Photoaffinity Labeling and Site-Directed Mutagenesis of Rat Squalene Epoxidase

Squalene epoxidase (SE) (EC 1.14.99.7) is a flavin- requiring, non-cytochrome P-450 oxidase that cata- lyzes the conversion of squalene to (3S)-2,3-oxidosqualene. Photolabeling and site-directed mutagenesis were performed on recombinant rat SE (rrSE) to elucidate the location and roles of active-site residues important for catalysis. Two new benzophenone containing analogs of NB-598, a nanomolar inhibitor of vertebrate SE, were synthesized in tritium-labeled lower SE activity. Mutations at Cys-490 and Cys-557 produced proteins with negligible SE activity, implcating these residues as being either structurally or catalytically essential. Chemical modification of wild- type and Cys mutants with a thiol-modifying reagent support the existence of a disulfide bond between Cys- 490 and Cys-557. © 2000 Academic Press

Squalene epoxidase (SE)2 (EC 1.14.99.7) catalyzes the conversion of squalene to (3S)-2,3-oxidosqualene (1, 2). This reaction and the subsequent cyclization of (3S)-2,3-oxidosqualene by oxidosqualene cyclase (OSC) to lanosterol are the key steps in the conversion of acyclic lipids to sterols in plants, fungi, and vertebrates.

Vertebrate liver SE is a membrane-bound protein that requires molecular oxygen, flavin adenine dinu- cleotide (FAD), NADPH-cytochrome P-450 reductase, NADPH, and a soluble protein factor (SPF) (4). SPF, a 47-kDa protein, has been purified and shown to be involved in intermembrane squalene transport (5); it may be replaced by 0.1% Triton X-100 (TX-100) in enzyme assays. SE catalyzes the transfer of molecular oxygen to either squalene or 2,3-oxidosqualene (6), and this process occurs with retention of the C-3 olefinic hydrogen (7). Mammalian SE has been purified from rat (4) and pig (8, 9) liver microsomes. A cDNA for the 63-kDa rat SE has been cloned and a truncated 50-kDa His-tagged construct was overexpressed (10). Inhibition of SE has become an important pharmaceutical target for the regulation of cholesterol biogenesis in humans (3, 11). However, a detailed understanding of the protein–protein and protein–ligand interactions relevant to the catalytic mechanism still remains elu- sive. In order to characterize the substrate binding site, a variety of inhibitors based on squalenoid (12–18) and non-squalenoid (19 –21) structural templates, time-dependent inhibitors (22), and photoaffinity la- bels (23, 24) have been synthesized and evaluated in vitro.

In this study, we first describe the preparation of two new photoaffinity analogs based on the allylamine SE inhibitor NB-598, a nanomolar competitive inhibitor of rat SE (25) with in vivo cholesterol-lowering abilities in dogs and in vitro activity in Hep G2 cells (26). To localize the binding site of one of these inhibitors, the photocovalently modified protein was digested with en- doproteinase Lys-C. HPLC purification and mass spec- trometry indicated a covalent adduct between N-(6,6- dimethyl-2-hepten-4-ynyl)-N-ethyl-3-(aminomethyl)- benzophenone (PDA-I) and a tripeptide, Asp-Ile-Lys, a Lys-C fragment beginning at Asp-426 of rat SE (27). To address the importance of this site, we present the mutation of each of these residues to alanine, and the determination of the enzymatic activity of the resulting constructs.

Second, a novel conserved sequence motif was re- ported among the flavoprotein hydroxylases (28). On the basis of crystal structure and site-directed mu- tagenesis (29) studies of p-hydroxybenzoate hydroxy- lase (PHBH) from Pseudomonas fluorescens, several residues were identified that might show a dual func- tion in both FAD and NADPH binding. The conserved amino acids Asp-159, Gly-160, and Arg-166 are neces- sary for maintaining the structure. The backbone oxy- gen of Cys-158 and backbone nitrogens of Gly-160 and Phe-161 were found to interact indirectly with the phosphate moiety of FAD. Moreover, it was known from mutagenesis studies that the side chain of the moderately conserved His-162 was important for NADPH binding. From these observations and protein sequence alignments, we selected five corresponding conserved residues of SE for site-directed mutagenesis. Finally, we evaluated which cysteine residues were involved in disulfide bonds or were present as free thiols. For example, a Cys thiol has been implicated in SE function and in the sensitivity of rat and human SE to tellurium-derived species (30 –32). To address this question, each of the seven Cys residues of recombi- nant rat SE (rrSE) were mutated individually to Ala, and the SE activity of each mutant was examined. The number of free thiols in wild-type and mutant con- structs was then determined by colorimetric titration.

EXPERIMENTAL PROCEDURES

Chemicals. Sodium borotritide (NaBT4) was obtained from NEN Life Science Products, Inc. (Boston, MA). Nanopure (Barnstead) wa- ter was used in all experiments. Buffer salts and reagents were purchased from Sigma (St. Louis, MO), Fisher (Springfield, NJ), and Boehringer-Mannheim (Mannheim, Germany). DEAE-Sephacel and Blue-Sepharose were purchased from Pharmacia (Uppsala, Sweden). DTNB was purchased from Pierce (Rockford, IL). Liquid scintillation counting was performed on an LKB 1218 RackBeta instrument using Fisher Scintiverse II.

Synthesis of unlabeled compounds. PDA-I and N-(6,6-dimethyl- 2-hepten-4-ynyl)-N-ethyl-3-[3-(aminomethyl)phenoxy]methylben- zophenone (PDA-II) were synthesized in unlabeled form following standard protocols (27).

Synthesis of tritium-labeled PDA-I ([3H]PDA-I). First, 3-formyl- benzophenone 1a (2.8 mg, 13.5 µmol) in 2 ml of 95% ethanol was added to 500 mCi of NaBT4 (specific activity (SA) = 60 Ci/mmol) in 500 µl of 95% ethanol containing 0.01 N NaOH, and the reaction was monitored by thin-layer chromatography (TLC) (20% ethyl acetate/ hexane) until conversion to the primary alcohol was complete (20 min). An additional quantity (39 mg) of the aldehyde was added during 1.5 h to consume the remaining NaBT4. The mixture was concentrated, diluted with H2O (200 µl), neutralized with 6 M HCl, extracted with ether (3 × 200 µl), saturated sodium chloride solution (200 µl), dried over MgSO4, and concentrated. The residue was purified on silica gel (20% ethyl acetate/hexane) to give 26 mg of tritiated 3-hydroxymethylbenzophenone ([3H]3) (SA = 15 Ci/mmol). A portion (13 mg) of the tritium-labeled alcohol was dissolved in 500 µl CH2Cl2 and 118 mg (9 eq) of pyridinium chlorochromate (PCC) was added. The reaction was stirred overnight at room temperature, diluted with ether (1 ml), filtered through Florisil with 3 × 1 ml ether, concentrated, and purified on silica gel (10% ethyl acetate/ hexane) to give 11.8 mg of the [3H]1b (SA = 15 Ci/mmol). [3H]1b in 500 µl of THF/MeOH (20:1) was stirred at 4°C while excess ethylamine (condensed at 16°C from the gas) was added. The solution was acidified to pH 6.0 with acetic acid, NaBH3CN (5.2 mg, 82.9 µmol, 1.54 eq) was added, and the reaction was stirred over- night at room temperature. The mixture was concentrated, and the residue was dissolved in water and then extracted with ethyl acetate. The organic extract layer washed (water, brine), dried, concentrated, and purified on silica gel (20% ethyl acetate/hexane) to give 9.0 mg of the [3H]N-ethylamine derivative 2 (SA = 12.3 Ci/mmol).

[3H]2 in 500 µl of DMF containing K2CO3 (6.36 mg, 60 µmol, 1.5 eq) was stirred at 4°C as 1-bromo-6,6-dimethyl-2-hepten-4-yne (33) (8 mg, 40 µmol, 1 eq, freshly prepared) in 500 µl of DMF was added, and the reaction was stirred overnight at room temperature. The solvent was removed, the residue was partitioned between H2O and ether, the organic phase was washed (2% tartaric acid, satd. NaHCO3), dried, concentrated, and purified on silica gel (10% MeOH/ CH2Cl2) to give [3H]PDA-I (SA = 5 Ci/mmol).

Synthesis of tritium-labeled PDA-II ([3H]PDA-II). Tritiated alco- hol [3H]3 (10 mg) in 500 µl of CH2Cl2 was added to a solution of N-bromosuccinimide (15.7 mg, 88.2 µmol, 1.5 eq) in CH2Cl2, containing (CH3)2S (6.6 mg, 106 µmol, 1.5 eq) at —20°C (24). The reaction was warmed to 0°C, stirred for 3 h, then diluted with pentane, and poured over ice water, and the organic layer was washed with brine, dried, and concentrated. The residue was purified on silica gel using 20% ether in pentane to give 6.0 mg of bromide [3H]4 (SA = 15 Ci/mmol).

A solution of N-(6,6-dimethyl-2-hepten-4-ynyl)-N-ethyl-3-hydroxy- benzylamine 5 (5.2 mg, 19 µmol, 0.88 eq) (24) in THF (200 µl) was treated with NaH (60% suspension in oil, 37.7 µmol, 1.3 eq) at 0°C for 15 min. Tetra(n-butyl)ammonium iodide (0.82 mg, 2.9 µmol, 10 mol%) was then added, and the reaction was stirred for an additional 15 min at 0°C. [3H]3-Bromomethylbenzophenone 4 (6 mg, 22 µmol, 1.13 eq) in THF (100 µl) was added, and the reaction was stirred for 2 h at room temperature and then diluted with H2O and CH2Cl2. The aqueous phase was extracted twice with CH2Cl2, and the organic layer was dried and concentrated. The residue was purified on silica gel (20% ethyl acetate/hexane) to give 6 mg of [3H]PDA-II (SA = 14 Ci/mmol).

Photoaffinity labeling of native and recombinant rat liver SE. Partially purified rat liver SE (20 µl, 0.5 mg/ml) was diluted with buffer A to a final volume of 200 µl (final Triton X-100 concentration of 0.05%), and pre-incubated with NB-598 (30 µM) for 1 h at room temperature. Photoaffinity labels ([3H]PDA-I and [3H]PDA-II) were added to final concentrations of 0.1 to 100 µM, pre-incubated for 1 h at 4°C, and irradiated at 360 nm for 45 min. Since photoaffinity labeling is a non-equilibrium process, pre-incubation is required for specific labeling and competitive displacement experiments. Each solution was transferred to a microfuge tube, and 800 µl of ethanol was added to precipitate protein and remove non-covalently bound photoaffinity labels and detergent. The samples were separated by SDS–PAGE, and the gels processed for fluorography (34, 35) and exposed to X-ray film at —80°C for 3 weeks.

Truncated recombinant rat liver SE (Δ99 His-tag) (4.5 µg 1.3 mg/ ml) in 50 mM Tris–HCl, pH 7.4, containing 0.3 M KCl, 1% MEGA-8, 5% glycerol, 1 mM EDTA, and 1 mM DTT was diluted with 50 mM Tris–HCl, pH 7.4, to a final volume of 200 µl, including 1 µl 10% Triton X-100 (final Triton X-100 concentration of 0.05%), and pre- incubated with 30 µM of either NB-598 or trisnorsqualene cyclopro- pylamine (TNS-CPA) (13) for 1 h on ice. Either [3H]PDA-I or [3H]PDA-II was added to give final concentrations of 1.0 and 0.5 µM PDA-I, and 0.5 and 0.1 µM PDA-II, pre-incubated for 1 h on ice, and then irradiated for 45 min at 360 nm, as described above.

Assay method for rrSE at pH 7.4. SE inhibition and electro- phoretic separations were performed as described (9, 14, 15). Each purified mutant protein was reconstituted as follows: SE (50 µl; 0.55 mg/ml), 0.06 unit of NADPH-cytochrome P-450 reductase, 1 mM of NADPH, and 0.1 mM of FAD in a total volume of 200 µl of 20 mM Tris–HCl, pH 7.4. This reconstituted enzyme solution was pre-incu- bated at 37°C for 10 min. Then, 15,000 cpm of [14C]squalene in 1 µl of 2-propanol was added to the mixture (final substrate concentra- tion, 6.75 µM). Incubation was continued at 37°C for 50 min. The enzymatic reaction was quenched by addition of 200 µl of 10% KOH in methanol and incubated at 37°C for 30 min. The nonsaponifiable lipids were extracted with 1 ml of CH2Cl2 by vigorous vortexing and then centrifuged at 1,000g for 5 min. The organic extracts were evaporated in a Savant Speed-Vac for 15 min. The residue was dissolved in 50 µl of CH2Cl2 and applied to the pre-adsorbent layer on a silica gel thin layer plate (Whatman silica gel 60 Å, 20 × 20 cm, 250-µm layer). The plate was developed with 5% ethyl acetate in hexane. Radioactive regions in each lane were visualized using the IBM Bioscan Imaging Scanner System 200 with autoexchanger 3000. Rf values were 0.84 for squalene and 0.45 for oxidosqualene. All assays were carried out in duplicate.

Effects of metal ions and multiple inhibitors on rrSE. To test whether the His-tag could affect inhibition of SE activity, 4.5 µg of rrSE was incubated with 0.06 unit of NADPH-cytochrome P-450 reductase, 840 µM NADPH, and 29 µM FAD in a total volume of 200 µl of 20 mM Tris-HCl, pH 7.4, in the presence of 0.42–100 µM NiCl2 or ZnCl2. This solution was pre-incubated for 10 min at 37°C and assayed for SE activity as described.

To test the effects of PDA-I on inhibitory potency of NB-598, 4.5 µg of rrSE was incubated with 0.06 unit of NADPH-cytochrome P-450 reductase, 840 µM NADPH, and 29 µM FAD in a total volume of 200 µl of 20 mM Tris–HCl, pH 7.4, in the presence of 270 µM PDA-I. This solution was pre-incubated for 10 min at 37°C and assayed for SE activity as described.

Photoinactivation of rrSE. To determine if photoaffinity labels could block the active site of SE, 4.5 µg of rrSE was pre-incubated with PDA-I and PDA-II at final concentrations of 1 and 0.5 µM PDA-I, and 0.5 and 0.1 µM PDA-II, for 1 h on ice. The solutions were then irradiated at 360 nm for 45 min. An aliquot from each solution was assayed for SE activity and compared to control solutions with photolabels that were not irradiated, as well as controls that did not contain photolabels but were irradiated.

Site-directed mutagenesis. Mutagenesis of the Δ99 His-tag-rrSE containing plasmid (10, 36) was performed using Quikchange mu- tagenesis kit (Stratagene, La Jolla, CA). This truncated enzyme, which lacked the N-terminal membrane association domain, had full enzymatic activity and was substantially easier to express and pu- rify. The oligonucleotide primers (25–30 mer), each complementary to opposite strands of the vector, were designed for each mutation. For each reaction, the sample contained 50 ng of plasmid, 100 ng of oligonucleotide primers, reaction buffer, dNTP, and Pfu DNA poly- merase. Sixteen cycles of amplification after denaturing at 95°C, annealing at 55°C, and extension at 68°C (12 min per cycle) resulted in nicked circular strands. Following temperature cycling, the prod- uct was treated with DpnI at 37°C for 1 h to digest the parental DNA template and to select mutation-containing synthesized DNA. The nicked vector DNA incorporating the desired mutation was then used to transform competent XL-1 Blue cells. Several colonies from ampicillin plates were selected and grown in 5-ml cultures, and DNA was isolated by phenol extraction. The integrity of the mutated plasmid was confirmed by sequencing. All mutant proteins were expressed in the BL21 (DE3) strain of Escherichia coli.

Purification of the mutant rrSE. Luria broth (LB) media was inoculated with a given colony, incubated for 3 h, and induced by adding IPTG to a final concentration of 0.4 mM at 37°C. After 3 h of additional incubation, the cells were pelleted by gentle centrifuga- tion. Cell pellets were resuspended in 10 ml of binding buffer (5 mM imidazole, 0.5 M NaCl, and 20 mM Tris–HCl, pH 8.0) with additives (1 mM PMSF, 1 mg/ml lysozyme, 0.5% TX-100, and 5% glycerol) and stirred for 1 h on ice. This suspension was frozen, thawed, and sonicated. The lysate was clarified by centrifugation at 32,000g for 30 min and dialyzed against binding buffer including 0.5% TX-100 and 5% glycerol. Supernatant was applied to a Ni-NTA agarose column (bed volume, 3 ml), which was pre-equilibrated with binding buffer containing 0.5% TX-100 and 5% glycerol. The column was washed (60 mM imidazole, 0.5 M NaCl, and 20 mM Tris–HCl, pH 8.0), and the protein was eluted (1 M imidazole, 0.5 M NaCl, and 20 mM Tris–HCl, pH 8.0). The active fraction was dialyzed against buffer C (50 mM Tris–HCl 7.4, 1 mM EDTA, 5% glycerol, 0.5% TX-100, and 0.5 mM DTT). Purified SE mutants were analyzed by electrophoresis under standard conditions using SDS–PAGE gel (10%) in Tris-gly- cine buffer, followed by Coomassie blue staining.

Determination of the KM value for FAD. The purified mutant proteins were reconstituted as follows: SE (50 µl; 0.55 mg/ml), 0.12 unit of NADPH-cytochrome P-450 reductase, and 1 mM NADPH; the optimal FAD concentration was between 0 and 100 µM in a total volume of 200 µl of 20 mM Tris–HCl, pH 7.4. This reconstituted enzyme solution was pre-incubated at 37°C for 10 min. Then, 15,000 cpm of [14C]squalene in 1 µl of 2-propanol was added to the mixture (final substrate concentration, 6.75 µM), and incubation was contin- ued at 37°C for 50 min.
Colorimetric quantitation of free sulfhydryls. The purified wild- type and cysteine mutant constructs were used for determination of disulfide bond formation. DTNB reagent (50 µl) and 3 µg of protein were prepared in 0.1 M sodium phosphate buffer pH 8.0. The mixture was incubated at room temperature for 15 min, and then the absor- bance was measured at 412 nm and compared with wild-type protein.

RESULTS

Previous efforts to prepare photoaffinity analogs of NB-598 employed 4-benzoylbenzyl and 3-[(trifluoro- methyl)diazirine]benzyl substituents as photophoric groups (24). Each of these materials was essentially inactive, showing an IC50 value greater than 500 µM. Efforts with the [3H]-labeled diazirines from that study led to nonselective and nonspecific labeling of proteins in partially purified native rat and pig SE (H.-K. Lee and M. Ceruso, unpublished results). On the basis of patented compounds reviewed by Banyu scientists (37), it became clear that the benzoyl substituents should be relocated from the para-position to the meta- position to give the 3-benzoylbenzyl photophores. The structures and IC50 values for NB-598, TNS-CPA, PDA-I, and PDA-II with native rat SE. Photoaffinity labeling of native rat SE with [3H]PDA-I was observed at concentrations as low as 2.5 µM; however, each of the major protein bands in the partially purified preparation showed considerable nonspecific labeling. In contrast, [3H]PDA-II labeled only the three major SE bands (native form and pro- teolytic fragments) at 0.1 µM. No competition could be seen in the presence of excess NB-598 (30 µM) with either of these photolabels and the native enzyme (27). Since photoaffinity labeling is a nonequilibrium pro- cess, pre-incubation with inhibitors allowed occupancy of the substrate or inhibitor site prior to introduction of the tritium-labeled photoprobe.

An N-terminally truncated rrSE possessing a His- tag on the carboxy-terminal end (Δ99 His-tag) was la- beled only with [3H]PDA-I. No labeling was observed when up to 10 µM of [3H]PDA-II was used. Further- more, neither PDA-I nor PDA-II substantially inhib- ited rrSE. In order to determine whether the difference between the IC50 values of these photolabels and the native and recombinant forms of the enzymes was due to the absence of the first 99 residues of the N-termi- nus, this experiment was repeated using the untrun- cated His-tag rrSE. This recombinant form of SE pos- sesses the His-tag at the C-terminus and retains the membrane-spanning first 99 residues at the N-termi- nus. The same profile of SE activity as a function of PDA-I concentration was obtained with the truncated and full-length recombinant enzymes (38). The pres- ence of 270 µM PDA-I had no measurable effect on the inhibitory potency of NB-598 (IC50 = 4 nM) for the enzyme. Complexation of the His-tag by the addition of Ni2+ and Zn2+ also had no effect on the enhancement of rrSE activity by PDA-I. Irradiation of rrSE with unla- beled PDA-I and PDA-II at increasing concentrations, followed by a standard SE assay, showed a 10% reduc- tion in activity, equivalent to loss of activity in no- inhibitor control irradiations.

Despite these conflicting inputs on the inhibitory nature of these photoactivatable analogs of NB-598, experiments were conducted to localize the binding site for [3H]PDA-I. Thus, Δ99 His-tag-rrSE was photola- beled, the labeled protein was digested with endopro- teinase Lys-C, and the peptide fragments were sepa- rated by reversed-phase HPLC. Those peptides that contained radioactivity were analyzed by matrix as- sisted laser desorption ionization–time of flight (MALDI-TOF) mass spectrometry. In contrast to the result expected for nonspecific labeling, only a single major labeled peptide was isolated. The mass spectrum of this adduct appeared at m/z 749.8 and was consis- tent with only one possible Lys-C fragment, the tri- peptide Asp-Ile-Lys (DIK), resulting from cleavage of C-terminal to Lys-425 and to Lys-428. The calculated mass for a covalent adduct of the tripeptide plus the inhibitor would be m/z 733.4. The additional 16 mass units suggest that the allylamine moiety may have undergone oxidation to an N-oxide or an epoxide in the adduct, a predictable outcome in benzophenone pho- to affinity labelling studies.

Mutagenesis was performed using the Δ99 His-tag- rrSE (10, 36), a truncated enzyme lacking the N-terminal membrane association domain, which none the less had full enzymatic activity and was substantially easier to express and purify. Each mutant construct was purified using a Ni-NTA column, which selectively bound His-tagged proteins, followed by Blue-Sepha- rose column, which bound enzymes that use FAD and NAD(P) cofactors, as described previously in purification of native (9) and recombinant SE (36). The purity of each construct was confirmed using 10% SDS–PAGE (Fig. 3a). Purified mutant proteins were then assayed for SE activity.

Mutations of the charged residues in the Asp-Ile-Lys tripeptide labeled by the photoaffinity label PDA-I re- duced SE activity, while alteration of the hydrophobic residue resulted in no change in SE activity. Thus, D426A and K428A showed less than 25% SE activity relative to the wild-type rrSE (Fig. 4a). Kinetic studies were carried out for D426A and K428A, which showed higher KM values for FAD of 44.7 and 40.3 µM, respec- tively. The wild-type rrSE showed a KM value for FAD of 4.5 µM (Fig. 5a), while essentially no change was observed for the I427A mutant (data not shown). No difference was seen in the KM value for NADPH (38) because NADPH binds primarily to P450 reductase and not directly to SE. For electron transfer to occur, there must be electrostatic interaction between the cofactor and the protein (41). It thus appears reason- able that alteration of Asp-426 or Lys-428 would alter the affinity of the rrSE for FAD cofactor.

This mutagenesis data suggested that PDA-I was not directed to the substrate-binding site of SE, a re- sult that was further supported by the failure of com- petitive inhibitors of the enzyme to displace the photo- labeling (27). Ryder has proposed (42) that the al- lylamine inhibitors may bridge the substrate-binding site and the FAD-binding site. Our results provided partial support for this hypothesis, in that FAD bind- ing was impaired by mutation at PDA-I labeling site.

We therefore pursued mutagenesis of other residues expected to participate in FAD binding. Next, five highly conserved flavoprotein residues were selected on the basis of crystal structure and mutagenesis studies of PHBH (28, 29). The consensus sequence between FAD-utilizing enzymes is shown in Fig. 6. There are three common motifs found among flavoproteins: FAD fingerprint I, a conserved motif, and FAD fingerprint II. On the basis of the crystal structure of PHBH, one can deduce three residues im- portant for maintaining the structure in an analogous region of SE; each of these was selected for mutagen- esis (D284, G285, and R291). In addition, two more residues from FAD fingerprint II were selected for mu- tagenesis (G406 and D407). Each construct was puri- fied as described above (Fig. 3b). Mutations of the conserved residues resulted in 4-fold or greater loss of enzymatic activity relative to wild-type rrSE. Thus, D284A, G406L, and D407A showed less than 25% of wild-type rrSE (Fig. 4b). The KM value for FAD with activity relative to the wild-type rrSE, suggesting ei- ther an altered conformation or the potential involve- ment of one or more sulfhydryl groups in the catalytic function. For the C165A, C199A, and C379A con- structs, the enzymatic activity dropped to an average of 52% (47%, 56%, and 54%, respectively) in 1 day at 4°C and to 19% (22%, 18%, and 17%, respectively) in 2 days at 4°C. It is possible that one or more disulfide bonds were present, since both C490A and C557A constructs abolished the activity completely, while C500A and C533A both showed 50% reduction of activity com- pared with wild-type rrSE.

In order to explore the presence and location of di- sulfide bonds, Ellman’s reagent, 5,5′-dithiobis(2-nitro- benzoic acid) (DTNB), was employed to quantify the free sulfhydryl groups in wild-type and cysteine mu- tant proteins in the non-denatured. The enzymes were examined in their unreduced, non-denatured states. Table I presents the relative SE activity for each protein.

DISCUSSION

We have described photolabeling and site-directed mutagenesis experiments performed on rrSE to eluci- date the location and role(s) of active-site residues im- portant for catalysis. On the basis of allylamine SE inhibitor NB 598, we synthesized two new photoaffin- ity analogs. These new analogs may not be directed toward the substrate-binding site of the enzyme, as evidenced by the failure of competitive inhibitors of the enzyme to displace the photolabeling and the unex- plained differences in labeling efficiency for native and rrSE. Ryder has suggested (42) that the allylamine inhibitors could bind both the substrate-binding site and the FAD-binding site; subtle changes in post- translational modification or folding could then ac- count for the different IC50 values and labeling proper- ties of the native and rrSE proteins. In addition to the requirement for exogenous FAD, protein–protein inter- actions may be required with NADPH-cytochrome P-450 reductase for electron transfer leading to forma- tion of the putative flavin hydroperoxide intermediate. It appears that the “simple” epoxidation of squalene by activated molecular oxygen required many small mol- ecule and protein cofactors.

There are clearly many important questions that remain unresolved. What are the flavin–protein inter- actions that control the rate with which the reduced flavin reacts with O2? Rate differences spanning 6 or- ders of magnitude are found among flavoproteins. What factors determine whether flavin C(4a)-hy- droperoxide should be formed, and, if it is, what deter- mines its stability? Again, enormous differences in the rates of decay of the hydroperoxide are found with different flavoenzymes. The answers to these questions will undoubtedly come from multi-faceted approaches to the problem, where protein structural information is combined with detailed steady-state and rapid reaction kinetics studies and where the properties of wild-type and site-directed mutant forms of the enzymes are compared. The effect of replacement of specific active- site residues on the rate of the individual steps in catalysis and on protein structures has been deter- mined for a number of mutant forms (46, 47). These studies have already revealed important conceptual information about the reaction mechanism and strongly suggest that physical movement of the flavin in the active site plays an important role in catalysis (48).