Increased ATPase Activity Is Responsible for Acid Sensitivity of Nisin-Resistant Listeria monocytogenes ATCC 700302

Abstract

The growth of the foodborne pathogen Listeria monocytogenes can be controlled by nisin, an antimicrobial peptide. A spontaneous mutant of L. monocytogenes shows both resistance to nisin and increased acid sensitivity compared to the wild type. Changes in the cell membrane correlated with nisin resistance, but the mechanism for acid sensitivity appears unrelated. When hydrochloric or lactic acid is added to cultures, intracellular ATP levels drop significantly in the mutant (P < 0.01) compared to the results seen with the wild type. Characterization of the F₀F₁ ATPase, which hydrolyzes ATP to pump protons from the cell cytoplasm, shows that the enzyme is more active in the mutant than in the wild type. These data support a model in which the increased activity of the mutant ATPase upon acid addition depletes the cells' supply of ATP, resulting in cell death.

The natural preservative nisin can be effectively used to control the growth of Listeria monocytogenes, a foodborne pathogen. The psychrotrophic pathogen is capable of multiplying at refrigeration temperatures, altering its membrane fluidity by decreasing the length and degree of saturation of the membrane fatty acids (8). The presence of L. monocytogenes in hot dogs, smoked salmon, cheese, and ready-to-eat chicken and turkey, for example, has resulted in recalls of these products owing to the zero-tolerance standards for the organism in ready-to-eat foods (U.S. Food and Drug Administration, 2001 [www.fda.gov/oc/po/firmrecalls/archive.html]; U.S. Department of Agriculture Food Safety Inspection Service, 2001 [www.fsis.usda.gov/OA/recalls/recdb/rec2000.htm]).

The mode of action of nisin has been extensively researched. As a Class Ia bacteriocin, its bactericidal action results from the formation of transient pores in the target cell membranes (1, 21). Lipid II, a potential docking molecule for nisin, appears to enhance the activity of the 34-amino-acid peptide (4).

Nisin resistance is well documented, but the exact mechanism of nisin resistance remains to be elucidated. The frequency of nisin resistance ranges from 10⁻² to 10⁻⁷ but can be decreased by exposing cells to an acidic environment low in salt at 10°C (10). Several authors have correlated nisin resistance with a change in membrane composition. Others have noted the up or down regulation of proteins with functions that do not have a clear relationship to nisin resistance (11). Davies et al. (7) concluded that mutations causing nisin resistance were the result of “mechanism(s) other than an alteration of membrane composition,” citing cell wall modifications as a probable cause. However, Crandall and Montville (6) observed changes in the composition of the nisin-resistant cell membrane when cells were grown in the presence of nisin. Lipid analysis of membranes from nisin-resistant cells shows that there is an increase in the proportion of straight-chain and saturated fatty acids and that the ratio of C15:C17 is decreased, resulting in a more rigid membrane (14).

Our laboratory has found that a spontaneous nisin-resistant L. monocytogenes strain had increased sensitivity to acid (15). After finding that the changes in the membrane of the nisin-resistant strain did not alter the permeability to acid, we hypothesized that differences in acid sensitivity were due to differences in the energy status of the cells. Here we show that although there is no difference in acidification of the cytoplasm, the ATPase activity of the mutant is greater than that of the wild type under acidic conditions. Upon exposure to acid (either organic or inorganic), intracellular ATP levels rapidly drop. This suggests that when the nisin-resistant strain is treated with acid, the ATPase depletes cellular ATP levels and the cells are unable to survive.

MATERIALS AND METHODS

Bacterial strains and growth conditions.

L. monocytogenes Scott A (wild type) and NR 30 (ATCC 700302), the spontaneous nisin-resistant mutant of Scott A, were grown overnight at 30°C with agitation in brain heart infusion broth (BHI) (pH 6.5) buffered with 100 mM potassium phosphate. L. monocytogenes NR 30 was always grown with 200 IU of nisin (Novasin [a gift from Rhodia, Madison, Wis.] or pure nisin [a gift from Chr. Hansen, Horsholm, Denmark]) ml⁻¹.

ATP assay.

The determination of total and extracellular levels of ATP was accomplished using an ATP bioluminescence assay kit from Sigma and following the procedure of McEntire et al. (15). Briefly, cells were washed in 50 mM 2-(N-morpholino)ethanesulfonic acid (MES; pH 6.5) and resuspended to A₆₀₀ = 0.6. The washed cells were resuspended in MES and energized with 0.2% glucose for 20 min at 30°C. In one experiment, equimolar concentrations of hydrochloric (HCl) or lactic acid (40 mM) were added to 0.6 ml of energized cells. The final pH of the cell mixture was 4.3 ± 0.1. In another experiment, lactic acid was used to adjust the pH to 4.9. Total and extracellular ATP levels were measured before acid addition (time, −1 min) and at intervals up to 20 min after treatment. Total ATP levels were determined by adding 20 μl of treated cells to 80 μl of dimethyl sulfoxide, adding 5 ml of sterile water, and using 100 μl of the mixture in the assay with 100 μl of enzyme mixture. For determination of extracellular ATP levels, 100 μl of treated cells was added to a test tube. A total of 5 ml of sterile water was added, and a 100-μl sample was added to 100 μl of enzyme mixture. Intracellular ATP levels were calculated as the difference between the total and extracellular ATP levels at each time point. To standardize between experiments, cell dry weight was determined by drying 1 ml of cell suspension in saline solution in a preweighed aluminum dish overnight at 45°C. Each experiment was repeated at least three times, and the Student t test was used to determine significant differences after 20 min.

In vivo acid sensitivity.

Exponential phase cells (A₆₀₀ = 0.28) were centrifuged, concentrated 10-fold, and then diluted 100-fold in BHI at pH 6.5, BHI at pH 3.0, BHI at pH 3.0 plus 1 mM 1,3-dicyclohexylcarbodiimide (DCCD) (Sigma) (dissolved in ethanol as a 100× solution), or BHI at pH 3.0 plus 1% ethanol (as a solvent control for DCCD). Cells were incubated at room temperature for 1 h before enumeration by serial dilution in 0.1% peptone water and plating on BHI agar. Colonies were counted after 48 h at 30°C.

Isolation of membrane vesicles.

Cell membranes were collected on the basis of the method of Hicks and Krulwich (12); the method was modified for L. monocytogenes. Early-stationary-phase (A₆₀₀ = 1.2) cells were centrifuged at 9,500 × g for 10 min at 4°C and resuspended in osmotic buffer (0.25 M sucrose, 50 mM Tris HCl [pH 6.5], 200 mM KCl, 10 mM MgCl₂, 5 mM 3-aminobenzamide, 1 mM phenylmethylsulfonyl fluoride, DNase and RNase [0.3 mg each/liter of growth medium]). Cell pellets were frozen at −70°C until use. Cells (in a total volume of 40 ml) were broken in a precooled French pressure cell at 18,000 lb/in² and subsequently incubated at 30°C for 10 min. Broken cells were centrifuged twice at 9,500 × g for 10 min at 4°C to remove unbroken cells and debris. The membranes were pelleted by ultracentrifugation in a type Ti60 rotor at 100,000 × g for 90 min at 4°C. The pellet was washed with French press buffer minus sucrose and centrifuged for 60 min. A buffer consisting of 20 mM Tris HCl (pH 6.5), 40 mM KCl, 10 mM MgCl₂, 5 mM 3-aminobenzamide, and 0.2 mM phenylmethylsulfonyl fluoride was used for the final 60-min wash. Membranes were resuspended in 1 ml of this buffer and stored in 100-μl aliquots at −70°C.

Characterization of ATPase activity.

A method described by Hicks (12) and based on the method of Lebel et al. (13) was modified to determine the ATPase activity of L. monocytogenes. The reaction mixture consisted of 50 mM Tris HCl (adjusted to the desired pH), MgCl₂ (up to 30 mM), and Na₂ATP (pH 7.0) (up to 10 mM). For pH optimum experiments, 20 mM MgCl₂ and 10 mM ATP were used. To determine the magnesium requirement of the enzyme at pH 5.0, 5 mM ATP was used. Kinetic studies were performed at pH 5.0 with 20 mM MgCl₂. To ensure that activity was attributable to the F₀F₁ ATPase, up to 1 mM of the specific inhibitor DCCD was added. To initiate the reaction, 10 μl of thawed membranes was added to the reaction mixture and the tube was vortexed and incubated at 30°C for 30 min. The reaction was stopped by the addition of 50 μl of 50% trichloroacetic acid, and the precipitate was removed by centrifugation. The supernatant was added to an equal volume of the detection solution. The detection solution was a 1.6:2:2 ratio of LeBels reagent (3.6 M acetic acid, 0.66 M sodium acetate, 20 mM copper sulfate), ammonium molybdate (5% stock solution), and a 5% sodium sulfite-2% p-methylaminophenyl sulfate solution. Color development proceeded for exactly 5 min at room temperature before absorbance at 750 nm was recorded. A standard curve was obtained by adding potassium phosphate to the reaction mixture. For the blank and standards, membranes were not added until after trichloroacetic acid addition. Activity is expressed as micromoles of phosphate minute⁻¹ mg of protein⁻¹. The means and standard deviations of triplicate experiments are presented. EZ Fit software (Perrella Scientific) was used for the analysis of kinetic data.

Protein concentration was determined by the method of Bradford (3) and using Bio-Rad's protein assay dye reagent concentrate according to the manufacturer's directions. Bovine serum albumin (stock, 1 mg/ml) was used as the standard. Color development after exactly 5 min was measured at 570 nm in a Dynex 96-well plate reader.

RESULTS AND DISCUSSION

Effect of acid on ATP levels.

Previous work showed that as pH decreases, L. monocytogenes NR 30, the nisin-resistant mutant, grows more slowly than the wild type until no growth of the mutant is observed when the pH is adjusted below 4.9 with lactic acid (14). Since L. monocytogenes NR 30 is more sensitive to acid than the Scott A strain, the ATP assay was used to study the effect of HCl and lactic acid on cellular ATP levels. The addition of HCl to Scott A cells had very little effect on ATP levels (Fig. 1). The decrease in intracellular ATP levels of Scott A in the presence of lactic acid was not statistically significant at any time point. At 20 min after acid exposure, L. monocytogenes NR 30 experienced a significant (P < 0.05), decrease in intracellular ATP levels upon addition of either acid (Fig. 1). The effect was more pronounced for lactic acid than for HCl. Even though the level of the control for strain NR 30 appears higher, there is no significant difference between the final intracellular ATP levels of the strains in comparison to each other when they were treated with lactic acid at pH 4.3. Neither strain is capable of growth at this pH. The Scott A strain, but not the NR 30 strain, can replicate at pH 4.9 (15). After 20 min at pH 4.9, the intracellular ATP level of the NR 30 strain was significantly lower than that of the Scott A strain (P < 0.01) (Fig. 2). While the levels of extracellular ATP increased as the duration of acid treatment increased, these did not wholly account for the changes in intracellular ATP, indicating that ATP was also being hydrolyzed. These data show that at acid levels at which the wild-type but not the mutant strain can grow, there is greater ATP depletion in the mutant than in the wild type. This suggests that the depletion of energy is the cause of growth inhibition in the mutant strain.

FIG. 1.

Changes in intracellular (closed symbols) and extracellular (open symbols) ATP levels of the wild-type Scott A strain (A) and the nisin-resistant mutant NR 30 strain (B). Circles indicate controls, triangles represent the addition of HCl, and squares represent the addition of lactic acid. Addition of both acids resulted in a final pH of 4.3. Significant (P < 0.05) differences in intracellular ATP levels of all six conditions are indicated by different letters.

FIG. 2.

Effect of lactic acid treatment (triangles; pH 4.9) on the Scott A (A) and NR 30 (B) strains. Circles represent untreated cells. Closed symbols represent intracellular ATP, and open symbols represent extracellular ATP. Significant (P < 0.05) differences in intracellular ATP levels of the four conditions are indicated by different letters.

In vivo acid sensitivity and contribution of the ATPase.

To show the acid sensitivity of strain NR 30 and to assess the contribution of the F₀F₁ ATPase in the response of the cells to acid, cells were exposed to pH 3.0 with or without the specific inhibitor DCCD. The inhibitor blocks the proton channel of F₀ by covalently binding to amino acid that is the H⁺ binding site in the c subunit (known to be Glu-54 in Propionigenium modestum and Asp-61 in Escherichia coli) (5, 9). The inability of the cell to pump protons leads to increased cell death for both the wild-type and mutant strains (data not shown). Acid alone causes a reduction in the population of the wild-type Scott A strain of less than 1 log, while the nisin-resistant mutant NR 30 strain experiences a 2.7-log decrease in counts. DCCD treatment results in an additional reduction in population for the strains of 2 to 2.5 log, showing that the F₀F₁ ATPase is necessary for cell survival.

ATPase activity of wild-type and mutant cells.

Lactic acid bacteria use substrate-level phosphorylation to generate ATP. Since the cells lack an electron transport chain and do not generate ATP through oxidative phosphorylation, the primary function of the F₀F₁ ATPase is to control intracellular pH by pumping protons (5). Because of its role in maintaining a near-neutral cytoplasmic pH in an acidic environment, the enzyme has been studied in bacteria important in oral care (2, 18, 19). There are no data on the kinetics of ATPase activity in L. monocytogenes. It has a high degree of sequence homology to other organisms (5). Cotter et al. used degenerate primers to locate the alpha, beta, and gamma subunits of L. monocytogenes strains with more than 70% similarity with Bacillus species (5).

O'Sullivan and Condon showed a relationship between ATPase activity and intracellular ATP levels of Lactococcus lactis subsp. cremoris NCDO 712. When cells were grown in a chemostat, the cytoplasmic pH could be lowered to increase ATPase activity with the concurrent loss of intracellular ATP (16). Since total ATP levels in strain NR 30 dropped upon addition of acid, the activity of the membrane-bound ATPase was measured. The maximum detection of phosphate in both strains was obtained at pH 5.0 and decreased as the pH was raised until virtually no activity was detected at pH 8.0 (Fig. 3). Little activity was measured at pH 3.0. At pH 5.0, the ATPase activity of the nisin-resistant mutant was more than threefold higher than that of the wild-type strain.

FIG. 3.

Influence of pH on ATP hydrolysis when the reaction mixture included 10 mM ATP and 20 mM MgCl₂. Closed circles represent strain Scott A. Open circles represent strain NR 30. Error bars represent 1 standard deviation.

Many fermentative organisms show optimal ATPase activity at more-neutral pH compared to L. monocytogenes (18, 19). For L. lactis subsp. cremoris NCDO 712, the highest ATPase activity was observed at pH 5.5, the lowest pH tested (16). The relationship between activity and pH determined for L. monocytogenes Scott A seems to be most similar to data reported for L. casei ATCC 4644, an aciduric organism (2). Studies of the ATPase in other organisms show that the enzyme requires magnesium as a cofactor. The magnesium concentration required for optimal activity was 20 mM MgCl₂ when the ATP concentration was held constant at 5 mM (data not shown). This concentration was used for kinetic studies.

To ensure that the release of phosphate was the result of the activity of the F₀F₁ ATPase and not of that of other components of the membrane preparations, the effect of increasing concentrations of DCCD was determined. The decrease in activity observed when up to 1 mM DCCD was added confirmed that the F₀F₁ ATPase is responsible for activity (Fig. 4). DCCD did not inhibit 100% of activity even at the highest concentration tested, but this was not unexpected, as reports indicate that only 30 to 73% of F₀F₁ ATPase activity can be inhibited, depending on the organism and conditions (18). The maximum inhibition of the Scott A strain was 33%, and strain NR 30 activity was inhibited up to 39%, in similarity to results obtained by Sturr and Marquis for octylglucoside-extracted F₀F₁ complexes of L. casei 4646 (18).

FIG. 4.

DCCD inhibition of the Scott A (closed circles) or NR 30 (open circles) strain at pH 5.

Figure 5 shows that the apparent K_m values of ATP for the two strains were similar: 4.39 and 3.37 mM ATP for the wild-type and mutant strains, respectively. The apparent V_max value of the enzyme in the mutant NR 30 strain is almost double that of the wild-type strain (1.25 versus 0.69 μmol of phosphate min⁻¹ mg of protein⁻¹). There is a 2.4-fold increase in the specificity constant (V_max/K_m) of the mutant compared to the wild type. The increased catalytic efficiency and specificity constant of the mutant shows that the mutant has a more active F₀F₁ ATPase than the wild type. Very little work has been done to quantify and characterize the subunits of the ATPase in L. monocytogenes (5). In other organisms, particularly E. coli, the number of c subunits per ATPase has been shown to be variable, being greater when cells were grown on a fermentable carbon source where proton pumping would be required (17). The number of c subunits per E. coli ATPase is between 8 and 14. Between three and four protons are pumped per ATP synthesized, and the H⁺/ATP ratio can be manipulated by point mutations in the c subunit, but the exact numbers are unknown for enzymes working in the opposite direction (9, 17, 20). Work done using E. coli for examination of the biosynthetic capacity of the ATPase showed that ATP was synthesized more slowly in cells in which the enzyme had more c subunits compared to those having fewer subunits (17). A potential explanation of the increased ATPase activity of the mutant could be that the number of c subunits is decreased compared to the wild-type number. This could account for the increase in V_max and the depletion of ATP, since more ATP would be needed to pump the equivalent number of protons compared to the amount needed with the wild type. This model would be strengthened if the number of c subunits for each strain could be determined experimentally.

FIG. 5.

Kinetics of strain Scott A (A) and strain NR 30 (B) ATPase assayed at pH 5.0 with 20 mM MgCl₂. Error bars represent 1 standard deviation.

Upon exposure to acid (either organic or inorganic), intracellular ATP levels of the nisin-resistant mutant NR 30 strain rapidly drop. This is a result of the increased ATPase activity of the mutant. Kinetic analysis shows that while the affinity constants are similar, the catalytic efficiency of the mutant is higher. At pH 5.0, at which level maximum enzymatic activity is observed, L. monocytogenes NR 30 is more sensitive to acid than the wild type. While the ATPase hydrolyzes ATP to pump protons from the cytoplasm, the stoichiometry is variable (17). It is quite possible that Scott A has more c subunits in the ATPase and is therefore able to use ATP more efficiently to achieve the same internal pH. NR 30 may need to use more ATP to pump an equal number of protons; such an increase in the use of ATP results in the depletion of ATP and cell death. Further work is needed to support this hypothesis.