'Howard Hughes Medical Institute and Department of Physiologyand Biophysics, The University of Iowa College ojMedicine, Iowa City, Iowa 52242, USA 2FondationJean- Dausset-CEPH, 7501 0 Paris, France 3Git~&thon, 91 000 Evry, France University of Texas, Howard Hughes Medical Institute, Biopolymer Facility, Dallas, Texas 75235, USA jINSERM U153, 75005 Paris, France 6 ~ e n r y Ford Hospital, Division of Clinical and Molecular Genetics, C.F.P. 407, Detroit, Michigan 48202, USA Correspondence should be addressed to K.P.C. and J.S.B. +Thefirst three authors contributed equally to this work. P-sarcoglycan: characterization and role in limb-girdle muscular dystrophy linked to 4q12 Leland E. Liml*, Franck ~ u c l o s ' " , Odile Broux3*, Nathalie our^^, Yoshihide Sunadal, Valerie Allamand3, Jon ~ e ~ e r l , Isabelle ~ i c h a r d ~ , Carolyn Moomaw4, Clive Slaughter4, Fernando M.S. Tome5, Michel Fardeau5, Charles E. Jackson6, Jacques S. ~ e c k m a n n ~ > ~ & Kevin P. Campbell1 P-sarcoglycan, a 43 kDa dystrophin-associated glycoprotein, is an integral component of the dystrophin-glycoprotein complex. We have cloned human P-sarcoglycan cDlVA and mapped the P-sarcoglycan gene to chromosome 4q12. Pericentromeric markers and an intragenic polymorphic CA repeat cosegregated perfectly with autosomal recessive limb- girdle muscular dystrophy in several Amish families. A Thr-to-Arg missense mutation was identified within the P-sarcoglycan gene that leads to a dramatically reduced expression of P-sarcoglycan in the sarcolemma and a concomitant loss of adhalin and 35 DAG, which may represent a disruption of a functional subcomplex within the dystrophin- glycoprotein complex. Thus, the P-sarcoglycan gene is the fifth locus identified (LGMD2E) that is involved in autosomal recessive limb-girdle muscular dystrophy. The dystrophin-glycoprotein complex (DGC)1-4 is a large oligomeric complex of sarcolemmal proteins and glycoproteins, consisting of dystrophin, a large F-actin binding intracellular syntrophin, a 59 kDa intracellular proteins,9; adhalin, a 50 kDa transmem- brane glycoproteinlO; a 43 kDa transmembrane glyco- protein doublet (P-dystroglycan and A3b)2,3,"; a 35 kDa transmembrane glycoprotein; a 25 kDa trans- membrane protein; and a-dystroglycan, a large extra- cellular laminin-binding gly~oprotein"-'~. Together, the dystrophin-glycoprotein complex acts as a struc- tural link between the cytoskeleton and the extracellu- lar matrix, and is believed to confer stability to the sarcolemma and protect muscle cells from contrac- tion-induced damage and necrosis15. The DGC has been implicated in several forms of muscular dystrophy. In Duchenne muscular dystrophy (DMD), mutations in the dystrophin gene cause the complete absence of dystrophin and a dramatic reduc- tion of its associated glycoproteins at the sarcolemma resulting in severe muscular dystrophy2,5. In the milder Becker muscular dystrophy, mutations in dystrophin result in the ~ r o d u c t i o n of a dvsfunctional vrotein16. More recently, some cases of severe childhood autoso- ma1 recessive muscular dystrophy (SCARMD or LGMD2D) were shown to be caused by missense and null mutations in the adhalin genel7>l8, which result in the reduction or absence of adhalin at the sarcolem- ma19. Non-Fukuyama congenital muscular dystrophy (CMD) has recently been linked close to the merosin locus on chromosome 6q20,21, which is likely to be responsible for this disease. Thus, in these muscular dystrophies, mutations in one component of the DGC cause the disruption of the complex and consequently lead to the dystrophic process. The limb-girdle muscular dystrophies (LGMDs) represent a clinically heterogeneous group of diseases which are characterized by progressive weakness of the pelvic and shoulder girdle m u ~ c l e s ~ ~ , ~ ~ . These disor- ders may be inherited as an autosomal dominant or recessive trait, the latter being more common with an estimated prevalence of I in 100,000 indi~idua1.s~~. Several genes have been implicated in the aetiology of these disorders. The autosomal dominant form, LGMDlA, was mapped to 5q22-q34 (ref. 25), while four genes involved in the autosomal recessive forms were mapped to chromosomes 2~13-p16 (LGMD2B)26, 13q12 (LGMD2C)27,28, 15915.1 (LGMD2A)29, and 17q12-q21.33 (LGMD~D)". The genes responsible for LGMD2D and LGMD2A have been identified: adhalin17 and muscle-specific calpain ( c A N P ~ ) ~ ~ , respectively. Cases of autosomal recessive limb-girdle muscular dystrophy among members of the old order of Amish of northern and-southern Indiana were described by Iackson and Carev31 and Tackson and Strehler32. Most of the families of these communities are interrelated by multiple consanguineous links and common ancestry which can be traced to the 18th and 19th century in the canton of Bern. Switzerland31. In view of the high u degree of consanguinity and the similar clinical pre- sentation of all Amish LGMD patients, the demonstra- tion of genetic heterogeneity within this community was unexpected33. Although families from northern Indiana were shown to carry the same R769Q calpain 3 mutation30, involvement of this locus was excluded in nature genetics volume 11 november 1995 257 article ( U ~ O U T O ( U O T T T U T O M A O T O O C C % ~ ~ m ~ T A T C T o 1 1 l T ~ 360 D S U Z T B Z S O L L R T X ~ V S D U O 117 Q ~ T C C M C C T C ~ T M ~ T ~ ~ T T T C m U T C 420 V I B P L X X S T V G O R R N Z N L V I 1 3 7 W-CMCUOCCTATTGTTTTT---TWdA- 480 T a N N Q P I V I Q Q O T @ X L S V Z N 1 5 1 W T T C T A T T h C M G - X O O C A T ~ T T T T I ~ C C ( W W L C T - 540 W X T ~ I T S D I ~ U Q ~ ~ ~ P R T _ P ? 1 7 7 A T C T T A T T ~ G 1 S T A T G A M C T C A T ~ G ~ T C A T T T O C C M G T O ( U M ~ T 600 I L ~ S T D X Z T B E ~ B L P S G V X S 1 9 7 TTOMmTT-OCITCTXT~TTACUGCMTOCTXUOT~TTTLUT 660 L U V Q X A S T L R I T S P A T S D L N 217 AT-TTMTGOOCGTGCTATTOTOCOT-TOMOGTGTATTUTTATG- 7 2 0 I X V D G R A I V R O N L G V r I M G K 2 3 7 XUTIOMTTTUUTGOaT00TMTATO(UGTTLUOOCO6UMUOTATUTCCTA 780 T I L T B M O O N M L L X A Z N I I I L 257 MTWTCTGTGATGTUOUCCACCCOCCTACCCAOTTCCTCUOT-CWG 840 W G S V N V S T T R L P S S S S G D Q L 2 7 7 00TMiTQOTOLCTGffiTACDCTACUOCTCT~TOTMOCTGITW(UCOCTCTT~ 900 O S O D Y V R X X L C U C A D G T L I X 2S7 Q T O C U G T M C C M C U O M C A T O O O C T D C C M A T C T ~ C C C C T G T ~ C T 960 V Q V T S Q N M G C Q I S D N P C G N T 3 1 7 ULTTIIlJIUiMCCCU(UGGTCMCUUNTTTATATCTTb)rCTT01CTTTTTTATOU 1 0 2 0 B STOP 311 TOUUTCATTGTTTTTAUWTTTGTDLTMCTUTUTTATTTTMTOOCA~ 1080 Fig. 1 Human P-sarcoglycan cDNA and translated amino acid sequence. Human P-sarcoglycan cDNA is shown with the deduced amino acid sequence below. The predicted transmem- brane domain is indicated by the shaded box. Three consensus sites for N-linked glycosylation are indicated by asterisks, *. The intracellular consensus phosphorylation site by protein kinase C or casein kinase II is indicated by a #. Peptide sequence frag- ments identified in the amino acid sequence are underlined. The amino acid that is mutated in the Amish limb-girdle muscular dystrophy patients is circled. Amish families from southern Indiana, as none of the examined patients from these cohorts carried this mutation, even in a heterozygous state (data not shown). Furthermore, the role of known LGMD loci, as well as several other candidate regions involved in other neuromuscular disorders, were all excluded33. These results thus implied the existence of yet another locus, LGMD2E, involved in autosomal recessive limb- girdle muscular dystrophy. Here we report the cloning and characterization of human P-sarcoglycan, a 43 kDa component of the dys- trophin-glycoprotein complex, and demonstrate its involvement in LGMD2E. We show that P-sarcoglycan colocalizes with the DGC at the sarcolemma and is expressed ubiquitously, but predominantly in muscle. We report the primary localization of this gene to chromosome 4q12, the identification of an intragenic polymorphic CA repeat, and linkage of LGMD2E to the intragenic P-sarcoglycan microsatellite in southern Indiana Amish LGMD families. We have also identified a missense mutation in the P-sarcoglycan sequence, which is present in a homozygous state in all patients, and is apparently responsible for the muscular dystro- phy in all chromosome 4-linked Amish LGMD families screened. By immunofluorescence of affected individ- uals, we demonstrate a dramatic reduction in the level of the P-sarcoglycan protein at the sarcolemma, as well as a concomitant reduction of adhalin and the 35 kDa DAG, thereby disrupting the integrity of the DGC. Together, these data demonstrate that a defect in 0- sarcoglycan is responsible for LGMD2E. Because of this protein's tight association with other components of what has been called the sarcoglycan complex34, which includes adhalin, this 43 kDa protein, the 35 kDa DAG, and the 25 kDa DAP, the name 'P-sarcogly- can' has been proposed. This name was mutually agreed upon by us and the authors of an accompany- ing manuscript which describes similar results35. Under this nomenclature, a-sarcoglycan becomes syn- onymous with adhalin. P-Sarcoglycan cDNA a n d primary structure The dystrophin-glycoprotein complex (DGC) was identified in 1989 based on the ability of dystrophin to be retained on a wheat-germ agglutinin When P-dystroglycan, a 43 kDa DAG, was cloned1', the translated peptide sequence was compared with peptide sequence fragments obtained from the 43 kDa band of the purified DGC. Of these fragments, only one was found in the primary structure of P-dystrogly- can. This suggested the presence of another protein of similar molecular weight, consistent with the observed 43 kDa doublet in the DGC2,3. TO further investigate this protein, we searched the GenBank database of expressed sequence tags (dbEST) with the unidentified peptide sequences using the TBLASTN search pro- gram36. Several ESTs encoding peptide sequence frag- ments of the P-sarcoglycan protein were identified and isolated from a normalized human infant brain cDNA library constructed by Dr. Bento Soares et ale3'. Clones 22297 and 25556, from which ESTs were generated, were received from the IMAGE Consortium at the Lawrence Livermore National Laboratories. The larger of these clones, 25556, was fully sequenced on both strands. In addition, we isolated clones from a XZAPII human skeletal muscle cDNA library using the 1225- bp insert of clone 25556 as a probe. Sequence analysis revealed a single open reading frame that encodes a protein with a predicted molecular weight of 34,777 Da (Fig. 1). Several peptide fragments obtained from sequencing of the 43 kDa doublet were found in the primary structure of the protein. There was no other significant homology with any other known protein or functional protein domains in the database. Addition- ally, our sequence data provided evidence of alternate polyadenylation. Two distinct poly(At) tails were iden- tified, one that is about 300 bases downstream of the stop codon in brain, and one that is approximately 3 kb downstream in skeletal muscle (data not shown). Hydropathy analysis of the amino acid sequence revealed a single transmembrane domain and n d func- tional signal sequence at the N terminus, despite the presence of eight hydrophobic alanine residues follow- ing the initiator methionine. Thus, the small N-termi- nal domain of the protein is predicted to be intracellular, whereas the large C terminus is extracel- lular. This membrane topology is consistent with the location of the three putative N-linked glycosylation nature genetics volume 11 november 1995 article a s 5 . . . . - . . . . . . . . 7 5 . . - . . . . . ~ ~ . . . . . . . 4 4 c * . . , - m. ' . . . . - . . .- - ~. - . . . . - - . . . . . . . . . . . 5' probe - . . 2 4 ' -: . . . .- ~ . . - . . . . . . . . . . 1 . 4 . . . . , "' . . . . . . 4 4 F .) .. " full-length probe adult fetal Fig. 2 Expression of 0-sarcoglycan in adult and fetal tissues. RNA blots containing 2 pg of poly(A+) RNA from the indicated human adult and fetal tissues were hybridized as follows. Molec- ular size in kb is indicated on the left. a, Northern blots hybridized with a 334-bp 5' segment of the P-sarcoglycan cDNA. b , Northern blots hybridized with the 1225-bp insert of P- sarcoglycan clone 25556. sites. all of which are C-terminal to the transmem- brane domain. There are also five extracellular cysteine residues which may form disulfide bonds. There is also o n e potential intracellular consensus site for phospho- rylation by protein kinase C o r casein kinase I1 at Ser21. T h e predicted membrane organization is similar to that of P-dystroglycan a n d adhalin, both of which have single transmembrane domains a n d large extra- cellular a n d short intracellular domains'0,' '. T i s s u e d i s t r i b u t i o n of P - s a r c o g l y c a n To determine the tissue-specific expression of p-sarco- glycan, we performed RNA hybridization analysis. Human adult a n d fetal multiple tissue northern blots were probed in two different ways: once with a PCR fragment encompassing bases 132-465 of the p-sarco- glycan coding region, and a second time with the 1225 b p insert of clone 25556 described above (Fig. 2). T h e predominant transcript is approximately 4.4 kb in length; however, there are also weaker signals of 3.0 a n d 1.35 kb. Seauence data demonstrating evidence of " alternate polyadenylation can account for the smallest and largest transcript. P-sarcoglycan RNA is present in all tissues, a n d is particularly enriched in skeletal a n d cardiac muscle. This pattern of expression is different from adhalin, which is expressed only in muscle tis- ~ u e ~ ~ ~ ' ~ , but is similar to dystroglycan, which is ubiqui- tously e x p r e ~ s e d ' ~ . Interestingly, the northern blot results are different between t h e two probes. When probed with a 334-bp PCR fragment corresponding to nucleotides 132-465, the signals in the fetal liver a n d adult pancreas lanes are weak o r absent (Fig. 2a). However, when probed with the larger clone, which contains all the coding region and nearly 300 b p of 3' untranslated region, these signals are significantly stronger, particularly in the adult pancreas (Fig. 2b). This suggests that the 5' e n d of the coding region is alternatively spliced among different tissues. ed a glutathione-S-transferase (GST) fusion protein (FP-I) containing 64 residues C-terminal to the trans- membrane domain (Fig. 30). Anti-FP-I polyclonal antibodies. ~ r o d u c e d in rabbits. were used to detect B- , , sarcoglycan in isolated membranes a n d purified DGC. T h e antibody specifically recognizes a 43 kDa protein i n both crude sarcolemlna and purified DGC, but does not recognize GST alone (Fig. 3b). Sheep polyclonal antibodies produced against a peptide fragment (residues 42-52) of the P-sarcoglycan protein also rec- ognized a 43 kDa protein in the purified DGC (data not shown). Identification of P-sarcoglycan during the purification of the DGC demonstrates that P-sarcogly- can is an integral component of the DGC. To determine the subcellular localization of B-sarco- glycan by immunofluorescence, serial transverse cryosections of control human biopsied skeletal nius- cle were immunostained with anti-FP-I antibody pre- absorbed with GST, as well as with antibodies against other components of the DGC including dystrophin, P-dystroglycan, syntrophin, adhalin, and 35 kDa DAG. As shown in Fig. 4, the anti-FP-I antibody labelled throughout the entire sarcolemma a n d showed colo- calization of P-sarcoglycan with other components of the DGC. However, immunofluorescence from D M D muscle biopsy samples demonstrated a loss of the P- sarcoglycan protein at the sarcolemma along with all other components of the D G C (data not shown). L o c a l i z a t i o n to c h r o m o s o m e 4q12 Primers were designed to amplift a fragment of the h u m a n P-sarcoglycan gene from a panel of hunian- rodent somatic cell hybrids containing various combi- nations of h u m a n chromosomes. Restriction digests of the amplified product with TaqI specifically cleaved the human allele, a n d allowed assignment of the p-sarco- glycan gene to chromosome 4. To f i ~ r t h e r narrow the 1 GST I I I FP-I 89 152 Fig. 3 lmmunoblot analysis of P-sarcoglycan. a, Schematic model of human P-sarcoglycan cDNA. The transmembrane domain is shaded; three sites of potential N-linked glycosylation sites are indicated. A schematic representation of fusion protein I (FP-I) is shown below. b , Nitrocellulose transfer of SDS-PAGE gel stained with anti-FP-l antibody. Membrane contains the fol- lowina: KCI-washed microsomes (Mic), crude rabbit skeletal muscie sarcolemma (CSM). Duriied 'dvstro~hin-a~vcoorotein L o c a l i z a t i o n of P - s a r c o g l y c a n p r o t e i n complex (DGC), f ~ s i o n ' ~ r o t i i n i (FP-I), anh glutathiGe-~-trans- ferase (GST). Anti-GST antibodies were removed prior to stain- To confirm that the 'Ioned represents the 43 ing. Molecular weight in kDa is indicated on the left. Arrow on kDa dystrophin-associated glycoprotein, we construct- right indicates position of p-sarcoglycan protein. nature genetics volume 11 november 1995 259 article Dystrophin D-Dystroglycan Syntrophin sarcoglycan gene between nucleotides 438 and 439 of search using the highly' informative microsatellit& Fig. 4 lmmunohistochemical analysis of P-sarcoglycan in normal and Amish LGMD2E muscle. Skeletal muscle biopsy samples were stained with antibodies against DGC components as described in the Methods. DYS, anti-dystrophin; P-DG, anti P- dystroglycan; 35DAG, anti-35 kDa dystrophin-associated glyco- protein. Control, normal subject; patient, Amish LGMD2E patient. Bar, 50 LIM. chromosonlal region, the same process was used t o analyse DNA isolated from human-rodent somatic cell hybrids containing various fragments of cl~romosome 4 (ref. 38). P-sarcoglycan fragments could be amplified only from hybrids containing the region 4p14-q21.1, which overlaus the centromere. We isolated two cosmids spanning approximately 40 kb of the human P-sarcoglycan gene by screening a human chromosome 4 cosmid library. The smaller cosmid, which contained a 28.5 kb insert, was used as a probe for fluorescence i n s i t u hybridization (FISH), and resulted in the specific labelling of the centromere o n the long arm of chromosome 4, corresponding to band 4q12 (data not shown). We searched for polymorphic microsatellites within the P-sarcoglycan gene. Southern blots of restriction fragments of the c o s ~ n i d s and genomic PCR fragments were probed with oligonucleotides encoding a dinu- cleotide (CA) repeat and several tetranucleotide repeats. Only hybridization with the CA repeat oligonucleotide was detected. Sequencing subsequent- ly located a novel CA repeat within a n intron of the P- Fig. 5 Recombinant haplotypes in w -3 <\I-- LGMD2E families. The ordered marker loci m , r C1 r U have arbitrarily been represented as =. : 2 9 - - ; q ; s s z $ d ; equidistant. Loci bracketing the smallest ., X I " - - , .$ - z a g ~~~~~~ interval defined by recombination events ( 8 UA~I-a -a : z = = --- --- are noted in larger letters. The intragenic ,.::::: D-sarcoalvcan microsatellite is underlined. ..w.v -- S o l ~ d squares, affected lndlvlduals, open :;;;El, = - : : : - : = - L- squares, healthy carrlers The numbers .,,,,,,, -; I --; : indicate the family and the individual. a , \ 6 1 * 1 r a --ct . ...*~- - -I- Chromosomes segregating with the dis- i::::::: I-?<--= - = - ease allele or the normal allele are coded n A 6 L b l l ~ --*---- as solid or open circles, respectively. Thin horizontal lines represent the recombina- tion interval. Uninformative markers are coded by a line in place of a circle, and nongenotyped markers are left blank. The last two recombinants define the critical interval for the location of the morbid locus. a, Previously described Amish LGMD fami- lies. b, Additional southern Indiana Amish families. In autosomal recessive disorders, affected individu- als from consanguineous families often show homozy- gosity by descent at the region surrounding a disease locus40. Haplotypes were ~nanually constructed for the chromosome 4 markers assuming a minimal number " of recombinations. We identified a unique carrier hap- lotype segregating within all the southern Indiana Amish population (data not shown), suggestive of a unique founder effect, though different from the o n e found in the northern Indiana and Pennsylvania Amish LGMD2A families33. Six affected and one unaf- fected offspring showed informative crossovers (Fig. 50). This identified D4S1547 and 04.5'1627 as newr flanking loci which define a region of approximately 9 cM, based o n analyses of CEPH reference families". Five additional southern Indiana LGMD families also showed linkage to this new locus, thereby increas- ing the number of informative meioses. A maximum lod score of 11.97 at 0 = 0.0 was obtained with marker D4S518 (Table In). Genotyping of these families with new microsatellite markers allowed a further narrow- ing of the LGMD2E interval, flanked by markers D4S396 and D4S1630 (Fig. 5 b ) . Ho~nozygosity map- ping and reconstitution of historical crossing over events suggested that the LGMD2E interval is flanked by markers D4S396and D4S428 (data not shown). Based o n physical maps for chromosome 4 (ref. 41 and D. Cox. versonal communication^, CEPH YACs , L spanning this region were used to localize the p-sarco- glycan gene inside the LGMD2E interval, between markers D4S1577 a n d D4S1630. Genotyping of the intragenic microsatellite in LGMD2E families yielded a lod score of 7.21 at 0 = 0.0 (Table l b ) . This lower lod score, as compared t o D4S518, is d u e t o a lower observed heterozygosity and hence lower inforrnativity of the intragenic marker. Identification of a mutation in Amish patients \Ve performed northern blot analysis o n total RNA isolated from skeletal muscle biopsies of two affected siblings in these families to determine whether P- sarcoglycan mRNA size or abundance were affected. The major muscle P-sarcoglyca~~ transcript (4.4 kb) was present at normal levels and size in both affected sibs (one of which is shown in Fig. 6) c o ~ n p a r e d to an 260 nature genetics volume 11 november 1995 article -- Table 1 ~ i n k a ~ e of LGMDPE families to chromosome 4 markers 4), and laminin a 2 chain (not shown) were present at comparable levels with control muscle. However, the Z at recombination (0) of immunostaining of 0-sarcoglycan was greatly Onelodsupport decreased, with- a concomitant-reduction of adhalin Marker O.OO O.O1 0.05 O.1° 0.20 (') and the 35 kDa DAG. The apparent differences in mus- a 0481547 -m 8.24 8.51 7.69 5.46 8.64 (0.031) 0 , 0 0 4 ~ . 1 0 3 cle cell size between control and ~ a t i e n t samples is due 0481627 -m 2.06 2.96 2.91 2.17 2.99 (0.068) - to cell hypertrophy associated with dystrophic changes. 048401 -- 5.37 5.39 4.82 3.38 5.51 (0.026) 0.001-0.124 D i s c u s s i o n Limb-girdle muscular dystrophy in the Amish of Indi- ana was first described over 30 years ago3', but only recently has the nature of the disease been closely stud- ied. The LGMD in the northern Indiana Amish popu- lation is caused by mutations in calpain 3, a cysteine proteinase that niay be involved in muscle protein pro- cessing or signal t r a n s d u ~ t i o n ' ~ . However, analysis of southern Indiana Amish revealed no linkage to this locus3'. The genetic heterogeneity of LGMD in this population was unexpected, because of evidence of common ancestry and multiple consanguineous mat- ings within these populations. The human cDNA of P-sarcoglycan, a component of the dystrophin-glycoprotein complex, has now been cloned. The identity of this clone is confirmed by the identification of peptide sequences in the primary c T151R mutation -- - - - - - a, Maximum pairwise lod scores and their corresponding recombination fractions with one-lod support intervals obtained in ten southern Indiana families between the LGMD2E locus and chromosome 4 markers. b, The intragenic microsatellite. c, The TI51 R mutation. Marker loci are listed according to their order on the regional map of chromosome 4. unrelated control. This strongly suggested that the causative mutation was most likely to involve a small deletion, insertion, or base substitution. Fragments of the P-sarcoglycan cDNA were ampli- fied following reverse-transcription from total RNA prepared from biceps brachii muscle biopsies of these two affected sibs. Sequencing the RT-PCR products structure, recognition of a 43 kDa protein in purified DGC by a P-sarcoglycan-specific antibody, and sar- colemrnal colocalization of this protein with other components of the DGC. The predicted molecular weight of this translated product is 34,777 Da. Experi- ments with endoglycosidase F indicate that the dis- crepancy between the predicted and observed revealed a single C to G transversion at nucleotide 461 in both patients in a homozygous state (Fig. 6). The codon change is ACA to AGA and results in a Thr-to- Arg substitution at residue 151 (T151R). Segregation of this mutation was assessed in this family and in other Amish LGMD2E families by sequencing and 'touchdown' PCR". Results showed perfect cosegregation of this missense mutation with the disease in all southern Indiana Amish families test- molecular weights is primarily due to the glycosylation of this ~ r o t e i n (data not shown). The P-sarcoglycan gene has been mapped to chro- mosome 4q12, the same region to which we have linked LGMD in the southern Indiana Amish. In agreement with the common haplotype displayed by these families around the disease locus, a mutation common to all southern Indiana Amish LGMD2E patients tested has been found. This suggests that a unique founder effect may account for LGMD2E in this population, as has already been observed in the LGMD2A families of the northern Indiana and Penn- ed, as expected from the common haplotype at this locus (Fig. 7). To exclude the possibility that this mis- sense mutation might be a polymorphism, 122 unre- lated chromosomes taken from the CEPH reference families were tested; none showed this mutation, nor did any northern Indiana LGMD2A Amish patients (data not shown). Linkage analysis with the mutation was performed (Table l c ) . A maximum lod score of 13.43 at 0 = 0.0 a GO" 6"' ,.c‘ b 9 5 * Control 7 5 + S A I I C * . G T " l C I , C 1 1 A . S r C * . 4.4 * R t , 2.4 * Patlent between the mutation and the disease in these Amish families was obtained. In addition, assuming complete linkage disequilibrium, the calculated lod scores are significantly increased (data not shown), further con- firming that the mutation in P-sarcoglycan is responsi- ble for the disease in these patients. Fig. 6: Northern blot and sequence analyses of 13-sarcoglycan gene in family A623. a, An R N A blot containing 12 pg of human skeletal muscle total R N A per lane from an unrelated control and an affected member of family A623 hybridized with a human P- sarcoglycan cDNA probe, representing nucleotides 1 to 1225, and exposed overnight. b, Direct nucleotide sequence determi- nation of reverse-transcribed total R N A representing nucleotides 1 to 1132 performed on the same individuals described above and on another affected member of the family A623. Arrowheads indicate the position of the mutation. Both affected members of family A623 showed the same mutation. P-sarcoglycan deficiency in LGMD2E m u s c l e Skeletal muscle biopsy specimens from the two patients described above were examined by immuno- fluorescence to test the effects of the T151R mutation on P-sarcoglycan expression. Serial frozen sections were stained with antibodies against P-sarcoglycan (anti-FP-I) or other DGC components as described above. Dystrophin, P-dystroglycan, syntrophin (Fig. nature genetics volume 11 november 1995 26 1 article explain the long survivability a n d slow disease progres- sion of affected individuals. In the future. it will be interesting to determine whether patients with null mutations exhibit anv svlnvtoms related to the trunca- , , . tion of the P-sarcoglycan protein, particularly in the nonmuscle tissues in which the protein is also known to be expressed (Fig. 2). A feature colnmon t o LGMD2D and LGMD2E is a selective absence of P-sarcoglycan, adhalin, a n d the 35 kDa DAG at the sarcolemma. T h e close association of these proteins had previously been s ~ g g e s t e d ~ ~ , ~ ~ , ~ ~ , a n d recent immunoprecipitation experiments have further demonstrated the existence of a subcomplex composed of these proteins within the DGC ( u n p u b - lished data). The loss of all cotnponents of this sub- complex in both LGMD2D and LGMD2E indicates that mutations in o n e component can affect the stabili- ty o r targeting of other components, further support- ing the hypothesis that the missense mutation found in the southern Indiana Amish is the cause of LGMD, since this mutation affects the function of not just p- sarcoglycan, but adhalin and 35 kDa DAG as well. This also suggests yet another mechanism by which LGMD could occur, through a mutation in the 35 kDa DAG gene. The precise function of P-sarcoglycan and its sub- complex as a whole is currently unknown. In Duclienne muscular dystrophy, disruption of the DGC causes the loss of a functional link between the extra- cellular matrix a n d the cytoskeleton, and may result in sarcolemmal instability a n d greater susceptibility to stress-induced damage. It is likely that the B-sarcogly- Fig. 7 Illustration of the segregation of the Thr to Arg substitution in the southern Indi- canladhalinl35 kDa ~ A G complex plays a similar h e ana LGMD2E Amish population. a, Schematic representation of 'touchdown' PCR in maintaining the structural and functional integrity amplification. Products of 100-bp and 158-bp are produced using primer R461 cou- of the cell membrane. In the BIO 14.6 cardiotnvo~athic - . , . pled with primer m l and primer T461 coupled with primer m3, respectively. Primer hamster, a deficiency of this subcomplex is apparently sequences are given in the Methods. b, Family A620 pedigree with affected and unaf- fected individuals indicated by closed and open symbols, respectively. c, Duplex for the a n d c a r d i O m ~ O - deposit of "touchdown" PCR amplification products. d, Chromosome 4 haplotypes pathic changes in this Dystrophin and a- segregating within the family. Disease-bearing chromosomes are boxed. CA12T repre- dystroglycan are not tightly associated with the sents the intragenic microsatellite. sarcolemma, indicating that disruption of the sarcogly- sylvania Amish. T h e perfect cosegregation of the muta- tion with the disease trait within the LGMD2E Amish population a n d lack of the mutation in over a hundred control c h r o m o s o n ~ e s is strong evidence to support the hypothesis that this base change is responsible for the disease. These genetic data, together with o u r bio- chemical results, demonstrate the role of P-sarcoglycan i n the aetiology of this disorder. As yet, however, the functional importance of the region in which the mutation occurs has not been defmed. Bonnemann et al. have cloned the cDNA of the same protein, and their results are reported in a manuscript published concurrently with this35. They have identi- fied a compound heterozygote with mutations result- ing in significant truncations of P-sarcoglycan. This patient's clinical presentation is much niore severe than that of the Amish LGMD2E patients, indicating that the 11~11 mutations in this patient produce a much more deleterious effect. This pattern of severity, com- pared to the milder clinical course of the Amish patients, is similar to that seen in the spectrum of adhalin mutations that cause L G M D ~ D I ' S ~ ~ . Thus, the presence of a ~nissense rather than a nonsense muta- tion in the southern Indiana Amish population could can cotnplex can lead to the d e ~ t ~ b i l i z a t i o n of the dys- troglycan complex as well4" Beca~tse n o mutations have been found in the adhalin cDNA o r mRNA of the cardiomyopathic hamster4', mutations in P-sarcogly- can may be responsible for the loss of this cotnplex and the resulting phenotype. It is also possible, liowe\ler, that P-sarcoglycan alone serves an important function- al role in nonmuscle tissues as well as in muscle, as suggested by the presence of P-sarcoglycan transcript in all tissues studied. Because the pattern of adhalin expression is different from that of P-sarcoglycan, P- sarcoglycan may function in a manner independent of the other components of the complex. In nonmuscle tissues, P-sarcoglycan [nay be part of another subcom- plex. It will be of interest to determine the structure a n d cellular localization of P-sarcoglycan in these tis- sues a n d to identify its associated proteins. T h e pres- ence of P-sarcoglycan in brain is particularly interesting because of the involvement of dystroglycan in synapse f o r m a t i ~ n ~ ~ , ~ ~ , though its involvement in this o r in other processes in the nervous system remains to be determined. The genetic a n d clinical heterogeneity of the limb- girdle muscular dystrophies has broad implications for the diagnosis a n d management of persons afflicted 262 nature genetics volume 11 novernber 1995 artzcle with this disease. Because defects in several distinct proteins can cause phenotypically similar clinical and biochemical manifestations, a definitive diagnosis can only be achieved through genetic testing of all candi- date genes. At the current stage of neuromuscular dis- ease research and therapy, such measures can only provide information on the precise aetiology of LGMD, but this information will be important in directing proper therapy in the future. Methods Peptide sequencing and isolation of ESTs. The 43 kDa band of the purified dystrophin-glycoprotein complex2 was partially sequenced, and several peptide fragments were obtained and are indicated in Fig. 1. Peptide sequences were used to search the GenBank database of expressed sequence tags (dbEST) using the TBLASTN search program36. Several overlapping ESTs were identified that represented portions of the p-sarco- glycan cDNA. Human P-sarwglycan cDNA cloning a n d characterization a n d mRNA analysis. Two of the clones, 22297 and 25556, from which ESTs were generated, were obtained from the IMAGE Consortium at Lawrence Livermore National Laboratories. The larger of these clones, 25556, was fully sequenced o n both strands using an Applied Biosystems, Inc. automated sequencer. This clone was determined to contain the full cod- ing region of the P-sarcoglycan cDNA as well as about 300 bp of 3' untranslated region and a poly(A+) tail. Clones were iso- lated from a XZAPII human skeletal muscle cDNA library (Stratagene) using the 1225 bp insert of IMAGE clone 25556 as a probe. Primary structure and site detection analyses were performed using PCIGene software (IntelliGenetics). This sequence data is available from the GenBank database, acces- sion number U29586. CLONTECH adult and fetal human multiple tissue northern blots containing 2 pg of poly(A+) RNA per lane were probed with a 334-bp PCR-amplified probe that represents nucleotides 132 to 465 of the P-sarcoglycan sequence, and with the 1225-bp insert d o n e 25556. Fusion protein construct. A 192-bp region of the P-sarcogly- can cDNA downstream of the predicted transmembrane domain was amplified by PCR using the following primers: sense, 5'-GCCGGGATCCGTGATTCGCATTGGACCAAA-3'; antisense, 5'-GCGCGAATTCCTTTGTTGTCCCTTGCT- GAA-3'. Following restriction digest with BamHI and EcoRI, the product was subcloned into pGEX2TK50 and introduced into E. Coli DH5a cells. 50 ml overnight cultures were diluted 1:10 and induced with IPTG to promote fusion protein (FP-I) production. Fusion proteins were purified o n a glutathione- agarose column51. Antibodies. Anti-P-sarcoglycan antibodies were generated by intramuscular and subcutaneous injection of New Zealand white rabbits with 100 pg of purified FP-I in an emulsion of Freund's complete adjuvant. Rabbits were boosted two weeks later with a subcutaneous injection of 500 pg of FP-I in PBS (50 mM sodium phosphate, p H 7.4,0.9% NaC1). Rabbits were bled two weeks following boost and the serum was tested for the presence of anti-FP-I antibodies. The serum was cleared of anti-GST antibodies with a glutathione column and anti-FP-I antibodies were affinity-purified using Immobilon-P strips containing 250 pg of FP-I. Monoclonal antibodies VIA42 against dystrophin, and IVD31 against adhalin were previously ~ h a r a c t e r i z e d ~ , ~ ~ . Monoclonal antibody 8D5 against 0-dystro- glycan was kindly provided by Louise Anderson. An affinity- purified rabbit antibody against 35 kDa DAG (D. Jung, unpublished results) was also used in this study. Monoclonal antibody against human laminin a 2 chain was purchased from Chemicon. Western blot and immunofluorescence. KC1-washed micro- somes, crude rabbit skeletal muscle sarcolemma, and purified DGC were prepared as previously d e ~ c r i b e d ~ . ~ ~ . Proteins were resolved on a 3%12% SDS polyacrylamide and trans- ferred to nitrocellulose by electroblottingS5. Blots were incubat- ed overnight in a 1:20 dilution of affinity-purified anti-FP-I antibody in Blotto (5% nonfat dried milk in TBS [20 mM Tris- HCl, 200 mM NaC1, pH 7.4]), then incubated with a horserad- ish peroxidase-conjugated goat anti-rabbit IgG secondary antibody (Boehringer Mannheim) for 1 hr. Antibody staining was detected with H 2 0 2 in TBS with 4-chloro-1-naphthol as a substrate. For immunofluorescence, 7 p m transverse cryosec- tions were prepared from control and Amish LGMD muscle. The following procedures were performed at room tempera- ture. Sections were treated with AB blocking solutions (Vecter), blocked with 5% BSA in PBS for 30 min, and then incubated with a 1:20 dilution of affinity-purified anti-FP-I antibody for 90 min. Antibodies against the following components of the DGC were also tested: dystrophin, laminin a 2 chain, P-dystro- glycan, syntrophin, adhalin, and 35 kDa DAG. After extensive washing with PBS, sections were incubated with biotinylated secondary antibodies (1:500) for 30 min, washed with PBS, then incubated with FITC-conjugated streptavidin (1:1000) for 30 min. After rinse with PBS, sections were mounted with FITC-guard (Testog) and observed under a Zeiss Axioplan flu- orescence microscope. Photographs were taken under identical conditions with the same exposure time. P-sarcoglycan gene localization t o chromosome 4q12. Primers corresponding to human P-sarcoglycan cDNA nucleotides 291-312 (sense) and 413-429 (anti-sense) were used in PCR using DNA from a panel of 25 human-rodent somatic cell hybrids (BIOS Corporation) containing various combinations of human chromosomes. Subsequent restriction digest of the PCR products by TaqI was used to distinguish between the human and rodent alleles. Use of a somatic cell hybrid panel containing various regions of chromosome 4 (ref. 38) further narrowed the location of the gene using the same approach described above. A chromosome 4 cosmid library (kindly provided by Jeffrey Murray) was screened with a 32P-labelled PCR product repre- senting nucleotides 135-429. Two cosmids with inserts of 28.5 kb and 35 kb were obtained. CsC1-purified DNA from the smaller cosmid was used for fluorescence in situ hybridization mapping which was performed by Genome Systems. Families. Six previously described LGMD Amish families from southern Indiana (52 individuals, 13 affected)33 were analysed in the linkage search. Subsequently, DNA from 5 additional southern Indiana families were included in this study (39 indi- viduals, 13 affected). AU of these kindreds show multiple con- sanguineous links (data not shown). In these families, 11 patients were clinically reevaluated after exclusion of the cal- pain 3 locus. AU patients presented with proximal symmetrical weakness and atrophy of limb and trunk muscles. Clinical onset was at 7.6 years of age (range 4 to 12 years), loss of walk- ing at 26 years (range 12 to 38 years), with marked intrafamilial variability. Calf hypertrophy was similar to that observed in BMD and LGMD2D patients, but different from Reunion Island LGMD patients56 and LGMD2A patients from the northern Indiana Amish community. Genotyping and linkage analysis. Markers were selected from the microsatellite panel described39 or from CHLC maps5'. DNA (50 ng) was used as templates in a 50 pl polymerase chain reaction as described58. Southern blots of restriction fragments of cosmids were probed with (CA), and tetranucleotide oligonucleotide repeats labelled with y-32P ATP. The positive fragment was subcloned and sequenced. The identified intra- genic polymorphic CA repeat was amplified using the follow- ing primers: sense (5'-TATCTTCTAATGTCTTCTGTCTAT-3') nature genetics volume 11 novernber 1995 263 article and antisense (5'-GAAACAAGAATAACATGCCATTT-3'). PCR conditions for this marker were - denaturation at 94 OC for 1 min, annealing at 60 "C for 1 min, and extension at 72 OC for 1 min, for 30 cycles. Primer sequences, PCR conditions, and other information concerning the highly polymorphic microsatellites used in this study can be obtained from the Genome Database, John Hopkins University. Two-point and multipoint linkage analyses were carried out using the LINK- AGE software package, version 5.1 (ref. 59) assuming fully pen- etrant autosomal recessive inheritance with a gene frequency of 0.001. RNA isolation and reverse-transcription PCR. Total RNA was extracted from 20-30 mg of skeletal muscle from one control and two Amish LGMD2E patients from family A623 (ref. 33) using RNAzol (Tel-Test) according to manufacturer specifica- tions. RNA samples were run on 15% formaldehyde/l.5% agarose gels and transferred to Hybond N+ membrane (Amer- sham). Membranes were then hybridized with the PCR labelled cDNA as described above. Total RNA ( 1 pg) was used for reverse transcription with a specific primer representing nucleotides 11 13-1 132 (antisense) in the p-sarcoglycan 3' cDNA untranslated region, in a reaction mixture containing 6 mM MgC12, 200 mM dNTP, 50 mM KCI, 10 mM Tris pH 8.2, 40 U RNasin, 10 pmol specific primer, 6 U AMV reverse transcriptase, and incubated for 90 min at 42 "C. PCR on the reverse-transcribed product was performed using the same 3' primer and one of two 5' primers (representing nucleotides 1-18 and 47-68 respectively) to cover the p-sarco- glycan cDNA coding sequence. The RT-PCR amplification products were analysed by agarose gel electrophoresis and by direct sequencing. Touchdown PCR. DNA (50 ng) was subjected to touchdown P C R ~ * in a 50 p1 reaction mix containing 10 mM Tris-HC1, pH 8.8, 50 mM KC1, 1.5 mM MgCl,, 0.1% Triton X-100, 200 mM of each dNTP, 100 ng of each primer, and 2 U Taq Polymerase (Perkin Elmer). After 5 min denaturation at 96 "C, amplifica- tion cycles were carried out as follows: 40 s denaturation at 94 "C followed by 30 s annealing steps starting at 63 'C with a decrease of 1 "C every two cycles until 59 "C. Twenty-five addi- tional cycles of amplification consisting of 40 s at 94 OC and 30 s at 58 OC, were performed. Primer pairs R461Iml and T4611m3 were designed to yield, respectively, a 100-bp product from individuals carrying the mutation, and a 158-bp product from individuals not carrying the mutation. Primers sequences were: R461: 5'-GTTTTTCAGCAAGGGACAAG-3'; T461: 5'-GTTTTTCAGCAAGGGACAAC-3'; m l : 5'-TATTTTGAG- TCCTCGGGTCA-3'; m3: 5'-CTTTTCACTCCACTTGG- CAA-3'. PCR products were mixed in a 1:l volume ratio and run in a single lane on 4% agarose gels stained with ethidium bromide. Acknowledgments We thank M.B. Soares and G. Lennon of the IMAGE Consortium forproviding the clonesfrom the normalized infant brain library; J. Murray for the chromosome 4 specific cosmid library and critical comments on this manuscript; J. Chamberlain for thegift of the microsatellite repeatprobes; L. Anderson for the 8 0 5 P- dystroglycan antibody; D. Jung for the 35 DAG antibody; S.A. Cullen forperforming the muscle biopsies; C. Leveille, J. Lee, S. Cutshall, D. Venzke, and D. Tischfield for their expert technical assistance; V Shefield for helpful consultation; L. Eberly and D. Moser of the University of Iowa DNA Core for their sequencing assistance; N. Chiannilkulchai, D. Fugman, F. Quetier, J. Tischfield, and all members from our respective labs for all their help. We are particularly indebted to all patients and their families. This research was supported by grants from the Muscular Dystrophy Association, L'Association Fran~aise contre les Myopathies (AFM), and the Groupement de Recherche Europken sur le Gknome. L.E.L. is supported by a grant from the Iowa Afiliate of the American Heart Association. VA. is funded by a grantfrom AFM. K.J?C. is an Investigator of the Howard Hughes Medical Institute. Received 31 July; accepted 4 October 1995. nature genetics volume 11 november 1995 article 1. Campbell, K.P. & Kahl, S.D. Association of dystrophin and an integral membrane glycoprotein. Nature 338,259-262 (1 989). 2. Emasti, J.M., Ohlendieck, K., Kahl, S.D., Gaver, M.G. &Campbell, K.P. Deficiency of a glycoprotein component of the dystrophin complex in dystrophic muscle. Nature 345.315-31. (1990). 3. Yoshida, M. & Ozawa, E. Glycoprotein complex anchoring dystrophin to sarcolemma. J. Biochem. 108,748-752 (1990). 4. Ewasti, J.M. &Campbell. K.P. Membrane organization of the dystmphin- glycoprotein complex. Cell66,1121-1131 (1991). 5. Hoffman, E.P., Brown, R.H., Jr. & Kunkel. L.M. Dystrophin: the protein product of the Duchenne muscular dystmphy locus. Cell 51, 919-928 (1 987). 6. Kwnig, M.. Monaco, A.P. & Kunkel, L.M. The complete sequence of dystrophin predicts a rod-shaped cytoskeletal protein. Cell 53, 219-228 (1 988). 7. Hemmings, L., Kuhlman, PA. & Critchley, D.R. Analysis of the actin- binding domain of a-actinin by mutagenesis and demonstration that dystrophin contains a functionally homologous domain. J. CellBiol. 116, 1369-1 380 (1 992). 8. Adams, M.E. etal. Two forms of mouse syntrophin, a 58 kd dystrophin- associated protein, differ in primary structure and tissue distribution. Neuron 11,531 -540 (1993). 9. Yang, B., Ibraghimov-Beskrovnaya, O., Moomaw, C.R., Slaughter, C.A. & Campbell, K.P. Heterogeneity of the 59 kDa dystrophin-associated protein revealed by cDNA cloning and expression. J. biol. Chem. 269, 6040-6044 (1 994). 10. Roberds, S.L., Anderson, R.D., Ibraghimov-Beskrovnaya, 0 . &Campbell, K.P. Primary structure and muscle-specific expression of the 50-kDa dystrophin-associated glycoprotein (adhalin). J. biol. Chem. 268,23739- 23742 (1 993). 11. Ibraghimov-Beskrovnaya, 0 . et a/. Primary structure of dystrophin- associated glycoproteins linking dystrophin to the extracellular matrix. Nature 356,696-702 (1992). 12. Ewasti. J.M. & Campbell, K.P. A role for the dystrophin-glycoprotein complex as a transmembrane linker between laminin and actin. J. Cell Biol. 122,809-823 (1 993). 13. Gee, S.H. et al. Laminin-binding protein 120 from brain is closely related to the dystrophin-associated glycoprotein, dystroglycan, and binds with high affinity to the major heparin-binding domain of laminin. J. biol. Chem. 268,14972-1 4980 (1 993). 14. Sunada, Y., Bernier, S.M., Kozak, C.A., Yamada. Y. & Campbell, K.P. Deficiency of merosin in dystrophic dy mice and genetic linkage of laminin M chain gene to dy locus. J. biol. Chem. 269, 13729-13732 (1 994). 15. Campbell, K.P. Three muscular dystrophies: loss of cytoskeleton- extracellular matrix linkage. Cell80,675-679 (1995). 16. Hoflman, E. & Kunkel, L.M. Dystrophin abnormalities in DuchenneIBecker muscular dystrophy. Neuron 2,101 9-1029 (1 989). 17. Roberds, S.L. et a/. Missense mutations in the adhalin gene linked to autosomal recessive muscular dystrophy. Cell78,625-633 (1994). 18. Piccolo. F. et a/. Primary adhalinopathy: a common cause of autosomal recessive muscular dystrophy of varying severity. Nature Genet. 10,243- 245 (1 995). 19. Matsumura, K. et a/. Deficiency of the 50K dystrophin-associated glycoprotein in severe childhood autosomal recessive muscular dystrophy. Nature 359,320-322 (1992). 20. Tomb, F.M.S. et a/. Congenital muscular dystrophy with merosin deficiency. C. R. Acad. Sci. Paris 317,351 -357 (1994). 21. Hillaire, D. et a/. Localization of merosin-negative congenital muscular dystrophy to chromosome 6q2 homozygosity mapping. Hum. molec. Genet. 3,1657-1661 (1994). 22. Stevenson, A.C. Muscular dystrophy in Northern Ireland. Ann. Eugen. 18, 50-91 (1 953). 23. Walton, J.N. & Nattrass. F.J. On the classification, natural history and treatment of the myopathies. Brain 77.169-231 (1954). 24. Emery, A.E.H. Population frequencies of inherited neuromuscular diseases - a w ~ l d survey. Neuromusc. Disord. I , 19-29 (1 991). 25. Speer, M.C. et a/. Confirmation of genetic heterogeneity in limb-girdle muscular dystrophy: linkage of an autosomal dominant form to chromosome 5q.Am. J. hum. Genet. 50,121 1-121 7 (1992). 26. Bashir, R. et al. A gene for autosomal recessive limb-girdle muscular dystrophy maps to chromosome 2p. Hum. molec. Genet. 3, 455-457 (1 994). 27. Ben Othmane, K. et al. Linkage of Tunisian autosomal recessive Duchenne-like muscular dystrophy to the pericentromeric region of chromosome 13q. Nature Genet. 2,315-317 (1992). 28. Azibi, K. eta!. Severechildhwd autosomal recessivemuscular dystrophy with the deficiency of the 50 kDa dystrophin-associated glycoprotein maps to chromosome 13q12. Hum. molec. Genet. 2,1423-28 (1 993). 29. Beckmann, J.S. etal. Agene for limb-girdle muscular dystrophy maps to chromosome 15 by linkage. C. R. Acad. Sci. Paris312,141-148 (1991). 30. Richard, I. et a/. Mutations in the proteolytic enzyme, calpain 3, cause limb girdle muscular dystrophy type 2A. Cel187.27-40 (1995). 31. Jackson, C.E. &Carey, J.H. Progressive muscular dystrophy: autosomal recessive type. Pediatrics 28,77-84 (1 961). 32. Jackson. C.E. 8 Strehler, D.A. Limb-girdle muscular dystmphy: clinical manifestations and detection of preclinical disease. Pediatrics 41, 495- 502 (1 968). 33. Allamand, V. et a/. Genetic heterogeneity of autosomal recessive limb- girdle muscular dystrophy in a genetic isolate (Amish) and evidence for a new locus. Hum. molec. Genet. 4,459-464 (1995). 34. Mizuno, Y et a/. Selective defect of sarcoglycan complex in severe childhood autosomal recessive muscular dystrophy muscle. Biochem. biophys. Res. Commun. 203,979-983 (1994). 35. Bannemann, C.G. et a!. p-sarcoglycan (A3b) mutations causeautosomal recessive muscular dystmphy with loss of the sarcoglycan complex. Nature Genet. 11,266-273 (1 995). 36. Altschul, S.F., Gish, W., Miller, W., Myers, E.W. & Lipman, D.J. Basic local alignment search tool. J. molec. Biol. 215.403-410 (1990). 37. Bento Soares, M. etal. Construction and characterization of a normalized cDNA libraly. Proc. natn. Acad. Sci. U.S.A. 91,9228-9232 (1 994). 38. Mills, K.A. etal. Genetic and physical maps of human chromosome 4 based on dinucleotide repeats. Genomics 14,209-21 9 (1 992). 39. Gyapay, G. eta/. The 1993-94 Gbnethon human genetic linkage map. Nature Genet. 7,246-329 (1 994). 40. Lander, E.S. & Botstein, D. Homozygosity mapping: a way to map human recessive traits with the DNA of inbred children. Science 236,1567-1570 (1 987). 41. Cohen, D., Chumakov, I. & Weissenbach, J. A first generation physical map of the human genome. Nature 366,698-701 (1993). 42. Don, R.H.. Cox, P.T., Wainwright, B.J., Baker, K. 8 Mattick, J.S. "Touchdown" PCR to circumvent spurious priming during gene amplification. Nucl. AcidsRes. 19,4008 (1991). 43. Emasti, J.M., Kahl, S.D. &Campbell, K.P. Purification of dystrophin from skeletal muscle. J. biol. Chem. 266,9161 -9165 (1991). 44. Yoshida, M., Suzuki, A., Yamamoto, H., Mizuno, Y. & Ozawa, E. Dissociation of the complex of dystrophin and Itsassociated proteins into several unique groups by n-octyl P-D-glucoside. Eur: J. Biochem. 222, 1055-1 061 (1 994) 45. Roberds, S.L. etal. Disruption of the dystmphin-glycoprotein complex in the cardiomyopathic hamster. J. biol. Chem. 268.11496-1 1499 (1993). 46. Miwno, Y et al. Sarcoglycan complex is selectively lost in dystrophic hamster muscle. Am. J. Path. 146,530-536 (1 995). 47. Roberds, S.L. &Campbell, K.P. Adhalin mRNA and cDNAsequence are normal in the cardiomyopathic hamster. FEES Lett. 364,245-249 (1995). 48. Campanelli, J.T., Roberds, S.L., Campbell, K.P. &Scheller, R.H.Arolefor dystrophin-associated glycoproteins and utrophin in agrin-induced AChR clustering. Cell77,663-674 (1994). 49. Gee, S.H.. Montanan, F., Lindenbaum. M.H. & Carbonetto, S. Dystroglycan-a, a dystrophin-associated glycoprotein, is a functional agrin receptor. Cell 77.675-686 (1 994). 50. Smith, D.B. & Johnson, K.S. Single-step purification of polypeptides expressed in Escherichia coli as fusions with glutathione S-transferase. Gene 67,31-40 (1 988). 51. Smith, D.B. & Corcoran, L.M. In Current Protocols in Molecular Biology, (eds Ausubel, EM. et a / . ) 16.7.1-16.7.8. (Current Protocols, Brooklyn, New York, 1987). 52. Jorgensen, A.O. et a/. Identification of novel proteins unique to either transverse tubules (TS28) or the sarcolemma (SL50) in rabbit skeletal muscle. J. CellBiol. 110.11 73-1 185 (1990). 53. Ohlendieck, K., E~aSti, J.M.. Snook, J.B. &Campbell. K.P. Dystrophin- glycoprotein complex is highly enriched in isolated skeletal muscle sarcolemma. J. CeIlBiol. 112,135-148(1991). 54. Laemmli, U.K. Cleavage of structural proteins during the assembly of the head of bacteriophageT4. Nature 227,680-685 (1970). 55. Towbin, H., Staehelin, T. &Gordon, J. Electrophoretic transfer of proteins from polyacrylamide gels to nitrocellulose sheets: procedures and some applications. Proc. natn. Acad. Sci. U.S.A. 76,4350-4354 (1979). 56. Fardeau, M. et a/. Juvenile limb-girdle muscular dystrophy: clinical. histopathological, and genetics data on a small community living in the Reunion Island. Brain (in the press). 57. Murray, J.C. et a/. A comprehensive human linkage map with centimorgan density. Science 265,2049-2054 (1 994). 58. Fougerousse F. et al. Mapping of a chromosome 15 region involved in limb-girdle muscular dystrophy. Hum. molec. Genet. 3.285-293 (1 994). 59. Lathrop G.M.. Lalwel J.-M.. Julier, C. & Ott, J. Multilocus linkage analysis in humans: detection of linkage and estimation of recombination. Am. J. hum. Genet. 37,482-498 (1985). nature genetics volume 11 novernber 1995