Soybean fermentation: Microbial ecology and starter culture technology

Abstract

Fermented soybean products, including Soya sauce, Tempeh, Miso, and Natto have been consumed for decades, mainly in Asian countries. Beans are processed using either solid-state fermentation, submerged fermentation, or a sequential of both methods. Traditional ways are still used to conduct the fermentation processes, which, depending on the fermented products, might take a few days or even years to complete. Diverse microorganisms were detected during fermentation in various processes with Bacillus species or filamentous fungi being the two main dominant functional groups. Microbial activities were essential to increase the bean’s digestibility, nutritional value, and sensory quality, as well as lower its antinutritive factors. The scientific understanding of fermentation microbial communities, their enzymes, and their metabolic activities, however, still requires further development. The use of a starter culture is crucial, to control the fermentation process and ensure product consistency. A broad understanding of the spontaneous fermentation ecology, biochemistry, and the current starter culture technology is essential to facilitate further improvement and meet the needs of the current extending and sustainable economy. This review covers what is currently known about these aspects and reveals the limited available information, along with the possible directions for future starter culture design in soybean fermentation.

Keywords:

• Natto and Tempeh
• Sufu and Soya sauce
• anti-nutritive factors
• Bacillus fermentation
• starter cultures

Introduction

The main food plants for humans include soybeans, wheat, maize, rice, sorghum, potato, oats, cassava, and sugar beets. Among them, soybeans (Glycine max (L.) Merr.) are characterized by their high protein and oil concentrations (Hymowitz 2008). The production of soybeans worldwide increased from 160 to 350 million tons in 1998 and 2018, respectively, due to advances in genetically modified technology such as climate and herbicide-resistant crops (FAOStat 2021; Salvagiotti et al. 2008). In the west, soybeans are mostly processed into two products: soybean meal for the feed sector and soybean seed oil for salad dressings, margarine, and mayonnaise. In Asia, several traditional foods are made from soybeans, such as Soy sauce, Tofu, Miso, Tempeh, and Natto (Hymowitz 2008; Xu, Guo, and Zhang 2019).

Soybeans were first known in China, where several fermented soy products were created. They spread to the nearby nations, where other process improvements were developed. These fermented soybean products are considered a valuable source of protein and are widely consumed for their high value in nutrition, functionality, and low cost. Soybean seeds contain about 36.5% protein, 30% carbohydrate, 19.9% lipids, and 9.3% fiber (Vagadia, Vanga, and Raghavan 2017). Soybean oil contains 57.7% polyunsaturated fatty acids, 22.8% monosaturated fatty acids, and 15.6% saturated fatty acids. Soybean protein contains high levels of lysine, leucine, valine, isoleucine, phenylalanine, and threonine but lacks tryptophan, cysteine, and methionine (Wolf 1970). However, due to their off-flavor and the presence of anti-nutritional factors (ANFs) such as phytic acid, oligosaccharides, and trypsin inhibitors, there is still limited adoption of soybeans as a major protein source in the diet (Rehms and Barz 1995; Yang et al. 2018).

The soybean fermentation process, as an example of food fermentation, was originally considered a preservation technique but has now evolved into a way to create traditional nutritive food products with distinctive flavors, aromas, and textures (Bessada, Barreira, and Oliveira 2019). Soybean fermentation is generally performed either in solid-state or submerged modes or a combination of both processes (Abass et al. 2017; M. Adams and Mitchell 2002; Garrido-Galand et al. 2021; He and Chung 2020; Saharan, Sadh, and Duhan 2017; Sarkar and Nout 2014; Thirunathan and Manickavasagan 2019). Traditionally, the fermentation of soybeans occurs spontaneously, with a variety of microorganisms originating from the seed surface or the surrounding environment. Elucidation of the roles of indigenous microorganisms enabled the rational control of the fermentation process through its application in starter culture technology (Geisen 1993; Liu, Chen, et al. 2021; Liu, Chen, et al. 2022; McCue et al. 2004; Park and Kim 2020; Shin and Jeong 2015; Singracha et al. 2017; Wang, Zhang, et al. 2021). These will be covered by this review, focusing mainly on the analysis of the soybean fermentation natural microflora, its contribution to final product quality, and the potential transformation process into the starter culture technology.

Physical and chemical factors limiting the consumption of soybeans

The primary component of a soybean is a dicotyledon seed enclosed by a roughly spherical shell or hull, the form of which varies greatly depending on cultivar and growth conditions (Perkins and D. R. B. T.-P. H. of S. P 1995). Despite the high nutritive value of soybeans, they contain anti-nutritive components that limit their consumption, including trypsin inhibitors, lipoxygenase, phytic acid, and oligosaccharides, Figure 1. Studies, for instance, have linked the presence of trypsin inhibitors, a class of protease enzymes that block the function of proteolytic enzymes within the gastrointestinal system, to the poor nutritional value of soybeans (Embaby 2010). Several physical, chemical, biological and advanced techniques such as ohmic heating, microwave, ultrasound, high-pressure processing, radiofrequency, infrared, and gamma irradiation heating have been used to eliminate them or lower their amount. Although heat and chemical treatments are efficient and affordable inactivation methods, they also damage important nutrients like lysine and heat-labile vitamins (Huang, Kwok, and Liang 2004; Kaur et al. 2012; Kunitz 1947; Rouhana et al. 1996; Sessa, Haney, and Nelsen 1990; Yuan and Chang 2010). During seed germination, the endogenous proteases, which were shown to be more active at 35 °C, were primarily responsible for the decline in trypsin inhibitor levels (Avilés-Gaxiola, Chuck-Hernández, and Serna Saldívar 2018; V. Kumar et al. 2006; Nkhata et al. 2018). Trypsin inhibitors are also degraded by microbial isolates such as Rhizopus oligosporus, Actinomucor elegans, Lactiplantibacillus plantarum, and Bacillus subtilis during soybean fermentation (Adeyemo and Onilude 2013; Amadou et al. 2010; Dai et al. 2017; Huang et al. 2019).

Figure 1. Soybean consumption constraints and possible ways for their reduction/elimination. Photos of soybeans seeds dehull seeds and dicotyledon structures are included. This figure was created using BioRender (https://biorender.com/).

Similarly, several studies have highlighted the impact of other anti-nutritive components such as lipoxygenases, phytic acid, oligosaccharides, saponin, allergens, urease, lectin, and alkaloids (Aulitto et al. 2021; Dong and Saneoka 2020; Humer, Schwarz, and Schedle 2015; Karkle, Elisa Noemberg Lazzari and Beleia, 2010; Kumar et al. 2020; Shahidi and Oh 2020; Vagadia, Vanga, and Raghavan 2017; Wang, Zhang, et al. 2021; Wang et al. 1997). Various methods were applied to eliminate or minimize these factors including physical, chemical, and biological approaches (Chitra, Singh, and Rao 1996; Gao et al. 2013; Hu et al. 2019; Li et al. 2020; Shahidi and Oh 2020; Vagadia, Vanga, and Raghavan 2017; Villacrés et al. 2020). More information can be found elsewhere (Norozi, Rezaei, and Kazemifard 2022; Qin, Wang, and Luo 2022; Suprayogi et al. 2022). Overall, fermentation was able to degrade them while preserving the bioavailability of valuable and nutritive compounds, illustrating its potential application (Nkhata et al. 2018; Shirai et al. 1994).

Microbial ecology of traditional soybean fermentation

Each country produces its own artisan soybean products, using different soybean fermentation procedures. Several factors influence the distinct characteristics of flavor and texture, including the type and length of fermentation processes, salt concentration, and moisture content, as well as additional components (Devanthi and Gkatzionis 2019). The manufacturing-specific details, different ingredients, and names used in different countries are out of this review scope and can be found elsewhere (Erickson 2015; Garrido-Galand et al. 2021; Liu, Chen, et al. 2022; Luh 1995; Sarkar and Nout 2014). By understanding the microbial identities and their roles in different types of soybean fermentation, one can better control the fermentation processes and modulate the final product quality. Soybeans are generally subjected to either solid-state fermentation, submerged fermentation, or both. The microbial groups that lead the fermentation are mainly bacterial, fungal, or a mixture. Figure 2 shows the overall soybean process flow chart, the fermentation types, and the major microbial species that drive the processes. The basic concepts of these fermentation procedures are described in the following sections.

Figure 2. Process flow diagram of soybeans. This figure was created using BioRender (https://biorender.com/).

Soybean surface ecology

Limited knowledge was reported about the microbial endogenous community detected from the surface of the soybean seeds. An early report showed several fungal isolates including Fusarium spp. (including F. moniliforme, F. acuminatum, F. equiseti), Alternaria alternata and Diaporthe phaseolorum var. sojae were detected from the soybean surface (Miller and Roy 1982). For bacteria, the main predominant microorganisms were Pseudomonas whereas others belonging to Enterobacteriaceae and Proteobacteria were reported in low populations (Kuklinsky‐Sobral et al. 2004; Miller and Roy 1982; Wang, Zhang, et al. 2021). A study conducted by Mulyowidarso (1991) showed that bean surface colonization contained between 2–5 log CFU/g of the total microbial count. Several bacteria have been reported in relevant research, such as Lactobacillus spp., Streptococcus facecium, Staphylococcus epidermidis, Klebsiella ozaenae, Citrobacter diversus, and Enterobacter cloacae which were predominant, as well as yeasts such as Pichia burtonii, Candida diddensiae and Rhodotorula rubra and mold species belonging to genera Trichosporon, Aspergillus and Rhizopus which were present in lower populations (Mulyowidarso 1991; Wei, Regenstein, and Zhou 2021). The microorganisms associated with the seeds play potentially crucial roles in the fermented product quality (Parkouda et al. 2009). The surrounding environment such as the soil, water, air, workers, and fermentation equipment may also introduce several microorganisms that are detected as the fermentation progresses, such as Escherichia coli, Salmonella, Staphylococcus, Proteobacteria, Fusarium, Geotrichum, Monascus, Aspergillus, and Mucor species (Natarajan et al. 2016; Xie et al. 2019). Further investigation is required to better understand the ecology of the soybean surface and the type of microbiota present in order to manage the subsequent processes. Knowledge about the soybean surface microbiota, mainly heat-resistant thermophilic spores, is essential to optimize the sterilizing or cooking steps that are necessary to eliminate or reduce the initial contamination and to not compete with the desired microorganisms and produce unpredictable fermented products. Furthermore, although starter cultures often have the ability to lead the fermentation process, the diversity and contamination of the survival indigenous population may make them more competitive than the inoculum strains and may therefore reduce the starter culture’s ability to dominate the fermentation to 100%.

Microbiology of the soaking process

Soybeans either heat-treated or untreated are mostly subjected to a soaking step before fermentation. Natural acid fermentation develops during non-heat-treated beans causing an increase in acidity to a final pH of 4.5. Several isolates were detected during the soaking process belonging to LAB, yeasts, Enterobacteriaceae, and Bacillus species (Chauhan et al. 2022; Mulyowidarso 1991; Nurdini, Nuraida, and Suwanto 2015). Lactcaseibacillus casei, Lactiplantibacillus plantarum, Leconostoc mesenteriodes, Pedicoccus sp., and Enterococcus faecium were the predominant LAB species during the whole soaking process. Enterobacteriaceae such as C. diversus and Klebsiella pneumoniae were also detected during the early stage of the soaking process and declined at the end. Yeasts such as P. burtonii, C. diddensiae, and R. rubra, Saccharomyces dairesnsis were also detected with a weak growth rate at the initial of the soaking process. Studies have shown that the ecology of soybeans is the main origin of the leading microorganisms of acid fermentation and not the tap water. In addition, it is believed that the microbial metabolites produced during this fermentation diffuse into the beans and affect their chemical profiles (Chauhan et al. 2022; Nurdini, Nuraida, and Suwanto 2015). This soybean acidification may control microbial pathogens and undesirable microbial growth, ensuring the safety and sensory quality of the final product (Chauhan et al. 2022; Nurdini, Nuraida, and Suwanto 2015). Weak acidification was linked to high Bacillus cereus levels in the final soybean product after heat treatment (Nout, Beernink, et al. 1987; Nout, De Dreu, et al. 1987). Endogenous species such as Citrobacter freundii, Pseudomonas fluorescens, K. pneumoniae, and Streptococcus sp. were found to produce high levels of vitamin B12 (Denter and Bisping 1994). Furthermore, the soaking process was reported to reduce the level of water-soluble anti-nutritive factors such as phytic acids, but also reduce other water-soluble important components such as vitamins, minerals, and phytochemicals (Garrido-Galand et al. 2021; Reddy and Pierson 1994; Rehman et al. 2014). Heat treatment of soybean before soaking was reported to eliminate most of the LAB, yeasts, and Enterobacteriaceae species, leaving behind heat resistance spore-forming bacteria such as Bacillus brevis causing acidification failure (Wickware 2015). This promoted undesirable microbial contaminant growth during the fermentation process and adding acetic and lactic acids into soaking water was recommended as a countermeasure (Nout, Beernink, et al. 1987; Steinkraus et al. 1965; Tunçel and Göktan 1990). To the best of the authors’ knowledge, limited information is available on the acidic fermentation changes that may occur in the beans during the soaking, as well as possible strategies (i.e., a biological strategy of using LAB or yeasts) that may control this process without interfering with the subsequent processes, as will be discussed later under Bacillus fermentation.

Microbiology of soybean fermentation

After preparation, soybeans are subjected to either solid-state, submerged fermentation processes, or a sequential of both processes. Solid-state fermentation is an old method of subjecting moist solid food particles to fermentation using bacteria, yeast, or filamentous fungi to produce fermented food or metabolites such as enzymes, flavors, acids, etc. Whole soybeans or their residues are subjected to bacterial or fungal solid-state fermentation to produce various fermented products including Natto, Tempeh, and the first stage of Miso, Soya sauce, and Sufu (Pandey 2003). Each product has its characteristic flavor and texture because of the unique microbial populations found in the raw materials that dominate the fermentation. The submerged fermentation is a process conducted in the presence of free excess water (Singhania 2011). Two main soybean-fermented products are produced through a submerged fermentation stage, as part of the process in brine: Soya sauce and Sufu.

Natto

Natto is produced by bacterial fermentation and characterized by a sweet, umami, and slightly bitter taste with the texture of a viscous polymer (Kanno and Takamatsu 1987). Traditionally, it was produced by wrapping boiled soybeans in rice straw pretreated with boiling water to kill vegetative microorganisms, leaving behind spore-forming Bacillus including B. subtilis subsp. subtilis (natto), and incubating it for about 3 days (Sarkar and Nout 2014). Rice straw is also a natural habitat of other spore-forming Bacillus species that are believed to have a crucial impact on the fermented product, such as B. amyloliquefaciens B. pumilus, Bacillus licheniformis, and Bacillus atrophaeus (Kubo et al. 2011; Meerak et al. 2008). Bacillus spores are known for their resistance to environmental stress, and their vegetative cells have high enzymatic activities to degrade protein, carbohydrates, and lipids (Ouoba et al. 2003). During fermentation, fructan and polypeptides such as glutamic acid are produced which cover Natto with a stringy mucous material of unique flavor and aroma (Reddy, Pierson, and Salunkhe 2009). Furthermore, microbial lipolytic activities in soybeans cause the production of fatty acids such as oleic, linoleic, and linolenic acids with a significant impact on the final product’s sensory quality (Hu et al. 2010; Sharma et al. 2020). The endogenous Bacillus species were reported to biosynthesize a wide range of alkyl pyrazines such as methyl pyrazine, dimethyl pyrazine, trimethyl pyrazine, and tetramethyl pyrazine with known pleasant aroma (Kłosowski, Mikulski, and Pielech-Przybylska 2021). They also produce inhibitory metabolites which suppress the growth of foodborne pathogens such as Escherichia coli, Bacillus cereus, Staphylococcus aureus, and Salmonella typhimurium and mycotoxin-producing fungi (Guo et al. 2006; Kaboré et al. 2013; N’Dir et al. 1994). Kinema is a subtype of Natto that is also produced by spontaneous Bacillus fermentation. It has a sticky texture and ammoniacal odor (Sharma et al. 2015). B. subtilis was the predominant microbial isolate found during the Kinema fermentation and was believed to be attached to the bean surface. Besides, yeasts such as Candida parapsilosis and Geotrichum candidum, as well as bacteria, mainly Enterococcus fuecium, were detected as originating from wooden mortar and pestle that were used to prepare soybeans (Sarkar et al. 1994; Tamang 2003). The isolated Bacillus showed higher enzymatic activities including peptidase, phosphatase, lipase, and esterase compared with Enterococcus and the detected yeasts made the authors link the physicochemical changes that occurred in the fermented beans to Bacillus strains (Tamang 2003). Premature arrest of alkaline fermentation (stuck fermentation) might be one of the challenging problems in soybean fermentation. The presence of organic acids, either naturally occurring during the soaking period from LAB fermentation or artificially introduced as previously described, may prevent this Bacillus fermentation (Hosoi and Kiuchi 2008). To remove the organic acids and eliminate their negative impact on Natto production, flowing tap water is often applied after soaking. However, excessive flowing may affect the taste of Natto (Hosoi and Kiuchi 2008). The mechanisms of such effects are largely unexplored and require further research, maybe by either optimizing the manufacturing process or selecting acid-tolerant Bacillus isolates to conduct the fermentation.

Tempeh

In Tempeh, the most common molds found are Rhizopus oryzae, R. oligosporus, and Aspergillus oryzae, with a lower population of other species such as T. pullulans, A. niger, A. alternata, Mucor javanicus, and Cladosporium oxysporiumand (Ahnan‐Winarno et al. 2021; Romulo and Surya 2021). Bacteria and yeasts were also isolated in Tempeh with a population ranging from 6 to 9 logs CFU/g. For bacteria, high populations of B. licheniformis, Lactobacillus delbreuckii, and Limosilactobacillus mucosae, and low populations of Enterobacteriaceae species such as K. pneumoniae, C. freundii, and E. cloacae were detected. Yeasts including species of Trichosporon, Candida, Rhodotorula, Pichia, and Yarrowia were also reported in low populations (Kustyawati 2017; Kustyawati et al. 2018; Nurdini, Nuraida, and Suwanto 2015; Rizal et al. 2021; Romulo and Surya 2021; Suwanto 2021). During Tempeh fermentation, the fungal mycelium is attached to the soybean surface, and studies report varying penetration capability from only a few cell layers in depth to penetrating the bean cell wall and growing inside (Mulyowidarso 1991; Paredes-Lopez et al., 1987; Sudarmadji and Markakis 1978). The growth of these endogenous microorganisms brings about chemical changes within the fermented beans. Several enzymes were detected during tempeh fermentation, including lipases, proteases, cellulases, pectinases, and amylases (Ruiz‐Terán and David Owens 1996). Such enzymes were reported to degrade the crude lipid, triglycerides, proteins, and carbohydrates, producing free ammonia, free amino acids, polypeptides, free fatty acids, reducing sugars, and gradually raising the pH (Handoyo and Morita 2006; Mulyowidarso 1991; Sarkar and Nout 2014). Various compounds were detected in fermented beans that have desirable flavor and aroma such as pyridine, diacetyl, butanediol, and other higher alcohols and organic acids (Mulyowidarso 1991). Tempeh fermentation was also reported to decrease the level of trypsin inhibitors and phytic acid and increase the bean digestibility and the levels of riboflavin, and niacin concentration (Owens, Astuti, and Kuswanto 2015; Romulo and Surya 2021). These physiochemical changes occurred with the soybeans converting unpalatable beans into more palatable ones with distinct clean and characteristic aromas. Besides the reported roles of filamentous fungi, knowledge about the other microbial groups including bacteria and yeasts might be essential. These groups were frequently reported during the fermentation process with various populations and their roles were scarily studied. For example, Tempeh was associated with a variety of sensory characteristics, including bitterness, that was believed to be linked to bacterial and yeast activities rather than mold growth, illustrating their essential impact (Barus et al. 2008; Suwanto 2021). Future research employing advanced technologies like metagenomic, and proteomic analysis may deepen our understanding of microbial populations, including the non-mold species, to better control Tempeh fermentation and its sensory quality.

Soya sauce

Soya sauce is produced using a traditional fermentation process by mold spores (Koji), under a non-sterile environment which leads to the growth of indigenous microorganisms. Koji is produced by subculturing A. oryzae or Aspergillus sojae on cereals including steamed rice, wheat bran, or soybean flour, and incubated for 3 days in a tray or small boxes in a warm room (25–35 °C), with good aeration and low moisture to prevent the growth of undesirable microbial growth. The mature koji culture has a yellow to yellowish-green color coating the surface of the whole mass, which contains several enzymes including proteases, α-amylases, cellulases, and lipases. Leucine aminopeptidases and two endopeptidases (alkaline protease and neutral) were found with A. oryzae during this koji solid-state fermentation, whereas acid carboxypeptidases were also found with A. sojae (Ito and Matsuyama 2021; Kusumoto et al. 2021; Yamashita 2021). These enzymes were reported to hydrolyze the soybeans protein into polypeptides, peptides, and amino acids (Kusumoto et al. 2021; Yamashita 2021). Similarly, starch and cellulose are enzymatically broken down into disaccharides and monosaccharides by the extracellular α-amylase and cellulase respectively. Soybean lipid breakdown was not reported at this stage of fermentation and mainly occurred in the subsequent brine fermentation process (Nout and Aidoo 2011). Solid-state fermentation supports fungal growth and optimal enzyme production that would not be possible during submerged fermentation (Chancharoonpong, Hsieh, and Sheu 2012). By ensuring fungal growth, the growth of undesirable microorganisms is prevented (Adams and Moss 2008) and enzyme degradation processes can take place to provide substrates for microorganisms in the subsequent fermentation (Zhu, 2013). The resulting mass is submerged in brine (moromi) to produce Soya sauce (Allwood, Wakeling, and Bean 2021).

During the moromi submerged fermentation, lactic acid and yeast fermentation take place spontaneously for about one year, giving the final product its distinct sensory quality (Yong and Wood 1974). An early study showed that lactic acid fermentation occurs in the early stage, followed by yeast alcoholic fermentation (Devanthi and Gkatzionis 2019), while fungal growth was mostly inhibited by the high sodium chloride concentration in brine. Salt-tolerant yeasts were detected and identified as Pichia guilliermondii, Z. rouxii, Candida glabrata, and Pyrococcus furinosa (Song, Jeong, and Baik 2015). Characteristic yeast flavor metabolites were detected during the fermentation including alkyl phenols such as 2-phenyl ethanol, 4-ethylphenol, and 4-ethyl guaiacol, and high levels of glycerol (Song, Jeong, and Baik 2015). For bacteria, homofermentative, salt-tolerant, tetrad-forming cocci isolate known as Tetracoccus soyae and Pediococcus acidihctici var. soya have been identified, as well as low levels of Lactobacillus sp. (Lockwood and Smith 1950; Yong and Wood 1974). In addition, Bacillus species including B. amyloliquefaciens, B. licheniformis, B. pumilus, and B. subtilis were also detected during brine fermentation (Devanthi and Gkatzionis 2019; Matsumoto and Tomoyasu 1925; Yang et al. 2017; Yong and Wood 1974). More recently, species such as W. paramesenteroides, P. acidilactici, Pediococcus pentosaceus, T. halophilus, and B. amyloliquefaciens were detected at high abundance during the fermentation process (Sun, Li, and Sun 2009). Lactic fermentation is essential to improve the stability of brine fermentation, preventing undesirable flavors and adding desirable metabolites, mainly organic acids, improving the Soy sauce sensory quality (Harada et al. 2017, 2018; Sun, Li, and Sun 2009). Continuing the acidification process might cause a decline in the neutral and alkaline proteinase activities, leading to a reduction of protein solubility and elongating the process completion (Hoang et al. 2016). For non-LAB species, Bacillus has been reported, with high antifungal and high enzymatic activities - protease, fibrinolytic, and α-amylase which are believed to contribute to the Soya sauce flavor profile and improve its stability (Baek et al. 2011; Cho et al. 2009; Sun, Li, and Sun 2009; Yoo et al. 1999). These findings showed the ecological complexity of the two fermentation steps, koji, and moromi. Despite the fact that during mormori maturation the majority of soy sauce flavors are created, this process has received little research attention. To the best of our knowledge, there is still much to learn about the mechanism of flavor formation and how the microorganisms present in moromi fermentation function. Such knowledge is necessary to control the fermentation process, which could speed up the brine maturation process and enhance the quality of the finished product.

Miso

A consortium of both fungi and bacteria can also produce artisan fermented soybean products, which is the case of Miso solid-state fermentation. The soybeans are subjected to fungal fermentation for 48 h, then mixed with other ingredients including salt, and left to ferment for up to 24 months (Allwood, Wakeling, and Bean 2021; Onda, Yanagida, Uchimura, et al. 2003). Traditionally, mashed cooked beans were shaped into balls, and wrapped in rice straw which was modified later using rice koji that was used in Soya sauce production (Kusumoto and Rai 2017). Many varieties of fungal isolates were detected during Miso fermentation. However the leading molds were mainly A. oryzae, A. sojae, besides, A. tamarii, A. awamori, A. saitoi, and A. karwachi (Ashu et al. 2016; Sanjukta and Rai 2016). The enzymes produced during fungal fermentation include alkaline, acid, and neutral proteases, lipases, amylase, glutaminase, acid phosphatase, metallopeptidase, and phytase (Sourabh et al. 2015). These enzymes were linked to the hydrolysis of protein, lipid, starch, glutamine, phytic acid, and formation of free fatty acids, reducing sugars, as well as glutamic acid, that provide Miso with a characteristic umami flavor (Allwood, Wakeling, and Bean 2021). The addition of salt selects halophilic LAB species such as Tetragenococcus halophilus which will predominate the fermentation and the subsequent ripening processes (Allwood, Wakeling, and Bean 2021; Onda, Yanagida, Uchimura, et al. 2003). These bacteria were reported to produce organic acids causing color brightening, enhanced product stability, and improved yeast growth. In contrast, other LAB species including Pediococcus acidilactici, L. plantarum, and Fructilactobacillus fructivorans were also detected and believed to have a negative impact on the product quality, namely over-acidification, swelling, and inhibition of other desirable microbial growth including that of yeasts and filamentous fungi (Allwood, Wakeling, and Bean 2021; Ohnishi 1994; Onda, Yanagida, Tsuji, et al. 2003; Onda, Yanagida, Uchimura, et al. 2003). Yeasts such as Wickerhamomyces anomalus, Zygosaccharomyces rouxii, Debaryomyces nepalensis, Pichia membranifaciens, and Candida prapsilosis were previously detected in Miso (Allwood, Wakeling, and Bean 2021; Kusumoto et al. 2021). These yeasts produce desirable metabolites including ethyl alcohol, higher alcohols, and succinic acid linked to the desirable aroma of Miso, while other yeasts and bacteria may produce a thick film coating the product surface and affecting its odor. These findings suggest that the typical Miso is produced by the metabolic activities of various microbial activities. It is challenging to identify the usual Miso microbial profile and identify which microorganisms are responsible for the related texture and flavor changes because of the presence of a variety of microbial groups. Comparative research employing specific inhibitors might shed more light on the microorganisms that make up a typical Miso. For example, the fermentation processes of coffee and cocoa revealed a rich microbial ecology of several fungi and bacteria species (De Vuyst and Weckx 2016; Elhalis, Cox, and Zhao 2023; Evangelista et al. 2014). Understanding the significance of each microbial group and its functions required a comparative study. The authors used selective agents such as Natamycin and Nisin to selectively suppress the fungal and bacteria, respectively, and compared the resulting beans to a spontaneous fermentation process (Elhalis, Cox, Frank, et al. 2020; Ho, Zhao, and Fleet 2014; Ho, Fleet, and Zhao 2018). More specifically, to illustrate the role of yeasts in coffee and cocoa fermentation, researchers suppressed yeast growth with Natamycin and then compared the resulting fermenting beans in the absence of endogenous yeasts to a spontaneous fermentation where yeasts were presented. The authors concluded that yeasts are essential to produce high-quality beans and are considered a typical microflora during both fermentation processes (Elhalis, Cox, Frank, et al. 2020; Ho, Zhao, and Fleet 2014). In contrast, a related study that used Nisin and lysozyme to restrict LAB growth during cocoa fermentation showed the unnecessary involvement of this microbial group to produce high-quality beans (Ho, Zhao, and Fleet 2015). Similar approaches might be followed to advance our understanding of the impact of different microbial groups on producing high-quality Miso.

Sufu

Sufu is produced from soymilk curd which is subjected firstly to fungal solid-state fermentation (soymilk curd overgrown with a mold known as pehtze). A study conducted by Han et al. (2004) to monitor the endogenous microflora during this solid-state fermentation reported a wide range of microorganisms including LAB, spore-forming bacteria, and yeasts, in addition to filamentous fungi. Aerobic bacteria and spore-forming bacteria had an initial presence of about 4 log CFU/g which grew to a maximum population of 8–9 log CFU/g at the end. LAB presented with a similar initial count and grew steadily during the solid-state fermentation to reach a maximum count of log 7 CFU/g. Fungi had an initial count of less than 3 log CFU/g and increased to about 7 log CFU/g by the end. The inclusion of a salting process reported a significant reduction of the total population of aerobic bacteria, LAB, fungi, and a lower effect on the level of spore-forming Bacillus, similar to that reported with Miso as discussed above. The typical mold fermentation of Sufu is predominated by A. elegans, Mucor wutunkiao, Mucor racemosus, Rhizopus microsporus, R. oryzae, A. oryzae, G. candidum, and Trichosporon asahii (Moy, Lu, and Chou 2012; Wan et al. 2020; Xie et al. 2018). Several enzymatic activities including protease, α-amylase, lipase, α-galactosidase, esterase, and l-glutaminase were detected during the solid-state fermentation and associated with significant physiochemical changes in the fermented curd. These enzymes were reported to degrade the raw materials through a series of biochemical processes producing various fatty acids, amino acids, alcohols, organic acids, aldehydes, and esters, that impart a desirable impact on the product aroma and texture quality (He, Chen, and Chung 2022; Liu, Song, and Luo 2018; Xu et al. 2020). For example, Liu, Chen, et al. (2022) detected basic proteases, neutral proteases, and amino acid decarboxylase during Sufu fermentation which was believed to increase the levels of free amino acids, peptides, and free ammonia. Similarly, different molecular mass proteases (73.1, 76.4, and 109.1 kDa) were detected during Sufu fermentation and were reported to degrade the curd protein into ammonia, amines, amino acids, and peptides (Han et al. 2003). The authors believed such enzymes belong to different microbes but not molds, i.e., A. elegans, which was commonly reported with proteases with molecular mass ranging from 30–40 kDa. By the end of the fermentation, high levels of long-chain fatty acid ethyl esters such as ethyl dodecanoate, ethyl octadecenoate, and ethyl octadecadienoate were detected. These are likely formed by the reaction between metabolic products such as ethanol and acids by fungal lipases. Similarly, sulfur-containing compounds such as methional increased with the extended maturation which might be formed by the indigenous microbial populations, amino acids enzymatic degradation, or Maillard reaction (He, Chen, and Chung 2022; Moy, Lu, and Chou 2012; Ouoba et al. 2003).

The mold-fermented soymilk curd is subjected to submerged fermentation in different dressing mixtures with various salt concentrations. The microbial diversity was potentially affected by the climate, locations, and raw materials. During fermentation, the level of dominant endogenous aerobic and spore-forming bacteria present at the earlier stage of solid-state fermentation declines, and significant reduction was observed with higher salt content dressings (Han, Rombouts, and Nout 2001, Han et al. 2004). The dominant genera found were Leuconostoc citreum, Weissella paramesenteroides, Lactococcus lactis, E. faecium, B. anthracis, B. cereus, and others belong to genera of Lactobacillus, Tetragenococcus, Acinetobacter, Pseudomonas, and Chryseobacterium (Fei et al. 2018; Wan et al. 2020; Zhen-Dong et al. 2021). The fungal communities that were presented at the early stage declined as the fermentation progressed. Yeasts such as Pichia cactophila, Kodamaea ohmerivoucher, Magnusiomyces capitatus, and Candida etchellsii were also detected initially and survived during the fermentation (He and Chung 2020; Wan et al. 2020; Wu et al. 2018). The endogenous species such as Tetragenococcus, Lactobacillus, Candida, Debaryomyces, and Geotrichum, Actinomucor were correlated to various desirable compounds such as free amino acids, fatty acids, esters, higher alcohols, and aldehydes that are known for their fruity, sweet, and nutty aroma, enhancing the final product quality (He and Chung 2020; Wan et al. 2020; Wang, Sun, et al. 2019). Most recently, Candida tropicalis and Rhodotorula sp. were reported to produce nonanoic acid that added milk flavor notes to the final products, most likely through their lipolytic activities (Kim et al. 2019; Li et al. 2022). On the contrary, endogenous isolates such as Bacillus, Weissella, and Alternaria were believed to negatively contribute to the final product quality (He and Chung 2020). Red Sufu is a variant of Sufu prepared using comparable methods, but the final maturation period is dominated by Monascus fermentation (Wang, Sun, et al. 2019). Monascus species including M. purpureus, M. anka, and M. ruber have been used to ferment substrates, mainly rice, and produce various secondary metabolites including pigments such as Ankaflavin and Monascin (yellow); Monascorubrin and Rubropunctatin (orange); Monascorubramine and Rubropunctamine (red) (Srianta et al. 2016). During red Sufu maturation, various aroma compounds were produced and esters and alcohols were abundant volatiles (Chung 2000; Wang, Sun, et al. 2019). Esters such as ethyl acetate, ethyl 2-methyl butanoate, ethyl 2-methyl propanoate, ethyl butanoate, and ethyl 3-phenylpropionate, are associated with adding fruity notes to the final products. The presence of these esters compounds was linked to yeast metabolic activities of lipids which produce free fatty acids that result in the synthesis of corresponding esters (Gao et al. 2010; Wang, Sun, et al. 2019). Alcohols, such as ethanol, phenylethyl alcohol, 1-octen-3-ol, 2-methylbutanol, hexanol, and 3-methylbutanol give flowery and sweet aroma to the final products (Devanthi et al. 2018; Zhang et al. 2019). Besides, other components including 2,3-butanedione, 3-(methylthio) propanal, and benzene acetaldehyde, were detected and believed to enhance the red Sufu characteristic aroma profile (Gao et al. 2010; Wang, Sun, et al. 2019). Overall, identifying the crucial flavor-producing microbial isolates among the endogenous microbiota of Sufu fermentation is still challenging because of the microbial community diversity, and their low culturability, as well as the long maturation duration (He, Chen, and Chung 2022; Huang et al. 2018; Li et al. 2022). Therefore, the intimate knowledge of the interactions between the detected microorganisms, their synergistic or antagonistic, fermentation environments, and product quality using the current advanced technology will help improve the comprehension of Sufu fermentation and the ability to control its production.

Other products

Apart from these known fermented soybean products, other forms have also been subjected to fermentation to improve their nutrition and sensory quality. For instance, soybean paste (dajiang) was subjected to a spontaneous fermentation process to produce a high-value fermented product under a solid state (An et al. 2021; Wu et al. 2013). The microbiota present during the fermentation showed a high abundance of bacterial species, including T. halophilus, Enterococcus faecium, Leuconostoc mesenteroides, Leuconostoc gasicomitatum, Paenibacillus glycanilyticus, Staphylococcus epidermidis, and Bacillus firmus, which stayed predominant during the fermentation except for S. epidermidis. For fungi, several species including A. oryzae, A. elegans, Mucor wutongqiao, M. racemosus, R. oryzae, R. oligosporus, Zygosaccharomyces rouxii occurred in the early stage with a high population but mostly declined with the fermentation progression (An et al. 2021; Liu, Chen, et al. 2022; Wu et al. 2013; Yue et al. 2021). The correlation between the high abundance of the indigenous isolates including Tetragenococcus, Lactobacillus, and unclassified Enterocococcaceae, and the high levels of acetic acid, benzaldehyde, and esters linked to the desirable aroma and taste of soybean paste was observed by (An et al. 2021; Wu et al. 2013). Similarly, Soybean meal has been subjected to spontaneous solid-state fermentation processes to improve its sensory and nutritional quality. The studies showed reported a high abundance of Pseudomonas species during the early stage of fermentation which was replaced by Bacillus species including B. cereus, B. subtilis, and B. amyloliquefacien as fermentation progressed (Medeiros et al. 2018; Wang, Zhang, et al. 2021). Various enzymes were detected, including α-amylase, protease, and cellulase, which correlated to the increase in water content and water-soluble protein associated with the Bacillus growth. Relevant research also illustrated that the soybean meal supports the growth of R. oryzae and Ceratocystis fimbriata, producing several metabolites including acids, alcohols, ketones, esters, and aldehydes with strong fruity aroma (Bramorski, Christen, et al. 1998; Bramorski, Soccol, et al. 1998).

Comparison of microbial roles in spontaneously fermented soybean products

Based on the aforementioned discussions, similar microorganisms have been detected during multiple spontaneous fermentations of soybean products with a variety of functions. Table 1 shows selected microorganisms and the reported main impact on different fermented soybean products. Three different fermentation types occur in soybeans. The first is alkaline fermentation, which is dominated by a wide range of Bacillus species including B. subtilis, B. amyloliquefaciens, B. licheniformis, B. pumilus, and B. velezensis. These species share hydrolytic capabilities, mainly their proteases, and polyglutamic acid production, in addition to their strain-specific properties (Sarkar, Cook, and Owens 1993; Shrestha, Dahal, and Ndungutse 2013). The resulting fermented products, such as Natto, Kinema, and Douchi, all mainly have a sticky appearance and powerful pungent odor. These findings might help to comprehend the typical roles of Bacillus species that might occur in soybean fermentation, considering the other changes that are highlighted in the earlier discussion for each Bacillus strain. The second type is fungal fermentation, which is dominated by filamentous fungi including Aspergillus, Rhizopus, and Mucor species. Examples of this type include Tempeh, Miso, Sufu, and Soya sauce. In Tempeh and Miso, the fungal growth causes the beans to fuze together by the fungal mycelium and its extracted enzymes that facilitate the mycelium attachment to the bean surface give these products their typical appearance as mycelial biomass with a softening texture (Wikandari et al. 2021). Additionally, several fungal enzymes were detected, including proteolytic, lipolytic, carbohydrate hydrolytic enzymes, and esterase, all of which play similar roles in increasing the bean digestibility and creating volatiles, mainly belonging to aldehydes and alcohols (Ghosh and Ray 2011; Mei Feng, Ostenfeld Larsen, and Schnürer 2007). Miso, Sufu, and Soya sauce are subsequently subjected to the third type of fermentation (maturation), which continues as either brine fermentation (submerged) in Sufu and Soya Sauce or continues as salt solid-state fermentation in the former product. Both types are dominant with LAB such as T. halophilus and yeasts such as Z. rouxiiin, which exhibit common osmophilic properties (Lee, Lee, and D. K. B. T.-A. M. 2002). During the maturation process, desirable volatiles, mainly esters, alcohols, and organic acids, are generated that give the final product its distinct fruity aroma profile, as well as several enzymes that continue the degradation process of lipid, fiber, and protein and soften their texture (Nout, Sarkar, and Beuchat 2007). Other subtypes of fermented soybeans, such as Manascus, which produces the characteristic red-colored Sufu, are included in the third fermentation (Wang, Sun, et al. 2019). These findings illustrate the crucial roles of endogenous microorganisms in soybean fermentation that can be manipulated to better control the fermentation process and produce high-quality products, as well as create novel fermented foods.

Table 1. Comparison of microbial roles in spontaneously fermented soybean products.

Dominant microbial species found in fermented soy products	Fermentation types	Fermented soybean products	Main functions	References
Bacillus subtilis var. natto B. amyloliquefaciens, B. licheniformis B. pumilus B. velezensis	Alkaline fermentation	Natto Kinema Douchi	Produced polyglutamic acid, which was the major component of the viscous sticky mucilage that was found in Natto Produced distinct flavor compounds associated with pungent that gave Natto its characteristic aroma Exhibited amylolytic, proteolytic, and lipolytic activity that could break down carbohydrates, proteins, and lipids into smaller products	(Shin and Jeong 2015; Shrestha, Dahal, and Ndungutse 2013)
Mucor spp. Actinomucor spp.	Fungal fermentation	Sufu Tempeh	High protease production broke down proteins into smaller peptides and amino acids Formed a product with a soft cheese-like texture with a creamy consistency (Sufu)	(Wei, Regenstein, and Zhou 2021)
Aspergillus spp.	Fungal fermentation	Soya sauce Miso Sufu Tempeh	Produced enzymes such as cellulase, amylase, and proteases Produced glutaminase that converted glutamine to glutamic acid to provide a distinct umami flavor Produced 4-ethyl guaiacol (characteristic soy sauce aroma)	(Kusumoto et al. 2021; Yamashita 2021).
Rhizopus spp.	Fungal fermentation	Tempeh Sufu	Formed mycelia to fuze fermented soybeans together to form a ‘cake-like’ product Produced carbohydrate degradation enzymes, lipases, and proteases	(Handoyo and Morita 2006; Mulyowidarso 1991; Sarkar and Nout 2014).
Tetragenococcus halophilus	Third fermentation (Maturation)	Soy sauce Miso Sufu	Salt tolerant bacteria Produced lactic acid to reduce pH	(Allwood, Wakeling, and Bean 2021)
Zygosaccharomyces rouxiiin Millerozyma spp. Candida spp.	Third fermentation (Maturation)	Douchi Soy sauce Sufu Miso	Produced flavor compounds such as (3-octanol, 2-methoxy-phenol, 4-ethyl-phenol, and phenol) Produced esters that enhanced sweet and caramel-like aroma	Wang et al. 2022; Zhao et al. 2023)
Monascus purpureus	Third fermentation (Maturation)	Red Sufu	Produced red pigments that gave Sufu its characteristic reddish appearance	(Xie et al. 2018)

Starter culture technology for soybeans fermentation

Traditional spontaneous fermentations have long been used to produce fermented beverages and foods where the inoculum needed to kickstart the fermentation process consists of either the microorganisms found naturally in the product or the addition of a small amount of old fermented sample to a new batch (Caballero, Trugo, and Finglas 2003). Although these processes produce oriental and traditional sensory products, certain drawbacks such as contamination with undesirable microorganisms, and slow or failed fermentation can result in inconsistent fermentation products, and hence are not suitable for large-scale production facilities (Caballero, Trugo, and Finglas 2003). Nowadays with modern technology, starter cultures consisting of identified microorganisms are inoculated into food materials to produce fermented products, aiming to improve product stability, nutritional and sensory quality, consistency, and process economics (Hansen 2014). In soybean fermentation, several microorganisms including bacteria, yeasts, and filamentous fungi have been used as starter cultures and their roles have been studied.

Bacillus subtilis starter cultures

Soybeans fermented by B. subtilis inoculums produce Natto, as known in Japan, and similar products with different names in other countries, such as Douchi in China, Chung-kook-jong in Korea, and Kinema in Thailand. The fermented product is characterized by a slimy white appearance, soft texture, and powerful odor. Several conditions reported to affect Bacillus activity include inoculum size, water activity, salt concentration, pH, and temperature. Low inoculum size causes a delay in growth while higher levels cause exhaustion of the nutrients before bacterial enzyme production starts (Uyar and Baysal 2004). During fermentation, B. subtilis converts carbohydrates into organic acids causing a decline in pH at the beginning. Subsequently, soybean proteins are hydrolyzed by B. subtilis and this produces ammonia and amines that gradually increase pH to 8.34 in 48 h (Sarkar, Cook, and Owens 1993; Sarkar and Tamang 1995). The moisture content of the fermented beans is also increased through the Bacillus fermentation. The decline of soybean protein concentration is associated with the accumulation of microbial proteins, ending with a slight increase in the overall protein concentration at the end of the fermentation (Hu et al. 2010; Shrestha and Noomhorm 2001). Similarly, lipids and carbohydrates are hydrolyzed during fermentation, leading to degradation products such as free fatty acids and sugars (Sarkar, Cook, and Owens 1993; Sarkar and Tamang 1995). Further physiochemical changes also potentially impact the texture and sensory quality of the final products (Shrestha, Dahal, and Ndungutse 2013). Such changes are associated with improved bean digestibility, removal of beany odor, and reduction of indigestible oligosaccharides (Sarkar, Cook, and Owens 1993). Enzymatic activities have also been reported to produce several volatiles, including ketone, aldehydes, esters, and sulfur compounds that contribute significantly to the soybean sensory quality (Owens et al. 1997). The main volatiles found in fermented Natto were butanone, 2-methyl butanoic acid, trimethyl pyrazines, 2-pentrylfuran, 3-hydroxy butanone, and dimethyl disulfide that were believed to improve Natto sensory quality (Leejeerajumnean et al. 2001; Liu, Song, and Luo 2018). Furthermore, B. subtilis increases the total phenolic compounds, anthocyanins, and isoflavones, known for their beneficial health impact linked to glucosidase activities (Hu et al. 2010). Kinema was produced by subjecting soybeans to B. subtilis starter culture under different inoculum sizes, from about 2 to 7 log CFU/g, and monitoring under different incubation temperatures (18-52 °C) aiming to better control the fermentation process and improve the final product quality (Sharma et al. 2015). The authors found that optimum conditions were 3.0 log CFU/g, and fermentation at 37 °C for 48 h that produced a high-quality product with a nutty flavor, mild ammonia, and a highly sticky texture. It was also associated with a significant reduction in the concentrations of anti-nutritive factors including phytic acid, trypsin inhibitor activity, tannic, and total biogenic amines.

Other Bacillus species are also used to produce fermented soybeans, such as B. amyloliquefaciens which was isolated from a naturally fermented soybean process. Studies have shown that its high proteolytic and β-glucosidases activities decrease fermentation time, and lead to an increase in the total phenolic, flavonoid, and antioxidant levels associated with a reduction in the anti-nutritive factors (Cho et al. 2003; Jiang et al. 2019; Wang, Zhang, et al. 2021; Yang et al. 2019). Bacillus polyfermenticus was also isolated from Meju (a Korean soybean fermentation starter). It had high fibrinolytic and proteolytic enzymatic activities and higher poly-glutamate production, and the resultant fermented beans showed a different texture of lower viscosity than that produced with the traditional B. subtilis (Mo et al. 2010). Similarly, Bacillus velezensis was isolated from fermented soybeans ‘Douchi’ and was shown to eliminate glycinin and part of β- conglycinin (allergenic proteins) in soybean and hydrolysis soybean protein which significantly improves digestibility of the beans (Liu, Chen, et al. 2021). Furthermore, the authors reported their broad-spectrum antimicrobial effects on bacteria, yeast, and filamentous fungi which may suppress contaminant growth when fermentation is conducted under non-sterile conditions. Table 2 shows the main Bacillus species and their characteristics and contributions to soybean fermentation.

Table 2. Bacterial and yeast starter cultures diversity in soybean fermentation.

Starter cultures	Characteristics of the selected strains	Contribution to the fermentation process and soybeans quality	References
Bacillus subtilis	High resistance High enzymatic activities Anti-microbial	Desirable metabolites: butanone, 2-methylbutanoic acid, trimethyl pyrazines, 2-pentrylfuran, 3-hydroxybutanone, and dimethyl disulfide Increased phenolics Improved the digestibility, Removed the beany odor and indigestible oligosaccharides	(Leejeerajumnean et al. 2001; Liu, Song, and Luo 2018; Owens et al. 1997; Sarkar and Tamang 1995; Shrestha, Dahal, and Ndungutse 2013; Shrestha and Noomhorm 2001)
Bacillus polyfermenticus	High fibrinolytic High proteolytic High polyglutamate production	Low viscosity Functional food	(Mo et al. 2010)
Bacillus amyloliquefaciens	High proteolytic B- glucosidase	Accelerated the fermentation Increased phenolics Reduced the anti-nutritive	(Cho et al. 2003; Jiang et al. 2019; Wang, Zhang, et al. 2021; Yang et al. 2019)
Bacillus velezensis		Decreased allergenicity Improved digestibility Showed Antimicrobial effects	(Liu, Chen, et al. 2021)
Lactobacillus rhamnosus Tetragenococcus halophilus Lactiplantibacillus plantarum Levilactobacillus brevis Lactcaseibacillus paracasei Lactobacillus casei Liquorilactobacillus mali Bifidobacterium breve	Produce organic acids and antimicrobial metabolites to inhibit food contamination and foodborne pathogens Add desirable flavors Halophilic pH tolerance	Improved the product sensory quality Increased its antioxidant capacity Improved the product texture Increased the concentration of amino acids Reduced the concentration of biogenic amines Enhanced product stability, Desirable volatile compounds Encouraged yeasts growth	(Chen, Yanagida, and Hsu 2006; Holzapfel et al. 2006; Liu et al. 2012; Marazza et al. 2012; Nguyen et al. 2007; Ouoba et al. 2005; Sarkar and Nout 2014; Sarkar and Tamang 1994; Shimakawa et al. 2003; Sirilun et al. 2017; Yoon, Kim, and Hwang 2008)
Saccharomyces Klyveromyces Debaryomyces Hansenula Meyerozyma guilliermondii Rhodoturola Zygosaccharomyces	Organic acids tolerance Osmotolerance pH tolerance Production of various desirable metabolites Produce proteases, galactosidases, invertases, and inulinases	Increased protein contents Improved the antioxidant activities Reduced the crude fiber level	(Arneborg et al., 2000; Deak, 2007; Lee et al. 1997; Rashad et al. 2011; Sarkar and Tamang 1994; Song, Jeong, and Baik 2015; Sørensen et al. 2011)
Zygosaccharomyces	Grow at high salt content Produce desirable metabolites	Improved the final product sensory quality	(Kwon and Kim 2005)
Zygosaccharomyces rouxii M. guilliermondii	–	Produced desirable metabolites (furanone, and higher alcohols) Added sweet and caramel flavor notes	(Hayashida, Nishimura, and Slaughter 1997; Kim et al. 2011; Wah et al. 2013)
Kluyveromyces marxianus	–	Increased crude fat, soluble dietary fibers, and polysaccharides Reduced anti-nutritive factors Reduced beany odor	(Hu et al. 2019)
Lindnera saturnus	Metabolize aldehydes such as 2-heptenal and hexanal that are responsible for green beany flavor and bioconvert them into corresponding ester compounds ethyl hexanoate and ethyl heptanoate	Improved the bean quality Added fruity and sweet aroma	(Vong and Liu 2018)
Candida parapsilosis Geotrichum candidum	-	Created a low-quality final product with an undesirable rancid smell	(Sarkar and Tamang 1994)

LAB starter cultures

The most common products using LAB starter culture in soybean submerged fermentation are soy milk and fermented soya yogurt. L. plantarum is used to ferment soymilk and it improves the product texture through the production of exopolysaccharides. Similarly, Levilactobacillus brevis and L. plantarum inoculated into soybean milk produce a soy yogurt with significantly high concentrations of amino acids, aglycones, and high antioxidant capabilities (Hwang et al. 2021). Other LAB species such as Bifidobacterium breve, Lactobacillus rhamnosus, Lacticaseibacillus paracasei, Lc. casei, and Liquorilactobacillus mali have been reported to add pleasant flavors, aroma, and various bioactive compounds which improve human health (Marazza et al. 2012; Shimakawa et al. 2003; Sirilun et al. 2017). In Soya sauce, T. halophilus is inoculated into the submerged fermentation process (moromi, 12% NaCl), and its contribution to the finished product quality has been reported (Singracha et al. 2017). The microorganism is capable of reducing the level of biogenic amines and increasing desirable metabolite levels including acids, aldehydes, alcohols, ester, and phenol which improves the overall product safety and quality (Table 2). In detail, alcohols such as phenyl ethyl alcohol, 3-methyl-1-butanol, ethanol, and 2-methyl-1-butanol are among the highest alcohols (Zhao et al. 2018). For esters, ethyl acetate was the most common ester found, while 2-methyl butanal were aldehydes and ketones. These substances are correlated to fruity, sweet pleasing aroma characteristics. For phenols, guaiacol and 4-ethyl guaiacol were the most common flavor compounds detected that could contribute to the smoky aroma of Soya sauce (Zhao et al. 2018).

For solid-state fermentation, soybean paste has also been inoculated with T. halophilus, showing higher concentrations of lactic acid and acetic acid that enhance product stability, color, and prevention of rancidity (Cui et al. 2012). Various volatile compounds were also detected with higher concentrations compared to the spontaneously fermented paste, such as acetate, isoamyl alcohol, ethanol, and 4-ethyl phenol which improved the final product sensory quality and masked the undesirable soybean flavor. The authors also stated that the lactic acid produced during LAB activity created an optimal environment for yeast growth, performing a synergistic action between both microbial groups and enhancing the production of secondary desirable metabolites (Cui et al. 2012). Other LAB species such as L. plantarum, P. acidilactici, and Staphylococcus carnosus were applied to soybean paste and Miso fermentation processes and showed potential improvement in the fermented soybean product safety by reducing the concentration of biogenic amines (Lee et al. 2016; Nguyen et al. 2007; Zhao et al. 2020). In addition, higher levels of esters (isopentyl acetate), alcohols (phenyl ethyl alcohol and 2,3-butanediol), aldehydes (benzene acetaldehyde), aspartic acid, and glutamic acid were found in the inoculated beans compared to the non-treated beans, providing the fermented beans with fruity, nutty, sweet, and umami sensory quality (Zhao et al. 2020).

Yeast starter cultures

Yeasts are not the main fermenting microbial group in soybean fermentation, and their roles are not well-illustrated and sometimes controversial. Yeasts such as Saccharomyces, Klyveromyces, Debaryomyces, Hansenula, Candida, Pichia, Rhodoturola, and Zygosaccharomyces have been isolated during meju fermentation. They were reported to have high enzymatic activities such as protease, galactosidase, invertase, and inulinase that led to increased protein content, improved antioxidant activities, and reduced crude fiber levels in the resultant fermented beans (Lee et al. 1997; Rashad et al. 2011). On the other hand, kinema (a Nepalese fermented soybean product) produced by yeast fermentation such as C. parapsilosis and G. candidum had low quality and were associated with an undesirable rancid smell (Sarkar and Tamang 1994).

Overall, despite yeasts not being the main fermenting microorganism, recent attempts to inoculate yeasts in soybeans have still been reported, which shows their potential capability of improving the final product quality (Table 2). The organoleptic properties of Soya sauce were found to be improved using Z. rouxii individually or in co-culture with Meyerozyma guilliermondii and Candida variabilis after earlier LAB fermentation, producing desirable metabolites such as furanone, and higher alcohols with sweet and caramel flavor notes (Hayashida, Nishimura, and Slaughter 1997; Kim et al. 2011; Wah et al. 2013). A similar finding was also recorded using Z. rouxii individually or in co-culture with T. halophilus which showed potential flavor improvement during the Soya sauce maturation (Harada et al. 2017; Kwon and Kim 2005). Related studies showed Z. rouxii was found to produce a wide range of volatiles Among them, 4-hydroxy-5-methyl-3(3 H)-furanone was considered the crucial compounds that give Soya sauce and Miso their characteristic aroma (Ebine 2004; Hanya and Nakadai 2003). Similarly, the isolated C. versatilis was shown to produce various metabolites where higher alcohols, including isobutyl alcohol, isoamyl alcohol, and 2-phenyl ethanol, were essential aroma compounds of Soya sauce, but not with Miso (Aidoo, Nout, and Sarkar 2006; Ebine 2004).

For solid-state fermentation, Kluyveromyces marxianus was inoculated into Okara (the insoluble residues that are produced during soy milk and tofu manufacturing) and the fermented product showed higher levels of crude fat, soluble dietary fibers, and polysaccharides and a reduction in anti-nutritive factors such as phytic acid, trypsin inhibitors, and beany odor (Hu et al. 2019). A related study using Lindnera saturnus confirmed the hypothesis of metabolizing aldehydes such as 2-heptenal and hexanal which are responsible for the soybean green beany flavor, by bioconversion into the corresponding ester compounds, including ethyl hexanoate and ethyl heptanoate, which possess fruity and sweet aroma (Vong and Liu 2018).

Filamentous fungi starter culture

Among the filamentous fungi isolated during spontaneous soybeans fermentation, three main species belonging to Mucor, Aspergillus, and Rhizopus are commonly used in controlled solid-state fermentation. Table 3 shows the most common filamentous fungi and their roles in soybean fermentation. Overall, the contribution of the filamentous fungi in soybean is mostly observed during solid-state fermentation type and if submerged fermentation is subsequently included their growth is potentially declined, however, the excreted enzymes during the former fermentation continue their functional activities under submerged fermentation (Liu, Chen, et al. 2021).

Table 3. Filamentous fungi starter cultures diversity in soybean fermentation.

Filamentous fungi	Most characteristics of the selected strains	Contribution to the fermentation process and soybeans quality	References
Mucor racemosus Mucor Sufu Actinomucor elegans Actinomucor taiwanensis	Narrow temperature tolerance between 20-30^oC	Increased amino acids concentration Degraded the fiber content Masked the fishy smell Had smooth mouthfeel and sensory quality	(Cheng et al. 2009; Krishna 2005; Kronenberg and Hang 1984; Xu, Liu, and Zhou 2012)
Mucor flavus	Grow at low temperature (15^oC) Potential proteolytic enzyme activities	Increased amino acid levels in the final product	(Cheng et al. 2009)
Aspergillus oryzae Aspergillus sajae	Saccharification capabilities High proteolytic	Improved color aroma and sensory quality to the fermented beans Neither of them produced mycotoxins Produced various flavors and aroma compounds Produced high abundant glutamic acid (added umami taste)	(Kitamoto 2002; Ohata et al. 2007)
Rhizopus oligosporus Rhizopus oryzae	Growth at high temperatures makes it suitable for summer-season production High proteolytic and amylolytic activities Antimicrobial activities Desire volatiles producer	Improved the product regarding alkalinization, taste, and texture Increased the level of vitamins	(Baumann and Bisping 1995; Sparringa and Owens 1999)
Penicillium glabrum	α-ketoglutaric acid-derived amino acid synthesis	Modulated the nutrient profile compared to A. oryzae fermented beans	(Sun et al. 2019)

Mucor species such as M. racemosus, M. Sufu, and A. elegans are the most common and characterized by narrow temperature tolerance between 20–30 °C and a slow growth rate out of this temperature range that limited their application during summer (Li et al. 2008; Yamasaki and Konno 1991). The inoculated beans were first covered with small white colonies and were later fully covered with white to grey mycelium in 2–3 days. A. elegans is used to ferment Okara and the resultant fermented products showed alkaline pH, higher levels of amino acids, more degraded fiber content, and masking the fishy smell of soybeans with a smooth mouthfeel that improves their sensory quality (Krishna 2005; Kronenberg and Hang 1984; Xu, Liu, and Zhou 2012). Sufu is mainly produced by using A. elegans and A. taiwanensis during the solid-state fermentation step (Pehtze) at a low temperature of 15–20 °C (Cheng et al. 2009). This low temperature is more favorable to Mucor spp. and not for bacteria, yeasts, and other filamentous fungi, and is characterized by high Mucor enzymatic activities of proteases, lipases, amylases, and galactosidases (Liu, Chen, et al. 2021). The fungal enzymes produced during the solid-state fermentation were responsible for the continuous degradation of the bean ingredients during subsequent processes liberating degraded compounds including free amino acids, fatty acids, and sugars in addition to secondary microbial metabolites. These physiochemical changes are crucial to producing Sufu with traditional organoleptic quality (Han, Rombouts, and Nout 2001; Liu, Chen, et al. 2021). Out of this temperature range, fermentation takes a long time or stuck making it difficult to produce fermented soybean products using these Mucor strains. Screening for alternative strains that have the capability to grow out of this temperature range received much attention recently. Mucor flavus spores were spread onto fresh tofu surfaces to produce Sufu at a low temperature (15 °C) and showed a significant growth rate with potential proteolytic enzyme activities that increased amino acid levels in the final product (Cheng et al. 2009). For high temperatures during summer, Rhizopus spp., such as R. arrhizus, R. oligosporus, and R. chinensis perform with better growth making them suitable for the summer season production (Cui, Li, and Liu 2016; Ruiz‐Terán and David Owens 1996).

R. oligosporus and R. oryzae were isolated and evaluated to be used as a starter culture to produce Tempeh (Baumann and Bisping 1995). The two strains showed high proteolytic activities releasing a significant amount of amino acids that are crucial for Tempeh quality regarding alkalinization, taste, and sliceability (Baumann and Bisping 1995; Sparringa and Owens 1999). R. oligosporus was also reported to increase the level of vitamins during Tempe production including riboflavin, nicotinamide nicotinic acid, and vitamin B6, but not B12 (Keuth and Bisping 1993). Furthermore, they reported with high capability to degrade most of the soybean polysaccharides such as raffinose, stachyose, sucrose, and melibiose reducing their negative nutritive impact on soybean consumption (Handoyo and Morita 2006; Rehms and Barz 1995). Related studies illustrated the antimicrobial activities of the resultant fermented soybeans of R. oligosporus (Kusumah et al. 2020; McCue et al. 2004). The authors believed fatty acids such as linoleic acid and α-linolenic acid 1-monolinolenin and 2-monolinolenin exhibited antimicrobial activity against gram-positive food pathogens including S. aureus and B. subtilis.

For Aspergillus, Koji is a common mold starter culture that mainly contains Aspergillus species such as A. oryzae, A. kawachii, A. sojae, A. awamori, and A. shirousamii used to produce Soya sauce, Miso, and other fermented soybean products (Kitamoto 2002; Matsushita et al. 2009). A. oryzae or A. sajae are the most popular and important among them being potential saccharification, and diastatic agents and positively contributing to the color, aroma, and sensory quality of the fermented beans, and neither of them produces mycotoxins (Kitamoto 2002). Generally, the soybeans with wheat flour are mixed with 0.05–0.3% w/w of the fungal spores and loaded into trays with layers depth of 3–5 cm (Devanthi and Gkatzionis 2019; Yan et al. 2013). During the fermentation, the inoculated mold showed high enzymatic hydrolysis of soybean components including proteins into small peptides and amino acids by proteinases, and starch degraded into simple sugars by α-amylases. These enzymatic hydrolyzes decrease the acidity from pH 6.5 to 7.3 and increase the internal temperature, in addition, the fungal mycelia grow and sporulate on the soybean surface producing a greenish mass within three days (Devanthi and Gkatzionis 2019; Luh 1995). The resultant fermented beans were reported with several desirable flavors and aroma compounds including esters and higher alcohols similar to that reported with fungal fermented soybean products mentioned above. Moreover, glutamic acid, which is a key compound produced by glutaminase during koji fermentation, is a crucial component of the umami flavor (Ohata et al. 2007; Zhu, 2013). The degraded compounds of protein and polysaccharides by the koji enzymes were also suggested to contribute to several flavor compounds created after fermentation during the ripening process such as the Maillard reaction giving the final products their characteristic sensory quality (Ogasawara, Yamada, and Egi 2006). Relevant research used other isolates such as Aspergillus awamori and A. fumigatus showed their high enzymatic activities that were correlated to a significant increase in the levels of amino acids, desired flavor metabolites, and antioxidant activities of the fermented products (Ghanem et al. 2020; Wardhani, Vázquez, and Pandiella 2009).

Comparative studies were conducted to investigate the impact of different mold species inoculated individually or combined into a soybean fermentation process. The data showed that the fermented sample by A. elegans and R. arrhizus at a ratio of 5:1 showed a high level of key desirable volatiles, especially ethyl caproate which has a low threshold of fruity and pineapple odor notes (Cui, Li, and Liu 2016). R. oryzae by itself was associated with high levels of degraded compounds including fatty acids, and furans sulfur-containing volatiles. A. oryzae was strongly related to degraded sugars such as arabinose, mannose, and mannitol, as well as free amino acids, methionine, serine, and isoleucine while ketoglutaric acid-derived amino acids were detected with Penicillium glabrum (Park and Kim 2020; Sun et al. 2019). These studies illustrated that the quality of the fermented product is strongly correlated to the type of the microorganism present and its metabolic activities (Lee et al. 2017; Z. Li et al. 2017; Park and Kim 2020). Overall, the filamentous fungi strains that were applied to the fungal fermented soybean products mostly declined during the long subsequent brine fermentation process. Further research may be considered to identify novel strains and strategies for overcoming the long fermentation time, high salt and acid contents, and producing high-quality final products.

Criteria for selecting and developing starter cultures

According to the literature data, the microflora of spontaneous soybean fermentation is complex and its contribution to the final product quality is varied. To convert the spontaneous fermentation process into a controlled one using starter culture technology, three main areas need to be fulfilled (Vinícius de Melo Pereira et al. 2017). The first one is the ability of the selected microorganisms to carry out the desired fermentation biotransformation. The second is the final product quality that can be achieved and the last is the feasibility of a commercially developed process (see Figure 3).

Figure 3. Starter culture selective criteria in soybean fermentation. This figure was created using BioRender (https://biorender.com/).

As illustrated earlier, soybean ecology harbors a wide range of indigenous microorganisms, and these isolates might compete with the starter culture inoculum. Furthermore, foodborne pathogens, mycotoxin-producing fungi, and other contaminants that produce undesirable metabolites were also detected during the spontaneous fermentation process (Anal et al. 2020; Jeong et al. 2016; Woo et al. 2019; Yoon, Kim, and Hwang 2008). This type of complicated ecology must be considered during selecting and evaluating the microorganisms for starter culture applications. Firstly, the selected strains should be characterized by high adaptation to the fermentation environment and are potentially able to compete with the natural microflora. Such criteria might be studied by subjecting the selected isolates to grow under several stressors such as high temperature, wide pH values, organic acids, osmotic, and alcohols as a common feature of the fermentation environment and monitoring their tolerance capabilities. Successful growth under these harsh associated stressors is a potential indicator of the high fermentation performance of the isolates as starter cultures, considering that the performance is mostly strain specific. Similar approaches were applied for the early evaluation of the selected isolates before application in starter culture technology for other foods (dos Santos Cruxen et al., 2019; Elhalis et al. 2021a; Pereira et al. 2012). The inoculum density and timing of the inoculation need to be taken into consideration. Studies indicate that a high inoculum size increases the likelihood of successful fermentation and avoids the risk of stuck and sluggish fermentation under some circumstances (Lleixà et al. 2016; Papagianni and Mattey 2006). Timing is also crucial for fermentation success, especially with cocultures (Henick-Kling and Park 1994). With the increase interesting in novel fermented products, cocultures of fungi and bacteria can work. However, evaluation studies should be conducted to ensure whether they are more efficient if they are working together or inoculated sequentially. Under some circumstances, especially long-time fermentation such as cocoa and cheese, microorganisms are required to initiate the activity of another microbial group in balance throughout the fermentation process (Giraffa 2004; Ho, Zhao, and Fleet 2014; Schwan and Wheals 2004). This pattern should be considered during designing the starter cultures for similar soybean fermentation types, i.e., Soya sauce, ensuring the right inoculation timing for each microbial strain. Secondly, data from the literature show a high concentration of protein, lipid, and oligosaccharides in soybeans. These compounds require high microbial enzymatic activities including proteases, lipases, and amylases to be degraded. Various methods have been applied to screen these enzymes and measure their activities that were involved in the selection process of promising starter cultures in similar food products (Kasana, Salwan, and Yadav 2011; Li et al. 2021; Thomson, Delaquis, and Mazza 1999). The development of a novel procedure to monitor enzyme activity on a real plant matrix, like soybeans, is equally important, but challenging. Finally, the selected strain safety and its possible productivity of the antagonist compound against the natural microflora contaminate should be considered and evaluated (Abass et al. 2017; Cintas et al. 1998; dos Santos Cruxen et al., 2019; Elhalis et al. 2021a; Koutinas et al. 2009; Russo, Spano, and Capozzi 2017; Thumu and Halami 2020). For instance, literature data showed a wide range of LAB naturally presented with several fermented soybean products including Lv. brevis, L. plantarum, P. acidilactici, T. halophilus, and E. faecium (Liu et al. 2012; Ouoba et al. 2005; Sarkar et al. 1994). Studying these species showed that most of them are halophilic and can grow at high pH values; from pH 5–9, making them suitable to be used in soybean starter culture technology (Chen, Yanagida, and Hsu 2006; Holzapfel et al. 2006). In addition, they produce various secondary metabolites that were reported to impact fermented product sensory quality. Furthermore, they were potential producers of various organic acids and other antimicrobial metabolites that inhibit food contaminants and foodborne pathogens (Chen, Yanagida, and Hsu 2006; Holzapfel et al. 2006; Sarkar and Nout 2014). Such features make them suitable candidates and might be considered for starter culture trials.

One of the major soybean constraints for human consumption is the presence of anti-nutritive factors, in addition to their low digestibility and palatability (Handoyo and Morita 2006; Li et al. 2019; Mukherjee, Chakraborty, and Dutta 2015; Vagadia, Vanga, and Raghavan 2017). The roles of main microbial isolates involved in spontaneous soybean fermentation processes were elucidated through this literature review to facilitate the selection to overcome these limitations. For example, A. sojae and A. ficuum were reported to degrade phytic acid and oligosaccharides (Olukomaiya et al. 2020). Similarly yeasts such as Saccharomyces cerevisiae and C. utilis and LAB such as L. plantarum reduced oligosaccharide, trypsin inhibitors, and phytate (Adeyemo and Onilude 2013; Kasprowicz‐Potocka et al. 2018). So far, the possibility of reducing most of the anti-nutritive factors by fermentation is illustrated, thus, attention should be paid to the promising candidates with positive attributes of reducing or eliminating the anti-nutritive factors. Several traditional and advanced methods have been applied to monitoring the soy consumption constraints such as trypsin inhibitors (Hamerstrand, Black, and Glover 1981; Liu 2019), phytic acid (Haug and Lantzsch 1983; Skoglund, Carlsson, and Sandberg 1998), beany flavor (Lu et al. 2013; Surrey 1964). The microbial contribution to final product sensory quality is another important impact that should be investigated during the evaluation study. As explained in this literature review, during soybean fermentation microorganisms create a wide range of metabolites such as higher alcohols, ketones, aldehydes, and esters that potentially modulate the product taste. Each microorganism has been shown to produce distinct volatile and nonvolatile profiles associated with characteristics of sensory quality that are different from each product to another. Sensory analysis is usually conducted by determining the aroma profile of each isolate was performed using mainly headspace solid-phase micro-extraction (HS-SPME) gas chromatograph equipped mass spectrometer GCMS method (Elhalis et al. 2021b; Jacobsen and Hinrichsen 1997; Motlhanka et al. 2022; Sørensen et al. 2011; Wang et al. 1997). For nonvolatiles, high-performance liquid chromatography HPLC and ultra-high-performance liquid chromatography were commonly applied (Elhalis, Cox, and Zhao 2020; Yi and Hong 2021), while the electronic tongue and trained panel are involved at the end (Da Silva et al. 2012; Tuitemwong et al. 1993).

For commercial purposes, two main factors relating to the design of starter cultures should be fulfilled. First, the selected starter culture strains should be suitable to be cultivated on cheap substrates to decrease the overall production costs. Secondly, selected starter culture strains should tolerate the downstream processing including drying, packaging, freeze-drying, and rehydration, and have high stability during storage without losing their fermentation performance and properties (Fleet 2008; Soubeyrand, Julien, and Sablayrolles 2006). Such development of commercialization may be affected by the Nagoya Protocol on benefit sharing (Johansen 2017). Before starting any research and development work on biological resources, according to the concept of this protocol, prior informed consent by the provider country is needed. This can be completed based on mutually agreed terms describing access to the materials and benefit shares that may be monetary or non-monetary (Cooper 2023).

Further development in starter culture technology might also be achieved by genetic improvement using current molecular techniques. High-throughput screening for specific genes and metabolic pathways has been widely applied to achieve starter cultures strains with high capability to metabolize a wide range of inexpensive substrates such as lignocellulosic agricultural and forest residues (Hatti-Kaul et al. 2018). Furthermore, genes may be introduced to increase the productivity of desirable metabolites or deleted to avoid undesirable metabolic pathways, such as with LAB in dairy fermentations (Weckbecker and Hummel 2006; Xiao et al. 2010; Yu et al. 2008; Zhang et al. 2012). Such technology may enhance the positive contribution of starter cultures to food, decrease the production cost and improve stability, which is essential in today’s economic and sustainability goals, but the genetically modified organisms (GMOs) policy may raise some concerns.

Conclusion and future prospects

Soybeans are rich protein sources with a balanced amino acid content but lack sulfur-containing amino acids, combined with several consumption constraints such as beany flavor, trypsin inhibitors, phytic acid, and oligosaccharides. In Asia, they have been subjected to traditional fermentation processes and their microbial ecology is highly complicated and varied from one product to another. This literature review demonstrated the crucial roles of the fermentation process and the potential improvement of applying starter culture technology. Overall, more diverse fungi and bacteria were detected during the fermentation of Miso, Sufu, and Soya sauce while other products such as Tempeh and Natto are either mainly harbored by filamentous fungi and Bacillus ssp., respectively. The quality of the resultant fermented products varies depending on the natural microflora, the processing type of either solid-state, submerged state, and the processed ingredients. Several studies have confirmed the microbial capability to degrade soybean anti-nutritive factors, remove unpleasant beany flavors, and improve its digestibility. Starter culture applications have been an alternative way to control the fermentation process, improving the sensory quality, and enhancing its safety, nutritional, and health effects. Such strains should have high fermentation performance and enzymatic activities to tolerate the harsh fermentation process. They also should create the desired flavor, aroma, and texture and positively address various soybean constraints including oligosaccharides, trypsin inhibitors, phytates, and beany flavor. Furthermore, the ability to be commercialized and their viability during storage should be considered. Advanced understanding of spontaneous fermentation processes and starter culture technology could open a new way in directing future researchers to overcome the soybean consumption constraints, improving the final product safety, quality, nutritional and health benefits in more economic and sustainable aspects, as well as creating novel products. This understanding can be further improved using synthetic biology and strain engineering molecular techniques, as well as omics technologies, including metabolomics, which is not well applied in soybean fermentation. Further improvement that has not been well studied, to the best of the authors’ knowledge, is the application of bioreactors in soybean fermentation processes. Compared with other food fermentations such as yogurt and wines, bioreactors may offer potential capabilities to improve the overall process by reducing the cost, increasing safety, accelerating the process, and improving product consistency. Currently, most soybean-submerged fermentations are conducted in wood, concrete, epoxy cement, or steel tanks, while solid-state fermentations are conducted on individual trays and handled manually. Further investigation is required to improve the overall quality and increase the production scale in response to the current demand for sustainable protein sources.

Authors’ contributions

H.E.: compiled data from literature review and drafted the manuscript; C.X.H.: revised the manuscript; Y.C.: Supervision, wrote, Reviewed and Edited. All authors have read and agreed to the published version of the manuscript.

Acknowledgments

The authors gratefully acknowledge financial support from the Agency for Science, Technology and Research (A*STAR).

Disclosure statement

No potential competing interest was reported by the authors.

Additional information

Funding

This work was supported by the Agency for Science, Technology and Research (A*STAR) under the CRF-ATR Programme.