Figures
Abstract
The human GI tract is a complex and still poorly understood environment, inhabited by one of the densest microbial communities on earth. The gut microbiota is shaped by millennia of evolution to co-exist with the host in commensal or symbiotic relationships. Members of the gut microbiota perform specific molecular functions important in the human gut environment. This can be illustrated by the presence of a highly expanded repertoire of proteins involved in carbohydrate metabolism, in phase with the large diversity of polysaccharides originating from the diet or from the host itself that can be encountered in this environment. In order to identify other bacterial functions that are important in the human gut environment, we investigated the distribution of functional groups of proteins in a group of human gut bacteria and their close non-gut relatives. Complementary to earlier global comparisons between different ecosystems, this approach should allow a closer focus on a group of functions directly related to the gut environment while avoiding functions related to taxonomically divergent microbiota composition, which may or may not be relevant for gut homeostasis. We identified several functions that are overrepresented in the human gut bacteria which had not been recognized in a global approach. The observed under-representation of certain other functions may be equally important for gut homeostasis. Together, these analyses provide us with new information about this environment so critical to our health and well-being.
Citation: Dobrijevic D, Abraham A-L, Jamet A, Maguin E, van de Guchte M (2016) Functional Comparison of Bacteria from the Human Gut and Closely Related Non-Gut Bacteria Reveals the Importance of Conjugation and a Paucity of Motility and Chemotaxis Functions in the Gut Environment. PLoS ONE 11(7): e0159030. https://doi.org/10.1371/journal.pone.0159030
Editor: Francesco Cappello, University of Palermo, ITALY
Received: April 15, 2016; Accepted: June 24, 2016; Published: July 14, 2016
Copyright: © 2016 Dobrijevic et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Data Availability: All relevant data are within the paper and its Supporting Information files.
Funding: Dragana Dobrijevic is the recipient of a grant in the Marie Curie EC FP7 Cross-talk project (PITN-GA-2008-215553).
Competing interests: The authors have declared that no competing interests exist.
Introduction
The human gastrointestinal system, and especially the distal gut, is inhabited by one of the densest populations of microorganisms known. The importance of this community, the human gut microbiota, for human health and wellbeing is now well documented. While some of the cultivable bacteria living in the human gut have been studied for decades (for instance Escherichia coli), the development of new DNA sequencing technologies and the concept of metagenomics provided a paradigm-changing shift in the study of the human gut microbiota. The collective genetic information of the human gut microbiota, the human gut metagenome, is currently the focus of intense international sequencing and research efforts. A first catalogue of 3.3 million human gut microbial genes has been established in 2010 [1], and more recently an extensive update of this catalogue was published, combining data from different sources and containing nearly 10 million genes [2]. Apart from metagenomics projects, data derive from sequencing efforts targeting the genomes of specific bacteria from the human gut. So far, sequences of 778 gut-associated bacterial genomes are available through the Human Microbiome Project [3], thus giving access to an independent complementary line of investigation of the human gut microbiota.
More than 90% of gut bacteria are members of only two phyla, Bacteroidetes and Firmicutes, the relative proportions of which exhibit a continuous gradient within the human population [4, 5]. Within the borders of these phyla the microbiota composition is highly individual-specific, showing high variability at the species and strain levels. In contrast to this taxonomic diversity however, functional profiles are far less variable across healthy individuals [1, 6, 7], confirming the existence of a well-balanced host-microbial symbiosis.
The results of functional analyses of the human gut metagenome showed that the proteome of the human gut microbiota is enriched in proteins involved in carbohydrate metabolism, energy metabolism and storage, generation of short-chain fatty acids, amino acids metabolism, biosynthesis of secondary metabolites and metabolism of cofactors and vitamins [1, 4, 6–8]. In particular, the human gut microbiota provides a broad and diverse array of carbohydrate-active enzymes, many of which are not present in the human glycobiome [9]. Similar observations were made during genome analysis of some of the human gut-associated bacteria, such as Bacteroides fragilis, Bacteriodes thetaiotaomicron or Methanobrevibacter smithii [10, 11]. Additionally, many yet uncharacterized or completely novel protein families were shown to be specific to the human gut, suggesting that many unknown and uncharacterized processes are yet to be discovered in this environment [12].
The formentioned catalogue containing 3.3 million non-redundant microbial genes from the intestinal tract of 124 individuals [1] provided an opportunity to differentiate bacterial functions necessary for a bacterium to thrive in the gut environment, and therefore present in every gut bacterial species, from those involved in the homeostasis of the whole gut ecosystem, encoded across many bacterial species. Qin et al. referred to these functions as (bacterial) "minimal gut genome" and "minimal gut metagenome", respectively [1]. The minimal gut metagenome includes functions known to be important to the host—bacterial interaction, such as the capacities to metabolize complex polysaccharides and to synthesize short-chain fatty acids, indispensable amino acids and vitamins. It also includes a considerable fraction of functions (~45%) that were present in less than 10% of earlier sequenced, mainly non-gut, bacterial genomes. These "otherwise rare", gut-specific, functions mainly contained uncharacterized genes, underscoring our limited knowledge of gut functioning.
Here we present a different approach to the identification of functions that may inform us on the conditions of gut homeostasis, comparing the predicted proteomes of fully sequenced gut bacteria to those of closely related bacteria from other environments. The underlying hypothesis is that the comparison of metagenomic data from different ecosystems, for example gut and soil, may reveal functions that are characteristic of either ecosystem partly because of the different constraints of each system (i.e. adaptive functions) and partly because the two ecosystems are populated by different bacterial phyla with inherent differences between them (which may or may not be important in adaptation to the environment). Also, the global approach would not recognize the importance of certain functions in the gut environment if these functions also play a role in (other bacteria in) other environments. The comparison of closely related bacterial species from different environments allows focusing on differences that are directly related to the different environments.
Materials and Methods
Data preparation
Information on sequenced genomes was downloaded from the GOLD database (www.genomesonline.org) (version of 07 March 2013). Gene and protein sequences were downloaded from NCBI via iMOMi [13] except for those of Megamonas rupellensis DSM 1994 and Caloramator australicus KCTC 5601 which were downloaded from Integrated Microbial Genomes data warehouse (www.img.jgi.doe.gov).
Phylogenetic analysis
The 16S rRNA gene sequences used in this study were retrieved in fasta format from GenBank (NCBI-GenBank Flat File Release 195.0). Phylogenetic analysis was conducted using the SeaView 4.4.1 platform [14]. MUSCLE [15] was used for multiple sequence alignment with default parameters and blocks of evolutionary conserved sites were selected by Gblocks [16]. The tree was computed using phyML 3.0 [16] based on the Maximum Likelihood method and visualized using iTOL v3.0 [17]. The 16S rRNA gene sequence of Methanobrevibacter smithii ATCC 35061 was used as an outgroup for the analysis.
Protein functional annotation and localization prediction
Protein functional annotations were made by BLASTP search against the eggNOGv3.0 bactNOG catalog [18] and the best hit (e-value <10e-5) was retained. To predict protein localization, protein sequences were analyzed using SurfG+ [19]. Functional comparisons were performed using R (https://www.r-project.org): functional clustering of the bacterial species used in this study was performed with the heatmap function taking bactNOGs present in ≥ 2 species and ≤ 45 species into account (i.e. bactNOGs present in only one species or in all 46 species studied, which are not informative for clustering, were ignored); the distribution of functions among gut and non-gut genomes was visualized with the function hist2d of the gplots library (version 2.17.0), bin contents was summarized with a log function.
Results
Closely related gut and non-gut bacterial species constitute distinct functional groups
Bacterial diversity in the gut is largely restricted to two major phyla, Firmicutes and Bacteroidetes. In this study, we chose to focus on gut bacteria from the Firmicutes phylum and compare the functions encoded in their genomes with those encoded in closely related bacteria from other environments. Information on sequenced genomes was downloaded from the GOLD database (www.genomesonline.org) and filtered to select publicly available complete genomes from the Firmicutes phylum. Among these, "gut species" were identified by comparison with the gut metagenome data presented in [1], choosing species that in the latter study were marked "frequent" (i.e. highly abundant in the gut) and/or "common" (i.e. present in the gut of many individuals). "Non-gut" species, for which no association with the gut environment could be found in literature, were chosen in taxonomic groups as close as possible to the gut-species. To validate this choice, the gene repertoires of the selected species (one strain per species) were analyzed using BLASTn against the MetaHIT gut microbiota gene catalog [1] to identify genes present in this catalog (sequence identity ≥ 95% over ≥ 90% of the longest sequence length). A species was considered as “gut species” if at least 5% of the genes were present in the catalog, and as “non-gut” species if less than 0.5% of the genes were present in the catalog. This resulted in the constitution of two genome sets with “gut” or “non-gut” attributions, respectively, that were coherent between data obtained from literature and from metagenomic sequencing of fecal samples. The first set represents 23 bacterial species isolated from human stool samples. The second set represents 23 closely related species from other environments (Table 1, Fig 1). We refer to these sets of genomes as “gut” and “non-gut”, respectively. The majority of the selected species belongs to the class Clostridia, order Clostridiales (19 gut species and 14 non-gut species).
16S rRNA based tree by Maximum Likelihood method. Green, gut species; Violet, non-gut species. Clusters A and B make part of one functional cluster (Fig 2).
To evaluate the functional composition of the species studied, we attributed the predicted protein complement of each genome to orthologous groups (bactNOGs) using the eggNOG v.3.0 database [18]. Orthologous groups of proteins have proven useful for functional analyses, as orthologs tend to have equivalent functions [20]. The fraction of the proteins that could be attributed to bactNOGs varied from 48 to 81% depending on the species (Table 2), and was generally higher for non-gut species than for gut species (71% vs 60%, respectively, on an average). For each genome this resulted in a functional profile based on the presence of bactNOGs, which was subsequently used as a basis for functional genome clustering (Fig 2). With only few exceptions, the gut bacterial species studied appear to form a distinct cluster. Notably, the phylogenetically distinct gut bacteria clusters A and B (Fig 1) appear to be united in one functional cluster (Fig 2).
Clustering of 46 species on the basis of functional profiles (presence or absence of bactNOGs). Green, gut species; Violet, non-gut species. A and B represent phylogenetically distinct clusters (Fig 1).
Functional adaptation to the human gut environment through gut-specific functions
In order to examine potential gut adaptation among Firmicutes, we compared the distribution of bactNOGs among gut and non-gut bacteria (Fig 3). The majority of the 20,426 detected bactNOGs was present in only one of the 46 bacterial species studied (7924 bactNOGs, not shown), or shared between two (3527 bactNOGs) or three species (2164 bactNOGs) (Fig 3, bottom left). For further analysis, bactNOGs for which the number of gut genomes where the bactNOG is represented exceeds the number of non-gut genomes where the bactNOG is represented by at least 12 were regarded as “overrepresented” in genomes of gut Firmicutes. This means that the representation of functions that are regarded as overrepresented varies from "present in at least 12 of the 23 gut genomes and none of the 23 non-gut genomes" to "present in all 23 gut genomes and 11 or less of the 23 non-gut genomes" (Fig 3).
Squares indicate bactNOGs as a function of the number of gut genomes (horizontal axis) and the number of non-gut genomes (vertical axis) in which they are encoded. The colours of the squares indicate the numbers of different bactNOGs at each position. BactNOGs encoded in only one of the 46 genomes are not indicated. Diagonal lines separate bactNOGs that are overrepresented in the gut genomes (bottom right), bactNOGs that are underrepresented in the gut genomes (top left), and bactNOGs with an intermediate position (see text for details).
The majority of the 153 overrepresented bactNOGs have no known function, or a function involved in energy production, DNA metabolism, transcription or translation (Fig 4, S1 Table). Individual bactNOG descriptions reveal a number of functions that have earlier been associated with the gut environment, such as the degradation of conjugated bile acids (bactNOG15678, Choloylglycine hydrolase), cobalamin (vitamin B12) biosynthesis (bactNOG85989, Cobyrinic acid a, c-diamide synthase) [21] or iron acquisition (bactNOG99581, bactNOG38121, Ferric uptake regulator protein) [22, 23], emphasizing the importance of these processes in the gut environment.
BactNOGs represented in one or more of the 23 gut bacterial species in this study were attributed to one of three groups: overrepresented or underrepresented compared to non-gut species (see text for details), or neither over nor underrepresented (indicated by *). Within each of these groups, the number of different BactNOGSs attributed to a functional category is indicated as a percentage of the total number of BactNOGs in the group. Functional category descriptions are short forms of the full descriptions presented in tables “S1” and “S2” Tables.
Not less than 4 of the overrepresented bactNOGs point to an important role of conjugation in the gut environment (bactNOGs 07070, 26309, 08200, 44258), a function that had not been recognized as such in the earlier global approach to the identification of bacterial functions that are important in the gut environment described in [1]. Other functions that had not been recognized in the global approach are "sulfuric ester hydrolase" (sulfatase, bactNOG20561), which may play a role in the foraging of abundant sulfated glycans in the intestine, such as mucins or glycosaminoglycans [24], and "Sortase B" (bactNOG70972) which plays a role in the anchoring of a specific category of bacterial surface proteins [25].
Surprisingly, only 78 bactNOGs were shared among all the genomes in the two genome sets (Fig 3). This is far less than expected when compared with the approximately 250 bacterial protein coding genes that are considered as essential in the Gram-positive model bacterium B. subtilis [26, 27]. Also, certain bactNOGs that represent essential proteins (e.g. certain ribosomal proteins) appeared to be present in only a fraction of the genomes studied (results not shown). A closer look at the catalog of bactNOGs and their functional descriptions revealed that these observations could (in part) be explained by the fact that many functions are represented by several bactNOGs, i.e. several bactNOGs carry identical function descriptions. This is due to the way the non-supervised orthologous groups (NOGs) were constructed [18], where the authors chose for a high resolution, i.e. small precisely defined NOGs containing very similar proteins, to improve accuracy with the consequence that less similar proteins with the same function are attributed to different bactNOGs. We therefore decided to perform a second comparison of our sets of gut and non-gut genomes at the level of "groups of bactNOGs" where we considered bactNOGs with identical function descriptions as one group (in our dataset, a "group of bactNOGs" may contain from 1 to 606 bactNOGs, the latter representing "transcriptional regulator proteins"). By doing so, all the information on bactNOGs that had no description, accounting for 31.2% of the bactNOGs in our dataset, was lost. Other methods of protein clustering and protein family construction that are out of the scope of this study may be instructive in revealing the potential importance of these unclassified bactNOGs beyond the level of the single bactNOG comparisons described above [12, 28, 29].
166 functional “groups of bactNOGs” were shared among all the genomes in the two sets (data not shown). As expected, among these shared groups we find functions involved in DNA and RNA metabolism, transcription, translation, cell envelope, shape and division, as well as energy conversion and metabolism of nucleotides, coenzymes and carbohydrates.
44 groups of bactNOGs were considered as overrepresented in the genomes of the gut Firmicutes (Table 3). These include several functions that had already been identified in the single bactNOG comparisons described above, like "sulfuric ester hydrolase" (sulfatase, bactNOG20561), "sortase B" (bactNOG 70972), "proteins involved in unidirectional conjugation" (6 bactNOGs of which 3 had been identified in individual comparisons) and "proteins involved in conjugation with cellular fusion" (3 bactNOGs, of which 1 had been identified in individual comparisons). The latter two functions are represented in 87 and 74% of the 23 gut species, respectively, as opposed to 35 and 22% of the non-gut species (Table 3).
A new function which only becomes apparent with the grouped bactNOGs approach, and which again was not detected by the global approach in [1], is L-Fucose isomerase (bactNOGs 24561 and 09355), the enzyme catalizing the first step of fucose metabolism, found in 74% of the gut bacteria against 22% of the non-gut bacteria. L-Fucose is highly abundant in the intestine [30], present in dietary polysaccharides such as pectin but also in mucin glycoproteins overlying the intestinal epithelium. It can be cleaved from host glycans by multiple fucosidases produced by gut commensals such as Bacteroides thetaiotaomicron, resulting in a high availability in the intestinal lumen where it can be used as a carbon source by B. thetaiotaomicron itself [31] or by other resident bacteria such as Escherichia coli [32] or Roseburia intestinalis [33].
A certain number of functions that were detected as overrepresented in the single bactNOG comparisons do not appear as such in the analysis by groups of bactNOGs. Some group functions even appear in the list of overrepresented single bactNOGs (S1 Table) as well as in the list of underrepresented single bactNOGs (see below, S2 Table). This is for example the case for the function "Ferric uptake regulator protein" (represented by bactNOGs 38121 and 99581 in the list of overrepresented bactNOGs and by bactNOG31290 in the list of underrepresented bactNOGs). This type of result suggests that closely related gut and non-gut bacteria use distinct proteins for the same function or, alternatively, that the proteins represented by different bactNOGs in spite of identical descriptions exert different functions of which some are more important in the gut environment.
Gut adaptation through functional paucity
The direct comparison of closely related bacteria from the gut and other environments also provides a unique opportunity to focus on underrepresented functions, which may be equally informative of the functioning of the gut ecosystem. In contrast to the overrepresented functions, underrepresented functions included far less bactNOGs with unknown functions (Fig 4, S2 Table). Remarkably, one third of the underrepresented bactNOGs is involved in motility and signal transduction, including several chemotaxis related functions. Another remarkable underrepresented function is the Pur operon repressor protein (bactNOG16918), identified in only one of the 23 gut bacteria as opposed to 21 of the 23 non-gut bacteria. This repressor controls the transcription of the pur operon for purine biosynthetic genes, and its absence would be expected to result in the constitutive transcription of the operon. The ability to synthesize nucleotides was shown to be a prerequisite for successful colonization of the mouse intestine by E. coli [34]. The constitutive expression of genes involved in purine biosynthesis may thus give a competitive advantage during unsteady nucleotide supply in the human gut environment.
The analysis of underrepresented functions by groups of bactNOGs led to the same conclusion as the single-bactNOG comparisons: of the 67 functional groups that are underrepresented in the gut bacteria (Table 4), an astonishing 1/3 appears to be involved in motility and chemotaxis. These functions are, on an average, represented in only 16% of the 23 gut bacteria studied, as opposed to 77% of the non-gut bacteria.
Underrepresentation of secreted proteins in gut bacteria
The predicted proteomes of the strains in this study were analyzed using SurfG+ [19] to predict protein localization (Table 2). This analysis revealed a difference in the numbers of secreted proteins where on an average 3.1% (SEM 0.3) of proteins were predicted to be secreted in our set of gut bacteria as opposed to 4.3% (SEM 0.3) for the non-gut bacteria (not shown). Interestingly, this difference appears to be explained by the presence of relatively low and stable numbers of secreted proteins across differently sized gut bacterial genomes, while in the closely related non-gut bacteria the number of secreted proteins clearly correlates with the total number of proteins encoded in the genome (Fig 5). The numbers of predicted membrane proteins and surface exposed proteins are correlated to the total numbers of encoded proteins in both gut and non-gut bacteria (Fig 5).
The predicted numbers of (A) membrane (mem), (B) potentially surface exposed (pse), and (C) secreted (sec) proteins in a bacterial species are correlated to the total number of encoded proteins (Spearman's rank correlation test, p < 0.01), with the exception of the sec proteins in gut bacteria where no significant correlation is observed (p > 0.2). Red, gut bacteria; blue, non-gut bacteria.
Discussion
The human gut microbiota is increasingly recognized as a major health determining factor. As our knowledge on this microbial community and notably its bacterial component expands, it becomes clear that atypical microbiota compositions, dysbioses, are associated with a growing number of diseases, to an extent that microbiota composition can constitute a "signature" or bio-marker of a specific disease (e.g. [35, 36]). At least for some diseases, experiments in animals convincingly show that an atypical microbiota can be a driving force in the development of disease. In line with these observations, promising results have been reported with the use of fecal microbiota transplantation (in this case the transfer of fecal material from a healthy donor to the intestine of a patient) to improve the symptoms of inflammatory bowel disease in humans [37].
Yet, our knowledge on the bacterial properties that drive homeostasis, the equilibrium of the gut microbial ecosystem including bacteria—host interactions, is still limited. Seen from the bacterial side, which are the genes a bacterium needs to maintain itself in this ecosystem? Which are the bacterial genes that play a role in the functioning of the system, including interactions with the host, as a whole? Several approaches, in silico and through experimental screening, have been and are used to answer these questions. One of the main in silico approaches consists of a global comparison of functions encoded by the gut bacteria and functions encoded by non-gut bacteria, looking for what seems to be gut-specific, as examplified by the study presented in [1]. A possible drawback of this approach, however, is that the taxonomical composition of the gut microbiota and the non-gut reference data set may be largely different. As a consequence, observed differences may in part be due to inherent differences between bacterial taxons that are not necessarily relevant for the comprehension of the gut ecosystem. On the other hand, this approach may fail to detect obvious gut adaptations in a specific taxon if similar functions exist in a different taxon among the non-gut bacteria.
In the present study we therefore used a complementary approach and compared data from selected gut bacteria and closely related non-gut bacteria of the Firmicutes phylum, a procedure that should favor the detection of environment-specific adaptations. We observed a tendency of a relatively low and stable number of predicted secreted proteins across the gut bacteria that was not observed in the non-gut bacteria. It will be interesting to see if this tendency is confirmed when larger numbers of genomes will be analyzed. If so, this could mean that gut bacteria limit the number of secreted proteins, maybe in response to the intestinal flow as this type of proteins could easily become separated from, and thus be of limited advantage to, the secreting bacteria.
We identified several functions that may play an important role in the gut environment but had gone undetected by the global comparison approach described in [1]. For instance, our data strongly suggest that conjugation plays an important role in the gut environment. Conjugation is the most effective mechanism of horizontal gene transfer (HGT) where the exchange of genetic material can occur even between highly divergent bacterial species [38], and our conclusion is in line with earlier evidence of HGT in the gut environment [39–41]. The ability to acquire fitness genes by conjugation may provide gut bacteria with competitive advantages to thrive in this complex environment. Of clinical importance, elevated bacterial conjugation activity in the densely populated gut ecosystem, an environment recognized as a significant reservoir of antibiotic resistances [42], may also play an important role in the spread of antibiotic resistance genes. We further detected sulfatase and L-fucose isomerase as overrepresented functions in the gut bacteria. Sulfatases and their role in the foraging of abundant sulfated glycans in the gut have been described as critical for the fitness of Bacteroides thetaiotaomicron in the gut environment [24], but are far less studied in Firmicutes [24, 43]. Similarly, L-fucose isomerase is involved in the metabolism of L-fucose, a highly abundant sugar in the intestine [30], present in dietary polysaccharides such as pectin but also in mucin glycoproteins overlying the intestinal epithelium.
Together, these examples clearly illustrate the potential of our targeted comparative analysis, focusing on closely related bacteria from different environments, to identify functions that are important in the gut environment. This approach also permits to identify functions that are underrepresented among gut bacteria, an analysis that proved equally informative. We thus observed that an astonishing one third of the underrepresented functions appears to be involved in motility and chemotaxis, representing to our knowledge the first observation of this kind. Bacterial chemotaxis is the phenomenon whereby bacteria direct their movements according to certain chemical stimulants in their environment, and our observation may be explained by the fact that the majority of the bacteria from the “non-gut” set were isolated from water or soil. It is easy to imagine that in these environments it is important for bacteria to move towards the highest concentration of food or other essential molecules, or to flee from poisons. In the gut, the opposite is true as free molecules in transit pass by bacteria that are often adhering to the intestinal surfaces or food particles [44]. An alternative explication may be that these commensal bacteria have been selected for the absence of one of the best known immune modulatory bacterial cell surface proteins, flagellin, that interacts with TLR5 to induce an inflammatory response [45].
The present study can be regarded as a proof of principle demonstrating the potential of taxonomically targeted comparative analyses in the identification of functions that are important in a given ecosystem, in our case the human intestinal tract. The results of these analyses confirmed a number of earlier observations or intuitions about functions that are considered as key functions in the gut environment. The analyses not only identified new functions but also a relative paucity in some other functions, both of which appear to be important in the human gut environment and that, even if experimental evidence is still incomplete, intuitively seem to make sense. These results suggest that the identified "unknown functions" that are found to be overrepresented in the gut bacteria in our analysis are important too and worth further investigating.
In this pilot experiment we limited ourselves to completely sequenced genomes. Without this self-imposed limitation, which is probably not necessary, a wealth of additional data becomes available for analysis. Many more bacterial genomes have been sequenced to near completion since we started this study and ever more are becoming available, including "metagenomic species" genomes that are directly assembled from metagenomic data [46]. The use of these data will allow more robust studies with higher numbers of bacteria. Parallel developments see the pairwise comparison of two human gut microbiota types, typically patients and healthy subjects, rather than comparison of the gut microbiota with bacteria from completely different ecosystems. While not answering exactly the same questions, the different approaches are complementary and should together eventually lead to the unraveling of the critical factors in gut homeostasis that rule our health. The acquired knowledge may subsequently guide our choice of health-beneficial probiotics, screening for desired properties to restore or consolidate homeostasis and avoiding properties that are incompatible with homeostasis.
Supporting Information
S1 Table. BactNOGs overrepresented in gut bacteria.
https://doi.org/10.1371/journal.pone.0159030.s001
(DOCX)
S2 Table. BactNOGs underrepresented in gut bacteria.
https://doi.org/10.1371/journal.pone.0159030.s002
(DOCX)
Acknowledgments
We thank the INRA MIGALE bioinformatics platform (http://migale.jouy.inra.fr) and Metagenopolis (http://www.mgps.eu), INRA Jouy en Josas, for providing computational resources.
Author Contributions
Conceived and designed the experiments: DD ALA MVDG. Performed the experiments: DD ALA AJ. Analyzed the data: DD ALA MVDG. Contributed reagents/materials/analysis tools: EM. Wrote the paper: DD MVDG.
References
- 1. Qin J, Li R, Raes J, Arumugam M, Burgdorf KS, Manichanh C, et al. A human gut microbial gene catalogue established by metagenomic sequencing. Nature. 2010;464(7285):59–65. pmid:20203603; PubMed Central PMCID: PMC3779803.
- 2. Li J, Jia H, Cai X, Zhong H, Feng Q, Sunagawa S, et al. An integrated catalog of reference genes in the human gut microbiome. Nature biotechnology. 2014;32(8):834–41. pmid:24997786.
- 3. Group NHW, Peterson J, Garges S, Giovanni M, McInnes P, Wang L, et al. The NIH Human Microbiome Project. Genome research. 2009;19(12):2317–23. pmid:19819907; PubMed Central PMCID: PMC2792171.
- 4. Human Microbiome Project C. Structure, function and diversity of the healthy human microbiome. Nature. 2012;486(7402):207–14. pmid:22699609; PubMed Central PMCID: PMC3564958.
- 5. Lozupone CA, Stombaugh JI, Gordon JI, Jansson JK, Knight R. Diversity, stability and resilience of the human gut microbiota. Nature. 2012;489(7415):220–30. pmid:22972295; PubMed Central PMCID: PMC3577372.
- 6. Morgan XC, Segata N, Huttenhower C. Biodiversity and functional genomics in the human microbiome. Trends in genetics: TIG. 2013;29(1):51–8. pmid:23140990; PubMed Central PMCID: PMC3534939.
- 7. Turnbaugh PJ, Hamady M, Yatsunenko T, Cantarel BL, Duncan A, Ley RE, et al. A core gut microbiome in obese and lean twins. Nature. 2009;457(7228):480–4. pmid:19043404; PubMed Central PMCID: PMC2677729.
- 8. Gill SR, Pop M, Deboy RT, Eckburg PB, Turnbaugh PJ, Samuel BS, et al. Metagenomic analysis of the human distal gut microbiome. Science. 2006;312(5778):1355–9. pmid:16741115; PubMed Central PMCID: PMC3027896.
- 9. Tasse L, Bercovici J, Pizzut-Serin S, Robe P, Tap J, Klopp C, et al. Functional metagenomics to mine the human gut microbiome for dietary fiber catabolic enzymes. Genome research. 2010;20(11):1605–12. pmid:20841432; PubMed Central PMCID: PMC2963823.
- 10. Samuel BS, Hansen EE, Manchester JK, Coutinho PM, Henrissat B, Fulton R, et al. Genomic and metabolic adaptations of Methanobrevibacter smithii to the human gut. Proceedings of the National Academy of Sciences of the United States of America. 2007;104(25):10643–8. pmid:17563350; PubMed Central PMCID: PMC1890564.
- 11. Wexler HM. Bacteroides: the good, the bad, and the nitty-gritty. Clinical microbiology reviews. 2007;20(4):593–621. pmid:17934076; PubMed Central PMCID: PMC2176045.
- 12. Ellrott K, Jaroszewski L, Li W, Wooley JC, Godzik A. Expansion of the protein repertoire in newly explored environments: human gut microbiome specific protein families. PLoS computational biology. 2010;6(6):e1000798. pmid:20532204; PubMed Central PMCID: PMC2880560.
- 13. Pons N, Batto JM, Ehrlich SD, Renault P. Development of software facilities to characterize regulatory binding motifs and application to streptococcaceae. Journal of molecular microbiology and biotechnology. 2008;14(1–3):67–73. pmid:17957112.
- 14. Gouy M, Guindon S, Gascuel O. SeaView version 4: A multiplatform graphical user interface for sequence alignment and phylogenetic tree building. Molecular biology and evolution. 2010;27(2):221–4. pmid:19854763.
- 15. Edgar RC. MUSCLE: multiple sequence alignment with high accuracy and high throughput. Nucleic acids research. 2004;32(5):1792–7. pmid:15034147; PubMed Central PMCID: PMC390337.
- 16. Guindon S, Dufayard JF, Lefort V, Anisimova M, Hordijk W, Gascuel O. New algorithms and methods to estimate maximum-likelihood phylogenies: assessing the performance of PhyML 3.0. Systematic biology. 2010;59(3):307–21. pmid:20525638.
- 17. Letunic I, Bork P. Interactive Tree Of Life v2: online annotation and display of phylogenetic trees made easy. Nucleic acids research. 2011;39(Web Server issue):W475–8. pmid:21470960; PubMed Central PMCID: PMC3125724.
- 18. Powell S, Szklarczyk D, Trachana K, Roth A, Kuhn M, Muller J, et al. eggNOG v3.0: orthologous groups covering 1133 organisms at 41 different taxonomic ranges. Nucleic acids research. 2012;40(Database issue):D284–9. pmid:22096231; PubMed Central PMCID: PMC3245133.
- 19. Barinov A, Loux V, Hammani A, Nicolas P, Langella P, Ehrlich D, et al. Prediction of surface exposed proteins in Streptococcus pyogenes, with a potential application to other Gram-positive bacteria. Proteomics. 2009;9(1):61–73. pmid:19053137.
- 20. Koonin EV. Orthologs, paralogs, and evolutionary genomics. Annual review of genetics. 2005;39:309–38. pmid:16285863.
- 21. Nicholson JK, Holmes E, Kinross J, Burcelin R, Gibson G, Jia W, et al. Host-gut microbiota metabolic interactions. Science. 2012;336(6086):1262–7. pmid:22674330.
- 22. Kortman GA, Boleij A, Swinkels DW, Tjalsma H. Iron availability increases the pathogenic potential of Salmonella typhimurium and other enteric pathogens at the intestinal epithelial interface. PloS one. 2012;7(1):e29968. pmid:22272265; PubMed Central PMCID: PMC3260200.
- 23. Werner T, Wagner SJ, Martinez I, Walter J, Chang JS, Clavel T, et al. Depletion of luminal iron alters the gut microbiota and prevents Crohn's disease-like ileitis. Gut. 2011;60(3):325–33. pmid:21076126.
- 24. Benjdia A, Martens EC, Gordon JI, Berteau O. Sulfatases and a radical S-adenosyl-L-methionine (AdoMet) enzyme are key for mucosal foraging and fitness of the prominent human gut symbiont, Bacteroides thetaiotaomicron. The Journal of biological chemistry. 2011;286(29):25973–82. pmid:21507958; PubMed Central PMCID: PMC3138274.
- 25. Marraffini LA, Dedent AC, Schneewind O. Sortases and the art of anchoring proteins to the envelopes of gram-positive bacteria. Microbiology and molecular biology reviews: MMBR. 2006;70(1):192–221. pmid:16524923; PubMed Central PMCID: PMC1393253.
- 26. Juhas M, Reuss DR, Zhu B, Commichau FM. Bacillus subtilis and Escherichia coli essential genes and minimal cell factories after one decade of genome engineering. Microbiology. 2014;160(Pt 11):2341–51. pmid:25092907.
- 27. Kobayashi K, Ehrlich SD, Albertini A, Amati G, Andersen KK, Arnaud M, et al. Essential Bacillus subtilis genes. Proceedings of the National Academy of Sciences of the United States of America. 2003;100(8):4678–83. pmid:12682299; PubMed Central PMCID: PMC153615.
- 28. Hunter S, Apweiler R, Attwood TK, Bairoch A, Bateman A, Binns D, et al. InterPro: the integrative protein signature database. Nucleic acids research. 2009;37(Database issue):D211–5. pmid:18940856; PubMed Central PMCID: PMC2686546.
- 29. Punta M, Coggill PC, Eberhardt RY, Mistry J, Tate J, Boursnell C, et al. The Pfam protein families database. Nucleic acids research. 2012;40(Database issue):D290–301. pmid:22127870; PubMed Central PMCID: PMC3245129.
- 30. Robbe C, Capon C, Coddeville B, Michalski JC. Structural diversity and specific distribution of O-glycans in normal human mucins along the intestinal tract. The Biochemical journal. 2004;384(Pt 2):307–16. pmid:15361072; PubMed Central PMCID: PMC1134114.
- 31. Xu J, Bjursell MK, Himrod J, Deng S, Carmichael LK, Chiang HC, et al. A genomic view of the human-Bacteroides thetaiotaomicron symbiosis. Science. 2003;299(5615):2074–6. pmid:12663928.
- 32. Pacheco AR, Curtis MM, Ritchie JM, Munera D, Waldor MK, Moreira CG, et al. Fucose sensing regulates bacterial intestinal colonization. Nature. 2012;492(7427):113–7. pmid:23160491; PubMed Central PMCID: PMC3518558.
- 33. Scott KP, Martin JC, Campbell G, Mayer CD, Flint HJ. Whole-genome transcription profiling reveals genes up-regulated by growth on fucose in the human gut bacterium "Roseburia inulinivorans". Journal of bacteriology. 2006;188(12):4340–9. pmid:16740940; PubMed Central PMCID: PMC1482943.
- 34. Vogel-Scheel J, Alpert C, Engst W, Loh G, Blaut M. Requirement of purine and pyrimidine synthesis for colonization of the mouse intestine by Escherichia coli. Applied and environmental microbiology. 2010;76(15):5181–7. pmid:20562286; PubMed Central PMCID: PMC2916491.
- 35. Le Chatelier E, Nielsen T, Qin J, Prifti E, Hildebrand F, Falony G, et al. Richness of human gut microbiome correlates with metabolic markers. Nature. 2013;500(7464):541–6. pmid:23985870.
- 36. Qin N, Yang F, Li A, Prifti E, Chen Y, Shao L, et al. Alterations of the human gut microbiome in liver cirrhosis. Nature. 2014;513(7516):59–64. pmid:25079328.
- 37. Anderson JL, Edney RJ, Whelan K. Systematic review: faecal microbiota transplantation in the management of inflammatory bowel disease. Alimentary pharmacology & therapeutics. 2012;36(6):503–16. pmid:22827693.
- 38. Grohmann E, Muth G, Espinosa M. Conjugative plasmid transfer in gram-positive bacteria. Microbiology and molecular biology reviews: MMBR. 2003;67(2):277–301, table of contents. pmid:12794193; PubMed Central PMCID: PMC156469.
- 39. Nicolas P, Bessieres P, Ehrlich SD, Maguin E, van de Guchte M. Extensive horizontal transfer of core genome genes between two Lactobacillus species found in the gastrointestinal tract. BMC evolutionary biology. 2007;7:141. pmid:17708761; PubMed Central PMCID: PMC1994166.
- 40. Smillie CS, Smith MB, Friedman J, Cordero OX, David LA, Alm EJ. Ecology drives a global network of gene exchange connecting the human microbiome. Nature. 2011;480(7376):241–4. pmid:22037308.
- 41. Stecher B, Denzler R, Maier L, Bernet F, Sanders MJ, Pickard DJ, et al. Gut inflammation can boost horizontal gene transfer between pathogenic and commensal Enterobacteriaceae. Proceedings of the National Academy of Sciences of the United States of America. 2012;109(4):1269–74. pmid:22232693; PubMed Central PMCID: PMC3268327.
- 42. Francino MP. Antibiotics and the Human Gut Microbiome: Dysbioses and Accumulation of Resistances. Front Microbiol. 2016;6:1543. pmid:26793178; PubMed Central PMCID: PMCPMC4709861.
- 43. Berteau O, Guillot A, Benjdia A, Rabot S. A new type of bacterial sulfatase reveals a novel maturation pathway in prokaryotes. The Journal of biological chemistry. 2006;281(32):22464–70. pmid:16766528.
- 44. Macfarlane S, Macfarlane GT. Composition and metabolic activities of bacterial biofilms colonizing food residues in the human gut. Applied and environmental microbiology. 2006;72(9):6204–11. pmid:16957247; PubMed Central PMCID: PMC1563644.
- 45. Hayashi F, Smith KD, Ozinsky A, Hawn TR, Yi EC, Goodlett DR, et al. The innate immune response to bacterial flagellin is mediated by Toll-like receptor 5. Nature. 2001;410(6832):1099–103. pmid:11323673.
- 46. Nielsen HB, Almeida M, Juncker AS, Rasmussen S, Li J, Sunagawa S, et al. Identification and assembly of genomes and genetic elements in complex metagenomic samples without using reference genomes. Nature biotechnology. 2014;32(8):822–8. pmid:24997787.