A Major, Transformation-sensitive PKC-binding Protein Is Also a PKC Substrate Involved in Cytoskeletal Remodeling*

Protein kinase C (PKC) plays a major role in regulating cell growth, transformation, and gene expression; however, identifying phosphorylation events that mediate these responses has been difficult. We expression-cloned a group of PKC-binding proteins and identified a high molecular weight, heat-soluble protein as the major PKC-binding protein in REF52 fibroblasts (Chapline, C., Mousseau, B., Ramsay, K., Duddy, S., Li, Y., Kiley, S. C., and Jaken, S. (1996) J. Biol. Chem. 271, 6417–6422). In this study, we demonstrate that this PKC-binding protein, clone 72, is also a PKC substrate in vitro and in vivo. Using a combination of phosphopeptide mapping, Edman degradation, and electrospray mass spectrometry, serine residues 283, 300, 507, and 515 were identified as the major in vitro PKC phosphorylation sites in clone 72. Phosphorylation state-selective antibodies were raised against phosphopeptides encompassing each of the four phosphorylation sites. These antibodies were used to determine that phorbol esters stimulate phosphorylation of serines 283, 300, 507, and 515 in cultured cells, indicating that clone 72 is directly phosphorylated by PKC in living cells. Phosphorylated clone 72 preferentially accumulates in membrane protrusions and ruffles, indicating that PKC activation and clone 72 phosphorylation are involved in membrane-cytoskeleton remodeling. These data lend further evidence to the model that PKCs directly interact with, phosphorylate, and modify the functions of a group of substrate proteins, STICKs (substrates that interact withC-kinase).

exocytosis. The diversity of the observed responses has made it complicated to focus on particular phosphorylation events that mediate changes in biological activity. Identifying target proteins and determining how PKC phosphorylation alters their activities are keys to understanding the role of PKC in cell regulation.
Several approaches have been used to identify PKC-binding proteins. PICKs (proteins that interact with C-kinase) are PKC substrates identified by yeast two-hybrid screening (2). RACKs (receptors for activated C-kinase) are non-substrate proteins that bind to catalytically active PKC. These may target active PKCs to the vicinity of appropriate substrate proteins (3). Several years ago, we began using a PKC overlay assay to identify PKC-interacting proteins (4 -6). We found that many of the binding proteins identified by this method, including MARCKS and the related protein MacMARCKS, were PKC substrates (5,7). We therefore named this group of PKC-binding proteins/ substrates STICKs (for substrates that interact with C-kinase. In cultured cells, phorbol esters stimulate 32 PO 4 incorporation into STICKs (7,8), indicating that PKC activation increases their phosphorylation. However, since activating PKC can also lead to increased activity of other cellular kinases, additional evidence is needed to determine if the increased phosphorylation is directly mediated by PKC. This is accomplished by comparing in vitro and in vivo phosphorylation sites in target proteins. In recent studies, we determined that two of the PKC-binding proteins we identified, ␣-adducin and the novel ␥-adducin (8), are in vivo PKC substrates by demonstrating that Ser 660 is phosphorylated by PKC in vitro and that phorbol esters stimulate Ser 660 phosphorylation in living cells (9). Increased phosphorylation in transformed renal cells in culture and in renal tumors was also noted (9,10). The major PKCbinding protein/substrate detected by overlay assays in normal REF52 fibroblasts (REF A cells) is a Ͼ200-kDa heat-soluble PKC substrate referred to as clone 72 (11) and also as SSeCKS (12,13). Clone 72 is also closely related to gravin, an autoantigen identified in patients with myasthenia gravis (14). Clone 72 protein and message levels are down-modulated in oncogene-transformed cells, indicating that this protein may play an important role in normal growth regulation. In fact, reexpressing clone 72 (SSeCKS) in transformed cells has been shown to attenuate growth and transformation (15). The goal of the present studies was to determine if clone 72 is an in vivo PKC substrate by comparing in vitro and in vivo sites of phosphorylation. Our results indicate that clone 72 is phosphorylated by PKC at multiple sites in vitro and in vivo. These studies extend our evidence that PKC directly binds to and phosphorylates a group of target proteins (STICKs). These results contribute to the emerging model that STICKs are primary targets for PKC phosphorylation and that phosphorylation of STICKs is directly related to the role of PKC in regulating the complex processes of cell growth, death, and differentiation.

MATERIALS AND METHODS
Preparation of Clone 72 Mutants-Site-directed mutagenesis was used to introduce HindIII and Xma sites at the 5Ј-and 3Ј-ends of the binding and phosphorylation domain encompassing amino acid residues 496 -531 (Table I). In this process, Val 497 in the wild type sequence was converted to Leu. This conservative change did not alter PKC binding, PS binding, or PKC phosphorylation. The HindIII and Xma sites were used to insert double-stranded synthetic oligonucleotides that encode for the deletion mutant sequences 1-4 (Table I). Sitedirected mutagenesis was also used to prepare the single and triple amino acid changes in mutants 5-7 (Table I). Mutagenesis was done with double-stranded plasmid DNA using the Chameleon mutagenesis kit (Stratagene). Mutant proteins were expressed from pTrc (Invitrogen) as His-tagged fusion proteins. All sequences were verified by sequence analysis.
Phosphorylation Reactions-For quantitative measurements, purified, recombinant proteins and peptides were phosphorylated with purified recombinant PKC␣ (PanVera) at 30°C for 15 min using standard conditions (4). Phosphorylated proteins were separated on SDS-polyacrylamide gels and detected by autoradiography. For maximum phosphorylation needed for identifying phosphorylated residues, synthetic peptide C34E designed from clone 72 residues (Lys 496 -Glu 529 , with an N-terminal cysteine added to facilitate coupling) (3.5 M) was phosphorylated with PKC␣ (13.5 nM) for 2-4 h until maximum phosphorylation was reached as monitored by incorporation of radiolabeled phosphate. The phosphorylated synthetic peptide (6.25 g) was digested with 0.6 g of endoproteinase Lys-C overnight. The digested peptides were analyzed by liquid chromatography electrospray mass spectrometry (LC-ESMS). Synthetic peptide corresponding to clone 72 residues Arg 296 -Lys 305 (10 M) was phosphorylated with PKC␣ (3.0 nM). The undigested peptide was analyzed by LC-ESMS.
Peptide Synthesis-Automated FMOC solid phase synthesis was used to produce PKC phosphorylation substrate peptides R10K and C34E as well as calibration and tuning reagents for the mass spectrometer (peptides G9I, K13L, L20R, V26E, E10A, K6R, and bA5K). Peptides G9I (GRTGRRNSI), K13L (KTREGNVRVSREL), L20R (LKPVK-KKKIKREIKILENLR), V26E (VAMKKIRLEEEGVPSTAIREISLLKE), and C34E (CKVQSPLKKLFSSSGLKKLSGKKQKGKRGGGGDE) were synthesized at the 250-mol scale on HMP (Wang) resin using a Perkin-Elmer Applied Biosystems model 430A Peptide Synthesizer and the Applied Biosystems FastMoc synthesizer program, cleaved from the solid support with trifluoroacetic acid, and extracted with acetic acid as described previously (16). Phosphopeptides E10A (EPQpYEEIPIA-NH 2 ) (where pY represents phosphorylated tyrosine) and K6R (KRp-TIRR-NH 2 ) (where pT represents phosphorylated threonine) were also prepared on the model 430A instrument (300-mol scale) using the FastMoc program, Fmoc-phosphotyrosine (PO 3 H 2 ) Ϫ OH, Fmoc-phosphothreonine (O-benzyl), and Fmoc-4-methylbenzylhydrylamine (MBHA) rink amide resin. PKC substrate peptide R10K (RKKTSFKKSK) (100mol scale) and phosphopeptide bA5K (biotin-ApSSYK) (where pS represents phosphorylated serine) (50-mol scale) were synthesized with a Rainin Symphony/Multiplex Peptide Synthesizer using the synthesizer software (version 3.2) provided by the instrument manufacturer (Protein Technologies, Inc.) and Fmoc-phosphoserine (O-benzyl). Following automatic cleavage from the solid support on the Symphony instrument, peptides were precipitated from trifluoroacetic acid with ice-cold ether, washed two or three times with cold ether, and either vacuumdried or purified by RP-HPLC. Extended double coupling steps were used for Fmoc-phosphoamino acids on both instruments; two equimolar additions of the Fmoc-phosphoamino acid were made approximately 6 -12 h apart with a total coupling reaction time of 18 h. Biotin was coupled with a single coupling reaction of 12 h. Fmoc-phosphoamino acids, D-biotin, Wang resins (with or without preloaded Fmoc amino acids), and Fmoc-MBHA rink amide resin were from AnaSpec. Reagents and chemicals used with the Symphony instrument were from Rainin, and those used with the model 430A instrument were from PE Applied Biosystems and Burdick and Jackson. Synthetic peptides used for mass spectrometry tuning and calibration were purified by RP-HPLC using aqueous trifluoroacetic acid/acetonitrile solvents and a 10 Vydac C18 column (2.2 ϫ 25 cm). All synthetic peptides were quality control-analyzed by analytical RP-HPLC and amino acid analysis (17) in addition to ESMS. Phenylthiocarbamyl amino acid analysis was performed as described previously using an automated analysis system (Applied Biosystems model 420H/130A/920) (18).
Electrospray Mass Spectrometry-ESMS, LC-ESMS, and low energy collision tandem MS (MS/MS) were performed with a Perkin-Elmer Sciex API 300 triple quadrupole mass spectrometer (Thornhill, Canada) fitted with articulated ion spray plenums and atmospheric pressure ionization sources. Initial tuning and calibration was with a standard mixture of polypropylene glycol provided by PE Sciex. Resolution was adjusted to about 50% valley between adjacent isotope peaks in a singly charged cluster, allowing singly charged ions to be identified by apparent spacing between peaks and doubly charged ions to be distinguished from those with higher charge states. The nebulization gas for negative ion monitoring was bottled air (Medical Grade USP, Merriam-Graves, Claremont, NH), and for positive ion monitoring nitrogen was used at 40 p.s.i. Nitrogen was also used as the curtain gas and was supplied from a Dewar flask (XL-45, Taylor-Wharton) of liquid nitrogen (Merriam-Graves). For infusions in positive ion mode, spectra were acquired at an orifice potential of 60 V over a scan range of m/z 300 -1500 using 0.25-atomic mass unit steps and a total scan time of 4.8 s. For negative ion monitoring, the ion spray needle voltage was operated at Ϫ4.0 kV and the focusing ring at Ϫ230 V; for positive ion monitoring, the ion spray needle voltage was 4.3 kV and the focusing ring was at 300 V. Low energy collision MS/MS was performed with the RO2 voltage set between Ϫ55 and Ϫ70 V, nitrogen as the collision gas, and CAD gas settings between 3 and 4, and 50 -70 scans were accumulated. Parent ion transmission was maximized by setting Q1 to transmit about a 2-3 m/z window around the selected ion.
LC-ESMS with selective ion monitoring for PO 3 Ϫ at m/z 79 was performed in negative ion mode at high orifice potential (Ϫ190 V) over a narrow scan range of m/z 78.8 -79.2 using 0.1-atomic mass unit steps, a dwell time of 100 ms/step, and a total scan time of 0.5 s. In the same analysis, full scans in positive ion mode were acquired at a lower orifice potential (60 V) over the scan range m/z 300 -2000 using 0.25-atomic mass unit steps and a total scan time of 4 s. This was accomplished by looping the negative ion and positive ion experiments together with computer-controlled polarity reversal between each scan. The higher orifice potential in negative ion mode enhances collision-induced decomposition of phosphopeptides, producing the diagnostic PO 3 Ϫ fragment ion, thus allowing selective detection of phosphopeptides (19). The lower orifice potential used for full scans in positive ion mode allowed acquisition of mass spectra for all peptides (19). Seven synthetic peptides, including three phosphopeptides containing phosphotyrosine, phosphoserine, or phosphothreonine, were used to optimize negative ion detection sensitivity for the m/z 79 fragment ion or to serve as internal controls for PO 3 Ϫ monitoring during LC-ESMS or to verify correct calibration during LC-ESMS. Synthetic peptides used for mass spectrometry include G9I (M ϭ 1015. For LC-ESMS, the HPLC eluant was split, with 22% going to the mass spectrometer and the remainder collected in 1-min fractions. RP-HPLC for LC-ESMS was performed at a flow rate of 50 l/min on a 5 Vydac C18 microbore column (1 ϫ 250 mm) using an Applied Biosystems model 120A HPLC system equipped with a 75-l dynamic mixer and aqueous acetonitrile/trifluoroacetic acid solvents (17). Portions of select HPLC fractions were diluted 1:3 in 0.2% formic acid, 50% methanol and infused at 2 l/min for MS/MS analysis.
Clone 72 Antisera-Four phospho-serine peptides containing the phosphorylation sites identified in in vitro reactions were synthesized with N-terminal cysteines added to facilitate coupling to keyhole limpet hemocyanin and/or solid supports. The peptides were as follows: Asn 276 -Lys 286 included phosphoserine 283; Arg 296 -Lys 305 included phosphoserine 300; Lys 503 -Lys 512 included phosphoserine 507; and Leu 511 -Arg 523 included phosphoserine 515. Antisera were raised in rabbits and purified against the cognate peptides coupled to agarose.
Immunoblotting-REF52 cells were grown, and cell lysates were prepared as described previously (7). Aliquots (25-50 g of protein) were electrophoresed on SDS-PAGE gels and transferred to nitrocellulose. Blots were probed with immunopurified antisera preparations under standard conditions, and reactive bands were detected with ECL reagents. The phosphorylation state-insensitive antibody was raised to the recombinant fusion protein containing amino acid residues 388 -790 (clone 72 original) (11).
Immunofluorescence-REF52 cells were grown on glass coverslips. Cells were fixed in formaldehyde and permeabilized with methanol as described previously (20). Cells were incubated with affinity-purified primary antibodies, washed, and then incubated with FITC-conjugated species-specific secondary antibodies. Photomicrographs were taken on a Nikon Optiphot microscope.

RESULTS
Mapping PKC Phosphorylation Sites-In previous studies, we prepared eight overlapping clone 72 fragments that span the entire open reading frame. Of the eight recombinant proteins, only fragments 1, 2, and 3 (from the N terminus) were phosphorylated in vitro with PKC. 2 To identify sites of phosphorylation, the phosphorylated fragments were digested with proteases, and phosphopeptides were isolated by HPLC and identified by Edman degradation. Those studies identified three phosphopeptides containing putative PKC phosphorylation sites, Thr 280 -Phe 288 , Arg 296 -Asp 308 , and Arg 494 -Glu 530 . We have now extended these studies using mass spectrometry to identify the PKC phosphorylation sites in these peptides. The synthetic peptide C34E (designed from residues Lys 496 -Glu 529 with cysteine added) was phosphorylated with PKC and then digested with endoproteinase Lys-C. Identified peptides with their corresponding retention times and masses are summarized in Table I. Several lysyl peptides (representing partially digested fragments) containing a single serine residue (Ser 20 in peptide C34E) corresponding to serine 515 in clone 72 showed an increased mass of 80 Da, indicative of phosphate incorporation. Several lysyl peptides containing serine residues 507-509 also showed an 80-Da increase in mass, indicating that one of these serine residues is also a target. Observed masses of lysyl peptides containing serine 500 correlated with calculated masses; therefore, serine 500 is not a PKC target (Table I).
To determine which of serines 507-509 were phosphorylated, lysyl peptide LFSSSGLK (corresponding to residues 505-512 from clone 72) was further analyzed by low energy collision MS/MS (Fig. 1). The sequence was determined from the loss in fragment mass starting from the C terminus (y fragments) or N terminus (b fragments). This analysis confirms that the serine corresponding to Ser 507 in clone 72 is phosphorylated. Thus, Ser 507 and Ser 515 , but not Ser 500 are in vitro PKC phosphorylation sites in fragment 3 (Met 387 -Ile 790 ).
Phosphorylated residues in the other phosphorylation site found in fragment 2 (Gln 261 -Gly 532 ) of clone 72 were identified by phosphorylating the synthetic peptide corresponding to residues Arg 296 -Lys 305 . Phosphorylation resulted in an observed mass increase of 80 Da, indicating that only one of the three serines and threonines is the preferred site (Fig. 2, A and B). Low energy collision MS/MS analysis showed that phosphorylation occurred at serine 300 (Fig. 2C).
Comparison of PKC Binding and Phosphorylation Domains-Several mutants within the fragment 3 phosphorylation site were constructed to analyze the structural requirements for PKC binding and phosphorylation (Table II). Fragment 3 and a subfragment (subfragment 3, Ser 370 -Glu 620 ) had quantitatively similar PKC binding activity, indicating that the PKC binding motif is contained within this region (Fig.  3). Mutating the three tandem serine residues to alanines (S507-509A) did not influence binding, indicating that these residues are not critical for PKC binding. Phenylalanine residues in the MARCKS phosphorylation domain have been shown to be important for hydrophobic interactions with PS at the membrane interface (21). Replacing phenylalanine 506 with valine (F506V) severely attenuated PKC binding and PS binding (Fig. 3). Thus, interactions of the clone 72 and MARCKS phosphorylation domain peptides with PS may be somewhat analogous. Structural changes in this motif were introduced by replacing leucine 511 with proline (L511P). According to Chou-Fasman and Garnier-Robson rules, this mutation is predicted to decrease ␣-structure and increase ␤-structure. This mutation also coordinately decreased PKC and PS binding. These results identify the Lys 497 -Pro 511 region as a major determinant of phospholipid-dependent PKC binding. Neither of these mutations influenced the net positive charge of this domain, indicating that PKC and PS binding are primarily regulated by hydrophobic interactions and/or secondary structure.
As shown in Fig. 4, phosphorylation of S507-509A relative to the wild type sequence was decreased, which verifies that these residues are among the PKC targets. Surprisingly, the F506V and L511P mutants were relatively better PKC substrates than the wild type protein, despite their decreased PKC and PS binding. Kinetic analysis of L511P mutant phosphorylation demonstrated that this was due to an increased V max with no significant effect on K m (Fig. 5). The V max value was increased 2.2-fold for the F506V mutant compared with subfragment 3. Differences in K m values were Ͻ20%. One potential explanation for this result is that phosphorylatable residues in the wild type sequence are partially masked, and mutations in this region increase accessibility to PKC. Clone 72 Is a PKC Substrate in Vivo-To determine if the PKC phosphorylation sites identified in in vitro studies were actual sites in vivo, antisera were raised to phosphopeptides containing the putative phosphorylation sites Ser 300 , Ser 507 , and Ser 515 . In addition, an antiserum to Ser 283 was prepared, since preliminary in vitro phosphorylation studies identified Thr 280 -Phe 289 as a phosphopeptide. Antisera were affinitypurified against the cognate phosphopeptides. Specificities of the antibodies were tested by comparing their reactivities against the phosphorylated and unphosphorylated forms of the immunizing peptides. As shown in Fig. 6, each of the antibodies preferentially recognizes the corresponding phosphopeptide compared with the unphosphorylated peptide. The antibodies did not cross-react with unphosphorylated recombinant clone 72 proteins. Furthermore, each of the antibodies specifically recognizes its cognate peptide and does not cross-react with other Ser(P)-containing peptides. Thus, these antibodies recognize context-specific Ser(P) and not Ser(P) alone. These data indicate that these antibodies are phosphorylation state-selective and can be used to monitor clone 72 phosphorylation at selected residues.
To determine if clone 72 is phosphorylated by PKC in vivo, we prepared extracts from REF52 cells that had been treated with the PKC activator, PDBu. As shown in Fig. 7, phorbol esters stimulated phosphorylation at serines 283, 300, 507, and 515. The relative increases in antibody binding were determined by densitometry of immunoblots and were 634 Ϯ 351, 297 Ϯ 185, 1996 Ϯ 875, and 783 Ϯ 300 (means Ϯ S.D., n ϭ 4) for phosphoserines 283, 300, 507, and 515, respectively. The correlation between the phosphorylation sites identified in in vitro and in vivo assays, indicates that clone 72 is an endogenous PKC substrate. Clone 72 and Phosphorylated Clone 72 Subcellular Location-To gain insight into the function of clone 72 and the effects of PKC phosphorylation, we compared the localization of clone 72 and phosphorylated clone 72 by immunofluorescence. Each of the clone 72 antibodies used for immunofluorescence preferentially recognized only clone 72 on immunoblots of REF52 cell extracts (Fig. 8A), which demonstrates their suitability for immunolocalization studies. In unstimulated cells, total clone 72 (monitored with a phosphorylation state nonspecific antibody) is distributed throughout the cytoplasm and concentrated in membrane ruffles (Fig. 8B). In contrast, phosphoserine 300 and 515 antibodies did not detect substantial cytoplasmic staining. With these antibodies, staining was concentrated in membrane protrusions and ruffles. These data indicate that under physiological conditions, PKC activation and clone 72 phosphorylation are preferentially localized to the cell perimeter. The prominent nuclear staining observed with  Fig. 6 at 34.8 min (MH ϩ ϭ 918.5) was further analyzed by MS/MS. Fragment ion nomenclature is as described by Biemann (24). M nϩ ϭ (M ϩ nH) nϩ ; the asterisks indicate phosphate losses from the next higher mass (without H 3 PO 4 (*) and without HPO 3 (**)). pS, phosphorylated serine. the phosphoserine 515 antibody is not seen with other antibodies and appears to be an artifact of this particular antibody. PDBu caused a redistribution of total clone 72 toward the perinuclear space but did not substantially influence the staining intensity. In contrast, PDBu substantially increased the immunostaining seen with the phosphoserine 300 and 515 antibodies. Staining was concentrated in the perinuclear region of the cytoplasm but was also apparent in membrane protrusions and ruffles. PKC isozymes expressed in these cells (␣, ␦, ⑀, and ) are also found in these locations. However, the broad distribution patterns of PKCs and clone 72 do not permit the detailed comparison required to assess colocalization. Other approaches will be required to determine which PKCs interact with and phosphorylate clone 72 in vivo.
Since PDBu is cell-permeable, it can activate PKCs throughout the cell, and its actions are not limited to the cell surface. To investigate physiological PKC activation and clone 72 phospho-rylation, cells grown on coverslips were artificially wounded by scraping the monolayer with a sharp object. Cells were then stained with antibodies to actin and clone 72 (Fig. 9). The Wild type and mutant cDNAs were expressed in bacteria as His 6 -tagged proteins and purified by nickel affinity chromatography. Equal amounts of recombinant proteins were bound to nitrocellulose using a slot blot apparatus. In the clone 72 lane, the nitrocellulose was stained with affinity-purified clone 72 antibody to compare the amounts of the recombinant proteins. In the PKC overlay lane, PKC binding was monitored using the overlay assay (5,6). In the PS overlay lane, phosphatidylserine (PS) binding was monitored using 14 C-PS. Fragment 3 (72 original), residues Met 387 -Ile 790 ; subfragment 3, residues Ser 370 -Glu 620 . F506V, L511P, and S507-509A are mutants of subfragment 3 shown in Table II. double staining clearly demonstrates that phosphorylated clone 72 is concentrated in membrane ruffles and regions in which actin is depolymerized. These results demonstrate a strong correlation among PKC activation, clone 72 phosphorylation, and dynamic actin remodeling. DISCUSSION PKC phosphorylates a wide variety of proteins in in vitro reactions, many of which are not relevant physiological substrates in vivo. The relatively limited number of PKC phosphorylation events that occur in vivo indicates that PKC access to substrates is regulated in cells. One mechanism for limiting PKC phosphorylations to appropriate substrate proteins would be through direct, high affinity interactions of PKC with its substrates. We have cloned several PKC-interacting proteins based on their ability to bind PKC in an in vitro assay (4). Binding proteins identified by this method include the PKC substrates MARCKS and the related protein MacMARCKS (22) and two forms of adducin, ␣-adducin and the novel ␥-adducin (8). We named this class of PKC binding proteins/substrates STICKs (for substrates that interact with C-kinase.
Studying the effects of PKC phosphorylation on STICK functions provides an avenue for understanding the mechanisms by which PKC regulates cell growth, death, and gene expression. In the present study, we demonstrated that clone 72, the major PKC-binding protein in REF52 cells, is phosphorylated by PKC at the same sites in vitro and in vivo. In other studies, we found that overexpressing certain PKCs increases clone 72 phosphorylation, whereas inhibiting PKC with bisindoylmaleimide I decreases clone 72 phosphorylation. 3 We conclude that PKC is a clone 72 kinase. The selective localization of phosphorylated clone 72 compared with total clone 72 indicates that phosphorylation occurs at discrete sites and may be either a consequence or an effector of localized actin remodeling involved in formation of membrane protrusions and ruffles.
Clone 72 is related to gravin, which was also recently shown to be a PKC-binding protein in vitro (14). Gravin is Ͼ80% identical to clone 72 in the PKC binding domains/phosphoryl-  6. Characterization of phosphorylation state selective antisera. Antisera raised to phosphopeptides containing the PKC phosphorylation sites were affinity-purified against the cognate phosphopeptides. Purified antibodies (Ab) (1 g/ml) were reacted with 250 ng of the phosphopeptide or the unphosphorylated peptide slot blotted onto nitrocellulose. For phosphoserines 300, 507, and 515, the unphosphorylated peptide was a corresponding synthetic sequence as indicated. For phosphoserine 283, the unphosphorylated peptide was derived from the phosphopeptide after exhaustive digestion with alkaline phosphatase.

FIG. 7. Clone 72 is phosphorylated by PKC in vivo.
Lysates were prepared from control and PDBu-treated (100 nM, 10 min) cultures of REF52 cells. Aliquots (approximately 50 g of protein) were electrophoresed and blotted to nitrocellulose. Replicate blots were stained with the indicated affinity-purified antibodies. The phosphorylation stateinsensitive antibody was raised to recombinant fragment 3 (clone 72 original).
ation sites that we have identified in clone 72. Studies with gravin failed to detect any evidence for an in vivo complex formation between gravin and PKC (14). However, since gravin is highly homologous to clone 72 in the PKC phosphorylation sites/binding domains, it seems likely that gravin is also a PKC substrate that associates (at least transiently) with PKC in vivo. Although a fragment of gravin was reported to inhibit PKC phosphorylation of an exogenous peptide substrate, since clone 72 and (by analogy) gravin are PKC substrates, it seems likely that emphasis should be placed on the role of clone 72 and gravin as PKC substrates rather than inhibitors.
Mutational analysis was used in structure-function studies to define the PKC binding site in the fragment 3 region of clone 72. We found that the F506V and L511P mutants had decreased PS and PKC binding activity yet were improved PKC substrates due to an increase in V max . These results lend support for the model proposed by Ohno et al. (23) to explain why MARCKS is a high affinity (low K m ) yet slowly phosphorylated (low V max ) PKC substrate. It is possible that high affinity interactions of PKC with substrates are accomplished at the expense of the rate of the phosphorylation reaction. For example, the decreased binding of the F506V and L511P mutants would facilitate dissociation of product and increase enzyme turnover. Given the high affinity of PKC for STICKs and the less than optimal phosphorylation efficiency, it is possible that PKC-STICK interactions are important for localizing PKC.
In summary, these data provide additional evidence that PKC directly interacts with and phosphorylates a class of substrate proteins named STICKs. Thus, STICKs have two important activities: PKC binding partners and PKC substrates. Phosphorylation of certain STICKs, such as MARCKS and adducins, alters their functions with regard to association with cytoskeletal proteins and membranes (20,21). Thus, STICKs FIG. 9. Comparison of phosphorylated clone 72 and actin immunostaining. REF52 cells were grown on glass coverslips to confluence. The monolayer was artificially wounded by scraping with a sharp object. After 4 -8 h, coverslips were fixed and stained with mouse monoclonal antibody to ␤-actin and affinity-purified clone 72 rabbit antibodies. After washing, coverslips were stained with fluorescein-and rhodamine-conjugated species-specific secondary antibodies to compare actin and phosphorylated clone 72 subcellular locations.
FIG. 8. Immunofluorescence localization of clone 72 and phosphorylated clone 72. A, lysates were prepared from control and PDBu-treated cultures (200 nM, 10 min). Blots were stained with the phosphorylation state-insensitive clone 72 antibody (Ab) (Total) and the phosphorylation state-selective phosphoserine 300 and 515 antibodies (pSer300 and pSer525, respectively). Full-length, long exposure blots are shown to verify single band specificity and the suitability of these antibodies for immunofluorescence studies. B, REF52 cells were grown on glass coverslips, fixed, and stained as described. Where indicated, cells were treated with PDBu (200 nM) for 10 min.
are also effectors of PKC activation. Additional studies of PKC-STICK interactions and phosphorylation-induced changes in STICK functions should provide new insight into the initial events in PKC activation that contribute to downstream effects on cell growth, transformation, and gene expression.