Diversity of lactobacilli in the oral cavities of young women with dental caries

Abstract

For nearly a century, lactobacilli (LB) in the oral cavity have been generally associated with dental caries. Here, we characterized the LB isolated from the saliva of 6 women with active caries using genetic-based taxonomical identification methods. From each subject, 30 isolates growing on Rogosa medium and presumed to be LB were analyzed. Of the 180 isolates, 176 were further characterized by biotyping, DNA melting points, DNA chromosomal fingerprinting, genotyping, and phylogenetic cluster assessment. We found a total of 30 unique genotypes of LB in the saliva of caries-active women, with each woman harboring between 2 and 8 distinct genotypes. Although Lactobacillus vaginalis, L. fermentum, and L. salivarius were found in 4 of 6 of the subjects, results from other studies using comparable methods show an entirely different array of LB associated with caries. These collective observations lead us to surmise that LB associated with dental caries are likely exogenous and opportunistic colonizers, arising from food or other reservoirs outside the oral cavity.

Fig. 1

Chromosomal DNA fingerprinting of representative subsample of Lactobacillus spp. obtained from 6 young adult women. Chromosomal DNA was digested with HindIII and the resulting fragments electrophoresed in agarose gels, stained with ethidium bromide and photographed under UV light. The far left lane displays the size standard with size (in kb) in the margin.

Fig 2

Phylogenetic analysis of the V2--V3 region of 16S rDNA amplicons from 30 Lactobacillus clinical strains along with known sequences designated as a Bergey’s type strain from the Ribosomal Database II with the closest match to each cluster of clinical samples. Fragments were first aligned using ClustalW along with known reference strains. Type strain S. mutans UA159 served as the designated out-group and the Lactobacillus sequences constrained to be monophyletic to the out-group after a tree was constructed using the Neighbor-Joining algorithm. Bootstrap values from 1,000 replicates are shown at each node. Seven well-supported clades containing 9 taxonomical entities were evident. Along the right side of each clade is the average %G+C content (and SD) of whole genomic DNA from each of the 30 strains of LB.