Response to the criticisms of Richard P. Novick in his review 'Autoinduction and signal transduction in the regulation of staphylococcal virulence' (Novick, 2003, Mol Microbiol 48: 1429-1449).

Abstract

R. Novick said that our article ‘The two-component system ArlS–ArlR is a regulator of virulence gene expression in Staphylococcus aureus’ (Fournier et al., 2001, Mol Microbiol 41: 247–261) should be regarded as tentative because of substantial ‘uncertainties’. In this article published in 2001, and not in 2000, as stated several times by R. Novick in his review, we described the effect of the arl mutation on the expression of various virulence factor genes. The ‘uncertainties’ mentioned by R. Novick relate to the first part of our article, which corresponds to only 10% of the work. At the start of the study, we used strain ISP794, which is a derivative of strain 8325 carrying the three prophages φ11, φ12 and φ13. In the original article, we concluded from this part of the work that strain 8325 was not suitable to study the effect of the Arl system on virulence gene expression. Indeed, the production of secreted virulence factors in the presence of the arlS mutation decreased in the 8325 background, whereas it increased in the 8325-4 background, corresponding to the 8325 strain cured of the three prophages. Thus, the production of virulence factors differs in the 8325 and 8325-4 backgrounds, because of either prophages or undefined mutations in the 8325 background. We decided to study the action of the Arl system in the 8325-4 background. Thus, the effects of the Arl system on virulence factor expression, corresponding to the bulk of our results, were studied in the 8325-4 background, and the results obtained are therefore not tentative. (i) Comment: β-haemolysin activities are reported for a φ13 lysogen, although hlb is insertionally inactivated by the prophage. Response: strain 8325 carries the φ13 lysogen in the β-haemolysin gene. Table 2 should be corrected as follows: β-haemolysin activities were not determined whereas α-toxin activities were 69 ± 4 haemolytic units for strain BF28 (8325) and 6 ± 2 haemolytic units for strain BF29 (8325 arlS –). (ii) Comment: β-lactamase activities are reported for strains of the NCTC8325 series, which are β-lactamase negative. Response: strain 8325 does not carry a β-lactamase gene. However, to make it possible to measure β-lactamase activity, we introduced a plasmid, pI258, encoding a β-lactamase. This is indicated in the Experimental procedures of our article (p. 258, first paragraph of secreted proteins). (iii) Comment: quantitative β-lactamase activities were determined by measuring minimum inhibitory concentrations (MICs). This is invalid owing to the profound inoculum effect seen with β-lactamase. Response: we used β-lactamase activity as a control for secreted protein because β-lactamase is not a virulence factor. We previously determined MICs because it had been reported that MICs determined for similar amounts of inoculum are correlated with the amount of β-lactamase produced by strains (Chapman and Steigbigel, 1983, J Infect Dis 147: 1078–1089). We investigated the correlation between MICs and the amount of β-lactamase in more detail by determining MICs for various inocula and β-lactamase activities in the culture supernatant, using nitrocefin, as described previously (Fournier et al., 2000, J Bacteriol 182: 664–671). At 105 cfu, MICs were well correlated with β-lactamase activities for strains BF23 (8325-4), BF24 (8325-4 arlS –) and BF29 (8325 arlS –) but not for strain BF28 (8325). At 104 cfu, MIC and β-lactamase activity were well correlated for all tested strains. MIC and β-lactamase activity of the wild-type strain BF23 were similar to those of its arlS mutant BF24 in the 8325-4 background. MIC and β-lactamase activity of the wild-type strain BF28 were also similar to those of its arlS mutant BF29 in the 8325 background. As we demonstrated in our article and confirm here by means of the previous results, virulence factor production such as α-toxin was modified by the arlS mutation. The production of extracellular β-lactamase, which is not a virulence factor, was not affected by the arlS mutation in either 8325 strains carrying the three prophages or 8325-4 strains cured of the three prophages. Thus, only the production of virulence factors was impaired in the absence of arlS. (iv) Comment: the authors of this paper have been unwilling to make their strains available to other investigators. Response: when R. Novick was writing his review (and more particularly when the review was accepted on 18 February, 2003), we were collaborating together: our team was constructing constitutive mutants of the Arl system that R. Novick's team were then to test in an arl mutant background. We were waiting until our part of the work was completed before sending him all the mutants.