Volatile and non-volatile compounds in ripened cheese : their formation and their contribution to flavour

Flavour is one of the most important attributes of cheese. Cheese flavour is the result of the breakdown of milk protein, fat, lactose and citrate due to enzymes from milk, rennet and microorganisms during production and ripening of cheese. For a large part the development of flavour during the ripening of cheese is determined by the process of proteolysis of caseins. Over the past years proteolysis has been studied very extensively and as a result a wealth of information about this process has been obtained. The ultimate products of proteolysis, amino acids, are prime flavour precursors in cheese and the formation of a "correct" pool of free amino acids is essential for a balanced development of cheese flavour. Amino acids undergo enzymic as well as chemical conversions to essential flavour compounds. However, the formation of cheese flavour compounds by lactic acid bacteria and their enzymes probably is much more important than the formation of flavour compounds by pure chemical reactions, e.g. coupling of carbonyls with amino acids.Until recently, the role of mesophilic starter lactococci, as present in hard-type cheeses (e.g. Gouda), in the process of amino acid degradation was regarded to be limited. Nevertheless, Gouda cheese develops an intense flavour and numerous volatile compounds derived from catabolism of amino acids can be found in this type of cheese.The purpose of the research described in this thesis was to investigate flavour compounds in the water-soluble fraction of cheese and to elucidate the formation of some of these compounds by lactococcal enzymes.Chapter 2 describes the production of water-soluble fractions (WSFs) from various types of cheese and the subsequent fractionation of these WSFs by serial ultrafiltration (UF) with membranes of different molecular weight cut-off. As a result fractions containing water-soluble components of various molecular weights were obtained (UF<500 Da, 500<UF<5000 Da and UF>5000 Da). Sensory analysis revealed that lowmolecular-weight (<500 Da) compounds were responsible for flavour in WSFs of Cheddar, Gouda, Gouda 20 +Gruyère, Maasdam, Parmesan and Proosdij cheese. The larger peptides and proteins in the UF>500 Da fractions did not contribute directly to the actual flavour of these cheeses, although some flavour attributes, e.g. bitter, are ascribed to larger peptides.The UF<500 Da fractions of the various cheeses include small peptides (probably not larger than tetra peptides), amino acids, fatty acids and further breakdown products of these compounds. In the WSFs of Gruyère, Proosdij and Parmesan cheese large amounts of small peptides and free amino acids were detected, probably due to the action of proteolytic enzymes from thermophilic lactobacilli.The direct contribution of free amino acids to the actual cheese flavour probably is limited. They more likely act as precursors for cheese flavour compounds. Our taste evaluation results with fractions containing mainly small peptides (<500 Da), obtained from WSF by Sephadex G-10 gel filtration (Gouda cheese) and Sep-Pak C 18 chromatography (various cheese-types), indicated that such peptides, together with amino acids, must be mainly responsible for basic flavours (e.g. brothy, savoury, sweet). There is not necessarily a relation between cheese flavour and concentration of total free amino acids.Free fatty acids were detected in relatively high amounts in WSF of Parmesan (butyric acid), Gruyère, (butyric and propionic acid) and Maasdam cheese (propionic acid). In these cheeses fatty acids probably play an important role in flavour.Chapter 3 describes the isolation, identification and possible origin of volatile compounds in the WSFs of 8 hard-type cheeses. The analysis was performed by gas chromatography-mass spectrometry. The cheeses used and the procedure for preparation of WSF (having a distinct cheese-like taste) were the same as applied in Chapter 2.The volatiles identified belonged to six major groups: fatty acids, esters, aldehydes, alcohols, ketones and sulphur compounds. The flavour attributes of various constituents of each of these groups have been described. Most of the compounds detected were present in the WSF of all eight types of cheese, although their concentrations showed distinct differences. From this it can be concluded that there is not a single compound or class of compounds which is responsible for the full flavour of cheese. Numerous volatile components contribute to the flavour of cheese and our results support the "component balance theory" postulated some 40 years ago. A consequence of this is that it is not possible to describe the flavour of cheese in precise chemical terms.A considerable portion of the volatiles identified during our study originated from fatty acids (e.g methyl ketones and secondary alcohols) and amino acids (e.g. branchedchain aldehydes and alcohols and sulphur compounds). The breakdown of fatty acids and amino acids is probably governed primarily by enzymic processes and the starter enzymes are a major source of the enzymes involved. Non-starter organisms (e.g. moulds and bacterial surface flora), present in certain types of cheese, naturally also contribute to the formation of flavour compounds.In hard-type cheeses, such as Gouda and Cheddar, proteolytic enzymes from mesophilic starter lactococci play a crucial role in the formation of free amino acids during ripening. The results of the study described in Chapter 4 indicate that enzymes from mesophilic lactococci are also very important for the formation of (volatile) flavour components from amino acids. In the literature little information is available concerning the significance of amino acid degradation by mesophilic starter bacteria, although the formation of flavour components by certain mesophilic starter strains has been reported.We incubated cell-free extract (CFE), containing all soluble enzymes from Lactococcus lactis subsp. cremoris B78, a Gouda cheese starter organism, with methionine, methionine-containing peptides (e.g. fragment α s1 -CN(f24-199) from α s1 -casein or a mixture of peptides of molecular weight 500 - 5000 Da isolated from Gouda cheese) and fragment α s1 -CN(f1-23) from α s1 -casein (containing no methionine). These peptides, but also methionine, in itself were tasteless. Sensory analysis showed that a cheese- like flavour only developed during the incubations with methionine and methionine-containing peptides. The formation of relatively large amounts of volatile sulphur compounds, such as dimethyldisulphide and dimethyltrisulphide, from methionine during these incubations could be demonstrated by gas chromatographymass spectrometry. The use of heat-treated CFE in incubation experiments with methionine did not result in formation of a cheese-like flavour indicating that enzymic activity is necessary. Apart from sulphur compounds other volatiles produced during incubation of CFE with amino acids were identified. An example is 3-methylbutanal originating from leucine.The experiments in Chapter 4 unquestionably demonstrated that the conversion of methionine by Lactococcus lactis subsp. cremoris B78 is, at least partially, an enzymic process. Chapters 5 and 6 describe the purification and characterization of enzymes from this organism involved in the conversion of methionine. Two enzymic routes for formation of volatile products from methionine were resolved and in Figure 1 the reactions identified and possible follow-up reactions are shown.A direct demethiolation reaction of methionine is performed by cystathionine β-lyase. This pyridoxal-5'-phosphate-dependent enzyme is able to catalyse α,β-elimination as well as α,γ-elimination reactions. The latter process (indicated in Figure I by dashed lines) results in the production of methanethiol, a very potent flavour compound, from methionine. The physiological role of cystathionine β-lyase is however the conversion of cystathionine during the process of methionine biosynthesis. Although cystathionine β-lyase prefers to catalyze the α,β-elimination reaction (e.g. on lanthionine and cystathionine) α,γ-elimination on methionine does occur under conditions prevailing in ripening cheese, such as a high salt concentration and a low pH. An enzyme similar to cystathionine β-lyase, cystathionine γ-lyase (γ-CTL) was purified from Lactococcus lactis subsp. cremoris SK11 by Bruinenberg et al. This enzyme only catalyses the α,γ-elimination of cystathionine and not α,β-elimination. SK11 γ-CTL is also able to convert methionine by α,γ-elimination. However, in contrast to the Lactococcus lactis subsp. cremoris B78 enzyme, SK11 γ-CTL is unable to degrade L-homoserine and shows relatively high (α,β-elimination activity) toward L- cysteine.Figure 1. Pathways of the formation of volatile sulphur compounds from methionine by enzymes from Lactococcus lactis subsp. cremoris .The transamination of methionine by branched-chain aminotransferases from Lactococcus lactis subsp. cremoris B78 provided evidence for the existence of an alternative route for the formation of volatile sulphur compounds. This route comprises the conversion of methionine to 4-methylthio-2-ketobutyric acid (KMBA) in the presence of an α-keto acid, e.g. α-ketoglutaric acid (Figure 1). The intermediate KMBA is converted to methanethiol, which can be further converted to other volatile compounds important for cheese flavour e.g. dimethyldisulphide and dimethyltrisulphide. The aminotransferases described in this thesis, AT-A and AT-B, have a broad substrate specificity for both the amino-group donor and the amino-group acceptor. Branched-chain amino acids and α-ketoglutaric acid respectively were the most preferred substrates in this respect. Recently Yvon et al. reported the purification and characterization of an aromatic-amino-acid-converting aminotransferase from Lactococcus lactis subsp. cremoris NCDO 763. The enzyme was able to convert aromatic amino acids but also leucine and methionine and an important role in cheese flavour formation was assumed.The route of conversion of KMBA is uncertain, however. In eukaryotes the decarboxylation of KMBA by a branched-chain 2-oxo dehydrogenase complex has been reported. The product of decarboxylation of KMBA, i.e. methional (see Figure 1), has been implicated as an important factor in cheese flavour. The route of breakdown of methional however remains uncertain; a direct conversion of KMBA to methanethiol cannot be ruled out (Figure 1). Although up to now we could not establish the formation of methional from KMBA (Figure 1), our experiments have shown that (a) methional is produced during incubation of CFE from Lactococcus lactis subsp. cremoris B78 with methionine and (b) enzymes from this organism are involved in the conversion of KMBA. The 20-37 % ammonium sulphate fraction obtained from CFE namely facilitated the breakdown of KMBA. Additional experiments showed that thiamine pyrophosphate, a cofactor required for activity of decarboxylases, stimulated the enzymic breakdown of KMBA.A transaminative route of degradation of leucine similar to that of methionine (i.e. comprising transamination and decarboxylation) was also suggested by Braun and Olson. The products ultimately produced were 3-methylbutanal and 3-methylbutanol. These compounds, together with analogous compounds produced from isoleucine and valine, are important flavour compounds in various types of cheese. Because the aminotransferases described in Chapter 6 also displayed high activities towards branched-chain amino acids and aldehydes were produced during incubation of CFE with amino acids, it seems that at least two enzymic steps (transamination and decarboxylation) in the catabolism of these amino acids and methionine are mediated by similar enzymes from Lactococcus lactis subsp. cremoris B78.In Chapter 7 the results are shown of some preliminary experiments with cheese pastes and curds aimed at testing the role of methionine in combination with starter enzymes under more practical conditions. The observed flavours in pastes to which CFE of Lactococcus lactis subsp. cremoris B78 had been added were definitely cheese-like and volatile degradation products of methionine were detected with the aid of gas chromatography. The cheese-like flavour of the curds to which CFE and methionine had been added was undoubtedly more intense after 13 weeks than that of regular 13-weeks old Gouda cheese. From the results it is concluded that enzymes in CFE are able to convert methionine in the cheese-like systems, despite the unfavourable conditions e.g. low pH and high salt concentration. This finding is in line with the results of Chapters 5 and 6 which show that both the branched-chain aminotransferases and cystathionine β-lyase are still active under conditions prevailing in cheese.The results obtained in this study demonstrate that:(a) The direct contribution of non-volatile compounds formed during the process of cheese ripening, e.g. amino acids and small peptides, to the actual cheese flavour is limited. Volatile compounds formed during the process of cheese ripening undoubtedly are crucial for a proper cheese flavour. These volatiles mainly originate from casein, fat, lactose and citrate.(b) Amino acid-converting enzymes from starter lactococci, present in hard-type cheeses such as Gouda, play an essential role in the formation of cheese flavours. Currently, the cloning of the genes coding for the enzymes discussed in this thesis is in progress. The construction of genetically modified strains will facilitate further studies to elucidate the importance of these enzymes during cheese ripening.Although evidence has been provided that in cheese-like systems conversion of methionine to methanethiol and dirmethyIdisulphide takes place, future studies have to focus on amino acid converting processes in cheese itself. For diversification of cheese flavour not only the conversion of methionine is of importance, but also the conversion of other amino acids present in ripening cheese. Another important aspect is the role of lysis of starter cells during ripening. To assure an adequate interaction between substrates and enzymes lysis of cells, leading to the release of intracellular enzymes into the cheese matrix, is considered to be essential.