Significant variations in the metabolism of various drugs and environmental chemicals which are metabolized via the cytochrome P450 (CYP) enzymes exist between humans on an individual and population scale. Many of these interindividual variations are attributed to polymorphisms in the 2C subfamily of enzymes. CYP2C subfamily of enzymes are responsible for the metabolism of a number of therapeutic agents such as S-mephenytoin, omeprazole, proguanil, certain barbiturates, diazepam, propranolol, diclofenac, tolbutamide, imipramine, and taxol. The overall objectives of this work is to elucidate the molecular and metabolic basis of CYP related polymorphisms, characterization of the substrate-structure metabolism relationships, and the determination of critical protein structures of CYP enzymes that infer substrate specificity and metabolic activity. Methods to study the metabolism and kinetics of the anticancer drug taxol and the anticonvulsant drug mephenytoin were developed in order to identify the polymorphisms, their functional significance, and the substrate recognition site (SRS) in CYP enzymes. Thirty four single nucleotide polymorphisms (SNPs) of CYP3A5 including 27 previously unidentified SNPS were identified by direct sequencing of CYP3A5 exons, intron-exon junctions and the 5?-upstream region of DNA from 92 ethnically and racially diverse individuals (24 Caucasians, 24 Africans, 24 Asians, and 20 individuals of unknown racial origin). Four new CYP3A5 SNPs produced coding changes: R28C, L82R, A337T, and F446S. CYP3A5 R28C was identified in African populations (allelic frequency of 4%). CYP3A5 A337T occurred in Asians (allelic frequency of 2%), CYP3A5 L82R occurred in the racially unidentified group (allelic frequency of 2%), and CYP3A5 F446S (linked to the splice variant CYP3A5*3) was identified in Caucasians (allelic frequency of 2%). The newly identified alleles were constructed from CYP3A5*1 by site-directed mutagenesis and expressed in an Escherichia coli cDNA expression system. CYP3A5*1 exhibited the highest intrinsic clearance for testosterone followed by CYP3A5 A337T> CYP3A5 R28C>> CYP3A5 F446S. CYP3A5*1 exhibited higher nifedipine oxidation than CYP3A5 R28C > CYP3A5 A337T >> CYP3A5 F446S. The Vmax of CYP3A5 A337T and CYP3A5 R28C were ~40-60% lower than that of CYP3A5*1 for both testosterone and nifedipine, while CYP3A5 F446S exhibited a >95% decrease in the intrinsic clearance for both 6?-hydroxytestosterone and nifedipine oxidation compared to that of CYP3A5*1. Of the three new potentially defective coding alleles, CYP3A5 F446S is predicted to be more defective than the CYP3A5*3 splice change alone.