INTRODUCTION

Top Abstract Introduction Materials and Methods Results and Discussion References

Members of the lactic acid bacteria play an essential role in a wide variety of food fermentations, with Lactococcus lactis representing the primary starter cultures exploited for cheddar cheese manufacture. In addition, a population of adventitious microflora, otherwise known as nonstarter lactic acid bacteria (NSLAB), can proliferate during ripening and often form the dominant flora in the cheese. The precise role of NSLAB strains in flavor development remains unclear; however, whether their effect is positive or negative, they certainly contribute to the unpredictability associated with cheddar cheese quality (7). Importantly, NSLAB have also been associated with defects in cheese such as the formation of calcium lactate crystals (27) and slit formation.

Given increasing consumer demand, the challenge now facing the cheesemaker is the development of a safe, well-flavored, consistent cheddar. One approach to improving cheese flavor has been the deliberate addition of selected adjunct lactobacilli during manufacture which would impart beneficial qualities (7). Although some studies reported cheese with improved flavor, in other studies adjunct cultures were responsible for flavor defects (13, 14, 20, 23). In addition, comparisons with control vats were difficult to interpret due to contamination with the adjunct strain or with adventitious NSLAB, which eventually reach levels similar to those in the test vat (1, 28). Hence, in order to study the relevance of NSLAB, as well as examine the role of adjunct cultures in cheddar flavor, there is a requirement for an effective and simple method of controlling adventitious strains during manufacture and ripening.

In previous studies, we reported on a lacticin 3147-producing lactococcal starter which is capable of inhibiting the growth of NSLAB during cheese ripening (6, 25). Lacticin 3147 is a lantibiotic-type, two-component (17) bacteriocin which is inhibitory to a broad spectrum of gram-positive bacteria (24-26). The genetic determinants encoding lacticin 3147 are located on a 60.3-kb conjugative plasmid (5) which has been transferred from the parent L. lactis DPC3147 to a range of cheesemaking lactococcal starters. One such strain previously used to manufacture full-fat cheddar, L. lactis DPC4275, exhibited control over the developing NSLAB population (25). In a further study, this strain was successfully used to manufacture a low-fat cheddar, which was subsequently ripened at the elevated temperature of 12°C in order to accelerate the ripening process while suppressing wild NSLAB populations (6). Importantly, the DPC4275 strain may also be used to improve the safety of some cheese products (16). Although this transconjugant strain has been used successfully in these trials, the parent strain, L. lactis DPC3147, is a more efficient producer of lacticin 3147 and is therefore more effective in reducing NSLAB populations.

It has been observed that some bacteria are capable of developing phenotypes of resistance to bacteriocins. For example, nisin-resistant mutants can be generated from sensitive strains by exposure to sublethal concentrations of the bacteriocin (10). The aim of the present study was to generate an adjunct culture which was more resistant to higher concentrations of lacticin 3147 and to analyze its performance when used in combination with a lacticin 3147 starter. It was envisaged that the resistant strain might survive the levels of bacteriocin produced in the cheese during manufacture, while the growth of adventitious NSLAB would be inhibited, thus allowing for analysis of the performance of the resistant strain throughout ripening.

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results and Discussion References

Bacterial strains and cultivation conditions. L. lactis DPC3147 (a natural producer of lacticin 3147), L. lactis DPC4275 (a lacticin 3147-producing transconjugant of a commercial strain) (25), and L. lactis HP (a sensitive indicator strain) were routinely propagated at 30°C in M17 (Difco Laboratories, Detroit, Mich.) supplemented with 0.5% lactose or in 10% reconstituted skim milk (RSM). Lactobacillus paracasei subsp. paracasei strains were maintained by weekly subculture in MRS (Difco Laboratories) and grown at 37°C. All strains were stocked at 80°C in 40% glycerol and are held in the Dairy Products Research Centre (DPC) collection at Moorepark, Fermoy, Ireland.

Isolation of Lactobacillus strains and development of increased resistance. Lactobacillus strains were isolated from a well-flavored, commercial, mature cheddar cheese. The cheese was initially macerated in 10-fold-sterile, distilled water, plated on lactobacillus selective agar, which is selective for lactobacilli (Becton Dickinson Microbiology Systems), and incubated for 5 days at 37°C. Four isolated colonies were selected and propagated in MRS broth. These strains were genetically fingerprinted (see below), following which two of the isolates were further selected and identified by sodium dodecyl sulfate-polyacrylamide gel electrophoresis protein profiling, as outlined by Pot et al. (22).

View larger version (28K): [in this window] [in a new window]   FIG. 1.   (A) The agar plates contain isolated colonies of the lacticin 3147-producing strain, L. lactis DPC3147, which are overlaid with a lawns of L. paracasei DPC5336 and DPC5337. The absence of zones of inhibition in the DPC5337 lawn indicates that this strain is more resistant to the levels of lacticin 3147 produced by L. lactis DPC3147 colonies, while strain DPC5336 remains sensitive. (B) Bactericidal activity of lacticin 3147 on L. paracasei DPC5336 () and DPC5337 (). Bacteriocin was added at a final concentration of 1,000 AU/ml to washed log-phase cells. Controls to which no bacteriocin was added are represented by DPC5336 () and DPC5337 (×). (C) Difference in flocculation during growth between L. paracasei DPC5336 () and DPC5337 ().

Genetic typing of Lactobacillus strains. Randomly amplified polymorphic DNA (RAPD)-PCR was subsequently used to verify that these more-resistant strains gave the same genetic fingerprint as the initial L. paracasei DPC5336 strain. The Lactobacillus strains were grown in MRS overnight at 37°C, and 1 ml of each culture was microcentrifuged for 1 min. The genomic DNA from each resultant pellet was isolated by resuspending the pellets in 200 µl of the surfactant, Tergitol (Sigma-Aldrich Company Ltd., Poole, Dorset, England) at a final concentration of 10% in sterile distilled water. Amplification of DNA was performed with a Perkin-Elmer DNA Thermal Cycler using R1 primer (5' ATGTAACGCC 3') (11) and P5 primer (5' ACGCGCCCT 3') (12). Taq DNA polymerase was added during the first temperature cycle (hot start), and DNA was amplified for 35 cycles. Each cycle involved 1 min of denaturation at 93°C followed by an annealing step at 44°C for 1 min and an extension step of 72°C for 1 min. Of the final mixture, 10 µl was analyzed on 1.5% (wt/vol) agarose (Sigma-Aldrich) gels with ethidium bromide staining.

Estimating the relative sensitivities of the resistant variants. To determine the potency of concentrated lacticin 3147 against L. paracasei DPC5336 and compare it with the potency against the resistant L. paracasei DPC5337 variant strain, the agar well diffusion assay as outlined by Ryan et al. (26) was adopted. The sensitivities of the Lactobacillus strains were expressed relative to the sensitivity of the indicator strain, L. lactis HP. To determine if the increased resistance is a stable feature of these variants, the strains were subcultured daily for 1 week at 37°C in the absence of lacticin 3147 for selection and sensitivity was again determined by the agar well diffusion assay (26).

Effect of concentrated lacticin 3147 on sensitive and resistant strains in buffer. Both L. paracasei DPC5336 and DPC5337 were grown for 4 h at 37°C (optical density at 600 nm, 0.32), and 2 ml of each was collected in triplicate by centrifugation. These cells were washed in 20 mM sodium phosphate buffer (pH 7) and resuspended at approximately 10⁷ CFU/ml in 2 ml of buffer. Lacticin 3147 solution was then added to both strains at a final concentration of 1,000 AU/ml, and the strains were incubated at 37°C. Samples were taken at appropriate intervals over a 90-min period to determine the viable cell count in MRS agar. Control experiments were performed in an identical manner with the exception that no bacteriocin was added. This experiment was performed in triplicate, and error bars in Fig. 1B represent the standard deviations of the counts.

Analysis of fatty acids in the membranes of Lactobacillus strains. The procedure for the determination of the fatty acid composition of the membranes was based on the method previously described by van Schaik et al. (29). Both L. paracasei DPC5336 and DPC5337 strains were inoculated at 2% in 1-liter volumes of MRS broth and incubated for 16 h at 37°C. Fatty acids were derivatized to fatty acid methyl esters (FAME), which were subsequently quantified by capillary gas-liquid chromatography using a Varian (Harbor City, Calif.) 3500 gas-liquid chromatograph fitted with a flame ionization detector. Standard FAME C_4:0 to C_20:2 (all of >99% purity) were obtained from Sigma Chemical Co. (St. Louis, Mo.). The response factors were calculated relative to the area of C_18:0, which was assigned a response factor of 1.00, and quantities were expressed as grams per 100 g of FAME.

Cheesemaking. For cheesemaking, bulk starters were cultivated in 10% RSM which had been heat treated at 90°C for 30 min and cooled to 21°C before inoculation. All starters were grown separately in 10% RSM and added to the cheese milk at a final rate of 1.4% (vol/vol). Adjunct Lactobacillus strains DPC5336 and DPC5337 were grown overnight at 37°C in MRS broth to approximately 10⁸ CFU/ml and were added at approximately 10³ CFU/ml to vats 2 and 3, respectively. Milk was pasteurized (72°C, 15 s) and cooled to 30°C prior to inoculation. Cheese was made as described by Ryan et al. (25) and ripened for 6 months at 7°C.

Analysis of cheese. The pH-versus-time profiles were monitored during cheesemaking. Samples were also removed during manufacture, and bacteriocin activity was assessed by the method outlined by Ryan et al. (25) with L. lactis HP again being used as the indicator strain.

Compositional analysis and cheese grading. The composition (pH, fat, protein, salt, and moisture) was analyzed by the method of Guinee et al. (9). The cheeses were assessed by a commercial grader from a local cheddar cheese factory. The maximum scores for flavor and aroma and body and texture were 45 and 40, respectively. For cheeses to be described as commercial grade, the minimum scores required are 38 for flavor and aroma and 32 for body and texture.

	RESULTS AND DISCUSSION

Top Abstract Introduction Materials and Methods Results and Discussion References

The variable nature of NSLAB populations probably reflects many of the inconsistencies often associated with cheddar flavor and quality, although this has not been proven experimentally (7). The use of lacticin 3147-producing starters in previous cheesemaking studies resulted in significant reductions in NSLAB counts during ripening (6, 25). In this study, we utilized these starters in combination with a lacticin 3147-resistant adjunct strain, which was found to form the dominant flora in the cheese, thereby showing that this strategy is a method of achieving complete control and manipulation of the developing flora in cheddar cheese.

Isolation of Lactobacillus strains. Four colonies, morphologically identical and randomly selected, were isolated from a commercial cheddar cheese and analyzed using both RAPD-PCR and sodium dodecyl sulfate-polyacrylamide gel electrophoresis protein profiling (22). While all were typed as L. paracasei subsp. paracasei by protein profiling, RAPD-PCR did indicate that one of the four was a different strain (results not shown). One of the three dominant isolates was chosen for further study and was designated L. paracasei subsp. paracasei DPC5336.

Characterization of the more-resistant Lactobacillus strain. Following the initial generation of DPC5337, a number of biochemical tests on both resistant and parent strains were performed. Previous studies have reported on the development of acquired or spontaneous resistance by Listeria subsp. (3, 10, 18, 19) to nisin, a commercially available broad-spectrum bacteriocin. Studies of Listeria mutants suggest that resistance may be caused by alterations in the fatty acid composition of the cell membrane (18, 19, 30). However, evidence which suggests that cell wall adaptations may also be responsible has also been presented (3, 4, 15). An important distinction between strains DPC5336 and DPC5337 was observed: the resistant derivative did not flocculate as readily during growth (Fig. 1C); this indicates that the resistant phenotype may be associated with changes on the cell surface. Furthermore, preliminary results from adsorption assays suggest that the resistant cells adsorb less bacteriocin than the sensitive cells (results not shown). Gas chromatographic analysis was used to determine whether the increased resistance was associated with changes in the membrane lipids. However, no significant differences between the membranes of DPC5336 and DPC5337 were observed, implying that the membrane lipids remain unaffected by the development of resistance to lacticin 3147 (Table 1).

Inhibition of resistant strain by concentrated lacticin 3147. Lacticin 3147 was shown to cause rapid cell death when added to log-phase cells of both DPC5336 and DPC5337, in contrast to results for control cultures to which no bacteriocin was added (Fig. 1B). However, the rate of killing and degree of inhibition were far greater for the sensitive strain than for the more-resistant strain. The addition of bacteriocin to approximately 10⁷ actively growing cells resulted in >99.9% killing of sensitive cells, whereas only approximately 40% of the more-resistant cells were killed under identical conditions.

Cheesemaking using resistant lactobacilli as adjuncts. A pilot cheddar cheese trial was initially carried out using L. lactis DPC3147 as the starter strain, even though previous trials using this strain suggest that it is associated with the development of off-flavors in cheese. This strain, however, produces sufficient acid to manufacture cheddar and is also the most efficient producer of lacticin 3147 identified to date. As observed in the previous trial, L. lactis DPC3147 produced sufficient acid to reduce the pH to 5.2 within the desired manufacturing time. Lacticin 3147 activity was assayed during manufacture and reached approximately 1,280 to 2,560 AU/g of cheese. This level of activity was maintained throughout ripening and correlated with a reduction in the growth rate of NSLAB in the control cheeses (Fig. 2). During the first 3 months of ripening, the levels of lactobacilli in vats 1 and 2 remained relatively low at approximately 3 × 10³ CFU/g. In contrast, the experimental vat had at least 100-fold more lactobacilli at day 1, and this value increased to 6 × 10⁶ CFU/g after 50 days. This represented a 1,000-fold increase in lactobacillus numbers over those found in control vats. Given that the initial inoculum of each adjunct culture was only 10³ CFU/ml, this indicated that the more-resistant strain grew (vat 3) during cheese manufacture whereas growth of the sensitive strain was inhibited (vat 2).

View larger version (17K): [in this window] [in a new window]   FIG. 2.   (A) Growth of NSLAB during cheese ripening in trial 1. The adjunct cultures were added to vats 2 and 3 at a level 103 CFU/ml (arrow). No adjunct was added to vat 1. The cheese in vat 1 () was manufactured with L. lactis DPC3147, the natural producer of lacticin 3147; that in vat 2 () was manufactured with L. lactis DPC3147 and L. paracasei DPC5336, and that in vat 3 () was manufactured with L. lactis DPC3147 and L. paracasei DPC5337. (B) Residual lacticin 3147 activity in the cheese after 6 months of ripening is observed as zones of inhibition against the indicator, L. lactis HP. V1 to V3, vats 1 to 3, respectively.

View larger version (45K): [in this window] [in a new window]   FIG. 3.   (A) Growth of NSLAB during cheese ripening in trial 2. The adjunct cultures were added to vats 2 and 3 at a level 103 CFU/ml (arrow). No adjunct was added to vat 1. The cheese in vat 1 () was manufactured with L. lactis DPC4275, that in vat 2 () was manufactured with L. lactis DPC4275 and L. paracasei DPC5336, and that in vat 3 () was manufactured with L. lactis DPC4275 and L. paracasei DPC5337. (B) Residual lacticin 3147 activity in the cheese after 6 months of ripening is observed as zones of inhibition against the indicator, L. lactis HP. V1 to V3, vats 1 to 3, respectively. (C) RAPD-PCR profiles of a representative number of NSLAB isolates from each of the cheeses at months 2 and 4. Lanes S and T, profiles of DPC5336 and DPC5337, respectively; lane M, 100-bp ladder.

Compositional and sensory analysis. The gross compositions of all cheeses were within the parameters set out by Giles and Lawrence (8) for cheddar cheese. There was no noteworthy difference in water-soluble nitrogen, phosphotungstic acid-nitrogen, and free amino acid composition between the control cheeses and cheese made with the adjunct cultures.

Conclusions. This study illustrates how lacticin 3147, in combination with resistant strains, can be applied as a tool to control the microbial flora of cheese products. In this way, proliferation of adventitious gram-positive flora can be limited, allowing both the producing starters and resistant adjuncts to flourish. This system offers obvious advantages over techniques such as aseptic production and antibiotic addition for NSLAB control, in that it can be adapted to large-scale production, as well as allowing sensory analysis of the final product. A number of potential adjuncts may be considered in the future, and, apart from flavor-enhancing strains, adjuncts with probiotic characteristics might also be worthy of analysis, given their associated health benefits. If such strains are to be examined, it should be ensured that the acquisition of lacticin 3147 resistance does not somehow compromise the health-promoting properties of the strain.