Characteristics comparative of gut microbiota in four small mammal species co-occurring in the Hengduan Mountains: differences in environmental adaptation strategies between alien and endemic species

Abstract

The composition of gut microbiota is jointly determined by the host’s food habit, ecological niche, and genetic background, serving as a direct reflection of the host’s adaptation to its environment and evolutionary pressures. To investigate the distinct adaptation mechanisms of four sympatric small mammal species in the Hengduan Mountains region, this study compared the environmental adaptation strategies of endemic and alien species. This study collected ten wild specimens each of the Eothenomys miletus, Eothenomys oliter, Tupaia belangeri, and Apodemus chevrieri from Yunlong County, Yunnan Province. Using 16SrRNA gene sequencing technology, we analyzed the gut microbial composition, abundance, and community structure across species, investigating the distinct gut microbial community characteristics between the endemic species (E. miletus and E. oliter) and the alien species (T. belangeri and A. chevrieri.). Results indicate: E. miletus and E. oliter possess complex and diverse gut microbial communities with plant-degrading functions. These communities with sparse interactions are capable of utilizing multiple plant sources for nutrition and exhibit strong resilience against environmental disturbances. In contrast, T. belangeri and A.chevrieri exhibit simple, specialized, yet tightly cooperative omnivorous gut microbial communities. While capable of utilizing diverse food resources within specific adaptive environments, they demonstrate extreme specificity in adaptation to particular habitats or survival strategies and are relatively sensitive to external disturbances. Furthermore, the E. miletus, widely distributed across the Hengduan Mountains, its gut bacterial community is dominated by stochastic processes. In contrast, the T. belangeri shows a positive correlation with various gut bacteria associated with omnivorous characteristics. The distinction between these two distinct environmental adaptation strategies is particularly pronounced. In summary, Among these four sympatric small mammal species in the Hengduan Mountains, the gut microbiota of endemic and alien species showed high similarity respectively and exhibited convergence.

Introduction

Gut microbiota are intrinsically linked to the host’s vital functions, co-evolving alongside the host to form an interactive symbiotic functional unit [1]– [2]. By participating in nutrient metabolism, immune regulation, and barrier formation, the microbial communities within the host gut exert decisive influence on physiological health and environmental adaptation [3]– [4]. Meanwhile, The composition of host gut microbes is not random but directly reflects their environmental conditions and evolutionary pressures [5]. The host’s diet, ecological niche, and phylogenetic history collectively shape its unique gut microbial structure, making it a biological manifestation of the host’s adaptation to specific survival strategies [6]– [7]. Even species sharing identical geographic conditions exhibit markedly distinct gut microbial structures due to differing feeding habits [8]. For instance, among the sympatric argali (Ovis ammon hodgsoni) and blue sheep (Pseudois nayaur) on the Qinghai-Tibet Plateau, argali intestines harbor greater diversity of Firmicutes and cellulose-degrading bacterial genera, while blue sheep intestines exhibit higher Bacteroidetes abundance and stronger carbohydrate synthesis and metabolic functions [7]. Studies on four sympatric lemur species inhabiting the eastern rainforests of Madagascar have demonstrated that the frugo-folivory (sifakas), young-leaf folivory (indri and woolly lemurs), and mature-leaf folivory (sportive lemurs) harbor species-specific microbiomes, metagenomes, and metabolomes that are precisely tailored to their dietary specializations [9]. Furthermore, under the strong selective pressure of dietary convergence, the gut microbial communities of different species may also exhibit a convergent pattern. Studies on the gut microbiomes of wild wolves, captive wolves, and domestic dogs have demonstrated that captive environments drive convergence in the compositions of carbohydrate-active enzymes (CAZymes) and virulence factors (VFs) between captive wolves and domestic dogs, with both groups co-enriched in CAZymes involved in plant carbohydrate degradation (e.g., GH13, GH3) as well as VFs associated with adhesion, invasion, and antibiotic resistance [10]. Therefore, analyzing the gut microbial composition of co-occurring species allows for inferences about resource utilization, environmental adaptation, and evolutionary status among species.

Endemic species refer to biological taxa that survive only within a limited geographic range and do not occur naturally elsewhere. Endemicity classified by spatial scope or taxonomic level, is a property that can change over time [11]. Typical examples include Bradypus variegatus, Tamandua tetradactyla, and Pusa caspica [12]– [13]. Their existence often reflects relatively recent speciation or long-term adaptive stability between species and their local environments [14]. In this study, the E. miletus and the E. oliter are endemic spcies on the Hengduan Mountains [15]– [16]. Molecular phylogenetic studies indicate that the Eothenomys likely originated in the Hengduan Mountains and surrounding regions during the late Pliocene to early Pleistocene, evolving from Arvicolinae that more ancient and widely distributed [17, 18]. The earliest fossils of this genus date back to the Tertiary period, showing little morphological difference from extant species [17]. Over time, orogenic movements in the Himalayan-Hengduan Mountains created multiple vertical ecological zones within the region [19]. Variations in latitude and elevation contributed to the speciation of the Eothenomys [20]. Now, the E. oliter population is restricted to higher-elevation mountainous areas like Zhaotong in northeastern Yunnan, while the E. miletus is widely distributed across the Yungui Plateau and surrounding mountain ranges [21]– [22]. In this study, alien species refer to taxa whose natural distribution range does not cover the Hengduan Mountains, the study area, and which have entered this region via natural dispersal or anthropogenic mediation and successfully established self-sustaining wild populations [23]. The T. belangeri and the A. chevrieri represent typical alien species in the Hengduan Mountains [24]– [25]. The T. belangeri represents the northernmost distribution of modern Scandentia, with its northernmost range potentially extending to the Yunnan Plateau and Hengduan Mountains in southwest China [25]. Currently, two competing hypotheses exist regarding the origin of extant Scandentia: the “island origin hypothesis” and the “mainland origin hypothesis” [26]. The island origin hypothesis holds that species first originated on islands and then migrated to and established themselves in other regions via transoceanic dispersal and other means. In contrast, the continental origin hypothesis proposes that species originated on ancient continents and later spread to islands and persisted there to the present day due to geological changes or climatic drivers [27]. Some evidence such as adaptive thermogenesis of the liver in T. belangeri during cold acclimation supports the “island origin hypothesis,” [28] suggesting Scandentia originated on Borneo island [29]. Fossil evidence discovered in various regions worldwide, such as India, Pakistan [30], and Lufeng, Yunnan, China [31], leans more toward the “mainland origin hypothesis.” However, due to limited studies of museum samples their evolutionary origin cannot be ascertained. Nevertheless, existing research consistently indicates that the Scandentia migrated northward from tropical regions. The Apodemus is a typical Palearctic species, exhibiting only marginal distribution in the Oriental realm [25]. It most likely originated within Europe or along the Asian periphery near Europe [32, 33]. The divergence between A. chevrieri and Apodemus agrarius occurred approximately 0.4 million years ago. It is likely that A. agrarius colonized the Hengduan Mountains during its radiation dispersal and subsequently expanded its population size continuously [34, 35].

The Hengduan Mountains lie at the junction of the Palearctic and Oriental realm, ranking among the world’s biodiversity hotspots [36]. Characterized by significant elevation gradients and pronounced seasonal variations in climate and vegetation, this region has developed complex ecosystems and serves as a center of habitat differentiation. It represents the most biodiverse and endemically rich area in China and across the entire Eurasian continent [37]. The region boasts a rich diversity of mammals, with a high proportion of ancient and endemic species [38]. Small mammals alone account for over 70% of China’s total mammal species, constituting a vital component of biodiversity and resources in the Himalayan-Hengduan region [39]. To elucidate the differences in gut microbiota characteristics among four sympatric small mammal species in the Hengduan Mountains, particularly whether convergent adaptation occurs between endemic species (the E. miletus and the E. oliter) and alien species (the T. belangeri and the A. chevrieri) in environmental adaptation? It holds significant importance for mammal diversity conservation in China. We predict that the gut microbiota compositions of the three rodent species will be more similar to each other, while the T. belangeri will show greater differences from the other three species.

Materials and methods

Experimental samples

In the same area of Yunlong County, Yunnan Province (98°89′56″E; 26°12′42″N), 10 adult healthy specimens of the E. miletus (n = 10; ♂:♀=6:4;Mean BW:40.60 g), E. oliter (n = 10; ♂:♀=6:4; Mean BW:35.45 g), T. belangeri (n = 10; ♂:♀=7:3; Mean BW:105.44 g) and A. chevrieri (n = 10; ♂:♀=5:5; Mean BW:42.10 g) were collected during the non-breeding season using mouse traps. All collected specimens underwent morphological identification in the field based on external characteristics, following the methods described by Yang Qisen et al. and Xia Lin et al. [40, 41].

Experimental methods

After disinfection and flea elimination, experimental animals were euthanized in a chamber filled with compressed carbon dioxide gas at a volume displacement rate of 70% [42]. Take 0.1 g of rectal feces and extract total DNA enriched on the filter membrane using the DNeasy^® PowerSoil^® Kit (Germany). The concentration of purified PCR products was measured using a Nanodrop 2000 spectrophotometer. Samples with nucleic acid concentrations above 10 ng/µL and purity ratios (A260/A180) greater than 1.8 were considered valid. DNA samples meeting the concentration criteria were mixed in equimolar ratios and sequenced using the Illumina MiSeq platform (Illumina, San Diego, CA, USA). Sequencing was performed using the Illumina MiSeq platform to generate 2 × 250 bp paired-end reads, which were processed and analyzed using the QIIME 1.8 platform. First, Flash 1.2.111 software was employed to assemble the paired-end reads, followed by assignment of a unique barcode label to each sample. Low-quality sequences (length < 300 bp or base quality < 30) were removed during assembly. Usearch 7.0 software was employed to eliminate chimeras from the sequences. Subsequently, the Uclust algorithm clustered all sequences into operational taxonomic units (OTUs) at 97% similarity. The longest sequence in each OTU was selected as the representative sequence, and taxonomic annotation was performed using the Ribosomal Database Project (RDP) database (Version 2.14). Finally, to enable comparative analysis across all samples, we standardized sequences using the “Daisychopper” script, setting a standard of 7,597 sequences per sample.

Data analysis

Use Origin 2018 to create a percentage stacked bar chart to describe the bacterial community composition of different species. The Alpha diversity was estimated using four diversity indices—Chao 1, Observed OTUs, Shannon, and Simpson to assess gut bacterial community diversity, with box plots created in Origin 2018 for visualization. Beta diversity was characterized using QIIME 1.8 and Origin 2018 to describe community structure. Principal coordinate analysis (PCoA) was then employed to visualize beta diversity across all samples. Intergroup differences were calculated via PERMANOVA based on both unweighted and weighted UniFrac distance matrices. Unweighted UniFrac distances depend on phylogenetic relationships and OTU species abundance, while weighted UniFrac distances account for species presence/absence and phylogenetic relationships. Microbial phyla and genera exhibiting significant intergroup distribution differences were identified and visualized using heatmaps generated with the R packages “vegan,” “permute,” and “gplots.” Single-factor analysis of variance was employed to compare differences in microbial abundance distribution among groups, with species serving as the variable. Visualization was achieved by generating heatmaps using the R packages “vegan,” “pheatmap,” and “igraph.” Using R 3.6.2 and Gephi v.0.9.2 software, we further analyzed OUTs detected in samples with relative abundance in the top 0.01% and above 20%, generating a collinearity network analysis (P < 0.05, |R|>0.6). Using R 3.6.2 software, we calculated observed beta diversity (βM), constructed a zero model to estimate expected beta diversity (βR), and then computed the Minimum Spanning Tree (MST) index to distinguish the contributions of deterministic processes (MST < 0.5) and random processes (MST > 0.5) to community assembly. Pearson correlation analysis and Mantel tests were employed to examine the associations between differential bacteria and gender and body weight.

Results

This study collected 40 intestinal samples—10 each from the E. miletus, E. oliter, T. belangeri and A. chevrieri, totaling 40 samples. DNA extraction and PCR amplification were performed, followed by 16 S rRNA sequencing. After removing low-quality sequences, chimeras, monomers, and chloroplasts, each sample was standardized to 7,597 sequences.

Composition of gut bacterial communities in different species

At the phylum level, the gut bacterial community of the E. miletus was dominated by Bacteroidetes, with a relative abundance of 38.9%. followed by Firmicutes (29.8%) and Proteobacteria (26.8%); that of the E. oliter was dominated by Firmicutes, with a relative abundance of 40.3%, followed by Bacteroidetes (24.0%) and Proteobacteria (23.2%); while both the A. chevrieri and T. belangeri exhibited gut bacterial communities dominated by Proteobacteria, which accounted for 77.7% and 85.8%, followed by the relative abundances of Bacteroidetes and Firmicutes in. A. chevrieri were 11.2% and 7.9%, while those in T. belangeri were 5.2% and 1.7%. At the phylum level, the gut bacterial communities of the E. miletus and E. oliter were relatively similar, while those of the T. belangeri and A. chevrieri were relatively similar.

At the genus level, the dominant genera in the gut bacterial communities of the E. miletus and E. oliter were S24-7 (UG), with relative abundances of 33.2% and 22.0%. Next in abundance were Escherichia, accounting for 23.6% and 14.2%, and Clostridiales (UG), accounting for 8.1% and 11.4% in the two rodent species, respectively. In contrast, both the A. chevrieri and T. belangeri had Escherichia as the dominant genus, with high relative abundances of 74.2% and 76.2%, while the abundances of other genera were relatively low. Similarly, at the genus level, the intestinal bacterial communities of the E. miletus and E. oliter were more similar, while those of the T. belangeri and A. chevrieri were more similar (Fig. 1).

Fig. 1.

Gut bacterial phylum (a) and genus (b) level community composition.Plot percentage stacked bar charts of gut bacterial community composition of different species using Origin 2018

,

Analysis of alpha diversity in gut bacterial communities across species

The alpha diversity of gut bacterial communities across different species is shown in Fig. 2. All four indices (Chao1, Observed OTUs, Shannon diversity and Simpson diversity) indicate that the gut bacterial communities of the E. miletus and E. oliter exhibit higher alpha diversity, showing significant differences (P < 0.05) compared to the gut bacterial communities of the T. belangeri and A. chevrieri.

Fig. 2.

Alpha diversity of gut bacterial communities (a: Chao1; b: Observed OTUs; c: Shannon; d: Simpson) across different species. (Different lowercase letters indicated significant differences between groups (p < 0.05), while the same letter denotes no significant difference.) Alpha diversity of gut bacterial communities was estimated using four indices and visualized with box plots in Origin 2018

Analysis of beta diversity in gut bacterial communities across species

The PCoA plot reveals that both weighted and unweighted Unifrac matrices show a clustering trend in gut bacterial beta diversity among different species. Further PERMANOVA analysis revealed significant differences in gut bacterial community structure between the T. belangeri and the other three species (P < 0.05; Table 1). No significant differences in gut bacterial beta diversity were observed among the other species (P > 0.05) (Fig. 3).

Table 1.

PERMANOVA test for differences in gut bacterial beta diversity among species. Intergroup differences were calculated via PERMANOVA based on unweighted and weighted UniFrac distance matrices, where the former reflects species presence/absence and phylogenetic relationships, and the latter phylogenetic relationships and OTU abundance. * Indicating significant differences

PERMANOVA	Weighted Unifrac			Unweighted Unifrac
	F	R2	P	F	R2	P
EM vs. EO	0.373	0.020	0.680	0.219	0.012	0.907
EM vs. AC	0.303	0.017	0.711	0.619	0.033	0.516
EM vs. TB	5.459	0.233	0.026*	4.011	0.182	0.033*
EO vs. AC	0.895	0.047	0.373	0.836	0.044	0.439
EO vs. TB	4.666	0.206	0.029*	5.253	0.226	0.012*
AC vs. TB	8.222	0.314	0.011*	7.403	0.291	0.012*

Fig. 3.

Beta diversity of gut bacteria across species. Beta diversity was characterized by QIIME 1.8 and Origin 2018. PCoA was used to visualize beta diversity across all samples

Differential analysis of intestinal microbiota across species

The heatmap displays gut bacteria exhibiting significant species-specific differences at the phylum and genus levels. At the phylum level, Proteobacteria predominantly enriched in the intestines of the T. belangeri and A. chevrieri, while Bacteroidetes, Firmicutes, and Actinobacteria primarily enriched in the intestines of the E. miletus and E. oliter (P < 0.05). At the genus level, the dominant genus Escherichia was enriched primarily in the intestines of the T. belangeri and A. chevrieri, while genera such as Desulfovibrio, Ruminococcus, and Christensenella were enriched mainly in the intestines of the E. oliter. Genera including Oscillospira, Coprococcus, Bacillus were enriched in the intestines of the E. miletus and E. oliter (P < 0.05). Overall, at the phylum and genus levels, most bacterial genera were significantly enriched in the intestines of the E. miletus and E. oliter, while a smaller number were significantly enriched in the intestines of the T. belangeri and A. chevrieri (Fig. 4).

Fig. 4.

Differential analysis of gut bacteria across species at the phylum (a) and genus (b) levels. Significantly differentially distributed microbial phyla and genera among groups were identified and visualized via heatmaps generated with R packages vegan, permute and gplots; one-way ANOVA was used to compare intergroup differences in microbial abundance distribution with species as the variable

Co-occurrence networks of gut bacterial communities across species

The correlation network of dominant OTUs in the intestines of different species is shown in Fig. 5. This network analysis included OTUs with relative abundances in the top 0.01% and detected in at least 20% of samples. After screening for correlation coefficients |R| > 0.6 and P < 0.05 using R, the correlation network of dominant OTUs in the intestines of different species was constructed using Gephi 0.9.2. The co-occurrence network of dominant OTUs in the gut microbiota of the E. miletus comprised 1118 nodes and 716 edges, including 689 positive edges and 27 negative edges. The co-occurrence network of dominant OTUs in the gut microbiota of the E. oliter comprised 1078 nodes and 579 edges, including 564 positive edges and 15 negative edges. The co-occurrence network of dominant OTUs in the gut of the A. chevrieri comprised 842 nodes and 4,337 edges, with 4,304 positive edges and 33 negative edges. The co-occurrence network of dominant OTUs in the gut of the T. belangeri comprised 526 nodes and 11,099 edges, with 11,044 positive edges and 55 negative edges. Bacterial interactions in the co-occurrence networks of dominant OTUs across different species were predominantly cooperative, primarily involving interactions among Firmicutes, Bacteroidetes, and Proteobacteria.

Fig. 5.

Co-occurrence networks of dominant OTUs across species guts. Heatmaps were generated for visualization with R packages vegan, pheatmap and igraph. Using R 3.6.2 and Gephi v0.9.2, we analyzed OTUs with relative abundance in the top 0.01% (≥20%) across samples and constructed a co-occurrence network analysis (P<0.05, |R|>0.6)

Construction mechanism of gut microbiota in different species

The construction mechanisms of gut bacterial communities in different species are illustrated in Fig. 6. Among them, the gut bacterial community of the E. miletus is dominated by random processes, while those of the E. oliter, A. chevrieri and T. belangeri are dominated by deterministic processes.

Fig. 6.

Mechanisms of gut microbiota assembly in different species. Using R 3.6.2, we calculated observed β-diversity (βM), constructed a null model to estimate expected β-diversity (βR), and computed the Minimum Spanning Tree (MST) index to distinguish deterministic (MST < 0.5) and stochastic (MST > 0.5) processes in community assembly

Correlation between differential phylum and gender or weight

The correlations between physiological indicators of different species and their differential phyla and genera are shown in Fig. 7. At the phylum level, gender showed no significant correlation with differential bacteria, while body weight exhibited significant positive correlations with most differential bacteria (P < 0.05). At the genus level, gender showed significant positive correlations with Pilimelia and Luteimonas, while body weight exhibited significant positive correlations with Microlunatus, Rubrobacter, Bradyrhizobium, Rhodoplanes, and Escherichia.

Fig. 7.

Correlation between bacterial differences at the genus level (a) and species level (b) and gender and body weight. Pearson correlation analysis and Mantel tests were used to explore the associations between differential bacteria and host gender and body weight

Discussion

Differences in gut microbiota community composition among four small mammal species

Based on the results of gut microbiota composition and differential analysis, we can conduct a comprehensive analysis of the physiological functions, ecological adaptability and evolutionary relationships of the four mammal species. First, the analysis of the gut microbiota composition and differential analysis reveals that the gut microbiota of E. miletus and E. olitor exhibit high similarity, with both dominated by the phyla Bacteroidetes and Firmicutes. At the genus level, they are co-enriched with gut microbiota such as Oscillospira, Coprococcus, and Ruminococcus, which are closely associated with functions like dietary fiber fermentation and short-chain fatty acid synthesis [43]. This strongly suggests that these two species have evolved digestive strategies highly adapted to a diet primarily consisting of cellulose-rich plant materials such as stems and leaves. This efficient fiber-degrading capability is crucial for them to obtain energy from tough plant tissues [43, 44]. It also reflects the long-term stable adaptation of these two endemic species to the geographical characteristics of the Hengduan Mountains region, where plant-based food resources are abundant and diverse [45]. In addition, the specific enrichment of Desulfovibrio and Christensenella in the gut of E. olitor represents its fine-tuned adaptation to specific habitat conditions (e.g., soil components and altitude) or energy metabolism requirements [46, 47]. This aligns with the ecological trait of E. olitor, which mainly inhabits the Hengduan Mountains at an altitude of 800–2600 m and has precise adaptations to cold climates [21]. In sharp contrast, the gut microbiota structures of A. chevrieri and T. belangeri are both dominated by the phylum Proteobacteria, particularly the genus Escherichia. The enrichment of Proteobacteria is often linked to gut environment instability and omnivorous diets [48]. The dominance of Escherichia—a genus of opportunistic pathogens—not only indicates that these two species may have more complex feeding habits and fluctuating gut environments but also implies that their immune systems face diverse challenges [49, 50]. Thus, it can be concluded that these two species possess more unstable gut microbial systems and likely adopt an omnivorous survival strategy of expanding food sources continuously to avoid core competition. Second, from the perspective of ecological niche and evolution, the gut microbiota of E. miletu and E. olitor, as well as A. chevrieri and T. belangeri, show distinct community similarity at both the phylum and genus levels. This not only confirms the inherent genetic and phylogenetic connections between E. miletus and E. olitor but also reflects the strong selective pressure of dietary convergence acting on the gut microbiota of endemic and alien species respectively.

In summary, the results of composition and differential analysis of the gut microbiota of the four species clearly depict a gut microbiota adaptation profile dominated by dietary niches. The gut microbiota of the endemic species (E. miletus and E. olitor) and the alien species (T. belangeri and A. chevrieri) have each evolved convergence respectively due to long-term occupation of similar ecological niches. E. miletus and E. olitor are representative of endemic environmental adaptations specialized in plant fiber degradation, while A. chevrieri and T. belangeri represent non-native opportunistic omnivores more adept at adjusting their ecological niches.

Differences in gut microbiota community characteristics among four small mammal species

Through the analysis of community characteristics including diversity, co-occurrence networks, and assembly mechanisms of gut bacteria in the four species, we can comprehensively infer the differences in structural and functional stability of their gut microbiota and the potential ecological significance. First, the results of alpha diversity show that the gut microbiota of E. miletus and E. olitor have significantly higher alpha diversity indices. A higher alpha diversity index indicates that their gut microecosystems are more complex and stable [46]. In contrast, beta diversity indices reveal that the microbiota structure of T. belangeri differs significantly from that of the other three species, while no such significant difference exists among the latter three. This strongly suggests that although A. chevrieri and T. belangeri appear similar in core microbiota, T. belangeri—as a unique representative of the order Scandentia—is relatively distantly related to the other three species. Its overall microbiota composition is profoundly shaped by its unique host genetics, immune system, and more specialized ecological niche, resulting in significant differences in gut microbial diversity from other species.

The proportion of positive edges in the co-occurrence networks of all four species exhibit a relatively high level, indicating that mutualism among microorganisms is predominant within the gut microbiota of each species. This facilitates the formation of functionally complementary symbiotic systems, enhancing metabolic stability and environmental adaptability among gut microbes [51]. Among them, the co-occurrence networks of gut microbiota in E. miletus and E. olito have significantly more nodes but fewer edges. This indicates that their gut microbiota are divided into more independent modules with strong modularity. Such a community structure may exhibit higher resilience when facing external environmental changes. In contrast, the co-occurrence networks of T. belangeri and A. chevrieri have fewer nodes but significantly more edges than those of the two vole species. This suggests that the microbiota of these two species have more refined internal division of labor and closer interactions, characterized by higher network complexity and connectivity. It implies that their communities are more suitable for efficiently performing specific functions in stable environments but may have lower flexibility and be more sensitive to external disturbances.

Further combined with the analysis of community assembly mechanisms, the microbiota assembly of E. olitor, A. chevrieri, and T. belangeri is dominated by deterministic processes. This indicates that deterministic factors such as environmental filtering, host selection, and interspecific competition play a major role in shaping their microbiota structure, making their gut microbiota more predictable. This is because E. olitor is only distributed in local areas such as Zhaotong, with a narrow vertical altitude range of 800–2600 m in humid shrublands [21], while A. chevrier is concentrated in meadows and coniferous forests above 1500 m in the Hengduan Mountains [24]. Their narrow distribution ranges and concentrated food resources result in more pronounced environmental filtering effects. In particular, the combination of T. belangeri’s unique diurnal arboreal lifestyle (completely different from the other three species), high network connectivity, and deterministic assembly mechanisms strongly suggests that its gut microbiota has high structuralization and functional stability, which is highly adapted to a more specific or stable ecological niche. In contrast, the gut microbiota assembly of E. miletus is dominated by stochastic processes, indicating that its microbiota composition is susceptible to random dispersal, individual differences, or stochastic events, and its microbiota functions may be more difficult to predict. This may be related to the wide distribution of E. miletus in the Hengduan Mountains region, which spans a broad vertical altitude range of 600–4200 m [52]. The sources of herbivorous food vary greatly across different altitudes, leading to stochastic colonization sources of gut microbiota and the absence of fixed environmental filtering rules.

Based on the above results, we conclude that the microbiota structure of each of the four species is influenced by the host’s own evolutionary history and specific adaptations, and each exhibits unique characteristics as a direct reflection of the host’s distinct living environment and evolutionary pressure. Nevertheless, the endemic species (E. miletus and E. olitor) and the alien species (A. chevrieri and T. belangeri) still show strong similarities in gut microbiota community characteristics. The gut microbiota of E. miletus and E. olitor are more complex, with a larger number of microbial taxa but fewer connections and collaborations among them. The advantage of this microbiota structure is stronger stress resistance in the face of complex and variable environmental changes, but it also means that the host lacks strong adaptability to specific environmental conditions. In particular, E. miletus, which is widely distributed in the Hengduan Mountains, has a microbiota assembly dominated by stochastic processes, further indicating its broad environmental tolerance. In contrast, the gut microbiota of A. chevrieri and T. belangeri have closer connections and internal cooperation, exhibiting strong gut microbial functionality and adaptability to specific environments. Their community assembly is naturally driven by strong host regulation and environmental adaptation. These results strongly corroborate the convergent evolution phenomenon existing in the gut microbiomes of endemic and alien species respectively, and also reflect the divergence in adaptive strategies between the two groups in terms of evolutionary history, feeding habits, habitat selection, and immune regulation in the Hengduan Mountains. Together, these factors determine the coexistence and breeding strategies of different species in the complex habitat of the Hengduan Mountains, characterized by large altitude differences and significant seasonal variations in climate and vegetation.

Conclusion

In conclusion, through multi-dimensional analysis of the gut microbiota of four small mammal species, this study found that the gut microbiomes of endemic species (E. miletus and E. olitor) and alien species (A. chevrieri and T. belangeri) have undergone convergent evolution respectively. Combined with the experimental result that body weight is significantly positively correlated with most differential bacterial taxa (P < 0.05), we speculate that this convergent evolution is caused by dietary overlap. Meanwhile, the similarity of gut microbiomes differs from our initial prediction: specifically, E.miletus and E. olitor show high similarity, and their microbial functions are mainly manifested in the digestion of cellulose-rich plant roots and stems; whereas A. chevrieri and T. belangeri have high similarity, and the enrichment of gut microbiota in their bodies indicates that the hosts have relatively complex feeding habits. Gut microbial diversity and community co-occurrence networks together indicate that the modular structure of gut microbiota in E. miletus and E. olitor endows them with strong resilience in response to environmental changes. In contrast, the gut microbiota of A. chevrieri and T. belangeri have lower species richness but close interactions among taxa, efficiently performing specific adaptations to particular living environments. In terms of assembly mechanisms, the gut microbiota of E. olitor, A. chevrieri, and T. belangeri are dominated by deterministic processes, resulting in highly predictable microbiota that are adapted to narrow habitats or concentrated food sources. In contrast, E. miletu is dominated by stochastic processes, which is consistent with its wide distribution and diverse food sources, conferring strong flexibility in environmental adaptation. Overall, as endemic species of the Hengduan Mountains, E. miletus and E. olitor represent a type of species that utilizes diverse plant-based food resources with strong resilience to environmental changes and broad environmental tolerance. This is their long-term environmental adaptation strategy to the complex habitat of the Hengduan Mountains, where climate and vegetation exhibit significant seasonal variations. In contrast, the later immigrant alien species (A. chevrieri and T. belangeri) have evolved omnivorous adaptive strategies targeting specific environmental conditions to occupy ecological niches in the Hengduan Mountains.

This study has the following limitations and shortcomings. First, due to the limited sample size, it is impossible to effectively verify the effects of host sex and body weight on their gut microbiomes, nor to explore the underlying mechanisms of such effects in depth. Second, the abnormally high proportion of positive edges in the co-occurrence network may be due to insufficient sequencing depth or inadequate filtering of low-abundance taxa. We expect that future studies will expand the sample size and improve the sequencing depth to accurately investigate the correlations between sex, body weight, and host gut microbiota and obtain an accurate co-occurrence network, thereby providing a more sufficient theoretical basis for research on the gut microbiota, niche characteristics, and evolutionary history of these four mammal species.

Abbreviations

SCFAs

short-chain fatty acids

RDA

redundancy analysis

ANOVA

analysis of variance

mean ± SE

mean±standard error

MST

Minimum Spanning Tree Index

EM

Eothenomys miletus

EO

Eothenomys oliter

TB

Tupaia belangeri

AC

Apodemus chevrieri

Authors’ contributions

Zhu WL conceived the study, participated in design, and coordination and drafted the manuscript. Ma Y and Jia T carried out the studies of body mass, gut microbiome and physiological indexes. Fan LX carried out the morphological studies and revised the manuscript. All authors contributed to the article and approved the submitted version.

Funding

This work was financially supported by the National Natural Scientific Foundation of China (32560262), Yunnan Fundamental Research Projects (202401AS070039).

Data availability

The 16 S rRNA sequences used in this study can be found on the European Nucleotide Archive (http://www.ebi.ac.uk/ena) under accession number: PRJEB72774.

Declarations

Ethics approval and consent to participate

The study was approved by the Animal Care and Use Committee of Yunnan Normal University (Accession number: 13-0901-011). In addition, the method of euthanasia on Eothenomys miletus, Eothenomys oliter, Tupaia belangeri, and Apodemus chevrieri was performed in accordance with the American Veterinary Medical Association (AVMA) guidelines for the euthanasia of animals (2020). Carbon dioxide have been used as euthanasia methods with low concentrations in AVMA guidelines for euthanasia.These four species is listed as Least Concern in the International Union for Conservation of Nature (lUCN) red list status (https://www.iucnredlist.org) and not endangered or protected in China. All sampling was permitted by the local level authority in scientific research.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.