Intracellular trafficking of Gag and Env proteins and their interactions modulate pseudotyping of retroviruses

Abstract

Glycoproteins derived from most retroviruses and from several families of enveloped viruses can form infectious pseudotypes with murine leukemia virus (MLV) and lentiviral core particles, like the MLV envelope glycoproteins (Env) that are incorporated on either virus type. However, coexpression of a given glycoprotein with heterologous core proteins does not always give rise to highly infectious viral particles, and restrictions on pseudotype formation have been reported. To understand the mechanisms that control the recruitment of viral surface glycoproteins on lentiviral and retroviral cores, we exploited the fact that the feline endogenous retrovirus RD114 glycoprotein does not efficiently pseudotype lentiviral cores derived from simian immunodeficiency virus, whereas it is readily incorporated onto MLV particles. Our results indicate that recruitment of glycoproteins by the MLV and lentiviral core proteins occurs in intracellular compartments and not at the cell surface. We found that Env and core protein colocalization in intracytoplasmic vesicles is required for pseudotype formation. By investigating MLV/RD114 Env chimeras, we show that signals in the cytoplasmic tail of either glycoprotein differentially influenced their intracellular localization; that of MLV allows endosomal localization and hence recruitment by both lentiviral and MLV cores. Furthermore, we found that upon membrane binding, MLV core proteins could relocalize Env glycoproteins in late endosomes and allow their incorporation on viral particles. Thus, intracellular colocalization, as well as interactions between Env and core proteins, may influence the recruitment of the glycoprotein onto viral particles and generate infectious pseudotyped viruses.

FIG. 1.

Schematic representation of RD114/MLV Env chimeras showing the domain organization of parental Env and chimeras. The open and solid boxes represent domains derived from RD114 (RD) and amphotropic MLV glycoproteins, respectively. The cytoplasmic region consists of the cytoplasmic tails (T) and the p2R peptides (R). E, ectodomain; M, transmembrane domain. The first and last amino acids of each domain are indicated.

FIG. 2.

Results of infection assays. TE671 target cells were infected using diluted supernatants harvested from 293T cells transfected with GFP marker-gene-expressing SIV or MLV vectors and with the indicated viral glycoproteins and core components. The transduction efficiency, expressed as the percentage of GFP-positive cells, was measured by fluorescence-activated cell sorter analysis 72 h postinfection and was used to determine the number of infectious viral particles present in the producer cell supernatants. The infectious-virus titers were then expressed as the number of GFP infectious units per milliliter of viral supernatant. The values are the means plus standard deviations of up to six independent experiments. Similar results were obtained with COS-7 cells as the producer cells (not shown).

FIG. 3.

Characterization of glycoproteins and pseudotypes. Detection of precursor (Pr), surface (SU), and two transmembrane forms (TM and TM*, deleted from the p2R carboxy-terminal peptide) of the RD114 and RD/TR^MLV glycoproteins and of SIV or MLV core proteins (Gag) was performed by Western blotting in crude cell lysates (Total cell), cell surface biotinylated proteins (Cell surface), and viral particles purified by ultracentrifugation of the cell supernatants on sucrose cushions (Pellet). As a control, the CD25 interleukin-2 receptor gamma chain that is not expressed in the producer cells was coexpressed with the viral proteins. The viral glycoproteins were expressed in 293T cells individually (A) or in the presence of Gag proteins derived from SIV (B) or MLV (D) or of a myristoylation-defective MLV G2A-Gag mutant (C). In separate experiments, the samples were digested with PNGaseF before being immunoblotted with anti-SU antibodies to readily distinguish SU from precursor proteins and to facilitate the relative quantification of the different Env forms (Pr, SU, and TM) as displayed in Fig. 4. Thus, upon deglycosylation by PNGaseF, the SU and Pr proteins appeared as 60- and 40-kDa bands (e.g., viral particles in B and C, insets). In this transient expression system, the budding of MLV particles was ∼5-fold more efficient than that of lentiviral particles, as evidenced both by different ratios between intracellular and virion-associated Gag proteins (data not shown) and by the difference between the infectious-virus titers of the two virus types pseudotyped with VSV G. Thus, to facilitate the comparison of the incorporations of the different glycoproteins on either type of virus core, equivalent amounts of vector particles were loaded on gels, using capsid proteins as the standard. The processing of the RD114 and RD/TR^MLV TM in TM* was blocked when the producer cells were treated with saquinavir, an inhibitor of the retroviral proteases (not shown). Similar results were obtained when COS-7 cells were used as the producer cells (not shown).

FIG. 4.

Quantification of mature glycoproteins. Upon expression with the indicated core components (Gag) in 293T cells, the SU, TM, and precursor forms of the RD114 and RD/TR^MLV glycoproteins were quantified by Western blot analysis performed in total cell lysates, cell surface, and purified viral particles (Fig. 3). Scanning of Western blot membranes was performed with the Storm 860 device, and signal densities were calculated with the ImageQuant program. Transfection efficiencies were normalized using the CD25 interleukin-2 receptor gamma-chain marker. The ratios of the mature (processed) glycoproteins to total glycoproteins are provided. na, not applicable.

FIG. 5.

Detection of viral glycoproteins and core proteins by confocal microscopy. (A) COS-7 cells expressing either the RD114 and RD/TR^MLV glycoproteins (Env) or the MLV and SIV core proteins (Gag), as indicated, were fixed, permeabilized, and stained using RD114 SU antibodies, MLV capsid protein antibodies, or SIV matrix protein antibodies. The transfected cells were grown on glass coverslips, fixed in paraformaldehyde, stained at room temperature, and imaged by confocal microscopy. The arrows indicate intracytoplasmic vesicles in which the RD/TR^MLV glycoprotein localized. Similar staining patterns were observed with 293T cells (not shown). Expression of myristoylation-defective MLV Gag proteins (G2A Gag mutant) resulted in diffuse staining throughout the cytoplasm and an absence of punctate staining (not shown). (B to E) Colocalization of Env and Gag viral components. COS-7 cells coexpressing the RD114 (B and D) or RD/TR^MLV (C and E) glycoprotein and the MLV (D and E) or SIV (B and C) core protein, as indicated, were fixed, permeabilized, and stained using RD114 SU antibodies (B to E), SIV matrix protein antibodies (B and C), or MLV capsid protein antibodies (D and E). Env and Gag coexpression are shown in the red (Env detection) and green (Gag detection) channels, respectively. The arrows in the red channel show intracytoplasmic vesicles in which the RD114 glycoprotein localized upon expression of MLV Gag protein (D). Colocalization of both viral components was assessed by confocal microscopy analysis (Merge). The arrows indicate intracytoplasmic vesicles in which the Env and Gag protein colocalized. Similar colocalization patterns were detected in 293T cells (not shown). Neither Env/Gag colocalization nor redistribution of Env localization was detected upon expression of myristoylation-defective MLV Gag proteins (data not shown).

FIG. 6.

Colocalization of Env and cellular markers of the endosomal pathway. COS-7 cells coexpressing MLV Gag proteins and RD/TR^MLV (A) or RD114 (B) glycoproteins were imaged by confocal microscopy analysis. Localization of the Env glycoproteins detected in the red channel (Env) in marked intracellular vesicles shown in the green channel (Marker) was assessed upon coexpression of vectors encoding the GFP-tagged cellular markers cellubrevin (Cellubrevin-GFP) and TI-VAMP (TI-VAMP-GFP), which localize in recycling and late endosomes, respectively, or upon costaining with Lamp-1 antibodies (α-Lamp-1) for detection of lysosomes. Colocalization of both Env and cellular markers is shown (Merge). The arrows indicate localization of Env in recycling or late endosomes.