Gut microbiome signatures of extreme environment adaption in Tibetan pig

1050 SGBs clustered from 8210 high-medium quality metagenome-assembled genomes (MAGs)

Over 779 Gb metagenomic sequence data was generated by Illumina HiSeq × Ten platform from 65 TPs samples (Supplementary Table 1). We integrated the above sequencing raw reads with a published metagenomic sequence dataset of 287 captive pigs from Denmark, France, and China²¹ to generate the gut genome catalog of pigs. Using a single-sample assembly strategy, we reconstructed a total of 8210 MAGs with a quality threshold of completeness >75% and contamination <10%^22,23 (see “Methods”; Fig. 1A). In total, 3807 of these MAGs were high-quality genomes with >90% completeness with <5% contamination (Fig. 1B). After de-replication at an average nucleotide identity (ANI) threshold of 95%²⁴, 1050 SGBs were identified for further analysis (see “Methods”). We used at least 40% genome coverage to determine the presence of SGBs in each sample, and 1048 representative SGBs were finally obtained, of which 623 SGBs (59.45%) were high-quality genomes (>90% completeness and <5% contamination) (Supplementary Table 2). Each SGB was supported by an average of 7.8 MAGs and 57.25% of SGBs contained at least two MAGs (Supplementary Table 2). We used the genome taxonomy database toolkit (GTDB-Tk)²⁵ to perform the taxonomic assignment of the SGBs (see “Methods)”. The results showed they were classified into 20 bacterial phyla and one archaeon phylum, 90.74% of SGBs were assigned to known genera, and 73.47% of SGBs were unclassified species (named uSGBs) (Fig. 1C). Besides, 45.04% of 1048 SGBs were assigned into Firmicutes A, 25.86% to Bacteroidetes, 6.97% to Firmicutes, and 5.63% to Proteobacteria (Supplementary Table 3). The prevalence and classification of SGBs in TPs, EPs, and CPs were shown in Fig. 1D, indicating the differences in the microbial community at the phylum level between the three groups. Additionally, functional gene profiles of 1048 SGBs were predicted using MetaGeneMark (v.3.38)²⁶ (see “Methods”). All gene annotations were performed by using the Carbohydrate-active enzymes (CAZymes)²⁷ and Kyoto Encyclopedia of Genes and Genomes (KEGG) Orthology (KO)²⁸ databases (see “Methods”).

Chen et al.²⁹ identified 6339 MAGs (>50% completeness and <5% contamination) and 2673 SGBs (ANI ≥ 95%) from the gut microbiome dataset of 500 Chinese pigs. To explore the uniqueness of our identified SGBs, we integrated our 8210 MAGs with Chen’s dataset and kept the MAGs with >75% completeness and <5% contamination for the identification of SGBs at the threshold of ANI ≥ 95%. A total of 2266 SGBs were finally obtained. 1248 (55.08%) of them were unique to Chen’s study, 519 (22.90%) to this study, and 499 (22.02%) were overlapped (Supplementary Fig. 1). This finding indicates that ongoing efforts are needed for understanding the pig gut microbial diversity.

The differences in microbial community between TPs, EPs, and CPs

The alpha and beta diversity analyses of gut microbiota from four groups (TPs, CPs, EPs-Denmark, EPs-France) were performed based on the prevalence or abundance of SGBs (see “Methods”). The results demonstrated that Species Richness, Shannon, and Simpson diversity indexes were significantly increased in TPs in comparison to other groups (Fig. 2A, Wilcoxon rank-sum test, p < 0.001). The principal co-ordinates analysis (PCoA) (based on the Weighted UniFrac distance matrix) and non-metric multidimensional scaling (NMDS) plotting analysis (based on Jaccard distance matrix) consistently indicated a clear separation of TPs from the other groups (Fig. 2B, Supplementary Fig. 2A, B). Moreover, the difference in the microbial diversity between sample types (fecal and cecum) was significantly lower than that between host groups (Supplementary Fig. 2C–F), suggesting that the TPs harbor distinct gut microbiota. For instance, 14.12%, 8.97%, and 1.43% of all SGBs were present uniquely in TPs, EPs, and CPs, respectively (Fig. 2C). According to the presence or absence of each SGB in each sample, we found that 464 SGBs were significantly associated with TPs, 209 SGBs with EPs, and 146 SGBs with CPs (Supplementary Table 4, Wilcoxon rank-sum test, p < 0.05, FDR corrected). At the phylum level, the prevalence of Fibrobacterota and Elusimicrobia was significantly higher in TPs than that in EPs and CPs (Fig. 2D; Fig. 2E, Fisher’s test, p < 0.001). Instead, Methanobacteriota known for its role in methanogenesis³⁰ was extremely significantly less in TPs than in other groups (Fisher’s test, p < 0.001). CPs had the highest prevalence of Methanobacteriota (Fisher’s test, p < 0.001). Both TPs and CPs had a higher prevalence of Actinobacteriota and Verrucomicrobiota than EPs (Fisher’s test, p < 0.01). The Desulfobacterota and Planctomycetotah existed mainly in CPs (Fisher’s test, p < 0.001), and the Deferribacterota and Verrucomicrobiota_A were prominently present in EPs (Fisher’s test, p < 0.05).

A Comparison of alpha diversity indices of gut microbiome among TPs, EPs, and CPs. The violin plot and box plot represent the Species richness, Shannon, and Simpson diversity index, respectively. Species Richness bar plot based on the presence or absence of SGBs in each sample, Shannon and Simpson diversity index based on the relative abundance of SGBs in each sample. The colors indicate host groups. The related source data is provided as a source data file. B Comparison of beta diversity among TPs, EPs, and CPs. PCoA was based on a weighted UniFrac distance matrix among samples. The colors and shapes indicate host groups and sample types, respectively. PCoA plotting to show the microbial community of TPs separated from those of EPs and CPs (Per mutational multivariate analysis of variance, p = 0.001, adjusted R² = 0.44). B NMDS plotting analysis was based on the Jaccard distance matrix among samples. The colors and shapes indicate host groups and sample types, respectively. NMDS plotting demonstrates the microbial community of TPs separate from EPs and CPs (stress=0.13). The related source data is provided as a source data file. C Distribution of SGBs among TPs, EPs, and CPs. Venn diagram showing the prevalence of SGBs in TPs, EPs, and CPs. The colors indicate the sample group. D Distribution of microbiota at phylum-level among TPs, EPs, and CPs. Pie chart showing the prevalence of phyla in TPs, EPs, and CPs. E, Comparison of enrichment analysis at phylum-level among TPs, EPs, and CPs. Bubble plot color and size correspond to the prevalence of SGBs in a certain phylum. The heat map shows the enrichment significance level (no * representing p > 0.05), and the colors indicate enriched groups.

Profiling of CAZymes in host-groups-associated SGBs

To investigate the potential of carbohydrate utilization of host-groups-associated SGBs, we performed CAZyme annotation using dbCAN2³¹ and performed enrichment analysis of CAZymes via Fisher’s test (see “Methods”). The results demonstrated that TPs-associated SGBs encoded much more carbohydrate utilization genes than captive pigs-associated SGBs, and the genes mainly belonged to glycoside hydrolases (GHs) families, carbohydrate esterases (CEs) families and polysaccharide lyases (PLs) families (Supplementary Table 5 and Fig. 3, Fisher’s test, p < 0.05). For example, TPs-associated SGBs included more genes in GH43, CE12, GH105, GH88, GH35, and GH51 than CPs and EPs (Fisher’s test, p < 0.001). They also included more genes in CE1, PL1, and GH53 than CPs, and genes in GH127 and GH146 than EPs, respectively (Fisher’s test, p < 0.05). Based on the CAZy and dbCAN-PUL database, we found that all the CAZymes of GH43, CE12, GH105, GH35, and GH51 were associated with the degradation of plant polysaccharides, such as cellulose, pectin, chitin, xylan, and other glucans. GH88 is related to the utilization of N-glycan and glycosaminoglycan. CE1, PL1, GH53, GH127, and GH146 are involved in the degradation of multiple polysaccharides, such as cellulose, hemicellulose, and other beta-glucans. These findings suggest that the TPs gut microbiome has significantly stronger polysaccharide utilization capability than captive pigs.

Heat map showing the significant difference in CAZymes (except glycosyltransferases) between TPs, EPs, and CPs, and their related potential substrates. It only demonstrates the significantly enriched CAZymes in the former host group.

Functional potential profile of TPs-associated bacteria

The Fibrobacterota and Elusimicrobia existed mainly in TPs rather than other groups (Fisher’s test, p < 0.001). Notably, the SGBs in the two bacteria phyla were all of high quality with > 90% completeness and < 5% contamination (Supplementary Table 2). Hence, we performed KO annotation and enrichment analysis of KEGG pathways to decipher their functional potential. The results showed that there were 38 KEGG pathways significantly enriched in the SGBs of Fibrobacterota and 39 pathways in the SGBs of Elusimicrobia, respectively (Supplementary Table 6, Fisher’s test, p < 0.05, FDR corrected). Overall, the two phyla bacteria exhibited different metabolic characteristics, even though the most enriched ten pathways in them both included translation, replication and repair, signal transduction, and energy metabolism (Fig. 4, Fisher’s test, p < 0.001). Fibrobacterota was particularly involved in the metabolism of cofactors and vitamins, as well as amino acid metabolism, and Elusimicrobia mostly participated in glycan biosynthesis and metabolism, as well as carbohydrate metabolism. Additionally, enriched pathways involved in amino acid metabolism were distinct between Fibrobacterota and Elusimicrobia (Fig. 4 and Supplementary Table 6, Fisher’s test, p < 0.05, FDR corrected). For example, Fibrobacterota was significantly enriched in the pathways of valine, leucine, and isoleucine biosynthesis (Fisher’s test, p = 0.002), alanine, aspartate, and glutamate metabolism (Fisher’s test, p = 0.010), arginine biosynthesis (Fisher’s test, p = 0.014), phenylalanine, tyrosine, and tryptophan biosynthesis (Fisher’s test, p = 0.023), lysine biosynthesis (Fisher’s test, p = 0.025), histidine metabolism (Fisher’s test, p = 0.025), cysteine and methionine metabolism (Fisher’s test, p = 0.032). Elusimicrobia preferred lysine biosynthesis (Fisher’s test, p = 0.0174), and glycine, serine, and threonine metabolism (Fisher’s test, p = 0.0277).

Bubble plot showing the significantly enriched KEGG pathways in Elusimicrobia and Fibrobacterota. Enrichment significance (p-value) was measured with Fisher’s test (see “Methods”). Bubble color responds to the enrichment significance and bubble size is related to the ratio of the number of genes mapped to a certain pathway. The same color of metabolism pathways on the right indicates the same pathway module.

Synthesis of energy sources and key precursor metabolites of SGBs in phyla Fibrobacteres and Elusimicrobiota

Metabolic pathway analysis demonstrated that the SGBs in Fibrobacteres and Elusimicrobiota contained a series of key enzyme genes in the glycolysis pathway (10 genes) and pyruvate metabolism pathway (2 genes) for catalyzing the conversion of glucose 6- phosphate to pyruvate and further into short-chain fatty acids (SCFAs), medium-chain fatty acids (MCFAs), lactate, ethanol and so on (Fig. 5, Supplementary Fig. 3 and Supplementary Table 7). Moreover, they encoded numerous enzyme genes for fatty acids synthesis, the conversion of acetyl-CoA to malonyl-CoA, as well as the synthesis of octanoic acid, decanoic acid, and dodecanoic acid. Nevertheless, the SGBs in Fibrobacteres encoded more enzyme genes in these pathways than in Elusimicrobiota, except in the synthesis pathways of lactate, butanoate, ethanol, and propanoate. It was remarkable that the SGBs in two phyla encoded some unique enzyme genes to participate in a certain catalytic process. For example, the SGBs in Fibrobacteres were solely involved in acetic acid synthesis through two pathways. One was the conversion of pyruvate directly into acetic acid. Another was pyruvate into acetyl adenylate then into acetic acid. The SGBs in Elusimicrobiota were merely involved in the direct conversion of pyruvate into lactate. Certainly, some pathways might require inter-cooperation of the two phyla, such as the butanoate synthesis. The key enzymes trans-2-enoyl-CoA reductase [EC:1.3.1.44] and butyrate kinase [EC:2.7.2.7] were annotated from the SGBs in Fibrobacteres and in Elusimicrobiota, respectively.

Elusimicrobia and Fibrobacterota were involved in polysaccharide degradation, membrane transport, glycolysis, and TCA cycle, as well as anabolism of fatty acids, amino acids, and B vitamins and cofactors. DHAP dihydroxyacetone phosphate, GAP glyceraldehyde 3-phosphate, PRPP 5-phosphoribosyl diphosphate, 3PG 3-phosphoglycerate, PEP phosphoenolpyruvate, AIR amino imidazole ribonucleotide, IMP inosine monophosphate, GTP guanosine 5′-triphosphate, D-Ru5P d-ribulose 5-phosphate, FAD flavin adenine dinucleotide, FMN riboflavin-5-phosphate, DHF 7,8-Dihydrofolate, THF tetrahydrofolate, AL (S)-2-acetolactate, OIV 2-oxo isovalerate, NAD+ nicotinamide adenine dinucleotide, NADP+ nicotinamide adenine dinucleotide phosphate.

Additionally, abundant enzyme genes catalyzing the process of non-oxidative branches of the pentose phosphate pathway were encoded by the SGBs in Fibrobacteres and Elusimicrobiota (Supplementary Fig. 3). For example, transketolase [EC:2.2.1.1], ribulose-phosphate 3-epimerase [EC:5.1.3.1], ribose 5-phosphate isomerase A [EC:5.3.1.6], and ribose-phosphate pyrophosphokinase [EC:2.7.6.1] catalyze the conversion of fructose 6-phosphate to erythrose 4-phosphate (erythrose-4P) and to 5-Phospho-alpha-D-ribose 1-diphosphate (PRPP). It is known that erythrose-4P is the essential substrate for chorismate synthesis through the shikimate pathway. Chorismate can serve as a precursor for the biosynthesis of tyrosine, phenylalanine, and tryptophan. And PRPP is the key precursor for histidine synthesis and formation of amino imidazole ribonucleotide (AIR) and guanosine triphosphate (GTP), which are used for the biosynthesis of vitamin B1, B2, B9, flavin adenine dinucleotide (FAD), flavin mononucleotide (FMN) and tetrahydrofolate (THF).

Synthesis of amino acids and vitamins (or cofactors) of SGBs in phyla Fibrobacteres and Elusimicrobiota

The enrichment analyses on metabolic pathways of Fibrobacteres and Elusimicrobiota illustrated that the two phyla both were involved in the synthesis of 20 essential amino acids (Fig. 5, Supplementary Fig. 4, and Supplementary Table 7). The SGBs in Fibrobacteres encoded more related enzyme genes than the ones in Elusimicrobiota. For instance, the SGBs in Fibrobacteres contained all of the key enzyme genes that catalyzed PRPP to form histidine (9 genes), oxaloacetate to glutamate and further to arginine (11 genes), oxaloacetate to aspartate (1 gene), aspartate to lysine (7 genes) and to threonine(5 genes) as well as to glycine (7 genes), 3-phosphoglycerate to serine (3 genes), pyruvate to alanine (2 genes) and to valine (5 genes), erythrose-4P to chorismate and further to tryptophan (12 genes). The results indicated Fibrobacteres had a stronger capacity to synthesize amino acids than Elusimicrobiota. The SGBs in Elusimicrobiota encoded all seven key enzyme genes in the lysine synthesis pathway (7 genes) and all the key enzyme genes catalyzed the conversion of threonine to glycine and further to serine, such as threonine 3-dehydrogenase [EC:1.1.1.103].

In addition, the SGBs in Fibrobacteres and Elusimicrobiota contained multiple enzyme genes in the synthesis pathways of B vitamins and their active forms with the exception of pyridoxine and cobalamin (Supplementary Fig. 4). For example, the SGBs in two bacteria phyla included 10 genes in the synthesis pathway of thiamine, 13 genes in niacin, nicotinamide, and nicotinamide adenine dinucleotide (NAD+), 11 genes in pantothenate and coenzyme A (CoA), 13 genes in biotin, as well as 19 genes in heme. It was remarkable that Fibrobacteres had all essential enzyme genes for catalyzing GTP to form riboflavin (7 genes), folate, and THF (12 genes), indicating their important roles in the B vitamins synthesis. The SGBs in Elusimicrobiota contained more enzyme genes in niacin and nicotinamide synthesis.

Source

Janice

Janice is a full-time mom who likes to write on a range of topics in her spare time. She specializes in the Home, Garden, and Recycling topics. Janice is our Lifestyle and positive vibe expert. She keeps the office running.