Combining chemoinformatics, molecular modeling and simulation methods: an alternative screening strategy in the search for new agricultural fungicides

MARÍA LILIANA MIRANDA SANGUINO; VICTORIA RICHMOND; MARCELO JAVIER OTERO

Ciudad de Córdoba, Argentina

Congreso; Reunión Anual de la Sociedad Argentina de Biofísica; 2023

Sociedad Argentina de Biofísica

COMBINING CHEMOINFORMATICS, MOLECULAR MODELING ANDSIMULATION METHODS: AN ALTERNATIVE SCREENING STRATEGY IN THE SEARCH FOR NEWAGRICULTURAL FUNGICIDESMaria L. Miranda1,2, Victoria Richmond1,3, Marcelo Otero4 1UMYMFOR(CONICET-UBA), 2Depto. de Química Biológica, FCEyN-UBA 3Depto.de Química Orgánica, FCEyN-UBA, 4IFIBA (CONICET-UBA), Depto. deFísica, FCEyN-UBA,  One of the main causesof food crop losses is due to plant diseases, mainly caused by phytopathogenicfungi. Most of these diseases cannot be effectively controlled, especiallybecause of accelerated development of fungal resistance to commercialfungicides. Thus, there is a pressing demand to develop new fungicides,particularly in an agro-exporting country like Argentina. In that sense, therepurposing of bioactive compounds would significantly accelerate thedevelopment process of new fungicides since safety information and productionmethods are already known. Inorder to identify novel fungicides, we started a cheminformatic study based onapproved compounds in the DrugBank database. We calculated a set ofphysicochemical descriptors for every compound and projected them over thechemical space defined by current approved fungicides, reported by theFungicide Resistance Action Committee (FRAC). It was possible via PrincipalComponents Analysis (PCA), a dimensionality reduction method. This step allowedus to identify those compounds that, according to their properties, share thesame chemical space with current fungicides whose have a known moleculartarget. Then, we propose as hypothesis that the higher the similarity in termsof physicochemical properties between the compounds in the database and theantifungal compounds in the FRAC, the more likely they are to be active.Oneof the targets of chemical fungicides and the first one we focused on is β-tubulin,which polymerizes with alpha α-tubulinto form microtubules. Microtubules are one of the cytoskeleton components andplay important roles in many processes, such as supporting cell structure andcell division, becoming essential for fungal viability. Thus, combination ofcheminformatics (PCA) and computational modeling tools like molecular docking,allowed us to filter the drug library and reduce the number of candidates to10% of the original database. Crossing this reduced base with fungicidalactivity data available in the literature, it was possible to identify sevenpotential inhibitors for β-tubulin, whose affinity for the target was evaluatedthrough molecular dynamics (MD) simulations. In all cases, we obtained anaffinity value comparable or even higher than reference compounds such asnocodazole or other fungicides currently used and aimed at inhibiting β-tubulinpolymerization. Now, we consider reapplying this screening workflow to identifypotential inhibitors of a novel fungal target as chitin deacetylase.