In vitro validation of a protein homology model is critical for determining the predictivity of a computationally generated structure. Here we discuss the generation, validation, and application of a homology model for CYP1A1. Validation of the CYP1A1 homology model, generated using the highly homologous crystal template of human CYP1A2 (pdb 2HI4), was achieved using the prototypic substrate 7-ethoxyresorufin (Eres). The model was subsequently applied to generate CYP1A1 mutants with increased catalytic efficiency (Vmax/Km) towards the anticancer prodrug dacarbazine (DTIC). Thirty-three directed CYP1A1 mutants were generated and expressed in E. coli; six of these were generated to rationalise docking data obtained from in silico experiments using Eres. DTIC N-demethylation by the CYP1A1 E161K, E256K, and I458V mutants exhibited Michaelis-Menten kinetics, with decreases in Km that doubled the catalytic efficiency relative to wild-type (P < 0.05). As a chemotherapeutic agent, DTIC has relatively poor clinical activity in human malignancies and exhibits numerous adverse effects, which presumably arise from bioactivation in the liver and other tissues resulting in systemic exposure to the cytotoxic metabolite. The successful generation of CYP1A1 enzymes with catalytically enhanced DTIC activation highlights their potential use as a strategy for P450-based gene directed enzyme prodrug therapy (GDEPT) in the treatment of metastatic malignant melanoma. Moreover, the combination of in vitro kinetic analyses with in silico docking data from a validated homology model has allowed interpretation of the structure-activity relationships of this enzyme-substrate pair.