For this study, we selected all the S. aureus isolates presenting resistance towards ciprofloxacin received by the Bacteriology Laboratory of one of the largest hospitals in Portugal during a four months period. These corresponded to a collection of 52 S. aureus isolates.

Efflux activity amongst these 52 ciprofloxacin resistant isolates was assessed by means of a fast and practical test, the Ethidum Bromide-agar Cartwheel (EtBrCW) Method that provides information on the capacity of each isolate to extrude EtBr from the cells by efflux, on the basis of the fluorescence emitted by cultures swabbed in EtBr-containing agar plates. Those cultures showing fluorescence at lower EtBr concentrations have potentially less active efflux systems than those for which fluorescence is only detected at higher concentrations of EtBr 11 12 . The application of this method allowed the selection of 12 S. aureus isolates showing increased EtBr efflux activity when compared to the non-effluxing control strain ATCC25923 and to the efflux-positive control strain ATCC25923_EtBr 13 . These 12 isolates were designated EtBrCW-positive isolates, whereas the remaining 40 isolates were considered to have no or intermediate efflux activity and therefore designated as EtBrCW-negative isolates (Table 1).

Based upon these results, we continued the study by further analyzing the 12 EtBrCW-positive isolates, as well as a group of representative 13 EtBrCW-negative isolates, as controls.

In order to characterize the efflux activity of the cells, we used a semi-automated fluorometric method previously developed by our group 14 , which allows monitoring, on a real-time basis, the accumulation of EtBr inside the bacterial cells, followed by its efflux.

The first step of this technique is to establish the ideal conditions for EtBr accumulation inside the cells. Thus, assays were initially performed to determine the EtBr concentration above which there is detectable accumulation and to select the most effective efflux inhibitor; that is the EI that promotes the highest EtBr accumulation. The EtBr accumulation assays showed that the two groups of isolates previously established by the EtBrCW Method differed with respect to their capacity to accumulate EtBr, with EtBrCW-negative isolates retaining more EtBr than the EtBrCW-positive isolates (Figure 1-A). The same result was observed for the reference strain ATCC25923. These differences were reflected in the minimum EtBr concentration required for detectable accumulation, which was higher for the EtBrCW-positive isolates. The accumulation assays performed in the presence of several EIs showed that verapamil was the most effective in promoting accumulation of EtBr, for either EtBrCW-positive isolates, EtBrCW-negative isolates or the reference strain (Figure 1-B).

The conditions established by the accumulation assays were then used to load cells with EtBr and perform efflux assays. The assessment of EtBr efflux on a real-time basis (during a 10 min frame) detected a considerable difference between EtBrCW-positive isolates, which showed a pronounced efflux pump activity, with a prompt and significant decrease in fluorescence and the EtBrCW-negative isolates, that showed only basal efflux pump activity, similar to the one presented by the reference strain (Figure 1-C). These results confirm the presence of increased efflux activity in the EtBrCW-positive isolates relatively to the EtBrCW-negative isolates.

As expected, since all clinical isolates were selected on the basis of resistance to ciprofloxacin, they all presented high MIC values for fluoroquinolones. Nevertheless, the majority of the EtBrCW-positive isolates displayed higher MIC values for the fluoroquinolones tested and EtBr, whilst the EtBrCW-negative isolates presented significantly lower values, although some overlap exists between the two sets of MIC values (Table 1). The EIs reduced the MIC values for fluoroquinolones and EtBr of the EtBrCW-positive isolates to the values presented by the EtBrCW-negative isolates, confirming the presence of an active efflux component in those isolates (Table 1). The EIs thioridazine (TZ) and chlorpromazine (CPZ) were the most effective in reducing the MIC values. Verapamil (VER) and reserpine (RES) showed a smaller or absent inhibitory effect, while carbonyl cyanide m-chlorophenylhydrazone (CCCP) showed no effect on the MIC values for the compounds tested (data not shown). However, no full reversion of the fluoroquinolone resistance phenotype was obtained with any of the EIs tested, suggesting the contribution of other mechanisms to this resistance, namely, mutations in the target genes.

The 25 isolates representing both EtBrCW-positive and negative isolates were screened for the presence of chromosomal mutations most commonly associated with fluoroquinolone resistance in S. aureus, namely the ones occurring in the QRDRs of both grlA and gyrA genes 3 5 15 16 . All isolates tested carried mutations in grlA and gyrA related to fluoroquinolone resistance, in six different combinations at the protein level (Table 1). The majority of the isolates presented a double mutation in GrlA together with a single mutation in GyrA, with 12 isolates carrying the GrlA and GyrA mutations S80Y/E84K and S84L, respectively; three isolates carrying mutations GrlA S80F/E84K and GyrA S84L; and one isolate carrying mutations GrlA S80Y/E84G and GyrA S84L. The other nine isolates screened showed a single mutation in both GrlA and GyrA, in three distinct arrangements (Table 1).

The overall analysis of these results reveals a clear distinction between the EtBrCW-positive and the EtBrCW-negative isolates, with each group showing a relatively homogeneous profile, both in terms of efflux capacity and mutations in the genes related to fluoroquinolone resistance. In order to test if such homogeneity would be the result of clonal expansion of specific S. aureus clones, the isolates were then typed by macrorestriction analysis.

The clonality of the S. aureus clinical isolates was assessed by pulsed-field gel electrophoresis (PFGE) analysis of SmaI macrorestriction profiles. According to the criteria of Tenover et al 17 , six clones were found among the entire collection. The two predominant clones, A and E, included several sub-clones and comprised 25 and 18 isolates, respectively. The remaining clones B, C, D and F, were represented by 1 to 6 isolates (representative data is presented in Table 1 and Figure 2).

Of the 12 EtBrCW-positive isolates, 10 belonged to clone A, one to clone B and one to clone C. On the other hand, the 40 EtBrCW-negative isolates included all isolates from clone E (18 isolates) plus isolates from clone A (15), clone B (5), clones D and F (1 isolate each).

The presence of EP genes was assessed by PCR. All S. aureus isolates carried the five chromosomal genes tested (norA, norB, norC, mepA and mdeA) and one isolate, SM52, carried the plasmid encoded smr gene, whereas no isolate was found to carry the plasmid encoded qacA/B gene.

To assess the contribution of each individual pump to the overall efflux activity presented by each strain, ten isolates representative of each clone or sub-clone (six EtBrCW-positive and four EtBrCW-negative,) plus reference strain ATCC25923 (also EtBrCW-negative), were selected for expression analysis by RT-qPCR of EP genes.

When determining the expression level of efflux pump genes for the clinical isolates in comparison to the reference strain ATCC25923 in drug-free media, no overexpression of these genes was detected (data not shown). When these clinical isolates and ATCC25923 were exposed to an efflux pump substrate, either ciprofloxacin or EtBr, at ½ their MICs, and gene expression levels determined against the respective unexposed condition, overexpression of efflux pump genes was detected in six clinical isolates, three EtBrCW-negative and three EtBrCW-positive as well as in the reference strain itself (Table 2).

The majority of the isolates showed overexpression of a single efflux pump gene, most frequently, norB or mdeA. One isolate showed overexpression of two efflux pump genes (norB/norC) and another one overexpressed three EP genes (norB/norC/mepA). Overall, isolates showed to be more responsive to ciprofloxacin. The smr gene was found to be overexpressed only in the presence of EtBr, in accordance to the substrate specificity described in the literature for this pump 18 . These same agents had a distinct effect on ATCC25923, which showed significant overexpression of all efflux pump genes tested in the presence of EtBr, and a higher overexpression of norB when exposed to ciprofloxacin (Table 2).

The effect of drug exposure on the expression level of the efflux pump genes was further explored by increasing the ciprofloxacin concentration to ¾ the MIC. Isolates that showed EP gene overexpression with ½ the MIC of ciprofloxacin showed either an increase in that expression level or the overexpression of additional genes. For instance, EtBrCW-positive isolate SM50 overexpressing norB/norC with ½ MIC of ciprofloxacin, now showed even higher expression of norB (37.05 ± 18.67) and norC (83.98 ± 19.98) and de novo overexpression of norA (8.36 ± 4.63) and mepA (45.86 ± 13.86). Likewise, exposure of the EtBrCW-negative SM2 to higher ciprofloxacin concentrations resulted in increased levels of norB expression (4.48 ± 2.48) that was accompanied by de novo overexpression of norC (5.33 ± 0.73) and mepA (10.58 ± 0.73).

Strains were grown in tryptic soy broth (TSB) at 37°C with shaking or in tryptic soy agar (TSA) (Oxoid Ltd., Basingstoke, UK). Strain ATCC25923_EtBr was grown in TSB or TSA supplemented with 50 mg/L of EtBr. For determination of minimum inhibitory concentrations (MICs), cultures were grown in Mueller-Hinton broth (MH, Oxoid) at 37°C.

Antibiotics in powder form were purchased from different sources, as follows: nalidixic acid (Sigma-Aldrich, St. Louis, MO, USA); norfloxacin (ICN Biomedicals Inc., Ohio, USA); ciprofloxacin (Fluka Chemie GmbH, Buchs, Switzerland). EtBr was acquired in powder form from Sigma (Madrid, Spain).

Carbonyl cyanide m-chlorophenylhydrazone (CCCP), thioridazine (TZ), chlorpromazine (CPZ), verapamil (VER) and reserpine (RES) were purchased from Sigma. Solutions of TZ, CPZ and VER were prepared in desionized water; RES was prepared in dimethylsulfoxide (DMSO) and CCCP in 50% methanol (v/v). All solutions were prepared on the day of the experiment and kept protected from light.

This simple method tests the presence of active efflux systems 11 12 23 , being an update of the already described, EtBr-agar screening method 23 24 . It provides information on the capacity of each isolate to extrude EtBr from the cells by efflux pumps, on the basis of the fluorescence emitted by cultures swabbed in EtBr-containing agar plates. Briefly, each culture was swabbed onto TSA plates containing EtBr concentrations ranging from 0.5 to 2.5 mg/L (0.5 mg/L increments). S. aureus cultures ATCC25923 and ATCC25923_EtBr were used as negative and positive controls for efflux activity, respectively 13 . The plates were incubated at 37°C during 16 hours, after which the minimum concentration of EtBr associated with the bacterial mass that produced fluorescence under UV light was recorded in a Gel-Doc XR apparatus (Bio-Rad, Hercules, CA, USA). Isolates showing fluorescence at lower EtBr concentrations have potentially less active efflux systems than isolates for which fluorescence is only detected at higher concentrations of EtBr 11 12 23 24 . Isolates showing emission of fluorescence only at the maximum concentration of EtBr tested (2.5 mg/L) were considered to have potential active efflux systems.

This method allows the real-time fluorometric detection of the accumulation of a given efflux pump substrate (in this case, EtBr) inside cells and its efflux, using a Rotor-Gene 3000™ thermocycler, together with real-time analysis software (Corbett Research, Sydney, Australia) 14 27 28 . Accumulation assays allow to assess the EtBr concentration above which detectable EtBr accumulation occurs and to select the most effective efflux inhibitor; that is the EI that promotes the highest EtBr accumulation 14 . These conditions can then be used to load bacterial cells with EtBr and follow its efflux.

For the accumulation assays, the cultures were grown in TSB medium at 37°C with shaking until they reach an optical density at 600 nm (OD_{600 nm}) of 0.6. To prepare the cellular suspension, the cells were collected by centrifugation at 13, 000 rpm for 3 minutes and the pellet washed twice with a 1X Phosphate Buffered Saline (PBS) solution. The OD_{600 nm} of the cellular suspension was then adjusted to 0.6 in 1X PBS. To determine the EtBr concentration where there is detectable accumulation, several assays were prepared in 0.1 mL (final volume) containing 0.05 mL of the cellular suspension (final OD_{600 nm} of 0.3) and 0.05 mL of 2X EtBr stock solutions (final concentrations of 0.25, 0.5, 1, 2, 3, 4 and 5 mg/L). To determine the most effective EI, assays were prepared in a final volume of 0.1 mL containing 0.05 mL of the cellular suspension (final OD_{600 nm} 0.3) and 0.05 mL of a solution containing 2X the EtBr concentration previously selected and 2X the EI concentration to be tested (final concentrations of TZ: 12.5 mg/L, CPZ: 25 mg/L, VER: 200 mg/L, RES: 20 mg/L). All assays included control tubes containing only the isolate (0.05 mL of cellular suspension at OD_{600 nm} of 0.6 plus 0.05 mL of 1X PBS) and only the EtBr concentration to be tested (0.05 mL of 2X EtBr stock solution plus 0.05 mL of 1X PBS). The assays were then run in a Rotor-Gene 3000™ at 37°C, and the fluorescence of EtBr was measured (530/585 nm) at the end of every cycle of 60 seconds, for a total period of 60 minutes.

For the efflux assays, EtBr-loaded cells were prepared by incubating a cellular suspension with an OD_{600 nm} of 0.3 with either 0.25 or 1 mg/L EtBr for EtBrCW-negative or positive cultures, respectively and 200 mg/L VER at 25°C for 60 minutes. After EtBr accumulation, cells were collected by centrifugation and re-suspended in 1X PBS to an OD_{600 nm} of 0.6. Several parallel assays were then run in 0.1 mL final volume corresponding to 0.05 mL of the EtBr loaded cells (final OD_{600 nm} of 0.3) incubated with 0.05 mL of (1) PBS 1X only; (2) glucose 0.8% only (final concentration of 0.4%); (3) 2X VER only (final concentration of 200 mg/L); (4) glucose 0.8% (final concentration of 0.4%) plus 2X VER (final concentration of 200 mg/L). These efflux assays were conducted in the Rotor-Gene 3000™ at 37°C, and the fluorescence of EtBr was measured (530/585 nm) at the end of every cycle of 10 seconds, for a total period of 10 minutes. The raw data obtained was then normalized against data obtained from non-effluxing cells (cells from the control tube with only 200 mg/L VER), at each point, considering that these correspond to the maximum fluorescence values that can be obtained during the assay. The relative fluorescence thus corresponds to the ratio of fluorescence that remains per unit of time, relatively to the EtBr-loaded cells.

Isolates were typed by pulsed-field gel electrophoresis (PFGE) analysis, using well-established protocols. Briefly, agarose disks containing intact chromosomal DNA were prepared as previously described 29 and restricted with SmaI (New England Biolabs, Ipswich, MA, USA), according to the manufacturer's recommendations. Restriction fragments were then resolved by PFGE, which was carried out in a contour-clamped homogeneous electric field apparatus (CHEF-DRIII, Bio-Rad), as previously described 29 . Lambda ladder DNA (New England Biolabs) was used as molecular weight marker. PFGE types were defined according to the criteria of Tenover et al. 17 .

Genomic DNA was extracted with the QIAamp DNA Mini Kit (QIAGEN, Hilden, Germany), with an additional step of 30 minutes digestion with lysostaphin (Sigma) (200 mg/L) prior to extraction.

The QIAprep Spin Miniprep kit (QIAGEN) was used, with the following modification: prior to extraction, cells were incubated with lysostaphin (35 mg/L) at 37°C for 90 minutes, as previously described 30 .

Internal fragments comprising the QRDR of grlA and gyrA genes were amplified using the primers described in Table 3. The reaction mixture (50 μL) contained 2.5 U of Taq Polymerase (Fermentas Inc., Ontario, Canada), 1X Taq buffer (Fermentas); 25 pmol of each primer; 0.2 mM of dNTP and 1.75 mM of MgCl₂. The PCR reactions were conducted in a thermocycler Mastercycler personal 5332 (Eppendorf AG, Hamburg, Germany). The amplification conditions were as follows: DNA was denatured at 94°C for 4 minutes, followed by 35 cycles of denaturation at 94°C for 30 seconds, annealing at 50°C for 30 seconds and extension at 72°C for 1 minute, followed by a step of final extension at 72°C for 5 minutes. Amplification products were purified and sequenced in both strands using the same set of primers. Sequences were analyzed and aligned using the freeware programs BioEdit and ClustalW, respectively.

DNA fragments internal to five chromosomal and two plasmid encoded efflux pump genes were separately amplified by PCR, using the primers described in Table 3. Reaction mixtures were prepared as described above. Amplification conditions were as follows: DNA was denatured at 94°C for 4 minutes, followed by 35 cycles of denaturation at 94°C for 30 seconds, annealing at 45°C (norA) or 53°C (norB, norC, mdeA, mepA) for 30 seconds and extension at 72°C for 1 minute, followed by a step of final extension at 72°C for 5 minutes. The PCR reactions for genes qacA/B and smr were conducted under the following conditions: DNA was denatured at 95°C for 1 minute, followed by 30 cycles of denaturation at 95°C each for 1 minute, annealing at 40°C (qacA/B) or 48°C (smr) for 1 minute and extension at 72°C for 1 minute, followed by a step of final extension at 72°C for 5 minutes. The amplification products were visualized by 1% agarose gel electrophoresis.

For RNA extraction, strains were cultured in TSB media containing ciprofloxacin or EtBr, at ½ their MIC for each strain or in drug-free TSB, and grown until an OD_{600 nm} of 0.6. Total RNA was extracted with the RNeasy Mini Kit (QIAGEN), following the manufacturer's instructions. Before extraction of total RNA, cultures were treated with the RNAprotect bacterial reagent (QIAGEN). Contaminating DNA was removed with RNase-free DNase (QIAGEN) by a two hours on-column digestion at room temperature.

Quantitative RT-PCR (RT-qPCR) was performed using the QuantiTect SYBR Green RT-PCR Kit (QIAGEN). The primers used in these assays are described in Table 3. The relative quantity of mRNA corresponding to genes norA, norB, norC, mepA, mdeA and smr was determined by the comparative threshold cycle (C_T ) method 31 in a Rotor-Gene 3000™ thermocycler with real-time analysis software. Relative expression of the efflux pump genes was assessed by two approaches: (i) comparison of the relative quantity of the respective mRNA in the S. aureus isolates to the one present in a reference strain, ATCC25923; (ii) comparison of the relative quantity of the respective mRNA in the presence of ciprofloxacin or EtBr (at ½ the MIC) to the drug-free condition. For each strain, three assays were conducted, corresponding to three independent total RNA extractions. Negative controls and genomic DNA contamination controls were included. 16S rDNA was used as reference. Genes showing increased expression of at least four-fold, when compared to the drug-free condition, were considered to be overexpressed 10 .