Striated Muscle Preferentially Expressed Genes α and β Are Two Serine/Threonine Protein Kinases Derived from the Same Gene as the Aortic Preferentially Expressed Gene-1*

Aortic preferentially expressed gene (APEG)-1 is a 1.4-kilobase pair (kb) mRNA expressed in vascular smooth muscle cells and is down-regulated by vascular injury. An APEG-1 5′-end cDNA probe identified three additional isoforms. The 9-kb striated preferentially expressed gene (SPEG)α and the 11-kb SPEGβ were found in skeletal muscle and heart. The 4-kb brain preferentially expressed gene was detected in the brain and aorta. We report here cloning of the 11-kb SPEGβ cDNA. SPEGβ encodes a 355-kDa protein that contains two serine/threonine kinase domains and is homologous to proteins of the myosin light chain kinase family. At least one kinase domain is active and capable of autophosphorylation. In the genome, all four isoforms share the middle three of the five exons of APEG-1, and they differ from each other by using different 5′- and 3′-ends and alternative splicing. We show that the expression of SPEGα and SPEGβ is developmentally regulated in the striated muscle during C2C12 myoblast to myotube differentiationin vitro and cardiomyocyte maturation in vivo. This developmental regulation suggests that both SPEGα and SPEGβ can serve as sensitive markers for striated muscle differentiation and that they may be important for adult striated muscle function.

Vascular smooth muscle cells (VSMCs) 1 play an important role in regulating vascular tone by contraction and relaxation under normal physiological conditions. Their excessive growth and extracellular matrix secretion, however, significantly con-tribute to various occlusive vascular diseases such as arteriosclerosis and restenosis after balloon angioplasty (1,2). Although the molecular mechanisms regulating these VSMC phenotypic modulations have not been fully elucidated, one theory holds that VSMCs are not terminally differentiated and that their phenotype can change from a quiescent and contractile state to a proliferative and synthetic state in vascular pathogenesis (1,3).
In contrast to the plastic phenotype of VSMCs, the striated skeletal muscle cells and cardiomyocytes undergo terminal differentiation during perinatal and neonatal stages (4 -6). In skeletal muscle, the terminal differentiation of myoblasts into myotubes is marked by the permanent withdrawal from the cell cycle and the induction of muscle-specific structural genes by myogenic transcriptional factors (e.g. MyoD, myogenin, Myf5, and MRF4) (4,7). In the heart, the maturation of cardiomyocytes from hyperplasia to hypertrophy takes place soon after birth, and they also permanently withdraw from the cell cycle (6,8,9). Multiple myofibrillar protein genes are regulated during maturation and terminal differentiation of striated muscles. Much of this regulation involves switches of various myofibrillar proteins from the fetal/neonatal isoforms to the adult isoforms. Isoforms of these myofibrillar proteins may derive from multigene families or from alternative splicing and/or use of different promoters in the same gene (for review, see Ref. 10). For instance, the cardiac troponin T (cTnT) is expressed in the human heart as four isoforms (cTnT 1 through cTnT 4 ) generated by alternative splicing of two 5Ј exons. The cTnT 3 is the dominant adult isoform in the heart, whereas the other three isoforms are expressed in the fetal heart (11). It is interesting that the fetal cTnT 4 isoform is also re-expressed in failing adult hearts (11).
Aortic preferentially expressed gene (APEG)-1 was originally cloned by differential mRNA display as a 1.4-kb mRNA preferentially expressed in adult VSMCs (12). It is a single copy gene, and analysis of its promoter demonstrates that a 2.7-kb mouse APEG-1 5Ј-flanking sequence confers VSMC-specific reporter gene expression (13). The expression of APEG-1 mRNA rapidly disappears in primary culture of VSMCs at early passages (12). In the rat carotid artery balloon injury model APEG-1 mRNA is significantly decreased 2 and 4 days after injury but recovers slightly after 8 days (12). On the basis of these observations we propose that APEG-1 serves as a sensitive marker specific for VSMC differentiation.
We report here the identification of three additional APEG-1 isoforms expressed in the striated muscles and in the brain. These isoforms, termed striated muscle (S)PEG␣, SPEG␤, and brain (B)PEG, were identified by using the 5Ј-end portion of APEG-1 cDNA in tissue Northern blot analysis and were de-* This work was supported by a grant from the Bristol-Myers Squibb Co. and by National Institutes of Health Grants HL10113 (to M. D. L.), HL60788 (to M. A. P.), and GM53249 (to M.-E. L. and M. A. P.). 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) AF215896.
We rived from the same gene locus as APEG-1. The largest 11-kb SPEG␤ isoform was cloned and characterized. The SPEG proteins contained a functional serine/threonine kinase domain. In the heart, the SPEG␣ and SPEG␤ were expressed only after birth, as the APEG-1 was down-regulated. This isoform switch from APEG-1 to SPEG␣ and SPEG␤ correlated with the timing of neonatal heart maturation. The developmental regulation of both SPEG␣ and SPEG␤ suggested that they can serve as sensitive differentiation markers for the striated muscles and that they may be important for adult striated muscle function.

EXPERIMENTAL PROCEDURES
Cell Culture and Reagents-C2C12 myoblasts were obtained from the American Type Cell Culture Collection, and 293T cells were obtained from Dr. Hamid Band (Dana Farber Cancer Institute, Boston) (14). C2C12 cells were grown in Dulbecco's modified Eagle's medium (Life Technologies, Inc.) supplemented with 20% fetal calf serum (Hy-Clone), 4 mM L-glutamine, penicillin (100 units/ml), and streptomycin (100 g/ml) in a humidified incubator (37°C, 5% CO 2 ). 293T cells were grown similarly except that they received 10% fetal calf serum and were maintained in a 10% CO 2 incubator. C2C12 cells were differentiated into myotubes in Dulbecco's modified Eagle's medium supplemented with 2% horse serum (HyClone).
SPEG cDNA Cloning and Sequencing-A mouse APEG-1 cDNA 5Јend probe that contained the open reading frame was used to screen an adult mouse heart 5Ј-stretch plus cDNA gt11 library (CLONTECH). Positive clones were purified, and cDNA inserts were excised by EcoRI digestion and ligated into plasmid vectors. Subsequent screening of the same cDNA library was carried out with the newly obtained cDNA fragments that extended at either the 5Ј-or the 3Ј-end. Both 5Ј-and 3Ј-RACE was performed according to the protocols from the manufacturer (Life Technologies, Inc.). The primers used in 5Ј-RACE experiment are as follows: 5Ј-GCTGAAGGCCCTGTCATCCC-3Ј, 5Ј-CT-GGGTCCCCGGCACATCGC-3Ј, and 5Ј-TGATGTCATCCTCGGCAGT-C-3Ј. The primers used in 3Ј-RACE were 5Ј-GGGTCGCTTTGATGCC-TTCCAG-3Ј and 5Ј-CGCCGCCAGACACTCACCTTCA-3Ј.
Genomic Library Screening and Sequence Analysis-A 129/SvJ mouse genomic library in FIX II vector (Stratagene) was screened by a mouse SPEG␤ cDNA 5Ј probe. The genomic DNA was sequenced by the dideoxy chain termination method with primers designed from the mouse SPEG␤ cDNA sequence to determine the exon-intron junctions. The intron sizes were determined by polymerase chain reactions (PCR) with primers in the flanking exons and by direct sequencing.
RNA Extraction and Northern Blot Analysis-Total RNA from cultured cells was prepared by the RNeasy Mini Kits (Qiagen). Tissue RNA was extracted from adult male Harlan Sprague-Dawley rats or from adult male Balb/c mice (Taconic Farms) as described (12). Rat hearts at different developmental stages were obtained from fetuses (17-19 days post-coitum) and from rats 2, 14, and 28 days after birth. Ten g of RNA was fractionated on 1.3% formaldehyde-agarose gels and transferred to nitrocellulose filters. The filters were then hybridized with [ 32 P]dCTPlabeled, random-primed APEG-1 or SPEG cDNA probes as described in the figure legends. The RNA blot hybridization was performed as described previously (12). To correct for differences in RNA loading, the same blots were also hybridized with an 18 S rRNA oligonucleotide probe.
Construction of an SPEG Minikinase-A portion of SPEG cDNA containing the internal kinase domain (amino acid residues 1598 -1862) was amplified by reverse transcription-PCR using the following two primers, SPEGF4935 5Ј-ggatccCGAGGAAGAAGACTTAGCGAC-3Ј (the ggatcc was an added BamHI site not present in the cDNA sequence) and SPEGR5729 5Ј-TTCTGTTTTGAACCAAGGATG-3Ј. This cDNA fragment was fused by ligation to an amino-terminal FLAG sequence in the plasmid vector pCMV-FLAG-2 (Sigma) to make the SPEG minikinase construct.
In Vitro Kinase Assay-293T cells were transfected with plasmid DNA of the FLAG-tagged SPEG minikinase, the FLAG-tagged Akt, or the empty pFLAG-CMV-2 vector (Sigma) by LipofectAMINE (Life Technologies, Inc.) according to the manufacturers' protocol. Twenty four h after transfection cells were harvested in the lysis buffer (10 mM Tris-HCl, pH 7.6, 1 mM EDTA, pH 8.0, 150 mM NaCl, 1% Triton X-100). The soluble cell lysates were used in kinase assays as described previously (15). In brief, protein concentration was determined by the Bradford dye-binding procedure (Bio-Rad). One mg of protein was immunoprecipitated by 10 g of M2 monoclonal antibody to FLAG conjugated to agarose beads (Sigma) at 4°C overnight. The beads were washed three times in the lysis buffer and once in the kinase reaction buffer (10 mM MgCl 2 , 3 mM MnCl 2 , 10 mM Tris-HCl, pH 7.2). The in vitro kinase reaction was done in the presence of 10 Ci of [␥-32 P]ATP (PerkinElmer Life Sciences) at 30°C for 10 min and then terminated by the addition of Laemmli sample buffer and boiled for 3 min. The samples were separated on a 10% Tricine/SDS-polyacrylamide gel and visualized by autoradiography.
Immunohistochemical Analysis-Adult mouse hearts were fixed in methyl Carnoy's fixative (60% methanol, 30% chloroform, 10% acetic acid) for 3 h and embedded in paraffin. Heart sections were processed for immunostaining essentially as described (16). In brief, sections were incubated with an affinity-purified antibody to APEG-1/SPEG that was generated in rabbit against a FLAG peptide-tagged APEG-1 fusion protein. For negative control staining, the same antibody was preabsorbed with the FLAG-APEG-1 fusion protein (1:1 by weight) at 4°C overnight before incubation. After washing, a biotinylated goat antibody to rabbit IgG (Vector Laboratories) and an avidin-horseradish peroxidase were sequentially applied to the sections. Staining results were developed by a peroxidase DAB kit (Vector Laboratories).
Immunofluorescent Staining-Differentiated C2C12 cells (for 5 days) grown on glass coverslips and mouse heart sections (see above) were used for staining as described (12). The C2C12 cells were fixed by 4% paraformaldehyde. The heart sections and C2C12 cells were incubated with both the affinity-purified rabbit antibody to APEG-1/SPEG and a mouse monoclonal antibody to desmin (Sigma) at 4°C overnight. After washing, an Alexa 488-conjugated goat antibody to mouse IgG and an Alexa 546-conjugated goat antibody to rabbit IgG were applied to the sections and coverslips at room temperature for 2 h before washing. Localization of desmin and SPEG protein signals was observed by fluorescent microscopy.

RESULTS
Identification of BPEG, SPEG␣, and SPEG␤-By using a full-length APEG-1 cDNA probe, we identified four different isoforms in adult rat and mouse tissues by Northern analysis (Fig. 1, A and B, respectively). In addition to the 1.4-kb APEG-1 mRNA, two messages of 9 and 11 kb were found in adult tissues containing striated muscles such as skeletal muscle and heart ( Fig. 1). A 4-kb message was also seen in the brain and aorta (Fig. 1). The 9-and 11-kb messages were not identified previously by an APEG-1 3Ј cDNA probe obtained by differential mRNA display (data not shown and also Fig. 5A, IV). We termed the 9-and 11-kb messages SPEG␣ and SPEG␤, respectively, for their preferential expression in the striated muscle, and the 4-kb message BPEG for its expression in the brain. Since we have shown that APEG-1 is a single copy gene in the genome (13), we speculated that these isoforms were products of the same gene locus as APEG-1.
Cloning and Analysis of SPEG␤ cDNA-To obtain the SPEG cDNA sequence, we screened an adult mouse heart cDNA library using an APEG-1 cDNA 5Ј-end probe. An 11-kb SPEG␤ cDNA contig was assembled from several cDNA library screenings and both 5Ј-and 3Ј-RACE ( Fig. 2A). We repeated the 5Ј-RACE twice using different sets of primers and confirmed the same 5Ј-end of the SPEG␤ cDNA (data not shown). An open reading frame was identified between base pairs 144 and 9929 with no other upstream ATG initiation codon. Multiple stop codons were also found in all three reading frames in the 5Ј-untranslated region (data not shown). The deduced 3262residue SPEG␤ peptide sequence (Fig. 2B) had a predicted molecular mass of 355 kDa. The APEG-1 peptide sequence was found at residues 855-965 of the SPEG␤ protein with the exception of its last two amino acids (Fig. 2B). A PROSITE data base search revealed that the SPEG␤ protein contained two putative serine/threonine kinase catalytic sites and one ATPbinding site (Fig. 2B) (17). A BLAST search in the nonredundant protein data base identified a partial human KIAA1297 protein (18) with a 75% identity to the SPEG␤ protein, suggesting that the KIAA1297 is a human counterpart of the SPEG␤. Furthermore, significant homology was identified between the SPEG␤ and proteins of the myosin light chain kinase (MLCK) family including MLCK, titin, twitchin, projectin, and death-associated protein kinases. By using ProDom and PFam data base analyses (19,20), we found that the homology between SPEG␤ and these proteins was not limited to the kinase domains but extended to multiple immunoglobulin and fibronectin structural domains ( Fig. 2B and 3A). These domains have been implicated in homophilic or heterophilic interaction of other myosin-binding proteins (21). We hypothesized that the SPEG␤ and potentially the other three isoforms (APEG-1, BPEG, and SPEG␣) were part of the functionally and structurally diverse MLCK protein family (22).
Activity of the SPEG Kinase Domain-We compared the kinase domains of SPEG␤ and several other homologous proteins (Fig. 3A). This region is defined to contain 11 subdomains and is conserved among the serine/threonine kinases (23,24). The internal putative kinase domain of SPEG␤ (SPEG-i) showed a 37% identity (98 of 268 residues) to the consensus sequence of the other kinases (Fig. 3A). In addition, all but one of the 15 conserved residues within the kinase domain are identified in the internal kinase domain of SPEG␤ (Fig. 3A, see asterisks) (23,24). Due to the divergence in the ATP-binding region (subdomains I and II) and subdomain VII in the carboxylterminal putative kinase domain of SPEG␤ (SPEG-c), there was only a 24% identity (64 of 268 residues) to the other kinases (Fig. 3A). We therefore chose to test the biochemical activity of the internal putative kinase domain of SPEG␤. A SPEG minikinase plasmid that expressed a FLAG-tagged SPEG internal kinase domain was constructed. The sequence of the SPEG minikinase included 29 serine and threonine residues, and in an in vitro kinase assay this minikinase was capable of autophosphorylation (Fig. 3B). As a positive control we used a non-related FLAG-tagged Akt, which is also a serine/ threonine protein kinase, in the same experiment. This FLAG-Akt fusion protein showed autophosphorylation (Fig. 3B) as previously reported (25).
Genomic Analysis of the SPEGs-We previously reported that APEG-1 is a single copy gene (13). We also found that sequences corresponding to the first and the last exons of APEG-1 were absent in the 11-kb SPEG␤ cDNA sequence, suggesting that SPEG␤ was produced by another upstream promoter and by alternative splicing in a tissue-specific manner. To study how all four isoforms were transcribed from a single copy gene, we obtained additional genomic DNA flanking the APEG-1. From the genomic DNA clones we identified multiple SPEG␤ exons both upstream and downstream of APEG-1 (Fig. 4). The identified exon-intron junctions (Fig. 4) were all in agreement with the consensus 5Ј-GT and 3Ј-AG sequences (26). In addition, the first exon of APEG-1 was located in an ϳ9-kb SPEG␤ intron, and the last APEG-1 exon was within another downstream intron (Fig. 4). The APEG-1 promoter (13) was located within the ϳ8-kb intronic sequence between the first exon of APEG-1 and the immediate upstream SPEG␤ exon.
Northern Analysis of the APEG-1 and Its Isoforms by Exonspecific cDNA Probes-To elucidate further the use of different exons by the four different isoforms, we performed a series of Northern blot analyses with various APEG-1 and SPEG cDNA probes (Fig. 5A). The last exon of APEG-1 hybridized only to the APEG-1 and BPEG messages (Fig. 5A, IV). This exon contained an in-frame stop codon at its 5Ј-end and a polyadenylation signal, and was likely the last exon of the BPEG as it is of the APEG-1 (13). All four isoforms (APEG-1, BPEG, SPEG␣, and SPEG␤) were recognized by a cDNA probe derived from the middle three exons of APEG-1 (Fig. 5A, III). This result was consistent with our initial finding (Fig. 1) and implied that these three exons were used by all isoforms, since excluding any of the three exons would have caused a shift in the reading frame. The probe derived from the first APEG-1 exon hybridized to the 1.4-kb APEG-1 and the 9-kb SPEG␣ (Fig. 5A, II), suggesting that these two isoforms initiated their transcription from the intronic APEG-1 promoter. The immediate upstream SPEG␤ exon hybridized only to the 4-kb BPEG and the 11-kb SPEG␤ (Fig. 5A, I), indicating that the transcription for these two isoforms initiated from at least one other promoter upstream of the APEG-1. We had also tested a SPEG␤ cDNA probe further upstream of the APEG-1 that gave results similar to those in panel I of Fig. 5A (data not shown). Two cDNA probes corresponding to the two putative SPEG kinase domains downstream of APEG-1 hybridized to both SPEG␣ and SPEG␤ (Fig. 5A, V and VI), demonstrating that these two isoforms contained both kinase domains and that neither the APEG-1 nor the BPEG did. Taken together, we hypothesized that the selective use of two transcription start sites and alternative splicing in tissue-specific manners gave rise to the four different isoforms (Fig. 5B). Besides their tissue FIG. 1. Identification of APEG-1 isoforms. Ten micrograms of total RNA obtained from adult rat (A) or mouse tissues (B) was used for Northern analysis with a full-length APEG-1 cDNA probe of the same species. In addition to the 1.4-kb APEG-1 message, the cDNA probe also recognized two additional messages of 9 and 11 kb mainly in tissues containing striated muscle cells (SPEG␣ and SPEG␤, respectively) and a message of ϳ4 kb in the brain and aorta (BPEG). The positions of 28 S and 18 S rRNA are indicated. The same blots were hybridized with an end-labeled 18 S rRNA oligonucleotide probe to show RNA loading in each lane. distribution, the distinction between SPEG␣ and SPEG␤ and between APEG-1 and BPEG was their 5Ј-end promoter selection. Moreover, alternative splicing differentiated SPEG␤ from BPEG and SPEG␣ from APEG-1. However, we could not exclude the possibility that SPEG␤ and BPEG used different promoters due to their distinct tissue distribution.
Induction of SPEG Expression during C2C12 Myoblast Differentiation-We previously showed that APEG-1 is a sensitive marker for VSMC differentiation (12). We speculated that SPEG expression is also regulated by striated muscle differentiation, and we examined the expression of SPEG␣ and SPEG␤ during C2C12 cell differentiation from myoblasts to myotubes (Fig. 6). In the rapidly growing C2C12 myoblasts, both SPEG␣ and SPEG␤ were minimally expressed (Fig. 6A). Increased expression of the 9-kb SPEG␣ could be detected when the C2C12 cells became confluent (Fig. 6A). A significant induction of SPEG␣ mRNA was observed when these cells were induced to differentiate and formed myotubes. The 11-kb SPEG␤ message, however, remained at a minimal level (Fig. 6A). We next confirmed the induction of SPEG␣ protein by an affinity-purified antibody to APEG-1/SPEG in Western blot analysis. This polyclonal antibody was generated in rabbit against a FLAGtagged APEG-1 fusion protein. It cross-reacted with the SPEG␣, SPEG␤, and BPEG because of their shared peptide sequences. 2 We found that an ϳ250-kDa protein was induced by the C2C12 differentiation (Fig. 6B). The size of this identified protein was in agreement with what we had predicted for the SPEG␣ protein (residues 855-3262 of SPEG␤, Fig. 2B). To assess further the association between SPEG␣ induction and C2C12 differentiation, we examined the expression of skeletal myosin heavy chain (Fig. 6B), a differentiation marker for skeletal muscle (27). We found that myosin heavy chain was expressed as early as 1 day and continued to increase through 5 days of differentiation, showing a concordant expression pattern between SPEG␣ and myosin heavy chain. The induction of SPEG␣ protein in differentiated C2C12 myotubes was further demonstrated by immunofluorescent staining of these cells by

FIG. 3. Sequence alignment between SPEG␤ and MLCK family proteins and in vitro kinase assay. A, the internal (SPEG-i) and
carboxyl-terminal (SPEG-c) ATP-binding and serine/threonine kinase domains of the SPEG proteins were aligned with proteins of the MLCK family, including human death-associated protein kinase I (DAPK1, GenBank TM accession number AAC35001), human titin (GenBank TM accession number I38344), human smooth muscle MLCK (GenBank TM accession number Q15746), Caenorhabditis elegans twitchin (GenBank TM accession number S07571), and Drosophila melanogaster projectin (GenBank TM accession number AAC27550). Solid boxes and shaded boxes highlight amino acid residues that were identical or similar in at least four proteins, respectively. The ATP-binding site (solid line) and serine/threonine kinase catalytic site (dashed line) are shown. The 11 subdomains conserved within serine/threonine kinases are indicated above the sequences, and the 15 amino acids conserved in most protein kinase domains (23,24) are indicated by asterisks. B, in vitro kinase assay of the internal SPEG kinase domain. A FLAG-tagged SPEG minikinase (SPEGmk) plasmid was transfected into 293T cells as described under "Experimental Procedures." An empty vector (Control) and a FLAG-tagged Akt (Akt) plasmid served as negative and positive controls, respectively. One mg of total cell lysate was immunoprecipitated by an agarose bead-conjugated anti-FLAG M2 monoclonal antibody, and the immunoprecipitated protein was used in the in vitro kinase assay (IP ϩ kinase assay). Fifty g of the cell lysate was analyzed by Western blot with the M2 monoclonal antibody to FLAG peptide to demonstrate expression of the fusion proteins (Western blot). The two FLAG-tagged fusion proteins (arrows) and the molecular mass (in kilodaltons) are shown. An asterisk indicates a smaller Akt fusion protein that may derive from protein degradation or from a translation initiation at an internal ATG codon. the antibodies to APEG-1/SPEG and desmin (Fig. 6C). We observed that SPEG␣, as recognized by the antibody to APEG-1/SPEG, was found only in differentiated C2C12 myotubes but not in the surrounding individual cells. The expression pattern of the SPEG␣ protein in the differentiated and fused myotubes was identical to that of the desmin protein, a sarcomeric protein expressed in differentiated striated muscles. Thus, SPEG␣ may serve as a sensitive marker for skeletal muscle differentiation at both the mRNA and the protein levels.
SPEG␣/␤ and APEG-1 Expression during Maturation and Terminal Differentiation of the Heart-To test whether the induction of SPEG␣ and SPEG␤ also occurred during cardiomyocyte maturation, we isolated RNA from rat heart at different developmental stages and performed Northern analysis (Fig.  7A). By using a probe that recognized all four isoforms, we detected strong APEG-1 and minimal BPEG expression in fetal hearts (17-19-day rat fetuses). Neither SPEG␣ nor SPEG␤ was detected at this stage, which suggested that APEG-1 was the predominant embryonic isoform expressed in the heart, despite its preferential expression in adult VSMCs (Fig. 1). After birth, the APEG-1 mRNA decreased significantly in neonatal hearts, and the two SPEG messages began to increase (Fig. 7A). In the 2-day-old neonatal hearts, APEG-1 expression was reduced by 25% (Fig. 7B). Two and four weeks after birth, APEG-1 expression decreased further to less than 20% of its prenatal level (Fig. 7). On the contrary, the two SPEG messages continued to increase, and together they reached close to 70% of the prenatal APEG-1 expression level at 4 weeks (Fig. 7B). The time course of the isoform switch from APEG-1 to SPEG␣ and SPEG␤ showed a strong correlation with neonatal cardiomyocyte maturation (28). This finding was consistent with the aforementioned induction of SPEG␣ during skeletal muscle differentia- Western blot analysis of SPEG␣ protein expression during C2C12 myoblast to myotube differentiation. Fifty g of cell lysate was separated in a Tris glycine-SDS 4 -20% polyacrylamide gradient gel and transferred by electroblotting onto a nitrocellulose membrane. This blot was then probed sequentially with an affinity-purified antibody to APEG-1/ SPEG, a monoclonal antibody to myosin heavy chain, and a monoclonal antibody to ␣-tubulin. C, immunofluorescent staining of desmin (green) and SPEG␣ (red) proteins in C2C12 cells differentiated for 5 days. Nuclear DNA was counter-stained by Hoechst 33258 (blue) to identify all cell nuclei. Phase-contrast microscopy was used to demonstrate cell morphology. Original magnification, ϫ 400. tion in vitro and implied a functional importance of SPEG␣ and SPEG␤ in the maturation and terminal differentiation of cardiomyocytes.
SPEG Proteins Colocalized with Desmin to the Sarcomeric Z Bands in Cardiomyocytes-We immunohistochemically stained the mouse heart to localize the distribution of SPEG proteins. The antibody to APEG-1/SPEG used in this study recognized both SPEG␣ and SPEG␤ proteins. We observed a prominent striated staining pattern and some longitudinal staining in the atrium and ventricle (Fig. 8A). The lack of any significant staining by an antigen-preabsorbed control antibody demonstrated that the observed staining patterns were specific (Fig. 8A). Since the striated distribution of SPEG proteins resembled that of other sarcomeric proteins, we next used immunofluorescent staining by antibodies to both desmin and APEG-1/SPEG to test whether the SPEG proteins colocalized with desmin in the sarcomeric Z bands (Fig. 8B). Staining with antibody to desmin produced a typical striated pattern specific to the sarcomeric Z bands (Fig. 8B). The striated pattern of the desmin protein overlapped with that of the SPEG proteins (Fig. 8B), demonstrating that the SPEG proteins were localized to the sarcomeric Z bands. DISCUSSION We identified three additional isoforms of APEG-1 termed SPEG␣, SPEG␤, and BPEG. SPEG␣ and SPEG␤ were demonstrated to be serine/threonine kinases (Fig. 3B) derived from the same gene locus as APEG-1 by both differential promoter usage and alternative splicing (Fig. 5). Our results suggest that SPEG␣ and SPEG␤ expression, like APEG-1 expression in VSMCs, serve as sensitive markers for striated muscle cell differentiation (Figs. 6 and 7). SPEG␣ was up-regulated during C2C12 myotube formation at both mRNA and protein levels. This implicated SPEG␣ as a differentiation marker for skeletal muscle cells. Interestingly, the induced SPEG␣ expression and the low level of SPEG␤ in differentiated C2C12 myotubes were contrary to their respective expression levels in adult skeletal muscle in vivo (Fig. 1). This may reflect that C2C12 myotubes, despite their extensive fusion and occasional spontaneous contraction, were not as terminally differentiated as skeletal muscle in vivo. Moreover, both SPEG␣ and SPEG␤ were expressed at similar levels in adult heart. This would argue that SPEG␤ is also a marker for striated muscle differentiation and may represent a differentiation state of striated muscle later than that represented by SPEG␣. Indeed, in the embryonic stem cell to embryoid body differentiation model, SPEG␣ expression was detected 2 days prior to SPEG␤ expression (7 and 9 days after differentiation, respectively) (data not shown). FIG. 7. Isoform switch from APEG-1 to SPEG␣ and SPEG␤ during neonatal heart maturation. A, Northern analysis of rat heart RNA collected from 17 to 19-day-old fetuses (17-19 d fetus) and from rats 2, 14, and 28 days after birth. Ten g of total RNA was separated by electrophoresis, transferred onto nitrocellulose membrane, and hybridized with a rat APEG-1 cDNA probe that recognized all isoforms. An 18 S rRNA hybridization was used to show loading of RNA. B, quantitative analysis of APEG-1 and SPEG␣ and ␤ expression during heart maturation. The expression levels of APEG-1 and SPEG␣ and ␤ were first normalized by their corresponding 18 S rRNA signals. APEG-1 expression in the embryo heart was then set as 100% and used to compare with the expression of APEG-1 (solid columns) and the SPEG␣ and ␤ (white columns) at different stages of heart maturation.
FIG. 8. Immunohistochemical analysis of the SPEG protein expression in the mouse heart. A, immunohistochemical staining was performed as described under "Experimental Procedures." Sections of mouse heart atrium and ventricle were incubated with an affinitypurified antibody to APEG-1/SPEG (SPEG␣/␤). To demonstrate specific staining, adjacent sections were incubated with the same antibody that had been preabsorbed by the antigen as described under "Experimental Procedures" (Control). Original magnification, ϫ 1000. B, immunofluorescent staining shows the localization of the desmin (green, desmin) and the SPEG proteins (red, SPEG␣/␤) in the same mouse cardiomyocytes. The colocalization of the desmin and SPEG proteins at the Z-bands is shown by superimposing these two images (yellow, Combined). Original magnification, ϫ 1000.
Since APEG-1 expression is down-regulated by vascular injury (12), we had speculated that SPEG expression in the heart might also be suppressed by cardiomyopathy. However, we detected little change in SPEG expression in a cardiac hypertrophy and dilated cardiomyopathy animal model (data not shown) (29). It is interesting that the APEG-1 mRNA level was increased by 2-fold in the hypertrophic heart (data not shown). Although this increase may have been a result of increased vascularization in cardiomyopathy, it was also possible that the cardiomyocytes re-expressed APEG-1 in response to hypertrophic cardiomyopathy. The latter is consistent with our observation that APEG-1 was the dominant embryonic isoform in muscles (Fig. 7) 3 and that some fetal isoforms of heart proteins such as cTnT 4 , ␣-skeletal actin, and sarcomeric tropomyosin are re-expressed in the failing heart (11,30,31).
The SPEG proteins are homologous to the MLCK family, which includes titin, twichin, and MLCK ( Fig. 3A (22)). We found that the protein sequences of APEG-1 and SPEG␣/␤ shared significant homology with those of telokin and MLCK, respectively. Both APEG-1 and telokin are proteins that contain an immunoglobulin-like domain ( Fig. 2B and (32)) and are a small portion of a larger muscle kinase (SPEG␣/␤ and MLCK, respectively). telokin is also transcribed from a promoter located within a 3Ј-end intron of MLCK (33), which is similar to the genomic organization of APEG-1 and SPEG (Fig. 4). Telokin can accelerate myosin light chain dephosphorylation and relaxation (34), reduces MLCK binding to myosin filaments, and modulates the oligomeric state of MLCK (35). These direct functional modulations of MLCK by telokin, however, may not apply to APEG-1 and SPEG proteins due to their differences in developmental expression and adult tissue distribution (Figs. 1 and 7). Since APEG-1 was expressed in embryonic hearts and later switched to SPEG␣ and SPEG␤ during neonatal heart maturation (Fig. 7), it is still possible that the APEG-1 and SPEG proteins interact with similar proteins.
Although many proteins of the MLCK family contain a calcium/calmodulin-binding domain that regulates their kinase activity, we did not find any amphiphilic ␣-helical structure (36) within the SPEG protein sequence that might bind calcium/calmodulin. The presence of immunoglobulin, fibronectin type III, and kinase domains in SPEG proteins is reminiscent of the much larger protein titin. The immunoglobulin and fibronectin type III domains have been implicated in homophilic and heterophilic interaction between proteins (21). Since SPEG proteins and the amino terminus of titin are colocalized to the sarcomeric Z bands (Fig. 8 (37)), it is interesting to speculate whether an interaction between SPEG proteins and titin exists in striated muscle cells. Future studies will also focus on understanding the biological significance of the various isoforms of APEG-1 and whether the isoform switch from APEG-1 to SPEG␣ and SPEG␤ is important for the terminal differentiation of the striated muscle cells.