Soybean Molecular Breeding Through Genome Editing Tools: Recent Advances and Future Perspectives

Abstract

Soybean (Glycine max L.) is an essential crop for global food, feed, and industrial applications, but its production is increasingly challenged by climate change and environmental stresses. Traditional breeding and transgenic approaches have contributed to improvements in yield and quality; however, limitations in genetic diversity and regulatory hurdles for genetically modified organisms (GMOs) underscore the need for innovative strategies to address these challenges. Genome editing technologies, particularly CRISPR/Cas9, have revolutionized soybean molecular breeding by enabling precise modifications of genes related to key agronomic traits such as yield, seed composition, and stress tolerance. These advances have accelerated the development of soybean varieties with enhanced nutritional value and adaptability. Recent progress includes improvements in editing efficiency, specificity, and the ability to target multiple genes simultaneously. However, the application of genome editing remains concentrated in a few model cultivars, and challenges persist in optimizing transformation protocols, minimizing off-target effects, and validating edited traits under field conditions. Future directions involve expanding the genetic base, integrating genome editing with synthetic biology, and addressing regulatory and public acceptance issues. Overall, genome editing offers significant potential for sustainable soybean improvement, supporting food security and agricultural resilience in the face of global challenges.

1. Introduction

Legumes are an essential source of dietary protein for vegetarians and play a crucial role in fulfilling global protein demands. Among legumes, soybean (Glycine max L. Merrill.) is a globally cultivated and nutritionally important staple crop. It is grown under a wide range of conditions, using diverse management strategies to satisfy varying end-user needs [1]. Soybean seed production, along with key by-products such as meal and oil, played a major role in driving a nearly 350% increase in crop output in 1987 [1]. Dry soybean seeds contain approximately 40% protein and 20% oil, making them a vital source of nutrition for humans, forage for livestock and fish, as well as raw material for biofuel production [2,3,4]. Soy meals are closely linked to food supply, serving directly as a protein source for humans and indirectly as a major component of livestock feed. Soy oil is used in diverse applications such as food products, polymers, building materials, cosmetics, beverages, waxes, and fuels [1].

The cultivated soybean originated in eastern Asia (approximately 6000–9000 years ago), was farmed in China for a millennia, and is considered a domesticated form of the wild soybean (Glycine soja Sieb. & Zucc.) [1,5,6]. Although China and other Asian countries are no longer the leading producers, they continue to consume large quantities of both traditional and new soy products. In 2018, China was the largest importer of whole soybeans from the United States, with imports valued at over $3 billion [1]. Soybeans rank as the ninth most widely produced crop globally, with the United States, Brazil, Argentina, China, and India among the leading exporters [7]. In 2018–19, the United States produced 120.52 million metric tons of soybeans, making it the world’s largest producer at that time. In 2020–21, Brazil surpassed the United States to become the world’s leading soybean producer, with a production of 126 million metric tons as of May 2020 [7]. Soybeans were harvested from an average of 39.6 million hectares (98 million acres) in the US per year, with average yields of 2.7 to 3.4 tons per hectare. Annual soybean production is valued at approximately $40 billion, based on an average market price of $330 per ton over the past 20 years [8].

The limited availability of genetic resources is a major obstacle to soybean improvement, particularly in coping with climate change, farmland scarcity, and increased abiotic and biotic stresses. These factors have imposed substantial limits on agricultural production [4,9]. In addition to flooding and drought, biotic stresses represent a serious threat, causing pest-related soybean yield losses of approximately 21.4% from 2010 to 2014 [10]. Given the high volatility in the soybean market and the extensive land dedicated to its cultivation, farmers are particularly concerned with minimizing yield losses in soybeans.

Conventional breeding has allowed for the development of improved soybean cultivars, leading to enhanced food security through higher yields, increased resistance to biotic and abiotic stresses, and improved nutrient content. However, its effectiveness is often limited by the narrow genetic diversity available, highlighting the need to develop new approaches to broaden genome variation. Moreover, breeders face growing challenges as they are expected to meet rising consumer demands while adapting to environmental changes.

Soybean is a diploid (2n = 40), self-fertilizing crop that evolved from an ancient tetraploid ancestor, resulting in a homologous gene-rich and highly complex genome. For instance, 48,378 genes have been identified in the Williams 82 cultivar, with up to 63,703 genes reported in the Jack cultivar [11,12]. The soybean genome duplicated approximately 59 and 13 million years ago. These events were followed by subsequent gene duplications, diploidizations, and chromosomal rearrangements, which have contributed to its large genome size and extensive gene redundancy [13]. Significantly, most homologous genes are located on two chromosomes (61.4%), but many are also found on three (5.63%) or four chromosomes (21.53%) [13]. This high gene count and duplicated genome structure are considerable challenges for both conventional and molecular breeding.

In the early 1980s, genetically modified (GM) technology, which enabled the insertion of foreign DNA into plant genomes, marked a major milestone and paved the way for modern plant genome editing. Despite their significant success, GM methods continue to face challenges in public acceptance and scientific approval, mainly due to concerns about transgene outcrossing and other biosafety issues [14,15,16]. The advancement of non-GM crops, which exclude foreign DNA from their genome, is essential to overcome the limitations associated with foreign gene insertion, made possible by recent progress in targeted genome editing [16,17].

Genome editing (GE) has provided powerful tools for the precise design of crops with improved traits by incorporating elite alleles to enhance commercial breeding programs [18,19]. This approach has demonstrated great value in both basic and applied research across various crops due to its efficiency, precision, and rapid gene modification capabilities [20,21,22]. First-generation genome editing tools, such as zinc-finger nucleases (ZFNs) and transcription activator-like endonucleases (TALENS), which use Fok-I endonuclease to knock out the target gene by inducing double-strand breaks, have been largely replaced by the more efficient and simpler Clustered Regularly Interspaced Short Palindromic Repeats (CRISPR/Cas9) technology. This RNA-guided DNA editing system employs guide RNAs (gRNAs) complementary to the target genes, together with Cas9 endonucleases, delivered in a single construct to transfer, integrate, and precisely edit the host genome. A major advantage of gene-targeting systems like CRISPR/Cas9 is their capacity to introduce precise genetic changes without permanently incorporating foreign DNA into the plant genome. Following genome editing, traditional breeding methods can segregate out the CRISPR/Cas9 transgene, resulting in plants that carry only the intended genetic changes without any transgenic component [23]. Unlike transgenic crops, GE crops have better consumer acceptance and face fewer regulatory barriers, as demonstrated by the first GE crop released to market, the γ-aminobutyric acid (GABA)-enriched tomato [24].

This review highlights recent advances in soybean molecular breeding, utilizing genome editing technologies. Over the past decade, significant progress has been achieved in soybean genome editing, particularly using CRISPR/Cas9 systems. These developments have facilitated precise genetic modifications to improve soybean traits such as seed composition, yield, stress tolerance, and functional qualities. Versatile genome editing tools are essential to overcome current challenges and improve key agronomic traits in soybeans. These advancements will help increase yields to satisfy global market demands, even under unstable climate conditions.

2. CRISPR-Based Advances: Precise Molecular Breeding in Soybean

As the primary biotechnological tool for soybean improvement, the CRISPR/Cas9 system has been utilized for precise gene editing in soybean since successful crop applications were reported in Arabidopsis thaliana [25], Nicotiana tabacum [25,26], Oryza sativa [27], and Triticum aestivum [28]. In addition, the CRISPR/Cas12a (Cpf1) system has emerged as another powerful tool for plant genome editing, offering a simpler design and recognition of T-rich PAM sequences. It has been successfully applied to soybean [29] and O. sativa [30]. Here, we summarize the list of targeted trait improvements in soybeans through CRISPR-based tools over the past decade (

Table 1. List of the targeted trait improvement in soybeans through CRISPR tools.

Trait		Target Gene(s)/Loci	References
Seed contents	Oil composition	GmFAD2-1	[31,32,33,34,35,36,37,38]
		GmFAD2-2	[32,34,36,39]
		GmKASI	[40]
		GmPDCT	[41,42]
		GmFATA	[42]
		GmFAD3	[38]
		GmSFAR4	[43]
	Nutritional content (protein, amino acid, sugar)	GmKASI	[40]
		GmAIP2	[44]
		GmCG-1	[45]
		GmRAV1; GmWRKY27	[46]
		GmSWEET10	[47]
		GmTCP670	[48]
Seed yield and development	Seed size	GmDMEa	[49]
	Seed number	GmFT2a; GmFT5a	[50]
		GmJAG1	[51]
		EIL3; EIL4; EIN2L	[52]
		miR396	[53]
		GmAOC4H8	[54]
	Modified pigmentation	GmDCL2	[55]
	Wrinkled shape	GmKASI	[40]
Abiotic stresses	Drought	gma-miR398c	[56]
		GmPLA	[57]
		GmHdz4	[58]
		GmLCL	[59]
		GmNAC8 *	[60]
		GmNAC12 *	[61]
	Salt	GmAITR	[62]
		GmCG-1	[45]
		GmZAT10-1	[63]
		GsSOS1 *; GsNSCC *	[64]
		GmAKT1 *	[65]
	Heat	GmSPL2b *	[66]
		GmBSK1 *	[67]
	Multiple	GmArm	[68]
Biotic stresses and interactions	Disease resistance	GmArm	[68]
		GmF3H	[69]
		GmUGT	[70]
		GmTAP1	[71]
		GmMLO	[72]
		GmSNAP02	[73]
		GmTCP19L *	[74]
	Herbicide resistance	GmAHAS4	[75]
		GmALS1; GmALS3	[76]
	Nodulation	Rfg1	[77,78]
		GmRIC1; GmRIC2	[79]
		GmNMHC5	[80]
		GmNLP1; GmNLP4	[81]
		GmbZIP4	[82]
		GmWRKY17	[83]
Flowering time	Early flowering	E1	[84,85,86]
		EIL3; EIL4; EIN2L	[52]
		E1Lb	[86]
		GmNF-YC4	[87]
	Late flowering	GmFT2a	[88,89,90]
		GmFT5a	[89,91]
		GmFT4	[90]
		GmFT5b	[92]
Plant architecture	Growth enhancement	GmSPL9	[93]
		GmFT2a; GmFT5a	[91]
	Dwarf phenotype	GmLHY	[94]
		E1	[85]
		GmDWF1	[95]
		GmMET1	[96]
		GmDRM	[97]
	Pod morphology	GmPPD	[98]
Functional property	Digestibility enhancement	GmGOLS	[99]
		GmRS	[100]
		GmKTI	[101]
		GmKTI3	[102]
	Flavor optimization	GmLOX-2	[103]
		GmBADH	[104,105]
		GmBAS1; GmBAS2	[106]
	Nutritional improvement	GmIPK1	[107,108]
		GmGLY1	[109]
	Allergen elimination	GmLOX	[110]
		Gly m Bd 28K	[111]
		Gly m Bd 30K	[111,112,113]
		lincCG1	[114]
		GmP34	[102]

* Susceptibility.

).

2.1. Seed Contents, Development, and Yield

Soybeans contain important nutritional components, including seed oil with a complex fatty acid profile and high-quality plant proteins, which provide all nine essential amino acids that humans require. Thus, precise molecular breeding efforts have been made to improve oil composition by increasing oleic acid content or enhancing nutrition through increased protein content. As one of the five major crops in global agriculture and food supply, soybeans have also been the focus of efforts to increase yield and optimize seed development for cultivation in various climates.

2.1.1. Seed Contents

Soybeans contain various fatty acids, including polyunsaturated fatty acids (PUFAs), monounsaturated fatty acids (MUFAs), and saturated fatty acids (SFAs), as well as rich proteins. Extensive research was conducted to regulate their fatty acid content. The relatively high concentration of PUFAs, linoleic acid (C18:2), and α-linolenic acid (C18:3) heavily contributes to oxidative rancidity, flavor reversion, and the short shelf-life of soybean oil [115]. The key enzymes involved in regulating fatty acid content are the fatty acid desaturase (GmFAD) and fatty acyl–acyl carrier protein thioesterase A (GmFATA) [116,117]. GmFAD converts the MUFA oleic acid into the PUFA linoleic acid by introducing an additional double bond. Omega-3 fatty acid desaturase (GmFAD3) catalyzes the conversion of linoleic acid (C18:2) to α-linolenic acid (C18:3) in the PUFAs biosynthetic pathway during seed development. It mainly produces α-linolenic acid (7–10% of the seed oil composition) in soybeans [118]. GmFATA ends the fatty acid synthesis pathway by cleaving fatty acids from their acyl-ACP forms, such as C18 SFAs or MUFA-like oleic acid (C18:1-ACP), to release free fatty acids. Phosphatidylcholine:diacylglycerol cholinephosphotransferase (GmPDCT) catalyzes the interconversion between phosphatidylcholine (PC) and diacylglycerol (DAG) by transferring the acyl groups from DAG to PC, which are then desaturated or modified before being incorporated back into PC [119].

Extensive gene-editing studies have been carried out on GmFAD genes in various soybean cultivars, including Williams 82 [31,37], Jack [33], Jinong 18 [36], Jinong 38 [32,35,36,39], Maverick [32], and Wandou 28 [38]. CRISPR/Cas9-based editing of GmFAD2-1 has been performed [31,32,33,34,35,36,37,38]. Studies editing GmFAD2-2 [32,35,36,39] and GmFAD3 have also been reported [38]. These studies have demonstrated that targeted editing of GmFAD2 genes can significantly alter fatty acid profiles in soybean cultivars, leading to increased oleic acid (MUFA) and reduced linoleic acid (PUFAs) [31,32,33,34,35,36,37,38,39]. Furthermore, the modifications of GmFAD3 genes were found to be stable across generations, indicating their potential for use in soybean breeding programs [38].

In addition, an engineered variant of Streptococcus pyogenes Cas9 (SpCas9), known as SpRY, was developed to be nearly PAM-free, greatly expanding the target range and improving NGG site recognition [120]. Reflecting the benefits of PAM-free, SpRY was employed for GmFAD2-1A and GmFAD2-1B editing, resulting in high oleic acid levels in the GmFAD2-1A/1B double mutant (T2 line) [37]. CRISPR/Cas9-mediated knockout of GmPDCT; gmpdct1 gmpdct2 double mutants, unlike the previous editing mutants of GmFAD genes, showed a significant increase in triacylglycerol (TAG) containing MUFAs by 57.01% in m3-1-1 and 33.13% in m6-2-1. At the same time, PUFAs, particularly C18:2 and then C18:3, were markedly reduced [41]. CRISPR/Cas9 knockout of GmFATA1 and GmFATA2 reduced fatty acid content in leaves and seeds compared to the wild type, with seeds showing increased oleic acid and decreased linoleic and linolenic acid, altering oil composition without affecting plant growth [42].

The seed fatty acid reducers (SFARs) gene belongs to the GDSL lipases/esterases family and regulates fatty acid storage in Arabidopsis seeds. Overexpression of SFAR significantly decreased the fatty acid content in seeds and caused clear changes in fatty acid composition. Furthermore, these plants exhibited higher germination rates and were less affected by stress conditions compared to the wild type. Conversely, loss of SFAR function resulted in a notable increase in total seed fatty acid content, from 9.1% to 16.9% [121,122]. In Williams 82 soybeans, CRISPR/Cas9-generated gmsfar4a/b double mutants showed an 8% to 17% increase in total fatty acids under both greenhouse and field conditions. Both the wild-type and gmsfar4a/b mutants demonstrated a decrease in fatty acid content during germination. However, the gmsfar4a/b mutants had a slower reduction rate, indicating delayed triacylglycerol degradation due to the GmSFAR4 knockout [43]. Collectively, these studies successfully generated soybean lines with elevated oleic acid and reduced linoleic acid content, thereby improving seed oil composition.

Several studies have investigated the differential modification of nutritional components in soybeans. An initial editing study in the Bert cultivar revealed that knocking out β-ketoacyl–acyl carrier protein synthase 1 (GmKASI) increased sucrose content from 5.97% to 11.50% while reducing oil content from 18.76% to 5.63% compared to the wild type [40]. Subsequently, it was reported that the knockout of the ABI3-interacting protein 2 (GmAIP2) gene in the Jack cultivar resulted in an 8% increase in protein content [44]. CRISPR/Cas9-mediated knockdown of GmCG-1, which encodes β-conglycinin α′ subunit, along with its paralogues GmCG-2 and GmCG-3, reduced β-conglycinin, increased the 11S/7S globulin ratio, elevated total protein content, and enhanced sulfur-containing amino acids, alongside unexpected salt tolerance during germination and seedling stages [45].

To enhance soybean seed protein content, researchers used CRISPR/Cas9 to edit the GmNF-YC4-1 promoter, specifically targeting the binding motifs for the transcriptional repressors GmRAV1 (Related to ABI3/VP1) and GmWRKY27 [123,124,125]. This modification led to a two-fold increase in GmNF-YC4-1 expression in both leaves and seeds, resulting in a 6–11% increase in seed protein content compared to the wild type, demonstrating the effectiveness of precise promoter editing for nutritional improvement [46].

GmSWEET10a encodes a member of the Sugars Will Eventually be Exported Transporters (SWEET) family, which allocates sugars to the embryo and promotes fatty acid biosynthesis. GmSWEET10b has functional overlaps with GmSWEET10a and is considered a promising target for improving oil content and seed size. Depending on sequence variation in the C-terminal region of GmSWEET10, the protein structure can be classified as either the Higher Oil (HO) type, which increases oil and decreases protein content, or the Lower Oil (LO) type, which decreases oil and increases protein content. CRISPR/Cas9-mediated GmSWEET10a/b generated the indel variants that produced protein structures resembling either the HO type or LO type. In the GmSWEET10a-edited line, both oil and protein contents were significantly increased compared to the wild type, while in the GmSWEET10b-edited line, oil content increased and protein content decreased [47].

More recently, CRISPR/Cas9-mediated GmTCP670 with excised large fragments increased sulfur-containing amino acid (Met and Cys) 3.16% higher than the DN50 cultivar wild type [48]. These GmCGs and GmTCP670-edited lines are valuable genetic resources for improving the overall amino acid quality of soybean proteins. Thus, these precise gene editing studies demonstrate that various nutritional components, including fatty acids, protein, and sucrose, are the primary traits targeted for improvement in soybean molecular breeding.

2.1.2. Seed Yield and Development

Based on the Arabidopsis study, DEMETER (DME) encodes a DNA glycosylase with nuclear localization domains, primarily expressed in the central cell, the precursor to endosperm in the female gametophyte. AtDME is essential for reproductive gene imprinting; maternal DME mutations cause seed abortion [126,127]. In the soybean DN50 cultivar, a gmdmea mutant was generated through the C-terminal modification of the GmDMEa protein. The gmdmea mutant exhibited significant improvements in traits, including increased seed length, width, and thickness, as well as higher overall seed yield compared to the wild type. The enhancement resulted from enlarged cell sizes due to the GmDMEa knockout. Notably, despite these morphological improvements, seed composition remained unchanged, as evidenced by the lack of significant differences in oil and protein content compared to the wild type [49].

Flowering in plants is tightly regulated by FLOWERING LOCUS T (FT), which promotes the transition from vegetative to reproductive growth, and TERMINAL FLOWER 1 (TFL1), which acts antagonistically [128]. In soybeans, among ten FT homologs, GmFT2a and GmFT5a are key players in promoting flowering. When both GmFT2a and GmFT5a were knocked out via CRISPR tools, the double mutant gmft2a gmft5a exhibited enhanced vegetative growth. It developed new branches instead of flowers in the leaf axils under short-day conditions. The gmft2a gmft5a mutants led to a substantial increase in pod number, 54.7 ± 12.6 compared to 11.9 ± 3.2 of the wild type, and in seed number 106.3 ± 22.8, approximately 4-fold higher than the 27.7 ± 6.4 observed in the wild type [50]. This study demonstrated that suppressing these flowering regulators can dramatically increase seed yield by promoting branch and pod formation.

Soybean leaflet shape and seed yield are closely linked through the Ln locus, which encodes the transcription factor GmJAGGED 1 (GmJAG1), a homolog of Arabidopsis JAGGED that regulates lateral organ development and fruit patterning [129]. The recessive ln allele causes a transition from broad to narrow leaflets. It is associated with a significant increase in the number of seeds per pod, which could potentially improve yields [130]. CRISPR/Cas9-mediated knockout of GmJAG1 in Huachun 6 cultivar produced the gmjag mutant, which maintained normal growth period, plant height, number of branches, and number of nodes compared to the wild type [51]. Although the total number of pods remained unchanged, GmJAG soybean mutants developed 4-seeded pods, which are absent in the wild type, and the proportion of 3-seeded pods increased to more than half, while the proportion of 2-seeded pods decreased [51]. These changes resulted in a significant increase in total seeds per plant and an 8.67–8.81% rise in overall seed yield in both spring and summer field trials. This demonstrated the utility of ln-related traits for enhancing yield, especially in low-latitude cultivars [51].

Ethylene is a key phytohormone involved in growth, development, and stress responses, with its signaling pathway mediated by ETHYLENE INSENSITIVE (EIN) proteins such as EIN2, EIL3, and EIL4. Using multiplex CRISPR/Cas9 gene editing, simultaneous knockout of EIN2, EIL3, and EIL4 in soybean resulted in triple mutants with earlier flowering by about 7 days, accelerated podding and maturation, and improved pod set, particularly at the top of the plant [52]. These mutants exhibited a dramatic increase in reproductive output. The average number of pods per plant of the triple mutant rose to 133.65 (a 1.5-fold increase) compared to 81.18 in the wild type. The seed number per plant increased to 293.29 from 164.69, representing a 1.78-fold increase in the triple mutant. The overall seed yield per plant improved by 65%, highlighting the effectiveness of targeting ethylene signaling components to create high-yielding soybean germplasm through enhanced floral organ, pod, and seed development [52].

The miR396 family targets GROWTH-REGULATING FACTOR (GRF) transcription factors, with GRF1 regulating stem elongation [131]. In Arabidopsis, the miR396-GRF module influences both development and stress responses, with the overexpression of AtGRF1, AtGRF2, and AtGRF5 leading to increased seed size [132]. Soybean miRNAs show tissue-specific expression patterns across seed development, flowering, and meristem function [133]. CRISPR/Cas12SF01 editing of six miR396 genes in Zhonghuang 302 soybean produced larger-seeded mutants across two regions. Shandong-grown mutants showed seed dimension increases of 7.6% length, 6.4% width, and 3.8% thickness, while Hainan mutants exhibited more substantial improvements of 13.2% length, 10.7% width, and 6.3% thickness. Specific combinations (mir396acd, mir396adf, and mir396cdf in Shandong; mir396abcdf and mir396bcdfi in Hainan) demonstrated significant yield enhancements, confirming that targeted miR396 editing effectively improves soybean seed size and productivity [53].

The allene oxide cyclase 4 (GmAOC4) gene has been identified as a negative regulator of seed germination in soybeans, with expression analysis revealing higher levels in cultivars with poor germination rates. Previous research has shown that GmAOC4 overexpression in soybean hairy roots leads to jasmonic acid accumulation and reduces germination rates when expressed in Arabidopsis [134]. CRISPR/Cas9-generated knockout lines (gmaoc4-1 and gmaoc4-2) exhibited accelerated germination kinetics with significantly improved germination potential and index compared to wild-type plants. These mutants also showed slightly higher overall germination rates, providing direct genetic evidence that GmAOC4 functions as a suppressor of seed vigor and that its targeted modification can enhance germination performance without compromising other agronomic traits [54].

Seed morphology and composition are intrinsically linked through complex genetic networks, as demonstrated by several CRISPR/Cas9 studies targeting different aspects of seed development. GmDCL2a and GmDCL2b encode dicer-like enzymes responsible for 22-nucleotide siRNA production. When these genes were knocked out using CRISPR/Cas9, the reduction in siRNAs targeting chalcone synthase (CHS) resulted in elevated CHS mRNA levels in the seed coat, leading to a dramatic color shift from yellow to brown or black [55]. In a separate study, GmKASI knockout mutants developed wrinkled and shriveled seeds as a direct consequence of profound metabolic alterations, including a drastically reduced oil content (from 18.76% to 5.63%) and increased sucrose contents (from 5.97% to 11.50%) compared to wild-type plants. These findings collectively demonstrate that seed morphological traits such as shape, size, and color are not independent characteristics but rather manifestations of underlying metabolic and developmental processes that can be precisely manipulated through targeted gene editing [40].

2.2. Environmental Stresses: Abiotic Stress

Recent studies on soybeans have made significant progress in elucidating the molecular mechanisms underlying responses to environmental stresses, including both abiotic and biotic stresses, with a focus on advances in precise molecular breeding. Key gene editing studies are summarized, highlighting experimental outcomes and insights into the mechanisms of stress tolerance and disease resistance, along with their implications for breeding applications.

2.2.1. Drought Tolerance

CRISPR/Cas9 genome editing has revealed critical gene functions in soybean drought tolerance, with targeted knockouts of stress-responsive regulators demonstrating either enhanced resilience or heightened vulnerability.

In the Williams 82, editing of the gma-miR398c via CRISPR/Cas9 upregulated the expression levels of GmCSD1a/b, GmCSD2a/b/c, and GmCCS, resulting in higher O₂ content and reduced stomatal aperture under drought stress [56]. Similarly, CRISPR/Cas9 knockout of two homologous Phospholipase A (GmPLA) genes, GmpPLA-II ε and ζ, exhibited higher chlorophyll content under drought stress, with some knockout Tianlong No. 1 soybeans showing increased shoot and root fresh weights compared to the wild type [57]. CRISPR/Cas9-generated GmHdz4 knockout mutants in the soybean Tianlong No. 1 cultivar exhibited significantly enhanced drought tolerance compared to both the GmHdz4 overexpression lines and controls. The gmhdz4 soybean showed notably improved root architecture, including increased total root length, root surface area, and root tip number, leading to enhanced water uptake efficiency that contributes to drought resilience. The edited soybeans exhibited increased osmolyte accumulation and elevated antioxidant enzyme activities, including superoxide dismutase (SOD), catalase (CAT), and peroxidase (POD), which reduced reactive oxygen species (ROS)-induced cellular damage under drought stress [58].

CRISPR/Cas9-mediated knockout of the soybean circadian clock genes GmLCLa1, GmLCLa2, GmLCLb1, and GmLCLb2, orthologs of LHY-CCA1-LIKE, generated a quadruple mutant (gmlclq). Under 20 h dehydration treatment, gmlclq exhibited slower wilting, 15% less electrolyte leakage, and reduced water loss over time compared to the wild type. Furthermore, thermal imaging showed that gmlclq leaves maintained higher temperatures than the wild type under both early morning and noon conditions, with the temperature difference being especially pronounced at noon, indicating that the mutant experiences significantly less water loss during peak transpiration periods [59]. In contrast, CRISPR/Cas9-edited gmnac8 and gmnac12 soybean lines exhibited reduced SOD activity and proline content, as well as decreased drought survival rates, demonstrating that GmNAC8 and GmNAC12 contribute to drought tolerance in soybean [60,61].

2.2.2. Salt Tolerance

GmAITR knockout mutants generated using CRISPR/Cas9 exhibited significantly enhanced salt tolerance, demonstrating higher survival rates and improved growth under saline conditions without compromising growth in non-saline soil. These GmAITR-edited lines also showed lower levels of malondialdehyde (MDA) and hydrogen peroxide (H₂O₂) under salt stress, indicating reduced oxidative damage [62].

CRISPR/Cas9-mediated knockout of Gm7Sα′ (CG-1) and Gm7Sα (CG-2, CG-3) in the Jack cultivar not only increased total protein- and sulfur-containing amino acid content but also conferred enhanced salt tolerance. Under control conditions, the knockout line and wild type exhibited similar hypocotyl lengths; however, upon exposure to 150 mM NaCl, the knockout line showed a significant increase in hypocotyl length. Unlike the wild type, the knockout line did not display elevated MDA levels under salt stress and demonstrated significantly higher activities of antioxidant enzymes SOD, POD, and CAT. Additionally, the knockout plants had longer aboveground and underground parts, as well as a higher K⁺ ion content, than the wild type, collectively indicating improved physiological and biochemical responses that contribute to enhanced salt tolerance [45].

The Cys2His2 (C2H2)-type zinc-finger protein (ZFP) family in Arabidopsis comprises 176 members, 143 of which are plant-specific. These proteins play key roles in regulating plant development and stress responses through transcriptional control [135]. In the soybean cultivar Lee 68, CRISPR/Cas9-mediated knockout of the conserved C₂H₂-type zinc-finger protein GmZAT10–1 resulted in enhanced salt tolerance without affecting growth under normal conditions. Although KO-GmZAT10–1 and wild-type plants showed similar heights and fresh weights in the absence of stress, exposure to 120 mM NaCl for 7 days led to significantly greater growth in the knockout plants. Furthermore, these mutants exhibited higher relative water content and chlorophyll levels in their leaves, as well as reduced relative electrolyte leakage and MDA in both leaves and roots compared to the wild type. These results demonstrate that targeted mutagenesis of GmZAT10–1 can improve physiological and cellular resilience to salt stress in soybean [63].

By contrast, CRISPR/Cas9-mediated knockout of the GsSOS1 and GsNSCC genes in soybean resulted in disrupted Na⁺/K⁺ homeostasis under 200 mM NaCl treatment, leading to elevated Na⁺/K⁺ ratios and heightened salt sensitivity compared to wild-type plants [64]. Similarly, as one of the major osmoregulatory ions in plant cells, K⁺ is essential for enzyme activation, protein synthesis, and cell elongation, with the K⁺ transporter AKT1 playing a key role in uptake and salt stress tolerance [136,137]. Under normal K⁺ conditions (20 mM), the Arabidopsis akt1 mutant showed no phenotypic differences from the wild type or the complementary lines (akt1/GmAKT1 1 and akt1/GmAKT1 2). However, under low K⁺ (0.25 or 0.5 mM), the fresh weight of akt1 mutant decreased significantly while the complementary lines remained similar to the wild type. When exposed to 150 mM NaCl, GmAKT1 knockout lines exhibited increased ion leakage, reduced fresh weight, lower K⁺ and higher Na⁺ levels in both leaves and roots, and a Na⁺/K⁺ ratio more than two-folds higher than the wild type, highlighting the critical role of GmAKT1 in maintaining ion homeostasis and salt stress tolerance [65].

2.2.3. Heat Tolerance

CRISPR/Cas9-mediated knockout of GmSPL2b (Williams 82) in soybean cytoplasmic male sterility (CMS)-based restorer lines (N4608, YY6) produced frame-shift mutations that truncate the SPL functional domain, generating gmspl2b mutants (NJCMS1A × W82) with reduced male fertility under high-temperature (HT) stress; anther indehiscence appeared only after prolonged heat exposure, indicating GmSPL2b is one of several miR156b targets controlling fertility under heat stress [66,138]. Under HT conditions, the expression of heat-shock protein (HSP) genes GmHSP17.6 and GmHSP83 declined in gmspl2b mutants, suggesting that GmSPL2 plays a role in activating HSP pathways to maintain fertility [66]. Recent work has also highlighted a brassinosteroid-signaling kinase, GmBSK1, as a positive regulator of heat resilience in soybeans. CRISPR/Cas9-generated gmbsk1 mutants exhibited increased ROS accumulation and decreased ROS-scavenging enzyme activities under HT, whereas GmBSK1 overexpression upregulated GmBES1.5, which binds to the E-box motif in the promoters of heat-responsive genes to enhance ROS detoxification and stress responses [67].

2.2.4. Multiple Tolerance

As a multiple stress-responsive gene, GmARM contains several hormone-response and stress-regulatory elements, including an ABRE and a TCA element in the promoter region [68]. CRISPR/Cas9-based knockout of GmARM in the soybean DN50 cultivar conferred broad-spectrum stress tolerance. The gmarm mutants showed a higher survival rate under salt stress (30–43% vs. 3%) and alkali stress (43–74% vs. 3%) compared to the wild type. The edited soybeans also exhibited medium to high resistance against Phytophthora sojae, whereas the wild-type DN50 remained highly susceptible. Expression profiling in gmarm soybean revealed upregulated key stress-responsive genes (GmNHX1, GmAKT1, GmNCED1, GmPR10, and GmWRKY40), which coordinate ionic homeostasis, osmoprotectant production, and pathogen defense under combined stresses [68].

2.3. Environmental Stresses: Biotic Stresses and Interactions

2.3.1. Disease Resistance

CRISPR/Cas9-mediated multiplex editing of three flavonoid pathway genes, GmF3H1, GmF3H2, and GmFNSII-1, in soybeans, resulted in significant isoflavone accumulation and enhanced resistance against Soybean Mosaic Virus [69]. Similarly, GmUGT knockout plants accumulated non-glucosylated flavonoids and showed upregulated expression of both flavonoid biosynthesis and immune response genes, conferring improved resistance against the leaf-chewing insect Helicoverpa armigera [70]. In a separate study targeting susceptibility factors, researchers identified six soybean MLO genes (GmMLO2, GmMLO10, GmMLO18, GmMLO19, GmMLO20, and GmMLO23) with close homology to Arabidopsis AtMLO2, AtMLO6, and AtMLO12—known powdery mildew susceptibility genes. CRISPR/Cas9 editing of these six GmMLO genes achieved a substantial 19.1 to 40% reduction in Erysiphe diffusa conidial infection without compromising plant growth or biomass production [72].

Intrinsic resistance to soybean cyst nematode (SCN, Heterodera glycines Ichinohe) is mediated by functional changes in housekeeping genes, particularly α-SNAP genes GmSNAP18a and GmSNAP18b, where copy number variation at the Rhg1 locus (chromosome 18) determines resistance. The identification of a quantitative trait locus (QTL) for SCN resistance on chromosome 2 implicated GmSNAP02 as a candidate gene, which was subsequently targeted for CRISPR/Cas9-mediated loss-of-function mutations [73]. Upon SCN infection, GmSNAP02 expression significantly increased in susceptible cv. Peking, moderately increased in moderately resistant line 81-4, but showed no change in highly resistant PI 90763 and PI 437654. CRISPR-edited GmSNAP02 knockout lines (T4 + T3 and T5 + T7) in Peking exhibited significantly reduced SCN counts compared to wild type. Crucially, editing GmSNAP02 in the highly resistant PI 90763 caused no significant difference in SCN resistance compared to the wild type. GmSNAP02 disruption specifically enhances resistance in susceptible backgrounds but is redundant in naturally resistant genotypes [73].

Another study tested Phytophthora sojae resistance in GmARM knockout lines (QF01-06) [68]. Susceptible controls developed severe symptoms, including stem decomposition, discoloration, and eventual plant death. In contrast, resistant mutants exhibited only localized discoloration at inoculation sites with normal growth. Five of the six knockout lines (excluding QF02) showed significantly enhanced resistance, achieving 40–60% survival post-inoculation, compared to complete mortality in the controls, validating GmARM as a promising target for resistance engineering [68]. Editing GmTCP19L in Williams 82 soybean increased susceptibility to P. sojae [74], whereas GmTAP1 editing in the same background significantly reduced both lesion size and pathogen biomass. Notably, gmtap1 mutants showed no differential expression in PTI markers (GmPR1, GmACS2, and GmERF) following XEG1/flg22 elicitation, indicating that defense functions occur through alternative pathways. These mutants maintained normal growth and yield [71], demonstrating disease resistance without agronomic penalties.

2.3.2. Herbicide Resistance

Acetolactate synthase (ALS) catalyzes the first step in branched-chain amino acid biosynthesis, making it the primary target of many commercial herbicides. To confer herbicide resistance, GmALS1 and GmALS3 in soybean were simultaneously edited using a CRISPR/Cas12-SF01 system, generating a gmals1/gmals3 double mutant. After seven days of standard ALS-inhibitor treatment, wild-type plants exhibited severe damage, whereas the gmals1/gmals3 mutant remained phenotypically healthy. Moreover, herbicide treatment did not compromise the mutant’s seed yield, plant height, node number, or grain weight relative to the untreated wild type [75]. Similarly, acetohydroxyacid synthase (AHAS), another key enzyme in branched-chain amino acid biosynthesis, is also inhibited by numerous herbicides. A GmAHAS4 mutant created via CRISPR/Cas9 displayed full tolerance to a panel of AHAS-targeting herbicides, bispyribac-sodium, imazapyr, chlorsulfuron, flucarbazone-sodium, and flumetsulam, without any observable phenotypic deviation from untreated wild-type soybean plants. Under 1 mg/L chlorsulfuron treatment, wild-type soybeans failed to survive, whereas GmAHAS4 mutants showed normal growth; importantly, seed number and grain weight remained unchanged in the mutant compared to the wild type [76]. By maintaining yield and architecture under herbicide stress, these alleles offer valuable genetic resources for breeding next-generation, herbicide-tolerant soybean cultivars optimized for high-input agricultural systems.

2.3.3. Nodulation

Nodulation is a highly specific interaction in legumes whereby compatible rhizobia infect root cortical cells, inducing dedifferentiation and the formation of nodule primordia. Once endocytosed, the bacteria convert atmospheric N₂ into ammonium within symbiosomes, supporting the host’s nitrogen nutrition [139]. The Rfg1 gene confers nodulation resistance in the Williams 82 and PI 377578 cultivars. CRISPR/Cas9-mediated editing of Rfg1 enabled the Sinorhizobium fredii USDA193 strain, which was previously unable to form root nodules, to initiate nodule formation successfully [77,78]. CRISPR/Cas9 editing in soybean revealed key regulators of nodule quantity: knockout of the paralogous GmRIC1 and GmRIC2 (CLAVATA3/EMBRYO SURROUNDING REGION-RELATED, CLE peptide genes) generated hypernodulation mutants forming 381 and 348 nodules per plant (vs. 228 in the wild type), confirming their role as negative regulators [79]. Conversely, simultaneous disruption of the three GmRdn1-1/1-2/1-3 genes (homologs of Medicago truncatula RDN) produced mutants with only 91–100 nodules (vs. >300 in the wild type), demonstrating their positive role in nodulation promotion [79]. These results revealed the core nodulation mechanism underlying nodule number regulation via CRISPR/Cas9 editing. In the Jack cultivar, the nmhC5 gene, a MADS-box family member orthologous to root-expressed Arabidopsis AGL17 subfamily, was shown to promote lateral root and nodule growth in a soybean root transformation system. Compared to the wild type, the Gmnmhc5 mutant showed reduced nodule numbers due to the induction of GmGAI, a repressor of gibberellin signaling [80].

CRISPR editing also enabled optimization of soybean symbiosis under environmental stress. Disruption of the nitrate-responsive GmNLP1 and GmNLP4 genes in the Huachun 6 cultivar preserved nodulation under high soil nitrate levels, indicating enhanced nitrogen tolerance [81]. Cultivar-specific signaling pathways were identified through knockout studies: disruption of cytokinin-related GmbZIP4a/b (Williams 82 background) reduced nodules to one-third of the wild type, while loss of GmWRKY17 in hairy-root assays reduced nodulation to one-tenth [82,83]. These advances demonstrate the power of CRISPR/Cas9 to optimize nitrogen fixation, enhance stress resilience, and tailor symbiosis for diverse agroecosystems.

2.4. Flowering Time

Soybean flowering time, regulated by photoperiod and temperature, has a significant influence on yield, regional adaptation, and risk of stress, such as frost. Traditional breeding for flowering adaptation is slow due to the complex genetic network, including major genes such as E1–E4 and FT homologs. Since photoperiod sensitivity restricts soybean to certain latitudes, developing cultivars with altered or insensitive flowering responses is essential for expanding cultivation, improving yield stability, and reducing climate vulnerability. Flowering is governed by intricate photoperiod and temperature pathways, with transcription factors like FT (activator), TFL1 (suppressor), LEAFY (LFY), and SOC1 playing central roles [140].

2.4.1. Early Flowering

In soybean, E1 functions as a photoperiod-responsive repressor of GmFT2a/5a under long-day conditions, with recessive alleles promoting earlier flowering [141]. CRISPR/Cas9-mediated knockout of E1 significantly accelerated flowering across different cultivars, although the magnitude of this effect varied among them [84,85,86]. In the Jack cultivar, E1 knockout lines (L7, L9, and L16) flowered at 39, 38, and 37 days, respectively, compared to 57 days in wild-type plants under long days, with no difference under short days [84]. Tianlong No. 1 E1 mutants exhibited reduced photoperiod sensitivity, reaching the R8 stage 20 days earlier and flowering at double the speed under long-day conditions [85]. In the 06KG218440 cultivar, CRISPR/LbCas12a-generated E1/E1lb knockouts further demonstrated variable acceleration: wild-type plants flowered at 57.2 ± 1.2 days, E1 substitution mutants (SYKE00D, E1-Sub/E1Lb-WT) at 48.2 ± 1.0 days, and E1 loss-of-function mutants (SYKE00A, E1-Lof/E1Lb-WT) at 26.6 ± 1.4 days. E1lb mutations had weaker effects, with in-frame deletions producing weak alleles that enabled fine-tuning of flowering time by 3–9 days, depending on the allele combinations [86]. These findings demonstrate that targeted editing of E1 and E1lb enables precise control of soybean flowering time for breeding adaptation.

Ethylene pathway components EIN2L, EIL3, and EIL4 have a significant influence on flowering and development. CRISPR/Cas9 triple mutants (Z4) in Williams 82 showed comprehensive acceleration: the early flowering (R1) stage was 7 days earlier, the early podding (R3) stage 6 days earlier, the full pod (R4) stage 8 days earlier, and the maturation (R8) stage 7 days earlier than the wild type. This timeline acceleration resulted in a 65% increase in seed production, demonstrating the potential of the ethylene pathway for controlling flowering and enhancing yield [52].

Beyond ethylene manipulation, researchers developed precise acceleration and delay methods through targeted editing. GmNF-YC4 suppresses flowering by inhibiting GmFT2a and GmFT5a under long-day conditions. CRISPR/Cas9 gmnfyc4 mutants exhibited photoperiod-specific responses, with short-day flowering times (22.1 ± 1.3, 21.9 ± 0.9 days) matching those of the wild type (22.2 ± 0.8 days). In contrast, long-day flowering (33.2 ± 1.4, 33.4 ± 1.5 days) occurred earlier than in the wild type (39.8 ± 0.9 days), confirming the photoperiod-sensitive repressor function [87].

2.4.2. Late Flowering

Flowering activator knockouts consistently delayed flowering across cultivars. Jack cultivar GmFT2a knockout mutants flowered seven days later than controls, failing to initiate flowering until the wild type reached pod development [88]. Large-scale GmFT2a deletion (>1600 bp) generated pronounced delays: 31.1 ± 0.8, 30.6 ± 1.2 days short-day versus 25.1 ± 1.2 wild type, and 55.9 ± 1.7, 54.4 ± 1.6 days long-day versus 50.7 ± 1.8 controls [89]. Base editing using Cas9n(D10A) nickase produced milder delays (34.3 ± 1.7 short-day, 44.9 ± 1.8 long-day) compared to complete knockout (36.6 ± 1.3 short-day, 47.5 ± 2.6 long-day), demonstrating fine-tuning through partial function preservation [90].

Tianlong No. 1 soybean studies revealed cultivar-specific and additive effects of FT knockout. Both gmft2a and gmft5a mutants showed severe delays exceeding 50 days under long-day conditions, while gmft2a gmft5a double mutants exhibited 2.5-fold longer delays [91]. GmFT5b knockout in Jack cultivar produced modest delays: no short-day difference (26.3–26.1 vs. 26.6 days) but 1.8–2.1 day natural long-day delays (29.8–30.1 vs. 28.0 days) and 3.0–3.2 day controlled long-day delays (47.8–48.0 vs. 44.8 days) [92]. These studies demonstrated that CRISPR/Cas9 editing of FT homologs yields predictable late-flowering phenotypes with cultivar-dependent severity, enabling precise agricultural manipulation.

2.5. Plant Architecture

Plant architecture, defined as the three-dimensional organization of leaves, flowers, and branches, is a major determinant of plant type and crop yield [142]. Branching patterns, originating from axillary bud development, play a central role in morphogenesis [143].

2.5.1. Growth Enhancement

The SQUAMOSA Promoter-Binding Protein-Like (SPL) gene family encodes transcription factors that regulate architecture, reproductive development, and responses to gibberellin. In Arabidopsis, miR156 post-transcriptionally regulates SPL genes, thereby influencing shoot branching [144,145]. In soybean, CRISPR/Cas9 editing of four sites in the GmSPL9 gene in the Williams 82 cultivar generated quadruple mutants (spl9a/spl9b-1/spl9c/spl9d) with significantly increased branch numbers, including new secondary branches. Triple mutants (spl9a^−/−b-1^−/−c^+/− and spl9a^−/−b-2^−/−c^+/−) also showed increased main stem nodes (16.3% and 7.7%), total node number (73.7% and 36.3%), branch number (72.5% and 57.8%), and dry weight (52.2% and 15.2%) compared to wild type, highlighting the crucial role of GmSPL9 in soybean plant architecture [93].

In the Tianlong No. 1 cultivar, late flowering was confirmed in the ft2a mutant, ft5a mutant, and ft2a ft5a double mutant. These mutants also caused changes in plant architecture. Under SD conditions, the plant height, number of nodes, and branches of the ft2a mutant were almost the same as those of the wild type. However, the ft5a and ft2a ft5a double mutants showed a significant increase. Under LD conditions, these differences were more pronounced, with the plant height, number of nodes, and branches of the ft2a, ft5a, and ft2a ft5a double mutants significantly increased compared to the wild type. The length and width of leaves were also considerably increased. As a result, the fresh aerial biomass increased by 162.3% in the ft2a ft5a double mutant and by 59.9% in the ft2a or ft5a single mutants compared to the wild type [91].

2.5.2. Dwarf Phenotype

LATE ELONGATED HYPOCOTYL (LHY) and its functional homolog CIRCADIAN CLOCK ASSOCIATED 1 (CCA1) form the central circadian oscillator [146]. Mutations in AtLHY suppress circadian gene expression and leaf movement, resulting in photoperiod-independent flowering, which demonstrates that constitutive LHY expression disrupts the plant’s circadian clock [147].

In the Harosoy cultivar, CRISPR/Cas9-mediated loss-of-function GmLHY quadruple mutants (Gmlhy1a1b2a2b) exhibited a 30% reduction in plant height under LD conditions, with no significant difference in node number but with markedly shorter internodes compared to the wild type. The reduced height was most evident 20–35 days after emergence. The application of exogenous GA3 restored the mutant’s height to wild-type levels, and endogenous GA3 levels were lower in the mutant, indicating a disruption in the GA3 biosynthesis pathway [94].

In the Tianlong No. 1 cultivar, CRISPR/Cas9-mediated E1 knockout mutants exhibited reduced photoperiod sensitivity, with a plant height ~60 cm shorter under SD versus LD conditions. Two independent E1 mutants exhibited a similar height reduction of ~25 cm under SD conditions compared to LD conditions, confirming E1’s role in photoperiod response and growth regulation. In the Tianlong No. 2 cultivar, CRISPR-edited GmDWF1 single mutants (dwf1a-1, dwf1b-1) and the double mutant (dwf1a-1 dwf1b-1) all displayed dwarfism at 21 days post-growth, with similar height reductions in single mutants. The double mutant exhibited an additional 15% height reduction compared to the single mutants, demonstrating the additive effects of GmDWF1 loss-of-function [95].

DNA methylation—an essential epigenetic mechanism regulating gene expression, development, and genome stability in plants—operates through distinct pathways: CG methylation (mCG) is maintained by DNA METHYLTRANSFERASE 1 (MET1), while de novo methylation across CG, CHG, and CHH contexts is driven by domains rearranged methyltransferases (DRMs) via the RNA-directed DNA methylation (RdDM) pathway [96]. In the Williams 82 cultivar, CRISPR/Cas9-edited GmMET1 mutants revealed distinct phenotypes: GmMET1a^+/− GmMET1b^+/− showed reduced plant height; GmMET1a^−/− GmMET1b^+/+ had shorter seeds; GmMET1a^+/+ GmMET1b^−/− exhibited reduced stem diameter and seed weight. Strikingly, GmMET1a^+/− GmMET1b^−/− mutants displayed reduced height, smaller leaves, fewer branches, early flowering, abnormal pods, and lower seed yield, indicating that GmMET1b plays a dominant role in development [96]. Additionally, GmDRM genes display tissue-specific expression (GmDRM2a/b in flowers/pods; GmDRM2c/3a/3b in seeds), and their knockout causes developmental defects like dwarfism and abnormal seeds. While a Gmdrm2c^−/−3a^−/−3b^−/− triple mutant showed no methylation loss, partial quintuple mutants (Gmdrm2a^−/−2b^−/−2c^−/−3a^−/−3b^+/− and Gmdrm2b^−/−2c^−/−3a^−/−3b^−/−2a^+/−) reduced mCG (3.6%), mCHG (5.7%), and mCHH (16%) methylation. Complete quintuple knockout decreased methylation by 9.6% (mCG), 12.1% (mCHG), and 10.5% (mCHH), confirming that simultaneous DRM2/3 inactivation disrupts methylation essential for development [97].

2.5.3. Pod Morphology

In Arabidopsis, mutations in the PEAPOD (PPD) genes result in plants with enlarged, dome-shaped leaves [148]. In the Kariyutaka cultivar, CRISPR-generated ppd1 mutants increased leaf size by 50% and seed weight by 21% (though with fewer seeds, reducing total yield), while ppd2 increased leaf size by 28%. The ppd knockout line developed twisted pods and severe seed loss, confirming the pleiotropic role of GmPPD in leaf, seed, and pod development [98].

2.6. Functional Properties

Recent advances in soybean breeding have shifted the focus from traditional traits, such as oil and protein content, to consumer-oriented qualities, including improved digestibility, flavor, nutrition, and safety (reduced allergens). With CRISPR/Cas9 technology, the direct modification of genes related to palatability is now enabled.

2.6.1. Digestibility Enhancement

Raffinose family oligosaccharides (RFOs), which are indigestible by humans due to the lack of α-galactosidase, cause intestinal flatulence and reduce digestive efficiency [149]. Targeted knockout of the galactinol synthase (GOLS) gene in the DT26 cultivar led to significant reductions in RFOs: the gmgols1A single mutant and gmgols1A gmgols1B double mutant contained 45.14 mg/g DW (30.2% decrease) and 41.95 mg/g DW (34.1% decrease) RFOs, respectively, compared to 64.7 mg/g DW in the wild type, without affecting sucrose or fructose levels [99]. Similarly, in Williams 82, CRISPR/Cas9-mediated knockout of raffinose synthase genes RS2 and RS3 increased sucrose content in rs2 and rs2 rs3 mutants (10.2% and 8.7%, respectively, vs. 7.7% in the wild type), while reducing raffinose content to 20–30% of wild-type levels and decreasing stachyose, demonstrating the potential for improving soybean seed digestibility and consumer acceptance through precise genome editing [100].

Beyond raffinose, soybean digestion is hindered by antinutritional factors (ANFs)—secondary metabolites, including tannins, phytic acid, flatulence-causing oligosaccharides, and trypsin inhibitors (TIs) that protect plants from biotic stresses. TIs critically impair protein digestion by strongly inhibiting pancreatic enzymes, such as trypsin and chymotrypsin, which significantly reduces dietary protein absorption efficiency, even in the presence of abundant digestive enzymes [150]. CRISPR/Cas9 knockout of Kunitz trypsin inhibitor genes KTI1 and KTI3 in the Williams 82 cultivar reduced KTI expression to <20% of wild-type levels in both kti1 single mutants and kti1 kti3 double mutants, with trypsin inhibition activity (TIA) minimized in the double mutant [101]. Similarly, KTI3 knockout in the Bert cultivar substantially decreased KTI protein production [102]. These findings demonstrate that targeted editing of KTI genes effectively reduces trypsin inhibitor content and activity, addressing a major digestive limitation in soybean consumption [101,102].

2.6.2. Flavor Optimization

To address undesirable flavors caused by lipoxygenase (Lox)-generated volatile compounds in soybeans, CRISPR/Cas9 editing targeted Lox genes across cultivars. In Huachun 6, knockout of GmLOX1/2/3 yielded three mutant lines (GmLOX-25, -40, 60), with colorimetric assays confirming complete loss of GmLOX1/2 activity in all lines and GmLOX3 inactivation specifically in GmLOX-25 and GmLOX-60 (evidenced by no substrate color change versus wild type) [103]. Similarly, in the SL1074 cultivar, Lox2 knockout reduced isozyme activity across five mutant lines from 325.74 ± 29.93 to 329.91 ± 26.00 units/g, representing a 25.93% decrease from wild-type levels of 439.81 ± 37.90 units/g. These results demonstrate the effective reduction in Lox activity or improvement in flavor through targeted gene editing [110].

2-Acetyl-1-pyrroline (2AP), identified in the 1980s as the primary aroma compound in aromatic rice, correlates with low BADH2 activity [151]. CRISPR editing of GmBADH genes enhanced soybean aroma: In Tianlong No. 1 (Cas9-edited), gmbadh2 mutants showed significantly elevated seed 2AP [104]. In Xudou 20 (Cas12i3-edited), gmbadh2 mutants increased 2AP, while gmbadh1 mutants showed no change; critically, gmbadh1 gmbadh2 double mutants exhibited four-fold higher 2AP in both leaves and seeds than gmbadh2 alone [105]. This confirms dual GmBADH1/2 knockout as an optimal strategy for flavor enhancement [104,105].

Soyasaponins impart a bitter aftertaste in soy foods, and CRISPR/Cas9 knockout of ß-amyrin synthase genes (GmBAS1/2) reduced saponin content. Specifically, only the GmBAS2 mutant exhibited significantly reduced levels of soyasapogenol A and B in seeds compared to the wild type. Conversely, stems and leaves of GmBAS2 mutants exhibited increased soyasapogenol A and B content, indicating altered transport to seeds [106].

2.6.3. Nutritional Improvement

GmIPK1 encodes inositol 1,3,4,5,6-pentakisphosphate 2-kinase, which converts IP5 to phytic acid (PA), the major phosphorus storage form in soybeans that also reduces mineral bioavailability. CRISPR/Cas9 editing of GmIPK1 in Huachun 6 and Kwangan cultivars significantly reduced PA content in knockout lines compared to the wild type, with similar reductions observed in GmMRP5 knockouts in Huachun 6 [107,108].

CRISPR/Cas9 knockout of GmGLY1 (encoding a P450 enzyme in glycitein; O-methylated isoflavone, biosynthesis) significantly increased daidzein content (DAC) in soybean mutants, 1.80-fold in KO-5 and 1.60-fold in KO-27 versus wild type. At the same time, genistin and total isoflavone levels remained unchanged. Conversely, glycitein-related compounds decreased markedly: 6″-O-malonylglycitin (11.49 to 13.29-fold lower), glycitin (2.37 to 4.13-fold lower), and glycitein (3.96 to 6.27-fold lower); 6″-O-acetylglycitin was unaffected. These results demonstrate that GmGLY1 disruption enhances daidzein, a compound with anticancer, antioxidant, and anti-inflammatory properties [152].

2.6.4. Allergen Elimination

Soybeans contain major allergens Gly m Bd 28K and Gly m Bd 30K, which have been shown to trigger IgE binding in 23% and 66.5% of soybean-allergic patients, respectively, as determined in a group of 69 individuals with atopic dermatitis and confirmed soybean sensitivity [153]. CRISPR/Cas9 editing of these genes in Enrei and Kariyutaka cultivars eliminated Gly m Bd 30K in single mutants and both proteins in double mutants, as shown by immunoblot, with RT-PCR confirming reduced gene expression [111]. Similar results were observed for Gly m Bd 30K editing in Yukihomare and Jack cultivars [112,113].

Soybean harbors 6018 long intergenic noncoding RNA (lincRNA) loci, with 23 linked to agronomic traits [154]. CRISPR/Cas9 knockout of lincCG1 in the DN50 cultivar downregulated six β-conglycinin genes (a major allergen) and upregulated six glycinin genes, inhibiting β-conglycinin synthesis [114]. Separately, triple knockout of lectin (LE), trypsin inhibitor (KTi3), and allergen-inducing P34 in the Bert cultivar reduced P34 protein levels, decreasing allergenicity [102]. These findings demonstrate that gene editing is being successfully applied across various soybean cultivars, such as DN50 and Bert, to improve functional properties, including reduced allergenicity and enhanced nutritional quality.

Overall, these results demonstrate that gene editing is being applied to a variety of genes across different cultivars. It is used to achieve better functional properties, such as enhanced digestibility, improved flavor, and reduced allergens.

3. Future Perspectives: Expanding the Frontiers of Soybean Genome Editing for Sustainable Agriculture

Recent advances in CRISPR-based genome editing have unlocked unprecedented opportunities for targeted trait improvement in soybean, a critical crop for global food security (Figure 1). The integration of next-generation editing tools and multifunctional enzyme engineering contributes to addressing longstanding challenges in soybean productivity, climate resilience, and nutritional quality. Here, we describe emerging strategies for soybean transformation to extend the applicability of cultivars and outline future directions for CRISPR applications in soybean breeding, drawing insights from foundational studies on soybean biology and climate adaptation.

3.1. Genome-Editable Soybean Cultivars: Limited Genetic Resources

Globally, there are more than 2500 officially registered or commercial soybean cultivars, reflecting the extensive genetic diversity available for breeding and research [155,156]. Despite this, actual cultivation in each country is typically concentrated among 10 to 30 leading commercial cultivars, which dominate the majority of planted acreage due to their agronomic performance and market preference.

Over the past decade, the application of CRISPR-based genome editing tools in soybeans has focused primarily on a limited set of cultivars (

Table 2. A list of the genome-edited soybean cultivars in reported studies.

Soybean Cultivar *	Target Gene(s)/Loci	Editing Tool	References
Bert	GmKASI	CRISPR/Cas9	[40]
	GmLE; GmKTI3; GmP34		[102]
Daewon	GmFAD2-1A, GmFAD2-1B	CRISPR/LbCpf1	[34]
DongNong 50 (DN50)	GmPDCT1; GmPDCT2	CRISPR/Cas9	[41]
	GmDMEa		[49]
	GmArm		[68]
	lincCG1		[114]
	GmTCP670		[48]
	GmAKT1		[65]
DT26	GmGOLS	CRISPR/Cas9	[99]
	GmMLO		[72]
Enrei	Gly m Bd 28K; Gly m Bd 30K	CRISPR/Cas9	[111]
Fukuyutaka	Gly m Bd 30K loci	CRISPR/Cas9	[113]
GL3510	GmBAS1; GmBAS2	CRISPR/Cas9	[106]
Harosoy	GmLHY	CRISPR/Cas9	[94]
	GmLCL		[59]
Huachun 6	GmRIC	CRISPR/Cas9	[79]
	GmLOX		[110]
	GmJAG1		[51]
	GmIPK1; GmMRP5		[108]
	GmNLP1; GmNLP4		[81]
	GmSWEET10		[47]
Jack	GmFT2a	CRISPR/Cas9	[50,88,89]
		Cas9n (D10A)	[90]
	GmFT5a	CRISPR/Cas9	[88,89]
	E1		[84]
	GmF3H		[69]
	GmFT4		[90]
	Gly m Bd 30K		[112]
	GmFAD2-1a; GmFAD2-1b		[33]
	GmAIP2		[44]
	GmNMHC5		[80]
	GmCG-1		[45]
	GmNF-YC4		[87]
	GmFT5b		[92]
	GmGLY1		[109]
Jinong 18	GmFAD2	CRISPR/Cas9	[36]
Jinong 38	GmFAD2	CRISPR/Cas9	[32,35,36,38]
Kariyutaka	Gly m Bd 28K; Gly m Bd 30K	CRISPR/Cas9	[111]
	GmPPD		[98]
Kefeng No.1	GmAOC4H8	CRISPR/Cas9	[54]
Kwangan	GmFAD2	CRISPR/LbCpf1	[34]
	GmIPK1	CRISPR/Cas9	[108]
Lee 68	GmZAT10-1	CRISPR/Cas9	[63]
Maverick	GmFAD2	CRISPR/Cas9	[32]
	GmGOLS1		[99]
Peking	GmSNAP02	CRISPR/Cas9	[73]
PI 377578	Rfg1	CRISPR/Cas9	[78]
SL1074	GmLOX-2	CRISPR/Cas9	[103]
Tianlong No. 1	GmDCL2	CRISPR/Cas9	[55]
	GmNAC8		[60]
	GmPLA		[57]
	GmHdz4		[58]
	GmNAC12		[61]
	GmUGT		[70]
	E1		[85]
	GmBADH2		[104]
	GmAHAS4	Cytosine base editor (CBE3)	[75]
	GmFT2a; GmFT5a	CRISPR/Cas9	[91]
Tianlong No. 2	GmDWF1	CRISPR/Cas9	[95]
Wandou 28	GmFAD2; GmFAD3	CRISPR/Cas9	[38]
Williams 82 (Reference genome)	GmSFAR4	CRISPR/Cas9	[122]
	Rfg1		[77]
	GmFAD2		[31]
		CRISPR/LbCpf1	[34]
		CRISPR/SpRY	[37]
	GmSPL9	CRISPR/Cas9	[93]
	gma-miR398c		[56]
	GmAITR		[62]
	Gly m Bd 30K		[112]
	GmTCP19L		[74]
	GmRS2		[100]
	EIN2L; EIL3; EIL4		[52]
	GmSPL2b		[66]
	GmTAP1		[71]
	GmKTI1; GmKTI3		[101]
	GmFATA		[42]
	GmRAV1		[46]
	GmbZIP4		[82]
	GmBSK1		[67]
	GmWRKY17		[83]
	GmMET1		[96]
	GmDRM		[97]
Xudou 18	GmALS1; GmALS3	Cas12-SF01-based cytosine base editor	[76]
Xudou 20	GmBADH1; GmBADH2	CRISPR/Cas12i3	[105]
Yukihomare	Gly m Bd 30K	CRISPR/Cas9	[113]
Zhonghuang 302	miR396	CRISPR/Cas12SF01	[53]
ZYD1239	GsSOS1; GsNSCC	CRISPR/Cas9	[64]
06KG218440	E1; E1Lb	CRISPR/LbCas12a	[86]

* Soybean cultivars are in alphabetical order.

). Among these, the reference genome cultivar Williams 82 stands out as the most frequently used for genome editing studies, mainly due to the comprehensive genomic information available and its amenability to transformation and genetic manipulation. We provide a summary of soybean cultivars that have been successfully edited using CRISPR technologies from 2015 to 2025 (Table 2). This overview highlights the frequent use of Williams 82, Jack, and Tianlong No. 1 as the most popular soybean cultivars for genome editing. It also emphasizes the need to expand genome-editable resources to a broader range of commercial soybean cultivars.

3.2. Major Traits of Soybean Editing for the Commercial Market

On 15 September 2021, the world’s first GE crop approved using CRISPR/Cas9 technology was the high-GABA tomato, which was commercialized in Japan. Although significant progress has been made in gene editing of soybeans with a growing number of traits targeted for improvement, commercial adoption remains limited. As of June 2025, only a handful of gene-edited soybean lines have received regulatory clearance, and none have reached the market except for Plenish high-oleic soybeans, which were developed using gene silencing rather than CRISPR-based technologies. The current landscape, as reflected in the USDA’s Am I Regulated (AIR) database, shows that most genome editing efforts have focused on traits such as altered oil composition, improved stress tolerance, disease resistance, and enhanced nutritional profiles. Here is a summary of the list of AIR entities with GE soybeans as of June 2025 (

Table 3. A summary of the AIR * genome-edited soybean.

USDA’s Response Date	Description of the Edited Trait	Institution
January 2014	Altered-flavonoid profile	University of Georgia
May 2015	FAD2KO	Cellectis Plant Sciences
May 2015	FAD3KO	Cellectis Plant Sciences
October 2017	Drought and salt tolerance	USDA ARS
June 2019	Changes in seed composition	University of Minnesota
June 2019	Changes in petiole length	University of Minnesota
March 2020	Resistance to soybean cyst nematode (SCN)	Evogene, Ltd.
May 2020	High oleic and low linolenic acid	Calyxt, Inc.
July 2020	High oleic acid	ToolGen, Inc.
August 2020	Increased carbon flux	University of Georgia
August 2020	Increased ketocarotenoids	University of Georgia
September 2020	Altered seed composition	University of Missouri
September 2020	Changes in leaf size and seed weight	University of Missouri
September 2020	Increased oil and protein content	Corteva Agriscience
March 2025	Herbicide resistance	Inari Agriculture, Inc.
April 2025	Improved architecture	Bayer Crop Science

* Am I Regulated (AIR) letters of Inquiry table from USDA (Last Modified: 5 June 2025).

). Among the 212 total GE applications, 11 institutes hold 16 AIR letters specifically concerning precise GE soybeans.

3.3. Editing Unexplored Traits for Next-Generation Soybean Resilience

Genome editing in soybeans has been successfully applied to various traits, particularly those related to seed composition and functional characteristics. However, several critical areas remain underexplored. Specifically, genome editing approaches to enhance photosynthetic efficiency, precisely to regulate rhizosphere microbiome interactions, and develop novel mechanisms of herbicide resistance represent significant research gaps, despite being key challenges for future soybean improvement.

3.3.1. Direct/Indirect Biological Regulation of Photosynthetic Efficiency

Photosynthesis is essential for seed contents and yield; the Calvin–Benson–Bassham (CBB) cycle has emerged as a prime target for enhancing photosynthetic capacity. Overexpression of cyanobacterial bifunctional fructose-1,6/sedoheptulose-1,7-bisphosphatase (FBP/SBPase) in soybean demonstrated robust improvements in RuBP regeneration under elevated CO₂ and high-temperature conditions, mitigating yield losses by up to 22% compared to wild-type plants [157]. Recent studies further highlight the potential of multiplexed editing of CBB cycle genes (e.g., GmSBPase, GmFBPA) to boost photosynthetic output while balancing ATP/NADPH stoichiometry synergistically [158].

Photosynthetic efficiency can be enhanced through indirect biological regulation rather than direct modulation of the Calvin–Benson–Bassham (CBB) cycle. A particularly notable example involves the regulatory relationship between nodulation and photosynthetic performance. CRISPR/Cas9-generated soybean gmric1a/2a-1 and gmric1a/2a-2 mutants displayed significantly increased nodule formation [159]. This enhanced nodulation resulted in increased nitrogen and carbon content in seeds, along with substantial improvements in key physiological parameters. Specifically, these mutants exhibited significantly higher percentages of nitrogen derived from the atmosphere, enhanced photosynthetic rates, and increased stomatal conductance compared to wild-type plants [159]. These findings demonstrate that nodules can effectively regulate photosynthetic processes, underscoring the critical importance of root-rhizosphere microbiome interactions in determining overall plant performance.

3.3.2. Root Exudate Engineering: Targeting Microbiome Composition

The interaction between root exudates and soil pore structure plays a crucial role in determining the composition and functional activity of the rhizosphere microbiome [160]. Currently, the strategic modulation of rhizosphere composition through targeted root exudate modification remains largely underexplored, and studies investigating the directed recruitment of specific beneficial rhizosphere microorganisms are limited in scope. Maize research has demonstrated that stalk rot-resistant and susceptible cultivars exhibit distinct rhizosphere microbiome profiles, with different root exudate components triggering specific variations in microbial community structure and functional gene abundance [161]. As evidenced by these root exudate–microbial interactions in maize, strategically engineered exudate compositions in CRISPR-edited soybeans could potentially induce the acquisition of enhanced disease resistance mechanisms. This approach represents a particularly promising strategy for developing disease resistance, offering substantial benefits in both agricultural productivity and commercial applications through the targeted manipulation of plant–microbiome interactions.

3.3.3. Engineering Detoxification Pathways: Herbicide/Metal Tolerance

Commercial herbicides are widely applied in crop cultivation, and most engineered herbicide resistance relies on modifying the herbicide’s target enzyme. However, this strategy often leads to the emergence of resistant weed biotypes following repeated use of herbicides [162]. In plants—including soybean—the Phase I and Phase II detoxification enzymes cytochrome P450 monooxygenases (P450s), glutathione S-transferases (GSTs), and UDP-glycosyltransferases (UGTs) are pivotal for herbicide metabolism [163,164,165,166] and have also been implicated in tolerance to heavy metals such as aluminum and cadmium [167,168,169]. Endogenous negative regulators restrain the expression of these detoxification enzymes. Thus, CRISPR/Cas9-mediated knockout of those repressors offers a means to upregulate P450s, GSTs, and UGTs. By enhancing both herbicide and heavy-metal detoxification, this genome editing approach could enable soybean cultivation on contaminated soils and bolster food security through expanded arable land use.

3.4. Overcoming Recalcitrance: Next-Generation Trait Stacking and Alternative Genome Editors

Climate change is drastically reshaping agricultural environments, manifesting as higher temperatures, altered rainfall patterns, more frequent droughts and floods, and intensified pressures from pests and diseases, which collectively threaten global crop productivity. To secure stable yields under these multifaceted stresses, future agricultural strategies must build crops with simultaneous resistance to both abiotic and biotic challenges.

Trait stacking—the concurrent incorporation of multiple tolerance genes into a single cultivar via advanced breeding or genome editing techniques—offers a powerful solution. By pyramiding, for example, drought-responsive transcription factors, heat-shock protein genes, salinity-tolerance transporters, and pathogen-resistance alleles, breeders can generate crop varieties that endure heatwaves, water scarcity, soil salinity, and disease outbreaks simultaneously. This integrative approach is crucial for developing the next generation of climate-resilient, high-yielding cultivars.

Despite these advances, applying genome-editing strategies to soybeans remain challenging. Soybeans are notably recalcitrant to transformation and regeneration among dicots [170], which constrain the deployment of larger, more complex editors, such as base and prime editors, in addition to CRISPR/Cas9. As a result, trait stacking has progressed more rapidly in rice and wheat, where multiplexed edits have generated broad-spectrum resistance or biofortified cultivars [28,171]. In comparison, soybean research has largely focused on single traits, such as oil and protein content (Table 2), often achieved through single-gene edits. To secure sustainable soybean production, it is therefore imperative to pursue new approaches, including multiplex trait stacking and alternative genome editing platforms beyond CRISPR/Cas9.

4. Conclusions

Recent advances in genome editing, particularly CRISPR-based technologies, have fundamentally transformed the landscape of soybean molecular breeding by enabling precise, efficient, and targeted modifications of genes associated with yield, seed composition, stress tolerance, and consumer-oriented traits. These breakthroughs have accelerated the development of soybean cultivars with enhanced nutritional value, improved resilience to environmental stresses, and reduced allergenicity, addressing both global food security and evolving consumer demands. In addition, precise molecular breeding has progressed further with the emergence of genome editing methods that avoid introducing foreign genes. DNA-free approaches, such as ribonucleoprotein (RNP)-based delivery and grafting transgenic rootstocks, to enhance non-transgenic scions are expected to play a crucial role in future soybean improvement strategies [172,173]. The newly achieved techniques would offer the potential to circumvent regulatory hurdles and improve consumer acceptance of genome-edited soybean cultivars [174].

Despite these achievements, the application of genome editing remains largely confined to a limited number of model cultivars, emphasizing the urgent need to expand genetic diversity by incorporating wild relatives and landraces. This expansion is essential for developing soybean cultivars adaptable to shifting cultivation zones and for introducing traits such as lodging resistance and broader environmental adaptability. Overcoming the recalcitrance of soybean transformation and regeneration will be crucial for extending genome editing to a wider range of commercial cultivars and for enabling the stacking of multiple beneficial traits.

Future directions in soybean improvement should prioritize multiplex trait stacking, combining genes for yield, nutrition, and stress resilience, alongside the integration of synthetic biology approaches, such as engineering nitrogen fixation and designer metabolic pathways, to generate novel metabolites and further enhance crop value. Additionally, unexplored frontiers, including the targeted engineering of photosynthetic efficiency, rhizosphere microbiome interactions, and detoxification pathways, offer promising avenues for next-generation soybean improvement.

As regulatory frameworks evolve, continued progress in editing precision, safety, and public acceptance will be vital for the commercial adoption of genome-edited soybeans. Ultimately, the convergence of advanced genome editing, digital agriculture, and sustainable management practices will be crucial in building climate-resilient crops. These resource-efficient soybean production systems can meet the challenges of a rapidly changing agricultural landscape.