A fungal tolerance trait and selective inhibitors proffer HMG-CoA reductase as a herbicide mode-of-action

The escalating problem of herbicide resistance in weeds poses a serious threat to global food security. The efficacy of current herbicides has diminished significantly, highlighting the urgent need for novel modes of action. The development of new herbicides has been slow, with only one new mode of action introduced in the last 40 years. Existing herbicides targeting isoprenoid biosynthesis, such as clomazone and bixlozone, which inhibit 1-deoxy-D-xylulose-5-phosphate synthase, have shown some success but are not without limitations. Isoprenoid biosynthesis is essential for the production of various vital compounds, including lipids, hormones, vitamins, and defense compounds, across all kingdoms of life. However, the biosynthetic pathways differ, with plants uniquely utilizing both the mevalonate (MVA) and methylerythritol phosphate (MEP) pathways. The MVA pathway, a key focus of this research, has not yet been exploited as a target for commercial herbicides. HMGR, a highly regulated, rate-limiting enzyme in the MVA pathway, is responsible for the conversion of HMG-CoA to mevalonate, and is the target of statins, drugs used to treat hypercholesterolemia in humans. HMGR exists in two classes, with class I enzymes possessing an N-terminal membrane domain and varying NAD(P)H cofactor preferences. Regulation of HMGR appears conserved between humans and plants, but many of the regulatory proteins differ, presenting challenges in targeting the enzyme for herbicide development. Statins, while known to exhibit herbicidal activity against various plant species, haven’t been thoroughly investigated in terms of efficacy and species-specificity. The potential of HMGR as a herbicide target has been overlooked due to concerns over off-target effects and antimicrobial activity of statins, but recent development of selective insecticides against HMGR offers a new perspective.

Several studies have demonstrated the herbicidal activity of various statins against different plant species, including *Lemna gibba*, *Raphanus sativus*, *Scoparia dulcis*, and *Arabidopsis thaliana*. However, comparative data on the herbicidal efficacy of different statins, particularly second-generation, synthetic statins, are limited. The discovery of mevastatin and lovastatin from fungal sources, and subsequent development of synthetic statins, provided initial evidence of HMGR's potential as a herbicidal target. However, concerns regarding off-target effects and antimicrobial activity hampered the exploration of HMGR as a herbicide target. Recent research on the development of selective insecticides targeting HMGR in other organisms suggest it is possible to overcome these limitations.

The study involved multiple aspects: herbicidal activity assays, crystal structure determination, species-specific inhibitor design, and generation of a statin-tolerance trait. **Herbicidal Activity Assay:** *Arabidopsis thaliana* (dicot) and *Eragrostis tef* (monocot) were treated with eight commercially available statins (rosuvastatin, pravastatin, simvastatin, mevastatin, lovastatin, fluvastatin, atorvastatin, and pitavastatin) at various concentrations, pre- and post-emergence. Growth inhibition was quantified by measuring the green pixel area in images of the plants using ImageJ software. **Crystal Structure Determination:** The crystal structure of the core domain of *AtHMG1* was determined in both apo form and complexed with pitavastatin using X-ray crystallography. Diffraction data were processed using XDS and AIMLESS, molecular replacement was performed with PHASER, and manual building and refinement were carried out using Coot and REFMAC5. **Development of Plant-Specific Statins:** Based on the crystal structure insights, especially regarding the wider active site of AtHMG1 compared to human HMGR, analogues of atorvastatin were designed and synthesized. These analogues incorporated modifications to the isopropyl group on the central pyrrole ring. Herbicidal activity was evaluated using *A. thaliana*, and *in vitro* species-specificity was assessed using a fluorometric NADPH-depletion assay with both AtHMG1 and human HMGR. **Engineering Statin Tolerance:** The authors analyzed HMGR genes from fungal biosynthetic clusters for natural statins to identify mutations conferring statin resistance. A Leu558 to Thr mutation (L558T) found in a cluster-associated HMGR gene was introduced into recombinant AtHMG1. The effects of this mutation on statin resistance and enzyme activity were tested *in vitro*. Full-length AtHMG1 and AtHMG1-L558T were overexpressed in *A. thaliana* to assess the potential of the L558T mutation as a tolerance trait. Rosuvastatin resistance was then evaluated in transgenic plants using ImageJ analysis.

The study revealed a range of herbicidal activities among the tested statins, with rosuvastatin being the most potent. The crystal structure of AtHMG1 showed a wider active site than previously characterized HMGR structures, driven by an altered conformation of the Lβ2-Lα1 loop. This wider pocket offered potential for developing plant-specific inhibitors. The designed atorvastatin analogues showed differential activity, with compound 7 displaying over 20-fold higher selectivity for AtHMG1 over human HMGR (*in vitro*). A single amino acid change (L558T) identified in fungal HMGR conferred significant statin resistance *in vitro* and *in planta*, thus demonstrating the potential for this mutation as a tolerance trait. The L558T mutation in *A. thaliana* overexpressing full-length *AtHMG1* conferred greater resistance to rosuvastatin than wild-type plants or those overexpressing wild type *AtHMG1*.

This research successfully validates HMGR as a potential target for herbicide development. The structural differences between plant and human HMGR, particularly the wider active site of AtHMG1, provide a basis for developing species-specific inhibitors. The identification of the L558T mutation as a successful tolerance trait provides an avenue for developing herbicide-tolerant crops. The findings suggest the feasibility of developing effective herbicides with minimal off-target effects. The study also highlights the potential of utilizing natural product-based herbicides and integrating resistance traits derived from biosynthetic gene clusters.

This study establishes HMGR as a viable herbicide target, demonstrating the feasibility of designing species-specific inhibitors and engineering tolerance traits. The findings open new avenues for developing novel herbicides and improving weed management strategies. Future research should focus on optimizing the L558T tolerance trait, exploring other potential resistance mutations, evaluating the efficacy of these approaches in various crops, and investigating the impact of HMGR inhibition on plant sterol levels and seed production. Further investigation of the regulatory elements interacting with the N-terminal domain of HMGR could potentially lead to the development of even more species-specific inhibitors.

The study focused primarily on *Arabidopsis thaliana* and *Eragrostis tef*. The generalizability of the findings to other plant species requires further investigation. The *in vitro* assays and *in planta* experiments may not fully capture the complexities of herbicide action within the whole plant environment. Further research is needed to fully understand the impacts on plant growth and development.