p32 (gC1qBP) Is a General Protein Kinase C (PKC)-binding Protein

The aim of this study was to identify cellular proteins that bind protein kinase C (PKC) and may influence its activity and its localization. A 32-kDa PKC-binding protein was purified to homogeneity from the Triton X-100-insoluble fraction obtained from hepatocytes homogenates. The protein was identified by NH2-terminal amino acid sequencing as the previously described mature form of p32 (gC1qR). Recombinant p32 was expressed as a glutathione S-transferase fusion protein, affinity-purified, and tested for an in vitro interaction with PKC using an overlay assay approach. All PKC isoforms expressed in rat hepatocytes interacted in vitro with p32, but the binding dependence on PKC activators was different for each one. Whereas PKCδ only binds to p32 in the presence of PKC activators, PKCζ and PKCα increase their binding when they are in the activated form. Other PKC isoforms such as β, ε, and θ bind equally well to p32 regardless of the presence of PKC activators, and PKCμ binds even better in their absence. It was also found that p32 is not a substrate for any of the PKC isoforms tested, but interestingly, its presence had a stimulatory effect (2-fold for PKCδ) on PKC activity. We also observed in vivo interaction between PKC and p32 by immunofluorescence and confocal microscopy. A time course of phorbol ester treatment of cultured rat hepatocytes (C9 cells) showed that PKCθ and p32 are constitutively associated in vivo, whereas PKCδ activation is required for its association with p32. Our data also showed that phorbol ester treatment induces a transient translocation of p32 from the cytoplasm to the cell nucleus. Together, these findings suggest that p32 may be a regulator of PKC location and function.

Protein kinase C (PKC) 1 comprises a large family of serine/ threonine kinases that play key regulatory roles in growth, differentiation, cell survival, neurotransmission and carcinogenesis (1). Based on structural characteristics that determine differences in their lipid and calcium binding properties, PKC are divided into three groups: conventional, novel, and atypical PKC (2). These differences do not, however, adequately account for the isozyme specificity observed in vivo, because overlapping specificity of these kinases is commonly observed in vitro. On the other hand, the study of PKC function is complicated by the existence of many isoforms within the same cellular type that redistribute from soluble to membrane-bound intracellular compartments as a result of their activation (3)(4)(5).
Recent observations have revealed protein kinase-anchoring and scaffold proteins as essential elements in cell signaling (6 -8). Given the dynamic nature of protein phosphorylation reactions, coordinated control of both kinases and phosphatases is often required. Cells seem to have solved this problem by utilizing anchoring proteins that bind to subcellular structures and localize their complement of enzymes/adaptor proteins close to their sites of action. Indeed, genetic, biochemical, and cell-based screens have revealed a heterogeneous group of PKC-binding proteins that can influence its subcellular localization, substrate availability, exposure to allosteric activators, and activation-dependent relocation within cells. This group includes the following: substrate-binding proteins that interact with PKC prior to phosphorylation (9); cytoskeletal/vesicle interacting proteins (10 -12); AKAPs, named so as a result of their original discovery as cAMP-dependent (protein kinase A)-anchoring proteins (13); and a class of proteins referred to as RACKs (receptors for activated C-kinase) (14 -16).
Using overlay assays, we have identified in rat hepatocytes several proteins that interact with PKC (16,17) and purified some of them from the Triton X-100-insoluble fraction of these cells to apparent homogeneity (18). In the current work, we purified a 32-kDa PKC-interacting glycoprotein and identified it as the previously described receptor for the globular "heads" of complement component C1q (gC1q-R) (19). Originally, p32 was characterized as a component of the AS/SF2 splicing activity purified from HeLa cells (20). Subsequent studies have shown that the molecule interacts with several cellular and viral proteins that participate in mitochondrial function (21,22), transcription, and splicing factor modulation (23,24). Since the original study, it has become apparent that p32 is an evolutionarily conserved eukaryotic protein. Homologous genes have been identified in a number of eukaryotic species ranging from fungi to mammals. Recently, the interaction of p32 with * This research was supported in part by Grants 31714-N and 28083N from Consejo Nacional de Ciencia y Tecnología (CONACYT) and IN206298 from Dirección General de Asuntos del Personal Académico (DGAPA). The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
The nucleotide sequence(s) reported in this paper has been submitted to the GenBank TM /EBI Data Bank with accession number(s) AF300619 (p32 (gC1qBP)).
§ To whom correspondence should be addressed: PKC has been reported (25). In the B cell line SKW 6.4, p32 appears to function as a compartment-specific negative regulator of the kinase activity of PKC. Here we present evidence that p32 interacts in vitro with all PKC isoforms expressed in rat hepatocytes and stimulates PKC activity. We also show that PKC and p32 are constitutively associated in vivo and that treatment of intact cells with phorbol ester promotes p32-PKC␦ in vivo interaction and a transient redistribution of p32 to the cell nucleus.

EXPERIMENTAL PROCEDURES
Materials and Antibodies-Phorbol 12-myristate 13-acetate, histone H1-IIIS, phosphatidylserine, Triton X-100, and 1,2-diolein were from Sigma. Protein A-Sepharose and Ro 31-8220 were from Calbiochem, and DEAE-cellulose (DE-52) was from Whatman. [␥-32 P]ATP (6000 Ci/mmol) was from PerkinElmer Life Sciences. To obtain polyclonal antisera against p32, New Zealand rabbits were injected subcutaneously with 100 g of recombinant GST-p32 fusion protein in 500 l of PBS plus 1 volume of complete (or incomplete for boosts) Freund's adjuvant. Three further immunizations were done before the rabbits were bled. Isozyme-specific polyclonal antibodies against the COOH terminus of PKC isozymes were purchased from Santa Cruz Biotechnology Inc. Monoclonal antibodies against PKC and PKC␦ were obtained from Transduction Laboratories (Becton Dickinson). Goat antirabbit IgG-alkaline phosphatase conjugate was from Bio-Rad. FITCconjugated goat anti-rabbit and TRITC-conjugated goat anti-mouse antibodies were from Zymed Laboratories Inc.
Isolation of Rat Hepatocytes and Cell Culture-Hepatocytes were isolated from male Wistar rats (200 -250 g) by the method of Berry and Friend (26). For immunofluorescence studies, C9 cells (a rat hepatocyte line obtained from the American Type Culture Collection, Manassas, VA) were maintained in F12K medium (Invitrogen) supplemented with 10% veal serum (with iron), 1% insulin (Humulin*R, from Lilly) and 1% penicillin/streptomycin (10000 units/g/ml, from Invitro) and grown in a humidified atmosphere of 5% CO 2 and 95% air at 37°C.
Purification of p32 and Protein Sequencing-p32 was purified to homogeneity from the Triton X-100-insoluble fraction of rat hepatocytes homogenates as described previously (27). NH 2 -terminal sequence analysis was carried out on purified protein electroblotted on polyvinylidene difluoride membrane with a model LF3000 protein sequencer (Beckman) using chemicals and software supplied by the manufacturer.
Detection of Carbohydrates in Purified p32-The digoxigenin glycan detection kit from Roche Molecular Biochemicals was used according to the manufacturer's instructions. The principle of this technique is that adjacent hydroxyl groups in sugars of glycoconjugates are oxidized to aldehyde groups by mild periodate treatment. The spacer-linked steroid hapten digoxigenin is then covalently attached to these aldehydes via a hydrazide group. Digoxigenin-labeled glycoconjugates are subsequently detected in an enzyme immunoassay using a digoxigenin-specific antibody conjugated with alkaline phosphatase.
Cloning of p32 cDNA-A clone encoding p32(gC1q-R), derived from a mouse embryo cDNA library, was found by searching the EST (expressed sequence tag) data base (IMAGE clone 478889, GenBank TM accession number AA048814). This plasmid was purchased from Research Genetics Inc., and the nucleotide sequence of its insert was determined by the chain termination method (28). For the construction of the GST-p32 fusion protein, the 618-bp fragment resulting from EcoRI (SI nuclease trimmed) and DraI digestion of the p32 cDNA was subcloned at the SmaI site of the pGEX2T vector (Amersham Biosciences, Inc.). The proper orientation of the subcloned insert was verified by restriction digests, and the correct in-frame insertion was verified by nucleotide sequencing. Expression of the GST-p32 fusion protein was done in the protease-deficient BL21 strain of Escherichia coli (Novagen, Madison, WI), and its affinity purification was performed as described by Smith and Corcoran (29).
The eluate was concentrated with an Amicon device (YM-30 membrane) and used for overlay assays.
Determination of PKC Binding Activity by an Overlay Assay-PKC binding was determined as described previously (16). Briefly, sample protein was separated by SDS-PAGE (10% gels) and transferred to nitrocellulose membranes, which were then blocked with 3% (w/v) BSA. Partially purified and concentrated PKC (100 -200 units; 1 unit is defined as 1 pmol of 32 P transferred from [␥-32 P]ATP to histone/min/mg of protein; final protein concentration, 0.5 mg/ml) was incubated for 1 h at 30°C with the nitrocellulose strips in blotting buffer (50 mM Tris-HCl, pH 7.5, 1% polyethylene glycol, and 0.2 M NaCl) in the presence of cofactors (20 g/ml phosphatidylserine, 0.8 g/ml 1,2-diolein, and 1 mM CaCl 2 ). The PKC/blotting buffer mixture was removed, and the membranes were washed three times with blotting buffer. Bound PKC was detected using isozyme-specific anti-PKC antibodies and secondary goat anti-mouse or goat anti-rabbit IgG antibodies conjugated to alkaline phosphatase. Quantification of binding was performed by densitometric analysis using an Image Densitometer (BioRad GS-670).
PKC Immune Complex Kinase Assay-This assay was performed as described elsewhere (18). Briefly, aliquots of 1 ml of partially purified and concentrated PKC (1 mg/ml) in the presence of phosphatase inhibitors (10 mM ␤-glycerophosphate, 1 mM Na 3 VO 4 , 11 mM NaF, 10 mM sodium pyrophosphate, and 0.2 mg/ml phosphoserine) and in the absence of 2-mercaptoethanol were incubated at 4°C with 1 g/ml isozyme-specific PKC antibody for 1 h with gentle shaking. Then, 20 l of protein A-Sepharose (30%, Calbiochem) were added, and incubation was continued for 1 h. Immune complexes were then washed three times with buffer A (50 mM Tris-HCl, 0.6 M NaCl, 1% Triton X-100, 0.5% Nonidet P-40, pH 8.3) supplemented with 0.1 mg/ml trypsin inhibitor and 1 mM phenylmethylsulfonyl fluoride and once with kinase buffer (20 mM Tris-HCl, pH 7.5, 10 mM MgCl 2 , 0.5 mM CaCl 2 , 50 mM 2-mercaptoethanol). Kinase activity was initiated by resuspending the immunoprecipitates in 50 l of assay mixture (kinase buffer plus 20 g/ml phosphatidylserine, 0.8 g/ml 1,2-diolein, 10 M [␥-32 P]ATP (6000 Ci/ mmol), and 200 g/ml of protein substrate (histone H1-IIIS or recombinant GST-p32)). Reactions were allowed to proceed for 20 min at 30°C, terminated by the addition of 50 l of SDS-PAGE sample buffer, boiled for 5 min and analyzed by 12.5% SDS-PAGE and autoradiography. Data were quantified by densitometric analysis performed both in Coomassie-stained gels and the corresponding autoradiographies. The ratio of 32 P-labeled protein/dyed protein represents the total specific phosphorylation.
Immunofluorescence-C9 monolayers grown on glass coverslips were rinsed twice with PBS, fixed with methanol at Ϫ20°C for 10 min, and permeabilized for 3 min with 0.1% Triton X-100 in PBS. Cells were washed 5 times with PBS and then blocked for 30 min with 1% IgG-free BSA (Research Organics). The monolayers were incubated overnight at 4°C with a polyclonal antiserum against p32 and with monoclonal antibody against PKC␦ or PKC. After washing 5 times with PBS, the coverslips were incubated for 1 h at room temperature with secondary antibodies (FITC-conjugated goat anti-rabbit and TRITC-conjugated goat anti mouse). The monolayers on the glass coverslips were mounted with the antifade reagent Vectashield (Vector laboratories Inc., Burlingame, CA). The fluorescence of the monolayers was examined using a confocal microscope (MRC-600, Bio-Rad) with a krypton argon laser.

RESULTS
Purification and Identification of the p32-binding Protein-We have previously reported the existence of at least seven proteins that bind PKC in rat hepatocytes (16). In addition, we developed a rapid and simple purification method for these proteins taking advantage of their ability to precipitate with Triton X-100 (27). In the current study, we have employed this protocol to isolate a protein from the Triton-insoluble fraction of rat hepatocyte homogenates (Fig. 1A, lane 1) by DEAEcellulose chromatography followed by a second precipitation step with Triton X-100 and by electroelution (Fig. 1A, lane 2). The aim of the present study was to identify the isolated PKC-binding protein by NH 2 -terminal sequencing and to characterize its interaction with PKC. As shown in Fig. 1A, we purified to homogeneity a protein with an apparent molecular mass of 32 kDa.
The purified protein was electroblotted to polyvinylidene difluoride membranes and subjected to NH 2 -terminal sequence analysis. The sequence LHTEGDKAFVEFLTDEIKEE was ob-tained and was subsequently aligned with nonredundant SwissProt data base sequences. As shown on Fig. 1B, we found an almost perfect match with the NH 2 terminus of the mature chain of human pre-mRNA splicing factor SF2 p32 subunit originally characterized by Krainer et al. (20). As can be seen, 17 of 20 amino acids in the query were identical to this protein (85%), and the remaining three were conservative changes (100% positive). Subsequent reports have shown that p32 is dispensable for general splicing activity (31) and have instead described it as a receptor of complement component C1q (19) and as a multiligand protein (21)(22)(23)(24)(25). Indeed, alignment of our found sequence with the conceptual rat cDNA translated sequence reported by Lynch et al. (33) gave a perfect match with the gC1qBP, which represents the rat homologue protein of human SF2 p32 subunit (see Fig. 1B).
p32/gC1q-BP is a single-chain, highly acidic protein (pI 4.15) with a calculated mass of 23.8 kDa that differs substantially from the apparent size of 32 kDa. This is, however, not unprecedented, because discrepancies between real sizes and those determined by gel electrophoresis are often described for highly charged proteins and may be caused by the reduced binding of SDS to such proteins (32). On the other hand, the presence of carbohydrates on proteins also contributes to the production of an abnormal relative electrophoresis mobilization. Because the primary sequence of p32 predicts three putative N-linked glycosylation sites, we proceeded to search for their presence in p32. The purified protein was subjected to carbohydrate analysis with the digoxigenin detection kit as described under "Experimental Procedures." As observed in Fig. 1C (lane 2), p32 gave and even better carbohydrate signal than the positive control ovalbumin (lane 3), indicating therefore that p32 is a glycosylated protein.
Cloning and Expression of p32 as GST Fusion Protein-p32 cDNA was obtained from a clone derived from a mouse embryonic cDNA library (Research Genetics Inc). As Fig. 2A shows, the nucleotide sequence of this cDNA contains a complete reading frame encoding a 279-residue protein that shows minor discrepancies with the one reported by Lynch et al. (33). Differences occur at residue 11 (Ala instead of Ser) and at residue 160 (Leu instead of Arg). Additionally, our sequence has three alanines (residues 14 -16), instead of only two as reported by Lynch et al. (33) at the same position. The "mature" form of the protein corresponding to residues 74 -279 presumably is generated by site-specific cleavage and removal of the highly basic 73-residue-long NH 2 -terminal segment during post-translational processing.
The 618-bp fragment resulting from the EcoRI (SI nuclease trimmed) and DraI digestion of the plasmid containing p32 cDNA was subcloned in the pGEX2T vector. In Vitro Analysis of p32 Association with PKC Isoforms-Rat hepatocytes express seven PKC isoenzymes (18) that are coeluted by DEAE-cellulose chromatography: two conventional isoforms (␣ and ␤II), four novel isoforms (␦, ⑀, , and ), and one atypical isoform (). To explore which PKC isozyme binds to p32, we made use of this partly purified PKC extract as a probe in an overlay assay. Fig. 3A shows the Western blot analysis of the PKC preparation used, indicating that similar relative amounts of each PKC isoform were present in the overlay assay. Moreover, Fig. 3B (upper and lower) clearly indicate that all of the PKC isozymes tested bind to purified recombinant GST-p32. The binding is not due to secondary antibody interactions, and it is specific for p32 because no PKC became associated to other proteins such as BSA or GST alone under the conditions employed (see upper portion of Fig. 3B). The typical PKC substrate, histone H1, which associates with all PKC isoforms, served as the positive control (Fig. 3B, top). The data also indicate that p32 does not interact with all PKC isoforms in the same way, because dependence of the binding on PKC cofactors (phosphatidylserine, 1,2-diolein, and calcium) differs for each PKC isoform. Quantification of binding data, presented in Fig. 3C, demonstrates that whereas PKC␦ is capable of binding p32 only in its activated conformation, other isozymes like PKC ␣ and , although interacting better with p32 in a cofactor-dependent manner, can also bind to p32 in their absence. On the other hand, the association between p32 and PKC ␤, ⑀, and is equally efficient regardless of the presence of PKC activators, but it seems that PKC can bind p32 with great affinity in their absence. In summary, independence of binding to p32 on PKC activators ranges as follows: P32 Is Not Phosphorylated by PKC and Increases PKC Kinase Activity in Vitro-To explore whether p32 is an in vitro substrate for several PKC isoforms, we used PKC isozyme-specific immune complexes to measure their phosphorylation activity toward GST-p32 in comparison with a typical PKC substrate, histone H1 III-S. Fig. 4A shows a representative autoradiography with its corresponding dried and Coomassie Blue-stained gel. As can be observed, whereas histone H1 was efficiently phosphorylated (only that phosphorylated by PKC is shown, but identical results were obtained with the other immunoprecipitated PKC isozymes), p32 was not phosphorylated by any PKC isoform tested, because the band corresponding to GST-p32 present in the Coomassie Blue-stained gel is absent at the same position in the autoradiography. It can also be seen that other proteins that co-immunoprecipitate with PKC were also phosphorylated. We analyzed the immunoprecipitates with anti-PKC antibodies (results not shown) and observed that the two more prominent bands observed in the autoradiography at 55 kDa corresponded to the catalytic fragment of PKC. The identity of the other faint bands is unknown. In addition, because it has been recently reported that p32 inhibits PKC aldolase phosphorylation in vitro (25), we per- FIG. 3. Overlay assay to determine binding of PKC isozymes to purified p32 (gC1q-R). A, Western blot of the PKC preparation used as probe in overlay assays. Samples (70 g) of partially purified PKC by DEAE-cellulose chromatography from rat hepatocytes homogenates were separated by 10% SDS-PAGE and transferred to nitrocellulose membrane. Immunoblot analysis was performed with polyclonal isozyme-specific anti-PKC antibodies and developed with anti-rabbit IgG-alkaline phosphatase-conjugated second antibody. Positions of M r standards are indicated. B, determination of PKC isoform binding to p32 by overlay assay. Samples (12 g) of purified recombinant GST-p32, GST, or BSA (negative controls) or histone H1-IIIS (positive control) were subjected to SDS-PAGE (12% gels) and blotted onto nitrocellulose. PKC binding was determined by incubating the membranes in the absence or presence of partially purified PKC and in the absence or presence of cofactors (20 g/ml phosphatidylserine, 0.8 g/ml 1,2-diolein, and 0.5 mM CaCl 2 ) for 1 h. The mixture was removed, membranes were washed, and bound PKC was detected with isozymespecific anti-PKC antibodies (which recognize in all cases the COOH terminus) or with anti-panPKC (which recognize all isoforms) as indicated on the figure. Membranes were developed with an alkaline phosphatase-conjugated secondary antibody. Positions of M r standards are indicated; and the GST-p32 protein revealed by the PKC bound to it is also indicated with an arrow. Data are representative of four independent experiments. H-H1, histone H1-IIIS. C, quantification of PKC binding to p32 in the absence or presence of PKC cofactors. Band intensities were quantified by densitometric scanning, and data obtained in the presence of cofactors were taken as 100% for each PKC isozyme to normalize the results of different experiments. Values plotted are means Ϯ S.E. for at least three independent experiments.
formed the assay in the absence or presence of p32 and in the absence or presence of the selective and potent PKC inhibitor Ro 31-8220. The upper portion of Fig. 4B shows a representative example of the results obtained with PKC and the lower portion those with PKC␦. We selected these isoforms on the basis of the strong in vitro interaction observed with p32 in a cofactor-dependent or -independent manner, respectively, but similar results were obtained with the other five PKC isozymes expressed in rat hepatocytes (not shown). In Fig. 4C are presented the quantitative data of several experiments. As can be FIG. 4. In vitro phosphorylation studies. A, phosphorylation of p32 by PKC isozyme immune complexes. Kinase activity was measured with the PKC immunoprecipitates as described in detail under "Experimental Procedures" in the absence or presence of substrate added to the immunoprecipitates, histone H1 (H) or GST-p32. Reactions were allowed to proceed for 20 min at 30°C, were terminated by the addition of electrophoresis sample buffer, and were analyzed by 12.5% SDS-PAGE and autoradiography. A representative of three independent experiments is shown. The autoradiogram with its corresponding dried and Coomassie Blue-stained gels is shown. The black arrows indicate the position of GST-p32, and the empty arrows the position of 31-KDa histone H1, respectively. Size markers are included on the right. B, p32 increases substrate phosphorylation by PKC. PKC immune complex kinase activity toward histone H1-IIIS was determined as described previously under "Experimental Procedures" in the absence or presence of the substrate, histone H1 (ϪS or H), and in the absence or presence of 200 g/ml GST-p32. Reactions performed in parallel with histone in the absence or presence of GST-p32 and in the presence of the PKC-specific inhibitor Ro 31-8220 are also shown (HϩRO). Results obtained with PKC are presented in the upper portion and those with PKC␦ in the lower portion. The results shown are representative of three independent experiments. The black arrows indicate the position of GST-p32, and the empty arrows the position of 31-KDa histone H1, respectively. C, quantification of specific phosphorylation. Autoradiograms and their corresponding stained gels were quantified with an image densitometer, and the specific phosphorylation was determined as the ratio of phosphorylated protein to the total protein content. Values plotted are means Ϯ S.E. for three experiments with different cell preparations. observed, whereas PKC showed 1.5-fold activation in the presence of p32, PKC␦ was activated 2-fold (indicated with a black arrow in Fig. 4B), because histone H1 phosphorylation increased in the presence of p32 in comparison with the obtained in its absence (see empty arrows) but not when the specific PKC inhibitor Ro 31-8220 was added. As indicated above, when we measured the activity of the other PKC isoforms under the same conditions, identical results were obtained, even if other substrates such as aldolase (in the case of PKC) or a peptide derived from myelin basic protein were used (data not shown).
P32 Associates in Vivo with PKC␦ and PKC-Although it is clear that p32 binds all classes of PKC isoforms in vitro, it remained important to determine which isoform interacts with p32 inside the cells. Therefore, we examined in cultured rat hepatocytes (clone C9), by immunofluorescence and confocal laser microscopy, the possible subcellular co-localization of p32 with the previously selected PKC␦ and PKC isoforms. We also studied the effect of PKC activation on this localization using 1 M TPA (an optimal concentration reported for PKC activation on these cells) (34). Figs. 5 and 6 show representative experiments of the time course of treatment of subconfluent C9 cells with TPA and the dual immunostaining of p32 and PKC (Fig. 5) or p32 and PKC␦ (Fig. 6) with specific antibodies, respectively. In untreated cells, p32 and PKC␦ appear distributed throughout the cytoplasm, whereas PKC locates predominantly at the cytoplasm but also at the cell nucleus. Monolayers treated with TPA did not show a clear PKC redistribution as a result of its activation at the time points presented, although in some cells both PKC ␦ and concentrate at certain long and thin cell elongations. However, the cells treated with TPA changed their morphological appearance, as they lost their polygonal shape and retracted from each other (see Figs. 5 and 6). Furthermore, the addition of TPA induced a transient translocation of p32 to the cell nucleus, visualized 5 min after TPA treatment (insert and arrowheads in Fig. 5, arrows in Fig. 6), as 15 min after its addition, p32 is no longer found at the nuclei. Consistent with the data obtained in vitro, the immunofluorescence analysis shows a constitutive colocalization of p32 and PKC, mainly at the perinuclear region, but also when p32 is at the cell nuclei (Fig. 5, merged images (Merge), with or without the addition of TPA). In contrast, the images in Fig. 6 indicate that PKC␦ activation is required for its association with p32. Their co-localization increases with the time of TPA treatment, and it is also found mostly in the perinu-clear region (Fig. 6, arrows in merged images). Taken together, the results presented here indicate that p32 and PKC and ␦ interact together in vivo.

DISCUSSION
The coordinated interaction of kinases, phosphatases, and other regulatory molecules with scaffolding proteins is emerging as a major theme in intracellular signaling networks (6 -8).
There are now an increasing number of PKC-binding proteins believed to play a role in directing the location and function of individual PKC isoforms to particular subcellular locations within cells (9 -18). In this study, we have purified a PKCbinding protein with an apparent molecular mass of 32 kDa and identified it as the previously described p32 subunit of the pre-mRNA splicing factor SF2 (20) or the receptor of the globular heads of the complement factor C1q (19). As mentioned previously, recent evidence indicates that the molecule is a multifunctional protein with affinity for diverse ligands, which include thrombin, vitronectin, kininogen, a mitochondrial marker protein, and several viral proteins involved in viral gene expression (21)(22)(23)(24). Very recently, Storz et al. (25) reported for the first time the interaction of this protein with PKC, demonstrating that p32 associates in vivo with PKC at the mitochondrial membranes of the B cell line SKW 6.4. Here, we provide evidence that p32 interacts differentially with several PKC isoforms both in vitro and in vivo. Using an overlay assay approach, we have demonstrated that although p32 is able to bind specifically to all PKC isozymes expressed in rat hepatocytes (␣, ␤, ␦, ⑀, , , and ), their binding affinities depend on the presence of PKC activators (Fig. 3). These results suggest, therefore, that there may be isozyme-specific domains in PKC that are better exposed in a specific conformation that permits the interaction with p32.
We have also shown that although p32 binds to the kinase, it does not serve as a substrate for any PKC isoform tested (Fig.  4A). Interestingly, we demonstrated that p32 stimulates PKC activity in vitro, suggesting that it may also regulate positively the enzyme in vivo (Fig. 4, B and C). In this respect, our data are in agreement with those of Storz et al. (25), who report that PKC does not phosphorylate p32, but are in contrast with their finding that p32 inhibits aldolase phosphorylation by PKC. It is important to consider that the in vitro studies of Storz et al. (25) suggest that the presence of p32 does not inhibit the kinase FIG. 5. Association of p32 and protein kinase C in vivo. C9 monolayers grown in coverslips were incubated in the absence or presence of 1 M TPA during 5 or 15 min as indicated on the figure. Cells were fixed, permeabilized, and co-immunostained with antibodies against PKC and p32. Fluorescence was analyzed by laser confocal microscopy as described under "Experimental Procedures." P32 was visualized with FITC-conjugated goat anti-rabbit antibody and PKC with TRITC-conjugated goat anti-mouse antibody. P32 (green) colocalizes with PKC (red) as indicated by the yellow color in the merged images. The arrowhead indicates the nuclear localization of both p32 and PKC in cells treated for 5 min with TPA. Data are representative of four independent experiments. intrinsic ability, because autophosphorylation of PKC not only was unaffected but was slightly increased in the presence of p32. Furthermore, they also show that aldolase phosphorylation by PKC immunoprecipitates from the soluble cell fraction could be readily discerned. They explain these results by suggesting that p32 regulates PKC activity in a compartment-specific fashion and that binding of p32 to the kinase catalytic domain might only restrict, by steric hindrance, the substrate access to PKC (25). It is important to take into account that the p32 portion obtained from the construct employed by us corresponds to the mature form of p32 (predominantly found inside cells), which lacks a 73-residue signal sequence at the NH 2 terminus, necessary to target the protein to mitochondria (22). Thus, the discrepancy with our data can be explained by the fact that Storz et al. (25) in their experiments used a construct that included all of the open reading frame of p32. On the other hand, it has been reported recently (36) that 14-3-3 proteins stimulate classical and PKC isoforms up to 2.5-fold in an apparently nonspecific manner solely because of the acidic nature of these proteins, as other acidic proteins such as lactoglobulin and apotransferrin stimulated PKC in a similar fashion. Because p32 is a highly acidic protein having a pI of 4.15, it would be predicted that it may activate PKC activity in the same way, but it remains to be clarified as to which mechanism stimulates PKC activity. In this respect, it is noteworthy that despite the differences in affinity displayed by p32 in binding PKC isozymes, its effects on PKC activities are stimulatory for all isoforms.
Immunofluorescence experiments with cultured rat hepatocytes in our current work demonstrated that p32 co-distributes with PKC and PKC␦ in vivo. Our data indicate that whereas p32 and PKC are constitutively associated, PKC␦ activation is required for its association with p32 at the perinuclear region in intact cells. It is interesting to note how the same protein, p32, is able to behave as a scaffold (for PKC) or as a receptor for activated C-kinase (RACK) (for PKC␦) depending upon which PKC isoform associates with it. In addition, a salient finding obtained here is that p32 is transiently translocated to the cell nucleus in response to TPA treatment, co-localizing there with PKC, and detected as a nucleoplasmic speckled pattern, typical of nucleoplasmic small nuclear ribonucleoprotein particles. This location is therefore consistent with the functional proposed roles for p32 in transcriptional regulation of viral gene expression (24) and as a regulatory factor of RNA splicing by inhibiting ASF/SF2 RNA binding and phosphoryl-ation (23). There has been some controversy regarding the subcellular localization of p32. Although some authors have reported p32 to be present predominantly in the mitochondria (22,25), others have found it at the cytoplasm (21), cell surface (19), and nucleus (20,21). Recently, using immunogold electron microscopy to evaluate the p32(gC1q-BP) distribution in immunogold-labeled cells, Soltys et al. (37) have shown specific labeling of mitochondria, condensing vacuoles, endoplasmic reticulum, nuclei, and cell surface in several cultured cell lines and in rat tissues (including rat liver) under normal physiological conditions. Furthermore, the distribution of p32 appears to be altered during adenovirus infection, with p32 migrating to the nucleus together with a viral core protein (21), suggesting that p32 plays a bridging role in nucleus-mitochondrion interactions. Therefore, it seems that p32 does not has a unique location within a cell, and our immunofluorescence data support this notion. With respect to the p32 carbohydrate content, it is important to mention that although up until the mid 1980s it was widely believed that glycosylation was restricted largely to proteins localized on the cell surface, the widespread existence of complex glycoconjugated proteins within the nucleus and cytoplasm is now know (38).
Another remarkable issue is that p32 shares with another PKC-binding protein, calreticulin (18), the property to bind C1q complement factor. Several binding molecules and potential receptors for C1q have been described: a cell surface receptor, calreticulin; a 60-kDa homologue, termed cC1q-R, that binds to the collagenous region of C1q; and a 32-kDa glycoprotein with affinity for the globular heads, or gC1q-R (39). Ghebrehiwet et al. have found that p32 and cC1q-R often coelute during purification and are able to associate with each other forming a complex on the cell surface (39). The functional significance of the interaction between these two proteins and PKC remains to be established.
Finally, it is also astonishing that the solved crystal structure of the p32(gC1q-BP) mature form determined at 2.25 Å (35) revealed a doughnut-shaped protein composed of three monomers, in close resemblance with the propeller-like threedimensional structure of the receptors for activated protein kinase C, RACK1 and ␤Ј-COP, that in turn belong to the WD-40 family of regulatory proteins (15,18). Important questions remain concerning the biological significance of this similarity as well as the molecular mechanisms involved in the functional coupling between p32 and protein kinase C.
FIG. 6. Association of p32 and protein kinase C ␦ in vivo. C9 monolayers grown in coverslips were incubated in the absence or presence of 1 M TPA during 5 or 15 min and processed for immunostaining and confocal microscopy analysis as indicated under "Experimental Procedures." P32 was visualized with FITCconjugated goat anti-rabbit antibody and PKC␦ with TRITC-conjugated goat antimouse antibody. P32 (green) colocalizes with PKC␦ (red) as indicated by the yellow color in the merged images (arrows). The data are representative of four independent experiments.