Phosphorylation of annexin II tetramer by protein kinase C inhibits aggregation of lipid vesicles by the protein.

Annexin II tetramer (A-IIt) is a member of the annexin family of Ca2+ and phospholipid-binding proteins. The ability of this protein to aggregate both phospholipid vesicles and chromaffin granules has suggested a role for the protein in membrane trafficking events such as exocytosis. A-IIt is also a major intracellular substrate of both pp60src and protein kinase C; however, the effect of phosphorylation on the activity of this protein is unknown. In the current report we have examined the effect of phosphorylation on the lipid vesicle aggregation activity of the protein. Protein kinase C catalyzed the incorporation of 2.1 +/- 0.8 mol of phosphate/mol of A-IIt. Phosphorylation of A-IIt caused a dramatic decrease in the rate and extent of lipid vesicle aggregation without significantly effecting Ca(2+)-dependent lipid binding by the phosphorylated protein. Phosphorylation of A-IIt increased the A50%(Ca2+) of lipid vesicle aggregation from 0.18 microM to 0.65 mM. Activation of A-IIt phosphorylation, concomitant with activation of lipid vesicle aggregation, inhibited both the rate and extent of lipid vesicle aggregation but did not cause disassembly of the aggregated lipid vesicles. These results suggest that protein kinase C-dependent phosphorylation of A-IIt blocks the ability of the protein to aggregate phospholipid vesicles without affecting the lipid vesicle binding properties of the protein.

monomer and a heterotetrameric protein (A-IIt).' A-IIt consists of two copies of the 36-kDa protein and two copies of an 11-kDa protein. Proteolytic enzymes cleave these proteins into two distinct domains: a large protease-resistant core and a small N-terminal tail. The core domain contains the annexin repeats and displays the Ca2+ and phospholipid binding sites Glenney, 1986). The N-terminal tail, which is variable in length and sequence among various annexins, contains the phosphorylation sites (Gould et al., 1986;Glenney and Tack, 1985) and also appears to exert a regulatory function on the core. Furthermore, in the case of A-IIt, the first 17 residues of the N-terminal tail contain the binding site for the 11-kDa subunit.
The N-terminal domains of A-I and A-I1 appear to play a key role in the regulation of Ca2+ and phospholipid binding by these proteins. Proteolysis of the N-terminal domain of A-I appears to decrease the Aso%(Ca2+) of binding to phospholipid vesicles (Ando et at., 1989) without decreasing the total Ca2+ bound, whereas proteolysis of the N terminus of A-I1 heavy chain is required for the chromaffin granule aggregation activity of the protein (Drust and Creutz, 1988). Furthermore, an intact N terminus appears to be necessary for reconstitution of secretion by A-IIt (Ali et al., 1989). The N terminus of A-I1 heavy chain appears to also play a unique role in the function of A-IIt since it contains the binding site for the 11-kDa light chain of A-IIt ( p l l subunit). Comparison of the monomeric form of A-I1 with the tetramer (containing the p l l subunit) suggests that formation of the tetramer results in a molecule with a decreased requirement of Caz+ for both phospholipid binding (Powell and Glenney, 1987;Pigault et at., 1990) and that this molecule acquires the ability to aggregate chromaffin granules (Drust and Creutz, 1988). Results from this laboratory have suggested that although both A-I and A-I1 bind F-actin, only the A-I1 tetramer can bundle Factin (Khanna et al. 1990;Ikebuchi and Waisman, 1990). In contrast, both forms of A-I1 have been suggested to reconstitute secretion (Ali and Burgoyne, 1990;Ali et al., 1989), although these data are controversial (Wu and Wagner, 1991).
The N-terminal domain of A-I and A-I1 also contains the sites for phosphorylation of the proteins. Several of the annexins have been shown to be in vivo substrates of protein tyrosine kinases; A-I1 and the A-IIt are major substrates of p60"" (Gerke and Weber, 1985;, whereas A-I is a substrate of the epidermal growth factor receptor kinase (Fava and Cohen, 1984;De et al., 1986;. A-I, A-11, and A-IIt are also phosphorylated in vivo by protein kinase C (Michener et al., 1986;Gould et al., 1986). In previous studies of the in vitro phosphorylation of annexins by protein kinase C, we reported a stoichiometry nearing 1 mol of phos-
phate/mol of A-I and A-I1 protein and 0.4 mol of phosphate/ mol of A-IIt protein in vitro (Khanna et al., , 1987b. Phosphorylation of A-I appears to involve both serine and threonine residues, whereas A-IIt is phosphorylated only on serine residues (Gould et al., 1986;Schlaepfer and Haigler, 1988;Varticovski et al., 1988). Phosphorylation of A-I1 or A-I has previously been shown t o influence their lipid binding characteristics. Phosphorylation of A-I by the epidermal growth factor receptor (on tyrosine residues) reduced by &fold the A5,,a(Ca2+) of phospholipid vesicle binding (Schlaepfer and Haigler 1987), whereas phosphorylation of A-I1 by p60""'" decreased the binding of the protein to phospholipid vesicles at low Ca2+ concentrations (Powell and Glenney, 1987).
In the present paper we have investigated the effect of phosphorylation of A-IIt by protein kinase C. Our results suggest that phosphorylation of A-IIt inhibits the ability of the protein to aggregate phospholipid vesicles without affecting the binding of the protein to the phospholipid vesicle. This suggests that protein kinase C may play an important role in the regulation of the function of A-IIt.

EXPERIMENTAL PROCEDURES
Purification of Annenin ZZ-A-IIt was purified from frozen bovine lung (Khanna et al., 1990). All steps were carried out a 4 "C. Purified proteins were concentrated to 2-3 mg/ml and stored at -80 "C.
Purification of Protein Kinase C-Protein kinase C was purified from rat brain (Pel-Freez Biologicals) using the procedure established by Wooten et al. (1987), with the exception that a threonine affinity column (Kitano et al., 1986) was substituted for the protamine agarose column. Protein kinase C was stored at -80 "C in 5% glycerol and 0.05% Triton X-100.
Lipid Micelle and Vesicle Preparation-Micelles for assaying protein kinase C activity during purification of the enzyme were prepared by the method of Hannun et al. (1986). Two mg of phosphatidylserine (PS) and 0.5 mg of 1,2-diolein (DAG) (Serdary) were dried under nitrogen and then sonicated (three 5-s bursts at 70 watts) in 1 ml of 0.3% Triton X-100.
Lipid vesicles for annexin phosphorylation and aggregation reactions were prepared fresh daily (Reeves and Dowbin, 1969). PS (200 pl of 20 mg/ml) and DAG (20 pl of 20 mg/ml) in CH&I, were diluted with 2 volumes of methanol. The solvents were then evaporated under a stream of nitrogen. The lipids were rehydrated in 1 ml of 20 mM Tris-HC1 (pH 7.5) and then sonicated by two brief (5-s) bursts at 70 watts with a probe sonicator.
Protein Kinase CAssay of Column Fractions-The reaction mixture used to assay for protein kinase activity contained, in a 100-pl final volume, 25 mM Tris-HCI (pH 7.5), 10 mM MgC12, 0.5 mM EGTA, 0.501 mM CaC12, 50 pg of histone IIIS, 10 pl of micelles, and 25 p M [T-~'P]ATP (200-500 cpm/pmol). The reaction was terminated by the addition of 1 ml of 25% trichloroacetic acid containing 2% sodium pyrophosphate. The acid-precipitated protein was collected in a filter paper funnel (Isolab) and washed three times with 3 ml of 5% trichloroacetic acid containing 1% sodium pyrophosphate. The radioactivity was determined by scintillation counting. Protein kinase C activity was calculated by subtracting the activity measured in the absence of Ca2+ and phospholipid from the activity measured in the presence of Ca2+ and phospholipid.
Phosphorylation of Annexins-A-IIt, 35-60 pg/ml, was incubated a t 30 "C for the indicated times in 25 mM Tris-HC1 (pH 7.5), 10 mM MgCl,, 0.5 mM EGTA, 1.5 mM CaCI,, 1.0 pg/ml protein kinase C, and 200 pl/ml lipid vesicles (400 pg/ml PS and 40 pg/ml DAG). The reaction was initiated by the addition of 25 p~ ATP (200-2,000 cpm/ pmol [Y-~'P]ATP). In time course experiments 25 pl 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-HCI (pH 6.8), 10% SDS, 20% glycerol, 2 mM EGTA, 2 mM EDTA, 20 mM 8-mercaptoethanol) and analyzed Phosphoamino Acid Analysis-Phosphorylated annexin was diluted in SDS sample buffer and run on 11.5% SDS-PAGE. Protein bands were visualized by staining in Coomassie Blue. The bands were excised from the gels, washed three times in 20% methanol to remove by SDS-PAGE. SDS, and then dried by lyophilization. Protein in gel slices was digested by heating at 110 "C in 6 M HCl for 2 h. The samples were dried under nitrogen, resuspended in 50 mM ammonium bicarbonate (pH 8.0), and then lyophilized. Phosphoamino acids were resuspended in electrophoresis buffer (buffer D: pyridinelacetic acid/water, 33:1:40 (v/v)) containing a 1 mM concentration each of 0-phosphoserine, 0phosphothreonine, and 0-phosphotyrosine (Sigma). Samples were spotted on thin layer cellulose plates (Merck). The plates were run in buffer D (pH 3.5), cathode to anode at constant voltage (1,000 volts) for 30 min using phenol red as a marker. Reference amino acids were stained with ninhydrin (0.2% in acetone) and then incubated at 100 "C for 5 min. Labeled phosphoamino acids were identified by autoradiography of the cellulose plates.
Peptide Mapping-Tryptic peptide mapping of phosphorylated A-I1 heavy chain was performed as described by Sharma and Wang (1986). Phosphorylated A-I1 (10 pg) was subjected to SDS-PAGE, and protein bands were excised from the gel. Gel pieces containing A-IIt were washed three times in 50% methanol, lyophilized, and then rehydrated in 1.0 ml of 50 mM ammonium bicarbonate (pH 8.0). The gel was digested for 24 h at 37 "C with a total of 30 pg of L-1tosylamido-2-phenylethyl chloromethyl ketone-treated trypsin. Trypsin was added in three equal aliquots at time zero and after 4 and 16 h of incubation. The peptide fragments were lyophilized and suspended in buffer D, centrifuged, and dried with a gentle stream of nitrogen. The peptides were then resuspended in 20 pl of buffer D, and 2 p1 of sample was spotted onto thin layer cellulose plates (Merck). Electrophoresis in buffer D was carried out a 500 volts for 1.5 h. Ascending chromatography in the second dimension was carried out at 4.5 h in butanol/pyridine/acetic acid/water, 50:33:1:40.
Extraction of Phosphorylated Annexin-A-IIt was phosphorylated in a final volume of 4 ml as discussed above except the MgC1, concentration was reduced to 2 mM. After 20 min, 10 p1 was removed to determine the stoichiometry of phosphorylation. The reaction mixture was centrifuged at 13,000 X g, and the pellet, containing lipid and lipid-associated A-IIt, was resuspended in 25 mM Tris-HC1 (pH 7.5) and 12.5 mM EGTA. The lipid and protein mixture were sonicated in a bath type sonicator for 15 s and then incubated for 10 min at room temperature. The mixture was then centrifuged for 5 min at 13,000 X g. The supernatant, containing A-IIt, was removed and recentrifuged twice to remove residual lipid vesicles. Recovery of phosphorylated protein was determined by scintillation counting of trichloroacetic acid-precipitated protein. Control samples were incubated in the absence of ATP and recovered the same way. Recovery of phosphorylated A-IIt, which bound to and was pelleted with lipid vesicles in the presence of Ca2+, was 83% ? 6 (n = 5) as confirmed by SDS-PAGE analysis of the supernatant and pellet. The recovery of A-IIt in the EGTA-extracted and pelleted lipid vesicles, as determined by recovery of trichloroacetic acid-precipitable radioactive protein, averaged 45% -C 1 (n = 5).
Lipid Vesicle Aggregation-The lipid vesicle aggregation reaction was carried out in a final volume of 1 ml, and the final A-IIt concentration was 23 pmol/ml. Phosphorylated or control A-IIt was extracted as discussed above into 100 pl of 25 mM Tris-HC1 (pH 7.5) and 12.5 mM EGTA. This was diluted into 9.6 ml of 25 mM Tris-HCI (pH 7.5), 1 mM MgCl,, and 1.5 mM EGTA. Lipid vesicles (250 pl) were added, and the mixture was divided into 0.950-ml aliquots. Aggregation was initiated by the addition of CaC1, (50 p1 of CaCl, stocks typically ranging from 5 to 50 mM) to the required concentration and monitored by the increase in absorbance at 540 nm at 5-min intervals for a total period of 30 min. The final absorbance change after 30 min of A-IIt-induced aggregation varied between 0.140 and 0.160 absorbance units. Free Ca2+ concentrations were determined by the estimation of Fabiato and Fabiato (1979).
Phosphorylation and Aggregation Reaction-The aggregation reaction was carried out with 23 pmol of A-IIt and 0.5 pg of protein kinase C in 25 mM Tris-HC1 (pH 7.5), 2 mM MgClZ, ATP (50 p~) , and 1.5 mM EGTA at room temperature. Controls were carried out in the absence of ATP. Aggregation and phosphorylation were simultaneously initiated by the addition of CaC12. The increase in absorbance at 540 nm was monitored at 5-min intervals for 30 min. To measure the phosphorylation of A-IIt which occurred during aggregation, parallel experiments were carried out with [-y-"P] ATP, and the stoichiometry of phosphorylation measured as described above.
Miscellaneous Techniques-Protein concentration was measured using the Bradford (1976) Coomassie blue dye binding assay using bovine serum albumin as a standard. The A-IIt concentration was also determined spectrophotometrically using an extinction coeffcient of Azsonm of 0.65 for 1 mg/ml (Gerke and Weber, 1985). SDS-polyacrylamide gel electrophoresis was performed using the method of Laemmli (1970). The concentration of all CaCI2 stock solutions was determined by atomic ahsorption spectroscopy.

RESULTS
Phosphorylation of A-Ilt-A time course of the phosphorylation of A-IIt by rat brain protein kinase C is presented in Fig. 1. A-IIt was phosphorylated to 2.1 k 1.2 mol of phosphate/ mol of A-IIt (S.D., n = 19). Phosphorylation of A-IIt by protein kinase C was rapid, with half-maximal incorporation occurring within 2 min, and the reaction was essentially complete by 30 min. Phosphorylation of A-IIt by protein kinase C generated multiple forms of the heavy chain, which could be distinguished by reduced mobility on SDS-PAGE ( Fig. 1, inset A ) . As phosphorylation progressed, Coomassie Rlue staining of the protein band with identical mobility to nonphosphorylated protein (band a ) decreased in intensity, and there was a concurrent increase in the Coomassie Rlue staining intensity of bands b and c. The autoradiogram of the gel (Fig. 1, inset R ) shows that all of the Coomassie Rluestained bands were radioactive, suggesting that the timedependent increase in phosphorylation of annexin I1 heavy chain results in the generation of bands b and c. Comparison of the Coomassie Blue-stained bands a and b at 30 min ( Fig.  1, inset A ) suggests that although the staining intensity of each band is comparable, the incorporation of :"Pi into band b is greater (Fig. 1, inset R ) , inferring that the stoichiometry of phosphorylation of band b is greater than band a.
Phosphoarnino Acid and Phosphopeptides-Previous studies have suggested that protein kinase C phosphorylates A-I1 exclusively on serine residues in vivo and in vitro (Gould et al., 1986;; however, the in vitro studies have demonstrated low stoichiometry of phosphorylation (typically about 0.6 mol of Pi/mol of A-IIt). Considering the higher stoichiometry observed under optimal assay conditions ( Fig. l ) , the possibility existed that additional amino acid residues might he phosphorylated. The phosphoamino acids were analyzed, and the presence of only phosphoserine was detected (not shown). Phosphopeptide maps of SDS-PAGE bands a and b were generated (Fig. 2). In each digest, the major phosphopeptide was the same, but the relative proportions of the secondary and minor phosphopeptides were different. In digests of band a (Fig. 2 A ) , phosphopeptide 8 was the major phosphopeptide, whereas phosphopeptides 4, 7 , and 11 were secondary in intensity on autoradiograms. In digests of band h (Fig. 2R), phosphopeptide 8 was again the major phosphopeptide, whereas the relative proportion of phosphopeptide 4 to phosphopeptide 7 was increased, and a new phosphopeptide, 5 , appeared. These results suggest that the altered mobility of phosphorylated A-Ilt on SI)S-PAGE was caused by distinct phosphoforms of the protein.
Lipid Vesicle Aggregation bv Phosphonlated A-llt-To study the effect of phosphorylation on an in Llitro activity of A-IIt, we chose to examine lipid vesicle aggregation. The conditions of the aggregation assay and the composition of the vesicles used for aggregation were chosen not only for their aggregation properties but also for their ability to support protein kinase C activity. Aggregation of these vesicles, composed of PS and DAG, under our assav conditions typically resulted in an increase in light scattering (540 nm) of 5-7-fold over initial values. A typical experiment showing the effect of phosphorylation of A-IIt on the rate of lipid vesicle aggregation is shown in Fig. 3. Phosphorylation of A-IIt reduced the rate and extent of lipid vesicle aggregation. Conversely, in control experiments in which A-IIt was incubated in the phosphorylation reaction in the absence of added ATP, there was no apparent difference between it and untreated A-IIt in terms of the Ca" concentration required to induce lipid vesicle aggregation or in the rate or extent of lipid vesicle aggregation induced. Heavy chain hand n ( pnncl A 1 and hand h 1 pond H I were resnlved by SDS-PAGE, excised from gels, and suhjected to tryptic cleavage. Phosphopeptides were resolved on thin layer cellulose plates hy electrophoresis (500 V for 1.5 h) at pH 3.5 in the first dimension followed hy chromatography in the second dimension. The origin is circkrd. Electrophoresis was carried out in the horizontal dimension as indicated. The figure is representative of two separate tr>Ftic digests. T o measure lipid hinding by A-llt, the lipid vesicles were centrifuged after the aggregation reaction was complete, and the lipid-associated protein was analyzed by SIX-PAGE. The inset shows phosphorylated protein ( A ) and nonphosphorylated protein ( R ) that co-sedimented with the vesicles at pCa2+ of 6.2, 3.7, and 3.4. Fig. 4.

phorylat.ed and nonphosphorylated A-IIt is shown in
The phosphorylated protein aggregated the lipid vesicles to the same extent as nonphosphorylated protein, but higher concentrations of Ca'+ were required to induce the same level of aggregation. The Ca2+ concentration required to induce half-maximal aggregation increased from 0.18 PM 2 0.6 ( n = suggested that A-Ut-independent lipid vesicle aggregation only occurred at ca2+ concentrations greater than 6 mM ca'+.

3) for nonphosphorylated
To determine if the reduction of aggregation by phosphorylated A-IIt was because of a reduction of lipid binding activity, t h e lipid vesicle binding properties of phosphorylated and nonphosphorylated A-Ilt were examined. Following the 30min aggregat,ion reaction, the reaction mixture was centrifuged a t 100,000 x g, and the protein bound to lipid vesicles was assessed on SDS-PAGE. The inwt to Fig. 4 shows the recovery of phosphorylated protein and nonphosphorylated protein, recovered in the pellet, from lipid vesicles incubated at 0.6, 200, and 400 pM ca'+. In the absence of Ca", no protein was recovered in the pellet. Phosphorylation did not reduce the amount of protein recovered with lipid vesicles in the high speed pellet, indicating that the A-llt bound t o lipid vesicles at concentrations of Ca" that were insufficient t o support lipid vesicle aggregation.
I'hosphorylation during A,ggrcgation-To examine the effect on the A-IIt-mediated aggregation of lipid vesicles during simultaneous activation of both the phosphorylation of A-IIt by protein kinase and A-IIt-induced aggregation of lipid vesicles, A-IIt was incuhated in the presence of lipid vesicles and prot,ein kinase C, with or without added ATP (Fig. -5 ) . T h e phosphorylation and aggregation reactions were hoth initiated by the addition of CaCI2 (0.6 PM). T h e aggregation reaction was monitored spectrophotometrically. Initially, in both samples, aggregation was rapid, but as phosphorylation progressed, the relative rate of the reaction containing ATP was reduced. In parallel control experiments, in which the protein kinase C was omitted from the reaction mixture, the addition of ATP did not affect the rate or extent of vesicle aggregation. I'hosptzoplation and A-IIt Suhunit Structure--.lohnssnn et al. (1986) reported that monomeric A-I1 phosphorylated by protein kinase C resulted in multiple phosphofnrms of the protein on SDS-PAGE, but phosphorylated A -l l t prociuced only a single band on SDS-PAGE. This prments the possibility that disruption of the A-I1 tetramer upon phosphorylation by protein kinase C is responsihle for the effect nf phosphorylation on lipid vesicle aggregat inn. Therefnre, the mobility of A-IIt phosphorylated to 4 mol of phosphate/mol of protein was determined by gel permeation chromatography (Fig. 6). Phosphorylated A-IIt displayed a mobility indistinguishable from nonphosphorylated A-IIt, suggesting that the tetramer is intact following phosphorvlat ion.

DlSCllSSION
Annexin-IIt has been shown to mediate the Cn'+-dependent aggregation of phospholipid vesicles and chromaffin granules. These activities have been interpreted as evidence for a regulatory role for the protein in membrane trafficking events such as exocytosis. Although A-IIt has been identified as an in uiuo substrate of protein kinase C, the effect of phosphorylation of the protein on the activity of the protein has not been determined. In the current report we examine the effect of protein kinase C-dependent phosphorylation of A-Ilt on the lipid vesicle aggregation activity of the protein.
The stoichiometry of phosphorylation of A-IIt hy protein kinase C is much higher than our previous report (Khanna et al., 1987a. This is because of our observation that the 0.00 i . phosphorylation reaction is dependent on protein concentration; increasing the protein concentration above 0.7 nmol/ml resulted in a decrease in the stoichiometry of phosphorylation. The decline in stoichiometry of phosphorylation of A-IIt with increased protein concentration could not be reversed by an increase in the amount of lipid vesicles or by increasing the amount of protein kinase C (not shown). A-IIt was optimally phosphorylated by protein kinase C at protein concentrations below 0.7 nmol/ml. The average level of phosphorylation under our assay conditions was 2.1 +-1.2 mol of phosphate/ mol of protein ( n = 19). The light chain was not visualized on autoradiograms, suggesting that the phosphorylation sites on A-IIt were restricted to the heavy chain. Electrophoresis of phosphorylated A-IIt on SDS-polyacrylamide gels revealed two new forms of A-I1 heavy chain, with reduced electrophoretic mobility relative to nonphosphorylated heavy chain. All three forms of the heavy chain were detected upon autoradiography, suggesting multisite phosphorylation of the heavy chain of A-IIt by protein kinase C. A-I1 had previously been shown to be hyperphosphorylated by protein kinase C (Gould et al., 1986;Johnsson et al., 1986), but the heavy chain of A-IIt was not . Phosphorylation of A-IIt by protein kinase C did not induce subunit dissociation, suggesting that there are multiple phosphorylation sites on the 36-kDa protein complexed to the 11-kDa protein.
A-I1 is a substrate of both protein kinase C and pp60"" kinase, but the consequences of phosphorylation on in vitro activity have not been fully examined. Phosphorylation of A-I1 by pp60"" has been shown to reduce the Caz+-dependent binding of this protein to phosphatidylserine vesicles (Powell and Glenney, 1987). Differences in Ca2+-mediated phospholipid binding following tyrosine phosphorylation by epidermal growth factor receptor/kinase has also been demonstrated for A-I, a protein closely related to A-11. Specifically, it has been shown that phosphorylation of A-I reduced the Ca2+ required for binding to phospholipid columns (Ando et al., 1989) or to phospholipid vesicles (Schlaepfer and Haigler, 1987).
Our results suggest that the phosphorylation of A-IIt by protein kinase C decreases the rate and extent of A-IItmediated lipid vesicle aggregation by dramatically increasing the Ca2+ concentration requirement of A-IIt-mediated lipid vesicle aggregation. The Ca" concentration required to induce half-maximal lipid vesicle aggregation increased from 0.18 WM f 0.6 ( n = 3) for nonphosphorylated A-IIt to 0.65 mM f 0.4 ( n = 3) for phosphorylated protein. However, in contrast to the inhibition of lipid vesicle binding by pp6OnrC-phosphorylated A-IIt, binding of A-IIt to lipid vesicles was not decreased by phosphorylation of the protein by protein kinase C (Fig. 4,  inset). These results suggest that phosphorylation of A-IIt uncouples the lipid binding property of the protein from that of the lipid vesicle aggregation properties of the protein. The data presented in Fig. 4 clearly demonstrate that lipid vesicle binding by A-IIt can occur in the absence of lipid vesicle aggregation.
A-IIt-mediated lipid vesicle aggregation was also inhibited by the addition of protein kinase C and ATP. The decrease in the relative rate of lipid vesicle aggregation in the presence of protein kinase C and ATP did not directly parallel the incorporation of phosphate into A-IIt. Lipid aggregation in the presence of protein kinase C was essentially complete by 10 min; however, phosphate incorporation into A-IIt continued to increase for more than 20 min. The data in Fig. 5 show that although activation of phosphorylation during lipid vesicle aggregation results in inhibition of the rate and extent of aggregation, reversal of aggregation does not occur. The A-IIt population bound to lipid vesicles and participating in vesicle aggregation is either not accessible to protein kinase C, or phosphorylation of A-IIt blocks its ability to initiate lipid vesicle aggregation but does not reverse established lipid vesicle aggregation. The Ca2+ concentrations required to induce lipid vesicle binding and lipid vesicle aggregation by A-IIt are quite similar (Powell and Glenney, 1987;Blackwood and Ernst, 1990). However, the fact that phosphorylation of A-IIt inhibits the lipid vesicle aggregating activity at Ca2+ concentrations at which the A-IIt is found to bind to the lipid vesicles suggests that phosphorylation by protein kinase C differentially affects the two phenomena. A-IIt-induced lipid vesicle aggregation may be mediated by a single A-IIt protein interacting with two lipid vesicles. Alternatively a protein-protein bridge, formed by the interaction of two A-IIt molecules associated with distinct lipid vesicles, may be required to aggregate lipid vesicles. Zaks and Creutz (1991) found evidence that annexin IV, VI, or VI1 aggregates chromaffin granules by both mechanisms; protein-protein interaction was observed during chromaffin granule aggregation at low Ca2+ concentrations, whereas aggregation at high Ca2+ concentrations occurred in the absence of protein-protein interaction. The membrane, or lipid vesicle aggregation induced by A-IIt, like other annexins, may operate through distinct mechanisms at low Ca2+ and high Ca2+ concentrations. Phosphorylation of A-IIt by protein kinase C may therefore inhibit only the A-Ilt-induced lipid vesicle interaction which occurs at low Ca2+ concentrations, thus requiring increased Ca2+ concentrations to aggregate lipid vesicles with phosphorylated A-IIt. The second possibility is that a single annexin tetramer forms a bridge between two lipid. vesicles. The lipid vesicle binding sites is then reduced from two to one by protein kinase C-induced phosphorylation of A-IIt.
A-IIt and protein kinase C are both potentially involved in stimulus secretion coupling . The roles for protein kinase C in A-IIt-mediated secretion, however, are controversial. Ali and Burgoyne (1990) suggest that A-IIt can reconstitute secretion in protein-depleted adrenal cells in the presence of the protein kinase C inhibitor, staurosporin. Sarafian et al. (1991) maintain that to reconstitute secretion in permeabilized adrenal chromaffin cells, A-IIt must be phosphorylated by protein kinase C. With the highly phosphoryl-ated A-IIt, the effect of protein kinase C phosphorylation on membrane aggregation can be further investigated.