Extracts from herpes simplex virus-infected cells and from mock-infected cells have been compared for their ability to process at RNA poly(A) sites in vitro. Nuclear extracts from infected cells contain an activity that increases processing efficiency specifically at a late herpes simplex virus poly(A) site. By contrast, a second virus poly(A) site is processed with equal efficiency by nuclear extracts from infected and mock-infected cells. Using precursor RNAs containing these two virus poly(A) sites in tandem, which allows ready detection of the processing factor, we show that this specific activity is heat labile. Analysis of RNAs produced by virus recombinants that contain the poly(A) site sequences in tandem also indicates that increased processing at the late virus poly(A) site occurs in vivo.

Two ϳ135-nucleotide (nt) direct repeats flank the Rous sarcoma virus (RSV) oncogene src and are composed of two smaller repeats, dr1 (ϳ100 nt) and dr2 (ϳ36 nt). These sequences have been reported to contain cis-acting signals necessary for RNA packaging and elements that allow cytoplasmic accumulation of unspliced RNA (cytoplasmic transport elements). In this report, we show that avian fibroblasts infected with the Prague A strain of RSV with precise deletions of both dr1 elements express src and are transformed by this mutant virus but production of virus particles is very low and virus spread throughout the culture requires several weeks. We show that the replication defect is due to complex effects on viral RNA transport, viral RNA half-life, and virus particle assembly. The dr1 elements may contain binding sites for a permissive cell-specific factor(s) that facilitates efficient nuclear-cytoplasmic transport, RNA stability, and cytoplasmic utilization of unspliced viral RNA. The implications of these results for understanding the defects of nonpermissive virus infections in mammalian cells are discussed.

Rous sarcoma virus (RSV) contains two approximately 135-nt imperfect direct repeats composed of smaller repeats, dr1 (approximately 100 nt) and dr2 (approximately 36 nt), that are between the env and src genes and downstream of src in the 3' untranslated region, respectively. It has previously been shown that a Prague A RSV mutant in which both dr1 sequences are deleted is defective at several points in the virus life cycle, including unspliced RNA and env mRNA stability, unspliced RNA transport, and virus particle assembly. A defect in unspliced RNA transport occurs because a cytoplasmic transport element is present within the dr1. We have suggested that the defect of particle production may arise from the failure of the unspliced RNA to be targeted to sites in the cytoplasm where its translation is favorable for Gag protein assembly. In this report, we have further investigated the function of the direct repeats by comparing virus mutants containing either a single upstream or downstream dr1 sequence. Both mutants were delayed in replication compared to the wild-type; the mutant with a single upstream dr1 (delta DDR) is significantly more defective than the mutant with a single downstream dr1 (delta UDR). While both mutants appear capable of efficiently transporting unspliced RNA to the cytoplasm, the delta DDR mutant with only the upstream dr1 is defective in its ability to support Gag assembly and particle release. The replication defect cannot be repaired by placing the upstream dr1 at the location of the downstream dr1 in the 3' untranslated region. A single point mutation in the upstream dr1 (U to C) restored replication and particle production to near normal levels. The results suggest that unspliced RNA transport and Gag assembly functions may be mediated by different elements within the dr1 and that the Prague A upstream dr1 is defective in the latter but not the former function.

A portion of the avian sarcoma virus (ASV) primary RNA transcripts is alternatively spliced in chicken embryo fibroblast cells to two different messages, the src and env mRNAs. Frameshift mutations of the viral genome causing premature translation termination within the src gene result in a decreased steady-state level of the src mRNA. In marked contrast, frameshift mutations at various positions of the env gene do not decrease the level of the env mRNA. We show that the src gene product is not required in trans for splicing and accumulation of src mRNA. Conversely, the truncated Src proteins do not act negatively in trans to decrease specifically the levels of src mRNA. Taken together, these results indicate that the frameshift mutations act in cis to reduce src mRNA levels. A double mutant with a lesion in the src initiator AUG and a frameshift within the src gene demonstrated wild-type RNA levels, indicating that the src mRNA must be recognized as a translatable mRNA for the effect on src mRNA levels to occur. Our results indicate that the reduced levels do not result from decreased cytoplasmic stability of the mature src mRNA. We also show that the src gene frameshift mutations affect src mRNA levels when expressed from intronless src cDNA clones. We conclude that the reduction of src mRNA levels triggered by the presence of frameshift mutations within the src gene occurs while it is associated with the nucleus. Our data also strongly suggest that this occurs at a step of RNA processing or transport independent of RNA splicing.The processing of retrovirus RNA is unusual since only a fraction of the primary transcripts are spliced. For example, the primary transcript of avian sarcoma virus (ASV) is capped, polyadenylated, and methylated similarly to host cell premRNAs (for reviews, see references 23 and 27). However, only about 30% of the RNA is spliced in the nucleus of chicken embryo fibroblasts (CEF), using a common 5' splice site at nucleotide (nt) 398 and two alternative 3' splice sites at nt 5078 and 7054 to form the env and src mRNAs, respectively. The remainder of the unspliced RNA is transported to the cytoplasm, where it is packaged into virions and used as mRNA for translation of the Gag and the Gag-Pol precursor polyproteins. The gag-pol and env mRNAs have the same initiator AUGs because the 5' leader, which contains the gag AUG, is spliced in frame to the env coding sequences. The src mRNA contains the same spliced leader, but the gag AUG initiates a short nine-amino-acid reading frame which is terminated upstream of the src initiator AUG.It has previously been shown that nonsense and frameshift mutations introduced into the gag gene of ASV result in decreased levels of the unspliced viral RNA; the levels of spliced mRNAs are not significantly affected by such gag mutations (2, 3). This decrease in the level of unspliced RNA has been attributed to a cis-acting effect on the stability of the mutant unspliced RNA (3). Accumulated evidence has suggested that the presence of premature translation ter...