Both Reactivity and Accessibility Are Important in Cytochrome P450 Metabolism: A Combined DFT and MD Study of Fenamic Acids in BM3 Mutants

Abstract

Cytochrome P450 102A1 from Bacillus megaterium (BM3) is a fatty acid hydroxylase that has one of the highest turnover rates of any mono-oxygenase. Recent studies have shown how mutants of BM3 can produce metabolites of known drug compounds similar to those observed in humans. Single-point mutations in the binding pocket change the regioselective metabolism of fenamic acids from aromatic hydroxylation to aliphatic hydroxylation. This study is concerned with the individual contribution from accessibility and reactivity for drug metabolism with a future goal to develop fast methods for prediction. For a BM3 M11 mutant as well as the M11 V87F and M11 V87I mutants, we studied the metabolism of the nonsteroidal anti-inflammatory drugs (NSAIDs) mefenamic acid, meclofenamic acid, tolfenamic acid, and diclofenac. Density functional theory (DFT; B3LYP and B3LYP-D3) calculations for all possible reactions were performed using a porphyrin model reacting with the four substrates. Molecular dynamics (MD) simulations were used to determine the potential sites of metabolism that are accessible. Finally, we combine reactivity and accessibility for each potential site to interpret the experimentally determined metabolism. Generally, the 3 and 5 positions (on the ring containing the acidic substituent) and the 2′, 3′, and 4′ positions are most reactive, whereas 4, 5, 3′, and 4′ are most accessible. Combining reactivity and accessibility show that the 5, 3′, and 4′ positions are predicted to be most prone to be metabolized, in agreement with experimentally observed data. Reactivity seems to be the dominant factor in the CYP-mediated metabolism of these NSAIDs, which is consistent with previously published methods based solely on reactivity.

A Modified Arrhenius Approach to Thermodynamically Study Regioselectivity in Cytochrome P450‐Catalyzed Substrate Conversion

“…When plotting natural logarithms of the ratios of V max values of mefenamic acid metabolites against 1000/T,l inear curves were obtained, Figure 5A.T his observed correlation supports our assumption that the temperature dependence of steps 2-6 of the catalytic cycle in Figure 1w ill be similarf or the different hydroxylation paths andc ancelw hen relating V max ratios to the inverse temperature. The slopes (D = DDG bind + DE a )a nd intercepts of the thus obtainedm odified Arrhenius plots are summarizedi nT able 2a nd discussed below in our thermodynamic analysiso ft he observed regio-specificity in MF conversion.I n addition, DDG overall values as derived (using the Curtin-Hammett principle and Equation (6)) from pairs of V max values at 300 Ka re also reported in Table 2, as wella sd ifferences in activation energies DE a calculated by using SMARTCyp.N ote that we used two versions of SMARTCyp (versions2and 3), which gave identicalr esultsf or E a .T he small differences in activation barrierf or all three hydroxylationr eactions are in line with the similar values for E a calculated at the B3LYP level of Density Functional Theory (DFT) by Leth et al [65] Leth modeled compound Ia sap orphyrin moietyw ithout side chains and with axial coordinating O 2 À and CH 3 S À ligands, and showed for example ad ifferencei nt he range of À6t o3kJ mol À1 in activation barrier when comparing 3'-methyl-OH-MF and 4'-OH-MF formation.T his is to be compared with the corresponding SMARTCyp value of À2kJmol À1 (Table 2).…”

“…(5)] of 0.0 AE 0.6 kJ mol À1 ,T able 2. The predicted identical E a valuesf or both pathways ( Table 2) thus suggestasimilar binding free energy for the corresponding catalytically active poses (DDG bind = DÀDE a )w hereas the lower B3LYP-D3 value of Leth et al (by 11 kJ mol À1 ) [65] hints at preferred binding in the pose enabling4 ' hydroxylation.T he preference of 4' over 5h ydroxylation can again be understoodi nt erms of the higherp robability (lower entropic cost) of transition state formation for 4' hydroxylation compared to 5-OH product formation ( Table 2), probablya lso due to hydrogen bonding with Serine 72 in the catalytic-active binding pose in the formercase ( Figure S3).…”

“…using the B3LYP-D3 level of theory), resulting in lower E a value for 5-OH-MF formation by 14 kJ mol À1 . [65] In any case, from Table 2t he preference of 3'-methyl over 5-OH mefenamic acid formationb yP 450 BM3 M11c an be understood in terms of the higherp robability (lowere ntropic cost) of transition state formation for 3'-methyl hydroxylation,a si ndicated by the higher intercepto fo ur modified Arrhenius plots and the observed difference between D and the Curtin-Hammett estimate for DDG overall .T his observed difference is equal to TDDS°u nder the assumptiont hat trends in enthalpy and energy are equal,c f. Equation (8) This is in accord with our in silico data from refs. [32] and [33] showing strong hydrogen bondingi nteraction between mefenamic acid's carboxylate group and BM3 M11'sS er72 residue, which directss ubstrate-bindingo rientations for 3'-methyl hydroxylation to adopt ac atalytically active pose.…”

Use of Computational Biochemistry for Elucidating Molecular Mechanisms of Nitric Oxide Synthase

“…Numerous experimental data are available on the structure of the NOS heme domain in the closed state, but several aspects remain to be elucidated. For comparison with a homolog, extensive in silico studies (e.g., microsecond all-atom and coarse-grain MD, docking) have been reported on P450 enzymes [ [73] , [74] , [75] , [76] , [77] , [78] ], including the investigation of the structure of the active site, substrate tunneling and binding modes, mutational effects, and interaction with membranes. These show the usefulness of molecular modeling in a context similar to NOSox.…”

Fast Methods for Prediction of Aldehyde Oxidase-Mediated Site-of-Metabolism

“…The potential SOMs in the molecules are not equally accessible (cf. Table S2, Supplementary Material), but we have not been able to establish a correlation between reactivity and accessibility for the present dataset as we have previously done for CYP mediated metabolism [25,26]. The effect of the protein, i.e.…”