The unique gut microbiome of giant pandas involved in protein metabolism contributes to the host's dietary adaption to bamboo

Abstract

Background: The gut microbiota of the giant panda (Ailuropoda melanoleuca), a global symbol of conservation, are believed to be involved in the host's dietary switch to a fibrous bamboo diet. However, their exact roles are still largely unknown.

Results: In this study, we first comprehensively analyzed a large number of gut metagenomes giant pandas (n = 322), including 98 pandas sequenced in this study with deep sequencing (Illumina) and third-generation sequencing (nanopore). We reconstructed 408 metagenome-assembled genomes (MAGs), and 148 of which (36.27%) were near complete. The most abundant MAG was classified as Streptococcus alactolyticus. A pairwise comparison of the metagenomes and meta-transcriptomes in 14 feces revealed genes involved in carbohydrate metabolism were lower, but those involved in protein metabolism were greater in abundance and expression in giant pandas compared to those in herbivores and omnivores. Of note, S. alactolyticus was positively correlated to the KEGG modules of essential amino-acid biosynthesis. After being isolated from pandas and gavaged to mice, S. alactolyticus significantly increased the relative abundance of essential amino acids in mice jejunum.

Conclusions: The study highlights the unique protein metabolic profiles in the giant panda's gut microbiome. The findings suggest that S. alactolyticus is an important player in the gut microbiota that contributes to the giant panda's dietary adaptation by more involvement in protein rather than carbohydrate metabolism. Video Abstract.

Fig. 1

Schematic overview of the experimental design

Fig. 2

Taxonomic annotation and phylogenetic tree of 408 metagenome-assembled genomes (MAGs). A The outermost circle (green) and outer cycle (orange) represent completeness and contamination of MAGs. The different colors of the background of clades represent different bacterial phyla. The tree was constructed using PhyloPhlan (v3.0.2) and visualized using Interactive Tree of Life (iTOL, v6.5.2). B Distribution of completeness and contamination of these MAGs. C Taxonomic classification of MAGs at different levels. D Alignment rate to the MAGs for metagenomic clean reads (MG, N = 125) and meta-transcriptomic clean reads (MT, N = 14). E Alignment rate to the gene collection of giant pandas in the gut for metagenomic clean reads (MG, N = 125) and meta-transcriptomic clean reads (MT, N = 14)

Fig. 3

Composition and expression pattern of dietary fiber metabolic enzyme-related genes in the gut microbiota of different host species. Bray–Curtis diversity plot based on dietary fiber metabolic enzyme-related gene families. Bray–Curtis distance was measured by metagenomic abundance (A) and meta-transcriptomic expression (B) of these gene families. The ellipses were calculated and drawn with a 0.95 confidence level. Metagenomic abundance and meta-transcriptomic expression of crucial genes involved in hemicellulose degradation (EC 3.2.1.8, shown in C) and cellulase degradation (EC 3.2.1.4 and EC 3.2.1.20, shown in D and E) of different host species

Fig. 4

Abundance and expression pattern of genes involved in amino acid metabolism in gut microbiotas of different host species. Bray–Curtis distance between groups calculated with gene families encoding amino acid metabolic enzyme based on gene abundance from metagenomic (A AA degradation, C AA biosynthesis) and expression from meta-transcriptomic data (B AA degradation, D AA biosynthesis). The ellipses were calculated and drawn with a 0.95 confidence level

Fig. 5

Relative abundance of essential-amino biosynthesis KEGG pathway in giant pandas with different S. alactolyticus levels. The high, medium, and low on the x-axis of plots A, B, and C represent the high, medium, and low abundance of S. alactolyticus in giant pandas samples, respectively. Y-axis is the normalized total abundance of the KEGG module related to essential amino acid biosynthesis. Statistical significance (p value) was calculated by unpaired two-tailed t test

Fig. 6

Comparison of amino acid abundance for the different treatment groups in mice. A and B indicated the comparison of the total abundance of the non-essential amino acids (A) and essential amino acids (B) for different treatment mice groups. Boxplots of C-E showed the abundance of L-tyrosine, L-glutamic acid, and L-valine in different mice groups, respectively. Statistical significance (p value) was calculated by unpaired two-tailed t test