Kinetics of cytochrome P450 3A4 inhibition by heterocyclic drugs defines a general sequential multistep binding process

Abstract

Cytochrome P450 (P450) 3A4 is the enzyme most involved in the metabolism of drugs and can also oxidize numerous steroids. This enzyme is also involved in one-half of pharmacokinetic drug-drug interactions, but details of the exact mechanisms of P450 3A4 inhibition are still unclear in many cases. Ketoconazole, clotrimazole, ritonavir, indinavir, and itraconazole are strong inhibitors; analysis of the kinetics of reversal of inhibition with the model substrate 7-benzoyl quinoline showed lag phases in several cases, consistent with multiple structures of P450 3A4 inhibitor complexes. Lags in the onset of inhibition were observed when inhibitors were added to P450 3A4 in 7-benzoyl quinoline O-debenzylation reactions, and similar patterns were observed for inhibition of testosterone 6β-hydroxylation by ritonavir and indinavir. Upon mixing with inhibitors, P450 3A4 showed rapid binding as judged by a spectral shift with at least partial high-spin iron character, followed by a slower conversion to a low-spin iron-nitrogen complex. The changes were best described by two intermediate complexes, one being a partial high-spin form and the second another intermediate, with half-lives of seconds. The kinetics could be modeled in a system involving initial loose binding of inhibitor, followed by a slow step leading to a tighter complex on a multisecond time scale. Although some more complex possibilities cannot be dismissed, these results describe a system in which conformationally distinct forms of P450 3A4 bind inhibitors rapidly and two distinct P450-inhibitor complexes exist en route to the final enzyme-inhibitor complex with full inhibitory activity.

Keywords: UV–visible spectroscopy; cytochrome P450; enzyme inhibitor; enzyme kinetics; enzyme mechanism; pre–steady-state kinetics.

Figure 1

Structures of inhibitors used in this work. The formula weights and approximate molecular volumes (∼18 Å³/non-H atom) are also shown.

Figure 2

Inhibition of cytochrome P450 3A4-catalyzed 7-benzoyl quinoline O-debenzylation and testosterone 6β-hydroxylation. The steady-state rates are plotted as functions of (log₁₀) inhibitor concentration, with the inhibitor indicated on the x-axis labels. A, C, E, G, and I (left panels): 7-benzoyl quinoline O-debenzylation; B, D, F, H, and J (right panels): testosterone 6β-hydroxylation. IC₅₀ values are shown on the graphs. All fits had r² values of ≥0.94. The 95% confidence intervals were as follows: A, 0.07 to 0.17 μM; B, 0.08 to 0.22 μM; C, 0.05 to 0.23 μM; D, 0.05 to 0.24 μM; E, 0.22 to 0.60 μM; F, 0.22 to 0.36 μM; G, 0.14 to 0.31 μM; H, 0.52 to 1.07 μM; I, 1.1 to 3.8 μM; and J, 0.94 to 2.7 μM.

Figure 3

Kinetics of recovery of catalytic activity from cytochrome P450 3A4–inhibitor complexes upon addition of 7-benzoyl quinoline.A, control (no inhibitor); B, ketoconazole; C, clotrimazole; D, ritonavir; E, indinavir; and F, itraconazole. Linear extrapolations to zero product formation are shown (red lines) in parts C, D, and E.

Figure 4

Pre–steady-state kinetics of inhibition of cytochrome P450 3A4-catalyzed O-debenzylation of 7-benzoyl quinoline. Plots of ΔF_410/>510 are shown for A, ketoconazole; B, clotrimazole; C, ritonavir; D, indinavir; and E, itraconazole. The (final) inhibitor concentrations are shown on the individual graphs. The data were fit to the equation y = A(1 − e^−k₁^t) + k_sst (Table 1).

Figure 5

Pre–steady-state kinetics of inhibition of cytochome P450 3A4-catalyzed 6β-hydroxylation of testosterone. Plots of product formation are shown: A, no inhibitor added (●), 0.4 μM ritonavir (▪), and 1.2 μM ritonavir (▲). B, no inhibitor added (●, same plot as in part A), 1.5 μM indinavir (▪), and 4 μM ritonavir (▲). The data points for the uninhibited reaction were fit by linear regression. The data points for the inhibitors in parts A and B could be fit to the expression y = A (1 − e^−k₁^t) + k_sst: A, k₁ = 0.062 s⁻¹ (t_1/2 11 s) and k_ss = 0.086⁻¹ (28% of uninhibited rate) for 1.2 μM ritonavir; B, k₁ = 0.17 s⁻¹ (t_1/2 4.0 s) and k_ss = 0.13 s⁻¹ (42% of uninhibited rate) for 1.2 μM ritonavir.

Figure 6

Absorbance changes at 390 and 425 nm observed upon mixing ketoconazole with cytochome P450s (P450s).A, steady-state spectra of P450 3A4 (5 μM) in the absence and presence of 15 μM ketoconazole (in 100 mM potassium phosphate, pH 7.4). Addition of more ketoconazole did not produce further changes, as would be expected from the submicrometer IC₅₀ (Fig. 2) and K_d values for ketoconazole and P450 3A4. B, difference spectrum from part A, with the P450-only spectrum subtracted from the spectrum obtained in the presence of ketoconazole. C, P450 3A4 (2 μM, final) was mixed with ketoconazole (15 μM). The indicated traces were obtained when the P450s were mixed with only the buffer (100 mM potassium phosphate, pH 7.4). The data were collected in the OLIS Show Pre-trigger Mode, with 0.1 s of data from the previous run (completed reaction) shown prior to actual mixing. D, expansion of data from experiment in part C, with the reaction being observed after 0.1 s (100 ms). The time points (7 and 24 ms) are calibrated for time after the initiation of the reaction (100 ms). E, spectra obtained upon P450 3A4 and ketoconazole 7 and 24 ms after reaction (from parts C and D), with the 7 ms spectrum reflecting mostly unbound P450 and the 24 ms spectrum reflecting the first observed P450·ketoconazole complex. F, difference spectrum generated from part E by subtracting the 7 ms spectrum from the 24 ms spectrum.

Figure 7

Effect of cytochrome b₅on spectral interaction of cytochome P450s (P450s) with ketoconazole. P450 (2 μM, final) was mixed with ketoconazole in the absence or presence of 2 μM cytochrome b₅. A, P450 3A4 (±cytochrome b₅) mixed with 2 μM ketoconazole. B, second-order plot of binding data (from part A, the absence of cytochrome b₅). With equal amounts of P450 3A4 and ketoconazole, the reaction X + Y → Z (where X is P450 3A4, Y is ketoconazole, and Z is the complex) is mathematically equivalent to 2X → Z and a plot of 1/(unbound P450 [X]) versus time yields an apparent second-order rate constant as the slope (5 × 10⁵ M⁻¹ s⁻¹) (52).

Figure 8

Binding of ketoconazole to cytochome P450 (P450) 3A4. The final P450 3A4 concentration was 2 μM, and the final ketoconazole concentration was 15 μM (in 100 mM potassium phosphate buffer, pH 7.4). A, spectral traces at 390 and 425 nm. B, spectra acquired at the indicated times after mixing, as indicated. The data were collected in the OLIS Show Pre-trigger Mode, with 2.1 s of data from the previous run (completed reaction) shown prior to mixing. C, time course of early ΔA₃₉₀ − A₄₂₅ change in the early phase after mixing. D, traces of ΔA₄₂₅ − A₃₉₀ following the first 18 s after mixing, as a function of ketoconazole concentration. E, plots of k_obs values from biexponential fits of data of part D plotted versus final ketoconazole concentration.

Figure 9

Singular value decomposition analysis of binding of ketoconazole to cyotochrome P450 (P450) 3A4. The final concentrations (after mixing) of P450 3A4 and clotrimazole were 2 and 15 μM, respectively. The OLIS GlobalWorks model used was a three-species 1 → 2 → 3 (A → B → C in software) fast/slow rate model, where the unbound P450 3A4 is not included, and 1, 2, and 3 are three different P450 3A4·ketoconazole complexes (this sequence would begin ∼100 ms after mixing P450 3A4 and ketoconazole; Fig. 6, C–D). A, spectra of the three intermediate species (species 1—blue; species 2—red, and species 3—black); B, time courses of the three species (same color pattern), plus a plot of the total absorbance accounted for at each time point; C, total residual plot for the kinetics traces. The data were collected in the Show Pre-trigger Mode, with 2.1 s of data from the previous run (completed reaction) showing prior to mixing (not reflected in part B). See Table 1 for calculated rates.

Figure 10

Analysis of kinetics of inhibition of cytochrome P450 3A4-catalyzed inhibition of 7-benzoyl debenzylation by ketoconazole.A, basic model; B, ketoconazole (red: no inhibitor; green: 0.12 μM ketoconazole; and blue: 4 μM ketoconazole). The upper section (r/σ) is a residual plot. See similar analyses for ritonavir, indinavir, and itraconazole in Figures S11–S13, with values for k₋₃, k₄, and k₋₄ adjusted.

Figure 11

Scheme summarizing interaction of cytochome P450 3A4 with inhibitors. The times of appearance of individual species are indicated in blue. E, enzyme; I, inhibitor, S, substrate.