Immediate Communication

Molecular Psychiatry (2003) 8, 186–194. doi:10.1038/

Sodium channels SCN1A, SCN2A and SCN3A in familial autism

The AGRE Consortium:22 AGRE Scientific Advisory Board, Chairman, DH Geschwind, MD, PhD, UCLA, Los Angeles, CA, Members, M Bucan, PhD, University of Pennsylvania, Philadelphia, PA; WT Brown, MD, PhD, FACMG, NYS Institute for Basic Research in Developmental Disabilities, Long Island, NY; JD Buxbaum, PhD, Mt Sinai School of Medicine, New York; TC Gilliam, PhD, Columbia Genome Center, New York; DA Greenberg, PhD, Mt Sinai Medical Center, New York; DH Ledbetter, PhD, University of Chicago, Chicago, IL; B Miller, MD, UCSF, San Francisco, CA; SF Nelson, MD, UCLA School of Medicine, Los Angeles, CA; J Pevsner, PhD, Kennedy Krieger Institute, Baltimore, MD; JI Rotter, MD, Cedars-Sinai Medical Center, Los Angeles, CA; GD Schellenberg, PhD, University of Washington, Seattle, WA; CA Sprouse, EdD, Children's National Medical Center, Baltimore, MD; RE Tanzi, PhD, Massachusetts General Hospital, Boston, MA; KC Wilhelmsen, MD, PhD, University of California, San Francisco, CA, and AGRE Collaborator: JM Silverman, PhD, Mt Sinai Medical School, New York, NY.

L A Weiss1, A Escayg1, J A Kearney1, M Trudeau1, B T MacDonald1, M Mori2, J Reichert3, J D Buxbaum3 and M H Meisler1

  1. 1Department of Human Genetics, University of Michigan, Ann Arbor, MI, USA
  2. 2Department of Physiological Sciences, Graduate University for Advanced Studies, National Institute for Physiological Sciences, Okazaki, Japan
  3. 3Seaver Autism Research Center, Department of Psychiatry, Mount Sinai School of Medicine, New York, NY, USA

Correspondence: Dr MH Meisler, PhD, Department of Human Genetics, University of Michigan, 4708 Medical Science II, Ann Arbor, MI 48109-0618, USA. E-mail:

Received 23 May 2002; Revised 12 July 2002; Accepted 29 June 2002.



Autism is a psychiatric disorder with estimated heritability of 90%. One-third of autistic individuals experience seizures. A susceptibility locus for autism was mapped near a cluster of voltage-gated sodium channel genes on chromosome 2. Mutations in two of these genes, SCN1A and SCN2A, result in the seizure disorder GEFS+. To evaluate these sodium channel genes as candidates for the autism susceptibility locus, we screened for variation in coding exons and splice sites in 117 multiplex autism families. A total of 27 kb of coding sequence and 3 kb of intron sequence were screened. Only six families carried variants with potential effects on sodium channel function. Five coding variants and one lariat branchpoint mutation were each observed in a single family, but were not present in controls. The variant R1902C in SCN2A is located in the calmodulin binding site and was found to reduce binding affinity for calcium-bound calmodulin. R542Q in SCN1A was observed in one autism family and had previously been identified in a patient with juvenile myoclonic epilepsy. The effect of the lariat branchpoint mutation was tested in cultured lymphoblasts. Additional population studies and functional tests will be required to evaluate pathogenicity of the coding and lariat site variants. SNP density was 1/kb in the genomic sequence screened. We report 38 sodium channel SNPs that will be useful in future association and linkage studies.


sodium channel, autism, genetic susceptibility, neurogenetics, Chr 2, calmodulin



Autism [MIM 209850] is a complex psychiatric disorder characterized by impaired communication and social skills, and restricted and repetitive behaviors.1 Half of the children with autism are nonverbal and between one half and three-quarters are mentally retarded. Epilepsy is a complication in one-third of autistic children, with the most severe seizures occurring in adolescence.2 In one study, 82% of children with regressive autism exhibited epileptiform activity.3 The population incidence of autism is approximately 0.1%.4

Autism appears to have a strong genetic component, with an estimated heritability above 90%.5 The disease risk of siblings is 20–40 times higher than the general population.6 Genetic modelling suggests that epistatic interaction between 3 and 15 genes may be responsible for the disorder.6,7 The broad phenotypic variation in autism spectrum disorders is consistent with the evidence of genetic heterogeneity. Specific rearrangements of chromosomes 7, 15 and X have been observed in autistic individuals, and evidence for linkage to markers on chromosome 7q was obtained in large-scale genome scans.6 Mutations in the X-linked Rett syndrome gene MECP2 have been identified in some patients with infantile autism,8 but were not observed in another survey of 59 patients.9 One chromosome region that generated positive LOD scores in early genome scans of autistic families is on Chr 2q.10,11 In recent studies, LOD scores approaching 4 were obtained with markers located between 176 and 180 cM on chromosome 2.12,13,14 Restriction to cases with delayed onset of phrase speech (PSD) increased the evidence for linkage on chromosome 2.12

The sodium channel alpha subunit genes SCN1A, SCN2A and SCN3A are located at approximately 175 cM on chromosome 2, band 2q24.15 These voltage-gated sodium channels are expressed in neurons and glia throughout the central and peripheral nervous system16,17 and have been highly conserved during invertebrate and vertebrate evolution.18 Missense mutations in SCN1A and SCN2A have been identified in the inherited seizure disorder Generalized Epilepsy with Febrile Seizures Plus (GEFS+2, MIM 604236, 604233).19 De novo null mutations in SCN1A result in severe myoclonic epilepsy of infancy, a syndrome that includes mental retardation.20 Mice expressing a mutated SCN2A channel with persistent current exhibit seizures and repetitive behaviors.21 The association of seizures with mutations in the sodium channel genes, and their location near an autism susceptibility locus, suggested that they might play a role in susceptibility to autism.

The Autism Genetic Research Exchange (AGRE) was developed to facilitate collaborative genetic research into the etiology of autism, by making biomaterials from well-characterized families widely available to the scientific community.4,22 Each AGRE family includes two parents and 2–6 sibs, at least 2 of whom are affected. Genomic DNA and immortalized cell lines are available for several hundred families. We have studied these multiplex families in order to identify novel mutations in SCN1A, SCN2A and SCN3A and evaluate their cosegregation with the disorder.


Material and methods


DNA (10 mug) from one affected individual in each of 229 multiplex families was purchased from the AGRE. Diagnosis was based on the autism diagnostic interview-revised protocol (ADI-R)1 and reflects the accepted criteria of the DSM-IV/ICD-10.22 A representative subset of AGRE families was analyzed. Linkage data for these families were not available at the outset of the study. In all, 11 additional samples were obtained from patients with sib-pair analysis consistent with linkage to chromosome 2.12 polyA+ RNA was extracted from immortalized lymphoblast cell lines obtained from AGRE and cultured under standard conditions.

Mutation detection

A total of 26 coding exons of SCN1A were amplified in 28 PCR reactions and analyzed by conformation sensitive gel electrophoresis (CSGE) as previously described.15,23 PCR products containing variants were gel purified using Qiaex II or QIAquick (Qiagen) and manually sequenced with ThermoSequenase (USB). PCR primers flanking the exons of SCN2A and SCN3A were designed from intronic sequences in the public databases. The polymorphism C(6015+856)T in the 1.7 kb 3'UTR of SCN2A was detected in the 499 bp fragment obtained by amplification of genomic DNA with the primers F2 (AGG TTG TTT ACT ATT ATA TGT GAC) and R2 (TTT TTA TTT TAT GCC AAA GGACAC). SCN1A and SCN2A were screened in 117 AGRE families and SCN3A was screened in a subset of 32 families consistent with Chr 2 linkage.

RT-PCR of lymphoblast RNA and primer extension reaction

To evaluate the effect of the lariat mutation in intron 6, total RNA was isolated from lymphoblast cell lines. Reverse transcription of 5 mug of total RNA from lymphoblast cell lines was performed with poly(dT) and random hexamers as previously described.24 First-strand cDNA (1 mul) was amplified with the forward primer 6F in exon 6 (TCC TTT GAA ATA AAT ATC ACT TCC) and the reverse primer 10R in exon 10 (TAG ATG CTA CTG AAG AAC TCT CTG). For the primer extension reaction, RT-PCR was carried out with primers F2 and R2 (above). PCR products were gel purified using the Qiaquick gel extraction kit (Qiagen). The reverse primer R5 (GAC TGT GAA TGA CAA TGT AAT GGT A) was gel purified and used for the primer extension reaction with 1 mul of PCR product as template, using the ThermoSequenase kit (USB) in the presence of 0.1 muCi [alpha-33P] ddATP and unlabelled dCTP, dTTP and dGTP. Reaction products were resolved on a 12% denaturing polyacrylamide gel and visualized by autoradiography.

Calmodulin binding

The interaction between calmodulin and the synthetic peptide NaIQ was measured by fluorescence spectroscopy using dansylated calmodulin as previously described.25 Fluorescence spectra from the wild-type peptide were compared with a peptide containing the SCN2A variant R1902C. Peptide titration data were analyzed using the MicroCal Origin software and fitted with a two-site binding model to obtain values for the dissociation constants, using the formula y=Bmax1*x/(k1+x)+Bmax2*x/(k2+x).



Five rare coding variants in autism families

The 26 coding exons of SCN1A and SCN2A were screened in an affected individual from 117 AGRE families. The four exons containing variants were then screened in an additional 112 families. Five missense variants were identified, each present in 1/229 families: R542Q, I1034T, F1038L, I1955T and R1902C (Table 1). The positions of the amino acid substitutions in the transmembrane channel protein are shown in Figure 1. The mutation R542Q in SCN1A was inherited by 2 affected sibs (Figure 2), and was previously observed in a patient with juvenile myoclonic epilepsy.15 This mutation disrupts the predicted tyrosine kinase site RLTYEKRY in cytoplasmic loop 1. The mutation I1034T in SCN1A was present in 1 of 2 affected sibs and not in the unaffected sib (Figure 1 and Figure 2). The mutation F1038L in SCN1A was present in 2 affected sibs (Figure 2). The missense mutation R1902C in SCN2A was detected in 1 affected individual from family AU0247 but not in his affected sib (Figure 2) nor in 296 controls. Residue R1902 is conserved in mammalian and invertebrate sodium channels and is located within a calmodulin binding site (see below). The mutation I1955T was identified in an affected individual from family AU0203; the other members of this family were not available for testing (pedigree not shown).

Figure 1.
Figure 1 - Unfortunately we are unable to provide accessible alternative text for this. If you require assistance to access this image, please contact or the author

Positions of rare and polymorphic coding substitutions in three sodium channel proteins: L1, cytoplasmic loop 1; L2, cytoplasmic loop 2. SCN1A, circles; SCN2A, squares; SCN3A, stars. Exons are numbered.

Full figure and legend (31K)

Figure 2.
Figure 2 - Unfortunately we are unable to provide accessible alternative text for this. If you require assistance to access this image, please contact or the author

Segregation of rare coding variants in multiplex autism families. Beneath each pedigree, the results of conformation-sensitive electrophoresis are shown. Heterozygosity for sequence variation is presented for one individual from each family. Solid symbols, autism; shaded symbol, pervasive developmental disorder; m, mutant allele; N, nonverbal; V, verbal; R, regressive disorder.

Full figure and legend (123K)

The 26 coding exons of SCN3A were screened in a subset of 32 AGRE families; in these families, markers for chr 2q24 were concordant with autism. No rare coding variants were observed; two polymorphic coding variants of SCN3A are described in a later section (Table 1).

Functional effect of R1902C on calcium–calmodulin binding

R1902C is located within a 27 amino-acid peptide of SCN2A that contains a functional binding site for calmodulin.25 The mutant peptide was synthesized and the binding affinity for calcium–calmodulin was determined. The Kd was 304 nM for the mutant peptide and 141 nM for the wild-type peptide (Figure 3), demonstrating reduced affinity of the mutant peptide for the low-affinity binding site for calcium-bound calmodulin. The binding constant for calcium-free calmodulin was not affected (data not shown), consistent with evidence that interaction with calcium-free calmodulin occurs at the IQ motif residues that are not affected by this mutation.25

Figure 3.
Figure 3 - Unfortunately we are unable to provide accessible alternative text for this. If you require assistance to access this image, please contact or the author

Reduced binding of calcium–calmodulin because of the amino-acid substitution R1902C in SCN2A. Dansylated synthetic peptides containing the wild-type or mutant sequence were incubated with calcium–calmodulin as previously described.25 Closed circles, wildtype; open squares, mutant.

Full figure and legend (19K)

A branchpoint mutation in intron 6 of SCN2A

The nucleotide substitution A(970-32)G was identified in an affected individual from family AU0207. This variant alters the essential adenosine residue of a lariat consensus element (YNYRAY) that is located 32 bp upstream of exon 7 in SCN2A (Figure 4). Substitution of G for A at the lariat branchpoint is known to reduce splicing efficiency,26,27 suggesting that this intronic variant could have functional significance. We therefore screened an additional 112 patients and identified 2 more families carrying the same mutation (Figure 4). The heterozygote frequency was 3/229 in unrelated autism patients and 3/450 in a group of unrelated controls. In 2 families, the branchpoint variant was inherited by affected sibs only, but in family AU0509, 5 unaffected offspring also carry the mutation (Figure 4). It is interesting that almost all the male heterozygotes in these families are affected (7/8), while none of the female heterozygotes are affected (0/7) (P<0.01).

Figure 4.
Figure 4 - Unfortunately we are unable to provide accessible alternative text for this. If you require assistance to access this image, please contact or the author

Segregation of the branchpoint mutation A(970-32)G in three autism families. (a) Segregation, conformation-sensitive electrophoresis and heterozygous sequence; symbols as defined in Figure 2. (b) Position of the branch site mutation. (c) Genotypes of two additional families. N, nonverbal; V, verbal; R, regressive disorder.

Full figure and legend (106K)

To investigate the effect of the intron 6 branchpoint mutation on splicing of exon 7, we analyzed lymphoblast RNA from family AU0207. RT-PCR was carried out using a forward primer in exon 6 and a reverse primer in exon 10 (Figure 5a). The product obtained from II-1 was identical in length to the wild-type product from I-1, indicating that splicing is qualitatively normal (Figure 5b).

Figure 5.
Figure 5 - Unfortunately we are unable to provide accessible alternative text for this. If you require assistance to access this image, please contact or the author

Effect of the branchpoint mutation on splicing of exon 6. Total RNA was isolated from cultured lymphoblasts and analyzed by RT-PCR and primer extension (see text for details). (a) Location of primers in exons 6 and 10. (b) Length of RT-PCR product obtained with these primers is not affected by the branchpoint mutation, indicating that exon 6 is spliced normally. (c) Primer extension analysis of the effect of the branchpoint mutation on transcript abundance. Genotypes are indicated for the branchpoint nucleotide (A or G) and for the SNP in the 3' UTR (allele 1=C(6015+856), allele 2=T (6015+856). On the reverse strand, allele 1 incorporates a G at the variable position and gives an A-labelled product of 31 bp; allele 2 incorporates an A at the variable position to give a product of 27 bp. Primer extension products were separated on a 12% acrylamide gel.

Full figure and legend (66K)

To evaluate the effect of the branchpoint mutation on splicing efficiency, we identified individuals who were heterozygous for both the branchpoint mutation and an exonic SNP that would enable us to distinguish between the two transcripts. We sequenced a 499 bp fragment from the 3' UTR and determined that an affected male and his unaffected father (Figure 4a) are heterozygous for the variant C(6015+856)T. The affected male is also heterozygous for the branchpoint mutation while the father is homozygous for the wild-type A at the branchpoint. To compare the abundance of the two SCN2A transcripts in these individuals, the 3' UTR was amplified by RT-PCR and the products were analyzed with a primer extension assay using a primer downstream of the C/T variant in the presence of one radiolabelled dideoxy nucleotide triphosphate. The radiolabelled primer extension products were 31 bp for the C allele and 27 bp for the T allele. Analysis of genomic DNA served as a control demonstrating the products obtained from equal abundance of the two alleles (Figure 5c, lane 3). The same ratio was observed with RNA from the branchpoint A/G heterozygote and the A/A homozygote (Figure 5c). This result demonstrates that the branchpoint mutation did not alter splicing efficiency of the SCN2A transcript in lymphoblast cells.

Polymorphic coding variants (cSNPs)

Four polymorphic coding variants were observed at equal frequencies in patients and controls: T1067A, R19K, DeltaN43 and G1813S (Figure 1 and Table 1). The 2 alleles of T1067A in SCN1A exhibit frequencies of 0.67 and 0.33 in the patient population (Table 1), in agreement with our previous analysis of a control population.15 The variant R19K in SCN2A was detected in 6/117 affected individuals and 3/48 controls. The deletion DeltaN43 in SCN3A was present in 12/117 patients and 5/57 controls. The rat and mouse orthologs of human SCN3A contain the N43 residue, indicating the minor human allele is the product of a recent 3 bp deletion that occurred within an imperfect trinucleotide repeat (Figure 6). The residue G1813 in SCN3A is located in an evolutionarily conserved region of the C-terminal domain. Either glycine or alanine is present in all of the sequenced vertebrate and invertebrate sodium channels represented in GenBank. The variant G1813S was detected in 3/117 patients and in 2/121 controls. G1813S did not cosegregate with autism in the 3 patient families (not shown).

Figure 6.
Figure 6 - Unfortunately we are unable to provide accessible alternative text for this. If you require assistance to access this image, please contact or the author

Simple sequence surrounding the DeltaN43 variant of SCN3A. The wild-type allele in human, mouse and rat contains an amino acid that is deleted in the variant. The population frequency of the deletion allele is 0.05 (Table 1).

Full figure and legend (17K)

Noncoding variants in the sodium channel genes

The sodium channel primer pairs amplified a total of 27 kb of coding sequence and 3 kb of adjacent intronic sequence from genomic DNA. We detected 37 noncoding SNPs, 2 single-base insertions, 2 single-base deletions, and 1 dinucleotide deletion (Table 2 and Table 3), for a density of 1 variant per 700 bp. Many of the variants were restricted to 1 or 2 individuals. Haplotype frequencies for SCN1A in the AGRE families were in agreement with those previously described,15 except that a novel SCN1A haplotype was observed in 2 autistic individuals (Table 4).



Potential disease mutations

Six potential disease mutations were observed, each in only 1 AGRE family. The substitution of G for the branchpoint site A in intron 6 of SCN1A (TTTTAAT to TTTTAGT) alters the consensus sequence YNYTRAY and is predicted to reduce splicing efficiency.26,27 The association of haploinsufficiency of SCN1A with severe seizures20 supports the possibility that a hypomorphic allele of SCN2A could contribute to autism. However, this mutation did not reduce the efficiency of removal of intron 6 in lymphoblasts. In view of other examples of neuron-specific splicing defects, for example, in familial dysautonomia,28 we cannot exclude the possibility that this mutation reduces splicing efficiency in neurons. It is interesting that 7/8 of heterozygous males in the AGRE families are affected, while none of the 7 heterozygous females are affected (P<0.01) (Figure 4). If we include the 2 heterozygous female controls and 1 heterozygous male control, the incidence of autism is 7/9 in male heterozygotes and 0/9 in female heterozygotes. This may indicate that males are more sensitive to the branchpoint mutation.

The R1902C substitution is located within the Ca2+–calmodulin binding site of SCN2A, which was identified in a yeast 2-hybrid screen.25 The sodium channel/calmodulin complex assumes 2 alternative conformational states depending on the Ca(2+) concentration. The amino-acid substitution R1902C caused a two-fold reduction in Kd for Ca(2+)CaM with no change in binding of apoCaM. This small difference might affect subcellular localization or modulation of channel properties. R1902C was observed in 1/117 AGRE families and 0/296 controls. One affected sib did not carry the mutation. Mutation of a residue in the cardiac sodium channel that is located closely to the corresponding arginine alters channel inactivation and causes disease.29 Additional screening for R1902C in patients and controls would be worthwhile.

The 5 coding variants that were only detected in affected individuals may contribute to autism: R542Q, I1034T, F1038L, I1955T and R1902C (Table 1). The fact that I1034T and R1902C were not inherited by an affected sib does not rule them out as candidate mutations, since affected individuals in the same family may carry a different combination of predisposing mutations for a polygenic disorder. Follow-up experiments that could provide evidence for pathogenicity of these variants include screening more autism patients to find additional examples of each variant, functional analysis of channel kinetics, and generation of transgenic mice carrying the mutations. Although the variants are not located within transmembrane protein domains of known function, the >95% conservation of the cytoplasmic loops in human and rodent orthologs suggests that they contain as yet uncharacterized functional elements.18

Mutations affecting protein sequence or splice sites of the sodium channel genes were found in 6/117 AGRE families and are clearly not a major cause of autism in this group. Transcriptional regulatory elements were not included in the survey, and remain a potential source of quantitative variation and disease.

Sodium channel mutations and neurological disease

Small alterations in sodium channel properties have been associated with neurological disability in vivo. For example, a 10 mV shift in the voltage dependence of activation of mouse SCN8A alters the repetitive firing pattern of cerebellar Purkinje neurons and produces ataxia.30,31 A small increase in persistent current in SCN2A produces a seizure disorder in transgenic mice.21 SCN1A mutations responsible for GEFS+ cause only minor changes in channel kinetics.23,32 Polymorphic sodium channel variants, such as those described here, may contribute to the range of cognitive and emotional function in normal individuals. The combined effect of variants in multiple ion channel genes could also predispose to psychiatric disorders such as autism. This hypothesis is consistent with recent genetic models indicating that segregation of multiple rare alleles may account for common polygenic disorders in the human population.33 Because of the low frequency of each variant allele, large-scale comparisons of thousands of patients and controls will be required to obtain statistically significant association data for such variants. It will also be important to test the effect of each putative disease variant in functional assays of channel activity.



  1. Lord C, Cook EH, Leventhal BL, Amaral DG. Autism spectrum disorders. Neuron 2000; 28: 355–363. | Article | PubMed | ISI | ChemPort |
  2. Rapin I. Current concepts: autism. N Engl J Med 1997; 337: 97–104.
  3. Lewine JD, Andrews R, Chez M, Patil A-A, Devinsky O, Smith M et al. Magnetoenchephalographic patterns of epileptiform activity in children with regressive autism spectrum disorders. Pediatrics 1999; 104: 405–418.
  4. Liu J, Nyholt DR, Magnussen P, Parano E, Pavone P, Geschwind D et al. A genomewide screen for autism susceptibility loci. Am J Hum Genet 2001; 69: 327–340. | Article | PubMed | ISI | ChemPort |
  5. Bailey A, Le Couteur A, Gottesman I, Bolton P, Simonoff E, Yuzda E et al. Autism as a strongly genetic disorder: evidence from a British twin study. Psychol Med 1995; 25: 63–77. | PubMed | ISI | ChemPort |
  6. Lamb JA, Moore J, Bailey A, Monaco AP. Autism: recent molecular genetic advances. Hum Mol Genet 2000; 9: 861–868. | Article | PubMed | ISI | ChemPort |
  7. Risch N, Spiker D, Lotspeich L, Nouri N, Hinds D, Hallmayer J et al. A genomic screen of autism: evidence for a multilocus etiology. Am J Hum Genet 1999; 65: 493–507. | Article | PubMed | ISI | ChemPort |
  8. Lam CW, Yeung WL, Ko CH, Poon PM, Tong SF, Chan KY et al. Spectrum of mutations in the MECP2 gene in patients with infantile autism and Rett syndrome. J Med Genet 2000; 37: E41. | Article | PubMed | ChemPort |
  9. Vourc'h P, Bienvenu T, Beldjord C, Chelly J, Barthelemy C, Muh JP et al. No mutations in the coding region of the Rett syndrome gene MECP2 in 59 autistic patients. Eur J Hum Genet 2001; 9: 556–558. | Article |
  10. International Molecular Genetic Study of Autism Consortium (IMGSAC). A full genome screen for autism with evidence for linkage to a region on chromosome 7q. Hum Mol Genet 1998; 7: 571–578. | Article | PubMed | ISI |
  11. Philippe A, Martinez M, Guiloud-Bataille M, Gillberg C, Rastam M, Sponheim E et al. Genome-wide scan for autism susceptibility genes. Hum Mol Genet 1999; 8: 805–812. | Article | PubMed | ISI | ChemPort |
  12. Buxbaum JD, Silverman JM, Smith CJ, Kilifarski M, Reichert J, Hollander E et al. Evidence for a susceptibility gene for autism on chromosome 2 and for genetic heterogeneity. Am J Hum Genet 2001; 68: 1514–1520. | Article | PubMed | ISI | ChemPort |
  13. IMGSAC. A genomewide screen for autism: strong evidence for linkage to chromosomes 2q, 7q, and 16p. Am J Hum Genet 2001; 69: 570–581. | Article | PubMed | ISI |
  14. Shao Y, Raiford KL, Wolpert CM, Cope HA, Ravan SA, Ashley-Koch AA et al. Phenotypic homogeneity provides increased support for linkage on chromosome 2 in autistic disorder. Am J Hum Genet 2002; 70: 1058–1061. | Article | PubMed | ISI | ChemPort |
  15. Escayg A, Heils A, MacDonald BT, Haug K, Sander T, Meisler MH. A novel GEFS+ mutation in sodium channel SCN1A and prevalence of variants in epilepsy patients. Am J Hum Genet 2001; 68: 866–873. | Article | PubMed | ISI | ChemPort |
  16. Felts PA, Yokoyama S, Dib-Hajj S, Black JA, Waxman SG. Sodium channel alpha-subunit mRNAs I, II, III, NaG, Na6 and hNE (PN1); different expression patterns in developing rat nervous system. Mol Brain Res 1997; 45: 71–82.
  17. Whitaker WR, Clare JJ, Powell AJ, Chen YH, Faull RL, Emson PC. Distribution of voltage-gated sodium channel alpha-subunit and beta-subunit mRNAs in human hippocampal formation, cortex, and cerebellum. J Comp Neurol 2000; 422: 123–139. | Article | PubMed | ChemPort |
  18. Plummer NW, Meisler MH. Evolution and diversity of mammalian sodium channel genes. Genomics 1999; 57: 323–331. | Article | PubMed | ISI | ChemPort |
  19. Meisler MH, Kearney J, Ottman R, Escayg A. Identification of epilepsy genes in human and mouse. Annu Rev Genet 2001; 35: 567–588. | Article | PubMed | ISI | ChemPort |
  20. Claes L, Del-Favero J, Ceulemans B, Lagae L, Van Broeckhoven C, De Jonghe P. De novo mutations in the sodium-channel gene SCN1A cause severe myoclonic epilepsy of infancy. Am J Hum Genet 2001; 68: 1327–1332. | Article | PubMed | ISI | ChemPort |
  21. Kearney JA, Plummer NW, Smith MR, Kapur J, Cummins TR, Waxman SG et al. A gain-of-function mutation in the sodium channel gene Scn2a results in seizures and behavioral abnormalities. Neuroscience 2001; 102: 307–317. | Article | PubMed | ISI | ChemPort |
  22. Geschwind DH, Sowinski J, Lord C, Iversen P, Shestack J, Jones P et al. The autism genetic resource exchange: a resource for the study of autism and related neuropsychiatric conditions. Am J Hum Genet 2001; 69: 463–466.  | Article | PubMed | ISI | ChemPort |
  23. Escayg A, MacDonald BT, Meisler MH, Baulac S, Huberfeld G, An-Gourfinkel I et al. Mutations of SCN1A, encoding a neuronal sodium channel in two families with GEFS+2. Nat Genet 2000; 24: 343–345. | Article | PubMed | ISI | ChemPort |
  24. Escayg A, Jones JM, Kearney JA, Hitchcock PF, Meisler MH. Calcium channel beta4 (CACNB4): human ortholog of the mouse epilepsy gene lethargic. Genomics 1998; 50: 14–22.
  25. Mori M, Konno T, Ozawa T, Murata M, Imoto K, Nagayama K. Novel interaction of the voltage-dependent sodium channel (VDSC) with calmodulin: does VDSC acquire calmodulin-mediated Ca2+-sensitivity. Biochemistry 2000; 39: 1316–1323. | Article | PubMed | ISI | ChemPort |
  26. Gaur RK, McLaughlin LW, Green MR. Functional group substitutions of the branchpoint adenosine in a nuclear pre-mRNA and a group II intron. RNA 1997; 3: 861–869. | PubMed | ISI | ChemPort |
  27. Lund M, Tange TO, Dyhr-Mikkelsen H, Hansen J, Kjems J. Characterization of human RNA splice signals by iterative functional selection of splice sites. RNA 2000; 6: 528–544. | Article | PubMed | ISI |
  28. Slaugenhaupt SA, Blumenfeld A, Gill SP, Leyne M, Mull J, Cuajungco MP et al. Tissue-specific expression of a splicing mutation in the IKBKAP gene causes familial dysautonomia. Am J Hum Genet 2001; 68: 598–605. | Article | PubMed | ISI | ChemPort |
  29. Rook MB, Alshinawi CB, Groenewegen WA et al. Human SCN5A gene mutations alter cardiac sodium channel kinetics and are associated with the Brugada syndrome. Cardiovasc Res 1999; 44: 507–517.  | Article | PubMed | ISI | ChemPort |
  30. Raman IM, Sprunger LK, Meisler MH, Bean BP. Altered subthreshold sodium currents and disrupted firing patterns in purkinje neurons of Scn8a mutant mice. Neuron 1997; 19: 881–891. | Article | PubMed | ISI | ChemPort |
  31. Meisler MH, Kearney J, Escayg A, MacDonald BT, Sprunger LK. Sodium channels and neurological disease: insights from Scn8a mutations in the mouse. Neuroscientist 2001; 7: 136–145.
  32. Spampanato J, Escayg A, Meisler MH, Goldin AL. Functional effects of two voltage-gated sodium channel mutations that cause generalized epilepsy with febrile seizures plus type 2. J Neurosci 2001; 21: 7481–7490. | PubMed | ISI | ChemPort |
  33. Pritchard JK. Are rare variants responsible for susceptibility to complex diseases? Am J Hum Genet 2001; 69: 124–137. | Article | PubMed | ISI | ChemPort |


We gratefully acknowledge the resources provided by the participating AGRE families. This work was supported by the Wilson Family Research Foundation, The Seaver Autism Research Center, and Cure Autism Now. Fellowship support was provided by National Research Service Award NS10692 (JK) and the Genetics Training Program at the University of Michigan (T32 GM07544) (BM).