Inactivation of Escherichia coli O157:H7 in biofilm on stainless steel by treatment with an alkaline cleaner and a bacteriophage

Authors


L.R. Beuchat, Center for Food Safety, University of Georgia, 1109 Experiment Street, Griffin, GA 30223-1797, USA (e-mail: lbeuchat@uga.edu).

Abstract

Aims:  To determine the effectiveness of an alkaline cleaner used in food-processing plants and a lytic bacteriophage specific for Escherichia coli O157:H7 in killing wild type and rpoS-deficient cells of the pathogen in a biofilm.

Methods and Results:  Wild type and rpoS-deficient cells were attached to stainless steel coupons (c. 7–8 log CFU per coupon) on which biofilms were developed during incubation at 22°C for 96 h in M9 minimal salts media (MSM) with one transfer to fresh medium. Coupons were treated with 100 and 25% working concentrations of a commercial alkaline cleaner (pH 11·9, with 100 μg ml−1 free chlorine) used in the food industry, chlorine solutions (50 and 100 μg ml−1 free chlorine), or sterile deionized water (control) at 4°C for 1 and 3 min. Treatment with 100% alkaline cleaners reduced populations by 5–6 log CFU per coupon, a significant (P ≤ 0·05) reduction compared with treatment with water. Initial populations (2·6 log CFU per coupon) of attached cells of both strains were reduced by 1·2 log CFU per coupon when treated with bacteriophage KH1 (7·7 log PFU ml−1) for up to 4 days at 4°C. Biofilms containing low populations (2·7–2·8 log CFU per coupon) of wild type and rpoS-deficient cells that had developed for 24 h at 22°C were not decreased by more than 1 log CFU per coupon when treated with KH1 (7·5 log PFU ml−1) at 4°C.

Conclusions:  Higher numbers of cells of E. coli O157:H7 in biofilms are killed by treatment with an alkaline cleaner than with hypochlorite alone, possibly through a synergistic mechanism of alkaline pH and hypochlorite. Populations of cells attached on coupons were reduced by treating with bacteriophage but cells enmeshed in biofilms were protected.

Significance and Impact of the Study:  The alkaline pH, in combination with hypochlorite, in a commercial cleaner is responsible for killing E. coli O157:H7 in biofilms. Treatment with bacteriophage KH1 reduces populations of cells attached to coupon surfaces but not cells in biofilms.

Introduction

Biofilms formed by Escherichia coli O157:H7 on inadequately cleaned and sanitized contact surfaces may be a source of contamination of ground beef and deli meat in processing facilities as well as in food service settings. Refrigeration temperatures in these environments provide opportunities for E. coli O157:H7 originating from faecal material on carcasses and hides to survive, attach to surfaces such as stainless steel, and become persistent (Chmielewski and Frank 2003). Outbreaks of infections caused by E. coli O157:H7 in the US have been associated with the consumption of undercooked meat products (Mead et al. 1999), including dry salami (Centers for Disease Control and Prevention 1995), fresh produce (Beuchat 2002), unpasteurized juices (IFT/FDA 2001), and drinking water (Health Canada 2000). Infections with E. coli O157:H7 can lead to the development of haemolytic uremic syndrome, causing renal failure, and thrombocytopenic purpura (Mead and Griffin 1998; Meng et al. 2001).

Treatment of surfaces with alkaline solutions such as strongly alkaline chlorinated cleaners used to remove fats and proteins in meat processing facilities, will temporarily impart hydrophilic properties to stainless steel and may enhance or decrease bacterial attachment to surfaces. The effects of acid and alkaline pH on adhesion of Bacillus cereus spores to surfaces have been shown to depend on the strain of the micro-organism and composition of the cell or substratum surface (Lindsay et al. 2004). The adhesion of E. coli cells to surfaces has been reported to be inversely proportional to the negative charge on the surface of the cell (Chmielewski and Frank 2003). It has also been observed that biofilms formed by E. coli O157:H7 are more strongly attached to surfaces, mature more rapidly, and contain more extracellular polysaccharide (EPS) when grown in nutritionally limited environments (Dewanti and Wong 1995). Such environments would occasionally exist in food-processing facilities.

Cells attached to surfaces or enmeshed in biofilms may have altered sensitivities to cleaners and sanitizers compared with sensitivities of planktonic cells. Salmonella in biofilms has been reported to be more resistant than planktonic cells to acidic challenge, hypochlorite and iodophors (Joseph et al. 2001; Gawande and Bhagwat 2002). Biofilms may protect cells through a combination of mechanisms, including diffusional resistance of the EPS matrix, chemical and enzymatic inactivation of sanitizers and disinfectants, physiological changes in cells, and the induction of stress responses in cells (Gilbert et al. 2002). Strong alkaline cleaners containing hypochlorite have been shown to be effective in killing planktonic cells of E. coli O157:H7 (Sharma and Beuchat 2003) but little is known about the ability of alkaline cleaners to inactivate E. coli O157:H7 in biofilms.

The emergence of antibiotic-resistant strains of food-borne pathogens has raised increased interest in developing alternative bactericidal treatments, e.g. bacteriophages (Merrill et al. 1996). Bacteriophages have been commercially produced for use as antibacterials and continue to be used in eastern Europe (Sulakvelidze et al. 2001). The ability of bacteriophages to kill E. coli O157:H7 attached to nonfood surfaces and in biofilms, or a comparison of this ability to reductions in populations to those caused by cleaners used in food-processing plants, has not been described.

Adsorption of bacteriophages to cells through specific receptors is followed by infection that results in lysis and release of large numbers of phage particles capable of infecting and lysing more cells. The specificity of bacteriophages for a micro-organism and their ability to propagate at the site of infection distinguishes them from other antimicrobials (Campbell 2003). This specificity enables their use as biocides to inactivate pathogenic or spoilage bacteria in situations that rely on the presence of natural flora to achieve, for example, desired fermentation of meat products (Ammor et al. 2004). Bacteriophages are also thought to have less effect than some types of sanitizers on sensory characteristics of food (Kudva et al. 1999). They can be combined with competitive exclusion (CE) micro-organisms to reduce populations of food-borne pathogens in food-processing plants without inactivating CE cells.

Bacteriophages have been applied to various poultry products and fresh-cut produce for the purpose of inactivating food-borne pathogens. A lytic bacteriophage specific for Salmonella Enteritidis was shown to reduce populations of the pathogen on chicken skin (Goode et al. 2003). A reduction of 2 log10 CFU was achieved when the multiplicity of infection value (the number of phage particles needed to infect one bacterial cell) was 2. By comparison, treatment with bacteriophage has been reported to reduce populations of Salmonella by 1·5 log10 g−1 of vegetable seed sprouts (Pao et al. 2004). Lytic bacteriophage mixtures were shown to reduce populations of Salmonella by 3·5 log10 CFU per wound on honeydew melons but were not as effective on fresh-cut apples (Leverentz et al. 2001). Populations of Listeria monocytogenes on the surface of honeydew melons were also reduced by 2·0–4·6 log per sample by applying lytic bacteriophage (Leverentz et al. 2003).

Attention has also been given to the use of bacteriophage to degrade EPS through bacteriophage-associated polysaccharide depolymerase activity, and simultaneously killing cells through lytic activity (Hughes et al. 1998). Previous studies have identified bacteriophages specific for E. coli O157:H7 and shown reduction in populations in cultures and on meat surfaces (Kudva et al. 1999; O'Flynn et al. 2004). These bacteriophages have not been evaluated for their ability to kill E. coli O157:H7 cells attached to nonfood surfaces or in a biofilms, or compared with cleaners or sanitizers for efficacy in killing the pathogen. Other have shown that rpoS-deficient E. coli may form more EPS than wild type cells (Corona-Izquierdo and Membrillo-Hernandez 2002), which may alter sensitivity of cells to chemicals in cleaners used in food-processing plants.

The objective of this study was to determine the effectiveness of a commercial alkaline cleaner used in food-processing plants and a lytic bacteriophage in killing wild type and rpoS-deficient cells of E. coli O157:H7 in biofilms.

Materials and methods

Bacterial strains

Escherichia coli O157:H7 strain ATCC 43895 and FRIK 816-3 (an rpoS-deficient strain of ATCC 43895) were used in studies to evaluate the effectiveness of a commercial alkaline cleaner and chlorinated waters in killing cells attached to stainless steel coupons and in biofilms formed on coupons. The lethality of bacteriophage KH1 to attached cells and cells in biofilms on stainless steel coupons was also studied. Escherichia coli O157:H7 strain ATCC 43895 was used for bacteriophage propagation and titre determination.

Preparation of cells and attachment of cells to coupon

Cells from stock cultures of E. coli O157:H7 strains ATCC 43895 and FRIK 816-3 were surface plated on tryptic soy agar (TSA) (BBL/Difco, Sparks, MD, USA) and TSA supplemented with 100 μg ml−1 ampicillin (TSAA), respectively, and incubated at 37°C for 24 h. Cells from colonies of strains ATCC 43895 and FRIK 816-3 were inoculated into 10 ml of tryptic soy broth (TSB) (BBL/Difco) and TSB supplemented with 100 μg ml−1 ampicillin (TSBA) respectively. After incubating at 37°C for 24 h, 0·1 ml of culture was inoculated into 100 ml of TSB or TSBA and incubated at 37°C for 24 h to attain a stationary phase of growth (c. 9 log CFU ml−1). One hundred millilitres of each culture were centrifuged in 2 × 50-ml conical centrifuge tubes (VWR International, South Plainfield, NJ, USA) at 4000 g for 10 min (Marathon, Pittsburgh, PA, USA). The supernatant was decanted and cells were resuspended in 100 ml of sterile phosphate-buffered saline (PBS, pH 7·4), composed (per litre of deionized water) of 8 g of NaCl, 0·2 g of KCl, 1·44 g of Na2HPO4 and 0·24 g of KH2PO4. Cell suspensions were then diluted in 900 ml of PBS solution to yield a population of c. 8 log CFU ml−1. Thirty millilitres of cell suspension were then deposited into a sterile 50-ml centrifuge tube. Sterile stainless steel coupons (Type 304, #4 finish, 2 × 5 cm2), prepared as described by Ryu et al. (2004), were then immersed in cell suspensions and held at 4°C for 24 h to facilitate attachment of cells, i.e. adherence of cells without biofilm formation. Coupons were removed from cell suspensions of strains ATCC 43895 and FRIK 816-3 with a sterile forceps, gently rinsed in a circular motion in 400 ml sterile PBS for 15 s, and placed in 30 ml of M9 minimal salts medium (MSM) (BBL/Difco) or in MSM supplemented with 50 μg ml−1 ampicillin (MSMA) respectively. Coupons on which cells were attached were incubated in MSM or MSMA at 22°C for 48 h to allow biofilms to form. Coupons were then removed from MSM or MSMA, rinsed in PBS as described above, placed in 30 ml of fresh sterile MSM or MSMA, respectively, and incubated at 22°C for 48 h. Coupons were held and placed at 4°C for approx. 1 h before treatment with alkaline cleaner, chlorine solutions, or water (control).

Preparation of alkaline cleaner and free chlorine solutions

A commercial alkaline cleaner, Enforce® (Ecolab, Inc., St Paul, MN, USA), was diluted to give a 100 and 25% working concentrations (pH 11·9 at 23°C). According to the label, a 100% working concentration of the cleaner contains 11% sodium hydroxide and 1·8% total available chlorine. To prepare free chlorine solutions, sodium hypochlorite (NaOCl) was added to cold (4°C) sterile 0·05 mol l−1 potassium phosphate buffer (pH 6·8) to give a low (50 μg ml−1) and high (100 μg ml−1) concentrations of free chlorine. Free chlorine concentrations in Enforce and chlorine solutions were determined with a DR/820 colorimeter (Hach, Loveland, CO, USA). Sterile deionized water was used as a control. Thirty millilitres of treatment solutions or deionized water were dispensed in sterile 50-ml conical centrifuge tubes and held at 4°C until used to treat coupons within 5 min.

Treatment of biofilms of E. coli O157:H7 with alkaline cleaner, chlorine solutions and water

Coupons on which E. coli O157:H7 biofilms had formed were removed from MSM or MSMA, and gently rinsed in 400 ml of sterile PBS for 15 s immersed in 30 ml of 25% Enforce, 100% Enforce, 50 μg ml−1 free chlorine, 100 μg ml−1 free chlorine, or sterile deionized water at 4°C in 50-ml centrifuge tubes and placed in an a 4°C incubator at 4°C. Coupons immersed in treatment solutions or water for 1 or 3 min were removed using a sterile forceps and immediately placed in 30 ml of Dey-Engley (DE) neutralizing broth (Difco/BBL) in a sterile 50-ml sterile centrifuge tube containing 3 g sterile glass beads (425–600 μm diam.) (Sigma, St Louis, MO, USA). Tubes containing treated coupons and DE broth were then agitated for 1 min using a benchtop vortex (VWR, South Planfield, NJ, USA) to remove cells from the coupon. After agitation, populations of E. coli O157:H7 in DE broth of treated coupons were enumerated using an automated spiral plater (Spiral Biotech, Norwood, MA, USA). Either 50 μl (in duplicate) or 1 ml (0·25 ml, in quadruplicate) of each sample of DE broth containing strains ATCC 43895 or FRIK 816-3 were plated on TSA or TSAA respectively. Two coupons were subjected to each combination of treatment solution and time, and four replicate experiments were performed.

Preparation of plating E. coli O157:H7 for production of bacteriophage

Cells from a single colony of E. coli O157:H7 ATCC 43895 were inoculated into 50 ml of Luria–Betrani (BBL/Difco) medium supplemented with 5 mmol l−1 MgSO4 (LBM). Cultures were incubated at 37°C for 18 h on rotary shaker (250 rev min−1), then centrifuged at 4000 g for 10 min. The supernatant was decanted and cells were resuspended in 20 ml of sterile 0·01 mol l−1 MgSO4 solution (pH 5·4). The optical density (OD600) of the suspension was adjusted to 2·0. Cultures were diluted 1 : 10 in sterile MgSO4 buffer (MB, pH 7·9), which contains 5·8 g of NaCl, 2 g of MgSO4, 50 ml of Tris–HCl (pH 7·5), and 5 ml of 2% gelatin in 1 l of deionized water.

Preparation of bacteriophage KH1 stocks

Bacteriophage KH1, determined to be specific for E. coli O157:H7, was provided by Dr Carolyn Hovde at the University of Idaho, Moscow, ID, USA (Kudva et al. 1999). Bacteriophages for treating E. coli O157:H7 were prepared using a soft agar overlay technique. Briefly, 100 μl of bacteriophage KH1 stock suspension (c. 3 log PFU ml−1) was combined with 100 μl of a suspension of E. coli O157:H7 strain ATCC 43895 (c. 8 log CFU ml−1) in a 1·5-ml microcentrifuge tube (Fisher Scientific, Pittsburgh, PA, USA). Suspensions of cells and bacteriophage were mixed, held for 15 min at 37°C to allow bacteriophage to adsorb to cells, mixed with LBM soft agar (0·75%) in a sterile test tube, and overlaid on LBM agar in Petri dishes (100 × 15 mm). After incubating at 37°C for 16 h, soft agar from plates displaying plaques was removed using a sterile flat spatula and deposited in 1·5-ml microcentrifuge tubes. Samples were centrifuged at 12 000 g in a Marathon mini-centrifuge for 20 min at 22°C. Supernatant (lysates) from five tubes was collected and treated with 1 ml of chloroform (Fisher Scientific) for 15 min at 22°C before centrifuging again. Lysates were then filtered through a 0·2-μm filter (Corning Inc., Corning, NY, USA) and collected. Bacteriophage stocks were stored in 1 ml chloroform at 4°C. The titre of the bacteriophage stock suspension, determined by a soft agar overlay technique, was 9·1 log PFU ml−1.

Confluent lysis on soft agar plates was observed when 1 ml of bacteriophage suspension (7·1 log PFU ml−1) was mixed with 1 ml of a suspension of E. coli O157:H7 strain ATCC 43895 containing 8 log CFU ml−1 suspension. To prepare large volumes of bacteriophage suspensions needed for treatment of coupons, 1 ml of suspension (7·1 log PFU ml−1) was mixed with 1 ml of E. coli O157:H7 strain ATCC 43895 suspension (8 log CFU ml−1), absorbed, mixed with 10 ml of LBM soft agar, overlaid on LBM in large Petri plates (150 × 20 mm), and incubated at 37°C for 12 h. Twenty-five Petri plates were prepared in this manner. Soft agar (c. 10 g per plate) from plates was collected as described above, combined, and deposited in a sterile 250-ml polypropylene centrifuge bottle. Soft agar containing cells and bacteriophage was centrifuged at 12 000 g for 20 min at 4°C in a J2 Mini centrifuge (Beckman Coulter, Fullerton, CA, USA). Lysate containing bacteriophage was separated from the pellet containing agar and cells, and pellets were centrifuged again to remove and separate remaining lysate contained in the pellet. Lysates were combined to give c. 150 ml. Lysate was deposited in a sterile 250-ml Erlenmeyer flask containing 10 ml of chloroform and incubated at 22°C with agitation (250 rev min−1) for 20 min to kill viable cells remaining in the lysate. Lysate was separated from chloroform, centrifuged at 12 000 g for 20 min at 4°C, removed from the pellets containing dead cells and cellular debris, and filtered (0·2-μm filter, Corning). Bacteriophages in the filtrate were stored in 10 ml of chloroform at 4°C for several days until used to treat coupons on which E. coli O157:H7 cells had attached or formed biofilms.

Treatment of attached cells and cells in biofilms with bacteriophage

Cells attached to coupons and in biofilms on coupons were prepared as described above. For studies with attached cells, coupons were placed in a suspension containing 4 log CFU ml−1 of strains ATCC 43895 or FRIK 816-3 in PBS for 24 h at 4°C. Coupons were removed from suspensions, gently rinsed as described above, and deposited in 27 ml of sterile MB. Three millilitres of bacteriophage suspension with a titre of 9 log PFU ml−1 or 3 ml of MB buffer (control) were added to centrifuge tubes containing coupons with attached cells and 27 ml of MB buffer. Coupons were incubated for 15 min at 37°C before storing at 4°C for 1, 2, 3, or 4 days. The number of viable cells on coupons was determined by adding 3 g of sterile glass beads (425–600 μm) to each tube containing a coupon and 30 ml of bacteriophage suspension or MB and vortexing for 1 min. Populations of E. coli O157:H7 in suspension were determined by surface plating undiluted samples (0·25 ml in quadruplicate or 0·1 ml in duplicate) and samples serially diluted in 0·1% peptone water (0·1 ml in duplicate) on TSA and TSAA for strains ATCC 43895 and FRIK 816-3 respectively. Plates were incubated at 37°C for 24 h before colonies were counted. Tubes containing coupons on which cells of strains ATCC 43895 and FRIK 816-3 had attached were also enriched with, respectively, 15 ml of 3X Universal Pre-enrichment Broth (UPB) (Difco/BBL) or 3X UPB containing 150 μg ml−1 ampicillin and incubated for 24 h at 37°C. A loopful of enriched broth of strains ATCC 43895 or FRIK 816-3 was streaked on the surface of either sorbitol MacConkey agar (SMAC)(Difco/BBL) or SMAC containing 50 μg ml−1 ampicillin, respectively, and incubated at 37°C for 24 h. Presumptive E. coli O157:H7 colonies were then analysed using the O157 latex agglutination test (Oxoid, Basingstoke, UK).

Biofilms were also treated with lower numbers of bacteriophage to evaluate their relative effectiveness compared with the effectiveness of high numbers. Coupons were placed in suspensions (30 ml) containing strain ATCC 43895 or FRIK 816-3 at 4 log CFU ml−1 in a sterile 50-ml centrifuge tubes, respectively, for 24 h at 4°C. Coupons on which cells had attached were gently rinsed as described above, placed in either 30 ml of MSM or MSMA supplemented with 5 ml of 1 mol l−1 MgSO4 per litre, respectively, incubated for 24 h at 22°C, removed from broths, gently rinsed as described above, and placed in 27 ml of MB. Three millilitres of bacteriophage suspension (8·7 log PFU ml−1) or 3 ml of MB was added to 27 ml of MB in which a coupon was immersed. In the second study, 3 ml of bacteriophage at a slightly lower titre (8·5 PFU ml−1) or 3 ml of MB buffer (control) were added to tubes containing inoculated coupons on which biofilms had formed. Coupons treated with bacteriophage were incubated for 15 min at 37°C and stored at 4°C (adsorbed) or immediately placed at 4°C after the addition of bacteriophage (unadsorbed). Coupons in MB (control) were immediately placed at 4°C. Coupons treated with adsorbed or unadsorbed bacteriophage suspension or with MB were stored for up to 4 days at 4°C and populations of E. coli O157:H7 were enumerated on each day of storage.

To enumerate cells in the first study, 3 g of glass beads were added to directly to each tube containing a coupon and bacteriophage suspension or MB and vortexed for 1 min. Cell suspensions from these tubes were surface plated on TSA or TSAA. For the second study, coupons were transferred to 30 ml of MBSM buffer containing 3 g of sterile glass beads and vortexed for 1 min. Populations were enumerated by surface plating the MB (as described above) on days 1, 2, 3, and 4. Four replicate experiments (one coupon for each treatment) in each study were performed. Titres of bacteriophage in suspensions were also determined on days 1, 2, 3, and 4 of the second study for both strains treated with adsorbed and unadsorbed bacteriophage KH1.

Treatment of planktonic cells with bacteriophage

Cell suspensions of strains ATCC 43895 and FRIK 816-3 containing 4 log CFU ml−1 of PBS with 5 mmol l−1 MgSO4 were prepared as described above and diluted 1 : 10. Suspensions (27 ml) in sterile 50-ml centrifuge tubes were stored at 4°C for 24 h before adding 3 ml of bacteriophage suspension (9 log PFU ml−1). The mixture was incubated at 37°C for 15 min before storing at 4°C for 1, 2, 3, or up to 4 days. Populations of E. coli O157:H7 were determined by surface plating suspensions as described above.

Statistical analysis

Populations of E. coli O157:H7 recovered from untreated and treated stainless steel coupons were subjected to analysis of variance (anova) and the least significant difference test using Statistical Analysis Software (SAS, Cary, NC, USA) to determine significant difference (P ≤ 0·05).

Results

Treatment of biofilms with alkaline cleaner

All cleaner and chlorine treatments caused significant reductions (P ≤ 0·05) in the number of viable cells on stainless steel coupons (Fig. 1). Treatment of biofilms of strains ATCC 43895 and FRIK 816-3 with 100% Enforce killed significantly (P ≤ 0·05) more cells than treatment with 25% Enforce, 100 μg ml−1 free chlorine, or 50 μg ml−1 free chlorine. Treatment of both strains with 50 and 100 μg ml−1 free chlorine for 1 min was significantly (P ≤ 0·05) more effective in reducing populations than treatment with 25% Enforce. Populations recovered from coupons treated with 50 or 100 μg ml−1 free chlorine were not significantly different (P > 0·05).

Figure 1.

Mean populations (n = 8) of Escherichia coli O157:H7 (strain ATCC 43895 and strain FRIK 816-3) recovered from stainless steel coupons on TSA and TSAA, respectively, after growing at 22°C for 48 h and treatment with water (control) (open bars), 100% Enforce (solid bars), 25% Enforce (bars with diagonal lines), 100 μg ml−1 free chlorine (shaded bars), and 50 μg ml−1 free chlorine (bars with horizontal lines) for 1 and 3 min at 4°C. Within strain and treatment time, values that are not noted by the same capital letter are significantly different (P ≤ 0·05). Within strain and treatment, values not noted by the same lowercase letter are significantly different. The detection limit was 30 CFU per coupon (1·5 log10 CFU per coupon). Populations of strains ATCC 43895 and FRIK 816-3 before treatment were 8·0 log10 and 7·1 log10 CFU per coupon respectively

Populations of strain ATCC 43895 were more significantly reduced when cells were exposed to 25% Enforce for 3 min compared with 1 min. Populations of strain FRIK 816-3 were significantly lower on coupons treated with 25% Enforce, 50 μg ml−1, and 100 μg ml−1 for 3 min than on coupons receiving respective treatments for 1 min.

Treatment of attached cells with bacteriophage

Initial numbers of attached cells of strains ATCC 43895 and FRIK 816-3 attached to coupons were reduced by 1·2 log CFU per coupon upon exposure to bacteriophage KH1 for 1 day (Fig. 2). Populations of attached cells of strain ATCC 43895 treated with bacteriophage for 1, 3, and 4 days were significantly (P ≤ 0·05) lower than populations treated with MB (control). Populations of strain FRIK 816-3 treated with bacteriophage were significantly lower than those of cells treated with MB after 1, 2, 3, and 4 days. Surviving cells of both strains not detectable by direct plating were detected by enrichment. No differences in susceptibility of wild type and rpoS-deficient strains to bacteriophage were observed.

Figure 2.

Mean populations (n = 8) of Escherichia coli O157:H7 strain ATCC 43895 and FRIK 816-3 attached to stainless steel coupons recovered on TSA and TSAA, respectively, after incubation at 4°C for up to 4 days in MB (control) (bsl00046 and bsl00001 respectively), and populations of strain ATCC 43895 attached to stainless steel coupons recovered on TSA from bacteriophage KH1 suspension (bsl00067) incubated at 4°C for 4 days. Initial populations for strains ATCC 43895 and FRIK 816-3 were 2·8 and 2·6 log CFU per coupon, respectively, before treatment with bacteriophage (8 log PFU ml−1) or MB. The detection limit was 30 CFU per coupon (1·5 log CFU per coupon) (dashed line). Populations of strain FRIK 816-3 treated with bacteriophage KH 1 suspension were below the detection limit after treatment for 1, 2, 3 and 4 days. On days on which counts were below the detection limit, samples were analysed by enrichment

Treatment of biofilms with bacteriophage

Initial populations of strains ATCC 43895 and FRIK 816-3 in biofilms were 4·4 and 4·0 log CFU per coupon, respectively, in the first study (Fig. 3). Treatment of biofilms of strain ATCC 43895 with adsorbed or unadsorbed bacteriophage (7 log PFU ml−1) did not significantly reduce counts compared with the MB control, regardless of treatment times on any day. However, treatment of biofilms of strain FRIK 816-3 with adsorbed and unadsorbed bacteriophage significantly reduced (P ≤ 0·05) populations after day 1 (2·8 log CFU per coupon) and day 2 (2·7 log CFU per coupon) compared with populations detected in MB (4·6 and 4·4 log CFU per coupon respectively). Populations were also reduced significantly (P ≤ 0·05) by treatment with unadsorbed bacteriophage on day 1 (3·0 log CFU per coupon) compared with the population surviving in MB. The number of cells detected after treatment with unadsorbed bacteriophage was not significantly different (P > 0·05) than the number surviving treatment with adsorbed bacteriophage for any treatment time.

Figure 3.

Mean populations (n = 8) of Escherichia coli O157:H7 strain ATCC 43895 and strain FRIK 816-3 cells recovered from biofilms on stainless steel coupons on TSA and TSAA, respectively, after treatment with MB (control, bsl00066), adsorbed bacteriophage (bsl00001), or unadsorbed bacteriophage (bsl00046), and incubation at 4°C for up to 4 days. Initial populations of strains ATCC 43895 and FRIK 816-3 recovered were 4·0 and 4·4 log10 CFU per coupon, respectively, before treatment with 7·7 log PFU ml−1 of bacteriophage suspension or MB

In the second study, initial populations of strains ATCC 43895 and FRIK 816-3 were 2·7 and 2·8 log CFU per coupon, respectively, and were treated with bacteriophage suspension of 7·5 log PFU ml−1 (Fig. 4). Treatments of biofilms of strain ATCC 43895 with adsorbed or unadsorbed bacteriophage did not cause significant (P > 0·05) reductions in populations compared with treatment with MB. After 1 day, populations of FRIK 816-3 recovered from biofilms treated with adsorbed bacteriophage (2·3 log CFU per coupon) were significantly (P ≤ 0·05) lower than populations recovered from biofilms treated with MB (2·9 log CFU per coupon). Bacteriophage titres for adsorbed cells of strains ATCC 43895 and FRIK 816-3 cells ranged from 7·0 log PFU ml−1 on day 1 to 6·8 log PFU ml−1 on day 4 respectively. Titres for unadsorbed bacteriophage treatments ranged from 6·8 log PFU ml−1 on day 1 to 6·6 log PFU ml−1 on day 4.

Figure 4.

Mean populations (n = 8) of Escherichia coli O157:H7 strain ATCC 43895 and strain FRIK 816-3 recovered on TSA and TSAA, respectively, from coupons transferred to sterile MB from biofilms on coupons after treatment with MB (control, bsl00066), adsorbed bacteriophage (bsl00001), or unadsorbed bacteriophage (bsl00046), and stored at 4°C for up to 4 days. Initial populations of strains ATCC 43895 and FRIK 816-3 recovered from coupons were 2·7 and 2·8 log10 CFU per coupon, respectively, before treatment with 7·5 log PFU ml−1 of bacteriophage suspensions or MB

Treatment of planktonic cells with bacteriophage

Treatment of planktonic cells of strains ATCC 43895 and FRIK 816-3 with bacteriophage resulted in reductions in populations of 1·2 and 1·6 log CFU ml−1, respectively, on day 1 (Fig. 5). Significant differences (P ≤ 0·05) between populations of cells of strain ATCC 43895 treated with bacteriophage and [PBS containing 5 mmol l−1 MgSO4 (control)] were observed only after treatment at 4°C for 4 days on day 4; significant differences in populations of cells of strain FRIK 816-3 treated with bacteriophage and MB were observed after storage at 4°C for 2, 3, and 4 days. Two replicate experiments of each treatment of each strain (one coupon/treatment) were performed.

Figure 5.

Mean populations (n = 4) of Escherichia coli O157:H7 strain ATCC 43895 and strain FRIK 816-3 recovered from planktonic suspensions on TSA and TSAA after treatment with PBS containing 5 mmol l−1 MgSO4 (control) (bsl00046 and bsl00001 respectively) or bacteriophage suspension KH1 (bsl00067 and bsl00000 respectively) at 4°C for up to 4 days. Initial populations for ATCC 43895 and FRIK 816-3 were 2·9 and 2·8 log CFU ml−1 before treatment with 8 log PFU ml−1 bacteriophage or PBS containing 5 mmol l−1 MgSO4

Discussion

Hypochlorite solutions containing 100 μg ml−1 free chlorine (100–103 ppm) were not as effective in killing E. coli O157:H7 cells as was 100% Enforce containing 100 μg ml−1 available free chlorine, indicating that synergistic bactericidal activity may result from the alkaline pH and hypochlorite. Chlorine compounds may be inactivated by EPS and other organic materials present on the surface of biofilms. Alkaline cleaners may be more effective than chlorine in penetrating biofilms because of their peptizing action against EPS (Chmielewski and Frank 2003) and surfactant properties. The rapid lethality of 100% Enforce may also be attributed in part to the dissolution of polymeric substances that attach E. coli O157:H7 cells to the stainless steel coupon and the disruption of cytoplasmic membrane integrity through saponification (Ammor et al. 2004). Others have shown that lethality of cleaning agents and sanitizers to cells in biofilms is decreased by biofilm components (Stewart et al. 2001). Chlorine may have been neutralized more rapidly than NaOH in the alkaline cleaner by the EPS in the biofilm.

Free chlorine concentrations in treatment solutions did not decrease by more than 14 μg ml−1 after treatment of coupons (data not shown). This corresponds with observations that chlorine concentration did not change significantly upon contact with stainless steel coupons inoculated with L. monocytogenes and Flavobacterium spp. (Bremer et al. 2002). Strong alkaline cleaners (12·5) containing hypochlorite reduced populations of Gram-negative and Gram-positive bacteria which had formed biofilms on stainless steel (AISI, 304, 2B) at 26°C (Bredholt et al. 1999). Alkaline solutions (pH 11) containing 400 μg ml−1 chlorine were slightly more effective in reducing L. monocytogenes and Flavobacterium spp. attached to stainless steel than alkaline solution alone (Bremer et al. 2002). In our study, the pH of the 25 and 100% Enforce solutions at 22°C was the same (pH 11·9) but solutions contained different concentrations of free chlorine (measured as 21 and 101 μg ml−1 respectively), indicating that the higher concentration of free chlorine in 100% Enforce was more bactericidal than the lower concentration present in 25% Enforce. This may indicate that there is a threshold concentration of free chlorine needed for synergistic bactericidal action of chlorine and high pH (Sharma and Beuchat 2003). We observed that the population of strain FRIK 816-3 in biofilms was significantly reduced (c. 1 log CFU per coupon) by treatment with 50 μg ml−1 free chlorine compared with treatment with 25% Enforce for 1 or 3 min.

All biofilms examined in this study were formed using a single species and this uniformity may have rendered cells to be more susceptible to inactivation by cleaning and sanitizing solutions than cells in biofilms containing multiple species (Frank 2000). Sensitivity of biofilms to alkaline bactericidal treatments may also be affected by the age of the biofilm (Stopforth et al. 2002). Cells of S. Enteritidis in biofilms grown for 48 h, for example, showed increased sensitivity to trisodium phosphate (pH 12·5) compared with cells grown for 72 h (Korber et al. 1997). Others have shown that pretreating surfaces inoculated with Pseudomonas aeruginosa and Staphylococcus aureus with an alkaline detergent (pH 11·6) before pressure washing aided in decreasing the viability of attached cells but not their removal (Gibson et al. 1999).

Cells of strain FRIK 816-3 in biofilms were not more resistant than cells of strain ATCC 43895 to alkaline cleaners and free chlorine. It has been reported that rpoS-deficient cells of a nonpathogenic strain of E. coli formed more biofilm than wild type cells (Corona-Izquierdo and Membrillo-Hernandez 2002) after 48 h at 37°C, and this increase in biofilm material, presumably EPS, may have protected cells against otherwise bactericidal agents. If this protective mechanism exists, it may be masked by the high bactericidal activity of Enforce in our study.

This is the first report in which bacteriophage KH1 has been shown to reduce populations of E. coli O157:H7 attached on stainless steel. Other studies using bacteriophage specific for E. coli O157:H7 reduced the number of cells attached to beef surfaces at 37°C (O'Flynn et al. 2004). Previous studies have also shown that bacteriophage KH1 was effective in reducing populations of planktonic cells at 4 and 37°C (Kudva et al. 1999). We observed that reductions in the number of cells of E. coli O157:H7 attached to stainless steel coupons or in planktonic suspensions were similar, indicating that attachment under the conditions tested did not provide additional protection against bacteriophage attack. Previous work by Kudva et al. (1999) demonstrated that at least a 1000-fold more bacteriophage particles than E. coli O157:H7 cells are needed to achieve a reduction in population. Compared with the population of E. coli O157:H7 in biofilms treated with alkaline cleaner (c. 7–8 log CFU per coupon), initial populations of attached cells, planktonic cells, and cells in biofilms were relatively low (c. 3–4 log CFU per coupon) so that titres of bacteriophage of c. 7–8 log PFU ml−1 would be at least 1000-fold greater than the population of cells. Kudva et al. (1999) reported that bacteriophage are more effective in killing cells in the presence of magnesium and when adsorption to cells is promoted. Because adsorption may be less likely to occur in a food-processing environment, lethality of bacteriophage to cells without adsorption was evaluated. The results are inconclusive because large differences in populations resulting from control and bacteriophage treatments were not observed.

It is unclear why cells of strain FRIK 816-3 in biofilms were more sensitive than cells of strain ATCC 43895 to bacteriophage KH1 after 1 day of treatment in the second study. Differences between counts after control and bacteriophage treatments, although statistically significant, were not >1 log CFU per coupon. Compared with results from the attached cell study (Fig. 2), the lack of reduction in populations of E. coli O157:H7 may be due to the biofilm formation by attached cells growing on stainless steel at 22°C for 24 h. This is not inconsistent with previous observations that changes in lipopolysaccharide content in cells may provide increased resistance to lysis by bacteriophage KH1 (Kudva et al. 1999). Hughes et al. (1998) showed that bacteriophage specific for Enterobacter agglomerans disrupts biofilms through a combination of lytic activity against cells and degradation of EPS through a polysaccharide depolymerase associated with the bacteriophage particle. It is possible that bacteriophage KH1 has lytic activity that kills cells but does not possess the EPS-degrading ability exhibited by other lytic bacteriophages. If the bacteriophage KH1 does produce polysaccharide depolymerase, it may not be as active at 4°C as at higher temperatures. Biofilms are composed of various polysaccharides representing different types of EPS (Sutherland 2001) and it is possible that EPS produced by E. coli O157:H7 may not be a substrate for depolymerase associated with KH1. It is also possible that bacteriophage entered a state of lysogeny, infecting E. coli O157:H7 cells but not lysing them (Doolittle et al. 1995).

Cells that survive infection by bacteriophage are thought to remain resistant to bacteriophage but resistance has been shown to be transient and cells can revert to be phage susceptible (O'Flynn et al. 2004). Complete lethality of E. coli O157:H7, i.e. the inability to detect cells by enrichment, has usually involved treatment with a mixture of bacteriophages specific for the pathogen (Kudva et al. 1999; O'Flynn et al. 2004). Our study involved only a single bacteriophage.

A second study was carried out after observing that by surface plating diluted suspensions containing E. coli O157:H7 and bacteriophage on TSA or TSAA, lytic killing of cells may have occurred during incubation at 37°C, and therefore populations recovered after 24 h may have not represented populations of cells in biofilms treated with bacteriophage at 4°C. In this study, coupons were transferred to tubes containing MB before vortexing with glass beads and surface plating to dilute bacteriophage titres to numbers that would be ineffective in reducing populations on the surface of media during incubation. When examining populations of E. coli O157:H7 recovered after treatment of cells attached on coupons and held at 4°C, coupons were not transferred to MB before enumeration.

Differences in initial counts of E. coli O157:H7 in the first and second studies may be due in part to the method used to recover cells from biofilms. Agitating coupons in the presence of glass beads may not be optimal for retrieval of low numbers of cells. This recovery method may also contribute to a lack of statistical significance in populations of strain ATCC 43895 recovered from biofilms on coupons treated with bacteriophage or water, even though differences were larger than those observed with strain FRIK 816-3. Direct plating by applying agar on the surface of coupons as a technique to enumerate cells may also lead to an underestimation of the number of CFU because colonies may develop from clumps of cells rather than from a single cell.

In summary, a commercial alkaline cleaner, hypochlorite solution, and bacteriophage KH1 were evaluated for their lethality to E. coli O157:H7. Alkaline cleaner at a 100% concentration was more effective than a 25% concentration or 50 or 100 μg ml−1 free chlorine in killing cells of wild type and rpoS-deficient E. coli O157:H7. Populations of planktonic cells and cells attached to stainless steel coupons were reduced by treatment with bacteriophage KH1 but populations in biofilms were not reduced dramatically by treatment with adsorbed or unadsorbed bacteriophage, indicating that bacteriophage KH1 does not possess sufficient lytic or enzymatic activity at 4°C to cause lethality. Further study is needed to evaluate the synergistic bactericidal activity of alkaline pH and hypochlorite, and to evaluate bacteriophages specific for E. coli O157:H7 for their ability to kill cells in biofilms.

Ancillary