Water is a necessary resource that sustains life on earth. Unfortunately, climate change is severely affecting rainfall, leading to a higher incidence of droughts that have a major impact on rain-fed farming systems, which account for 80 % of the world’s cropland and produce about 60 % of global agricultural output. Therefore, solutions to improve water use efficiency in crops are becoming crucial to contribute to food security. Under drought stress, the inhibition of clade A subfamily of protein phosphatases type-2C (PP2Cs) modulates plant transpiration. Indeed, it has been extensively reported that the genetic inactivation of PP2Cs generates mutant plants more tolerant to drought and, in consequence, more productive than wild type plants in water limited conditions. However, this genetic approach often leads to plants with a growth penalty which limits their use. Alternatively, in this project, we propose a chemical approach taking advantage of artificial intelligence (AI) to design small molecules that directly inhibit PP2C activity thus activating drought response mechanisms in plants. Towards this end we have used AI through machine learning algorithms to screen millions of compounds in silico. In our lab, we have built a pipeline to analyze the potency, solubility, specificity and toxicity in vitro and in vivo of the putative PP2C inhibitors previously designed through AI and filtered by score and structural docking analysis. Our early results have shown a high success rate of the AI machine learning algorithms designing PP2C inhibitors. More than 85 % of the compounds were effective inhibiting PP2C phosphatases in vitro showing many of them (≈ 25 %) IC50s in the low micromolar range. However, given the high structural similarity between PP2C phosphatases the compounds were also effective inhibiting PP2Cs non-related to drought response mechanisms. We are applying structure-guided ligand design to improve the specificity and potency of our best candidates, especially those that induce a drought-related response observed in in vivo-experiments. In sum, our pipeline allows us to design and to test a significant number of PP2C inhibitors not only in vitro but also in vivo using whole crop plants such as tomato, thus increasing our chances to translate these findings into commercial agrochemicals.
Poster
