Mechanism of Regulation of Group IVA Phospholipase A2 Activity by Ser727 Phosphorylation*

Although group IVA cytosolic phospholipase A2 (cPLA2α) has been reported to be phosphorylated at multiple Ser residues, the mechanisms by which phosphorylation at different sites regulates cPLA2α activities are not fully understood. To explore the possibility that phosphorylation of Ser727 modulates cellular protein-protein interactions, we measured the effect of Ser727 mutations on the interaction of cPLA2α with a reported cPLA2α-binding protein, p11. In vitro activity assays and membrane binding measurements by surface plasmon resonance analysis showed that a heterotetramer (A2t) of p11 and annexin A2, but not p11 or annexin A2 alone, directly binds cPLA2α via Ser727, which keeps the enzyme from binding the membrane and catalyzing the phospholipid hydrolysis. Phosphorylation of Ser727 disrupts this inhibitory cPLA2α-A2t interaction, thereby activating cPLA2α. Subcellular translocation and activity measurements in HEK293 cells cotransfected with cPLA2α and p11 also showed that p11, in the form of A2t, inhibits cPLA2α by the same mechanism and that phosphorylation of Ser727 activates cPLA2α by interfering with the inhibitory cPLA2α-A2t interaction. Collectively, these studies provide new insight into the regulatory mechanism of cPLA2α through Ser727 phosphorylation.

Although group IVA cytosolic phospholipase A 2 (cPLA 2 ␣) has been reported to be phosphorylated at multiple Ser residues, the mechanisms by which phosphorylation at different sites regulates cPLA 2 ␣ activities are not fully understood. To explore the possibility that phosphorylation of Ser 727 modulates cellular protein-protein interactions, we measured the effect of Ser 727 mutations on the interaction of cPLA 2 ␣ with a reported cPLA 2 ␣-binding protein, p11. In vitro activity assays and membrane binding measurements by surface plasmon resonance analysis showed that a heterotetramer (A2t) of p11 and annexin A2, but not p11 or annexin A2 alone, directly binds cPLA 2 ␣ via Ser 727 , which keeps the enzyme from binding the membrane and catalyzing the phospholipid hydrolysis. Phosphorylation of Ser 727 disrupts this inhibitory cPLA 2 ␣-A2t interaction, thereby activating cPLA 2 ␣. Subcellular translocation and activity measurements in HEK293 cells cotransfected with cPLA 2 ␣ and p11 also showed that p11, in the form of A2t, inhibits cPLA 2 ␣ by the same mechanism and that phosphorylation of Ser 727 activates cPLA 2 ␣ by interfering with the inhibitory cPLA 2 ␣-A2t interaction. Collectively, these studies provide new insight into the regulatory mechanism of cPLA 2 ␣ through Ser 727 phosphorylation.
Group IVA cytosolic PLA 2 (cPLA 2 ␣) is a ubiquitous enzyme whose transcript is expressed at a fairly constant level in all human tissues (2). cPLA 2 ␣ is the only known PLA 2 with pronounced specificity for sn-2 arachidonoyl-containing phospholipids (5,6) and is generally thought to play a crucial role in maintaining cellular AA levels (2). Gene knock-out studies firmly established cPLA 2 ␣ as a key proinflammatory enzyme (7,8). cPLA 2 ␣ is composed of the NH 2 -terminal C2 domain and the COOH-terminal catalytic domain. The regulatory C2 domain contains Ca 2ϩ -and membrane-binding sites (9), whereas the catalytic domain of cPLA 2 ␣ has an active site serine and shows lysophospholipase and transacylase activities as well as PLA 2 activity (10).
Cellular cPLA 2 ␣ activities are tightly regulated by different factors, including Ca 2ϩ , phosphorylation (2,11,12), lipid mediators (13)(14)(15)(16)(17), and other cellular proteins (18 -22). It has been well established that Ca 2ϩ binds to the NH 2 -terminal C2 domain (9) and drives its membrane binding in vitro (23)(24)(25) and in the cell (26,27). Recently, it has been reported that in vitro and cellular membrane affinity and enzyme activity of cPLA 2 ␣ are regulated by phosphatidylinositol 4,5-bisphosphate (13,14) and ceramide 1-phosphate (15)(16)(17). Phosphatidylinositol 4,5-bisphosphate binds to the groove between the C2 and the catalytic domain and might activate cPLA 2 ␣ by inducing conformational changes of the enzyme and thereby juxtaposing its active site and the membrane surface (14). On the other hand, ceramide 1-phosphate binds to the cationic ␤-groove of the C2 domain (17) and thereby enhances the membrane affinity of cPLA 2 ␣ in vitro and may induce the targeting of cPLA 2 ␣ to the Golgi membrane in the cell.
Phosphorylation of different serines of cPLA 2 ␣ has also been shown to activate cPLA 2 ␣; however, the physiological significance of multisite phosphorylation of cPLA 2 ␣ and the mechanism by which the phosphorylation activates cPLA 2 ␣ have not been fully elucidated. It was initially found that the phosphorylation of Ser 505 by mitogen-activated protein kinases is essential for agonist-induced AA release in Chinese hamster ovary cells (28). The phosphorylation analysis of cPLA 2 ␣ expressed in baculovirus-infected insect Sf9 cells showed that the protein is phosphorylated on multiple Ser residues, including Ser 454 , Ser 437 , Ser 505 , and Ser 727 (29). It was reported that among these sites, Ser 454 , Ser 437 , and Ser 505 are constitutively phosphorylated in unstimulated Sf9 cells but that only Ser 505 phosphorylation partially contributes to cPLA 2 ␣ activation (27). Subsequent studies with human platelets, HeLa, and human embryonic kidney 293 (HEK293) cells indicated that Ser 505 and Ser 727 are phosphorylated in response to agonists and that phosphorylation of both sites plays an important role in agonist-induced AA release (30,31). More recently, it was reported that phosphorylation of Ser 515 by calcium-/calmodulindependent protein kinase II in vascular smooth muscle cells increased the enzymatic activity of cPLA 2 ␣ (32).
Among these reported phosphorylation sites, phosphorylations of Ser 505 and Ser 727 have been shown to enhance the cellular AA-releasing activity of cPLA 2 ␣ in different mammalian cells. A recent study indicated that phosphorylation of Ser 505 enhanced the activity of cPLA 2 ␣ by greatly increasing the membrane affinity of cPLA 2 ␣, particularly at submicromolar Ca 2ϩ , both in vitro and in HEK293 cells (33). Another study in Madin-Darby canine kidney cells suggested, however, that phosphorylation of Ser 505 might not have a direct effect on the membrane affinity (34). Interestingly, phosphorylation of Ser 727 also increased the cellular activity of cPLA 2 ␣ but had little effect on in vitro membrane affinity and enzyme activity of cPLA 2 ␣, suggesting that phosphorylation of Ser 727 might modulate the interaction of cPLA 2 ␣ with a (or more) cellular protein that affects the cellular cPLA 2 ␣ activity (33).
Various cellular proteins, including p11, annexin I, vimentin, PLIP, and NADPH oxidase, have been reported to bind cPLA 2 ␣ and regulate its cellular localization and activity (18 -22). However, the physiological relevance and significance of these findings remain controversial. Among these proteins, p11 has been most extensively studied in terms of its inhibitory interaction with cPLA 2 ␣. p11, also known as S-100A10 or calpactin I light chain, is a member of the S-100 family that is a multigenic family of low molecular mass (9 -11 kDa) calcium-binding proteins (35). p11 is unique among S-100 proteins, because its two EF-hands carry mutations that limit its ability to bind calcium (36,37). p11 is a natural ligand of annexin A2, and they form a heterotetramer (A2t) containing two annexin A2 and p11 molecules, respectively (38). It was reported that p11 interacts with the COOH-terminal region of cPLA 2 ␣ and inhibits its activity, resulting in reduced AA release (18). Antisense inhibition of p11 mRNA resulted in enhanced cPLA 2 ␣ activity and increased AA release, whereas p11 overexpression reduced cPLA 2 ␣ activity and AA release (39). Dexamethasone, a glucocorticoid, reduced cPLA 2 ␣ activity by up-regulation of p11 in human epithelial cells (39). It was also reported that transforming growth factor-␣ inhibited A23187-induced AA release by increasing p11-cPLA 2 ␣ binding (40,41). Collectively, these data suggest that p11 may play a role in regulating cPLA 2 ␣ activity and AA release.
In this study, we explored the possibility that phosphorylation of Ser 727 regulates the cellular cPLA 2 ␣ activity by modulating its interaction with p11 on the basis of the report that p11 interacts with the COOH-terminal region of cPLA 2 ␣ where Ser 727 is located. Extensive in vitro and cellular membrane binding and activity measurements using cPLA 2 ␣ wild type and mutants provide evidence that phosphorylation of Ser 727 regu-lates the cellular cPLA 2 ␣ activity by modulating its interaction with A2t but not p11 alone, which binds to the hydroxyl group of Ser 727 and thereby interferes with its membrane binding. Expression Vector Construction and Mutagenesis-Baculovirus transfer vectors encoding the cDNAs of cPLA 2 ␣ mutations were generated by PCR using the pVL1393-cPLA 2 ␣ plasmid as a template. Mutations were verified by DNA sequencing. Mammalian expression vectors of cPLA 2 ␣ and mutants with carboxyl-terminal enhanced yellow fluorescence protein (EYFP) tags were generated by subcloning the gene into the pEYFP C1 vector (Invitrogen). cDNA of p11 was subcloned into pFLAG-CMV TM -5.1 (Sigma) or pECFP C1 vector (Invitrogen) by PCR.

Materials-1,2-Di-O-hexadecyl-sn-glycerol-3-phosphocho-
Expression and Purification of Recombinant cPLA 2 ␣ and A2t-Baculovirus transfer vectors pVL1393 encoding the cDNAs of cPLA 2 ␣ and mutations, respectively, were used to infect Sf9 cells using a BD BaculoGold TM transfection kit (BD Biosciences Pharmingen). For protein expression, Sf9 cells were grown in a 400 ml of suspension culture with a density of 2 ϫ 10 6 cells/ml and infected by high titer recombinant baculovirus at a multiplicity of infection of 3. The cells were then grown at 27°C for 3 days. To boost the phosphorylation of Ser 727 of cPLA 2 ␣, some Sf9 cells were treated with 1 mM okadaic acid for 12 h prior to the harvest. For protein purification, cells were collected by centrifugation at 1000 ϫ g for 10 min, and the cell pellet was resuspended with 40 ml of the lysis buffer (20 mM Tris-HCl, pH 7.5, containing 0.1 M KCl, 1 mM EDTA, 2 mM dithiothreitol, 1% (v/v) Triton X-100, 0.5 mM phenylmethylsulfonyl fluoride, and one tablet of a protease inhibitor mixture (Roche Applied Science)). The suspension was homogenized on ice for 15 min and centrifuged at 100,000 ϫ g for 1 h at 4°C. To the supernatant, 1 ml of nickel-nitrilotriacetic acid-agarose (Qiagen) was added, and the mixture was incubated on ice while shaking at 80 rpm for 2 h. The mixture was poured into a 10-ml empty column, and the resin was washed first with 10 ml of 50 mM sodium phosphate, pH 8.0, containing 300 mM NaCl and 10 mM imidazole, and subsequently with 10 ml of 50 mM sodium phosphate, pH 8.0, containing 300 mM NaCl and 15 mM imidazole. The column was then washed with 6 ml of the same buffer containing 20 mM imidazole and next with 3 ml of the same buffer containing 25 mM imidazole. The protein was then eluted from the column with six 0.75-ml fractions of 50 mM sodium phosphate, pH 8.0, containing 300 mM NaCl and 300 mM imidazole. cPLA 2 ␣ fractions were concentrated and desalted in an Ultrafree-15 centrifugal filter device (Millipore).
Annexin A2 and p11 were bacterially expressed and equimolecularly mixed to form the A2t complex as described previously (42). p11 was also expressed as NH 2 -terminal glutathione S-transferase (GST)-tagged protein for GST pull-down experiments. Protein concentration was determined by the bicinchoninic acid method (Pierce).
In Vitro cPLA 2 ␣ Activity Assay-The enzymatic activity of cPLA 2 ␣ was assayed by measuring the initial rate of [ 14 C]SAPC hydrolysis as described (43). Assay mixtures (100 l) contained 20 mM HEPES buffer, pH 7.4, 0.16 M KCl, 50 M Ca 2ϩ , and 16 M BSA. Free Ca 2ϩ concentration was adjusted using a mixture of EGTA and CaCl 2 according to the method of Bers (44). Reactions were started by first adding cPLA 2 ␣ to the mixture (to a final concentration of ϳ30 nM) and then adding small unilamellar vesicles of [ 14 C]SAPC (typically 2.5 M final concentration). The reactions were quenched by adding 760 l of hexane/isopropanol/H 2 SO 4 (2:1.8:0.01, v/v/v) after a given period of incubation (15 min) at 37°C. For inhibition measurements, 30 nM to 1.5 M of annexin A2, p11, or A2t was preincubated with cPLA 2 ␣ for 5 min before adding the substrate. Liberated [ 14 C]AA was separated from the reaction mixture on small silica gel columns, and the radioactivity was measured by liquid scintillation counting.
Surface Plasmon Resonance (SPR) Analysis-Membrane binding of cPLA 2 ␣ was measured by SPR analysis at 23°C using a lipid-coated L1 chip in the BIACORE X system as described previously (45). The control surface was coated with 4000 resonance units of BSA and a synthetic cationic lipid, 1-palmitoyl-2-oleoyl-sn-glycero-3-ethylphosphocholine (Avanti). The active surface was coated with extruded large unilamellar vesicles of DHPC at a flow rate of 5 l/min until resonance units reached 4000. After lipid coating, 30 l of 50 mM NaOH was injected at 100 l/min to remove loosely bound vesicles. Typically, no further decrease in SPR signal was observed after one wash cycle. To probe for any uncovered surface, the flow cell was subjected to a 5-min injection of 0.1 mg/ml BSA. After coating, the drift in signal was allowed to stabilize below 0.3 resonance units/min before any kinetic experiments were done. All ensuing measurements were performed at a flow rate of 60 l/min. Ninety microliters of protein sample was injected for the association time of 90 s, whereas the dissociation was monitored for 4 min in the running buffer. For inhibition studies, cPLA 2 ␣ was preincubated with of annexin A2, p11, or A2t for 5 min before injection. After each protein injection, the lipid surface was regenerated with a 10-l pulse of 50 mM NaOH. The regeneration was repeated until the SPR signal reached the initial background value before the next protein injection. After each set of measurements, the entire sensor surface was removed with a 5-min injection of 40 mM CHAPS followed by a 5-min injection of 40 mM octyl-␤-D-glucopyranoside at 5 l/min, and the sensor chip was recoated for the next set of experiments. For each trial, the signal was corrected against the control surface in order to eliminate any refractive index changes due to buffer change.
Immunoprecipitation and Immunoblotting-Confluent HEK293 cells co-transfected with cPLA 2 ␣ (wild type or mutants) and FLAG-tagged p11 and cultured in a 10-cm 2 plate were washed three times with ice-cold phosphate-buffered saline (PBS), pH 7.4, lysed in cold lysis buffer (50 mM HEPES, pH 7.4, containing 100 mM KCl, 2 mM EGTA, 10 mM NaF, 1% Triton X-100, and a protease inhibitor mixture (one tablet per 10 ml of buffer)), sonicated for 10 s, and centrifuged for 30 min. The lysates of HEK293 cells were incubated with 10 l of anti-FLAG M2-agarose for 4 h, and the proteins bound to the beads were separated by SDS-PAGE under reducing conditions and transferred onto a nitrocellulose membrane for immunoblotting. Blots were treated with either 250 ng/ml mouse monoclonal anti-cPLA 2 ␣ antibody (Santa Cruz Biotechnology, Inc., Santa Cruz, CA) or 250 ng/ml mouse monoclonal anti-annexin A2 antibody (BD Transduction Laboratories), followed by the incubation with horseradish peroxidase-conjugated antimouse IgG antibody (0.01 ng/ml; Santa Cruz Biotechnology), and visualized with the ECL luminescence system (Amersham Biosciences).
GST Pull-down Assay-The recombinant GST-tagged A2t was prepared from annexin A2 and GST-tagged p11 and was functionally identical to the A2t prepared from the intact p11. 0.1 nmol of GST-A2t or GST was incubated with 1 g of cPLA 2 ␣ in PBS at 4°C for 3 h, and the protein mixture was mixed with the glutathione-Sepharose 4B resin and incubated for another 1 h. The glutathione beads with the bound proteins were washed three times with PBS containing 0.1% Triton X-100. The washed beads were then mixed with 5 times concentrated gel loading solution and were subjected to SDS-PAGE and immnunoblotting using the anti-GST antibody (1:1000 dilution; Santa Cruz Biotechnology) and the anti-cPLA 2 ␣ antibody.
Preparation of HEK293 Cells Stably Expressing BLT1 (HEK293-BLT1)-The stable HEK293 cell line expressing the LTB 4 receptor (BLT1) was generated using Lipofectamine TM 2000 (Invitrogen) according to the manufacturer's protocol. Two days after transfection, cells were treated with 1 mg/ml Geneticin. After culture for 3-4 weeks, individual clones were tested by Western blotting. The established clones were expanded and used for further experiments.
Cell Culture, Transient Transfection, and Protein Production-HEK293 EcR or HEK293-BLT1 cells were used for all cell studies. Cells were cultured in Dulbecco's modified Eagle's medium (DMEM) supplemented with 10% fetal bovine serum at 37°C in 5% CO 2 and 98% humidity. Cells were transfected with plasmids harboring cPLA 2 ␣ and p11 by using Lipofectamine TM reagent according to the manufacturer's instructions.
Microscopy-Subcellular localization of cPLA 2 ␣ was measured using a custom-built combination laser-scanning and multiphoton microscope, the configuration of which was described in detail previously (2). All microscopy experiments were carried out at the same laser power that was found to induce minimal photobleaching over 30 scans and with the same gains and offset setting on the photomultiplier tubes. Transfected cells were washed twice with the PBS and then overlaid with 400 l of HEPES-Hanks' balanced salt solution (ϩ) buffer containing 1.5 mM CaCl 2 . After initially imaging cells, 40 l of the above buffer containing 2 M ionomycin (HEK293) or 3 M LTB 4 (HEK293-BLT1) was added to cells, and the cPLA 2 ␣ translocation was monitored every 30 s. All microscopic manipulation and data acquisition were controlled by a program, SimFCS, kindly provided by Dr. Enrico Gratton.
Determination of [Ca 2ϩ ] i -The [Ca 2ϩ ] i was determined using Fluo-4 as an indicator. HEK293-BLT1 cells (10 7 cells/ml) were labeled with 2 M Fluo-4 for 10 min in Hanks' balanced salt solution containing 1.2 mM Ca 2ϩ and 1% BSA, and cells were incubated at 37°C with 5% CO 2 , as described previously (46). After washing once with Hanks' balanced salt solution containing 1.2 mM Ca 2ϩ , 3 M LTB 4 was added, and the fluorescence intensity of Fluo-4 was monitored with a 488-nm argon/krypton laser and a 530-nm long pass filter. [Ca 2ϩ ] i was calibrated as described previously using the reported calcium dissociation constant value (i.e. 345 nM) of Fluo-4 (47).
AA Release from HEK293 Cells-HEK293 cells transiently transfected with cPLA 2 ␣ or both cPLA 2 ␣ and p11 were incubated with 0.1 Ci/ml [ 3 H]AA (Amersham Biosciences) for 20 h at 37°C. Unincorporated [ 3 H]AA was removed by washing the cells three times with DMEM containing 0.2% BSA. Cells were then stimulated with 1 M ionomycin for 10 min or 3 M LTB 4 for 20 min in 0.5 ml of the same buffer solution containing 1% BSA. In some experiments, cells were pretreated for 15 min with 1 M phorbol 12-myristate 13-acetate (PMA; Sigma) prior to stimulation with 10 M ionomycin for 30 min. After cell activation, the supernatant was collected and briefly centrifuged to remove the cell pellet, and a 250-l aliquot was subjected to scintillation counting. Also, the cells were removed from the plate by applying 200 l of trypsin-EDTA and subjected to scintillation counting to determine total radioactivity of labeled cells. The AA release into the medium was then expressed in terms of the percentage of total counts of the cells. Data are expressed as mean Ϯ S.E. for each group. Individual statistical comparisons of paired data were assessed by Student's t test, and p Ͻ 0.05 was considered to be statistically significant.
Phosphorylation of cPLA 2 ␣ by PMA and Ionomycin-Two days after transfection, HEK293 cells were incubated with serum-free DMEM for 30 min and treated with 1 M PMA for 30 min, followed by 1 M ionomycin for 10 min. Cells were washed with ice-cold PBS two times, harvested, and subjected to immunoprecipitation with anti-GFP antibody (Anaspec). Precipitated cPLA 2 ␣ was subjected to SDS-PAGE and immunoblotting. Phosphorylation of cPLA 2 ␣ was accessed by blotting with the anti-Ser(P) antibody (Zymed Laboratories Inc.).
Identification of Ser 727 Phosphorylation by Mass Spectrometry (MS) Analysis-The regions corresponding to cPLA 2 ␣ bands were excised from SDS-polyacrylamide gels. The proteins in gels were reduced and alkylated and then digested with a 12.5 ng/l concentration of the sequencing grade trypsin (Promega). The digested peptides were extracted with 5% formic acid in 50% acetonitrile solution at room temperature for 20 min and desalted using C18 ZipTips (Millipore) before MS analysis. The resulting tryptic peptides were loaded onto a fused silica microcapillary column (15 cm ϫ 75 m) packed with the C18 resin (5 m, 300 Å; Alltech) and separated with a two-step linear gradient of Buffer B (3-40% in 80 min and 40 -90% in 13 min) at a flow rate of 200 nl/min. Buffer A was 0.1% formic acid in H 2 O, whereas Buffer B contained 0.1% formic acid in acetonitrile. Peptides eluting from the column were ionized by electrospray ionization and analyzed by an linear trap quadrupole ion trap mass spectrometer (Finnigan). The electrospray voltage was set at 2.1 kV, and the threshold for switching from MS to MS/MS was 250. The normalized collision energy for MS/MS was 35% of main RF amplitude, and the duration of activation was 30 ms. All spectra were acquired in a data-dependent mode. Each full MS scan was followed by nine MS/MS scans covering from the most intense peak to the ninth most intense peak of the full MS scan. The repeat count of peak for dynamic exclusion was 1, and its repeat duration was 30 s. The dynamic exclusion duration was set for 180 s, and exclusion mass width was Ϯ1.5 Da. The list size of dynamic exclusion was 50. For the data base search, all MS/MS spectra recorded were matched with cPLA 2 ␣ sequence using the SEQUEST algorithm (ThermoFinnigan). Precursor and fragment ion mass tolerances were 1.5 and 1 Da, respectively. Oxidation on Met (ϩ16 Da), carboxyamidomethylation on Cys (ϩ57 Da), and phosphorylation on Ser, Tyr, and Thr (ϩ80 Da) were selected as variable modifications (48).

RESULTS
A2t Inhibits cPLA 2 ␣ by Interacting with Ser 727 -It has been reported that p11 inhibits cPLA 2 ␣ in human epithelial cells by binding to its COOH-terminal region (18,39,49). However, direct inhibition of cPLA 2 ␣ by p11 has not been demonstrated in a cell-free assay system that includes pure recombinant proteins and lipid substrates. We therefore measured the inhibition of recombinant cPLA 2 ␣ by p11, annexin A2, and A2t, respectively, in an in vitro cPLA 2 ␣ activity assay (Fig. 1A). Annexin A2 and A2t were included, because p11 is known to spontaneously form a heterotetramer (A2t) with annexin A2 in mammalian cells (50). When the activity of cPLA 2 ␣ (30 nM) toward [ 14 C]SAPC was measured in the presence of p11, less than 20% inhibition was observed with up to 1 M p11. Similarly, annexin A2 caused a small degree of inhibition. In contrast, A2t showed much more potent inhibition; cPLA 2 ␣ activity was inhibited up to 50% with IC 50 of ϳ100 nM. To rule out the possibility that inhibition of cPLA 2 ␣ activity by A2t is due to a nonspecific substrate depletion effect (i.e. A2t covers the membrane surface), the cPLA 2 ␣ activity assay was performed with the substrate concentration in the range of 5-50 M. As shown in Fig. 1B, A2t (150 nM) was able to inhibit cPLA 2 ␣ (30 nM) activity to essentially the same degree regardless of the substrate concentration. This indicates that A2t, but not p11 alone, directly inhibits cPLA 2 ␣. This also suggests that the reported cellular cPLA 2 ␣-inhibitory of p11 is mediated through the formation of A2t.
The reports that p11 interacts with the COOH-terminal region of cPLA 2 ␣ (18,39,49) where Ser 727 is located suggest that cPLA 2 ␣ inhibition by A2t might be modulated by the phosphorylation-dephosphorylation cycle of Ser 727 . To explore the possibility that Ser 727 phosphorylation activates cPLA 2 ␣ by disrupting the inhibitory cPLA 2 ␣-A2t interaction and thereby allowing cPLA 2 ␣ to interact with its substrate in the membrane, we examined the A2t inhibition of the activity of cPLA 2 ␣ and its phosphorylation-mimicking (i.e. S727E) and dephosphorylation-mimicking (i.e. S727A) mutants (Fig. 1C). As reported previously, the MS analysis showed that phosphorylation of Ser 727 cPLA 2 ␣ Ser 727 Phosphorylation FEBRUARY 15, 2008 • VOLUME 283 • NUMBER 7 was not significant in cPLA 2 ␣ expressed in Sf9 cells under our conditions (see Fig. 2 and Table 1). It was thus expected that S727A would behave like cPLA 2 ␣ wild type. Surprisingly, however, S727A was not inhibited by up to 1 M A2t, whereas S727E was inhibited by 40% with IC 50 of ϳ100 nM. This apparently paradoxical finding suggests that the cPLA 2 ␣-A2t interaction is mediated by the hydrogen bonds via the free -OH group of Ser 727 . According to this mechanism, S727E would better simulate cPLA 2 ␣ with nonphosphorylated Ser 727 than S727A, because Ala in S727A cannot substitute for Ser, whereas Glu in S727E may be able to form hydrogen bonds to some degree. The mechanism also suggests that the primary effect of Ser 727 phosphorylation is to abrogate the cPLA 2 ␣-A2t hydrogen bonds.
To test this notion, we prepared and characterized cPLA 2 ␣ from Sf9 cells treated with 1 mM okadaic acid for 12 h prior to the cell harvest. This was based on the report that the treatment of Sf9 cells expressing cPLA 2 ␣ with okadaic acid significantly enhances the degree of Ser 727 phosphorylation (51). The MS analysis confirmed that cPLA 2 ␣ from okadaic acid-treated Sf9 cells has a much higher (i.e. Ͼ4 times) degree of phosphorylation at Ser 727 (see Fig. 2 and Table 1) than cPLA 2 ␣ from untreated Sf9 cells. Thus, the preparation of cPLA 2 ␣ from okadaic acid-treated Sf9 cells allowed us to directly measure the effect of Ser 727 phosphorylation on A2t inhibition of cPLA 2 ␣. As shown in Fig.  1C, although cPLA 2 ␣ from okadaic acid-treated Sf9 cells and untreated Sf9 cells had similar enzyme activity, the former was inhibited by A2t to a much lesser degree than the latter.
We also characterized the S727T mutant of cPLA 2 ␣, in which the hydroxyl group of Thr should be able to form hydrogen bonds with A2t as well as that of Ser in cPLA 2 ␣. Thr 727 in S727T should not be phosphorylated to a significant degree under our experimental conditions, just as Ser 727 in the wild type was not. Fig. 1C shows that S727T was as potently inhibited by A2t as the wild type cPLA 2 ␣, corroborating the notion that A2t interacts with the hydroxyl group of Ser 727 , presumably through the hydrogen bonds. Again, all cPLA 2 ␣ mutants, including S727T, showed essentially the same activity as the wild type, showing that Ser 727 is not directly involved in either membrane binding or catalytic steps.
It has been reported that Ser 505 phosphorylation plays a key role in cPLA 2 ␣ and that it often accompanies agonist-induced Ser 727 phosphorylation. To show that our observed results on cPLA 2 ␣-A2t interaction and its putative disruption by Ser 727 phosphorylation are independent of Ser 505 phosphorylation, we measured the inhibition of S505E and S505E/S727A by A2t (Fig. 1D). As reported previously, S505E that is a Ser 505 -phosphorylated mimetic of cPLA 2 ␣ was only slightly more active than cPLA 2 ␣ wild type, because Ser 505 is mostly constitutively phosphorylated in HEK293 cells overexpressing cPLA 2 ␣ (33). Also, S505E/S727A was as active as S505E. Importantly, S525E and S525E/S727A, respectively, behaved like cPLA 2 ␣ wild type and S727A, respectively, in terms of A2t inhibition. These results indicate that A2t inhibition of cPLA 2 ␣ is independent of Ser 505 phosphorylation.
A2t Interferes with the Membrane Binding of cPLA 2 ␣-To further investigate how A2t inhibits cPLA 2 ␣ and how Ser 727 phosphorylation relieves the inhibition, we first measured by SPR analysis the effects of A2t, annexin A2, and p11, respec- tively, on binding of cPLA 2 ␣ to the sensor chip coated with the vesicles made of a nonhydrolyzable synthetic phospholipid, DHPC. Preincubation of cPLA 2 ␣ with A2t greatly reduced the binding of cPLA 2 ␣ to DHPC vesicles at 20 M Ca 2ϩ (Fig. 3A), whereas neither annexin A2 or p11 had a significant effect on the membrane binding of cPLA 2 ␣ under the same conditions (data not shown). Control experiments showed that A2t, annexin A2, and p11 did not bind the DHPC-coated sensor chip (data not shown), precluding the possibility that the inhibition of membrane binding of cPLA 2 ␣ by A2t is due to a nonspecific lipid depletion effect.
We then measured the effect of A2t on membrane binding of cPLA 2 ␣ mutants and the cPLA 2 ␣ wild type purified from okadaic acid-treated Sf9 cells. When compared with cPLA 2 ␣ wild type purified from untreated Sf9 cells, these proteins showed comparable binding to DHPC vesicles. This again shows that Ser 727 is not directly involved in membrane binding, which is consistent with the location of Ser 727 remote from the membrane binding surface of cPLA 2 ␣. In agreement with our activity assay data, A2t significantly inhibited the membrane binding of S727T (Fig. 3B) and S727E ( Fig. 3C) but had much smaller inhibitory effects on S727A (Fig. 3D) and the cPLA 2 ␣ purified from okadaic acid-treated Sf9 cells (Fig. 3A). Therefore, these results indicate that binding of A2t to cPLA 2 ␣ via Ser 727 interferes with the membrane binding of cPLA 2 ␣, resulting in cPLA 2 ␣ inhibition. Ser 727 phosphorylation would activate cPLA 2 ␣ by disrupting the inhibitory cPLA 2 ␣-A2t interaction and thereby allowing cPLA 2 ␣ to interact with the membrane.
A2t Directly Interacts with Ser 727 of cPLA 2 ␣-To prove that A2t directly interacts with cPLA 2 ␣ via Ser 727 and that the phosphorylation of Ser 727 interferes with this protein-protein interaction, we measured the binding of cPLA 2 ␣ and mutants to A2t in vitro and in HEK293 cells. First, we performed an in vitro GST pull-down assay using purified cPLA 2 ␣ proteins and the GST-tagged A2t, prepared from the GST-tagged p11 and annexin A2. The GST-tagged A2t was functionally identical to the untagged A2t in the cPLA 2 ␣ inhibition assay (data not shown). As shown in Fig. 4A, the cPLA 2 ␣⅐A2t complex formation was clearly seen for wild type and S727T. S727E also formed a complex with A2t, albeit to a slightly lesser degree. Under the same conditions, however, S727A had a dramatically reduced tendency to form a complex with A2t. Also, the cPLA 2 ␣ wild type purified from okadaic acid-treated Sf9 cells showed a significantly lower degree of complex formation with A2t. When  compared with A2t under the same conditions, p11 showed much lower affinity for cPLA 2 ␣ (see Fig. 4B). We then collected the lysates of HEK293 cells co-transfected with FLAG-tagged p11 and cPLA 2 ␣ (annexin A2 is abundant in all mammalian cells, including HEK293 cells; see Fig. 5), immunoprecipitated cPLA 2 ␣⅐A2t complexes using an anti-FLAG antibody, and stained the bands on the gel with anti-cPLA 2 ␣ and anti-annexin A2 antibodies, respectively. Fig. 5 shows that cPLA 2 ␣ wild type, S727E, and S727T formed a complex with A2t, whereas S727A did not under the same conditions. Taken together, these results confirm that A2t directly interacts with cPLA 2 ␣ via Ser 727 both in vitro and in the cell and that Ser 727 phosphorylation disrupts the interaction.
A2t Inhibits Membrane Translocation and AA Release of cPLA 2 ␣ in HEK293 Cells-To further investigate the effect of A2t on cellular activities of cPLA 2 ␣ and how Ser 727 phosphorylation regulates the cellular cPLA 2 ␣-A2t interaction, we monitored the subcellular translocation and activity of cPLA 2 ␣ and mutants expressed in HEK293 cells in the presence of different concentrations of A2t. Our Western blotting analysis showed that HEK293 cells contain undetectable levels of endogenous cPLA 2 ␣ and p11 (data not shown) but have a significant level of endogenous annexin A2 (Fig. 5). We thus expressed the COOH-terminal EYFP-tagged cPLA 2 ␣ (wild type or mutant) to see the gain-of-function in HEK293 cells and then co-expressed p11 to observe the inhibition of cPLA 2 ␣ actions by A2t. It was reported that Ser 727 phosphorylation was negligible in cPLA 2 ␣ expressed in HEK293 cells grown in DMEM containing 10% fetal bovine serum in the absence of an agonist (31). Thus, it was expected that cPLA 2 ␣ expressed in HEK293 cells under our experimental conditions predominantly has a nonphosphorylated Ser 727 (see below). The cell populations expressing similar levels of cPLA 2 ␣ (wild type or a mutant) were selected by visual inspection of EYFP fluorescence intensity and used for translocation measurements. A minimum of quadruple measurements were performed for each protein with more than five cells monitored for each measurement. Typically, Ͼ80% of the cell population showed similar behaviors with respect to membrane translocation of cPLA 2 ␣. Fig. 6 shows the time lapse images of cPLA 2 ␣ and mutants in representative HEK293 (Fig. 6A) and HEK293-BLT1 (Fig. 6B) cells.
In quiescent HEK293 cells, cPLA 2 ␣ and mutants displayed random cytosolic distribution (first column of Fig. 6A). When HEK293 cells were stimulated with 10 M ionomycin, which was shown to cause a sustained increase of the intracellular Ca 2ϩ concentration ([Ca 2ϩ ] i ) to ϳ2 M under our experimental conditions (33), cPLA 2 ␣ and mutants all rapidly translocated to the perinuclear region. The membrane translocation was completed within the first 3 min.

TABLE 1
Relative abundance of Ser 727 -phosphopeptide and Ser 727 -nonphosphopeptide determined for cPLA 2 ␣ prepared from Sf9 cells and okadaic acid-treated Sf9 cells The phosphopeptide and nonphosphopeptide peaks corresponding to residues 726 -736 were identified in the ion chromatograms, and the integrated peak areas were normalized against those of trypsin autolysis peptides. The extent of Ser 727 phosphorylation was then calculated from the normalized areas of phosphopeptide and nonphosphopeptide peaks.

Proteins
Normalized cPLA 2 ␣ and mutants also showed homogenous cytosolic distribution when both a cPLA 2 ␣ protein and p11 were expressed in HEK293 cells (third column of Fig. 6A). In these co-transfected HEK293 cells, stimulation with 10 M ionomycin did not cause the perinuclear targeting of cPLA 2 ␣ wild type, S727E, and S727T; however, S727A translocated to the perinuclear region to a significant degree in the presence of p11. These results are consistent with our in vitro results showing that A2t binds to and inhibits the membrane binding of cPLA 2 ␣ wild type, S727E, and S727T but not S727A.
To measure the cellular inhibitory action of A2t at physiologically relevant [Ca 2ϩ ] i , we performed similar experiments in HEK293-BLT1 cells in which a transient increase of [Ca 2ϩ ] i to a submicromolar level can be induced by LTB 4 . In quiescent HEK293-BLT1 cells, cPLA 2 ␣ wild type and mutants were again distributed in the cytosol with (Fig. 6B, first column) or without (Fig. 6B, third column) p11 overexpression. When the cells were stimulated with 3 M LTB 4 , which raised [Ca 2ϩ ] i to ϳ500 nM (see Fig. 6C), all cPLA 2 ␣ proteins were recruited to the perinuclear region, albeit significantly more slowly than in HEK293 cells (Fig. 6B, second column). As was the case with HEK293 cells, the overexpression of p11 in HEK293-BLT1 cells prevented cPLA 2 ␣ wild type, S727E, and S727T from translocating to the perinuclear region in response to the [Ca 2ϩ ] i increase while having a minimal effect on the membrane targeting of S727A. These cell translocation data thus further support the physiological relevance of the inhibitory A2t-cPLA 2 ␣ interaction and its regulation by Ser 727 phosphorylation.
We then measured the effects of p11 overexpression on the cPLA 2 ␣ activity in HEK293 cells transfected with EYFP-tagged cPLA 2 ␣ wild type and mutants. Specifically, we labeled these HEK293 cells expressing comparable levels of cPLA 2 ␣ proteins (see Fig. 7A) with [ 3 H]AA and monitored the AA release for 30 min after ionomycin activation. The cPLA 2 ␣-overexpressing cells produced four fold more AA than the controls cells, and this increase was greatly (i.e. by 50%) reduced by the p11 overexpression (Fig. 7B). Similar inhibition by p11 was observed for S727T-overexpressing HEK293 cells. In agreement with in vitro studies, a significant but lesser degree of p11 inhibition was seen in S727E-expressing cells than in wild type-or S727T-expressing cells. Furthermore, p11 overexpression did not cause statistically significant inhibition of AA release in S727A-overexpressing HEK293 cells. As shown in Fig. 7C, essentially the same trend was observed in HEK293-BLT1 cells stimulated with LTB 4 . These cell translocation and activity data corroborate that A2t inhibits cPLA 2 ␣ by interacting with Ser 727 and thereby interfering with its membrane binding in the cell.
To further assess the role of Ser 727 phosphorylation on the cellular activity of cPLA 2 ␣, we measured the A2t inhibition of S505E and S505E/S727A in HEK293 cells treated with PMA. No agonist has been reported to specifically phosphorylate Ser 727 of cPLA 2 ␣ in mammalian cells. These two mutants were selected because they allow separate evaluation of the effect of Ser 727 phosphorylation using agonists, including PMA, that have been reported to induce phosphorylation at both Ser 505 and Ser 727 in mammalian cells (29). When HEK293 cells expressing wild type and S505E, respectively, were treated with 1 M PMA for 20 min, the level of phosphoserine was significantly increased as measured by immunoblotting using a phosphoserine-specific antibody (Fig. 8A). Under the same conditions, however, little increase in phosphoserine was detected in HEK293 cells expressing S505E/S727A, suggesting that the PMA treatment induced the Ser 727 phosphorylation for S505E. Interestingly, overexpression of p11 in HEK293 cells expressing FIGURE 4. In vitro binding of cPLA 2 ␣ with A2t. A, 0.1 nmol of GST-A2t (or GST) and 1 g of cPLA 2 ␣ (or mutants) were incubated in PBS at 4°C for 3 h, and the GST proteins were precipitated with the glutathione-Sepharose 4B beads. The proteins bound to the glutathione beads were separated by SDS-PAGE and immunoblotted (IB) with the anti-cPLA 2 ␣ antibody. The gel was also stained with Coomassie Blue to visualize GST-A2t and GST. WT, wild type; WT-OA, cPLA 2 ␣ purified from okadaic acid-treated Sf9 cells. B, 0.1 nmol of GST-A2t or GST-p11 (or GST) and 1 g of cPLA 2 ␣ were incubated in PBS at 4°C for 3 h, and protein complex was precipitated with the glutathione-Sepharose 4B beads. After electrophoresis, membrane was blotting with the anti-cPLA 2 ␣ and anti-GST antibody. FIGURE 5. Immunoprecipitation of the cPLA 2 ␣⅐A2t complex from HEK293 cells. Lysates from HEK293 cells co-transfected with cPLA 2 ␣ (wild type (WT) or mutants) and FLAG-tagged p11 were incubated with 10 l of anti-FLAG M2-agarose for 4 h. The proteins bound to the beads were separated by SDS-PAGE, and the gel was transferred onto a nitrocellulose membrane for immunoblotting (IB) with the indicated antibodies. The first two rows are immunoblotting images of the co-immunoprecipitated proteins, and the bottom three rows show immunoblotting of total cell lysates demonstrating comparable levels of protein expression. cPLA 2 ␣ wild type did not appreciably reduce the level of phosphoserine (Fig. 8B). This indicates that the kinase(s) responsible for Ser 727 phosphorylation has higher apparent affinity for cPLA 2 ␣ than A2t, which makes it possible to regulate cPLA 2 ␣-A2t interactions through Ser 727 phosphorylation. We then measured the effect of p11 overexpression on the AA-releasing activity from [ 3 H]AA-labeled HEK293 cells expressing wild type, S505E, and S505E/S727A, respectively, with and without PMA pretreatment (Fig. 8C). Consistent with the phosphorylation data, the PMA treatment caused an approximately 40% increase in net cPLA 2 ␣ activity (equal to wild type activity minus control). Importantly, the p11 overexpression inhibited the PMA-boosted cPLA 2 ␣ activity by only 20% while inhibiting 45% of the cPLA 2 ␣ activity in the absence of PMA stimulation. This is consistent with the notion that Ser 727 phosphorylation relieves the inhibitory cPLA 2 ␣-A2t interactions. Similarly, net S505E activity was enhanced by 30% by PMA, and this activity was inhibited only modestly (18%) by p11 in HEK293 cells, in contrast to 45% inhibition of S505E in the absence of PMA. Last, S505E/S727A was neither activated by PMA treatment nor inhibited by p11 overexpression. Thus, it appears that PMA activates cPLA 2 ␣ mainly by inducing the phosphorylation of Ser 727 under our experimental conditions presumably because Ser 505 is almost constitutively phosphorylated. Collectively, these results further support the notion that Ser 727 phosphorylation activates cPLA 2 ␣ in HEK293 cells by relieving inhibitory cPLA 2 ␣-A2t interactions.

DISCUSSION
Although it has been reported (18 -22) that the cellular activity and localization of cPLA 2 ␣ are modulated by specific protein-protein interactions, the physiological relevance and significance of these findings have not been established. Also, the mechanisms by which phosphorylation of different sites of cPLA 2 ␣ regulates its activity, individually or in combination, are not fully understood. This study addresses both questions by linking Ser 727 phosphorylation to the cPLA 2 ␣-A2t interaction (33). A main hypothesis of the present study is that Ser 727 phosphorylation disrupts the p11-cPLA 2 ␣ interaction that inhibits the cPLA 2 ␣ activity. The hypothesis is based on the previous reports indicating that p11 inhibits cPLA 2 ␣ in human epithelial cells by binding to its COOH-terminal region (18,39,49) where Ser 727 is located.
Our in vitro measurements demonstrate that it is not p11 but A2t that potently binds and inhibits cPLA 2 ␣. The observed inhibition of cPLA 2 ␣ by A2t is not due to the nonspecific substrate depletion, because the inhibition is independent of the substrate concentration and A2t has extremely low affinity for zwitterionic phosphatidylcholine vesicles. Indeed, both the GST pull-down assay and the immunoprecipitation in HEK293 cells show that A2t directly binds cPLA 2 ␣. The present study does not yield specific structural information about how exactly A2t inter- acts with cPLA 2 ␣. However, the finding that the S727T and S727E mutants with hydrogen bonding-forming Thr and Glu, respectively, at 727 can interact with A2t, whereas S727A with a nonhydrogen bond-forming Ala cannot bind A2t indicates that the hydroxyl group of Ser 727 is directly involved in hydrogen bond(s) with A2t.
MS data and A2t-binding properties of cPLA 2 ␣ wild type proteins prepared under different conditions clearly show that Ser 727 phosphorylation disrupts the inhibitory A2t-cPLA 2 ␣ interaction(s) and thereby activates cPLA 2 ␣. It has been generally thought that phosphorylation regulates the activity of cellular proteins either by inducing conformational changes or altering local electrostatics. According to either mechanism, FIGURE 7. Inhibition of cellular AA-releasing activity of cPLA 2 ␣ wild type and mutants by A2t. A, HEK293 cells transiently transfected with cPLA 2 ␣ wild type (WT) and mutants were washed and scraped in the lysis buffer. Total cell lysates (20 g) were resolved on an SDS-polyacrylamide gel, and the gel was transferred onto a nitrocellulose membrane for immunoblotting with the monoclonal cPLA 2 ␣ antibody. Three independent measurements gave essentially the same pattern.  A, HEK293 cells transiently transfected with cPLA 2 ␣ wild type (WT) and mutants were subjected to immunoprecipitation with the anti-GFP antibody. Aliquots of immunoprecipitates (IP) were resolved on an SDS-polyacrylamide gel and subjected to immunoblotting (IB) with the Ser(P) antibody and the cPLA 2 ␣ antibody, respectively. The data are representative of at least two experiments. B, HEK293 cells transiently transfected with cPLA 2 ␣ alone and with both cPLA 2 ␣ and p11 were subjected to the same analysis as described for A. C, the [ 3 H]AA release from HEK293 cells was measured as described for Fig. 7B, except that some cells were treated with 1 M PMA for 15 min prior to ionomycin stimulation. Each data point represents a mean and S.E. from triplicate measurements. *, pairs of data sets with p Ͻ 0.05. S727E would simulate the Ser 727 -phosphorylated form, whereas S727A and S727T would mimic the nonphosphorylated form. However, unexpected properties of S727A, S727E, and S727T as well as kinetic behaviors of cPLA 2 ␣ wild type proteins prepared under different conditions suggest that Ser 727 phosphorylation may activate cPLA 2 ␣ by a different mechanism (i.e. disruption of Ser 727 -A2t hydrogen bonds by introducing a bulky phosphate group) (Fig. 9). Although the verification of this mechanism would need further structural studies, it represents a new mechanism by which phosphorylation modulates the cellular protein activities.
Since Ser 727 is located in the flexible COOH-terminal end that is remote from either the active site or the membrane binding surface of cPLA 2 ␣, its mutations or phosphorylation has a minimal direct effect on enzyme activity or membrane affinity. Likewise, the interaction of cPLA 2 ␣ with a small protein via Ser 727 is not expected to significantly affect its membrane binding and lipid hydrolysis. This may explain why p11 shows much lower cPLA 2 ␣ inhibitory activity than A2t. The structure of A2t has not been solved yet, but the structure of p11 in complex with the NH 2 -terminal peptide of annexin A2 (38) shows that two p11 molecules form a tight dimeric interface upon which two annexin A2 molecules can be attached via their NH 2 termini. In this molecular arrangement, A2t will form a bulky complex, and the interaction of the hydroxyl group of Ser 727 of cPLA 2 ␣ with p11 in A2t would sterically interfere with the membrane binding of cPLA 2 ␣ (Fig. 9), which is confirmed by our SPR measurements (Fig. 3). The fact that the maximal inhibition of cPLA 2 ␣ by A2t is about 50% indicates that A2t partially blocks the membrane binding of cPLA 2 ␣. Phosphoryla-tion is known to modulate the cellular protein activity to varying degrees. For cPLA 2 ␣, phosphorylation has been reported to cause modest 2-3-fold activation under in vitro and physiological conditions (30,31,33). Thus, the proposed mode of cPLA 2 ␣ activation through Ser 727 phosphorylation, albeit not inducing dramatic activity changes, may serve a purpose of cellular regulation of cPLA 2 ␣ activities. Further, the lack of inhibition of Ser 727 phosphorylation by p11 overexpression (see Fig. 8B) supports the notion that a kinase(s) can readily modulate cPLA 2 ␣-A2t interactions through Ser 727 phosphorylation.
In HEK293 cells overexpressing cPLA 2 ␣ wild type and mutants, p11 overexpression inhibits the Ca 2ϩ -induced perinuclear translocation of and AA release by cPLA 2 ␣ wild type, S727E, and S727T but not by S727A. Also, PMAinduced Ser 727 phosphorylation in HEK293 cells greatly reduced the inhibition of cPLA 2 ␣ activity by p11 overexpression. The good correlation between our in vitro and cell data supports that phosphorylation of Ser 727 activates cPLA 2 ␣ by the same mechanism in the cell. It should be noted that overexpression of p11 is necessary for observing inhibition of overexpressed cPLA 2 ␣ in HEK293 cells, because HEK293 cells do not have a significant level of endogenous p11. However, a large portion of annexin A2 is known to exist as A2t in other mammalian cells under physiological conditions (50). Thus, inhibition of cPLA 2 ␣ by A2t is expected to be significant in quiescent mammalian cells before Ser 727 phosphorylation.
Among reported phosphorylation sites of cPLA 2 ␣, phosphorylation of Ser 505 and Ser 727 is known to have the most significant effect in agonist-induced AA release (30,31). It still remains unknown whether Ser 505 phosphorylation activates FIGURE 9. A proposed mechanism by which Ser 727 phosphorylation modulates the cPLA 2 ␣ activity. The hydroxyl group of Ser 727 of cPLA 2 ␣ forms hydrogen bonds with a part of p11 in A2t. This interaction with a bulky A2t complex sterically interferes with the membrane binding of cPLA 2 ␣, hence the cPLA 2 ␣ inhibition. Phosphorylation of Ser 727 abrogates the cPLA 2 ␣-A2t interaction, thereby allowing the enzyme to bind the membrane and catalyze the lipid hydrolysis. Phosphorylation of Ser 505 does not appear to influence the effect of Ser 727 phosphorylation on the cPLA 2 ␣-A2t interaction, but the cPLA 2 ␣-A2t binding may inhibit both phosphorylated and nonphosphorylated forms of Ser 505 . The schematic diagram of A2t is not drawn to scale. cPLA 2 ␣ by modulating membrane-protein (33) or protein-protein interactions (34), because conflicting effects have been reported in different cell lines. Regardless of the activation mechanism of Ser 505 phosphorylation, the present study clearly shows that the modulation of cPLA 2 ␣-A2t interactions by Ser 727 phosphorylation is independent of Ser 505 phosphorylation. This is consistent with the separate locations of Ser 505 and Ser 727 in the cPLA 2 ␣ structure (Fig. 9).
In summary, our in vitro and cell data indicate that phosphorylation of Ser 727 activates cPLA 2 ␣ by disrupting the inhibitory hydrogen bond interactions between cPLA 2 ␣ and A2t, thereby allowing cPLA 2 ␣ to bind the membrane and catalyze the phospholipid hydrolysis. It is possible that Ser 727 is involved in other protein-protein interactions that, whether activating or inhibitory, may be also regulated by Ser 727 phosphorylation. Further studies are necessary to fully address this question.