Comparison of the oral microbiome in mouthwash and whole saliva samples

Abstract

Population-based epidemiologic studies can provide important insight regarding the role of the microbiome in human health and disease. Buccal cells samples using commercial mouthwash have been obtained in large prospective cohorts for the purpose of studying human genomic DNA. We aimed to better understand if these mouthwash samples are also a valid resource for the study of the oral microbiome. We collected one saliva sample and one Scope mouthwash sample from 10 healthy subjects. Bacterial 16S rRNA genes from both types of samples were amplified, sequenced, and assigned to bacterial taxa. We comprehensively compared these paired samples for bacterial community composition and individual taxonomic abundance. We found that mouthwash samples yielded similar amount of bacterial DNA as saliva samples (p from Student's t-test for paired samples = 0.92). Additionally, the paired samples had similar within sample diversity (p from = 0.33 for richness, and p = 0.51 for Shannon index), and clustered as pairs for diversity when analyzed by unsupervised hierarchical cluster analysis. No significant difference was found in the paired samples with respect to the taxonomic abundance of major bacterial phyla, Bacteroidetes, Firmicutes, Proteobacteria, Fusobacteria, and Actinobacteria (FDR adjusted q values from Wilcoxin signed-rank test = 0.15, 0.15, 0.87, 1.00 and 0.15, respectively), and all identified genera, including genus Streptococcus (q = 0.21), Prevotella (q = 0.25), Neisseria (q = 0.37), Veillonella (q = 0.73), Fusobacterium (q = 0.19), and Porphyromonas (q = 0.60). These results show that mouthwash samples perform similarly to saliva samples for analysis of the oral microbiome. Mouthwash samples collected originally for analysis of human DNA are also a resource suitable for human microbiome research.

Fig 1. Alpha-diversity of oral bacterial communities in the paired mouthwash-saliva samples.

Bar plots of number of observed OTUs (a) and Shannon Index (b) in paired mouthwash-saliva samples in 10 subjects. These indices were calculated for 500 iterations of rarefied OTU table with minimum sequencing depth of 38,400 among all study subjects, with the average over the iterations taken for each participant. No differences were found between mouthwash and saliva samples in α-diversity (p from paired t-test = 0.33 for richness, and 0.51 for Shannon index).

Fig 2. Beta-diversity of oral bacterial communities in the paired mouthwash-saliva samples.

Hierarchical cluster analysis using JSD distance. AU (approximately unbiased) p-values, the unbiased bootstrap probability, ranged from 0.97 to 1.00 for all paired samples in hierarchical cluster analysis with number of 1,000 bootstrap replications. Cluster with AU ≥ 0.95 are considered to be strongly supported by data. S01-S10 indicate study subject 1 to 10. “M” indicates mouthwash sample and “S” indicates salivary sample.

Fig 3. Correlation of the centered Log-Ratio (clr) transformed count of major bacteria phyla and genera in the paired mouthwash-saliva samples.

Correlation of clr-transformed counts in mouthwash and saliva samples of major bacterial phyla (Panel A) and genera (Panel B). The x-axis represents the transformed counts in mouthwash samples, and the y-axis represents transformed counts in saliva samples. The straight line is the line of equality. All FDR adjusted q values from Wilcoxon signed-rank test for the comparison of the taxonomic abundance in paired samples were >0.05.