Figures
Abstract
The human intestine hosts a complex bacterial community that plays a major role in nutrition and in maintaining human health. A functional metagenomic approach was used to explore the prebiotic breakdown potential of human gut bacteria, including non-cultivated ones. Two metagenomic libraries, constructed from ileum mucosa and fecal microbiota, were screened for hydrolytic activities on the prebiotic carbohydrates inulin, fructo-oligosaccharides, xylo-oligosaccharides, galacto-oligosaccharides and lactulose. The DNA inserts of 17 clones, selected from the 167 hits that were identified, were pyrosequenced in-depth, yielding in total 407, 420 bp of metagenomic DNA. From these sequences, we discovered novel prebiotic degradation pathways containing carbohydrate transporters and hydrolysing enzymes, for which we provided the first experimental proof of function. Twenty of these proteins are encoded by genes that are also present in the gut metagenome of at least 100 subjects, whatever are their ages or their geographical origin. The sequence taxonomic assignment indicated that still unknown bacteria, for which neither culture conditions nor genome sequence are available, possess the enzymatic machinery to hydrolyse the prebiotic carbohydrates tested. The results expand the vision on how prebiotics are metabolized along the intestine, and open new perspectives for the design of functional foods.
Citation: Cecchini DA, Laville E, Laguerre S, Robe P, Leclerc M, Doré J, et al. (2013) Functional Metagenomics Reveals Novel Pathways of Prebiotic Breakdown by Human Gut Bacteria. PLoS ONE 8(9): e72766. https://doi.org/10.1371/journal.pone.0072766
Editor: Hauke Smidt, Wageningen University, Netherlands
Received: March 19, 2013; Accepted: July 13, 2013; Published: September 16, 2013
Copyright: © 2013 Cecchini 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.
Funding: This work was carried out with the financial support of Institut National de la Recherche Agronomique (INRA), and of the French National Agency for Research, in the frame of the ANR/DEDD/PNRA/PROJ/200206-01-01 project. ICEO is supported by grants from the Region Midi-Pyrénées, France, the European Regional Development Fund, and INRA, France. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
Competing interests: The authors confirm the affiliation of Patrick Robe, one of the authors, to the LibraGenSA commercial company, and confirmed in the Competing Interests section of the online manuscript form that this does not alter the authors' adherence to all the PLOS ONE policies on sharing data and materials: “LibraGenSA is a commercial company. However, in the work described in this manuscript, LibraGenSA acted as an technological expert in the field of construction of metagenomic libraries. This does not alter the authors' adherence to all the PLOS ONE policies on sharing data and materials.
Introduction
The human gut hosts a complex microorganism community that is dominated by bacterial phylotypes belonging to Firmicutes, Bacteroidetes and Actinobacteria. The bacterial density increases along the digestive tract to reach 107 and 1011 bacteria mg−1 of intestinal content in the distal parts of the small intestine (ileum) and of the large intestine (colon), respectively [1]–[4]. Collectively, gut bacterial genomes contain 150-fold more genes than the human genome equivalent [5], [6], [7], and code for many metabolic functions that make the gut microbiota a “virtual organ” which provides the host with missing metabolic capabilities that are crucial for maintaining its health and welfare [8], [9]. In particular, a huge diversity of enzymes involved in carbohydrate transport and metabolism are encoded by the gut metagenome, as revealed by Gill et al. [10]. Indeed, the intestinal microbiota contributes to host nutrition by harvesting energy from dietary fibers, these carbohydrates that are not digested in the upper part of the digestive tract because the human genome does not encode adequate carbohydrate active enzymes (CAZymes) [11]. Fibers are fermented by gut bacteria to produce organic acids (e.g., lactate, pyruvate, succinate), gases (e.g., H2, H2S, CO2 and CH4), and also short chain fatty acids (SCFA) (e.g., acetate, propionate and butyrate), these compounds contributing to the metabolic balance of the host [12]. Non-digestible carbohydrates include the natural constituents of vegetables, cereals, leguminous seeds and fruits, but also prebiotic oligosaccharides and polysaccharides [13]. Prebiotics are defined as food ingredients that fit to the three following criteria: 1) resistance to gastric acidity, to hydrolysis by mammalian enzymes, and to gastrointestinal absorption; 2) fermentation by intestinal microbiota; and 3) selective stimulation of the growth and/or activity of those intestinal bacteria that contribute to health and well-being [14], in particular bifidobacteria. For this last reason, the world market for prebiotics has grown rapidly in the last three decades [15], focusing on the production of compounds with established prebiotic effects (inulin, fructo-oligosaccharides, galacto-oligosaccharides and lactulose), as well as development and commercialization of other dietary carbohydrates, like resistant starch [16], xylo-oligosaccharides [17], gluco-oligosaccharides [18], [19], polydextrose [20], lactosucrose [21], pectin-derived [22] or soybean oligosaccharides [23].
However, the health benefits of prebiotics are controversial. Indeed, the causal relationship between increasing the number of bifidobacteria in the intestine and improving health status has not been established [24]. There is thus an urgent need to re-evaluate the impact of prebiotics on the intestinal microbiota as a whole, and not only on few cultivated Bifidobacterium and Lactobacillus species of whose growth is stimulated by prebiotics [25], [26], but which do not represent the ecosystem complexity. Several studies showed that other genera can be stimulated by prebiotics, directly or through cross-feeding of prebiotic breakdown products, such as Bacteroides [27], [28], Faecalibacterium [29], Roseburia, Ruminococcus and Eubacterium [30]–[32].
However, very little data are available regarding the enzymes that are directly and specifically involved in prebiotic metabolism by gut bacteria. Only a few studies based on sequence functional annotation [33], [34], transcriptomics [35]–[37], proteomics [38] or recombinant enzyme expression and in vitro characterization [26], [39]–[42] reported on the identification of genes and proteins involved in inulin, fructo-, xylo- and galacto-oligosaccharide metabolism by some cultivated Bifidobacterium, Lactobacillus Roseburia, Faecalibacterium and Bacteroides strains. Nevertheless, these strains constitute only a minor fraction of the gut microbial diversity.
In this paper, we present a screening-based functional metagenomics study to identify genes encoding key prebiotic degrading enzymes from both cultivated and non-cultivated gut bacteria. Both the ileum mucosa and fecal microbiota were targeted, to highlight the specific potential of prebiotic metabolization in two gut compartments and locations, which, to our knowledge, has never been studied. Our strategy was first based on the design of functional screens to guide the in-depth sequencing of metagenomic fragments encoding enzymes that hydrolyse five compounds sold as prebiotics for human consumption. Then, the occurrence in the human gut metagenome of the fished genes, obtained from two individuals, was studied in the light of the latest metagenomic data released from the Human Microbiome Project, to identify the genes that are shared by numerous individuals, whatever are their age or their geographical origin [43]. The results provided new insights on metabolism of inulin, fructo-oligosaccharides, lactulose, galacto-oligosaccharides, and xylo-oligosaccharides.
Results and Discussion
Screening the human gut metagenome for prebiotic-degrading activities
Twenty thousand clones of each of the two E. coli fosmid libraries constructed from fecal and ileum mucosa microbiota (named F and I libraries respectively), covering in total 1.4 Gb of metagenomic sequence, were screened for clone ability to degrade prebiotic compounds of various structures (Table 1). E. coli was chosen as the recombinant host because it was previously shown to allow expression of genes belonging to bacteria that are distantly related from a taxonomical point of view, like Bifidobacteria [44].
First, a high-throughput primary search for clones growing on xylo-oligosaccharides, lactulose and fructo-oligosaccharides was performed using solid selective media containing the oligosaccharides as sole carbon source. Because the expression host EPI100 E. coli strain used to construct the library was unable to metabolize galactose, another strategy presented below has been applied to search for galacto-oligosaccharide degrading clones. The 120, 000 growth assays that were carried out allowed the identification of 11 and 49 hit clones from the F and I libraries, respectively. Hit yields varied between 0.2 and 1.35‰, depending on the library and on the oligosaccharide structure (Table 1).
In a second step, we validated and discriminated the prebiotic oligosaccharide hydrolysis ability of selected clones. For this purpose we performed HPAEC-PAD analysis of the products obtained after incubation of the cytoplasmic extracts with the prebiotic compound. All the clones growing on media containing xylo-oligosaccharides, fructo-oligosaccharides and lactulose were able to release mono-saccharides from these substrates (Table 2, Figure S1). In addition, this analysis showed that the clones selected by using fructo-oligosaccharides were also able to degrade inulin. Regarding the search for galacto-oligosaccharide degrading clones, the 14 I library clones found on lactulose were tested for galacto-oligosaccharide hydrolysis. Eleven clones were able to degrade oligosaccharides that were present in the commercial galacto-oligosaccharide preparation. The 107 clones which were identified in a previous study by screening 136, 000 clones from the F library for AZCL-galactan breakdown [44] were tested for galacto-oligosaccharide hydrolysis. All of them were able to degrade galacto-oligosaccharides (Table 1). Depending on the HPAEC-PAD profiles of their prebiotic degradation products (amount and polymerisation degree of the hydrolysis products) the hit clones were classified into 19 clusters. The 17 clones that were representative of clusters containing the most efficient clones were chosen for subsequent pyrosequencing of metagenomic DNA insert (Table 2).
This study provides experimental evidence of the prebiotic degrading potential of non-cultivated gut bacteria, in particular those from ileum. This hydrolysis ability correlates with sequence-based metagenomic results which revealed that functions related to uptake and metabolism of relatively simple carbohydrates (mono- and di-saccharides) are well represented in the small intestine microbiota (6, 7). However, it also proves that ileum bacteria are also able to breakdown longer oligosaccharides (for instance fructo-oligosaccharides of polymerisation degree 5 or higher, like those present in inulin). Numerous prebiotics contain the same glycoside residues and the same glycosidic linkages as plant cell wall polysaccharides (cellulose, hemicellulose, pectin, galactan) or polysaccharides involved in plant energy storage (starch and inulin). Indeed, they are industrially produced by partial hydrolysis of these polysaccharides that also constitute natural dietary fibres and the main carbon sources for gut colonizing bacteria. Unlike the other prebiotics tested in this study, lactulose is not produced from plants; this is an artificial oligosaccharide that is chemically or enzymatically synthesized from lactose [45]. Lactulose is a structural isomer of lactose, a disaccharide that is present in most mammals' milk, including human [46] and bovine milks [47], and of galacto-disaccharides that derives from the enzymatic hydrolysis of the plant cell wall polysaccharide galactan. We thus propose that at least a part of the metabolic pathways of the natural dietary components lactose and galacto-oligosaccharides of gut bacteria is also responsible for the breakdown of the synthetic prebiotic lactulose.
Identification of gene clusters assigned to prebiotic catabolism
Pyrosequencing of metagenomic DNA from the 17 selected clones was performed with a high sequencing depth (between 45X and 171X), ensuring reliable sequence assembly. For each clone, one contig sizing between 13, 000 and 39, 000 bp was obtained, except for clones 9 and 10 which generated two contigs each. Partial or complete sequence redundancy was observed for 4 clones from the I library and 6 clones from the F library, respectively (Table 3). As most of these redundant sequences were not assignable from a taxonomic point of view (based on sequence homology analysis with sequenced genomes), it was impossible to determine if their bacterial origin corresponds to over-represented species in the samples used for library construction.
In total, 213, 780 bp (containing 172 ORFs) and 193, 640 bp (containing 158 ORFs) of metagenomic DNA were analysed for F and I libraries, respectively. On average, 26% of the detected ORFs were assigned to the G COG cluster, corresponding to proteins that are predicted to be involved in Carbohydrate Transport and Metabolism (Tables 3 and S1). The dominance of G COG cluster, which was not observed in the randomly sequenced human gut metagenome [5], [48], [49], reflects as previously described [44], the power of screening-based strategies to guide the sequencing of DNA fragments that are rich in genes of interest.
Among these 330 ORFs, 38 were annotated as CAZyme encoding genes. Thirty-three of these predicted CAZymes contained a glycoside-hydrolase (GH) module (Tables 3, S1 and S2), others being esterases and glycosyl-transferases, which are not involved in prebiotic breakdown.
Moreover, on 11 contigs, the putative GHs are encoded by operon-like multigenic systems encoding proteins involved in carbohydrate binding, metabolism and transport (belonging to the ABC transport family, to PTS systems or to SusD and SusC families), surrounded by putative transcriptional regulators (Figure 1, Table S1). Five of the discovered GHs share more than 90% sequence identity with enzymes from Bifidobacterium species, which were already shown to be involved in prebiotic breakdown: one GH32 (contig F2) involved in fructo-oligosaccharide degradation [50]–[52], three GH43, GH8 and GH120 (contig F4), synergistically involved in the metabolization of xylo-oligosaccharides [42], [53], [54], and the last two predicted enzymes assigned to families GH2 and GH42 (contig F6), involved in the hydrolysis of galacto-oligasaccharides [55]. For all the other pathways of carbohydrate transport and catabolism highlighted here (encoded by contigs F1, F3, F5, I7, I8, I9, I10, I11, I14, I12, and I13), the present screening results constitute the first biochemical evidence of their ability to participate to prebiotic metabolization. Moreover, these results contribute to indicate functions of putative CAZyme sequences belonging to families which are often multispecific, and which contain only between 3 (for GH2) and 11% (for GH32) of characterized enzymes.
The operon-like multigenic systems encoding proteins involved in carbohydrate binding, metabolism and transport surrounded by putative transcriptional regulators are shown. Genes are represented by arrows. Under the operon-like structures are reported the strains for which perfect synteny was identified between part of their genome and the present metagenomic sequences.
Which are the bacteria that possess these prebiotic catabolic machineries?
To get new insights into the relationships between bacteria taxonomy and their role in prebiotic catabolism, the origin of the metagenomic DNA sequences was predicted. All the contigs came from bacteria of which the genome has not yet been sequenced, as none of the contig sequences were perfectly identical to sequenced genomes. However, for the contigs F2, F4, F5, F6, I10, I11, I13 and I14, the high sequence similarity with reference genomes made it possible to identify at the species level the nearest homolog of the gut bacteria of from which came the metagenomic inserts (Table 3). A near perfect synteny was observed with the genomic sequences of Bifidobacterium longum NCC2705 (contigs F2 and F6), Bifidobacterium adolescentis ATCC 15703 (contig F4), Eubacterium rectale ATCC 33656 (contig I10a), Streptococcus thermophilus LMD9 and LMG 18311 (contigs I11 and I14) (Figure 1 and Table S1). Regarding the contig F5, synteny with Bacteroides vulgatus ATCC 8482 was interrupted in the region carrying the GHs proposed to be responsible for the screened activity. Synteny between contig I13 and the Faecalibacterium prausnitzii KLE1255 genome was also highly perturbed. The gut bacteria from which came the DNA inserts F2, F4, F5, F6, I10, I11, I13 and I14 are thus probably very near, from a taxonomical point of view, of strains which were already described as being probiotics, carbohydrate fermenters, or, as being able to metabolize xylo-, fructo-, galacto-oligosaccharides or inulin [22], [30], [32], [33], [55]–[60]. However, as mentioned upper, apart from those of Bifidobacterium longum NCC2705, Bifidobacterium adolescentis ATCC 15703 and Streptococcus thermophilus LMD9, the enzymes involved in prebiotic degradation were not experimentally identified to date. The present metagenomic sequences are thus highly useful to pinpoint the corresponding gene machinery in the genome of cultivated strains. This is of prime importance for the design of innovative synbiotics, these functional food mixtures containing both a probiotic strain and the prebiotic compound that it is able to metabolize.
Regarding this aspect, clone I13, which was able to degrade lactulose and galacto-oligosaccharides, is of particular interest. Indeed, a cluster containing 17 putative genes (ORFs CCG34912 to CCG34928) were annotated as being involved in cobalamin (B12 vitamin) synthesis (Table S1). This vitamin plays a key role in the brain and in the nervous system [61]. Its absorption occurs by receptor-mediated endocytosis in the terminal ileum [62]. Cobalamin was already shown to be synthesized by bacteria present in the distal gut, in particular Lactobacillus reuteri strains that are marketed as probiotics and commonly used in the food industry [63]. Here, we showed that i) an ileum colonizing strain possesses both the machineries for lactulose/galacto-oligosaccharides hydrolysis and for putative cobalamin synthesis, ii) proteins which would be involved in prebiotic degradation and in vitamin synthesis share more than 90% identity with proteins of Faecalibacterium cf. prausnitzii KLE1255. These results allow enlarging the panel of prebiotic compounds that could stimulate the growth of Faecalibacterium prausnitzii species that could beneficially impact health [29], [64]. In particular, it would be interesting to test both the ability of Faecalibacterium cf. prausnitzii KLE1255 to produce vitamin B12 and to hydrolyse lactulose and galacto-oligosaccharides, to evaluate its interest as a probiotic strain in synbiotics.
No significant sequence homology with lactobacilli genomic sequences was found, even if these bacteria are reported, with bifidobacteria, to be the most common prebiotic targets. This is probably due to the lower abundance of Lactobacillus species in the human gut microbiota, compared to Bifidobacterium species [2], [5], and to the fact that the gut metagenome diversity was not sufficiently covered by 40, 000 clones of 30 to 40 kbp each to isolate such low abundant sequences. Therein, based on 16S rRNA gene sequencing results, no Lactobacillus bacteria was found in the microbiota sample which was used to construct the F library [44], while Bifidobacterium bacteria represented 1% of this sample. No Lactobacillus bacteria was found in the ileum mucosa sample as well, but in this case, PCR based 16S rRNA gene taxonomic characterization was based only on 96 sequences, which is insufficient to be statistically reliable. One should thus not exclude that either cloning [65] or expression bias [66] may also have prevented the finding of lactobacilli sequences in this study, resulting from the use of E. coli as expression host.
Finally, the metagenomic inserts of clones F1, F3, I7, I8, I9, I12 were impossible to assign accurately, as their sequences were too distant from any available sequenced genome. However, Megan analysis of sequence homologies using low stringent criteria revealed that 5 of these sequences probably belong to bacteria of Clostridiales and Bacteroidales orders (Table 3). Moreover, the clone I9 metagenomic insert generated 2 contigs that were impossible to assemble. The first of these contigs (contig I9a), that encoded for a fructo-oligosaccharide metabolic gene cluster and a putative transposase (Figure 1, Table S1), presents a perfect synteny with a part of the Dorea longicatena DSM 13814 genome, while the second one (contig I9b) was highly homolog to a part of Clostridium nexile DSM 1788 genome. The PCR amplification of one gene in each contig allowed us to confirm that these two contigs came from one single metagenomic insert of one isolated clone. Dorea longicatena is one of the dominant species in human gut [5], and the DSM 13814 strain was previously shown to grow on inulin and on fructo-oligosaccharides [67], but we found no report on its eventual beneficial effects on human health. This chimeric construct could be an illustration of the human gut metagenome plasticity, mediated through horizontal gene transfers (HGTs) between nearly or distantly related bacteria, as previously described [44], [68]. Overall, these results highlight that growth of still uncharacterized commensal bacteria, of which the beneficial or deleterious effects for human health are unknown, could be stimulated by prebiotic oligosaccharides.
Occurrence in the human gut metagenome of genes assigned to prebiotic breakdown pathways
Finally, we compared our metagenomic sequences, obtained from our 2 different subjects, with three different gene catalogues obtained from the random sequencing of fecal metagenomes, sampled from 162 subjects of MetaHIT cohorts [5], [69]; 139 subjects of NIH-HMP cohort [70] and other 110 subjects from different cultures, ages and families [43]. Striking results were obtained. Indeed, 87% of the genes identified in this study were at least 90% identical to genes detected in the gut metagenomic sequences of at least one subject of these cohorts (62% of genes being 100% identical), and 53% were at least 90% identical to sequences obtained from at least 20 individuals (Tables 3 and S1).
Anyway, the 20 genes constituting the operon-like gene clusters encoding prebiotic degrading enzymes obtained from clones F1 (non assigned), F3 (non assigned), F5 (assigned to Bacteroides vulgatus) and I10 (assigned to Eubacterium rectale) had highest occurences, whatever the targeted cohort. Indeed, they all are at least 90% identical to genes found in at least 100 different individuals. These results indicate that the present function-driven metagenomics study allowed the discovery of some of the major prebiotic metabolization machineries along the human digestive tract.
Concluding remarks
In this study, we were able, by using a functional metagenomics approach, to directly access key enzymes involved in prebiotic metabolism by gut bacteria, including non-cultivated ones. We provided the first experimental evidence of the prebiotic hydrolytic potential in the ileum, and showed that in this gut compartment, bacteria are not only able to breakdown short carbohydrates like di-saccharides, but also longer ones, thanks to both exo- and endo-acting glycoside-hydrolases (like for instance polysaccharide degraders classified in the GH10 family).
Our results are consistent with the known ability of bifidobacteria to hydrolyse prebiotics, and allowed to evidence the prebiotic metabolization potential of strains currently used as probiotics or assumed to have health-promoting properties, which could thus be used to design novel functional foods like symbiotics. However, we also showed the high occurrence in human gut of still unknown bacteria possessing the enzymatic machinery to breakdown these compounds. Their growth could thus also be selectively stimulated by diet supplementation with prebiotics, thus generating effects on human health that remain to be determined. Prebiotic compounds, that were previously thought to direct a very specific response on human gut microbiota, exhibit a broader spectrum of action regarding target species but also gastro intestinal tract location. However, metagenome profiling, comparative metatranscriptomics and metaproteomics studies would have to be performed to evaluate the involvement level of the proteins that were identified here in real physiological conditions, with and without diet supplementation with prebiotics.
Materials and Methods
Metagenomic library construction
Two different gut microbiota were sampled as previously described [44], [71], in the frame of a study registered as a clinical trial (http://clinicaltrials.gov/, Calibrated Diets and Human Intestinal Microflora NCT00639561) and carried out at the Center for Clinical Investigation (CCI) of Grenoble University Hospital (France). The first one was a fecal sample of a healthy 30 years old male, who followed a vegetarian and fish-eating diet for twelve years before sampling. The distal ileum sample was obtained from a 51 years old male patient undergoing colonoscopy and surgery for lower colon cancer suspicion, after he has been submitted to a cleansing preparation. A segment of 2 cm2 was obtained, immediately frozen and kept at −80°C until processing. Both individuals showed an average body mass index. They did not eat any functional food such as prebiotics or probiotics, nor did receive any antibiotics or other drugs during 6 months before sampling. The sampling protocols were approved by the local ethics committee “Comité de Protection des Personnes Sud Est V” (Ref 07-CHUG-21, composed of B.Habozit, J.L.Crakowski, J.Juge, J.Grunwald, E.Svhan and E.Fontaine) and informed written consent was obtained from the subjects before sampling.
Metagenomic DNA was extracted, and fragments ranging in size from 30 to 40 kb DNA were isolated and cloned into pCC1FOS fosmid (Epicentre Technologies). EPI100 E. coli cells were then transfected to obtain a 156, 000 clone library from the fecal sample [44], and a 20, 000 clone library from the ileum sample [71]. Recombinant clones were transferred to 384-well microtiter plates containing Luria-Bertani medium, supplemented with 12.5 mg.l−1 chloramphenicol and 8% glycerol. After 22 h of growth at 37°C without any agitation, the plates were stored at −80°C.
Carbohydrate sources
Commercial prebiotic oligosaccharides and polysaccharides used for screening were the following (Table 1): xylo-oligosaccharides (Iro Taihe International, China), fructo-oligosaccharides (Actilight-FOS, Beghin Meiji, France), inulin (from dahlia tuber, Sigma, Germany), galacto-oligosaccharides (Vivinal-GOS, FrieslandFoods Domo, The Netherlands), lactulose (TEVA, France).
Xylo-oligosaccharide preparations were purified by preparative chromatography to remove xylose, as following: 250 ml of a 50% (w/v) carbohydrate solution was loaded in a preparative-column (bed volume 7.8l, diameter 8 cm, column length 100 cm) filled with an ionic exchange resin (Amberlite CR1320K). Oligosaccharides were eluted by using MilliQ H2O and the carbohydrate content of each fraction was analysed by high-performance anion-exchange chromatography coupled to pulsed amperometric detection (HPAEC-PAD) as described in the following section. Fractions containing carbohydrates of polymerisation degree ≥2 were pooled, and the resulting solution was freeze-dried with an SMH 15 freeze-dryer (Société Nouvelle Usifroid, France). The final purity of the oligosaccharide powder was determined by HPAEC-PAD.
High throughput functional screening
Twenty thousand clones from each library were gridded onto 22×22 cm trays containing minimum solid medium (Na2HPO4 6 g.l−1, KH2PO4 3 g.l−1, NH4Cl 1 g.l−1, NaCl 0.5 g.l−1, MgSO4 0.12 g.l−1, CaCl2·2H2O 0.015 g.l−1, FeSO4·7H2O 0.0042 g.l−1, thiamine hydrochloride 0.0005 g.l−1, leucine 0.04 g.l−1, agar type E 15 g.l−1 and chloramphenicol 12.5 mg.l−1) supplemented with prebiotic oligosaccharides as sole carbon source, using a QPixII colony picker (Genetix), with a density of 2, 304 clones per tray. Oligosaccharides were sterilized by membrane filtration, using 0.20 µm Minisart filters (Sartorius Stedim Biotech GmbH, Germany) and added aseptically to the medium. The final concentration of carbohydrates was 0.5% (w/v) for xylo-oligosaccharides and lactulose, and 1% (w/v) for fructo-oligosaccharides. The trays were incubated at 37°C between 1 and 3 weeks, depending on the time necessary to visualize the growth of hit clones. It was checked that the EPI100 strain harboring the empty pCC1FOS vector is unable to grow on these selective media.
The hit clones were then streaked on agar LB medium supplemented with chloramphenicol 12.5 mgl−1, and the resulting isolated clones were stored at −80°C in liquid LB medium supplemented with chloramphenicol 12.5 mg.l−1 and glycerol 15% (w/v).
Secondary screening
The prebiotic hydrolytic activities of the hit clones isolated from primary screening were further examined by HPAEC-PAD. The clones were grown at 37°C in 5 ml LB medium supplemented with chloramphenicol 12.5 mgl−1, with orbital shaking at 120 r.p.m. After 24 h, cells were harvested by centrifuging for 5 min at 5, 000 r.p.m., re-suspended in 1 ml potassium phosphate buffer 50 mM pH 7, containing 0.5 gl−1 lysozyme and incubated at 37°C for 1h. Cell lysis was completed with one freeze (−80°C) and thaw (30°C) cycle. Cell debris were centrifuged at 13, 200 r.p.m. for 10 min and the cytoplasmic extracts were filtered with a 0.20 μm Minisart RC4 syringe filter. Enzymatic reactions were carried out at 37°C by adding 0.2 ml of a 5% carbohydrate solution (w/v) to 0.8 ml of crude cytoplasmic extract. Reactions were stopped after 24 h by heating at 90°C for 5 min. Samples were diluted 200 times with MilliQ water and analysed by HPAEC-PAD on a Dionex ICS-3 000 system (Dionex Corp., Sunnyvale, CA) equipped with a CarboPac PA100 4×250 column connected to the corresponding guard column (Dionex).
The analyses were carried out at 30°C with a flow rate of 1 ml.min−1 with the following multistep gradient: 0–30 min (0–60% B), 30–32 min (60–90% B), 32–36 min (90–0% B) and 36–46 min (0% B). For analysing inulin hydrolysis products, samples were diluted twice in 1M NaOH, and the following multistep gradient was used: 0–15 min (0–30% B), 15–70 min (30–60% B), 70–125 min (60–90% B), 125–135 min (90% B), 135–150 min (90–0% B). Solvents were 150 mM NaOH (eluent A) and 150 mM NaOH, 500 mM CH3COONa (eluent B). We checked that the EPI100 E.coli host was unable to degrade the prebiotics tested in this study, as no prebiotic breakdown was detected by HPAEC-PAD, after having incubated during 24 h at 37°C the cytoplasmic and the total crude extracts of the EPI100 strain harboring the empty pCC1FOS vector with 5% of each prebiotic.
Pyrosequencing, read assembly and ORF detection
Fosmid pyrosequencing was performed using a 454 GS FLX System (454 Life Science, CT) by the INRA Genomic Platform (Auzeville, France). Read assembly was performed using CAP3 software [72]. Contigs showing a sequence length <1, 000 bp and a sequencing depth <8 were removed. Remaining contigs were cleaned from the vector pCC1FOS sequence using Crossmatch (http://bozeman.mbt.washington.edu/phredphrapconsed.html). Open reading frames (ORF) of at least 20 amino acids were predicted using MetaGene [73]. Annotated contig sequences were deposited in European Nucleotide Archive under accession numbers: HE663537 and HE717006 to HE717020 (Table S2).
ORF analysis
CAZyme encoding genes were identified by BLAST analysis of the predicted ORFs against the functional modules of glycoside hydrolases, polysaccharide lyases, carbohydrate esterases, carbohydrate-binding modules and glycosyltransferases included in the CAZy database (http://www.cazy.org) using a cut-off E-value of 7•10−6 followed by visual inspection and alignment with known CAZy families. Functions of other ORFs were inferred and manually annotated, based on BLASTX analysis against the NCBI non redundant and environmental database (E-value <10−8, identity >35%, query length coverage ≥50%). ORFs were assigned to clusters of orthologous groups of proteins (COGs) using RPS-BLAST analysis against the COG database (E-values ≤10−8). Finally, gene occurrence in the human gut microbiota was determined by TBLASTN comparisons of the predicted ORFs with the metagenomic data sets available, from fecal sampling of 162 subjects of the MetaHIT cohort (http://www.bork.embl.de/~arumugam/Qin_et_al_2010/http://www.bork.embl.de/Docu/Arumugam_et_al_2011/data/genes/; [5], [69]); of 139 U.S. subjects from the NIH-HMP cohort (http://www.hmpdacc.org/HMASM/; [70] and of the 110 subjects from different cultures, ages and families (http://metagenomics.anl.gov/projects/98; [43] (E-value = 0, identity ≥90% and = 100%).
Contig taxonomic assignment was based on ORF sequence similarity with sequenced genomes, using BLASTX analysis against the non-redundant NCBI database (E-value ≤10−8, identity ≥90%, query length coverage ≥50%). Contigs were assigned to a class, genus or species only if at least 50% of the ORFs were assigned to the same organism. Contigs containing ORFs assigned to different classes were not assigned. The most probable common ancestor of the organism from which the non-assignable contigs came from was retrieved using MEGAN v3.2.1, [74], based on BLASTX analysis against the non-redundant NCBI.
Supporting Information
Figure S1.
HPAEC-PAD analysis of reaction products resulting from prebiotic hydrolysis.
https://doi.org/10.1371/journal.pone.0072766.s001
(DOC)
Table S1.
Characteristics of metagenomic clones.
https://doi.org/10.1371/journal.pone.0072766.s002
(PDF)
Table S2.
Glycoside hydrolase families found in the study.
https://doi.org/10.1371/journal.pone.0072766.s003
(DOC)
Acknowledgments
The high-throughput screening work was performed at the Laboratory for BioSystems & Process Engineering (Toulouse, France) with the ICEO automated facility. We thank Sophie Bozonnet for her assistance.
Author Contributions
Conceived and designed the experiments: GPV MRS PM JD. Performed the experiments: DAC EL PR SL. Analyzed the data: DAC EL BH ML. Wrote the paper: DAC EL GPV.
References
- 1. Eckburg PB, Bik EM, Bernstein CN, Purdom E, Dethlefsen L, et al. (2005) Diversity of the human intestinal microbial flora. Science 308: 1635–1638.
- 2. Wang M, Ahrné S, Jeppsson B, Molin G (2005) Comparison of bacterial diversity along the human intestinal tract by direct cloning and sequencing of 16S rRNA genes. FEMS Microbiol Ecol 54: 219–231.
- 3. Booijink CC, Zoetendal EG, Kleerebezem M, de Vos WM (2007) Microbial communities in the human small intestine: coupling diversity to metagenomics. Future Microbiol 2: 285–295.
- 4. Rajilić-Stojanović M, Smidt H, de Vos WM (2007) Diversity of the human gastrointestinal tract microbiota revisited. Environ Microbiol 9: 2125–2136.
- 5. Qin J, Li R, Raes J, Arumugam M, Burgdorf KS, et al. (2010) A human gut microbial gene catalogue established by metagenomic sequencing. Nature 464: 59–65.
- 6.
Kleerebezem M (2010) Metagenomic approaches to unravel the composition and function of the human small intestine microbiota. In:Heidt PJ, Snel J, Midtvedt T, Rusch V editors. Intestinal microbiomics: novel indicators of health and disease. Old Herborn University. 27–42.
- 7. Zoetendal EG, Raes J, van den Bogert B, Arumugam M, Booijink C, et al. (2012) The human small intestinal microbiota is driven by rapid uptake and conversion of simple carbohydrates. ISME J 6: 1415–1426.
- 8. Lepage P, Leclerc MC, Joossens M, Mondot S, Blottière HM, et al. (2013) A metagenomic insight into our gut's microbiome. Gut 62: 146–158.
- 9.
Ottman N, Smidt H, deVos WM, Belzer C (2012) The function of our microbiota: who is out there and what do they do? Front Cell Infect Microbiol 2, DOI 10.3389/fcimb.2012.00104.
- 10. Gill SR, Pop M, Deboy RT, Eckburg PB, Turnbaugh PJ, et al. (2006) Metagenomic analysis of the human distal gut microbiome. Science 312: 1355–1359.
- 11. Flint HJ, Bayer EA, Rincon MT, Lamed R, White BA (2008) Polysaccharide utilization by gut bacteria: Potential for new insights from genomic analysis. Nat Rev Microbiol 6: 121–131.
- 12. Cummings JH, Macfarlane GT (1991) The control and consequences of bacterial fermentation in the human colon. J Appl Bacteriol 70: 443–459.
- 13. Grabitske HA, Slavin JL (2008) Low-digestible carbohydrates in practice. J Am Diet Assoc 108: 1677–1681.
- 14. Roberfroid M (2007) Prebiotics: the concept revisited. J Nutr 137: 830S–837S.
- 15. Figueroa-González I, Cruz-Guerrero A, Quijano G (2011) The benefits of probiotics on human health. J Microbial Biochem Technol S1: 003.
- 16. Bird AR, Conlon MA, Christophersen CT, Topping DL (2010) Resistant starch, large bowel fermentation and a broader perspective of prebiotics and probiotics. Benef Microbes 1: 423–431.
- 17. Jaskari J, Kontula P, Siitonen A, Jousimies-Somer H, Mattila-Sandholm T, et al. (1998) Oat beta-glucan and xylan hydrolysates as selective substrates for Bifidobacterium and Lactobacillus strains. Appl Microbiol Biotechnol 49: 175–181.
- 18. Djouzi Z, Andrieux C, Pelenc V, Somarriba S, Popot F, et al. (1995) Degradation and fermentation of alpha-gluco-oligosaccharides by bacterial strains from human colon: in vitro and in vivo studies in gnotobiotic rats. J Appl Bacteriol 79: 117–127.
- 19. Sarbini SR, Kolida S, Naeye T, Einerhand A, Brison Y, et al. (2011) In vitro fermentation of linear and alpha-1,2-branched dextrans by the human fecal microbiota. Appl Environ Microbiol 77: 5307–5315.
- 20. Herfel TM, Jacobi SK, Lin X, Fellner V, Walker DC, et al. (2011) Polydextrose enrichment of infant formula demonstrates prebiotic characteristics by altering intestinal microbiota, organic acid concentrations, and cytokine expression in suckling piglets. J Nutr 141: 2139–2145.
- 21. Ohkusa T, Ozaki Y, Sato C, Mikuni K, Ikeda H (1995) Long-term ingestion of lactosucrose increases Bifidobacterium sp. in human fecal flora. Digestion 56: 415–420.
- 22. Gullón B, Gullón P, Sanz Y, Alonso JL, Parajó JC (2011) Prebiotic potential of a refined product containing pectic oligosaccharides. LWT-Food Sci Technol 44: 1687–1696.
- 23. Hopkins MJ, Cummings JH, Macfarlane GT (1998) Inter-species differences in maximum specific growth rates and cell yields of bifidobacteria cultured on oligosaccharides and other simple carbohydrate sources. J Appl Bacteriol 85: 381–386.
- 24. Rask Licht T, Ebersbach T, Frøkiær H (2012) Prebiotics for prevention of gut infections. Trends in Food Science & Technology 23: 70–82.
- 25. Macfarlane S, Macfarlane GT, Cummings JH (2006) Review article: prebiotics in the gastrointestinal tract. Aliment Pharmacol Ther 24: 701–714.
- 26. Goulas T, Goulas A, Tzortzis G, Gibson GR (2009) Expression of four β-galactosidases from Bifidobacterium bifidum NCIMB41171 and their contribution on the hydrolysis and synthesis of galactooligosaccharides. Appl Microbiol Biotechnol. 84 (5): 899–907.
- 27. Rycroft CE, Jones MR, Gibson GR, Rastall RA (2001) A comparative in vitro evaluation of the fermentation properties of prebiotic oligosaccharides. J Appl Microbiol 91: 878–887.
- 28. Van der Meulen R, Makras L, Verbrugghe K, Adriany T, De Vuyst L (2006) In vitro kinetic analysis of oligofructose consumption by Bacteroides and Bifidobacterium spp. indicates different degradation mechanisms. Appl Environ Microbiol 72: 1006–1012.
- 29. Ramirez-Farias C, Slezak K, Fuller Z, Duncan A, Holtrop G, et al. (2009) Effect of inulin on the human gut microbiota: stimulation of Bifidobacterium adolescentis and Faecalibacterium prausnitzii. Brit J Nutr 101: 541–550.
- 30. Kleessen B, Hartmann L, Blaut M (2001) Oligofructose and long-chain inulin: influence on the gut microbial ecology of rats associated with a human faecal flora. Brit J Nutr 86: 291–300.
- 31. Langlands SJ, Hopkins MJ, Coleman N, Cummings JH (2004) Prebiotic carbohydrates modify the mucosa-associated microflora of the human large bowel. Gut 53: 1610–1616.
- 32. Duncan SH, Hold GL, Harmsen HJM, Stewart CS, Flint HJ (2002) Growth requirements and fermentation products of Fusobacterium prausnitzii, and a proposal to reclassify it as Faecalibacterium prausnitzii gen. nov., comb. nov. Int J Syst Evol Microbiol 52: 2141–2146.
- 33. Schell MA, Karmirantzou M, Snel B, Vilanova D, Berger B, et al. (2002) The genome sequence of Bifidobacterium longum reflects its adaptation to the human gastrointestinal tract. PNAS 99: 14422–14427.
- 34. Barrangou R, Altermann E, Hutkins R, Cano R, Klaenhammer RT (2003) Functional and comparative genomic analyses of an operon involved in fructooligosaccharide utilization by Lactobacillus acidophilus. PNAS 100: 8957–8962.
- 35. Goh JY, Zhang C, Benson KA, Schlegel V, Lee JH, et al. (2006) Identification of a putative operon involved in fructooligosaccharide utilization by Lactobacillus paracasei. Appl Environ Microbiol 72: 7518–7530.
- 36. Ryan MS, Fitzgerald FG, van Sinderen D (2005) Transcriptional regulation and characterization of a novel beta-fructofuranosidase-encoding gene from Bifidobacterium breve UCC2003. Appl Environ Microbiol 71: 3475–3482.
- 37. Scott KP, Martina JC, Chassarda C, Clergeta M, Potrykusa J, et al. (2011) Substrate-driven gene expression in Roseburia inulinivorans: Importance of inducible enzymes in the utilization of inulin and starch. PNAS 108: 4672–4679.
- 38. Majumder A, Sultan A, Jersie-Christensen RR, Ejby M, Schmidt BG, et al. (2011) Proteome reference map of Lactobacillus acidophilus NCFM and quantitative proteomics towards understanding the prebiotic action of lactitol. Proteomics 11: 3470–3481.
- 39. Imamura L, Hisamitsu K, Kobashi K (1994) Purification and characterization of beta-fructofuranosidase from Bifidobacterium infantis. Biol Pharm Bull 17: 596–602.
- 40. Van Laere KMJ, Hartemink R, Beldman G, Pitson S, Dijkema C, et al. (1999) Transglycosidase activity of Bifidobacterium adolescentis DSM 20083 α-galactosidase. Appl Microbiol Biotechnol 52: 681–688.
- 41. Sonnenburg ED, Zheng H, Joglekar P, Higginbottom SK, Firbank SJ, et al. (2010) Secificity of polysaccharide use in intestinal Bacteroides species determines diet-induced microbiota alterations. Cell 141: 1241–1252.
- 42. Lagaert S, Pollet A, Delcour JA, Lavigne R, Courtin CM, et al. (2011) Characterization of two β-xylosidases from Bifidobacterium adolescentis and their contribution to the hydrolysis of prebiotic xylooligosaccharides. Appl Microbiol Biotechnol 92: 1179–1185.
- 43. Yatsunenko T, Rey FE, Manary MJ, Trehan I, Dominguez-Bello MG, et al. (2012) Human gut microbiome viewed across age and geography. Nature 486: 222–228.
- 44. Tasse L, Bercovici J, Pizzut-Serin S, Robe P, Tap J, et al. (2010) Functional metagenomics to mine the human gut microbiome for dietary fiber catabolic enzymes. Genome Res 20: 1605–1612.
- 45. Aider M, de Halleux D (2007) Isomerization of lactose and lactulose production: review. Trends Food Sci Tech 18: 356–364.
- 46. Kunz C, Rudloff S, Baier W, Klein N, Strobel S (2000) Oligosaccharides in human milk: structural, functional, and metabolic aspects. Annu Rev Nutr 20: 699–722.
- 47. Urashima T, Saito T, Nakamura T, Messer M (2001) Oligosaccharides of milk and colostrum in non-human mammals. Glycoconjugate J 18: 357–371.
- 48. Kurokawa K, Itoh T, Kuwahara T, Oshima K, Toh H, et al. (2007) Comparative metagenomics revealed commonly enriched gene sets in human gut microbiomes. DNA Res 14: 169–181.
- 49. Turnbaugh PJ, Gordon JI (2009) The core gut microbiome, energy balance and obesity. J Physiol 587: 4153–4158.
- 50. Kullin B, Abratt VR, Reid SJ (2006) A functional analysis of the Bifidobacterium longum cscA and scrP genes in sucrose utilization. Appl Microbiol Biotechnol 72: 975–981.
- 51. Bujacz A, Jedrzejczak-Krzepkowska M, Bielecki S, Redzynia I, Bujacz G (2011) Crystal structures of the apo form of β-fructofuranosidase from Bifidobacterium longum and its complex with fructose. FEBS J 278: 1728–1744.
- 52. Jedrzejczak-Krzepkowska M, Tkaczuk KL, Bielecki S (2011) Biosynthesis, purification and characterization of β-fructofuranosidase from Bifidobacterium longum KN29.1. Process Biochem 46: 1963–1972.
- 53. Lagaert S, Van Campenhout S, Pollet A, Bourgois TM, Delcour JA, et al. (2007) Recombinant expression and characterization of a reducing-end xylose-releasing exo-oligoxylanase from Bifidobacterium adolescentis. Appl Environ Microbiol 73: 5374–5377.
- 54. Lagaert S, Pollet A, Delcour JA, Lavigne R, Courtin CM, et al. (2010) Substrate specificity of three recombinant α-L-arabinofuranosidases from Bifidobacterium adolescentis and their divergent action on arabinoxylan and arabinoxylan oligosaccharides. Biochem Bioph Res Co 402: 644–650.
- 55. González R, Klaassens ES, Malinen E, de Vos WM, Vaughan EE (2008) Differential transcriptional response of Bifidobacterium longum to human milk, formula milk, and galactooligosaccharide. Appl Environ Microbiol 74: 4686–4694.
- 56. Belenguer A, Duncan SH, Calder AG, Holtrop G, Louis P, et al. (2006) Two routes of metabolic cross-feeding between Bifidobacterium adolescentis and butyrate-producing anaerobes from the human gut. Appl Environ Microbiol 72: 3593–3599.
- 57. Rossi M, Corradini C, Amaretti A, Nicolini M, Pompei A, et al. (2005) Fermentation of fructooligosaccharides and inulin by Bifidobacteria: A comparative study of pure and fecal cultures. Appl Environ Microbiol 71: 6150–6158.
- 58. Conte MP, Schippa S, Zamboni I, Penta M, Chiarini F, et al. (2006) Gut-associated bacterial microbiota in paediatric patients with inflammatory bowel disease. Gut 55: 1760–1767.
- 59. Iyer R, Tomar SK, Maheswari TU, Singh R (2010) Streptococcus thermophilus strains: Multifunctional lactic acid bacteria. Int Dairy J 20: 133–141.
- 60. Rhimi M, Boisson A, Dejob M, Boudebouze S, Maguin E, et al. (2010) Efficient bioconversion of lactose in milk and whey: immobilization and biochemical characterization of a beta-galactosidase from the dairy Streptococcus thermophilus LMD9 strain. Res Microbiol 161: 515–525.
- 61.
Stabler SP (1999) B12 and nutrition. In: Barnejee R editor. Chemistry and biochemistry of B12. Wiley: New York. 343–365.
- 62. Barnerjee R (2006) B12 trafficking in mammals: A for coenzyme escort service. ACS Chem Biol 1: 149–159.
- 63. Santos F, Vera JL, van der Heijden R, Valdez G, de Vos WM, et al. (2008) The complete coenzyme B12 biosynthesis gene cluster of Lactobacillus reuteri CRL1098. Microbiology 154: 81–93.
- 64. Van Hylckama Vlieg JE, Veiga P, Zhang C, Derrien M, Zhao L (2011) Impact of microbial transformation of food on health – from fermented foods to fermentation in the gastro-intestinal tract. Curr Opin Biotechnol 22: 211–219.
- 65. Temperton B, Field D, Oliver A, Tiwari B, Muhling M, et al. (2009) Bias in assessments of marine microbial biodiversity in fosmid libraries as evaluated by pyrosequencing. Isme J 3: 792–796.
- 66. Gabor EM, Alkema WB, Janssen DB (2004) Quantifying the accessibility of the metagenome by random expression cloning techniques. Environ Microbiol 6: 879–886.
- 67. Taras D, Simmering R, Collins MD, Lawson PA, Blaut M (2002) Reclassification of Eubacterium formicigenerans Holdeman and Moore 1974 as Dorea formicigenerans gen. nov., comb. nov., and description of Dorea longicatena sp. nov., isolated from human faeces. Int J Syst and Evol Microbiol 52: 423–428.
- 68. Smillie CS, Smith MB, Friedman J, Cordero OX, David LA, et al. (2011) Ecology drives a global network of gene exchange connecting the human microbiome. Nature 480: 241–244.
- 69. Arumugam M, Raes J, Pelletier E, Le Paslier D, Yamada T, et al. (2011) Enterotypes of the human gut microbiome. Nature 473: 174–180.
- 70. Peterson J, Garges S, Giovanni M, McInnes P, Wang L, et al. (2009) The NIH Human Microbiome Project. Genome Res. 19: 2317–2323.
- 71. Gloux K, Berteau O, El Oumami H, Beguet F, Leclerc M, et al. (2011) A metagenomic β-glucuronidase uncovers a core adaptive function of the human intestinal microbiome. PNAS 108 Suppl 14539–4546.
- 72. Huang X, Madan A (1999) CAP3: A DNA sequence assembly program. Genome Res 9: 868–877.
- 73. Noguchi H, Park J, Takagi T (2006) MetaGene: prokaryotic gene finding from environmental genome shotgun sequences. Nucleic Acids Res 34: 5623–5630.
- 74. Huson DH, Auch AF, Qi J, Schuster SC (2007) MEGAN analysis of metagenomic data. Genome Res 17: 377–386.