INTRODUCTION

Cyt b₅¹ was first discovered in Cecropia silkworm larvae (1), but the functions of this membrane-bound heme protein have been most extensively studied in mammals (2, 3). A microsomal form of cyt b₅ is required for numerous biosynthetic and biotransformation reactions, which include P450-dependent reactions (3), desaturation of fatty acids (4), plasmalogen biosynthesis (5), and cholesterol biosynthesis (6, 7). A soluble form of cyt b₅ is involved in the reduction of methemoglobin in erythrocytes (8) and the biosynthesis of N-glyconeuraminic acid (9). A mitochondrial form, bound to the outer mitochondrial membrane, has been described in mammals as well (10). Cyt b₅-like sequences are also found as part of larger polypeptides such as flavocytochrome b₂, sulfite oxidase and nitrate reductase (11), probably as a result of gene fusion events. The role of cyt b₅ in microsomal P450-dependent monooxygenase reactions has been studied most extensively.

P450s are a large superfamily of heme proteins which play a crucial role in the biosynthesis of a number of endogenous compounds (steroid hormones, vitamins D₃, eicosanoids, and so forth) and in the activation or detoxification of a vast variety of xenobiotics. In many of these reactions, cyt b₅ is known to determine the fate of certain substrates by either stimulating (2, 3) or inhibiting (2, 12, 13) substrate metabolism, or even by influencing the type of reaction catalyzed (14). The stimulating effect of cyt b₅ has been thought to result from: 1) enhanced rate of the second electron transfer to P450 (12, 15, 16); 2) increased ``coupling'' of the reaction, i.e. inhibition of superoxide or hydrogen peroxide formation (2, 12, 17, 18); 3) allosteric effects (19, 20); and 4) stimulation of the first electron transfer from P450 reductase to some P450s (21). However, the exact mechanism by which cyt b₅ affects P450-dependent reactions remains unclear.

Insect P450s have been extensively studied because of their crucial role in the biosynthesis of hormones regulating insect growth, development, and reproduction (ecdysteroids and juvenile hormones) and in the biotransformation of foreign compounds of synthetic (insecticides) or natural (plant and microbial toxins) origin (22). Metabolism of insecticides by P450s is a major mechanism of insecticide resistance in insects (22, 23), and detoxification of plant toxins by P450s is thought to be an adaptation to the hazards of herbivory (24). Both CYP6A1, an insect P450, which is overproduced in insecticide-resistant strains of the house fly, and NADPH-dependent cytochrome P450 reductase, which provides electrons to P450s from NADPH, have been cloned from the house fly, Musca domestica (25, 26) and expressed in E. coli (27). We have found that epoxidation of the cyclodiene insecticide heptachlor by CYP6A1 is stimulated by rat microsomal cyt b₅ in a reconstituted system.² Furthermore, immunological evidence for the involvement of cyt b₅ in several P450-dependent monooxygenase activities in house fly microsomes has been reported (28).

Our concerted effort to clone, express, and reconstitute elements of the insect cytochrome P450 system therefore led us to clone house fly cyt b₅ cDNA. In addition to a better understanding of the role of cyt b₅ in P450-mediated reactions, it was felt that the characterization of house fly cyt b₅ would facilitate the study of the evolutionary conservation of structural and functional properties of this electron carrier. We report here that the cloned house fly cyt b₅ has been overexpressed in E. coli, purified to electrophoretic homogeneity, and extensively characterized both structurally by a number of spectroscopic techniques (absorption, NMR, EPR) and functionally (redox and electron-transfer properties). We reconstituted in vitro an insect P450 system consisting of P450 reductase, CYP6A1, cyt b₅, and phospholipid. CYP6A1-catalyzed epoxidation of the insecticide heptachlor is stimulated by cyt b₅ in this system. The mechanisms by which cyt b₅ stimulates P450 turnover are discussed.

EXPERIMENTAL PROCEDURES

Four-day-old larvae of the diazinon-resistant strain ``Rutgers'' of the house fly, M. domestica, were used as the source of poly(A)⁺ RNA. About 1 µg of poly(A)⁺ RNA was reverse transcribed using the Superscript RNase H- Reverse Transcriptase Kit (Life Technologies, Inc.) and an 18-mer oligo(dT) primer. The mRNA-cDNA duplex was used as template for PCR with AmpliTaq polymerase (Perkin Elmer) and a pair of degenerate oligonucleotide primers for the two heme-binding regions of known cyt b₅s. The forward primer (for KFLEEHPG) was 5-AA(A/G) TT(C/T) (C/T)TI GA(A/T) GA(A/T) CA(C/T) CCI GG(A/T/C) GG-3. The reverse primer (for ATENFEDVG) was 5-CCI A(C/T)(A/G) TC(C/T) TC(A/G) AA(A/G) T(C/T)I TCI GT(A/T/G) GC. Upon reamplification, DNA from a faint band of the expected size (86 base pairs) was purified using the Sephaglas BandPrep Kit (Pharmacia Biotech Inc.), cloned into the pCR II plasmid (Invitrogen) and sequenced with the Sequenase Version 2.0 DNA sequencing kit (U. S. Biochemical Corp.).

A house fly cDNA library in ZAP (25) (approximately 1 × 10⁶ plaques) was screened at high stringency with the partial house fly cyt b₅ cDNA obtained by RT-PCR. The ³²P-labeled probe was generated using Prime-It II kit (Stratagene) modified by replacing random primers with the degenerate primers used for RT-PCR (see above). Inserts were excised from purified phage, recircularized into pBluescript plasmids, and sequenced. Nucleotide sequences were analyzed with the BLAST program on the NCBI Blast Server (29).

Plasmid pCWori⁺ (kindly provided by Dr. F. W. Dahlquist, University of Oregon, Eugene, OR) was used for the expression of the house fly cyt b₅. This plasmid contains an NdeI site adjacent to the ATG codon. An NdeI site was inserted at the 5-end of the cyt b₅ coding region using PCR mutagenesis. An alanine codon (GCT) was inserted at position 2 because it is a preferred second codon for expression of the lacZ gene (30), and two silent changes at the 5-end of the cDNA were introduced to reduce the probability of secondary structure formation in the mRNA. The complete coding region of the house fly cyt b₅ cDNA was amplified in the process of PCR mutagenesis. The forward primer was 5-CAT ATG GCT TCT TCT GAA GAC GTT AAA TAC TTT ACC-3. The reverse primer was 5-AAG CTT TGT GTC TCT CTA TGC-3. This primer corresponds to a short sequence just 3 of the stop codon and introduces a HindIII site for further cloning. The PCR product was cloned into the pCR II vector and sequenced. The NdeI/HindIII-digested TA-clone containing the PCR product was then ligated with NdeI/HindIII-digested pCWori⁺ vector to give the expression plasmid pCb5 which was used to transform the E. coli strain BL21 for expression experiments.

An overnight culture of E. coli strain BL21 (pCb5) in Terrific Broth (31) modified with a mixture of trace elements (32) and containing 200 µg/ml ampicillin was diluted 100-fold into the same medium. The flasks were shaken at 125 rpm and 37 °C until the A₆₀₀ of the culture was 1.0-1.5. Isopropyl--D-thiogalactopyranoside was then added to a concentration of 0.1 mM, and the flasks were shaken at 125 rpm and 30 °C for 40-48 h. The cells were pelleted, washed with 10 mM sodium phosphate, pH 7.4, containing 0.15 M NaCl, resuspended in 75 mM Tris-HCl, pH 8.0, containing 0.1 mM EDTA, and stored frozen at 80 °C overnight. After thawing, phenylmethylsulfonyl fluoride was added to a concentration of 2 mM, and the cells were sonicated on ice. The suspension was centrifuged at 100,000× g for 1 h, and the pellet was resuspended in 20 mM Tris, pH 8.0, containing 10% glycerol and 0.5 mM EDTA (buffer A) to give a protein concentration of 3 mg/ml. CHAPS was then added to a concentration of 1% (w/v), and the suspension was stirred for 1 h on ice. The suspension was centrifuged at 100,000 × g for 1 h, and the resulting supernatant was loaded onto a DEAE-Sepharose column equilibrated with buffer A containing 0.5% sodium cholate (buffer B). The column was washed with buffer B with 0.2% Emulgen 911, then with buffer B, and the protein was eluted with 0.1 M NaCl in buffer B. The red-colored fractions were combined, diluted 10-fold with 0.1 M sodium phosphate, pH 7.4, 10% glycerol, 0.5 mM EDTA (buffer C) and loaded onto a phenyl-Sepharose column equilibrated with buffer C. The column was washed with 10 mM sodium phosphate, pH 7.4, 10% glycerol, 0.5 mM EDTA (buffer D). A linear gradient of buffer D and buffer D containing 0.5% sodium cholate and 0.1% Emulgen 911 was then applied. The red-colored fractions were combined and applied onto a DEAE-Sepharose column equilibrated with buffer A. The column was washed with buffer A with 0.1% Emulgen 911, then with buffer B (no Emulgen 911) containing increasing concentrations of NaCl and eluted with buffer B with 0.3 M NaCl. The protein was concentrated through a PM10 membrane (Amicon) replacing the buffer with 10 mM sodium phosphate, 10% glycerol, 0.5 mM EDTA.

The membrane fractions from E. coli cells expressing CYP6A1 or P450 reductase were obtained and solubilized as described previously (27). CYP6A1 was further purified by a combination of DEAE-Sepharose and octyl-Sepharose column chromatography to apparent electrophoretic homogeneity. P450 reductase was purified using phenyl-Sepharose chromatography followed by affinity chromatography on 2,5-ADP-agarose (33). Detergents were removed on a DEAE-Sepharose column.

The EPR spectrum was acquired using an ESP 300 E Bruker spectrometer operating at 9.38 GHz with 200 microwatts of microwave power. The modulation frequency was 100 kHz, and the modulation amplitude was 10 G. The spectrum was acquired at 4.2 K using the model ESR 900 Oxford Instrument Cryostat. ¹H NMR spectra were recorded at 25 °C on a Varian Unity 300 spectrometer operating in the quadrature mode with a proton frequency of 299.997 MHz. Recombinant house fly cyt b₅ solutions (1 mM) were obtained by repeated exchanges of a concentrated aqueous solution of the protein with 30 mM phosphate buffer in D₂O at pH* = 7.0. The pH* values were not corrected for the isotope effect. One-dimensional ¹H NMR spectra were obtained using a spectral bandwidth of 15 kHz using a 30° pulse width with presaturation of the residual HOD water peak during a 500-ms relaxation delay between transients; 1000-6000 transients were collected. The time-domain data were treated with an exponential multiplication function with a line broadening of 10 Hz before Fourier transformation. Spectra were referenced to the residual HOD water peak,  = 4.84 ppm.

Linear staircase cyclic voltammetry was carried out using a BAS electrochemical system utilizing a gold disk working electrode, platinum gauze counter electrode, and silver-silver chloride miniature reference electrode as described previously (34). Before each experiment the working electrode was cleaned and surface modified with -mercaptopropionic acid for 10 min (34). The cyt b₅ solutions used for these cyclic voltammetry studies were typically 100 µM in cyt b₅. Hexamminechromium(III) chloride (0.75 mM) was used as charge mediator. Phosphate buffer (30 mM, pH 7.0) was used as the electrolyte.

The CYP6A1 activity was assayed in a 0.1 M potassium phosphate buffer, pH 7.6, at 25 °C in a final volume of 100 µl. The reaction mixture contained 50 µM heptachlor as a substrate, NADPH-regenerating system (100 µM NADPH, 2.0 mM glucose 6-phosphate, 4.0 units/ml glucose-6-phosphate dehydrogenase) and final concentrations of CYP6A1, P450 reductase, and cyt b₅ as given in the legends to Figs. 9, 10, 11. Prior to assay, CYP6A1 was incubated for 15 min on ice with P450 reductase and, where indicated, cyt b₅ in the presence of 1 mg/ml L--dilauroyl-sn-glycero-3-phosphocholine and 0.2% CHAPS in a 0.1 M potassium phosphate buffer, pH 7.6 (enzyme mixture). The reaction was started by the addition of 10 µl of the enzyme mixture to 90 µl of the reaction mixture, and after 3-5 min the incubation was stopped by the addition of 20 µl of 5 N HCl. After extraction with 0.5 ml of iso-octane, the organic phase was dried over anhydrous sodium sulfate, and 2-µl aliquots were analyzed by gas chromatography with electron capture detection (27).

1

1

1

1

The enzyme mixtures were prepared essentially as described for the CYP6A1 activity assay and, after 15 min incubation on ice, were diluted to a 1.6-2.0-ml volume with 0.1 M potassium phosphate buffer, pH 7.6, and placed in the first syringe. The concentrations of the proteins used are given in the legends to the figures. A second syringe contained 100 µM NADPH, 2.0 mM glucose 6-phosphate, and 0.2 unit/ml glucose-6-phosphate dehydrogenase in the same buffer. When CYP6A1 reduction was measured, the reaction mixture also contained 50 µM heptachlor, and the buffer was saturated with carbon monoxide by a 30-min bubbling of CO gas. No special measures were taken to remove oxygen from the solutions in order to keep the conditions as close as possible to the conditions of the CYP6A1 activity assay.

Concentrations of cyt b₅ and P450 were determined by difference spectrophotometry (35). Recombinant house fly cyt b₅ is present in a reduced form in whole E. coli cells. Therefore, sodium dithionite and hydrogen peroxide were added to the sample and reference cuvettes, respectively, to determine cyt b₅ by reduced versus oxidized difference spectra. Molar concentration of recombinant house fly P450 reductase was calculated based on protein content (36).

RESULTS

Molecular Cloning and Expression in E. coli

Conserved sequences around the two His ligands of the heme in known cyt b₅s (Fig. 1) were used to design two oligonucleotide primers for RT-PCR from larval mRNA of the house fly. Cloning of the PCR products and sequencing of a few individual clones revealed the presence of two clones (1-2 and 1-7) containing cyt b₅-like sequences (EETLDEVAGRD and EEVLIEQAGKD, respectively). These short PCR products were 53% identical at the nucleotide level.

The inserts of clones 1-2 and 1-7 were used as probes to screen a house fly ZAP cDNA library, and six identical cDNA clones with inserts of approximately 1.35 kilobase pairs were isolated. One of these, clone 16-A2 was shown to contain an open reading frame of 402 base pairs encoding a protein of 134 amino acids (15.4 kDa) at the 5-end of the cDNA. The 5-end of clone 16-A2 containing the open reading frame was deposited in GenBank^TM under the accession number L38464[GenBank] (706 base pairs). Clone 16-A2 contained the exact sequence of the initial RT-PCR clone 1-7. No cDNA clone corresponding to the initial RT-PCR clone 1-2 was found.

Comparison of the coding region of the cDNA revealed a high similarity to known cyt b₅s. The deduced protein sequence was 48% identical to rat microsomal cyt b₅ and 46.5% identical to rat outer mitochondrial membrane cyt b₅. It was only 27% identical to a portion of a Drosophila virilis sequence deposited in GenBank^TM as a cyt b₅ (39). A phylogenetic tree based on an alignment of amino acid sequences of known cyt b₅s and cyt b₅-like proteins reveals that the putative house fly cyt b₅ clusters with all known cyt b₅s and clearly differs from cyt b₅-like proteins and the D. virilis cyt b-like protein (Fig. 2).

A plasmid for expression in E. coli (Fig. 3) was constructed as described under ``Experimental Procedures.'' We obtained expression levels of 9 µmol/L culture in small scale cultures (50 ml) and 4-4.5 µmol/L culture in large scale cultures (500-750 ml). House fly cyt b₅ was found in E. coli in the reduced Fe(II) form as was reported for rat and human cyt b₅s expressed in E. coli (40), and the protein gradually oxidized during purification. Table I shows purification steps of the recombinant house fly cyt b₅. After cell lysis by sonication, most of the cyt b₅ was found in the membrane fraction (Table I). CHAPS treatment solubilized about 65% of the cyt b₅. The solubilized protein was purified by a combination of ion-exchange and hydrophobic interaction chromatography to apparent electrophoretic homogeneity (Fig. 4) with a yield of 23% (Table I). The protein mobility under the conditions of SDS-PAGE corresponded to a polypeptide of about 19 kDa.

Table I. Purification of recombinant house fly cyt b5 from a 3-liter culture Fraction Total protein Cyt b5 specific content Cyt b5 total content Purification Recovery mg nmol/mg protein nmol fold % Whole cells 3160 4.1 13,100   100 100,000 g centrifugation  Supernatant     2,200   17  Pellet 1395 6.1 8,500 1.5 65 Solubilized membranes 752 7.6 5,700 1.9 43.5 1st DEAE-Sepharose eluate 106 50 5,300 12.2 40 Phenyl-Sepharose eluate 54 67 3,700 16.3 28 2nd DEAE-Sepharose eluate, concentrated 44 68 3,000 16.6 23

Spectral Characterization of Recombinant House Fly Cyt b₅

The absorbance spectra of oxidized and reduced recombinant house fly cyt b₅ (Fig. 5) show a wavelength maximum at 413 nm in the oxidized form and at 423, 526, and 556 nm in the reduced form. These characteristics are typical of cyt b₅s (40, 41, 42, 43).

The EPR spectrum of the recombinant cyt b₅ from house fly is shown in Fig. 6. The rhombic signal with three resolved g values (g₁ = 3.07, g₂ = 3.22, and g₃ = 1.35) is very similar to those of other cyt b₅s, including bovine microsomal (44), bovine erythrocyte (45), recombinant rat microsomal (46), and recombinant rat outer mitochondrial membrane (43) cyt b₅s. The NMR spectrum of the recombinant house fly ferricytochrome b₅ is shown in Fig. 7, where the conditions of spectral acquisition were optimized for observing the hyperfine-shifted heme resonances rather than the protein resonances. From the number of resonances observed and their positions as compared to the NMR spectra of other cyt b₅s (38, 43, 47, 48) summarized in Table II, it is clear that two forms, A and B, of the protein are present in approximately equal amounts. These forms result from the two possible ways that the noncovalently attached heme molecule can be found in the protein, which differ by a 180° rotation about the , meso axis of the heme (38, 49, 50, 51). Thus, based on the results of absorption, EPR, and NMR spectroscopy we conclude that the gene cloned in this study encodes a typical cyt b₅.

Table II. Proton NMR shifts of the heme methyl resonances and A:B ratios of various cyt b5s Heme methyl Chemical shift Chicken microsomala Bovine microsomala Recombinant bovine microsomalb Recombinant human erythrocytec Recombinant rat microsomala Recombinant rat outer mitochondriald Recombinant house fly microsomale ppm B3-Me 29.9 31.0 ~31.0 31.5 31.8 31.0 30.9 B8-Me 27.5 27.5 ~27.5 ~27f 24.3 25.9 23.1 A5-Me 23.5 21.8 ~22.3 21.6 20.4 22.5 20.0 A3-Me 13.2 14.4 ~14.5 ~14f 14.4 14.7 14.9g A1-Me 12.6 11.5 ~12.8 ~11f 10.7 13.2 14.8g A:B ratio 20:1 9:1 ~9:1 9:1 1.5:1 1:1 ~1:1 a  Data taken from Lee et al. (38). b  Approximate values taken from Burch et al. (47). c  Data taken from Lloyd et al. (48). d  Data taken from Rivera et al. (43). e  This work. Assignments are based on comparison with NMR spectra of other cyt b5s (38, 43, 47, 48). f  Approximate values taken from Lloyd et al. (48). g  Assignments not yet confirmed.

Functional Characterization of Recombinant House Fly Cyt b₅

In order to investigate the role of house fly cyt b₅ in insect cytochrome P450-dependent monooxygenase systems we have characterized the electron-transfer properties of the protein, i.e. redox potential and reduction by NADPH-dependent cytochrome P450 reductase. We have also studied the effect of house fly cyt b₅ on epoxidation of the cyclodiene insecticide heptachlor catalyzed by CYP6A1 in a reconstituted system.

Fig. 8 shows the cyclic voltammogram of the recombinant house fly cyt b₅ obtained on a gold electrode modified with -mercaptopropionate and in the presence of 0.75 mM hexamminechromium(III) chloride. The peak separation, E_p = 65 mV (Fig. 8), is very similar to that observed for recombinant rat outer mitochondrial membrane and bovine microsomal cyt b₅s under the same conditions (34). The calculated midpoint potential of house fly cyt b₅ is 26 mV versus SHE, approximately 26 mV more negative than that of bovine microsomal cyt b₅ and 44 mV more positive than that of rat outer mitochondrial cyt b₅ measured under the same experimental conditions (34).

Cyclic voltammogram of 100 µM recombinant house fly cyt b₅ obtained on a gold electrode modified with -mercaptopropionate in the presence of 0.75 mM hexamminechromium(III) chloride and 30 mM potassium phosphate buffer, pH 7.0.

Purified recombinant house fly cyt b₅ was rapidly reduced in our reconstituted system by house fly P450 reductase upon NADPH addition as shown by the absorbance increase at 423 nm measured by stopped-flow spectrophotometry. The reduction of cyt b₅ followed apparent first-order kinetics with a rate constant of 5.5 s¹ (Fig. 9A). At least 95% of cyt b₅ was reduced by P450 reductase under these conditions, as shown in the inset (Fig. 9A). An increase of P450 reductase concentration did not increase the rate of cyt b₅ reduction, indicating that the rate of 5.5 s¹ is the maximal rate of cyt b₅ reduction by P450 reductase in our reconstituted system.

CYP6A1 reduction by P450 reductase was measured at 448 nm as the CYP6A1 ferrous CO complex formation in the presence of heptachlor, a CYP6A1 substrate (Fig. 9B). Under these conditions, about 80% of CYP6A1 was reduced by P450 reductase with an apparent first-order rate constant of 3.0 s¹. In a separate experiment, flash-photolysis of the preformed ferrous CO complex showed that CO binding by the reduced CYP6A1 was completed within 10 ms. Thus, the rate constant of 3.0 s¹ obtained in the stopped-flow experiments represents the rate of electron transfer from P450 reductase to CYP6A1. Neither the rate nor the amount of CYP6A1 reduced changed when the P450 reductase concentration was increased 3-fold (data not shown), indicating that the rate of 3.0 s¹ for CYP6A1 reduction by P450 reductase (Fig. 9B) was the maximal rate of electron transfer in our reconstituted system.

The results of the stopped-flow experiments presented in Fig. 9 demonstrate that, under our reconstitution conditions, the three purified recombinant proteins from the house fly interact in a catalytically competent manner. We therefore used the reconstituted system as a model to study the effect of cyt b₅ on the catalytic turnover of CYP6A1 supported by P450 reductase.

Prior incubation of P450 reductase and CYP6A1 at high concentrations in the presence of L--dilauroyl-sn-glycero-3-phosphocholine and detergent was required for high rates of heptachlor epoxidation by CYP6A1. Co-reconstitution of cyt b₅ with P450 reductase and CYP6A1 resulted in a stimulation of the rate of heptachlor epoxidation at both limiting and high P450 reductase concentrations (Fig. 10). CYP6A1 stimulation by cyt b₅ was more prominent at lower P450 reductase concentrations (Fig. 10, inset).

The dependence of CYP6A1 turnover on the P450 reductase concentration in the absence and presence of 1.0 µM cyt b₅ is shown in Fig. 11. In both cases, the heptachlor epoxidation rate increased as the concentration of P450 reductase increased. The reaction obeyed simple hyperbolic kinetics (Fig. 11B) and could be described by a set of K_m and V_max values. Addition of cyt b₅ increased the V_max of the reaction from 24 min¹ to 57 min¹ while it decreased the K_m from 0.50 µM to 0.14 µM. Thus, cyt b₅ makes P450 reductase a more effective electron donor in the presence of cyt b₅ despite the fact that cyt b₅ drains electrons from P450 reductase.

DISCUSSION

The house fly cyt b₅ is similar to mammalian, avian, plant, and yeast cyt b₅ in that it is a protein with a hydrophobic C terminus and a highly conserved pair of histidine residues at the core of the N-terminal heme binding site. The noncovalently bound heme interacts with 15 amino acid residues in the crystal structure determined for the bovine enzyme (37, 38) (marked in Fig. 1). Of these, 14 are identical or highly conserved between the fly and the vertebrate proteins, and one (Ser-71) is replaced by Met in the fly protein. The ``b₅ fold'' described by F. S. Mathews (11) has 13 invariant residues, all but one of which (Ala-34 in the house fly cyt b₅, Fig. 1) are exactly conserved in the fly protein. Nonetheless, the overall amino acid positional identity of the fly and rat microsomal protein is only 48%. A partial amino acid sequence of the cyt b₅ isolated from house fly microsomes³ is identical to the amino acid sequence deduced from our cDNA, thus confirming that our cDNA in fact codes for a microsomal cyt b₅.

A phylogenetic analysis not only shows that all vertebrate cyt b₅s cluster together, as was expected, but also reveals that the vertebrate outer mitochondrial cyt b₅ is more closely related to vertebrate microsomal cyt b₅s than to fly, yeast, or plant cyt b₅s (Fig. 2). Thus, outer mitochondrial membrane cyt b₅ may have evolved from a common microsomal cyt b₅ ancestor, perhaps after the deuterostome-protostome divergence. Our analysis also indicates that the sequence from D. virilis, highly similar (75% identity) to a Drosophila melanogaster cyt b₅-like sequence (52) and reported earlier to be a cyt b₅ (39), has some common features with cyt b₅ sequences including the heme binding region, but must represent a cytochrome other than cyt b₅. These Drosophila sequences are clearly distinct from cyt b₅s and from proteins with a cyt b₅-like core such as flavocytochrome b₂, sulfite oxidase, and nitrate reductase.

Synthetic cyt b₅ genes and natural cDNAs for cyt b₅ have been expressed previously in E. coli, either constitutively (46, 53) or under the control of lacZ (54, 55) or T7 promoter (40, 43, 56). We have expressed the cDNA of house fly cyt b₅ in the protease deficient E. coli strain BL21, under control of the double cassette of the strong synthetic isopropyl--D-thiogalactopyranoside-inducible Tac promoter. The high expression level of up to 9 µmol/L culture enabled us to purify the protein by a combination of conventional column chromatography techniques with good yield (20-25%) and without the help of an N-terminal His-tag as was done for the rat and human enzymes (40). Upon disruption of E. coli cells by sonication, most of the cyt b₅ was found in the membrane fraction and could be solubilized by the detergent CHAPS, suggesting the presence of an intact C-terminal membrane-binding domain. This suggestion is supported by our NMR studies that show relatively broad proton resonances (Fig. 7). Because the widths of the peaks in NMR spectra are proportional to rotational correlation time of the tumbling species, which in turn is a function of its size (57), the broad peaks in the house fly cyt b₅ NMR spectrum (Fig. 7) are likely to be due to cyt b₅ aggregation, possibly caused by the hydrophobic C-terminal membrane-anchoring tail.

Although the house fly cyt b₅ cDNA encodes a protein of 15.4 kDa molecular mass, the purified protein has a decreased electrophoretic mobility under the conditions of SDS-PAGE and migrates as a protein of about 19 kDa (Fig. 4). A 19.7-kDa apparent molecular mass was estimated by SDS-PAGE for the cyt b₅ purified from house fly microsomes (58).

The availability of purified house fly cyt b₅ enabled us to investigate spectral and electron-transfer properties of the protein in detail. The spectroscopic properties of house fly cyt b₅ (absorption, NMR, and EPR) indicate that this is a typical cyt b₅ protein. The g values of the EPR spectrum of the house fly cyt b₅ (Fig. 6) are characteristic of bis-histidine-coordinated heme centers (59) in which the histidine imidazole planes are approximately parallel in orientation (60). Thus, the EPR spectrum indicates that the heme binding site of house fly cyt b₅ is essentially identical, in the presentation of the protein's histidine ligands to the metal, to those of other cyt b₅s.

NMR spectroscopy is extremely sensitive to the protein environment of the heme, the orientation of the axial ligands, and the strength of the hydrogen bonds of the N-H protons of histidine ligands to protein backbone residues (61). Because of the unpaired electron of the low spin Fe(III) heme center, many of the resonances of the protons of the heme are shifted well outside the 0-10-ppm region of the protein NMR spectrum (38, 49, 50, 51, 61, 62). The similarity of the relative intensities and chemical shifts of the heme resonances of recombinant house fly cyt b₅ to those of other cyt b₅s (38, 43, 47, 48), summarized in Table II, is striking. This finding indicates that the shape of the heme pocket of recombinant house fly cyt b₅ is similar to that of other cyt b₅s. The small differences in chemical shifts represent only very minor changes in the orientation of the heme group with respect to the planes of the histidine ligands (62). The NMR spectrum (Fig. 7) also reveals that heme is incorporated into house fly cyt b₅ with the two different orientations of the heme, both forms (A and B) being present about equally. Protein residues that line the heme binding pocket create different shapes for the heme pockets of the cyt b₅s from various organisms. Each protein has a different relative stability for these two forms as shown in Table II. Thus, house fly cyt b₅ is similar to both rat cyt b₅s in the lack of specificity of heme orientation. Since the recombinant bovine microsomal (47, 53) and human erythrocyte (48) cyt b₅s, both expressed in E. coli, each have approximately 9:1 ratios, we conclude, as did Lloyd et al. (48), that the nearly complete heterogeneity of heme orientation observed for two recombinant rat proteins and house fly cyt b₅ is clearly not a result of expression in E. coli but must be due to the sizes of the protein residues that line the heme pockets of each protein.

We have developed a modified electrode that allows rapid measurement of the midpoint potential by cyclic voltammetry (34). Using this method with 0.75 mM hexamminechromium(III) chloride as charge mediator, a reduction potential of 70 mV for recombinant rat outer mitochondrial cyt b₅ was observed,⁴ while the more time-consuming spectroelectrochemical method yielded a potential of 102 mV (34). Under the same conditions, trypsin-cleaved bovine cyt b₅ has a midpoint potential of 0 mV versus SHE (34), very similar to its potential measured by spectroelectrochemistry. In this work, we have measured the midpoint potential of the house fly cyt b₅ by cyclic voltammetry and found it to be 26 mV versus SHE. These results show that the house fly cyt b₅ is more similar in midpoint potential to the bovine microsomal protein (and also the rat microsomal protein) (63) than it is to the rat outer mitochondrial membrane cyt b₅ (34).

The Fe^III/Fe^II midpoint potentials of the mammalian microsomal and erythrocyte proteins measured thus far span a range of 9 to +5 mV (48, 64, 65, 66, 67), and those for microsomal mung bean and yeast cyt b₅s are 30 mV and 23 mV, respectively (68, 69). Values for proteolytically cleaved cyt b₅s purified from insect tissues are 57 mV in case of Spodoptera eridania (70) and +6 mV in case of M. domestica (71). The wide range of reported values may reflect different methods used to determine the midpoint potential. Changes in overall surface charge are known to cause shifts in reduction potential, with up to +8 to +12 mV shift in potential for every negative charge removed (63). However, the full-length house fly cyt b₅ has a net negative charge (10 at neutral pH) close to that of outer mitochondrial membrane cyt b₅ (9) yet the midpoint potentials for these two proteins are significantly different (26 and 70 mV, respectively, as measured by cyclic voltammetry). Thus, midpoint potential of cyt b₅s cannot be rationalized on the basis of net charge alone, and investigations of appropriately designed mutants using the same titration technique will be required in order to explain the wide range of midpoint potentials of cyt b₅s.

Mammalian microsomal cyt b₅s are known to be involved in P450-dependent monooxygenase reactions (2, 3), and studies with antibodies to house fly cyt b₅ have suggested that this protein plays a similar role in insect P450 systems (28). We have demonstrated here that recombinant house fly cyt b₅ can be efficiently reduced by house fly microsomal NADPH-dependent P450 reductase, and is able to stimulate the activity of house fly microsomal CYP6A1 in a reconstituted system. It is well established that two electrons are required for a complete turnover of P450 enzymes and that NADPH-dependent P450 reductase can provide both electrons. It is widely held that reduced cyt b₅ can serve as an alternative donor of the second electron (12, 15, 16).

Under our experimental conditions, P450 reductase is able to provide the first electron at a rate higher than the epoxidation rate catalyzed by CYP6A1. Indeed, the rate constant of CYP6A1 reduction by P450 reductase is 3.0 s¹ (Fig. 9B). With 1.0 µM P450 reductase and 0.05 µM CYP6A1 present in the reaction mixture, the rate of heme reduction is 3.0 µM/s, while the rate of heptachlor epoxidation does not exceed 0.05 µM/s even in the presence of 1.0 µM cyt b₅. Moreover, total CYP6A1 turnover including both productive (heptachlor epoxidation) and nonproductive (``uncoupled'') use of molecular oxygen occurs at much lower rates than heme reduction.⁵ Therefore, CYP6A1 turnover is limited by a step following the first electron transfer from P450 reductase. When reconstituted with 1.0 µM P450 reductase, CYP6A1 turnover increases from about 15 min¹ (without cyt b₅) to about 50 min¹ with 1.0 µM cyt b₅ (Fig. 11). Because the redox potential of cyt b₅ (26 mV) is not negative enough to allow efficient transfer of the first electron, and because cyt b₅ does not increase the rate of CYP6A1 reduction by P450 reductase,⁵ the stimulation of CYP6A1 turnover by cyt b₅ (Figs. 10 and 11) appears to be a result of an increased rate of transfer of the second electron.

Analysis of heptachlor epoxidation as a function of P450 reductase concentration showed that cyt b₅ lowers the apparent K_m of CYP6A1 for P450 reductase 3-fold. This novel observation can reflect: (a) different affinities of CYP6A1 for P450 reductase in the presence and absence of cyt b₅, (b) the formation of different P450 reductase-CYP6A1 complexes for first and second electron transfer which are unmasked because cyt b₅ changes the rate-limiting step of epoxidation, and (c) another unrecognized property of the reconstituted system. Experiments designed to distinguish between these possibilities are currently under way.