Retroviruses are defined by their ability to reverse the normal flow of genetic information from genomic DNA to mRNA (19). Although not unique (pararetroviruses such as hepatitis B virus also share this property), retroviruses nevertheless form a clearly defined viral genus. Distinguishing characteristics include the morphology and composition of retroviral virions and the possession of a single-stranded, positive-sense RNA genome. While retroviruses form a relatively homogeneous viral family, they have customarily been subdivided into three taxonomic groupings primarily on the basis of the in vivo and in vitro consequences of infection (16, 19). The oncovirus subgroup includes retroviruses able to cause neoplastic disease in the infected host animal but also includes some related, relatively benign viruses. The lentivirus subgroup includes retroviruses that cause slow, chronic diseases that generally, but not always, lack a neoplastic component. Members of the spumavirus sub-group cause a foamy cytopathic effect in tissue culture and have yet to be clearly associated with any disease. Retroviruses were the first oncogenic viruses to be identified, and it is this ability to transform cells which first attracted scientific attention. Interest in the mechanisms of retroviral transformation led to considerable research into the life cycle of animal retroviruses belonging to the onco-virus subgroup, focusing particularly on the prototypic mu-rine and avian leukemia viruses (MLV and ALV)(19). Retroviral replication is initiated by the intracytoplasmic penetration of the virion core, a process mediated by the specific interaction of the viral envelope glycoprotein with a host cell surface receptor. Subsequently, the virion-associ-ated reverse transcriptase transcribes the single-stranded viral RNA genome into a double-stranded linear DNA proviral intermediate. This proviral intermediate then mi-grates to the nucleus where the viral integrase enzyme acts to covalently link the retroviral genome to the host chromo-somal DNA, thereby forming the retroviral provirus. In its simplest form, as seen for example in MLV, retro-viral replication requires only three distinct virus-encoded genes (19)(Fig. 1). These are the gag gene, which encodes the virion structural proteins, the pol gene, which encodes the variousvirion-associated enzymes, and env, which en-codes the envelope glycoprotein.(A viral protease, required for the posttranslational processing of the Gag and Gag-Pol polyproteins, may be encoded in pol or may form part of gag.) In the integrated DNA provirus, these three genes are invariably arranged in the same order (5'-gag-pol-env-3') and are flanked by the characteristic long terminal repeats (LTRs) generated during the process of reverse transcrip-tion. The LTRs contain enhancer and promoter elements required for efficient transcription ofthe retroviral genome and also contain sequences important for efficient mRNA polyadenylation within the 3'LTR. In the case of MLV and the majority of other animal oncoviruses, the integrated provirus encodes only two distinct transcripts. These are the genomic RNA, which also functions as the mRNA for Gag and Pol synthesis, and a singly-spliced mRNA species that encodes Env. It therefore appeared that the life cycle of MLV, and by extension of retroviruses in general, was both simple and efficient. Viral gene products served structural or enzymatic functions, while regulation of viral gene expres-sion at both the transcriptional and posttranscriptional levels was controlled by the interplay of cis-acting viral DNA or RNA sequences with trans-acting factors encoded entirely by the host cell.

The discovery of directly pathogenic human …