Identification of a Chromosome 6-encoded Dystrophin-related Protein*

Dystrophin is the protein product of an X-linked locus which is disrupted to yield the phenotypes of Duchenne and Becker muscular dystrophies. Recently an autosomal transcript was identified (Love, D. R., Hill, D. F., Dickson, G., Spurr, N. K., Byth, B. C., Marsden, R. F., Walsh, F. S., Edwards, Y. H., and Davies, K. E. (1989) Nature 339, 55-58) with a carboxyl-terminal sequence similar to dystrophin. We have isolated part of this chromosome 6-encoded cDNA by polymerase chain reaction cloning and expressed it as a recombinant bacterial protein. Antibodies against this recombinant protein detected a large protein that exactly co-migrates with dystrophin yet is detectable in patients suffering from Duchenne and Becker muscular dystrophies. This protein of similar size to dystrophin may play a functional role similar to dystrophin, but unlike dystrophin, the protein is detected in tissues other than muscle and nerve. This newly identified protein is presumably a member of the spectrin/alpha-actinin/dystrophin family of proteins.

This protein of similar size to dystrophin may play a functional role similar to dystrophin, but unlike dystrophin, the protein is detected in tissues other than muscle and nerve. This newly identified protein is presumably a member of the spectrin/ a-actinin/dystrophin family of proteins.
Duchenne/Becker muscular dystrophy is one of the most common inherited neuromuscular disorders in humans. The etiological basis of this disorder is a quantitative or qualitative defect in dystrophin, which is the protein product of the Xlinked DMD locus (l-3). Dystrophin is a large cytoskeletal protein that forms a network beneath external membranes (4). Extensive sequence similarities exist between dystrophin and the spectrins and ol-actinins (5). Much of the sequence similarity is across the central "rod" domain of each protein which is proposed to form a conserved triple a-helical repeat motif (5). a-Actinin, P-spectrin, and dystrophin sequences show conservation at the amino terminus as well (6) where all three are thought to bind actin filaments. This has led to the hypothesis that the spectrins, a-actinins, and dystrophins are members of a superfamily (7,8 dystrophin and ly-actinin have been further reinforced by immunological cross-reactivity data (9). The spectrins and cu-actinins are themselves families of proteins that contain multiple isoforms (8, 10). Spectrin USUally exists as a heterodimer of LY and /3 subunits associated in an antiparallel manner. The functional multimeric state, however, is thought to be a tetramer of two dimers associated end-to-end. The spectrins are components of membrane cytoskeleton and have been best characterized in erythroid cells; however, many specific non-erythroid isoforms exist (10). The spectrins have been shown to bind F-actin, and at least in the case of the sea urchin this binding is modulated by calcium (11). The spectrin heterodimer usually consists of large M, 220,000-265,000 subunits. The a-actinins are also F-actin binding proteins, many isoforms of which have been defined in a number of species and cell types based upon their molecular mass and structure (8). Muscle and non-muscle cy-actinin isoforms appear to be somewhat functionally divergent based upon the calcium sensitivity of non-muscle isoforms toward binding actin and differential subcellular tissue distribution of isoforms (9). Non-muscle cy-actinin appears to consist of a homodimer with M, 95,000 subunits.
Since the spectrins and a-actinins have many isoforms, at least some of which seem functionally diverse, it is possible that proteins exist that are similar in structure and size to dystrophin. A strong candidate for such a dystrophin-related protein has come from the recent identification of an autosomally encoded partial cDNA (B3 cDNA) corresponding to part of a 13-kilobase mRNA species (12). Over the short stretch sequenced, the predicted protein shares 83% amino acid identity with the carboxyl terminus of dystrophin. To identify and characterize the putative protein product encoded by the B3 cDNA, we have amplified by polymerase chain reaction a 646-bp' region of the B3 cDNA and have directly cloned the amplified product in a bacterial protein expression vector. Antisera were raised against the bacterial fusion proteins and affinity-purified antibodies prepared which were specific for these fusion proteins. These antibodies were then used to study this protein.
We have found this newly identified protein to share the molecular weight and low level of cellular abundance of dystrophin (iVr 427,000 and ~0.01% total protein). However, unlike dystrophin this large dystrophin-relatedprotein (DRP) is present at normal levels in muscle obtained from both normal and Duchenne/Becker dystrophy-affected individuals. Interestingly, this protein, unlike dystrophin, is detected in all tissues studied and thus does not share the muscle and neuron specificity of dystrophin. Based on the previously determined sequence homology with dystrophin, together with our demonstration of similar size and abundance in muscle, we have named this new protein DRP. of a Dystrophin-related Protein BamHI and Hind111 sites were contained in the forward and reverse primers, respectively. The primers used were as follows.
Forward 5'-gggGGATCCagagcactatgacccctcacaatct-3' Reverse 5'-ggggAAGCTTgaactgtgggccggaggcatctgg-3' The amplified 646-bp region of B3 cDNA (12) corresponds to domain IV of dystrophin (5). Polymerase chain reaction products were visualized by ethidium bromide staining of agarose gels. Bands were excised and purified using Geneclean (Bio 101). Clones were treated with T4 polymerase to polish the ends, restricted to completion with appropriate enzymes to generate restriction sites, and ligated into Bluescribe and PATH (13) vectors for sequencing and protein expression, respectively. Verification of inserts was done by sequencing and chromosomal localization to a panel of somatic cell hybrids. The primers were designed to maintain an open reading frame in the PATH 2 vector, thus obviating the need for engineering sites and allowing rapid production of large amounts of fusion protein. Fusion protein was processed for antibody production in rabbits (Hazelton Laboratories) by methods described previously (1, 13). Antibodies specific for DRP were affinity-purified using an Affi-Gel 10 column (Bio-Rad) to which the fusion proteins had been cross-linked covalently, using methods similar to those suggested by the manufacturer (14). The globulin fraction was precipitated using ammonium sulfate. Antibodies to trpE were removed by incubating the antibodies with trpE lysate and centrifuging the precipitated antigen-antibody complexes.
Zmmutzoblotting-Tissue was processed for immunoblotting as previously described (1). Briefly, tissues were solubilized in sample buffer (10% SDS, 0.1 M Tris, pH 8.0, 10 mM EDTA, 50 mM dithiothreitol (l), boiled, and aliquots fractionated on a 3.5-12.5% gradient SDSpolyacrylamide gel. After electrophoresis, proteins were transferred to nitrocellulose filters. Post-transfer gels were stained with Coomassie Blue to control for gel loading, and some filters were subjected to staining with Ponceau S solution, which along with transfer of prestained molecular weight standards serves as a control for the efficiency of transfer.

RESULTS AND DISCUSSION
In order to identify and characterize the protein product of the previously identified "B3" cDNA (12), we used polymerase chain reaction-mediated cloning and fusion protein expression to generate antibodies to the encoded protein species. The resulting antibodies were affinity-purified on antigen columns and used in immunoblot analyses.
The affinity-purified antibodies recognized 10 ng of fusion protein on an immunoblot (Fig. l), indicating that they were a sensitive tool for further characterization of the putative protein product of the B3 cDNA. Immunoblotting of total human skeletal muscle in the same experiment revealed a strong immunoreactive band of very large molecular weight along with some minor low molecular weight cross-reactive bands of variable intensity (Fig. 1). Thus, the affinity-purified antibodies, much like those directed against dystrophin, chiefly detect a high molecular weight component of muscle, and this large protein was presumed to be the product of B3 cDNA.
The antibodies used were generated against a fusion protein having significant sequence homology to dystrophin. Thus, it was possible that the protein recognized by the antibodies might actually be dystrophin rather than the putative product of B3 cDNA. To test this possibility, we utilized these antibodies on immunoblots of muscle biopsies from both Duchenne dystrophy and Becker dystrophy patients who show dystrophin deficiency and abnormal molecular weight dystrophin, respectively. Parallel immunoblots containing these patient muscle samples were processed using either affinitypurified anti-dystrophin antibodies (1) or the anti-B3 antibodies described here. As shown in Fig. 2 (trpE) and fusion protein (trpE+DRP) are indicated. c, immunoblot (27) using affinity-purified antibodies to the fusion protein. Lanes are 10 ng of fusion protein (left lane) or 100 pg of total human muscle protein (right lane). The antibodies chiefly recognize a large molecular weight protein species in muscle. cDNA was synthesized by reverse transcription from human fetal muscle RNA. This was used as a template, and the polymerase chain reaction was performed. The amplified 646bp region of B3 cDNA (12) corresponds to domain IV of dystrophin (5) (a). Polymerase chain reaction products were spliced into Bluescribe and PATH (13) vectors for sequencing and protein expression, respectively. Fusion protein was induced (b) and processed for antibody production in rabbits by methods described previously (1, 13). Antibodies specific for DRP were affinity-purified and used for immunoblotting the fusion protein or skeletal muscle (c). Immunoblotting was as previously described (1). Briefly, aliquots of muscle were crushed in plastic weigh boats with a pestle at -20 "C. Pounded muscle was weighed, and mg amounts were solubilized in a known volume of sample buffer. An appropriate volume of this sample, calculated to give uniform loading was run in each well. The amount of fusion protein was quantitated by running aliquots of the protein preparation on a SDS-Polyacrylamide gel and comparing the intensity of Coomassie Blue staining with standards, as described (1). Molecular weight markers (MJ are followed by (k) denoting 1000 X in this figure and all subsequent figures.
antibodies recognized a single large molecular weight protein which was of similar size and intensity in all of these patients (Fig. 2). In addition, this newly identified protein differs from dystrophin in that it appears to be expressed more highly in fetal muscle tissue than in adult muscle (Fig. 2). To completely eliminate the possibility that we might be observing a crossreaction to dystrophin we tested a Duchenne dystrophy patient having a genetic deletion spanning the entirety of his dystrophin gene (15). This patient with the complete deletion also expressed apparently normal levels of this newly identified species (data not shown). Thus, we conclude that the protein species detected by our anti-B3 antibodies is in fact distinct from dystrophin. Given that this protein is large in size and has extensive sequence homology to dystrophin, we have named this protein DRP.
The cellular abundance of the transcript encoding DRP has been previously compared with that of dystrophin and found to be similar by Northern blot experiments (12) (16) it is probable that estimation of dystrophin and DRP's expression at a protein level would yield similar results. Consistent with this, dystrophin's cellular abundance at a protein level has been estimated and found to be quite low as well (1). We used similar methods to those used for dystrophin to approximate the relative cellular abundance of DRP in muscle tissue. We compared the immunoblot signal seen for 10 ng of fusion protein with that seen in DRP in 100 pg of total muscle protein.
As shown in Fig. 1, the signal intensities of these two lanes were similar, suggesting that the amounts of antigen in the two lanes were similar. Thus the relative cellular abundance of DRP is similar to that of dystrophin when determined under similar conditions (1). However, this analysis is at best a rough approximation because of the variability of transfer of proteins of drastically different molecular weights.
The data presented above suggest that dystrophin and DRP are of very similar molecular weight. To compare the molecular weight of the two proteins, we prepared parallel immunoblots from normal and Duchenne skeletal muscle. One blot was incubated with anti-dystrophin antibodies and the other anti-dystrophin antibodies together with anti-DRP antibodies. The DRP species was evident in the dystrophin-deficient patient and seemed to exactly co-migrate with dystrophin on gradient SDS-polyacrylamide gel electrophoresis (Fig. 3). Given the predicted molecular mass of dystrophin as approximately M, 427,000 (5) DRP must be considered to have nearly the same molecular weight.
Previously the B3 transcript had been reported to be present in muscular tissue and gut smooth muscle (12). With the ability to detect the encoded protein directly we attempted to see if the tissue distribution was similar to dystrophin or different. Since the DRP antisera detected the expected large protein in (dystrophin-deficient) mdx mice, we used tissues MI(k) of both the mdx mouse and normal mice to study tissue distribution of this protein, thereby eliminating the possibility of dystrophin complicating the analysis. A protein of M, similar to DRP was detected, being equally abundant in muscle, brain, kidney, and gut in both normal and mdx mice (Fig. 4). This protein was also detected in the spleen, liver, and testis (data not shown). Thus the large molecular weight protein detected by DRP antisera has a tissue distribution different than that found for dystrophin.
We have identified a new protein species which shares the molecular weight, cellular abundance, and at least partial sequence homology with dystrophin but differs significantly in its tissue distribution. It will be important to determine if DRP is homologous to dystrophin over its entire length and, therefore, a true member of the spectrin/a-actinin/dystrophin family. In this respect, it is interesting to note a recent publication which described a cross-reactive protein recognized by anti-dystrophin antibodies raised against either the amino or carboxyl termini of dystrophin (15). This crossreactive protein was of identical molecular weight as dystrophin yet, unlike dystrophin, it was present in Duchenne dystrophy muscle, mdx mouse muscle, and many non-muscle tissues. We feel that it is very likely that this cross-reactive protein detected by antibodies against dystrophin (15) is in fact the DRP species described in this communication. If Identification of a Dystrophin-related Protein DRP and the cross-reactive protein are indeed identical, then the fact that amino-terminal dystrophin antibodies appeared to recognize the same protein implies that the sequence homology between dystrophin and DRP exists not only at the carboxyl terminus (12) but might in fact extend through the amino terminus. Our present findings concerning the identification of DRP and the previous findings concerning cross-reaction of some dystrophin antisera suggest due caution be applied to the interpretation of dystrophin abnormalities in neuromuscular disease patients. Specifically, the inadvertent visualization of dystrophin-related protein could complicate the diagnostic interpretation of these biochemical assays, which are currently in widespread use. For example, if a poorly characterized anti-dystrophin antibody is used on muscle from a Duchenne patient and DRP is visualized, then the patient might be incorrectly interpreted as having significant levels of dystrophin and, therefore, diagnosed as a non-Duchenne patient. In this regard, it is important to note that a few recent papers using new previously uncharacterized dystrophin antisera have reported what was interpreted as dystrophin in Duchenne dystrophy patients (17,18). It is quite possible that these investigators were in fact visualizing the DRP we have described in this communication. Additionally, in view of this type of immunological cross-reactivity it is possible that DRP antisera may recognize other members of the superfamily such as dystrophin or hitherto uncharacterized proteins such as the 400-kDa protein from Torpedo electrolytes (24,25) or those visualized at the myotendenous junctions in mdx mice (26).
The identification of DRP brings to mind a number of questions concerning the possible role of this protein during development and health and its presence or absence during the continuum of disease. Based solely upon the limited sequence similarity reported, a role for DRP in modulating dystrophic phenotypes has been suggested (19,20). Though it is tempting to speculate that DRP deficiency could lead to a myopathic phenotype, its apparent widespread tissue distribution would suggest that a disorder caused by abnormal expression of DRP might manifest by involving more tissues than usually involved in most inherited neuromuscular disorders. We are currently screening patient biopsies to study these questions. It has been asked if DRP might interact with dystrophin by forming a heterodimer (19) much like the homologous (Y-and P-spectrins do. Dystrophin and DRP are indeed similar in size and possibly structure, but we have demonstrated differences in tissue distribution of DRP and dystrophin. The 01-or P-spectrins which are normally found as heterodimers are usually co-expressed in mammalian tissues (10). However, the mismatch of tissue distribution does not exclude this possibility, since it has been recently reported that only fi-spectrins are detectable postsynaptically at rat myotubular neuromuscular junctions (21). In addition, in chicken it has been shown that tissue-specific P-spectrin isoforms associate with the same cy-spectrin in different tissues (10); a process similar to this may be occurring with respect to dystrophin and DRP. Clearly, more information is needed about the functions of dystrophin, DRP, and other members of the spectrin superfamily before conclusions can be drawn about their functional relationships.