Retromer regulates HIV-1 envelope glycoprotein trafficking and incorporation into virions

Abstract

The envelope glycoprotein (Env) of the Human Immunodeficiency Virus Type-1 (HIV-1) is a critical determinant of viral infectivity, tropism and is the main target for humoral immunity; however, little is known about the cellular machinery that directs Env trafficking and its incorporation into nascent virions. Here we identify the mammalian retromer complex as a novel and important cellular factor regulating Env trafficking. Retromer mediates endosomal sorting and is most closely associated with endosome-to-Golgi transport. Consistent with this function, inactivating retromer using RNAi targeting the cargo selective trimer complex inhibited retrograde trafficking of endocytosed Env to the Golgi. Notably, in HIV-1 infected cells, inactivating retromer modulated plasma membrane expression of Env, along with Env incorporation into virions and particle infectivity. Mutagenesis studies coupled with coimmunoprecipitations revealed that retromer-mediated trafficking requires the Env cytoplasmic tail that we show binds directly to retromer components Vps35 and Vps26. Taken together these results provide novel insight into regulation of HIV-1 Env trafficking and infectious HIV-1 morphogenesis and show for the first time a role for retromer in the late-steps of viral replication and assembly of a virus.

Figure 1. Vps26 depletion increases plasma membrane expression of HIV-1 Env and incorporation into viral particles.

A) HeLa TZM-bl cells were transfected twice with control or Vps26 siRNA. Forty-eight hours after the second transfection cell lysates were prepared and subjected to SDS-PAGE and western blotting for Vps26 and actin as a loading control. Band intensities were quantified using ImageJ and the percentage of remaining Vps26 relative to actin was calculated. One representative Western blot from three independent experiments is shown. B) HeLa TZM-bl cells were infected with HIV-1 immediately after the second knockdown. Forty-eight hours later viral supernatants were pelleted by ultracentrifugation through sucrose and subjected to SDS-PAGE and western blotting for HIV-1 Env (gp120 and gp41) and Gag (p24CA). One representative Western blot from three independent experiments is shown. The band intensities from three independent experiments were quantified using ImageJ and the ratio of HIV-1 Env gp120 to Gag p24CA is shown. Error bars show the SEM. Cell lysates were also prepared and blotted for Gag p55 (lower panel). C) Equal volumes of viral supernatants from infected cells were used to infect Jurkat 1G5 reporter cells. Infectious virus titer is shown as luciferase relative light units (RLU). One representative of three independent experiments is shown. Errors bars show the SEM. D) Flow cytometry analysis of HIV-1 Env plasma membrane expression following Vps26 KD. One histogram representative of three independent experiments is shown. Cells were either left uninfected (filled grey), untreated and infected (solid grey line), control siRNA treated and infected (broken black line) or Vps26 siRNA treated and infected (solid black line) and surface stained for Env on ice prior to fixation. E) Pooled cell surface mean fluorescence intensity (MFI) of Env staining from three independent experiments. Error bars show the SEM. F) Cell lysates from HIV-1 infected siRNA treated cells were subjected to SDS-PAGE and Western blotting for Vps26, tubulin, HIV-1 Env and Gag. One representative of three independent experiments is shown. The band intensities from three independent experiments were quantified using ImageJ and the ratio of unprocessed (gp160) and processed (gp120) Env and Gag (p55) is shown. Error bars show the SEM.

Figure 2. Immunofluorescence staining of Env in HIV-1 infected cells.

A) HeLa TZM-bl cells were infected with HIV-1, fixed, permeabilized and stained for HIV-1 Env (green) and retromer component Vps26 or the Golgi marker giantin or (red). Panels are single xy slices and are representative examples from three independent experiments. Arrows highlight coincident labeling of Env and Vps26. Scale bar is 20 microns. The amount of immunoreactive Env colocalizing with Vps26 or giantin was calculated from at least 20 cells.

Figure 3. The cytoplasmic tail of HIV-1 Env mediates sensitivity to retromer-depletion.

A) Amino acid sequence of the NL4.3 HIV-1 cytoplasmic tail. YSPL in bold italics highlights the membrane-proximal YxxL endocytosis motif. The two leucines in bold show the site of truncations of the gp41 cytoplasmic tail in the CD8-gp41 constructs used in Figure 6. B) HeLa TZM-bl cells were transfected with control or Vps26 siRNA and infected with Δ144 NL4.3 lacking the C-terminal 144 amino acids of the Env cytoplasmic tail. Viral supernatants were pelleted and cell lysates prepared and subjected to SDS-PAGE and Western blotting for HIV-1 Env and Gag. One representative western blot from three independent experiments is shown. The band intensities from three independent experiments were quantified using ImageJ and the ratio of HIV-1 Env gp120 to Gag p24CA in virus preparations is shown. Error bars show the SEM. C) Flow cytometry analysis of HIV-1 Env plasma membrane expression following Vps26 KD. Cells were either left uninfected (filled grey), untreated and infected (solid grey line), control siRNA treated and infected (broken black line) or Vps26 siRNA treated and infected (solid black line). One representative from three independent experiments performed with HIV-1 Δ144 (top panel) and WT (lower panel) is shown.

Figure 4. The gp41 cytoplasmic tail mediates retromer-dependent Golgi retrieval.

A) Immunostaining of intracellular CD8-gp41CT (green) and Vps26 (red) in HeLa cells. Panels are single xy slices and are a representative example from three independent experiments. Coincident staining appears yellow. Scale bar is 10 microns. B) CD8-gp41CT expressing cells were treated with control or Vps26 siRNA. Cells were incubated with anti-CD8 monoclonal antibody at 4°C for 30 min, washed and incubated 37°C for 10 and 120 min. Cells were fixed, permeabilized and stained for the Golgi marker giantin (red) and fluorescently conjugated anti-mouse secondary antibody to localize the pool of internalized CD8-gp41CT (green). Panels are maximum intensity projections reconstructed from serial Z sections through the entire volume of the cell. Data are representative of three independent experiments. Scale bar is 20 microns. C) Immunostaining of total intracellular CD8-gp41CT (green) and Vps35 (red) in HeLa cells. Panels are single xy slices and are a representative example from three independent experiments. Coincident staining appears yellow. Scale bar is 15 microns. D) CD8-gp41CT expressing cells were treated with control or Vps35 siRNA and antibody-feeding and Golgi retrieval performed as described in B. Data are representative of three independent experiments. Scale bar is 20 microns. Western blot confirming efficient knockdown of Vps35 in HeLa cells following transfection with control or Vps35 siRNA as described in Figure 1. Band intensities were quantified using ImageJ and the percentage of remaining Vps35 relative to tubulin was calculated. One representative Western blot from three independent experiments is shown.

Figure 5. The gp41 cytoplasmic tail binds directly to retromer.

A) Native coimmunoprecipitation with anti-CD8 identifies retromer components Vps26 and Vps35 as interacting with the HIV-1 gp41CT. Cell lysates prepared from untransfected (UT), CD8-CIMPR or CD8-gp41CT expressing HeLa cells were incubated with anti-CD8 coated beads and co-IP proteins were subjected to SDS-PAGE and western blotting for Vps35 and Vps26. Untransfected HeLa cells were used as a negative control and CD8-CIMPR as a positive control. B) GST-pulldown confirms direct binding of the gp41CT to retromer. Purified recombinant FLAG-tagged retromer complex (3xFLAG-Vps26-Vps29-Vps35-His₆) was incubated with purified GST or GST fusion proteins containing the CIMPR (GST-CIMPR) or Envgp41 (GST-gp41CT) cytoplasmic tail and proteins were pulled down with glutathione-Sepharose 4B beads. Bound retromer components Vps26 and Vps35 were detected by immunoblotting.

Figure 6. The C terminal 100 amino acids of the Env cytoplasmic tail are required to bind retromer.

A) Cells expressing reporter constructs containing truncated EnvCT were treated with control or Vps26 siRNA and an antibody feeding assay was performed to follow Golgi retrieval of endocytosed protein as described in Figure 4. Golgi marker giantin (red) and internalized CD8-gp41CT (green). Panels are maximum intensity projections reconstructed from serial Z sections through the entire volume of the cell. Data are representative of three independent experiments. Scale bar is 20 microns. B) Cells from A stained for Vps26 show that full length and truncated mutants are all efficiently endocytosed and colocalize with intracellular Vps26. Scale bar is 20 microns. C) Truncation of the gp41CT abrogates co-IP of Vps26 and Vps35. Cell lysates prepared from untransfected (UT) or cells stably expressing CD8-reporter constructs were incubated with anti-CD8 coated beads and co-IP proteins were subjected to SDS-PAGE and Western blotting for Vps35 and Vps26. Untransfected HeLa cells were used as a negative control and CD8-CIMPR as a positive control. D) HeLa cells expressing CD8-reporter constructs were treated with cycloheximide for the indicated periods of time and cell lysates subjected to SDS-PAGE and western blotting. One representative of two independent experiments is shown.

Figure 7. Identification of two internal sequences within the EnvCT required for retromer binding.

A) Antibody-feeding and retrieval assay was performed as described in Figure 4. Panels are maximum intensity projections reconstructed from serial Z sections through the entire volume of the cell. Data are representative of two independent experiments. Scale bar is 20 microns. B) Cell lysates prepared from cells stably expressing CD8-gp41CT constructs were incubated with anti-CD8 coated beads and co-IP proteins were subjected to SDS-PAGE and Western blotting for Vps35 and Vps26. Deleting either of two internal regions of the EnvCT abrogates co-IP of Vps26 and Vps35.

Figure 8. The effect of retromer depletion on HIV-1 assembly in T cells.

A) Western blot analysis of Vps26 expression in Jurkat T cells stably expressing shRNA targeting Vps26 or control non-targeting shRNA. Quantification relative to tubulin was performed using ImageJ. B) Flow cytometry analysis of cell surface Env expression on Jurkat cells stably expressing control or Vps26 targeting shRNA following HIV-1 infection. C) Equal volumes of supernatants harvested from HIV-1 infected cells were used to infect reporter cells and HIV-1 infectivity was measured by luciferase assay (RLU, relative light units) (left panel) and Gag p24 was quantified by ELISA to measure the viral content of supernatants (right panel). D) Western blotting for Env (gp120 and gp41) and Gag (p24CA) in virions purified from Jurkat T cells expressing control or Vps26 shRNA. The band intensities from two western blots were quantified using ImageJ and the ratio of HIV-1 Env gp120 to Gag p24CA is shown. E) HIV-1 infected Jurkat cells were stained for Env (green) and Vps26 (red). DAPI is blue. Scale bar is 10 microns.