Abstract
The significance of effective and reliable prediction of ecotoxicity, particularly across various trophic levels, including humans, is gaining increasing prominence as ecosystems face new threats and challenges. Computational ecotoxicological predictive approaches have already been deemed as a swifter and economical feasible answer. This work presents a new proposal in that context, integrating structure-based virtual screening and quantitative structure-activity relationship (QSAR) methodologies to address the ecotoxicity of per- and poly-fluoroalkyl substances (PFAS) in aquatic organisms, such as zebrafish. By focusing on the interaction between PFAS and the zebrafish mitochondrial voltage-dependent anion channel (zfVDAC2), resembling bioaccumulation in low concentrations, we analyzed 123 PFAS compounds. Our findings reveal that the top-ranked docked PFAS exhibits a predominant affinity for van der Waal interactions, followed by fluorine (F)-halogen bonds and hydrogen bonds interactions. The latter suggests that PFAS interaction strength may influence mitochondrial ATP transport via zfVDAC2. Similarly, the derived QSAR models identified packing density index, a descriptor linked to van der Waal interactions, as the most significant PFAS factor. Moreover, the high predictive power and statistical robustness of these models positioning them as valuable tools for environmental risk assessment in PFAS applications, while offering mechanistic insights into ecotoxicity.