Downloaded from genesdev.cshlp.org on October 3, 2021 - Published by Cold Spring Harbor Laboratory Press REVIEW Regulation of HIV and HTLV gene expression Harold Varmus Departments of Microbiology and Immunology and Biochemistry and Biophysics, University of California School of Medicine, San Francisco, California 94143 USA Until recently, simplicity and uniformity of design ap­ derstanding of their functions has been impeded by con­ peared to be universal features of retroviruses. Most re­ flicting results, confusing nomenclatures, possible mul­ troviruses thrive with a few open reading frames in their tiplicities of action, and a dearth of mechanistically in­ genomes, all devoted to major constituents of virus par­ cisive experiments. In an effort to confront some of ticles: core proteins, envelope glycoproteins, and virion- these problems head-on, the Banbury Center at the Cold associated catalytic proteins (reverse transcriptase, inte- Spring Harbor Laboratory recently invited most of the grase, and protease). In the life cycle of the usual retro­ major players, as well as several investigators concerned virus, there is a straightforward division of labors: early with more general issues of transcriptional control or events depend largely upon viral reverse transcriptase retrovirus replication, to hash out differences and dis­ and integrase, whereas in the later stages components cuss new findings. Predictably enough, many important provided mainly by the host are harnessed for produc­ questions remained unresolved, but a semblance of con­ tion of viral proteins and progeny particles (Varmus and sensus also emerged about what could be accepted as Swanstrom 1982, 1985). fact, what names should be adopted for some of the Over the past few years, close study of two types of genes, and what issues should be given experimental pri­ human retroviruses, the T-cell leukemia viruses (HTLV- ority. The following is one observer's view of that con­ I and -II) and the immunodeficiency viruses (HIV-1 and sensus. -2), has uncovered a multiplicity of unexpected viral reg­ ulatory functions, now forcing a revision of this tradi­ tional view of the retrovirus life cycle. The newly dis­ Novel genes in the HIV genome covered functions are fascinating mechanistically and mysterious evolutionarily. Moreover, they offer promise There is now evidence for five or more novel genes in of understanding some perplexing physiological aspects HIV genomes, in addition to those {gag, pol, and env) of infection with major clinical consequences—latency encoding proteins found universally in retroviral par­ and reactivation of HIV and inefficient leukemogenesis ticles. At least three of these, tat, rev, and nef, have been by HTLV. strongly implicated in the regulation of HIV gene ex­ The most prominent regulatory genes in the two types pression and are discussed in detail below; others have of vims {tat and rev of HIV and tax and rex of HTLV) are no identified function as yet {vpr, vpx) or augment virus profoundly different from a biochemical perspective, but infectivity in unknown ways {vif). To the consternation their ultimate effects on gene control are provocatively of outside observers, some of these novel genes have similar: tat and tax augment viral gene expression gen­ been called by as many as a half dozen different names. erally, most clearly by increasing concentrations of viral Shortly after the Banbury Center meeting, an agreement RNA; rev and rex favor production of viral structural was reached on the terms to be used for these genes over regulatory proteins by altering the proportions of (Gallo et al. 1988). The recommendations are summa­ unspliced to fully spliced mRNA. These functions are rized in Table 1 and followed in the text. not confined to human retroviruses: Animal viruses re­ lated to the HTLVs [e.g., bovine leukemia virus (BLV) and simian T-cell leukemia viruses (STLV)] and to the Viral regulators HIVs [e.g., simian immunodeficiency virus (SIV) and tat visna virus] appear to encode proteins similar to those encoded by the human prototypes. The tat gene is a powerful traizs-activator of HIV gene Although the existence and importance of the newly expression at one or more levels of control (Sodroski et discovered retroviral regulators are beyond doubt, an un- al. 1985a; CuUen 1986; Rosen et al. 1986) and is essen­ tial for virus growth (Dayton et al. 1986; Fisher et al. 1986a). Unlike virtually all other retroviral genes, tat The control of HIV gene expression, organized by F. Wong-Staal, R. Franza, and B. Cullen, was held February 28 to March 2, 1988, at the splits its coding sequence into two exons (see Fig. 1), Banbury Center, Cold Spring Harbor Laboratory. A volume, The control which in the most commonly used virus strain encode of HTV expression (eds. B.R. Franza, Jr., B.R. CuUen, and F. Wong-Staal), an 86-amino-acid protein (Tat) found mostly in the nu­ containing about 30 papers written by contributors to the meeting was be published in July, 1988, by Cold Spring Harbor Laboratory and con­ cleus (Arya et al. 1985; Sodroski et al. 1985b; Hauber et tains hill citations for the large body of work summarized here. al. 1987). The protein sequence is rich in cysteines and GENES & DEVELOPMENT 2:1055-1062 © 1988 by Cold Spring Harbor Laboratory ISSN 0890-9369/88 $1.00 1055 Downloaded from genesdev.cshlp.org on October 3, 2021 - Published by Cold Spring Harbor Laboratory Press Vannus Table 1. Regulatory genes in human retroviral genomes metals (Rosen; Wong-Staal), but rather than form the zinc fingers characteristic of certain DNA-binding pro­ Virus group New names Old names teins, the metal binding appears to promote dimeriza- Immunodeficiency tion (Frankel et al. 1988). viruses tat tflt-III, TA Whether assessed for its effects on expression of an rev trs, art intact provirus or, more commonly, on expression of nef 3' orf, orf-B, orf-V, chimeric reporter genes [e.g., HIV LTR-CAT (chloram­ E, orf-1 phenicol transacetylase)], tat may show induction ratios vif sor, orf-A, orf-Q, P', as high as 1000 or more. The magnitude of the effect orf-l varies widely, however, and is especially dependent on vpr R vpx X choice of host cells. (Mouse fibroblasts allow much smaller inductions than do human T cells, and the T-cell leukemia human colon carcinoma line, SW480, may have an en­ X, x^ tat-l, tat-ll, viruses tax dogenous tat'likc activity that precludes a response to x-lor viral tat.) Results are also likely to be strongly in­ rex X™ fluenced by the concentration of an exogenous respon­ New names as proposed by Gallo et al. (1988). [tat] Trans-sicti- sive gene, the experimental strategy (e.g., transient vator; [rev] regulator of expression of virion proteins; {nef) nega­ versus stable transfections), and perhaps by other less tive factor; [vif] virion infectivity factor; [vpr] virion protein R; tangible factors. (vpx) virion protein X; [tax] trans-activator from X; (rex) regu­ lator of expression from x. These differences, often invoked in the Banbury dis­ cussions, may account for some radically different views of tat function. For example, effects upon the concentra­ basic residues, but most of the latter are dispensible for tion of reporter protein often exceed the increase in cog­ activity, whereas most of the cysteines appear, from re­ nate RNA, suggesting that there might be more than one cent site-directed mutagenesis experiments, to be cru­ mechanism by which Tat works; but in many other ex­ cial for function (F. Wong-Staal, National Institutes of periments (e.g., Rice and Mathews 1988), the concentra­ Health; C. Rosen, Roche Institute; Sadaie et al. 1988). tions of reporter RNA and protein rise in equal propor­ The cysteines are likely to mediate binding to heavy tion. Discordance between tat induction of RNA and HUMAN T-LEUKEMIA HUMAN IMMUNODEFICIENCY VIRUSES VIRUSES tatb tatgrevb gag pol vip vpr reva envnef tax Transcription Nucleus Transcription Nucleus I \ -^ RNA """""" *• RNA Splicing + Splicing + X transport • transport T j^^>ju/^^ Cytoplasm -l-_ ^ -JT + ^gag pol -• gag-pol tat (p14) -• vip tax p40 rex p27 oi Structural rev (p19) X proteins Structural proteins Maturation t nef (p27) Maturation Yoshida, 1988. Wong-Staal, 1988. Figure 1. The organization and expression of regulatory genes of human retroviruses. The diagrams (provided by M. Yoshida and F. Wong-Staal) illustrate many of the aspects of retroviral gene expression discussed in the text. Transcription of HTLV and HIV pro- viruses produces genome-length precursors, which are differentially spliced to generate mRNAs for structural proteins (encoded by gag, pol, env, and vif [labeled vip in this figure]) and nonstructural, regulatory proteins (encoded by tax, rex, tat, rev, and nef). Strong positive and negative effects upon steps in the life cycle are indicated by -\- and -, weak negative effects are indicated by (-). 1056 GENES & DEVELOPMENT Downloaded from genesdev.cshlp.org on October 3, 2021 - Published by Cold Spring Harbor Laboratory Press HIV and HTLV gene expression protein suggests that tat may affect other processes [e.g., out that the observed RNA species is predicted to have a RNA trafficking or translation efficiency (W. Haseltine, high degree of secondary structure and might be the ex­ Harvard)], as well as accumulation of RNA. pected residuum of degradation. In addition, an antiter- One way to approach the mechanism (or mechanisms) mination mechanism predicts that deletion of the termi­ by which Tat works is to define the phenomena more nation site will yield constitutive 'up' mutants, yet no carefully than has been done. It was disappointing to such mutants have been isolated. Thus, proposals that learn, for example, that no serious effort has been made Tat affects initiation, RNA processing, or RNA stability to follow the metabolism of RNA from a t^t-regulated still have currency—and may be testable with a Tat-re- gene, thereby defining its association with subcellular sponsive in vitro transcription system (Okamoto and structures, including polyribosomes.