Multilocus Sequence Typing ofLactobacillus caseiReveals a Clonal Population Structure with Low Levels of Homologous Recombination

Abstract
Robust genotyping methods forLactobacillus caseiare needed for strain tracking and collection management, as well as for population biology research. A collection of 52 strains initially labeledL. caseiorLactobacillus paracaseiwas first subjected torplBgene sequencing together with reference strains ofLactobacillus zeae,Lactobacillus rhamnosus, and other species. Phylogenetic analysis showed that all 52 strains belonged to a single compactL. casei-L. paracaseisequence cluster, together with strain CIP107868 (= ATCC 334) but clearly distinct fromL. rhamnosusand from a cluster withL. zeaeand CIP103137T(= ATCC 393T). The strains were genotyped using amplified fragment length polymorphism, multilocus sequence typing based on internal portions of the seven housekeeping genesfusA,ileS,lepA,leuS,pyrG,recA, andrecG, and tandem repeat variation (multilocus variable-number tandem repeats analysis [MLVA] using nine loci). Very high concordance was found between the three methods. Although amounts of nucleotide variation were low for the seven genes (π ranging from 0.0038 to 0.0109), 3 to 12 alleles were distinguished, resulting in 31 sequence types. One sequence type (ST1) was frequent (17 strains), but most others were represented by a single strain. Attempts to subtype ST1 strains by MLVA, ribotyping, clustered regularly interspaced short palindromic repeat characterization, and single nucleotide repeat variation were unsuccessful. We found clear evidence for homologous recombination during the diversification ofL. caseiclones, including a putative intragenic import of DNA into one strain. Nucleotides were estimated to change four times more frequently by recombination than by mutation. However, statistical congruence between individual gene trees was retained, indicating that recombination is not frequent enough to disrupt the phylogenetic signal. The developed multilocus sequence typing scheme should be useful for future studies ofL. caseistrain diversity and evolution.