Ethanol-Mediated Variations in Cellular Fatty Acid Composition and Protein Profiles of Two Genotypically Different Strains of Escherichia coli O157:H7

Abstract

Two strains of Escherichia coli O157:H7 were grown in tryptic soy broth (TSB, pH 7.1) supplemented with 0, 2.5, 5.0, 7.5, and 10% ethanol at 30°C for up to 54 h. Growth rates in TSB supplemented with 0, 2.5, and 5.0% ethanol decreased with an increase in ethanol concentration. Growth was not observed in TSB supplemented with 7.5 or 10% ethanol. The pH of TSB containing 5.0% ethanol decreased to 5.8 within 12 h and then increased to 7.0 at 54 h. The ethanol content in TSB supplemented with 2.5 or 5.0% ethanol did not change substantially during the first 36 h of incubation but decreased slightly thereafter, indicating utilization or degradation of ethanol by both strains. Glucose was depleted in TSB supplemented with 0, 2.5, or 5.0% ethanol within 12 h. Cells grown under ethanol stress contained a higher amount of fatty acids. With the exceptions of cis-oleic acid and nonadecanoic acid, larger amounts of fatty acid were present in stationary-phase cells of the two strains grown in TSB supplemented with 5.0% ethanol for 30 h than in cells grown in TSB without ethanol for 22 h. The trans-oleic acid content was 10-fold higher in the cells grown in TSB with 5.0% ethanol than those grown in TSB without ethanol. In contrast, cis-oleic acid was not detected in ethanol-stressed cells but was present at concentrations of 0.32 and 0.36 mg/g of cells of the two strains grown in TSB without ethanol. Protein content was higher in ethanol-stressed cells than in nonstressed cells. Sodium dodecyl sulfate-polyacrylamide gel electrophoresis protein profiles varied qualitatively as affected by the strain and the presence of ethanol in TSB. An ethanol-mediated protein (28 kDa) was observed in the ethanol-stressed cells but not in control cells. It is concluded that the two test strains of E. coli O157:H7 underwent phenotypic modifications in cellular fatty acid composition and protein profiles in response to ethanol stress. The potential for cross protection against subsequent stresses applied in food preservation technologies as a result of these changes is under investigation.

Ethanol can be found in foods, particularly in fruits and fruit products and in fermented foods and beverages, and in food-processing environments as a result of fermentation of sugars by naturally occurring microorganisms. Ethanol and cleaners containing ethanol and other alcohols are also used in some areas of food-processing plants to reduce or remove microorganisms on equipment and for promoting good worker hygiene. Occasionally, low concentrations of residual ethanol may be present on treated surfaces of equipment and in environmental niches that are not properly cleaned and sanitized. This provides an opportunity for pathogens to adapt and grow in environments with sublethal concentrations of ethanol.

Organic solvents are lethal to bacteria at various concentrations, depending upon their inherent toxicity and the level of intrinsic tolerance of cells. Gram-negative bacteria, including Escherichia coli, possess novel adaptive mechanisms that enable survival under stress conditions imposed by organic solvents (1, 19, 20). Enterohemorrhagic E. coli O157:H7 may also be able to adapt and grow in environments containing elevated concentrations of ethanol. The growth of E. coli O157:H7 in nutrient agar supplemented with ≤7.5% ethanol was shown not to be completely inhibited without the addition of NaCl to the medium (10). Control of the growth of E. coli O157:H7 in food-processing environments and in foods and beverages that may contain ethanol is crucial to minimizing the risk of human infections.

The responses of nonpathogenic E. coli upon exposure to ethanol and other stress conditions have been investigated. The effects of ethanol on steady-state growth and interference with cell division in E. coli K-12 have been described (8). It was proposed that the structure and functions of some membrane proteins are reversibly altered when cells are exposed to sublethal ethanol concentrations, allowing exponential growth. Growth of E. coli in the presence of ethanol has been reported to result in the synthesis of lipids containing increased proportions of unsaturated fatty acids (5). The phage shock protein operon (pspABCE) in E. coli is expressed in response to stress conditions such as heat shock, osmotic shock, filamentous phage infection, and exposure to ethanol and hydrophobic organic solvents (3, 14, 18, 27).

Ethanol is used in the food-processing industry largely for the purpose of killing microorganisms rather than as a preservative, although low concentrations of ethanol have been examined for controlling the growth of spoilage and pathogenic species (21). The efficacy of ethanol as a preservative in a wide range of foods was studied by Shibasaki (23). Shapero et al. (22) and Ballesteros et al. (2) concluded that growth inhibition of Staphylococcus aureus by ethanol was caused by factors other than reduced a_w. Treatment of a high-moisture bakery product with ethanol vapor delays the growth of and toxin production by Clostridium botulinum (6). The combined effects of ethanol, a_w, and pH on the probability of growth and toxin production by C. botulinum have been studied (7). Little is known about physiological changes in E. coli O157:H7 and other food-borne pathogens exposed to sublethal concentrations of ethanol. However, the prospect of using ethanol as a preservative in foods raises the need for more information on the behavior of spoilage and pathogenic microorganisms that may adapt to otherwise lethal concentrations of ethanol and subsequently exhibit altered survival or growth behavior and increased resistance to other environmental stresses imposed by traditional preservation technologies.

The physicochemical and functional characterization of E. coli O157:H7 grown in ethanol-containing substrates has not been described. It is important to gain a better understanding of the mechanisms that E. coli O157:H7 may possess to adapt and grow in these substrates in order to more accurately assess the level of safety hazard they may represent. In this study, two strains of E. coli O157:H7 isolated from unpasteurized apple juice and salami were grown in ethanol-supplemented tryptic soy broth (TSB) as a model food system. The growth patterns and changes in pH and glucose content in TSB as affected by ethanol content were determined. Early-stationary-phase cells grown in TSB supplemented with 5.0% ethanol were analyzed for fatty acid composition, protein content, and sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) protein patterns. Cells grown in TSB and in TSB supplemented with ethanol were subjected to pulsed-field gel electrophoresis (PFGE) analysis to compare genomic DNA fingerprints.

MATERIALS AND METHODS

E. coli O157:H7 strains used.

E. coli O157:H7 strain SEA 13B88, an isolate from unpasteurized apple juice, and E. coli O157:H7 strain 30-2C4, an isolate from salami, were studied. Both strains were grown on tryptic soy agar (TSA) (BBL/Difco, Sparks, Md.) and stored at 4°C until used.

Conditions for culturing E. coli O157:H7 in ethanol-supplemented broth.

From each strain grown on TSA, a loopful of inoculum was streaked onto a TSA plate and incubated at 30°C for 18 to 20 h. A loopful of inoculum was then transferred from a single colony to a 250-ml Erlenmeyer flask containing 100 ml of TSB (pH 7.3 ± 0.2) containing 0.25% glucose (BBL/Difco) and supplemented with 2.0% ethanol that had been filtered through a membrane (0.2 μm; Gelman Sciences, Ann Arbor, Mich.) and incubated on an orbital shaker (160 rpm) at 30°C for 48 h. The population of viable cells was 9.5 to 9.8 log₁₀ CFU/ml. These cultures were used as inocula for further study of the behavior of E. coli O157:H7 in TSB supplemented with higher concentrations of ethanol.

To prepare TSB containing various concentrations of ethanol, 30 g of TSB powder (formulated to prepare 1 liter of broth) was dissolved in 980, 954, 927, 901, or 875 ml of deionized water, from which 49.0, 47.7, 46.3, 45.1, and 44.8 ml, respectively, were withdrawn and dispensed in a series of 250-ml Erlenmeyer flasks for use in experiments to determine changes in pH and changes in ethanol and glucose contents resulting from growth. The same volumes of TSB were dispensed into a series of 300-ml sidearm Erlenmeyer flasks for the purpose of monitoring growth by repeated in situ measurements of broth turbidity. After autoclaving and cooling the TSB to 22 ± 2°C, 0, 1.3, 2.4, 4.0, or 5.3 ml of filtered 95% ethanol was added to the 250-ml and 300-ml flasks to give 0, 2.5, 5.0, 7.5, and 10% (vol/vol) ethanol concentrations, respectively, after adding 1 ml of inoculum, for a final volume of 50 ml. Inoculated broth was incubated on an orbital shaker (160 rpm) at 30°C for up to 54 h before analysis.

At 3- to 12-h intervals, 1 ml of culture was withdrawn from the 250-ml flasks; the pH of 0.2 ml was measured, and 0.8 ml was stored at −30°C until analyzed for ethanol and glucose contents. The pH of the broth (0.2 ml) deposited onto a paraffin film was measured with a pH meter (basic pH meter; Denver Instrument Co., Arvada, Colo.) equipped with a contact electrode (recorder 540 CD combination pH electrode; Sensorex, Garden Grove, Calif.). To monitor cell growth, the turbidity of the culture was determined by inserting the sidearm of each 300-ml flask into the tube holder in a spectrophotometer (Spectronic 20; Bausch and Lomb, Buffalo, N.Y.) and recording the optical density at 600 nm. All experiments were replicated three times.

Determination of ethanol and glucose contents in TSB.

Samples (0.5 ml) of TSB periodically withdrawn from cultures were combined with 9.5 ml of 60 mM NaOH and membrane filtered (0.2 μm). Ethanol and glucose contents in the filtrates were determined by high-pressure liquid chromatography (HPLC) (LC-6A; Shimadzu Co., Kyoto, Japan); the instrument was equipped with an ionic exchange column (RCX-10; Hamilton Co., Reno, Nev.). An isocratic state with 60 mM NaOH as a mobile phase, a flow rate of 1.0 ml/min, and a loading volume of 20 μl were used. A refractometer (differential refractometer R401; Waters Associates, Milford, Mass.) was used to monitor refractive intensity, and an integrator (C-R5A Chromatopac; Shimadzu) was used for quantitative estimation and qualitative characterization of ethanol and glucose in the effluents. A reference solution of 60 mM NaOH containing 50 μg of glucose/ml and 100 μg of ethanol/ml was chromatographed under identical conditions to determine ethanol and glucose contents, respectively, in each sample.

Procedure for harvesting cells.

The two strains of E. coli O157:H7 were grown in TSB and in TSB supplemented with 5.0% ethanol prepared as described above. Based on the growth curves obtained in preliminary studies, cells grown in TSB and ethanol-supplemented TSB reached the early stationary phase in 22 and 30 h, respectively. Cultures were centrifuged (10,000 × g, 10 min, 5°C) (Marathon 12 KBR benchtop refrigerated centrifuge; Hermle-Labortechnik, Gosheimerstr, Germany), and the cells were washed twice in deionized water and lyophilized (TriTis Sentry; TriTis Co. Inc., Gardiner, N.Y.) before storage at −30°C for further analyses.

Analysis of cells for fatty acid composition.

The procedure of Miller and Berger (17), with minor modifications, was used to determine fatty acid profiles of E. coli O157:H7 cells as affected by ethanol content in TSB. Each lyophilized sample (10 mg) was placed in a Teflon-lined screw-cap test tube (13 by 125 mm) to which 1 ml of saponification solution (15% NaOH in 50% [wt/vol] methanol) was added. After swirling to disperse and suspend the cells, tubes were sealed and heated in a heating block (Thermolyne 16500 Dri-Bath; Branstead Co., Dubuque, Iowa) set at 100°C for 30 min. After cooling to 22 ± 2°C, 2 ml of methylation reagent (6 N HCl-methanol, 65:55, vol/vol) was added to the mixture before the tubes were sealed, swirled, and placed into the heating block for methylation at 80°C for 10 min. After cooling to 22 ± 2°C, 1 ml of n-hexane was added to the mixture, and the tubes were sealed, followed by mild and intermittent vortexing for 5 min (3 to 4 s at 30-s intervals). After the mixtures were centrifuged at 500 × g for 3 min, the upper layer was withdrawn and placed in a 1.5-ml microcentrifuge tube (1.5 ml) containing 60 to 70 mg of anhydrous sodium sulfate. After vortexing the microcentrifuge tubes for 5 s and holding them for 5 min, we centrifuged the mixture (8,000 × g in a Micro16; Fisher Scientific Co., Pittsburgh, Pa.) for 2 min, and the upper layer was withdrawn, deposited in an amber vial, sealed, and stored at −30°C until analyzed.

A gas chromatograph (Varian CP 3800; Varian Inc., Palo Alto, Calif.) equipped with a flame ionization detector and a DB-5MS capillary column (30 m long by 0.25 mm inner diameter by 0.25-μm film thickness; J & W Scientific, Inc., Folsom, Calif.) and an integrator (C-R5A Chromatopac) were used to analyze cells for fatty acid composition. A temperature program initiated at 150°C for 4 min was linearly increased to 250°C at 4°C/min and held for 5 min. Hydrogen was used as a carrier gas. The flow rate of carrier gas, split ratio, and injector temperature were set at 1.2 ml/min, 20:1, and 260°C, respectively. A bacterial acid methyl ester mixture (Sigma-Aldrich, St. Louis, Mo.) was analyzed under identical conditions as a reference for fatty acid identification and quantitation. Each fatty acid in the sample was further confirmed by gas chromatography-mass spectrometry spectrophotometry (Hewlett Packard HP 6890, HP 5973N MSD, and ChemStation) equipped with a quadrupole mass analyzer (Agilent Technologies, Palo Alto, Calif.). The amount of each fatty acid was estimated by calculating the integrated area of each acid obtained on chromatograms in proportion to areas obtained from the reference compounds in the standard sample (10 mg/ml, total concentration of 26 bacterial fatty acids).

Nitrogen determination and cellular protein analyses.

From each freeze-dried cell sample, 10 to 15 mg was deposited in a furnace cup and subjected to nitrogen determination with a nitrogen analyzer (Leco 2100; Leco Corporation, St. Joseph, Mich.) following the operational procedure provided by the manufacturer. The crude protein content was estimated by multiplying the nitrogen content by 6.25.

For cellular protein extraction, 10 mg of freeze-dried cell sample was deposited in a 50-ml tube and mixed with 1.0 ml of lysis buffer containing 5 mM MgCl₂, 5 mM EDTA, 1% Nonidet P-40, and 1 mM phenylmethylsulfonyl fluoride (Sigma Co.) as a protease inhibitor, all dissolved in 10 mM Tris buffer (pH 7.5). The suspension was immersed in an ice bath and sonicated for 15 min with a probe sonifier (Sonifier 450; Branson UL Trasonics Co., Danbury, Conn.) set at 5 for the power outlet and 50% of the duty cycle. After sonication, suspensions were withdrawn and centrifuged (Micro 16; Fisher Scientific Co.) at 16,000 × g for 10 min. The protein contents in the supernatural fluids were determined with a DC protein assay kit (Bio-Rad, Hercules, Calif.), with a standard curve constructed with bovine serum albumin (Sigma). The supernatants were stored at −30°C until analyzed.

For SDS-PAGE analysis, the Laemmli method (15) was followed. Cell extract (7.5 μl) was mixed with 7.5 μl of Laemmli dye solution (Bio-Rad) supplemented to contain 10% (vol/vol) mercaptoethanol and heated in a water bath at 95°C for 5 min. After cooling to 22 ± 2°C, the mixture was deposited on 12% polyacrylamide gels and separated with a mini-Protean II (Bio-Rad) set at 20 mA constant current for the two gels for 15 min and then raised to 30 mA for 40 min. The gels were stained with colloidal Coomassie blue (contains 0.1% Coomassie brilliant G-250, 0.2% phosphoric acid, 10% ammonium sulfate, and 20% methanol) overnight and destained with a solution containing 1% acetic acid and 1% glycerol. A low-molecular-weight marker (Amersham Biosciences Co., Piscataway, N.J.) was run concurrently as a size marker.

PFGE genomic DNA fingerprinting.

Cells from a single colony formed on TSA incubated at 30°C for 18 to 20 h were inoculated into 10 ml of TSB and TSB supplemented with 5% ethanol. Inoculated TSB and TSB containing 5.0% ethanol were incubated at 30°C for 22 and 30 h, respectively. Cultures were centrifuged (1,000 × g, 5 min), and cells were washed in potassium phosphate-buffered saline (PBS, pH 7.4). The PFGE analysis procedure described by Brown et al. (4) was followed, with minor modifications. Briefly, the cells were washed three times in PBS, resuspended in 1.0 ml of PBS, and diluted 10-fold in PBS. The cell suspension was mixed with 25 μl of proteinase K (Bethesda Research Laboratory, Gaithersburg, Md.) and incubated at 37°C for 5 min before 1.0 ml was mixed with 0.5 ml of 1.0% PFGE-grade agarose solution in Tris-EDTA buffer (previously heated to completely solubilize and cooled to 55°C) and deposited in molds (Bio-Rad) for sample plug preparation. The agarose plugs were digested with proteinase K and XbaI (Bethesda Research Laboratory). After incubating the samples at 37°C for at least 3 h, plugs were embedded in 1.2% agarose gel in 0.5× Tris-borate-EDTA buffer and subjected to pulsed electrophoresis with a contour-clamped homogenous electric field (CHEF Mapper; Bio-Rad) for 18 h at 14°C, 6 V/cm, ±60^o angle, and 2.5 to 55 s of switch time. The gels were stained with ethidium bromide, and bands were visualized and photographed with a UV transilluminator. A reference strain of E. coli O157:H7, G5244, provided by the Centers for Disease Control and Prevention, Atlanta, Ga., was run concurrently for comparison of patterns.

RESULTS

Growth and change in pH as affected by ethanol.

Changes in the turbidity and pH of inoculated TSB containing 0, 2.5, 5.0, 7.5, and 10% ethanol and incubated at 30°C for 54 h are shown in Fig. 1. The growth rates of both strains decreased with increases in ethanol concentration up to 5.0%. The exponential growth rates calculated from the slopes of the growth curves decreased with each increase in ethanol concentration. E. coli O157:H7 did not grow in TSB containing 7.5 or 10% ethanol.

FIG. 1.

Changes in turbidity (A and B) and pH (C and D) of TSB supplemented with 0, 2.5, 5.0, 7.5, and 10% ethanol and inoculated with E. coli O157:H7 strain SEA 13B88 (A and C) and strain 30-2C4 (B and D). ○, 0% ethanol; ▪, 2.5% ethanol; ▴, 5.0% ethanol; □, 7.5% ethanol; ▵, 10% ethanol.

The pHs of inoculated TSB supplemented with 0, 2.5, and 5.0% ethanol decreased from 7.1 to below 6.2 during the first 12 h of incubation (Fig. 1C and D). The lowest pH (5.8) recorded was in TSB containing 5.0% ethanol. When cultures containing 2.5 or 5.0% ethanol were incubated for more than 12 h, pHs gradually increased. The pH of TSB not supplemented with ethanol was 8.1 after 54 h. A close correlation between the decrease in pH and the rate of cell growth occurred during the early stages of incubation. A lack of change in the pH of TSB supplemented with 7.5 or 10% ethanol was observed throughout the 54-h incubation period. Glucose, at an initial concentration of 0.25% in TSB, was fermented, causing the pH to decrease. When glucose and its intermediate metabolites were depleted, the carbon source was shifted to proteins. Breakdown of proteins resulted in alkaline metabolites that increased the pH of the TSB in the late hours of incubation.

The ethanol content of TSB initially containing 0, 2.5, 5.0, 7.5, and 10% ethanol, inoculated with E. coli O157:H7, and incubated for up to 48 h at 30°C is shown in Table 1. The initial concentration of 10% ethanol in TSB remained constant for 48 h. Since E. coli O157:H7 did not grow in TSB containing 10% ethanol (Fig. 1A and B) and the ethanol concentration did not change during incubation for 48 h, it is concluded that evaporation of ethanol was minimal. Ethanol was not detected in TSB not supplemented with ethanol, indicating that E. coli O157:H7 did not produce ethanol. The ethanol content in TSB supplemented with 2.5, 5.0, or 7.5% ethanol did not change in the first 24 h of incubation but decreased slightly between 24 and 48 h. This indicates that small amounts of ethanol may have been utilized for growth or converted to less-toxic metabolites, allowing the cells to adapt to the stress that it imposed.

TABLE 1.

Changes in ethanol content in TSB and in ethanol-supplemented TSB inoculated with E. coli O157:H7 and incubated at 30°C

Incubation time (h)	Ethanol concn (%, vol/vol) in TSBa
	0	2.5	5.0	7.5	10
0	ND/ND	2.5/2.5	5.0/5.0	7.5/7.5	10/10
12	ND/ND	2.5/2.5	5.0/5.0	7.5/7.5	10/10
24	ND/ND	2.5/2.5	5.0/5.0	7.5/7.5	10/10
36	ND/ND	2.1/2.2	4.5/5.0	6.8/7.4	10/10
48	ND/ND	1.9/2.2	4.4/4.4	6.5/7.2	10/10

^aValues are presented as E. coli O157:H7 SEA 13B88/E. coli O157:H7 30-2C4. ND, none detected.

When TSB was supplemented with ethanol at concentrations allowing growth of both strains, i.e., 0, 2.5, and 5.0%, glucose was depleted within 12 h. The rapid depletion was coincident with rapid cell growth and a rapid decrease in the pH (Fig. 1). Although there was no apparent growth in TSB supplemented with 7.5% ethanol (Fig. 1A and B), the glucose level decreased slightly within 48 h of incubation. In agreement with the observation that the pH decreased slightly within 48 h in TSB supplemented with 7.5% ethanol, both test strains, comparatively more obvious for E. coli O157:H7 strain SEA 13B88 (Fig. 1A) than for strain 30-2C4 (Fig. 1B), appeared to have adapted to the stress caused by the ethanol. In TSB supplemented with 10% ethanol, the glucose content did not decrease during 48 h of incubation, which concurs with the lack of growth of E. coli O157:H7 (Fig. 1A and B).

Cellular fatty acid composition.

The cellular fatty acid composition of E. coli O157:H7 grown in TSB for 22 h and in TSB supplemented with 5.0% ethanol for 30 h were, respectively, 3.50 and 4.70% (wt/wt) for E. coli O157:H7 strain SEA 13B88 and 3.84 and 4.74% (wt/wt) for E. coli O157:H7 strain 30-2C4 (Table 2). The total amount of fatty acids was significantly higher (α = 0.05) in the cells grown under ethanol stress compared to control cells.

TABLE 2.

Cellular fatty acid profiles and crude protein contents of E. coli O157:H7 grown at 30°C in TSB for 22 h and in TSB supplemented with 5.0% ethanol for 30 h

Fatty acid or protein	E. coli O157:H7 straina
	SEA 13B88		30-2C4
	Without ethanol	With 5.0% ethanol	Without ethanol	With 5.0% ethanol
Fatty acid (mg/g of freeze-dried cells)
Dodecanoate (n-12:0)	1.35 ± 0.10	2.22 ± 0.14	1.51 ± 0.09	2.54 ± 0.05
Tetradecanoate (n-14:0)	4.28 ± 0.18	4.66 ± 0.51	4.95 ± 0.45	5.15 ± 0.57
3-Hydroxytetradecanoate (3-OH 14:0)	2.50 ± 0.41	3.88 ± 0.62	2.99 ± 0.24	4.05 ± 0.16
cis-9-Hexadecenoate (cis-16:1, 9)	0.29 ± 0	2.13 ± 0.32	0.33 ± 0.02	1.68 ± 0.64
Hexadecanoate (n-16:0)	11.77 ± 0.37	15.00 ± 2.46	13.05 ± 0.98	15.61 ± 0.79
Heptadecanoate (n-17:0)	8.14 ± 0.21	9.02 ± 1.56	8.51 ± 0.67	9.28 ± 0.68
cis-9-Octadecenoate (cis-18:1, 9)	0.32 ± 0.04	ND	0.36 ± 0.06	ND
trans-9-Octadecanoate (trans-18:1, 9)	0.40 ± 0.02	5.54 ± 1.10	0.50 ± 0.02	4.55 ± 0.22
Nonadecanoate (n-19:0)	5.90 ± 0.32	4.50 ± 0.79	6.20 ± 0.41	4.50 ± 1.00
Total	34.95 (3.50%)	46.95 (4.70%)	38.40 (3.84%)	47.36 (4.74%)
Crude protein content (% [wt/wt] N × 6.25, freeze-dried cells)	82.81 ± 1.32	86.94 ± 0.93	88.44 ± 1.02	90.62 ± 1.22
Protein content in extracted cell lysates (mg/ml)	4.04 ± 0.16	6.82 ± 0.32	4.05 ± 0.15	5.45 ± 0.45

^aValues are means of two determinations ± the range. ND, none detected.

A comparison of the fatty acid content of cells grown in TSB and in TSB supplemented with ethanol revealed that, with the exceptions of cis-9-octadecenoate (cis-oleic acid, n-18:1, 9) and nonadecanoate (n-19:0), the amounts of most species were higher or unchanged in cells grown under ethanol stress. A comparison of cells grown in TSB with and without ethanol supplementation indicates that the trans-oleic acid (trans-9-octadecenoate) contents were 5.54 and 0.40 mg, respectively, per g of lyophilized E. coli O157:H7 strain SEA 13B88 cells and 4.55 and 0.50 mg, respectively, per g of lyophilized E. coli O157:H7 strain 30-2C4 cells. In contrast, cis-oleic acid was not detected in cells of the two strains grown under ethanol stress but was present at concentrations of 0.32 and 0.36 mg/g of lyophilized cells of strain SEA 13B88 and strain 30-2C4grown in TSB without ethanol, respectively,. A cis to trans isomerization of oleic acid was apparently induced in the cells grown under ethanol stress.

A comparison of the saturated fatty acid contents of E. coli O157:H7 cells grown in TSB with and without ethanol revealed that the nonadecanoic acid (n-19:0) content was relatively lower in the cells grown under ethanol stress (Table 2). With this exception, however, upon exposure to ethanol, inhibition of saturated fatty acid synthesis in E. coli O157:H7 was not observed. With the exception of nonadecanoic acid, the amounts of other saturated fatty acids with shorter chain lengths were not reduced in TSB supplemented with ethanol. Significantly increases in cis-9-hexasecenoate (cis-16:1, 9) and trans-9-octadecanoate (trans-18:1, 9) contents were observed in cells grown in TSB containing 5% ethanol compared to cells grown in TSB without ethanol. These shifts in the fatty acid profile may result in changes in the sensitivity of cells upon exposure to subsequent stress.

Protein content and SDS-PAGE patterns as affected by ethanol.

The protein contents of the freeze-dried E. coli O157:H7 cells were slightly higher in cells grown in TSB containing 5.0% ethanol than in cells grown in TSB without ethanol (Table 2). Protein concentrations in extracted cell lysates ranged from 4.05 to 6.82 mg/ml. For both strains, cells grown under ethanol stress yielded higher amounts of protein in lysate extracts. Since equal volumes of the cell lysates were loaded on gels for analysis, the intensity (quantity) of the protein bands was generally proportional to total cell protein content and protein content in the lysates shown in Table 2. More than 40 protein bands were observed by mini-gel electrophoresis of extracts of both strains (Fig. 2). The patterns of proteins with larger than 45 kDa were qualitatively different for E. coli O157:H7 strains SEA 13B88 and 30-2C4 (lanes 1 and 3, respectively) grown in TSB without ethanol. A comparison of cells of the two strains grown in TSB with and without 5.0% ethanol (lane 1 versus lane 2 for strain SEA 1388 and lane 3 versus lane 4 for strain 30-2C4) shows a unique protein band with an estimated size of 28 kDa in ethanol-stressed cells (indicated with arrows).

FIG. 2.

Cellular SDS-PAGE protein patterns of E. coli O157:H7. Lane M, low-molecular-weight protein markers; lane 1, E. coli O157:H7 SEA 13B88 grown in TSB without ethanol; lane 2, E. coli O157:H7 SEA 13B88 grown in TSB supplemented with 5.0% ethanol; lane 3, E. coli O157:H7 30-2C4 grown in TSB without ethanol; lane 4, E. coli O157:H7 30-2C4 grown in TSB supplemented with 5.0% ethanol. The arrows indicate a protein band with an estimated size of 28 kDa in cells grown in TSB supplemented with 5.0% ethanol but not in cells grown in TSB without ethanol.

Comparison of genomic DNA by PFGE.

The PFGE DNA fingerprint patterns of E. coli O157:H7 grown in TSB with and without 5.0% ethanol revealed 19 DNA fragments of various sizes in the reference strain (G5244) of E. coli O157:H7 (Fig. 3, lane 1). The genomic DNA fingerprint patterns of E. coli O157:H7 strain SEA 13B88 (lane 2) and E. coli O157:H7 strain 30-2C4 (lane 4) grown in TSB without ethanol were slightly different from that of the reference strain (lane 1) and also slightly different from each other, indicating that the three strains were not genotypically identical. A comparison of the DNA patterns of cells grown in TSB with that of cells grown in TSB containing 5.0% ethanol (lane 2 versus lane 4 for strain SEA 13B88 and lane 3 versus lane 5 for strain 30-2C4) revealed no difference caused by the presence of ethanol in TSB. This was not unexpected, because the inherent genomic DNA of each strain is unlikely to be altered by ethanol stress.

FIG. 3.

PFGE genomic DNA patterns of E. coli O157:H7. Lane 1, E. coli O157:H7 strain G5244 (reference strain); lane 2, E. coli O157:H7 strain SEA 13B88 grown in TSB without ethanol; lane 3, E. coli O157:H7 strain 30-2C4 grown in TSB without ethanol; lane 4, E. coli O157:H7 SEA 13B88 grown in TSB supplemented with 5.0% ethanol; lane 5, E. coli O157:H7 30-2C4 grown in TSB supplemented with 5.0% ethanol.

DISCUSSION

The inability of E. coli O157:H7 to grow in TSB containing 7.5 or 10% ethanol is in agreement with a report by Huang et al. (10), who observed that E. coli and E. coli O157:H7 were not able to grow in nutrient broth supplemented with 7.5% ethanol. Ethanol is an amphiphilic compound that functions as a general membrane perturbant to repress cell growth. Reduction of growth rate, prevention of cell division, and cell death of a wild-type strain of E. coli K-12, E102, exposed to ethanol have been reported (8).

Membrane fatty acid content and composition changed in E. coli O157:H7 as an adaptation response to the presence of ethanol. The effects of alcohols of various chain lengths on the fatty acid composition of E. coli K-12 were investigated by Ingram (11) and Sullivan et al. (24). They observed that fatty acid composition was radically altered when cells were grown in media containing alcohols. These adaptive changes compensate for the direct physicochemical interaction of alcohols with the cell membrane. The higher fatty acid content in ethanol-stressed cells observed in our study may result in less polarity of cell membranes, i.e., an increase in cell hydrophobicity, to compensate for the change in the polarity of TSB resulting from supplementation with ethanol.

This is the first report showing that cis-trans isomerization of oleic acid occurred in stressed E. coli O157:H7 cells. cis-trans isomerization of fatty acids has been observed to occur in Pseudomonas putida as a defense mechanism induced by heat, toluene, and alcohols with different chain lengths (16, 25, 26). Holtwick et al. (9) stated that cis-trans isomerization of membrane fatty acids in P. putida increased the rigidity of the cell envelope, thereby decreasing permeability, a known adaptation mechanism for bacteria grown under diverse environments. cis-trans isomerization of cellular fatty acids is one of the various novel mechanisms that enable cells to adapt and thrive in the presence of toxic organic solvents (19, 20, 25). In our study, cis-trans isomerization of oleic acid (18:1, 9) but not hexadecenoic acid (16:1, 9) was observed (Table 2). This is contrary to observations reported by Weber et al. (26) on P. putida cells, in which both trans-oleic acid and trans-hexadecanoic acid contents were higher in cells grown under alcohol stress.

Our observation that nonadecanoic acid (n-19.0) was lower in E. coli O157:H7 cells grown under ethanol stress tends to support observations on changes in the ratio of unsaturated to saturated fatty acids in ethanol-stressed E. coli (5, 11, 12) showing that growth of E. coli in the presence of ethanol results in extensive synthesis of lipids containing elevated amounts of unsaturated fatty acids. Ethanol acted at the level of fatty acid biosynthesis to alter lipid composition by decreasing the proportion of saturated acyl chains available for synthesis of phospholipids.

The 28-kDa protein detected in ethanol-stressed E. coli O157:H7 cells is identical to phage shock protein A (pspA) reported by Brissette et al. (3) and Kobayashi et al. (14). The phage shock operon (pspABCE) in E. coli is strongly expressed in response to environmental stresses such as heat shock, ethanol treatment, osmotic shock, filamentous phage infection, and exposure to n-hexane or cyclooctane (3, 14). The appearance of the phage shock protein (psp) operon in the late stationary phase of E. coli cells exposed to alkaline pH and protection of the cells against environmental challenge conditions have been demonstrated (18). Bacteria lacking the pspABC genes exhibit a substantial decrease in the ability to survive in stationary phase under alkaline conditions (pH 9) (27). In our study, the production of a 28-kDa protein by two strains of E. coli O157:H7 is interpreted as an adaptation response that may enable the cells to grow under the stress caused by ethanol.

As a general conclusion, the two E. coli O157:H7 strains originally isolated from apple juice and salami belong to the same serotype, but their genotypic DNA PFGE fingerprint patterns, phenotypic SDS-PAGE protein profiles, and cellular fatty acid composition are slightly different. However, both strains responded similarly to stress imposed by ethanol, as evidenced by changes in cellular fatty acid composition and protein profiles. Both strains had similar growth patterns when cultured in TSB containing 2.5 or 5.0% ethanol. The need for a better understanding of the adaptive mechanisms exhibited by E. coli O157:H7 under ethanol stress and their potential for inducing cross protection of the pathogen against heat, solvents, acid pH, reduced a_w, and other stress conditions deserves further research.