Water Exchange from the Buried Binding Sites of Cytochrome P450 Enzymes 1A2, 2D6, and 3A4 Correlates with Conformational Fluctuations

Abstract

Human cytochrome P450 enzymes (CYPs) are critical for the metabolism of small-molecule pharmaceuticals (drugs). As such, the prediction of drug metabolism by and drug inhibition of CYP activity is an important component of the drug discovery and design process. Relative to the availability of a wide range of experimental atomic-resolution CYP structures, the development of structure-based CYP activity models has been limited. To better characterize the role of CYP conformational fluctuations in CYP activity, we perform multiple microsecond-scale all-atom explicit-solvent molecular dynamics (MD) simulations on three CYP isoforms, 1A2, 2D6, and 3A4, which together account for the majority of CYP-mediated drug metabolism. The MD simulations employ a variety of positional restraints, ranging from keeping all CYP atoms close to their experimentally determined coordinates to allowing full flexibility. We find that, with full flexibility, large fluctuations in the CYP binding sites correlate with efficient water exchange from these buried binding sites. This is especially true for 1A2, which, when restrained to its crystallographic conformation, is unable to exchange water between the binding site and bulk solvent. These findings imply that, in addition to crystal structures, a representative ensemble of conformational states ought to be included when developing structure-based CYP activity models.

Effect of Lipid Bilayer Anchoring on the Conformational Properties of the Cytochrome P450 2D6 Binding Site

“…Snapshots for RMSF analysis were saved every 0.25 ns from the production phase of the MD simulations, resulting in 4000 snapshots for each 1-μs production simulation (“run”). C α RMSF values were computed on a per-run basis by first aligning all 4000 snapshots from the run to a reference structure using least-squares structural alignment minimizing the RMSD of the C α atoms for residues 54–497; other N-terminal residues were excluded on the basis of prior aqueous CYP2D6 simulations showing high structural variability in the N-terminal region . Reference structures for the aqueous and the bilayer-anchored systems were the coordinates from the first of the triplicate runs for each system after completion of the heating and equilibration phases, corresponding to t = 0 of the first run’s production stage and displayed in Figure .…”

“…Protein parameters were version 36m, − POPC lipid parameters were version 36, K + and Cl – ion parameters were as described in references , and water parameters were the TIP3P model modified for the CHARMM force field . Force field parameters for iron­[III] heme with the Cys443 side chain liganded in the thiolate form, as used for both the bilayer-anchored and the aqueous system simulations, are available as Supporting Information for reference .…”