Mutations in LGI1 cause autosomal-dominant partial epilepsy with auditory features

Abstract

The epilepsies are a common, clinically heterogeneous group of disorders defined by recurrent unprovoked seizures. Here we describe identification of the causative gene in autosomal-dominant partial epilepsy with auditory features (ADPEAF, MIM 600512), a rare form of idiopathic lateral temporal lobe epilepsy characterized by partial seizures with auditory disturbances. We constructed a complete, 4.2-Mb physical map across the genetically implicated disease-gene region, identified 28 putative genes (Fig. 1) and resequenced all or part of 21 genes before identifying presumptive mutations in one copy of the leucine-rich, glioma-inactivated 1 gene (LGI1) in each of five families with ADPEAF. Previous studies have indicated that loss of both copies of LGI1 promotes glial tumor progression. We show that the expression pattern of mouse Lgi1 is predominantly neuronal and is consistent with the anatomic regions involved in temporal lobe epilepsy. Discovery of LGI1 as a cause of ADPEAF suggests new avenues for research on pathogenic mechanisms of idiopathic epilepsies.

Fig. 1

Transcript map of the genetically defined interval for ADPEAF on chromosome 10q24. The minimal genetic region believed to harbor the ADPEAF gene is shown. a,b, The speckled bars denote minimal genetic regions defined by the linkage studies of Ottman et al. (a; top bar) and Poza et al. (b; bottom bar). The 3-cM region of overlap was flanked by locus D10S200 on the centromeric boundary and by D10S577 at the telomeric boundary. Radiation hybrid mapping analysis predicted that the interval would span 16 cR, which corresponds to 4 Mb on human chromosome 10. Physical mapping further refined the region to 4.2 Mb. c, A transcript map of the region constructed by combining data from Human Gene Map 99 from NCBI with basic local alignment search tool predictions generated by comparison of 10q24 genomic DNA with the dbEST database. A total of 47 independent ESTs were identified, 28 of which contained unambiguous ORFs. These putative genes are listed in red, and microsatellite markers in black. Arrows indicate the direction of transcription. Black arrows indicate that the candidate gene was screened for ADPEAF-related mutations. SLIT1, a gene with a gene structure and expression similar to that of LGI1, is located in the interval approximately 3.2 Mb telomeric to LGI1.

Fig. 2

Segregation of putative disease alleles in families with ADPEAF. Family 6610 is the original family used to establish linkage between 10q24 DNA markers and disease. Filled bars mark the presence and boundaries of disease-related haplotypes, defined in families A, B and C by markers D10S185, D10S200, D10S198, D10S603, D10S192, D10S222 and D10S566, and in family 6610 by these markers excluding D10S192. The markers span a sex-averaged distance of 10.8 cM, according to the Marshfield map. Individuals who did not carry mutations are denoted by +/+, and those who carried one mutant and one normal allele by M/+. Original sequence tracings used to detect putative disease alleles are shown to the right of each family pedigree. Variant alleles are denoted by red arrows. Families 6610, A and B had insertion or deletion mutations, and the sequence traces show ‘signatures’ characteristic of heterozygous changes by which the diploid sequences became asynchronous, beginning with the mutant nucleotide. To be certain of the actual mutation, LGI1 DNA was subcloned and sequenced as haploid DNA from each of the families, and haploid cell lines were derived for family 6610 (GMP Companies; data not shown). Filled symbols represent individuals with idiopathic epilepsy; symbols containing a ‘?’ represent individuals classified as ‘unknown’, either because they had symptomatic epilepsy (sympt. epil.), febrile seizures (febrile sz.), alcohol-related acute symptomatic seizures (alcohol-related sz.) or isolated unprovoked seizure (isol. unprov. sz.), because they were under 20 y at the time of clinical assessment or because their final diagnosis was ‘possible’ (rather than definite) epilepsy (poss. epil.). A detailed clinical description of each family has been presented separately,,.

Fig. 3

Predicted effect of ADPEAF mutations on Lgi1. a, Structural organization of LGI1. LGI1 spans 36.9 kb and consists of eight exons ranging in size from 72 bp (exons 2–5) to 1,197 bp (exon 8). The mRNA transcript consists of a 262-bp 3′ UTR and a 1,674-bp ORF. Exon 1 includes a signal peptide, and exons 3–5 each contain a full 24-aa LRR motif (spanning from the fifth amino acid of one repeat to the fifth amino acid of the next). Putative mutations are illustrated by their corresponding family identifiers. Three were located in exon 8, one in exon 6 and one (from family B) in the 95-bp intron between exons 3 and 4. b, Predicted sequence motifs in Lgi1. This figure is a modification of those previously presented,. Beginning at the amino terminus (left), the protein contained a predicted signal peptide (filled arrow; positions 1–35), an N-terminal cysteine-rich LRR flanking sequence (LRRNT; residues 45–71), three LRR repeat sequences (orange boxes; residues 90–113, 114–137 and 138–161), a C-terminal cysteine-rich LRR flanking sequence (LRRCT; residues 173–222), two direct repeat sequences (Rep1 and 2; residues 226–361 and 420–549, respectively) and a putative membrane-spanning segment (yellow rectangle). A 22-bp transmembrane region (residues 288–309) was previously reported. We found support for this prediction using several methods, including the dense alignment surface (DAS) method for predicting integral membrane proteins,, although other methods (PRED-TMR) fail to predict the region. A fourth putative LRR repeat, predicted by Sommerville et al., resided N-terminally to the other three repeats and was encoded by exon 2. Although it seemed to share a common origin with the other repeats, its divergence at canonical residues made its role as a functional LRR repeat ambiguous. Chernova et al. predicted two potential N-glycosylation sites at Asn192 and Asn277 and potential phosphorylation sites for cAMP-dependent protein kinase (Ser 313), tyrosine kinase (Tyr 384) and several sites for PKC and casein kinase II. Lgi1 is predicted to consist of 557 amino-acid residues and to encode a slightly alkaline 60-kD protein (after removal of the signal peptide). c, Predicted effect of mutations on protein sequence. The predicted effect of each of the five presumptive ADPEAF mutations is depicted relative to the ‘normal’ protein shown to encode 557 amino-acid residues. The single–base pair insertion in family 6610 would predictably encode an Lgi1 protein with a normal complement of the first N-terminal 546 amino acids, followed by seven missense residues and four truncated residues. The Lgi1 extracellular region and the C-terminal region were highly conserved between human and mouse. Likewise, the frameshift mutations in families A and C would predictably alter protein structure and function dramatically. Family B showed a C→A transversion at position −3′ relative to the exon 4 acceptor splice site. Whereas positions −1 and −2 are typically conserved and often lead to cryptic site usage or exon skipping when altered, little is known about the effects of alteration at the −3 position. The presumptive mutation led to retention of intron 3 in a portion of LGI1 transcripts from affected individuals (Fig. 4). Family D showed a non-conservative missense mutation in exon 8.

Fig. 4

Aberrant LGI1 splicing in family B. a, LGI1 oligonucleotide primers specific for exon 3 and exon 6 were amplified using RT–PCR and mRNA isolated from lymphoblastoid cell lines. Lane 1, control sample; lanes 2–5, samples from four affected individuals in family B; lane 6, 100-bp size standards. All sample lanes showed a 286-bp fragment corresponding to the normal transcript. Samples from affected individuals also showed a second band that included a 72-bp insertion corresponding to the intact intron 3. b, Alignment of DNA sequencing traces corresponding to genomic DNA from a control individual (top panel) and an individual from family B with epilepsy (middle panel), together with the sequence trace corresponding to the aberrant RT–PCR fragment (bottom panel) derived from the same cell line represented in the middle panel. As expected, the genomic trace from the individual with epilepsy showed a C/A heterozygote at IVS3(–3), whereas the aberrant RT–PCR fragment contained exclusively adenine at the same position. c, Sequence alignment of the aberrant (top) and normal (bottom) LGI1 amplification fragments. A complete copy of intron 3 was retained in the aberrant transcript, introducing a putative stop codon (blue type) at the start of intron 3. The presumptive splice-site mutation is shown in red.

Fig. 5

Analysis of Lgi1 expression in the adult mouse brain by chromogenic RNA in situ hybridization. a–c,f, An antisense riboprobe (a–c) or a sense (control) riboprobe (f) of Lgi1 was hybridized to 16-μm thin coronal cryosections of brains harvested from 10-wk mice (see Methods). Shown are representative sections from the anterior to the posterior extent of the temporal cortex (×40). Bregma values refer to the classical coordinates for vertebrate sections along the anterior–posterior axis. In a, the arrow marks the piriform cortex; in b, the open arrow marks the amygdala and the closed arrows the CA3 region and the dentate gyrus of the hippocampal formation. In c, the arrow marks the location of the mouse auditory cortex. As seen in the low-power images of a–c, Lgi1 was expressed more in ventral cortical structures than dorsal structures. Careful inspection of these sections, however, revealed Lgi1-expressing cells in distinct areas of the dorsal cortex. d,e, Brain maps of the piriform cortex (d) and the dentate gyrus and amygdala (e). g–i, Enlargements of Lgi1 expression in the piriform cortex, dentate gyrus and amygdala, respectively (×200). The relative location of each of these brain regions is identified by rectangular boxes on the brain maps depicted in d and e. Stars in g–i denote molecular areas or fiber tracts that seemed to be generally devoid of Lgi1-expressing cells, whereas areas that were densely packed with neurons demonstrated high Lgi1 expression (arrows in g–i). This restricted expression of Lgi1 was especially apparent in the hippocampal formation (h). Here Lgi1 expression was constrained to the granular cells of the dentate gyrus (closed arrows), to large-bodied cells within the hilus of the dentate gyrus (open arrows) and to the pyramidal cells of the CA3 region (solid arrows) but was absent from the molecular areas (stars in h). The in situ analysis also revealed the presence of distinct mRNA levels in individual cells in both the piriform cortex (g) and amygdala (i). Cells within the basolateral nuclei of the amygdala (BLP, basolateral amygdala posterior; BLA, basolateral amygdala anterior) seemed to express distinctly larger amounts of Lgi1 than did cells within the lateral nuclei of the amygdala (LA in i).