Combined Effect of High-Pressure Treatments and Bacteriocin-Producing Lactic Acid Bacteria on Inactivation of Escherichia coli O157:H7 in Raw-Milk Cheese

Abstract

The effect of high-pressure (HP) treatments combined with bacteriocins of lactic acid bacteria (LAB) produced in situ on the survival of Escherichia coli O157:H7 in cheese was investigated. Cheeses were manufactured from raw milk inoculated with E. coli O157:H7 at approximately 10⁵ CFU/ml. Seven different bacteriocin-producing LAB were added at approximately 10⁶ CFU/ml as adjuncts to the starter. Cheeses were pressurized on day 2 or 50 at 300 MPa for 10 min or 500 MPa for 5 min, at 10°C in both cases. After 60 days, E. coli O157:H7 counts in cheeses manufactured without bacteriocin-producing LAB and not pressurized were 5.1 log CFU/g. A higher inactivation of E. coli O157:H7 was achieved in cheeses without bacteriocin-producing LAB when 300 MPa was applied on day 50 (3.8-log-unit reduction) than if applied on day 2 (1.3-log-unit reduction). Application of 500 MPa eliminated E. coli O157:H7 in 60-day-old cheeses. Cheeses made with bacteriocin-producing LAB and not pressurized showed a slight reduction of the pathogen. Pressurization at 300 MPa on day 2 and addition of lacticin 481-, nisin A-, bacteriocin TAB 57-, or enterocin AS-48-producing LAB were synergistic and reduced E. coli O157:H7 counts to levels below 2 log units in 60-day-old cheeses. Pressurization at 300 MPa on day 50 and addition of nisin A-, bacteriocin TAB 57-, enterocin I-, or enterocin AS-48-producing LAB completely inactivated E. coli O157:H7 in 60-day-old cheeses. The application of reduced pressures combined with bacteriocin-producing LAB is a feasible procedure to improve cheese safety.

Escherichia coli O157:H7 is one of the foodborne pathogens of concern for the dairy industry. Raw milk has been identified as an important vehicle for the transmission of this pathogen, and milk and different dairy products, including cheeses, have been implicated in E. coli O157:H7 outbreaks (14, 27).

Survival of E. coli O157:H7 during the manufacture and ripening of some cheese varieties has been investigated. Numbers of E. coli O157:H7 increased during the manufacture of cottage cheese, but death occurred during heating of the curd at 57°C for 90 min (2). In cheddar cheese, E. coli O157:H7 decreased by only 2 log units after 60 days of ripening (28). This pathogen also survived the manufacture of Camembert and feta cheeses and reached counts higher than those present in milk after 75 and 65 days of storage, respectively (27). Maher et al. (19) observed that the manufacturing procedure of a raw-milk smear-ripened cheese encouraged substantial growth of E. coli O157:H7 to levels that permitted its survival during ripening.

High-pressure (HP) processing is a nonthermal method of food preservation which could represent an alternative to milk pasteurization in the manufacture of raw-milk cheeses. Pressurization of cheese affects the activity of enzymes involved in cheese ripening, potentially altering the flavor characteristics of the cheese (20). Treatment at 400 MPa for 5 min caused a reduction in casein degradation with respect to control cheese, which became more marked as cheese aged (30). Therefore, milder HP treatments would be preferable, provided that the microbiological quality of cheese was significantly improved.

Many factors affect the sensitivity of microorganisms to HP treatments, including the magnitude of the pressure, the pressurization time and temperature, the microbial species and strain, the cell growth phase, and the suspending medium. Differences in resistance to HP between different strains of E. coli O157:H7 in cell suspensions have been detected (1, 3). Linton et al. (18) reported differences in pressure resistance of E. coli strains in skimmed milk, observing, however, that O157:H7 strains were not more resistant to HP than other E. coli serotypes. Patterson et al. (25) showed that E. coli O157:H7 NCTC 12079 was more resistant to pressurization when treated in ultra-high-temperature milk than in poultry meat or buffer. Inactivation of E. coli, as high as 6.9 log units when treated at 400 MPa in phosphate buffer, was only 0.7 log units when treated in milk (8). According to Hauben et al. (13), calcium provides protection against pressure inactivation of E. coli, explaining the baroprotective properties of food substrates such as milk. However, HP treatments at 400 to 500 MPa for 5 to 15 min at temperatures of 2 to 25°C achieved reductions of 6.7 to 8.7 log units in the population of E. coli in Mató cheese (6).

According to the hurdle concept of food preservation, HP treatments can be used in combination with other preservation strategies to increase microbial lethality. The effect of HP treatments in combination with the lactoperoxidase system (9), lactoferrin and lactoferricin (21), lysozyme (12), and bacteriocins (12, 15, 17, 23), on different foodborne pathogens and spoilage bacteria has been investigated. The effects of pressurization parameters and of the addition of pediocin AcH on the lethality of pathogenic and spoilage bacteria were studied by Kalchayanand et al. (15), who recorded a higher reduction in the numbers of HP-injured E. coli O157:H7 cells in the presence of the bacteriocin. A disruption of the E. coli outer membrane after HP treatments of 180 to 320 MPa, causing sensitization to nisin, has been reported (12). Combinations of nisin with HP showed strong synergistic effects against E. coli (32), which were enhanced at low temperatures. Also, the lethality of E. coli was increased by nisin addition before HP treatments in milk (8) and in a meat model system (10). The combination of HP and lacticin 3147 was proposed by Morgan et al. (23) for improving the quality of minimally processed foods at lower hydrostatic pressure levels. Pressurization at 250 MPa (2.2 log unit reduction) combined with lacticin 3147 (1 log unit reduction) lowered Staphylococcus aureus counts by more than 6 log units in skimmed milk.

In this study, the antimicrobial effects of two HP treatments (300 and 500 MPa) on E. coli O157:H7 in cheese at two ripening times (2 days and 50 days) combined with different bacteriocin-producing lactic acid bacteria (LAB) added as adjuncts to the starter culture were investigated in order to determine synergisms in the inactivation of the pathogen.

MATERIALS AND METHODS

Microorganisms and culture conditions.

E. coli O157:H7 ATCC 43894 was grown on Tryptic soy broth (Biolife, Milan, Italy) at 37°C for 18 h. Lactococcus lactis subsp. cremoris TAB 24 (producing lacticin 481), L. lactis subsp. lactis TAB 26 (producing nisin Z), L. lactis subsp. lactis TAB 50 (producing nisin A), L. lactis subsp. lactis biovar diacetylactis TAB 57 (producing a noncharacterized bacteriocin, TAB 57), Enterococcus faecium TAB 7 (producing a noncharacterized bacteriocin, TAB 7), E. faecalis TAB 52 (producing enterocin I), and E. faecalis INIA 4 (producing enterocin AS-48) from the INIA (Instituto Nacional de Investigación y Tecnología Agraria y Alimentaria, Madrid, Spain) culture collection were used as bacteriocinogenic cultures in cheesemaking experiments. They were grown in MRS broth with Tween 80 (Biolife) at 30°C for 18 h. Clostridium tyrobutyricum NZ 8 (NIZO Food Research, Ede, The Netherlands), grown in Clostridium broth (Biolife) at 37°C for 48 h in anaerobic conditions, was used as the indicator strain in bacteriocin assays (29).

The strains were maintained in a frozen stock with 15% glycerol at −80°C. E. coli O157:H7 and bacteriocin-producing strains were subcultured twice in reconstituted skim milk supplemented with 0.3% yeast extract at 30°C for 16 h before use in cheesemaking.

Manufacture of cheese.

Cheeses were manufactured from raw milk in duplicate experiments carried out on two different days. In each experiment, raw milk at 32°C was distributed into eight 12-liter vats. Milk in all vats was inoculated with E. coli O157:H7 at approximately 10⁵ CFU/ml and with 0.1% (approximately 10⁶ CFU/ml) of a commercial mesophilic starter culture (MA 016; Texel, Dangé Saint-Romain, France). Seven vats were individually inoculated with 0.1% (approximately 10⁶ CFU/ml) of a culture of L. lactis TAB 24, L. lactis TAB 26, L. lactis TAB 50, L. lactis TAB 57, E. faecium TAB 7, E. faecalis TAB 52, or E. faecalis INIA 4. Rennet (0.015 g/liter, Maxiren 150; Gist-brocades, Delft, The Netherlands) was added to milk 20 min after inoculation. The curds were cut 40 min after rennet addition into 6- to 8-mm cubes and heated at 37°C for 25 min. Whey was drained off, and curds were distributed into plastic cylindrical molds. Five cheeses (approximately 260 g in weight) were obtained from each vat. Cheeses were pressed at 0.7 kg/cm² for 4 h, salted in 20% NaCl brine for 30 min, kept at 20°C for 16 h, vacuum packaged in Cryovac plastic bags, and ripened at 12°C for 60 days.

HP treatments.

The five cheeses obtained per vat were HP treated at 10°C as follows: (i) nonpressurized (NP) cheese; (ii) 2-day, 300-MPa cheese, pressurized on day 2 at 300 MPa for 10 min; (iii) 2-day, 500-MPa cheese, pressurized on day 2 at 500 MPa for 5 min; (iv) 50-day, 300-MPa cheese, pressurized on day 50 at 300 MPa for 10 min; and (v) 50-day, 500-MPa cheese, pressurized on day 50 at 500 MPa for 5 min.

HP treatments were performed at the Universitat Autònoma de Barcelona Pilot Plant with a discontinuous isostatic press from ACB Gec-Alsthom (Nantes, France). The pressure chamber was cooled at 10°C with a constant flow of ethylene glycol-water (1:1). Temperature of pressurization fluid was monitored with a thermocoupler.

Microbiological analysis.

Nonpressurized cheeses were sampled on days 3, 20, 51, and 60; cheeses pressurized on day 2 were sampled on days 3, 20, and 60, and cheeses pressurized on day 50 were sampled on days 51 and 60. Cheese samples (10 g) were homogenized in 90 ml of 2% (wt/vol) sterile sodium citrate solution with a Stomacher 400 (Seward Laboratory, London, United Kingdom), and decimal dilutions were prepared in sterile peptone (0.1%) water. Viable counts of E. coli O157:H7 were determined on duplicate plates of MacConkey sorbitol agar (Scharlau Chemie, Barcelona, Spain) with selective cefixime-tellurite supplement (Oxoid, Basingstoke, United Kingdom) after incubation at 37°C for 24 h. If necessary, E. coli O157:H7 enrichment from cheese samples (1 g) was made in 9 ml Tryptic soy broth with selective cefixime-tellurite supplement at 42°C for 6 h, followed by 48 h at 37°C. Enrichment broth was streaked after incubation at 37°C for 24 and 48 h on duplicate plates of MacConkey sorbitol agar with selective cefixime-tellurite supplement that were examined after incubation at 37°C for 24 h for characteristic colonies.

Cheese pH.

Cheese pH was measured in duplicate after 8 h and on days 3, 20, 51, and 60 directly with a penetration electrode (model 52-3,2; Crison Instruments S.A., Barcelona, Spain) by means of a Crison GPL 22 pH meter.

Detection of bacteriocin activity in cheese.

Cheese samples (5 g) were homogenized with 10 ml of sterile 0.02 N HCl at 50°C. Homogenates were centrifuged (12,000 × g, 20 min, 4°C) and supernatants frozen at −40°C in Eppendorf tubes. After thawing, pH of supernatants was adjusted to pH 6 with 1 N NaOH. Supernatants (25 μl) were placed in triplicate into wells (5-mm diameter) made in pour plates of Clostridium agar (Biolife) inoculated with 1% of a 16-h culture of C. tyrobutyricum NZ 8. After incubation at 37°C for 48 h in anaerobic conditions, the diameter of the zone growth inhibition was measured and bacteriocin activity expressed in millimeters.

Statistical analysis.

Multivariate analysis of variance with LAB culture, high-pressure treatment, days of ripening and cheesemaking experiment as main effects, and their interactions on E. coli O157:H7 counts and cheese pH was performed using the general linear model procedure of SPSS program Win 9.0 software (SPSS Inc., Chicago, IL). Significant differences in E. coli counts of cheese manufactured with different bacteriocin-producing LAB for each HP treatment and time of ripening were assessed by Tukey's test at P < 0.01 (31).

RESULTS

The population of E. coli O157:H7 in cheese was significantly (P < 0.001) influenced by the bacteriocin-producing LAB, the HP treatment, and the ripening time, according to the analysis of variance. E. coli O157:H7 counts on the day after pressure treatments and during the rest of the ripening period are summarized in Tables 1 and 2.

TABLE 1.

Survivors of E. coli O157:H7 (log CFU/g) in cheeses with bacteriocin-producing LAB pressurized on day 2^a

LAB	Bacteriocin	Nonpressurized			300 MPa at 10°C for 10 min			500 MPa at 10°C for 5 minb
		3 days	20 days	60 days	3 days	20 days	60 days	3 days	20 days	60 days
None		6.80C	6.22E	5.11D	5.50H	4.58D	3.18D	3.07C	ND	ND
L. lactis TAB 24	Lacticin 481	6.29B	5.63C	4.66AB	3.63B	3.63A	1.88B	NDA	ND	ND
L. lactis TAB 26	Nisin Z	6.31B	6.11DE	4.84BCD	3.83C	3.60A	2.96CD	NDA	ND	ND
L. lactis TAB 50	Nisin A	6.06A	5.58BC	4.66AB	4.23F	4.00B	1.81B	1.92B	ND	ND
L. lactis TAB 57	TAB 57	6.25AB	5.89D	4.59AB	4.04E	3.43A	0.90A	NDA	ND	ND
E. faecium TAB 7	TAB 7	6.41B	5.34B	4.25A	3.42A	3.98B	2.79C	NDA	ND	ND
E. faecalis TAB 52	Enterocin I	6.44B	6.21E	5.10CD	4.78G	4.34C	2.94CD	NDA	ND	ND
E. faecalis INIA 4	Enterocin AS-48	6.44B	5.02A	5.15D	3.97D	3.88B	1.94B	NDA	ND	ND

^aValues in the same column with different superscripts are significantly (P < 0.01) different.

^bND, absence of E. coli O157:H7 after enrichment in two 1-g samples.

TABLE 2.

Survivors of E. coli O157:H7 (log CFU/g) in cheeses with bacteriocin-producing LAB pressurized on day 50^a

LAB	Bacteriocin	Nonpressurized		300 MPa at 10°C for 10 minb		500 MPa at 10°C for 5 minc
		51 days	60 days	51 days	60 days	51 days	60 days
None		5.82D	5.11D	2.04C	+B	ND	ND
L. lactis TAB 24	Lacticin 481	5.40A	4.66AB	+A	+B	ND	ND
L. lactis TAB 26	Nisin Z	5.69C	4.84BCD	+A	+B	ND	ND
L. lactis TAB 50	Nisin A	5.56B	4.66AB	+A	NDA	ND	ND
L. lactis TAB 57	TAB 57	5.50AB	4.59AB	0.55B	NDA	ND	ND
E. faecium TAB 7	TAB 7	5.40A	4.25A	1.88C	+B	ND	ND
E. faecalis TAB 52	Enterocin I	5.92D	5.10CD	+A	NDA	ND	ND
E. faecalis INIA 4	Enterocin AS-48	5.51AB	5.15D	NDA	NDA	ND	ND

^aValues in the same column with different superscripts are significantly (P < 0.01) different.

^b+, presence of E. coli O157:H7 after enrichment in at least one of two l-g samples.

^cND, absence of E. coli O157:H7 after enrichment in two l-g samples.

The E. coli O157:H7 count in inoculated milk was on average 4.9 log CFU/ml. In NP cheese from milk not inoculated with bacteriocin-producing LAB, E. coli O157:H7 counts were 6.8 log CFU/g on day 3 and decreased to 5.1 log CFU/g on day 60 (Table 1).

Effect of bacteriocin-producing LAB.

The addition of bacteriocin-producing LAB to milk resulted in a slight inhibition of E. coli O157:H7 in NP cheeses on day 3 (Table 1), with counts 0.4 to 0.7 log units lower than those in cheeses without bacteriocin-producing LAB. On day 60, these differences were significant in cheeses with lacticin 481-, nisin A-, TAB 57-, and TAB 7-producing strains, with counts 0.5 to 0.9 log units lower than those in cheeses without bacteriocin-producing LAB. However, no significant differences were observed in NP cheeses with nisin Z-, enterocin I-, and enterocin AS-48-producing strains with respect to NP cheese without bacteriocin-producing LAB.

Effect of HP treatments.

Differences in E. coli O157:H7 counts between HP treatments (nonpressurized, 300 MPa and 500 MPa) for cheeses of the same age were in all cases (P < 0.01) significant (Tables 1 and 2). HP treatment of cheeses manufactured without bacteriocin-producing LAB at 300 MPa on day 2 (Table 1) resulted in a reduction of E. coli O157:H7 counts, with levels 1.3 log units lower on average than those in NP cheeses on day 3 and 1.9 log units lower on day 60. A higher inactivation of E. coli O157:H7 was achieved with HP treatment on day 2 at 500 MPa (Table 1), with a level 3.7 log units lower than that in NP cheese on day 3 and the complete inactivation of the pathogen from day 20 to the end of the ripening period.

Application of HP treatments on day 50 to cheeses without bacteriocin-producing LAB was much more effective than the application of HP treatments on day 2 (Table 2). HP treatment at 300 MPa on day 50 led to numbers of E. coli O157:H7 3.8 log units lower than those in NP cheeses on day 51 and to E. coli O157:H7 counts detected only after enrichment of 1 g on day 60. At 500 MPa, the elimination was complete on day 51, with no survivors in 60-day-old cheeses.

Combined effect of bacteriocin-producing LAB and HP treatments.

Pressurization at 300 MPa on days 2 and 50 and at 500 MPa on day 2 combined with bacteriocin-producing LAB resulted in a higher lethality of the pathogen than if HP was applied individually. Counts of E. coli O157:H7 in 2-day, 300-MPa cheeses were on day 3 (Table 1) significantly (P < 0.01) lower than those of pressurized cheese without bacteriocin-producing LAB. The highest rate of inactivation for HP treatments at 300 MPa was achieved in combination with bacteriocin TAB 7, resulting in a 3.4-log-unit reduction with respect to NP cheese without bacteriocin-producing LAB. This reduction was synergistic, as the combined effect of the two treatments (3.4 log units) was higher than the sum of reductions achieved by HP at 300 MPa (1.3 log units) and bacteriocin TAB 7 (0.4 log units). Synergistic effects were also observed on day 3 for the combinations of HP at 300 MPa with the other bacteriocin-producing LAB. The synergy persisted on day 60 only in cheeses with lacticin 481-, nisin A-, bacteriocin TAB 57-, and enterocin AS-48-producing LAB, in which the reduction achieved by the combined effect of HP at 300 MPa and bacteriocin was higher than the sum of reductions achieved by the individual effects. In some 2-day, 500-MPa cheeses, synergistic effects of the combination of HP and bacteriocins were also recorded. E. coli O157:H7 was completely inactivated on day 3 in all pressurized cheeses with bacteriocin-producing LAB, excepting those with nisin A-producing L. lactis TAB 50. On day 20, E. coli O157:H7 was not detected in any of 2-day, 500-MPa cheeses and the inactivation persisted to the end of the ripening period.

In 50-day, 300-MPa cheeses with bacteriocin-producing LAB (Table 2), E. coli O157:H7 was completely inactivated on day 51 in cheeses with enterocin AS-48-producing E. faecalis INIA 4 and detected after enrichment of 1 g in cheeses with lacticin 481-, nisin Z-, nisin A-, and enterocin I-producing LAB. Synergistic effects of HP treatment at 300 MPa and bacteriocins were recorded on day 51 for all cheeses, excepting that with E. faecium TAB 7. After 60 days, E. coli O157:H7 was completely inactivated in cheeses with nisin A-, bacteriocin TAB 57-, enterocin I-, and enterocin AS-48-producing LAB and detected after enrichment of 1 g in the rest of the cheeses. No survivors were detected in any of 50-day, 500-MPa cheeses after 51 and 60 days of ripening.

Cheese pH.

Cheese pH was significantly (P < 0.001) influenced by bacteriocin-producing LAB, HP treatment, and ripening time (Tables 3 and 4). Values of pH were between 5.00 and 5.27 on day 3 and between 5.09 and 5.35 at the end of the ripening period. After pressurization on day 2, an increase in cheese pH was recorded the day after treatment with respect to NP cheeses. Mean pH value (5.17) of 2-day, 300-MPa cheeses was on day 3 significantly (P < 0.01) higher than the mean pH value (5.07) of NP cheeses and the mean pH value (5.24) of 2-day, 500-MPa cheeses significantly (P < 0.01) higher than that of 2-day, 300-MPa cheeses. When pressurization was applied on day 50, the opposite phenomenon was observed, with a mean pH value (5.03) of 50-day, 300-MPa and 50-day, 500-MPa cheeses on day 51 significantly (P < 0.01) lower than the mean pH value (5.12) of NP cheeses.

TABLE 3.

Values of pH in cheeses with bacteriocin-producing LAB pressurized on day 2^a

LAB	Bacteriocin	Nonpressurized			300 MPa at 10°C for 10 min			500 MPa at 10°C for 5 min
		3 days	20 days	60 days	3 days	20 days	60 days	3 days	20 days	60 days
None		5.07BC	5.03D	5.22B	5.15A	5.17A	5.34D	5.23B	5.14B	5.09A
L. lactis TAB 24	Lacticin 481	5.04B	5.02C	5.25C	5.16AB	5.20C	5.30C	5.25D	5.14B	5.21C
L. lactis TAB 26	Nisin Z	5.08C	5.03D	5.24BC	5.16AB	5.28E	5.31C	5.27E	5.26E	5.26D
L. lactis TAB 50	Nisin A	5.08C	5.00A	5.22B	5.18ABC	5.32F	5.31C	5.23B	5.26E	5.23C
L. lactis TAB 57	TAB 57	5.09C	5.04E	5.28D	5.20BC	5.26D	5.35D	5.24C	5.24D	5.29D
E. faecium TAB 7	TAB 7	5.00A	5.01B	5.23BC	5.17ABC	5.18B	5.30C	5.23B	5.16A	5.15B
E. faecalis TAB 52	Enterocin I	5.10C	5.11F	5.24BC	5.20BC	5.35G	5.26B	5.27E	5.35F	5.34E
E. faecalis INIA 4	Enterocin AS-48	5.07BC	5.04E	5.15A	5.17ABC	5.29F	5.20A	5.22A	5.22C	5.27D

^aValues in the same column with different superscripts are significantly (P < 0.01) different.

TABLE 4.

Values of pH in cheeses with bacteriocin-producing LAB pressurized on day 50^a

LAB	Bacteriocin	Nonpressurized		300 MPa at 10°C for 10 min		500 MPa at 10°C for 5 min
		51 days	60 days	51 days	60 days	51 days	60 days
None		5.16B	5.22B	5.10G	5.22E	5.10E	5.28F
L. lactis TAB 24	Lacticin 481	5.09A	5.25C	5.01D	5.17B	5.04C	5.21C
L. lactis TAB 26	Nisin Z	5.13B	5.24BC	5.07F	5.21DE	5.09D	5.21C
L. lactis TAB 50	Nisin A	5.13B	5.22B	4.99C	5.18BC	4.99B	5.18AB
L. lactis TAB 57	TAB 57	5.14B	5.28D	4.98B	5.20CD	4.99B	5.25E
E. faecium TAB 7	TAB 7	5.09A	5.23BC	5.07F	5.21DE	5.07D	5.24D
E. faecalis TAB 52	Enterocin I	5.09A	5.24BC	5.02E	5.20CD	5.02C	5.23D
E. faecalis INIA 4	Enterocin AS-48	5.13B	5.15A	4.97A	5.14A	4.96A	5.12A

^aValues in the same column with different superscripts are significantly (P < 0.01) different.

Bacteriocin activity.

Bacteriocin activity was detected throughout the whole ripening period in NP, 2-day, 300-MPa and 50-day, 300-MPa cheeses made with bacteriocin-producing LAB (Tables 5 and 6). When 500 MPa was applied on day 2 (Table 5), bacteriocin activity was detected on day 3 only in cheeses made with nisin A- or nisin Z-producing LAB, whereas on day 60, bacteriocin activity was detected in all cheeses made with bacteriocin-producing LAB. In cheeses pressurized at 500 MPa on day 50 (Table 6), all bacteriocins except lacticin 481 and bacteriocin TAB 7 were detected on days 51 and 60.

TABLE 5.

Bacteriocin activity (halo [mm]) in cheeses with bacteriocin-producing LAB pressurized on day 2^a

LAB	Bacteriocin	Nonpressurized			300 MPa at 10°C for 10 min			500 MPa at 10°C for 5 min
		3 days	20 days	60 days	3 days	20 days	60 days	3 days	20 days	60 days
None		ND	ND	ND	ND	ND	ND	ND	ND	ND
L. lactis TAB 24	Lacticin 481	7.8 ± 0.5	7.0 ± 0.7	8.4 ± 0.7	8.5 ± 0.6	S	8.0 ± 1.0	ND	ND	6.1 ± 0.8
L. lactis TAB 26	Nisin Z	9.5 ± 0.8	7.1 ± 0.6	7.5 ± 1.0	6.5 ± 0.5	7.0 ± 0.6	7.9 ± 0.6	10.1 ± 0.7	ND	6.5 ± 0.8
L. lactis TAB 50	Nisin A	15.0 ± 0.8	15.1 ± 1.0	16.5 ± 1.3	9.3 ± 0.8	15.5 ± 1.2	7.1 ± 0.7	7.5 ± 0.6	ND	7.5 ± 0.9
L. lactis TAB 57	TAB 57	6.3 ± 0.4	S	8.1 ± 1.0	S	S	8.0 ± 0.9	ND	ND	6.0 ± 0.6
E. faecium TAB 7	TAB 7	8.3 ± 0.6	7.0 ± 0.6	7.0 ± 0.7	7.3 ± 0.5	7.1 ± 0.6	8.0 ± 1.0	ND	ND	6.5 ± 0.5
E. faecalis TAB 52	Enterocin I	16.0 ± 1.1	16.5 ± 1.1	9.8 ± 0.7	15.8 ± 0.9	17.5 ± 1.0	7.0 ± 0.7	ND	ND	6.0 ± 1.0
E. faecalis INIA 4	Enterocin AS-48	8.5 ± 0.8	9.0 ± 1.0	7.3 ± 0.6	10.8 ± 1.0	11.5 ± 0.9	10.0 ± 1.4	ND	ND	7.1 ± 0.9

^aMeans ± standard deviations from triplicate determinations on samples from two experiments are given. ND, not detected; S, slight, nonmeasurable inhibition.

TABLE 6.

Bacteriocin activity (halo [mm]) in cheeses with bacteriocin-producing LAB pressurized on day 50^a

LAB	Bacteriocin	Nonpressurized		300 MPa at 10°C for 10 min		500 MPa at 10°C for 5 min
		51 days	60 days	51 days	60 days	51 days	60 days
None		ND	ND	ND	ND	ND	ND
L. lactis TAB 24	Lacticin 481	7.0 ± 0.6	8.4 ± 0.9	7.1 ± 0.7	7.0 ± 0.9	ND	ND
L. lactis TAB 26	Nisin Z	9.5 ± 1.0	7.5 ± 0.9	6.5 ± 0.5	6.5 ± 0.8	7.0 ± 0.6	S
L. lactis TAB 50	Nisin A	18.5 ± 1.6	16.5 ± 1.5	6.5 ± 0.8	7.1 ± 1.1	7.1 ± 0.7	6.0 ± 0.8
L. lactis TAB 57	TAB 57	7.1 ± 0.7	8.1 ± 1.0	6.0 ± 0.4	7.0 ± 0.7	6.0 ± 0.5	6.0 ± 0.7
E. faecium TAB 7	TAB 7	6.5 ± 0.8	7.0 ± 0.9	6.5 ± 0.5	7.0 ± 0.6	ND	ND
E. faecalis TAB 52	Enterocin I	14.3 ± 1.0	9.8 ± 0.7	13.5 ± 0.9	7.6 ± 0.7	6.1 ± 0.8	6.0 ± 1.0
E. faecalis INIA 4	Enterocin AS-48	9.3 ± 0.7	7.3 ± 1.1	8.8 ± 0.9	5.8 ± 0.4	6.0 ± 0.4	6.0 ± 0.5

^aMeans ± standard deviations from triplicate determinations on samples from two experiments are given. ND, not detected; S, slight, nonmeasurable inhibition.

DISCUSSION

The effect of HP treatments in combination with bacteriocin-producing LAB added as adjuncts to the starter culture on the inactivation of E. coli O157:H7 throughout the ripening of raw-milk cheeses was investigated. The increase in counts of E. coli O157:H7 from milk to 3-day-old cheese of approximately 2 log units is in agreement with previous data obtained for other cheese varieties (2, 27). Levels of E. coli O157:H7 decreased only 1.7 log units from day 3 to day 60 in NP cheese made without bacteriocin-producing LAB, indicating that E. coli O157:H7 may survive in raw-milk semihard cheeses made with commercial starter cultures. The ability of this serotype to survive in ripened cheeses has been demonstrated previously (11, 27, 28).

The reductions in E. coli O157:H7 populations by HP (1.3 to 5.8 log units) on the day after treatment (Tables 1 and 2) were lower than those reported for E. coli in Mató cheese (6.7 to 8.7 log units) treated at 400 to 500 MPa for 5 to 15 min at 2 to 25°C (6) and in cheddar cheese (7 log units) treated at 400 MPa for 20 min at 20°C (24).

Our results showed a higher inactivation of E. coli O157:H7 by HP at 500 MPa than at 300 MPa. This is in agreement with the results obtained by García-Graells et al. (8), with reductions of 0.5 log units and 1.4 log units for E. coli K12 pressurized in milk at 300 and 500 MPa, respectively, during 15 min at 20°C. Linton et al. (18) also observed higher inactivation rates at 500 MPa than at 300 MPa in skim milk treated during 15 min at 20°C for 10 out of 12 pathogenic E. coli strains. O'Reilly et al. (24) demonstrated that increasing the pressure from 300 to 400 MPa resulted in a higher degree of inactivation than extending the length of treatment.

Inactivation of E. coli O157:H7 by pressurization at both 300 and 500 MPa was higher when the treatment was applied on day 50 than if applied on day 2 (Tables 1 and 2). In skim milk, the pressure resistance of E. coli cells at 600 MPa at 20°C for 10 min did not change for the first 5 days after the onset of the stationary phase, after which their resistance decreased (18). In the present work, variations in cell inactivation by HP treatment at different times of ripening could be related to differences in the physiological status of cells in the stressful environment of cheese. Another plausible explanation for variations in cell inactivation is the concentration of easily available substrates needed for damage repair and cell recovery after sublethal HP treatments. Carbohydrates such as lactose, glucose, and galactose are at much higher concentrations in 2-day-old cheese than in 50-day-old cheese, since they are exhausted by cheese microbiota as ripening proceeds (7). HP-treated cells seeded in a selective medium would grow more efficiently if they had repaired cell damage during the previous 24 h in cheese, and 2-day-old cheese offers energy sources for repair which are not available in 50-day-old cheese. Also, stress of cheese and HP treatment could lead to sublethally injured cells that would be killed by plating on selective media but might be recovered if previously resuscitated.

The increase in cheese pH observed after HP treatment on day 2 has been previously described for goats' milk cheese pressurized at 400 MPa 1 day after salting (30) and for Gouda cheese pressurized during brining from 200 MPa onwards (22). Contrarily, a decrease in cheese pH caused by HP treatment such as that found for cheeses pressurized on day 50 (0.09 pH units on average) has not been previously reported.

Gram-negative bacteria are generally not inhibited by bacteriocins of LAB (4). Our results showed a slight inhibitory effect of lacticin 481, nisin A, bacteriocin TAB 7, and bacteriocin TAB 57 in NP cheeses on the survival of E. coli, which cannot be related to differences in pH values. It could be attributed to a higher sensitivity to bacteriocins of injured cells of gram-negative bacteria (16), which in our work might be due to a prolonged acid injury at the low pH values of cheese.

Bacteriocin activity was detected in cheeses pressurized at 300 MPa throughout the ripening period. However, the application of 500 MPa diminished bacteriocin activity. Cheeses pressurized at 500 MPa on day 2 exhibited low bacteriocin activity on day 60 but did not show bacteriocin activity on day 20, a fact which could be attributed to the recovery of bacteriocin-producing LAB during ripening. The loss of bacteriocin activity resulting from HP treatment at 500 MPa observed in the present work has not been previously reported. On the contrary, Morgan et al. (23) observed an increase in the inhibitory activity of lacticin 3147, but not of enterocin 1146 or pediocin, at 400 to 800 MPa.

According to our results, an increase in the inactivation rate of HP treatments on E. coli O157:H7 in cheese was achieved in the presence of bacteriocins produced by LAB. Sublethal damage of the outer membrane of gram-negative cells (12) or changes in membrane fluidity (32) by HP treatments could facilitate the access of bacteriocins to the cytoplasmic membrane (16). The higher activity of HP in combination with nisin on the inactivation of E. coli in buffer (21) and of HP in combination with pediocin PA-1 in peptone water (15) have been described previously. The use of HP in combination with nisin was an effective method to extend the shelf life of a fresh cheese variety (5) and of liquid egg (26). Furthermore, pressurization at 400 MPa for 10 min at 17°C of a model meat system with nisin A led to a greater inactivation of E. coli population (>6 log units) that persisted throughout 60 days of storage at 4°C, suggesting that the injured survivors became sensitive to nisin after pressurization (10).

A different behavior of E. coli O157:H7 in cheese after pressurization in the presence of various bacteriocins produced in situ was observed in the present work. Kalchayanand et al. (17) reported a different behavior of E. coli O157:H7 and Salmonella enterica serovar Typhimurium M1 after HP treatments in the presence of pediocin AcH or nisin. Pressurization of a meat model system with nisin A also led to a higher inactivation of E. coli than if sakacin K, enterocins A and B, or pediocin AcH were added (10).

HP treatments of semihard cheeses at 300 MPa for 10 min at 10°C on day 50 applied in combination with nisin A-, bacteriocin TAB 57-, enterocin I-, or enterocin AS-48-producing LAB led to the complete inactivation of E. coli O157:H7 in 60-day-old raw-milk cheeses. Mild HP treatments applied to cheeses after the first weeks of ripening would minimize changes in the maturation process. The use of moderate pressures combined with bacteriocin-producing LAB, within the hurdle concept of food preservation, is a feasible procedure to control the growth of pathogenic microorganisms coming from postpasteurization contamination or present in the milk used to manufacture raw-milk cheeses.