Identification of a major protein kinase C-binding protein and substrate in rat embryo fibroblasts. Decreased expression in transformed cells.

We have used an interaction cloning strategy to isolate cDNAs for sequences that interact with protein kinase C (Chapline, C., Ramsay, K., Klauck, T., and Jaken, S. (1993) J. Biol. Chem. 268,6858-6861). In this paper, we report a novel sequence, clone 72, isolated according to this method. Clone 72 has a 4.8-kilobase pair open reading frame; antibodies to clone 72 recognize a >200-kDa protein in cell and tissue extracts. Clone 72 message and protein are detected in a variety of tissues. Immunoprecipitation studies demonstrate that clone 72 is the major >200-kDa binding protein described previously in REF52 fibroblasts (Hyatt, S. L., Liao, L., Aderem, A., Nairn, A., and Jaken, S. (1994) Cell Growth & Differ. 5, 495-502). Expression of clone 72 message and protein are decreased in progressively transformed REF52 cells. Since clone 72 is both a protein kinase C (PKC)-binding protein and substrate, decreased levels of clone 72 may influence both the subcellular location of endogenous PKCs as well as signaling events associated with clone 72 phosphorylation. Our results emphasize that the role of PKCs in carcinogenesis may involve several factors, including the quantity and location of the PKCs isozymes and their downstream targets.

Protein kinase C (PKC) 1 is a family of phospholipid-dependent kinases involved in basic cellular functions, including regulation of growth, differentiation, and gene expression (1,2). The role of individual PKCs in these processes is not yet known; however, since most cells express more than one type of PKC, it seems likely that individual PKCs have unique rather than overlapping functions. All of the PKCs require phosphatidylserine for maximal activity; however, PKCs can be grouped according to differences in their dependence on other activators. In addition to phosphatidylserine, conventional PKCs require calcium and diacylglycerol, novel PKCs require only diacylglycerol, and atypical PKCs require nothing more. Several other lipid modifiers of PKC activity have also been identified, and there is some evidence that they may selectively influence individual isozyme activities (1). Selective isozyme activation in response to physiological agonists has been noted and may be a result of the recognized differences in cofactor requirements among the PKCs (3)(4)(5)(6).
In addition to isozyme selective activation, isozyme-specific functions may depend on selective substrate recognition. However, only minor differences in substrate specificity among the isozymes have been observed in in vitro assays (7,8). Recently, immunofluorescence studies have demonstrated unique subcellular localizations for individual PKCs (9 -11) (data not shown). Thus, targeting of individual PKCs to specific subcellular addresses may be a means of restricting accessibility to substrates and, thereby, provide the mechanism for isozyme-selective phosphorylation events in vivo. To isolate proteins that interact with PKCs with high affinity, we developed an assay for identifying PKC-binding proteins (12)(13)(14). Subsequently, this assay was adapted to screen expression libraries and isolate cDNA clones for PKC-binding proteins (15). Binding proteins isolated according to this strategy are also substrates (16). In this manuscript, we report the full-length sequence of one of these binding proteins, clone 72. The results indicate that clone 72 is widely expressed and is a major PKC-binding protein in REF52 fibroblasts.

MATERIALS AND METHODS
Library Screening-A gt11 REF52 cDNA library was prepared and screened for PKC-binding proteins as described in Ref. 15. Positive colonies were plaque-purified and sequenced using an automated sequencer. Two overlapping clones (53 original and 72 original, see Fig. 1) were identified as partial cDNAs and served as the starting point for obtaining the entire open reading frame. Two rounds of 5Ј-RACE were used to derive the 5Ј-extended (EXT) sequences 53EXT1 and 53EXT2. This same sequence was independently obtained by screening a rat brain cDNA library with a 32 P-labeled clone 53 cDNA probe (brain cDNA). The 3Ј extensions (72EXT1-5) were derived from the REF52 library using pairs of primers of gene-specific and sequences in PCR reactions. PCR products were cloned into pCRI (Invitrogen) or pSK (Stratagene) vectors for sequencing. Two to three clones of each PCR product were sequenced to identify and correct for PCR-generated sequencing errors. Homology searches were done using the Blast program (17) and motif searching was done using Blocks Searcher (18).
Expression of Recombinant Sequences and Production of Antisera-53ORIG and 72ORIG cDNAs were isolated and ligated in frame into the pQE bacterial expression vector (Qiagen) to produce recombinant His-tagged fusion proteins. The expressed sequences were isolated by nickel-nitrilotriacetic acid chromatography according to the manufacturer's instructions. The purified antigens were used to raise antisera in rabbits. Antisera were purified by affinity chromatography using the expressed sequences coupled to Sepharose.
Northern Blots-Blots containing 2.0 g of poly(A) mRNA from multiple tissues were purchased from Clontech. Oligo(dT)-selected RNA was prepared from the normal and transformed REF52 cells. Aliquots (2.5 g) were electrophoresed and blotted to Hybond N nylon membranes (Amersham Corp.). Random labeled cDNA probes of 53ORIG and 72ORIG were prepared with [ 32 P]dCTP. Prehybridization conditions were 3 h at 42°C in 5 ϫ SSPE and 2 ϫ Denhardt's in the presence of 100 g/ml denatured salmon sperm DNA, 2% SDS, and 50% formamide. Blots were hybridized for 48 h at 42°C in the prehybridization * This work was supported by National Institutes of Health Grants CA53841 and GM50152 (to S. J.). 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.
Immunoblots-Growth and properties of REF52 cells have been described previously (19,20). Cell lysates and tissue homogenates were prepared as described (16). Aliquots (20 -50 g of protein) were separated by SDS-polyacrylamide gel electrophoresis and blotted to nitrocellulose. Blots were probed with affinity-purified clone 72 antiserum as described (16). The PKC overlay assay is described in Ref. 13.
Immunofluorescence-REF52 cells were grown on glass coverslips. Cells were fixed in formaldehyde, permeabilized in methanol, and incubated with first and second antibodies as described previously (21).
Phosphorylation of Clones 53 and 72 Expressed Sequences-Purified recombinant clone 72ORIG expressed sequence (40 and 80 g, respectively) were incubated with PKC purified from rabbit brains (22) (4 and 8 g, respectively) in the presence of calcium, phosphatidylserine, and radiolabeled ATP as described previously (22). The phosphorylated proteins were digested with V8 protease overnight at 37°C. Peptides were separated on a C8 reversed phase column with an acetonitrile gradient in the presence of 0.01% trifluoroacetic acid. Phosphopeptides were identified by liquid scintillation counting.

Isolation of Clone 72
Open Reading Frame-In order to identify PKC-binding proteins and substrates that mediate PKC action in cells, a gt11 cDNA library prepared from REF52 cells was screened for PKC interacting proteins as described previously (15). Two clones isolated during this screen, clones 53 and 72, were overlapping (Fig. 1). Additional 5Ј sequence was obtained by 5Ј-RACE using primers to the clone 53 sequence and by screening a rat brain library with a clone 53 cDNA probe. Both of these approaches led to identification of the apparent 5Ј end of the coding sequence. An open reading frame with a start methionine and a consensus Kozak sequence begins at position 34. Additional 3Ј sequence was obtained by PCR screening the REF52 library with clone 72 primers. A summary of the individual clones obtained with these methods is shown in Fig. 1. The nucleic acid and predicted amino acid sequences are shown in Fig. 2. Database searches indicated clone 72 had substantial homology to the recently reported clone 322 (23); however, clone 72 had an additional ϳ1000 bp of open reading frame in the 5Ј region and ϳ400 bp less in the 3Ј region due to the stop codon at position 4821. Several differences in primary sequence were also noted. No overall homologies to other nucleotide or amino acid sequences were detected. However, a search for domain similarities indicated significant homology with two well characterized PKC substrates, neuromodulin and MARCKS. Residues 286 -319 have significant homology to the PKC phosphorylation site in neuromodulin (and MARCKS) (24 -26). Two other peptides (148 -178 and 502-532) also have significant homology to the PKC phosphorylation site in MARCKS. In addition to the phosphorylation domain homology, four peptides (residues 221-263, 317-359, 404 -446, 614 -656) have significant homology to an acidic domain in neuromodulin (residues 120 -162 in rat neuromodulin). Repeats of these highly acidic peptides (more than 30% acidic residues) are likely to be important for clone 72 structure.
Expression of Recombinant Clone 72-The original clone 72 partial cDNA was expressed in bacteria as a recombinant Histagged fusion protein. The recombinant protein had an apparent molecular mass of 97 kDa, which is significantly larger than the predicted molecular mass of 44 kDa (Fig. 3). In general, anomolously slow migration on denaturing gels is characteristic of PKC-binding proteins (data not shown), including MARCKS (16), and may reflect an extended, rod-like structure. Antisera raised to the purified recombinant protein recognized the expressed sequence and furthermore recognized a high molecular mass (Ͼ200 kDa) protein in REF52 cell extracts (Fig.  3). The antisera immunoprecipitated a Ͼ200-kDa protein from REF52 cell extracts which reacted with affinity-purified clone 72 antibody on immunoblots (Fig. 4A). The clone 72 antibody immunoprecipitates also contained a Ͼ200-kDa PKC-binding protein detected by PKC overlays (Fig. 4B). Thus, the cDNA isolated according to PKC binding activity in the interaction cloning assay appears to code for a PKC-binding protein expressed in REF52 cells.
Tissue Distribution of Clone 72-The relative expression of clone 72 in various tissues was compared by Northern and Western blots. The major form of the detectable message was 6 kb and was abundantly expressed in testes, heart, skin, and brain (Fig. 5A). Smaller message sizes were also detected in skin, liver, and heart. Since identical results were obtained with blots probed with clone 72 and clone 53 cDNA probes (data not shown), these bands appear to be specific and most likely represent message species related to clone 72.
Immunoblots of rat tissues demonstrated a ladder of immunoreactive proteins Ͼ200 kDa in most tissues (Fig. 5B). Expression levels were highest in testes and brain. Smaller bands that may be related either to the smaller, related messages or to proteolysis were also detected in most tissues. Thus, clone 72 encodes a high molecular weight PKC-binding protein that is widely expressed in mammalian tissues.
Clone 72 Is a PKC Substrate-Clones 72ORIG and 53ORIG expressed sequences were purified and used in in vitro kinase assays with PKC. After phosphorylation in the presence of radiolabeled ATP, the proteins were digested with V8 protease, peptides were separated by reversed phase HPLC, and phosphopeptides were identified by liquid scintillation counting. Only one V8 peptide from clone 72ORIG contained phosphate; this was identified as 494R(35)E529 by automatic Edman degradation (underlined in Fig. 2). A synthetic peptide corresponding to 496V(33)E529 was also phosphorylated by PKC in vitro with a stoichiometry of 0.72. A second V8 peptide in 53ORIG was also phosphorylated. Further digestion with chymotrypsin indicated that the phosphorylation site was within the peptide 296R(13)D308 (underlined in Fig. 2). Both of these sequences were identified in the domain search as regions homologous to the phosphorylation motifs in neuromodulin and/or myristoylated alanine-rich C kinase substrates (see above). These results are consistent with previous studies that established a strong correlation between PKC-binding proteins and substrates (16).
Clone 72 Expression Is Transformation-sensitive-REF52 (REFA) cells are the parental cell line for progressively transformed SV40 derivatives, REFB, REFC, and REFD. REFB cells FIG. 1. Contig map of clone 72 cDNA clones and PCR products. The original cDNAs isolated in the interaction cloning screen (53ORIG and 72ORIG) were overlapping cDNAs. 5Ј-RACE of clone 53 produced the 500-bp extension 53EXT1. 5Ј-RACE of 53EXT1 was used to produce 53EXT2. Concurrently, a rat brain library was screened with a clone 53 cDNA probe, and brain cDNA (Br cDNA) was isolated. Additional 3Ј sequence was obtained by PCR from the REF52 cDNA library using 3Ј gene-specific and sequences as primer pairs. The resulting products (72ext1-5) contained an additional 2.5 kb of coding sequence and 416 bp of 3Ј-noncoding sequence.
are morphologically transformed, REFC cells are capable of growth in soft agar, and REFD cells are fully transformed as demonstrated by their ability to produce tumors in nude mice (12,13,16). Northern blots demonstrated the progressive loss of clone 72 message with progressive transformation of REF52 cells (Fig. 6). Immunoblots also indicate decreased expression of clone 72 protein with progressive transformation of REF52 cells (Fig. 7A). Loss of immunoreactive clone 72 correlates with the progressive loss of the major Ͼ200-kDa PKC-binding protein with progressive transformation of REF52 cells (Fig. 7A). Decreased expression of clone 72 in transformed cells appears to be general rather than specific for SV40 transformation, Decreased PKC-binding Proteins in Transformed Cells 6420 since clone 72 protein was not detected in ras-transformed REF52 cells. DISCUSSION We have used the PKC overlay assay to identify PKC-binding proteins in cell extracts and to isolate cDNAs for these binding proteins. Clone 72 is a novel sequence isolated from a REF52 cell expression library according to its PKC binding properties. Considerable effort was required to isolate the entire open reading frame of this large protein. Clone 72 may be identical to a recently reported sequence, clone 322, which appears to be missing approximately 1000 bp of coding sequence in the 5Ј end (23). Other differences between the reported clone 322 and clone 72 sequences may be attributed to differences between manual and automated sequencing. 2 The translated sequence does not share significant homology with other proteins in the data bases.
Northern analysis with clone 53 and clone 72 cDNA probes indicate that clone 72 has a broad tissue distribution. These data were substantiated by immunoblot analysis of various tissues with affinity-purified antibody to the clone 72 expressed sequence. Previous studies with the related sequence, clone 322, did not detect a similar tissue distribution profile. The reasons for these discrepancies are not clear at this time. Both studies identified abundant expression of the 6-kb transcript in testes and skin, whereas abundant expression in heart was detected only in this study. The presence of additional mes- FIG. 3. Production of antisera to clone 72. Clone 72ORIG (see Fig.  1) was expressed as a bacterial His-tagged fusion protein and purified by nickel affinity chromatography. The purified protein was used to raise antisera in rabbits. Antisera were immunopurified against the clone 72ORIG expressed sequence. A, Coomassie Blue-stained gel of the clone 72ORIG expressed sequence. B, immunoblots of clone 72ORIG expressed sequence; C, REF52 cell extract probed with affinity-purified antibody raised against clone 72ORIG expressed sequence.
FIG. 4. Immunoprecipitation of clone 72 from REF52 cell extracts. Antisera to clone 72ORIG were used to immunoprecipitate proteins from REF52 cell extracts. Immunoprecipitated proteins were blotted to nitrocellulose which was stained with clone 72ORIG antibody (A) or assayed for PKC-binding proteins using the PKC overlay assay (B).
FIG. 5. Tissue distribution of clone 72. A, aliquots of mRNA (2.5 g) from various tissues were separated by electrophoresis, blotted, and hybridized to a 32 P-labeled DNA probe of clone 72ORIG. The blot shown was exposed to film for X days. Positions of molecular weight markers in kilobase pairs are shown on the right. B, aliquots of protein (50 g) from various tissues were separated by electrophoresis, blotted, and stained with clone 72ORIG antibody. Positions of molecular weight markers in kilodaltons are shown on the right. Br, brain; H, heart; Ki, kidney; Li, liver; Lu, lung; Sk, skin; Sp, spleen; Te, testes.

FIG. 2-continued
Decreased PKC-binding Proteins in Transformed Cells sages at 3.0 and 1.2 kb may indicate that alternate forms or closely related sequences are expressed in some tissues.
Several motifs in clone 72 share homology with phosphorylation domains in the PKC substrates neuromodulin and MARCKS (24 -26). In agreement with this, we found that clone 72 was phosphorylated by PKC in vitro. Although preliminary studies have demonstrated that PKC activation stimulates phosphate incorporation into clone 72 in REF52 cells (data not shown), further studies are required to establish that the phosphorylation is directly due to PKC. Taken together, these data indicate that clone 72 is an abundant PKC substrate and may be important for mediating PKC signals in a variety of cell types.
Immunoblot analysis demonstrated a graded loss of clone 72 protein in SV40-and ras-transformed fibroblasts. These results raise the intriguing possibility that differences not only in the expression of PKCs themselves, but also in the expression of PKC substrates, may contribute to disordered signaling through PKC pathways in transformed cells. Furthermore, since clone 72 can bind PKC with sufficient affinity to be detected in an in vitro overlay assay, it is also feasible that clone 72 functions as a PKC-targeting molecule in vivo. Preliminary immunolocalization studies indicate that clone 72 and certain PKCs both align with fibers in REF52 cells; however, colocalization studies have not yet been done. If clone 72 binds PKC in vivo, loss of clone 72 in transformed cells would also result in loss of PKC targeting. In fact, differential localization of PKCs in normal and transformed cells has been noted (12,16). Further studies are required to determine if the differential localization is due to differential expression of PKC-binding proteins in the normal and transformed cells. In principle, inappropriate localization could lead to promiscuous phosphorylation by PKCs and consequently contribute to disordered signaling and loss of growth control.
In summary, our results underscore that understanding the role of PKC signaling in transformation requires analysis not only of the expression of the PKC isozymes and their activation in response to exogenous stimuli, but also analysis of the expression of the downstream targets that mediate the effects of PKC activation on cellular functions, including growth, differentiation, and gene expression.
FIG. 6. Transformation-sensitive expression of clone 72 message in REF52 cells. Aliquots of mRNA (2.5 g) from REFA, REFB, REFC, and REFD cells were separated by electrophoresis, blotted, and hybridized to a 32 P-labeled clone 72ORIG cDNA probe. The blot shown was exposed to film for days. Positions of molecular mass markers are shown on the right.