Social behavior of musk deer during the mating season potentially influences the diversity of their gut microbiome

The study area and sample collection

The study area was the MCP in Sichuan, China. Fresh feces from each individual were collected in the morning in July 2018 and placed into a 15 ml sterile tube. Thus, each individual had one sample tube. Because the deer live alone, the feces could be assigned to each individual and detailed information (age and sex) was available for each sample. We collected 179 fecal samples from 179 individuals (141 adult deer) among four of nine regions in MCP (Fig. 1A): 42 from Region 7 (25 females and 17 males, a total of 31 adults); 43 from Region 12 (16 females and 27 males, a total of 38 adults); 47 from Region 13 (25 females and 22 males, a total of 37 adults); and 47 from Region 15 (31 females and 16 males, a total of 35 adults). We collect one fecal sample for each deer. All feces were preserved in refrigerators at −20 °C and shipped to the lab with dry ice. All the collections were consistent with animal welfare.

16S rRNA gene sequencing and analysis

We extracted DNA from the feces using a QIAamp DNA Stool Mini kit (Qiagen, Valencia, CA, USA). The V4 region of the 16S rRNA gene was amplified using 515F (5′-GTGCCAGCMGCCGCGG-TAA-3′) and 806R (5′-GACTACHVGGGTWTCTAAT3′) primer pair (Caporaso et al., 2012). PCR was performed in a 20-µl flask using 10 ng of the DNA template, 2.5 mM dNTPs, 5 µM for each primer, 5 ×FastPfu buffer, and FastPfu polymerase. PCR thermo cycling: an initial denaturation at 95 °C for 5 min followed by 35 cycles for amplification (95 °C for 30 s, 55 °C for 30 s, and 72 °C for 45 s), with a final extension step at 72 °C for 10 min. The PCR amplification products were sent to Hangzhou Mingke Biotechnology Co., Ltd. for Illumina MiSeq high-throughput sequencing.

We processed the raw sequence data using QIIME1.9 (Caporaso et al., 2010). The function search was used to detect chimerism and remove low-quality sequences (default parameters: Window size: 20 base pair; Minimum read length: 50 base pair) (Edgar, 2010). The operational taxonomic unit (OTU, or the sequences with >97% identity) was classified by annotation against the SILVA132 database (Christian et al., 2013). In order to decrease sequencing depth bias, we rarefied our sequencing depth at about 24,770 sequences per sample based on the smallest number of sequences of the sample. The alpha diversity (e.g., Shannon index) was calculated for each sample. To evaluate the effect of region, age, and sex on the Shannon index, we performed general linear model analysis in SPSS (SPSS, 1990). To determine the dissimilarity in the gut microbiome community between the different regions, we performed pairwise comparisons among regions using unweighted Unifrac distance (Lozupone et al., 2011). Moreover, we performed one-way PERMANOVA (9,999 permutations) on Bray–Curtis dissimilarities in species abundance in PAST3 (Hammer, Harper & Ryan, 2001) to evaluate the effect of region, age, and sex on the gut microbiota composition.

The composition of captive musk deer

The dominant gut microbiome phyla in MCP were Firmicutes, Bacteroidetes, and Proteobacteria, which made up over 80% of total sequences. The dominant genera included Christensenellaceae R 7 group, Ruminococcaeae UCG 005, Bacteroides, and Acinetobacter. Captivity has the potential effect on the mammal gut microbiome composition compared to that of the wild field (Clayton et al., 2016). Here, the gut microbiome composition of the captive musk deer was similar across different captive centers (Sun et al., 2020; Zhao et al., 2019). However, wild musk deer showed significant enrichment of Firmicutes, and the relative abundance of Bacteroidetes was very low (Sun et al., 2020), which may be caused by different living environments (e.g., diet, and habitat).

The significant effect of breeding regions on the alpha diversity of the musk deer

The general linear model revealed the effects on the Shannon index of the captive musk deer due to the breeding region (F = 9.930, P = 0.000), age (F = 2.157, p = 0.035), and the interception (Region*Age*Sex) (Fig. 2A). Considering that the highest F value and lowest p value were found in the breeding region variable, we suggested that the breeding cage might have a more profound effect on the alpha diversity of the musk deer gut microbiome than the other factors (Figs. 2A and 2B). Previous research in another captive center also found no significant difference in Shannon index between the sexes (Zhao et al., 2019). Thus, there may be no significant differences in alpha diversity of the captive musk deer gut microbiome linked to sex. However, there was a difference among different breeding groups.

Social behavior of musk deer during the mating season potentially contributes to the gut microbiome dissimilarity among the breeding regions in the same center

PCoA using the unweighted Unifrac distance revealed some differences in the gut microbiome community among the breeding regions (Fig. 2C). One-way PERMANOVA test revealed the effects on the gut microbiome community due to the breeding region (F = 5.043, p = 0.000) and sex (F = 2.603, p = 0.006). The breeding region, with the largest F vale and the lowest p value, may have a more profound effect on gut microbiota than the sex of the deer. Furthermore, the unweighted Unifrac distance within each breeding region was smaller than those between the breeding regions (pairwise comparisons, Fig. 2D), which indicated high similarity in the gut microbiome community within each breeding region and high dissimilarity among the breeding regions.

These breeding regions were located in the same center and had similar living conditions (e.g., same diet). Due to their solitary character, the captive adult musk deer are singly raised in the cage with no direct social contact most of time (Meng et al., 2006). However, during the mating season (November to January) in MCP, about 40% of the deer take part in breeding within each region, and one adult male and six adult females live together in one large cage for about three months. We observed that many individuals displayed social behavior such as grooming, body contact, and mating during this period. Social behaviors have been found to shape the animal gut microbiome in many species (Antwis et al., 2018; Moeller et al., 2016; Moeller et al., 2018; Tung et al., 2015). Thus, we speculated that the social behavior during musk deer mating season led to the similarity in the gut microbiome community with each breeding region. Here, we described a simple and clear correlation between social behavior and the gut microbiome community in musk deer under the same living conditions (e.g., same center and same diet).

Some additional factors may have influenced the gut microbiome community within each breeding region. When living in the same cage, resulting in exposure to feces and coprophagy (Bo et al., 2020; Ridaura et al., 2013), mice have been found to share the same gut microbiome, leading to the similarity in the gut microbiome community. Here, each breeding region has several activity areas for the doe after their baby deer weaning (about September), and each activity area has about five mother deer and they live together for recovering (about two months) until the next mating season. No coprophagy was observed in the musk deer in this study. Thus, we deduced that possible contact with the feces from the other deer in the common area also led to the putative gut microbiome transmission and the similarity in the gut microbiome community among the individuals within each breeding region. The similarity of the musk deer gut microbiome within the same breeding region was stable, even after breeding season.