Splice site skipping in polyomavirus late pre-mRNA processing

Polyomavirus late nuclear primary transcripts contain tandem repeats of the late strand of the viral genome, as a result of inefficient transcription termination and polyadenylation. Pre-mRNA processing involves the splicing of short noncoding late leader exons to each other (removing genome-length introns) and the splicing of the last leader to a coding body exon (such as for the major virion structural protein, VP1). As a result, cytoplasmic mRNAs contain 1 to 12 tandem leader exons at their 5' ends that are followed by a single coding exon. To understand more about how polyomavirus exons are spliced together, we studied a double-genome construct consisting of two tandem but nonidentical polyomavirus late transcription units. The alternating leader exons are distinguishable from one another but retain identical flanking RNA-processing signals, as for the alternating VP1 exons. We transfected this construct and derivatives of it into mouse cells and determined which leader exons are spliced to which others and which VP1 exons are utilized. Results showed that leader exons are almost never skipped during splicing and are spliced sequentially to one another. On the other hand, VP1 exons were often skipped, with the VP1 exon closest to the polyadenylation site splicing to the nearest upstream leader exon. Splice site replacement experiments showed that VP1 exon skipping is not due to a relative weakness of its 3' splice site or to any sequence upstream of the VP1 3' splice site. Exon skipping is also not the result of sequences within the VP1 exon. Rather, VP1 3' splice site skipping can be eliminated by replacing the inefficient late polyadenylation signal with an efficient one, or by inserting a 5' splice site between the VP1 3' splice site and the late polyadenylation site. Thus, sequences that compose the distal border of the VP1 exon can influence usage of the upstream 3' splice site.