HIV-1 assembly: viral glycoproteins segregate quantally to lipid rafts that associate individually with HIV-1 capsids and virions

Abstract

HIV-1 assembly depends on its structural protein, Gag, which after synthesis on ribosomes, traffics to the late endosome/plasma membrane, associates with HIV Env glycoprotein, and forms infectious virions. While Env and Gag migrate to lipid microdomains, their stoichiometry and specificity of interaction are unknown. Pseudotyped viral particles can be made with one viral core surrounded by heterologous envelope proteins. Taking advantage of this property, we analyzed the association of HIV Env and Ebola glycoprotein (GP), with HIV-1 Gag coexpressed in the same cell. Though both viral glycoproteins were expressed, each associated independently with Gag, giving rise to distinct virion populations, each with a single glycoprotein type. Confocal imaging demonstrated that Env and GP localized to distinct lipid raft microdomains within the same cell where they associated with different virions. Thus, a single Gag particle associates "quantally" with one lipid raft, containing homogeneous trimeric viral envelope proteins, to assemble functional virions.

Figure 1

Analysis of Buoyant Density Gradient Fractionated Env/GP Pseudotyped Lentiviral Vectors 293T cells were transfected with the packaging vector pCMVΔR8.2, pHR′-CMV-Luciferase, as well as the Env + GP expression vector. (A) Virus supernatant was harvested 60 hr later for buoyant density gradient analysis. (B) Rabbit anti-GP, anti-gp120, and anti-Gag were used for subsequent western blot detection of GP, Env, and Gag, respectively, in each fraction. (C) Comparable expression of GP and Env was confirmed on viral producer cells transfected singly (GP or Env) or doubly (GP/Env) with the indicated spikes.

Figure 2

Specificity of Lentiviral Vectors Determined by Inhibition of Gene Transfer with Neutralizing Antibodies and Biochemical Evidence of Segregation of Viral Envelopes (A) Pseudotyped lentiviral vectors produced from single (Env or GP) or doubly (GP/Env) transfected 293T cells infect corresponding HIV (MT2) and Ebola (786-O) target cells. Error bars indicate standard error of the mean of at least three independent transductions. (B) Antibody neutralization of pseudotyped lentiviral vectors produced from doubly transfected 293T cells and fractionated in buoyant density gradient. For antibody neutralization, viruses were incubated at 37°C for 1 hr with 2F5 and 2G12 (5 μg/ml each), or for GP with KZ52 (20 μg/ml) before infecting the target cells.

Figure 3

Evidence of Segregation in Virions Made from Producer Cells Expressing Two Alternative Viral Spikes by Immunodepletion and Biochemical Analyses (A) Immunodepletion of Optiprep band-purified virions prepared from double expression plasmid (Env/GP) with anti-Env (2F5 and 2G12) or anti-GP (KZ52) as indicated. Error bars indicate standard error of the mean of at least three independent transductions. (B) Primary (lanes 1–18) and secondary immunoprecipitation (lanes 19–36) of purified virions with heterologous antibody, anti-GP or anti-Env, respectively, followed by western blotting for the indicated gene products (right), providing biochemical evidence for the absence of chimeric virus. For immunodepletion, virus supernatants were incubated with 2F5 and 2G12 (5 μg/ml each) or for GP with KZ52 (20 μg/ml) at 4°C for 60 min before addition of biomagnetic particle-conjugated anti-human IgG for another 3 hr. Immune complexes were held in place with a strong magnet during the removal of the virus supernatant. Human IgG served as negative control. Immunoprecipitation of Env containing pseudotyped lentiviral vector was performed by incubating with human HIV Ig_b12 (1 μg/ml) and GP containing pseudotyped lentiviral vector with 13C6 (1 μg/ml) at 4°C for 60 min before adding protein A/G for another 3 hr. Immune complexes were separated from the virus supernatant by centrifugation at 4000 × g for 3 min. Between primary and secondary immunoprecipitation, the virus supernatants were precleared by incubating with protein G agarose for 3 hr before removal of the agarose beads. mAb 149, rabbit anti-gp120, and anti-Gag (Advanced Biotechnologies, Columbia, MD) were used for subsequent western blot detection of GP, Env, and Gag, respectively, in each immune complex.

Figure 4

Confocal Microscopy of Transfected MAGI-CCR5 Cells Revealed that GP Did Not Colocalize with Env, and This Pattern Did Not Change in the Presence or Absence of Gag (A) MAGI-CCR5 cells were transfected with myc- or HA-Env expression vectors and tagged within the V1 region (see Experimental Procedures), which showed colocalization and served as a positive control (left panels). MAGI-CCR5 cells were transiently transfected with GP and HA-Env-expression vectors (right panels). Immunofluorescent staining was performed 40 hr later. Z stack sectioning of the cells was collected and analyzed by Imaris software to obtain 3D reconstruction and colocalization statistics. (B) MAGI-CCR5 cells were transiently transfected with Env, GP, or double expression plasmid (Env/GP) in the presence or absence of packaging vector pCMVΔR8.2, which contains Gag. Immunofluorescent staining was performed 40 hr later. Z stack sectioning of the cells was collected and analyzed by Imaris software.

Figure 5

Lack of Colocalization of Heterologous Envelope Proteins, Env and GP, in Producer Cells Confocal microscopy revealed that GP did not colocalize with Env. MAGI-CCR5 cells were transiently transfected with packaging vector pCMVΔR8.2 and GP- and HIV-Env_ADA-expressing vectors. Immunofluorescence staining was performed 40 hr later. Z stack sectioning of the cells was collected and analyzed by Imaris software to obtain iso-surface models. (A) Expanded view of a thick section of a single cell is presented here as an iso-surface model depicting Env, GP, and Gag distribution in the vicinity of the plasma membrane. (B)–(D) depict two of the three viral proteins presented in (A) for clear visual comparisons. Descriptions of the method used to identify membrane versus cytoplasmic proteins with Imaris software are described further in the Experimental Procedures and Figure S4.

Figure 6

Distribution of Env and Gag in HIV-Infected Cells (A) Confocal imaging confirms the expression of Env in discrete locations intracellularly, largely separate from Gag in MAGI-CCR5 and αCD3/CD28 stimulated human CD4 lymphocytes. MAGI-CCR5 or human CD4 cells were infected with HIV-1 89.6 and analyzed by immunofluorescent staining 48 hr later. Approximately one percent of cells were infected by this criteria. (B) Z stack sectioning of the cells was collected and analyzed by Imaris software to obtain iso-surface models as in Figure 5. A thick section of a single cell is presented here as iso-surface model for each cell type depicting Env and Gag distribution within the cytoplasm and plasma membrane as well as areas of colocalization.

Figure 7

Models of Segregation of Viral Spikes with Lentiviral Ribonucleoprotein Complexes While it was expected that cells expressing two different trimeric viral spikes could give rise to virions containing both proteins, trafficking of such spikes to distinct lipid rafts that associate individually with Gag particles would give rise to virions containing one or the other viral glycoprotein, a model supported by the data in this study.