brief communications - Nature

© 2001 Nature Publishing Group http://genetics.nature.com

brief communications

Mutations in SEPN1 cause congenital muscular dystrophy with spinal rigidity and restrictive respiratory syndrome


Published online: 20 August 2001, DOI: 10.1038/ng713 One form of congenital muscular dystrophy, rigid spine syndrome (MIM 602771), is a rare neuromuscular disorder characterized by early rigidity of the spine and respiratory insufficiency. A locus on 1p35–36 (RSMD1) was recently found to segregate with rigid spine muscular dystrophy 1 (ref. 1). Here we refine the locus and find evidence of linkage disequilibrium associated with SEPN1, which encodes the recently described selenoprotein N (ref. 2). Our identification and analysis of mutations in SEPN1 is the first description of a selenoprotein implicated in a human disease.

The RSMD1 locus is a 3 cM interval3 that includes about 100 putative genes, most of which are unknown. To reduce this interval we developed eight new CA-repeat microsatellite markers with P1-derived artificial chromosome (PAC) sequences mapped to 1p35–36 (Fig. 1a). As the analysis of these new markers in RSMD1 families did not show any sign of a recombination event, we searched for evidence of linkage disequilibrium in families originating from the same geographic area. Three Turkish and two Iranian families shared the same alleles for markers D1S3766, D1S3767, D1S3768 and D1S3769, which suggests a linkage disequilibrium among these families (Web Fig. A). We thus focused our attention on this area, which contains about 20 genes including SEPN1. SEPN1 encodes selenoprotein N, which is expressed in skeletal muscle2. Previous studies suggest a role for selenium in the physiopathology of the striated muscles, showing an association between selenium deficiency and muscular dystrophy in livestock4,5. In humans, selenium deficiency is associated with a cardiomyopathy known as Keshan disease6. Selenium is incorporated into selenoproteins as selenocysteine

at the UGA codon, normally a termination codon. The recognition of UGA as a selenocysteine codon requires a secondary structure called SECIS (selenocysteine

insertion sequence) that is located in the 3′ UTR of the transcripts of these genes7–9. SEPN1 contains an active SECIS element in its 3′ UTR that allows selenocysteine incorporation2; it was therefore a good candidate gene for RSMD1. We characterized the full-length cDNA of SEPN1 by 5′ RACE and RTPCR, and established its genomic structure using this cDNA sequence and the genomic sequence of the PAC dj317E23. SEPN1 contains 13 exons spanning 18.5 kb (Fig. 1b). It produces a 4.5 kb transcript, with an open reading frame of 1,770 nucleotides encoding a 590 amino-acid protein. The coding sequence and SECIS element of SEPN1 were analyzed by PCR-SSCP and sequencing. We identified two frameshift, one nonsense and three missense mutations

a

b

c calcium-binding site

Fig. 1 a, PAC sequences were assembled into contigs by the Sanger centre. We ordered these contigs using radiation hybrid localization of STSs on different PACs. The order of contigs 26 and 29 may be reversed. The position of microsatellite markers on different contigs is shown. *, newly developed markers. b, Genomic structure of SEPN1. Exons are indicated as black boxes, intronic sequences as thin bars. Arrows show exons in which mutations were found. The two inframe TGA codons are located at positions 379 and 1384, respectively, on the cDNA sequence, and the SECIS element is between nucleotides 2867 and 2927. c, Predicted amino acid sequence of human SEPN1 and homology comparison with its orthologs. Alignment was performed by the tblastn program against nonhuman ESTs. Identical amino acids are in bold. A calcium-binding site was predicted by the pfscan program (www.expasy.ch). Mutations are shown by arrows. -, alignment gaps; =, sequence portions unavailable in EST databases; U and Sec, selenocysteine; Fs, frameshift.

nature genetics • volume 29 • september 2001

17

brief communications © 2001 Nature Publishing Group http://genetics.nature.com Table 1 • Mutations in SEPN1 Number of affected children

Origin

1809

3

Morocco

1

22dup10bp

frameshift at G7

13369

1

Algeria

1

22dup10bp

frameshift at G7


Family

Exon

Nucleotide change

Amino acid change

12566

2

Iran

6

817G→A

G273E

E1

1

Iran

6

817G→A

G273E

from the Institut National de la Santé et de la Recherche Médicale (INSERM), Association Française contre les Myopathies (AFM), the European Commission and a Muscular Dystrophy Campaign Grant. B.M. was supported by grants from La Fondation pour la Recherche Médicale (FRM) and La Fondation Bettencourt Schueller.

E13

1

Turkey

6

817G→A

G273E

Received 28 February; accepted 26 June 2001.

T21

1

Turkey

6

817G→A

G273E

T2

1

Turkey

6

817G→A

G273E

E8

2

Italy

Behzad Moghadaszadeh1, Nathalie Petit1, Céline Jaillard1, Martin Brockington2, Susana Quijano Roy3, Luciano Merlini4, Norma Romero1, Brigitte Estournet3, Isabelle Desguerre5, Denys Chaigne6, Francesco Muntoni2, Haluk Topaloglu7, Pascale Guicheney1

7

878A→G*

H293R

11

1397G→A*

R466Q

1385G→A

14961

1

France

10

T17

1

Turkey

11

1446delC

U462stop frameshift at L482

Note: * heterozygous mutation

in ten families (Table 1, Fig. 1b,c and Web Fig. B). The nonsense mutation identified in the French family transforms the selenocysteine codon (TGA) into a stop codon (TAA), preventing selenocysteine incorporation and leading to a shorter protein. The missense mutations detected in RSMD1 patients change residues that are conserved in vertebrates (Fig. 1c). Two of these mutated residues are clustered in one region, the other one being in close proximity to the selenocysteine residue. Two isoforms of SEPN1 are predicted from EST databases and RT-PCR experiments. Isoform 1 corresponds to the fulllength transcript, whereas in isoform 2, exon 3 is spliced out. Both transcripts were detected in skeletal muscle, brain, lung and placenta, but isoform 2 is always predominant. The exon 3 sequence corresponds to an Alu cassette and contains a second in-frame selenocysteine codon. So far, ten selenoprotein families have been characterized in mammals10,11. All contain selenium as selenocysteine, and four have known enzymatic functions. The selenocysteine residue is always located in the active site. Its selenol group is ionized at physiological pH and confers more reactivity to the proteins than a thiol group. SEPN1 contains two selenocysteines and seven cysteines, two of them close to the selenocysteine residues; thus, this protein may also function as an enzyme. Both selenocysteine-containing

18

motifs (USSC and CUG) might correspond to catalytic sites of SEPN1; for example, selenoproteins such as thioredoxin reductases present the CUG motif in their C-terminal catalytic portion12. In contrast to congenital muscular dystrophy patients with merosin deficiency (MIM 156225)13, those with mutations in SEPN1 are ambulant and present a milder muscular dystrophy with no basal membrane alteration and almost normal levels of serum creatine kinase. They all present reduced vital capacity and progressive nocturnal hypoventilation requiring ventilatory support14,15, which might be a direct consequence of SEPN1 dysfunction. SEPN1 may thus play a key part in the physiology of skeletal muscles such as the diaphragm, by maintaining the redox environment of the cell and preventing it from oxidant damage. Note: Supplementary information is available on the Nature Genetics web site (http:// genetics.nature.com/supplementary_info/).

1INSERM

U523, Institut de Myologie, Groupe Hospitalier Pitié-Salpêtrière, 47 boulevard de l’Hôpital, 75651 Paris CEDEX 13, France. 2Department of Paediatrics and Neonatal Medicine, Imperial College School of Medicine, Hammersmith Hospital, London, UK. 3Hôpital Raymond Poincaré, Garches, France. 4Rizzoli Orthopedic Institute, Bologna, Italy. 5Service de Neuropédiatrie, Hôpital Saint-Vincent-de-Paul, Paris, France. 6Clinique Sainte Odile, Strasbourg, France. 7Department of Paediatric Neurology, Hacettepe Children’s Hospital, Ankara, Turkey. Correspondence should be addressed to P.G. (e-mail: p.guicheney@ myologie.chups.jussieu.fr). 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11.

Acknowledgments

We thank the patients and their families for their participation; K. Schwartz, M. Fiszman, F. Tomé and M. Fardeau, for continuous support and constructive discussions; A. Lescure, V. Allamand and U. Wewer for comments and suggestions on the manuscript; G. Boccara, M. Bonay, E. Leguern, B. Riou and the AFM BTR for providing human tissues. This work was supported by funds

12. 13.

14. 15.

Moghadaszadeh, B. et al. Am. J. Hum. Genet. 62, 1439–1445 (1998). Lescure, A., Gautheret, D., Carbon, P. & Krol, A. J. Biol. Chem. 274, 38147–38154 (1999). Flanigan, K.M. et al. Ann. Neurol. 47, 152–161 (2000). Lamand, M.C.R. Acad. Sci. Hebd. Sciences Acad. Sci. D. 270, 417–420 (1970). Oldfield, J.E. Biol. Trace. Elem. Res. 20, 23–29 (1989). Ge, K., Xue, A., Bai, J. & Wang, S. Virchows Arch. A. Pathol. Anat. Histopathol. 401, 1–15 (1983). Berry, M.J., Banu, L., Harney, J.W. & Larsen, P.R. EMBO J. 12, 3315–3322 (1993). Walczak, R., Westhof, E., Carbon, P. & Krol, A. RNA 2, 367–379 (1996). Gu, Q.P. et al. Gene 193, 187–196 (1997). Holben, D.H. & Smith, A.M. J. Am. Diet. Assoc. 99, 836–843 (1999). Flohe, L., Andreesen, J.R., Brigelius-Flohe, R., Maiorino, M. & Ursini, F. IUBMB Life 49, 411–420 (2000). Zhong, L., Arner, E.S. & Holmgren, A. Proc. Natl. Acad. Sci. USA 97, 5854–5859 (2000). Tomé, F.M.S., Guicheney, P. & Fardeau, M. Neuromuscular disorders: Clinical and molecular genetics (ed. Emery, A.E.H.) 21–57 (John Wiley & Sons, Chichester, Sussex, England, 1998). Merlini, L., Granata, C., Ballestrazzi, A. & Marini, M.L. J. Child. Neurol. 4, 274–282 (1989). Moghadaszadeh, B. et al. Neuromuscul. Disord. 9, 376–382 (1999).

nature genetics • volume 29 • september 2001