Intracellular distribution of human T-cell leukemia virus type 1 Gag proteins is independent of interaction with intracellular membranes

Molecular Determinants of Human T-cell Leukemia Virus Type 1 Gag Targeting to the Plasma Membrane for Assembly

Assembly of human T-cell leukemia virus type 1 (HTLV-1) particles is initiated by the trafficking of virally encoded Gag polyproteins to the inner leaflet of the plasma membrane (PM). Gag-PM interactions are mediated by the matrix (MA) domain, which contains a myristoyl group (myr) and a basic patch formed by lysine and arginine residues. For many retroviruses, Gag-PM interactions are mediated by phosphatidylinositol 4,5-bisphosphate [PI(4,5)P2]; however, previous studies suggested that HTLV-1 Gag-PM interactions and therefore virus assembly are less dependent on PI(4,5)P2. We have recently shown that PI(4,5)P2 binds directly to HTLV-1 unmyristoylated MA [myr(-)MA] and that myr(-)MA binding to membranes is significantly enhanced by inclusion of phosphatidylserine (PS) and PI(4,5)P2. Herein, we employed structural, biophysical, biochemical, mutagenesis, and cell-based assays to identify residues involved in MA-membrane interactions. Our data revealed that the lysine-rich motif (Lys47, Lys48, and Lys51) constitutes the primary PI(4,5)P2-binding site. Furthermore, we show that arginine residues 3, 7, 14 and 17 located in the unstructured N-terminus are essential for MA binding to membranes containing PS and/or PI(4,5)P2. Substitution of lysine and arginine residues severely attenuated virus-like particle production, but only the lysine residues could be clearly correlated with reduced PM binding. These results support a mechanism by which HTLV-1 Gag targeting to the PM is mediated by a trio engagement of the myr group, Arg-rich and Lys-rich motifs. These findings advance our understanding of a key step in retroviral particle assembly.

M-Sec induced by HTLV-1 mediates an efficient viral transmission

Human T-cell leukemia virus type 1 (HTLV-1) infects target cells primarily through cell-to-cell routes. Here, we provide evidence that cellular protein M-Sec plays a critical role in this process. When purified and briefly cultured, CD4⁺ T cells of HTLV-1 carriers, but not of HTLV-1^- individuals, expressed M-Sec. The viral protein Tax was revealed to mediate M-Sec induction. Knockdown or pharmacological inhibition of M-Sec reduced viral infection in multiple co-culture conditions. Furthermore, M-Sec knockdown reduced the number of proviral copies in the tissues of a mouse model of HTLV-1 infection. Phenotypically, M-Sec knockdown or inhibition reduced not only plasma membrane protrusions and migratory activity of cells, but also large clusters of Gag, a viral structural protein required for the formation of viral particles. Taken together, these results suggest that M-Sec induced by Tax mediates an efficient cell-to-cell viral infection, which is likely due to enhanced membrane protrusions, cell migration, and the clustering of Gag.

In the present study, we identified the cellular protein M-Sec as a host factor necessary for de novo infection of human T-cell leukemia virus type 1 (HTLV-1), the causative retrovirus of an aggressive blood cancer known as adult T-cell leukemia/lymphoma. The inhibition or knockdown of M-Sec in infected cells resulted in a reduced viral infection in several culture models and a mouse model. We recently demonstrated a similar role of M-Sec in macrophages infected with another human retrovirus HIV-1, but it has been generally thought that M-Sec is not related to HTLV-1 infection because of the lack of its expression in CD4⁺ T cells, the major target of HTLV-1. In this study, we revealed that CD4⁺ T cells of HTLV-1 asymptomatic carriers, but not those of HTLV-1^- individuals, expressed M-Sec, and that the viral protein Tax mediated the induction of M-Sec. Thus, M-Sec is a new and useful tool for further understanding the process of HTLV-1 transmission.

Structural Insights into the Mechanism of Human T-cell Leukemia Virus Type 1 Gag Targeting to the Plasma Membrane for Assembly

Retroviral Gag targeting to the plasma membrane (PM) for assembly is mediated by the N-terminal matrix (MA) domain. For many retroviruses, Gag-PM interaction is dependent on phosphatidylinositol 4,5-bisphosphate (PI(4,5)P₂). However, it has been shown that for human T-cell leukemia virus type 1 (HTLV-1), Gag binding to membranes is less dependent on PI(4,5)P₂ than HIV-1, suggesting that other factors may modulate Gag assembly. To elucidate the mechanism by which HTLV-1 Gag binds to the PM, we employed NMR techniques to determine the structure of unmyristoylated MA (myr(-)MA) and to characterize its interactions with lipids and liposomes. The MA structure consists of four α-helices and unstructured N- and C-termini. We show that myr(-)MA binds to PI(4,5)P₂ via the polar head and that binding to inositol phosphates (IPs) is significantly enhanced by increasing the number of phosphate groups on the inositol ring, indicating that the MA-IP binding is governed by charge-charge interactions. The IP binding site was mapped to a well-defined basic patch formed by lysine and arginine residues. Using an NMR-based liposome binding assay, we show that PI(4,5)P₂and phosphatidylserine enhance myr(-)MA binding in a synergistic fashion. Confocal microscopy data revealed formation of puncta on the PM of Gag expressing cells. However, G2A-Gag mutant, lacking myristoylation, is diffuse and cytoplasmic. These results suggest that although myr(-)MA binds to membranes, myristoylation appears to be key for formation of HTLV-1 Gag puncta on the PM. Altogether, these findings advance our understanding of a key mechanism in retroviral assembly.

Structural and molecular determinants of HIV-1 Gag binding to the plasma membrane

Targeting of the Gag polyprotein to the plasma membrane (PM) for assembly is a critical event in the late phase of immunodeficiency virus type-1 (HIV-1) infection. Gag binding to the PM is mediated by interactions between the myristoylated matrix (MA) domain and PM lipids. Despite the extensive biochemical and in vitro studies of Gag and MA binding to membranes over the last two decades, the discovery of the role of phosphatidylinositol-4,5-bisphosphate [PI(4,5)P₂] in Gag binding to the PM has sparked a string of studies aimed at elucidating the molecular mechanism of retroviral Gag–PM binding. Electrostatic interactions between a highly conserved basic region of MA and acidic phospholipids have long been thought to be the main driving force for Gag–membrane interactions. However, recent studies suggest that the mechanism is rather complex since other factors such as the hydrophobicity of the membrane interior represented by the acyl chains and cholesterol also play important roles. Here we summarize the current understanding of HIV-1 Gag–membrane interactions at the molecular and structural levels and briefly discuss the underlying forces governing interactions of other retroviral MA proteins with the PM.

A reduction of viral mRNA, proteins and induction of altered morphogenesis reveals the anti-HTLV-1 activity of the labdane-diterpene myriadenolide in vitro

Background

Human T-lymphotropic virus 1 (HTLV-1) has been associated with leukemia/lymphoma (ATL) and myelopathy/tropical spastic paraparesis (HAM/TSP), in addition to other inflammatory diseases as well as infection complications. Therapeutic approaches for HTLV-1-related pathologies are limited. The labdane diterpene myriadenolide (AMY) is a natural product that exhibit biological activities, such as anti-inflammatory and antiviral activity as reported for HIV and herpesvirus.

Results

We demonstrated that this natural product was able to inhibit the expression of gag-pol mRNA and substantially reduced the expression of the structural proteins p19 and gp46. Comparison of treated and untreated cells shows that AMY alters both the morphology and the release of viral particles. The Atomic Force Microscopy assay showed that the AMY treatment reduced the number of particles on the cell surface by 47%.

Conclusion

We demonstrated that the labdane diterpene myriadenolide reduced the expression of the structural proteins and the budding of viral particles, besides induces altered morphogenesis of HTLV-1, conferring on AMY a new antiviral activity that may be useful for the development of new compounds with specific anti-HTLV-1 activity.

Interrelationship between cytoplasmic retroviral Gag concentration and Gag-membrane association

The early events in the retrovirus assembly pathway, particularly the timing and nature of Gag translocation from the site of protein translation to the inner leaflet of the plasma membrane, are poorly understood. We have investigated the interrelationship between cytoplasmic Gag concentration and plasma membrane association using complementary live-cell biophysical fluorescence techniques in real time with both human T-cell leukemia virus type 1 (HTLV-1) and human immunodeficiency virus type 1 (HIV-1) Gag proteins. In particular, dual-color, z-scan fluorescence fluctuation spectroscopy in conjunction with total internal reflection fluorescence and conventional, epi-illumination imaging were utilized. Our results demonstrate that HTLV-1 Gag is capable of membrane targeting and particle assembly at low (i.e., nanomolar) cytoplasmic concentrations and that there is a critical threshold concentration (approaching micromolar) prior to the observation of HIV-1 Gag associated with the plasma membrane. These observations imply fundamental differences between HIV-1 and HTLV-1 Gag trafficking and membrane association.

Cross-reactive antibodies to target proteins are dependent upon oligomannose glycosylated epitopes in HTLV-1 associated neurological disease

Our lab recently identified a cross-reactive antibody response between human T-lymphotropic virus type-1-p24-(gag) (HTLV-1-p24-(gag)) and peroxiredoxin-1 (PrX-1) as potentially contributing to the pathogenesis of HTLV-1 associated neurological disease via molecular mimicry. These targets proteins were glycosylated, yet the glycan side chains immunoreactive with the immunoglobulins were unknown. Using a combination of lectin isolation and serial enzymatic deglycosylation of glycoproteins, we determined that the immunoreactive epitopes contained branched oligomannose side chains. These data suggest that post-translational glycosylation specifically related to oligomannose immunoreactivity to both the infecting and host antigens may contribute to molecular mimicry and be important in the pathogenesis of HTLV-1 associated neurological disease.

Characterization of cytoplasmic Gag-gag interactions by dual-color z-scan fluorescence fluctuation spectroscopy

Fluorescence fluctuation spectroscopy (FFS) quantifies the interactions of fluorescently-labeled proteins inside living cells by brightness analysis. However, the study of cytoplasmic proteins that interact with the plasma membrane is challenging with FFS. If the cytoplasmic section is thinner than the axial size of the observation volume, cytoplasmic and membrane-bound proteins are coexcited, which leads to brightness artifacts. This brightness bias, if not recognized, leads to erroneous interpretation of the data. We have overcome this challenge by introducing dual-color z-scan FFS and the addition of a distinctly colored reference protein. Here, we apply this technique to study the cytoplasmic interactions of the Gag proteins from human immunodeficiency virus type 1 (HIV-1) and human T-lymphotropic virus type 1 (HTLV-1). The Gag protein plays a crucial role in the assembly of retroviruses and is found in both membrane and cytoplasm. Dual-color z-scans demonstrate that brightness artifacts are caused by a dim nonpunctate membrane-bound fraction of Gag. We perform an unbiased brightness characterization of cytoplasmic Gag by avoiding the membrane-bound fraction and reveal previously unknown differences in the behavior of the two retroviral Gag species. HIV-1 Gag exhibits concentration-dependent oligomerization in the cytoplasm, whereas HTLV-1 Gag lacks significant cytoplasmic Gag-Gag interactions.

New insights into HTLV-1 particle structure, assembly, and Gag-Gag interactions in living cells

Human T-cell leukemia virus type 1 (HTLV-1) has a reputation for being extremely difficult to study in cell culture. The challenges in propagating HTLV-1 has prevented a rigorous analysis of how these viruses replicate in cells, including the detailed steps involved in virus assembly. The details for how retrovirus particle assembly occurs are poorly understood, even for other more tractable retroviral systems. Recent studies on HTLV-1 using state-of-the-art cryo-electron microscopy and fluorescence-based biophysical approaches explored questions related to HTLV-1 particle size, Gag stoichiometry in virions, and Gag-Gag interactions in living cells. These results provided new and exciting insights into fundamental aspects of HTLV-1 particle assembly-which are distinct from those of other retroviruses, including HIV-1. The application of these and other novel biophysical approaches promise to provide exciting new insights into HTLV-1 replication.

Gag localization and virus-like particle release mediated by the matrix domain of human T-lymphotropic virus type 1 Gag are less dependent on phosphatidylinositol-(4,5)-bisphosphate than those mediated by the matrix domain of HIV-1 Gag

The human immunodeficiency virus type 1 (HIV-1) Gag matrix (MA) domain facilitates Gag targeting and binding to the plasma membrane (PM) during virus assembly. Interaction with a PM phospholipid, phosphatidylinositol-(4,5)-bisphosphate [PI(4,5)P(2)], plays a key role in these MA functions. Previous studies showed that overexpression of polyphosphoinositide 5-phosphatase IV (5ptaseIV), which depletes cellular PI(4,5)P(2), mislocalizes HIV-1 Gag to the cytosol and greatly reduces HIV-1 release efficiency. In this study, we sought to determine the role of the MA-PI(4,5)P(2) interaction in Gag localization and membrane binding of a deltaretrovirus, human T-lymphotropic virus type 1 (HTLV-1). We compared the chimeric HIV-1 Gag (HTMA), in which MA was replaced with HTLV-1 MA, with wild-type HIV-1 and HTLV-1 Gag for PI(4,5)P(2) dependence. Our results demonstrate that, unlike HIV-1 Gag, subcellular localization of and VLP release by HTLV-1 and HTMA Gag were minimally sensitive to 5ptaseIV overexpression. These results suggest that the interaction of HTLV-1 MA with PI(4,5)P(2) is not essential for HTLV-1 particle assembly. Furthermore, liposome-binding analyses showed that both HTLV-1 and HTMA Gag can bind membrane efficiently even in the absence of PI(4,5)P(2). Efficient HTLV-1 Gag binding to liposomes was largely driven by electrostatic interaction, unlike that of HIV-1 Gag, which required specific interaction with PI(4,5)P(2). Furthermore, membrane binding of HTLV-1 Gag in vitro was not suppressed by RNA, in contrast to HIV-1 Gag. Altogether, our data suggest that Gag targeting and membrane binding mediated by HTLV-1 MA does not require PI(4,5)P(2) and that distinct mechanisms regulate HIV-1 and HTLV-1 Gag membrane binding.

The HTLV-1 Virological Synapse

Human T-lymphotropic virus-1 (HTLV-1) spreads efficiently between T-cells via a tight and highly organized cell-cell contact known as the virological synapse. It is now thought that many retroviruses and other viruses spread via a virological synapse, which may be defined as a virus-induced, specialized area of cell-to-cell contact that promotes the directed transmission of the virus between cells. We summarize here the mechanisms leading to the formation of the HTLV-1 virological synapse and the role played by HTLV-1 Tax protein. We propose a model of HTLV-1 transmission between T-cells based on the three-dimensional ultrastructure of the virological synapse. Finally, in the light of recent advances, we discuss the possible routes of HTLV-1 spread across the virological synapse.

Development of a cytotoxic T-cell assay in rabbits to evaluate early immune response to human T-lymphotropic virus type 1 infection

Human T-lymphotropic virus type 1 (HTLV-1) infection causes adult T-cell lymphoma/leukemia (ATL) following a prolonged clinical incubation period, despite a robust adaptive immune response against the virus. Early immune responses that allow establishment of the infection are difficult to study without effective animal models. We have developed a cytotoxic T-lymphocyte (CTL) assay to monitor the early events of HTLV-1 infection in rabbits. Rabbit skin fibroblast cell lines were established by transformation with a plasmid expressing simian virus 40 (SV40) large T antigen and used as autochthonous targets (derived from same individual animal) to measure CTL activity against HTLV-1 infection in rabbits. Recombinant vaccinia virus (rVV) constructs expressing either HTLV-1 envelope surface unit (SU) glycoprotein 46 or Tax proteins were used to infect fibroblast targets in a (51)Cr-release CTL assay. Rabbits inoculated with Jurkat T cells or ACH.2 cells (expressing ACH HTLV-1 molecule clone) were monitored at 0, 2, 4, 6, 8, 13, 21, and 34 wk post-infection. ACH.2-inoculated rabbits were monitored serologically and for viral infected cells following ex vivo culture. Proviral load analysis indicated that rabbits with higher proviral loads had significant CTL activity against HTLV-1 SU as early as 2 wk post-infection, while both low- and high-proviral-load groups had minimal Tax-specific CTL activity throughout the study. This first development of a stringent assay to measure HTLV-1 SU and Tax-specific CTL assay in the rabbit model will enhance immunopathogenesis studies of HTLV-1 infection. Our data suggest that during the early weeks following infection, HTLV-1-specific CTL responses are primarily targeted against Env-SU.

The inner loop of tetraspanins CD82 and CD81 mediates interactions with human T cell lymphotrophic virus type 1 Gag protein

The tetraspanin superfamily proteins play important roles in organizing membrane protein complexes, modulating integrin function, and controlling T cell adhesion. Tetraspanins such as CD82 contain two extracellular loops with its N terminus, C terminus, and inner loop exposed to the cytoplasm. The matrix (MA) domain of human T cell lymphotrophic virus, type 1 (HTLV-1), Gag interacts with the cytoplasmic face of the plasma membrane and is concentrated at tetraspanin-enriched microdomains. To understand the basis of this association, we generated site-directed mutations in the various domains of CD82 and used coimmunoprecipitation and colocalization approaches to examine interactions with HTLV-1 MA. The large extracellular loop of CD82, which is important for interactions with integrins, was not required for the association with HTLV-1 MA. The cytoplasmic N terminus and C terminus of CD82 were also dispensable for CD82-MA interactions. In contrast, mutations of conserved amino acids in the inner loop of CD82 or of palmitoylated cysteines that flank the inner loop diminished CD82 association with MA. HTLV-1 MA also interacted with the inner loop of CD81. Thus, association of HTLV-1 Gag with tetraspanin-enriched microdomains is mediated by the inner loops of CD81 and CD82.

Human Dlg protein binds to the envelope glycoproteins of human T-cell leukemia virus type 1 and regulates envelope mediated cell-cell fusion in T lymphocytes

Human homologue of the Drosophila Dlg tumor suppressor (hDlg) is a widely expressed scaffold protein implicated in the organization of multi-protein complexes at cell adhesion sites such as the neuronal synapse. hDlg contains three PDZ domains that mediate its binding to the consensus motifs present at the C-termini of various cell surface proteins, thus inducing their clustering and/or stabilization at the plasma membrane. Using a yeast two-hybrid screen, we identified hDlg as a cellular binding partner of a viral membrane integral protein, the envelope glycoprotein (Env) of human T-cell leukemia virus type 1 (HTLV-1). HTLV-1 is a human retrovirus that infects CD4+ T lymphocytes and is preferentially transmitted via direct contacts between infected and target cells, through a structure referred to as the virological synapse. Here, we demonstrate that hDlg interacts with a classical PDZ domain-binding motif present at the C-terminus of the cytoplasmic domain of HTLV-1 Env and conserved in the related HTLV-2 virus. We further document that, in HTLV-1 infected primary T cells, hDlg and Env are concentrated in restricted areas of the plasma membrane, enriched in molecules involved in T-cell contacts. The presence of Gag proteins responsible for viral assembly and budding in these areas indicated that they constitute platforms for viral assembly and transmission. Finally, a mutant virus unable to bind hDlg exhibited a decreased ability to trigger Env mediated cell fusion between T lymphocytes. We thus propose that hDlg stabilizes HTLV-1 envelope glycoproteins at the virological synapse formed between infected and target cells, hence assisting the cell-to-cell transmission of the virus.

Retrovirus budding

The release of retrovirus particles from the infected cell is greatly stimulated by short motifs, known as "late" or "L" domains, present within the Gag precursor protein. Three distinct classes of L domains have been identified; these bear the core sequence: Pro-Thr/Ser-Ala-Pro [P(T/S)AP], Pro-Pro-x-Tyr (PPxY), or Tyr-Pro-x-Leu (YPxL). A number of recent studies have demonstrated that L domains function by interacting with components of the machinery responsible for sorting cellular proteins into the multivesicular body (MVB) pathway. This review traces the history of L domain discovery and characterization, and highlights the relationship between L domain activity, retrovirus release, and the host endosomal sorting machinery.

Both the PPPY and PTAP motifs are involved in human T-cell leukemia virus type 1 particle release

In retroviruses, the late (L) domain has been defined as a conserved motif in the Gag polyprotein precursor that, when mutated, leads to the emergence of virus particles that fail to pinch off from the plasma membrane. These domains have been observed to contain the PPXY, PTAP, or YXXL motifs. The deltaretroviruses, which include bovine leukemia virus (BLV) and human T-cell leukemia virus type 1 (HTLV-1) and HTLV-2, have a conserved PPPY motif in the C-terminal region of the matrix (MA) domain of Gag, while HTLV-1 also encodes a PTAP motif in MA. In this study, we analyzed the roles of the PPPY and PTAP motifs in the C terminus of MA in HTLV-1 particle release. Mutation of either motif (i.e., PPPY changed to APPY or PTAP changed to PTRP) reduced budding efficiencies. Particle buds and electron-dense regions of plasma membrane were observed by electron microscopy. When the locations of PPPY and PTAP were switched, particle release was eliminated. Intriguingly, the replacement of the PTAP motif with either the PPPY or YPDL motifs did not influence the release of virus particles, but the replacement of the PPPY motif with either PTAP or YPDL eliminated particle production. This indicates that the role that PPPY plays in HTLV-1 budding cannot be replaced with either PTAP or YPDL. A similar observation was made with the BLV PPPY motif. Finally, HTLV-1 particle release was found to be sensitive to proteasome inhibitors, implicating a role for ubiquitin in HTLV-1 budding. In summary, our observations indicate that (i) the PPPY motif plays a crucial role in virus budding and (ii) the PTAP motif plays a more subtle role in HTLV-1 particle release. Each of these motifs may play an important role in virus release from specific cell types and therefore be important in efficient virus spread and transmission.

Alpha interferon inhibits human T-cell leukemia virus type 1 assembly by preventing Gag interaction with rafts

Alpha-2a interferon (IFN-alpha2a) has beneficial clinical effects on human T-cell leukemia virus type 1 (HTLV-1) infection, but its antiviral mechanism of action is unknown. Antiviral effects of IFN-alpha2a were studied in 293T cells expressing HTLV-1 proviral DNA and in HTLV-1-infected cells (HOS/PL, MT2, and HUT102). In 293T cells, an 50% inhibitory concentration of 10 U of IFN-alpha2a/ml was determined by p19 antigen ELISA. Analysis of IFN-treated cells demonstrated no defect in viral protein synthesis but did show a decrease in the level of released virus, as determined by immunoblot assays. Electron microscopy studies of IFN-treated cells revealed neither a defect in the site of virus budding nor tethering of virus particles at the plasma membrane, thus arguing against an effect on virus release. Cell fractionation studies and confocal microscopy showed no effect of IFN on Gag association with membranes. However, the level of Gag association with lipid rafts was decreased, suggesting a role of IFN in inhibiting HTLV-1 assembly.

Spread of HTLV-I between lymphocytes by virus-induced polarization of the cytoskeleton

Cell contact is required for efficient transmission of human T cell leukemia virus- type 1 (HTLV-I) between cells and between individuals, because naturally infected lymphocytes produce virtually no cell-free infectious HTLV-I particles. However, the mechanism of cell-to-cell spread of HTLV-I is not understood. We show here that cell contact rapidly induces polarization of the cytoskeleton of the infected cell to the cell-cell junction. HTLV-I core (Gag protein) complexes and the HTLV-I genome accumulate at the cell-cell junction and are then transferred to the uninfected cell. Other lymphotropic viruses, such as HIV-1, may similarly subvert normal T cell physiology to allow efficient propagation between cells.

Intracellular localisation of Fv1

Retroviral resistance mediated by the murine Fv1 gene is believed to result from a direct interaction between the Fv1 gene product and the viral capsid protein. To study the mechanism of Fv1 action, the expression and intracellular localisation of the Fv1 protein were examined. Only very low levels of protein expression seem necessary for virus restriction but the site of expression appears crucial. Active Fv1 was found in association with tubules of the trans-Golgi network, whereas an inactive form was localised in the endoplasmic reticulum. We hypothesize that Fv1 is compartmentalised in the cell on the pathway taken by virus en route to the nucleus, suggesting that incoming virus must pass the trans-Golgi network during its transit to the nucleus.

Time course of Gag protein assembly in HIV-1-infected cells: a study by immunoelectron microscopy

Recent biochemical studies have identified high molecular complexes of the HIV Gag precursor in the cytosol of infected cells. Using immunoelectron microscopy we studied the time course of the synthesis and assembly of a HIV Gag precursor protein (pr55gag) in Sf9 cells infected with recombinant baculovirus expressing the HIV gag gene. We also immunolabeled for pr55gag human T4 cells acutely or chronically infected with HIV-1. In Sf9 cells, the time course study showed that the first Gag protein appeared in the cytoplasm at 28-30 h p.i. and that budding started 6-8 h later. Colloidal gold particles, used to visualize the Gag protein, were first scattered randomly throughout the cytoplasm, but soon clusters representing 100 to 1000 copies of pr55gag were also observed. By contrast, in cells with budding or released virus-like particles the cytoplasm was virtually free of gold particles while the released virus-like particles were heavily labeled. Statistical analysis showed that between 80 and 90% of the gold particles in the cytoplasm were seen as singles, as doublets, or in small groups of up to five particles probably representing small oligomers. Clusters of gold particles were also observed in acutely infected lymphocytes as well as in multinuclear cells of chronically infected cultures of T4 cells. In a few cases small aggregates of gold particles were found in the nuclei of T4 lymphocytes. These observations suggest that the Gag polyprotein forms small oligomers in the cytoplasm of expressing cells but that assembly into multimeric complexes takes place predominantly at the plasma membrane. Large accumulations of Gag protein in the cytoplasm may represent misfolded molecules destined for degradation.