Chem. Res. Toxicol. 2002, 15, 843-853
843
Site-Directed Mutagenesis of Cytochrome P450eryF: Implications for Substrate Oxidation, Cooperativity, and Topology of the Active Site Kishore K. Khan,* You Ai He, You Qun He, and James R. Halpert Department of Pharmacology and Toxicology, University of Texas Medical Branch, 301 University Boulevard, Galveston, Texas 77555-1031 Received April 4, 2002
The role of five active-site residues (Phe-78, Gly-91, Ser-171, Ile-174, and Leu-175) has been investigated in P450eryF, the only bacterial P450 known to show cooperativity. The residues were selected based on two-ligand-bound P450eryF structures and previous mutagenesis studies of other cytochromes P450. To better understand the role of these residues in substrate catalysis and cooperativity, each mutant was generated in the wild-type and A245T background, a substitution that enables P450eryF to oxidize testosterone and 7-benzyloxyquinoline (7-BQ). Replacement of Phe-78 with tryptophan decreased cooperativity of 9-aminophenanthrene binding, with little effect on testosterone binding or oxidation. Interestingly, substitution of Gly-91 with alanine or phenylalanine abolished the type-I spectral change elicited by testosterone and significantly decreased testosterone hydroxylation. However, G91A/A245T showed a 4-fold higher kcat value with 7-BQ compared with A245T. Replacement of Ser-171 with alanine or phenylalanine did not alter cooperativity of testosterone binding but significantly decreased binding affinity and oxidation of testosterone and 7-BQ. The only mutant that exhibited an increased testosterone binding affinity and increased rates of testosterone and 7-BQ oxidation was I174F. Substitution of Ile-175 with phenylalanine decreased testosterone and 7-BQ oxidation. Reaction with phenyldiazene showed that P450eryF may be much more open above pyrrole ring B than other cytochromes P450 and indicated significant changes in active-site topology in some of the mutants. The study suggests a crucial role of residues Ser-171, Ile-174, and Leu-175, which are part of a distal ligand site, in addition to the proximal Gly-91 in determining the oxidative properties of P450eryF. Cytochromes P450 (P450)1 are a superfamily of hemeproteins containing more than a thousand members according to a recent estimate (1). These enzymes are involved in the oxidation of a wide variety of exogenous and endogenous compounds such as drugs, carcinogens, steroids, eicosanoids, alkaloids, pesticides, and environmental pollutants (2, 3). Although turnover numbers vary from 5000 min-1 (4), most of these reactions are highly regio- and stereospecific. Additionally, in the case of enzymes such as human P450 3A4 the presence of one substrate can either stimulate, inhibit, or have no effect on the metabolism of another substrate (5-13). Thus, the vast number of enzymes and their apparent complexity have made a proper understanding of the molecular basis of substrate specificity and selectivity of individual P450s very difficult. A thorough knowledge of the molecular interactions that control substrate specificity and catalytic activity is essential for the development of drugs/inhibitors with beneficial therapeutic effects and for the redesign of P450 enzymes to oxidize unnatural substrates (14). * To whom correspondence should be addressed. E-mail: kkkhan@ utmb.edu. Phone: (409) 772-9677. Fax: (409) 772-9642. 1 Abbreviations: P450, cytochrome P450; 6-DEB, 6-deoxyerthronolide B; 7-BQ, 7-benzyloxyquinoline; IPTG, isopropyl-β-D-thiogalactopyranoside; δ-ALA, δ-aminolevulinic acid.
Cytochrome P450eryF (CYP107A1) is a bacterial P450 that catalyzes the hydroxylation of 6-deoxyerythronolide B (6-DEB) at the 6S position, a necessary step in the synthesis of erythromycin (15, 16). P450eryF lacks the highly conserved threonine residue found in the I-helix of other P450s. The crystallographic structure of 6-DEBbound P450eryF suggests that the enzyme uses a water molecule in place of the threonine residue to deliver the proton required for conversion of iron-linked dioxygen to ferryl oxygen (17). This specialized design of the active site has made P450eryF ineffective in oxidizing even some closely related derivatives of the physiological substrate (18). P450eryF is also the only bacterial P450 to show cooperativity of substrate binding, and a recent crystallographic study of P450eryF provided the first structural proof of the simultaneous presence of two ligand molecules in a P450 active site (19). Recently, it has been shown that the substitution of Ala-245 with threonine leads to a significant gain-of-function that confers on P450eryF the ability to oxidize testosterone (20) and 7-benzyloxyquinoline (7-BQ) (21). The steadystate kinetics of oxidation of these substrates shows sigmoidal behavior. These unique properties make P450eryF an ideal model to study P450 cooperativity and structure-function relationships.
10.1021/tx025539k CCC: $22.00 © 2002 American Chemical Society Published on Web 05/24/2002
844
Chem. Res. Toxicol., Vol. 15, No. 6, 2002
Khan et al.
Figure 1. Diagram showing the residues of P450eryF selected for the present mutagenesis study. Since the crystal structure of ligand free P450eryF is not available, the diagram was generated from the androstenedione-bound P450eryF structure with the ligand molecules deleted for clarity.
In contrast to extensive structure-function studies on bacterial P450cam and P450BM-3 (22-24), the role of various P450eryF active-site residues in substrate oxidation and cooperativity has not been explored, with the exception of a single residue, Ala-245 (20, 21, 25). In the present study, we have investigated five P450eryF activesite residues (Phe-78, Gly-91, Ser-171, Ile-174, and Leu175) for their role in substrate oxidation and cooperativity (Figure 1). Phe-78, Ser-171, Ile-174, and Leu-175 are located near the distal ligand in the two-ligand-bound P450eryF crystal structures and might be expected to play a significant role in cooperative binding. In contrast, Gly-91 is close to the proximal ligand and the heme and could be crucial for substrate binding and/or oxidation. Initial characterization of seven mutants by spectral binding showed significant changes in their binding affinities. To better understand the effect of these mutations on testosterone and 7-BQ metabolism, each mutant was generated in the A245T gain-of-function background. The alteration in the active-site topology was further examined by determination of N-phenylprotoporphyrin IX regioisomer ratios (26).
Experimental Procedures Materials. Testosterone, androstenedione, 9-aminophenanthrene, imidazole, triethylamine, potassium ferricyanide, and horse skeleton muscle myoglobin were purchased from Sigma Chemical Co. (St. Louis, MO). 7-BQ and 7-hydroxyquinoline were purchased from Gentest Corp. (Woburn, MA). Ketoconazole was obtained from Spectrum Quality Products, Inc. (Gardena, CA). Phenyldiazene carboxylate azo ester was purchased from Research Organics, Inc. (Cleveland, OH). Oligonucleotide primers for PCR were either obtained from The University of Texas Medical Branch Molecular Biology Core Laboratory (Galveston, TX) or from Sigma Genosys (Woodlands, TX). PfuTurbo DNA polymerase and Rapid Ligation kits were obtained from Stratagene (La Jolla, CA) and Roche (Indianapolis, IN), respectively. Restriction enzymes were purchased from GIBCO-BRL (Grand Island, NY). The GeneClean kit was from BIO 101(Carlsbad, CA). Ni2+-NTA affinity resin was purchased from Qiagen (Valencia, CA). All other chemicals were of the highest grade available and were obtained from standard commercial sources. Construction of Single P450eryF Mutants (F78W, G91A, G91F, S171A, S171F, I174F, and L175F). All single mutants were generated by overlap extension PCR using forward and
Figure 2. Primers used for generation of P450eryF mutants. The mutated nucleotides are underlined. reverse primers containing the desired residue changes (Figure 2), His-tagged P450eryF as a template (21), and PfuTurbo DNA polymerase. Reaction conditions were as follows: one cycle of 95 °C for 30 s, 16 cycles of 95 °C for 30 s, 58 °C for 1 min, and 68 °C for 10 min. The PCR products generated contained fulllength constructs with mutations at the desired residues. The PCR products were treated with DpnI to digest the methylated template DNA and transformed directly into DH5R competent cells. DNA was isolated and analyzed for the desired alterations. All constructs were sequenced to ensure that only the desired mutation was present (Protein Chemistry Core Laboratory, The University of Texas Medical Branch, Galveston, TX). Generation of P450eryF Double Mutants (A245T/F78W, A245T/G91A, A245T/G91F, A245T/S171A, A245T/S171F, A245T/I174F, and A245T/L175F). The P450eryF cDNA contains two BlpI sites, one before the codon for Phe-78 and another between the codons for Leu-175 and Ala-245. Therefore, each double mutant was made by subcloning this small segment of the single mutants into the previously constructed A245T (21). The orientation of the inserted fragment and the presence of mutations were verified by sequencing. Expression and Purification. The details of wild-type and A245T expression and purification are described previously (21). For the expression and purification of the other P450eryF mutants, plasmids containing the mutant P450eryF cDNA were transformed into DH5R host cells. The transformed colonies were grown in 3 mL of LB media containing 50 mg/L ampicillin at 37 °C overnight and subsequently used for inoculation of 250 mL of TB media. The culture was grown at 37 °C with shaking at 240 rpm until OD (600 nm) reached 0.8-1.2, and 1 mM isopropyl-β-D-thiogalactopyranoside (IPTG) and 80 mg/mL δ-aminolevulinic acid (δ-ALA) were added. The culture was allowed to grow for an additional 48-72 h at 30 °C with shaking at 190 rpm. The cells were harvested by centrifugation (3840g, 10 min) and resuspended in 20 mM Tris-HCl, pH 8.0, and 500 mM KCl (buffer A). The resuspended cells were lysed by sonication (5 × 5 pulses for 5 s each, 60% capacity, flat tip, output 50%). The lysed cells were centrifuged (113000g, 35 min), and P450 in the supernatant was determined using reduced-CO difference spectra. For purification, a nickel affinity column (Qiagen, Valencia, CA) was packed (approximately 100 nmol of P450/mL of resin) and equilibrated using buffer A. The loaded column was washed with buffer A containing increasing concentration of imidazole (0-20 mM) and the protein eluted with buffer A containing 200
Probing the P450eryF Active Site by Site-Directed Mutagenesis mM imidazole. The eluted protein was dialyzed against two changes of 2 L of 20 mM Tris-HCl, pH 8.0, 1 mM EDTA, and 1 mM DTT. Spectral Binding Studies. Binding spectra were recorded on a Shimadzu-2600 spectrophotometer fitted with a temperature controller (TCC-240A). For titration of ligands other than 7-BQ, the spectra were recorded in the difference binding mode. A 2.0 mL solution containing 2 µM protein in 50 mM Tris-HCl, pH 7.5, and 10 mM EDTA was divided into two quartz cuvettes (10 mm path length), and a baseline was recorded between 350 and 500 nm. An aliquot of substrate in a suitable solvent (water or methanol) was added to the sample cuvette and the same amount of the solvent added to the reference cuvette. The maximal methanol concentration used was 2%. The difference spectra were obtained after the system reached equilibrium (3 min). All spectra were recorded at 25 °C. In the case of 7-BQ, titration in the difference mode did not allow an accurate estimation of binding affinity due to interference from substrate absorbance at lower wavelengths. Therefore, all 7-BQ titrations were carried out in the absolute mode. In this case the sample chamber contained 1 µM protein in 50 mM Tris-HCl, pH 7.5, and 200 mM KCl, and the reference chamber contained the buffer. An aliquot of substrate (in methanol, 2% final) was then added to the sample cuvette, the same amount of the solvent added to the reference cuvette, and spectra were recorded. All the spectra were analyzed using principle component analysis (27, 28) with the help of standard spectra of P450eryF low-spin, high-spin, and 7-BQ. 7-BQ Oxidation Assay. The details of the 7-BQ oxidation assay have been recently elaborated (21). Briefly, a fixed concentration of 7-BQ (in methanol, 2% final) was added to 4 µM enzyme in a buffer containing 50 mM Tris-HCl, pH 7.5, and 200 mM KCl. The mixture was incubated for 5 min at room temperature before starting the reaction by adding 10 mM H2O2 (final). The reaction was stopped after 5 min by adding 50 µL of 6.8 unit/µL catalase. Two mL of a buffer containing 50 mM Tris-HCl, pH 7.5, and 200 mM KCl was subsequently added to this reaction mixture, which was briefly vortexed, before determining the total amount of the product formation by fluorescence. The excitation and emission wavelengths were 410 and 510 nm, respectively, and the bandwidth was 20 nm in each case. The final activity of the various proteins was calculated by comparison to a standard curve. Testosterone Oxidation Assay. Testosterone hydroxylation assay conditions were essentially similar to those described in the literature (20). A fixed concentration of testosterone (in methanol, 2% final) was added to 4 µM enzyme in a buffer containing 50 mM Tris-HCl, pH 7.5, and 200 mM KCl. The mixture was incubated for 5 min at room temperature before starting the reaction by adding 10 mM H2O2 (final). The final volume of the reaction mixture was 100 µL. The reaction was stopped after 5 min by adding 50 µL of 6.8 unit/µL catalase, and 1.5 mL of CH2Cl2 was added. The mixture was vortexed and centrifuged at 3000 rpm for 5 min. The lower organic layer was removed and evaporated to dryness under nitrogen. The extracted metabolites were resuspended in 200 µL of the mobile phase (65% methanol containing 0.1% triethylamine, pH 7.0). Fifty-µL of the mixture was loaded for HPLC analysis. The HPLC system used for testosterone metabolite separation was the same as described previously (29, 30). The separation was achieved using an ultrasphere ODS column (5 µm × 250 mm × 4.6 mm, Beckman, Fullerton, CA) with an ultrasphere C18 guard column (5 µm × 7.5 mm × 4.6 mm, Alltech, Deerfield, IL). The separation of the metabolites was achieved isocratically using the mobile phase (65% methanol containing 0.1% triethylamine, pH 7.0). The flow rate was 1.0 mL/min, and the UV detector was set at 254 nm. All chromatographic separations were performed at room temperature. Formation and Determination of N-Phenylprotoporphyrin IX Isomers. Formation and determination of Nphenylprotoporphyrin IX isomers ratio were carried out as described earlier (31, 32). Briefly, 2.5 nmol of P450 in 500 µL of
Chem. Res. Toxicol., Vol. 15, No. 6, 2002 845 100 mM potassium phosphate buffer (pH 7.4) was treated with 10 µL of a stock solution of methyl phenyldiazene carboxylate azo ester, prepared by dissolving 2 µL of azo ester with 200 µL of 1 N potassium hydroxide. After standing at room temperature for 1 h, the solution was treated with 4 × 2 µL of potassium ferricyanide solution (62.5 mM in 100 mM potassium phosphate buffer, pH 7.4) at 5 min intervals. The solution was allowed to stand another 10 min after the final addition. At this stage, 5 mL of freshly prepared 5% sulfuric acid in acetonitrile (v/v) was added to the reaction mixture, and the mixture was left overnight at 4 °C. The mixture was concentrated under reduced pressure to a volume of 1-2 mL, 2 mL of aqueous sulfuric acid (5% v/v) was added, and the resulting solution was extracted three times with 1 mL of CH2Cl2. The extract was washed once with 1 mL of water and dried in vacuo. The dried sample was resuspended using 100 µL of solvent A for HPLC analysis. The separation of the isomers was achieved using a Partisil ODS-3 column (5 µm × 250 mm × 4.6 mm, Alltech, Deerfield, IL) with a Partisil ODS-3 guard column. Separation was achieved using the following gradient: 10% solvent B [10:1 CH3OH/CH3COOH (v/v)] in solvent A [6:4:1 CH3OH/H2O/CH3COOH (v/v/v)] for 10 min, a 1-min gradient to 30% B, 30% B for 30 min, and then a 1-min gradient to 100% B, which was maintained for 10 min. The flow rate was 1.0 mL/min and the UV detector was set at 416 nm. All chromatographic separations were performed at room temperature.
Results Expression and Purification of Single and Double P450eryF Mutants. The original vectors containing cDNA of the wild-type and A245T (16, 25) were modified to put a 4-His tag at the C-terminus for easier purification of the expressed proteins (21). As detailed in Materials and Methods, the same template was also used for generation of various mutants. The expression of mutant proteins were maximized by adding 80 mg/L of δ-ALA and 1 mM of IPTG during culture growth as in the case of wild-type P450eryF, and purification of the protein was achieved using a single metal-affinity column. The final yield of various purified single and double mutants varied between 410-1125 and 430-1325 nmol/L, respectively, with an A418/A280 ratio of 1.2-1.8. The yield for the wildtype enzyme varied between 800 and 1200 nmol/L, with an A418/A280 ratio of 1.5-1.7 (21). Spectral Binding Studies of Single P450eryF Mutants. The single mutants generated were initially tested for their binding affinity with testosterone and 9-aminophenanthrene. Both these compounds are known to show cooperative binding behavior with wild-type P450eryF (19-21). While the steroid produces a type-I spectrum, 9-aminophenanthrene induces type-II binding. Although only the crystal structure of an androstenedione- and not testosterone-bound P450eryF is available, we preferred to use testosterone because the assay conditions for testosterone metabolism are well established (20) and the n value for the cooperative binding of wild-type P450eryF with testosterone is much higher than with androstenedione (see Table 1). For simplicity, we assumed that the binding of testosterone with P450eryF would be similar to androstenedione, and that active-site residues involved in protein-substrate interaction would be the same in both cases. The replacement of Phe-78 with tryptophan caused no significant change in testosterone binding (Table 1). This is expected based on structural analogy with androstenedione-bound P450eryF, as this residue is not within the van der Waals interaction distance of either the proximal or distal ligand (19). However, the substitution of this
846
Chem. Res. Toxicol., Vol. 15, No. 6, 2002
Khan et al.
Table 1. S50 or KD Values, Hill Coefficients (n Values), and Maximal Absorbance Changes from Spectral Titrations of Wild-Type P450eryF and Mutants with Various Ligands testosteronea ∆Amax (µM-1 cm-1) WT
0.030c (0.003) 0.037 (0.006)
n
G91A
1.64 (0.13) 1.59 (0.12) d
G91F
d
F78W
S171A S171F I174F L175F
0.037 (0.010) 0.054 (0.025) 0.024 (0.001) 0.042 (0.003)
1.40 (0.15) 1.47 (0.16) 2.01 (0.10) 1.48 (0.07)
androstenedionea
9-aminophenanthrenea
S50 (µM)
∆Amax (µM-1 cm-1)
n
S50 (µM)
∆Amax (µM-1 cm-1)
278 (35) 309 (45)
0.070c (0.005)
1.20 (0.07)
261 (32)
0.011 (0.002)
0.98 (0.08) d
427 (121)
0.064 (0.010) 0.058 (0.004) 0.063 (0.003) 0.058 (0.004) 0.042 (0.001) 0.089 (0.002) 0.042 (0.002)
438 (147) 788 (389) 139 (9.2) 259 (23)
ketoconazoleb
S50 (µM)
∆Amax (µM-1 cm-1)
KD (µM)
1.84 (0.07) 1.24 (0.06) 1.82 (0.10) d
10.3 (0.2) 55.7 (6.8) 33.4 (2.0)
0.071c (0.008)
5.1 (2.0)
0.044 (0.001) 0.044 (0.001)
16.2 (1.31) 1.67 (0.12)
1.53 (0.11) 1.76 (0.08) 1.84 (0.01) 1.39 (0.08)
29.7 (3.1) 35.6 (1.3) 13.3 (0.4) 27 (2)
n
a The absorbance changes were fit to the Hill equation. Values in parentheses show the average deviation from the fit. b The absorbance changes were fit to the tight binding equation (∆A ) ∆Amax/2E(E + S + KD - [{E + S + KD}2 - 4ES]1/2)), where E and S represent the concentrations of P450eryF and substrate, respectively, ∆A and ∆Amax are the changes in the absorption at the substrate concentration S and a saturating substrate concentration, respectively, and KD is the equilibrium dissociation constant of the enzyme-substrate complex. Values in parentheses show the average deviation obtained from the fit. c Reference 21. d Atypical spectra were obtained in these cases (see Figure 3 for details).
residue significantly decreased the n value and increased the S50 value for 9-aminophenanthrene binding to the enzyme. The substitution of Gly-91 with alanine or phenylalanine had a dramatic effect on binding affinity (Figure 3 and Table 1). The addition of increasing concentrations of testosterone failed to give the expected type-I binding spectra, as evident from Figure 3, panels A and B. The titration was repeated using androstenedione to investigate whether the mutants show similar behavior upon titration with other type-I substrates. Increasing concentrations of androstenedione caused a very small spectral change in case of G91A (
