Guo and Kemphues reported on the first successful use of in vitro synthesised antisense RNA to phenocopy a maternal-effect embryonic mutant phenotype, that of the
par-1 gene (Cell, 81: 611-620, 1995). More recently, a number of other labs have demonstrated the general usefulness of this technique in generating phenocopies of other maternal-effect mutations. Here, it is reported that 1) Antisense RNA can phenocopy zygotic mutations, which have effects at different developmental stages 2) Antisense RNA can phenocopy predicted loss-of-function phenotypes of C. briggsae homologues of C. elegans genes 3) The severity of phenocopies produced by antisense RNA is influenced by position and 4) Antisense RNA directed against an intron sequence fails to produce a mutant phenocopy. These experiments were performed using the sex-determining gene
tra-2, which promotes hermaphrodite development in XX animals; XX
tra-2(lf) mutants are transformed into pseudomales. Klass et al. (Dev. Biol. 69: 329-335,1976) have shown that wild-type
tra-2 activity is required throughout larval development, because animals carrying the
tra-2(
b202ts) allele develop as intersexes with variably masculinised somatic gonad, tail, and germ line depending on the timing of temperature shifts. No maternal-effects are associated with
tra-2 mutants. The
tra-2 antisense RNA studies were initiated to show that the
Cb-tra-2 homologue promotes hermaphrodite sex determination in C. briggsae. This approach was adopted because the sequences of Cb-TRA-2A and Ce-TRA-2A are extremely diverged (only 43% identical) and it was not possible to show that transgenic
Cb-tra-2 has cross-species feminising activity in C. elegans. The progeny of C. briggsae hermaphrodites injected with antisense RNA (~1 kb) corresponding to a 3' region of the 4.7 kb
Cb-tra-2 mRNA showed masculinisation of the somatic gonad, vulva, and tail (truncated spike), including the absence of a vulva. Those animals injected with 5' antisense RNA (~1 kb) also produced masculinised progeny, however the degree of masculinisation was more extensive. For example, some animals injected with 5'
tra-2 antisense RNA developed male tails with a fan and ray, whereas those injected with 3'
tra-2 antisense showed only truncation of the tail spike. Studies similar to those described above were also performed using
Ce-tra-2 to gain more information about the general utility and properties of antisense RNA, because
Ce-tra-2 is a large genetically well characterised gene whose wild-type activity is required throughout larval development and in a variety of tissues. As summarised in the figure below, antisense RNA was synthesised against four non-overlapping regions (labeled A-D) of the 4.7 kb (and 1.8 kb)
tra-2 mRNA and the progeny of injected mothers were examined. Although this study is complicated by the presence of at least 2
tra-2 mRNAs (4.7 and 1.8 kb), it was previously shown that the 4.7 kb
tra-2 mRNA encodes the primary somatic feminising activity of the
tra-2 locus. As shown, there is a strong correlation between the 5' - 3' location of the sequences chosen for generating antisense RNA and the extent of somatic masculinisation. Region A antisense RNA injections produced pseudomales, similar in phenotype to XX
tra-2(lf) mutants. Region B and C antisense RNA produced intersex animals, and Region D antisense RNA produced animals with a female soma and male germ line (Mog phenotype). Finally, given the rapid progress of the Sequencing Consortium in generating genomic C. elegans gene sequences, it was of interest to determine whether antisense RNA directed against intron sequences could also produce a mutant phenocopy. In the case of
Ce-tra-2, antisense RNA (~600 nt) directed against a
tra-2 intron sequence failed to produce somatically masculinised progeny in 9/9 injected animals. This result suggests that antisense RNA is likely to disrupt wild-type gene activity at the post-transcriptional level. To conclude, these results indicate that antisense RNA can phenocopy not only maternal-effect mutations, but also zygotic mutations that may have their effects at different stages of larval development. It is surprising that
tra-2 antisense RNA shows such perdurance. However, as the antisense RNAs used in these experiments were synthesised without capping, which is a modification of the Guo and Kemphues protocol, it is possible that uncapped RNAs might escape degradation - a hypothesis that remains to be tested. These experiments also suggest that it is preferable to synthesise antisense RNA corresponding to either a 5' region of an mRNA or to use exon, but not intron sequences when using genomic templates. Thanks to Tim Schedl and Allen Jones for sharing their unpublished antisense RNA results.