Agricultural biotechnology has a well-worn playbook for finding herbicide resistance genes. You cast a wide net, you mutagenize or screen thousands of soil microbes, plant metagenomes, or existing enzyme libraries, and you test each candidate against a panel of chemicals until something sticks. It is brute force, and it has worked well enough. But it leaves a nagging question unanswered: what if the best solution is hiding in a place you would never think to look?
A study published July 25 in Nature Communications by researchers at Shandong University and Qingdao KingAgroot CropScience offers an answer that flips that playbook on its head. Rather than screening blindly, the team used a structural insight rooted in fundamental biochemistry to predict exactly which enzyme class could degrade a family of widely used herbicides. The result is not only a transgenic rice line resistant to multiple auxin herbicides, but a demonstration that mechanism-driven discovery can uncover solutions that random screening would almost certainly miss.
A Shared Molecular Handshake
The story begins not with a herbicide, but with the chemistry of how proteins grab hold of small molecules. The authors noticed something that had been sitting in the literature for years: auxin herbicides such as 2,4-D, MCPA, and fluchloraminopyr all bind their protein targets through a terminal carboxyl group. Whether docking into the TIR1 receptor, the RdpA enzyme, or the Arabidopsis GH3.15 protein, the carboxyl group forms a salt bridge with a conserved arginine residue or a hydrogen bond with a serine. It is a predictable, well-characterized recognition motif.
Separately, a family of bacterial cytochrome P450 enzymes called CYP152 peroxygenases also use a terminal carboxyl group to anchor their substrates. A conserved arginine in their active site grips the carboxyl, positioning the molecule for oxidative cleavage. The structural parallel was striking, but no one had connected the dots.
The Shandong team did. They reasoned that if both herbicides and CYP152 substrates rely on the same carboxyl-anchoring mechanism, CYP152 enzymes might recognize auxin herbicides as legitimate substrates and degrade them.
From Hypothesis to Degradation
To test this, the researchers assembled three CYP152 peroxygenases, OleT_JE, P450 BSbeta from Bacillus subtilis, and P450 SPalpha, and challenged them against seven auxin herbicides using an in situ hydrogen peroxide generation system. Two of the three enzymes, OleT_JE and P450 BSbeta, catalyzed C-O bond cleavage across multiple herbicides, converting compounds like 2,4-D, MCPA, and fluchloraminopyr into nonherbicidal phenols. P450 SPalpha showed negligible activity.
More surprising was what happened next. A single P450 BSbeta enzyme performed not one but two sequential reactions on the same substrate: first cleaving the C-O bond to produce dichlorophenol, then hydroxylating the aromatic ring to generate dichlorocatechol. Two-step consecutive degradation by a single enzyme had not been reported before in the context of herbicide resistance.
Opening the Gate
P450 BSbeta, it turned out, had a problem. Its active site was choked by a narrow substrate access channel lined with five bulky phenylalanine residues, creating a bottleneck that limited activity. The team systematically replaced each phenylalanine with alanine, a smaller amino acid, and every single mutant showed improved performance.
The best variant carried a single substitution at position 46. P450 BSbeta-F46A achieved 86 percent conversion of fluchloraminopyr with a turnover number of 7,443 and a catalytic efficiency of 12.3 per millimolar per minute. For the herbicides 2,4-DB and MCPB, conversion reached nearly 100 percent, with turnover numbers exceeding 8,700 and catalytic efficiencies above 80. The broadest substrate scope of any reported auxin herbicide-degrading enzyme.
Following the Electron
To understand exactly how the reaction worked, the team turned to quantum mechanics/molecular mechanics calculations. The simulations revealed a precise choreography: the herbicide’s carboxyl group forms a salt bridge with Arg242 in the enzyme’s active site. Hydrogen peroxide enters and undergoes heterolytic O-O cleavage to form the reactive Compound I intermediate. Compound I abstracts a hydrogen atom from the carbon adjacent to the oxygen bridge, the alpha-carbon, with an energy barrier of about 21.1 kilocalories per mole. A hydroxyl group rebounds onto the radical intermediate, and a water-mediated proton transfer drives spontaneous C-O bond cleavage.
Labeling experiments using oxygen-18 confirmed that the oxygen atom in the released phenol originates from the herbicide itself, not from hydrogen peroxide or water. A kinetic isotope effect of 3.0 with deuterated substrate indicated that the alpha-carbon hydrogen abstraction is the rate-limiting step. The arginine anchor proved essential: a R242A mutant completely abolished activity.
From Bench to Field
The team introduced the codon-optimized F46A mutant into rice variety Jingeng 818, fusing it to a bacterial cytochrome P450 reductase domain for electron supply, a chloroplast localization signal, and a maize ubiquitin promoter. Western blot confirmed protein expression, and GFP imaging showed the enzyme successfully targeted to chloroplasts.
In plate assays, 0.03 micromolar of the pro-herbicide fluchloraminopyr-methyl severely inhibited root growth in wild-type rice. Transgenic plants carrying the P450 BSbeta-F46A gene showed significantly less impairment. In spray assays at 40 and 80 grams of active ingredient per hectare, transgenic T1 plants displayed no notable growth defects while wild-type plants were strongly inhibited.
Beyond the Trait
The practical outcome, a rice variety resistant to multiple auxin herbicides, is certainly valuable. But the deeper significance of this study lies in how the researchers got there. They did not discover an obscure gene from a deep-sea vent or a hypersaline lake through heroic metagenomic sequencing. They found a common bacterial enzyme doing something it was structurally predestined to do, guided by a simple observation about how proteins recognize carboxyl groups.
This approach represents a shift in strategy for agricultural biotechnology. Random screening discovers what is abundant; mechanism-driven design discovers what fits. The two are not the same, and the gap between them is where the CYP152 family had been sitting, unnoticed, waiting for someone to notice the handshake.
The authors report that the terminal carboxyl group is essential for activity, analogs lacking it showed no conversion. That is not a limitation; it is a design principle. Now that it has been articulated, the same logic could be applied to search for enzymes that degrade other classes of agrochemicals by matching structural recognition patterns rather than casting ever-wider nets.
The Shandong study is not the end of screening in biotechnology. But it marks the moment when rational design, built on fundamental biochemical understanding, began to catch up.
References
Jiang, Y., Zhang, J., Li, Z. et al. Mechanism-driven discovery of a bacterial P450 peroxygenase conferring broad-spectrum auxin herbicide resistance through consecutive two-step degradation. Nature Communications (2026). DOI: 10.1038/s41467-026-75811-1