Regulation BRYAN

of HIV-1

gene

expression

R. CULLEN

Howard Hughes Medical Carolina27710, USA

Institute

and Department

of Microbiology

ABSTRACT The quantity and quality of HIV-1 gene expression is temporally controlled by a cascade of sequential regulatory interactions. Basal HIV-1 transcription is determined by interaction of cellular regulatory proteins with specific DNA target sequences within the HIV-1 long-terminal repeat. The most notable of these protein:DNA interactions involves NF-xB, a transcription factor that plays a pivotal role in the activation of genes important for cellular responses to infection and inflammation. A second level of control involves the virally encoded Tat trans-activator. Tat, in combination with as yet unidentified cellular proteins, activates HIV-1 gene expression through a specific interaction with the viral TAR RNA stem-loop target sequence. A final level of regulation is mediated by the viral Rev protein. Rev acts posttranscriptionally to induce the expression of HIV-1 structural proteins and thereby commits HIV-1 to the late, cytopathic phase of the viral replication cycle. Rev activity appears to require a critical, threshold level of Rev protein expression, thus preventing entry into this late phase in cells that are unable to support efficient HIV-1 gene expression. In total, this cascade of regulatory levels allows the HIV-1 provirus to respond appropriately to the intracellular milieu present in each infected cell. In activated cells, the combination of Tat and Rev can stimulate a very high level of viral gene expression and replication. In quiescent or resting cells, in contrast, these same regulatory proteins are predicted to maintain the HIV-1 provirus in a latent or nonproductive state. Cullen, B. R. Regulation of HIV-1 gene expression. FASEB J. 5: 2361-2368; 1991. K#{128}y Words:

protein

HIV-1

gene expression

‘NF-xB

Tai

Rev

RNA:

binding

ARE DEFINED BY their ability to reverse transcribe a single-stranded RNA genome into a doublestranded DNA intermediate that is then integrated into the host cell genome as a provirus. In its simplest form, as seen for example with the avian (ALV)1 and murine (MLV) leukemia viruses, the viral proteins required for the retrovirallifecycle are encoded by only three separate genes (reviewed in ref 1).These are the gag gene, which encodes structural proteins that constitute the virion; the pol gene, which encodes the various enzymatic activities essential to the process of reverse transcription and proviral integration; RETROVIRUSES

and

last,

the

env gene,

which

encodes

glycoproteins

present

on the virion surface that confer the ability to bind to and fuse with the appropriate target cells. Complex retroviruses differ from simple retroviruses, such as ALV, in that the ga pol, and env gene products, although still essential, are no longer sufficient for viral replication. Human immunodeficiency virus type 1 (HIV-1), which may be viewed as the prototype complex retrovirus, encodes at least six such auxiliary gene products. These novel retroviralproteins are

and Immunology,

Duke University Medical

Center, Durham,

North

now known to play critical roles in both regulation of HIV-1 gene expression and in morphogenesis and release of infectious HIV-1 virions. Expression of these additional HIV-l gene products both requires and facilitates a more complex genomic organization in HIV-1 than is observed in simple retroviruses such as ALV (Fig. 1). The genome of ALV contains a single splice donor and acceptor combination and encodes only two viral mRNA species. In contrast, the HIV-1 genome contains at least four known splice donor and six known splice acceptor sequences (Fig. 1). More than 30 distinct processed HIV-1 mRNA species have now been characterized and these include examples in which almost every viral gene is positioned as the first open reading frame (2). Exceptions are the viral pol gene, which is translated as the result of a regulated ribosomal frame-shift event to give the Gag-Pol polyprotein (3), and the viral eno gene, which appears to be translated from a bicistronic mRNA that also encodes the vpu gene product (2). The various HIV-1 mRNA species can be divided into three distinct size classes. These are unspliced, genomic RNA, which also functions as the mRNA for gag and pol, a singly spliced, -4-kb class that encodes the viral env, v/ vpr, and vpu gene products, and a doubly spliced, -2-kb class that encodes the viral Tat, Rev, and Nef proteins. As discussed in detail below, expression of these different viral mRNA species is differentially regulated in vivo by action of the viral Rev trans-activator.

THE ROLE FACIORS

OF

CELLULAR

TRANSCRIPTION

The integrated HIV-1 provirus, like all other retroviral proviruses, is flanked by long-terminal repeats (LTR5) generated during the process of reverse transcription (Fig. 1). The 5’ LTR functions to promote proviral transcription whereas

the

3’ LTR

is required

for

efficient

polyadenylation

of the resultant transcripts.The recognition sequences for several constitutively expressed or inducible host cell transcription factors have been identified within the HIV-1 LTR promoter element (Fig. 2). Of particular interest are the two binding sites specific for NF-xB, an inducible transcription factor that plays a central role in activation of many genes involved in the cellular response to infection or injury (4). These genes include cytokines such as 3-interferon, GMCSF, IL 2, TNF-a, and IL 6, immunologically relevant cellsurface receptors such as the IL 2 receptor a chain and class I MHC antigens, and even acute-phase proteins such as serum amyloid A.

‘Abbreviations: ALV, avian leukemia virus; HIV-1, human immunodeficiency virus type 1; LTR, long-terminal repeat; HTLV, human T cell leukemia virus; TAR, Tat response element; RRE, Rev response element.

0892-6638/91/0005-2361/$01.50. © FASEB 2361 w.fasebj.org by Univ of So Dakota Lommen Hlth Sci Library (192.236.36.29) on September 14, 2018. The FASEB Journal Vol. ${article.issue.getVolume()}, No. ${article.issue.getIssueNum

D

I

LTR

ALV

ENV

POL LTR

02D3

D4

The constitutively expressed cellular transcription factors Spi and TFIID play an important role in mediating promoter function in many cellular and viral genes. The HIV-1 LTR possesses three functional Spl binding sites as well as a typical TATA element (8, 9) (Fig. 2). These c#{252}acting sequences appear to constitute the basal HIV-1 promoter element. Both Spi and TFIID binding interactions are critical for HIV-1 LTR promoter function and deletion of these binding domains results in a defective HIV-1 provirus

HIV-

Al

Figure 1. Comparison The figure contrasts

A6

of the genetic complexity the genomic organization

of ALV and HIV-1. of a representative in this case avian leukemia virus, with that of are the location of known viral genes as well as

simple retrovirus, HIV-1. Indicated known splice donor (D) and acceptor (A) sites. Also shown are the location of the RNA target sites for the HIV-1 Tat and Rev transactivators. LTR, long-terminal repeat; TAR, Tat response element; RRE, Rev response element (reproduced by permission from Cullen and Greene, Virology, Vol. 178, pp. 1-5, 1990).

In most cells, NF-xB is expressed in an inactive cytoplasmic form that consistsof three distinctsubunits of -50 kDa, -65 kDa, and -37 kDa (reviewed in ref 5). The 37-kDa protein,

termed

IxB,

serves

to anchor

this

protein

complex

in the cytoplasm in an inactive form. Activation of NF-xB appears to involve phosphorylation of the IxB subunit. This is followed by release of the 50-kDa and 65-kDa polypeptides from IxB and their subsequent nuclear migration. The 50-kDa NF-xB subunit, which is believed to encode the

DNA binding specificityof NF-xB, has recently been shown to belong to a family of eukaryotic regulatory proteins that also includes the rd oncogene and the Drosophiladorsal gene product (5). In T cells, activation of NF-xB can result from a wide range of stimuli,These include T cellmitogens such as lectins,phorbol esters,and bacteriallipopolysaccharide,several viral trans-activators(see below), and cytokines such as TNF-a and IL 1. Recently, it has been suggested that intracellular thiols andlor reactive oxidative intermediates may play a central role in mediating activation of NF-xB by agents such as phorbol estersand TNF-a (6).Reactive oxidative intermediates are observed at relativelyhigh levelsat sites of tissue inflammation, and IL 1 and TNF-a (which are effectiveinducers of NF-xB activity)are also highly effective inducers of the inflammatory response in vivo.Activation of NF-xB-being an entirely posttranslational event-may therefore serve as one of the earliestcellularresponses to inflammation. HIV-1 has apparently effectively subverted this defensive host response by using activation of NF-xB to enhance retroviral gene expression, and hence replication, at sites where inflammation occurs (6). A second nuclear factor induced by activation of T cells, termed NFAT-1, also interacts with a specific binding site within the HIV-1 LTR (Fig. 2). Although it is tempting to speculate that NFAT-1 may also play a role in activation of HIV-1 gene expression, littleevidence supporting the functional importance of NFAT-1 currently exists(7). Similarly, the importance of a binding sitefor the inducible transcription factor AP-1, located toward the 5’ end of the LTR U3 region, remains to be established (7).

2362

Vol. 5

July 1991

(8).

Two additional constitutiveHIV-1 LTR-specific DNA:protein binding interactions involving cellular transcription factors USF and LBP have also been defined (Fig. 2). Both of these protein:DNA binding events are readily detectable in vitro by using gel retardation and DNAse footprinting analysis, and in vivo by using DNase hypersensitivity analysis (9, 10). Although the functional significance of these interactions remains unclear, some data suggest that the HIV-1 LTR USF site may function as a weak negative regulatory element (7). Mutational analyses have so far failed to demonstrate any direct role for the tripartite LBP interaction in regulating HIV-1-specific gene expression in vivo (10). In addition to transcription factors encoded by the host cell itself, superinfecting viruses may also encode factors able to enhance HIV-1-specific gene expression in trans. Viral gene products encoded by the human T cell leukemia virus (HTLV), several members of the herpes virus family, the adenoviruses, and hepatitus-B virus have all been shown to enhance HIV-1 LTR-dependent gene expression in transfected cells(11, 12). Because of the somewhat narrow cell tropism of HIV-1, it has remained unclear whether these observed activation events have any relevance to the spread of HIV-1 in vivo. Two viruses known to infect human T cell populations, and hence strong candidates as cofactors in the pathogenesis of HIV-1 induced disease, are HTLV and the human herpes virus 6 (HHV-6). The HTLV Tax transactivator has been shown to induce NF-xB activity in HTLV-1 infected T cells (12) whereas herpes virus early gene products are believed to activate HIV-1 LTR-specific gene expression via both NF-xB-dependent and -independent pathways (11).Dual infection of cultured CD4 human Tcells with HHV-6 and HIV-1 results in enhanced replication of HIV-1 and in accelerated cell death (13).

THE

HIV-1

TAT

TRANS-ACTIVATOR

The cellular transcription factors described previously sufficient to induce a relatively low basal level of HIV-1

-455

+1

Host DNA

+98

+182

(TAR) [R

1)3

4

USF

are gene

Provirsi DNA

I

1)5

LBP

AP.t

Nrk8

WIlD

2. The HIV-1 LTR promoter element. Scale representation of the unique 3 (U3), repeat (R), and unique 5 (U5) subregions of the HIV-1 proviral LTR. Basepair coordinates are given relative to the transcription start site which is designated +1. DNA binding sites for known cell transcription factors are indicated by boxes, with inducible factors shown by hatching (adapted with permission from Cullen and Maim, Nucleic Acids and Molecular Biology, Vol. 4, pp. 176-184, 1990). Figure

The FASEB Journal

CU LIEN

w.fasebj.org by Univ of So Dakota Lommen Hlth Sci Library (192.236.36.29) on September 14, 2018. The FASEB Journal Vol. ${article.issue.getVolume()}, No. ${article.issue.getIssueNum

in TAR serve primarily a structural role in the appropriate presentation of essential primary sequence information located in the six nucleotide terminal loop and the three nucleotide bulge of TAR (20). Evidence that Tat possesses the ability to specifically bind the TAR RNA stemioop has been provided by several groups (21-23). However, attempts to correlate the effect of TAR mutations on in vivo function with their effect on Tat binding in vitro have met with mixed success. Although mutations that disrupted the TAR stem structure of that affected the three nucleotide bulge were found to reduce both in vivo function and in vitro Tat binding, equally deleterious mutations in the terminal loop of TAR were found to have little or no effect on Tat binding (20, 21). These observations have been explained by demonstrating that the interaction between Tat and TAR ocis flanked at its 3’ end by a conserved translation termination curs entirely at the site of the three nucleotide bulge (22, 23) signal, is sufficient to encode a fully active Tat protein (15). of Tat appears to be both necesAt leasttwo distinctfunctional domains have been identified (Fig. 3). The basic domain sary and sufficient for binding to TAR (23), thus placing Tat in Tat. A highly conserved motif containing seven cysteine into the arginine-rich class of sequence-specific RNA bindresidues has been proposed to bind metal ions and may ing proteins that also includes the HIV-1 Rev trans-activator. mediate protein:protein interactions in vivo (16). A COOHDespite general agreement that Tat shows sequenceterminal domain rich in lysine and arginine residues is respecific binding to TAR in vitro, several lines of evidence quired for the nuclear and nucleolar localization of Tat (17) and also serves as a sequence specific RNA binding domain suggest that this direct interaction constitutes only one of the components involved in the in vivo interaction of Tat with (see below). TAR. The most significant of these is the finding, noted The TAR element is a 59 nucleotide RNA stem-loop previously, that mutations to the terminal loop of TAR that structure located at the 5’ end of all HIV-1 transcripts (18, 19) strongly inhibit in vivo function have little or no effect on in (Fig. 3). Both the location and orientation of TAR are critivitro Tat binding. It therefore appears probable that a cellucal for function (18, 19). Extensive mutational analyses suplar cofactor must be involved in mediating the TatTAR inport the hypothesis that the double-stranded RNA segments teraction in vivo. The best current candidate for this role is a nuclear protein of -68 kDa that has been shown to bind LoopBinding specifically to the terminal loop sequence of TAR in vitro of the terminal loop that affect in vivo TAR 68 kD Protein (24). Mutations function were observed to resultin an appropriate reduction in the in vitro binding affinity of p68 for TAR. In addition, partially purified pfiS protein was observed to enhance the in A-U BulgeBinding vitro trans-activation of HIV-l LTR specific transcription by Tat (25).It istherefore hypothesized that Tat trans-activation Tat Protein G-C of the HIV-1 LTR requires the cooperative binding of both A-U Tat and a 68-kDa cellular protein to two different, adjacent G-C loop structures present near the apex of the TAR RNA eleA-U ment (Fig. 3). C-G Recently, it has been suggested that trans-activation of the HIV-1 LTR by Tat requires activation of cellular protein A C-G kinase C (26). This activation, which can result from a numMG-C ber of mitogenic stimuli, does not lead to phosphorylation of A-U Tat itself, thus suggesting that one or more cellular factors reU-A quired for the Tat response may be present in an inactive U-A form in resting cells. These observations therefore imply that G-C mitogenic stimulation of infected cells may lead to an inducG-U tion of HIV-1 LTR-specific gene expression via mechanisms U-A involving both cellular DNA sequence-specific transcription C-G factors such as NF-xB and as yet undefined cellular proteins involved in mediating trans-activation by Tat. U-G C-G Many studies have shown that Tat increases the steadystate level of transcripts derived from genes linked to the HIV-1 LTR (15, 27). There has also been general agreement that this increase results from an enhancement in the rate of G-C transcription of such genes rather than from specific stabiliG-C zation of mRNAs containing TAR (17). However, the m7GpppG C mechanism by which Tat enhances the rate of transcription has been more controversial. One possibility is that Tat could Figure 3. The Tat/TAR interaction. The sequence and structure of act to increase the rate of transcription initiation (17), thus the 59 nucleotide TAR RNA element are shown together with the making TAR the RNA equivalent of a DNA enhancer seputative binding sites of Tat and of a cellular 68-kDa protein. This quence. Alternatively, Tat could function to prevent premacooperative interaction is hypothesized to be essentia for in vivo Tat function (reproduced with permission from Cullen, Cell, Vol. 63, ture termination of transcripts initiated within the HIV-1 pp. 655-657, 1990). LTR (28). This latter hypothesis was originally derived from

expression in infected cells. These initial transcripts reach the infected cell cytoplasm exclusively in the form of the fully spliced, 2-kb class of viral mRNAs that encode the viral regulatory proteins Tat, Rev, and Nef (see below). The role of the viral Tat protein is to trans-activate HIV-1 LTRdirected gene expression, thereby establishing a powerful positive feedback loop that can lead to very high levels of HIV-1-specific RNA and protein synthesis. A functional copy of the viral tat gene is essential for HIV-1 replication in vitro as is an intact copy of the cit-acting viral tat response (TAR) element (14). The tat gene is divided into two coding exons, which together predict the synthesis of an 86 amino acid protein. However, the 72 amino acid first coding exon of Tat, which

present

HIV-1 EXPRESSION 2363 w.fasebj.org by UnivGENE of So Dakota LommenREGULATION Hlth Sci Library (192.236.36.29) on September 14, 2018. The FASEB Journal Vol. ${article.issue.getVolume()}, No. ${article.issue.getIssueNum

the observation that Tat had little effect on the level of RNA polymerase density adjacent to the site of transcription initiation but dramatically increased the rate of transcription of sequences distal to the HIV-1 LTR (28). It was also reported that the basal HIV-1 LTR promoter gave rise to a high level of a prematurely terminated, -59 nucleotide TAR RNA that was reduced upon coexpression of the Tat trans-activator (19, 28). It was therefore proposed that Tat activated the HIV-1 LTR by relieving a specific block to transcription elongation through the TAR sequence. Unfortunately, the obvious prediction of this hypothesis, i.e., that deletion of TAR should enhance the level of HIV-l LTR-dependent gene expression, proved not to be valid (15, 18, 19). However, the observation that Tat promotes elongation has now been confirmed both in vivo and in vitro (25, 39, 30). These more recent results indicate that the transcription termination observed in the absence of Tat occurs at multiple, possibly random locations in viral or heterologous sequences linked to the HIV-1 LTR, thus suggesting that Tat acts by increasing processivity rather than by preventing a specific termination event (29, 30) (Fig. 4). Accumulation of short TAR-specific RNAs in the absence of Tat is now thought to result from the resistance of these structured RNAs to the action of a 3’ exonuclease that is believed to degrade the heterogeneous, prematurely terminated RNA species that are transcribed from the basal HIV-l LTR promoter (19, 30). The resultsdiscussed previously suggest that transcription complexes initiating in the HIV-1 LTR are only poorly able to elongate through adjacent DNA templates. This lack of processivity is rectified by interaction of the Tat protein with the nascent TAR element, and potentially with the transcription complex itself (Fig. 4). Neither the reason for the high incidence of premature termination nor the mechanism by which rescue of these transcription complexes occurs is currently understood. However, it appears likely that this poor processivity must be encoded within the HIV-1 LTR promoter element itself. It remains unclear whether the

A)

RSV LTR

B)

EHVLTR TAR

C)

HIVLTR TAR

Figure 4. Schematic representation of HIV-1 Tat function. A) Strong constitutive promoters, such as the LTR of Rous sarcoma virus, efficiently assemble transcription complexes capable of both effective transcription initiation and elongation. B) The basal HIV-1 LTR promoter, in contrast, appears to assemble initiation complexes efficiently yet these complexes terminate spontaneously before completing transcription of the provirus. C) The HIV-1 Tat trans-activator is hypothesized to interact with the nascent TAR RNA stem-loop, and via an unknown mechanism, modify initiated transcription complexes to a termination-resistant form, thus activating viral gene expression.

HIV-1 LTR simply lacks the ability to assemble a complete transcription complex or whether one or more factors that interact with LTR promoter sequences specify inefficient elongation. However Tat, in combination with TAR, can at least modestly trans-activate transcription from such standard promoters as the LTR of Rous sarcoma virus (15, 18). It therefore seems possible that the poor processivity observed during basal HIV-1 LTR-driven transcription may simply be an extreme example of a relatively general phenomenon (Fig. 4). Although results from several groups support the hypothesis that transcriptional trans-activation by Tat results primarily from enhanced transcription elongation, evidence has also been presented suggesting that in at least some experimental settings, Tat can also act to increase the level of transcription initiation (29). Therefore it appears that the increased processivity of transcription complexes formed in the presence of Tat might be correlated with an increased ability to assemble a functional transcription complex at the HIV-1 LTR promoter, i.e., to enhance HIV-1 LTR transcription initiation via an interaction with the TAR RNA element. However, the question of whether Tat affects mRNA synthesis at both the level of initiation and elongation, and whether these two effects are functionally interlinked, remains to be fully resolved. The recent demonstration of an in vitro transcription system that appears to faithfully reproduce the TAR-dependent trans-activation of the HIV-1 LTR by Tat appears likely to represent a key step toward eventually unraveling the mechanism of action of this novel regulatory protein (25). Whereas activation of viral RNA transcription is the major action of Tat in most experimental systems, it is clearly not the only effect of this small trans-activator. It has been noted in a number of reports that the effect of Tat on the level of expression of genes linked to the viral LTR, when measured at the protein level, can be significantly higher than the effect determined at the level of steady-state mRNA (14, 15, 27, 31).Although the molecular basis for thissecond, posttranscriptional component of the bimodal action of Tat remains unclear, it also appears to be mediated by the sequence-specific interaction of Tat with the viral TAR RNA element (32). Results obtained using microinjection of preformed TAR containing RNA molecules into Xenopus 00cytes show that this posttranscriptional effect also occurs in the cell nucleus, yet can be segregated from the transcriptional action of Tat. Thus far, Tat has not been shown to modulate the nuclear export of TAR-containing RNA species and thus a more complex mechanism of action appears likely (17, 32). One hypothesis is that Tat could affect the cytoplasmic compartmentalization, and hence the translational utilization, of TAR-containing transcripts (17, 32).

TRANSITION EXPRESSION

TO

VIRAL

STRUCTURAL

GENE

The mRNA species encoded by HIV-1 can be divided into two classes based on their temporal expression during the HIV-1 replication cycle. The early class of viral mRNAs consists of the multiply spliced, -2-kb mRNA species that encode the viral regulatory proteins Tat, Nef, and Rev. The late class of viral mRNAs consists of the unspliced (-9 kb) and singly spliced (-4 kb) transcripts that encode the virion structural proteins. In the absence of functional Rev protein, only the fully spliced class of HIV-1 mRNAs is expressed (31, 33-35) (Fig. 5). In fact, Rev mutants of HIV-1 are incapable of inducing synthesisof the viralstructuralproteins,and are

CULLEN The FASEB Journal 2364 Vol. 5 july 1991 w.fasebj.org by Univ of So Dakota Lommen Hlth Sci Library (192.236.36.29) on September 14, 2018. The FASEB Journal Vol. ${article.issue.getVolume()}, No. ${article.issue.getIssueNum

1

3

2

kb

-9.1

kb

kb

agents that result in activation of the HIV-1 LTR also induces expression of the viral structural proteins (37). It is therefore hypothesized that latency in this context is due to expression of a subcritical level of the Rev trans-activator, a level that in turn reflects a lack of cellular transcription factors critical for efficient HIV-1 LTR-dependent gene expression (37). The primary role of the Rev regulatory pathway may therefore be to prevent premature progression of the viral replication cycle to the late or lytic phase in cells which are incapable of supporting a sufficient level of viral mRNA and protein synthesis. Although Rev is absolutely required for the cytoplasmic expression of unspliced HIV-1 RNA species, it appears to have little effect on the pattern of HIV-1 RNA expression in the cell nucleus (33, 34). In particular, high levels of Unspliced viral transcripts can be detected in the nucleus even in the absence of Rev (Fig. 6). While it appears that the splice sites present in HIV-1, like those present in other retroviral transcripts (38), are inefficiently utilized by the cell splicing machinery (34, 39), it has been less clear why these incompletely spliced viral transcripts remain sequestered in the cell nucleus in the absence of Rev. One hypothesis suggests that the gag; pal, and env genes might contain multiple cit-acting repressive sequences which function to retain these RNAs in the nucleus in the absence of Rev (40). An alternative hypothesis argues that it is, in fact, the intact splice sites present in these incompletely spliced mRNAs that act as nuclear retention signals (34, 39). It has been suggested that splicing factors may be able to assemble on the primary HIV-1 transcript but are then only poorly able to carry out the actual splicing step. Instead, this interaction results in the retention of incompletely spliced viral transcripts within the nucleus (34, 39). The Rev protein is believed to function by activating the nuclear export of these sequestered viral RNA species either by antagonizing their interaction with these splicing factors (39) or by directly facilitating their in-

IwcL1

Figure 5. Cytoplasmic HIV-1 mRNA expressionpatternsin the presence and absence of Rev. This Northern analysis demonstrates that

the

mRNA sence

singly-spliced,

-4.3

kb

species are not expressed of Rev (Rev).

This

and

unspliced,

-9.1

in the cell cytoplasm

Rev negative

state

occurs

kb

HIV-1

Vpu Env

VII

--------cii

=

-lit.

in the ab-

naturally

early

ThAN3G

in the HIV-1 replication cycle but can also be reproduced, as here, by mutation of the Rev gene in an otherwise intact provirus.

therefore replication defective (31, 35). An analysis of the time course of HIV-1 infection of human T lymphocytes reveals a similar phenomenon (36). Initially, only the 2-kb class of viral mRNAs is detected in the cytoplasm of HIV-l infected cells; however as the level of viral gene expression increases (due to the action of the Tat protein), a switch to synthesis of the viral structural gene mRNAs is observed. This effect, which reflects the action of the viral Rev transactivator, occurs concomitantly with an essentially equivalent reduction in the synthesis of the fully spliced mRNA species that encode the viral regulatory proteins (31, 34) (Fig. 5). Therefore, Rev functions as a negative regulator of its own synthesis. The switch from the early, regulatory phase of HIV-1 gene expression to the late, structural phase appears to require expression of a critical level of the Rev protein (37). Several cell lines nonproductively infected by HIV-1 have been shown to constitutively express a low level of viral mRNA that is primarily of the 2-kb class. Treatment of these cells with

a

E

oIcI

Gag. Pol

Env. Vif. Vpu. Vpr, Tat (p14)

2115 GLASS

Tat (p 16). Rev. Nsf

Figure 6. Schematic representation of HIV-1 Rev function. The nuclear and cytoplasmic pattern of HIV-1 mRNA expression observed in the absence of Rev (i.e., early gene expression) and in the presence of Rev (i.e., late gene expression) is indicated. Tat (p16) represents the full length, two-exon form of Tat while Tat (p14) represents the truncated, one-exon form of Tat. See text for detailed discussion.

HIV-1 GENE EXPRESSION REGULATION 2365 w.fasebj.org by Univ of So Dakota Lommen Hlth Sci Library (192.236.36.29) on September 14, 2018. The FASEB Journal Vol. ${article.issue.getVolume()}, No. ${article.issue.getIssueNum

teraction with a component of a cellular RNA transport pathway (33, 34). In contrast, in the absence of Rev, viral mRNAs are eventually fully spliced before their transport to the cytoplasm (Fig. 6). The action of the Rev trans-activator is specific for unspliced HIV-1 transcripts and appears not to affect the splicing or transport of any cellular RNAs. The specificity of this response is conferred by a highly structured 234-nt RNA target sequence, the Rev response element or RRE, that is located within the envelope gene of HIV-1 (34, 40) (Fig. 1, Fig. 7). The Rev protein has been shown to bind to the RRE with high affinity in vitro (41-43). Although the entire 234-nt RRE is required for full biological activity in vivo, it is now clear that a 66-nt stem-loop subdomain of the RRE is both necessary and sufficient for high-affinity binding of Rev and also sufficient for partial biological activity in vivo (41) (Fig. 7). It is hypothesized that the remainder of the RRE functions to stabilize the RNA structure of the Rev binding site and/or to facilitate presentation of this RNA sequence in vivo (41). The rev gene consists of two coding exons that together predict a protein of 116 amino acids. Rev is localized to the nuclei, and particularly the nucleoli of expressing cells (33). Rev is phosphorylated at two serine residues in vivo; however this posttranslational modification does not appear relevant to Rev function (44). Two distinct protein domains that are essential for Rev function have been defined (Fig. 8). The first is an -40 amino acid NH2-terminal sequence characterized by an arginine-rich central core that has been shown to function as the Rev protein nuclear/nucleolar localization signal (44). This sequence also mediates the sequence-

Cu

A

8O

A .CGA

Primary

Rev

Binding SIte

I0

/

230 structure

7. The Rev/RRE of the

234

interaction. The sequence and predicted

nucleotide

RRE

RNA

target

sequence

shown together with the approximate location of the primary ing site of the viral Rev protein (modified with permission Maim et al., Cell, Vol. 60, pp. 675-683, 1990).

2366

100

116

I

I

I

-NL-.+RNA

BINDING MULTIMERIZATION

ACTIVATION DOMAIN

Figure 8. Domain structure of the HIV-1 rev protein. Functional domains present within the 116 amino acid Rev protein are identified (see text for details). The leucine-rich activation domain (amino acids 75-83) plays no role in RRE binding or Rev ultimerization but is essential for in vivo function. The arginine-rich motif (amino acids 35-50) appears both necessary and sufficient for Rev nuctear/nucleolar localization and specific binding to the RRE. Sequences flanking the arginine-rich motif between approximately amino acid 18 and 56 are essential for multimerization of Rev on the RRE target sequence. The Rev protein sequences located between approximately amino acid 56 and 75 may serve primarily to space these flanking functional domains appropriately, as this region is highly tolerant of missense mutations but not of de!etion mutations. The NH2- and COOH-terminal Rev sequences indicated by stippling are dispensable for in vivo Rev function (Reproduced with permission from Malim and Cullen, Cel4 Vol. 65, pp. 241-248, 1991).

specific interaction of Rev with the RRE (42, 45). Flanking the arginine-rich core are sequences that facilitate the multimerization of Rev, a process that appears important for in vivo Rev function (42, 45). Mutations within any part of this sequence element result in Rev proteins displaying a recessive negative phenotype (42, 44). It seems probable that Rev also contains a protein sequence element that interacts directly with a component of the nuclear RNA transport or splicing machinery (44). Mutational analysis has suggested that a leucine-rich domain centered on amino acid 80 may serve this function (44). Rev proteins mutated in this latter domain retain full RRE binding and multimerization activity, and yet are not only defective but also inhibit wild-type Rev function in trans (42, 44). This trans-dominant inhibition could result from competition between wild-type and mutant Rev proteins for binding to the viral RRE, whereupon the bound, mutant Rev proteins are unable to interact with cellular factors involved in RNA transport from the nucleus. Alternatively, these mutant Rev proteins could inhibit wild-type Rev by forming inactive mixed multimers. In either case, definition of the in vivo role of this leucine-rich motif, and identification of the cellular factor (or factors) that interact with this domain in vivo, appear critical to the full resolution of the mechanism of action of Rev.

AUXILIARY

PROTEINS

OF

HIV-1

0

AG GAGC U UAC UCCUOGACA Figure

/

OTHER

\u

I

/

I

80

Vol. 5

July 1991

are

bindfrom

In addition to Tat and Rev, HIV-1 encodes four other auxiliary proteins named Nef, Vpr, Vif, and Vpu. Nef, the third early gene product of HIV-1, is a myristylated phosphoprotein that is associated with cytoplasmic membrane structures (46). Nef has been reported to possess the GTPase, autophosphorylation, and GTP-binding properties typicalof the G protein family of signal transduction proteins, but this observation has not been confirmed. Unlike Tat and Rev, the

The FASEB Journal

CULLEN

w.fasebj.org by Univ of So Dakota Lommen Hlth Sci Library (192.236.36.29) on September 14, 2018. The FASEB Journal Vol. ${article.issue.getVolume()}, No. ${article.issue.getIssueNum

Nef gene product is not required for HIV-1 replication in culture. It has, in fact, been proposed that expression of Nef results in inhibition of HIV-1 LTR-specific gene expression and viral replication (47, 48). However, these negative effects of Nef remain controversial as others have observed no effect of the Nef protein on either viral replication or gene expression (46, 49). The role of the nef gene product in the HIV-1 replication cycle therefore remains unclear. However, the fact that the Nef open reading frame is reasonably conserved in all primate lentiviruses suggests that this protein is likely to play a significant role in the viral life cycle in the infected host. Recent data in fact suggest that a functional nef gene product can markedly enhance viral replication and pathogenicity in rhesus macaques infected with a cloned isolate of SIV (H. Kestler and R. Desrosiers, personal communication). The HIV-1 Vif, Vpu, and Vpr gene products are cytoplasmic proteins that are expressed late in the HIV-1 replication cycle in a Rev-dependent manner (reviewed in ref 50). As predicted by their temporal expression and subcellular localization, the role played by these proteins is primarily structural rather than regulatory. Vif and Vpu have been shown to function in the morphogenesis and release of infectious HIV-1 virions while the Vpr protein appears to be virion-associated. All three of these proteins significantly enhance replication of HIV-1 in culture, although the effect is relatively modest in the case of Vpu and Vpr.

PERSPECTIVE The importance of HIV-1 as a human pathogen has led to an extraordinary effort to understand the regulatory mechanisms that govern HIV-1 gene expression in the infected cell. Whereas these efforts were launched primarily in the hope of revealing novel avenues for chemotherapeutic intervention into HIV-1 induced disease, they have also led to an extraordinary series of discoveries in the more general area of eukaryotic gene regulation. Among the most notable of these is an increased understanding of the mechanism of action and significance of NF-xB, a cellulartranscription factor that appears to play a central role in mediating host response to inflammation and injury. Of perhaps even more interest are the HIV-1 Tat and Rev trans-activators, which have been shown to utilize entirely novel regulatory mechanisms mediated by sequence-specific interactions with discrete viral RNA target sequences. Despite these advances, much remains to be accomplished. Of particular importance, both in terms of understanding HIV-1 gene regulation in particular and in defining the significance of these novel regulatory pathways in general, is identification of the cellular proteins that mediate the Tat and Rev response. Overall, it appears likely that the field of HIV-1 gene regulation will continue to lead to insights that are of broad significance in the area of eukaryotic molecular biology.

REFERENCES 1. Varmus, H., and Brown, P. (1989) Retroviruses. In Mobile DNA (Berg, D. E., and Hose, M. M., eds) pp. 53-108, American Society for Microbiology, Washington, D.C. 2. Schwartz, S., Felber, B. K., Fenyo, E. -M., and Pavlakis, G. N. (1990) Env and Vpu proteins of human immunodeficiency virus type 1 are produced from multiple bicistronic mRNAs. j Virol. 64, 5448-5456 3. Jacks, T., Power, M. D., Rasiarz, F. R., Luciw, P. A., Barr, HIV-1 GENE EXPRESSION REGULATION

P. J., and Varmus, H. (1988) Characterization frameshifting in HIV-l gag-pol expression. Nature 280-283 4. Libermann, T. A., and Baltimore, D. (1990) interleukin-6 gene expression through the NF-xB factor. Mol. CelL Biol. 10, 2327-2334 5. Gilmore, T. D. (1990) NF-xB, KBF1, dorsal, and ters. Cell 62, 841-843 6. Staal, F. J. T., Roederer, M., Herzenberg, L. A.,

of ribosomal (London) 231, Activation transcription related

of

mat-

and Herzenberg, L. A. (1990) Intracellular thiols regulate activation of nuclear factor xB and transcription of human immunodeficiency virus. Proc. Nati. Acad. Sci. USA 87, 9943-9947 7. Lu, Y., Touzjian, N., Stenzel, M., Dorfman, T., Sodroski, J. G., and Haseltine, W. A. (1990) Identification of cis-acting repressive sequences within the negative regulatory element of human immunodeficiency virus type 1. j Viral. 64, 5226-5229 8. Jones, K. A., Kadonaga, J. T., Luciw, P. A., and Tjian, R. (1986) Activation of the AIDS retrovirus promoter by the cellular transcription factor, Spl. Science 232, 755-759 9. Garcia, J. A., Wu, F. K., Mitsuyasu, R., and Gaynor, R. B. (1987) Interactions of cellular proteins involved in the transcriptional regulation of the human immunodeficiency virus. EMBO j 6, 3761-3770 10. Maim, M. H., Fenrick, R., Ballard, D. W., Hauber, J., Bohn1cm, E., and Cullen, B. R. (1989) Functional characterization of a complex protein:DNA binding domain located within the HIV-1 LTR leader region. j Virol. 63, 3213-3219 11. Nabel, G. J., Rice, S. A., Knipe, D. M., and Baltimore, D. (1988) Alternative mechanisms for activation of human immunodeficiency virus enhancer in T cells. Science 239, 1299-1302 12. Siekevitz, M., Josephs, S. F., Dukovich, M., Peffer, N., WongStaal, F., and Greene, W. C. (1987) Activation of the HIV-l LTR by T cell mitogens and the trans-activator protein of

HTLV-1. Science 238, 1575-1578 13. Lusso, P., Ensoli, B., Markham, P. D., Ablashi, D. V., Salahuddin, S. Z., Wong-Staal, F., and Gallo, R. C. (1989) Productive dual

infection

of human

CD4

T lymphocytes

by HIV-1

and

HHV-6. Nature (London) 337, 370-373 14. Rosen, C. A., Sodroski, J. G., Goh, W. C., Dayton, A. I., Lippke, J., and Haseltine, W. A. (1986) Post-transcriptional regulation accounts for the trans-activation of the human T-lymphotropic virus type III. Nature (London) 319, 555-559 15. Cullen, B. munodeficiency 973-982

R.

(1986) virus occurs

Trans-activation via a bimodal

of human immechanism. Cell 46,

16. Frankel, A. D., Bredt, D. S., and Pabo, C. 0. (1988) Tat protein from human immunodeficiency virus forms a metal-linked dimer. Science 240, 70-73 17. Hauber,J., Perkins, A., Heimer, E. P., and Cullen, B. R. (1987) Trans-activation of human immunodeficiency virus gene expression is mediated by nuclear events. Proc. Natl. Acad. Sci. USA 84, 6364-6368 18. Rosen, C. A., Sodroski,

location

of cit-acting

lymphotropic

repeat. 19. Selby, (1989) determine

virus

J.

G., and

regulatory type

III

Haseltine,

sequences

W. A. (1985)

in the human

(HTLV-III/LAV)

long

The

T cell terminal

Cell 41, 813-823 M. J., Bain, E. S., Luciw, P. A., and Peterlin,B. M. Structure, sequence, and position of the stem-loop in tar transcriptional

elongation

by tat through

the

HIV-l

long terminal repeat. Genes Dcv. 3, 547-558 20. Feng, S., and Holland, E. C. (1988) HIV-1 tat trans-activation requires the loop sequence within tar. Nature (London)334, 165-167 21. Dingwall, C., Ernberg, I., Gait, M. J., Green, S. M., Heaphy, S., Karn, J., Lowe, A. D., Singh, M., Skinner, M. A., and Valerio, R. (1989) Human immunodeficiency virus 1 tat protein binds ttuns-activation-responsive Proc. Nail. AcacL Sci. USA

region

(TAR)

RNA

in vitro.

86, 6925-6929

22. Roy, S., Delling, U., Chen, C. -H., Rosen, C. A., and Sonenberg, N. (1990) A bulge structure in HIV-1 TAR RNA is required for Tat binding and Tat-mediated trans-activation. Genes

Dcv. 4,

1365-13

73

23. Weeks, K. M., Ampe, C., Schultz, Crothers, D. M. (1990) Fragments

S. C., Steitz, T. A., and of the HIV-l Tat protein 2367

w.fasebj.org by Univ of So Dakota Lommen Hlth Sci Library (192.236.36.29) on September 14, 2018. The FASEB Journal Vol. ${article.issue.getVolume()}, No. ${article.issue.getIssueNum

specifically bind TAR RNA. Science 249, 1281-1285 24. Marciniak, R. A., Garcia-Blanco, M. A., and Sharp, P. A. (1990) Identification and characterization of a HeLa nuclear

protein

that

specifically

binds

to the trans-activation-response

element of human immunodeficiency virus. Proc. Natl. Acad. Sci. USA 87, 3624-3628 25. Marciniak, R. A., Calnan, B. J., Frankel, A. D., and Sharp, P. A. (1990) HIV-1 Tat protein trans-activates transcription in vitro. Cell 63, 791-802 26. Jakobovits, A., Rosenthal, A., and Capon, D. J. (1990) Transactivation of HIV-1 LTR-directed gene expression by tat requires protein kinase C. EMBO,J 9, 1165-1170 27. Wright, C. M., Felber, B. K., Paskalis, H., and Pavlakis, G. N. (1986) Expression and characterization of the trans-activator of HTLV-III/LAV virus.Science 234, 988-992 28. Kao, S. -Y., Calman, A. F., Luciw, P. A., and Peterlin, B. M. (TAR)

(1987) Anti-termination of transcription within the long terminal repeat of HIV-1 by tat gene product. Nature (London) 330, 489-493 29. Laspia, M. Tat protein

F., Rice, increases

A. P., and Mathews, M. transcriptional initiation Cell59, 283-292

B. (1989) HIV-1 and stabilizes

elongation. 30. Feinberg, M. B., Baltimore, D., and Frankel, A. D. (1991) The role of Tat in the human immunodeficiency virus life cycle indicates a primary effect on transcriptional elongation. Proc. NatI. Acad. Sd. USA 88, 4045-4049 31. Feinberg, M. B., Jarrett, R. F., Aldovini, A., Gallo, R. C., and Wong-Staal, F. (1986) HTLV-III expression and production involve complex regulation at the levels of splicing and translation of viral RNA. Cell 46, 807-817 32. Braddock, M., Thorburn, A. M., Chambers, A., Elliott, G. D., Anderson, G. J., Kingsman, A. J., and Kingsman, S. M. (1990) A nuclear translational block imposed by the HIV-1 U3 region is relieved by the Tat-TAR interaction. Cell 62, 1123-1133 33. Felber, B. K., Hadzopoulou-Cladaras, M., Cladaras, C.,

Copeland,

T., and Pavlakis,

G. N. (1989) Rev protein

of human

immunodeficiency virus type 1 affects the stability and transport of the viral mRNA. Proc. NatL Acad. Sci. USA 86, 1495-1499 34. Malim, M. H., Hauber,J., Le, S.-Y., Maizel,J. V., and Cullen, B. R. (1989) The HIV-1 rev trans-activator acts through a structured target sequence to activate nuclear export of unspliced viral mRNA. Nature (London) 338, 254-257

35. Sodroski,J.,

Goh, W. C., Rosen, C., Dayton,

A., Terwilliger,

and Haseltine, W. A. (1986) A second post-transcriptional activator gene required for HTLV-III replication. Nature don) 321, 412-417 36. Kim, S., Byrn, R., Groopman, J., and Baltimore, D.

Temporal

aspects

immunodeficiency

2368

Vol. 5

of DNA virus

July 1991

and RNA

infection:

synthesis

evidence

during

E., trans(Lon-

(1989)

human for differential gene

expression. j ViroL 63, 3708-3713 37. Pomerantz, R. J., Trono, D., Feinberg, M. B., and Baltimore, D. (1990) Cells nonproductively infected with HIV-1 exhibit an aberrant pattern of viral RNA for latency. Cell 61, 1271-1276

expression:

a molecular

model

38. Katz, RNA

R. A., and Skalka, A. M. (1990) Control of retroviral splicing through maintenance of suboptimal processing signals. Mol. CelL BioL 10, 696-704 39. Chang, D. D., and Sharp, P. A. (1989) Regulation by HIV Rev depends upon recognition of splice sites. Cell 59, 789-795

40. Rosen, C. A., Terwilliger, E., Dayton, A., Sodroski, J. G., and Haseltine, W. A. (1988) Intragenic cis-acting art gene-responsive sequences of the human immunodeficiency virus. Proc. NatI. Acad. Sci. USA 85, 2071-2075 41. Malim, M. H., Tiley, L. S., McCarn, D. F., Rusche, J. R., Hauber, J., and Cullen, B. R. (1990) HIV-1 structural gene expression requires binding of the Rev trans-activator to its RNA target sequence. Cell 60, 675-683 42. Olsen, H. S., Cochrane, A. W., Dillon, P. J., Nalin, C. M., and Rosen, C. A. (1990) Interaction of the human immunodeficiency virus type 1 Rev protein with a structured region in env mRNA is dependent on multimer formation mediated through a basic stretch of amino acids. Genes Dcv. 4, 1357-1364 43. Zapp, M. L., and Green, M. R. (1989) Sequence-specific RNA binding by the HIV-1 Rev protein. Nature (London) 342, 714-716 44. Malim, M. H., Bohnlein, S., Hauber, J., and Cullen, B. R.

(1989) Functional dissection of the HIV-1 Rev trans-activator-derivation of a trans-dominant repressor of Rev function. Cell 58, 205-214 45. Malim, M. H., and Culleri, B. R. (1991) HIV-l structural gene expression requires the binding of multiple Rev monomers to the viral RRE: implications for HIV-1 latency. Cell 65, 241-248 46. Hammes, S. R., Dixon, E. P., Malim, M. H., Cullen, B. R.,

and

Greene,

W.

C.

(1989)

Nef

protein

of

human

im-

munodeficiency virus type 1: evidence against its role as a transcriptional inhibitor. Proc. Natl. Acad. Sci. USA 86, 9549-9553 47. Ahmad, N., and Venkatesan, 5. (1988) Nef protein of HIV-1 is a transcriptional repressor of HIV-l LTR. Science 241, 1481-1485 48. Niederman, T. M. J., Thielan, B. J., and Ratner, L. (1989) Human immunodeficiency virus type 1 negative factor is a transcriptional silencer. Proc. NaIL Acad. Sci. USA 86, 1128-1132 49. Kim, S., Ikeuchi, K., Byrn, R., Groopman, J., and Baltimore,

D. (1989) Lack of a negative influence on viral growth by the nef gene of human immunodeficiency virus type 1. Proc. NatI. Acad. Sci. USA 86, 9544-9548 50. Cullen, B. R., and Greene, W. C. (1990) Functions of the auxilIary gene products of the human 1. Virology 178, 1-5

The FASEB Journal

immunodeficiency

virus

type

CULLEN

w.fasebj.org by Univ of So Dakota Lommen Hlth Sci Library (192.236.36.29) on September 14, 2018. The FASEB Journal Vol. ${article.issue.getVolume()}, No. ${article.issue.getIssueNum