Identification of a protein kinase from Dictyostelium with homology to the novel catalytic domain of myosin heavy chain kinase A.

Myosin II assembly and localization into the cytoskeleton is regulated by heavy chain phosphorylation in Dictyostelium. The enzyme myosin heavy chain kinase A (MHCK A) has been shown previously to drive myosin filament disassembly in vitro and in vivo. MHCK A is noteworthy in that its catalytic domain is unrelated to the conventional families of eukaryotic protein kinases. We report here the cloning and initial biochemical characterization of another kinase from Dictyostelium that is related to MHCK A. When the segment of this protein that is similar to the MHCK A catalytic domain was expressed in bacteria, the resultant protein displayed efficient autophosphorylation, phosphorylated Dictyostelium myosin II, and also phosphorylated a peptide substrate corresponding to a portion of the myosin II tail. We have therefore named this gene myosin heavy chain kinase B. These results provide the first confirmation that sequences in other proteins that are related to the MHCK A catalytic domain can also encode protein kinase activity. It is likely that the related segments of homology present in rat eukaryotic elongation factor-2 kinase and a putative nematode eukaryotic elongation factor-2 kinase also encode the catalytic domains of those enzymes.


Myosin II assembly and localization into the cytoskeleton is regulated by heavy chain phosphorylation in
Dictyostelium. The enzyme myosin heavy chain kinase A (MHCK A) has been shown previously to drive myosin filament disassembly in vitro and in vivo. MHCK A is noteworthy in that its catalytic domain is unrelated to the conventional families of eukaryotic protein kinases. We report here the cloning and initial biochemical characterization of another kinase from Dictyostelium that is related to MHCK A. When the segment of this protein that is similar to the MHCK A catalytic domain was expressed in bacteria, the resultant protein displayed efficient autophosphorylation, phosphorylated Dictyostelium myosin II, and also phosphorylated a peptide substrate corresponding to a portion of the myosin II tail. We have therefore named this gene myosin heavy chain kinase B. These results provide the first confirmation that sequences in other proteins that are related to the MHCK A catalytic domain can also encode protein kinase activity. It is likely that the related segments of homology present in rat eukaryotic elongation factor-2 kinase and a putative nematode eukaryotic elongation factor-2 kinase also encode the catalytic domains of those enzymes.
Myosin II has central roles in a variety of cellular motility events, including cytokinesis, cell locomotion, and morphological changes that occur during development. In Dictyostelium, myosin heavy chain (MHC) 1 phosphorylation on mapped threonine residues in the tail seems to be a mechanism regulating bipolar myosin filament assembly and localization into the cytoskeleton (1).
Two distinct MHC kinases have been previously implicated in myosin filament regulation. In vitro each of these enzymes is capable of driving myosin filament disassembly. A kinase that has been named MHC-PKC was purified from developed Dictyostelium cells (2) and subsequently cloned (3). This enzyme has a catalytic domain similar to members of the protein kinase C family and also contains a distinct domain with significant similarity to the catalytic domain of diacylglycerol kinases (4). This enzyme is expressed in a developmental-specific manner and seems to be involved in regulating myosin assembly/disassembly during chemotactic cell migration (5,6). The enzyme MHCK A was first purified from growth-phase cells (7) and is expressed during both growth and development.
Molecular analysis of the MHCK A gene has revealed a highly novel three-domain structure (8). MHCK has an aminoterminal domain of ϳ70 kDa that has a predicted coiled-coil structure, a central domain of ϳ30 kDa, and a carboxyl-terminal domain of ϳ30 kDa that consists of 7 WD or G ␤ -like repeats (9,10). The deduced polypeptide sequence of MHCK A displays no detectable similarity to conventional eukaryotic protein kinases or to members of the histidine protein kinase family (11). Truncation analysis, coupled to expression in Escherichia coli, was used to demonstrate that the central domain of MHCK A (residues 550 -841) contains all of the catalytic protein kinase functions (12). This segment of MHCK A displays substantial identity (ϳ42%) to a Caenorhabditis elegans cDNA (yk3f11) that has been sequenced as part of the nematode genome project (GenBank accession numbers U10414 and D27775). More recently, a cDNA sequence has been reported (13) for rat skeletal muscle eukaryotic elongation factor-2 (eEF-2) kinase (also known as calcium/calmodulin-dependent kinase III). The deduced protein sequence of the eEF-2 kinase also displays a high degree of identity to the catalytic domain of MHCK A (43%; see Ref. 12). Based upon similarity in the flanking regions, it seems that the C. elegans yk3f11 cDNA may encode a nematode homologue of eEF-2 kinase. As with MHCK A, neither rat eEF-2 kinase nor the nematode gene displays any significant similarity to conventional eukaryotic protein kinases.
The presence of sequences related to the novel MHCK A catalytic domain in this diverse set of organisms suggests a possibly widespread new family of protein kinases. We have performed polymerase chain reaction (PCR) experiments using primers matching regions conserved between MHCK A and the putative C. elegans eEF-2 kinase to search for additional genes in Dictyostelium that may encode members of this novel group of proteins. We report here the molecular cloning and primary biochemical characterization of an additional member of the MHCK A-related protein kinase family.

MATERIALS AND METHODS
PCR and Cloning-Several oligonucleotide primers were designed based on sequences conserved between MHCK A and the C. elegans * This work was supported by National Institutes of Health Grant GM50009 (to T. T. E.). 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) U90946.
‡ These authors contributed similarly to this work and are listed alphabetically.
¶ T. T. E. is a recipient of an American Cancer Society Junior Faculty Research Award. 1 The abbreviations used are: MHC, myosin heavy chain; MHCK, MHC kinase; eEF-2, eukaryotic elongation factor-2; PCR, polymerase chain reaction; BSA, bovine serum albumin; bp, base pairs. putative eEF-2 kinase. The pair used for successful amplification was: (a) KinU2 (5Ј-CAGAATTCACNCCNCARGCNTTYTC, in which N ϭ (G, A, T, and C), R ϭ (G and A), and Y ϭ (T and C)); and (b) KinD2 (5Ј-CAGAATTCGTRTGDATYTGNGGGTC, in which D ϭ (G, A, and T)). The KinU2 oligonucleotide corresponds to the translated sequence TPQAFS (single-letter amino acid code) of MHCK A, and the KinD2 oligonucleotide corresponds to the translated sequence DPQIHT of MHCK A. PCR was performed using 100 ng of genomic DNA from a MHCK A -Dictyostelium cell line in which the entire gene for MHCK A was deleted (14). Amplification was performed as follows: 94°C denaturation (30 s), 37°C annealing (30 s), and 72°C extension (1 min) for 3 cycles, and then 94°C denaturation (30 s), 37°C annealing (30 s), and 72°C extension (1 min) for 30 cycles. A single product band of the expected size (ϳ120 bp) was obtained and cloned into vector pGEM7 (Promega). Sequence analysis revealed that this product was related to MHCK A.
The cloned PCR fragment was labeled with 32 P and used as a probe for library screening. A 4-h developed cDNA library (Clontech) was probed using standard conditions (15). Two overlapping clones were isolated that upon sequence analysis seemed to encode the entire MHCK B open reading frame. The compiled sequence was determined from both clones, using a combination of automated fluorescent sequencing and manual dideoxy sequencing. All portions of the gene were sequenced at least once on both strands.
Fusion Protein Expression and Phosphorylation Assays-A segment of the cDNA clone spanning predicted amino acid residues 31-387 was cloned into an E. coli expression vector that attaches a carboxyl-terminal 6X histidine tag (pET21c; Novagen). The resultant protein has a predicted mass of 44 kDa. A protein of approximately this size was found at high levels in the inclusion bodies of cells expressing this construct when isopropyl-1-thio-␤-D-galactopyranoside inductions were performed at 37°C. Isopropyl-1-thio-␤-D-galactopyranoside inductions were therefore performed at room temperature, which resulted in lower yield, but a substantial portion of the protein product remained soluble under these conditions. Lysis and Ni-affinity chromatography were performed with a kit, according to the manufacturer's instructions (Novagen). The final protein was stored at Ϫ80°C in 10 mM Hepes, pH 7.5, 50 mM KCl, and 30% glycerol. Final protein concentration was determined by performing SDS-polyacrylamide gel electrophoresis and Coomassie staining of the expressed protein in parallel with a standard concentration series of bovine serum albumin (BSA; Pierce). Densitometry was performed on the resulting gels to determine the concentration Myosin phosphorylation assays were performed in a 50-l reaction volume containing 2 g of Dictyostelium myosin, either 30 or 300 ng of the MHCK B fusion protein, and typically 1 g of BSA as a negative internal control. Omission of BSA has no effect on autophosphorylation or phosphorylation of myosin. Myosin phosphorylation reactions were performed in 10 mM Tris, pH 7.0, 20 mM KCl, 2 mM MgCl 2 , 0.5 mM ATP, and 5 Ci of [␥-32 P ]ATP at 23°C for 60 min.
Peptide phosphorylations were performed with the peptide MH-1 (RKKFGESEKTKTKEFL-amide; (16)), which corresponds to the major MHCK A target site on the myosin heavy chain at residue 2029 (underlined T in the peptide). These reactions were performed in 10 mM Hepes, pH 7.5, 2 mM MgCl 2 , 1 mM ATP, 1 mM dithiothreitol, 50 M MH-1 peptide, and 0.25 Ci/l [␥-32 P ]ATP. The recombinant MHCK B protein was allowed to autophosphorylate in the presence of nonradioactive ATP for 25 min before initiation of the peptide phosphorylation reactions. Omission of this autophosphorylation step resulted in much lower activity toward the peptide during the peptide phosphorylation assay. Assays were conducted at 23°C for 30 min, and incorporation was measured using a filter binding assay (16).

RESULTS AND DISCUSSION
In a search for new genes related to MHCK A, PCR studies were performed using primers to regions highly conserved between MHCK A and the related C. elegans yk3f11 cDNA. With this approach we isolated cDNA clones for a gene in Dictyostelium with substantial similarity to MHCK A in the central (catalytic) and carboxyl (WD repeat) domains. Based upon the deduced amino acid sequence, domain organization, and biochemical activity of the encoded protein (discussed below), we have named this gene myosin heavy chain kinase B (MHCK B). The complete cDNA for MHCK B encodes a protein of 733 amino acids with a predicted mass of 83 kDa (Fig. 1A). MHCK B has an amino-terminal segment of ϳ100 residues with no significant similarity to any data base sequences, a central segment of ϳ250 residues with ϳ54% identity to the catalytic domain of MHCK A, and a carboxyl-terminal segment of ϳ260 residues consisting of WD repeats that displays 43% identity to the corresponding WD repeat domain of MHCK A (Fig. 1B, p). The segment of MHCK B that is related to the catalytic domain of MHCK A shows a slightly lower level of identity to rat EF-2 kinase and to the putative C. elegans eEF-2 kinase (ϳ48%). MHCK B lacks the amino-terminal predicted coiled-coil domain present in MHCK A. The conserved carboxyl-terminal domain of rat eEF-2 kinase and the putative C. elegans eEF-2 kinase (Fig. 1B, checked box) displays no detectable similarity to MHCK A or MHCK B.
Multiple sequence alignments indicate conservation of MHCK B to the other members of this group throughout the region that corresponds to the catalytic domain of MHCK A, but the conservation is particularly strong to the carboxylterminal half of the MHCK A catalytic domain (Fig. 2). It is noteworthy that a possible nucleotide-binding motif described in the original MHCK A sequence (8) is conserved in all 4 sequences in this alignment (residues 298 -303 of MHCK B).
A segment of the MHCK B cDNA corresponding to amino acid residues 31-387 was expressed in E. coli with a 6X histidine tag. The resultant 44-kDa protein was purified using Ni-chelation chromatography and assayed for protein kinase activity. This protein was capable of active autophosphorylation and was found to phosphorylate Dictyostelium myosin (Fig. 3). Under these reaction conditions, it was found that the MHCK B fusion protein transferred ϳ0.6 mol of phosphate/mol of MHC. In experiments similar to that shown, we have observed MHC phosphorylation stoichiometry as high as 1.2 mol of P i /mol of MHC by the MHCK B fusion protein. Phosphorylation reactions were also performed using Dictyostelium myosin in which the 3 mapped target sites for MHCK A (at positions 1823, 1833, and 2029 of MHC) were converted to alanine residues (3X ALA MHC) (1). With this substrate, the MHCK B fusion protein reproducibly incorporated less phosphate into the MHC (0.12 mol of P i /mol of MHC in the presented experiment). The MHCK B fusion protein autophosphorylated to a level of ϳ1.2 mol of P i /mole of fusion protein under tested conditions. Although the expressed segment of MHCK B phosphorylated MHC, the activity of the protein toward MHC was significantly lower than the activity observed previously with a similar recombinant segment of the MHCK A protein (12). Approximately 300 ng of the expressed MHCK B was required to achieve transfer of 1 mol of P i /mol of MHC in reactions containing 2 g of myosin.
Catalytic activity was also assessed using the peptide substrate MH-1 (16), which corresponds the previously characterized target site for MHCK A at residue 2029 of the Dictyostelium myosin tail. Minimal activity against this peptide was observed when the purified MHCK B fusion protein was added directly to peptide phosphorylation reactions. However, if a 25-min autophosphorylation step was performed first, in the presence of nonradioactive ATP, the resulting autophosphorylated MHCK B fusion protein displayed significant kinase activity toward the peptide MH-1. Under similar assay conditions, the expressed catalytic domain of MHCK A displayed a K m of 380 M for this peptide and a corresponding turnover (k cat ) of 4.8/s (12). In a similar analysis with the expressed segment of MHCK B, we were not able to obtain saturating substrate conditions (Fig. 3C). At 650 M MH-1, reaction velocity still displayed a near linear increase with increasing substrate, indicating that the K m of the expressed segment of MHCK B for the peptide MH-1 is probably greater than 1000 M. At the highest tested MH-1 concentration, the reaction velocity or turnover corresponded to 0.29/s, approximately 6% of the maximal turnover of the expressed catalytic domain of MHCK A under V max conditions. These experiments demonstrate that the MHCK B gene encodes a protein kinase that is capable of phosphorylating Dictyostelium myosin to stoichiometric levels. These results also suggest that this enzyme has at least partial specificity to the same target sites phosphorylated by MHCK A, because the 3X ALA MHC is a poorer substrate than wild-type myosin. Given the demonstrated role of MHCK A in regulating myosin localization in vivo (1,14), these results imply that MHCK B may also participate in control of myosin localization.
Alignment of the conserved portions of MHCK A, MHCK B, rat eEF-2 kinase, and the putative C. elegans eEF-2 kinase (Fig. 2) reveals a series of conserved regions, including a possible nucleotide-binding motif (GXGXXG at residues 298 -303 in MHCK B), and a conserved pair of cysteines (residues 316 and 320 in MHCK B). The presence of a GXGXXG motif at the extreme carboxyl-terminal end of the conserved portion of each sequence further suggests that these proteins may have a three-dimensional structure unrelated to the conventional eukaryotic protein kinases, in which the GXGXXG nucleotidebinding motif is located near the amino-terminal portion of the catalytic domain (17).
Although Redpath and colleagues (13) have suggested that the catalytic domain of rat eEF-2 kinase lies in the carboxylterminal portion of that protein, our biochemical analysis of MHCK A (12) and MHCK B (reported here) strongly suggests that the amino-terminal segment of eEF-2 kinase, which is approximately 50% identical to the catalytic domains of MHCK A and MHCK B, encodes the catalytic functions of rat eEF-2 kinase.
Further biochemical mapping of the catalytic sequences of MHCK B and other members of this novel family of protein kinases will clearly be important. Additional studies are in progress to determine the substrate specificity of MHCK B and to determine whether this enzyme plays a physiological role in the regulation of myosin function in Dictyostelium and/or whether MHCK B has other in vivo substrates of importance.
Note Added in Proof-Dr. Alexy Ryazanov and colleagues have recently demonstrated that the C. elegans gene discussed in this publication encodes a bona fide eEF-2 kinase and that closely related genes are present in mice and humans (Proc. Natl. Acad. Sci. U. S. A., in press).