Syncoilin, a novel member of the intermediate filament superfamily that interacts with alpha-dystrobrevin in skeletal muscle.

Dystrophin coordinates the assembly of a complex of structural and signaling proteins that are required for normal muscle function. A key component of the dystrophin protein complex is alpha-dystrobrevin, a dystrophin-associated protein whose absence results in neuromuscular junction defects and muscular dystrophy. To gain further insights into the role of alpha-dystrobrevin in skeletal muscle, we used the yeast two-hybrid system to identify a novel alpha-dystrobrevin-binding partner called syncoilin. Syncoilin is a new member of the intermediate filament superfamily and is highly expressed in skeletal and cardiac muscle. In normal skeletal muscle, syncoilin is concentrated at the neuromuscular junction, where it colocalizes and coimmunoprecipitates with alpha-dystrobrevin-1. Expression studies in mammalian cells demonstrate that, while alpha-dystrobrevin and syncoilin associate directly, overexpression of syncoilin does not result in the self-assembly of intermediate filaments. Finally, unlike many components of the dystrophin protein complex, we show that syncoilin expression is up-regulated in dystrophin-deficient muscle. These data suggest that alpha-dystrobrevin provides a link between the dystrophin protein complex and the intermediate filament network at the neuromuscular junction, which may be important for the maintenance and maturation of the synapse.

The muscular dystrophies are a group of clinically and genetically heterogeneous diseases characterized by progressive muscle wasting. While mutations in many muscle genes are known to be responsible for various forms of muscular dystrophy, the most common form is caused by mutations in the Duchenne muscular dystrophy gene that encodes dystrophin (1)(2)(3). Dystrophin is a large, actin-binding protein that is pivotal in assembling a complex of membrane-associated and cytoplasmic proteins, which make up the dystrophin protein com-plex (DPC) 1 (4,5). The DPC can be divided into three subcomplexes: the transmembrane dystroglycan complex, the membrane-spanning sarcoglycan complex, and the ␣-dystrobrevin-syntrophin cytoplasmic complex (5,6). All of these subcomplexes are thought to act in concert to provide a link between the extracellular matrix and the actin cytoskeleton (7) and thereby provide mechanical support to the muscle membrane (8,9). Loss of the DPC from the sarcolemma through the absence of dystrophin is thought to disrupt membrane integrity, ultimately leading to loss of muscle function (10).
␣-Dystrobrevin is a member of the dystrophin family of proteins and a key component of the DPC that binds directly to dystrophin and syntrophin (11)(12)(13)(14)(15)(16). In mammalian muscle, there are at least three isoforms, ␣-dystrobrevin-1 (94 kDa), ␣-dystrobrevin-2 (62 kDa), and ␣-dystrobrevin-3 (42 kDa), which are generated by alternative splicing of the single ␣-dystrobrevin gene (15,17,18). While ␣-dystrobrevin-1 and -2 contain dystrophin and syntrophin binding sites, ␣-dystrobrevin-3 lacks both of these sites (19). The exact function of the individual isoforms remains unclear, but the recent characterization of the ␣-dystrobrevin knock-out mouse has shed new light on the combined functions of these proteins in skeletal muscle. ␣-Dystrobrevin-deficient mice have neuromuscular junction (NMJ) abnormalities (20). While NMJs form normally, the postsynaptic apparatus matures abnormally, resulting in destabilized acetylcholine receptor (AChR) clusters and altered synaptic morphology. ␣-Dystrobrevin-deficient mice also develop mild muscular dystrophy, similar to that of the dystrophin-deficient mdx mouse (21). Unlike the mdx mouse, however, where the loss of dystrophin results in the loss of other components of the DPC at the sarcolemma (22), loss of ␣-dystrobrevin results in neither the perturbation of the DPC nor the disruption of the structural integrity of the muscle membrane (21). This result implies that other mechanisms, including aberrant cell signal transduction, are likely to contribute to the pathogenesis of muscular dystrophy (21,23). Thus, it has been proposed that the ␣-dystrobrevins have both a structural role at the NMJ and a signaling role in muscle.
To further dissect these roles of ␣-dystrobrevin in skeletal muscle, we have screened for new interacting proteins using the yeast two-hybrid system. Here we describe the interaction between ␣-dystrobrevin and a novel member of the intermediate filament superfamily, which we have named syncoilin. Syncoilin is concentrated at the NMJ of normal skeletal muscle, where it colocalizes and coimmunoprecipitates with ␣-dystro-brevin-1. This interaction is likely to be important in maintaining the integrity of the NMJ in normal muscle and provides a link between the DPC and the intermediate filament cytoskeletal network.

EXPERIMENTAL PROCEDURES
Library Construction and Yeast Two-hybrid Screening-The H2K myoblast cell line was used to generate a cDNA library for yeast two-hybrid screening (24,25). Briefly, H2K myoblasts were differentiated for 6 days into myotubes prior to harvest and generation of poly(A) ϩ RNA using the Fast Track 2.0 system (Invitrogen). cDNA was generated using methods described in Ref. 15. BstXI/EcoRI adaptors were ligated onto the cDNA ends and subsequently cloned into the BstXI sites of the yeast expression vector pYESTrp2 (Invitrogen). This fuses the cDNAs with the B42 activation domain. The cDNA library was then transformed into chemically competent Escherichia coli XL-1Blue, and purified plasmid DNA was extracted directly from the plated library. The entire ␣-dystrobrevin-2 coding sequence including the muscle-expressed variable region-3 (vr3) sequence was PCR-amplified using the following primers: forward, 5Ј-GTGAATTCAGAATGATTGAAGAT-AGTGGA-3Ј; reverse, 5Ј-CACTCGAGTGTTAGTTAAGACCTGCAGTA-3Ј. The amplified PCR product was subcloned into the EcoRI/XhoI sites of the yeast bait vector pHybLex/Zeo to generate ␣DB2:pHybLex. This placed the ␣-dystrobrevin-2 sequence in frame with the DNA binding domain of the LexA transcriptional activator. A bait strain was created by transforming ␣DB2:pHybLex into Saccharomyces cerevisiae strain L40 as described in the manufacturer's instructions (Hybrid Hunter; Invitrogen). The bait strain was cotransformed with the H2K cDNA library, and transformants were plated onto minimal yeast medium lacking histidine, tryptophan, uracil and lysine, containing 300 g/ml Zeocin (Invitrogen) and 5 mM 3-aminotriazole. Plates were incubated at 30°C for 5 days, and yeast colonies that grew on histidine-deficient medium were restreaked onto fresh selective plates and assayed for ␤-galactosidase activity as per the manufacturer's instructions. Prey plasmids were isolated from yeast and electroporated into E. coli XL-1Blue electrocompetent cells (Stratagene). The 5Ј-end of each clone was sequenced using a vector primer. To confirm the interaction in yeast, purified prey plasmids were retransformed with the ␣DB2:pHybLex bait and with the empty bait vector pHybLex/Zeo and tested for growth on selective plates and ␤-galactosidase activity. Other control yeast strains used were the positive control strain containing pHybLex/Zeo-FOS2 and pYESTrp-Jun plasmids (Invitrogen) and the negative control containing the empty bait and prey vectors pHybLex/Zeo and pYESTrp2.
cDNA Library Screening-The H2K cDNA library was plated onto Hybond-N membranes (1 ϫ 10 6 colonies) and screened by hybridization with the 32 P-labeled 5Ј EcoRI/BamHI fragment of the SNIP4 cDNA clone (see below). Hybridizing clones were isolated, purified, and sequenced using standard methods.
Northern Blotting-Mouse multiple tissue Northern blots (CLON-TECH) were probed with the entire 32 P-labeled SNIP4 sequence, generated by an EcoRI digest. The membrane was hybridized overnight and then washed to a stringency of 0.25ϫ SSC (37.5 mM NaCl, 3.75 mM sodium citrate), 0.1% (w/v) SDS at 65°C and exposed overnight to film.
Antibodies-The entire open reading frame of syncoilin was produced as a thioredoxin fusion protein by subcloning the 1.8-kb EcoRI insert from the SNIP4 cDNA clone into the EcoRI site of pET32c (Novagen). This construct was transformed into E. coli BL21 (DE3) Gold (Stratagene), and SYNC-trx fusion protein synthesis was induced by overnight growth. SYNC-trx was purified under denaturing conditions using the Talon resin (CLONTECH) as per the manufacturer's instructions. Purified protein was used to immunize New Zealand White rabbits using standard protocols. Affinity purification of the SYNC-FP antiserum was performed by coupling the SYNC-trx fusion protein to SulfoLink coupling gel (Pierce) according to the manufacturer's instructions. Eluted SYNC-FP antibodies were further preabsorbed against purified thioredoxin coupled to SulfoLink. Affinity-purified and preabsorbed SYNC-FP antiserum was used in all subsequent applications. The specificity of this antibody was determined by preabsorption of SYNC-FP antiserum diluted 1:200 in phosphate-buffered saline (PBS) with a 1 mM concentration of the immunizing fusion protein SYNC-trx for 1 h at room temperature. The preabsorbed serum was used to label skeletal muscle quadricep sections as described below. The following polyclonal antibodies are described elsewhere: dystrobrevin antibodies ␣1CT-FP (26), ␤CT-FP (26), ␣2-PEP (27), and ␣-PAN (27); the anti-dystrophin antibody 2166 (28); the anti-utrophin antibody URD40 (28); and the ␤2-syntrophin antibody 2045 (29). The 9E10 monoclonal anti-Myc antibody was purchased from Covance (Berkeley, CA). The MANDRA1 anti-dystrophin monoclonal antibody was kindly supplied by Prof. Glenn Morris (30). For immunofluorescence studies, rhodamine Red-X-conjugated donkey anti-rabbit IgG (Jackson Immunoresearch), Alexa 488-conjugated goat anti-mouse IgG (Molecular Probes, Inc., Eugene, OR), and Alexa 488 ␣-bungarotoxin (Molecular Probes) were used. For Western blot analysis, horseradish peroxidaseconjugated donkey anti-rabbit IgG (Jackson Immunoresearch) was used.
Generation and Transfection of Expression Constructs-The syncoilin expression construct sync:pCIneo was made by cloning the entire SNIP4 cDNA as an EcoRI fragment into the EcoRI site of pCI-neo (Promega). The Myc-tagged ␣-dystrobrevin-1 and -2 expression constructs containing the muscle-specific vr3 sequence were generated by digesting the cDNA clones m24 (murine ␣-dystrobrevin-1 (15)) and m32 (murine ␣-dystrobrevin-2 (15)) with BglII and inserting a doublestranded DNA insert encoding the Myc epitope (EQKLISEEDL), generated by annealing a pair of phosphorylated oligonucleotides (5Ј-GATCTGGAGCAAAAGCTCATTTCTGAAGAGGACTTG-3Ј, 5Ј-GATC-CAAGTCCTCTTCAGAAATGAGCTTTTGCTCCA-3Ј). The vr3 sequence was generated by PCR on reverse-transcribed first strand cDNAs from skeletal muscle and inserted using the restriction enzymes Bsu361 and ClaI into the appropriate region of m24 and m32, replacing the Bsu361/ClaI fragment without the vr3 sequence with the vr3containing sequence. The modified cDNA clones were excised by SpeI-NotI digestion and inserted in the sense orientation into pCI-neo cut with NheI and NotI. The constructs pDp116 (encoding rat Dp116) and ␤2syn:pCIneo (encoding ␤2-syntrophin) are described elsewhere. 2 All expression constructs were prepared in E. coli XL-1Blue and purified using the Qiagen Plasmid Maxikit ready for transfection. 24 h prior to transfection, COS-7 cells were seeded onto coverslips in six-well plates at a density of 1.5 ϫ 10 5 cells/well or 2 ϫ 10 6 cells/15-cm dish. Cells were transfected with 1 g of plasmid DNA/well or 15 g of DNA/dish using Fugene-6 (Roche Molecular Biochemicals) following the manufacturer's instructions. After 24 h, cells on coverslips were washed twice in PBS, fixed in ice-cold 4% (w/v) paraformaldehyde in PBS for 15 min, and permeabilized in 0.1% (v/v) Triton X-100 in PBS for 10 min prior to being processed for immunofluorescence microscopy as described below. Cells plated on dishes were washed three times in PBS prior to being processed for immunoprecipitation analysis.
Protein Preparation and Western Blotting-Skeletal muscle, heart, kidney, liver, lung, and brain were dissected from a normal adult C57 BL/10 mouse, flash-frozen in liquid nitrogen, and subsequently homogenized separately in 8 ml of SDS/urea buffer. In addition, skeletal muscle extracts were prepared from five different strains of mice: a C57 normal mouse, a dystrophin-deficient mdx mouse, a utrophin-deficient (utrnϪ/Ϫ) mouse (31), a dystrophin-and utrophin-deficient (dko) mouse (32), and an mdx mouse transgenic for full-length utrophin (Fio) (33) and the tissue treated as above. These mice were all ϳ10 weeks old. 25 g of total protein or 10 l of each immunoprecipitation experiment was separated on 10% SDS-polyacrylamide gels and electrophoretically transferred onto nitrocellulose membranes (Schleicher and Schuell). After blocking the membranes for 1 h in Tris-buffered saline containing 0.1% (v/v) Tween 20 (TBS-T) and 5% (w/v) nonfat dry milk, blots were incubated for 1 h with the following primary antibodies: SYNC-FP (1:500). Membranes were washed twice in TBS-T and once in blocking buffer prior to incubation for 1 h with horseradish peroxidase-conjugated donkey anti-rabbit IgG diluted 1:3000 in blocking buffer. Membranes were washed extensively in TBS-T and were developed using the BM chemiluminescence substrate system according to the manufacturer's instructions (Roche Molecular Biochemicals). The Seescan Optical Density System was used to measure the optical density of bands obtained on exposed film. The mean optical density was measured for each band by drawing a standardized box around the band. This value was normalized by subtracting the mean background optical density measured immediately above and below the band.
Immunofluorescence Microscopy-8-m cryosections of adult skeletal muscle from normal C57 BL/10 and mdx mice were prepared as described previously (31). Sections and transfected COS-7 cells were blocked in 10% (v/v) fetal calf serum PBS for 30 min at room temperature prior to incubation with the primary antibodies. Radiation Hybrid Mapping-The T31 Mouse-Hamster Radiation Hybrid Mapping Panel (Research Genetics) (34) was screened to determine mouse chromosomal localization of the syncoilin gene. The forward primer (5Ј-ATGAGAAGCCAGCACTGACA-3Ј) and the reverse primer (5Ј-TCGGTGTTGGGAATCGGTTGA-3Ј) were designed against the SNIP4 cDNA sequence and resulted in a PCR product of 816 base pairs. Ninety-three cell lines were analyzed and scored based on the presence or absence of a PCR product. Data were submitted to the Jackson Laboratory Mapping Panel (available on the World Wide Web) for multipoint analysis. The human chromosomal localization of the syncoilin gene was determined using the G3 Human-Hamster Radiation Hybrid Panel (Research Genetics). Eighty-three cell lines were screened by PCR using the forward primer (5Ј-ATCCATGAGCTTGTATTGCTC-3Ј) and the reverse primer (5Ј-ACTGGGCAGAGAGTCGC-3Ј). These primers were designed against the corresponding human sequence (GenBank TM accession number AL138800) and gave a PCR product of 393 base pairs. Data were submitted to the Stanford Human Genome Center RH Data base (available on the World Wide Web) for analysis.

RESULTS
Identification of a Novel ␣-Dystrobrevin-binding Protein Using Yeast Two-hybrid Analysis-Previously, we have described three ␣-dystrobrevin isoforms, ␣-dystrobrevin-1, -2, and -3, in murine skeletal muscle, and these are shown schematically in Fig. 1A (15,17,19). To identify proteins that interact with ␣-dystrobrevin in skeletal muscle, we performed a yeast twohybrid screen using the bait construct ␣DB2:pHybLex, encoding the entire ␣-dystrobrevin-2 open reading frame including the muscle-expressed vr3 sequence (15). Attempts to produce a bait containing the full-length ␣-dystrobrevin-1 sequence resulted in the nonspecific activation of the yeast HIS3 and lacZ reporter genes. The ␣DB2:pHybLex bait plasmid was screened against a cDNA library generated from mouse H2K myotubes. 1.2 ϫ 10 6 yeast colonies were screened, and a total of 28 positive clones were identified that activated the lacZ and HIS3 reporter genes. Among the prey plasmids isolated were dystrophin, utrophin, and syntrophin cDNAs, demonstrating the success of this screen in identifying known binding partners for ␣-dystrobrevin (12,14,16,35). Importantly, no known components of signaling cascades were identified. Rather, we identified three identical cDNA clones, SNIP4, SNIP12, and SNIP88, that showed strong interaction with ␣DB2:pHybLex in this system. These 2.1-kb clones were found to encode the entire open reading frame of a novel protein that we have called syncoilin. To confirm the syncoilin/␣-dystrobrevin-2 interaction in the yeast system, the SNIP4 cDNA was retransformed into the yeast strain L40 together with the bait plasmid ␣DB2: pHybLex. The resulting cotransformants grew vigorously on medium lacking histidine and also demonstrated ␤-galactosidase activity (Fig. 1B). Cotransformation of SNIP4 with the empty bait vector pHybLex/Zeo confirmed that syncoilin did not nonspecifically activate the yeast reporter genes (Fig. 1B). Control experiments for the yeast two-hybrid system utilized the transcription factors Fos and Jun as a positive interaction and the empty bait and prey vectors that failed to transactivate the yeast reporter genes (Fig. 1B).
Syncoilin Encodes a Novel Member of the Intermediate Filament Superfamily-The complete nucleotide sequence of the cDNA clone SNIP4 predicts an uninterrupted open reading frame from its 5Ј-end with a stop codon at position 1422. The fourth predicted amino acid encodes a methionine residue with an accompanying Kozak consensus sequence. To generate additional 5Ј sequence of the syncoilin transcript and to confirm that this residue is indeed the first methionine of the syncoilin protein, the H2K cDNA library was rescreened with the 5Ј EcoRI/BamHI fragment of SNIP4. This screen identified a single clone that contained an additional 43 nucleotides at its 5Ј-end compared with the original SNIP4 cDNA clone. Within this additional sequence are two in-frame stop codons, confirming that the start of translation does indeed occur at the methionine present in the SNIP4 sequence. The nucleotide sequence and complete open reading frame of syncoilin is shown in Fig. 2A (GenBank TM accession number AJ251641). This gives a protein with a predicted molecular mass of 53.6 kDa and pI of 4.5.
To determine whether the syncoilin sequence is similar to other known proteins, a gapped BLAST search (36) of the nonredundant protein sequence data base was performed at the National Center for Biotechnology Information (available  (60); vr3, variable region 3; cc, coiled-coil domain (13); Y, unique tyrosine kinase substrate domain of ␣-dystrobrevin-1. The dystrophin binding site (DBS) (13,16) and syntrophin binding sites (SBS 1 and 2) (61) are also shown. B, ␤-galactosidase filter assays and growth assays on yeast cotransformed with the bait and prey constructs indicated. Yeast containing interacting clones demonstrated ␤-galactosidase activity with the production of the blue coloration and also grew vigorously on histidine-deficient medium (ϪHIS). on the World Wide Web) using the SEG filter option. This search showed that syncoilin is a novel sequence. Using the COILS prediction program (available on the World Wide Web), syncoilin was found to have a region containing coiled-coils. These coiled-coils were significantly similar to those found in intermediate filament proteins. The top 107 highest scoring proteins in the BLAST search (E Ͻ 4 ϫ 10 -6 ) were all members of the intermediate filament family. Syncoilin was most similar in sequence to rat ␣-internexin (E ϭ 5 ϫ 10 -16 ), a class IV intermediate filament protein (37). Similarity to class III or IV intermediate filaments such as NF-L (E ϭ 5 ϫ 10 -14 ), peripherin (E ϭ 9 ϫ 10 -11 ), glial fibrillary acidic protein (E ϭ 4 ϫ 10 -10 ), vimentin (E ϭ 1 ϫ 10 -9 ), and synemin (E ϭ 3 ϫ 10 -9 ) was also apparent (Fig. 2B). Although coiled-coils are present in several different nonhomologous proteins, such as myosins, kinesins, and desmosomal proteins, the greatest similarity of syncoilin to these 107 intermediate filament sequences strongly suggests that it is a novel member of the intermediate filament protein family. All intermediate filament proteins possess a central ␣-helical coiled-coil region containing ϳ300 amino acid residues, with nonhelical and divergent N-and C-terminal regions (38). It should be noted that the length of the coiled-coil region in syncoilin (296 amino acids) and the alignment, without insertions or deletions, of its sequence against the intermediate filament-diagnostic helices 1A, 1B, 2A, and 2B (38) are entirely consistent with its assignment as an intermediate filament homologue (Fig. 2B). Syncoilin has a N-terminal head domain of 156 residues and a short C-terminal tail domain of 18 amino acids. Both head and tail regions do not show any sequence similarity to sequences in the data base.
The human orthologue of syncoilin was identified by a BLAST search of the human expressed sequence tag data base and the human high throughput genomic sequence data base at the National Center for Biotechnology. The sequence of human syncoilin was assembled from a human genomic sequence (chromosome 1 clone RP5-887B19, map p34.1-34.3, Gen-Bank TM Accession Number AL138800) that overlaps with three expressed expressed sequence tags (GenBank TM accession numbers AI281998, W24127, and AI074110), and the conceptual translation is shown in Fig. 2B. Comparison of the human and mouse sequences reveals an overall identity of 76%.
Syncoilin Is Expressed in Mammalian Skeletal Muscle and Heart-To determine the expression pattern of syncoilin, we hybridized a mouse multiple tissue Northern blot with the entire SNIP4 cDNA clone. As Fig. 3A illustrates, syncoilin transcripts were found to be highly expressed in skeletal muscle and heart, two tissues known to be rich in ␣-dystrobrevin transcripts (15). Two syncoilin transcripts were detected in skeletal muscle and heart, one of ϳ3.0 kb and a larger transcript of ϳ4.5 kb. On longer exposure, weak signals of 3.0 and 4.5 kb were also detected in lung and testes. The same Northern blot was also hybridized with a ␤-actin cDNA control probe to confirm that each lane contained a similar amount of mRNA (Fig. 3A, lower panel).
To investigate the protein products of these transcripts, a syncoilin-specific polyclonal antibody called SYNC-FP was generated. The entire open reading frame of syncoilin contained within the SNIP4 cDNA clone was fused in frame to thioredoxin and expressed in E. coli, and the purified fusion protein was used for immunizing rabbits. The resulting SYNC-FP antiserum was affinity-purified against the thioredoxin-syncoilin fusion protein and subsequently preabsorbed against purified thioredoxin. Fig. 3B shows that probing a mouse multiple tissue Western blot with SYNC-FP led to the detection of two protein products in the heart of ϳ55 and 64 kDa. The presence of two bands suggests that syncoilin may undergo alternative splicing. Alternatively, syncoilin may be subject to proteolytic processing, resulting in the production of a smaller protein species in the heart. The 64-kDa protein was also detected strongly in skeletal muscle, while a faint 55-kDa band was detected in the lung. To determine which of these protein products corresponds to the syncoilin sequence encoded by SNIP4, the SNIP4 cDNA was cloned into pCI-neo to generate the expression construct sync:pCIneo, which was then transfected into COS-7 cells. Western blots of transfected COS-7 cell extracts probed with SYNC-FP revealed that the relative mobility of sync:pCIneo is identical to the 64-kDa protein detected in skeletal muscle and heart (Fig. 3B, final lane). This result suggests that the larger protein product detected on the multiple tissue Western blot corresponds to the protein identified in the yeast two-hybrid screen.
In Vitro Binding of ␣-Dystrobrevin and Syncoilin in Transfected COS-7 Cells-To confirm the interaction between ␣-dystrobrevin-2 and syncoilin, both proteins were expressed in COS-7 cells. First, sync:pCIneo and the Myc-tagged ␣-dystrobrevin expression constructs m32Myc:pCIneo (encoding fulllength ␣-dystrobrevin-2 and including the vr3 sequence) and m24Myc:pCIneo (encoding full-length ␣-dystrobrevin-1 and including the vr3 sequence) were separately transfected into COS-7 cells. Expression of syncoilin was detected using the SYNC-FP polyclonal antibody, and the expression of dystrobrevin was detected using the anti-Myc monoclonal antibody 9E10. As shown in Fig. 4, the transfection of syncoilin (Fig. 4A) and ␣-dystrobrevin-2 (Fig. 4B) and ␣-dystrobrevin-1 (Fig. 4C) into COS-7 cells each resulted in a very similar pattern of protein expression: strong cytoplasmic labeling particularly concentrated around the nucleus, coupled with intense staining of the cell periphery. ␣-Dystrobrevin-1 and -2 immunoreactivity was also detected strongly in the nucleus of transfected cells (Fig. 4, B and C). It should also be noted that overexpression of syncoilin in COS-7 cells did not result in the formation of intracellular filaments. This implies that syncoilin does not self-assemble into 10-nm intermediate filaments under these conditions. Instead, a second intermediate filament protein may be required for heteropolymeric filament formation.
Identical observations were made when sync:pCIneo and m24Myc:pCIneo were cotransfected into COS-7 cells (Fig. 4, G-I), indicating that syncoilin colocalizes with both ␣-dystrobrevin-1 and -2 in this system. To investigate whether syncoilin colocalized with Dp116, a C-terminal dystrophin isoform containing related sequences to ␣-dystrobrevin, pDp116 was cotransfected with sync:pCIneo. The expression of Dp116 in  transfected cells was detected with the anti-dystrophin monoclonal antibody MANDRA1. In this case, syncoilin and Dp116 immunoreactivity did not colocalize in cotransfected cells (Fig.  4, J-L). In control experiments, COS-7 cells were cotransfected with m24Myc:pCIneo and a known interactor, ␤2-syntrophin which was also expressed in pCI-neo. ␤2-Syntrophin was detected with the polyclonal antibody 2045. Clearly, ␤2-syntrophin expression colocalizes with ␣-dystrobrevin-1 expression in transfected COS-7 cells (Fig. 4, M-O). However, these data do not demonstrate a true interaction between syncoilin and ␣-dystrobrevin, since both proteins are shown to be targeted to the same intracellular compartments (Fig. 4, A-C).
Colocalization of ␣-Dystrobrevin and Syncoilin in Skeletal Muscle-To determine the localization of syncoilin in skeletal muscle, normal C57 mouse quadricep cryosections were labeled with the antibody SYNC-FP. Fig. 6, A, E, and I, illustrates that syncoilin was concentrated at NMJs of normal muscle, as identified by double labeling the sections with ␣-bungarotoxin (Fig.  6, B, F, and J). On longitudinal sections, syncoilin appears to be localized to the contractile apparatus in a stripelike pattern, consistent with the labeling of Z-lines (Fig. 6, I and J). Some weak sarcolemmal labeling was also observed, predominantly on type I fibers as determined by NADH staining (data not shown). This syncoilin immunostaining was completely abolished by incubating the antibody with the immunizing fusion protein (Fig. 6, M and N). The labeling pattern of syncoilin most closely resembles that of ␣-dystrobrevin-1, which was concentrated at the NMJs of normal muscle and also weakly detected at the sarcolemma (Fig. 6, C, D, G, H, K, and L). Fig.  6, E and G, show serial muscle sections labeled with SYNC-FP and ␣1CT-FP, respectively. The staining patterns demonstrate that ␣-dystrobrevin-1 and syncoilin immunoreactivity colocalize in skeletal muscle. Using the antibody ␣2-PEP, ␣-dystrobrevin-2 immunoreactivity was also shown to be enriched at NMJs, although strong sarcolemmal labeling was also observed (Fig. 6, O and P). Thus, it is possible that syncoilin is also associated with ␣-dystrobrevin-2 at NMJs.
In Vivo Association of ␣-Dystrobrevin-1 and Syncoilin-To demonstrate an in vivo association between syncoilin and ␣-dystrobrevin, a series of immunoprecipitation experiments were performed using extracts from cultured myotubes and mouse skeletal muscle. Cultured H2K myotubes were extracted in solubilization buffer, and immunoprecipitations were performed using the antibodies SYNC-FP and ␣1CT-FP (Fig. 7, A and B). Using SYNC-FP as the precipitating anti-body, a small amount of ␣-dystrobrevin-1 was detected in the resulting immune complexes (Fig. 7A). Reciprocal experiments using ␣1CT-FP for immunoprecipitation resulted in the detection of a faint syncoilin band (Fig. 7B). These precipitated complexes were also tested for the presence of utrophin and dystrophin, which are known binding partners for ␣-dystrobrevin in skeletal muscle. Utrophin was found to be strongly immunoprecipitated with ␣1CT-FP, but only a very weak signal was detected in the SYNC-FP precipitates (Fig. 7C). Meanwhile, dystrophin was only weakly detected in these myotube extracts and was not detected at all in SYNC-FP immunoprecipitates (Fig. 7D). A faint dystrophin band was observed in ␣1CT-FP immune complexes, demonstrating an interaction between ␣-dystrobrevin-1 and dystrophin.
Immunoprecipitation experiments performed using skeletal muscle extracts and the SYNC-FP antibody resulted in the precipitation of a small amount of ␣-dystrobrevin-1 (Fig. 7E). A reciprocal immunoprecipitation using ␣1CT-FP did not result in the detection of syncoilin (data not shown). The anti-utrophin antibody URD40 was included in this set of immunoprecipitations as a control, and it can be seen that this antibody results in the precipitation of ϳ7.5 times as much ␣-dystrobrevin-1 as SYNC-FP. Thus, only a small proportion of syncoilin is associated with ␣-dystrobrevin and vice versa. It was also noticeable that the SYNC-FP antibody only resulted in the precipitation of the larger ␣-dystrobrevin-1 isoform containing the vr3 sequence in myotubes and skeletal muscle (Fig. 7, A and E). This result suggested that syncoilin might interact with ␣-dystrobrevin-1 via the vr3 sequence. We tested this possibility using the yeast two-hybrid system by generating a vr3 bait and cotransforming yeast with this new bait and the SNIP4 prey. The vr3 bait did not interact with syncoilin in this system (data not shown), implying that syncoilin does not bind to the alternatively spliced region of ␣-dystrobrevin in skeletal muscle.
Syncoilin Expression Is Increased in Dystrophic Muscle-To investigate the localization of syncoilin in dystrophin-deficient muscle, mdx and normal C57 quadricep sections were immunostained with the SYNC-FP antibody. As Fig. 8A shows, syncoilin immunoreactivity in mdx muscle was found to be maintained at NMJs but was increased at both the sarcolemma and within muscle fibers. This increase in muscle fiber staining appeared to correlate with regenerating muscle fibers, as identified by the presence of centrally located nuclei (data not shown). This observation contrasts markedly with the localization of ␣-dystrobrevin in mdx muscle, which, like other components of the DPC, has greatly reduced immunoreactivity at the sarcolemma (22). Fig. 8A illustrates such immunolabeling of normal and dystrophin-deficient muscle sections with the antibody ␣-PAN, which detects all three ␣-dystrobrevin isoforms in skeletal muscle. Like syncoilin, ␣-dystrobrevin remains enriched at NMJs in mdx muscle.
To determine whether this increase in syncoilin immunoreactivity in mdx muscle is reflected in an increase in levels of the protein, Western blots of total protein from skeletal muscle of normal and mdx muscle were probed with SYNC-FP. In addition, muscle extracts from a utrophin-deficient mouse (utrnϪ/Ϫ) that does not have muscular dystrophy (31), a dystrophin-and utrophin-deficient double knock-out mouse (dko) that has severe muscular dystrophy (32), and an mdx mouse expressing a full-length utrophin transgene that largely rescues the dystrophic phenotype (Fio) (33) were also included. Fig. 8B shows the resulting Western blot and reveals a clear correlation between levels of syncoilin protein and muscle damage; the more dystrophic the muscle, the greater the levels of syncoilin. It is estimated using mean optical density measurements of the resulting bands that levels of syncoilin in mdx  D). Syncoilin was strongly precipitated by ␣1CT-FP (A), and in reciprocal immunoprecipitations, ␣-dystrobrevin-1 was precipitated with the SYNC-FP antibody (B). Syncoilin was also precipitated with ␤CT-FP, the antibody detecting ␣-dystrobrevin-2 (C). In reciprocal immunoprecipitation, ␣-dystrobrevin-2 was obscured by IgG heavy chains (D) but can be detected in the COS-7 cell lysate (lys). In contrast, syncoilin was not immunoprecipitated by the anti-dystrophin antibody 2166 in cells cotransfected with pDp116 and sync:pCIneo (E). As a positive control, ␣-dystrobrevin-1 was strongly immunoprecipitated with the ␤2-syntrophin antibody 2045 (F). lys, COS-7 cell lysate; sync, syncoilin; ␣DB1, ␣-dystrobrevin-1; ␣DB2, ␣-dystrobrevin-2. *, rabbit IgG heavy chain. muscle are ϳ6.4-fold greater than in normal control muscle and in dko muscle that levels are increased 12-fold. In Fio muscle, levels of syncoilin are similar to the control, presumably reflecting the rescue of the dystrophic phenotype by the utrophin transgene. These results for syncoilin contrast with changes in the levels of the ␣-dystrobrevins in dystrophic muscle. While levels of ␣-dystrobrevin-1 are largely unaffected in mdx and dko muscle, levels of ␣-dystrobrevin-2 are significantly reduced (Fig. 8C).
Syncoilin Is Located on Mouse Chromosome 4 and Human Chromosome 1-To determine the mouse and human chromosomal localizations of syncoilin, radiation hybrid panels were screened by PCR using primers specific to the syncoilin gene. The mouse syncoilin gene is located on chromosome 4 at position 59.9-60 centimorgans, between markers D4Mit203 (LOD 18.4) and D4Mit72 (LOD 14.0). The human syntenic region is 1p32-36, and mapping to chromosome 1 was confirmed by the G3 Human-Hamster Hybrid Panel. The syncoilin gene was located between markers SHGC-30801 (LOD 12.98) and SHGC-57292 (LOD 12.32), which correspond to D1S2569 and D1S2676, respectively. D1S2676 has been mapped by fluorescence in situ hybridization to chromosome 1p33-34.3 (Sanger Center Cytogenetics), and a BLAST data base search with the mouse cDNA matched a human chromosome 1 sequence in the interval p34.1-34.3 (GenBank TM accession number AL138800). The localization of syncoilin to human chromosome 1p makes it a candidate gene for muscle-eye-brain disease, which has been mapped to 1p32-34 by linkage analysis (39). DISCUSSION Here we describe the interaction between ␣-dystrobrevin and syncoilin, a novel member of the intermediate filament superfamily. Syncoilin is so named due to its localization at the neuromuscular synapse and the presence of predicted coiledcoils within its primary sequence. Using a combination of yeast two-hybrid analysis, in vitro expression assays, and in vivo colocalization and coimmunoprecipitation experiments, we propose that ␣-dystrobrevin-1 and syncoilin form a complex at the NMJ, suggesting a link between the intermediate filament network and the subsynaptic DPC.
The intermediate filaments, along with actin-containing microfilaments and tubulin-containing microtubules, make up the three major classes of cytoskeletal filaments in multicellular animals (37,38). There are over 50 members of the intermediate filament protein superfamily, and these have been result in the formation of intracellular filaments, indicating that syncoilin cannot self-assemble into 10-nm structures under these conditions. One explanation for this may be that syncoilin is a component of heterotypic polymers, in which case it would require a second intermediate filament family member to form functional filaments. A number of classes of intermediate filament proteins, including the class I and II keratins and the class IV neurofilament proteins, are known to form obligate heteropolymers in vivo, resulting in intermediate filaments that consist of at least two different proteins (38). Furthermore, recent evidence has shown that desmin, the major intermediate filament protein in mature muscle, forms heteropolymeric filaments with a newly described member of the intermediate filament superfamily called synemin (40). Since both desmin and the large intermediate filament protein nestin are localized to the NMJ of normal skeletal muscle, these are candidates for associating with syncoilin to form heterofila- FIG. 7. In vivo association of syncoilin and ␣-dystrobrevin-1 in cultured myotubes and skeletal muscle. Solubilized proteins extracted from H2K myotubes (A-D) and skeletal muscle (E) were subject to immunoprecipitation analysis using antibodies SYNC-FP and ␣1CT-FP. A, immunoblot probed with ␣1CT-FP showing that ␣-dystrobrevin-1 was precipitated using SYNC-FP as the immunoprecipitating antibody in H2K myotubes. B, in the reciprocal immunoprecipitation probed with SYNC-FP, syncoilin was precipitated using ␣1CT-FP. C, immune complexes were also probed with the anti-utrophin antibody URD40 and resulted in the detection of a faint utrophin band in SYNC-FP immunoprecipitates. A strong utrophin band was detected in ␣1CT-FP immunoprecipitates. D, immune complexes probed with the anti-dystrophin antibody 2166. A faint full-length dystrophin band was detected in ␣1CT-FP immunoprecipitations, along with a substantial band of ϳ70 kDa. This band possibly corresponds to the detection of the dystrophin isoform Dp71 that is not expressed in normal adult muscle, although it is not detected in the H2K protein lysate (H2K Lys). Alternatively, this band may correspond to a dystrophin breakdown product. E, ␣-dystrobrevin-1 was immunoprecipitated by SYNC-FP and URD40 in mouse skeletal muscle. Immune complexes were probed with ␣1CT-FP to reveal ␣-dystrobrevin-1 complexed with syncoilin and utrophin. It is notable that utrophin was associated with ␣-dystrobrevin-1 isoforms with and without the alternatively spliced vr3 sequence, while syncoilin was only complexed with ␣-dystrobrevin-1 containing the vr3 sequence. SM Lys, skeletal muscle protein lysate; sync, syncoilin; ␣DB1ϩvr3, ␣-dystrobrevin-1 containing the vr3 sequence; ␣DB1Ϫvr3, ␣-dystrobrevin-1 without the vr3 sequence; Utr, utrophin; Dys, dystrophin. *, rabbit IgG heavy chains.
FIG. 8. Syncoilin expression in dystrophic muscle. A, localization of syncoilin and ␣-dystrobrevin in normal C57 and mdx quadricep cryosections. In normal muscle, SYNC-FP immunoreactivity was concentrated at NMJs, as identified by double labeling with ␣-bungarotoxin (␣-BT). mdx muscle demonstrated increased immunostaining at the sarcolemma and within dystrophic muscle fibers (arrowhead). NMJs were still labeled in mdx muscle but were harder to identify due to the overall increase in immunostaining. Using the antibody ␣-PAN, which detects all three ␣-dystrobrevin isoforms in skeletal muscle, ␣-dystrobrevin immunoreactivity was depleted from the sarcolemma of mdx muscle fibers while maintained at NMJs. Scale bar, 50 mm. B, Western blot of total protein extracted from skeletal muscle of a normal C57 mouse, an mdx mouse, a utrophin-deficient (utrnϪ/Ϫ) mouse, a dystrophin and utrophin double knock-out (dko) mouse, and an mdx mouse transgenic for full-length utrophin (Fio), probed with SYNC-FP. Syncoilin levels were increased 6.4-fold in mdx muscle and 12-fold in dko muscle compared with C57 controls as assessed by optical density measurements. In the Fio mouse, levels of syncoilin were reduced to almost control levels. C, an identical Western blot probed with ␤CT-FP, which detects ␣-dystrobrevin-1 and -2 in skeletal muscle. While levels of ␣-dystrobrevin-1 were largely unaffected in dystrophic muscle, levels of ␣-dystrobrevin-2 were reduced substantially in mdx and dko muscle. Molecular mass markers, in kDa, are shown. ments (41)(42)(43)(44). A second explanation for syncoilin's lack of self-assembly is that this protein may not participate in filament formation per se but instead associates with intermediate filaments to perform an accessory role. Synemin, for example, was initially proposed to play a role in cross-linking intermediate filaments, prior to its identification as a bona fide intermediate filament protein (45).
Syncoilin is highly expressed in skeletal and cardiac muscle, tissues both rich in ␣-dystrobrevin. The loss of ␣-dystrobrevin through targeted disruption of the ␣-dystrobrevin gene results in a number of NMJ abnormalities including the irregular distribution of AChRs and a 50% reduction in the number of junctional folds (20). These results imply that there must be a molecular link between ␣-dystrobrevin and elements important for maintaining synaptic structure. Here we have shown that in skeletal muscle, syncoilin immunoreactivity is concentrated at the NMJ, where it colocalizes with ␣-dystrobrevin-1.
In addition, immunoprecipitation analysis showed that a proportion of syncoilin and ␣-dystrobrevin-1 associate in vivo. This association at the postsynaptic membrane may offer a route for understanding the effects of ␣-dystrobrevin deficiency at the NMJ.
Circumstantial evidence supports the idea that the syncoilin/ ␣-dystrobrevin-1 interaction may be required for NMJ integrity and for the correct localization of AChRs. Intermediate filament proteins have been implicated in synapse organization. For example, desmin, a protein enriched at the NMJ and found near the AChR-rich crests of the junctional folds (41,43), together with actin, has been proposed to form a submembraneous support for AChRs at the NMJ (46). This desmin-containing network is suggested to be important in mediating the excitation of AChRs to the sarcomeric contraction system (46). Similarly, in the central nervous system, the NMDA receptor subunit NR1 interacts directly with the intermediate filament protein neurofilament subunit NF-L. This interaction was also identified using yeast-two hybrid screening and is thought to be important for anchoring or localizing NMDA receptors in the neuronal plasma membrane (47). Thus, there is a clear precedent for the association of intermediate filaments with neurotransmitter receptors and receptor-associated proteins.
Most DPC proteins, including ␣-dystrobrevin-1 and -2, are depleted from the sarcolemma of dystrophin-deficient muscle (19,35). Syncoilin however, shows increased immunoreactivity within mdx muscle fibers and at the sarcolemma. Increased syncoilin levels also appear to correlate with the severity of the muscular dystrophy, implying that syncoilin may be a marker for muscle fiber damage or regeneration. Interestingly, the muscle intermediate filament proteins desmin and vimentin, which are also up-regulated in regenerating muscle fibers, are thought to be part of the regenerative process, rather than representing an immediate response to muscle damage (48). Syncoilin is not the first DPC-associated protein that has been found to be up-regulated in dystrophin-deficient muscle. Others include utrophin (49,50), the sarcoglycan-binding protein filamin-2 (51), and biglycan, which binds ␣-dystroglycan (52).
The involvement of intermediate filament proteins in muscle disease has been described. Mice that lack desmin develop skeletal and cardiac myopathy (53), and recent studies have shown that mutations in the human desmin gene cause a distinct "desmin myopathy," a disease that is often associated with cardiomyopathy (54). In a number of the cases examined, a mutant desmin protein was produced and interfered with the normal assembly of intermediate filaments, resulting in fragility of the myofibrils (54). Another example of the role of intermediate filaments in muscle disease is illustrated by Emery Dreifuss muscular dystrophy. The autosomal dominant form of Emery Dreifuss muscular dystrophy is caused by mutations in the lamin A/C gene (55). Lamin A and C are class V intermediate filament proteins of the nuclear lamina, a lattice structure localized between chromatin and the inner nuclear membrane (38). While it is not understood how mutations in the lamin A/C gene result in muscular dystrophy, it has been suggested that lamins A and C are involved in protecting the nuclear membrane from the mechano-physical stress encountered during muscle contraction and relaxation (56). Finally, loss of the intermediate filament-associated protein plectin also results in muscular dystrophy and epidermal bullosa in humans and mice (57,58).
Mapping of the syncoilin gene to human chromosome 1p34.1-34.3 makes it a candidate gene for muscle-eye-brain disease (39). This disease is characterized by congenital muscular dystrophy with consistent ocular and central nervous system involvement that commonly manifests itself in myopia and severe mental retardation (59). While syncoilin was not detected in adult mouse brain by Northern or Western analysis, there is a possibility that low levels of this gene are expressed in the brain and eye. Alternatively, syncoilin may be developmentally expressed. The association of syncoilin with ␣-dystrobrevin and the involvement of different intermediate filament proteins in muscle disease makes syncoilin a good functional as well positional candidate for this disorder.
In summary, we have shown that ␣-dystrobrevin-1 binds to a novel member of the intermediate filament superfamily, syncoilin, that is highly expressed in skeletal and cardiac muscle. It is possible that this interaction is important for the maintenance of the mature NMJ and in efficient synaptic transmission. The marked up-regulation of syncoilin in dystrophic muscle suggests that this protein may also be involved in the pathogenesis of muscular dystrophy. Further molecular and cellular studies will be required to understand the interactions and function of syncoilin within the intermediate filament network, the importance of its association with the DPC through ␣-dystrobrevin, and its possible involvement in muscle-eyebrain disease.