Introduction

The extreme cold, high ultraviolet radiation, low oxygen content, and poor food resources of high-altitude environments pose great challenges to animal survival1,2. The gut microbiota plays an important role in a variety of physiological activities in animals, such as cardiovascular activity, food digestion, energy metabolism, nutrient homeostasis, immune regulation, and maintenance of body temperature3,4,5,6,7,18,19. Among them, the production of SCFAs is related to diet, microbiota composition, and the host19. The macronutrient composition of the diet determines the amount and source of fermentable substrates for the gut microbiota and is a major driver of gut microbiota structure and function19,20. The SCFAs produced by gut microbiota when fermenting host foods are mainly acetate, propionate, and butyrate, which mediate energy metabolism and physiological regulatory processes in the host21,22. There are also a few branched-chain fatty acids (BCFAs), that although in low abundance, can also mediate metabolic processes in the host23,24,25. The gut microbiota of high-altitude ruminants helps the host adapt to the high-altitude environment by reducing the emission of methane and elevating the production of volatile fatty acids11.

Nonhuman primates (NHPs) are the closest biological relatives of humans, originated from tropical lowlands and are primarily found in warm tropical and subtropical environments, with a few species having radiated into temperate alpine forests26,27. The structure and composition of gut microbiota in NHPs are closely related to factors, such as dietary composition28,29, habitat occupancy30, social interactions30, eco-environmental factors31, and health status32,33. The mechanisms of high-altitude adaptation in NHPs, including physiological, behavioral, and genetic, have been widely studied34,35,36. However, less is known about the role of gut microbiota in high-altitude adaptation.

The gut microbiota of NHPs responds to changes in season and differences in habitat environment9,28. The plasticity of the gut microbiota in Ethiopian geladas (Theropithecus gelada) largely provides dietary and metabolic flexibility to the host and may be a key factor in allowing them to thrive in a changing environment37. There have been many reports on the interaction between gut microbiota composition and environmental adaptation in nonhuman primates based on the 16S rRNA genes of the microbiota; however, studies on the interaction between the functional gene composition of gut microbiota and metabolite composition in the environmental adaptation of NHPs are largely lacking.

Rhesus macaques (RMs) are the most widely distributed small omnivorous NHPs and are currently a widely used biological model in medical and biological research given their high physiological similarity to humans14. In China, wild RMs are widely distributed from sea level to the Qinghai Tibetan Plateau (altitude exceeding 3000 m). Given its strong adaptability, the RM is an ideal model for investigating the adaptation mechanisms involving primate gut microbiota in the host to high-altitude environments38. Previous studies have also found significant differences in the composition of the gut microbiota between RMs from different habitats, with the highest diversity and the largest number of specific microbiota in high-altitude populations9,14. Related studies are mainly based on 16S rRNA gene sequencing and the characterization of gut microbiota composition in RMs at the genus or family level. Less is known about the functional and metabolite composition of gut microbiota, or the interactive mechanisms between gut microbiota and host adaptation to high-altitude environments. Therefore, we collected fresh fecal samples in March 2021 from four wild RM populations on the southeast edge of the Qinghai Tibetan Plateau, from Pamlin (HA population) and ** the host improve the metabolic capacity of food19,20,21,22,23,24. The pyruvate, propionate, butyrate, starch, sucrose, and pentose phosphate pathways are important for the fermentation of unabsorbed carbohydrates into SCFAs25. Meanwhile, butyrate is preferentially used as an energy source for the intestinal mucosa, propionate contributes to gluconeogenesis in the liver, and acetate has the highest concentration in blood61. Acetate production pathways are widely distributed in anaerobic bacteria15, while the production pathways of propionate, butyrate, and lactic acid are more conservative and substrate-specific62. Among the SCFAs produced by gut microbiota, butyrate is the most preferred source of energy in this respect, but its synthesis depends mainly on carbohydrates in the intestine22,63. In human gut microbiota, a variety of fermentation strategies have been developed to further generate energy. Pyruvate can be decomposed and metabolized into succinate, lactic acid, or acetyl-CoA. These intermediates can be further metabolized by hosts to produce acetate, propionate, and butyrate at the same time23,64. In this study, we found that high-altitude populations, acting through the m00422, m00631, and m00003 modules, can produce the energy substrates acetyl-CoA and pyruvate, mediate the energy metabolism of the host and provide energy compensation to the host. Meanwhile, the gut microbiota of RMs had a significantly higher potential to synthesize acetate at high altitudes, and a significantly higher potential to synthesize butyrate at low altitudes. However, the gene enrichment of the microbiota itself for these SCFA breakdown-related modules did not differ significantly between high- and low-altitude populations. Further, metabolite analysis did not detect these energy substrates, and the detected SCFAs did not differ significantly between the high- and low-altitude populations, indicating that these SCFAs and energy substrates are more likely to be absorbed and utilized by the host. These results are compatible with food acquisition and environmental stress in populations of wild RMs from different altitudes. Low-altitude populations are more likely to obtain foods with a high carbohydrate content and have more residual carbohydrates in their intestines than high-altitude populations; therefore, the potential to catalyze the formation of butyrate is higher as a result of the richer carbohydrate in the food of low-altitude populations. The gut microbiota at high altitudes shows high potential for acetate production. Meanwhile, pyruvate and acetyl-CoA are produced through the gut microbiota own metabolic pathways, which provide energy substrates for the host’s tricarboxylic acid cycle and fatty acid metabolic pathways. This also corresponds to the fact that the abundance of the microbiota associated with acetate production was significantly higher in the high-altitude populations. A recent study in other mammals also found that elevated acetate levels induced by altered microbiota drive hosts to seek food and energy accumulation by activating the parasympathetic nervous system to promote the secretion of insulin and gastric hunger hormones65. Another study in obese mice also found that oral butyrate reduced body weight largely due to increased energy expenditure and lipid oxidation66. This also appears to account for the higher potential for gut microbiota in high-altitude populations to directly provide energy compensation to the host through acetate synthesis and by stimulating food intake, further promoting energy accumulation. Moreover, the gut microbiota of high-altitude populations had a significantly higher abundance of genes associated with the metabolism of methanol than that of low-altitude populations. Methanol, as a by-product of microbiota metabolism, can serve as a substrate for microbiota methanogenesis and acetate64,67. Our results shows that the potential of microbiota in high-altitude populations to use methanol decomposition to produce energy is also higher than that in low-altitude populations, and their utilization of intestinal metabolites is more refined than that in low-altitude populations.

Wildlife gut microbiota is commonly exposed to environmental contaminants and is involved in the process of degrading environmental contaminants68. The gastrointestinal tract is the main route for xenobiotics to enter the body, and the gut microbiota has a high metabolic potential for xenobiotics. Gut microbiota can directly metabolize xenobiotic compounds of exogenous origin or affect the absorption, distribution, metabolism, and elimination of xenobiotics in the host, thus changing the toxicity of xenobiotics to the host68,69,70. PAHs are a broad class of organic pollutants present in the environment and toxic to humans and animals71. Soils act as a reservoir of PAHs mainly derived from atmospheric deposition and plant to microbiota synthesis72,73. Toluene is a common volatile organic compound (VOC) that is toxic to animals. Dead plant litter produces toluene, which is then transported to the soil74,75,76. In this study, we performed functional enrichment analysis of differential metabolites derived from microbiota production and found that among 4 statistically significant metabolic pathways, metabolites associated with cyclic aromatic hydrocarbon degradation and toluene degradation were significantly enriched in high-altitude populations compared with low-altitude populations. Based on biological and statistical analysis, we found that multiple responses of various bacterial communities (such as Streptomyces vietnamensis, Tsukamurella paulometabola, Nocardia brasiliensis, Candidatus Formimonas warabiya, and Rhizobium leguminosarum) in these two pathways play an important role,, in promoting the degradation of these exogenous toxins through R01631, R01635, and R05375 reactions in high-altitude populations. Meanwhile, our metagenomic screening based on KO genes in microbiota also revealed significantly higher tphB (K18076) gene richness in the polycyclic aromatic hydrocarbon degradation pathway (terephthalate degradation M00624) module that prompted the R01633 reaction to proceed. These results indicate that the degradation of PAHs, occurs at a significantly higher rate in high-altitude populations than in low-altitude populations. This means that high-altitude RMs can eat many plant roots and stems in cold seasons when there is a lack of food. As we observed during the long-term monitoring of the HB population60 and sample collection, high-altitude populations will dig and eat many plant roots when food resources are scarce in winter, and PAHs and toluene are widely enriched in soil and plant roots. The degradation of these exogenous toxicants by gut microbiota in high-altitude populations is important. In addition, our gut microbe-based functional enrichment analysis also revealed that high-altitude populations were significantly enriched in genes involved in the conversion of L-serine to L-cysteine in the cysteine biosynthesis (M00021) module, while related studies found that L-cysteine, an amino acid detoxification drug, is involved in cellular reduction processes and phospholipid metabolism in the liver, protects hepatocytes from damage, and promotes liver function recovery pharmacological effects77, implying an important role for their gut microbiota in the detoxification of the host’s own products via metabolites. These results suggest that as part of RMs adaptation to high-altitude environments, the gut microbiota contributes to hosts degradation of exogenous toxicants.

In conclusion, this study revealed the synergistic involvement of the gut microbiota in the adaptation of wild RMs to high-altitude environments (Fig. 7). Wild RMs adapted to high energy demands and high-quality food deprivation in a high-altitude environment had a diverse gut microbiota. And the abundance of multiple species of bacteria associated with acetate synthesis and those associated with fiber degradation was significantly higher than in low-altitude populations. We found a stronger potential for gut microbiota to synthesize the energy substrate acetyl-CoA using CO2 and the energy substrate pyruvate using oxaloacetate, as well as a stronger potential to transform acetyl-CoA to the energy substrate acetate in high-altitude populations. These energy substrates provide the host with enhanced energy compensation at high altitudes. Acetate production by gut microbiota promotes host insulin and gastric ghrelin secretion, driving food behaviors seeking and further energy accumulation. Meanwhile, in high-altitude populations, the gut microbiota can effectively help the host degrade PAHs consumed when feeding on complex foods such as plant rhizomes. In addition, we revealed biological and statistical links between multiple distinct gut microbiota and metabolites, providing fundamental data for understanding the ways gut microbiota affect the host.

Fig. 7: Study design and main results.
figure 7

cooF anaerobic carbon-monoxide dehydrogenase iron sulfur subunit, cdhD acetyl-CoA decarbonylase/synthase, CODH/ACS complex subunit delta, cdhE acetyl-CoA decarbonylase/synthase, eda 2-dehydro-3-deoxyphosphogluconate aldolase, kdgK 2-dehydro-3-deoxygluconokinase, E4.1.1.32 phosphoenolpyruvate carboxykinase, E4.1.1.49 phosphoenolpyruvate carboxykinase, ppdK pyruvate, orthophosphate dikinase, korA 2-oxoglutarate/2-oxoacid ferredoxin oxidoreductase subunit alpha, ACSS1_2 acetyl-CoA synthetase, EC:2.3.1.8 putative phosphotransacetylase. Red indicates that relative abundance was significantly higher in high-altitude populations. Green indicates that abundance was significantly higher in low-altitude populations. These macaque photos were taken by ourselves. These images and every element of these images were created by us.

Full size image

Methods

Fecal sample collection

Fecal samples from RMs were collected from the wild natural habitat in March 2021. The samples were collected from Pamulin (HA; 101.182061°E, 30.101192°N) and **’elou Town (HB; 100.715649°E, 29.936961°N) in Yajiang County, Ganzi Prefecture, Sichuan Province at an altitude of more than 3000 m and from Baidicheng (LA; 109.577491°E, 31.040904°N) and Simianshan (LB; 106.405604°E, 28.644834°N) of the Chongqing municipality at an altitude of less than 1000 m (Fig. 8a). Yajiang County is located on the southeast edge of the Qinghai Tibet Plateau, with high-altitude climate characteristics of low temperature, low oxygen, and high ultraviolet radiation. The vegetation in the RMs habitat is mainly alpine coniferous forest, alpine hardwood forest, and alpine meadow, so it is difficult for RMs to obtain food. Simianshan and Baidicheng in Chongqing are located on the southeast edge of the Sichuan Basin, a subtropical humid monsoon climate area with a mild climate and high oxygen content. The vegetation in this habitat is mainly broad-leaved evergreens and broad-leaved deciduous trees, and it is less difficult for RMs to obtain food. Comparison of the monthly average temperature and precipitation data for the 4 sampled sites based on ArcGIS 10.0 and WorldClim’s (version 2.1)78 published monthly climate data for 1970–2000 (resolution 30 s) revealed that the climate of the habitats of the two high-altitude populations (Yajiang Pamulin Temple and **’elou Town) was similar, with the lowest average temperature in January (−6.1 °C/−4.9 °C) and the highest average in July (−10.7 °C/−11.7 °C), and precipitation mainly falling in 6–9 months, with the highest monthly precipitation of 150 mm/152 mm showing distinct high-altitude climatic features. Although separated by a wide geographic distance, the habitats of the two low-altitude populations (Baidicheng and Simianshan, Chongqing) also have similar climates, with the lowest average temperature in January (5.5 °C/3.8 °C), the highest average temperature in July/August (27.6 °C/21.2 °C), the highest monthly precipitation of 212 mm/174 mm, and significantly higher temperature and humidity than the two high-altitude population habitats (Fig. 8b). According to the observations made during sampling, the total number of RM groups sampled at the four sites ranged from 40 to 50, and all sampled individuals were considered adult individuals based on follow-up observations and comparison of stool size. After defecation by the wild RMs, fecal samples were immediately collected with sterile gloves then stored in liquid nitrogen and transported to the laboratory. The samples for DNA preparation and metabolome detection for metagenomic sequencing were taken from the center of the fecal samples under sterile conditions, while samples for DNA preparation for individual identification and gender identity were taken from the surface. After collection, the samples were transported in liquid nitrogen and stored at -80 °C until analysis. There was no direct contact with RM during fecal collection, and our study had no impact on the health or welfare of RMs.

Fig. 8: Distribution of sampling sites and climate overview.
figure 8

a Distribution of sample collection sites. b Monthly average temperature and precipitation data for 1970–2000 at the sampling site obtained based on WorldClim (version 2.1). The boundary data comes from the data published by China National Basic Geographic Information Center. Triangles represent sample points.

Full size image

Samples collected from the four locations were analyzed for microsatellite molecular markers for individual identification and for SRY/ZFX genes for sex determination. The partial SRY gene (364 bp) on the Y chromosome and the partial ZFX gene (183 bp) on the X chromosome were used for sex determination. The primers were designed by Primer Premier 5 based on sequences in the GenBank database (accession numbers AF284311 and XM_028842412). The six microsatellite loci used for individual identification were amplified using the primers designed in our previous study for nested PCR. The primer pairs (Supplementary Table 4) under their respective annealing temperatures were used with 1.0 μl of template DNA or the product of the first round of PCR in a 20 μl PCR mixture (95 °C for 10 min initial denaturation; 35 cycles of denaturation at 94 °C for 30 s, 45 s annealing, and 30 s extension at 72 °C; followed by final extension for 10 min at 72 °C). All PCR products were examined by electrophoresis in 1.5% agarose gels containing ethidium bromide. Capillary electrophoresis analysis was used for PCR products of microsatellite loci. Sex determination was performed with five independent experiments per sample. Ten individual samples were screened per population by a male to female ratio, and a total of 40 fresh fecal samples were subjected to subsequent metagenomic and metabolomic testing (Fig. 7, Supplementary Table 5).

Metagenomic sequencing and quality control of raw data

DNA samples with a total content >25 ng, dispersion concentration above 500 bp, and no serious agglomeration below 500 bp were used for library preparation. The qPCR concentration of the final library was greater than 1.5 nm. There was no splice sequence, heteropeaks, or small fragments, and the library had a fragment size of 350 bp. After we completed Dnbseq-T7 platform (MGI Tech Co., Ltd. China) sequencing (Novogene Co., Ltd. China), adapters and low-quality reads in raw data were removed by Trimmomatic79, and potential RM sequences were removed by Bowtie280 based on the RM reference genome (assembly Mmul_10; GCA_003339765.3).

Assembly, taxonomic annotation, functional prediction, and quantification of genes

After quality control of the data obtained and the removal of host genes, mate pair assembly was performed using MEGAHIT (1.2.9)81, and the quality of the assembly results was assessed using QUAST (5.0.2)82. Gene prediction was done using Prodigal (2.6.3)83 with the parameter ‘-p meta’. The gene model of all genes predicted by CD-hit (4.8.1)84 was used to construct a non-redundant gene catalog parameter of ‘- as 0.9 - c 0.95 - G 0 - g 0 - T 0 - M 0’. SALMON (1.3.0)85 was used to quantify the non-redundant genes. Functional assignments of the protein sequences was done based on DIAMOND alignment against the eggNOG 5.0 database using eggNOG Mapper (2.0.0)86 by taking the best hit with the criterion of an E value <1e−3. The functional description of the portal gene responding to the searched sequence was the final annotation result to calculate the difference in gene composition; the next difference analysis was performed using normalized counts (TPM values). STAMP (2.1.3)87 was used to analyze the differences in gene composition between the low-altitude and high-altitude populations. After false discovery rate (FDR) corrections, P < 0.05 was used as the threshold for significance, revealing the differential genes of the gut microbiota between high- and low-altitude populations of RMs. The 8 GB MiniKraken database, pre-built using the abundance estimates of Kraken2 (2.1.1)88, was used to test the classification analysis based on non-redundant genes. LEfSe analysis was completed using LEfSe (1.1.2). Taxonomic composition, alpha diversity, beta diversity analysis, and visualization of all results were done using R (4.1.2). In order to investigate pathway modules known in microorganisms, we collected KEGG modules listed under “Microbial metabolism in diverse environments” (map01120)89.

Metabolite profiling from stool samples

Faecal samples (100 mg) were individually ground with liquid nitrogen, and the homogenate was resuspended with prechilled 80% methanol and 0.1% formic acid using vortexing. The samples were incubated on ice for 5 min then centrifuged at 15,000 g and 4 °C for 20 min. Some of the supernatants were diluted with UPLC-MS/MS grade water to a final concentration of 53% methanol. The samples were subsequently transferred to a fresh Eppendorf tube then centrifuged at 15,000 g and 4 °C for 20 min. Finally, the supernatant was injected into the UPLC-MS/MS system90.

UHPLC-MS analyses were performed using a Vanquish UHPLC system (Thermo Fisher, Germany) coupled with an Orbitrap Q ExactiveTM HF mass spectrometer (Thermo Fisher, Germany) at Novogene Co., Ltd. (Bei**g, China). Samples were injected onto a Hypersil gold column (100 ×2.1 mm, 1.9 μm) using a 17-min linear gradient at a flow rate of 0.2 mL/min. The eluents for the positive polarity mode were eluent A (0.1% FA in water) and eluent B (methanol). The eluents for the negative polarity mode were eluent A (5 mM ammonium acetate, pH 9.0) and eluent B (methanol). The solvent gradient was set as follows: 2% B, 1.5 min; 2%–100% B, 12.0 min; 100% B, 14.0 min; 100%–2% B, 14.1 min; and 2% B, 17 min. The Q Exactive TM HF mass spectrometer was operated in positive/negative polarity mode with a spray voltage of 3.2 kV, a capillary temperature of 320 °C, a sheath gas flow rate of 40 arb, and an aux gas flow rate of 10 arbs.

The raw data files generated by UHPLC-MS/MS were processed using Compound Discoverer 3.1 (CD3.1, Thermo Fisher) to perform peak alignment, peak picking, and quantitation for each metabolite. The main parameters were set as follows: retention time tolerance, 0.2 min; actual mass tolerance, 5 ppm; signal intensity tolerance, 30%; signal/noise ratio, 3; and minimum intensity. After that, peak intensities were normalized to the total spectral intensity. The normalized data was used to predict the molecular formula based on additive ions, molecular ion peaks, and fragment ions, then peaks were matched with the mzCloud (https://www.mzcloud.org/), mzVault, and MassList databases to obtain accurate qualitative and relative quantitative results. Statistical analyses were performed using the statistical software R (R version 3.4.3), Python (Python version 2.7.6), and CentOS (CentOS release 6.6). When the data were not normally distributed, normal transformations were attempted using the area normalization method.

Metabolite statistical analysis

Data on the fecal metabolome were processed using the Novogene platform. The metabolites were annotated using the KEGG (https://www.genome.jp/kegg/pathway.html) and HMDB (https://hmdb.ca/metabolites) databases. Principal components analysis (PCA) and PLS-DA were performed using metaX91 (a flexible and comprehensive software for processing metabolomics data). We applied univariate analysis (t-test) to calculate the statistical significance (P-value). The metabolites with VIP > 1, a P-value <0.05, and a fold change (FC) > 1.5 or <0.7 were considered differential metabolites. Volcano plots were used to filter metabolites of interest, and were based on the log2 (FC) and -log10(P-value) of metabolites by ‘ggplot2’ in R. For clustering heat maps, the data were normalized using z-scores of the intensity areas of differential metabolites and were plotted using the ‘pheatmap’ package in R. The correlation between differential metabolites was calculated using the function ‘cor’ in R (method = Pearson). The statistical significance of correlations between differential metabolites was calculated using ‘cor.mtest’ in R. A P-value greater than 0.05 was considered as statistically significant, and correlation plots were plotted using the ‘corrplot’ package in R. The functions of these metabolites and metabolic pathways were studied using the KEGG database. Metabolic pathway enrichment was analyzed: when the ratio was x/n > y/N, the metabolic pathway was considered as enriched, and when the P-value of the metabolic pathway was <0.05, the metabolic pathway was considered as significantly enriched.

Metabolite traceability and association analysis with microbiota

The bioinformatics analysis process on the MetOrigin platform92 was used to trace the source of metabolites by searching and integrating the databases of seven metabolites (KEGG, HMDB, BIGG, ChEBI, FoodDB, drug database, and toxin and toxin target database [T3DB]). Further, functional enrichment analysis was performed according to the different sources of metabolites92. The reference metabolic pathway of the host was from the KEGG database, and the reference metabolic pathway of the microbiota community was from more than 6,800 microbiotas in the integrated database of the MetOrigin platform. The reference metabolic pathway shared by the host and the microorganism was obtained by integrating these two pathways. Combined with the differential microbiota data, the statistical links between the differential fecal metabolites and differential microbiota were first revealed by Spearman analysis, then the KEGG database was utilized to search for microbiotas that might be involved in the associated metabolic responses and to perform associations with metabolites. Sankey network analysis was further utilized to visually demonstrate the biological and statistical links between gut microbiota and fecal metabolites92. Finally, network summary analysis was performed to further reveal the statistical and biological significance of correlations between microbiota and metabolites in specific functional pathways92. We mapped the KO genes and metabolites with significant differences in the feces of RMs at high and low altitudes together in the “microbial metabolism in diverse environments” pathway using the KEGG Mapper–Color tool (https://www.genome.jp/kegg/mapper/color.html) to reveal the residues of metabolites in feces produced by the main modules of the functioning of RMs gut microbiota.

Reporting summary

Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article.