Expression and release of proteolytic enzymes of Lactococcus lactis : ripening of UF-cheese

Semi-hard cheese types, such as Gouda, cannot be satisfactorily produced when using ultrafiltration technology. Although the cheese yield increases using this method, the higher financial return is completely lost by the (poor) quality of the cheese. The work described in this thesis is directed at improving, by microbiological methods, the quality of e.g. Gouda cheese made from ultrafiltered milk. In the course of the work, some fundamental questions were raised on growth behaviour of lactic acid bacteria in UF concentrated milk in relation to regular milk, on survival and stability of the bacteria in regular cheese and on the principles of flavour development during cheese ripening.Growth characteristics of starter bacteria in milk concentrated by ultrafiltration. Chapter 2 describes the growth behaviour of Lactococcus lactis in UF concentrated milk in relation to regular milk. The total amount of biomass of L. lactis subsp. cremoris E8 and the mixed strain starter culture Bos decreased gradually by 25 and 40%, respectively, when growing in UF-retentates with increasing concentration factors up to a factor 3.6 compared with the total growth reached in regular milk. The cause of the decreased outgrowth found in UFretentates is not precisely known, but it is clearly related to the increased concentration of whey proteins in the UF-retentate.Regulated production of proteolytic enzymes. The enzymes of the proteolytic system, composed of the extracellular serine proteinase and the intracellular peptidases, hydrolyse in concerted action the milk proteins into amino acids. The proteolytic activity of lactococci is crucial for growth in milk. In cheese milk the degradation of casein is hydrolysed by the combined activity of chymosin and proteolytic enzymes. During maturation of cheese the pool of amino acids contribute, either directly or as precursor for flavour compounds, to the final cheese flavour (8). In contrast to the wealth of knowledge on the biochemical and genetic characterization of the proteolytic enzymes, little is known about the regulation of these enzymes, e.g. their medium and growth dependent activity.Chapter 3 describes the use of a reporter gene, β-glucuronidase (gusA) of Escherichia coli , to study expression of the prtP and prtM genes under different conditions. Both, prtP and prtM promoters, were stringently controlled by the peptide content of the medium. Specifically, addition of the peptides leucylproline or prolylleucine to the growth medium negatively affected the expression level of the prtP-gusA fusions. In mutants defective in the uptake of di-tripeptides the repression by these dipeptides was not observed, which suggests a role of the di-tripeptide transporter as a sensor for the extracellular small peptides.Chapter 4 describes the regulation of the extracellular PrtP and two intracellular peptidases, aminopeptidase N (PepN) and X-prolyl-dipeptidyl aminopeptidase (PepXP), in two different host strains, L. lactis subsp. lactis MG1363 and L. lactis subsp. cremon's SK1128, both containing plasmid pNZ521, which encodes the PrtP from strain SK110. Production levels of all three enzymes were found to be highest in the late exponential phase of growth. The production was only slightly affected by the growth rate, PrtP and PepN production levels increased with increasing growth rates whereas PepXP showed an optimum at growth rate of 0. 22 h -1. The PrtP production level showed a medium-dependent control, which correlated with the controlled expression of the prt promoters. Highest production level was observed during growth of L. lactis cells in milk, lowest levels during growth in a peptide rich medium. The two peptidases were found to be regulated in a similar way as PrtP in strain MG1363, while in host strain SKI128 no regulation was observed. The regulating effect of the dipeptide prolylleucine appeared to be independent of the growth rate of the cells.The basic mechanism for the controlled production of the proteolytic enzymes is not yet clarified. Deletion and mutation analyses of the prt promoter region revealed that a 90 bp sequence (operator), which contains the prtP and prtM promoter, is sufficient for their full expression and regulation (5). As already speculated in Chapter 3 , a putative negative regulator may bind to this prt operator region. The affinity of the regulator for the prt operator region is increased after a conformational change induced by interaction with an effector, resulting in repression of transcription (Fig. 1). Specific dipeptides, such as prolylleucine, are supposed to act as the effector molecules. Whether the dipeptide plays a direct role in the conformational change of the regulator protein, or that the activity of the di-tripeptide transport system itself facilitates the conformational change of the regulator, is still an intriguing question.The basic knowledge of the control mechanism of proteolytic enzyme production can be used to influence the proteolytic enzyme activity in starter bacteria using cultivation media with varying peptide concentrations.Chapter 5 describes the controlled production of proteolytic enzyme activity in L. lactis cell grown in different pre-treated milk media, when milk was subjected to increased heat treatments and higher UF concentration factors. Cells of L. lactis showed decreased activity of PrtP, PepN and PepXP when grown in milk with increased heat treatments and milk with different concentration factors concentrated by ultrafiltration. This medium-dependent regulation of PrtP was confirmed at the level of transcription initiation. Analysis of the peptide composition of the heat treated milk showed higher concentrations of small, probably hydrophobic, peptides, than in non-treated milk. Therefore, it is suggested that small peptides present in the milk medium, due to the heat treatment of the milk, control the production of the different proteolytic enzymes. It is speculated that the control of proteolytic enzyme production in UF-retentates is directed via the same mechanism. The observation that the increase in soluble N is much slower during ripening of UF-cheese than in traditional manufactured cheese is in agreement with the reduced proteolytic activity of the starter cell grown in UF concentrated milk (4, 7).Figure 1. Proposed model for the medium-dependent regulation of PrtP in L. lactis during growth in milk. (Opp = oligopeptide transport system, DtpT = di-tripeptide transport system).Lysis of starter bacteria in relation to flavour development in cheese. During maturation of cheese the starter cells are metabolically inactive. This excludes the energy driven transport of oligopeptides, degraded from the milk protein by the hydrolytic activity of PrtP, into the cell. To assure the production of amino acids from the oligopeptides intracellular peptidases have to be released into the cheese matrix by lysis of the starter bacteria (2).Chapter 6 describes the use of the lysogenic L. lactis subsp. cremoris SK110 to study the influence of different growth conditions on lysis. Lysis was induced via a temporary increase in growth temperature from 30°C to 40°C for 2.5 It. Highest sensitivity of the lactococcal cell for induced lysis was observed at neutral pH values and at high growth rates. Using electron microscopy, it was confirmed that lysis was indeed a result of prophage induction. The induced lysis resulted in an increased release of peptidases from the cytoplasm. Lysis induced in starter culture SKI 10 during cheese manufacturing leads to an enhanced pool of amino acids and a clearly distinguishable cheese flavour in the matured cheese, compared to the control cheese.Remarkably, the reduction in number of viable cells as a result of induced lysis, is not quantitatively reflected by the increased amount of released intracellular enzymes. This suggests the existence of non-viable, stable protoplasts after the prophage-induced lysis. Subsequently, the 6 times higher release of intracellular proteolytic enzymes in a prophageinduced culture of strain SK110, just enhances the amount of free amino acids 1.4-fold in cheese. This clearly demonstrates that the release of intracellular peptidolytic enzymes is a rate limiting step in flavour development in cheese. However, elimination of this rate limiting step is not sufficient to ensure satisfactory ripening. It may be speculated that another rate limiting step become apparent, for instance the degradation of the free amino acids into volatile flavour compounds.Finally, Chapter 7 deals with the relationship between the cell wall composition and the susceptibility of the cell for lysis. Cheese was manufactured with strain L. lactis SK 110 and its transconjugant containing the mutated nisin transposon Tn5276, which encodes for nisin immunity but not production, bacteriophage resistance and the sucrose operon. The bitter score was rather high in cheese produced with the transconjugant compared to the cheese made with strain SK110. Cells of transconjugant SK110::Tn5276-NI showed less susceptibility for (induced) lysis than cells of strain SK110. It was observed that the peptidoglycan of the transconjugant SKI 10::Tn5276-NI was less sensitive to mutanolysin than the parental strain. The peptidoglycan of the transconjugant SK110::Tri5276-NI showed a significantly higher amount of tetrapeptides, involved in cross-linking of the glycan strands, than the peptidoglycan of strain SK110. The changed peptidoglycan composition of transconjugant SK 110:: Tn52 76-NI could decrease the susceptibility of the cell wall for lytic enzymes. This explains the observed higher bitter score via a reduced release of (debittering) intracellular peptidases during cheese ripening.The observation that the presence of the nisin-sucrose transposon Tn5276-NI affects the lactococcal cell wall composition, suggests either a disturbed gene expression in the host strain due to the specific integration site of Tn5276-NI in the genome of the host strain, or the presence of genes on Tn5276-NI that play a role in peptidoglycan synthesis. The first option can be revealed by using different, non-isogenic, host strains for Tri5276-NI, the second via inactivation studies of the various genes located on Tri5276-NI. Until now nothing is known about the control of peptidoglycan synthesis in lactococci. For application in the dairy industry, this knowledge would allow the development of nisin-immune, industrial strains which are still able to develop proper flavour characteristics.In general, it can be concluded that the release of peptidolytic enzymes, due to lysis of the cell, is an important, rate limiting, factor in cheese ripening. Since lysogenity seems to be wide-spread among lactococci, it is interesting to speculate that the required lysis of the starter culture during maturation of cheese is based on (spontaneous) induction of the prophage in the early stage of ripening. The fact that a lysogenic strain is immune to infection with its own phage, prevents lysis of the whole starter population and leads to the desired balance between lysed and intact cells. Therefore, induction of lysis may well be come a strong tool to accelerate the ripening of cheese and to alter the flavour characteristics of the product. However, induction of lysis via a heat treatment during cheese manufacturinginduces considerably enhanced syneresis, which may prevent attaining the desired moisture content of the cheese. Alternative inducers to be considered are treatments with salt, as already used during brining of the cheese, high pressure or UV-light. Another possibility is the use of the controlled expression of lytic enzymes. Recently, de Ruyter et al. have developed a controlled expression system using the autoregulated promoter of the nisin operon for overexpression of bacteriophage lysins (3).Prospects of UF-cheese ripening. The studies described in this thesis can be related to various aspects encountered during ripening of UF-cheese. The less favourable growth characteristics of starter cells grown in UF-retentates, compared to normal milk (Chapter 2) gives rise to (i) a lower total proteolytic activity expressed per g cheese, (ii) reduced production levels of the different proteolytic enzymes which is due to the changes in growth behaviour and the changes in cultivation medium of the starter bacteria (Chapter 4, 5), and (iii) to a reduced sensitivity for lysis of the starter culture, which results in a reduced release of intracellular flavour generating enzymes into the cheese matrix (Chapter 6).Other studies, directed at elucidating the technological problems encountered during UF-cheese manufacturing, showed that the relative activity of chymosin in UF cheese-milk gradually decreased with increasing the concentration factor of the milk (1). To improve the ripening of UF-cheese it is important, therefore: (i) to increase the total addition of chymosin to the UF-cheese milk, (ii) to increase the inoculation size of the starter culture to the UF-cheese milk, (iii) to select starter cultures with high production levels of proteolytic enzymes, and (iv) to select lysogenic starter cultures, which gives the possibility to induce lysis during cheese manufacturing. Preliminary results showed that these measures can significantly enhance the organoleptic quality of UF-cheese.Another promising possibility for manufacturing UF-cheese is the use of thermophilic strains as an additional starter culture. Thermophilic strains have been successfully used because of their debittering activity (6), which is probably due to their high proteolytic activity and their high susceptibility for release of the intracellular peptidolytic activity. Although, the use of thermophilic strains gives rise to particular organoleptic characteristics, which deviate from the traditional Gouda cheese flavour, these strains are very successfully used for rapid flavour development in semi-hard cheeses.