Correspondence: Gregory G. Anderson, Department of Biology, Indiana University Purdue University Indianapolis, 723 West Michigan St., SL 320, Indianapolis, IN 46202, USA. Tel.: 317 278 3896; fax: 317 274 2846; e-mail: firstname.lastname@example.org
Staphylococcus aureus forms pathogenic biofilms. Previous studies have indicated that ethanol supplementation during S. aureus biofilm formation results in increased biofilm formation and changes in gene expression. However, the impact of alcohols on preformed S. aureus biofilms has not been studied. In this study, we formed S. aureus biofilms on PVC plastic plates and then treated these preformed biofilms with five different alcohols. We observed that alcohol treatment of preformed S. aureus biofilms led to significant increases in biofilm levels after 24 h of treatment. Many bacteria within these biofilms were found to be alive and metabolically active. Alcohol treatment also resulted in increased transcription of the biofilm-promoting genes icaA and icaD, as well as several antibiotic resistance genes. These results demonstrate that treatment of S. aureus preformed biofilms with alcohols enhances biofilm levels if maintained for extended periods. Thus, alcohols might be of limited usefulness for the eradication of preformed S. aureus biofilms.
Staphylococcus aureus, a Gram-positive bacterium, is both a human commensal organism, found in the nose, ears, and other body locations (Chiller et al., 2001), and an opportunistic pathogen. Staphylococcus aureus is implicated in many disease states, including skin infections, septic arthritis, meningitis, endocarditis, septicemia, toxic shock syndrome, and medical device-related infections (Lowy, 1998; Chiller et al., 2001; Foster, 2005; Iwatsuki et al., 2006). The ability of S. aureus to form self-associated communities called biofilms can enhance the persistence of the microorganism at infection sites or on medical devices (Gotz, 2002). Furthermore, the formation of S. aureus biofilms on or in the human host indicates a switch from commensalism to a more pathogenic state (Gotz, 2002). Biofilms can also serve as a reservoir of pathogens for maintaining chronic infection (Gotz, 2002).
Unfortunately, bacteria within biofilms display increased tolerance to antimicrobial substances. For instance, biofilms can be 10–1000 times more tolerant to antibiotics (Mah & O'Toole, 2001). Likewise, biofilm bacteria can resist the antimicrobial activity of toxic chemicals and can also resist the clearance mechanisms of the immune system (Donlan & Costerton, 2002). Thus, it is difficult to eradicate bacteria within biofilms.
Alcohols have widely been used as antimicrobial substances for sanitizing or sterilizing surfaces. Ethanol and isopropanol, in particular, have been used for disinfection in medical and home settings. It is generally thought that alcohols act by disrupting membranes and denaturing proteins (McDonnell & Russell, 1999; Gilbert & McBain, 2003). These actions could lead to subsequent perturbations of cellular metabolism, and death (McDonnell & Russell, 1999). We were interested in determining whether ethanol, isopropanol, and other alcohols could eradicate S. aureus growing within biofilms. Intriguingly, several studies have suggested that treatment with low concentrations of alcohols can enhance biofilm formation by Staphylococcus species (Knobloch et al., 2002; Kogan et al., 2006; Korem et al., 2007; O'Gara, 2007; Milisavljevic et al., 2008). For example, during initial biofilm formation, alcohol supplementation increases Staphylococcus epidermidis biofilm production, in part by the expression of the ica genes for polysaccharide adhesin production (Knobloch et al., 2002). Furthermore, using microarrays, Korem et al. found enhanced expression of ica and other biofilm-promoting genes after ethanol treatment (2.4% w/v) of S. aureus during growth in early exponential phase (Korem et al., 2010). These changes correlated with significant increases in initial biofilm formation (Korem et al., 2010).
The preceding reports demonstrated that certain alcohols, at low concentrations, can enhance the initial formation of staphylococcal biofilms, when many of the bacteria were likely in a planktonic, free-swimming state. However, bacteria within preformed, mature biofilms are in a physiologically and phenotypically distinct state (Costerton et al., 1995), and it is unclear whether preformed S. aureus biofilms also display unique responses to alcohol treatments. Furthermore, it is possible that much higher alcohol concentrations could overcome biofilm-enhancement signaling and destroy biofilms. To investigate these issues, we formed biofilms with three different S. aureus strains and then measured biofilm levels after further treatment with five different alcohols.
Materials and methods
For this study, we used S. aureus strains Newman (Baba et al., 2008), RN6390 (Peng et al., 1988), and MZ100 (Shanks et al., 2005). These strains are laboratory strains that have been used in previous biofilm studies (Shanks et al., 2005; Johnson et al., 2008; Trotonda et al., 2008). The genome sequence of strain Newman has been published (Baba et al., 2008). Bacteria were grown in LB using standard microbiology techniques.
To form biofilms, S. aureus strains were diluted 1 : 100 in LB from overnight-shaken cultures and dispensed in 100 μL volumes into the wells of a polyvinylchloride (PVC) 96-well microtiter plate (BD Falcon #35391). Bacteria were incubated in these microtiter plates statically at 37 °C for 24 h, as previously described (O'Toole & Kolter, 1998; Shanks et al., 2005). These ‘preformed’ biofilms were then washed 2–3 times with 150–200 μL phosphate-buffered saline (PBS), and then 100 μL of LB or the indicated alcohols was added per well. We tested the effect of ethanol, methanol, isopropanol, isoamyl alcohol, and n-butanol because of their practical use and different chemical structures. Alcohols were used at 100% strength or diluted in sterile deionized water to the indicated percentage (v/v). After alcohol addition, the plates were incubated at 37 °C for an additional 4 h or 24 h, as indicated. These times were chosen to measure the effects of short-term and long-term alcohol exposure, respectively. After alcohol treatment, biofilms were assessed by washing the plates twice in water, staining for 10 min in 0.1% crystal violet, and then washing four times in water. Biofilms were evident as purple-stained films on the bottom of wells. Quantification of biofilms was carried out by dissolving the stain in 30% acetic acid and measuring OD540 nm (Conover et al., 2011) in a SpectraMax M2 spectrophotometer, as previously described (Shanks et al., 2005). Reported data are representative of three independent experiments performed with triplicate or quadruplicate samples.
Live/dead staining and microscopy
For fluorescent differentiation of live and dead bacteria within treated biofilms, preformed biofilms were treated as indicated and then washed twice with 150–200 μL PBS. 3 μL of a 1 : 1 mixture of SYTO®9 and propidium iodide fluorescent dyes from the BacLight Viability Kit (Invitrogen, Carlsbad, CA) was added to 1 mL PBS. 100 μL of this solution was added to each PBS-washed biofilm well, and the plates were incubated for 15 min in the dark at room temperature. Bacteria were imaged using an EVOSfl epifluorescence microscope. Images presented are representative of three independent experiments.
Measuring metabolic activity with alamarBlue®
After alcohol treatment of preformed biofilms, metabolic activity of biofilm bacteria was measured by alamarBlue® (AdB Serotech, Raleigh, NC) (Pettit et al., 2005). We used Greiner Bio-One sterile 96-well microtiter plates (#655161) for these experiments because the bottoms of the wells provide optical clarity necessary for spectrophotometric readings. Treated biofilms were washed twice with 150–200 μL PBS. 100 μL alamarBlue®, diluted 1 : 10 in LB, was added per well, and the plates were incubated at 37 °C for approximately 2 h. Calculation of biofilm metabolic activity as a percentage of metabolic activity in untreated biofilms was made by measuring OD570 nm and OD600 nm in a SpectraMax M2 spectrophotometer, using LB as a blank, and applying the following calculation: 100*([117 216*OD570 nm]−[80 586*OD600 nm])/([117 216*OD570 nm untreated sample]]−[80 586*OD600 nm of untreated sample]), according to the manufacturer's protocol. Reported data are representative of three independent experiments performed with quadruplicate samples.
Changes in the transcription of S. aureus genes known to be involved in biofilm formation were evaluated by semi-quantitative RT-PCR. In order to isolate sufficient RNA for analysis, biofilms were grown in 6-well plates (BD Falcon #353046). Preformed biofilms were treated with each alcohol at 100% strength for 24 h, after which RNA was isolated using the RNeasy Mini protocol (Qiagen, Valencia), and cDNA was generated from RNA using SuperScript III First-Strand Synthesis System (Invitrogen)(Anderson et al., 2008). Gene transcription was detected by PCR with the following primers: IcaAF 5′ TATGAACCGCTTGCCATGTG 3′, IcaAR 5′ TCACGCGTTGCTTCCAAAG 3′, IcaDF 5′ GGGTGGATCCTTAGTGTTACAATTTT 3′, IcaDR 5′ TGACTTTTTGGTAATTCAAGGTTGTC 3′, GyrBF 5′ ATCGGTGGCGACTTTGATCTA 3′, and GyrBR 5′ CCACATCGGCATCAGTCATAA 3′ (Cassat et al., 2006; Korem et al., 2010). Results were compared to levels of the gyrB transcript, which has previously been used as a control transcript for RT-PCR analyses of S. aureus transcriptional regulation following ethanol supplementation (Cassat et al., 2006; Korem et al., 2010). Reported data are representative of three independent experiments. For quantitative analysis of gene transcription, we performed qRT-PCR using an Applied Biosystems 7300 Real Time PCR System, as we have previously described (Anderson et al., 2010). In addition to the primers mentioned above, we also assayed for antibiotic resistance genes using the following primers: MmplRTFor 5′ GGAATGACATCTACAGAAGTAGGC 3′, MmpLRTRev 5′ AACTGCTAGTCCAATCATTACGG 3′, VraSRTFor 5′ AGTGCCGATGAAAGTTGTGC 3′, VraSRTRev 5′ TTTTGTACCGTTTGAATGACG 3′, MepARTFor 5′ AAGCCATTGTTGAGATGGCGAC 3′, MepARTRev 5′ AATTGTAGCACCACGACCTTGC 3′. Values were normalized with transcript levels of gyrB, and thus, graphs depict relative gene expression levels.
We performed Student's t-tests to detect statistical differences between the mean biofilm levels of samples (Walker et al., 2005; Kadouri et al., 2007) and to compare mean expression levels in qRT-PCR experiments. P values < 0.05 were considered significant.
Alcohol treatment enhances the level of preformed biofilms
We preformed biofilms of S. aureus strains MZ100, RN6390, and Newman in 96-well plates at 37 °C for 24 h, as described in 'Materials and methods'. These biofilms were treated with ethanol, methanol, isopropanol, isoamyl alcohol, and n-butanol for 24 h at 37 °C. Each alcohol was employed at five different percentage levels (20, 40, 60, 80, and 100% v/v) and compared to an untreated control. We found that the alcohol treatments were unable to fully eradicate preformed biofilms of strain Newman (Supporting Information, Fig. S1). Similarly, alcohol treatment failed to destroy biofilms of strains RN6390 and MZ100 (data not shown). We had originally used LB incubation as our untreated control conditions. We figured that LB would give the biofilm bacteria the greatest growth advantage for comparisons to alcohol-treated samples. Indeed, we found that 24-h incubation of preformed S. aureus biofilms with sterile water (which was used as the diluent for alcohols) resulted in significantly less biofilm compared to LB incubation, or similar biofilm levels in the case of the MZ100 strain (Fig. S2).
Quantification of biofilm levels after alcohol treatment revealed significantly greater staining after the treatment with 40% or greater concentrations of ethanol (Fig. 1). We observed a similar pattern after treatment of preformed biofilms with different concentrations of methanol, isopropanol, isoamyl alcohol, and n-butanol (Fig. S3).
We considered the possibility that biofilm levels were enhanced immediately after alcohol treatment and that further treatment destroyed the bacteria while leaving the accumulated biofilm biomass intact. In this scenario, enhanced biofilm would be evident after 24 h of alcohol treatment, but the bacteria would be nonviable. To investigate this hypothesis, we examined biofilm levels 4 h after alcohol treatment of S. aureus preformed biofilms. Because 100% alcohol had the greatest effect in our assays, we used this concentration for subsequent experiments. Whereas 24-h incubation led to significant biofilm enhancement for all 100% alcohol treatments (Fig. 1), 4-h incubation resulted in biofilm enhancement for only a portion of the alcohol treatments (Fig. 2). All strains displayed greater biofilm after 4-h isoamyl alcohol and n-butanol treatment, but other alcohols exerted varying effects (Fig. 2).
Alcohol treatment results in live and metabolically active bacteria
We used a Live/Dead fluorescent stain to determine whether any bacteria were alive in alcohol-treated or air-dried biofilms. In water-incubated controls for all three S. aureus strains, we found numerous live (green) and dead (red) bacteria at both 4 h and 24 h of incubation (Fig. 3, top row). For alcohol treatment, we chose to narrow our focus for Live/Dead staining to ethanol. All of the alcohols we have tested resulted in the same effects on biofilm formation (Fig. 1), and ethanol is arguably the most medically relevant alcohol used herein. After ethanol treatment, we observed many dead and live bacteria, but there appeared to be a greater number of live bacteria by 24 h, compared to 4-h ethanol treatment (Fig. 3, bottom rows). We also examined the viability of bacteria within alcohol-treated biofilms with alamarBlue®, which measures metabolic activity (Pettit et al., 2005). After 4 h, there was a general decrease in metabolic activity of alcohol-treated biofilms of strains RN6390 and Newman, compared with water-incubated controls, while biofilms of strain MZ100 displayed insignificant decreases (Fig. 4a). After 24 h of alcohol treatment, bacterial metabolism of treated biofilms had increased to the level of controls, and in many cases significantly greater metabolism was evident, compared to controls (Fig. 4b). The increase in bacterial metabolism of treated biofilms from 4 h of alcohol treatment to 24 h was even more pronounced when compared to controls incubated with LB (Fig. S4).
Alcohol treatment increases transcription of biofilm genes
To investigate the molecular basis of biofilm enhancement after alcohol treatment of preformed S. aureus biofilms, we utilized quantitative real-time PCR (qRT-PCR) to examine the changes in expression level of icaA and icaD. These genes assist in the production of the polysaccharide intercellular adhesin (PIA) and are thus vitally important for S. aureus biofilm formation (Archer et al., 2011). We found that ethanol treatment resulted in significantly greater transcript levels of both icaA and icaD, compared to LB treatment alone, in preformed biofilms of strain Newman (Fig. 5) as well as strains MZ100 and RN6390 (Fig. S5a–b). Furthermore, a similar pattern of expression was observed in biofilms of all three strains after the treatment with methanol, isopropanol, isoamyl alcohol, and n-butanol (Fig. S5c).
Alcohol treatment increases transcription of antibiotic resistance genes
We also investigated the effect of ethanol on transcription of genes associated with S. aureus antibiotic resistance, specifically the putative multidrug efflux pump gene mmpL (Cuaron et al., 2012), the efflux pump gene mepA (Kaatz et al., 2006), and the sensor histidine kinase gene vraS (Boyle-Vavra et al., 2006). These genes displayed significantly increased transcription in ethanol-treated preformed biofilms of S. aureus strains Newman (Fig. 6–a–c) and RN6390 (Fig. 6c), while in strain MZ100, there was increased transcription of mmpL and vraS only (Fig. S6a). To further explore this apparent increase in antibiotic gene expression, we compared biofilm formation in LB, in LB with 200 μg mL−1 carbenicillin, and in ethanol with 200 μg mL−1 carbenicillin. We found significantly greater Newman and MZ100 biofilm formation after treatment with ethanol and carbenicillin, compared to carbenicillin treatment alone (Figs 6d and S6b). Strain RN6390 showed a similar trend, but this difference was not statistically significant (Fig. S6d).
We found that alcohol treatment of preformed S. aureus biofilms resulted in enhanced biofilm formation (Fig. 1). While previous studies observed increased initial biofilm formation after ethanol or isopropanol treatment of planktonic bacteria (Knobloch et al., 2002; Kogan et al., 2006; Korem et al., 2007; O'Gara, 2007; Milisavljevic et al., 2008), our results demonstrate that ethanol and isopropanol promote further biofilm development of established biofilms. Additionally, these effects were observed with methanol, isoamyl alcohol, and n-butanol, and we demonstrated that biofilm enhancement could be accomplished at much higher alcohol concentrations (up to 100%) than previously tested (Fig. 1). Different antimicrobial stresses are known to affect S. aureus biofilm formation, in part through enhanced production of exopolysaccharides and other biofilm-promoting factors (Archer et al., 2011). Strikingly, we observed transcriptional enhancement of icaA and icaD (Fig. 5). Expression of icaADBC-encoded enzymes results in production and secretion of PIA, which enhances staphylococcal biofilm levels (Fitzpatrick et al., 2005; Resch et al., 2005; Archer et al., 2011). It is likely that increased production of adhesion proteins after alcohol treatment leads to greater bacterial binding to the plastic surface, or to neighboring bacteria, and thus greater biofilm biomass. Whether this increased biomass comes from growth of biofilm bacteria or from planktonic bacteria in the supernatant is not clear. Importantly, these same genes (icaA and icaD) were also transcriptionally increased in planktonic S. aureus treated with ethanol, which led to increased biofilm formation (Korem et al., 2010). Thus, these genes may be contributing to enhanced formation of nascent and established S. aureus biofilms under alcohol treatment, and they might be important in the maintenance of these treated biofilms (Fig. 1).
We also found that many bacteria within biofilms survived alcohol treatment (Fig. 3) and were metabolically active (Fig. 4). This result demonstrates that at least some of these bacteria were, in fact, living despite the treatment with high levels of alcohols. Notably, the relative level of bacterial metabolism, compared to untreated controls, increased in the time period between 4 h of treatment to 24 h of treatment (Fig. 4). These results mirror the increase in biofilm enhancement over untreated control samples between 4 h (Fig. 2) and 24 h (Fig. 1) of treatment. In other words, whereas after 4 h of alcohol treatment there was little enhancement of biofilm levels over untreated controls, and relatively significantly decreased metabolism, by 24 h, biofilm levels were greatly increased compared to untreated controls, and metabolism had also increased. It is tempting to speculate that the biofilm milieu provided an environment suitable for the development of an adaptive response to the alcohols, which subsequently led to greater biofilm formation.
Part of that adaptive response appears to be induction of antibiotic resistance gene transcription. We found enhanced transcription of mmpL, vraS, and mepA after treatment with ethanol (Fig. 6). mmpL and mepA encode putative (Cuaron et al., 2012) and characterized (Kaatz et al., 2006) antibiotic efflux pumps, respectively. The vraS gene encodes a sensor histidine kinase controlling the bacterial genetic response to cell wall-active antibiotics, and VraS activity has been found to be required for S. aureus resistance to oxacillin (Boyle-Vavra et al., 2006). It is possible that ethanol-stimulated enhancement of antibiotic resistance gene expression could overcome the inhibitory effects of antibiotic treatment (Fig. 6d). In the future, it will be interesting to extend our analysis to include antibiotic-resistant clinical isolates, including methicillin-resistant strains.
It is generally thought that treatment of biofilms with toxic chemicals will lead to, if not a complete biofilm eradication, at least a decrease in biofilm levels due to some bacterial killing. However, our results suggest that alcohols might be a poor disinfectant choice for the eradication of well-established S. aureus biofilms, such as might be found on heavily contaminated surfaces. Furthermore, because we found that biofilm enhancement increased with increasing alcohol concentration (Fig. 1), simply adding more alcohol might hinder disinfection. Further studies will elucidate the mechanisms of how preformed S. aureus biofilms sense and respond to alcohol treatment.
We thank G. O'Toole for the kind donation of the S. aureus laboratory strains used in this study. We also thank Joel Reyes and Beverly Ransdell for technical support. This work was supported by GK-12 Fellowship from the NSF to C. Redelman, and RSFG from IUPUI and PRF from Purdue University to G. Anderson.