Tyrosine phosphorylation of annexin II tetramer is stimulated by membrane binding.

In the present article we have examined if the interaction of the Ca2+-binding protein, annexin II tetramer (AIIt) with the plasma membrane phospholipids or with the submembranous cytoskeleton, effects the accessibility of the tyrosine phosphorylation site of AIIt. In the presence of Ca2+, pp60c-src catalyzed the incorporation of 0.22 ± 0.05 mol of phosphate/mol of AIIt (mean ± S.D., n = 5). The Ca2+-dependent binding of AIIt to purified adrenal medulla plasma membrane or phosphatidylserine vesicles stimulated the pp60c-src-dependent phosphorylation of AIIt to 0.62 ± 0.04 mol of phosphate/mol of AIIt (mean ± S.D., n = 5) or 0.93 ± 0.07 mol of phosphate/mol of AIIt (mean ± S.D., n = 5), respectively. Phosphatidylserine- or phosphatidylinositol-containing vesicles but not vesicles composed of phosphatidylcholine or phosphatidylethanolamine, stimulated the phosphorylation of AIIt. In contrast, the binding of AIIt to F-actin resulted in the incorporation of only 0.04 ± 0.04 mol of phosphate/mol of AIIt (mean ± S.D., n = 5). These results suggest that the interaction of AIIt with plasma membrane and not the submembranous cytoskeleton, activates the tyrosine phosphorylation of AIIt by inducing a conformational change in the protein resulting in the enhanced exposure or accessibility of the tyrosine-phosphorylation site.

The annexins (reviewed in Refs. [1][2][3][4][5] are a family of Ca 2ϩbinding proteins that bind acidic phospholipids. Members of this family contain a region of amino acid homology called the annexin fold (6). Annexin II tetramer (AIIt) 1 is an abundant annexin which is composed of two M r 36,000 annexin II subunits and two M r 11,000 subunits. The 36-kDa subunit consists of two functional domains. The first, the amino-terminal regulatory domain contains the first 30 amino acids of the amino terminus of the heavy chain, incorporates the serine and tyrosine-phosphorylation sites and the binding site for the p11 light chain. The remaining carboxyl domain, comprises the binding sites for Ca 2ϩ , phospholipid, and F-actin (reviewed in Ref. 2). AIIt is associated with the cytosolic surface of the plasma membrane in association with the submembranous cytoskeleton of many secretory cells where the protein has been shown to form links between the plasma membrane and secre-tory granules (6,7). Although the exact physiological function of AIIt is unclear, the protein is thought to play a role in Ca 2ϩ -dependent exocytosis or endocytosis (2).
AIIt has been shown to be phosphorylated in vivo by protein tyrosine kinases. For example, the expression of transforming protein tyrosine kinases in a variety of cells has been shown to correlate with the appearance of phosphotyrosine in AIIt (8,9) and in many cells AIIt is a major in vivo substrate of pp60 v-src (10 -12). Activation of transmembrane protein kinase receptors, such as the platelet-derived growth factor receptor, has also been shown to result in the tyrosine phosphorylation of AIIt (13)(14)(15). The phosphorylation of AIIt in pp60 v-src transformed cells or in cells activated by platelet-derived growth factor is identical to the site phosphorylated on the protein in vitro by pp60 v-src , namely tyrosine 23 (13).
In previous work, we examined the consequences of the tyrosine phosphorylation of AIIt on the biological activities of the protein (16). We reported that pp60 c-src -phosphorylated AIIt did not bind to heparin or bind to or bundle F-actin. Furthermore, native AIIt but not tyrosine-phosphorylated AIIt could promote the formation of a plasma membrane-AIIt-chromaffin granule complex in vitro (16). We therefore concluded that the tyrosine phosphorylation of AIIt was an inhibitory signal. Consistent with these results, we have shown that activation of adrenal chromaffin cells, with acetylcholine, results in the dephosphorylation of AIIt concomitant with the release of catecholamine (2).
In the present article we have examined the kinetics of phosphorylation of AIIt by pp60 c-src . The results suggest that the binding of AIIt to the phospholipid component of the plasma membrane stimulates the phosphorylation of the protein by pp60 c-src .

MATERIALS AND METHODS
Phosphorylation of Annexin II Tetramer-AIIt phosphorylation reactions were performed according to Ref. 16. AIIt at a final concentration of 60 g/ml, was incubated at 30°C for 30 min in 25 mM HEPES (pH 7.5), 10 mM MgCl 2 , 0.5 mM EGTA, 0.6 mM CaCl 2 , 1.5 g/ml baculovirus produced human recombinant pp60 c-src , and 100 l/ml lipid vesicles which were taken from a stock containing 200 g/ml phosphatidylserine, 200 g/ml phosphatidylcholine, and 40 g/ml diolein. When incubations were performed in the presence of plasma membrane, 0.5 mM orthovanadate and 3 mg/ml para-nitrophenyl phosphate were included as phosphatase inhibitors. The reaction was initiated by addition of 25 M ATP (200 -2000 cpm/pmol [␥-32 P]ATP). To quantify the stoichiometry of phosphorylation of AIIt, 25 l was removed from the reaction mixture and either precipitated with 25% trichloroacetic acid and 2% sodium pyrophosphate and subjected to scintillation counting, or alternatively, boiled with 1 volume of SDS-PAGE sample buffer (0.25 M Tris-HCl, pH 6.8, 10% SDS, 20% glycerol, 2 mM EGTA, 2 mM EDTA, 20 mM ␤-mercaptoethanol) and analyzed by SDS-PAGE (17). For experiments testing the phospholipid specificity of phosphorylation (Table  II), phospholipid liposomes were prepared by hydrating 4 mg of phospholipid (previously dried from a chloroform solution under N 2 ) with 50 mM HEPES (pH 7.5). The suspensions were mixed in a vortex and sonicated three times for 15 s. Since phosphatidylethanolamine did not readily form stable liposomes when used as a pure phospholipid, phosphatidylethanolamine/phosphatidylcholine (5:1) liposomes were used to measure the stimulation of phosphorylation of AIIt by these liposomes.
Miscellaneous Techniques-AIIt was prepared from bovine lung (18) and stored in 50 mM KCl at Ϫ70°C. The protein was essentially homogeneous as determined by SDS-PAGE. Adrenal medulla plasma membranes were purified according to Ref. 16. Protein concentration was measured with the Bradford Coomassie Blue dye binding method using bovine serum albumin as a standard (19). Alternatively, protein concentrations were determined spectrophotometrically using appropriate extinction coefficients (16). Lipid vesicles for the annexin phosphorylation reaction were prepared fresh daily according to Ref. 20. Human pp60 c-src was made using a baculovirus vector and purified by hydroxyapatite and immunoaffinity chromatography (21). Partially proteolyzed pp60 c-src was produced by omitting protease inhibitors during the final Mono-Q FPLC column purification step (21). This resulted in proteolysis of the pp60 c-src preparation without any loss in enzyme activity. The partially proteolyzed pp60 c-src was chromatographed on a monoclonal antibody affinity column (utilizing Mb 203 07D10-Quality Biotech). This monoclonal antibody recognizes an epitope at amino acids residues 2-17 of pp60 c-src . The column flow-through was pooled and SDS-PAGE analysis identified a major 51-kDa protein band. This protein reacted with the 327 monoclonal antibody but not the 203 07D10 monoclonal antibody. This suggested that the partially proteolyzed pp60 c-src had lost part of the NH 2 terminus including the myristate that is normally attached to Gly-2. The proteolyzed pp60 c-src was extensively characterized 2 and found to have similar activity to the native enzyme. Calcium concentrations were determined by Ca 2ϩ electrode and FURA measurements (20). Fig. 1 presents the time course of phosphorylation of AIIt by pp60 c-src . At low Ca 2ϩ concentration (0.1 M) a slow rate of incorporation of phosphate into AIIt was observed with about 0.09 Ϯ 0.08 mol of phosphate/mol of AIIt (mean Ϯ S.D., n ϭ 5) incorporated by pp60 c-src in 60 min (Fig. 1). When 100 M Ca 2ϩ was added to the reaction, the rate of phosphorylation of AIIt by pp60 c-src was also slow and after 60 min only 0.22 Ϯ 0.05 mol of phosphate/ mol of AIIt (mean Ϯ S.D., n ϭ 5) was incorporated. Increasing the Ca 2ϩ concentration did not alter the phosphorylation rate and at 1 mM Ca 2ϩ about 0.21 Ϯ 0.01 mol of phosphate/mol of AIIt (mean Ϯ S.D., n ϭ 5) was incorporated. The addition of phospholipid vesicles to the reaction mixture stimulated the rate of phosphorylation of AIIt and at equilibrium, 0.84 Ϯ 0.05 mol of phosphate/mol of AIIt (mean Ϯ S.D., n ϭ 5) was incorporated. In the presence of 100 M Ca 2ϩ and phospholipid vesicles, the phosphorylation of AIIt was very rapid and at equilibrium about 0.93 Ϯ 0.07 mol of phosphate/mol of AIIt (mean Ϯ S.D., n ϭ 9) was incorporated.

Time Course of Phosphorylation of AIIt by pp60 c-src -
The results presented in Fig. 1 suggested that the addition of phospholipid vesicles to the reaction mixture stimulated the rate and extent of phosphorylation of AIIt by pp60 c-src . In order to determine if the stimulation of AIIt phosphorylation was due to the interaction of phospholipid vesicles with AIIt or pp60 c-src , the effect of phospholipid vesicles on pp60 c-src activity was examined. The pp60 c-src activity in these experiments was measured using a peptide to the phosphorylation site of AIIt (KLSLEGDHSTPPSAYGSVKAYT). This peptide exhibited the following enzymatic parameters: K m ϭ 198.4 Ϯ 25.5 M (mean Ϯ S.D., n ϭ 5); V max ϭ 41.3 Ϯ 2.7 pmol of phosphate⅐min Ϫ1 ⅐pmol Ϫ1 (mean Ϯ S.D., n ϭ 5) which compared favorably with the pp60 c-src -dependent phosphorylation of the CDC 2 peptide (residues 6 -20) (K m ϭ 245.5 Ϯ 26.6 M; mean Ϯ S.D., n ϭ 5; V max ϭ 61.8 Ϯ 4.1 pmol of phosphate⅐min Ϫ1 ⅐pmol; mean Ϯ S.D., n ϭ 5) (22). As shown in Table I, pp60 c-src exhibited a maximal initial rate of phosphorylation of the AIItphosphorylation site peptide in the absence of added Ca 2ϩ or phospholipid vesicles. The addition of Ca 2ϩ , phospholipid vesicles or both, resulted in a small decrease in the initial phosphorylation rate. In contrast, the initial rate of phosphorylation of AIIt was maximal in the presence of both Ca 2ϩ and phospholipid vesicles. The phosphorylation rate in the presence of Ca 2ϩ was only 12% of the rate of phosphorylation of AIIt in the presence of both Ca 2ϩ and phospholipid vesicles. Furthermore, at a Ca 2ϩ concentration of 0.1 M, the initial rate of phosphorylation of AIIt was stimulated almost 13-fold by the addition of phospholipid vesicles. These results therefore suggest that the stimulation of phosphorylation of AIIt observed in the presence of phospholipid vesicles was not due to the stimulation of pp60 c-src activity by the phospholipid vesicles but due to a phospholipid-induced increased exposure of the tyrosinephosphorylation site of AIIt.
Stimulation of pp60 c-src -dependent Phosphorylation of AIIt by Plasma Membrane-As shown in Fig. 2, at equilibrium, about 0.62 Ϯ 0.04 mol of phosphate (mean Ϯ S.D., n ϭ 5) were incorporated into AIIt by pp60 c-src in the presence of purified adrenal medulla plasma membrane (Fig. 2, inset). This compared to 0.22 Ϯ 0.05 mol of phosphate/mol of AIIt (mean Ϯ S.D., n ϭ 5) incorporated in the absence of plasma membrane. The  stimulation of the phosphorylation of the pp60 c-src -dependent phosphorylation of AIIt by plasma membrane could be due to the interaction of AIIt with phospholipid or F-actin components of the plasma membrane. However, as shown in Fig. 2, the interaction of AIIt with F-actin did not activate the phosphorylation of AIIt by pp60 c-src . In contrast, the interaction of AIIt with phospholipid stimulated the rate (Table I) and extent (Fig.  2) of phosphorylation of AIIt. These results therefore suggest that the binding of AIIt to the phospholipid component of the plasma membrane is responsible for stimulation of the phosphorylation of AIIt. Phosphorylation of AIIt by Partially Proteolyzed pp60 c-src -Both AIIt and pp60 c-src are membrane-associated proteins that are known to bind acidic phospholipids (23)(24)(25)(26). Therefore, it was possible that the phospholipid-dependent activation of phosphorylation of AIIt could be due to recruitment of AIIt and pp60 c-src to the phospholipid vesicles, thereby increasing the proximity of pp60 c-src and AIIt. This would result in an increase in the effective concentration of AIIt and pp60 c-src in the phosphorylation reaction and could produce an enhanced phosphorylation rate. To examine this possibility, we prepared a partially proteolyzed pp60 c-src . The partial proteolysis of the pp60 c-src results in the loss of a portion of the NH 2 -terminal domain of the enzyme and the loss of the NH 2 -terminal myristic acid. It was therefore expected that the loss of this domain of pp60 c-src would inhibit the binding of the enzyme to phospholipid liposomes. Therefore, native and partially proteolyzed pp60 c-src were incubated with phosphatidylserine liposomes and following centrifugation, the pellets were examined for pp60 c-src activity. As shown in Fig. 3, the partially proteolyzed pp60 c-src did not bind to phosphatidylserine liposomes. In contrast, the native pp60 c-src bound to the phospholipid liposomes.
Since the partially proteolyzed pp60 c-src did not bind to phospholipid liposomes it was possible to directly examine the role of phospholipid in the phosphorylation of AIIt without the potential complication of the binding of both pp60 c-src and AIIt to the phospholipid liposomes. We therefore examined the specificity of the phospholipid stimulation of phosphorylation of AIIt by partially proteolyzed pp60 c-src . As shown in Table II, phosphatidylserine and phosphatidylinositol liposomes stimulate the phosphorylation of AIIt by partially proteolyzed pp60 c-src . In contrast, phosphatidylcholine or phosphatidylethanolamine liposomes do not stimulate the phosphorylation of AIIt. Therefore, this data provides direct evidence that phospholipid-induced conformational change in AIIt stimulates the tyrosine phosphorylation of AIIt by pp60 c-src . Furthermore, since AIIt binds phosphatidylserine and phosphatidylinositol but not phosphatidylcholine or phosphatidylethanolamine (23), these results suggest that phospholipid binding and not a nonspecific effect of phospholipid is responsible for the phospholipid stimulation of phosphorylation of AIIt by pp60 c-src .
Dephosphorylation of AIIt by Plasma Membrane Phosphatases-Our results suggest that the binding of AIIt to plasma membrane activates a conformational change in AIIt which enhances exposure of the tyrosine-phosphorylation site of the protein, resulting in the stimulation of the phosphorylation of the protein by pp60 c-src . Previous work from our laboratory has shown that the tyrosine phosphorylation of AIIt induces a conformational change in AIIt which results in the inability of the protein to bind F-actin or heparin or to bridge biological membranes (16). Therefore, it was unclear from these studies if the tyrosine-phosphorylation site of plasma membrane-bound, tyrosine-phosphorylated AIIt, was accessible to plasma membrane-associated protein tyrosine phosphatases. Since the binding of AIIt to plasma membrane stimulated the phosphorylation of AIIt by pp60 c-src , it was important to ascertain if the plasma membrane bound AIIt was a substrate for plasma membrane associated phosphotyrosine phosphatases. As shown in Fig. 4, the incubation of pp60 c-src -phosphorylated AIIt with plasma membrane resulted in the dephosphorylation of AIIt. The dephosphorylation was very rapid and was essen-  pp60 c-src (lanes a and b) and partially proteolyzed pp60 c-src (lanes c and d) were autophosphorylated in the presence of phospholipid liposomes ("Materials and Methods") at 30°C. After 30 min the reaction mixture was mixed and aliquots were taken prior to centrifugation and disrupted with SDS-PAGE sample buffer (lanes a and c). The remaining phospholipid liposomes were pelleted by centrifugation (10,000 ϫ g for 5 min), resuspended in 25 mM HEPES (pH 7.5), 10 mM MgCl 2 , 0.5 mM EGTA, 0.6 mM CaCl 2 , repelleted, and then resuspended in SDS-PAGE sample buffer (lanes b and d). The samples were analyzed by SDS-PAGE on a 9% polyacrylamide gel. An autoradiogram is shown.

FIG. 2. Activation of the tyrosine phosphorylation of annexin II tetramer by plasma membrane binding.
Annexin II tetramer (50 g/ml) was incubated with 2.6 g/ml pp60 c-src at 37°C for 60 min in the presence of 0.5 mM orthovanadate and 3 mg/ml para-nitrophenyl phosphate as described under "Materials and Methods." Additions included 0.6 mM CaCl 2 (Ca 2ϩ ), 100 l/ml phospholipid vesicles (PL), 100 M F-actin (Ac), or 0.16 mg/ml adrenal medulla plasma membrane (PM). Inset, after 60 min, aliquots were removed and analyzed by SDS-PAGE and autoradiography. The arrow indicates the position of AIIt. tially complete after 10 s of incubation of tyrosine-phosphorylated AIIt with plasma membrane. This indicated that the tyrosine-phosphorylation site of plasma membrane-bound AIIt was accessible to membrane phosphatases and could be regulated by cycles of tyrosine phosphorylation and dephosphorylation.

DISCUSSION
Studies of the x-ray crystallographic analysis of annexin V have provided valuable information on the possible structure of all members of the annexin family of Ca 2ϩ -binding proteins (27)(28)(29). These studies support the prediction that the p36 heavy chain is a planar molecule composed of opposing concave and convex surfaces. The convex surface is predicted to lie along the plane of the phospholipid membrane and contain the Ca 2ϩ -binding sites. The concave surface of the protein faces the cytosol and contains the amino-terminal and carboxyl-terminal domains. The exposure of the amino-terminal domain of the p36 heavy chain to the cytosol results in the accessibility of the Tyr-23 and Ser-25 phosphorylation sites to pp60 src and protein kinase C, respectively, and also allows binding of the p11 light chains to the first 12 amino acids of the amino-terminal domain. These studies have not addressed the issue of whether or not the accessibility of the Tyr-23 and Ser-25 phosphorylation sites are influenced by the binding of AIIt to Ca 2ϩ or plasma membrane.
Predictions based on the x-ray crystallographic structure of annexin V as well as fluorescence spectroscopic analysis of AIIt have suggested that AIIt can exist in several distinct conformations. AIIt binds Ca 2ϩ with a K d (Ca 2ϩ ) of about 0.5 mM Ca 2ϩ . Ca 2ϩ binding results in a decrease in the ␣-helical content of AIIt (30) and the exposure of Trp-212 to a more hydrophobic environment (30 -32). In the presence of phosphatidylserine, AIIt binds 15 mol of Ca 2ϩ /mol of AIIt with a K d (Ca 2ϩ ) of 1.3 M (33, 34). X-ray crystallographic analysis has suggested that phosphatidylserine binding will cause a substantial conformational change in AIIt. The phosphatidylserine-induced conformation change results from phosphatidylserine linking two Ca 2ϩ -binding sites that are about 8.8 Å apart and from the interaction of phosphatidylserine with amide group of Gly-206 and the hydroxyl group of Thr-207. The phosphoryl oxygen of phosphatidylserine is involved in the coordination of AB site Ca 2ϩ while the serine carboxylate oxygen of phosphatidylserine coordinates the ABЈ Ca 2ϩ . However, there are significant differences between the Ca 2ϩ -induced conformational change that occurs when Ca 2ϩ is added to AIIt at millimolar concentration in the absence of phosphatidylserine and the conformational change that occurs when AIIt binds Ca 2ϩ at micromolar Ca 2ϩ in the presence of phosphatidylserine (34). Therefore, it has been concluded that the conformation of AIIt, induced by binding of Ca 2ϩ to its Ca 2ϩ -binding sites on AIIt differs from the conformation of AIIt induced by the binding of Ca 2ϩ to its Ca 2ϩ -binding sites in the presence of phosphatidylserine.
Although the structure of AIIt-F-actin complex is unknown, it is reasonable to suggest that this conformation of AIIt will be distinct from other AIIt conformations. Within the context of our studies, our data suggests that the Tyr-23 phosphorylation site of AIIt is not fully exposed or accessible to pp60 c-src when AIIt is in the Ca 2ϩ -induced conformation or in the F-actininduced conformation. In contrast, when AIIt is in the phosphatidylserine-induced conformation, the tyrosine-phosphorylation site is accessible to pp60 c-src . The binding of Ca 2ϩ and phosphatidylserine to AIIt, which induces the Ca 2ϩ -and phosphatidylserine-induced conformation of AIIt, results in only a modest increased exposure of the tyrosine-phosphorylation site compared to the phosphatidylserine-induced conformation of AIIt.
We have also examined the possibility that the stimulation of the pp60 c-src -dependent AIIt phosphorylation by phosphatidylserine or plasma membrane is not due to a phospholipid-induced conformational change in AIIt but is due to the binding and therefore the enhanced concentration of pp60 c-src and AIIt on the phosphatidylserine liposomes. However, as shown in Table II, the phospholipid-dependent stimulation of phosphorylation of AIIt is also observed in the presence of a partially proteolyzed pp60 c-src . Since this truncated enzyme does not bind to phospholipid liposomes (Fig. 3), the phospholipid-stimulation of phosphorylation of AIIt must be due to a direct effect of phospholipid on AIIt. Furthermore, since the phosphorylation of AIIt is stimulated only by phospholipids that bind to AIIt (Table II), such as phosphatidylserine and phosphatidylinositol, our data suggest that a phospholipid-dependent conformational change in AIIt is responsible for stimulation of the phosphorylation of AIIt by pp60 c-src .
The phosphorylation of other substrates of pp60 src have also been shown to be accelerated upon plasma membrane binding. Vinculin binds phosphatidylserine, phosphatidylglycerol, and phosphatidic acid and the binding of these phospholipids to vinculin correlates with a increase in the phosphorylation of vinculin by pp60 v-src (35). The stimulation of vinculin phosphorylation was shown to be due to a phospholipid-induced conformational change in the protein and these investigators concluded that phospholipid binding resulted in the increased accessibility of a phosphorylation site of vinculin to pp60 v-src .
Within the cell, AIIt has been shown to be localized to the cytoplasmic surface of the plasma membrane in the submembranous cytoskeleton (34). Furthermore, AIIt has been shown to interact with endosomes and transfection of Madin-Darby  4. Dephosphorylation of annexin II tetramer. pp60 c-src was incubated in the presence (a) or absence (b) of 60 g/ml AIIt, as described under "Materials and Methods." The reaction mixture was dialyzed extensively against 40 mM Tris (pH 7.5) and 1 mM dithiothreitol to remove [␥-32 P]ATP. The dialyzed mixture was then incubated at 37°C in the absence (a and b) or presence of 0.2 mg/ml purified adrenal medulla plasma membrane for 10 s (c), 15 s (d), 30 s (e), or 1 min (f). Aliquots were subjected to SDS-PAGE and autoradiography. These results are typical of those seen in three experiments. canine kidney cells with a dominant negative mutant form of AIIt causes translocation of AIIt and early endosomes to the cytoplasm (36,37). Similarly, pp60 c-src has also been shown to be localized at the inner surface of the plasma membrane (38 -40) and more recently this enzyme has been localized to the endosomes of fibroblasts (41). Therefore, AIIt and pp60 c-src exist in the same intracellular compartment. The intracellular localization of AIIt has been suggested to be due to two properties of AIIt observed in vitro, namely phospholipid binding and F-actin binding. Although it is possible that the binding of AIIt to biological membranes could utilize plasma membraneassociated F-actin as the AIIt-binding site, our results (Fig. 2) suggest that AIIt bound to F-actin is not appreciably phosphorylated by pp60 c-src . This result therefore suggests that AIIt bound to the phospholipid component of the plasma membrane is regulated by tyrosine phosphorylation.
Previous results from our laboratory (2,16) have established that the tyrosine phosphorylation of AIIt produces a large conformational change in AIIt which results in the alteration or inhibition of many of the biological activities of AIIt. It was unclear from these previous studies if the phosphorylation site of the plasma membrane-associated tyrosine-phosphorylated AIIt was accessible to phosphotyrosine phosphatases. The results presented in Fig. 4 suggest that the plasma membranebound form of AIIt is readily dephosphorylated by plasma membrane-associated phosphotyrosine phosphatases. Therefore, considering the established association of pp60 c-src with the plasma membrane, it appears that the key elements for the regulation of AIIt by phosphorylation and dephosphorylation are present at the plasma membrane.