Potential of gut microbiota for lipopolysaccharide biosynthesis in European women with type 2 diabetes based on metagenome

Abstract

The abnormal accumulation of lipopolysaccharide (LPS) plays a crucial role in promoting type 2 diabetes (T2D). However, the capability of the gut microbiota to produce LPS in patients with T2D is still unclear, and evidence characterizing the patterns of gut microbiota with LPS productivity remains rare. This study aimed to uncover the profiles of LPS-biosynthesis-related enzymes and pathways, and explore the potential of LPS-producing gut microbiota in T2D. The gut metagenomic sequencing data from a European female cohort with normal glucose tolerance or untreated T2D were analyzed in this study. The sequence search revealed that the relative abundance of the critical enzymes responsible for LPS biosynthesis was significantly high in patients with T2D, especially for N-acetylglucosamine deacetylase, 3-deoxy-D-manno-octulosonic-acid transferase, and lauroyl-Kdo2-lipid IVA myristoyltransferase. The functional analysis indicated that a majority of pathways involved in LPS biosynthesis were augmented in patients with T2D. A total of 1,173 species from 335 genera containing the gene sequences of LPS enzymes, including LpxA/B/C/D/H/K/L/M and/or WaaA, coexisted in controls and patients with T2D. Critical taxonomies with discriminative fecal abundance between groups were revealed, which exhibited different associations with enzymes. Moreover, the identified gut microbial markers had correlations with LPS enzymes and were subsequently associated with microbial pathways. The present findings delineated the potential capability of gut microbiota toward LPS biosynthesis in European women and highlighted a gut microbiota-based mechanistic link between the disturbance in LPS biosynthesis and T2D. The restoration of LPS levels through gut microbiota manipulation might offer potential approaches for preventing and treating T2D.

Keywords: enzyme; gut microbiota; lipopolysaccharide; metagenome; type 2 diabetes.

FIGURE 1

Relative abundance of LPS-biosynthesis-related enzymes in the gut of individuals with NGT and patients with T2D. (A) Schematic diagram of the direct synthetic process and transformation pathways for LPS biosynthesis. Key enzymes catalyzing the progress from UDP-GlcNAc to Kdo2-lipid A were crucial and necessary for LPS biosynthesis, and genes encoding these enzymes were targeted for exploring their abundance. LpxA, UDP-N-acetylglucosamine acyltransferase; LpxC, UDP-3-O-[3-hydroxymyristoyl] N-acetylglucosamine deacetylase; LpxD, UDP-3-O-[3-hydroxymyristoyl] glucosamine N-acyltransferase; LpxH, UDP-2,3-diacylglucosamine hydrolase; LpxB, lipid-A-disaccharide synthase; LpxK, tetraacyldisaccharide 4′-kinase; WaaA, 3-deoxy-D-manno-octulosonic-acid transferase; LpxL, Kdo2-lipid IVA lauroyltransferase/acyltransferase; LpxM, lauroyl-Kdo2-lipid IVA myristoyltransferase. (B) Box plots depicting the relative abundance levels of LPS-biosynthesis-related enzymes between individuals with NGT and patients with T2D based on the genes encoding the enzymes (orange, T2D; blue, NGT). Boxes represent the interquartile ranges; the inside lines indicate the median; circles are outliers, and red asterisk represents statistical difference at p < 0.05 (Wilcoxon rank-sum test). (C) Enrichment of LPS-producing enzymatic genes according to the encoding genes was evaluated using Log2 (fold change T2D/NGT, FC). Log2 (FC) > 0 and Log2 (FC) < 0 indicate the enrichment of enzymes in patients with T2D and individuals with NGT, respectively. The colors from dark purple to light yellow indicate the scales of p-value. (D) Co-occurrence network of the LPS-biosynthesis-related enzymes in the study cohort. The numbers labeled in the enzymes indicate the number of connections with others. The thresholds calculated via Spearman’s correlation analyses were p-value < 0.05 and |r| >0.4. The pink lines connecting two enzymes denote a positive correlation. Line thickness and node size represent the |r| value of correlation and relative abundance of enzymatic genes, respectively. T2D, type 2 diabetes; NGT, normal glucose tolerance.

FIGURE 2

Gut microbial KOs implicated in LPS biosynthesis in individuals with NGT and patients with T2D. (A) KOs involved in synthetic progress and transformation pathways for LPS biosynthesis were identified through related compounds and enzymes. Different compounds or enzymes are shown in different colors. The outer circle in gray indicates the six KOs not detected in the present cohort. (B) UpSet plot and Venn diagram analysis showed the coexisting KO distributions in LPS-biosynthesis-related compounds and enzymes. The bar plot at the bottom left shows the number of KOs implicated in each compound and enzyme. The dot and line at the bottom right represent the unique and shared KO intersection by different compounds and enzymes, respectively. The number of specific KOs and shared KO intersections are shown in the black histogram above. Venn diagrams at the upper right show the number of annotated KOs involved in LPS-biosynthesis-related compounds and enzymes. One KO was shared by the LPS-biosynthesis-related enzymes and UDP-2,3-diacylglycosamine, and two KOs are shared by Kdo2-lipidA and enzymes. (C) PCA analysis based on the 45 annotated KOs across individuals with NGT and patients with T2D. p = 0.016 between groups, ANOSIM test. The distribution of patients with T2D according to KOs was significantly different from that in NGT at the PCA1 axis (p = 0.0089, Wilcoxon rank-sum test). (D) Correlation network reflecting the interaction between LPS-biosynthesis-related enzymes (circle) and KOs (square). Spearmen correlation analysis, |r| >0.4, p < 0.05. Lines connecting nodes indicate positive (solid lines in red) or negative (dotted lines in blue) correlations, and the thickness of lines is proportional to the correlation coefficients. T2D, type 2 diabetes; NGT, normal glucose tolerance.

FIGURE 3

Disparate abundance of gut microbial KOs associated with LPS biosynthesis between individuals with NGT and patients with T2D. (A) Enrichment of the LPS-biosynthesis-related KOs with distinct abundance between individuals with NGT and patients with T2D. Log2 (OR score) < 0 (enriched in NGT, blue); Log2 (OR score) > 0 (enriched in T2D, orange), p < 0.05. (B) Box plots illustrating the relative abundance of KOs varied significantly between the NGT and T2D groups. Horizontal lines within the box plots represent median values, upper and lower ranges of the box represent 75% and 25% quartiles, respectively, and circles represent outliers. T2D, type 2 diabetes; NGT, normal glucose tolerance.

FIGURE 4

Taxonomic identification and the distribution of bacteria harboring LPS-biosynthesis-related enzymatic genes in individuals with NGT and patients with T2D. (A,B) UpSet plot and Venn diagram showing the distribution of bacteria harboring LPS-biosynthesis-related enzymatic genes, shared and unique in the NGT and T2D groups at the genus, and species levels, respectively. Bar charts above represent the number of genera, and species harboring LPS-biosynthesis-related enzymatic genes, shared and specific between groups. At the bottom left, bar charts indicate the bacteria that harbored LPS-biosynthesis-related enzymatic genes in each group (orange, T2D; blue, NGT). The line between nodes at the bottom right denotes the intersected taxon within each group. (C,D) Stack columns depicting the relative abundance for the top 10 genera and species with LPS-producing potentials (enzymatic genes) in NGT and T2D. T2D, type 2 diabetes; NGT, normal glucose tolerance.

FIGURE 5

Abundance comparisons of candidate LPS-producing bacteria across the NGT and T2D groups. (A,B) Twenty-two genera, and 89 species harboring LPS-biosynthesis-related enzymatic genes were identified to be significantly different between individuals with NGT and patients with T2D using volcano plots, respectively. Orange and blue indicate significant enrichment and reduction in the T2D group, respectively; gray denotes a non-significant difference between the two groups. (C,D) Box plots showing the relative abundance of the top 10 gut genera and species with both genes for LPS-biosynthesis-related enzymes, and statistical difference between the NGT and T2D groups. Boxes represent the interquartile ranges, the inside lines represent the median, and circles represent outliers. Orange and blue represent enrichment in patients with T2D and individuals with NGT, respectively. T2D, type 2 diabetes; NGT, normal glucose tolerance.

FIGURE 6

Correlation of LPS-biosynthesis-related enzymes and harboring gut bacteria. (A) Spearmen correlation analysis of the relationship between potential LPS-producing genera (22 bacteria discriminative in patients with T2D and individuals with NGT) and enzymatic genes. Statistical differences in correlation coefficients between LPS-producing genus and one of the enzyme genes were shown (B) Heat map depicts the association between LPS-biosynthesis-related enzymes and top 30 altered species harboring enzymatic genes. Blue, negative correlation; red, positive correlation; ^* p < 0.05; ^** p < 0.01; ^*** p < 0.001. Correlation with p ≥ 0.05 is not shown. T2D, type 2 diabetes; NGT, normal glucose tolerance.

FIGURE 7

Interrelationship across gut microbes, LPS-biosynthesis-related enzymatic genes, and respective KOs. Sankey diagram based on Spearman correlation analysis. The gut microbes with the potential to produce LPS were related to LPS-biosynthesis-related enzymes and further linked to microbial KOs during LPS biosynthesis. Red, positive correlation (p < 0.05); blue, negative correlation (p < 0.05); gray, non-significant correlation (p ≥ 0.05). Square in yellow, enriched in T2D; square in white, enriched in NGT. T2D, type 2 diabetes; NGT, normal glucose tolerance.