In C. elegans the pre-mRNAs of some genes receive the spliced leader SL1 while others receive SL2 .SL1 is trans-spliced onto conventional monocistronic pre-mRNAs containing an "outron" (Conrad et al.,1991, MCB 11:1921). SL2 ,however, appears to be specialized for trans-splicing onto the pre-mRNAs of downstream genes in polycistronic transcription units (i.e. operons) (see WBG 12:3, p. 72). We've been using the
gpd-21 /gpd-3 "operon" to study SL2 -specifictrans-splicing. Ralph Hecht kindly provided us with genomic sequence surrounding
gpd-2 &3and pointed out the presence of an open reading frame upstream of
gpd-2 (Lee et al., 1992, Gene: In press). We have now shown by RT-PCR and the isolation of cDNA clones that this ORF represents a real gene. It has been named
mai-1 (for mitochondrial ATPase inhibitor) because of its homology to a mammalian gene which specifies a nuclear-encoded subunit of the mitochondrial ATP synthetase (Interestingly, the worm version doesn't appear to begin with a mitochondrial import peptide, as the mammalian gene does.) The structure of
mai-1 ,shown in the top line of the figure, has been determined from genomic and cDNA sequences. Introns and exons are indicated in the traditional fashion with the coding portions of exons shaded. The intercistronic spaces between
mai-1 and
gpd-2 ,and between
gpd-2 and
gpd-3 are represented by narrow solid bars. We selected six cDNA clones using a fragment of genomic clone containing the second intron and third exon of
mai-1 .The structures of the 5 types of clone are diagrammed below (clone
cm1 was isolated twice.) Clone
cm6 may be full length based on Northern analysis. The 3' ends of
cm1 & 6 are just downstream of the unusual cleavage and polyadenylation signal, AGUMA. These two clones place the 3' end of
mai-1 only about 100 bp upstream of
gpd-2 ,suggesting that
mai-1 could also be part of the
gpd-21 gpd-3 operon. Additional evidence for this possibility is provided by clones
cm2 & 7 which both contain part of
mai-1 and all of
gpd-2 on the same clone. In both clones, introns were properly spliced, and
cm7 at least is polyadenylated at the normal site at the end of
gpd-2 .However, the entire intercistronic space between
mai-1 and
gpd-2 remains intact in both clones, so the introns were removed but trans-splicing of
gpd-2 didn't occur. [See Figure] PCR experiments suggest that the abundance of clones from polycistronic RNAs in this cDNA library (provided by Bob Barstead) represents the relative abundance in the total RNA population. We attribute this abundance of an apparent RNA processing intermediate to the unusual
mai-1 polyA signal (AGUAM) which may be an inefficient cleavage signal. We hypothesize that these cDNA clones representing the polycistronic pre-mRNA accumulate both because cleavage is slow and because trans-splicing of the resultant polycistronic pre-mRNA is a naturally slow event. Perhaps the most interesting clone is
cm5 .Its 3' end is one base upstream of the trans-splice site for
gpd-2 ,suggesting that the process of trans-splicing separated the polycistronic RNA, not the cleavage that normally precedes polyA addition. Furthermore, this clone is polyadenylated even though there is no MUAAA just upstream (the closest is the AGAUUU at the 3' end of
mai-1 ,127 nt upstream). Our results support the idea that
mai-1 ,
gpd-2 and
gpd-3 constitute a 3 gene "operon". Clones
cm2 ,5 & 7 offer strong evidence for co-transcription of
mai-1 and
gpd-2 .Our model then predicts that
gpd-2 should receive SL2 because it is a downstream gene in a polycistronic transcription unit. Although
gpd-2 has previously been reported to be an SL1 -acceptor,we have performed two types of experiments (RTPCR with either SL1 or SL2 as upstream primers and primer extension following RNaseH cleavage with SL1 -or SL2 -specificoligos) indicating that some
gpd-2 pre-mRNAs receive SL1 while others receive SL2 .This suggests the possibility that transcription of
gpd-2 /gpd-3 might originate from two promoters. One would be upstream of
mai-1 such that its transcript would include
mai-1 ,
gpd-2 ,and
gpd-3 .In this case
gpd-2 would be expected to receive SL2 because it is a downstream gene. Clones
cm2 ,5, and 7 indicate this configuration of
mai-1 ,
gpd-2 transcription exists. A polycistronic RNA containing all three has not been detected, presumably because of the normal, efficient, cleavage signal at the end of
gpd-2 .The other promoter may lie within
mai-1 ,such that
gpd-2 is at the 5' end of the transcript and thus it receives SL1 because it begins with an outron. Deletions of
mai-1 in which only a part of the last intron and the last exon remain show good expression of
gpd-2 and
gpd-3 ,providing evidence for this transcript. Clearly what we thought was a simple system to study SL2 -specifictrans-splicing (of
gpd-3 )has become more complex because the "operon" actually is composed of three genes and multiple promoters.