-
Minireviews
HIV Accessory Proteins: EmergingTherapeutic Targets
Roger H. Miller and Nava SarverTargeted Interventions Branch,
Basic Sciences Program, Division ofAIDS, National Institute of
Allergy and Infectious Diseases, NationalInstitutes of Health,
Bethesda, Maryland, U.S.A.
HIV ACCESSORY PROTEINSIn addition to the capsid (gag),
polymerase (pol),and envelope (env) genes common to all
retrovi-ruses, human immunodeficiency virus type 1(HIV-1) possesses
six additional genes (i.e., nef,rev, tat, vif, vpr, and vpu) that
encode regulatoryproteins (Fig. 1). Four of these proteins, Nef,
Vif,Vpr, and Vpu, were originally termed "accesso-ry" proteins due
to the fact that their function invitro appeared to be nonessential
for HIV repli-cation (reviewed in Ref. 1). However, it is
nowevident that these proteins play an importantrole in viral
replication in vivo, and some may beintimately involved in HIV
pathogenesis and dis-ease progression.
The following review is based on a recentsymposium (HIV
Accessory Proteins: Therapeu-tic Opportunities, February 2, 1995,
Washington,DC) focusing on the therapeutic opportunitiesafforded by
HIV accessory proteins.
NEFNef, a viral protein expressed early in infection,possesses
two distinct functions: the ability toenhance HIV replication and
the ability to down-regulate the T cell surface CD4 molecule
(re-viewed in Ref. 2). Enhancement of virus replica-tion results
from the presence of virion-associated Nef that increases the
infectivity ofcell-free virus. This was revealed by studies
com-paring the replication capabilities of infectiousprovirus
plasmids and virus particles in CEM
Address correspondence and reprint requests to: NavaSarver,
Targeted Interventions Branch, Basic Sciences Pro-gram, Division of
AIDS, NLAID, NIH, Bethesda, MD 20892-7620.
Copyright 1995, Molecular Medicine,
1076-1551/95/$10.50/0Molecular Medicine, Volume 1, Number 5, July
1995 479-485
cells (a human T cell line) in a single-cycle assaywhere
subsequent rounds of infection were in-hibited by the addition of
neutralizing antibody.Transfection of CEM cells with identical
amountsof either a nef deletion (A) mutant or wild-type(wt)
proviral DNA yielded similar expression lev-els of the HIV capsid
protein p24. However,when CEM cells were infected with equal
inoc-ula of Anef and wt virus (standardized for theirp24 content),
Anef virus yielded 3- to 5-fold lessp24 than did the wt HIV- 1.
Thus, the presence ofparticle-associated Nef in wt virus resulted
in anincrease in p24 production in infected CEM cell.
The mechanism whereby Nef confers repli-cative advantage on the
virus is not yet estab-lished. Comparison of virions with or
withoutNef revealed no difference in reverse tran-scriptase (RT)
activity, viral binding, or viral en-try into host cells. The
reduction in infectivity ofAnefvirus does not seem to reflect on an
inabilityto degrade CD4 (see below). When cells engi-neered to
express a truncated CD4 molecule,lacking virtually the entire
cytoplasmic domainrequired for Nef-induced CD4 degradation,
wereinfected with Anefor wt HIV, attenuation of Anefvirus growth
was still observed.
The severe combined immunodeficientmouse implanted with human
thymus and livertissue (SCID-thy/liv) is an important animalmodel
for studying HIV infection and pathogen-esis. Mice are generated by
surgical transplanta-tion of human fetal liver and thymus
fragmentsunder the kidney capsule. After several months,the
implanted tissues become fused together toform a graft, which is
histologically and physio-logically indistinguishable from a normal
humanthymus. Direct inoculation of the graft with HIVresults in
infection of human CD4-positive cellsand precursor CD4/CD8-double
positive cells.
479
-
480 Molecular Medicine, Volume 1, Number 5, July 1995
HIV-1 PROVIRAL GENOME
S;P.
POL
GAG F
PR
RT
ENV
REV EXII
IVPI
ITAT EX)
'PU
IN
FIG. 1. Diagrammatic representation of the HIV-1 genome.The long
terminal repeats (LTRs), divided into the unique 3' (U3), repeat
(R), and unique 5' (U5) domains, flankthe protein encoding regions
of the virus genome (light blue). The large precursor proteins
encoded by the capsid(GAG), polymerase (POL), and envelope (ENV)
genes are proteolytically cleaved as depicted by the vertical
lines.The GAG protein is cleaved into the smaller proteins p17,
p24, and p6; the POL protein is cleaved into the pro-tease (PR),
reverse transcriptase (RT), and integrase (IN) proteins; and the
ENV protein, with a signal peptide(S.P.) sequence at the amino
terminus, is cleaved into the gp 120 (surface) and gp4l
(transmembrane) proteins.Exons (EX) 1 and 2 of the rev and tat
genes as well as the genes encoding the accessory proteins are
highlighted bycolor.
The SCID-thy/liv model was used to evaluate theeffect of HIV
regulatory proteins on CD4-positivecells. CEM cells were
transfected with infectiousplasmids with deletions in the nef, vif,
vpr, or vpugenes, and the virus produced was titered onactivated
human peripheral blood lymphocytes(PBLs). SCID-thy/liv mice were
then inoculatedwith equal amounts of infectious virus. Suchanalysis
revealed a hierarchy in the abilities ofthe four deletion mutant
viruses to replicate andinduce CD4 depletion in the graft relative
toinfection with wt HIV as follows: Avpr > Avpu >Avif>
Anef. Deleting the nefgene was thus moredeleterious to virus
replication than deleting anyof the other regulatory genes
studied.
Another well-established function of Nef isits ability to
down-regulate CD4, a surface mol-ecule crucial for helper T cell
signaling and aprincipal cellular receptor for HIV. When nefgene
constructs were expressed in CD4 positivecells, the half-life of
CD4 was found to decreasefrom 24 to 4-6 hr. While expression of CD4
wasunaffected, once the protein reached the cyto-plasmic membrane
Nef induced rapid endocyto-sis followed by lysosomal degradation of
CD4 (3).
Recent work indicates that Nef may disrupt theinteraction
between CD4 and the p56Ick tyrosinekinase (4). Site-directed
mutagenesis of the cy-toplasmic tail of CD4 (the binding site of
p56lckand Nef) reveal that mutated CD4 moleculeswith decreased
responsiveness to Nef also havereduced capacity to associate with
p561ck. Aminoacids 407 and 410 within a hydrophobic domainof CD4
are especially important. Thus, it is likelythat Nef, or a
Nef-recruited factor, competes withp561ck for a binding site on the
cytoplasmic tail ofCD4 to disrupt CD4/p561ck interaction,
therebyinducing CD4 endocytosis and its subsequentdegradation.
VPUVpu is a regulatory protein encoded by HIV- 1 butnot by HIV-2
or most of the simian immunode-ficiency viruses. This
phosphorylated transmem-brane protein has two distinct and
separablefunctions in the HIV replication cycle: enhance-ment of
viral export from infected cells and deg-radation of CD4 in the
endoplasmic reticulum
5' LTR 3' LTR5
-
Minireview 481
(ER). The requirement for Vpu in particle releaseis cell-type
specific. Pulse-chase experiments uti-lizing constructs expressing
HIV p6 Gag (a pro-cessed Gag peptide derived from the
carboxyl-terminus of p55 Gag) and/or Vpu indicate thatthe latter
facilitates particle release in HeLa butnot in COS cells. This
finding implies the pres-ence of a Vpu-interacting protein in some
cellsbut not in others. Additional support for a cellu-lar
Vpu-interacting protein comes from a studyshowing that Vpu
increases the release of heter-ologous retrovirus particles.
Using the yeast two-hybrid system a Vpubinding protein, or UBP,
was identified. The pro-tein, approximately 69 kD in size, is
encoded bya 2.9-kb cellular transcript expressed in a varietyof
human tissues and cells, including PBLs. Par-tial analysis
indicates that UBP shares certainsequence similarities with the
immunophilin su-perfamily of proteins, which includes the
cyclo-philins and the FK506 binding proteins (FKBPs).It is still
unclear what role UBP plays in mediat-ing Vpu function(s) and
whether UBP/Vpu in-teraction is required for HIV replication.
Insight into the mechanism of Vpu-mediatedviral export comes
from solution NMR experi-ments demonstrating that the
transmembranedomain of oligomeric Vpu has the potential toform an
ion conductive membrane pore. Exper-iments in Xenopus oocytes
verified the ion chan-nel activity of synthetic Vpu.
Interestingly,FKBP12, a member of the immunophilin familyrelated to
UBP, also interacts with a cellular pro-tein involved in ion
channel formation. Whilemutation of the transmembrane domain of
Vpuresulted in the loss of Vpu-mediated ion channelformation
activity and caused a reduction in vi-rus release, there was no
effect on CD4 degrada-tion, indicating that these two functions of
Vpuare separate.
The mechanism for Vpu-mediated CD4 deg-radation differs from
that described for Nef. InHIV-infected cells, gp 160 Env and CD4
formcomplexes that become sequestered in the roughER, preventing
further transport of gp 160 andCD4 to the cellular membrane. Vpu
reverses thisinhibition of gp 160 sequestering by
selectivelydegrading CD4 (5). Site-directed mutagenesis ofVpu
suggests that both the anchor and cytoplas-mic domains of the
protein are required for thisprocess.
As indicated, both the enhancement of viralsecretion and the
induction of CD4 degradationby Vpu involve distinct molecular
mechanisms.It is not clear at this time whether blocking either
function alone will have an impact on viralgrowth and
pathogenesis or whether both func-tions will need to be targeted
for maximal ther-apeutic benefit.
VIFSeveral independent studies have shown that theVif protein
significantly increases the infectivityof HIV- 1 particles. The
importance of this proteinis underscored by the observation that
virtuallyall lentiviruses examined (the single exceptionbeing
equine infectious anemia virus) possess avif gene. While Vif was
previously thought to befound only in the cytoplasm of infected
cells (6),recent experiments have revealed the existenceof Vif in
mature virions. Studies with anti-Vifantibodies show that Vif
exists within the viruscore rather than on the virus surface, and
it isestimated that 20 Vif molecules are encapsidatedper
virion.
Within the cytoplasm of infected cells, Vifexists both as
soluble and as membrane-associ-ated forms. Site-directed
mutagenesis of vif re-vealed that the basic amino acids at the
carboxylterminus of the molecule (amino acids 157-160and 173-184)
are important for association withcell membranes. Interestingly,
pretreatment ofmembranes with trypsin abolishes the ability ofVif
to bind, suggesting that Vif interacts with amembrane-associated
protein which stably an-chors it to the membrane surface.
Again, these studies indicate an associationbetween an HIV- 1
regulatory protein, in this caseVif, with a cellular protein(s).
Such proteins arelikely to be required for the proper functioning
ofthe regulatory proteins in viral replication andmay provide
additional targets for therapeuticintervention. One advantage in
targeting a cel-lular "partner" over a viral gene product is that
itobviates potential emergence of HIV escape mu-tants resistant to
specific anti-viral agents.
In additional studies of the role of Vif in virusreplication,
Vif mutant virions purified fromCEM cells (but not those purified
from SupTIcells) were shown to exhibit a major defect inendogenous
reverse transcriptase (RT) activity.Radiolabeled nucleotides were
incorporated pri-marily into low molecular weight DNA frag-ments in
the mutant virus. This suggests that ayet unidentified component of
the virus coreinvolved in reverse transcription is a target forVif
function. Although Vif does not effect RNAdimerization, the level
of unprocessed p55 Gag
-
482 Molecular Medicine, Volume 1, Number 5, July 1995
in Avif virions was elevated compared with thatin wt virions. At
the present time, it is not clearwhat this effect on Gag processing
means withregards to the role of Vif in HIV infectivity.
A second function of Vif involves reorgani-zation of
cytoskeletal elements, consisting largelyof microtubules and
intermediate filaments.While microtubules are composed of actin
ortubulin, intermediate filaments are exclusivelymade of several
proteins, notably vimentin. Ex-periments examining the cellular
localization ofVif in HeLa cells transfected with an
expressionplasmid revealed that Vif associates specificallywith
intermediate filaments comprised of vimen-tin. Treatment of HeLa
cells with brefeldin A, adrug that mediates the trafficking of
membraneproteins from the ER to the Golgi, produced pe-rinuclear
caps of vimentin filaments. Vif ex-pressed in HeLa cells
colocalized specifically tosuch structures following drug
treatment. WhileVif does not affect the phosphorylation or
solu-bility of vimentin, it causes a dramatic reorgani-zation of
the cytoskeleton. Vif may, therefore,also function to enhance the
movement of thevirion to the nucleus by reorganizing the
cellularcytoskeletal elements. One hypothesis is that af-ter the
virion enters the host cell and uncoatingbegins, Vif attaches to
vimentin and initiates thereorganization of intermediate filaments.
As-sisted by a vimentin-associated factor capable ofmoving along
the filaments, the Vif-containingviral pre-integration complex is
efficiently trans-ported to the nucleus via a
"three-dimensional"process, rather than a "two-dimensional"
pro-cess in the absence of Vif. Conceivably, therapeu-tic agents
that block Vif's ability to interact withvimentin are likely to
interfere with the trans-port of the pre-integration complex to the
nu-cleus, and dramatically affect the ability of HIV toefficiently
infect lymphocytes.
VPRVpr exists as an oligomer (possibly a tetramer) inthe nucleus
of infected cells. Vpr is incorporatedinto virions via a specific
mechanism involvingp6 Gag (7). Virus strains encoding Vpr
replicatefaster and to higher levels than strains unable toencode a
functional Vpr protein. Structural anal-ysis of Vpr has focused on
two fronts: determin-ing the location of the nuclear localization
signal(NLS) that facilitates transport of the nascentprotein into
the cell nucleus, and identifying thesequence(s) responsible for
Vpr oligomerization
(see below). Typical NLS elements consist of astretch of
positively charged amino acids. TheNLS of Vpr is atypical in that
it is located in anegatively charged domain at the amino termi-nus
of the molecule. The Vpr NLS resides in adomain coincident with a
region required for theinteraction of the protein with a 180-kD
cellularfactor that may facilitate transport of Vpr into
thenucleus. Recombinant Vpr expressed in Esche-richia coli is
capable of forming oligomeric struc-tures and preliminary work
indicates that aminoacids at the amino terminus of the molecule
arerequired for oligomerization. Mutagenesis exper-iments mapped
the crucial domain to amino ac-ids 36-43. Speculations as to the
purpose of Vproligomerization include expanding the capabili-ties
of the small (i.e., 14 kD) monomeric proteinand/or increasing the
packaging efficiency of theprotein.
The role of Vpr as an integral component ofthe virus may be
explained by its profound bio-logical effect on cells; Vpr was
shown to regulatecell proliferation as well as cell
differentiation(8,9). As Vpr is released into the cell
duringuncoating of the virion, it is presumed to "pre-pare" an
environment favorable to HIV replica-tion and gene expression. In
one study, an infec-tious molecular HIV clone was altered to
expressthe murine Thy protein, a cell surface moleculethat can
serve as a convenient marker for mon-itoring virus gene expression.
Using a dual stain-ing method to analyze concomitantly the
cellcycle status and the infection status of individualcells,
infection with the altered HIV recombinantwas shown to induce the
accumulation of cellsarrested in G2 in a variety of cell lines
(PBL,SupTl, HeLa, and COS cells). In contrast, mockinfected cells
had twice as many cells in Gl thanin G2. When the vpr gene was
mutated, therecombinant infectious HIV could no longer in-duce G2
arrest. Interestingly, the vpr mutant vi-rus was still capable of
spreading and inducingsyncytia. Another study examined the effects
ofbaculovirus-expressed Vpr on tumor cell prolif-eration. While
purified Vpr caused a dramaticinhibition of proliferation, latently
infected celllines treated with Vpr exhibited reactivation ofvirus
replication. Such effects were the samewhether recombinant Vpr or
Vpr purified fromHIV-infected sera were used. Overall, these
stud-ies implicate Vpr in inducing a long-lasting in-crease in
cellular permissiveness to HIV replica-tion by modulating the
differentiation stage ofthe host cell.
These studies are consistent with recent find-
-
Minireview 483
ings demonstrating that viruses containing anintact vpr gene are
unable to establish a chronicinfection of cultured lymphocytes due
to thedeath of the infected cells (10). This effect isobserved late
in infection and occurs even inPBLs transfected with a provirus
with a func-tional vpr gene and a defective env gene.
Sincecell-to-cell spread does not occur under thesecircumstances,
cell death must be mediated viaVpr synthesized de novo from the
integrated pro-virus and not from import of virion-associatedVpr
into the cell. Combined with other results,the data are consistent
with the hypothesis thatVpr either prevents cells from continued
prolif-eration or induces their terminal differentiation.The latter
could explain why primary macro-phages, which are differentiated
cells, are capa-ble of long-term expression of HIV without
sig-nificant cell death. Additional experiments arenecessary to
determine the relevance of the invitro observations to HIV
replication in vivo.
How does Vpr exert such profound effects oncells? A recent
report suggests that Vpr is linkedto the glucocorticoid steroid
pathway (11). Anewly discovered cellular protein which associ-ates
with Vpr provides the link to this importantregulatory pathway.
Vpr-interacting protein 1(Rip-i) is a 41 -kD cytosolic protein that
is ex-pressed in a wide variety of tissues and cells,including
those that support HIV replication.Rip- I is translocated to the
nucleus by both Vpror glucocorticoid receptor (GR)
-II-stimulatingsteroids. The finding that Vpr and Rip-I
coim-munoprecipitate with the human GR provides apotential
biochemical mechanism for Vpr's activ-ity and suggests the possible
use of anti-glucocor-ticoid agents as inhibitors of HIV
replication.
PROSPECTS FOR THERAPYAn emerging theme is that Nef, Vif, Vpr,
and Vpuare involved in crucial aspects of HIV replicationand
release, cell susceptibility to infection, andmodulation of signal
transduction (Table 1). It isalso apparent that several of these
regulatoryproteins share common functions mediated bydifferent
mechanisms, such as the downregula-tion of CD4 by both Nef and Vpu.
While Nefpromotes the internalization of surface CD4 andsubsequent
degradation in lysosomal vesicles,Vpu targets CD4 sequestered in
the ER via aCD4-gp 160 complex. Why does the virus devotetwo gene
products to CD4 down-regulation? Ithas been shown that removal of
the HIV receptor
from the cell surface benefits HIV replication bypreventing
superinfection and by enhancing therelease of progeny virions;
additional studies willdetermine whether CD4 down-regulation
isimportant in other aspects of HIV infection ofhumans.
One of the most intriguing findings pre-sented at the meeting
was that the HIV regula-tory proteins interact with specific
cellular factorsto perform their functions. Whatever the
precisemechanisms involved, Nef, Vif, Vpr, and Vpu andtheir
associated cellular proteins afford highlypromising yet
underexploited therapeutic oppor-tunities. Conceivably, targeting
cellular factorsmay be more effective than targeting viral
fac-tors. This premise is founded on the observationthat cellular
genes replicate with considerablymore fidelity than HIV genes,
thereby providingconserved targets for therapeutic
interventions.Thus, the serious and prevalent problem associ-ated
with the evolution of virus variants resistantto the inhibitory
effect of the drug(s) could bealleviated. Although a downside to
targeting cel-lular factors is the potential cellular toxicity,
itshould be possible, nonetheless, to strike a clin-ical balance
with significant antiviral effect andmanageable toxicity.
Importantly, it is possiblethat a synergistic inhibitory effect may
beachieved in targeting the viral and cellular pro-teins
simultaneously using combination thera-pies. Overall, efforts to
exploit these dual targetsshould be explored in devising therapies
againstHIV infection which may be used alone or incombination with
other anti-HIV therapeuticstrategies.
Understanding the mode of action of the HIVregulatory proteins
and their associated cellularfactors opens a window for deciphering
the in-tricate mechanism(s) associated with HIV-medi-ated cellular
dysfunction and dysregulation. Thisknowledge, in turn, is
imperative for developingeffective therapies against HIV infection,
eitherby identification of specific inhibitory agents inbiochemical
screening assays, a rational drug de-sign approach driven by
structure-function rela-tionship, or by novel therapeutic
strategies (e.g.,gene-based inhibition).
ACKNOWLEDGMENTSWe thank the following investigators for
theirpresentations at the meeting, some of which aresummarized
herein: Grace Aldronvandi, UCLASchool of Medicine and Jonsson
Comprehensive
-
484 Molecular Medicine, Volume 1, Number 5, July 1995
TABLE 1. HIV accessory proteins
InteractingMolecular CellularWeight Cellular Virus Factor(s)
Known/Putative(kD) Modifications Location Association (size, kD)
Functions
Nef 27 Myristoylated Cytoplasm Yes Ick (56)a * Increases
virioninfectivity
* Down-regulatessurface CD4
Vif 23 ? Cytoplasm Yes (?)b * Increases virioninfectivity
* Reorganizescytoskeletalelements
Vpr 14 ? Nucleus Yes Rip-I (41)C *
Promotescellulardifferentiation
RIP (180)d 0 Arrests cells inG2 phase of cellcycle
* Interacts withglucocorticoidsteroid pathway
* Prevents cellproliferationduring chronicinfection
Vpu 16 Phosphorylated Cytoplasm ? UBP (69)e 0 Enhances
viralexport frominfected cells viaion channelformation
* Promotes CD4degradation inER
aNef appears to compete for a CD4 binding site with the p561ck
tyrosine kinase to induce endocytosis of CD4.bVif may interact with
two uncharacterized cellular factors: (i) a membrane-associated
protein which stably anchors Vif on themembrane surface and (ii) a
vimentin-associated factor capable of moving a Vif-containing
pre-integration complex along inter-mediate filaments.'Rip-1 is a
cytosolic protein that is translocated to the nucleus by Vpr. Rip-I
coprecipitates with Vpr and the glucocorticoid recep-tor.dRIP is
also reported to be involved in Vpr nuclear transport but the exact
relationship of Rip-I and RIP has not yet been estab-lished. The
fact that both Rip-I and RIP appear to be involved in Vpr nuclear
transport has led to the speculation that the latterrepresents a
multimeric form of Rip-i.eUBP shares similarities with the
immunophilin family of proteins that includes the cyclophilins and
the FK506 binding proteins.It has been speculated that UBP may be
involved in the ion channel formation capability of Vpu.
Cancer Center; Dana Gabuzda, Dana Farber Can-cer Institute; John
Guatelli, UCSD and the SanDiego Veterans Affairs Medical Center; M.
AbdulJabbar, The Cleveland Clinic Foundation; JohnKappes, UAB;
Nathaniel Landau, Aaron Dia-mond AIDS Research Center; Jon Marsh,
Na-tional Institute of Mental Health; Antonio Pan-
ganiban, University of Wisconsin; VincentePlanelles, UCLA School
of Medicine and JonssonComprehensive Cancer Center; Ulrich
Schubert,National Institute of Allergy and Infectious Dis-eases;
Jacek Skowronski, Cold Spring HarborLaboratories; Klaus Strebel,
National Institute ofAllergy and Infectious Diseases; Didier
Trono,
-
Minireview 485
The Salk Institute for Biological Studies; RobertVigne, INSERM;
David Weiner, University ofPennsylvania; and Lingjun Zhao,
University ofKansas Medical Center. We also thank OpendraSharma for
helpful suggestions on the programand Carl Dieffenbach for critical
reading of themanuscript.
REFERENCES1. Subbramanian RA, Cohen EA. (1994) Mo-
lecular biology of the human immunodefi-ciency virus accessory
proteins. J. Virol. 68:6831-6835.
2. Cullen BR. (1994) The role of Nef in thereplication cycle of
the human and simianimmunodeficiency viruses. Virology 205:
1-6.
3. Aiken C, Konner J, Landau NR, LenburgME, Trono D. (1994) Nef
induces CD4 endo-cytosis: Requirement for a critical dileucinemotif
in the membrane-proximal CD4 cyto-plasmic domain. Cell 76:
853-864.
4. Salghetti S, Mariani R, Skowronski J. (1995)Human
immunodeficiency virus type 1 Nefand p56Ick protein-tyrosine kinase
interactwith a common element in CD4 cytoplasmictail. Proc. Natl.
Acad. Sci. U.S.A. 92: 349-353.
5. Bour S, Schubert U, Strebel K. (1995) Thehuman
immunodeficiency virus type 1 Vpuprotein specifically binds to the
cytoplasmic
domain of CD4: Implications for the mecha-nism of degradation.
J. Virol. 69: 1510-1520.
6. Goncalves J, Jallepalli P, Gabuzda DH.(1994) Subcellular
localization of the Vifprotein of human immunodeficiency virustype
1. J. Virol. 68: 704-712.
7. Paxton W, Connor RI, Landau NR. (1993)Incorporation of Vpr
into human immuno-deficiency virus type 1 virions: Requirementfor
the p6 region of gag and mutational anal-ysis. J. Virol. 67:
7229-7237.
8. Levy DN, Refaeli Y, MacGregor BR, WeinerDB. (1994) Serum Vpr
regulates productiveinfection and latency of human
immunode-ficiency virus type 1. Proc. Natl. Acad. Sci.U.S.A. 91:
10873-10877.
9. Levy DN, Refaeli Y, Weiner DB. (1995) Ex-tracellular Vpr
protein increases cellular per-missiveness to human
immunodeficiency vi-rus replication and reactivates virus
fromlatency. J. Virol. 69: 1243-1252.
10. Rogel ME, Wu LI, Emerman M. (1995) Thehuman immunodeficiency
virus type 1 vprgene prevents cell proliferation duringchronic
infection. J. Virol. 69: 882-888.
11. Refaeli Y, Levy DN, Weiner DB. (1995) Theglucocorticoid
receptor type II complex is atarget of the HIV-1 vpr gene product.
Proc.Natl. Acad. Sci. U.S.A. 92: 3621-3625.