Synergy of Epidermal Growth Factor and 12(S)-Hydroxyeicosatetraenoate on Protein Kinase C Activation in Lens Epithelial Cells*

12(S)-Hydroxyeicosatetraenoic acid (12(S)HETE) is a bioactive metabolite of arachidonic acid synthesized by 12-lipoxygenase. The 12-lipoxygenase blocker, baicalein, prevents epidermal growth factor (EGF)-induced activation of protein kinase C (PKC) α and β in lens epithelial cells, whereas supplementation with 12(S)HETE reverses this effect, suggesting that EGF and 12(S)HETE may work together to activate PKC. This study investigates the mechanism of PKCβ activation by EGF and 12(S)HETE. 12(S)HETE alone directed translocation of PKCβ through the C1 rather than the C2 domain, without activating phosphoinositide 3-kinase (PI3K) or MAPK signaling or increasing intracellular calcium concentration. In the presence of baicalein, EGF triggered an asymmetric phosphorylation of the EGF receptor initiating signaling through PI3K and MAPK, but not PLCγ. Together, 12(S)HETE and EGF synergistically increased phosphorylation of PKCβ in the activation loop and C terminus as well as PKCβ-specific activity. PI3K inhibitors blocked phosphorylation, but MEK1 inhibitors did not. Microvesicles containing phosphatidylinositol 3,4,5-trisphosphate mimicked the action of EGF on PKCβ activity in the presence of 12(S)HETE. Kinase-inactive PKCβ mutations in either activation loop or C terminus were effectively translocated by 12(S)HETE, as was PKCβ in the presence of chelerythrine or Gö-6983. These findings indicate that unphosphorylated PKCβ is translocated to the membrane by 12(S)HETE and phosphorylated by EGF-dependent PI3K signaling, to generate catalytically competent PKCβ.

12(S)-Hydroxyeicosatetraenoic acid (12(S)HETE) 1 is a bioactive metabolite of arachidonic acid, which evokes a wide variety of cellular responses, ranging from survival and proliferation to invasion and metastasis (1,2). Although 12(S)HETE is synthesized primarily in platelets and leukocytes, a number of other cell types have some capacity to make this hydroxylipid, including the epithelial cells of the lens and cornea (3)(4)(5). Previous studies from this laboratory found that inhibitors of endogenous 12(S)HETE synthesis prevent EGF-dependent DNA synthesis and c-fos mRNA induction in cultured lens epithelial cells. This effect was specifically reversed by exogenous 12(S)HETE, but not by closely related HETEs, suggesting that 12(S)HETE plays an essential role in regulating lens cell proliferation (6,7). Further investigation of the role of 12(S)HETE showed that the selective lipoxygenase inhibitor, baicalein, prevented EGF-induced activation of classic PKC isoforms, PKC␣ and PKC␤ (8), raising the possibility that a cooperative effect of EGF and 12(S)HETE is needed for full activation of PKC in these cells. In addition, because inhibition of the classic PKC isoforms was sufficient to block both c-fos mRNA induction and DNA synthesis, these findings pinpointed PKC as an important target of 12(S)HETE action in regulating lens epithelial cell proliferation (8).
Structural studies of the classic PKC isoforms (PKC␣, ␤, and ␥) have identified several functional domains (9,10). These include an autoinhibitory pseudosubstrate domain at the N terminus, the C1 domain, containing a diacylgycerol binding site (11), the C2 domain, containing binding sites for both anionic lipids and calcium (12), an activation loop adjacent to the active site, and the C-terminal domain. The classic PKC isoforms require diacylglycerol and calcium as well as PtdSer for full activity. In contrast, amino acid replacements at certain key residues in the C2 domain of the novel PKC isoforms (PKC␦, ⑀, , and ) has made these isoforms insensitive to calcium signals, whereas changes in both the C1 and C2 domains of the atypical isoforms (PKC and ) have rendered these isoforms insensitive to both calcium and diacylglycerol (9,10). However, all isoforms require PtdSer or other acidic phospholipids for activity. Activation of PKC involves both phosphorylation of the enzyme at the activation loop and C terminus and translocation to the membrane, where it interacts with its lipid cofactors, PtdSer and diacylglycerol. Phosphorylation at the activation loop seems to be catalyzed by the phospholipiddependent kinase, PDK1 (13)(14)(15), whereas the two phosphorylations in the C terminus appear to be autophosphorylations (16 -18). Phosphorylation of PKC is thought to introduce a conformational change, which allows it to respond to lipid second messengers, such as diacylglycerol (18). Upon binding at the membrane, an additional conformational change removes the autoinhibitory substrate domain from the active site and the enzyme becomes catalytically active (9,10).
Because previous studies of lens epithelial cells had suggested that EGF and 12(S)HETE may cooperate in some way to * 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.

Materials
Site-directed Mutagenesis and Cloning-The preparation of constructs with mutations in either activation loop or C terminus of PKC␤-GFP was done with the QuikChange site-directed mutagenesis kit from Stratagene Corp. (La Jolla, CA). The changed bases are underlined. The oligonucleotides used for introducing the T500A, T641A, and S660A mutations in pEGFP-PKC␤ were as follows (5Ј 3 3Ј): for T500A, GGT-GACAACCAAGGCATTCTGTGGCACTCC ( For introducing the double mutation in the C terminus, mutant T641A was used as a template for synthesis of the S660A mutation. The T500A, T641A, and T641A/S660A mutation sites were confirmed by full sequencing and then prepared in large quantity. For constructing truncated C1 and C2 domains of PKC␤, the two fragments were amplified by PCR with pfu DNA polymerase (Stratagene, La Jolla, CA) using primers that introduced flanking restriction sites for EcoRI and BglII. The primers for amplification of C1 (residues 17-172) and C2 (residues 171-267) domains are as follows (5Ј 3 3Ј): for C1, sense/ATGGGGGC-CCGGTACCTTGGGGCGGGCCAC and antisense/CGGAATTCCCCAC-GCCGCAAAGGGAGGGCACGCTGCGCAC; for C2, sense/GAAGATCT-ATGGGGGCCCTAGAGCGCCGTGGACGTCTGC and antisense/CGG-AATTCTGGTACCTCTCGCCCTCCTCCTGGTTCAGTAACTTG.
The two PCR products were purified and cut with EcoRI and BglII, and the digested fragments were cloned into pEGFP-N3 vector (Clontech, Palo Alto, CA).
Expression of PKC␤-GFP and Its Mutants in N/N1003A Cells-Transient transfection into N/N1003A cells was performed by FuGENE transfection reagent according to the manufacturer's standard protocol (Roche Molecular Biochemicals, Indianapolis, IN). Plasmids encoding wild-type or mutated PKC␤-GFP or GFP-tagged truncated C1 or C2 domains were transfected into 6 ϫ 10 6 cells. After the transfection, cells were cultured at 35°C to obtain the optimal fluorescence of GFP, and experiments were performed 2 days after the transfection.
Establishment of Stable Transfectants of Wild Type and Mutated PKC␤-GFP-N/N1003A cells were transiently transfected with PKC␤ wild-type, T500A, T641A, and T641A/S660A, respectively, by FuGENE transfection reagent (Roche Diagnostics Corp., Indianapolis, IN). Clones were selected based on G418 (400 g/ml) resistance, and GFP fluorescence was checked under a fluorescence microscope (Axioscope, Carl Zeiss, Jena, Germany) to confirm the stability of expression.
Cell Lysis and Subcellular Fractionation-Whole-cell lysates were prepared by lysing cells in phosphate-buffered saline containing 1.0% Triton X-100 (v/v), 1% (w/v) sodium deoxycholic acid, and 1% sodium dodecyl sulfate (w/v). Cells were scraped off the plate, transferred to a microcentrifuge tube on ice, and sonicated in a cold-water bath. Lysates were kept on ice for 15 min and centrifuged at 14,000 ϫ g for 10 min at 4°C. The resultant supernatants were stored at Ϫ80°C for immunoblot analysis. Subcellular fractions for immunoblotting and PKC activity assays were prepared as previously described (8). Immunoprecipitation was performed as previously described (19).
Translocation of Wild-type and Mutated PKC␤-GFP-PKC␤-GFPtransfected cells were spread onto the glass-bottomed chamber (LabTek-II, Ashland, MA) and cultured for at least 36 h. Serum-containing medium was replaced with serum-free DMEM 24 h before experiments. PKC␤-GFP fluorescence was measured by confocal laser scanning microscopy (Leica TCS SP2, Leica Microsystems) using 488-nm argon laser excitation, a 500-nm RSP dichroic filter, and a 500to 550-nm emission spectrum. Reagents were diluted directly into the serum-free medium to obtain the appropriate final concentration, and real-time images were collected to monitor PKC␤-GFP movement. Images were collected at room temperature.
Immunoblot Analysis-Protein concentration was measured by the bicinchoninic acid method (BCA Protein Assay Reagent kit, Pierce, Rockford, IL). Aliquots of fractions containing 20 g of protein were mixed with an equal volume of 2ϫ loading buffer, electrophoresed on 4 -20% SDS-polyacrylamide gels, and transferred to nitrocellulose membranes (0.45-m pore size, Novex, San Diego, CA) for 90 min at 400 mA as described previously (20). After transfer, membranes were blocked for 1 h at room temperature in 5% skim milk (Difco, Detroit, MI) in TBST (15 mM Tris-HCl and 150 mM NaCl, pH 7.5, with 0.05% Tween 20). After blocking, membranes were probed with specific primary antibodies according to the manufacturer's recommendations. Antibodies used were as follows: PKC␤ (1:500) mouse monoclonal (Transduction Laboratories, Lexington, KY), rabbit polyclonal antibody to Erk1/2 and phospho-44/42 Erk1/2 (T202/Y204) (1:1000), and rabbit polyclonal antibody to Akt or phospho-Akt (New England BioLabs, Beverly, MA). The immunoblots were incubated with primary antibodies for 1 h at room temperature on a shaking platform, washed three times with TBST, and incubated with the appropriate horseradish peroxidase-conjugated secondary antibody, either anti-rabbit or anti-mouse IgG, (1:2500, New England BioLabs) for 30 min at room temperature. Specific immunoreactive bands were detected by enhanced chemiluminescence (ECL Plus, Amersham Biosciences, Buckinghamshire, UK). Chemiluminescence was quantified by densitometric scanning of x-ray films with image analysis software (ImageQuaNT Scientific Software, version 5.0, Amersham Biosciences, Piscataway, NJ).
PepTag Assay for Nonradioactive Detection of PKC Activity-The PepTag assay utilizes a brightly colored, fluorescent peptide substrate that is highly specific to PKC (Promega). Phosphorylation by PKC changes the net charge of the substrate from ϩ1 to Ϫ1, thereby allowing the phosphorylated and nonphosphorylated versions of the substrate to be separated on an agarose (0.8%) gel. The phosphorylated species migrates toward the positive electrode, whereas the nonphosphorylated substrate migrates toward the negative electrode. The phosphorylated peptide in the band can then be visualized under UV light. Immunoprecipitated PKC was incubated with PKC reaction mixture (25 l) according to the manufacturer's protocol (Promega) at 30°C for 30 min. The reactions were stopped by placing the tubes in a boiling water bath for 10 min. After adding 80% glycerol (1 l), the samples were loaded onto an agarose gel (0.8% agarose in 50 mM Tris-HCl, pH 8.0). The samples were separated on the agarose gel in the same buffer at 75 V for 25 min, and the bands were visualized under UV light and quantified by ImageQuaNT Scientific Software (version 5.0, Amersham Biosciences).
Confocal Ca 2ϩ Imaging-Intracellular Ca 2؉ was monitored with Fluo-3 AM, a membrane-permeable long wavelength fluorescence indicator. N/N1003 cells were plated in glass-bottomed chambers (LabTek-II Chamber Coverglass, VWR International, Bridgeport, NJ) and cultured in the standard medium for 24 h. The cells were loaded with acetoxymethyl ester of Fluo-3 (5 M) for 30 min; after loading, a period of at least 30 min elapsed before experimentation to allow for deesterification of the intracellularly accumulated Fluo-3 AM. A Leica TCS-SP2 laser scanning confocal microscope (Leica Microsystems, Germany) was used to visualize Ca 2ϩ -mediated fluorescence in the cells. Fluo-3 was excited with the 488-nm line of an argon laser, and Fluo-3 fluorescence was collected between 524 and 540 nm. Scanning was performed every 10 s for a total of 20 min after treatment with 12(S)H-ETE (800 nM) or EGF (15 ng/ml). For quantitative measurements, scanning was performed every 30 s for 40 min, and the ratio of fluorescence intensity to initial fluorescence intensity (F/F 0 ) was calculated at each point. Data were collected on 14 cells, and the results were averaged.

12(S)HETE and EGF Act in Synergy to
Increase PKC␤-specific Enzymatic Activity-To examine the effect of 12(S)HETE and EGF on PKC␤ activation, we immunoprecipitated both endogenous PKC␤ and exogenous GFP-tagged PKC␤ from the membrane fraction of N/N1003 cells and measured its activity using the PKC peptide substrate, PepTag. We also calculated the specific enzymatic activity of PKC␤ by normalizing the relative level of phosphorylated PepTag to the relative amount of PKC␤ in the immunoprecipitate, as determined by immunoblotting. The results showed that EGF (15 ng/ml) in the pres-ence of baicalein (30 M) did not significantly increase either the phosphorylation of the PepTag substrate (Fig. 1A, upper panel) or the amount of PKC␤ in the membrane fraction ( Fig.  1A, lower panel). Treatment with 12(S)HETE (300 nM) increased the amount of PKC␤ in the membrane fraction and increased the phosphorylation of the PepTag substrate somewhat but produced no increase in the specific enzymatic activity of PKC␤ ( Fig. 1, A and B). However, combining 12(S)HETE (300 nM) with EGF (15 ng/ml) significantly enhanced PepTag phosphorylation and produced an ϳ2-fold increase in the specific enzymatic activity of PKC␤ (Fig. 1, A and B). These results confirm that 12(S)HETE and EGF have a synergistic effect on PKC␤ activation. For further studies of this effect, we established a cell line that stably expresses GFP-tagged PKC␤. The GFP tag was fused with the C terminus of PKC␤ ( Fig. 2A), because this fusion protein has previously been shown to retain full catalytic competence (21). Immunoprecipitation of PKC␤-GFP with monoclonal anti-GFP antibody confirmed that its behavior was indistinguishable from that of endogenous PKC␤ with respect to activation in the presence and absence of 12(S)HETE (Fig. 1B, lower panel). 12(S)HETE Promotes Translocation of PKC␤-GFP-Because translocation of PKC to the membrane is often considered sufficient for its activation, we next examined the effect of 12(S)HETE and EGF on PKC␤ translocation, using real-time imaging to follow the movement of PKC␤-GFP in living cells. idly delivered GFP-tagged PKC␤ to the plasma membrane (Fig.  2B). PKC␤-GFP fluorescence began to localize at the plasma membrane about 10 min after 12(S)HETE addition and reached a plateau after about 15 min. In contrast, in the presence of baicalein, EGF (15 ng/ml) had no effect on PKC␤-GFP translocation, whereas 12(S)HETE plus EGF showed a time course of PKC␤ translocation very similar to that seen with 12(S)HETE alone (Fig. 2B).
To quantify the time course of PKC translocation induced by 12(S)HETE and EGF, we calculated a translocation factor (R), defined as the difference between fluorescence intensity in the plasma membrane (I M ) and cytoplasm (I C ) divided by the intensity in the cytoplasm (I C ). Thus, A time series of confocal images was taken, and the value of R was calculated automatically by comparing the changes of the fluorescence intensity in two pre-assigned areas: one in the membrane (representing I M ), another in the cytosol (representing I C ) (Fig. 2C). The results demonstrated that the rate and extent of translocation produced by 12(S)HETE and EGF in combination was indistinguishable from that produced by 12(S)HETE alone (Fig. 2C).
To investigate the mechanism underlying the 12(S)HETEdependent translocation of PKC␤, we tested the effect of 12(S)HETE on GFP-tagged constructs of the truncated C1 and C2 domains of PKC␤. Previous studies have identified these domains as membrane targeting modules that mediate PKC translocation by binding with diacylglycerol and calcium, respectively (11,12,22,23). In control experiments, TPA (1.5 g/ml) induced a translocation of the C1-GFP domain to the membrane, where it remained for more than 20 min (not shown). Similarly, control experiments with ionomycin (0.5 M) induced translocation of GFP-C2 (Fig. 2D). In this case, however, the time course of translocation was very rapid, with membrane-associated fluorescence reaching a peak and returning to baseline within 60 s in most cells (Fig. 2D). Interestingly, ionomycin (0.5 M) also caused translocation of GFP-C1 to the membrane (data not shown). Because previous studies of the isolated C1 domain indicate that it binds to the membranes in a calcium-independent manner (12), the response to ionomycin may reflect a calcium-dependent change in membrane lipids. Incubating cells in 300 nM 12(S)HETE, a concentration that promoted translocation of full-length PKC␤-GFP (30/34 cells imaged) within about 14 min, had a weak effect on C1-GFP membrane translocation (2/15 cells imaged) and no effect on C2-GFP (data not shown). Increasing the concentration of 12(S)HETE to 800 nM strongly directed C1-GFP to the membrane (14/16 cells imaged) but had no effect on C2-GFP translocation (Fig. 2D). These findings suggest that 12(S)HETE promotes the translocation of PKC␤ by affecting either the lipid composition of the membrane or the lipid binding capacity of PKC␤, rather than by mobilizing Ca 2ϩ . In addition, the observation that higher concentrations of 12(S)HETE are needed for translocation of the isolated C1 domain than for the full-length protein suggests that interactions between C1 and other regions of PKC␤, such as the recently reported interaction between C1 and C2 (24), may facilitate membrane binding.
In view of the inability of 12(S)HETE to direct the translocation of the calcium-responsive C2 domain, we next tested whether 12(S)HETE has any effect on cytoplasmic Ca 2ϩ levels. To measure changes in cytoplasmic calcium, we loaded N/N1003A cells with the membrane-permeable calcium indicator, Fluo-3 ester, and monitored the Ca 2ϩ -dependent fluorescence by real-time imaging, sampling every 10 s for 20 min. Application of 12(S)HETE (800 nM) had no effect on intracellular Ca 2ϩ levels (Fig. 2D). Thus, 12(S)HETE seems to mediate translocation of PKC␤ through a preferential effect on the C1 domain, which is not accompanied by changes in intracellular calcium.
EGF Induces PI3K and MAPK Signaling Cascades But Not the PLC-␥ Pathway-The EGF receptor undergoes tyrosine autophosphorylation in its intracellular domain after binding with EGF (25). The phosphorylated tyrosine residues then serve as docking sites for recruitment of different signaling molecules, leading to the activation of signaling through PI3K, MAPK, and PLC-␥ (25). Phosphorylation of tyrosine 1068 leads to activation of PI3K and MAPK through the adaptor proteins FIG. 3. Asymmetric phosphorylation of EGFR initiates PI3K and MAPK, but not PLC-␥, signaling pathways. Phosphorylation of the EGFR intracellular domain at different tyrosine residues was determined with site-specific anti-phosphotyrosine antibodies. Activation of PI3K and MAPK signaling pathways was assayed by monitoring Akt and Erk1/2 phosphorylation, respectively, after treatment with EGF (15 ng/ml). Cytosolic Ca 2ϩ was monitored by confocal microscopy using Fluo-3 AM fluorescence (excitation at 514 nm and emission between 524 and 560 nm) as an indicator for PLC-␥ pathway activation. A, Western blots demonstrating asymmetric phosphorylation of tyrosine residues at 845, 1068, and 992. B, Western blots indicating total and phosphorylated Akt at threonine 308 and serine 473, respectively (left panel), and total as well as phosphorylated Erk1/2 (right panel) after EGF (15 ng/ml) application. Grb2 or Gab1 (26,27), whereas phosphorylation at tyrosine 992 or 1173 activates PLC-␥ (28, 29) (Fig. 3A, upper panel). In addition, EGF receptor phosphorylation by c-Src at tyrosine 845 increases the receptor kinase activity (30). To investigate the signaling pathways activated by EGF in N/N1003A cells, we examined the phosphorylation state of each of these tyrosines using phospho-specific antibodies. The results confirmed that the EGF receptor is expressed (Fig. 3A) and undergoes phosphorylation at tyrosine 845 and 1068 in response to EGF (15 ng/ml). However, no phosphorylation of tyrosine 992 was observed (Fig. 3A). Consistent with these findings, EGF treatment triggered strong phosphorylation of Akt, a kinase downstream of PI3K, and of Erk1/2 (Fig. 3B), indicating that both the PI3K and MAPK signaling cascades are activated.
The lack of autophosphorylation at tyrosine 992 suggested that EGF may fail to activate signaling through PLC-␥ in N/N1003 cells under the present conditions. In this case, EGF would also not produce the increase in cytosolic calcium that is the expected consequence of PLC-␥ activation. To test this possibility, we examined the ability of EGF (15 ng/ml) to deliver FIG. 4. Co-application of 12(S)HETE (300 nM) with EGF (15 ng/ml) has no synergistic effect on PI3K or MAPK signaling. Akt and ErK1/2 phosphorylation were determined as indicators for PI3K and MAPK signaling cascades activation, respectively. LY294002 (10 M) or wortmannin (100 nM) was utilized to inhibit PI3K activity, whereas PD98059 (10 M) was used for MEK inhibition. 12(S)HETE, alone or in combination with EGF, had no effect on Akt phosphorylation (left panel) or Erk1/2 phosphorylation (right panel). Phosphorylation of Akt was abolished by Ly294002 (10 M) or wortmannin (100 nM) and phosphorylation of Erk1/2 by PD98059 (10 M) confirming the involvement of PI3K and MEK, respectively.

FIG. 5. Co-application of 12(S)HETE (300 nM) with EGF (15 ng/ml) synergistically affects PI3K-dependent phosphorylation of endogenous PKC␤ in the activation loop and C terminus.
Endogenously expressed PKC␤ from either membrane-associated fraction (A) or whole cell lysate (B) were immunoprecipitated with rabbit anti-PKC␤. Total PKC␤ was immunoblotted with mouse monoclonal anti-PKC␤. PKC␤ phosphorylation in the activation loop and C terminus was determined with phosphospecific antibodies. Immunoblots were quantified by densitometry and results were expressed as percentage of control. Five separate experiments were performed (mean Ϯ S.E.). A, representative immunoblots showing total PKC␤ and phosphorylation in either activation loop or C terminus from the membrane-associated fractions. Bar graphs of quantitative results show a PI3K-dependent synergistic effect of 12(S)HETE and EGF on PKC␤ phosphorylation in activation loop and C terminus. B, representative immunoblots showing total PKC␤ and phosphorylation in either activation loop or C terminus from whole cell lysate. Bar graphs of quantitative results show a PI3K-dependent synergistic effect of 12(S)HETE and EGF on PKC␤ phosphorylation in activation loop and C terminus. the calcium-responsive GFP-tagged C2 domain of PKC␤ to the membrane. EGF failed to cause translocation of the GFPtagged C2 domain (Fig. 3C, upper panel), although ionomycin (0.5 M) produced translocation within 40 s after EGF was washed out (not shown).
To further define the effect of EGF on the PLC␥ signaling pathway, we monitored changes in free cytosolic Ca 2ϩ level using the membrane-permeable calcium indicator, Fluo-3 ester. Cells were treated with EGF (15 ng/ml), and fluorescence was monitored every 10 s for 20 min. We observed no increase in cytosolic Ca 2ϩ level within this time period (Fig. 3C, lower  panel). In contrast, when EGF was washed out and replaced by ionomycin (0.5 M), Ca 2ϩ levels were elevated within 20 s (Fig.  3C, lower panel). For a more quantitative measure of Ca 2ϩ levels, the ratio of fluorescence intensity to initial fluorescence intensity (F/F 0 ) was measured every 30 s for 40 min. Data were collected from fourteen individual cells, and the results were averaged (Fig. 3D). We detected no increase in F/F 0 , although this ratio increased sharply when EGF was washed out and replaced by ionomycin. Thus, under these experimental conditions, the PLC-␥ pathway appears to be silent after treatment with EGF (15 ng/ml). This may explain the unusual finding that EGF (15 ng/ml) alone is unable to induce PKC␤ translocation ( Fig. 2A).

12(S)HETE Does Not Activate PI3K and MAPK Signaling
Cascades-To test the effect of 12(S)HETE and EGF on the PI3K and MAPK signaling cascades, N/N1003 cells were treated with baicalein, followed by exposure to 12(S)HETE, EGF, or both, for 20 min, and specific antibodies were used to detect the activated forms of Akt and Erk1/2. As expected, EGF treatment led to phosphorylation of both Thr-308 and Ser-473 of Akt, as well as phosphorylation of Erk1/2 (Fig. 4). In contrast, treatment with 12(S)HETE did not lead to phosphoryla-tion of either Akt or Erk1/2. Adding both agents in combination did not increase the extent of phosphorylation above the levels produced by EGF alone. Phosphorylation of Akt and Erk were blocked by the PI3K inhibitor, Ly294002, and the MEK1 inhibitor, PD98059, respectively, confirming that the observed phosphorylations resulted from signaling through PI3K and MEK1. These results imply that EGF is able to initiate signaling through these pathways, whereas 12(S)HETE is not.
Synergistic Effect of EGF and 12(S)HETE on PKC␤ Phosphorylation-Because PKC phosphorylation is regarded as an important factor in regulating PKC activity, we next tested the respective roles of EGF and 12(S)HETE in phosphorylating the activation loop and C terminus of PKC␤, using specific antibodies against the phosphorylated motifs in the activation loop DGVTTKpTFCGTPD and C terminus NFDKFFTRGQPVLpT-PPDQLVIANIDQSDFE. PKC␤ was immunoprecipitated from both membrane fraction and whole cell lysate then analyzed by immunoblotting to determine its phosphorylation state. Unstimulated N/N1003A cells contained very low levels of phosphorylation in the activation loop (T500) and C terminus (T641) in either the membrane fraction or whole cell lysate as determined by immunoblotting with phospho-specific antibodies (Fig. 5, A and B). Incubating baicalein-treated cells with EGF (15 ng/ml) alone for 20 min produced approximately a 70% increase in PKC␤ phosphorylation at both sites, whereas 12(S)HETE (300 nM) alone had no significant effect on PKC␤ phosphorylation under the same conditions (Fig. 5, A and B). In contrast, the combination of EGF and 12(S)HETE had a strong synergistic effect on PKC␤ phosphorylation in both activation loop and C terminus with approximately a 4-to 5-fold increase of PKC␤ phosphorylation (Fig. 5, A and B). To investigate the signaling cascades required for this synergistic effect on PKC phosphorylation, cells were pretreated with PI3K inhibitor

FIG. 6. Synergy effect of 12(S)HETE (300 nM) and EGF (15 ng/ml) on phosphorylation of stably expressed exogenous PKC␤-GFP in both activation loop and C terminus is PI3K-dependent.
A stably transfected line of N/N1003A cells expressing PKC␤-GFP was established by resistance to G418 (400 g/ml). The exogenously expressed PKC␤-GFP fusion protein was immunoprecipitated with mouse monoclonal anti-GFP antibody. Experiments were performed as described in the legend to Fig. 5 and expressed as percentage of control. Results represent mean of five separate experiments, Ϯ S.E. A, representative immunoblots and quantitative results for PKC␤-GFP from the membrane-associated fractions. B, representative immunoblots and quantitative results for PKC␤-GFP from whole lysate. LY294002 (10 M) or MEK inhibitor PD98059 (10 M) for 30 min. Pretreatment with LY294002 (10 M) abolished the phosphorylation of PKC produced by combined treatment with EGF and 12(S)HETE, whereas the MEK inhibitor, PD98059, had no effect (Fig. 5, A and B), suggesting that signaling through PI3K is required.
To confirm that EGF and 12(S)HETE have a synergistic effect on phosphorylation of the PKC␤-GFP fusion protein, as well as the endogenous enzyme, PKC␤-GFP was immunoprecipitated from stably transfected cells using anti-GFP antibody. The phosphorylation state of its activation loop and C terminus were then examined by immunoblotting with phospho-specific antibodies. Once again, the combination of 12(S)H-ETE and EGF had a synergistic effect on phosphorylation of the activation loop and C terminus (Fig. 6, A and B), which was prevented by pretreatment with PI3K blocker, LY294002 (10 M), but not by the MEK inhibitor, PD98059 (Fig. 6, A and B).
Synergy of 12(S)HETE and EGF on PKC␤ Activation Is PI3Kdependent-The above results imply that the combined effect of EGF and 12(S)HETE on PKC␤ activation may be due to EGFdependent phosphorylation of PKC␤ following its translocation to the membrane by 12(S)HETE. As a direct test of this possibility, we measured the effect of PI3K inhibitors on the specific enzymatic activity of PKC␤ after stimulating with 12(S)HETE, EGF, or both agents combined. If phosphorylation in either the activation loop or C terminus is essential to the increase in specific enzymatic activity of PKC␤ produced by these agents in combination, co-treatment with these inhibitors should block this effect. The results indicate that this is the case (Fig. 7). The synergistic effect of 12(S)HETE and EGF on the specific enzymatic activity of PKC␤ was prevented by pretreatment with PI3K blockers, wortmannin (100 nM) or LY294002 (10 M), but not by the MEK inhibitor, PD98059 (10 M) (Fig. 7A). The same result was obtained using PKC␤-GFP immunoprecipitated from stably transfected N/N1003 cells with monoclonal anti-GFP antibody (data not shown). Because 12(S)HETE had no effect on PI3K signaling (Fig. 4), these results imply that the synergistic effect of EGF and 12(S)HETE on PKC␤ activation requires EGF-dependent signaling through PI3K.
One of the principal products of phosphoinositide phosphorylation by PI3K is PtdIns(3,4,5)P 3 (31). This lipid generates a membrane docking site for a variety of pleckstrin homology domain proteins, including PDK1, the kinase implicated in PKC phosphorylation (13)(14)(15). To confirm that the EGF-dependent activation of PI3K is responsible for the synergistic effect of EGF and 12(S)HETE on PKC␤ activation, we loaded N/N1003A cells with PtdIns(3,4,5)P 3 microvesicles and measured the specific enzymatic activity of PKC␤ in the presence or absence of 12(S)HETE. PtdIns(3,4,5)P 3 had no effect on the specific enzymatic activity of PKC␤ when added alone (Fig. 7B). However, in combination with 12(S)HETE, PtdIns(3,4,5)P 3 increased the specific enzymatic activity by about 2-fold (Fig. 7B). Thus, when combined with 12(S)HETE, PtdIns(3,4,5)P 3 is able to enhance the specific activity of PKC␤ to about the same extent as EGF. These findings strengthen the view that, when 12(S)HETE and EGF act together to increase the specific enzymatic activity of PKC␤, the role of EGF is to increase PKC␤ phosphorylation via PI3K signaling.
12(S)HETE Delivers Unphosphorylated PKC␤ to the Membrane-These findings implied that the contribution of 12(S)H-ETE to the synergistic effect with EGF lies in its ability to deliver unphosphorylated PKC␤ to the plasma membrane, where it can be phosphorylated by EGF-dependent activation of the PI3K cascade. To explore this idea in greater detail, we generated stably transfected N/N1003A cell lines expressing PKC␤-GFP containing specific mutations at phosphorylation sites in the activation loop (T500A) and C terminus (T641A and T641A/S660A) (Fig. 8A). As expected, neither the T500A mu- tation nor the T641A/S660A double mutation was phosphorylated in the combined presence of EGF and 12(S)HETE, although the wild-type PKC␤ construct was significantly phosphorylated under these conditions (Fig. 8B). We also found no phosphorylation in the C terminus when the phosphorylatable threonine residue in the activation loop (T500) was replaced with alanine (data not shown), as expected if these sites are autophosphorylated following phosphorylation of the activation loop. In addition, immunoprecipitates of the mutated proteins showed no kinase activity toward the PepTag substrate. This confirms that phosphorylation at these sites is required for activity, as suggested by previous studies (17,32,33). Using real-time fluorescence imaging, we next inquired whether 12(S)HETE directed translocation of the T500A and T641A/S660A mutated proteins to the plasma membrane (Fig.  8D). Both mutated proteins were effectively translocated, with a time course similar to that of the wild-type protein (compare with Fig. 2). Thus, 12(S)HETE is able to direct membrane translocation of inactive, unphosphorylated forms of PKC␤. Finally, we tested whether abolishing PKC␤ activity with the inhibitors chelerythrine or Gö-6983 interfered with the ability of 12(S)HETE to direct PKC␤ to the membrane (Fig. 8E). Although both inhibitors effectively blocked PKC␤ kinase activity toward the PepTag substrate (not shown), they had no effect on the ability of 12(S)HETE to direct translocation to the plasma membrane (Fig. 8E). TPA was also able to direct translocation of PKC␤ in the presence of chelerythrine (Fig. 8E) or Gö-6983 (not shown). Together these findings confirm that 12(S)HETE is able to direct membrane translocation of unphosphorylated, inactive forms of PKC␤ and demonstrate that translocation and phosphorylation are separable events. DISCUSSION The present findings indicate that submicromolar concentrations of 12(S)HETE are sufficient to cause translocation of PKC␤ to the plasma membrane. Moreover, 12(S)HETE also directs translocation of the kinase inactive construct PKC␤(T500A), which can not be phosphorylated in either the transactivation loop or C-terminal domain. Because this construct would be expected to retain the conformation of the unphosphorylated enzyme, as previously shown for PKC␤(T500V) (17,34), conversion to the mature, phosphorylated conformation of the enzyme is apparently not required for 12(S)HETE-dependent translocation. These findings contrast with previous findings indicating that PKC␤ must be fully phosphorylated to respond to lipid cofactors such as diacylglycerol or TPA (18) and suggest that certain lipids can recruit the unphosphorylated enzyme to the membrane for further processing by phosphorylation.
Although the mechanism responsible for 12(S)HETEdependent translocation is unclear, it does not seem to involve calcium mobilization, because 12(S)HETE had no effect on the calcium-responsive C2 domain of PKC␤ and did not produce a detectable change in cytosolic calcium. Indeed, the preferential effect on the lipid-responsive C1 domain of PKC␤, suggests that 12(S)HETE in some way modifies the membrane lipid environment to facilitate PKC␤ binding. Interestingly, a recent report indicates that low concentrations of arachidonic acid (10 -30 nM) have a similar ability to translocate PKC in human polymorphonuclear neutrophils (44). Translocation in response to arachidonic acid, like translocation in response to 12(S)HETE, did not require calcium mobilization.
The present results confirm that EGF is unable to activate PKC␤ in lens epithelial cells if endogenous 12(S)HETE synthesis is blocked and suggest that endogenously synthesized 12(S)HETE may participate in PKC activation in lens epithelial cells by promoting translocation. Interestingly, this may provide a mechanism for activating cPKC isoforms that does not require calcium mobilization. Exactly how endogenously synthesized 12(S)HETE might participate in this process is not yet clear, however. One possibility is that 12(S)HETE may be esterified to phospholipids, then released in response to EGF by the action of cPLA2, which is activated by MAPK (35,36). Alternatively, arachidonic acid might be released by cPLA2 then converted to 12(S)HETE by 12-lipoxygenase. Because our results suggest that EGF does not activate PLC␥ in this cell type, it seems unlikely that formation of a 12(S)HETE-esterified diacylglycerol is involved. Nevertheless, the related hydroxylipid, 15(S)HETE, has been shown to form 15(S)HETEesterified diacylglycerols that specifically activate PKC␣ in human tracheal epithelial cells (37).
Several reports in the literature indicate that there are specific cell surface receptors for 12(S)HETE (38 -40). In support of this view, addition of 12(S)HETE has been shown to activate PKC, PLC␥, MAPK, and PI3K in a variety of cell types (38,40,41). In contrast, we found that 12(S)HETE had no effect on signaling via MAPK, PI3K, or intracellular calcium and was unable to activate PKC␤ without the cooperative action of EGF. The difference between these findings and those reported for other cell types suggests that 12(S)HETE may have multiple modes of action. Indeed, binding curves for 12(S)HETE are complex and provide evidence for both high affinity and low affinity receptors, further supporting the possibility that 12(S)HETE may have various modes of action (39,40). Moreover, the apparent lack of downstream signaling in response to 12(S)HETE in the present study raises the possibility that some of its effects may be receptor-independent and may result from its ability to modify the lipid environment of the membrane.
In most adherent cell types, PKC is highly phosphorylated in both the activation loop and C terminus, even under unstimulated conditions (10,15) or after serum starvation (15). Under these circumstances, cPKC translocation is sufficient for its activation, because membrane binding induces the conformational change needed to release the catalytic core from the autoinhibitory pseudosubstrate (10,12,34). In contrast, serumdeprived lens epithelial cells have very low levels of PKC␤ phosphorylation, making it possible to separate translocation and activation. Our data suggest a model in which serumdeprived lens epithelial cells contain both phosphorylated and unphosphorylated PKC␤, which is distributed between the membrane and cytoplasm, with the bulk of the enzyme in the cytoplasm. Upon addition of exogenous 12(S)HETE, both phosphorylated and unphosphorylated PKC␤ are translocated from the cytoplasm to the membrane (Fig. 9A). The phosphorylated PKC␤ present in this fraction produces an increase in membrane-associated kinase activity. However, because there is no new phosphorylation of PKC␤, there is no increase in the specific enzymatic activity. On the other hand, if endogenous FIG. 9. Schematic of the synergistic effect of 12(S)HETE and EGF on PKC activation. A, when endogenous 12(S)HETE synthesis is blocked, most PKC␤ is unphosphorylated and cytoplasmic. Application of exogenous 12(S)HETE (300 nM) recruits PKC␤ to the cytoplasmic membrane without changing its phosphorylation state. B, under the same experimental conditions, application of EGF (15 ng/ml) triggers an asymmetric autophosphorylation of EGF receptor (EGFR) leading to activation of PI3K and MAPK signaling pathways but not the PLC␥ pathway. This increases phosphorylation of PKC␤ that was already membrane-bound (presumably through the action of PDK-1); however, because most PKC␤ is cytoplasmic, there is little increase in PKC␤ activity. Previous results have shown that cell proliferation does not occur under these conditions, although MAPK signaling is activated (6 -8). C, application of exogenous 12(S)HETE (300 nM) and EGF (15 ng/ml) together leads to 12(S)HETE-dependent translocation of PKC␤ as well as EGF-dependent PI3K activation. The resulting synthesis of PtdIns(3,4,5)P 3 may recruit PDK-1 to the membrane, increasing the probability of interaction between PDK-1 and PKC␤. Under these conditions, a large proportion PKC␤ is activated by phosphorylation and mitogenesis occurs (6 -8).
12(S)HETE synthesis is blocked and EGF is added, activation of PI3K signaling may phosphorylate the small amount of unphosphorylated PKC␤ already associated with the membrane fraction of the serum-deprived cells but is unable to promote translocation (Fig. 9B). Phosphorylation of PKC␤ in the membrane increases both the activity of the membrane fraction and the specific enzymatic activity, but the effect is small, because the amount of enzyme associated with the membrane is small. Previous results have shown that cell proliferation does not occur under these conditions, although MAPK signaling is activated in response to EGF (6 -8). In contrast, when both 12(S)HETE and EGF are present, a much larger amount of unphosphorylated PKC␤ is associated with the membrane, due to the ability of 12(S)HETE to direct the translocation of unphosphorylated, as well as phosphorylated, enzyme (Fig. 9C). The newly translocated, unphosphorylated enzyme is phosphorylated by EGF-dependent PI3K signaling, producing a synergistic effect on membrane-associated activity and specific enzymatic activity. Under these conditions, a large proportion of PKC␤ is activated by phosphorylation and mitogenesis occurs (6 -8). Importantly, unphosphorylated PKC␤ is not phosphorylated in response to EGF unless it is translocated to the membrane. Thus, our data support the view that localization of PKC␤ at the membrane brings it into close proximity with PDK1, the kinase thought to be responsible for its phosphorylation (13,14,42). In a similar manner, membrane localization of Akt (also known as PKB) has been shown to facilitate its phosphorylation by PDK-1 (43). Although some studies have suggested that PDK1 may be constitutively active in cells, even following serum starvation (15), in the present study its activity appears to be regulated by PI3K in response to EGF. Thus, the results of this study support the view that PI3K-dependent activation of PDK1 and colocalization of PDK1 and PKC␤ at the membrane are both important for efficient phosphorylation of PKC␤.