Characterization of Staphylococcus aureus mutants expressing reduced susceptibility to common house-cleaners

Abstract

Aims

To characterize mutants of Staphylococcus aureus expressing reduced susceptibility to house cleaners (HC), assess the impact of the alternative sigma factor SigB on HC susceptibility, and determine the MIC of clinical methicillin-resistant S. aureus (MRSA) to a HC.

Methods and Results

Susceptibility to HC, HC components, H₂O₂, vancomycin and oxacillin and physiological parameters were determined for HC-reduced susceptibility (HC^RS) mutants, parent strain COL and COLsigB::kan. HC^RS mutants selected with three HC expressed reduced susceptibility to multiple HC, HC components, H₂O₂ and vancomycin. Two unique HC^RS mutants also lost the methicillin resistance determinant. In addition, all HC^RS mutants exhibited better growth at two temperatures, and one HC^RS mutant expressed reduced carotenoid production. COLsigB::kan demonstrated increased susceptibility to all HC and many HC components. sigB operon mutations were not detected in one HC^RS mutant background. Of 76 clinical MRSA, 20 exhibited reduced susceptibility to a HC.

Conclusions

HC^RS mutants demonstrate altered susceptibility to multiple antimicrobials. While sigB is required for full HC resistance, one HC^RS mechanism does not involve sigB operon mutations. Clinical MRSA expressing reduced susceptibility to a common HC were detected.

Significance and Impact of the Study

This study suggests that HC^RS mutants are not protected against, nor selected by, practical HC concentrations.

INTRODUCTION

A working definition for house cleaners (HC) employed in this investigation are formulations that can contain antimicrobial plant extracts (e.g. pine), surfactants, alcohols and other constituents. Some HC are advertised as disinfectants, while other HC formulations are not advertised as disinfectants and make effective household cleaning agents. Research has demonstrated that HC are effective in destroying bacterial pathogens (Rutala et al. 2000) and that household antibacterial use might reduce the number of pathogens lurking within domicile environments (Cole et al. 2003).

While the majority of Staphylococcus aureus infections occur in health care settings, this pathogen is also a cause of life-threatening disease in the community (Centers for Disease Control 1999; Martinez-Aguilar et al. 2003; Salgado et al. 2003). It has been estimated that one-third of the US population is colonized with S. aureus (Salgado et al. 2003) and household pets can also act as reservoirs for this pathogen (Manian 2003).

The S. aureus sigB operon encodes for the alternative transcription factor SigB, which is intimately involved with the general stress response of this organism (Wu et al. 1996; Kullik and Giachino 1997; Chan et al. 1998). The sigB operon also produces RsbU, RsbV, RsbW, which collectively regulate SigB. RsbW acts as an anti-sigma factor by interacting with SigB and blocking SigB interactions with the RNA polymerase core enzyme. Activated RsbV binds to RsbW, causing the release of functional SigB from RsbW. RsbU acts as a phosphatase that removes an inhibitory phosphate group from inactivated RsbV-Pi, thus activating its anti-RsbW activity (Wu et al. 1996; Kullik and Giachino 1997; Bischoff and Berger-Bächi 2001).

Mutants of S. aureus demonstrating reduced susceptibility to Pine-Sol® (PS; Clorox Company, Oakland, CA, USA) also demonstrate reduced susceptibility to the cell wall active antibiotics vancomycin and oxacillin (Price et al. 2002). PS-reduced susceptibility (PS^RS) and innate resistance to PS is dependent on the alternative sigma factor SigB (Price et al. 2002).

In an effort to further understand the effects of HC formulations on bacterial cells, spontaneous mutants of S. aureus demonstrating reduced susceptibility to one of four HC formulations were selected and phenotypically characterized. The applied goal of this study is to determine if in vitro selected HC-reduced susceptibility (HC^RS) mutants are resistant to practical HC concentrations and demonstrate reduced susceptibility to multiple antibacterials. In addition, the impact of sigB on the expression of HC^RS was further investigated and 76 clinical methicillin-resistant S. aureus (MRSA) strains were screened for PS^RS.

MATERIALS AND METHODS

Bacterial strains and growth conditions

The MRSA strain COL used for HC^RS selection, PS-reduced suceptibility (PS^RS) strains CP170 and CP171, and isogenic strain COLsigB::kan have been previously described (Price et al. 2002). Seventy-six clinical MRSA strains were collected from two El Paso medical centres from June to August 2002. All strain working stocks, including the HC^RS mutants produced in this study were maintained on Difco^™ Luria broth (LB) base (Becton Dickinson and Company, Sparks, MD, USA) agar (LBA) at 4°C or stored following growth in LB and glycerol addition (20% v/v final concentration) at −20 or −80°C. LB or LBA (Becton Dickinson) were used in all experiments unless otherwise noted. MIC/MBC determinations (minimum inhibitory concentration/minimum bactericidal concentration) employed Mueller Hinton broth (MHB) and Mueller Hinton agar (MHA) (Becton Dickinson). All liquid cultures were rotated at 200 rev min⁻¹ and most experiments were carried out at 37°C. Growth curves were performed in 20 ml LB cultures in 50 ml capped Erlenmeyer flasks inoculated with diluted overnight cultures (16 h) to reach an initial starting O.D.₅₈₀ of 0·01 for all cultures examined, and the O.D.₅₈₀ at 37 or 42 °C was then monitored over a 24-h period.

HC formulations and chemicals

House cleaners were randomly selected. Information describing the constituents of these HC is given on the container supplied by the manufacturers: Pine-Sol® contains 15% pine oil, isopropanol and cleaning agents called surfactants; Orange Clean® Super Concentrate (OC) (Orange Glo International Inc., Littleton, CO, USA) contains 30% cold pressed orange oil, 5–10% dodecylbenzenesulphonic acid triethanolamine and ingredients not precisely identified that are propriety or nonhazardous; Simple Green® (SG; Sunshine Makers Inc., Huntington Harbor, CA, USA) is a blend of water, surfactant (surface active) agents, wetting agents, emulsifiers, colour and scent; and Parsons® Lemon Ammonia (PA) (Procter & Gamble, Cincinnati, OH, USA) contains ammonium hydroxide solution, anionic surfactant, nonionic surfactant, perfume, colour clarifying agent and inert salts. All the formulations above were stored in their original containers at room temperature and single lots of each were used for all experimentation.

Pine oil (PO), tea tree oil (TTO) (Frontier Natural Brands Inc., Boulder, CO, USA) and H₂O₂ (Sigma Aldrich, St Louis, MO, USA) were obtained in amber glass bottles and maintained at 4°C. α-terpineol (Aldrich Chemical Company Inc., Milwaukee, WI, USA) was obtained in an amber glass bottle and maintained at room temperature. Vancomycin, oxacillin, isopropanol, ethanol and sodium dodecyl sulphate were all obtained from Sigma Aldrich. Stocks of vancomycin and oxacillin were made up in ddH₂O and stored at −20°C until required.

HC^RS spontaneous mutant isolation

One hundred microlitre aliquots of an overnight 20 ml LB COL culture grown in a 50 ml screw capped Erlenmeyer flask were inoculated onto LBA containing 0·8% PS, 0·15% OC, 10% SG, or 0·4% PA (v/v). After 24–72 h at 37°C, spontaneous HC^RS mutants appeared on the HC containing LBA at a mutation frequency of 10⁻⁶ to 10⁻⁸. All spontaneous HC^RS mutants were then passed through HC-free LB at least three times before characterization.

Antimicrobial susceptibility testing

For qualitative assessment of HC^RS expression, the gradient plate technique was utilized. HC gradient plates were prepared in square plates (90 mm × 90 mm) with LBA. Initially 40 ml of molten LBA (50°C) was poured into a square plate raised at one side by a 1 ml glass pipette (6 mm) and allowed to solidify overnight. The next day, 40 ml of molten LBA containing one of the HC being analysed (see Table 1 for gradient concentrations) was poured into a plate containing the first slanted layer of LBA lying flat, and allowed to solidify. All plates containing the continuous HC gradients were then dried open face upside down at 37°C for 2 h before being inoculated. Overnight LB cultures inoculated with single colonies were diluted to an O.D.₆₂₅ of 0·1 with LB and streaked three times in a single line onto freshly prepared gradient plates with sterile cotton swabs. Inoculated plates were incubated at 37°C and MIC were measured as the point at which confluent bacterial growth halted following 48-h incubation.

Table 1.

HC gradient plate results

Strain	Growth (mm) along HC gradients (n = 3 ± S.D.)
	0 → 0·7% v/v PS	0 → 0·15% v/v OC	0 → 13·0% v/v SG	0 → 0·45% v/v PA
COL	20 ± 0·6	16 ± 0	17 ± 0·6	26 ± 1·7
COLsigB::kan	0 ± 0 ↓	0 ± 0 ↓	0 ± 0 ↓	20 ± 0·1↓
PS12	51 ± 1 ↑	44 ± 0·6 ↑	32 ± 0·6 ↑	35 ± 0·1↑
OC11	26 ± 1·5 ↑	37 ± 0·6 ↑	>90 ± 0 ↑	25 ± 0·1 −
SG12	22 ± 1 −	40 ± 3·6 ↑	>90 ± 0 ↑	30 ± 0·1↑
PA11	0 ± 0 ↓	10 ± 1 ↓	0 ± 0 ↓	>90 ± 0 ↑

PS, Pine Sol®; OC, Orange Clean® Super Concentrate; SG, Simple Green®; PA, Parsons® Lemon Ammonia.

Arrows indicate increases (↑) or decreases (↓) in distances grown by mutant compared with COL; (−) indicates no difference in distances grown by mutant and COL.

MIC/MBC for HC, HC components and H₂O₂ were determined in 2 ml (final volume) of MHB in screw top test tubes (150 mm × 10 mm). Growth in all tubes was initiated by combining 1 ml of diluted overnight MHB cultures (final O.D.₆₂₅ = 0·01) with 1 ml of MHB containing the antibacterial analysed. MICs were determined after 24 h static incubation at 37°C. MBCs were determined by spotting 5 μl aliquots from the MIC and all tubes above the MIC (previously vortexed for 5 s to resuspend cells) onto MHA and scoring for visible growth following overnight incubation (24 h). MIC/MBCs for HC and HC components were determined utilizing: 0·09–0·33% v/v PS, in 0·03% increments; 0·005–0·05% v/v OC, in 0·005% increments; 6–20% v/v SG, in 2% increments; 0·3–1·2% v/v PA, in 0·1% increments; 1·5–6·0% v/v PO, in 0·5% increments; 0·025–0·25% v/v α-terpineol, in 0·025% increments and 0·025–0·25% v/v TTO, in 0·025% increments. MICs only were determined for sodium dodecylsulphate (SDS) (0·005–0·05% w/v SDS, in 0·005% increments) ethanol and isopropanol (3–12%, v/v in 1% increments) and hydrogen peroxide (0·3–3 mmol l⁻¹, in 0·3 mmol l⁻¹ increments).

Vancomycin and oxacillin resistance population analyses were performed essentially as described by Pfeltz et al. (2001). Staphylococcus aureus isolates were grown overnight in 20 ml of LB, diluted to an O.D.₅₈₀ of 1, and then serially diluted 10⁻¹ to 10⁻⁸ in sterile LB. Ten microlitre aliquots of all the dilutions were then pipetted at separate positions on the surface of LBA plates containing various concentrations of vancomycin or oxacillin, and allowed to absorb into the agar surface. Plates were then incubated for 48 h at 37°C and CFUs within the individual dilution positions were then scored. Kirby–Bauer disc susceptibility tests were performed as previously described (NCCLS 1997) using Remel (Lenexa, KS, USA) oxacillin discs.

Carotenoid production

Production of the orange carotenoids was measured essentially as described by Morikawa et al. (2001). Initially 5 ml of overnight cultures grown in screw capped test tubes were diluted (final O.D.₅₈₀ of 2) and 1 ml aliquots of the diluted cultures were harvested by centrifugation (16 000 × g, 3 min, 25°C). The resulting cell pellet was then resuspended in 450 μl of methanol and placed at 55°C for 3 min. Following heat treatment, the sample was pelleted and the O.D.₄₆₅ of the supernatant was determined. Data presented in results section is in triplicate (n = 3) with the standard deviation.

Detection of the methicillin resistance determinant, mecA

The primers mecA-1 (5′-TGC TAT CCA CCC TCA AAC AGG-3′) and mecA-2 (5′-AAC GTT GTA ACC ACC CCA AGA-3′) and PCR parameters for detection of a 280 bp internal mecA amplicon were performed as previously described (Ito et al. 2001). Chromosomal DNA used in the mecA-specific PCR reactions was isolated from 20 ml overnight LB cultures by glass rod spooling from lysostaphin (Sigma Aldrich) (10 mg l⁻¹ final concentration) treated cell lysates as previously described (Sambrook et al. 1989). Methicillin-susceptible strain BB255 (Price et al. 2002) was used as a mecA negative control.

Sequencing and comparison of sigB operon sequences

Initially, chromosomal DNA from overnight 5 ml tryptic soy broth cultures of COL, CP170 and CP171 were isolated using the Wizard genomic DNA isolation kit (Promega Corporation, Madison, WI, USA) according to the manufacturer’s instructions. Based on the S. aureus COLsigB operon sequence (NCBI Y09929), primers sigB-1 (5′-AGG CTT TAG GTA TTC CAT CCT A-3′) and sigB-8 (5′-ATC GCG ACA TTT ATG TGG AT-3′) were designed to amplify the entire sigB operon and 97 bp upstream of the rsbU start site and 62 bp downstream of the sigB stop codon (2833 bp total). sigB amplicons were generated with the GeneAmp amplification system (Applied Biosystems, Foster City, CA, USA) using Taq DNA polymerase (Promega) and then purified for sequencing using the QIAquick PCR purification kit (Qiagen, Valencia, CA, USA). Purified sigB amplicons were then sequenced using the ABI Prism BigDye v3·1 kit and the ABI Prism 310 genetic analyzer (Applied Biosystems). Comparison of sigB operon sequences was carried out using the ‘blast2sequence’ at http://www.ncbi.nlm.nih.gov/blast/bl2seq/bl2.html (Tatusova and Madden 1999).

Determination of PS MICs for clinical MRSA

Pine-Sol® MICs for the clinical MRSA, COL and PS12 were determined using an agar diffusion method utilizing MHA infused with 0·01% to 0·8% v/v PS, in 0·01% increments until 0·1%, then 0·1% increments until 0·8%. Inocula were prepared by diluting 3 ml overnight MHB cultures grown in screw capped test tubes to an O.D.₆₂₅ of 0·1 in sterile MHB and then diluting this culture an additional 10-fold in MHB. Two μl inocula of the diluted cultures were then carefully spotted (at times inocula were repelled by PS) and allowed to absorb into the surface of MHA plates containing different PS concentrations. PS MICs were scored after 24-h incubation at 37°C.

RESULTS

HC gradient plates, and HC and HC component MIC/MBCs

Mutants of COL demonstrating reduced susceptibility to PS (PS12), OC (OC11), SG (SG12) and PA (PA11) were successfully selected (Tables 1 and 2). Qualitatively compared with COL, all HC^RS mutants demonstrated reduced susceptibility to their respective mutant-selecting HC on continuous HC gradients (Table 1). With the exception of SG12 and OC11 on PS and PA gradients, respectively, PS12, OC11 and SG12 demonstrated reduced susceptibility to all HC investigated compared with COL. PA11 demonstrated reduced susceptibility to its mutant-selecting HC, PA, but increased susceptibility to PS, OC and SG, compared with COL. COLsigB::kan demonstrated reduced susceptibility to all HC investigated via the gradient plate technique compared with COL (Table 1).

Table 2.

MIC/MBC for HC, HC components and H₂O₂

Substance	Strain COL	COLsigB::kan	PS12	OC11	SG12	PA11
* PS
MIC	0·15	0·12 ↓	0·18 ↑	0·27 ↑	0·27 ↑	0·27 ↑
MBC	0·33	0·18 ↓	>0·33 ↑	0·33 −	>0·33 ↑	0·33 −
OC*
MIC	0·035	0·025 ↓	0·035 −	0·035 −	0·035 −	0·035 −
MBC	0·035	0·035 −	0·045 ↑	0·060 ↑	>0·065 ↑	0·065 ↑
SG*
MIC	12	6 ↓	18 ↑	20 ↑	>20 ↑	12 −
MBC	14	10 ↓	20 ↑	>20 ↑	>20 ↑	14 −
PA*
MIC	0·5	0·4 ↓	1·1 ↑	0·7 ↑	0·7 ↑	0·9 ↑
MBC	1	0·7 ↓	>1·2 ↑	>1·2 ↑	>1·2 ↑	>1·2 ↑
PO*
MIC	5·5	2 ↓	>6 ↑	6 ↑	>6 ↑	5 ↓
MBC	>6	>6 −	>6 −	>6 −	>6 −	>6 −
α–T*
MIC	0·150	0·125 ↓	0·150 −	0·125 ↓	0·100 ↓	0·100 ↓
MBC	0·200	0·125 ↓	0·250 ↑	0·250 ↑	0·250 ↑	0·225 ↑
TTO*
MIC	0·125	0·075 ↓	0·175 ↑	0·175 ↑	0·150 ↑	0·125 −
MBC	0·200	0·150 ↓	0·225 ↑	>0·250 ↑	0·250 ↑	0·250 ↑
SDS†
MIC	0·025	0·020 ↓	0·030 ↑	0·030 ↑	0·035 ↑	0·025 −
ETH*
MIC	7	7 −	10 ↑	10 ↑	10 ↑	9 ↑
ISO*
MIC	10	6 ↓	9 ↓	10 −	10 −	8 ↓
H2O2
MIC	0·6 mmol l−1	0·6 mmol l−1 −	1·2 mmol l−1 ↑	1·2 mmol l−1 ↑	1·2 mmol l−1 ↑	1·5 mmol l−1 ↑

PS, Pine Sol®; OC, Orange Clean® Super Concentrate; SG, Simple Green®; PA, Parsons® Lemon Ammonia; PO, pine oil; α-T, α-terpineol; TTO, tea tree oil; ETH, ethanol; ISO, isopropanol; SDS, sodium dodecylsulphate.

Arrows indicate increases (↑) or decreases (↓) in MIC/MBC expressed by mutant compared with COL; (−) indicates no difference between the MIC/MBC of mutant and COL.

^*%v/v.

^†%w/v.

For quantitative comparisons, MICs and MBCs to HC and HC components were determined for the strains investigated. In general, all HC^RS mutants demonstrated elevated MICs and/or MBCs to HC other than the mutant-selecting formulation, including PA11 (Table 2). For instance, PS12 demonstrated elevated MICs and MBCs to all HC compared with parent strain COL, except the OC MIC, which was the same for both strains. In contrast, PA11 did not demonstrate increased OC and SG MICs and PS and SG MBCs compared with COL. OC MICs for all HC^RS mutants were the same as parent strain COL; however OC MBCs for all HC^RS mutants was greater than COL.

MICs and MBCs for a random selection of potential HC components were also determined (Table 2). With the exception of PA11, the PO MICs were elevated in all HC^RS mutants. PO MBCs for all strains investigated was >6%. α-Terpineol is an antimicrobial terpene found in PO (Kartnig et al. 1991). All HC^RS mutants demonstrated similar or reduced α-terpineol MICs compared with COL, but increased MBCs to the substance. The MIC/MBCs to another antimicrobial terpene mixture, TTO (Cox et al. 2000), was also determined for all strains. All HC^RS mutant TTO MICs and MBCs were elevated compared with COL, except for the TTO MICs of PA11. SDS MICs were higher for all HC^RS mutants compared with COL, except PA11. Ethanol and H₂O₂ MICs were elevated for all HC^RS mutants while isopropanol MICs were the same or lower compared with COL. COLsigB::kan demonstrated reduced MIC/MBCs to all HC and HC components tested compared with COL, except for the OC MBC, and ethanol and H₂O₂ MICs which were the same as COL (Table 2).

Growth curves and carotenoid production

All HC^RS mutants investigated grew slightly faster at 37 and 42°C for the first 8 h, compared with COL (Fig. 1a,b), yet all strains reached a similar final growth density. COLsigB::kan grew faster than COL at 37 and 42°C for 8 h, yet the final density of COLsigB::kan obtained in both growth temperatures was less than COL at 24 h (Fig. 1a,b).

Fig. 1.

Twenty-four hour growth curves of strains at 37°C (a) and the heat shock temperature of 42°C (b). Strains are represented as follows: COL, open circle ○; COLsigB::kan, closed circle ●; PS12, open square □; OC11, open triangle △; SG12, open diamond ◇; PA11, cross +

Strains COL, OC11, SG12 and PA11 all had similar levels of carotenoid production (0·041 ± 0·005, 0·048 ± 0·004, 0·040 ± 0·001 and 0·039 ± 0·01, respectively), while PS12 (0·014 ± 0·006) demonstrated reduced carotenoid production compared with COL. Two additional PS^RS mutants (PS10 and PS11) also demonstrated reduced carotenoid production (0·027 ± 0·006 and 0·007 ± 0·005, respectively). COLsigB::kan demonstrated a dramatic reduction in carotenoid production (0·002 ± 0·0005) compared with isogenic strain COL.

Vancomycin and oxacillin resistance population analyses

All HC^RS mutants produced more colonies on increasing vancomycin concentrations compared with COL. While colonies of parent COL stopped appearing on 1·4 mg l⁻¹ vancomycin, colonies of HC^RS mutants continued to appear on higher vancomycin concentrations (1·6–2·0 μg l⁻¹) in a strain-dependent manner (Fig. 2a). OC11, SG12 and PA11 did not produce colonies on elevated oxacillin concentrations as well as COL or PS12 during oxacillin resistance population analysis (Fig. 2b). Both PS12 and COL produced colonies at a similar level on increasing oxacillin concentrations (Fig. 2b). COLsigB::kan did not produce as many colonies as COL on increasing concentrations of vancomycin or oxacillin (Fig. 2a,b).

Fig. 2.

Vancomycin resistance (a) and oxacillin resistance (b) population analyses. Strains are represented as follows: COL, open circle ○; COLsigB::kan, closed circle ●; PS12, open square □; OC11, open triangle △; SG12, open diamond ◇; PA11, cross +

Kirby–Bauer susceptibility testing revealed that COL, PS12 and PA11 were oxacillin-resistant, while methicillin-susceptible BB255, and SG12 and OC11 were susceptible to oxacillin. Furthermore, chromosomal DNA isolated from COL, PS12 and PA11 produced 280 bp mecA-specific amplicons following PCR with mecA-specific primers, while SG12, OC11 and negative control BB255 DNA did not (data not shown).

sigB operon comparison

No alterations in the entire 2833 bp sigB operon sequence was observed between our COL, and isogenic PS^RS mutants CP170 and CP171. We did however note two alterations in our strains compared with the COL sigB operon in the NCBI database. Compared with the NCBI COL sigB operon sequence, our COL, CP170 and CP171 all harbour two-point mutations (based on Y09929 positions, C¹⁴³² → T¹⁴³² transition and C²⁰⁵³ → A²⁰⁵³ transversion) that lead to an Arg²¹⁷ → Cys²¹⁷ mutation in RsbU, and an Ala⁵⁰ → Glu⁵⁰ mutation in RsbV.

PS MIC of El Paso MRSA

The PS MICs (v/v) for COL and PS12 using this agar diffusion method were 0·3% and 0·7% respectively. Therefore, clinical MRSA isolates with PS MIC ≥0·7% were considered to express reduced susceptibility to PS. The El Paso MRSA isolates demonstrated a diverse range of PS MICs from 0·01% to >0·8% (v/v). The PS MIC and number of strains exhibiting these MICs were: 0·01% (1), 0·06% (2), 0·1%, (2), 0·2% (5), 0·3% (9), 0·4% (15), 0·5% (6), 0·6% (16), 0·7% (7), 0·8% (12) and >0·8% (1). The majority of clinical strains (57) investigated expressed PS MICs greater than the MIC of COL. Twenty El Paso MRSA exhibited PS MIC of ≥0·7%.

DISCUSSION

Overall, HC^RS mutants PS12, OC11 and SG12 expressed reduced susceptibility to multiple HC, PO, α-terpineol, TTO, SDS, ethanol, H₂O₂ and vancomycin. The elevated HC, α-terpineol and TTO MBCs observed for the HC^RS mutants (including PA11) demonstrates that these mutants can resist the killing action of these antibacterial substances as well.

Compared with the other HC^RS mutants, PA11 did not demonstrate reduced susceptibility to PS, OC and SG on gradient plates, nor increased SG, PO, TTO and SDS MICs, or a SG MBC. PA contains ammonium hydroxide, which is known to be bactericidal (Scanlon and Quinn 2000). Perhaps the mutation(s) leading to the PA-selected HC^RS phenotype are in part selected by ammonium hydroxide, which is not listed as a component of the other HC investigated. PA11 did however demonstrate the highest H₂O₂ MIC of all strains investigated (Table 2). As the components within the HC formulations investigated are different, it is not surprising that the phenotypes of the HC^RS mutants selected differ.

The manufacturers recommended concentrations of the HC investigated for household use varies with regard to the cleaning strength required. For heavy cleaning, all the product bottles recommend use at full strength. For general disinfection with PS, the manufacturers recommend full strength, after removal of excess dirt. For ‘light duties/light cleaning/general cleaning’, the manufacturers of OC, SG and PA all recommend a working concentration of 3·0% v/v in water, while the producers of PS recommend a 1·6% v/v concentration.

In this study, OC demonstrated the strongest anti-staphylococcal activity, while SG displayed the weakest activity (Table 2). The PS, OC and PA MICs and MBCs of all strains investigated including HC^RS mutants, and the HC^RS mutant-selecting concentrations for PS (0·8%), OC (0·15%) and PA (0·4%) were all well below the general cleaning concentrations suggested for PS, OC and PA. Therefore, we speculate that PS-, OC- and PA-reduced susceptibility mutants or the acquisition of mutations by S. aureus leading to reduced susceptibility, probably do not pose any additional threat to households using these products. SG demonstrated the poorest anti-staphylococal activity with MICs ranging from 6 to >20% and MBCs ranging from 10 to >20%. It needs to be noted that OC, SG and PA are not advertised on their containers as disinfectants.

Staphylococcus aureus sigB mutants exhibit increased susceptibility to PS, the glycopeptide antibiotic vancomycin, the β-lactam antibiotic oxacillin, and H₂O₂ (Wu et al. 1996; Sieradzki and Tomasz 1998; Giachino et al. 2001; Morikawa et al. 2001; Price et al. 2002; Singh et al. 2003); lower growth yields (Giachino et al. 2001; Horsburgh et al. 2002); disabled heat shock response (Kullik and Giachino 1997; Chan et al. 1998) and reduced biosynthesis of carotenoids (Kullik et al. 1998). In addition, we now show that COLsigB::kan exhibits increased susceptibility to OC, SG and PA, PO, α-terpineol, TTO, SDS and isopropanol compared with COL. This evidence expands the role that sigB plays in the survival of S. aureus growing in the presence of antibacterials. As expected, COLsigB::kan also demonstrated increased susceptibility to vancomycin and oxacillin, reduced final growth densities at 37 and 42°C, and a reduction in carotenoid synthesis. Giachino et al. (2001) reported that sigB mutants demonstrated lowered H₂O₂ MBC, but had MIC similar to parent strains. As expected, the H₂O₂ MIC was the same for COLsigB::kan and COL in this study.

PS12, OC11 and SG12 exhibited reduced susceptibility to multiple HC, HC components, vancomycin and H₂O₂ and altered oxacillin resistance levels. In addition, compared with COL, most HC^RS mutants exhibit faster growth at 37 and 42°C. We also revealed a reduction in carotenoid synthesis in the PS^RS mutants isolated. As our HC^RS mutants expressed alterations in sigB-mediated phenotypes, especially the PS^RS mutants, we hypothesized that the spontaneous PS^RS mutants might harbour mutations within the sigB operon. Sequencing and comparison of the sigB operons of our COL and PS^RS mutants CP170 and CP171 proved this hypothesis to be false. Comparison of our COL, CP170 and CP171 to the NCBI COL sigB operon sequence did however reveal that our strains harbour two-point mutations altering the primary sequence of both RsbU and RsbV. It is unclear what effects these changes in RsbU and RsbV may have on SigB regulation, if any. Although PS^RS is SigB-dependent, low-level PS^RS mutants can still be selected from COL sigB::kan without altering the sigB::kan allele, indicating that loci outside of sigB can also mediate PS^RS (Price et al. 2002). Our laboratory is currently investigating if the sigB operon is upregulated in our HC^RS mutants.

Oxacillin resistance is mediated by mecA, which encodes the low-affinity penicillin binding protein PBP2a (Matsuhashi et al. 1986). The loss of oxacillin resistance and the mecA determinant in HC^RS mutants OC11 and SG12 was unexpected. In vitro selection of the vancomycin intermediate (vancomycin MIC ≥4 mg l⁻¹) phenotype in S. aureus can result in the loss of PBP2a expression because of mutation within mecA (Sieradzki and Tomasz 1999). As vancomycin susceptibility is reduced in HC^RS mutants (but not near the intermediate range), perhaps mutations leading to the vancomycin intermediate, and the OC- and SG-tolerant phenotypes occur in similar chromosomal loci that affect mecA stability.

Previously we reported that multiple PS-selected HC^RS mutants demonstrated elevated oxacillin resistance levels compared with COL using the gradient plate method to determine resistance levels (Price et al. 2002). In this study however, COL and PS12 survived at a similar level in oxacillin resistance population analysis. This discrepancy is the result of the fact that the gradient plate technique can detect small reductions in antibiotic susceptibility, because the organisms are exposed to a continuous gradient of the antibiotic being investigated. The antibiotic resistance population analysis used in this study however utilized twofold increases in antibiotic concentration and therefore is not as discriminatory in comparison of strain resistance levels as the gradient plate technique.

We were surprised to find such a diverse range of PS MICs (0·01% to >0·8% v/v) in the 76 clinical MRSA isolates investigated. Our results are similar to Cole et al. (2003) who reported a variable range of PO MICs (0·039–1·25% v/v) in 11 strains of S. aureus isolated from domicile environments. Our PO MICs (2% to >6% v/v) were much higher than those of Cole et al. (2003), but this is probably reflected in different MIC determination protocols and PO suppliers. Twenty of the MRSA investigated exhibited PS MICs of ≥0·7%, suggesting that reduced susceptibility to PS might be common among MRSA isolates. In fact, the vast majority of clinical strains (57) investigated, expressed PS MICs greater than COL (MIC = 0·3%). Further investigation is required to determine if our clinical isolates expressing elevated PS MIC also display cross-reduced susceptibility to other HC and HC components, or if they to can easily obtain higher HC^RS via single step selection.

Further studies are also required to determine the mechanism of HC^RS in S. aureus. As many of the components of HC formulations, including detergents, probably target the cell membrane (for review see, McDonnell and Russell 1999), we speculate that cell wall alterations are involved with this mechanism.