Chickens are major source of food and protein worldwide. Feed conversion and the health of chickens relies on the largely unexplored complex microbial community that inhabits the chicken gut, including the ceca. We have carried out deep microbial community profiling of the microbiota in twenty cecal samples via 16S rRNA gene sequences and an in-depth metagenomics analysis of a single cecal microbiota. We recovered 699 phylotypes, over half of which appear to represent previously unknown species. We obtained 648,251 environmental gene tags (EGTs), the majority of which represent new species. These were binned into over two-dozen draft genomes, which included Campylobacter jejuni and Helicobacter pullorum. We found numerous polysaccharide- and oligosaccharide-degrading enzymes encoding within the metagenome, some of which appeared to be part of polysaccharide utilization systems with genetic evidence for the co-ordination of polysaccharide degradation with sugar transport and utilization. The cecal metagenome encodes several fermentation pathways leading to the production of short-chain fatty acids, including some with novel features. We found a dozen uptake hydrogenases encoded in the metagenome and speculate that these provide major hydrogen sinks within this microbial community and might explain the high abundance of several genera within this microbiome, including Campylobacter, Helicobacter and Megamonas.
Citation: Sergeant MJ, Constantinidou C, Cogan TA, Bedford MR, Penn CW, Pallen MJ (2014) Extensive Microbial and Functional Diversity within the Chicken Cecal Microbiome. PLoS ONE 9(3): e91941. https://doi.org/10.1371/journal.pone.0091941
Editor: John Parkinson, Hospital for Sick Children, Canada
Received: October 14, 2013; Accepted: February 18, 2014; Published: March 21, 2014
Copyright: © 2014 Sergeant 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: The work was funded by the BBSRC. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
Competing interests: Mike Bedford is employed by AB Vista Feed Ingredients. There are no patents, products in development or marketed products to declare. This does not alter the authors' adherence to all the PLOS ONE policies on sharing data and materials, as detailed online in the guide for authors.
The domestic chicken occupies a special place in science and society. It is the most common domesticated animal, the most important food production animal and the most abundant and widely distributed bird in the world, scattered across islands and continents from the Arctic to the Falklands. Chickens provide a case study in domestication, a process began >7,000 years ago in South-East Asia . Chickens have also played a key role in the history of science, from Aristotle's description of a chick embryo  to Darwin's account of their domestication . The chicken has been adopted as a laboratory organism for research in genetics, embryology, development, immunology, virology and cancer . The completion of the chicken genome sequence was a landmark in basic and applied biology . However, the nuclear and mitochondrial genomes of the chicken represent just a small part of the gene pool associated with this bird.
Chickens, like other birds, have proportionally smaller intestines and shorter transit digestion times than mammals, but do not appear to any less efficient at digestion than their mammalian counterparts . This may be explained, in part, by the fact that the chicken gastrointestinal tract is home to a complex microbial community, the chicken gut microbiota, which underpins the links between diet and health in this bird. The healthy chicken gut microbiota plays important roles in assimilating nutrients from food, in producing vitamins and essential amino acids and in preventing harmful pathogens from gaining a foothold . The chicken gut microbiota also acts as source of human infections and as a reservoir of antibiotic-resistance determinants . An optimal gut microbiota can increase agricultural productivity, as evidenced by the ability of antibiotics to promote growth in chicks . This microbiota is also home to a rich collection of genes, the chicken gut microbiome, likely to include many sequences of scientific interest and biotechnological potential.
The chicken gastrointestinal tract consists of a crop, where food can be stored, followed by a proventriculous or true stomach which leads to the muscular gizzard, where digesta are ground, before entering the small and finally large intestine. The most densely populated microbial community within the chicken gut is found in the ceca, a pair of blind-ended sacs that open off the large intestine . At the entrance to each cecum, long intertwining villi sieve out coarse particulate material, allowing only soluble or fine particulate digesta to enter the lumen. Whereas the transit time through the upper intestine is only 2.5 hours, digesta dwell in the ceca for 12–20 hours, potentially allowing a longer time for digestion and absorption of nutrients. Various roles have been assigned to the cecal microbiota in wild birds  but its function in the modern commercial broiler is unclear and, in some experiments, the caeca have been removed without impairing the chicken's development . However, this microbiota has been implicated in many processes; nitrogen recycling by breakdown of uric acid , supplying B vitamins to their hosts  and producing essential amino acids required by their host . Particularly relevant to the intensive farming of chickens is the cecum's role in digestion of non-starch polysaccaharides NSPs , which are found in the grains used in commercial chicken feed. Some authors suggest as much as 10% of energy needs is recovered from a well-functioning cecum , , with chicken fed high fiber showing a dramatic increase in the size of this organ . The fermentation of breakdown products of NSPs produces short chain volatile fatty acids (SCFA), which are absorbed across the mucosa and catabolised by the host . Such SCFA not only contribute to the nutrition of the chicken, they also lower pH which can inhibit acid-sensitive pathogens and improve mineral absorption . In particular, butyrate has been shown to reduce Salmonella infection  and it is also able to stimulate epithelial cell growth . Molecular hydrogen produced during the fermentation process  if allowed to accumulate can lead to inhibition of this process. Therefore, for efficient fermentation this gas needs to be removed by hydrogen-consuming bacteria . These interactions are poorly understood and therefore the current study used metagenomics to identify pathways in the chicken ceca microbiome for NSP breakdown, SCFA production and hydrogen consumption.
Previous culture-based studies have established that, as in other vertebrates, the phylums Bacteroidetes and Firmicutes predominated in the chicken cecal microbiota –. Attempts at culture-independent phylogenetic profiling using 16S rRNA gene sequences have yielded insights into the composition of the chicken cecal microbiome – and more recent studies have shown the domination by Fimicutes, especially of the Clostridiales order . Similarly, studies using metagenomics, the whole-scale sequencing of DNA extracted from a microbial community without target-specific amplification or culture, have provided glimpses into the function of this complex microbial community , . However, most previous studies have examined pooled samples, thus ignoring variation in individual chickens, and have not achieved sufficient depth of coverage to provide a global picture of the taxonomic diversity and functional repertoire of the whole microbiome. This study aimed to address these points and analyzed cecal contents from 10 broiler chickens from the same flock on a commercial farm to an unprecedented depth, and on one sample produced the largest metagenomic analysis to date in order to illuminate the function of the cecal microbiota in chickens.
In accordance with the United Kingdom Animal (Scientific Procedures) Act of 1986, this study did not require a Home Office project license because no regulated procedures were carried out. Chickens were humanely killed at a designated establishment by cervical dislocation, which is an appropriate method under Schedule 1 of the Act.
Samples collection, DNA extraction and PCR amplification
Cecal samples were collected from Ross broilers at 42 days, housed indoors under standard commercial conditions. They had been fed on a wheat based diet with 5% maize which contained ionophores but no antibiotics. The feed conversion ratio was standard for the breed and there was no feed withdrawal period before the birds were sacrificed. The cecal contents were a typical brown colour and like toothpaste in constituency. Birds were euthanized by cervical dislocation and the ceca removed and transported to the laboratory on ice. The cecal surface was sterilized with 70% ethanol, a longitudinal incision was made with a scalpel and the edges pulled back. Contents were removed into a sterile tube and flash frozen in liquid nitrogen and stored at −70°C. DNA was extracted and PCR amplification of 16S rRNA sequences was performed as described previously  using primers F20 and R19 flanking the variable regions V1–V3 and an annealing temperature of 55°C. Sequencing of amplicons was carried out on the Roche FLX Titanium instrument as described previously .
Phylogenetic analysis of 16S rRNA gene sequences
Reads were truncated where flow signals were less than 0.7 and all sequences were trimmed to 600 flows (around 350 bp in length). Clustering of the reads into Operation Taxonomic Units (OTUs) which share 97% identity was achieved using Uparse . Other methods were also attempted (Text S1 and Figure S5). The following modifications to the standard Uparse pipeline were carried out. Pre-clustering (-cluster_smallmem) was included, which clustered sequences exhibiting two or less nucleotide difference. In addition, an extra chimera checking step (-uchime_ref) against the default reference database was performed. The resulting sequences were taken as defining putative operational taxonomic unites (OTUs) or phylotypes. OTU sequences were assigned to a taxonomic lineage using QIIME (assign_taxonomy.py) . In addition, all OTU sequences were BLAST-searched against the NCBI nr database.
The resultant sequences, frequencies and taxonomic assignments were formatted into an OTU table that was compatible with the QIIME pipeline . QIIME was used to calculate Simpsons' index and construct rarefaction curves. Dendograms depicting the similarity of bacterial communities were constructed via the jackknifed_beta_diversity script from QIIME, exploiting the Bray Curtis method to compute a similarity matrix and then the un-weighted pair-group method with arithmetic mean (UGPMA) to cluster the results. Jacknifed support was included to account for the different sampling depths. A hundred rarefied tables were generated at sample size that corresponded to the number of reads in the smallest sample to produce bootstrap values.
Metagenome sequencing and analysis
DNA extracted from cecum 1A was used for metagenome sequencing. Library preparation and sequencing on one channel of the Illumina MiSeq 2000 were performed according to the manufacturer's instructions by Source BioScience, Nottingham UK.
Preliminary analysis of the data involved blasting 260 000 reads against the NCBI nr database and obtaining taxonomic from the output using MEGAN(45). A comprehensive analysis was then carried out and is summarized in Figure S4a. Initially, the reads were filtered and trimmed with CLC workbench (CLC bio, Denmark) using the default settings and reads below 60 bases were removed. The 81,772,788 reads that remained were assembled into contigs (environmental gene tags) using CLC work bench under the default settings. Coding sequences (CDSs) were predicted using MetaGene  and translated into protein sequences which were searched against the NCBI NR database using BLASTP. BLAST results were fed into MEGAN , which provided KEGG and SEED annotations as well as assigning a taxonomic origin using the software's LCA algorithm. Coverage of each environmental gene tag (EGT) was assessed by mapping reads to the EGT using BWA .
A taxonomic assignment was made for each EGT whenever >50% of the CDSs from the EGT had been assigned by MEGAN to the same taxon, otherwise the EGT's taxonomic assignment was scored as unknown. Functional domains were assigned to each predicted protein using Hidden Markov Models (HMMs). This was achieved using hmmsearch from HMMER  against the PFAM  and TIGRFAMS databases. Signal sequences were predicted using SignalP  and antibiotic resistance assigned using the ARDB package . Known patterns were identified in the data using stand alone tool from the PROSITE web site . All the information was stored in a MySQL database (Figure S4A) for easy retrieval and linking of the different threads of information.
Binning of EGTs
Z scores for the 256 tetra-nucleotides in each EGT were calculated as described previously . Pairwise comparisons of EGT Z scores were performed using the Pearson correlation coefficient and the results assembled into a matrix. The matrix was then used to cluster EGTs at various identity levels using the cluster.classic command from MOTHUR . Initial seed clusters were obtained at 0.25 identity. Clusters were merged into putative genomes if they had similar depth-of-coverage and an identical taxonomic assignment at the family level. The validity of putative genomes was assessed by determining whether the house-keeping genes, rpoB, recA, groEL and gyrB were single copy. Taxonomic status confirmed by when bacterial ‘core’ genes from the human microbiome project (http://hmpdacc.org/tools_protocols/tools_protocols.php) were blasted against putative genomes with a cut off value of 1e−10. Completeness of putative genomes was assessed by comparing the size of the genome with that of the closest relative in the GenBank.
Functional domain analysis
Most glucosyl Hyrdrolase (GH) domains were detected using PFAM HMMs as described previously . To detect GH domains without an HMM, metagenomic protein sequences were blasted against a set of GH domains described previously  with an e-value cut off 10−6. Xylanase, licheninase and cellulase sequences were detected using predictions of enzyme activities from the SEED and KEGG annotations, combined with the appropriate GH domains as specified by the Cazy Website  (Table S6). Gene contexts were visualised in Artemis.
Fermentation to acetate was examined by looking for putative acetate kinase (TIGR0016) and phosphate acetyltransferase (TIGR00651) that were adjacent to each other in EGTs. To investigate fermentation to butyrate, 3-hydroxybuturyl-CoA dehydrogenase (BCD), models were constructed using HMM, based on 10 BCD proteins from divergent prokaryotes. Putative BCD genes were indentified using this model and were further verified by looking for an adjacent Electron Transfer Flavoprotein (PF01021). Putative butyryl-CoA:acetate Coa transferase genes was found by using the conserved sequence AFVDIAKAG . The alternative pathway was identified by adjacent transbutyrylase (TIGR02706) and butyrate kinase (TIGR02707) genes. Fermentation to propionoate was investigated by searching for the methyl-Malonyl-CoA:pyruvate transcarboxylase gene using as a SEED/KEGG annotation of EC 184.108.40.206 and the epimerase gene was putatively identified using a SEED annotation of EC 220.127.116.11 and TIGR03081. Methylmalonyl decarboxylase genes were identified by a SEED/KEGG annotation of EC18.104.22.168 and TIGR01117.
Putative uptake hydrogenases were identified with the patterns [EGMQS]RXC[GR][IV]CXXX[HT]XXX[AGS]X(0,4)[VANQD] or [AFGIKLMV][HMR]XX[HR][AS][AFLY][DN]PC[FILMV]XC[AGS]XH . The dissimilatory 5′ phophosulfate resuctase (APS) present in suphate reducers was putatively identified if genes with TIGR02064 and TIG02066 models were adjacent to each other. The AcetylCoa Synthase beta subunit responsible for the CO dehydrogenase reaction in acetogens was identified by the conserved sequences LCGAVSW and PMTSCGC . A search for the methyl-coenzyme M reductase present in methanogens was performed by looking for adjacent genes containing TIGR03256 and TIG03257 domains.
Results and Discussion
Microbial diversity in the chicken cecum
From the ten chickens, we recovered 414,070 16S sequences, which were classified into 699 phylotypes using a 97% sequence identity threshold. Rarefaction curves suggest that sampling has reached saturation and that the number of phylotypes in individual chickens ranges from ∼200 to ∼350 (Figure 1A). Sequences from 232 phylotypes showed less than 97% identity to any sequence in GenBank (Data Set S1) and therefore potentially represent new species. The majority of these OTUs contained large numbers of reads and hence are unlikely to be due to sequencing or PCR errors (Figure S1).
(A) Rarefaction curves of the OTUs clustered at 97% sequence identity for each chicken, based on the averages of the four replicates per bird. (B) Hierarchical clustering visualizing similarities among cecal samples. The number in each label indicates each individual chicken and the letter each individual ceca (either A or B). Technical replicates of each ceca were carried out. (C) Bar chart showing the proportion of reads assigned to the top 5 most abundant OTUs in each chicken.
Previous studies have shown large inter-chick variation of microbiota community structure with both T-RLFP  and quantitative PCR , so we were curious to determine how much variation occurs within the same region of the gut. We analysed biological replicates from each bird (two ceca) and technical replicates for each sample and found that samples from the same bird clustered together, with the exception of bird 1(Figure 1B). Furthermore, in all but one case (bird 4), individual ceca from the same chicken could be differentiated with technical replicates clustering together. Simpson's diversity index varied between ceca and birds, with the highest value (i.e. least evenness) in bird 1 (Table S1), which was dominated by Megamonas and Alistipes (Figure 1C).
Curiously, in our samples, Lachnospriceae, Ruminococcaceae and Bacteroides appear to be less abundant than might be expected from previous studies (Figure S2). This can largely be explained by the unprecedented numerical dominance of the class Negativicutes (over 20% of all sequences) and in particular, by the high prevalence of a single novel species of Megamonas, which accounts for nearly a third of the reads in one of the birds, chicken 9 (Figure 1C).
The chicken cecal metagenome
We obtained a hundred and twenty million 110-base reads from the contents of a single cecal sample (cecum 1A). A taxonomic analysis of a subset of reads (Table S2) showed that, of the reads that could be assigned to a taxon, >98% were bacterial, with <1% assigned to the chicken and fewer than one in a thousand assigned to archaea, other eukaryotes (e.g. the cereal fungus Magnaporthe orazae and the alga Chlamydomonas reinharrdii) or to viruses (bacteriophages and Fowl Adenovirus).
Reads were incorporated into a 337-Mbp assembly, with an N50 of 647 bp and consisting of 648,251 environmental gene tags (EGTs). From this assembly, we predicted 108,103 complete coding sequences (CDS) and 595,178 partial CDSs. Of the EGTs greater than 2 kb about 13% could be assigned to a known species (Figure S3A), suggesting that the majority of EGTs that we have recovered represent new species. Seven percent of EGTs could not even be assigned to a known domain (Figure S3A). The majority of these showed, on homology searching, just a few weak matches to phage genes. Since phages largely contain genes with unknown functions  these EGTs probably represent phages.
When the taxonomic distribution predicted from the 16S analysis is compared with that from the metagenomics data, there is a discrepancy in the ratio of Firmicutes to Bacteroidetes (Figure S3b). This is largely due to differences in the ratio of reads from the genus Alistipes, which belongs to the Bacteroidetes, to those from Megamonas, which belongs to the Firmicutes, (19.9 and 16.3% in the 16S data set; 32% of and 9% in the metagenome), probably reflecting differential amplification in the 16S PCR.
EGTs were binned according to Z-score, coverage and taxonomic assignment, allowing us to reconstruct two-dozen genomes larger than 1Mbp (Table 1). Only five genomes showed a ≥99% match within the rpoB gene to sequences in GenBank; suggesting most of the draft genomes we have recovered represent new species.
Antibiotic resistance genes were investigated using the ARDB database pipeline  and the results are shown in data S2. The most common predicted resistance was against tetracycline and bacitracin. In addition a vancomycin resistance operon in Enterococcus was detected.
The grain that forms the basis of most commercial chicken diets is rich in non-starch polysaccharides, with β-glucans, including cellulose, predominating in barley and oats and arabinoxylans abundant in wheat, maize and rye. In comparison to a set of other known metagenomes, the chicken cecal microbiome contains the largest proportion of sequences (1.5%) representing glycosyl-hydrolase (GH) domains (Table 2). Within this microbiome, we found sequences from over two hundred different non-starch polysaccharide-degrading enzymes (Figure 2).
The figure shows each class of enzyme as judged by SEED/KEGG annotation and GH (see methods) for three types of NSP. The size of pie chart reflects abundance of the enzyme class; numbers indicate quantity of genes assigned to each bacterial Taxon at the class level.
In line with the expectation that less soluble, larger polymers are excluded from the cecum, we found that, in comparison to other microbial communities, the chicken cecal microbiome (along with the panda microbiome) harbored the lowest proportion of cellulases and endohemicellulases (4%), but the largest proportion of oligosaccharide degrading enzymes (38%) within GH-contiaing genes (Table 2). In particular, glucanases acting on oligosaccharides (E.C. 22.214.171.124 and E.C. 126.96.36.199) predominate over endoglucanases (EC 188.8.131.52 and E.C 184.108.40.206) acting on full-length polymers of cellulose and xylan (Figure 2). We found no evidence of any sequences representing the exo-acting cellobiohydralase activity (E.C. 220.127.116.11) associated with the degradation of crystalline cellulose . We found a greater abundance of sequences involved in degradation of xylans than for degradation of β-glucans in the ceca, which contrasts with the relative dominance of β glucanase activity over xylanase in the small intestine .
Enzymes involved in xylan and cellobiose degradation are similarly distributed across three main bacterial classes, Actinobacteria, Clostridia and Bacteroidia (Figure 2). However, among endoglucanases, there appears to be a relative under-representation of the Actinobacteria, with greater contributions from the Negativicutes and Lentisphaerae (in particular the species Victivallis vadensis).
In order to reap the benefits of extracellular non-starch polysaccharides (NSP) degradation, bacteria exploit multi-protein cellulosomes, characterized by dockerin motifs, or isolated cellulases which use carbohydrate binding modules (CBM) to attach to the substrate , . However, in our dataset, only 51 of 9498 GH domains were attached to binding modules and only two were attached to dockerin motifs (Table S3), again suggesting an absence or paucity of conventional cellulase activity.
Another complex enabling bacteria to exploit NSP degredation, which are present in Bacteroides, are polysaccharide utilization systems (PUL), which digest and import the products of polysaccharide degradation . We found evidence of over 500 such systems in the Bacteriodetes from the chicken cecum, including representatives of a novel class of PUL recently identified in the Tammar Wallaby foregut that incorporates β-1,4 endoglucanases and β-1,4 endoxylanases  (Figure 3A).
NSP degrading genes were identified by SEED/KEGG annotation and GH domain (see methods). (A) Gene clusters encoding putative PUL systems from various Bacteroidetes. (B) Gene cluster encoding putative integrated polysaccharide degradation and utilization systems from various Firmicutes. (C) NSP degrading enzymes associated with sporulation genes * indicates predicted signal peptide.
We found additional genetic evidence of co-ordination of polysaccharide degradation with sugar transport and utilization (Figure 3B) – for example, in Megamonas, we found a gene cluster encoding secreted endo-glucanases, a cellobiose phosphotransferase system and a 6-phospho-beta-glucocidase, which together probably form an extracellular complex capable of degrading NSPs to cellobiose, which is then imported and cleaved to give D-glucose.
Some endoglucanases were not associated with regulatory genes or sugar transporters and therefore may not be involved in the exploitation of NSP for energy. For example, it has been suggested that some cellulase genes might be involved in biosynthesis of cellulose , as evidenced by their association in operons with cellulose synthesis genes. We found no evidence of biosynthetic cellulose activity in the chicken microbiome. However, we found two endoglucanase genes in a Clostridiales genome that were associated with sporulation genes (Figure 3C), which leads us to speculate that such endoglucanase genes may be implicated in breakdown of the cell wall or formation of the spore cortex.
Production of short-chain volatile fatty acids
Short-chain volatile fatty acids (SCFAs), such as acetic acid, propionic acid and butyric acid, are produced by microbial fermentation of the sugars released from non-starch polysaccharides and are found in the chicken ceca . SCFAs elicit several beneficial effects. Absorption of SCFAs across the gut mucosa contributes to the nutrition of the chicken (6). In addition, SCFAs can inhibit the growth of pathogens and improve mineral absorption .
We found abundant evidence of acetate fermentation in the chicken cecal metagenome, with over 30 acetate kinase/phosphotransferase sequences (Table S4). One of pathways to propionate fermentation, the Wood-Werkman cycle, is poorly represented, with two methylmalonyl-CoA carboxytransferase genes found at low coverage (Figure 4A). However, an alternative propionate fermentation pathway, which exploits methylmalonyl-CoA decarboxylase and methylmalonyl-CoA epimerase appears to be well represented in this microbial community, with many genes encoding these enzymes (Figure 4A). As expected, we found that gene clusters from the Bacteroidetes that encode the alpha subunit also encode the beta, gamma and delta subunits of methylmalonyl-CoA decarboxylase together with methylmalonyl epimerase (Figure 4B). However, similar gene clusters from Megamonas and Dialister appear to lack genes for the beta subunit, which couples decarboxylation to export of Na+ across the membrane, and the delta subunit, which holds the complex together . Instead these clusters encode a ferredoxin gene, which, we speculate, may be involved in coupling decarboxylation to electron transport in a novel propionate fermentation pathway.
(A) Pathways involved in propionoate production showing the putative genes identified coding for the enzymes involved. Size of pie chart reflects abundance of the gene class; numbers indicate quantity of genes assigned to each taxon. (B) Operon structure of two Methylmalonyl decarboxylase loci.
The first committed step in butyrate production is the conversion of crotonoyl-CoA to butyryl-CoA by 3-hydroxybutyryl-CoA dehydrogenase (BCD) . Assignment of this enzyme to KEGG or SEED categories is hampered by similarities of this enzyme to acyl-CoA dehydrogenases involved in lipid metabolism. In butyrate-producing bacteria the BCD gene is found in an operon adjacent to electron transferring flavoproteins (ETF). The BCD:ETF complex conserves the energy of the reaction, which is believed to be used to generate proton-motive force .
We searched the chicken cecal microbiome for BCD sequences, using a protein-sequence-based hidden Markov model, twinned with evidence from genetic context (co-occurrence with adjacent ETF genes). This approach yielded nineteen sequences, mostly from Bacteroidetes and Firmicutes, but also a single sequence from Escherichia spp (Figure 5A). In one Firmicutes EGT, there was enough adjacent sequence to reveal genomic context, which included the classic BDH operon structure  (Figure 5B). However in Bacteroidetes and Escherichia, the order of the BCD and ETF genes differed and they were not adjacent to other genes in the butyrate fermentation pathway.
(A) Pathways involved in butyrate production showing the putative genes identified coding for the enzymes involved. Size of pie chart reflects abundance of the gene class; numbers indicate quantity of genes assigned to each taxon. (B) Operon structure of two butyryl-CoA:acetate-CoA transferase (BCD) loci.
After the formation of butyryl–CoA, butyrate fermentation can proceed either by butyryl-CoA:acetate:CoA transferase or via the enzymes phosphotransbutyrylase and butyrate kinase (Figure 5A). The former pathway is by far the most commonly used by butyrate-producing bacteria isolated from the human gut . We detected only four sequences encoding butyryl-CoA:acetate:CoA transferase in our chicken cecal metagenome. This is in line with results from a recent study of butyrate-producing bacteria cultured from chicken ceca , where fewer than half the isolates contained the butyryl-CoA:aceate: CoA transferase gene, leading to speculation that another as yet unidentified acyl transferase might substitute for the canonical butyryl-CoA: acetate CoA transferase in some butyrate fermentation pathways. We found nineteen sequences corresponding to phosphotransbutylase and butyrate kinase, including seven from the Bacteroidetes. As Bacteroidetes appear to contain an energy conserving BCD/ETF complex and phosphotransbutylase/butyrate kinase genes, we speculate that they might have the potential to produce butyrate. In this context, it is worth noting that, although butyrate production is not generally associated with Bacteroidetes grown in culture, butyrate has been shown to be produced by Bacteroides spp, in colonized germ free mice .
Hydrogen sinks within the cecal microbiota
One of the goals of metagenomics, and of microbial ecology in general, is to explain why some taxa are so much more abundant than others in a given microbiology community. We speculated that high species abundance might be explained in the chicken cecal microbiome by the need for hydrogen sinks within this ecosystem.
During fermentation, NADH formed through glycolysis needs to be recycled to NAD+. This can be achieved through reduction of pyruvate or acetyl CoA to SCFAs. Fermentation to acetate produces the most ATP by substrate level phosphorylation, but this route allows little NADH to be recycled. Hydrogenases provide an alternative route for NADH recycling, but one that leads to the production of molecular hydrogen . Hydrogenases are inhibited by high levels of their own product, so without a sink for hydrogen within an ecosystem, accumulation of hydrogen leads to reduced fermentation and/or less energy-efficient fermentation to substances such as ethanol, butyrate and propionate , . The presence of bacteria that can act as a hydrogen sink not only results in a switch to the more productive fermentation to acetate, but also increased production of SCFAs which bring benefits to the host.
In some enteric microbiotas, methanogens act as major consumers of hydrogen. However, in our dataset, we found no evidence of potential methanogens (members of the Euryarchaeota), even though they have been reported in chicken cecal samples in other studies . Sulfate reducers are also potential hydrogen consumer, but we found no evidence of Desulfovibrio spp. in our 16S and metagenomic data. Acetogenesis, the reduction of CO2 to acetyl-CoA, provides another potential hydrogen sink, but we detected only six acetyl CoA synthase genes, (from the Ruminococcae and Lachnospiraceae) in our dataset (table S5 and Fig. 6).
The key genes involved in each pathway that could potentially use hydrogen are shown. The size of pie chart reflects abundance of the gene class; numbers indicate quantity of genes assigned to each taxon.
Bacterial uptake NiFe hydrogenases link the oxidation of H2 to the reduction of anaerobic electron acceptors such as NO3−, SO4−, fumarate or CO2 . Bacteria that possess these enzymes (such as Wollinella and Campylobacter) could potentially provide a significant hydrogen sink capable of facilitating acetate production in microbial ecosystems , . We found sequences for twelve uptake hydrogenases in our chicken cecal metagenome (Figure 6 and table S5), in several of the most abundant genera in this ecosystem (Megamonas, Helicobacter, Campylobacter). We speculate that the high abundance of these genera in the chicken cecum might, at least in part, be explained by their potential to remove hydrogen, which is likely to benefit not only the other members of the microbial community, but also, indirectly, the host, by improving recovery of energy from food.
Relationship Between Size of OTU and Percentage Identity to Genbank Sequences. Graph showing relationship between the number of reads in an OTU and its percentage identity to the closest match in the NCBI nr database. Those OTUs showing limited homology still were comprised of many reads hence increasing the validity of the OTU.
Summary of Previous Studies on the Chicken Metagenome. Graph showing the relative amounts of different taxa within the chicken ceca from a number of different studies using different techniques (1). J.H.A. Apajalahti and A. Kettunen p. 124–126 in G. Perry, ed. Avian Gut Function in Health and Disease,2006. (2). W.E. Holben, K.P. Feris, A. Kettunen and J.H.A Apajalahti Appl. Envron. Microb, 70:2263–2270, 2004. (3). J.R. Lu, U. Idris, B. Harmon, C. Hofacre, J.J. Maurer and M.D. Lee. App. Environ. Microb. 69:6186–6823, 2003. ().G.R. Siragusa and M.G. Wise J. Appl. Microb. 102:1138–1149, 2007. (5). M.D. Cressman, Z. Yu, M.C. Nelson, S.J. Moeller, M.S. Liburn and H.N. Zerby. Appl Environ Microbiol 76:6572–6582 2010. (6). J.L. Danzeisen, H.B. Kim, R.E. Isaacson, Z.J. Tu and T.J. Johnson. Plos One 6:e27949. 2011 (7). This study.
Taxonomic Assignment of Metagenomic data. (A) The percentage of each EGT that could be assigned to a given taxonomic level. EGTs were assigned taxonomy based on the protein coding sequences they contained (see methods). (B) Comparison of phylogenetic assignment from pyrosequenced 16S amplicons and EGTs. (i) 260 000 110 bp reads assigned by MEGAN using the Least Common Ancestor algorithm. (ii) Taxonomic assignment of protein sequences from the assembled metagenomic data, abundance corrected by coverage. (iii) Taxonomic distribution of the genomes, abundance corrected by coverage and size of genome. (iv). Taxonomy of 16S, assigned using the rdp database project taxonomy.
Organization of the Metagenomic Data. (A) Flow diagram of the procedures used to analyse the metagenomic data. Grey boxes show the programs that were used for each step (B) Database scheama showing the structure of the MySQL database holding the metagenomic information.
Comparison of OTU pipelines. (A) Rareaction curves for each OTU picking method. The four reps for each chicken were merged (i) Mothur (ii) Custom (iii) Uparse. (B) PCA plots showing differences in OTU composition of each chicken (colour) when analysed by different pipelines (circle Uparse, triangle custom method and square Mothur). OTUs were merged according to taxonomic assignment by the RDP classifier at the family level. The axes represent the two principal components which are responsible for the most variation in the samples.
Summary of the reads obtained for each of the 10 chickens.
Broad Taxonomic assignment of a subset of reads from the chicken metagenome.
Genes which contain both GH motifs and carbohydrate binding motifs (CBMs) or dockerin motifs.
Predicted acetate kinase/phosphotransferase loci present in the chicken metagenome.
Predicted Acetyl CoA synthase genes and the large subunit of uptake hydrogenases present in the Chicken metagenome.
The assignment of putative NSP degrading enzymes based on SEED/KEGG annotation and GH domains.
Comparisons of different techniques to cluster reads into OTUs from raw reads.
List of all OTUs showing assigned taxonomy and their relative abundance in each sample.
We thank Jan Kreft for suggestions that we should investigate the hydrogen metabolism of the chicken cecal microbiota.
Data deposition: The sequence reads reported in this paper have been deposited in the Sequence Read Archive (Accession no SRP019776). The EGTs are deposited the WGS database at NCBI (Accession no. SUB195828). Annotated EGTs and OTU information can also be accessed at the chicken microbiome project website (http://chickenmicrobiome.xbase.ac.uk/).
Conceived and designed the experiments: MJP CWP. Performed the experiments: MJS CC. Analyzed the data: MJS. Contributed reagents/materials/analysis tools: TAC. Wrote the paper: MJS CWP MRB MJP.
- 1. Fumihito A, Miyake T, Takada M, Shingu R, Endo T, et al. (1996) Monophyletic origin and unique dispersal patterns of domestic fowls. Proc Natl Acad Sci U S A 93: 6792–6795.
- 2. Harre R (1981) Aristotle: the embryology of the chick. Great scientific experiments Oxford: Phaidon. 31–38.
- 3. Darwin C (1868) The variation of animals and plants under domestication. London: Murray.
- 4. Burt DW (2007) Emergence of the chicken as a model organism: implications for agriculture and biology. Poult Sci 86: 1460–1471.
- 5. Hillier LW, Miller W, Birney E, Warren W, Hardison RC, et al. (2004) Sequence and comparative analysis of the chicken genome provide unique perspectives on vertebrate evolution. Nature 432: 695–716.
- 6. McWhorter TJ, Caviedes-Vidal E, Karasov WH (2009) The integration of digestion and osmoregulation in the avian gut. Biol Rev 84: 533–565.
- 7. Apajalahti J (2005) Comparative gut microflora, metabolic challenges, and potential opportunities. J Appl Poultry Res 14: 444–453.
- 8. Zhou W, Wang Y, Lin J (2012) Functional Cloning and Characterization of Antibiotic Resistance Genes from the Chicken Gut Microbiome. Appl Environ Microb 78: 3028–3032.
- 9. Huyghebaert G, Ducatelle R, Van Immerseel F (2011) An update on alternatives to antimicrobial growth promoters for broilers. Vet J 187: 182–188.
- 10. Clench MH, Mathias JR (1995) The Avian Cecum – a Review. Wilson Bull 107: 93–121.
- 11. Karasawa Y (1999) Significant role of the nitrogen recycling system through the ceca occurs in protein-depleted chickens. J Exp Zool 283: 418–425.
- 12. Mcnab JM (1973) Avian Ceca – Review. World Poultry Sci J 29: 251–263.
- 13. Obst BS, Diamond JM (1989) Interspecific Variation in Sugar and Amino-Acid-Transport by the Avian Cecum. J Exp Zool: 117–126.
- 14. Jozefiak D, Rutkowski A, Martin SA (2004) Carbohydrate fermentation in the avian ceca: a review. Animal Feed Science and Technology 113: 1–15.
- 15. Hegde SN, Rolls BA, Coates ME (1982) THe effects of the gut microflora and dietary fibre on energy utilization by the chick. Br J Nutr 48: 73–80.
- 16. Jorgensen H, Zhao XQ, Knudsen KE, Eggum BO (1996) The influence of dietary fibre source and level on the development of the gastrointestinal tract, digestibility and energy metabolism in broiler chickens. Br J Nutr 75: 379–395.
- 17. Van Immerseel F, Boyen F, Gantois I, Timbermont L, Bohez L, et al. (2005) Supplementation of coated butyric acid in the feed reduces colonization and shedding of Salmonella in poultry. Poultry Sci 84: 1851–1856.
- 18. Kien CL, Blauwiekel R, Bunn JY, Jetton TL, Frankel WL, et al. (2007) Cecal infusion of butyrate increases intestinal cell proliferation in piglets. J Nutr 137: 916–922.
- 19. Stams AJM (1994) Metabolic Interactions between Anaerobic-Bacteria in Methanogenic Environments. Anton Leeuw Int J G 66: 271–294.
- 20. Salanitr JP, Blake IG, Muirhead PA (1974) Studies on Cecal Microflora of Commercial Broiler Chickens. Appl Microbiol 28: 439–447.
- 21. Salanitro JP, Blake IG, Muirhead PA, Maglio M, Goodman JR (1978) Bacteria Isolated from Duodenum, Ileum, and Cecum of Young Chicks. Appl Environ Microb 35: 782–790.
- 22. Mead GC (1989) Microbes of the Avian Cecum – Types Present and Substrates Utilized. J Exp Zool: 48–54.
- 23. Zhu XY, Zhong TY, Pandya Y, Joerger RD (2002) 16S rRNA-based analysis of microbiota from the cecum of broiler chickens. Appl Environ Microb 68: 124–137.
- 24. Lan PT, Hayashi H, Sakamoto M, Benno Y (2002) Phylogenetic analysis of cecal microbiota in chicken by the use of 16S rDNA clone libraries. Microbiology and Immunology 46: 371–382.
- 25. Lu JR, Idris U, Harmon B, Hofacre C, Maurer JJ, et al. (2003) Diversity and succession of the intestinal bacterial community of the maturing broiler chicken. Appl Environ Microb 69: 6816–6824.
- 26. Amit-Romach E, Sklan D, Uni Z (2004) Microflora ecology of the chicken intestine using 16S ribosomal DNA primers. Poultry Sci 83: 1093–1098.
- 27. Wei S, Morrison M, Yu Z (2013) Bacterial census of poultry intestinal microbiome. Poult Sci 92: 671–683.
- 28. Danzeisen JL, Kim HB, Isaacson RE, Tu ZJ, Johnson TJ (2011) Modulations of the Chicken Cecal Microbiome and Metagenome in Response to Anticoccidial and Growth Promoter Treatment. Plos One 6: e27949.
- 29. Qu A, Brulc JM, Wilson MK, Law BF, Theoret JR, et al.. (2008) Comparative Metagenomics Reveals Host Specific Metavirulomes and Horizontal Gene Transfer Elements in the Chicken Cecum Microbiome. Plos One 3.
- 30. Kim T, Mundt E (2011) Metagenomic analysis of intestinal microbiomes in chickens. Methods Mol Biol 733: 185–194.
- 31. Sergeant MJ, Constantinidou C, Cogan T, Penn CW, Pallen MJ (2012) High-Throughput Sequencing of 16S rRNA Gene Amplicons: Effects of Extraction Procedure, Primer Length and Annealing Temperature. Plos One 7: e38094.
- 32. Edgar RC (2013) UPARSE: highly accurate OTU sequences from microbial amplicon reads. Nat Methods 10: 996–998.
- 33. Caporaso JG, Kuczynski J, Stombaugh J, Bittinger K, Bushman FD, et al. (2010) QIIME allows analysis of high-throughput community sequencing data. Nat Methods 7: 335–336.
- 34. Noguchi H, Park J, Takagi T (2006) MetaGene: prokaryotic gene finding from environmental genome shotgun sequences. Nucleic Acids Res 34: 5623–5630.
- 35. Huson DH, Auch AF, Qi J, Schuster SC (2007) MEGAN analysis of metagenomic data. Genome Res 17: 377–386.
- 36. Li H, Durbin R (2009) Fast and accurate short read alignment with Burrows-Wheeler transform. Bioinformatics 25: 1754–1760.
- 37. Eddy SR (1998) Hidden Markov models and genome sequence analysis. Faseb J 12: A1327–A1327.
- 38. Punta M, Coggill PC, Eberhardt RY, Mistry J, Tate J, et al. (2012) The Pfam protein families database. Nucleic Acids Res 40: D290–D301.
- 39. Petersen TN, Brunak S, von Heijne G, Nielsen H (2011) SignalP 4.0: discriminating signal peptides from transmembrane regions. Nat Methods 8: 785–786.
- 40. Liu B, Pop M (2009) ARDB-Antibiotic Resistance Genes Database. Nucleic Acids Res 37: D443–D447.
- 41. de Castro E, Sigrist CJA, Gattiker A, Bulliard V, Langendijk-Genevaux PS, et al. (2006) ScanProsite: detection of PROSITE signature matches and ProRule-associated functional and structural residues in proteins. Nucleic Acids Res 34: W362–W365.
- 42. Teeling H, Meyerdierks A, Bauer M, Amann R, Glockner FO (2004) Application of tetranucleotide frequencies for the assignment of genomic fragments. Environ Microbiol 6: 938–947.
- 43. Schloss PD, Westcott SL, Ryabin T, Hall JR, Hartmann M, et al. (2009) Introducing mothur: Open-Source, Platform-Independent, Community-Supported Software for Describing and Comparing Microbial Communities. Appl Environ Microb 75: 7537–7541.
- 44. Warnecke F, Luginbuhl P, Ivanova N, Ghassemian M, Richardson TH, et al. (2007) Metagenomic and functional analysis of hindgut microbiota of a wood-feeding higher termite. Nature 450: 560–U517.
- 45. Cantarel BL, Coutinho PM, Rancurel C, Bernard T, Lombard V, et al. (2009) The Carbohydrate-Active EnZymes database (CAZy): an expert resource for Glycogenomics. Nucleic Acids Res 37: D233–D238.
- 46. Louis P, Flint HJ (2007) Development of a semiquantitative degenerate real-time PCR-based assay for estimation of numbers of butyryl-coenzyme A (CoA) CoA transferase genes in complex bacterial samples. Appl Environ Microb 73: 2009–2012.
- 47. Vignais PM, Billoud B (2007) Occurrence, classification, and biological function of hydrogenases: An overview. Chem Rev 107: 4206–4272.
- 48. Gagen EJ, Denman SE, Padmanabha J, Zadbuke S, Al Jassim R, et al. (2010) Functional Gene Analysis Suggests Different Acetogen Populations in the Bovine Rumen and Tammar Wallaby Forestomach. Appl Environ Microb 76: 7785–7795.
- 49. Torok VA, Allison GE, Percy NJ, Ophel-Keller K, Hughes RJ (2011) Influence of antimicrobial feed additives on broiler commensal posthatch gut microbiota development and performance. Appl Environ Microbiol 77: 3380–3390.
- 50. Siragusa GR, Wise MG (2007) Quantitative analysis of the intestinal bacterial community in one- to three-week-old commercially reared broiler chickens fed conventional or antibiotic-free vegetable-based diets. J Appl Microbiol 102: 1138–1149.
- 51. Clokie MR, Millard AD, Letarov AV, Heaphy S (2011) Phages in nature. Bacteriophage 1: 31–45.
- 52. Warren RAJ (1996) Microbial hydrolysis of polysaccharides. Annu Rev Microbiol 50: 183–212.
- 53. Hubener K, Vahjen W, Simon O (2002) Bacterial responses to different dietary cereal types and xylanase supplementation in the intestine of broiler chicken. Arch Anim Nutr 56: 167–187.
- 54. Bayer EA, Belaich JP, Shoham Y, Lamed R (2004) The cellulosomes: Multienzyme machines for degradation of plant cell wall polysaccharides. Annu Rev Microbiol 58: 521–554.
- 55. Pope PB, Denman SE, Jones M, Tringe SG, Barry K, et al. (2010) Adaptation to herbivory by the Tammar wallaby includes bacterial and glycoside hydrolase profiles different from other herbivores. P Natl Acad Sci USA 107: 14793–14798.
- 56. Medie FM, Davies GJ, Drancourt M, Henrissat B (2012) Genome analyses highlight the different biological roles of cellulases. Nat Rev Microbiol 10: 227–U.
- 57. Meimandipour A, Soleimanifarjam A, Azhar K, Hair-Bejo M, Shuhaimi M, et al. (2011) Age effects on short chain fatty acids concentrations and pH values in the gastrointestinal tract of broiler chickens. Arch Geflugelkd 75: 164–168.
- 58. Bott M, Pfister K, Burda P, Kalbermatter O, Woehlke G, et al. (1997) Methylmalonyl-CoA decarboxylase from Propionigenium modestum cloning and sequencing of the structural genes and purification of the enzyme complex. Eur J Biochem 250: 590–599.
- 59. Colby GD, Chen JS (1992) Purification and Properties of 3-Hydroxybutyryl-Coenzyme-a Dehydrogenase from Clostridium-Beijerinckii (Clostridium Butylicum) Nrrl-B593. Appl Environ Microb 58: 3297–3302.
- 60. Louis P, Flint HJ (2009) Diversity, metabolism and microbial ecology of butyrate-producing bacteria from the human large intestine. Fems Microbiol Lett 294: 1–8.
- 61. Louis P, McCrae SI, Charrier C, Flint HJ (2007) Organization of butyrate synthetic genes in human colonic bacteria: phylogenetic conservation and horizontal gene transfer. Fems Microbiol Lett 269: 240–247.
- 62. Eeckhaut V, Van Immerseel F, Croubels S, De Baere S, Haesebrouck F, et al. (2011) Butyrate production in phylogenetically diverse Firmicutes isolated from the chicken caecum. Microb Biotechnol 4: 503–512.
- 63. Eastwood M (1995) Dietary fibre: mechanisms of action in human physiology and metabolism.
- 64. Macfarlane S, Macfarlane GT (2003) Regulation of short-chain fatty acid production. P Nutr Soc 62: 67–72.
- 65. Saengkerdsub S, Anderson RC, Wilkinson HH, Kim W-K, Nisbet DJ, et al. (2007) Identification and quantification of methanogenic archaea in adult chicken ceca. Appl Environ Microb 73: 353–356.
- 66. Howlett RM, Hughes BM, Hitchcock A, Kelly DJ (2012) Hydrogenase activity in the foodborne pathogen Campylobacter jejuni depends upon a novel ABC-type nickel transporter (NikZYXWV) and is SlyD-independent. Microbiology 158: 1645–1655.
- 67. Gross R, Simon J (2003) The hydE gene is essential for the formation of Wolinella succinogenes NiFe-hydrogenase. Fems Microbiol Lett 227: 197–202.
- 68. Allgaier M, Reddy A, Park JI, Ivanova N, D'Haeseleer P, et al. (2010) Targeted discovery of glycoside hydrolases from a switchgrass-adapted compost community. Plos One 5: e8812.
- 69. Zhu LF, Wu Q, Dai JY, Zhang SN, Wei FW (2011) Evidence of cellulose metabolism by the giant panda gut microbiome. P Natl Acad Sci USA 108: 17714–17719.
- 70. Hess M, Sczyrba A, Egan R, Kim TW, Chokhawala H, et al. (2011) Metagenomic Discovery of Biomass-Degrading Genes and Genomes from Cow Rumen. Science 331: 463–467.