F429 Regulation of Tunnels in Cytochrome P450 2B4: A Top Down Study of Multiple Molecular Dynamics Simulations

Abstract

The root causes of the outcomes of the single-site mutation in enzymes remain by and large not well understood. This is the case of the F429H mutant of the cytochrome P450 (CYP) 2B4 enzyme where the substitution, on the proximal surface of the active site, of a conserved phenylalanine 429 residue with histidine seems to hamper the formation of the active species, Compound I (porphyrin cation radical-Fe(IV) = O, Cpd I) from the ferric hydroperoxo (Fe(III)OOH-, Cpd 0) precursor. Here we report a study based on extensive molecular dynamic (MD) simulations of 4 CYP-2B4 point mutations compared to the WT enzyme, having the goal of better clarifying the importance of the proximal Phe429 residue on CYP 2B4 catalytic properties. To consolidate the huge amount of data coming from five simulations and extract the most distinct structural features of the five species studied we made an extensive use of cluster analysis. The results show that all studied single polymorphisms of F429, with different side chain properties: i) drastically alter the reservoir of conformations accessible by the protein, perturbing global dynamics ii) expose the thiolate group of residue Cys436 to the solvent, altering the electronic properties of Cpd0 and iii) affect the various ingress and egress channels connecting the distal sites with the bulk environment, altering the reversibility of these channels. In particular, it was observed that the wild type enzyme exhibits unique structural features as compared to all mutant species in terms of weak interactions (hydrogen bonds) that generate a completely different dynamical behavior of the complete system. Albeit not conclusive, the current computational investigation sheds some light on the subtle and critical effects that proximal single-site mutations can exert on the functional mechanisms of human microsomal CYPs which should go rather far beyond local structure characterization.

Fig 1. A schematic catalytic cycle for CYPs, showing all the key species, and the uncoupling process.

Fig 2. a,b) Cartoon representation of the CYP-2B4 crystal structure (PDB ID 1SUO) with specific domains coloured and labeled according to Scott et al.;[12] the heme group, the peroxide anion and Cys426 are shown as ball and stick. The enzyme is shown looking from the distal (a) or proximal (b) side of the heme group. To be coherent with the literature, the same color map reported in the original experimental work has been used. (c,d) Detail of the active site.

The protein backbone is shown as a transparent cyan ribbon, while Cpd0 is shown in and stick style, together with the various rotamers of residue 429. Cys 436, residue 429 and the A and D propionate groups are labeled.

Fig 3. Clustering of CYP2-2B4 conformations.

Distribution of structures from 17500 frames artificial trajectory (frame set #2) in clusters (y axis) and simulations (x axis) using GROMOS. Snapshots originally sampled in the 15 trajectories are represented as black circles (WT), blue squares (F429A), magenta triangles (F429E), orange squares (F429H) and green down pointing triangles (F429L).

Fig 4. Principal component analysis.

Projection of the trajectories of Cα atoms from the larger artificial trajectory in the volume spanned by the first 3 eigenvectors obtained by the principal component analysis, from 2 different points of view. Frames have been represented as dots coloured in black (WT), blue (F429A), magenta (F429E), orange (F429H) and green (F429L).

Fig 5. Root Mean Square Fluctuations in clusters and specific systems.

a) Cα atom RMSF in the 3 clusters obtained by clustering. Cyan dashed line: cluster #1 (i.e. frames from the F429A, F429E and F429H simulations); full black line: cluster #2 (WT); magenta dot-dashed line: cluster #3 (F429L). b) difference of RMSF calculated from the trajectories of 4 mutants with respect to the WT trajectory. Full blue line: F429A; magenta dashed line: F429E; orange dot-dashed line: F429H; green dotted line: F429L.

Fig 6. Per residue averaged root mean square fluctuation.

Average structures of the 3 clusters, depicted using ribbon colours and thickness proportional to the RMSF. a) First cluster (F429A, F429E, F429H). b) Second cluster (Wild Type enzyme). c) Third cluster (F429L).

Fig 7. Cys436 hydration.

Radial distribution functions of water molecules (center of mass) around the Sγ atom in C436 calculated from the WT (full black line), F429A (blue dashed line), F429E (magenta dot-dashed line), F429H (orange dashed line) and F429L (green dotted line).

Fig 8. Tunnel towards Cys436.

The heme group is shown in purple and the protein backbone as a cyan transparent ribbon. Cys436 is always shown in ball and stick. a) Tunnels in the centroid of clusters #1. The bottleneck residues E429 and L437 are represented in licorice. b) Tunnel of the F429H simulation; residues G438 and L437 are also shown.

Fig 9. Peroxide anion hydration.

Radial distribution function of water molecules (center of mass) around the OOH^- anion calculated for the WT, F429A, F429E, F429H and F429L, respectively.

Fig 10. Bottleneck radii of pathways frame set #2.

a) bottleneck radius of pathways #1 (blue), #5 (orange) and #6 (magenta). b) bottleneck radius of pathway #2 (black), #3 (red) and #4 (green). Any given configuration may be easily related to the simulation it belongs to dividing the graph in windows 150 ns wide (see also the graph top).

Fig 11. Pathways to the distal site.

The heme group, OOH^- and Cys436 are shown in magenta. The enzyme and tunnel coordinates shown are those of the corresponding centroid as in Fig 8. a) Pathways #1 (red), #2 (green) have been represented; coordinates from the F429H centroid were used. Bottleneck lining residues for pathway #1 and #2 have been represented in ball stick or licorice, respectively. b) Pathway #3 (F429L simulation) and corresponding bottleneck residues. c) Pathways #4 (WT simulation) and corresponding bottleneck residues.