Tyrosine Kinase Activity Modulates Catalysis and Translocation of Cellular 5-Lipoxygenase*

Tyrosine kinase activity, a determinant of Src homol- ogy domain interactions, has a prominent effect on cellular localization and catalysis by 5-lipoxygenase. Six separate inhibitors of tyrosine kinase each inhibited 5( S )-hydroxyeicosatetraenoic acid formation by HL-60 cells stimulated with calcium ionophore, in the presence or absence of exogenous arachidonic acid substrate, in- dicating that they modulated cellular 5-lipoxygenase activity. The tyrosine kinase inhibitors also blocked the translocation of 5-lipoxygenase from cytosol to membranes during cellular activation, consistent with their effects on its catalytic activity. These results fit a model which postulates that Src homology domain interac- tions are a molecular determinant of the processes which coordinate the subcellular localization and func- tions of 5-lipoxygenase. In addition, we demonstrate that activated leukocytes contain two molecularly dis- tinct forms of 5-lipoxygenase: a phosphorylated form and a nonphosphorylated form. In activated HL-60 cells the pool of phosphorylated 5-lipoxygenase accumulates in the nuclear fraction, not with the membrane or cyto- solic fractions. The amount of phosphorylated 5-lipoxy-genase is a small fraction of the total. Overall, equilib- rium reactions involving the nuclear localizing sequence, the proline-rich SH3 binding motif, and the phosphorylation state of 5-lipoxygenase may each influ- ence its partnership with other cellular proteins and any novel functions derived from such partnerships. on prominent sources 5-LO and FLAP, were incubated with 0–300 (cid:109) M tyrosine kinase inhibitors for 3 min at 37 °C. Cells were then stimulated for 10 min with 2.5 (cid:109) M A23187. Cell suspensions were quenched with 5 m M EDTA; acidified to pH 1–2; diluted with 1 ml 0.9%, w/v, NaCl; and extracted with ethyl acetate/hexane (3 (cid:51) 2 ml, 1/1, v/v). After evapora- tion of the organic solvent under N 2 , the residue was dissolved in 200 (cid:109) l of ethanol/H 2 O, 1/1, v/v, containing 1 nmol of prostaglandin B 2 as an internal standard for quantitative analysis of 5,12-diHETEs, LTB 4 , and 5( S )-HETE by reverse phase-HPLC (30). In corresponding experiments we determined the effect of tyrosine kinase inhibitors on 5-LO activity in cells stimulated with 2.5 (cid:109) M A23187 in the presence of 20 (cid:109) M arachidonic acid, added exogenously, to bypass the phospholipase-de-*

Tyrosine kinase activity, a determinant of Src homology domain interactions, has a prominent effect on cellular localization and catalysis by 5-lipoxygenase. Six separate inhibitors of tyrosine kinase each inhibited 5(S)-hydroxyeicosatetraenoic acid formation by HL-60 cells stimulated with calcium ionophore, in the presence or absence of exogenous arachidonic acid substrate, indicating that they modulated cellular 5-lipoxygenase activity. The tyrosine kinase inhibitors also blocked the translocation of 5-lipoxygenase from cytosol to membranes during cellular activation, consistent with their effects on its catalytic activity. These results fit a model which postulates that Src homology domain interactions are a molecular determinant of the processes which coordinate the subcellular localization and functions of 5-lipoxygenase. In addition, we demonstrate that activated leukocytes contain two molecularly distinct forms of 5-lipoxygenase: a phosphorylated form and a nonphosphorylated form. In activated HL-60 cells the pool of phosphorylated 5-lipoxygenase accumulates in the nuclear fraction, not with the membrane or cytosolic fractions. The amount of phosphorylated 5-lipoxygenase is a small fraction of the total. Overall, equilibrium reactions involving the nuclear localizing sequence, the proline-rich SH3 binding motif, and the phosphorylation state of 5-lipoxygenase may each influence its partnership with other cellular proteins and any novel functions derived from such partnerships. 1 catalyzes the formation of leukotriene (LT) mediators of inflammation (1,2). In resting neutrophils, 5-LO is usually confined, in an inactive state, in the cytosol. Agonist stimulation initiates the translocation of 5-LO from the cytosol to cell membranes where it can associate with an activating protein, termed FLAP (3)(4)(5)(6)(7)(8)(9)(10)(11). Translocation and interaction with FLAP are determinants of 5-LO activity in the simplified model of cellular LT formation (7). This model explains the mechanism of action of certain anti-inflammatory agents (12)(13)(14); however, it is imprecise in five respects. First, there is no evidence for a direct 5-LO-FLAP interaction; all data supporting their interaction are correlative (12)(13)(14). Second, stabilization of 5-LO by phospholipids, in vitro, fully accounts for effects originally attributed to FLAP (15,16). Third, in cells lacking FLAP, 5-LO still translocates from the cytosol to the membrane, implying that 5-LO can bind to membrane components other than FLAP (7). Fourth, in certain leukocytes, 5-LO occurs in the cell membrane or nucleus in the resting state (8,11,17). Thus, in nonactivated cells, cytosolic localization of 5-LO is not a general rule, and membrane localization of 5-LO does not necessarily correspond with FLAP interaction (18). Fifth, LT formation by receptor mediated events, such as IgE cross-linking in mast cells, depends on an unidentified signal, not a Ca 2ϩ threshold (19,20). Thus, the oversimplified model of cellular LT formation prompted us to investigate molecular determinants governing the redistribution or activation of 5-LO. We recently reported that 5-LO associates with other proteins via Src homology domain interactions, specifically via interactions with SH3 domains. We now report that cellular tyrosine kinase activity, a determinant of SH2 domain interactions, has a prominent, previously overlooked, influence on activity and cellular localization of 5-LO and that activated leukocytes contain two molecularly distinct forms of 5-LO, a phosphorylated form, and a nonphosphorylated form.
Effect of Tyrosine Kinase Inhibitors on Cellular 5-LO Activity-Neutrophils or differentiated HL-60 cells (1.0 ml, 2 ϫ 10 7 cells/ml), both prominent sources of 5-LO and FLAP, were incubated with 0 -300 M tyrosine kinase inhibitors for 3 min at 37°C. Cells were then stimulated for 10 min with 2.5 M A23187. Cell suspensions were quenched with 5 mM EDTA; acidified to pH 1-2; diluted with 1 ml 0.9%, w/v, NaCl; and extracted with ethyl acetate/hexane (3 ϫ 2 ml, 1/1, v/v). After evaporation of the organic solvent under N 2 , the residue was dissolved in 200 l of ethanol/H 2 O, 1/1, v/v, containing 1 nmol of prostaglandin B 2 as an internal standard for quantitative analysis of 5,12-diHETEs, LTB 4 , and 5(S)-HETE by reverse phase-HPLC (30). In corresponding experiments we determined the effect of tyrosine kinase inhibitors on 5-LO activity in cells stimulated with 2.5 M A23187 in the presence of 20 M arachidonic acid, added exogenously, to bypass the phospholipase-de-pendent liberation of arachidonic acid.
Effect of Tyrosine Kinase Inhibitors on Translocation of 5-LO from Cytosol to Membrane-Differentiated HL-60 cells or human neutrophils (3 ml, 2 ϫ 10 7 cells/ml) were incubated with tyrosine kinase inhibitors for 5 min at 37°C, then stimulated with 1 M A23187 for 10 min. Samples were quenched with EDTA, 5 mM final concentration, then centrifuged at 100 ϫ g for 12 min. The cells were resuspended in 1.0 ml, 0.05 M phosphate, pH 7.1, containing 0.1 M NaCl, 2 mM EDTA, 0.5 mM phenylmethylsulfonyl fluoride, 60 g/ml soybean trypsin inhibitor and disrupted by sonication for 3 ϫ 20 s (75% duty cycle, power setting 3). The homogenate was centrifuged at 12,000 ϫ g for 12 min at 4°C. The 12,000 ϫ g supernatant fraction was isolated and centrifuged at 100,000 ϫ g for 1 h at 4°C to isolate microsomal and cytosolic fractions. A portion of the 100,000 ϫ g microsomes (200 l, 4 -5 g/l) and cytosolic fraction (200 l, 1-2 g/l) was immediately mixed with 50 l of 5 ϫ electrophoresis buffer (250 mM Tris-HCl, pH 6.8, 5% w/v SDS, 50% glycerol, 100 mg/ml dithiothreitol, and a trace of bromphenol blue), boiled for 5 min, and then held at 4°C while the protein concentration was determined. Samples with equal amounts of protein were fractionated by SDS-PAGE on a 10% polyacrylamide gel with a 4% stacking gel (10 -20 g of protein/lane), then transferred to nitrocellulose membranes for Western analysis. The nitrocellulose membrane was first saturated with 5% non-fat dried milk for 1-3 h to reduce nonspecific binding, then incubated with anti-5-LO antiserum (4), 1:400 dilution for 2 h. Following equilibration, the nitrocellulose membrane was washed four ϫ 250 ml ϫ 15 min in Tween-TBS and incubated for 1 h with goat anti-rabbit antiserum conjugated with horseradish peroxidase (1:1,000,000 in Tween-TBS containing 1%, v/v, fish skin gelatin). The nitrocellulose membrane was subsequently washed 4 ϫ 250 ml ϫ 15 min in TBS with Triton X-100, 0.1%, v/v. Washed membranes were incubated in ECL reagents for 60 s and exposed with Kodak XAR-5 film.
In certain experiments we determined the subcellular distribution of 5-LO among the cytosolic, plasma membrane, and the nuclear soluble and particulate fractions of resting and A23187-stimulated HL-60 cells. Differentiated HL-60 cells were disrupted by nitrogen cavitation. The cytosol, membrane, and nuclei were isolated, and 5-LO was visualized by immunoblotting as described (11).
Phosphorylation of Proteins in Activated Leukocytes-Neutrophils or HL-60 cells (2 ϫ 10 7 cells/ml) incubated with 2.5 M A23187 for 0, 0.5, 1, 2.5, 5, 10, and 20 min were sedimented (12,000 ϫ g, 30 s) on a microcentrifuge. The cell pellet was mixed with 1 ϫ electrophoresis buffer (250 l) and boiled for 5 min. Samples with equal amounts of protein were fractionated by gel electrophoresis; transferred to nitrocellulose membranes and prepared for immunoblot analysis as described above. Proteins on nitrocellulose were equilibrated for 12-16 h at 4°C with mouse monoclonal anti-phosphotyrosine antibody (Oncogene Science, 2 g/ml in TBS, containing 0.05% Tween 20 and 0.1% fish skin gelatin). Following equilibration the nitrocellulose was washed and incubated with goat anti-mouse antibody conjugated with biotin (1:25,000) and avidin-horseradish peroxidase conjugate (1:1,000,000). The membrane was washed, incubated with ECL reagents, and exposed with Kodak XAR-5 film.
HL-60 cells were differentiated into granulocytes in the presence of [ 32 P]orthophosphate. Six flasks, each with 20 ml, 2 ϫ 10 5 cells/ml in phosphate-free RPMI 1640 media, 10%, v/v, fetal bovine serum, 1.25% Me 2 SO, 2 mM L-glutamine, and penicillin/streptomycin were incubated for 12 h in 5% CO 2 . After confirming cell viability at 12 h, [ 32 P]orthophosphate (2.5 Ci/ml) was added to each flask. After 6 days HL-60 cells were isolated by centrifugation at 100 ϫ g, washed once with phosphate-free RPMI medium, 10 mM HEPES, 1 mM CaCl 2 , pH 7.4, and resuspended in this buffer (2 ϫ 10 7 cell/ml) containing [ 32 P]orthophosphate (100 Ci). Cells were stimulated with 2.5 M A23187 or 2 M PMA for 10 min, 37°C, then quenched with 5 mM EDTA. Cells were centrifuged (100 ϫ g) and washed 2 ϫ 2 ml of phosphate free buffer. The cell pellet was suspended in 500 l of lysis buffer (phosphate-buffered saline, 1 mM sodium orthovanadate, 1% Triton X-100, 0.1% sodium deoxycholate, 0.01% SDS, 5 mM EGTA, 5 mM EDTA, 50 mM sodium pyrophosphate, and 5 mM benzamidine), frozen on dry ice, and thawed. After four freeze-thaw cycles, the sample was centrifuged at 12,000 ϫ g on a microcentrifuge for 10 min. In this case, cell lysate supernatant fraction contains both microsomal and cytosolic 5-LO; the cell lysate pellet contains genomic DNA and any nuclear associated proteins. The cell lysate supernatant fraction, containing the microsomal and cytosolic 5-LO, was removed and transferred to a separate microcentrifuge tube containing 15 l of anti-5-LO antiserum (1/67 final dilution). The cell lysate pellet, containing genomic DNA, was washed using minimal shear stress, with 3 ϫ 1 ml with 0.05 M Tris-HCl, pH 7.5, 10 mM MgCl 2 , then suspended forcefully in 500 l of buffer containing DNase I (50 l, 0.2 g/l, 89 units/g). After digestion of DNA for 1 h at 25°C, the lysate pellet was centrifuged at 12,000 ϫ g ϫ 10 min on a microcentrifuge. The supernatant fraction (500 l), containing proteins originally associated with intact DNA, was mixed with 2 ϫ lysis buffer (500 l) and 15 l of anti-5-LO antiserum. The digested DNA fraction (cell lysate pellet) and the soluble microsome/ cytosol fraction (cell lysate supernatant) were incubated for 3 h at 25°C to permit immune complex formation between anti-5-LO antibody and any 5-LO in the samples. Protein A-agarose was then added (40 l, 1:1, w/v, in 1 ϫ lysis buffer); the sample was incubated for 1 h at 25°C and then the resin containing 5-LO⅐anti-5-LO complexes was isolated by centrifugation for 30 s at 12,000 ϫ g on a microcentrifuge. The agarose resin was washed 3 ϫ 1 ml with lysis buffer, suspended in 20 l of 2.5 ϫ electrophoresis buffer, and boiled for 5 min. The entire immunoprecipitate was fractionated by SDS-PAGE on a 10% polyacrylamide gel with a 4% stacking gel. The supernatant fraction from the protein A-agarose immunoprecipitate was also fractionated by SDS-PAGE. Following electrophoresis, gels were fixed for 30 min with 10% acetic acid, 30% ethanol, then in 0.5% glycerol/water prior to drying. Proteins containing 32 PO 4 were detected by exposing gels to Kodak XAR-5 film 10 -20 days at Ϫ80°C. Prior to immunoprecipitation with the anti-5-LO antiserum all samples were "cleared" of proteins which interacted nonspecifically with normal rabbit serum (1/67 dilution) and protein Aagarose. The tracking dye front, which contained a majority of the free 32 P background, was removed from the gel prior to exposure. However, the exposure time necessary to detect the immunoprecipitated 78-kDa 5-LO band still produced substantial background from 32 P-labeled oligomeric DNA fragments in the section of the gels from 39 kDa to the tracking front. For 10 -20 day exposures we were unable to eliminate this background by extensively washing the protein A-agarose immunoprecipitate without jeopardizing the 5-LO/LO-32 binding to the resin.

Effect of Tyrosine Kinase Inhibitors on Cellular 5-Lipoxy-
genase Activity-Six separate inhibitors of tyrosine kinase, including 2,5-DHC, genistein, herbimycin, tyrphostin, lavendustin, and compound 5, each inhibited 5-HETE formation by HL-60 cells stimulated with 1 M A23187 (Fig. 1A). Potencies ranged from 2 to 80 M (Table I). To clarify whether these six agents acted by reducing substrate availability or by inhibiting cellular 5-LO activity, we examined their effect on 5-HETE formation by HL-60 cells stimulated with A23187 in the presence of 20 M arachidonic acid, added exogenously. Under these conditions, where substrate availability is not limited by phospholipase A 2 activity, all six agents still inhibited 5-HETE formation in a concentration-dependent manner. The rank order of inhibitor potency was identical, with or without exogenous arachidonic acid, implying that they inhibited 5-LO (Fig.  1B). Results were similar for human neutrophils (data not shown). Next, we determined if the six agents inhibited only 5-LO catalytic activity, only 5-LO translocation, or both.
Effect of Tyrosine Kinase Inhibitors on Isolated 5-Lipoxygenase-Tyrosine kinase inhibitors inhibited purified 5-LO in a concentration-dependent manner (Fig. 2). The rank order of inhibitor potency differed between cells and purified enzyme. Genistein and herbimycin inhibited the purified enzyme least potently, but they inhibited cellular 5-LO activity more potently. Conversely, lavendustin inhibited purified 5-LO most potently, but it inhibited cellular 5-LO least potently (Table I). This reversal in the rank order of potency between isolated enzyme and cellular enzyme suggests that tyrosine kinase inhibitors affected other cellular processes contributing to 5-LO activity. The redistribution of 5-LO from cytosol to the membrane during cellular activation is notable in this context.

Effect of Tyrosine Kinase Inhibitors on Translocation of 5-LO
from Cytosol to Membrane-5-LO occurs predominantly in the cytosol of leukocytes under "resting" conditions (Fig. 3, lane 1

, panels A-F). Elevation of cytosolic [Ca 2ϩ ] by stimulation with 1
M A23187 promotes its translocation from the cytosol to cell membranes (Fig. 3, lane 2, panels A-F). We verified that MK-886, a novel leukotriene biosynthesis inhibitor, blocks the translocation of 5-LO and its subsequent activation by FLAP (Fig. 3, lane 6,  We verified by immunoblotting that 5-LO in resting HL-60 cells was most abundant (80% of total) in the cytosolic pool; however, 5-LO also occurred in the soluble nuclear protein fraction. This is analogous to results reported for rodent cells (11,17). When HL-60 cells were activated with A23187, 5-LO redistributed from the cytosolic pool to the 100,000 ϫ g nuclear particulate fraction. Enzymatic assays for 5-LO activity supported the immunoblotting experiments. For instance, we detected 5-LO activity in the cytosol and nuclear soluble fraction of resting cells and the 5-LO activity redistributed to the nuclear particulate fraction in A23187-stimulated cells (data not shown).  Phosphorylation of Cellular 5-LO-Immunoblot analysis with anti-phosphotyrosine antibody confirmed that tyrosine phosphoprotein formation increased in HL-60 cells stimulated with A23187, consistent with reports by others (32,33). Tyrosine phosphoproteins at 78 -80 kDa coincide with 5-LO; however, we were unable to detect 32 PO 4 incorporation in 5-LO immunoprecipitated from cell lysates containing the 100,000 ϫ g microsomal and cytosol fractions of HL-60 cells stimulated with A23187 or PMA. Recent data show that 5-LO is abundant in a nuclear pool under some circumstances (9 -11, 17). Consistent with these data, we detected 32 PO 4 incorporation in the 5-LO immunoprecipitated from the 12,000 ϫ g fraction containing genomic DNA and any associated proteins. The phosphorylated 5-LO in the immunoprecipitate of the DNA fraction from HL-60 cells stimulated with A23187 (Fig. 4, lane 2) was more readily detectable than the phosphorylated 5-LO in the DNA fraction of control cells (Fig. 4, lane 1) or cells stimulated with PMA (Fig. 4, lane 3). Among several replicate experiments we could detect phosphorylated 5-LO in the PMA-treated and control cells but it was typically 3-10-fold less abundant compared with cells stimulated with A23187. Thus, 5-LO is phosphorylated under conditions associated with its translocation to the nucleus, and the phosphorylated pool of 5-LO accumulates with genomic DNA, not with membrane or cytosolic proteins. The amount of phosphorylated 5-LO is a small fraction of the total, based on its apparent specific activity.
When purified 5-LO isolated from human leukocytes, or recombinant 5-LO, was incubated with [␥-32 P]ATP, then analyzed by SDS-PAGE and autoradiography, an indistinct band of 32 PO 4 co-migrated with the purified protein at 78 kDa. The signal/noise ratio was always low and barely distinguishable above the background. This indicates that tight association of [␥-32 P]ATP with a putative ATP binding site in 5-LO does not account for the radiolabeled phosphoprotein observed in the 5-LO immunoprecipitate depicted in Fig. 4. Phosphorylation of 5-LO by a cellular kinase best explains the data.
Neither the serine/threonine kinases, MAP and Cdc-2, nor the tyrosine kinase, Lyn, catalyzed phosphorylation of isolated 5-LO in vitro under the conditions described. Autophosphorylation of the kinases occurred, indicating that they were catalytically active. DISCUSSION Compartmentalization and functions of 5-LO depend on processes other than binding to FLAP (8 -11). Modulation of 5-LO translocation and catalysis by tyrosine kinase inhibitors fits a model which postulates that Src homology domain interactions are a molecular determinant of these processes (34). There are two types of Src homology domains (35). The SH2 domain, consisting of approximately 100 amino acid residues, binds to tyrosine residues phosphorylated by tyrosine kinase. Intermolecular binding between SH2 domains and tyrosine phosphate residues initiates the cellular redistribution of proteins. The SH3 domain, consisting of approximately 60 amino acid residues, binds to proline-rich regions of amino acid residues (36). Notably, 5-LO contains a proline-rich motif that interacts with the SH3 domain of certain signaling proteins (34). Signaling proteins, adaptor proteins, or Src kinases usually contain both SH2 and SH3 domains, enabling them to assemble multimeric complexes (35). Our model suggests how cellular tyrosine kinase and Src homology domain interactions could coordinate 5-LO compartmentalization and activity (Fig. 5). For instance, cytosolic 5-LO, via its proline-rich motif, could equilibrate with a protein containing an SH3 domain (34). If this protein also contained an SH2 domain then the 5-LO⅐SH3⅐SH2 complex, via its unoccupied SH2 domain, could equilibrate with tyrosine phosphate residues on proteins distributed in the cytoskeleton, membrane, or nucleus. The order in which these reactions occur and the specific proteins involved requires further investigation.
As depicted in our model, 5-LO, itself, need not be phosphorylated to influence its distribution and activity. However, we demonstrate, directly and for the first time that cells can phosphorylate 5-LO. Our results contradict Rouzer and Kargman who concluded that phosphorylation of 5-LO does not occur (3). We attribute the differences to adequate labeling of cells with carrier-free [ 32 P]orthophosphate and to examination of the nuclear fraction from cellular homogenates. Investigators often neglected the nuclear fraction of cells until experiments showed that 5-LO can translocate and accumulate at the nuclear membrane (8 -11, 17, 37). Although novel, the significance of our observation is uncertain. Phosphorylation of 5-LO may influence its translocation and function. For instance, the nuclear localizing sequence, the proline-rich SH3 binding motif, the FLAP-binding domain, plus the phosphorylation state of 5-LO may determine its partnership with other cellular proteins and whether it has functions other than lipid mediator catalysis. It is notable that phosphorylated 5-LO is most abundant in the nuclear fraction, not the cytosol or membrane fraction of cells activated by A23187; however, this is a small fraction of the total 5-LO. The amount of phosphorylated 5-LO and its specific activity were insufficient to determine if it contains phosphoserine, phosphotyrosine, or phosphothreonine. 5-LO contains several consensus sequences for tyrosine phosphorylation typified by DYI, at residues 20 -22 and residues 94 -96 and EYL at residues 538 -540. 5-LO also contains three tyrosines clustered at residues 660 -662, about 12-15 residues from the C terminus. These sequences are unique to 5-LO, not 12-LO and 15-LO (38). If 5-LO were tyrosine phosphorylated it might interact directly with SH2 domains. 5-LO also contains a consensus sequence, YLSP, at residues 662-665 for phosphorylation by MAP kinase (39) and a consensus sequence, RKSS, at residue 521-524, for phosphorylation by S6 kinase (40). There is also a consensus sequence for Cdc-2 (p34 cdc-2 ) kinase, SPDR at residues 664 -667 in 5-LO. All three mammalian lipoxygenases contain consensus sequences for phosphorylation by protein kinase A (41). These are: RRCT at residues 247-250 in 5-LO, RRST at residues 242-245 in 12-LO, and RRSA at residues 242-245 in 15-LO. However, lavendustin inhibits protein kinase A and C poorly.
To fortify certain conclusions we used a panel of chemically distinct agents which inhibit receptor and Src tyrosine kinase activity by different mechanisms. Herbimycin, a benzoquinone, reacts irreversibly with sulfhydryl groups near the active site on Src kinase (22,23). Lavendustin competitively inhibits both ATP and substrate binding; (26) tyrphostin-25 competitively inhibits only substrate binding; genistein, an isoflavone, competitively inhibits only ATP binding to receptor tyrosine kinases, typified by the epidermal growth factor receptor. 2,5-DHC inhibits substrate binding competitively and ATP binding noncompetitively. These agents can affect cellular 5-LO catalysis, independent of any tyrosine-kinase mediated processes; however, this effect alone would not account for two observa- tions. First, the rank order of potency for inhibition of cellular 5-LO and isolated 5-LO reverses for certain inhibitors, e.g. lavendustin, compound 5, and tyrphostin (Table I). Second, inhibitors of 5-LO catalysis or redox status do not inhibit translocation or interaction with FLAP. The respective effects of 2,5-DHC, herbimycin, tyrphostin, genistein, and lavendustin on 5-LO translocation correspond best with their effects on cellular tyrosine kinase activity. This conforms with precedents showing that tyrosine kinase coordinates the activation and localization of other enzymes involved in lipid mediator biosynthesis and leukocyte activation (42)(43)(44)(45). It is possible that all six tyrosine kinase inhibitors act like MK-886 and block 5-LO-FLAP interactions. However, the divergence among their chemical structures and their relative potencies as inhibitors of translocation, which approximate their potencies for inhibition of cellular tyrosine kinase, make this unlikely. We are presently unable to test this without access to radiolabeled FLAP ligands.
Although the effect of the tyrosine kinase inhibitors on isolated 5-LO may seem "nonspecific" it is predictable and interpretable from two perspectives. First, lavendustin, compound 5, genistein, and 2,5-DHC each inhibit ATP binding to tyrosine kinase. 5-LO, like tyrosine kinase, requires ATP as a co-factor for catalysis (15). The sequence and locale of the ATP binding site on 5-LO are uncertain; however, our results are compatible with a similarity between the ATP binding site on tyrosine kinases and the site on 5-LO. Second, all of these inhibitors are "redox"-sensitive. Certain redox sensitive agents inhibit 5-LO catalysis via effects on the oxidation state of the hemoprotein (46). We stress, again, that inhibitors of 5-LO catalysis or redox status rarely modulate its translocation or interaction with FLAP. Thus, the effect of tyrosine kinase inhibitors on 5-LO translocation was not predictable and is most compatible with a process involving tyrosine kinase activity.
In summary: (i) tyrosine kinase inhibitors uniformly inhibited the activity of both isolated and cellular 5-LO, (ii) tyrosine kinase inhibitors uniformly inhibited the translocation of 5-LO in leukocytes stimulated with A23187, and (iii) 5-LO occurs as a phosphoprotein in the nuclear fraction of activated leukocytes. These results add to the notion that 5-LO may have a novel role within the nucleus (11,17,34).