Mutagenesis Study of Asp-290 in Cytochrome P450 2B11 Using a Fusion Protein with Rat NADPH-Cytochrome P450 Reductase

ARCHIVES OF BIOCHEMISTRY AND BIOPHYSICS

Vol. 326, No. 1, February 1, pp. 85–92, 1996 Article No. 0050

Mutagenesis Study of Asp-290 in Cytochrome P450 2B11 Using a Fusion Protein with Rat NADPH-Cytochrome P450 Reductase Greg R. Harlow1 and James R. Halpert Department of Pharmacology and Toxicology, College of Pharmacy, University of Arizona, Tucson, Arizona 85721

Received July 31, 1995, and in revised form November 7, 1995

Asp-290 of the phenobarbital-inducible dog liver cytochrome P450 (P450) 2B11 was mutated to nine other amino acid residues by site-directed mutagenesis, and the functional significance of the unique negative charge in P450 2B11 at that position was studied. To facilitate the analysis of mutated P450 2B11 enzymes heterologously expressed in Escherichia coli, an enzymatically active fusion enzyme was genetically engineered between the cDNAs for P450 2B11 and rat liver NADPH-cytochrome P450 reductase using a Ser-Thr linker as previously described (Fisher et al., 1992, Proc. Natl. Acad. Sci. USA 89, 10817–10821). Sonicated whole-cell lysates of E. coli cells expressing the wildtype fusion protein were able to catalyze the 16-hydroxylation of androstenedione (AD) in the absence of added reductase, and exhibited activities and androstenedione metabolite profiles very similar to those of purified and reconstituted enzyme preparations. The substitution of Ala, Glu, Gly, Met, Asn, Arg, Ser, Thr, or Val for Asp-290 of P450 2B11 resulted in decreased AD hydroxylase activities as assessed using solubilized membranes. Replacement of Asp-290 with Glu yielded the highest activity (55% of wild type), while substituting the positively charged amino acid Arg created an enzyme with the lowest activity (õ1% of wild-type activity). Regioselectivity of AD hydroxylation was not affected although the stereoselectivity of hydroxylation at the 16 carbon position was altered in some cases. The use of the fused enzyme to study the effects of site-directed mutagenesis has resulted in the demonstration of the importance of size and charge at position 290 for enzymatic activity of P450 2B11. q 1996 Academic Press, Inc.

Key Words: P450 2B11, fusion, properties, site-directed mutants, expressed in E. coli; cytochrome P450, fusion, properties, site-directed mutants, expressed in E. coli; NADPH-cytochrome P450 reductase, fusion; androstenedione, hydroxylation.

Cytochromes P4502 comprise a superfamily of hemoproteins that are capable of metabolizing a wide variety of endogenous and exogenous compounds. Despite the wide range of substrates recognized by P450 enzymes as a group, various isoforms show very strict stereoand regiospecificity toward some substrates, particularly steroids (1). Studies of various members of the P450 2 family (2–12) have so far confirmed that the residues that govern substrate specificity are found within or near one or more of six predicted substrate recognition sites (SRSs) (13). A major focus of this laboratory has been the identification of structural determinants of substrate specificity in the P450 2B subfamily. Studies of rat P450 2B1, rabbit P450s 2B4 and 2B5, and dog P450 2B11 have led to the identification of a number of amino acid residues within these enzymes that affect stereo- and regiospecificity of hydroxylation (14–19). These studies were done using allelic variants, enzymes from related species, chimeric enzymes, or site-directed mutants. Site-directed mutagenesis coupled with heterologous expression has created the ability to generate and analyze many mutated forms of individual enzymes. The use of a more randomized P450 mutagenesis method and development of a rapid functional screen to identify mutant proteins with altered function would facilitate the molecular genetic analysis of P450 substrate specificity. Recently, expression methods have been developed (20) that allow high levels of functional P450 2B11 pro1 To whom correspondence and reprint requests should be addressed. Fax: (520) 626-2466; E-mail: [email protected]. 2 Abbreviations used: P450, cytochrome P450; AD, androstenedione; CHAPS, 3-[(3-cholamidopropyl)dimethylammonio]-1-propanesulfonate; -OH, hydroxy; DLPC, dilauryl-L-3-phosphatidylcholine; PCR, polymerase chain reaction; TLC, thin-layer chromatography; reductase, NADPH-cytochrome P450 reductase; IPTG, isopropyl-bD-thiogalactopyranoside; MOPS, 4-morpholinepropanesulfonic acid; SRS, substrate recognition site.

85

0003-9861/96 $12.00 Copyright q 1996 by Academic Press, Inc. All rights of reproduction in any form reserved.

/ m4949$9273

12-20-95 01:12:15

arca

AP: Archives

86

HARLOW AND HALPERT

tein to be synthesized in Escherichia coli. P450 enzymes receive electrons from a flavoprotein, NADPHcytochrome P450 reductase (reductase). In vitro, reductase is normally supplied exogenously to reconstitute an active complex with an isolated P450 enzyme. Fusion proteins between P450 and reductase have previously been engineered (21–25) in an attempt to emulate the naturally occurring fusion protein P450BM-3 , which has the highest turnover number of any known P450 (ú4000 nmol myristic acid hydroxylated/min/ nmol P450). Although turnover numbers obtained with these artificial P450–reductase fusions have not reached the level of P450BM-3 , the enzymes are capable of carrying out substrate metabolism in the absence of exogenously added reductase (21–25). Three residues in P450 2B11, Val-114, Asp-290, and Leu-363, were recently identified as determinants of substrate specificity (18). The replacement of Asp-290 in P450 2B11 with Ile, which is the residue present at that position in rat P450 2B1, caused the altered metabolism of androstenedione (AD), testosterone, 7ethoxycoumarin, (R)- and (S)-warfarin, and 2,2*,4,4*, 5,5*-hexachlorobiphenyl. In the present study, heterologous expression of site-specifically mutated forms of a fusion protein between dog P450 2B11 and rat reductase was used to study in more detail the functional significance of the charged Asp-290 of P450 2B11. The negative charge on Asp-290 is especially interesting because none of the substrates used in the previous study contain a positively charged group, nor do other P450 2B enzymes have a charged residue in this position. Here we describe the generation and characterization of nine substitutions of P450 2B11 residue 290. A method is presented that uses sonicated E. coli extracts of P450 2B11 fusion-expressing cells to rapidly characterize the effects of the Asp-290 substitutions on enzyme function. EXPERIMENTAL PROCEDURES Materials. Restriction endonucleases were purchased from GIBCO-BRL (Grand Island, NY) and Stratagene Cloning Systems (La Jolla, CA). Taq DNA polymerase was obtained from Boehringer Mannheim (Indianapolis, IN). Growth media for E. coli were from Difco (Detroit, MI). Androstenedione, NADPH, d-aminolevulinic acid, and IPTG were purchased from Sigma Chemical Company (St. Louis, MO). [4-14C]Androstenedione was from DuPont–New England Nuclear (Boston, MA). TLC plates were obtained from J.T. Baker, Inc. (Phillipsburg, NJ). Oligonucleotide primers were synthesized by the University of Arizona Macromolecular Structure Facility (Tucson, AZ). Construction of 2B11–reductase fusion. The E. coli strain DH5a was used for all transformations and plasmid preparations. The plasmid pCWori/::bov17A-rORfus which encodes a fusion enzyme between an N-terminally modified P450 17A and an N-terminally modified rat NADPH-P450 reductase was kindly provided by Dr. R. W. Estabrook (University of Texas Southwestern Medical Center, Dallas, TX) (22). The dog P450 2B11–rat NADPH-P450 reductase fusion protein was originally constructed in the T7 expression vector pET29a(/) (Novagen, Madison, WI) in five steps (see Fig. 1) as fol-

/ m4949$9273

12-20-95 01:12:15

arca

lows. Step 1 involved the introduction of a ‘‘silent’’ EcoRI restriction site into the 2B11 coding sequence by changing the codon for amino acid 282 from GAG to GAA. This was accomplished by using the oligonucleotide primer 5*-GGAATTCGCTGCTGGG-3*, corresponding to amino acids 284–279, as the reverse primer in a PCR amplification reaction with the N-terminal modification primer (20) as the forward primer. The DNA template for the PCR reaction was the modified 2B11 expression plasmid pKKm66c (20). The resulting amplification product was cut with the restriction enzymes NcoI and EcoRI and cloned into the corresponding sites in pET29a(/) to form pET29a(/)2B11(KK-RIS). pET29a(/)2B11(KK-RIS) has the N-terminal portion of 2B11, corresponding to amino acids 1–283, inserted in frame with and downstream of the S-Tag and thrombin cleavage coding sequences found in pET29a(/). In step 2, a SalI–HindIII fragment carrying the cDNA encoding the modified rat NADPHcytochrome P450 reductase was moved from pCWori/::bov17A-rORfus to the corresponding sites in pET29a(/)2B11(KK-RIS) to form pET29a(/)2B11(KK-RIS)(OR). For step 3, the remainder of the 2B11 cDNA was then inserted into pET29a(/)2B11(KK-RIS)(OR) by first PCR-modifying the C-terminus of 2B11 with the oligonucleotide 5*GGAATTCAAGTCGACCCGCACCCTCC-3*. In the reverse orientation, this oligonucleotide hybridizes to the sequence encoding amino acids 492–494 and changes the 2B11 stop codon (TGA) to Gly (GGG), adds codons for Ser (TCG) and Thr (ACT) followed by a new stop codon (TGA), and introduces sites for the restriction endonucleases SalI and EcoRI (underlined). The forward primer used in the reaction was 5*-TTCACGGTGTACCTGGGGCCA-3*, which hybridizes to pKKm66c at a position corresponding to P450 2B11 amino acids 66– 72. The resulting PCR fragment was cut with PstI and EcoRI and inserted into pET29a(/)2B11(KK-RIS)(OR) at the corresponding sites to form pET29a(/)2B11(KK-RIS)(STOR). Step 4 was performed to eliminate the 22-bp fragment containing the engineered stop codon and an EcoRI site; this was accomplished by cutting pET29a(/)2B11(KK-RIS)(STOR) with SalI, followed by agarose gel purification and religation to give pET-29a(/)2B11fusion. Although high levels of immunologically reactive 2B11–reductase fusion protein could be expressed in E. coli using the T7 promoter of pET29a(/)2B11fusion, cell lysates gave no detectable reduced CO/reduced difference spectrum. Therefore, in step 5, a SacII–HindIII fragment from pET29a(/)2B11fusion was used to replace the SacII–HindIII fragment of pKKm66c creating a 2B11–reductase fusion construct, pKK2B11fusion, that gave spectrally detectable expression levels of 150–200 nmol/liter culture. The portion of pKK2B11fusion that was made by PCR (from SacII to SalI) was confirmed by DNA sequencing. Mutagenesis of 2B11 residue 290. Mutations were introduced into pKK2B11fusion at amino acid position 290 by using the mixed-sequence oligonucleotide 5*-CGAATTCCACCATCGGAACCTCATA{G or A}{G, A, T, or C}{G or C}ACGG-3*, corresponding to amino acids 281–292. This oligonucleotide, which can create 16 different codon combinations representing 12 different amino acids, was used as the forward primer in a PCR reaction with the 2B11 C-terminal modification primer (described above). The amplified product was cut with EcoRI and SalI and inserted into the corresponding sites on pCW(KK-RIS)(OR) which was created by replacing the NdeI– HindIII fragment of pCWori/::bov17A-rORfus with the NdeI–HindIII fragment of pET29a(/)2B11(KK-RIS)(OR). Resulting clones were sequenced to identify the altered codon at amino acid 290. Portions of plasmids with a desired change were subcloned as SacII– HindIII fragments from pCW(KK-RIS)(OR) to pKK2B11fusion. Two amino acid substitutions of interest, Ser and Asn, were not identified by this method. Ser and Asn were subsequently made by using the mixed-sequence primer 5*-GCGAATTCCACCATCGGAACCTCATAA{G or A}CACGGCGC-3* using the PCR amplification and cloning strategy described above. All D290 mutants were confirmed by sequencing the entire PCR-amplified region. The expressed fusion proteins used in this study represent the following codons for amino

AP: Archives

STRUCTURE AND FUNCTION OF CYTOCHROME P450 2B11

87

FIG. 1. Scheme depicting construction of P450 2B11 plasmids used in this work.

acid 290: Ala(GCC), Asp(GAC), Glu(GAG), Gly(GGC), Met(ATG), Asn(AAC), Arg(AGG), Ser(AGC), Thr(ACC), Val(GTC). Enzymatic assays. Protein expression was carried out in the E. coli strain TOPP3 purchased from Stratagene Cloning Systems. For whole-cell lysate androstenedione hydroxylation assays, 5-ml cultures of TB / ampicillin (12 g tryptone, 24 g yeast extract, 100 mg ampicillin, 4 ml glycerol/liter) were started in 15 1 150-mm glass culture tubes from single colonies growing on recently streaked LB plates. Cultures were grown for 5 h at 377C on a roller mechanism (40 rpm), then induced with 1 mM IPTG and placed at a 457 angle in racks in a 307C shaking incubator set at 190 rpm. After 24 h at 307C, d-aminolevulinic acid was added to a final concentration of 80 mg/ml and the cultures were grown an additional 48 h. Cells were then pelleted by centrifuging at 5000 rpm for 10 min in a Beckman

/ m4949$9273

12-20-95 01:12:15

arca

TJ-6 centrifuge at room temperature. The pellet was resuspended in 0.5 ml MOPS buffer (100 mM MOPS, pH 7.3, 10% glycerol, 0.2 mM dithiothreitol, 1 mM EDTA), placed on ice, and sonicated briefly by giving five 2-s pulses using the microtip attachment on a Heat Systems–Ultrasonic, Inc., W-380 sonicator (No. 5 output, 60% duty cycle). Androstenedione hydroxylase activities were determined by mixing 50 ml of the sonicated cell lysate with 40 ml of 62.5 mM AD (22.5 mCi/mmol) in 125 mM Hepes (pH 7.6), 37.5 mM MgCl2 , and 250 mM EDTA. Reactions were preincubated at 377C for 5 min before adding 10 ml 50 mM NADPH and incubating for an additional 30 min. Reactions were stopped by adding 50 ml tetrahydrofuran. Radioactive metabolites were separated by TLC, scraped from TLC plates, and assayed by liquid scintillation counting as described previously (15). Solubilized E. coli membranes were prepared as previously de-

AP: Archives

88

HARLOW AND HALPERT

FIG. 2. Autoradiogram of androstenedione metabolites produced by sonicated cell lysates of E. coli-expressed P450 2B11 fusion and nonfused P450 2B11. Sonicated whole-cell lysates containing P450 (10 pmol) were reconstituted in a 20-ml volume with or without 20 pmol reductase and 0.72 mg DLPC. Reactions were incubated with 25 mM [14C]androstenedione for 10 min at 377C in 100 ml of 11 Hepes buffer (50 mM Hepes, pH 7.6, 15 mM MgCl2 , 0.1 mM EDTA, 3 mg/ml DLPC). NADPH was added, as indicated, to a final concentration of 1 mM. Samples were quenched by the addition of 50 ml of tetrahydrofuran, and 50-ml aliquots were applied to a TLC plate. F, P450 2B11– reductase fusion enzyme; NF, P450 2B11 nonfused enzyme. The dark band at the top of the TLC plate is the parent compound, [14C]androstenedione. Metabolites are 16a-OH- and 16b-OH-androstenedione.

scribed (20). The fusion protein was affinity purified as described by Fisher et al. (25) with the following modifications. Buffer A (50 mM potassium phosphate, pH 7.8/20% glycerol/0.5 mM EDTA/0.1 mM dithiothreitol/0.5% CHAPS) with CHAPS replacing 0.2% Emulgen was used to equilibrate a Whatman DE-52 column and to dilute solubilized membranes prior to sample application. Elution from the DE52 column and affinity purification on (2*,5*-ADP)–Sepharose 4B were as described (25). Following elution from (2*,5*-ADP)–Sepharose 4B, pooled fractions were concentrated in a Centricon 30 microconcentrator (Amicon, Beverly, MA) and were not processed further.

RESULTS

Enzymatic activity of 2B11 fusion protein and nonfused 2B11 in sonicated bacterial cell lysates. To study the effects of fusing the N-terminal-modified coding sequence of rat reductase (25) to dog P450 2B11, two different modified forms of the 2B11 expression plasmid pKKm66c (25) were made and expressed in E. coli. The nonfused construct, pKK2B11(STOR), contains a P450 2B11 coding sequence modified at the Cterminus to include three extra codons, Gly-Ser-Thr; a SalI site; and a translation termination codon preceding the N-terminal-modified rat reductase coding sequence. pKK2B11–fusion has the three added C-terminal amino acids but has the termination codon removed by excision of a short SalI fragment. This results in an in-frame gene fusion of the 2B11 P450 coding sequence to reductase capable of encoding a single 127-kDa protein (confirmed by SDS–PAGE immunoblotting, data not shown). E. coli cells expressing either construct were sonicated and assayed for the ability to hydroxylate the steroid substrate AD (Fig. 2).

/ m4949$9273

12-20-95 01:12:15

arca

When cells expressing either the fused or the nonfused enzyme were sonicated and incubated with or without added reductase, little or no AD metabolism could be detected (see Fig. 2, additions: none and /FP) in the absence of added NADPH. When NADPH was added to the reaction mix (additions: /NADPH) the activity of the fusion protein was greatly increased, while the nonfusion enzyme still did not produce detectable levels of metabolites. Only when both reductase and NADPH were added to the reaction mix (additions: /FP /NADPH) was the nonfusion enzyme able to metabolize AD. The addition of reductase and NADPH to the fusion protein cell lysate did not result in an increase in metabolism of AD over that obtained by adding NADPH alone, indicating that the reductase domain and the P450 domain may be tightly associated in the E. coli membrane. Reductase can associate with nonfused 2B11 protein in sonicated E. coli cell lysates, suggesting that the inability of added reductase to increase the activity of 2B11 fusion is not due to a sequestering of the enzyme within the membrane. Furthermore, the addition of three amino acids, Gly, Ser, and Thr, to the C-terminus of P450 2B11, as judged by the AD metabolite profile of the fully reconstituted nonfused enzyme, had no effect on stereo- or regiospecificity of AD hydroxylation by the P450 2B11 enzyme (18, 26, 27). Titration of 2B11 fusion with added reductase. As a measure of fusion protein function, increasing amounts of purified rat reductase were exogenously added to preparations of P450 2B11 fusion protein. The results of this set of titrations are presented in Fig. 3. The AD 16b-hydroxylase activity of various preparations of 2B11 fusion, representing different levels of purification, was measured in reactions supplemented with an up to fourfold molar excess of added rat reduc-

FIG. 3. Effects of added NADPH-P450 reductase on 2B11 fusion AD metabolism. P450 2B11 (10 pmol) was reconstituted in 50 ml with the indicated amount of rat NADPH-P450 reductase (FP) and 1.5 mg DLPC. Reactions were performed at 377C for 5 min in 100 ml using 25 mM [14C]androstenedione in 11 Hepes buffer (50 mM Hepes, pH 7.6, 15 mM MgCl2 , 0.1 mM EDTA, 3 mg/ml DLPC). Values are the average of duplicate determinations.

AP: Archives

STRUCTURE AND FUNCTION OF CYTOCHROME P450 2B11

tase. The addition of reductase did not increase enzyme activity for sonicated whole-cell preparations; in fact, it slightly decreased enzyme activity. However, added reductase did increase activity for solubilized membrane and affinity-purified preparations of the fusion protein. These results suggest that the process of solubilizing and purifying the fused enzyme results in the disruption of functional contacts between P450 and reductase domains within at least some of the fusion molecules so that exogenously added reductase can associate with the P450 domain. Effect of fusion protein concentration on enzyme activity. The ability of added reductase to increase enzyme activity in solubilized membrane and affinity-purified preparations of P450 2B11 fusion suggested that the P450 2B11 domains were not absolutely coupled with the fused reductase. If the reductase were associated with the fused P450 domain with 100% efficiency, then the substrate turnover rate would be a constant, independent of enzyme concentration. If the reductase domain behaved like free reductase in solution, essentially 100% uncoupled to its fused P450 partner, then the rate of reaction would increase by the square of the increase in enzyme concentration. The ability of affinity-purified P450 2B11 fusion protein to metabolize AD was therefore measured at three concentrations of enzyme; 50, 100, and 200 nM. The results (Fig. 4) demonstrate that the turnover number was not constant for these three concentrations. For each doubling of the fusion enzyme concentration, the reaction rate increased by a factor of approximately 4, the square of the enzyme concentration increase. Use of the fusion enzyme to analyze changes introduced at residue 290 in P450 2B11. A mixed oligonu-

FIG. 4. Effect of P450 2B11 fusion concentration on AD metabolism. 5, 10, and 20 pmol of affinity-purified P450 2B11 fusion protein was preincubated in 20 ml 11 Hepes buffer (50 mM Hepes, pH 7.6, 15 mM MgCl2 , 0.1 mM EDTA) at 257C for 10 min. Enzyme concentrations during the preincubation correspond to 0.25, 0.5, and 1.0 mM, respectively. Reactions were performed at 377C for 5 min in 100 ml 11 Hepes buffer containing 25 mM [14C]androstenedione. Enzyme concentrations in the final reaction volume were plotted versus the rate of formation of the 16b-OH AD metabolite. Values are the average of duplicate determinations.

/ m4949$9273

12-20-95 01:12:15

arca

89

FIG. 5. Autoradiogram of the TLC showing the metabolites of AD formed by sonicated E. coli whole-cell lysates expressing wild-type and mutant 2B11 fusion enzymes. Site-directed mutants of the P450 domain of the 2B11 fusion enzyme are designated using the single letter code for the amino acid replaced, the position in the sequence, and the designation of the new residue, in that order. The cell lysates were obtained and assayed for androstenedione hydroxylase activity for 30 min as described under Experimental Procedures. Steroid substrate and its metabolites were extracted from the reaction with 500 ml ethyl acetate and dried by evaporation. Samples were dissolved in 75 ml ethyl acetate and one-third of the reaction was analyzed by TLC. Individual lanes are labeled with the designation for the mutated enzyme that catalyzed the reaction. 2B11 NON-FUSION refers to cells expressing a P450 2B11 enzyme lacking the fused reductase domain. SHAM refers to cells expressing the pKK233-2 plasmid with no P450 coding sequence inserted. The dark band at the top of the TLC plate is the parent compound, [14C]androstenedione. Metabolites are 16a-OH- and 16b-OH-androstenedione; the 11 ratios are shown in Table I.

cleotide was used to introduce changes by PCR into P450 2B11 at amino acid position 290. Nine mutants at this position were expressed as fusion proteins in 5ml cultures and were analyzed initially as sonicated whole-cell preparations. The changes at residue 290 were designed to test the significance of the negatively charged Asp residue by replacing it with: (1) a similarly charged Glu residue, (2) the uncharged polar residues Asn, Gly, Ser, and Thr, (3) the hydrophobic residues Ala, Met, and Val, and (4) the positively charged residue Arg. Cells were harvested at 72 h by brief centrifugation, resuspended in MOPS buffer, briefly sonicated, and used directly in an AD hydroxylation reaction adding only AD in buffer and NADPH. The AD metabolite profiles (Fig. 5) show that this form of assay can be used to give rapid information about the effects of a particular mutation. Differences in relative amounts of metabolites could be due to different levels of expression or stability in E. coli or to differences in enzyme activities. Each reaction was performed using the same volume of sonicated cells and was not standardized for P450 content. Of note is the lack of detectable metabolites for the charge change mutation, D290R, and the

AP: Archives

90

HARLOW AND HALPERT TABLE I

Metabolism of Androstenedione by Solubilized Membrane Preparations of Wild Type and Site-Directed Mutants of P450 2B11 Expressed in Escherichia coli nmol product/min/nmol P450 Amino acid at 290

16b-OH-AD

16a-OH-AD

Total

A D (WT) E G M N R S T V

0.24 1.29 1.01 0.45 0.81 0.15 ND 0.31 0.39 0.14

0.17 1.16 0.32 0.37 0.29 0.11 ND 0.24 0.47 0.04

0.4 2.5 1.4 0.9 1.1 0.3 ND 0.55 0.86 0.18

% WT activity 16 100 56 36 44 12 õ1 22 34 7

16b/16aa

Sonicated cells 16b/16ab

1.4 1.1 3.2 1.2 2.8 1.4 ND 1.3 0.8 3.5

1.2 1.1 3.2 0.8 2.8 1.4 ND 1.2 0.8 3.6

Note. 10 pmol P450 2B11 fusion from CHAPS-solubilized membrane preparations was reconstituted in 50 ml and assayed for 30 min in 100 ml 11 Hepes buffer (50 mM Hepes, pH 7.6, 15 mM MgCl2, 0.1 mM EDTA) containing 25 mM [14C]androstenedione. ND, not detectable or metabolites less than twice background. Values are the average of duplicate determinations. a Ratios of 16b-OH/16a-OH AD metabolites formed by enzyme preparations from solubilized membrane preparations. b Ratios of 16b-OH/16a-OH AD metabolites formed by enzyme preparations from sonicated E. coli cells.

relatively abundant production of metabolites for WT and D290E, a conservative amino acid substitution. None of the mutations alter the regiospecificity of the enzyme. Like the wild-type enzyme, each mutant enzyme hydroxylated AD only at the 16 carbon position, although the stereospecificity as given by the ratios of the 16b-OH:16a-OH metabolites was affected by some mutations (Table I, last column). Enzymatic activities of solubilized membrane preparations. Each of the mutant fusion proteins was grown in large culture and processed into a solubilized membrane preparation. AD hydroxylase assays were performed over the course of 30 min using 10 pmol of P450 per reaction in equal reaction volumes (Table I). The rate of metabolite formation was constant over the 30-min incubation for the wild-type enzyme (data not shown). None of the mutants gave activity equal to or higher than the wild-type enzyme. The conservative substitution mutant, in which Asp-290 is replaced with Glu (D290E), has only 56% activity, due mostly to a decreased ability to form the 16a-OH metabolite (reduced to 28% WT level), although it still forms the 16b-OH metabolite at near-WT levels (78% WT). The reduced ability of D290E to form the 16a-OH metabolite results in a 16b-OH:16a-OH ratio of 3.2 versus 1.1 for the WT enzyme. The polar uncharged amino acid substitutions, D290G, D290N, D290S, and D290T, yield lower overall activities, 36, 12, 22, and 34% of WT, respectively, but have 16b-OH:16a-OH ratios similar to WT. Though lacking the negative charge, these residues have sidechains that are approximately the same size or smaller than Asp. Introduction of the hydrophobic residues Ala and Val at position 290 further

/ m4949$9273

12-20-95 01:12:15

arca

reduced the overall activity of the enzyme, 16 and 7% of WT, respectively, and caused the 16b-OH:16a-OH ratio to favor the 16b-OH metabolite (16b-OH:16a-OH ratios were A, 1.4; V, 3.5). The 16b-OH:16a-OH AD ratio for D290I was previously determined to be Ç9.0 (18). Interestingly, insertion of the large hydrophobic amino acid Met at 290, D290M, did not negatively affect overall activity (43% of WT) as severely as D290A or D290V. Replacement of the negative charge with a positively charged residue, as in the D290R mutant, reduced the activity to õ1% WT levels. DISCUSSION

In an ongoing study of structure–function relationships in the P450 2B subfamily, we have created a P450 2B11 fusion protein that efficiently metabolizes androstenedione when assayed as a sonicated wholecell preparation. The fused enzyme was further used to analyze the effect of changing the negatively charged P450 2B11 residue Asp-290 to one of nine other amino acids. The presence or absence of charge and the size of the residue at position 290 were shown to be important for maintaining the overall activity and stereoselectivity of AD hydroxylation. A number of previous studies with P450–reductase fusions (21–24, 28, 29) led to our choice of fusion methodology and heterologous expression in E. coli. The fusion of the P450 domain to the modified reductase domain was accomplished by changing the P450 2B11 stop codon to Gly and using a short Ser-Thr linker similar to one described previously for the fusion of P450 17A to rat NADPH-P450 reductase (25). We have

AP: Archives

STRUCTURE AND FUNCTION OF CYTOCHROME P450 2B11

shown high activities for the 2B11 fusion protein when assayed as very crude sonicated whole-cell extracts, which should facilitate more rapid screening of P450 2B11 mutants for altered enzyme function. Currently, we create specific mutations by site-directed mutagenesis, confirm the change(s) by sequencing, then express the mutant protein and use solubilized membranes or more highly purified preparations to assess the effect of the changes. The methods presented here could be used to directly screen mutants created by cassette mutagenesis (8) or random mutagenesis for changes in substrate metabolism, making it feasible to characterize the functional significance of a large number of residues from specific regions [e.g., each of the six SRSs (13)]. The ability of these P450 2B11 fusion-expressing crude cell extracts to metabolize AD efficiently stands in contrast to the 17A–reductase fusion, in which membrane-bound enzyme preparations showed turnover numbers 5- to 10-fold lower than the purified fusion protein for the steroid substrates progesterone and pregnenolone (23). The reason for the inhibition of P450 17A fusion metabolism by the sonicated cell lysate was not known. The difference most probably reflects inherent properties of the two P450 enzymes, although different E. coli strains were used, TOPP3 in our studies versus DH5a, as well as different steroid substrates, androstenedione versus progesterone or pregnenolone. Fusions between P450 C17 and yeast NADPH-P450 reductase expressed in yeast cells were previously used to detect hydroxylation of progesterone, which had been added directly to the culture media (24). The inability of exogenously added reductase to increase turnover in sonicated whole-cell extracts suggests that in the membrane-bound state the P450 domain of the 2B11 fusion enzyme is tightly associated with the reductase domain. Added reductase was able to provide reducing equivalents to the nonfused P450 enzyme, suggesting that P450 is accessible to the reductase in these sonicated cell preparations. Our results demonstrate an approximately 2-fold stimulation of AD 16-hydroxylation by purified P450 2B11 fusion upon addition of a 4-fold molar excess of purified rat reductase (Fig. 3). The solubilization and purification process apparently causes the P450 domain of the fusion protein to become less tightly associated with the reductase domain. One interpretation is that the reconstitution conditions were not optimal; although the presence or absence of DLPC and cytochrome b5 had no effect on enzyme activity (data not shown), other reconstitution conditions were not exhaustively tested. Another explanation is that aggregation of P450 2B11 fusion protein molecules into multimeric complexes, which was recently demonstrated for the naturally occurring fusion protein P450 BM-3 (30) and for E. coliexpressed P450s 2E1 and 2B4 (31), may be prevented by the CHAPS detergent that is present in the more

/ m4949$9273

12-20-95 01:12:15

arca

91

highly processed enzyme preparations. It should be noted that fusion proteins made with different P450 domains have been shown to react differently to exogenously added reductase. For example, the ability of a bovine 17A–reductase fusion protein to metabolize progesterone or pregnenolone was completely unaffected by added reductase (25), while the rate of lauric acid v-hydroxylation by a rat 4A1 fusion protein was increased 3-fold by adding a 2-fold molar excess of purified rat reductase (25), and the rate of testosterone 6bhydroxylation by a human 3A4 fusion protein could be stimulated 10-fold by addition of a 4-fold molar excess of reductase (32). Our results suggest that the function of Asp-290 in AD hydroxylation is to coordinate the position of amino acid residues that interact with the AD substrate to produce the 16a-OH metabolite. A previous study demonstrated that replacement of Asp-290 with Ile affected the enzymatic activity of P450 2B11. In the present study we have substituted nine additional residues for Asp-290 and analyzed their affect on the ability of P450 2B11 to metabolize AD. From this we conclude that Asp-290 is probably not in direct contact with the substrate given the lack of charge on AD and the range of amino acid replacements that can be introduced without drastically altering the 16b-OH:16a-OH ratio (Table I). For example, 2B11 D290G has nearly the same 16b-OH:16a-OH ratio as the WT enzyme but has only a single hydrogen atom for a sidechain. 2B11 D290E produces nearly WT levels of the 16b-OH metabolite but has reduced 16a-hydroxylase activity. Presumably, the extra carbon on the 2B11 D290E sidechain distorts the active site, either directly or through longer range effects as seen with mutants of rabbit P450s 2C1 and 2C2 (33), enough to hinder the 16a-hydroxylation of AD more than the 16b-hydroxylation. 2B11 D290M, in which the substituted amino acid residue is larger than Asp and hydrophobic, is surprisingly active toward hydroxylation of AD. Like D290E, D290M favors 16bover 16a-hydroxylation. It is of interest to note that mouse P450 2B9 has Met at position 290, suggesting that although large and hydrophobic, Met may make alternate contacts in P450 2B11 that compensate for its lack of negative charge. As a result of the range of mutations introduced at position 290 in P450 2B11 we propose that both the negative charge and the sidechain size of the amino acid residue at 290 are important for maintaining the stoichiometry of 16b:16ahydroxylation of AD. Increasing the size of the residue causes a shift to predominantly 16b-hydroxylation, while substitution of nonnegatively charged residues at 290 leads to a general loss of enzyme activity toward AD hydroxylation. We have evidence to suggest that 2B11 D290 forms a salt bridge with the positively charged Lys residue at position 242 (manuscript in preparation). The disruption of the charge pair may

AP: Archives

92

HARLOW AND HALPERT

lead to decreased enzyme stability. Heat-inactivation studies (data not shown) indicate that P450 2B11 stability is greatly affected by the presence of unpaired charges at either position 242 or position 290. ACKNOWLEDGMENTS We thank Dr. Ronald Estabrook of the University of Texas Southwestern Medical Center for a generous gift of the plasmid encoding the bovine 17A–rat NADPH-P450 reductase fusion construct. This work was supported in part by National Institutes of Health Grant ES04995 (J.R.H.), Center Grant ES06694 (University of Arizona), and National Institutes of Environmental Health Sciences Training Grant PS07091 (G.R.H.).

REFERENCES 1. Waxman, D. J. (1988) Biochem. Pharmacol. 37, 71–84. 2. Lindberg, R. L. P., and Negishi, M. (1989) Nature 339, 632–634. 3. Juvonen, R. O., Iwasaki, M., and Negishi, M. (1991) J. Biol. Chem. 266, 16431–16435. 4. Iwasaki, M., Darden, T. A., Pedersen, L. G., Davis, D. G., Juvonen, R. O., Sueyoshi, T., and Negishi, M. (1993) J. Biol. Chem. 266, 759–762. 5. Aoyama, T., Korzekwa, K., Nagata, K., Adesnik, M., Reiss, A., Lapenson, D. P., Gillette, J., Gelboin, H. V., Waxman, D. J., and Gonzalez, F. J. (1989) J. Biol. Chem. 264, 21327–21333. 6. Kronbach, T., and Johnson, E. F. (1991) Biochemistry 266, 6215– 6220. 7. Kronbach, T., Kemper, B., and Johnson, E. F. (1991) Biochemistry 30, 6097–6102. 8. Straub, P., Lloyd, M., Johnson, E. F., and Kemper, B. (1993) J. Biol. Chem. 268, 21997–22003. 9. Imai, Y., and Nakamura, M. (1989) Biochem. Biophys. Res. Commun. 158, 717–722. 10. Hsu, M.-H., Griffin, K. J., Wang, Y., Kemper, B., and Johnson, E. F. (1993) J. Biol. Chem. 268, 6939–6944. 11. Kaminsky, L. S., de Morais, S. M. F., Faletto, M. B., Dunbar, D. A., and Goldstein, J. A. (1992) Mol. Pharmacol. 43, 234–239. 12. Fukuda, T., Imai, Y., Komori, M., Nakamura, M., Kusunose, E., Satouchi, K., and Kusunose, M. (1993) J. Biol. Chem. 113, 7– 12. 13. Gotoh, O. (1992) J. Biol. Chem. 267, 83–90.

/ m4949$9273

12-20-95 01:12:15

arca

14. Luo, Z., He, Y., and Halpert, J. R. (1994) Arch. Biochem. Biophys. 309, 52–57. 15. Kedzie, K. M., Balfour, C. A., Escobar, G. Y., Grimm, S. W., He, Y.-A., Pepperl, D. J., Regan, J. W., Stevens, J. C., and Halpert, J. R. (1991) J. Biol. Chem. 266, 22515–22521. 16. He, Y., Luo, Z., Klekotka, P. A., Burnett, V. L., and Halpert, J. R. (1994) Biochemistry 33, 4419–4424. 17. He, Y. A., Balfour, C. A., Kedzie, K. M., and Halpert, J. R. (1992) Biochemistry 31, 9220–9226. 18. Hasler, J. A., Harlow, G. R., Szklarz, G. D., John, G. H., Kedzie, K. M., Burnett, V. L., He, Y.-A., Kaminsky, L. S., and Halpert, J. R. (1994) Mol. Pharmacol. 46, 338–345. 19. Halpert, J. R., and He, Y. A. (1993) J. Biol. Chem. 268, 4453– 4457. 20. John, G. H., Hasler, J. A., He, Y., and Halpert, J. R. (1994) Arch. Biochem. Biophys. 314, 367–375. 21. Murakami, H., Yabusaki, Y., Sakaki, T., Shibata, M., and Ohkawa, H. (1987) DNA 6, 189–197. 22. Shet, M. S., Fisher, C. W., Holmans, P. L., and Estabrook, R. W. (1993) Proc. Natl. Acad. Sci. USA 90, 11748–11752. 23. Shet, M. S., Fisher, C. W., Arlotto, M. P., Shackleton, C. H. L., Holmans, P. L., Martin-Wixtrom, C. A., Saeki, Y., and Estabrook, R. W. (1994) Arch. Biochem. Biophys. 311, 402–417. 24. Shibata, M., Sakaki, T., Yabusaki, Y., Murakami, H., and Ohkawa, H. (1990) DNA Cell Biol. 9, 27–36. 25. Fisher, C. W., Shet, M. S., Caudle, D. L., Martin-Wixtrom, C. A., and Estabrook, R. W. (1992) Proc. Natl. Acad. Sci. USA 89, 10817–10821. 26. Graves, P. E., Elhag, G. A., Ciaccio, P. J., Bourque, D. P., and Halpert, J. R. (1990) Arch. Biochem. Biophys. 281, 106–115. 27. Duignan, D. B., Sipes, I. G., Leonard, T. B., and Halpert, J. R. (1987) Arch. Biochem. Biophys. 255, 290–303. 28. Yabusaki, Y., Murakami, H., Sakaki, T., Shibata, M., and Ohkawa, H. (1988) DNA 7, 701–711. 29. Sakaki, T., Shibata, M., Yabusaki, Y., Murakami, H., and Ohkawa, H. (1990) DNA Cell Biol. 9, 603–614. 30. Black, S. D., and Martin, S. T. (1994) Biochemistry 33, 12056– 12062. 31. Pernecky, S. J., Olken, N. M., Bestervelt, L. L., and Coon, M. J. (1995) Arch. Biochem. Biophys. 318, 446–456. 32. Shet, M. S., Faulkner, K. M., Holmans, P. L., Fisher, C. W., and Estabrook, R. W. (1995) Arch. Biochem. Biophys. 318, 314–321. 33. Ramarao, M. K., Straub, P., and Kemper, B. (1995) J. Biol. Chem. 270, 1873–1880.

AP: Archives