Cooperative Binding of Acetaminophen and Caffeine within the P450

Sep 26, 2007 - ReceiVed March 5, 2007 .... Medicinal Chemistry, University of Washington, Box 357610, Seattle, ... 2007 American Chemical Society...
0 downloads 0 Views 578KB Size
Download PDF
1434

Chem. Res. Toxicol. 2007, 20, 1434–1441

Cooperative Binding of Acetaminophen and Caffeine within the P450 3A4 Active Site Michael D. Cameron, Bo Wen, Arthur G. Roberts, William M. Atkins, A. Patricia Campbell, and Sidney D. Nelson* Department of Medicinal Chemistry, UniVersity of Washington, Seattle, Washington 98195 ReceiVed March 5, 2007

Acetaminophen (N-acetyl-p-aminophenol, APAP) is a commonly used analgesic/antipyretic. When oxidized by P450, a toxic APAP metabolite is generated. Human P450 3A4 was expressed in Escherichia coli, purified, and reconstituted using artificial liposomes. Oxidation of APAP by P450 3A4, as detected by the formation of its glutathione adduct, was found to exhibit negative homotropic cooperativity with a Hill coefficient of 0.7. In the presence of caffeine, the observed kinetics were close to classical Michaelis–Menten kinetics with a Hill coefficient approaching 1. In order to probe for a potential repositioning of APAP within the P450 3A4 pocket in the presence of caffeine, NMR T1 paramagnetic relaxation techniques were used to calculate distances from the P450 3A4 heme iron to protons of APAP alone and in the presence of caffeine. Both APAP and caffeine were found to bind at the active site in proximity to the heme iron. When APAP was incubated with P450 3A4, the acetamido group of APAP was found to be closest to the heme iron consistent with the amide group of APAP weakly associating with the heme iron. The addition of caffeine disrupted the ability of APAP to coordinate with the heme iron of P450 3A4 and enhanced the rate of oxidation to its toxic metabolite. Introduction Acetaminophen is widely used throughout the world because of its analgesic and antipyretic properties, and is generally considered safe, although acetaminophen overdose can lead to hepatotoxicity and death (1). After therapeutic doses, most of a dose of APAP is metabolized and excreted through sulfation or glucuronidation pathways; however, a small proportion is metabolized by cytochrome P450s (1, 2). Different cytochrome P450s have been shown to metabolize APAP to two primary metabolites in differing ratios, the nontoxic 3-OH-APAP and the toxic metabolite N-acetyl-p-benzoquinone imine (NAPQI) (2–9). Metabolism of APAP by P450 3A4 almost exclusively yields the toxic NAPQI (9). In the presence of caffeine, the rate of NAPQI production in rat liver microsomes is increased about 3-fold (10). Activation of the rat isoform P450 3A2 is believed to be responsible. Similar increases in APAP metabolism were observed in the presence of R-naphthoflavone (11). The rate of APAP oxidation by human P450 3A4 is also activated in the presence of caffeine. The formation of multi-substrate or substrate/activator complexes within the active site of P450 3A4 has generated considerable scientific interest (12–16). Recently solved X-ray crystal structures of P450 3A4 have shown that the active site can be altered to accommodate diverse ligands (17–19) and can simultaneously accommodate two molecules of ketoconazole (19). In line with this thinking, APAP and caffeine may simultaneously occupy the active site. Additionally, binding of caffeine may reposition APAP within the active site giving rise to the increased rate of oxidation. APAP and caffeine are an attractive allosteric combination to study because caffeine causes kinetic changes in APAP

oxidation but does not effectively compete for the reactive oxygen species generated by P450 3A4. While caffeine may be hydroxylated at the 8 position, it is a very poor substrate with both a high KM (≈50 mM) and a low VMax (20). Because of this, caffeine metabolism can be ignored in the kinetic analysis of APAP oxidation. Several researchers have used NMR T1 relaxation rate measurements to look at the positioning of a substrate with respect to a paramagnetic center, in this case, the iron center of the heme. The T1 relaxation measurements of APAP with P450 1A1 and P450 2B1 have been previously reported (21, 22), as has caffeine with P450 1A2 (23). More recently, NMR T1 relaxation experiments have shown that the drug dapsone alters the orientation of flurbiprofen within the active site of P450 2C9 (24) and that testosterone and R-naphthoflavone alter the positioning of midazolam within the active site of P450 3A4 (25). Relaxation measurements are well suited for the examination of APAP and caffeine in the P450 3A4 active site because both substrates are in fast exchange on the NMR timescale, they do not have overlapping or complicated NMR spectra, and they are both soluble at millimolar concentrations. Moreover, T1 relaxation measurements provide relative distances of each substrate to the heme iron and thus are particularly suited for probing P450 3A4 allosterism as they can detect changes in the relative orientation of the first substrate (APAP) in the presence of the second (caffeine). Ultimately, this information allows for a more structurally detailed and conceptually simpler model than that provided by the kinetic studies previously described in the literature.

Materials and Methods * To whom correspondence should be addressed. Department of Medicinal Chemistry, University of Washington, Box 357610, Seattle, Washington 98195. Tel: (206) 543-1419. Fax: (206) 685-3252. E-mail: [email protected].

Chemicals Used. Deuterated solvents were purchased from Cambridge Isotope Laboratories (Andover, MA). Chelex 100 resin was purchased from Bio-Rad, (Hercules, CA). Lipids were pur-

10.1021/tx7000702 CCC: $37.00  2007 American Chemical Society Published on Web 09/26/2007

Simultaneous Binding of APAP and Caffeine by P450 3A4 chased from Avanti Polar Lipids Inc. (Alabaster, AL). All other chemicals were purchased from Sigma Chemical Co. (St. Louis, MO), in the highest purity available. All solutions were prepared with Mili-Q treated water with a specific resistance of 17.8 MΩ or greater. Buffers were double Chelex treated with the Chelex column regenerated between passages over the column. All solutions used for NMR experiments were stored over regenerated Chelex and passed the ascorbic acid stability test for transition metals (26). Production of Purified P450 3A4. Recombinant P450 3A4 was produced in Escherichia coli DH5R using the expression vector pCW 3A4-His6 kindly provided by Dr. Ron Estabrook. In a 2.8 L Fernbach flask, cells were shaken at 150 RPM and 27 °C for 48 h. Media and remaining expression conditions are as described by Gillam et al. (27). Pelleted cells were resuspended in resuspension buffer: 100 mM Tris-HCl, pH 7.4, 50µM testosterone, and 20% glycerol with the addition of protease inhibitor cocktail from Sigma Chemical Co. (1 mL per liter of initial culture volume). Lysozyme 5 mg/L of culture was added and allowed to stir at 4 °C for 1 h. Cells were homogenized and spun at 150,000g, and the yellowish supernatant was discarded. The pellet was resuspended in resuspension buffer by homogenization and allowed to stir at 4 °C for 1 h after the addition of 1% Emulgen 911. The solution was centrifuged at 150,000g. The red/orange supernatant had imidazole added to a final concentration of 25 mM and was directly loaded onto a column of ProBond nickel resin from Invitrogen. The column was washed with 20 column volumes of wash buffer: 100 mM TrisHCl, pH 7.4, 20% glycerol, 40 mM imidazole, 0.05% cholate, and 50 µM testosterone. The column was eluted with a minimal volume of elution buffer: 100 mM Tris-HCl, pH 7.4, 20% glycerol, 500 mM imidazole, and 0.02% cholate. The eluted protein was dialyzed against 100 mM potassium phosphate at pH 7.4 in 20% glycerol and stored at -80 °C. Final yields after purification were typically 200 nmol/L of culture. Expression and Purification of Cytochrome b5 and NADPH-Cytochrome P450 Oxidoreductase. Both cytochrome b5 and NADPH-cytochrome P450 oxidoreductase were expressed in E. coli. Human cytochrome b5 expression plasmid was provided by Dr. Ron Estabrook and expressed in BL21-DE3 cells using previously described conditions (28). Purification of cytochrome b5 was similar to the purification of P450 3A4. Expression and purification of NADPH-cytochrome P450 oxidoreductase was carried out as described by Shen et al. (29). Incubations. Premixes for P450 3A4 were prepared using the protocol described by Shaw et al. (30). Briefly, two stock solutions, a 5× buffer premix and a 5× protein premix, were prepared and subaliquotted into 500 µL fractions, which were frozen at -80 °C. The final concentration of the 5× buffer premix was 12 mM GSH and 150 mM Mg2+ in 200 mM potassium HEPES buffer at pH 7.4. Final concentrations for the 5× protein premix were 0.5 µM P450 3A4, 1.0 µM NADPH-P450 reductase, 0.5 µM cytochrome b5, 0.5 mg/mL CHAPS, 0.033 mg/mL of L-R-dilauroyl-sn-glycero3-phosphocholine, 0.033 mg/mL of L-R-dioleoyl-sn-glycero-3phosphocholine, 0.033 mg/mL of L-R-dilauroyl-sn-glycero-3phosphoserine, and 3 mM GSH in 50 mM potassium HEPES buffer at pH 7.4. Stock solutions were freshly made the morning the premixes were prepared. To prepare the protein premix, the three phospholipids, freshly prepared in buffer using minimal sonication, were mixed with the P450, B5, and reductase on ice with gentle agitation for 30 min. A total volume of 5 mL was prepared at twice the concentration listed above. The mixture was dialyzed in a SlideA-Lyzer dialysis cassette (Pierce, Rockford, IL) three times against 1 L of 50 mM potassium HEPES buffer, pH 7.4, at 4 °C for 1 h to remove the glycerol from the enzyme storage buffers. The enzyme/ phospholipid solution was removed from the dialysis cassette and an aliquot was removed to spectroscopically determine the P450 concentration. This dialyzed enzyme/phospholipid solution was mixed with an appropriate volume of 2.5 mg/mL CHAPS and 15 mM GSH, prepared in 50 mM potassium HEPES buffer, pH 7.4, and diluted to the targeted concentrations listed above with HEPES buffer.

Chem. Res. Toxicol., Vol. 20, No. 10, 2007 1435 Incubations were performed for 5 min at 37 °C by adding 20 µL of both premixes, 7 µL of 100 mM GSH, and a combined total of 43 µL water/substrate. The reactions were initiated by the addition of 10 µL of 10 mM NADPH. After 5 min, the reaction was stopped by adding 10 µL of a 70% w/v perchloric acid and centrifuged for 5 min at 14,000g to spin down protein. APAP metabolism was determined by measuring its glutathione conjugate (GS-APAP) using HPLC. An Econosil C18 100 Å 5 µm (7 × 33 mm) column (Alltech, Deerfield, IL) was used with a mobile phase of 25 mM ammonium phosphate, pH 5.0, and 10% methanol. The GS-APAP derivative was detected using a HP 1049A electrochemical detector at 0.40 V. Difference Spectra. Dissociation constants were determined using a Cary 3E spectrophotometer by titrating APAP or caffeine into cuvettes containing purified P450 3A4. The sample and reference cuvettes each contained 0.25 µM P450 3A4 in 100 mM potassium phosphate buffer at pH 7.4. Upon the addition of APAP or caffeine to the sample cuvette, an equal volume of buffer was added to a reference cuvette. The absorbance differences between the peak and trough were plotted and corrected for dilution. Dissociation constants in D2O were assumed to be equal to those in H2O. P450 3A4 spin state was evaluated under conditions similar to those used in the NMR experiments using spectroscopy and a split cell cuvette. The cuvette was divided into two chambers, each having a path length of 5 mm, by a quartz dividing wall, half the height of the cuvette. One chamber contained 2 µM P450 3A4, while the other contained an equal volume of 20 µM APAP, caffeine, or both APAP and caffeine. A background spectrum was obtained prior to mixing. The cuvette was inverted three times to mix the components of the two chambers, and the spectrum was reacquired. Changes in P450 3A4 spin state were estimated using an extinction coefficient of 126 mM-1 cm-1 (31). The initial spin state was estimated to be 19% high spin based upon the second derivative of the Soret peak (32). NMR Relaxation Study. Proton NMR measurements were carried out at 500 MHz using a Bruker DRX499 spectrometer. The purified P450 3A4 was exchanged into deuterated 100 mM potassium phosphate at pH 7.4. The exchange buffer was previously prepared by treating a 100 mM potassium phosphate solution twice with Chelex-100 to remove trace amounts of transition metals in the buffer. The treated phosphate buffer was then lyophilized and resuspended in D2O. This was repeated three times to exchange the hydrogen for deuterium. The deuterated phosphate buffer was exchanged into the sample using an Amicon Ultra 10,000 MW cut-off spin concentrator. It was observed that upon removing glycerol, the P450 protein occasionally aggregated. To eliminate aggregation, P450 3A4 was reconstituted with freshly prepared DLPC in D2O at a molar ratio of 1:300. A final P450 3A4 concentration of 1 µM was used for all NMR relaxation studies. The sample was purged with N2 gas to remove any dissolved oxygen in the buffer immediately before NMR spectra were taken. The sample (0.6 mL) was contained in a standard 5 mm NMR tube. An inversion–recovery pulse sequence with 10 t values was used to measure the observed longitudinal relaxation time (T1,obs). The t values were as follows: 0.1 s, 0.2 s, 0.3 s, 0.5 s, 0.75 s, 1 s, 1.5 s, 2 s, 4 s, and 8 s. The proton relaxation time measurement was carried out at a temperature of 298 K. For each t value, eight scans were performed with four dummy scans to reduce the proton signals from water. The water proton signals were suppressed using a watergate pulse sequence (33). A brief explanation of equations used to measure protein–heme distances is given below. Subtraction of the relaxation rate of a diamagnetic control from the total relaxation rate was used to determine the paramagnetic contribution to relaxation (34). In our experiments, the T1 relaxation time was measured before and after conversion of the P450 to its diamagnetic sodium dithionite-reduced carbon monoxide complex. The paramagnetic relaxation time (T1,P) was calculated from T1,obs(Fe3+) and T1,obs(Fe2+) according to equation 1 (21, 23).

1436 Chem. Res. Toxicol., Vol. 20, No. 10, 2007

Cameron et al.

1/T1,p)1/T1,obs(Fe3+) - 1/T1,obs(Fe2+)

(1)

T1,P may be determined in this manner because the contribution of unbound substrate and other diamagnetic effects of substrate binding cancel out. Distance and T1p. The distance to the heme iron can be solved directly using the Solomon–Bloembergen equation to relate paramagnetic relaxation times and distances to the paramagnetic iron (34–36),

(

1 1 )R T1P T1M + τm and

)

(

7τc 3τc 1 2γ12g2β2S(S + 1) ) + 2 2 6 T1M 15r 1 + ωI τc 1 + ωs2τc2

(2)

)

(3)

where R represents the fractional binding of ligand to total ligand concentration, γ1 is the nuclear gyromagnetic ratio, g is the isotropic electronic g factor, β is the Bohr magneton, S is the electronic spin of the paramagnetic center (5/2 for high and ½ for low spin iron), r is the distance of the proton from the heme iron, wI and wS are the nuclear and electron Larmor frequencies, respectively, and tc is the correlation time of the electron and proton dipolar interaction. T1P and T1M are related by the percentage of substrate bound and the lifetime of the substrate–enzyme complex (34). The R value is the fraction of ligand bound to total ligand concentration. Under fast exchange, τm, which is the lifetime of the substrate–enzyme complex, is much smaller than T1M, and τm drops out of equation 2. The Solomon–Bloembergen equation can be simplified for our experiments because wI2tc2 , 1 and ws2tc2 . 1. Additionally, γ12g2β2 is equal to 2.47 × 1017 s-2 Å-6. Thus, eqs 2 and 3 can be combined and simplified to the following:

1 9.87 × 10 S(S + 1) )R (τc) T1P r6

between the symmetric protons of the aromatic ring) away from the heme iron. In this case, 1/T1p1 * 1/T1p2 let us assume that 1/T1p1 and 1/T1p2 correspond to the closer and farther proton, respectively. Examination of equation 4 shows that 1/T1p is proportional to 1/r6. Therefore,

1 T 1 1p1 1 1 1 1 >> and ≈ ≈ T1p1 T1p2 T1p,apparent 2 2 T1p1

()

(7)

For example, if the nearest proton was 6 Å from the heme iron, then the distant proton would be 10.2 Å from the heme iron. Then the distant proton would only contribute ∼4% to the observed T1P. Fractional Binding of Ligand. The R value represents the fractional binding of ligand to total ligand concentration. For the binding of APAP to P450 3A4, the R value is shown in the following equation:

R)

[E0] [EA] ) [A0] KDA + [A0]

(8)

where A is APAP, A0 is total APAP, E is P450 3A4, E0 is total P450 3A4, EA is the P450 3A4–APAP complex, and KDA is the dissociation constant of APAP. When caffeine is introduced into the system, APAP and caffeine can bind simultaneously. This binding can be modeled with the Adair–Pauling binding model. In this case, the R value and equilibria will be as follows:

[E0] [E0][C0] + KDA KDAKDCA [EA] + [ECA] (9) R) ) [A0] [C0] [A0][C0] [A0] 1+ + + KDA KDC KDAKDCA where C is caffeine, KDC is the caffeine dissociation constant, and

16

(4)

A tc value of 3 × 10-10 was used for the calculations. This value represents an average of all tc values quoted for P450s in the literature (most recent tc values reported for P450 fall within the range of 1 × 10-10 to 6 × 10-10) and corresponds to the value our laboratory found for P450 1A2 (23). Orientation, fractional binding, and spin state all have significant effects on the apparent 1/T1p and are examined in more detail below. Orientation Dependence of 1/T1p. The apparent paramagnetic relaxation rate for a given proton resonance represents the weighted average of the paramagnetic relaxation rates of the individual protons. These paramagnetic relaxation rates can differ considerably between individual protons. In the case of APAP, the aromatic protons at positions 2/6 and 3/5 on the phenyl ring are chemically equivalent on opposite sides of the aromatic ring. The contribution of the individual protons on the 1/T1p will depend on its relative orientation. In the current study, we calculated distances using the two extremes: APAP binding with the aromatic ring protons parallel to the heme plane and equidistant to the heme iron and with binding perpendicular to the heme plane. When APAP aromatic ring protons are equidistant to the heme iron, the weighted average of these contributions will be as follows:

1 1 + T1p1 T1p2 ) T1p,apparent 2 1

S(S + 1) ) fHSSHS(SHS + 1) + fLSSLS(SLS + 1)

(10)

S(S + 1) ) 8.75fHS + 0.75fLS

(11)

where fHS and fLS are the fraction of high and low spin heme iron, respectively.

Results (5)

In the parallel orientation it will be as follows:

1 1 1 1 1 ) and ) ) T1p1 T1p2 T1p,apparent T1p1 T1p2

KDCA is the dissociation constant for caffeine when APAP is bound to the enzyme. The Kd values used in the article were experimentally determined from spin-state titration studies. Concentrations unless otherwise noted were 1 µM P450 3A4, 10 mM APAP, and 10 mM caffeine. Spin State. To account for the mixed spin state of the enzyme, the S(S + 1) component of eq 4 can be approximated as a linear combination of high (SHS ) 5/2) and low (SLS ) 1/2) spin states as follows:

(6)

When the aromatic protons of APAP are perpendicular to the heme plane, there will be a proton closer and 4.2 Å (the distance

Spin-state changes upon the addition of APAP and caffeine were examined using visible spectroscopy. Difference spectra were obtained upon the addition of APAP and caffeine to P450 3A4. APAP caused a low-spin transition with a ∆Amax at 420 nm and a ∆Amin at 390 nm (Figure 1A), consistent with a reverse type-I ligand. In contrast, caffeine caused a high-spin transition with a ∆Amax at 393 nm and a ∆Amin at 427 nm, consistent with a type-I ligand (Figure 1B). When these absorbance

Simultaneous Binding of APAP and Caffeine by P450 3A4

Chem. Res. Toxicol., Vol. 20, No. 10, 2007 1437

Figure 2. Difference spectra upon APAP, caffeine, or APAP-caffeine co-addition to P450 3A4. Split-cell cuvettes were used with 2 µM P450 3A4 on one side of the partition and 20 mM APAP, 20 mM caffeine, or 20 mM of both APAP and caffeine on the other side of the cuvette. The background spectrum was aquired prior to mixing the two chambers by inverting the cuvette three times. Final concentrations were 1 µM P450 3A4 and 10 mM APAP, caffeine, or both. Each condition was done in triplicate, and representative spectra are shown.

Figure 1. P450 3A4 difference spectra upon APAP or caffeine titration. Equal volumes of 0.25 µM P450 3A4 in 100 mM phosphate buffer, pH 7.4, were added to the sample and reference cuvette. Aliquots of APAP (Figure 1A) or caffeine (Figure 1B) were added to the sample cuvette, and an equal volume of buffer was added to the reference cuvette. Absorbance readings were taken every 0.2 nm from 500 to 350 nm. The resulting ∆A of the P450 3A4 heme soret difference spectra were fit to Lineweaver–Burke Plots and are shown as inserts. The maximum ∆A values observed for the titrations of APAP and caffeine were 2.7 and 9.7 milli-absorbance units, respectively.

changes were fit to a Lineweaver–Burke plot, a linear relationship was observed for APAP, and the KS was determined to be 77 ( 9 µM. Two distinct phases were observed for caffeine. The caffeine-induced spin-state change data was fit to an ordered two-site kinetic model, and KS values of 62 ( 33 µM and 12.9 ( 2.9 mM were determined. In the presence of 10 mM APAP, caffeine-induced spin-state change fit best to a single-site model and had a calculated KS value of 740 ( 50 µM. Difference spectra were also obtained to determine the change in spin state upon the addition of 10 mM APAP, 10 mM caffeine, and 10 mM of both APAP and caffeine to solutions containing P450 3A4 (Figure 2). The calculated spin states were used in the calculations of subsequent NMR studies. The addition of 10 mM APAP to P450 3A4 caused a low-spin transition, corresponding to a decrease in the percentage of highspin heme iron from 19% to 11%. Conversely, the addition of 10 mM caffeine caused a high-spin transition with an associated increase in the percentage of high-spin ferric iron from 19% to 48%. Interestingly, the addition of both APAP and caffeine caused the greatest increase in the percentage of high-spin iron (19% to 52%), even though APAP is typically associated with low-spin transitions in P450 3A4. The cooperative relationship of APAP and caffeine in the promotion of high-spin P450 3A4 was observed with CPY3A4 from three separate expression

Figure 3. Kinetics of GS-APAP formation and effect of caffeine. Conditions for the incubation of various APAP and caffeine concentrations with reconstituted P450 3A4 are as described in the Materials and Methods section. A 24 × 4 matrix of APAP and caffeine concentrations was used. Calculated rates from incubations with APAP (b) and with APAP + 10 mM caffeine (O) are shown fit to an ordered-two-site model (solid line) and a Michaelis–Menten model (dashed line).

batches on separate days to ensure the observed interaction was not due to random error. While APAP-induced spin-state changes were found to be monophasic, APAP oxidation was found to be biphasic with respect to APAP concentration. Figure 3 plots the velocity of GS-APAP formation versus APAP concentration. APAP metabolism was increased by the addition of caffeine when APAP concentrations were above 200 µM; however, caffeine had a negative effect on the rate of APAP oxidation at APAP concentrations below 100 µM (Figure 3). Higher caffeine concentrations reduced the biphasic nature of APAP metabolism, as can be seen by the increased linearity of the Eadie–Hofstee plots as caffeine concentrations were increased (Figure 4). The Hill coefficients increased from 0.70 to 0.95 upon the addition of caffeine. Calculated kinetic constants are given in Table 1. For samples without caffeine or containing 1 mM caffeine, the results fit best to an ordered two-site kinetic model (37). Classical Michaelis–Menten equations gave a poor fit to the data at low APAP concentrations. Samples containing 5 or 10 mM caffeine fit reasonably well to the Michaelis–Menten equation (Figure 3). Calculated kinetic constants indicate that

1438 Chem. Res. Toxicol., Vol. 20, No. 10, 2007

Cameron et al.

Figure 4. Eadie–Hofstee Plots of GS-APAP formation by P450 3A4 in the presence of various levels of caffeine. GS-APAP formation was measured from 5 min incubations of P450 3A4 as described in the Materials and Methods section. Incubations were done in triplicate with 24 APAP concentrations and 4 caffeine concentrations. Hill coefficients were calculated by nonlinear regression. The data was weighted 1/X2, and the fits are shown in the inserts.

Table 1. Kinetic Constants for GS-APAP Formationa kinetic constants

KM1 VMax1 KM2 VMax2

APAP (mM)

APAP + 1 mM caffeine (mM)

APAP + 5 mM caffeine (mM)

APAP + 10 mM caffeine (mM)

0.097 ( 0.015 1.1 ( 0.1 3.7 ( 0.3 9.8 ( 0.7

0.12 ( 0.02 1.3 ( 0.1 3.8 ( 0.3 10 ( 1

0.51 ( 0.10 3.6 ( 0.6 4.5 ( 1.1 11 ( 2

0.76 ( 0.16 4.9 ( 1.0 4.8 ( 1.6 11 ( 2

a The Kinetic constants were determined by curvefitting the kinetic data to the equation V ) VMax × ([APAP]/KM1 + b × [APAP]2/(KM1 × KM2))/(1 + [APAP]/KM1 + [APAP]2/(KM1 × KM2)). The units for the KM values are in mM, and VMax values are in nmol GS-APAP × nmol P450 3A4-1 × min-1. Curve fitting was weighted by 1/V to minimize the relative errors.

Table 2. T1 Times of Free APAP and APAP in the Presence of Paramagnetic and Diamagnetic P450 3A4 observed proton relaxation times position 10 mM APAP

Figure 5. Structures of acetaminophen and caffeine.

the affinity of the tighter binding APAP is decreased 8-fold in the presence of elevated caffeine concentrations, while KM2 was minimally changed (Table 1). NMR T1 relaxation studies were performed in order to measure the distance from the heme iron to the protons of APAP and caffeine (Figure 5). Observed T1 relaxation times are reported in Table 2. Calculated distances to the 2-H and 3-H protons of APAP in the absence of caffeine indicate that the acetamido group of APAP is closer to the heme iron than is the phenol group, consistent with the amide group of APAP coordinating to the heme (Table 3). This is consistent with the increase in the low-spin state of the heme detected via optical spectroscopy. Samples containing 250 µM APAP gave distance calculations similar to those of 10 mM APAP samples.

APAP 2-H APAP 3-H APAP methyl 10 mM APAP + APAP 2-H 10 mM caffeine APAP 3-H APAP methyl caffeine 1′-methyl caffeine 3′-methyl caffeine 7′-methyl caffeine 8-H

T1 observed T1 observed paramagnetic(s) diamagnetic(s) 3.21 3.64 1.43 2.80 2.72 1.33 1.60 1.56 1.53 4.56

4.61 4.63 1.58 4.54 4.50 1.56 1.67 1.71 1.63 5.64

However, because of the better signal to noise ratio associated with the 10 mM samples and thus their smaller associated errors, only 10 mM APAP data is presented and discussed in this article. It is important to point out that the distances between the heme iron and the protons of the phenol ring of APAP are the closest approach to the heme iron and do not necessarily represent the average positioning within the active site. This is because distance has a sixth order effect on relaxation times, while the time an individual proton resides at a particular distance has a first order effect (34).

Simultaneous Binding of APAP and Caffeine by P450 3A4 Table 3. Distances Calculated for T1 Dataa calculated proton to heme iron distances position parallel APAP binding perpendicular APAP binding

APAP 2-H APAP 3-H APAP methyl APAP 2-H APAP 3-H APAP methyl caffeine 1′-methyl caffeine 3′-methyl caffeine 7′-methyl caffeine 8-H

10 mM APAP (Å)

10 mM APAP + 10 mM caffeine (Å)

6.0 ( 0.1 6.5 ( 0.1 6.4 ( 0.2 5.4 ( 0.1 5.8 ( 0.1 6.4 ( 0.2

6.8 ( 0.2 6.7 ( 0.2 7.0 ( 0.2 6.1 ( 0.2 6.1 ( 0.2 7.0 ( 0.2 8.9 ( 0.3 7.9 ( 0.2 8.4 ( 0.2 8.3 ( 0.3

a

Distance calculations were done using 1 µM P450 3A4, 10 mM APAP, 10 mM caffeine, 77 µM as the Kd for APAP, 740 µM as the Kd for caffeine, and 11% and 52% as the percent high-spin P450 3A4 in the APAP and APAP + caffeine samples, respectively. Protons on the APAP phenol ring were fit to the parallel and perpendicular binding models described in the Materials and Methods section.

Two models were used to evaluate the closest approach distance of APAP because the APAP protons at positions 2/6 and 3/5 of the phenyl ring are chemically (and NMR) equivalent but on opposite sides of the aromatic ring. Distances were calculated using the two extreme binding orientations where APAP binds parallel to the heme with the distance to both protons equivalent and with APAP binding perpendicular with the maximum discrepancy between the two equivalent protons (Table 3). When APAP and caffeine were co-incubated with P450 3A4, there did not appear to be a preferred orientation, as seen by the equivalent distances for all of the APAP protons. This is also consistent with optical spectroscopy results where caffeine decreased the low-spin APAP-induced transition because of the coordination of the APAP amide group to the heme iron. Calculated distances from the heme iron to the protons of caffeine were on average 1.5 to 2.5 Å longer than those calculated for APAP (Table 3). The shorter distance calculated for the 2-H proton of APAP in the absence of caffeine was confirmed for over 20 samples using varying APAP concentrations and with P450 3A4 from three different batches. In all cases, this preferred orientation was lost upon the addition of caffeine.

Discussion P450 3A4 is the major human liver enzyme responsible for xenobiotic oxidation in terms of both its concentration and its range of possible substrates. In order to accommodate the range of diverse compounds oxidized by P450 3A4, it is logical that the active site is both large and flexible. While this allows for a wider range of substrates, it also yields rather complicated and confusing kinetics and substrate–substrate interactions. A

Chem. Res. Toxicol., Vol. 20, No. 10, 2007 1439

recent P450 3A4 crystal structure demonstrated large differences in P450 3A4 active site geometry with the binding of ketoconazole and erythromycin (19). The same work also found that erythromycin had multiple binding modes implying that there are several potential sites for protein–ligand interaction within the active site. We realize that the amino acids of the protein are likely to be the dominant parameter influencing substrate binding in the active site of P450 3A4; however, at this stage, it is unclear that the active site geometry would be conserved between samples containing APAP, caffeine, or both. For this reason, we have limited our discussion to the geometric relationship of caffeine and APAP to the heme iron and not to the overall active site. The observed spin-state changes in cytochromes P450 are not simply a function of substrate binding within the P450 pocket. Spin-state changes arise from changes in the coordination sphere of the heme iron. While oversimplified with regard to P450, one can conceptually think of the pentacoordinate heme iron as high spin and the hexacoordinate heme iron as low spin. In the absence of substrate, water may form a sixth ligand with the heme iron of P450 3A4 resulting in a low-spin state. The binding of many substrates cause high-spin transitions in P450 by precluding the coordination of water and thus increasing the pentacoordinate nature of the heme iron. In the case of caffeine, the biphasic nature of the spin-state changes in P450 3A4 indicate a two-step process. This can be rationalized by the first caffeine being less effective at excluding the water ligand from coordinating the heme iron and the second caffeine as being more effective at excluding the water, resulting in a greater transition to high spin upon the binding of the second caffeine. APAP, however, can be visualized as increasing the percentage of hexacoordinate heme iron by the direct coordination of the APAP amide group to the heme iron, which increases the percentage of low-spin heme iron. The addition of caffeine decreased the allosteric nature of APAP oxidation by P450 3A4. Without the addition of caffeine, APAP significantly deviated from classical Michaelis–Menten kinetics and had a Hill coefficient of 0.70. It is notable that the concentration of caffeine where the Hill coefficient for APAP oxidation approaches 1 is more similar to the binding affinity for the second caffeine (KM2). The presence of this second caffeine appears to facilitate APAP oxidation. Since caffeine activation of APAP oxidation is similar to what is seen with R-NF (10), it is possible that the catalytically important caffeine resides in the region that R-NF is reported to occupy (38–40). Calculated distances for APAP are consistent with the amide group interacting with the heme iron at both 250 µM and 10 mM concentrations. A generalized model of APAP coordinating to the heme iron is depicted in Figure 6A. In the presence of

Figure 6. Schematic of APAP–heme orientation with (B) and without caffeine (A).

1440 Chem. Res. Toxicol., Vol. 20, No. 10, 2007

high concentrations of caffeine, it appears that APAP is precluded from direct coordination with the heme iron, as indicated by the increased APAP atom distances from the heme iron and the loss of specificity of orientation of the 2-H and 3-H protons. More than one model fits the observed results. Caffeine may overlay and interact with APAP to preclude its coordination with the heme iron, or it may modify the substrate binding pocket leading to a different APAP orientation. Figure 6B presents one possible orientation of APAP in the presence of caffeine. The NMR calculated distances for APAP and caffeine indicate that the two drugs are in proximity to each other and that interaction between them is possible. Although the calculated distances are too close to accommodate van der Waal’s contacts, the distances reflect the closest approach and not equilibrium distances. Caffeine is more removed from the heme iron compared to APAP as the calculated distances of closest approach for caffeine protons to the iron are 1.8 to 2.7 Å further than those calculated for APAP protons. This could account for the slow rate of caffeine oxidation by P450 3A4. The reorientation of APAP may provide more room near the face of the heme and facilitate oxygen binding or better positioning of APAP to interact with the reactive oxygen species generated by P450 3A4. Because both the reduced ferrous heme and the oxygen-bound perferryl heme are diamagnetic, it is not possible to examine these states by NMR T1 paramagnetic relaxation experiments. The changes in APAP orientation within the active site are more likely factors in the altered kinetics than facilitation of electron transfer from the reductase. While it is true that the rate of electron transfer in P450 is greater for high-spin P450–substrate complexes, we would expect a larger increase in the rate of APAP oxidation if heme reduction were the ratelimiting step. Additionally, the decreased rate of APAP oxidation observed in the presence of high caffeine concentrations and APAP concentrations below 100 µM does not agree with an explanation of enhanced electron transfer from the reductase. Spectroscopic titration indicates that one molecule of APAP is coordinated to the heme iron. The kinetics of APAP metabolism indicate the possibility that a second APAP binds and modulates the reaction rate. It is likely that there is a high affinity/low velocity site of APAP binding and a lower affinity but higher velocity binding site. Examination of the kinetics along with the NMR data would support a model where the high affinity/low velocity APAP binding site corresponds to the heme-ligated APAP. In the presence of caffeine, direct heme ligation is precluded, indicated by the enzymatic spin state and the equivalent proton distances for the 2-H and 3-H protons of APAP. This would account for the observed results where caffeine caused a decrease in the rate of APAP oxidation at low APAP concentrations but an increase rate at high APAP concentrations. In this study, we were able to show that the binding of caffeine disrupts the coordination of APAP to the heme iron. This disruption resulted in changes in the APAP orientation within the active site and decreased the homotropic allosterism exhibited by APAP. The altered positioning of one substrate by a second substrate/effector shown in this work is a likely explanation for much of the allosterism reported for P450 3A4. Acknowledgment. This work was supported by NIH Grant GM32165 (to S.D.N.) and the UW NIEHS sponsored Center for Ecogenetics and Environmental Health: NIEHS P30ES07033. We thank Dr. Matt Cheeseman for helpful discussions regarding

Cameron et al.

the expression and purification of recombinant P450 and Dr. Ronald Estabrook for providing the vector for P450 3A4 expression.

References (1) Prescott, L. F. (1996) Paracetamol (Acetaminophen): A Critical Bibliographic ReView. Taylor & Francis, London. (2) Nelson, S. D. and Bruschi, S. (2003) in Mechanisms of AcetaminophenInduced LiVer Disease. Drug-Induced LiVer Disease (Kaplowitz, N., and DeLeve, L. D., Eds.) 1st ed., pp 287–326, Marcel Dekker, New York. (3) Raucy, J. L., Lasker, J. M., Lieber, C. S., and Black, M. (1989) Acetaminophen activation by human liver cytochromes P450IIE1 and P450IA2. Arch. Biochem. Biophys. 271, 270–283. (4) Patten, C. J., Thomas, P. E., Guy, R. L., Lee, M., Gonzalez, F. J., Guengerich, F. P., and Yang, C. S. (1993) Cytochrome P450 enzymes involved in acetaminophen activation by rat and human liver microsomes and their kinetics. Chem. Res. Toxicol. 6, 511–518. (5) Lee, S. S., Buters, J. T., Pineau, T., Fernandez-Salguero, P., and Gonzalez, F. J. (1996) Role of CYP2E1 in the hepatotoxicity of acetaminophen. J. Biol. Chem. 271, 12063–12067. (6) Tonge, R. P., Kelly, E. J., Bruschi, S. A., Kalhorn, T., Eaton, D. L., Nebert, D. W., and Nelson, S. D. (1998) Role of CYP1A2 in the hepatotoxicity of acetaminophen: investigations using CYP1A2 null mice. Toxicol. Appl. Pharmacol. 153, 102–108. (7) Chen, W., Koenigs, L. L., Thompson, S. J., Peter, R. M., Rettie, A. E., Trager, W. F., and Nelson, S. D. (1998) Oxidation of acetaminophen to its toxic quinone imine and nontoxic catechol metabolites by baculovirus-expressed and purified human cytochromes P450 2E1 and 2A6. Chem. Res. Toxicol. 11, 295–301. (8) Dong, H., Haining, R. L., Thummel, K. E., Rettie, A. E., and Nelson, S. D. (2000) Involvement of human cytochrome P450 2D6 in the bioactivation of acetaminophen. Drug Metab. Dispos. 28, 1397–1400. (9) Thummel, K. E., Lee, C. A., Kunze, K. L., Nelson, S. D., and Slattery, J. T. (1993) Oxidation of acetaminophen to N-acetyl-p-aminobenzoquinone imine by human CYP3A4. Biochem. Pharmacol. 45, 1563– 1569. (10) Lee, C. A., Manyike, P. T., Thummel, K. E., Nelson, S. D., and Slattery, J. T. (1997) Mechanism of cytochrome P450 activation by caffeine and 7,8-benzoflavone in rat liver microsomes. Drug Metab. Dispos. 25, 1150–1156. (11) Li, Y., Wang, E., Patten, C. J., Chen, L., and Yang, C. S. (1994) Effects of flavonoids on cytochrome P450-dependent acetaminophen metabolism in rats and human liver microsomes. Drug Metab. Dispos. 22, 566–571. (12) Kenworthy, K. E., Bloomer, J. C., Clarke, S. E., and Houston, J. B. (1999) CYP3A4 drug interactions: correlation of 10 in vitro probe substrates. Br. J. Clin. Pharmacol. 48, 716–727. (13) Kenworthy, K. E., Clarke, S. E., Andrews, J., and Houston, J. B. (2001) Multisite kinetic models for CYP3A4: simultaneous activation and inhibition of diazepam and testosterone metabolism. Drug Metab. Dispos. 29, 1644–1651. (14) Shou, M., Dai, R., Cui, D., Korzekwa, K. R., Baillie, T. A., and Rushmore, T. H. (2001) A kinetic model for the metabolic interaction of two substrates at the active site of cytochrome P450 3A4. J. Biol. Chem. 276, 2256–2262. (15) Shou, M., Grogan, J., Mancewicz, J. A., Krausz, K. W., Gonzalez, F. J., Gelboin, H. V., and Korzekwa, K. R. (1994) Activation of CYP3A4: evidence for the simultaneous binding of two substrates in a cytochrome P450 active site. Biochemistry 33, 6450–6455. (16) Shou, M., Mei, Q., Ettore, M. W., Jr., Dai, R., Baillie, T. A., and Rushmore, T. H. (1999) Sigmoidal kinetic model for two co-operative substrate-binding sites in a cytochrome P450 3A4 active site: an example of the metabolism of diazepam and its derivatives. Biochem. J. 340, 845–853. (17) Williams, P. A., Cosme, J., Vinkovic, D. M., Ward, A., Angove, H. C., Day, P. J., Vonrhein, C., Tickle, I. J., and Jhoti, H. (2004) Crystal structures of human cytochrome P450 3A4 bound to metyrapone and progesterone. Science 305, 683–686. (18) Yano, J. K., Wester, M. R., Schoch, G. A., Griffin, K. J., Stout, C. D., and Johnson, E. F. (2004) The structure of human microsomal cytochrome P450 3A4 determined by X-ray crystallography to 2.05-A resolution. J Biol. Chem. 279, 38091–38094. (19) Ekroos, M., and Sjogren, T. (2006) Structural basis for ligand promiscuity in cytochrome P450 3A4. Proc. Natl. Acad. Sci. U.S.A. 103, 13682–13687. (20) Ha, H. R., Chen, J., Krahenbuhl, S., and Follath, F. (1996) Biotransformation of caffeine by cDNA-expressed human cytochromes P-450. Eur. J. Clin. Pharmacol. 49, 309–315. (21) Myers, T. G., Thummel, K. E., Kalhorn, T. F., and Nelson, S. D. (1994) Preferred orientations in the binding of 4′-hydroxyacetanilide

Simultaneous Binding of APAP and Caffeine by P450 3A4

(22)

(23)

(24) (25)

(26) (27)

(28)

(29) (30)

(acetaminophen) to cytochrome P450 1A1 and 2B1 isoforms as determined by 13C- and 15N-NMR relaxation studies. J. Med. Chem. 37, 860–867. van de Straat, R., de Vries, J., de Boer, H. J., Vromans, R. M., and Vermeulen, N. P. (1987) Relationship between paracetamol binding to and its oxidation by two cytochromes P-450 isozymes: a proton nuclear magnetic resonance and spectrophotometric study. Xenobiotica 17, 1–9. Regal, K. A., and Nelson, S. D. (2000) Orientation of caffeine within the active site of human cytochrome P450 1A2 based on NMR longitudinal (T1) relaxation measurements. Arch. Biochem. Biophys. 384, 47–58. Hummel, M. A., Gannett, P. M., Aguilar, J. S., and Tracy, T. S. (2004) Effector-mediated alteration of substrate orientation in cytochrome P450 2C9. Biochemistry 43, 7207–7214. Cameron, M. D., Wen, B., Allen, K. E., Roberts, A. G., Schuman, J. T., Campbell, A. P., Kunze, K. L., and Nelson, S. D. (2005) Cooperative binding of midazolam with testosterone and alphanaphthoflavone within the CYP3A4 active site: a NMR T1 paramagnetic relaxation study. Biochemistry 44, 14143–14151. Buettner, G. R. (1988) In the absence of catalytic metals ascorbate does not autoxidize at pH 7: ascorbate as a test for catalytic metals. J. Biochem. Biophys. Methods 16, 27–40. Gillam, E. M., Baba, T., Kim, B. R., Ohmori, S., and Guengerich, F. P. (1993) Expression of modified human cytochrome P450 3A4 in Escherichia coli and purification and reconstitution of the enzyme. Arch. Biochem. Biophys. 305, 123–131. Holmans, P. L., Shet, M. S., Martin-Wixtrom, C. A., Fisher, C. W., and Estabrook, R. W. (1994) The high-level expression in Escherichia coli of the membrane-bound form of human and rat cytochrome b5 and studies on their mechanism of function. Arch. Biochem. Biophys. 312, 554–565. Shen, A. L., Christensen, M. J., and Kasper, C. B. (1991) NADPHcytochrome P450 oxidoreductase. The role of cysteine 566 in catalysis and cofactor binding. J. Biol. Chem. 266, 19976–19980. Shaw, P. M., Hosea, N. A., Thompson, D. V., Lenius, J. M., and Guengerich, F. P. (1997) Reconstitution premixes for assays using

Chem. Res. Toxicol., Vol. 20, No. 10, 2007 1441

(31)

(32) (33) (34) (35) (36)

(37) (38)

(39)

(40)

purified recombinant human cytochrome P450, NADPH-cytochrome P450 reductase, and cytochrome b5. Arch. Biochem. Biophys. 348, 107–115. Harlow, G. R., and Halpert, J. R. (1998) Analysis of human cytochrome P450 3A4 cooperativity: construction and characterization of a sitedirected mutant that displays hyperbolic steroid hydroxylation kinetics. Proc. Natl. Acad. Sci. U.S.A. 95, 6636–6641. Guengerich, F. P. (1983) Oxidation-reduction properties of rat liver cytochromes P-450 and NADPH-cytochrome p-450 reductase related to catalysis in reconstituted systems. Biochemistry 22, 2811–2820. Piotto, M., Saudek, V., and Sklenar, V. (1992) Gradient-tailored excitation for single-quantum NMR spectroscopy of aqueous solutions. J. Biomol. NMR 2, 661–665. Villafranca, J. J. (1989) Paramagnetic probes of macromolecules. Methods Enzymol. 177, 403–413. Solomon, I. (1955) Relaxation processes in a system of two spins. Phys. ReV. 99, 559–565. Modi, S., Primrose, W. U., Boyle, J. M., Gibson, C. F., Lian, L. Y., and Roberts, G. C. (1995) NMR studies of substrate binding to cytochrome P450 BM3: comparisons to cytochrome P450 cam. Biochemistry 34, 8982–8988. Segel, I. H. (1975) Enzyme Kinetics: BehaVior and Analysis of Rapid Equilibrium and Steady-State Enzyme Systems, John-Wiley & Sons, New York. Domanski, T. L., Liu, J., Harlow, G. R., and Halpert, J. R. (1998) Analysis of four residues within substrate recognition site 4 of human cytochrome P450 3A4: role in steroid hydroxylase activity and alphanaphthoflavone stimulation. Arch. Biochem. Biophys. 350, 223–232. Harlow, G. R., and Halpert, J. R. (1997) Alanine-scanning mutagenesis of a putative substrate recognition site in human cytochrome P450 3A4. Role of residues 210 and 211 in flavonoid activation and substrate specificity. J. Biol. Chem. 272, 5396–5402. Hosea, N. A., Miller, G. P., and Guengerich, F. P. (2000) Elucidation of distinct ligand binding sites for cytochrome P450 3A4. Biochemistry 39, 5929–5939.

TX7000702