Molecular organization at the glycoprotein-complex-binding site of dystrophin : Three dystrophin-associated proteins bind directly to the carboxy-terminal portion of dystrophin

Eur. J. Biochem. 220, 283-292 (1994) 0 FEBS 1994

Molecular organization at the glycoprotein-complex-binding site of dystrophin Three dystrophin-associated proteins bind directly to the carboxy-terminal portion of dystrophin

Atsushi SUZUKI, Mikiharu YOSHIDA, Kensuke HAYASHI, Yuji MIZUNO, Yasuko HAGIWARA and Eijiro OZAWA Department of Cell Biology, National Institute of Neuroscience, National Center of Neurology and Psychiatry, Tokyo, Japan

(Received September 13/November 30, 1993) - EJB 93 1393/3

Direct interaction between the C-terminal portion of dystrophin and dystrophin-associated proteins was investigated. The binding of dystrophin to each protein was reconstituted by overlaying bacterially expressed dystrophin fusion proteins onto the blot membranes to which dystrophin- associated proteins were transferred after separation by SDSPAGE with the following results. (a) Among the components of the glycoprotein complex which links dystrophin to the sarcolemma, a 43-kDa dystrophin-associated glycoprotein binds directly to dystrophin. Although at least one of the binding sites of this protein resides within the cysteine-rich domain of dystrophin, a contribution of additional amino acid residues within the first half of the C-terminal domain was also suggested for more secure binding. (b) Two other proteins also directly bind to dystrophin. Their binding sites are suggested to reside within the last half of the C-terminal domain which is alternatively spliced depending on the tissue type.

Previously, based on the enzyme digestion experiments, we showed that the binding site for the glycoprotein complex on dystrophin is present within the cysteine-rich domain and the first half of the C-terminal domain [Suzuki, A., Yoshida, M., Yamamoto, H. & Ozawa, E. (1992) FEBS Lett. 308, 154-1601. Here, we have extended this work and found that the region which is involved in interaction with the complex is widely extended to the entire length of this part of the molecule. On the basis of the present results, we propose a model of molecular architecture at the binding site for the complex on dystrophin.

Skeletal muscle dystrophin is a large spectrin-like membrane-skeletal protein with a molecular mass of 427 kDa [ l - 41. It is encoded on the Duchenne muscular dystrophy gene on the X chromosome and defects in it give rise to muscle wasting [5] . Ever since its cDNA was cloned in 1987 [6], extensive studies have been carried out to investigate how the lack of dystrophin results in the phenotype [7-91.

On the basis of amino acid sequence analysis, Kunkel and co-workers predicted that the dystrophin molecule can be divided into four distinct domains [l , 101 : (a) an N-terminal domain with sequence similarity to the actin-binding region of a-actinin and 8-spectrin ; (b) a large, central triple-helical domain; (c) a cysteine-rich domain, which is similar in sequence to the C-terminus of a-actinin and has two incomplete Ca2+-binding motifs, and (d) a C-terminal domain, which bears no apparent resemblance to any previously reported protein. Among these domains, the functional importance of the last two domains, located at the C-terminal portion of dystrophin, has been specifically suggested because the se-

Correspondence to A. Suzuki, Department of Cell Biology, National Institute of Neuroscience, NCNP, 4-1 -1 Ogawa Higashi, Kodaira, Tokyo, Japan 187

Abbreviations. GPC, glycoprotein complex ; DAP, dystrophin- associated protein.

Enzymes. Glutathione S-transferase (EC 2.5.1.18); calpain (EC 3.4.22.17); chymotrypsin (EC 3.4.21.1).

quences of these domains are highly conserved [ l l ] and genomic deletions in these domains are tightly linked to phenotypes causing severe muscle wasting [12- 141. Meanwhile, we demonstrated that 35 -38-kDa dystrophin fragments, generated by calpain digestion, which originate from the cysteine-rich domain and the first half of the C-terminal domain, bind to the sarcolemmal glycoprotein complex (GPC) [ 151. This provided the first direct evidence for the GPC-binding site of dystrophin and finally settled the problems of membrane anchoring of the C-terminus portion [ 16- 181. Interest- ingly, the GPC-binding region we recognized coincides well with the part of the molecule the deletion of which has been shown always to result in phenotypes causing severe muscle wasting [12]. Therefore, our finding suggested that the lack of interaction between dystrophin and GPC through this region is one of the essential causes of the disease.

GPC has been shown to consist of at least four integral glycoproteins, 50DAG(A2), 43 DAG(A3 a), A3 b, 35 DAG (A4) and one extracellular proteoglycan, 156DAG [19-221. However, the molecular organization of this complex in- teracting with dystrophin still remains the subject of specula- tion [22]. In addition, there is little information about the location of other dystrophin-associated proteins (DAPs), A0 and 59DAP(a- and P-Al). In this paper, we prepared a dystrophin fusion protein in Escherichia coli which contains the cysteine-rich domain and the entire C-terminal domain

284

and investigated its interaction with each DAP separated on a blot membrane. Monitoring the effect of C-terminal truncation of this fusion protein also yielded information about the region responsible for each binding. From this study, we found that, among the components of GPC, 43 DAG(A3 a) is the protein which directly binds to dystrophin, and at least one of its binding sites is confined to the cysteine-rich domain. Direct interactions between dystrophin and other DAPs, A0 and P-Al, around this GPC- binding site were also identified.

EXPERIMENTAL PROCEDURES

Construction of plasmids encoding dystrophin C-terminus and its mutants

A set of pGEX vectors (Pharmacia) was used to express various dystrophin sequences as E. coli fusion proteins. The correct colonies were verified by restriction mapping of the recombinant plasmids or immunoscreening with region-specific anti-dystrophin antibodies (see Fig. 1 a). The currently available human dystrophin cDNA clones, cDMD8 and cDMD9-14, encoding bp 7003-7879 and 7876-13957, respectively [61, were used to make the following constructs (Fig. 1 a). pGEX/DCR encoding dystrophin amino acid residues 2266 -2557 fused to glutathione S-transferase was made by ligating cDMD8, into the EcoRI site within pGEX-2T (Pharmacia). To construct pGEX/DCT685 encoding the C-terminal residues 3026- 3685, the excised Hind111 fragment of cDMD9-14 was ligated into the EcoRI site of pGEX-3X by using HindIIVEcoRI adaptors made by mixing EcoRUSmaI and HindIIIISmaI adaptors (Takara). The C-terminal deletion mutants, pGEXDCT442 and pGEX/DCT264, were constructed by cutting the pGEX/DCT685 at a con- venient XbaI site within the 3‘ untranslated region and at a 5‘ NheI site or an AccIII site, respectively. The resultant plasmids were blunt-ended and religated. These constructions encode dystrophin residues 3026- 3442 and 3026 - 3264, respectively.

Expression and purification of fusion proteins

Recombinant proteins were expressed and purified from the soluble fraction of cell lysates using the glutathione- Sepharose column essentially as described by Kennedy et al. [23], with slight modification. To minimize degradation of the product, Zon- mutant (ME8426) of E. coli was used as a host strain. Disruption of the bacteria was simply performed by sonication in a buffer containing 50mM Tris, 50mM NaCl, 1 mM EDTA, 1 mM dithiothreitol, 10 pg/ml leupeptin, 2 pg/ml aprotinin, 1 mM benzamidine and 0.5 mM phenylmethylsulfonyl fluoride, pH 8.0. An additional washing procedure of the resin with an ATP-containing buffer (3 mM ATP, 3 mM MgC12, 50 mM Tris/HCl, 1 mM dithiothreitol, 2 g / m l leupeptin, 0.5 pg/ml aprotinin, 1 mM benzamidine, and 0.1 mM phenylmethylsulfonyl fluoride, pH 7.6) was performed after loading cell lysate (see the results). The fusion proteins other than DCR (a control fusion protein, Fig. l a ) were further purified by gel filtration on Superose 12 (Pharmacia) equilibrated with a reconstitution buffer (20 mM Hepes, 150 mM NaC1, 2 mM MgC12, 1 mM dithiothreitol, 2 pg/ml leupeptin, 0.5 pg/ml aprotinin, 1 mM benzamidine, and 0.1 mM phenylmethylsulfonyl fluoride, pH 7.5).

The fusion protein, DCT685, was treated with N-ethylmaleimide as follows. Glutathione- Sepharose eluate of DCT685 was dialyzed against the dithiothreitol-depleted reconstitution buffer and incubated with 1 mM N-ethylmaleimide for 15 min on ice. The reaction was stopped by adding 1OmM glutathione, and the solution was subjected to gel filtration as described above. The concentrations of the fusion proteins were determined with a protein assay kit (Pierce) using bovine serum albumin as a standard.

Overlay binding assay of fusion proteins to dystrophin-associated proteins

The complex of dystrophin and its associated proteins purified from rabbit skeletal muscle, as described previously [20], was electrophoretically separated on 4 - 15 % SDSI polyacrylamide gels and transferred onto polyvinylidene difluoride membranes. The membranes were washed with phosphate-buffered saline (150 mM NaC1, 10 mM NaH,PO,/ NaOH, pH 7.5) with 0.1% Tween 20 to remove remaining SDS, and blocked overnight at 4°C in the same buffer containing 0.1% casein and 0.1% gelatin. Blots were incubated with purified fusion proteins (4-30 pg/ml) in the reconstitution buffer containing 0.1% Tween 20 for 2 h at room tem- perature with mild agitation. This incubation period was confirmed to be sufficient for > 10 pg/ml of each fusion protein to reach binding saturation (data not shown). After extensive washing with the phosphate-buffered saline containing 0.1 % Tween 20, the membranes were processed for immuno- blotting with antibodies to dystrophin.

Electrophoresis and immunoblot SDS/PAGE was performed as described by Laemmli

[24]. In some experiments, we used an automatic electrophoresis system, Phast System (Pharmacia), using precast mini- gels (PhastGel) for convenience and saving samples. Two- dimensional PAGE developed by O’Farrell [25] was also performed using the Phast System according to the manufac- turer’s instruction (Phast System Technical Note no.2). For the initial one-dimensional isoelectric focusing, PhastGel Dry IEF rehydrated with 8 M urea, 0.5% Triton X-100 and 7.5% Pharmalyte 3-10 was used. Samples were prepared as described previously [26]. Electrophoresis under nondenaturing conditions was performed at 4°C using Davis’ buffer system containing 0.1 % digitonin [15, 271. Separated proteins were stained by Coomassie blue or transferred electrophoretically onto polyvinylidene difluoride membranes using the buffer system described by Kyhse-Anderson [28]. Immuno- staining was performed as described previously, using Vecs- tain ABC kit or the Elite ABC kit (Vector Lab) to increase sensitivity.

Antibodies The anti-dystrophin antibodies used were as follows.

Polyclonal antibodies P23a, P30b, P31b, P33c and P34c, which were previously reported [15, 20, 291, were raised against synthetic polypeptides corresponding to the amino acid sequences of human dystrophin 2360-2409, 3085 - 3097, 3186-3200, 3373-3391 and 3495-3544, respectively. Monoclonal antibody Dy8/6C5, which was raised against a synthetic polypeptide consisting of the last 17 amino acids at the C-terminus of the human dystrophin sequence, was purchased from Novocastra Laboratories Ltd. Polyclonal an-

285

Fig. 1. Expression and purification of recombinant dystrophin fusion proteins. (a) The largest bar represents the carboxy-terminal half of the dystrophin molecule showing its domain organization predicted by Kunkel and co-workers. [l, 101. For convenience, the C-terminal domain is separated into two domains, the first half of the C-terminal domain (FHC) and the last half of the C-terminal domain (LHC). CYR is the cysteine-rich domain. A jagged line indicates the main alternative splicing site [38]. The restriction endonuclease sites used to construct truncated dystrophin expression vector and the amino acid scale are illustrated over the dystrophin diagram. Vertical lines under the diagram show the position of the cysteine-residues [l] and a hatched bar shows the position of 35-38-kDa fragments, produced by calpain digestion, containing GPC-binding site [15]. A set of smaller bars represents the constructed fusion proteins and black blocks above these show the positions of the peptides used to generate the antibodies in this study. Dy8/6C5 is abbreviated as DY8 and other antibodies are represented without the last small characters. (b-e) The left-hand sections show a purification (CBB = Coomassie blue staining) and the right-hand sections a characterization using the region-specific antibodies of each fusion protein: (b) DCR; (c) DCT685; (d) DCT442; (e) DCT264. Lane 1, glutathione-Sepharose-adsorbed fraction (no Mg-ATP washing); lane l’, the same fraction after 3 mM Mg-ATP washing; lane 2, 3 mM Mg-ATP washing eluate after concentration; lane 3, pooled fractions after gel filtration on Superose 12. The molecular mass standards are indicated on the left in kDa.

286

Fig.2. Specific bindings of DCT685 to several DAPs. (a) Dystrophin-DAP complex (1.4 pg) was separated on 10% SDSPAGE and transferred onto polyvinylidene difluoride membranes. After incubation with the fusion proteins (DCT685 or DCR), the blot was processed by immunostaining with region-specific anti-dystrophin antibodies. Lane 1, Coomassblue staining of a blot which is not processed by overlay; lanes 2-6, dose-dependence of the bindings of DCT685: 0 (lane 2 ) , 1 pg/ml (lane 3), 3 pg/ml (lane 4), 10 pg/ml (lane 5 ) and 20 pg/ml (lane 6); lanes 7 and 8, 20 pg/ml DCT685 was overlaid in the presence (lane 7) or the absence (lane 8) of 0.1% purified glutathione transferase; lane 9, 30 pg/ml DCR overlay for control. Used antibodies are as follows: lanes 2-6, Dy8/6C5; lanes 7 and 8, P31 b; lane 9, P23a. Positions of molecular mass standards (in kDa) and dystrophin-associated proteins (DAPs) are indicated. (b) Overlay binding assay of DCT685 onto dystrophin-DAP complex separated by two-dimensional electrophoresis. Top, the result of immunostaining with mixture of the polyclonal antibodies PAI, PA2 and PA3a. Bottom, the result of 10 pg/ml DCT685 overlay. Detection was carried out with the anti-dystrophin antibody DY8/6CS. Asterisks indicate the PA3 a-reactive 97-kDa protein which is discussed in the text referring to Fig. 3 b.

tibodies, PA1, PA2 and PA3a against 59DAP(a- and P-Al), 50DAG(A2) and 43 DAG(A3 a), respectively, were also previously described [26, 301. It should be noted that PA3a was shown to react with only 43DAG(A3a) but not with A3b [30]. In addition, PA1 reacted with both P-A1 and a-A1 and cross-reacted with A0 [26].

RESULTS Expression and purification of fusion proteins

All constructed fusion proteins were recovered from the soluble fraction of E. coli cell lysate by using the glutathione- affinity column. However, a fusion protein DCT685, which contains the cysteine-rich domain and the entire C-terminal domain, and its C-terminal deletion mutants, DCT442 and DCT264, were highly susceptible to endogenous bacterial proteinases (data not shown). To minimize this degradation, we used the lon- mutant of E. coli (ME8426) as the host strain and obtained certain amounts of intact product for each fusion protein (Fig. 1 b-e, lane 1’). The following gel filtration enabled us to purify these products almost as a single band (Fig. 1 b-e, lane 3). This procedure also revealed the physical state of each C-terminal product: the smallest fusion protein, DCT264, was produced in a monomeric state, whereas preparations of DCT685 and DCT442 contain aggregated material as well as monomeric one, this being more prominent for the latter. In the following assay, we used a monomer fraction for each product, although aggregated material could not be completely excluded in the case of

DCT442. Immunoblot analysis showed that each product was reasonably recognized using predictable region-specific anti- dystrophin antibodies at the predicted molecular masses (Fig. 1 b-e).

Milner et al. reported that the E. coli DnaK (product of dnaK gene and analog of heat-shock protein 70) contami- nated their preparation of dystrophin fusion proteins [31]. Since a similar contaminant was associated with all our products, we eliminated it by routinely subjecting all samples to an Mg-ATP washing procedure (lane 2).

Overlay binding assay of fusion proteins to DAPs To examine binding of dystrophin C-terminal portion to

each DAP, purified dystrophin-DAP complex was separated electrophoretically, transferred to polyvinylidene difluoride membranes and overlaid with the constructed fusion proteins. The fusion proteins bound to DAPs were detected by region- specific anti-dystrophin antibodies whose epitopes are shown in Fig. l a .

When low concentrations (0.01 -0.2 pM) of the fusion protein DCT685, which expresses the cysteine-rich domain and the entire C-terminal domain, were incubated with the membrane in the presence of 1 mM dithiothreitol, several bands other than dystrophin, corresponding to AO, 59DAP(Al) and 43DAG(A3a), were strongly reactive to anti-dystrophin antibodies (Fig. 2a). On the other hand, the members of GPC other than 43DAG, namely SODAG, A3b and 35 DAG, were not specifically stained. This staining pattern did not change even when a detergent in the reconsti-

287

Fig. 3. Region specificity of the DAP-binding activity of dystrophin fusion proteins. (a) Overlay binding assay was performed as described in the legend to Fig. 2, except that 4-15% gel was used to sharply delineate the staining on 97-kDa protein. Used fusion proteins and antibodies were as indicated. The concentration of each fusion protein overlaid was 30 pg/ml (= 300 nM) for DCT685, 4 pg/ml (= 50 nM) for DCT442 and 20 pg/ml (= 400 nM) for DCT264. The positions of molecular mass standards (in kDa) and dystrophin- associated proteins (DAPs) are indicated. (b) Identical blots of dystrophin-DAP complex separated on 10% SDS/polyacrylamide gel were processed by overlay binding assay with 5 pg/ml DCT685 (left) or by immunostaining with the anti-43DAG antibody, PA3a (right). In this case, the mixed DCT685, a mixture of aggregated and monomeric fractions of DCT685, was used to amplify the staining of the 97-kDa protein. Antibody DY8/6C5 was used to detect bound DCT685.

tution buffer was removed (data not shown) and did not depend on the antibodies used to detect DCT685 (typical data using antibody Dy8/6C5 and P31 b are shown in lanes 5 and 7). Analysis by two-dimensional PAGE revealed that it was P-A1 but not a-A1 to which DCT685 bound (Fig. 2b) [26].

For the following reasons, we concluded that the observed staining on AO, P-A1 and 43DAG(A3a) was not due to nonspecific binding of DCT685 but arose from specific interaction of dystrophin C-terminal residues to these DAPs. (a) The staining pattern is irrelevant to the amount of each DAP. (b) Purified glutathione transferase (0.1 %) did not compete with DCT685 for the binding, suggesting that binding is not nonspecific through glutathione transferase (Fig. 2a, compare lanes 7 and 8). (c) DCR, a control fusion protein which comprise the triple-helical domain sequence (Fig. 1 a), did not show specific bindings to any DAPs (lane 9).

As shown in Fig. 3a, the staining pattern on DAPs varied depending on the fusion proteins overlaid. When another fusion protein, DCT442, which lacks the last half of the C-terminal domain, was overlaid, the staining on A0 and P- A1 disappeared, while that of 43DAG remained. In addition, sharp staining at a molecular mass of 97'kDa became con- spicuous. As in the case of DCT685, these results did not depend on the antibodies used (compare lanes 4 and 5). 43DAG and the 97-kDa protein also reacted with the smallest fusion protein, DCT264, which lacks the whole C-terminal domain and shows no tendency to aggregate. In the case

of DCT685, the staining on 97-kDa protein was difficult to discern from the broad staining on AO, but could be distin- guished if amplified by mixing the aggregated fraction of DCT685 with a monomer fraction (Fig. 3 b). We further found that this 97-kDa protein cross-reacted with anti- 43DAG antibody (PA3 a) raised against its cytoplasmic sequence (Fig. 2b, Fig. 3b lane 2). Judging from the amount of its protein and the results of a competition assay (see below), we concluded that the staining on it also resulted from the specific binding of the C-terminal residues.

Therefore, the above results indicate that (a) the region containing amino acid residues 3026- 3264, mainly corresponding to the cysteine-rich domain, can bind to 43DAG and 97-kDa protein, and (b) for the bindings to A0 and P-A1, the sequence 3443-3685, corresponding to the last half of the C-terminal domain, is essential.

The region encompassing the cysteine-rich domain and the first half of the C-terminal domain (amino acid residues 3080- 3360) is rich in cysteine residues highly conserved among species (Fig. 1 a). Therefore, we examined the effect on the fusion protein DCT685 of N-ethylmaleimide treatment which should modify the cysteine residues specifically and thus disrupt the protein structure around them. As shown in Fig. 4, the treatment was found to reduce greatly the binding activity of DCT685 to 43DAG and the 97-kDa protein. On the other hand, the staining on A0 and P-A1 was not affected at all. We also tested the effect of Ca2+ concentration on the staining pattern and confirmed that the presence or absence

288

Fig. 4. The effects of N-ethylmaleimide treatment on DAP-binding activity of a dystrophin fusion protein. Mixed DCT685 (5 pg/ ml) untreated (lane 1) or treated (lane 2) with 1 mM N-ethylmaleimide (NEM) was overlaid on a blot membrane, as described in the legend to Fig. 2. Bound DCT685 was detected with antibody DY8/ 6C5. The result of immunostaining of an identical blot using the anti-43DAG antibody (PA3a) is also shown for comparison (lane 3 ) .

of 1 mM EGTA or the presence of 1 mM Caz+ in the reconstitution buffer did not change the results (data not shown).

As stated above, we could not remove aggregated material completely from the preparation of one of the fusion proteins, DCT442. This material increased the sensitivity of DCT442 to the antibodies, which made it difficult in our assay system to compare quantitatively the DAP-binding affinities among our fusion proteins. Thus, we compared their affinities qualitatively by a competition binding assay shown in Fig. 5. In this assay, antibody Dy8/6C5, which reacts with the extreme C-terminal portion of dystrophin, was used to detect selectively the binding of DCT685. As expected, coexisting DCT442 inhibited the binding of DCT685 to both 43DAG and the 97-kDa protein at a concentration of 0.5 pM, while it did not affect those to A0 and P-Al. On the other hand, the smallest fusion protein, DCT264, did not inhibit the binding of DCT685, even at a 20-fold molar excess of DCT685. This suggests that, although at least one of the binding sites for 43DAG and the 97-kDa protein is located in the region between residues 3026- 3264, additional amino acid residues between 3265 -3442 are required for sufficient binding activity.

Chymotrypsin digestion of dystrophin-DAP complex Previously, we demonstrated that when dystrophin-DAP

complex is digested with calpain, the first half of the C- terminal domain is excised into 35 -38-kDa fragments together with the cysteine-rich domain [15]. Fig. 6a shows that this is not specific for calpain but chymotrypsin also generates a similar 35-kDa fragment (lane$ 3-6). In addition, as previously reported in the case of calpain digestion [32], this fragment was stable only when it was bound to GPC, other- wise it was completely digested in the early stage of the reaction (lanes 7 and 8). These results indicate that not only the cysteine-rich domain but also the first half of the C-termi-

Fig. 5. A competition assay between fusion proteins for DAP- binding activity. The mixed DCT685 (5 pg/ml) was overlaid on a blot membrane in the absence (lane 1) or the presence (lanes 2-5) of excess amount of other fusion proteins. Coexisting fusion proteins are as follows: lanes 2 and 3, DCT442 (18 and 72 pg/ml, respectively); lane 4, DCT264 (50 p g h l ) ; lane 5 , DCR(5O pg/ml). Antibody DY8/6C5 was used to selectively detect the binding of DCT685. Positions of molecular mass standards (in kDa) and some dystrophin-associated proteins (DAPs) are indicated.

nal domain is protected from proteinase attack through interaction with GPC directly or indirectly. We also confirmed that 43 DAG shows relative resistance against chymotryptic/ calpain digestion (Fig. 6c, lanes 1 and 2) and, even after digestion, it remains associated with dystrophin fragments containing the GPC-binding site (Fig. 6b and c). These results are consistent with the idea that among the members of GPC, 43DAG is the protein which directly interacts with the gen- erally excised 35-kDa fragment of dystrophin.

DISCUSSION

The overlay binding assay technique has been success- fully applied to detect specific protein-protein interactions which are of physiological significance [33-351. In this study, by combining this with the technique of bacterial recombinant protein production, we not only found specific interactions between the C-terminal portion of dystrophin and several DAPs, but also obtained information about the sequence of dystrophin involved in each interaction. Based on the results obtained here, we have presented a model of molecular organization at the GPC-binding site of dystrophin (Fig. 7; see below).

Previously, we have shown that 35 -38-kDa fragments of dystrophin which originate from the cysteine-rich domain and the first half of the C-terminal domain remain bound to GPC after calpain digestion of dystrophin-DAP complex [15]. That was the first report which provided direct bio- chemical evidence about the GPC-binding site on dystrophin. Here, in turn, we have demonstrated that among the members of GPC, 43DAG(A3a) is the glycoprotein which directly binds to dystrophin. The region responsible for its binding

289

Fig. 6. Protection of 35-kDa dystrophin fragment from proteinase digestion through association with GPC. Dystrophin-DAP complex or isolated dystrophin free from DAPs was digested with chymotrypsin (1/100, by mass) for 10 min or with calpain (1/100, by mass) for 50 min at 25°C. The reaction conditions were as described previously [32]. (a) Lanes 1 and 2, SDSPAGE patterns of the chymotryptic digests of dystrophin-DAP complex (4 pg; lane 1) or isolated dystrophin (4 pg; lane 2) stained with Coomassie blue (CBB). Lanes 3-8, immunoblot analysis of chymotryptic digests of dystrophin-DAP complex (1 pg; lanes 3-6) or isolated dystrophin (1 pg; lanes 7 and 8). Used anti-dystrophin antibodies were as indicated. Lane 9, immunostaining pattern of calpain digest of dystrophin-DAP complex stained with P33c is shown for comparison. (b) Analysis of calpain (lanes 1-3) or chymotryptic (lanes 4-6) digests of dystrophin-DAP complex by 3.5-6% nondenaturing PAGE. Lanes 1 and 4, silver staining; lanes 2 and 5 , immunostaining with anti-43DAG antibody (PA3a); lanes 3 and 6, immunostaining with P33c. (c) Two-dimensional PAGE analysis of the proteinase-resistant complexes. The calpain (top) or chymotryptic (bottom) digest separated by 3.5 -6% nondenaturing PAGE were laterally developed by 12.5% SDSPAGE. Shown are the results of the immunoblot analysis stained with mixture of the anti-dystrophin and anti-43DAG antibodies (P33c and PA3 a, respectively). On the left side are shown PA3a-staining patterns of dystrophin-DAP complex (lane 1) and the digest (lane 2) not subjected to the initial electrophoresis. Small arrows indicate the positions of dystrophin-derived fragments containing GPC-binding site, which were identified as PA3a-nonreactive bands in lane 2.

confirmed here (residues 3026-3442; see below) shows good agreement with the previous work, supporting the con- clusion that this binding is specific and relevant to the native linkage of dystrophin. There still remains a possibility of direct interactions between dystrophin and other members of GPC which were not detected under the present experimental

conditions. However, considering that 43DAG is encoded by a single mRNA together with the extracellular laminin-binding proteoglycan, 156DAG [22, 351, it is very plausible that 43DAG serves as an essential link between dystrophin and the extracellular matrix by binding to both dystrophin and 156DAG. This is also consistent with the recent immunohis-

290

Fig. 7. A schematic illustration showing a model of molecular organization at the glycoprotein-complex-binding site of dystrophin. The binding sites of P-A1 and A0 are tentatively localized at the last half of the C-terminal domain on the basis of the result of N- ethylmaleimide treatment (see text). The location of a-Al-binding site is not determined. FHC, LHC and CYR represent parts of the dystrophin molecule as defined in Fig. 1 .

tochemical studies on other kinds of muscular dystrophy, which are considered to arise from a deficiency of DAPs but not of dystrophin [36, 371. In these dystrophic skeletal muscles, 43DAG and dystrophin are preserved almost at normal level despite the loss of 50DAG and the great reduction of 35DAG.

In the previous work, we were uncertain that 43 DAG (A3 a) persisted in the calpain-resistant components of GPC because we did not have anti-43DAG antibody. However, we have now confirmed its relative resistance to proteinase digestion and established the consistency between our previous work and the above idea that 43DAG is the protein which directly mediates the binding between dystrophin and GPC.

By monitoring the effect of C-terminal truncation of the fusion proteins, a somewhat complicated feature of the interaction between dystrophin and 43 DAG was suggested. Since the smallest fusion protein, DCT264, which lacks the whole C-terminal domain, could bind to 43DAG, at least one of the 43 DAG-binding sites should reside within the cysteine-rich domain. However, the competition assay between the constructed fusion proteins also suggested that additional amino acid residues between 3265 - 3442, which roughly correspond to the first half of the C-terminal domain, contrib- ute to a more secure binding to 43DAG. To reveal the nature of this contribution of the first half of the C-terminal domain, more systematic study is needed. However, because we ini- tially observed that a fusion protein expressing residues 3265 - 3685 which corresponds to the entire C-terminal domain did not bind to 43DAG (Suzuki et al. unpublished results), we think it implausible that there is a second 43DAG-binding site in this region. We prefer the idea that this region plays a structural role in helping the cysteine-rich domain into a conformation for secure 43DAG binding. In fact, we extended our previous work on the limited digestion of dystrophin-DAP complex [32] and-demonstrated that not only the cysteine-rich domain but also the first half of the C- terminal domain is closely involved in interaction with GPC.

The good agreement of these results obtained from two independent experiments leads us to propose the model of dystrophin-GPC interaction shown in Fig. 7. In this model, the region (amino acid residues 3080- 3408) encompassing

the cysteine-rich domain and the first half of the C-terminal domain, which contains highly conserved cysteine residues with high frequency, forms a functional unit to interact with GPC, and this interaction is essentially mediated by 43DAG (Fig. 7). This idea, that the cysteine-rich domain and the first half of the C-terminal domain function as a single unit to interact with GPC, is also supported by the following fea- tures of this region. (a) The sequence (3123-3418) approxi- mately corresponding to this region, shows an extremely high degree of amino acid conservation during evolution (99% identity between chicken and human) [ l l ] . (b) In contrast to the various splicing patterns in the last half of the C-terminal domain (3409-3685), the first half of the C-terminal domain is commonly shared, together with the cysteine-rich domain, by all dystrophin transcripts so far detected except for the smooth-muscle-type isoform [38 -401. (c) As previously pointed out, this region is coincident with the part of the molecule which, when missing, necessarily leads to phenotypes for severe forms of the disease [12-141.

This extended, highly conserved, binding region implies a highly sophisticated interaction between dystrophin and GPC which is structurally controlled. In fact, we observed that the 43DAG-binding activity of the fusion protein, DCT685, is very unstable and tends to be damaged rapidly (data not shown). As we demonstrated by N-ethylmaleimide treatment (Fig. 4), the cysteine residues concentrated in this region must play a key role in this interaction.

Recently, Matsumura et al. showed that, not only the components of GPC, but also 59DAP is drastically reduced in patients where the C-terminal portion is absent, probably due to the lack of interaction with dystrophin [41]. In the present work, we directly demonstrated that A0 and P-A1 (one of the components which were named 59DAP by Camp- bell’s group) bind to the C-terminal portion of dystrophin and that their binding site is distinct from the GPC-binding site. Considering the absence of the effects of N-ethylmaleimide treatment, the binding site for A0 and P-A1 might be far from the cysteine-rich region (3080-3360) and reside within the last half of the C-terminal domain. Although we cannot define it at this stage, it is very interesting that the binding of these proteins to dystrophin requires the last half of the C-terminal domain which is alternatively spliced de-

291

pending on tissue types and developmental stages [38-401. This suggests that these DAPs may be responsible for modification of the dystrophin function.

Finally, we observed an extremely minor component in dystrophin-DAP complex at a molecular mass of 97 kDa. It cross-reacted with anti-43 DAG antibody and behaved the same as 43DAG with regard to dystrophin binding. This component was reproducibly observed in SDSPAGE even after conventional sample denaturation under reducing conditions. Therefore, we are of the opinion that this is not an artefact resulting from dimerization of 43DAG but is an im- munochemically related protein of 43DAG which forms a minor subpopulation of the dystrophin-DAP complex. On the other hand, it should be noted that A3b, which migrates very closely to 43DAG in one-dimensional SDSPAGE, did not show dystrophin-binding activity at all. Campbell’s group does not distinguish this protein from 43DAG [35] but, since we have shown it is not a degradation product of 43DAG [30], our results suggest that A3b is distinct from 43DAG.

We thank Dr A. Nishimura (National Institute of Genetics) for providing us with lon- mutant of E. coli (ME8426). This work was supported by a grant [2-I] from the National Center of Neurology and Psychiatry of the Ministry of Health and Welfare, Japan.

REFERENCES 1 . Koenig, M., Monaco, A. P. & Kunkel, L. M. (1988) The com-

plete sequence of dystrophin predicts a rod-shaped cytoskele- tal protein, Cell 53, 219-228.

2. Arahata, K., Ishiura, S., Ishiguro, T., Tsukahara, T., Suhara, Y., Eguchi, C., Ishihara, T., Nonaka, I., Ozawa, E. & Sugita, H. (1988) Immunostaining of skeletal and cardiac muscle surface membrane with antibody against Duchenne muscular dystrophy peptide, Nature 333, 861 -863.

3. Watkins, S. C., Hoffman, H. P., Slayter, H. S. & Kunkel, L. M. (1 988) Immunoelectron microscopic localization of dystrophin in myofibres, Nature 333, 863 -866.

4. Zubrzycka-Gaarn, E. E., Bulman, D. E., Karpati, G., Burghes, A. H. M., Belfall, B., Klamut, H. J., Talbot, J., Hodges, R. S., Ray, P. N. & Worton, R. G. (1988) The Duchenne muscular dystrophy gene product is localized in sarcolemma of human skeletal muscle, Nature 333, 466-469.

5. Moser, H. (1984) Duchenne muscular dystrophy: pathogenetic aspects and genetic prevention, Hum. Genet. 66, 17-40.

6. Koenig, M., Hoffman, E. P., Bertelson, C. J., Monaco, A. P., Feener, C. & Kunkel, L. M. (1987) Complete cloning of the Duchenne muscular dystrophy (DMD) cDNA and preliminary genomic organization of the DMD gene in normal and affected individuals, Cell 50, 509-517.

7. Franco, Jr A. & Lansman, J. B. (1990) Calcium entry through stretch-ion channels in mdx myotubes, Nature 344, 670-673.

8. Menke, A. & Jockusch, H. (1991) Decreased osmotic stability of dystrophin-less muscle cells from the mdx mouse, Nature 349, 69-71.

9. Petrof, B. J., Shrager, J. B., Stedman, H. H., Kelly, A. M. & Sweeney. H. L. (1993) Dystrophin protects the sarcolemma from stresses developed during muscle contraction, Proc. Natl Acad. Sci. USA 90, 3710-3714.

10. Hoffman, E. P. & Kunkel, L. M. (1989) Dystrophin abnormali- ties in Duchennemecker muscular dystrophy, Neuron 2, 1019- 1029.

11. Lamaire, C., Heilig, R. & Mandel, J. L. 41988) The chicken dystrophin cDNA: striking conservation of the C-terminal coding and 3’ untranslated regions between man and chicken,

12. Koenig, M., Beggs, A. H., Moyer, M., Scherpf, S., Heindrich, K., Bettecken, T., Meng, G., Muller, C. R., Lindlof, M., Kaari- ainen, H., de la Chapelle, A,, Kiuru, A., Savontaus, M-L.,

EMBO J. 7,4157-4162.

Gilgenkrantz, H., Recan, D., Chelly, J., Kaplan, J-C., Covone, A. E., Archidiacono, N., Romeo, G., Liechti-Gallati, S., Schneider, V., Braga, S., Moser, H., Darras, B. T., Murphy, P., Francke, U., Chen, J. D., Morgan, G., Denton, M., Greenberg, C. R., Wrogemann, K., Blonden, L. A. J., van Paassen, H. M. B., van Ommen, G. J. B. & Kunkel, L. M. (1989) The molecular basis for Duchenne versus Becker muscular dystrophy: correlation of severity with type of deletion, Am. J. Hum. Genet. 4.5, 498-506.

13. Roberts, R. G., Bobrow, M. & Bentley, D. R. (1992) Point mut- ations in the dystrophin gene, Proc. Natl Acad. Sci. USA 89,

14. Bies, R. D., Caskey, C. T. & Fenwick, R. (1992) An intact cysteine-rich domain is required for dystrophin function, J. Clin. Invest. 90, 666-672.

15. Suzuki, A., Yoshida, M., Yamamoto, H. & Ozawa. E. (1992) Glycoprotein-binding site of dystrophin is confined to the cysteine-rich domain and the first half of the carboxy-terminal domain, FEBS Lett. 308, 154-160.

16. Hoffman, E. P., Garcia, C. A., Chamberlain, J. S., Angelini, C., Lupski, J. R. & Fenwick, R. (1991) Is the carboxyl-terminus of dystrophin required for membrane association? A novel, severe case of Duchenne muscular dystrophy, Ann. Neurol. 30, 605-610.

17. Recan, D., Chafey, P., Leturcq, F., Hugnot, J.-P., Vincent, N.. Tome, F., Collin, H., Simon, D., Czernichow, P., Nicholson, L. V. B., Fardeau, M., Kaplan, J.-C. & Chelly, J. (1992) Are cysteine-rich and COOH-terminal domains of dystrophin crit- ical for sarcolemmal localization? J. Clin. Invest. 89, 712- 716.

18. Helliwell, T. R., Ellis, J. M., Mountford, R. C., Appleton, R. E. & Moms, G. E. (1992) A truncated dystrophin lacking the C-terminal domain is localized at the muscle membrane, Am. J. Hum. Genet. 50, 508-514.

19. Campbell, K. P. & Kahl, S. D. (1989) Association of dystrophin and an integral membrane glycoprotein, Nature 338, 259- 262.

20. Yoshida, M. & Ozawa, E. (1990) Glycoprotein complex anchoring dystrophin to sarcolemma, J. Biochem. (Tokyo) 108,748- 752.

21. Ervasti, J. M., Ohlendieck, K., Kahl, S. D., Gaver, M. G. & Campbell, K. P. (1990) Deficiency of a glycoprotein component of the dystrophin complex in dystrophic muscle, Nature

22. Ervasti, J. M. & Campbell, K. P. (1991) Membrane organization of the dystrophin-glycoprotein complex, Cell 66, 1121 - 1 1 31.

23. Kennedy, S. P., Warren, S. L., Forget, B. G. & Morrow, J. S. (1991) Ankyrin binds to the 15th repetitive unit of erythroid and nonerythroid P-spectrin, J. Cell. Biol. 115, 267-277.

24. Laemmli, U. K. (1970) Cleavage of structural proteins during the assembly of the head of bacteriophage T4, Nature 227, 680-685.

25. O’Farrell, P. H. (1975) High resolution two-dimensional electrophoresis of proteins, J. Biol. Chem. 250, 4007-4027.

26. Yamamoto, H., Hagiwara, Y., Mizuno, Y., Yoshida, M. & Ozawa, E. (1993) Heterogeneity of dystrophin-associated proteins, J. Biochem. (Tokyo) 114, 132-139.

27. Davis. B. (1964) Disc electrophoresis-11. Method and applicat- ion to human serum proteins, Ann. NY Acad. Sci. 121, 404- 427.

28. Kyhse-Anderson, J. (1984) Electroblotting of multiple gels: a simple apparatus without buffer tank for rapid transfer of proteins from polyacrylamide to nitrocellulose, J. Biochem. Bio- phys. Methods 10, 203-209.

29. Tanaka, H., Yoshida, M., Ishiguro, T., Eguchi, C., Nonaka, I. & Ozawa, E. (1989) Expression of dystrophin of the cell surface membrane of intrafusal fibers of human skeletal muscle, Pro- toplasma 152, 109-I l l .

30. Yoshida, M., Mizuno, Y., Nonaka, I. & Ozawa, E. (1993) A dystrophin-associated glycoprotein, A3 a (one of 43DAG doublets), is retained in Duchenne muscular dystrophy muscle, J. Biochem. (Tokyo) 114, 634-639.

2331 -2335.

345, 315-319.

292

31. Milner, R. E., Busaan, J. & Michalak, M. (1992) Isolation and characterization of different C-terminal fragments of dystrophin expressed in E. coli, Biochem. J. 288, 1037-1044.

32. Yoshida, M., Suzuki, A., Shimizu, T. & Ozawa, E. (1992) Proteinase-sensitive sites on isolated rabbit dystrophin, J. Biochem. (Tokyo) 112, 433-439.

33. Smalheiser, N. R. & Schwartz, N. B. (1987) Cranin: a laminin- binding protein of cell membranes, Proc. Nut1 Acad. Sci. USA 84, 6457 -6461.

34. Jarrett, H. W. & Madhavan, R. (1991) Calmodulin-binding proteins also have a calmodulin-like binding site within their structure, J. Biol. Chem. 266, 362-371.

35. Ibraghimov-Beskrovnaya, O., Ervasti, J. M., Leveille, C. J., Slaughter, C. A, Sernett, S. W. & Campbell, K. P. (1992) Primary structure of dystrophin-associated glycoproteins link- ing dystrophin to the extracellular matrix, Nature 355, 696- 702.

36. Matsumura, K., Tome, F. M. S., Collin, H., Azibi, K., Chaouch, M., Kaplan, J-C., Fardeau, M. & Campbell, K. P. (1992) Defi- ciency of the 50 K dystrophin-associated glycoprotein in severe childhood autosomal recessive muscular dystrophy, Nature 359, 320-322.

37. Yamanouchi, Y., Mizuno, Y., Yamamoto, H., Takemitsu, M., Yoshida, M., Nonaka, I. & Ozawa, E. (1994) Selective defect

in dystrophin-associated glycoproteins 5ODAG (A2) and 35DAG (A4) in dystrophic hamster: an animal model for severe childhood autosomal recessive muscular dystrophy (SCARMD), Neuromuscular Disorders, in the press.

38. Feener, C. A., Koenig, M., & Kunkel, L. M. (1989) Alternative splicing of human dystrophin mRNA generates isoforms at the carboxy terminus, Nature 338, 509-511.

39. Bies, R. D., Phelps, S. F., Cortez, M. D., Roberts, R., Caskey, C. T. & Chamberlain, J. S. (1992) Human and murine dystrophin mRNA transcripts are differentially expressed during skeletal muscle, heart, and brain development, Nucleic Acids Res. 20, 1725 - 1731.

40. Tinsley, J. M., Blake, D. J. & Davies, K. E. (1993) Apo- dystrophin-3 : a 2.2 kb transcript from the DMD locus encoding the dystrophin glycoprotein binding site, Hum. Mol. Genet. 2, 521 -524.

41. Matsumura, K., Tome, F. M. S., Ionasescu, V., Ervasti, J. M., Anderson, R. D., Romero, N. B., Simon, D., Recan, D., Kaplan, J.-C., Fardeau, M. & Campbell, K. P. (1993) Defi- ciency of dystrophin-associated proteins in Duchenne muscular dystrophy patients lacking COOH-terminal domains of dystrophin, J. Clin. Invest. 92, 866-871.

Documents

Molecular organization at the glycoprotein-complex-binding site of dystrophin : Three dystrophin-associated proteins bind directly to the carboxy-terminal portion of dystrophin