Folate Biofortification in Soybean: Challenges and Prospects

Abstract

Folate deficiency is a significant global health issue that affects millions of people and causes severe adverse effects. Major staple crops, which provide significant amounts of calories, often contain inadequate folate levels. Synthetic fortification has contributed to a reduction in low-folate populations, but a more sustainable solution is needed. Biofortification, or the breeding of crops to naturally increase their nutrient content, is a promising alternative. Soybean is a highly nutritious crop and a good candidate for folate biofortification. However, studies on folate have been limited due to the challenges in folate analysis. The development of sensitive and selective tools, reference materials, and studies on the stability of folate vitamers in crops has facilitated the development of improved folate determination methods. Additionally, the soybean folate biofortification program can be improved by leveraging previous studies in major cereals, common bean and pea, as well as combining conventional breeding with new genomics approaches. In this review, we discuss the folate content, composition, and analytical challenges in soybean and suggest possible frameworks and strategies for folate biofortification in soybean. We also conducted an in silico analysis of key folate biosynthesis enzymes in soybean.

1. Introduction

Folate (vitamin B9) is an essential water-soluble vitamin in plants and microorganisms that plays a role in one-carbon metabolism [1,2]. It functions as a cofactor in the synthesis of nucleic acids, metabolism of amino acids, and methylation of hormones, lipids, proteins, and chlorophyll [3]. Folate is particularly important for cell division in pregnant and lactating women [4]. However, humans cannot synthesise folate de novo and must obtain it from dietary sources, such as crops, animal-based foods, or nutritional supplements [5]. Despite this, folate intake in humans often falls below the recommended levels of 400 μg/day for adults and 600 μg/day for pregnant women [6]. The folate content of the major staple crops, including wheat, maize and rice [7], ranges from 10 to 91 μg/100 g [8,9,10,11], 33 to 129 μg/100 g [7,8] and 11 to 111 μg/100 g [7,8,12,13], respectively. To address the low folate levels of most populations, as well as the low folate concentrations in most staple crops, many developed countries have implemented the mandatory fortification of cereal products. However, mandatory fortification in developing and some developed countries can be hindered by financial constraints, imbalances in distribution, and the absence of industrial food systems, limiting its sustainability. Additionally, there have been concerns about the potential carcinogenic effects of chronic exposure to synthetic folic acid [2].

Micronutrient malnutrition, also known as hidden hunger, affects two billion people globally and is caused by the insufficient intake of micronutrients [14]. Hidden hunger may be partly attributed to the one-sided focus on increasing the caloric and macro-nutrient content of staple crops over the years, which has resulted in inadequate levels of micronutrients in these crops. The major hidden hunger deficiencies include vitamin A deficiency, which can cause blindness and impaired immunity; iron deficiency, which can cause anaemia and impaired cognitive function; zinc deficiency, which can cause growth retardation and impaired wound healing; and folate deficiency, which can cause birth defects, neurological and cognitive problems, anaemia and fatigue. In 2000, the United Nations adopted the Millennium Development Goals (MDGs) to address major global issues, including child mortality and maternal health, which are key indicators of micronutrient deficiencies. In 2015, the United Nations established 17 Sustainable Development Goals (SDGs), with a focus on zero hunger (SDG2) and good health and well-being (SDG3) (Global Nutrition Report 2018) [15]. As Ban Ki-Moon, United Nations 8th Secretary General, stated, “Improved nutrition is the platform for progress in health, education, employment, empowerment of women and the reduction of poverty and inequality, and can lay the foundation for peaceful, secure and stable societies”. Enhancing the nutritional content of staple and major crops is, therefore, a key priority, as outlined in SDGs 2 and 3. By improving the nutritional content of crops through biofortification, we can provide access to diverse and nutritious diets and prevent hidden hunger and other deficiencies.

Biofortification is the process of increasing the nutritional content of crops through selective breeding, genetic modification, fermentation, and or the application of fertilizers or other nutrients [3,16]. Its goal is to address specific nutrient deficiencies or improve the overall nutritional quality of crops. It can serve as a sustainable and cost-effective intervention for addressing nutrient deficiencies and improving nutrition in low-income and resource-poor areas where access to diverse diets may be limited and other interventions, such as supplementation or fortification programs, may be impractical [17]. Biofortification can help to improve access to essential micronutrients and contribute to the achievement of SDGs 2 and 3.

Soybean (Glycine max L. Merrill) originated from China and is an inexpensive legume consumed worldwide as a major source of vegetable protein and oil [18]. In recent years, soybean has gained much interest due to their reported benefits on human nutrition and health [19]. Additionally, soybean can be transformed into various food products, including tempeh, tofu, soymilk, soybean sprouts, soy sauce, vegetable soybean (edamame or maodou), and natto. Therefore, soybean is cultivated across a wide range of latitudes and can be considered a staple in some countries. According to FAOSTATS 2019, soybean production exceeds 300 million tons per annum, with the highest production in Brazil (114 million tons), the United States of America (96 million tons), Argentina (55 million tons), China (15 million tons) and India (13 million tons). Soybean is an ideal crop for biofortification due to its high protein and oil content, versatility as a food and feed ingredient, and adaptability to a range of environments. With a relatively short growing season, soybean is also suitable for biofortification efforts in various regions. The biofortification program can be utilised in soybean to increase the levels of micronutrients and to improve the bioavailability of nutrients [20]. The combination of soybean’s potential for biofortification and its high protein and oil content, as well as its versatility, make it a valuable asset in efforts to enhance global nutrition.

This review aims to provide an overview of the nutritional profile, folate content and composition, analytical challenges, biosynthetic pathway, and the advances and prospects for the biofortification of folates in soybean.

2. Soybean as a Vital Source to Address Malnutrition

Vegetarian diets have become increasingly popular globally for ethical, ecological and health reasons. As a result, soybean, which is an excellent source of protein (ranging from 31.70–57.90%), having all essential amino acids for humans [21,22] (

Table 1. The nutritional profile of the soybean seed.

Nutrient (Unit)	Concentration	Reference
Major nutrients
Protein (%)	31.70–57.90	USDA-ARS; [21,25]
Oil (%)	6.50–25.60	USDA-ARS; [26]
Oleic acid (%)	4.64–36.31	[23,25]
Linoleic acid (%)	7.38–63.90	[23,25]
Linolenic acid (%)	1.38–12.80	[23,25]
Stearic acid (%)	0.70–7.50	[23,25]
Palmitic acid (%)	2.02–15.20	[23,25]
Carbohydrate (%)	30.16–35	USDA
Soluble sugar (%)	7.42–14.42	[25,27]
Micronutrients
Vitamin B1-Thiamin (mg/100 g)	0.90–8.00	[28,29,30,31]
Vitamin B2-Riboflavin (mg/100 g)	0.20–0.33	[28,29,31]
Vitamin B3-Niacin (mg/100 g)	1.60–4.00	[28,29,31]
Vitamin B5-Panthothenic acid (mg/100 g)	0.34–1.14	[28,29,31]
Vitamin B6-Pyridoxine (mg/100 g)	0.22 –1.09	[28,29,31]
Vitamin E-Tocopherol (mg/100 g)	10.00–36.00	[25,32,33]
Vitamin B9-Folate (mg/100 g)	0.06–0.69	[25,34]
Minerals
Potassium (mg/100 g)	1800–2301	[31]
Phosphorus (mg/100 g)	630–704	[31]
Magnesium (mg/100 g)	257–296	[25,31]
Sodium (mg/100 g)	2.00–3.70	[31]
Calcium (mg/100 g)	201–317	[25,31]
Zinc (mg/100 g)	2.32–4.89	[25,31]
Iron (mg/100 g)	7.30–15.7	[25,31]
Manganese (mg/100 g)	2.40–3.60	[25,31]
Bioactive compounds
Isoflavone (mg/100 g)	74.50–525.39	[25,35,36]
Carotenoid (mg/100 g)	0.32–2.92	[25,37]
Saponin (mg/100 g)	444.60–464.00	[38,39]
Phospholipid (mg/100 g)	38,100–45,000	[40]
Sterol (mg/100 g)	205.00–287.00	[41]

USDA-ARS—U.S. Department of Agriculture-Agricultural Research Service.

), has gained popularity. Soybean oil is high in polyunsaturated fatty acids, with a moderate amount of monounsaturated fatty acids and no trans fats [23]. Soybean oil provides over 40% of essential fatty acids, including omega-3 fatty acid, α-linolenic acid and omega-6 fatty acid linoleic acid. Compared to most crops and legumes, soybeans are relatively low in carbohydrates, with a very low glycaemic index, and are a rich source of micronutrients, including vitamins and minerals (Table 1). Soybean is a good source of folate, with levels typically ranging from 64.51–691.24 μg/100 g (Table 1). This is significantly higher than other crops (

Table 2. The folate content of various crops.

Crop	Folate Content (μg/100 g)	Reference
Chickpea	351.00–589.00	[42,43]
Common bean	113.00–296.00	[42,43,44]
Lentils	136.00–361.00	[42,43,45]
Maize	33.40.00–129.00	[7,8]
Pea	19.50–55.00	[42,43,46]
Peanut	81.00–240.00	[47,48]
Rice	11.00–111.00	[7,8,12,13]
Soybean	64.51–691.24	[25,34]
Tomato	14.00–46.00	[49]
Wheat	10.00–91.00	[8,9,10,11]

), making soybean a valuable source of this micronutrient. Soybean is also the richest source of isoflavones, with concentrations ranging from 74.50–525.39 mg/100 g (Table 1). Additionally, the soybean seed contains other phytochemicals, including saponin, lunasin, and sterols [24], making it a potential source of essential nutrients and nutraceuticals with implications for human health.

3. Structure, Distribution, Content and Composition of Folate in Soybean

3.1. Structure and Distribution of Folate Vitamers in Soybean

Folate comprises a pterin moiety attached by a methylene bridge to para-aminobenzoic acid coupled to one or more glutamyl residues. In vivo, folates exist as tetrahydrofolate (H₄folate) and its derivatives, which vary in the state of oxidation of the pteridine ring, single carbon substituents linked at the 5 and 10 positions, and with a variable number of glutamyl residues [50]. In theory, over 150 folate derivatives exist, but a few exist in plants and humans.

Recent chromatographic methods have indicated the folate vitamers present in soybean seeds include H₄folate (THF), 5-CH₃-H₄folate (5MTHF), 5-CHO-H₄folate (5FTHF), H₂folate (DHF), 10-CHO-H₄folate, PteGlu (FA), 5,10-CH=H₄folate (5,10-MTHF), 10-CHO-PteGlu (10FFA), and MeFox (Figure 1). However, the distribution of these vitamers in soybean seeds has not been frequently reported. Soybean vitamer distribution can be influenced by factors such as storage, plant developmental stage, cultivar, accession type, environment and analytical method. Some studies have reported THF as the most abundant folate vitamer in soybean [51,52], while other studies have identified 5FTHF as the most abundant [53,54]. In our recent study, 5FTHF accounted for more than 60% of the total folate content of soybean accessions grown in southern China [34]. Climate conditions, particularly temperature and humidity, may influence folate synthesis and accumulation. For example, lower temperatures have been shown to increase 5MTHF accumulation in sweet corn seedlings [55], while higher temperatures can lead to a significant decrease in the folate content of lettuce [56].

Whereas a few studies have reported THF as the most dominant vitamer in soybean [51,52], that was not observed in our recent studies [34]. THF is one of the most labile vitamers that can easily degrade or convert right from harvest. We assume that THF might have been converted into FA or degraded right after harvest and during post-harvest storage. Studies on the homeostasis and dynamics of plant folates during storage have been limited. Biosynthesis may occur during post-harvest storage in some organisms, while variations in the folate pool in other organisms may be due to spontaneous or enzymatic reactions [57]. Further studies will be necessary to confirm vitamer distribution during post-harvest storage in soybean seeds.

3.2. The Folate Content and Composition of Soybean and Soy-Based Products

Soybean can be processed into various forms through germination, soaking, boiling, and fermentation [58,59]. However, folates are sensitive to light, air, heat and pH; therefore, their content and composition may be indirectly affected during processing. Mature unprocessed soybean seeds contain total folate levels ranging from 64.51–691.24 µg/100 g FW and 199 to 464 µg/100 g DW (Table 3). The differences in folate contents can be attributed to cultivar type, environment, and analytical methods used. In our recent study, we identified a 10-fold variation in the total folate content of over 1000 germplasm consisting of landraces and cultivars [34]. However, most studies have been limited to a few cultivars. A larger sample size is needed to evaluate the variation among soybean folates and, most importantly, different accession types (wild-type, landraces and cultivars).

Cooked soybean seeds contained lower levels of folate (44.70–77.90 µg/100 g) at a retention rate of 24–45%, depending on the cooking treatment [60]. Similarly, significant losses were observed during the preparation of soymilk and tofu [59]. The significant loss of folates during tofu and soymilk processing was mainly caused by soaking and boiling, as most folates were recovered in the cooking or soaking media. Thus, the major cause of folate loss is leaching and, to a certain extent, oxidation. To avoid such losses, shorter boiling times and possible consumption of the cooking media are recommended.

Folate in tempeh ranged from 149.30 to 416.40 µg/100 g. Fermentation contributes to the increase in folate during tempeh preparation. During fermentation, folate compounds may be liberated by the actions of enzymes produced by the microorganism, leading to increased folate concentrations. The de novo formation of folate compounds during fermentation may also increase folate content [54]. However, the extent of the increase of folates will depend on the microorganism involved [58].

Seed germination is an age-old practice used to improve the nutritional value of crops, especially legumes [61]. Germination affects folate profiles and alters the distribution of individual vitamers [62,63]. In soybean, folate content increased 3.5 to 3.7-fold from 230.50 to 815.20 μg/100 g in Bangladesh soybean-4 and from 202.90–759.50 μg/100 g in Heinong 48 [51]. In germinated soybean, approximately 80% of the total folate content is 5MTHF, the most active folate vitamer.

In our recent study, the total folate content of immature soybean seeds at the R6–R7 stage, a typical stage for edamame or maodou, ranged between 344.06–685.81 μg/100 g FW among 12 soybean cultivars, with 5MTHF contributing approximately 70% of the total folate content and 5FTHF contributing approximately 15% [25].

Table 3. The folate content of soybean seeds and soy-based products.

Sample	THF	5MTHF	5,10-MTHF	10FFA	5FTHF	DHF	FA	Total Folate	References
Soybean seed		16.90		53.80	121.00			199.00–464.00	[53,54,59,64,65]
Soybean seed	20.00–75.00 *	28.00–205.74 *	5.00–28.60	11.00–71.06	160.00–590.56	2.90–29.44	28.50–34.40 *	64.51–691.24 *	[34,51,52,60]
Soybean seed (cooked)								44.70–77.90 *	[60]
Vegetable soybean	12.55 *	356.18 *	10.17 *	4.33 *	75.07 *	2.98 *	1.00 *	344.06–685.81	[25]
Soymilk								34.00–276.00	[8,64]
Tofu								15.00–127.30	[8,58,64]
Tempeh					231.80			149.30–416.40	[58,64]
Soybean sprouts								759.50–815.20 *	[51]

The folate content of soybean seeds and soy-based products μg/100 g on dry weight basis. Values with * indicate folate content on fresh weight basis.

Table 3. The folate content of soybean seeds and soy-based products.

Sample	THF	5MTHF	5,10-MTHF	10FFA	5FTHF	DHF	FA	Total Folate	References
Soybean seed		16.90		53.80	121.00			199.00–464.00	[53,54,59,64,65]
Soybean seed	20.00–75.00 *	28.00–205.74 *	5.00–28.60	11.00–71.06	160.00–590.56	2.90–29.44	28.50–34.40 *	64.51–691.24 *	[34,51,52,60]
Soybean seed (cooked)								44.70–77.90 *	[60]
Vegetable soybean	12.55 *	356.18 *	10.17 *	4.33 *	75.07 *	2.98 *	1.00 *	344.06–685.81	[25]
Soymilk								34.00–276.00	[8,64]
Tofu								15.00–127.30	[8,58,64]
Tempeh					231.80			149.30–416.40	[58,64]
Soybean sprouts								759.50–815.20 *	[51]

The folate content of soybean seeds and soy-based products μg/100 g on dry weight basis. Values with * indicate folate content on fresh weight basis.

4. Analytical Challenges in Folate Quantification

Folate analysis in foods is challenged by the multiplicity of folate forms, the stability requirements of different folate vitamers, the low concentration of naturally occurring folate and the complexity of sample matrices. In an interlaboratory analysis of soybean flour, a wide variation was found in the folate concentrations of the flour samples from different laboratories, highlighting the limitations of current folate extraction and quantification methods [65]. Folate analysis involves the liberation of folates in the matrix, deconjugation of folate polyglutamates into monoglutamates, and quantifying folates by microbiological assay or chromatographic methods. Folate liberation into the sample matrix is influenced by the pH and concentration of the extraction buffer and by heat.

The pH of the extraction buffer plays a critical role in folate liberation because the stability of folate vitamers differs significantly with pH. Studies on the stability of folate standards in different pHs have been widely reported, but the stability of folates in a sample matrix has rarely been studied. In our study of seven extraction buffer pH conditions (4.5, 5.5, 6.5, 7, 7.5, 8.5 and 9) in the soybean seed matrix, the total folate content was highest at pH5.5 (Figure 2A) [34]. The highest total folate recovery was due to the highest recoveries of THF, 5MTHF, and 5FTHF at pH5.5, which corroborated a previous study in mung bean that reported the highest total folate content within pH4.5–5.5 [66]. Similarly, an extraction buffer of pH6 was reported to be stable for THF in food matrices [67]. Lower pHs contribute to the precipitation of proteins, which may reduce matrix influences during extraction.

Heating can aid in liberating folates in the cell matrix, but the effectiveness may depend on temperature and time. For instance, folate recoveries for food matrices were higher at 100 °C for 15 min than at 70 °C for 60 min [68]. Similarly, the total folate content was higher when soybean samples were boiled at 100 °C for 15 min than for 10 and 5 min (Figure 2B) [34]. In our study, recovery for DHF, the least occurring vitamer, decreased with increased boiling times, confirming the extreme lability of this folate vitamer. To improve the stability of folates during extraction, the combined use of thiols and ascorbic acid has been recommended [5].

In previous years, the tri-enzyme treatment comprising α-amylase, protease and conjugase was preferred for folate extraction. α-amylase breaks down starch, protease degrades proteins, and conjugase converts folate polyglutamates to monoglutamates. However, recent studies have suggested that treatment with α-amylase and protease may not be necessary. The addition of these enzymes leads to longer extraction times and possible oxidations and interconversions. Most naturally occurring folates are polyglutamates, and the deconjugation step is necessary for monoglutamate quantification. Accordingly, the combined usage of chicken pancreas and rat serum as conjugase has proven to be as effective as the tri-enzyme method or even four enzymes in legume crops (Figure 2C) [34,43,69]. Furthermore, extraction time is reduced using conjugase treatment alone. Studies indicate that chicken pancreas, which contains endogenous enzymes, such as protease and α-amylase, can convert polyglutamates into diglutamates, while rat serum converts the diglutamates to monoglutamates. Recently, plant-based and human-based conjugases have been reported, and their efficiency is much higher than animal-based conjugases [67,70,71]. While higher volumes of animal-based conjugases are used, fewer plant-based or human-based conjugases are required to effectively deconjugate polyglutamates in legumes. However, animal-based conjugases are cheaper and more readily available than plant-based or human-based conjugases (according to our experience). Making plant-based and human-based conjugases readily available in the markets at low costs will be important for folate quantification in crops.

Traditionally, folate quantification was performed using the microbiological assay (MA) method. The MA method, which utilises estimation of the growth of bacteria, such as Lactobacillus sp., is a very sensitive method but is tedious, time-consuming, and unable to differentiate between different folate vitamers. The introduction of chromatographic methods coupled with detectors, such as ultra-violet (UV), fluorescence (FLD), and mass-spectrometry (MS) or tandem mass spectrometry (MS/MS), overcame the disadvantages of the MA by being able to quantify all folate forms separately in a single run. HPLC coupled with UV and fluorescence detectors (HPLC-UV/FLD) and HPLC coupled with tandem mass spectrometry (HPLC-MS/MS) have been used in soybean. While difficulties in separating masked target peaks with matrix interferences have been reported using HPLC-UV/FLD, mass spectrometric detection is the optimal choice for folate analysis due to its sensitivity, selectivity and applicability to various folate forms. Internal standards are used in MS/MS quantification to compensate for analyte losses during sample preparation and injection. However, due to the high costs of isotope-labelled internal standards, folate analogues, such as methotrexate (MTX), have been used as internal standards in some studies [72]. Despite this, stable isotope-labelled internal standards are preferred for very sensitive or food-grade experiments. Additionally, certified reference materials (e.g., BCR485 and BCR487) are recommended to ensure the trueness of the protocol [73]. This will potentially reduce the variations and discrepancies in reported values.

5. Folate Biosynthesis

In soybean, just as in plants, there are three pathways involved in the synthesis of folates (Figure 3). The precursors, pteridine and para-aminobenzoate (pABA), are produced from guanosine triphosphate (GTP) in the cytosol and chorismate in the plastids, respectively, and are then transported to the mitochondrion to form polyglutamate tetrahydrofolate [74,75]. The biosynthesis of pteridine starts in the cytosol, where the enzyme guanosine triphosphate cyclohydrolase 1 (GTPCHI) hydrolyses GTP to form 7,8-dihydroneopterin 3′-triphosphate (DHN-P₃) However, it is unclear how DHN-P₃ is converted to DHN, as some studies suggest that the conversion is non-enzymatic, while others propose that it occurs in two steps involving the removal of pyrophosphate and the hydrolysis of dihydroneopterin monophosphate to DHN [74,76]. According to the KEGG database [77], there are no specific enzymes in the two-step process to convert DHN-P₃ to DHN. However, an uncharacterised alkaline phosphatase was shown to directly convert DHN-P₃ to DHN. This remains to be further investigated. The enzyme dihydroneopterin aldolase (DHNA) cleaves the side chain DHN to 6-hydroxymethyl-7,8-dihydropterin (HMDHP) and glycolaldehyde. It also enables the epimerisation of C-2 carbon of the DHN side chain to form 7,8-dihydromonapterin (DHM), which can be further cleaved to form HMDHP [78,79].

In order to synthesise pABA from chorismate in the plastids, two enzymes, amino deoxychorismate synthase (ADCS) and amino deoxychorismate lyase (ADCL) catalyse the process in two steps. The first step involves the transfer of the amide nitrogen of glutamine to chorismate to form 4-amino-4-deoxychorismate (ADC). In the second step, pyruvate is eliminated and aromatised to give rise to pABA. The HMDHP and pABA moieties are assembled into the mitochondrion to form tetrahydrofolate (THF, H₄folate). HMDHP is converted to dihydropteroate (DHP) in two steps, both of which are mediated by the enzymatic activities of HMDHP pyrophosphokinase (HPPK) and dihydropteroate synthase (DHPS). In soybean, these two enzymes are found to be coupled on a single polypeptide. Next, dihydrofolate synthase (DHFS) couples DHP to glutamate to form DHF. In soybean, folylpolyglutamate synthase (FPGS) is also present with DHFS to form DHF, although they function separately, contrary to the bifunctional FPGS-DHFS in bacteria [80]. The exact role of FPGS in this reaction may be the production and attachment of glutamate residues to be used by DHFS to form DHF but may require further studies.

Finally, in the penultimate step of folate biosynthesis, the bifunctional enzyme dihydrofolate reductase-thymidylate synthase (DHFR/TS) reduces DHF to THF. The polyglutamate tail of THF-Glu_n is formed when γ-lined glutamate residues are sequentially added to THF by FPGS. The polyglutamate tails can be hydrolysed by γ-glutamyl hydrolases (GGHs), which are important in the regulation of folate homeostasis.

In Silico Analysis of Major Folate Biosynthesis Enzymes in Soybean

To identify the putative orthologs of key folate enzymes in soybean, the phytozome database (https://phytozome-next.jgi.doe.gov/ accessed on 30 October 2022) was queried using Arabidopsis thaliana GTPCHI, ADCS, HPPK-DHPS, DHFR/TS, FPGS and GGH. Genes with similarities >70% were selected (Table S1). The genomic and protein sequences of selected genes were downloaded, and the gene structures were drawn using TBtools [81]. Analysis of conserved amino acids was conducted using NCBI Conserved Domain Database (https://www.ncbi.nlm.nih.gov/Structure/cdd/cdd.shtml/ accessed on 30 October 2022), and the conserved motifs were identified using the MEME suite [82]. Multiloc2 [83] was used to determine the most probable subcellular location of the predicted proteins.

Our results showed that most folate biosynthesis enzymes in soybean are encoded by multiple copies of genes (Table S1) [84]. From our search, we found five isoforms of GTPCHI and FPGS, four isoforms of DHFR/TS, three isoforms of GGH, and two isoforms of ADCS and HPPK-DHPS in soybean. The architecture and structure of the folate biosynthesis genes were studied to gain a better understanding of their evolution, which could lead to future genome-wide studies. The structures of GTPCHI, ADCS, and DHFR/TS genes were similar among their isoforms, which may indicate structural similarity during evolution. The 5 isoforms of FPGS were divided into 2 groups, with 1 group containing 15 exons with longer genes and the other having 14 exons and shorter genes (Figure 4A). Both isoforms of HPPK-DHPS contained two exons but different introns and gene lengths. GGH isoforms were all different from each other and may require genome-wide studies to understand their distribution in soybean.

The conserved domain analysis showed that all GTPCHI isoforms contained two domains of the tunnelling fold (T-fold) superfamily (Figure 4B). The analysis also revealed that both isoforms of ADCS contained two domains: the chorismate bind superfamily and the glutamine amidotransferase (GATase) superfamily. The presence of an HPPK superfamily domain and pterin-binding superfamily confirmed the bifunctionality of HPPK-DHPS genes. Similarly, all four isoforms of DHFR/TS contained two domains: a TS superfamily and a DHFR superfamily, confirming the bifunctional function of this gene in soybean. No conserved domains were identified for FPGS genes, and only one of the three GGH isoforms, GmGGH-1, contained a GATase superfamily. The motif patterns were well conserved for most genes and their isoforms (Figure 4C), except for GGH and ADCS. The consistency of the protein domains and motifs among all isoforms suggests that the selected genes are reliable.

Prediction of the subcellular localisation showed that soybean GTPCHI is cytoplasmic, consistent with recent studies and reports in other plants [85]. In Arabidopsis, ADCS was targeted to only the chloroplast [86]. However, our studies suggest that in soybean, ADCS can possess both chloroplast and mitochondria targeting signals. In our studies, HPPK-DHPS was predicted to be localised in the cytoplasm and mitochondria. DHFR/TS was predicted to be targeted to multiple compartments, including the cytoplasm, chloroplast, and mitochondria, which is consistent with previous studies in Arabidopsis [87]. All soybean isoforms of FPGS proteins were predicted to be cytoplasmic, and for GGH, the analysis revealed they localised in secretory pathways, as has been reported in other plants [88,89].

6. Prospects for Biofortification of Folates in Legumes

Biofortification can be classified based on a variety of approaches, including agronomic and genetic methods and the consumption of functional foods, such as fermented foods and sprouts [3,17]. Agronomic biofortification involves using agricultural techniques to increase the nutrient content in crops. This can include applying fertilizers, such as zinc or iron, or soil microbes to the soil in which the crops are grown [90,91,92,93]. The soil is the primary source of nutrients, and the abundance of nutrients in the soil and their availability to plants determine the synthesis of plant metabolites [94]. Agronomic biofortification is relatively simple and straightforward but can be expensive and time-consuming [95]. This is because this approach can be limited by the inherent variability in the soil nutrient availability, which is affected by factors such as soil type, climate and crop requirements. As a result, the effectiveness of agronomic biofortification may vary and may require extensive maintenance and input. To date, agronomic fortification has not been studied for folates in soybean. However, research has shown that the folate content of plants can increase when they have access to nutrients, such as phosphorus, nitrogen, and boron, suggesting that agronomic fortification of folates in soybean may be worth investigating in the future [96].

Genetic biofortification is the most well-known form of biofortification and has been widely studied and researched by scientists and plant breeders. It involves using genetic techniques to increase the nutrients in crops and can be achieved through several approaches, including conventional breeding, genomics-assisted breeding, and metabolic engineering.

Conventional breeding involves using traditional breeding techniques, such as crossbreeding and selection, to develop new varieties of crops that are rich in nutrients such as folate. Conventional breeding typically involves several steps, including identifying genetic variation in a germplasm pool, selecting elite-folate accessions and crossing these accessions with local varieties to create new varieties with improved nutrient content. One of the main advantages of conventional breeding is that it is a relatively low-cost and low-risk approach to biofortification. However, it can be a slow process, as it can take several years to develop and test new varieties of crops. Additionally, crops with low genetic variation for folate [57] may not be suitable for conventional breeding as there may not be sufficient genetic diversity to create new varieties with improved folate content.

Overall, conventional breeding can be an effective approach to folate biofortification, particularly in crops that have a high level of genetic variation, such as rice, potato and soybean [34,97,98]. Additionally, wild and landrace accessions of crops are often adapted to local growing conditions and may contain important genes that can be used to improve the nutrient content of cultivated species [99,100]. For example, wild lentils and soybean landraces have been found to contain higher folate content than cultivated accessions and introgressing these traits into cultivated varieties could help to increase the overall folate content of these crops [34,45].

Genomics-assisted breeding involves the use of genomic tools and technologies such as marker-assisted breeding (MAB), marker-assisted selection (MAS), marker-assisted backcrossing (MABC) and marker-assisted recurrent selection (MARC) to improve crop breeding programs [101] (Figure 5). These techniques can be used in conjunction with multi-omics data, databases, and genes generated by genomics to better understand the genetic basis of traits such as folate and to develop new varieties with improved folate profiles. High-throughput genotyping methods, such as genotyping by sequencing (GBS), can be used to genotype a large number of accessions [102], while phenotyping methods can be used to measure the folate content of these accessions. Linkage mapping, genome-wide association studies (GWAS) [103], and bulked segregant analysis-sequencing (BSA-Seq) [104] can then be used to identify quantitative trait loci (QTL) [105] or quantitative trait nucleotides (QTN) [106] that are associated with folate content. RNA-Seq can be combined with identified QTL or QTN and analysed using weighted gene co-expression network analysis (WGCNA) to group genes with similar expression patterns and identify candidate genes [107]. After the identification of candidate genes, functional analysis and metabolic pathway analysis can be conducted to further understand their roles in folate biosynthesis. MAB can then be utilised to develop new hybrids with improved folate content.

Many QTL and genes associated with folate content have been identified in various crops using these techniques, including pea [46], common bean [44,108], rice [97], maize [109], potato [110] and sweet corn [111]. For example, a genome-wide association study in 85 pea accessions revealed 9 SNPs significantly associated with 5MTHF, 5FTHF, THF and total folate, with two SNPs linked to higher levels of 5MTHF and total folate content. Linkage mapping has also been used to identify 4 QTL for 5MTHF in common bean, while a GWAS in 96 common bean genotypes identified 6 QTL for folate accumulation, including QTL on chromosome 11 that occurred in genomic regions syntenic to already reported QTL in maize, rice and common bean [44] (

Table 4. QTL identified for folate content in legumes.

Crop	Population Type	Population Size	Total Number of QTL Identified	PVE (%)	Model of Analysis	Reference
Pea	Natural population	85	9	-	MLM	[46]
Common bean	Bi-parental	6	4	8–19	SMA	[108]
Common bean	Natural population	96	6	-	Fast-LMM/EMMA	[44]

PVE—phenotypic variation explained; MLM—mixed linear model; SMA—single marker analysis; Fast-LMM—factored spectrally transformed linear mixed model; EMMA—efficient mixed model analysis.

). However, to date, there have been no studies on QTL/QTN related to folate components in soybean.

One way to improve the folate content of crops is through metabolic engineering. This involves modifying the genetic makeup of crops to increase the production of specific nutrients or other bioactive compounds. Conventional breeding techniques may not always be sufficient to improve the folate content of crops in which case metabolic engineering can be used [17]. The metabolic engineering of folates has focused on enhancing folate synthesis [3]. Most of the enzymes that partake in the biosynthesis of folates have been characterised and cloned. However, transcriptomic and gene expression studies have revealed that the key enzymes in folate synthesis include GTPCHI, ADCS, HPPK/DHPS, FPGS and GGS. GTPCHI, a homolog of the folE gene in E. coli, is the first enzyme of the de novo biosynthesis pathway of folates in bacteria, fungi, and plants. Hence, it has been mostly targeted for metabolic engineering because it is thought to be the first rate-determining step controlling flux into the folate pathway [112]. Similarly, ADCS catalyses the first step of pABA synthesis in plants.

Regarding the metabolic engineering of folates in legumes, a single study has been conducted in the common bean with no reported investigation in soybean. The Mexican common bean (Phaseolus vulgaris L.) was metabolically engineered by overexpressing GTPCHI, which enhanced folate levels in the seeds by 3-fold and pteridine levels by 150-fold [113]. In other studies, two key soybean folate biosynthesis genes, GmGCHI (GTPCHI) and GmADCS (ADCS) were cloned and co-overexpressed in maize and wheat [114]. Transgenic maize and wheat grains had folate content increased by 4.2-fold and 2.3-fold, respectively. A subsequent co-expression of GCHI from soybean and ADCS from tomato significantly increased folate levels in wheat (Gm8gGCHI⁺/LeADCS⁺).

Studies in other species and crops have shown that the metabolic engineering of folate biosynthesis genes enhances folate levels. The overexpression of the folE gene in microorganisms resulted in increased folate levels [115]. The ectopic expression of GTPCHI increased folate levels in rice by 3.3 to 6.1-fold [116], increased the folate content of tomato by an average of 2-fold and pteridine levels by 140-fold and increased lettuce folates from 2.1 to 8.5-fold [117]. A 1250-fold increase in pterins and a 2- to 4-fold enhancement of folates in E. coli folE overexpressed Arabidopsis plants have been reported [112]. Folate levels were also doubled by the overexpression of E. coli folE in a simultaneous biofortification study with vitamins β-carotenoid and vitamin C [118]. A 1.5–1.8-fold increase was reported in transgenic AtADCS lines [116].

Increasing the pteridine levels of tomato by the overexpression of GTPCHI resulted in the depletion of pABA, increased pteridine levels, and a 2-fold increase in folates [119]. The depleted pABA levels indicated that the pABA supply limits further accumulation. Further exogenous application of pABA in pABA-depleted crops increased the folate levels [119,120]. Thus, a combined engineering of p-ABA and pteridine production in tomatoes achieved 25-fold higher folates than controls and increased pABA and pteridine levels [121]. Similarly, the two-gene strategy for metabolic engineering of pterin and pABA in rice resulted in a 100-fold increase in folates [122]. However, co-overexpressing GTPCHI with ADCS could not increase folate content in potato and Arabidopsis [123], suggesting the need to engineer other pathways.

HPPK/DHPS is bifunctional and combines activities catalysing two consecutive steps to form dihydropteroate from HMDHP. HPPK/DHPS performs the condensation reaction of pterin and pABA in the mitochondria, while in E. coli, HPPK and DHPS are monofunctional enzymes, with their encoding genes being folk and folP, respectively. However, these enzymes are coupled as one protein in plants, protozoa, and fungi. Wheat HPPK/DHPS has been singly overexpressed in rice, which increased folate content 2-fold [124]. However, overexpressing AtHPPK/DHPS in rice resulted in no difference in folate levels [116].

FGPS is the enzyme that catalyses the attachment of the glutamate tail to the THF molecule. Polyglutamylation by FPGS is regarded as an essential regulatory point in folate metabolism [5,125]. This is because the overexpression of GGH in animal systems reduces polyglutamate abundance and intracellular folate levels, whereas increased FPGS enhances intracellular folate levels. For instance, over-expression of GGH in Arabidopsis and tomato resulted in reduced folate levels [125]. Thus, FGPS plays an important role in folate homeostasis. Overexpression of AtFPGS in rice increased seed folate content by 7.50 to 19.90% and 4.30–45.50% [116]. In a recent study, the overexpression of foxtail millet FPGS gene SiFPGS2 in Arabidopsis increased folate content [126]. Owing to the previous unsuccessful attempt to enhance folates in potatoes by the two-gene approach (GTPCHI and ADCS), the four-gene approach overexpressing GTPCH1, ADCS, HPPK/DHPS and FPGS was studied, resulting in a 12-fold increase in folate content [127].

Soybean is a promising candidate for folate biofortification due to the variation and diversity of its folate content, as revealed in recent studies. Additionally, the availability of the soybean reference genome, along with advanced next-generation sequencing technologies and developed omic databases, has facilitated a deeper understanding of the genetics underlying various agronomic traits in soybean [128,129]. This understanding can be leveraged to develop new varieties with enhanced folate content through approaches such as conventional breeding, genomics-assisted breeding and metabolic engineering.

Some functional foods can be used for biofortification, such as fermented foods and sprouts. For example, traditional fermented soy-based foods, such as tempeh, natto, miso, soy sauce, douchi, and fermented soymilk, can help to increase the bioavailability of nutrients, making them more easily absorbed and used by the body [16,130,131]. Fermentation can also help to increase the overall nutrient content of these foods, as many microorganisms used in fermentation can synthesise essential vitamins and minerals. Fermentation has been shown to increase the folate content of some fermented soy-based foods such as tempeh and soymilk [58,59,132]. Similarly, soybean sprouts, which are made by germinating soybean seeds, are rich source of folates and contain higher amounts of the active component, 5MTHF [51]. By including these functional foods in the diet, it is possible to increase the intake of folate and other essential nutrients, helping to improve overall nutrition and address nutrient deficiencies.

7. Conclusions and Future Prospect

Over the years, much attention has been given to the significant impact of micronutrient deficiencies on human health, leading to the development of new technologies to address these problems. Folate, an essential micronutrient, has been well studied in the health sector. However, research on folate in soybean has been limited. Recent studies on the optimisation of folate extraction and quantification have provided new technologies for accurate and sensitive profiling of folate using methods such as HPLC-MS or MS/MS. However, the lability of folate forms can pose challenges during extraction, requiring the optimisation of parameters such as pH, heat and enzyme treatments. Using reference materials and stable-isotope-labelled standards may also improve extraction efficiency and serve as quality controls.

Soybean seeds and soy-based products contain relatively high amounts of folate compared to most staple crops, and the variation in the folate content among soybean plants suggests that breeding and biofortification of folate in soybean is a feasible strategy. In the short term, the consumption of soy-based products can provide necessary levels of folate, while breeding for increased folate content in soybean is a viable strategy for a sustainable future. Although QTL mapping and metabolic engineering of folate content in soybean has been less studied, the use of omics tools and next-generation genotyping technologies can help to better understand the genetics of folate in soybean and to identify candidate genes for folate biofortification. In conclusion, soybean is a rich source of folate and other phytonutrients, and increasing research focus on soybean folate will be crucial in the fight against malnutrition.