Sustainable flavor engineering in fermented foods: biochemical routes and technological applications of lactic acid bacteria

1 Introduction

Fermented foods form an important part of the human diet and have a long history of consumption since ancient times (Rezac et al., 2018; Cuamatzin-García et al., 2022). For instance, traces of cheese making have been identified in Northern Europe dating to the sixth millennium BC, as revealed by milk fat residues found in distinctive ceramic sieves designed for the separation of curds and whey (Thierry et al., 2015). Historically, fermentation progressed through spontaneous microbial action of autochthonous microorganisms. Over time, humans gradually developed empirical methods to regulate these processes, primarily for two purposes: (i) prolonging the shelf life of easily spoilable raw materials like milk, fish, meat, grains, and vegetables, and (ii) improving and broadening the organoleptic qualities of foods. Each community developed its own distinctive fermented products based on the locally available resources, which have gradually evolved into culturally significant heritage foods (Tamang et al., 2016; Lee et al., 2024). Among the sensory qualities, flavor remains one of the most critical determinants of consumer acceptance. However, flavor formation is a multifaceted process involving complex biochemical transformations of food constituents, and understanding these mechanisms is essential for directing and optimizing flavor in foods. Lactic acid bacteria (LAB) constitute an essential component of the dominant microbial group associated with most fermented foods, operating either as the sole fermentative agent or in association with yeasts, molds, and other bacteria.

While LAB are central to flavor development in fermented foods, their biochemical routes differ markedly from the pathways employed by yeasts, filamentous fungi, and non-LAB bacteria (Malekijahan et al., 2025). LAB-driven flavor formation is predominantly governed by amino acid catabolism via transamination and decarboxylation, lipid transformations mediated by strain-specific esterases, aminotransferases, and dehydrogenases and carbohydrate fermentation to lactic acid (Ağagündüz et al., 2025). These reactions primarily yield organic acids, acetoin/diacetyl, short-chain alcohols, and mild esters that impart buttery, creamy, and acidic notes characteristic of dairy and vegetable fermentations. In contrast, yeasts such as Saccharomyces cerevisiae and Kluyveromyces marxianus rely on alcoholic fermentation and Ehrlich pathways, producing higher alcohols, acetate and ethyl esters, and volatile sulfur derivatives that generate fruity, floral, or wine-like aromas (Malekijahan et al., 2025; Wang H. et al., 2025). Filamentous fungi (e.g., Penicillium, Aspergillus) contribute extensive lipolysis, β-oxidation, and secondary metabolite formation, giving rise to methyl ketones, lactones, and aromatic compounds that dominate the flavor profiles of mold-ripened cheeses and soybean-based fermentations. Non-LAB bacteria, including Brevibacterium and Propionibacterium, further diversify flavor through pathways such as methanethiol release from sulfur amino acids, propionate biosynthesis, and advanced lipolytic activity (Piwowarek et al., 2017). Collectively, LAB contribute foundational acidification and precursor generation, while yeasts, fungi, and non-LAB bacteria introduce complexity, higher-volatility esters, and signature aromatic compounds. Understanding the complementarity of these pathways is essential for designing mixed-culture fermentations and optimizing sustainable flavor engineering strategies.

In addition, many LAB species exhibit diverse metabolic activities that contribute directly to flavor development (Anumudu et al., 2024; Malekijahan et al., 2025). Although several studies have described individual biochemical routes or LAB functions in flavor development, current literature does not provide an integrated synthesis of (i) pathway-level mechanisms, (ii) strain-level enzymatic variability, and (iii) technological and ecological factors that modulate LAB-driven aroma formation (Pereira et al., 2025). Moreover, the flavor-forming potential of LAB from traditional and region-specific fermented foods remains insufficiently characterized (Hu et al., 2022). This review addresses these gaps by providing a consolidated biochemical and technological perspective on LAB-mediated flavor generation.

This study emphasizes the function of LAB in the development of flavor compounds within fermented delicacies. It outlines the major metabolic routes by which LAB influence flavor, as well as the environmental and biological factors that determine their contribution (Sawant et al., 2025). Advances in omics and meta-omics technologies that provide new insights into flavor biogenesis are also discussed (Chen Q. et al., 2025). However, the scope does not include applications of genetically modified LAB or their industrial use outside food systems for aroma production. A bibliometric survey of publications from the past three decades related to LAB and flavor formation in fermented foods reveals that about half of these studies focus on dairy products, as illustrated in Figure 1. Consequently, many examples discussed in this chapter are drawn from dairy systems.

Although dairy products remain the most extensively studied systems for LAB-mediated flavor formation, important contributions of LAB are also evident in a wide range of non-dairy fermented foods (Ağagündüz et al., 2025). In cereal-based fermentations such as sourdough, idli, and fermented millet products, LAB contribute sour, malty, and buttery notes through organic acid and aldehyde production (Tsafrakidou et al., 2020). In fermented vegetables including kimchi, sauerkraut, and pickles, LAB generate acids, esters, sulfur compounds, and alcohols that shape pungent and fresh flavor profiles (Gangakhedkar et al., 2025). Fermented bamboo shoots such as Khorisa, widely consumed across Northeast India and Southeast Asia, also depend heavily on microbial activity for the development of characteristic sour, earthy, and sulfurous aromas (Borthakur et al., 2025a). Fish-based fermentations such as Teles, Ngari and Hentak, rely on microbes for the production of savory, cheesy, and sulfurous volatiles (Borthakur et al., 2025b; Sharma et al., 2025; Borthakur et al., 2026). Similarly, soybean fermentations, cocoa fermentations, kombucha, coffee fermentations, and fermented beverages involve complex interactions between LAB and other microorganisms that contribute significantly to flavor complexity (Malekijahan et al., 2025). These examples highlight the broad relevance of LAB beyond dairy systems and underscore the need for further investigation of flavor formation in diverse traditional and region-specific fermented foods.

2 Flavor and odor compounds

Flavor refers to the combined perception of flavor and aroma stimuli generated by a food or beverage within the oral cavity and is typically assessed through organoleptic analysis (Le Quéré and Schoumacker, 2023). It originates from two primary sources: (i) a small group of hydrophilic compounds responsible for the fundamental taste modalities comprising saltiness, sweetness, bitterness, sourness, and umami—and (ii) a diverse array of odor-active volatile compounds that impart the characteristic flavors of fermented products.

2.1 Diversity and analytical approaches to volatile compounds

Fermented delicacies are particularly rich in volatile substances, with more than 550 identified in cocoa, about 200 in fermented sausages, around 300 in bread, nearly 200 in Cheddar cheese, and up to 300 in Comté cheese (Thierry et al., 2015; Bis-Souza et al., 2019). Microorganisms are capable of producing several hundred volatile organic compounds (VOCs), most of which have molecular masses ranging between 100 and 300 g/mol (Poveda, 2021). To characterize these molecules, gas chromatography coupled with mass spectrometry (GC–MS) is widely employed, following extraction of volatiles from the food matrix. The extraction technique strongly influences both the yield and the spectrum of compounds detected (Belwal et al., 2018). Among these, headspace methods have become the most prevalent, as they are comparatively straightforward, reproducible, and more closely aligned with sensory perception—since they mainly capture the most volatile, aroma-relevant compounds (McAdam et al., 2020). Recent advancements in GC–MS technology include improved separation by comprehensive two-dimensional gas chromatography and accelerated data processing through metabolomics-based workflows (Higgins Keppler et al., 2018).

2.2 Analytical techniques for detection and characterization of volatile compounds

Accurate detection and characterization of volatile organic compounds (VOCs) in fermented foods require sensitive analytical platforms capable of resolving complex mixtures of odor-active metabolites (Epping and Koch, 2023). GC–MS remains the most widely used tool for VOC analysis, offering high sensitivity and robust identification based on mass spectral libraries. Sample preparation methods strongly influence the VOC profile obtained. Among these, headspace solid-phase microextraction (HS-SPME) is particularly favored due to its solvent-free operation, ability to concentrate trace volatiles, and strong alignment with sensory perception (Gao et al., 2024). Dynamic headspace extraction, although more labor-intensive, enables the recovery of a broader range of compounds, particularly those with lower vapor pressures.

In most studies investigating LAB-derived volatile compounds, identification is performed using gas chromatography–mass spectrometry (GC–MS) coupled with electron ionization (EI, 70 eV). The separation of volatiles is typically achieved using non-polar capillary columns such as DB-5, HP-5, or RTX-5, which provide robust separation of alcohols, esters, ketones, and hydrocarbons, or polar columns such as DB-Wax, Carbowax, or FFAP, which are particularly effective for resolving organic acids, aldehydes, and polar esters. Quadrupole mass analyzers are most commonly employed due to their reliability and compatibility with spectral libraries, whereas time-of-flight (TOF) mass analyzers are frequently used in comprehensive two-dimensional GC (GC×GC–MS) for enhanced resolution of complex matrices. Compound identification is generally based on comparison of mass spectra with reference libraries such as NIST, retention index matching using homologous n-alkane series, and, confirmation with authentic standards.

To identify aroma-active compounds specifically responsible for sensory attributes, gas chromatography–olfactometry (GC–O) is employed. This technique combines chromatographic separation with human sensory detection, facilitating the identification of “character-impact” molecules that may be present at low concentrations but possess very low odor thresholds (Li L. et al., 2024). Recent advances in comprehensive two-dimensional gas chromatography coupled with time-of-flight mass spectrometry (GC×GC–TOFMS) have further enhanced chromatographic resolution, enabling the detection of co-eluting or matrix-bound volatiles that are often missed in conventional systems (Zhang et al., 2023). Metabolomics-assisted workflows—based on high-throughput data processing, chemometric analysis, and feature-based annotation—provide additional capability for profiling VOCs at scale. These approaches support rapid identification of biomarkers associated with fermentation stage, microbial activity, or flavor defects.

Because extraction and detection techniques differ in sensitivity and selectivity, the choice of method substantially influences the VOCs detected. Highly volatile, aroma-relevant compounds are preferentially extracted by HS-SPME, whereas more hydrophobic or thermally labile molecules may require dynamic headspace or solvent-based extraction (Dadalı and Elmacı, 2019). Therefore, analytical method selection must be aligned with the specific chemical classes of interest and the sensory objectives of the study. The analytical techniques described here complement the biochemical pathways discussed in the subsequent sections and provide a foundation for understanding how instrumental data correlate with sensory perception.

2.3 Role of volatile compounds in flavor development

Not all volatile molecules are odor-active, and only a subset contributes to the perception of flavor. To identify these key compounds, analytical techniques must be combined with sensory evaluation. Gas chromatography coupled with olfactometry (GC–O) is a commonly employed technique to pinpoint aroma-active volatiles in foods, leading to the establishment of databases of “character-impact” aroma molecules (Chen X. et al., 2025). These compounds span several chemical families, such as ketones, acids, aldehydes, esters, and sulfur-containing volatiles (Brattoli et al., 2013; Lo et al., 2024). Table 1 presents selected aroma compounds generated by LAB in fermented delicacies, whereas Table 2 summarizes major odor-active compounds, their precursors, metabolic pathways, producing microorganisms, sensory descriptors, and associated fermented foods. A comprehensive list of reported LAB-derived volatiles is provided in Supplementary Table 1.

Table 1. Volatile flavor compounds generated by LAB and their occurrence in various fermented dairy matrices.

Table 2. Major odor-active compounds produced during LAB-mediated fermentation, their precursors, metabolic pathways, producing microorganisms, sensory descriptors, and associated fermented foods.

Table 1 reveals that LAB-driven flavor formation in fermented dairy and non-dairy matrices is dominated by five major classes of volatile and semi-volatile compounds: organic acids, alcohols, esters, carbonyl compounds, and sulfur-containing metabolites. The structures of the volatile organic compounds have been tabulated in Supplementary Table 2. Among these, short-chain organic acids (e.g., lactic, acetic, butyric acids) contribute foundational sour, buttery, and tangy notes that define early-stage fermentation profiles. Alcohols and esters, particularly ethanol, isoamyl alcohol, ethyl acetate, and ethyl hexanoate, emerge largely through heterofermentative pathways and amino acid transamination, imparting fruity, floral, and creamy nuances. Carbonyl compounds such as diacetyl, acetoin, and aldehydes derived from lactose and citrate metabolism provide critical buttery, green, and malty attributes, often serving as high-impact flavor molecules even at low thresholds. Volatile sulfur compounds, though produced in trace quantities, exert disproportionate sensory influence by contributing characteristic savory, meaty, or sulfurous notes typical of advanced proteolysis. Collectively, these chemical families demonstrate the central role of LAB in modulating flavor complexity by integrating carbohydrate, lipid, and protein catabolic routes, thereby shaping the sensory identity of diverse fermented foods.

Aroma thresholds can vary substantially, often differing by several orders of magnitude. For instance, ethyl butanoate can be detected at about 1 ng/g in water, whereas ethanol requires concentrations closer to 10⁵ ng/g. In foods, detection thresholds are influenced not only by intrinsic air thresholds but also by factors such as solubility in water and lipids, and the physicochemical composition of the food matrix. For instance, lipid content can serve as a reservoir that retains volatile molecules, whereas pH affects the dissociation and availability of organic acids. Consequently, an identical array of odor compounds—whether inherently present or generated in-situ by LAB—may be experienced in a distinctly different manner depending on the food environment where they are present.

As the concentration of a volatile compound alone does not necessarily reflect its sensory importance, modern flavor studies increasingly use odor activity value (OAV) and relative odor activity value (ROAV) approaches to identify key aroma contributors (Lester et al., 2021; Ma et al., 2025). Representative OAVs of major LAB-derived aroma compounds are summarized in Table 3. OAV is defined as the ratio between the concentration of a compound and its odor threshold in a given matrix, whereas ROAV provides a relative ranking of aroma importance within a product. Compounds with OAV values greater than 1 are generally considered capable of contributing directly to aroma perception, while those with very high OAV values often act as character-impact compounds (Cen et al., 2026). For example, diacetyl, methanethiol, ethyl butanoate, dimethyl disulfide, and 3-methylbutanal may exert stronger sensory effects than compounds present at much higher concentrations because of their extremely low odor thresholds. Therefore, the integration of GC–MS, GC–O, OAV, and ROAV analyses is essential for linking analytical chemistry with actual sensory perception in fermented foods (Tang et al., 2024). The relative contribution of major LAB-derived odor-active compounds across different fermented food matrices is illustrated in Figure 2 (Ganesan et al., 2007; Papaioannou et al., 2021; Song et al., 2021; Wang J. et al., 2025).

Table 3. Representative odor activity values (OAVs) of key LAB-derived flavor compounds.

The sensory impact of a given volatile compound is also highly context-dependent: it may impart desirable or undesirable notes as influenced by the type of product and the concentration levels involved. For instance, short-chain free fatty acids are major contributors to the pungent notes of many cheeses, as tabulated in Table 1, but when present in excess, cause rancid off-flavors (Štefániková et al., 2020; Xu et al., 2025). Similarly, 3-methylbutanal, derived from leucine catabolism, enhances desirable notes in cheeses such as Proosdij, yet is regarded as a defect-causing compound in Cheddar (Meng et al., 2021; Zheng et al., 2021). Interestingly, many volatile organic compounds typically associated with microbial spoilage of raw meat are chemically like those that positively shape the aroma of ripened cheeses (Bleicher et al., 2022; Ritschard and Schuppler, 2024).

2.4 Origin of aroma compounds

Microbial enzymes and naturally occurring endogenous enzymes both contribute significantly to the formation of aroma compounds in fermented foods, yet their roles differ in timing, substrate specificity, and biochemical impact. Endogenous enzymes present in raw materials such as milk lipoprotein lipase, plant polyphenol oxidases, cereal amylases, and meat proteases initiate early hydrolysis of lipids, proteins, and complex carbohydrates before microbial populations become fully established (Mótyán et al., 2013). These enzymes generate initial pools of free fatty acids, peptides, amino acids, and simple sugars that serve as essential precursors for downstream microbial transformations. As fermentation progresses, microbial enzymes produced by LAB become the dominant drivers of biochemical conversion. LAB proteases and aminopeptidases continue peptide breakdown, releasing amino acids that enter aminotransferase, decarboxylase, or lyase-mediated reactions to produce key aroma intermediates such as aldehydes, branched-chain acids, and sulfur compounds (Mannaa et al., 2021). Similarly, microbial esterases and lipases transform endogenous free fatty acids into methyl ketones, short-chain organic acids, secondary alcohols, and esters, many of which are highly odor-active. Together, endogenous and microbial enzyme systems operate sequentially and synergistically: endogenous enzymes liberate primary substrates, while LAB- derived enzymes convert these intermediates into the wide spectrum of volatile and non-volatile compounds that define the characteristic flavor profile of fermented foods.

Flavor development in fermented foods arises from the coordinated action of endogenous enzymes naturally present in raw materials and microbial enzymes expressed by fermenting microorganisms, particularly lactic acid bacteria. Endogenous enzymes such as milk lipoprotein lipase, plant amylases, polyphenol oxidases, and tissue proteases initiate early hydrolysis of lipids, proteins, and complex carbohydrates, releasing free fatty acids, peptides, amino acids, and fermentable sugars that serve as essential precursors for microbial metabolism. As fermentation progresses, LAB-derived enzymes become the primary drivers of aroma formation. Proteases and peptidases liberate free amino acids that are converted by aminotransferases, decarboxylases, and lyases into aldehydes, alcohols, branched-chain acids, and sulfur-containing volatiles, while esterases and alcohol acyltransferases catalyze the formation of fruity esters from organic acids and alcohols. Lipid-derived flavor compounds arise largely from the transformation of free fatty acids—often released by endogenous lipases—into methyl ketones, secondary alcohols, and esters through LAB-mediated β-oxidation and esterification reactions. Thus, endogenous enzymes predominantly govern precursor availability, whereas microbial enzymes determine the diversity, intensity, and specificity of flavor compounds formed during fermentation.

3 Flavor modulation in fermented delicacies by LAB

LAB rank among the predominant microbes involved in the fermentation of foods, with numerous species fulfilling critical functional roles. Their presence in food matrices can arise from: (i) natural contamination from raw ingredients or the surrounding environment, (ii) transfer from earlier fermented batches through back-slopping, (iii) deliberate introduction via defined starter cultures, or (iv) a combination of these approaches, as summarized in Table 4. In numerous traditional fermented products, particularly in developing regions, fermentation is still predominantly spontaneous, occurring without the inoculation of starter cultures (Siddiqui et al., 2023; Sionek et al., 2023). Conversely, contemporary food fermentations have increasingly adopted the systematic use of starter cultures to achieve uniformity, safety, and product quality. At present, defined starter cultures are commonly applied in the generation of different fermented delicacies, like sourdough bread, dairy products and dry-fermented sausages (Zabat et al., 2018).

Table 4. LAB species linked with different fermented food products derived from various raw materials.

Lactobacillus is regarded as the most widespread genus of LAB, with species such as L. plantarum detected in nearly all types of fermented foods (Fusco et al., 2023; Zhao et al., 2024). Numerous other LAB are also involved in food fermentations: S. thermophilus, L. lactis, and Streptococcus gallolyticus subsp. macedonicus; L. mesenteroides and Leuconostoc lactis; Pediococcus spp.; Carnobacterium maltaromaticum; and Enterococcus spp. in dairy systems. In alcoholic fermentations, Oenococcus oeni plays a crucial role in wine and beer, while L. mesenteroides is typical of sauerkraut (Thierry et al., 2015; Link et al., 2021; Ashaolu et al., 2025). In cereal-based fermentations, Weissella species are common (Song et al., 2021; Zhao and Gänzle, 2025), whereas Tetragenococcus halophilus dominates fish sauce fermentations (Zheng et al., 2024; Tigrero-Vaca et al., 2025). Food microbiology has been revolutionized by recent developments in molecular biology tools and genomic sequencing technologies, which have enabled detailed characterization of microbial consortia. High-throughput tools such as PCR-based rDNA profiling and, more recently, shotgun metagenomics have revealed the microbial ecology of fermented foods with unprecedented resolution (Agyirifo et al., 2019; Maoloni et al., 2020). These approaches have uncovered microbial diversity in products ranging from kimchi to cocoa beans (Chandel, 2021; Mayo et al., 2021). Remarkably, even in extensively studied systems like cheese, new microbial players have been identified that had previously gone unnoticed with culture-dependent techniques (Li et al., 2023).

The presence of LAB does not automatically mean that they are the main drivers of aroma formation. Their actual contribution depends on three factors: (i) their metabolic capacity to generate aroma molecules from food substrates, (ii) the extent to which they express these activities under in situ conditions, and (iii) their relative importance in comparison with other microbes or enzymes active in the same ecosystem. These aspects are explored in greater depth in the following sections.

3.1 Biochemical mechanisms of flavor development in foods and the functional role of LAB

The enzymatic pathways responsible for the conversion of carbohydrates, proteins, and lipids into volatile flavor compounds by LAB are highly coordinated and proceed through distinct but interconnected metabolic routes. Carbohydrate-derived flavor formation begins with glycolysis, in which glucose is converted to pyruvate through the actions of hexokinase, phosphofructokinase, and aldolase (Bergman et al., 2016). Pyruvate then serves as a central branching metabolite and is further transformed into diacetyl, acetoin, and 2,3-butanediol via α-acetolactate synthase, α-acetolactate decarboxylase, and diacetyl reductase (McAuliffe et al., 2019). Additional enzymes such as pyruvate decarboxylase, alcohol dehydrogenase, acetate kinase, and alcohol acyltransferases facilitate the formation of ethanol, acetate, and short-chain esters. In parallel, heterofermentative LAB utilize the phosphoketolase pathway in which phosphoketolase cleaves xylulose-5-phosphate to yield acetyl-phosphate and CO₂, leading to the production of acetic acid, ethanol, and additional esters (Mailloux, 2024).

Protein-derived flavor compounds arise from proteolysis, where proteases and peptidases release free amino acids that undergo transamination by aminotransferases to yield α-keto acids (Chen et al., 2023). These α-keto acids are subsequently decarboxylated or reduced by decarboxylases and dehydrogenases to generate aldehydes, alcohols, and organic acids, while sulfur-containing amino acids such as methionine are cleaved by methionine γ-lyase or cystathionine lyase to produce methanethiol and its derivatives (DMDS, DMTS) (Pott et al., 2019).

Lipid-derived flavor compounds originate from lipase- and phospholipase-mediated hydrolysis of triglycerides and phospholipids to produce free fatty acids (FFAs), which undergo β-oxidation via acyl-CoA dehydrogenase and related enzymes to form methyl ketones, secondary alcohols, and lactones (Malekijahan et al., 2025). FFAs may also react with alcohols through the action of esterases and alcohol acyltransferases to yield a diverse array of short-chain esters. Collectively, these carbohydrate-, amino acid-, and lipid-derived pathways, and the specific enzymes that drive them, contribute to the complex bouquet of flavor-active metabolites that characterize LAB-driven fermentations.

LAB are equipped with enzymatic systems that can degrade diverse food constituents such as proteins, complex carbohydrates, lipids and peptides—into flavor precursors, which are then converted into volatile aroma compounds (Malekijahan et al., 2025). Mapping these metabolic routes and identifying enzymatic bottlenecks are key to pinpointing the genes and enzymes that control aroma production. This knowledge also enables systematic genome mining to evaluate the metabolic potential of different species and even strain-level variations. While the extracellular degradation of proteins and polysaccharides is carried out by secreted or cell-surface-associated enzymes, most subsequent reactions take place within the cell (Dominici et al., 2022). For certain transformations, such as esterification, the precise cellular location remains uncertain. Interestingly, although some flavor molecules correspond to end-products of primary carbohydrate metabolism, the majority are secondary metabolites whose physiological functions in bacteria are still unclear and whose biosynthetic pathways are often only partially understood (Iarusso et al., 2025).

LAB-mediated flavor formation proceeds through three interconnected metabolic routes involving carbohydrate, amino acid, and lipid metabolism. Carbohydrate catabolism of fermentable sugars via the Embden–Meyerhof–Parnas or phosphoketolase pathways converges at pyruvate, which serves as a central metabolic node. Pyruvate is primarily reduced to lactic acid but may also be redirected toward aroma-active metabolites, including diacetyl, acetoin, and 2,3- butanediol, through α-acetolactate–dependent reactions. In heterofermentative LAB, acetyl-phosphate further contributes to the formation of acetic acid, ethanol, and short-chain esters.

Amino acid–derived flavor formation originates from proteolysis and peptidolysis, releasing free amino acids that undergo transamination to α-keto acids. These intermediates are subsequently converted into aldehydes, alcohols, organic acids, and sulfur-containing compounds through decarboxylation, reduction, or lyase-mediated reactions, with pathway flux strongly influenced by strain-specific enzymatic capacity.

Lipid-derived pathways are initiated by lipase- and esterase-mediated hydrolysis of triglycerides, yielding free fatty acids that are either directly aroma-active or further transformed into methyl ketones, secondary alcohols, lactones, and esters via β-oxidation and esterification reactions. Collectively, these pathways illustrate how primary metabolism in LAB is rerouted toward secondary reactions that generate the volatile compounds responsible for the characteristic flavors of fermented foods.

3.2 Production of flavor metabolites through carbohydrate fermentation by LAB

During food fermentation, LAB primarily utilize naturally occurring sugars present in raw materials, including lactose in milk, fructose, sucrose and glucose in plant-based ingredients (Robles-Iglesias et al., 2023). In some cases, carbohydrates are also generated during food processing through enzymatic hydrolysis of polysaccharides, such as starch breakdown in cereals or pectin degradation in fruits and vegetables. Carbohydrate fermentation by LAB primarily leads to lactic acid production, thereby reducing the pH and contributes to the characteristic sour taste of many fermented foods. Beyond acidification, the fermentation process also generates additional metabolites that directly or indirectly impact flavor. Depending on the metabolic pathway employed—homofermentative or heterofermentative—different by-products are produced.

• Homofermentative LAB predominantly metabolize hexoses to lactic acid through the Embden–Meyerhof–Parnas (EMP) pathway, leading to a clean acidic taste but relatively few aroma-active volatiles.
• Heterofermentative LAB, in contrast, metabolize sugars through the phosphoketolase pathway, leading to lactic acid accumulation along with acetic acid, ethanol, CO₂, and various volatile compounds that enrich the aroma profile of fermented products.

Additionally, sugar metabolism provides important precursors for the synthesis of esters, aldehydes, ketones, and other volatile compounds. For instance, heterofermentative metabolism contributes significantly to the fruity and buttery notes in certain dairy fermentations and vegetable-based products. The diversity and intensity of these flavor-active metabolites depend not only on the type of LAB species and their metabolic pathways but also on environmental factors such as sugar availability, fermentation temperature, pH, and interactions with co-existing microorganisms.

3.2.1 Formation of flavor compounds through carbohydrate fermentation by LAB

LAB primarily utilize the sugars naturally occurring in raw materials as carbon sources in fermented foods, including lactose in milk and fructose, glucose and sucrose in plant-based substrates (Blaya et al., 2018) and, in some cases, added sugars such as glucose in fermented sausages. Sugars may also be liberated during fermentation through the enzymatic breakdown of complex carbohydrates. For example, amylolytic LAB, including strains of L. plantarum, Limosilactobacillus fermentum, and related species, are capable of degrading starch in cereal-based fermentations (Wang J. et al., 2021; Fusco et al., 2023). Beyond sugars, LAB can metabolize alternative carbon sources including organic acids such as citrate and malate, polyols, and amino acids, either in combination with sugars or, in some cases, as sole energy substrates (Vermelho et al., 2025). The precise substrates exploited by certain groups, such as non-starter LAB (NSLAB) that proliferate in the later stages of cheese ripening, following lactose depletion by starter cultures, remain a topic of debate (Eicher et al., 2024). The range of carbon sources available largely determines the dynamics of LAB populations and their succession during food processing.

Carbohydrate metabolism in LAB can proceed via either homofermentative or heterofermentative routes, depending on the strain, the environmental conditions and the available substrate. Homofermentative metabolism produces lactic acid almost exclusively, which contributes sourness but little aromatic diversity. By contrast, heterofermentative metabolism yields lactic acid alongside other products like acetic acid, CO₂, ethanol, and key flavor-active volatiles. For example, acetaldehyde plays an important role in yogurt and cheese flavor; acetic acid provides complexity in many cheeses; and ethanol serves as a precursor for fruity ethyl esters (Verma et al., 2022). The sequential bioconversion of glucose and citrate into lactic acid and acetoin is illustrated in the metabolic framework of LAB, as shown in Figure 3.

These carbohydrate-mediated pathways are not only central to the generation of key flavor molecules such as diacetyl, acetoin, and esters but also support sustainability by reducing the need for external flavor additives. Because LAB synthesize these compounds in situ during fermentation, processors can minimize synthetic flavor supplementation, leading to cleaner labels and lower environmental footprint associated with manufacturing, packaging, and transport of chemical additives (Tian et al., 2025).

Citrate, naturally present in milk, fruits, and vegetables, is an important co-substrate for citrate-utilizing LAB. Its metabolism leads to pyruvate accumulation, which is then diverted into compounds such as diacetyl, 2,3-butanediol, acetoin and to preserve intracellular redox equilibrium (Cha et al., 2019). Some strains can also synthesize 2,3-pentanedione from pyruvate and threonine (Bork et al., 2023). Diacetyl and 2,3-pentanedione impart buttery notes and are key contributors to the sensory profile of dairy products like butter, yogurt as well as cheese (Bergamaschi et al., 2015; Yuan et al., 2023). While diacetyl is a positive attribute in wine, it is considered a spoilage marker in beer fermentations. These compounds are widely produced across LAB genera, including Lactococcus, Leuconostoc, Streptococcus, Pediococcus, Lactobacillus, and Oenococcus.

α-Dicarbonyls such as diacetyl, glyoxal, 2,3-pentanedione, and methylglyoxal are also important odor precursors. Through non-enzymatic Maillard and Strecker reactions with amino acids—under moderate conditions comparable to those observed in cheese or wine—these intermediates yield highly odor-active molecules (Sıçramaz et al., 2022; Mirza Alizadeh et al., 2025). For instance, furan derivatives with sweet, caramel-like notes have been reported in a range of cheeses (Taivosalo et al., 2018; Hernandez-Valdes et al., 2020a), in cultures of L. helveticus and L. delbrueckii (Zhang X. et al., 2022).

3.3 Formation of flavor compounds via amino acid catabolism by LAB

Amino acid (AA) metabolism is another major source of aroma-active compounds in LAB. Because of their multiple auxotrophies, LAB rely on an efficient proteolytic system to access amino acids, which has been extensively reviewed (Zhao et al., 2019). Proteins are enzymatically hydrolyzed into peptides and free amino acids, both of which influence taste. While large hydrophobic peptides can impart undesirable bitterness, smaller peptides and amino acids play a key role in developing the characteristic taste of cheese. During ripening, cell lysis releases intracellular peptidases from LAB, which continue to degrade bitter hydrophobic peptides in the cheese matrix, thereby improving flavor by reducing bitterness (Zhao S. et al., 2025).

3.3.1 Amino acid conversion and aroma formation by LAB

LAB possess a diverse set of amino acid–converting enzymes that transform aromatic amino acids (Phe, Tyr), branched-chain amino acids (Leu, Ile, Val), methionine, and threonine into volatile flavor-active metabolites (Al-Kharousi, 2025). The first step in these pathways is typically initiated by transamination, or in the case of methionine and threonine, by elimination reactions catalyzed by aminotransferases and lyases, respectively. During cheese ripening, a complex cascade of proteolysis, peptidolysis, and amino acid conversion occurs, collectively shaping the characteristic flavor profile of the product. Among these processes, transamination has been recognized as the major rate-limiting step (Mu et al., 2021; Ter-Ovanessian et al., 2025). This process requires an α-keto acid co-substrate, with α-ketoglutarate typically serving as the main keto donor for LAB aminotransferases, as illustrated in Figure 4.

Concurrently, carbohydrates, lipids, and proteins undergo metabolism through glycolysis, lipolysis, and proteolysis, yielding key intermediates such as pyruvate, fatty acids, and amino acids. Through subsequent reactions, these intermediates give rise to a broad spectrum of volatile and non-volatile flavor compounds, such as organic acids, alcohols, aldehydes, esters, sulfur compounds, and diacetyl. The diagram further emphasizes the pivotal enzymatic functions of hydrolases, esterases, proteases, aminotransferases, and decarboxylases in driving the biosynthesis of these flavor compounds during cheese maturation. The major biochemical routes by which LAB convert amino acids into volatile flavor compounds are shown in Figure 5.

Recent studies have demonstrated that the formation of potent aroma compounds in cheese could be significantly enhanced either by supplementing curd with exogenous α-ketoglutarate or by employing a genetically engineered L. lactis strain that produces α-ketoglutarate from glutamate via glutamate dehydrogenase (Gdh) activity (Shahidi and Hossain, 2022). Thus, Gdh activity is considered a key functional trait for selecting LAB strains with high aroma-forming potential (Fadimu et al., 2022). Numerous studies have clarified that CitP, the citrate uptake system, aids the uptake of external α-ketoglutarate, with oxaloacetate serving as an alternative keto donor in L. lactis (Thierry et al., 2015).

The α-keto acids generated through transamination are not odor-active themselves but serve as central intermediates that can undergo several downstream reactions. These include: (i) reduction to α-hydroxy acids, which are generally flavor-neutral; (ii) decarboxylation to aldehydes, which may subsequently be reduced to alcohols; and (iii) oxidative decarboxylation to acyl-CoA derivatives, which can be converted into carboxylic acids, as illustrated in Figure 4. The enzymes mediating these conversions have been thoroughly described in several reviews (Chandra et al., 2020; Zareie et al., 2023). Their activity varies widely among LAB species and strains. In addition, certain α-keto acids, especially those derived from aromatic amino acids and methionine, can undergo non-enzymatic oxidation under cheese ripening conditions. LAB may facilitate such chemical transformations by releasing catalytic metal ions (e.g., manganese) during cell lysis (Thierry et al., 2015).

Beyond their sensory significance, amino acid–derived pathways contribute directly to the sustainability of fermented food systems. The in-situ generation of branched-chain aldehydes, acids, aromatic compounds, and sulfur volatiles reduces the requirement for externally added flavor enhancers, many of which have high manufacturing and transport footprints. LAB-mediated proteolysis and transamination also valorize native nitrogen pools in raw materials by converting peptides and amino acids—often released through endogenous proteases—into high-impact aroma compounds (Grabner et al., 2017). This intrinsic biochemical recycling minimizes wastage of proteinaceous substrates and supports cleaner-label formulations. Moreover, strains with efficient amino acid catabolism can shorten ripening periods in cheese and meat fermentations, thereby lowering the energy inputs associated with prolonged maturation and cooling. Collectively, LAB protein-derived pathways support more resource-efficient and additive-free flavor development in sustainable fermentation processes.

Overall, amino acid catabolism by LAB gives rise to a wide spectrum of odor compounds: aldehydes imparting malty character, branched-chain acids with cheesy notes, aromatic aldehydes and alcohols with floral qualities, and volatile sulfur-containing compounds reminiscent of cabbage, garlic or cooked potato, as tabulated in Table 1. The in-situ generation of amino acid–converting enzymes is modulated by numerous technological and environmental factors.

3.4 Flavor compounds produced from lipid metabolism by LAB

Lipids present in raw materials also serve as important precursors of aroma compounds in fermented delicacies. Through triglyceride hydrolysis, free fatty acids (FAs) are liberated, particularly medium-and short-chain FAs, which impart sharp, pungent, and sometimes rancid aroma notes that are characteristic of many ripened cheeses (Sarnaik et al., 2023). Even a modest degree of lipolysis—for example, the release of only ~0.5% of total fat in cheese—is sufficient to generate FA concentrations high enough to noticeably influence flavor. Beyond their direct sensory impact, free FAs act as precursors for a variety of secondary odor compounds, such as methyl ketones, esters, lactones and secondary alcohols (Anal, 2019). The release of free FAs is mediated by lipolytic esterases (lipases), enzymes classified as carboxylesterases that act on long-chain acylglycerols insoluble in water (Reyes-Reyes et al., 2022). However, most LAB strains harbor only intracellular esterases, with the notable exception of L. fermentum, where a cell surface-associated esterase has been identified (Lee et al., 2020). Consequently, lipid hydrolysis by LAB is limited during active growth and becomes significant only upon cell lysis, when esterases are released into the food matrix.

LAB esterases are generally weakly lipolytic and display limited activity on natural food lipids. Their enzymatic activity has been predominantly examined on soluble synthetic esters, whereas their capacity to hydrolyze triglycerides or diglycerides is considerably lower. For example, crude extracts of L. lactis and S. thermophilus exhibit minimal activity toward triglycerides containing fatty acids longer than C4, or diglycerides containing fatty acids longer than C6. Despite this limited primary lipolysis, LAB play an important indirect role in lipid-derived aroma formation. The free FAs released by endogenous milk lipases, or by coexisting lipolytic microbes in mixed fermentations, provide essential precursors that LAB can transform through esterification, β-oxidation, or reduction into a wide array of aroma-active volatiles. LAB esterases are also active on monoglycerides containing medium- or long-chain fatty acids (Wang et al., 2022b). Consequently, the overall lipolytic activity of LAB depends on the type of lipids present and the extent to which they have been pre-hydrolyzed, for example, by endogenous milk lipoprotein lipase.

In addition to hydrolysis, LAB esterases can catalyze the direct ester formation through transferase (alcoholysis) reactions in aqueous systems, whereby fatty acyl groups from glycerides are transferred to alcohols. This enzymatic esterification represents an important but still underexplored mechanism in LAB-mediated aroma formation (Thierry et al., 2015). Similarly, the specific pathways by which LAB generate other lipid-derived volatile compounds remain an area requiring further investigation. The major enzymatic routes by which LAB convert free fatty acids into methyl ketones, lactones, and esters are illustrated in Figure 6.

Lipid-derived pathways also offer notable sustainability advantages. The ability of LAB to transform free fatty acids—many of which originate from endogenous lipoprotein lipases in milk or naturally occurring lipases in plant materials into methyl ketones, secondary alcohols, and short-chain esters provides a biological alternative to chemically synthesized flavoring agents (Logsdon et al., 2025). These microbial conversions diversify flavor without requiring high-energy chemical synthesis steps or solvent-intensive extraction processes typically used in the flavor industry. Enhanced lipolytic or esterifying activity in selected LAB strains can additionally reduce the aging time required to achieve mature flavor profiles in cheese and fermented meats, thereby lowering energy consumption during cold storage and long ripening cycles. The synergy between endogenous and microbial lipases represents a low-input strategy that leverages natural biochemical potential to generate complex flavor bouquets with minimal environmental burden.

3.5 Biosynthetic pathways for ester production

Various alcohols and acids generated through carbohydrate, amino acid, and lipid metabolism serve as substrates for esterification reactions. Numerous esters, often responsible for fruity aromas, are detected in various fermented foods. Esterases from LAB are capable of catalyzing ester formation from free alcohols and acids (Zhu et al., 2022). Cheeses typically contain organic acids at high concentrations (10–10⁴ μg/g), while ethanol, the main alcohol, is present in smaller amounts. The limited availability of ethanol can constrain the formation of ethyl esters in cheeses including Cheddar and Emmental (Abeijón Mukdsi et al., 2009).

3.6 Interspecies and intraspecies variability in aroma production

The capacity of LAB to generate flavor compounds varies significantly between species, with some enzymatic activities restricted to only a few species. For instance, branched-chain α-keto acid decarboxylase, which converts branched-chain amino acids into malty aldehydes, has been identified in L. lactis but not in the Leuconostoc or Lactobacillus isolates tested (Guo et al., 2025). Within a species, considerable strain-to-strain variability exists, both at the genomic and phenotypic levels.

To explore this variety, strategies combining genomic and phenotypic analyses have been developed. Comparing complete genome sequences of numerous strains within the same species or genus provides new insights into their aroma-forming potential (Voidarou et al., 2020). The number of sequenced LAB genomes continues to grow rapidly, with over 140 draft or complete sequences publicly available (Luo et al., 2018; Chen et al., 2024). However, in situ enzymatic activity cannot always be reliably predicted from genomics or transcriptomics data alone. Esterases require direct biochemical testing on food lipids; of the 12 predicted esterases in P. freudenreichii, only a single enzyme could hydrolyze milk triglycerides (Thierry et al., 2015).

The primary aim is to choose synergistic species and strain combinations to modulate flavor profiles in fermented delicacies (Mahdizade Ari et al., 2025; Wang F. et al., 2025). This approach has been demonstrated in cheese using L. lactis strains with complementary amino acid catabolic pathways. By combining strains with high proteolytic activity, aminotransferase capacity, or α-keto acid decarboxylase activity, the formation of branched-chain aroma compounds, particularly 3-methylbutanal, was enhanced compared with single-strain fermentations (TruChado et al., 2020; Pazos-Rojas et al., 2023).

4 Effects of biotic and abiotic factors on LAB-associated flavor formation

4.1 General insights into in situ flavor development in fermented foods

The generation of flavor in fermented delicacies stems from the interplay between microbes and their environment. The microbial ecosystem in fermented delicacies is influenced by both the raw material composition and processing conditions, with raw materials providing substrates for microbial metabolism as well as endogenous enzymes and incidental microbes. The processing parameters, encompassing sequential operations like heat treatment, mixing, and fermentation, and the deliberate addition of starter or adjunct cultures, as well as exogenous enzymes—further guide the development of fermentation and, consequently, flavor formation, as illustrated in Figure 7.

LAB express flavor-forming enzymes in actively growing cells, resting cells, and non-culturable populations (TruChado et al., 2020; Bork et al., 2023). Additionally, enzymes released from lysed cells contribute significantly to aroma generation. Both biotic and abiotic factors influence the activity of LAB in food ecosystems by modulating their growth, metabolic activity, survival, and cell lysis, as well as by providing or limiting substrates necessary for flavor-generating reactions. A viable LAB community of approximately 10⁶ cells per gram is typically considered the minimum threshold to exert a perceptible impact on the flavor of the food. Ultimately, the impact of LAB on flavor formation is contingent upon the presence and activity levels of other microorganisms or enzymatic systems within the food matrix that may dominate aroma production.

4.2 Factors influencing the flavor-forming functions of LAB

The production of aroma compounds by LAB is affected by multiple abiotic factors in fermented delicacies, such as the food constituents (nutrient availability, pH, water activity), physical parameters (temperature, oxygen levels), and changes that occur throughout processing. Such factors act as a “selective funnel” favoring microorganisms, including LAB, that can thrive and remain metabolically active under the given conditions. Processing steps, such as mixing (increasing oxygen), heat treatments, salting, or drying (reducing water activity), can further affect microbial survival and metabolism.

The depletion of primary carbon sources significantly alters the metabolism of LAB. For instance, carbon starvation can induce a viable but non-culturable (VBNC) state in L. lactis, which is linked with alternative metabolic capabilities that enhance survival, including the catabolism of branched-chain amino acids (Fleischmann et al., 2021; Ren et al., 2022). Non-proliferating LAB populations, including VBNC cells, have been detected in cheese and plant-based fermented foods, suggesting that flavor formation should consider all LAB populations, whether actively growing or non-culturable (Trejo-González et al., 2021; Peres Fabbri et al., 2024). Autolytic processes in LAB are initiated by the enzymatic cleavage of cell wall peptidoglycan by autolysins or peptidoglycan hydrolases following cell death (Kurnianto et al., 2025). Lysis releases intracellular enzymes that contribute to flavor formation differently than active cells. Autolysis during cheese ripening promotes the degradation of bitter hydrophobic peptides, while in fermented meats like salami, LAB lysis is thought to enhance flavor formation (Huang et al., 2025). Studies using bacteriocin-producing or autolysis-variant L. lactis strains in model cheeses have shown that lysis enhances early amino acid conversion (transamination) and certain amino acid–derived aroma compounds, whereas reactions requiring intact cells, such as oxidative or reductive transformations, may decrease at high lysis levels. Overall, controlled use of autolytic as well as non-autolytic strains appears beneficial (Barzideh et al., 2022).

The salt content also modulates LAB flavor activity. For example, T. halophilus isolated from 25% NaCl fish sauce displayed strain-dependent increases in free amino acids and related aldehydes (Kaushik and Anand, 2025). Biotic factors, particularly microbial interactions, further influence LAB activity. These interactions can be cooperative, neutral, or inhibitory, shaping the succession and metabolic output of microbial communities (Anastasiou et al., 2022). Interactions occur within starter strains, between non-starter and starter LAB, and between LAB and other microbes (Li et al., 2021). Microbial interactions can also affect cell lysis; for instance, autolysis of L. helveticus strains in Swiss cheese can trigger simultaneous lysis of S. thermophilus (Thierry et al., 2015).

4.3 Influence of microbial interactions and environmental factors on flavor formation

The complexity and intensity of fermented delicacy flavors are often correlated with extensive microbial variation, both at the species and strain levels, as demonstrated in cheeses from raw, pasteurized, or microfiltered milk (Ritschard and Schuppler, 2024). NSLAB grow during the process of cheese ripening, often reaching 10⁸ CFU/g, including in pasteurized milk cheese (Randazzo et al., 2021). NSLAB contribute positively to the development of mature cheese flavors but can also negatively influence flavor on the basis of the species, strains, and type of cheese (Peralta et al., 2016; Liu et al., 2025). Consequently, adjunct LAB cultures are frequently employed to control NSLAB development and reduce variability in cheese quality (Abbate et al., 2023).

Microbial interactions can also enhance flavor formation by providing substrates. For instance, ethanol produced by yeasts or heterofermentative LAB increases ethyl ester synthesis (Li et al., 2025), and wild LAB strains with glutamate dehydrogenase activity provide α-ketoglutarate, promoting transamination reactions as well as aroma compound formation (Canon et al., 2020). Designing complementary strain combinations to complete metabolic pathways is a promising strategy to optimize flavor development (Wang et al., 2022a; Murtaza et al., 2024). Despite their importance, the molecular mechanisms governing these interactions are poorly understood, with volatile profiles often differing between monocultures and mixed cultures (Bulat and Topcu, 2020).

4.4 Mixed fermentations: synergy and process control

Mixed fermentations involving LAB in combination with yeasts or filamentous fungi are increasingly recognized as important drivers of flavor complexity in fermented foods (Sidari and Tofalo, 2024). In many traditional and industrial fermentations, LAB rarely function alone; instead, they interact metabolically with co-occurring microorganisms that contribute complementary enzymatic activities and expand the spectrum of aroma-active compounds produced (Song et al., 2021). Such synergistic interactions are particularly important in products such as kefir, yogurt, sourdough, wine, beer, cocoa, cheese, and fermented soybean products.

LAB–yeast interactions are especially relevant in fermented dairy and cereal systems. In kefir and sourdough, yeasts such as Saccharomyces cerevisiae, Kluyveromyces marxianus, and Debaryomyces hansenii metabolize sugars into ethanol, CO₂, and higher alcohols, while LAB simultaneously acidify the matrix through lactic acid production (Yerlikaya et al., 2025). The ethanol generated by yeasts serves as a precursor for ethyl ester synthesis by LAB esterases, enhancing fruity and floral notes in fermented products. Yeasts may also release vitamins, amino acids, and other growth-promoting factors that stimulate LAB growth and metabolic activity. Conversely, LAB reduce pH and suppress spoilage organisms, thereby creating conditions that favor yeast-mediated aroma production. In cocoa and wine fermentations, sequential interactions between yeasts, LAB, and acetic acid bacteria are essential for the generation of esters, aldehydes, alcohols, and organic acids that define the final flavor profile (Besançon et al., 2024; Zhao LZ. et al., 2025).

Interactions between LAB and molds are equally important in cheese ripening and soybean fermentations. Filamentous fungi such as Penicillium roqueforti, Penicillium camemberti, and Aspergillus oryzae contribute strong proteolytic and lipolytic activities, releasing peptides, amino acids, and free fatty acids from proteins and lipids (Pouris et al., 2024). These compounds subsequently serve as substrates for LAB-mediated transamination, decarboxylation, esterification, and redox reactions, leading to the formation of branched-chain aldehydes, sulfur compounds, methyl ketones, secondary alcohols, and esters. LAB therefore provide acidification and precursor generation, while molds contribute extensive substrate liberation and secondary metabolite formation, together producing more complex flavor profiles than monocultures.

Despite these benefits, uncontrolled mixed fermentations may lead to several technological problems. Excessive yeast activity can result in overproduction of ethanol, CO₂, and volatile acids, causing gas defects, package swelling, undesirable alcoholic notes, or excessive sourness. Similarly, excessive mold growth may increase lipolysis and proteolysis beyond desirable levels, generating rancid, bitter, or ammoniacal off-flavors. Competition between microbial groups for nutrients may also suppress the growth of desirable starter cultures or alter expected flavor outcomes. Therefore, successful mixed fermentations require careful control of inoculation ratios, fermentation temperature, oxygen availability, pH, salt concentration, and processing time. Sequential inoculation strategies, where one microbial group is introduced before another, may also improve process stability and allow better control over the formation of key aroma compounds.

4.5 Meta-omics insights into LAB-mediated aroma biogenesis

Recent omics-driven tools have significantly expanded current understanding of the molecular processes underlying flavor formation by LAB. Genomics enables the identification of core and accessory genes encoding key flavor-forming enzymes such as aminotransferases, decarboxylases, alcohol dehydrogenases, and esterases. Comparative genomics has revealed substantial strain-level variation in these metabolic modules, explaining why closely related LAB strains often produce markedly different volatile profiles. Transcriptomics provides insight into the dynamic regulation of flavor-forming genes during fermentation, showing how environmental factors such as pH, temperature, carbohydrate availability, and citrate concentration modulate the expression of enzymes involved in pyruvate conversion, amino acid catabolism, and lipid metabolism.

Evidence from recent omics-driven investigations has directly demonstrated the contribution of LAB to flavor biogenesis in fermented foods. Comparative genomic analyses of Lactococcus lactis strains have identified marked strain-level variation in genes encoding aminotransferases, α-keto acid decarboxylases, and esterases, which correspond to differences in branched-chain aldehyde and ester production during cheese ripening. Transcriptomic studies performed at advanced stages of cheese maturation have further shown coordinated induction of amino acid catabolic pathways, particularly those involving branched-chain and sulfur-containing amino acids, concurrent with the accumulation of malty and sulfurous aroma compounds. Complementary metabolomic profiling using GC–MS in fermented milk systems inoculated with distinct Lactobacillus helveticus strains has revealed strain-specific volatile signatures, notably in diacetyl, acetoin, and amino acid–derived aldehydes, underscoring the value of integrating multi-omics approaches to resolve pathway regulation and flavor outcomes in situ.

Proteomics complements transcriptomics by profiling the abundance and activity of intracellular and extracellular enzymes, including peptidases, lipases, and redox-active dehydrogenases that directly participate in aroma biogenesis. Metabolomics, particularly GC–MS and LC–MS-based untargeted workflows, allows high-resolution mapping of pathways from primary precursors—carbohydrates, amino acids, and lipids—towards end-product volatiles such as esters, aldehydes, ketones, and sulfur compounds. When coupled with isotopic labeling, metabolomics can trace precursor-to-volatile flux, clarifying rate-limiting steps.

At the community level, meta-genomics and meta-transcriptomics have provided new insights into how LAB interact with yeasts and other bacteria in mixed fermentations, revealing cooperative or competitive metabolic exchanges that shape flavor outcomes. These approaches also identify low-abundance but functionally important non-starter LAB that become active during maturation and significantly contribute to aroma. Collectively, omics and meta-omics platforms offer an integrated systems-level understanding of LAB flavor biosynthesis, enabling predictive modelling and the rational selection or engineering of starter cultures with enhanced aroma-forming potential.

4.6 Factors determining the contribution of LAB to food flavor

The real impact of LAB on flavor depends on the enzymatic activities of LAB relative to other microorganisms and non-microbial enzymes present in the food. For example, in cheese, LAB esterases contribute little to lipolysis; most free fatty acids originate from milk lipase activity during initial processing. In Cheddar cheese ripened exclusively by LAB, the extent of lipolysis is similar to that of milk (FA/fat ≈ 0.4%), indicating that ~80% of free FAs are already present before ripening (Xiang et al., 2019). In cheeses ripened with propionibacteria and LAB (e.g., Emmental), lipolysis primarily results from propionibacteria activity, while in blue cheeses and other fungi-ripened varieties, fungal activity drives high lipolysis (Ritschard and Schuppler, 2024). Similarly, in fermented sausages, endogenous meat proteases dominate over bacterial proteases, and aroma formation by L. sakei is minor compared to coagulase-negative staphylococci (Roy et al., 2021). This accounts for why L. sakei can be employed for the preservation of meat or seafood without significantly affecting flavor.

In summary, LAB contributions to flavor are substantial in foods fermented predominantly by LAB, such as yogurt and certain cheeses. However, when LAB coexist with other enzymatically active microbes (yeasts, fungi) or flavor-rich food matrices (e.g., baked goods, kimchi), their direct impact may be limited. Nevertheless, LAB remain crucial for primary acidification, reducing redox potential and creating conditions that indirectly influence flavor reactions (Leeuwendaal et al., 2022).

4.7 Industrial relevance, current challenges, and future perspectives

LAB-mediated flavor formation holds substantial industrial relevance across dairy, meat, cereal, vegetable, and beverage fermentations. Flavor-active LAB strains are widely used as starter, adjunct, and bio-protective cultures to enhance sensory attributes, ensure process consistency, shorten ripening time, and improve the overall quality and safety of fermented foods (Shah et al., 2023). The ability of LAB to generate desirable metabolites such as diacetyl, acetoin, short-chain esters, branched aldehydes, and sulfur compounds makes them indispensable for achieving characteristic flavor profiles in yogurt, cheese, sourdough, fermented vegetables, sausages, and traditional artisanal products. The increasing demand for natural, clean-label, minimally processed foods further elevates the importance of LAB as biological flavor-forming agents, reducing the industry’s reliance on synthetic additives.

Despite this relevance, several challenges limit the full exploitation of LAB in industrial flavor design. Significant strain-to-strain variability in metabolic capacity leads to inconsistent flavor outcomes, particularly when starter cultures are applied across diverse raw materials. Mixed microbial ecosystems such as dairy or vegetable fermentations often involve complex LAB–yeast interactions that are difficult to standardize at scale. Environmental fluctuations in pH, oxygen availability, temperature, and salt concentration can dramatically alter gene expression and enzymatic activity, causing variability in volatile profiles (Coelho et al., 2022). The stability of industrial starter cultures during freeze-drying, storage, and transportation remains another constraint, as does the regulatory framework governing the approval of novel LAB strains, particularly those improved via genome editing or adaptive evolution.

Future opportunities lie in integrating systems biology, multi-omics, and process-engineering approaches to improve strain selection and fermentation control. Genomics and metabolomics can guide the discovery of high-performing cultures with strong flavor-forming potential, while CRISPR-based genome editing and adaptive laboratory evolution can be used to enhance specific metabolic pathways. Advances in precision fermentation, bioreactor design, and predictive modelling using machine learning will enable better control over flavor formation in real time. Combining selected LAB strains with co-cultures, protective cultures, or bioengineered variants may allow industry to customize and intensify flavor formation while maintaining process robustness. Together, these technological developments provide a roadmap for overcoming current challenges and unlocking the full industrial potential of LAB-driven flavor biosynthesis.

4.8 Safety, quality, and regulatory considerations

The industrial use of LAB for flavor production requires careful evaluation of safety and regulatory compliance. Many commonly used LAB species, including Lactococcus lactis, Lactiplantibacillus plantarum, Streptococcus thermophilus and Leuconostoc mesenteroides, possess GRAS status from the U.S. Food and Drug Administration or QPS status from the European Food Safety Authority (Sivamaruthi et al., 2025). However, the approval is usually strain-specific because strains within the same species may differ in safety and metabolic traits.

The novel LAB strains isolated from unconventional sources or developed through adaptive evolution or genome editing require detailed genomic and toxicological evaluation before approval. Important safety concerns include the production of biogenic amines such as histamine and tyramine, the presence of antimicrobial resistance genes, mobile genetic elements, and virulence-associated traits. Therefore, industrial strains should be screened for antibiotic resistance, toxin production, hemolytic activity, and genetic stability.

Product labeling and regulatory classification are also important considerations for LAB-derived flavor systems. The terms such as “natural flavor,” “fermented,” and “clean label” are commonly applied to foods containing LAB-derived flavor compounds because these metabolites are generated naturally during fermentation and reduce the need for synthetic additives. However, products manufactured using genetically modified or genome-edited LAB may face stricter regulatory requirements and additional labeling obligations, particularly in regions with strict GMO regulations (Ahmad et al., 2021). Consequently, robust genomic screening, toxicological validation, traceability, and strain deposition in recognized culture collections are essential for the safe industrial use of LAB starter cultures.

5 Discussion

Recent advances in microbial ecology, multi-omics technologies, and systems-level analyses have significantly expanded current understanding of LAB-mediated flavor formation. The integration of genomics, transcriptomics, proteomics, and metabolomics now allows deeper resolution of strain-level differences, key enzymatic steps, and pathway regulation underlying the conversion of carbohydrates, amino acids, and lipids into diverse aroma-active compounds. These insights provide a more predictive framework for selecting or engineering LAB strains with enhanced flavor-forming potential, while also informing process design and optimization in industrial fermentations.

Fermented foods are attracting renewed interest from both consumers and the food industry and offer opportunities to support sustainable food systems amid growing pressures on global resources (Canon et al., 2020; Sidari and Tofalo, 2024). Enhancing and diversifying flavor in these products requires further research to identify key aroma compounds, the principal biological agents responsible for their formation, and the complex interactions within microbial ecosystems.

The influence of LAB on flavor formation depends both on their intrinsic ability to produce odor compounds and on the extent to which this potential is realized during fermentation. Investigating LAB species and strains requires a combination of approaches, including in-silico genome mining, transcriptomic and proteomic analyses, and high-throughput metabolomics-based profiling of volatiles. For in-vitro studies, culture conditions critically influence outcomes, making it essential to consider the factors that modulate LAB activity in-situ. Results from screening experiments should be benchmarked against real food systems. Strong correlations have been observed between in-vitro results and actual cheese production, both under cheese-relevant culture conditions and in microcheese models (Thierry et al., 2015). Nonetheless, linking insights from isolated LAB strains to final product flavor remains challenging due to complex microbial interactions, as well as the competitive effects of non-starter bacteria (Eicher et al., 2024; Pouris et al., 2024).

The preservation of microbial diversity is also critical. A variety of LAB and non-LAB strains have been recovered from ethnic, premium fermented delicacies or natural environments represent valuable biological resources. Depositing these strains in Biological Resource Centers (BRCs) ensures long-term conservation and provides a reservoir for flavor diversification. In products such as cheeses produced with simplified starter cultures and under improved sanitary conditions, declines in flavor richness and intensity have often been observed, highlighting the potential of BRC-preserved strains to restore or enhance sensory complexity.

The development of effective LAB-based flavor cultures will benefit from multiple directions. Advances in molecular tools now allow detailed characterization of microbial diversity, including active non-viable populations, using both targeted and untargeted techniques. Metagenomics and metatranscriptomics are increasingly applied to study food microbiomes, while metabolomics provides deeper insight into in situ microbial metabolism and the spectrum of metabolites produced. These approaches will benefit from improved functional annotation of key aroma-forming enzymes, and from more effective integration of data from various methodologies, which remains an ongoing challenge (Aguiar-Pulido et al., 2016; Butowski et al., 2025).

In addition to expanding scientific understanding, the biochemical pathways described throughout this review highlights the substantial industrial and sustainability implications of LAB-driven flavor generation. The ability of LAB to synthesize key volatiles in situ reduces reliance on synthetic additives and resource-intensive chemical flavor production. Moreover, valorization of traditional fermented foods as reservoirs of diverse, high-performing LAB strains supports regional biodiversity, clean-label innovation, and low-input fermentation practices. Nevertheless, challenges such as process variability, strain instability, and limited predictive control remain. Future progress will depend on integrating omics-guided strain selection, precision fermentation, adaptive evolution, and AI-driven modelling to achieve robust, scalable, and environmentally efficient flavor formation in diverse food matrices.

Generative AI statement

The author(s) declared that generative AI was not used in the creation of this manuscript.

Any alternative text (alt text) provided alongside figures in this article has been generated by Frontiers with the support of artificial intelligence and reasonable efforts have been made to ensure accuracy, including review by the authors wherever possible. If you identify any issues, please contact us.

Publisher’s note

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.

Supplementary material

The Supplementary Material for this article can be found online at: https://www.frontiersin.org/articles/10.3389/finmi.2026.1834783/full#supplementary-material

Supplementary material Supplementary material at the source

Supplementary material Supplementary material at the source

Footnotes

Edited by: Teresa Semedo-Lemsaddek, University of Lisbon, Portugal

Reviewed by: Dibyendu Mondal, Kalsec Inc, United States

Qiaozhi Song, Chinese Academy of Agricultural Sciences (CAAS), China