Staphylococci in high resolution: Capturing diversity within the human nasal microbiota

Abstract

Staphylococci include both nasal commensals and opportunistic pathogens, globally responsible for a large proportion of infection-related deaths, especially in S. aureus carriers. To understand staphylococcal temporal dynamics within the nasal microbiota, we employed Staphylococcus-targeted sequencing in two cohorts from Denmark and Germany. We identified two major staphylococcal community state types (sCSTs)-one dominated by S. aureus and one dominated by S. epidermidis-and eight subgroups defined by co-colonizing coagulase-negative staphylococci. The distribution of sCSTs was similar between the two cohorts. Predominance of either S. aureus or S. epidermidis was highly persistent over time, whereas co-colonizing staphylococcal species were transient with varying stability among the sCST subgroups. Detection of S. aureus by culture was positively associated with absolute abundance by qPCR. S. aureus domination was diminished when Dolosigranulum and Corynebacterium co-occurred. Our findings could inform efforts to reduce S. aureus nasal colonization and infection.

Keywords: CP: Microbiology; Staphylococcus aureus; Staphylococcus epidermidis; antagonist; carriage; community state type; microbial abundance; microbiome; nasal microbiota; staphylococci.

Figure 1.. Study outline and analysis workflow

Nasal swabs from both cohorts were subjected to staphylococcus-specific tuf gene (staphylome) sequencing, followed by assessment of alpha and beta diversity and staphylococcal community state typing (black arrows). Additional analyses were performed only on the Danish samples (red arrows). Staphylome sequencing was performed on longitudinal samples, which was used for assessing intra-individual temporal stability of the staphylococcal composition. Further, 16S rRNA sequencing and qPCR analysis were used to determine bacterial absolute abundance and nasal microbiota community state types (CSTs). The tuf gene sequencing approach was validated by S. aureus culture results. sCST, staphylococcal community state type; S, Staphylococcus; ASV, amplicon sequence variant; MS, mass spectrometry. Created with BioRender. See also Figures S1, S2, S4, and S7 and Table S2.

Figure 2.. Diversity of staphylococci in the cross-sectional dataset

(A) A non-metric multidimensional scaling (NMDS) plot based on Jensen-Shannon divergence. An ellipse of 95% confidence based on multivariant t distribution was used. (B) Alpha diversity measurement comparison based on staphylococcal species between the Danish and the German cohorts (displayed as median) by Kruskal Wallis test with the following significance levels: *p ≤ 0.05, **p ≤ 0.01, ***p ≤ 0.001; ns, not significant. See also Figure S3.

Figure 3.. Pairwise correlation of staphylococcal species abundance

Spearman’s rank correlations were calculated between the relative abundance of the 10 most abundant staphylococcal species in the cross-sectional data from Denmark and Germany (n = 1,518). Only correlations with a correlation coefficient ρ < −0.1 or ρ > 0.1 are visualized. p values were corrected for multiple testing and are only displayed for correlations with a correlation coefficient ρ < −0.1 or ρ > 0.1. Significant p values are marked in the plot. Purple indicates a positive correlation, and green indicates a negative correlation. See also Table S3.

Figure 4.. Defining the sCSTs

This analysis was done on all samples (n = 1,863). (A) Hierarchical clustering and relative abundance of the 10 most abundant staphylococcal species in the eight defined sCSTs. “Others” represents all other detected staphylococci. (B) NMDS plot based on Jensen-Shannon divergence of the eight sCSTs based on species level, with each point indicating a sample, and a 95% confidence ellipse with a multivariate t distribution. (C) Shannon staphylococcal diversity, displayed as median per sCST. See also Table S4.

Figure 5.. Alluvial plot and fitted Markov chain models showing transitions in sCSTs over time

(A) Stability of sCSTs based on samples from three time points (available for 79 individuals). (B) A fitted Markov chain model showing the transition between sCST1 and sCST2 based on data from three time points (available for 79 individuals) and transition probabilities from time point 1 to time point 2 and from time point 2 to time point 3. (C and D) A fitted Markov chain model built from three time points, showing transition between sCST1, sCST2_Sa and other sCST2 subgroups (C) and showing transition between S. aureus-associated (sCST1_{Sa_H}, sCST1_{Sa_M}, and sCST2_Sa) and CoNS-associated sCSTs (D), with sCST2_Sl and sCST2_Se representing their own groups. See also Figure S5.

Figure 6.. Correlation between sCSTS and CSTs in 846 Danish cross-sectional nasal samples

Shown is a mosaic plot illustrating the contingency table of bacterial CSTs and sCSTs. The size of each tile is proportional to the number of samples it presents, with tile widths representing samples per bacterial CST and tile heights representing samples per sCST. Pearson residuals are shown, where blue indicates the given combination of CST and sCST occurring more often than if the data were random (positive correlation), and red indicates the given combination of CST and sCST occurring less often than if the data were random (negative correlation). Only CSTs with at least 30 samples and at least one correlation with a Pearson residual >2/<−2 are displayed. See also Figure S6.

Figure 7.. Abundance data for a subset of 846 Danish cross-sectional samples

Abundance data were derived by integrating 16S rRNA gene sequencing data, qPCR-derived 16S rRNA gene copy number, and the relative abundance of staphylococcal species obtained from tuf amplicon data in relation to the sCSTs. (A) The median bacterial, staphylococcal, S. aureus, and S. epidermidis absolute abundances are visualized for each sCST. (B) The staphylococcal relative abundance was obtained from 16S rRNA gene sequencing data. For distribution data (interquartile ranges), see Table S5.