Contrasting Vaginal Bacterial Communities Between Estrus and Non-estrus of Giant Pandas (Ailuropoda melanoleuca)

Bacterial infection and imbalance of bacterial community in the genitourinary system of giant panda could affect the reproductive health. In severe cases, it can also lead to abortion. In this study, 13 of vaginal secretions in the estrue (E) group and seven of vaginal secretions in the non-estrue (NE) group were used to study the composition and diversity of vaginal bacterial communities between estrus and non-estrus by 16S rRNA gene sequencing analysis. The results showed that the vaginal microbiome in giant pandas shared the same top five abundant species between estrus and non-estrus at the phylum level. However, the vaginal microbiome changed significantly during estrus at the genus level. In top 10 genera, the abundance of Escherichia, Streptococcus, and Bacteroides in the E group was significantly higher than that in the NE group (p<0.05); Azomonas, Porphyromonas, Prevotella, Campylobacter, and Peptoniphilus in the NE group was significantly higher than that in the E group (p<0.05). The richness and diversity of vaginal microbiome in giant panda on estrus were significantly lower than those on non-estrus (p<0.05). It is noteworthy that the abundance of Streptococcus, Escherichia, and Bacteroides of vagina in giant pandas maintained low abundance in the daily. Whereas, they increased significantly during estrus period, which may play an important role in female giant pandas during estrus period. It was hypothesized that hormones may be responsible for the changes in the vaginal microbiome of giant pandas between estrus and no-estrus stages.


INTRODUCTION
The community of microorganisms associated with the host have been shown to affect physiology, immunity and metabolism (Smith and Ravel, 2017). In vagina, microbes exist in co-operative relationship with the host and provide the first line of defense against the migration of opportunistic pathogens. Previous researches have indicated that the imbalance of the vaginal bacteria maybe implicated in genital tract infections, even can result in reproductive failure (Stumpf et al., 2013;Payne and Bayatibojakhi, 2014). Thus, it is important to study the composition of bacterial microbiome and compare the diversity of bacterial communities of the giant panda's vagina on estrus and non-estrus, which may aid in the prevention and treatment of genital tract infections and negative pregnancy outcomes.
The giant panda (Ailuropoda melanoleuca), as a flagship species for wildlife conservation, is a national treasure in China and also well-known in the world (Peng et al., 2007). As an umbrella species, it protects other endangered wildlife in the same nature reserve. Since the giant panda population remains threatened by environmental and anthropogenic pressure (Wei et al., 2017), maintaining stable development of the population has been the prior conservation work. Moreover, the genital health is crucial to the giant panda breeding, which is the key to keep population growth, thus the research of genital tract cannot be ignored.
So far, the major vaginal microbiome studies carried out conventional culture-dependent method, biochemical identification method and high-throughput sequencing method. However, the first two methods could not illustrate the vaginal microbiome of giant panda, since they left many uncultured and unidentified bacteria. Considering the cost, accuracy, efficiency and other factors, 16S rRNA gene sequencing analysis was applied to investigate the vaginal microbiome of giant panda. Recently, the study of composition and diversity of vaginal bacterial communities in women and animals used high-throughput sequencing of 16S rRNA gene sequencing method (Clemmons et al., 2017;Smith and Ravel, 2017).
The composition of vaginal microbiome varies in human and other animals. In human, Lactobacillus spp., Atopobium vaginae, and Streptococcus spp. were considered to be major microbiome in the vagina (Yamamoto et al., 2009). Lactobacillus spp. have been consistently linked with good vaginal health (Verhelst et al., 2004). The composition of vaginal bacterial communities is different among cows, which is affected by various factors such as species, fitness, geographical distance and differences in animal handling (Otero et al., 2000;Giannattasio-Ferraz et al., 2019). There are few records on the bacterial communities of giant panda's vagina, mainly focusing on the differences of composition and diversity between vagina and uterine (Yang et al., 2017), as well as the differences in region and age (Zhang et al., 2020). In this study, our aim is to examine the composition and diversity of bacterial communities between estrus and non-estrus in giant pandas' vagina by using high-throughput sequencing. It helps to better understand the composition and dynamics of the vaginal microbiome, and the factors associated with the effect of estrus on bacterial communities of giant panda.

Sample Collection
In this study, the samples of giant panda were collected from Chengdu Research Base of Giant Panda Breeding from September 2019 to April 2020. The giant pandas are all 6 to 26-year-old female adult individuals. The animal handling procedures including anesthesia and sampling were approved by Institutional Animal Care and Use Committee (IACUC) of the Chengdu Research Base of Giant Panda Breeding (No. 2019006) for giant panda. The disposable swabs were extended into the vagina of female giant panda to take vaginal secretion samples. A total of 13 samples of vaginal secretions of giant pandas were collected with artificial insemination operation on estrus and, respectively, named E1-E13 belong to the estrus (E) group. The other seven samples were collected on non-estrus when the giant pandas were carried out other examination and, respectively, named NE1-NE7 belong to the non-estrus (NE) group. The summary of the giant panda was shown in Table 1.

DNA Extraction of Bacteria
A total of 20 samples of vaginal secretions of giant pandas were collected in this experiment. Thirteen samples of them on estrus and seven samples of them in non-estrus were sent to BGI Co., Ltd., China (Shenzhen, China) for DNA extraction and 16S rRNA gene sequencing analysis. Bacterial DNA extraction was carried out according to the requirements of BGI Co., Ltd., China (Shenzhen, China).

PCR Amplification of 16S rRNA V4 Region
The 16S rRNA gene sequencing targeting V4 hypervariable region was amplified from samples of vaginal secretion. In brief, the primers 515F (5'-GTGCCAGCMGCCGCGGTAA-3') and 806R (5'-GGACTACHVGGGTWTCTAAT-3') were designed to amplify the V4 of 16S rRNA gene sequence from the single vaginal microbiome sample. The targeted gene was amplified

Bioinformatics Analysis
To get the clean data, the raw data was filtered to eliminate adapter contamination and low-quality readings with following methods: FLASH (fast length adjustment of short reads, v1.2.11) was used to merge the coincident clean paired-end reads into the tag (Magoc and Salzberg, 2011

Statistical Analyses
Finally, the Venn diagram and bar graph of each group was displayed by R software (v3.1.1). The Alpha diversity index was calculated by Mother (v.1.31.2), and the rarefaction cure was drawn by R software (v3.1.1) to judge the reasonable degree of sequencing data, and the Wilcoxn rank sum test were performed on Alpha diversity index between groups to obtain the difference of species diversity between two groups.
UPGMA cluster analysis with weighted Unifrac distance matrix was performed by Phytools and R software (V3.5.1). The cluster tree was used to demonstrate the results. Using iterative algorithm in QIIME (v1.80), in the case of weighted abundance species classification information, sampling analysis is carried out using 75% of the sequence number of all samples with the least sequence number. After 100 iterations, the final statistical analysis result table and Principal Co-ordinates Analysis (PCoA) display diagram are obtained, which are used to study the similarity or dissimilarity of sample community composition between samples. R (v3.4.1) was used to calculate the significance of the difference test between the two groups for the bacteria in the top 10 abundances. LEfSe software 1 was used to make LEfSe cluster diagram and LEfSe linear discriminant analysis (LDA) diagram. Wilcoxn test was used to look for differences between groups by R (v3.4.1). The function was shown graphically.

Sequencing Data
After the raw data was filtered and spliced, 1,143,725 highquality tags were produced in the E and NE groups. The sequencing quality was calculated by rarefaction curve. All of the curves were flat and the number of operational taxonomic units (OTUs) was close to saturation ( Figure 1A), indicating that the sequencing depth was sufficient to cover all species in the microbiome community. The total numbers of OTUs obtained was 969, among which 286 OTUs were shared by both groups, 300 and 383 OTUs were uniquely identified in the E and the NE groups, respectively ( Figure 1B).

OTU Abundance Analysis
The result of PCoA showed the clustering (p < 0.05) was identified for the vaginal secretions of giant panda on estrus and non-estrus (Figure 2A). UPGMA cluster number results also showed that most of the samples in the same group have similar branches ( Figure 2B). Futhermore, the results showed an obvious clustering pattern of the E group bacterial community in giant pandas' vagina, the same as the NE group, indicative of distinct differences between the vaginal microbiome on estrus and that non-estrus of giant pandas.

Community-Composition Analysis
The taxonomic classification of the sequences from the study samples resulted in 15 different phyla. The sequences that  The graph shows the LDA scores obtained from linear regression analysis of the significant microorganism groups in the two groups. When the default LDA value is more than 2 and the p value is less than 0.05, the result corresponds to a differential species.
cannot be assigned were categorized as "Unclassified. " When comparing at the phylum level, the NE and E groups shared the same top five bacterial species with little difference in overall abundance. The most dominant phylum of samples in the NE group was Proteobacteria (31.6%), followed by Firmicutes (29.0%), Bacteroidetes (18.8%), Actinobacteria (18.2%), and Fusobacteria (0.6%; Figure 3A). Proteobacteria (57.3%) was also the most abundant phylum in the E samples, followed by Firmicutes (24.0%), Bacteroidetes (11.1%) and Actinobacteria (5.4%) Fusobacteria (2.1%; Figure 3A). The estimated cumulative abundance of these five dominant phyla was above 98% of the identified OTUs. The vaginal bacterial community of giant panda in the NE and E groups were similar in composition. However, there was the discrepancy in their relative abundance at phyla level. Further analysis of the relative abundance showed that the abundance of Proteobacteria in the E group was significantly higher than that in the NE group (p < 0.01); whereas, the abundance of Actinobacteria in the NE group was significantly higher than that in the E group (p < 0.01; Figure 4A). At the genus level, a total of 244 bacterial genera were identified in our study, 171 in the E group and 212 in the NE group. Giant pandas in the the NE group had the highest Azomonas (14.8%) content, followed by Porphyromonas (12.6%), Corynebacterium (8.5%) and Campylobacter (4.7%). While giant pandas in the E group displayed the highest amount of Escherichia, accounting for 34.5% of the total microbial abundance, the second most abundant was Streptococcus (18.2%), followed by Bacteroides (8.8%) and Corynebacterium (4.3%; Figure 3B). However, the vaginal bacterial communities in the E group of giant pandas were notably different compared to the NE groups and had fewer the number of species. Our results showed that the female giant panda on estrus can lead to changes in the vaginal bacterial community at the genus level.
In accordance with analysis of the relative abundance at genus level, the top 10 species were selected to evaluate the significance of the difference test between the two groups. The results showed that there were differences in Escherichia, Streptococcus, Azomonas, Porphyromonas, Bacteroides, Prevotella, Campylobacter, and Peptoniphilus between these two groups. The abundance of Escherichia, Streptococcus, and Bacteroides in the E group was significantly higher than that in the NE group (p < 0.05); Azomonas, Porphyromonas, Prevotella, Campylobacter, and Peptoniphilus in the NE group was significantly higher than that in the E group (p < 0.05; Figure 4B).

LEfSe Analysis
The cladogram of the LEfSe analysis was showed in Figure 5. The green and red parts represented the groups of the NE and E, respectively. The red and green nodes in the cladogram denoted the bacteria playing a critical role in the NE and E group, respectively. The yellow nodes corresponded to the bacteria that did not have an important role in each group. The LDA score was showed in Figure 6, which obtained through LDA. The results showed that the microorganisms were significant difference in the vaginal secretions from giant panda on non-estrus (in the NE group) belonged to Porphyromonadaceae, Porphyromonas, Pseudomonadaceae, Azomonas, and Pseudomonadales. Morever, the results showed that those were significant difference in the vaginal secretions from giant panda on estrus (in the E group) belonged to Escherichia, Enterobacteriales, Enterobacteriaceae Streptococcus, and Streptococcaceae.

Microbial Diversity Analysis
In Figure 7, the Shannon index, Simpson index, Chao1 index and ACE index of each group were calculated to analyze α-diversity. The calculated ACE index ( Figure 7A) and Chao1 index ( Figure 7C) were positively associated with the richness of the bacterial species in each group. Our results found that the ACE and Chao1 index in the NE group were significantly higher than that in the E group (P ACE < 0.05; P Chao1 < 0.05). These results indicate that estrus reduced the overall number of bacterial species and thus the richness of the vaginal bacterial community in giant pandas.
The calculated Shannon index, (Figure 7B) was positive associated with the diversity of the microbiome, while Simpson index ( Figure 7D) was negative associated with the diversity of the microbiome in each group. Our results found that the Shannon index in the NE group were significantly higher than that in the E group (P Shannon < 0.001). The Simpson index in the NE group were significantly lower than that in the E group (P Simpson < 0.001) which indicate that estrus decreases the diversity of the vaginal microbial community in giant pandas.

Predictive Function Gene Analysis
The functional genes of the samples were analyzed by Kyoto Encyclopedia of Genes and Genomes (KEGG) database. The levels of carbohydrate metabolism and membrane transport in the E group were obviously higher than those in the NE group (p < 0.05; FDR < 0.05). The level of transcription in the E group was lower than that in the NE group (p < 0.05; FDR < 0.05; Figure 8), and the levels of galactose metabolism, phosphotransferase system (PTS), pentose and glucuronate interconversions and so on in the E group were higher than those in the NE group (p < 0.05; FDR < 0.05; Figure 9).

DISCUSSION
There are some reports on the vaginal microbiome composition of giant pandas. However, there is no report to compare the difference of vaginal microbiome of giant panda between estrus and non-estrus according to our knowledge. In this study, we explored the differences about the composition and diversity of giant pandas' vaginal microbiome between estrus and non-estrus (p < 0.05). Our results of bacterial composition showed that the vaginal microbiome shared the same top five abundant species between estrus and non-estrus at the phylum level. Whereas, at the genus level, the top 10 abundant bacterial species and their relative abundance were significant difference between estrus and non-estrus. At genus level, microbial diversity and richness of vaginal microbiome of giant panda were decrease during estrus. However the abundance of Escherichia, Streptococcus, and Bacteroides increased and became the main component of the vaginal microbiome as the results showed in Figure 4. The result of LEfSe analysis was consistent with it. It's hypothesized that the change of microbiome structure makes the vaginal bacterial communities of giant panda more adaptable to the physical changes of giant panda during estrus (Figure 5). Streptococcus was one of main bacteria in the vagina of giant panda during estrus which were significant increased compared with giant panda on non-estrus. Yang et al. (2017) and Zhang et al. (2020) have also report that Streptococcus was an important genus of bacteria living in the vagina and uterus of giant pandas on estrus. However, this is the first report describing the change of Streptococcus in vagina of giant panda between estrus and non-estrus and illustrated it may play an important role on the estrus of giant pandas. Therefore, it was hypothesized that Streptococcus maintained low daily abundance when the giant panda entered estrus. It was significantly increased and performed a lactic acid-producing function, which is similar to the effect of Lactobacillus on human. Interesting that Lactobacillus was the most common bacteria in human, but its relative abundance was very low in the vaginal microbiome of giant pandas no matter on estrus or on non-estrus which was similar with the result of Yang et al. (2017), moreover Streptococcus and Lactobacillus belongs to the same orders, Lactobacillales.
There are also some reports to support this conclusion. Streptococcus probably can decompose glycogen and utilize its degradation products to produce lactic acid and other bacteriostatic substances, which maintains the health of the host by reducing environmental PH (Thomas et al., 1979;Abbott et al., 2010) and other pathways (Gezginc et al., 2015;Wang et al., 2017). In particular, our results of functional prediction showed that the carbohydrate metabolism, galactose metabolism and other level of microorganisms in vaginal secretions on estrus were higher than that on non-estrus, which may associate with the change of Streptococcus.
The results of Escherichia and Bacteroides also interesting. Escherichia was the highest amount in the vagina of giant panda during estrus and significantly increased compared with giant panda on non-estrus. It should be one of the important probiotics in the vagina of giant pandas. Other studies have shown that some bacteria belong to Escherichia have probiotic effects. It can inhibit the growth of pathogenic microorganisms by metabolizing glucose into acidic products, and competing with other pathogenic bacteria for colonization and other ways (Hejnova et al., 2005;Nag et al., 2018). In the study of Yang et al. (2017), Escherichia are also dominant bacteria in the vagina of giant pandas. Bacteroides was the third of main bacteria in the vagina of giant panda during estrus. It also significantly increased compared with giant panda on non-estrus. The bacteria belong to Bacteroides also have the potential to be probiotics, which can maintain the health of the host by regulating host immunity, competitively inhibiting pathogens and secreting antibacterial substances (Sun et al., 2019).
Hormones may be responsible for the changes in the vaginal microbiome of giant pandas between estrus and no-estrus. Studies have shown that in female human beings, the secretion of estrogen will promote the proliferation of Lactobacillus (Pekmezovic et al., 2019). Nugeyre et al. (2019) reported that progesterone can affect the abundance of some microorganisms in the vagina of female cynomolgus macaques. Messman (Messman et al., 2020) reported that estradiol had a certain effect on the bacterial composition of Brangus heifers vagina. However, the corresponding relationship between hormones and vaginal microbiome in giant pandas still needs to be revealed by further study.

CONCLUSION
The vaginal microbiome in giant pandas shared the same top five bacterial species between estrus and non-estrus at the phylum level. However, the vaginal microbiome changed significantly during estrus at the genus level. The richness, diversity and the bacterial species at genus level of vaginal microbiome in giant panda on estrus were significantly lower than those on non-estrus. It is noteworthy that the abundance of Streptococcus, Escherichia, and Bacteroides of vagina in giant pandas increased significantly during estrus, which may play an important role in female giant pandas during estrus.

DATA AVAILABILITY STATEMENT
The datasets presented in this study can be found in online repositories. The names of the repository/repositories and accession number(s) can be found at: https://www.ncbi.nlm. nih.gov/, SAMN19090826-19090845.

ETHICS STATEMENT
The animal study was reviewed and approved by Institutional Animal Care and Use Committee (IACUC) of the Chengdu Research Base of Giant Panda Breeding (No. 2019006).

AUTHOR CONTRIBUTIONS
CY, SL, and XM contributed to conception and design of the study. SL and SC plays a guiding role in carrying out the experiment. CY, XL, and XS organized the database. XY, YUL, JA, YLi, and LL collected samples. LX, DS, and ZS performed FIGURE 9 | Wilcox Test result map of pathway difference (KEGG level 3). The relative abundance bar chart of each group is shown on the left; the log2 value of the relative abundance mean ratio of the same pathway in the two groups is shown in the middle; and the value of p and FDR values obtained by wilcox test are shown on the right. If the value of p and FDR values are less than 0.05, the pathway is significantly different between the two groups.