Use of Mutator Strains for Characterization of Novel Antimicrobial Agents

A recent minireview by Martinez and Baquero (8) provides a useful discussion on various aspects of mutational resistance to antibiotics in bacteria. As noted by these authors, bacteria displaying strong mutator (hypermutator) phenotypes exhibit significantly increased rates of mutation conferring antibiotic resistance (up to 10,000-fold that of the wild type). We would like to further highlight the value of such strains for basic antimicrobial drug discovery research, an aspect that was only partially addressed in the minireview.

Novel antimicrobial drug candidates are invariably evaluated with respect to the frequency with which resistant bacterial genotypes arise in vitro (3, 4, 7, 14). This provides an indication of whether resistance to the agent is likely to arise rapidly, either during therapy or within the environment. In addition, mutants recovered during such determinations may be important for elucidation of the drug's mode of action (4, 7, 12, 13) and for predicting the mechanism of resistance that may arise in the clinical setting.

We would like to stress the point touched upon in the minireview that hypermutators, e.g., Escherichia coli and Salmonella enterica with defects in the mismatch repair pathway (5, 9), should be used alongside wild-type isolates to examine the frequency with which drug resistance to a particular agent arises. This will yield mutation frequencies that represent worst-case scenarios. In turn this allows expression of the frequency of mutations conferring resistance as a range, not as a single value.

The rationale is that populations of pathogenic bacteria do not exhibit homogeneous mutation rates. For example, >1% of natural pathogenic E. coli and S. enterica populations exhibit a strong mutator phenotype (5). In addition, 0.0001 to 0.001% of some, and possibly all, bacterial populations are hypermutators (6), and a single selection event (e.g., antibiotic selection) can enrich the mutator population to 0.5% of the total (6). As Martinez and Baquero (8) point out, it is therefore erroneous to assume that a bacterial population exhibits uniform mutation rates. This could be particularly relevant during infection when in vivo mutation rates may be elevated (1).

Hypermutator strains may also be used to enhance the recovery of rare resistance mutations, e.g., for elucidation of modified drug targets within the cell. We have established that a fully grown 2YT or TB (11) culture of E. coli reaches cell densities of about 10¹⁰ CFU/ml (unpublished data). Resuspension of this culture in 1/10 the volume and incorporation of 1-ml aliquots in 10 agar pour plates allow mutants arising at frequencies approaching 10⁻¹² to be detected. Using E. coli hypermutators such as mutS or uvrD mutants, which exhibit 1,000-fold increases in mutation rate under certain conditions, allows detection of drug-resistant mutants that effectively occur at frequencies as low as 10⁻¹⁵. Indeed, we have used this approach to detect rare ampC promoter mutations in E. coli that confer increased ampicillin resistance (unpublished data).

There is little doubt that new antimicrobial agents are needed to combat the growing problem of antibiotic-resistant bacteria (2, 10). We suggest that hypermutator strains have an important role in the evaluation of such new agents.