Abstract

   Trillions of microbes colonize the ungulate gastrointestinal tract,
   playing a pivotal role in enhancing host nutrient utilization by
   breaking down cellulose and hemicellulose present in plants. Here,
   through large-scale metagenomic assembly, we established a catalog of
   131,416 metagenome-assembled genomes (MAGs) and 11,175 high-quality
   species-level genome bins (SGBs) from 17 species of ungulates in China.
   Our study revealed the convergent evolution of high relative abundances
   of carbohydrate-active enzymes (CAZymes) in the gut microbiomes of
   plateau-dwelling ungulates. Notably, two significant factors contribute
   to this phenotype: structural variations in their gut microbiome
   genomes, which contain more CAZymes, and the presence of novel gut
   microbiota species, particularly those in the genus Cryptobacteroides,
   which are undergoing independent rapid evolution and speciation and
   have higher gene densities of CAZymes. Furthermore, these enrichment
   CAZymes in the gut microbiomes are highly enrichment in known metabolic
   pathways for short-chain fatty acid (SCFA) production. Our findings not
   only provide a valuable genomic resource for understanding the gut
   microbiomes of ungulates but also offer fresh insights into the
   interaction between gut microbiomes and their hosts, as well as the
   co-adaptation of hosts and their gut microbiomes to their environments.

   Subject terms: Microbial genetics, Metagenomics, Molecular evolution

Introduction

   The gut microbiota plays a crucial role in host energy metabolism,
   immune homeostasis, and significantly influences host behavior, health,
   and diseases by regulating the gut-organ axis^[50]1–[51]8. Notably,
   ungulates such as horses, wild asses, camels, cattle, and sheep
   primarily subsist on vegetation with limited nutritional content,
   particularly grass^[52]9,[53]10. The dominant components, such as
   (hemi)cellulose in grass, pose challenges for digestion, compelling
   these ungulates to rely predominantly on metabolites like short-chain
   fatty acids (SCFAs) and volatile fatty acids (VFAs) synthesized through
   gut microbiome fermentation for their survival and reproductive
   processes^[54]9,[55]11. According to previous studies, SCFAs/VFAs serve
   up to 70% of the remints’ energy needs^[56]12. The Qinghai-Tibet
   Plateau (QTP), characterized by low temperature, hypoxia, and intense
   ultraviolet radiation, exhibits sparse vegetation compared to other
   plain areas in China^[57]13,[58]14. The harsh climate and limited food
   resources pose significant challenges to the survival of local
   ungulates. In this context, the gut microbiota plays an indispensable
   role for plateau ungulates, enabling them to maintain highly efficient
   energy metabolism essential for survival in the high-altitude harsh
   environment^[59]15.

   The carbohydrate-active enzymes (CAZymes) present in rumen gut
   microbiomes play a crucial role in various metabolic pathways involved
   in carbohydrate degradation^[60]12. They employ several mechanisms,
   including the biosynthesis and hydrolysis of glycosidic bonds,
   beta-elimination, and the removal of ester bonds. CAZymes convert
   cellulose, hemicellulose, and lignin into monosaccharides, such as
   glucose, galactose, and mannose, significantly contributing to the
   production of SCFAs/VFAs^[61]16,[62]17. Both monosaccharides and
   SCFAs/VFAs serve as essential energy sources for the survival and
   reproduction of rumen microorganisms, while also supporting their
   health and environmental adaptation^[63]15–[64]18. Carbohydrate-active
   enzymes (CAZymes) are mainly divided into six major groups: glycoside
   hydrolases, glycosyltransferases, polysaccharide lyases, carbohydrate
   esterases, auxiliary activities, and carbohydrate-binding
   modules^[65]12. Rumen gut microbiomes convert cellulose, hemicellulose,
   and lignin into monosaccharides through the cooperation of CAZymes,
   which are involved in most metabolic pathways for SCFA and VFA
   synthesis^[66]9,[67]11,[68]12. Therefore, the efficiency of SCFA and
   VFA production relies on the composition and abundance of CAZymes and
   gut microbiomes at the molecular and cellular levels, respectively.
   Notably, previous researches have demonstrated that Tibetan ungulates
   exhibit higher efficiency in SCFA and VFA production compared to their
   lowland relatives, which may be attributed to their unique gut
   microbiome community compositions^[69]15,[70]19–[71]23. Additionally,
   high CAZyme abundances have also been found in other Tibetan plateau
   mammals such as Tibetan pigs and horses^[72]15,[73]20,[74]23–[75]25.
   However, little is known about the gut microbiome compositions, genomic
   characterizations, and evolutionary patterns underlying this phenotype
   in these ungulates. In previous studies, frequent genetic variations
   such as structural variations (SVs) and copy number variations (CNVs)
   in the genomes of gut microbiomes, mainly caused by widespread
   horizontal gene transfer (HGT) events and high mutation rates, are
   highly associated with the gain and loss of CAZymes and the overall
   rate of SCFA production in gut microbiomes^[76]26–[77]30. On the other
   hand, hosts and their gut microbiomes form closely related systems that
   interact tightly in response to changing environments^[78]31,[79]32.
   Recent studies on apes and hominids and their gut microbiomes suggest a
   co-evolution and co-speciation relationship between gut microbiomes and
   their hosts. Based on these findings, hosts and their gut microbiomes
   constitute a stable system adapted to environments, regardless of their
   internal interactions^[80]33–[81]35. However, some studies also
   observed widespread host selection, host swapping, and extinction in
   the evolution and speciation of gut microbiomes with their hosts,
   suggesting that co-evolution and co-speciation are not always
   dominant^[82]36–[83]40. In other words, special evolution and
   speciation patterns of gut microbiomes may occur in the dynamics of
   their hosts and external environments. Consequently, we propose that
   the high SCFA/VFA production efficiency phenotypes observed in the gut
   microbiomes of Tibetan ungulates may be related to the convergent
   evolution of higher abundances of CAZymes in some of their core
   SCFA/VFA-producing gut microbiome species compared to closely related
   species in lowland hosts. Specifically, this may be associated with
   three potential factors: first, the higher abundances of some core gut
   microbiomes with higher CAZyme activities in plateau hosts; second, the
   adaptive genomic structural variations with higher CAZyme coding genes
   in some core gut microbiomes in plateau hosts; and third, the
   establishment of unique gut microbiomes in plateau hosts through other
   evolutionary and speciation patterns which have higher CAZyme coding
   genes.

   Unlike traditional metagenomic strategies that focus on gene sets
   within assembled contigs/scaffolds, the metagenomic binning strategy
   groups these structures into clusters representing the whole genome of
   an organism, defined as metagenome-assembled genome
   (MAG)^[84]41–[85]43. This approach facilitates evolutionary and
   functional analysis of microbiota (including numerous novel
   microbiomes) at the species or strain level based on comparative
   genomes. In our study, we sequenced and assembled MAGs from the gut
   microbiotas of 1167 individuals across 17 plateau and lowland ungulate
   species. Using these MAGs, we conducted comparative genomic and
   phylogenetic analyses to explore factors associated with the high
   SCFA/VFA production efficiency phenotypes in plateau ungulates. We
   further elucidated the relationship among functional genes, genomic
   structural variations, and evolution patterns of gut microbiomes in
   response to their energy metabolic phenotypes, shedding more light on
   the co-adaptation mechanisms of plateau ungulates and their gut
   microbiomes to harsh plateau environments.

Results

Distribution and survival status of 17 species of common ungulates

   Among the 17 common ungulates, two species in one family belong to
   perissodactyl, and 15 species in two families (four subfamilies) belong
   to artiodactyl (Supplementary Data [86]1). Noteworthy species include
   the kiang (Equus kiang), white-lipped deer (Cervus albirostris),
   Chinese red deer (Cervus elaphus), domestic yak (Bos grunniens), wild
   yak (Bos mutus), goitered gazelle (Gazella subgutturosa), Tibetan
   gazelle (Procapra picticaudata), Przewalski’s gazelle (Procapra
   przewalskii), Tibetan antelope (Pantholops hodgsonii), bharal (Pseudois
   nayaur), and argali (Ovis ammon), which inhabit the Qinghai-Tibet
   Plateau (QTP) at altitudes exceeding 3000 m. Conversely, the horse
   (Equus caballus), sika deer (Cervus nippon), eastern roe deer
   (Capreolus pygargus), cattle (Bos taurus), goat (Capra hircus), and
   sheep (Ovis aries) dwell in vast plains in China at altitudes below
   1000 m (Fig. [87]1). Specifically, the kiang, white-lipped deer, wild
   yak, goitered gazelle, Tibetan gazelle, Przewalski’s gazelle, and
   Tibetan antelope are endemic to the Tibetan plateau; the sika deer and
   eastern roe deer are common wild ungulates, while the horse, domestic
   yak, cattle, goat, and sheep are common livestock. According to the
   IUCN Red List of threatened species
   ([88]https://www.iucnredlist.org/en), Przewalski’s gazelle is
   endangered (EN); wild yak and goitered gazelle are vulnerable (VU);
   Tibetan gazelle, Tibetan antelope, and argali are near threatened (NT);
   sika deer, eastern roe deer, and bharal are least concern (LC)
   (Supplementary Data [89]1).

Fig. 1. The distribution of 17 common ungulates in our study.

   [90]Fig. 1
   [91]Open in a new tab

   The left tree was constructed based on the geographic distance of fecal
   samples from 17 ungulates’ species. the Qinghai-Tibet Plateau (QTP) at
   altitudes exceeding 3000 m, while brown branches represent ungulates
   inhabiting the vast plains in China at altitudes below 1000 m. The
   right map is color-coded by altitude, and the silhouettes of ungulates
   depict the locations of their fecal samples.

Metagenome-assembled genomes of ungulates’ gut microbiota

   Following Illumina NovaSeq 6000 sequencing, we generated 9.15 Tb of
   high-quality metagenomic sequencing data from fecal samples of 10
   ungulate species. Additionally, 9.19 Tb of metagenomic sequencing data
   from the gut microbiota of seven ungulate species were downloaded from
   NCBI (Supplementary Data [92]2). In total, we accumulated 18.34 Tb of
   sequencing data encompassing 17 ungulate species (1167 individuals;
   15.72 ± 2.35 Gb per sample). Post quality control, we removed 101.17 Gb
   of reads originating from the hosts’ genome, retaining 18.01 Tb of
   host-free reads (99.45%; averaging 15.43 Gb per sample) for further
   analysis (Supplementary Data [93]3).

   All host-free reads underwent assembly and classification, resulting in
   the generation of 131,416 metagenome-assembled genomes (MAGs). This set
   included 88,343 low-quality MAGs (completeness < 50% or
   contamination ≥ 5%), 20,075 medium-quality MAGs (completeness ≥ 50% and
   contamination < 5%), and 22,998 high-quality MAGs (completeness ≥ 80%
   and contamination < 5%) (Fig. [94]2a and Supplementary Data [95]4).
   Among the high-quality MAGs, the average completeness and contamination
   were 90.51% and 1.44%, respectively (Supplementary Data [96]5).
   Notably, 83.71% of high-quality MAGs exhibited an assembly N50
   length > 10 kb (Supplementary Data [97]5). Subsequently, we assessed
   the correlation (Spearman’s r) between relative abundances and
   completeness of MAGs. The results revealed a positive relationship
   between relative abundances and completeness of MAGs (r = 0.23,
   p < 0.01 for high-quality MAGs, r = 0.25, p < 0.01 for all MAGs) (Fig.
   [98]2c and Supplementary Fig. [99]1). Remarkably, through high sequence
   depth (15.72 ± 2.35 Gb per sample), many high-quality MAGs (n = 14,867)
   with a low relative abundance (log[10] ^relative abundance < 1) were
   completely assembled (Fig. [100]2c and Supplementary Data [101]4,
   [102]5). All highlight the overall high assembly quality of these MAGs.

Fig. 2. Metagenome-assembled genomes in the gut microbiota of the 17-ungulate
species.

   [103]Fig. 2
   [104]Open in a new tab

   a The number of high-quality, medium-quality, and low-quality MAGs. b
   The number of species-level genome bins (SGBs) in different hosts. c
   Correlation between the relative abundances and completeness of MAGs.
   The correlation (Spearman’s r) was measured for the high-quality MAGs.
   d The phylogenetic tree of 8489 nonredundant SGBs, including 8387
   bacteria and 102 archaea, with taxonomies assigned by GTDB-Tk and
   colored with circles at the phylum level. The legend is arranged in
   decreasing order (top to bottom) of the number of bacteria and archaea
   detected in the corresponding phyla, respectively.

   For redundancy removal, high-quality MAGs in each ungulate gut
   microbiota were dereplicated separately at the strain and species
   levels by dRep^[105]44,[106]45. This process in total yielded 16,638
   strain-level MAGs (ANI ≤ 99%) and 11,175 species-level genome bins
   (SGBs; ANI ≤ 95%). Based on GTDB-Tk assignments, 4084 SGBs had
   references in the Genome Taxonomy Database (GTDB) and were defined as