A late region deletion mutant of simian virus 40 (dl5) was previously shown to be deficient in the transport of nuclear RNA. This is a splice junction deletion that has lost the 3' end of an RNA leader, an intervening sequence, and the 5' end of the splice acceptor site on the body of the mRNA. In this report, we analyzed the steady-state structure of the untransported nuclear RNA. The 5' ends of this RNA are heterogeneous but contain a prominent 5' end at the normal position (nucleotide 325) in addition to several other prominent 5' ends not seen in wildtype RNA. The 3' end of this RNA does not occur at the usual position (nucleotide 2674) of polyadenylation; instead, this RNA is non-polyadenylated, with the 3' end occurring either downstream or upstream of the normal position.Eucaryotic mRNA production occurs via a surprisingly complicated pathway. The sequence of events leading to stable mRNA includes promotion (which in itself may be multifunctional), 5' capping, 3' polyadenylation, splicing, and finally transport from the nucleus to the cytoplasm (for review, see references 9 and 14). Of these post-transcriptional events, the least investigated is the question of how nuclear transcripts are transported into the cytoplasm. Not all nuclear RNA sequences are destined to become cytoplasmic mRNA. It was reported that the sequence complexity of heterogeneous nuclear RNA is measurably greater than that of cytoplasmic RNA (31). It could be argued, however, that this may be because transcription runs well past the polyadenylation site and that subsequent cleavage and polyadenylation could result in the excision and degradation of downstream sequences, thereby accounting for the different complexities observed.It seems unlikely, however, that such an explanation could account for the transport behavior of several constructed and naturally occurring deletion mutations in viral and cellular genes (7,8,12,16,(18)(19)(20)(21)(22)30; R. T. White, Ph.D. thesis, Stanford University, Stanford, Calif., 1980). There are now numerous examples of genes deleted downstream from promoters and upstream from polyadenylation sites that do not code for the production of stable mRNA. These deletions generally affect splice junctions. An early example of a transport deficiency mutation was constructed in the late region of simian virus 40 (SV40), in which an intervening sequence was precisely deleted (18, 19). The relation between splicing and transport, however, is complicated. There are, for example, deletion mutants which efficiently transport unspliced RNA (8,15,30; White, thesis), indicating that splicing per se need not occur to allow transport. Along these lines, a hypothesis that the intervening sequences were directly involved in the retention within the nucleus of nuclear RNA was proposed and tested (28). The results of those experiments indicated that retention of nuclear RNA is probably not simply the result of the presence of intervening sequences, as transcripts containing intronic sequences were efficiently transported...