A Mobile Element in mutS Drives Hypermutation in a Marine Vibrio

Abstract

Bacteria face a trade-off between genetic fidelity, which reduces deleterious mistakes in the genome, and genetic innovation, which allows organisms to adapt. Evidence suggests that many bacteria balance this trade-off by modulating their mutation rates, but few mechanisms have been described for such modulation. Following experimental evolution and whole-genome resequencing of the marine bacterium Vibrio splendidus 12B01, we discovered one such mechanism, which allows this bacterium to switch to an elevated mutation rate. This switch is driven by the excision of a mobile element residing in mutS, which encodes a DNA mismatch repair protein. When integrated within the bacterial genome, the mobile element provides independent promoter and translation start sequences for mutS-different from the bacterium's original mutS promoter region-which allow the bacterium to make a functional mutS gene product. Excision of this mobile element rejoins the mutS gene with host promoter and translation start sequences but leaves a 2-bp deletion in the mutS sequence, resulting in a frameshift and a hypermutator phenotype. We further identified hundreds of clinical and environmental bacteria across Betaproteobacteria and Gammaproteobacteria that possess putative mobile elements within the same amino acid motif in mutS In a subset of these bacteria, we detected excision of the element but not a frameshift mutation; the mobile elements leave an intact mutS coding sequence after excision. Our findings reveal a novel mechanism by which one bacterium alters its mutation rate and hint at a possible evolutionary role for mobile elements within mutS in other bacteria.

Importance: DNA mutations are a double-edged sword. Most mutations are harmful; they can scramble precise genetic sequences honed over thousands of generations. However, in rare cases, mutations also produce beneficial new traits that allow populations to adapt to changing environments. Recent evidence suggests that some bacteria balance this trade-off by altering their mutation rates to suit their environment. To date, however, we know of few mechanisms that allow bacteria to change their mutation rates. We describe one such mechanism, driven by the action of a mobile element, in the marine bacterium Vibrio splendidus 12B01. We also found similar mobile genetic sequences in the mutS genes of many different bacteria, including clinical and agricultural pathogens. These mobile elements might play an as yet unknown role in the evolution of these important bacteria.

FIG 1

Serial selection for salt tolerance identifies a hypermutator phenotype, with a distinct mutation profile, in Vibrio splendidus 12B01. (a) We grew eight independent lineages on hypersaline media and sequenced genomes from each selection round. Strict SNP calling indicated that two hypermutator lineages had rapidly accumulated a large number of mutations, but all lineages had accumulated many more mutations (37 to 1,802) than the one or two mutations expected, given literature averages of spontaneous mutation (2). (b) The number of new mutations varied greatly across selection rounds in both mutator and hypermutator lineages. (c) This variability was much higher than the expected Poisson distribution variance, and this disparity was largely driven by transition mutations. (d) Selection rounds with more mutations tended to have larger ratios of transition versus transversion mutations; these ratios far exceeded averages from the literature (28), indicated by a dashed blue line, and the transition-to-transversion ratio of 12B01 compared with a closely related strain, V. splendidus 12F01.

FIG 2

Excision of a mobile element within mutS disrupts the mutS genetic sequence. (a) We identified a mobile element adjacent to mutS that was missing in both hypermutator lineages. Further inspection revealed that, when present, the mobile element appeared integrated within the mutS sequence, separating the original host-encoded mutS starting sequence. (b) When integrated, the mobile element provided a new start and upstream regulatory region to the mutS coding sequence. After excision, the mobile element left a 2-bp frameshift deletion in the host’s mutS sequence, resulting in a premature stop codon. (c) We designed a PCR assay to detect the excision of this mobile element. When the mobile element is integrated into the host mutS sequence, the left (attL) and right (attR) attachment site junctions of the mobile element and host genome are amplified. When the mobile element is excised, the rejoined host mutS gene (attB) and the circular excised mobile element (attP) are amplified. Expected amplicon lengths: attL = 819 bp, attR = 836 bp, attB = 613 bp, and attP = 1,042 bp. (d) We found that in rich medium (LB), the mobile element excised itself at low frequency. Sanger sequencing of PCR products attB and attP confirmed the 2-bp frameshift deletion in the host mutS sequence and the transfer of these base pairs to a circularized mobile element. In hypermutator lineages (e.g., lineage 2), we could no longer detect the mobile element, only the scarred host mutS sequence. (e) qPCR assays indicated that the frequency of excision was approximately 1/10,000 genomes, with moderately higher excision frequency during the stationary phase.

FIG 3

Mobile elements within mutS occur across Vibrio, Betaproteobacteria, and Gammaproteobacteria. (a) Phylogeny of close relatives of 12B01 (>98% similarity in 16S rRNA). Strains with mobile elements within mutS are not a monophyletic clade, and comparison between host phylogenies and mobile element phylogenies indicates that these elements have been horizontally transferred (dotted lines). (b) Broader BLAST searches identified other bacteria with mobile elements within mutS, including many human pathogens (Table S2). Phylogenetic tree built using 16S rRNA sequences. (c) Using a PCR assay similar to what we used for 12B01, we found mobile element excision in some, but not all, of a small subset of these bacteria when grown to the stationary phase. Sanger sequencing of the attB and attP PCR products from E. coli 536, P. putida F1, and B. multivorans CF2 indicated that mobile element excision in these bacteria did not result in any deletions or frameshift mutations. Expected amplicon lengths: E. norvegicus FF-162 attL = 894 bp, attR = 891 bp, attB (hypothetical) = 380 bp, and attP (hypothetical) = 1,405 bp; E. coli 536 attL = 404 bp, attR = 416 bp, attB = 393 bp, and attP = 427 bp; P. putida F1 attL = 601 bp, attR = 583 bp, attB = 596 bp, and attP = 588 bp; and B. multivorans CF2 attL = 811 bp, attR = 723 bp, attB = 451 bp, and attP = 1,083 bp. (d) These mobile elements were all integrated into the HTPMMQQ amino acid motif in MutS, although the precise location varied.