Characterization of the Activity of Purified Recombinant Human 6-Lipoxygenase in the Absence and Presence of Leukocyte Factors*

Purified recombinant human 5-lipoxygenase was used to investigate the catalytic properties of the protein in the presence and absence of leukocyte stimulatory factors. Recombinant human 5-lipoxygenase was purified to apparent homogeneity (95-99%) from a high expression baculovirus system by chromatography on ATP-agarose with a yield of 0.6 mg of protein per 100 ml of culture (2 X 10’ cells) and a specific activity of 3-6 rtmol of 6-hydroperoxyeicosatetraenoic acid (5-HPETE) per mg of protein in the presence of ATP, Ca2+, and phosphatidylcholine as the only factors. In the absence of leukocyte factors, the reaction catalyzed by the purified recombinant enzyme showed a half-time of maximal 5-HPETE formation of 0.5-0.7 min and was sensitive to the selective 5-lipoxygenase inhibitors BW756C (ICao = 13 MM) and L-656,224 (ICao = 0.8 p ~ ) . The reaction products of arachidonic acid oxidation were 6-HPETE and 6-trans- and 12-epi-6-tram-leukotriene B4, the nonenzymatic hydrolysis products of leukotriene A4 (LTA4),

Purified recombinant human 5-lipoxygenase was used to investigate the catalytic properties of the protein in the presence and absence of leukocyte stimulatory factors. Recombinant human 5-lipoxygenase was purified to apparent homogeneity (95-99%) from a high expression baculovirus system by chromatography on ATP-agarose with a yield of 0.6 mg of protein per 100 ml of culture (2 X 10' cells) and a specific activity of 3-6 rtmol of 6-hydroperoxyeicosatetraenoic acid (5-HPETE) per mg of protein in the presence of ATP, Ca2+, and phosphatidylcholine as the only factors. In the absence of leukocyte factors, the reaction catalyzed by the purified recombinant enzyme showed a half-time of maximal 5-HPETE formation of 0.5-0.7 min and was sensitive to the selective 5-lipoxygenase inhibitors BW756C (ICao = 13 MM) and L-656,224 (ICao = 0.8 p~) . The reaction products of arachidonic acid oxidation were 6-HPETE and 6-trans-and 12-epi-6tram-leukotriene B4, the nonenzymatic hydrolysis products of leukotriene A4 (LTA4), indicating that the purified protein expressed both the 6-oxygenase and leukotriene A4 synthase activities (ratio 6:l). The microsomal fraction and the 60-90% ammonium sulfate precipitate fraction from sonicated human leukocytes did not increase product formation by the isolated enzyme when assayed in the presence of ATP, Ca2+, and phosphatidylcholine. These factors were found to stabilize 6-lipoxygenase during preincubation of the enzyme at 37OC with the assay mixture but they failed to stimulate enzymatic activity when added at the end of the preincubation period. The results demonstrate that human 5-lipoxygenase can be isolated in a catalytically active form and that protein factors from leukocytes protect against enzyme inactivation but are not essential for enzyme activity.
The 5-lipoxygenase from leukocytes is the first enzyme involved in the conversion of arachidonic acid to leukotrienes. The enzyme catalyzes both the oxygenation of arachidonic * The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
3 To whom correspondence and reprint requests should be addressed. acid to 5-HPETE' (5-oxygenase activity) and the further conversion of 5-HPETE to unstable allylic epoxide LTA, (LTA, synthase activity) (Yamamoto, 1989;Samuelsson and Funk, 1989). LTA, then serves as substrate for enzymatic hydrolysis to LTB, and conjugation with glutathione to yield cysteinyl-peptidoleukotrienes or can be nonenzymatically degraded to trans-LTB, isomers (Borgeat et aZ., 1985). LTA, synthesis by 5-lipoxygenase represents the first step in the biosynthesis of leukotrienes, a group of mediators which has been implicated in the pathophysiology of inflammatory and allergic reactions (Brain and Williams, 1990;Lewis and Austen, 1984;Ford-Hutchinson, 1989).
Mammalian 5-lipoxygenases have been shown to require ATP and Ca2+ for activity (Yamamoto, 1989) unlike other lipoxygenases from plants and animals. The rat and human 5-lipoxygenases have been shown to translocate from the soluble to the membrane fraction during leukocyte activation Wong et al., 1988), and, recently, a membrane-bound 18-kDa protein has been identified as an essential protein for leukotriene biosynthesis . The cDNA for human 5-lipoxygenase has been cloned Matsumoto et al., 1988) and encodes a 674-amino acid protein of molecular weight 78,000. The amino acid sequence shows 93% identity with that of rat 5-lipoxygenase and also indicates the presence of two regions of weak homology to a 17-amino acid consensus sequence of Ca2+dependent membrane-binding proteins (Balcarek et al., 1988).
A hydrophobic region with homology to other lipoxygenases has been identified which contains conserved histidine residues and is postulated to represent an iron-binding domain (Shibata et al., 1988). No ATP-binding site could be predicted from the sequence information.

Properties of Purified Recombinant
Human 5-Lipoxygenme 5073 lipoxygenase cDNA (Noguchi et al., 1989;Nakamura et al., 1990). However, studies on purified human 54ipoxygenase have proven to be extremely difficult because of the tediousness of the purification procedures, the low recoveries of activity (1-4%), and the extreme lability of the purified enzyme. Furthermore, previous studies have shown that the purified protein from human leukocytes had very low levels of activity unless a membrane fraction and other fractions generated during the purification procedure were added to the assay mixtures Samuelsson, 1985, 1990). Thus, most of our information on the reaction catalyzed by human 5-lipoxygenase has been obtained from activity measurements in leukocyte extracts or in the presence of stimulatory fractions and inference from studies of the soybean lipoxygenase and purified 5-lipoxygenase from other sources.
In this study, we describe the isolation and characterization of recombinant human 5-lipoxygenase from a baculovirus high expression system. The purified recombinant human 5lipoxygenase was found to be catalytically active, which has allowed for the study of the reaction catalyzed by the isolated protein and a re-examination of the effect of various protein stimulatory factors from leukocytes.

MATERIALS AND METHODS
Cells and Virus-Spodoptera frugiperda (Sf91 insect cells and Autographa californica nuclear polyhedrosis virus were obtained from C. Richardson, Biotechnology Research Institute, Montreal, Quebec. Cell counts and percent cell viability were determined in 0.2% trypan blue. Cells were cultured in Grace's complete medium (Gibco Laboratories) supplemented with 10% fetal bovine serum (Flow Laboratories Inc.), TC Yeastolate and TC lactalbumin hydrolysate (Difco Laboratories), 50 pg of gentamicin sulfate per ml and 2.5 pg of amphotericin B (Fungizone) per ml in either Falconware T flasks (Becton Dickinson Labware) or spinner flasks (Bellco Glass, Inc.) at 28°C following the procedures of Summers and Smith (1987). Escherichia coli, JM107, was obtained from Bethesda Research Laboratories. The bacterial cells were transformed according to published methods (Hanahan, 1985).
Construction of Recombinant Transfer Vector pJVETLZH5LO-The baculovirus expression vector pJVETLZ was generously provided by J. Vialard and C. Richardson (Biotechnology Research Institute, Montreal, Quebec). This vector differed from the recently published vector pJV(Nhe1) (Vialard et al., 1990) in that the P10 promoter was replaced by the ETL promoter (Crawford and Miller, 1988) of wild type baculovirus. The vector, pGBT-H5LO, was provided by T. Nguyen, Merck Frosst Centre for Therapeutic Research. A 2.1-kb EcoRI-StuI restriction fragment from pGBT-H5LO containing 11 bp of 5'-noncoding, the entire coding region, and 85 bp of 3'-noncoding sequence of the human 5-lipoxygenase cDNA  was blunt-end-ligated into NheI-cut pJVETLZ. Miniprep DNA from JM107 transformants was isolated, and the orientation of the inserted fragment with respect to the polyhedrin promoter was confirmed by digestion with EcoRV. The 16.1-kb vector, PJVETLZ-H~LO, was isolated from minipreps and purified on NACS columns (Bethesda Research Laboratories) and was subsequently used for co-transfection.
DNA Transfections and Plaque Assays-Wild type A. californica nuclear polyhedrosis virus DNA (1 pg) was mixed with pJVETLZ-H5LO (2 pg) and co-transfected into Sf9 cells by the calcium phos-phate method (Summers and Smith, 1987). After 7 days, the cellular debris was spun down, and the supernatant was used as the source of recombinant virus for the first plaque assay. Plaque assays were carried out in culture dishes (100 X 15 mm) as previously described (Summers and Smith, 1987). The infected cells were overlaid with 10 ml of 10% SeaPlaque agarose (FMC Corp., Marine Colloid Div., Rockland, ME) diluted in Grace's complete medium plus 100 pg/ml Bluo-Gal (Bethesda Research Laboratories) (50 mg/ml in dimethylformamide). Blue plaques indicative of the presence of @-galactosidase were visible within 4-6 days. The plaques were picked with a Pasteur pipette and placed in 1 ml of Grace's complete medium. The virus was allowed to elute from the agarose plug overnight at 4°C. From this initial plaque assay, 12 blue plaques were subjected to dot-blot analysis (see below) and probed with a fragment of the human 5lipoxygenase cDNA. Seven of the twelve plaques (58%) which expressed @-galactosidase activity also contained the 5-lipoxygenase cDNA (data not shown). Recombinant virus from three of the seven 5-lipoxygenase positive plaques was used for the second round of plaque assays and performed with 1-, lo2-, and 103-fold dilutions of virus, from which pure recombinant virus containing both 5-lipoxygenase and &galactosidase was obtained. Usually, two to three rounds of plaque assays were sufficient in order to purify recombinant virus from wild type virus. The recombinant virus was used to infect Sf9 cells so as to produce viral stocks with titers of 10' to lo9 plaqueforming units/ml. All three separate isolates of recombinant virus, upon infection of Sf9 cells, produced virtually equivalent amounts of 5-lipoxygenase (data not shown). One of them, designated rvH5L0(8-

I), was used in subsequent experiments.
Nucleic Acid Dot-Blot Hybridizations-24-Well culture plates were seeded with 3-5 X lo5 cells in 1 ml of Grace's complete medium and infected with virus derived from blue plaques. Infected cells from week-old cultures were used for the dot-blot assay according to a published method (Summers and Smith, 1987). The probe, an 813bp ClaI-BamHI 5-lipoxygenase cDNA fragment  was labeled with [a-32P]dCTP using the Multiprime labeling system (Amersham, Canada) to specific activities of >lo9 dpm/pg. The membranes were exposed to X-Omat AR film (Kodak) for 2-4 h at Polyacrylamide Gel Electrophoresis and Immunoblots-Various protein samples were prepared in electrophoresis loading buffer (0.1 M Tris-HC1, pH 6.8, 4% SDS, 10% @-mercaptoethanol, 30% glycerol, 0.05% bromphenol blue) and were applied to 10% polyacrylamide gels according to the method of Laemmli (1970). For immunoblot analysis, proteins were electrophoretically transferred to nitrocellulose (Schleicher & Schuell/Baxter, Montreal, Quebec) and probed with the rabbit anti-5-lipoxygenase LO-33 antiserum, generated against the 5-lipoxygenase purified from human leukocytes according to the procedure of Rouzer and Samuelsson (1985). The antigen-antibody complexes were detected with "'1-protein A (Towbin et al., 1979). The nitrocellulose membrane was exposed to X-Omat AR film (Eastman Kodak Co.) with intensifying screens for 2-6 h at -70 "C.
Preparation of Lysates from Infected Cells-Sf9 cells were grown at 27 "C in 100-ml spinner flasks to a cell density of 1.5-2 X lo6 cells/ ml and infected for 44-48 h with rvH5L0(8-1). The cells were then collected by centrifugation (900 X g for 10 min, at 20"C), washed twice with Dulbecco's phosphate-buffered saline (pH 7.4) (25 m1/2 X 10' cells) and resuspended at 1.2 X lo7 cells/ml in a homogenization buffer containing 50 mM potassium phosphate (pH 7.9) 2 mM EDTA, 1 mM dithiothreitol, 0.5 mM phenylmethylsulfonyl fluoride, and 60 pg/ml soybean trypsin inhibitor. The cells were then lysed by sonication at 4 "C using a Cole Parmer (4710 Series) Ultrasonic homogenizer (3 to 5 bursts of 10 s with 30-5 lags, pulse mode, 70% duty cycle, and output setting at 3). The preparations were examined under the microscope to achieve efficient cell lysis (>go%) with minimal sonication. The lysate was then centrifuged at 100,000 X g for 1 h (Beckman L5-65, 60 Ti rotor) at 4 "C and the resulting supernatant (S-100 fraction) brought to 24 pg/ml PC by the addition of a 250-fold concentrated solution in ethanol. The S-100 fraction (1-3 mg/ml) was stable for several hours at 4 "C and could be stored for several months at -70 "C in 20% ethylene glycol with about 50% recovery of activity. Affinity Purification of Recombinant Human 5-Lipoxygenme-A chromatography column (0.7 X 15 cm) containing 2 ml of ATPagarose gel was equilibrated at 4 'C in 0.1 M Na+-HEPES buffer (pH 7.31, containing 1 mM EDTA, 1 mM dithiothreitol, and 24 pg/ml PC. The S-100 fraction (8 ml) was applied at a flow rate of 5 ml/h, and the column was then washed at a flow rate of 10 ml/h with 2 ml of the equilibration buffer. After a high salt wash (10 ml of 0.5 M NaCl -70 "C.

Properties of Purified
Recombinant Human 5-Liprygenase in equilibration buffer), the column was re-equilibrated with 5 ml of buffer before elution of 5-lipoxygenase with 20 mM ATP in equilibration buffer (Wiseman, 1989). All buffers were filtered through 0.22pm membranes (Millex-GS, nonsterile, Millipore) and purged with helium gas. Fractions containing 5-lipoxygenase activity were identified using the spectrophotometric assay and pooled. The enzyme could be stored as aliquots in 20% ethylene glycol at -70°C for several months with approximately 50% recovery of activity. RP-HPLC Assay for S-Lipoxygenase Actiuity-The enzyme reaction was performed in 0.025 M Na'/K*-phosphate buffer (pH 7.3) containing 0.5 mM CaCI,, 1 mM ATP, 24 pg/ml PC, and 20 p M arachidonic acid in a final volume of 200 pl. Arachidonic acid was added as 2 pl of a 2 mM solution in ethanol prior to the initiation of the reaction with the addition of enzyme. After an incubation of 5 min at room temperature, the reaction was stopped by mixing with 0.8 ml of diethyl ether/methanol/l M citric acid (304:l) containing 100 ng of 13-HOD as internal standard. The mixture was centrifuged at 1,000 X g for 5 min. The upper phase was collected and dried under NZ, and the residue was redissolved in 100 pl of the RP-HPLC solvent system. This sample was injected on a Nova-Pak C18 column (Waters) eluted isocratically with acetonitrile/water/acetic acid (6535:O.l) at 2 ml/min. The effluent was monitored at 235 nm for the detection of 5-HETE and 5-HPETE which eluted at 3.0 and 3.4 min, respectively. These producta were quantitated from an absorbance area relative to the 13-HOD standard and were corrected for a low level of background using control samples incubated in the absence of enzyme or arachidonic acid. The reaction products were 5-HPETE for the purified enzyme and a mixture of 5-HPETE and 5-HETE for assays of the S-100 fraction or for the purified enzyme in the presence of protein factors. The 5-lipoxygenase activity is expressed as the maximal amount of 5-H(P)ETE accumulated after incubation (Rouzer and Samuelsson, 1985) under conditions where arachidonic acid is not limiting (<40% conversion). The amount of protein for the calculation of specific activity was measured using the Bio-Rad dye-binding assay (Bradford, 1976) and bovine serum albumin as standard.
Spectrophotometric Assay of 5-Lipoxygenase Activity-The kinetics of 5-HPETE production was measured from the increase in Am upon incubation of the purified enzyme with arachidonic acid, as previously described for the assay of porcine 5-lipoxygenase (Riendeau et al., 1989a). The reaction was measured at room temperature in 0.05 M sodium phosphate buffer (pH 7.4), 0.4 mM CaCI2, 24 pg/ml PC, 20 p~ arachidonic acid and was initiated by the addition of enzyme. ATP concentrations were adjusted to 0.2 mM taking into account the contribution of ATP from the enzyme solution. The optimal velocity of the reaction was determined from the maximal rate of conjugated diene formation (Am), using an t = 23,000 M" cm" for 5-HPETE (Gibian and Vanderberg, 1987).
Resolution of the Reaction Products of 5-Lipoxyge~se and Measurement of LTA, Synthase Actiuity-The reactions were performed using purified recombinant human 5-lipoxygenase under the conditions described for the spectrophotometric assay. The reaction (200 p l ) was stopped after a 4-min incubation by the addition of 400 pl of diethyl ether/methanol/l M citric acid (3041). The upper phase was removed and dried under Nl, and the residue was dissolved in 100 pl of methanol. NaBH, (0.01%) was added (1 pl of a 1% solution in methanol), and the mixture was brought to 0.04% acetic acid after 2 min. Samples were then injected onto a Nova-Pak Cla column at a flow rate of 1 ml/min using methanol/water/acetic acid (75:25:0.01) as eluting solvent (Ueda et aL, 1986). This chromatography system resolves ti-trans-LTB,, 12-epi-6-trans-LTB4, LTB,, (5S,6R)-di-HETE, and 5-HETE with retention times of 5.0, 5.5, 5.9, 10.7, and 23.2 min, respectively. The elution of 5-HETE was followed by monitoring absorbance a t 235 nm and that of the trienes at 270 nm.

Expression of Human 5-Lipoxygenclse cDNA in Insect
Cells-Time course analysis of human 5-lipoxygenase expression in insect Sf9 cells after infection with the recombinant virus rvH5L0(8-1) was undertaken to optimize conditions for the production of soluble and active enzyme. At various time points over the course of the 96 h of infection, aliquots were removed and analyzed by SDS-PAGE and immunoblot. A 75-kDa protein, expressed in infected Sf9 cells, was readily detected on a Coomassie Blue-stained gel 40 h postinfection (Fig. 1) and was present as early as 24 h as detected by The S-100 ( S ) and P-100 (P) fractions were prepared from 1.8 X 10" cells at 24 h or 0.9 X 10' cells at 40,48.56, 72. and 96 h postinfection. The proteins were subjected to SDS-PAGE, transferred to nitrocellulose membranes, and successively incubated with the LO-33 antiserum (1/300 dilution) and "'1-protein A as described under "Materials and Methods." The immunoblot was exposed to x-ray film with intensifying screens for either 25 h were determined using aliquots from a 100-ml culture removed at the indicated times postinfection.
immunoblot using antisera generated against the purified human 5-lipoxygenase (see Fig. 2A In order to determine the proportion of 5-lipoxygenase in the soluble fraction, S-100 and P-100 fractions from sonicated cells were analyzed by immunoblot ( Fig. 2A). Maximal level of 5-lipoxygenase was reached in the S-100 fraction by 48 h, with about half of the amount of the protein also found in the P-100 fraction at this time. The level of soluble 5-lipoxygenase then decreased with time such that, by 96 h, greater than 90% of the protein resided in the P-100. These data correlated well with the activity profile of the soluble fractions, showing a peak at 44-48 h (Fig. 2 8 ) where high specific activities (300-600 nmol of 5-H(P)ETE/mg of protein) were observed. The activity decreased rapidly by 56 h, along with cell viability (50%) and the redistribution of 5-lipoxygenase from the S-100 to the P-100 fraction. The P-100 fraction contained barely detectable levels of 5-lipoxygenase activity. Therefore, in order to maximize the amount of soluble 5-lipoxygenase that could be purified from synchronously infected Sf9 cells, the cultures were processed 48 h postinfection.
Affinity Purification of Recombinant Human 5-Lipoxygenase by ATP-Agarose Chromatography-The enzyme from the infected cell extracts was purified by chromatography on ATPagarose according to a procedure similar to that described for the partial purification of rat 5-lipoxygenase (Wiseman, 1989) with the addition of PC to the buffers to stabilize the enzyme. The elution profile of Fig. 3 shows that 5-lipoxygenase can be selectively adsorbed onto this gel with most of the proteins eluting in the flow-through or after a high salt (0.5 M NaCI) wash of the column. The 5-lipoxygenase was eluted using a buffer containing 20 mM ATP with recoveries of activity of 30-50% and specific activities for different enzyme preparations ranging from 3 to 6 pmol of 5-HPETE per mg of protein. From specific activity values and a molecular weight of 78,000 for the enzyme, it can be calculated that the 5-lipoxygenase reaction would undergo a maximum of about 500 turnovers before enzyme inactivation. Analysis of the various fractions by SDS-PAGE shows that the elution of enzyme activity with ATP coincided with that of a highly purified 75-kDa protein as revealed by Coomassie Blue staining (Fig. 4A) and immunoblot analysis with the antiserum against the leukocyte 5lipoxygenase (Fig. 4B). About 0.6 mg of purified 5-lipoxygenase could be recovered from 100 ml of infected cell culture (2 X 10' cells).
The migration of the purified enzyme from insect cells on SDS-PAGE was identical with that of the 5-lipoxygenase as  peripheral hlood leukocytes using the procedure of Rouzer and Samuelsson (1985) ( I x u k . .5-I.[)). The proteins were analvzed by SIIS-PACE and visualized either hv Coomassie Blue staining or autoradiography of the immunohlot proteins using a 1/500 dilution of the L O 4 3 antiserum.
isolated from human leukocytes according to the procedure described by Rouzer and Samuelsson (1985). These results are shown in Fig. 5 , using both the Coomassie Blue staining and immunoblotting techniques to identify t h e protein. \'arious enzyme preparations were estimated to he 95 to homogenous after electrophoresis by scanning densitometry of the stained gel.
Kinetics of Arachidonic Acid Oxidation by Purified Human 5-Lipoxygenase-The kinetics of 5-HPETE production by the purified 5-lipoxygenase was followed spectrophotometrically from the increase in AZ3, (conjugated diene) upon incubation of the enzyme with arachidonic acid, ATP, Ca2+, and PC (Fig.  6). The reaction progress curve shows the typical profile for 5-lipoxygenase with a rapid apparent first order decay in activity (ts,z = 0.6-0.7 min) and maximal product formation being reached by 3 min, before complete substrate consumption. Removal of Ca2+ ions from the incubation caused a very large decrease in the optimal rate of the reaction, which was attained only after a long (>1 min) lag phase (Fig. 6). Calcium concentrations had to be >0.1 mM for maximal activity (data not shown). Less than 10% activity was observed in the absence of PC (data not shown), and no activity could be detected with the omission of both Ca2+ and PC (Fig. 6). The reaction was inhibited by the selective 5-lipoxygenase inhibitors L-656,224 and BW755C, with a short lag phase being also observed during inhibition by BW755C (Fig. 6). The ICso values for the inhibition of the purified human 5-lipoxygenase by L-656,224 and BW755C were 0.8 @VI and 13 PM, respectively.
The optimal velocity of the reaction, as estimated from the initial rate of conjugated diene formation by spectrophotometry, varied linearly with enzyme concentration (0.2-5 pg/ml) (data not shown). The specific activity of the freshly purified enzyme using this assay was 3-6 pmol of 5-HPETE/min/mg of protein. RP-HPLC analysis of the reaction mixture after incubation indicated that 5-HPETE is the predominant product that absorbs at 234 nm, with only traces of 5-HETE (<2%) being detectable (data not shown).
Reaction Products of Arachidonic Acid Oxidation by Human 5-Lipoxygenuse-The products of the 5-lipoxygenase reaction were resolved by RP-HPLC to determine whether the purified enzyme also possesses LTA, synthase activity in the absence of protein stimulatory factors. Fig. 7 shows the chromatogram of the 5-lipoxygenase reaction products obtained after sodium borohydride reduction and as monitored by absorption at 235  and 270 nm. Under these conditions, 5-HPETE is reduced to 5-HETE which was detected as a major peak eluting at the position of the synthetic 5-HETE standard (23 min) and having an absorption maximum at 235 nm (data not shown). Two other peaks which had a higher absorption at 270 nm were found to elute at the positions of 6-trans-LTB4 and 12epi-6-trans-LTB,. These peaks were about equal in intensity and showed the UV spectra characteristic of that of trienes (not shown) and correspond to the known nonenzymatic hydrolysis products of LTA4. Other very minor peaks were not identified; one of them eluted at the position of (5S,6R)-diHETE and could have arisen from the 6R-oxygenase activity of the enzyme, as dcmonstrated for the porcine 5-lipoxygenase (Ueda and Yamamoto, 1988). The ratio of the products from the 5-lipoxygenase activity (5-HETE from 5-HPETE) to that of the LTA, synthase activity (trans-LTB, isomers) was about 6:l.
Effect of Protein Factors from Human Leukocytes on the Actiuity of Purified 5-Lipoxygenase-The high recovery of 5lipoxygenase activity after the rapid affinity purification raises questions on the requirement of the various protein factors from human leukocytes that have been reported to stimulate enzyme activity (Rouzer and Samuelsson, 1985). Two of the factors were the microsomal (P-100) pellet and the proteins precipitating at 60-90% saturation of ammonium sulfate (60-90% precipitate). A stimulation of the activity of the purified recombinant enzyme by these factors could be observed using identical amounts of protein factors and assay conditions to those described by Rouzer et al. (1988a) and Rouzer and Kargman (1988). These data are summarized in Table I (assay condition 1). Each of the 60-90% precipitate or P-100 factors caused a significant increase in 5-lipoxygenase activity with a maximal 12-fold stimulation being observed in the presence of both factors. However, this higher level of specific activity could also be measured in the absence of protein factors under the standard assay conditions used in the present work (assay condition 2), which are somewhat milder than those used for the leukocyte enzyme (no prein-

TABLE I Influence of assay conditions on the requirement of leukocyte protein factors for maximal activity of purified human 5-lipoxygenase
The specific activity of purified human recombinant 5-lipoxygenase was determined in the absence or the presence of the indicated protein cofactors under different sets of conditions. The amount of protein factors added and assay condition 1 were as described by Rouzer et al. (1988b) and Rouzer and Kargman (1988), using an incubation mixture (200 pl) consisting of 100 p~ arachidonic acid, 2 mM ATP, 2 mM CaCI2, 1.6 mM EDTA, 1 mM dithiothreitol, 2 p~ 13-HPOD, 1.6 mg/ml P-100, 0.5 mg/ml 60-90% precipitate, and 11 pg/ml affinitypurified 5-lipoxygenase. The assay was performed at 37 "C using a preincubation time of 5 min and a reaction time of 10 min. Reactions were initiated by the addition of a concentrated solution (2 pl) containing arachidonic acid and 13-HPOD in ethanol. The incubation mixture for assay condition 2 contained 20 pM arachidonic acid, 1 mM ATP, 0.5 mM CaC12, 24 pg/ml PC, 4.8 pg/ml affinity-purified 5lipoxygenase, and the same amounts of protein factors as under condition 1 when included. The reaction time was 5 min at room temperature, and no preincubation was used. Protein factors contained no appreciable amount of activity when tested individually. The specific activity is reported as nmol of 5-H(P)ETE formed per mg of protein at the end of the incubation period (means rt S.E.). ND, not detectable.  cubation, stabilization by PC, room temperature, and a lower substrate concentration). Under these conditions, the specific activity could be increased another 50% by raising the temperature to 37 "C, although the results were more variable than at room temperature (see Table I). The presence of the factors was strongly inhibitory at either temperature, which could be due to a nonspecific binding of arachidonic acid (initially 20 p~) at high protein concentration. These results demonstrate that the stimulatory effect of the leukocyte factors are dependent on assay conditions, but that the factors are not essential for the optimal activity of human 5-lipoxygenase. Additional experiments showed that neither of the 60-90%

Assay conditions Protein factors added
precipitate nor the mixture 60-90% precipitate + P-100 could stimulate the activity when added after the enzyme had been preincubated at 37°C (Table 11). Furthermore, half of the maximal activity can be measured in the absence of protein factors by eliminating the preincubation period, as compared to 8% using a preincubation time of 5 min (Table 11). Therefore, the stimulatory effects of the protein factors on the purified human 5-lipoxygenase can be accounted for by a protection of the enzyme against rapid loss of activity during incubation in the reaction medium.

DISCUSSION
Using improved procedures for the expression and affinity purification of recombinant human 5-lipoxygenase, it has been possible to isolate the enzyme in a catalytically active form and further investigate the effects of leukocyte factors on enzyme activity. The expression of high levels of soluble binant Human 5-Lipoxygenme 5077 enzyme was achieved in Sf9 cells infected with recombinant baculovirus carrying the cDNA for human 5-lipoxygenase  downstream of the strong polyhedrin promoter. The production of 5-lipoxygenase in the S-100 fraction from Sf9 cells infected with rvH5L0(8-1) (300-600 nmol of 5-HPETE/mg of protein) represents a 10-fold improvement in yield over the previously described preparation from insect cells . The reasons for this increase are not clear since the transfer vectors used in both systems were essentially identical in the nucleotide sequences spanning the polyhedrin start codon. One difference was that the 5-lipoxygenase cDNA used by Funk et al. (1989) contained 34 bp of 5'-noncoding and 162 bp of 3"noncoding sequence versus only 11 and 85 bp of 5'-noncoding and 3'-noncoding sequences, respectively, used here. Perhaps translational efficiency was suboptimal since the amount of expressed protein was lower than would be expected from the abundance of the 5-lipoxygenase mRNA they detected . It is possible that the removal of most of the GC-rich 5"noncoding region of the cDNA contributed to the increase in translated protein. Differences in mRNA structure have been shown to affect translational efficiency (Kozak, 1983, and references therein). Human 5-lipoxygenase was purified by a single chromatography step on ATP-agarose to a specific activity of 3-6 pmol of 5-HPETE/mg of protein, a value in the range of, or slightly higher than, those reported for the purified human enzyme expressed in E. coli (Noguchi et al., 1989), yeast (Nakamura et al., 1990), and other mammalian 5-lipoxygenase (Yamamoto, 1989), with 30-50% recovery of activity. The latter value is significantly higher than those of 1.5% and 1% obtained for the recombinant enzymes from bacterial and yeast extracts, respectively, and of that of 4.4% for the human enzyme purified from leukocytes (Rouzer and Samuelsson, 1985) which required several chromatography steps. The adsorption of human 5-lipoxygenase on ATP-agarose and its selective elution with ATP but not with high salt concentrations, as reported for the rat enzyme (Wiseman, 1989), suggests that adsorption on ATP-agarose occurs through a specific ATP-binding site, presumably at the site involved in the stimulation of enzyme activity by ATP. The localization of this site, as well as the mechanism by which ATP stimulates the reaction, remains to be characterized since sequence comparisons have failed to reveal any homology between 5-lipoxygenases and ATP-binding domains of other proteins Matsumoto et al., 1988). The catalytic properties of purified recombinant human 5lipoxygenase in the presence of ATP, Ca2+, and PC as the only factors were similar to those reported for the leukocyte enzyme when measured in extracts or in the presence of various protein fractions from leukocytes. Homogeneous preparations of recombinant 5-lipoxygenase expressed both oxygenase and LTA, synthase activities, as reported for either the purified leukocyte enzyme assayed in the presence of protein factors (Rouzer et al., 1986) or the recombinant enzyme from E. coli (Noguchi et al., 1989). Furthermore, the ICs0 values for the inhibition of the purified recombinant enzyme by L-656,224 (0.8 p M ) were similar to that reported for cell-free extracts from leukocytes (0.4-1 p~) , human 5lipoxygenase (Belanger et al., 1987), or purified porcine 5lipoxygenase (Riendeau et al., 1989a). The RP-HPLC profile of the reaction products of arachidonic acid oxidation by isolated human 5-lipoxygenase is almost identical with that of the enzyme from porcine leukocytes, for which a 6:l ratio between 5-lipoxygenase and LTA, synthase reaction products was measured (Ueda et al., 1986) and low amounts of (5S,6R)-diHETE originated from the 6R-oxygenase activity on the 5-

Stabilization of purified 5-lipoxygenme during incubation with leukocyte protein factors
The activity of 5.lipoxygenase was determined using 11 pg/ml purified recombinant 5-lipoxygenase and the assay condition 1, as described in the legend to Table I, with the enzyme (5-lipoxygenase) and leukocyte fractions (60-90% precipitate, P-100) being added either at the beginning of the preincubation period at 37 "C (t = -5 min) or immediately prior to the initiation of the reaction with arachidonic acid (t = 0). Enzyme activity is expressed as the total amount of 5-H(P)ETE accumulated at the end of the 10-min incubation period (means & S.E.). ND, not detectable. HPETE product (Ueda and Yamamoto, 1988). It has been reported that 5-lipoxygenase, after purification from human leukocytes, requires the presence of other protein stimulatory factors from leukocyte extracts for maximal activity. However, the present values for the unstimulated enzyme are in the range of those reported for the purified leukocyte enzyme stimulated by protein factors (3 pmollmg of protein, Rouzer et al., 1988a; 12 pmol/mg of protein, Rouzer and Samuelsson, 1985). (The latter high number may suggest that a certain amount of inactive protein may be present in most enzyme preparations.) Three stimulatory fractions were initially resolved the microsomal pellet (P-loo), the 60-90% ammonium sulfate precipitate (60-90% precipitate), and a flow-through fraction from anion exchange chromatography (Rouzer and Samuelsson, 1985), the latter fraction being optional depending on the purification scheme . From the results with the recombinant human 5-lipoxygenase, it can be concluded that the P-100 fraction and the 60-90% precipitate are not required for high activity, but that they do protect the enzyme against rapid loss of activity after dilution with the assay mixture under certain conditions. However, these factors have been shown to stimulate the kinetics of the reaction in crude preparations where the enzyme is not as labile (Rouzer et al., 1988b). The effect of the P-100 fraction on enzyme activity in crude preparations has been shown to be mimicked by replacement with PC and other membrane lipids (Puustinen et al., 1988), although we have observed that PC was not as efficient as the P-100 fraction in stabilizing the purified enzyme during preincubation in the assay mixture (data not shown).
The demonstration that human 5-lipoxygenase is catalytically active in the absence of leukocyte factors suggests that the human enzyme is more similar to other mammalian 5lipoxygenases than originally thought, in agreement with recent data that revealed a 93% identity in the amino acid sequences of the rat and human 5-lipoxygenases (Balcarek et al., 1988). However, it is possible that some other factors from leukocytes could stabilize the 5-lipoxygenase and stimulate the activity under certain specific conditions. Intuitively, one could expect that the activity of such an unstable and selfinactivating enzyme might be regulated by other components. Another unidentified protein, purified from rat leukocyte extracts by gel filtration, has been shown to replace the hydroperoxide activation of 5-lipoxygenase (Riendeau et al., 198913). Recent results have also demonstrated the presence of an 18-kDa protein from human and rat leukocytes which can modulate leukotriene biosynthesis in intact cells Dixon et al., 1990). Clearly, there is still much to be learned about the effects and the role of these various proteins and which assay conditions are the most representative of the cellular and in vivo situations.
The 5-lipoxygenase is the first enzyme of the leukotriene biosynthesis pathway and as such is an important potential drug target (Fitzsimmons and Rokach, 1989). Using the present procedure, about 0.6 mg of recombinant 5-lipoxygenase has been purified from a 100-ml suspension culture of infected Sf9 cells, this yield being at least 50-fold higher than those reported for recombinant enzymes from E. coli (90 pg/liter) (Noguchi et al., 1989) and yeast cultures (50 pglliter) (Nakamura et al., 1990). The ability to produce and readily purify large amounts of active recombinant human 5-lipoxygenase should greatly facilitate not only studies aimed at elucidating the catalytic and structural properties of the enzyme but also will allow for more detailed evaluation of novel potential inhibitors of 5-lipoxygenase to modulate leukotriene biosynthesis in human diseases.