Lack of Evidence for Involvement of Hypermutability in Emergence of Vancomycin-Intermediate Staphylococcus aureus

In a recent publication, Schaaff et al. demonstrate that selection of laboratory mutants resistant to vancomycin occurs more readily in a mutS knockout mutant of Staphylococcus aureus than in the parental strain from which it was derived (10). On the basis of these findings, the authors speculate that the emergence of vancomycin-intermediately resistant S. aureus (VISA) may also occur in clinical isolates that possess elevated mutation frequencies. Their arguments, in part, are also based on a report that the naturally occurring VISA strain Mu50 apparently contains a frame shift in mutS that would potentially confer a mutator phenotype (1). During the period that the manuscript by Schaaff et al. (10) was in preparation, we published a paper that does not support a role for mutators in the emergence of VISA (8).

Strains exhibiting elevated mutation frequencies have recently been reported among natural populations of pathogenic Escherichia coli, Salmonella enterica, Pseudomonas aeruginosa, Neisseria meningitidis, and Helicobacter pylori (6, 8). The majority of naturally occurring mutators contain defects in the methyl-directed mismatch repair (MMR) system, with mutations in mutS predominating (4, 7). MMR-deficient strains possess superior genetic backgrounds for the selection of some antibiotic resistance mutations, since mutation frequencies up to 1,000-fold higher than that of normal strains have been reported, and resistance levels achieved in mutators can be greater than those arising in nonmutator hosts (6, 7).

Intermediate resistance to vancomycin in S. aureus probably involves sequential mutations in a number of genes affecting peptidoglycan synthesis (1, 2). Since mutations in mutS are responsible for hypermutable phenotypes in other species, it was not unreasonable for Schaaff et al. (10) to propose that VISA strains might have emerged in staphylococcal mutators deficient in MutS activity. However, in our opinion, evidence for this theory is lacking.

Using an approach similar to that described by Schaaff et al. (10), we created a derivative of S. aureus RN4220 disrupted in mutS that displayed elevated mutation frequencies for resistance to a number of antibiotics (8). We agree that such a phenotype has the potential to accelerate evolution in S. aureus for adaptation to selective processes (in this case, antibiotics). Our work and that of Schaaff et al. (10) therefore indicate that if MutS-deficient strains arose in the clinical setting, they could contribute to the emergence of antibiotic resistance, including intermediate-level vancomycin resistance.

However, on the basis of quantitative mutation frequency determinations, none of the seven VISA strains we have examined, including Mu50, demonstrate any phenotypic evidence for mutator status (8, 9), and there is currently no experimental evidence to suggest that they have ever exhibited such a phenotype. Furthermore, resequencing of a portion of mutS from Mu50 indicates that this gene is intact (8).

In support of their theory for involvement of mutators in emergence of VISA, Schaaff et al. (10) also note that vancomycin-resistant mutants selected in S. aureus RN4220ΔmutS exhibit higher levels of resistance to the antibiotic (MIC, 32 μg/ml) than mutants derived from the parental strain RN4220 (MIC, 4 μg/ml). However, resistant mutants for which vancomycin has high MICs (up to 100 μg/ml) have been selected in the laboratory from other S. aureus strains (3, 11). Since there is no reason to believe that these strains are mutators, it is apparent that mutator status is unlikely to be a requirement for the generation of high-level vancomycin resistance in S. aureus.

Whether VISA strains have passed through a hypermutable state during which increased accumulation of mutations has enabled strains to circumvent growth inhibition by vancomycin is unknown. However, our analysis of several VISA strains has demonstrated that none currently exhibits a mutator phenotype. Furthermore, the apparent frame shift in the mutS gene of VISA strain Mu50 is a sequencing artifact, and the mutation frequency of this strain is that of the wild type. In addition, other multiply resistant S. aureus strains we have tested show no increases (8) in their mutation frequencies relative to those of antibiotic-sensitive isolates or laboratory strains. Mutation frequencies in all the naturally occurring isolates that we examined, including strains recovered from cystic fibrosis patients, were significantly lower than those observed in a strain in which mutS had been disrupted (8). Our finding that no mutators were present in nearly 500 S. aureus isolates, including many methicillin-resistant S. aureus (MRSA) strains, argues against the authors' suggestion that mutator clones may be highly prevalent among MRSA.