DIK, a Novel Protein Kinase That Interacts with Protein Kinase Cδ

A novel serine/threonine kinase, termed DIK, was cloned using the yeast two-hybrid system to screen a cDNA library from the human keratinocyte cell line HaCaT with the catalytic domain of rat protein kinase Cδ (PKCδcat) cDNA as bait. The predicted 784-amino acid polypeptide with a calculated molecular mass of 86 kDa contains a catalytic kinase domain and a putative regulatory domain with ankyrin-like repeats and a nuclear localization signal. Expression of DIK at the mRNA and protein level could be demonstrated in several cell lines. Thedik gene is located on chromosome 21q22.3 and possesses 8 exons and 7 introns. DIK was synthesized in an in vitro transcription/translation system and expressed as recombinant protein in bacteria, HEK, COS-7, and baculovirus-infected insect cells. In the in vitro system and in cells, but not in bacteria, various post-translationally modified forms of DIK were produced. DIK was shown to exhibit protein kinase activity toward autophosphorylation and substrate phosphorylation. The interaction of PKCδcat and PKCδ with DIK was confirmed by coimmunoprecipitation of the proteins from HEK cells transiently transfected with PKCδcat or PKCδ and DIK expression constructs.

The members of the PKC 1 family, because of structural and enzymatic differences, can be subdivided into several groups (for reviews see Refs. 1 and 2). PKC␦, a member of the so-called nPKC subfamily, has attracted the interest of an increasing number of research groups over the last years and presumably is one of the most thoroughly studied PKC isoenzymes (for a review see Ref. 3). Like all the other PKC isoforms, PKC␦ is thought to play an individual role in various signaling pathways and to specifically affect diverse cellular processes, such as growth, differentiation, apoptosis, and tumorigenesis (4 -16). This specific action is likely to afford a sophisticated network of regulation of PKC␦ activity, subcellular localization, and substrate phosphorylation. Beside the well known regulation of enzyme activity by signal-induced second messengers, such as diacylglycerol, PKC␦ is regulated by up-and down-modulation of its expression (4,17,18), by phosphorylation (19 -29), and presumably by interaction with other proteins involved in signal transduction, such as other protein kinases and anchor or docking proteins (30 -38). Particularly the latter is essential for a specific subcellular localization of the enzyme and a selective phosphorylation of substrate proteins. Our knowledge of PKC␦-protein interactions, however, is rather scanty. This holds true not only for the interaction with physiologically relevant substrate proteins but also for the interaction with other proteins that might affect PKC␦ signaling.
We therefore attempted to clone proteins interacting with PKC␦ by using the yeast two-hybrid system (YTHS). Here, we describe the cloning, expression, and characterization of a novel serine/threonine kinase, termed DIK (PKC-delta-interacting protein kinase), which is coimmunoprecipitated with PKC␦ and the catalytic fragment of PKC␦ from cell extracts.
Yeast Two-hybrid Screen-The bait plasmid for the two-hybrid screen contained the catalytic domain of PKC␦. Therefore, PCR was used to specifically amplify the catalytic domain of rat PKC␦ cDNA (Pwo-Polymerase; Roche Diagnostics GmbH) with primers harboring EcoRI (5Ј) and BamHI (3Ј) sites. The PCR product was purified and digested with EcoRI/BamHI. Then the cDNA insert was ligated to pGBT9 (CLONTECH, Palo Alto, CA) to obtain pGBT9-rPKC␦ cat . The in-frame fusion to the DNA-binding domain vector was confirmed by sequencing, and the expression of the fusion protein was confirmed by immunoblotting. This bait construct did not activate the reporter genes by itself.
For the two-hybrid screen, pGBT9-rPKC␦ cat was cotransformed with the human keratinocyte MATCHMAKER cDNA library (CLONTECH) in the yeast strain PJ69-2A. Transformants that were able to grow in medium containing 2 mM 3-amino-triazole were analyzed according to the manufacturer's instructions. True positives were further analyzed by DNA sequencing (Thermo Sequenase polymerase; Amersham Pharmacia Biotech).
Molecular Cloning of DIK-The lacking 5Ј-end of the DIK cDNA isolated in the YTHS was determined by 5Ј-rapid amplification of cDNA ends (Roche Diagnostics GmbH) according to the manufacturer's instructions using three nested DIK sequence-specific primers and RNA of HEK cells. The resulting DNA fragment thus obtained was cloned into pCR-XL-TOPO (Invitrogen, San Diego, CA) and sequenced giving rise to a partial cDNA clone of 272 nucleotides derived from the 5Јregion of DIK. This cDNA clone was combined with the DIK cDNA clone isolated in the YTHS using a unique BsrGI restriction site within the overlapping region. The resulting full-length open reading frame (ORF) with 5Ј-and 3Ј-untranslated regions was cloned into the vector pBlue-* This work was supported by Wilhelm Sander-Stiftung Grant 97.090. 19. 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) AJ278016.
Plasmid Constructs-For in vitro transcription/translation the complete DIK cDNA was amplified by PCR (Pwo-Polymerase; Roche Diagnostics GmbH) with primers harboring EcoRI sites and then subcloned into pGEM3Z (Promega). To construct a GST-DIK fusion vector, the ORF of the DIK cDNA was amplified by PCR with primers harboring EcoRI sites and then subcloned into pGEX2T (Amersham Pharmacia Biotech). For expression of FLAG-tagged DIK, the ORF of the DIK cDNA was subcloned into the eucaryotic expression vector pFLAG (Sigma) using PCR and primers harboring EcoRI (5Ј) and BglII (3Ј) sites, respectively. For expression of recombinant DIK in eucaryotic cell lines, the ORF of the DIK cDNA was amplified by PCR with primers harboring EcoRI sites and then subcloned into pCDNA3 (Invitrogen). To construct a His-DIK fusion vector, the ORF of DIK was amplified by PCR with primers harboring EcoRI (5Ј) and XhoI (3Ј) sites, respectively, and then subcloned into pBac-1 (Novagen, Schwalbach). For eucaryotic expression of PKC␦ and the catalytic (PKC␦ cat ) and regulatory (PKC␦ reg ) domain of PKC␦, the cDNAs of PKC␦, PKC␦ cat , and PKC␦ reg were amplified using PCR and primers harboring EcoRI (5Ј) and BamHI (3Ј) or XhOI (3Ј) sites and then cloned into the eucaryotic expression vector pCDNA3 (Invitrogen). The nucleotide sequences of all constructs were confirmed by sequencing, and the expression of recombinant proteins was confirmed by immunoblotting.
Preparation of RNA-Cells were washed in sterile phosphate-buffered saline, lysed in the presence of RNAClean (Hybaid-AGS, Heidelberg, Germany), and extracted with chloroform.
Reverse Transcription-PCR Analysis-Total RNA was treated with 10 units of RNase-free DNase 1 (Roche Diagnostics GmbH)/g RNA. One g of total RNA was reverse-transcribed into cDNA using MuLV Reverse Transcriptase (PerkinElmer Life Sciences) and oligo(dT) primers. PCR was performed on a 1 ⁄10 aliquot of the reverse transcription mixture for 35 cycles at 94°C for 30 s, 60°C for 30 s, and 72°C for 4 min using Taq DNA polymerase (Appligene Oncor, Heidelberg, Germany). The nucleotide sequences of PCR products were confirmed by sequencing. Negative control PCR reactions were run with RNA that was not treated with reverse transcriptase.
In Vitro Transcription and Translation-Plasmid pGEM3Z-DIK was analyzed for in vitro transcription/translation by using the TNT-coupled reticulocyte lysate system (Promega). Plasmid DNA (1 g) was incubated with a rabbit reticulocyte lysate for 1 h at 30°C in the presence of T7 RNA polymerase and 40 Ci of [ 35 S]methionine. The reaction was terminated by addition of SDS sample buffer, and aliquots were analyzed by SDS-PAGE. Gels were dried, and the products were visualized by autoradiography.
DNA Transfection of HEK Cells-HEK cells were cultured in Eagle's minimum essential medium (Bio Whittaker Europe, Verniers, Belgium) supplemented with 10% fetal calf serum, 100 units/ml penicillin, 100 g/ml streptomycin, and 2 mM L-glutamine. The cells were transiently transfected with 10 g of expression plasmid (in coexpression studies 10 g of each plasmid were used) by calcium phosphate precipitation according to the manufacturer's instructions (Stratagene). Cells were supplemented with fresh medium 16 h after transfection and analyzed for expression of the recombinant proteins 2 days later.
Expression of His-tagged DIK in Baculovirus-infected Insect Cells and Purification of the Recombinant Kinase-Spodoptera frugiperda cells (Sf158) were grown at 27°C as monolayer cultures in Sf900 SFM medium (Life Technologies, Inc.) supplemented with 10% heat-inactivated fetal calf serum, 50 units/ml penicillin, 50 g/ml streptomycin, and 125 g/1iter amphotericin B.
Recombinant baculovirus was generated by cotransfecting pBac-DIK and Bac-Vector-3000 Triple Cut virus DNA (Bac vector transfection kit; Novagen, Schwalbach) according to the manufacturer's instructions. Expression of His-DIK was monitored by SDS-PAGE and immunoblotting. Cells were extracted in buffer C (50 mM Na 3 PO 4 ϫ 12 H 2 O, pH 8.0, 150 mM NaCl, 10% glycerol, 1% Triton X-100, 1 tablet of EDTA-free protease inhibitor mixture "complete" (Roche Diagnostics GmbH)) containing 5 mM imidazole as described before (see "Preparation of Cell Extracts"). His-DIK was purified from cell extracts by affinity chromatography on nickel-nitrilotriacetic acid agarose (Qiagen) according to the manufacturer's protocol. Upon washing the agarose with buffer C, His-DIK was eluted with 50 -500 mM imidazole in buffer C. The fraction eluted with 200 mM imidazole was used for the kinase assays and stored at Ϫ70°C.
Antibodies and Immunoblot Analysis-Immunoblotting was performed as described previously (40). DIK was detected by a polyclonal rabbit antiserum that was raised against the peptide AHINLQSLK-FQGGHGPAATLL (amino acids 759 -779 of DIK) coupled to keyhole limpet hemocyanine (Eurogentec, Seraing, Belgium). Preimmune serum was prepared from the blood of rabbits prior to their treatment with the antigen. The anti-DIK antibody was not suitable for immunoprecipitation of DIK. For detection and immunoprecipitation of FLAG-DIK a monoclonal anti-FLAG antibody (Sigma) was used. PKC␦ and PKC␦ cat were immunoprecipitated, and PKC␦ cat was also detected by a polyclonal anti-PKC␦ antibody (C17; Santa Cruz) recognizing the C terminus of PKC␦. For the detection of PKC␦ a monoclonal anti-PKC␦ antibody (Dianova, Hamburg, Germany) was used that recognizes the N terminus of PKC␦. As secondary antibodies alkaline phosphataseconjugated or horse-radish-peroxidase-conjugated goat anti-rabbit antibodies were used (Dianova).
Immunoprecipitation-Cells in 100-mm dishes were lysed with 600 l of ice-cold buffer D (20 mM Tris/HCl, pH 7.5, 2 mM EDTA, 5 mM EGTA, 0.2% Triton X-100, 150 mM NaCl, 5% glycerol, 1 g/ml each of pepstatin, leupeptin, aprotinin, 1 mM phenylmethylsulfonyl fluoride), sonicated using a Branson sonifier, and centrifuged at 100,000 ϫ g for 30 min at 4°C. The protein concentration of the supernatant was determined with protein-dye reagent concentrate from Bio-Rad, using bovine serum albumin as standard. The lysate (800 g of protein) was incubated at 4°C for 2 h under gentle shaking with 10 g of anti-FLAG antibody or 2 g of anti-PKC␦ antibody and for another 12 h with 20 l of protein G-or protein A-Sepharose beads (Roche Diagnostics GmbH), respectively. Upon washing the Sepharose beads four times with 500 l of ice-cold buffer E (20 mM Tris/HCl, pH 7.5, 2 mM EDTA, 5 mM EGTA, 150 mM NaCl, 5% glycerol), the immunoprecipitated proteins were dissolved by boiling the beads for 5 min in 90 l of SDS sample buffer and subsequently separated by SDS-PAGE.
Kinase Assays-Autophosphorylation of His-DIK was carried out in a total volume of 100 l containing buffer F (10 mM Tris/HCl, pH 7.2, 10 mM MgCl 2 , 3 mM MnCl 2 ), 37 M ATP containing 20 Ci of [␥-32 P]ATP, and 15 l of the purified kinase. After incubation at 30°C for 30 min, the reaction was terminated by addition of 10% trichloroacetic acid. Precipitated proteins were redissolved in SDS sample buffer, separated by SDS-PAGE, and visualized by autoradiography. Autophosphorylation of His-PKC␦ was performed essentially as that of His-DIK. However, the assay contained in a total volume of 100 l of buffer G (20 mM Tris/HCl, pH 7.5, 20 mM ␤-mercaptoethanol), 4 mM MgCl 2 , 10 g of phosphatidylserine, 100 nM 12-O-tetradecanoylphorbol-13-acetate, and 37 M ATP containing 10 Ci of [ 32 P]ATP.
Substrate phosphorylation was carried out essentially as described for autophosphorylation. However, 5 g of a substrate protein were added to the assay mixture.
Phosphoamino Acid Analysis-The analysis was essentially performed as described by Boyle et al. (41). Briefly, purified DIK (0.6 g) was autophosphorylated, precipitated, and washed with 10% trichloroacetic acid, and hydrolyzed with 75 l of 6 N HCl at 110°C for 2 h. The sample was dried by lyophilization in a Speed-Vac, resuspended in 20 l of deionized water containing 0.6 mg/ml each of the phosphoamino acid markers phosphoserine, phosphothreonine, and phosphotyrosine. Upon loading 7 l of the sample on a TLC plate, the sample was separated by two-dimensional electrophoresis. For the first dimension, we used formic acid/glacial acetic acid/water (22.5:78:899.5; pH 1.9), 1.5 kV for 60 min. For the second dimension, we used glacial acetic acid/pyridine/ water (52:5:94.3; 0.1 M EDTA; pH 3.5), 1.3 kV for 30 min. 32 P-Labeled phosphoamino acids were visualized by autoradiography, and the phosphoamino acid markers were visualized by treatment with ninhydrin.
dik Gene Analysis-Searching of the EMBO data base with the software program BLASTN using the DIK cDNA sequence resulted in the identification of the dik gene on the human BAC clone AP001743 (approximately 219 kilobases). This clone has been mapped by the Human Genome Consortium to chromosome 21q22.3. The gene locus encompasses bp 180529 -152939 of the BAC clone on the antisense strand. The exon-intron boundaries of this gene were determined by comparison with the cDNA using the DNA analysis program SIMILAR-ITY. Preliminary promoter characterization was performed on 600 bp of DNA sequence upstream of the initiation codon using the TRANSFAC data base (42).

RESULTS AND DISCUSSION
Cloning of a cDNA Encoding DIK and Demonstration of Its Cellular Expression-The catalytic domain (amino acids 326 -673) of rat PKC␦ cDNA was used as a bait in the YTHS to screen a cDNA library from the human keratinocyte cell line HaCaT. Clones were selected and tested using standard procedures. The deduced amino acid sequence of three identical cDNA clones exhibited several domains characteristic for protein kinases. Based on these and other data (see below) the encoded protein of these clones was termed DIK. The DIK clones lacked the 5Ј-end. To obtain the full-length ORF, the cDNA was extended by 5Ј-rapid amplification of cDNA ends. The nucleotide sequence of the full-length DIK cDNA (3879 bp) and the deduced amino acid sequence of DIK (784 amino acids) are shown in Fig. 1.
The ORF contains 2352 bp beginning with an initiation codon at position 49 and ending with a TAG stop codon at nucleotide 2401. An in-frame stop codon is located 12 bp 5Ј of the initiation codon. There are no other candidate initiation codons between this stop codon and the ATG at position 49, which suggests that this ATG codon is the true translation starting site. The ORF is flanked by a 5Ј-untranslated region (48 nucleotides) and a 3Ј-untranslated region (1573 nucleotides). A potential polyadenylation signal (AATAAA) is located at position 3838. The predicted 784-amino acid polypeptide has a calculated molecular mass of 86 kDa. As indicated in Fig. 1 (A  and B), it contains 12 subdomains in the N-terminal region (amino acids 22-276) that, according to Hanks and Quinn (43), are highly conserved in all protein kinases. They include an ATP binding site (GXGX 2 GXVX 14 K) in subdomain I and II, the catalytic loop region DLKPAN in subdomain VIB, and the highly conserved DFG triplet in subdomain VII. The sequence of the catalytic loop region, particularly the conserved lysine, points to DIK as a serine/threonine kinase (43). Based on these structural characteristics alone, however, the possibility cannot be excluded that DIK is a dual specificity protein kinase. Outside the catalytic domain, DIK exhibits 10 ankyrin-like repeats (amino acids 438 -768) and the potential nuclear localization signal RRX 10 RR (amino acids 469 -482; Fig. 1, A and  B), the latter indicating that DIK might be localized in the nucleus. Ankyrin-like repeats are thought to play a role in protein/protein interactions (44).
When looking for homologies of DIK with other proteins, predominantly protein kinases and proteins with ankyrin-like repeats were found. DAP, another protein kinase containing ankyrin-like repeats (45), showed the highest sequence similarity (49%) with DIK. Unlike DAP, however, DIK contains neither a death domain nor a calmodulin-binding domain. Expression of DIK mRNA was observed in various human cell lines, as determined by reverse transcription-PCR (Fig. 2). DIK mRNA could not be detected in the murine cell line MSCP5, possibly because of poorly fitting primers that were based on the human sequence of DIK. In accordance with the expression of DIK mRNA, expression of DIK protein could be demonstrated in various cell extracts by immunoblotting (Fig. 3). An anti-DIK antibody was applied that had been raised in rabbits against a peptide with the amino acid sequence corresponding to amino acids 759 -779 of DIK.
Chromosomal Localization and Analysis of the dik Gene-While this paper was in preparation, the complete sequence of the human chromosome 21q became available (46). According to these data (EMBL sequence data bank accession number AP001743), the dik gene is located on chromosome 21q22.3. The dik gene is around 27.7 kilobases in size and possesses 8 exons and 7 introns (Fig. 1C). Analysis of the region approximately 600 bp upstream of the initiation codon showed no putative tataaa box sequence. However, about 100 bp upstream of the start codon is a highly GC-rich region containing two consensus sp1 sites, perhaps indicative of a non-tataaa box containing promoter region.
In Vitro Transcription/Translation of DIK cDNA and Ex- Total RNA was prepared and transcribed into cDNA serving as template for PCR, as described under "Experimental Procedures." Negative control PCR reactions (Control) were run with RNA that was not treated with reverse transcriptase. DIK-specific primers were applied to amplify a 340-bp fragment (A). Specific primers for amplification of GAPDH were applied serving as internal standard (B).

FIG. 3. Detection of DIK in various cell extracts by immuno-
blotting. 2 ϫ 10 6 cells were lysed in 500 l of sample buffer, and 100 l were applied to SDS-PAGE. DIK was detected by immunoblotting using polyclonal anti-DIK antibody. The HEK control was performed using a preimmune serum.

pression of Recombinant DIK in Bacteria, HEK Cells, and
Baculovirus-infected Insect Cells-In vitro transcription/translation of the DIK cDNA in a rabbit reticulocyte lysate system resulted in the synthesis of several forms of the DIK protein with apparent molecular masses ranging from around 95 to 106 kDa, as demonstrated by SDS-PAGE and autoradiography of the 35 S-labeled proteins (Fig. 4A). These proteins were not synthesized in the vector control. The appearance of various protein forms might indicate that DIK is post-translationally modified and that different stages of modification may be observed in this in vitro system. Additional studies (see below, Fig. 4, B-D) supported this assumption. However, a partially incomplete transcription/translation and/or degradation of the DIK protein cannot be excluded.
A GST-DIK fusion protein was expressed in bacteria and identified in a bacterial extract upon PAGE by immunoblotting using GST and DIK antibodies (Fig. 4B). In contrast to the heterogeneous appearance of the in vitro synthesized protein, just one homogeneous protein band of GST-DIK was observed in the bacterial extract. The vector control did not contain this protein band. In good agreement with the calculated value of 86 kDa, the relative molecular mass of bacterially expressed GST-DIK was 110 kDa, i.e. 85 kDa plus 25 kDa because of the GST tag. Thus, as could be expected, DIK was synthesized in bacteria in a form that we assume to be the unmodified form.
Upon transfection of HEK cells (or COS-7 cells, not shown) with the constructs pCDNA/DIK or pFLAG/DIK, in addition to endogenous DIK, tag-less DIK or FLAG-DIK were expressed in these cells, as shown in Fig. 4C. Endogenous DIK (see particularly the vector controls, pCDNA and pFLAG, and compare also the expression of endogenous DIK in other cells, as shown in Fig. 3), as well as tag-less DIK had an apparent molecular mass of 106 kDa. FLAG-DIK exhibited a somewhat larger molecular mass because of the tag (around 2 kDa). The 106-kDa form was likely to represent the final stage of the diverse modified forms of DIK that were observed in the in vitro tran-

FIG. 4. In vitro transcription/translation of DIK cDNA (A), expression of GST-DIK in bacteria (B), expression of DIK and FLAG-DIK in HEK cells (C), and expression of His-DIK in baculovirus-infected insect cells (D).
A, the reticulocyte lysates were incubated with the vector pGEM3Z alone (left lane) or with pGEM3Z/DIK (right lane) as described under "Experimental Procedures" and analyzed by SDS-PAGE. In vitro synthesized 35 S-labeled proteins were visualized by autoradiography. B, bacteria were transfected with pGEX (vector control) or pGEX/DIK ("Experimental Procedures"). Bacterial extracts were applied to SDS-PAGE and immunoblotted using an anti-GST antibody (left panel) or the polyclonal anti-DIK antibody (right panel). C, HEK cells were transfected with pCDNA or pFLAG (vector controls) and pCDNA/DIK or pFLAG/DIK. Cell extracts were applied to SDS-PAGE and immunoblotted using the anti-DIK antibody. The DIK and FLAG-DIK protein bands are indicated by DIK. D, recombinant baculovirus (pBac/DIK) was generated, and Sf158 cells were grown, infected with the virus, and extracted as described under "Experimental Procedures." His-DIK was purified from cell extracts by affinity chromatography on nickel-nitriloacetic acid agarose as described under "Experimental Procedures." 40 l of the cell extract (pBac/DIK, left panel) and 15 l each of the fractions eluted from the agarose with 50 -500 mM imidazole (right panel) were applied to SDS-PAGE, and His-DIK was visualized by immunoblotting using the anti-DIK antibody. Control, extract from uninfected cells. scription/translation system (see Fig. 4A). This assumption was strongly supported by the fact that also in baculovirus-infected insect cells DIK is expressed exclusively as a 106-kDa protein (see below, Fig. 4D). In addition to this completely modified form, putatively partially modified forms of DIK and FLAG-DIK with lower molecular masses (95 and 97 kDa, respectively), similar to those observed in the in vitro system, were expressed in the transiently transfected cells (Fig. 4C). Upon transfection of the cells with the tag-less DIK construct, the lower molecular mass form of DIK was even the predominant form. An answer to the question of whether this is a partially modified protein and whether DIK is indeed post-translationally modified has to await further studies.
For the production of larger amounts of the 106 kDa DIK protein that presumably was completely modified and thus could be expected to be enzymatically active the baculovirusinfected insect cell system was chosen, because bacteria had been found unable to express DIK in this form. A His-tagged DIK protein, with six histidine residues at the C terminus, was efficiently expressed in these cells, as shown in Fig. 4D (left  panel). The apparent molecular mass of around 106 kDa (the molecular mass is not increased significantly by the short His tag) indicated that, like endogenous DIK in HEK, COS, and other cells, His-DIK was present in the infected insect cells predominantly as the completely modified protein. His-DIK was purified by metal chelate affinity chromatography on nickel-nitrilotriacetic acid-agarose (Fig. 4D, right panel). Upon extensive washing, His-DIK was eluted from the agarose with 50 -500 mM imidazole. The fraction eluted with 200 mM imidazole was used for the kinase assays.
DIK Is a Protein Kinase-As mentioned above, DIK contains in its N-terminal domain all characteristic subdomains that, according to Hanks and Quinn (43), are conserved in protein kinases. Therefore, it was of major interest to demonstrate that DIK indeed exhibited kinase activity. Recombinant His-DIK purified from baculovirus-infected insect cells (see above) was used for the kinase assays. In the presence of [ 32 P]ATP His-DIK was able to autophosphorylate, as shown in Fig. 5A (lane  3 and 4). Upon PAGE and autoradiography, incorporation of labeled phosphate into His-DIK was clearly visible. No phosphorylation of His-DIK was observed in a control sample that had been prepared from an extract of uninfected cells according to the purification procedure of His-DIK (Fig. 5B, left lane). In the presence of 1 M staurosporine, that is known to inhibit many protein kinases including PKCs (47), autophosphorylation of His-DIK was not suppressed, whereas autophosphorylation of His-PKC␦ was almost completely abolished (Fig. 5A). Thus, staurosporine may become a valuable tool for differentiating DIK kinase activity from that of other protein kinases, particularly PKCs. According to a phosphoamino acid analysis (Fig. 6), autophosphorylation was predominantly on serine and, to some extent, on threonine but not on tyrosine residues.
In addition to its autophosphorylation capacity, His-DIK was able to phosphorylate substrate proteins, such as histone. Incubation of purified His-DIK with histone III-S in the presence of [ 32 P]ATP resulted in an efficient incorporation of phosphate into the histone (Fig. 5B). When using the control sample from uninfected cells instead of His-DIK, phosphorylation of histone was lacking. This indicated that the substrate phosphorylation was indeed due to His-DIK. Several other proteins served as substrates for His-DIK, such as aldolase and myelin basic protein (data not shown). Similar to the autophosphorylation (Fig. 6), the substrate proteins histone III-S and myelin basic protein were phosphorylated by DIK exclusively on serine and threonine residues (data not shown). These results clearly show that DIK is a serine/threonine kinase, being in accordance with the amino acid sequences of the kinase subdomains that contain motifs characteristic for serine/threonine kinases  6. Phosphoamino acid analysis of autophosphorylated DIK. Purified His-DIK was autophosphorylated, and the phosphoamino acid analysis was performed as described under "Experimental Procedures." The sample was hydrolyzed, markers were added, and the phosphoamino acids were separated by two-dimensional electrophoresis. 32 P-Labeled phosphoamino acids were visualized by autoradiography. The locations of the phosphoamino acid markers phosphoserine (P-ser), phosphothreonine (P-thr), and phosphotyrosine (P-tyr) that had been visualized by treatment with ninhydrin are indicated by arrows.
(see above). As yet we have not successfully phosphorylated PKC␦ with His-DIK. Similarly, we have not been able so far to unequivocally demonstrate phosphorylation of His-DIK by PKC␦. A major problem is the autophosphorylation of His-DIK that cannot be easily differentiated from a putative phosphorylation of His-DIK by PKC␦. Therefore, it has to be considered that the interaction of both protein kinases might not be due to an enzyme-substrate relationship but might rather have another function, for instance participation of both proteins in a signaling complex.
These results prove that DIK is a protein kinase able to autophosphorylate and phosphorylate various proteins in vitro. Bacterially expressed GST-DIK did not exhibit any protein kinase activity (data not shown). DIK is synthesized in bacteria in a form that is probably unmodified, as shown above (Fig.  4B). Thus, it is likely that the modification observed upon synthesis of DIK in eucaryotic cells is required for the catalytic competence of the kinase.
Interaction of DIK with PKC␦ and the Catalytic Domain of PKC␦ (PKC␦ cat )-DIK had been cloned using the YTHS with the cDNA of PKC␦ cat as bait and thus could be assumed to interact with this domain of PKC␦. To prove this interaction, HEK cells were transiently cotransfected with FLAG-DIK and PKC␦ cat or PKC␦ expression constructs. Extracts of the transfected cells expressing FLAG-tagged DIK and PKC␦ cat or PKC␦ were used for immunoprecipitation either applying an anti-FLAG (Fig. 7, A, B, and E) or an anti-PKC␦ antibody (Fig. 7, C  and D). The immunoblots of these immunoprecipitates clearly demonstrated coimmunoprecipitation of PKC␦ cat (Fig. 7A) and PKC␦ (Fig. 7E) with FLAG-DIK as well as of FLAG-DIK with PKC␦ cat (Fig. 7C), thus unequivocally indicating an interaction of DIK with PKC␦ and PKC␦ cat . The intensities of the PKC␦ cat and PKC␦ bands in Fig. 7 (A and E), respectively, appear to indicate that DIK interacts better with the catalytic domain of PKC␦ than with the holoenzyme. Panels B and D of Fig. 7 serve as controls proving that FLAG-DIK and PKC␦ cat are indeed immunoprecipitated with the FLAG and PKC␦ antibody, respectively. No unspecific precipitation of PKC␦, PKC␦ cat , and FLAG-DIK was observed with extracts from cells transfected with the empty vectors (pFLAG and pCDNA). Intriguingly, in the YTHS interaction of DIK with the catalytic domain of PKC␦ only and not with its regulatory domain (PKC␦ reg ) was observed (data not shown). The specificity of the interaction of DIK with the catalytic domain of PKC␦ was confirmed by the coimmunoprecipitation assay. Upon transfection of HEK cells with FLAG-DIK and PKC␦ reg expression constructs, no coimmunoprecipitation of either FLAG-DIK with PKC␦ reg or PKC␦ reg with FLAG-DIK could be demonstrated (data not shown). The immunoprecipitation assay was performed essentially as described in Fig. 7, except that an anti-PKC␦ antibody (P36520; Transduction Laboratories) was used that recognizes the N terminus of PKC␦.
Taken together, we cloned a novel protein kinase, termed DIK, that contains a catalytic kinase domain and a putative regulatory domain with ankyrin-like repeats and a nuclear localization signal and that associates with PKC␦ in vivo. The role of this protein kinase in signal transduction and its association with PKC␦ is not yet clear and awaits further investigation, for example studies with DIK mutants (e.g. kinase negative and ankyrin repeat deletion mutants) and searching for other DIK interacting proteins, particularly physiologically relevant substrates.