Computing pathogenicity of mutations in human cytochrome P450 superfamily

Abstract

Cytochrome P450 (CYP) are heme-containing enzymes involved in the metabolism of drugs and endogenous compounds. Humans have 57 known CYPs, with >200 mutations linked to severe disorders, highlighting their role in disease etiology. To investigate their pathogenicity, we performed an in-depth computational study, comparing pathogenic mutations with non-pathogenic ones. First, we evaluated the effects of all possible mutations across 26 known CYPs structures using five structure-and five sequence-based methods, aiming to reproduce pathogenesis. We computed 23,46,410 mutations, categorized into: all possible, non-pathogenic, and pathogenic. Comparisons consistently revealed a meaningful stability pattern: non-pathogenic > all > pathogenic. Second, we found that a significantly higher number of pathogenic mutations were buried in CYP structures, relating to their higher pathogenesis potential. Third, we analyzed mutation allele frequency relative to solvent accessibility. Fourth, we computed change in 48 amino acid properties upon mutation and identified three distinguishing features: Gibbs free energy, isoelectric point and volume. Fifth, the positive residue content was reduced significantly in diseased mutations, with arginine mutations being the main culprit, directly linked to isoelectric point change. Sixth, we found a higher propensity for pathogenic mutations in conserved sites, suggesting disruption of CYP function. Finally, analysis of heme versus substrate sites showed a higher frequency of pathogenic mutations in heme site, with arginine being the main mutating residue, possibly disrupting arginine-heme interactions. We provide the first comprehensive analysis of mutation effects across multiple CYPs. We model the chemical basis of CYP-related pathogenicity, paving way for a semi-quantitative model to predict diseases.