INTRODUCTION

Top Abstract Introduction Materials and Methods Results Discussion References

The dystrophin-glycoprotein complex (DGC) is a multisubunit protein complex composed of integral membrane, peripheral membrane, and cytoplasmic proteins expressed at the sarcolemma of striated muscle fibers (see references 7, 41, 54, and 62 for reviews). Isolation and cloning of proteins within this complex have provided key insights into the function of the DGC and its role in normal muscle physiology. The skeletal muscle DGC is composed of dystrophin (9, 31); the syntrophins (22); - and -dystroglycan (33); -, -, -, and -sarcoglycan (3, 36, 42, 48, 50, 51, 55, 56); and sarcospan (14). Recently, the DGC within smooth muscle fibers has been characterized (64). This complex differs from the striated muscle DGC in at least two respects. First, -sarcoglycan replaces -sarcoglycan within the sarcoglycan subcomplex, and second, differences occur in the glycosylation pattern of -dystroglycan in the smooth muscle DGC.

One likely function of the DGC is to provide a structural link between the extracellular matrix and the actin cytoskeleton, thereby maintaining the stability of the sarcolemma under contractile forces (10, 21). This link occurs through dystrophin, which binds to filamentous actin (31, 58, 59), and dystroglycanwith its -subunit binding to dystrophin (37) and its -subunit interacting with the extracellular matrix component laminin-2 (24). The DGC likely plays other roles in normal muscle physiology by interacting with cell signaling molecules or other proteins at the sarcolemma. The skeletal muscle DGC has been the most fully characterized in this respect, and several laboratories have demonstrated interactions between DGC components and known signaling molecules (1, 4, 5, 11, 12, 25, 28, 34, 44, 60, 71). Additionally, the skeletal, cardiac, and smooth muscle-specific forms of the DGC may have different functions and specific transient binding partners in each tissue.

Sarcospan is a transmembrane protein found in skeletal, cardiac, and smooth muscle DGCs (14, 15, 64). It was identified as a component of the rabbit skeletal muscle DGC by peptide sequence analysis of the 25-kDa DGC component (14). Primary sequence analysis predicts that sarcospan contains four transmembrane domains with cytoplasmic NH₂ and COOH termini (14, 29, 61). This makes sarcospan unique among other integral membrane components of the DGC, which contain single-transmembrane domains. Protein sequence analysis designates sarcospan as a member of the tetraspan superfamily, most closely related to Rom-1, peripherin, and uroplakin (14). The tetraspan proteins are thought to function as molecular facilitators, mediating interaction between proteins at the plasma membrane. This is an attractive function for a component of the DGC: to facilitate protein-protein interactions within the plane of the plasma membrane. The tetraspan proteins have also been implicated in cell signaling activities including cell adhesion, migration, and proliferation (45, 70). This suggests that sarcospan may be an important mediator of signaling functions within the DGC.

Sarcospan is tightly associated with the sarcoglycans to form a subcomplex within the DGC, the sarcoglycan-sarcospan (SG-SSPN) complex (15). In animal models for the limb-girdle muscular dystrophies, namely, in the -, -, and -sarcoglycan knockout mice (2, 13, 17, 18) and the cardiomyopathic hamster (63), there is a concomitant loss of sarcospan with the sarcoglycans at the sarcolemma. When the DGC has been successfully restored in Sgca-deficient mice or cardiomyopathic hamsters by adenoviral injections of either -sarcoglycan or -sarcoglycan, respectively, sarcospan expression is also restored (15, 17). Biochemical evidence further demonstrates this association. The SG-SSPN complex cofractionates when the DGC is disrupted by alkaline treatment or in the mdx mouse, a dystrophin-deficient animal model for Duchenne muscular dystrophy. Additionally, sarcospan can be coimmunoprecipitated with the sarcoglycans in an in vitro cell expression system (15) and from skeletal muscle membrane preparations (C. S. Lebakken and K. P. Campbell, unpublished data). One function of the SG-SSPN complex is to stabilize -dystroglycan at the membrane (32). Furthermore, the SG-SSPN complex may interact with other sarcolemma proteins and be involved in cell signaling processes.

Primary genetic defects in several components of the DGC lead to muscular dystrophies. Duchenne and Becker muscular dystrophies are caused by mutations in the dystrophin gene. Four forms of autosomal recessive limb-girdle muscular dystrophies (types 2C, 2D, 2E, and 2F) are caused by primary mutations in each of the four sarcoglycan genes (reviewed in references 8, 41, and 62). Additionally, mutations in the laminin-2 chain cause a severe form of congenital muscular dystrophy. To date, no human mutations have been found in the dystroglycans, dystrobrevins, syntrophins, or sarcospan, nor have any unclassified muscular dystrophies been mapped to their chromosomal locations. Human-null mutations in the dystroglycan gene would likely lead to an early embryonic lethality. Recent data suggest that dystroglycan is important in basement membrane formation (30), and dystroglycan-null mice die at a very early embryonic stage (68). -Dystrobrevin-deficient mice maintain expression of DGC components, however, and develop a mild muscular dystrophy (26). In contrast, disruption of the 1-syntrophin gene in mice does not result in muscle degeneration (38).

To address the consequences of sarcospan deficiency, we have generated a sarcospan-deficient mouse through standard techniques of homologous recombination in embryonic stem (ES) cells. In this study, we demonstrate that the Sspn-deficient mice do not show any hallmarks of a muscular dystrophy phenotype. Surprisingly, the other components of the DGC are maintained at the sarcolemma in the absence of sarcospan. The muscle appears normal by histological analysis, and it maintains sarcolemma integrity as determined by serum creatine kinase measurements and the absence of Evans blue dye uptake. Additionally, the muscle maintains normal force and power generation capacity. These data suggest either that sarcospan is not required for normal DGC function or, more likely, that the muscle compensates for the loss of sarcospan in some manner.

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results Discussion References

Isolation of murine sarcospan genomic and cDNA clones. The murine sarcospan genomic DNA sequence was obtained by screening a 129/SvJ mouse genomic library (Lambda FIX II; Stratagene, La Jolla, Calif.) using a mouse sarcospan expressed sequence tag (EST) (IMAGE clone 406138; obtained from Research Genetics, Inc., Huntsville, Ala.) containing a portion of the sarcospan 3' untranslated region (UTR). Clones containing the sarcospan insert were analyzed by restriction enzyme mapping and DNA sequencing using dye terminator cycling on a 373 Stretch Fluorescent Automated sequencer (Applied Biosystems). Twenty-four kilobases of the sarcospan genomic DNA sequence was analyzed and found to contain exons 2 and 3 and flanking regions. Murine sarcospan cDNA was isolated by screening a mouse skeletal muscle library (a generous gift of Jeffrey Chamberlain, University of Michigan) with the same EST. Clones were analyzed by DNA sequencing as described above and are identical to that previously published by Scott and colleagues (61) (GenBank accession no. U02487).

Generation of the Sspn targeting plasmid. The Sspn targeting plasmid was constructed using a modified version of the positive-negative selection vector pPNT (obtained from Richard Mulligan, Whitehead Institute, Cambridge, Mass.) (67). A neomycin cassette flanked by loxP sites replaced the neomycin cassette from the original pPNT vector. The short arm of homology in the targeting vector was a 2.1-kb BsaHI fragment containing a 3' portion of intron 1. This fragment was inserted into the modified pPNT vector in the ClaI cloning site upstream of the phosphoglycerate kinase (PGK)-neomycin resistance cassette. Orientation of this fragment was confirmed by restriction enzyme digestion and sequencing. To generate the long-arm region of homology, a 10-kb NheI/NotI fragment containing a portion of the 3' UTR, located in exon 3, and downstream sequences was isolated from a genomic clone and subcloned into pBluescript KS(+) (pBS). From this, a 6.5-kb BamHI fragment was subsequently isolated and inserted into the BamHI cloning site distal to the PGK-neomycin resistance cassette. Orientation of this fragment was also confirmed by restriction enzyme digestion and sequencing. The vector contained a thymidine kinase cassette distal to the long arm. The mutant gene therefore lacked 7.6 kb, which included 223 bp of intron 1, exon 2, intron 2, and 1.8 kb of exon 3 (which encompasses the entire coding region of exon 3).

Generation of Sspn null mice. The targeting construct was linearized with NotI and introduced into 2 × 10⁷ R₁ ES cells by electroporation (Bio-Rad Gene Pulser; Hercules, Calif.). The ES cells were maintained on feeder layers under the selective pressure of G418 and ganciclovir. Colonies surviving the selection were isolated, expanded, and screened for targeting fidelity by Southern blot analysis (Fig. 1). Nine targeted clones were obtained from 236 analyzed. These clones were subsequently reconfirmed by Southern blot analysis by digestion with XbaI (X, Fig. 1A) and hybridization with a second probe (probe 2, Fig. 1A) to verify correct targeting. Finally, Southern blot analysis was performed with a PGK-neomycin resistance gene probe to confirm a single targeting event. Cells from two targeted clones (U104 and U202) were microinjected into C57BL/6J blastocysts and transferred into pseudopregnant recipients. Chimeric animals resulting from the microinjections were bred to C57BL/6J mice, and agouti pups were screened for germ line transmission of the mutant allele. The genotypes from these matings and all subsequent matings were determined by PCR on DNA from tail biopsy specimens (Fig. 1). The following oligonucleotide primers were used: SPNi1FA (primer 1), ACTCCCTGGAATACAGAGGAACT; SPNi2RA (primer 2), TACAAGGGGACAGACACTCAGAC; and NEOspnKO (primer 3), TTTCTCTTGATTCCCACTTTGTG. PCR conditions were as follows: denaturation at 94°C for 2 min, followed by 35 cycles of 1 min at 94°C, 1 min at 55°C, and 1.5 min at 72°C. Southern blot analysis was periodically performed to confirm the fidelity of the PCR. We have obtained identical results from both targeted lines (U104 and U202). Analysis thus far has been carried out on the hybrid C57BL/6-129-SV/J background. However, initial results analyzing Sspn-deficient mice on the 129-SV/J background are consistent with the data presented in this report. All animals were housed in the animal care unit of the University of Iowa College of Medicine according to animal care guidelines.

View larger version (26K): [in this window] [in a new window]   FIG. 1.   Generation of sarcospan-deficient mice by homologous recombination in ES cells. (A) Restriction map of the wild-type murine Sspn locus, the targeting construct, and the targeted locus. A 7.6-kb region including exon 2 (E2) and the coding region of exon 3 (E3) (black boxes) was deleted and replaced by a PGK-neomycin cassette (Neo) flanked by loxP sites (arrowhead). (B) Southern blot analysis. Using probe 1 (as illustrated in panel A), the wild-type locus contains a 16.2-kb EcoRV (RV) fragment, whereas the targeted allele contains a 7.5-kb EcoRV (RV) fragment. (C) Genotype analysis by PCR. Primer sites are shown in panel A, denoted by arrows; using primers 1 and 2, the wild-type allele (+/+) corresponds to an 840-bp PCR product; using primers 1 and 3, the null allele (/) corresponds to a 425-bp product. (D) Northern blot analysis of total RNA isolated from skeletal muscle of wild-type, heterozygous, and Sspn-deficient mice using a cDNA probe spanning the entire sarcospan coding region. A 4.5-kb transcript is detected in the wild type (+/+) and heterozygous (+/) mice but is not detected in the Sspn-deficient mice (/). (E) Immunoblot analysis. KCl-washed skeletal muscle microsomes were separated by SDS-polyacrylamide gel electrophoresis, transferred to nitrocellulose, and probed for sarcospan using the polyclonal R256 antibody, generated against the N-terminal cytoplasmic domain of murine sarcospan. A 20-kDa band is detected in the wild-type and heterozygous mouse preparations (+/+ and +/) but is not detected in the Sspn-deficient mouse preparations (/).

Antibodies. Polyclonal antibodies against mouse sarcospan were generated by injecting New Zealand White rabbits at intramuscular and subcutaneous sites with a C-terminal sarcospan-glutathione-S-transferase fusion protein (amino acids 186 to 216 of mouse sarcospan) (R235) (15) or with an N-terminal sarcospan-glutathione-S-transferase fusion protein (amino acids 1 to 25 of mouse sarcospan: MGRKPSPRAQELPEEEARTCCGCRF) (R256). Affinity purification of sarcospan antibodies was accomplished using Immobilon-P (Millipore, Burlington, Mass.) strips containing the C-terminal or N-terminal sarcospan maltose binding protein-fusion proteins. Antibody specificity was verified for immunofluorescence and immunoblotting by competition experiments utilizing the corresponding fusion proteins and peptides synthesized to these regions. Polyclonal antibodies to -sarcoglycan (rabbit 98) (55), -sarcoglycan (goat 26) (17), -sarcoglycan (rabbit 245) (17), -sarcoglycan (rabbit 214) (17), -sarcoglycan (rabbit 232) (17), dystrophin (rabbit 31) (53), the laminin-2 chain, -dystroglycan (rabbit 83), and -dystroglycan (affinity purified from goat 20) (33) were used as previously described (17). Monoclonal antibodies IIH6, against -dystroglycan (20), and 8D5, against -dystroglycan (42), were previously characterized. Monoclonal antibodies Ad1/20A6, against -sarcoglycan; BSarc1/5B1, against -sarcoglycan; and 35DAG/21B5, against -sarcoglycan, were generated in collaboration with L. V. B. Anderson (Newcastle General Hospital, Newcastle upon Tyne, United Kingdom).

RNA isolation and Northern blot analysis. Total RNA from control, heterozygous, and homozygous-null skeletal muscle was extracted using RNAzol (Tel-Test, Friendswood, Tex.) according to manufacturer specifications. Twenty micrograms of total RNA was run on a 1.25% agarose gel containing 5% formaldehyde and transferred to a Hybond N⁺ membrane (Amersham Corp., Arlington Heights, Ill.). RNA was cross-linked to the membrane using a Stratagene UV cross-linker. Membranes were prehybridized and hybridized using standard methods. Washes were carried out at 65°C in 1× SSC (0.15 M NaCl plus 0.015 M sodium citrate)-1% sodium dodecyl sulfate (SDS) followed by 0.1× SSC-0.1% SDS.

Immunofluorescence and histological analysis. Skeletal, cardiac, and smooth muscle from the small intestine was dissected from wild-type, heterozygous, and Sspn-deficient mice and rapidly frozen in liquid nitrogen-cooled isopentane. Seven-micrometer cryosections were analyzed by immunofluorescence as previously described (17) or by staining with hematoxylin and eosin (H&E).

Measurement of serum creatine kinase levels. Serum creatine kinase levels were quantified using a kinetic determination of creatine kinase activity in serum obtained from the retro-orbital sinus. The creatine kinase assay was carried out on a Hitashi 917 spectrophotometer by the University of Iowa Department of Pathology.

Immunoblot analysis of membrane preparations. KCl-washed membranes from skeletal muscle were prepared as previously described (52) with the addition of two protease inhibitors, calpeptin and calpain inhibitor I (17) (Calbiochem-Nova-Biochem Corp., San Diego, Calif.). Proteins were resolved by SDS-polyacrylamide gel electrophoresis on 3 to 15% linear gradient gels and transferred to nitrocellulose membranes. Immunoblot analysis was performed as previously described (52), except that the blots were developed using enhanced chemiluminescence (Supersignal; Pierce Chemical Co.).

Contractile properties. Contractile properties of the extensor digitorum longus (EDL) and soleus muscles of wild-type, heterozygous, and Sspn-null mice (n = 9 for EDL and n = 10 for soleus muscles) were measured in vitro. Mice were anesthetized with avertin (13 to 17 µl/g of body weight). Muscles were isolated and removed carefully from the anesthetized mice and immersed in an oxygenated (95% O₂ and 5% CO₂) bath containing a buffered mammalian Ringer's solution, pH 7.4, which included curare. The solution was maintained at 25°C. The tendons of the muscles were tied securely to a force transducer and a fixed post. Muscles were stimulated directly by the current flow between two large platinum electrodes (6). The voltage of the stimulator was set to provide maximum twitch force, and the muscle length was set at optimum length for force development. With the muscle at optimum length, the frequency of stimulation was increased until force plateaued at maximum isometric tetanic force (P_o). Each muscle was removed from the bath and blotted, and the tendons were trimmed from the muscle prior to weighing to obtain the muscle mass. Total fiber cross-sectional area and specific P_o were calculated based on the direct measurements of muscle mass, muscle length, fiber length, and P_o (6). The force (kilonewtons) was divided by the total fiber cross-sectional area (square meters) to obtain an estimate of the specific force (kilonewtons per square meter) of the EDL and soleus muscles. Each data set was analyzed by a two-way analysis of variance appropriate for unequal sample sizes (Statistical Analysis System, Cary, N.C.). Significance was set a priori at P < 0.05.

	RESULTS

Top Abstract Introduction Materials and Methods Results Discussion References

Generation of sarcospan-deficient mice. To generate sarcospan-deficient mice, we first isolated genomic DNA clones from a 129/SvJ genomic library by hybridization screening with a mouse EST corresponding to a portion of the 3' UTR of mouse sarcospan. Isolated clones were analyzed by restriction enzyme digestion and DNA sequencing. From these genomic clones, a targeting vector was designed to replace exon 2 and the coding sequence of exon 3 of sarcospan, which encode three of the four transmembrane domains, the extracellular loop, and the C-terminal intracellular domain of sarcospan (Fig. 1A). Southern blot analysis of 236 neomycin-resistant ES cells demonstrated homologous recombination in nine independent clones (shown for one clone in Fig. 1B). Two of these clones were used to generate chimeric founder mice. Heterozygous mice from the F₁ generation were identified by PCR analysis and were crossed to obtain Sspn-deficient mice (PCR screen is demonstrated in Fig. 1C). Sspn-deficient mice are born from heterozygous matings in a predicted Mendelian inheritance frequency. Northern blot analysis of skeletal muscle RNA from wild-type, heterozygous, and homozygous Sspn-deficient mice revealed a 4.5-kb sarcospan-specific transcript in the wild-type and heterozygous mice which was absent in the Sspn-deficient skeletal muscle (Fig. 1D).

View larger version (80K): [in this window] [in a new window]   FIG. 2.   Sarcospan-deficient mice do not express sarcospan protein. Muscle cryosections (7 µm) were stained with the R256 polyclonal antibody generated against the N-terminal cytoplasmic domain of mouse sarcospan. Sarcospan is expressed in skeletal, cardiac, and smooth muscle of wild-type animals but is not detected in the Sspn-deficient tissues. Shown are results of immunofluorescence analysis of wild-type (A, C, and E) and Sspn-deficient (B, D, and E) quadriceps skeletal muscle (A and B), smooth muscle of the small intestine (C and D), and cardiac muscle (E and F). Bars, 50 µm.

Other components of the DGC are maintained at the sarcolemma in Sspn-deficient animals. We have previously documented that sarcospan is tightly associated with the sarcoglycans to form a subcomplex within the DGC. From analysis of several animal models of sarcoglycanopathies, we have demonstrated that sarcospan expression at the sarcolemma is dependent upon expression of an intact sarcoglycan complex (15, 17, 63). To evaluate whether loss of sarcospan affects the expression of the sarcoglycans and other components of the DGC, immunofluorescence analysis was performed on cryosections of quadriceps muscle from control and Sspn-deficient mice. The sarcoglycans (, , , and ), - and -dystroglycan, and dystrophin are all expressed at the sarcolemma in both control and Sspn-deficient mice (Fig. 3A). This demonstrates that although the localization of sarcospan to the sarcolemma is dependent on the sarcoglycans, the reverse is not the case. Namely, expression of the sarcoglycans is not dependent on sarcospan. These results were identical with mice ranging from 1 day to 8 months of age. To verify this result, immunoblot analysis was performed on KCl-washed microsomes obtained from wild-type, heterozygous, or Sspn-deficient skeletal muscle (Fig. 3B). Consistent with immunofluorescence analysis, the sarcoglycans (, , , , and ), - and -dystroglycan, and dystrophin are expressed at similar levels in membrane preparations.

View larger version (94K): [in this window] [in a new window]   FIG. 3.   Components of the DGC are maintained at the sarcolemma in Sspn-deficient mice. (A) Immunofluorescence analysis of wild-type (+/+) and Sspn-deficient (/) quadriceps muscle for components of the DGC. Expression levels of the sarcoglycans (-, -, -, and -SG), - and -dystroglycan (-, and -DG), and dystrophin (DYS) as well as the laminin-2 chain (LAM2) are unaffected in the Sspn/ muscle. Several muscle groups in mice up to 8 months of age have been analyzed with identical results. Bar, 100 µm. (B) Immunoblot analysis of wild-type (+/+), heterozygous (+/), and Sspn-deficient (/) mouse KCl-washed microsome preparations from skeletal muscle. Equal protein loading (100 µg) in each lane was verified by Coomassie blue staining of lanes that were identically loaded (data not shown). Expression levels of each of the tested proteins are unaltered in the Sspn-deficient mice. Abbreviations are defined in the legend to panel A.

Sarcospan-deficient animals do not develop a muscular dystrophy phenotype. Although Sspn-deficient mice maintain proper expression and localization of the DGC at the sarcolemma, this does not demonstrate that the DGC remains functionally intact in these animals. If the DGC was not functionally intact, the animals might exhibit a phenotype similar to the loss of the complex (i.e., a muscular dystrophy) or a milder myopathy as is observed in the -dystrobrevin-deficient mice (26). To test this, we first examined the skeletal and cardiac muscle morphology of Sspn-deficient mice. Cryosections (10 µm) of quadriceps, biceps, diaphragm, neck, and cardiac muscles were stained with H&E. The skeletal muscle from the Sspn-deficient mice appears normal and does not display any classical hallmarks of a muscular dystrophy phenotype such as necrosis, fibrosis, centrally located nuclei, calcification, or fat infiltration (shown for quadriceps, Fig. 4A and B). Analysis of the cardiac tissue also indicates that there are no observable dystrophic alterations in the heart. We have analyzed mice up to 8 months in age, and muscle from the Sspn-deficient animals is indistinguishable from control muscle.

View larger version (95K): [in this window] [in a new window]   FIG. 4.   Skeletal muscle from Sspn-deficient mice has normal morphology and maintains membrane integrity. The figure shows H&E staining of cryosections from the quadriceps of littermate, sex-matched wild-type (A) or Sspn-deficient (B) mice, Evans blue uptake in Sgca-deficient (C) or Sspn-deficient (D) skeletal muscle, and measurement of serum creatine kinase levels (E). The presence of the muscle-specific enzyme creatine kinase in the blood is another indication of membrane permeability. As a positive control, the serum creatine kinase levels of mdx mice were compared to those of wild-type (+/+), heterozygous (+/), and Sspn-deficient (/) mice. No significant differences in the serum creatine kinase levels between the Sspn-deficient mice and their controls have been detected. However, the mdx mice exhibit very high serum creatine kinase levels. Error bars represent the standard errors of the means.

Sspn-deficient muscle maintains normal force and power generation capabilities. To examine if the Sspn-deficient mice had physiological muscle defects, we analyzed force and power generation capabilities of isolated EDL and soleus muscles. Total fiber cross-sectional area and specific P_o were calculated based on the direct measurements of muscle mass, muscle length, fiber length, and P_o (6). The force (kilonewtons) was divided by the total fiber cross-sectional area (square meters) to obtain an estimate of the specific force (kilonewtons per square meter) of the EDL and soleus muscles. We found no differences in the muscle mass, P_o, or specific P_o among the wild-type, heterozygous, or Sspn-deficient mice (Fig. 5).

View larger version (23K): [in this window] [in a new window]   FIG. 5.   Sarcospan-deficient muscle maintains normal power and force generation capabilities. Contractile properties of the EDL and soleus muscles of homozygous (+/+), heterozygous (+/), and Sspn-deficient (/) male mice were measured in vitro. No significant differences in the normalized power, absolute force, or specific force generation between the groups of mice (n = 9 soleus muscles from each genotype and n = 10 EDL muscles from each genotype) were found. Error bars represent the standard errors of the means.

	DISCUSSION

Top Abstract Introduction Materials and Methods Results Discussion References

Sarcospan is a unique integral membrane component of the DGC found in skeletal, cardiac, and smooth muscle cells. We have generated sarcospan-deficient mice through standard homologous recombination within the Sspn gene. No sarcospan transcript or protein is detected in these animals. Surprisingly, the mice maintain expression of the other integral membrane components of the DGC and do not develop a muscular dystrophy phenotype. Analysis of skeletal and cardiac muscle from these mice demonstrates normal histological parameters. The sarcolemma of the Sspn-deficient mice is not perturbed, as demonstrated by normal levels of serum creatine kinase and the lack of Evans blue-positive fibers in the Evans blue uptake assay. Additionally, Sspn-deficient muscle maintains normal power and force generation capabilities. These data suggest that a functional DGC is formed and that either sarcospan is not required for normal DGC function or the Sspn-deficient muscle is compensating for the loss of sarcospan.

This is the first report describing a null mutation in an integral membrane component of the DGC in mice that does not result in either muscular dystrophy or, as in the case of dystroglycan, early embryonic lethality. Mice with null mutations in -, -, -, and -sarcoglycan have recently been described (2, 13, 17, 18, 27). Each of these mice develops a progressive muscular dystrophy with a dramatic reduction in the expression of the other sarcoglycans at the sarcolemma, as is observed in patients with sarcoglycan mutations (reviewed in references 7, 41, and 62). Additionally, in the -, -, and -sarcoglycan-deficient mice sarcospan expression is also dramatically reduced at the sarcolemma. The SG-SSPN complex is also lost in the smooth muscle of - and -sarcoglycan-deficient mice. Disruption of the smooth muscle DGC accelerates the progression of the muscular dystrophy and leads to the cardiomyopathy observed in these animals (13, 18). Recently, the targeted disruption of the 1-syntrophin gene has been reported. As in the Sspn-deficient mice, there are no gross skeletal muscle defects observed in these animals, and presumably the rest of the DGC is maintained (38). In contrast, although the DGC is maintained at the sarcolemma in -dystrobrevin-deficient mice, they do develop a mild muscular dystrophy and cardiomyopathy (26).

The finding that Sspn-deficient mice maintain normal muscle viability, integrity, and function is important and suggests that human muscle may also be able to compensate for the loss of sarcospan. This may explain why there have been no mutations found in the sarcospan gene which cause a muscular dystrophy phenotype. It is possible, however, that dominant mutations in the sarcospan gene that result in expression of a nonfunctional sarcospan protein might result in disruption of the complex and may be manifested in a muscular dystrophy phenotype, perhaps in a dominant form of limb-girdle muscular dystrophy.

Sarcospan is an integral component of the DGC by all tested criteria: (i) sarcospan expression is lost in Duchenne muscular dystrophy patients and in the mdx mouse, (ii) sarcospan copurifies with the DGC and additionally cofractionates with the complex on sucrose gradient columns (14), and (iii) sarcospan binds with the DGC to laminin affinity columns (R. H. Crosbie and K. P. Campbell, unpublished results). Although we hypothesize that there is a protein functionally replacing sarcospan within the DGC, we cannot rule out the possibility that sarcospan is not required for a functional complex. It is also possible that the role of sarcospan is to link the DGC to other proteins at the plasma membrane. Candidates for such interactions include but are not limited to the 51 and 71 integrins, both of which are important for normal muscle physiology (47, 66).

Sarcospan and the sarcoglycans are tightly associated and form a subcomplex within the DGC. In animal models for limb-girdle muscular dystrophy types 2D, 2B, and 2F, expression of sarcospan at the sarcolemma is severely reduced along with that of the sarcoglycans (2, 13, 17, 18). We have analyzed this tight association biochemically by demonstrating that sarcospan cofractionates with the sarcoglycans after they have been dissociated from the dystroglycans and dystrophin by alkaline treatment and by immunoprecipitation analysis (15). Given this tight association between sarcospan and the sarcoglycans, the finding that the sarcoglycans remain expressed in the Sspn-deficient mice is surprising. We hypothesize that there is a protein functionally replacing sarcospan within the DGC of the Sspn-deficient mice. The identity of this protein remains unknown. Perhaps a homologue of sarcospan, which has not been previously identified, is compensating for the loss of sarcospan. It is also interesting to speculate that an unrelated protein may be functionally replacing sarcospan within the complex.

Sarcospan is a member of the tetraspan superfamily of proteins, which is a large family of proteins expressed in nearly all cell types (45, 70). Each tetraspan protein has four conserved hydrophobic membrane-spanning domains and intracellular N- and C-terminal domains. Across species, the most conserved regions of sarcospan lie within its transmembrane domains, suggesting that these domains are functionally important. It is possible that another tetraspan family member is carrying out the function of sarcospan in its absence. Alternatively, many other four-transmembrane proteins have been demonstrated to mediate protein-protein interactions within the plasma membrane. Examples of such proteins include the connexins (43, 72, 73), which associate to form gap junctions; the claudins (23, 49), which mediate interactions within tight junctions; and the gamma subunits of voltage-gated calcium channels (19, 35, 40, 69), which associate with the ₁, ₂-, and -subunits of the calcium channel. Future work may lead us to the identification of a protein compensating for the loss of sarcospan in the Sspn-deficient muscle.