No abstract

The effect of Rev on cytoplasmic accumulation of the singly spliced human immunodeficiency virus type 1 (HIV-1) vif, vpr, and env/vpu RNAs was examined by using a quantitative RNA polymerase chain reaction (PCR) analysis following transfection of complete proviral molecular clones into lymphoid cells. Previously published studies using subgenomic env constructs in nonlymphoid cell types concluded that Rev was necessary for cytoplasmic accumulation of high levels of unspliced env RNA and that, by analogy, Rev must be necessary for the cytoplasmic accumulation of all HIV-1 RNAs that contain the Rev-responsive element (RRE). We confirm those results in COS cells. Unexpectedly, in lymphoid cells, we find that although Rev acts somewhat to increase the cytoplasmic level of full-length HIV-1 RNA, Rev has little or no effect on cytoplasmic accumulation of singly spliced HIV-1 RNAs. However, Env protein expression was greatly reduced in the absence of Rev. Analysis of the cytoplasmic RNA revealed that in the absence of Rev or the RRE, the cytoplasmic vif, vpr, and env/vpu 2 RNAs were not associated with polysomes but with a complex of 40S-80S in size. Consequently, efficient expression of the Vif, Vpr, Vpu, and Env proteins from these RNAs is dependent on Rev. These results exclude a mechanism whereby the sole function of Rev is simply to export RNAs from nucleus to cytoplasm. We discuss other models to take into account the dependence on Rev for efficient translation of cytoplasmic HIV-1 RNAs.

Only a fraction of retroviral primary transcripts are spliced to subgenomic mRNAs; the unspliced transcripts are transported to the cytoplasm for packaging into virions and for translation of the gag and pol genes. We identified cis-acting sequences within the gag gene of Rous sarcoma virus (RSV) which negatively regulate splicing in vivo. Mutations were generated downstream of the splice donor (base 397) in the intron of a proviral clone of RSV. Deletion of bases 708 to 800 or 874 to 987 resulted in a large increase in the level of spliced RSV RNA relative to unspliced RSV RNA. This negative regulator of splicing (nrs) also inhibited splicing of a heterologous splice donor and acceptor pair when inserted into the intron. The nrs element did not affect the level of spliced RNA by increasing the rate of transport of the unspliced RNA to the cytoplasm but interfered more directly with splicing. To investigate the possible role of gag proteins in splicing, we studied constructs carrying frameshift mutations in the gag gene. While these mutations, which caused premature termination of gag translation, did not affect the level of spliced RSV RNA, they resulted in a large decrease in the accumulation of unspliced RNA in the cytoplasm.Retroviral pre-mRNA splicing differs dramatically from that of cellular mRNAs. Nearly all eucaryotic transcripts either are spliced to completion, so that no unspliced RNA is detectable in the cytoplasm, or do not contain introns and so require no splicing (reviewed in references 14 and 27). Retroviruses, on the other hand, produce a single primary transcript which is utilized in both spliced and unspliced forms; unspliced RNAs serve as genomic RNA and also as mRNA for the gag and pol genes, while spliced mRNAs are translated to form env products (32,36). In the case of the avian transforming retrovirus Rous sarcoma virus (RSV), alternative splicing of the primary transcript from a common splice donor also generates the src mRNA (32, 36). Retroviral RNA must be spliced to an intermediate degree rather than to completion, so that the proper balance of unspliced and spliced RNAs reaches the cytoplasm. Since retroviruses must use the same splicing machinery as the eucaryotic cells they infect, the retroviral transcripts need to be partially spliced by a machinery which seems geared to all-or-none splicing.There may be regulatory functions within retroviruses which mediate incomplete splicing (reviewed in references 10 and 32). One possible mechanism is that the retroviral splice donors and acceptors may be recognized inefficiently by splicing factors. Alternatively, cis-acting sequences elsewhere in the viral genome may directly interfere with splicing processes or may facilitate transport of the unspliced precursor RNA away from the splicing machinery, indirectly resulting in a decrease in the level of spliced RNA. A third possibility is that viral proteins may be involved in regulating the level of splicing. A final possibility is that the regulation occurs at the level of RNA stability as...

A cis-acting enhancer element has been detected within the gag gene of several avian retroviruses, including Rous sarcoma virus, Fujinami sarcoma virus, and the endogenous Rous-associated virus-0. A consensus enhancer core sequence, GTGGTTTG, is present in all of these viral genomes, approximately 900 bases downstream from the site of initiation of transcription. When an internal fragment derived from the gag gene of any of these viruses (spanning nucleotides 533 to approximately 1149) was inserted into a plasmid containing the chloramphenicol acetyltransferase (cat) gene under control of the simian virus 40 promoter, 9-or 21-fold enhancement of CAT expression was observed after transfection into mouse L cells and chicken embryo fibroblasts, respectively. This enhancement was not dependent on the position of insertion of the gag fragment into the plasmid. However, there was a strong dependence on orientation, with higher levels of CAT expression in constructs in which the 5' end of the gag fragment was nearest to the promoter, suggesting a possible negative regulatory element at the 3' end of this fragment. Deletion of the 3' end of the insert resulted in a gag fragment, containing nucleotides 533 to 1017, which enhanced expression equally in either orientation. When the gag fragment was inserted into a plasmid containing the cat gene under the control of an intact Rous sarcoma virus long terminal repeat, it induced a two-to threefold increase in CAT activity and CAT mRNA levels.Translation of the gag fragment did not appear to be necessary for the observed enhancement, since two insertional mutations resulting in frameshifts in the gag insert did not affect CAT expression. However, deletion of a 330-base internal fragment from the gag insert restored a basal level of CAT activity. These results suggest that retroviruses have regulatory elements within their genes distinct from those in the long terminal repeats that flank the genes.Previous work on the regulation of retroviral gene expression has focused on the long terminal repeats (LTRs) at both ends of the integrated proviral DNA. The LTRs contain sequences regulating both initiation and termination of transcription (46). Both transcriptional enhancers and promoters have been characterized within the unique 3' (U3) region of retroviral LTRs (6,13,24,25,28,46). In addition to promoting transcription of viral genes, the LTRs of a number of different retroviruses are capable of activating the transcription of contiguous cellular genes, sometimes resulting in neoplasia (16,32). When the strengths of different avian retroviral LTR enhancers are compared, the LTR of the endogenous avian virus Rous-associated virus-0 (RAV-0) is less than 10% as active as the LTR of Rous sarcoma virus (RSV), whereas the LTR of Fujinami sarcoma virus (FSV) has an intermediate level of expression (5, 7, 44, 45, 51).There are several suggestions that sequences within the retroviral genome may also be important in regulating viral gene expression. For example, sequences in the 3' noncod...

Severe combined immunodeficient (SCID) mice inoculated intracerebrally (i.c.) with HIV-infected human monocytes develop brain pathology similar to that in humans with HIV encephalitis. This includes HIV-positive macrophages and multinucleated giant cells, astrogliosis, microglial nodules, and neuronal dropout. These xenografts survive about 1 month. To develop a model of chronic HIV encephalitis and to assay the resulting behavioral abnormalities, we reinoculated SCID mice i.c. every 4 weeks for 3 months with either HIV-infected human monocytes (n = 5) or uninfected human macrophages (n = 4) or administered no inoculation (n = 6); these three groups were monitored for behavioral abnormalities. Tests of cognitive function in a Morris water maze 3.5 months after the first inoculation suggested that HIV-infected mice performed poorly compared with controls. Following testing in the water maze on days 4 and 5 of acquisition, motor activity of infected mice was reduced in comparison with that of controls. Retention of goal location when tested 1 week later was impaired in HIV-infected mice compared with controls. Histopathologic analysis of brains revealed significant astrogliosis and strongly suggested higher numbers of major histocompatibility complex (MHC) class II-positive multinucleated macrophages in HIV-infected compared with control mice. Thus, our preliminary studies indicate that SCID mice with HIV encephalitis develop behavioral abnormalities reminiscent of human disease. These behavioral abnormalities are associated with significantly increased astrogliosis, the presence of HIV, and probably multinucleated giant cells. These studies further support the use of this SCID animal model system for studies of the pathogenesis of HIV encephalitis and for drug interventions.