Scientists Engineer Supercharged Herbicide-Resistant Crops by Editing Protein and Promoter Together

A team of plant scientists in China has unveiled a dual-layer genome-editing strategy that rewrites both the protein-coding sequence and the regulatory DNA of a single crop gene, producing rice and wheat plants that shrug off herbicide doses far exceeding what farmers normally spray in the field. The study, published in Nature Plants by Rui Zhang, Zhuangzhuang Chai, Meihua Wang, Sha Chen and colleagues under the leadership of Caixia Gao at the Institute of Genetics and Developmental Biology of the Chinese Academy of Sciences, tackles one of the most persistent bottlenecks in agricultural biotechnology: how to create elite crop alleles that deliver powerful traits without undermining the essential functions of the proteins they modify.

The target of the engineering campaign was 4-hydroxyphenylpyruvate dioxygenase, or HPPD, an enzyme that sits at a critical junction in plant metabolism. HPPD catalyzes a key step in the tyrosine degradation pathway, converting 4-hydroxyphenylpyruvate into homogentisate, which in turn feeds the production of plastoquinone and tocopherols. Because plastoquinone is indispensable for carotenoid biosynthesis and for protecting the photosynthetic apparatus, blocking HPPD with herbicides such as mesotrione, isoxaflutole and fenpyrazone causes treated plants to bleach and die. This vulnerability makes HPPD inhibitors among the most valuable weed killers in modern agriculture, but it also means that any crop engineered to tolerate them must maintain a delicately balanced enzyme that still performs its metabolic duties while refusing to bind the herbicide.

Previous attempts to generate herbicide-tolerant HPPD variants relied largely on directed evolution, randomly mutating the gene and screening survivors for resistance. That approach can find resistance mutations, but it offers no guarantee that the resulting enzyme retains full catalytic activity, and it typically operates on a single genetic axis: the protein sequence itself. The Gao laboratory chose a fundamentally different route, beginning with structure. The team solved the crystal structure of rice HPPD, OsHPPD, depositing it in the Protein Data Bank under accession 21WF, and used the three-dimensional map of the enzyme to identify amino acid positions where substitutions could plausibly disrupt herbicide binding while leaving the catalytic machinery untouched. This structure-guided design narrowed the search space dramatically, focusing attention on residues near the herbicide-binding pocket rather than across the entire protein.

The structural analysis pointed to a cluster of glycine residues and a phenylalanine at position 378 as prime candidates. The researchers designed a panel of coding variants, including F378A, G414A, G415V and a series of substitutions at position 417, each chosen because the atomic geometry suggested the mutation would either enlarge or reshape the herbicide-binding cavity without disturbing the iron-coordinating active site or the substrate-channeling architecture of the enzyme. Biochemical assays confirmed the logic: the engineered enzymes retained their dioxygenase activity against the natural substrate while showing markedly reduced sensitivity to mesotrione, the HPPD-inhibiting herbicide used as the primary test compound throughout the study.

Installing these mutations in the rice genome required precision rather than brute force. The team deployed prime editing, a CRISPR-derived technology that uses a modified Cas9 fused to a reverse transcriptase and an extended guide RNA carrying the desired edit, to write the designed substitutions directly into the endogenous OsHPPD locus. Prime editing was chosen over conventional CRISPR-Cas9 knockout approaches because it makes exact sequence changes without creating double-strand DNA breaks, avoiding the large deletions, rearrangements and chromosomal damage that have been documented as unintended consequences of break-based editing. Using both standard pegRNA vectors and engineered epegRNA systems, the researchers generated transgene-free rice lines carrying each of the designed resistance alleles, confirming the edits by Sanger sequencing and verifying that the plants grew normally in the absence of herbicide.

The coding mutations alone conferred meaningful tolerance, but the team suspected that a second lever could push resistance to a level that would survive real-world field conditions. Their attention turned to cis-regulatory DNA, the non-coding sequences that govern how strongly a gene is expressed. Drawing on expression-modulating elements, short enhancer-like motifs that can boost transcription when placed near a promoter, the researchers screened candidate sequences in rice protoplasts using a dual-luciferase reporter assay. One sequence stood out: a compact 22-base-pair element, dubbed EME, that significantly increased expression when inserted upstream of the OsHPPD core promoter and 5-prime untranslated region. Computational analysis with the PlantPAN 4.0 database predicted that the element recruits transcription factors capable of amplifying gene activity, and RNA-folding simulations suggested the short sequence adopts a defined secondary structure that may facilitate its integration during prime editing.

The mechanism of EME insertion is itself a technical tour de force. The 22-base-pair element contains a palindromic sequence, meaning it reads the same in both directions, and the researchers showed that during the flap-equilibration step of prime editing this palindrome can exist either as a linear strand, producing a precise replacement, or as a structured three-prime flap whose microhomology with the genomic target allows flexible insertion. By tuning the editing system, including the use of an enhanced prime editor called ePPEplus, the team achieved efficient installation of the EME at the OsHPPD locus in whole plants, generating alleles that combined a resistance mutation in the coding region with an expression-boosting element in the regulatory region, all within a single gene and without any foreign DNA.

The results were striking. Rice plants carrying both layers of engineering, a resistance mutation such as G417A or G414A plus the EME element, showed elevated OsHPPD mRNA and protein abundance and tolerated mesotrione at rates up to four times the recommended field application dose. The plants also withstood isoxaflutole and fenpyrazone, demonstrating that the tolerance extends across the HPPD-inhibitor class rather than being specific to a single chemistry. Crucially, multi-season field trials in Hainan, China, in 2025 and 2026 showed that the engineered lines matched or exceeded wild-type plants in every major agronomic trait measured, including plant height, grain number per panicle, seed-setting rate, thousand-grain weight, tiller number and grain yield per plant. The absence of yield penalties is the detail that separates this work from many earlier trait-engineering efforts, where resistance often came at the cost of vigor.

To test whether the framework was a rice-specific trick or a general principle, the team turned to wheat, a hexaploid crop whose genome carries three homoeologous copies of the HPPD gene, TaHPPD-A1, TaHPPD-B1 and TaHPPD-D1. Multiplex prime editing was used to introduce glycine-to-alanine substitutions at the positions corresponding to rice G414 across all three homoeologs, and a duplicated EME element was then inserted into the regulatory region of one TaHPPD-B1 allele. The resulting wheat plants, designated 2EME-G404A, showed increased TaHPPD transcript and protein levels and enhanced mesotrione tolerance relative to plants carrying only the coding mutation. The demonstration that the dual-layer strategy functions in a polyploid genome, with its added complexity of multiple gene copies and larger regulatory landscapes, is strong evidence that the approach can be ported across species and genome architectures.

Beyond herbicide tolerance, the study articulates a template that the authors argue is broadly applicable to crop improvement. Most engineered traits to date have manipulated either a protein or a regulatory element, and the single-axis approach often hits a ceiling: a resistance enzyme may be too weak, or an overexpressed wild-type enzyme may still be inhibited. By coordinating structure-guided protein design with precise cis-regulatory tuning at the same locus, the dual-layer framework allows each layer to compensate for the limitations of the other, producing alleles that are simultaneously resistant and abundantly expressed. The strategy also sidesteps the transgene silencing, position effects and regulatory burdens associated with inserting foreign genes, because every change is written into the crop’s own DNA. With deep-sequencing data deposited under NCBI BioProject PRJNA1417282 and analysis code released through GitHub, the work provides both the biological blueprint and the computational toolkit for breeding programs seeking to engineer durable, penalty-free traits in the crops that feed the world.

Subject of Research: Dual-layer genome editing of HPPD coding and regulatory sequences to engineer herbicide-tolerant rice and wheat

Article Title: Coordinating structure-guided design and cis-regulatory optimization to engineer elite herbicide-tolerance alleles in crops

Article References: Coordinating structure-guided design and cis-regulatory optimization to engineer elite herbicide-tolerance alleles in crops. (n.d.). https://doi.org/10.1038/s41477-026-02423-4

Image Credits: AI Generated

DOI: 10.1038/s41477-026-02423-4

Keywords: genome editing, prime editing, herbicide tolerance, HPPD, rice, wheat, cis-regulatory elements, mesotrione, crop breeding, protein design, CRISPR, plant biotechnology