Calsarcin-3, a novel skeletal muscle-specific member of the calsarcin family, interacts with multiple Z-disc proteins.

The Z-disc is a highly specialized multiprotein complex of striated muscles that serves as the interface of the sarcomere and the cytoskeleton. In addition to its role in muscle contraction, its juxtaposition to the plasma membrane suggests additional functions of the Z-disc in sensing and transmitting external and internal signals. Recently, we described two novel striated muscle-specific proteins, calsarcin-1 and calsarcin-2, that bind alpha-actinin on the Z-disc and serve as intracellular binding proteins for calcineurin, a calcium/calmodulin-dependent phosphatase shown to be integral in cardiac hypertrophy as well as skeletal muscle differentiation and fiber-type specification. Here, we describe an additional member of the calsarcin family, calsarcin-3, which is expressed specifically in skeletal muscle and is enriched in fast-twitch muscle fibers. Like calsarcin-1 and calsarcin-2, calsarcin-3 interacts with calcineurin, and the Z-disc proteins alpha-actinin, gamma-filamin, and telethonin. In addition, we show that calsarcins interact with the PDZ-LIM domain protein ZASP/Cypher/Oracle, which also localizes to the Z-disc. Calsarcins represent a novel family of sarcomeric proteins that serve as focal points for the interactions of an array of proteins involved in Z-disc structure and signal transduction in striated muscle.

The Z-disc of striated muscle cells is a highly complex and specialized three-dimensional structure, consisting of dozens of different proteins assembled into a multiprotein complex. The precise interactions, mechanisms of assembly, and identities of Z-disc proteins are poorly understood (for review, see Refs. 1 and 2). The Z-disc delineates the border of the individual sarcomeric unit and cross-links thin filaments via antiparallel binding of filamentous actin to ␣-actinin, thereby providing a means to transmit force longitudinally. The Z-disc not only serves as the interface of the sarcomere and the cytoskeleton via association with intermediate filament proteins such as desmin, but it also resides in close proximity to the plasma membrane and its associated receptors and ion channels, suggesting that it has additional functions in the transmission and sensing of external and internal signals. Given this central role in the structural integrity and function of striated muscle, it is not surprising that mutations in several proteins that are either Z-disc components or bind to Z-disc proteins have been implicated in the development of dilated cardiomyopathy and/or muscular dystrophy, including actin (3), desmin (4), dystrophin (5), myotilin (6), and telethonin/T-Cap (7). Moreover, several loss of function studies in genetically engineered mice demonstrate the central importance of additional Z-disc proteins in muscle; targeted ablation of muscle LIM protein (MLP) leads to severe myofibrillar disarray with subsequent dilated cardiomyopathy (8), while a deficiency of actinin-associated LIM protein (ALP), conveys a predominantly right ventricular cardiomyopathy (9). Lack of another recently identified striated muscle-specific PDZ-LIM protein, ZASP 1 /Cypher/Oracle, is associated with a severe congenital myopathy with disorganized and partially disrupted Z-discs (10).
Calcineurin is a calcium/calmodulin-dependent serine/threonine phosphatase, which plays an important role in striated muscle signal transduction (for review, see Ref. 11). Its function has been studied most extensively in T-lymphocytes, where it dephosphorylates members of the nuclear factor of activated T-cell (NFAT) family of transcription factors, resulting in their nuclear translocation and activation of target genes. In cardiac muscle, calcineurin has been shown to be activated by hypertrophic agonists such as angiotensin II (12), as well as intrinsic defects of the sarcomere (13). Constitutive activation of calcineurin in hearts of transgenic mice is sufficient to induce severe hypertrophy, ultimately leading to heart failure and sudden death. In skeletal muscle, calcineurin activation has been shown to promote differentiation and fibertype specialization toward a slow-twitch program (14,15).
In an effort to identify tissue-specific modulators of calcineurin signaling, we previously conducted a yeast two-hybrid screen with the catalytic subunit of calcineurin as bait and discovered the first two members of a novel striated musclespecific protein family, calsarcin-1 and calsarcin-2 (16). Calsarcin-1 is expressed in cardiac muscle and oxidative skeletal muscle fibers (types I and IIa), while in the adult, calsarcin-2 is exclusively expressed in skeletal muscle and predominantly in fast-twitch fibers. Calsarcin-2 has also been independently reported as FATZ (filamin-, actinin-, and telethonin-binding protein of the Z-disc) (17) and myozenin (18). Calsarcin-1 and calsarcin-2 proteins bind calcineurin and are localized to the Z-disc, suggesting that they direct its subcellular localization in striated muscle cells, which may be critical to activate calcineurin and to associate it with its downstream effectors.
Here we describe calsarcin-3, the third member of the calsarcin family, which is specifically expressed in skeletal muscle and is enriched in fast-twitch muscle fibers. Similar to calsar-cin-1 and calsarcin-2, calsarcin-3 interacts with calcineurin and Z-disc proteins such as ␣-actinin, ␥-filamin, and telethonin. In addition, we performed a yeast two-hybrid screen to identify novel calsarcin-binding proteins in skeletal muscle and found that calsarcins also interact with ZASP (19)/Cypher (20)/Oracle (21). Taken together, our data suggest that calsarcin family proteins may serve a dual role by linking key Z-disc proteins such as ␣-actinin, ␥-filamin, and telethonin and localizing calcineurin signaling to the sarcomere.

EXPERIMENTAL PROCEDURES
Bioinformatics and cDNA Library Screening-A human genomic DNA data base (GenBank TM ) was searched with the calsarcin-1 protein sequence using the tBlastN algorithm of the BLAST search program to identify potential calsarcin-related sequences. A cDNA fragment encoding calsarcin-3 was cloned by polymerase chain reaction (PCR) utilizing a human skeletal muscle library (CLONTECH). This fragment was subsequently used to screen 5 ϫ 10 6 clones of the same library for calsarcin-3 full-length cDNA. In addition, a mouse skeletal muscle bacteriophage cDNA library (CLONTECH, TriplEx2) was screened for mouse calsarcin-3 with a probe corresponding to its open reading frame.
Northern Blot Analysis-Multiple tissue Nothern Blots (CLON-TECH) containing mouse and human poly(A) ϩ RNA were hybridized overnight at 65°C with [ 32 P]dCTP-labeled cDNA probes corresponding to the open reading frame of mouse and human calsarcin-3, respectively. Serial washes were conducted with 2 ϫ SSC, 0.1% SDS and 0.2 ϫ SSC, 0.1% SDS at 65°C. Autoradiography was performed at Ϫ80°C for 24 -48 h with an intensifying screen.
Generation of Calsarcin-3-specific Antiserum-A peptide consisting of the NH 2 -terminal 20 amino acids (NH 2 -MIPKEQKEPVMAVPGDLA-EPVP-COOH) of calsarcin-3 was synthesized (Biosynthesis) and used to generate antisera in rabbits. The NH 2 terminus was chosen to generate an isoform-specific antiserum, since this part of the protein does not display significant homology to the other two calsarcin family members (Fig. 1B). IgG was purified from rabbit serum using protein A-Sepharose beads (Amersham Biosciences) and subsequently used for immunostaining of skeletal muscle cryosections.
Yeast Two-hybrid Screen-A full-length human calsarcin-1 cDNA, fused to the GAL4 DNA-binding domain (plasmid pAS1, CLONTECH), was used as bait in a yeast two-hybrid screen of approximetaly 1 ϫ 10 6 clones of a human skeletal muscle cDNA library (CLONTECH), as described (12). Briefly, clones displaying differential growth on selective plates, lacking histidine, leucine, and tryptophan, were picked and replated for ␤-galactosidase assays. Positive clones were grown in selective medium lacking leucine, and plasmid DNA was isolated and subsequently electroporated into DH10B Escherichia coli (Invitrogen). The obtained clones were sequenced and retransformed with the calsarcin-1 construct to confirm the interaction. In addition, one of the identified clones, full-length ZASP/Cypher-2/Oracle, was cloned into the GAL4-DNA-binding domain-containing vector pAS1 and tested for ␤-galacosidase activation with human calsarcin-1, calsarcin-2, calsarcin-3, and ␣-actinin, which were cloned into pACT2 (CLONTECH), containing a cassette encoding the GAL4-transactivation domain.

Cloning of Human and Mouse
Calsarcin-3 cDNA-An in silico search of the data base (GenBank TM ) using the tBLASTN algorithm with the calsarcin-1 protein sequence revealed two genomic fragments on human chromosome 5 that contained two highly homologous sequence stretches encoding a calsarcin-related protein we named calsarcin-3 (GenBank TM accession numbers AC022091 and AC008453). Primers matching these sequences were designed to PCR-amplify a partial cDNA (ϳ350 bp) from human skeletal muscle cDNA. Subsequently, this fragment was used as a probe to isolate a full-length cDNA from a human skeletal muscle library. We identified and sequenced multiple identical cDNA clones, encompassing a 1.0-kb transcript (GenBank TM accession number AF480443), that predicts an open reading frame of 251 amino acids (Fig. 1). The 5Ј-end of the open reading frame matches a human brain expressed sequence tag (GenBank TM accession numbers AL037981/AL037982). However, this expressed sequence tag predicts a premature stop codon, presumably due to the presence of unspliced intronic sequences. According to the draft of the human genome, calsarcin-3 maps to chromosome 5q31 and encompasses 7 exons, spanning ϳ19 kb (data not shown).
Mouse calsarcin-3 cDNA was cloned by screening a mouse skeletal muscle bacteriophage library. Four independent overlapping clones of a combined length of 913 bp were identified, predicting a single open reading frame of 245 amino acids (GenBank TM accession number AF480442).
Sequence comparison revealed 75% identity between human and mouse calsarcin-3 (Fig. 1A). As previously described for calsarcin-1 and calsarcin-2 (16), calsarcin-3 also displays the highest amino acid homology to the other calsarcins at its NH 2 and COOH termini, while the intervening sequences are less well conserved (Fig. 1B). Interestingly, none of the highly conserved calsarcin domains shows any significant homology to other proteins in the database.
Skeletal Muscle-specific Expression of Calsarcin-3-Probing of multiple tissue Northern blots with specific probes for human and mouse calsarcin-3 revealed a skeletal muscle-restricted pattern in both species (Fig. 2), while no detectable expression was observed in the adult heart. Mouse calsarcin-3 displayed a single major transcript of ϳ3.5 kb and a minor transcript of ϳ4.0 kb. In contrast, several human calsarcin-3 transcripts were detected, suggesting more complex regulation of the human calsarcin-3 gene. However, our cDNA library screens did not identify additional or alternative coding sequences.
Calsarcin-3 Coimmunoprecipitates with Calcineurin and Several Z-disc Proteins-We previously reported that calsarcin-1 and calsarcin-2 interact with calcineurin as well as the Z-disc proteins ␣-actinin-2/3 and the muscle-specific filaminisoform, ␥-filamin (16). Another sarcomeric protein, telethonin/ T-Cap, which is also localized to the Z-disc (22,23), has also been shown to interact with calsarcin-2 (17). We therefore investigated whether calsarcin-3 is similarly capable of medi- FIG. 1. Protein sequence alignments of calsarcin-3. A, the deduced amino acid sequences of human and mouse calsarcin-3 are shown. A comparison of human and mouse calsarcin-3 reveals an amino acid identity of 75%. B, alignment of human calsarcin-3 with calsarcin-1 and calsarcin-2, revealing highly homologous stretches near the NH 2 and COOH termini of the proteins, whereas the intervening sequences are less well conserved. Conserved amino acids are highlighted.
Calsarcin-3 Interaction Domain Mapping-To map the domains of calsarcin-3 that mediate its many protein-protein interactions, we created several deletion mutations and tested their abilities to coimmunoprecipitate with calcineurin, telethonin, ␣-actinin-2, or ␥-filamin. Deletion of the COOH-terminal 141 amino acids of human calsarcin-3 (mutant 1-110) did not affect the ability to interact with any of these partners (Fig.  4). These interactions were also unaffected by a further deletion of the first 36 amino acids (mutant 37-110), which are not highly conserved among calsarcin isoforms. In contrast, deletion of the NH 2 -terminal 109 amino acids (mutant 109 -251) eliminated binding to ␥-filamin, calcineurin, and telethonin, but it did not abolish the interaction with ␣-actinin-2. Since this mutant did not overlap with mutants 1-110 or 37-110, these results suggested the existence of two independent ␣-actinin-binding domains. Further deletions revealed that amino acids 186 -207, which are highly conserved, are necessary for ␣-actinin binding of the COOH-terminal portion of the protein (compare mutants 109 -207 and 109 -186).
To further pinpoint the protein-protein interaction regions, we added back regions of the protein to mutant 109 -251, which was only able to mediate interaction with actinin. Addition of amino acids 73-109 to this construct (mutant 73-251) restored the ability to bind ␥-filamin, but not calcineurin or telethonin. Further extension of this construct to residue 50 (mutant 50 -251) restored calcineurin and telethonin binding. This same region also mediated binding to ␣-actinin-2 in the absence of the more COOH-terminal binding site (mutant 50 -186) located between amino acids 186 -207. However, further deletion of this construct to residue 110 (mutant 50 -110) rendered a molecule that still bound calcineurin, telethonin, and ␥-filamin, but was not sufficient to interact with ␣-actinin. The latter finding may be due to the small size of the mutant peptide or to altered secondary structure, since both NH 2 -and COOH-terminal extensions of this mutation confer again the ability to bind ␣-actinin.
In our previous study, we mapped the calcineurin-binding domain on calsarcin-1 to the COOH terminus of the molecule (16). To directly compare the calcineurin binding properties of calsarcin-1 and calsarcin-3, we generated an additional series of calsarcin-1 deletion mutants and performed coimmunoprecipitations. While calsarcin-3 mutant 109 -251 did not bind calcineurin, the corresponding calsarcin-1 construct (amino acids 100 -264) was able to bind calcineurin (data not shown), suggesting a structural difference between these calsarcin isoforms accounts for these findings. Moreover, further deletion of the last 34 amino acids of calsarcin-1 abolished calcineurin binding, but not ␣-actinin binding (data not shown), consistent with our previous mapping of the COOH-terminal calcineurin and ␣-actinin interaction domains (16). A calsarcin-1 deletion construct encompassing amino acids 1-100 was also able to mediate interaction with both calcineurin and ␣-actinin-2 (data not shown), indicating that, similar to calsarcin-3, calsarcin-1 contains an additional NH 2 -terminal binding site for these molecules.
Calsarcins Interact with ZASP/Cypher/Oracle-To identify additional novel protein-protein interactions for calsarcin-3, we fused the full-length calsarcin-3 protein and several NH 2 -and COOH-terminal deletion mutants to the GAL4-DNA-binding domain to perform yeast two-hybrid screens of a skeletal muscle cDNA library. However, all calsarcin-3 constructs tested autoactivated the GAL4-dependent ␤-galactosidase reporter in yeast and therefore could not be used for a library screen (data not shown). Therefore, we used a calsarcin-1 bait that had previously been used to successfully screen a cardiac library (16) to screen a human skeletal muscle library, assuming that potential newly identified partners might also bind calsarcin-3. From this screen, we identified nine clones of ZASP/Cypher/ Oracle, a recently described striated-muscle specific protein (19 -21), containing a PDZ domain and three LIM domains.
Several splice variants of ZASP have been identified, all of which share the NH 2 -terminal PDZ-domain. However, all clones identified in our yeast screen encoded for the shortest isoform (ZASP/cypher-2), consisting of 283 amino acids and lacking the three COOH-terminal LIM domains. All nine clones strongly interacted with calsarcin-1 in a ␤-galactosidase assay in yeast. Moreover, in a reverse experiment in which ZASP was fused to the GAL4-DNA-binding domain, we detected a strong interaction with all three calsarcins as well as ␣-actinin-2, which served as a positive control, since it had been shown previously (20) to interact with ZASP (Table I). None of these calsarcin or ␣-actinin GAL4-transactivation domain fusions displayed any background activation of ␤-galactosidase when cotransformed with a plasmid encoding the GAL4-DNAbinding domain only (Table I). We also attempted to assess the interaction of the longest ZASP/Cypher splice variant, encompassing 723 amino acids (ZASP/Cypher-1) with calsarcins in yeast. However, a fusion of Cypher-1 and the GAL4 DNAbinding domain autoactivated the ␤-galacosidase reporter and therefore could not be tested (data not shown).
Calsarcin-3 Is Localized to the Z-disc and Is Preferentiallly Expressed in Fast-twitch Skeletal Muscle Fibers-To analyze the subcellular localization of calsarcin-3 in skeletal muscle, we used an NH 2 -terminal peptide of mouse calsarcin-3 to generate an antibody that did not recognize other calsarcins. Antiserum BSYN2021 revealed a strong Z-disc staining pattern in cryosections of mouse hindlimb skeletal muscle, as assessed by colocalization with sarcomeric ␣-actinin (Fig. 6A, panels a-c). Preimmune serum was used as a negative control (data not shown). Since the BSYN2021 antiserum worked only poorly in Western blot experiments, we further assessed its specificity by immunostaining of tissue culture cells. BSYN2021 antiserum strongly recognized mouse calsarcin-3 overexpressed in COS or C2C12 cells, while neither human calsarcin-3 or the other calsarcin isoforms were detected, confirming its specific immunoreactivity with mouse calsarcin-3 (data not shown).
We previously demonstrated that in addition to its cardiac expression, calsarcin-1 is specifically expressed in oxidative skeletal muscle fibers (type I slow and IIa fast), while calsarcin-2 appears to be a predominantly fast-twitch isoform (16). To determine whether calsarcin-3 is expressed in a fiber typespecific fashion, we examined cross-sections of mouse hindlimb muscles by immunofluorescence. Costaining with calsarcin-3 antiserum and antibodies directed against slow (Fig. 6B, panels  b and c) and fast (Fig. 6B, panels e and f) skeletal muscle myosin revealed a preferential expression of calsarcin-3 in fasttwitch myofibers.

TABLE I Interaction of calsarcins with ZASP/Cypher in yeast two-hybrid assay
A ZASP splice variant (amino acids 1-283) that lacks the three COOH-terminal LIM domains (ZASP/Cypher-2) was identified in a yeast two-hybrid screen of a skeletal muscle library using calsarcin-1 as a bait. To assess the ability of this protein to also bind calsarcin-2 and calsarcin-3, ZASP/Cypher-2 was fused to the GAL4-DNA-binding domain (DBD) and tested against calsarcin 1-3, as well as ␣-actinin-2, which were fused to the GAL4-transactivation domain. ZASP/Cypher-2 strongly activated a ␤-galactosidase reporter when cotransformed with calsarcin proteins and ␣-actinin. In contrast, when cotransformed with the GAL4-DNA binding domain alone, no background activity was observed. 3. Coimmunoprecipitation of calsarcin proteins with calcineurin, ␣-actinin, telethonin, and ␥-filamin. COS cells were transiently transfected with expression vectors encoding Myc-tagged calsarcin-1/-2/-3 and FLAG-tagged calcineurin, ␣-actinin, telethonin, and ␥-filamin (amino acids 2133-2725), as indicated. Immunoprecipitations and Western blots were performed as described under "Experimental Procedures." Top, anti-FLAG immunoblots of anti-Myc immunoprecipitates show the association of calsarcins with calcineurin, ␣-actinin, telethonin, and ␥-filamin. Middle, anti-FLAG immunoblots of transfected COS cell extracts demonstrate the presence of FLAG-tagged input proteins, as indicated. Bottom, anti-Myc-immunoblots of cell extracts demonstrate the presence of calsarcin proteins. Fig. 3 and the text, COOH-terminal Myc-tag fusions of calsarcin-3 deletions were immunoprecipitated with FLAG-tagged ␣-actinin, ␥-filamin, calcineurin, and telethonin, as indicated, and interactions were analyzed by Western blot. Two independent domains are sufficient to bind ␣-actinin. Amino acids 186 -207 are necessary for the COOH-terminal interaction, while an NH 2 -terminal domain (amino acids 50 -67) also mediates association with ␣-actinin. The latter domain is also required for calcineurin and telethonin binding, while amino acids 67-110 mediate the binding of calsarcin-3 to ␥-filamin. ND, not determined.

DISCUSSION
This study presents the initial description of a novel member of the calsarcin family, calsarcin-3, which is exclusively expressed in skeletal muscle and is enriched in fast-twitch muscle fibers. Like calsarcin-1 and calsarcin-2, calsarcin-3 is localized to the Z-disc and has the ability to bind several other Z-disc proteins, in particular ␣-actinin, ␥-filamin, telethonin, and calcineurin A. In addition, we demonstrate that ZASP (Cypher, Oracle), which has also been shown to bind ␣-actinin and to be localized to the Z-disc (20), is a calsarcin-binding protein as well. Consistent with the latter finding, Zhou et al. (10) demonstrated recently that ZASP mutants lacking the actininbinding PDZ domain still localize to the Z-disc, suggesting that it may associate with additional Z-disc proteins. Our data suggest that calsarcins may provide this link.
The Many Partners of Calsarcins-Given the relatively small size of calsarcin proteins, it is surprising how many different protein-protein interactions they are capable of mediating. Our domain mapping studies, which demonstrate overlap of several protein interaction domains on calsarcin-3, suggest that some of these interactions could be mutually exclusive rather than simultaneous. It has recently been shown that the phosphorylation status of telethonin's COOH terminus controls its interaction with the potassium channel subunit minK (24), suggesting that signal-dependent mechanisms may influence proteinprotein interactions among Z-disc proteins. We are currently investigating whether similar mechanisms regulate calsarcin's choice of partners.
Calsarcins display considerable homology in their aminoand carboxyl-terminal regions, suggesting a conservation of functional properties and protein-binding domains. For example, we demonstrated that a COOH-terminal domain, which is highly conserved between calsarcin-3 (amino acids 186 -207) and calsarcin-1 (amino acids 179 -200), mediates ␣-actinin binding on both proteins. In contrast, we previously mapped the calcineurin binding region of calsarcin-1 to the COOH terminus of the protein (16), while the domain mapping studies of calsarcin-3 indicate that calcineurin binds the region from residues 50 -110, but not the COOH terminus. To address this apparent discrepancy we investigated additional calsarcin-1 deletions and found that, while we were able to confirm the COOH-terminal calcineurin binding site on calsarcin-1, both calsarcin-isoforms share an additional NH 2 -terminal calcineurin-binding domain. Moreover, the COOH terminus of calsarcin-2 (residues 163-299) was shown to be sufficient for ␥-filamin binding (18), while we were unable to demonstrate interaction of the COOH-terminal 141 amino acids of calsarcin-3 with ␥-filamin. It is currently unclear whether these findings also reflect distinct properties of calsarcin-3 as compared with the other calsarcins or whether differences in the assays used for mapping are responsible for the apparent discrepancies (i.e. yeast versus mammalian cells, NH 2  In contrast, a directly adjacent section stained with calsarcin-3 antiserum (panel d) and a fast myosin antibody (panel e) shows that their staining overlaps (f), suggesting that calsarcin-3 is predominantly expressed in fast-twitch skeletal muscle fibers. All images were taken at a magnification of ϫ40. calsarcin-2 will therefore be necessary to further dissect the complex protein-protein interactions of the calsarcin family.
Potential Functions of Calsarcins-What could be the function of calsarcin proteins in cardiac and skeletal muscle in vivo? The colocalization and direct interaction with several key Zdisc proteins such as ␣-actinin and ␥-filamin suggests that calsarcins serve to cross-link these proteins, thereby likely contributing to the formation and maintanance of the Z-disc. Furthermore, the interaction with the phosphatase calcineurin, a key transducer of calcium signals that control many aspects of muscle growth and function (reviewed in Ref. 11), points to an additional role of calsarcins in cellular signaling. In light of the vast differences of calcium ion concentrations in muscle cells not only between contraction and relaxation but also among different compartments of the myocyte, it is likely that calcineurin's subcellular localization is critical for its proper activation by calcium/calmodulin. Calsarcins could serve this role by tethering calcineurin to the Z-disc, which may also be important for the association with its substrates. The latter notion is supported by a recent report showing that the transcription factor NFATc, which is dephosphorylated by calcineurin, is also localized to the Z-disc in unstimulated skeletal muscle fibers and becomes translocated to the nucleus in response to chronic electric stimulation (25). The localization of calcineurin to the Z-disc in cardiomyocytes has been confirmed by others (26). It has also been shown that isoproterenol stimulation of cardiomyocytes, which induces hypertrophy through activation of calcineurin, results in translocation of calcineurin from the Z-disc to the nucleus (26).
The association of signaling molecules with the Z-disc is not without precedence and is perhaps not surprising given the close spatial association of the Z-disc with the sarcolemma/ttubule system. Certain protein kinase C isoforms have also been been localized to the Z-disc (27). Of note, the LIM domains of ZASP/Cypher are capable of directly binding protein kinase C (20), suggesting that ZASP may link protein kinase C to the Z-disc in an analagous fashion to the association of calsarcin and calcineurin. Moreover, the finding that telethonin and the potassium channel minK can interact (24) adds another layer of complexity to this multiprotein complex and further suggests a close relationship of the Z-disc and signal transduction cascades. It is therefore interesting to speculate that calsarcins might play a role in mechanosensing and mediation of stretchassociated signals in striated muscle tissue.
The importance of the Z-disc in striated muscle is further supported by the association of several Z-disc proteins with muscle disease. Limb-girdle muscular dystrophies (LGMDs) are a clinically heterogenous group of diseases, characterized by muscle weakness, eventually progressing to wasting of limb muscles (reviewed in Ref. 28). Genetically, dominant (LGMD1) and recessive (LGMD2) forms can be differentiated. Interestingly, the calsarcin-interacting Z-disc protein telethonin has been shown to cause LGMD2G (7). Moreover, several proteins associated with the muscle-specific filamin isoform, ␥-filamin, have been shown to result in muscular dystrophy as well, including another novel Z-disc protein, myotilin (LGMD1A (6, 29)), and ␥and ␦-sarcoglycan (LGMD 2C and LGMD 2F (30 -32)). The close proximity and/or direct interaction of calsarcins with these disease gene products raises the possibility that mutations in calsarcins may also cause muscular dystrophies. Intriguingly, a distinct myopathy, vocal cord and pharyngeal weakness with autosomal dominant distal myopathy (VCPDM), has been mapped to chromosome 5q31 (33), where according to the draft of the human genome the human calsarcin-3 gene is predicted to be located.
Taken together, we propose a dual role for calsarcin-3, and calsarcins in general, which is to support the structural integrity of the Z-disc and to target calcineurin-dependent signaling pathways to the sarcomere. Loss of function studies will allow further dissection of calsarcins' functions in vivo.