Recently, it was found that mutations in the binding cavity of drug-metabolizing Cytochrome P450 BM3 mutants can result in major changes in regioselectivity in testosterone (TES) hydroxylation. In the current work, we report the intrinsic reactivity of TES' C-H bonds and our attempts to rationalize experimentally observed changes in TES hydroxylation using a protein structure-based in silico approach, by setting up and employing a combined Molecular Dynamics (MD) and ligand docking approach to account for the flexibility and plasticity of BM3 mutants. Using this approach, about 100,000 TES binding poses were obtained per mutant. The predicted regioselectivity in TES hydroxylation by the mutants was found to be in disagreement with experiment. As revealed in a detailed structural analysis of the obtained docking poses, this disagreement is due to limitations in correctly scoring hydrogen-bonding and steric interactions with specific active-site residues, which could explain the experimentally observed trends in regioselectivity in TES hydroxylation.