Role of lactic acid bacteria on the yogurt flavour: A review

ABSTRACT

Considerable knowledge has been accumulated on the lactic acid bacteria (LAB) that affect the aroma and flavour of yogurt. This review focuses on the role of LAB in the production of flavour compounds during yogurt fermentation. The biochemical processes of flavour compound formation by LAB including glycolysis, proteolysis, and lipolysis are summarised, with some key compounds described in detail. The flavour-related activities of LAB mostly depend on the species used for yogurt fermentation, and some strategies have been developed to obtain more control of the flavour-forming process. Metabolic engineering can be a powerful tool to reroute the metabolic flux towards the efficient accumulation of the desired flavour compounds with the knowledge of the complex network of flavour-forming pathways and the availability of genetic tools. Further progress made in the omics-based techniques and the use of systems biology approaches are needed to fully understand, control, and steer flavour formation in yogurt fermentation processes.

KEYWORDS:

• Flavour
• Glycolysis
• Lactic acid bacteria
• Lipolysis
• Metabolic engineering
• Proteolysis Yogurt

Introduction

Yogurt is one of the most popular fermented dairy products, and its consumption is increasing worldwide.^[1] According to the Euromonitor database, the yogurt production in 2015 reached 27.7 million metric tonnes, a 1.2-fold increase compared with the yield in 2010.^[2] Its popularity is not only due to various health claims and therapeutic effects but also for its sensory properties. The primary sensory attributes of yogurt include texture, colour, and flavour.^[3,4] Yogurt is typically characterised as a smooth, viscous gel with a characteristic taste of sharp acid and a green apple flavour.^[5] Among these attributes, flavour has played an important role in determining the acceptability and preference of yogurt for customers.

Flavour is the sensory impression of food or other substances and is determined primarily by the chemical senses of taste and odour. Taste is the sensation produced when a substance in the mouth reacts chemically with taste receptor cells located on taste buds in the oral cavity and can be expressed as sweet, sour, salty, bitter, and umami.^[6] Regarding odour, this is caused by one or more volatilised chemical compounds, generally at a very low concentration, that humans or other animals perceive by the sense of olfaction.^[7] Taste and odour are complex phenomena in themselves, and their interaction with other sensory properties increases the complexity of human perception. Furthermore, other factors, for example, textural properties, sources of milk and fat content, also affect and interact with the flavour properties of yogurt, which can be found in recent reviews.^[8,9]

For dairy products, the sensory properties depend largely on the relative balance of flavour compounds derived from carbohydrate, protein or fat in the milk. The flavour components of yogurt include the volatile and non-volatile compounds already present in the milk and specific compounds produced from milk fermentation.^[5,10] It has been suggested that more than 90 different volatile compounds have been identified in yogurt, including carbohydrates, alcohols, aldehydes, ketones, acids, esters, lactones, sulphur-containing compounds, pyrazines, and furan derivative.^[10–12] In Table 1, a number of descriptions are given of some flavour components and their thresholds in water (if known). However, not all of the volatile components found in yoghurt are of sensory importance, as these components differ in concentration, only when its concentration exceeds the threshold, the components can be perceived. Noticeably, the aroma threshold values vary by several orders of magnitude (Table 1), which influences their perception by the human senses. Although no final conclusion has yet been drawn, the major compounds commonly reported responsible for imparting desirable flavour to yogurt are lactic acid, acetaldehyde, diacetyl, acetoin, and 2-butanone.^[5,8,13] Good flavoured yogurt results when proper levels of these compounds are produced. For example, the optimal concentrations of acetaldehyde in yogurt ranged from 14 to 20 mg/kg, while less than 8.0 mg/kg resulted in weak flavour and too much acetaldehyde would lead to an “astringent” off-flavour yogurt.^[14] Furthermore, like many other dairy products, yogurt is prone to deterioration, especially under improper storage conditions. Generation of volatile by-products leads to off-flavours and makes the product unsatisfactory for consumers.^[15,16] Accordingly, a desirable flavour in yogurt means yogurt containing major flavour compounds in proper levels without off-flavours.

Table 1. List of flavour compounds, description, and odour thresholds identified in yogurts.

Flavour compounds^a	Description^b	Odour threshold(mg/kg)^c
Aldehydes
Acetaldehyde	Ethereal, fresh, green, pungent	0.0079–0.039
Propanal	Solvent, pungent	0.081–0.16
Butanal	Pungent, green	0.007–0.017
3-Methylbutanal	Cocoa, almond, green, malty, unripe, cocoa, malty	0.006
Hexanal	Green, cut-grass	0.0091
Octanal	Fat, soap, lemon, green	0.0009
Furfural	Bread, almond, sweet	0.77
Benzaldehyde	Almond, burnt sugar	1–4.6
Methional	Cooked potato	0.0002
Pentanal	Fruit, bread, sweet	0.022
2-Methylbutanal	Fermented	0.003
Heptanal	Green, sweet	0.06–0.55
Nonanal	Sweet, floral, citrus, grass-like	0.04
(E)-2-Nonenal	Green, fatty	0.0004
Decanal	Fatty	0.03
Undecanal	Fatty	0.01
Phenylacetaldehyde	Floral sweet	0.004
Ketones
Diacetyl	Buttery, creamy, vanilla	0.0011
Acetone	Sweet, fruity	0.832
Acetoin	Buttery	0.014
2-Propanone	Sweet, fruity	0.832
2-Butanone	Varnish-like, sweet, fruity	17–32
2-Pentanone	Sweet, fruity, cheesy	1.38
2,3-Pentanedione	Butter, vanilla, mild	0.02
2-Hexanone	Floral, fruity	0.56
2-Undecanone	Floral, rose-like, herbaceous	0.041–0.082
1-Nonen-3-one	Sweet,waxy	0.000000008
3-Penten-2-one	Sweet, fruity	1.2
2-Heptanone	Fruity, spicy	0.001–0.01
3-Octanone	Mushroom, fruity	0.028
Nonanone	Grassy-herbal, green-fruity, floral	0.041–0.082
2-Dodecanone	Sweet, waxy	0.042–0.083
3-Hexanone	Sweet, fruity	0.041–0.081
1-Octen-3-one	Mushroom-like, earthy, fruity	0.0005
Acids
Octanoic acid	Wax, soap, goat, musty, rancid, fruity	0.91–1.9
Propionic acid	Vinegar, pungent, sour milk	2
Acetic acid	Vinegar, pungent, acidic	25.59–26
Pentanoic acid	Sweat	0.28
Hexanoic acid	Pungent, rancid, flowery	1.8–1.84
Heptanoic acid		0.64–0.91
Decanoic acid	Rancid	0.13
Butanoic acid	Sharp, cheesy, rancid, sweaty, sour, putrid	1.40
Butyric acid	Rancid, cheese, sweat	1.40
3-Methylbutanoic acid	Rancid cheese, sweaty, putrid	0.1
Nonanoic acid	Green, fat	4.6–9.0
Formic acid	Pungent	1240–2440
Isovaleric acid	Rotten fruit, mild, sweaty, rancid, faecal, putrid, flowery	0.1
Isobutyric acid	Sweet, mild, rotten apple	0.05
Alcohols
1-Butanol	Fruity, alcoholic	4.33
2,3-Butanediol	Buttery	20–50
Methanol	Alcoholic	800–1200
1-Propanol	Alcoholic, pungent	5.7–7.8
2-Propanol	Alcoholic	40–78
Ethanol	Mild, ether	200
2-Heptanol	Earthy, oily	0.041–0.081
2-Butanol	Wine	3.3
2-Furanmethanol	Burnt, sweet, caramellic, brown	1.9–2.0
1-Penten-3-ol	Fish, pungent	0.4
1-Pentanol	Alcoholic, iodoform-like	0.12
3-Methyl-2-butenol	Sweet, fruity	0.25
3-Methylbutanol	Fusel, alcoholic, pungent, ethereal	3.06
1-Octen-3-ol	Mushroom-like	0.01
Guaiacol	Bacon, phenolic, smoked, spicy	0.02
Esters
Ethyl acetate	Solvent-like, fruity, pineapple	3.28–3.3
Methyl acetate	Ethereal and solvent-like, estery, fruity, winey	1.5–47
Butyl acetate	Pineapple	0.01–0.1
δ-Dodecalactone	Fruit, sweet	0.06
γ—Dodecalactone	Sweet, fruit, flower	0.007
Ethyl hexanoate	Apple peel, fruit	0.001
Ethyl butanoate	Fruity, sweet, banana	0.00018
Ethyl octanoate	Fruity, banana, apple	0.015
Aromatic hydrocarbons
Benzene	Aromatic	0.072
Toluene	Paint	0.024
Ethylbenzene	Phenol, spice	0.0024
1,3-Dimethylbenzene	Plastic	1.1
1,4-Dimethylbenzene	Leathery	1
Trimethylbenzene	Leathery	0.5
Hydrocarbon
Dichloromethane		5.6
Limonene	Lemon, orange	0.2
Ethenylbenzene	Balsamic, gasoline	0.0036
Heterocyclic
2-Methylthiophene	Sulphur	3
Pyrazine	Roasted	300
Methylpyrazine	Roasted	30
1 H-pyrrole	Sweet, warm, nutty, ethereal	20
Furans
Furan	Sweet	4.5–6
2-Methylfuran	Ethereal, acetone, chocolate	3.5–4
2-Pentylfuran	Green bean, butter	0.006
Sulphur compounds
Dimethyl sulphide	Intense, lactone-like, sulphurous, cabbage	0.01
Dimethyl disulphide	Boiled cabbage, cauliflower, garlic	0.0003
Dimethyl trisulphide	Sulphurous, faecal	0.01

^aFlavour compounds are from the following literature: Cheng,^[5] Routray & Mishra,^[8] Beshkova et al.^[58] and Thierry et al.^[20]

^bOdour descriptions are mainly sourced from the following website: http://flavornet.org/flavornet.html and http://www.thegoodscentscompany.com/

^cThresholds are mainly obtained from the literature which applied water as the matrix.^[91]

A great majority of the flavour compounds produced in yogurt result from the activity of microorganisms in starter cultures. The predominant organisms in these starter cultures are lactic acid bacteria (LAB), for example, Lactococcus lactis, Lactobacillus species, Streptococcus thermophilus, Bifidobacterium species, and Leuconostoc species.^[17] During fermentation, these bacteria perform three major biochemical conversions of milk components: (i) conversion of carbohydrate into lactic acid or other metabolites (glycolysis), (ii) hydrolysis of caseins into peptides and free amino acids (proteolysis), and (iii) breakdown of milk fat into free fatty acids (lipolysis).^[18] These reactions also lead to the production of various flavour and off-flavour compounds for yogurt catalysed by microbial enzymes.^[19] In addition, a few non-enzymatic reactions also occur, but they mainly derive from milk processing, for example, after heating.^[20]

Although the starter cultures could produce most key flavour compounds in yogurt, the amounts would be far from sufficient during the fermentation process that adding food flavours is the conventional approach to reach the demand of flavour acceptance by customers. However, there is a certain gap between the external flavours and natural milk aroma.^[14] In addition, food flavours do not meet the needs of consumers for natural and environmental attributes of food products. In these days, the increase in intrinsic potential to produce flavour compounds by LAB without adding food flavours (clean label) is the new trend for yogurt-making. Other LAB are combined with yogurt starter cultures to produce yogurt with enhanced nutritional properties and characteristic flavour. Metabolite formation and the effects of different LAB on flavour compounds in yogurt have been investigated.^[21,22] Furthermore, metabolic engineering has also been applied in laboratory to change genetic and regulatory processes in the cells to increase their potential for the formation of flavour compounds.^[23,24] However, there is still a long distance between research and application for this technology because of the legal restriction and the lack of consumer acceptance.

The aim of this article was to give an in-depth review of the role of LAB to yogurt flavour, including flavour and off-flavour production from microorganism-mediated glycolysis, proteolysis and lipolysis, the effect of different starter cultures on yogurt flavour, and the current development of metabolic engineering for flavour enhancement. Finally, the recent literatures on methods to steer and control flavour formation by LAB are also reviewed.

Biochemical routes of flavour compounds formation in yogurt

Flavour compounds produced from carbohydrate fermentation by LAB

In milk, lactose is present in substantial amounts in nature, and it is also the major energy and carbon source for the growth of LAB. LAB convert lactose into lactic acid, which gives yogurt the characteristic acidic taste. Variations in the metabolic products of lactose metabolism have yielded two main categories of fermentation, homofermentation, and heterofermentation, depending on the LAB species, the substrate, and the environmental conditions.^[25] Homofermentative pathways generate lactic acid as the main end-product, whereas heterofermentative metabolism results in some other metabolites, such as ethanol, carbon dioxide, or acetic acid.^[26] In addition, under certain conditions (carbon limitation, carbon excess of slowly metabolised sugars, aerobic conditions), a homofermentative metabolism can be shifted to a mixed-acid metabolism with various metabolites.^[27–29] These metabolic products include several aroma compounds or aroma precursors, such as acetaldehyde, ethanol, and diacetyl.

The typical aroma of yogurt is characterised chiefly by acetaldehyde, which exhibits a green apple or nutty flavour. Its concentration ranges from 2.0 to 41 mg/kg, depending on the strains and process factors used for yogurt fermentation. It has been reported that good flavour resulted only when greater than 8.0 mg/kg of acetaldehyde was produced in yogurt.^[23,30] Some possible pathways of acetaldehyde synthesis have been proposed (Fig. 1).^[5,13] An important metabolic precursor is pyruvate, which can be catalysed by α-carboxylase or aldehyde dehydrogenase to generate acetaldehyde.^[13] However, some researchers have found that acetaldehyde could not be formed from pyruvate in yogurt, as no activity of these enzymes was detected in S. thermophilus and Lb. bulgaricus.^[31]

Figure 1. General metabolic pathways for the formation of the main flavour compounds of carbohydrate fermentation by LAB. The pathways are generated according to the following literature: Cheng,^[5] Thierry et al.^[20] and Tamime & Robinson.^[13] The main flavour compounds are in bold. Key enzymes: CL, citrate lyase; OAD, oxaloacetate decarboxylase; LDH, lactate dehydrogenase; PDC, pyruvate decarboxylase; ALS, α-acetolactate synthase; PFL, pyruvate formate lyase; PDH, pyruvate dehydrogenase; ACDH, acetaldehyde dehydrogenase; ADHE, alcohol dehydrogenase; ACK, acetate kinase; ALD, α-acetolactate decarboxylase; DR, diacetyl reductase; AR, acetoin reductase; Tppi, thiamine pyrophosphate.

In fermented dairy products, various compounds with four carbon atoms are responsible for the typical aroma of yogurt with its butter-like flavour. Collectively known as C4 compounds, they include diacetyl, acetoin, and 2,3-butanediol.^[32] They can be generated from glycolysis or citrate metabolism of several LAB (Fig. 1), for example, Lactococcus, Leuconostoc, and Weissella species. Further details on the biochemical processes in the production of C4 compounds in yogurt can be found in recent reviews.^[3,5,8,17]

Among these C4 compounds, diacetyl is the most important flavour compound due to its low threshold. The important effect of diacetyl on the aroma of milk has been recognised since 1929, when it was shown that the distinctive aroma of fermented milk could be sensed just as the concentration of diacetyl reached 1 mg/kg.^[33] The typical concentrations of diacetyl in yogurt range from 0.2 mg/kg to 3 mg/kg.^[5,34] Both S. thermophilus and Lb. bulgaricus are able to produce diacetyl, and strains of Lc. lactis subsp. lactis biovar. diacetylactis can accumulate significant amounts of diacetyl due to their high capacity to metabolise citrate.^[35] Acetoin is the reduced form of diacetyl, and its flavour is considerably weaker than that of diacetyl. However, acetoin is significant for reducing the harshness of diacetyl and contributes to the mild creamy flavour.^[5] 2,3-Butanediol is the reduced form of acetoin, and it makes limited contribution to the creamy or buttery attribute.^[36]

Flavour compounds from amino acid conversion by LAB

The proteolysis of caseins is also an important biochemical pathway for flavour formation in yogurt. The conversion comprises two steps: proteolysis and amino acid degradation. Proteolysis of casein by LAB is initiated by cell-envelope proteinases (CEPs), which degrade the protein into oligopeptides (Fig. 2). For some LAB, CEP activation requires the help of an extra-cytoplasmic chaperon called PrtM, which is a lipoprotein essential to the autocatalytic maturation of CEPs.^[37] The second stage of protein degradation is the transport of di-, tri-, and oligopeptides into the cell (Fig. 2). Once casein-derived peptides are taken up by LAB cells, they are further hydrolysed to amino acids by the actions of peptidases. Multiple copies of peptidases can be encoded in one bacterial genome: for example, Lb. plantarum WCFS1 has 19 genes encoding intracellular peptidases of different specificity,^[38] while the genome of Lb. helveticus DPC 4571 has been found to include 24 genes encoding peptidases.^[39]

Figure 2. General conversion pathways of proteins relevant for flavour formation by LAB. The pathways are generated according to the following literature: Smit et al.,^[17] Sadat-Mekmene et al.^[48] and Aleksandrzak-Piekarczyk et al.^[25].Key enzymes: CEP, cell envelope proteinase; PrtM, proteinase maturation protein; Opp, oligopeptide transporter; DtpT, ion-linked transporter for di-and tripeptides; Dpp, ATP-driven transporter for di-and tripeptides; Pep, peptidase; AT, aminotransferase; HycDH, hydroxyacid dehydrogenase; KaDH, α-keto acid dehydrogenase complex; PTA, phosphotransacylase; ACK, acyl kinase; KdcA, α-keto acid decarboxylase; AlcDH, alcohol dehydrogenase; AldDH, aldehyde dehydrogenase; EstA, esterase A.

Free amino acids produced by proteolysis may be converted to various flavour compounds including ammonia, amines, aldehydes, phenols, indole, and alcohols, all of which may contribute to yogurt flavour (Fig. 2). In particular, branched-chain amino acids (e.g. Val, Leu, Ile), aromatic amino acids (e.g. Phe, Tyr, Trp), and sulphuric amino acids (e.g. Cys, Met) are the main sources of flavour compounds derived from milk protein.^[40] The first step of amino acid catabolism is transamination to their corresponding α-keto acids. The α-keto acids can further undergo different enzymatic reactions: either a reduction generating flavourless α-hydroxy acids, or a decarboxylation forming aldehydes, which can be further reduced to alcohols, or undergo oxidative decarboxylation forming acyl-CoA and then carboxylic acids.^[25] Esters or thio-esters are then formed in reactions between alcohols and carboxylic acids, catalysed by esterases or acyltransferases.^[17] Another important conversion route of amino acids is initiated by lyases, such as cystathionine β- and γ-lyases. These lyases are able to convert methionine to methanethiol, which in turn can be converted into dimethyl sulphide, dimethyl disulphide and dimethyl trisulphide via oxidation or thioesters via esterase-catalysed reactions.^[41] Although these sulphur compounds have significant impacts on cheese sensory profiles, some of them are identified as off-flavours for yogurt products.^[20] Threonine aldolase belongs to another class of lyases and is able to convert threonine directly to acetaldehyde.^[42] This lyase pathway is thought to contribute greatly to the acetaldehyde pool of yogurt.^[43] Furthermore, chemical conversions also play an important role in flavour formation: for example, phenylpyruvic acid (the α-keto acid of Phe) can be chemically converted to benzaldehyde.^[44]

Flavour compounds from lipids in LAB

Lipids are also a source of aroma compounds for yogurt. Lipid breakdown is the main source for the accumulation of free fatty acids (FAs), because most of the free FAs are derived from the decomposition of triglycerides. Considerable quantities of short-chain FAs that are precursors of other aroma compounds are produced from saturated FAs (Fig. 3).^[45] Unsaturated FAs are oxidised in the presence of free radicals to form hydroperoxides, which rapidly decompose to form hexanal or unsaturated aldehydes (Fig. 3).^[5] Unsaturated FAs also lead to the formation of 4- or 5-hydroxyacids, which readily cyclise to cyclic compounds.^[46] The principal cyclic compounds in yogurt are γ- and δ-lactones, which have 5- and 6-sided rings, respectively, and are stable and strongly fruity flavoured. However, excessive or unbalanced lipid oxidation and lipolysis can also lead to off-flavour especially during storage. Some products of lipid oxidation (e.g. aldehydes and ketones) give dairy products the stale and “oxidised” flavours.

Figure 3. General degradation pathways of lipids during milk fermentation by LAB. The pathways are generated according to the following literature: Cheng^[5] and Mcsweeney et al.^[45].

LABs are a source of esterase or lipases.^[47,48] However, most LAB possess only intracellular esterases, except Lb. fermentum and Micrococcus spp., in which extracellular esterases have been isolated.^[49,50] As a result, most esterases from LAB are unable to hydrolyse food lipids until their release from lysed cells. Consequentially, the contribution of lipolysis to the flavour of yogurt is limited, compared with its contribution to long-ripened cheeses.

Esterases of LAB can also catalyse the direct synthesis of flavour-active esters from glycerides and alcohols via a transferase reaction (alcoholysis).^[51,52] Research has shown that LAB can esterify ethanol with butyric acid and hexanoic acid to form ethyl butanoate and ethyl hexanoate.^[53] Interestingly, similar to what is found for hydrolysis, di- and monoglycerides are preferred substrates, compared with triglyceride, for the alcoholysis of esterases.^[54]

Effects of different LAB on yogurt flavour

As yogurt-making is a complex process of milk transformation, the flavour of yogurt is influenced by several factors, such as the starter cultures, processing parameters, source of milk and chemicals, and additives used.^[8,55] Among these factors, the starter cultures used for fermentation make key contributions to the formation of the flavour compounds. Generally, the traditional yogurt culture is composed of S. thermophilus and Lb. delbrueckii subsp. bulgaricus, and in several countries, the name “yogurt” is only allowed for those products that are produced with starters containing strains of both species.^[56] Although the two microorganisms are able to grow individually in milk, they have a symbiotic interaction called “proto-cooperation” in mixed cultures, which means that they are mutually beneficial during fermentation.^[56,57] It has been suggested that the level of flavour compounds is much greater in mixed cultures than either of the two single cultures due to their associative growth and mutual stimulation.^[13] Beshkova et al. measured the production of flavour compounds when mixed cultures were used during lactic acid fermentation and found that the maximum concentration of flavour compounds was reached within 22 to 31 h of cooling stage. In the samples with 22 h cooling time, acetaldehyde predominated (1415.0–1734.2 μg per 100 g), followed by diacetyl (165.0–202.0 μg per 100 g), acetoin (170.0–221.0 μg per 100 g), acetone (66.0–75.5 μg per 100 g), ethanol (58.0 μg per 100 g), and 2-butanone (3.6–3.8 μg per 100 g).^[58] Some researchers have shown that in the mixed starter cultures, Lb. bulgaricus is mainly responsible for acetaldehyde production, but others hold the opposite view.^[10] In fact, both strains are able to produce acetaldehyde, and the ability is strain specific.^[59] Benozzi et al.^[60] monitored the lactic fermentation driven by different yogurt commercial starters using proton transfer reaction mass spectrometry. It was found that nine volatile compounds showed considerable differences (statistically significant) among the four starters, due to the starters’ different biosynthetic capacities for the formation of volatiles. De Bok et al.^[21] also found that the levels of the methylated sulphides and dimethyl trisulphide were relatively low in the monocultures of Lb. bulgaricus and S. thermophilus, which suggests that the increased levels in the mixed culture are the result of interaction between the two species. In addition, proteolytic activity plays an important role in the proto-cooperation of the mixed cultures. However, Settachaimongkon et al.^[61] showed that only the non-proteolytic S. thermophilus strain performed proto-cooperation with Lb. bulgaricus. This combination produced more aroma volatiles and non-volatile metabolites than pure cultures or cultures with proteolytic S. thermophilus. From these results, it has been found that the association of these two microorganisms affects the production of volatile and non-volatile molecules involved in flavour development.

To increase the flavour of yogurt, strains belonging to the genera Leuconostoc and Lactococcus are often incorporated as adjunct cultures. These strains can metabolise citrate to produce C4 compounds.^[35] The major compounds related to the utilisation of Leuconostoc are diacetyl, acetic acid, and ethanol, while the flavour metabolites of Lc. lactis are diacetyl and acetoin. The most used strains are Lc. diacetylactis and Leuconostoc citrovorum due to their high capacity to metabolise citrate.^[62] Incorporation of these strains into yogurt could enhance the butter-like flavour and improve the yogurt flavour quality.

Incorporation of probiotics into yogurt is a recent trend in functional dairy production. Most commercial probiotics used in yogurt are strains belonging to the genera Lactobacillus and Bifidobacterium. In addition to their proposed health benefits, they also influence the flavour of yogurt into which they are incorporated. As a well-known probiotic strain, Lb. rhamnosus GG has been added to yogurt in many countries. Although it does not significantly influence the major aroma-forming volatile metabolites, it contributes to the yield of volatile organic acids and alcohols during fermentation and increases the formation of non-volatile organic acids and free amino acids during refrigerated storage.^[63] Some studies have compared the properties of commercial fermented dairy products, such as Yakult^@ and Actimel^TM, when fermented either by Lb. casei alone or with the incorporation of yogurt cultures. For samples fermented only with Lb. casei, the predominant compounds in the volatiles were acetic acid, acetoin, butyric acid, caproic acid, 2-pentanone, and 2-butanone, while the volatile compounds typical of yogurt were absent.^[64] For samples fermented with Lb. casei and yogurt cultures, the levels of 3-hydroxy-2- butanone and hexanoic acid greatly increased.^[65] As for other LAB, Lb. acidophilus has the ability to produce more acetic acid and acetaldehyde^[66,67] and Lb. pentosus has been shown to efficiently produce ethanol, 2,3-butanedione and acetic acid.^[22]

Bifidobacteria have been widely incorporated into dairy products in the last decade due to their potential health benefits. Bifidobacteria use a distinctive “bifid shunt” for carbohydrate fermentation and convert lactose into acetic acid and lactic acid in the proportion of 3:2.^[68] Thus, the final products obtained with bifidobacteria often exhibit a characteristic aroma and a slightly acidic flavour.^[69,70] High levels of acetic acid impart a “vinegary” taste that may not be accepted by consumers.^[13] However, the incubation temperature greatly influences the ratio of acetic to lactic acid. In products fermented at 30 or 37°C, the ratio between lactate and acetate was 3:2, but when strains were incubated at 42°C or 45°C, the ratio was lower than the theoretical one.^[63,71] Furthermore, bifidobacteria have also been reported to contribute to acetaldehyde and acetoin formation in yogurt.^[72]

These results suggest that in co-fermentation with traditional starters, the probiotic strains do not influence the key aroma-forming volatile metabolites in yogurt, but produce different amount of metabolic products after a defined fermentation time and temperature. A number of environmental factors including composition of the culture medium, nutrient competition, and the interaction between microorganisms affect the flavour formation in yogurt. It is essential to control the use level of probiotics and the fermentation parametersto optimise the organoleptic effect of probiotic yogurt.

Metabolic engineering: Application for flavour enhancement

With the advent of molecular biology, various strategies have been applied to LAB to genetically engineer strains that can over-produce certain flavour metabolites, such as acetaldehyde, diacetyl, and esters. As stated before, there are many pathways for the production of acetaldehyde by yogurt bacteria, among which acetaldehyde production from threonine catalysed by threonine aldolase is thought to be the key pathway. In S. thermophilus, its serine hydroxymethyltransferase (SHMT), encoded by the glyA gene, also possesses threonine aldolase activity. Chaves et al. found that inactivation of the glyA gene resulted in a complete loss of acetaldehyde formation, and overexpression of the gene in S. thermophilus results in an increase in acetaldehyde production by 80–90%.^[23,73] Another attempt to improve acetaldehyde production through metabolic engineering for Lc. lactis was reported by Bongers et al.^[74] In that work, overexpression of pyruvate decarboxylase (pdc) from Zymomonas mobilis in Lc. lactis, controlled by the NICE system, rerouted pyruvate metabolism towards acetaldehyde. To further increase pyruvate availability, the NADH oxidase gene (nox) was also overexpressed. The double mutant converted almost 50% of the consumed glucose to acetaldehyde under anaerobic conditions.

Due to its importance in yogurt flavour, efficient diacetyl production has been the aim of several metabolic engineering strategies for LAB. The key enzymes involved in diacetyl formation include α-acetolactate synthase (encoded by the als or ilvBN genes), lactate dehydrogenase (encoded by the ldh gene), and α-acetolactate decarboxylase (encoded by the aldB gene). Therefore, several approaches to improve diacetyl production in Lc. lactis have been developed, such as the overexpression of the als genes or ilvBN gene^[75] and the inactivation of the ldh^[62] or aldB^[76,77] genes to increase the metabolic flux towards diacetyl. However, the effects are limited, as the precursors of diacetyl, such as α-acetolactate and pyruvate are at the centre of the metabolic pathways of organisms, and the change of a single gene may not be enough to greatly direct the metabolic flux towards diacetyl production.

Consequently, various combined strategies have been developed. In a study with Lc. lactis, ldh inactivation was combined with overexpression of the als gene in an attempt to produce high yields of diacetyl. However, the result of the engineering was the production of very high amounts of acetoin rather than diacetyl.^[78] Thus, the activity of the ALDB enzyme, which catalyses the conversion of α-acetolactate to acetoin, should be prevented. As attempts to combine both LDH and ALDB inactivation had found no success, Monnet et al.^[79] selected mutants using random mutagenesis that were deficient in α-acetolactate decarboxylase and had low lactate dehydrogenase activity. The mutants were able to overproduce diacetyl, with the maximum amount being 0.6 mmol/L. In addition, a combination of aldB deletion and increased expression of ilvBN in a Lc. diacetylactis strain was found to increase the level of diacetyl to 0.53 mmol/L.^[80]

Furthermore, the glycolytic flux in LAB is affected by the redox balance between NAD⁺ and NADH. NADH oxidase specifically utilises NADH and provides an extra route for the regeneration of NAD⁺ under aerobic conditions, thus leading pyruvate to be rerouted towards NADH-independent pathways, such as the formation of α-acetolactate and diacetyl.^[81,82] In 1998, Lopez de Felipe et al.^[27] demonstrated that overproduction of NADH oxidase can change Lc. lactis from a homolactic bacterium to a highly diacetyl-producing bacterium. Hugenholtz et al.^[83] described the combination of NOX overproduction with ALDB inactivation in Lc. Lactis. Under aerobic conditions, 80% of the carbon flux was found to be rerouted via α-acetolactate to the production of diacetyl, and the yield reached 1.6 mmol/L. To avoid the disadvantages brought by the massive expression of controlling enzyme and elimination of the branching flux, Guo et al. reported a novel strategy for fine tuning of lactate and diacetyl production in Lc. lactis through promoter engineering.^[24] Using selected promoters for the constitutive expression of the nox gene, NOX activity increased by 58.17-fold compared with the wild-type strain under aerobic conditions. Meanwhile, the corresponding diacetyl production increased from 1.07 ± 0.03 mM to 4.16 ± 0.06 mM, which is the highest value of diacetyl formation obtained by genetically modified LAB.

In metabolic engineering, little attention has been paid to dairy microorganisms for the production of esters, although this approach has great potential for controlling ester biosynthesis in fermented dairy products. An example of this is the cloning and overexpression of the esterase gene in Lc. lactis, with a nisin-controlled expression system.^[84] The enzyme can catalyse ester synthesis via an alcoholysis reaction in an aqueous environment. As a result, the recombinants produced 2–5-fold ester yields compared with the non-induced cells.^[85]

Flavour improvement strategies and future prospects

The contribution of LAB to the formation of yogurt flavour depends on their intrinsic potential to produce aroma compounds and on the way that this potential is (or is not) revealed during the fermentation process. Although much has been done for increasing the intrinsic potential for flavour enhancement of LAB, there are still some conundrums hindering the possible application in yogurt-making. Firstly, several reviews have summarised the flavour-forming pathways of LAB,^{[5,13,17,25,48]} however, many of the known flavour compounds cannot be traced back to their metabolic precursors due to our limited knowledge of the complex network of flavour-forming pathways. Such issue will benefit from improvements of functional annotation of the key enzymes in the formation of flavour compounds. Furthermore, the technological progress made in high-throughput analysis methods, such as genomics, proteomics, and metabolomics,^[59,86,87] and the use of genome-scale metabolic models^[88] has been driving the development of new approaches to understand, control and steer aroma formation in dairy fermentation processes.

Second, the flavour-related activities of LAB mostly depend on the species used for yogurt fermentation; however, some specific activities are found only in a few species. As an example, only some specific variants of Lc. lactis (e.g. Lc. diacetylactis) are capable of citrate utilisation and diacetyl accumulation. As a consequence, even small differences in the initial composition of a starter culture can have profound effects on the profile of aroma compounds produced during fermentation.^[21] Strategies have been developed to explore the diversity on both genomic and phenotypic levels. New insights into the potential of LAB to produce flavour compounds are expected from the comparison of complete genome sequences of members of the same bacterial group, genus, or species.^[89] In addition, the actual composition of mixed and complex starter cultures is highly dynamic during the process of dairy fermentation, and the microbe–microbe interactions are thought to be crucial for obtaining the desired product characteristics. The classic example is the proto-cooperation and population dynamic of S. thermophilus and Lb. bulgaricus during growth in the yogurt, as stated before. Metagenomics and metatranscriptomics are increasingly applied in food microbiology, whereas application of metabolomics approaches to food products enlarge our view of in situ microbial metabolism and the varieties of metabolites.^[20] The ultimate aim is to obtain more control of the flavour-forming process executed by the fermenting LAB. To achieve this target, some strategies are suggested: i) screening new bacteria for the detection of the presence and activity of key enzymes involved in aroma production using omics-based approaches;^[86] ii) changing the ratio among different strains in mixed starter cultures to achieve the relative abundance of specific aroma-forming strains at different steps in the fermentation process; and^[19] iii) varying physicochemical parameters to influence the microbial physiology leading to modulation of aroma production.^[60]

The limited knowledge of the complex network of flavour-forming pathways of LAB also hampers the progress of genetically modified LAB as starter cultures for industrial production of flavour compounds.^[83,89] For instance, although several metabolic engineering strategies have been designed to improve diacetyl production by LAB, the effects are limited, as diacetyl is an intermediate metabolite, and its synthesis and decomposition are controlled by many pathways and influenced by redox balance and chemical modifications.^[90] A genome-scale metabolic model that includes carbon and nitrogen metabolism and flavour-forming pathways is crucial for understanding and designing metabolic engineering strategies.^[88] The combination of these strategies and the availability of numerous genetic tools for these microorganisms will help reroute the metabolic flux towards the efficient accumulation of the desired flavour compounds.

Conclusion

In conclusion, yogurt flavour development is a complex and dynamic biochemical process influenced by the LAB used as starter cultures. The flavour-related compounds in yogurt are primarily derived from microorganism-mediated glycolysis, proteolysis, and lipolysis. For flavour enhancement to yogurt-making, strains of other LAB are incorporated as adjunct cultures and metabolic engineering are applied in laboratory for the production of flavour-related compounds. Although the factors that determine the formation of these compounds through yogurt fermentation remain poorly understood, the use of systems biology approaches will help reveal the complex physiology of LAB and steer flavour formation during yogurt-making.

Funding

This work was sponsored by Shanghai Rising-Star Program (No. 17QB1404200), “Shu Guang” project (No. 16SG50) supported by Shanghai Municipal Education Commission and Shanghai Education Development Foundation and the Ability Construction Project of the Science and Technology Commission Foundation of Shanghai (No. 15590503500).

Additional information

Funding

This work was sponsored by Shanghai Rising-Star Program (No. 17QB1404200), “Shu Guang” project (No. 16SG50) supported by Shanghai Municipal Education Commission and Shanghai Education Development Foundation and the Ability Construction Project of the Science and Technology Commission Foundation of Shanghai (No. 15590503500).