CRISPR/Cas gene editing for plant resistance to pests, diseases, and herbicide: application, risk assessment, and safety evaluation

ABSTRACT

Global food security faces growing threats from substantial crop losses attributable to insect damage, pathogenic diseases, and herbicide-related impacts. Developing new tools to enhance crop production has become a critical challenge. The CRISPR/Cas system represents a breakthrough in precision genome editing that operates without requiring transgene integration, fundamentally transforming both plant science research and crop enhancement strategies. However, the application of this approach in five major crops (rice, wheat, maize, potato, and soybean), faces hurdles stemming from public acceptance and regulatory uncertainty. This review provides a systematic analysis of CRISPR/Cas technology applications for developing resistance against insects, diseases, and herbicides in essential food crops. We evaluate both the promising potential and significant challenges of this technology, particularly off-target effects and environmental safety. Furthermore, we investigate the evolving regulatory frameworks for gene editing crops across different countries, as regulatory clarity represents a critical determinant for their commercial development.

Introduction

Global agricultural production confronts unprecedented challenges. Projected population growth to 9.6 billion by 2050 necessitates a 60% increase in staple crop output.¹ Declining yield gains from Green Revolution technologies, compounded by climate constraints, demands cultivars with enhanced resilience, productivity, and nutritional quality.² Conventional breeding approaches remain labor-intensive and time-prohibitive, underscoring the need for efficient alternatives.³ Rapid advances in sequencing have enabled precise genome editing technologies, facilitating predictable genetic modifications for trait development. These innovations establish precision breeding platforms that define next-generation plant improvement strategies.⁴ The CRISPR (clustered regularly interspaced short palindromic repeats)/Cas system has become the premier platform for crop genome engineering.⁵ It induces targeted double-strand breaks (DSBs) repaired through two primary pathways: error-prone nonhomologous end-joining (NHEJ), often generating indels or substitutions; and homology-directed repair (HDR) when donor templates are available.⁶ Leveraging programmable RNA-guided DNA recognition, CRISPR/Cas enables cost-effective DSB induction at any genomic locus.⁷ Since its first applications in plants in 2013,^8–10 CRISPR/Cas technology has rapidly expanded and been deployed across numerous crop species, introducing highly valuable agronomic traits into many of these crops.¹¹ The ongoing development of CRISPR-based editing systems has progressively expanded the technical capabilities beyond conventional DSB-dependent approaches.⁴ Newly emerging platforms, including base editors (BEs) and prime editors, enable precise genome modifications without relying on double-strand breaks. Specifically, base editing technology facilitates irreversible nucleotide conversions in the absence of DSBs or donor templates, as demonstrated by the established cytidine base editors (CBEs) and adenine base editors (ABEs).^12,13 These capabilities establish CRISPR/Cas as an unparalleled agricultural genome manipulation tool.

CRISPR/Cas technology offers an alternative method of genome editing in crops, accelerating the development of improved varieties.¹⁴ Rice (Oryza sativa), wheat (Triticum aestivum), maize (Zea mays), potato (Solanum tuberosum), and soybean (Glycine max) are major food crops worldwide.¹⁵ Maize, rice and wheat are the world’s three major cereal crops. Together they form a vital part of the human diet, contributing approximately 87% of global grain production and providing around 43% of all calories consumed worldwide.¹⁶ Maize is the most extensively cultivated crop worldwide. Its planting area has continued to expand since 2010, rising by 46%. This growth is supported by the crop’s use in various non-food sectors, including biofuels and animal feed.¹⁵ In 2023, rice and wheat production reached similar volumes, at approximately 800 and 799 million tonnes respectively. However, their growth patterns since 2010 have differed: rice production increased by 15%, while wheat output grew by 25%.¹⁵ Breeding high-yielding and resilient varieties of both rice and wheat is therefore essential to achieving global food security. Potato is the most produced commodity in the roots and tubers crops, with 383 million tonnes in 2023, up 17% compared to 2010¹⁵ and is the world’s most important tuber crop. It is produced on a large scale, widely consumed, and remains affordable and readily available in the open market.^17,18 Soybean is a key source of protein and oil for food and feed. While genome editing offers great potential for its genetic improvement, application of this technology in soybean lags behind crops like rice.¹⁹ This is largely due to inefficient Agrobacterium-mediated transformation and regeneration, which restricts the use of modern biotechnologies, including genome editing. In addition, soybean is a paleotetraploid-derived diploid with high gene redundancy, complicating functional studies and targeted mutagenesis. These factors make genome editing in soybean more challenging than in rice and other model plants.^20–22 This review synthesizes current applications of CRISPR/Cas-based genome editing for trait enhancement in four globally essential food crops: rice, wheat, maize, potato, and soybean. We focus on improvements in resistance to insects, disease, and herbicides. The article also addresses biosafety considerations and how different countries regulate gene-edited crops. By providing a broad overview of applications and risk assessment, we aim to support the responsible development and adoption of this promising technology.

Genetic Improvement for Insect Resistance in Crop Plants

Insects represent a major cause of global crop loss, acting as both direct pests and vectors of plant diseases.²³ Genetic engineering provides powerful tools to develop effective resistance against these biotic threats. An alternative strategy involves editing plant genes to enhance resistance. Modifying key components of plant immunity including susceptibility (S) genes, resistance (R) genes, and genes mediating insect effector interactions may provide durable protection against crop pests (Table 1).³² A two steps strategy that involves editing insect genes in the laboratory followed by field release of modified individuals shows promise for pest control.³³ This approach can suppress insect populations by causing lethality, skewing sex ratios, or reducing reproductive capacity (Table 2).³² CRISPR/Cas9 provides an alternative strategy for insect control by editing genes that regulate essential behaviors. Targeting traits such as chemical sensing and mate recognition can disrupt key interactions between insects and their host plants.⁵⁴ Currently, most research has focused on optimizing the CRISPR/Cas system in specific crop species. However, growing insights into plant – insect interactions now offer promising opportunities to apply genome editing in developing insect-resistant crops.

Table 1.

CRISPR/Cas gene editing in plants for insect pest management.

Crops	Target insect	Target gene	Editing	Delivery of CRISPR components	References
Rice	Nilaparvata lugens and Chilo suppressalis	OsCYP71A1	Deletion	Cas9 mRNA and sgRNA	24
	Nilaparvata lugens	OsLRR2	Knockout	Cas9 and sgRNA	25
		OsWRKY36	Knockout	Cas9 and sgRNA	26
		ACS2	Knockout	Cas9 and sgRNA	27
	Cnaphalocrocis medinalis	OsJAZ10	Frameshift mutation	Cas9 protein and sgRNA	28
Maize	Helicoverpa armigera and Spodoptera frugiperda	ZmPP2C45	Knockout	Cas9 and sgRNAs	29
Soybean	Spodoptera litura	GmCDPK38	Knockout	Cas9 and sgRNA	30
	Helicoverpa armigera and Spodoptera litura	GmUGT	33bp deletion and 1bp insertion	Cas9 and sgRNA	31

Table 2.

CRISPR/Cas gene editing in insects for insect pest management.

Target insect	Target gene	Mutation type	Delivery of CRISPR components	References
Nilaparvata lugens	NlugGr23a	Knockout	Cas9 mRNA and sgRNA	34
	Nl-cn and Nl-w	Knockout	Cas9 mRNA and sgRNA	35
Agrotis ipsilon	AiTH	Knockout	sgRNA and Cas9 mRNA	36
Helicoverpa armigera	NPC1b	Knockout	RNP complex	37
	HaCad	Knockout	sgRNAs and Cas9 mRNA	38
	HaABCA2	Knockout	Cas9 mRNA and sgRNA	39
	eptor 16 (OR16)	Knockout	Cas9 mRNA + sgRNA and RNP complex	40
	white, ok, brown, and scarlet	Knockout	Cas9 mRNA	41
	cluster of nine P450 genes	Knockout	Cas9 protein and multiple sgRNAs	42
Mythimna separata	NPC1b	Knockout	RNP complex	43
Spodoptera litura	Slabd-A	Knockout	Cas9 mRNA and sgRNA	44
	Orco	Knockout	Cas9 mRNA and sgRNA	45
	SlitBLOS2	Knockout	Cas9 mRNA and sgRNA	46
Spodoptera frugiperda	BLOS2 E93 TO	Knockout	Cas9 protein and multiple sgRNAs	47
	Sfabd-A	Knockout	RNP complex	48
Leptinotarsa decemlineata	vestigial gene (vest)	Knockout	RNP complex	49
Locusta migratoria	Orco	Knockout	mRNA	50
	Crz	Knockout	xCas9	51
Tetranychus urticae	PSST	Knockout	Plasmid	52
	phytoene desaturase	Knockout	RNP complex	53

Genome Editing in Crop Plants

Insects depend on essential chemical compounds present in plants for their growth, immunity, and observed behaviors in rice.²⁴ Research indicates that most insects rely on essential plant compounds and volatiles for the development of their immune system.⁵⁵ Using CRISPR/Cas systems, plants can be engineered to produce or suppress specific enzymes that either repel certain insect pests or attract natural predators to control pest populations.⁵⁶ Editing key plant immunity genes provides a reliable strategy for developing insect-resistant plant varieties and enhancing host defense against pests. For instance, genome editing can improve pest resistance by knocking out S genes in plants. When Lu et al. ²⁴ used CRISPR/Cas9 to disrupt the OsCYP71A1 gene in rice, serotonin biosynthesis was inhibited, resulting in increased resistance against both the brown planthopper Nilaparvata lugens and the Asiatic rice borer Chilo suppressalis, two devastating rice pests in Asia and other rice-growing regions. Similarly, Kuai et al. ²⁵ found that knocking out OsLRR2 significantly enhances rice resistance to the brown planthopper and rice blast, improves response to the bacterial elicitor flg22, and increases rice yield. Lignin, which acts as a physical barrier against pathogens and pests, represents another target for achieving balanced improvements in yield and resistance. Early studies showed that the transcription factor OsMYB30 activates phenylalanine ammonia-lyase (PAL) genes OsPAL6/8, promoting lignin production and enhancing resistance to the brown planthopper.⁵⁷ More recently, Liu et al. ²⁶ revealed that OsWRKY36 downregulates OsPAL expression, suppresses lignin synthesis, and reduces parenchyma tissue thickness, thereby lowering resistance to pests and diseases. Conversely, knocking out OsWRKY36 releases the suppression of both PAL genes and key yield regulators IPA1 and MOC2, simultaneously enhancing resistance to N. lugens, Sogatella furcifera, Laodelphax striatellus, rice blast, and bacterial leaf blight (BLB), while improving yield traits. This dual regulatory role identifies OsWRKY36 knockout as a promising target for synergistically enhancing disease resistance and productivity in rice. Multiplex CRISPR/Cas9 editing offers the possibility to stack insect resistance with other agronomic traits in a single event. Qin et al. ²⁷ concurrently edited ACS2 (for brown planthopper resistance), along with Bsr-D1, ERF922, Pi21 (for blast resistance) and Xa5 (for bacterial blight resistance) in rice. The resulting triple-mutant lines exhibited significantly enhanced resistance to the brown planthopper, rice blast, and bacterial blight without compromising agronomic traits, demonstrating that insect resistance can be successfully combined with disease resistance through multiplex gene editing. This knockout strategy extends beyond rice to other crops. In maize, knocking out the ZmPP2C45 suppressor gene, which regulates benzoxazinoid (BZX) defensive metabolite biosynthesis, increased BZX content by more than threefold and significantly inhibited insect herbivore growth.²⁹ According to Li et al.³⁰ the GmCDPK38 mutant with the Hap3 deletion in soybeans showed significant resistance to common cutworms. Additionally, the GmUGT gene deletions of 1bp and 33bp were made in soybeans to improve their resistance to S. litura and H. armigera.³¹

Beyond editing S genes, CRISPR/Cas9 can also generate beneficial neoproteins via frameshift mutations. Jasmonate (JA) signaling, which regulates lignin accumulation via PAL genes, is central to rice defense. Previous work showed JA signaling mediates resistance to Cnaphalocrocis medinalis via defense compounds.⁵⁸ A recent study targeting the JA negative regulator JAZ10 revealed an unexpected outcome, a specific frameshift mutant produced a new protein (FJ10) that enhanced both growth and brown planthopper resistance by modulating DELLA and PAL pathways.²⁸ Field trials confirmed FJ10-expressing lines maintained higher yields under pest pressure. This case demonstrates frameshift mutations can generate neoproteins that resolve growth-defense trade-offs.

Genome Editing in Insects

In many insects, mating behavior is guided by pheromones: females release chemical signals that males detect and use to identify receptive partners.⁵⁹ The gustatory receptor gene NlugGr23 was successfully disrupted in N. lugens using CRISPR/Cas9 technology.³⁴ Mutant males with impaired pheromone detection showed complete mating avoidance toward homozygous females and ultimately caused infertile egg laying, indicating a novel population control tactic.³⁴ Olfactory receptors (ORs) play a central role in the insect chemosensory system.⁵⁵ When N. lugens had the NICSAD gene knocked out, it led to reduced oviposition and hatching rates.⁶⁰ In a study aimed at advancing functional genomics in N. lugens, the CRISPR/Cas9 system was used to knock out the cinnabar (Nl-cn) and white (Nl-w) genes, both involved in eye pigmentation. Successful knockout of these genes was confirmed through subsequent RNAi-mediated knockdown experiments, which reproduced the expected loss-of-pigmentation phenotypes.³⁵ All these results offered valuable data for the study of functional genomics and pest management in this planthopper species.

Lepidopteran pests cause devastating losses to crop yields. However, the overuse of chemical controls against major species such as Spodoptera litura, S. littoralis, Plutella xylostella, and Helicoverpa armigera has driven the widespread evolution of resistance to conventional pesticides. The S. litura Abdominal-A (Slabd-A) gene is critical in S. litura embryogenesis,⁴⁴ and introduction of Cas9 mRNA along with Slabd-A-targeting sgRNA into early embryos resulted in a distinct loss-of-function phenotype.⁴⁴ The resulting larvae displayed clear defects in segment patterning and abnormal pigmentation.⁴⁴ By employing CRISPR/Cas9 to generate olfactory receptor coreceptor (Orco) gene knockouts in S. littoralis, Koutroumpa et al. ⁴⁵ demonstrated that this gene is indispensable for the sense of smell in Lepidoptera. Mutant insects lacking this gene could no longer detect plant odors or sex pheromones. Wang et al. ⁶¹ utilized CRISPR/Cas9 to disrupt the ABCC2 gene in Ostrinia furnacalis via a 8-bp deletion. The resulting homozygous OfC2-KO strain, which produces a truncated protein, was subsequently demonstrated to confer high-level resistance to the Cry1Fa toxin. Diamide insecticides target insect ryanodine receptors (RyRs). Zuo et al. ⁶² created a strain of S. exigua carrying the RyRG4946E mutation using CRISPR/Cas9 system. This genetic modification made the insects highly resistant to diamide insecticides. The black cutworm (Agrotis ipsilon) feeds on many important vegetables and grain crops. To study its biology, researchers have created mutations in the tyrosine hydroxylase (AiTH) gene through CRISPR/Cas9 system. Some mutants were observed to develop a narrowed eggshell while the embryos inside developed fully, but they failed to hatch. Other embryos that did hatch experienced severe dehydration and died within one day.³⁶

The insect membrane protein NPC1b represents a potential target for pest control. However, such targeting strategies may be susceptible to off-target effects.³⁷ Research using CRISPR/Cas9 has confirmed that the HaCad protein is a critical receptor required for resistance to the Cry1Ac toxin in H. armigera.³⁸ Separately, the CRISPR/Cas9 enables a novel pest control strategy by manipulating mating timing through antagonist-mediated optimization, thereby maximizing fecundity and offering a novel pest suppression strategy.³⁹ To verify the resistance to the Bt toxin Cry2Ab in H. armigera, researchers used CRISPR/Cas9 to generate HaABCA2 knockout strains.⁴⁰ These mutants exhibited strong resistance to both Cry2Aa and Cry2Ab toxins, thereby confirming that HaABCA2 is essential for mediating their toxicity.⁴⁰ In the same species, CRISPR/Cas9 was used to disrupt four key pigment genes, including white, ok, brown and scarlet, producing mutants with distinct physiological phenotypes, somatic knockouts of white block pigmentation of the egg, first instar larva and adult eye, but germ-line knockouts of white are recessive lethal in the embryo.⁴¹ To investigate how insects respond to insecticides and to develop improved pest control strategies, researchers used CRISPR/Cas9 in H. armigera to knockout a cluster of nine P450 genes. This study identified key metabolic genes, yielding crucial leads for managing resistance in this major pest.⁴² The NPC1b gene encodes a protein critical for sterol absorption and trafficking in pest Mythimna separata, and Tang et al. ⁴³ established an efficient editing protocol by microinjecting eggs with preassembled Cas9-sgRNA ribonucleoprotein (RNP) complexes, and mutants were detected among G1 progenies, which demonstrated that the mutation generated by the Cas9/sgRNAs system on M. separata was heritable. This approach provides a foundation for developing new genetic control strategies against this pest.

The Colorado potato beetle (CPB, Leptinotarsa decemlineata) is a major agriculture pest that infests potato and other solanaceous plants.⁶³ Researchers successfully replicated a previously observed RNAi-induced wing deformity by creating vestigial gene mutants (vest) with CRISPR, thereby validating this gene’s function. This achievement established a vital genetic toolkit for CPB, facilitating the creation of more sustainable pest control methods.⁴⁹

The locust (Locusta migratoria) severely compromises global food security worldwide.⁶⁴ To investigate its functional genetics, researchers used CRISPR/Cas9 to target the odorant receptor co-receptor (Orco) gene, achieving high mutation efficiency.⁵⁰ Yan et al. ⁵¹ demonstrated the efficacy of xCas9 by generating a heritable homozygous Crz–/– locust strain. This mutant strain showed compromised physiological performance under low-temperature stress, manifesting as prolonged lifespan, reduced body weight, and smaller body size. Collectively, this work extends the application of xCas9-based gene editing to non-model insects.

Future pest management strategies should prioritize the advancement of CRISPR/Cas-based technologies. For instance, the use of CRISPR/Cas9 for transcriptional regulation represents a promising strategy. However, in such cases, the resulting organisms would not be considered “transgene-free,” since they would still contain integrated sequences encoding dCas9 and guide RNA. While genome editing has been used to alter insect pests for better plant resistance, the approach of directly engineering plants to manage pests remains less explored. The genetic modification of insects presents a promising yet complex strategy, which demands meticulous trait selection to guarantee environmental safety and preserve ecological integrity. Furthermore, extending these techniques to non-model plant species faces significant challenges, including diverse genetic backgrounds, specific tissue culture needs, and variable transformation efficiency.⁶⁵ A critical priority is to expand and characterize the pool of resistance genes through extensive phenotyping. This effort should include systematically evaluating available germplasm and wild crop relatives for their insect resistance. Multi-omics approaches can then help identify key stress-responsive genes. Ultimately, multiplex editing of these genes via high-throughput transformation techniques will form a core strategy for managing insect pests with genome editing.

Genetic Improvement for Disease Resistance in Crop Plants

Recent advances in plant-pathogen interactions, combined with progress in biotechnology, have offered novel strategies for engineering disease-resistant crops. Several of these strategies have already demonstrated success under field conditions (Table 3). Unlike the introduction of dominant resistance genes, which may drive the evolution of resistant pathogen strains, targeted disruption of host susceptibility factors using CRISPR/Cas represents an alternative approach for developing sustainable crop protection.

Table 3.

CRISPR/Cas gene-editing in crops for disease management.

Crop plants	Target disease	Target gene	Mutation Type	References
Rice	Xanthomonas oryzae pv. oryzae	SWEET11, SWEET13, SWEET14	Knockdown	66–68
		OsERF922	Knockdown	69
		Xa13	Knockout	70
		Bsr-d1, Pi21 and ERF922	Knockout	71
Wheat	Blumeria graminis f. sp. tritici	EDR1	Knockout	72
		MLO	Knockout	73
		Tamlo-R32	Knockout	74
	Puccinia striiformis f. sp. tritici	TaCIPK14	Knockout	75
		TaCIPK14	Knockdown	76
	Polymyxa graminis	TaPDIL5-1	Knockdown	77
		Importin-a	Knockout	78
		TaNFXL1	Knockout	79
Maize	Maize rough dwarf disease	ZmGDIa-hel	Knockout	80
		ZmNANMT	Knockout	81
	Ustilago maydis	LOX3	Knockout	82
	Fusarium graminearum	jaz15	Knockout	83
Potato	Phytophthora infestans	StDND1, StCHL1 and StDMR6-1	Knockout	84
	Potato late blight	StCCoAOMT	Knockout	85
Soybean	Phytophthora sojae	GmTAP1	Knockout	86
		GmARM	Knockout	87

Rice

BLB, triggered by infection with Xanthomonas oryzae pv. Oryzae (Xoo), constitutes a serious constraint to sustainable rice production, usually resulting in a 10–20% yield loss in rice, and up to 50% or more in areas with severe occurrences.⁸⁸ During the infection process, Xoo releases effector molecules that induce the expression of specific SWEET genes in the host plant, thereby enhancing susceptibility to the disease. By employing promoter editing of diverse OsSWEET family members, researchers have successfully engineered rice cultivars exhibiting broad-spectrum resistance against leaf blight.^{66,67,89–91} Oliva et al.⁶⁷ achieved this by using CRISPR/Cas9 gene editing in three genes that are recognized by the pathogen TALE as SWEET11, SWEET13, and SWEET14, among other mutations that were introduced in the promotor of EBE gene, and the study showed that editing the known SWEET EBE would confer broad resistance to BLB in rice. Additionally, this study applied a multiplex genome-editing approach to systematically interfere with all major TALE EBEs that induce SWEET genes. This effort successfully produced Kitaake rice lines resistant to currently prevalent Xoo strains. Targeted editing of host susceptibility genes in rice with the CRISPR/Cas9 system conferred disease resistance by compromising the plant’s inherent susceptibility to pathogens. One such gene, OsERF922, encodes an ethylene-responsive transcription factor that is upregulated during infection by the rice blast pathogen, and its disruption has been reported to impart enhanced rice disease-resistance.⁹² The targeted disruption of the rice gene OsERF922 by CRISPR/Cas9 effectively enhanced blast resistance without compromising major agronomic characteristics. This breakthrough creates valuable new genetic material for crop improvement.⁶⁹

The induction of the susceptibility gene Os-8N3 (Xa13) by a TAL effector from Xoo is a critical step in BLB development.⁹³ Os-8N3 mediates the redistribution of toxic copper, thereby enabling bacterial colonization within the xylem, enhancing nutrient uptake, and promoting disease progression.⁹⁴ Li et al. ⁷⁰ used CRISPR/Cas9 to target and delete part of the promoter sequence of the Xa13 gene, including the pathogenic bacteria induced expression element, to obtain rice lines without affecting yield and with enhanced resistance to leaf blight, which offers a novel strategy for advancing molecular breeding. By targeting the Bsr-d1, Pi21, and ERF922 loci with CRISPR/Cas9, Zhou et al. ⁷¹ developed single and triple mutant pure lines. The study revealed that mutations in Pi21 and ERF922 not only enhanced rice blast resistance but also provided resistance to leaf blight, thereby establishing a rapid and effective pathway for engineering rice with dual resistance.

Wheat

Wheat powdery mildew, is a widespread and destructive airborne disease which caused by the biotrophic fungus Blumeria graminis f. sp. Tritici (Bgt).⁹⁵ Enhanced disease resistance 1 (EDR1) gene encodes a MAP kinase that suppresses defense responses against this disease.⁹⁶ CRISPR/Cas-mediated knockout of all three wheat EDR1 homologs resulted in plants with significantly enhanced resistance to the pathogen.⁷² Similarly, disrupting all three homologs of the mildew-resistance locus O (MLO) conferred broad-spectrum resistance against powdery mildew.⁷³ However, this mutant also showed early senescence and yield reduction, which limit its wide application in production. Recently, researchers reported a new Tamlo-R32 super-allele genotype in wheat that is both resistant to powdery mildew and does not cause yield loss, and the rapid creation of superior new wheat lines that combine broad-spectrum resistance to powdery mildew with high yields using CRISPR/Cas9.⁷⁴

Wheat stripe rust, induced by Puccinia striiformis f. sp. tritici (Pst), caused major biotic threat to global wheat production.⁹⁷ The TaCIPK14 gene encodes a member of the CBL-interacting protein kinase (CIPK) family, which functions in calcium-mediated signal transduction.⁹⁸ As a specific component of this signaling pathway, TaCIPK14 participates in regulating wheat’s defense mechanisms against biotic stress. Wang et al. ⁷⁵ identified the susceptibility gene TaPsIPK1 (Puccinia striiformis-Induced Protein Kinase 1) in wheat for the first time, which is targeted and hijacked by pathogen effectors. They subsequently generated a knockout mutant of this gene using CRISPR/Cas9 technology. The edited plants exhibited broad-spectrum resistance against both stripe rust and leaf rust pathogens without adversely affecting key agronomic traits. Research has confirmed that stripe rust resistance in wheat is substantially increased through CRISPR/Cas9-mediated suppression of TaCIPK14, thereby highlighting its importance in plant immunity.⁷⁶ Utilizing this susceptibility gene for wheat improvement departs from the traditional paradigm of breeding with R genes. This breakthrough opens an innovative avenue for developing disease-resistant wheat varieties through modern biotechnological breeding.

Wheat yellow mosaic virus (WYMV), a soil-borne pathogen transmitted by the protist Polymyxa graminis, poeses a serious threat to wheat production.⁹⁹ Infection typically leads to yellow streaking on leaves, stunted growth, and considerable yield loss.¹⁰⁰ Kan et al. ⁷⁷ utilized CRISPR/Cas9 technology to simultaneously edit three copies of the TaPDIL5-1 gene across the A, B, and D subgenomes of hexaploid wheat, thereby conferring resistance to yellow mosaic disease. This study confirmed the first known gene conferring resistance to wheat viral diseases. In 2024, Wang et al. ⁷⁸ demonstrated the important role of wheat Importin-a gene in the pathogenicity of BYDV-GAV, and created a series of Importin-a knockout mutants by CRISPR/Cas9 technology, from which they identified two new wheat germplasm with significantly enhanced resistance to yellow dwarf disease, which have potential breeding applications.

Fusarium head blight (FHB) severely threatens wheat production, with the disease being triggered by a complex of pathogens primarily from the Fusarium genus.^101,102 This disease results in substantial reductions in wheat yields, as infected grains become shriveled and shrunken, exhibiting diminished quality and weight, whilst also containing mycotoxins harmful to human and animal nutrition.¹⁰³ Deoxynivalenol (DON), a mycotoxin produced by Fusarium species, supports fungal colonization within wheat floral tissues.⁷⁹ Studies indicate that exposure to DON can upregulate certain plant genes, including TaNFXL1.⁷⁹ Disruption of TaNFXL1 using CRISPR/Cas9 enhanced resistance to fungal infection, indicating its applicability as a viable approach to breed disease-resistant wheat varieties.⁷⁹

Maize

Maize rough dwarf disease (MRDD), a widespread and destructive viral disease affecting maize, leads to substantial crop losses.¹⁰⁴ The recessive resistance allele against this disease has been identified and functionally characterized.¹⁰⁵ By creating null mutants via CRISPR/Cas9, Liu et al. ⁸⁰ demonstrated that the resulting lines exhibited significantly enhanced MRDD resistance compared to the natural ZmGDIα-hel allele. Recently, Li et al.⁸¹ demonstrated that ZmNANMT functions as a novel susceptibility gene in maize. CRISPR/Cas9-mediated editing of this gene quantitatively enhanced MDR without compromising agronomic performance. The biotrophic fungus Ustilago maydis triggers tumorigenic growth throughout the aerial architecture of maize plants. Pathi et al. ⁸² validated LIPOXYGENASE 3 (LOX3) as a key susceptibility factor for Ustilago maydis by creating LOX3 mutants via Cas endonuclease technology, which led to a significant decrease in infection susceptibility upon pathogen challenge. Gibberella stalk rot (GSR), induced by Fusarium graminearum, is a major cause of maize yield reduction worldwide.¹⁰⁶ Through CRISPR/Cas9-mediated knockout and Mutator insertion, Ma et al. ⁸³ found that mutation of coi1a conferred resistance to GSR, whereas disruption of jaz15 enhanced susceptibility. These successful applications demonstrate the efficacy of engineering disease resistance in maize through the targeted disruption of susceptibility genes.

Potato

Potato late blight (PLB) caused by the oomycete Phytophthora infestans, remains a destructive disease worldwide.^107–109 Tiwari et al. ¹¹⁰ identified several resistance genes in potato, including R3a, RGA2, RGA3, R1B-16, Rpi-blb2, Rpi, and Rpi-vnt1. Modifying host S genes also offers a promising route to enhance resistance against this pathogen.¹¹¹ Kieu et al. ⁸⁴ edited seven S genes (StMLO1, StHDS, StTTM2, StDND1, StCHL1, StDMR6-1 and StDMR6-2) in tetraploid potato “Desiree” and “King Edward” using CRISPR/Cas9 technology, and found that there were differences in editing efficiency among different genes with the highest StCHL1 (18%) and the lowest StHDS (0%); The efficiency of 1 to 3 alleles edited ranged from StDND1 (1.5%) to StCHL1 (58%). Disruption of StMLO1, StHDS, StTTM2, or StDMR6-2 did not significantly improve late blight resistance; however, the abolition of all four alleles of StDND1, StCHL1, or StDMR6-1 did so markedly. From the analysis of the growth phenotypes of the mutant after gene knockout, the phenotypes of StMLO1, StTTM2, StCHL1, and StDMR6-1 were consistent with those of the wild type, and there was no obvious change. Heterozygous StHDS mutants begin to show albino seedlings and stop growing after 2 weeks; Homozygous StDND1 mutants showed slow growth accompanied by thinning stem elongation and shrinking leaves. StDMR6-2 mutant plants became shorter, fresh weight decreased, and more importantly, the “King Edward” background mutant had no obvious red eyes than the wild type. The CRISPR/Cas9 knockout of StCHL1 and StDMR6-1 therefore presents a promising strategy for engineering late blight-resistant potato varieties with negligible impact on normal growth. A four-year field study demonstrated that CRISPR-edited potato plants with disrupted STDMR6-1 gene confer durable, broad-spectrum resistance against late blight, early blight, and common scab, while also showing improved tolerance to drought and salinity, all without yield penalty or quality trade-offs.¹¹² These findings establish STDMR6-1 mutation as a promising strategy for sustainable potato cultivation.

The StCCoAOMT gene encodes the enzyme caffeoyl-CoA-O-methyltransferase (CCoAOMT), which is critical for lignin synthesis in the potato.¹¹³ Cultivars susceptible to late blight, such as Russet Burbank, exhibit polymorphism due to a single nucleotide polymorphism (SNP) mutation on the StCCoAOMT gene, resulting in premature termination of transcription.⁸⁵ Carboxyl terminal truncated proteins cannot perform normal functions.⁸⁵ RNA-seq data provided a specific location of the SNP variation within this gene, which was precisely repaired by CRISPR/Cas9 homologous recombination, enabling the StCCoAOMT gene to encode a functioning cafeyl-coenzyme A methyltransferase. The repair of StCCoAOMT gene improves the transcription level of downstream biosynthesis-related genes, resulting in a substantial thickening of cell wall. After inoculation, the expression of multiple phenylalanine pathway genes and accumulation of disease-resistant metabolites in the mutant strains significantly increased, thus significantly inhibiting the growth of late blight lesions and enhancing the resistance levels of the varieties against late blight.⁸⁵

Soybean

Soybean production around the world is challenged by increasing negative impacts of fungus, bacterium, phytoplasma, nematode, and virus infections. Zhao et al. ¹¹⁴ investigated nucleotide fixation of pathogen resistance in wild and cultivated varieties, and their study revealed that Glyma20g08290 (homolog of Arabidopsis thaliana RPM1 gene) is a naturally selected locus, which is associated with Pseudomonas syringae in soybean ¹¹⁵ and found in wild soybean varieties.^114,115 Marker-assisted selection provides causal QTLs for vertical and horizontal resistance. It unravels major R genes, which maintain vertical resistance in soybean to soybean cyst nematode (Rhg), Phytophthora root and stem rot (Rps), soybean rust (Rpp), frog eye leaf spot (Rcs), bacterial blight (Rpg), and soybean mosaic virus (Rsv and Rsc).^{114,116–118} Horizontal resistance is controlled by multiple minor effect genes and confers resistance against many soybean diseases such as sudden death syndrome, Sclerotinia stems rot, root-knot nematode, and most Pythium species.¹¹⁸ Phytophthora soyae is the primary pathogen responsible for soybean root rot. Upon infecting soybean plants, this pathogen secretes numerous toxic effectors that hijack host targets to facilitate its own invasion. Genetically engineering these host targets represents an effective strategy for enhancing soybean resistance to root rot. Liu et al. ⁸⁶ firstly enhanced soybean resistance to root rot disease by knocking out targets of Phytophthora effectors using CRISPR/Cas9 gene editing technology. It has successfully generated novel germplasm materials resistant to root rot disease, holding significant practical value for achieving molecular design breeding of disease-resistant soybeans. Luo et al. ⁸⁷ identified the stress-response gene GmARM and generated its CRISPR-Cas9 mutants, which exhibited significantly enhanced tolerance to salt, alkali, and Phytophthora sojae. Further analysis revealed that the GmARM mutation also modulates the expression of related resistance genes, demonstrating its role in conferring broad-spectrum stress resistance.

The CRISPR/Cas technologies enable targeted disruption of host – pathogen interactions, thereby enhancing broad-spectrum disease resistance. By directly modifying susceptibility alleles, this approach can introduce pathogen resistance without the need for lengthy breeding cycles. Further expanding its potential, base editing and prime editing technologies are emerging as prospective strategies for engineering disease-resistant crops.^119,120 Nevertheless, several constraints currently limit its broader implementation. Key factors influencing editing efficiency include PAM specificity, sgRNA design, promoter selection, and off-target effects.¹²¹

Genetic Improvement for Herbicide Resistance in Crop Plants

Addressing the global increase in weed pressure, the development of herbicide-resistant germplasms represents a practical means to secure crop yields and minimize land degradation. CRISPR/Cas gene editing technologies now enable the precise development of such crops, as illustrated in Table 4. The most documented herbicide tolerance traits in agricultural biotechnology are those targeting the key enzymes acetolactate synthase (ALS), acetyl-CoA carboxylase (ACCase), and 5-enolpyruvylshikimate-3-phosphate synthase (EPSPS). Among these, resistance to ALS-inhibiting herbicides appears particularly common, representing a substantial proportion of known weed resistance cases. As the first enzyme in the biosynthetic pathway for branched-chain amino acids, ALS is inhibited by several herbicide families, including sulfonylurea (SU), imidazolinone (IMI), pyrimidinylthiobenzoates (PTB), sulfonylaminocarbonyltriazolinones (SCT), and triazolopyrimidines (TP).¹⁵¹ ACCase is a crucial enzyme in lipid biosynthesis and another valuable herbicide target. By disrupting fatty acid synthesis, ACCase-inhibiting herbicides like aryloxyphenoxypropionate (APP), cyclohexanedione (CHD), and phenylpyrazoline (PPZ) ultimately cause plant death.¹⁵² Similarly, the broad-spectrum herbicide glyphosate functions through the inhibition of EPSPS, a pivotal enzyme in the aromatic amino acid biosynthesis pathway. Conventional transgenic methods confer herbicide resistance by introducing foreign genes like bar into crops. By contrast, CRISPR/Cas achieves endogenous resistance through precise editing of herbicide target genes, offering a faster, more flexible, and transgene-free pathway.

Table 4.

The herbicide-resistant plants generated by CRISPR/Cas gene-editing.

Crop plants	Target weed	Target gene	Repair pathway	References
Rice	ALS- inhibiting herbicide	ALS	CBE	122–124
			HDR	125–129
	ACCase-inhibiting herbicide	ACCase	ABE	130,131
	Glyphosate	EPSPS	NHEJ	132
			HDR	133
			Base editing	134
		TubA2		135
	Trifluralin	AFB4	HDR	136
	Butafenacil	MPK6	HDR	137
	Herboxidiene	HPPD	HDR	138
Wheat	Sulfonylurea-, Imidazolinone-, Aryloxyphenoxy, Quizalofop	ALS, ACCase	Base editing	139
Maize	Glufosinate	ALS	HDR	131
	Quizalofop, 2,4-D	ACCase	HDR	140
	chlorsulfuron	ALS2	CRISPR/Cas9	141
	/	ALS	PE	142
Potato		ALS	CRISPR/Cas9 or TALEN	143,144
			TALEN	145
	chlorsulfuron		CBE	146
			Prime Editing	147
	glyphosate	CP4-EPSPS	HDR	148
Soybean	AHAS-inhibiting herbicide	GmAHAS	CBE	149
	chlorsulfuron	ALS1	CRISPR/Cas9	150

Rice

Researchers applied CBEs to design and introduce precise C-to-T base substitutions in OsALS.^122–124 This strategy successfully generated herbicide-resistant rice lines while preserving the essential enzymatic activity of ALS.^122–124 Previous studies have used CRISPR-Cas9 with the HDR pathway to develop herbicide-resistant rice. By introducing specific mutations at sites 548 and 627 in the ALS gene, they created new rice varieties that are resistant to bispyribac-sodium.^125,126 Subsequently, a chimeric Cas9-VirD2 protein was engineered to improve HDR efficiency in plants by leveraging VirD2’s strand-cleavage activity at Ti plasmid borders. This system successfully achieved precise modification of OsALS alleles, producing herbicide-resistant rice.¹²⁷ Similarly, ALS inhibitor resistance was alternatively achieved in rice using cgRNA to direct precise edits via homologous recombination.¹²⁸ Researchers also successfully replaced the wild-type ALS gene in rice with an edited variant harboring two specific point mutations to confer herbicide resistance utilized the CRISPR/Cpf1 system.¹²⁹ Early efforts, however, were constrained by the generally low efficiency of homology-directed repair in plants. Subsequent optimizations to the CRISPR/Cpf1 platform significantly increased editing efficiency, enabling reliable production of herbicide-resistant rice lines.¹⁵³

ABE was utilized to generate a C2186R mutation in the OsACCase gene, yielding rice strains resistant to the herbicide haloxyfop-R-methyl.¹³⁰ Meanwhile, the dual base editing system (STEMEs) was developed, which combines cytidine and adenosine deaminase activities to enable efficient saturated mutagenesis of endogenous genes.¹³¹ This dual base-editing system introduced diverse mutations across the rice ACCase gene, enabling the discovery of variants that confer tolerance to the herbicide haloxyfop.¹³¹ Applying this system to the carboxyltransferase domain of the rice ACCase gene, researchers designed 141 sgRNAs to screen for resistance-conferring mutations. This approach identified a novel W2125S mutation through CRISPR-mediated directed evolution, that confers resistance to APP herbicides in rice.¹³¹

To confer glyphosate resistance in rice, previous studies have utilized the NHEJ pathway to introduce TIPS double amino acid substitutions into the OsEPSPS gene, which successfully resulted in the intended trait.¹³² Similarly, Sony et al. ¹³³ employed CRISPR/Cas9 to create glyphosate-resistant (GR) rice lines. Their strategy involved fragment knockout and knock-in of an HDR template carrying specific mutations to achieve site-specific amino acid substitutions and modification of the phosphoenolpyruvate-binding site in the native OsEPSPS gene. Moreover, to achieve more efficient and versatile base editing, researchers created STCBE-2, a system that combines different deaminases with nCas9-NG.¹³⁴ This surrogate editor enables enhanced C-to-T and A-to-G conversion across an expanded genomic window. The application of this system enabled near-saturated mutagenesis of the OsEPSPS gene, which resulted in reliable glyphosate tolerance in rice.¹³⁴ Reports also indicate that TubA2, AFB4, MPK6, and 4-hydroxyphenylpyruvate dioxygenase (HPPD) can respectively impart tolerance against glyphosate, trifluralin, butafenacil, and herboxidiene (GEX1A).^135–138

Wheat

Herbicide-tolerant wheat germplasms have also been generated through base editing of ALS and ACCase genes, conferring resistance to SU, IMI, and APP herbicide classes.¹³⁹ Separately, quizalofop resistance was introduced in wheat by creating a targeted A1992V substitution in the ACCase gene.¹³⁹

Maize

Early CP4-EPSPS transgenic maize lines exhibited glyphosate resistance but suffered from reduced male fertility.¹⁵⁴ To circumvent this, the first-generation GA21 trait employed a rice actin promoter to drive a modified maize EPSPS gene.^155,156 Successive development led to the NK603 event, which contained two optimized CP4-EPSPS copies and conferred robust tolerance across all growth stages. This trait was later stacked with glufosinate and insect resistances for commercial use.¹⁵⁶ Separately, glufosinate-resistant maize was marketed from 1996 onward, eventually combined with glyphosate resistance as a “double-stacked” product.^156, To confer herbicide resistance, researchers have also utilized CRISPR/Cas9-mediated base editing platforms to introduce precise point mutations into endogenous genes like ZmALS1 and ZmALS2.¹³¹ Additionally, Svitashev et al.¹⁴¹ edited the endogenous ALS2 gene in maize, resulting in P165S modifications that conferred resistance to chlorsulfuron. Jiang et al.¹⁴² utilized a PE system targeting the ZmALS1 and ZmALS2 genes, generating maize HR lines that harbored the W542L and/or S621I mutations.

2,4-D is a selective auxin herbicide that primarily affects broadleaf weeds. Resistance to both 2,4-D and a class of ACCase-inhibiting “fop” herbicides (APPs, such as quizalofop) can be conferred by specific aryloxyalkanoate dioxygenase (AAD) enzymes.¹⁴⁰ These AADs metabolize both herbicide classes by targeting a common chemical bond shared between them. Concerns regarding the proliferation of GR grass weeds ahead of the commercialization of dedicated 2,4-D-resistant plants consequently prompted the development and promotion of 2,4-D-resistant maize as a strategy to manage GR grasses.

Potato

Research on herbicide-related genes in potato has primarily focused on assessing transformation efficiency or developing new gene-editing methodologies. For instance, Butler et al.^143,144 introduced mutations into the ALS1 gene using CRISPR/Cas9 or TALEN combined with a geminivirus replicon (GVR) as a donor template. The efficacy of transient TALEN expression was further confirmed in another study, where this non-integrating platform mediated efficient gene editing in tetraploid plant cultures, targeting the same locus.¹⁴⁵ Similarly, researchers have applied a CBE delivered via Agrobacterium to modify the ALS gene in potato, achieving precise cytidine substitutions and generating chlorsulfuron-resistant plants.¹⁴⁶ Perroud et al.¹⁴⁷ successfully employed prime editing in potato to introduce specific nucleotide transversions in the ALS gene. Beyond ALS-targeting efforts, several studies have aimed to confer glyphosate resistance. Bakhsh et al. ¹⁴⁸ developed GR potato cultivars by integrating a bacterial CP4-EPSPS gene. This enzyme, derived from Agrobacterium, exhibits structural insensitivity to glyphosate, thereby enabling herbicide tolerance.

Soybean

Soybean is easily distributed by weeds because the seeds are planted at a wide interval to form branches and to allow the canopy to expand fully in the later stage of growth. Weeds cause the largest yield loss in soybean production. HDR-meditated gene editing shows great potential in developing HR crops. Li et al.¹⁵⁰ applied a Cas-gRNA-directed soybean system to target the GmALS1 gene and successfully obtained soybean plants with a P178S mutation for chlorsulfuron resistance. Subsequently, Wei et al.¹⁴⁹ used a CBE system to target the GmAHAS4 gene and generated a novel AHAS-inhibiting resistance soybean through a P180S substitution.

A key challenge in developing herbicide-resistant crops lies in identifying new genetic targets suitable for the CRISPR/Cas system. So far, however, little research has been dedicated to discovering and characterizing such genes. To date, successful generation of herbicide-resistant germplasms has been largely confined to crops resistant to ALS-inhibiting, ACCase-inhibiting, EPSPS-inhibiting, and glyphosate-based herbicides. By contrast, resistance to herbicides that inhibit HPPD or protoporphyrinogen oxidase remains less advanced, with limited studies on their broad applicability and efficacy in weed management.

Biosafety Aspects of Genetically Engineered Crops

Unintended Off-Target Effects

In CRISPR-mediated editing, two principal forms of off-target alterations can occur: sgRNA-dependent and sgRNA-independent off-target edits.¹⁵⁷ sgRNA-dependent off-target edits arise when the editing machinery acts at genomic sites that are not perfectly complementary to the sgRNA sequence. According to whole-genome sequencing analyses, the CRISPR/Cas system appears to cause no substantial sgRNA-independent off-target mutations in plants such as rice and cotton.^158,159 In contrast, certain CBEs have been shown to promote widespread sgRNA-independent off-target mutations in rice.^160,161 These mutations are attributed to the uncontrolled activity of cytidine deaminase on single-stranded DNA regions distributed throughout the genome. The nature of off-target risk differs significantly between simple and complex genomes. Off-target edits in diploid crops like rice are typically random and distributed across the genome, increasing the risk of disrupting unrelated genes with unpredictable consequences. In contrast, for polyploid crops such as hexaploid wheat, a common type of “off-target” activity is the predictable editing of homologous genes, functional copies that live on different subgenomes. As noted in the studies in barley, brassica, rice, and wheat, off-target modifications are frequently found in protein-coding regions that are homologs of the target gene. This is an expected result given the high sequence conservation among these homologues. As a result, while the number of off-target sites in complex genomes may be higher due to redundancy, the associated biological risk is often more predictable and manageable, as it is primarily achieved through multiplex editing of genes with similar functions rather than random mutation. The primary challenge in polyploids shifts from preventing random off-targets to intentionally designing for or against the concurrent modification of these homologous copies.

One approach to enhance the specificity of CRISPR/Cas systems involves directly engineering the Cas enzyme, such as converting key catalytic residues in the wild-type protein from an endonuclease to a nickase, which requires two adjacent gRNAs for activity, thereby reducing off-target mutations. Alternatively, guide RNA (gRNA) design can be optimized, for instance, by truncating the 5’ end by 2–3 base pairs. Other effective methods include the transient expression of CRISPR/Cas9 components, as demonstrated in wheat and maize,^162,163 the application of rationally designed guide RNAs,¹⁶⁴ and the use of engineered high-fidelity variants of Cas9, Cas12a, and deaminases,^139,165 These techniques collectively minimize off-target effects and, in some cases, lower the integration rate of exogenous DNA. However, Li et al.²⁸ found that mutating the OsJAZ10 gene in the JA signaling pathway via CRISPR-Cas9 genome editing did not affect classical JA signaling. However, one type of mutant harboring an INDEL generated a novel frameshift protein, designated FJ10 (JAZ10 frameshift mutant protein), which exhibited enhanced rice growth and improved resistance against C. medinalis. This work first investigated into whether proteins produced by frameshift mutations possess biological function, and served as a reminder to researchers that when employing gene editing to mutate genes, they must select the correct mutation type to ensure that the mutant’s phenotype is indeed caused by the mutation in that specific gene.

To improve targeting precision of CRISPR/Cas system, diverse systems are being developed. These include engineering novel CRISPR/Cas systems and optimizing existing editors such as BEs. BEs are fusion proteins composed of a catalytic deaminase domain and a DNA-binding module, capable of converting adenine or cytidine without producing DSBs.¹⁶⁶ By avoiding DSB formation, these editors reduce the incidence of random indels at on-target sites and lower off-target effects.^13,167 Additionally, xCas9 and Cas9-NG,^168,169 Cas9 Nickase (Cas9n) and Dead Cas9 (dCas9)^170–173 also are under development to facilitate the precise genetic changes critical for improving crop traits. In addition, researchers have developed numerous methods for identifying off-target effects. For example, whole-genome sequencing (WGS) provides an unbiased approach for comprehensively assessing genomic alterations.¹⁷⁴ This technique has been widely used to profile Cas9-induced off-target mutations in diverse plant species, such as Arabidopsis,¹⁷⁵ rice,¹⁷⁶ tomato¹⁷⁷ and cotton.¹⁵⁸ Additionally, many algorithms and tools, such as Primer-Extension-Mediated Sequencing (PEM-seq),¹⁷⁸ CRISPR-PLANT v2,¹⁷⁹ CCTop,¹⁸⁰ CROP-IT ¹⁸¹ and so on (Table 5), have been employed to predict off-target effects. Concurrently, many highly specific Cas9 proteins have been successfully developed to mitigate their nonspecific editing activity. In most cases, selecting sgRNAs with lower predicted off-target rates and relying on high-quality reference genomes can effectively minimize off-target mutations.¹⁷⁷

Table 5.

Algorithms for detection of off-target effects.

Algorithm	Description	Web source	References
PEM-seq	Detection of off-target effects. Simultaneously determines the editing efficiency and specificity of CRISPR/Cas9	–	178
CRISPR-PLANT v2	CRISPRPLANT v2 detects every off-target	http://www.genome.arizona.edu/crispr2/.	179
CCTop	Employs position-dependent weight coefficients in their off-target scoring algorithms	http://crispr.cos.uni-heidelberg.de.	180
CROP-IT	Scoring potential off-target sites by the division of protospacer into three segments with weight coefficients optimized/rained with ChIP-Seq data	http://www.adlilab.org/CROP-IT/homepage.html.	182
CHOPCHOP	Rapid and easy selection of the optimal CRISPR/Cas9 or TALEN target sequences in genes from various organisms	https://chopchop.cbu.uib.no/., https://chopchop.rc.fasharvard.edu.	183
CHOPCHOP v2	Web-based tool for GE based on TALEN and CRISPR. A powerful and intuitive tool that serves both beginners and experienced users	http://chopchop.cbu.uib.no.	184

Following the selection of a reference genome, key steps involve selecting a suitable sgRNA design tool and establishing an efficient delivery system for introducing functional genome editing molecules into target cells while minimizing somatic mutations. Several software tools are now available for designing highly specific sgRNAs with minimal off-target risk. For instance, CRISPR-P 2.0 is capable of designing sgRNAs for a comprehensive suite of plant genomes, including numerous major crops like rice, cotton, maize, and wheat, for which high-quality genome assemblies are available.¹⁸⁵ The establishment of globally standardized, prescriptive protocols for off-target detection in gene-edited crops remains elusive, with a clear divergence between regulatory philosophies and scientific best practice. While international bodies like the OECD and Codex Alimentarius endorse a flexible, “case-by-case” risk assessment based on a weight-of-evidence approach, a powerful de facto scientific standard has emerged through industry and academic consensus. This standardized workflow is inherently tiered: it mandates rigorous in silico prediction and high-fidelity nuclease use for all projects, triggers targeted deep sequencing only when bioinformatics flags high-risk sites, and ultimately relies on phenotypic screening across breeding generations as the final filter for deleterious effects. Consequently, national guidelines range from the US model, which treats off-target analysis as a developer’s due diligence rather than a regulatory requirement, to the EU’s EFSA, which may explicitly request such data for complex edits. Thus, the current landscape is defined not by a universal checklist but by a scientifically rigorous, risk-proportionate framework that developers are expected to follow to demonstrate product safety and meet evolving, albeit varied, regulatory expectations.

Exogenous Gene Residues

CRISPR/Cas9 constructs are typically introduced into plants via Agrobacterium-mediated transformation.¹⁸⁶ The resulting transformed plants are then screened to identify mutant lines. When using editing vectors that carry sgRNA and Cas9, fragments of the vector may be randomly inserted into the plant’s genomic DNA, resulting in the presence of exogenous genes.⁶⁵ The presence of exogenous gene components can cause a series of unforeseen and difficult to solve biosafety problems. For example, random insertion of exogenous genes can disrupt endogenous genes or regulatory elements, complicating phenotypic analysis and potentially compromising plant fitness. Furthermore, constitutive expression of Cas nuclease increases the duration of exposure, elevating the risk of off-target mutations.¹⁸⁷ Researchers can screen and identify plants free of exogenous gene components in progeny species like rice, but in some crops that do not undergo genetic segregation through sexual generations, such as potato, it is not possible to avoid the phenomenon of exogenous gene residues. To address these challenges, various delivery methods have been developed and optimized, each with distinct advantages and limitations (Table 6). To enable DNA-free genome editing, a transitory transformation strategy for Agrobacterium-mediated CRISPR/Cas9 delivery was developed. Chen et al. ¹⁸⁸ reported 47.5% editing efficiency in tobacco without antibiotic selection, with 8% of mutants being DNA-free. Huang et al.¹⁸⁹ employed a CBE/Cas12a co-editing system, targeting both the ALS gene for herbicide resistance and a gene of interest, achieving 1.9–42.1% DNA-free mutants among herbicide-resistant lines. Similar transient approaches in potato and poplar yielded 10% and 7% transgene-free mutants, respectively.^146,190 These advances highlight the promise of Agrobacterium-mediated transient transformation for DNA-free editing, particularly in vegetatively propagated, genetically complex, or slow-maturing species. However, major technical challenges remain in efficiently recovering transgene-free edited plants without antibiotic selection.

Table 6.

Comparison of CRISPR/Cas system delivery methods.

Delivery Method	Representative Species	Delivery Form/Cargo	Typical Editing Efficiency	Advantages	Disadvantages	References
Agrobacterium-mediated transient	Tobacco, potato, poplar	T-DNA (transient expression)	1.9–47.5%	No protoplast regeneration required; Established protocols; works in asexually propagated/perennial species; Non-antibiotic enrichment via ALS herbicide resistance	Labor-intensive screening; T-DNA integration risk cannot be excluded; Selection requires target sites amenable to gain-of-function mutations	146,188–190
Particle bombardment	Wheat, maize	DNA plasmid	1.0–9.5%	Host-independent; works in recalcitrant cereals; established transformation infrastructure	Random plasmid/chromosomal DNA insertions at target sites (>14%); large transgene arrays; genomic rearrangements; increased off-target risk	191–194
	Wheat	RNA	~1%	• No DNA integration risk • Safer than DNA delivery	• RNA instability • Requires high-quality RNA synthesis	193
	Wheat; Maize	Ribonucleoprotein (RNP)	4–5%	• No off-target effects detected • Rapid degradation of reagents • Reduced off-target effects compared to DNA	• Lower efficiency than DNA plasmid • Requires high-purity Cas9 protein and sgRNA • Equipment-intensive; tissue damage • Efficiency requires optimization	162,163
PEG-protoplast	Arabidopsis, tobacco, lettuce, rice, potato, cabbage, bok choy, carrot, wheat, grape, soybean, citrus	Ribonucleoprotein (RNP)	Not specified	• No foreign DNA involvement • Minimized off-target effects • Successfully regenerated in multiple species • Applicable to asexually propagated and perennial crops	Protoplast regeneration bottleneck; labor-intensive protoplast isolation; species-dependent; technical expertise required	195
Protoplast transfection (Non-PEG)	Cabbage, soybean	Ribonucleoprotein (RNP) – Electroporation	3.4–3.8%	Alternative to PEG; precise electroporation control; promising transformation strategy	No complete plants regenerated to date; efficiency lower than PEG; requires optimization	196,197
	Tobacco, citrus	Ribonucleoprotein (RNP) – Lipofection	Up to 6%	High delivery efficiency (66% with Lipofectamine 3000); gentle; low cytotoxicity	No complete plants regenerated; still in early development phase; cost of lipofectamine reagents	198,199
Negative-sense RNA virus	Nicotiana benthamiana, tobacco, tomato, chili pepper, sweet pepper, habanero pepper, ground cherry, peanut	Viral RNA (carrying Cas9 + sgRNA)	Effective (not specified)	Systemic delivery without DNA integration; carries both Cas9 and sgRNA; broad host range (TSWV: >1000 species)	Host range limitations; cargo size constraints; viral pathogenicity concerns; requires specialized virus handling	200,201
Geminivirus (DNA virus)	Tomato, tobacco, potato, wheat, rice, cotton, Arabidopsis	Viral DNA replicon	High	High loading capacity; efficient replication; high copy number; facilitates homologous recombination	Viral DNA integration into plant genome; difficult to eliminate viral particles in progeny; not optimal for DNA-free editing	202–207
Positive-sense RNA virus	Wide range (requires Cas9-OE plants: rice, tobacco, Arabidopsis, soybean, potato, tomato, wheat, corn)	Viral RNA (carrying sgRNA only)	Variable	Well-established VIGE platforms; broad host range; no DNA integration from virus; sgRNA systemic delivery	Requires Cas9 transgenic plants (stable transformation); not applicable to transformation-recalcitrant species; cannot deliver Cas9 (size limitation); limited meristem penetration	208–216
Positive-sense RNA virus + FT fusion	Cas9-OE plants	Viral RNA + FT-sgRNA fusion	Increased heritable mutation frequency	FT-sgRNA moves to meristem; overcomes meristem penetration limitation; enables heritable mutations	Still requires Cas9-OE plants; requires fusion construct optimization; two-component system	217,218
Positive-sense RNA virus	Nicotiana benthamiana	Viral RNA (carrying Cas9 + sgRNA)	Not specified	Simultaneous delivery of Cas9 and sgRNA; No DNA integration; fully DNA-free (100%); Multiple platforms (FoMV, PVX, BaMV)	Cargo capacity severely limited Host range restricted Requires further optimization Recent development; limited validation	214,219,220

Non-integrative viral vectors (e.g., tobacco mosaic virus or bean yellow dwarf virus) facilitate high-efficiency editing via systemic infection while typically avoiding genomic integration, allowing for the recovery of edited plants free of viral sequences.²⁰⁰ Additionally, transient delivery methods, such as the direct introduction of preassembled Cas9–sgRNA ribonucleoprotein (RNP) complexes via particle bombardment or polyethyleneglycol (PEG)-mediated transfection of protoplasts, allow efficient DNA cleavage without genomic traces, as the components are rapidly degraded.²²¹ Besides the PEG-mediated transient transfection method, electroporation and liposome transfection are also effective for introducing DNA into protoplasts,¹⁸⁷ it has been applied in cabbage,¹⁹⁶ soybean,¹⁹⁷ tobacco,¹⁹⁸ and citrus.¹⁹⁹ The strategic imperative to employ these DNA-free or non-integrative methods is most acute for clonally propagated crops like potato, cassava, and many fruit trees. For these species, the inability to purge unwanted vector sequences through genetic crossing makes the avoidance of integration not merely an ideal but a practical prerequisite for commercial and regulatory acceptance. However, significant technical hurdles remain. RNP delivery is often constrained by the need for efficient protoplast regeneration systems, which are underdeveloped in many recalcitrant species. Viral vectors, while powerful for systemic delivery, face limitations in cargo capacity (precluding the delivery of larger editors like prime editing systems) and host range. Consequently, the development of robust, standardized protocols to generate edits free of exogenous DNA in clonally propagated crops remains a critical and active frontier in plant genome editing.

Toxicity Impacts on Human/Animal Health

The toxicity associated with CRISPR/Cas application may be caused by its components, the exposure period, and/or depending on the delivery methods.²²² Since CRISPR was first used in plant cells, whole plants have been regenerated from edited tissues, indicating that CRISPR components are not inherently toxic to plants.²²³ However, depending on the CRISPR/Cas strategy and target gene, serious pleiotropic effects can occur.²²⁴ Knocking out plant S genes, for instance, may reduce pathogen susceptibility but often impairs normal growth and development.^84,225,226

Nevertheless, the mode of delivery of CRISPR/Cas system components seems to stand out as one of the main factors of toxicity in plants (Table 7). Current gene delivery strategies span from conventional tools like Agrobacterium transformation, abiotic methods, particle bombardment, and PEG to modern platforms such as engineered viral vectors and nanoscale carriers.²²⁷ In plant genetic transformation research, the predominant delivery vehicle relies on biological mediation. Despite its prevalence, this method is suitable for only a limited range of plant species and often proves inefficient at integrating exogenous genes into the host genome. Viral vectors offer high transfection efficiency and high throughput, but are oncogenic, prone to immune responses and limited by the size of Cas9-sgRNA. Particle bombardment is often delivered to plant species or cells with the aid of external forces, but this delivery method has limitations and often results in damage to plant tissues due to improper external manipulation. The advantages of non-viral vectors are low cytotoxicity, low insertion mutation rate and high biocompatibility, but the packaging efficiency and targeting phase rate need to be improved. Nanoparticles and similar novel materials demonstrate considerable potential for facilitating the delivery of genome-editing components,²²⁸ including mesoporous silica nanoparticle (MSN),²²⁹ carbon nanotube (CNT),²³⁰ layered double hydroxide (LDH),²³¹ and magnetite nanoparticle (MN).²³² MSN as a delivery vehicle to biologically deliver proteins to selected genes or reporter genes in plant cells avoids the creation of precision modified non-transgenic plants.²³⁰ CNT plays a marginal role in selectively delivering plasmid DNA to chloroplasts in different plants without the help of biology or chemistry.²³¹ Nontoxic, degradable LDH nanosheets can facilitate dsRNA delivery in tobacco cells. Utilizing magnetic-loaded nanoparticles, exogenous DNA can be delivered into pollen under a magnetic field, and the subsequent use of this magnetized pollen for pollination yields transgenic crops.²³² Nanoparticle technology is no longer restricted by the host, can selectively and efficiently deliver in the cells of plants according to the characteristics of the material, and can deliver biomolecules in a targeted way to the target to play their editorial role.²³³

Table 7.

Toxicity of delivery methods for CRISPR/Cas systems.

Delivery Method	Toxicity	Advantages	Limitations	Key Features	References
Agrobacterium-mediated transformation	Low cytotoxicity	Mature technology; widely used	Limited host range; inefficient integration in many species	Predominant delivery vehicle in plant research	227
Engineered viral vectors	Oncogenic; prone to immune responses	High transfection efficiency; high throughput	Cargo size limitation (Cas9-sgRNA); biosafety concerns	–	227
Particle bombardment (biolistics)	Tissue damage due to external mechanical forces	No host restriction; works in recalcitrant species	Equipment-intensive; improper manipulation exacerbates damage	Often requires external force assistance	227
Polyethylene glycol (PEG)-mediated transformation	Low cytotoxicity	Simple; low cost	Packaging efficiency and targeting efficiency need improvement	Protoplast-based; regeneration bottleneck	227
Mesoporous silica nanoparticle (MSN)	Low cytotoxicity; high biocompatibility	Delivers proteins; avoids creating transgenic plants	Packaging efficiency needs optimization	Enables precision-modified non-transgenic plants	228–230
Carbon nanotube (CNT)	Low cytotoxicity	Selectively delivers plasmid DNA to chloroplasts; no biological/chemical assistance required	Marginal role; efficiency to be improved	Works in different plant species	228,230,231
Layered double hydroxide (LDH)	Nontoxic; degradable	Facilitates dsRNA delivery	Targeting efficiency needs improvement	Demonstrated in tobacco cells	228,231
Magnetite nanoparticle (MN)	Low cytotoxicity	Delivers DNA into pollen under magnetic field; magnetized pollen yields transgenic crops	Requires magnetic field equipment	Enables host-independent, targeted delivery	228,232

Safety Evaluation System: Different Attitudes of Relevant Countries and Regions Toward CRISPR/Cas Gene Editing Technology

Compared to GM technology, gene editing emerged later. Consequently, most jurisdictions have adapted existing GM crop safety frameworks to accommodate their unique characteristics, resulting in hybrid assessment systems. Globally, two primary legislative models exist for gene-edited safety assessment as shown in Table 8: product-oriented evaluation (focusing on inherent crop traits) and process-oriented evaluation (regulating crop development methodologies). Compared to process-based assessments, product-centered approaches inherently possess greater regulatory constraints.

Table 8.

Regulatory policies for genome-edited crops in selected countries and regions (January 2026 update).

Country/region	Regulatory approach	Regulatory policies	Key regulatory principles	Approved crops
USA	Product-based	SECURE Rule (2020)	Exemption if foreign DNA-free and non-harmful traits	Soybean, Potato, Wheat, Maize, etc
Canada	Product-based	Plant Breeding Innovation Guidance (2022)	Trait-based risk assessment	Potato
European Union	Processed-based	Directive 18/2001/EC (2001), European Court of Justice 2018	Full GMO process under CJEU 2018 ruling	–
UK	Product-based	Genetic Technology Act 2023	Precision editing exempt, Simplified field trials	–
Australia	Product-based	Gene Technology Regulations 2021	Exemption SDN-1 edited products	Banana
China	Product-based	Safety Evaluation Guidelines (2022 trail)	Tiered management (type 1 simplified)	Soybean, Maize, Wheat, Cotton
Japan	Product-based	Genome-Edited Food Guidelines (2019)	Exemption for foreign DNA-free crops with disclosure	Tomato, Maize
India	Product-based	Revised regulations by MoEF&CC (2022)	SDN-1 and SDN-2 that do not contain foreign gene are exempt from GMO regulation	Rice
Philippines	Product-based	Rules and Procedures for Evaluating Plant-Breeding Innovations (2022)	Exemption for foreign DNA-free crops	Banana
Nigeria	Product-based	National Biosafety Guidelines for Gene Editing (2020)	Exemption if foreign genetic material-free	–
Kenya	Product-based	Guidelines for Determining the Regulatory Process of Genome Editing Techniques (2022)	Exemption if foreign genetic material-free	Maize, Sorghum
Malawi	Product-based	the Genome Editing Guidelines (2022)	Exemption for recombinant DNA-free crops	–
Ghana	Product-based	the Genome Editing Guidelines (2023)		-
Brazil	Product-based	Normative Resolution No. 16	Foreign DNA-free products may be exempt from regulation	Soybean, Sugarcane

USA

The United States actively promotes biotechnology development through a facilitative legal framework. Adopted in 1986, the Coordinated Framework for Regulation of Biotechnology governs both GM crops and CRISPR/Cas products under a unified evaluation system focused on crop traits rather than production methods. This framework assigns different regulatory functions to the United States Department of Agriculture (USDA), the Food and Drug Administration (FDA), and the U.S. Environmental Protection Agency (EPA).²³⁴ Specifically, the USDA is responsible for assessing whether novel gene-edited crops are effective in preventing agricultural pests and diseases; The EPA is responsible for evaluating whether any pesticide components potentially present in novel gene-edited crops demonstrate adequate safety for human health and the environment; The FDA is responsible for evaluating whether novel gene-edited crops, when used as food or feed, are as safe as traditionally bred crops. According to the Plant Protection Act, transgenic plants and microorganisms are subject to the Animal and Plant Health Inspection Service (APHIS) of the USDA.²³⁵

Generally, if there is a potential risk of plant pests, relevant regulations require that developers must first obtain permission before introducing certain gene-edited plants and microorganisms.²³⁶ In this case, APHIS has extensively interpreted the regulations, expanded its regulatory authority to all biotech products that may pose risks to plant health or agriculture, and finalized its own regulatory scope in the SECUR rule.²³⁷ The rule means sustainable, ecological, consistent, unified, responsible and efficient,²³⁷ which greatly affects the production and circulation safety evaluation system of gene editing crops in the United States. Under this regulatory framework, the USDA has successively announced that multiple gene-edited plants meet the criteria for exemption and are therefore exempt from regulatory oversight. For example, on 2 May 2023, APHIS announced that a gene-edited potato had met the criteria for exemption, which possesses anti-browning properties. On 8 April 2024, APHIS announced that 13 genetically edited soybean varieties meet the criteria for exemption. FDA issued an industry guidance document on premarket voluntary notifications for gene-edited plant foods on 22 February 2024. This guidance outlines the process by which companies may voluntarily engage with FDA prior to marketing gene-edited plant foods, encompassing premarket voluntary consultations and premarket voluntary meetings. This framework facilitates streamlined market access procedures for gene-edited plant foods while ensuring the FDA’s safeguards remain in place. On September 8, 2025, FDA granted authorization for the gene-edited maize variety VPM-0006, developed by Corteva, to be used in food and feed. This variety utilizes CRISPR/Cas system to editing the Wx-1 gene, thereby increasing its amylopectin content.²³⁸ This trait is of significant agricultural and industrial importance, as waxy maize starch offers superior properties for food processing (e.g., as a thickener and stabilizer) and is highly digestible in animal feed, potentially enhancing feed efficiency. By September 2025, FDA has authorized six gene-edited plants, including corn, rice, potato, canola, strawberry, and mustard greens.²³⁹ On 25 May 2023, EPA issued a notice and guidance clarifying the approved exemption conditions for two categories of gene-edited crops employing novel biotechnologies. The publication of these requirements signifies that EPA will refine its regulatory framework for plant-incorporated protectants (PIPs) to promote compliance within the gene-editing breeding sector. Unlike USDA, EPA requires both a thorough assessment of the safety of novel PIPs and proof that these PIPs can also be achieved through conventional breeding methods. That is to say, EPA still requires applicants to demonstrate that novel PIPs are not substantially different in safety from conventionally bred crops. To this end, EPA has also issued detailed regulatory exemption principles.

Canda

Canadian and US regulatory approaches are largely aligned. Canada’s regulations on GM crops exist in the Seeds Regulations, part V of which defines the criteria and regulatory agencies for plants with new traits, which are defined as any plant trait new to the Canadian environment with potential environmental or human health impacts, irrespective of their development through conventional, organic, or biotechnological methods, and requires Canadian Food Inspection Agency (CFIA) approval if its expression exceeds conventional varieties by ≥30%. Health Canada complements this framework through its 2021 Guidance for Novel Food Regulations, which addresses plant breeding techniques and incorporated public consultation with industry stakeholders and citizens. Both nations employ industry-oriented governance for gene-edited crops. While criticized for regulatory opacity, including concerns about undermining consumer right to know and public health protections, this approach has accelerated gene-editing research and agricultural sustainability in North America. In May 2023, CFIA released the latest version of its Seed Regulations Guide. Part V of the guide, updated by the CFIA, states that gene-edited seeds and plant material will no longer be classified as GM but treated as conventional crops. This will provide farmers with varieties better equipped to withstand extreme temperatures, precipitation, and insects, thereby aiding adaptation to climate change, feeding a growing population and reducing food costs for consumers. Furthermore, Canada will implement measures to enhance transparency regarding innovative plant breeding products and allocate funding for a review of the Canadian Organic Standards to safeguard the organic sector’s credibility. The guidelines stipulate that organic farmers may utilize conventional seeds but are prohibited from employing gene-edited seeds.

EU

Unlike the US and Canada, the EU prioritizes regulating crop production processes over end-product characteristics. Despite French scientist Emmanuelle Charpentier’s co-discovery of CRISPR/Cas9, the stringent regulatory framework of EU has led to significantly fewer CRISPR patents compared to China and the US. On 5 July 2023, the European Commission published a proposal to amend the GM crops regulation (EU 2017/625), regulating food, feed and other products containing, consisting of or produced from plants derived from new genome editing technologies (NGTs). The primary goals of the proposal are to ensure a high standard of protection for human and animal health and the environment, enable the advancement of a range of plant species, and cultivate a conducive environment for research and innovation, particularly for small and medium-sized enterprises. The proposal applies solely to NGT plants obtained through directed mutagenesis and cis-mutagenesis techniques. Plants obtained by introducing genetic material from non-interspecific species via NGT remain subject to the GMO Regulation. In January 2024, the proposal passed by a narrow majority of 307 votes in favor to 263 against but has yet to be implemented.

England

On March 23, 2023, the Genetic Technology (Precision Breeding) Bill was enacted into UK law following Royal Assent. This legislation authorizes the application of gene editing in England to enable precise, targeted modifications of an organism’s genetic code. The legislation also permits precision breeding in animals through techniques like gene editing, thereby shielding them from certain diseases. Such plants and animals produced via precision breeding will no longer fall under the UK’s regulatory requirements for GMOs. However, the bill stipulates the establishment of two separate information disclosure systems for research and commercial purposes respectively.

Australia

The Office of the Gene Technology Regulator (OGTR) in Australia primarily oversees the product development process. In 2016, the third amendment to the Gene Technology Act was launched, which stipulated that Site-Directed Nucleases (SDN)-1 products are exempt from regulation due to their similarity to natural mutations, whereas SDN-2 and SDN-3 products, which involve homologous recombination, significant alterations to the genome sequence, or the introduction of new genetic material, require oversight by relevant government departments.²⁴⁰ On 10 April 2019, Australia introduced new gene-editing regulatory rules, adopting a middle ground approach.²⁴¹ Australian authorities have exempted certain applications of gene editing from their regulatory framework. This policy applies to modifications in plant, animal, and human cellular subjects, provided the techniques refrain from incorporating any novel genetic elements. The updated guidelines specifically exclude techniques that rely on a cell’s natural repair mechanisms after DNA is cut at targeted locations, provided no engineered template is used to direct the repair. According to Australian regulatory authorities, gene editing techniques that do not involve template insertion are environmentally and biologically safe, showing no difference from naturally occurring mutations. In contrast, any genetic modification process that introduces external DNA templates or additional genetic elements will remain under the supervision of the Government Commissioner’s Office for Genetic Technology Regulation.

China

Like the mentioned countries and regions, China also regulates CRISPR-Cas gene editing and its products based on its existing transgenic technology legal framework. The current regulatory system, primarily established by the Biosafety Law, the Regulations on Agricultural GMO Safety Management, and related implementation measures (covering safety evaluation, labeling, and import management), relies on the Biosafety Law as its programmatic foundation. However, the system overall lacks specific provisions addressing the environmental and ecological impacts of GM technology, genetic contamination, and product safety assessments.

In January 2022, Ministry of Agriculture and Rural Affairs, PRC issued the Guidelines for Safety Evaluation of Gene-Edited Plants for Agricultural Use (Trial) (the Safety Evaluation Guidelines),^242, aligning with strategic interests of China and international legislative trends. A flow chart summarizing this evaluation process is presented in Figure 1. According to the Guidelines, the definition of gene-edited plants is “Plants and their products for agricultural production or agricultural product processing obtained by targeted modification of specific genomic sites using genetic engineering technology,”²⁴² explicitly excluding those with introduced exogenous genes. It shows that China has recognized the uniqueness of gene-edited crops at the legal level, distinguished gene editing from transgenic crops, and introduced safety assessment and declaration regulations suitable for the nature of gene-edited crops according to their own characteristics. Notably, the introduction of exogenous DNA refers to SDN3 crop products, indicating that the crop types applicable to the Safety Evaluation Guidelines are SDN1 and SDN2 crops. In addition, the Safety Evaluation Guidelines also stipulates the specific process of safety declaration and divides the declaration process into four situations: the cultivation goal will not increase the environmental safety risk and food safety risk, only increase the environmental safety risk, only increase the food safety risk, both increase the environmental safety risk and increase the food safety risk. The regulatory model of China represents a compromise between US and EU approaches, whose strength lies in adequately addressing public health risks for this developing technology while minimizing stakeholder errors and fraudulent practices; however, documentation and procedures still have streamlining potential. This differentiated regulatory approach is supported by scientific evidence showing that novel crops can be reliably assessed using established frameworks. For example, a safety evaluation of transgenic miR-14 rice successfully adapted Bt crop risk assessment protocols to confirm its safety for the non-target parasitoid Cotesia chilonis.²⁴³ Although this case involves a transgenic RNAi plant, it demonstrates that gene-silencing-based crops can be rigorously evaluated for ecological safety. Such precedents provide scientific justification for China’s policy of applying streamlined regulation to gene-edited crops (SDN1/SDN2) that contain no exogenous DNA.

Figure 1.

Safety evaluation process for gene-edited agricultural crops in China.

Japan

The Ministry of Health, Labour and Welfare (MHLW) and the Ministry of Agriculture, Forestry and Fisheries (MAFF) established the policy, which says that site-directed nucleases (SDNs), including CRISPR/Cas9, are not subject to the stringent pre-market safety assessment for conventional breeding. Developers of SDN-1 products must only submit a notification to the relevant authorities. An example is the high-GABA tomato, which was recognized as the first gene-edited tomato variety in the world.²⁴⁴ On 20 March 2023, Japanese authorities cleared the high-starch maize variety as the fourth genome-edited food, exempt from GMO rules.²⁴⁵ Gene-edited products developed using SDN-3 techniques are regulated under existing GM legislation. The policy exempts SDN-1-derived foods, which contain no recombinant DNA, from mandatory labeling requirements. This regulatory strategy aims to facilitate precision breeding, while maintaining oversight of novel genetic combinations.

India

On 30 March 2022, the Indian Ministry of Environment, Forest and Climate Change announced that genome-edited plants without foreign genes will no longer be classified as transgenic products.²⁴⁶ As defined in the Memorandum, SDN1 and SDN2 products that contain no exogenous DNA are excluded from mandatory biosafety evaluation (Rule 20 of the 1989 “Manufacture, Use, Import, Export and Storage of Hazardous Microorganisms/Genetically Engineered Organisms or Cells Rules”). On 17 May 2022, the final guidelines for the safety assessment of genome-edited plants were issued, outlining a roadmap for the technology’s development and sustainable application. This streamlined approach was demonstrated in May 2025 when India approved two indigenous gene-edited rice varieties, Pusa rice DST1 and DRR Dhan 100, for large-scale field trials, marking the country’s first approval for large-scale testing of gene-edited crops.²⁴⁷ It reflects a strategic shift to fast-track innovation in crop breeding to address climate change and food security challenges.

Philippines

The Philippines Department of Agriculture issued Memorandum Circular No. 08 Series of 2022,²⁴⁸ Rules and Procedures for Evaluating Plant-Breeding Innovations (PBIs), which took effect on 19 May 2022.²⁴⁸ This pivotal circular establishes a process whereby gene-edited plants deemed to lack foreign DNA and which could have been produced through conventional breeding may be exempted from the standard GM regulatory requirements, signaling a strategic shift toward a more product-based oversight approach that aims to streamline the innovation pathway while upholding biosafety commitments. The Bureau of Plant Industry in the Philippines decided that a reduced-browning banana developed by Tropic is not a GMO in April 2023.²⁴⁹ This banana variety was the first to pass the national regulatory review for gene-edited agricultural products.

African

With the promulgation of its National Biosafety Guidelines for Gene Editing in December 2020, Nigeria pioneered the establishment of such a regulatory framework in Africa. These guidelines apply to all experiments, commercialization, food, feed, processing, and import/export of gene-edited plants, animals, and microorganisms within Nigeria (excluding pharmaceuticals). Under the guidelines, gene-edited products containing recombinant DNA or novel genetic material are classified as genetically engineered products.

With the National Biosafety Authority’s publication of genome editing guidelines in March 2022, Kenya reached a regulatory milestone, becoming the continent’s second nation to establish such a framework. The new guidelines demarcate the regulatory pathway by specifying which edited products require oversight under the Biosafety Act versus those considered equivalent to conventional varieties.²⁵⁰ The National Biosafety Authority of Kenya stands as the sole African nation to have reached three decisions under the early consultation approach outlined in the Genome Editing Guidelines. These encompass gene-edited nitrogen-fixing bacteria, necrosis-resistant gene-edited maize, and Striga-resistant sorghum.

In August 2022, Malawi approved the Genome Editing Guidelines, fostering a biosafety regulatory environment.²⁵⁰ The guidelines review and evaluate genome editing applications and their products, providing a step-by-step procedure governing genome editing and its products, clarifying which products may be exempt from regulation as GMOs. According to the guidelines, any product with recombinant DNA is subject to GMO regulations. These provisions provide sufficient safeguards and ensure high biosafety standards. They also give clear guidance to regulators when assessing any genome editing application.

On 30 October 2023, Ghana’s National Biosafety Authority (NBA) issued the Genome Editing Guidelines, providing procedural guidance for potential applicants regarding genome-edited organisms and/or their product categories subject to regulation under the Biosafety Act 2011 (Act 831).²⁵¹ With this publication, Ghana became the fourth African nation to possess effective genome editing guidelines.

South America

On 23 March 2023, the National Technical Commission on Biosafety of Brazil (CTNBio) classified a drought-tolerant soybean developed using gene-editing technology as a conventional soybean.²⁵² Developed by the Brazilian Agricultural Research Corporation’s Soybean Research Institute, this soybean was engineered to possess drought-tolerant traits through CRISPR/Cas9 technology. Previously, CTNBio evaluated each product individually, with assessments tailored to its specific traits. In 2018, it issued Normative Resolution No. 16, declaring that gene-edited products containing no foreign genes are exempt from GMO regulations.²⁵³ Argentina, Chile, Colombia and other countries have adopted regulatory approaches relatively similar to Brazil’s, focusing on the final product for oversight. Evaluation is conducted according to the case-by-case analysis principle, whereby developers determine whether their product possesses novel attributes. Should the product involve DNA recombination and new traits, it automatically triggers regulatory scrutiny.

Regulations are needed to make sure crops are safe and validated. Including transparency and stakeholder engagement will help people accept these crops. Regulatory frameworks also need to attach importance to public acceptance of such crops. Effective outreach from the scientific community is crucial to encourage public acceptance of gene-edited crops. Addressing challenges associated with lack of technology, production, sales, safety supervision, and standards is vital to securing public confidence and, ultimately, consumer endorsement of gene-edited crops.

Conclusions and Perspectives

CRISPR-Cas technology has transformed plant biology through its capacity for highly precise and adaptable genome editing. These tools are now employed for diverse objectives, such as increasing tolerance to environmental and biological stresses, enhancing nutritional value, prolonging shelf life, and reducing post-harvest waste. The CRISPR toolkit is continuously being augmented through the development of novel nucleases, including sophisticated base and prime editors, alongside the AI-facilitated engineering of innovative Cas variants. These tools have generated hundreds of improved crop varieties and revolutionized breeding paradigms. The journey of gene-edited crops from research to commercialization follows a structured, multi-stage pipeline integrating scientific innovation with rigorous regulatory and safety oversight (Figure 2). Despite these advancements, challenges such as transformation and editing efficiency, genotype dependence, and regulatory hurdles remain. The efficiency of CRISPR/Cas systems varies considerably across different plant species, a challenge rooted not only in technical limitations but also in the complex biological interplay between the editing system and host plant. While high editing efficiency is routinely achieved in diploid model systems such as Arabidopsis and rice, many crops present substantial hurdles rooted in their unique genomic and cellular traits. For example, low transformation and regeneration rates in legumes like soybean create primary bottlenecks, while the dominance of NHEJ over HDR in maize biases outcomes toward indels rather than precise substitutions. A particularly defining factor is plant ploidy, which critically modulates the performance of advanced editing tools including BEs, PEs, and engineered Cas variants (e.g., xCas9). In polyploid species such as hexaploid wheat and tetraploid potato, genomic redundancy and allelic heterogeneity significantly constrain editing uniformity and efficiency. Studies demonstrate that BE4 editors exhibit markedly reduced and uneven editing across homeologs of the TaALS gene in wheat, while PE-mediated herbicide-trait integration in potato often yields chimeric plants due to inconsistent editing across alleles. Although variants like xCas9 offer broader PAM compatibility, they still struggle with specificity and equitable activity in polyploid backgrounds. This variability necessitates the optimization of both editing tools and delivery strategies. This includes advancing versatile systems such as prime editing, HDR-based gene targeting, and emerging methods like DNA polymerase editors, transposons, integrases,^254,255 click editing,²⁵⁶ and bridge editing,²⁵⁷ while also developing genotype-independent, transgene-free delivery approaches, such as biolistic RNA, to circumvent transformation recalcitrance, simplify operations, and address regulatory and public concerns. By integrating tailored molecular tool design with innovations in cellular and developmental biology, the potential of genome editing can be expanded across diverse plant species.

Figure 2.

The process of gene-edited crops from research to commercialization. The templates were obtained from biorender (https://www.biorender.com).

While CRISPR-mediated genome editing holds immense promise for crop improvement, the potential for off-target effects necessitates consideration. Current technologies can alleviate but not eliminate these effects, and the downstream significance of residual off-targets remains unclear. In plants, evidence for widespread, deleterious off-target activity is limited. This risk must be contextualized within the history of crop breeding, which has long utilized random mutations from conventional methods. A multi-faceted strategy addresses these challenges. This includes engineering high-fidelity nucleases and base editors, optimizing guide RNA design and delivery, employing a tiered evaluation protocol that ranges from in silico prediction to comprehensive whole-genome sequencing. The scientific community prioritizes phenotypic screening as the definitive validation of product safety. Moving forward, regulators should support research on off-target detection and unintended effects to establish safety assessments based on molecular and phenotypic data. Scientists must advance the technology and clearly communicate product benefits and progress in crop gene editing. These efforts are essential to improve public understanding and facilitate the responsible translation of edited crops to the field, ensuring their safety and acceptance in the global food system.

The evolving regulations and public perception of genome-edited crops continue to involve a complex relationship between scientific development, regulatory adaptation and societal views. The rapid development of CRISPR technologies presents a continuous challenge to global regulatory bodies, which are progressively shifting from rigid, blanket policies toward more nuanced, science-based frameworks. These emerging systems increasingly recognize the spectrum of edits possible, from those mimicking natural mutations to more complex alterations, necessitating a tiered and proportionate regulatory approach. Concurrently, while public awareness remains a work in progress, the generally more favorable perception of genome editing compared to conventional genetic modification offers a crucial social license for its development. Achieving tangible global benefits from this technical and social potential, such as enhanced food security, improved nutrition, and sustainable agricultural practices, requires a deliberate and collaborative path forward. This entails fostering transparent, internationally coherent regulations that both ensure safety and encourage innovation. It equally demands proactive, evidence-based communication and inclusive dialogue that engages diverse stakeholders, especially communities in regions most vulnerable to food insecurity. The pace and success of the coming agricultural revolution will therefore be determined not by science alone, but by our collective ability to align technological capability with responsive governance and societal trust.

Funding Statement

This work was supported by Fundamental Research Funds for the Central Universities [226-2025-00004] and Biological Breeding-Major Projects [2023ZD04062].

Disclosure statement

No potential conflict of interest was reported by the author(s).