Inhibitory amino-acid receptors, including those for GABA and glutamate, play an important role in the functioning of the nematode nervous system, involved in such activities as locomotion, defaecation and pharyngeal pumping (McIntire et. al, 1993; Avery et. al, 1995; Laughton et. al, 1995). PCR amplifications on C. elegans cDNA using oligonucleotide primers matching conserved regions found in invertebrate GABA-A receptors resulted in the isolation of a number of different partial cDNAs with clear homology to GABA-A receptor subunits. 5' and 3' RACE reactions were used to amplify a full-length cDNA from one of these sequences, named Cegbr 2 (accession no.U40573). This cDNA encodes a protein of 416 amino acids which has all the characteristics of an inhibitory amino acid receptor subunit, including a disulphide-bonded loop in the extracellular N-terminal domain and four hydrophobic, predicted transmembrane regions towards the C-terminus. The spliced leader sequence, SL1, is found at the N-terminus immediately upstream of the start methionine. Interestingly, the 3' untranslated region of this cDNA contains sequence very similar to that encoding the 4 transmembrane regions. During subsequent PCR reactions we isolated an additional cDNA, named Cegbr 3 (ac! cession no. U41113) which encodes a receptor with an identical extracellular domain to Cegbr 2 linked directly to the transmembrane sequence found in the 3'UTR of Cegbr 2 (fig 1). To verify that both transcripts exist in mRNA we performed Northerns using polyA+ RNA isolated from a developmentally mixed population of C. elegans. Radiolabelled cDNA probes specific to the unique transmembrane III-IV intracellular loop sequences of Cegbr 2 or Cegbr 3 were hybridised to duplicate filters and washed under high stringency. As expected, the Cegbr2-specific probe bound to a single transcript of ~2.2Kb (Cegbr 2) while the Cegbr 3 probe bound to two transcripts of ~2.2Kb (Cegbr 2) and ~1.2Kb (Cegbr 3). From these results we predicted that Cegbr 2 and Cegbr 3 are alternatively spliced products of a single gene. To investigate this further we performed genomic PCR across the extracellular/transmembrane boundaries of both subunits. As predicted, both boundaries corresponded with intron splice sites (fig 2). As far as we are aware, this pattern of splicing is unique to nematode fast ion channel receptors, and suggests that these animals will possess receptors with identical ligand binding sites but slightly different channel properties. A similar genomic organisation has been found for a related gene,
unc-49, during the C. elegans genome sequencing project. Here, a common N-terminus may be spliced onto three alternative transmembrane domains. This work is supported by the Wellcome Trust.