De Novo Exon Duplication in a New Allele of Mouse Glra1 (Spasmodic)

Abstract

The novel neurological mutant Cincinatti arose by genomic duplication of exon 5 in the glycine receptor gene Glra1. The mutant transcript results in premature protein truncation. A direct repeat of the pentamer GGGGC is present adjacent to the breakpoints and may have mediated the duplication event by a replication slippage mechanism.

EXON duplication has been a significant source of novel gene function during evolution, with many well-characterized examples in invertebrate and vertebrate genes. In the sodium channel gene family, for example, alternative copies of exon 5 and exon 18 arose by exon duplication and are conserved in several mammalian genes (Plummer and Meisler 1999). A genomewide in silico search for adjacent exons with sequence similarity concluded that 10% of mammalian genes contain duplicated exons (Letunic et al. 2002). Traces of the molecular mechanisms involved in ancient exon duplications have been erased from modern genomic sequences. We describe a novel example of gene duplication resulting in a null allele of the Glra1 gene. Because of the recent origin of this mutation on the homozygous, sequenced genomic background of inbred strain C57BL/6J, insight into the molecular mechanism can be obtained from the sequence of the mutated chromosome.

Cincinnati is a spontaneous neurological mutation that arose in the mouse colony of Keith Cox at the University of Cincinnati. The mutant was recognized by the generation of multiple affected offspring from one mating cage. At 2 weeks of age, affected mice exhibit impaired righting reflex and tremor, which progress with loss of neurological function and death by 3 weeks of age.

The mutation arose on a C57BL/6J line congenic for a knockout allele of the LCAD gene on chromosome 1, which proved to be unlinked to the mutation. For genetic mapping, an obligate heterozygote was crossed to strain C3H/HeNCr1BR and the F₁ offspring were randomly intercrossed. Among the litters containing affected offspring, the proportion of affected mice was 79/329, consistent with the expected 25% for a single, recessively inherited autosomal locus (P > 0.67). To map the locus, 15 affected F₂ mice were genotyped with a whole-genome SNP panel of 394 markers (Moran et al. 2006). The data localized the mutation to a 13.8-Mb nonrecombinant region on chromosome 11 between the flanking SNP markers 11.044.080 at 44.1 Mb and 11.057.906 at 57.9 Mb (mouse genome sequence build 33 (http://www.genome.ucsc.edu). The haplotype of the mutation-carrying chromosome was consistent with origin on a C57BL/6J chromosome. The nonrecombinant interval was reduced to 1.9 Mb between rs13481034 at 54.9 Mb and D11Mit164 at 56.6 Mb by genotyping 92 affected and unaffected F₂ mice for additional markers on chromosome 11. There are 31 annotated genes in the nonrecombinant region, including the nine-exon gene Glra1 encoding the glycine channel α-subunit.

Glycine-gated channels are the major ligand-gated channel in the spinal cord and brain and respond to release of the inhibitory neurotransmitter glycine (Lynch 2004). Two mutant alleles have been previously characterized in the mouse. The null mutant oscillator closely resembles the mutant described here (Buckwalter et al. 1994). The spasmodic mouse carries a missense allele with partial loss of function that results in a viable neurological disorder with hyperekplexia (exaggerated startle) and hypertonia (Ryan et al. 1994). Twenty different mutations of human GLRA1 have been identified in families with inherited hyperekplexia (Bakker et al. 2006).

To evaluate Glra1 as a candidate gene, we carried out RT–PCR on brain RNA. Amplification of the cDNA with a forward primer in exon 4 (CCT ACA ATG AAT ACC CTG ATG ACT) and a reverse primer in exon 8 (ACA GCC ATC CAA ATG TCA ATA GCT) produced the predicted 646-bp product from wild-type RNA, but an abnormal product of 729 bp from mutant RNA. The abnormal cDNA fragment contains two tandem copies of the 83-bp exon 5 (Figure 1A). A stop codon at the junction between the two copies of exon 5 results in the nonsense mutation F187X. The predicted protein lacks the C-terminal 262 residues of the 449-residue protein and is likely to be a null allele, consistent with the phenotypic similarity to oscillator, which lacks the C-terminal half of the protein.

Figure 1.—

Molecular characterization of the Glra1 mutation. (A) Partial sequence of the abnormal transcript. The transcript was amplified by RT–PCR with a forward primer in exon 4 and a reverse primer in exon 8. Exon 5 is represented twice in the sequence, generating a stop codon (F187X) at the exon 5/exon 5′ junction. The cDNA sequence is numbered from the ATG initiation codon (NM_020492). (B) Genomic structure of the mutant chromosome containing two copies of exon 5. The 537-bp duplicon begins with nucleotide −279 of the 2-kb intron 4 and ends with nucleotide 175 of intron 5 and includes 279 bp from intron 4, 83 bp from exon 5, and 175 bp from intron 5. (C) Genomic sequence around the duplication breakpoint in intron 4. The sequence of the breakpoint in the mutant chromosome is aligned with intron 4 (above) and intron 5 (below). A single T nucleotide was inserted at the breakpoint. Intron 4 is numbered from the 5′ end of exon 5 (−279) and intron 5 is numbered from the 3′ end of exon 5 (+175) (mouse sequence build 33).

To identify the genomic mutation underlying the abnormal transcript, we carried out PCR of genomic DNA using two complementary primers within exon 5, the forward primer GTA CAG ACG TGT ATC ATG CAA CTC and the reverse primer GAG TTG CAT GAT ACA CGT CTG TAC (Figure 1B, arrows). These primers are not expected to amplify wild-type DNA, and no wild-type product was obtained. Amplification of mutant DNA with these primers generated a 561-bp product containing the 537-bp duplicated genomic fragment composed of the final 279 bp of intron 4, the intact exon 5 (83 bp), and the first 175 bp of intron 5 (Figure 1B). The new allele carrying this duplication is designated Glra1^dup5.

Analysis of the genomic sequence of intron 4 and intron 5 did not identify any SINE or LINE elements or other shared repetitive sequences that would be expected if the mutation were caused by unequal recombination due to misaligment of the two introns. There is, however, a perfectly aligned direct repeat of the pentamer GGGGC located 4 bp from the breakpoints in intron 4 and intron 5 (boxed in Figure 1C). In a study of 80 genomic duplications, Chen et al. (2005) observed short repeats of 2–11 bp at the site of 47 duplications. Since these are unlikely to be long enough to mediate unequal recombination, it was proposed that a replication slippage mechanism could explain some genomic duplications. The structure of the Glra1 breakpoint is consistent with this model.

Inclusion of two copies of exon 5 in the Glra1 transcript results in a frameshift mutation and predicted loss of function. Similar exon duplications have been identified recently as human disease mutations in Duchenne's muscular dystrophy (White et al. 2006) and familial polyposis (McCart et al. 2006). In the context of evolutionary time, a pair of duplicated exons can be converted to mutually exclusive alternative exons by accumulation of point mutations in splice sites, intronic splice enhancers, or exonic splice enhancers, thereby correcting the loss of function caused by the in-frame translation of the duplicated exons (Kondrashov and Koonin 2001). Later divergence of the originally identical alternative exons would then lead to acquisition of novel function (Letunic et al. 2002; Xing and Lee 2006). The duplicated copies of exon 5 and exon 18 in the sodium channel genes have diverged with respect to developmental regulation and amino acid sequence, resulting in alternative expression of fetal and adult forms of exon 5, as well as a tissue-specific form of exon 18 encoding a truncated protein (Plummer and Meisler 1999; V. L. Drews and M. H. Meisler, unpublished data). A pair of alternative exons in the human glycine receptor GLRA2 similarly have undergone amino acid sequence divergence (Letunic et al. 2002). The Glra1 mutation described here supports the model that alignment of very short direct repeats in adjacent introns can mediate exon duplication, a process that has contributed significantly to the evolution of functional complexity in the human genome.