Identification, Characterization, and Functional Analysis of Heart-specific Myosin Light Chain Phosphatase Small Subunit*

Myosin light chain phosphatase consists of three subunits, a 38-kDa catalytic subunit, a large 110–130-kDa myosin binding subunit, and a small subunit of 20–21 kDa. The catalytic subunit and the large subunit have been well characterized. The small subunit has been cloned and studied from smooth muscle, but little is known about its function and specificity in the other muscles such as cardiac muscle. In this study, cDNAs for heart-specific small subunit isoforms, hHS-M 21 , were isolated and characterized. Evidence was obtained from an analysis of genome to suggest that the small subunit was the product of the same gene as the large subunit. Using permeabilized renal artery preparation and permeabilized cardiac myocytes, it was shown that the small subunit increased sensitivity to Ca 2 1 in muscle contraction. It was also shown using an overlay assay that hHS-M 21 bound the large subunit. Mapping experi- ments demonstrated that the binding domain and the domain involved in the increasing Ca 2 1 sensitivity mapped to the same N-terminal region of hHS-M 21 . These observations suggest that the heart-specific small

been characterized (14,22,23). There are, however, only a few reports about the MLCP in the cardiac muscle (24,25). In addition, it has been considered that the Ca 2ϩ sensitivity of contraction in the striated muscle is mainly regulated by troponins (26), and the significance of MLC phosphorylation in the Ca 2ϩ sensitization of the striated muscle remains to be elucidated. However, a recent report has demonstrated a possible involvement of MLC phosphorylation in the development of cardiac hypertrophy (27).
Recently, another gene for MBS, MYPT2, was isolated by Fujioka et al. (25). Toward the N terminus and C terminus, MBS encoded by MYPT2 has seven ankyrin repeats and three leucine zipper motifs, respectively, which are highly homologous to the relevant sequences of MBS encoded by MYPT1. Because MYPT2 is expressed preferentially in the striated muscles, especially in the cardiac muscle, it was suggested that the function of MYPT2 might be related with the regulation of MLC phosphorylation in the heart (25). The functions of MLC and MLCP in the cardiac muscle, then, should be unraveled for better understanding of the regulation of muscle contractility in the heart.
We report here the isolation of cDNAs for heart-specific isoform of M20 subunit (hHS-M 21 ), which were obtained in the process of isolating novel genes being preferentially expressed in the cardiac muscle. From an analogy of function for sm-M20 in the smooth muscle, hHS-M 21 was suggested to participate in the regulation of MLC phosphorylation in the cardiac muscle. We, then, determined genomic structure of the gene for hHS-M 21 and analyzed the expression profiles of hHS-M 21 . In addition, the role of hHS-M 21 in the Ca 2ϩ sensitivity of muscle contraction was investigated, along with its interaction with MBSs encoded by MYPT1 and MYPT2.

EXPERIMENTAL PROCEDURES
Isolation of a Heart-specific cDNA-To obtain information on genes preferentially expressed in the cardiac muscle, a normalized subtraction PCR-cDNA library between mRNAs of cardiac and skeletal muscles was constructed using a PCR-Select cDNA subtraction kit (CLON-TECH). Randomly selected clones from this library were determined for their sequences, and these sequences were compared with the primate nucleotide sequences in the GenBank data base. Several cDNA fragments that were found more than two times in the sequenced cDNA fragments and had no identity in the data base except for human expressed sequence tags were investigated for tissue specificity of gene expression by an RT-PCR analysis. In the RT-PCR analysis, mRNAs from various human tissues including fetal heart, adult heart, skeletal muscle, brain, kidney, liver, uterus, lung, spleen, thymus, small intestine, and colon (CLONTECH) were examined. A cDNA fragment (HS602) that was expressed only in the heart was investigated further. To isolate full-length cDNA clones encoding the HS602 gene, a human heart cDNA library in gt11 (CLONTECH) was screened using HS602 fragment as a probe according to the standard methods (28). The cDNA inserts from positive recombinant phages were subcloned into pBluescript KS (Ϫ) (Invitrogen) for sequence determination.
Expression and Purification of Recombinant Proteins-Various fragments of the hHS-M 21 cDNA were expressed as hexahistidine-tagged fusion proteins in a prokaryote expression system using pQE30 (Qiagen). Recombinant MBS proteins encoded by MYPT1 and MYPT2 were obtained as flag-tagged fusion proteins using a pQE30-flag vector in which the hexahistidine tag was replaced by a flag tag. The pQE30-flag vector was constructed as follows. At first, PCR amplification with primers, 5Ј-CCTTTCGTCTTCACCTCGAG-3Ј (sense) and 5Ј-GGATCC-CTTATCGTCATCGTCCTTGTAATCTCTCATAGTTAATTTCTCCT-C-3Ј (antisense, containing a flag sequence), was performed using pQE30 as a template. The PCR product was then cloned into pCR2.1 (Invitrogen) to be confirmed for the sequences. The insert was excised by digestion with EcoRI and BamHI and cloned back into the EcoRI-BamHI-cleaved pQE30 to obtain pQE30-flag.
The recombinant hHS-M 21 proteins generated were hHS-M 21 A and hHS-M 21 B (two constructs for major isoforms; corresponding to exons 16 -23 and 25 or exons 16 -24 of MYPT2, respectively), hHS-M 21 t-1 (deletion of the leucine zipper motifs; corresponding to exons 16 -23) MBSs encoded by MYPT1 and MYPT2 were obtained by the RT-PCR from human heart cDNA. The amplicons were once cloned into pCR2.1 for sequence confirmation and the insert cDNAs were excised by digestion with BamHI and SalI (for all constructs of hHS-M 21 and MYPT2, and MYPT1 A and MYPT1 M) or with SphI and SalI (for MYPT1 P). These restriction sites were included in the design of sense and antisense primers of which sequences are available upon request. The excised DNA fragments were cloned into pQE30 (hHS-M 21 ) or pQE30flag (MYPT1 and MYPT2). The recombinant constructs were used to transform Escherichia coli M15[pREP4], and the expression of recombinant proteins in the pQE system (Qiagen) was performed according to the manufacturer's instructions. After the induced expression of recombinant proteins, the E. coli cell pellets were washed with ice-cold phosphate-buffered saline (PBS) (116 mM NaCl, 10 mM Na 2 HPO4, and 3 mM K 2 HPO 4 (pH 7.4)) and stored at Ϫ80°C until preparation of recombinant proteins.
The preparation of recombinant proteins from the cell pellets was done as follows. After thawing at 4°C, cells were sonicated in 6 M urea, 5 mM imidazole, 500 mM NaCl, 5 mM phenylmethanesulfonyl fluoride, 5 mM N-ethylmaleimide, and 20 mM Tris-HCl (pH 7.5) (buffer A) and centrifuged at 10,000 ϫ g for 15 min at 4°C. Supernatants containing recombinant MBS proteins were stored at Ϫ80°C until use in binding assay. On the other hand, the supernatants containing recombinant hHS-M 21 proteins were loaded into columns in which the resin (Novagen) charged with Ni 2ϩ and equilibrated with buffer A was bedded. After the columns were washed extensively with 6 M urea, 60 mM imidazole, 500 mM NaCl, and 20 mM Tris-HCl (pH 7.5), bound proteins were eluted by 6 M urea, 400 mM imidazole, 500 mM NaCl, and 20 mM Tris-HCl (pH 7.5). The eluates were dialyzed against 0.1% acetic acid (pH 4.0), and the dialysates were clarified by centrifugation at 10,000 ϫ g for 15 min at 4°C. The supernatants were freeze-dried and dissolved in 0.1% acetic acid (pH 4.0) for concentration. The recombinant hHS-M 21 proteins were further purified by using an HPLC system consisting of a pump (LC-10AD; Shimazu), a column (TSKgel G2000SW XL ; Toyosoda), and an UV detector (SPD-10AVP; Shimazu).
Far Western Analysis-Equal amounts of recombinant MBS proteins were applied to SDS-PAGE in 12% acrylamide gels, and transferred to PVDF membranes (Millipore). The membranes were blocked with 5% nonfat dry milk, incubated for 4 h at room temperature with hexahistidine-tagged hHS-M 21 recombinant proteins (A, B, t-1, t-2, t-3, t-4, t-5, o-1, and o-2) solubilized at a concentration of 10 g/ml in Tris-buffered saline (TBS), washed once with TBS, and once with PBS. Subsequently, the membranes were soaked in 100 ml of 0.5% formaldehyde solution in PBS, then in 2% glycine/PBS. After washing twice with PBS and twice with TBS, the membranes were incubated with horseradish peroxidaseconjugated rabbit anti-His 6 tag polyclonal antibody (Santa Cruz) for 1 h. The immunocomplex was detected by enhanced chemiluminescence (Pierce). The experiments were repeated at least three times to confirm the binding results.
Permeabilization of Porcine Renal Artery and Rat Cardiac Myocytes and Force Measurements-Measurements of alteration in Ca 2ϩ sensitivity of muscle contraction using permeabilized porcine renal artery were performed as described previously (21,29,30). In brief, renal arterial rings were permeabilized with 1% Triton X-100 in the Ca 2ϩ -free cytoplasmic substitution solution (CSS) (10 mM EGTA, 100 mM potassium methanesulfonate, 3.38 mM MgCl 2 , 2.2 mM Na 2 ATP, 10 mM creatine phosphatase, 2 M calmodulin, 50 units/ml creatine phosphokinase, and 20 mM Tris-maleate (pH 6.8)) at 24 -25°C for 20 min. The recombinant hHS-M 21 proteins were dissolved in 100 mM potassium methane sulfonate, 20 mM Tris-maleate (pH 6.8), and applied directly in CSS. The permeabilized arterial rings were activated by various concentrations of Ca 2ϩ from 0.05 M to 10 M, and the developed force was monitored. Ca 2ϩ -CSS containing the indicated concentration of free Ca 2ϩ was prepared by adding appropriate amount of CaCl 2 , using a EGTA-Ca 2ϩ binding constant of 10 6 /M (31).
The measurements of alteration in Ca 2ϩ sensitivity using permeabilized rat cardiac myocytes were done as reported previously (32). Single cardiac cells were prepared from rat left ventricular myocardium (33) and permeabilized with 2.5% ␤-escin for 10 min in a relaxing solution (110 mM potassium methane sulfonate, 5 mM magnesium methane sulfonate, 5 mM Na 2 ATP, 10 mM creatine phosphate, 4 mM EGTA, and 20 mM PIPES-KOH (pH 7.1)) at 25°C. In experiments for pCa-tension relation, skinned cells were treated first with 30 M ryanodine and 30 mM caffeine in pCa 6 solution for 4 min to impair the Ca 2ϩ releasing function of the sarcoplasmic reticulum (34) so that the effect of the released Ca 2ϩ from the sarcoplasmic reticulum on pCa-tension relation could be neglected. Ryanodine and caffeine was then washed out with the relaxing solution and the cells were relaxed completely. After this procedure, the striation uniformity remained unchanged, and the initial resting sarcomere length was set at 2.03 Ϯ 0.03 m. The permeabilized cells were then activated by various concentrations of Ca 2ϩ cumulatively from 0.05 to 10 M, and the developed force was monitored as sarcomere length during the activation. The sarcomere length did not vary at ϾpCa 6, varied by less than 0.2 m at pCa 6, and varied by more than 0.2 m at ϽpCa 6 due to a significant internal shortening with obscured striation pattern. After the cell was relaxed completely with the relaxing solution, however, sarcomere length was returned to the initial value. At 10-min intervals, the second activation was repeated in the absence (control) or presence of hHS-M 21 recombinant proteins (A, B, or t-1) solubilized in 20 mM PIPES-KOH (pH 7.1). If the sarcomere length at pCa 6 varied more than 0.2 m, or the sarcomere length after the first activation varied more than 5% from the initial one, the records were discarded (35). Using the measured values of cell width and length, each maximum tension was expressed as the force/ cross-sectional area based on the assumption of cylindrical cell geometry, because cell depth could not be measured due to technical difficulties. The activating solutions of the desired Ca 2ϩ concentrations (0.05-10 M) were prepared by adding the appropriate amounts of calcium methane sulfonate to the relaxing solution.
Other Procedures-Nucleotide sequences were determined using dye terminator cycle sequencing pre-mix kit (Amersham Pharmacia Biotech) and ABI373A automated DNA sequencer. SDS-PAGE and the immunoblotting technique were carried out according to the standard procedures (36,37). Concentration of protein was measured by the Bradford method (38) with bovine serum albumin (Pierce) as standard.
Data Analysis-The extent of force development was expressed as a percentage, assigning values in Ca 2ϩ -free buffer (resting state) and in 10 M Ca 2ϩ buffer (maximum contraction) to be 0% and 100%, respectively. The EC 50 value, a concentration of Ca 2ϩ required to induce the 50% force of the maximum response, was determined by fitting the Ca 2ϩ concentration-tension response curves to a four-parameter logistic model (39). The measured values were expressed as means Ϯ S.E. Effects of recombinant hHS-M 21 proteins on Ca 2ϩ -induced force were statistically analyzed by analysis of variance. The obtained EC 50 values from pCa-tension relations were compared in the absence or presence of the hHS-M 21 fragments using Student's t test for paired values, and the P values of less than 0.05 were considered to be statistically significant.

Isolation of hHS-M 21 cDNAs-
We constructed a normalized subtraction PCR-cDNA library between the mRNAs from cardiac and skeletal muscles to obtain heart-specific cDNA fragments. Randomly selected 1,021 clones were sequenced, and the sequence data were compared with the GenBank data base. It was revealed that 243 (23.8%) clones were not matched to the known human genes in the data base except for expressed sequence tags. Forty-seven cDNA fragments from unmatched genes were tested for their expression in various human tissues by the RT-PCR analysis. Among them, three cDNA fragments were expressed only in the heart (data not shown), and one of these heart-specific clones, which were 209 bp of length, was designated as HS602.
The HS602 fragment was used as a probe to isolate cDNA clones from a human heart cDNA library. Restriction mapping and end sequencing of the isolated cDNAs showed that there were at least three types of cDNAs different by a small deletion or insertion. Representative clones of each type, 602-4, 602-6, and 602-7, were completely determined for their sequences, and it was revealed that 602-6 and 602-7 had an insertion of 181 bp and a deletion of 3 bp, respectively, as compared with 602-4 ( Fig. 1). From an estimation by the numbers of isolated cDNA clones corresponding to these three types, it was suggested that two of them, 602-4 type and 602-6 type, were major isoforms and another type, 602-7 type, was a minor isoform.
Full sequences of HS602 cDNAs for two major isoforms were determined from the isolated overlapping clones. These two isoforms span a total of 2,046 and 2,227 bp with the open reading frame predicting to encode for 208 and 224 amino acid residues, respectively (Fig. 2). The full sequence included 827 bp in the 5Ј-untranslated region and 595 or 731 bp in the 3Ј-untranslated region for two different termination codons in the HS602 cDNAs.
Characteristic feature of the predicted proteins was the presence of leucine zipper motifs at the C-terminal end. Although the long isoform, hHS-M 21 B, had different C-terminal structures by the 181-bp insertion from the short isoform, hHS-M 21 A, leucine zipper motifs were found in both isoforms (Fig. 2). A data base search revealed that the amino acid sequence of hHS-M 21 A was 76.8% identical to the 20-kDa smooth muscle small subunit (sm-M20) of chicken myosin phosphatase (14) and 50.9% identical to C-terminal part of MBS (MYPT1) of human myosin phosphatase (15). Of particular interest was that the C-terminal half sequence of hHS-M 21 A was 92.4% and 69.6% identical to the relevant sequence of sm-M20 and MBS (MYPT1), respectively. It was, then, suggested that the HS602 gene encoded for a heart-specific 21-kDa small subunit (hHS-M 21 ) of human myosin phosphatase, and that the hHS-M 21 had two major isoforms; the short type (hHS-M 21 A) and the long type (hHS-M 21 B). To our surprise, hHS-M 21 A cDNA sequence was identical to 3Ј-terminal one-third sequence of MYPT2 cDNA for another MBS (25), except that 5Ј-sided 355-bp sequence of hHS-M 21 cDNA was lacked from the MYPT2 cDNA. In addition, the initiation codon of hHS-M 21 was corresponding to codon 775 of MYPT2, i.e. aa 1-208 of hHS-M 21 A was exactly identical to aa 775-982 of MBS encoded by MYPT2. These observations strongly suggested that the MYPT2 gene was a multi-functional gene encoding for both the striated muscle type MBS and heart-specific M 21 . To obtain a structural evidence of this multicoding feature, we determined the genomic organization of the human MYPT2 gene.
Genomic Organization of Human MYPT2 Gene-To determine exon-intron organization of the human MYPT2 gene, a human genomic DNA library was screened with 602-7 as a probe. Several different genomic clones were obtained and EcoRI or XhoI fragments hybridized to 602-7 were subcloned for the sequencing analyses from primers designed in the MYPT2 cDNA sequence. Representative genomic clones, 602G2, 602G4, 602G3, and 602G1, are shown in Fig. 3 along with their subclones, 602gs2, 602gs7, 602gs4, 602gs3, 602gs1, 602gs11, and 602gs10, containing exons 14 -18, 20, 21, and 23-25. Because the screening of genomic DNA library has failed to isolate clones corresponding to the other exons, the remaining genomic organization was determined by sequencing of overlapping LA-PCR products (Fig. 3). Sequencing analyses of genomic subclones and LA-PCR products have revealed that the MYPT2 gene consists of 25 exons as shown in Fig. 3 and Table I. The hHS-M 21 B cDNA was the product of alternative splicing at exon 24, and the 3-bp deletion in cDNA clone 602-7 was suggested to be a differential splicing product to a minor acceptor site in exon 22 (Table I). It also was revealed that the ankyrin repeats in the N-terminal part of MBS (MYPT2) molecule were encoded by exons 1-7, and two mutually exclusive leucine zipper motifs at the C-terminal part of the hHS-M 21 A and B cDNAs were encoded by exon 25 and 24, respectively (Fig. 3). Quite interestingly, the 5Ј sequence (1-355 bp) of hHS-M 21 cDNA (exon 1 of hHS-M 21 ) was found in intron 13 (1 kbp upstream of the exon 14) of the MYPT2 gene and exons 14 -25 were exactly matched to the hHS-M 21 sequences (Fig. 3). In addition, several cis-elements that were observed in promoter regions of heart-specific genes, GATA binding site, MEF-2-like binding motif, E-box, AP2 and M-CAT, were found in upstream of the hHS-M 21 coding exon 1 (data not shown). These results indicate that the hHS-M 21 cDNAs are transcribed from a heart-specific promoter in intron 13 of the MYPT2 gene.
Expression of MYPT1, MYPT2, and hHS-M 21 in Human Tissues-Expression of MYPT2 (this means MBS coding region of the MYPT2 gene here) and hHS-M 21 in various human tissues was investigated using the RT-PCR analysis and compared with that of MYPT1 (MBS coding region of the MYPT1 gene). As shown in Fig. 4A, hHS-M 21 mRNA was expressed only in the heart, while MYPT2 expression was detected in several tissues, preferentially in heart, skeletal muscle, and brain. It was confirmed by the RT-PCR analysis that exon 24 of the MYPT2 gene was alternatively spliced in encoding for the MBS molecule (Fig. 4A), as is the case for encoding for the hHS-M 21 molecule (Fig. 1). Expression of mRNA skipped for exon 24 was relatively abundant as compared with mRNA utilizing exon 24 in most tissues, whereas similar quantities of both type mRNA were found in the skeletal muscle (Fig. 4A).
A competitive RT-PCR analysis showed the different expres-sion level of MYPT1 and MYPT2 in human tissues (Fig. 4B). The expression of MYPT2 mRNA was higher than that of MYPT1 in the striated muscles, such as heart and skeletal muscle. In contrast, the amount of MYPT1 mRNA was more abundant than that of MYPT2 mRNA in brain and other tissues including lung, uterus, and small intestine. It should be noted here that the expression level of MYPT1 mRNA was relatively constant but somewhat different in tissues tested here (Fig. 4B).

Effect of hHS-M 21 Fragments on the Ca 2ϩ Sensitivity of Contraction in Permeabilized Porcine Renal Artery and Rat
Cardiac Myocytes-To analyze the function of hHS-M 21 in muscle contraction, various recombinant hHS-M 21 proteins were prepared (Fig. 5A) and used in permeabilized cell assays. Fig. 5 (B and C) shows the effect of hHS-M 21 proteins on the Ca 2ϩinduced contraction in 1% Triton X-permeabilized porcine renal artery and 2.5% ␤-escin-permeabilized rat cardiac myocytes, respectively. In these protocols, contractions were monitored by stepwise increases in Ca 2ϩ concentration in the presence of 3 M recombinant hHS-M 21 proteins in permeabilized porcine renal artery and 1 M recombinant hHS-M 21 proteins in permeabilized rat cardiac myocytes. In the Ca 2ϩfree buffer, application of hHS-M 21 , even up to 10 M, was not able to produce any significant tension development in both permeabilized porcine renal artery and rat cardiac myocytes (data not shown). Then, the observed contraction reflected the Ca 2ϩ sensitization effects by the hHS-M 21 proteins.
In permeabilized porcine renal artery, application of 3 M recombinant proteins, hHS-M 21 A, hHS-M 21 B, and hHS-M 21 t-1, caused augmentation of the Ca 2ϩ -induced contractions, and the [Ca 2ϩ ] i force relation curves were shifted to left as compared with that in the controls where no recombinant proteins were added (Fig. 5B). The EC 50 (Table II).
In permeabilized rat cardiac myocytes, the [Ca 2ϩ ] i force relationships obtained in the presence of these recombinant proteins were shifted leftward (Fig. 5C), as observed in the permeabilized porcine renal artery. Although the EC 50  These results indicated that the exogenously added hHS-M 21 proteins increase the Ca 2ϩ sensitivity of the contractile apparatus in the permeabilized smooth muscle (porcine renal artery) and cardiac muscle (rat cardiac myocytes). The effect of hHS-M 21 was prominent in the smooth muscle, but it was significantly observed also in the cardiac muscle. Because the presence or absence of C-terminal leucine zipper motifs showed a little but with no statistically significant difference in the effect of muscle contraction, it was suggested that the main functional domain of hHS-M 21 was not located in the C-terminal leucine zipper motifs.
Interaction of hHS-M 21 with MYPT1 and MYPT2-As demonstrated in the previous section, hHS-M 21 showed a prominent effect in the smooth muscle and, to less extent, in the cardiac muscle. This phenomenon might be curious, because the hHS-M 21 gene is expressed only in the cardiac muscle and not in the smooth muscle. However, it can be explained if hHS-M 21 exhibits its function mainly through interaction with MBS encoded by MYPT1, and not with MBS encoded by MYPT2. To investigate the interaction of hHS-M 21 with MBSs encoded by MYPT1 and MYPT2, an overlay assay was used to evaluate the binding affinity between them. Both isoforms of full-length hHS-M 21 (A and B) showed a binding ability to C-terminal one-third of MYPT1-MBS (MYPT1 P), while its binding to the corresponding part of MYPT2-MBS (MYPT2 P) was extremely low (Fig. 6A). In addition, the hHS-M 21 proteins did not bind to N-terminal two-thirds of MBS encoded by either MYPT1 or MYPT2 (Fig. 6A).
To  (Table II). On the other hand, the [Ca 2ϩ ] i force relationship in the presence of 3 M hHS-M 21 recombinants, t-3, t-4, t-5, and o-2, overlapped with that in the control (data not shown) and the EC 50 values were not significantly different from the control value (Table II). Therefore, the main active domain of hHS-M 21 was mapped in the similar region as the binding domain, although the enhancing activity of o-1 (residues 1-56) was less than that of t-2 (residues 1-110).

DISCUSSION
In the present study, we isolated and characterized cDNA clones for the human myosin light chain phosphatase (MLCP) small subunit (hHS-M 21 ) that was expressed only in the heart. Genomic organization of the human MYPT2 gene was determined, and it was revealed that one of the large subunits (MYPT2-MBS) and the small subunit of MLCP in the heart are the products of the same gene. In addition, we investigated the function of hHS-M 21 as measured by the Ca 2ϩ sensitivity of contraction in the permeabilized porcine renal artery and rat cardiac myocytes. Moreover, the interaction of the hHS-M 21 with MBSs encoded by MYPT1 and MYPT2 was investigated by the overlay assay using recombinant proteins. We found that: 1) hHS-M 21 induced an additional contraction at a constant Ca 2ϩ concentration and shifted the [Ca 2ϩ ] i force relationship to the left; 2) the fragments containing the N-terminal half conferred this action of hHS-M 21 , while the deletion of the N-terminal 56 residues completely abolished the action; 3) hHS-M 21 bound to the C-terminal one-third of MBS encoded by MYPT1 with high affinity, and only to a little extent, bound to the same region of MBS encoded by MYPT2; and 4) the binding domain of hHS-M 21 to MYPT1-MBS was mapped to the same region as the main effective domain for the Ca 2ϩ sensitivity on the Ca 2ϩ -induced contraction in permeabilized porcine renal artery.
The MLCP holoenzyme is composed of three subunits; catalytic subunit and small and large regulatory subunits (5,6). The MYPT2 gene locates on chromosome 1q32 and encodes for MBS of an approximately 125-kDa protein in the cardiac muscle, of which sequence is 61% identical to that of the human MYPT1 gene expressed in the smooth muscle (25). Unexpectedly, the C-terminal region (residues 775-982) of the MYPT2-MBS was 100% identical to the full-length amino acid sequence of the hHS-M 21 A, one of the two major isoforms of human hHS-M 21 , and moreover, residues 775-954 of the MYPT2-MBS were 100% identical to residues 1-180 of the hHS-M 21 B, another isoform of human hHS-M 21 . These observations led us to speculate that the hHS-M 21 isoforms may be products of the MYPT2 gene. To confirm this speculation, we carried out determination of genomic organization of the MYPT2 gene and investigated their expressions by the RT-PCR analyses. It was revealed that sequence of 5Ј-untranslated region of the hHS-M 21 cDNA was present in intron 13 of MYPT2, and that the two major isoforms of human hHS-M 21 were derived from the alternative splicing of exon 24 of MYPT2 (Figs. 3 and 4). These results have demonstrated that the C-terminal 208 (or 224) residues of MYPT2-MBS are expressed in the heart as a separate protein, termed as hHS-M 21 , of which mRNAs are transcribed from a heart-specific promoter located within an intron of the MYPT2 gene. We also have confirmed that the hHS-M 21 protein is expressed in the human heart by a Western blot analysis using rabbit antisera raised against the recombinant hHS-M 21 A protein (data not shown).
It has recently been suggested that the sm-M20 subunit from chicken gizzard may be produced from an avian orthologue of MYPT2 (40), while the sm-M20 subunit has not been isolated yet from human tissue. However, the C-terminal 120 residues of chicken sm-M20 (residue 67-186) showed 91% identity to the C-terminal 120 residues of hHS-M 21 , whereas the homology in N-terminal residues between chicken sm-M20 and hHS-M 21 were only about 50%. It is of interest that intron 18 of the human MYPT2 gene is long (more than 40 kbp) as is the case with intron 13 (about 20 kbp) where hHS-M 21 exon 1 exists and that the high sequence similarity of hHS-M 21 with chicken sm-M20 was found after the sequence of exon 19. If mRNA of human sm-M20 subunit were transcribed from intron 18 of the MYPT2 gene, it would be a very unique gene that encodes for three different MLCP regulatory subunits: MBS, hHS-M 21 , and sm-M20.
It might be unusual for one gene to produce two or more proteins with different function in the same tissue by utilizing different promoters, but MLCK gene is known to generate two different proteins in the smooth muscle. The C-terminal 154 residues of the smooth muscle MLCK is expressed as an independent protein, telokin (41,42), and the telokin cDNA is transcribed from a promoter within an intron of the MLCK gene (42,43). Telokin binds specifically to dephosphorylated MLC of smooth muscle and inhibits MLC phosphorylation by MLCK (44,45). In addition, telokin induces Ca 2ϩ desensitization by enhancing the MLCP activity in the smooth muscle (46). The function of hHS-M 21 appears opposite to that of telokin, but the production of hHS-M 21 from the MYPT2 gene is anal- FIG. 4. RT-PCR analyses of MYPT1, MYPT2, and hHS-M 21 in various human tissues. A, RT reaction was performed using total RNA from fetal heart, adult heart, skeletal muscle, brain, kidney, liver, uterus, lung, spleen, thymus, small intestine, and colon, followed by PCR to detect expression of MYPT2, hHS-M 21 , and glyceraldehyde phosphodehydrogenase. Two bands of MYPT2 (1376 and 1195 bp) correspond to insertion and deletion of MYPT2 exon 24, respectively. B, comparison of mRNA expression of MYPT1 and MYPT2 in various human tissues. The RT reaction was performed using total RNA from adult heart, skeletal muscle, brain, uterus, lung, and small intestine. Upper band (486 bp) and lower band (179 bp) represent MYPT2 and MYPT1, respectively. Glyceraldehyde phosphodehydrogenase (GAPDH) represents the external control showing that similar amounts of cDNAs were analyzed. All the RT-PCR analyses were done within the course of exponential amplification.
ogous to the situation of telokin produced by the MLCK gene.
Although MYPT2 mRNA of 11.4 kbp was abundantly found in the heart and skeletal muscle (25), the MYPT2 protein (MYPT2-MBS) was reported to be expressed only in the heart and brain (25), in contrast to MYPT1-MBS, which is widely distributed among chicken or human tissues except for skeletal muscle (25,47). In this study, we investigated the expression of MYPT2-MBS and hHS-M 21 in various human tissues at the mRNA level by the RT-PCR analysis (Fig. 4). The amount of MYPT2 mRNA was slightly more abundant than that of MYPT1 mRNA in the cardiac and skeletal muscles, while the MYPT1 mRNA was much abundant as compared with MYPT2 mRNA in the brain and smooth muscle tissues (Fig. 4B). These findings were consistent with that the MYPT1-MBS was present in most tissues except for skeletal muscle, and that it appeared more abundant in the smooth and cardiac muscles than in the other tissues (47). Although both MBSs encoded by MYPT1 and MYPT2 increase the activity of MLCP, MYPT1-MBS was a more efficient activator than MYPT2-MBS, because MYPT2-MBS required about 10-fold higher concentration to achieve the same extent of activation as MYPT1-MBS for which maximum activation was found at approximately equimolar ratio of MYPT1-MBS and catalytic subunit (25). These observations are in good agreement with the present findings that MYPT1-MBS, and not MYPT2-MBS, was a main regulatory/ target subunit bound by hHS-M 21 to modulate the MLCP activity in the cardiac muscle, as in the smooth muscle.
The function of MYPT2-MBS is yet unclear. It was observed in this study that hHS-M 21 proteins bound to the C-terminal one-third of MYPT2-MBS to a lesser extent (Fig. 6). This binding appears not to be background, because hHS-M 21 proteins did not bind at all to the N-terminal two-thirds of MYPT2-MBS under the same conditions (Fig. 6A) and several truncated hHS-M 21 mutants, t-1, t-2, and t-4, showed a weak binding to the C-terminal MYPT2-MBS (Fig. 6B). It also may be worth noting that a weak binding of the hHS-M 21 t-4 protein to MYPT2-MBS was observed to a similar level as that to MYPT1-MBS. Although we could not demonstrate the Ca 2ϩ -sensitizing effect of hHS-M 21 t-4 in the smooth muscle (Table II), we cannot exclude the possibility that there might be other function(s) than the Ca 2ϩ sensitization of contraction, which is conferred by interaction between the C-terminal half of hHS-M 21 and MYPT2-MBS, in the cardiac muscle. Further studies will be required to elucidate the function of MYPT2-MBS.
The C terminus of MYPT1-MBS was reported to contain binding sites for sm-M20 (12,18,19), Rho A (48), arachidonic acid (49), and acidic phospholipids (50). We have demonstrated  here that hHS-M 21 interacts with the C-terminal one-third of MYPT1-MBS and increases the Ca 2ϩ sensitivity of muscle contraction. There might be no evidence for the direct involvement of hHS-M 21 in the Ca 2ϩ sensitization effect. However, it was reported that the Ca 2ϩ -calmodulin MLCK complex induced an increase in Ca 2ϩ sensitivity in rat single skinned cardiac cells (51), and the movements of tension/pCa relationships were in similar extent to that observed in this study (Fig. 5C). It is therefore likely that the increase in the Ca 2ϩ sensitivity by hHS-M 21 is a reflection of the increased MLC phosphorylation.
In addition, the binding domain to MYPT1-MBS and the main activating domain in Ca 2ϩ sensitization were mapped to the N-terminal half region of the hHS-M 21 in this study. It is suggested, then, that the exogenously applied recombinant hHS-M 21 proteins may bind to endogenous MYPT1-MBS and exhibit the inhibitory action on the MLCP activity. On the other hand, the N-terminal 56 residues of hHS-M 21 (representative of hHS-M 21 o-1) were sufficient for full activity of binding to the C-terminal one-third of MYPT1-MBS (Fig. 6), but they were not enough to confer full activity in increasing the Ca 2ϩ sensitivity as demonstrated by hHS-M 21 t-2 (residues 1-110) (Table II). However, neither binding activity nor Ca 2ϩ sensitization effect was found with hHS-M 21 o-2 that encompassed residues 57-110 ( Fig. 6 and Table II). These results suggest that the main binding domain and the main active domain of hHS-M 21 are overlapped considerably in the Nterminal 56 residues and that the main active domain is extended to C-terminal side of 57th residue but not exceeds the 110th residue. It is noteworthy that the main active domain of sm-M20 was mapped in the N-terminal half region (21). Because the C-terminal halves of hHS-M 21 and sm-M20 have virtually identical amino acid sequences and no enhancing function of Ca 2ϩ sensitivity, these observations suggest that the C-terminal halves of MLCP small subunits are dispensable for their functions in regulation of muscle contraction. In turn, these findings indicated that the N-terminal halves of hHS-M 21 and sm-M20 exhibit their functions despite the low similarity in amino acid sequences. It will be interesting to determine which motifs in the N-terminal half of MLCP subunit would confer the function. Further investigations including site-directed mutagenesis of MLCP small subunit genes will be needed to demonstrate the functional motif(s).
The effect of hHS-M 21 on the Ca 2ϩ sensitivity was prominent in the smooth muscle as compared with in the cardiac muscle. The apparent difference in the efficiency of Ca 2ϩ sensitization by hHS-M 21 in these muscles may be due to the difference in the amount of recombinant proteins used in the assays. This possibility is unlikely because the increases in Ca 2ϩ sensitivity with the different amounts of hHS-M 21 were constant within a range from 1 to 10 M in the assay with the porcine renal artery and within a range from 1 to 3 M in the assay with rat cardiac myocytes (data not shown). The other possibilities for the difference are the species difference, porcine versus rat, in the assay system and the different expression or activity level of MYPT1-MBS between the smooth and cardiac muscles because the expression of MYPT1-MBS was a little more abundant in the smooth muscle than in the cardiac muscle. It also is possible that the contraction of tissue (porcine renal artery) can be measured more prominently than that of single cells (rat cardiac myocytes), because the contraction power from tissue is a summation of that from single cells. In support of the last possibility, the effect of MLCK on force development found in demembraned heart muscle strips (52,53) was stronger than that in single-skinned cardiac cells (51).
In the cardiac muscle, it has been poorly understood about the role of MLC phosphorylation. However, a recent report has focused on the relevance of MLC phosphorylation system in the cardiac hypertrophy (27), and our observations highlight the role of MLCP system in the cardiac muscle contraction via identification and functional analysis of the heart-specific MLCP subunit. This is the first report demonstrating that hHS-M 21 , a heart-specific MLCP small subunit, plays a role in the regulation of Ca 2ϩ -dependent contraction in the cardiac muscle via binding to C-terminal one-third of MYPT1-MBS.