Characterization of the gut microbiota in the golden takin (Budorcas taxicolor bedfordi)

The gut microbiota of mammals is a complex ecosystem, which is essential for maintaining gut homeostasis and the host’s health. The high throughput sequencing allowed us to gain a deeper insight into the bacterial structure and diversity. In order to improve the health status of the endangered golden takins, we first characterized the fecal microbiota of healthy golden takins using high throughput sequencing of the 16S rRNA genes V3–V4 hypervariable regions. Our results showed that, Firstly, the gut microbiota community comprised 21 phyla, 40 classes, 62 orders, 96 families, and 216 genera. Firmicutes (67.59%) was the most abundant phylum, followed by Bacteroidetes (23.57%) and Proteobacteria (2.37%). Secondly, the golden takin maintained higher richness in spring than in the winter while community diversity and evenness was not significantly different. Thirdly, four female golden takins demonstrated highly similar microbiota and the five golden takin males had relatively highly similar microbiota. All of our results might indicate that the fecal microbiota of golden takins were influenced by the season and the animal’s sex. The findings provided theoretical basis regarding the gut microbiota of golden takins and may offer new insights to protect this endangered species.


Introduction
The takin (Budorcas taxicolor) is a large Himalayan ungulate in the family Bovidae, which is classified into four subspecies based on distinct differences in physical characteristics and geographic location, including the golden takin (B. t.bedfordi), the Mishmi takin (B. t.taxicolor), the Sichuan takin (B. t. Tibetana), and the Bhutan takin (B. t.whitei). The golden takin and the Sichuan takin are endemic subspecies of China. Furthermore, relative to the other subspecies, the golden takin is listed as vulnerable by the International Union for Conservation of Nature (IUCN 2008) and is under state protection (category I) in China. The golden takin is a type of social bovid herbivore. As a ruminant, this animal primarily forages on the browse of trees and shrubs (Schaller et al. 1986). In the spring and summer, golden takins undertake seasonal migration to the higher elevations of the uppermost tree line. During the autumn and winter, they move to lower elevations with smaller groups (Ming et al. 2003).
The gut microbiota offers substantial benefits to the host, such as aiding in digestion, promoting immune system development and competing for niches with pathogens (Cadwell 2014). To date, there have been many studies investigating the gut microbiota in herbivores (Bian et al. 2013;Costa et al. 2012;Liu et al. 2014). For instance, Firmicutes and Bacteroidetes, represented by the Ruminococcaceae, Lachnospiraceae, Rikenellaceae and Prevotellaceae families (Bian et al. 2013), dominate the microbiota in the healthy rhinoceros. Costa et al. (2012) described the predominance of Firmicutes (68%) in the feces of healthy horses, followed by Bacteroidetes (14%) and Proteobacteria (10%). The core microbiota of donkeys is dominated by Firmicutes (64% males and 64% females) and Bacteroidetes (23% males and 21% females), followed by Verrucomicrobia, Euryarchaeota, Spirochaetes and Proteobacteria (Liu et al. 2014). Considerable evidence supports the complexity of the gut microbiota in many mammals and the necessity of gut bacterial communities to maintain gut homeostasis.
However, surprisingly, little research was on the gut microbiota of the golden takin. Therefore, in present research, we characterized the fecal microbiota of nine golden takins by performing high-throughput sequencing. On this basis, we also compared gut microbiota variation from winter to spring, which is the season of animals likely to get sick. Our result indicated that season was a factor that might impact the composition of the gut microbiota.

Animal selection
Nine healthy golden takins were enrolled in this study (Table 1). Four fresh fecal samples (Group A) were collected at Jinan wild animal park during the late morning in December 2015. Five fresh fecal samples (Group B) were collected at Jinan Animal Park during the late morning in April 2016. Group A and group B respectively represented the winter and spring feces samples. All of the animals were about 8 years old and the two environments were roughly same. The experiment was approved by the Qufu Normal University Animal Care and Use Committee. None of the animals had received anti-inflammatory drugs or antimicrobials within the prior 4 months, and none had a gastrointestinal-related disease. Samples were collected off the ground immediately after defecation, and each sample was instantly transferred into plastic tubes. The samples were stored at 4 °C with long-term storage at −80 °C for further analysis.

DNA extraction, 16S rRNA gene PCR and sequencing
Total genomic DNA was isolated from fecal samples using a QIAamp DNA Stool Mini Kit (Qiagen, Germany) according to the manufacturer's recommended protocol. We used a UV-Vis spectrophotometer (NanoDrop 2000c, USA) to determine DNA quantification and quality. DNA concentrations were measured to ensure final concentrations greater than 20 ng/μl for PCR amplification of the V3-V4 hypervariable regions of the 16SrRNA gene using the primers 341F (CCTAYGGGRBGCASCAG) and 806R (GGACTACNNGGGTATCTAAT), as previously described (Hugerth et al. 2014). Amplification was performed in a 30-μl volume, and the amplified fragment size was approximately 410 bp. Specifically, PCR reactions contained 15 μl of Phusion Master Mix (2er), 1.5 µl of each primer, 10 µl of microbial genomic DNA, and 2 µl ddH 2 O to constitute a final volume of 30 µl. The following PCR protocol was used: 1 min at 98 °C for initial denaturation, 35 cycles of 10 s at 98 °C for denaturation, 30 s at 55 °C for annealing, and 30 s at 72 °C for elongation, and a final extension at 72 °C for 5 min. PCR products were mixed with same volume of 1× loading buffer (containing SYBR green) and separated by electrophoresis on a 2% agarose gel. Samples containing a bright band between 400 and 450 bp were chosen and mixed in equidensity ratios. Then, this mixture of PCR products was purified with a Qiagen Gel Extraction Kit (Qiagen, Germany). Sequencing libraries were generated using a TruSeq ® DNA PCR-Free Sample Preparation Kit (Illumina, USA) following the manufacturer's recommendations, and index codes were added. Library quality was assessed with a Qubit@ 2.0 Fluorometer (Thermo Scientific) and an Agilent Bioanalyzer 2100 system. Finally, the DNA library was sequenced on an IlluminaHiSeq 2500 platform, and 250-bp paired-end reads were generated.

Bioinformatics analysis
Paired-end reads were assigned to samples based on their unique barcodes and truncated by cutting off the barcodes and primer sequences. Paired-end reads were then merged using FLASH (V1.2.7) (Magoč and Salzberg 2011). The raw tags were subjected to specific filtering conditions to obtain high-quality clean tags using the QIIME (V1.7.0) (Caporaso et al. 2010) quality control process. The tags were compared with the reference database (Gold database) using the UCHIME algorithm (UCHIME Algorithm) (Edgar et al. 2011) to detect chimeric sequences, which were then removed. Finally, effective tags were obtained. Sequence analysis was performed using Uparse software (Uparse v7.0.1001) (Edgar 2013). Sequences demonstrating greater than or equal to 97% similarity were assigned to the same operational taxonomic unit (OTU). For each representative sequence, the GreenGene Database (Desantis et al. 2006) was used and the RDP classifier algorithm (Version 2.2) (Wang et al. 2007) was employed to annotate taxonomic information. To study the phylogenetic relationships between different OTUs and identify dominant species in different samples (groups), multiple sequence alignment was carried out using MUSCLE software (Version 3.8.31) (Edgar 2004). Ultimately, we obtained 54,748 sequences for diversity analysis. Alpha diversity was evaluated by calculating six indices, specifically the Observed-species, Chao1, Shannon, Simpson, ACE and Good-coverage indices. All indices were calculated

Results
During quality control, a total of 670,356 high-quality sequences with an average length of 410 bp per sample were obtained from the nine fecal samples. The statistical estimates of species richness for the sequences from the nine samples that demonstrated a genetic distance of 3%, the total number of sequences, the coverage, and the number of OTUs were shown in Table 2. The rarefaction curves and rank abundance curves were generated using R software (Version 2.15.3). The rarefaction curves tended to approach the saturation plateau, while the rank abundance curves revealed a few dominant taxa and many low-abundance taxa comprising the bacterial community of the golden takin ( Fig. 1). Based on ANOSIM analysis, significant differences were observed in bacterial community structure between group A and group B (P < 0.05). When analyzing differences between the two groups, the Alpha diversity index and box-plots intuitively reflected the median, degree of dispersion, maximum, minimum and outlier of species diversity within  1 Rarefaction curve analysis of nine golden takins. Repeated samples of OTU subsets were performed to evaluate whether further sampling would likely yield additional taxa, as indicated by whether the curve reached a plateau value. The y-axis indicates the number of OTUs detected, and the x-axis indicates the number of taxa in the sequence subset analyzed. Rank abundance curves were used to estimate the richness and evenness of taxa present in the samples. The y-axis indicates the relative abundance of OTUs, and the x-axis indicates the number of OTUs according to the relative abundance from large to small. The larger the span curve on the x-axis, the higher the species richness. The smoother the curve on the y-axis, the more even the species distribution groups. The number of species in group A was lower than that in group B (Fig. 2), and this difference was significant (P < 0.05). However the difference of the community diversity and the species distribution uniformity between two groups were not significant (P > 0.05).
At the genus level, the percentage of unclassified bacteria averaged 42.52% (ranging from 36.7 to 50.27%), and the remainder were classified into 216 genera. The distribution proportion was greater than 0.5% in 10 genera (Fig. 3c). Among these 10 genera, Ruminococcaceae_UCG-010 was predominant, with an abundance of 21.44%, followed by Ruminococcaceae_UCG-005 (13.93%), Rikenellaceae_ RC9_gut_group ( To know the bacterial community profiles at the genus level, we plotted a clustered heatmap. The result indicated that the sample J1, J2, J3 and J4 (group A) grouped together, while the sample J5, J6, J7 and J8 (group B) clustered together (Fig. 4). To compare the microbial communities across the different samples, we plotted another heatmap based on Weighted Unifrac and Unweighted Unifrac distances (Fig. 5). The similarity of the samples was consistent with the clustering results. In addition, the PCA score plot revealed the separation of group A from group B (Fig. 6), which was consistent with the heatmap results.

Analysis of species abundance
We performed t test and Mann-Whitney U test to evaluate the significant differences (P < 0.05) among species. At the phylum level, Actinobacteria was significantly more abundant in group B than in group A. At the class level, Coriobacteriia and Deltaproteobacteria were significantly less abundant in group A than in group B. At the order level, Coriobacteriales was also significantly higher in group B. At the family level, Coriobacteriaceae was significantly higher in group B, while Ruminococcaceae and Bacteroidales_S24-7_group were higher in group A (Table 3). Eighteen of the 216 genera were significantly different (P < 0.05) between group A and group B (Table 4). Among these 18 genera, Parabacteroides, Sphaerochaeta, Ruminiclostridium_1 and Phascolarctobacterium were significantly higher in group A, while the others were significantly higher in group B. Additionally, 13 of the 18 genera belonged to Firmicutes.

Discussion
In present study, we first performed high-throughput sequencing to characterize the microbial community in the golden takin. Similar to previous researches (Costa et al. 2012;Oikonomou et al. 2013), the golden takin gut  (Ley et al. 2008). The advantage exhibited by Firmicutes might be related to the anatomical physiology and diet of specific species that feed primarily upon insoluble fiber and in which the cecum and colon were the major sites of fermentation (Costa et al. 2012). Within Firmicutes, Clostridiales (65.42%) dominated in the feces of golden takins. According to previous studies, Clostridiales accounted for 24-59% of human fecal samples and was also the dominant order in the canine    (Wang et al. 2005). Additionally, Ruminococcaceae dominated in the nine golden takins, and Lachnospiraceae was present at elevated levels; these bacteria were regarded as fiber-degraders in the herbivore rumen and hindgut (Bian et al. 2013;Jami and Mizrahi 2012). Trace amounts of another major fiber-degrader in the rumen, Fibrobacter, were detected in the feces of the nine golden takins. Secondly, Bacteroidetes were the secondmost predominant components of the gut flora in the nine golden takins, which were consistent with previous researches (Bian et al. 2013;Costa et al. 2012;Oikonomou et al. 2013). Conversely, Bacteroidetes constitutes the dominant group in the rumens of other ruminants (Bhatt et al. 2013;Jami and Mizrahi 2012). Bacteroidetes degraded high-molecular-weight organic matter, such as protein and carbohydrates, and complemented the host genome to degrade resistant foods such as plant cell wall compounds (including cellulose, pectin, and xylan) (Thoetkiattikul et al. 2013). Bacteroides also degraded host sources of carbohydrates, primarily derived from gastrointestinal tract secretions, such as mucins and chondroitin sulfate polysaccharides (Salyers et al. 1977). Additionally, Bacteroidetes produces butyrate, which was the product of colonic fermentation. Bacteroidetes possessed antitumor properties, which played important roles in maintaining intestinal health (Kim and Milner 2007), and participated in toxic or mutagenic bile acid metabolism and transformation (Smith et al. 2005). In our study, the families Rikenellaceae, Prevotellaceae and Bacteroidaceae dominated within this phylum. Thus, Bacteroidetes played an equally important role as Firmicutes in the digestive system of herbivorous animals. Thirdly, Proteobacteria constituted a large group of bacteria comprising various bacterial taxa that are capable of catabolizing a wide range of feedstuff components (Evans et al. 2010), which advantages these bacteria in the panda (Wei et al. 2012). These differences may be caused by different community structures in the host diet and bowel. In our study, at the genus level, the highest species distribution proportions were observed for Ruminococ-caceae_UCG-010 and Ruminococcaceae_UCG-00, which belonged to Ruminococcaceae and were involved in fiber digestion. Fourthly, unclassified bacteria in the nine samples accounted for 18.82-22.48% of sequences. According to previous studies, many animal gut microbiomes contained certain proportions of unclassified bacteria (Bian et al. 2013;Costa et al. 2012;Wu et al. 2016). The reason of this might due to the limited database of 16S RNA gene sequences and little research on classification of fecal microbes. Unclassified bacteria and their functions in different hosts required further research. The PCA and heatmap analyses indicated that the separation was observed between the microbiota of group A and group B. However, within groups, there was a high percent identity, suggesting the fecal microbiota of golden takins may be influenced by different seasons. Moreover, rarefaction curves revealed lower phylogenetic diversity and numbers of observed species (OTUs) in the winter microbiota than in the spring microbiota. At the same time, the spring microbiota exhibited higher numbers of OTUs with significant differences (Fig. 2). Sun et al. (2016) also found that the dietary shift from winter to spring affected the Tibetan macaque gut microbial composition and diversity. For our research, the diets of the nine golden takins in winter and spring were roughly the same. The significant differences (P < 0.05) among species showed that the relative abundances of Firmicutes and Bacteroidetes in spring were slightly decreased compared with those in winter, Acidobacteria differed significantly between the two seasons and was significantly higher in the spring at the phylum level. According to previous studies, Acidobacteria was present in many wetland soils and contained a number of genes encoding cellulases and hemicellulases. Thus, these bacteria might play an important roles in plant residue degradation (Kanokratana et al. 2011;Pankratov et al. 2011). At the class level, Coriobacteriia and Deltaproteobacteria were significantly more abundant in the spring than in the winter. Coriobacteriia belonged to Actinobacteria, and Deltaproteobacteria belonged to Proteobacteria. Actinobacteria was an important secondary metabolite (enzyme and antibiotics) producer that also played this role in animal intestines, and the secondary metabolites of Actinobacteria are potent antibiotics (Bull et al. 2005;Cundliffe 2006). We hypothesized that the rise in Coriobacteriia facilitated better resistance to harmful microorganisms external to the intestinal environment, and the rise in Deltaproteobacteria was attributable to decreased food quality. Instestingly, 18 genera were significantly different (P < 0.05) between spring and winter, and 13 of these 18 genera belonged to Firmicutes; among these 13 genera, two genera exhibited higher abundances in the winter than in the spring. In contrast, the other 11 genera were less abundant in the winter than in the spring. Thus, Firmicutes might play different roles in different seasons (Carey et al. 2012;Ley et al. 2008;Taglialatela et al. 2009). The weather in spring was highly variable, which was suitable for the proliferation of a variety of pathogenic microorganisms; thus, food might contain greater numbers of pathogenic microorganisms. We suspected that the seasonal differences might be due to the feeding quality, and the intestinal flora of the golden takin might change in the spring to enhance resistance and facilitate the ability to withstand changes in the outside world.
Previous researches also showed that the fecal microbiota of many mammals might be influenced by sex (Bian et al. 2013;Liu et al. 2014;Wu et al. 2016). In present study, the four female golden takins demonstrated highly similar microbiota and the five golden takin males had relatively highly similar microbiota, although J6 and J7 were relatively far from the other three samples, which might indicate that the fecal microbiota of golden takins might also be influenced by the animal's sex. Sex-related differences might be induced by sex hormones and their effects on gene expression as well as the immune system but also might be attributable to innate physiological differences between males and females (Mcclelland and Smith 2011;Zhao et al. 2013). However, the physiological mechanisms underlying sexual influence were uncertain.
In conclusion, we describe the predominant fecal bacterial populations in golden takin and offered the taxonomic baseline for further research investigating the intestinal ecosystems in golden takins. We also found that the composition of gut microbiota was influenced by the season which might help to strengthen the management of feed quality in the spring to prevent disease in these animals. These observations also increased our understanding of the bacterial ecosystems in this vulnerable animal.