Myotonia Congenita: Mutation Spectrum of CLCN1 in Spanish Patients

Journal of Genetics (2019) 98:71 © Indian Academy of Sciences https://doi.org/10.1007/s12041-019-1115-0

RESEARCH ARTICLE

Myotonia congenita: mutation spectrum of CLCN1 in Spanish patients

CARMEN PALMA MILLA1∗ , CARMEN PRIOR DE CASTRO1, CLARA GÓMEZ-GONZÁLEZ1, PALOMA MARTÍNEZ-MONTERO1, SAMUEL I. PASCUAL PASCUAL2 and JESÚS MOLANO MATEOS1

1Instituto de Genética Médica y Molecular, CIBERER, IdiPAZ, Hospital Universitario La Paz, Madrid 28046, Spain 2Servicio de Neurología Pediátrica, Hospital Universitario La Paz, Madrid 28046, Spain *For correspondence. E-mail: [email protected].

Received 15 January 2019; revised 2 May 2019; accepted 6 May 2019

Abstract. Myotonia congenita (MC) is a Mendelian inherited genetic disease caused by the mutations in the CLCN1 gene, encoding the main skeletal muscle ion chloride channel (ClC-1). The clinical diagnosis of MC should be suspected in patients presenting myotonia, warm-up phenomenon, a characteristic electromyographic pattern, and/or family history. Here, we describe the largest cohort of MC Spanish patients including their relatives (up to 102 individuals). Genetic testing was performed by CLCN1 sequencing and multiplex ligation-dependent probe ampliﬁcation (MLPA). Analysis of selected exons of the SCN4A gene, causing paramyotonia congenita, was also performed. Mutation spectrum and analysis of a likely founder effect of c.180+3A>T was achieved by haplotype analysis and association tests. Twenty-eight different pathogenic variants were found in the CLCN1 gene, of which 21 were known mutations and seven not described. Gross deletions/duplications were not detected. Four probands had a pathogenic variant in SCN4A. Two main haplotypes were detected in c.180+3A>T carriers and no statistically signiﬁcant differences were detected between case and control groups regarding the type of haplotype and its frequencies. A diagnostic yield of 51% was achieved; of which 88% had pathogenic variants in CLCN1 and 12% in SCN4A. The existence of a c.180+3A>T founder effect remains unsolved.

Keywords. myotonia congenita; founder effect; paramyotonia congenita; CLCN1 gene.

Introduction slightly elevated (≤ 3–4x the upper limits of normal) and electromyography (EMG) discloses characteristic showers Myotonia congenita (MC) is the most common form of of spontaneous electrical activity (myotonic bursts) (Dunø nondystrophic myotonia and is caused by mutations in and Colding-Jørgensen 2005). the CLCN1 gene (Koch et al. 1992; Suetterlin et al. 2014), Other forms of muscle skeletal channelopathies should which encodes the skeletal muscle voltage-gated chloride be considered in differential diagnosis of MC, espe- channel (ClC-1) and affects muscle excitability (Lipicky cially paramyotonia congenita (PMC, OMIM: 168300) et al. 1971). Inheritance mode may be dominant (Thom- and sodium channel myotonia (SCM, OMIM: 170500 & sen disease, OMIM: 160800) or recessive (Becker disease 613345), caused by pathogenic variants in the SCN4A or recessive generalized myotonia, OMIM: 255700), with gene, which encodes the alpha subunit of the skeletal a more severe phenotype in the latter form (Pusch 2002; muscle voltage-gated sodium channel. Although, some- Lossin and George 2008; Dunø and Colding-Jørgensen times PMC may be difﬁcult to distinguish from MC, since 2005). both conditions are presented with episodes of general- MC should be suspected in individuals with the ized stiffness, in individuals with PMC, repeated muscle following clinical ﬁndings: episodes of muscle stiffness contractions may aggravate stiffness (also termed paradox- (myotonia), alleviation of stiffness by brief exercise (warm- ical myotonia) conversely, to individuals with MC. Both up effect), myotonic contraction elicited by percussion phenotypes (PMC and SCM) have a dominant mode of of muscles and a family history consistent with either inheritance, and result in a gain of channel function (Pusch autosomal dominant or autosomal recessive inheritance. 2002; Jurkat-Rott et al. 2010; Suetterlin et al. 2014). It is Additionally, serum creatine kinase concentration may be estimated that up to 20% of the individuals suspected of

0123456789().: V,-vol 71 Page 2 of 10 Carmen Palma Milla et al. having MC may in fact have pathogenic variants in SCN4A was specifically conducted for myotonic signs such as per- (Trip et al. 2008). In the differential diagnosis of myotonic cussion myotonia and other clinical symptoms (warm-up syndromes, it is imperative to distinguish those associated effect, EMG myotonic bursts etc.). Extramuscular man- with dystrophy from the nondystrophic forms. Myotonic ifestations of DM1 and DM2 were also considered for dystrophy type 1 (DM1, OMIM: 160900) and myotonic a differential diagnosis. Informed consent was obtained dystrophy type 2 (DM2, OMIM: 602668) should always from all patients for genetic testing according to Interna- be considered in the differential diagnosis of MC. DM1 tional Ethical Guidelines and Declaration of Helsinki. and DM2 are multisystem disorders in which late-onset progressive wasting of bulbar and distal muscles usually Genetic diagnosis predominates over a relatively mild degree of myotonia. Inheritance is dominant, highly penetrant and present Genomic DNA was extracted from peripheral blood, genetic anticipation (Dunø and Colding-Jørgensen 1993). collected in EDTA-containing tubes by standard proce- MC was originally estimated to occur with a frequency dures. In a first approach, the whole coding region of of 1 : 23,000 for autosomal dominant MC and 1 : 50,000 CLCN1 (NM_000083.2) was sequenced followed by gene- for the autosomal recessive form (Becker 1979). ClC-1 is − targeted deletion/duplication analysis if no pathogenic nearly exclusively expressed in skeletal muscle, where Cl variant or just one recessive variant was found. For this channels have a critical role in electrically stabilized and purpose of deletion/duplication analysis, SALSA-P350- repolarization of skeletal muscle membranes (Jentsch et al. B1 CLCN1-KCNJ2 was used. When no mutations were 2005). Myotonia-inducing mutations in chloride channels detected in CLCN1, analysis of selected-exons in SCN4A are loss-of-function mutations that reduce the stabilizing − was performed. In this case, exons 9, 13, 22 and 24 were Cl conductance (Pusch 2002). analysed, as recognized hot-spot regions in the SCN4A Until now more than 200 pathogenic variants have been gene (Lehmann-Horn et al. 2008; Matthews et al. 2008; identified over the entire length of this channel, most Simkin and Bendahhou 2011). of them being point mutations, especially missense and Proband’s parents and other family members who are at nonsense mutations. Molecular testing approach includes risk were also analysed whenever possible as part of genetic sequence analysis of CLCN1 followed by gene-targeted counselling. deletion/duplication analysis if no pathogenic variant is Variant interpretation was made according to Ameri- found. can College of Medical Genetics and Genomics (ACMG) In this study, we provide a diagnostic approach to MC practice guidelines (Richards et al. 2015). Different aspects in the largest cohort of Spanish patients, including CLCN1 regarding pathogenicity were considered: predicted impact, sequencing and multiplex ligation-dependent probe ampli- result of computational (in silico) predictive programmes, fication (MLPA) in addition to selected exons (recognized computational and predictive data (PolyPhen2 (Adzhubei as mutation ‘hot spots’) sequencing of SCN4A,aspart et al. 2010), SIFT (Li et al. 2009), LRT (Chun and Fay of differential diagnosis. We also provide an in-depth 2009), mutation taster (Schwarz et al. 2014), mutation characterization for some ClC-1 mutations described assessor (Reva et al. 2011), human splicing finder (Desmet for the first-time (c.181-1G>T; c.434-5_434-4insGCA; et al. 2009), conservation (Radial SVM and LR) and c.563G>T: p.Gly188Val; c.890T>A: p.Phe297Tyr; c.1649 segregation analysis. Data were collected from the fol- C>G: p.Thr550Arg; c.1712A>G: p.Gln571Arg; c.2096 lowing electronic databases: dbSNP (Sherry et al. 2001), delG: p.Gly699Alafs*95). Finally, we performed a founder ExAC (‘ExAC Browser’ 2016, http://exac.broadinstitute. effect analysis for the recurrent mutation in Spanish pop- org/about), 1000G (Auton et al. 2015), ExomeVariant ulation c.180+3A>T. Server (Tennessen et al. 2012), ClinVar (Landrum et al. 2015) and Human Gene Mutation Database (HGMD) professional version (Stenson et al. 2014). Patients and methods Newly described pathogenic variants were incorporated into the Leiden Open Variation Database 3.0 (‘CLCN1 Clinical diagnosis gene homepage - Shared database,’ n.d.). We investigated a large cohort of Spanish patients who had all been referred to our centre with a clinical diagnosis of Founder effect MC (102 individuals, of which 72 were probands and 30 were their relatives). We investigated the putative founder effect of the c.180 Patients with clinical diagnosis of MC were referred +3A>T mutation through the construction of SNP hap- to our lab who were from different neurology units scat- lotypes throughout the CLCN1 gene. We analysed intra- tered all over the country. Patients, as well as their genic SNPs: rs6962852, rs2272253, rs2272252, rs2272251, first-degree relatives underwent clinical examination and rs13438232 and rs56680997, all were diallelic and infor- electromyography evaluation. Neurological examination mative markers (MAF > 0.20) (Gabriel et al. 2002). For Mutation spectrum of CLCN1 in Spanish patients Page 3 of 10 71 haplotype identification, the Haploview program was used Founder effect (Barrett et al. 2005). A total of 18 individuals carriers of c.180+3A>Tmuta- Six intragenic single-nucleotide polymorphisms (SNPs) tion (four individuals carried the c.180+3A>T mutation were selected for the haplotype construction. They were in both alleles and 14 individuals were heterozygote) and chosen based on informativity from public databases with 18 individuals genotyped in 1000 Genome as control pop- available genotyping data. Controls and MC patients were ulation from Iberian population in Spain were included genotyped and Haploview software (Barrett et al. 2005) for this haplotype analysis. showed that five SNPs lie within the same 7-kb haplotype block: rs2272253, rs2272252, rs2272251, rs13438232 and rs56680997 while rs6962852 is localized in a different one Results (figure 2). Eight different haplotypes were detected in the group Distribution of CLCN1 mutations in the MC cohort of affected individuals and two of them were clearly pre- dominant: C-GCTCC (47%) and C-ATCTG (37%). The CLCN1 gene mutations were detected in 33 patients with program Haploview was adapted to find significant dif- clinical diagnosis of MC (46%) of the 72 probands (67% ferences between the genotypes in the group of affected male and 33% female subjects). We also analysed 30 rela- individuals and the group of controls. No significant tives as part of genetic counselling, both symptomatic and differences were detected between these two groups for asymptomatic subjects. Of the 33 probands, we found 28 individual markers or haplotype blocks results. different mutations scattered across the 23 exons of the CLCN1 gene, 21 of them had been already reported and seven were novel (table 1). Discussion Additionally, four of the 39 probands with normal results for CLCN1 had a pathogenic variant in SCN4A Our genetic screening of 72 probands is the most extended (6% of the cohort). Mutations detected in the SCN4A gene cohort of Spanish individuals with suspicion of MC pub- were: c.3917G>T (p.Gly1306Val), c.3917G>C (p.Gly1306 lished until now. We identified 33 CLCN1 mutated, unre- Ala), c.4342C>T (p.Arg1448Cys) and c.4300T>C (p.Ser lated probands. Among the different mutations detected, 1434Pro); the last one was previously described by our seven were novel and 21 were previously reported. The group (Palma et al. 2017). diagnostic yield reached 51% (37 of 72 probands) with No gross deletions/duplications were detected in none our algorithm of CLCN1/SCN4A analysis; 45% of them of the patients tested (0/35 patients). harboured CLCN1 mutations and 6% SCN4A mutations. Including selected SCN4A exons, analysis increased the diagnostic yield in those patients but a high percentage Novel mutations of individuals remained undiagnosed. This fact could be due to an erroneous diagnosis because of vagueness of the Novel mutations included four missense mutations, one symptoms or, on the other hand, limitations of detection splicing mutation, one deletion and one insertion of three methods. It should be noted that some of the undiag- nucleotides in splicing region. Clinical findings and com- nosed patients could still have pathogenic variants in the plete genotype of those new mutations are described in SCN4A gene and this issue should be satisfied in further table 2. analysis. SCM, MC and PMC are primary skeletal mus- One of those families carrying mutation T550R is cle channelopathies, and have similar clinical presentations depicted in pedigrees shown in figure 1. where both myotonia and weakness are present with differ- Effects of the novel mutations predicted by computa- ent intensity in a continuous spectrum. Associated genes tional and predictive tools are summarized in table 3, (CLCN1 and SCN4A) encode ion channels involved in where conservation of mutated residues is also shown. For skeletal muscle excitability and mutations on them can lead two mutations: c.563G>T (p.Gly188Val), and c.1649C>G to a similar phenotype. (p.Thr550Arg), all in silico tests agree on their pathogenic- No gross deletions were detected in the CLCN1 gene ity. DNA variants c.890T>A (p.Phe297Tyr) and by MLPA. We found a high heterogeneity in CLCN1 c.1712A>G (p.Gln571Arg) have variable results with mutations, scattered across most of the 23 exons of respect to their pathogenicity, when analysed with predic- the CLCN1 gene, excluding 1, 2, 9, 10, 16, 20 and 21 tive bioinformatic algorithms. However, for most of the exons. Most mutations are missense (64%) and splicing algorithms (Poly Phen2, LRT, Mutation Taster, Mut Pred, (19%). Phylo and Siphy) results indicate a pathogenic effect. For We identified two highly prevalent mutations in this deletion and splicing mutations, just one predictive tool Spanish cohort, in agreement with previously published was used (MutationTaster and HSF, respectively), in these data (Mazón et al. 2012): c.180+3A>T and c.1453A>G cases, a pathogenic impact was predicted. (p.Met485Val), both were associated with autosomal 71 Page 4 of 10 Carmen Palma Milla et al. 1 1 )2 )3 )9 )3 . (2012) 1 1995 1995 1995 1995 ( ( ( ( ) )2 )1 )1 )3 )1 )1 )3 et al )2 ) )3 et al. et al. et al. et al. 2012 2012 2012 2012 2012 2012 2012 2012 ( ( ( ( ( ( ( ( . (2013) 1 . (1994) 1 . (1994) 2 ., (1994) 1 2008 2008 2001 ( ( ( et al et al et al et al. et al. et al. et al. et al. et al. et al. et al. et al et al. et al. et al. Mazón Mazón Trip Mazón Mazón Meyer-Kleine Meyer-Kleine Meyer-Kleine Trip Meyer-Kleine Mazón Mazón Mazón Sun Mazón . (2002) and UniProt UniProt Consortium (2015), (‘UniProt: a hub for protein information’ et al mutations detected in Spanish MC patients. C K Helix i10 Becker A N Helix i13 Becker T C terminus i18 Becker T E Helix (Selectivity filter part 2) i5 Becker > > A N terminus i2 ? > T N terminus i1 Becker Present study 2 A; p.Gly416Glu K-L Linker 11 ? A; p.Gly482Arg M-N Linker 13 Becker T; p.Arg496Ser N Helix 14 Becker Lorenz A; p.Val851Met C terminus 22 ? G; p.Met485Val N Helix 13 Becker C; p.Gln788Pro C terminus 19 ? T N terminus i1 Becker Sloan Brown and George (1997) 7 G; p.Thr550ArgA; p.Asp567Lys P Helix Q Helix 15 15 Becker Thomsen Present study 1 G; p.Gln571Arg Q Helix 15 Thomsen Present study 1 T; p.Arg894* C terminus 23 Becker/Thomsen George > CLCN1 > > A; p.Gly270Asp G Helix 7 ? T; p.Gly233Val F Helix (Selectivity filter part 2) 6 ? A; p.Gly190Arg D Helix (Selectivity filter part 1) 5 Morrow T; p.Lys248* F Helix 6 Gurgel-Giannetti T; p.Gly188Val D Helix (Selectivity filter part 1) 5 Becker Present study 2 G; p.Phe167Leu C Helix 4 Becker George A; p.Phe297Tyr H-I Linker 8 Becker Present study 1 > > > > > > > > > > > > > > > > > > c.1167-10T c.890T c.1247G c.696+1G c.1444G c.1453A c.1471+1G c.809G c.1262insC; p.Glu422Argfs*8 L Helix 12 Becker c.698G c.742A c.563G c.568G c.434-5insGCAc.501C B Helix i3 ? Present study 2 c.1488G c.1701C c.1649C c.2363A c.302-1G c.181-1G Table 1. Mutation180+3A Domain Exon/intron PhenotypeDomains were inferred from bibliography and structure proposed by Dutzler Reference Freq. 2014). Newly described mutations are in bold letters. Freq, frequency of apparition in the cohort. c.2096delG; p.Gly699Alafs*95c.2178delT; p.Pro727Leufs*67 C terminus C terminus 17 18 Becker ? Present study 1 c.1712A c.2284+5C c.2551G c.2680C Mutation spectrum of CLCN1 in Spanish patients Page 5 of 10 71 . (2014). et al EMG: myotonic runs Warm-up phenomenon. EMG: myotonic runs. Percussion myotonia. Muscular hypertrophy myotonic runs Warm-up phenomenon. Muscular hypertrophy. No cold-sensitive myotonia Warm-up phenomenon. Abnormal levels of creatine kinase in blood. No muscular EMG: hypertrophy. myotonic runs Warm-up phenomenon. Percussion myotonia. Cold-sensitive myotonia. Muscular hypertrophy myotonic runs Age at (years) Family history Clinical features diagnosis Sporadic Early onset No Percussion myotonia. Recessive 9 Yes Myotonia. EMG: G G > > (p.Met485Val) (p.Phe167Leu) – – 7 – Percussion myotonia. – Sporadic Early onset No Muscular stiffness. – – – – Myotonia A c.501C T c.1453A > T Recessive 55 No Myotonia. EMG: > T G > T > > > (p.Gly188Val) (p.Arg496Ser) (p.Met485Val) c.1453A c.563G c.2284+5C – – Dominant Early onset Yes EMG: myotonic runs. c.1471+1G T c.181-1G G G > T A > > > > 1st Mutation 2nd Mutation 3rd Mutation Inheritance (p.Phe297Tyr) (p.Gly188Val) (p.Gly699Alafs95*) (p.Gln571Arg) (p.Thr550Arg) Genotype and clinical features in patients in whom newly-described mutations were detected. Phenotype data standardized with HPO terms Köhler E c.890T D c.563G BC c.434-5insGCA – c.434-5insGCA c.1488G – Sporadic 12 No Muscular stiffness. Table 2. Patient/ family A c.181-1G H c.2096delG G c.1712A F c.1649C 71 Page 6 of 10 Carmen Palma Milla et al.

Figure 1. Genetic tree of F and I families. Circles indicate females; squares indicate males. Arrows: propositus. Solid squares and circles indicate affected individuals. Dotted symbols indicates carrier clinically unaffected.

recessive inheritance. Thus, despite a large The novel variant c.181-1G>T was detected in heterogeneity of mutations observed in patients with MC, homozygous state in a 55 year-old patient and a Becker a limited number of mutations accounted for a large disease phenotype. Both parents were consanguineous proportion of cases and could be used for designing a and from the same small town in Cáceres (Spain). They diagnostic strategy for Spanish patients with myotonia. were asymptomatic carriers of this alteration. The muta- The c.180+3A>T mutation has also been shown fre- tion occurs in the late intronic position of intron 1, and quent in other South-European cohorts (Brugnoni et al. affects the wild-type cannonical acceptor splice site, most 2013); however, c.1453A>G has never been reported as probably affecting splicing process. This variant has not a frequent mutation in other European cohorts of MC been reported before in consulted databases. According patients. to the ACMG standards for sequence variants interpre- Among all MC patients with genetic diagnosis, we tation, this variant could be classified as a pathogenic found an autosomal recessive phenotype (Becker disease) variant (Richards et al. 2015). Nevertheless, splicing in in 80% of patients, most of them been composed heterozy- vitro assays should be performed to confirm the splicing gotes. In three of these cases, patients harboured three effect. different mutations: c.501C>G(;)313C>T(;)180+3A>T, The insertion of three nucleotides in splicing region c.[1471+1G>A];[501C>G;1649C>G] and c.2284+5C c.434-5_434-4insGCA has been detected in two different >T(;)1453A>G(;)2096delG. probands, both with early onset of the disease: one of Thomsen disease (autosomal dominant) was detected them in composed heterozygosity with c.1488G>Tand in 20% of the patients. Although MC was originally the other one just in heterozygosis. This variant has been estimated with a frequency of 1:23,000 for autosomal reported in ExAc database (frequency 0.0002%) and in dominant MC and 1:50,000 for the autosomal recessive dbSNP database as rs753470655. This splice acceptor vari- form (Becker 1979), our results of a higher prevalence ant is also named as c.434-2_434dupAGC and it could of autosomal recessive phenotype is in agreement with lead to the insertion of one amino acid residue in the suggestions of subsequent studies (Meyer-Kleine et al. CLCN1 protein (p.Ala145dup), preserving the integrity 1995; Lehmann-Horn and Jurkat-Rott 1999; Papponen of the reading frame. The human splicing finder predic- et al. 1999; Sun et al. 2001; Fialho et al. 2007; Dunø and tor assess that this variant would likely affect splicing Colding-Jørgensen 2005). process by altering an exonic ESE site. According to the Regarding novel mutations, we describe the following ACMG standards, this variant could be classified as a : c.181-1G>T, c.434-5_434-4insGCA, c.563G>T (p.Gly variant of uncertain significance, nevertheless additional 188Val), c.890T>A (p.Phe297Tyr), c.1649C>G (p.Thr550 splicing assays would be necessary to fully assess this vari- Arg), c.1712A>G (p.Gln571Arg) and c.2096delG (p.Gly ant. 699Alafs*95). These novel mutations were not described The c.563G>T (p.Gly188Val) variant was present in a in control population databases, or in some cases these Spanish patient with Maroccan ancestry, in homozygous are already included with an extremely low frequency state. This substitution affects the highly conserved 188 (ExAC (Consortium et al. 2015) and 1000 Genome (Auton residue, which is part of the five-nucleotides sequence that et al. 2015)). All missense novel changes affect highly con- makes the first part of selectivity filter of ClC-1 channel. In served residues with a predicted functional role in the addition, nucleotide 563 is the first one of exon 5, therefore protein. Computational tools were also employed to assess it could also affect splicing process. All the in silico tools pathogenicity. agreed in a pathogenic effect except for the predictor HSF. Mutation spectrum of CLCN1 in Spanish patients Page 7 of 10 71 value, P eutral). SIFT: D, deleterious 5.02 9.869 18.413 5.57 9.170 – 5.02 9.869 18.413 splicing acceptor site, most probably affecting splicing probably no impact on splicing. values reliability of the prediction, so that higher the

P Figure 2. Linkage disequilibrium pattern scheme generated using Haploview. The default algorithm is taken from Gabriel et al. (2002). 95% conﬁdence bounds on D are generated and each comparison is called ‘strong LD’, ‘inconclusive’ or ‘strong recombination’. A block is created if 95% of informative (i.e. noninconclusive) comparisons are ‘strong LD’. Rate (D of LD is represented by different colours (highest rate of LD in dark grey, lower LD in light grey, and white for no LD).

Mutation Assessor Mut Pred HSF GERP++ PhyloP Siphy A different missense change in homozygosis at this amino acid residue had been reported before. The substitution of glycine by alanine at this amino acid residue also resulted in a pathogenic effect (Brugnoni et al. 2013). 0.909). For mutation Taster y MutPred > ≥ The c.890T A (p.Phe297Tyr) variant was detected in

Mutation Taster composed heterozygosis with c.1453A>G (p.Met485Val) in a MC patient. Mostly in silico tools agreed in a pathogenic effect of this variant affecting a conserved residue and been part of the H–I α-helix union which form the channel interface in combination with P–Q α-helix. In addition, this intracellular domain connecting H and I helix interacts with CBS2 domain which could be related to the common gate of the channel (Estévez et al. 2004). Another missense change in this position was described before: c.890T>C; p.Phe297Ser, this case was associated with Thomsen phenotype. Differences in electrophysio-

NA NA NAlogical P (1.00)properties NA has been seen NA between NA p.Phe297Tyr and - - - p.Phe297Ser (Fialho et al. 2007). This different behaviour 0.05). PolyPhen2: D, probably damaging (

> could be explained by the different physico-chemical properties of Tyr and Ser amino acids, where Tyr is a more conservative change, Phe and Tyr are aromatic amino acids with a likely structural role (based in Consurf prediction, data not shown (Berezin et al. 2004)). The c.1649C>G (p.Thr550Arg) mutation was detected T – – – – – – Alteration of the WT G p.Gln571Arg T (0.10) D (0.99) P P (1.00) M P (0.59) NA 5.02 9.869 18.413 G p.Thr550Arg D (0.00) D (1.00) P P (1.00) H P (0.92) NA 6.04 9.331 16.576 Computational tools employed to assess pathogenicity and conservation of newly described mutations. >

T p.Gly188Val – D (1.00) Pin family P (1.00) G (ﬁgure H1), which P (0.99) is This mutation a has paternally inherited muta- A p.Phe297Tyr D (0.03) D (0.99) P P (1.00) L P (0.88) NA 5.91 7.811 20.313 > > > > 0.05); T, tolerated (sift tion detected in two sibs, diagnosed at age of nine and 10 ≤ years old, respectively, in combination with c.501C>Gand p.Gly699Alafs*95 c.563G Table 3. Mutationc.181-1G SIFT Poly Phen2 LRT Results for different computational tools are indicated. P,pathogenic; H/M, functional and L/N means nonfunctional (H, high; M, medium; L, low; N, n (sift higher the reliability. NA, non applicable; WT, wild-type. c.434-5insGCA – – – N (0.99) – – Potential alteration of c.2096delG c.1649C c.1712A c.890T in trans disposition with c.1471+1G>A. This mutation 71 Page 8 of 10 Carmen Palma Milla et al.

Table 4. Haplotypes association studies.

Block Haplotype Frequency Case–control frequencies Chi-square P value

1 C 0.814 0.806; 0.824 0.037 0.8467 T 0.186 0.194; 0.176 0.037 0.8467 2 GCTCC 0.470 0.415; 0.529 0.914 0.3389 ATCTG 0.369 0.330; 0.411 0.501 0.479 ATCTC 0.043 0.083; 0.000 2.96 0.0853 GTCTG 0.043 0.083, 0.000 2.96 0.0853 GCCCC 0.030 0.002; 0.059 2.007 0.1566 ATTTG 0.016 0.031; 0.000 1.052 0.3051 GTTTC 0.014 0.028; 0.000 0.958 0.3277 GTCCG 0.014 0.028, 0.000 0.958 0.3277 occurs in a conserved position of P α-helix. All the in sil- sample size, in such a way we found for instance: 0% ico tools agree with a pathogenic effect. A variant affecting frequency in Iberian population in Spain or 1.5% in the same residue: c.1649C>T; p.Thr550Met (rs762754992) Mexican ancestry from Los Angeles, USA. Our group was described before with some controversial interpreta- previously reported a frequency in Spanish control pop- tion as it has been associated with Thomsen (Lossin and ulation, based on 340 individuals around 10 times more George 2008), mild Thomsen (Colding-Jørgensen 2005)or than the global frequency (∼0.3%) (Mazón et al. 2012). Becker phenotypes (Pusch 2002). To investigate the possibility of a founder effect regard- The c.1712A>G (p.Gln571Arg) mutation was detected ing this variant, we genotyped six intragenic SNPs of in heterozygous state in a 13 year-old patient with a positive CLCN1: rs6962852, rs2272253, rs2272252, rs2272251, family history (mother and maternal aunt) and character- rs13438232 and rs56680997. We observed two haplo- istic symptoms of autosomal dominant MC. This variant types blocks: one with SNP1 and the second com- affects to a conserved residue, with a likely functional role, posed by SNP2–SNP6 and two representative haplo- localized in the last transmembrane helix (Q). This variant types: C-GCTCC and C-ATCTG (figure 1), both present has never been observed in control populations and in sil- in MC patients and control population. In a second ico tools agree in a pathogenic prediction for this variant. approximation, a case–control analysis was performed. The c.2096delG (p.Gly699Alafs*95) deletion was detec- We did not find differences between both groups, nei- ted in a sporadic case of early onset MC. This is one of ther in markers individually nor in haplotypes blocks the patients with three CLCN1 mutations detected in this (table 4). cohort, with genotype: c.2284+5C>T(;)1453A>G(;)2096 Regarding this data, although the c.180+3A>T patho- delG. Unfortunately, we do not know the phase of these genic allele is prevalent in Spanish population, is not three variants. This mutation would produce a trun- associated to a common allele and the existence of a cated protein with lack of CBS2 and C-terminal domains. founder effect remained unconfirmed. Nevertheless, some This may be a similar product than the c.2178delT limitations are noted in this haplotype study, since all (p.Pro727leufs*67) associated with autosomal recessive the informative markers studied were located downstream MC which would also produce a truncated protein with the target mutation and not around it. Thus, analysis of a stop codon in the same position (amino acid 793). The additional markers could be done in future as well as an only in silico tool which predicts effects of deletion vari- increased number of patients studied. ants (MutationTaster), predicts a pathogenic effect for this In conclusion, different aspects of molecular genetics variant. of MC disease have been addressed in this study: novel It is interesting that in this cohort we have detected variants description, segregation analysis, Spanish MC c.568G>A (p.Gly190Arg) pathogenic variant acting as cohort description and identification of highly prevalent a dominant and recessive mutation in the same fam- mutations in CLCN1 gene, different phenotypes preva- ily (figure 1, family I). c.568G>A variant (rs369773321, lence (Becker vs Thomsen) and finally a founder effect MAF 0.008%) was previously described in a recessive study of the c.180+3A>T variant. pedigree (Skálová et al. 2013) and has been classified as This data will contribute to a better knowledge of the a likely pathogenic variant in Human Gene Mutation genotypic and phenotypic characteristics of MC patients. Database. This variant affects the glycine residue form- − ing the highly conserved Cl selectivity filter. The variant Acknowledgements c.180+3A>T has a global frequency at around 0.03– 0.04% (ExAc, Exome Variant Server and 1000 Genomes). We thank Ruben de Sancho and Amparo García Cardenal The reported frequency for c.180 + 3A>T in specific for their technical contribution in carrying out the experi- populations looks like artefactual, may be because a small ments and Rocio Mena and Maria Victoria Gomez for capillary Mutation spectrum of CLCN1 in Spanish patients Page 9 of 10 71 sequencing work. We also thank Will Brooks for improving the Koch M. C., Steinmeyer K., Lorenz C., Ricker K., Wolf F., Otto English language. This work was supported by a fellowship from M. et al. 1992 The skeletal muscle chloride channel in domi- the ‘Fundación J.L. Castaño’ in 2013 to C. Palma. CLCN1 nant and recessive human myotonia. Science 257, 797–800. gene homepage - Shared database [WWW Document], n.d. Landrum M. J., Lee J. M., Benson M., Brown G., Chao C., Chi- URL https://databases.lovd.nl/shared/genes/CLCN1(accessed4. tipiralla S. et al. 2015 ClinVar: public archive of interpretations 18.18). of clinically relevant variants. Nucleic Acids Res. 44, D862- D868. References Lehmann-Horn F. and Jurkat-Rott K. 1999 Voltage-gated ion channels and hereditary disease. Physiol. Rev. 79, 1317–1372. Lehmann-Horn F., Jurkat-Rott K. and Rüdel R. 2008 Diagnos- Adzhubei I. A., Schmidt S., Peshkin L., Ramensky V.E., Gerasi- tics and therapy of muscle channelopathies–guidelines of the mova A., Bork P. et al. 2010 A method and server for predicting Ulm muscle centre. Acta Myol. 27, 98–113. damaging missense mutations. Nat. Methods 7, 248–249. LiB.,KrishnanV.G.,MortM.E.,XinF.,KamatiK.K., Auton A., Abecasis G. R., Altshuler D. M., Durbin R. M., Bent- Cooper D. N. et al. 2009 Automated inference of molecular ley D. R., Chakravarti A. et al. 2015 A global reference for mechanisms of disease from amino acid substitutions. Bioin- human genetic variation. Nature 526, 68–74. formatics 25, 2744–2750. Barrett J. C., Fry B., Maller J. and Daly M. J. 2005 Haploview: Lipicky R. J., Bryant S. H. and Salmon J. H. 1971 Cable param- analysis and visualization of LD and haplotype maps. Bioin- eters, sodium, potassium, chloride, and water content, and formatics 21, 263–265. potassium efflux in isolated external intercostal muscle of nor- Becker P. E. 1979 Heterozygote manifestation in recessive gener- mal volunteers and patients with myotonia congenita. J. Clin. alized myotonia. Hum. Genet. 46, 325–329. Invest. 50, 2091–2103. Berezin C., Glaser F., Rosenberg J., Paz I., Pupko T., Fariselli Lossin C. and George A. L. 2008 Myotonia congenita. Adv. P. et al. 2004 ConSeq: the identification of functionally and Genet. 63, 25–55. structurally important residues in protein sequences. Bioinfor- Matthews E., Tan S. V, Fialho D., Sweeney M. G., Sud R. and matics 20, 1322–1324. Haworth A. 2008 What causes paramyotonia in the United Brugnoni R., Kapetis D., Imbrici P., Pessia M., Canioni E., Kingdom? Common and new SCN4A mutations revealed. Colleoni L. et al. 2013 A large cohort of myotonia congenita Neurology 70, 50–53. probands: novel mutations and a high-frequency mutation Mazón M. J., Barros F., De la Peña P., Quesada J. F., Escudero region in exons 4 and 5 of the CLCN1 gene. J. Hum. Genet. 58, A., Cobo A. M. et al. 2012 Screening for mutations in Spanish 581–587. families with myotonia. Functional analysis of novel mutations Chun S. and Fay J.C. 2009 Identification of deleterious mutations in CLCN1 gene. Neuromuscul. Disord. 22, 231–243. within three human genomes. Genome Res. 19, 1553–1561. Meyer-Kleine C., Steinmeyer K., Ricker K., Jentsch T. J. and Colding-Jørgensen E. 2005 Phenotypic variability in myotonia Koch M. C. 1995 Spectrum of mutations in the major human congenita. Muscle Nerve 32, 19–34. skeletal muscle chloride channel gene (CLCN1) leading to Consortium E. A., Lek M., Karczewski K., Minikel E., Samocha myotonia. Am. J. Hum. Genet. 57, 1325–1334. K., Banks E. et al. 2015 Analysis of protein-coding genetic Papponen H., Toppinen T., Baumann P., Myllylä V., Leisti J., variation in 60,706 humans. Nature 536, 285–291. Kuivaniemi H. et al. 1999 Founder mutations and the high Desmet F.-O., Hamroun D., Lalande M., Collod-Béroud G., prevalence of myotonia congenita in northern Finland. Neu- Claustres M. and Béroud C. 2009 Human Splicing Finder: an rology 53, 297–302. online bioinformatics tool to predict splicing signals. Nucleic Pusch M. 2002 Myotonia caused by mutations in the muscle chlo- Acids Res. 37, e67. ride channel gene CLCN1. Hum. Mutat. 19, 423–434. Dunø M. and Colding-Jørgensen E. 1993 Myotonia congenita, Reva B., Antipin Y.and Sander C. 2011 Predicting the functional GeneReviews. University of Washington, Seattle (https:// impact of protein mutations: application to cancer genomics. databases.lovd.nl/shared/variants/CLCN1/unique). Nucleic Acids Res. 39, e118. Dunø M. and Colding-Jørgensen E. 2005 Myotonia congénita, Richards S., Aziz N., Bale S., Bick D., Das S., Gastier-Foster J. GeneReviews. University of Washington, Seattle (https://www. et al. 2015 Standards and guidelines for the interpretation of ncbi.nlm.nih.gov/books/NBK1355/). sequence variants: a joint consensus recommendation of the Estévez R., Pusch M., Ferrer-Costa C., Orozco M. and Jentsch American College of Medical Genetics and Genomics and the T. J. 2004 Functional and structural conservation of CBS Association for Molecular Pathology. Genet. Med. 17, 405– domains from CLC chloride channels. J. Physiol. 557, 363– 423. 378. Schwarz J. M., Cooper D. N., Schuelke M. and Seelow D. Fialho D., Schorge S., Pucovska U., Davies N. P., Labrum R., 2014 MutationTaster2: mutation prediction for the deep- Haworth A. et al. 2007 Chloride channel myotonia: exon 8 sequencing age. Nat. Methods 11, 361–362. hot-spot for dominant-negative interactions. Brain 130, 3265– Sherry S. T., Ward M. H., Kholodov M., Baker J., Phan L., 3274. Smigielski E. M. and Sirotkin K. 2001 dbSNP: the NCBI Gabriel S. B., Schaffner S. F., Nguyen H., Moore J. M., Roy J., database of genetic variation. Nucleic Acids Res. 29, 308–311. Blumenstiel B. et al. 2002 The structure of haplotype blocks in Simkin D. and Bendahhou S. 2011 Skeletal muscle na channel dis- the human genome. Science 296, 2225–2229. orders. Front. Pharmacol. 2, 63. (https://doi.org/10.3389/fphar. Jentsch T. J., Poet M., Fuhrmann J. C. and Zdebik A. A. 2011.00063). 2005 Physiological functions of Clc Cl-Channels gleaned from Skálová D., Zídková J., Vohánkaˇ S., Mazanec R., Mušová Z., human genetic disease and mouse models. Annu. Rev. Physiol. Vondrácekˇ P. et al. 2013 CLCN1 Mutations in Czech patients 67, 779–807. with myotonia congenita, in silico analysis of novel and known Jurkat-Rott K., Holzherr B., Fauler M. and Lehmann-Horn F. mutations in the human dimeric skeletal muscle chloride chan- 2010 Sodium channelopathies of skeletal muscle result from nel. PLoS One 8, e82549. gain or loss of function. Pflügers Arch. Eur. J. Physiol. 460, Stenson P.D., Mort M., Ball E. V,Shaw K., Phillips A., Cooper D. 239–248. N. et al. 2014 The Human Gene Mutation Database: building 71 Page 10 of 10 Carmen Palma Milla et al.

a comprehensive mutation repository for clinical and molec- Tennessen J. A., Bigham A. W., O’Connor T. D., Fu W., Kenny ular genetics, diagnostic testing and personalized genomic E. E., Gravel S. et al 2012 Evolution and functional impact of medicine. Hum. Genet. 133, 1–9. rare coding variation from deep sequencing of human exomes. Suetterlin K., Männikkö R. and Hanna M. G. 2014 Muscle chan- Science 337, 64–69. nelopathies: recent advances in genetics, pathophysiology and Trip J., Drost G., Verbove D. J., van der Kooi A. J., Kuks J. B. therapy. Curr. Opin. Neurol. 27, 583–590. M., Notermans N. C. et al. 2008 In tandem analysis of CLCN1 Sun C., Tranebjaerg L., Torbergsen T., Holmgren G. and Van and SCN4A greatly enhances mutation detection in families Ghelue M. 2001 Spectrum of CLCN1 mutations in patients with non-dystrophic myotonia. Eur. J. Hum. Genet. 16, 921– with myotonia congenita in Northern Scandinavia. Eur. J. 929. Hum. Genet. 9, 903–909.

Corresponding editor: H. A. Ranganath