Gut Microbiome Composition and Metabolomic Proles of Wild and Captive Chinese Monals (Lophophorus Lhuysii)

Background (cid:0) The Chinese monal (Lophophorus lhuysii) is an endangered bird species, with a wild population restricted to the mountains of southwest China, and only one known captive population in the world. We investigated the fecal microbiota and metabolomics of wild and captive Chinese monals to explore differences and similarities in nutritional status and digestive characteristics. An integrated approach combining 16S ribosomal rRNA (16S rRNA) gene sequencing and ultra-high performance liquid chromatograph (UHPLC) based metabolomics were used to examine the fecal microbiome composition and the metabolomic prole of Chinese monals. Results: The results showed that the alpha diversity of gut microbes in the wild group were signicantly higher than that in the captive group and the core bacterial species in the two groups showed remarkable differences at all levels. Metabolomic proling revealed a concurrent difference, mainly related to galactose, starch and sucrose metabolism, fatty acid, bile acid biosynthesis and bile secretion. Furthermore, these metabolites in difference are have a strong correlation with the main microbe in genus level. Conclusions: Various factors related to diet and environmental conditions played a crucial role in shaping the gut microbiome composition and metabolomic prole. Through this study, we have established a baseline for a normal gut microbiome and metabolomic prole for wild Chinese monals, thus allowing us to evaluate if differences seen in captive specimens has an impact on their overall health and reproduction.

The gut microbiota is recognized as a coevolutionary partner with an ecosystem that is metabolically adaptable, rapidly renewable, and metabolically exible [5,6]. It performs numerous bene cial functions for the host, such as nutrient acquisition, immunomodulation and physiogenesis in response to profound lifestyle changes [7,8]. Numerous factors in uence the unique and variable community of gut microbiota present in each individual, such as host's age [9], diet composition (Guan et al. 2017), social interactions [11], gut morphology [9] and health status [12]. Thus, the study of intestinal microbial activity is helpful to understand the health and nutritional status of the host.
Meanwhile, the gut microbiota is closely related to the host's metabolic phenotype, which is a product of resident microbial communities integrated with host cells [13]. Any differences in the microbial community may have a signi cant effect on the metabolite pro les of the host's blood, urine, and fecal extracts [14], the in uence of which can be explored through untargeted metabolomics that offers new opportunities to investigate individual needs, foods and nutritional functions [5]. The combined study of fecal 16S ribosomal rRNA (16S rRNA) sequencing and untargeted metabolomics promises to uncover the inherent associations between the microbiome and the host metabolic phenotypes [14]. The two techniques have been used to elucidate the relationship between environmental and ecological factors and gut microbes [15], as well as changes in gut microbes and host metabolic phenotypes in the context of different food composition [16,17].
There are currently reports on the use of fecal 16S rRNA sequencing in wildlife, including birds [12,[18][19][20] and mammals [10,11,21,22], these studies all have observed considerable differences between wild and captive individuals, but few studies on untargeted metabolomics in wild animals, especially birds, have been reported. Therefore, the objectives of the present study were to investigate the fecal microbiota and metabolomics of wild (WCM) and captive Chinese monal (CCM).

Monal species identi cation
There were 6 and 37 samples of the feces in Yuancaopo and Hongshanding belonging to the Chinese monal, respectively, according to the species identi cation. The remaining samples were identi ed as 22 blood pheasants (Ithaginis cruentus), 6 temminck's tragopan (Tragopan temminckii), and 8 samples without ampli cation results.

Microbial Community Composition
A total of 1,466,455 high-quality reads with an average sequencing depth of 35,767 ± 1,265 reads per WCM fecal sample was conducted and classi ed into 2,132 operational taxonomic units (OTUs). Whereas, 580,736 high-quality reads with an average sequencing depth of 36,296 ± 1,062 reads per CCM fecal sample was conducted and classi ed into 4,899 OTUs. The number of OTUs present in both the wild and captive groups was 1,234, with 898 unique OTUs in the wild group, and 3,665 unique OTUs in the captive group ( Fig. 2A). The species accumulation curves (Fig. 1A) and rank abundance curves ( Fig. 1B) of the WCM and CCM fecal samples convey the richness and evenness of the microbial species in the samples, and the curve had reached a plateau. These ndings demonstrated that the sequencing data was acceptable and re ected the diversity of the gut microbiome.
The gut microbiome alpha diversity indexes in WCM were signi cantly higher than that of the CCM (P Shannon = 0.0066, P Chao1 = 0.000011) at the microbial species level. The discrepancy between the WCM and CCM was further analyzed via the partial least squares discrimination analysis (PLS-DA) plot as illustrated in Fig. 2B. The distance between the orange and blue color markers demonstrated a unique bacterial community structure in WCM vs. CCM, suggesting that there was a signi cant difference in gut microbial composition between the groups.

Microbiome-metabolome Association
Strong associations between microbiome composition (at genus level) and metabolite perturbations were revealed by a Mantel test (r = 0.825,P = 0.001). Signi cant correlations could be detected between the fecal microbiome and metabolite based on Spearman correlation coe cients (r > 0.5 or r < -0.5, P < 0.05) [15]. Among the metabolites, fatty acids, bile acid derivatives, and carbohydrate were involved in more than half of the metabolic pathways. Metabolites that were remarkably different were highly correlated with microorganisms. The correlations between speci c fecal metabolites and the top 10 most common fecal bacteria genera were shown in Fig. 6A.

Discussion
Through fecal analysis, we studied the gut microbiome structure and the metabolomic pro les in both wild (WCM) and captive Chinese monals (CCM). The results of this research indicated distinct differences of the fecal bacterial communities between the two groups, furthermore, we identi ed signi cant correlations between the fecal microbiota and metabolites.
The levels of microbiome alpha and beta-diversity from WCM were signi cantly higher than CCM, indicating a more diverse microbial community in the wild population. These differences may be the result of differences in diet composition, geographical ranges, energy utilization, climate conditions, and stress exposure in the WCM vs. the CCM [9,11].
Proteobacteria was the dominant phylum in the feces of WCM followed by Firmicutes, while Firmicutes was relatively more abundant than Proteobacteria in the feces of the captive group. The characterization of gut bacteria in several studies reveals a dominance of Firmicutes and Bacteroidetes, a feature which appears to be typical for birds. The dominance of these phyla has been described in the ceca of turkeys (Meleagris gallopavo) [18], the feces of penguins (Spheniscidae) [23] and Japanese quail (Coturnix japonica) [24], as well as in the crop of hoatzin (Opisthocomus hoazin) [25]. However, one study reports that Proteobacteria and Firmicutes are the dominant phyla in the feces of the Lady Amherst's pheasant (Chrysolophus amherstiae), Reeves's pheasant (Syrmaticus reevesii), and Cabot's tragopan (Tragopan caboti) [26], which is more consistent with our observations of the microbiome composition in the Chinese monal. Moreover, Bacteroidetes just accounted for 3.8% and 1.6%, respectively, in the WCM and CCM unlike what was observed in the turkey, penguin, quail, and hoatzin studies. Based on the results of this study, it would appear that the gut microbial composition of the Chinese monal is most re ective of that which has been detected in other species of the family Phasianidae; however, in the aforementioned cited research only two samples per group were collected and study details are lacking which highlights the need for further research on this subject.
A large diversity of gut microbes has been described as adaptive and bene cial, with diet being considered one of the most critical factors shaping gut microbial structures. The Chinese monal is an omnivorous bird. For wild populations, food is scarce in the winter months, due to their natural environment being at a high altitude and low temperatures. Whereas, the small captive population of monals receives a relatively steady diet with no change in environment except for exposure to outdoor seasonal climate conditions. Our results indicated a high abundance of Proteobacteria in WCM. Proteobacteria have greatly variable morphology and versatile functions [27], previous studies have demonstrated that increased an richness of Proteobacteria in the gut microbial ora is mainly related to energy accumulation [28][29][30], and is more abundant when the host animal has prolonged exposure to cold climates [29]. We suspect our ndings in WCM may be in response to their comparatively complex dietary composition, as well as to help them with cold weather and scarce food to get more energy.
In contrast to WCM, Firmicutes was identi ed as the dominant bacteria phylum in CCM. Compared with the WCM, we suspect captive individuals are fed a diet with a comparatively higher lipid content. Diet composition, as indicated in Table 1 showed a 6.2% crude fat in the commercial pelleted feed, which provides the base of the diet, in addition to corn which contains 3.6-5.3% crude fat (The data provided at the China Feed-database Information Network Centre: http://www.chinafeeddata.org.cn/picture/pdf/CFIC2019_1.pdf). In a 2014 study, mice were fed a high-fat diet, and the Firmicutes: Bacteroidetes ratio in mice gut microbiome increased after three weeks, with the abundance of Firmicutes increased and Bacteroidetes depleted [31]. This nding is in accordance with many former studies, suggesting that a high-fat content in diet is one of the primary factors responsible for changes observed in gut microbe composition [7,32]. Thus, the increased abundances of Firmicutes and the Firmicutes: Bacteroidetes ratio in CCW group (CCM 33.45 versus WCM 6.25) is likely the result of a higher lipid content in the diet.
Concerning the metabolome, we detected 58 signi cant metabolites that associated with 20 metabolic pathways. We found that a signi cantly higher fatty-acid content in feces of the CCM was related to unsaturated fatty acid and fatty acid biosynthesis. A previous study found that different dietary triacylglyceride composition has an in uence on fatty acid content in the feces of human neonates [33].
Thus, our nding may be related to the nutrient composition of the diet, which contains a higher level of fat and a different lipid composition than what wild specimens are likely consuming. Otherwise, our results showed that the primary microbial genera in Chinese monal feces are signi cantly related to the metabolites we identi ed; therefore, microbial metabolism is an essential aspect that needs to be evaluated. Microbial biosynthesis of fatty acid may be due to cellular structure or storage, de novo synthesis from glucose, or incorporation of exogenous fatty acids directly into lipid structures [34,35]. The content of metabolites in the feces relating to bile acid biosynthesis and bile secretion had a statistically signi cant difference between the WCM and CCM. Our results showed differences in metabolism of bile acid, an important component of bile, which plays a role in digestion and metabolism of dietary lipids and cholesterol [36,37]. Chenodeoxycholate, taurochenodeoxycholate and taurocholate are bile salts that are hydrolyzed into free bile acids by bile salt hydrolase [38]. Chenodeoxycholate amount was higher in the feces of CCM, whereas taurochenodeoxycholate and taurocholate were more abundant in the WCM. Taurodeoxycholic acid and tauroursodeoxycholic acid are taurine conjugated bile acids that were expectedly more abundant in the feces of WCM. The type of bile salt that is most abundant may re ect how the corresponding bile acids play a major role in digestion which is dependent on the dietary composition of each individual bird [39]. In the wild, Chinese monals primarily consume seeds and roots of shrubs in the winter when food is scarce, but they also consume a small amount of moss, earthworms and insect pupa [40,41]. High concentrations of total non-structural carbohydrate (TNC) reserves are usually found in root tissues of plants [42][43][44], and may serves as one of the main energy sources for WCM. Because of the seed consumption, the WCM also need to secrete bile acids for fat digestion thus providing another valuable energy source. We also found a high correlation between the bile acid metabolites and gut microbial ora. There exists a mutual regulation between intestinal microbes and bile acids, for microbiome structure appear to be a major regulator of bile acid pool size and composition [45]. Kim G et. al. (2005) found that several bacterial genera in Firmicutes and Actinobacteria can hydrolyze bile salt [38], among which Enterococcus was signi cantly related to the above metabolites in this study.
A signi cantly higher abundance of carbohydrates was found in the feces of WCM as opposed to the CCM, including galactinol, ra nose, stachyose, sucrose, and cellobiose. In plants, ra nose and stachyose are oligosaccharides synthesized from sucrose, subsequently galactitol and galactose are also involved [46]. Animals don't secrete enzymes that utilize ra nose-series oligosaccharides, which are likely to be digested by microbial enzymes at the end of the gastrointestinal tract [47]. Coon et al. (1990) reported that ileal digestibility of ra nose and stachyose was less than 1.0% in roosters; however, the digestibility, determined with excreta collection, reached 90.5% and 83.8%, respectively [48]. Previous studies have also indicated several Passerine bird species have a low digestibility of sucrose [49,50] as fecal sugar content increased after consuming sucrose solutions [50]. Cellobiose is an important hydrolytic product of cellulose degradation [51]. Plants contain signi cant quantities of cellulose, which is di cult to naturally digest by most birds species, but can be utilized by bacteria inhabiting the intestinal tract, especially within the ceca [52]. Yuzhang Wang (2007) reported that the cecum of the Chinese monal is well developed, accounting for 24.09% of the total length of the intestine [53], which may help the species utilize cellulose and other indigestible components of the diet. Microorganisms in genus that have signi cant positive correlation with these metabolites may play a big role in these processes.
Indole-containing metabolites were signi cantly different in fecal samples of CCM, such as 5hydroxyindoleacetic acid (5-HIAA) and indole-2-carboxylic acid, which are nearly 100 and 300 times higher than that in the wild, respectively. Indole and its derivatives widely exist in animals, microorganisms and plants [54], and are the main metabolites of serotonin [55]. Degg et. al. (2002) found that excretion of indole containing metabolites, such as 5-HIAA, increased after humans ingested high serotonin-containing foods including tomatoes, bananas, etc. [56]. Our results which indicate higher levels of 5-HIAA and indole-2-carboxylic acid in CCM feces is likely associated with the incorporation of tomatoes in their routine diet. These two altered indole-containing metabolites were also highly correlated with fecal microbial ora. Previous studies have suggested that intestinal bacteria can conversion of tryptophan to indole through enzymatic processing, then form indole-containing metabolites [37,57]. These studies showed that metabolic processes in intestinal bacteria are essential for the synthesis of indole-containing metabolites, therefore, it may be a speci c indicator of intestinal bacteria differences.
Finally, the last metabolite with a signi cant difference observed between the two groups is sinigrin (2propenyl glucosinolate). Sinigrin was several thousand times higher in the feces of WCM than that observed in CCM. Sinigrin, which is enriched in plants such as cruciferous vegetables, leaf mustard, and horseradish, has been associated with carbohydrate regulation and lipid metabolism and has also been shown to have anti-neoplasia and anti-microbial properties [58]. In vitro experiments con rmed that sinigrin was degraded by rat intestinal microbiota [54]. Enterococcus and Bacteroides in the human intestine can degrade glucosinolates [59], and have a signi cant positive correlation with sinigrin in this study. Therefore, based on our ndings of high levels of sinigrin in WCM feces, we can speculate that the diet of WCM also contains a large amount of sinigrin. Although the intestinal ora plays a role in the digestion process, it is more excreted through feces.

Conclusion
This is the rst study to evaluate the fecal microbiome and the fecal metabolic pro les in the Chinese monal. The results clearly demonstrate the differences of the gut bacterial composition and metabolism of the host and microbiome between wild and captive Chinese monals. The results of this study, will without a doubt, enhance our knowledge of wild and captive Chinese monals which may help guide recommendations that will affect the health and fecundity of birds in the captive breeding program. Further studies are required to determine if there is a difference in the fecal microbiome and `metabolic pro les in spring and summer months when the dietary composition is likely to change.
In November 2018, fecal samples of 11 captive birds (3 males and 8 females) were collected at the Chinese Monal Conservation and Research Center, where the altitude is 1582.0 meters and the temperature range from 9 ºC -13 ºC. These monals are maintained in four enclosures, with three enclosures containing one male each paired with one or two females, while one enclosure contains only three sub-adult females. Feces was collected by placing clean cardboard beneath the nightly roosting perches, overnight. The following morning at 8 AM, four fecal samples per enclosure were collected from the cardboard with an attempt to collect feces with varying colors and shapes.
Approximately 1 g of fecal material was collected from the central section of each fecal dropping, from both wild and captive animals, and placed into 1.5 mL sterile collection tubes (Corning, New York, USA) and immediately frozen in liquid nitrogen. Thereafter, the samples were shipped to the Chengdu Research Base of Giant panda Breeding within 24 h., where they were stored at -80 ºC for subsequent analysis.

Measurement of the main components of the commercial pellet feed
The commercial pellet (New hope group, Chengdu, China) being fed to the CCM was analyzed to measure the primary nutritional components according to Chinese national standards. Moisture was determined by direct drying method [60]. Crude fat was determined by petroleum ether extraction method [61]. Crude protein was determined by Kjeldahl method [62]. Crude ber was determined by Soxhlet extraction method [63]. The raw sequencing data were processed using the Quantitative Insights Into Microbial Ecology (QIIME, v1.8.0) software package. The low-quality sequences were ltered through criteria as previously described [64], and following chimera detection, the remaining high-quality sequences were clustered into

Monal Species Identi cation
OTUs with a threshold of 97% sequence similarity. OTU taxonomic classi cation was conducted via BLAST searching the representative sequences set against the Greengenes Database [65] using the best hit [66].
Sequence data analyses were primarily calculated using QIIME and R packages (v3.2.0). Generated OTUlevel ranked abundance curves to compare the richness and evenness of OTUs among samples. Venn diagram was obtained to visualize the mutual and unique OTUs between groups using R package "VennDiagram". Alpha diversity was performed and discrepancies analyzed between the two groups using the wilcoxon test. Beta diversity analysis was performed to investigate the structural variation of microbial communities across samples using UniFrac distance metrics [67] and visualized via PLS-DA. The microbiota compositional pro les could be achieved at different taxonomic levels, including phylum, class, order, family and genus, wherewith the taxa abundances at these levels were statistically compared among samples or groups by Metastats [68].

Metabolomic Analyses
Nine samples were selected from WCM and CCM, respectively, for untargeted metabolomics analyses, which was performed with liquid chromatography/mass spectrometry (LC/MS). An Agilent 1290 In nity LC ultra-high performance liquid chromatograph (UHPLC) (Agilent, Palo Alto, USA), after sample pretreatment. The mass spectrometer (MS) used was a Triple TOF 6600 system (AB/Sciex, Foster City, USA) equipped with an electrospray ionization (ESI) source in positive (ESI+) and negative (ESI-) ion modes. Samples were mixed in equal amounts to prepare quality control (QC) samples, which were spaced evenly among the injections. All experimental samples were randomly distributed throughout the run in order to monitor the precision and stability of the method during its operation. The raw LC/MS data were converted to mzXML les using Proteo Wizard software. The data were subsequently processed using XCMS for peak detection, alignment, and data ltering [69,70]. Mass accuracy tolerance within 25 ppm was used as the mass window for the database search and secondary spectral map matching, and the self-built database of the laboratory was retrieved. For the data extracted by XCMS, ion peaks of group sum > 50% were deleted. Pattern recognition was carried out by applying the software SIMCA-P 14.1 (Umetrics, Umea, Sweden).
PLS-DA was performed for the supervised multivariate statistical analysis, and heat maps were generated using a hierarchical clustering algorithm to visualize the metabolite difference within the data set. Singledimensional statistical analysis includes Student's t-test and multiple of variation analysis. The differential metabolites were submitted to the KEGG website (http://www.genome.jp/kegg/pathway.html) for related pathway analysis.
The correlation matrix between the gut micro ora-related metabolites and gut bacterial species was generated using Pearson's correlation coe cient. Ethics approval and consent to participate

Abbreviations
The experimental protocol was approved by the Chengdu Research Base of Giant Panda Breeding and in accordance with the current laws on animal welfare and research in China.

Consent for publication
Not applicable.

Availability of data and materials
The datasets used and/or analysed during the current study are available from the corresponding author on reasonable request.

Figure 3
Page 19/25 The fecal microbiome composition pro les at the phylum (A) and genus (B) level in the captive (CCM) and wild Chinese monal (WCM).