Worm Breeder's Gazette 12(2): 57 (January 1, 1992)
These abstracts should not be cited in bibliographies. Material contained herein should be treated as personal communication and should be cited as such only with the consent of the author.
We have sequenced the point mutations associated with the let-23 ( n1045 )and the dpy-10 ( e128 )alleles. In both cases the "G" at the 3' end of one intron (i.e., the splice acceptor site) has been altered to an "A." The dinucleotide AG is found at the 3' end of all C. elegans and non-IG eukaryotic pre-mRNA introns.
[See Figure 1]
Based on results from in vitro and in vivo splicing studies in other eukaryotic systems in which the acceptor AG site has been mutated, we expected any of three consequences of the n1045 and e128 mutations: (1) the mutated intron is not spliced out; (2) use of the splice acceptor of the next downstream intron (exon skipping); or (3) splicing at an AG not normally used as a splice acceptor site (i.e., a cryptic AG). In either case when splicing occurs (i.e., (2) and (3)), the acceptor site chosen is still an AG. Our results, described below, suggest that splicing in C. elegans may occur somewhat differently than other eukaryotes because we find splicing at non-AG acceptor sites.
The let-23 gene encodes a member of the EGF receptor tyrosine kinase family and is necessary for differentiation of the hermaphrodite vulva. The dpy-10 gene encodes a cuticle collagen that, when altered, can cause Dpy, DLR, or LR phenotypes. Neither the let-23 ( n1045 )allele nor the dpy-10 ( e128 )allele is null.
The allele let-23 ( n1045 )alters the splice acceptor of intron 16 (53 nts long) from AG to AA (there are a total of 18 exons and 17 introns in let-23 )whereas the allele dpy-10 ( e128 )alters the splice acceptor of intron 2 (48 nts long) from AG to AA (there are a total of 4 exons and 3 introns in dpy-10 ).To understand the effect of these mutation on the products produced, RNA from homozygous n1045 and e128 hermaphrodites were isolated and reverse transcribed (RT).
For let-23 ( n1045 ),PCR amplification of the RT RNA was performed using primers that amplify from exon 16 to exon 18. The PCR amplification of n1045 RT RNA resulted in four bands (shown below). The parallel experiment with RNA isolated from wild-type hermaphrodites resulted in only one band.
[See Figure 2]
The wild type-sized band in the n1045 RNA was unexpected. Sequencing eleven subclones from this band showed the presence of six different, but similarly sized, transcripts, only one of which uses an AG acceptor site (shown below). The other transcripts use AA, AT, TG, and GG acceptor sites, virtually unprecedented in the splicing literature. These unusual acceptor sequences, however, all share either the penultimate A or the ultimate G of the wild-type AG acceptor site. About half of the splices were made at the same location as the wild-type acceptor site, even though it was now an AA. These results are not likely to be artifactual. First, parallel RT-PCR analysis of wild-type RNA results in only wild-type RNA product, verified by sequencing of four subclones. Second, both RNAse protection and quantitative PCR analysis suggest that the levels of let-23 transcript in wild type and n1045 are similar. Therefore, the phenomenon is not a minor effect. Third, the presence of the bands is RT-dependent and DNAse-independent. Fourth, splicing of n1045 RNA at other introns occurs normally.
[See Figure 3]
For dpy-10 ( e128 ),PCR amplification of the RT RNA was performed using primers that amplify from exon 1 to exon 3. RT-PCR revealed two splicing products on an ethidium stained gel: one that appeared unspliced and one that appeared normally spliced (relative intensity of about 2:1 respectively). Thirteen cDNA clones of the apparently normally sized product were sequenced (see below). Two were normally spliced despite the presence of the mutant acceptor; however the other 11 were spliced near, but not at, the correct site. As in the case of n1045 ,splicing can occur at non-AG splice acceptors. If unspliced, the 16aa encoded by the intron would be present in the dpy-10 collagen in e128 animals. The level of dpy-10 message was assayed in e128 worms by competitive PCR and found to be virtually identical to wild-type levels (<2 fold difference).
[See Figure 4]
Taken together, these results suggest that the splicing machinery in C. elegans may use a different mechanism to recognize intro-exon boundaries than that used in mammalian and yeast systems. Two other observations that support this notion are: (1) about 3/4 of C. elegans introns are shorter than the 80 nt minimum required for introns to be spliced in mammalian cells; (2) incorrect splicing of C. elegans introns at the 3' site when spliced in mammalian cells even after correcting for intron length requirement (Kay et al., NAR 15: 3723). One possible mechanism could derive intron-exon border information from AT vs. GC content (suggested by T. Blumenthal).