Introduction

The structural polymers of plant cell walls are the most abundant source of energy from primary producers, but vertebrate herbivores lack the enzymes necessary to digest these plant polymers. They can only nutritionally exploit them through symbiosis with microorganisms that can ferment structural carbohydrates into energy-rich byproducts such as short-chain fatty acids that provide energy for the host [26, 33]. Bacteria, protozoa, and fungi constitute the microbiota of the herbivore gut, where fermentative digestion depends on the efficiency of the gut microbiota activities and on the retention time of the digesta. Thus, the digestive tract of herbivores contain a voluminous fermentation chamber [21], and on the basis of the location of this chamber, herbivores can be divided into two groups: foregut and hindgut fermenters. While foregut fermenters degrade cellular soluble and structural plant carbohydrates into volatile fatty acids before they can reach the small intestine, hindgut fermenters degrade the remaining structural carbohydrates that bypassed digestion in the small intestine [1, 32].

In foregut fermenters, the fermentation chamber is anterior to the acid stomach, as in cows, camels, hippopotamuses, colobine monkeys, sloths, marsupials, and, in a single case in birds, the hoatzin [26]. In hindgut fermenters, the chamber is posterior to the acid stomach. Hindgut fermentation occurs either in the expanded colon of generally big mammals (horses, rhinoceros, elephants, tapirs, manatees, lagomorphs, howler monkeys) or in the cecum of generally small animals such as some arboreal marsupials and rodents [12, 32]. However, the biggest rodent, the capybara, Hydrochoerus hydrochaeris [32], is also a cecal fermenter.

Capybaras belong to the caviomorphs, a rodent group found only in South and Central America within the suborder Hystricomorpha [39]. They inhabit the seasonally flooded savannas and wetlands of Venezuela, Colombia, Brazil, Argentina, and other South American countries (except Chile). Adult capybaras are social animals, with little or no sexual dimorphism in size and a body weight of ~50 kg [10], being the largest living rodent and the largest known cecum-fermenting mammal [2]. They are grazers, with an exclusively herbivorous diet dominated by grasses [10]. They spend the early afternoon in the water, while in the evening and night they alternately graze and rest outside the water [15].

As with all caviomorphs, capybaras are coprophagous and re-ingest the special morning feces directly from the anus [15]. Coprophagy in capybaras is more frequent in the dry season, coinciding with the time of fewer food resources and highest cellulose content in the grass biomass. The capybara is adapted to grazing, with a fermenting cecum from which cultivable bacteria has previously been characterized [25]. The aim of the present study was to perform a molecular characterization of the bacterial microbiota in the capybara cecum.

Materials and Methods

Animals and Sampling

Five adult capybaras were captured at Hato Santa Luisa, a cattle ranch located in Apure State, Venezuela (8°19′ N, 70°16′ W), in March 2009, during the dry season. The sampling took place during the capybara culling, carried out annually under a permit from the Venezuelan Ministry of the Environment, in which a large number of animals are hunted, mostly for their meat [24]. Immediately after the animals were killed (by gunshot), we sampled three males (identified as C311, C571, and C591) and two females (C122 and C551) by dissecting the gut and accessing the cecum. We placed 1.5 ml of cecal contents into a sterile microcentrifuge tube and immediately stored it in liquid nitrogen, to be later transferred to an ultralow temperature freezer until analysis.

DNA Extraction and Amplification

DNA was extracted from ~200 mg cecal contents from each animal, using the PowerSoil DNA Isolation Kit (MO BIO Inc., Carlsbad, CA, USA), and was kept frozen (−20°C) until use. Cecal DNA was PCR-amplified using 16S ribosomal primers 27F (5′-AGR GTT TGA TCM TGG CTC AG) and 1492R (5′-GGT TAC CTT GTT ACG ACT T), using eight different annealing temperatures (gradient from 48°C to 58°C, extension at 72°C). The PCR mix contained 25 μl of PCR Master Mix Cat No. M7505 (Promega, Madison, WI, USA) with ~50 ng of DNA template and 10 pmol of each primer using the same condition as described in [13].

Pooled products from each of the eight different annealing temperatures were purified using a PCR purification kit (Qiagen, Valencia, CA, USA).

PhyloChip DNA Hybridization

The 16S rDNA was hybridized onto the PhyloChip (Affymetrix, CA, USA) as previously described [3]. The G2 PhyloChip microarray has 506,944 probes representing ~8,700 bacterial and archaeal taxa [7]. The microarray has been validated and used to characterize bacterial communities [3, 7]. Its high sensitivity has been demonstrated, detecting 2.5-fold more diversity than cloning [4], although there are no species-level taxa obtained, as with Sanger sequencing. Instead, each operational taxonomic unit (OTU) is based on an average of 25 probe pairs, each consisting of a perfectly matched and a mismatched probe, and represents 16S rRNA gene sequences with 0–3% sequence divergence [7]. The sample amplicons were fragmented (to 50–200 bp) using DNase I (0.02 U/mg DNA; Invitrogen, USA) and One-Phor-All buffer (NJ, USA). Biotin labeling was performed with deoxyribonucleotransferase (Promega, USA). DNA was denatured at 99°C for 5 min and hybridization onto the PhyloChip was performed overnight at 48°C at 60 rpm. Scanning of the arrays was done using the GeneArray Scanner (Affymetrix, CA, USA) as previously reported [3, 13]. A taxon was reported present if at least 90% of the probe pairs in the set (probe fraction) had a perfect match probe with at least 130 times the square of background intensity and 1.3 times the mismatch probe intensity.

Data Analysis

To determine the common bacterial community composition, a core taxa (shared OTUs between at least four animals) was determined. Rank abundance curves were drawn from the data to visualize species richness and overall diversity using the vegan package in “R” http://www.R-project.org. A heatmap was calculated with hybridization fluorescence scores for the obtained OTUs using Pearson’s correlation as the similarity metric and average linkage clustering in “R” using the package made4 [9].

In order to compare the individual cecal communities, community analyses were performed using FastUniFrac [14]. The community analyses are based on a clearcut tree [11, 31] generated from our data, an environment file with the PhyloChip-detected taxa for each individual, and a category mapping file consisting of a table with the sample names and its metadata. To determine the raw distances between bacterial communities in each pair of capybara cecal samples, we used the UniFrac metric (unweighted), resulting in a distance matrix where the smaller distances indicate the most similar pairs.

Principal coordinate analysis was performed unweighted to determine if the samples were distributed along any axes of variation that could be interpreted easily.

Results and Discussion

This work characterizes the composition and structure of the cecal community of the wild capybara. We found a total of 933 bacterial OTUs belonging to 182 families in 21 phyla. Figure 1 shows the individual rank abundance curves. The richness in the capybara cecum is lower than the ~1,400 OTUs observed in the hoatzin crop, also using the PhyloChip [13]. Consistent with previous studies of the human colon [28] or the hoatzin crop [13] microbiomes, the capybara cecal bacterial community has a high inter-individual bacterial variation. Only half of the OTUs (464 out of 933 OTUs) were shared by all individuals. The individual variation was evident in the pattern and intensity of the different OTUs in Fig. 2, in which two males that cluster together (and share 74% of the OTUs) and one male and two females are together in a separate cluster (and share 69% of the OTUs). The clusters differ in the presence of Chlamydiae and phyla Thermodesulfobacteria, TM7, OP10, marine group A, and NC10. Beta-diversity analysis of the cecal bacterial communities using principal coordinate analysis shows the individual Unifrac distances (Fig. S1; Table S2), showing closer distances (higher similarity) between the communities of the two female individuals. However, the low number of animals does not allow one to draw any conclusion in relation to sex differences.

Figure 1
figure 1

Rank–abundance curves of cecal bacterial communities for each capybara. Relative abundances were based on values of hybridization fluorescence in the PhyloChip

Full size image
Figure 2
figure 2

Heatmap and dendograms with bidirectional clustering of all 933 bacterial OTUs and specific host communities, showing the inter-host similarities. The color codes red/green indicate, respectively, a high/low in relative abundance for each OTU (shown here in columns). The pie charts on the bottom depict the phyla composition of the two main taxa clusters (here divided by a dashed yellow line): A—327 OTUs in 16 phyla that are in higher abundance in most samples and B—606 OTUs in 20 phyla that are less abundant, except for some OTUs in capybara male C311

Full size image

The majority of OTUs detected in the cecum of the capybara belonged to the Firmicutes (322 of 933 OTUs detected in the chip) and Proteobacteria (301 OTUs). The other phyla present in the capybara cecum were the Bacteroidetes (76 OTUs), the Actinobacteria (69 OTUs), and the Sphirochaetes (37 OTUs) (Table 1). The core bacterial microbiota was composed by 575 OTUs in 20 phyla (excluding Chlamydiae) with the same dominance of the Firmicutes (225 OTUs) and the Proteobacteria (155 OTUs), followed by Bacteroidetes, Actinobacteria, and Spirochaetes (Fig. 3). The phylum Firmicutes comprised 26 families with high numbers of taxa in the Clostridiaceae, Lachnospiraceae, and Peptostreptococcaceae (80, 70, and 26 OTUs, respectively) (Table 1). Within this phylum, the Clostridium, Butyrivibrio, and Eubacterium genera include known cellulolytic species commonly found in other mammalian intestines [6, 18, 19, 23], in rumen [17, 27], in the crop of hoatzins [13], as well as in the termite gut [16]. The detection of cellulolytic activity in the cecum of the capybara is consistent with the reported cellulolytic activity in cultures of isolated cecum contents [2].

Table 1 Numbers of operational taxonomic groups in 21 bacterial phyla from the cecal contents of five wild capybaras
Full size table
Figure 3
figure 3

Phyla-level composition of the core microbiota in the Capybara cecum. The figure represents the richness of OTUs per phylum

Full size image

There was high richness of the Proteobacteria in the cecum of the capybara. In our study, we only approximate the estimations of abundance through the intensity of the fluorescence of hybridized probes and cannot directly compare with the abundance estimations based on the sampling of bacterial DNA in Sanger or 454 sequencing procedures. This phylum is abundant in other mammals, ranking third (after Firmicutes and Bacteroidetes) in abundance among digestive bacteria in the hindgut of horses, pigs, rabbits, or humans [6, 18, 19, 23, 28]. In the capybara, the majority of cecal Proteobacteria OTUs belong to orders Bradyrhizobiales and Rhizobiales from Alphaproteobacteria class (Table 1). Proteobacteria in these groups are responsible for atmospheric nitrogen fixation in herbivorous ants and in rhizosphere of potato cultivars [29, 30, 34, 37]. It has been suggested that N2 fixation by the Spirochetes phylum is important to termite nitrogen economy [20]. In mammals, bacterial nitrogen fixation has been reported in rodents consuming low-nitrogen diets (such as moles) as a mechanism for nitrogen supplementation [35]. In voles and in the European beaver, atmospheric nitrogen fixed in bacteria is utilized nutritionally via coprophagy [22, 36]. This might be the case of the capybara since grasses have low N2 content and coprophagy may contribute to the capybara’s nitrogen economy. However, future studies are needed to further explore the role of N2-fixing species in the capybara nutrition.

Interestingly, Helicobacteraceae, a family commonly associated with gastrointestinal diseases in mammals [38], was one of the most predominant families from Epsilonproteobacteria in the capybara cecum (Table 1; Table S1), consistent with the report of the genus Helicobacter in the intestine of wild rodents [5, 8].

Among the less represented phyla, the Bacteroidetes included 76 OTUs with ~32% belonging to unclassified families (Table S1). Zoo capybaras seem to have a dominance of fecal Bacteroidetes (~60% of all sequences), according to Ley et al. [19], based on ~300 16S rDNA sequences. Our results are not directly comparable to Ley’s due to differences in methodology, but differences in bacterial communities of the microbiome in wild and captive animals have been reported in mammals and birds [13, 19], highlighting the importance of pursuing studies in wild animals to understand the structure of un-impacted microbiotas.

The results presented here on the structure of the cecal bacterial communities in the capybara are an important first step towards a more comprehensive understanding of the digestive physiology of this large rodent. This is a pioneer but preliminary work, and the results lead to ecological questions and hypotheses testing that will benefit from using current sequencing technologies that are more robust and informative. Future studies monitoring the effects of temporal and spatial factors on the cecal bacterial communities using metagenomics will allow further understanding of community composition and function-informative genes involved in the digestion of cellulose and other substrates in the cecum of the capybara.