Cloning and chromosomal location of a novel member of the myotonic dystrophy family of protein kinases.

We have cloned a novel serine/threonine protein kinase (PK428) which is highly related (65%) within the kinase domain to the myotonic dystrophy protein kinase (DM-PK), as well as the cyclic AMP-dependent protein kinase (33%). Northern blots demonstrate that PK428 mRNA is distributed widely among tissues and is expressed at the highest levels in pancreas, heart, and skeletal muscle, with lower levels in liver and lung. Two PK428 mRNAs 10 and 3.8 kilobase pairs in size are seen in a number of cell lines, including hematopoietic and breast cancer cells. An antibody generated to a glutathione S-transferase-PK428 fusion protein detects a 65-kDa protein in these cell lines, and a similarly sized protein when the cloned cDNA is transiently expressed in Cos 7 cells. Immunoprecipitation of the transiently expressed PK428 protein and incubation with [gamma-32P]ATP demonstrate that it is capable of autophosphorylation. In addition, immunoprecipitates of the PK428 protein kinase also phosphorylated histone H1 and a peptide encoding a cyclic AMP-dependent protein kinase substrate. The gene corresponding to the 3.8-kb PK428 mRNA, and its corresponding 65-kDa protein, was isolated by polymerase chain reaction screening of a P1 phage human genomic library. Using this P1 phage clone as a probe, the PK428 gene was located on 1q41-42, a possible location for a human senescence gene, a gene associated with Rippling muscle disease, as well as a region associated with genetically acquired mental retardation.

male baldness (1,2), with an incidence of 1 in 8000 (1). This disease can present either as a slowly progressive form in adults or as a congenital form in childhood (1,2). The molecular basis of DM involves the mutation and expansion of a trinucleotide sequence (CTG) n located in the 3Ј-untranslated region of the corresponding mRNA (3,4). The DM gene encodes a putative serine-threonine protein kinase (DM-PK) which is located in chromosome 19q13.3 (3)(4)(5). This protein kinase is predicted to contain a protein kinase catalytic domain near the amino terminus, a central ␣-helical coiled domain, and a potential carboxyl-terminal transmembrane domain (3). Consistent with the systemic nature of this disease, transcripts of the gene are expressed in various tissues, including the heart, skeletal muscle, liver, and brain, both in human and mouse (6,7).
Another protein kinase related to the DM-PK, p160 ROCK , has been shown to bind the small GTP-binding protein Rho, and it contains a protein kinase domain that shares 44% similarity to the DM-PK, an amphipathic ␣-helix, a pleckstrin domain, and a cysteine-rich region (8 -10). Since the activation of Rho regulates the actin cytoskeleton, it is possible that the DM-PK has a similar or parallel function. The warts protein kinase, which is found in Drosophila melanogaster, is also related to the DM-PK (46% identity); however, the biologic function of this protein is unknown. Deletion of the warts gene in Drosophila leads to the formation of cell clones that are fragmented, rounded, and greatly overgrown (11).
Signaling through the granulocyte-macrophage colony-simulating factor (GM-CSF) receptor is mediated by two subunits, an ␣ subunit, which binds GM-CSF and interacts with the ␤ subunit containing the binding site for the JAK2 protein kinase. The JAK2 protein kinase has been shown to regulate at least some of the signaling mediated by this receptor. The ␣ subunit has a short intracytoplasmic carboxyl-terminal tail (54 amino acids), which is essential for GM-CSF-mediated growth stimulation (12)(13)(14)(15). Within this 54-amino acid tail is a short stretch of prolines, a sequence that is conserved in the ␣ subunits of both the interleukin-3 and -5 receptors. Mutation of one or more of these proline residues blocks the ability of these hormones to signal and act as stimulators of growth (14,15). Because the intracytoplasmic region of the ␣ subunit is essential for the function of GM-CSF, we have used this portion of the receptor in a two-hybrid yeast interaction screen to search for proteins that may be essential for signal transduction by this hormone.
We have cloned a novel protein kinase, PK428, which is highly related to the DM-PK amino-terminal kinase domain, but which diverges in its carboxyl terminus. This novel protein kinase is highly expressed in skeletal muscle and heart, but it is also expressed in the brain, pancreas, and at lower levels in the lung. Northern blots demonstrate that this kinase is ex-* This work was supported in part by United States Public Health Service Grants DK44741 (to A. S. K.) and GM 44088 (to V. J. K.), National Institutes of Health Cancer Center Support Core Grant CA 21765 to St. Jude Children's Research Hospital, and by support from the American Lebanese Syrian Associated Charities (ALSAC) (to V. J. K.). 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  1 The abbreviations used are: DM, myotonic dystrophy; PK, protein kinase; GM, granulocyte-macrophage; CSF, colony-stimulating factor; PCR, polymerase chain reaction; FISH, fluorescence in situ hybridization; GST, glutathione S-transferase; PAGE, polyacrylamide gel electrophoresis; DAPI, 4Ј,6-diamidine-2Ј-phenylindole dihydrochloride. pressed as two separate mRNAs, 3.8 and 10 kb in size, in both leukemic and breast cancer cells, but not in HeLa (cervical carcinoma) cells. Transfection of this cDNA into Cos 7 cells results in the production of a 65-kDa protein, which is capable of autophosphorylation, as well as phosphorylation of histone H1 and a peptide substrate that contains a cyclic AMP-dependent protein kinase phosphorylation site. An antibody generated to the PK428 protein kinase immunoprecipitated a 65-kDa protein (corresponding closely to the predicated size of the open reading frame in a 3.8-kb PK428 mRNA) in both hematopoietic and breast cancer cells, as well as a larger 200-kDa protein (which presumably corresponds to a larger 10-kb PK428 mRNA). The chromosomal gene corresponding to the 3.8-kb PK428 mRNA was isolated by PCR screening of a human genomic P1 phage library. Two identical clones were isolated, and their identity was confirmed by partial DNA sequence analysis. Using fluorescence in situ hybridization (FISH) of the PK428 P1 phage clones, we found that the PK428 gene encoding the 3.8-kb mRNA and 65-kDa protein localizes to the long arm of human chromosome 1, specifically the q41-42 region. This area has previously been implicated in some forms of genetically acquired mental retardation (16), as a possible location for a second human senescence gene localized to the long arm of chromosome 1 (17), and most interestingly, as a location of a gene(s) involved in Rippling muscle disease (18).

MATERIALS AND METHODS
Cloning of Full-length cDNA-The PK428 cDNA fragment was identified using the yeast two-hybrid system (a gift of Dr. Steven Elledge). The 54-amino acid intracytoplasmic tail of the GM-CSF ␣ subunit was cloned into plasmid pAS1-CYH2 as the bait. The yeast Y190 cells were cotransformed with a peripheral B lymphocyte library fused to the DNA activation domain of GAL4 that is contained within the plasmid pACT II. The two-hybrid screen was carried out as described previously (19). This screening yielded 10 positive cDNA clones, of which three were identified as PK428 cDNA clones. To obtain a full-length cDNA, the largest PK428 fragment was used as probe to screen a ZAP expression cDNA library constructed from five breast cancer cell lines (a gift of Dr. M. Ruppert, University of Alabama) using the ZAP Express TM cDNA synthesis kit (Stratagene). The prehybridization and hybridization were carried out in blotto buffer (0.05% heparin, 1% SDS, 0.5% non-fat dry milk, 6% PEG8000, 5 ϫ SSPE, 10% formamide, 2 mg/ml salmon sperm single-strand DNA) at 65°C for 3 and 16 h, respectively. The filters were washed with 0.1 ϫ SSC, 0.3% SDS for 5 min at room temperature and 0.3 ϫ SSC, 0.3% SDS for 60 min at 65°C. DNA sequence analysis was performed by the dideoxy method using Sequenase 2.0 according to the manufacturer's (U. S. Biochemical Corp.) specifications. Homology searches were performed with the BLAST network service at National Center for Biotechnology Information.
Plasmid Construction-The pcDNA3PK428/2.8 plasmid was constructed by digesting pBK/CMV PK428, the longest clone from ZAP breast cancer cell line cDNA library screening, with BamHI/XhoI and inserting the released 2.8-kb fragment into the BamHI/XhoI site of pcDNA3. The pcDNA3PK428/2.0 was constructed by digesting pcDNA3PK428/2.8 with SmaI/XhoI and inserting the about 2.0-kb fragment into EcoRV/XhoI site of pcDNA3. The pcDNA3PK428/1.5 was constructed by digesting the pcDNA3PK428/2.8 with AflIII followed by Klenow enzyme treatment, and the resulting product was then digested with XhoI. The released 1.5-kb fragment was then ligated into the EcoRV/XhoI site of pcDNA3. The GSTPK428 used for raising polyclonal antibody was constructed by inserting a PCR fragment from nucleotide 1748 -2818 of pcDNA3PK428/2.8 with BamHI and EcoRI sites at its 5Ј and 3Ј ends, respectively, into the BamHI/EcoRI site of pGEX2T (Pharmacia Biotech Inc.). The plasmid pGEX2TK encoding a fusion protein between GST and a cAMP-dependent protein kinase substrate (20) (GST-TK) was kindly provided by Dr. C. Franklin (University of Colorado, Denver).
The plasmids encoding either single proline mutants or an 8-amino acid deletion in the ␣ subunit intracytoplasmic domain (pAS/␣7400, pAS/␣7401, and pAS/␣7402) were constructed by first cloning the intracytoplasmic domain of the receptor into M13mp19 phage, which was then used to prepare single-stranded DNA for site-directed mutagenesis (Amersham Corp.). The oligonucleotides used for the mutagenesis were as follows: GTCTTTGATCTGTATCCTAAGGAACC (7402), GGA-ACTGGACCGAACAGC (7401), and CTTTGATCTGACCAACTGGCGG (7400). The cytoplasmic regions of the resulting mutants were cloned into the NcoI/BamHI site of pAS/CYH2 plasmid by polymerase chain reaction.
Northern Blotting Analysis-A human tissue blot with mRNAs from different tissues was purchased from Clontech (a gift from Dr. J. Han, University of Alabama at Birmingham). Each lane contained 2 g of poly(A) ϩ mRNA. Total RNA was prepared from various cell lines by lysing the cells in guanidinium isothiocyanate solution and pelleting the RNA through a cesium chloride cushion. The human cell lines used were U937 (a histiocytic lymphoma), K562 (a chronic myelogenous leukemia), MDA-MB-231 (a breast cancer cell line), A549 (a lung cancer cell line), G-401 (a Wilms' tumor cell line), Colo 320 (a colon cancer cell line), SK-OV-3 (an ovarian adenocarcinoma cell line), MOLT-4 (T cell line), HL-60 (a promyelocytic leukemia cell line), PLB985 (a myeloid leukemia cell line), HeLa (an epithelial cancer), and IE8 and OB5 BLIN-1 (pre-B cell lines). All cell lines were obtained from the ATCC, with the exception of IE8 and OB5 which were a gift of Dr. M. Cooper, (University of Alabama at Birmingham). 20 g of total RNA were electrophoresed on a 1% agarose-formaldehyde gel and transferred onto nylon membranes. The filters was prehybridized and hybridized as described previously (12).
Transfection of PK428 cDNA into Cos 7 Cells-Cos 7 cells were maintained in Dulbecco's modified Eagle's medium containing 10% bovine calf serum. For transfections, 3-5 g of the PK428 cDNA plus 20 g of Lipofectin reagent (Life Technologies, Inc.) were incubated with 7 ϫ 10 5 Cos 7 cells for 6 h. Forty-eight hours after transfection, the cells were lysed, and the lysate was used as a source of the expressed PK428 protein.
Immunoprecipitation and Western Blotting-An antibody to the PK428 protein was raised by injecting an SDS-gel-purified fusion protein of glutathione S-transferase and amino acids 154 -496 of PK428 into a New Zealand White rabbit every 3 weeks. The method for immunogen injection has been described previously (13).
Immunoprecipitations were carried out by lysing the Cos 7 cells in TNE buffer (50 mM Tris, pH 8.0, 150 mM NaCl, 1% Nonidet P-40 2 mM EDTA) containing inhibitors (12). The anti-PK428 antibody was added to the supernatant of the cell lysate at a 1:100 dilution and incubated at 4°C for 1 h. 30 l of a protein A-Sepharose (Pharmacia) slurry were used to adsorb the immune complexes, and the beads were then washed five times in lysis buffer prior to their elution in Laemmli sample buffer. The eluted proteins were electrophoresed on an 8% SDS-PAGE and, for Western blots, electrophoretically transferred to a polyvinylidene difluoride membrane filter (Millipore, Bedford, MA). The blotted filter was incubated in TBS (20 mM Tris, pH 7.6, 137 mM NaCl) containing 3% bovine serum albumin (fraction V, Sigma) for 1 h. The filter was incubated with the PK428 antibody at 1:200 dilution. Bound antibodies were visualized by using the enhanced chemiluminescence (ECL) system (Amersham Corp.).
In Vivo Phosphorylation, Autophosphorylation, and Protein Kinase Assays-Cos 7 cells were transfected with pcDNA3PK428/1.5, pcDNA3PK428/2.0, or pcDNA3 vector only, and labeled with 0.5 mCi/ml [ 32 P]orthophosphate (Amersham) for 4 h prior to cell lysis and then immunoprecipitated with the PK428 antibody or preimmune serum. For autophosphorylation, the immunoprecipitated PK428 protein was washed twice in TNE lysis buffer and twice in kinase buffer (20 mM Tris, 100 mM NaCl, 1 mM dithiothreitol) followed by incubation in 50 l of the assay buffer containing 20 mM Tris, pH 7.5, 100 mM NaCl, 10 mM MgCl 2 , 1 mM dithiothreitol, 10 M cold ATP, and 5 mCi of [␥-32 P]ATP at a specific activity of 3000 mCi/mM (Amersham) for different time periods at 30°C. The immunokinase assay was performed at 30°C using the assay buffer described for autophosphorylation with or without 5 g of each substrate. The reactions were stopped by dilution with Laemmli sample loading buffer. The phosphorylated proteins were then separated by SDS-PAGE and visualized by autoradiography.
Isolation of the Human PK428 Gene-A human genomic P1 phage library (Genome Systems, St. Louis, MO) was screened by PCR using a pair of oligonucleotide primers (forward, 5Ј-TATTCATTAATGATG-CAACCGGAT-3Ј, and reverse, 5Ј-GTTATTCAAACAACTGTCATGCAA-3Ј), the sequences of which were derived from the PK428 cDNA corresponding to nucleotides 870 -893 and 1162-1185, respectively. These sequences are located in the 5Ј-untranslated region of the cDNA and give rise to a single 315-base pair PCR product using normal human genomic DNA (data not shown). Conditions for the genomic PCR using the PK428 primers were as follows: 0.25 g of template DNA, 1000 ng of each primer, 200 M of each dNTP, and 1.5 mM MgCl 2 in a 50-l reaction volume. The PCR cycles used for these reactions were 95°C initial denaturation to 3 min, 95°C denaturation for 30 s, 62°C anneal-ing for 1 min, and a 72°C extension for 30 s for 35 cycles. Two P1 phage clones containing the human PK428 gene (11499 and 11500) were isolated and analyzed further by direct DNA sequence analysis to confirm their identity (data not shown).
Fluorescence in Situ Hybridization-Bromodeoxyuridine-synchronized and phytohemagglutinin-stimulated peripheral blood lymphocytes from a normal male donor were used as a source of metaphase chormosomes. Purified DNA from the P1 phage clones containing the human PK428 gene (11499 and 11500) were labeled for FISH analysis by nick translation with digoxigenin-11-UTP (Boehringer Mannheim). To localize the human PK428 gene, one set of metaphase chromosomes was simultaneously hybridized with a control genomic centromeric clone corresponding to alphoid sequences of human chromosome 1, D1Z1 (Oncor, Inc.). Specific hybridization signals were detected by applying fluorescein-conjugated sheep antibodies to digoxigenin (Boeh-ringer Mannheim) and avidin-Texas Red (Vector Laboratories, Inc., Burlingame, CA), followed by counterstaining in DAPI (Sigma). Fluorescence microscopy was performed with a Nikon microscope equipped with a cytovision image analysis system (Applied Imaging, Pittsburgh, PA) and a fluorescence filter wheel that sequentially captures the fluorescein, Texas Red, and DAPI images and electronically superimposes them.

RESULTS
Isolation of a Full-length cDNA Clone for PK428 -The PK428 clone was isolated using the yeast two-hybrid screen designed to identify proteins that bind to the intracytoplasmic portion of the ␣ subunit of the GM-CSF receptor. This 54-amino acid intracytoplasmic domain of the GM-CSF ␣ subunit ap- pears to play an important role in signal transduction, although the proteins that bind to this domain remain unknown (12,13). We fused the 54-amino acid intracytoplasmic domain to the DNA binding portion of GAL4 as the bait for the twohybrid analysis, and a B cell library fused to the GAL4 DNA transactivation domain was used as the prey. Ten positive clones were obtained, three of which encoded PK428. The PK428 cDNA fragment obtained from the yeast two hybrid screening was used as a hybridization probe to isolate additional cDNAs from a human breast cancer cell cDNA library. Nine positive clones were isolated and analyzed by DNA sequencing. The PK428 insert contains a 2,818-base pair cDNA fragment (GenBank TM accession no. U59305) with a large open reading frame found from nucleotide 1289 -2776, encoding a 496-amino acid protein (Fig. 1). The designated ATG is likely to be correct, since several stop codons in all three reading frames are found 5Ј of this ATG codon. Analysis of the amino acid sequence of PK428 demonstrates that it contains a complete protein kinase catalytic domain which conserves all 11 kinase subdomains ( Figs. 1 and 2) (21). In addition to the high homology kinase region between DM-PK and PK428, the aminoterminal region of the deduced PK428 amino acid sequence is also related (to a lesser extent) to the DM-PK, suggesting that the two proteins are unique. Hydrophobicity analysis (22) shows that the PK428 amino acid sequence contains a helical region following the kinase domain, as well as a hydrophobic domain. Both of these regions, which are similar to those found in DM-PK, suggest that PK428 is most likely a protein kinase related to the DM-PK (Fig. 1).
The catalytic domain of the PK428 protein kinase corresponds to a 263-amino acid region that is located in the amino terminus of the protein. The GXGXXG sequence motif characteristic of many protein kinases is found in subdomain I. An invariant lysine at position 106 of the protein that is necessary for the ATP binding is also present (15), as well as the conserved amino acids DIKPEN in subdomain VI and GTPDYL-SPE in subdomain VIII. These features of the predicted PK428 open reading frame strongly suggest that this protein is a serine/threonine protein kinase. A sequence homology search demonstrated that the kinase domain of PK428 was 65% homologous to the kinase domain of the myotonic dystrophy protein kinase (23), 50% homologous to the kinase domain of p160 ROCK (10), 51% homologous to the rat ROK␣ protein (8), 46% to the D. melanogaster tumor suppressor protein kinase warts (11), 40% homologous to the Cot-1 protein kinase from Neurospora crassa (24), 43% homologous to the serine/threonine protein kinase from Mesembryanthenum crystallinum, 2 and 33% homologous to the catalytic subunit of protein kinase A (25) (Fig. 2).
Tissue and Cell Type Expression of PK428 -To examine the tissue expression of PK428 mRNA, Northern blot analysis of poly(A) ϩ RNA from different human tissues and cell lines was performed. A 10-kb mRNA was most abundant in the heart, brain, skeletal muscle, kidney, and pancreas, with little or no expression in the lung and liver (Fig. 3A). Northern blot analysis of various human cell lines demonstrates the presence of 3.8-and 10-kb hybridizing bands (Fig. 3B) To verify the identity and molecular weight of this putative protein kinase, a polyclonal antibody was generated using a GST-PK428 fusion protein. MDA-MB-231 and U937 cells were lysed, and PK428 protein was immunoprecipitated and Western blotted with this antibody (Fig. 4A). In both cell lines a protein with a apparent molecular mass of 65 kDa, as well as a less intense band at 78 kDa, were immunoprecipitated. In addition, in the U937 cells, a 200-kDa band was also seen. To determine which of these polypeptides was encoded by the cloned cDNA, and its molecular weight on SDS-PAGE gels, Cos 7 cells were transfected with each of two PK428 cDNA clones (2.0 and 1.5 kb), which contained the entire PK428 predicted open reading frame and varying lengths of the 5Ј-untranslated region. A 65-kDa protein was immunoprecipitated from Cos 7 cells transfected with both of the PK428 cDNAs, suggesting that they both utilized an identical site for initiation of translation (Fig. 4B). These analyses demonstrated that the PK428 cDNA clones that were isolated corresponded to the 65-kDa protein species identified by the PK428 antibody in human cell lines.
Protein Kinase Activity Associated with the PK428 Protein-Since many protein kinases are capable of autophosphorylation, the in vivo phosphorylation state of the PK428 protein was examined. Cos 7 cells were transfected with the PK428 cDNA and labeled with [ 32 P]orthophosphate, and the PK428 protein was immunoprecipitated with the PK428-specific antibody. A 65-kDa phosphorylated protein was detected in the Cos 7 cells transfected with the cDNA (Fig. 5A, lane 3), but it was not present in either preimmune cells (lane 2) or when this antibody was used for immunoprecipitation from the untransfected cells (lane 1). To demonstrate in vitro that this protein  lanes 1 and 4), pcDNA3PK428/2.0 (lane 2), or pcDNA3 alone (lane 4). The cells were lysed, the PK428 protein was immunoprecipitated with either the 2188 antibody or preimmune serum, the proteins were transferred to a membrane, and a Western blot was performed with the 2188 antibody.
was capable of autophosphorylation, the PK428 protein from transfected and unlabeled Cos 7 cells was immunoprecipitated with the PK428 antibody, and the resulting protein was incubated in a kinase reaction buffer containing [␥-32 P]ATP. In this autophosphorylation assay, a 65-kDa band was seen within 1 min, and the appearance of this 65-kDa band increased in intensity with continued, longer incubation times (Fig. 5B). Antiphophotyrosine Western blotting of PK428 immunoprecipitates demonstrated that this protein is not phosphorylated on tyrosine residues.
Since the PK428 protein was homologous to the protein kinase domain to the cyclic AMP-dependent protein kinase, the ability of the PK428 protein kinase to phosphorylate a cyclic AMP protein kinase substrate was examined (25). Demonstration of PK428 protein kinase activity was measured by the incorporation of [ 32 P]phosphate from [␥-32 P]ATP into either histone H1 or a peptide substrate of cyclic AMP-dependent protein kinase encoded as part of a GST fusion protein (GST-TK) (20). PK428 immunoprecipitates phosphorylated the GST-TK fusion protein (Fig. 5C), but did not phosphorylate GST alone (Fig. 5C, lane 5). When the PK428 antiserum was replaced with normal rabbit serum, no phosphorylation of the substrates was seen (data not shown). Likewise, the PK428 immunoprecipitates phosphorylate histone H1 when it is used as a substrate (Fig. 5C), while no phosphorylation of histone H1 was seen if the preimmune serum was used as a control (Fig.  5C), or if histone H1 is not included in the reaction (data not shown).
Isolation of the Human PK428 Gene and Its Chromosome Localization-Using oligonucleotide primers specific for a region of the 5Ј-untranslated region of the PK428 cDNA, nucleotides 870 -1185, a 315-bp specific band was amplified from normal genomic DNA (data not shown). This confirmed the specificity and continuity of the selected cDNA sequences in genomic DNA, allowing the isolation of the PK428 gene by genomic PCR screening a gridded human P1 phage genomic library (Genome Systems). The corresponding human PK428 gene was isolated from this human P1 phage library (clone identities 11499 and 11500), verified by direct sequencing of the P1 DNA with the same primers, and then used to determine the location of the human gene by FISH. Both PK428 P1 phage clones localized to the same location, human 1q41-42 (Fig. 6). Previously, others have shown that this terminal region of the long arm of chromosome 1 is associated with certain forms of mental retardation (16) and with the possible location of a second human senescence loci (17). Of more interest, the gene(s) which is associated with Rippling muscle disease has also been localized to this same region of chromosome 1 (18). DISCUSSION A novel protein kinase whose kinase domain is highly related, but not identical, to the DM-PK has been isolated by the two-hybrid analysis using a 54-amino acid region of the GM-CSF ␣ subunit. Regions outside of the protein kinase domain demonstrate little amino acid homology with the DM-PK, although a short hydrophobic stretch of amino acids in the carboxyl-terminal portion of the protein is also conserved. In ad- The arrows indicate the signal due to specific hybridization with clone 11499 (generated by the fluorescein antibody; green dots), which contains the PK428 gene, while the arrowheads indicate the signal created by hybridization with the PUC 1.77 control DNA (generated by the Texas Red antibody; red dots), which is specific for human chromosome 1q heterochromatin. Also seen in the upper left hand corner is an interphase nucleus hybridized with the same probes. dition to these sequence differences between the DM-PK and PK428 protein kinases, additional data demonstrate that the PK428 protein is unique. The PK428 protein kinase described here has a molecular mass of 65 kDa and undergoes autophosphorylation as well as phosphorylating histone H1 and a cyclic AMP-dependent protein kinase substrate.
In addition to the 65-kDa form of the PK428 protein, a 70 -80-kDa protein was immunoprecipitated from breast and hematopoietic cells. The myotonic dystrophy protein kinase is 71-80 kDa in size (26), and it is possible that the PK428 antibody detects this protein as well. Furthermore, the PK428 antibody detects a 200-kDa protein in U937 cells. This larger protein could be another member of the DM-PK family. Recently, a 160-kDa protein with significant homology to the DM-PK kinase domain, p160 ROCK , has been shown to interact with the small GTP-binding protein Rho, suggesting that there are larger members of this family of protein kinases (8 -10). It is also possible that this 200-kDa protein is derived by alternative splicing of the PK428 gene, or that it corresponds to a PK428-related gene and protein. In fact alternative splicing of the DM-PK gene has been reported (27). The identity of these PK428-related mRNAs may be revealed by analysis of its corresponding gene and/or related genes.
Protein kinase activity of the PK428 protein was demonstrated by immunoprecipitation of PK428 produced by transient transfection. This kinase activity stimulated autophosphorylation and demonstrated that PK428 protein is phosphorylated in vivo. This PK428 protein kinase is also capable of phosphorylating a substrate for the cyclic AMP-dependent protein kinase and histone H1. However, we have not eliminated the possibility that another protein that coimmunoprecipitates with this protein kinase is responsible for some of the effects we have described. Although difficult to quantify, the phosphorylation of cyclic AMP-dependent protein kinase substrate that was seen with the PK428 immunoprecipitates was relatively weak when compared with the catalytic subunit of the cyclic AMP-dependent kinase (data not shown). These results suggest that "better," more physiological, substrates for this protein kinase may exist.
Using FISH we have mapped this gene to human chromosome 1q41-42. A number of diseases have been localized to this region, including Usher's syndrome type IIa, which is associated with hearing loss, and retinitis pigmentosa (28), a syndrome of mental retardation associated with trisomy 1q42 (16), arrythmogenic right ventricular cardiomyopathy (29), and a Rippling muscle disease gene (18). Because of the similarity of PK428 with the DM-PK it is interesting to speculate that PK428 gene may also encode a protein which has an important role in muscle physiology.
Overexpression of the DM-PK has been shown to induce a skeletal muscle phenotype in BC 3 H1 cells (30). The neuromuscular junction location of the DM-PK protein (31) also suggests that it may be involved in signaling integrating extracellular events to cytoskeleton and the nucleus. Mutations in either the N. crassa Cot-1 or the D. melanogaster warts genes, which are related to DM-PK, result in abnormal cell growth and changes in cell morphology, supporting a possible role for this protein kinase in cytoskeletal morphology and signal transduction to the nucleus. In addition, the Rho protein regulates stress fibers and focal contacts in cells, presumably through its interaction with a protein kinase that shares sequence homology with the DM-PK. These signals appear to help to control cell shape. Furthermore, the DM-PK-like protein kinase that binds to Rho, p160 ROCK , may coordinate responses to specific external stimuli such as lysophosphatidic acid (32). Thus, it would appear that this family of protein kinases regulates responses to ex-ternal stimuli that are involved in potential changes of the actin cytoskeleton and, subsequently, cell morphology. Identification of PK428 by two-hybrid interactive cloning, using the short intracytoplasmic tail of the GM-CSF ␣ subunit as a bait, suggests that this protein kinase may be relevant to GM-CSF ␣ subunit signaling. The observation that deletion of the prolinerich segment or mutation of specific prolines to glycines blocks this interaction (data not shown) suggests that PK428 binds to a portion of this receptor, which is critical for hematopoietic signaling.
The addition of GM-CSF ␣ subunit to hematopoietic cell types induces marked changes in cell shape and changes in locomotion, cell division, and differentiation. However, although this PK428 protein kinase was isolated due to its ability to bind to the intracytoplasmic tail of the GM-CSF ␣ subunit in yeast, demonstration of a strong interaction (i.e. coimmunoprecipitation) between the two proteins in hematopoietic cells has proven to be difficult. The exact role of PK428, if any, in hematopoietic signaling must await further analysis.