Regulation of Leukotriene A4 Hydrolase Activity in Endothelial Cells by Phosphorylation*

Endothelial cells contain leukotriene (LT) A4 hydrolase (LTA-H) as detected by Northern and Western blotting, but several studies have been unable to detect the activity of this enzyme. Since LTA-H could play a key role in determining what biologically active lipids are generated by activated endothelium during the inflammatory process, we studied possible mechanisms by which this enzyme may be regulated. We find that LTA-H is phosphorylated under basal conditions in human endothelial cells and in this state does not exhibit epoxide hydrolase activity (i.e. conversion of LTA4 to LTB4). LTA-H purified from endothelial cells is efficiently dephosphorylated by incubation with protein phosphatase-1 in the presence of an LTA-H peptide substrate and not at all in the absence of substrate. Under conditions that lead to dephosphorylation, protein phosphatase-1 activates the epoxide hydrolase activity of LTA-H. Using peptide mapping and site-directed mutagenesis, we have identified serine 415 as the site of phosphorylation of LTA-H by a kinase found in endothelial cell cytosol. In parallel, we have studied a human lung carcinoma cell line that expresses active LTA-H. Although these cells have cytosolic kinases that phosphorylate recombinant LTA-H, they do not target serine 415 and thus do not inhibit LTA-H activity. We believe that LTA-H is regulated in intact cells by a kinase/phosphatase cycle and further that the kinase in endothelial cells specifically recognizes and phosphorylates a regulatory site in the LTA-H.

Immunohistochemical studies have suggested that the enzyme is widely spread, if not ubiquitous (11,12). It is interesting to note, however, that 5-lipoxygenase, the enzyme that generates LTA 4 , has a much more restricted distribution than LTA-H, and thus, many more cell types appear to have the ability to metabolize LTA 4 than have the capacity to synthesize it. This has led several groups to study the transfer of LTA 4 from PMNL to cells that lack the 5-lipoxygenase including erythrocytes (13,14), platelets (15,16), endothelial cells (16 -19), and vascular smooth muscle (20). Similar transcellular metabolism has been demonstrated in perfused organ studies, suggesting a role in cardiac and pulmonary function (21)(22)(23)(24).
We have focused on endothelial cells (EC), because the interaction of PMNL with the endothelium is the critical initiating step of the inflammatory process and because LTB 4 induces PMNL accumulation in the extravascular space and initiates the PMNL-dependent component of edema formation. Our earliest studies demonstrated that porcine aortic endothelial cells could metabolize LTA 4 to LTC 4 but not to LTB 4 (17), and later work extended these observations to other endothelial cell types including human umbilical vein endothelial cells (HU-VEC) (25). If EC contain active LTA-H, the presentation of LTA 4 by activated, marginal PMNL could provide an important source of substrate for EC LTB 4 synthesis. Unregulated, this pathway could lead to uncontrolled synthesis of this bioactive lipid and result in inappropriate accumulation of PMNL and other manifestations of inflammation. Therefore, it has been proposed that EC may have a mechanism for regulating the activity of LTA-H (19). It is this question that we address in the present report.

EXPERIMENTAL PROCEDURES
Cell Culture-Human umbilical vein endothelial cells were cultured as previously reported (25). Cells were maintained in M-199 supplemented with 20% fetal bovine serum, endothelial cell growth supplement (3 g/ml), and heparin (2 units/ml) in Petri dishes coated with 0.2% gelatin. Cells were grown at 37°C in a humidified atmosphere of 95% air and 5% CO 2 . HUVEC were used in first to third passages.
EA.hy 926 and A549 cells were kindly provided by Dr. Cora-Jean Edgell (University of North Carolina, Chapel Hill, NC). EA.hy 926 is a hybridoma line formed by the fusion of HUVEC with the human lung carcinoma cell line, A549. These cells have proven to be a good model for HUVEC in many studies (26,27). EA.hy 926 cells were maintained in Dulbecco's modified Eagle's medium (DMEM) supplemented with 10% fetal calf serum (27). A549 cells were cultured in the presence of 6thioguanine (0.1 M) in DMEM supplemented with 5% fetal bovine serum as described earlier (28).
Northern and Western Blots-Proteins, separated on 7.5% gels by SDS-PAGE, were transferred to nitrocellulose according to the method of Towbin (29). Immunostaining of the blots was done with commercially available alkaline phosphatase-linked reagents. The primary antibodies were rabbit anti-human LTA-H (raised against recombinant LTA-H) or the monoclonal anti-FLAG antibody, M2 (Eastman Kodak Co.).
Total RNA (20 g) extracted from cultured HUVEC, EA.hy 926 cells, and A549 cells according to the acid guanidinium isothiocyanate-phenol-chloroform method (30) was fractionated on a 1% denaturing aga-rose gel, blotted to Hybond-N (Amersham Corp.) by capillary action, and cross-linked with UV light.
The blots were probed with an EcoRI fragment excised from the LTA-H expression vector, pEX85 (31). The probes were labeled by random priming according to the manufacturer's instructions (Boehringer Mannheim). Radiolabeled probe was added to the hybridization solution to a final concentration of approximately 10 6 cpm/ml and incubated with the blot overnight at 42°C. The membrane was washed once with 2 ϫ SSC (0.15 M NaCl, 15 mM sodium citrate) with 0.5% SDS for 30 min at room temperature and three times for 20 min at 65°C in 1 ϫ SSC with 0.1% SDS. After drying, the blot was exposed to x-ray film overnight.
LTA-H Assay in Intact Cells-EC or A549 monolayers were rinsed twice with phosphate-buffered saline and bathed in the same buffer containing 0.5% human serum albumin (2 ml). LTA 4 (5 M; prepared as the lithium salt by saponification of LTA 4 methyl ester (32); Biomol) was added, and each dish was gently rocked for 30 min at 37°C in a 5% CO 2 incubator. The reaction was terminated with ice-cold ethanol (10 ml) containing PGB 2 as internal standard (0.3 nmol), and the samples were kept at Ϫ20°C for at least 30 min. Any precipitate was removed by centrifugation, and the sample was evaporated under reduced pressure. The residue was dissolved in mobile phase and analyzed by reverse phase HPLC (RP-HPLC) using a Nucleosil C 18 column (3 m; 3 ϫ 100 mm) eluted at 0.4 ml/min with methanol/water/acetic acid (70:30:0.1, v/v/v). Metabolites of LTA 4 were detected by UV absorbance at 270 nm.
Purification of LTA 4 Hydrolase from EC-EC were scraped from culture dishes with a rubber policeman and suspended in buffer A (50 mM Tris-Cl, pH 8.0, 5 mM EDTA, 5 mM ␤-mercaptoethanol, 2 g/ml leupeptin) at 10 6 cells/ml. The cells were disrupted by sonication with a Branson sonifier model S-250 (three times, 1 min each; duty cycle 70%, output 3). The cell supernatant was obtained by centrifuging the sonicate at 10,000 ϫ g for 15 min at 4°C. The enzyme was partially purified by ammonium sulfate fractionation (40 -70% saturation). The resulting precipitate was resuspended in buffer A and dialyzed against several changes of the same buffer overnight at 4°C. The dialyzed sample was filtered (0.45 m) and applied to an anion exchange column (Mono Q) equilibrated with 20 mM Tris-Cl, pH 8.0, on a Pharmacia FPLC system. Proteins were eluted with a linear gradient of KCl from 0 to 0.15 M over 60 min at a flow rate of 2 ml/min. Fractions containing LTA-H were identified by measuring the aminopeptidase activity of the enzyme as described below. Typically, LTA-H was found in fractions 36 -41 (approximately 65-85 mM KCl). Active fractions were pooled, and the buffer was changed to 50 mM Tris-Cl, pH 8.0, by ultrafiltration (50,000 molecular weight cut-off; Centricon).
LTA-H was further purified from the Mono Q fractions by chromatofocusing on a Mono P column (HR 5/5, Pharmacia Biotech Inc.). Samples were applied in 25 mM Bis-Tris, pH 6.7, and eluted with a linear gradient of Polybuffer 74 (pH 3.7; 10%, v/v) at a flow rate of 0.5 ml/min. Fractions were tested for aminopeptidase activity, and active fractions were pooled, concentrated by ultrafiltration, and reconstituted in 50 mM Tris-Cl, pH 8.0.
Purification of Recombinant LTA 4 Hydrolase-Recombinant LTA-H was expressed in bacteria as described elsewhere (31). The enzyme was purified from sonicated bacterial cultures essentially as described above.
LTA-H Assay with Purified Enzyme-Recombinant LTA-H (1-2 g) was incubated in 50 mM Tris-Cl, pH 8.0, with human serum albumin (2 mg/ml; 50 l) and LTA 4 (25-100 M) at 37°C for 1 min. The reaction was terminated by adding ice-cold methanol (150 l) containing PGB 2 (3.3 nmol) as internal standard. The sample was stored at Ϫ20°C for at least 30 min, and precipitated protein was removed by centrifugation. LTB 4 was quantified by RP-HPLC on a Nucleosil C 18 column eluted with methanol/water/acetic acid (65:35:0.01, v/v/v) at a flow rate of 0.4 ml/min. The eluant was monitored at 270 nm, and LTB 4 was quantified by comparison with a standard curve generated from known amounts of LTB 4 and PGB 2 . The identity of LT peaks was verified by parallel analysis of synthetic standards.
Aminopeptidase Activity Assay-The aminopeptidase activity was determined by incubating recombinant LTA-H (1-2 g) with alaninep-nitroanilide (1 mM) in 50 mM Tris-Cl, pH 8.0, with bovine serum albumin (1 mg/ml) in 96-well microtiter plates at 37°C for 30 min. Formation of p-nitroaniline was monitored at 405 nm and quantified versus a standard curve on a Multiscan plate reader (model EL312, BIO-TEK Instruments). Spontaneous hydrolysis of the substrate was corrected for by subtracting the absorbance of control incubations without enzyme.
Phosphoamino Acid Analysis-Confluent cultures were washed three times with phosphate-buffered saline and then incubated in 20 mM HEPES-buffered, phosphate-free DMEM. After 4 h, the monolayers were washed twice more and incubated with this medium (2.4 ml/ 100-mm dish) plus dialyzed serum (0.3 ml) and [ 32 P]orthophosphate (0.3 ml; stock 10 mCi/ml; NEN Life Science Products). After 1 h at 37°C, the labeling medium was removed, and the cells were washed with ice-cold phosphate-buffered saline. The cells were scraped into Netal's buffer (0.5 ml of 10 mM NaCl, 5 mM EDTA, 50 mM Tris-Cl, pH 8.0, 1 mg/ml bovine serum albumin, and 0.1% Triton X-100) with added sodium orthovanadate (1 mM) and sodium fluoride (25 mM). The cells were further disrupted by three freeze/thaw cycles, and the supernatant was obtained by centrifugation of the suspension at 10,000 ϫ g for 30 min. The supernatant was precleared by incubation with normal rabbit serum (1:200) followed by protein A-agarose (Oncogene Science). One portion of the resulting supernatant was mixed with recombinant LTA-H (10 g) before both were incubated with LTA-H antibody (1:200) on a rotating table for 1 h at 4°C. Protein A-agarose was added again to each tube, and the incubation was continued overnight at 4°C. The immunoprecipitates were collected by centrifugation and washed three times with Netal's buffer without bovine serum albumin and with 1% Triton X-100. The beads were boiled in sample buffer, and the eluted proteins were separated by SDS-PAGE.
Proteins were transferred to Problott, and the specific LTA-H band was identified and excised from the blot. The strip of membrane containing the labeled LTA-H was rinsed with water, and the protein was hydrolyzed with 6 N HCl (200 l) at 110°C for 1 h. The hydrolysate was transferred to a clean tube and reduced to dryness in a SpeedVac. The residue was dissolved in pH 1.9 buffer (5 l of 2.2% formic acid, 7.8% glacial acetic acid (v/v)) containing unlabeled phosphoamino acid standards (0.5 g each), and the phosphoamino acids were separated by thin layer electrophoresis according to the method of Hunter (33). The first dimension was run for 20 min at 1.5 kV in pH 1.9 buffer. After thoroughly drying and rotating the plate, a second dimension was run in pH 3.5 buffer (5% glacial acetic acid, 0.5% pyridine (v/v)) for 20 min at 1.3 kV. The plate was dried, sprayed with ninhydrin solution (0.25% in acetone), and developed at 65°C for 15 min. The location of the standards was noted and the plate was exposed to x-ray film at Ϫ70°C overnight.
In Vitro Phosphorylation of LTA-H by EC Lysate and Peptide Mapping-EC extracts were prepared from confluent EA.hy 926 cell cultures. The cells were washed three times with phosphate-buffered saline and scraped off the plates with a rubber policeman. EC were recovered in a pellet after centrifugation at 4°C and were resuspended in kinase buffer (100 l/100-mm plate of 20 mM HEPES, pH 7.5, 0.34 mM EDTA, 0.34 mM EGTA, 1.67 mM CaCl 2 , 1.0 mM DTT, 10 mM MgCl 2 , 1 mM sodium orthovanadate). The cells were sonicated three times for 15 s each time (duty cycle 30%, output 4), and a cytosolic extract was obtained by centrifugation at 10,000 ϫ g at 4°C for 30 min. [␥-32 P]ATP (150 M, 20 -50 Ci; NEN Life Science Products) was added to the cell lysate (300 l; approximately 1.4 mg of total protein), and the reaction was begun by the addition of recombinant LTA-H. Labeled LTA-H was separated by SDS-PAGE, blotted to nitrocellulose, and used for peptide mapping (34).
The band corresponding to LTA-H was identified, excised from the membrane, and cut into smaller pieces. The membrane pieces were incubated with polyvinylpyrrolidone (0.5%, w/v) in acetic acid (100 mM) for 30 min at 37°C to block nonspecific binding. The acid was aspirated, and the membrane pieces were washed five times with water. The immobilized proteins were then carboxymethylated with iodoacetate (100 mM) for 10 min at room temperature in the dark. After the membranes were washed three times with water, the proteins were denatured with urea (1.5 M) in borate buffer (100 l; 50 mM, pH 8.0), and the protein was digested with sequencing grade trypsin or endoproteinase Lys-C (10 g; Boehringer Mannheim) at 37°C overnight. The digested peptides were eluted from the membrane by sonication, transferred to a clean tube, and lyophilized. The residue was reconstituted in 0.1% trifluoroacetic acid, and the digest was fractionated by RP-HPLC on a Vydac semimicro C 18 column (2.1 ϫ 250 mm). Peptides were eluted with a linear gradient from 0.1% trifluoroacetic acid in water to 0.1% trifluoroacetic acid in acetonitrile over 140 min at a flow rate of 1 ml/min. The eluant was monitored at 220 nm, and 30-s fractions were collected. Radioactivity in each fraction was determined by Cerenkov counting.
Creation of FLAG-LTA-H and Site-directed Mutants-pEX85, the bacterial expression vector that contains the cDNA for human LTA-H, was used as a template to amplify the 2-kilobase insert by polymerase chain reaction (PCR). The original vector encoded a fusion protein with 10 extraneous amino acids at the N terminus. Primers were designed to replace this section with the FLAG epitope (DYKDDDDK) and were synthesized in the Columbia University DNA Core. The sense primer, ACCATGGACTACAAGGACGACGATGACAAGCCCGAGATAGTGG, contains a Kozak translation initiation consensus sequence followed by the coding sequence for the FLAG epitope and then a stretch of LTA-H sequence. The antisense sequence was 5Ј-TCAATACGCAGGTCTTTA-ATCCACTTTTAAGTCTTTCCCC-3Ј. The sense primer was phosphorylated at its 5Ј-end by T4 polynucleotide kinase.
A 20-cycle PCR reaction was run containing deoxynucleotides (25 M each), pEX85 (100 ng), sense and antisense primers (50 pmol each), and Taq DNA polymerase (5 units) in PCR buffer (50 l of 10 mM Tris-Cl, pH 8.3, 50 mM KCl, 2.5 mM MgCl 2 , 0.001% gelatin). The initial denaturing step was 2 min at 94°C followed by 19 cycles (94°C for 1 min, 55°C for 1 min, 72°C for 1 min). The final extension reaction was 7 min at 72°C. The PCR product was directly ligated to the PCR3.1 vector at 14°C overnight, and a portion of the ligation mixture (2 l) was used to transform Escherichia coli. Transformants were selected on LB plates containing kanamycin (50 g/ml). Several clones were picked and expanded, and the plasmid DNA was purified with Magic minipreps (Promega). The orientation of the insert was confirmed by restriction digestion of the plasmid, and positive clones were sequenced to ensure that the insert had no mutation.
FLAG-LTA-H cDNA was subcloned into a pET vector (Promega) for expression in bacteria and for site-directed mutagenesis. The mutation of serine 415 to alanine was accomplished with the PCR-based QuikChange system according to the manufacturer's instructions (Stratagene), and the wild-type and mutant proteins were expressed in E. coli. Bacteria were grown in LB/AMP medium to a density of 0.6 at 600 nm and were induced by the addition of isopropyl-1-thio-␤-D-galactopyranoside (0.4 mM). FLAG-LTA-H and S415A-FLAG-LTA-H were purified from the 10,000 ϫ g supernatants of sonicated E. coli by ammonium sulfate precipitation and anion exchange chromatography as described above, followed by affinity chromatography over an M2 column (Kodak). The column was washed with three aliquots of glycine, (5 ml, 0.1 M, pH 3.5), followed by three washes with Tris-buffered saline (5 ml of 50 M Tris, pH 7.4, 150 mM NaCl). Partially purified enzyme was applied to the column, which was then washed three times with Trisbuffered saline (12 ml/wash). The FLAG-tagged proteins were then eluted by the application of excess FLAG peptide (five 1-ml aliquots at 100 g/ml). Fractions were tested for aminopeptidase activity as described above. The active fractions were pooled, and the buffer was changed to 50 mM Tris-Cl, pH 8.0, by ultrafiltration. The purity of the protein was checked by SDS-PAGE stained with Coomassie Blue. In some experiments, the identity of the FLAG-tagged protein as LTA-H was verified by immunostaining with anti-LTA-H antibody.
EC Kinase Phosphorylation of FLAG-tagged LTA-H-FLAG-LTA-H or S415A-FLAG-LTA-H (8 g) was incubated with EC or A549 cell lysate as described above. Typically, the reaction was carried out at 37°C for 3 min and was stopped by rapid dilution (10 volumes of ice-cold kinase buffer). FLAG-LTA-H was purified from the reaction mixture by immunoaffinity chromatography over an M2 column and analyzed by SDS-PAGE. Gels were dried and exposed to x-ray film to localize the radioactive bands, or radioactivity was quantified on a PhosphorImager (Molecular Dynamics).
Statistical Analyses-Data are expressed as the mean Ϯ S.E. Differences between the means were determined by a paired t test or, where appropriate, by a repeated measures analysis of variance. Postexperiment comparisons of individual means were performed using Tukey's procedure.

Presence of LTA-H in Endothelial
Cells-Despite the fact that LTA-H activity was not detected in various endothelial cells, we tested for the presence of messenger RNA and enzyme protein in these cells. Both Northern and Western blots demonstrated the presence of LTA-H in HUVEC and in the endothelial cell line, EA.hy 926. Similarly, the human lung carcinoma line, A549, that served as the fusion partner for the EA.hy 926 was positive for LTA-H message and protein (Fig. 1).
We next retested these cells for the presence of LTA-H activity in an assay that measures the production of LTB 4 from exogenous LTA 4 (Fig. 2C).
To be sure that the apparent absence of LTA-H activity was not due to some inability of the cells to transport LTA 4 across the cell membrane, we repeated these assays in cell lysates and on purified enzyme obtained from these cultures. The results confirmed our findings in the intact cells. LTA-H purified from HUVEC and EA.hy 926 was inactive, while the same enzyme obtained from A549 cells readily converted LTA 4 to LTB 4 (data not shown).
Phosphorylation of the Native Endothelial Cell LTA-H-A brief examination of the deduced amino acid sequence of LTA-H showed that the enzyme contains multiple phosphorylation sites. To determine if the inactivity of the purified EC enzyme might be due to a post-translational modification by a kinase, we incubated HUVEC cultures with [ 32 P]orthophosphate, lysed the cells, and recovered the endogenous LTA-H by immunoprecipitation. The precipitate was fractionated by SDS-PAGE, blotted to nitrocellulose, and exposed to x-ray film. LTA-H incorporated radioactive phosphate under these conditions, confirming that the native, inactive enzyme is phosphorylated in intact HUVEC (data not shown). Enzyme purified from [ 32 P]orthophosphate-loaded HUVEC was hydrolyzed in 6 N HCl for phosphoamino acid analysis. The only radiolabeled amino acid in this preparation was phosphoserine (Fig. 3). Identical results were obtained with EA.hy 926 cultures (data not shown).
Peptide Mapping of Phosphorylated LTA-H-To identify the site(s) of LTA-H phosphorylation, we analyzed the peptide map of recombinant LTA-H phosphorylated in vitro by the kinase(s) in cell lysates prepared from EA.hy 926 cell cultures. Phosphorylated LTA-H was digested either with trypsin or Lys-C as detailed under "Experimental Procedures," and the resulting peptides were fractionated by RP-HPLC. Digestion with either proteinase yielded one major phosphopeptide (Fig. 4, A and B). These peptides were indistinguishable by HPLC analysis. However, the Lys-C digest of LTA-H phosphorylated by A549 lysate was clearly different (Fig. 4, B and C). Although phosphorylation was always on a serine (data not shown), these data indicate that the kinases of the two cell types targeted different sites in the LTA-H.
We focused our attention on the 14 peptides that overlap in the trypsin and Lys-C peptide maps and contain a serine residue. The radiolabeled peptide obtained from enzymatic digests of LTA-H phosphorylated by EC lysate was identified by mi- FIG. 1. Northern and Western analysis from cultured HUVEC,  EA.hy 926, and A549 cells. A, Northern blot. Total RNA was extracted from cultures as described under "Experimental Procedures." 20 g of RNA from each cell type was fractionated in a 1% formaldehyde-agarose gel, transferred by capillary action to a nylon membrane, and fixed by UV irradiation. The blot was hybridized with an EcoRI fragment from pEX85 (nucleotides 1055-1880) labeled by random priming with [ 32 P]dCTP. After washing, the membrane was dried and exposed to x-ray film. The positions of the 28 and 18 S ribosomal RNA bands are marked. The LTA-H transcript is approximately 2.2 kb as reported by others (44 -46). crosequencing to be the tetrapeptide FSYK, and thus the phosphorylated residue was serine 415.
Site-directed Mutagenesis of the Putative Phosphorylation Site-The recombinant LTA-H used in our earliest studies was expressed in E. coli as a fusion protein that contained 10 extraneous amino acids on the N terminus (31). Prior to initiating site-directed mutagenesis studies, the expression vector for this recombinant LTA-H, pEX85, was reengineered to replace this extraneous sequence with the FLAG epitope. FLAGtagged recombinant LTA 4 hydrolase (FLAG-LTA-H) was expressed and purified over an anti-FLAG immunoaffinity column.
Using a PCR-based mutagenesis, the serine corresponding to residue 415 of the wild-type enzyme was converted to alanine, and the sequence of the resulting expression vector was verified by cDNA sequencing. Mutant LTA-H (S415A-FLAG-LTA-H) was expressed in bacteria and purified over an anti-FLAG immunoaffinity column. The addition of the FLAG tag and the S415A mutation slightly reduced the specific activity of the recombinant LTA-H in the LTB 4 synthesis assay and in the aminopeptidase assay of this enzyme (Table I), but these changes were not significant.
Both the FLAG-LTA-H and S415A-FLAG-LTA-H were incubated with cell lysates from EA.hy 926 or A549 cultures in an in vitro phosphorylation assay in the presence of [␥-32 P]ATP. FLAG-tagged proteins were recovered from an immunoaffinity column and fractionated by SDS-PAGE. Wild-type LTA-H was phosphorylated by EA.hy 926 lysate (Fig. 5, lane 1), whereas the S415A mutant was not a substrate for the kinase in this lysate (Fig. 5, lane 2). A549 lysate phosphorylated LTA-H at a different site and, as expected, catalyzed the incorporation of phosphate into both wild-type (Fig. 5, lane 3) and mutant LTA-H (Fig. 5, lane 4).
Studies of the EA.hy 926 Kinase-mediated Phosphorylation of LTA-H-Despite the fact that the kinase source for these experiments is a crude cell lysate, we determined some of the characteristics of this enzyme preparation. Peak phosphorylation occurred at 3 min and then fell off slightly over the next 10 min, indicating that the lysate probably contained a phosphatase activity (Western analysis shows the presence of protein phosphatase-1 in these lysates; data not shown).

FIG. 3. Phosphoamino acid analysis of LTA-H isolated from [ 32 P]orthophosphate-loaded HUVEC.
One 100-mm dish of HUVEC (primary culture) was washed with phosphate-buffered saline and incubated at 37°C for 4 h in 20 mM HEPES-buffered phosphate-free DMEM. The medium was replaced with fresh phosphate-free DMEM supplemented with dialyzed serum (final concentration, 10%) containing [ 32 P]orthophosphate (1 mCi/ml). One hour later, the cells were lysed in Netal's buffer, and the LTA-H was recovered by immunoprecipitation with rabbit anti-LTA-H antibody and protein A-agarose as described under "Experimental Procedures." A parallel immunoprecipitation was run in the presence of excess recombinant LTA-H (10 g) to verify the specificity of any immunoprecipitated band. The precipitate was washed, boiled in SDS-PAGE sample buffer, and fractionated by SDS-PAGE on a 7.5% gel. Proteins were transferred to a Problott membrane, and the radioactive bands were identified after exposure of the blot to x-ray film. The specific LTA-H band was cut from the blot and hydrolyzed with 6 N HCl at 110°C for 1 h. The hydrolysate was lyophilized and dissolved in pH 1.9 buffer containing phosphoamino acid standards. The sample was applied to a microcrystalline cellulose thin layer plate and fractionated by two-dimensional electrophoresis as described under "Experimental Procedures." Phosphoamino acids were visualized with a ninhydrin (0.25% in acetone) spray, and the plate was exposed to x-ray film to locate the radioactive spots. The locations of the standard phosphoamino acids have been circled. This result was typical of three similar experiments. umn purification. Control incubations were carried out with the mutant S415A-FLAG-LTA-H treated under identical conditions. In these experiments, phosphorylation caused a 32 Ϯ 5% decrease in LTA 4 hydrolase activity (range, 17-40%; p Ͻ 0.01; Fig. 6, A and B, and Table II), while the activity of the mutant enzyme was not significantly altered (Fig. 6, C and D, and Table II). In addition, there was no change in the aminopeptidase activity after incubation with the lysate.
Phosphatase Treatment Activates Native LTA-H from Endo-thelial Cells-We next purified native LTA-H from HUVEC cultures and treated this enzyme with protein phosphatase-1 to determine whether dephosphorylation could reactivate the inactive native form of the enzyme. Phosphatase treatment was carried out under conditions that were shown to dephosphorylate recombinant LTA-H that had been phosphorylated in vitro (data not shown). The ability of the native EC enzyme to convert LTA 4 to LTB 4 was then retested. We found that EC LTA-H activity was significantly increased (p Ͻ 0.05, n ϭ 3) after incubation with phosphatase (Fig. 7B).

DISCUSSION
Several studies of LT metabolism in EC have failed to demonstrate the conversion of LTA 4 to LTB 4 (16 -19). In startling disagreement, experiments by Claesson and Haeggström (35) found evidence of active LTA-H in primary HUVEC cultures. Our findings confirm that EC contain LTA-H as assessed by Northern and Western blotting; however, LTA-H purified from EC was unable to catalyze the metabolism of LTA 4 . Immunoassays (12,36) and enzymatic assays (11,37) have shown that LTA-H is a widely distributed protein, but enzyme activity measurements detected a considerable tissue-dependent variation in specific activity (11). These data lead us to speculate that a system for the cellular regulation of LTA-H activity may exist (19). Until now, only indirect evidence has been generated to address this hypothesis.
What is the significance of cellular regulation of LTA-H, especially in EC that do not synthesize LTA 4 ? It is clear that EC can utilize LTA 4 generated by PMNL in a process known as transcellular metabolism (17). In fact, PMNL, which synthesize both LTA 4 and LTB 4 , have been shown to export as much as 50% of the LTA 4 generated after activation by the calcium ionophore, A23187 (38,39). Thus, the capacity of EC to take up and metabolize LTA 4 generated by PMNL, especially when the two cell types are in intimate contact, can lead to physiologically significant quantitative or qualitative changes in LT production by the system. Evidence for EC production of LTC 4 via transcellular metabolism has been presented and discussed elsewhere (16 -19). What if EC could also generate LTB 4 when in contact with activated PMNL? LTB 4 is a potent stimulus for PMNL adhesion to EC (5,40). At nanomolar concentrations, this lipid induces rolling PMNL to stick to postcapillary venules, thereby inducing prominent plasma leakage at these sites (41) and stimulates transendothelial migration in vitro (42). Similar observations of the biological activity of LTB 4 have been reported and reviewed (43), but all indicate that the production of LTB 4 by EC could be an important feed-forward signal during the development of inflammation by generating an intense local gradient of this chemotactic, PMNL-activating   4 hydrolase or S415A-FLAG-LTA-H (4 g) was incubated for 3 min at 37°C with cell lysate obtained from EA.hy 926 or A549 cultures (roughly equivalent to the 14,000 ϫ g supernatant from half of a 100-mm plate) and [␥-32 P]ATP (150 M; 50 Ci/ml). The reaction was terminated by the addition of 10 volumes of cold buffer followed by rapid purification of the FLAG-tagged proteins over an immunoaffinity column. FLAG-tagged proteins were analyzed by SDS-PAGE, and the dried gel was exposed to x-ray film to localize radioactive bands. lipid. Although we have not yet identified a specific signal or set of signals that activates LTA-H in intact EC, we are confident that such a signal exists for two reasons. First, the findings of Claesson and Haeggström (35) clearly demonstrate that this enzyme can exist in an active form within EC. Second, we can activate LTA-H purified from EC by treatment with protein phosphatase (Fig. 7). This reasoning led us to concentrate on possible mechanisms by which EC regulate LTA-H.
In EC, under basal conditions, we find that LTA-H is phosphorylated and that this post-translational modification inhibits LTA-H catalyzed conversion of LTA 4 to LTB 4 . Phosphorylation does not affect the aminopeptidase activity of the enzyme. In our studies, LTA-H purified from endothelial cell cultures was inactive, explaining the apparent lack of enzyme activity in intact cells. Treatment of purified endothelial cell LTA-H with a protein phosphatase restored activity to the enzyme, providing support for a kinase/phosphatase-dependent regulation of LTA-H.
Protein kinases are ubiquitous and exist in many varieties.
We find that LTA-H is a substrate for one or more kinase found in the cytosol of at least two cell types, A549, a human lung carcinoma cell line, and EA.hy 926, an endothelial cell-like hybridoma line derived from the fusion of A549 with HUVEC. However, the phosphorylation in these cases occurs at distinct sites, and it is only the EC kinase that inhibits LTA-H activity.
In addition, A549 cells contain active LTA-H, suggesting that the LTA-H kinase is differentially distributed. This may account for the variations in specific activity of this enzyme seen in various tissues. What is the kinase in EC that phosphorylates LTA-H? Current data leads us to predict that this enzyme is not a known kinase and suggest that it may be a previously undescribed enzyme. For example, by the comparison of peptide maps, it is  7. Phosphatase treatment activates LTA-H purified from HUVEC. LTA-H was purified from HUVEC as described under "Experimental Procedures." This enzyme (1 g) was incubated in phosphatase buffer (50 mM Tris-HCl, pH 7.0, 0.1 mM CaCl 2 , 1 mM NiCl 2 , bovine serum albumin 0.5 mg/ml) in the presence (B) or absence (A) of protein phosphatase-1 (0.2 units) at 37°C for 30 min. Both incubations contained p-nitroaniline (the aminopeptidase reaction product, 1 mM), which had been shown earlier to enhance the ability of protein phosphatase-1 to dephosphorylate LTA-H. After this, LTA 4 (50 M) was added, and the reaction was run for an additional 1 min at 37°C. The reaction was stopped with ice-cold methanol containing the internal standard (PGB 2 ; 3.3 nmol), and the samples were analyzed as described in the legend to clear that the EC LTA-H kinase is not one of the ␣-, ␤-, or ␥-isoforms of protein kinase C. By this criterion, the kinase identified in A549 lysates is probably one of these isoforms, since it phosphorylates the LTA-H at the same site as commercially prepared protein kinase C. 2 In addition, LTA-H kinase is stable at 4°C for weeks and is not dependent upon calcium or added phospholipid. This tends to rule out the protein kinase C family as candidates. We have observed that LTA-H is not a substrate for PKA in vitro, and although LTA-H is phosphorylated by casein kinase and the insulin receptor kinase, both produce a very weak signal and have no effect on LTA-H activity. 3 Finally, the sequence at which the LTA-H is phosphorylated does not fit with any known consensus kinase sequence. Although this analysis is limited, we are proceeding with our effort to purify the EC kinase under the assumption that it is a novel protein kinase.
Finally, by site-directed mutagenesis, we have unequivocally identified the site at which the EC LTA-H kinase phosphorylates and thereby regulates the enzyme. The change of serine 415 to alanine does little to alter the activity of the enzyme; therefore, it is not clear what role, if any, this residue may play in the catalysis of epoxide hydrolysis. However, the mutant LTA-H is not a substrate for the EC LTA-H kinase, and its activity is not altered by incubation with the cell lysate under conditions known to result in phosphorylation and inhibition of the wild-type protein.
We have consistently obtained modest levels of inhibition of LTA-H in the studies described in this report, and our phosphorylation conditions yield a product that incorporates only about 0.1 mol of phosphate/mol of enzyme. In a limited number of trials, we have observed that the inhibition of LTB 4 synthesis doubled with a doubling of the incorporation of phosphate (n ϭ 3). 2 It seems likely that there are at least two independent reasons for the low level of phosphate incorporation in these experiments. First, our studies of phosphate incorporation versus time indicate a peak at 3 min that falls slowly after 10 min. This result is probably caused by the action of a protein phosphatase in the kinase preparation. This is further supported by studies with a more purified kinase preparation isolated from EC supernatant (and able to phosphorylate wild type but not S415A-LTA-H). These studies have produced a somewhat higher stoichiometry of phosphate incorporation (0.5 mol/mol). 2 However, it is clear from the protein phosphatase experiments that the phosphorylation site is somehow protected within the three-dimensional structure of the enzyme. Thus, protein phosphatase-1 alone will not strip radiolabeled phosphate from LTA-H. The phosphatase is active only in the presence of the aminopeptidase substrate or product. This result suggests that the mere presence of a protein phosphatase in the cell cytosol is insufficient to activate the native, inactive LTA-H, but rather a second event (i.e. the presence of substrate) is also required.
The protection of the phosphorylation site from the action of phosphatase implies that this site may also be protected from the LTA-H kinase. This may explain our routine finding of 30 -40% inhibition of recombinant enzyme phosphorylated by cell lysate. The native EC enzyme, on the other hand, may be phosphorylated as a nascent protein and thus would be synthesized in an inactive state. These data suggest that some mechanism exists to regulate LTA-H in endothelial cells and presumably in other tissues as well. Based on our findings, we believe that LTA-H is tightly regulated by phosphorylation by a specific LTA-H kinase.