The bacterial species profiles of the lingual and salivary microbiota differ with basic tastes sensitivity in human

Abstract

Taste perception is crucial and impairments, which can be linked to pathologies, can lead to eating disorders. It is triggered by taste compounds stimulating receptors located on the tongue. However, the tongue is covered by a film containing saliva and microorganisms suspected to modulate the taste receptor environment. The present study aimed to elucidate the links between taste sensitivity (sweetness, sourness, bitterness, saltiness, umami) and the salivary as well as the tongue microbiota using shotgun metagenomics. 109 bacterial species were correlated with at least one taste. Interestingly, when a species was correlated with at least two tastes, the correlations were unidirectional, indicating a putative global implication. Some Streptococcus, SR1 and Rickenellaceae species correlated with five tastes. When comparing both ecosystems, saliva appears to be a better taste predictor than tongue. This work shows the implication of the oral microbiota in taste and exhibits specificities depending on the ecosystem considered.

Figure 1

Flow chart of the different steps of the study.

Figure 2

Taste sensitivity of the panel (A) Biplot representation of the distribution of the participants (N = 100). Each dot represents the average of 3 tests. Females are represented by circles, and males by squares. The intensity of colour decreases with age. (B) Average taste sensitivity scores for 4 age classes (N = 25). Different letters indicate significant differences (p ≤ 0.05) between the age groups for a given taste. (C) Average sensitivity scores for males (N = 50) and females (N = 50). A star indicates a significant difference (p ≤ 0.05) between groups. A high score signifies a high taste sensitivity.

Figure 3

Bacterial composition of saliva and lingual film. (A) Venn diagrams showing the number of features (resp. phylum, genus and species) unique to the saliva (in blue) or lingual film (in red) and common to both (in purple), with (above) or without (below) occurrence filtering. (B) Relative abundance of the different phyla in the saliva and lingual film at the group level on the left (relative abundance average) and at the individual level on the right. (C) Box plot representation of the Spearman correlation coefficients. The green boxplot displays correlations between saliva samples; the blue box displays correlations between lingual samples, and the red box displays correlations between saliva and lingual samples from the same subject. A star indicates a significant difference (p ≤ 2.22–16) between samples.

Figure 4

Heatmap showing the Spearman correlation among differential phylum, genus, species expression and taste sensitivity scores in saliva and lingual film. Phyla, genera and species highlighted in yellow are common to the saliva and lingual film. Positive correlations are represented in red, and negative ones in blue. A positive correlation means that a higher abundance of the microbial taxa corresponds to a higher sensitivity to the taste. (p value < 0.05).

Figure 5

Boxplots showing the distribution of sensitivity scores of the participants according to their high (N = 50) or low (N = 50) abundance in the four species correlated significantly to the sensitivity to the five tastes in saliva. The number of stars (*p < 0.05; **p < 0.01; ***p < 0.001; ns: nonsignificant) indicates the significance level of the Wilcoxon test depicting the difference between the two groups.

Figure 6

Box plot representation of taste-prediction performances. The upper part of the plot displays the prediction model performances when predicting taste scores with saliva taxa. The lower part displays the prediction model performances when predicting taste scores with lingual taxa.