The few studies on efflux among S. aureus clinical isolates use the decrease of antibiotic MICs in the presence of EIs, particularly reserpine, as indicative of efflux activity . This approach is laborious and dependent on the susceptibility of the efflux system(s) to reserpine, which varies considerably . More recently, Patel and colleagues have proposed the use of EtBr MICs to identify S. aureus effluxing strains . This approach has the advantage of assessing efflux activity using a broad range efflux pump substrate, EtBr, which is pumped out by most efflux systems described for S. aureus, and thus, is an useful marker for the detection of efflux pump activity [7, 12, 14, 20]. Nevertheless, it is still an indirect assessment of efflux activity.
In the present study, we have applied two methods for the direct assessment of efflux activity among a collection of 52 ciprofloxacin resistant S. aureus clinical isolates, both also based on EtBr efflux. We first applied the EtBr-agar Cartwheel Method to select isolates with increased efflux activity. The presence of increased efflux in the 12 isolates selected was supported by the data collected from the semi-automated fluorometric method, which demonstrated that EtBrCW-positive isolates had a higher efflux activity than the EtBrCW-negative isolates. Thus, both methods proved to be adequate to assay efflux activity in S. aureus cells. In particular, the EtBrCW method proved to be a valuable tool for the rapid screening of efflux pump activity, allowing its application to screen large collections of clinical isolates. Furthermore, the use of a broad range efflux pump substrate such as EtBr warrants its wider application as compared to the analysis of EIs effect on MIC values, which can be severely impaired by the susceptibility of each efflux system to the EI being used and for which the mechanism of action at the cellular level remains, in most cases, to be clarified.
In addition, the semi-automated fluorometric method also allowed the characterization of this efflux activity, in terms of maximal concentration of EtBr that the cells were able to extrude without observable accumulation over a 60 min period and susceptibility toward several EIs. The results obtained clearly showed a distinct capacity of the two groups of isolates to extrude EtBr from their cells, with the EtBrCW-positive isolates being able to handle higher EtBr concentrations with no detectable accumulation. It was also observed that for both groups of isolates, EtBr extrusion/accumulation was most affected by the EI verapamil.
The efflux assays further demonstrated the higher efflux capacity of the EtBrCW-positive isolates, with a pronounced decreased of EtBr fluorescence (80%) within a 2 min interval, whereas the EtBrCW-negative isolates showed a smaller decrease of EtBr fluorescence (40%) over a 10 min interval.
These results were then complemented with MIC determination in the presence of EIs, leading to the observation that the efflux-mediated resistance is an important component of the level of fluoroquinolone resistance. In fact, not only the 12 EtBrCW-positive isolates presented higher MIC values towards the several fluoroquinolones, also these MIC decreased to levels similar to those of the EtBrCW-negative isolates in the presence of TZ and CPZ, even for isolates sharing the same QRDR mutations (Table 1). Altogether, these data demonstrate that mutations in the QRDR of grlA and gyrA genes confer resistance up to a certain level (8-32 mg/L for ciprofloxacin), above which resistance is mainly efflux-driven. This implies that although the inhibition of the efflux component by EIs does not bring resistance down to the susceptibility level, it promotes a significant decrease in this resistance.
In the MIC assays TZ and CPZ were the two EIs with the highest effect, whereas in the fluorometric assay, EtBr extrusion/accumulation was most affected by verapamil. This should reflect differences in the mechanism of action of each molecule, as well as to the characteristics of each assay. We have recently observed the same type of results with isolates of Mycobacterium smegmatis . The absence of efflux inhibitory effect of CCCP at sub-MIC concentrations for S. aureus strains has been discussed in a previous study .
For the analysis of gene expression, we first compared our clinical isolates to a fully-antibiotic susceptible reference strain, S. aureus ATCC25923, following the rationale of previous studies, [10, 20, 22]. However, in contrast to these earlier studies, no EP gene was found to be overexpressed. Consequentially, we explored the effect of exposing the isolates to ½ the MIC of the antimicrobial compounds used previously as selective markers, ciprofloxacin and EtBr, using the isolates grown in a drug-free condition as a reference for determining the gene expression level. Using this approach, we were able to detect overexpression of EP genes, albeit at levels lower than the ranges described in literature [10, 20, 22]. These differences could, in some extent, reflect the different approaches used, including the use of a different reference strain for gene expression assays. Nevertheless, the different methodological approaches do not explain all the results and since EtBrCW-positive isolates showed a strong involvement of efflux in the resistance phenotype, the absence of high levels of efflux pump genes expression suggests that the isolates could be already primed to respond to these noxious compounds. Clinical isolates are under a constant pressure by antimicrobial compounds, such as antibiotics and biocides. Since the expression of efflux pumps provides the cell with the means to cope with these compounds, it could be expected that those clinical isolates already have in their cell membrane the necessary number of efflux pump proteins, thus, increases in efflux pump genes expression may have already taken place. Also, no significant differentiation could be established between EtBrCW-positive and EtBrCW-negative isolates at the level of individual EP gene expression (Table 2). On the other hand, ATCC25923, which showed only basal efflux activity on the fluorometric assay, responded to drug pressure in a completely different manner, showing a significant overexpression of all efflux pump genes tested in the presence of EtBr and the highest expression level of norB following exposure to ciprofloxacin (Table 2). The distinct behavior observed for the clinical isolates as compared to the antibiotic fully susceptible reference strain further support the hypothesis that the clinical strains are primed to efflux noxious substances.
Increasing the concentration of ciprofloxacin to ¾ of the MIC augmented the expression rate of the already overexpressed genes with the additional overexpression of other efflux pump genes. These results show a clear concentration level above which there is an inducement of expression of the same or additional efflux pump genes. This response could reflect the involvement of these genes in a global stress response regulon, or simply be the result of a substrate-responsive regulation. Future work should clarify this aspect.
A previous study described the predominance of norB overexpression among a collection of S. aureus bloodstream isolates. For this collection, when a single efflux pump gene was overexpressed, it corresponded mostly to norA, whereas norB and norC were prevalent when two or more efflux pump genes were overexpressed . In our work, amongst the clinical isolates that overexpressed efflux pump genes, four showed overexpression of a single gene, either norB, mdeA or mepA. Only two isolates showed overexpression of more than one efflux pump gene. Remarkably, norA was the only gene for which no overexpression was detected among the clinical isolates, suggesting that other efflux pumps can have a more relevant role in the resistance to fluoroquinolones and EtBr in S. aureus than the one attributed to date. Nevertheless, exposure of ATCC25923 to EtBr, resulted in the overexpression of all efflux pump genes tested, including norA. This result does not oppose to our previous finding that the prolonged exposure of this strain to increasing concentrations of EtBr resulted in high overexpression of solely norA , inasmuch as it strengthens the premise that exposure of the same strain to a given drug over different ranges of concentrations and/or time may result in the activation of different efflux systems. Our data also revealed that the same clinical isolate can respond differently at the gene expression level, to the presence of two inducers, ciprofloxacin and ethidium bromide, both common substrates of the main multidrug efflux pumps in S. aureus. For example, the EtBrCW-negative isolate SM2 exposed to ciprofloxacin showed only norB overexpression, whilst in the presence of EtBr, it overexpressed norB, norC and mepA. In the particular case of strain SM52, the plasmid encoded Smr pump was only overexpressed upon exposure to EtBr, whereas when challenged with ciprofloxacin, the strain responded with the overexpression of mepA. Our data also demonstrates that isolates from the same clone, as defined by PFGE, can have distinct levels of efflux activity and respond to the same agent through the activation of different efflux pumps (cf Tables 1 and 2).