Oligomerization of ß-Dystroglycan in Rabbit Diaphragm and Brain As Revealed by Chemical Crosslinking

Biochimica et Biophysica Acta 1370Ž. 1998 325±336

View metadata, citation and similar papers at core.ac.uk brought to you by CORE

provided by Elsevier - Publisher Connector Oligomerization of b-dystroglycan in rabbit diaphragm and brain as revealed by chemical crosslinking

Denise M. Finn, Kay Ohlendieck ) Department of Pharmacology, UniÕersity College Dublin, Belfield, Dublin 4, Ireland Received 24 November 1997; accepted 11 December 1997

Abstract

The surface component b-dystroglycan is a member of the dystrophin±glycoprotein complex providing a trans-sarcolemmal linkage between the actin membrane cytoskeleton and the extracellular matrix component laminin-a 2. Although abnormalities in this complex are involved in the pathophysiology of various neuromuscular disorders, little is known about the organization of dystrophin-associated glycoproteins in diaphragm and brain. We therefore investigated the oligomerization of b-dystroglycan and its connection with the most abundant dystrophin homologues in these two tissues. Employing detergent solubilization and alkaline extraction procedures of native membranes, it was confirmed that b-dystroglycan behaves like an integral surface molecule as predicted by its cDNA sequence. Immunoblot analysis following chemical crosslinking of native membranes showed that b-dystroglycan has a tendency to form high-molecular-mass complexes. Within these crosslinkable complexes, immuno-reactive overlaps were observed between b-dystroglycan, a-dystroglycan, laminin and 427 kDa dystrophin in diaphragm and skeletal muscle. In synaptosomes, the major brain dystrophin isoform Dp116 also exhibited an immuno-reactive overlap with members of the dystroglycan complex. These findings demonstrate that b-dystroglycan does not exist as a monomer in native membranes and imply that certain dystrophin isoforms and dystrophin-associated components interact with this surface protein in diaphragm and brain as has been previously shown for skeletal and heart muscle. q 1998 Elsevier Science B.V.

Keywords: Dystroglycan; Dystrophin; Dp-116; Dystrophin±glycoprotein complex; Laminin; Crosslinking

1. Introduction are proposed to provide a trans-sarcolemmal linkage between the basement membrane component laminin- Dystrophin, initially identified as the primary de- a 2 and the actin membrane cytoskeleton in skeletal r wx fect in Duchenne Becker muscular dystrophy 1 , is muscle fibreswx 6±9 . While peripheral a-dystroglycan now well established to be associated with a variety wx of apparent 156 kDa represents a novel non-integrin of surface proteins 2,3 and was shown to exist in class of laminin-binding proteins, 43 kDa b-dys- numerous isoforms in many tissues including brain wx troglycan is an integral dystrophin-associated glyco- 4,5 . Dystrophin-associated components belonging to proteinwx 10 . Both dystroglycan proteins are encoded the highly glycosylated surface dystroglycan-complex by a single gene whose expressed precursor protein product is post-translationally processed into two ) wx Corresponding author. Fax: q353-1-269-2749; E-mail: glycoproteins 11,12 . Molecular studies have shown [email protected] that the extreme COOH-terminus of b-dystroglycan

0005-2736r98r$19.00 q 1998 Elsevier Science B.V. All rights reserved. PII S0005-2736Ž. 97 00283-6 326 D.M. Finn, K. OhlendieckrBiochimica et Biophysica Acta 1370() 1998 325±336 constitutes a unique binding site for the second half trophin isoform Dp116 exhibits biochemical proper- of the hinge-4 region and the cysteine-rich domain of ties typical of a membrane cytoskeletal proteinwx 32 . dystrophinwx 13 and that a-dystroglycan can interact Thus, although Dp116 is lacking the N-terminal with the a 2-chain of laminin-2, formerly referred to actin-binding domain of 427 kDa dystrophin, it ap- as merosinwx 11,14,15 . Both proteins co-localize with pears to be a component of the sub-plasma membrane dystrophin to the muscle surfacewx 16±18 , clearly cytoskeleton possibly involved in anchoring proteins co-purify in subcellular fraction studies with the sar- in the periphery of neurons. Another member of the colemma fractionwx 16,19 and contain asparagine-lin- superfamily of spectrin-like components, autosoma- ked oligosaccharideswx 20 . Serinerthreonine-linked lly-encoded utrophin of 395 kDa, has an even broader oligosaccharides on a-sarcoglycan are proposed to distribution in brainwx 33 and may function in the confer a stiff conformation to the peptide core and generation and maintenance of regional substratum- protease resistance to this peripheral cell surface pro- associated membrane specialisations at the blood± teinwx 20 . Although the principal sub-plasmalemmal brain barrierwx 34 . Utrophin was originally described cytoskeleton link to the basement membrane is prob- as a neuromuscular junction-specific dystrophin iso- ably provided by these two tightly associated dystro- formwx 35±37 . Since the utrophin-associated compo- glycans, other sarcolemmal proteins might also be nent a-dystroglycan was found to be a functional involved in this process, i.e., components of the agrin-receptor, the utrophin±glycoprotein complex is sarcoglycan sub-complex ranging in relative molecu- implicated in the anchoring of the nicotinic acetyl- lar mass from apparent 25 kDa to 50 kDawx 2,3 . choline receptor at the neuromuscular junctionw 38± Following the detailed analysis of the spatial con- 40x . Although a- and g-sarcoglycan appear to be figuration of the muscle dystrophin±glycoprotein muscle-specific proteinswx 41 , other dystrophin-asso- complex, it was established that primary or secondary ciated proteins were clearly established to be present defects in the dystrophin±glycoprotein complex lead in neuronal tissueswx 42±45 . In neural retina, the to a variety of neuromuscular disorders, i.e., dystroglycan complex is expressed in the inner limit- DuchennerBecker muscular dystrophy, congenital ing membrane and around blood vessels where it muscular dystrophy and certain forms of limb-girdle co-localises with laminin and utrophinwx 43 . Both muscular dystrophywx 8,9,21±24 . Recently, b-dys- dystroglycans are also present in the outer plexiform troglycan was described as defective in a novel form layer of retina, which is devoid of laminin, but of muscular dystrophywx 25 . However, although a contains dystrophin and utrophinwx 43 . Studies by large number of patients afflicted with Duchenne Mummery et al.wx 42 showed that b-dystroglycan is muscular dystrophy suffer from respiratory complica- enriched in synaptic membranes of adult rat forebrain tionswx 26 , relatively little work has been done on the and that a novel b-dystroglycan-related protein of biochemical and pathophysiological characterization apparent 164 kDa is enriched in the postsynaptic of dystrophin-associated components in the di- density. In addition, analysis of dystrophin-associated aphragm. In addition, the association of brain dys- proteins in rat cerebellum by Tian et al.wx 44 support trophins with other neuronal surface components is the hypothesis that dystrophin interacts with dystro- not well understood and the normal function and glycans in Purkinje neurons. Binding studies with potential involvement of these isoforms in brain dis- purified brain a-dystroglycan established that it can orders is mostly unknown. Since one third of interact with a 2-chain-containing laminin-2wx 44 . In Duchenne patients suffer from non-progressive men- addition, syntrophins were recently shown to co- tal retardation, mutations in brain dystrophins may be localize with various dystrophin isoforms and dystro- the primary or contributing factor leading to these glycans in mouse brainwx 45 . clinical symptomswx 27±29 . Biochemical evidence for a tight association be- Brain dystrophin of 427 kDa has been localised to tween skeletal muscle surface glycoproteins and dys- the postsynaptic densitywx 27 and is present in a wide trophin isoforms exists from chromatographic tech- range of neurons including cerebellar Purkinje cells niques, analytical gradient centrifugation and domain and pyramidal neurons in the cerebral cortexwx 30,31 . binding studieswx 2,3,13,15,20 . Although the dys- Recently, we could show that the major brain dystrophin±glycoprotein complex can be isolated by D.M. Finn, K. OhlendieckrBiochimica et Biophysica Acta 1370() 1998 325±336 327 laminin affinity columns and labeled laminin binds nize the major brain isoform Dp116, were produced strongly to a-dystroglycan in overlay assayswx 11,15 , as previously described in detailwx 32 . Polyclonal the purified skeletal muscle dystrophin±glycoprotein antibody to crude laminin preparations from base- complex contains no detectable laminin-a 2 chain ment membranes of Englebreth±Holm±SwarmŽ. EHS wx2,3 . This loss in laminin-a 2 may be due to break- mouse sarcoma, monoclonal antibody VC1.1 to the age of weak non-covalent bonds during salt washes HNK-1 antigen, as well as monoclonal antibody SB- to remove actomyosin contaminations from vesicular SP-1 to human erythrocyte spectrin were purchased muscle membrane preparations. Thus, to more di- from Sigma Chemical CompanyŽ. Poole, Dorset, UK . rectly determine the oligomerization of the principal The laminin antibody, although produced against trans-sarcolemma spanning component of the dys- laminin-1Ž. previously named EHS-lamininwx 49 , rec- trophin±glycoprotein complex, b-dystroglycan, and ognizes both laminin-1 and laminin-2Ž previously its association with laminin isoforms and dystrophin, named merosin. wx 49 isoforms. Monoclonal antibody we performed a chemical crosslinking analysis. This NCL-43 to b-dystroglycan was from Novocastra is an extremely powerful technique for the determina- LaboratoriesŽ. Newcastle upon Tyne, UK , while tion of tight neighborhood relationships between monoclonal antibodies VIA41 to a-dystroglycan, wx membrane proteins 46 and this method was ex- XIXC2 to 427 kDa dystrophin and C464.6 to a1- ploited in this study to investigate potential protein± NaqrKq-ATPase were from Upstate Biotechnology protein interactions between key components of the Ž.Lake Placid, NY, USA . dystrophin±glycoprotein complex in diaphragm and brain. Crosslinking analysis was performed under 2.3. Membrane preparations optimized conditionswx 47 with the hydrophobic 12 AÊ crosslinker dithiobis-succinimidyl propionatewx 48 and Homogenates from diaphragm, hind leg muscle changes in the electrophoretic mobility of compo- and heart from New Zealand white rabbits were used nents of the diaphragm and brain dystrophin±glyco- to prepare crude microsomal membranes using stan- protein complex were determined using immuno- dard differential centrifugation techniqueswx 16,47 . blotting with highly specific antibodies to b-dys- Rabbit brain synaptosomes were prepared according troglycan, a-dystroglycan, laminin, dystrophin and to the method of Gordon-Weekswx 50 . In order to Dp116. minimize protein degradation, all isolation procedures were performed at 0±48C and all buffers contained 2. Experimental procedures the following protease inhibitors: 0.15 mM aprotinin, 0.3 mM E-64, 1 mM leupeptin, 0.2 mM pefabloc, 1.4 2.1. Materials mM pepstatin A, 0.5 mM soybean trypsin inhibitor, Dithiobis-succinimidyl propionate was obtained and 1 mM EDTA. Protein concentration was deter- wx from Pierce & WarrinerŽ. Chester, Chesire, UK . Per- mined by the method of Bradford 51 with rat my- oxidase-conjugated and fluorescein-labeled secondary ofibrillar proteins as a standard. antibodies to rabbit or mouse IgG, as well as chemicals for enhanced chemiluminescence detection and 2.4. Extraction procedures protease inhibitors were purchased from Boehringer MannheimŽ. Lewis, East Sussex, UK . Immobilon-NC To determine the basic biochemical properties of nitrocellulose was from MilliporeŽ Bedford, MA, b-dystroglycan, diaphragm and brain membranes USA. . All other chemicals were of analytical grade were extracted by detergent solubilization using 1% and purchased from Sigma Chemical Company Ž.vrv Triton X-100wx 19,52 and alkaline extraction Ž.Poole, Dorset, UK . was performed at pH 11 as described previously in wx 2.2. Antibodies detail 19,52 . Following solubilization, membranes were centrifuged for 20 min at 150,000=g and then Affinity-purified, polyclonal antibodies to the ex- separated fractions were analyzed by immuno- treme COOH-terminus of dystrophin, which recog- blotting. 328 D.M. Finn, K. OhlendieckrBiochimica et Biophysica Acta 1370() 1998 325±336

2.5. Chemical crosslinking troglycan and a polyclonal antibody to a peptide representing the last 15 amino acids of human skele- Following isolation, membranes were immediately tal muscle dystrophinwx 32 . Primary and secondary chemically crosslinked and then electrophoretically antibodies were used at a dilution of 1:100 and 1:150, analyzed. Membrane vesicles were incubated at room respectively. Preparation, blocking, incubation and temperature for 30 min with 10 to 200 mg crosslinker washing of cryosections, as well as photography was per mg protein at pH 8.0 as previously described in carried out by standard procedures as already de- detailwx 47 . To prepare stock solutions of the 12 AÊ scribed in detail elsewherewx 35 . probe dithiobis-succinimidyl propionatewx 48 , hydrophobic crosslinker was dissolved in dimethyl sulfoxide. The final dimethyl sulfoxide concentration in 3. Results r membrane suspensions did not exceed 4%Ž. v v. 3.1. Characterization of the dystrophin-associated m Crosslinking was terminated by the addition of 50 l glycoprotein b-dystroglycan of 1 M ammonium acetate per ml reaction medium and then an equal volume of sodium dodecyl sulfate- Since investigation of the trans-plasmalemmal containing sample bufferwx 53 was added prior to linkage between dystrophin and laminin-a 2 via the electrophoretic separation. dystroglycan complex up to now has been mainly limited to skeletal muscle fibers, we examined the 2.6. Gel electrophoresis and immunoblot analysis basic biochemical properties of b-dystroglycan in the diaphragm and the brain. Standard extraction proce- Electrophoretic separation of proteins was per- dures using non-ionic detergent or alkaline treatment formed for 2000 V h in a Hoefer SE-600 system confirmed that this dystrophin-associated glyco- ŽHoefer Scientific Instruments, San Francisco, CA, protein exhibits properties of a typical integral mem- USA. using sodium dodecyl sulfate-containing 4± brane protein also in these two tissues. As illustrated 16%Ž. wrv gradient polyacrylamide gelswx 53 . Since in Fig. 1a, incubation with Triton X-100 solubilized very large protein complexes did not properly enter b-dystroglycan from diaphragm microsomes, but al- the separating gel after chemical crosslinking, routine kaline extraction left it with the bilayer fraction. The gel electrophoretic analysis of crosslinked proteins same results were obtained with solubilization experi- was carried out in resolving gels which lacked a ments using rabbit brain membrane preparations. Im- stacking gel system. Titin, nebulin and myosin from muno-reactive bands representing the dystrophin-as- rat myofibrillar preparations served as high molecular sociated glycoprotein of 43 kDa were mostly detected mass markerswx 47 . Electrophoretically separated pro- in the detergent-extracted fraction and the synaptoso- teins were transferred to nitrocellulosewx 54 using a mal pellets following alkaline treatmentŽ. Fig. 1b . TE-52X transfer unit from Hoefer Scientific Instru- Extensive cellular localization studies of the dys- mentsŽ. San Francisco, CA, USA . Blocking and incu- trophin- and utrophin±glycoprotein complexes are bation of nitrocellulose sheets was performed by mostly restricted to skeletal muscles. Thus prior to established procedureswx 16,47 whereby primary and our crosslinking analysis of diaphragm membranes, secondary antibodies were used at a dilution of 1:1000 we verified the plasmalemmal localization of b-dys- and 1:3000, respectively. Immuno-reactive bands troglycan in this tissue. Indirect immunofluorescence were visualized using the enhanced chemilumines- microscopy clearly revealed that the antigen recog- cence kit from Boehringer MannheimŽ Lewis, East nized by monoclonal antibody NCL-43 is not only Sussex, UK. . localized to the sarcolemma of gastrocnemius fibers Ž.Fig. 2d , but is also almost exclusively found in the 2.7. Immunofluorescence microscopy cell periphery of rabbit diaphragmŽ. Fig. 2b . As already established for rabbit skeletal muscle, b-dys- Staining of 10 mm transverse cryosections from troglycan co-localized with the dystrophin of 427 rabbit diaphragm and gastrocnemius muscle was per- kDa in the plasma membrane of transversely cut formed with monoclonal antibody NCL-43 to b-dys- diaphragm cryosectionsŽ. Fig. 2a,b . Surface staining D.M. Finn, K. OhlendieckrBiochimica et Biophysica Acta 1370() 1998 325±336 329

of the diaphragm was specific, since control sections incubated with fluorescein-labeled secondary antibodies only, did not exhibit significant background stain- ingŽ. not shown . It has previously been shown that b-dystroglycan is associated with blood vessels in brainwx 34,44 . However, using monoclonal antibody NCL-43, we did not observe enough specific staining patterns in brain for proper interpretation of the subcellular localization of b-dystroglycan. Since immunofluorescence labeling of rabbit brain sections did not result in sufficient staining above background levels, this project was not further pursued.

b Fig. 1. Immunodetection of -dystroglycan from rabbit di- 3.2. Chemical crosslinking of b-dystroglycan from aphragm and brain following non-ionic detergent treatment or alkaline extraction. Shown are immunoblots of microsomal mem- diaphragm, skeletal muscle and heart branes isolated from rabbit diaphragmŽ. a and brain Ž. b ho- b mogenates stained with monoclonal antibody NCL-43 to -dys- The dystrophin±glycoprotein complex is of ex- troglycan. The majority of b-dystroglycan is found in the detergent-extracted supernatant fractionŽ. lane 3 and the non-alkaline tremely large size and therefore our oligomerization sensitive bilayer fractionŽ. lane 4 as compared to the untreated studies had to take into account the potentially very microsomal fractionŽ. lane 1 , the detergent-insoluble pellet Ž lane limited electrophoretic mobility of chemically 2.Ž. and the alkaline-extracted supernatant fraction lane 5 . This crosslinked membrane complexes. Initial studies relative insolubility in alkaline solutions and easy extraction with showed that conventional gel systems employing a non-ionic detergents confirms that b-dystroglycan behaves bio- chemically like a typical integral membrane protein, as predicted stacking gel did not reveal satisfactory results since by its cDNA sequencewx 11 . Arrows indicate the position of the dystrophin-associated components did not properly immunodecorated 43 kDa band of b-dystroglycan Ž.b-DG . enter the resolving gel andror were not sufficiently

Fig. 2. Distribution of b-dystroglycan in rabbit diaphragm and skeletal muscle. Indirect immunofluorescence labeling of transverse cryosections from rabbit diaphragmŽ. a, b and gastrocnemius muscle Ž. c, d was performed with primary antibodies to dystrophin Ž. a, c and b-dystroglycanŽ. b, d and fluorescein-conjugated secondary antibodies. Both primary antibodies stained in both tissues almost exclusively the cell periphery, while control experiments with secondary antibody only did not result in plasmalemma labeling. Bar equals 20 mm. 330 D.M. Finn, K. OhlendieckrBiochimica et Biophysica Acta 1370() 1998 325±336 transferred from the stacking gel system onto nitro- 3.3. Chemical crosslinking analysis of b-dystrogly- cellulose. However, since the unequivocal identifica- can, a-dystroglycan, dystrophin and laminin in tion of b-dystroglycan depended on immunoblotting diaphragm and skeletal muscle of electrophoretically separated protein complexes, a gel system lacking a stacking gel was tried out and Crosslinker to protein ratios were chosen so as to found to be superior in the detection of very high avoid artifacts of general protein clustering. At a molecular mass complexes. While increasing amounts concentration of 10 to 200 mg crosslinker per mg of crosslinker resulted, in both gel systemsŽ not membrane protein no general shifts to higher molecu- shown. , in a clearly detectable decrease in the b-dys- lar masses were observed. Coomassie-stained protein troglycan monomer of apparent 43 kDa, only the gels revealed only decreases in the relative elec- immunoblots from gels which lacked a stacking gel trophoretic mobility of distinct subsets of membrane exhibited high molecular mass complexes of this proteinsŽ. not shown . Incubation of diaphragm mem- glycoproteinŽ. Fig. 3a±c . Since the lack of linearity branes with 10 mg dithiobis-succinimidyl propionate between the relative electrophoretic mobility and the per mg protein resulted in a dramatic decrease in the molecular mass of very large protein bands prevented 43 kDa monomer of b-dystroglycan and in the con- a proper determination of the apparent size of the current appearance of a very high molecular mass oligomeric complexes, the relative molecular mass of protein band recognized by monoclonal antibody the crosslinked complexes could only be estimated to NCL-43Ž. Fig. 4a . Similar shifts to a greatly reduced be approximately 2000 kDa. Diaphragm, skeletal electrophoretic mobility were observed for peripheral muscle and heart microsomes exhibited a very com- a-dystroglycan of 156 kDa, the basement component parable shift of b-dystroglycan to a much reduced laminin and the membrane cytoskeletal protein dys- electrophoretic mobility following crosslinking with a trophin of 427 kDaŽ. Fig. 4b±d . All four key compo- hydrophobic 12 AÊ crosslinkerŽ. Fig. 3a±c . nents of the dystrophin±glycoprotein complex exhib-

Fig. 3. Chemical crosslinking of b-dystroglycan in microsomes from rabbit diaphragm, skeletal muscle and heart. Shown are immunoblots of crosslinked microsomes from rabbit diaphragmŽ. a , skeletal muscle Ž. b and heart Ž. c stained with monoclonal antibody NCL-43 to b-dystroglycan Ž.b-DG . Lanes 1 to 5 represent 0, 10, 50, 100 and 200 mg of crosslinker per mg membrane protein, respectively. Nitrocellulose transfers were obtained from non-reducing 4±16%Ž. wrv gradient gels run without a stacking gel system. The arrow indicates the position of the 43 kDa b-dystroglycan Ž.b-DG monomer, while an arrow head marks the position of high molecular mass complexes following crosslinking. Molecular mass markers Ž=10y3. are indicated on the left. D.M. Finn, K. OhlendieckrBiochimica et Biophysica Acta 1370() 1998 325±336 331

Monomers of a-dystroglycan, laminin and dystrophin were relatively weakly labeled by their re- spective antibodies, possibly due to their low abundance in the diaphragm membrane preparations. However, for an accurate crosslinking analysis it was necessary to perform our experiments with crude, non-salt washed microsomes so as to avoid potential artifacts introduced by excessive density gradient

Fig. 4. Chemical crosslinking of the dystroglycan complex in microsomes from rabbit diaphragm. Shown are immunoblots of crosslinked microsomes from rabbit diaphragm stained with monoclonal antibodies NCL-43 to b-dystroglycan Ž.Ž.b-DG a , VIA41 to a-dystroglycan Ž.Ž.a-DG b and XIXC2 to 427 kDa dystrophin Ž.Ž.DYS-427 d , as well as a polyclonal antibody to laminin Ž.Ž.LAM c . Lanes 1 to 5 represent 0, 10, 50, 100 and 200 mgof crosslinker per mg membrane protein, respectively. Nitrocellu- lose transfers were obtained from non-reducing 4±16%Ž. wrv gradient gels run without a stacking gel system. Arrows indicate the position of monomers, while arrow heads mark the position of high molecular mass complexes following crosslinking. Fig. 5. Chemical crosslinking of the dystroglycan complex in Molecular mass markers Ž=10y3. are indicated on the left. microsomes from rabbit skeletal muscle. Shown are immunoblots of crosslinked microsomes from rabbit skeletal muscle stained with monoclonal antibodies NCL-43 to b-dystroglycan Ž.b-DG ited a distinct overlap in immunoreactivity. Im- Ž.a Ža .Ž. a a , VIA41 to -dystroglycan -DG b and XIXC2 to 427 kDa munoblotting with antibodies to adhalin Ž - dystrophinŽ.Ž. DYS-427 d , as well as a polyclonal antibody to sarcoglycan. , a member of the sarcoglycan subcom- lamininŽ.Ž. LAM c . Lanes 1 to 5 represent 0, 10, 50, 100 and 200 plex, did not result in strong enough labeling for mg of crosslinker per mg membrane protein, respectively. Nitro- proper interpretation of oligomerization following cellulose transfers were obtained from non-reducing 4±16% r crosslinkingŽ. not shown . Increasing amounts of Ž.w v gradient gels run without a stacking gel system. Arrows indicate the position of monomers, while arrow heads mark the crosslinker caused no further decrease in elec- position of high molecular mass complexes following cross- trophoretic mobility of the members of the dys- linking. Molecular mass markers Ž=10y3. are indicated on the trophin±glycoprotein complex studiedŽ. Fig. 4a±d . left. 332 D.M. Finn, K. OhlendieckrBiochimica et Biophysica Acta 1370() 1998 325±336

3.4. Chemical crosslinking analysis of b-dystroglycan, a-dystroglycan, Dp116 and laminin in brain

In synaptosomes, b-dystroglycan of 43 kDa exhibited a distinct shift to very high molecular mass complexes after chemical crosslinking. A decrease in monomer staining and an increase in labeling of oligomeric complexes was synchronous in appear- anceŽ. Fig. 6a . Labeling of the 156 kDa monomer of a-dystroglycan was extremely weakŽ. not shown , but a clear overlap of immuno-reactivity was observed for both dystroglycans following crosslinkingŽ Fig. 6b. . A shift of a-dystroglycan to higher molecular mass complexes following crosslinking was indicated by immunoblotting with a monoclonal antibody which recognizes the HNK-1 epitopeŽ. Fig. 6c . The fact that a-dystroglycan contains the HNK-1 epitope was recently established by Yamada et al.wx 55 . However, other proteins such as the neuronal cell adhesion molecule might also be recognized by this monoclonal antibody. Antibodies to laminin showed sufficient staining of monomers and oligomers of this basement membrane component and the oligomeric band exhibited an overlap with a- and b-dystrogly- canŽ. Fig. 6d . With respect to brain dystrophin isoforms, Dp116 appears to be of relatively high abun- Fig. 6. Chemical crosslinking of the dystroglycan complex in dance in synaptosomal vesicles. Our affinity-purified microsomes from rabbit brain. Shown are immunoblots of antibody to the extreme COOH-terminus of dys- crosslinked synaptosomes from rabbit brain stained with mono- trophin recognized this protein of apparent 116 kDa b Ž.Ž.b clonal antibodies NCL-43 to -dystroglycan -DG a , VIA41 extremely wellŽ. Fig. 6e . Increasing amounts of a Ž.Ž.a to -dystroglycan -DG b and VC1.1 to the HNK-1 antigen chemical crosslinking of synaptosomes resulted in a Ž.Ž.HNK-1 c , as well as polyclonal antibodies to laminin Ž. LAM Ž.d and 116 kDa brain dystrophin Ž Dp116 .Ž. e . Lanes 1 to 5 marked decrease in the immunolabeling of the represent 0, 10, 50, 100 and 200 mg of crosslinker per mg monomer band and concurrently produced the ap- membrane protein, respectively. Nitrocellulose transfers were pearance of a high molecular mass oligomeric com- obtained from non-reducing 4±16%Ž. wrv gradient gels run plex. The Dp116 band with decreased electrophoretic without a stacking gel system. Arrows indicate the position of mobility showed a potential overlap with the im- monomers, while arrow heads mark the position of high molecular mass complexes following crosslinking. Molecular mass muno-reactive bands representing laminin and the markers Ž=10y3. are indicated on the left. dystroglycansŽ. Fig. 6a±d . Immunoblotting of crosslinked brain membranes with antibodies to 427 centrifugation or subcellular affinity purification. In kDa dystrophin and 395 kDa utrophin did not reveal skeletal muscle, very comparable results to those of satisfactory labelingŽ. not shown and the potential the diaphragm were obtained. Chemical crosslinking complex formation of this minor brain dystrophin caused a shift to higher relative molecular masses for isoform and its autosomal homologue with the dys- all members of the dystrophin±glycoprotein complex troglycan proteins could thus not be investigated. analyzedŽ. Fig. 5a±d . A definite overlap between Results from control experiments as illustrated in immunoreactive bands representing b-dystroglycan, Fig. 7 demonstrate the specificity of the immunologi- a-dystroglycan, laminin and dystrophin was estab- cal overlap between members of the dystrophin± lished. glycoprotein complexes in skeletal muscle, di- D.M. Finn, K. OhlendieckrBiochimica et Biophysica Acta 1370() 1998 325±336 333

tion in immunodecoration of the apparent 220 kDa bandŽ. Fig. 7d . Thus, brain spectrin does not appear to be extensively crosslinkable using 10 to 100 mg DSP per mg protein which clearly induces complex formation of dystrophin-associated glycoproteinsŽ Fig. 6. . The lack of high molecular mass bands containing spectrin could be due to the failure of very large spectrin complexes to be able to enter the gel system employed or possibly the epitopeŽ. s recognized by monoclonal antibody SB-SP-1 were destroyed during the crosslinking process. The fact that the immunological overlap between dystrophin and dystrophin- associated glycoproteins is specific is furthermore demonstrated by previously published findingswx 47 . Crosslinking of rabbit skeletal muscle membranes, under the same conditions as employed here, does not result in an immunoreactive overlap between dystrophin-associated proteins and various muscle membrane proteins present in crude microsomal fractions including the surface dihydropyridine receptor. Com-

plexes between ryanodine receptor tetramers and a1- subunits of the transverse tubular dihydropyridine receptor are larger in relative molecular masswx 47 than the dystrophin±glycoprotein complexes described here. With respect to brain membranes, stain- qr q ing of the a1-subunit of the Na K -ATPase showed Fig. 7. Chemical crosslinking of microsomes from rabbit skeletal a marked decrease of the 110 kDa monomer follow- muscle, diaphragm and brain. Shown are control immunoblots of ing chemical crosslinking using dithiobis-succi- crosslinked membranes from rabbit skeletal muscleŽ. a , di- nimidyl propionateŽ. Fig. 7c . However, as already aphragmŽ. b and brain Ž c, d . stained with monoclonal antibodies qqr observed for spectrin, no appearance of a high molec- SB-SP-1 to spectrinŽ. a, b, d and C464.6 to a1-Na K -ATPase Ž.c . Lanes 1 to 5 represent 0, 10, 50, 100 and 200 mgof ular mass band containing this ubiquitous surface crosslinker per mg membrane protein, respectively. Nitrocellu- membrane marker was observed in the region of lose transfers were obtained from non-reducing 4±16%Ž. wrv crosslinked b-dystroglycanŽ. Fig. 7c . Therefore, the gradient gels run without a stacking gel system. Arrows indicate results presented in Figs. 4±6 strongly suggest that y3 the position of monomers. Molecular mass markers Ž=10. are b-dystroglycan does not exist as a monomer in native indicated on the left. membranes and that the immunological overlap with other members of the dystrophin±glycoprotein com- aphragm and brain following chemical crosslinking plex is a specific result of chemical crosslinking of Ž.Figs. 4±6 . Immunodecoration of the 220 kDa surface membrane proteins which have a very close monomer of the membrane cytoskeletal component neighbourhood relationship. spectrin decreases with increasing amounts of crosslinker used, but no high molecular mass band is detectable in the region of the crosslinked b-dystroglycan protein band. While the crosslinking- 4. Discussion induced decrease in the relative abundance of spectrin monomers is very pronounced for skeletal muscle Although early symptoms of Duchenne muscular Ž.Fig. 7a and diaphragm Ž. Fig. 7b membranes, brain dystrophy include a decrease in strength of limb and synaptosomes exhibit a much less prominent reduc- torso muscles, the great majority of patients die of 334 D.M. Finn, K. OhlendieckrBiochimica et Biophysica Acta 1370() 1998 325±336 respiratory and cardiac failurewx 26 . In addition, cog- surface membrane system that links the sub-sarco- nitive impairments may be the primary manifestation lemmal membrane cytoskeleton and the extracellular of the disease process preceding muscular weakness basement membranewx 7±9 . Approximately the same wx26 . This lends central importance to studies into size dystrophin-associated complexes were found in dystrophin-associated glycoproteins in diaphragm and skeletal muscle and diaphragm preparations follow- brain with respect to understanding the overall patho- ing crosslinking, indicating that very similar spatial genesis of muscular dystrophieswx 8,9 . Considering configurations might be present in the two different the world-wide efforts to develop suitable therapeutic tissues. Therefore dystrophin may have the same strategies for the treatment of inherited diseases such function in diaphragm as in skeletal and heart muscle as muscular dystrophieswx 56 , it is important to take and disruption of its proper organization in muscular into account the complex nature of the pathophysio- dystrophy could be the initial step in the pathophysiological mechanisms occurring down-stream from the logical processes rendering all muscular tissues more initial genetic defectŽ. s . Efficient up-regulation of susceptible to necrosiswx 8,9 . dystrophin and its isoforms by myoblast or gene In contrast to skeletal muscle fibers, very little is transfer therapies in muscular dystrophy patients is known about the function of brain dystrophins and still elusivewx 57,58 . For gene therapy to be successful dystrophin-associated proteins and their fate during the pathophysiological problems in tissues such as muscular dystrophieswx 27±34,42±45 . The cross- the diaphragm and nervous system have also to be linking analysis presented here indicates that in addressed. In this respect, elucidation of the normal synaptosomes both a-dystroglycan of 156 kDa and cellular functions of diaphragm and brain dystrophins b-dystroglycan of 43 kDa form a complex. In anal- and its associated glycoproteins might shed light on ogy to the muscle dystrophin±glycoprotein complex, respiratory and neuronal abnormalities associated with laminin was also found to exist in a close neighbour- certain neuromuscular disorders. hood relationship to the dystroglycans. It is unclear Based on biochemical studies and its cDNA se- how and with which brain laminin isoform a-dys- quence, skeletal muscle b-dystroglycan is predicted troglycan and thus indirectly b-dystroglycan may to be a glycosylated transmembrane proteinw 2,10± interact in synaptosomes. Since Tian et al.wx 44 could 12,20x . Here, we could clearly show that diaphragm show that brain a-dystroglycan and a 2-chain-con- and brain b-dystroglycans also exhibit properties typ- taining laminin-2 can interact, we presume that both ical of integral membrane proteins. With respect to dystroglycans and the laminin-2 isoform are integral solubility in non-ionic detergent and resistance to parts of this proposed brain complex. However, stud- alkaline extraction procedures, diaphragm and brain ies by Mummery et al.wx 42 indicate that b-dys- b-dystroglycan behave in an exactly opposite way to troglycan, although present in synaptic membranes, is dystrophinwx 19 , utrophin wx 52 and the dystrophin not detectable in the postsynaptic density fraction of brain isoform Dp116wx 32 . These findings agree with rat forebrain. This would exclude a major link be- the current model of the dystrophin±glycoprotein tween 427 kDa brain dystrophin and b-dystroglycan complexwx 7±9 which assumes a membrane cyto- in the postsynaptic density. On the other hand, little skeletal function for dystrophin isoforms and a trans- is known about the subcellular localisation of the plasmalemma spanning role for b-dystroglycanwx 10 . other brain dystrophin isoform, Dp116, which was Binding of laminin-a 2 by peripheral a-dystroglycan investigated in this report. Dp116 might have a has been demonstrated by overlay assays, but highly broader synaptic distribution than the less abundant purified preparations of the dystrophin±glycoprotein 427 kDa dystrophin. complex from skeletal muscle fibers are essentially Since immuno-decoration for Dp116 overlapped devoid of the a 2-chain-containing laminin-2wx 2,20 . with that of the dystroglycans, these components However, here we could show by chemical cross- might form a complex within a synaptic region other linking of skeletal muscle microsomes that laminin- than the postsynaptic density. However, currently, it a 2 and the dystroglycans exist in a very close neigh- cannot be concluded that b-dystroglycan, a-dys- borhood relationship. This confirms the prediction troglycan, Dp116 and utrophin are all closely linked that 427 kDa dystrophin is an integral part of a at only one specific subcellular region in brain tis- D.M. Finn, K. OhlendieckrBiochimica et Biophysica Acta 1370() 1998 325±336 335 sues. It is however clear that Dp116, since it is teins are complexed at which synaptic regions re- lacking the N-terminal actin-binding domain of 427 mains to be determined. kDa skeletal muscle dystrophinwx 6 , cannot be involved in anchoring a cortical actin membrane cytoskeleton to the brain cell surface. A potential indirect Acknowledgements connection of Dp116 with the extracellular matrix of brain cells could have a similar function as proposed Research was funded by the Health Research for the muscular dystrophin±glycoprotein complex Board, Dublin and Muscular Dystrophy Ireland. We and the neuromuscular utrophin±glycoprotein com- wish to thank John Perryman for technical assistance. plex, namely stabilization of the surface membrane andror anchoring of peripheral receptor molecules wx8,9 . Our crosslinking studies of synaptosomes iso- References lated from brain tissue is in agreement with reports wx wx 1 A.P. Monaco, R.L. Neve, C. Coletti-Feener, C.J. Bertelson, by Yamada et al. 55,59 who described a complex D.M. Kurnit, L.M. Kunkel, Nature 323Ž. 1986 646±650. formation of Dp116 in peripheral nerve using laminin wx2 J.M. Ervasti, K. Ohlendieck, S.D. Kahl, M.G. Gaver, K.P. affinity chromatography. Thus, Dp116 appears to Campbell, Nature 345Ž. 1990 315±319. wx form a linkage with some of the glycoproteins origi- 3 M. Yoshida, A. Suzuki, H. Yamamoto, S. Noguchi, Y. Ž. nally described in skeletal muscle both in the central Mizuno, E. Ozawa, Eur. J. Biochem. 222 1994 1055±1061. wx4 D.J. Blake, J.M. Tinsley, K.E. Davies, Trends Cell Biol. 4 and peripheral nervous system. Recently, it was sug- Ž.1994 19±23. gested that this interaction between a-dystroglycan wx5 J.M. Tinsley, D.J. Blake, M. Pearce, A.E. Knight, J. and laminin is at least partially mediated by complex Kendrick-Jones, K.E. Davies, Curr. Opin. Genet. Dev. 3 carbohydrate moities on dystroglycanwx 60 . Ž.1993 484±490. wx The overall spatial configuration of dystrophin± 6 A.H. Ahn, L.M. Kunkel, Nat. Genet. 3Ž. 1993 283±291. wx7 J.M. Ervasti, K.P. Campbell, Curr. Opin. Cell Biol. 5Ž. 1993 glycoprotein complexes, based on extensive studies 82±87. into the structure of dystrophin and its many associ- wx8 K.P. Campbell, Cell 80Ž. 1995 675±679. ated surface componentswx 6±10 , is confirmed by the wx9 K. Ohlendieck, Eur. J. Cell Biol. 69Ž. 1996 1±10. chemical crosslinking data as presented in this report. wx10 M.D. Henry, K.P. Campbell, Curr. Opin. Cell Biol. 8Ž. 1996 625±631. In the periphery of fibers from skeletal muscle, heart wx and diaphragm, the plasmalemma-spanning dystro- 11 O. Ibraghimov-Beskrovnaya, J.M. Ervasti, C.J. Leveille, C.A. Slaughter, S.W. Sernett, K.P. Campbell, Nature 355 glycan complex is in close contact with sarcoglycans, Ž.1992 696±702. dystrophin and the cortical actin membrane cyto- wx12 O. Ibraghimov-Beskrovnaya, A. Milatovich, T. Ozcelik, B. skeleton, as well as laminin-a 2. Besides dystrophin, Yang, K. Koepnick, U. Franke, K.P. Campbell, Hum. Mol. a major cytoskeletal anchor in the neuromuscular Genet. 2Ž. 1993 1651±1657. wx junction appears to be utrophin which exists also in 13 D. Jung, B. Yang, J. Meyer, J.S. Chamberlain, K.P. Camp- bell, J. Biol. Chem. 270Ž. 1995 27305±27310. close association with the same set of surface glyco- wx wx 14 G. Dickson, A. Azad, G.E. Morris, H. Simon, M. Nour- proteins 61 . In contrast, sarcoglycans do not appear sadeghi, F.S. Walsh, J. Cell Sci. 103Ž. 1992 1223±1233. to be present in brain tissueswx 41 and possibly the wx15 J.M. Ervasti, K.P. Campbell, J. Cell Biol. 122Ž. 1993 809± major brain dystrophin isoform Dp116 is an integral 823. wx part of a surface membrane complex including 16 K. Ohlendieck, J.M. Ervasti, J.B. Snook, K.P. Campbell, J. Ž. laminin-a 2 and the two dystroglycans. Since di- Cell Biol. 112 1991 135±148. wx17 M.J. Cullen, J. Walsh, L.V.B. Nicholson, Acta Neuropathol. aphragm dystrophin of 427 kDa is missing in muscu- 87Ž. 1994 349±354. lar dystrophy patients, a disruption of the dystrogly- wx18 T.R. Helliwell, T.M. Nguyen, G.E. Morris, Neuromuscular can-complex might destabilize the diaphragm fiber Disord. 4Ž. 1994 101±113. periphery and cause an impaired respiratory function. wx19 K. Ohlendieck, K.P. Campbell, FEBS Lett. 283Ž. 1991 Whether abnormalities in various brain dystrophin 230±234. wx20 J.M. Ervasti, K.P. Campbell, Cell 66Ž. 1991 1121±1131. isoforms play a direct or indirect role in cognitive wx21 K. Ohlendieck, K. Matsumura, V.V. Ionasescu, J.A. Tow- impairment of Duchenne muscular dystrophy patients bin, E.P. Bosch, S.L. Weinstein, S.W. Sernett, K.P. Camp- and which dystrophins and dystrophin-associated pro- bell, Neurology 43Ž. 1993 795±800. 336 D.M. Finn, K. OhlendieckrBiochimica et Biophysica Acta 1370() 1998 325±336 wx22 S.L. Roberds, F. Leturcq, V. Allamand, F. Piccolo, M. wx40 M.A. Bowe, K.A. Deyst, J.D. Leszyk, J.R. Fallon, Neuron Jeanpierre, R.D. Anderson, L.E. Lim, J.C. Lee, F.M.S. 12Ž. 1994 1173±1180. Tome, N.B. Romero, M. Fardeau, J.S. Beckmann, J.C. wx41 H. Yamomoto, Y. Mizuno, N. Ikuya, M. Yoshida, E. Ozawa, Kaplan, K.P. Campbell, Cell 78Ž. 1994 625±633. J. Biochem. 115Ž. 1994 162±167. wx23 C.G. Bonnemann,È R. Modi, S. Noguchi, Y. Mizuno, M. wx42 R. Mummery, A. Sessay, F.A. Lai, P.W. Beesley, J. Neu- Yoshida, E. Gussoni, E.M. McNally, D.J. Duggan, C. An- rochem. 66Ž. 1996 2455±2459. gelini, E.P. Hoffman, E. Ozawa, L.M. Kunkel, Nat. Genet. wx43 F. Montanaro, S. Carbonetto, K.P. Campbell, M. Linden- 11Ž. 1995 266±272. baum, J. Neurosci. Res. 42Ž. 1995 528±538. wx24 S. Noguchi, E.M. McNally, K.B. Othmane, Y. Hagiwara, Y. wx44 M. Tian, C. Jacobson, S.H. Gee, K.P. Campbell, M. Jucker, Mizuno, M. Yoshida, H. Yamamoto, C.G. Bonnemann,È E. Eur. J. Neurosci. 8Ž. 1996 2739±2747. Gussoni, P.H. Denton, T. Kyriakides, L. Middelton, F. wx45 D.C. Gorecki, H. Abdulrazzak, K. Lukasiuk, E.A. Barnard, Hentai, M.B. Hamida, I. Nonaka, J.M. Vance, L.M. Kunkel, Eur. J. Neurosci. 9Ž. 1997 965±976. E. Ozawa, Science 270Ž. 1995 819±822. wx46 S.S. Wong, Chemistry of Protein Conjugation and wx25 M.A. Salih, Y. Sunada, M. Al-Nasser, C.O. Ozo, M.H. Crosslinking, CRC Press, Boca Raton, FL, 1991. Al-Turaiki, M. Akbar, K.P. Campbell, Ann. Neurol. 40 wx47 B.E. Murray, K. Ohlendieck, Biochem. J. 324Ž. 1997 689± Ž.1996 925±928. 696. wx26 A.E.H. Emery, Duchenne muscular dystrophy, Oxford wx48 A.G. Lomant, G. Fairbanks, J. Mol. Biol. 104Ž. 1976 243± Monographs on Medical Genetics No. 24, Oxford Univ. 261. Press, Oxford, 1993. wx49 S.K. Powell, H.K. Kleinman, Int. J. Biochem. Cell Biol. 28 wx27 T.W. Kim, K. Wu, J.L. Xu, I.B. Black, Proc. Natl. Acad. Ž.1997 401±414. Sci. U.S.A. 89Ž. 1992 11642±11644. wx50 P.R. Gordon-Weeks, in: A.J. Turner, H.S. BachelardŽ. Eds. , wx28 N. Bresolin, E. Castelli, G.P. Comi, G. Felisari, A. Bardoni, Neurochemistry, a Practical Approach, IRL Press, Oxford, D. Perani, F. Grassi, A. Turconi, F. Mazzuchelli, D. Gal- pp. 1±26. lotti, M. Moggio, A. Prelle, C. Ausenda, G. Fazio, G. wx51 M.M. Bradford, Anal. Biochem. 72Ž. 1976 248±254. Scarlato, Neuromuscular Disord. 4Ž. 1994 359±369. wx52 K. Ohlendieck, Biochim. Biophys. Acta 1283Ž. 1996 215± wx29 T.W. Kim, K. Wu, I.B. Black, Ann. Neurol. 38Ž. 1995 222. 446±449. wx53 U.K. Laemmli, Nature 227Ž. 1970 680±685. wx30 H.G.W. Lidov, T.J. Byers, S.C. Watkins, L.M. Kunkel, wx54 H.T. Towbin, T. Staehlin, J. Gordon, Proc. Natl. Acad. Sci. Nature 348Ž. 1990 725±728. U.S.A. 76Ž. 1979 4350±4354. wx31 H.G.W. Lidov, T.J. Byers, L.M. Kunkel, Neuroscience 54 wx55 H. Yamada, A. Chiba, T. Endo, A. Kobata, L.V. Anderson, Ž.1993 167±187. H. Hori, H. Fukuta-Ohi, I. Kanazawa, K.P. Campbell, T. wx32 D.M. Finn, K. Ohlendieck, Neurosci. Lett. 222Ž. 1997 25± Shimizu, K. Matsumura, J. Neurochem. 66Ž. 1996 1518± 28. 1524. wx33 T.S. Khurana, S.C. Watkins, L.M. Kunkel, J. Cell Biol. 119 wx56 C.G. Bonnemann,È E.M. McNally, L.M. Kunkel, Curr. Opin. Ž.1992 357±366. Pediatr. 8Ž. 1996 569±582. wx34 T.S. Khurana, L.M. Kunkel, A.D. Frederickson, S. Car- wx57 T.A. Partridge, K.E. Davies, Br. Med. Bull. 51Ž. 1995 bonetto, S.C. Watkins, J. Cell Sci. 108Ž. 1995 173±185. 123±137. wx35 K. Ohlendieck, J.M. Ervasti, K. Matsumura, S.D. Kahl, C.J. wx58 K. Inui, S. Okada, G. Dickson, Brain Dev. 18Ž. 1996 Leveille, K.P. Campbell, Neuron 7Ž. 1991 499±508. 357±361. wx36 T.M. Nguyen, J.M. Ellis, D.R. Love, K.E. Davies, K.C. wx59 H. Yamada, A.J. Denzer, H. Hori, T. Tanaka, L.V. Ander- Gatter, G.E. Morris, J. Cell Biol. 115Ž. 1991 1695±1700. son, S. Fujita, H. Fukuta-Ohi, T. Shimizu, M.A. Ruegg, K. wx37 F. Pons, N. Augier, J.O.C. Leger, A. Roberts, F.M.S. Tome, Matsumura, J. Biol. Chem. 271Ž. 1996 23418±23423. M. Fardeau, T. Voit, L.V.B. Nicholson, D. Mornet, J.J. wx60 A. Chiba, K. Matsumura, H. Yamada, T. Inazu, T. Shimizu, Leger, FEBS Lett. 282Ž. 1991 161±165. S. Kusunoki, I. Kanazawa, A. Kobata, T. Endo, J. Biol. wx38 S.H. Gee, F. Montanaro, M.H. Lindenbaum, S. Carbonetto, Chem. 272Ž. 1997 2156±2162. Cell 77Ž. 1994 675±686. wx61 K. Matsumura, J.M. Ervasti, K. Ohlendieck, S.D. Kahl, K.P. wx39 J.T. Campanelli, S.L. Roberds, K.P. Campbell, R.H. Campbell, Nature 360Ž. 1992 588±591. Scheller, Cell 77Ž. 1994 663±674.