Impact of a vegan diet on the human salivary microbiota

Abstract

Little is known about the effect of long-term diet patterns on the composition and functional potential of the human salivary microbiota. In the present study, we sought to contribute to the ongoing elucidation of dietary effects on the oral microbial community by examining the diversity, composition and functional potential of the salivary microbiota in 160 healthy vegans and omnivores using 16S rRNA gene amplicon sequencing. We further sought to identify bacterial taxa in saliva associated with host inflammatory markers. We show that compositional differences in the salivary microbiota of vegans and omnivores is present at all taxonomic levels below phylum level and includes upper respiratory tract commensals (e.g. Neisseria subflava, Haemophilus parainfluenzae, and Rothia mucilaginosa) and species associated with periodontal disease (e.g. Campylobacter rectus and Porphyromonas endodontalis). Dietary intake of medium chain fatty acids, piscine mono- and polyunsaturated fatty acids, and dietary fibre was associated with bacterial diversity, community structure, as well as relative abundance of several species-level operational taxonomic units. Analysis of imputed genomic potential revealed several metabolic pathways differentially abundant in vegans and omnivores indicating possible effects of macro- and micro-nutrient intake. We also show that certain oral bacteria are associated with the systemic inflammatory state of the host.

Figure 1

Differential abundance of operational taxonomic units. Volcano plot of estimated log₂ fold difference in operational taxonomic unit (OTU) abundance between vegans and omnivores and corresponding Benjamini-Hochberg adjusted P-values (Q) from negative binomial Wald tests as implemented in the DESeq2 R package. The red dotted line indicates the 10% false discovery threshold. Prevalence indicates percentage of participants in which a given OTU is present. Abundance indicates mean relative abundance (‰) of a given OTU. Name of OTUs differentially abundant at an FDR ≤ 5% are given at the lowest classified rank in Greengenes [Greengenes ID]. See Supplementary Table S2 for a full list of OTUs differentially abundant at an FDR < 10%. p, phylum, o, order. f, family. g, genus. s, species.

Figure 2

Salivatypes in vegans and omnivores. (A) Principal coordinates analysis visualizing salivatype clusters based on partitioning around medoids of Jensen-Shannon distance (JSD). Ellipses cover 67% samples in each cluster. (B) The ratio of Neisseria to Prevotella along the first principal coordinate axis of Jensen-Shannon distances used to build the salivatype clusters. (C–F) Relative abundance of the main contributors to each salivatype cluster. Differential abundance of each genus on which the clusters were build was tested using a Wilcoxon rank-sum test and the genera with P values < 10⁻¹⁰ are depicted. Boxes represent interquartile range (IQR), with the inside line representing the median. Whiskers represent values within 1.5 × IQR of the first and third quartiles. Circles represent outliers beyond the whiskers.

Figure 3

Effect of diet principal components on alpha and beta diversity. (A) Forest plot of effect sizes and corresponding 95% confidence intervals of the first ten diet principal components (PC) on observed richness, estimated (Chao1) richness, Shannon’s diversity index, and Simpson’s reciprocal index. Associations were tested using multiple regression including all ten PCs simultaneously. (B) Heatmap of variance in beta diversity explained by each of the ten first diet PCs as estimated by permutational analysis of variance. Q values are given for associations significant at an FDR ≤ 10% only.

Figure 4

Differential abundance of KEGG pathways in vegans and omnivores. Orthologous groups were collapsed at pathway level. Pathways within the overall topics of metabolism, environmental information processing, genetic information processing and cellular processes present in ≥10% of individuals were tested for differential abundance in vegans and omnivores using the DESeq2 package. Only metabolic pathways differentially abundant at an FDR ≤ 5% are depicted. Bars indicate mean (standard deviation) proportion of differentially abundant pathways in omnivores. Forest plot indicates log₂ fold difference and corresponding 95% confidence intervals for the difference in pathway abundance between vegans and omnivores. Pathways are coloured by functional category and ordered by decreasing log₂ fold-difference within each category.