layers of SSFs, and constitute a valuable contribution to optimizing the design and operation of biofilters in full-scale DWTPs.

Multi-approach assessment of groundwater biogeochemistry: Implications for the site characterization of prospective spent nuclear fuel repository sites

BackgroundThe interaction mechanism between microbial communities and environment is a key issue in microbial ecology. Microbial communities usually change significantly under environmental stress, which has been studied both phylogenetically and functionally, however which method is more effective in assessing the relationship between microbial communities shift and environmental changes still remains controversial.ResultsBy comparing the microbial taxonomic and functional shift pattern along heavy metal contamination gradient, we found that both sedimentary composition and function shifted significantly along contamination gradient. For example, the relative abundance of Geobacter and Fusibacter decreased along contamination gradient (from high to low), while Janthinobacterium and Arthrobacter increased their abundances. Most genes involved in heavy metal resistance (e.g., metc, aoxb and mer) showed higher intensity in sites with higher concentration of heavy metals. Comparing the two shift patterns, there were correlations between them, because functional and phylogenetic β-diversities were significantly correlated, and many heavy metal resistance genes were derived from Geobacter, explaining their high abundance in heavily contaminated sites. However, there was a stronger link between functional composition and environmental drivers, while stochasticity played an important role in formation and succession of phylogenetic composition demonstrated by null model test.ConclusionsOverall our research suggested that the responses of functional traits depended more on environmental changes, while stochasticity played an important role in formation and succession of phylogenetic composition for microbial communities. So profiling microbial functional composition seems more appropriate to study the relationship between microbial communities and environment, as well as explore the adaptation and remediation mechanism of microbial communities to heavy metal contamination.Electronic supplementary materialThe online version of this article (doi:10.1186/s12866-016-0731-6) contains supplementary material, which is available to authorized users.

In the present study, the complete genome of a filamentous fungus Aspergillus terreus ATCC 20542 was sequenced, assembled, and annotated. This strain is mainly recognized for being a model wild-type lovastatin producer and a parental strain of high-yielding industrial mutants. It is also a microorganism with a rich repertoire of secondary metabolites that has been a subject of numerous bioprocess-related studies. In terms of continuity, the genomic sequence provided in this work is of the highest quality among all the publicly available genomes of A. terreus strains. The comparative analysis revealed considerable diversity with regard to the catalog of biosynthetic gene clusters found in A. terreus. Even though the cluster of lovastatin biosynthesis was found to be well-conserved at the species level, several unique genes putatively associated with metabolic functions were detected in A. terreus ATCC 20542 that were not detected in other investigated genomes. The analysis was conducted also in the context of the primary metabolic pathways (sugar catabolism, biomass degradation potential, organic acid production), where the visible differences in gene copy numbers were detected. However, the species-level genomic diversity of A. terreus was more evident for secondary metabolism than for the well-conserved primary metabolic pathways. The newly sequenced genome of A. terreus ATCC 20542 was found to harbor several unique sequences, which can be regarded as interesting subjects for future experimental efforts on A. terreus metabolism and fungal biosynthetic capabilities.Key points• The high-quality genome of Aspergillus terreus ATCC 20542 has been assembled and annotated.• Comparative analysis with other sequenced Aspergillus terreus strains has revealed considerable diversity in biosynthetic gene repertoire, especially related to secondary metabolism.• The unique genomic features of A. terreus ATCC 20542 are discussed.

The quorum-sensing system, consisting of an autoregulator synthase (AfsA or Aco homolog) and an autoregulator receptor (ArpA homolog), has been reported to be universally involved in regulating secondary metabolism in streptomycetes. Although the autoregulator synthase is thought to activate antibiotic production, the activation pathway remains poorly understood. Streptomyces cyaneogriseus ssp. noncyanogenus NMWT1 produces nemadectin, which is widely used as a biopesticide and veterinary drug due to its potent nematocidal activity. Here, we identified the Aco/ArpA-like system ScyA1/ScyR1, the ArpA homolog ScyR2 and the AfsA/ArpA-like system ScyA3/ScyR3 as important regulators of nemadectin production in NMWT1. Genetic experiments revealed that these five genes positively regulate nemadectin production, with scyA1 and scyR1 having the most potent effects. Importantly, ScyA1 is an upstream regulator of scyR1 and promotes nemadectin production and sporulation by activating scyR1 transcription. Intriguingly, scyR1 silencing in NMWT1 up-regulated 12 of the 17 secondary metabolite biosynthetic core genes present in the NMWT1 genome, suggesting that ScyR1 mainly to be a repressor of secondary metabolism. In conclusion, our findings unveiled the regulatory pathways adopted by the quorum-sensing system, and provided the basis for a method to enhance antibiotic production and to activate the expression of cryptic biosynthetic gene clusters.

Article Figures and data Abstract Introduction Results Discussion Materials and methods Data availability References Decision letter Author response Article and author information Metrics Abstract The biosynthetic capabilities of microbes underlie their growth and interactions, playing a prominent role in microbial community structure. For large, diverse microbial communities, prediction of these capabilities is limited by uncertainty about metabolic functions and environmental conditions. To address this challenge, we propose a probabilistic method, inspired by percolation theory, to computationally quantify how robustly a genome-derived metabolic network produces a given set of metabolites under an ensemble of variable environments. We used this method to compile an atlas of predicted biosynthetic capabilities for 97 metabolites across 456 human oral microbes. This atlas captures taxonomically-related trends in biomass composition, and makes it possible to estimate inter-microbial metabolic distances that correlate with microbial co-occurrences. We also found a distinct cluster of fastidious/uncultivated taxa, including several Saccharibacteria (TM7) species, characterized by their abundant metabolic deficiencies. By embracing uncertainty, our approach can be broadly applied to understanding metabolic interactions in complex microbial ecosystems. https://doi.org/10.7554/eLife.39733.001 Introduction Metabolism, in addition to enabling growth and homeostasis for individual microbes, contributes to the organization of complex, dynamic microbial communities. Within these communities, different microbes have diverse metabolic capabilities that lead to interactions driving microbial community structure and dynamics at multiple spatial and temporal scales (Ponomarova and Patil, 2015; Phelan et al., 2012; Watrous et al., 2013; Harcombe et al., 2014; Embree et al., 2015). For example, through cross-feeding, a compound produced by one species might benefit another, leading to a network of metabolic interdependences (Embree et al., 2015; Goldford et al., 2017; Mee et al., 2014; Pande et al., 2015; D'Souza et al., 2018; Zengler and Zaramela, 2018; Pacheco et al., 2019; Mee and Wang, 2012). This type of interaction has been proposed as one of the main reasons for the prevalence, in natural microbial communities, of uncultivated (or fastidious) microbes (Stewart, 2012; Epstein, 2013; Pande and Kost, 2017; Staley and Konopka, 1985). These microbes do not grow in pure culture on standard laboratory conditions as they may depend on diffusible metabolites produced by neighboring microbes (Pande and Kost, 2017). The prominence of uncultivated/fastidious microbial organisms across the tree of life and their potential importance in microbial community structure is highlighted by the recent identification of the candidate phyla radiation – a large branch of the tree of life consisting mainly of uncultivated organisms with small genomes and unique metabolic properties (Kantor et al., 2013; Brown et al., 2015; Hug et al., 2016). Efforts towards understanding this important component of microbial communities require further knowledge of metabolic interdependencies driven by biosynthetic deficiencies. Some of the most promising strides in understanding metabolic interdependences between microbes have been taken in the study of the human oral microbiome. The human oral microbiome serves as an excellent model system for microbial communities research, due to its importance for human health and ease of access for researchers (Dewhirst et al., 2010; Wade, 2013). For example, the order of colonization of species in dental plaque has been characterized physically (Kolenbrander et al., 2010) and metabolically (Mazumdar et al., 2013), and visualized microscopically (Mark Welch et al., 2016). The human oral microbiome consists of roughly 700 different microbial species, identified by 16S rRNA microbiome sequencing and cataloged in the human oral microbiome database (Dewhirst et al., 2010; Chen et al., 2010). Importantly, 63% of species in the human oral microbiome have been sequenced, including several uncultivated and recently-cultivated strains implicated in oral health and disease (Krishnan et al., 2017; Siqueira Jr and Rôças, 2013). Exciting recent work has led to successful laboratory co-cultivation of at least three previously uncultivated organisms, the Saccharibacteria (TM7) phylum taxa: Saccharibacteria bacterium HMT-952 strain TM7x (Bedree et al., 2018; He et al., 2015; Bor et al., 2016; Bor et al., 2018), Saccharibacteria bacterium HMT-488 strain AC001 (Collins et al., 2019a), and Saccharibacteria bacterium HMT-955 strain PM004 (Collins et al., 2019b). Saccharibacteria are prominent in the oral cavity and relevant for periodontal disease (Brinig et al., 2003; Ouverney et al., 2003). Due to their importance, they were among the first uncultivated organisms from the oral microbiome to be fully sequenced via single-cell sequencing methods (Marcy et al., 2007), and represent the first co-cultivated members of the candidate phyla radiation (He et al., 2015). Thus, their metabolic and phenotypic properties are of great interest for oral health and microbiology in general. In parallel to achieving laboratory growth of diverse and uncultivated bacteria, a major unresolved challenge is understanding the detailed metabolic mechanisms that may underlie their dependencies. Ideally, one would want to computationally predict, directly from the genome of an organism, its biosynthetic capabilities and deficiencies, so as to translate sequence information into mechanisms and community-level phenotypes (Widder et al., 2016). A number of approaches, based on computational analyses of metabolic networks, have contributed significant progress towards this goal (Schuster et al., 2000; Oberhardt et al., 2009; Lewis et al., 2012). At the heart of these methods are metabolic network reconstructions, formal encodings of the stoichiometry of all metabolic reactions in an organism, that are readily amenable to multiple types of in silico analyses and simulations (Feist et al., 2009). Recent exciting progress has led to the automated generation of ‘draft’ metabolic network reconstructions for any organism with a sequenced genome (Henry et al., 2010), opening the door for the quantitative study of large and diverse microbial communities. The most commonly used metabolic network analysis methods – flux balance analysis (FBA) (Orth et al., 2010a) and its dynamic version (dFBA) (Mahadevan et al., 2002) – have been extensively applied to study microbial communities (Harcombe et al., 2014; Embree et al., 2015; Pacheco et al., 2019; Magnúsdóttir et al., 2017; Magnúsdóttir and Thiele, 2018; Zarecki et al., 2014; Stolyar et al., 2007; Klitgord and Segrè, 2010; Freilich et al., 2011; Zelezniak et al., 2015; Cook and Nielsen, 2017; Biggs et al., 2015; Zomorrodi and Segrè, 2016). However, FBA and dFBA are not easily applicable to automatically-generated draft metabolic networks due to gaps (missing or incorrect reactions) in the metabolic network, and are thus difficult to scale to large and diverse microbial communities. Methods for ‘gap-filling’ draft reconstructions can address this problem, and ensemble methods potentially present a promising approach (Biggs and Papin, 2017; Machado et al., 2018). However, any gap-filling comes at the expense of an increased risk for false positive predictions. Additionally, gap-filling typically requires specific knowledge or assumptions on the growth media composition – which are often difficult to obtain for diverse environmental isolates and by definition unknown for uncultivated organisms. Alternatively, topology-based metabolic network analysis methods, such as network expansion (Ebenhöh et al., 2004) and NetSeed-based methods (Borenstein et al., 2008), are less dependent on gap-filling and have been applied to the analysis of draft metabolic reconstructions. These methods have provided valuable large-scale insight into metabolic processes in microbial communities, including the biosynthetic potentials of organisms and metabolites (Basler et al., 2008; Matthäus et al., 2008), the chance of cooperation or competition between species (Carr and Borenstein, 2012; Kreimer et al., 2012; Levy et al., 2015; Opatovsky et al., 2018), and the relationship between organisms and environment (Borenstein et al., 2008; Freilich et al., 2009; Handorf et al., 2008), for example in the human gut microbiome (Levy and Borenstein, 2013). While all of these approaches are promising, an additional issue that continues to limit the use of metabolic network analysis for prediction of biosynthetic capabilities is the difficulty of generating these predictions when the chemical environment of the microbes is unknown. In complex microbial communities, such as the human microbiome, the exact chemical composition of the environment is difficult to estimate, due both to the molecular complexity of the environment itself, and to the likely prevalence of secretions, lysing and cross-feeding within the community. Thus, the capacity to provide metabolic predictions based on unelaborated genome annotation, and on limited knowledge about an organism’s growth environment remains an important open challenge. Here we introduce a new method, which begins to address the above limitations, and provides a novel prediction of an organism’s biosynthetic capabilities. Our method applies a probabilistic approach to define and compute a metric that estimates which metabolites, such as biomass components, are robustly synthesized by a given metabolic network and which would likely need to be supplied from the environment/community. Discrepancies in these calculated estimates between organisms can be used to generate hypotheses regarding microbial auxotrophy and metabolic exchange in microbial communities. Importantly, our metric has the capacity to estimate biosynthetic capabilities in spite of uncertainty about environmental conditions by randomly sampling many different possible nutrient combinations. In this study, we first demonstrated our method on E. coli to clarify its performance and interpretation. Next, we applied our method to a large number of organisms from the human oral microbiome, and predicted broad trends in biosynthetic capabilities associated with taxonomy and microbial co-occurrence. We further focused our analysis on uncultivated microorganisms, including three recently co-cultivated Saccharibacteria (TM7) strains. In addition to highlighting their biosynthetic deficiencies, we developed specific hypotheses for their metabolic exchange with growth-supporting partner microbes. Analysis method Our newly developed method quantifies the robustness with which a given metabolic network can produce a given metabolite from variable metabolic precursors. In essence, we quantify a metabolic network specific metric for metabolite producibility by probabilistically sampling sets of possible environments. While the probabilistic sampling can be adjusted to reflect a specific environment, its power lies largely in the capacity to explicitly incorporate in a statistical way the lack of knowledge about environmental composition. The inspiration for this method comes from the statistical physics concept of percolation. Percolation theory has been applied in a wide range of fields, including the study of cascading metabolic failure upon gene deletions in metabolism (Smart et al., 2008; Barabási, 2015). In percolation theory the robustness of a network can be characterized by randomly adding or removing components (nodes or edges) of a network and assessing network connectivity (Barabási, 2015). The smaller the number of components that need to be randomly added to the network before it becomes connected, the more robust it is to perturbations. We utilized this concept to characterize the network robustness of a particular metabolic network towards producing a specified target metabolite by randomly adding input metabolites to the network and assessing the network’s ability to produce the target. To implement our method, we first introduced a probabilistic framework for analyzing metabolic networks (Figure 1 and Figure 1—figure supplement 1). In this framework, every metabolite can be considered to be drawn from a Bernoulli distribution, i.e. present in the network with a given input probability (Pin). These probabilities could represent beliefs about the environment, chances of metabolites being available from a host organism, or any arbitrary prior assumption on metabolite inputs. Throughout the majority of our analyses we have assigned Pin to be an identical value for all input metabolites. However, as illustrated in an example in our results section (Metabolite producibility in a protein vs. carbohydrate-enriched environment) this probabilistic framework can utilize Pin values that vary across metabolites. Following the assignment of Pin, the network structure is used to calculate the output probability (Pout) of some specified target metabolite. In practice, random sampling of probabilistically drawn input metabolite sets is used to calculate the probability of producing the target metabolite. For each random sample, a modified version of FBA (Orth et al., 2010a) is used to assess the network's ability to produce the target metabolite (for a complete explanation of how FBA is implemented in this context, see methods section: Algorithm functions, feas). Figure 1 with 2 supplements see all Download asset Open asset A probabilistic framework for calculating the producibility metric (PM). (A) Random samples of input metabolites are added to the metabolic network with probability Pin. Samples are shown here with gray or red circles. Sampled input metabolites are then used to calculate if a specified target output metabolite can be produced or not. Here the solid red circled sample leads to production of the target metabolite while the dotted gray circled samples do not. The probability of producing the target output metabolite (Pout) is calculated by taking many random samples at a specified Pin. (B) A producibility curve is calculated which represents Pout as function of Pin. Points along this curve are sampled by assigning the Pin value and estimating Pout. The Pin value at which Pout = 0.5 (Pin,0.5) is used to define the producibility metric (PM) as PM = 1-Pin,0.5. https://doi.org/10.7554/eLife.39733.002 Using the above probabilistic framework, we defined a novel metric quantifying biosynthetic capabilities, the producibility metric (PM) (Figure 1B). The PM is calculated as follows: First, a producibility curve describing Pout as a function of Pin is generated for a given metabolic network and metabolite target. This curve can be estimated by sampling input metabolites for different values of Pin (between 0 and 1), and calculating Pout. Next, we calculated the Pin value along the producibility curve at which Pout is equal to 0.5 (Pin,0.5, analogous to the Km in the Michaelis-Menten curve). Finally, PM is defined as PM = 1-Pin,0.5, such that larger PM values correspond to increased robustness. Our method calculates PM efficiently by random sampling and a nonlinear fitting algorithm (for details, see methods section: Algorithm functions calc_PM_fit_nonlin). In addition to calculating PM computationally for arbitrary metabolic networks and metabolites, we also derived a way to calculate PM analytically using combinatorial equations. The combinatorial equations are built up from simple scenarios to the most general in Figure 1—figure supplement 2. This analytical result, verified in detail for one specific pathway (Figure 2—figure supplement 2) clarifies the connection between our metric and the concept of minimal precursor sets (Andrade et al., 2016). It describes mathematically how the PM captures the multiplicity of routes through which a given target metabolite can be produced, and could serve as the basis for further theoretical work on the fundamental properties of metabolic networks. The algorithms used to implement our method are written in MATLAB and designed as a set of modular functions that interface with the COBRA toolbox – a popular metabolic modeling software compendium (Schellenberger et al., 2011; Heirendt et al., 2019). The methodology behind each function is further explained in the methods section. The code is freely available online at https://github.com/segrelab/biosynthetic_network_robustness (Bernstein, 2019; copy archived at https://github.com/elifesciences-publications/biosynthetic_network_robustness). Results Using the E. coli core metabolic network to demonstrate features of metabolite producibility Before applying our approach to the systematic study of genome-scale metabolic networks from the human oral microbiome, we used the model organism E. coli to illustrate the performance and interpretation of our method. We started with the E. coli core metabolic network, a simplified network consisting of central carbon metabolism and lacking peripheral metabolic pathways, such as amino acid or cofactor biosynthesis (Orth et al., 2010b). We calculated the PM for all intracellular metabolites in this network using a uniform ensemble of environments (as described in the methods). The results are shown in Figure 2A, overlaid on the E. coli core metabolic network itself, with each node’s color indicating its PM value and node size indicating its degree of connectivity. Consistent with the high connectivity of the E. coli core metabolic network, most metabolites have high PM values (PM >0.950). For example, the metabolites H+ and pyruvate are both highly connected in the metabolic network and have high PM (PM = 0.968 and 0.952 respectively). However, the network also contains several metabolites that are well connected, but have lower PM values. These include, for example, the cofactors AMP/ADP/ATP and NAD+/NADH, which have PM values of ~0.7 and ~0.5 respectively, because they can be produced from each other, but not biosynthesized in this network. The network also includes several examples of metabolites that are poorly connected but have high PM values. One example is D-lactate, which is produced only via Lactate Dehydrogenase from the high PM metabolites Pyruvate and H+ (Figure 2B). This reaction also consumes NADH and produces NAD+ but because these cofactors can be easily recycled from each other by a large number of different reactions, their relatively low PM (as described above) has minimal influence on the PM value of D-lactate (Figure 2B). This example demonstrates the fact that our metric captures metabolites which are easily produced because their precursors are easily produced, and that the PM of recycled cofactors has minimal influence on the PM of a target metabolite. Overall, there is also no significant correlation between the PM values and the node degree of a metabolite in the network (Figure 2—figure supplement 1), indicating that our metric describes a more complex property of a metabolite in a network that is not captured simply by node degree. Figure 2 with 2 supplements see all Download asset Open asset E. coli core metabolic network metabolite producibility. (A) The E. coli core metabolic network is represented as a bipartite graph with metabolites shown as circles and reactions shown as squares. Reactions shown with a black border are irreversible in the model, those with no border are reversible. All intracellular metabolites are colored based on their PM value (low – blue, high – red). Reactions and metabolite nodes are sized based on their total node degree. Several key metabolites of interest are highlighted with their corresponding PM values shown. Central metabolites such as H+ and Pyruvate have high degree and high PM. Cofactors such as AMP/ADP/ATP and NAD+/NADH have high degree but low PM, as they cannot be synthesized in this network. Oxygen is an example of a PM=0 metabolite that cannot be produced from any other metabolites in this network. D-lactate is an example of a metabolite with low degree and high PM that is it is easily produced but not well-connected. (B) The lactate dehydrogenase reaction producing D-Lactate is shown as an example to illustrate that poorly connected metabolites can display a high PM, and how recycled cofactors have minimal impact on PM values. Lactate dehydrogenase produces D-lactate and NAD+ from pyruvate, H+ and NADH. The metabolite D-lactate has high PM despite being produced only by this one reaction in the metabolic network because it can be produced from the high PM metabolites pyruvate and which are produced from a large number of possible precursors. NADH is also used to produce D-lactate, and has a relatively low PM in this core model, it has minimal impact on the PM of D-lactate as NADH can be recycled from NAD+ by a large number of reactions by the at the of the and thus production of NADH is not for the production of of metabolites from pathway and captures minimal precursor set structure We applied our method in detail to a specific biosynthetic pathway within a genome-scale model to demonstrate how our PM provides information that is can be from simply the of reactions present in a given biosynthetic we the biosynthetic pathway in the E. coli genome-scale metabolic network (Orth et al., and how the methods in their capacity to the of reaction along the pathway (Figure 2—figure supplement The PM is more pathway as it captures features the of reactions in the biosynthetic For different in the biosynthetic pathway distinct reactions) the PM is to the of the reaction from the target metabolite the would be the of for each (Figure 2—figure supplement 2B). This capacity of PM to of the of reactions in a pathway is also by a analysis of biosynthesis across all oral microbiome draft metabolic networks microbiome network and analysis described in the (Figure 2—figure supplement In in with the of the biosynthetic the PM on the pathway the number of different routes through which the target metabolite can be synthesized the minimal precursor et al., 2016). This property from the way the PM is and is explained by our combinatorial theory (Figure 1—figure supplement While our computational estimate of the PM is based on sampling the of possible precursor the combinatorial theory provides an exact value for the producibility of a with a given minimal precursor set structure. The between the PM and the combinatorial theory for the biosynthetic pathway (Figure 2—figure supplement that the PM captures the complex multiplicity of for producing a given metabolite. analysis to reactions to flux balance analysis One of the we to address with our method is the of robust about the metabolic capabilities of different organisms in spite of reactions – a often upon metabolic networks from newly sequenced To assess the performance of our approach in this context, we it with flux balance analysis (FBA) for a genome-scale metabolic networks with a given number of randomly In we applied both FBA and our method to the E. coli genome-scale metabolic network, which we by removing an number of randomly In this performance the metabolic network used as a which the predictions of our method and FBA on metabolic networks were Figure the of both FBA and the PM as a function of the of reactions from the metabolic network. While the output of our method PM for any is different from that of FBA flux through all one can use the PM values across all biomass components as a for the growth capacity of an organism, a metric that is with the biomass production The specific used to the PM and FBA predictions for biomass production are described further in the Figure One can see that both the FBA and the PM predictions as the metabolic networks are further However, the PM predictions are more to reactions the FBA predictions. While the FBA production of biomass becomes for the majority of the metabolic networks removing less of the reactions, the PM results when removing up to of the This analysis provides insight into the of our method for analyzing metabolic networks with such as draft metabolic networks produced through automated Figure with 1 supplement see all Download asset Open asset The of flux balance analysis and the producibility metric for different E. coli genome-scale metabolic networks. Reactions were randomly from the E. coli metabolic network generating different networks at different of reaction These networks were then with the producibility metric (PM) and flux balance analysis (FBA) in a minimal and complete The of the PM and FBA results were through different and as a function of the number of reactions on a (A) – The based on the of the between the network metric and the randomly network For FBA the as the value of the between the biomass flux of the network and the network. For the PM the calculated as the of the value of the between each PM The for both then and from one to a of The of different randomly networks at different reaction is shown with connected by solid on minimal FBA on complete The standard of the metric is shown as a the (B) production – The by the of randomly metabolic networks that were of producing For FBA this calculated as the of networks of producing biomass flux above of the biomass flux on minimal FBA on complete For the PM, the biomass production calculated as the of networks of producing all biomass components above a specified PM The PM or PM producibility to metabolic mechanisms for E. coli a first of our approach in its capacity to provide metabolic insight about of inter-microbial interactions, we used the PM to estimate the capacity of different E. coli to for each metabolic In we data from of E. coli from and with corresponding PM in silico the specific strains used in this work on the E. coli metabolic we calculated the PM for all biomass components in each and the PM values to the growth of (Figure supplement 1). by PM, were to based on the pathway of the and with in different of the biosynthetic pathway a in PM for the corresponding biomass to in our biosynthetic pathway analysis in Figure 2—figure supplement 2. The between PM values with that with different biosynthetic capabilities could each growth (Figure supplement Several examples and that further this are highlighted in Figure supplement and This analysis also the to in more the capacity of our approach to provide insight into with PM

Antibiotic resistance and the golden age of microbiology

Siderophore-mediated iron acquisition has been well studied in many bacterial pathogens because it contributes to virulence. In contrast, siderophore-mediated iron acquisition by saprophytic bacteria has received relatively little attention. The independent identification of the des and cch gene clusters that direct production of the tris-hydroxamate ferric iron-chelators desferrioxamine E and coelichelin, respectively, which could potentially act as siderophores in the saprophyte Streptomyces coelicolor A3(2), has recently been reported. Here it is shown that the des cluster also directs production of desferrioxamine B in S. coelicolor and that very similar des and cch clusters direct production of desferrioxamines E and B, and coelichelin, respectively, in Streptomyces ambofaciens ATCC 23877. Sequence analyses of the des and cch clusters suggest that components of ferric-siderophore uptake systems are also encoded within each cluster. The construction and analysis of a series of mutants of S. coelicolor lacking just biosynthetic genes or both the biosynthetic and siderophore uptake genes from the des and cch clusters demonstrated that coelichelin and desferrioxamines E and B all function as siderophores in this organism and that at least one of these metabolites is required for growth under defined conditions even in the presence of significant quantities of ferric iron. These experiments also demonstrated that a third siderophore uptake system must be present in S. coelicolor, in addition to the two encoded within the cch and des clusters, which show selectivity for coelichelin and desferrioxamine E, respectively. The ability of the S. coelicolor mutants to utilize a range of exogenous xenosiderophores for iron acquisition was also examined, showing that the third siderophore-iron transport system has broad specificity for tris-hydroxamate-containing siderophores. Together, these results define a complex system of multiple biosynthetic and uptake pathways for siderophore-mediated iron acquisition in S. coelicolor and S. ambofaciens.

Coastal ecosystems are increasingly threatened by anthropogenic activities, including sewage discharge and tourism-related pollution, which alter microbial diversity and biochemical cycles. This study applied molecular techniques to examine the coral microbial diversity, and metabolite composition of seawater across five sites (A–E) in Dapeng Cove, South China Sea, to assess pollution impacts. Sites A and B, within the yacht tourism area, exhibited high microbial diversity, dominated by Synechococcus and Rhodobacteraceae, with minimal pollution effects. Site C, inside a domestic drainage channel, showed moderate pollution, with elevated nitrite (NO₂) and nitrate (NO₃) levels, microbial taxa linked to organic matter degradation, and increased hydroxy acids and indoles. Sites D and E, located in main sewage channels, experienced severe pollution, characterized by high salinity, low dissolved oxygen, and dominance of pollution-tolerant bacteria such as Exiguobacterium and Tepidibacter. Metabolite analysis revealed elevated fatty acyls, organonitrogen compounds, and amino acids at these sites, highlighting strong anthropogenic influence. Beta diversity analysis (NMDS and ANOSIM) confirmed distinct microbial community structures, while KEGG pathway analysis indicated shifts in metabolic functions, with enrichment in xenobiotic biodegradation and anaerobic respiration in sewage-impacted areas. These findings underscore the detrimental effects of wastewater discharge on microbial ecology and biochemical functions. Urgent interventions, including improved wastewater management and regular environmental monitoring, are recommended to mitigate pollution effects. Future research integrating multi-omics approaches is necessary to evaluate the long-term ecological consequences of pollution and climate variability on coastal microbial communities.

Mulching to conserve moisture has become an important agronomic practice in saline soil cultivation, and the effects of the dual stress of salinity and microplastics on soil microbes are receiving increasing attention. In order to investigate the effect of polyethylene microplastics on the microbial community of salinized soils, this study investigated the effects of different types (chloride and sulphate) and concentrations (weak, medium, and strong) of polyethylene (PE) microplastics (1% and 4% of the dry weight mass of the soil sample) on the soil microbial community by simulating microplastic contamination in salinized soil environments indoors. The results showed that:PE microplastics reduced the diversity and abundance of microbial communities in salinized soils and were more strongly affected by sulphate saline soil treatments. The relative abundance of each group of bacteria was more strongly changed in the sulphate saline soil treatment than in the chloride saline soil treatment. At the phylum level, the relative abundance of Proteobacteria was positively correlated with the abundance of fugitive PE microplastics, whereas the relative abundances of Bacteroidota, Actinobacteriota, and Acidobacteria were negatively correlated with the abundance of fugitive PE microplastics. At the family level, the relative abundances of Flavobacteriaceae, Alcanivoracaceae, Halomonadaceae, and Sphingomonasceae increased with increasing abundance of PE microplastics. The KEGG metabolic pathway prediction showed that the relative abundance of microbial metabolism and genetic information functions were reduced by the presence of PE microplastics, and the inhibition of metabolic functions was stronger in sulphate saline soils than in chloride saline soils, whereas the inhibition of genetic information functions was weaker than that in chloride saline soils. The secondary metabolic pathways of amino acid metabolism, carbohydrate metabolism, and energy metabolism were inhibited. It was hypothesized that the reduction in metabolic functions may have been caused by the reduced relative abundance of the above-mentioned secondary metabolic pathways. This study may provide a theoretical basis for the study of the effects of microplastics and salinization on the soil environment under the dual pollution conditions.