PII: S0960-8966(97)00135-1


b-Sarcoglycan: genomic analysis and identification of a novel missense
mutation in the LGMD2E Amish isolate

F. Duclos 1a, O. Broux1 b, N. Bourg1 b, V. Straub a, G.L. Feldman c, Y. Sunada a, L.E. Lim a,
F. Piccolo d, S. Cutshall a, F. Gary b, F. Quetier b, J.-C. Kaplan d,

C.E. Jacksonc, J.S. Beckmann b , e, K.P. Campbell a ,*
aHoward Hughes Medical Institute, Department of Physiology and Biophysics and Department of Neurology,

University of Iowa College of Medicine, Iowa City, IA 52242, USA
bCNRS URA 1922, France, Généthon, 91000 Evry, France

cHenry Ford Hospital, Department of Medical Genetics, Detroit, MI 48202, USA
dHopital Cochin Maternités, INSERM U129, 75014 Paris, France

eFondation Jean Dausset-CEPH, 75010 Paris, France

Received 11 August 1997; revised version received 30 October 1997; accepted 12 November 1997

Abstract

The sarcoglycan complex is involved in the etiology of four autosomal recessive limb-girdle muscular dystrophies (LGMD2C–F). A
missense mutation (T151R) in the b-sarcoglycan gene on chromosome 4q12 has been shown to cause a mild form of LGMD2E in 11
families from a Southern Indiana Amish community sharing a common haplotype. We now report that two sibs from another Amish family
with mild LGMD2E are compound heterozygotes for chromosome 4q12 markers. In order to characterize the genetic defect in this new
family, we determined the genomic organization of the b-sarcoglycan gene. A second missense mutation (R91C) has now been identified in
this LGMD2E Amish family. This mutation is also present in the homozygous state in another family of probable Amish ancestry. Finally,
analysis of all the components of the dystrophin-glycoprotein complex was performed for the first time on a biopsy from a patient
homozygous for the b-sarcoglycan mutation (T151R). Interestingly, in addition to the loss of the entire sarcoglycan complex, we detected
a reduction of a-dystroglycan which suggests a role for the sarcoglycan complex in stabilizing a-dystroglycan at the sarcolemma.  1998
Elsevier Science B.V.

Keywords: Limb-girdle muscular dystrophy; b-Sarcoglycan; Sarcoglycan complex; a-Dystroglycan

1. Introduction

In the dystrophin-glycoprotein complex (DGC) [1–4], b-
sarcoglycan, a 43 kDa glycoprotein, is closely associated
with a-, g-, and d-sarcoglycan [5–7]. This subcomplex
can be distinguished from a-, and b-dystroglycan [8], the
syntrophins [9,10] and from dystrophin [11] by differential
solubilization [6] or by immunoprecipitation [12]. Together,
the DGC confers a structural link between laminin 2 in the
extracellular matrix, via a-dystroglycan, and the cytoskele-

ton, via dystrophin, and is believed to protect muscle cells
from contraction-induced damage [13,14].

The DGC has been implicated in several forms of mus-
cular dystrophy. In Duchenne muscular dystrophy (DMD),
mutations in the dystrophin gene result in a severe dys-
trophic phenotype [2,11], whereas in Becker muscular dys-
trophy, dystrophin mutations result in a dysfunctional
dystrophin and are associated with a milder phenotype
[15]. Severe and mild forms of congenital muscular dystro-
phy have recently been characterized by mutations in the
laminin a2-chain gene [16–18].

Limb-girdle muscular dystrophies (LGMDs) represent a
clinically and genetically heterogeneous group of diseases
which are characterized by progressive weakness of the
pelvic and shoulder girdle muscles [19–21]. The genes

Neuromuscular Disorders 8 (1998) 30–38

0960-8966/98/$19.00  1998 Elsevier Science B.V. All rights reserved
P I I S 0 9 6 0 - 8 9 6 6 ( 9 7 ) 0 0 1 3 5 - 1

* Corresponding author. Tel.:+1 319 3357867; fax: +1 319 3356957;
e-mail: kevin-campbell@uiowa.edu

1 These authors contributed equally to this work.



responsible for LGMD2C–F have been identified and
encode for g-sarcoglycan [22], a-sarcoglycan [23], b-sarco-
glycan [24,25], and d-sarcoglycan [26,27], respectively. In
LGMD2C–F caused by sarcoglycan mutations, a mutation
in one of the sarcoglycan proteins leads to the concomitant
loss or reduction of the other sarcoglycans at the sarco-
lemma [7,22–32].

To facilitate the analysis of LGMD2E patients and to
characterize the b-sarcoglycan gene, we performed an
extensive analysis of 36 kb of genomic sequences overlap-
ping the b-sarcoglycan gene on chromosome 4q12. Using
various predictive algorithms, we analyzed the entire geno-
mic sequence and detected two putative exons that could be
part of tissue specific isoforms of b-sarcoglycan. We have
also found several additional polyadenylation sites that may
account for the two transcripts of 3.0 and 4.4 kb which have
not been yet characterized. A second LGMD2E mutation
was identified in an Amish family with two affected sibs
carrying the common haplotype [24] at the heterozygous
state. This second mutation was also present in the homo-
zygous state in patients belonging to another family with
probable Amish origins. In addition, we screened for a-
sarcoglycan deficiency 107 cases diagnosed with myopathy,
primarily characterized with normal immunohistochemical
findings for dystrophin. We identified 10 sporadic cases
with these criteria and screened them for b-sarcoglycan
mutations. A novel missense mutation I119F was identified
in the heterozygous state in the b-sarcoglycan gene of one of
these patients. Of particular interest, we demonstrated with
immunofluorescence analysis, a reduction in the expression
level of both the sarcoglycan complex and of a-dystrogly-
can at the sarcolemma of one individual carrying a homo-
zygous b-sarcoglycan T151R mutation.

2. Materials and methods

2.1. Subcloning and sequencing of cosmid Cos4

DNA of cosmid Cos4 (cloning vector, Supercos 1) was
sonically sheared and subcloned into a M13 vector. Insert
DNA was amplified by PCR under standard conditions
using 1 ml of overnight culture in a 50 ml total volume
containing the M13 forward and M13 reverse primers.
After a 5 min hot start at 94°C, 2 units of Taq Polymerase
(Perkin-Elmer) were added, followed by 35 cycles of 60 s
denaturation at 94°C, 60 s annealing at 55°C, and 90 s
extension at 72°C. The PCR amplification products were
analyzed by agarose gel electrophoresis and only those dis-
playing an insert size higher than 150 bp were sequenced on
one strand of DNA for sizes ,600 bp or on both DNA
strands for size .600 bp.

2.1.1. EcoRI/NotI restriction map of Cos4
Restriction fragments of Cos4 obtained upon EcoRI and

NotI digestion were separated by agarose gel electrophor-

esis and capillary transferred to (Amersham) nylon mem-
brane and hybridized with T3 and T7 primers for mapping.

2.1.2. Sequence analysis
The sequences were analyzed and alignments performed

using XBAP software of the STADEN package (version
94.2). Gaps between sequence contigs were filled by walk-
ing with internal primers and by sequencing both ends of
several EcoRI-EcoRI, NotI-NotI and EcoRI-NotI restriction
fragments. The total consensus sequence was scanned using
the following software programs: Grail2a and XGrail2a
(version 1.3b); FEXH, HEXON, FGENEH, TSSG, TSSW
of the BCM Gene finder package (version 2.10.5.94); and
Promoter Scan program (version 1.7). Accession number
EMBL# Y09781.

2.2. Immunofluorescence analysis

Amish patients were ascertained through the Henry Ford
Hospital, at the Department of Medical Genetics, Detroit,
MI, USA. Skeletal muscle tissue from LGMD patients were
obtained from diagnostic biopsies. Human control muscle
was received from patients undergoing orthopedic surgery
not related to neuromuscular disorders. Antibody staining
on 7 mm transverse cryosections was performed at room
temperature. Sections were treated with AB blocking solu-
tions (Vector), blocked with 3% bovine serum albumin
(BSA) in phosphate-buffered saline (PBS) for 30 min, and
then incubated with a dilution of the primary antibody for
90 min. Antibodies against the following components of the
DGC were tested: dystrophin, laminin a2-chain, a-, b-, g-,
d-sarcoglycan, and a- and b-dystroglycan. Sections were
then treated and analyzed as previously described [24].
Monoclonal antibodies VIA42 against dystrophin and
IVD31 against a-sarcoglycan were previously characterized
[2,33]. Monoclonal antibody 5B1 against b-sarcoglycan and
21B5 against g-sarcoglycan were produced in collaboration
with Dr. Louise Anderson (Muscular Dystrophy Research
Labs, Newcastle General Hospital, Newcastle upon Tyne,
NE4 6BE, UK). Monoclonal antibody 8D5 against b-dys-
troglycan was kindly provided by Dr. Louise Anderson. An
affinity-purified d-sarcoglycan antibody (rabbit 215) was
produced against the N-terminal sequence (MMPQE-
QYTHHRSTMPGM) of human d-sarcoglycan. This syn-
thetic peptide (Research Genetics) with an additional
cysteine at the N terminus was conjugated with the N-term-
inal cysteine of keyhole limpet hemocyanin (Pierce) using
m-maleimidobenzoic acid-N-hydroxysuccinimide ester
(Pierce), mixed with Freund’s adjuvant (Sigma) and
injected into rabbit 215. Polyclonal antibody against the
peptide was affinity purified from crude sera using BSA-
conjugated peptide as described previously [12]. Antibodies
against different components of the dystrophin-glycoprotein
complex were produced in a goat (Goat 20) using the pur-
ified rabbit skeletal muscle DGC [2] as previously described
[27]. A specific antibody against a-dystroglycan was affi-

31F. Duclos et al. / Neuromuscular Disorders 8 (1998) 30–38



nity-purified from Goat 20 antiserum using Immobilon P
(Millipore) strips of a-dystroglycan fusion protein-D [8].
Monoclonal anti-human laminin a2 chain antibody (clone
5H2) was purchased from Chemicon.

2.3. Mutation screening of b-sarcoglycan gene

For b-sarcoglycan cDNA fragments were amplified fol-
lowing reverse-transcription from total RNA prepared from
biceps brachii muscle biopsies of patients [24]. Primer
sequences used for these experiments are reported in
Table 1. The RT-PCR products were directly sequenced
and mutations confirmed by allele specific oligonucleotide
hybridization or restriction digest using Hinf1 due to the
R91C mutation leading to the loss of a restriction site for

this enzyme. Exons 2-6 of the b-sarcoglycan gene including
flanking intronic sequences were amplified from patients’
DNA and sequenced. The absence of variants in 100 control
chromosomes from the normal population was tested on
DNA from the CEPH reference families.

3. Results

3.1. b-Sarcoglycan gene: exon-intron structure and
genomic sequence

To facilitate the mutation search and the identification of
regulatory domains in the b-sarcoglycan gene, we have
determined its genomic sequence and structure. We had

Table 1

Primers used for RT-PCR and genomic analysis of b-sarcoglycan

Gene/fragment Sequences of primers Length (bp) Annealing T°C

Exon 1–3 BF0: GGA.CAG.TCG.GGC.GGG.GA 365 58
(RTPCR) BR0: CTT. TGT.TGT.CCC.TTG CTG.A
Exon 2–6 BF9: AGT.TCC.AAT.GGT.CCT.GTA.AAG.A 1085 58
(RTPCR) BR8: TTA.GGC.TCT.CTG.AGA.AGA.TT
Exon 2 BF9: AGT.TCC.AAT.GGT.CCT.GTA.AAG.A 286 58

BR1: AAT.AAA.GTC.TAT.GAT.TTC.ACT.ATA
Exon 3 BF3: GGC.CTA.CCA.ACA.GCG.GAC 316 55

BR3: AAT.ATC.TCT.TTC.CAA.TAA.TCA.ACT.AG
Exon 4 BF4: TTC.AGG.AAT.TTG.TTT.GCA.GTC.T 205 54

BR4: ACT.CTA.GAG.AAT.AAT.TCT.CTC
Exon 5 BF5: TCT.GAT.AAC.AAT.TCT.CAC.ATT.AAT 254 54

BR5: TAT.GGA.TTT.ATG.TAC.CCA.AGA.A
Exon 6 BF6: CCT.CAT.GGA.GGT.AAA.TAA.ATC 291 55

BR6: AAG.TCA.AGT.CAA.GAT.ATA.AA

Primers designed to amplify b-sarcoglycan cDNA or b-sarcoglycan exons are indicated with the annealing temperature used for PCR amplification of the
corresponding fragment. Positions are indicated according to the sequence reported for b-sarcoglycan cDNA (GenBank U29586).

Table 2

Analyses of the human b-sarcoglycan genomic sequence

Splice acceptor Score
(%)a

Exon
(size, bp)

Exon~position Splice donor Score
(%)a

Intron
(size, bp)

Power of prediction algorithmsc

Grail2a FEXH HEXON

Exon-intron organization of the b-sarcoglycan gene corresponding to the cDNA
1 (75) 6958–7032 AACAGgtctgt 70.43 1 (4569) Med Good –

tatttcatttttagCAAAG 84.00 2 (210) 11602–11811 TAATAgtgagt 76.46 2 (3546) Exc Exc Exc
GtgttcttttcagATAAC 93.34 3 (186) 15358–15543 AGCCTgtaagt 79.00 3 (756) Exc Med Med
TgttttttttgaagATTGT 84.79 4 (192) 16300–16491 AAAGGgtatga 75.18 4 (629) Good – –
ttttaaattttcagATTAC 90.54 5 (132) 17121–17252 AGGCGgtgagt 88.50 5 (3800) Exc Exc Exc
TattctcctaatagGAAAA 88.70 6 (201) 21053–21256 Exc Good –

Additional predicted exonsb

actgtattgtttagGGAAT 75.10 A (207) 10829–11035 TACAGgtgcgt 85.22 – – Good Good
tttcctttttccagGCATT 97.81 B (501) 27595–28095 CAAGGgtgagt 88.86 – Exc Exc Exc

~ Position 1 corresponds to the first nucleotide on the sense strand of the sequenced cosmid Cos4 (Accession # Y03781).
aA score expressing adherence to the consensus was calculated for each size according to Shapiro and Senapathy [34].
bOnly those exons showing at least one Excellent or two Good scores are listed.
cMed, Good, Exc and – refer to mediocre, good, excellent and no detected scores, respectively.
The values given by BCM were arbitrarily translated into qualitative scores as follows, exc ≥ 9; 5.5 . good . 9; med ≤ 5.5. The scores listed for Grail2a
correspond to the best score obtained using either Grail2a and XGrail2a (version 1.3b).

32 F. Duclos et al. / Neuromuscular Disorders 8 (1998) 30–38



previously isolated a cosmid (Cos4) for the b-sarcoglycan
gene [24]. The Cos4 insert (nucleotides 1–36 159; Acces-
sion EMBL# Y09781) was entirely sequenced and it will be
referred to hereafter as the sense strand. The consensus
sequence has been determined with an average 9× coverage
value (minimal local value 3×). The genomic organization
of the b-sarcoglycan gene was determined from comparison
of the published cDNA sequences [24,25] with the genomic
sequence using several prediction algorithms. A comparison
of the relative power of the latter demonstrated the benefits
that can be obtained in the absence of prior exon informa-
tion by the use of all existing algorithms: individually, none
of these programs were able to identify all six exons. Since
the 5′ stretch of the cDNA published by Bönnemann et al.
[25] extended upstream of that of Lim et al. [24], the former
was taken into account. Furthermore, local discrepancies in

the translated parts were easily ascertained by genomic
sequencing (6× coverage in these regions).

The human b-sarcoglycan gene (Table 2 and Fig. 1) is
organized into 6 exons, which range in size from 75 to 210
bp, separated by 5 introns (from 629 to 4569 bp). All donor
and acceptor splicing sequences display the GT/AG consen-
sus. The translation start codon is the ATG at position 7000,
and the stop codon at position 21 254 is followed by several
putative polyadenylation sites (Fig. 2) which could account
for several transcript isoforms. The possible biological roles
of these isoforms either in regulating the stability or addres-
sing of the mRNA [35] remain to be elucidated. Analysis of
the public domain expressed sequence tag (EST) databases
and of the Northern blot expression patterns [24] bring
further credence to the biological relevance of these alter-
native polyadenylation sites. Indeed, we reported the pre-

Fig. 1. Exon-intron structure of the b-sarcoglycan gene. EcoRI restriction map of Cos4 DNA was determined. Size in bases of EcoRI fragment is indicated at
the top of the cosmid DNA box. Below is the exon-intron organization of the b-sarcoglycan gene. The genomic structure was inferred from comparison with
the published cDNA sequences [24,25] and upon analyses of the genomic sequence with several algorithms. The human b-sarcoglycan gene is organized in 6
exons (none of the programs identified the six) whose size range from 75 to 210 bp, separated by 5 introns (from 629 to 4569 bp).

Fig. 2. Analysis of the 3′ untranslated region of the b-sarcoglycan gene. The top lane represents the Cos4 DNA region corresponding to the 3′-UTR of the b-
sarcoglycan gene. The scale is in kilobase. The different putative polyadenylation sites AATAAA, the stop codon in exon 6 and predicted exon B are
indicated with their position. The arrows represent ESTs and their orientation with the name of the cDNA clones from which they were obtained. The position
of the 4.4 and 1.35 kb transcripts detected by Northern blot [24] are shown.

33F. Duclos et al. / Neuromuscular Disorders 8 (1998) 30–38



sence of mRNAs of 4.4, 3.0 and 1.35 kb, respectively [24].
The 4.4 and 1.35 kb molecules could result from the use of
these alternative polyadenylation sites (Fig. 2), whereas the
3.0 kb transcript still remains unaccounted for. The latter
may correspond to the 3′ end EST of the cDNA clones
197959 and 230147 (see Fig. 2), although no polyadenyla-
tion site was identified in the corresponding genomic
sequence.

Use of the Promoter Scan algorithm predicted a promo-
ter-like structure (from nucleotide 6946 to 7195; score
18.18) upstream of the gene, containing a GC box (position
−41 with respect to the initiator ATG), and spanning a 250
bp long CpG rich island (82%) covering exon 1 and part of
intron 1. These deductions were in agreement with the pre-
dictions of the TSSW program of BCM Gene finder (Human
Genome Center, Baylor College of Medicine, Houston, TX,
USA).

Three prediction algorithms, Grail 2A, FEXH BCM and
HEXON BCM were applied to the genomic DNA
sequences; Table 2 shows that none of them detected all
known exons with at least a good score, and that several
exons were not detected by one or two of these algorithms.
In addition, these programs also revealed additional poten-
tial exons, two of which had good or excellent scores,
respectively. Exon A was ranked good by two algorithms
and mapped between exon 1 and exon 2 and would lead, if
authentic, to another isoform of this gene. However, its
location within an Alu area and its content in repetitive
DNA diminished the likelihood that it is translated. In con-
trast, exon B was scored excellent by the three algorithms.
This 501 bp long exon B lies downstream of the recognized
b-sarcoglycan cDNA sequence on the same strand (Fig. 2).
It could therefore correspond to an as yet undocumented
isoform. Alternatively, it could be part of another gene,
although no promoter or initiating ATG was significantly
identified in silicio. To test these hypotheses, Northern blot
as well as RT-PCR experiments on human skeletal, cardiac
and brain mRNA were conducted. All tests performed to
visualize the corresponding transcripts were negative (data
not shown). Either the corresponding transcript was present
at too low levels to be detected, or its restricted spatio-tem-
poral pattern of expression prevented its detection. Whereas
no homologous amino-acid sequence was recognized upon
screening public domain (TIGR, Genbank, EST, C.elegans
and yeast) databases, two regions showed amino-acid
homology with sequences internal to the b-sarcoglycan
exon 4. The most significant was the RxxNILF motif
which is also found in Mycoplasma (Accession: U02188);
Escherichia coli (Accession: AE000317), and Chloroplast
30S ribosomal protein (e.g. Accession: P36471). Although
this homology does not imply an actual functional role for
exon B, it seems likely that the latter is somehow related to
the b-sarcoglycan gene. Given the difficulty to assess
whether exon B is an authentic exon, it may be worthwhile
to investigate the corresponding sequence in other mam-
mals. Additionally, one exon was scored excellent with pro-

grams FEXH, HEXON and FGENEH of BCM Gene Finder
and good with Grail 2A (reverse position: 25 750–25 805,
data not shown), on the opposite strand of the cosmid insert.
This putative ‘orphan’ exon showed no hit with any pub-
lished sequence and hence was not integrated onto the map
(Table 2).

3.2. b-Sarcoglycan gene mutations in the LGMD2E Amish
community

In agreement with the common haplotype displayed by
the first 11 families analyzed, a missense mutation (T151R)
common to all Southern Indiana Amish patients tested has
been found and was associated with a mild form of muscular
dystrophy [24]. No evidence of cardiomyopathy has been
found, and joint contractures have been infrequent in these
patients. Since then, two sibs from an Amish family exhibit-
ing heterozygosity for the common haplotype with chromo-
some 4q12 markers and a second distinct haplotype were
identified among the community of Southern Indiana. Sub-
sequently, mutation screening of these patients revealed in
addition to the common T151R mutation, a C to T transition
corresponding to the substitution of arginine 91 by a
cysteine. These two patients (2a and 2b) had a late onset
of the disease (Table 3), characterized by muscle weakness
and enlargement of calves. At age 52 and 57 years, respec-
tively, both individuals were still able to carry on work
requiring physical exertion and were still able to walk.
Five other children in the family have not been found to
be affected. It is unclear whether the unaffected members
of this family are carriers of mutations (i.e. a penetrance
issue) or just unaffected genetically (i.e. carrying one or
two copies of the normal allele). Surprisingly, two other
familial cases not belonging to the Amish community car-
ried this mutation in the homozygous state (Table 3, cases
3a and 3b). No relationship was ascertained between the
father and mother, however Amish names were reported
in their ancestry. Patient 3a and her affected sister (3b)
presented leg cramps and weakness after exercise at age 4
and 11 years, respectively. Calf hypertrophy was noted in
early childhood in both patients. Patient 3a was still able to
ambulate at age 40 years whereas her affected sister has
used a walker since age 37 years. They have two unaffected
brothers. The status of the R91C mutation was analyzed by
restriction digest of PCR amplified genomic DNA using
Hinf1, which is ablated as a result of the mutation. Analysis
of 100 chromosomes from the CEPH reference families
confirmed that the R91C was a mutation and not a poly-
morphism.

3.3. Analysis of the expression of the components of the
DGC

Since the original characterization of the b-sarcoglycan
gene defect in 11 LGMD2E Amish families from Southern
Indiana [24], two additional sarcoglycan genes have been

34 F. Duclos et al. / Neuromuscular Disorders 8 (1998) 30–38



cloned and new antibodies developed against these mole-
cules. We examined the expression of the sarcoglycan and
dystroglycan complexes, dystrophin, and the a2 chain of
laminin 2 in a biopsy from an Amish patient homozygote
for the T151R mutation. This analysis extended previous
examinations [24,25,30] by using a monoclonal antibody
against g-, and d-sarcoglycan. In addition, a newly affinity
purified anti-a-dystroglycan antibody was developed. The
antibody against a-dystroglycan allowed us for the first time
to specifically look at the expression of one of the key
proteins of the DGC on skeletal muscle cryosections of
LGMD patients. The expression level of dystrophin, b-dys-
troglycan, and the laminin a2 chain in the Amish patient
was comparable to the levels found in normal human control
muscle. On the other hand, expression of the sarcoglycan
complex was strongly reduced in the patient’s biopsy com-
pared to control muscle. All sarcoglycan proteins showed a
similar level of reduction at the sarcolemma. In addition to
the secondary loss of a-, g-, and d-sarcoglycan, we also
found a reduced expression of the a-subunit of the dystro-
glycan complex. This unexpected and interesting finding
may unveil for the first time a direct or indirect link between
the sarcoglycan complex and a-dystroglycan.

3.4. Patient screening for b-sarcoglycan gene mutations

In order to identify additional patients carrying a primary
b-sarcoglycan deficiency, we examined a total of 107 new
independent cases of autosomal recessive myopathy char-
acterized by normal immunostaining of dystrophin and
markedly elevated serum creatine kinase level. We primar-
ily screened these patients by immunostaining of skeletal
muscle cryosections for a-sarcoglycan deficiency and iden-
tified 10 new cases. All these sporadic cases originate in the
USA. In all patients, the serum creatine kinase levels were
elevated but the clinical pattern was of variable severity,
ranging from severe childhood to late onset progressive
muscular dystrophy (data not shown). There were no intel-
lectual impairments, nor cardiomyopathy. Skeletal muscle
sections of these patients were further analyzed by immuno-
fluorescence staining for the other components of the DGC
and a concomitant loss or a drastic reduction of b, g and d-
sarcoglycan was observed (data not shown). The immuno-
fluorescence data did not indicate the primary defect caus-
ing the observed phenotype.

All sporadic LGMD cases were analyzed for mutations in
the b-sarcoglycan gene as well as in the other known sar-
coglycan genes (data not shown), due to the absence of
linkage data. For one of the sporadic cases (T044), only
one mutant allele corresponding to the substitution of an
isoleucine by a phenylalanine at position 119 (I119F), was
found in the b-sarcoglycan gene although all coding
sequences were investigated. This patient developed a prox-
imal muscle weakness as well as progressive scoliosis with
onset in early childhood. At age 13 years, he was confined to
a wheelchair and presented more severe clinical symptoms

as compared to the Amish patients. In this isolated case, the
second mutation could be in the unexplored untranslated
region or possibly in one of the alternative exons. Allele
specific hybridization with either a control oligonucleotide
or an oligonucleotide including this latter mutation,
revealed that this substitution was not found in 100 inde-
pendent chromosomes from CEPH reference families (data
not shown) and that the sporadic LGMD patient was thus in
all likelihood a compound heterozygote for the I119F and a
second, as yet unidentified mutation. We also determined
the primary sarcoglycan gene defect in 7 out of 9 remaining
isolated cases: four patients have a mutated a-sarcoglycan
gene and three have a mutated g-sarcoglycan gene (data not
shown).

4. Discussion

In order to facilitate further gene expression analysis and
the identification of functional regulatory elements, we per-
formed sequence analysis of a 36 kb genomic fragment
containing the b-sarcoglycan gene. Interestingly, several
putative additional exons were predicted by the computer
programs XGrail and/or BCM in the 3′-UTR of the b-sar-
coglycan gene, which may be part of different tissue-speci-
fic transcripts. One GC box was detected at position 6952
using Promoter scan, and several polyadenylation sites were
found (Table 2 and Fig. 2). The relationship between these
putative polyadenylation sites and the different transcripts
detected by Northern hybridizations at 4.4, 3.3, 1.3–1.4 kb
requires further probing with specific 3′-UTR fragments to
be ascertained. The tissue-specificity of expression should
also be determined. Although experiments to confirm the
existence of the two predicted exons gave negative results
it was not possible to conclude whether these two exons
form part of real mRNA. This study constitutes a good
example of what to expect from the human genome sequen-
cing project, since it illustrates how far one can get inferring
relevant genomic information just from sequencing data. On
the one hand it highlights some of the limits of the current
prediction algorithms, and on the other hand it points also to
the presence of potential false positives.

Although a common chromosome 4q11–q12 haplotype is
displayed by most of the Southern Indiana Amish LGMD2E
families around the disease locus [24], two patients from
one family were compound heterozygotes in this region.
Mutation screening revealed that these patients carried the
common T151R mutation [24] together with another mis-
sense mutation, R91C. In addition, we identified the same
mutation (R91C) in an unrelated familial case (Table 3,
Amish 3) who does not belong to the Southern Indiana
Amish community, but whose ancestors have Amish
names on both paternal and maternal sides. The presence
of a second mutation in the Amish community was unex-
pected, although the presence of multiple mutations in a
small ethnic community has already been observed before.

35F. Duclos et al. / Neuromuscular Disorders 8 (1998) 30–38



The presence of several mutations in the calpain 3 gene
among LGMD2A patients from La Réunion Island was puz-
zling [36]. A hypothesis that could account for this and other
related observations has been reported [36–38]. The identi-
fication of a second mutation in another family of probable
Amish origin suggests a relatively ancient occurrence of this
mutation, in contrast to the possibility of a high mutation
rate coupled with a recent occurrence of these mutations in
this isolate as proposed by Zlotogora et al. [39] for Hurler
syndrome, metachromatic leukodystrophy and LGMD2A.
In such highly inbred populations, one mutation associated
with a founder haplotype may become frequent by result of
a genetic drift. This frequent haplotype/mutation can now
uncover other rare events leading possibly to these multiple
disease alleles we observe now in the Amish isolate from
Southern Indiana. These homozygotes (R91C; codon CGC
mutated to TGC) displayed a mild muscular dystrophy phe-
notype (Table 3, Amish 3a and 3b). Surprisingly,
Bönnemann et al. [30] reported a R91P mutation (codon
CGC changed to CCC) in three homozygous sibs from a
Brazilian family associated with a severe form of LGMD.
Secondary structure prediction analysis suggested that the
mutation R91C did not affect the protein structure whereas
the proline at position 91 did. These findings suggest that the
arginine 91 in b-sarcoglycan may constitute a critical site of
interaction in the complex. These homozygote cases may be
helpful in unraveling some genotype/phenotype correlations
in LGMD2E and in providing clues on the sarcoglycan/
dystroglycan interaction.

We have also analyzed 107 patients with normal dystro-
phin and identified 10 sporadic cases with a sarcoglycan
complex deficiency. These cases represent 10% of the mus-
cular dystrophy patients tested and this correlates with the
ratio obtained by a recent study [31]. These patients were
investigated for sarcoglycan mutations in the a-, b-, g- and
d-sarcoglycan genes and a sarcoglycan molecular defect
was revealed in 8 out of 10 (data not shown). These two
latter may carry mutations in another yet unidentified sar-

coglycan gene. We identified one LGMD patient (Table 3,
case T044) carrying on one chromosome a novel missense
mutation (I119F) in the b-sarcoglycan gene, the second
allele remaining unidentified despite the fact that all coding
sequences were investigated. Interestingly, many of the
T151R and R91C Amish cases are associated with a milder
course of the disease in contrast to the severe phenotype of
LGMD2E patients reported [25,30] and one LGMD2E case
(T044) reported in this study. All mutations mapped to the
extracellular domain of the protein. The b-sarcoglycan gene
mutation R91C (CGC to TGC) abolished a CpG site. The
functional importance of the region in which the mutation
occurred has not yet been defined. On the other hand, mis-
sense mutations can cause abnormal protein processing or
stability, as was seen in the majority of cases with LGMD2
[22–32]. In LGMD2D, a broad spectrum of missense muta-
tions in the a-sarcoglycan gene have been reported which
affected protein targeting to the sarcolemma and resulted in
muscular dystrophy of varying degrees of severity [29,32].
This phenomenon has already been observed with other
transmembrane proteins, including CFTR [40].

The close association of the components of the sarco-
glycan complex [6,12] and their concomitant loss in
LGMD2C–F have been reported. In the BIO14.6 cardio-
myopathic hamster, an animal model of d-sarcoglycan defi-
cient LGMD [41], a secondary reduction of a-dystroglycan
expression had been reported [42]. To test whether the sta-
bility of a-dystroglycan is also affected in human sarcogly-
can deficient skeletal muscle, we used a new anti-a-
dystroglycan antibody and demonstrated for the first time
that a missense mutation in the b-sarcoglycan gene (T151R)
can modify a-dystroglycan expression (Fig. 3). This result
suggests a possible link between the sarcoglycan complex
and the peripheral membrane protein a-dystroglycan. This
link could either be mediated through direct interaction or
via an intermediate. The maintenance of b-dystroglycan
expression in the sarcolemma of this patient supports the
hypothesis that the sarcoglycan complex is required for high

Table 3

Clinical and biochemical data of patients with LGMD

Case no. Age (years) Sex Onset (years) CK value
at (age)

Calf muscle
hypertrophy

Activity b-Sarcoglycan
gene defect

Amish 1 9 F Presymptomatic 12480 (5) + Ambulant T151R
Amish 2a 54 F 24 1160 (36) + Ambulant T151R–R91C
Amish 2b 57 F 43 ? ? Ambulant T151R–R91C
Amish 3a 39 F 11 ? + Walker R91C
Amish 3b 29 F 4.5 6140 (26) + Ambulant R91C
T044 15 M ? ? ? Wheelchair I119F–?

Creatine kinase levels (CK) measured at (age) of the patients are indicated. Normal CK values are estimated to be lower than 100. Severity of the condition of
the patients at present is indicated by their activity. No cardiomyopathy nor intellectual impairment have been reported in these cases. The second allele
mutation in the sporadic case T044 was not determined despite the analysis of the whole coding sequence of the b-sarcoglycan gene. Amish 1 refers to one
patient from one of the 11 families originally characterized with the common mutation T151R for which we had a biopsy available. Amish 2 refers to the two
LGMD2E patients from an Amish family heterozygote for the common T151R and the R91C mutations. Amish 3 refers to two affected sibs from a LGMD2E
family of probable Amish descent which was identified independently, and shown to carry an homozygous R91C mutation.

36 F. Duclos et al. / Neuromuscular Disorders 8 (1998) 30–38



affinity interaction of a-dystroglycan with the sarcolemma.
Further investigation of muscle biopsies from sarcoglycan
deficient patients are currently being performed.

The genetic and clinical heterogeneity of the limb-girdle
muscular dystrophies has broad implications for the diag-
nosis and management of persons afflicted with this disease.
Because defects in several distinct proteins can cause simi-
lar clinical symptoms and biochemical manifestations, a
definitive diagnosis can so far only be achieved through
genetic testing of all candidate genes. Analysis of additional
cases will provide more insights in genotype/phenotype cor-
relations and key domains of the proteins.

Acknowledgements

We are grateful to Delphine Samson, Corinne Cruaud,
Colleen Campbell, Chris Snyder, Cindy Leveille and
Herve Crespeau for their help. We are grateful to all the
clinicians who provided the patient samples for our study.

V.S. is supported by a grant from the Deutsche Forschungs-
gemeinschaft. C.E.J., G.L.F., J.-C.K., F.P. and K.P.C. are
supported by three grants from AFM (Association Francaise
contre les Myopathies). L.E.L. is supported by a grant from
the Iowa Affiliate of the American Heart Association. This
work was also supported by the Muscular Dystrophy Asso-
ciation. K.P.C. is an Investigator of the Howard Hughes
Medical Institute.

References

[1] Campbell KP, Kahl SD. Association of dystrophin and an integral
membrane glycoprotein. Nature 1989;338:259–262.

[2] Ervasti JM, Ohlendieck K, Kahl SD. et al. Deficiency of a glycopro-
tein component of the dystrophin complex in dystrophic muscle.
Nature 1990;345:315–319.

[3] Yoshida M, Ozawa E. Glycoprotein complex anchoring dystrophin to
sarcolemma. J Biochem 1990;108:748–752.

[4] Ervasti JM, Kahl SD, Campbell KP. Purification of dystrophin from
skeletal muscle. J Biol Chem 1991;266:9161–9165.

Fig. 3. Immunofluorescence analysis of DGC components in human control and Amish LGMD2E muscle. The LGMD2E patient (Patient) carried a
homozygous T151R mutation. Seven micrometer cryosections were stained with dystrophin (Dys), a-dystroglycan (a-DG), b-dystroglycan (b-DG), laminin
a2 (LAM-a2), a-, b-, g-, and d-sarcoglycan antibodies (a-SG, b-SG, g-SG and d-SG) as described in Section 2. The entire sarcoglycan complex is greatly
reduced at the sarcolemma of skeletal muscle from the b-sarcoglycan mutated patient compared to control. Additionally, the sarcolemmal staining for a-
dystroglycan is reduced. Bar, 50 mM.

37F. Duclos et al. / Neuromuscular Disorders 8 (1998) 30–38



[5] Ervasti JM, Campbell KP. Membrane organization of the dystrophin
glycoprotein complex. Cell 1991;66:1121–1131.

[6] Yoshida M, Suzuki A, Yamamoto H. et al. Dissociation of the com-
plex of dystrophin and its associated proteins into several unique
groups by n-octyl-D-glucoside. Eur J Biochem 1994;222:1055–
1061.

[7] Mizuno Y, Noguchi S, Yamamoto H. et al. Selective defect of sar-
coglycan complex in severe childhood autosomal recessive muscular
dystrophy muscle. Biochem Biophys Res Commun 1994;203:979–
983.

[8] Ibraghimov-Beskrovnaya O, Ervasti JM, Leveille CJ. et al. Primary
structure of dystrophin-associated glycoproteins linking dystrophin
to the extracellular matrix. Nature 1992;355:696–702.

[9] Adams ME, Butler MH, Dwyer TM. et al. Two forms of mouse
syntrophin, a 58 kD dystrophin-associated protein, differ in primary
structure and tissue distribution. Neuron 1993;11:531–540.

[10] Yang B, Ibraghimov-Beskrovnaya O, Moomaw CR. et al. Hetero-
geneity of the 59 kDa dystrophin-associated protein revealed by
cDNA cloning and expression. J Biol Chem 1994;269:6040–6044.

[11] Hoffman EP, Brown RH Jr, Kunkel LM. Dystrophin: the protein
product of the Duchenne muscular dystrophy locus. Cell
1987;51:919–928.

[12] Jung D, Leturcq F, Sunada Y. et al. Absence of g-sarcoglycan
(35DAG) in autosomal recessive muscular dystrophy linked to chro-
mosome 13q12. FEBS Lett 1996;381:15–20.

[13] Ervasti JM, Campbell KP. A role for the dystrophin-glycoprotein
complex as a transmembrane linker between laminin and actin. J
Cell Biol 1993;122:809–823.

[14] Campbell KP. Three muscular dystrophies: loss of cytoskeleton-
extracellular matrix linkage. Cell 1995;80:675–679.

[15] Hoffman E, Kunkel LM. Dystrophin abnormalities in Duchenne/
Becker muscular dystrophy. Neuron 1989;2:1019–1029.

[16] Helbling-Leclerc A, Zhang X, Topaloglu H. et al. Mutations in the
laminin a2chain gene (LAMA2) cause merosin-deficient congenital
muscular dystrophy. Nature Genet 1995;11:216–218.

[17] Nissinen M, Helbling-Leclerc A, Zhang X. et al. Substitution of a
conserved cysteine-996 in a cysteine-rich motif of the laminin a2-
chain in congenital muscular dystrophy with partial deficiency of the
protein. Am J Hum Genet 1996;58:1177–1184.

[18] Allamand V, Sunada Y, Salih MAM. et al. Mild congenital muscular
dystrophy in two patients with an internally deleted laminin a2-
chain. Hum Mol Genet 1997;6:747–752.

[19] Stevenson AC. Muscular dystrophy in Northern lreland. Ann Eugen
1953;18:50–91.

[20] Walton JN, Nattrass FJ. On the classification, natural history and
treatment of the myopathies. Brain 1954;77:169–231.

[21] Jackson CE, Strehler DA. Limb-girdle muscular dystrophy: clinical
manifestations and detection of preclinical disease. Pediatrics
1968;41:495–502.

[22] Noguchi S, McNally EM, Ben Othmane KB. et al. Mutations in the
dystrophin associated protein g-sarcoglycan in chromosome 13 mus-
cular dystrophy. Science 1995;270:819–822.

[23] Roberds SL, Leturcq F, Allamand V. et al. Missense mutations in the
adhalin gene linked to autosomal recessive muscular dystrophy. Cell
1994;78:625–633.

[24] Lim LE, Duclos F, Broux O. et al. b-Sarcoglycan: characterization
and role in limb-girdle muscular dystrophy linked to 4q12. Nature
Genet 1995;11:257–265.

[25] Bönnemann CG, Modi R, Noguchi S. et al. Mutations in the dystro-
phin-associated glycoprotein b-sarcoglycan (A3b) cause autosomal
recessive muscular dystrophy with disintegration of the sarcoglycan
complex. Nature Genet 1995;11:266–273.

[26] Nigro V, de Sa Moreira E, Piluso G. et al. Autosomal recessive limb-
girdle muscular dystrophy, LGMD2F, is caused by a mutation in the
d-sarcoglycan gene. Nature Genet 1996;14:195–198.

[27] Jung D, Duclos F, Apostol B. et al. Characterization of d-sarcogly-
can, a novel component of the oligomeric sarcoglycan complex
involved in limb-girdle muscular dystrophy. J Biol Chem
1996;271:32321–32329.

[28] Matsumura K, Tomé FMS, Collin H. et al. Deficiency of the 50K
dystrophin-associated glycoprotein in severe childhood autosomal
recessive muscular dystrophy. Nature 1992;359:320–322.

[29] Piccolo F, Roberds SL, Jeanpierre M. et al. Primary adhalinopathy: a
common cause of autosomal recessive muscular dystrophy of vary-
ing severity. Nature Genet 1995;10:243–245.

[30] Bönnemann CG, Passos-Bueno MR, McNally EM. et al. Genomic
screening for b-sarcoglycan gene mutations: missense mutations may
cause severe limb-girdle muscular dystrophy type 2E (LGMD2E).
Hum Mol Genet 1996;5:1953–1961.

[31] Duggan DJ, Gorospe JR, Fanin M. et al. Mutations in the sarcoglycan
genes in patients with myopathy. N Engl J Med 1997;336:618–624.

[32] Carrie A, Piccolo F, Leturcq F. et al. Mutational diversity and hot
spots in the a-sarcoglycan gene in autosomal recessive muscular
dystrophy (LGMD2D). J Med Genet 1997;34:470–475.

[33] Jorgensen AO, Arnold W, Shen AC. et al. Identification of novel
proteins unique to either transverse tubules (TS28) or the sarco-
lemma (SL50) in rabbit skeletal muscle. J Cell Biol
1990;110:1173–1185.

[34] Shapiro MB, Senapathy P. RNA splice junctions of different classes
of eukaryotes: sequence statistics and functional implications in gene
expression. Nucleic Acids Res 1987;15:7155–7174.

[35] Kislauskis EH, Zhu X, Singer RH. Sequences responsible for intra-
cellular localization of beta-actin messenger RNA also affect cell
phenotype. J Cell Biol 1994;127:441–451.

[36] Richard I, Broux O, Allamand V. et al. Mutations in the proteolytic
enzyme calpain 3, cause limb girdle muscular dystrophy type 2A.
Cell 1995;87:27–40.

[37] Allamand V, Broux O, Bourg N, et al. Genetic heterogeneity of
autosomal recessive limb-girdle muscular dystrophy in a genetic
isolate (Amish) and evidence for a new locus. Hum Mol Genet
1995;4(3):459–464.

[38] Beckmann JS. The Réunion paradox and the digenic model. Am J
Hum Genet 1996;59:1400–1402.

[39] Zlotogora J, Gieselmann V, Bach G. Multiple mutations in a specific
gene in a small geographic area: a common phenomenon? Am J Hum
Genet 1996;58:241–243.

[40] Welsh MJ, Smith AE. Molecular mechanisms of the CFTR chloride
channel dysfunction in cystic fibrosis. Cell 1993;73:1251–1254.

[41] Nigro V, Okasaki Y, Belsito A. et al. Identification of the Syrian
hamster cardiomyopathy gene. Hum Mol Genet 1997;6:601–607.

[42] Roberds SL, Ervasti JM, Anderson RD. et al. Disruption of the
dystrophin-glycoprotein complex in the cardiomyopathic hamster. J
Biol Chem 1993;268:11496–11499.

38 F. Duclos et al. / Neuromuscular Disorders 8 (1998) 30–38