Reply to “Lack of Evidence for Reduced Fitness of Clinical Staphylococcus aureus Isolates with Reduced Susceptibility to Triclosan”

REPLY

We welcome the opportunity to reply to the comments by Oggioni et al. (3) concerning our recent article (2), but we feel it necessary to point out that our article focused on small colony variants (SCV) and that at no point did we describe our strains as fabI mutants.

The authors cite a single article by Suller and Russell (5) to support their assertion that our data “are in discordance with previous reports.” However, the Suller and Russell paper did not present data for “virulence-associated phenotypes,” growth rate, biofilm formation, fabI mutation, or the expression of virulence factors. The highest MIC reported for their clinical isolates of S. aureus was 1.0 mg/liter, and there was no correlation between MIC values and kill rates in activity tests. It is therefore debatable whether any of their isolates were actually resistant, assuming that term can be applied meaningfully to biocides at concentrations considerably below use levels.

The recently published survey of S. aureus biocide susceptibility that was coauthored by Oggioni (1) does not mention SCV, present biofilm formation data, or consider virulence factors. The assertion that their in vitro-selected mutants showed none of the phenotypes we listed is therefore baseless, and in any case, direct comparisons of these two articles with our own cannot be made due to fundamental differences in focus and methodology. The method they used to select “multistep” mutants (serial passage in liquid medium) (1) is likely to select against SCV and, for that matter, any mutant with a markedly reduced planktonic growth rate, which is an important measure of fitness. This is a central theme of our recently published article (2). Moreover, assessing the true incidence of SCV is complicated by the fact that the “nicest” colonies are apparently commonly chosen when archiving clinical isolates.

We did not suggest, as has been inferred, that reduced susceptibility to triclosan renders bacteria avirulent; we suggested that the adoption of an SCV-like phenotype is not necessarily associated with increased virulence and that triclosan adaptation “may be associated with deficiencies in growth and virulence.”

The isolation of a mutant from a clinical sample provides no indication of its relative fitness, and we are not aware of any reliable means by which triclosan exposure and susceptibility can be correlated in clinical isolates. SCV form relatively small colonies due to metabolic impairment, and it seems obvious to us that the phenotypic implications of this and other forms of adaptation will vary considerably between strains. This is, in fact, illustrated by the data presented by Oggioni et al. (3) in which interstrain variation is apparent in the fabI mutants (see Fig. 1 in the comment letter). Furthermore, assuming the lowest possible virulence for ATCC 6538 (since the data lines are partially obscured), it appears that the death rates for the fabI mutants CR001 and D7 are significantly smaller than that for the mother strain (ATCC 6538).

Our paper partly addresses questions posed by Seaman et al. (4) regarding the fitness of SCV selected by triclosan exposure. As Oggioni et al. have noted, we tested one “mutant,” so the wider significance of our observations remains unclear. We did, however, demonstrate that a number of phenotypes linked to virulence and environmental survival were attenuated in our SCV strain.

The hypothesis that reduced susceptibility to triclosan does not affect fitness in clinical Staphylococcus aureus isolates apparently remains untested. The Galleria mellonella data presented by Oggioni et al. (3) serve only to illustrate differences in comparative larval virulence between (some) fabI mutants and the SCV that were the focus of our paper (2).