Composition and diversity of the bacterial community in snow leopard (Uncia uncia) distal gut

Abstract

Intestinal microflora influences many essential metabolic functions, and is receiving increasing attention from the scientific community. However, information on intestinal microbiota, especially for large wild carnivores, is insufficient. In the present study, the bacterial community in the feces of snow leopards (Uncia uncia) was described based on 16S rRNA gene sequence analysis. A total of 339 near-full-length 16S rRNA gene sequences representing 46 non-redundant bacterial phylotypes (operational taxonomical units, OTUs) were identified in fecal samples from four healthy snow leopards. Four different bacterial phyla were identified: Firmicutes (56.5 %), Actinobacteria (17.5 %), Bacteroidetes (13 %), and Proteobacteria (13 %). The phylum Actinobacteria was the most abundant lineage, with 40.4 % of all identified clones, but Clostridiales, with 50 % of all OTUs, was the most diverse bacterial order. The order Clostridiales was affiliated with four families: Clostridiaceae I, Lachnospiraceae, Peptostreptococcaceae, and Ruminococcaceae. Lachnospiraceae was the most diverse family with 17 OTUs identified. These findings were basically consistent with previous reports on the bacterial diversity in feces from other mammals.

Introduction

The animal gastrointestinal tract harbors a complex microbial ecosystem, as has been proven by both classical culture-based and molecular biology techniques. During the long-term co-evolution between host and microorganisms, indigenous microbial communities have become a crucial part of the host, and alterations of this complex ecosystem have been associated with the host’s age, diet and health (Ley et al. 2008b; Tilg and Kaser 2011). A large number of clinical trials have shown that imbalances in the gut microbiota can be correlated to many gastrointestinal diseases, such as inflammatory bowel diseases (Frank et al. 2007a; Xenoulis et al. 2008). Consequently, research into the composition and diversity of host gut microbiota is imperative to understanding the developmental mechanisms of several gastrointestinal diseases (Frank et al. 2007b; Xenoulis et al. 2008).

Previous research, beginning with that of Rahner (1901), applied cultivation methods, which provided much information pertaining to the microbial diversity of gut ecosystems. However, culture-dependent methods are extremely time consuming, and inefficient in identifying the anaerobic bacteria that are dominant within intestinal microbiota. With the development of molecular biology techniques, some new molecular tools, such as denaturing gradient gel electrophoresis (DGGE), restriction fragment length polymorphism (RFLP), PCR-amplification of 16S rRNA genes, and metagenomic approaches have been applied successfully to evaluating the microbial composition and diversity of natural ecosystems. Particularly, 16S rRNA gene sequence analysis has proved to be an efficient and sensitive molecular method, and is now frequently applied in gastrointestinal microbial research.

Recently, several studies have used molecular biology techniques to take a deep look inside the gastrointestinal tract, and have yielded valuable results. It has been recognized that the host’s diet and phylogeny can both influence the composition and diversity of the gut bacterial community, which increases from carnivory to omnivory to herbivory (Ley et al. 2008a). Recently, a study using massive parallel 16S rRNA gene pyrosequencing revealed that feline and canine intestinal tracts harbor quite different bacterial communities. At the same time, inter-individual differences in the relatives abundance of major bacterial groups were less in cats than in dogs (Handl et al. 2011). In addition, several potential pathogens, such as Clostrium perfringens, Escherichia coli, Pseudomonas spp., and Enterococcus spp., are commonly detected in healthy animal intestines. This finding should demonstrate that potential pathogens are possibly indigenous members of the intestinal microbiota of healthy animals (Handl et al. 2011; Wu et al. 2012). To date, several studies concerned the microbial ecosystem of humans and of domestic animals while the intestinal bacterial community of large wild carnivores remains poorly described.

The snow leopard (Uncia uncia) is a mysterious feline that inhabits the high, remote mountains of Central Asia at altitudes of 3,000–4,500 m (Jackson 1996). Unfortunately, increasing human activities have caused this species to meet the criteria for endangered status in the IUCN Red List. Until recently, studies on the snow leopard have focused mainly on habitat utilization, population conservation, and phylogenetic classification, but no study has assessed the microbial communities in the intestinal tract. Therefore, our objective in this study was to describe the microbial diversity in fecal samples from healthy snow leopards in comparison to gut microbiota of other animals using 16S rRNA gene-based analysis.

Materials and methods

Sample collection

Fresh fecal samples were collected (within half an hour after defecation) from four (three males and one female) unrelated snow leopards that were raised in Xining Zoo of Qinghai, China during May 2012. The age of the animals ranged from 4 to 8 years old, and their weight from 32 kg to 48 kg. The tested animals were raised semi-freely singly and fed a diet based on fresh raw meat (mutton and beef) and live hare. All snow leopards were fed by the same personnel at the same time each day. All the animals were clinically healthy with no history of antibiotic or probiotic therapy 3 months prior to sample collection. Fecal samples were collected immediately after spontaneous defecation using sterile sampling bags. Feces from the same individual were seen as one fecal sample. Samples were frozen immediately and preserved at −80 °C without any additives or pretreatment until further analysis.

DNA extraction

An aliquot from each fecal sample (250 mg) was homogenized in a sterile tube and genomic DNA was extracted using a QIAamp® DNA Stool Mini Kit (Qiagen, Hilden, Germany) under the guidance of the QIAamp® DNA Stool Handbook. To avoid bias, genomic DNA was extracted in duplicate for each sample and extracts from the same sample were pooled. Purified DNA was stored at −20 °C until use.

PCR amplification and purification

PCR amplification was carried out individually on all samples. The 16S rRNA gene was amplified using universal bacterial primer F (5′-GAGAGTTTGATCCTGGCTCAG-3′, E. coli position 7–27) and primer R (5′-TACGGCTACCTTGTTACGAC-3′, E. coli position 1493–1511). Both primers were purchased from the Sunbiotech (Beijing, China). DNA was amplified using the following reaction conditions: 2 μL of each primer (10 μM), 3 μL template DNA, 5 μL 10 × Ex PCR buffer, 4 μL of dNTPs (10 mM each), 1 μL Ex Taq DNA polymerase (5U/μL) and 1 μL 20 mg/ml bovine serum albumin in a 50 μL reaction volume. The above reagents were all purchased from TaKaRa (Dalian, China). The DNA was then amplified in a DNA Thermal Cycler (GeneAmp® PCR System 9700, Applied Biosystems, Foster City, CA), using the following PCR protocol: initial denaturing at 94 °C for 5 min, 35 cycles of denaturation at 94 °C for 30 s, annealing at 54 °C for 30 s, extension at 72 °C for 1 min 30 s, and a final extension at 72 °C for 10 min. For all samples, between four and eight independent PCR reactions were performed.

PCR products belonging to the same sample were pooled and visualized on agarose gels electrophoresis (0.5 × TBE buffer). The were extracted with a QIAquick® Gel Extraction Kit (Qiagen) following the manufacturer’s instructions. Approximately equal amounts of purified PCR products of each sample were then pooled and stored at 4 °C in sterilized double distilled water.

Amplicons cloning and sequencing

PCR products were ligated into pMD^TM18-T Vector and then transformed into DH5α^TM-T1^R competent E. coli cells (TaKaRa) according to the manufacturer’s instructions. Clones were grown and selected (blue/white) on Luria-Bertani medium, containing ampicillin (100 μg/mL), X-gal (100 μg/mL) and IPTG (0.5 mM) at 37 °C overnight. Positive clones were stored in glycerol at −80 °C for future plasmid extraction.

Plasmids were extracted using the Perfectprep® BAC 96 plasmid purification kit (Eppendorf, Shanghai, China). The 16S rRNA gene insert was sequenced using an ABI PRISM Big Dye Terminator Cycle Ready Reaction Kit and an ABI PRISM 3730 DNA Sequencer (Applied Biosystems) with the bacterial universal primers (E. coli 27F and 1492R). The sequences were assembled using vector NTI advance software (version 10.0).

Sequence analysis

All near-full-length sequences obtained were edited to exclude the PCR primer-binding sites using the VecScreen available on NCBI. Additionally, sequences were checked by BELLEROPHON (Huber et al. 2004) available through the Ribosomal Database Project (RDP), and putative chimeras were excluded from further analysis.

Sequences were aligned with the CLUSTAL_W, which is contained in MEGA software package (version 5.1 Bata) (Tamura et al. 2011). A PHYLIP distance matrix was generated and used as an input file for the DOTUR software (version 1.5) (Schloss and Handelsman 2005) to determine phylotypes (operational taxonomical units, OTUs) at 97 % similarity.

The non-chimeric sequences were compared with existing 16S rRNA gene sequences using GenBank and RDP on an 80 % confidence threshold, and the closest 1–3 neighbor(s) for each sequence was obtained. Distal gut bacterial diversity in the snow leopard was demonstrated by constructed phylogenetic trees based on the neighbor-joining algorithm using the MEGA software. Evolutionary distances were inferred by the Jukes-Cantor model and branch stability was assessed by 1,000 replicates bootstrap analysis. The Archaea Aquifex pyrophilus (GenBank accession number: M83548) was selected as out-group.

Statistical analysis

The coverage of the clone library was calculated using the formula C = [1 − (n/N)] × 100 according to Good (1953), where n is the number of OTUs represented by one sequence and N is the total number of sequences of the clone library.

Bacterial community diversity was calculated by the Shannon–Weaver diversity index, which is defined as H’ = −∑p _i ln (p _i), where p _i is the proportion of individual bacteria found in a certain species.

Nucleotide sequence accession numbers

Sequences submitted to the GenBank database in this study were prefixed by UUF (Uncia uncial fecal, e.g., UUF001). A total of 339 near-full-length 16S rRNA gene sequences were deposited with the GenBank database (accession numbers: KC245156–KC245494).

Results

A total of 500 clones were selected randomly. Of these, 382 clones contained an insert with a sequence of adequate quality, of which 43 were identified as possible chimeras and excluded from further analysis. With 97 % sequence similarity, a total of 46 OTUs representing 339 near-full-length sequences were used in the subsequent phylogenetic analysis. The Coverage C and Shannon Weaver index (H') of the 16S rDNA clone library were 94.4 % and 2.75, respectively.

BLAST analysis revealed that 27 OTUs (58.7 % of all OTUs) showed <98 % sequence similarity with existing 16S rRNA gene sequences in the GenBank database. Four major phylogenetic lineages were identified: Actinobacteria (17.5 %), Bacteroidetes (13.0 %), Firmicutes (56.5 %), Proteobacteria (13.0 %). Additionally, further classification indicated that 23 OTUs (50.0 %) were affiliated with the order Clostridiales of the phylum Firmicutes. The phylogenetic positioning of OTUs in each phylum is shown in Figs. 1, 2, 3 and 4.

Fig. 1

Dendrogram illustrating the phylogenetic affiliation of operational taxonomical units (OTUs) isolated from the snow leopard fecal samples for Actinobacteria. The tree was inferred using neighbor-joining algorithm. Near-full-length 16S rDNA sequences were aligned to their closest neighbour(s) in RDP database. Aquifex pyrophilus was used as out-group

Fig. 2

Dendrogram showing the phylogenetic affiliation of OTUs isolated from the snow leopard gastrointestinal tract for Bacteroidetes. See legend of Fig. 1 for explanation

Fig. 3

Dendrogram showing the phylogenetic affiliation of OTUs isolated from the snow leopard gastrointestinal tract for Firmicutes. See legend of Fig. 1 for explanation

Fig. 4

Dendrogram showing the phylogenetic affiliation of OTUs isolated from the snow leopard gastrointestinal tract for Proteobacteria. See legend of Fig. 1 for explanation

Actinobacteria

A total of 137 reads representing eight phylotypes were classified within the phylum Actinobacteria (Fig. 1). Two different bacterial genera within the family Coriobacteriaceae were identified. The genus Collinsella was the predominant subgroup with 136 clones representing seven OTUs. One OTU was affiliated within the genus Slackia. According to the Blast analysis, most clones affiliated within this phylum showed < 98 % sequence similarity with the GenBank database entries (Table 1).

Table 1 BLAST results and sequence analysis of Actinobacteria representative sequence

Bacteroidetes

A total of 26 clones representing six individual OTUs were identified within the phylum Bacteroidetes (Fig. 2). This phylum was exclusively represented by the family Bacteroidaceae. One OTU represented by two clones had 99 % similarity to Bacteroides massiliensis type strain (AY126616); one OTU had 99 % similarity to Bacteroides vulgatus (CP000139) and represented nine clones (Table 2).

Table 2 BLAST results and sequence analysis of Bacteroidetes representative sequence

Firmicutes

More than half of all the OTUs (56.5 %) containing 98 clones were classified within the phylum Firmicutes (Fig. 3), which was the most diverse phylum in the feces of the snow leopard. Three different bacterial classes were identified: Clostridia, Erysipelotrichia, and Negativicutes. Further classification showed that these classes were respectively represented by the orders Clostridiales, Erysipelotrichales, and Selenomonadales. A total of 83 clones, representing 23 phylotypes, were affiliated to the order Clostridiales. Only one clone fell into the order Erysipelotrichiales and it was ascribable to the family Erysipelotrichaceae. Finally, the order Selenomonadales was represented by the families Acidaminococcaceae and Veillonellaceae, each containing one OTU.

Clostridiales, the largest order in the Firmicutes, was divided into four families: Clostridiaceae I, Ruminococcaceae, Peptostreptococcaceae, and Lachnospiraceae. At the family level, Lachnospiraceae was predominant in this phylum with 63 clones representing 17 OTUs identified. In this family, one OTU showed 99 % similarity with Blautia hansenii (AB534168). In the family Clostridiaceae, two OTUs displayed 99 % and 98 % similarity with Clostridium hiranonis (JN713315), respectively, and one OTU showed 99 % similarity with Clostridium perfringens (CP000246) (Table 3).

Table 3 BLAST results and sequence analysis of Firmicutes representative sequence

Proteobacteria

A total of 78 clones representing six OTUs were affiliated with the phylum Proteobacteria (Fig. 4). Among four different identified classes in this phylum, the class Gammaproteobacteria was the most diverse group with 65 clones representing three individual OTUs. Within this class, the family Pseudomonadaceae was the predominant group with 64 clones representing two OTUs. At the genus level, Pseudomonas was the most common representative subgroup in family Pseudomonadaceae, followed by the genus Sutterella represented by nine clones in class Betaproteobacteria. The two last OTUs were affiliated to the class Alphaproteobacteria (one clone) and Epsilonproteobacteria (three clones), respectively.

Based on the BLAST results, OTUs in this phylum all showed >97 % sequence similarity to GenBank database strain entries. Within the most common class Gammaproteobacteria, the OTUs were closely related to previously cultured bacteria including Escherichia coli (AP010953), Pseudomonas aeruginosa (HE978271) and Pseudomonas fluorescens (D86001). Additionally, OTUs belonging to classes Alphaproteobacteria, Betaproteobacteria and Epsilonproteobacteria were classified as Phyllobacterium myrsinacearum (AY785315), Sutterella stercoricanis (AJ566849) and Campylobacter upsaliensis (JX912527), respectively (Table 4).

Table 4 BLAST results and sequence analysis of Proteobacteria representative sequence

Discussion

According to our results, four major phylogenetic lineages were identified: Actinobacteria, Bacteroidetes, Firmicutes and Proteobacteria. In the present study, no Fusobacteria were identified in the clone libraries. However, this phylum was found as a major component of the intestinal microbial community of dogs (Suchodolski et al. 2008) and wolves (Zhang and Chen 2010), while it was less represented both in humans (Wang et al. 2005) and in different animal species such as cats (Ritchie et al. 2008), pigs (Leser et al. 2002), and horses (Daly et al. 2001). It could be hypothesized that Fusobacteria, if present in low numbers within the complex gut community of snow leopards, have been underestimated by the PCR procedure used here.

Phylogenetic analysis revealed that the Firmicutes was the most diverse phylum in the snow leopard distal intestinal tract and Clostridiales was the most diverse bacterial order. These findings are consistent with previous studies on other mammals (Mentula et al. 2005; Ritchie et al. 2008; Desai et al. 2009; Garcia-Mazcorro et al. 2012), in which Firmicutes has been frequently reported to be among the dominant bacterial groups in various segments of the gastrointestinal tract or in feces.

Furthermore, within the order Clostridiales, Lachnospiraceae was the most diverse family in the feces of the snow leopard, with 17 OTUs identified. This finding was consistent with a previous report in wolves (Zhang and Chen 2010) showing that Lachnospiraceae sequences were predominant within Clostridiales in wolf feces. On the contrary, data from humans (Eckburg et al. 2005), cats (Ritchie et al. 2010) and dogs (Mentula et al. 2005; Suchodolski et al. 2008) suggested that, within the phylum Firmicutes, Clostridium spp. was the most common genus in feces or in intestinal tract. Therefore, we speculate that the family Lachnospiraceae may account for a larger proportion of gut microorganisms in wild carnivores compared to humans or domestic animals.

In the present study, Actinobacteria was the most abundant phylum in the intestinal bacterial community (40 % of all identified sequences compared to Firmicutes 29 %, Proteobacteria 23 %, and Bacteroidetes 8 %). However, this phylum generally makes up a small proportion of the gut bacterial community in humans and animals (Wang et al. 2005; Suchodolski et al. 2008; Middelbos et al. 2010). Only a limited number of studies, like the present one, have reported Actinobacteria as dominant in the mammal gut (Andersson et al. 2008; Desai et al. 2009). In the present study, Collinsella was the most frequent genus in the Actinobacteria.

Sequences belonging to the phylum Bacteroidetes were less abundant (7.7 % of all sequences) in snow leopard feces. Moreover, previous studies in humans and animals (Wang et al. 2005; Andersson et al. 2008; Handl et al. 2011; Tun et al. 2012) have shown that the relative abundance of Bacteroidetes varies significantly at intra- and inter-species level.

The Proteobacteria (including E. coli-like organisms) was the other phylum identified in the fecal microbial community of the snow leopard. These data support the results of studies on feline gut microbiota (Ritchie et al. 2008; Desai et al. 2009). Moreover, we evidenced the presence of potential pathogen species ascribable to the genus Pseudomonas. Interestingly, the four snow leopards participating in our study were all clinically healthy and showed no signs of any gastrointestinal diseases. Other authors (Handl et al. 2011) detected several potential pathogens species in feces of healthy cats and dogs and they hypothesized that these populations are indigenous components of the gut microbiota of healthy animals. However, the exact role of these microorganisms in gastrointestinal diseases requires further study.

It is recognized that differences in gastrointestinal microbiota might be due to adaption to the diet and gut morphology of the host. The snow leopard’s principal natural prey species are blue sheep (Pseudois nayaur) and ibex (Capra sibirica), whose distribution coincides closely with snow leopard habitat. Snow leopard also preys on marmot (Marmota spp.), hare (Lepus spp.) and small rodents (Jackson 1996). The diet administered to the tested animals is basically consistent with natural feeding. Thus, we consider our results representative of the composition and diversity of wild snow leopard fecal microbiota.

In conclusion, in the present study, the composition and diversity of snow leopard fecal microbiota were evaluated using 16S rRNA gene analysis. We also compared our study results with related data from other mammals. The results obtained facilitate the next step in understanding of the composition and diversity of the microbial community in the snow leopard’s intestinal tract. Moreover, further studies are warranted to provide a more detailed description of the intestinal microbiota of wildlife and of the contribution of different gastrointestinal bacterial populations to digestion, immunology and nutrition.