CRISPR/Cas-Mediated Optimization of Soybean Shoot Architecture for Enhanced Yield

CRISPR was first noticed as unusual direct repeats in Escherichia coli [10] and later shown to colocalize with Cas genes [11]. Mojica and colleagues linked CRISPR spacers to invading phage DNA, proposing an RNA-mediated defense system [12]. This hypothesis was mechanistically confirmed when [13] demonstrated that the Cas9–CRISPR RNA (CrRNA)–trans-activating CRISPR RNA (tracrRNA) complex introduces DSBs at sites flanked by an NGG protospacer-adjacent motif (PAM). Engineering a single-guide RNA (sgRNA) of 20 nt streamlined the system for laboratory use and, within one year, Zhang and Church laboratories showed its functionality in mammalian cells [14,15], igniting a genome-editing revolution.

Subsequent engineering has broadened the precision and scope of the CRISPR platform. CRISPR systems are broadly classified based on effector module architecture, with Class 1 systems (Types I, III, and IV) employing multi-subunit protein complexes and Class 2 systems (Types II, V, and VI) utilizing single effector proteins [16,17] (Figure 1). High-fidelity Streptococcus pyogenes Cas9 (SpCas9) variants, such as enhanced specificity Cas9 (eSpCas9) and hyper-accurate Cas9 (HypaCas9), carry structure-guided mutations that suppress off-target cleavage [18,19], addressing one of the primary concerns in precise genome editing applications. The recently developed SpCas9 with relaxed PAM recognition (SpRY) variant further expands the targetable genomic regions by relaxing PAM requirements, enabling plant genome editing at sites previously inaccessible to conventional SpCas9 [20]. Catalytically impaired derivatives, nickase Cas9 (nCas9) and nuclease-dead Cas9 (dCas9), enable homology-directed repair bias or programmable transcriptional/epigenetic regulation [21,22] (Figure 2). CRISPR/dCas9-based transcriptional regulation systems represent another frontier in the genetic modulation of crops, with activators and repressors enhancing gene activation and interference [23]. These systems are particularly valuable for modulating gene expression without permanently altering DNA sequences, allowing for more nuanced control of plant traits. Building upon these modifications, precision editors emerged by fusing deaminases or reverse transcriptase to nCas9, leading to cytosine/adenine base editing [24,25] and prime editing for small insertions, deletions, or substitutions [26].

The parallel discovery of Class 2 alternatives has significantly increased targeting flexibility beyond that of the traditional SpCas9 system. Cas12a (Cas12a formerly Cpf1) recognizes 5′-TTTV PAMs and generates staggered ends [27], offering advantages in T-rich genomic regions that are common in plant promoters and introns. The more recently characterized Cas12h is a compact DNA cutter for only one strand of double-stranded DNA, allowing for precise gene editing [28]. It has been successfully engineered into an ABE-Cas12h for efficient A-to-G conversion. Cas13 targets RNA rather than DNA [29], opening new possibilities for the transient modification of gene expression. New variants of the CRISPR system are continuously emerging to improve the efficiency and precision of gene editing and reduce potential risks, thereby broadening the application prospects of CRISPR technology (Table 1).

Applying CRISPR to plant genomes requires adaptations to address polyploidy, cell wall composition, and genotype-dependent transformation. Early successes in Arabidopsis (Arabidopsis thaliana) and rice have been extended to diverse crops, with pooled CRISPR-Cas9 approaches enabling multiplex mutagenesis for large-scale functional genomic studies [56,57]. Compact nucleases like Cas12f derived from Acidibacillus sulfuroxidans, one of the smallest Cas proteins, can be maintained in Potato Virus X (PVX) vectors for plant genome editing [58]. Additionally, “CRISPR-Combo” systems integrate multiple editing capabilities, expanding the plant genome-editing toolbox [59]. While these innovations show remarkable potential, their agricultural applications hinge on overcoming delivery barriers through advanced transformation engineering.

Recent advancements in genetic manipulation techniques are closely linked to progress in transformation engineering methods. For instance, BBM and Wuschel (WUS) have significantly improved transformation efficiency in crops, such as maize (Zea mays), rice, and wheat [60]. In addition, the Wuschel-Related Homeobox 5 (TaWOX5) gene in wheat has been identified as an enhancer of transformation and regeneration [61]. To overcome the limitations of traditional transformation methods, novel strategies have been developed, including Ribonucleoprotein based CRISPR/Cas delivery systems and geminivirus replicons for the transient amplification of donor templates. These approaches have demonstrated high efficiency in crops [62,63,64,65]. Tissue culture-free genetic modifications have been achieved through virus-induced genome editing, grafting, and cut-dip-budding techniques [63,66]. Although transformation efficiency varies significantly among crops, the ongoing optimization of nuclease engineering, multiplex design, and innovative delivery methods is enhancing the toolkit for crop genetic improvement, unlocking the full potential of CRISPR technology in agriculture.

Node number in soybean—a critical determinant of yield—is regulated by a complex network of developmental pathways involving floral transition, vegetative growth duration, and stem architecture [106]. Advances in CRISPR/Cas9 genome editing have enabled precise manipulation of these pathways, offering new strategies for optimizing plant structure and productivity.

Central to this regulation are florigen genes, particularly Flowering Locus T (FT) homologs. The FT protein, and in some cases, its mRNA, functions as mobile flowering signals that primarily transported from leaves to the shoot apical meristem to trigger the reproductive transition [107]. In soybean, GmFT2a acts as a classical photoperiod-sensitive florigen primarily promoting flowering under short-day conditions by interacting with transcription factors such as Floral Development Locus 19 (GmFDL19) to activate downstream floral integrator genes including Apetala1 (GmAP1) and Suppressor of Overexpression of Constans 1 (GmSOC1) [108,109,110]. CRISPR/Cas9-mediated knockout of GmFT2a delayed flowering, extending the vegetative phase and increasing node number [109,110,111]. GmFT5a functions as both a floral promoter and determinacy regulator, modulating shoot architecture by disrupting the Dt1–AP1 feedback loop [112]. Loss-of-function GmFT5a mutations delay flowering and can disrupt post-flowering stem termination, leading to a shift from determinate to indeterminate growth habits, which enables continued node production beyond the reproductive phase, particularly under specific photoperiod conditions [109,110]. The sequential activation of these florigen genes creates a developmental window during which node formation continues, directly linking flowering gene expression kinetics to final node number.

In parallel, the circadian clock machinery contributes to architectural regulation through its influence on growth patterns. Mutations in Late Elongated Hypocotyl (GmLHY), a key circadian regulator, resulted in reduced plant height and shortened internodes, likely via disruption of gibberellin (GA) signaling pathways [113]. Direct manipulation of transcription factors involved in developmental transitions has also produced substantial architectural changes. Targeted editing of AP1 homologs generated quadruple mutants that displayed delayed flowering under short-day conditions, increased plant height, and resulted in a greater internode length and a higher number of nodes [114]. Similarly, CRISPR/Cas9-mediated mutagenesis of the Squamosa Promoter Binding Protein-like 9 (GmSPL9) gene family (GmSPL9a–d) enhanced branching and increased node numbers, with each homolog contributing differentially to overall plant architecture [115].

Collectively, these studies highlight how integrated control of flowering, stem growth, and circadian regulation underlies node number variation in soybean. The precision of CRISPR/Cas9 technology in targeting these interconnected pathways presents a powerful approach for developing soybean cultivars with improved yield potential and adaptability.

Optimizing soybean plant architecture requires precision engineering of internode length, as this directly affects plant height and lodging resistance. The GA pathway serves a central regulatory function in plant height, with its biosynthesis, metabolism, and signaling pathways being critical determinants of stem elongation [116,117]. Loss-of-function mutations in Dwarf1 (GmDW1), identified from EMS-induced mutants, result in dwarf phenotypes in soybean by disrupting GA biosynthesis [118]. Advances in CRISPR/Cas technology have facilitated the targeted modification of key GA-related genes to achieve desired plant heights. For instance, CRISPR-mediated mutagenesis of GmLHY genes, which encode myeloblastosis (MYB) transcription factors, significantly reduces plant height by modulating the GA pathway in soybean [113]. Additionally, AP1 homologs in soybean, such as GmAP1s, influence internode length through GA biosynthesis; a quadruple gmap1 mutant showed longer internodes and increased plant height, while overexpression of GmAP1a led to shorter internodes [114]. Conversely, targeting GA catabolism genes provides an alternative method for height control. Studies on Gibberellin 2-oxidase 8 (GA2ox8A and GA2ox8B), which encode GA2-oxidases involved in degrading active GAs, indicate that higher copy numbers of these genes are associated with reduced shoot length, a trait selected during soybean domestication [119].

Light conditions, particularly in dense plantings, trigger shade-avoidance responses that can result in undesirable internode elongation. Blue light photoreceptors, such as Cryptochrome 1s (GmCRY1s) in soybean, play a crucial role in repressing excessive stem elongation under low blue light conditions by stabilizing the basic Leucine Zipper (bZIP) transcription factors TGACG-motif-binding Factor 1 and 2 (STF1 and STF2), which, in turn, upregulate GA2ox expression and reduce the levels of bioactive GAs [120]. The photomorphogenic repressor Constitutive Photomorphogenic 1b (GmCOP1b) promotes stem elongation under light conditions; CRISPR-created gmcop1b mutants exhibited reduced shade responses and improved performance in high-density planting [121]. The Plant Height 13 (PH13) gene and its paralog PHP (PH13 Paralog Homolog) are also key regulators. PH13 affects internode length by interacting with GmCOP1 proteins to modulate the accumulation of STF1/STF2, which are vital for GA metabolism [122,123]. Knocking out PH13 and PHP using CRISPR/Cas9 resulted in optimized internode lengths, enhancing lodging resistance and increasing yields in dense cultivation [124]. Furthermore, the reduced internode 1 (rin1) gene, which encodes a SUPPRESSOR OF PHYA 105 3a (SPA3a) protein, impacts internode length by regulating GA2ox7 expression via STF1/STF2, linking light signaling components to GA-mediated growth control [123].

Additional pathways and environmental responses also regulate internode length. The brassinosteroid (BR) pathway is critical for cell elongation and division. For example, CRISPR/Cas9-mediated editing of DWARF1(GmDWF1) genes (GmDWF1a and GmDWF1b), which encode key enzymes in BR biosynthesis, has successfully induced dwarfism in soybean, highlighting the potential for BR pathway manipulation to create compact plant architectures suitable for high-density planting [125]. Moreover, plant responses to Ultraviolet-B (UV-B) radiation can influence plant height. In soybean, the E3 ubiquitin ligase Increased Leaf Petiole Angle 1 (GmILPA1) is upregulated under UV-B exposure, promoting the ubiquitination and degradation of the GA catabolic enzyme GmGA2ox-like. This action counters the UV-B-induced reduction in bioactive GAs, thus promoting plant height. Conversely, gmilpa1 mutants exhibit dwarfism under UV-B conditions, underscoring GmILPA1’s role in maintaining GA homeostasis and growth under UV-B stress [126]. These findings illustrate the diverse molecular targets available for fine-tuning soybean internode length and overall plant height through precision gene editing.

In the protein degradation pathway, PH13 and its paralog PHP function as key regulators of soybean branching architecture. PH13 encodes a WD40-repeat domain protein homologous to Arabidopsis Suppressor of Phya-105 (SPA) family proteins. Additional pathways and environmental responses also regulate internode length. The GmAP1 genes act as negative regulators of branch formation, as quadruple gmap1 mutants exhibit significantly increased branch numbers, whereas GmAP1a-overexpressing lines show reduced branching [127].

Branching architecture fundamentally determines soybean canopy structure, light interception efficiency, and yield potential [128]. CRISPR/Cas technology has emerged as a powerful tool for the precise modification of genes that control branch development.

Several key regulatory genes have been identified as targets for optimizing branch patterns. The SPL9 gene family functions as a negative regulator of branch formation in soybean. While single mutants showed limited phenotypic changes, higher-order mutants exhibited significantly increased node and branch numbers [115]. The Dt2–GmAP1 regulatory module represents a key mechanism controlling shoot branching in soybean. Through a genome-wide association study of 2409 soybean accessions, Dt2 was identified as a major locus on chromosome 18 associated with branch number [127]. Functionally, Dt2 interacts with Agamous-like 22 (GmAGL22) and GmSOC1a to form a transcriptional complex that directly binds to the promoters of GmAP1a and GmAP1d, activating their expression.

It forms an E3 ubiquitin ligase complex with GmCOP1 proteins to target specific transcription factors, such as STF1/2, for degradation, thereby modulating plant architecture. The phd (PH13/PHP double mutant) plants exhibited multiple advantageous traits under high-density planting conditions, including increased branch number, more pods per plant, thicker stems, and a 15.8% improvement in grain yield per plant compared to the wild-type cultivar JY202 [124].

Hormone transport and signaling also play crucial roles in branch development. Pinformed 1(GmPIN1) genes encode auxin efflux carriers that control hormone distribution patterns in soybean. Zhang et al. found that gmpin1abc and gmpin1bc multiple mutants displayed more branches than wild-type plants [129]. In addition, these mutants showed shorter shoot stature and, in some lines, higher podding rates, indicating that GmPIN1 regulates both branching and broader aspects of plant architecture.

The capacity of CRISPR/Cas9 technology to simultaneously target multiple genes represents a significant advantage in soybean enhancement. This multiplexing capability is essential for addressing genetic redundancy within the complex soybean genome, where the modification of a single gene often leads to minimal phenotypic changes due to the compensatory functions of homologous genes.

Optimizing leaf orientation and petiole inclination is crucial for maximizing light interception and photosynthetic efficiency in soybean canopies [130]. However, conventional breeding efforts are often constrained by pleiotropic loci that impose trade-offs between leaf morphology and yield [131]. The advent of CRISPR/Cas9 genome editing offers new avenues to target these multifunctional genes with high precision, potentially circumventing such limitations.

A paradigmatic example is the JAGGED1 (Ln) locus, which encodes a transcription factor that simultaneously regulates leaflet width and the number of seeds per pod (NSPP). The dominant Ln allele confers broad leaflets but reduces NSPP, whereas the recessive ln allele results in narrow leaflets accompanied by increased seed set employed CRISPR/Cas9 in the low-latitude soybean variety ‘Huachun 6′ to knock out GmJAG1 using dual sgRNAs targeting two exons [132,133]. The resultant gmjag lines displayed the characteristic narrow-leaflet morphology and achieved an 8.7–8.8% yield increase under field conditions. While this work confirms the pleiotropic role of GmJAG1, it also illustrates that leaflet morphology and seed number remain coupled in the mutant lines, indicating that further allelic refinement or regulatory-level editing will be necessary to fully decouple these traits [133].

Beyond transcriptional regulators, hormone transport pathways also provide promising targets for architectural optimization. Zhang et al. demonstrated that multiplex CRISPR/Cas9 mutagenesis of the GmPIN1 auxin efflux carriers disrupted polar auxin distribution at the petiole base, leading to more erect leaf angles favorable for high-density planting [129]. This work underscores the utility of targeting gene families with redundant functions in polyploid genomes to sculpt plant architecture.

Similarly, GmILPA1 plays a critical role in modulating leaf petiole angle through control of pulvinus cell proliferation. A gamma ray-induced deletion in the fourth exon of GmILPA1 was shown to increase petiole angles, and natural variation in its expression is inversely correlated with angle magnitude across diverse soybean accessions [134]. These findings point to GmILPA1 as a promising candidate for promoter or base editing to achieve graded phenotypic outputs without altering protein coding sequences.

Collectively, these case studies highlight the transformative potential of CRISPR/Cas-based strategies in soybean canopy engineering. Representative target genes involved in regulating plant architecture and yield potential are summarized in Table 3. Through targeted gene disruption, multiplex editing, and regulatory sequence manipulation, researchers are now equipped to rationally redesign canopy traits to enhance light capture and yield efficiency in intensive cropping systems.

The regulatory network of soybean plant architecture is intricate, with key genes often showing pleiotropic effects. For instance, the transcription factor GmMYB14 influences leaf angle and enhances drought resistance by regulating stress-responsive pathways [141], while the soybean Dt1 gene interacts with Flowering Locus D-class 1 (FDc1), FT5a, and AP1 to form a complex network that governs reproductive transition and shoot apical meristem activity [112]. Chromatin topology and long-range chromosomal interactions also play roles in regulating gene expression during soybean polyploidization, diploidization, and domestication [119]. This complexity has spurred multi-omics advances; for instance, Li et al. developed an integrated database combining the deep sequencing of 984 germplasms, 2500+ phenotypic images, and existing multi-omics data to enable trait gene mining [142]. Concurrently, a cellular atlas of soybean was constructed using single-nucleus RNA sequencing, encompassing approximately 120,000 nuclei from 10 distinct organs and morphological structures [143], and another study mapped organ development via RNA sequencing [144]. Currently, multiple soybean multi-omics platforms have been developed to support soybean genetic research and breeding work. The main platforms established include SoyOmics, SoyFGB, SoyOD, SoyMD, etc. Integrating multidimensional omics data with artificial intelligence enables the identification of high-confidence target genes for CRISPR-based editing by reducing false positives, decoding complex regulatory landscapes, and facilitating precise functional validation. This approach supports precision breeding by prioritizing candidate genes across diverse biological layers [145].

The spatiotemporal regulation of plant growth genes is essential for organ development and environmental adaptation, with promoter regions critically orchestrating the timing and location of gene expression. For example, the Small Heat Shock Proteins (SmsHSP24.1) promoter in eggplant (Solanum melongena) harbors cis-regulatory elements that govern its spatiotemporal expression. A 2 kb promoter fragment was shown to drive heat-inducible expression in sensitive tissues, such as anthers and young leaves, while truncation analysis revealed a 300 bp core region responsible for basal activity and a distal 1.2 kb segment containing heat-responsive elements, illustrating promoter-mediated integration of environmental cues with tissue-specific expression [146]. Building on this understanding, CRISPR/Cas systems have enabled a range of promoter engineering strategies, from precise modifications of existing elements to the rational design of novel regulatory circuits.

A primary strategy for promoter engineering involves the targeted modification or disruption of existing cis-regulatory elements to precisely adjust a gene’s native expression pattern. This approach effectively fine-tunes transcriptional output. For instance, CRISPR/Cas9-mediated edits to the Nuclear Factor Y subunit C4 (NF-YC4) promoter in rice and soybean enhanced protein accumulation, leading to improved yield [147]. Similarly, precise editing of cis-regulatory elements within the promoter of the heat-inducible Casein Kinase I (GhCKI) gene in cotton (Gossypium hirsutum), using both CRISPR/Cas9 and CRISPR/Cpf1, disrupted MYB transcription factor binding sites. This modification attenuated high-temperature-induced GhCKI activation, thereby preserving male fertility under thermal stress [148].

A more advanced strategy focuses on designing and assembling synthetic regulatory circuits. This approach treats known cis-regulatory elements as modular units, combining them to create predictable and tunable expression patterns. Such modular design enables de novo programming of gene expression. A notable example is the use of prime editing, a highly precise CRISPR-based technique, to insert a carefully designed 10 bp heat-shock element into the targeted promoter region of the endogenous cell-wall invertase gene Cell Wall Invertase 5 (LIN5) in tomato, as well as in rice. This precise insertion rewired the native promoter, conferring heat-inducible expression that significantly enhanced yield resilience under thermal stress [149]. This work exemplifies the power of CRISPR technologies in enabling tailored genetic modifications and highlights the synthetic biology principle of modular design, offering a blueprint for engineering climate-resilient soybean cultivars.

The expanding toolkit of promoter engineering—from targeted edits of single elements to the modular construction of novel regulatory functions—provides unprecedented control over plant trait programming. Enabled by continuous advancements in CRISPR technologies, these strategies are pivotal for developing high-performing, resilient soybean cultivars capable of addressing future agricultural challenges.

Soybean genome editing faces challenges due to its paleopolyploid genome, with comparative analyses indicating that approximately 75% of soybean genes exist in multiple copies [150]. On the one hand, this genetic redundancy necessitates simultaneous editing of homologous genes to observe phenotypic changes, as demonstrated in studies on GmLHY and GmSPL9 genes. Quadruple mutants of GmLHY genes showed reduced plant height and shortened internodes [113], whereas higher-order mutants of GmSPL9 genes exhibited altered plant architecture with increased node and branch numbers [115]. However, functional differentiation of homologous genes complicates selective editing [151,152]. The editing efficiency of CRISPR/Cas9 systems is constrained by both the GC content in regions proximal and distal to the PAM sequence and the preference for NGG PAM motifs, which together impose significant limitations on target-site selection [153]. The development of new editing tools has overcome the limitations of editing site selection. Next-generation tools like Cas12a and PAM-less SpRY offer expanded targeting capabilities [154,155] but introduce new challenges, such as potential off-target effects [156]. Additionally, although the study by Chen et al. has shown that the editing efficiency of the CRISPR-SpRY system in soybeans can reach up to 57.7%, there is still significant room for improvement in multiple aspects. First, target dependency is evident, with editing efficiency varying significantly among targets. Second, base editing efficiency remains low: A-to-G editing is only 5.0–5.6%, and C-to-T editing is mostly around 10%. Further, multi-gene editing efficiency is limited: While single-target editing is efficient, editing 4–6 genes simultaneously reduces biallelic mutation rates to 4–17% [156]. Further refinement is required for precise and efficient genome engineering of soybean.

The application of CRISPR-Cas in soybeans has long been hindered by transformation inefficiencies and genotype-dependent regeneration challenges [157]. Developmental genes such as GROWTH-REGULATING FACTOR5 have shown promise in enhancing soybean transformation [158,159]. This approach remains limited by chimerism that affects inheritance and the potential pleiotropic effects of GRF overexpression on plant development [159]. These limitations have spurred the exploration of alternative delivery systems such as virus-induced genome editing (VIGE).

VIGE has emerged as a promising strategy to overcome the need for stable transgene integration and to reduce regulatory barriers to generate gene-edited DNA-free plants [160,161,162]. In soybean, the Apple Latent Spherical Virus (ALSV) has been adopted as the first viral vector for gRNA delivery, successfully editing endogenous genes such as Phytoene desaturase [163]. VIGE systems demonstrate competitive editing efficiencies: In Solanaceous plants, Tobacco Rattle Virus (TRV)-based systems achieve 40.3% and PVX systems reach 36.5%, while the ALSV-based CCAC (Cas9-based Csy4-processed ALSV Carry) system shows 45.3% editing efficiency specifically in soybean hairy roots [107,163]. Beyond numerical efficiency, VIGE offers significant temporal advantages by eliminating the tissue culture period required for antibiotic selection in conventional Agrobacterium-mediated approaches. The CCAC system enables efficient gRNA delivery once a stable Cas9-expressing line is established, offering significant advantages for crops with low transformation rates like soybeans [163].

However, VIGE application in soybeans remains in its infancy, facing critical challenges, such as the efficient delivery of large Cas9 proteins and optimization of viral vectors for stable editing. The size constraints of viral cargo—Cas9 coding sequence (~4.2 kb) often exceeds viral packaging limits—pose significant hurdles, as does the requirement for transient expression without permanent genomic integration. The CCAC system addresses these through a hybrid approach: Cas9 is delivered via traditional Agrobacterium transformation while gRNAs are delivered through ALSV infection, utilizing the Csy4 endonuclease for precise gRNA release [163]. Additional strategies include engineering compact Cas variants (e.g., CasPhi), implementing split-Cas systems for reconstitution in planta, utilizing dual-virus approaches, and optimizing viral vector design through capsid engineering [164,165].